基于WGCNA联合Lasso回归构建结直肠癌ICI治疗预测标志物并验证  

Predictors of ICI treatment for colorectal cancer constructed and validated based on WGCNA combined with Lasso regression

作　　者：张虎 吴红雁 赵小芳 杜欣娜 ZHANG Hu;WUHong-Yan;ZHAO Xiao-Fang(Department of Physiology and Biochemistry,Faculty of Basic Medicine,Jiangsu Vocational College of Medicine,Yancheng 224005,Jiangsu,China)

机构地区：[1]江苏医药职业学院基础医学部生理生化教研室,江苏盐城224005 [2]江苏医药职业学院生物医药研究院,江苏盐城224005 [3]江苏医药职业学院基础医学部病理与病理生理学教研室,江苏盐城224005

出　　处：《中国老年学杂志》2025年第6期1296-1302,共7页Chinese Journal of Gerontology

基　　金：国家自然科学基金面上项目(82274339);江苏省青蓝工程(2021);江苏省产学研合作项目(FZ20210156);江苏省教育厅科研项目(22KJD310001);盐城市工程中心(YC2022808);江苏医药职业学院校企横向合作项目(2021320906000215,2023320906000197);江苏医药职业学院校院横向合作项目(2021320906000004);江苏医药职业学院校级科研项目(20214401)。

摘　　要：目的构建一种结直肠癌(CRC)免疫检查点抑制剂(ICI)治疗的预测标志物并验证。方法首先利用加权基因共表达网络分析(WGCNA)获得CRC免疫浸润相关基因模块。然后,利用单因素生存分析和Lasso-Cox回归构建CRC预后标签及其风险分数的计算公式,并利用Kaplan-Meier和受试者工作特征(ROC)曲线评估该标签预后效能。接下来,利用肿瘤免疫功能障碍和排斥(TIDE)在线工具评估各CRC数据集的免疫治疗效果,分析标签风险分数对上述治疗的预测作用。最后,基因集富集分析(GESA)被用于分析预后标签的潜在分子机制。结果WGCNA筛选获得了免疫相关模块基因614个,其中63个基因与CRC预后有关(均P<0.05)。Lasso-Cox回归构建了一个包含17个基因的预后标签,低风险组CRC患者总体预后显著优于高风险组(P<0.001),且预后敏感度和特异度较高。进一步地,ICI治疗预测显示GSE39582中低风险组响应率显著高于高风险组(P<0.001),4个外部数据集验证结果均显示,低风险组ICI治疗响应率显著高于高风险组(P<0.001,P<0.05)。最后,GSEA显示G2M检查点、HEDGEHOG信号通路、干扰素反应、缺氧和上皮间质转化等信号通路可能介导预后标签调节免疫的功能。结论基于免疫相关基因构建的预后模型不仅可以有效预测CRC患者总体生存率,还可以预测免疫治疗疗效。Objectivee To develop and validate a predictive biomarker for immune checkpoint inhibitor(ICI)therapy in colorectal cancer(CRC).Methods Firstly,weighted gene co-expression network analysis(WGCNA)was used to identify CRC immune infiltration-related gene modules.Then,univariate survival analysis and Lasso-Cox regression were employed to construct a CRC prognostic signature and its riskscore calculation formula.Kaplan-Meier and receiver operating characteristic(ROC)curves were utilized to evaluate the prognostic performance of the signature.Next,tumor immune dysfunction and rejection(TIDE)was used to assess the efficacy of immunotherapy in various CRC datasets,analyzing the predictive role of the signature's riskscores on the aforementioned treatments.Finally,gene set enrichment analysis(GSEA)was conducted to explore the potential molecular mechanisms of the signature.Results 614 immune-related module genes were identified by WGCNA,of which 63 genes were associated with CRC prognosis(all P<0.05).A prognostic signature comprising 17 genes were constructed by Lasso-Cox regression,with the low-risk group of CRC patients showing significantly better overall prognosis than that in the high-risk group(P<0.001),and the signature exhibiting high sensitivity and specificity.Furthermore,ICI therapy prediction revealed that the response rate in the medium-,low-risk groups of GSE39582 was significantly higher than that in the high-risk group(P<0.001).Validation in four external datasets consistently showed a significantly higher ICI therapy response rate in the low-risk group compared to the high-risk group(P<0.001,P<0.05).Finally,GSEA indicated that signal pathways such as the G2M checkpoint,HEDGEHOG signaling pathway,interferon response,hy-poxia,and epithelial-mesenchymal transition might mediate the immunoregulatory functions of the signature.Conclusions The prog-nostic model based on immune-related genes could not only effectively predict the overall survival rate of CRC patients,but also predict the efficacy of immunothera

关 键 词：结直肠癌(CRC) 加权基因共表达网络分析(WGCNA) Lasso-Cox回归 免疫检查点抑制剂(ICI)治疗 

分 类 号：R735.35[医药卫生—肿瘤]