基于m6A和m5C相关lncRNA的胰腺导管腺癌预后模型及其与免疫微环境的关系

Construction of a prognostic model for pancreatic ductal adenocarcinoma based on m6A- and m5C-related lncRNAs and its relationship with the immune microenvironment

摘要:

背景与目的 胰腺导管腺癌（PDAC）是一种恶性程度高、预后极差的消化系统肿瘤，其早期诊断和治疗仍面临巨大挑战。近年来，RNA甲基化修饰［如N⁶-甲基腺苷（m6A）和5-甲基胞嘧啶（m5C）］在肿瘤发生发展中的作用备受关注，但其在PDAC中的调控机制及临床应用价值尚不明确。本研究旨在筛选与m6A和m5C关联的PDAC预后相关lncRNA，构建可靠的预后预测模型，并探讨其与肿瘤免疫微环境的关系。方法 基于TCGA-PDAC队列的RNA-seq数据，通过差异表达分析和Pearson相关性分析鉴定m6A和m5C相关差异表达lncRNA（DElncRNA）。将样本随机分为训练集（n=89）和验证集（n=89）。采用LASSO-Cox回归筛选关键DElncRNA并构建预后模型，根据风险评分将患者分为高、低风险组。通过Kaplan-Meier生存分析、受试者工作特征曲线（ROC）及多因素Cox回归评估模型的预测效能。进一步结合CIBERSORT算法和ESTIMATE评分分析高、低风险组患者的免疫细胞浸润特征及肿瘤微环境（TME）差异。结果 研究筛选出4个m6A和m5C相关DElncRNA（LINC00857、LINC02038、TSPOAP1-AS1、TRPC7-AS1）构建预后模型。高风险组患者总体生存率显著低于低风险组（P<0.05），且风险评分是PDAC预后的独立预测因素（HR=1.551，95% CI=1.297~1.854，P<0.001）。ROC曲线分析结果显示，风险评分模型在训练集和验证集中均显示高预测效能（1、3、5年曲线下面积分别为0.766、0.875、0.879；0.685、0.711、0.792）。免疫分析显示，高风险组M0巨噬细胞浸润增加，TME评分降低（均P<0.05），提示免疫抑制性微环境特征。结论 本研究成功构建了基于m6A和m5C相关DElncRNA的PDAC预后模型，证实其具有独立预测价值；高风险患者呈现M0巨噬细胞富集及免疫抑制性微环境特征，可能为不良预后的潜在机制。

Abstract:

Background and Aims Pancreatic ductal adenocarcinoma (PDAC) is a highly malignant digestive system tumor with an inferior prognosis, and its early diagnosis and treatment remain significant challenges. In recent years, RNA methylation modifications (such as m6A and m5C) have attracted considerable attention for their roles in tumor development; however, their regulatory mechanisms and clinical significance in PDAC remain unclear. This study was conducted to identify prognosis-related long noncoding RNAs (lncRNAs) associated with m6A and m5C in PDAC, construct a reliable prognostic prediction model, and explore their relationship with the tumor immune microenvironment.Methods Based on RNA-seq data from the TCGA-PDAC cohort, differentially expressed lncRNAs (DElncRNAs) related to m6A and m5C were identified through differential expression analysis and Pearson correlation analysis. The samples were randomly divided into a training set (n=89) and a validation set (n=89). Key DElncRNAs were selected using LASSO-Cox regression to construct a prognostic model, and patients were categorized into high- and low-risk groups based on risk scores. Kaplan-Meier survival analysis, ROC curves, and multivariate Cox regression were used to evaluate the model's predictive performance. Furthermore, CIBERSORT and ESTIMATE scores were used to analyze immune cell infiltration characteristics and tumor microenvironment (TME) differences between the high- and low-risk groups.Results To construct the prognostic model, four m6A- and m5C-related DElncRNAs (LINC00857, LINC02038, TSPOAP1-AS1, and TRPC7-AS1) were identified. Patients in the high-risk group had significantly lower overall survival than those in the low-risk group (P<0.05), and the risk score was an independent prognostic factor for PDAC (HR=1.551, 95% CI=1.297-1.854, P<0.001). ROC curve analysis showed that the risk score model exhibited high predictive efficiency in both the training and validation sets (AUC values for 1, 3, and 5 years: 0.766, 0.875, 0.879; 0.685, 0.711, 0.792, respectively). Immune analysis revealed increased infiltration of M0 macrophages with lower TME scores in the high-risk group (all P<0.05), suggesting an immunosuppressive microenvironment.Conclusion This study successfully established a PDAC prognostic model based on m6A- and m5C-related DElncRNAs and confirmed its independent predictive value. High-risk patients exhibited M0 macrophage enrichment and immunosuppressive microenvironment characteristics, possibly contributing to poor prognosis.