基于肿瘤异质性的耐药亚型识别及候选药物筛选

Drug resistance is one of the main reasons for the failure of cancer treatment, and genetic heterogeneity leads to different molecular mechanisms of drug resistance, which makes drug-resistant tumors more difficult to treat. Therefore, accurate classification of drug-resistant tumor subtypes is an important requisite for precise cancer treatment. In this project, we will develop a novel systems biology method to predict drug-resistant tumor subtypes through integrating drug response, somatic mutation, gene/miRNA/lncRNA expression and transcription factor-noncoding RNA coordinate regulatory network. Furthermore, we aim to design a new measurement to identify the specific subnetwork markers and consequently uncover the specific molecular regulatory mechanisms for each drug-resistant tumor subtype, as well as construct the classification to predict the drug-resistant tumor subtypes based on these subnetwork markers. In addition, based on gene expression profiles of drug perturbation and manually curated information, drug sensitizer will be filtered from the aspects of gene, miRNA and lncRNA, and candidate drug combinations for treating drug-resistant tumors will be proposed. Finally, we will develop the precision medicine platform for drug-resistant tumor and conduct preliminary experiments to verify the above predictions (Taking lung cancer as example). In a word, this project will contribute to elucidate the molecular regulatory mechanisms of drug resistance and provide novel ideas and tools for precise cancer treatment.

药物耐药性是癌症治疗失败的主要原因之一，而患者间的遗传异质性又导致个体产生耐药的分子机制不尽相同，从而使耐药肿瘤的治疗更加困难。因此，准确区分肿瘤的耐药亚型是癌症精准治疗的重要前提。本课题将开发新颖的系统生物学方法，通过整合分析患者的药物反应、体细胞突变、基因/miRNA/lncRNA表达以及转录因子-非编码RNA协同调控网络，预测肿瘤耐药亚型；设计新的测度识别耐药亚型特异的调控子网标志物，解析亚型特异的分子调控机制，并以子网为特征构建耐药亚型分类器；利用药物干扰的基因芯片数据以及文献挖掘的相关信息，分别从基因、miRNA和lncRNA的角度筛选亚型特异的增敏剂，构建候选的治疗耐药肿瘤的联合用药模式；最终，开发耐药肿瘤精准医疗数据平台，并以肺癌为例，对预测的结果进行初步的生物学实验验证。综上，本项目的完成将有助于深入解析肿瘤耐药性复杂的分子调控机制，为癌症的精准治疗提供新的思路和分析工具。

肿瘤细胞的耐药性极大地降低了抗癌药物的疗效，而患者间的遗传异质性又导致个体产生耐药的分子机制不尽相同，从而使耐药肿瘤的治疗更加困难。因此，肿瘤亚型的精确识别、肿瘤分子标志物的预测以及联合用药模式的筛选对于肿瘤耐药分子调控机制的研究以及新型抗癌药物的研发均具有重要意义。首先，本项目在138种药物处理的19种癌症类型的细胞系中识别了耐药性相关的ceRNA（竞争性内源RNA）调控模块，并从泛癌症（pan-cancer）和泛药物（pan-drug）的角度探讨了ceRNA影响耐药性的潜在分子机制；然后，本项目通过分析乳腺癌的单细胞测序数据，重构了乳腺癌的分子亚型，解析了乳腺癌的瘤内异质性，识别了异质性相关的关键调控子；接下来，本项目还提出了基于网络传播和通路交互的耐药性相关miRNA、lncRNA和生物学通路的识别算法；进一步，为了预测治疗耐药肿瘤的候选药物，本项目构建了影响lncRNA表达的小分子化合物数据库D-lnc，提出了逆转失调基因表达的药物筛选策略；此外，本项目还从DNA甲基化、m6A修饰、非编码RNA编码肽、铁死亡非编码RNA等方面进行了扩展性研究，为耐药肿瘤的多组学数据分析奠定了基础。最后，本项目在顺铂和阿霉素耐药的细胞系和小鼠模型中验证了耐药基因的作用机制及联合用药对耐药肿瘤的抑制作用。. 课题组已经发表标注本项目资助的SCI论文11篇；编写生物信息学教材1部；培养博士研究生7人（已毕业1人），硕士研究生11人（已毕业5人）；承办国内学术会议4次；获软件著作权6项。本项目的研究成果有助于揭示肿瘤耐药的分子机制，解析肿瘤耐药的遗传异质性，为耐药肿瘤的精准治疗提供科学依据。