ALBERT

Description: Introduction : Protein phosphatase 2A (PP2A) is a Ser/Thr phosphatase negatively regulates a diverse set of signaling pathways promoting tumor growth. While PP2A functions as tumor suppressor in multiple types of cancer, here, we demonstrate an unexpected pro-survival role of PP2A in pre-B acute lymphoblastic leukemia (ALL) cells. By calibrating PI3K-AKT-mTOR signaling strength, PP2A regulates glycolysis rate and thereby balances energy demands against anti-oxidant protection of pre-B ALL cells. In addition, PP2A reinstates activity of FOXO factors by dephosphorylation and thereby enables the anti-oxidant function of FOXO1 and FOXO3. Consistent with previous studies, we find PP2A is dispensable for the survival of myeloid leukemia cells which indicates a lineage-specific role of PP2A. Combined with genetic study and small molecule inhibitor, we verify that regulation in patient-derived xenografts and highlight PP2A as a therapeutic target in pre-B ALL. Results: Consistent with a divergent role of PP2A in pre-B ALL (compared to CML), we found that high mRNA levels of PP2A subunits at the time of diagnosis predict poor outcome of children (COG P9906; n=207) and adults (ECOG 2993; n=215) with ALL. Consistent with these findings, mutations in PP2A subunits are extremely rare in B cell malignancies but relatively common in solid tumors and myeloid malignancies (COSMIC). We therefore, studied the function of PP2A in a genetic mouse model for Cre-induced deletion of Ppp2r1a in BCR-ABL1 (Ph+) ALL. Inducible activation of Cre reduced protein expression of the targeted PP2A subunit A and the catalytic subunit C, which leads to near-complete loss of PP2A phosphatase activity. Conversely, Cre-mediated deletion increased phosphorylation levels of FoxO1, FoxO3a, p70S6K and S6 ribosomal protein, which indicated elevated PI3K-Akt-mTOR signaling. Acute deletion of Ppp2r1afl/fl in B cell-lineage ALL cells dramatically affected survival and colony formation, both of which could be rescued by overexpression of wildtype PP2A. PI3K and mTOR inhibitors also have rescue effect on PP2A deficient ALL cells in growth-competition assay. Luciferase-labeled PP2A-deleted ALL cells showed reduced cell growth and leukemia progression after being transplanted into recipient mice. However, Cre-mediated deletion had no deleterious effects in a Ppp2r1afl/fl CML model. This lineage-specific role of PP2A was verified by inducible CEBPα expression to reprogram B cell lineage ALL cells into myeloid cells. Interestingly, inducible deletion of PP2A caused profound imbalances of glucose metabolism in Ph+ ALL but not in CML cells. Upon PP2A-deletion, ALL cells showed higher glycolytic flux shunted into lactate rather than NADPH production. By employing glucose flux metabolic profiling assay using [1,2-13C2]-D-glucose tracer, we found elevated glycolysis and repressed pentose phosphate pathway (PPP) flux in PP2A-deleted pre-B ALL cells. Lower NADPH/NADP ratio and higher reactive oxygen species level in PP2A-deleted pre-B ALL cells, together with decreased anti-oxidant gene expression, increased DNA damage, including H2AX phosphorylation and p53 expression. The mechanistic role of ROS downstream of PP2A was supported by a strong rescue effect of overexpression of the antioxidant catalase in PP2A-deleted cells. The unexpected role of PP2A in Ph+ ALL was further validated by CRISPR-Cas9 mediated disruption of PPP2R1A in ALL xenografts derived from three patients. In addition, a PP2A specific inhibitor LB-100 (in clinical trial for solid tumors) was employed to pharmacologically inhibit PP2A activity. Low micromolar concentrations of LB-100 induced cell death in patient-derived ALL xenografts in parallel with ROS-accumulation and increased S6 and H2AX phosphorylation. Conclusion: Here we revealed an unexpected role of PP2A in maintaining redox homeostasis in pre-B ALL cells. By regulating AKT-mTOR signaling, PP2A keeps the balance of glycolysis and PPP to meet the energy demands of pre-B ALL cells and avoids extreme levels of oxidative stress. We confirmed this pro-survival role of PP2A in both genetic mouse ALL model and human Ph+ ALL-patients derived leukemia cells. These findings highlight PP2A as a therapeutic target and suggest that agents like the PP2A inhibitor LB-100 may be of interest for pre-clinical development and testing. Disclosures No relevant conflicts of interest to declare.

Description: Background and significance: The B-lymphoid transcription factors IKZF1 and PAX5 are essential for early B-cell development but also function as metabolic gatekeepers by restricting glucose uptake across the cell membrane (Chan et al., Nature 2017; Xiao et al., Cell 2018). Paraoxonase 2 (PON2) is used as a diagnostic marker and included in a 15 gene diagnostic LDA panel (NCT02883049) for the identification of Ph+ and Ph-like ALL, a B-ALL subgroup with poor outcome and frequent deletion of IKZF1. Pon2 is a member of detoxifying enzymes that are located to the mitochondrial membrane, and hydrolyze lactone metabolites. While PON2 is typically not expressed in normal B-cell precursors, we studied the functional role of high expression levels of Pon2 in B-ALL. Results: Indeed, high Pon2 mRNA levels at diagnosis predicted poor clinical outcomes for both children (COG P9906; n=207; P=1.09e-05) and adults (ECOG; n=215; P=0.003). While PON2 mRNA and protein levels are barely detectable in normal B-cell precursors we found 〉10-fold increased mRNA and protein levels in multiple B-ALL. Pon2-deficient mice have a number of subtle defects, including propensity to neurogenerative disease and atherosclerosis. However, deletion of Pon2 had profound effects in two models for B-ALL (BCR-ABL1- and NRASG12D). Compared to Pon2 wildtype, Pon2-/- B-ALL cells failed to form colonies in semisolid agar, were prone to cell cycle arrest in G0/G1 phase. Consistent with these observations, Pon2-/- B-ALL cells had a substantially reduced proliferation rate and expressed Arf and p21 checkpoint proteins at substantially increased levels compared to Pon2+/+ ALL cells. These in vitro findings suggest an important role of PON2 in B-lymphoid leukemogenesis, which was confirmed in mouse transplant experiments for BCR-ABL1 and NRASG12D B-ALL models. Pon2-deficiency substantially prolonged survival of recipient mice of BCR-ABL1 B-ALL cells (P=0.0001). Strikingly, NSG mice transplanted with Pon2-deficient NRASG12D B-ALL cells did not develop any leukemia and survived for indefinite periods of time. Based on engineered expression of Cas9 and PON2-specific cRNAs, genetic deletion of PON2 replicated these findings in patient-derived B-ALL xenografts (PDX). Mechanism: Reflecting PAX5 and IKZF1 function, B-lymphoid cells exhibit transcriptional repression of glucose transport (e.g. repression of INSR, GLUT1). Interestingly, deletion of PON2 decreased glucose uptake even further. Intracellular ATP levels in Ph+ ALL cells were drastically reduced and activated the AMPK energy-stress sensor in both murine BCR-ABL1 and NRASG12D B-ALL models and human Ph+ B-ALL cells. Reconstitution of PON2 restored ATP production as early as 3 hours after PON2 induction. Mechanistically, PON2 was previously identified as an inhibitor of STOM and GLUT1 interactions. By disrupting STOM: GLUT1 interactions, PON2 can restore glucose transport. STOM-knockdown in PON2-/- ALL cells replenished cellular ATP levels as did PON2 reconstitution. We conclude that PON2 serves a critical role in B-ALL by subverting the metabolic gatekeeper function of IKZF1 and PAX5 through interference with STOM:GLUT1 interactions. Therapeutic targeting of PON2: While PON2 activity typically results in detoxification of lactone-metabolites, lactone-hydrolysis of the N-(3-oxododecanoyl)-homoserine lactone (3OC12) prodrug results in cytotoxic byproducts. Hence, we tested the concept of leveraging PON2 lactonase activity in a prodrug-approach as a synthetic lethal in Pon2+ B-ALL cells. 3OC12 had profound cytotoxic effects in wildtype but not Pon2-/- B-ALL cells while reconstitution or overexpression of PON2 restored cytotoxic responses to 3OC12. Conclusion: We describe the previously unrecognized function of the detoxifying PON2 lactonase as a central factor in glucose and energy supply in B-ALL, hence subverting metabolic gatekeeper functions of PAX5 and IKZF1. In addition to identifying PON2 as the biomarker for for Ph+ and Ph-like ALL and outcome predictor for patients with B-ALL, our results indicate PON2 facilitate glucose uptake and ATP production via its interaction with STOM. PON2 protects B-ALL cells against glucose depletion and enables malignant growth. On the other hand, we demonstrate a targeted approach that leverages PON2-lactonase as a synthetic lethal to the lactone-prodrug 3OC12. Disclosures No relevant conflicts of interest to declare.

Description: Background and hypothesis: CD25 (IL2RA, interleukin 2 receptor α chain) is a transmembrane protein with a 13aa cytoplasmic tail. CD25 cooperates with β- and γ-chains in binding IL-2, but does not contribute to cytokine signaling. During normal B cell development, CD25 is specifically upregulated on the surface of IL7-dependent pre-B cells and is also expressed on the surface of a subset of human pre-B ALL cases. CD25-expressing ALL is typically associated with poor clinical outcome. For these reasons, we studied the functional significance of CD25 expression on human pre-B ALL cells. Results: Flow cytometry and immunohistochemistry staining on a large panel of patient samples (n=416; MDACC, ECOG) revealed specific cell surface expression of CD25 in Ph+ ALL and Ph-like ALL, which are both high-risk subtypes of ALL. In agreement with selective expression on high-risk subsets, high expression levels of CD25 at the time of diagnosis were predictive of poor overall clinical outcome in these studies (P=0.005). BCR-ABL1 in Ph+ ALL and related tyrosine kinases in Ph-like ALL strongly activate STAT5, which then induces transcriptional activation of the IL2RA locus. Since Stat5 is also active during normal pre-B cell differentiation, we first analyzed B cell development in Il2ra-/- mouse bone marrow. Il2ra-/- B cell development was blocked at the pre-B cell stage, consistent with specific upregulation of CD25 on pre-B cells. In human Ph+ ALL cells, we found IL2RB and IL2RG were not co-expressed with CD25, suggesting a function of CD25 in Ph+ALL that is distinct from IL2 signaling. To test the biological significance of tyrosine kinase/STAT5-induced activation of CD25, we developed an Il2ra-/- mouse model for BCR-ABL1 pre-B ALL. Interestingly, the cytoplasmic tail of CD25 includes phosphorylation sites (S268 and T271) that are known substrates for serine/threonine phosphorylation by PKCα, which was reported to regulate protein phosphatase 2A (PP2A). To investigate interacting proteins with the cytoplasmic tail of CD25, we performed immune precipitation (IP) against the flag-tagged CD25-tail in primary Ph+ ALL cells which were transduced with either a CD25-tail-flag or an EV-flag vector. 2D mass spectrometry and Western blot on the IP products confirmed strong interactions with PKCα and PP2A. Weestern blot analysis confirmed additional interactions with inhibitory phosphatases including PTEN, PTPN6 (SHP1) and Inpp5d (SHIP1) in human Ph+ALL cells. In addition, both 2D MS and Westernblot showed recruitment of the Stat5-feedback inhibitors CISH, SOCS2 and SOCS3 at the CD25 cytoplasmic tail. Studying functional parameters of Il2ra-/-BCR-ABL1 ALL cells, we found impaired proliferation and colony formation capacity and drastically increased increased phosphorylation levels of pABLY412, pSTAT5Y694, pERKT202/Y204, pAKTS473, pP38T180/Y182 and p53. Reconstitution of CD25 expression restored normal phosphorylation levels of these molecules, as well as proliferation and colony formation.In a serial transplant setting, we observed that leukemia initiation in transplant recipients from Il2ra-/- BCR-ABL1 ALL cells required 10- to 100-times higher cell numbers, suggesting that CD25 contributes to leukemia initiation. In addition, CD25 expression is associated with a higher level of drug-resistance: In patient-derived pre-B ALL cells with mixed CD25Low and CD25High populations, the standard chemotherapy agent vincristine selectively induced apoptosis of in CD25Low but not CD25High ALL cells. An anti-CD25 immunotoxin drugs efficiently eradiated CD25High leukemia cells and thereby overcame drug-resistance against vincristine. Conclusions: Our studies identified CD25 as a surface receptor that mediates membrane recruitment of PP2A and CISH, SOCS2, negative feedback regulators of STAT5. CD25 is transcriptionally activated by STAT5 and therefore specifically expressed on high-risk ALL subtypes with oncogenic activation of the Stat5 pathway (Ph+ ALL and Ph-like ALL). We propose that CD25-mediated negative feedback control stabilizes oncogenic tyrosine kinase signaling and mediates drug-resistance in Ph+ ALL and Ph-like ALL cells. Targeted inhibition using CD25-directed immunotoxins may be useful in new approaches to overcome drug-resistance in Ph+ ALL and Ph-like ALL. Disclosures No relevant conflicts of interest to declare.

Description: Background and hypothesis: CD25 (IL2RA) represents the α chain of the interleukin 2 receptor on T cells and plays an important role in the maintenance of regulatory T (Treg) cells, hence preventing T cell autoimmunity. In a comprehensive gene expression analysis, we found that CD25 is specifically upregulated by pre-B cell receptor (pre-BCR) signaling during early B cell development and oncogenic tyrosine kinase that mimic pre-BCR signaling (e.g. in Ph+ ALL and Ph-like ALL). In adults with Ph+ ALL (ECOG; MDACC) and children with Ph-like ALL (P9906) patients with CD25 expression at the time of diagnosis have a particularly poor outcome (n=416; P=0.005). For these reasons, we studied the function of CD25 in B cell development and leukemia in a series of genetic experiments. Results: Unlike T cells, CD25 (IL2RA) does not function as IL2 receptor chain in B cells and B-lineage ALL: CD25 expressed on B-lineage cells did not pair with IL2Rb and g-chains and was not responsive to IL2. Il2ra-/- B cells were arrested at the pre-B cell stage with hyperactive pre-BCR downstream signaling including SRC, BTK and ERK. In the presence of CD25, Il2ra+/+ B cells responded to engagement of the pre-BCR with phosphorylation of pre-BCR downstream tyrosine kinases and coordinated release of Ca2+ from cytoplasmic stores. In the absence of CD25 (Il2ra-/-), the pre-BCR signals autonomously, resulting in uncoordinated Ca2+ oscillations of variable duration. While CD25 does not function as IL2 receptor chain in B cells, it coordinates pre-BCR-dependent signal transduction and regulates its intensity. The pre-BCR related tyrosine kinase BTK is phosphorylated by BCR-ABL1 in Ph+ ALL and other tyrosine kinase oncogenes in Ph-like ALL (Chen et al., 2015). Interestingly, overexpression of a constitutively active form of BTK resulted in strong upregulation of CD25 surface expression. Conversely, the BTK-inhibitor ibrutinib abolished CD25 expression suggesting that feedback control between pre-BCR signaling and CD25 requires BTK. The ability of CD25 to stabilize oncogenic signaling strength in Ph+ ALL and Ph-like ALL was important for leukemia-initiation and development of fatal disease. In the absence of CD25, Il2ra-/- ALL cells showed impaired proliferation and colony formation. Serial transplantation experiments revealed a profound defect of Il2ra-/- ALL cells to initiate leukemia. 100-times more cells were required to cause fatal disease. In addition, CD25 expression mediated drug-resistance in ALL cells: In patient-derived pre-B ALL cells with heterogeneous CD25 expression, vincristine selectively induced apoptosis in CD25Low cells but spared CD25High ALL cells. Combination with an anti-CD25 immunotoxin efficiently eradiated CD25High leukemia cells and sensitized the ALL cell population to treatment with vincristine. To elucidate the mechanism of how CD25 coordinates negative feedback control of pre-BCR signaling or its oncogenic mimics, we focused on its short (13aa) cytoplasmic tail, which includes two phosphorylation sites (S268 and T271) that are known substrates for serine/threonine protein kinase, PKCα. To identify cytoplasmic interaction partners of CD25, we overexpressed a Flag-tagged truncated form of CD25 including a myristoylation signal for constitutive membrane localization, transmembrane domain and cytoplasmic tail. Immunoprecipitation (IP; Flag) followed by 2D mass spectrometry revealed strong interactions of PP2A with cytoplasmic tail of CD25. Western blots showed additional strong interactions of the cytoplasmic tail of CD25 with inhibitory phosphatases PTEN, SHP1 and SHIP1. Importantly, reconstitution of myristoylated CD25 tail but not a mutant construct lacking the serine/threonine motif (S268A/T271A) rescued proliferation and survival defects of Il2ra-/- ALL cells. Conclusion: We identified CD25 as a surface receptor that mediates membrane recruitment of PP2A, PTEN, SHP1 and SHIP1, which balances fluctuations in signaling output from a pre-B cell receptor or its oncogenic mimic in ALL cells (e.g. BCR-ABL1 in Ph+ ALL). We propose that CD25-mediated negative feedback control stabilizes oncogenic tyrosine kinase signaling and mediates drug-resistance in Ph+ ALL and Ph-like ALL cells. Targeted inhibition using CD25-directed immunotoxins may be useful in new approaches to overcome drug-resistance in Ph+ ALL and Ph-like ALL. Disclosures No relevant conflicts of interest to declare.

Description: Chronic graft-versus-host disease (GVHD) in patients with allogeneic hematopoietic cell transplantation (HCT) is an immune-mediated syndrome. Both CD4+ T and B cells are required for disease induction, but details of this interaction have not been fully defined. Other groups have reported that germinal centers are enlarged in mice with chronic GVHD and that GC formation is required for induction of the disease. On the other hand, patients with chronic GVHD have a predominance of IgG1 in the serum with little evidence of IgG somatic hypermutation by one year after HCT, suggesting a lack of GC formation (Suzuki et al: Blood. 1996;87:1873-1880; Glas et al: Blood. 2000; 96:1064-1069). Consistently, we observed that GCs were not detectable in mice with chronic GVHD, including MHC-matched C57BL/6 (H-2b) recipients given LP/J (H-2b) grafts and BALB/c (H-2d) recipients given DBA/2 (H-2d) or B10D2 (H-2d) grafts, and MHC-mismatched BALB/c (H-2d) recipients given C57BL/6 (H-2b) grafts. Furthermore, BALB/c recipients given grafts from BCL6-deficient (BCL6 fl/fl Mb1-Cre) C57BL/6 (H-2b) donors lacking the ability to form GC developed chronic GVHD with similar severity to recipients given control BCL6-sufficient grafts. Consistent with lack of GC formation, recipients of BCL6-deficient grafts had profound reduction of serum IgG2b and IgG2c but moderate reduction of IgG1 as compared to recipients given control BCL6-sufficient grafts. The presence of IgG1 indicates extrafollicular CD4+ T and B interaction, which is Stat-3 dependent. Further experiments showed BALB/c recipients given grafts from Stat3-deficient (Stat-3 fl/fl CD4-Cre) C57BL/6 donors did not develop chronic GVHD. Taken together, these results indicate that GC formation is not required for development of chronic GVHD and that extrafollicular CD4+ T and B interaction can induce chronic GVHD. This work is supported by NIH R01 AI066008 and The Beckman Research Institute Excellence Award (to D. Zeng). Disclosures Forman: Mustang: Research Funding; Amgen: Consultancy. Martin:Neovii: Research Funding; RegImmune: Research Funding; Enlivex: Consultancy; Janssen: Consultancy; Pfizer: Consultancy; Pharmacyclics: Consultancy.

Description: Background and Hypothesis: B-cell identity is determined by a set of B-cell transcription factors including PAX5, IKZF1 and EBF1. However, B-lineage leukemia clones often carry secondary genetic lesions that result in reduced activity or inactivation of these transcription factors. Studying patient samples from clinical trials for B-lineage childhood (P9906; n=187) and adult (MDACC; n=92) leukemia, we found that genetic defects in one or more B-cell transcription factors represent near-obligate lesions in human acute lymphoblastic leukemia (209 of 279 B-lineage ALL cases). While previously of unknown significance, we found that adult ALL cases with known lesions in one or more of these transcription factors had higher activity of lactate dehydrogenase, and phospho-states indicated higher activity of IRS1, PDK1, and AKT, which contribute to glucose uptake. For this reason, we investigated whether B-cell transcription factors set metabolic constraints of oncogenic signaling in leukemia. Results: Reconstitution of wildtype PAX5 and IKZF1 in patient-derived B-lineage ALL cells carrying PAX5 and IKZF1deletions decreased phospho-AKT levels. Furthermore, protein levels of multiple positive regulators involved in glucose uptake and metabolism (including insulin receptor, glucose transporters and hexokinases) were downregulated upon reconstitution of PAX5 and IKZF1. Conversely, protein levels of negative regulators of glucose uptake and metabolism including NR3C1 and TXNIP were upregulated upon reinstatement of PAX5 and IKZF1 function. Analysis of ChIPseq data of human B cells revealed binding of multiple B-cell transcription factors including PAX5, IKZF1 and EBF1 to promoter regions of genes encoding positive regulators of glucose uptake and metabolism. Binding peaks for B-cell transcription factors were also observed at genes that encode negative regulators of glucose uptake and metabolism (NR3C1 and TXNIP). In addition, direct recruitment of PAX5 was confirmed by single-locus quantitative chromatin immunoprecipitation (qChIP). Reconstitution of wildtype PAX5 and IKZF1 in patient-derived B-lineage ALL cells caused depletion from the cell culture in competitive growth assays, in parallel with reduced glucose consumption and depletion of cellular ATP levels. On the contrary, overexpression of dominant-negative PAX5-ETV6 and IK6 in patient-derived B-lineage ALL cells expressing wildtype PAX5 and IKZF1 resulted in a net survival advantage, concomitant with increases in both glycolytic activity and cellular ATP levels. PAX5-mediated impaired survival fitness was significantly rescued by CRISPR-based activation of gene expression of insulin receptor, glucose transporter and hexokinases in patient-derived B-lineage ALL cells. Conversely, CRISPR/Cas9-mediated deletion of NR3C1 and TXNIP largely reversed the effects of PAX5. Finally, reduced survival fitness upon reconstitution of wild-type PAX5 and IKZF1 in patient-derived pre-B ALL cells was mostly rescued by metabolites that can enter the TCA cycle and thus provide ATP. Conclusions: In summary,B-cell-specific restriction of glycolytic energy supply represents a previously unrecognized metabolic barrier against malignant transformation. Our findings suggest a causative link between impaired glucose uptake and metabolism caused by B-cell transcription factors as well as cell death, and the known tumor suppressive function of PAX5 and IKZF1 in B-lineage ALL. Disclosures No relevant conflicts of interest to declare.

Description: Background and Hypothesis: The transcriptional repressor and proto-oncogene BCL6 is a therapeutic target in subtypes of diffuse large B cell lymphoma (DLBCL) and modulates drug-resistance in Philadelphia chromosome-positive acute lymphoblastic leukemia (Ph+ ALL; Duy et al., Nature 2011). BCL6 was shown to be a critical factor that bypasses p53-dependent senescence and thereby enables RAS-driven transformation of mouse embryonic fibroblasts (Shvarts et al., Genes Dev. 2002). Given that ~50% of pediatric ALL cases carry genetic lesions that lead to hyperactivated RAS-ERK signaling (Zhang et la., Blood 2012), we examined the role of BCL6 in RAS-driven pre-B ALL and identified a novel mechanism by which RAS-ERK signaling can mediate BCL6 expression. Results: Using a doxycycline-inducible TetOn- NRASG12D vector system, we found that inducible activation of RAS-ERK signaling strongly upregulated BCL6 expression at both the mRNA (~350-fold) and protein (~50-fold) levels in murine pre-B cells. Increases in BCL6 expression were abrogated upon treatment with a MEK inhibitor (PD325901). In addition, Cre-mediated deletion of Mapk1 suppressed upregulation of BCL6 expression upon imatinib treatment in BCR-ABL1-driven pre-B ALL cells. These findings suggested that elevated expression of BCL6 is a consequence of ERK activation. Previously, we demonstrated that BCL6 expression is negatively regulated by STAT5 in BCR-ABL1 pre-B ALL (Duy et al., Nature 2011). Interestingly, oncogenic NRASG12D inhibited phosphorylation of STAT5-Y694 by activating the inhibitory protein tyrosine phosphatase Ptpn6. Cre-mediated deletion of Ptpn6 induced STAT5 activity. Furthermore, loss of Ptpn6 function abrogated upregulation of BCL6 expression induced by imatinib in BCR-ABL1 pre-B ALL. Taken together, RAS-ERK signaling induces BCL6 expression by suppressing STAT5 activity. To directly test the role of BCL6 in RAS-transformed pre-B ALL, we generated a novel mouse model for inducible Cre-mediated deletion of Bcl6 exons 5-10, flanked by loxP sites. Inducible deletion of Bcl6 in NRASG12D-transformed pre-B ALL cells led to rapid depletion from the cell culture and reduced colony forming ability in vitro. These findings suggested that BCL6 is required for maintenance of fully established RAS-transformed ALL. Notably, we found that initiation of NRASG12D-driven leukemia in vivo depends on BCL6 as NRASG12D ALL failed to give rise to leukemia in the absence of Bcl6 in transplant recipient mice. Studying a diagnostic (KRAS wild-type) and a relapsed (KRASG12V) sample from one pre-B ALL patient revealed increased BCL6 expression in KRASG12V relapsed ALL cells. In addition, selective sensitivity to PD325901 and a retro inverso BCL6 peptide inhibitor (RI-BPI) was observed in KRASG12V relapsed ALL cells. Finally, RI-BPI prolonged overall survival of recipient mice transplanted with KRASG12V relapsed ALL cells in vivo. Conclusions: In summary, we demonstrated a novel mechanism by which oncogenic RAS signaling induces expression of BCL6, and showed that BCL6 is critical for RAS-driven transformation in pre-B ALL. Importantly, ALL clones often acquire drug resistance and activating mutations in the RAS pathway (Bhojwani and Pui, Lancet Oncol. 2013). Our findings suggest that pharmacological inhibition of BCL6 may provide a novel therapeutic avenue to overcome drug-resistance and prevent leukemia relapse after initial remission in RAS-driven ALL. Disclosures Melnick: Janssen: Research Funding.

Description: Background and rationale. Leukemia-initiating cells (LIC) were extensively studied in AML (Bonnet 1997) and CML (Graham 2002), while LIC populations in B-ALL remained elusive. In the absence of a functional definition, targeted approaches for LIC-eradication are not feasible in B-ALL. Some studies suggested immunophenotypes for B-ALL LIC-populations (Cox 2004, Castor 2005, Wang 2007). However, other groups demonstrated that tumor-initiation in B-ALL is not limited to rare populations or developmental hierarchy (Kelly 2007; Le Viseur 2008; Rehe 2013; Aoki 2015). Concept: We hypothesize that self-renewal in the B-cell lineage is induced by positive selection and antigen-receptor (BCR) signaling, i.e. encounter of cognate antigen. Self-renewal at this stage leads to clonal expansion and survival. Unlike stemness in AML and CML, which is determined by a developmental hierarchy, we propose that self-renewal in the B-cell lineage is transient and driven by environmental antigen and the ability of BCRs to bind with high affinity. In B-cell malignancies, positive B-cell selection events, resulting in Lgr5 surface expression, are mimicked by transforming oncogenes (e.g. BCR-ABL1, NRASG12D, MYD88L265P). Results: Combining flow cytometry and genetic approaches, we identified surface expression of the leucine-rich repeat containing G-protein coupled receptor 5 (Lgr5) as new biomarker of positively selected pre-B cells in the bone marrow. Conversely, conditional ablation of Lgr5 during earliest stages of B-cell development resulted in near-complete failure to develop a mature B cell pool (reduced by 2-3 log orders). Lgr5 is a Wnt target gene and an established cancer stem cell marker for epithelial cancers (e.g. colorectal cancer and mammary tumors), however, a role for Lgr5 in normal and malignant hematopoiesis is not known. Importantly, Lgr5 represents a previously unrecognized predictor of poor clinical outcome in children and adults with pre-B ALL, including worse overall survival and higher risks of drug-resistance and relapse. Limiting dilution transplant experiments showed that Lgr5-overexpression increased LIC-frequencies in NSG recipient mice. Inducible activation of Cre in Lgr5fl/fl mouse models for BCR-ABL1- or NRASG12D-driven B-ALL resulted in cell cycle arrest, abolished colony forming capacity and compromised the ability of leukemia cells to initiate fatal disease in NSG transplant recipients. Deletion of Lgr5 in pre-B ALL cells caused massive accumulation of nuclear β-catenin and increased expression of β-catenin target genes. Phosphoproteomic analyses revealed increased levels of β-catenin S675-phosphorylation, which increases β-catenin transcriptional activity (Taurin 2006; Hino 2015). Inducible activation of a gain-of-function mutant of β-catenin revealed that pre-B ALL cells are extremely sensitive to β-catenin activation. Thus, Lgr5 enables positive selection and self-renewal of B-ALL cells by curbing β-catenin activity. Therapeutic implication: To assess Lgr5 surface expression on B-ALL as a target for antibody-drug conjugate (ADC), we treated refractory B-ALL PDX with the Lgr5-MMAE ADC. Single-agent treatment with Lgr5-MMAE significantly reduced B-ALL leukemia burden. Treatment with dexamethasone not only enforced persistent surface expression of Lgr5, but also potentiated efficacy of by Lgr5-MMAE. Conclusion: Unlike self-renewal in myeloid leukemia that is determined by a developmental hierarchy, our results here show that self-renewal in the B-cell lineage is transient and driven by the ability of BCRs to bind antigen with high affinity. Lgr5 is a biomarker of this selection event, critical for the initiation of B-ALL and other B-cell malignancies in transplant recipients. Given that positive B-cell selection events, resulting in Lgr5 surface expression are mimicked by transforming oncogenes, Lgr5 also represents a promising target for ADC therapy for instance Lgr5-MMAE (Genentech). Disclosures No relevant conflicts of interest to declare.

Description: Studying gene expression and clinical outcome data from 136 clinical trials for patients with cancer (~21,000 patients with 26 cancer types), we found CD25 as one of the strongest predictors of poor clinical outcome in patients with B-cell malignancies, but not in other cancer types. This was unexpected because CD25 is known as one of three chains of the IL2 receptor on T-cells and NK-cells. Interleukin-2 (IL2) functions as essential T-cell growth factor. IL2 signals through β- and γ-, but not α-chains (CD25) of its heterotrimeric receptor. CD25-deficiency causes lymphoproliferation and autoimmunity, however, its mechanistic role is unclear. Our experiments based on genetic mouse models and engineered patient-derived B-cell leukemia and lymphoma xenografts revealed that CD25 expressed on B-cells is not an IL2 receptor chain, but in fact binds the B-cell receptor (BCR) to regulate its activity. Suggesting IL2-independent functions, defects in CD25-/-B-cells were not replicated in IL2-deficient mice. CD25 bound the BCR but not IL2Rβ- and IL2Rγ-chains. IL2Rβ- and IL2Rγ-chains can pair with other chains to form receptors for different cytokine-ligands. However, CD25 represents the first example of a cytokine receptor chain that binds to the BCR for negative feedback regulation. Likewise, in T-cells, CD25 had a bifunctional role and either functioned as IL2 receptor chain or as negative feedback regulator of T-cell receptor signaling. CD25-function was regulated by cell-membrane translocation, which required phosphorylation of its cytoplasmic tail at S268 (see schematic, left). In a family with monogenic autoimmunity, a mutation immediately preceding S268 compromised CD25-surface translocation, which was restored by homology-directed repair of the S268 motif. CD25-interactome analyses identified PKCd as critical effector molecule downstream of CD25 to mediate B-cell selection during normal B-cell development and calibrate oncogenic BCR signaling in B-cell tumors. In B-cell malignancies, BCR-dependent survival and proliferation signals are often substituted by oncogenic BCR-mimics (e.g. BCR-ABL1, JAK2, BRAFV600E, LMP2A, CD79B mutations; see schematic, right). Accordingly, we identified CD25 surface-expression as biomarker of oncogenic BCR-signaling and predictor of poor clinical outcomes. CD25-/-B-cell leukemia failed to initiate fatal disease in transplant recipients. Owing to imbalances of oncogenic BCR-signaling and p53-checkpoint activation, CD25-/- B-cell leukemia failed to initiate fatal disease in transplant recipients. In patient-derived xenograft models of drug-resistant B-cell malignancies, treatment with a CD25-specific antibody drug-conjugate (ADCT-301) extended survival of transplant recipients or eradicated disease. These findings identified CD25 as previously unrecognized feedback regulator of oncogenic BCR-signaling and provide a rationale for therapeutic targeting of CD25 in refractory B-cell malignancies. Figure. Figure. Disclosures Forman: Mustang Therapeutics: Other: Licensing Agreement, Patents & Royalties, Research Funding. Weinstock:Genentech/Roche, Monsanto: Consultancy; Novartis: Consultancy, Research Funding; Novartis, Astra Zeneca, Abbvie, Aileron, Surface Oncology, Daiichi Sankyo: Research Funding; Novartis, Dragonfly, Travera, DxTerity, Travera: Consultancy; Travera: Equity Ownership; Astra Zeneca, JAX, Samumed, Regeneron, Sun Pharma, Prescient: Patents & Royalties. Uzel:Novartis: Research Funding.

Description: Background & Hypothesis: B cell receptor (BCR) signaling and oncogenic tyrosine kinases that mimic BCR-signaling in B-lineage leukemia and lymphoma depend on assembly of membrane proximal signaling complexes. Signalosomes in normal BCR- and oncogene (e.g. BCR-ABL1, RAS-pathway lesions) signal transduction are recruited to phospholipid anchors in lipid rafts. The robustness of these complexes depends on cholesterol accumulation in lipid rafts. Here we identified the interferon-induced transmembrane protein IFITM3 as a central regulator of cholesterol in lipid rafts. Results: IFITM3 is mostly localized to endosomal compartments. By antagonizing VAP-A and oxysterol-binding protein 1 (OSBP1), IFITM3 promotes cholesterol accumulation and solidifies the endosomal membrane. This mechanism is particular important in anti-viral immunity, to "trap" intraluminal viral particles for lysosomal degradation. In B-cells, IFITM3 can translocate to the cell membrane and form a complex with the BCR and its co-receptors CD19, CD81 and CD21. While the functional significance of membrane expression of IFITM3 on B-cells was not known, we found that higher IFITM3 mRNA levels at the time of diagnosis represents a strong predictor of poor clinical outcome for children (COG P9906; P=0.006; n=207) and adults (ECOG E2993; P=0.014; n=215) with B-ALL. In addition, higher than median IFITM3 mRNA levels at the time of diagnosis were associated with a higher risk of relapse and positive MRD status at the end of induction chemotherapy in B-ALL and other B-cell malignancies. Interestingly, IFITM3 is a transcriptional target and strongly repressed by IKZF1 (Ikaros) a potent tumor suppressor in B- ALL and high IFITM3 mRNA levels represents a biomarker for patients with IKZF1-deletion. While its membrane-topology can vary in different cell types, we found that IFITM3 functions as a dual-pass transmembrane protein in tight association with CD19 and the Iga and Igb signaling chains of the BCR in B-ALL and B-cell lymphoma cells. To study the function of Ifitm3 in a model for human pre-B ALL, pre-B cells from Ifitm3-/- mice were transformed with BCR-ABL1 or oncogenic NRASG12D. Strikingly, deletion of IFITM3 resulted in destabilization of lipid rafts, loss of CD19 surface expression and loss of PI3K signaling. Ifitm3-/- leukemia cells could not sustain oncogenic signaling from BCR-ABL1 or oncogenic NRASG12D and failed to initiate fatal leukemia in transplant recipient mice. These changes were paralleled by G0/1 cell cycle arrest (P