Does Age of Patients Influence the Composition of Gene Mutations in Myeloid Neoplasms?

Introduction: Incidences of myeloid neoplasms, i.e. AML, MDS, MDS/MPN overlap, increase with age. Cytogenetic aberrations are still the hallmark for diagnosis and prognostication especially in AML and MDS. Different patterns of relations between chromosomal aberrations and age have been described. However, in recent years, gene mutations have been depicted to further discriminate patients with respect to their diagnosis and are increasingly used for prognostication. In parallel, recent studies (Jaiswal, NEJM 2014) have demonstrated that mutations in genes occurring in hematological neoplasms are also observed in healthy individuals and increase in frequency with age.

Aim: To investigate if specific molecular markers, such as DNMT3A, ASXL1, and TET2, increase in frequency with age in myeloid neoplasms as recently shown for healthy individuals.

Patients and methods: We investigated 1639 patients (pts) between 20 and 93 years (yrs), 578 with de novo AML (median age: 63 yrs), 846 with MDS (median age: 73 yrs), and 215 with CMML (median age: 75 yrs). In all cases, we followed the diagnostic criteria of the WHO classification based on morphology. All patients have also been investigated by cytogenetics and for disease-oriented molecular mutations (15-36 genes/pt: 15 in AML, 36 in MDS, and 21 in CMML). Analyses were performed by melting curve analysis, gene scan, Sanger sequencing, or next generation sequencing.

Results: In total we detected 3089 mutations (range: 0-9/pt) spread over all except for seven analyzed genes. Grouping the entity-specific cohorts by age of the patients into decades revealed a steady increase of the prevalence of mutations with age in MDS (at least one mut/pt, 25% in 20-29 to 93% in >80 yrs; p<0.001), less prominent in AML (77% in 20-29 to 100% in >80 yrs, p=0.007), but not in CMML (96%-100% in all decades). However, the number of mutations per patient increased according to age in all three entities, significantly in MDS (p<0.001) and AML (p<0.001). Considering AML patients separated into three cytogenetic classes (Grimwade, Blood 2010) resulted in the same findings for the intermediate risk (p=0.012) and adverse risk group (p<0.001), while the good risk group showed no change in mutation numbers over decades (median: 1 mut/pt, range 0-3). This indicates that in good risk AML (PML-RARA, CBFB-MYH11, RUNX1-RUNX1T1) only very few additional mutations are needed for AML initiation. In contrast, an age-dependent increasing incidence of gene mutations is specific in normal karyotype and in adverse cytogenetics.

We next focused on specific gene mutations according to age <60 vs ≥60 yrs within all three entities. In addition, AML patients where again subgrouped by cytogenetics. In AML good and adverse risk groups no age-dependent significant increase of specific gene mutations occurred, while in the intermediate risk group mutations in ASXL1 (3/160 vs 32/114, p<0.001), MLL-PTD (3/160 vs 17/214, p=0.01), RUNX1 (32/160 vs 20/214, p=0.001), and TET2 (4/159 vs 26/214, p=0.001) were significantly more frequent at higher age. In contrast, NRAS mutations appeared more often in younger AML patients (32/160 vs 20/214, p=0.004). In MDS, mutations in SF3B1 (27/115 vs 253/731, p=0.019), SRSF2 (10/115 vs 133/730, p=0.011), TET2 (10/115 vs 250/731, p<0.001), and TP53 (2/115 vs 50/731, p=0.035) were more frequently observed in older patients. In CMML only TET2 mutations occurred more often in older patients (5/12 vs 135/190, p=0.026).

Focusing on the genes recently described to be mutated in healthy individuals showed that all of the above mentioned mutations found in myeloid neoplasms (except MLL-PTD and RUNX1) are comprised in the 10 most frequently mutated genes in the healthy aging population. However, the fact that the frequencies of these mutations are not age-dependent in some entities, e.g. ASXL1 only age-dependent in AML but not in CMML and MDS, might indicate different roles of these mutations in the pathogenesis, i.e. driver mutations independent of age, as well as their contribution to accumulation of mutations and onset of a myeloid neoplasm.

Conclusions: 1) The number of mutations significantly increase with age in AML and MDS and non-significantly in CMML. 2) Several genes show age-dependent frequencies, which differ between AML, MDS, and CMML and are also related to the cytogenetic background. 3) Based on molecular mutations healthy aging and myeloid neoplasms are neighbouring scenarios.