Extrachromosomal DNA in cancer cells can strengthen MYC activity

Extrachromosomal DNA in cancer cells may behave less like a collection of isolated rings and more like parts of a shared molecular hub. In a colorectal cancer cell line, these rings came together inside the nucleus and created a particularly favorable regulatory environment for MYC—but not for every copy of the oncogene.

Extrachromosomal DNA, or ecDNA, consists of large circular pieces of the genome that exist separately from conventional chromosomes. Cancer cells often use these circles to carry extra copies of oncogenes together with regulatory elements that control their activity. This has made gene dosage the obvious explanation for high expression: more DNA copies should produce more RNA and, ultimately, more protein.

Copy number, however, is not the whole story. The three-dimensional arrangement of the circles—and the regulatory regions they physically contact—may also determine how strongly an oncogene is expressed.

DNA rings that act together

The study focused on COLO320-DM cells, a laboratory colorectal cancer line containing MYC-bearing ecDNA with several MYC- and PVT1-derived regions. The investigators built a minimal polymer-physics model in which each DNA circle was treated as a flexible chain. BRD4-binding sites were positioned along that chain, while BRD4-containing complexes could connect distant sites.

BRD4 is associated with active regulatory regions of the genome. In the model, BRD4 complexes acted as transient molecular bridges: they brought distant segments of the same circle together and also linked separate circles. Once enough of these interactions formed, dispersed ecDNA rings collapsed into a dense cluster.

The authors describe this transition as phase separation. In simple terms, molecules that were distributed throughout the nucleus become concentrated in a localized, membrane-free compartment. The important point is that contacts within this condensate were neither random nor evenly distributed.

The model predicted specialized interaction neighborhoods between different DNA circles, termed in-trans associated contact domains, or I-TADs. They emerged around particular arrangements of strong BRD4-binding sites. As a result, MYC copies located at PVT1–MYC fusion regions repeatedly encountered regulatory elements. A canonical MYC copy present on the same ecDNA circles did not receive the same advantage.

More than extra copies

That selectivity matters. If clustering merely packed many equivalent circles into the same space, each MYC copy should benefit to roughly the same degree. Instead, the model indicated that both copy number and the placement of regulatory elements on each circle determine the transcriptional effect.

As cluster size increased from one to ten ecDNA rings, regulatory contacts involving the PVT1–MYC regions rose by about threefold. Most of the gain was already achieved when roughly six circles had assembled, after which the effect plateaued. The proposed explanation is steric crowding: once a condensate becomes dense, additional rings have limited access to the most favorable interaction sites.

The predicted architecture was compared with two independent datasets measuring active-chromatin contacts: H3K27ac HiChIP and RNA polymerase II ChIA-PET. In the second comparison, the model and experimental contact maps showed correlations of 0.72 by Pearson analysis and 0.67 by Spearman analysis. This does not mean that the minimal model captures every molecular component, but it indicates that the model reproduces a substantial part of the observed ecDNA contact pattern.

The researchers then tested a further prediction. JQ1 prevents BRD4 from binding efficiently to acetylated chromatin. If BRD4-mediated bridging holds the circles together, inhibiting that interaction should dissolve the cluster and reduce transcription of genes located within I-TADs.

In COLO320-DM cells, RNA levels of MYC and two additional ecDNA genes declined in a nonlinear, switch-like manner, with a sharp transition around 100 nanomolar JQ1. A comparable response was not seen in the related COLO320-HSR line, where amplified MYC copies are integrated into a chromosome rather than carried on ecDNA. Additional controls indicated that this difference could not be explained simply by total BRD4 abundance or by the overall fraction of BRD4 bound to chromatin.

What the study does not establish

JQ1 was primarily a mechanistic probe in this work, not a demonstrated cancer treatment. The experiments were performed in cultured tumor cells, and the study does not show that disrupting ecDNA condensates would be safe or effective in patients. BRD4 controls many genomic programs, so systemic BET inhibition is not selective for ecDNA.

The central framework also remains a simplified physical model of one well-characterized ecDNA architecture. Real nuclear condensates probably involve additional components, including Mediator, RNA polymerase II and possibly nascent RNA. ecDNA structures vary widely among tumors, so the same contact pattern cannot be assumed to occur in every cancer.

There is also a technical ambiguity. Because ecDNA sequences closely resemble the chromosomal regions from which they originated, sequencing-based contact maps cannot always distinguish interactions between ecDNA circles from contacts involving the corresponding chromosome. Some of the functional support for I-TADs also came from aligning the model with previously published super-enhancer perturbation experiments rather than directly disrupting the predicted domains in this study.

The main contribution is therefore not the identification of a ready-made therapy. It is a more basic principle: the same number of oncogene copies can produce different transcriptional outputs depending on how those copies are organized inside the nucleus. ecDNA circles can build selective three-dimensional regulatory hubs, making genome architecture an additional mechanism for amplifying MYC activity.