What exactly is cancer?
We’re asked this question a lot.
The simple answer is that cancer is a disease caused by normal cells changing so that they grow in an uncontrolled way. The uncontrolled growth can cause a lump called a tumour to form, or rogue immune cells to build up in the blood.
By this definition, cancer is an incremental process during which healthy cells turn bad rather than a ‘thing’ that arrives fully fledged. But to a patient, cancer is very much a thing. Something affecting their health and their body, and something they want to be rid of.
We’re also all familiar with the language used to describe cancer – breast cancer, brain tumour, lung cancer, and many more. Now though, slowly but surely, researchers are chipping away at the traditional organ-based definition of tumours and dramatically changing our view of how cancer should be diagnosed and treated.
A huge US-led project involving over 250 scientists from across the globe recently released a series of scientific publications that typify this shift.
In this post, we look at how the key findings of the studies fit into the wider picture of cancer research and the work of our own scientists. Finally, and most importantly, we look at what this means for how patients could be treated in the future.
The research is part of a project launched in 2006 by the National Institutes of Health in America called The Cancer Genome Atlas (TCGA – a nod to the four ‘letters’ of DNA). This ambitious venture is probing the molecular characteristics of thousands of tumours to better understand the alterations that lead to cancer.
Through these painstaking analyses, TCGA researchers are building detailed molecular profiles of more than 20 different types of cancer. Last week’s publications come from their analysis of the first 12.
In what they’ve dubbed a Pan-Cancer initiative, the researchers have compiled information on several thousand tumours, developing a bespoke computer system to analyse and share this information across the world.
Beyond the genome
The result is a phenomenally detailed map of molecular changes seen in some of the most common cancers. And it throws up many promising leads for new strategies to tackle cancer.
We’ve written before about the DNA-scanning ‘genomics’ technologies that are helping to uncover new insights about cancer, and this latest work adds several more layers of valuable information beyond just the genome.
The researchers used some eye-wateringly complex technology to chart several types of molecular changes in tumour cells. They looked not just at the basic DNA code, but also the scale of gene activity (known as gene expression), and the suite of proteins that were present in the tumour cells (the so-called proteome).
They also recorded which tumours had missing or extra sections of DNA, known as copy number changes. And, finally, they looked at the spectrum of DNA ‘tags’ that can switch genes on and off, which scientists call epigenetic marks.
Under the hood
The project’s results are published in a group of papers in the prestigious journal Nature Genetics. One study, led by Dr Chris Sander from the Memorial Sloan-Kettering Cancer Centre in New York, applied a computational process to distil many thousands of genetic and epigenetic changes in tumours into around 500 ‘functional events’ – in other words, the ones that are likely to be important in contributing to cancer development.
The researchers then grouped individual tumours according to the ‘signature’ of functional events they shared. Surprisingly, they found that tumours tended to fit into two broad groups – those with non-inherited faults in the genetic code (known as somatic mutations) and those that had unusual copy number changes.
If a tumour had lots of somatic changes, it was almost certain to have very few copy-number changes, and vice versa. This unexpected trend could give insight into the fundamental processes behind how different types of cancer develop.
And within the two broad groups, tumours clustered into subgroups according to their shared genetic signature. For example, a type of lung cancer shared characteristics with a previously unrelated type of head and neck cancer.
The broad message from the publications coming out of the Pan-Cancer project is clear – just as two different models of car can be the same under the hood, tumours originating in different parts of the body can be driven by the same genetic engine.
The wider picture – transcending tissue
But what does this mean for cancer patients? Huge genetic projects aren’t merely an intricate exercise in biological stamp collecting – the more we understand about the shared characteristics of seemingly different cancers, the better equipped we are to treat them.
Many of the buzzwords of cancer research at the moment – such as ‘precision’ medicine and ‘personalised’ treatment – stem from a growing ambition to treat cancer not just by where it appears in the body, but according to its molecular make-up.
For some cancers, we’re already doing this. For example, the drug Glivec, which has been hailed as a ‘magic bullet’ for chronic myeloid leukaemia, was designed to work in patients with a particular genetic fault.
And the widely used breast cancer treatment Herceptin is given to women with a particular subtype of the disease that has an overabundance of a protein called HER2. Interestingly, Herceptin isn’t targeted ‘at breast cancer’ as such, but targets cancer cells that produce too much of HER2. As a result, it’s since been found to work in other HER2-containing cancer types such as stomach cancer.
Old dog, new tricks
But beside some very noteworthy successes, scientists have so far struggled to bring the promise of truly personalised treatment to the majority of cancer patients.
This could change in the coming years. Massive projects such as the Pan Cancer Initiative apply the logic of molecular-based tumour classification on a much larger scale and are helping researchers pinpoint similarities between different tumour types that have been missed by smaller studies.
One result is that we’re finding already available drugs could be used in different types of cancer, as well as discovering new molecular weaknesses in cancer that we can exploit.
Clinical trials 2.0
Another implication of such large-scale molecular-fingerprinting studies is that the way clinical trials are designed in future needs careful thought.
Our scientists are pioneering the design of ever-more sophisticated clinical trials that take into account the molecular differences between tumours of the same cancer type.
For example, Professor Tim Maughan in Oxford is running a flagship clinical trial for bowel cancer patients that will give participants different combinations of treatments based on their molecular profile and assess these against the current best treatment.
The design and analysis of such trials is hugely challenging, but crucial to translate the huge volume of molecular information researchers are generating into meaningful patient benefit.
The International Cancer Genome Consortium
The American-led TCGA is part of a larger global coordinated effort to catalogue the molecular changes in thousands of tumours, called the International Cancer Genome Consortium (ICGC). Our scientists are leading ICGC projects in prostate and oesophageal cancer, funded by the Dallaglio Foundation and the Catalyst Club, respectively.
These groundbreaking projects will yield valuable information about the molecular faults driving these cancers – diseases that weren’t included in the US team’s analysis.
Another distinction is that our scientists are using whole-genome sequencing technology, which scans the whole DNA code in cells rather than just the bits that contain instructions for cell components. The Pan Cancer studies used ‘exome sequencing’, which homes in on the 1.5 per cent of our DNA that encodes our genes.
Research this month has already highlighted that these non-coding (sometimes mistakenly called ‘junk’) DNA sections of the genome may actually have some very important functions in cancer, so it will be interesting to see what our teams uncover.
Tracing the evolution of cancer
Our scientists are also leading pioneering work to understand how the genetic landscape of tumours evolves over time. It’s becoming clear that as tumours grow and spread, their genetic make-up changes, so different parts of the same cancer can become genetically distinct in the same person.
This ties in to the idea of cancer as a process rather than a thing, and we need to better understand this evolutionary process in order to find ways to stop it.
Our researchers are on the case – for example, Professor Charles Swanton at our London Research Institute is running a multi-million pound project to track how lung cancer evolves over time as it is treated. His work will give us an unprecedented window into what makes this cancer so difficult to treat.
Know thy enemy
Our research, along with the work of many others across the globe, is painting an ever more complex picture of cancer. And the results are already highlighting the inadequacy of simple anatomical explanations of the disease.
But it’s unlikely we will ever completely move away from describing the disease as ‘breast’, ‘lung’ or ‘bowel’ cancer, as these labels are clearly useful descriptions for patients and doctors alike.
What matters to patients is they have cancer and they want it cured. And the crucial point is that research from the Pan Cancer initiative and the work of our own scientists is edging us ever closer to a new age of cancer classification, which will undoubtedly improve how we treat people with the disease.