The 3D genome in transcriptional regulation across the postnatal life span, with implications for schizophrenia and bipolar disorder

项目总结 复杂神经精神疾病的全基因组关联研究，包括精神分裂症和 双相情感障碍(BD)，已经确定了许多主要位于非编码区的风险基因座， 有必要对非编码调控元件进行系统研究。此外，还已确定SCZ风险 基因座优先位于神经元的启动子和增强子调控序列中，它们共同作用于神经元。 用表达数量性状基因座(EQTL)定位，从而暗示特定的基因。然而，要做到这一点 到目前为止，时空分辨率有限：(1)只有少数皮质区域 检查，(2)3D基因组在整个生命周期中对转录调控的影响从未被 检查，以及(3)研究局限于匀浆脑组织或仅包括广义的 神经性和非神经性群体。为了解决这些限制，我们将生成细胞类型--大脑区域- 和特定年龄段的高维数据，将告诉我们3D基因组对 转录调控，并将调控元素与特定的转录产物联系起来。在目标1中，我们将研究 SCZ和BD风险变异对3D基因组结构和转录调控的影响。我们将使用 荧光激活核团分选法分离谷氨酸和氨基丁酸能神经元 与SCZ和BD相关的5个人脑皮质和皮质下区域的少突胶质细胞和星形胶质细胞核 在出生后的五个年龄段。然后我们将为基因表达生成特定于细胞类型的注释，并 增强子RNA(RNA-SEQ和CAGE-SEQ)、开放染色质(ATAC-SEQ)、绝缘体(CTCF芯片-SEQ)、活性 增强子和启动子(H3K27ac和H3K4me3芯片序列)，以及染色质环相互作用(HIC和 Capture-C)。使用得到的数据，我们将描绘与3D相关的顺式转录调控 基因组(包括启动子-增强子环)并揭示SCZ和BD风险的功能后果 增强子-转录单位的定位。在目标2中，我们将研究SCZ和BD风险变异对细胞的影响 类型特异性基因表达和表观基因组QTL。我们将在单个细胞中映射RNAseq和ATACseq 水平，并将使用特定于细胞类型的标记和去卷积方法来处理现有的大规模 转录组和表观基因组数据集，来自CommonMind联盟、MinenCODE和其他项目，在 以产生细胞类型特异性表达和表观基因组QTL。然后，我们将共同本地化SCZ和BD风险 具有表达和精细图谱的表观基因组QTL来定义与疾病相关的增强子-转录单位。 最后，在目标3中，我们将通过风险基因座的表观基因组编辑来验证疾病相关的增强子-转录本单元 在IPC来源的细胞中。我们将应用CRISPR/Cas9来激活(P300)或抑制(KRAB)增强子 疾病相关的增强子-转录本单位(Aims1-2)。最后，我们将介绍表观微扰和 研究HiPSC来源细胞中的基因表达、染色质可及性和染色质环相互作用。它 我们的期望是，这些综合分析将使我们能够在SCZ内分配特定的监管单位 和BD风险单倍型到特定的细胞类型、脑区域和年龄窗口，从而提供对 SCZ和BD的遗传风险机制。

PROJECT SUMMARY Genome wide association studies of complex neuropsychiatric diseases, including schizophrenia (SCZ) and bipolar disorder (BD), have identified numerous risk loci that are mostly situated in non-coding regions, necessitating a systematic study of non-coding regulatory elements. It has also been established that SCZ risk loci are preferentially located within promoter and enhancer regulatory sequences of neurons and that they co- localize with expression Quantitative Traits Loci (eQTL), thus implicating specific genes. However, work that has been performed to-date has limited spatiotemporal resolution as: (1) only a few cortical regions have been examined, (2) the effect of 3D genome on transcriptional regulation across the lifespan has never been examined, and (3) studies have been limited to homogenate brain tissue or include only broadly defined neuronal and non-neuronal populations. To address these limitations, we will generate cell type-, brain region- and age period-specific high-dimensional data that will inform us of the effect of 3D genome on the transcriptional regulation and will link regulatory elements with specific transcripts. In Aim 1, we will examine the impact of SCZ and BD risk variants on 3D genome structure and transcriptional regulation. We will use fluorescence activated nuclei sorting to isolate glutamatergic and GABAergic neuronal as well as oligodendrocyte and astrocyte nuclei from five human cortical and subcortical regions relevant to SCZ and BD across five postnatal age periods. We will then generate cell-type specific annotations for gene expression and enhancer RNA (RNA-seq and CAGE-seq), open chromatin (ATAC-seq), insulators (CTCF ChIP-seq), active enhancers and promoters (H3K27ac and H3K4me3 ChIP-seq), and chromatin loop interactions (HiC and Capture-C). Using the resulting data, we will delineate cis transcriptional regulation associated with the 3D genome (including promoter-enhancer loopings) and uncover the functional consequences of SCZ and BD risk loci on enhancer-transcript units. In Aim 2, we will examine the impact of SCZ and BD risk variants on cell type-specific gene expression and epigenome QTLs. We will map RNAseq and ATACseq at the single cell level and will use cell type-specific markers and deconvolution approaches to the existing large scale transcriptome and epigenome datasets, from CommonMind consortium, psychENCODE and other projects, in order to generate cell type-specific expression and epigenome QTLs. We will then co-localize SCZ and BD risk loci with expression and fine map epigenome QTLs to define disease-associated enhancer-transcript units. Finally, in Aim 3, we will validate disease-associated enhancer-transcript units by epigenomic editing of risk loci in iPCS-derived cells. We will apply the CRISPR/Cas9 to activate (p300) or inhibit (KRAB) enhancers of the disease-associated enhancer-transcript units (Aims1-2). Lastly, we will introduce epigenomic perturbations and characterize gene expression, chromatin accessibility and chromatin loop interactions in hiPSC-derived cells. It is our expectation that these integrated analyses will enable us to assign specific regulatory units within SCZ and BD risk haplotypes to specific cell types, brain regions and age windows, thereby providing insight into the mechanisms of genetic risk for SCZ and BD.