Human cells have twenty-three pairs of chromosomes but in cancer, genes can be amplified in chromosomes or in circular extrachromosomal DNA (ECDNA), whose frequency and functional significance are not understood1–4. We performed whole genome sequencing, structural modeling and cytogenetic analyses of 17 different cancer types, including 2572 metaphases, and developed ECdetect to conduct unbiased integrated ECDNA detection and analysis. ECDNA was found in nearly half of human cancers varying by tumor type, but almost never in normal cells. Driver oncogenes were amplified most commonly on ECDNA, elevating transcript level. Mathematical modeling predicted that ECDNA amplification elevates oncogene copy number and increases intratumoral heterogeneity more effectively than chromosomal amplification, which we validated by quantitative analyses of cancer samples. These results suggest that ECDNA contributes to accelerated evolution in cancer.

Summary Paragraph Medulloblastoma is a highly malignant paediatric brain tumour currently treated with a combination of surgery, radiation, and chemotherapy, posing a considerable burden of toxicity to the developing child. Genomics has illuminated the extensive intertumoural heterogeneity of medulloblastoma, identifying four distinct molecular subgroups. Group 3 and Group 4 subgroup medulloblastomas account for the majority of paediatric cases; yet, oncogenic drivers for these subtypes remain largely unidentified. Here we describe a series of prevalent, highly disparate genomic structural variants, restricted to Groups 3 and 4, resulting in specific and mutually exclusive activation of the growth factor independent 1 family protooncogenes, GFI1 and GFI1B. Somatic structural variants juxtapose GFI1/GFI1B coding sequences proximal to active enhancer elements, including super-enhancers, instigating oncogenic activity. Our results, supported by evidence from mouse models, identify GFI1 and GFI1B as prominent medulloblastoma oncogenes and implicate ‘enhancer hijacking’ as an efficient mechanism driving oncogene activation in a childhood cancer.

Drugs that modify the epigenome are powerful tools for treating cancer, but these drugs often have pleiotropic effects, and identifying patients who will benefit from them remains a major clinical challenge. Here we show that medulloblastomas driven by the transcription factor Gfi1 are exquisitely dependent on the enzyme lysine demethylase 1 (Kdm1a/Lsd1). We demonstrate that Lsd1 physically associates with Gfi1, and that these proteins cooperate to inhibit genes involved in neuronal commitment and differentiation. We also show that Lsd1 is essential for Gfi1-mediated transformation: Gfi1 proteins that cannot recruit Lsd1 are unable to drive tumorigenesis, and genetic ablation of Lsd1 markedly impairs tumor growth in vivo. Finally, pharmacological inhibitors of Lsd1 potently inhibit growth of Gfi1-driven tumors. These studies provide important insight into the mechanisms by which Gfi1 contributes to tumorigenesis, and identify Lsd1 inhibitors as promising therapeutic agents for Gfi1-driven medulloblastoma.