南方医科大学学报

南方医科大学学报 ›› 2022, Vol. 42 ›› Issue (5): 681-689.doi: 10.12122/j.issn.1673-4254.2022.05.08

• • 上一篇    下一篇

m7G相关lncRNAs是影响结肠癌患者预后和肿瘤微环境的潜在生物标志物

陈曙冉，董 锐，李 艳，吴华彰，刘牧林   

  1. 蚌埠医学院第一附属医院胃肠外科，生命科学学院癌症转化医学安徽省重点实验室，第一附属医院妇科肿瘤，安徽 蚌埠 233000
  • 出版日期:2022-05-20 发布日期:2022-06-02

m7G-lncRNAs are potential biomarkers for prognosis and tumor microenvironment in patients with colon cancer

CHEN Shuran, DONG Rui, LI Yan, WU Huazhang, LIU Mulin   

  1. Department of Gastroenterology, the First Affiliated Hospital of Bengbu Medical College, School of Life Science, Anhui Province Key Laboratory of Translational Cancer Research, Department of gynecologic oncology，the First Affiliated Hospital of Bengbu Medical College Bengbu Medical College, Bengbu 233000, China
  • Online:2022-05-20 Published:2022-06-02

RichHTML

39

PDF

776

摘要： 目的 探讨m7G-lncRNAs能否作为结肠癌患者预后及肿瘤微环境的生物标志物。方法 TCGA数据库筛选m7G-lncRNAs（|Pearson R|>0.4，P<0.001），多因素Cox分析构建m7G-lncRNAs风险模型。使用ROC和C-index曲线对风险模型进行验证。构建诺莫图和诺莫图的校准曲线用于预测结肠癌患者的预后。点柱图和K-M生存曲线评估风险打分对患者临床分期和预后的影响。CIBERSORT和ESTIMATE探究高低风险组患者肿瘤微环境和免疫细胞浸润程度的联系，同时分析风险打分对结肠癌患者微卫星不稳定性，干细胞指数和免疫检查点表达的影响。使用相互作用基因搜索工具（STRING）构建蛋白质-蛋白质相互作用网络，挖掘m7G-lncRNAs调控的关键靶点。最后，使用蛋白印迹实验在4对结肠癌组织与癌旁正常组织中验证这些关键靶点的表达。结果 从TCGA数据库鉴别出1722个m7G-lncRNAs。多因素Cox分析筛选出12个lncRNAs用于构建风险模型，其中AC003101.2、AC005014.2、AC008760.1、AC092944.1、AL1161729.4、AL301422.4、AP001619.1、AP003355.1和ZEB1-AS1为高风险lncRNAs，AC025171.4、AC073957.3及TNFRSF10A-AS1为低风险lncRNAs。ROC曲线显示风险模型对患者1年、3年、5年生存预测的AUC值分别为0.727、0.747、0.794。诺莫图预测患者预后的AUC值为0.794，校准曲线显示诺莫图对患者生存的预测与患者实际的生存基本一致。高风险组患者的T分期（T1~T2 vs T3~T4：P=0.034）、N分期（N0 vs N2：P=7.8e-08； N1 vs N2：P=0.00081）以及M分期（M0 vs M1：P=0.007）均高于低风险组患者。低风险组患者常伴随高微卫星不稳定状态（MSS vs MSI-H：P=0.034）。肿瘤干性指数与风险得分呈负相关（r=-0.19；P=7.3e-05）。高风险组患者基质细胞打分（P=0.0028）以及总打分（P=0.007）明显高于低风险组患者较高，激活的肥大细胞（r=-0.11；P=0.045）和静息CD4+T细胞（r=-0.14；P=0.01）的表达也较低。多数免疫检查点在高风险患者中高表达（P<0.05）。蛋白印迹实验表明m7G-lncRNAs调控的关键靶点ATXN2 （P=0.006）and G3BP1（P=0.007）在4对结肠癌组织中表达均高于配对的癌旁正常组织。结论 12个m7G-lncRNAs构建的风险模型对结肠癌具有重要的预后价值，同时也能反映结肠癌患者肿瘤微环境及免疫治疗的疗效。

关键词: 结肠癌；m7G；lncRNAs；预后模型；肿瘤微环境

Abstract: Objective To assess the value of m7G-lncRNAs in predicting the prognosis and microenvironment of colorectal cancer (CRC). Methods We screened m7G- lncRNAs from TCGA to construct an m7G-lncRNAs risk model using multivariate Cox analysis, which was validated using ROC and C-index curves. Calibration and nomogram were used to predict the prognosis of CRC patients. Point-bar charts and K-M survival curves were used to assess the correlation of risk scores with the patients' clinical staging and prognosis. CIBERSORT and ESTIMATE were used to explore the association between the tumor microenvironment and immune cell infiltration in patients in high and low risk groups and the correlation of risk scores with microsatellite instability, stem cell index and immune checkpoint expression. A protein-protein interaction network was constructed, and the key targets regulated by m7G-lncRNAs were identified and validated in paired samples of CRC and adjacent tissues by immunoblotting. Results We identified a total of 1722 m7G-lncRNAs from TCGA database, from which 12 lncRNAs were screened to construct the risk model. The AUCs of the risk model for predicting survival outcomes at 1, 3 and 5 years were 0.727, 0.747 and 0.794, respectively. The AUC of the nomogram for predicting prognosis was 0.794, and the predicted results were consistent with actual survival outcomes of the patients. The patients in the high- risk group showed more advanced tumor stages and a greater likelihood of high microsatellite instability than those in the low-risk group (P<0.05). The tumor stemness index was negatively correlated with the risk score (r=- 0.19; P=7.3e-05). Patients in the high-risk group had higher stromal cell scores (P=0.0028) and higher total scores (P=0.007) with lowered expressions of activated mast cells (r=-0.11; P=0.045) and resting CD4+ T cells (r=-0.14; P=0.01) and increased expressions of most immune checkpoints (P<0.05). ATXN2 (P= 0.006) and G3BP1 (P=0.007) were identified as the key targets regulated by m7G-lncRNAs, and their expressions were both higher in CRC than in adjacent tissues. Conclusion The risk model based on 12 m7G-lncRNAs has important prognostic value for CRC and can reflect the microenvironment and the efficacy of immunotherapy in the patients.

Key words: colon cancer; m7G; long non-coding RNAs; prognostic model; tumor microenvironment