【作者】 王晓东；

【导师】 姜宗来；

【作者基本信息】 上海交通大学 ， 生物医学工程， 2014， 博士

【摘要】 血管重建（remodeling）是很多心脑血管疾病，如高血压、动脉粥样硬化、脑卒中等的共同发病基础以及其病理的基本过程。研究表明，机械应力诱导的血管重建在动脉粥样硬化的发生、发展过程中起着非常重要的作用。探索机械应力诱导动脉血管重建的分子机制，寻找生理和病理过程中关键的标记分子和潜在的药物靶标，对于动脉粥样硬化及其相关疾病的病理研究及临床治疗具有非常重要的意义。本研究首先在低切应力（5dyn/cm2）与正常切应力（15dyn/cm2）条件下分别体外培养大鼠血管，应用双向电泳结合质谱分析的方法，鉴定出血管组织具有显著性表达差异的43个蛋白质。对这些蛋白采用基于Gene Ontology和IPA的功能与信号通路分析，预测了4个切应力相关的信号网络，发现了LaminA，LOX，RACK1，Rab28等新的力学敏感分子。这些分子在应力诱导血管重建中的作用及其分子机制尚未见到研究报道。然后，我们采用平行平板流动腔系统对联合培养的内皮细胞（endothelial cells，ECs）与平滑肌细胞（vascular smoothmuscle cells，VSMCs）加载切应力，应用western blotting方法检测不同水平切应力条件下细胞内LaminA和LOX表达。结果证实，LaminA和LOX为力学刺激敏感分子，低切应力抑制ECs和VSMCs的LaminA表达，而促进LOX表达。应用RNA干扰和重组蛋白刺激等分子生物学方法，验证了IPA预测的切应力信号通路网络中与LaminA和LOX相关的部分关键节点表达及其调控关系。由于绝大多数信号分子是通过磷酸化／去磷酸化修饰而改变其活性来发挥作用，我们又应用稳定同位素标记氨基酸活细胞培养（SILAC）、色谱分离磷酸化蛋白和质谱技术，得到了周期性张应变加载条件下，VSMCs的磷酸化蛋白质随时间的表达变化及相应的磷酸化位点。基于上述数据，应用聚类分析和功能分析等生物信息学与计算生物学方法，分析了张应变条件下VSMCs蛋白质磷酸化的表达模式，以及在不同时间点差异表达的磷酸化蛋白质所参与的细胞功能和信号通路，并构建了VSMCs在张应变刺激下的动态磷酸化信号网络。研究结果除揭示了新的、可能参与应力细胞内传导的信号途径外，还提示，细胞骨架结构可能直接感受应力刺激并调控多种结合（binding）蛋白活性，进而调控细胞功能，为血管细胞将细胞外机械应力信号转化为细胞内生物化学信号的“应力感受器”研究提供了新的依据。为进一步深入分析上述蛋白质组学数据，揭示更多的、有价值的机械应力作用下血管细胞内信号网络信息，尤其是直接调控基因表达的转录因子信息，我们探索了应用大规模基因微阵列数据推断转录调控网络的新方法，并应用于我们的蛋白质组学数据分析及力学信号网络构建。首先提出了平均三元互信息（AMI3）与网络辅助回归分析算法，这一从大规模基因微阵列芯片数据重建基因调控网络的新方法，并应用大肠杆菌、酵母菌和人的B细胞微阵列数据，对该方法的准确性和可靠性进行验证。然后，我们构建了人脐静脉ECs的转录调控网络，并结合低切应力条件下ECs的基因微阵列数据与血管细胞的蛋白质组学数据，应用AMI3结合网络辅助回归分析，构建了ECs响应切应力的信号转导与转录调控耦合网络，更全面地描述了机械应力在血管细胞内从信号转导到基因表达的调控过程。然后，我们提出网络中节点之间的图元相互作用的概念，并预测了心血管疾病相关基因。研究发现，疾病基因之间的图元相互作用与普通基因之间具有明显差异。根据这种差异用图元相互作用构造新的得分指数，可用来鉴定疾病基因。我们基于OMIM数据库中已知疾病基因信息，采用弃一法交叉检验，发现基于图元相互作用的疾病基因鉴定方法与之前的随机游走和Endeavour等方法相比，具有更高的准确度和稳定性。最后，我们将经过验证的上述方法用于预测高血压和冠状动脉粥样硬化相关的基因，得到了NOS2与VEGFR1等12个新基因。上述研究基于蛋白质组学与计算生物学方法，构建机械应力作用下血管细胞内信号转导与转录调控网络，发现了新的可能参与机械应力感受、细胞内信号转导和转录调控的分子，并应用分子生物学方法对其中部分分子的表达水平和相互关系进行验证。本研究为全面、深入了解应力调控血管重建的分子机制提供了新的研究手段和研究方向，为动脉粥样硬化等心血管疾病的诊断和药物治疗提供了潜在靶向。更多还原

【Abstract】 Vascular remodeling is the crucial pathological process during manydiseases such as hypertension and artherosclerosis. Studies have shownthat mechanical stress played an important role in the onset anddevolopment of vascular remodeling. It is significant to explore themolecular mechanism of the vascular remodeling and discover the keybiomarker and drug targets for the pathological research and clinicaltreatment of artherosclerosis.Using two dimensional electrophoresis (2-DE) and massspectrometry,43differentially expressed proteins between vasculartissues cultured under between normal (15dyn/cm2) and low (5dyn/cm2)shear stress were identified. Gene ontology and IPA were used to analyzethe functions and signal pathways of the above defferencially epressedproteins, and4probable signal networks related to shear stress wereinfered. Lamin A, LOX, Rack1, Rab28, etc were found as newmechano-sensitive proteins, whose expression and mechanism in themechano-induced vascular remodeling had not been reported. Then, theshear stress was exerted to the co-cultured vascular smooth muscle cells (VSMCs) and endothelial cells (ECs) by flow chamber system. Westernblotting was used to detect the expression of Lamin A and LOX underdifferent shear stress, and the result show that low shear stress inhabitedthe expression of Lamin A and increase the expression of LOX in ECsand VSMCs. Afterwards, RNA interference and recombination proteinwere used to verified some key nodes and edges related to Lamin A andLox in the predicted network.Phosphorylation/dephosphorylation cycling is one of the mostimportant modifications that regulate the activation of signal molecules.Therefore, in the following research, the time series profiles ofphosphorated proteins and the coresponding phosphorylated sites inVSMCs subjected to cyclic strain were investigated by stable isotopelabelling amino acid in cell culture (SILAC), chromatogram and massspectrometry. Then, the clustering, function and pathway analysis wereapplied to analyze the expression model, GO functions, signal pathwaysand phosphorylated network of the differential expressed phosphorylatedproteins in different time points. The results discovered new signalpathway which maybe participated in the mechano-transduction, andsuggested that cytoskeleton may directly sense mechanical stress,regulate many binding protein, and regulate cell function. It offered a newdirection of the research of mechano-sensor which transferredextracellular mechanical stress into intracellular biochemical signal. In order to further analyze the above proteomic data, discover morevaluable mechano-transduction network, especially the information oftranscription factors which regulate gene expression, we investigated newmethods, applied from large-scale microarray data, and used it to discovertranscriptionalnetwork based on the above proteomic data. First, weintroduced the averaged three-way mutual information (AMI3) andnetwork assisted regression algorithm, a new approach which could infertranscriptional network from large-scale microarray. The algorithm wasverified by E.coli and S.cerevisiae and human B cell data. Then, applyingthe AMI3and network assisted regression algorithm, the transcriptionalnetwork of human umbilical vein ECs was constructed, and thetranscriptional combined with signal transduction network in ECs werealso constructed based on the microarray and2-DE data of vascular cellsstimulated by low shear stress. This network, including both signaltransduction to transcriptional regulation, systematically describedmechano-transduction process of vascular cells.Furthermore, the graphlet interaction was introduced andcardiovascular disease genes were predicted. Our study found that thegraphlet interaction between disease genes and normal genes weresignificantly different. Accordingly, new score was calculated based onthe graphlet interaction to identify disease genes. Leave-one-outcross-validation was applied to evaluate the performance based on known disease genes in OMIM database. Compared with random walk andEndeavour, graphlet interaction obtained higher precision and stability.Using this method,12new cardiovascular diseases genes, such as NOS2and VEGFR1, were predicted.In summary, using proteomic analysis combined with computationalmethods, we constructed signal transduction and gene regulatory networkof vascular cells respond to mechanical stress, discovered new moleculeswhich were probable participated in mechano-transduction. Usingmolecular and biological experiments, part of the predicted molecules andthe regulatory relation were verified. This study provided new researchapproaches and direction for investigating the molecular mechanism ofvascular remodeling induced by mechanical stress, and suggestedpotential biomarker for diagnosis and treatment of cardiovasculardiseases including atherosclerosis.更多还原