Skip to main content

Together we are beating cancer

Donate now
  • Science & Technology

Translating breast cancer’s life history

by Henry Scowcroft | Analysis

22 May 2012

0 comments 0 comments

The Wellcome Trust Sanger Institute

The Wellcome Trust Sanger Institute in Cambridge

Nestling in the Cambridgeshire countryside, the Wellcome Trust Sanger Institute is a hot-bed of cutting-edge genetic research.

The Institute played a key role in deciphering the human genome at the turn of the millennium, and is continuing to make huge strides in mapping humanity’s evolutionary history, and the diseases that affect us.

As you might expect, a substantial portion of the Institute’s work focuses on cancer. Its director, Professor Mike Stratton, has a long history in this field, having helped lead the efforts to hunt down genes like BRCA2 and BRAF (work we’re proud to have supported).

And thanks to places like the Sanger, and other labs and institutes around the world, it’s an incredibly exciting time in cancer genetics research.

Just last month, we wrote about a landmark Cancer Research UK-funded study called METABRIC, led by researchers at our Cambridge Research Institute. METABRIC redefined breast cancer as ten separate diseases, based on how genetic mutations driving each tumour were related to how long women survived their disease.

This week, the Sanger’s researchers have built on this work, releasing three tour de force research papers, providing even more detail of what’s turning out to be a extraordinarily complex disease.

Nine new breast cancer genes

As we reported on our news feed, the first paper (published in Nature) described the discovery of nine new genes involved in breast cancer. It also confirmed one of METABRIC’s key findings – that breast tumours are phenomenally genetically diverse, but important similarities exist between them.

While METABRIC focused mainly on large scale flaws in DNA – so-called ‘copy-number variations’ (where whole stretches of DNA are repeated or missing) this study looked in more detail, looking for tiny faults in the entire complement of active genes – known as a tumour’s ‘exome’ – as well as copy number changes, across 100 different breast cancer samples.

Every cancer they analysed was different, and no two women’s tumours were 100 per cent identical. But as well as this diversity (“a sobering perspective”, in the words of Professor Stratton) there was also some good news.

Many of the faulty genes were found to belong to the same molecular pathway or process. For example, several genes belonged to a signalling network called ‘JUN signalling’. Others regulated the winding and unwinding of DNA – a process called chromatin remodelling.

Focusing on such similarities, rather than the daunting diversity, will help further refine how researchers and doctors classify and treat breast cancers in future.

Molecular mechanisms and family trees

But how do all these DNA faults arise, and in what order? The Sanger Institute’s researchers have begun to sketch out some tentative answers in two more papers, both published in the journal Cell.

These studies focused on a repeated DNA analysis of a single tumour from a woman with oestrogen receptor-positive breast cancer – the most common form of the disease.

The Sanger team mapped out DNA errors in samples of this tumour again and again, 188 times in all, cataloguing individual mutations present in each sample.

In the first analysis, they then sifted through the resulting data mountain using powerful computer programmes, originally developed as facial recognition software.

This allowed them to compare the context of each error in the DNA code – in other words, to look at the ‘letters’ either side of each ‘spelling mistake’ – to see if patterns emerged.

And they did. The researchers identified five clear signatures, each corresponding to a different biological process at work as the tumour developed. And they then looked for – and found – the same signatures at work in a further twenty tumours.

They also spotted an entirely new and perplexing phenomenon, which they termed ‘kataegis’ (after the Greek for ‘thunderstorm’), where specific regions of the tumour’s DNA had mutated extraordinarily rapidly at a particular time in its history.

But what was causing these signatures?

Detailing the damage


The researchers discovered 'thunderstorms' in the genome

The researchers compared what they’d seen to what’s currently known about DNA error and repair pathways, to see if they could work out what was going on.

But they could only definitively link one of the five observed signatures to a process known to be involved in cancer – the conversion of a modified version of a DNA ‘letter’ called cytosine (a ‘C’ in the DNA code) to thymine (a ‘T’), when the ‘C’ comes before a ‘G’ – a guanine.

They have their suspicions about what might be behind the other four (and excitingly, one of the prime suspects is a process we’ve blogged about before), but this is the starting point for a very exciting period of research. Knowing the precise molecular forces at work as a cancer develops is exactly the sort of intelligence that could transform our ability to fight cancer.

But the researchers didn’t stop there. Their second analysis tried to find out when particular mutations arose in a cancer’s history.

It’s all in the timing

If tumours contain genetically different populations of cells, that evolved from a common ancestor, then each sample from a tumour should give a slightly different set of results when it’s DNA is analysed.

So measuring how often different genes are mutated in different samples, mapping these gene mutations to different regions of chromosomes, and applying some pretty sophisticated maths to the results, should allow a tumour’s genetic past to be reconstructed.

And so, again, the researchers started off with data from the single tumour they’d analysed 188 times, but now they looked at how often each mutation or DNA rearrangements occurred in each of the 188 analyses.

And again, the researchers turned to sophisticated computer algorithms – this time to look for patterns to help them reconstruct the tumour’s ‘family tree’. And just as before, they repeated their analysis in twenty other tumours, testing and confirming their findings.

As anticipated, each tumour seemed to be made up of different populations (‘subclones’) of cells, each stemming from a common ancestor unique to each patient, and yet each subtly different. They also found that each tumour had a dominant population of cells, making up at least half of the tumour. And they were able to get a sense of the time-span over which these populations evolved.

Putting all this together, the researchers deduced that a tumour can exist for some time as a tiny colony of subtly different mutated, growing cells, until one colony, for reasons unknown, suddenly takes over, growing rapidly and – ultimately – causing symptoms.

But a second crucial finding, one that gives hope to cancer researchers everywhere, is that mutations in so-called ‘driver’ genes tended to occur early in a tumour’s lifetime – in the ‘trunk’ of the tumour’s family tree. This suggests that treatments aimed at these gene faults will continue to show promise.

What happens next?

More work is now needed to understand exactly how tumours originate, grow, evolve and diversify.

Professor Carlos Caldas, who led the METABRIC study with colleagues in Canada and contributed to last week’s Nature paper, said the new data would be vital to future genetic cataloguing of breast cancer.

“This will feed into other researchers’ work, including that of my own group”, he told us. In Caldas’s view, cataloguing and exploring these mutations in larger numbers of tumours will be essential.

As he points out, the researchers analysed a lot of DNA, but it only came from a relatively small number of patients – these findings need to be scaled up to thousands or even tens of thousands of samples. Although this research is groundbreaking stuff, our knowledge of how all this fits together is still, Caldas says, “preliminary”.

“We’re still only seeing part of the picture. We need to be looking at how cancer continues to evolve after its spread,” he argued, which means analysing samples of secondary tumours from patients with advanced disease.

And he also pointed out that current DNA sequencing technology isn’t perfect, and this meant some types of tumour couldn’t be analysed – particularly smaller tumours, or tumours with a high proportion of non-cancer cells. “We’ve not even begun to look at the whole spectrum of breast cancers – just the ones we can analyse using the tools we have available,” he said.

Nevertheless, he said, this is “an extremely important piece of work, and it’s exciting that so many recent discoveries are converging around the same ideas,” of diversity and underlying similarity.

Professor Charles Swanton, the London-based Cancer Research UK scientist whose work on evolutionary diversity in kidney cancer caused such excitement earlier this year, agrees that the work “provides crucial insight” into the range of genetic diversity in individual tumours.

“Importantly, these papers begin to shed light on the way in which this diversity appears – a key step towards working out how to stop it arising in the first place,” he told us. Ultimately, Swanton hopes, we’ll use this knowledge to improve outcomes for patients with breast cancer and other cancers. After all, the principles at work here are likely to be common to more than just one form of the disease.

So what do all these genetic studies say about our progress against this terrible disease?

We know that cancer is incredibly complex, yet driven by common principles.

We also know that tumours are diverse, ever-changing collections of genetically damaged cells, but that by aiming our expertise and technical knowhow at them, we can begin to understand, unpick – and hopefully conquer – the processes that cause them.

And we know that a new era of genomics research is finally allowing us to read cancer’s inner secrets, deciphering their life history in ever greater detail.

As the Sanger team said in their Cell paper, “The cancer genome is like a palimpsest, an ancient parchment that was frequently reused, each time retaining traces of what had previously been written”.

It’s a parchment we urgently need to translate, so that we can convert this breathtaking explosion of new scientific knowledge into treatments to help the cancer patients of the future.



  • Stephens, P. et al (2012). The landscape of cancer genes and mutational processes in breast cancer Nature DOI: 10.1038/nature11017
  • Nik-Zainal, S. et al (2012). Mutational Processes Molding the Genomes of 21 Breast Cancers Cell DOI: 10.1016/j.cell.2012.04.024
  • Nik-Zainal, S. et al (2012). The Life History of 21 Breast Cancers Cell DOI: 10.1016/j.cell.2012.04.023