EPEN-14. EXPLORATION OF INTRA- AND INTER-TUMOR CELLULAR HETEROGENEITY IN SUPRATENTORIAL EPENDYMOMA. (23rd April 2019)

• Record Type: Journal Article
• Authors: Donson, Andrew; Gillen, Austin; Riemondy, Kent; Hesselberth, Jay; Amani, Vladimir; Griesinger, Andrea; Sanford, Bridgit; Jones, Ken; Hankinson, Todd; Handler, Michael; Venkataraman, Sujatha; Vibhakar, Rajeev; Foreman, Nicholas
• Is Part Of: Neuro-oncology. Volume 21(2019)Supplement 2
• Page Start: ii80
• Page End: ii80
• Publication Date: 2019-04-23
• Subjects: Brain Neoplasms -- Periodicals; Brain -- Tumors -- Periodicals; Brain -- Cancer -- Periodicals; Nervous system -- Cancer -- Periodicals; 616.99481
• DOI: 10.1093/neuonc/noz036.071
• Languages: English
• ISSNs: 1522-8517

Abstract

Abstract: Cellular heterogeneity is a critical determinant of ependymoma biology. Analysis of bulk tumor samples therefore presents a significant barrier to fully understanding ependymoma biology. Single cell RNAseq analysis (scRNAseq) addresses this problem, and has recently provided us with unparalleled insights into the biology of posterior fossa ependymoma (EPNPF). In the present study we have applied this approach to explore the intra- and inter-tumor cellular heterogeneity of supratentorial ependymoma (EPNST). Using the 10x Chromium Drop-seq platform we analyzed approximately 1500 cells from 3 EPNST C11orf95-RELA fusion patient samples. These cells were compared to approximately 2000 cells from 10 EPNPF patient samples. Unsupervised tSNE clustering showed that the EPNST cells clustered separately from EPNPF. Examination of EPNST marker gene expression identified RELA and L1CAM as significantly overexpressed in all EPNST versus EPNPF. Analysis of single cell regulatory pathways demonstrated upregulation of the RELA regulatory network in EPNST. Despite clustering differences, key EPNPF genes were also expressed by EPNST, most notably IGF2 and midkine (MDK). In EPNPF, IGF2 and MDK are expressed by a subpopulation of tumor cells with an undifferentiated phenotype that are associated with a worse clinical outcome. IGF2 and MDK may therefore underlie the malignant phenotype of both EPNPF and EPNST. Whereas all EPNPF patient samples clustered predominantly together, EPNST patient samples did not, suggesting greater inter-tumor heterogeneity in EPNST than EPNPF. These findings increase our biological understanding of EPNST with RELA-fusions, and reveal molecular features that potentially underlie oncogenesis in both EPNST and EPNPF.