Abstract

Acute myeloid leukemia (AML) carrying nucleophosmin (NPM1) mutations and cytoplasmic NPM (NPMc+ AML) accounts for about one-third of all AML patients, and exhibits distinctive biological and clinical features. The role of NPM1 mutations in leukemogenesis remains elusive. Mathematical models have been developed that, starting from cancer incidence data, allow to infer the somatic mutation rate, or the number of genetic events required to cause cancer. We collected data on age at diagnosis of AML patients from four centers in three different countries, and calculated age-specific rates of NPMc+ AML. A total of 4,155 AML patients were investigated. NPM1 mutations these were detected in 1288. Patients carrying NPM1 mutations with age below 20 years and above 59 years were excluded from the study because of the low number of younger cases and because older patients are not always referred to major institutions for diagnosis and treatment. To investigate NPMc+ AML we adapted one-mutation model published by Michor et al (

PNAS
,
2006
;
103
:
14931
). The mathematical model consider a population of N (hemopoietic stem) cells that at beginning are wild-type. These cells proliferate according to the Moran process. The growth follows a logistic law with a saturation term. Our process follows the “classical” Moran process up to the appearance of a successful mutant. After that, the clone expands to a limiting population size. This is done to account for the dramatic expansion of the initial compartment peculiar of AML. Finally the rate of AML detection is proportional to the number of mutated cells. Experimental incidence curves of AML in Germany (Ge), Netherlands (Nl), and Italy (It) plotted simultaneously with predicted one-mutation model estimates are shown in Fig. 1. Linear regression of curves representing age-specific rate of diagnoses per year showed similar slopes (about 4 on a double-log scale) in different countries. The one-event model reproduces well the “exponential phenotype” of NPMc+ AML. In conclusion the model is in accordance with the hypothesis that NPM1 mutations by themselves are sufficient to cause NPMc+ AML. Alternatively, it is still possible that NPM1 mutations might cooperate with other molecular alterations to cause AML. In particular, since NPM1 mutations cause haploinsufficiency of wild-type NPM in leukemic cells and in knock-out mice NPM haploinsufficiency results in a MDS-like syndrome and given that the NPM1 mutant has oncogenic properties, these alterations could act in concert to cause AML. Indeed, the effect of these two alterations occurring simultaneously could be seen as a single genetic event.

Author notes

Disclosure: No relevant conflicts of interest to declare.