The Leukemic Stem Cell in Polycythemia Vera and Primary Myelofibrosis Is Distinct From the Initiating JAK2 V617F-Positive Hematopoiertic Stem Cell

Abstract 613

The chronic myeloproliferative disorders (MPD), polycythemia vera (PV) and primary myelofibrosis (PMF) are clonal disorders involving a multipotent hematopoietic stem cell (HSC) but the identity of the involved HSC is a matter of controversy. For example, involvement of lymphoid cells in the MPD clone as determined by JAK2 V617F expression implies that these disorders arise in a pluripotent HSC. However, evidence has been presented that PV CD34+ HSC expressing JAK2 V617F are predisposed to erythroid differentiation, while mathematical modeling of JAK2 V617F-positive MPD suggests that these disorders arise in a hematopoietic progenitor cell that acquires JAK2 V617F, and then a mutation conferring self-renewal. Recently, a high level of aldehyde dehydrogenase (ALDHhigh) activity was used to distinguish normal CD34+CD38− HSC from committed hematopoietic progenitor cells, which lack ALDH activity, and also to separate leukemic stem cells (LSC) from CD34+/CD38−(ALDHhigh) HSC due to the presence of lower ALDH activity(ALDHint) in the LSC. We, therefore, sought to determine whether ALDH activity could be used to define the HSC involved in JAK2 V617F-positive PV and PMF, and whether leukemic transformation in these disorders was associated with a change in ALDH activity in the involved stem cell. To this end, we studied the immunophenotypic characteristics and ALDH activity of circulating CD34+ cells from 6 PV, 6 PMF (3 PMF and 3 post PV/MF) and 3 post MPD acute myelogenous leukemia (AML) patients (2 PV and 1PMF). CD34+ cells were isolated from peripheral blood using immunomagnetic bead technology with a purity of 95 % and a viability of 98 %, and analyzed for CD34 and CD38 expression and ALDH activity by flow cytometry. Cell sorting was carried out for measurement of the JAK2 V617F allele burden in discrete cell populations using purified genomic DNA and either a quantitative allele-specific assay or pyrosequencing. Where informative, sorted cell populations were analyzed by FISH for chromosomal abnormalities. As a per cent of total leukocytes, the median CD34+ cell fraction was 0.07 (range 0.01–0.2) for PV; 0.5 (range 0.4–1.3) for PV/MF and 2.3 (range 0.2–2.9) for PMF. The CD34+CD38− fraction was 17.8% of total PV CD34+ cells, 20.7% of total PV/MF CD34+ cells (p = 0.69) and 51.6 % of total PMF CD34+ cells (p= 0.003 and p= 0.007 respectively), indicating greater expansion of the PMF CD34+CD38− cell population, even though there was not a significant difference in disease duration between the three groups. ALDH activity was high in the CD34+CD38− cell population of all three groups. In 9/9 patients studied, the JAK2 V617F mutation was present in the CD34+CD38−(ALDHhigh) cell population at approximately 80 % of the patients' neutrophil JAK2 V617F allele burden, substantiating that the JAK2 V617F mutation was present in primitive PV and PMF HSC. We next analyzed the CD34+CD38− cell population for ALDH activity in 1 PMF and 2 PV patients who had transformed to AML. In addition to an CD34+/CD38−(ALDHhigh) cell population, all three patients had an additional CD34+/CD38− population with lower ALDH activity (ALDHint), identical to that of LSC. Importantly, this latter CD34+/CD38−(ALDHint) cell population was not present before AML transformation in any patient. In 2 of 2 patients studied, the CD34+/CD38−(ALDHint) cell population expressed JAK2 V617F with an allele burden comparable to the CD34+/CD38−(ALDHhigh) population. Significantly, in one of the patients, who had acquired a 5q- deletion, the chromosomal abnormality was present only in the CD34+CD38−(ALDHint) cell population. In conclusion, these data support the contention that in PV and PMF, JAK2 V617F is acquired in a primitive CD34+CD38−(ALDHhigh) HSC. In addition, in contrast to PV and PV/MF, expansion of total CD34+ cells in PMF was also accompanied by differential expansion of this primitive CD34+CD38−(ALDHhigh) population. Furthermore, AML transformation in PV and PMF appeared to occur in an LSC, which had an aberrant ALDH activity pattern (ALDHint) identical to that seen in de novo AML LSC. The 5q- deletion, a molecular marker characteristic of AML, also tracked with the LSC cell population but not with the CD34+CD38−(ALDHhigh) cell population, while JAK2 V617F was equivalently expressed by both cell populations. This observation supports the contention that with respect to AML transformation in PV and PMF, JAK2 V617F is essentially a passenger lesion.