Many cancer cells contain numerous copies of small DNA circles that are not part of the chromosomes. This unusual ‘extrachromosomal DNA’ (ecDNA) carries cancer-promoting genes and has been associated with tumour drug resistance and poor patient outcome. The presence and amount of ecDNA, as well as the specific gene sequences each circle contains, vary substantially from one cancer cell to the next even within the same patient and are important factors that can influence tumour growth. However, the lack of a method for directly observing the behaviour of ecDNA in living cells has limited our ability to understand how it contributes to cancer progression.

DNA sequencing has shown that the ecDNA circles contain copies of chromosomal segments that are connected by ‘breakpoints.’ Because the breakpoint sequences are unique features of ecDNA compared with parental chromosomes, we developed ecTag, a method that generates breakpoint-specific fluorescent signals that can be tracked in live cells. The workflow of this method includes several steps. First, ecDNA-specific breakpoint sequences in both patient-derived cancer cells and a cancer cell line are identified and confirmed. Then, Casilio, a CRISPR-based genome imaging tool, combined with a programmable Pumilio RNA-binding protein system, is designed to fluorescently label the breakpoint sequences for live cell imaging.

Using ecTag to track ecDNA in live cancer cells, we made two significant observations. First, we directly observed that when a cell divides, ecDNA circles are not equally distributed between the two daughter cells, starkly contrasting the behaviour of the chromosomes in dividing cells. Thus, substantial variation in the amount of ecDNA among cells is obtained within only a few cell divisions. Second, multiple ecDNA circles frequently gather together and form ecDNA ‘hubs’ where the cellular machinery for gene transcription becomes concentrated in the cell nucleus. As a result, cancer cells that contain ecDNA hubs are able to transcribe more cancer-promoting genes than those cells in which ecDNAs do not form hubs. These findings shed light on how ecDNA contributes to substantial variation among cancer cells not only at the genetic level but also at the transcriptional level.

ecDNA contributes to substantial variation among cancer cells

While the dearth of knowledge on how ecDNA forms overall limits the development of therapeutic methods that could prevent ecDNA formation, a fitting alternative would be to specifically perturb formation of ecDNA hubs. Uncovering the mechanisms of ecDNA hub formation will require a live-cell imaging-based approach to capture the dynamic behaviour of ecDNAs. Therefore, the ecTag method is well suited for this important future work.