类风湿关节炎的潜在生物标志物及其免疫调控机制:基于GEO数据库

Identification of potential biomarkers and immunoregulatory mechanisms of rheumatoid arthritis based on multichip co-analysis of GEO database

目的 探讨类风湿关节炎(RA)的潜在发病机制,挖掘可以作为RA早期诊断的生物标志物,并探究其可能的免疫调控机制。方法 依托GEO数据库多个基因组数据,利用limma包、RRA方法、批次矫正方法与clusterProfiler包筛选差异表达基因并预测其作用功能;通过STRING在线数据库得到蛋白质互作网络,利用Cytoscape3.8.0软件与GeneMANIA数据库筛选差异表达关键基因并预测其互作机制;构建ROC工作曲线验证关键基因诊断模型的信效度,并通过机器学习方法筛选疾病特征免疫细胞;使用corrplot包分析特征免疫细胞与疾病关键基因的相关性;使用GSEA富集方法探究关键基因生物学功能和相关性;通过构建胶原诱导性关节炎(CIA)大鼠模型探究具有较高诊断价值的差异基因在RA滑膜组织中的表达情况。结果 得到与RA密切相关核心关键基因9个,分别为:CD3G、CD8A、SYK、LCK、IL2RG、STAT1、CCR5、ITGB2、ITGAL,富集分析结果表明差异基因可能主要通过细胞因子相关通路调控滑膜炎症;ROC工作曲线分析表明9个核心基因的预测模型信效度较高,其中STAT1预测模型的AUC值最高(0.909);相关性分析结果显示,CD3G、ITGAL、LCK、CD8A和STAT1等5个核心基因与疾病特征免疫细胞相关性较强,其中STAT1与M1型巨噬细胞的相关性联系最强(R=0.68,P<0.001);CIA大鼠膝关节的HE、番红-O固绿染色结果表明关节炎大鼠造模成功且药物治疗有效,免疫荧光结果表明STAT1和磷酸化的STAT1在CIA大鼠踝关节滑膜组织中表达高于正常组和治疗组;免疫印迹实验结果表明,STAT1蛋白表达量在滑膜成纤维细胞的细胞核(P=0.0002)与细胞质(P<0.0001)中存在差异,且经过治疗,细胞核内p-STAT1与STAT1的蛋白表达量显著降低(P<0.0001)。结论 CD3G、CD8A、SYK、LCK、IL2RG、STAT1、CCR5、ITGB2与ITGAL等9个基因可作为RA的早期诊断生物标志物,CD3G/CD8A/LCK-γδ T细胞、ITGAL-Tfh细胞、STAT1-M1型巨噬细胞等基因-免疫细胞途径可能与RA的发生发展密切相关。

Objective To identify the biomarkers for early rheumatoid arthritis(RA)diagnosis and explore the possible immune regulatory mechanisms.Methods The differentially expressed genesin RA were screened and functionally annotated using the limma,RRA,batch correction,and clusterProfiler.The protein-protein interaction network was retrieved from the STRING database,and Cytoscape 3.8.0 and GeneMANIA were used to select the key genes and predicting their interaction mechanisms.ROC curves was used to validate the accuracy of diagnostic models based on the key genes.The disease-specific immune cells were selected via machine learning,and their correlation with the key genes were analyzed using Corrplot package.Biological functions of the key genes were explored using GSEA method.The expression of STAT1 was investigated in the synovial tissue of rats with collagen-induced arthritis(CIA).Results We identified 9 core key genes in RA(CD3G,CD8A,SYK,LCK,IL2RG,STAT1,CCR5,ITGB2,and ITGAL),which regulate synovial inflammation primarily through cytokines-related pathways.ROC curve analysis showed a high predictive accuracy of the 9 core genes,among which STAT1 had the highest AUC(0.909).Correlation analysis revealed strong correlations of CD3G,ITGAL,LCK,CD8A,and STAT1 with disease-specific immune cells,and STAT1 showed the strongest correlation with M1-type macrophages(R=0.68,P=2.9e-08).The synovial tissues of the ankle joints of CIA rats showed high expressions of STAT1 and p-STAT1 with significant differential expression of STAT1 between the nucleus and the cytoplasm of the synovial fibroblasts.The protein expressions of p-STAT1 and STAT1 in the cell nuclei were significantly reduced after treatment.Conclusion CD3G,CD8A,SYK,LCK,IL2RG,STAT1,CCR5,ITGB2,and ITGAL may serve as biomarkers for early diagnosis of RA.Gene-immune cell pathways such as CD3G/CD8A/LCK-γδT cells,ITGAL-Tfh cells,and STAT1-M1-type macrophages may be closely related with the development of RA.