健康与遗传心理学研究室知识库

阿尔茨海默病（Alzheimer’s disease，AD）是一种由于中枢神经系统紊乱导致的神经退行性疾病。近数十年来对AD 机制的研究已经取得了显著的成就，通过遗传研究发现AD 是一种多基因复杂疾病，受到多种微效基因的影响。但是，AD 的病理进程仍不明确。基因表达研究在阐释基因的功能方面具有非常重要的作用，为阐释疾病的致病分子进程提供重要的线索。到目前为止，已经积累了丰富的关于AD 患者解剖脑的全基因组表达谱数据。但是，如何从这大量的数据中挖掘出AD 相关的风险基因及分子过程，是目前AD 机制研究中所面临的挑战之一。
因此，为了帮助阐释AD 相关的内在分子机制，本研究对包含76 例AD 患者和40 例健康对照者的三套海马区的全基因组表达谱数据进行整合分析。首先，我们构建线性混合效应模型控制一些协变量因素的影响（包括年龄，性别，解剖时间和批次效应），最终计算得到80 个在AD 病例和健康对照组中存在表达差异的基因，我们进一步对这80 个差异表达基因进行验证，包括jackknife 交叉验证，与原始数据集和AlzBase 数据库的结果进行比较，进而发现有八个下调的差异表达基因（GAD2，RPH3A，GAD1，SST，GABBR2，NUDT11，DLGAP2 和PCLO）和一个上调的差异表达基因（ITGB5）被这三种方法所共同验证。进一步地，为更好地了解差异表达基因所行使的功能及其相互作用关系，本研究对80 个差异表达基因进行了通路富集分析和蛋白-蛋白相互作用网络分析。通过通路富集分析得到了一条显著富集的KEGG 通路—“GABAergic synapses”和22 条GO 术语，涉及的GO 主要包含神经元、分泌小泡、突触信号、突触传递、细胞连接以及突触囊泡代谢等生物过程，这些过程都与阿尔茨海默病的病理特征存在一定的联系，表明GABA 能系统、神经元及突触相关的功能可能在阿尔茨海默病的发病机制中受到显著的影响。我们基于80 个差异表达基因所编码蛋白质的相互作用关系，得到了180 个与差异表达基因存在功能连接的拓展基因，并构建了蛋白-蛋白相互作用网络，其中，位于网络中心位置的核心基因是CDC42，该差异表达基因被jackknife 交叉验证方法和AlzBase 数据库所验证。最后，基于收敛功能基因组学（Convergent Functional Genomics，CFG）方法，我们对80 个差异表达基因和180 个拓展基因进行排序，进而探索得到多证据支持的候选基因，结果发现分值最高的是两个拓展基因ETS1 和PAK4。这些基因在AD 的致病过程中可能发挥重要的作用，值得我们进行深层次的关注和探索。
综上所述，本研究基于生物信息学方法分析AD 相关的表达谱数据，获得差异表达基因及其生物学功能，并进一步得到多证据支持的候选基因，为阿尔茨海默病的后续研究提供可靠的证据支持

Alzheimer’s disease (AD) is a neurodegenerative disease caused by neurological disorders. In recent decades, remarkable achievements have been made in the study of AD mechanisms. Genetic studies have found that AD is a complex disease affected by a variety of genes with small effects. However, the pathogenesis of AD remains unclear. Gene expression studies play an important role in explaining the function of genes and providing significant clues for the interpretation of the biological mechanisms of diseases. So far, there was a wealth of genome-wide expression profiling performed on the postmortem brain tissues of AD patients. However, how to dig out AD-related risk genes and molecular processes from this large amount of data is one of the challenges faced in the current research on AD mechanisms.
Therefore, to characterize intrinsic molecular processes related to Alzheimer’s disease (AD), a combined analysis of three genome-wide expression profiling in the hippocampus containing 76 AD patients and 40 normal controls was performed. Firstly, the effects of covariates (including age, gender, postmortem interval (PMI) and batch effect) were controlled using a linear mixed-effects model. 80 differentially expressed genes (DEGs) between the AD patients and age-matched controls were best identified, then validated these genes from three aspects, including jackknife cross-validation, an original study comparison and an AlzBase comparison. Eight down-regulated DEGs (GAD2, RPH3A, GAD1, SST, GABBR2, NUDT11, DLGAP2 and PCLO) and one up-regulated DEG (ITGB5) were all validated. Furthermore, to better understand the functions and interactions of these DEGs, functional pathway enrichment analysis and protein-protein interaction (PPI) network were performed on the 80 DEGs. In the pathway enrichment analysis, the DEGs were significantly enriched in one KEGG pathway, GABAergic synapses, and 22 GO terms; the main GO function of these genes mainly involved in neuron, secretory vesicle, synaptic signaling, synaptic transmission, cell junction, and synaptic vesicles metabolism. These processes were all linked to the pathological features of Alzheimer's disease, demonstrating that GABAergic system, neurons, and synaptic function might be significantly affected in the pathogenesis of AD. Based on the interactions of the proteins encoded by 80 DEGs, 180 extended genes that interact with the DEGs were obtained, and then the protein-protein interaction network was constructed. The hub gene occupied in the most central position in this study was CDC42, which was validated by jackknife cross-validation and an AlzBase comparison. Finally, using the convergent functional genomics (CFG) method, the 80 DEGs and 180 extended genes were ranked. We found that the two genes with the highest scores were ETS1 and PAK4, both of which were the extended genes in the network. These genes might play an important role in the pathogenesis of AD, deserving our deep attention and exploration.
In conclusion, gene expression profiling studies of AD were analyzed using bioinformatics methods. Differentially expressed genes and biological functions were obtained, and candidate genes supported by multiple evidences were also acquired, which provided reliable evidence for the follow-up research in AD.