HGG-57. WHOLE-GENOME SEQUENCING, METHYLATION ANALYSIS, AND SINGLE-CELL RNA-SEQ DEFINE UNIQUE CHARACTERISTICS OF PEDIATRIC TREATMENT-INDUCED HIGH-GRADE GLIOMA AND SUGGEST ONCOGENIC MECHANISMS
John Lucas,John DeSisto,Ke Xu,Andrew Donson,Tong Lin,Bridget Sanford,Gang Wu,Quynh Tran,Dale Hedges,Chih-Yang Hsu,Gregory Armstrong,Michael Arnold,Smita Bhatia,Patrick Flannery,Rakeb Lemma,Lakotah Hardie,Ulrich Schuller,Lindsey Hoffman,Kathleen Dorris,Jean Levy,Todd Hankinson,Michael Handler,Arthur Liu,Nicholas Foreman,Rajeev Vibhakar,Kenneth Jones,Sariah Allen,Jinghui Zhang,Suzanne Baker,Thomas Merchant,Brent Orr,Adam Green
DOI: https://doi.org/10.1093/neuonc/noaa222.336
2020-01-01
Neuro-Oncology
Abstract:Abstract BACKGROUND Pediatric treatment-induced high-grade glioma (TIHGG) is among the most severe late effects observed in childhood cancer survivors and is uniformly fatal. We previously showed that TIHGG are divergent from de novo pediatric high-grade glioma (pHGG) and cluster into two gene expression subgroups, one stemlike and the other inflammatory. Here we systematically compared TIHGG molecular profiles to pHGG and evaluated expression and single cell sequencing profiles in order to identify oncogenic mechanisms and the cellular basis for the observed TIHGG gene expression subgroups. MATERIALS/ METHODS 450/850K methylation and mutational signature analysis was conducted in 36 TIHGG samples. Resultant data were analyzed for the presence of chromothripsis, distinct molecular alterations, and mutational signatures in a subset of 10 samples with whole genome sequencing data. Five TIHGGs underwent single-cell RNA-Seq analysis (scRNAseq). RESULTS 26/36 TIHGG clustered with the pedRTK1 methylation class. TIHGG were characterized by an increased frequency of chromothripsis relative to pHGG (67% vs. 31%, p=0.036). FISH and WGS revealed frequent PDGFRA amplification secondary to enrichment in ecDNA. TIHGG were enriched for COSMIC mutational signatures 5 and 19 (p=0.0003) relative to pHGG. scRNAseq data showed that TIHGG tumors are composed of stem-like, neuronal, and inflammatory cell populations which may contribute to the previously described dominant expression profiles. CONCLUSIONS TIHGG represents a distinct molecular subtype of pHGG. Chromothripsis, leading to enriched expression of genes in extrachromosomal DNA, likely contribute to TIHGG oncogenesis. The dominant cell type (stem-like vs. inflammatory) may define the expression subgroup derived from bulk RNA-seq in heterogeneous tumors.