Identification of Potential Biomarkers Associated with Impaired Fatty Acid Oxidation in Aged Skeletal Muscle Using...

Objective: Impaired fatty acid oxidation (FAO) is considered an important metabolic mechanism underlying skeletal muscle aging and sarcopenia; however, the key regulatory molecules involved in this process remain incompletely defined. This study aimed to identify candidate biomarkers associated with impaired FAO in aged skeletal muscle, characterize their potential biological functions and regulatory features through integrated bioinformatics and machine learning analyses, and preliminarily validate their expression patterns in in vivo and in vitro aging models. Methods: Skeletal muscle aging transcriptomic datasets GSE1428 and GSE674 were obtained from the Gene Expression Omnibus database. FAO-related genes were retrieved from GeneCards. Differentially expressed FAO-related genes (DE-FAOGs) were identified through differential expression analysis and were further analyzed by Gene Ontology and Kyoto Encyclopedia of Genes and Genomes enrichment analyses. Random forest, Boruta, and protein-protein interaction (PPI) network analyses were used to screen hub genes, and an artificial neural network (ANN) model was constructed. Single-cell RNA sequencing analysis, gene set enrichment analysis, ceRNA network construction, drug prediction, molecular docking, and molecular dynamics simulation were further performed. Hub gene expression was validated by qRT-PCR in naturally aged mice and D-galactose-induced senescent C2C12 cells. Results: A total of 69 DE-FAOGs were identified and were mainly enriched in mitochondrial function, electron transport chain, and energy metabolism-related pathways. Three hub genes, creatine kinase, mitochondrial 2 (CKMT2), actin alpha cardiac muscle 1 (ACTC1), and forkhead box O3 (FOXO3), were identified by random forest, Boruta, and PPI analyses. Receiver operating characteristic (ROC) analysis showed good discriminatory performance for these genes. The three-gene ANN model achieved area under the curve (AUC) values of 0.992 and 0.964 in the training and validation datasets, respectively. Gene set enrichment analysis (GSEA) suggested that the hub genes were closely associated with mitochondrial energy metabolism, lipid metabolism, and stress regulation. qRT-PCR confirmed decreased Ckmt2 expression and increased Actc1 and Foxo3 expression under aging conditions, consistent with the bioinformatics results. Conclusions: CKMT2, ACTC1, and FOXO3 are potential biomarkers associated with impaired FAO in aged skeletal muscle. The ANN model based on these three genes showed good predictive performance and may provide new insights into the metabolic mechanisms and therapeutic targets of sarcopenia.