Low protein expression of Janus kinase 2 (JAK2) is prognostically adverse and independent of mutation status in acute myelogenous leukemia

Abstract INTRODUCTION: Mutations in Janus Kinase (JAK2) occur in <5% of Acute Myelogenous Leukemia (AML) cases and are associated with a worse prognosis. The most common genetic variant, JAK2V617F, leads to a gain of function (GOF) of the protein resulting in increased cell proliferation, thus supporting leukemogenesis. We hypothesized that levels of wild-type (WT) JAK2 protein could have prognostic impact similar to mutants (MUT). Methods: We measured the levels of 434 proteins using Reverse Phase Protein Arrays in 806 newly diagnosed, fresh, pre-treatment and >95% blast enriched AML samples. Protein expression was normalized to non-G-CSF treated, normal bone marrow (NBM)-derived CD34+ cells. JAK2 mutation status, treatment, and outcome data were known for 559 patients, of which 29 (~5%) had JAK2 mutation (MUT) and 530 were WT. LogRank tests were used to compared outcomes; Fisher's Exact, Pearson's Chi-squared or Wilcoxon tests for comparing variables; Pearson's correlation for protein correlation (p<0.01 and R>0.3); Wilcoxon tests adjusted by FDR for differential expression (p<0.05 and LFC>0.5); and Cox proportional hazards models (CoxPH) for Uni-(UV) and Multi-variate (MV) analysis. Results: The cohort was divided into quintiles and regrouped into upper 1/5th, named High (WT N=113, MUT N=3) and lower 4/5th, termed Low (WT N=417, MUT N=26). Interestingly, MUT JAK2 cases were more frequent than expected in the JAK2 Low group. Low JAK2 cases had less WBC and blood blasts, and a higher frequency of unfavorable cytogenetics, complex karyotype (CK), -5/5q- and mutations TP53, but lower inv16, FLT3-TKD, and DNMT3a, NPM1 and RAS MUT (P<0.001, <0.001, 0.005, 0.005, 0.007, <0.001, <0.001, 0.002, 0.02, <0.001, 0.03). JAK2 MUT had worse Overall Survival (OS) compared to WT, independent of protein levels, but WT-High performed significantly better than WT-Low (5ys OS: MUT-Low=3.8%, MUT-High=0%, WT-Low=20%, WT-High=37%; P=0.0002). Moreover, patients responded different to therapy depending on the levels of JAK2 WT: High expression patients performed well with AraC-based intensive chemotherapy (IC), but did poorly with Venetoclax plus Hypomethylating agent (VH), while Low expression patients had poor OS independent of therapy (5ys OS: IC WT-Low=30%, IC WT-High=57%, VH WT-Low=24%, VH WT-High=0%; P=0.001). A similar pattern was observed for Remission Duration (RD), with JAK2 High patients performing better with IC compared to VH and Low ones doing worse independent of therapy (5ys RD: IC WT-High=69%, IC WT-Low=53%, VH WT-High=10%, VH WT-Low=45%; P=0.006). For the CoxPH models of OS, we separated JAK2 WT patients by expression levels and therapy, resulting in 4 prognostic groups: IC High, IC Low, VH High and VH Low. In the UV CoxPH model of OS, all JAK2 WT treatment groups were prognostic, along with other clinical, cytogenetic and molecular features (e.g. age, CK, unfavorable cytogenetics, 2nd AML, mutations in CEBPA, FLT3-ITD, NPM1, IDH2, etc.). In the MV model, only IC treated JAK2 groups were significant, as well as age, 2nd AML, and CEBPA and IDH2 MUT. We also compared the protein signature of JAK2 WT-Low patients with JAK2-MUT to search for similarities and differences between them. The differential expression (DE) analysis, yielded only two proteins (IFITM3 and SOD2), which were downregulated in WT-Low. On the other hand, protein-protein correlations between JAK2 and the other 433 in our database revealed, 13 significantly correlated proteins in both WT-Low and MUT JAK2, which are related to histone H3K4 methylation, cell cycle and DNA Damage Response, SUMO E3 ligases and ErbB signaling, suggesting more commonalities than differences between JAK2 WT Low and JAK2 MUT. Conclusions: Patients with JAK2 WT-Low expression had a similarly bad prognosis compared to JAK2 MUT cases, independent of therapy. In contrast, JAK2 WT-High cases performed better, but only when treated with IC, and not VH, suggesting that correct therapy recommendation is crucial for these patients. Since JAK2 MUT directed-therapy are currently being investigated for AML, we propose that targeting signaling pathways that showed commonalities between JAK2 MUT and WT Low could be used in a much larger number of patients (~5% MUT vs 75% MUT+WT-Low), greatly increasing their utility. Finally, proteomics could be used to prospectively triage patients based on JAK2 levels to identify the WT cases that could benefit from JAK2-directed therapy.