肿瘤中lncRNA融合事件的挖掘及功能研究

Fusion genes resulting from chromosomal rearrangements played important roles in cancers. In recent years, lncRNAs have been revealed to be critical in cancer. Nevertheless, genome-wide long noncoding RNA (lncRNA) fusions and their roles in cancers have been less elucidated. In this study, we developed a computational pipeline and identified genome-wide fusion events in pan cancer, a considerable fraction of which involved lncRNAs. By collecting RNA sequencing data from The Cancer Genome Atlas (TCGA), and lncRNAs from the MiTranscriptome database, we can identify lncRNA fusion events. And through analysis of Breast cancer samples, We found that lncRNAs influenced differential expression, immune response and prognosis through fusion. Similar results could be reproduced in another two cancer types. We have identified abundant fusion events involving 5138 protein-coding genes and 2620 lncRNAs in breast cancer. Our analysis revealed that protein-coding fusion partners of lncRNAs were enriched in cancer-associated processes. Combined with chromatin interaction data, we revealed that fusions were associated with chromosomal spatial organization at whole-genome scale in both cancer and normal cells, and organization in normal cell was more predictive to cancer fusions. These results provide comprehensive insights into fusion features, and reveal the previously unappreciated roles of lncRNA fusions in breast cancer. The pan cancer analysis would revealed specificities of lncRNA fusion landscape. Our study will provided evidence for the notion that lncRNA fusions have functional consequence. Importantly, we will find the lncRNA fusions that lead to long survival time. These findings will provide insights into the previously unappreciated regulatory roles of lncRNA fusions, and will have implications in cancer therapeutic targets. We believe that this is one of the most significant findings on mechanisms of lncRNA function in cancer.

癌症的发病率与死亡率一直处于各种疾病之首。新一代高通量测序技术为研究肿瘤发病机理提供了崭新的手段，越来越多的研究表明，长非编码 RNA（Long non-coding RNA，lncRNA）与癌症发生密切相关，融合基因与多种肿瘤的形成和发展密切相关。目前学术界对LncRNA的基因融合事件及其发生的规律都鲜有研究，为填补这方面研究的不足，本研究以TCGA数据库的肿瘤转录组（RNAseq）数据为研究对象，基于MiTranscriptome数据库的肿瘤lncRNA注释，检测lncRNA所参与的基因融合事件并探索其在肿瘤中的调控作用，挖掘肿瘤中基因融合的特点和规律，同时构建lncRNA融合事件数据库及多组学交互分析平台、构建基于基因融合的肿瘤预测模型。前期分析结果表明：lncRNA通过与蛋白编码基因形成融合，来调控编码基因的转录，从而影响预后和生存期，在癌症发生过程中扮演重要作用。

通过对不同癌症的lncRNA 融合景观进⾏了系统研究，确定了>30000 个的⾼置信度的肿瘤特异的lncRNA 融合事件(采⽤了 8284 个肿瘤样本和 6946 个正常样本)。融合事件与 DNA 损伤和肿瘤⼲性呈正相关，并且在微卫星不稳定(MSI)和受病毒感染的肿瘤中发⽣的数量特别低。另外，融合在不同的肿瘤亚型中的分布不同，但在微卫星稳定(MSS)、体细胞拷贝数 (SCNA)⾼以及⽣存率低的肿瘤中具有共同的富集。发现了⼀种由增强⼦(eRNA) 介导的潜在融合基因⽣成的新机制，这种机制产⽣的次级融合⽹络的中⼼基因拥有多个 FDA 批准的靶向药物。最后，我们通过实验验证了两种来源于mRNA-lncRNA 融合的促肿瘤嵌合蛋⽩（KDM4B-G039927 和 EPS15L1-lnc0R7C2-1）的功能。其中，EPS15L1 融合蛋⽩潜在地可以调控对细胞焦亡和抗肿瘤免疫⾄关重要的GSDME(Gasdermin E)。本项目研究完成了对肿瘤中的融合景观，揭⽰了⼀种新的融合机制，并且丰富了 lncRNA 融合在肿瘤发⽣和癌症发展中的功能。