The crosstalk between exercise and ncRNAs in NAFLD: bioinformatic evidence and regulatory network analysis

Non-alcoholic fatty liver disease (NAFLD) is associated with metabolic dysfunction, high-fat diets, and sedentary lifestyle, leading to chronic inflammation, hepatic fibrosis, and an increased risk of cirrhosis and hepatocellular carcinoma. Given the limitations of drug therapies, exercise, as an accessible approach, plays a key role in improving lipogenesis control, reducing inflammation, and enhancing mitochondrial function. The aim of this review–bioinformatic study was to examine the interactions between exercise and non-coding RNAs (ncRNAs)—miRNAs, lncRNAs, and circRNAs—in the context of NAFLD, to identify ncRNAs modulated by exercise, and to reconstruct competitive endogenous RNA (ceRNA) networks related to lipid metabolism, inflammation, and bioenergetics. To achieve this, murine models (high-fat diet with/without exercise) were retrieved from GEO (GSE239729 and GSE226132), followed by a systematic search of PubMed, Scopus, Web of Science, and Google Scholar (2019–2025) using standard keywords. Preprocessing and normalization were performed in R (v4.5.0) with DESeq2, focusing on ncRNAs meeting |log2FC| ≥ 2 and FDR < 0.05. Annotations were obtained from Ensembl, NONCODE, LNCipedia, circBase, and circAtlas; miRNA targets were predicted using TargetScan and miRWalk; circRNA–miRNA interactions were obtained from starBase; and ceRNA networks were reconstructed in Cytoscape (v3.10.1). Functional enrichment (GO/KEGG) was conducted with clusterProfiler and DAVID. Integrated findings from the literature review and reanalysis showed that exercise consistently decreased miR-34a and increased miR-122; among lncRNAs, MALAT1 and HOTAIR decreased; and among circRNAs, circRNA_002581 decreased while circScd1 increased. The ceRNA networks predominantly targeted AMPK, SIRT1/PGC-1α, and PPARα axes, consistent with reduced lipogenesis (↓SREBP1/FASN), enhanced fatty-acid oxidation (↑CPT1/FAO), attenuated inflammation (↓NF-κB/TNF-α), and modulation of TGF-β signaling. Concordance of patterns between the two datasets confirmed the robustness of the exercise-driven, ncRNA-centered response, and sensitivity analyses with more permissive thresholds preserved key results. In summary, exercise remodels NAFLD liver by coordinated ncRNA reprogramming that reshapes gene-network architecture governing metabolism, inflammation, and mitochondrial performance. By integrating literature evidence with RNA-Seq reanalysis and ceRNA reconstruction, this study provides a practical conceptual framework for targeting ncRNAs, developing biomarkers of exercise response, and designing RNA-based interventions that mimic the effects of exercise. By integrating the systematic review with reanalysis of existing datasets and ceRNA network building, we show that the benefits of exercise are mediated through AMPK, SIRT1/PGC-1α, and PPARα pathways and align with ncRNA signatures (miR-34a↓, miR-122↑, MALAT1/HOTAIR↓, circScd1↑). This network map has clear clinical applications: (1) exercise-response biomarkers, and (2) RNA-based exercise-mimetic interventions targeting these hubs to optimize therapy and improve metabolic/inflammatory clinical outcomes in NAFLD.