Tumour evolution - Professor Charles Swanton
Charles - We have to determine how representative a single tumour biopsy was of the entire tumour genomic landscape. And by that I mean if you identified mutation in one part of the tumour, how likely are you to identify that mutation in another part of the tumour? Because clearly, if we're going to identify new ways of predicting drug response, we have to go for genetic events that are common throughout all parts of the tumour so that a random biopsy could pick those genetic events up.
And what we found actually is that, in the study when we took multiple biopsies from several primary renal cancers, and in some cases their metastatic sites, that the number of genomic aberrations or the pattern of genomic aberrations very often wasn't shared across all biopsies. And in fact, when we counted the number of mutations, about two-thirds of the entire mutational load in the first two tumours as we sequenced were not shared across every biopsy taken from those two tumours.
Kat - So this is quite a big difference from the original tumour. You find differences in the gene faults within the tumour and also, in secondary cancers where it spread. Presumably this almost blows personalized medicine out of the water, what do you think is going on here?
Charles - No, I definitely don't think this blows personalized medicine out of the water. I think what this does is actually begins to inform us that if we consider tumours as trees where we have the common mutations represent the trunks of the tree, these mutations are present at every site of the disease. And the mutations that differ from one region to the next represent the branches.
Now, in terms of personalized medicine, one can perhaps think of HER2 or the BRAF mutation as the mutation or the amplification event that occurs at every sites of the disease and is therefore a good drug target.
Kat - So that's the trunk?
Charles - The trunk, exactly. That controls the disease effectively because that mutation or genetic aberration is present at every site of the disease.
The model taken through is sort of logical conclusion also helps to perhaps think about how drug resistance is acquired, that low frequency events within the tumour that may confer drug resistance upon the tumour. Our equivalent to the mutations perhaps that occur in the branches of this tumour to take the tree analogy further. And during treatment, those resistance mutations may be selected out enabling the tumour to evade drug treatment.
Kat - So when you cut off some branches it just keeps growing.
Charles - Exactly. That's right. It's like cutting off some branches.
Kat - What does this mean for the field of genomics research in cancer?
Charles - First and foremost I think we need to sequence fewer tumours in greater depth and at multiple regions, comparing primary with multiple metastatic sites to really truly identify those driver trunkal mutations that are present and responsible for the disease biology at every site of the disease. Those can be very efficient drug targets one would imagine.
It also tells us that actually understanding what's driving the diversity, in terms of the tree model, what's driving the changes between the trunk and the branches, actually maybe very important for trying to restrict underlying tumour diversity, to stop, if you like, the target from moving. Because it's the diversity within the tumour that is likely responsible for tumour adaptation, resistance to treatment, and potentially the acquisition of mutations that allow tumours to grow at sites distant from the original tumour.
I hope in the next 10 to 15 years, considering tumours as trees, trunks and branches, it might help to improve the development and discovery of new drugs and enable us to understand how resistance to treatment is acquired by studying how these branches are cut off.
Kat - You mentioned that you're sequencing the whole genomes of several samples of a single cancer, how have advances in technology allowed us to do this kind of research?
Charles - I mean we wouldn't be able to do this research two or three years ago. I mean the advances in sequencing machines, in computer technology, have been immense over the last two to three years and continue to be dramatic and will continue in the future to change and enable us to sequence tumours and process the data at an unprecendented rate
We're already seeing in some centres, in the US and elsewhere offer patients access to whole genome sequencing facilities as part of their treatment. So already we're seeing human genomes sequencing impact upon patient care.
Kat - And finally for you, where next for this research and for your team?
Charles - What we're very excited about is the implications of this research to our understanding of how resistance is acquired during treatment. Because it's only by understanding resistance to treatment that we will really make progress in prolonging patient benefit from the drug treatments we already have in the clinical setting.
So it's really understanding what's going on in branches of the tumours, it's understanding how the branches changed through therapy, and most importantly I think, understanding what initiates the change from the trunk to the branches, what is basically creating the underlying tumour diversity.
And when we've addressed to all of those problems, I think we'll have a much greater understanding of kidney cancer as a human disease and our, hopefully ability to target it much more effectively.