OTEH-12 - Artem Berezovsky.m4v
OTEH-12 Assessing Adaptive Responses to Loss of Extrachromosomal DNA Amplification
Contact Presenter
Artem Berezovsky1,2, Indrani Datta1, Ruicong She1, Laura Hasselbach1, Laila Poisson1, Ana C. deCarvalho1,2
1Henry Ford Health System, Detroit, Michigan, USA. 2Wayne State University, Detroit, Michigan, USA
Background: Oncogene activation through somatic gene amplification happens frequently in GBM, with over 70% of these tumors presenting amplification of at least one putative driver gene, oftentimes in small extrachromosomal circular DNA segments composed of chromatin (ecDNA). A molecularly diverse and representative panel of GBM patient-derived cancer stem-like cells (CSC) and orthotopic mouse xenografts (PDX), which retain the original genomic abnormalities and ecDNA amplifications, was employed to assess adaptive response to the absence of ecDNA amplification. Methods: We have isolated ecDNA negative cell populations from two patient-derived models. HF3035 harbors a MET amplification and HF3253 harbors a PDGFRA constitutively active genomic rearrangement and extrachromosomal amplification. We conducted paired, whole RNA-sequencing on 20 HF3253 populations (ecDNA+/-: 6 clones from 3 biological replicate PDXs and 4 clones from 4 in vitro technical replicates) and 12 HF3035 population (ecDNA+/-: 6 clones from 3 biological replicate PDXs). Results: Nonparametric differentially expressed gene (DEG) analysis using NOISeqBio (R/Bioconductor), identified 564 differentially expressed genes (482 upregulated in ecDNA(-)) employing a stringent false discovery rate of 0.05. Genes significantly associated with PDGF stimulation, central carbon metabolism, and H3K27me3 were downregulated in ecDNA(-), while genes significantly associated with astrocytic processes, neuronal differentiation, and EGFR signaling were upregulated in ecDNA(-) (EnrichR). We employed an additive linear model with PDX serving as a blocking factor to compare ecDNA+ and ecDNA- populations in both models (R/edgeR). 2071 genes were upregulated in ecDNA+ PDX specimens and 2365 genes were downregulated. Specifically, E2F targets were highly enriched in ecDNA+ populations, in addition to mRNA pre-processing. ecDNA loss primarily targeted glycogen metabolism, NTRK signaling, and inositol phosphate catabolism. Conclusions: We have identified PDX-specific and non-specific features to an adaptive response to the loss of ecDNA amplification. Notably, a signature adaptation is an upregulation of seemingly redundant receptor tyrosine kinases.
Contact Presenter
Artem Berezovsky1,2, Indrani Datta1, Ruicong She1, Laura Hasselbach1, Laila Poisson1, Ana C. deCarvalho1,2
1Henry Ford Health System, Detroit, Michigan, USA. 2Wayne State University, Detroit, Michigan, USA
Background: Oncogene activation through somatic gene amplification happens frequently in GBM, with over 70% of these tumors presenting amplification of at least one putative driver gene, oftentimes in small extrachromosomal circular DNA segments composed of chromatin (ecDNA). A molecularly diverse and representative panel of GBM patient-derived cancer stem-like cells (CSC) and orthotopic mouse xenografts (PDX), which retain the original genomic abnormalities and ecDNA amplifications, was employed to assess adaptive response to the absence of ecDNA amplification. Methods: We have isolated ecDNA negative cell populations from two patient-derived models. HF3035 harbors a MET amplification and HF3253 harbors a PDGFRA constitutively active genomic rearrangement and extrachromosomal amplification. We conducted paired, whole RNA-sequencing on 20 HF3253 populations (ecDNA+/-: 6 clones from 3 biological replicate PDXs and 4 clones from 4 in vitro technical replicates) and 12 HF3035 population (ecDNA+/-: 6 clones from 3 biological replicate PDXs). Results: Nonparametric differentially expressed gene (DEG) analysis using NOISeqBio (R/Bioconductor), identified 564 differentially expressed genes (482 upregulated in ecDNA(-)) employing a stringent false discovery rate of 0.05. Genes significantly associated with PDGF stimulation, central carbon metabolism, and H3K27me3 were downregulated in ecDNA(-), while genes significantly associated with astrocytic processes, neuronal differentiation, and EGFR signaling were upregulated in ecDNA(-) (EnrichR). We employed an additive linear model with PDX serving as a blocking factor to compare ecDNA+ and ecDNA- populations in both models (R/edgeR). 2071 genes were upregulated in ecDNA+ PDX specimens and 2365 genes were downregulated. Specifically, E2F targets were highly enriched in ecDNA+ populations, in addition to mRNA pre-processing. ecDNA loss primarily targeted glycogen metabolism, NTRK signaling, and inositol phosphate catabolism. Conclusions: We have identified PDX-specific and non-specific features to an adaptive response to the loss of ecDNA amplification. Notably, a signature adaptation is an upregulation of seemingly redundant receptor tyrosine kinases.