Shedding Light on ecDNA Behavior Using CRISPR-Based Live-Cell Imaging

Cancer is considered equivalent to a quasi-species that is subject to Darwinian evolutionary processes, as laid out by Peter Nowell in 1976 (1). This prevailing model of cancer evolution is based on the assumption that all genetic cancer alterations are inherited equally and stably to daughter cells during mitosis. Extrachromosomal DNA (ecDNA) has emerged as a major cause of oncogene amplification that challenges these assumptions (2), as it lacks a centromere and is detached from its chromosome of origin (3), allowing unequal segregation to daughter cells during mitosis (4). Relieved from chromosomal positional constraints, ecDNA can change its position within a cell and has been observed to interact with other ecDNAs and chromosomal DNA in trans (5) and to be preferentially ejected from nuclei into micronuclei following genotoxic treatments. Furthermore, ecDNA can reintegrate into chromosomes, disrupting genome stability (6). Although imaging and sequencing methods for the detection of ecDNA exist, such conventional methods offer only a static readout and do not allow tracking the aforementioned spatial and genomic ecDNA mobility over time.

In this issue of Cancer Discovery, Yi and colleagues (7) present a CRISPR-based method, termed ecTag, for the detection and tracking of ecDNA in live cancer cells (Fig. 1). This method is based on the Casilio system that combines dead-Cas9 and Pumilio RNA binding for the recruitment of fluorescent clover protein molecules to a single-guide RNA (sgRNA; ref. 8). By designing sgRNAs targeting ecDNA-specific breakpoint sequences and employing Casilio, the authors visualize ecDNA harboring the EGFR oncogene in cancer cell lines. Comparing ecTag with signals from ecDNA labeled by fluorescence in situ hybridization (FISH), the authors confirmed the accuracy and specificity of their method. Next, the authors combined ecTag with live-cell imaging to investigate the mitotic segregation behavior of ecDNA. In line with previous cytogenetic reports (4), the authors observed that whereas chromosomal DNA segregates evenly to daughter cells during mitosis, ecDNA follows uneven segregation patterns. Intriguingly, the authors also observed that ecDNA clustered into hubs in interphase nuclei, which was not observed for telomeres labeled using the Casilio system. The authors reasoned that this clustering may contribute to the aberrantly high oncogene expression from ecDNA previously observed in cancers (9). To test this, colocalization of ecDNA hubs with known nuclear bodies of high transcriptional output was examined. Indeed, both Cajal and promyelocytic leukemia bodies significantly interacted with ecDNA hubs compared with chromosomal DNA, and this was accompanied by colocalization of hubs with RNA polymerase II. The overall RNA polymerase II signal associated with ecDNA hubs was positively correlated with the number and size of ecDNA hubs. A positive correlation was also observed for 5-ethynyl uridine–labeled nascent RNA in ecDNA hubs, suggesting that ecDNA hubs represent sites of active transcription. Last, the authors tested whether ecDNA hubs could be an origin of high oncogene transcription by measuring EGFR mRNA abundance in ecDNA hubs. Indeed, EGFR mRNA positively correlated with the size of ecDNA hubs in cells harboring EGFR on ecDNA, further supporting the authors' conclusion that ecDNA hubs serve as sites of active oncogene transcription.

The method described in this study has several advantages over conventional ecDNA imaging techniques. First and foremost, ecTag allows quantification of ecDNA copy number over time, which will enable investigations of ecDNA dynamics under changing environments, such as anticancer therapy. ecTag may also be used for investigations into the mechanisms contributing to ecDNA mobility within a cell and could help uncover factors contributing to ecDNA expulsion into micronuclei. As the use of ecTag is not restricted to the labeling of large, copy number–amplified ecDNA, it should in principle also be applicable to other DNA such as small extrachromosomal circular DNAs, which may enable investigations into their largely unknown functions.

This study raises several exciting questions fundamental to the biology of ecDNA. For example, although the work here suggests that ecDNA segregates unequally but stably to daughter cells, it will be important to understand how cells ensure that ecDNA are not lost during cell division (e.g., in micronuclei). Tethering of ecDNA to chromosomes during mitosis, as observed cytogenetically in the past, could represent one possibility through which stable ecDNA transmission is ensured. In addition, the functional impact of the observed intercellular ecDNA copy number heterogeneity, resulting from unequal mitotic segregation, still remains to be determined. Given that oncogenes expressed on ecDNA are critical for the malignant phenotype of cancer cells, it will be paramount to study how differences in oncogene expression that result from ecDNA-driven copy-number heterogeneity affect cancer cell behavior.

The publication of the article by Yi and colleagues (7) was concurrent with a publication by Kung and colleagues (10), both describing, for the first time, that ecDNA cluster in transcriptionally active hubs. This concept challenges our understanding of oncogene regulation, as it argues for ecDNA hubs and not individual ecDNA molecules as the functional unit of aberrant oncogene expression in cancer. Future studies interrogating the factors involved in hub formation, maintenance, and function will likely reveal important insights into oncogene regulation, which may represent important therapeutic targets in ecDNA-harboring cancers.

Overall, the work of Yi and colleagues (7) not only provides a valuable method for the detection and tracking of ecDNA in live cancer cells but also presents compelling evidence that ecDNA are randomly segregated during mitosis and organize in transcriptionally active hubs.

No disclosures were reported.