Timing cancer evolution
Timing the acquisition of key driver mutations not only provides meaningful biological insights, but can help patients understand the often long latency in tumour development and reconcile any possible delays in presentation to healthcare professionals.
As an oncological surgeon, I am extremely privileged to help patients make important decisions in regard to how best to manage kidney tumours. These consultations often involve discussing the initial diagnosis of cancer, before mapping out all of the possible treatment pathways available. One important question commonly asked by patients centres around how long the tumour might have been present, and it is easy to imagine ongoing ruminations on ignored symptoms or any possible delays in the diagnostic or treatment pathways.
Inferring when key mutational events occur, through "molecular archaeological" techniques might just help answer some of those questions.
One of the aims of the TRACERx consortium, a large group of clinicians and scientists, was to understand tumour evolution and progression through the comprehensive collection and genetic sequencing of well-annotated tumour samples over a large number of patients.
I was fortunate to be asked to analyse these multi-regional whole genome sequencing data at the Wellcome Sanger Institute.
As targeted multi-regional sequencing and single region whole genome sequencing data had already been published in kidney cancer, I was interested to find out what insights the combination of these techniques could show.
We saw that the hallmark genetic event (an event that occurs in almost all of the most common type of kidney cancer), loss of most of one arm of chromosome 3, often occurred in conjunction with the gain of another chromosomal arm. This allowed us to time when the event occurred. Staggeringly, our modelling showed that this key event appeared to occur in late adolescent years, often 4-5 decades prior to clinical presentation.
As loss of the short arm of chromosome 3 occurred so early in life, we pondered whether the die was cast at such an early age, whereby any person who has such an event will be sentenced to developing kidney cancer in the future. Alternatively, these chromosomal events could be relatively commonplace in small numbers of cells and they may not confer a greater risk of forming kidney cancer in our advanced years.
To help answer this conundrum we took inspiration from the assumptions that Knudson used when he was formulating the 2-hit hypothesis. In his modelling, he used epidemiological data and an estimated number of retinal ganglion cells to calculate the mutation rate.
In our dataset, we know the mutation rate of the key cancer gene in kidney cancer (VHL) and we can therefore use epidemiological data for inherited and sporadic kidney cancer to model how many cells are lose the short arm of chromosome 3 and are at risk of tumour development.
Again, we were surprised by our findings, estimating that by the end of our adolescence each of our kidneys contain several hundred cells that contain the hallmark genetic event associated with the formation of kidney cancer.
Although I remain unable to definitively answer questions as to how long tumours have been present, these findings are beginning to unpick evolutionary trajectories and give an indication of the lengthy time-scales of tumour formation.