Search

Supplementary Table S1. Genomic characteristic of ALL PDXs including ALL subtype and key genomic alterations. Supplementary Table S2. Top 5 enriched signaling pathways from the 52 query proteins, as analysed by GO and KEGG pathway analysis platforms. Supplementary Table S3. In vitro combination effects of BMS-754807 and ponatinib. Supplementary Table S4. Differentially expressed genes between Combination and Control (FDR < 0.05). Supplementary Table S5. Differentially expressed genes between BMS-754807 and Control (FDR < 0.05). Supplementary Table S6. Differentially expressed genes between Ponatinib and Control (FDR < 0.05). Supplementary Table S7. KEGG pathway analysis of the differentially expressed genes caused by each treatment condition against control (FDR < 0.25). Supplementary Table S8. Plasma concentration of BMS-754807 and ponatinib as measured by LC-MS/MS.

S1. In vitro effects of TSLP on target cells as assessed by Alamar Blue assay. S2. Demographics of phosphosites identified from MHH-CALL-4 and MUTZ-5 cells in global P-Tyr profiling. S3. In vitro single agent efficacy of BMS-754807 and ponatinib against CRLF2r cell lines and PDXs. S4. siRNA knockdown of IGF1R and FGFR1 in MHH-CALL-4 and MUTZ-5 cell lines. S5. Tolerability testing of ponatinib in combination with BMS754807 in naïve NSG mice as assessed by percentage of body weight change. S6. Ex vivo combination effect of BMS-754807 and ponatinib over time against CRLF2r Ph-like ALL PDXs. S7. Plasma concentrations of BMS-754807 and ponatinib at specific timepoints.

Supplementary Materials and Methods. Supplementary Table S1. QC of Affymetrix U133_Plus_2.0 Data. Supplementary Table S2. Summary of the Objective Response Measure (ORM) scoring methodology. Supplementary Table S3. Correlation of xenograft with parent U133_Plus_2.0 expression. Supplementary Table S4. Probe sets where parent expression is always greater than xenograft. Supplementary Table S5. Probe sets where xenograft expression is always greater than parent. Supplementary Table S6. Complete summary of in vivo AZD1480 single agent responses for individual mice. Supplementary Table S7. In vitro Combination Indices (CIs) for AZD1480 and selumetinib. Supplementary Table S8. In vivo responses of JAK-mutated ALL xenografts to AZD1480/selumetinib combination treatment. Supplementary Table S9. Complete summary of in vivo AZD1480 single agent and AZD1480/selumetinib combination responses for individual mice. Supplementary Figure S1. Chemical structures of (A) AZD1480 and (B) selumetinib. Supplementary Figure S2. Relationship between high CRLF2 expression and engraftment potential of the P9906 cohort. Supplementary Figure S3. Probe set expression by range. Supplementary Figure S4. Heatmap of microarray gene expression data corresponding to the qRT-PCR data shown in Figure 1. Supplementary Figure S5. In vitro cytotoxicity of AZD1480 against ALL xenograft cells. Supplementary Figure S6. In vivo efficacy of single agent AZD1480 against ALL xenografts. Supplementary Figure S7. Time of drug exposure affects synergistic interactions of AZD1480 and selumetinib against JAK-mutated ALL xenografts. Supplementary Figure S8. The extent of inhibition of the JAK/STAT and MAPK signaling pathways depends on the length of exposure to AZD1480/selumetinib.

Intensities of all Class I phosphosites p(STY) light (PDX) and heavy (SILAC) identified in 16 PDXs

Legend for Supplementary Figures S1-S9.

Description of additional methods and procedures used in the study. Also includes Supplementary References.

Phosphotyrosine sites identified by immunoaffinity profiling in ALL PDXs (S1); Comparison of phosphotyrosine enrichment from fresh and frozen PDX (S2); Optimization of quantitative phosphotyrosine profiling of pediatric ALL PDXs (S3); Number of identified Class I phosphosites in each of the PDXs (S4); Unsupervised hierarchical clustering of Class I phosphosites in ALL PDXs segregated based on molecularly-defined subgroups (S5); Comparison of ETP-ALL, Ph-like-ALL, BCP-ALL and T-ALL PDXs by gene expression profiling (S6); Comparison of phosphorylation levels of pTyr sites with gene expression profiling (S7); In vitro sensitivity of ALL PDXs to Sunitinib (S8); Venn diagram showing the number of Light PDX phosphopeptides and Heavy SILAC labelled phosphopeptides identified in each of the 16 PDXs (S9).

Activating mutations in tyrosine kinases (TK) drive pediatric high-risk acute lymphoblastic leukemia (ALL) and confer resistance to standard chemotherapy. Therefore, there is urgent need to characterize dysregulated TK signaling axes in patients with ALL and identify actionable kinase targets for the development of therapeutic strategies. Here, we present the first study to quantitatively profile TK activity in xenografted patient biopsies of high-risk pediatric ALL. We integrated a quantitative phosphotyrosine profiling method with “spike-in” stable isotope labeling with amino acids in cell culture (SILAC) and quantified 1394 class I phosphorylation sites in 16 ALL xenografts. Moreover, hierarchical clustering of phosphotyrosine sites could accurately classify these leukemias into either B- or T-cell lineages with the high-risk early T-cell precursor (ETP) and Ph-like ALL clustering as a distinct group. Furthermore, we validated this approach by using specific kinase pathway inhibitors to perturb ABL1, FLT3, and JAK TK signaling in four xenografted patient samples. By quantitatively assessing the tyrosine phosphorylation status of activated kinases in xenograft models of ALL, we were able to identify and validate clinically relevant targets. Therefore, this study highlights the application and potential of phosphotyrosine profiling for identifying clinically relevant kinase targets in leukemia. Cancer Res; 76(9); 2766–77. ©2016 AACR.

Philadelphia chromosome-like acute lymphoblastic leukemia (Ph-like ALL) is a high-risk ALL subtype with high rates of relapse and poor patient outcome. Activating mutations affecting components of the JAK-STAT signaling pathway occur in the majority of Ph-like ALL cases. The use of JAK inhibitors represents a potential treatment option for Ph-like ALL, although we and others have shown that CRLF2-rearranged Ph-like ALL responds poorly to single-agent JAK inhibitors in the preclinical setting. Therefore, the aim of this study was to identify effective combination treatments against CRLF2-rearranged Ph-like ALL, and to elucidate the underlying mechanisms of synergy. We carried out a series of high-throughput combination drug screenings and found that ruxolitinib exerted synergy with standard-of-care drugs used in the treatment of ALL. In addition, we investigated the molecular effects of ruxolitinib on Ph-like ALL by combining mass spectrometry phosphoproteomics with gene expression analysis. Based on these findings, we conducted preclinical in vivo drug testing and demonstrated that ruxolitinib enhanced the in vivo efficacy of an induction-type regimen consisting of vincristine, dexamethasone, and L-asparaginase in 2/3 CRLF2-rearranged Ph-like ALL xenografts. Overall, our findings support evaluating the addition of ruxolitinib to conventional induction regimens for the treatment of CRLF2-rearranged Ph-like ALL.

Philadelphia (Ph)-like acute lymphoblastic leukemia (ALL) is characterized by aberrant activation of signaling pathways and high risk of relapse. Approximately 50% of Ph-like ALL cases overexpress cytokine receptor-like factor 2 (CRLF2) associated with gene rearrangement. Activated by its ligand thymic stromal lymphopoietin (TSLP), CRLF2 signaling is critical for the development, proliferation, and survival of normal lymphocytes. To examine activation of tyrosine kinases regulated by TSLP/CRLF2, phosphotyrosine (P-Tyr) profiling coupled with stable isotope labeling of amino acids in cell culture (SILAC) was conducted using two CRLF2-rearranged (CRLF2r) Ph-like ALL cell lines stimulated with TSLP. As a result, increased P-Tyr was detected in previously reported TSLP-activated tyrosine kinases and substrates, including JAK1, JAK2, STAT5, and ERK1/2. Interestingly, TSLP also increased P-Tyr of insulin growth factor 1 receptor (IGF1R) and fibroblast growth factor receptor 1 (FGFR1), both of which can be targeted with small-molecule inhibitors. Fixed-ratio combination cytotoxicity assays using the tyrosine kinase inhibitors BMS-754807 and ponatinib that target IGF1R and FGFR1, respectively, revealed strong synergy against both cell line and patient-derived xenograft (PDX) models of CRLF2r Ph-like ALL. Further analyses also indicated off-target effects of ponatinib in the synergy, and novel association of the Ras-associated protein-1 (Rap1) signaling pathway with TSLP signaling in CRLF2r Ph-like ALL. When tested in vivo, the BMS-754807/ponatinib combination exerted minimal efficacy against 2 Ph-like ALL PDXs, associated with low achievable plasma drug concentrations. Although this study identified potential new targets in CRLF2r Ph-like ALL, it also highlights that in vivo validation of synergistic drug interactions is essential. Implication: Quantitative phosphotyrosine profiling identified potential therapeutic targets for high-risk CRLF2-rearranged Ph-like ALL.

Purpose: Directing nanoparticles to cancer cells without using antibodies is of great interest. Subtle changes to the surface chemistry of nanoparticles can significantly affect their biological fate, including their propensity to associate with different cell populations. For instance, nanoparticles functionalized with thiol-reactive groups can potentially enhance association with cells that over-express cell-surface thiol groups. The potential of such an approach for enhancing drug delivery for childhood acute lymphoblastic leukemia (ALL) cells has not been investigated. Herein, we investigate the impact of thiol-reactive star polymers on the cellular association and the mechanisms of uptake of the nanoparticles.Methods: We prepared fluorescently labeled star polymers functionalized with an mPEG brush corona and pyridyl disulfide to examine how reactivity to exofacial thiols impacts cellular association with ALL cells. We also studied how variations to the mPEG brush composition could potentially be used as a secondary method for controlling the extent of cell association. Specifically, we examined how the inclusion of shorter diethylene glycol brush moieties into the nanoparticle corona could be used to further influence cell association.Results: Star polymers incorporating both thiol-reactive and diethylene glycol brush moieties exhibited the highest cellular association, followed by those functionalized solely with thiol reactive groups compared to control nanoparticles in T and B pediatric ALL patient-derived xenografts harvested from the spleens and bone marrow of immunodeficient mice. Transfection of cells with an early endosomal marker and imaging with correlative light and electron microscopy confirmed cellular uptake. Endocytosis inhibitors revealed dynamin-dependent clathrin-mediated endocytosis as the main uptake pathway for all the star polymers.Conclusion: Thiol-reactive star polymers having an mPEG brush corona that includes a proportion of diethylene glycol brush moieties represent a potential strategy for improved leukemia cell delivery.

Antibody fragments are promising building blocks for developing targeted therapeutics, thus improving treatment efficacy while minimising off-target toxicity. Despite recent advances in targeted therapeutics, patients with Philadelphia-like acute lymphoblastic leukemia (Ph-like ALL), a high-risk malignancy, lack specific and effective targeted treatments. Cytokine receptor-like factor 2 (CRLF2) is overexpressed in 50% of Ph-like ALL cases, conferring the survival of leukemia blasts through activation of the JAK/STAT signalling pathway. Targeting such a vital cell-surface protein could result in potent anti-leukaemic efficacy and reduce the likelihood of relapse associated with antigen loss. Herein, we developed a novel single-chain variable fragment (scFv) against CRLF2 based on a monoclonal antibody raised against the recombinant extracellular domain of human TSLPRα chain. The scFv fragment demonstrated excellent binding affinity with CRLF2 protein in the nanomolar range. Cellular association studies in vitro using an inducible CRLF2 knockdown cell line and ex vivo using patient-derived xenografts revealed the selective association of the scFv with CRLF2. The fragment exhibited significant receptor antagonistic effects on STAT5 signalling, suggesting possible therapeutic implications in vivo. This study is the first to describe the potential use of a novel scFv for targeting Ph-like ALL.

Introduction: Philadelphia-like acute lymphoblastic leukaemia (Ph-like ALL) is a high-risk ALL subtype characterized by an inferior survival rate and chemotherapeutic drug resistance (Tasian et al, Blood 130: 2064-2072, 2017). Around 50% of Ph-like ALL cases harbour gene rearrangements leading to the overexpression of cytokine receptor-like factor 2 (CRLF2) (Loh et al, Blood 121: 485-488, 2013). CRLF2 (also known as thymic stromal lymphopoietin receptor, TSLPR) heterodimerizes with the interleukin-7 receptor alpha (IL-7Rα) subunit to form the functional TSLPR. Upon TSLP binding, CRLF2 activates the JAK/STAT signalling pathway, leading to enhanced proliferation and survival of leukemia cells resulting in poor outcomes in patients (Harvey et al, Blood 115: 5312-5321, 2010). The extracellular overexpression of CRLF2 and association with poor prognosis suggest the translational value of designing precision-based therapeutics targeting CRLF2 in Ph-like ALL. Although conventional immunotherapy using full-sized antibodies is a promising treatment strategy that can improve treatment efficiency and minimize off-target toxicity, their clinical translation is challenging due to the high production cost and large size affecting targeting efficiency (Holliger et al, Nat Biotech 23: 1126-1136, 2005). Herein, we validated a novel single-chain variable fragment against CRLF2 (CRLF2-ScFv) for targeting Ph-like ALL cells. Methods: A CRLF2-rearranged Ph-like ALL cell line (MHH-CALL-4) was lentivirally transduced with a CRLF2-targeting shRNA driven by an inducible promoter, enabling the inducible knockdown of CRLF2. CRLF2 expression in MHH-CALL-4 cells after shRNA induction (KD-CALL-4) was evaluated using fluorescence-activated cell sorting (FACS). The cellular association of the CRLF2-ScFv was determined in MHH-CALL-4 and KD-CALL-4 at 4° and 37°C using an indirect immunofluorescence assay. Confocal laser scanning microscopy was used to visualize and compare the cellular association of CRLF2-ScFv in MHH-CALL-4 and KD-CALL-4. The cellular association of CRLF2-ScFv was also investigated ex vivo using a small panel of Ph-like and non-Ph-like ALL patient-derived xenografts (PDXs), representing similar immunophenotype and genetic characteristics to their original disease subtypes (Liem et al, Blood 103: 3905-3914, 2004), and peripheral blood mononuclear cells (PBMCs) to investigate the non-selective association. CRLF2 expression in MHH-CALL-4 and Ph-like ALL PDX cells was quantified using indirect immunofluorescence assay. The downstream impact of CRLF2-ScFv on STAT5 phosphorylation (pSTAT5) was determined by FACS either with or without TSLP stimulation. The statistical significance was tested using Unpaired unequal variances t-test or one-way ANOVA followed by multiple comparisons test. Statistical significance was considered when P ≤ 0.05. All experiments were performed in triplicate. Results: KD-CALL-4 showed a 75% reduction in CRLF2 expression compared with MHH-CALL-4 cells (P = 0.0009). CRLF2-ScFv exhibited a 94% higher association with MHH-CALL-4 compared with KD-CALL-4 cells at 37°C (P = 0.0013). The association of CRLF2-ScFv with MHH-CALL-4 cells was reduced by 75% at the non-proliferating state of cells at 4°C compared to 37°C (P = 0.006). Orthogonally viewed confocal microscopy images showed 82% higher intracellular uptake of CRLF2-ScFv in MHH-CALL-4 compared to KD-CALL-4 cells (P = 0.0003). CRLF2-ScFv showed >80% higher association with a Ph-like PDX sample compared with a control CRLF2low PDX and PBMCs (P < 0.001). Of note, a Ph-like ALL PDX exhibited only one-third of the association with CRLF2-ScFv compared with MHH-CALL-4 cells (P = 0.04), corresponding to the significant difference in CRLF2 surface expression (P = 0.01). CRLF2-ScFv significantly reduced pSTAT5 expression in MHH-CALL-4 cells (P = 0.05) and prevented TSLP-induced STAT5 phosphorylation (P = 0.01), suggesting competition with the TSLP binding site. Conclusion: CRLF2-ScFv is a selective targeting moiety for CRLF2 with a significant internalization potential and receptor antagonistic effect, highlighting the therapeutic implications for targeting Ph-like ALL. Keywords: CRLF2, ScFv, STAT5 phosphorylation, Patient-Derived Xenografts. Disclosures No relevant conflicts of interest to declare.

Genome-wide studies have identified a high-risk subgroup of pediatric acute lymphoblastic leukemia (ALL) harboring mutations in the Janus kinases (JAK). The purpose of this study was to assess the preclinical efficacy of the JAK1/2 inhibitor AZD1480, both as a single agent and in combination with the MEK inhibitor selumetinib, against JAK-mutated patient-derived xenografts. Patient-derived xenografts were established in immunodeficient mice from bone marrow or peripheral blood biopsy specimens, and their gene expression profiles compared with the original patient biopsies by microarray analysis. JAK/STAT and MAPK signaling pathways, and the inhibitory effects of targeted drugs, were interrogated by immunoblotting of phosphoproteins. The antileukemic effects of AZD1480 and selumetinib, alone and in combination, were tested against JAK-mutated ALL xenografts both in vitro and in vivo. Xenografts accurately represented the primary disease as determined by gene expression profiling. Cellular phosphoprotein analysis demonstrated that JAK-mutated xenografts exhibited heightened activation status of JAK/STAT and MAPK signaling pathways compared with typical B-cell precursor ALL xenografts, which were inhibited by AZD1480 exposure. However, AZD1480 exhibited modest single-agent in vivo efficacy against JAK-mutated xenografts. Combining AZD1480 with selumetinib resulted in profound synergistic in vitro cell killing, although these results were not translated in vivo despite evidence of target inhibition. Despite validation of target inhibition and the demonstration of profound in vitro synergy between AZD1480 and selumetinib, it is likely that prolonged target inhibition is required to achieve in vivo therapeutic enhancement between JAK and MEK inhibitors in the treatment of JAK-mutated ALL. Mol Cancer Ther; 14(2); 364–74. ©2014 AACR.

Summary Glucocorticoids play a critical role in the treatment of lymphoid malignancies. While glucocorticoid efficacy can be largely attributed to lymphocyte-specific apoptosis, its molecular basis remains elusive. Here, we studied genome-wide lymphocyte-specific open chromatin domains (LSOs), and integrated LSOs with glucocorticoid-induced RNA transcription and chromatin modulation using an in vivo patient-derived xenograft model of acute lymphoblastic leukemia (ALL). This led to the identification of LSOs critical for glucocorticoid-induced apoptosis. Glucocorticoid receptor cooperated with CTCF at these LSOs to mediate DNA looping, which was inhibited by increased DNA methylation in glucocorticoid-resistant ALL and non-lymphoid cell types. Our study demonstrates that lymphocyte-specific epigenetic modifications pre-determine glucocorticoid resistance in ALL and may account for the lack of glucocorticoid sensitivity in other cell types.

Glucocorticoids are well known for their immunosuppressive activity, and play a critical role in the treatment of lymphoid malignancies. However, the development of resistance remains a significant barrier to cure and the mechanisms are poorly defined. Moreover, glucocorticoids are rarely efficacious in treating myeloid and other non-lymphoid malignancies, and the mechanisms of lymphocyte-specific efficacy are unclear. To address these issues, we first carried out a global analysis of DNase I hypersensitive sites in 18 lymphoid and 64 non-lymphoid cell types to map lymphocyte-specific open chromatin domains (LSOs). We then integrated these domains with genome-wide glucocorticoid-induced gene transcription and epigenetic modulation in an in vivo patient-derived xenograft (PDX) model of acute lymphoblastic leukemia (ALL). Performing chromatin immunoprecipitation and assays for transposase-accessible chromatin (ATAC), we determined a strong correlation between glucocorticoid receptor (GR) binding and chromatin accessibility, acetylated histone marks and binding of a DNA structural protein (CTCF), and identified 1,536 GR bound LSOs. We next analyzed RNA-seq data in glucocorticoid sensitive and resistant ALL PDXs for expression changes in genes located within 100 kb of the LSOs after glucocorticoid treatment in vivo, and identified four groups comprising 389 genes that were significantly differentially expressed. Of the 198 up-regulated genes, 143 showed increased H3K27Ac enrichments at 177 LSOs. Forty-two LSOs showed the increase only in glucocorticoid sensitive but not resistant ALLs. Applying this to an extended panel of ALL PDXs, basal DNA methylation at the 42 LSOs was significantly higher in resistant ALLs despite no difference within the gene bodies, whereas the basal chromatin accessibility indicated by ATAC abundance was diminished. Pathway analysis indicated that the LSO regulated genes were involved in repressing B- and T- cell receptor signaling pathways and activating the apoptotic pathway. One such LSO was at the pro-apoptotic BIM gene locus, where CTCF binding was found only in lymphocytes but not in other cell types. The GR cooperated with CTCF to mediate interactions between the promoter and the BIM LSO to direct DNA looping, thus triggering BIM transcription. Importantly, this LSO was heavily methylated in resistant ALLs as well as non-lymphoid cells. Azacitidine, a DNA demethylating drug that is routinely used in the clinic, could partially reverse these changes and restore glucocorticoid sensitivity. Taken together, this study demonstrated for the first time that lymphocyte-specific epigenetic modifications pre-determine glucocorticoid resistance in ALL and may account for the lack of glucocorticoid sensitivity in other cell type. Reversal of these epigenetic changes may lead to improvements in the use of glucocorticoids in the clinic. Citation Format: Duohui Jing, Yizhou Huang, Xiaoyun Liu, Keith Sia, Rebecca C. Poulos, Miriam Span, Chao Zhang, Jianqing Mi, Jason WH Wong, Dominik Beck, John E. Pimanda, Richard B. Lock. Lymphocyte-specific chromatin accessibility predetermines glucocorticoid resistance in acute lymphoblastic leukemia [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2018; 2018 Apr 14-18; Chicago, IL. Philadelphia (PA): AACR; Cancer Res 2018;78(13 Suppl):Abstract nr 3173.

Introduction: While cure rates for children with acute lymphoblastic leukemia (ALL) are approaching 90% with conventional chemotherapeutic regimens, certain high-risk patient subsets such as early T-cell precursor ALL (ETP-ALL) and Philadelphia Chromosome-like (Ph-like) ALL have an aggressive disease profile and poor prognosis. More recently whole genome and transcriptome sequencing of these high-risk subtypes have revealed several activating gene fusions, alterations and mutations that could result in constitutively activated tyrosine kinases (TKs). Activated TKs are then capable of phosphorylating downstream substrates and impacting several key signaling pathways, resulting in increased cell survival, proliferation and differentiation. Further, the highly heterogeneous nature of these subtypes, along with activating fusions/mutations, makes them refractory to standard chemotherapy. Consequently, there is an urgent need to develop tailored therapeutic strategies for the treatment of these high-risk ALL subtypes. Recent advances in mass-spectrometry and the use of anti-phosphotyrosine antibodies for enrichment of tyrosine phosphorylated peptides have greatly facilitated characterization of the global tyrosine phosphorylation state in cancer cells and identified activated TKs that could be therapeutically targeted. Here we present the first study to quantitatively profile TK activity in xenografted patient biopsies of high-risk pediatric ALL. Methods: In this study, we have established an MS-based phosphotyrosine profiling approach in patient derived xenografts (PDXs) of high-risk pediatric ALL patients and integrated it with a spike-in SILAC quantitative tool to identify and quantify dysregulated TK activity across 16 PDXs. We further extended our study on markedly altered tyrosine phosphorylation in 4 PDXs to assess the therapeutic potential of specific TK inhibitors (TKIs). Immunoblots were performed to validate activated sites and their dephosphorylation upon TKI treatment. RT-PCR and Exome sequencing was carried out to detect novel fusion partners and point mutation sites to validate the activated TK profiles in these PDXs. In vitro cytotoxicity was assessed by mitochondrial metabolic activity assay (Alamar blue) following 48h drug exposures. PDXs were established from ETP-ALL, Ph-like ALL, B-cell precursor (BCP)-ALL, or T-lineage ALL (T-ALL) bone marrow or peripheral blood (PB) biopsies in immune-deficient (NOD/SCID or NSG) mice. Engraftment and in vivo drug responses were assessed by enumeration of the proportion of human versus mouse CD45+cells in the murine PB. Results: Using a quantitative phosphotyrosine profiling method in 16 PDXs, we mapped close to 1900 class I phosphosites with >0.75 localization probability and 99% confidence, of which 1394 tyrosine phosphorylated sites had a heavy SILAC partner that allowed quantification. Such profiling could accurately classify the leukemias into either T or B-cell lineages with the high-risk ETP and Ph-like ALL clustering as a distinct group. In particular, PDXs with activated tyrosine phosphorylation profiles of ABL1, FLT3 and JAK were targeted with commercially available TKIs both in vitro and in vivo. Subsequent analysis to investigate the aberrant ABL1 and FLT3 signaling showed a NUP214-ABL1 translocation unique to BCP-ALL in one PDX, and a novel Y572S FLT3 mutation in another. Importantly, using a pre-clinical in vivo xenograft model, the activated JAK-STAT signaling observed in one ETP-ALL PDX was targeted with the JAK1/2 inhibitor, ruxolitinib, leading to a significant decrease in the leukemic blast population in the murine PB. Aberrant ABL1 kinase signaling indicated dasatinib treatment in a Ph+-ALL PDX and a PDX with high phospho-ABL1 (harboring the NUP214-ABL1 translocation), and a complete response and significant progression delay, respectively, were achieved in vivo. Similarly, the uniquely activated FLT3 in one PDX (Y572S mutation) correlated with an in vivoobjective response to the multi-kinase inhibitor sunitinib. Conclusions: This study demonstrates the direct application of an unbiased and quantitative tool to identify aberrant TK signaling in high-risk ALL PDXs and highlights its potential to identify tractable drug targets. This research was supported by NCI NO1CM42216 and by the Australian National Health and Medical Research Council. Disclosures No relevant conflicts of interest to declare.

Abstract 3562 Relapsed/refractory pediatric acute lymphoblastic leukemia (ALL) remains a continuing challenge to treat with currently available therapies, and new treatments are urgently required for the management of these high-risk cases. Activating mutations in the pseudokinase or kinase domains of Janus kinases (JAKs) 1, 2 and 3 are present in approximately 10% of high-risk pediatric ALL and are associated with high expression of cytokine receptor-like factor 2 (CRLF2) and poor outcome. These mutations can lead to continuous activation of JAKs, resulting in constitutive activation by phosphorylation of downstream signaling, including the signal transducer and activator of transcription (STAT), AKT, and mitogen-activated protein kinase (MAPK) pathways. The availability of specific JAK inhibitors, developed primarily for the treatment of JAK-mutated myeloproliferative diseases (MPDs), represents an opportunity to improve the treatment options for JAK-mutated pediatric ALL. AZD1480, a potent ATP-competitive small-molecule JAK2 inhibitor that also exhibits inhibitory activity against JAK1, is in solid tumor clinical trials. The purpose of this study was to gain a greater understanding of a potential role for AZD1480 in the treatment of JAK-mutated pediatric ALL either as a single agent or in rational drug combinations, using a preclinical model of xenografts established in immune-deficient mice from direct patient explants. As part of the Pediatric Preclinical Testing Program (PPTP) we previously showed that AZD1480 administered at 10 mg/kg twice daily × 5 then at 15mg/kg once daily × 2 via oral gavage for an intended three weeks significantly delayed the progression of only one in five JAK-mutated xenografts, with no tumor regressions observed. We now show that the relative insensitivity of JAK-mutated ALL xenografts to AZD1480 is a cell-intrinsic phenomenon, since 6/7 JAK1- or JAK2-mutated xenografts exhibited ex vivo IC50 values >2 μM following 72 h drug exposures, as assessed by mitochondrial function cell viability (MTT) assay. In order to gain a greater understanding of the underlying mechanisms for the lack of AZD1480 single-agent efficacy against JAK-mutated xenografts we analyzed intracellular signaling pathways and their responses to AZD1480 treatment. In contrast with “Typical” B-cell precursor (BCP)-ALL xenograft cells, JAK-mutated xenografts exhibited constitutive JAK pathway activation, as assessed by increased levels of phospho-JAK1 (pJAK1), pJAK2, pSTAT1/3/5, pAKT, pMAP2K1/2 (MEK1/2) and phospho-extracellular signal-regulated kinase 1/2 (pERK1/2). Ex vivo exposure of JAK-mutated xenografts to 1 μM AZD1480 caused rapid (within 1 h) and sustained (up to 24 h) decreases in pSTATs, but minimal reduction in pMEK1/2 and pERK1/2. These results indicate that AZD1480 alone selectively inhibits JAK downstream signaling pathways, which may be insufficient to delay leukemia progression in vivo or induce cell death ex vivo. Moreover, they provide a rationale for dual targeting of the JAK and MAPK pathways to elicit synergistic anti-leukemic cell killing in JAK-mutated ALL. Ex vivo exposure of two JAK2-mutated xenografts to 1 μM of the MEK1/2 inhibitor AZD6244 (selumetinib) caused a profound decrease in pERK1/2, and the combination of 1 μM each of AZD1480 and AZD6244 resulted in reductions of both pSTATs and pERK1/2. Moreover, fixed-ratio MTT cytotoxicity assays using these two JAK2-mutated xenografts demonstrated very strong synergy between AZD1480 and AZD6244, with Combination Indices for each xenograft of 0.36 and 0.098 at the ED50; 0.23 and 0.015 at the ED75; and 0.15 and 0.002 at the ED90. This strong synergistic effect was observed despite AZD1480 and AZD6244 exerting minimal cell killing activity against the xenograft cells when used as single agents. In conclusion, our data indicate that AZD1480 is unlikely to exert significant single-agent activity in the treatment of JAK-mutated pediatric ALL, and that future efforts focusing on dual targeting of the JAK/STAT and MAPK offer a potential pathway to achieving clinical efficacy. Disclosures: No relevant conflicts of interest to declare.

Summary To design rational therapies for JAK2-driven hematological malignancies, we functionally dissected the key survival pathways downstream of hyperactive JAK2. In tumors driven by mutant JAK2, Stat1, Stat3, Stat5, and the Pi3k and Mek/Erk pathways were constitutively active, and gene expression profiling of TEL-JAK2 T-ALL cells revealed the upregulation of prosurvival Bcl-2 family genes. Combining the Bcl-2/Bcl-xL inhibitor ABT-737 with JAK2 inhibitors mediated prolonged disease regressions and cures in mice bearing primary human and mouse JAK2 mutant tumors. Moreover, combined targeting of JAK2 and Bcl-2/Bcl-xL was able to circumvent and overcome acquired resistance to single-agent JAK2 inhibitor treatment. Thus, inhibiting the oncogenic JAK2 signaling network at two nodal points, at the initiating stage (JAK2) and the effector stage (Bcl-2/Bcl-xL), is highly effective and provides a clearly superior therapeutic benefit than targeting just one node. Therefore, we have defined a potentially curative treatment for hematological malignancies expressing constitutively active JAK2., Graphical Abstract, Highlights • Bcl-2 and Bcl-xL are the key survival factors downstream of oncogenic JAK2 • Combined targeting of JAK2 and Bcl-2/Bcl-xL is highly efficacious in vivo • Combination therapy is well tolerated in preclinical models of JAK2-driven ALL • Combination therapy can overcome and circumvent resistance to JAK2 inhibitors, Better therapeutic options are required for mutant JAK2-driven hemopoietic malignancies to overcome intrinsic and acquired therapy resistance. Johnstone and colleagues functionally dissect signaling pathways downstream of hyperactive JAK2 and define the STAT5-Bcl-2/Bcl-xL axis as crucial for tumor cell survival and development of acquired JAK inhibitor resistance. Combined inhibition of this oncogenic JAK2 signaling network at two critical points—JAK2 and Bcl-2/Bcl-xL—proved highly efficacious in vivo and was able to circumvent and overcome acquired resistance to single-agent JAK inhibitors.