Chimeric JAK2 Kinases Trigger Non-uniform Changes of Cellular Metabolism in BCR-ABL1-like Childhood ALL

lmost a century ago, Otto Warburg described increased glucose uptake and induced glycolysis in malignant cells, a metabolic alteration, which is now considered a general feature of tumor cells. 1 Aerobic glycolysis (so-called Warburg effect) was regarded as a privileged energetic pathway despite its lower energy production in comparison to oxidative phosphorylation (OXPHOS), the bioenergetic process generally used in normal cells. However, recent studies suggest that malignancies are metabolically very heterogeneous, with some relying on glycolysis while other preferentially on OXPHOS. 2,3 Malignant transformation is a process characterized by the acquisition of multiple genetic aberrations that enable uncontrolled proliferation of cancer cells and their invasion of other tissues. Some variants affecting cellular signaling may be essential in altering cellular metabolism in favor of increased energetic demands of malignant cells. 4–6 The investigation of different metabolic changes driven by specific genetic backgrounds of individual cancers can contribute to our better understanding of the process of tumorigenesis and identify novel cancer vulnerabilities possibly exploitable in therapy. In this study, we focused on acute lymphoblastic leukemia (ALL) and studied the impact of specific genetic aberrations on the metabolism of leukemic cells. ALL represents the most common pediatric cancer with an overall good but strikingly heterogeneous prognosis. One of the subgroups with less favorable

A lmost a century ago, Otto Warburg described increased glucose uptake and induced glycolysis in malignant cells, a metabolic alteration, which is now considered a general feature of tumor cells. 1 Aerobic glycolysis (so-called Warburg effect) was regarded as a privileged energetic pathway despite its lower energy production in comparison to oxidative phosphorylation (OXPHOS), the bioenergetic process generally used in normal cells. However, recent studies suggest that malignancies are metabolically very heterogeneous, with some relying on glycolysis while other preferentially on OXPHOS. 2,3 Malignant transformation is a process characterized by the acquisition of multiple genetic aberrations that enable uncontrolled proliferation of cancer cells and their invasion of other tissues. Some variants affecting cellular signaling may be essential in altering cellular metabolism in favor of increased energetic demands of malignant cells. [4][5][6] The investigation of different metabolic changes driven by specific genetic backgrounds of individual cancers can contribute to our better understanding of the process of tumorigenesis and identify novel cancer vulnerabilities possibly exploitable in therapy.
In this study, we focused on acute lymphoblastic leukemia (ALL) and studied the impact of specific genetic aberrations on the metabolism of leukemic cells. ALL represents the most common pediatric cancer with an overall good but strikingly heterogeneous prognosis. One of the subgroups with less favorable prognosis is defined as BCR-ABL1-like (Ph-like) ALL and accounts for up to 10%-15% of all pediatric ALL cases and for up to 20%-25% of ALLs in adults. 7,8 Unlike the BCR-ABL1positive ALLs, the BCR-ABL1-like ALLs do not contain the BCR-ABL1 fusion, yet their gene expression profile is similar to that of BCR-ABL1-positive ALL. The majority of BCR-ABL1-like cases harbor primary genetic lesions affecting kinase signaling, one of the examples being JAK2 fusions causing constitutive activation of the Janus kinase 2 (JAK2). The most common fusion partner of JAK2 in BCR-ABL1-like ALL is PAX5; however, the repertoire of fusion partners is still growing as less frequent fusions continue to be discovered. 8 In line with this fact, we recently identified a novel JAK2 fusion, NPAT-JAK2, in a patient with BCR-ABL1like ALL. 9 Here we describe this fusion, study its impact on JAK/ STAT kinase signaling and its oncogenic potential, and compare it with that of PAX5-JAK2. Importantly, we describe the impact of both fusions on cellular metabolism.
Molecular characterization of the NPAT-JAK2 fusion gene showed that, similarly to other described JAK2 fusions, NPAT-JAK2 involves an intact JAK2 kinase domain (for details of NPAT-JAK2 sequence and structure see Suppl. Figure S1). In transiently transfected HEK293T cells, NPAT-JAK2 was translated into an in-silico-predicted chimeric protein (molecular weight 74 kDa), which was localized both in the cytoplasm and the nucleus (predominantly in the cytoplasm) and was phosphorylated on tyrosines corresponding to Y1007/Y1008 of wild-type JAK2 ( Figure 1A). Introduction of an inactivating mutation (NPAT-JAK2 K882E ) to the ATP-binding site within the JAK2 moiety (corresponding to K882 of wild-type JAK2) prevented NPAT-JAK2 phosphorylation, which strongly suggested that the chimeric protein was autophosphorylated upon constitutive activation of the JAK2 kinase domain. Thus, we confirmed that the JAK2 moiety within NPAT-JAK2 preserved its ATP-binding and catalytic function ( Figure 1A).
We also studied the oncogenic potential of NPAT-JAK2 and its impact on JAK/STAT pathway signaling. Using a lentiviral vector, NPAT-JAK2 was introduced into IL-3-dependent Ba/F3 cells, where it exhibited prosurvival and proproliferative capacities upon IL-3-withdrawal ( Figure 1B). This effect was not reproduced in the NPAT-JAK2 K882E mutant cells, indicating its dependence on the NPAT-JAK2 kinase activity. Transformation of Ba/F3 cells by NPAT-JAK2 was significantly faster compared with PAX5-JAK2 ( Figure 1B). Interestingly, the transformed PAX5-JAK2 cells under IL-3 starvation (Figure 2Dii) showed a slower growth rate 1 CLIP -Childhood Leukaemia Investigation Prague, Czech Republic than cells cultured in the presence of IL-3 while such a difference was not observed in the NPAT-JAK2 model (Figure 2Di). 10 These results suggest a stronger oncogenic potential of NPAT-JAK2.
Western blot analysis of the NPAT-JAK2-transformed Ba/ F3 cells showed increased phosphorylation of STAT1, STAT3, and STAT5 ( Figure 1C; Suppl. Figure S2), which was inhibited by the JAK1/2 inhibitor Ruxolitinib (Figure 1Cii; Suppl. Figure  S3). Importantly, Ruxolitinib blocked the proproliferative effect of NPAT-JAK2-transformed Ba/F3 cells ( Figure 1D). Ectopic expression of PAX5-JAK2 resulted in constitutive phosphorylation of STAT1 and STAT5 and with significantly less phosphorylation of STAT3 comparing to cells with NPAT-JAK2 expression (Figure 1Ci; Suppl. Figure S3). Analogously, all phosphorylated STATs were decreased after the treatment with JAK2-inhibitors (Figure 1Cii; Suppl. Figure S3), which also inhibited the proliferation of PAX5-JAK2-transformed cells. 10  To determine the metabolic and energetic changes induced by NPAT-JAK2 and PAX5-JAK2, we employed the extracellular flux analysis, which allows real time quantification of extracellular acidification rate (ECAR) representing glycolytic function and oxygen consumption rate (OCR) representing cellular respiration (OXPHOS) in live cells. Using the Cell Energy Phenotype Test, we showed the NPAT-JAK2-transformed Ba/F3 cells without IL-3 in comparison to controls with IL-3 supplementation (Ba/ F3 cells transduced with empty vector or NPAT-JAK2) increase both glycolysis and mitochondrial respiration and acquire a more "energetic" phenotype ( Figure 2Ai). This metabolic phenotype shift of NPAT-JAK2-transformed Ba/F3 cells was enhanced after stress induction with combination of oligomycin and carbonyl cyanide-p-trifluoromethoxyphenylhydrazone (FCCP) (inhibitor of ATP-synthase and mitochondrial uncoupler, respectively), which enable to determine the cells' maximal bioenergetic capacity. These findings were accompanied by significantly increased ATP production together with increased NAD+/NADH ratio as assessed by a luminescence assay in the NPAT-JAK2-transformed cells when compared with controls ( Figure 2C). Intriguingly, PAX5-JAK2-transformed cells displayed an opposite metabolic shift toward a more quiescent phenotype with decrease of both OCR and ECAR (Figure 2Aii). Importantly, we employed stable isotope tracing with C 13 -glucose. We confirmed higher activity of glycolysis in NPAT-JAK2-tranduced cells in comparison to PAX5-JAK2 transduced cells. It was represented by increased labeled pyruvate and lactate. Further, we showed that NPAT-JAK2-transformed cells increased labeled citrate and malate, intermediates of Krebs cycle ( Figure 2B). Nevertheless, their ATP production and NAD+/NADH ratio was similar to that of NPAT-JAK2-transformed cells and significantly higher compared with controls ( Figure 2C). Furthermore, we used 2-deoxy-D-glucose (2DG), a competitor of glucose (which is one of the main substrates used in bioenergetic pathways of cancer cells) to test the dependence of the NPAT-JAK2 and the PAX5-JAK2 transforming potential on bioenergetic rewiring. The proliferation of NPAT-JAK2-transformed cells treated with 2DG was dramatically inhibited, compared with control cells. In contrast, the proliferation of PAX5-JAK2-transformed cells was affected only moderately ( Figure 2D). To support our findings, we introduced NPAT-JAK2 and PAX5-JAK2 fusion genes into primary murine c-kit-positive bone marrow cells (c-kit BMCs) and determined their metabolic phenotype. We observed that NPAT-JAK2 transduced c-kit BMCs had more energetic metabolic phenotype in comparison to more quiescent phenotype detected in PAX5-JAK2 transduced c-kit BMCs; the result was concordant with the data of transformed Ba/F3 cells (Figure 2Aiii).
Metabolic rewiring, a cancer hallmark, is one of the key mechanisms that underlie tumorigenesis, tumor progression, and chemoresistance. 11 Although hematological malignancies mainly use aerobic glycolysis, some favor OXPHOS to satisfy their energy demands. 2,12 Here, we show that lymphoid cells transformed by NPAT-JAK2, a novel fusion, which we found in BCR-ABL1-like ALL, increase both their glycolytic activity and mitochondrial respiration. These results are in concordance with the previously published study showing elevated glycolysis and increased OXPHOS in an in vivo model of myeloproliferative neoplasms driven by JAK2-activating mutation. 13 Surprisingly, we observed that PAX5-JAK2-transformed cells change their metabolism in an opposite manner, that is, they lower their glycolytic and respiratory functions. This quiescent metabolic phenotype may potentially mirror the relatively weaker transforming capacity of PAX5-JAK2 resulting in a lower proliferation rate compared with NPAT-JAK2. Suppressed STAT3 signaling in PAX5-JAK2-transformed cells could also contribute to lower glycolytic function. 14 Hypothetically, the metabolic phenotype may be significantly influenced by the biological impact of the PAX5 moiety. It has been shown that the PAX5-JAK2 protein is localized exclusively in the nucleus and binds to wild-type PAX5 target loci and may activate PAX5 target genes' expression, although to a lesser extent compared with wild-type PAX5. 10 Wild-type PAX5 represses glucose and energy metabolism 15 ; a partial preservation of this metabolic impact of PAX5 within the PAX5-JAK2 fusion could explain the low glycolytic activity of PAX5-JAK2-transformed cells.
Interestingly, despite the differences in glycolytic and OXPHOS activity, both JAK2 fusions showed increased levels of ATP and NAD + /NADH. The elevated ATP production could represent a more general feature of the overall metabolic impact of activating JAK2 lesions, as it was previously also observed in JAK2-mutated myeloproliferative neoplasms. The increased ratio of NAD + /NADH in NPAT-JAK2-transformed cells is likely a result of enhanced mitochondrial respiration, while its nature in the PAX5-JAK2-transformed cells is unclear.
In conclusion, we studied the impact of JAK2 fusions on metabolic rewiring of leukemic cells; we chose two representatives, the most frequent fusion and a novel previously undescribed fusion, and we showed that JAK2 fusions indeed alter energy metabolism of leukemic cells, but surprisingly in a very distinct manner. This discordance between both fusions may result from differences in their overall biological impacts, which are indicated by the variations in transformation potential and subcellular localization and are likely caused by the JAK2 partner gene character. The BCR-ABL1-like ALL includes leukemias with functionally distinct classes of genetic aberrations (such as ABL versus JAK/ STAT class) resulting in a nonnegligible biological but also clinical heterogeneity of this ALL subtype. Yet, our data show that even fusions involving the same kinase gene may exert considerable biological variety, and thus, although we believe that metabolism-targeting drugs may be therapeutically relevant in this unfavorable ALL subtype and deserve further studies, the use of such drugs may require precise tailoring with respect to the specific metabolic impact of individual genetic aberrations.