基于转录组数据的糖尿病肾病系统生物学研究

The incidence of diabetic nephropathy (DN) increased year by year and it has became an important disease that influence China's population health. Since the lack of knowledge about the molecular mechanisms of DN, there is no effective prognostic biomarker, and the intervention methods in clinical are still inadequate until now. Therefore, the key issues in this field was to explore more molecular mechanisms and to search for effective prognostic biomarkers and therapeutic drugs. Based on the whole genome gene expression profiles of the glomerular from the earlier work, we will utilize Weighted Gene Co-expression Network Analysis ( WGCNA ) to construction a weighted gene network and to identify modules, analyze the association between the modules and the clinical trait, and reveal the genes and pathways which are responsible for the initiation and progression of the DN. We will utilize the lasso regression analysis and the stepwise regression analysis to find the DN prognostic molecular biomarkers. The 1-year decline percentages in estimated glomerular filtration rate (eGFR) were used as a response variable, while the whole genome gene expression levels were used as predictive variables. The molecular biomarkers will be validated in the kidney tissue slices. We will utilize the Connectivity Map (CMAP) database and the bioinformatics methods to search for candidate drugs, then we will validates them in the mouse models of DN. This research will utilize systems biology approaches to investigate the whole genome gene expression profiles in order to explore the molecular mechanisms and key molecules of DN and seek effective prognostic biomarkers and candidate drugs, which will help us to develop promising strategy for diagnosis, prevention and treatment.

糖尿病肾病（DN）的发病率逐年升高，已成为影响我国人口健康的重要疾病，但目前临床缺乏有效的预后分子标志物和基于分子机制的干预药物。本课题将在前期研究获得DN患者肾小球转录组数据基础上，构建基因共表达网络并识别模块，通过与临床指标关联分析，寻找与DN发生发展密切相关的模块，进一步分析这些模块中基因的通路和功能，阐明DN发生发展的分子机制；以DN患者eGFR年下降幅度为响应变量，以肾小球基因表达数据为预测变量，通过lasso回归和逐步回归法，寻找DN预后分子标志物，并在病理标本中验证；利用CMAP（The Connectivity Map）数据库和生物信息学方法寻找DN候选药物，并在DN小鼠模型中初步验证。本课题基于转录组数据，利用系统生物学方法探索DN发生和发展的分子机制，并寻找可能用于临床的预后标志物和治疗药物，有助于发展更好的DN诊断、预防和治疗策略。

糖尿病肾病（diabetic nephropathy, DN）是糖尿病（diabetes mellitus, DM）最常见的并发症，也是糖尿病患者的主要死亡原因之一。DN已成为导致我国终末期肾功能不全（End stage renal diseases，ESRD）的最主要继发性肾脏病之一。. 本课题在DN病人肾小球转录组数据和临床数据的基础上，运用生物信息学方法深入挖掘这些数据之间的内在联系和相互关系，揭示DN的关键分子及其调控机制，寻找预后标志物和治疗药物。. 建立糖尿病肾病（DN）肾小球基因共表达网络：根据基因共表达关系识别模块，将基因模块与DN患者临床指标作相关性分析，得到与临床指标密切相关的模块。Turquoise、Green 和Red 模块与DN 患者的蛋白尿和eGFR以及eGFR 年下降百分比高度相关。GO 分析和KEGG 通路分析发现Turquoise 模块主要和细胞骨架相关（其中包括很多足细胞特异性表达的基因），Red 模块主要和免疫相关，Green 模块主要和细胞外基质相关。. SRGAP2是Turquoise模块中的一个枢纽基因，在DN患者中下调，与患者蛋白尿负相关（R = -0.81，P = 9.61 ×10-15）。免疫荧光染色和Western blot表明SRGAP2主要在肾小球足细胞中表达。SRGAP2能够通过结合RhoA/Cdc42抑制足细胞迁移保护足细胞。. 利用DN患者基线肾小球转录组数据和病人随访得到疾病进展指标，通过lasso回归和逐步回归法，我们找到了DN的预后分子标志物PTH1R，在独立队列中得到验证。.转录组数据进行基因差异表达分析后，得到DN晚期组与早期组之间的差异表达基因，利用CMAP数据库 和生物信息学方法筛选出荜茇酰胺（piperlongumine）、15d-PGJ2（15-脱氧前列腺素J2）、伏立诺他（vorinostat）和曲古抑菌素A（trichostatin A）等候选药物。. 本课题是转录组学和生物信息学在复杂疾病研究中的应用，得到的结果将进一步揭示DN分子发病机制，发展更好的DN诊断、预防和治疗策略。