In 1969, William Dameshek proposed that paroxysmal nocturnal haemoglobinuria (PNH) should be considered a candidate myeloproliferative neoplasm (MPN) based on the evidence of red blood cell abnormalities in the context of bone marrow pancytopenia, low leucocyte alkaline phosphatase and the potential for leukaemic transformation (Dameshek, 1969). Further similarities between MPN and PNH include expansion of subsequent or simultaneous myeloid clones and an increased frequency of abdominal vein thrombosis. PNH is an acquired, stem cell disorder associated with mutations of the PIGA gene that encodes a protein required for the synthesis of glycosylphosphatidyl-inositol anchored proteins (GPI-AP). PNH blood cells are therefore deficient in cell surface proteins that use the GPI anchor and may be detected immunophenotypically by absent or reduced expression of CD55 and CD59 on red blood cells (RBC) and CD16, CD24, CD66b and FLAER (fluorescence labelled aerolysin) on granulocytes. PNH clones are also present in a significant number of patients with the related marrow failure syndromes aplastic anaemia (AA) and hypoplastic myelodysplastic syndrome (MDS). Although PNH originates from a multipotent haematopoietic stem cell, PIGA mutations are insufficient to drive clonal expansion and several models have attempted to explain this phenomenon: GPI-AP may be immunogenic and their deficiency enables evasion from immunological surveillance or the possibility that PIGA mutations themselves confer an intrinsic resistance to apoptosis (Brodsky, 2008). Another theory postulates the acquisition of a second genetic event that enables clonal expansion in co-operation with the PIGA mutation, some evidence for which comes from the deregulated expression of HMGA2 in a minority of PNH patients (Inoue et al, 2006).
A recent report identified the JAK2 V617F mutation in GPI-AP deficient cells of three PNH patients and suggested this coexistence could explain clonal expansion (Sugimori et al, 2012). These findings prompted investigation into the frequency of the JAK2 V617F in other patients harbouring a PNH clone. Archival DNA samples from blood (n = 3) and bone marrow aspirate smears (n = 19) from 22 patients (PNH, n = 5; AA, n = 13; hypoplastic MDS, n = 4) with an immunophenotypically detectable PNH clone of both RBC and granulocytes were retrospectively analysed for JAK2 V617F by qualitative allele-specific polymerase chain reaction (PCR) (Baxter et al, 2005) with a sensitivity of 2% mutant alleles. The JAK2 V617F mutation was detected in one patient with AA. This female patient, who presented with severe AA and no evidence of a PNH clone (nor retrospectively the JAK2 V617F), was initially successfully treated with two courses of antithymocyte globulin (ATG) and cyclosporin. She relapsed in pregnancy 2 years later and was treated with immunoglobulin, and RBC and platelet transfusions. Post partum treatment included G-CSF, ATG (2 courses) and ciclosporin, the latter being discontinued after achieving RBC and platelet transfusion independence. Ten years after the initial presentation, progressive pancytopenia suggested relapse of AA with a PNH clone detected within the granulocytes (17·7%) and RBCs (5·4%). At this relapse, the JAK2 V617F was detected in the bone marrow and confirmed by quantitative PCR (qPCR; Larsen et al, 2007) at a reproducible level of 1·8% of total JAK2 alleles (Fig 1). Ciclosporin was recommenced with a good response attained. 3 years post-relapse, the PNH clone remained stable in granulocytes (18·9%) and RBCs (6·0%) with the JAK2 V617F not detected in the bone marrow by both allele-specific and qPCR. At no point during the clinical course did the patient have a persistent leucocytosis, erythrocytosis or thrombocytosis. The patient was subsequently diagnosed with malignant melanoma to which she succumbed. Diagnostic or pre-allogeneic haematopoietic stem cell transplantation samples from a further 38 AA patients without evidence of PNH were screened for the JAK2 V617F, none of which were found to harbour this mutation.
We describe the hitherto unreported detection of the JAK2 V617F in a PNH-related bone marrow failure syndrome at a similarly low frequency to that previously described in classical PNH (Fouassier et al, 2009; Sugimori et al, 2012). As this case was identified retrospectively, existence of the JAK2 V617F in the PNH clone could not be verified by fluorescence-activated cell-sorting of the GPI-AP-negative population (Sugimori et al, 2012). However, the transient nature of the JAK2 mutation in the presence of a relatively stable PNH clone, taken together with the evidence of different JAK2 V617F and PNH clone sizes implies two distinct populations as opposed to the co-existence of mutations within one clone. The phenomenon of transient JAK2 V617F-positivity may be explained by the acquisition of the mutation in a more differentiated haematopoietic progenitor than that in patients with MPN. If this progenitor had reduced self-renewal potential and capacity for differentiation then the clone generated might therefore be short-lived and not result in a clinical phenotype (Passamonti et al, 2007).
Characterization of the role of JAK2 signalling in PNH and PNH-related disorders may provide some insight into disease development, maintenance or progression. The detection of the JAK2 V617F and other molecular lesions that contribute to clonal expansion, such as deregulation of HMGA2 expression, evident in PNH (Inoue et al, 2006) and both typical MPN and atypical myeloid malignancies (Odero et al, 2005; Bruchova et al, 2008) suggests some common pathogenetic mechanisms of PNH and MPN that warrant further investigation.