Myeloproliferative neoplasms (MPNs) are clonal disorders of hematopoietic stem cells (HSCs) that are most frequently caused by acquired somatic mutations in JAK2. A number of conditional mouse models of JAK2-V617F-driven MPN have been generated that rely on Cre-LoxP-mediated activation, resulting in polyclonal disease. To more closely mimic the monoclonal origin of human MPN, transplantations of single purified JAK2-mutant HSCs or bone marrow (BM) at limiting dilutions into lethally irradiated recipient mice have been previously performed. However, irradiation is known to alter the BM microenvironment and also to induce transient aplasia accompanied by elevated cytokine levels that promote the expansion of the mutant clone. To overcome these limitations, we examined whether JAK2-V617F-mutant HSCs are able to engraft and initiate MPN in non-conditioned recipients. We found that BM from two different MPN models, one expressing the human JAK2-V617F, and another expressing the mouse Jak2-V617F, efficiently engrafted and initiated MPN in non-irradiated immunocompromised Rag2 -/- recipients. MPN evolved even in transplantations at limiting dilutions, showing the high competitiveness of single JAK2-mutant HSCs. In contrast, BM from mice expressing the human JAK2-V617F failed to engraft in non-conditioned immunocompetent C57BL/6 mice, while BM from mice expressing the mouse Jak2-V617F engrafted and initiated MPN, suggesting that mouse JAK2-V617F protein, which differs from the endogenous JAK2 in only one amino acid, was tolerated. Our results show that JAK2-V617F mutant HSCs can outcompete resident non-mutated HSCs even in the absence of elevated cytokine levels and without the need of emptying stem cell niches by irradiation.

Myeloproliferative neoplasms (MPN) are clonal disorders of hematopoietic stem cells (HSC) that are most frequently caused by acquired somatic mutations in JAK2 . A number of conditional mouse models of JAK2- V617F-driven MPN have been generated that rely on Cre-LoxP mediated activation, resulting in polyclonal disease. To more closely mimic the monoclonal origin of human MPN, transplantations of single purified JAK2 -mutant HSCs or bone marrow (BM) at limiting dilutions into lethally irradiated recipient mice have been previously performed. However, irradiation is known to alter the BM microenvironment and also to induce transient aplasia accompanied by elevated cytokine levels that promotes the expansion of the mutant clone. To overcome these limitations, we examined whether JAK2- V617F-mutant HSCs are able to engraft and initiate MPN in non-conditioned recipients. We found that BM from two different MPN models, one expressing the human JAK2 -V617F, and another expressing the mouse Jak2 -V617F, efficiently engrafted and initiated MPN in non-irradiated immunocompromised Rag2−/− recipients. MPN evolved even in transplantations at limiting dilutions, showing high competitiveness of single JAK2 -mutant HSCs. Thus, JAK2 -V617F mutant HSCs can outcompete resident non-mutated HSCs in the absence of elevated cytokine levels and without the need of emptying stem cell niches by irradiation. However, only BM from mice expressing the mouse Jak2 -V617F engrafted and initiated disease in non-conditioned C57BL/6 mice, while BM from mice expressing the human JAK2 -V617F was rejected, indicating that mouse Jak2 -V617F is ignored by the immune surveillance. These results provide a possible explanation why JAK2 -V617F is so frequently found in healthy individuals with clonal hematopoiesis. ### Competing Interest Statement R.C.S. is a scientific advisor/SAB member and has equity in Ajax Therapeutics, he consulted for and/or received honoraria from Novartis, BMS/Celgene, AOP, GSK, Baxalta and Pfizer. N.H. owns stocks in the company Cantargia. The remaining authors declare no competing financial interests. Swiss National Science Foundation, 31003A_166613, 310030_185297/1, 310030_185297/2 Swiss Cancer Research foundation, KFS-3655-02-2015, KFS-4462-02-2018 Stiftung für Hämatologische Forschung Cancer Prevention and Research Institute of Texas, RR240024 Research Fund of the University of Basel for Junior Researchers Jubiläumsstiftung von Swiss Life Jacques and Gloria Gossweiler Foundation, https://ror.org/046keqp87

Therapy with pegylated interferon alpha (pegIFNα) can induce a deep molecular response in a subset of patients with myeloproliferative neoplasms (MPN). Here we investigated the role of Socs2, a negative regulator of cytokine signaling, in modulating the response to pegIFNα in a JAK2-V617F mouse model of MPN. Deleting Socs2 in JAK2-V617F mice resulted in increased sensitivity to cytokines, without causing significant alterations in the MPN phenotype. When subjected to pegIFNα, the loss of Socs2 enhanced the depletion of JAK2-mutant hematopoietic stem cells (HSCs), evidenced by reduced chimerism in peripheral blood and bone marrow compared to vehicle controls. Additionally, pegIFNα-treated Socs2-deficient JAK2-mutant HSCs exhibited functional impairments in secondary transplantations, reflecting long-term detrimental decline of their stemness. These findings demonstrate that loss of Socs2 enhances the effectiveness of pegIFNα in depleting the JAK2-mutant HSC clone. In line with the genetic ablation of Socs2, the SOCS2 inhibitor MN714 combined with IFNα exhibited better efficacy than IFNα alone in reducing the output of CD34+ cells from PV patients in vitro. Targeting SOCS2 could therefore improve therapeutic responsiveness in MPN patients receiving interferon therapy.

•In a mouse model of oligo-clonal MPN, IL-1β favors disease initiation by promoting early expansion of a sub-clinical JAK2-V617F clone•Anti-IL-1β antibody treatment during the early expansion phase of the JAK2-mutant clone reduced the frequency of MPN disease initiation

Pegylated interferon alpha (pegIFNα) can induce molecular remissions in JAK2-V617F-positive myeloproliferative neoplasms (MPN) patients by targeting long-term hematopoietic stem cells (LT-HSCs). Additional somatic mutations in genes regulating LT-HSC self-renewal, such as DNMT3A, have been reported to have poorer responses to pegIFNα. We investigated if DNMT3A loss leads to alterations in JAK2-V617F LT-HSCs functions conferring resistance to pegIFNα treatment in a mouse model of MPN and in hematopoietic progenitors from MPN patients. Long-term treatment with pegIFNα normalized blood parameters, reduced splenomegaly and JAK2-V617F-chimerism in single-mutant JAK2-V617F (VF) mice. However, pegIFNα in VF;Dnmt3aΔ/Δ (VF;DmΔ/Δ) mice worsened splenomegaly and failed to reduce JAK2-V617F-chimerism. Furthermore, LT-HSCs from VF;DmΔ/Δ mice compared to VF were less prone to accumulate DNA damage and exit dormancy upon pegIFNα treatment. RNA-sequencing showed that IFNα induced stronger upregulation of inflammatory pathways in LT-HSCs from VF;DmΔ/Δ compared to VF mice, indicating that the resistance of VF;DmΔ/Δ LT-HSC was not due to failure in IFNα signaling. Transplantations of bone marrow from pegIFNα treated VF;DmΔ/Δ mice gave rise to more aggressive disease in secondary and tertiary recipients. Liquid cultures of hematopoietic progenitors from MPN patients with JAK2-V617F and DNMT3A mutation showed increased percentages of JAK2-V617F-positive colonies upon IFNα exposure, whereas in patients with JAK2-V617F alone the percentages of JAK2-V617F-positive colonies decreased or remained unchanged. PegIFNα combined with 5-azacytidine only partially overcame resistance in VF;DmΔ/Δ mice. However, this combination strongly decreased the JAK2-mutant allele burden in mice carrying VF mutation only, showing potential to inflict substantial damage preferentially to the JAK2-mutant clone.

Introduction. Myeloproliferative neoplasms (MPN) are clonal diseases that originate from a single hematopoietic stem cell (HSC). The JAK2-V617F mutation is found in ~70% of all MPN patients. To date, it remains unclear why some patients with this JAK2-V617F mutation develop essential thrombocythemia (ET), with hyperproliferation of the megakaryocyte lineage, while others develop polycythemia vera (PV), characterized by an overproduction of both platelets and erythrocytes. In this project, we explored the hypothesis that acquiring the JAK2-V617F mutation in a subset of HSCs that are biased towards the megakaryocyte lineage will lead to ET, whereas acquiring the mutation in unbiased HSCs will lead to PV. Methods. To identify megakaryocyte-biased HSCs, we used a mouse reporter line that expresses the green fluorescent protein (GFP) under the promoter of the von Willebrand Factor (vWF). This reporter allows for the identification of a subset of HSCs with high GFP expression (vWF-GFPhi HSCs) which, in wildtype (WT) mice, showed bias towards megakaryocyte and platelet production in vivo (Sanjuan-Pla et al., Nature 2013). We crossed this vWF-GFP reporter with our SclCreER;JAK2-V617F MPN mice (VF), which display marked erythrocytosis and thrombocytosis after activation of the transgene by tamoxifen (Kubovcakova et al., Blood 2013). To compare the in vivo functional potential of FACS-sorted vWF-GFPlo and vWF-GFPhi HSCs isolated from WT or VF donor mice, we performed competitive transplantations of 100 HSCs (defined as Lin-c-Kit+Sca-1+CD48-CD150+Epcr+) mixed with 1x106WT bone marrow competitor cells into lethally irradiated WT recipients. Results.Recipient mice transplanted with vWF-GFPhi HSCs from WT donors exhibited higher donor chimerism in platelets than in other peripheral blood lineages, while recipients of vWF-GFPlo HSCs showed no megakaryocyte bias, as previously reported (Sanjuan-Pla et al., Nature 2013). Mice transplanted with vWF-GFPhi HSCs from VF donors developed PV phenotype but surprisingly lost their megakaryocyte bias, whereas recipients of vWF-GFPlo HSCs had low overall chimerism without MPN phenotype and no platelet bias. Thus, in the presence of the JAK2-V617F mutation, only the vWF-GFPhi HSCs were capable of initiating MPN, while losing their preference for the megakaryocyte lineage. In both VF and WT mice, we also found that 40% of vWF-GFPhi HSCs showed high surface expression of the integrin CD41 (CD41hi), which was reported to be associated with megakaryocyte-biased HSCs (Gekas and Graf, Blood 2013), while 60% were CD41lo. Mice transplanted with vWF-GFPhi /CD41lo HSCs showed better engraftment compared to recipients of vWF-GFPhi /CD41hi HSCs, but neither fraction showed bias towards megakaryocyte and platelet production. Conclusion.Only the vWF-GFPhi subset of HSCs from VF mice was capable of initiating MPN phenotype in competitive stem cell transplantations into irradiated WT recipient mice, irrespective of the CD41 marker expression. Furthermore, expression of JAK2-V617F abolished megakaryocyte and platelet bias of the vWF-GFPhi subset of HSCs that was observed in WT mice. Additional investigations are underway to dissect the contribution of vWF-GFPlo versus vWF-GFPhi HSCs in our VavCre;JAK2-V617F mice that develop an ET-like phenotype with high platelets but normal red cell parameters.

JAK2 -V617F is the most frequent somatic mutation causing myeloproliferative neoplasm (MPN). JAK2 -V617F can be found in healthy individuals with clonal hematopoiesis of indeterminate potential (CHIP) with a frequency much higher than the prevalence of MPNs. The factors controlling the conversion of JAK2 -V617F CHIP to MPN are largely unknown. We hypothesized that interleukin-13 (IL-13) -mediated in flammation can favor this progression. We established an experimental system using bone marrow (BM) transplantations from JAK2- V617F and GFP transgenic ( VF;GFP ) mice that were further crossed with IL-1 beta -/ - or IL-1R1 - / - mice. To study the role of IL -13 and its receptor on monoclonal evolution of MPN, we performed competitive BM transplantations at high dilutions with only 1 to 3 hematopoietic stem cells (HSCs) per recipient. Loss of IL-1 beta in JAK2 -mutant HSCs reduced engraftment, restricted clonal expansion, lowered the total numbers of functional HSCs, and decreased the rate of conversion to MPN. Loss of IL -1R1 in the recipients also lowered the conversion to MPN but did not reduce the frequency of engraftment of JAK2 -mutant HSCs. Wild -type ( WT ) recipients transplanted with VF;GFP BM that developed MPNs had elevated IL -13 levels and reduced frequencies of mesenchymal stromal cells (MSCs). Interestingly, frequencies of MSCs were also reduced in recipients that did not develop MPNs, had only marginally elevated IL -13 levels, and displayed low GFP-chimerism resembling CHIP. Anti -IL-13 antibody preserved high frequencies of MSCs in VF;GFP recipients and reduced the rate of engraftment and the conversion to MPN. Our results identify IL-1 beta as a potential therapeutic target for preventing the transition from JAK2 -V617F CHIP to MPNs.

Hyperproliferation of myeloid and erythroid cells in myeloproliferative neoplasms (MPN) driven by the JAK2-V617F mutation is associated with altered metabolism. Given the central role of glutamine in anabolic and catabolic pathways, we examined the effects of pharmacologically inhibiting glutaminolysis, that is, the conversion of glutamine (Gln) to glutamate (Glu), using CB-839, a small molecular inhibitor of the enzyme glutaminase (GLS). We show that CB-839 strongly reduced the mitochondrial respiration rate of bone marrow cells from JAK2-V617F mutant (VF) mice, demonstrating a marked dependence of these cells on Gln-derived ATP production. Consistently, in vivo treatment with CB-839 normalized blood glucose levels, reduced splenomegaly and decreased erythrocytosis in VF mice. These effects were more pronounced when CB-839 was combined with the JAK1/2 inhibitor ruxolitinib or the glycolysis inhibitor 3PO, indicating possible synergies when cotargeting different metabolic and oncogenic pathways. Furthermore, we show that the inhibition of glutaminolysis with CB-839 preferentially lowered the proportion of JAK2-mutant hematopoietic stem cells (HSCs). The total number of HSCs was decreased by CB-839, primarily by reducing HSCs in the G1 phase of the cell cycle. CB-839 in combination with ruxolitinib also strongly reduced myelofibrosis at later stages of MPN. In line with the effects shown in mice, proliferation of CD34+ hematopoietic stem and progenitor cells from polycythemia vera patients was inhibited by CB-839 at nanomolar concentrations. These data suggest that inhibiting GLS alone or in combination with inhibitors of glycolysis or JAK2 inhibitors represents an attractive new therapeutic approach to MPN.

JAK2-V617F is the most frequent driver gene mutation in patients with myeloproliferative neoplasms (MPN) but the phenotypic manifestation is heterogeneous with some patients presenting as polycythemia vera (PV), whereas others with essential thrombocythemia (ET) or myelofibrosis (MF). JAK2-V617F is acquired in a hematopoietic stem cell (HSC) and leads to clonal expansion of HSCs and progenitor cells that gain dominance over unmutated hematopoiesis. We tested the hypothesis that the preferential expansion of megakaryopoiesis in ET versus erythropoiesis in PV could be due to the acquisition of JAK2-V617F in different subsets of HSCs with inherent bias towards megakaryopoiesis or erythropoiesis. In addition, we addressed the question of whether sensitivity or resistance to interferon-α (IFNα), currently the only treatment that can induce a deep molecular remission in some MPN patients, could be also due to heterogeneity in HSC subpopulations. We used a mouse model of JAK2-V617F driven MPN in combination with molecular barcoding that allowed us to monitor expansion and lineage contribution of individual HSC subclones. First, we genetically barcoded HSCs from our tamoxifen-inducible SclCre; JAK2-V617F ( VF) mice with a lentiviral vector, and transplanted 2'000 HSCs into lethally irradiated recipient mice. After 12 weeks we sequenced the barcodes in bone marrow (BM) progenitors by next generation sequencing (NGS). We observed a strong selection of 1-2 HSCs clones per recipient mouse, which alone contributed to > 80% of all the myeloid cells. Since cytokine storm after irradiation and aplasia during the reconstitution with lentivirally transduced BM may favor such oligo-clonal dominance observed in this experiment, we characterized the clonal composition of JAK2-mutant HSCs in non-transplanted MPN mice. To this end, we crossed a CRISPR-based genetic barcoding mouse line ( CARLIN) (Bowling, S. et al, Cell 2020) with our VF mice to obtain VF;CARLIN mice. Cas9-mediated barcoding in VF;CARLIN or in CARLIN mice with wildtype Jak2 (WT;CARLIN) was induced with doxycycline and immediately followed by the activation of the JAK2-V617F by tamoxifen (Figure 1A). After 12 weeks, when the mice developed full PV phenotype (hemoglobin 195 g/L; platelets 4.3x10 12 /L and neutrophils 10.5x10 9/L), we performed scRNAseq on lin- cKit+ BM cells using the 10X-platform to retrieve the cellular barcodes and transcriptomic profiles. MPN hematopoiesis in these non-transplanted VF;CARLIN mice was polyclonal (Figure 1B). A total of > 500 barcoded clones per mouse were detected and in both VF;CARLIN and WT;CARLIN mice, but ~5 individual barcodes accounted for 25% of the cells in all progenitor subsets analyzed (Figure 1B). Overall, clonal composition was similar between WT;CARLIN and VF;CARLIN mice, indicating that clonal selection was not more prominent in JAK2-mutant mice. Importantly, in the offspring of HSCs that accounted for 25% of the cells, we did not observe preferential presence of some barcodes in MkP versus EryP and other barcodes in EryP versus MkP, arguing against the model predicting that erythrocytosis and thrombocytosis originated in different subsets of biased HSCs. To investigate the question of whether sensitivity or resistance to IFNα could be also due to heterogeneity in HSC subpopulations, we transplanted BM cells from these barcoded VF;CARLIN mice together with an excess of WT competitor cells into irradiated WT recipient mice, and treated them with IFNα or vehicle for 16 weeks. IFNα normalized hemoglobin (122 g/L) compared to vehicle (177 g/L) and reduced VF;CARLIN chimerism in granulocytes from 85% to 44%. BM cells were harvested and the clonal composition is currently being analyzed by NGS in order to identify IFNα resistant and sensitive subclones. Our barcoding analyses of individual HSCs showed that, during recovery after BM transplantations, a few VF clones dominated, whereas induction of VF expression in situ without transplantation resulted in a polyclonal disease that showed a similar number of HSCs actively contributing to hematopoiesis as in a control experiment with WT mice. The lack of substantial preference in the clonal contribution of individual HSCs to erythropoiesis versus megakaryopoiesis in VF mice suggests that PV and ET are not caused by acquiring JAK2-V617F in HSCs with a pre-existing lineage bias.

Myeloproliferative neoplasms (MPNs) are caused by a somatic gain-of-function mutation in 1 of the 3 disease driver genes JAK2, MPL, or CALR. About half of the MPNs patients also carry additional somatic mutations that modify the clinical course. The order of acquisition of these gene mutations has been proposed to influence the phenotype and evolution of the disease. We studied 50 JAK2-V617F-positive MPN patients who carried at least 1 additional somatic mutation and determined the clonal architecture of their hematopoiesis by sequencing DNA from single-cell-derived colonies. In 22 of these patients, the same blood samples were also studied for comparison by Tapestri single-cell DNA sequencing (scDNAseq). The clonal architectures derived by the 2 methods showed good overall concordance. scDNAseq showed higher sensitivity for mutations with low variant allele fraction, but had more difficulties distinguishing between heterozygous and homozygous mutations. By unsupervised analysis of clonal architecture data from all 50 MPN patients, we defined 4 distinct clusters. Cluster 4, characterized by more complex subclonal structure correlated with reduced overall survival, independent of the MPN subtype, presence of high molecular risk mutations, or the age at diagnosis. Cluster 1 was characterized by additional mutations residing in clones separated from the JAK2-V617F clone. The correlation with overall survival improved when mutation in such separated clones were not counted. Our results show that scDNAseq can reliably decipher the clonal architecture and can be used to refine the molecular prognostic stratification that until now was primarily based on the clinical and laboratory parameters.

JAK2-V617F mutation causes myeloproliferative neoplasms (MPN) that can manifest as polycythemia vera (PV), essential thrombocythemia (ET) or primary myelofibrosis (PMF). PV patients at diagnosis already exhibited iron deficiency, whereas ET patients had normal iron stores. We examined the influence of iron availability on MPN phenotype in mice expressing JAK2-V617F and in mice expressing JAK2 with a N542-E543del mutation in exon 12 (E12). At baseline on control diet, all JAK2-mutant mouse models with PV-like phenotype displayed iron deficiency, although E12 mice maintained more iron for augmented erythropoiesis than JAK2-V617F mutant mice. In contrast, JAK2-V617F mutant mice with an ET-like phenotype had normal iron stores comparable to wildtype (WT) mice. On low-iron diet, JAK2-mutant mice and WT controls increased platelet production at the expense of erythrocytes. Mice with PV phenotype responded to parenteral iron injections by decreasing platelet counts and further increasing hemoglobin and hematocrit, whereas no changes were observed in WT controls. Alterations of iron availability primarily affected the pre-megakaryocyte erythrocyte progenitors (pre-MegE), which constitute the iron-responsive stage of hematopoiesis in JAK2-mutant mice. The orally administered ferroportin inhibitor vamifeport and the minihepcidin PR73 normalized hematocrit and hemoglobin levels in JAK2-V617F and E12 mutant mouse models of PV, suggesting that ferroportin inhibitors and minihepcidins could be used in the treatment of PV patients.

Background: Myeloproliferative neoplasms (MPNs) are clonal disorders of the hematopoietic stem cell (HSC), caused by somatic mutations in JAK2, MPL or CALR . Myelofibrosis, characterized by increased deposition of reticulin and/or collagen fibers, is found in advanced stages of MPN. Pentraxin-2 (PTX2, serum amyloid P component/SAP) belongs to the family of short pentraxins and acts as an inhibitor of fibrocyte differentiation and modulator of macrophage polarization. Zinpentraxin alfa (PRM-151, ZPN), a recombinant form of human PTX2, was reported to reduce myelofibrosis in a retroviral model of MPN driven by MPL-W515L (Verstovsek S et al, J Exp Med 2016). ZPN has also been investigated as monotherapy and in combination with ruxolitinib (RUX) in a phase 2 clinical study in patients with myelofibrosis (NCT01981850; Verstovsek S et al, Haematologica 2023). Evidence of clinical activity and tolerable safety as monotherapy and in combination with RUX was shown in that open-label, non-randomized trial. Here, we examined the effects of ZPN alone or in combination with RUX in a mouse model of MPN driven by Cre-inducible expression of human JAK2-V617F (Tiedt R et al, Blood 2008). Methods: To obtain sufficient numbers of mice for drug testing, bone marrow cells from JAK2-V617F mice were transplanted into lethally irradiated C57BL/6 recipients (Kubovcakova L et al., Blood 2013), and groups of 6 mice were sacrificed at 16, 20, and 24 weeks to determine the histological grade of reticulin fibrosis. 24 weeks post transplantation, grade 1-2 fibrosis was confirmed in these satellite mice and treatment was initiated. During treatment, weight and complete blood counts were monitored (n=8 mice per group). At terminal work-up, spleen weight and fibrosis grade were determined, along with flow cytometry of bone marrow and peripheral blood, single cell RNA sequencing of bulk bone marrow, and proteomics analysis by mass spectrometry of bone marrow and plasma. Additionally, we characterized the plasma pharmacokinetics (PK) in these mice following single IP (10 mg/kg) administration of ZPN. Results: Systemic exposure after IP administration was confirmed in each animal (T max at 4 h, C max 27.9 ug/mL,AUC 0-48 381 h*ug/mL). No weight loss or mortality were observed in ZPN monotherapy cohorts, and the slight weight loss observed in RUX treated animals was not potentiated by the addition of ZPN. A trend towards lower platelet, monocyte and total leukocyte counts was observed for ZPN treatment groups compared to vehicle controls (Figure 1A). RUX alone largely normalized hemoglobin values, while ZPN alone or in combination with RUX had less effect on hemoglobin. Terminal work-up showed grade 1-2 fibrosis in the vehicle group, whereas reticulin fibrosis decreased in the majority of mice with ZPN monotherapy or in combination with RUX (Figure 1B). ZPN monotherapy did not reduce spleen weight. Single cell RNA sequencing and proteomics analyses are currently being analyzed and data will be shown. Conclusion: ZPN treatment in a JAK2-V617F mouse model of MPN with myelofibrosis was well tolerated as monotherapy and in combination with RUX. A reduction in the grade of myelofibrosis was observed in all ZPN treatment groups. ZPN showed promising trends in reducing platelet and monocyte counts, while the decrease in hemoglobin by RUX was in part prevented in combination with ZPN.

Patients with myeloproliferative neoplasms (MPN) frequently carry a somatic mutation in the JAK2 gene (JAK2-V617F), which causes an overactivation of the JAK-STAT signaling pathway that leads to erythrocytosis, thrombocytosis and/or myelofibrosis. JAK2-V617F is acquired in a hematopoietic stem cell (HSC). Therefore, in order to develop novel effective therapies, it is imperative to specifically target the mutant HSCs, as these cells are the reservoir that maintains MPN. To date, pegylated-interferon-alpha (pegIFNα) is the only treatment known to induce molecular remission in some patients with MPN. The mechanism behind this effect is not completely understood, but recent studies suggest that pegIFNα selectively pushes mutant HSCs into cell cycle, which leads to their differentiation and exhaustion. We hypothesized that disrupting components of the negative feedback loop to JAK-STAT signaling in HSCs would lead to JAK-STAT hyperactivation and thereby sensitize HSCs to the effects of pegIFNα. To investigate which genes are governing the feedback loop, we first performed single-cell RNA-sequencing on HSCs from our JAK2-V617F MPN mouse model. We found that Socs2, a JAK-STAT negative regulator gene, was one of the top upregulated genes in MPN HSCs, while expression of other genes known to be involved in JAK-STAT negative feedback loop was unaltered by the JAK2 mutation. We examined the effects of loss of Socs2 by crossing our JAK2-V617F mice (VF) with a constitutional Socs2 knockout strain (Metcalf et al. Nature, 2000), to obtain VF;Socs2-/- double mutant mice. We performed competitive bone marrow transplantations by mixing bone marrow cells from VF mice that also express a GFP reporter gene (VF;GFP) at 1:1 ratio with bone marrow cells from VF;Socs2-/-mice, and added a 10-fold excess of bone marrow cells from wildtype CD45.1 mice. This 1:1:20 mixture was then transplanted into lethally irradiated wildtype CD45.1 recipient mice. These mice were allowed to engraft for 4 weeks and were then treated for 16 weeks with pegIFNα (25 ug/kg s.c. once weekly) or vehicle (Figure 1A). CD45.2 was used to track total VF plus VF;Socs2-/- chimerism and GFP distinguished between VF;GFP and VF;Socs2-/- cells. The ratio between VF;GFP and VF;Socs2-/- chimerism remained stable in the vehicle group during the course of the experiment (Fig.1B). However, the VF;Socs2-/-chimerism was reduced in the pegIFNα-treated mice, both in peripheral blood granulocytes (7.6% compared to 41% in the vehicle group) and in the bone marrow. Specifically, HSC chimerism of double mutant cells dropped from 43% in the vehicle arm to 15% in the treated mice (Fig.1B). We also performed transplantation of unfractionated BM cells from VF;Socs2-/- or VF mice in competition with BM from GFP-expressing wildtype mice at a ratio of 1:10. These transplanted mice were randomized and treated for 16 weeks with either pegIFNα or vehicle (n=12 per arm). Although both single and double mutant groups had similar hematologic responses, the VF;Socs2-/-arm had a better molecular response than VF (data not shown). 41% of the mice from the double mutant group had molecular response in peripheral blood and in HSCs, compared to only 16% in the VF group. Our data show that releasing the brake applied by Socs2 on Jak2 signaling magnifies the pegIFNα-induced effects on VF;Socs2-/- HSCs, leading to a better molecular response. Thus, hyperactivating JAK2-STAT signaling in HSCs by inhibiting Socs2 could potentially improve the molecular response caused by pegIFNα. Figure 1View largeDownload PPTFigure 1View largeDownload PPT Close modal

Type I JAK inhibition (JAKi) represents a mainstay therapy for myelofibrosis and hydroxyurea-resistant high-risk polycythemia vera. Current type I JAKi, including ruxolitinib (RUX) and fedratinib, improve symptoms and outcomes in MPNs; however, JAK2VF allele burden remains essentially unchanged, and efficacy wanes over time. Sustained JAK/STAT signaling plays a critical role in MPN cell persistence in the setting of type I inhibition. CHZ868, a type II JAKi that binds the inactive conformation of the kinase domain, overcomes RUX persistence in vitro and reduces Jak2VF allele fraction in vivo suggesting improved JAK2 targeting might enhance clinical efficacy. However, current type II JAKi, including CHZ868, are limited by lack of kinome specificity and off-target toxicity. We therefore sought to develop novel type II JAK2i with improved potency and selectivity. Computational free energy perturbation and structure-activity relationship-based methods were used to identify lead type II inhibitor compounds. This work, followed by Absorption, Distribution, Metabolism and Excretion modeling, led to the development of AJ1-10502. Kinome selectivity profiling of AJ1-10502 revealed potent JAK2 selectivity with minimal kinase cross-reactivity compared to CHZ868, including among other JAK family tyrosine kinases (Figure 1A). Assessment of in vitro efficacy of AJ1-10502 in SET2 inhibitor naïve and SET2 RUX persistence (RUXper) cells revealed concentration-dependent inhibition of proliferation of RUXper cells comparable to that of CHZ868. The IC50 for AJ1-10502 in RUXper cells was similar to the IC50 observed in inhibitor naïve cells. We next evaluated AJ1-10502 in vivo using a novel dual Dre/Cre-recombinase Jak2VF knock-in/knock-out model allowing for the comparison of JAKi to Jak2VF genetic deletion. AJ1-10502 demonstrated reductions in leukocytosis comparable to that of RUX (K/uL: VEH 18.7 vs. RUX 10.1 vs. AJ1-10502 9.3) but dose-dependent improvements in hematocrit, platelet levels, and spleen weights superior to RUX, with reductions in spleen weights on par to that of Jak2VF deletion (mg: VEH 480 vs. RUX 255 vs. AJ1-10502 145 vs. del 107, p<0.05). A greater degree of restoration of splenic architecture was also observed with AJ1-10502 compared to type I JAKi. Critically, we observed reductions in peripheral blood (PB) and bone marrow (BM) mutant allele fraction with AJ1-10502 not observed with RUX, including within Mac1+Gr1+ myeloid cell fractions (VEH 94% vs. RUX 92% vs. AJ1-10502 78.5%, p<0.05), consistent with a mutant-biased reduction in myeloid output. In a separate series of competitive transplants using a Cre-inducible human JAK2VF transgenic mouse line, we validated the phenotypic changes in regard to leukocytosis, hematocrit, and spleen weight reduction seen with the DreCre model. Most importantly, we confirmed significant reductions in mutant cell fraction within the hematopoietic stem (HSC) compartment of BM and spleen, including among granulocytic-monocytic progenitor (GMP) and long-term HSC (LT-HSC) populations (Figure 1B). Finally, a separate study comparing AJ1-10502 to CHZ868 revealed no significant weight loss with AJ1-10502 (-0.5g) compared to CHZ868 (-2.2g, p<0.05) despite similar reductions in Jak2VF allele fraction suggesting comparable efficacy without systemic toxicity. In sum, AJ1-10502 is a potent, selective type II JAK2i with improved efficacy compared to RUX and an enhanced safety profile compared to previous non-selective type II JAKi. Most importantly, AJ1-10502 results in superior reductions in PB and BM mutant cell fraction in vivo not observed with type I JAKi. These data demonstrate the preclinical utility of type II JAKi with AJ1-10502 and inform a path to clinical development of type II JAKi for MPN patients. Figure 1View largeDownload PPTFigure 1View largeDownload PPT Close modal

Interleukin-1β (IL-1β) is a master regulator of inflammation. Increased activity of IL-1β has been implicated in various pathological conditions including myeloproliferative neoplasms (MPNs). Here we show that IL-1β serum levels and expression of IL-1 receptors on hematopoietic progenitors and stem cells correlate with JAK2 -V617F mutant allele fraction in peripheral blood of patients with MPN. We show that the source of IL-1β overproduction in a mouse model of MPN are JAK2 -V617F expressing hematopoietic cells. Knockout of IL-1β in hematopoietic cells of JAK2 -V617F mice reduces inflammatory cytokines, prevents damage to nestin-positive niche cells and reduces megakaryopoiesis, resulting in decrease of myelofibrosis and osteosclerosis. Inhibition of IL-1β in JAK2 -V617F mutant mice by anti-IL-1β antibody also reduces myelofibrosis and osteosclerosis and shows additive effects with ruxolitinib. These results suggest that inhibition of IL-1β with anti-IL-1β antibody alone or in combination with ruxolitinib could have beneficial effects on the clinical course in patients with myelofibrosis.

Pegylated interferon alpha (IF) can induce molecular remissions in subset of JAK2-V617F-positive MPN patients by targeting long-term hematopoietic stem cells (LT-HSCs). Patients with additional somatic mutations in genes involved in LT-HSC self-renewal have been reported to have poorer responses to IF. We found that loss of Dnmt3a increases competitiveness and self-renewal of JAK2-V617F-positive LT-HSCs and confers resistance to IF treatment in MPN mouse models and in HSCs and progenitor cells from MPN patients. Here we examined whether the resistance of double mutant JAK2-V617F;Dnmt3aΔ/Δ hematopoietic cells to IF can be overcome by addition of arsenic trioxide (At), or 5-azacytidine (Az). To test the effects of the combination treatments in vivo, we generated cohorts of mice expressing JAK2-V617F (VF) alone, or in combination with a homozygous deletion of the Dnmt3a gene (Dnmt3aΔ/Δ). We used bone marrow cells from VF or VF;Dnmt3aΔ/Δ mice that express a GFP reporter gene mixed with a 10-fold excess of bone marrow cells from a wildtype (WT) mouse to perform competitive transplantations into lethally irradiated WT recipient mice. After 7 weeks, recipient mice were randomized into 6 treatment groups including a vehicle control group. These mice were treated for 12 weeks with IF (25µg/kg; s.c. once per week), At (5mg/kg; i.p. every second day), or Az (2mg/kg; i.p. daily for two weeks followed by a break of 2 weeks), or combinations of IF+At and IF+Az (Figure 1). Peripheral blood parameters were normalized by At and Az treatment arms in both single-mutant and double-mutant mice (Figure 1A). Spleen and liver weight was decreased in all treatment groups and genotypes, except in VF;Dnmt3aΔ/Δ mice treated with IF alone, which showed a trend towards increased spleen and liver weights. Expression of GFP allowed us to follow the contribution of cells derived from the VF or VF;Dnmt3aΔ/Δ donor mice, respectively. In VF recipient mice, a pronounced decrease in GFP chimerism of peripheral blood lineages was observed in the groups treated with a combination of IF+At, or IF+Az (Figure 1A). LT-HSC from bone marrow and spleen showed reduction in GFP chimerism upon treatment with Az alone, compared to vehicle and further reduction below 10% was observed in combination of Az and IF (Figure 1B). Double-mutant VF;Dnmt3aΔ/Δ recipient mice remained resistant to the single agent regiments, but showed a significant reduction of GFP chimerism in peripheral blood and in LT-HSCs when treated with a combination of IF+Az and to a lesser degree also when treated with IF+At (Figure 1B). Thus, a combination of pegIFNa with 5-azacytidine is a promising approach to target MPN cells carrying mutations in JAK2 and Dnmt3a genes that could be also considered as a treatment option for therapy-resistant forms of MPN in patients carrying JAK2-V617F and loss-of-function DNMT3A mutations. Figure 1View largeDownload PPTFigure 1View largeDownload PPT Close modal

Metabolic reprogramming is one of the hallmarks of cancer, as these rapidly dividing cells need to adapt their metabolism to cope with an increased energy demand. We have previously showed that in mouse models of myeloproliferative neoplasms (MPN), JAK2-mutant cells display metabolic alterations, including increased oxidative phosphorylation and glycolysis. The glutamine-glutamate-alphaKetoglutarate (aKG) axis, besides fueling the Krebs cycle and anabolic processes, contributes to the synthesis of the heme precursor 5-aminolevulinic acid (5-ALA), thus making glutaminolysis a potential target for MPN therapy.

We studied a subset of hematopoietic stem cells (HSCs) that are defined by elevated expression of CD41 (CD41hi) and showed bias for differentiation toward megakaryocytes (Mks). Mouse models of myeloproliferative neoplasms (MPNs) expressing JAK2-V617F (VF) displayed increased frequencies and percentages of the CD41hi vs CD41lo HSCs compared with wild-type controls. An increase in CD41hi HSCs that correlated with JAK2-V617F mutant allele burden was also found in bone marrow from patients with MPN. CD41hi HSCs produced a higher number of Mk-colonies of HSCs in single-cell cultures in vitro, but showed reduced long-term reconstitution potential compared with CD41lo HSCs in competitive transplantations in vivo. RNA expression profiling showed an upregulated cell cycle, Myc, and oxidative phosphorylation gene signatures in CD41hi HSCs, whereas CD41lo HSCs showed higher gene expression of interferon and the JAK/STAT and TNFα/NFκB signaling pathways. Higher cell cycle activity and elevated levels of reactive oxygen species were confirmed in CD41hi HSCs by flow cytometry. Expression of Epcr, a marker for quiescent HSCs inversely correlated with expression of CD41 in mice, but did not show such reciprocal expression pattern in patients with MPN. Treatment with interferon-α further increased the frequency and percentage of CD41hi HSCs and reduced the number of JAK2-V617F+ HSCs in mice and patients with MPN. The shift toward the CD41hi subset of HSCs by interferon-α provides a possible mechanism of how interferon-α preferentially targets the JAK2 mutant clone.