High-throughput cellular genetics to connect noncoding variants to coronary artery disease genes

摘要 尽管有有效的治疗方法来控制血压和胆固醇，冠状动脉疾病（CAD）仍然是 是美国和全世界的主要死因最近的全基因组关联研究（GWAS） 已经确定了超过200个与CAD显著相关的遗传位点。这些基因座中的大多数与 与高血压或高脂血症，但含有基因表达的血管细胞，表明存在的 未发现的CAD疾病机制通过血管壁细胞起作用。识别生物 这些基因座的机制是困难的，因为大多数与CAD相关的常见变体是非编码的， 很可能通过增强子发挥作用，增强子可以远距离调节多个基因，并且是高度细胞型的。 特定.这减缓了发现的速度，因此只有少数这样的基因座被“解决”，阻止了实现 遗传学研究在开发急需的CAD新疗法方面的全部潜力。 我们需要的是更高通量、无偏见的方法来系统地剖析这些基因座， 疾病的分子和细胞机制。为了实现这一目标，我们开发并验证了两个 新的高通量方法，能够量化的分子和形态学的影响， CAD基因座中的数千个基因：优化的Perturb-seq（将基因与转录表型联系起来）和合并 扰动细胞绘画（将基因与形态效应联系起来）。 我们建议使用这些技术，以及新的和新颖的计算方法，以：1）分析 干扰CAD基因座基因在血管内皮细胞（EC）中的转录组学效应， 与动脉粥样硬化相关的刺激，2）量化干扰所有CAD基因座基因的形态学效应 在EC中，3）将这些数据与表观基因组和遗传数据整合以提名致病基因并建立假设 将变异与基因、基因与疾病相关的细胞表型联系起来，然后测试前5名， 这些假设中的每一个都使用变体编辑和EC功能测定。这包括几个优先的因果关系 调节KLF 2表达上游EC剪切应力反应的变体假说。 这些研究的完成将加速对CAD遗传学与疾病之间联系的理解， 并提供了对CAD中EC功能的深入了解，这可能会为开发新的治疗类别提供信息。 更广泛地说，这项研究将建立新的可扩展的实验和计算方法， 基因和细胞表型的疾病变异，这可能会大大加快变异功能 心血管和其他常见疾病的研究。

ABSTRACT Despite effective therapies to control blood pressure and cholesterol, coronary artery disease (CAD) remains the leading cause of death in the United States and the world. Recent genome-wide association studies (GWAS) have identified >200 genetic loci significantly associated with CAD. The majority of these loci are not associated with hypertension or hyperlipidemia, but contain genes expressed in vascular cells, suggesting the presence of undiscovered CAD disease mechanisms operating through cells of the blood vessel wall. Identifying the biologic mechanisms of these loci is difficult because most of the common variants associated with CAD are non-coding, likely functioning through enhancers that can regulate multiple genes at great distances, and be highly cell-type specific. This has slowed discovery, so that only a few such loci have been ‘solved’, preventing the realization of the full potential of genetic studies for the development of much needed new classes of therapies for CAD. What is needed are much higher-throughput, unbiased methods to systematically dissect these loci, to identify molecular and cellular mechanisms of disease. To accomplish this goal, we have developed and validated two novel high-throughput approaches with the ability to quantify the molecular and morphological effects of thousands of genes in CAD loci: optimized Perturb-seq (to link genes to transcriptional phenotypes) and pooled perturbation Cell Painting (to link genes to morphological effects). We propose to use these technologies, together with new and novel computational approaches, to: 1) analyze the transcriptomic effects of perturbing CAD-locus genes in vascular endothelial cells (ECs) subjected to four stimuli associated with atherosclerosis, 2) quantify the morphological effects of perturbing all CAD-locus genes in ECs, 3) integrate these data with epigenomic and genetic data to nominate causal genes and build hypotheses connecting variants to genes, and genes to disease-associated cellular phenotypes, and then test the top 5 of each of these hypotheses using variant-editing and EC-functional assays. This includes several prioritized causal variant hypotheses that regulate EC shear stress response upstream of KLF2 expression. The completion of these studies will accelerate understanding of the links between CAD genetics and disease, and provide insights into EC functions in CAD that may inform the development of new classes of therapies. More broadly, this study will establish new scalable experimental and computational methods to link noncoding disease variants to genes and cellular phenotypes, which could dramatically accelerate variant-to-function studies for cardiovascular and other common diseases.