Next Generation Sequencing Facilitates Quantitative Analysis of HIV-1 Latency in Central Memory T Cells

Purpose: Next-generation sequencing (NGS) has become a powerful microscope to study cell models of HIV. The goals of this study is to analyze latent HIV infection in the TCM model of HIV latency described in Martins et al 2015 and to evaluate potential markers for HIV infection. Methods: Peripheral blood mononuclear cells were collected from 4 healthy donors. Naive CD4 T cells were isolated and utilized to generate a model of latent HIV infection. During model generation, cells were sorted and collected for RNA extraction 10 days post infection along with their uninfected counterparts. 2 days after this, cells were activated with CD3/CD28 stimulation and collected for study for a total of 16 samples from 4 donors. After cell collection, RNA was extracted from infected cells and their uninfected counterparts for deep sequencing by Expression Analysis. Sequence reads that passed quality filters were mapped using Tophat and counted using HTSeq. In addition to human transcripts, we utilized 92 ERCC spikes as negative controls. Any human gene which did not achieve at least 1 count per million reads in at least 4 samples or any ERCC that did not achieve at least 5 reads in at least 4 samples was discarded. Differential expression for activated samples was not performed, but rather used to demonstrate upregulation of T-cell activation markers along with changing to the type and abundance of HIV transcripts produced. Results: Using an custom built data analysis pipeline, 82 million reads per sample to the human genome (build hg38) and identified 13,534 human transcripts 67 ERCC spike in transcripts, and HIV NL43 transcripts were identified with the Tophat/HTSeq workflow. Differential expression analysis was performed between uninfected and latently infected cells. 826 genes were found to be differentially expressed, 275 downregulated and 551 upregulated (FDR < 0.05) with EdgeR. GO and Pathway analysis of differential expressed genes revealed significant dysregulation of genes involved in the p53 signaling pathway. Subsequent studies with pifithrin demonstrated a reduction in latently infected cell during model generation. Conclusions: This study represents the first detailed analysis of HIV latency with this cell model using next generations sequencing. These results demonstrate that the TCM model of HIV latency of Martins et al 2015 is truly reflective of HIV latency. This study also provides a framework for which to analyze future cell models of HIV latency using next generation sequencing. Finally, this work demonstrates that the p53 signaling pathway is a key pathway dysregulated in latency for this cell model with several genes dysregulated in HIV latency. Overall design: TCM model of HIV latency mRNA profiles of 4 donors in 4 conditions were generated by deep sequencing by Expression Analysis.

研究目的：下一代测序（next-generation sequencing, NGS）已成为研究HIV细胞模型的强力工具，如同高性能光学显微镜。本研究旨在分析Martins等于2015年报道的TCM型HIV潜伏感染模型中的潜伏HIV感染特征，并评估HIV感染的潜在生物标志物。

方法：从4名健康供者体内采集外周血单个核细胞（peripheral blood mononuclear cells, PBMC），分离初始CD4 T细胞并用于构建HIV潜伏感染细胞模型。在模型构建阶段，于感染后10天分选收集感染细胞及其未感染对照细胞，进行RNA提取。后续2天，通过CD3/CD28磁珠刺激活化细胞，最终共收集来自4名供者的16份样本开展后续研究。样本收集完成后，由Expression Analysis公司对感染细胞及对应未感染对照细胞的RNA进行深度测序。通过质量过滤的序列读段（reads）采用Tophat软件进行基因组比对，并通过HTSeq工具进行转录本计数。除人类转录本外，本研究引入92个外部RNA对照联盟（External RNA Controls Consortium, ERCC）外参作为阴性对照。随后对数据进行初步筛选：剔除至少4份样本中每百万读段计数不足1的人类基因，以及至少4份样本中读段数不足5的ERCC外参。本研究未对活化样本开展差异表达分析，而是利用其验证T细胞活化标志物的上调现象，以及所产生HIV转录本的类型与丰度变化。

结果：通过自主搭建的数据分析流程，将每份样本的8200万条序列读段比对至人类基因组（hg38版本），共鉴定得到13534个人类转录本、67个ERCC外参转录本以及HIV NL43病毒转录本，上述鉴定流程均通过Tophat/HTSeq组合完成。对未感染细胞与潜伏感染细胞开展差异表达分析，借助EdgeR软件鉴定得到826个差异表达基因，其中275个基因表达下调，551个基因表达上调（错误发现率<0.05，false discovery rate, FDR）。对差异表达基因进行基因本体（Gene Ontology, GO）富集分析与通路富集分析后发现，p53信号通路相关基因存在显著表达失调。后续采用pifithrin进行的功能验证实验表明，模型构建过程中潜伏感染细胞的比例出现显著降低。

结论：本研究首次利用下一代测序技术对该细胞模型的HIV潜伏感染特征开展详细分析。研究结果证实，Martins等于2015年报道的TCM型HIV潜伏感染模型可真实模拟体内HIV潜伏感染状态。本研究同时为后续利用下一代测序技术分析HIV潜伏感染细胞模型提供了标准化研究框架。最后，本研究证实p53信号通路是该细胞模型潜伏感染过程中失调的核心通路，多个基因在HIV潜伏感染状态下存在表达异常。

整体实验设计：通过Expression Analysis公司的深度测序，获取了4名供者在4种培养条件下的TCM型HIV潜伏感染模型的mRNA表达谱。