Using high dimensional molecular data to decipher gene dynamics underlying pathogenic synovial fibroblasts

项目摘要 类风湿关节炎周围滑膜组织中成纤维细胞的病理性扩张 关节炎(RA)和骨关节炎(OA)。最近的研究已经确定了分子和功能上不同的 滑膜成纤维细胞表型的单细胞RNA测序(ScRNAseq)。其中一种表型， 仅在滑膜的衬里间隔中发现，并在RA和OA中扩展 与体内组织破坏有关。先前的研究表明滑膜成纤维细胞的表型是 塑料，使它们有可能通过生物疗法诱导，但很难在体外研究，因为它们失去了 他们的表型在体外。我提出了两种新的策略来模拟诱导和维持 滑膜衬里表型。初步分析优先考虑转化生长因子𝛽，一种已知的推动成纤维细胞的细胞因子 在这两种战略中，都存在差异化。第一种策略建立在成纤维细胞表型存在于 在人体组织中存在动态平衡和诱导的多个阶段。目标2将对这些状态进行建模 用新的RNA速度算法推断108例滑膜供体活检组织中的100,000多个成纤维细胞 成纤维细胞的分化过程和驱动基因的命名。AIM 2将执行108个单独的 通过Meta分析进行组合的分析，或与Cresendo进行一次联合分析，将在目标1中开发 作为首例多供体RNA测速分析。第二个战略建立在初步数据的基础上，这些数据表明 在表型诱导中被激活的基因在体外表型丧失过程中被失活。Aim 3将直接 体外实验测定15万成纤维细胞表型缺失的动态变化 使用scRNAseq进行积分。目的3a将测试外源性转化生长因子𝛽刺激维持衬里的效果。 体外表型。目标3b将从生成的复杂分析中提名和测试更多路径 时程数据。总之，这些目标将确定衬里表型的分子驱动因素，并推动新的 靶向衬里成纤维细胞的治疗方法研究。 我在单细胞计算生物学和滑膜成纤维细胞基因组学方面有专长。我开发了 用于单细胞集成的流行和声算法，发表在《自然方法》杂志上，并与人合著了 详细介绍了体内关节炎疾病所必需的一种新的成纤维细胞亚型的诱导，在新闻发布会上 大自然。完成建议的研究将帮助我建立时间进程数据分析的分析技能 并在实验成纤维细胞生物学方面发展无价的技能。我会训练Soumya Raychaudhuri博士 统计分析，共同导师迈克尔·布伦纳博士，滑膜成纤维细胞生物学专家，顾问彼得博士 卡尔琴科，RNA速度的开发者，顾问菲奥娜·鲍里博士，肯尼迪研究所所长 风湿病和顾问克里斯托弗·巴克利博士，滑膜成纤维细胞生物学专家。有了这个多- 学科培训，我将成为一名应用计算和实验方法的首席研究员 转化性风湿病研究。

Project Summary Pathological expansion of fibroblasts in the synovial tissue surrounding the joint drive disease in rheumatoid arthritis (RA) and osteoarthritis (OA). Recent studies have identified molecularly and functionally distinct phenotypes of synovial fibroblasts using single cell RNA sequencing (scRNAseq). One of the phenotypes, found exclusively in the lining compartment of the synovium and expanded in both RA and OA, has been implicated in tissue destruction in vivo. Previous studies have shown that synovial fibroblast phenotypes are plastic, making them potentially inducible with biological therapies but difficult to study in vitro, as they lose their phenotypes ex vivo. I propose two novel strategies to model the induction and maintenance of the synovial lining phenotype. Preliminary analyses prioritized TGF𝛽 , a cytokine known to drive fibroblast differentiation, in both strategies. The first strategy builds on the notion that fibroblast phenotypes are in dynamic equilibrium and exist at multiple stages of induction in human tissue. Aim 2 will model these states in over 100,000 fibroblasts from 108 synovial donor biopsies with the novel RNA velocity algorithm to infer lining fibroblast differentiations processes and nominate driver genes. Aim 2 will either perform 108 separate analyses combined through meta-analysis or do one joint analysis with Crescendo, to be developed in aim 1 as the first multi-donor RNA velocity analysis. The second strategy builds on preliminary data that show that genes activated in phenotype induction are inactivated during phenotype loss ex vivo. Aim 3 will directly experimentally assay the dynamics of phenotype loss ex vivo, profiling 150,000 fibroblast at multiple time points with scRNAseq. Aim 3a will test the efficacy of exogenous TGF𝛽 stimulation to maintain the lining phenotype ex vivo. Aim 3b will nominate and test more pathways from sophisticated analysis of the generated time-course data. Together, these aims will identify molecular drivers of the lining phenotype and fuel novel research on therapeutics to target lining fibroblasts. I have expertise in single cell computational biology and synovial fibroblast genomics. I developed the popular Harmony algorithm for single cell integration, published in Nature Methods and co-first authored a paper detailing the induction of a novel fibroblast subtype necessary for arthritic disease in vivo, in press at Nature. Completing the proposed research will help me build my analytical skills with time-course data analysis and develop invaluable skills in experimental fibroblast biology. I will train Dr. Soumya Raychaudhuri, in statistical analyses, co-mentor Dr. Michael Brenner, expert in synovial fibroblast biology, advisor Dr. Peter Kharchenko, developer of RNA velocity, advisor Dr. Fiona Powrie, director of the Kennedy Institute for Rheumatology, and advisor Dr. Christopher Buckley, expert in synovial fibroblast biology. With this multi- disciplinary training, I will be become a principal investigator applying computational and experimental methods to translational rheumatology research.