I'm constantly fascinated by the fact that the cells that make up a cancerous tumor are each undergoing their own private evolution every time they divide. Unlike most normal cells, cancer cells are so wacky that even a small batch can morph into a highly variable mass within a few generations. As I wrote in a story last week:
In many ways, cancer cells represent biology's wild west. These cells divide rampantly in the absence of normal biological checkpoints, and, as a result, they mutate or even lose genes at much higher rate than normal. As errors accumulate in the genome, things go ever more haywire.
Recently, oncologist Hanlee Ji, MD, the senior associate director of the Stanford Genome Technology Center, and postdoctoral scholar Noemi Andor, PhD, devised a way to measure the extent of these differences among individual cancer cells and to associate their effect with the virulence of the disease as a whole. They published their results today in Nature Medicine.
As Ji, who is also a member of the Stanford Cancer Institute, explained in an email to me:
Until recently the scientific community believed that a typical tumor was composed of malignant cells with very similar genomes. The advent of next-generation sequencing technologies has revealed that this is not the case, and that most tumors are a heterogeneous product of ongoing evolution. This genetic heterogeneity also explains why therapeutic interventions in advanced cancers are often unsuccessful: some cells within a tumor develop resistance to therapies. Understanding the extent of tumor heterogeneity and how it leads to drug resistance is a major challenge in cancer biology research.
The researches used data from the National Institutes of Health's The Cancer Genome Atlas to compare portions of the genome in cells from over 1000 cancers of 12 main types. They then meshed their findings with data about each patient's outcome. As Ji explained:
We identified subpopulations of cells across 12 tumor types to reveal that the coexistence of multiple distinct subpopulations is a general phenomenon in tumor progression. We found that the presence of more than two subpopulations in the same tumor is prognostic of an increased risk of mortality. However, additional diversification beyond four subpopulations does not impart further risk. Our results suggest this nonlinear association is the consequence of a trade-off between the chance of acquiring a mutation that triggers the emergence of a new, fitter subpopulation and the risk of the mutation reducing the fitness of an already existing subpopulation. Now that we know this trade-off exists, cancer researchers may be able to target it to slow or rein in tumor evolution and prevent tumors from acquiring drug resistance.
Previously: Experts at Big Data in Biomedicine: Bigger, better datasets and technology will benefit patients and Is cancer too complex for targeted therapies?
Image by PeteLinforth
