Exploring the role of mitochondrial dysfunction-related genes in osteoporosis using weighted gene co-expression network analysis and machine learning

作者：Hengxi Zeng, Xiang-tao Yuan, Lianguo Wu, Ruikun Zhang, Kun Yan, Xiaobo Lai, Peng Xu, Xinmiao Yao · 发表于：European journal of medical research · 年份：2025 · DOI：10.1186/s40001-025-03675-0 · 被引用次数：1 · 研究领域：Bone Metabolism and Diseases、Bone health and osteoporosis research、Bioinformatics and Genomic Networks

BACKGROUND: Osteoporosis (OP) is a prevalent age-related disease characterized by bone loss and increased fracture risk. Mitochondrial dysfunction is tightly associated with OP pathogenesis. This study aimed to explore the role of mitochondrial dysfunction-related genes (MDRGs) in OP. METHODS: Differentially expressed genes (DEGs) were identified from OP transcriptomic data sets, and weighted gene co-expression network analysis was performed to determine OP-associated modules. Machine learning algorithms (Random Forest, Support Vector Machine, and Lasso) were applied to identify hub MDRGs. Gene set enrichment analysis (GSEA) and regulatory network analysis were performed to elucidate their functions and regulatory interactions. Hub gene's expression was validated in tissues using immunofluorescence, and osteogenic induction assays were conducted to validate the biological role and mechanism of the key gene VPS35 in osteoblast differentiation. RESULTS: ALAS1, HSPB1, and VPS35 were identified as hub MDRGs, with VPS35 showing good diagnostic performance. GSEA highlighted potential molecular pathways associated with hub genes in OP, such as TNF-NFκB, RAS-ERK, and RAS-PI3K signaling pathways. In addition, a regulatory network involving hub genes, miRNAs, and lncRNAs was preliminarily constructed, providing insights into possible regulatory mechanisms underlying OP. Experimental validation demonstrated that VPS35 overexpression inhibited osteoblast differentiation by suppressing th...