JAK2 E846D Germline Mutation Associated with Erythrocytosis Causes Increased and Prolonged Epo-Induced Activation of STAT5

Until now, disease-causing JAK2 mutations have been associated with clonal myeloproliferative neoplasms (acquired mutations) or with polyclonal hereditary thrombocytosis (congenital mutations). Here we present the first example of a JAK2 mutation associated with congenital erythrocytosis.

We studied a young patient with congenital erythrocytosis who inherited two heterozygous mutations, an E846D mutation in JAK2 from the mother and a Q157H in PHD2 from the father. Each mutation apparently sufficed to induce EPO-hypersensitivity in BFU-E progenitors, since BFU-Es from both parents were hypersensitive to EPO, as were those of the patient. Strikingly, only the patient, but not his parents exhibited an erythrocytosis phenotype.

Expression analysis of granulocyte mRNA revealed markers of inappropriately increased signaling mediated by hypoxia sensing pathway (HIF) (Kapralova, Blood, 2014,16;123:391-4) in the propositus and his father. This result suggested that the polycythemia phenotype of the propositus could result from a combined impact of otherwise clinically silent defects co-inherited from both parents and affecting EPOR/JAK2 and HIF signaling pathways.

Using a Ba/F3-EPOR cellular model we have characterized the impact of JAK2 E846D on EPO-induced JAK2-mediated signaling, cell cycle progression and cell responsiveness to limited concentrations of EPO. We demonstrated prolonged, EPO-dependent JAK2 and STAT5 activation caused by JAK2 E846D, which was mitigated by a JAK2 inhibitor, AZ-960. However, in contrast to the oncogenic V617F mutation, E846D did not trigger factor-independent constitutive signaling. Also unlike the oncogenic V617F or the JAK2 germline mutations associated with hereditary thrombocytosis, the E846D did not activate STAT1, in agreement with differential signaling of STATs in different clinical phenotypes associated with activating JAK2 mutations. The thrombocytosis phenotype (ET for acquired and hereditary thrombocytosis for germline) is promoted by STAT1 signaling, while STAT5 signaling results in erythroid hyper-proliferation.

Using BrdU incorporation assay and growth factor starved Ba/F3 cells we also demonstrated that JAK2 E846D - when compared with JAK2 wild type - stimulates more rapid re-entry of the transfectants into the S phase in response to low concentration of EPO. In agreement with immunoblot analyses JAK2 E846D did not support growth factor-independent proliferation in cytokine-free media; however it provided improved survival in limiting concentrations of EPO. The proliferation advantage of JAK2 E846D transfectants was diminished by AZ-960. These results showed that E846D substitution lacks features of a constitutive gain-of-function mutation and were consistent with its germline transmission and with non-clonal hematopoiesis observed in the JAK2 E846D-positive mother.

Structural modeling based on the X-ray crystal structures of inactive and active kinase domain explain the mechanism of hypersensitivity and prolonged signaling. The homologous residue of E846 in the JAK2 pseudokinase domain (JH2) is N542, a residue that is targeted by the exon 12 activating mutations in polycythemia vera, and a close residue to this homologous position in JH2 is R564 recently found to be mutated in hereditary thrombocytosis. Thus the E846D mutation belongs to a conformational hotspot in kinases, but this mutation is only inducing hypersensitivity and prolonged action and not constitutive activation.

In conclusion, JAK2 E846D mutation associated with congenital erythrocytosis confers hypersensitivity to EPO, prolonged and selective but not constitutive STAT5 activation, and requires additional cooperating event, likely affecting the HIF signaling, for its full clinical expression.

(KK, MH, JK contributed equally to this work.)

Funding: Czech Science Foundation P301/12/1503, Czech Ministry of Health NT13587-4, Education for Competitiveness Operational Program CZ.1.07/2.3.00/20.0164 and Palacky University (IGA_LF_2014_011). JK was supported by CZ.1.07/2.3.00/30.0041, EL by a FRIA PhD fellowship. SNC was supported from the F.R.S.-F.N.R.S., Belgium, the Salus Sanguinis Foundation, the Action de Recherche Concertée projects MEXP31C1 and ARC10/15-027 of the University catholique de Louvain, Brussels, the Fondation contre le Cancer, Brussels, the PAI Programs BCHM61B5 and Belgian Medical Genetics Initiative.