A study led by CBGP researchers shows that interactions among cancer mutations determine tumor properties and patient survival

Cancer is the consequence of an evolutionary process that takes place within a human lifetime. The process starts when some cells acquire mutations that allow them to divide faster by overturning the mechanisms that restrain cell proliferation. As a result, the number of mutant cells increases, facilitating the appearance of more mutations. Eventually, this may end up with a tumor that invades the surrounding tissues and spreads to other parts of the body. Over the last decade, cancer researchers have started applying tools and concepts from evolutionary biology to understand how tumors develop.

In a recent study with participation of the National Institutes of Health and Georgia Tech (USA), researchers from the CBGP investigated the interactions among different mutations (a phenomenon called epistasis) and the effect of such interactions on cancer progression. "When one looks at the genomes of cancer cells, one finds that some mutations tend to appear together and others tend to be mutually exclusive", says Jaime Iranzo, junior group leader at the CBGP and main author of the study. "We wanted to understand why such patterns arise and what is their effect on the development of tumors".

The authors focused on a phenomenon called conditional selection. "If a mutation promotes cancer growth, we will observe it in tumors more often than we would expect by chance. Thus, by looking at the abundances of mutations, we can infer which ones really contributed to cancer", explains Jorge Calle, postdoctoral fellow at the CBGP and coauthor of the study. By applying this principle, the researchers found that some mutations only promote tumor growth if they occur together with mutations in a different gene. On the other side, mutations that would lead to cancer by themselves can lose their effect if other mutations are present. Such interactions of dependency and antagonism affect most of the key genes known to intervene in cancer. In words of Jaime Iranzo, "these results imply that cancer prognosis and response to gene-targeting therapies are better predicted by the combined patterns of mutation in multiple genes than by the presence or absence of mutations in single key genes". From a developmental perspective, the mutations that occur early in the tumor determine which mutations will come next, driving tumor development in one direction or another. Thus, long-term tumor properties, such as aggressiveness and metastatic potential, could be forecasted by identifying specific mutations in early pre-cancerous lesions.

Beyond cancer research, the tools developed in this study can be applied to understand how conditional selection shapes other evolutionary processes of clinical, biotechnological, or environmental relevance, such as the development of resistance to therapies by pathogens or the adaptation of crops to climate change.