For experiments using human derivatives, materials, methods, ethical considerations and reasons for exemption from research subject consent were reviewed by the institutional review board of Chungnam National University Hospital, and exemption from review was approved on the condition that human derivatives were provided from a biobank (approval no. CNUH 2022-11-087; Daejeon, South Korea). All human plasma samples were provided by the biobank of Chungbuk National University Hospital. All human plasma samples were collected and stored in advance by the biobank with informed patient consent. Normal and patient plasma samples were obtained from individuals with disease codes Z00 (people without symptom complaints or reported diagnoses) and N18.5 (stage 5 chronic kidney disease), respectively, according to the Korean Standard Classification of Diseases (www.kcdcode.kr). The patient group was further adjusted to those with a history of diabetes. In addition, clinical information on the age, disease code, and additional diseases of the normal and patient groups was provided by the biobank. A total of 1 ml plasma was received from each of the 20 normal subjects (average age, 23.4 years; 10 men and 10 women) and 19 patients (average age, 64.8 years; 9 men and 10 women).

The Institutional Animal Care and Use Committee of Chungnam National University approved all animal management and experiment protocols (approval nos. 202305A-CNU-083 and 202309A-CNU-155; Daejeon, South Korea). All mice were bred and maintained in a controlled environment (free access to food and water, 12-h light/dark cycle, 50–60% humidity, 22°C ambient temperature). A total of 18 male C57BL/6 mice, 10 to 12 weeks old, were supplied by Narabiotech (Seoul, South Korea). A total of five 12-week-old male PHF20-overexpressing C57BL/6J mice that we had previously generated and maintained were used (17). At the end of the study, mice were euthanized by CO₂ gas inhalation at a CO₂ filling rate of 30% of the chamber volume/min. Before cardiac perfusion of all the mice, the mice were anesthetized with Avertin (200 mg/kg) and perfused first with PBS and then with 4% paraformaldehyde. To ensure sacrifice, euthanasia was performed using CO₂ at the aforementioned rate before specimen collection. The gastrocnemius, tibialis and soleus muscles of the mice were collected and used in the experiments. All mice experiments were conducted at animal facilities in accordance with institutional guidelines.

Human plasma stored at −80°C was rapidly thawed at room temperature, centrifuged at 4°C for 10 min at 2,000 × g, and the supernatant was transferred to a sterile 1.5-ml tube. Subsequently, 15 µl concentrated cfRNA was extracted using the Quick-cfRNA Serum & Plasma Kit (cat. no. 1059; Zymo Research Corp.) according to the manufacturer's instructions, and was immediately stored at −80°C.

Blood effects were minimized through cardiac perfusion before muscle collection. The gastrocnemius, soleus and TA muscles located on the hind limb of the mice were collected, and any fat tissue was carefully removed without further injuring the muscle tissues. For cfRNA isolation, fresh skeletal muscle tissues were collected from mice and immediately placed in a 3.5-cm dish containing Dulbecco's Modified Eagle's Medium (cat. no. LM001-06; Welgene, Inc.) supplemented with 10% exosome-depleted FBS, 4 mM L-glutamine and 5.5 mM D-glucose. The tissues were incubated in a humidified incubator at 37°C and 5% CO₂ for 24 h, allowing them to release muscle-derived cfRNA under conditions resembling the natural environment within the body. Subsequently, 15 µl concentrated cfRNA was extracted using the Quick-cfRNA Serum & Plasma Kit, according to the manufacturer's instructions. In addition, total RNA was extracted from the gastrocnemius and TA muscles using TRIzol according to the manufacturer's protocol.

cDNA synthesis was performed using SuperScript™ II Reverse Transcriptase (cat. no. 18064022; Invitrogen; Thermo Fisher Scientific, Inc.) according to the manufacturer's instructions. For cDNA synthesis, 300 ng human plasma-derived cfRNA, 500 ng mouse gastrocnemius and tibialis muscle-derived cfRNA, 300 ng mouse soleus muscle-derived cfRNA, and 1 µg mouse gastrocnemius and tibialis anterior (TA) muscle-derived total RNA samples were used. cDNA was diluted 1:50 for qPCR on a Bio-Rad CFX96 Real-Time PCR Detection System (cat. no. 3600037; Bio-Rad Laboratories, Inc.) using GoTaq^® qPCR Master Mix (cat. no. A6001; Promega Corporation). The expression levels of each mature mRNA and precursor of hsa-miRNA and mmu-miRNA were normalized to the expression levels of Gapdh for mature mRNAs, to U6 for hsa-miRNA and to small nucleolar RNA (snoRNA)202 for mmu-miRNAs. The qPCR thermal cycling conditions were as follows: Initial heat activation at 95°C for 2 min, followed by 40 cycles at 95°C for 15 sec and 60°C for 1 min. The Cq value was defined using Bio-Rad CFX Maestro 2.8 software (Bio-Rad Laboratories, Inc.). The relative RNA expression levels were quantified using the 2^−ΔΔCq method (ΔCq=Cq Target gene-Cq Normalization gene) (18). The primer sequences are listed in Table I.

Human skeletal muscle cells obtained at passage two (cat. no. CC-2561; Lonza Group, Ltd.) were cultured in SkBM Basal Medium (cat. no. CC-3161; Lonza Group, Ltd.) containing SkGM SingleQuots Supplements and Growth Factors (cat. no. CC-4139; Lonza Group, Ltd.). Exosome-depleted FBS (10%; cat. no. A2720801; Gibco; Thermo Fisher Scientific, Inc.) was added to the cell medium. The cells were cultured at 37°C in a humidified incubator containing 5% CO₂. Cells were subcultured at 80% confluence. Human skeletal muscle cells were differentiated when they reached 60% confluence for 5 days using Human Skeletal Muscle Cell Differentiation Medium (cat. no. 151D-250; Cell Applications, Inc.). The medium was changed every day during differentiation. The morphology of differentiated muscle cells was confirmed using an optical microscope. Cells were used for experiments for up to 5 passages. Mycoplasma negativity was confirmed before all cell experiments.

As described previously (19), after the completion of experiments for western blot analysis, cells and muscles from mice were placed on ice and proteins were extracted using PRO-PREP™ Protein Extraction Solution (cat. no. 17081; Intron Biotechnology, Inc.). The lysates were then centrifuged for 30 min, 20,000 × g at 4°C. The extracted proteins was quantified using the Bradford method and 25 µg protein was loaded per lane. Quantified proteins were separated by SDS-PAGE on 10.0% gels and were transferred to PVDF blotting membranes (cat. no. 10600023; Cytiva). Subsequently, the transferred membranes were blocked in 1X Tris-buffered saline [140 mM NaCl, 2.7 mM KCl and 250 mM Tris-HCl (pH 7.4)] containing 5% skimmed milk and 0.2% Tween-20 for 1 h at room temperature. Blocked membranes were then incubated with primary antibodies (dilution ratio, 1:1,000) overnight at 4°C, followed by incubation with secondary antibodies (dilution ratio, 1:10,000) for 1 h at room temperature. The following primary antibodies were used: Anti-MYOD (cat. no. sc-377460; Santa Cruz Biotechnology, Inc.), anti-PHF20 (cat. no. 3934S; Cell Signaling Technology, Inc.), anti-muscle RING-finger protein-1 (MuRF1; cat. no. sc-398608; Santa Cruz Biotechnology, Inc.), anti-tumor susceptibility gene 101 (TSG101; cat. no. sc-7964; Santa Cruz Biotechnology, Inc.), anti-calnexin (cat. no. 2679; Cell Signaling Technology, Inc.) and anti-GAPDH (cat. no. A19056; ABclonal Biotech Co., Ltd.). The following secondary antibodies were used: HRP-conjugated anti-rabbit (cat. no. 7074V; Cell Signaling Technology, Inc.) and HRP-conjugated anti-mouse (cat. no. 31430; Invitrogen; Thermo Fisher Scientific, Inc.). Protein expression was visualized using ProNA ECL (cat. no. TLP-112.1; TransLab, Co., Ltd.) according to the manufacturer's instructions. All protein expression levels were normalized to the expression levels of the loading control GAPDH. ImageJ software (version 1.49; National Institutes of Health) was used for semi-quantification of blotting.

Two 10-cm plates were used for exosome isolation. ExoQuick-TC™ (cat. no. EXOTC50A-1; System Biosciences, LLC) was used to isolate exosomes from the growth media, cultured for 24 h after media change, of pre-differentiation and post-differentiation skeletal muscle cells. Concentrated media (20 ml) were centrifuged at 3,000 × g for 15 min at room temperature to remove cells and cell debris. Subsequently, 4 ml ExoQuick-TC was added to the supernatant. After inverting at least five times, the samples were incubated at 4°C overnight (≥12 h), then centrifuged at 1,500 × g for 30 min at room temperature. The supernatant was discarded and the pellet was resuspended in TRIzol^® (cat. no. 15596026; Invitrogen; Thermo Fisher Scientific, Inc.) for precursor miRNA profile analysis. After resuspending the exosomes in TRIzol, they were immediately stored at −80°C before starting precursor miRNA profiling analysis.

The miRNeasy^® Serum/Plasma Kit (cat. no. 217184; Qiazen, Inc.) was used to prepare RNA samples. The RNA isolated from each sample was used to construct sequencing libraries with the SMARTer smRNA-Seq Kit for Illumina (Takara Bio, Inc.), following the manufacturer's protocol. Briefly, input RNA was first polyadenylated in order to provide a priming sequence for an oligo-(dT) primer. cDNA synthesis was primed by the 3′ smRNA dT Primer, which incorporated an adapter sequence at the 5′ end of each first-strand cDNA molecule. When the MMLV-derived PrimeScript™ Reverse Transcriptase (cat. no. 2680A; Takara Bio Inc.) reached the 5′ end of each RNA template, it added non-templated nucleotides, which are bound by the SMART smRNA Oligo-enhanced with locked nucleic acid (LNA) technology for greater sensitivity. In the template-switching step, PrimeScript RT used the SMART smRNA Oligo as a template for the addition of a second adapter sequence to the 3′ end of each first-strand cDNA molecule. In the next step, full-length Illumina adapters (including index sequences for sample multiplexing) were added during PCR amplification. The Forward PCR Primer bound to the sequence added by the SMART smRNA Oligo, while the Reverse PCR Primer bound to the sequence added by the 3′ smRNA dT Primer. Resulting library cDNA molecules included sequences required for clustering on an Illumina flow cell. The libraries were validated by checking the size, purity and concentration on the Agilent Bioanalyzer. The libraries were pooled in equimolar amounts, and sequenced on an Illumina HiSeq 2500 instrument. Image decomposition and quality values calculation were performed using the modules of the Illumina pipeline All exosomal precursors of miRNA sequencing procedures were performed by Macrogen, Inc. Sequence alignment and detection of known and novel precursors of miRNA were performed using the miRDeep2 software algorithm (Ver. 2.0.0.8; Max Delbrück Center). In the read alignment to precursor result of the Quantifier module result of miRDeep2, the read aligned to each precursor was selected. Among the selected reads, reads that were assigned to mature miRNA were excluded, and reads with a length of >50% of the precursor were selected. Reads per million (RPM) was calculated using the total sum of reads for each sample obtained from the selected results. All exosomal precursors of miRNA sequencing procedures were performed by Macrogen, Inc. From the analysis results of precursor miRNA sequencing, as a cutoff point, abundant precursor miRNAs >700 RPM were selected as biomarker candidates.

To measure hindlimb grip strength, mice at 12 weeks old were placed on a grip strength meter, while their tail and nape were held and pulled gently. Each mouse was allowed to perform the test five times and was given a 5-min break between each test. Values for grip strength were recorded and normalized by dividing by whole body weight.

To evaluate the relationship between the expression of genes in skeletal muscle, Gene Expression Profiling Interactive Analysis 2 (GEPIA2) (http://GEPIA2.cancer-pku.cn/index.html) (20) was used, and a Spearman correlation analysis was performed. Only human skeletal muscle tissue databases (n=396) from GTEx (21) were used for the analysis.

Gastrocnemius and tibialis muscle tissues from mice were fixed with 4% paraformaldehyde for 16 h at 4°C and 4-µm paraffin-embedded sections were generated. The paraffin-embedded sections were stained with hematoxylin (for 5 min at room temperature) and eosin (for 1 min at room temperature) according to standard protocols. Light microscopy was used for histological observation. Measurements of cross-sectional area (CSA) were performed using ImageJ software (version 1.49; National Institutes of Health). For each CSA, six different views were randomly selected for measurement.

The right hind limbs of 13 10-week-old male C57BL/6 mice supplied by Narabiotech (Seoul, South Korea) were used. As described in a previous study, surgical tape was wrapped around starting from the distal aspect of the foot to the ankle area. Subsequently, Velcro was wrapped around the hindlimb, starting at the distal end of the foot. Velcro was replaced if adverse effects (e.g., skin injury and edema) or loose Velcro was observed. The forepaw and left hindlimb were free, so that the mice had free access to food and water.

Using Human TargetScan (Ver 8.0; http://www.targetscan.org/vert_80/) and mouse TargetScan (Ver 8.0; http://www.targetscan.org/mmu_80/), target mRNAs for microRNAs were scrutinized.

To transfer miR-6516 to TA muscles of immobilization-injured mice, mimic-miR-6516 (cat. no. SMM-003; 20 nM; Bioneer Corporation) was mixed with RNAiMAX (cat. no. 13778030, Invitrogen; Thermo Fisher Scientific, Inc.), as described previously (22). For the control group, mimic-miR-control (cat. no. SMC-2003; Bioneer Corporation) was used and a mixture was prepared in the same manner as the mimic-miR-6516 mixture. The resulting mixture (50 µl) was injected on the 5th and 10th day during the 14-day immobilization period, and the Velcro was replaced after injection.

Data are expressed as the mean and standard error of the mean (SEM) from at least three separate experiments. GraphPad Prism (version 8.1.1; Dotmatics) was used for statistical analysis. Quantitative data are presented as the mean ± SEM unless indicated otherwise. Comparisons between two groups were evaluated using unpaired Student's t-test or Mann-Whitney U test, depending on the results of the Shapiro-Wilk test. For multiple comparisons, one-way ANOVA was performed, followed by Dunnett's post hoc test. P<0.05 was considered to indicate a statistically significant difference.

It has previously been shown that miRNA expression levels are not always proportional to those of precursor miRNAs (14), suggesting that the changes in miRNA levels observed as muscular atrophy develops may differ from the changes in precursor miRNA levels. To profile the skeletal muscle-derived miRNA precursors secreted into the plasma (possible biomarkers of muscle atrophy), the exosomal miRNA precursors (cfRNAs) of human muscle cells were first sequenced. Muscle cell differentiation was confirmed by the gross changes in fused cell morphology and increased expression of MYOD (Fig. 1A and B). Exosomes were extracted using ExoQuick, and western blotting detected the positive exosomal marker TSG101 but not the negative marker calnexin (Fig. 1C). Exosomal precursor miRNA sequencing results confirmed that precursor miRNAs were abundant in exosomes derived from pre- and post-differentiated muscle cells (Table II). Of such precursor miRNAs, any that overlapped with other genes or were difficult to analyze via qPCR were excluded. The precursor of miR-206, known to be a muscle-only miRNA (23), and the precursor of miR-133b, which is downregulated in the plasma of patients with sarcopenia (24), were analyzed in the present study. Previous studies have shown that some snoRNAs also function as precursors of miRNAs (25–28). Precursors of miRNAs that overlapped with snoRNAs that were not identified as producing miRNAs were excluded from the analysis, even if they did not overlap in sequence with a specific mRNA. Another study reported that miR-1291 and miR-3651 are derived from snoRA2C and snoRD84, respectively, and these snoRNAs are very similar in terms of their sequences to the precursors of the corresponding miRNAs (29). For this reason, the precursors of miR-1291 and miR-3651 were not excluded from the analysis, despite their significant sequence overlap with snoRA2C and snoRD84. In addition, the precursor of miR-4449 was excluded from the analysis, as primer design and RT-qPCR were difficult to perform given the 86% G+C content. In summary, in human plasma-derived cfRNAs, the precursors of miR-206, miR-12136, miR-1291, miR-664a, miR-3651 and miR-133b were selected for analysis. Given the low number of patients diagnosed with sarcopenia, patients with diseases likely to cause sarcopenia as a complication were included. Previous studies have revealed that the incidence of sarcopenia is high in patients with diabetes and chronic kidney disease (30,31). Therefore, the biomarker candidates were quantitatively analyzed in the plasma of patients with either of these diseases, and of normal people aged 20–30 years. Of the biomarker candidates, the expression levels of precursors of miR-206 and miR-664a were significantly increased in the patient group with diabetes and chronic kidney disease (Fig. 1D). By contrast, there were no significant differences in the expression levels of the precursors of miR-133b, miR-1291, miR-3651 and miR-12136 between patients and control individuals (Fig. 1E). These results suggested that increased expression of the precursors of miR-206 and miR-664a in circulating plasma cfRNAs may be associated with sarcopenia, which is characterized by muscle atrophy.

The patients analyzed in the present study were expected to develop muscle atrophy as a complication of diabetes or chronic kidney disease; however, atrophy was not definitively confirmed. Therefore, the present study analyzed muscle-derived cfRNAs in a mouse model of muscular atrophy. Induction of atrophy was confirmed in mice overexpressing PHF20. In a previous study, PHF20 exerted a negative effect on muscle differentiation via positive regulation of YY1 (17). Subsequently, correlations between PHF20 levels and those of gene products associated with muscle differentiation, namely MYH2, TNNT1 and TNNT3, were assessed (32). The relationships between the levels of the muscle atrophy markers MuRF1 (TRIM63) and atrogin-1 (FBXO32), and their upstream transcription factor FOXO3a (33,34), were also investigated. These studies confirmed the negative association between the levels of PHF20 and those of muscle differentiation genes, and the positive association between the levels of PHF20 and muscle atrophy genes. The correlation pattern between PHF20 and each marker gene was similar to the correlation pattern between YY1 and each marker gene (Fig. 2A). The correlation of genes was analyzed using the dataset from GTEx. These findings indicated the possibility that PHF20 overexpression may inhibit muscle differentiation and facilitate muscle atrophy. A previous study confirmed that overexpression of PHF20 inhibits muscle differentiation and triggers defects in muscle morphology in vivo (17). This suggests that overexpression of PHF20 may induce muscle atrophy; additional experiments were thus performed in the present study to assess this. In PHF20 TG mice, the expression levels of both PHF20 and the muscle atrophy marker MuRF1 were increased (Fig. 2B), as were the mRNA expression levels of Phf20 and Trim63, which encode PHF20 and MuRF1 (Fig. 2C). The grip strength test was used to evaluate the muscle strength of PHF20 TG mice (35), and the strength was significantly lower than that of wild-type mice (Fig. 2D). Additionally, the muscle CSA was significantly reduced in PHF20 TG mice compared with that in wild-type mice (Fig. 2E). These results provided clear evidence that overexpression of PHF20 induced muscular atrophy in vivo.

Muscle atrophy attributable to muscle disuse is a common clinical problem and one of the major causes of muscle atrophy (36,37). Therefore, to identify candidate biomarkers of muscle atrophy not only in the PHF20-overexpressing mouse model but also in a mouse model of muscle atrophy induced by muscle disuse (immobilization), muscle atrophy was induced in wild-type mice via Velcro immobilization. In previous studies, it was confirmed that muscle atrophy was induced when the hind limb was immobilized using Velcro for 2 weeks (38,39). Therefore, the right hind limbs of 10-week-old male wild-type mice were immobilized for 2 weeks to induce muscle atrophy (Fig. 3A). Notably, the muscle CSA level of the hind leg that was fixed with Velcro was decreased compared with that in non-treated wild type mice (Fig. 3B). Skeletal muscle atrophy causes biochemical and physiological changes in the muscles, leading to alterations in gene expression (40). These characteristics suggest that atrophied muscles in PHF20-overexpressing mice and Velcro-fixed mice may synthesize cfRNAs, the levels of which may differ from those of non-fixed wild-type mice; therefore, further analysis of muscle atrophy biomarker candidates was performed to identify changes in expression levels.

As aforementioned, to obtain cfRNAs from atrophied muscles, the right hind limbs of five 10-week-old male wild-type mice were fixed with Velcro for 2 weeks to induce muscle atrophy; subsequently, the gastrocnemius, soleus and TA muscles were dissected, and cfRNAs were extracted from each muscle. Similarly, the gastrocnemius, soleus and TA muscles of 12-week-old male PHF20 TG mice were prepared, and cfRNAs were extracted from each muscle. For the same reasons presented when selecting precursor miRNAs to be analyzed in human plasma, the precursors of miR-206, miR-6516, miR-1291, miR-23a and miR-664 were selected for mouse analysis. In mice, miR-664 is not divided into miR-664a and miR-664b; therefore, the precursor of miR-664 was included in the analysis. In the immobilized mice, the expression levels of the precursor of miR-206 were increased in cfRNAs derived from the gastrocnemius, soleus and TA muscles compared with wild type mice; also the expression levels of the precursor of miR-6516 were decreased (Fig. 4A-C). In addition, the expression levels of the precursors of miR-1291 and miR-23a were decreased in the TA muscle-derived cfRNAs from immobilized-mice compared with those from wild-type mice (Fig. 4B). In the PHF20 TG mice, the expression levels of the precursor of miR-206 were increased in cfRNAs derived from the TA muscles compared with those from wild-type mice (Fig. 4B); also the levels of the precursor of miR-6516 decreased (Fig. 4B). In addition, the expression levels of the precursors of miR-1291 and miR-23a were decreased in gastrocnemius muscle-derived cfRNAs from PHF20 TG mice compared with those in wild-type mice (Fig. 4A). The expression levels of the precursors of miR-1291, miR-23a and miR-664 were also decreased in the TA muscles of PHF20 TG mice compared with those in wild-type mice (Fig. 4B). It has previously been reported that miR-206 is a skeletal muscle-specific miRNA (23). Thus, as a mature miRNA is cleaved from a precursor of that miRNA, the precursor of miR-206 is also muscle-specific. Taken together, these experiments indicated that the expression levels of the precursor of miR-206 may increase in mouse muscle-derived cfRNAs during muscle atrophy and in plasma cfRNAs obtained from patients at risk of sarcopenia.

Notably, a decrease in the expression levels of the miR-6516 precursor was observed in all muscle-derived cfRNAs of Velcro-immobilized mice with muscle disuse atrophy. Based on these results, additional experiments were performed to determine whether miR-6516 administration in muscles affected muscle disuse-induced atrophy. During the 14-day period of muscle atrophy induced via Velcro immobilization of the right hind limb, a mimic-miR-6516 was injected into the TA muscle on days 5 and 10, and the mice were sacrificed on day 14 (Fig. 5A). It was confirmed that the muscle CSA of the mice administered mimic-miR-6516 was significantly greater than that of mice administered the mimic-miR-control, indicating that miR-6516 inhibits the progression of muscle atrophy caused by muscle disuse (Fig. 5B). The genes targeted by miR-6516 when inhibiting muscle atrophy were subsequently investigated. Using TargetScan human (Ver. 8.0) and TargetScan mouse (Ver. 8.0), genes that were predicted to be common targets of miR-6516-3p or −5p in humans and mice were identified, according to a cumulative weight context score of ≤-0.4. Additionally, since muscle atrophy was suppressed upon injection of mimic-miR-6516, only genes positively correlated with the muscle atrophy markers MuRF1 (TRIM63), Atrogin1 (FBXO32) and FOXO3a (FOXO3a) were selected as possible target genes. Only CDKN1B (p27^kip1) and ubiquitin specific peptidase 25 (USP25; USP25) satisfied the aforementioned conditions, and they were predicted to be target genes of miR-6516 in the context of inhibition of muscle atrophy (Table III). The mimic-miR-6516 was injected to determine whether this affected the expression levels of Usp25 and Cdkn1b. In mice injected with mimic-miR-6516, the expression levels of Cdkn1b and Usp25 were significantly reduced (Fig. 5C). A previous study showed that p27^kip1 levels negatively regulate skeletal muscle satellite cell proliferation and muscle regeneration (41). Additionally, a decrease in muscle fiber diameter has been confirmed in mice overexpressing Cdkn1b (42). As expected, the expression of early muscle regeneration markers Pax7 and Myod1 was increased in muscles administered mimic-miR-6516 (Fig. 5D). However, additional investigations are required because no association between suppression of muscle atrophy and reduction of USP25 expression was conclusively shown (data not shown). Taken together, these data suggested that injection of mimic-miR-6516 during immobilization inhibits muscle rest-induced muscle atrophy by suppressing expression of Cdkn1b and promoting skeletal muscle satellite cell proliferation and muscle regeneration (Fig. 6).