Study provides evidence for theory on tumor evolution and its relationship to clinical outcomes

Keratin skin tumors

Keratin skin tumors
Photo Credit: Elisabetta Palazzo, Ph.D., NIAMS Laboratory of Skin Biology

For the first time, researchers used tumor samples to support a long-held theory about tumor evolution: At a critical point, mutation load, which accounts for the number of mutations in an organism and is a balance between driver and passenger mutations, leads to maximum tumor fitness. In this scenario, natural selection is neutral. Below that critical point, mutation load and selection favor tumor growth, while above it, the opposite is true. Tumors become less robust, improving clinical outcomes for patients.

The study, a collaboration between researchers from the Cancer Data Science Laboratory in the NCI’s Center for Cancer Research (CCR) and the National Center for Biotechnology Information (NCBI) at NIH, was published in PNAS on November 7, 2018.

Led by CCR’s Eytan Ruppin, M.D., Ph.D., of the Cancer Data Science Laboratory and NCBI’s Erez Persi, Ph.D., and Eugene V. Koonin, Ph.D., the team used samples from 6,721 tumors representing 23 cancer types in The Cancer Genome Atlas tumor repository. By analyzing the relationship between mutation load and clinical outcomes in all cancer types, researchers found that cancers can be divided into three groups: low-, medium- and high-mutation load. To demonstrate the action of natural selection, the study focused on acute myeloid leukemia as an example of an extreme low mutation-load cancer type and metastatic melanoma as an example of an extreme high mutation-load cancer type.

After performing a statistical analysis for each cancer type separately, researchers uncovered two opposing trends. Low mutation-load cancers have high numbers of driver mutations, which often lead to cancer and result in a poor prognosis; however, high mutation-load cancers have high numbers of passenger mutations, resulting in a better prognosis. Similarly, some tumors in the low mutation-load group displayed signatures of natural selection, and some tumors in the high mutation-load group, particularly melanoma, displayed evident purifying selection, with inhibited tumor fitness. Mutation load, however, has a greater impact on clinical outcomes, as most tumors evolve in conditions where selection is neutral.

“The accumulation of driver mutations in low mutation-load cancer types increases tumor fitness,” said Persi. “In contrast, high mutation-load tumors have acquired more passenger mutations. Eventually, the number of passenger mutations overwhelms the cancer genome, leading to tumor regression, presumably through mutational meltdown.”

These findings have important clinical implications. They help explain why immunotherapy is effective for high mutation-load cancers like melanoma. The number and type of mutations work in the patient’s favor, making it easier for the therapeutic agents to locate and attack the cancer. 

“This study opens the door to other avenues of research,” said Ruppin. “Over time, the model may be used to study the transition of tumors from primary to metastatic states and the impact of therapy on the relationship between mutation load and selection.”  

Summary Posted: 11/2018