JAK2^V617F mutation in leukaemic transformation of philadelphia-negative chronic myeloproliferative disorders

The natural history of Philadelphia-negative chronic myeloproliferative disorders (Ph-CMPD) is characterised by a chronic phase that, in a fraction of cases, may progress to acute leukaemia (Jaffe et al, 2001). The risk of leukaemic transformation is time-, age- and treatment-dependent, and is approximately 1–5% for essential thrombocythaemia (ET), 5–10% for polycythaemia vera (PV) and 10–30% for idiopathic myelofibrosis (IMF) (Jaffe et al, 2001).

The JAK2^V617F mutation occurs in 70–90% of PV and 30–50% ET and IMF (Baxter et al, 2005; Kralovics et al, 2005) in chronic phase. The aim of this study was to verify the prevalence of the JAK2^V617F mutation in acute myeloid leukaemia (AML) evolved from Ph-CMPD.

Samples were derived from five PV, ten ET and five IMF patients at the time of transformation to AML. In four cases (cases 12A/12B, 13A/13B, 14A/14B, 15A/15B), the corresponding chronic phase was also available. Bone marrow (BM) samples were collected at diagnosis of chronic and/or leukaemic phase. Chronic phases of PV, ET and IMF and AML were diagnosed according to the World Health Organization criteria (Jaffe et al, 2001). In all cases, the presence of BCR/ABL rearrangement was ruled out. For each patient, BM samples were simultaneously investigated for the JAK2^V617F mutation by cDNA sequencing and mutation-specific polymerase chain reaction (Baxter et al, 2005; Kralovics et al, 2005). All patients provided informed consent.

The JAK2^V617F mutation was identified in 5/20 (25·0%) AML secondary to Ph-CMPD (Table I). In all mutated cases, the normal JAK2 sequence could be detected in addition to the mutated allele. JAK2^V617F mutation occurred in 3/5 (60·0%) cases of AML transformed from PV, 2/5 (40·0%) AML transformed from IMF and 0/11 AML transformed from ET (Table I). In four AML transformed from ET (cases 12B, 13B, 14B, 15B), the corresponding chronic phase sample was also available (cases 12A, 13A, 14A, 15A). In three of these patients (case 12A/12B, 14A/14B, 15A/15B), both the ET and the AML phase were negative for the JAK2^V617F mutation. One patient carried the JAK2^V617F mutation in the ET phase (13A), but not in the AML phase (13B) (Table I). This same patient was karyotyped and investigated for p53 mutations in both ET and AML phases. During the ET phase, the karyotype was 45,XY,−4,−20 and p53 mutations were absent. In the AML phase, the karyotype was 47,XY,+mar and p53 analysis revealed a Ser→Thr mutation at codon 303.

Several observations suggest that the JAK2^V617F mutation may not play a substantial role in leukaemic transformation of PV, IMF and ET. First, the prevalence of the JAK2^V617F mutation in AML transformation from PV and IMF was similar to that observed in chronic phase PV and IMF (Baxter et al, 2005; Kralovics et al, 2005). Second, AML cases with JAK2^V617F mutation carried a normal JAK2 allele in addition to the mutated allele, indicating that the JAK2^V617F mutation burden does not increase during leukaemogenesis. Of note, similar observations have been recently reported in a series of AML post-IMF (Mesa et al, 2006). Third, in our series, all cases of AML post-ET were devoid of the JAK2^V617F mutation, a prevalence that was markedly different from that expected on the basis of JAK2^V617F mutation analysis in chronic phase ET (Baxter et al, 2005; Kralovics et al, 2005).

Two alternative hypotheses may be envisaged to justify the absence of the JAK2^V617Fmutation in AML evolving from ET. One hypothesis is that only cases of ET that are devoid of JAK2^V617Fmutation are prone to AML transformation. However, clinical studies do not identify JAK2 mutation status as a risk factor for AML transformation (Wolanskyj et al, 2005). On the other hand, one may speculate that a JAK2^V617F-negative subclone emerges during AML transformation of JAK2^V617F-positive ET. This hypothesis is supported by paired sample analysis of case 13A/13B, which carried the JAK2^V617F mutation at ET diagnosis but not at the time of AML transformation. The possibility that, in case 13A/B, AML derived from a subclone that was genetically different from the predominant chronic phase population was also supported by the appearance during the AML phase of a less complex karyotype, characterised by the disappearance of chromosome 4 and 20 deletions that were present in the corresponding ET chronic phase.

Although limited to a single patient, blastic transformation of a JAK2^V617F-negative subclone in the context of JAK2^V617F-positive ET is in accordance with the emerging concept that, at least in a fraction of Ph-CMPD, the JAK2^V617F mutation may represent a ‘second hit’ lesion (Kralovics et al, 2006).