To examine SCMs after injury, we injured the tibialis anterior (TA) muscles by BaCl₂ injection (Figure 1A) and labeled the differentiating SCMs using an antibody against NCAM, a cell adhesion molecule highly expressed in these cells (Capkovic et al., 2008), and ghost fibers using an antibody against Laminin, a major component of the BM (Webster et al., 2016; Figure 1B). Since macrophages are present in the ghost fibers to clear the necrotic debris of the damaged myofibers (Collins et al., 2024), we also labeled macrophages with an antibody against MAC-2, a member of the lectin family expressed on the cell surface of macrophages (Hohsfield et al., 2022). At dpi 2.5, the differentiating SCMs and macrophages were two major cell populations residing within the ghost fibers, occupying 39.9% and 47.8% of the total volume, respectively. By dpi 3.5, SCMs filled 98.2% of the ghost fiber volume, whereas macrophages only accounted for 1.8% (Figure 1C), with most of the macrophages residing in the interstitial space outside of the ghost fibers (Figure 1B), which would account for the high overall number of macrophages in the regenerating muscle tissues in this time period (Collins et al., 2024). Of note, our confocal (Figure 1—figure supplement 1A and B and Videos 1 and 2) and transmission electron microscopy (TEM) (Figure 1—figure supplement 1C) analyses of regenerating TA muscles at dpi 3 also revealed narrow openings (~1 μm diameter) on the BM of the ghost fibers, through which macrophages (MAC-2⁺) with an ~20 μm diameter (as a round cell) were traversing, suggesting that macrophages enter and/or escape the ghost fibers by squeezing through tiny openings on the BM. By dpi 4.5, most of the SCMs have fused into multinucleated primary myofibers (Figure 1B; Collins et al., 2024).

Given that the Arp2/3 complex-mediated branched actin polymerization is required for myoblast fusion during mouse embryogenesis (Lu et al., 2024), we asked whether the Arp2/3 complex is required for skeletal muscle regeneration in adults. Toward this end, we generated satellite cell-specific, tamoxifen-inducible KO mice for ArpC2, a subunit of the Arp2/3 complex (Goley and Welch, 2006), by breeding Pax7^CreERT2 mice (Lepper et al., 2009) with Arpc2^fl/fl mice (Wang et al., 2016). The conditional knockout (cKO) mouse line Pax7^CreERT2; Arpc2^fl/fl will be referred to as Arpc2-cKO hereafter. The littermates of the Pax7^CreERT2 mice without the floxed Arpc2 allele were used as wild-type controls. To induce genetic deletion of Arpc2 in satellite cells, we performed intraperitoneal injection of tamoxifen to the control and mutant mice every two days over a period of ten days (Figure 2A). The Arpc2 KO in satellite cells was confirmed by western blot (Figure 2—figure supplement 1). Satellite cell-specific Arpc2 KO did not affect TA muscle weight and size in uninjured muscle (Figure 2—figure supplement 2). However, muscle injury by BaCl₂ resulted in a significant decrease (87.7 ± 2.0%) in the cross-sectional area (CSA) of regenerated myofibers in Arpc2-cKO mice compared to their littermate controls at dpi 14 (Figure 2B and C) and dpi 28 (Figure 2—figure supplement 3). Consistent with this, the frequency distribution of CSA displayed a significant shift toward the small size in the mutant mice (Figure 2D). Taken together, these data demonstrate that the Arp2/3-mediated branched actin polymerization is essential for skeletal muscle regeneration.

To pinpoint the specific step of skeletal muscle regeneration – satellite cell proliferation, differentiation, migration, and SCM fusion – in which branched actin polymerization is required, we performed immunostaining using antibodies that specifically mark these steps. As shown in Figure 3—figure supplement 1, the percentages of muscle cells positive for the proliferation marker (Ki67) and the muscle differentiation marker (MyoG) in the injured TA muscles were similar between control and Arpc2-cKO mice (Figure 3—figure supplement 1A–E). In addition, live imaging analysis showed that cultured Arpc2-cKO SCMs exhibited normal migration and cell–cell contact behaviors (Video 3). These results demonstrate that branched actin polymerization is dispensable for satellite cell proliferation, differentiation, and migration during skeletal muscle regeneration. Thus, the reduced muscle size in the Arpc2 mutant mice is likely due to defects in SCM fusion.

To monitor the SCM fusion phenotypes, we examined the regenerating TA muscles of the control and Arpc2-cKO mice at dpi 4.5, when myoblast fusion leading to primary myofiber formation is mostly completed (Figure 1B; Collins et al., 2024). While most of the SCMs within the ghost fibers had fused in the control animals, the ghost fibers in the Arpc2-cKO mice contained differentiated (NCAM⁺), but mostly unfused, SCMs, which were readily observed in cross-sections, comparable to those in the Mymx-cKO mice (Figure 3A–C, Figure 3—figure supplement 2A). Consistent with this, the frequency distribution of SCM numbers in a cross-section per ghost fiber in the Arpc2-cKO and Mymx-cKO mice displayed a dramatic shift toward higher numbers compared with that of the wild-type mice (Figure 3D). These results indicate that the actin cytoskeleton plays an essential role in SCM fusion as the fusogenic proteins. Interestingly, expression levels of the fusogenic proteins, MymK and MymX, in the TA muscle of Arpc2-cKO mutant mice were either similar to or higher compared to that of wild-type mice (Figure 3—figure supplement 1F–H), suggesting that the fusion defect in the Arpc2-cKO mutant mice was not due to a lack of fusogen expression. Consistent with this, cultured Arpc2-cKO SCMs exhibited a severe fusion defect despite undergoing normal differentiation (Figure 3E–G) and pharmacologically inhibiting Arp2/3 with CK666 also led to a similar fusion defect (Figure 3—figure supplement 2B–D). In addition, cell-mixing experiments using wild-type and Arpc2-cKO SCMs showed that Arpc2-cKO SCMs failed to fuse with wild-type cells (Figure 3H and I), indicating that branched actin polymerization is required in both fusing partners. Taken together, these results demonstrate that branched actin polymerization is required in SCMs for their fusion during skeletal muscle regeneration. The severe myoblast fusion defects observed in early stages of regeneration (e.g., dpi 4.5) provide a good explanation for the presence of thin muscle fibers in Arpc2-cKO mice at dpi 14 (Figure 2B and C) and dpi 28 (Figure 2—figure supplement 3). These thin muscle fibers could be either elongated mononucleated muscle cells or multinucleated myofibers each containing a small number of nuclei due to occasional fusion events (comparable to those in Mymx-cKO muscles) (Figure 2B and C; Figure 2—figure supplement 3). Whether Arp2/3 and branched actin polymerization may play a role in the growth and/or maintenance of post-fusion multinucleated myofibers requires future loss-of-function studies to inactivate Arpc2 using a myofiber-specific Cre driver.

Activation of the Arp2/3 complex requires the actin NPFs, including the WASP and WAVE family of proteins (Goley and Welch, 2006). To examine their potential functions in mammalian muscle regeneration, we generated single and double cKO mice for N-WASP [the WASP family member with high expression in SCMs Lu et al., 2024] and CYFIP1 [a subunit of the WAVE complex Eden et al., 2002], respectively. The cKO mouse line Pax7^CreERT2; Nwasp^fl/fl will be referred to as Nwasp-cKO, Pax7^CreERT2; Cyfip1^fl/fl as Cyfip1-cKO, and Pax7^CreERT2; Nwasp^fl/fl; Cyfip1^fl/fl as dcKO hereafter. Target protein knockouts in SCMs were confirmed by western blot (Figure 2—figure supplement 1).

For the single cKO mice, immunostaining revealed a moderate but significant reduction of TA myofiber CSA at dpi 14 by 48.2 ± 0.1% in Nwasp-cKO and 67.7 ± 3.3% in Cyfip1-cKO mice, respectively, compared to their littermate controls (Figure 2B and C). The myofiber CSA of dcKO mice further decreased to 80.9 ± 1.8%, comparable to the 87.7 ± 2.0% observed in the Arpc2-cKO mice (in which both N-WASP and WAVE complexes are defective) and to the 89.3 ± 1.9% in the Mymx-cKO mice (Figure 2B and C). Moreover, the Nwasp and Cyfip1 single KO mice exhibited moderate myoblast fusion defects at dpi 4.5, which were exacerbated in dcKO mice (Figure 3B–D), despite normal satellite cell proliferation and differentiation, as well as the persistent fusogenic protein expression in the dcKO mice (Figure 3—figure supplement 1A–H). Thus, our data revealed partially redundant functions between N-WASP and WAVE NPFs in promoting myoblast fusion during skeletal muscle regeneration.

To investigate the mechanism by which branched actin polymerization regulates SCM fusion during muscle regeneration, we first examined the cellular structure at the fusogenic synapse of cultured SCMs. Live cell imaging of cultured SCMs expressing Arp2-mNeongreen (mNG) and LifeAct-mScarleti (mScar) at day 2 in differentiation medium (DM) revealed Arp2- and F-actin-enriched finger-like protrusions projecting from the invading cells into their fusion partners (receiving cells) at the fusogenic synapse prior to cell membrane fusion (Figure 4A and A’ and Video 4). Consistent with this, TEM analysis of the TA muscle at dpi 3.5 in wild-type mice revealed finger-like protrusions projected by SCMs invading their neighboring cells (20.3 ± 16.5% of SCMs exhibited invasive protrusions, n=83 SCMs from 20 ghost fibers examined) (Figure 4B and C). The average length and width of the invasive finger-like protrusions were 422±200 nm and 121±73 nm, respectively (Figure 4—figure supplement 1, n=32 invasive protrusions examined). In contrast, muscle cells in Arpc2-cKO mice seldom projected invasive protrusions (Figure 4B and C, 0.5 ± 2.2% of SCMs exhibited invasive protrusions, n=147 SCMs from 20 ghost fibers examined), whereas protrusions in SCMs of Mymx-cKO mice appeared normal (Figure 4B and C, 24.1 ± 15.6% of SCMs exhibited invasive protrusions, n=93 SCMs from 20 ghost fibers examined; Figure 4—figure supplement 1, n=29 invasive protrusions examined). Therefore, branched actin polymerization, but not the fusogenic protein MymX, is required for invasive protrusion formation to promote myoblast fusion during adult muscle regeneration.

C57BL/6J (stock: 000664) and Pax7^CreERT2 (stock: 012476) (Lepper et al., 2009) mice were obtained from the Jackson Laboratory. The Arpc2^fl/fl (Wang et al., 2016), Nwasp^fl/fl (Cotta-de-Almeida et al., 2007), and Cyfip1^fl/fl (Habela et al., 2020) mice were previously described. The Mymx^fl/fl line (Bi et al., 2018) was generously provided by Dr. Eric N. Olson. The control and mutant male littermates were used in each cohort of experiments.

Tamoxifen (Sigma; T5648) was dissolved at 20 mg/ml in corn oil. 100 μl tamoxifen/corn oil solution was administered by intraperitoneal injection to 2-month-old male mice as schematized in the figures. To induce muscle injury, BaCl₂ (Sigma; 342920) was dissolved in sterile saline to a final concentration of 1.2%, aliquoted, and stored at −20°C. Mice were anesthetized by isoflurane inhalation, the legs were shaved and cleaned with alcohol, and TA muscles were injected with 50 μl of BaCl₂ with a 28-gauge needle.

Satellite cells were isolated from limb skeletal muscles of 2-month-old male mice. Briefly, muscles were minced and digested in 800 U/ml of type II collagenase (Worthington; LS004196) in F-10 Ham’s medium (Thermo Fisher Scientific; 11550043) containing 10% horse serum for 90 minutes at 37°C with rocking to dissociate muscle fibers and dissolve connective tissues. The dissociated myofiber fragments were collected by centrifugation and digested in 0.5 U/ml dispase (Gibco; 17105041) in F-10 Ham’s medium for 30 minutes at 37°C with rocking. Digestion was stopped with F-10 Ham’s medium containing 20% FBS. Cells were then filtered from debris, centrifuged, and resuspended in growth medium (GM: F-10 Ham’s medium supplemented with 20% FBS, 4 ng/ml FGF2, 1% penicillin–streptomycin and 10 mM HEPES). The cell suspension from each animal was pre-plated twice on the regular 100 mm tissue culture-treated dishes for 30 minutes at 37°C to eliminate fibroblasts. The supernatant containing mostly myoblasts was then transferred into collagen-coated dishes for culture in GM. To validate the KO efficiencies of the target genes, skeletal muscle from one to six mice of each genotype was pooled for satellite cell isolation. To induce myogenic differentiation, satellite cells were cultured in DM (DMEM supplemented with 2% horse serum, 1% penicillin-streptomycin, and 10 mM HEPES).

To pharmacologically inhibit branched actin polymerization in SCMs, the Arp2/3 inhibitor CK666 (50 μM) was added into the DM at day 0 of differentiation of wild-type SCMs. After 48 hours, the cells were fixed in 4% paraformaldehyde (PFA) and stained with anti-MHC and DAPI to assess their differentiation and fusion index.

To delete Arpc2 in SCMs in vitro, satellite cells isolated from Arpc2-cKO mice were cultured in GM supplemented with 2 μM 4-hydroxytamoxifen (Sigma; H6278) for 10 days. Subsequently, the cells were trypsinized and plated at 70% confluency in DM. After 48 hours, the cells were fixed in 4% PFA and stained with anti-MHC and DAPI to assess their differentiation and fusion index.

The cytosolic GFP, cytosolic mScarleti, LifeAct-mScarleti, and Arp2-mNeongreen constructs were described in the previous study (Lu et al., 2024), and assembled into the retroviral vector pMXs-Puro (Cell Biolabs; RTV-012) using the NEBuilder HiFi DNA Assembly Cloning Kit (NEB; E2621L). To package the retrovirus, 2 μg of retroviral plasmid DNA was transfected into platinum-E cells (Cell Biolabs; RV-101) using the FuGENE HD transfection reagent (Promega, E2311). Two days after transfection, the virus-containing medium was filtered and concentrated with Retro-X Concentrator (Clontech, PT5063-2) following the manufacturer’s protocol. The concentrated retroviruses were diluted in GM (with a 1:1000 dilution), mixed with polybrene (7 μg/ml), and used to infect cells. One day after infection, cells were washed with PBS and cultured in fresh GM.

To co-stain NCAM, MAC-2, and Laminin, the 4% PFA fixed TA muscles were dehydrated in 30% sucrose at 4°C overnight. The specimens were embedded in Tissue-Plus O.C.T. Compound (Fisher Scientific; 23-730-571) and 12 μm cryosections were collected onto Superfrost Plus Microscope Slides (Fisher Scientific; 12-550-15). Then, the cryosections were incubated with blocking buffer (PBS containing 2% BSA and 0.1% Triton X-100) for 20 minutes at room temperature (RT), followed by overnight incubation with rabbit anti-NCAM (1:200; Millipore; AB5032), rat anti-MAC-2 (1:200; Biolegend; 125401), and rat anti-Laminin-2 (1:500; Sigma; L0663) at 4°C. To stain for dystrophin, the freshly dissected TA muscles were snap frozen in Tissue-Plus O.C.T. Compound and 12 μm cryosections were collected onto Superfrost Plus Microscope Slides. Next, the sections were fixed in 4% PFA for 12 minutes at RT, washed three times with PBS, and incubated with blocking buffer for 20 minutes at RT, followed by overnight incubation with rabbit anti-dystrophin (1:200; Abcam; ab15277) at 4°C. To co-stain Pax7, MyoG, Laminin, and Ki67, the freshly dissected TA muscles were snap frozen in Tissue-Plus O.C.T. Compound and 12 μm cryosections were collected onto Superfrost Plus Microscope Slides. Then, the sections were fixed in 2% PFA for 5 minutes at RT, washed three times with PBS, and incubated with blocking buffer supplemented with M.O.M blocking reagent (1:25; Vector; MKB-2213-1) for 60 minutes at RT, followed by overnight incubation with mouse anti-Pax7 (1:2; DSHB; Pax7), mouse anti-MyoG (1:2; DSHB; F5D), rat anti-Laminin-2 (1:500; Sigma; L0663), and rat anti-Ki67 (1:500; Thermo Fisher Scientific; 14-5698-82) at 4°C. After the incubation with primary antibodies, the sections were extensively washed with PBS and then incubated with Alexa Fluor-conjugated secondary antibodies for 1 hour at RT. Subsequently, the sections were washed with PBS and subjected to imaging using a Leica TCS SP8 inverted microscope.

For western blots, proteins were isolated from the cultured SCMs or TA muscle using ice-cold RIPA buffer (150 mM NaCl, 1% NP40, 0.1% SDS and 50 mM Tris, pH 7.4) containing protease and phosphatase inhibitors (Cell Signaling Technologies; 5872) for 20 minutes. The supernatants were collected by centrifugation at 140,000 × g for 15 minutes. Protein concentrations were determined using the Bradford Protein Assay Kit (Bio-Rad; 5000201). 10–30 μg total protein was loaded for each sample and separated by 10% SDS-PAGE gel and transferred to PVDF membranes (Millipore; GVHP29325). Then, the membranes were blocked for 1 hour at RT in PBS containing 5% nonfat dry milk and 0.1% Tween-20 (PBSBT) and subsequently were incubated with primary antibodies diluted at 1:1000 in PBSBT overnight at 4°C. The membranes were then washed with PBST and incubated with appropriate HRP-conjugated secondary antibodies diluted in PBSBT for 1 hour at RT. After extensive washes with PBST, the membranes were developed with the ECL western blotting substrate (Thermo Fisher Scientific; 32209). The following primary antibodies were used: sheep anti-ESGP/MymX (1:1000; R&D Systems; AF4580), mouse anti-MymK (Zhang et al., 2020b) (1:1000), and rabbit anti-β-Tubulin (1:1000; Cell Signaling Technologies; 2146).

Time-lapse imaging of cells incubated in 5% CO₂ at 37°C was performed on a Nikon A1R confocal microscope with a Nikon Biostation CT. The satellite cells were seeded on fibronectin-coated cover glass (MATTEK; P35G-0-14C) and imaged using a 40× (0.4 NA) objective at indicated time points after switching from GM to DM. The cells were imaged at 2- or 5-minute intervals. After time-lapse imaging, ImageJ (NIH, 64-bit Java 1.8.0_172) was used to project the z-stacks in 2D, using maximum intensity projection, and the resulting 2D images were assembled into a time-lapse video.

To observe the invasive protrusions at the contact sites of SCMs during muscle regeneration in vivo, TA muscle at dpi 3.5 was fixed in a solution containing 3% PFA, 2% glutaraldehyde, 1% sucrose, 3 mM CaCl₂ in 0.1 M sodium cacodylate buffer (pH 7.4) overnight at 4°C. Samples were subsequently washed with 0.1 M cacodylate buffer containing 3% sucrose and 3 mM CaCl₂, and post-fixed with 1% osmium tetroxide in 0.1 M sodium cacodylate buffer for 1.5 hours on ice. The muscle samples were stained with 2% uranyl acetate, dehydrated, and embedded in EPON resin as previously described (Zhang and Chen, 2008). The embedded samples were then cut into 70-nm-thick sections using LEICA ultramicrotome (UC6) and collected on copper slot grids. These sections were post-stained with 2% uranyl acetate and Sato’s lead solution and examined using a JEOL 1400 transmission electron microscope.

Statistical significance was determined using a two-tailed Student’s t-test conducted using the GraphPad Prism 8 software. The sample sizes and number of replicates are indicated in the figure legends. All experiments were repeated in at least three independent biological replicates. The investigators were not blinded to allocation during the experiments and outcome assessment. No data were excluded from the analyses. For the in vivo studies, age-matched animals were randomly assigned to experimental and control groups. No statistical methods were used to predetermine sample sizes, but our sample sizes are similar to those reported in previous publications (Millay et al., 2014; Bi et al., 2018; Duan et al., 2018). Data distribution was assumed to be normal, but this was not formally tested.