Lan Zhibin, Yang Yang, Sun Rui, Lin Xue, Xue Di, Wu Zhiqiang, Jin Qunhua
The third ward of orthopaedic department, Institute of Osteoarthropathy, Institute of Medical Sciences, General Hospital of Ningxia Medical University, 804 Shengli South Street, Yinchuan, Ningxia Hui Autonomous Region, 750004, PR China; Ningxia Key Laboratory of Clinical and Pathogenic Microbiology, General Hospital of Ningxia Medical University, 804 Shengli South Street, Yinchuan, Ningxia Hui Autonomous Region, 750004, PR China.
Ningxia Key Laboratory of Clinical and Pathogenic Microbiology, General Hospital of Ningxia Medical University, 804 Shengli South Street, Yinchuan, Ningxia Hui Autonomous Region, 750004, PR China.
Comput Biol Med. 2025 Mar;187:109816. doi: 10.1016/j.compbiomed.2025.109816. Epub 2025 Feb 11.
The objective of this study was to identify aging-related immunophenotypic biomarkers associated with osteoarthritis (OA) using advanced machine learning techniques. We employed a combination of lasso regression and random forest algorithms to analyze transcriptomic data obtained from OA patients. Differential expression analysis and functional enrichment analysis were conducted to identify aging-related differentially expressed genes (ag-DEGs) and annotate their biological functions. Furthermore, correlation analysis among hub genes and immune cell infiltration analysis were performed to understand the molecular phenotypes of OA. Our analysis identified 43 ag-DEGs enriched in immune-related biological processes and pathways. Lasso regression and random forest analysis narrowed down the gene pool to three hub genes: CACNA1A, FLT1 and KCNAB3. These genes exhibited differential expression between normal and OA groups and demonstrated high accuracy in distinguishing between them. Clustering analysis revealed two distinct molecular phenotypes of OA: an "immune-activated subgroup" and an "immune-suppressed subgroup." Experimental validation confirmed the expression patterns of hub genes. This study identified biomarkers associated with the aging-related immune phenotype in OA, shedding light on potential targets for immunotherapy and personalized medical treatments. Characterized by CACNA1A, FLT1, and KCNAB3, clustering analysis suggests that OA can be divided into two molecular phenotypes: an "immune-activated subgroup" and an "immune-suppressed subgroup." The findings may contribute to the development of novel therapeutic strategies aimed at modulating immune responses in OA patients, ultimately improving treatment outcomes and prognosis.
本研究的目的是使用先进的机器学习技术，识别与骨关节炎（OA）相关的衰老相关免疫表型生物标志物。我们采用套索回归和随机森林算法相结合的方法，分析从OA患者获得的转录组数据。进行差异表达分析和功能富集分析，以识别衰老相关的差异表达基因（ag-DEGs）并注释其生物学功能。此外，还进行了枢纽基因之间的相关性分析和免疫细胞浸润分析，以了解OA的分子表型。我们的分析确定了43个富集于免疫相关生物学过程和途径的ag-DEGs。套索回归和随机森林分析将基因库缩小到三个枢纽基因：CACNA1A、FLT1和KCNAB3。这些基因在正常组和OA组之间表现出差异表达，并在区分两者方面显示出高准确性。聚类分析揭示了OA的两种不同分子表型：“免疫激活亚组”和“免疫抑制亚组”。实验验证证实了枢纽基因的表达模式。本研究确定了与OA中衰老相关免疫表型相关的生物标志物，为免疫治疗和个性化医疗的潜在靶点提供了线索。以CACNA1A、FLT1和KCNAB3为特征，聚类分析表明OA可分为两种分子表型：“免疫激活亚组”和“免疫抑制亚组”。这些发现可能有助于开发旨在调节OA患者免疫反应的新型治疗策略，最终改善治疗效果和预后。
