Interplay between myofibers and pro-inflammatory macrophages controls muscle damage in mdx mice

Adult skeletal muscle completely regenerates after an injury thanks to the muscle stem cells (MuSCs) or satellite cells that sustain adult myogenesis. This process involves activation, proliferation, commitment into terminal myogenesis, differentiation and fusion to form new myofibers. The regeneration power of adult skeletal muscle is very high, and damaged muscle fully recovers its functions within a few weeks. While MuSCs are absolutely indispensable for skeletal muscle regeneration, surrounding cells also play important roles in this process, including endothelial cells, fibro–adipogenic progenitors (FAPs) and immune cells (Bentzinger et al., 2013).

Macrophages are particularly crucial for efficient muscle regeneration. As in other damaged tissues, the first macrophages to invade the injured area are pro-inflammatory damage-associated macrophages that mainly arise from circulating monocytes. Inflammatory damage-associated macrophages express pro-inflammatory effectors that stimulate MuSC proliferation (Saclier et al., 2013) while limiting FAP expansion (Juban et al., 2018; Lemos et al., 2015). Phagocytosis of debris contributes to the resolution of inflammation, which is crucial to start tissue repair and regeneration (Arnold et al., 2007). Resolution of inflammation is characterized by a progressive switch of the inflammatory profile exhibited by macrophages, which then become anti-inflammatory or recovery macrophages. This process is controlled by several molecular effectors, including the metabolic sensor AMP-activated protein kinase (AMPK) (Mounier et al., 2013), the IGF1 pathway (Tonkin et al., 2016), the p38 MAPK regulator MKP1 (also known as DUSP1) (Perdiguero et al., 2011), and the transcription factors C/EBPβ (also known as CEBPB) (Ruffell et al., 2009) and NFIX (Saclier et al., 2020). These anti-inflammatory macrophages exert a variety of effects on surrounding cells: they stimulate the last steps of myogenesis, including differentiation and fusion (Saclier et al., 2013), they activate FAPs to remodel the extracellular matrix (ECM) (Juban et al., 2018; Lemos et al., 2015) and they promote angiogenesis, which occurs concomitantly to myogenesis (Latroche et al., 2017).

Post-acute-injury skeletal muscle regeneration is a highly regulated process, during which the various stages described above must be well coordinated in space and time. The situation is drastically different in degenerating myopathies such as Duchenne muscular dystrophy (DMD), during which multiple injuries occur asynchronously throughout the muscle. Skeletal muscle in degenerative myopathies fails to regenerate because of these permanent and asynchronous cycles of degeneration and regeneration due to mutations in the dystrophin–glycoprotein complex, which ensures myofiber integrity (Lapidos et al., 2004). In these disorders, recurrent myofiber damage leads to chronic inflammation and the eventual establishment of fibrosis. Indeed, it has been shown that asynchronous injuries in normal muscle trigger a chronic inflammatory status in the muscle associated with a fibrotic signature, reminiscent of what is observed in degenerative myopathies (Dadgar et al., 2014).

Macrophages are associated with fibrosis in mdx muscle, the mouse model for DMD (Vidal et al., 2008). Inhibition of the pro-inflammatory NF-κB pathway is associated with a lower macrophage number and a better phenotype of the muscle (Acharyya et al., 2007). Similarly, preventing the entry of circulating monocytes into the muscle temporarily improves muscle histology and function (Liang et al., 2018; Mojumdar et al., 2014; Wehling et al., 2001). However this is detrimental over a longer term (Zhao et al., 2016), showing the pivotal role of macrophages during muscle regeneration and the necessity of targeting the right subset of cells to prevent disease progression. Recently, we have shown that fibrosis is associated with pro-inflammatory macrophages in an experimentally fibrotic mdx mouse model (Juban et al., 2018). Functionally, these macrophages stimulate matrix production by fibroblastic cells through the abnormal secretion of latent TGFβ, due to increased expression of the gene encoding latent-transforming growth factor beta-binding protein 4 (LTBP4), which is involved in latent TGFβ export into the ECM. Interestingly, pharmacological skewing of these macrophages towards an anti-inflammatory phenotype using an AMPK activator improves muscle phenotype and function (Juban et al., 2018).

The standard treatment for DMD patients consists of glucocorticoids (GCs), which have been shown to delay disease progression and ambulation loss (McDonald et al., 2018). However, chronic daily exposure to GCs is associated with adverse effects including obesity, metabolic changes and muscle atrophy (Schakman et al., 2013; Wood et al., 2015). Therefore, alternative therapeutic strategies are required. Hydrogen sulfide (H₂S) is an endogenously synthetized small gaseous signaling molecule that freely permeates across cell membranes (Wang, 2014). H₂S is involved in various biological processes, including inflammation. Alteration of H₂S levels has been associated with several diseases in human and animal models, including sepsis, diabetes and rheumatoid arthritis (Fagone et al., 2018). H₂S has emerged as a promoter of the resolution of inflammation through the regulation of macrophage inflammatory state (Sun et al., 2020; Wallace et al., 2012). Interestingly, H₂S-releasing molecules, such as NaHS, have been shown to reduce skeletal muscle atrophy in a mouse model of diabetes (Lu et al., 2020) and to decrease muscle fibrosis after contusion-induced injury (Zhao et al., 2020). Moreover, H₂S-releasing molecules have shown promising results in treatment of cardiovascular disorders and arthritis, with reduced adverse effects (Wallace et al., 2020, 2018).

In this study, we investigated the interaction between myofibers and macrophages in the context of degenerative myopathies. As the macrophage phenotype is modulated by their close environment, we tested the hypothesis that dystrophic muscle fibers may themselves alter macrophage status, using ex vivo co-culture experiments. Then, we analyzed whether the macrophage phenotype could be modulated in vivo using NaHS, an H₂S donor, to ameliorate the dystrophic muscle phenotype. Finally, we identified the metabolic sensor AMPK as an essential NaHS mediator in macrophages.

The effect of myofibers on macrophage polarization was investigated by co-culturing wild-type bone marrow-derived macrophages (BMDMs) with single fibers isolated from wild-type and dystrophic muscle (Fig. 1A). The presence of wild-type myofibers increased the expression of two out of four pro-inflammatory markers analyzed in resting macrophages as compared with expression in controls [23% increase for TNFα (also known as TNF), 37% increase for COX2 (also known as PTGS2); Fig. 1B,C]. However, in the presence of mdx myofibers, the number of BMDMs expressing pro-inflammatory markers was strongly increased for three out of four pro-inflammatory markers: TNFα, CCR2 and COX2 (45%, 30% and 58% increase as compared with controls, respectively; Fig. 1B,C). Importantly, this was not associated with a difference in mortality between wild-type and mdx myofibers after 3 days of culture (Fig. S1). No modification of the expression of the anti-inflammatory markers CD301 (CLEC10A) and ARG1 by BMDMs was observed when these cells were co-cultured with wild-type or mdx myofibers (Fig. 1B,C). These results show that mdx myofibers skew macrophages towards a pro-inflammatory phenotype. Thus, the inflammation in mdx muscles is not only sustained by the cycles of injury and regeneration, but also by the dystrophic myofibers themselves, highlighting the relevance of modulating inflammation as a therapeutic strategy.

In an attempt to modulate the macrophage inflammatory status, we treated 3-month-old mdx mice with NaHS, an H₂S donor. Because of the short half-life of H₂S in vivo (Calvert et al., 2010), NaHS was delivered intraperitoneally daily for 3 weeks (Fig. 2A). NaHS treatment induced an important decrease in the number of macrophages per myofiber in the tibialis anterior (TA) muscle (42% decrease; Fig. 2B) and in the diaphragm (56% decrease; Fig. 2C). More precisely, in TA muscle from NaHS-treated mice, a strong decrease in the number of macrophages expressing pro-inflammatory markers was observed [35% and 31% decrease for TNFα and iNOS (NOS2), respectively; Fig. 2D,E], together with a decrease of the fibrosis-associated ARG1 marker (35% decrease; Fig. 2D,E) and an increase of the anti-inflammatory marker CD206 (also known as MRC1; 61% increase; Fig. 2E). The expression of CCR2 and CD301 was not affected (Fig. S2). Similarly, the number of macrophages expressing pro-inflammatory markers was reduced in the diaphragm (23% and 31% decrease for TNFα and iNOS, respectively; Fig. 2F), although the expression of ARG1 and CD206 was not affected in this muscle (Fig. 2F). These results show that NaHS treatment specifically reduces the number of pro-inflammatory macrophages in the muscles of mdx mice.

To study the impact of pro-inflammatory macrophage reduction in dystrophic muscles, several myogenic parameters were analyzed after NaHS treatment. Immunostaining of TA transversal sections from treated mdx muscles did not show any modification of the number of satellite cells per fiber, as determined by Pax7 immunostaining (Fig. S3A), or of the proportion of newly formed myofibers, as determined by expression of embryonic myosin heavy chain (eMHC, also known as MYH3; Fig. S3B). However, NaHS treatment led to a decreased number of nuclei per myofiber in the TA (20–33% decrease depending on the section level; Fig. 3A,B; Fig. S3C), as well as in the diaphragm (32% decrease; Fig. 3C), suggesting a reduction in cell fusion. Importantly, this was confirmed by longitudinal analyses on mono-myofibers isolated from the TA, which showed a decrease in the number of nuclei per micrometer along the length of myofibers from NaHS-treated animals (8% decrease; Fig. 3D,E). Muscle regeneration after an injury is associated with myofiber branching (Pichavant and Pavlath, 2014), which is characterized by myofibers containing two or more cytoplasmically continuous strands. Mdx mice harbor an increased proportion of branched myofibers as compared with wild-type mice (Bockhold et al., 1998), likely due to the permanent cycles of injury and regeneration, which is correlated with a decrease in muscle force production (Chan et al., 2007). Interestingly, isolated myofibers from NaHS-treated mdx mice were characterized by a decrease in myofiber branching (16% decrease; Fig. 3F,G). Taken together, these results show that dampening of the pro-inflammatory status of macrophages is associated with a reduction of myofiber branching and of the number of myonuclei per myofiber, indicative of a reduction in muscle damage.

Next, we evaluated the consequences of dampening the macrophage pro-inflammatory status by NaHS treatment on dystrophic muscle features. Hematoxylin–Eosin (HE) staining of muscle transversal sections showed a strong decrease in the damaged area in NaHS-treated mice for both the TA (50–55% decrease depending on the section level; Fig. 4A,B; Fig. S4A) and the diaphragm (58% decrease; Fig. 4C). This was associated with a decrease in muscle fibrosis, as shown by a reduction in the area expressing collagen 1 (Coll1) in the TA (9–10% decrease depending on the section level; Fig. 4D,E; Fig. S4B) but not in the diaphragm (Fig. 4F). Finally, we observed that the mean cross-sectional area (CSA) of myofibers, which is correlated with muscle force production, was increased in both the TA (18–27% increase depending on the section level; Fig. 4G,H; Fig. S4C) and the diaphragm (25% increase; Fig. 4I). Specifically, a decreased proportion of the smallest fibers (28% decrease in fibers with CSA <500 μm² for both TA section levels; 23% decrease in fibers with CSA <250 μm² in the diaphragm; Fig. 4G–I; Fig. S4C) and a concomitant increased proportion of larger fibers (40–47% increase in fibers with CSA >5000 μm² depending on the TA section level; 155% increase in fibers with a CSA of 2000–2250 μm² in the diaphragm; Fig. 4G–I; Fig. S4C) were observed in NaHS-treated mdx mice.

Taken together, these results show that the reduction in pro-inflammatory macrophages upon NaHS treatment is associated with an amelioration of the dystrophic muscle phenotype.

The AMPK metabolic sensor plays a critical role in macrophage phenotypic shift during skeletal muscle regeneration (Mounier et al., 2013). Because AMPK has previously been shown to be activated by NaHS in a microglia cell line (Zhou et al., 2014), we asked whether NaHS-mediated pro-inflammatory macrophage dampening was mediated through AMPK activation. Treatment of wild-type BMDMs with NaHS increased the phosphorylation at Thr172 of the AMPK catalytic α subunit after 1 and 2 h (61% and 89% increase, respectively, as compared with the untreated control; Fig. 5A,B), as did treatment with the AMPK allosteric activator 991 (80% and 110% increase at 1 and 2 h, respectively, as compared with untreated control; Fig. 5A,B), indicating AMPK activation. Next, we investigated the functional role of AMPK activation in the NaHS-mediated skewing of macrophages towards an anti-inflammatory phenotype. To this aim, we used AMPKα1^−/− BMDMs, which harbor a knockout mutation of the gene encoding the AMPK α1 subunit (Prkaa1), the only catalytic subunit expressed in macrophages (Sag et al., 2008) (Fig. 5C). As expected, NaHS-induced AMPK activation was associated with a decreased expression of the pro-inflammatory markers TNFα and CCL3 (14% and 11% decrease, respectively; Fig. 5D), and an increased expression of the anti-inflammatory markers CD206 and CD163 (19% and 11% increase, respectively; Fig. 5E). These NaHS-mediated effects on macrophage inflammatory status were completely abrogated in AMPKα1^−/− BMDMs (Fig. 5D,E). Finally, we used an in vitro model of adult myogenesis where conditioned medium generated from BMDMs was transferred onto MuSCs to stimulate their fusion into multinucleated myotubes (Fig. 5F). Conditioned medium from NaHS-treated wild-type BMDMs increased MuSC fusion into myotubes (22% increase; Fig. 5G,H), whereas this was not observed with conditioned medium derived from NaHS-treated AMPKα1^−/− BMDMs (Fig. 5G,H). These results identify AMPK activation through phosphorylation of the α1 catalytic subunit as a critical downstream target of NaHS in macrophages.

In this study, we analyzed the interaction between myofibers and macrophages in the context of DMD. We showed that the dystrophic myofibers themselves skew resting macrophages towards a pro-inflammatory phenotype. As alterations in the secretome of myotubes generated by myoblasts isolated from mdx mice (Duguez et al., 2013) or DMD patients (Lecompte et al., 2017) have already been described, this suggests that in the context of degenerative myopathies it is not only the degeneration and regeneration events that lead to chronic inflammation – the myofibers themselves also contribute to maintenance of the pro-inflammatory environment that alters macrophage properties and favors the generation of a pro-fibrotic population. Although the molecular effectors secreted by mdx myofibers contributing to the sustained inflammation remain to be precisely characterized, Duguez et al. have shown an increased secretion of various proteins, including secreted protein acidic and rich in cysteine (SPARC), by myotubes derived from mdx myoblasts (Duguez et al., 2013). As in vitro treatment of peritoneal macrophages with recombinant SPARC has been shown to stimulate the expression of pro-inflammatory genes such as Tnfa (Toba et al., 2015), its increased secretion by mdx myofibers could contribute to the skewing of resting macrophages towards a pro-inflammatory phenotype and, therefore, to fibrosis establishment in DMD muscle. Indeed, we have previously shown that pro-inflammatory macrophages promote fibrosis in DMD muscle (Juban et al., 2018). Moreover, modulating the inflammatory status of this population towards an anti-inflammatory phenotype improves DMD muscle function (Juban et al., 2018). Therefore, identification of specific anti-inflammatory strategies appears to be a relevant therapeutic approach to alleviate muscle damage.

H₂S is a gasotransmitter that exhibits anti-inflammatory properties, in part through the regulation of macrophage inflammatory status (Sun et al., 2020; Wallace et al., 2012). H₂S donors have been widely used in vivo to treat various pathological conditions in animal models (Fagone et al., 2018). This includes use of the H₂S donor NaHS to reduce skeletal muscle atrophy in a mouse model of diabetes (Lu et al., 2020) and to decrease muscle fibrosis after contusion-induced injury (Zhao et al., 2020). Interestingly, various therapeutics, including zofenopril, have been shown to exert their effect through the indirect release of H₂S (Fagone et al., 2018; Wallace et al., 2018). Recently, clinical studies have shown the safety and efficiency of H₂S-releasing molecules in humans. Patients with congestive heart failure exhibit a deficit of H₂S in the blood. Treatment with the H₂S donor SG1002 attenuates the level of brain natriuretic peptide, a stress marker of cardiac function, thus reducing the severity of, and potentially preventing, heart failure (Polhemus et al., 2015). Similarly, ATB-346, an H₂S-releasing derivative of the nonsteroidal anti-inflammatory drug naproxen, reduces the severity of symptoms in osteoarthritis patients with reduced adverse effects as compared to naproxen alone (Wallace et al., 2018, 2020).

At the molecular level, the H₂S donor NaHS has been shown to favor microglia inflammatory shift towards an anti-inflammatory phenotype (Du et al., 2014; Zhou et al., 2014) through the activation of AMPK (Zhou et al., 2014). Interestingly, our previous work identified AMPK as a crucial activator of macrophage inflammatory shift during normal skeletal muscle regeneration (Mounier et al., 2013). In the present study, we identified AMPK as a critical mediator of NaHS-induced macrophage skewing that allowed the acquisition of a phenotype associated with muscle repair. Although the molecular cascade leading to AMPK activation by NaHS in macrophages remains to be established, the resolvin annexin A1 (ANXA1) may be involved in this process. Indeed, NaHS-induced downregulation of pro-inflammatory marker expression in BMDMs following lipopolysaccharide (LPS) stimulation requires the resolvin annexin A1 (Brancaleone et al., 2014), and we have recently shown that ANXA1 triggers AMPK activation in macrophages through the binding of its G-protein-coupled formyl peptide receptor, FPR2 (McArthur et al., 2020).

Thus, in an attempt to counteract the exacerbated macrophage pro-inflammatory phenotype induced by dystrophic myofibers, we treated mdx mice with NaHS. We showed that NaHS-treated mdx mice exhibited a decrease in the total number of macrophages in the TA and the diaphragm. More specifically, in the TA, we observed a reduction in the number of macrophages expressing pro-inflammatory markers as well as the ARG1 fibrotic marker, together with an increase in the anti-inflammatory CD206 marker, suggesting a skewing towards an anti-inflammatory phenotype and a reduction of the pro-fibrotic macrophage population (Juban et al., 2018). This was associated with a global muscle phenotype amelioration characterized by a reduced myofiber necrosis and reduced fibrosis, together with an increased myofiber size. In the diaphragm, NaHS treatment also induced a reduction of macrophages expressing pro-inflammatory markers, but no effect was observed on ARG1 and CD206 markers. This was associated with a decreased myofiber necrosis and an increased myofiber size, but no fibrosis reduction. These discrepancies observed between the TA and the diaphragm may be explained by the more severe phenotype exhibited by the diaphragm, notably the higher level of fibrosis that may require extended treatment duration in order to be reduced. Globally, these results are in accordance with the beneficial effect observed on dystrophic muscle function by preventing the entry of circulating monocytes at early time points (Liang et al., 2018; Mojumdar et al., 2014; Wehling et al., 2001) or by skewing macrophages towards an anti-inflammatory phenotype (Juban et al., 2018). These results confirm that the modulation of macrophage inflammatory status is a relevant therapeutic approach to relieve inflammation and muscle damage in degenerative myopathies.

The amelioration of the muscle phenotype by NaHS treatment was characterized by a decrease in the number of nuclei per myofiber together with a reduction of myofiber branching. Branched myofibers are barely observed in normal skeletal muscle, but they appear after regeneration following acute injury (Pichavant and Pavlath, 2014). They are observed in DMD patients (Bell and Conen, 1968) and in DMD fibers generated in vitro (Al Tanoury et al., 2021), as well as in the mdx mouse model (Head et al., 1992; Lefaucheur et al., 1995). Branched myofibers are characterized by an impairment of Ca²⁺ signaling and excitation–contraction coupling, causing myofiber function defects and higher fragility (Head, 2010; Lovering et al., 2009). Although the mechanisms resulting in myofiber branching are still unclear (and could involve a defect in final fusion between forming myofibers or a normal process that is exacerbated in dystrophic muscle due to continuous regeneration), a recent study of mdx extensor digitorum longus (EDL) muscle has shown that the increased number of myonuclei per myofiber is concomitant with myofiber branching (Massopust et al., 2020). Accordingly, the concomitant reduction of the number of nuclei per myofiber and of branching observed in NaHS-treated mice suggests an improvement in muscle homeostasis.

To conclude, in this study we have identified a novel property of dystrophic myofibers in the DMD mouse model mdx that is associated with a regulatory loop between myofibers and macrophages. Dystrophic myofibers skew macrophages towards a pro-inflammatory phenotype that in turn contributes to myofiber damage, consequently increasing myofiber branching and fragility, thus favoring muscle damage and fibrosis. Finally, this work has identified treatment with H₂S-releasing molecules as a potential therapeutic strategy to dampen inflammation in DMD and improve the dystrophic muscle phenotype through AMPK activation.

DMD^mdx4Cv (mdx) (Chapman et al., 1989) mice on C57BL/6J background and 129;B6-Prkaa1^tm (AMPKα1^−/−; Jørgensen et al., 2004) were bred and used according to French legislation. The protocols were approved by the Ethical committee from Université Claude Bernard Lyon 1 (CEEA55). Experiments were conducted on males at 10–12 weeks of age. NaHS (1 mg/kg; Sigma-Aldrich) or phosphate-buffered saline (PBS) were injected daily intraperitoneally in mdx mice for 3 weeks, and muscles were harvested the day after the last injection.

The fascia of TA muscles was removed, then the TA and diaphragm were frozen in nitrogen-chilled isopentane and kept at −80°C until use. 8 µm-thick cryosections were prepared for staining with HE (Sigma-Aldrich) and immunolabeling. For immunolabeling, cryosections were permeabilized for 10 min in 0.5% Triton X-100 and saturated in 2% bovine serum albumin (BSA) for 1 h at room temperature. For identification of regenerating myofibers, saturated cryosections were colabeled with primary antibodies directed against embryonic myosin (eMHC; sc-53091, Santa Cruz; 1:100) and laminin (L9393, Sigma-Aldrich; 1:200). For macrophage double immunolabeling, cryosections were labeled with antibodies (from Abcam unless indicated) against F4/80 (ADGRE1; ab6640 or ab74383; 1:400) overnight at 4°C and labeling using the second antibody was performed for 2 h at 37°C. Antibodies were directed against arginase 1 (sc-18355, Santa Cruz; 1:100), CCR2 (ab32144; 1:200), CD206 (sc-58987, Santa Cruz; 1:100), CD301 (ab59167; 1:200), collagen I (ab292, ab34710 or Biotech 131001; 1:200), COX2 (ab2372; 1:50), iNOS (ab15323; 1:25), laminin (ab78287 or L9393, Sigma-Aldrich; 1:200) and TNFα (ab34839; 1:50). Secondary antibodies were coupled to FITC, Cy3 or Cy5 (Jackson ImmunoResearch Inc.; 1:200). Muscle stem cells were labeled as previously described (Théret et al., 2017) using an antibody directed against Pax7 (AB_528428, Developmental Studies Hybridoma Bank; 1:10). HE–stained muscle sections were recorded with a Nikon E800 microscope at 20× magnification connected to a QIMAGING camera. Damaged areas were measured as previously described (Juban et al., 2018). Images of fluorescent immunolabeling were recorded with a DMI 6000 Leica microscope connected to a Coolsnap camera (Photometrics) at 20× magnification. For each condition of each experiment, at least 8–10 randomly chosen fields of view were counted. The number of labeled macrophages was calculated using ImageJ software (NIH, Bethesda, MD) and was expressed as a percentage of total macrophages. Areas of collagen1 were calculated using ImageJ software as previously described (Juban et al., 2018). Cross-sectional area (CSA) was determined on whole muscle sections labeled by anti-laminin antibody using the Open-CSAM program, as previously described (Desgeorges et al., 2019). For quantification in the TA of the number of nuclei per myofiber, damaged areas, fibrotic areas and CSA, two different cross sections were analyzed. The first and second cross sections were harvested 5 and 10 mm from the tendon base, respectively.

Macrophages were derived from murine bone marrow precursors as previously described (Mounier et al., 2013). Briefly, total bone marrow was obtained by flushing the femur and tibiae from wild-type C57BL/6J mice with Dulbecco's Modified Eagle's Medium (DMEM; Gibco). Cells were cultured in DMEM containing 20% heat-inactivated fetal bovine serum (FBS; Sigma-Aldrich), 30% of conditioned medium derived from the L929 cell line (European Collection of Authenticated Cell Cultures; enriched in CSF-1), 2.5 μg/ml of Fungizone (Gibco) and 100 U/ml of penicillin-streptomycin (Gibco) for 6–7 days. To generate conditioned medium, 110,000 cells/cm² were treated with 500 µM NaHS for 2 days. Cells were then washed twice with PBS and incubated for 24 h in DMEM F-12 medium (Gibco) without serum. Medium was collected, centrifuged for 10 min at 500 g and supplemented with 2% horse serum (HS; Gibco) before transfer onto MuSC cultures (see below).

54,000 cells/cm² were seeded on glass coverslips and treated with 500 µM of NaHS for 2 days. Cells were fixed for 10 min in 4% formaldehyde, permeabilized for 10 min in 0.5% Triton X-100 and saturated in 2% BSA for 1 h at room temperature. Cells were then labeled as described above using anti-TNFα (ab34839; 1:50), anti-CCL3 (sc-1383, Santa Cruz; 1:50), anti-CD206 (ab64693; 1:50) or anti-CD163 (bs-2527R, Bioss; 1:100) primary antibodies revealed by secondary antibodies coupled to FITC or Cy3 (Jackson ImmunoResearch Inc.; 1:200). After staining of nuclei using Hoechst 33342 (Sigma-Aldrich), cells were mounted in Fluoromount (Interchim), and 10-12 randomly chosen fields of view were imaged on a Nikon E800 microscope at 10× magnification connected to a QIMAGING camera.

Single myofibers of extensor digitorum longus (EDL) muscle were isolated from wild-type and mdx mice as previously described (Le Grand et al., 2012) and were co-cultured as 15 myofibers/well in 96-well plates containing 14,000 cells/cm² of wild-type BMDMs (prepared as described above) for 3 days. Macrophages were immunolabeled as described above for the immunofluorescence on muscle sections. For each condition of each experiment, ∼10–12 randomly chosen fields of view were recorded with a DMI 6000 Leica microscope connected to a Coolsnap camera at 20× magnification.

The fascia of TA muscles was removed, and muscles were fixed for 45 min in 4% formaldehyde. After two quick washes and a 30 min wash in PBS, muscles were transferred into PBS containing 0.5% Triton X-100. A total of 30–40 single myofibers were gently isolated by manual dissociation using tweezers, incubated with 1 µg/ml of Hoechst 33342 to label their nuclei, and mounted on slides. Slides were observed on an Axio Observer.Z1 (Zeiss) connected to a Coolsnap HQ2 CCD Camera (Photometrics), and the branched myofibers were numerated. For the quantification of the number of nuclei, 30–40 images corresponding to 7000–9000 nuclei from 6–10 randomly chosen single myofibers were acquired, and the number of nuclei was normalized to the myofiber length.

Muscle hindlimbs from adult males were dissociated and digested first for 45 min at 37°C under agitation in Ham's F-10 medium (Gibco) containing 10% HS (Gibco) and 300 U/ml collagenase II (Gibco), followed by a second digestion for 30 min at 37°C under agitation in Ham's F-10 medium (Gibco) containing 10% HS (Gibco), 150 U/ml collagenase II (Gibco) and 1.7 U/ml dispase (Gibco). MuSCs were isolated using a Satellite Cell Isolation Kit (Miltenyi Biotec) with LS Columns (Miltenyi Biotec) on a MidiMACS separator (Miltenyi Biotec), and were seeded on gelatin-coated flasks at 2000 cells/cm² in DMEM F-12 (Gibco) supplemented with 20% FBS (Sigma-Aldrich), 2% Ultroser G (Pall Gelman Sciences) and 100 U/ml of penicillin-streptomycin (Gibco).

MuSCs were seeded on Growth Factor Reduced Matrigel (Corning) at 10,000 cells/cm² in DMEM F-12 (Gibco) supplemented with 20% FBS (Sigma-Aldrich), 2% Ultroser G (Pall Gelman Sciences) and 100 U/ml of penicillin-streptomycin (Gibco) overnight, and culture medium was then replaced with macrophage-derived conditioned medium (see above). After 48 h, cells were fixed for 10 min in 4% formaldehyde, permeabilized for 10 min in 0.5% Triton X-100 and saturated in 2% BSA for 1 h at room temperature. They were then labeled and mounted, as described above, using an anti-desmin antibody (ab32362, Abcam; 1:200). Then, 10–12 randomly chosen field of view were imaged on a Nikon E800 microscope at 10× magnification connected to a QIMAGING camera.

Two million macrophages were seeded in 6-well plates. At 1 h before treatment, the culture medium was replaced with DMEM without Phenol Red (Gibco) supplemented with 10% charcoal-stripped FBS (Gibco), and cells were treated with 10 µM 991 (SpiroChem) or 500 µM NaHS (Sigma-Aldrich) for 1 or 2 h. Proteins were isolated in lysis buffer containing 50 mM Tris-HCl (pH 7.5), 1 mM EDTA, 1 mM EGTA, 0.27 M sucrose, 1% Triton X-100, 20 mM glycerol-2-phosphate disodium, 50 mM NaF, 0.5 mM PMSF, 1 mM benzamidine, 1 mM Na₃VO₄ and 1% phosphatase inhibitor cocktail 3 (Sigma-Aldrich; P0044) for 30 min on ice and centrifuged for 10 min at 16,100 g to remove debris. Then, 25 μg of protein was subjected to SDS–PAGE and transferred onto a nitrocellulose membrane, which was probed with antibodies against phosphorylated AMPKα (p-AMPKα; #2531, Cell Signaling Technology; 1:1000), AMPKα (#2532, Cell Signaling Technology; 1:1000) or β-actin (A5316, Sigma-Aldrich; 1:2000). Blots were analyzed on a Chemidoc imager (Bio-Rad) using SuperSignal West Femto Maximum Sensitivity Substrate (Thermo Fisher Scientific), and signal intensity was quantified using Image Lab Software (Bio-Rad).

All experiments were performed using at least three different cultures or animals in independent experiments. Statistical tests were performed using GraphPad Prism. Depending on the experimental setup, a two-tailed, unpaired Student's t-test, or a one-way or two-way ANOVA with Tukey's multiple comparison correction was used for statistical analyses.

The authors thank the ALECS-SPF and SCAR animal facilities (SFR Santé Lyon-Est) for breeding and housing mice.

The peer review history is available online at https://journals.biologists.com/jcs/article-lookup/doi/10.1242/jcs.258429