一种基于炎症反应相关 lncRNAs 的膀胱癌新型分子亚型和风险模型。
A novel molecular subtypes and risk model based on inflammatory response-related lncrnas for bladder cancer.
机构信息
Department of Urology, The Eighth Affiliated Hospital, Sun Yat-sen University, No. 3025, Shennan Zhong Road, Shenzhen, 518033, Guangdong, China.
The Sixth Clinical College of Guangzhou Medical University, Guangzhou, 511436, Guangdong, China.
出版信息
Hereditas. 2022 Aug 13;159(1):32. doi: 10.1186/s41065-022-00245-w.
BACKGROUND
Inflammation and long noncoding RNAs (lncRNAs) are gradually becoming important in the development of bladder cancer (BC). Nevertheless, the potential of inflammatory response-related lncRNAs (IRRlncRNAs) as a prognostic signature remains unexplored in BC.
METHODS
The Cancer Genome Atlas (TCGA) provided RNA expression profiles and clinical information of BC samples, and GSEA Molecular Signatures database provided 1171 inflammation-related genes. IRRlncRNAs were identified using Pearson correlation analysis. After that, consensus clustering was performed to form molecular subtypes. After performing least absolute shrinkage and selection operator (LASSO) and multivariate Cox regression analyses, a risk model constructed based on the prognostic IRRlncRNAs was validated in an independent cohort. Kaplan-Meier (KM) analysis, univariate and multivariate Cox regression, clinical stratification analysis, and time-dependent receiver operating characteristic (ROC) curves were utilized to assess clinical effectiveness and accuracy of the risk model. In clusters and risk model, functional enrichment was investigated using GSEA and GSVA, and immune cell infiltration analysis was demonstrated by ESTIMATE and CIBERSORT analysis.
RESULTS
A total of 174 prognostic IRRlncRNAs were confirmed, and 406 samples were divided into 2 clusters, with cluster 2 having a significantly inferior prognosis. Moreover, cluster 2 exhibited a higher ESTIMATE score, immune infiltration, and PD-L1 expression, with close relationships with the inflammatory response. Further, 12 IRRlncRNAs were identified and applied to construct the risk model and divide BC samples into low-risk and high-risk groups successfully. KM, ROC, and clinical stratification analysis demonstrated that the risk model performed well in predicting prognosis. The risk score was identified as an independently significant indicator, enriched in immune, cell cycle, and apoptosis-related pathways, and correlated with 9 immune cells.
CONCLUSION
We developed an inflammatory response-related subtypes and steady prognostic risk model based on 12 IRRlncRNAs, which was valuable for individual prognostic prediction and stratification and outfitted new insight into inflammatory response in BC.
背景
炎症和长链非编码 RNA（lncRNA）在膀胱癌（BC）的发展中逐渐变得重要。然而，炎症反应相关 lncRNA（IRRlncRNA）作为一种预后标志物在 BC 中的潜力仍未被探索。
方法
癌症基因组图谱（TCGA）提供了 BC 样本的 RNA 表达谱和临床信息，GSEA 分子特征数据库提供了 1171 个炎症相关基因。使用 Pearson 相关分析鉴定 IRRlncRNA。然后，进行共识聚类以形成分子亚型。在进行最小绝对收缩和选择算子（LASSO）和多变量 Cox 回归分析后，基于预后 IRRlncRNA 构建的风险模型在独立队列中进行了验证。 Kaplan-Meier（KM）分析、单变量和多变量 Cox 回归、临床分层分析和时间依赖性接收器操作特征（ROC）曲线用于评估风险模型的临床有效性和准确性。在聚类和风险模型中，使用 GSEA 和 GSVA 进行功能富集分析，使用 ESTIMATE 和 CIBERSORT 分析进行免疫细胞浸润分析。
结果
共鉴定出 174 个预后 IRRlncRNA，406 个样本分为 2 个簇，簇 2 的预后明显较差。此外，簇 2 表现出更高的 ESTIMATE 评分、免疫浸润和 PD-L1 表达，与炎症反应密切相关。进一步，鉴定出 12 个 IRRlncRNA 并应用于构建风险模型，成功将 BC 样本分为低风险和高风险组。KM、ROC 和临床分层分析表明，风险模型在预测预后方面表现良好。风险评分被确定为一个独立的显著指标，富集于免疫、细胞周期和凋亡相关途径，与 9 种免疫细胞相关。
结论
我们基于 12 个 IRRlncRNA 开发了一种炎症反应相关的亚组和稳定的预后风险模型，该模型对个体预后预测和分层具有重要价值，并为 BC 中的炎症反应提供了新的见解。
Zotero 插件