Genomic dissection of tumor heterogeneity and progression

该项目的第一个目标是使用单细胞测序来了解儿童实体肿瘤中存在的细胞类型和细胞间异质性的复杂性。在这项工作中，我们专注于生成NF1患者肿瘤中存在的细胞的全面基因表达谱。迄今为止，我们收集并分析了丛状神经纤维瘤（PN）、非典型神经纤维瘤（AN）和恶性周围神经鞘肿瘤（MPNST）的手术标本。在过去的两年中，我们对来自24例NF1神经肿瘤患者的600,000个单细胞进行了高基因覆盖率测序，包括组织学验证的PN， ANF和MPNST。我们迄今为止生成的数据集包括PN和ANF的单细胞测序。从这些实验中，我们捕捉到这些肿瘤内细胞异质性的景观。在PNs中，我们已经确定了至少21种独特的细胞群，包括各种基质细胞和免疫细胞类型。这个丰富的数据集详细介绍了这些群体的转录谱，并突出了已知和新的细胞类型。按照相同的程序，我们从3个MPNST肿瘤和10个PDX MPNST模型的300,000个细胞中生成了scRNAseq数据。该数据集的初步示例今年与纪念斯隆凯特林和华盛顿大学的研究人员合作发表。这些工作发现了8号染色体扩增在MPNST中的复发和重要性，以及具有癌症干细胞样特性的有趣细胞群。此外，我们已经取得了进展，将人类的发现与基于NF1的神经肿瘤的基因工程模型系统相结合。这个描述这些肿瘤中大量细胞间相互作用的初始数据集被接受发表。我们也一直在积极地寻找新的方法来阻止这些肿瘤的恶性进展。目前的努力是开发一个体外系统来了解骨髓细胞群在肿瘤微环境中的极化。如果成功，这项工作可以提供特定的脆弱性，可以与FDA批准的MEK抑制剂联合使用，进一步改善丛状神经纤维瘤患者的肿瘤缩小。鉴于很难从实体瘤患者获得多个连续肿瘤活检标本，我们开展了一个项目，通过测定游离细胞DNA来评估NF1患者的疾病状态。在这项工作中，我们开发了一种将低通全基因组测序与选定基因的NF1特异性靶向捕获深度测序相结合的检测方法。这项工作在我们发表的“分类器”中达到了顶峰，该分类器使用血液标本中的肿瘤部分来提供一种临床工具，可用于区分MPNST和良性丛状神经纤维瘤。目前，我们正在使用积极招募的临床试验SARC031 （NCT03433183）“MEK抑制剂Selumetinib （AZD6244）与mTOR抑制剂西罗莫司联合治疗MPNST患者”收集的样本评估我们的检测值。游离细胞DNA技术的另一个应用是将循环肿瘤DNA中观察到的变化与再分期成像中观察到的变化联系起来，作为肿瘤对治疗反应的衡量标准。这些努力目前正在进行中。正在探索将循环肿瘤DNA与正在治疗的肿瘤活检配对，观察循环肿瘤DNA的相关性和描述全球基因组变化，以发现肿瘤进化的机制。最后，我们继续努力提高细胞游离DNA检测的灵敏度和特异性。目前的努力包括片段大小的更深层次的分析，通过基于捕获或全基因组的方法进行更深层次的测序，以及甲基化分析等替代测序策略。这项工作的第二个目标是开发和采用新的条形码策略与单细胞测序相结合，可用于临床前模型系统来模拟肿瘤细胞的耐药性和生存。单细胞测序可以同时剖析数千个细胞的基因表达谱，但在选择性压力下追踪细胞群的能力有限。在今年，我们已经完成了一些实验，这些实验突出了在我们的实验系统中添加一个功能的必要性，该功能可以随着时间的推移跟踪特定的细胞并进行复制。目前的实验正在采用单细胞条形码策略以及将CRISPr基因敲除纳入实验工作流程。我们现在已经完成了一项针对MPNST细胞中多种表观遗传复合物（包括SWI/SNF和PRC1）敲除的筛选。这些实验揭示了MPNST细胞的新漏洞。正在进行的工作旨在验证这些观察结果并描述这些观察结果背后的机制。

Project Summary The first aim of this project is to use single cell sequencing to understand the complexities of cell types and cell to cell heterogeneity that is present within pediatric solid tumors. In this work, we are focusing on generating comprehensive gene expression profiling of the cells present within tumors that occur in patients with NF1. To date we have collected and analyzed surgical specimens from Plexiform Neurofibromas (PN), Atypical Neurofibromas (AN) and Malignant Peripheral Nerve Sheath Tumors (MPNST). Over the past two years, we have generated high gene coverage sequencing on 600,000 single cells from 24 patients with NF1 nerve tumors including histologically validated PN, ANF and MPNST. Our dataset generated to date includes single-cell sequencing of PN and ANF. From these experiments, we capture the landscape of cellular heterogeneity within these tumors. Within PNs we have identified at least 21 unique cell populations including a variety of stromal and immune cell types. This rich dataset details the transcriptional profile of each of these populations and highlights both known and novel cell types. Following the same procedure, we have generated scRNAseq data from 300,000 cells from three MPNST tumors and 10 PDX MPNST models. Preliminary examples of this dataset were published this year in collaborative efforts with investigators at Memorial Sloan Kettering and Washington University. These works discovered the recurrence and importance of chromosome 8 amplification in MPNST and an interesting cellular population with cancer stem cell like properties. Additionally, we have made progress integrating the human findings with genetically engineered model systems of NF1 based nerve tumors. This initial dataset describing the multitude of cell-to-cell interactions within these tumors is accepted for publication. We also have actively been working to discover novel ways to arrest the malignant progress in these tumors. A current effort is to develop an in vitro system to understand the polarization of myeloid cell populations in the tumor microenvironment. If successful, this work could deliver specific vulnerabilities that could be used in combination with the FDA approved MEK inhibitors to further improve tumor shrinkage in patients with plexiform neurofibroma. Given the difficulty in obtaining multiple sequential tumor biopsy specimens from patients with solid tumors, we have undertaken a project to assay cell free DNA to assess disease status in NF1 patients. In this work, we have developed an assay that marries low pass whole genome sequencing with NF1 specific targeted capture deep sequencing of selected genes. This work has culminated in publication of our "classifier" which uses tumor fraction from blood specimens to provide a clinical tool that could be used to differentiate between MPNST and the benign Plexiform Neurofibroma. Currently we are evaluating our assays value using samples collected on the actively enrolling clinical trial SARC031 (NCT03433183) "MEK Inhibitor Selumetinib (AZD6244) in Combination with the mTOR Inhibitor Sirolimus for Patients With MPNST". Another application of cell free DNA technology is correlating the changes observed in the circulating tumor DNA with the changes observed on re-staging imaging as a measure of response of a tumor to therapy. These efforts are currently underway. Secondary efforts to pair the circulating tumor DNA with an on-treatment tumor biopsy to observe correlation and description of the global genomic changes in the circulating tumor DNA to discover mechanisms of tumor evolution are being explored. Finally, we have continued efforts to improve the sensitivity and specificity of the cell free DNA assay. Current efforts include deeper profiling of fragment size, deeper sequencing through capture based or whole genome methods, and alternative sequencing strategies such as methylation profiling. The second aim of this work is to develop and employ novel barcoding strategies married with single cell sequencing that can be used in preclinical model systems to model tumor cell resistance and survival. Single cell sequencing can dissect the gene expression profile of thousands of cells in parallel but is limited in its ability to track populations of cells under a selective pressure. Within the current year we have completed experiments that highlighted the need to add a feature to our experimental system which enabled tracking a particular cell over time and as it replicated. Current experiments are employing the single cell barcoding strategy as well as incorporation of CRISPr gene knockouts to the experimental workflow. We now have completed a screen targeting knockout of multiple epigenetic complexes including SWI/SNF and PRC1 in MPNST cells. These experiments have delivered novel vulnerabilities of the MPNST cell. Ongoing work is designed to validate these observations and describe the mechanisms underlying these observations.