基于TCGA和GEO数据库探索结肠癌肿瘤微环境中的免疫相关预后因子

Exploring immune-related prognostic factors in the tumor microenvironment of colon cancer based on TCGA and GEO databases

目的 构建和评估结肠癌预后模型,探索肿瘤免疫微环境的特征,并分析其在结肠癌的免疫治疗中的作用。方法 从癌症基因组图谱(The Cancer Genome Atlas, TCGA)中下载结肠癌及其正常组织数据作为训练数据集,从基因表达综合数据库(Gene Expression Ominibus, GEO)中下载结肠癌及其正常组织数据作为验证数据集,从免疫学数据库和分析平台(Immunology Database and Analysis Portal, Immport)中下载免疫相关基因列表。基于以上数据集,利用生物信息学的方法构建并验证预后风险模型。另外,基于预后风险模型进行肿瘤样本分组,探讨高风险评分组和低风险评分组的肿瘤浸润水平差异、肿瘤突变负荷(tumor mutation burden, TMB)和免疫检查点抑制剂相关靶点的表达水平差异。结果 构建的预后模型ROC曲线下面积(AUC^(ROC))在3年时为0.668,4年时为0.699,5年时为0.696。此外,低风险评分与良好的预后显著相关(P<0.000 1),而高风险评分与较高的肿瘤分期显著相关(P<0.05),并在验证数据集中得到进一步验证。同时,构建的列线图模型的C-index从0.63增加至0.75。基于建立的预后风险模型,高风险组的免疫治疗相关靶基因的表达水平显著升高(P<0.05),有6种免疫细胞的浸润程度在高风险组和低风险组中的差异有统计学意义(P<0.05)。同时,较高的TMB与较差的预后显著相关(P<0.05)。结论 构建的结肠癌预后模型在评估结肠癌患者复发风险分层、肿瘤分期等方面具有潜在的临床意义,有助于结肠癌的肿瘤免疫微环境的探索及结肠癌患者的免疫治疗。

Objective To construct and evaluate the prognosis model of colon cancer,explore the characteristics of tumor immune microenvironment,and analyze their roles in the immunotherapy of colon cancer.Methods The data of colon cancer and its normal tissues downloaded from the Cancer Genome Atlas(TCGA)and the Gene Expression Omnibus(GEO)were regarded as the training data set and validation data set,respectively.The list of immune related genes(IRGs)was downloaded from the Immunology Database and Analysis Portal(Immport).Based on the above data sets,the prognostic risk model was constructed and verified by the bioinformatics method.Tumor samples were grouped based on the prognostic risk model to explore the differences in tumor invasion level,tumor mutation burden(TMB)and the expression level of immune checkpoint inhibitor related targets between the high-risk score group and low-risk score group.Results The areas under the ROC curve of the constructed prognostic model were 0.668,0.699 and 0.696 at 3,4 and 5 years,respectively.The low-risk score was significantly correlated with good prognosis(P<0.0001),and the high-risk score was significantly correlated with higher tumor stage(P<0.05),which was further verified in the validation data set.In addition,the c-index of the constructed nomogram model increased from 0.63 to 0.75.Based on the established prognostic risk model,the expression levels of immunotherapy related target genes in the high-risk group were significantly increased(P<0.05),and there was significant difference in the invasion degree of six kinds of immune cells between the high-risk group and the low-risk group(P<0.05).Moreover,higher TMB was significantly correlated with poor prognosis(P<0.05).Conclusion The constructed prognosis model of colon cancer has potential clinical significance in evaluating the recurrence risk stratification and tumor stage of colon cancer patients,which may be helpful to explore the tumor immune microenvironment of colon cancer and immunotherapy of colon cancer patients.