Identifying aging-related biomarkers in adipose tissue using integrative bioinformatics and machine-learning approaches: discovery of ELN, MXD1, and FGF21 as key genes.

BACKGROUND: Adipose tissue plays a critical role in aging and age-related diseases. However, the specific molecular and cellular alterations associated with aging in adipose tissue remain incompletely understood. METHODS: Aging-related differentially expressed genes (DEARGs) were identified by intersecting differentially expressed genes (DEGs) in adipose tissue, age-related genes (ARGs), and human genes linked to aging. Functional enrichment analysis was conducted to explore the potential roles of these DEARGs. Protein-protein interaction (PPI) networks were analyzed using STRING, and hub DEARGs were identified via least absolute shrinkage and selection operator (LASSO) analysis. Oil Red O staining was used to confirm adi-pocyte differentiation, and D-galactose treatment induced cellular senescence. Validation of hub DEARG expression was conducted in an independent dataset and confirmed using quantitative polymerase chain reaction (qPCR) both in vitro and in vivo. RESULTS: Forty-nine DEARGs were identified, with functional enrichment analyses revealing significant roles in glucose homeostasis and key aging pathways, including the FoxO and JAK-STAT signaling pathways, Th17 cell dif-ferentiation, growth hormone signaling, the adiponectin pathway, and AMPK pathway. Five hub genes (PCK1, ELN, MXD1, STAT3, and FGF21) were selected through interaction network anal-ysis and LASSO regression. Expression levels of three DEARGs (ELN, MXD1, and FGF21) were validated by qPCR and an independent dataset. CONCLUSIONS: This study identified three DEARGs (ELN, MXD1, and FGF21) as potential biomarkers of adipose tissue aging, suggesting their role in organismal aging and age-related disease pathways.