Multi-omics analysis identifies CD48 and ITIH5 as potential diagnostic biomarkers and therapeutic targets for metabolic disorders associated with obesity.

BACKGROUND: Obesity-driven metabolic disorders involve pathological remodeling of white adipose tissue (WAT), wherein fibroblasts coordinate fibrosis and inflammation. This study aimed to characterize fibroblast transcriptional heterogeneity, identify fibrosis-inflammation mediators, and discover diagnostic and therapeutic targets. METHODS: Single-cell (GSE155960) and bulk (GSE205668) RNA-seq datasets from human WAT were integrated. Cell-cell communication was inferred using CellChat, and fibroblast gene co-expression modules were built via high-density weighted gene co-expression network analysis (hdWGCNA). Candidate biomarkers were identified by intersecting bulk differentially expressed genes (DEGs) with top module hub genes, refined by four machine learning models (LASSO, glmBoost, stepwise bidirectional/backward regression), and evaluated by ROC curves. Functional enrichment used CIBERSORT and GSEA. For in vivo validation, epididymal white adipose tissue from diet-induced obese mice was analyzed by qRT-PCR, western blot, and immunohistochemistry/immunofluorescence. In vitro, 3 T3-L1 cells were treated with 0.2 mM palmitic acid (PA). Drug candidates were predicted by DSigDB and molecular docking. RESULTS: Fibroblasts were the primary communication hubs, showing the most frequent and strongest interactions with endothelial and immune cells. CD48 and ITIH5 were identified as key biomarkers, both transcriptionally upregulated in obese adipose tissue, with AUC values of 0.82 (CD48) and 0.81 (ITIH5) for distinguishing obese from non-obese individuals. CD48 expression was associated with neutrophil infiltration and immune activation pathways, whereas ITIH5 was associated with lysosomal organization and extracellular matrix remodeling. In obese mice, both genes were elevated at the mRNA and protein levels, accompanied by increased CD48 and ITIH5 expression in both immune cells and stromal fibroblasts. PA-treated 3 T3-L1 cells showed similar upregulation. Molecular docking suggested raloxifene as a candidate CD48 ligand and bucladesine for ITIH5, both with favorable binding affinities. CONCLUSION: Multi-omics and machine learning establish fibroblasts as central hubs in obese WAT and identify CD48/ITIH5 as potential dual-axis biomarkers linking immune dysregulation and matrix remodeling in obesity-related metabolic disorders.