关键词: 骨关节炎, 特征基因, 生物信息学, 孟德尔随机化, 药食同源中药, 分子对接, 分子动力学模拟

Abstract: BACKGROUND: Early diagnosis and treatment of osteoarthritis remain a significant challenge due to the lack of highly specific biomarkers.  
OBJECTIVE: To screen characteristic genes of osteoarthritis, predict potential food-medicine homology traditional Chinese medicine and their core components, and validate their therapeutic potential through molecular docking and molecular dynamics simulations.
METHODS: This study is based on three datasets (GSE55235, GSE169077, and GSE55457) from the GEO database (a public database maintained by the National Center for Biotechnology Information in the United States, which contains a large amount of gene expression and related high-throughput experimental data), including a total of 25 normal samples and 26 osteoarthritis samples. It combines genes extracted from the eQTL database (a publicly available genome-wide database constructed by an international multi-center research team) as exposure factors and osteoarthritis data extracted from the IEU openGWAS database (a large-scale resource library publicly available from the MRC Integrative Epidemiology Unit at the University of Bristol in the UK) as outcome factors (including 407 746 samples). Least Absolute Shrinkage and Selection Operator, random forest, and support vector machine algorithms are used to determine core biomarkers. Subsequently, CIBERSORT is used to assess immune infiltration characteristics and single-gene Gene Set Enrichment Analysis is conducted. Potential traditional Chinese medicines are predicted based on the Coremine Medical database (a biomedical database developed by PubGene Technology Company) and the HERB database (a traditional Chinese medicine database jointly established by several top universities in China), and food-medicine homologous traditional Chinese medicines and their core components are screened. Molecular docking and molecular dynamics simulations are further conducted for verification.
RESULTS AND CONCLUSION: (1) The solute carrier family 2 member 3 and atypical chemokine receptor 1 genes are identified as characteristic genes of osteoarthritis. These genes have good diagnostic efficacy in osteoarthritis and are involved in metabolic regulation, cell signaling, and inflammatory response processes, closely related to glucose metabolism, immune regulation, and inflammatory signaling pathways. (2) Through drug prediction, seven traditional Chinese medicines with food and medicinal properties, including Evodia rutaecarpa, Perilla frutescens seed, Ganoderma lucidum, Gastrodia elata, Prunus armeniaca seed, Syzygium aromaticum, and Rehmannia glutinosa, were screened out. Their core components are β-sitosterol and stigmasterol. Molecular docking and molecular dynamics simulation results show that stigmasterol has the best affinity with solute carrier family 2 member 3, and the complex shows high stability. (3) This study systematically reveals the key role of solute carrier family 2 member 3 and atypical chemokine receptor 1 in the pathogenesis of osteoarthritis and confirms the possibility that traditional Chinese medicines with food and medicinal properties can intervene in the pathological process of osteoarthritis through multiple targets and multiple pathways. The research results not only provide new molecular basis for the early diagnosis and targeted treatment of osteoarthritis, but also provide theoretical support for the prevention and treatment strategies of osteoarthritis. This study, from the perspective of traditional Chinese medicine and modern molecular biology, provides certain references for the subsequent clinical application of traditional Chinese medicine in osteoarthritis.

Key words: osteoarthritis, characteristic genes, bioinformatics, Mendelian randomization, food-medicine homology traditional Chinese medicine, molecular docking, molecular dynamics simulation