Topographic, cell type and molecular pathway characterization ofAlzheimer's disease using single cell transcriptomics and epigenomics

摘要（项目2） 阿尔茨海默病（AD）是一种高度渗透性的神经退行性疾病，预计影响1380万人 到2050年，如果不开发治疗方法，美国将花费1.1万亿美元。AD的特征是刻板的 进行性神经变性和两种错误折叠的蛋白质在大脑区域的积累， 认知和记忆。过度磷酸化tau蛋白的神经元缠结（NFT）遵循以下进展， 神经变性，而细胞外淀粉样蛋白β（Aβ）斑块最初在皮质和深部脑中检测到 结构.目前尚不清楚这些病理是因果关系还是其他潜在过程的影响， 目前没有抗tau或抗Aβ疗法停止或逆转AD。AD的基因表达研究主要是 在组织或细胞群上进行，并且神经元损失和神经胶质增生对这些结果的影响是未知的。 表观遗传修饰也与AD相关，尽管甲基化研究产生了相互矛盾的结果， 结果表明，没有出现与AD进展直接相关的表观遗传失调的明确模式。 本研究采用了最近开发的高通量，单细胞的转录组学方法， 表观遗传分析，以确定“临床典型”AD的分子和基因调控标志， 显著的合并症。建立在对神经典型成体细胞类型的详细了解基础上，该项目旨在 鉴定与AD严重程度增加相关的特定细胞类型或类别中的转录变化 受疾病影响的不同大脑区域的病理学，然后确定基因和染色质的可及性 脆弱细胞群的病理变化。该项目将初步优化单核RNA-seq 和表观遗传学方法用于不同病理学和组织质量的死后样品，并产生 用于在AD中分析的脑区域的参考数据集。低成本、基于液滴的单核RNA-seq将 然后用于分类和表征受tau和淀粉样蛋白病理学差异影响的区域中的细胞类型 来自许多跨越AD进展的供体，具有定量的tau和Aβ病理。更广泛的大脑 然后将对具有一致AD相关表型的病例子集进行调查，以了解是否 在整个大脑区域中存在共同的AD特征，以及是否可以在治疗之前检测到AD的特征。 神经病理学的出现最后，更高分辨率的方法将针对转录组和表观遗传 与特定细胞类型的病理学和疾病诊断相关的AD变化，旨在实现 对AD表型的机械理解。使用这种设计，本项目可以直接探测失调 第一次在受影响的细胞类型内的基因网络，提供了遗传或 表观遗传状态和由此产生的基因表达。由此产生的数据集和平台将产生有价值的 深入了解AD的细胞和分子基础，并将通过Data Core公开访问。

ABSTRACT (PROJECT 2) Alzheimer's disease (AD) is a highly penetrant neurodegenerative disease projected to affect 13.8 million cases in the US by 2050 at a cost of $1.1 trillion if no treatment is developed. AD is characterized by stereotyped progressive neurodegeneration and accumulation of two misfolded proteins in brain regions important for cognition and memory. Neurofibrillary tangles (NFTs) of hyperphosphorylated tau follow a progression like neurodegeneration, while extracellular amyloid beta (Aβ) plaques are initially detected in cortical and deep brain structures. It is unclear whether these pathologies are causal or effects of other underlying processes and currently no anti-tau or anti- Aβ therapies stop or reverse AD. Gene expression studies of AD have largely been performed on tissue or cell populations, and impact of neuronal loss and gliosis on these results is unknown. Epigenetic modifications are also associated with AD, though methylation studies have produced conflicting results and no clear pattern of epigenetic dysregulation associated directly with AD progression has emerged. The present study adapts recently developed high-throughput, single-cell methods for transcriptomic and epigenetic analysis to the identify molecular and gene regulatory hallmarks of “clinically typical” AD without significant co-morbidities. Building off a detailed understanding of neurotypical adult cell types, the project aims to identify transcriptional changes in specific cell types or classes correlated with increasing severity of AD pathology in different brain regions affected by the disease, and then identify gene and chromatin accessibility changes with pathology in vulnerable cell populations. This project will initially optimize single nucleus RNA-seq and epigenetics methods for use with postmortem samples of varying pathology and tissue quality, and generate reference datasets for brain regions to be analyzed in AD. Low-cost, droplet-based single nucleus RNA-seq will then be used to classify and characterize cell types in regions differentially affected by tau and amyloid pathology from many donors spanning AD progression with quantified tau and Aβ pathologies. A broader set of brain regions will then be surveyed on a subset of cases with consistent AD-related phenotypes to understand whether there is a common AD signature across brain regions, and whether signatures of AD can be detected prior to the emergence of neuropathology. Finally, higher-resolution methods will target transcriptomic and epigenetic changes in AD associated with pathology and disease diagnosis in specific cell types, aimed at achieving a mechanistic understanding of AD phenotypes. Using this design, this project can directly probe dysregulated gene networks within affected cell types for the first time, providing a potential causal link between genetic or epigenetic states and resulting gene expression. The resulting datasets and platform will produce valuable insights into the cellular and molecular basis of AD and will be made publicly accessible through the Data Core.