应用“GSEA-WGCNA-验证”整合策略分析类风湿关节炎肝肾亏虚证的生物标志物研究

Identification of novel biomarkers for rheumatoid arthritis with liver⁃kidney deficiency pattern by a“GSEA⁃WGCNA⁃validation”integrated strategy

目的应用基因集富集分析(GSEA)和加权基因共表达网络分析(WGCNA)发现类风湿关节炎(RA)肝肾亏虚证的候选生物标志物,并进行临床验证,探讨RA肝肾亏虚证的生物学基础。方法临床收集RA肝肾亏虚证(3例)、7个非肝肾亏虚证(各3例)及健康志愿者(4名)的全血样本,开展转录组学测序。以健康志愿者为对照,筛选RA肝肾亏虚证的差异表达基因集进行富集分析和功能注释。以RA非肝肾亏虚证患者和健康志愿者为对照,对转录组表达谱开展GSEA和WGCNA联用挖掘,将获取的关键差异表达基因作为RA肝肾亏虚证的候选生物标志物。利用独立临床验证集样本(每组≥12例)对候选生物标志物的表达水平进行qPCR检测,采用受试者操作特征(ROC)曲线等评价其辨证效能。结果RA肝肾亏虚证的差异表达基因富集于“免疫-炎症”相关通路、细胞调控相关通路和代谢相关通路,同时,还参与肝肾发育和代谢等生物过程。转录组表达谱的GSEA富集结果表明,与非肝肾亏虚证组相比,肝肾亏虚证组的差异表达基因更明显地参与肝功能(脂质、血液)代谢调节,肾功能(水盐、激素)代谢调节和神经系统调节相关的作用通路。WGCNA分析使转录组表达谱的17010个基因被分为了19个特征模块,其中3个特征模块与肝肾亏虚证呈明显正相关(r>0.300,P<0.05),其生物功能以“免疫-炎症”调节为主。整合GSEA和WGCNA分析结果后,选取变异系数、作用通路和生物模块代表性均排名前50%的3个关键基因[花生四烯酸5-脂氧合酶(ALOX5)、含Patatin样磷脂酶结构域蛋白8(PNPLA8)及抗沉默功能1(ASF1A)]作为RA肝肾亏虚证的候选生物标志物。验证结果显示:ALOX5、PNPLA8和ASF1A的辨证敏感度分别为88.89%、100.00%和100.00%,特异度分别为84.51%、76.47%和78.69%,准确度分别为85.00%、79.49%和80.00%,精确度分别为88.89%、100.00%和100.00%,ROC曲线下面积值分别为0.860、0.910和0.900。结论应用“GSEA-WGCNA-验证”整合策略,发现了RA肝肾亏虚证的新型生物标志物,有助于提高RA核心证候临床精准化诊断的水平和证候客观化研究的深度。

Objective We aimed to(i)explore the biological basis of rheumatoid arthritis(RA)with liver⁃kidney deficiency(LKD)pattern using gene set enrichment analysis(GSEA)and weighted gene co⁃expression network analysis(WGCNA)method and(ii)identify and clinically validate candidate biomarkers.Methods Transcriptome sequencing was carried out on whole blood samples from RA patients with LKD pattern(3 cases),RA patients with other seven TCM patterns(3 cases each),and healthy volunteers(4 persons).Differentially expressed gene(DEG)sets of RA with LKD pattern were screened using healthy control samples and other TCM patterns as controls.Then,biological functions of DEGs were investigated by enrichment analysis and functional annotation.After that,the gene expression profiles were mined by GSEA and WGCNA to obtain key DEGs as candidate biomarkers for RA with LKD pattern.The expression levels of the candidate biomarkers were experimentally determined by qPCR using an independent clinical cohort(not less than 12 cases/group),and their clinical efficacy was assessed using receiver operating characteristic(ROC)curve analysis.Results DEGs of RA with LKD pattern were most significantly enriched in the“inflammation⁃immune”,cell regulation,and metabolic pathways,and also involved in biological processes such as liver and kidney development and metabolism.The GSEA result of the gene expression profiles indicated that the DEGs of RA with LKD pattern were more significantly involved in hepatic function(lipid,blood)metabolic regulation⁃,renal function(water,salt,hormone)metabolic regulation⁃,and neurological regulation⁃related pathways than those of RA with other TCM patterns.The expression profiles of 17010 genes were categorized into 19 functional modules through WGCNA,three of which were significantly positively correlated with LKD(r>0.300,P<0.05),and their biological functions mainly included“immune⁃inflammatory”regulation.After integrating the GSEA and WGCNA result,three key genes(ALOX5,PNPLA8,ASF1A)that ranked in the top 50%in terms of coefficient of variation and representativeness of pathway and biological modules were selected as candidate biomarkers for RA with LKD pattern.Further validation of the clinical independent samples and evaluation of the ROC model showed that the sensitivities of ALOX5,PNPLA8,and ASF1A were 88.89%,100.00%,and 100.00%,their specificities were 84.51%,76.47%,and 78.69%,their accuracies were 85.00%,79.49%,and 80.00%,their precision was 88.89%,100.00%,and 100.00%,and their values of area of ROC curve were 0.860,0.910,and 0.900,respectively.Conclusion This study applied the“GSEA⁃WGCNA⁃validation”integration strategy to identify novel biomarkers of RA with LKD patterns.The validation result of the independent sample set showed that they have good clinical efficacy,which may help improve the accuracy of clinical diagnosis of core RA patterns and the depth of objective research on the TCM patterns.