基于N6-甲基腺苷相关lncRNA表达的喉鳞状细胞癌预后分析*

Prognostic analysis of laryngeal squamous cell carcinoma based on the expression of N6-methyladenosine-related lncRNAs

摘要:

目的：探索N6-甲基腺苷（m6A）相关长链非编码核糖核酸(long non-coding Ribonucleic Acid，lncRNA)与喉鳞状细胞癌（LSCC）的预后关系及其临床意义。 方法：获取癌症基因组图谱 （The Cancer Genome Atlas ，TCGA）数据库喉鳞状细胞癌（laryngeal squamous cell carcinoma，LSCC）的转录组数据和临床数据，共表达分析筛选m6A基因相关的lncRNA，单变量cox分析m6A相关lncRNA与LSCC预后关系、套索回归及交叉验证法迭代分析构建LSCC预后模型。采用实时荧光定量聚合酶链式反应 （qPCR） 验证构建预后模型的lncRNA在LSCC中的转录水平。通过预后模型计算风险评分，将LSCC区分为高、低风险病人，高、低风险病人之间进行基因集富集分析（GSEA）。风险评分与浸润LSCC的免疫细胞进行免疫相关性分析。 结果：m6A相关基因与lncRNA的共表达分析筛选出169个与m6A基因相关的lncRNA（相关系数〉0.4，p〈0.001），通过单变量cox分析确定了LSCC预后相关lncRNA：ALOX12-AS1（P<0.05）、LINC00528(P<0.05)、STAG3L5P-PVRIG2P-PILRB(P<0.05)、MNX1-AS1(P<0.01)和LINC02043(P<0.05)。 套索回归、交叉验证迭代分析结果：变量为4时模型的均方根误差最小，即5个预后相关lncRNA仅有ALOX12-AS1、LINC00528、STAG3L5P-PVRIG2P-PILRB与MNX1-AS1可作为模型变量。预后模型：风险评分=（0.176723096228585*MNX1-AS1表达量）+（-0.614916717648596*ALOX12-AS1表达量）+（-0.814201385798827*LINC00528表达量）+（-0.436537752110547*STAG3L5P-PVRIG2P-PILRB表达量）。qPCR 结果ALOX12-AS1（P<0.0001）、LINC00528(P<0.01)、STAG3L5P-PVRIG2P-PILRB（P<0.0001）、MNX1-AS1（P<0.0001）较于癌旁组织，在喉鳞癌组织表达水平上调。GSEA分析结果：高风险病人在细胞-基质粘附（KEGG_FOCAL_ADHESION）（P<0.05）与细胞外基质受体相互作用（KEGG_ECM_RECEPTOR_INTERACTION）（P<0.05）的通路下调，低风险病人在碱基切除修复（KEGG_SPLICEOSOME）（P<0.05）、剪接体（KEGG_SPLICEOSOME）（P<0.05）通路上调。风险评分与免疫浸润细胞相关性分析，风险评分与效应B细胞（Plasma cells）（R=-0.24,P=0.017）、细胞毒性T细胞（T cells CD8）（R=-0.26,P=0.0091）和T滤泡辅助细胞（T cells follicular helper）（R=-0.22,P=0.028）成负相关。 结论：基于m6A相关LncRNA表达量构建的喉鳞状细胞癌预后模型可作为预测LSCC患者预后的有效工具。

Abstract:

Abstract Objective: To explore the prognostic relationship between m6A-related lncRNAs and laryngeal squamous cell carcinoma and its clinical significance. Methods: Transcriptome data and clinical data of laryngeal squamous cell carcinoma (LSCC) were obtained from the Cancer Genome Atlas (TCGA) database, lncRNA associated with the m6A gene was screened by co-expression analysis, and a prognostic model of laryngeal cell carcinoma was constructed by univariate cox analysis of the relationship between m6A-associated lncRNA and LSCC survival, LASSO regression analysis, and iterative analysis of cross-validated methods. Real-time polymerase chain reaction (RT-qPCR) was used to validate the transcript levels of lncRNA, which builds a prognostic model, in LSCC. Risk scores were calculated by a prognostic model to distinguish LSCC into high- and low-risk groups, and gene set enrichment analysis (GSEA) was performed between high- and low-risk groups. Tumor tissues were analyzed for immune infiltration, and the correlation between the risk score value and the degree of infiltrating immune cell infiltration was understood by correlation analysis between the risk score and immune infiltrating cells. Results: Co-expression analysis of m6A-related genes with lncRNA screened 169 lncRNA associated with m6A genes (correlation coefficient > 0.4, p < 0.001), and prognostic associated lncRNA in LSCC was determined by univariate cox analysis: ALOX12-AS1 (P < 0.05), LINC00528 (P < 0.05), STAG3L5P-PVRIG2P-PILRB (P < 0.05), MNX1-AS1 (P < 0.01), and LINC02043 (P < 0.05). Results of lasso regression, cross-validation iterative analysis: The root mean square error of the model was the smallest when the variable was 4, that is, only ALOX12-AS1, LINC00528, STAG3L5P-PVRIG2P-PILRB and MNX1-AS1 could be used as model variables for the five prognostically relevant lncRNAs. The prognostic model: risk score = (0.176723096228585MNX1-AS1 expression) + (-0.614916717648596ALOX12-AS1 expression) + (-0.814201385798827LINC00528 expression) + (-0.436537752110547STAG3L5P-PVRIG2P-PILRB expression).RT-qPCR results showed that ALOX12-AS1 (P < 0.0001), LINC00528 (P < 0.01), STAG3L5P-PVRIG2P-PILRB (P < 0.0001), and MNX1-AS1 (P < 0.0001) were up-regulated in LSCC tissues compared with those in tissues.GSEA analysis results: high-risk KEGGFOCALADHESION（P<0.05） and KEGGECMRECEPTORINTERACTION（P<0.05） pathways were down-regulated, and low-risk KEGGSPLICEOSOME pathway（P<0.05） and KEGGSPLICSOME pathway（P<0.05） were up-regulated.Correlation analysis between risk score and immune infiltrating cells: risk score and plasma cell plasma.Cells (R = -0.24, P < 0.05), CD8 T cells T.Cells.CD8 (R = -0.26, P < 0.01) and T cell follicular helper cells T.Cells.Follicular.Helper (R = -0.22, < 0.05) was negatively correlated. Conclusion: The prognostic model of laryngeal squamous cell carcinoma constructed based on m6A-associated lncRNA expression can be used as an effective tool to predict the prognosis of patients with LSCC.