Engineering ecDNA: creativity, collaboration and the first ecDNA conference

Andrea Ventura’s lab at Memorial Sloan Kettering Cancer Center has always been interested in developing mouse models to try and understand how cancer develops. In addition to their work on microRNAs, the lab is known for inventive ways of modelling mutations in mice, having previously developed a clever way to use the CRISPR-Cas9 system to make chromosomal rearrangements such as inversions and deletions. But, as Andrea puts it, ‘What was bugging me was this class of mutations, gene amplifications, that nobody could model, that was a big gap in the mouse genetics field.’ 

Back in 2020 during the pandemic, unable to go to the lab and stuck at home, Andrea had time to step back and think about what was next for his lab. He was reading old papers when he came across one from Paul Mischel’s lab describing ecDNA. He remembers being surprised that nobody had modelled ecDNA: ‘Immediately I knew how to do it and that this was the way to model gene amplification. Plus, we already had all the tools in the lab!’

When making knockout models using the Cre-lox system, it was known that whatever gets removed from the genome also get circularized in the process. But usually, these byproduct circles are lost. This is where Andrea’s idea came in: ‘I thought maybe if we make a large enough circle that contains important genes, they will not be lost, they will become extrachromosomal DNA and will actually amplify.’ 

Now the team has published its work in Nature, finally providing a way to model gene amplification that more accurately reflects the evolutionary process taking place in patients, as well as reflecting the peculiarities of ecDNA that are emerging.

The team has harnessed the Cre-lox system to induce ecDNA with precise spatiotemporal control in both cell lines and mice. Andrea and Rui’s initial experiments were a success- in cell lines, by using a combination of fluorophores, the system simultaneously allows the tracking of both the ecDNA and the chromosomal DNA that it was originally excised from. This provides a useful setting to study ecDNA dynamics under a variety of conditions.

eDyNAmiC went on to make two mouse models of the most common ecDNAs found in cancer. Rui generated the first genetically engineered mice with inducible Myc ecDNAs before graduating, which were then characterised by Davide. Minsi, a radiation oncologist with an interest in sarcomas, generated the second key mouse model harbouring inducible Mdm2 ecDNAs.

Using embryonic fibroblasts from these mice, Davide could show that ecDNAs can and will accumulate very quickly in primary non-transformed cells, given the right selective pressure. Andrea remarks, ‘We didn't know previously whether normal cells could tolerate the presence of ecDNA.’ The work also emphasises that formation of ecDNA per se is not tumourigenic, and that it needs many rounds of cell division and continuous selection for gene amplification to occur. Andrea comments, ‘ecDNA are really unique compared to all other mutations we see in cancer, because they're not an on/off event.’ 

Minsi then demonstrated that Mdm2 ecDNA can cooperate with HRAS expression to drive immortalisation in vitro. Critically, the last piece of the puzzle was provided by graduate student Melissa who, together with Minsi, went on to show that Mdm2 ecDNA can also drive tumour formation in an autochthonous fashion, providing the second hit in MYC-driven hepatocellular carcinoma. The work thus provides important evidence for an early role of ecDNAs in tumour initiation.  

ecDNA also amplifies immunomodulatory genes, but this phenomenon has so far only been seen in patients in the presence of an immune system providing the necessary selective pressure. Critically the mice the team has now generated are immunocompetent, so with these models, and others in development, it can now ask how does ecDNA modulate the immune system? Andrea muses, ‘In a way these ecDNAs look a lot like viruses, these circular genomes. And so you could imagine that the immune system might find ways to fight cells that have ecDNA.’

The team is also working to model multiple ecDNAs simultaneously in vivo to follow how they evolve, cooperate and co-segregate, as well as incorporating the fluorophore tracking system within the mouse models.

The system will also be invaluable for combining with CRISPR screening approaches in vitro in order to look for factors involved in ecDNA formation and maintenance, and then subsequently to screen the small molecules that eDyNAmiC is developing to target these processes.  

The generation of these mice is an important step along the path to advancing potential therapeutics into the clinic, as they provide important pre-clinical models of ecDNA amplified cancer. Andrea’s lab has deposited the mouse models generated to the Jackson laboratory where they will be available for researchers worldwide. Andrea does this for all the mouse models his lab creates, ‘It’s very important to us that the research community has immediate access, the more people that use them, the better.’

Andrea’s attitude matches that of the team, ‘eDyNAmiC has been incredibly welcoming and open, so collaborative, and it's been really fun to do science in this context.’ He has also benefited from the input of the team’s patient advocates, ‘You get obsessed about the little experiments you're doing, and it's very easy to lose track of why you're doing them. The patient advocates remind you it's much more important than publishing a paper in Nature, it's about finding new ways to treat patients and to improve their prognosis.’ 

And in the same vein, Cancer Grand Challenges is delighted to announce the first edition of its conference series, which is also the very first meeting focused on ecDNA. Taking place on 11-13 June in London, the purpose of the meeting, in addition to sharing cutting-edge science, is to further enhance and build upon an open, collaborative, and inclusive community that will advance the science rapidly for the benefit of patients.

The conference is being co-organised by Paul Mischel, Howard Chang and Mariam Jamal-Hanjani, with Andrea part of the stellar line up of invited speakers, who are importantly from both within and beyond the eDyNAmiC team.

As Andrea puts it, ‘There's been this flurry of really ground-breaking papers that have answered some questions, but at the same time, uncovered new questions that need to be answered, so the conference comes at the right time. My hope is that we'll expand the field, and there will be new people that make the community even bigger and more exciting than it already is.’ 

-----------------------------------------------------------