Six-week-old male C57BL/6 mice were purchased from Jicui Yaokang Biological Technology Co., (Nanjing, China). All mice were maintained under standard conditions of 22°C ± 2°C, 50%–60% relative humidity, and alternate dark/light cycles. Mice were fed a high-fat diet (HFD) (60 kcal%, D12492, Research Diets) for 10 weeks to induce obesity. Control mice were fed a basal diet (10 kcal%, D12450B, Research Diets). Both groups of mice were sacrificed at 16 weeks of age, and their gastrocnemius muscles were dissected for experimental analyses. Skeletal muscle triglyceride (TG) contents were measured using a Triglyceride Quantification Kit (Nanjing Jiancheng, China). All experiments in this study were authorized by the Ethics Committee of Jiangnan University [NO: JN. No20211030c0700625 (426)].

C2C12 mouse myoblast cells were purchased from the ATCC (CRL-1772, United States). The C2C12 cell line was cultured in growth medium (GM) supplemented with basic DMEM (11995–065, Gibco, United States), 10% FBS (1009–141, Gibco, United States), and 1% penicillin plus 1% streptomycin (SV30010, Hyclone, United States) at 37°C and 5% CO₂. GM was replaced with differentiation medium (DM) containing DMEM with 2% horse serum (HS, 16050–130, Gibco, United States) and incubated for 4 days in order to induce C2C12 cell differentiation when C2C12 cells reached 90% confluency.

To explore the effect of Apol9a on C2C12 cells, small interfering RNA (siRNA, GenePharma, Shanghai, China) was used to knockdown intracellular Apol9a, and the overexpression plasmid (GENEWIZ) was used to overexpress Apol9a. C2C12 cells were seeded into 6-well plates and cultured for 8 h (equivalent to a cell density of 30%–40%). Next, siRNA and the overexpression plasmid were transfected into cells at a concentration of 50 nM. The jetPRIME^® transfection reagent (114–15, Polyplus transfection) was used to transfer siRNAs, the overexpression plasmid, and the empty control plasmid. At 12 h after transfection, the transfection medium was replaced with DM to induce differentiation for 4 days. Specific knockdown of Apol9a was validated using two different siRNAs. The two siRNA sequences for mouse Apol9a was 5′-UUGUAUCCAAGGCCAAGUUGUTT-3′and 5′-AGC​CCU​UGA​GCA​GCA​CAU​GAA​TT-3′. A random siRNA sequence (A06001, GenePharma, shanghai, China) was applied as control. To inhibit the ERK1/2 signaling pathway, C2C12 myoblasts were treated with 50 μm PD98059 (HY-12028, Med Chem Express) in DM for 4 d (after 8 h transfection). Dimethyl sulfoxide (DMSO, D8418, Sigma-Aldrich) was used as a solvent control.

Myotube diameter measurements were obtained using ImageJ software. The diameter was measured at the widest region of each myotube, and the mean diameter size was compared between conditions.

Total RNA was isolated from both C2C12 myoblasts and skeletal muscle tissue by using the RNA extraction kit (K101, JN. BIOTOOLS, Wuxi, China). cDNA was synthesized using the BTS I 1st Strand cDNA Synthesis Kit (K102, JN. BIOTOOLS, Wuxi, China), and 1 μg of total RNA to be used for quantitative real-time PCR (qPCR) was reverse transcribed. qPCR was performed using the Power SYBR Green Master Mix kit (4367659, Invitrogen, United States) in a CFX 96^TMRealTime PCR Detection System (Bio-Rad, United States). The amplification conditions for qPCR were as follows: 94°C for 5 min, followed by 45 cycles of 94°C for 15 s and 45 s at 59°C. The fold change of gene expression was analyzed according to the 2^−ΔΔCT method. β-actin was utilized as the internal control for normalization. The sequences of all primers used in this research are presented in Supplementary Table S1.

Total proteins were extracted from the cell lysate and determined using the Quick BCA Protein Assay Kit (SYW3-1, Solarbio, China). The lysate proteins were resolved through SDS-PAGE with a 10% gel, transferred to PVDF membranes (IPVH00005, Merck Millipore), blocked, and incubated with primary antibodies against Apol9a (AC-15–1076, Ango Biotechnology, 1:1000); myosin heavy chain (MyHC) (MF-20, 1:1000, Developmental Studies Hybridoma Bank); myogenin (MyoG) (ab 1835, Abcam, 1:2000); myoblast determination protein (MyoD) (18943-1-AP, 1:1000, Proteintech); Myf5 (BD-PT2930, 1:1000, Biodragon); phospho-Erk1/2 (4370, 1:1000, CST); Erk1/2 (4695, 1:1000, CST); phospho-JNK (4668, 1:1000, CST); JNK (9252, 1:1000, CST); phospho-p38 (4511, 1:1000, CST); p38 (9212, 1:1000, CST); and β-actin (ab8227, 1:2000, Abcam). After all membranes were washed, they were treated with specific secondary antibodies (AS014, AS003, Abclonal, China). Specific protein bands were detected using an ECL kit (WBKLS0100, Merck Millipore). The images were observed using the Bio-Rad ChemiDoc MP Imaging System, and then, image bands were quantified through densitometry by using ImageJ plus software. The expression level of each target protein was normalized to that of β-actin, which was used as an internal control.

C2C12 myoblasts were fixed with paraformaldehyde (4%) for 20 min and treated with 1% Triton X-100 for 5 min. The fixed cell samples were incubated with blocking buffer (A8010, Solarbio, China) in phosphate-buffered saline (PBS) and then incubated with anti-MyHC (MF20, 1:50, DSHB) antibody at 4°C overnight. After washing the samples with PBS, the cell samples were treated with a 1:100 dilution of secondary antibody, which was derived from goat anti-mouse IgG (AS001, Abclonal) and conjugated with FITC, for 2 h. Cell nuclei were counterstained with DAPI staining solution (C1002, Beyotime, China). Finally, the stained cells were viewed using a fluorescence microscope (Leica DM2500, Leica Microsystems), and the images were captured. The fusion index was calculated as the percentage of nuclei in fused myotubes out of the total nuclei. The number of nuclei in each image was evaluated using the ImageJ plus software.

Fatty acid extraction and methyl-esterification were performed as previously described (Zhu et al., 2022a). Briefly, the recovered fatty acid methyl esters were examined through GC-MS (QP2010 ultra mass spectrometer, GC 2010 plus, Thermo Scientific). The electron energy was fixed at 70 eV, and the temperature programming profile was as follows: kept for 5 min at 60°C, up to 120°C in increments of 10°C/min, maintained for 5 min at 120°C, up to 190°C in increments of 5°C/min, kept for 7 min at 190°C, up to 230°C in increments of 2°C/min, and kept for 10 min at 230°C. A 5-MS column (Restek, United States) was used to resolve all samples. The detector and ion source were run at 240°C and 230°C, respectively. The peaks obtained were identified by comparing their retention times with those of known standards (Sigma). Pentadecanoic acid (C15:0) was used as the internal standard, and each sample was normalized to total cellular protein concentration, as measured using the Quick BCA protein assay kit (SYW3-1, Solarbio, China).

Cell samples were produced in the manner previously described (Zhu et al., 2022b). Lipidomic analysis was conducted using LC-MS (QExactive Plus Orbitrap mass spectrometer, Thermo Scientific). Acetonitrile:MilliQ water (6:4 v/v) and isopropanol:acetonitrile (9:1 v/v) were used as solvents A and B, respectively; both solvents contained 10 mm ammonium acetate. Column chromatography was performed using the Waters ACQUITY UPLC CSHTM C18 column. The gradient profile was as follows: 32%–100% solvent B over 24 min, back to 32% solvent B and 6 min before the next injection, and equilibrate the column. LipidSearch program v4.1.16 (Thermo Scientific) was used to identify the type of lipids. A pool of all lipid extracts was prepared and used for quality control. SIMCA-P (Sweden) was used to import the raw MS data, and an orthogonal projection was performed for latent structures-discriminant analysis (OPLS-DA). The resulting heatmap plots were drawn using R software.

RNA sequencing (RNA-seq) was conducted as previously described (Zhang et al., 2021; Zhu et al., 2022c). Briefly, total RNA of C2C12 myoblasts transfected with si-NC or si-Apol9a was isolated using the RNA extraction kit (K101, JN. BIOTOOLS, China). Library construction and paired-end sequencing were performed by GENEWIZ Biotech (Suzhou, China). Raw data files were matched to the mouse reference genome by using STAR software (http://www.code.google.com/p/rna-star/). For each sample, fold change was estimated using the fragments per kilobase per million reads values, and differential expression analysis was performed using the DESeq2 package. The standard of fold change ≥1.5 and p < 0.05 were defined to screen differentially expressed genes (DEGs). Gene ontology (GO) analysis was conducted to perform the functional enrichment analysis of specific DEGs using Metascape (metascape.org). All RNA-seq data files have been deposited in the SRA database (Accession: PRJNA845924).

Statistical analysis was performed using SPSS software version 22.0 and GraphPad Prism 8.0. The results are shown as mean ± SEM from at least three independent experiments. Differences between groups were evaluated using an unpaired Student’s t-test (between two groups) or one-way ANOVA (between multiple groups). *p < 0.05, **p < 0.01, and ***p < 0.001 were considered statistically significant.

To evaluate the effects of obesity on skeletal muscle differentiation, we used the HFD-induced obesity mouse model (Feraco et al., 2021). Body weight and TG content in skeletal muscle tissue were considerably higher in HFD mice than in normal diet (ND) mice (Figures 1A,C). Meanwhile, the HFD mice exhibited a lower gastrocnemius muscle mass (Figure 1B) than the ND mice. Furthermore, a significant decrease in the mRNA expression levels of myogenesis-related genes (including Myf5, MyoD, MyoG, and MyHC) was observed in the gastrocnemius muscle from HFD mice (Figure 1D). Based on Pearson’s correlation analysis, we investigated the correlations between body weight, gastrocnemius muscle mass, expression levels of myogenesis-related genes, and TG levels in HFD mice (Figure 1E). A notable negative correlation was observed between TG content and gastrocnemius muscle mass. Together, these results demonstrate that muscle mass and myogenesis capacity were decreased in obese mice.

To determine the critical genes involved in myogenic differentiation, we compared the transcriptome of differentiated C2C12 cells (an established myoblast cell model) with that of undifferentiated C2C12 cells (Yaffe and Saxel, 1977). After myogenic differentiation, C2C12 cells became fused and the myotube diameter increased (Figure 2A). As demonstrated in the volcano plot (Figure 2B), 1353 genes were upregulated following myogenic differentiation and 2731 genes were downregulated (fold change ≥1.5). The upregulated genes were subjected to GO analysis, and the top 10 enriched GO terms were all related to muscle differentiation (Figure 2C). Among the upregulated gene sets, we observed that many apolipoprotein family genes were significantly upregulated following myogenic differentiation (Figure 2D). The most significant change involved Apol9a, an understudied cytoplasmic, interferon-inducible gene with antiviral activity (Kreit et al., 2015). To our knowledge, a role for Apol9a in myogenic differentiation has not been previously reported. Using qPCR, we confirmed that Apol9a mRNA expression levels increased following myogenesis (Figure 2E). However, Apol9a mRNA expression levels decreased in obese mice (Figure 2F). Together, these results indicate that Apol9a may play a role in myogenic differentiation.

To further investigate the function of Apol9a in the progress of myogenesis, siRNA was used to interfere with Apol9a expression. C2C12 cells were transfected with Apol9a siRNA or NC siRNA, and then, the GM was replaced with DM for 4 days. Apol9a mRNA expression levels were reduced approximately 60% in Apol9a siRNA-transfected cells compared with the si-NC control group (Figure 3A). Myogenic differentiation is known to start with the induction of a specific set of transcription genes known as myogenic regulatory factors (MRFs), which include Myf5, MyoD, and MyoG (Chargé and Rudnicki, 2004; Ciciliot and Schiaffino, 2010). Comparison with the si-NC control group, the protein and mRNA expression levels of several myogenic differentiation markers, including Myf5, MyoD, MyoG, and MyHC, were remarkably decreased in differentiated C2C12 cells after si-Apol9a transfection (Figures 3B–D). To validate the effect of Apol9a deficiency on myotube formation, we evaluated the expression of MyHC protein through immunofluorescence in differentiated C2C12 cells and a quantitative analysis of the fusion index. As shown in Figure 3E, Apol9a knockdown cells displayed a different morphology and appeared to have shorter myofibers compared with the control group. Moreover, the fusion index was significantly decreased after Apol9a silencing (Figure 3E). Furthermore, we explored the effect of Apol9a overexpression on myogenic differentiation. Compared with the empty vector group, the Apol9a mRNA expression level was markedly increased in pcDNA3.1-Apol9a-transfected cells. Compared with the empty vector group, the Apol9a mRNA expression level was significantly increased in pcDNA3.1-Apol9a-transfected cells (Figure 4A), and the myotube diameter was significantly longer after Apol9a overexpression (Figure 4B). Western blot analysis showed that Apol9a overexpression significantly upregulated MyoG and MyHC protein levels (Figures 4C,D). Together, these results demonstrate that Apol9a knockdown inhibited myogenesis and Apol9a overexpression promoted myogenic differentiation.

Apol9a is an apolipoprotein family member, and previous research has suggested that Apol9a has a role in lipid transport (Arvind and Rangarajan, 2016). To identify changes in the types of fatty acids and changes in lipid composition induced by Apol9a knockdown after cell differentiation, the types and contents of fatty acid in C2C12 cells with or without Apol9a knockdown after myogenic differentiation were determined through GC-MS and LC-MS, respectively. After Apol9a knockdown, the levels of various fatty acids (C16:0, C18:0, C18:2, and C22:6) significantly decreased (in comparison with the si-NC transfected cells) (Figure 5A). To further analyze changes in lipid composition, we applied LC-MS-based lipidomics to identify the specific types of lipids whose levels decreased in Apol9a knockdown cells. Moreover, orthogonal partial least squares-discriminant analysis (OPLS-DA) indicated different clustering of lipids from NC and Apol9a-silenced groups (Figure 5B). While the intensity levels of C16:0 fatty acid were mainly decreased in phosphatidylcholine (PC) and phosphatidylethanolamine (PE), those of C18:0 fatty acid were reduced in TG and PE (Figure 5D). In addition, a significant decrease was observed in the intensity levels of C18:2 fatty acid, especially in PC and TG (Figures 5C,E). Elevated skeletal muscle TG is reported to be related to insulin resistance in obesity and impaired skeletal muscle differentiation (Wu and Ballantyne, 2017), (Eum et al., 2020). Furthermore, our heatmap analysis revealed that many TG species (including C18:0 and C18:2 fatty acids) demonstrate a remarkable decrease (VIP >1 and p < 0.05) in C2C12 cells after Apol9a knockdown and myogenic differentiation (Figure 5F). Together, these results provide evidence that inhibition of myogenesis following Apol9a knockdown may be related to changes in lipid composition, especially changes in TG species.

To elucidate the underlying molecular mechanisms of Apol9a regulation of myogenesis, RNA sequencing was used to evaluate gene expression differences in C2C12 cells with or without Apol9a knockdown. Principal component analysis (PCA) was then performed to discriminate changes in expression between experimental groups. The PCA results revealed that the two groups could be completely separated based on their transcriptomes (Figure 6A). Based on the criteria of fold change ≥1.5 and p < 0.05, 580 differentially expressed genes (DEGs) between the two groups were identified, including 282 downregulated genes and 298 upregulated genes (Figure 6B). Then, GO analysis was applied to find the potential functions of the upregulated and downregulated genes. As shown in Figure 6C, significantly enriched functional terms, including “tissue homeostasis,” “response to lipoprotein particle,” “negative regulation of cell differentiation,” and “MAPK cascade,” were identified using the upregulated gene set. However, no specifically enriched biological processes were identified using the downregulated gene set (data not shown). The MAPK cascade pathway is essential for the regulation of myogenic differentiation (Jones et al., 2001; Keren et al., 2006; Xie et al., 2018; Boyer et al., 2019). We hypothesized that the ERK1/2 signaling pathway may play a role in the inhibition of myogenic differentiation induced by Apol9a knockdown. This hypothesis was subsequently confirmed by western blotting analyses of the ERK1/2 phosphorylation status and protein levels in C2C12 cells. Thus, while Apol9a knockdown markedly increased ERK1/2 phosphorylation during myogenesis, other MAPKs (JNK and p38) remained unchanged (Figure 6D). Moreover, phosphorylated ERK1/2 protein levels were significantly reduced after C2C12 differentiation (Figure 6E). In summary, our data demonstrate that Apol9a knockdown activates the ERK1/2 signaling pathway during myogenic differentiation.

To test the hypothesis that Apol9a regulates myogenic differentiation via the ERK1/2 signaling-dependent pathway, we applied PD98059, a specific inhibitor of ERK1/2, to the C2C12 cells. In agreement with our hypothesis, Apol9a knockdown-induced inhibition of myogenesis was almost completely reversed by PD98059 (Figure 7A). Furthermore, western blot analysis and immunofluorescence experiments demonstrated that ERK inhibition blocked the downregulation of myogenic genes (MyHC, MyoD, and MyoG) otherwise observed during Apol9a knockdown (Figures 7B,C). Together, these results provide evidence that Apol9a regulates myogenic differentiation via the ERK pathway.