Extrachromosomal DNA drives cancer evolution

Chromosomal instability promotes oncogene amplification through extrachromosomal DNA (ecDNA), but the underlying mechanisms have been elusive. Recent data indicate that ecDNA originating from the rupture of micronuclei is retained throughout mitotic rounds and actively promotes oncogenic transcriptional programs, establishing ecDNA as a bona fide driver of cancer evolution.

Extrachromosomal DNA (ecDNA) consists of circular acentromeric DNA elements with an accepted role in high-level oncogene amplification during oncogenesis. ecDNA is detected in ~15%–20% of tumors, especially advanced or recurrent lesions, correlating with increased intratumoral heterogeneity and poor clinical outcome.¹ ecDNA appears to be preferentially associated with chromosomal instability (CIN) and has been proposed to support oncogenesis by sustaining cancer cell plasticity and adaptation to therapeutic challenges.¹ However, ecDNA lacks centromeres and segregates in a non-Mendelian manner during mitosis, at least hypothetically pointing to limited stability throughout tumor progression.¹

Many aspects of ecDNA biology have remained obscure. First, although CIN-associated processes such as micronucleation and chromothripsis have been associated with ecDNA, its origins have generally been inferred from sequencing data,² and the molecular mechanisms promoting ecDNA formation have been elusive. Second, while ecDNA can persist during mitosis upon association with chromosomes³ and can be stabilized by clustering,⁴ the molecular actors promoting ecDNA retention are unclear. Third, while ecDNA-driven oncogenesis involves processes other than gene amplification, including broad transcriptional rearrangements and chromatin reorganization,⁵ the underlying mechanisms remain incompletely understood. Three recent studies now shed light on ecDNA initiation, inheritance, and oncogenic functions using complementary mechanistic approaches.^6,7,8

Krupina and colleagues dissected the molecular mechanisms linking micronuclear rupture and ecDNA formation in cancer cells with elevated CIN.⁶ An unbiased, imaging-based siRNA screen targeting 204 known/predicted human nucleases in DLD1-Y^MN cells (a colorectal cancer cell line enabling inducible Y chromosome missegregation into micronuclei) followed by low-throughput validation by DNA-FISH identified NEDD4 binding protein 2 (N4BP2) as a cytoplasmic nuclease driving DNA fragmentation in micronuclei and consequent ecDNA formation.⁶

Mechanistically, N4BP2 was shown to translocate to ruptured micronuclei and localize to DNA damage sites. Genetic inhibition of N4BP2 suppressed micronuclear DNA damage without affecting rupture. Moreover, enforced targeting of N4BP2 to intact nuclei was sufficient to induce extensive chromosome fragmentation, demonstrating its intrinsic chromatin-destructive capacity.⁶ N4BP2 expression was positively correlated with chromothripsis, complex chromosomal rearrangements, and elevated ecDNA levels across human tumors, including glioblastoma. Finally, in a mouse model of glioma, N4bp2 deletion reduced micronuclear DNA damage and ecDNA formation, resulting in decelerated disease progression.⁶

Sankar and collaborators explored ecDNA inheritance in chromosomally instable malignancies.⁷ After confirming ecDNA co-segregation with mitotic chromosomes across multiple tumor models, these authors developed a genome-wide screen to identify human DNA fragments retained within episomal plasmids throughout mitosis. This approach revealed thousands of bona fide retention elements distributed across the genome,⁷ which (1) are preferentially mapped to active chromatin sites (especially promoters and enhancers), (2) associate with RNA polymerase II, active histone marks, and increased CpG content, and (3) are depleted from heterochromatin and replication origins. Episomal persistence was correlated with the number of retention elements, supporting a sequence-encoded basis for ecDNA inheritance.⁷

Live-cell imaging showed that retention elements tether episomal DNA to mitotic chromosomes, reducing its loss per cell division by > 2-fold. Retention elements were found to preferentially contact accessible chromatin enriched for mitotic bookmarking factors (i.e., transcription factors and other chromatin regulators that remain associated with DNA during mitosis), suggesting that ecDNA retention exploits enhancer–promoter-like interactions in trans.⁷ High mitotic retention was required to maintain ecDNA amplification across divisions, even under strong positive selection. Whole-genome analyses of human tumors revealed that oncogene-containing ecDNA often harbors multiple retention elements and is large, suggesting that co-amplification shapes ecDNA size and structure. Finally, retention elements were focally hypomethylated, and enforced CpG methylation disrupted their association with chromosomes and episomal retention, identifying DNA methylation as a crucial regulator of ecDNA persistence.⁷

Taghbalout and co-authors explored how ecDNA reshapes transcription in cancer by mapping ecDNA interactions with the nuclear genome.⁸ Thousands of reproducible transcription-linked ecDNA–chromosome contacts were identified across several tumor models, globally revealing that each ecDNA molecule simultaneously engages multiple chromosomal loci. Live-cell imaging confirmed the proximity between ecDNA and predicted targets. Most ecDNA-associated genes formed multivalent interaction clusters, with a core of 2299 genes conserved across cancer types.⁸ Notably, ecDNA interaction clusters were enriched at enhancer-like loci bound by mediator complex subunit 1 (MED1), a transcriptional coactivator, and exhibiting active enhancer histone modifications. Live-cell imaging revealed frequent colocalization of ecDNA with MED1-positive nuclear puncta, pointing to a condensate-like organization. Moreover, ecDNA-encoded super-enhancers (ecSEs) mediated most ecDNA–chromosome contacts and presented cancer-specific connectivity patterns preferentially associated with MED1-rich regulatory regions.⁸

Pharmacological disruption of nuclear condensates rapidly dissolved ecDNA hubs, redistributing MED1 to chromosomal regulatory elements, reducing ecDNA-mediated chromatin interactions and cell viability, and altering gene expression (but not ecDNA copy number). Epigenetic silencing of ecSEs not only reduced enhancer activity and ecDNA-dependent chromatin interactions, but also silenced ecDNA-associated target genes, resulting in impaired cancer cell proliferation. These findings support a model of ecDNA-associated oncogenesis involving the MED1- and ecSE-dependent formation of ecDNA condensates that transactivate genes driving proliferation.⁸

Taken together, these studies shed new lights on the molecular mechanisms underlying ecDNA origin, inheritance and oncogenic activity (Fig. 1), de facto repositioning ecDNA from a passive by-product of CIN to an active driver of cancer evolution and a therapeutically actionable vulnerability. It will be important to determine whether ecDNA molecules that are most efficiently retained through mitotic rounds exhibit prominent enhancer activity. Moreover, while mitotic tethering mechanisms can enhance ecDNA retention, CIN is expected to continuously promote ecDNA copy number changes and de novo generation, reshaping the ecDNA pool across cellular divisions. This suggests that additional buffering or stabilizing mechanisms may contribute to the long-term persistence of ecDNA-associated phenotypes. In this context, continuous karyotypic shifts may also alter the nuclear environment for ecDNA condensation, implying that ecDNA transcriptional hubs are dynamic rather than fixed structureds. Finally, large-scale genomic analyses indicate that ecDNA frequently amplifies genes associated with poor T cell infiltration,⁹ suggesting a potential role for ecDNA in immunoevasion.¹⁰ Whether targeting ecDNA might increase tumor immunogenicity, however, remains to be determined.

Fig. 1: ecDNA biogenesis, inheritance, and function.

a Chromosoma instability (CIN) promotes extensive chromosome missegregation coupled with the formation of relatively unstable micronuclei. Upon rupture of their envelope, the cytosolic nuclease N4BP2 acquires access to micronuclear DNA, resulting in extensive DNA damage, catastrophic chromatin fragmentation, and the generation of extrachromosomal DNA (ecDNA). b ecDNA persists across cell divisions through epigenetically regulated retention elements that enable tethering to chromosomes. c During interphase, ecDNA integrates into transcriptional condensates via ecDNA-encoded super-enhancers (ecSEs) that preferentially associate with the transcriptional coactivator MED1, reshaping higher-order nuclear architecture and activating oncogenic transcription programs.

Despite these uncertainties, ecDNA stands out as an active, targetable driver of cancer evolution rather than a passive consequence of genome instability, with significant therapeutic implications.