Background and purpose: Chronic myeloproliferative neoplasms (MPN) may transform into secondary acute myeloid leukemia (sAML). The genetic evolution pattern is important in understanding the pathogenesis of sAML transformation. We aimed to determine the profile of genetic evolution by examining a cohort of paired samples collected at both MPN and sAML phases using a panel of commonly mutated genes involved in myeloid neoplasms.

Methods: Mutational analyses of 14 genes (JAK2V617F, CALR, ASXL1, IDH1, IDH2, TET2, DNMT3A, EZH2, KRAS, RUNX1, TP53, WT1, CBL, and SRSF2) were performed on the paired matched samples at diagnosis and at sAML transformation from 22 patients with MPN (6 polycythemia vera, 5 essential thrombocythemia, and 11 primary myelofibrosis). The samples were analyzed by polymerase chain reaction (PCR)-based assays followed by direct sequencing. For CALRmutational analysis, the PCR products of exon 9 were analyzed by GeneScan for screening deletions and insertions, then validated by Sanger sequencing. The allele frequencies of all mutated genes were determined by pyrosequencing assays in samples harboring any detectable mutations.

Results: At the MPN phase, JAK2V617F was the most common mutation (N=12), followed by CALR (N=9). At sAML phase, JAK2V617F was not detected in 4 patients who had JAK2V617F at initial diagnosis while all 9 patients retained CALR mutations. Mutations of epigenetic modifier genes were detected in 13 patients (59%) at MPN phase, including TET2 (N=5), EZH2 (N=4), ASXL1 (N=2), DNMT3A (N=2), IDH1 and IDH2 (N=1 for each). One MPN sample was also found carrying RUNX1 mutation. Acquisition of 12 mutations was found in 11 patients at the sAML phase, including RUNX1 (N=4), KRAS (N=3), IDH2 (N=2), and one each of CBL, ASXL1 and TP53. Except JAK2V617F and TET2 mutations, the mutations present in the MPN phase remained detectable at sAML transformation. The allele frequencies of IDH2 and EZH2 mutant clones were apparently increased at sAML phase. Of note, two patients had double TET2 mutations at the MPN phase, clonal selection of TET2 mutants was explicitly demonstrated in both cases, with disappearance of one clone and expansion of the other during sAML transformation. The median time from MPN to sAML was 71.6 months and the median overall survival was 73.2 months for the cohort of MPN patients who had sAML transformation later on. Female patients had a trend towards lower incidence of CALR mutation (2/11 vs. 7/11, P=0.08) and significantly shorter overall survival (43.3 vs. 129.2 months, P<0.0001) and leukemia-free survival (36.4 vs. 116.9 months, P<0.0001). The number of mutations other than CALR was higher among females (mean 2.4 vs.1.3, P=0.013). The overall survival and time to sAML were not affected by age at diagnosis of MPN, degree of bone marrow fibrosis, leukocyte or platelet counts. Patients with CALR mutation had a trend towards longer survival (120.5 vs. 67.8 months, P=0.051) and a longer time to sAML (114.2 vs. 67.7 months, P=0.051). No other mutated genes examined or the number of mutations harbored by MPN patients had significant impact on survival or time to sAML.

Conclusions: Our results showed that in addition to JAK2V617F and CALR mutations, mutations of epigenetic modifiers were present in more than half of MPN patients at diagnosis who later on had a sAML transformation. During sAML transformation, non-JAK2V617F clone evolved in one third of patients and acquisition of RUNX1, KRAS and IDH2 mutations recurrently emerged. Clonal selection and expansion of TET2 mutations might play a vital role in leukemogenesis of MPN.

Grant support: NSC96-2314-B-182-003, CMRPG330303, OMRPG3C0021 and DOH102-TD-C-111-006

Disclosures

No relevant conflicts of interest to declare.

Author notes

*

Asterisk with author names denotes non-ASH members.

Sign in via your Institution