Abstract 6945: Dissecting the interplay of cell identity and extrachromosomal DNA in glioblastoma

Abstract

Abstract The five-year survival rate of glioblastoma multiforme (GBM) is 5.8%, and its standard of care has not changed since 2005. This stagnation in therapeutic development is due, in part, to its characteristic heterogeneity–multiforme–highlighting a critical need to understand the biological processes underlying this GBM hallmark. Despite the well-documented individual relationships between extrachromosomal DNA (ecDNA), GBM, and heterogeneity, the interplay between the three is not well understood. The Furnari lab has generated a library of cancer avatars, or human induced pluripotent stem cells (hiPSCs) engineered to harbor cancer driving mutations. Single cell RNA sequencing (scRNA-seq) was used to characterize the longitudinal evolution of the first avatars engineered to have TP53−/−;PDGFRAΔ8-9 and PTEN−/−;NF1−/− genotypes. The TP53−/−;PDGFRAΔ8-9 avatar acquired ecDNA, while the PTEN−/−; NF1−/− avatar did not. Re-analysis of the scRNA-seq data identified a subpopulation of cells in the TP53−/−;PDGFRAΔ8-9 avatar that had high expression of CDK4 and PDGFRA, two genes that are commonly co-amplified, found on ecDNA, and drive cellular states. This subpopulation of cells also had highly expressed genes near the CDK4 locus indicating a structural variant and possibly ecDNA. This population diverged transcriptionally from other cells and was associated with the OPC-like cellular state, which has been shown to be associated with high levels of PDGFRA and, at times, co-expression with CDK4. While TP53−/−;PDGFRAΔ8-9 tumors contained cells across all four GBM cellular states, PTEN−/−;NF1−/− tumors had a significantly less diverse composition of cellular states suggesting that ecDNA may facilitate cellular state diversity. With the eventual goal to directly test whether ecDNA enables transcriptional state heterogeneity compared to chromosomally inherited amplifications, CRISPR-Cas9 and CRISPR-C were implemented to engineer EGFRvIII ecDNA in U87. While inverse PCR confirmed the successful generation of ecDNA in U87, ecDNA was lost following long term in vitro cell passaging. Current follow up work aims to identify different intrinsic (oncogene amplified, genetic background, cell type, cell line) and extrinsic (in vitro conditions, orthotopic engraftment) contexts that promote the selection and amplification of engineered ecDNA, as it would be would be valuable to have genotypically identical cancer models with different classes of amplifications (i.e., engineered ecDNA vs. lentiviral overexpression, which is chromosomally inherited). Together, these results highlight the potential to use hiPSC-derived models and single cell methods to investigate gliomagenesis and cell identity. Citation Format: Brett Taylor, Brandon Jones, Daisuke Kawauchi, Yohei Miyake, Raghav Vadla, Nathan Jameson, Shunichiro Miki, Tomoyuki Koga, Bing Ren, Frank Furnari. Dissecting the interplay of cell identity and extrachromosomal DNA in glioblastoma [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2024; Part 1 (Regular Abstracts); 2024 Apr 5-10; San Diego, CA. Philadelphia (PA): AACR; Cancer Res 2024;84(6_Suppl):Abstract nr 6945.