Scientists Unwind the Evolutionary History of Cancer

It’s easy to think of “having cancer” as a binary situation: you either have it or you don’t. But cancer is far more complicated than that: it’s a disease caused by our own cells run amok, not some foreign invader entering our system. A tumor represents the buildup of cancer over time. Individual cells acquire certain mutations, sometimes very slowly, and manage to evade destruction by our immune system. Eventually those mutations can snowball, with cells taking on more and more of the traits of cancer. Finally, the cells grow out of control, having more of an impact in our bodies. That’s usually when cancer is finally diagnosed, though the process leading up to it may have been going on for years.

But scientists and doctors know that any glimpse of cancer is just a snapshot in time; what they’d really like to do is rewind the show and observe the development of each person’s cancer, noticing which mutations appeared when, where, and how often. Information like that could be incredibly valuable for predicting the path of a specific cancer — how aggressively it will grow, whether it will spread, whether it’s likely to be resistant to common treatments.

This information has been largely inaccessible until now. In a fascinating new study published in the journal Nature, scientists from Europe and the U.S. developed and evaluated a new method to reverse-engineer the evolutionary history of cancer based on its current molecular profile.

The approach relies on DNA methylation, an intimidating term that’s actually really easy to understand. You’ve heard that DNA is like a biological code, written with four building blocks (usually shortened to their initials: A, C, G, and T). Methylation is another layer of code sitting on top of our DNA. It doesn’t change the underlying DNA, but it does change how that DNA is read by our cells. Methylation regulates which genes or parts of genes are read, silencing certain genes and activating others. And here’s the important part for the current study: methylation changes over time, affected by our environments, stress levels, and more.

In this new research, scientists took advantage of that, using methylation changes to unwind the evolutionary clock in cancer cells. They identified “evolving barcodes” based on low-cost methylation analysis technology and created a computational tool to use those barcodes to infer the biological history of cancer cells.

With this new method, the scientists analyzed some 2,000 blood cancer samples collected from patients with known clinical outcomes. Their question: could the evolutionary history gleaned from methylation data allow them to predict cancer progression? Results showed a strong correlation between cancer’s history and each patient’s progression.

For example, they found that pediatric blood cancers were much more aggressive than adult blood cancers — a trend that matches what’s seen in the clinic. Even narrowly defined subtypes of cancer could be distinguished by their methylation profile, with high-risk subtypes associated with faster growth than subtypes known to be lower-risk. For one particular cancer trait that’s known to give patients extremely poor outcomes, the evolutionary analysis revealed that the molecular underpinnings of that trait had been present in the patients more than 30 years before the cancer was diagnosed.

This project was a research study, and it could be years before it has any impact for patients. But it’s a really promising report that could one day make it possible for oncologists to hone the prognosis for each patient and possibly even select more tailored treatments based on the history of that person’s cancer.