A 21st Century Voyage on the HMS Beagle: The Study of Evolution in AML

It is now widely embraced that tumor initiation and progression follows a Darwinian process of natural selection of tumor clones and sub-clones derived from somatic cell mutations.¹  The phenotype and natural history of cancers are shaped by several factors including a spectrum of tumor-specific genetic and epigenetic changes that occur over time, the tissue microenvironment that may act to either promote or constrain tumor growth, and external “artificial selection” pressures, such as radio/chemotherapy. Using the language of evolutionary theory, cancers exist as part of complex ecosystems, wherein cancer sub-clones that are best adapted to their geographic niche are more likely to survive and reproduce.²  Selective pressures may promote the expansion or extinction of specific cancer cell populations, resulting in a clonal architecture that mimics Darwin’s branching evolutionary tree of species.

Death in AML is principally related to relapse due to the emergence of clones that have escaped the body’s mechanisms to suppress cancer, developed resistance to treatment, or both. In order to better characterize the molecular basis for AML transformation, Li Ding and colleagues from Washington University, St. Louis, utilized whole-genome sequencing to profile genetic mutations in paired samples taken at the time of diagnosis and at the time of relapse. Two basic patterns of clonal evolution emerged from investigation of paired specimens from eight AML patients: in the first, relapse was related to the acquisition of new mutations in the dominant clone identified in the primary leukemia sample; in the second, a minor subclone of the founding clone at the time of diagnosis gained additional mutations and expanded at relapse. For example, in one patient, the known recurrent AML mutations DNMT3A, NPM1, PTPRT, SMC3, and FLT3 were identified in the dominant clone at diagnosis. A minor sub-clone with a distinct cluster of mutations expanded at relapse due to the acquisition of additional mutations in ETV6, MYO18B, and the fusion gene WNK1-WAC. The pattern of relapse described in the first was identified in three patients, and the remaining five patients conformed to the second pattern. Despite chemotherapy, the founding clone was observed in all eight patients at both diagnosis and relapse. In addition, analysis of relapse versus diagnosis-specific mutations revealed a statistically significant increase in transversions. Compared with transitional mutations in which one purine or pyrimidine nucleotide is substituted for another, transversional mutations involve substitution of a purine nucleotide for a pyrimidine nucleotide or vice versa, and this latter type of mutation is associated with DNA damage mediated by cytotoxic therapy.