基于GWAS的肺癌预后相关miRNA协同调控网络的识别及其调控机制研究

Lung cancer is by far the leading cause of cancer-related mortality in most populations worldwide. Effective molecular markers are the keys to develop targeted therapy and prognostic prediction for lung cancer. miRNAs, which synergistically target to mRNA sequences, are associated with tumor biology. In this study we analyze miRNA synergistic network and the relationship between lung cancer survival and miRNA SNPs based on GWAS and expression profile data. First, gene sets are used to identify prognosis related miRNA SNPs in GWAS data, then, miRNAs and mRNA expression profiles in lung cancer tissues are used to identify survival associated miRNA target–dysregulated network and construct the support vector machine classifier. The SVM classifier is applied to prioritize lung cancer prognosis related miRNAs. After that, a genetic risk score (GRS) of miRNA SNPs is conducted by the methods of principal component analysis and random survival forests. Then we demonstrate the utility of miRNA-GRS to predict survival time in another lung cancer population. miRNA synergistic network are identified by integrating analysis of the co-expression relationship and the predicted targets information. Flow cytometry is used to detect the cell apoptosis and cell cycle, and real-time PCR are employed to measure expressive levels of targets after up-regulation of miRNAs using miRNA mimics together or respectively. The pmirGLO Dual-Luciferase Vector contains 3′UTR cloned of predicted targets are used to analyze miRNA co-targets. Results of this study will provide theoretical basis and practical guidance for prognostic prediction and targeted therapy of lung cancer.

肺癌死亡率位于恶性肿瘤死因首位。研究预后相关遗传因素，筛选有效分子标志物是开展肺癌预后预测和靶向治疗的必要手段。miRNA通过协同作用参与肿瘤发生发展的生物过程，是潜在的预后分子标志物。本研究拟整合GWAS和表达谱数据优化筛选预后相关miRNA，分析miRNA SNPs对肺癌预后的影响及miRNA协同调控作用。首先利用基因集分析GWAS数据中miRNA SNPs与肺癌预后相关性，利用表达谱分析miRNA-mRNA失调控关系，建立miRNA分类器优化筛选预后相关miRNA；进一步利用主成分分析等方法建立miRNA-GRS肺癌预后模型，评价并验证其预测效果；进而利用miRNA共表达和靶基因调控关系识别肺癌预后相关miRNA协同调控网络，利用相关实验探讨miRNA协同调控网络对靶基因表达、细胞凋亡侵袭等能力的影响及其机制。研究结果将为肺癌的预后诊断和肺癌靶向性治疗提供理论和实验依据。

肺癌是发病率和死亡率均居首位的肿瘤。深入研究影响肺癌预后的个体遗传因素，筛选有效的肺癌预后分子标志物，是开展肺癌预后预测和靶向治疗的必要手段。通过生物信息学方法在全基因组范围内寻找可能影响miRNA功能的多态位点，利用基因表达谱分析影响预后的基因调控网络变化，进而从功能变化结果入手分析GWAS的miRNA SNPs数据并利用系统生物学方法进行优化筛选，将有助于优化肺癌预后相关遗传标志，使基于miRNA的肺癌预后风险预测成为可能。本研究利用生物信息学基于核函数的机器学习方法按基因集在全基因组范围内筛选肺癌预后关联SNPs，使用SNAP数据库提取预后关联位点的LD关联位点，利用主成分分析等及机器学习的方法筛选肺癌相关mRNA-SNPs，识别了miR-499 rs3746444、HMGB1 rs1412125、MIAT rs1061541等肺癌相关的SNP位点。进一步利用系统综述和病例对照设计的实验验证方法对筛选肺癌相关的miRNAs及其SNP位点标志物并进行了验证。研究利用TCGA、GEO等公开数据库筛选肺癌预后相关差异表达基因，并进行预后生存分析，同时结合课题组收集的历史队列的随访信息，分别构建了用于肺癌预后预测和脑转移预测的列线图（nomogram），提示了较好的预测诊断价值。为探索miRNA对肺癌的作用机制，研究进一步利用miRNA与mRNA共表达数量关系，构建失调控网络对miRNA进行模块化功能分析，并利用细胞转染实验初步验证mir-133a-5p和mir-139-5p两个miRNA对相关基因表达的调控作用。同时尝试利用ceRNA机制对可能影响miRNA协同调控关系的机制进行了探索，初步识别了MIAT可能作为ce RNA，通过与miR-133a-5p结合而影响下游mRNAs (MYO1B, SGK1 及 WNT9A)的表达。研究实现了宏观的流行病学方法与微观的分子生物学相结合，生物信息学的计算和预测与功能学实验的验证相结合。研究按照计划进行，基本实现了研究目标。同时在原计划基础上，进一步尝试了miRNA协同调控关系的分析思路在ce RNA竞争机制上的拓展。研究结果如能向临床转化，将对优化非吸烟女性肺癌患者的预后判断和靶向分子治疗方式的选择提供有益的数据支持。