JAK2 impairs stem cell function?

Several mouse models have been generated to mimic the myeloproliferative neoplasm (MPN) phenotypes observed in patients. Retroviral transduction of bone marrow cells with JAK2-V617F cDNA resulted in a transplantable MPN phenotype with polycythemia vera (PV) and variable degrees of myelofibrosis.^2,,,–6  Transgenic mouse models expressing the JAK2-V617F displayed a phenotype resembling essential thrombocythemia (ET) or PV, depending in part on the relative expression levels of mutant and wild-type Jak2.^7,–9  Three groups have now generated knock-in mice by targeting the mouse Jak2 locus to introduce the V617F mutation.^1,10,11  These 3 models differ in the targeting strategy and, interestingly, also in the observed phenotypes (Table 1). Two of them are inducible using the Cre-loxP system,^1,10  but only 1 allows breeding the mice to homozygosity by providing a wild-type Jak2 transcript in the noninduced configuration.¹⁰  In 2 studies, the mouse Jak2-V617F was used and in both cases a strong PV phenotype with high hematocrit, thrombocytosis (3- to 5-fold normal) and neutrophilia was observed.^10,11 

In contrast, Li and colleagues inserted a human JAK2-V617F cDNA into the start codon of the mouse Jak2 gene.¹  The peripheral blood phenotype was the mildest of all the models made so far, with a modest increase in platelets (1.3-fold normal) and hematocrit with no signs of myelofibrosis. Over 26 weeks, approximately 10% of the mice developed a more pronounced increase in hematocrit resembling PV with a concomitant decrease in platelets; in 1 mouse myelofibrosis was noted. The reasons for the variations in phenotype within the cohort of JAK2-V617F knock-in mice are unclear. Thus, the genotype-phenotype correlations remains puzzling even after having 3 knock-in mouse strains at hand. There could be differences between the human and mouse Jak2-V617F in respect to how they signal in mouse hematopoietic cells, which may explain the mild ET phenotype of the human Jak2-V617F in the model by Li and colleagues.¹  However, at similar expression levels, the human JAK2-V617F transgene caused a more pronounced phenotype ranging from ET to PV.⁷  Surprisingly, the 2 knock-in models that used the mouse Jak2 had similar phenotypes, despite Jak2-V617F expression levels of only 50% of wild-type Jak2 in the report by Akada.¹⁰  A 1:1 ratio between mutant and wild-type Jak2 was found by Marty and colleagues.¹¹  A side-by-side comparison of the different mouse strains will be necessary to resolve open issues.

Li and colleagues have gone a step further in the analysis of their mouse model and examined stem cell function and accumulation of DNA damage. Unexpectedly, they found reduced numbers of Lineage⁻Sca-1⁺c-Kit⁺ (LSK) cells, which contain the early progenitors and the hematopoietic stem cells. No difference was present at 6 weeks, but a 50% reduction became apparent at 26 weeks.¹  At this stage, the hematopoietic progenitors were also decreased. Bone marrow transplantation into sublethally irradiated hosts resulted in the same mild phenotype as in nontransplanted mice, but competitive bone marrow transplantation showed a reduced capacity of the JAK2-V617F bone marrow cells to outgrow wild-type competitor cells. This difference was even more pronounced when secondary transplantations were performed, suggesting that the JAK2-V617F stem cells show signs of senescence and exhaustion. Several questions arise: why does the mouse model with the mildest phenotype show such a severe reduction of self-renewal potential in the stem cell compartment? The Jak2-V617F mouse model by Akada et al actually showed a 4-fold increase in LSK numbers.¹  Because the genetic background of the knock-in mice by Li and colleagues was not uniform, it remains possible that weak performance in the competitive repopulation of sublethally irradiated hosts could be due to immunologic rejection.

Since JAK2-V617F has been reported to induce DNA damage, Li et al measured γ-H2AX levels, a histone variant that is used as an indicator for the presence of double-strand DNA breaks. The JAK2-V617F has been shown to increase homologous recombination and genetic instability.¹²  Indeed, the LSK cells from JAK2-V617F mice showed increased γ-H2AX levels at 26 weeks and also reduced numbers of apoptotic cells, as measured by annexin-V/7-aminoactinomycin D. These results are very interesting and merit further investigations. Whether mutations in additional genes found to be mutated in MPN, such as TET2, ASXL1, CBL, IDH and IKZF1, may provide additional clonal advantage and how they influence MPN stem cell function remains to be determined. The mouse models will be important tools to dissect these functional relationships.