Studying precancer to prevent cancer: ecDNA in Barrett’s oesophagus

Two things to know about cancer: it starts when the DNA in our cells mutates, and then it makes more mutations possible. 

It’s a moving target.  

Take any tumour. It will have grown from one type of cell, so it can seem like it’s all one thing. But, as it grows, that tumour can change to include dozens of altered versions of its first cell. Each one will have its own strengths and weaknesses. 

To treat cancer, doctors need to find a way to kill or control all those changing cells. By that point, it’s often too late for them to tell which mutations led to the disease and which ones followed on from it.  

Then again, for a doctor, the mutations are almost always bad news. Treatment is fighting the fire, not working out who lit it. 

But it doesn’t always have to be. Fires can burn differently depending on how they started – and, though they might try to destroy everything, even arsonists leave clues.  

By looking more closely at how cancers start, our researchers are working out how to stop them. 

The discovery of ecDNA in cancer 

Those researchers are part of eDyNAmiC, one of the Cancer Grand Challenges teams we funded with the National Cancer Institute in the US last year. Team lead Professor Paul Mischel is heading the investigation. He’s a doctor and a researcher. He knows all about fighting fires that are already burning, but he also thinks we can catch arsonists before they do any damage. 

The more we know about what mutations cause cancer, he says, the better we’ll be at treating and preventing it. 

And Mischel’s spent the last decade trailing some suspects: rings of genetic information that have broken free from their proper place in our cells – extrachromosomal DNAs (ecDNAs). Keep your enemies close; Team eDyNAmiC (extrachromosomal DNA in Cancer) is named after them. 

Scientists first spotted pairs of ecDNAs loitering in some tumour cells in the 1960s. Those duos looked a bit like handcuffs, which is odd, because, for the next half-century, they never got caught. Extrachromosomal DNAs – whether alone, in pairs or big clusters – show up in half of all cancer types, but they always seem to come with an alibi. No one could make a convincing case that they were causing the disease. 

How ecDNA links Barrett’s oesophagus and oesophageal cancer

In their latest paper, eDyNAmiC turned their attention to a precancerous condition (or precancer) called Barrett’s oesophagus. That’s when cells in the oesophagus, the tube connecting our mouths and stomachs, change to look more like cells in other parts of our digestive system.  

It’s hard to know exactly how common Barrett’s oesophagus is, but the latest estimates suggest it affects fewer than 2 in every 100 people. Very few of those cases (between 3% and 13%) turn into oesophageal cancer. Still, that outcome is so serious that people with Barrett’s oesophagus are routinely screened to spot if the cells are getting more abnormal (dysplastic).

Using screening data from patients in the UK and the US, the team looked at the changes that took place as some precancers transitioned to oesophageal cancer. They also compared them to the precancers that didn’t become malignant. 

That tactic has paid off. It looks like eDyNAmiC might have found a smoking gun. When patients have ecDNAs in their abnormal oesophageal cells, they’re very likely to develop oesophageal cancer.  

“This really shows that ecDNA is not just a latecomer of the genomic instability linked to cancer,” says Mischel. “We need to do more experiments, but it looks like it’s causing the development of the cancer. So, as we keep working on ways to target ecDNA with treatments, we may be able to think about the concept of cancer prevention, or early treatment.  

“We could have a way of intercepting cancer before it becomes a lethal disease.” 

Evolution, but not as we know it

Helpfully, this team works fast. It’s only been a year since Cancer Grand Challenges awarded eDyNAmiC £20 million to investigate the role of ecDNA in cancer. They’ve already discovered a drug that could be used to stop the vicious circles forming in the first place.  

We’ll keep you updated about that in future articles. But to understand just what this means, it’s worth taking a step back.  

Just 10 years ago, our reliance on chromosome-based cancer mapping tools meant no one even knew how to look at ecDNA. Now it’s turning 200 years of biology on its head. 

One of the other tools scientists use to understand how cancers develop is a version of Charles Darwin’s ‘tree of life’ – a kind of family tree for evolution. It tracks how slight changes to the genes on our chromosomes build up generation by generation until, eventually, one species splits into two. 

The clues in Barrett’s oesophagus: looking at precancer to understand cancer

Well, that’s how it should be.  

Because it’s broken loose from our chromosomes, ecDNA doesn’t have to pay much attention to ‘should’. 

Mischel knows that better than anyone. This study was about seeing what was actually happening – not just the chromosomal changes that conventional technologies make it easy to track. So eDyNAmiC looked to patients for answers.  

They found them at the Fred Hutchinson Cancer Centre in the US. Over multiple years, researchers there collected data from people with Barrett’s oesophagus. By looking back over it with the tools they’ve designed to find ecDNAs, eDyNAmiC were able to work out, step by step, what happened to turn some of those precancers into cancer.  

“It was a big moment, because rarely do you have the opportunity to see cancer developing,” says Mischel. “Usually, it’s a fully-fledged cancer by the time you see it. It’s genomic chaos, and you’re trying to backtrack and work out the order of events, which is why we make our phylogenetic trees. But that’s the problem. In this case, the tree’s wrong, because of the chromosomal assumption.” 

The patient samples showed that ecDNA can be a truncal driver of oesophageal cancer.  

That had been hidden because, unlike chromosomal DNA, ecDNA isn’t split equally every time a cell copies its DNA and divides. In fact, the inheritance of this abnormal DNA is shockingly random.

One mother cell with 6 circles of ecDNA can split into one daughter cell with 12 and another with 0, one with 5 and one with 7, or any other combination. That means closely related cells can begin to look very different, very quickly. The trunk can hide among the branches.  

More than that, because ecDNA can drop to undetectably low levels in some cells and then quickly start multiplying, every branch has the potential to become a trunk.  

At that point, when the entire tumour can rapidly change its DNA composition, there’s not much sense in talking about trees at all. 

One step ahead

This doesn’t change everything. Most cancers aren’t driven by ecDNA. But it’s a feature of many of the most aggressive and hard-to-treat tumours.  

It might seem discouraging to find out that some of our most established ways of studying cancer can’t keep up with ecDNA, but Mischel is optimistic. The ground might be changing under us, but with every step forward we’re pushing back against cancer. 

And this new ground could be the foundation for other breakthroughs, too. Evolutionary trees and chromosome-focused cancer maps are ways to deal with fires that are already raging, frameworks for studying things that have already been lost. Team eDyNAmiC have developed new tools that let us look forward instead.  

Now we can see what ecDNA is doing and work out specific ways to stop it. And, as eDyNAmiC begin to focus on targeting ecDNAs with drugs, this new way of understanding cancer could become a way to intercept and prevent it, too.  

“We’re really learning from patients,” says Mischel. “Thanks to them, our biological knowledge has jumped out ahead of our therapeutic toolkit – but the toolkit will be catching up.”

Tim