STEM-22. SINGLE-CELL LINEAGE TRACING IN PRIMARY GLIOBLASTOMA REVEALS DISTINCT PROGENITOR SUBTYPES DRIVING INTRATUMORAL HETEROGENEITY

Abstract

Abstract Glioblastoma (GBM) is the most common type of adult primary brain tumor with nearly a 100% rate of recurrence. One of the key clinical challenges is the vast degree of inter- and intra-tumoral variation in cell types and cell states, which change throughout the course of disease. Single-cell RNA-sequencing (scRNA-seq) studies in GBM samples have illuminated the complex landscape of cell type composition in primary and recurrent GBM, however these studies do not address how this diverse array of cell types developed or which cell populations are principally responsible for generating the heterogeneity and treatment resistance in the tumor. In this study, we implement a scRNA-seq compatible lineage tracing strategy, CellTagging, to track the behavior of individual tumor cells over time. Human GBM samples are dissociated immediately upon resection and infected with the CellTag DNA barcode library. The labeled cells are transplanted onto human cortical organoids, which facilitate the crucial normal-tumor cell interactions that enable GBM growth and invasion. After a period of proliferation, tumor cells are dissociated from the organoid and a fraction are harvested for scRNA-seq for transcriptomic and DNA barcode analysis. The remaining tumor cells are re-transplanted onto new organoids and harvested for scRNA-seq at later timepoints. Clonal relationships are determined by identifying cells with the same DNA barcode profile, and cell type characterizations are assigned via transcriptomic analysis. Transcriptomic profiling demonstrates the existence of multiple subtypes of tumor progenitors across timepoints. Strikingly, clonal analysis indicates that each progenitor subtype gives rise to numerous mature tumor cell types, suggesting multiple distinct lineages that contribute to tumor heterogeneity. Through differential expression analysis, we can predict which progenitors are most likely to drive which cell states. By introducing a treatment paradigm into our lineage tracing protocol, we identify tumor progenitor populations that contribute to treatment resistance.