Image from Flickr under CC BY-NC-ND 2.0

The study of early human remains – and the DNA within them – has given scientists vital insights into our origins and the migration of our ancestors across the globe. Some tribes flourished and went on to conquer entire continents. Others floundered and eventually died out.

But what we see in the changing human population over thousands of years also holds true for evolution on a much smaller, more personal timescale. And the same tactics can be used to understand the evolution and spread of cancer within the body.

We’ve written before about research on tumour evolution. Scientists around the world are now revealing an increasingly detailed picture of how the DNA in the different ‘tribes’ of cells in a patient’s tumour change over time as the disease migrates around the body.

As a result, in recent years we’ve seen huge leaps forward in our understanding of how this happens in a range of different cancer types including kidney, lung, ovarian and pancreatic cancers, to name just a few.

One of the driving forces behind all this has been the International Cancer Genome Consortium (ICGC) – a collaborative global project to catalogue the DNA changes across dozens of cancer types in unprecedented detail.

Since 2011, Cancer Research UK has supported ICGC projects looking at oesophageal and prostate cancers. And these projects are now starting to make significant inroads into our understanding of these diseases.

The prostate cancer team is led by scientists at The Institute of Cancer Research in London, the Wellcome Trust Sanger Institute in Cambridge, and the University of Tampere in Finland. Just a few weeks ago, their analysis of cancer’s DNA revealed new insights into the disease’s origins within a man’s prostate.

Now they’ve published another new finding in the journal Nature – this time about how the disease spreads around the body, and how it evolves to become resistant to treatment.

This is vital work – prostate cancer is the fourth leading cause of cancer death in the UK. And although several new treatments have recently become available, there’s still an urgent need to work out how to treat men when the disease has spread.

So let’s have a look at what they found, and the implications it has for how prostate cancer is managed and treated.

Who do you think you are, prostate cancer?

To start with, let’s put this in context. Conventional wisdom – and current research – suggests (or at least assumes) that a patient’s cancer starts from one rogue cell. This splits into two bad cells, which split into four, then eight, and so on.

Over time, these evolve and change into clusters of genetically different cancer cells that can spread through the body. So a genetic family tree tracing this evolution would have just one ‘founding father’.

a prostate cancer cell

Prostate cancer cell

But prostate cancer is often different. All too frequently, the disease appears as pockets of genetically different cancers in more than one area of the prostate at once – so-called ‘multifocal’ prostate cancer.

In other words, the ‘family tree’ has multiple roots. And scientists have been puzzling over what causes this for some time.

But in their study published in early March this year, the ICGC team seemed to have solved this mystery. They revealed that, over time, normal-looking prostate cells accumulate lots of background genetic faults, even though they appear healthy under the microscope and are not overtly cancerous.

This means that any of these damaged cells can be nudged down the road to become a cancer. And it goes a long way to explaining why prostate cancer is frequently found to be a hotchpotch of distinct tumours within the prostate.

But how does this relate to how the disease migrates around the body? Do different ‘tribes’ lead to settlements in various organs? Or is there one ‘master race’ that emerges from the prostate to spread and colonise different parts of the body?

To find out, the ICGC researchers have meticulously pieced together what happens at the genetic level as prostate cancer evolves within the body, tracing out the ancestral history of a patient’s disease.

This was a huge undertaking. The researchers analysed the entire DNA sequence of multiple different cell samples, taken from 10 men with advanced prostate cancer. For each man, the researchers were able to analyse samples from his original prostate tumour and from the sites to which it later spread: 51 samples in all.

It’s the first time multiple different metastatic prostate cancer samples from the same individuals have been analysed in such detail.

Armed with this wealth of DNA data, they turned to a series of sophisticated computer tools to map out each patient’s prostate cancer’s ancestral history, showing how each cancer had evolved and spread beyond the prostate.

And this analysis has revealed three important things.

We are family

First, the different secondary tumours from each man’s cancer all tended to share a very similar set of ancestral DNA errors (although these differed from patient to patient).

In other words, all the secondary ‘colonies’ had the same ‘founders’, from the same ancestral site in the man’s prostate.

This suggests that, even though prostate cancers may begin independently in different parts of the prostate gland from different sets of genetic changes, only one or a few of these ‘founder clones’ manages to migrate and settle elsewhere. Gaining this ability to go on the move is a relatively rare event and not common to all cancer cells in the prostate – most stay put.

Although this is based on results from only 10 men, the researchers found the same process at work in all of them. And a further study from some of the same team, also published today, adds further weight to these results – again they see just one, or a few or the ‘tribes’ of cells in the prostate spreading then gaining footholds in other places.

The next finding relates to the mechanics of how prostate cancer spreads.

Current thinking about cancer spread holds that after a tumour arises, genetic changes occur that allow its cells to start moving round the body. Once they find a suitable location, they settle down, forming secondary tumours that stay put in their new home.

But, for prostate cancer at least, the ICGC team’s results suggest otherwise. In each patient, the secondary tumours were more closely related to each other than to the primary tumour. This suggests that – as the diagram below shows – once the first few first cancer cells escape from the prostate, they build a new settlement, carry on growing and evolving, and then spread further afield.


The researchers’ third observation raises some important questions about how prostate cancer cells colonise new locations in the body.

Although all the cells in a secondary tumour shared a common ancestor, in at least half of the 10 men studied, each secondary was made up from several different, less-related groups of cancer cells. This is fascinating, and suggests that they came from different places, and somehow met up in their new location.

Whether this is coincidence – for example, there’s something in their genetic wiring that independently attracts them to the same place – or whether the groups of cells somehow communicate and cooperate to form a new tumour is still unknown.

Close relations

As well as making important discoveries about how the disease spreads, the ICGC team also made a big step forward in understanding another important problem in prostate cancer: the evolution of drug resistance.

Because most prostate cancers grow in response to male hormones, such as testosterone, men are often treated with hormone therapy. This is designed to lower their testosterone levels, halting the disease’s growth.

Unfortunately, this treatment can eventually stop working and the cancer starts growing again. This is known as a relapse.

The men included in this study had all been given hormone therapy, and had all relapsed.

The scientists wanted to find out when resistance to therapy appeared. Did it arise during the course of disease, as a result of genetic changes picked up by certain groups of cancer cells? Or was it an in-built capacity for resistance, present in a subset of cancer cells at the start?

When they looked closely at the genetic data, they found that different groups of cells had become resistant in a range of ways in each secondary tumour.

Some of the cancer cells had gene faults that meant they could make more of the molecule called the androgen receptor. This allows them to respond to testosterone-related substances in the blood, helping them grow in lower hormone levels (as encountered during hormone therapy).

Other groups of tumour cells contained genetic faults that allowed them to override the need for testosterone altogether.

This suggests that hormone therapy resistance develops quite late on in the disease. Had it been an early event, all the tumour samples would share common mistakes in genes involved in hormone responses.

But although the specific genetic faults were different, they all affected molecules involved in how cancer cells sense and respond to testosterone levels.

So it seems that hormone therapy treatment causes an evolutionary pressure that drives cancer cells to evolve in the same direction – namely, to overcome the lack of testosterone and keep growing – albeit via different routes.

Cutting down the tree

This study has allowed researchers to track an individual patient’s disease through its evolutionary journey, tracing it right from its roots as it spreads and becomes resistant to hormone therapy.

Importantly, there are some intriguing implications from this research that might help to treat prostate cancer more effectively in the future.

The fact that most prostate cancer cells don’t gain the ability to move narrows down the ‘dangerous’ cells – the ones that can spread – to one or two ‘tribes’ that share the same ancestral origin. This means there might be a number of shared mutations in the dangerous cells that could be targeted with drugs – effectively cutting the main ‘trunk’ of the family tree to stop it spreading.

Furthermore, because each man’s cancer seemed to have a unique ‘family tree’, it adds further weight to the idea that tailoring therapy according to the genetic makeup of an individual patient’s disease – so called stratified or personalised medicine – is the way forward.

And finally, it suggests that prostate cancer drugs targeting different parts of the hormone pathway – including the new treatments that have recently become available, such as abiraterone (Zytiga) – might work better in combination than one after another. This is an idea that is being tested in clinical trials, for example our STAMPEDE trial.

And this last point is vital. Although these new drugs have made a big difference to survival from prostate cancer over recent years, we still need to do much more to improve how we use them to treat men whose cancer has spread.

Understanding the evolution of prostate cancer has given scientists a brand new insight into the way the disease grows and spreads within each individual man – information that may be relevant for other types of cancer too. Uncovering the biology underpinning how this happens is essential if we’re to find more effective ways to treat late-stage prostate cancer and save more lives.



  • Gundem, et al. (2015) The Evolutionary History of Lethal Metastatic Prostate Cancer. Nature. DOI: 10.1038/nature14347
  • Hong, et al. (2015) Tracking the origins and drivers of subclonal metastatic expansion in prostate cancer. Nature Communications. DOI: 10.1038/ncomms7605