Single-cell RNA Profiling Identifies Diverse Cellular Responses to EWSR1-FLI1 Down-regulation in Ewing Sarcoma

Abstract

Single-cell analyses provide insight into time dependent behaviors in response to dynamic changes of oncogene expression. We developed an unbiased approach to study gene expression variation using a model of cellular dormancy induced via EWSR1-FLI1 down-regulation in Ewing sarcoma (EWS) cells. We propose that variation in the expression of EWSR1-FLI1 over time determines cellular responses. Cell state and functions were assigned using random forest feature selection in combination with machine learning. Notably, three distinct expression profiles were uncovered contributing to Ewing sarcoma cell heterogeneity. Our predictive model identified ~1% cells in a dormant-like state and ~2-4% with higher stem-like and neural stem-like features in an exponentially proliferating EWS cell line and EWS xenografts. Following oncogene knockdown, cells re-entering the proliferative cycle have greater stem-like properties, whereas for those remaining quiescent, FAM134B-dependent dormancy provides a survival mechanism. We also show cell cycle heterogeneity related to EWSR1-FLI1 expression as an independent feature driving cancer heterogeneity, and drug resistance.SIGNIFICANCE We show that time-dependent changes induced by suppression of oncogenic EWSR1-FLI1 induces dormancy, with different subpopulation dynamics, including stem-like characteristics and prolonged dormancy. Cells with these characteristics are identified in exponentially growing cell populations and confer drug resistance, and could potentially contribute to metastasis or late recurrence in patients.