Skeletal muscle releases tiny particles with unique proteins and microRNAs that act within the muscle environment

To characterize EVs secreted from both human and mouse skeletal muscle cells, we first isolated EVs from mouse C2C12 myoblasts and myotubes, and human induced pluripotent stem cell (hiPSC)-derived myocytes (hiPSC-myocytes) by a standard ultracentrifugation protocol (25). C2C12 myoblasts were differentiated into myotubes (Fig. S1A and B) and incubated for 48 h in a differentiation medium containing EV-free horse serum (HS) before isolating EVs from the conditioned medium. To isolate C2C12 myoblast EVs, the cells were incubated for 48 h in an EV-free FBS medium. In addition, we used two lines of hiPSC-myocytes, 414C2^tet-MyoD and 409B2^tet-MyoD. These hiPSC lines harbor tetracycline-inducible human MYOD1 expressing piggyBac vector, and thus adding doxycycline (Dox) into culture medium induces MyoD1 expression, initiating myogenic differentiation. After 5 to 6 days after Dox addition, these hiPSCs differentiated into myocytes (Fig. S1C) expressing skeletal muscle cell marker proteins, including myosin heavy chain (MyHC), myogenin, and caveolin-3 but no longer expressing the iPSC marker proteins Nanog, OCT-4A, and Sox 2 (Fig. S1D). After differentiation, hiPSC-myocytes were incubated in a medium supplemented with EV-free HS for 48 h to isolate EVs from a conditioned medium. To observe the morphology of isolated EVs, we first performed transmission electron microscopy (TEM) analysis on EVs from C2C12 myoblasts (C2C12-MB-EVs), C2C12 myotubes (C2C12-MT-EVs), and hiPSC-myocytes (hiPS-MC-EVs). Figure 1A shows typical images of C2C12-MB-EVs, C2C12-MT-EVs, and hiPS-MC-EVs. The diameters of C2C12-MB-EVs and C2C12-MT-EVs were 48.8 ± 16.4 nm and 61.4 ± 22.4 nm, respectively (Fig. 1B). C2C12-MT-EVs were statistically larger than C2C12-MB-EVs. The average size of hiPS-MC-EVs was approximately 58 nm in both 414C2^tet-MyoD and 409B2^tet-MyoD lines (Fig. 1B). The sizes are all within the range of typical exosomes. Together, these results showed that both human and mouse skeletal muscle cells release small EVs with similar size.

We next examined the presence of the exosome marker proteins in the isolated EVs. The results show that C2C12-MB-EVs and C2C12-MT-EVs contained the well-defined exosome markers, including Alix, TSG101, CD81, and HSP90 (Fig. 1C). hiPS-MC-EVs also contained these exosome markers (Fig. 1D). In addition to these typical markers, we found that C2C12-MT-EVs and hiPS-MC-EVs but not C2C12-MB-EVs, contained caveolin-3, a protein highly expressed in skeletal muscle and cardiomyocytes, and its contents increased by differentiation. The results suggest that skeletal muscle cells release EVs harboring skeletal muscle-specific proteins.

To determine proteomic profiling of EVs released by skeletal muscle cells, we first performed quantitative shotgun proteomic analyses on C2C12-MB-EVs and C2C12-MT-EVs isolated by ultracentrifugation (Table S1). The analyses identified 894 and 933 proteins in C2C12-MB-EVs and C2C12 MT-EVs, respectively, which cover 983 different proteins (Fig. 2A). Previously, Forterre et al. identified 455 proteins as those found in EVs secreted from C2C12 myoblasts and myotubes (15). Of the 455 proteins, 354 proteins (78%) were also found in our results (Fig. S2A). The current results thus revealed 629 additional C2C12-MB/MT-EVs proteins not identified previously.

Next, the same proteomic analysis was performed on hiPS-MC-EVs from both 409B2^tet-MyoD and 414C2^tet-Myo-D lines (Table S2), and 624 and 628 proteins were detected in those from 409B2^tet-MyoD and 414C2^tet-Myo-D, respectively, showing over 98.5% of EV proteins overlaps (Fig. S2B). Among the 619 proteins identified in hiPS-MC-EVs, the 586 proteins (95%) are covered by the EV database Vesiclepedia (26) (http://microvesicles.org↗), validating isolated EVs of quality (Fig. 2B). Of the 619 proteins, 78% (481 proteins) were also found in EVs isolated from either C2C12-MB-EVs or C2C12-MT-EVs (Fig. 2A). The results also identified 435 proteins that overlap among C2C12-MB-EVs, C2C12-MT-EVs, and hiPS-MC-EVs. Thirty-six proteins were found in both C2C12-MT-EVs and hiPSM-EVs but not in C2C12-MB-EVs (Table S3), suggesting a distinct protein profile of myotube-derived EVs. On the other hand, 40 proteins were found only in C2C12-MB-EVs (Table S4). We next compared our results with published results which reported the identification of 954 proteins secreted from differentiating human myoblast through either classical secretory pathway or unconventional mechanisms that include EVs (14). Of the 954 proteins, 238 proteins (25%), and 340 proteins (36%) were found in our hiPS-MC-EVs (Fig. S2C) and C2C12-MB/MT-EVs (Fig. S2D), respectively. Our proteomics data along with others suggest that human and mouse SkM-EVs exhibit similar protein profiles.

To annotate identified EV proteins, we classified these proteins based on Gene Ontology (GO) using an integrative platform, DAVID (27, 28). The results showed that in hiPS-MC-EVs, C2C12-MB-EVs, and C2C12-MT-EVs, proteins belonging to the term “Extracellular Exosome” in “Cellular Components” were highly enriched, confirming that isolated EVs are of good quality (Fig. 2C). Consistent with previous studies (14), among the established EV/exosome markers tetraspanins, CD9, CD63, and CD81 (4, 7, 10), our proteomics identified only CD81 in both C2C12-MB/MT-EVs (Table S1) and hiPS-MC-EVs (Table S2). For the “Biological processes” term, proteins classified into “Muscle contraction” were significantly enriched in all three EV samples, which indicates that SkM-EVs contain proteins unique to skeletal muscle. Together, all these results suggest that SkM-EVs display a distinct protein signature.

We next sought to identify potential marker proteins that mark EVs released from skeletal muscle cells. To this end, we searched proteins highly expressed in skeletal muscle from our proteome data obtained from hiPS-MC-EVs and C2C12-MT-EVs. As mentioned above, we identified 36 potential MT-EV proteins (Fig. 2A, Table S3). To assess their specificity, we searched specific proteins using the Gene Ontology Consortium's Community Annotation Wiki for Muscle Biology (http://wiki.geneontology.org/index.php/Muscle_Biology↗) and confirmed that many of these proteins, including Nebulin, KLHL41, MYH1, TRIM72, ACTA1, and MYBPH are predominantly expressed in skeletal muscle. Furthermore, based on The Human Protein Atlas and The Genotype-Tissue Expression (GTEx) databases, we selected 10 proteins as potential marker proteins for SkM-EVs (Fig. 3A). To confirm whether these proteins are included in EVs, C2C12-MT-EVs were isolated by three methods, standard ultracentrifugation (UC-EVs), the Tim4-based method, which is based on the high-affinity binding of Tim4 to phosphatidylserine (PS) (PS⁺ EVs) (29), and immuno-isolation with anti-CD81 antibody (CD81⁺ EVs), and subjected to immunoblot analysis. Due to the availability and/or validity of antibodies, six out of ten proteins were analyzed. The results show that in addition to the typical exosome marker proteins (Alix, CD81, and Flotillin-1), C2C12-MT-EVs contain the skeletal muscle proteins, ATP2A1, β-enolase, calsequestrin 2, caveolin-3, and desmin (Fig. 3B), validating our proteomic analysis. None of the EVs isolated by the three methods contained detectable levels of calnexin (an ER marker), GM130 (a Golgi marker), and Tom20 (a mitochondrial marker). Intriguingly, CD81⁺ EVs contained much less ATP2A1 compared to UC-EVs and PS⁺ EVs but were highly enriched with calsequestrin 1, suggesting that CD81⁺ EVs show a distinct protein profile. Among the skeletal muscle proteins, ATP2A1, β-enolase, and desmin are predominantly expressed in skeletal muscle tissues (Fig. S3).

Recent reports showed that apart from plasma, the interstitium of tissues such as the liver and lung contain significant amounts of EVs (30, 31). To determine whether the skeletal muscle cells release EVs in vivo, we isolated EVs from both plasma and skeletal muscle (tibialis anterior, gastrocnemius, soleus, and quadriceps) interstitium of mice using the Tim4-based method (Fig. 4A). We validated the quality of isolated EVs by TEM and confirmed that PS⁺ EVs from both the plasma and SkM-interstitium show similar morphology (Fig. 4B). Plasma and SkM-interstitium PS⁺ EVs were similar in size, ranging from 30 to 150 nm (Fig. 4C). Scanning electron microscopy (SEM) analysis showed that the skeletal muscle interstitium contains EV-like vesicles with a diameter of 50 to 500 nm, which are attached to extracellular matrix (ECM)-like structures (Fig. 4D). We next asked whether plasma and SkM-interstitium EVs contain SkM-EV markers identified above. We isolated EVs from plasma and SkM-interstitium using PS-affinity or CD81-affinity beads used to isolate C2C12-MT-EVs (Fig. 4A). To validate and compare the two methods, we first examined whether plasma and interstitial EVs isolated by these methods were contaminated with non-EV membranous structures. The results showed that non-EV proteins (calnexin, GM130, and Tom20) were undetectable in isolated EVs (Fig. 4E). We also examined the presence of lipoprotein markers in these fractions. The results showed that apolipoprotein A-I (apoA-I), a major apolipoprotein of high-density lipoprotein, was detected in plasma CD81⁺ EVs but not in the interstitial CD81⁺ EVs (Fig. S4A). ApoA-I was also found in plasma and interstitial PS⁺ EVs (Fig. S4B). ApoE was hardly detected in both plasma and interstitial EVs isolated by either method. Although whether apoA-I found in EV fractions are contamination or an EV component cannot be determined at present (32–35), we conclude that both methods can efficiently isolate EVs from plasma and interstitium. As expected, the typical EV markers Alix and CD81 were found in both plasma and SkM-interstitium EVs regardless of the isolation methods (Fig. 4E). In addition, potential SkM-EV marker proteins (ATP2A1, β-enolase, calsequestrin 1, calsequestrin 2, caveolin-3, and desmin) were detected at a different degree in PS⁺ and CD81⁺ EVs isolated from SkM-interstitium. In contrast, the SkM-EV markers were much less or undetectable in the plasma EVs. ATP2A1, calsequestrin 2, and desmin were slightly detected in plasma EVs, suggesting that SkM-EVs only partly enter the bloodstream. Calsequestrin 2-positive EVs found in the plasma could be derived from the heart where its expression is much higher than in skeletal muscle (Fig. S3). Altogether, our results indicate that ATP2A1, β-enolase, and desmin may serve as reliable SkM-EV protein markers in vivo and that SkM-EVs are highly concentrated in skeletal muscle tissues but are not major populations in the circulation.

To further determine the physiological importance of SkM-EVs in vivo, we examined the effect of exercise on SkM-EVs. Whether exercise increases circulating EV contents is currently controversial (21–23, 36–38). Moreover, even though exercise increases circulating EVs, their origin(s) is not fully characterized. We took advantage of our newly identified SkM-EV marker proteins to clarify this issue. After mice were subjected to exhaustive endurance running on a treadmill (Fig. S5A and B), we immediately harvested blood from the heart and skeletal muscle tissues from a hind limb, and prepared plasma EVs and SkM-interstitium EVs, respectively, using PS-affinity beads. The results showed that the single bout of acute exercise does not alter the concentration of proteins in either plasma EVs (Ctrl, 340.3 ± 22.5 mg/mL; Exercise, 328.4 ± 8.0 mg/mL; P = 0.30) or SkM-interstitium EVs (Ctrl, 371.4 ± 19.5 mg/mL; Exercise, 378.7 ± 23.6 mg/mL; P = 0.35). We also assessed levels of marker proteins in plasma and the interstitium EVs. Neither typical EV markers nor SkM-EV markers in plasma and interstitium EVs were significantly changed by the exercise, while trends in the increase by the exercise of the typical EV markers CD81 (1.4-fold; P = 0.25) and Alix (1.4-fold; P = 0.22) in plasma EVs and CD81 (2.0-fold; P = 0.17) in interstitium EVs were observed (Fig. 5A and B). CD81 and Alix expression decreased in the muscle after the exercise through an unknown mechanism (Fig. S5C). On the other hand, the exercise did not influence the expression of SkM-EV marker proteins in skeletal muscle (Fig. S5C). These results suggest that the acute exercise does not influence EV release from skeletal muscle. Meanwhile, we observed positive correlations between CD81 and β-enolase, caveolin-3, or ATP2A1 contained in the interstitium EVs (Fig. 5C, Table S5). It could be consistent with previous studies that skeletal muscle cells preferentially release CD81-positive EVs (12, 14). Furthermore, we noticed subtle but significant increases in the size of the interstitium EVs but not of plasma EVs upon exercise (Fig. 5D and E). Collectively, our results show that SkM-EVs are mainly accumulated in the interstitium and that they do not account for the major proportion of circulating EVs. Our findings also suggest that exercise does not promote the release of EVs from skeletal muscle.

EVs are characterized as the vehicle for miRNAs. Therefore, we finally investigated myomiR profiles of SkM-interstitium EVs and plasma EVs. The results show that all the four miRNAs (miRs-1, -206, -431, and -486) abundantly expressed in the muscle are markedly concentrated in the interstitium EVs (Fig. 6A). In particular, miR-1 and miR-206 in the interstitium EVs were 45- and 20-fold higher than those in plasma EVs, respectively, confirmaing the intramuscular accumulation of SkM-EV detected by our protein-based analysis. Together, these results demonstrate that SkM-interstitium EVs display unique protein and miRNA profiles that are distinct from plasma EVs. MyomiRs play important roles in skeletal muscle homeostasis, including the regulation of myogenesis by targeting the paired box transcription factor Pax7, a master regulator for myogenesis (39–41). Our results led us to hypothesize that SkM-EVs predominantly play their roles within the intramuscular microenvironment. To test this, we asked whether SkM-interstitium EVs isolated from mice modulate the expression of genes involved in myogenesis. Figure 6B shows that C2C12 myoblasts uptake the interstitium EVs, suggesting that SkM-EVs function in these cells. We next determined mRNA levels that regulate myoblast differentiation. The results show that the interstitium EVs suppress Pax7 expression but increase MyHC expression, a marker for myoblast differentiation (Fig. 6C). Although the downregulation of Pax7 by interstitium EVs was not as robust as overexpression of myomiRs in myoblasts, our results were largely consistent with previous studies showing the repression of Pax7 mRNA expression by myomiRs (39–41). The current results thus suggest that SkM-interstitium EVs regulate myogenesis at least in part by suppressing Pax7 expression within the muscle microenvironment (Fig. 6D).

Fetal bovine serum (FBS) and horse serum (HS) were obtained from Gibco. FBS and HS were heat-inactivated before use. EV-free FBS and HS were prepared as described (25). Briefly, FBS and HS were spun at 2,000 g for 10 min followed by centrifugation at 100,000 g for 70 min. The supernatant was further centrifuged at 100,000 g for 16 h. The supernatant was filtrated with a 0.20 μm filter (Advantec) and used as EV-free FBS or HS. EV-free FBS and HS were stored at −80°C until use.

C2C12 mouse myoblasts (obtained from ATCC) were maintained at low cell density in growth medium (DMEM supplemented with 10% FBS). For differentiation to myotubes, C2C12 myoblasts were seeded into a 6-well plate at a density of 1.5 × 10⁵ cells per well and grown for 2 days in a growth medium. Afterward, cells were incubated for 4 days in a differentiation medium (DMEM supplemented with 2% HS). The medium was changed every other day. For isolation of EVs, cells were incubated in 2% EV-free HS for 48 h. Human iPS cell (hiPSC) lines, 414C2^tet-Myo-D and 409B2^tet-MyoD were maintained in StemFit AK02N (Ajinomoto) as described (49). These hiPSCs were differentiated into myocytes by a published protocol (49). Briefly, on day 0, hiPSCs were seeded into a Matrigel-coated 6-well plate at a density of 3 to 4 × 10⁵ cells/well and grown overnight in StemFit medium with 10 μM Y-27632. On day 1, the medium was switched to Primate ES Cell Medium (Reprocell) containing 10 μM Y-27632. On day 2, cells were incubated in Primate ES Cell Medium containing 1 μg/mL doxycycline (Dox) to induce MyoD1 expression. On day 3, the medium was changed to αMEM containing 5% KnockOut Serum Replacement (Gibco) and Dox (1 μg/mL) and incubated for 2 to 3 days. After differentiation, hiPSC-myocytes were incubated in DMEM containing 2% EV-free HS for 48 h to isolate EVs.

All protocols for animal procedures were approved by the Animal Care and Use Committee of the University of Tokyo, which are based on the Law for the Humane Treatment and Management of Animals (Law No. 105, 1 October 1973, as amended on 1 June 2020). C57BL/6 J male mice at 8 weeks old were obtained from Japan Clea. Mice were housed in a 12 h-light/12 h-dark schedule at 23 ± 2°C and 55 ± 10% humidity and fed ad libitum with a standard chow diet (Labo MR Stock, Nosan Corporation) and water. Mice at 9 to 10 weeks old were randomly assigned to either exercise or sedentary groups. After mice were adapted to the treadmill (5 m/min for 10 min per day) for 4 days, they were subjected to exhaustion running for up to 90 min using a ramped treadmill exercise protocol starting at 10 m/min and increasing by 2 m/min every 10 min (22) using a treadmill (MK-680C, Muromachi Kikai). Mice were defined as the exhausted state when they stopped running on a treadmill for more than 5 s despite gentle encouragement. Immediately after exercise, blood was collected by cardiac puncture under anesthesia with isoflurane. Afterward, mice were perfused through the left ventricle with PBS for 2 min at a rate of 1 mL/min to remove blood from the tissue, and skeletal muscle (tibialis anterior, gastrocnemius, soleus, and quadriceps) and other tissues were then harvested.

We used three methods to isolate EVs. Method I: EVs were isolated by ultracentrifugation according to a method previously described (25). Briefly, conditioned media (typically 12 mL from 6 wells) where cells were incubated in EV-free medium for 48 h was spun sequentially at 300 g for 10 min, 2,000 g for 10 min, and 10,000 g for 30 min. After each centrifugation step, the supernatant was transferred to a new centrifuge tube. The 10,000 g-supernatant was filtered through a 0.20 μm filter (Advantec) to obtain small EVs. Afterward, the supernatant was ultracentrifuged at 100,000 g for 70 min at 4°C using an MLA-55 rotor (Beckman Coulter) and an Optima MAX-TL Ultracentrifuge (Beckman Coulter). Pellet was washed once with PBS (2 mL/tube) and EVs were pelleted by ultracentrifugation at 100,000 g for 70 min at 4°C again. Resulting pellet was resuspended in 150 μL of PBS. Method II: The 10,000 g-supernatant was prepared as described above. After filtration with a 0.20 μm filter and concentration with Amicon Ultra-15 (Merck), EVs were isolated using by MagCapture Exosome Isolation Kit PS (Fujifilm-Wako) according to the manufacture instruction. This method is based on the ability of Tim4 protein to bind phosphatidylserine (PS), which localizes on the exosome surface (29). In brief, medium concentrated (1 mL) as above was mixed with 0.6 mg of streptavidin magnetic beads bound to 1 μg of biotinylated mouse Tim4- Fc and incubated in the presence of 2 mM CaCl ₂ for 16 to 18 h with rotation at 4°C. After washing beads three times with 1 mL of washing buffer (20 mM Tris-HCl pH 7.4,150 mM NaCl, 2 mM CaCl₂, and 0.0005% Tween20), EVs were eluted twice with 50 μL of elution buffer (20 mM Tris-HCl pH 7.4,150 mM NaCl, and 2 mM EDTA). Method III: The 10,000 g-supernatant prepared was filtrated and concentrated as Method II. Afterward, EVs were isolated using Exosome Isolation Kit CD81, mouse (Miltenyi Biotec) according to the manufacture's instruction. In brief, medium concentrated (1 mL) was mixed with CD81-binding magnetic beads and incubated for 1 h with rotation at room temperature. Magnetically labeled EVs were loaded onto a μ Column placed in the magnetic field of MACS Separator (Miltenyi Biotec). After washing the column, EVs were eluted with Isolation Buffer (100 μL/sample). EV protein contents were determined by Micro BCA Protein Assay Kit (Thermo Fisher). EVs were stored at −80°C until use.

Plasma EVs were isolated by MagCapture Exosome Isolation Kit PS (Fujifilm-Wako) and Exosome Isolation Kit CD81, mouse (Miltenyi Biotec) as described above. Plasma (300 μL) from a mouse was mixed with PBS (600 μL) and spun at 10,000 g for 30 min. After filtration of the supernatant with a 0.20 μm filter, plasma was subjected to the isolation of EVs using MagCapture Exosome Isolation Kit PS or Exosome Isolation Kit CD81, mouse with elution volume of 100 μL per 300 μL plasma. Skeletal muscle interstitium EVs were isolated according to a method recently reported (30, 50). Skeletal muscle tissues (tibialis anterior, gastrocnemius, soleus, and quadriceps) from hind limbs of a mouse (approximately 300 mg in total) were combined and digested with collagenase (10 mg/mL, Sigma) and dispase II (10,000 PU/mL, Wako-Fujifilm) for 1 h at 37°C in HEPES buffer (100 mM HEPES and 2.5 mM CaCl₂) immediately (within 1 h) after harvesting, to avoid excessive cell death. To avoid disruption of cells, tissues were minced gently. Afterward, one volume of PBS containing 2 mM EDTA was added to the sample, and the sample was passed through a 100 μm cell strainer (Corning). Samples were then centrifuged at 600 g for 5 min at 4°C, 2,000 g for 10 min, and 10,000 g for 30 min. The supernatant was filtrated with a 0.20 μm filter and concentrated using Amicon Ultra-15. EVs were then isolated by MagCapture Exosome Isolation Kit PS or Exosome Isolation Kit CD81, mouse as described above. EVs were eluted with 100 μL of elution buffer per 300 mg tissue.

Cells were lysed with urea buffer (8 M Urea, 50 mM Na-phosphate pH 8.0, 10 mM Tris-HCl pH 8.0, and 100 mM NaCl) containing protease inhibitor cocktail (Nacalai Tesque) as described (51). Tissue homogenates were prepared in radioimmunoprecipitation assay buffer (50 mM Tris-HCl pH 7.4, 150 mM NaCl, 1 mM EDTA, 1% Nonidet P-40, and 0.25% sodium deoxycholate) supplemented with a protease inhibitor mixture (Nacalai Tesque) and phosphatase inhibitor mixture (Sigma). Protein concentration was determined by BCA Protein Assay (Thermo Fisher). Cell lysate, tissue homogenate, and EVs were mixed with Laemmli buffer and heated at 95°C for 3 min. Aliquots were subjected to SDS-PAGE and immunoblot analysis according to a standard protocol. The expression of a protein was analyzed by Image J software or Evolution-Capt software (Vilber Lourmat). Antibodies used were obtained from commercial sources as follows: anti-caveolin 3 (sc-5310), anti-calsequestrin 1 (sc-137080), anti-calsequestrin 2 (sc-390999), anti-CD81 (sc-166029), anti-β enolase (sc-100811), anti-HSP90 (sc-13119), anti-Calnexin (sc-46669), anti-TSG101 (sc-7964) antibodies from Santa Cruz Biotechnology; anti-flotillin 1 antibody (ab41927) from Abcam; anti-CD81 (#10037), anti-Alix (#92880), anti-desmin (#5332), anti-Tom20 (#42406), anti-ATP2A1 (#12293), anti-GAPDH (#5174), HRP-linked anti-mouse IgG (#7076), and HRP-linked anti-rabbit IgG (#7074) antibodies from Cell Signaling Technology; anti-apolipoprotein A-I rabbit mAb (14227–1-AP) from Proteintech; anti-apolipoprotein E mouse mAb (NB110-60531) from Novus Biologicals; anti-GM130 (610823) from BD Bioscience; anti-myosin heavy chain, clone # MF20 (MAB4470) from R&D Systems; Mouse TrueBlot: Anti-Mouse Ig HRP (18–8817–31) from Rockland Immunochemicals.

EVs were solubilized in 50 mM Tris-HCl pH 9.0 containing 5% sodium deoxycholate, reduced with 10 mM dithiothreitol for 60 min at 37°C, and alkylated with 55 mM iodoacetamide for 30 min in the dark at 25°C. The reduced and alkylated samples were diluted 10-fold with 50 mM Tris-HCl pH 9.0 and digested with trypsin at 37°C for 16 h (trypsin-to-protein ratio of 1:20 (w/w)). An equal volume of ethyl acetate was added to each sample solution and the mixtures were acidified with the final concentration of 0.5% trifluoroacetic acid. The mixtures were shaken for 1 min and centrifuged at 15,700 g for 2 min. The aqueous phase was collected and desalted with C18-StageTips. LC-MS/MS analysis was performed using an UltiMate 3000 Nano LC system (Thermo Fisher Scientific) coupled to Orbitrap Fusion Lumos hybrid quadrupole-Orbitrap mass spectrometer (Thermo Fisher Scientific) with a nano-electrospray ionization source. The sample was injected by an autosampler and enriched on a C18 reverse-phase trap column (100 μm × 5 mm length, Thermo Fisher Scientific) at a flow rate of 4 μL/min. The sample was subsequently separated by a C18 reverse-phase column (75 μm × 150 mm length, Nikkyo Technos) at a flow rate of 300 nL/min with a linear gradient from 2% to 35% mobile phase B (95% acetonitrile and 0.1% formic acid). The peptides were ionized using nano-electrospray ionization in positive ion mode. The raw data were analyzed by Mascot Distiller v2.3 (Matrix Science), and peak lists were created based on the recorded fragmentation spectra. Peptides and proteins were identified by Mascot v2.3 (Matrix Science) using UniProt database with a precursor mass tolerance of 10 ppm, a fragment ion mass tolerance of 0.01 Da and strict trypsin specificity allowing for up to 1 missed cleavage. The carbamidomethylation of cysteine and the oxidation of methionine were allowed as variable modification.

Specimens for transmission electron microscopy (TEM) were prepared at room temperature. An aliquot of EV sample was pipetted onto a copper grid with carbon support film and incubated for 10 min. After the excess liquid was removed, a grid was briefly placed on 10 μL 2% uranyl acetate (w/v, Merck). Images were acquired under a JEM-1010 electron microscope (JEOL) operated at 100 kV with a Keen view CCD camera (Olympus Soft Imaging Solution). The size of EVs was measured using Image J software.

For scanning electron microscopy (SEM) analysis, skeletal muscle tissue (approximately 3 × 3 mm in size) was fixed with 10% neutral buffered formalin for 1 h and with 0.2% glutaraldehyde and 2% paraformaldehyde in PBS for 1 h. After post-fixation with 1% osmium tetroxide in PBS, samples were dehydrated in ethanol series (70%, 90%, 95%, 99.5%, and 100%) for 10 min each, treated with tert-butyl alcohol for 10  min twice and freeze-dried. The dried specimen was applied onto a carbon double side-tape with silver paste and sputter coated with platinum palladium. Images were acquired under a Hitachi S-4800 scanning electron microscope with a secondary electron in-lens detector.

EVs were labeled using ExoSparkler Exosome Protein Labeling Kit-Red (Dojindo Laboratories) according to the manufacturer's instruction. C2C12 myoblasts seeded in a 35-mm film bottom dish (Matsunami) were incubated without or with the labeled EVs (4 μg protein per 2 mL). Cells were fixed with 4% paraformaldehyde (Fujifilm-Wako) for 10 min and then permeabilized with 0.1% Triton X-100 in PBS for 5 min at room temperature. After nuclei were stained with DAPI, specimens were mounted with ProLong Gold Antifade Reagent (Thermo Fisher). Cell images were acquired by an LSM800 confocal laser microscope (Carl Zeiss) equipped with a Plan-Apochromat 63×/1.4 objective. Images were processed with a Zen software (Carl Zeiss).

C2C12 myotubes and hiPSC-myocytes (both differentiated in a 35-mm film bottom dish) were fixed and permeabilized as above. After blocking with 5% FBS in PBS for 1 h, specimens were incubated with anti-MHC antibody (1:100 dilution, R&D Systems) for 1 h followed by incaubation with Alexa Fluor 568-conjugated anti-mouse IgG (1:800 dilution, Thermo Fisher) for 45 min. Nuclei were stained with DAPI, and specimens were mounted as above. Cell images were obtained by an LSM800 confocal laser microscope (Carl Zeiss).

Total RNA was isolated using ISOGEN (NIPPON GENE), according to the manufacturer's instructions. The high-capacity cDNA reverse transcription kit (Applied Biosystems) was used to synthesize cDNA from total RNA. Quantitative real-time PCR (qPCR) analyses were performed using an Applied Biosystems StepOnePlus. mRNA levels were normalized to 18S ribosomal RNA levels. The primers used for qPCR analysis are described in SI Appendix, Table S6.

Plasma and SkM-interstitium EVs were isolated from two mice and pooled for miRNA analyses. Total RNA was extracted from EVs (20 μg protein) using miRNeasy Mini Kit (Qiagen). RNA was then reverse-transcribed using TaqMan MicroRNA Reverse Transcription Kit (Applied Biosystems) according to the manufacturer's protocol. qPCR was then performed using TaqMan MicroRNA Assay (miR-1, Assay ID: 002222; miR-206, Assay ID: 000510; miR-431, Assay ID: 001979; miR-486, Assay ID: 002093; miR-16, Assay ID: 000391; miR-21, and Assay ID: 000397) (Applied Biosystems). Exosomal miRNA levels were normalized by the mean value of miR-16 and miR-21 as described (31).

Results are expressed as mean ± SEM from at least three independent biological replicates. Statistical analyses were performed using the two-tailed, unpaired Student's t-test. P values less than 0.05 were considered statistically significant.