基于网络推断的功能性异常可变剪接体识别及其在血液系统肿瘤中的应用研究

Spliceosomal gene mutations are frequently observed in hematologic malignancies and contribute to disease initiation and development. In the proposed project, we will use transcriptomes from SF3B1(K700E), U2AF1(S34F) and SRSF2(P95H) mice as an entry point to systematically identify functional dysregulated splicing isoforms by network-based inference, and validate results with experiments and clinical sequencing data. First, we will computational reconstruct a gene regulatory network using multi-conditional expression profiles of hematopoietic stem/progenitor cells. Leveraging the gene regulatory network and mutant mouse transcriptomes, we are going to infer the regulators with changed regulatory patterns. Furthermore, molecular annotations of splicing isoforms will be integrated to identify the dysregulated isoforms whose expressions are the major contribution to the change of regulatory patterns. The functions of these isoforms will be experimental validated in vitro. Finally, we will compare the composition of dysregulated isoforms among the three spliceosomal gene mutations. Meanwhile, the prevalence and prognostic significance of identified dysregulated isoforms will be determined in patients through clinical sequencing data. The proposed project will improve our understanding of the pathogenesis of splicing gene mutations and promote the discovery of new drug targets. It will also provide a new method to study other human diseases involved in splicing isoforms.

剪接子的编码基因在血液系统肿瘤中存在高频率突变，与疾病的起始和发展密切相关。本项目将以SF3B1(K700E)、U2AF1(S34F)和SRSF2(P95H)三种剪接子突变模型鼠转录组为切入点，通过网络推断系统地识别功能性异常可变剪接体并通过实验和临床测序数据进行验证。首先，利用多种条件下的造血干/祖细胞表达谱，计算重建基因调控网络。借助该调控网络和模型小鼠的转录组，推断调控模式变化的表达调控基因。进一步，对于选出的每个调控基因，整合可变剪接体的分子注释，识别对调控模式变化有主要贡献的异常可变剪接体并通过体外实验验证。最后，我们将比较这三种剪接子突变对应的异常可变剪接体之间的异同点，并用临床测序数据评估异常可变剪接体在患者中的分布和预后意义。本项目的实施将促进对剪接子基因突变致病机制的理解和新药靶点的发现，也为研究与可变剪接体有关的其它人类疾病提供一种新的方法。

可变剪接子编码基因SF3B1、SRSF2和U2AF1在血液系统肿瘤中频繁发生突变，导致肿瘤细胞中mRNA前体的剪接异常，出现很多在正常细胞中不表达或者低表达的异常mRNA可变剪接体。鉴定功能性的异常可变剪接体对于理解这些突变的生物学效应，寻找新的治疗靶点具有重要意义。本项目开发了一个新的生物信息学方法，通过结合差异可变剪接信息和基因表达调控网络的扰动去识别功能性的异常可变剪接体。我们将方法运用于分析SF3B1、SRSF2和U2AF1突变的小鼠模型以及骨髓增生异常综合征患者的RNA-seq数据，成功识别了已知的异常可变剪接体，而且发现来自ATRX，APC和ASXL2等基因的异常可变剪接体可能和突变的致病性有关。尽管SF3B1、SRSF2和U2AF1所导致的异常可变剪接体并不相同，但它们都干扰了RUNX1控制的转录调控子网络。我们构建了SF3B1突变的K562细胞并进行ATAC-seq测序和转录因子足迹分析，验证了我们的发现。进一步，结合82位骨髓增生异常综合征患者的生存数据，我们识别了表达水平与患者预后显著相关的异常可变剪接体，为临床研究提供了潜在的生物学标志物。