Abstract PO-016: Revealing the cellular ecosystem and signaling pathways of posterior fossa childhood ependymoma with single-cell transcriptomic profiling

Abstract

Abstract Childhood ependymoma is a cancer of the central nervous system marked by its relapsing pattern. Currently, there are no effective chemotherapies for this devastating disease and standard of care remains neurosurgery followed by radiation. The most prevalent and aggressive pediatric subtype is posterior fossa ependymoma type A (PF EPN A) with a 10-year progression free survival rate less than 20%. Extensive bulk RNA-sequencing, DNA methylation, and whole-genome sequencing analyses have revealed a distinct transcriptomic profile for PF EPN A tumors driven by abnormal DNA methylation and histone modifications rather than single-nucleotide or copy number alterations. However, deciphering the specific tumorigenic mechanisms regulating the progression of these tumors and their relation with the tumor microenvironment remains an open question that is essential for expanding treatment options. To remedy this lack of knowledge, we have generated single-cell transcriptomic and immunohistochemistry (IHC) data from 6 primary and 3 metastatic PF EPN A specimens. Upon clustering and differential expression analysis, we show that PF EPN A tumors consist of transformed radial glia-like cells at varying stages of differentiation, recapitulating the differentiation of radial glia into ependymal cells, neurons, and astrocytes during neurodevelopment. Additionally, we show evidence that the cell-of-origin of these tumors is a LGR5+ radial glia driven by WNT/β-catenin signaling. Upon tumor progression, our data suggests that tumor stem cells associated with hypoxic conditions acquire an invasive and motile phenotype derived from a TGF-β-driven reactive astrogliosis program. Tumor-derived reactive astrocytes recapitulate much of the gene expression programs involved in the mesenchymal phenotype of grade IV astrocytoma, such as the expression of CD44, MET, VEGFA, and CHI3L1. To strengthen the statistical power of our study, we used our single-cell data to infer the abundance of cell types in a cohort of PF EPN A tumors profiled with bulk RNA-sequencing. We demonstrate that PF EPN A tumors can be stratified into two subtypes: stem-cell- and ependymal-cell-like tumors. Stem-cell-like tumors have a higher abundance of tumor-infiltrating microglia with a pro-inflammatory phenotype, younger age of diagnosis, and mesenchymal gene expression signature. These results were confirmed in an independent cohort of PF EPN A tumors profiled for gene expression with DNA microarrays. Overall, we have begun to uncover the cellular microenvironment and oncogenic pathways associated with PF EPN A tumorigenesis and progression. Our results lay the foundation for the development of therapies that target dysregulated pathways in childhood PF EPN A tumors. Citation Format: Rachael G. Aubin, Emma C. Troisi, Adam Alghalith, MacLean Nasrallah, Mariarita Santi, Pablo G. Camara. Revealing the cellular ecosystem and signaling pathways of posterior fossa childhood ependymoma with single-cell transcriptomic profiling [abstract]. In: Proceedings of the AACR Virtual Special Conference on Tumor Heterogeneity: From Single Cells to Clinical Impact; 2020 Sep 17-18. Philadelphia (PA): AACR; Cancer Res 2020;80(21 Suppl):Abstract nr PO-016.