Search

Cdc42 was originally discovered as a key regulator of bud site assembly and polarity in S. cerevisiae. Recent genetic studies have shown that the function of Cdc42 in regulating cell polarity appears highly conserved from budding yeast to humans. The role of Cdc42 in hematopoietic cell transformation and leukemia progression has been studied in an acute myeloid leukemia model using the MLL-AF9 oncogene-induced transformation and a Cdc42 conditional gene-targeted mouse model. Here we describe the leukemia cell polarity and division symmetry assays in the context of leukemia cell fate determination.

Hematopoietic stem and progenitor cells (HSPCs) reside in a specialized niche that regulates their proliferative capacity and their fate. There is increasing evidence for similar roles of marrow niches on controlling the behavior of leukemic cells; however, whether normal hematopoietic stem cell (HSC) and leukemic cells reside in or functionally compete for the same marrow niche is unclear. We used the mixed lineage leukemia-AF9 (MLL-AF9) murine acute myeloid leukemia (AML) in a competitive repopulation model to investigate whether normal HSPC and leukemic cells functionally compete for the same marrow niches. Irradiated recipient mice were transplanted with fixed numbers of MLL-AF9 cells mixed with increasing doses of normal syngeneic whole bone marrow (WBM) or with purified HSPC (LSK). Survival was significantly increased and leukemic progression was delayed proportional to increasing doses of normal WBM or normal LSK cells in multiple independent experiments, with all doses of WBM or LSK cells studied above the threshold for rapid and complete hematopoietic reconstitution in the absence of leukemia. Confocal microscopy demonstrated nests of either leukemic cells or normal hematopoietic cells but not both in the marrow adjacent to endosteum. Early following transplantation, leukemic cells from animals receiving lower LSK doses were cycling more actively than in those receiving higher doses. These results suggest that normal HSPC and AML cells compete for the same functional niche. Manipulation of the niche could impact on response to antileukemic therapies, and the numbers of normal HSPC could impact on leukemia outcome, informing approaches to cell dose in the context of stem cell transplantation. Stem Cells 2015;33:3635–3642

Thymocyte development is regulated by complex signaling pathways. How these signaling cascades are coordinated remains elusive. RhoA of the Rho family small GTPases plays an important role in actin cytoskeleton organization, cell adhesion, migration, proliferation, and survival. Nonetheless, the physiological function of RhoA in thymocyte development is not clear. By characterizing a conditional gene targeting mouse model bearing T cell deletion of RhoA, we show that RhoA critically regulates thymocyte development by coordinating multiple developmental events. RhoA gene disruption caused a strong developmental block at the pre-TCR checkpoint and during positive selection. Ablation of RhoA led to reduced DNA synthesis in CD4−CD8−, CD4+CD8−, and CD4−CD8+ thymocytes but not in CD4+CD8+ thymocytes. Instead, RhoA-deficient CD4+CD8+ thymocytes showed an impaired mitosis. Furthermore, we found that abrogation of RhoA led to an increased apoptosis in all thymocyte subpopulations. Importantly, we show that the increased apoptosis was resulted from reduced pre-TCR expression and increased production of reactive oxygen species (ROS), which may be because of an enhanced mitochondrial function, as manifested by increased oxidative phosphorylation, glycolysis, mitochondrial membrane potential, and mitochondrial biogenesis in RhoA-deficient thymocytes. Restoration of pre-TCR expression or treatment of RhoA-deficient mice with a ROS scavenger N-acetylcysteine partially restored thymocyte development. These results suggest that RhoA is required for thymocyte development and indicate, to our knowledge, for the first time that fine-tuning of ROS production by RhoA, through a delicate control of metabolic circuit, may contribute to thymopoiesis.

As a central regulator of cell polarity, the activity of CDC42 GTPase is tightly controlled in maintaining normal hematopoietic stem and progenitor cell (HSC/P) functions. We found that transformation of HSC/P to acute myeloid leukemia (AML) is associated with increased CDC42 expression and activity in leukemia cells. In a mouse model of AML, the loss of Cdc42 abrogates MLL-AF9-induced AML development. Furthermore, genetic ablation of CDC42 in both murine and human MLL-AF9 (MA9) cells decreased survival and induced differentiation of the clonogenic leukemia-initiating cells. We show that MLL-AF9 leukemia cells maintain cell polarity in the context of elevated Cdc42-guanosine triphosphate activity, similar to nonmalignant, young HSC/Ps. The loss of Cdc42 resulted in a shift to depolarized AML cells that is associated with a decrease in the frequency of symmetric and asymmetric cell divisions producing daughter cells capable of self-renewal. Importantly, we demonstrate that inducible CDC42 suppression in primary human AML cells blocks leukemia progression in a xenograft model. Thus, CDC42 loss suppresses AML cell polarity and division asymmetry, and CDC42 constitutes a useful target to alter leukemia-initiating cell fate for differentiation therapy.

Despite significant advances in AML therapy over the past decades, 30-40% of all young patients with AML will suffer a relapse, after which achieving long-term disease-free survival remains challenging. In pediatric and young adult patients with AML, curative-intent treatment at relapse involves intensive induction therapy to achieve CR followed by consolidation with HSCT. Historically, initial response rates to salvage therapy were limited. Recently, Children's Oncology Group presented results from a Phase II study of CPX-351 (liposomal cytarabine:daunorubicin at a synergistic ratio) in pediatric patients with AML in first relapse demonstrating an impressive CR/CRp/CRi rate of 81.3%. Separately, we presented data from our first-in-pediatrics CPX-351 Phase I trial showing 48% M1 marrow response in a population consisting of children with multiply relapsed and refractory (R/R) AML patients. We also obtained single cell RNA sequencing on study patients' peripheral blasts before, during, and after CPX-351 treatment (i.e.: Days 0, 1 and 5). Preliminary data showed that CPX-351 differentially impacts p53 targets within leukemic blast clusters over time, with an enrichment for genes regulating apoptosis (e.g.: FAS, BAX) in most clusters. This suggests that these blasts may be primed for apoptosis following CPX-351 treatment. Anti-apoptosis signaling in leukemias is an established mechanism of chemoresistance. Venetoclax is an orally available, small molecule inhibitor of anti-apoptotic protein BCL-2, which is frequently overexpressed in leukemias. Based on published preclinical data and single cell RNA sequencing data from our CPX-351 trial, we developed a Phase I study to investigate the combination of CPX-351 and venetoclax in the treatment of young patients with R/R acute leukemias. The V2 trial is a single-institution Phase I pilot study (NCT03826992) designed to test the safety and tolerability of combining venetoclax with CPX-351 in patients ages 1-39 years for the treatment of R/R acute leukemias. Inclusion diagnoses are based on preclinical sensitivities and include AML, MPAL, or AUL in first or greater relapse, or ETP-ALL, MLL-rearranged ALL, or T-ALL in second or greater relapse. Patients must have measurable disease. Exclusion criteria include CNS status 3, bone marrow failure syndromes, and prior anthracycline and mediastinal radiation exposures above acceptable cardiotoxicity risk thresholds. Subjects receive a single course of study therapy consisting of daily venetoclax and CPX-351 at the FDA approved dose for adults on Days 1, 3, and 5 (Fig. 1) with the goal of CRMRD- to allow for off-study HSCT. The venetoclax dose exploration phase begins with a starting dose of 400 mg (or the allometrically-scaled equivalent) for 21 days with a 3-day ramp to target dose. The dose exploration phase proceeds on a rolling 6 design and will require 4-12 patients to determine RP2D. Provided a tolerable RP2D is found, a dose expansion phase will open. Primary endpoints are determination of RP2D and description of the safety profile including time to hematologic recovery. Secondary endpoints are estimations of CR/CRp/CRi +/- MRD- within the context of a phase I study, and evaluation of cancer therapy-related cardiac dysfunction (CTRCD) rates. Correlative studies include evaluation of venetoclax PK with concomitant CPX-351 using modeling & simulation with a combined limited sampling and pediatric opportunistic PK sampling (POPS) strategy, and exploration of the correlations between BCL-2 family expression and pediatric tumor genetics with treatment response. Figure 1 Disclosures O'Brien: BTG: Research Funding; Pfizer: Research Funding; Celgene: Research Funding; AbbVie: Research Funding; Amgen: Research Funding; BMS: Research Funding. Phillips:Novartis: Membership on an entity's Board of Directors or advisory committees. Vinks:Myriad Genetics: Consultancy, Patents & Royalties. Perentesis:Kurome Therapeutics: Consultancy. Absalon:Jazz Pharmaceuticals: Other: CPX-351 for clinical trial . OffLabel Disclosure: This clinical trial involves the off-label use of CPX-351 and venetoclax for relapsed/refractory pediatric and young adult leukemias.

Aberrant activation of Rho guanine nucleotide exchange factors (RhoGEFs) is a chief mechanism driving abnormal activation of their RhoGTPase targets in cancer. Thus, small molecule inhibitor of RhoGEF activities can be used as a drug lead to treat leukemia and other malignancies. We have identified an active small molecule, IODVA1, in several xenograft mouse models of cancer including Ras-driven cancers. Here, we used cellular and mouse models of Ph+(BCR-ABL1) B-ALL to identify Vav3, the multi-domain tyrosine phosphorylation-dependent RacGEF as the target of the small molecule IODVA1. IODVA1 specifically reduces the proliferation and survival of p190-BCR-ABL but not of empty vector expressing human CD34+ blood cells. IODVA1 binds tightly to recombinant Vav3 (Kd= 400 nM) but not to Rac1 or to the RhoGEF LARG. In Ba/F3 cells expressing p190-BCR-ABL, IODVA1 inhibits Rac activation and signaling within minutes of exposure. The decrease in Rac activity is not due to the activation of p50GAP or RhoGDI1 and is accompanied by a decrease in the activity of the pro-survival effectors PAK, JNK, and 4EBP and an increase in pro-apoptotic BAD activity. IODVA1prevents leukemia-related death and eliminates the leukemia burden in a BCR-ABL-induced murine model with no apparent toxicity. It eradicates leukemic propagating activity assessed by serial transplantation. Most importantly, IODVA1 increases the survival of a mouse model of TKI-resistant p210-BCR-ABL1(T315I) B-ALL better than ABL1-TKI imatinib and eliminates leukemic burden. p210-T315 leukemic mice survive 45 days after treatment has ended. Vav3- and Rac1/Rac2-deficient leukemic cells do not respond to IODVA1 in colony formation assay consistent with Vav3 being IODVA1's target. A mouse model of BCR-ABL leukemia deficient in Vav3 also does not respond to IODVA1. Cells from PDX models representing pediatric ALL patients including Ph+ (BCR-ABL(T315I) were found to be highly sensitive to IODVA1 ex vivo. Based on the described mechanism of action, we propose that IODVA1 is an allosteric inhibitor of Vav3 that holds promise as an anti-leukemic agent. Disclosures Cancelas: Hemanext: Consultancy, Research Funding; Fresenius-Kabi: Research Funding; Cerus Co.: Research Funding; TerumoBCT: Consultancy, Research Funding; Velico: Consultancy, Research Funding; Macopharma Inc: Research Funding; Cytosorbents: Research Funding; Cellphire: Research Funding.

Background: Resistance to chemotherapy-induced apoptotic death is a major mechanism responsible for the failure of AML therapies. Levels of anti-apoptotic proteins BCL2 and MCL1 are increased in relapsed AML samples. Venetoclax, a BH3 mimetic that binds to BCL-2, was recently granted accelerated approval for use in AML in combination with cytotoxic chemotherapy. However, MCL-1 and BCL-XL are known pathways of resistance to BCL-2 directed therapy. AMG-176 is a first-in-class MCL-1 specific inhibitor that induces specific and significant apoptosis-mediated toxicity against leukemia cell lines, tumor xenograft models, and primary patient samples. We compared anti-leukemic synergy of Venetoclax and AMG-176 together and individually with cytotoxic chemotherapy used in AML. We sought to identify optimal combinations for efficacy against varying molecular signatures of AML. Methods: To examine the potential of pairing Venetoclax with AMG-176 we performed a series of in vitro 8x8 dose response matrices on a panel of 11 leukemias spanning AML and Acute Lymphocytic Leukemia (ALL) (Table). This included 3 relapsed pediatric Patient Derived Xenograft (PDX) AML lines propagated in the presence of cytokines to better model inflammation, stem cell function, and niche derived support of leukemic blasts. An additional 8 established cell lines representing key AML & ALL molecular classes were employed. We used the Delta Bliss (DB) method for scoring matrices and report the DBsum, the summation of all of the synergistic(-) and antagonistic(+) well scores in an 8x8 matrix. Drugs tested in combinations with Venetoclax and/or AMG-176 included cytarabine and daunorubicin. Results: The Figure shows average DBSum for the Venetoclax+AMG-176 combination was -3.41 in the 8 commercial lines, whereas combining the anthracycline Daunorubicin+AMG-176 combination yielded a DBsum score of -0.34, with the score being skewed by a marked antagonism observed with the t(8;21) AML line Kasumi-1 with a DBSum score of 2.13. The 3 relapsed pediatric PDX derived AML lines continued the trend of strong synergy with the Venetoclax+AMG-176 combination with an average score of -3.68 with no observed antagonistic interactions. For context, a synergistic DBSum score of -2.33 would be considered noteworthy from a combination dataset of over >4000 discreet combinations. The Venetoclax+AMG-176 combination exhibited the highest degree of synergy broadly across AML and ALL molecular subgroups. Different patterns of cytotoxicity of combination of chemotherapy with BCL2 or MCL1 inhibition were observed in leukemias with varying molecular backgrounds and histology, including patterns of strong synergy, additivity or antagonism. However, leukemias carrying KMT2A rearrangements exhibited strong synergy or additivity of BCL2 or MCL1 combinations with either daunorubicin or cytarabine, regardless of histology. Conclusions: We identified broad synergy in high risk genetic subtypes of AML and ALL with the combination of BCL-2 and MCL-1 inhibition, and the effect was superior to either agent combined with chemotherapy individually. We identified potential genetic signatures associated with response to chemotherapy cytotoxicity with either Venetoclax or AMG-176. In vivo experiments are in progress to evaluate the efficacy of this combination and better determine genetic subtypes who would most benefit from this treatment approach. Figure Disclosures Perentesis: Kurome Therapeutics: Consultancy.

Therapies for pediatric acute myeloid leukemia (AML) remain unsatisfactory and generally do not incorporate molecularly-targeted agents aside from FLT3 inhibitors outside of the relapse setting. Patient-derived xenograft (PDX) models of AML are increasingly accessible for the preclinical evaluation of targeted therapies, though the degree to which these systems recapitulate the disease state as found in patients has not been well defined for AML. Gene expression profiling of patient blasts has been successfully used to discriminate distinct subtypes of AML, to uncover sub-type specific vulnerabilities, and to predict response to therapy and outcomes. We sought to systematically examine PDX models of pediatric AML for their ability to replicate global gene expression patterns and preserve mutational signatures found in patients. In addition, we conducted in-depth bioinformatic analyses of samples with cryptic CBA2T3-GLIS2 fusion generated by the inv(16)(p13.3q24.3) for identification of potential novel targeted therapies. We performed detailed analyses of RNA sequencing data from a diverse series of 24 pediatric AML PDX models established from samples obtained from patients with relapse and refractory disease. Initially we compared our PDX data against 49 selected relapse and refractory patient sample data files found in the NCI TARGET dataset of pediatric AML. When applying unsupervised hierarchical clustering to the PDX samples, we found that clustering was associated with MLL status. Clustering of the combined sets of samples by MLL status showed integration of samples according to mutation profile, regardless of data source (PDX or patient). The expression levels of all detectable transcripts were highly conserved between PDX and patient MLL-r samples. Separate analysis of each dataset yielded MLL specific gene lists that included a subset of overlapping genes which may point to a unique relapse and refractory pediatric MLL-r signature. This list contains several interesting new targets for further study. A subset of 12 PDX models were compared directly to the matched patient sample from which they were established. This analysis revealed strong similarity, with each PDX most closely related to its matched patient sample, suggesting retention of sample-specific gene expression in immune deficient mice. We set up our PDX models in NSG mice with transgenic expression of human myelo-supportive cytokines SCF, GM-CSF, and IL-3 in order to promote the most efficient and robust engraftment of precious patient material. In order to detect any skewing effects due to the host mouse strain, we compared NSGS PDX RNA sequencing data to 10 matched NSG PDX models. This comparison revealed consistent differences in only 9 transcripts, which were almost entirely related to increased JAK/STAT signaling and macrophage activation pathways in NSGS mice relative to NSG mice. Interestingly, during this analysis we observed a distinct PCA-driven clustering of a pair of PDX samples with previously clinically unidentified driver mutations. Reanalysis of the RNA sequencing data revealed evidence of a cryptic GLIS2 rearrangement (found in ~1% of pediatric AML cases) as the driver mutation, which was subsequently confirmed by RT-PCR in both samples. The unique CBFA2T3/GLIS2 RNA signature was mined to guide the composition of a focused 75-molecule in vitro drug screen against ex vivo PDX samples with an emphasis on the SHH, WNT, and BCL2 pathways. This screen identified the Wnt-C59 PORCN inhibitor as having specific activity against CBFA2T3/GLIS2+ AMLs. Further testing of C-59 in combinatorial studies revealed enhanced effects with the addition of the BCL2 inhibitor, venetoclax. In vivo experiments are currently underway to determine the pre-clinical efficacy of this novel combination. In summary, we found highly significant fidelity of gene expression in PDX models of relapse and refractory pediatric AML. Analysis of this dataset has led to several insights, including potential targeted therapies, highlighting how this system could be a valuable tool for discovery of novel targeted therapies, especially for very rare, distinct subtypes of disease. Disclosures Perentesis: Kurome Therapeutics: Consultancy.

Hematopoietic stem cells (HSC) mainly engage glycolysis while leukemia stem cells (LSC), such as in acute myeloid leukemia (AML), heavily rely on mitochondrial (Mt) respiration (i.e. oxidative phosphorylation, OxPhos) to fuel energy. Growing evidence suggests this metabolic reprogramming confers therapeutic vulnerabilities in AML. BCL2 is overexpressed in LSC from AML patients, while BCL2 inhibitors, such as venetoclax (VEN), have been shown to suppress OxPhos in LSC and to eradicate LSC. In clinical practice, half of the patients fail to respond to VEN. VEN is a costly medication and patients who are resistant to VEN forgo alternative treatment at that time. Prediction of the response to VEN and strategies to circumvent resistance are urgently needed. BCL2 is shown to increase Mt Ca2+ levels ([Ca2+]m), which enhance OxPhos through activation of the Ca2+-sensitive dehydrogenases within the tricarboxylic acid (TCA) cycle. G protein-coupled receptor 68 (GPR68) is a proton sensor, activating phospholipase C that leads to releasing of Ca2+ from the endoplasmic reticulum (ER) to the cytosol and elevation of cytosolic Ca2+ levels ([Ca2+]c). This prompted us to examine the cooperative effect of GPR68 and BCL2 on the Ca2+/OxPhos pathway in AML, and particularly in LSC. Expression of leukemic oncogenes (i.e. MLL-AF9 and HRASG12D) in mouse hematopoietic stem and progenitor cells, such as Lineage-Sca-1+cKit+ (LSK) cells or granulocyte-monocyte progenitor cells, promote leukemogenesis as evidenced by serial colony formation in vitro and leukemia development in vivo. We found significantly reduced colonies in oncogene-expressing LSK cells from Gpr68 knockout mice compared to wild type mice. Deletion of Gpr68 reduced [Ca2+]c in oncogene-expressing LSK cells. We next examined the function of GPR68 in human AML cell lines. Knockdown of GPR68 with shRNA reduced cell growth and colony formation, and induced apoptosis in AML cells. Knockdown of GPR68 also reduced [Ca2+]c and Mt membrane potential (Δψm) in AML cells, indicating reduced Mt OxPhos. These results suggest that GPR68 regulates the Ca2+/OxPhos pathway in AML cells. We next examined the cooperative effect of GPR68 and BCL2 by jointly inhibiting their activities with pharmacological agents (i.e. GPR68 antibody and VEN, respectively) in AML cell lines and AML patient-derived xenograft models. Of note, the expression of GPR68 was positively correlated with the sensitivity to VEN in AML cells. For AML cells that were resistant to VEN, GPR68 antibody but not an unrelated antibody increased the sensitivity to VEN by enhancing apoptosis, indicating that GPR68 and BCL2 co-regulate AML cell survival. We next examined the mechanism of this synthetic lethality by measuring cellular respiration. Single treatment with VEN reduced Δψm, ATP production and O2 consumption in AML cells. Cotreatment with VEN and GPR68 antibody further reduced Δψm, ATP production and O2 consumption in AML cells, indicating that GPR68 and BCL2 co-regulate Mt OxPhos. Given that GPR68 releases Ca2+ from ER to cytosol, while BCL2 maintains [Ca2+]m by inhibiting its extrusion from Mt, we hypothesize that the GPR68/BCL2 axis relocates Ca2+ from ER to Mt. As expected, treatment with GPR68 antibody reduced [Ca2+]c. Cotreatment with VEN and GPR68 antibody increased [Ca2+]c, indicating enhanced extrusion of Ca2+ from Mt to cytosol by VEN. Consistently, cotreatment with VEN and GPR68 antibody reduced the activity of isocitrate dehydrogenase, the rate limiting enzyme in the TCA cycle, in AML cells. These results indicate that GPR68 and BCL2 co-regulate the Ca2+/OxPhos pathway in AML cells and that co-inhibition of GPR68 and BCL2 may enhance lethality and overcome VEN resistance. In summary, our study suggests that the GPR68/BCL2 axis co-regulates AML cell survival by relocating Ca2+ from ER to Mt thus enhancing Mt OxPhos, and that disruption of the GPR68/BCL2 axis provides a novel therapeutic strategy to overcome resistance to VEN. Disclosures No relevant conflicts of interest to declare.

AML1-ETO (AE) is a fusion product of translocation (8;21) that accounts for 40% of M2 type acute myeloid leukemia (AML). In addition to its role in promoting preleukemic hematopoietic cell self-renewal, AE represses DNA repair genes, which leads to DNA damage and increased mutation frequency. Although this latter function may promote leukemogenesis, concurrent p53 activation also leads to an increased baseline apoptotic rate. It is unclear how AE expression is able to counterbalance this intrinsic apoptotic conditioning by p53 to promote survival and self-renewal. In this report, we show that Bcl-xL is up-regulated in AE cells and plays an essential role in their survival and self-renewal. Further investigation revealed that Bcl-xL expression is regulated by thrombopoietin (THPO)/MPL-signaling induced by AE expression. THPO/MPL-signaling also controls cell cycle reentry and mediates AE-induced self-renewal. Analysis of primary AML patient samples revealed a correlation between MPL and Bcl-xL expression specifically in t(8;21) blasts. Taken together, we propose that survival signaling through Bcl-xL is a critical and intrinsic component of a broader self-renewal signaling pathway downstream of AML1-ETO–induced MPL.

Background: CPX-351 (Vyxeos®) is a liposomal nanoparticle of cytarabine and daunorubicin in a 5:1 molar ratio currently FDA approved for treatment of adults with newly-diagnosed therapy-related acute myeloid leukemia (AML) or AML with myelodysplasia-related changes. The agent also has potent activity against acute lymphoblastic leukemia (ALL) in murine xenograft models [H. Carol et al. Pediatr Blood Cancer 2015;62: 65-71]. We conducted the first phase I/pilot study to evaluate the safety and describe the efficacy of CPX-351 in children and adolescents with relapsed or refractory (R/R) acute leukemia (NCT01943682). Methods: In the Dose Exploration Phase participants age 12 months to 21 years were eligible to receive a single course of CPX-351 administered once daily on Days 1, 3, and 5. Enrollment was initiated at the recommended dose for study in adults (Dose Level 1: 100 Units/m2/dose= cytarabine 100 mg/m2 and daunorubicin 44 mg/m2 per dose) with provisions to explore either a single dose escalation to 134 Units/m2/dose (Dose Level 2) or de-escalation to 67 Units/m2/dose. Response was determined after Day 28. The upper age for eligibility was increased to 30 years and an additional 18 subjects were enrolled into the Expanded Phase of the study. Results: Nine children and adolescents with R/R AML (n=6) or ALL (n=3) were enrolled and evaluable for toxicity in the Dose Exploration Phase. All 9 experienced Grade 3 or 4 hematologic toxicity including 8 with fever with neutropenia or infection. Dose Limiting Toxicity (DLT) was not observed in the 4 patients enrolled at Level 1 but Grade 3 pain/headache lasting greater than 48 hours did occur in 2 of the 5 patients treated at Dose Level 2. One patient with AML was unevaluable for response in this phase due to removal from protocol therapy to receive bone marrow transplantation while in aplasia. All others were evaluable for response. One patient with ALL had stable disease, 2 with AML achieved a complete remission (CR), and 1 with AML achieved a CR in blood and marrow but developed low level disease in the cerebral spinal fluid (CSF). One patient with mixed phenotype acute leukemia and 17 with R/R AML were subsequently enrolled into the Expanded Phase of the study to receive CPX-351 at Dose Level 1. CPX-351 treatment was well tolerated in this population with a similar toxicity profile observed in the Dose Exploration Phase. Dose limiting prolonged aplasia occurred in one patient but no other DLTs were observed. The response has yet to be determined in one patient. To date, seven patients with AML (41%) achieved a CR or complete remission with incomplete blood recovery (CRi). Another patient achieved a CR in marrow but with low level disease detected in the CSF similar to the patient described in the Dose Exploration phase. The remaining 9 patients failed to achieve a clinical response. Pharmacokinetic (PK) parameters of total cytarabine and daunorubicin were analyzed for 21 of these pediatric patients and the results were similar to those observed in adults. However, due to the small sample size of each age group, these interim PK results will need to be validated as data in pediatric patients become available from future clinical studies. Conclusions: CPX-351 100 Units/m2/dose on Days 1, 3, and 5 was well tolerated in this pediatric population and has promising activity in children and young adults with R/R AML. Additional CNS directed therapy is recommended for pediatric patients at high risk for failure in the CNS. Disclosures Phillips: Novartis AG: Consultancy. An:Jazz Pharmaceuticals: Employment.

Each case of Acute Myeloid Leukemia (AML) represents a unique ecosystem. Despite recent advances in our understanding of the genetic landscape of AML, this information remains insufficient to accurately match patients with targeted therapies. While pediatric and adult AML share phenotypic similarities, pediatric AML represents a genetically distinct disease from adult AML, and will benefit from independent genomic studies and novel therapeutic strategies. Real-time ex-vivo functional screening can identify mechanisms underpinning drug response and diversity between tumors, aiding in patient stratification. We established an in vitro drug screening system that incorporates cytokine signaling to better model inflammation, stem cell function, and niche derived support of leukemic blasts. This approach provided insight into variability between patients who currently would be placed on the same therapeutic regimen. Samples were acquired from 12 pediatric AML patients, after informed consent was obtained. Samples were enriched for blasts, and cultured in the presence of SCF, TPO, FLT3-L, IL-3, and IL-6 (KTF36). A panel of 38 drugs was selected from a larger screen of 1839 compounds done on commercially available hematological malignancy cell lines. Drugs included standard chemotherapy agents used in AML and drugs currently under clinical development. Cells were exposed to drugs for 72hrs. An MTS assay was performed and results reported as % of viable cells remaining, after normalization to vehicle control wells. Targeted DNA NGS sequencing of 406 genes, 31 introns, and RNA sequencing of 265 genes was performed for genetic characterization. In vitro drug screening revealed variations in drug sensitivity between samples and revealed time ex-vivo and cytokine milieu to be important factors affecting response of the same cells to the same drugs. Engraftment of long term cultures into immunodeficient mice produced aggressive disease in all cases, indicating robust support of stem cell function via addition of KTF36 cytokines. When possible, clinical response to therapy was compared with in vitro response to the same drugs. This screening approach highlighted well established agents that showed significant activity in highly refractory disease providing rationale for further clinical trial development. An unsupervised clustering showed drug sensitivity primarily correlated with the presence of MLL-X fusion, NRAS/KRAS and PTPN11 alterations. Linear Regression with interaction effects showed drug sensitivity/resistance to be highly selective for single/signature of specific molecular alterations. Figure. Figure. Disclosures No relevant conflicts of interest to declare.

Despite the improvement of chemotherapy and targeted therapy, drug resistance still remains a challenge for long term disease free survival in aggressive leukemia patients. Recently, enhanced glycolysis is observed in acute myeloid leukemia (AML), and in association with poor clinical outcomes and chemoresistance. The byproducts of glycolysis include lactate and protons (H+), which contribute to intracellular acidosis. The extrusion of protons further results in extracellular acidosis. A group of G protein-coupled receptors (GPCRs), including GPR4, GPR65 (TDAG8), GPR68 (OGR1) and GPR132 (G2A), have been demonstrated to respond to extracellular acidosis, resulting in activation of downstream signaling pathways that regulate pleotropic cellular processes. However, it remains unclear whether these proton-sensing GPCRs contribute to the etiology of AML. Here, we performed genomic examination of leukemia (via cBioPortal). Among 660 leukemia patients, only one patient exhibited deletion of GPR132. Other than this single case, we found no genetic mutations or cytogenetic abnormalities pertaining to proton-sensing GPCRs. Examination of transcripts of these proton-sensing GPCRs revealed that GPR68 was upregulated in both pediatric and adult AML. AML patients with higher levels of GPR68 were associated with shorter overall survival. To understand the function of GPR68 in AML, we knocked down GPR68 in AML cell lines with shRNA targeting GPR68 (shGPR68). GPR68 knockdown markedly induced apoptosis, and reduced colony formation and proliferation in AML cells. This result indicates that myeloid malignancies acquire a dependency on GPR68 function. In response to extracellular H+ or overexpression, GPR68 activates Ca2+ pathway. To determine the molecular mechanism by which GPR68 overexpression supports leukemia cell growth and survival, we examined the intracellular Ca2+ levels (i.e. [Ca2+]i) in primary AML samples. Compared with CD34+ normal hematopoietic cells, all primary AML specimens tested exhibited increased [Ca2+]i, consistent with GPR68 overexpression in AML cells. Meanwhile, shGPR68 reduced [Ca2+]i in all AML cell lines tested, indicating that overexpressed GPR68 activates the Ca2+ pathway in AML. Given that enhanced glycolysis leads to extracellular acidosis, we tested whether glycolysis-mediated local acidosis could also explain enhanced GPR68 activation in AML. Indeed, inhibition of glycolysis by 2-deoxyglucose (2-DG) reduced [Ca2+]i in most of the AML cell lines tested, indicating that glycolysis is likely responsible for enhanced GPR68 activation in AML as well. Next, we attempted to identify the Ca2+-dependent molecular mechanism that mediates the prosurvival effects due to GPR68 activation. We screened a series of pharmacological inhibitors for their efficacy in reducing cell growth and inducing apoptosis. Among the inhibitors tested, only a calcineurin (CaN) inhibitor, Cyclosporine, dramatically reduced cell growth and induced apoptosis in AML cells. This finding raises the possibility that GPR68 promotes AML cell survival through activating the Gq/11/Ca2+/CaN pathway. In summary, we find that the myeloid malignancies acquire a dependency on GPR68 signaling pathway, and inhibition of GPR68 might provide a novel therapeutic strategy for AML, especially in those developing chemoresistance. Disclosures No relevant conflicts of interest to declare.

Macrophage Activation Syndrome (MAS) is a form of secondary hemophagocytic lymphohistiocytosis (HLH) that frequently complicates various rheumatoid disorders. Similar to HLH, these disorders present with chronic inflammation and immune dysfunction. Most typically, an identifiable infection triggers an immune response that is either ineffective due to genetic inactivating mutations that affect lymphocyte cytotoxicity, as in primary HLH, or uncontrolled and characterized by expansion of activated T and macrophage cells as in the case of MAS. While lymphocyte function is generally thought to be a crucial aspect of the disease state, abnormal myeloid cell activation by genetic activation of inflammasomes or repeated TLR stimulation was recently shown to result in MAS. These studies have implicated several cytokines, including SCF, IL-3, and GM-CSF in the disease process. A lymphocyte-independent model could be useful to isolate the myeloid contribution to MAS and would allow the testing of additional therapeutic strategies that target the effector cells directly. Such novel therapies are needed to address refractory cases that often prove fatal. Transgenic expression of key myelosupportive human cytokines in immune-deficient mice corrects for the lack of cross species activities of SCF, IL-3, and GM-CSF. When engrafted with human umbilical cord blood (UCB), these triple transgenic mice produce BM and spleen grafts with much higher myeloid composition, relative to non-transgenontrols. After engraftment with UCB, these mice develop severe, fatal MAS characterized by a progressive drop in rbc numbers, increased reticulocyte counts, decreased rbc half-life, progressive cytopenias and evidence of chronic inflammation. The BM becomes strikingly hypocellular and spleens are significantly enlarged with evidence of extramedullary hematopoiesis. Activated macrophages engaged in hemophagocytosis are readily detectable in the spleen and liver. Mice also show elevated body temperature and sCD25, indicating the presence of activated lymphocytes. Treatment of mice with the CD52 inhibitor alemtuzumab eliminated the human graft and resolved MAS, implicating a human cell and demonstrating the reversibility of the phenotype. However, there was no response to lymphocyte suppressive therapies such as steroids, intravenous immunoglobulin, or antibody-mediated ablation of human B and T cells, demonstrating a lymphocyte independent mechanism of action. In contrast, elimination of human myeloid cells using gemtuzumab ozogamicin (anti-CD33) completely reversed the disease, demonstrating the utility of targeting the effector macrophage cells directly in this model. Using a multiplex ELISA assay, TNF alpha and IFN gamma were low to absent in affected mice, however, IL-6, IL-10, MIP-1 alpha, and MIP-1 beta were consistently upregulated. IL-6, MIP-1 alpha, and MIP-1 beta were lost after gemtuzumab ozogamicin therapy, but not after combined B/T cell ablation with rituximab and OKT-3. Interestingly, IL-10 and sCD25 was suppressed by either treatment. Since the IL-6R antibody tocilizumab is currently being evaluated in a clinical trial for HLH, we examined the efficacy of tocilizumab injections in our model and found delayed progression of anemia and a prolonged lifespan for affected mice. This novel model of MAS provides a new tool for the investigation of the mechanisms driving MAS and for the testing of myeloid-directed therapies in a human cell model. Disclosures No relevant conflicts of interest to declare.

Leukemic stem cells (LSCs) are capable of limitless self-renewal and indefinitely propagating leukemia. Eradication of LSCs is the ultimate goal of treating acute myeloid leukemia (AML). Using a mouse model of AML induced by the MLL-fusion protein MLL-AF9, we recently showed that the combined loss of Runx1/Cbfb inhibited the development of leukemia in vivo (Goyama S…Mulloy JC. Transcription factor RUNX1 promotes survival of acute myeloid leukemia cells. Journal of Clinical Investigation 123(9): 3876-3888, 2013). However, LSC-enriched cells with immature surface phenotype (cKit+Gr1-) remained viable in Runx1/Cbfb-deleted MLL-AF9 cells, indicating that RUNX targeting may not eradicate the most immature LSCs. Gene expression analyses of Runx1/Cbfb-deleted MLL-AF9 cells revealed the upregulation of thrombin pathway genes including a thrombin-activatable receptor PAR-1. Interestingly, both overexpression and knockout of PAR-1 inhibit leukemogenesis but do so through distinct mechanisms. Similar to the effect of Runx1/Cbfb-depletion, PAR-1 overexpression induced p21 expression and attenuated proliferation in MLL-AF9 cells. To our surprise, PAR-1-deficiency also prevented leukemia development induced by a small number of MLL-AF9 LSCs in vivo. Re-expression of PAR-1 in PAR-1-deficient cells combined with a limiting-dilution transplantation assay demonstrated the cell-dose dependent role of PAR-1 in MLL-AF9 leukemia: PAR-1 inhibited rapid leukemic proliferation when there are a large number of LSCs, while a small numbers of LSCs required PAR-1 for their growth. Mechanistically, PAR-1 increased adhering properties of MLL-AF9 cells and promoted their engraftment to bone marrow. PAR-1-deficiency also reduced leukemogenicity of AML1-ETO-induced leukemia. Together, these data reveal a multifaceted role for PAR-1 in leukemogenesis, and highlight this receptor as a potential target to eradicate primitive LSCs in AML. Disclosures No relevant conflicts of interest to declare.

As acute myeloid leukemia (AML) xenograft models improve, the potential for using them to evaluate novel therapeutic strategies becomes more appealing. Currently, there is little information on using standard chemotherapy regimens in AML xenografts. Here we have characterized the immunodeficient mouse response to combined Ara-C (cytarabine) and doxorubicin treatment. We observed significant toxicity associated with doxorubicin that required optimization of the route of injection as well as the maximum-tolerated dose for immunodeficient strains. Mice treated with an optimized 5-day induction protocol showed transient weight loss, short-term reduction of peripheral blood cell and platelet counts, and slight anemia. Considerable cytotoxicity was observed in the bone marrow (BM), with primitive LSK cells having a significant survival advantage relative to more mature cells, consistent with the idea of chemotherapy targeting actively growing cells. Treated leukemic mice demonstrated reduced disease burden and increased survival, demonstrating efficacy. AML cells showed significantly increased sensitivity to doxorubicin-containing therapy compared with murine BM cells. Although early treatment could result in some cures, mice with significant leukemia grafts were not cured by using induction therapy alone. Overall, the data show that this model system is useful for the evaluation of novel chemotherapies in combination with standard induction therapy.

The Rac family of small Rho GTPases coordinates diverse cellular functions in hematopoietic cells including adhesion, migration, cytoskeleton rearrangements, gene transcription, proliferation, and survival. The integrity of Rac signaling has also been found to critically regulate cellular functions in the initiation and maintenance of hematopoietic malignancies. Using an in vivo gene targeting approach, we demonstrate that Rac2, but not Rac1, is critical to the initiation of acute myeloid leukemia in a retroviral expression model of MLL-AF9 leukemogenesis. However, loss of either Rac1 or Rac2 is sufficient to impair survival and growth of the transformed MLL-AF9 leukemia. Rac2 is known to positively regulate expression of Bcl-2 family proteins toward a prosurvival balance. We demonstrate that disruption of downstream survival signaling through antiapoptotic Bcl-2 proteins is implicated in mediating the effects of Rac2 deficiency in MLL-AF9 leukemia. Indeed, overexpression of Bcl-xL is able to rescue the effects of Rac2 deficiency and MLL-AF9 cells are exquisitely sensitive to direct inhibition of Bcl-2 family proteins by the BH3-mimetic, ABT-737. Furthermore, concurrent exposure to NSC23766, a small-molecule inhibitor of Rac activation, increases the apoptotic effect of ABT-737, indicating the Rac/Bcl-2 survival pathway may be targeted synergistically.

One key limitation of hematopoietic xenograft mouse models is the relative lack of effective myelopoiesis compared with human hematopoiesis. NSGS mice expressing human myelo-supportive cytokines allow for improved myeloid cell output upon human hematopoietic cell transfer. Surprisingly, we found that NSGS mice developed a severe, eventually fatal anemia upon humanization with umbilical cord blood (UCB) CD34+ cells. Anemia was characterized by a consistent, progressive drop in red blood cell numbers and hemoglobin that was correlated with higher MCV and an increased reticulocyte production at early time points. Eventually total white blood cell and platelet counts were also affected and reticulocyte numbers declined to or below normal. These effects were only apparent in mice with human grafts, and were also not observed in NSG mice with human grafts. Bone marrow cellularity was 40% of control humanized NSG mice and spleens were significantly larger, implying bone marrow dysfunction and extra medullary hematopoiesis. Bone marrow sections from anemic mice clearly showed a general hypocellularity with abundant fatty tissue filling the majority of the intramedullary space. We were able to recover red blood cell counts and hemoglobin to some degree upon transfusion of non-humanized mouse blood, however, the effect lasted only 7 days. CFSE labeling of anemic blood allowed us to detect a decreased half-life of red blood cells of 20 days compared to 30 days for humanized NSG mice. Treatment of mice with Campath, a CD52 antibody, completely reversed the phenotype at the expense of the entire human graft, demonstrating clearly that this was a graft-mediated process. However, specific ablation of B and T cells with rituximab and OKT3 did not alter the phenotype, even when administered soon after CD34+ cell transplant to suppress development of lymphocytes. Additionally, dexamethasone and IV/IG were ineffective in reversing the disease. These data, coupled with negative findings from tests to detect mouse reactive human antibodies, point towards myeloid cell involvement. Indeed, further experiments revealed that specific ablation of human myeloid cells effectively reversed the anemia. In summary, these data highlight a potential pitfall for effective myelopoiesis in the xenograft setting. More broadly, these findings demonstrate the possibility for myeloid driven bone marrow dysfunction and may warrant investigation of myeloid cells in cases of idiopathic aplastic anemia that fail to respond to therapies directed at autoimmune lymphocyte activity. Disclosures: No relevant conflicts of interest to declare.

The leukemic stem cell (LSC) depends on specific interactions with extracellular matrix, soluble factors, and cellular components of the microenvironment, or niche. These interactions promote LSC self-renewal and survival, thus contributing to chemoresistance and treatment failure. Understanding the signaling pathways that promote LSC maintenance in response to niche interactions may reveal novel targets for therapy. Recent studies indicate a critical role for the small Rho GTPase, Cdc42, in the maintenance of normal hematopoietic stem and progenitor cells (HSPCs). Cdc42 coordinates actin cytoskeleton organization, adhesion, migration, self-renewal, cell polarity, proliferation, and survival of normal HSPCs in response to niche signaling through multiple cell surface receptors, including CXCL12/CXCR4, SCF/KIT, and fibronectin/integrin. Cdc42 activity is increased in both murine and human models of MLL-AF9 (MA9) acute myeloid leukemia (AML). Cdc42 expression is also increased in human patient AML samples across cytogenetic subtypes, compared to normal hematopoietic cell subsets, in analysis of curated datasets in the HemaExplorer database. In earlier work, we have shown that Cdc42 inhibition leads to peripheral mobilization of leukemia cells out of the marrow niche (Blood 114, 13). In the present study, we investigate whether Cdc42 inhibition also disrupts intrinsic LSC self-renewal. To interrogate Cdc42 in LSC self-renewal, MA9 cell lines were established following transduction of bone marrow HSPCs harvested from tamoxifen-inducible Cdc42 knockout mice, with Cre null donors as controls. Upon tamoxifen (TAM) treatment, Cdc42KO-MA9 cells had decreased CFU and small, diffuse colony morphology. Mice transplanted with untreated MA9 cells were divided to receive injections of TAM vs control. The Cdc42KO-MA9 cohort remains alive at over 180 days post-transplant, whereas vehicle control mice died of AML with latency similar to Cre null MA9 cell recipients (p Disclosures: No relevant conflicts of interest to declare.

Abstract 4897 Much is known regarding the location, cellular composition, signaling pathways, and functional role of the normal hematopoietic stem cell (HSC) niche in the bone marrow microenvironment. Microenvironmental cells including osteoblasts, other specialized mesenchymal cells, and vascular endothelial cells exert control over HSC self-renewal, differentiation, and engraftment. Niche occupancy appears to be competitive and limiting in terms of controlling the number of HSCs per organism. Leukemia stem cells (LSCs), through their inherent properties of quiescence and resistance to chemotherapeutic agents, are thought to be one of the principal mechanisms underlying disease relapse in patients. Much less is known regarding the interaction of LSCs and the marrow microenvironment. It is not clear whether LSCs localize to the same niches as HSCs, compete with HSCs for niche occupancy, or share dependence on niche signals, and whether those signals affect tumor responses to chemotherapy. Using a human pre-B ALL xenograft mouse model, Colmone et al (Science 2008) recently showed that leukemic cells may alter the normal microenvironment, resulting in initial homing of transplanted normal HSPCs in distinct atypical niches. Shiozawa et al (JCI 2011) showed that metastatic prostate cancer cells, a tumor type known to target bone, impeded HSC engraftment in a murine model, suggesting competition for the same niche. To investigate the relationship between HSC and LSC niche localization and functional occupancy, we used murine progenitor cells transduced with an MLL-AF9 vector expressing GFP in a murine syngeneic competitive transplantation model. MLL-AF9 cells are highly enriched for LSCs, particularly the c-kit+ compartment (Somervaille Cancer Cell 2006). We found that between approximately 21% and 24% of cells were c-kit+ by FACS in 2 separate experiments. In our model, mice transplanted with unsorted MLL-AF9 cells (1×107) died of AML with a latency of 11–14 days. We cotransplanted a fixed number of MLL-AF9-GFP cells (1×106) with increasing numbers of normal mouse whole bone marrow (WBM) cells, derived from dsRed transgenic mice to facilitate distinction from the GFP+ MLL-AF9 cells, into mice irradiated with 1000 rads: 1×105 [group 1], 1×106 [group 2], 1×107 [group 3], 5×107 [group 4]. Control groups received 1×105 and 1×106 normal WBM cells only. Survival was monitored daily. The control group receiving 1×105 cells only all died with median time to death of 16.5 days from lack of count recovery, those receiving 1×106 cells are still alive 35 days after transplant, indicating that 1×106 cells is adequate to rescue from irradiation. Mice were bled weekly until death and samples were analyzed by flow cytometry. Complete blood counts, blood smears, and splenic sections were obtained from these mice. As expected, there were no circulating blasts detected 7 days post transplant and all mice were healthy. However, 14 days after transplant the percentages of GFP+ leukemic cells detected in the blood were inversely proportional to the number of normal dsRed WBM cells transplanted (group 1 vs. group 2 vs. group 3 vs. group 4 mean percentage of GFP+ cells, 83.97 v 66.53 v 18.73 v 9.275 p< 0.0001). At day 15, mice from group 1, but not from groups 2 to 4, became moribund and were sacrificed. Spleens in this group were heavier than in those mice transplanted with 1×105 normal WBM cells alone and 2 out of 3 showed leucocytosis compared to leucopenia in all mice in the group transplanted with normal cells alone. When mice in the other groups had blood samples taken for analysis while moribund, GFP+ cells were greater than 80% suggesting that mice in group 1 died from complications relating to leukemic infiltration. Confocal microscopy confirmed the colocalization of normal HSPCs and MLL-AF9-GFP LSCs in the niche. Most interestingly, survival was proportional to the numbers of normal WBM cells transplanted, with a continuous delay in leukemic death proportional to the number of normal WBM cells cotransplanted with the same dose of MLL-AF9 cells (Figure 1). Hence, this murine model of leukemia suggests that normal and leukemic cells compete for the same functional niche, that manipulation of the niche could impact on response to anti-leukemic therapies, and that cell dose in the context of stem cell transplantation for leukemia may have an impact on outcome via niche competition. Figure 1 Figure 1. Disclosures: No relevant conflicts of interest to declare.

Abstract 1536 The Rac family of small Rho GTPases has attracted interest as a therapeutic target in hematologic malignancies due to their central role in coordinating diverse cellular processes such as adhesion, cytoskeletal organization, proliferation, and survival. Rac activity is increased in MLL-AF9 (MA9) acute myeloid leukemia, and Rac inhibition using the small-molecule NSC23766 induces apoptosis in MA9 cells. We recently found that loss of Rac2 delays development of MA9 leukemia in a murine genetic model. Furthermore, latency of disease can be rescued by ectopic expression of Bcl-xL in the Rac2 knockout cells implicating Rac2 regulation of pro-survival Bcl-2 proteins in MA9 leukemogenesis. Whether Rac survival signaling through Bcl-2 proteins can be exploited therapeutically in the developed MA9 leukemia is untested. We hypothesized that Rac2 signaling is critical for MA9 leukemia cell survival, and that inhibition of Rac or downstream Bcl-2 proteins could be an effective therapeutic strategy alone or in combination. We tested our hypothesis in MA9 cells derived from human CD34+ umbilical cord blood (UCB) cells expressing MA9 by retroviral transduction. Lentiviral knockdown was used to determine the specific contribution of Rac2 to MA9 cell survival and maintenance. Decreased levels of Bcl-xL and Bcl-2 were seen in MA9 cells expressing Rac2-targeting shRNA, compared to non-targeting control. Rac2 knockdown induced apoptosis and impaired growth of MA9 cells in culture. Furthermore, Rac2 deficiency reduced in vitro colony-forming ability indicating impairment of the clonogenic MA9 cell. MA9 cells expressing two different Rac2 shRNA vs. non-targeting control were injected into NOD/LtSz-scid-SGM3 (NSS) mice to determine whether Rac2 deficiency impairs engraftment and progression of the MA9 leukemia stem cell in a xenotransplantation assay. Flow cytometric analysis of bone marrow aspirates showed markedly reduced MA9 engraftment in the Rac2 knockdown groups. Whereas all mice in the control group eventually died of MA9 leukemia (N=5), only one death from MA9 cells expressing Rac2 shRNA was seen in the knockdown groups (N=10). We next evaluated the effects of direct inhibition of Bcl-2 proteins downstream from Rac using the BH3-mimetic ABT-737. Three different MA9 cell lines, as well as the THP-1 cell line bearing an MLL-AF9 fusion, were highly sensitive (IC50 ∼30 nM) to ABT-737, with no toxicity seen in control UCB cells in the dose range tested. To determine in vivo efficacy, ABT-737 was administered to NSS mice engrafted with human MA9 cells. End of treatment aspirates showed a marked decrease in leukemia engraftment in the ABT-737 treatment group compared to vehicle control ( Disclosures: No relevant conflicts of interest to declare.

Abstract 3304 Although significant progress has been made in the treatment of leukemia, relapse continues to be a major problem, particularly in acute myeloid leukemia (AML). The prognosis for relapsed leukemia is poor, indicating an area for potential improvements. However, animal models to study the response of human AML to chemotherapeutics and subsequent relapse are lacking. Recently we developed an improved NOD/SCID mouse with IL2RG knockout and transgenic expression of myelo-supportive cytokines SCF, GM-CSF, and IL-3 (the NSGS mouse). This mouse is remarkable in its ability to accept human AML grafts more efficiently than all other available strains. When coupled with in vitro derived AML cells, the NSGS mouse allows for a more predictable AML model with shorter latency and smaller range of death than in other mouse strains, including NSG mice. Importantly, very low numbers of cells reliably generate fatal AML in roughly 40 days, even in non-irradiated NSGS mice, allowing for rapid experimental conclusions and reduced toxicity. With the benefits of these unique tools, we sought to develop a model system to evaluate the efficacy of chemotherapeutic agents on human AML cells in vivo. Engrafted mice received a chemotherapy regimen over a 5-day treatment period consisting of a daily dose of cytarabine with simultaneous injection of doxorubicin during the first three days. Treated mice experienced striking weight loss during the treatment period with a nadir at days 8–10 post-treatment. Mice recovered body weight within 3 weeks. Serial complete blood counts indicated a rapid transient drop in total white blood cell and neutrophil counts and a delayed transient drop in red blood cell and platelet numbers, reminiscent of the effects observed in patients undergoing chemotherapy. The drugs successfully targeted the cells of the bone marrow, as evidenced by a profound loss of cellularity in treated mice relative to controls. When mice harboring N-Ras(G12D) positive AML cells were treated at early time points post-transplant, a significant reduction of tumor burden was observed in the BM and PB, with the grafts of treated mice essentially undetectable for weeks after treatment cessation. Nevertheless, treated mice inevitably succumbed to disease, although with a significantly prolonged latency compared to mock treated mice. However, when AML cells containing the FLT3-ITD mutation were used, a shift in disease latency was not reproducibly seen. This data correlates well with patient data showing that FLT3-ITD mutant AML has a worse prognosis than AML samples with N-Ras mutations. Importantly, the reappearance of AML within weeks of treatment affords the opportunity to model drug resistance and relapse, as well as the potential synergistic effects of experimental compounds used in combination with traditional chemotherapy. Additionally, the period following treatment may allow for studies of minimal residual disease as well as the testing of potential maintenance therapies. Finally, this approach permits a detailed analysis of the critical few cancer stem cells that remain after induction therapy with the goal of identifying novel compounds capable of targeting these cells. Disclosures: No relevant conflicts of interest to declare.

Abstract 3524 Poster Board III-461 Several strains of immunodeficient mice have been employed in the effort to model both normal and aberrant growth and differentiation of human hematopoietic cells in vivo. Despite numerous successes in this field, room remains for improvement. For example, robust models of in vivo differentiation and expansion of human erythroid and T cells are lacking or challenging at best. With the goal to further improve in vivo human hematopoietic cell study, we have crossed the NOD/SCID IL2Rγ knockout mice (NSG) with the NOD/SCID-SGM3 mice (NS-SGM; transgenic expression of human cytokines SCF, GM-CSF, and IL-3) to create a new strain of immunodeficient mice (NSG-SGM3). Uncultured umbilical cord blood CD34+ cells (UCB) were used to initiate xenografts in these three strains as well as NOD/SCID (NS) mice for comparison. Although xenografts with high human composition can be set up in NS mice, the graft, which is B-lymphoid biased without any measurable T cell component, tends to fade over time. In comparison, NSG mice offer longer sustained engraftment with a predominantly B-lymphoid bias and an inconsistent minor T cell component that is absent in NS mice. These mice also display increased levels of human cells in the periphery, allowing for easier measurement of the graft. Similar to NS grafts, NS-SGM3 grafts are initiated at high levels of chimerism, decrease significantly over time, lack T cells, and have a relatively low peripheral involvement. However, the graft consists of a more balanced myeloid/lymphoid distribution. The NSG-SGM3 combines some of the useful properties of the parent strains. Grafts remain at high, sustained levels through 16 weeks while maintaining a fairly balanced myeloid/lymphoid contribution. Chimerism in the NSG-SGM3 mice is readily detectable in the periphery as well as the BM. Remarkably, as early as 8 weeks post-transplant, we observed consistently high human T cell engraftment and expansion far superior to what we observed in NSG mice in parallel experiments using the same UCB samples. Flow cytometric analysis revealed the presence of numerous T cell sub-populations, suggesting significant recapitulation of human T cell development and function in NSG-SGM3 mice. Although it remains unclear as to why the unique combination of SGM3 cytokine expression and IL2Rγ knockout promotes enhanced human T cell engraftment and expansion, this mouse presents an improved opportunity to examine human T cell development and function in a xenograft setting as well as modeling T cell associated diseases including HIV infection, GVHD, and T cell leukemia. Disclosures: No relevant conflicts of interest to declare.

Abstract 1971 Poster Board I-994 Rac GTPases have been found to play an important role in MLL-AF9 (MA9) acute myeloid leukemia. Rac activity is increased in MA9 leukemia cells, and inhibition of Rac by shRNA or the Rac specific inhibitor NSC23766 induces rapid apoptosis in MA9 cells but not in normal cord blood cells or cells expressing the t(8;21) fusion protein AML1-ETO (Wei et al, Cancer Cell 2008). More recently, we demonstrated that MA9-mediated leukemogenesis is significantly delayed in Rac2 knockout (Rac2 KO) low-density bone marrow (LDBM) cells in contrast to the rapid development of leukemia seen after transplantation of wild-type (WT) murine LDBM expressing MA9. Loss of Rac2 has previously been shown to affect Bcl-2 family expression, specifically resulting in decreased levels of the pro-survival protein Bcl-xL and increased levels of the pro-apoptotic protein Bad (Yang et al, Immunity 2000). Here we investigate the role of Bcl-xL in mediating the effects of Rac signaling in MA9 leukemia. Interestingly, treatment with NSC23766 inhibits Bcl-xL expression in MA9 cells. This effect was also seen in Bcl-xL-transduced MA9 cells, implicating a role for Rac in post-transcriptional regulation of Bcl-xL expression in addition to the transcriptional effect that has been previously reported. Significantly fewer Annexin V+ cells were found in NSC23766-treated MA9 cells expressing ectopic Bcl-xL protein when compared to control MA9-pBabe cells, demonstrating that excess Bcl-xL expression can partially rescue the phenotype associated with Rac inhibition. To target the Bcl-2 survival signals in MA9 cells, we used the small-molecule pan-Bcl-2 inhibitor GX15-070 and found a significant inhibition of survival and growth of treated MA9 cells at much lower doses than required for control cells not expressing MA9. Similarly, shRNA knockdown of Bcl-xL in MA9 cells inhibited growth. To determine whether ectopic expression of Bcl-xL could compensate for loss of Rac2 in vivo and restore short latency MA9-mediated leukemogenesis, LDBM was harvested from Rac2 KO mice and transduced with a retroviral vector expressing MA9 and a second retrovirus to express either Bcl-xL or empty vector control. Unsorted cells were transplanted into congenic CD45.1 mice to also establish a competitive assay between singly transduced cells expressing MA9 alone and doubly transduced cells co-expressing both MA9 and Bcl-xL. All recipients of Rac2 KO cells from the MA9+empty vector transduction group remain alive at 300 days from transplant. All recipients of Rac2 KO cells from the MA9+Bcl-xL transduction group died of leukemia with a mean latency of 184 days (range 102 — 280 days), and analyzable tumors showed co-expression of MA9 and Bcl-xL. Two mice that died later in this cohort (day 224 and day 280) also showed a second population of cells expressing MA9 without co-expression of Bcl-xL. Experiments are underway to determine which population has leukemogenic ability upon secondary transplant. Mice that received WT LDBM transduced with MA9+Bcl-xL died of leukemia with the expected mean latency of 190 days for this model (range 130 — 250 days). In this group, the number of WT MA9 tumors co-expressing Bcl-xL was in equal proportion to tumors co-expressing empty vector, suggesting the advantage bestowed by Bcl-xL overexpression is more pronounced in the context of Rac2 deficiency. Together, these findings suggest a critical role for Rac signaling through the Bcl-xL pathway in MA9 leukemia and support therapeutic approaches targeting this pathway in MLL leukemia. Disclosures: No relevant conflicts of interest to declare.

Abstract 13 A large body of studies has proposed that leukemia initiating cells (LICs) are the culprit of leukemia relapse from conventional therapies. Like normal hematopoietic stem/progenitor cells (HSPCs), LICs are thought to reside in the bone marrow (BM) endothelial and/or endosteal niche. Although the full nature of LIC-niche interaction remains elusive, it has been postulated as a useful target for leukemia therapy based on a dual rationale: on one hand, the survival of LICs may depend upon interactions with specific niche, while on the other hand, chasing LICs out of the BM niche may drive quiescent LICs into active cell cycle, sensitizing them for conventional chemotherapy. Recent progresses in studying the PML tumor suppressor and the CXCR4-antagonist AMD3100 have provided strong support for such a rationale (Blood 113, 6215; Nature 453, 1072). The Rho GTPase family member Cdc42 is a central regulator of cell proliferation, adhesion, and migration by integrating signals from multiple cell surface receptors. Ablation of Cdc42 in HSPCs in a conditional knockout mouse model leads to massive egress of HSPCs from BM to the peripheral blood (PB), a phenotype attributable to deficiencies in HSPC adhesion, migration, and F-actin polymerization. We recently demonstrated that pharmacological targeting of Cdc42 by a small molecule Cdc42 activity-specific inhibitor (CASIN) transiently and specifically inhibits Cdc42 activity and mimics the Cdc42 knockout mobilization phenotype by suppressing HSPC adhesion, migration, and F-actin polymerization. CASIN appears to be active in this regard on both murine and human blood progenitors in vitro and in xenografted mice (Blood 112: 68a, Nature Biotechnology under revision). In the present studies, we hypothesize that pharmacological targeting of Cdc42 by CASIN is effective in mobilizing LICs from the BM niche, thus providing a new method for combinatory therapy against LICs. To test this hypothesis, we used MLL-AF9/N-Ras human AML cells (MA9/N-Ras), whereby the N-Ras G12D oncogene was introduced into MLL-AF9 transduced human CD34+ umbilical cord blood (HCB) cells. These cells grow vigorously in vitro independent of cytokine supplementation and induce AML readily in humanized NOD/SCID-SGM3 (SGM3) mice. Xenotransplant experiments confirmed that MA9/N-Ras cells are clonal inducers of leukemia with the property of LICs, as animals transplanted with either bulk cultures or with single cell derived cultures succumbed to AML with similar latencies. In MA9/N-Ras cells CASIN effectively inhibited downstream effectors of Cdc42 such as p-PAK, p-MLC and p-FAK in a dose-dependent manner. In the SGM3 mouse xenograft, we found CASIN administration (1.2 mg/kg, IV) transiently elicited mobilization of LICs from BM to PB by 20 minutes post injection. This was correlated with in vitro suppression of SDF-1α induced F-actin polymerization detected by FACS analysis and directional migration detected by a transwell assay upon CASIN treatment of the LICs (5-10 μM). Similar observations were made using Cdc42-specific shRNA knockdown of endogenous Cdc42 in the LICs. Continuous CASIN infusion into the xenografted mice for 5 days (1.2 mg/kg, IV, once daily) led to a potent induction of apoptosis of LICs detected by AnnexinV/7AAD staining. Significantly, the CASIN infusion showed no effects on the survival of HCB cells in xenografted SGM3 mice. In addition to a potential niche-dependent survival mechanism, the LICs, not normal HCB cells, appear to directly depend on Cdc42 for survival signals as further in vitro culture studies found that a 24-hour CASIN treatment resulted in a dose-dependent apoptosis of MA9/N-Ras cells, but not of normal HCBs. Finally, mouse genetic studies using MA9 transduced Mx-cre;Cdc42lox/lox BM cells transplanted into congenic BoyJ recipients showed that none of the mice with deleted Cdc42 upon pIpC injection developed AML while all mock-injected mice die from leukemia with less than 4 weeks latency, providing genetic evidence that Cdc42 is required for MA9-induced initiation of AML. Whether CASIN is effective in sensitizing the LICs to conventional chemotherapy in a combinatory regiment is currently under investigation. Our studies present a novel concept that pharmacological targeting of the intracellular signal transducer Cdc42 may have therapeutic value in eradicating LICs. Disclosures: No relevant conflicts of interest to declare.

The t(9;11) translocation fusion gene MLL-AF9 (MA9) is commonly found in acute myeloid and lymphoid leukemia and is associated with intermediate to poor outcome. The specific signaling pathways downstream of MA9 are still poorly understood. It has recently been reported that MA9 leukemia cells express higher levels of the small GTPase protein Rac and CDC42 when compared to in vitro MA9 immortal cells in a murine model. To determine the importance of Rac GTPase signaling in MA9-induced transformation, we used an MA9 leukemia model we recently established involving MA9 fusion gene expression in human CD34+ cells. Treatment with the Rac specific inhibitor NSC23766, or transient knockdown of Rac expression by RNAi, induced rapid apoptosis in MA9 cells but not in normal cord blood or t(8;21) translocation fusion gene AML-ETO expressing cells. These data demonstrate that the Rac signaling pathway plays a critical role in the growth and survival of MA9 leukemia cells. To extend this work to an in vivo genetic model, we used mice deficient in Rac2 (Rac2-KO) or with floxed alleles of Rac1 in mice transgenic for Mx-Cre. Leukemia development was compared in mice transplanted with MA9-transduced wild type, Rac1 or Rac2 deficient low density bone marrow cells. Poly I:C injections were performed 2 weeks after transplantation to delete Rac1. Rac deletion was confirmed by PCR and western blot analysis. Mice that received either WT or Rac1−/ − MA9 expressing cells uniformly developed AML and died at 3 to 5 months. Mice transplanted with Rac2-KO cells expressing MA9 showed a decreased incidence and increased latency (6 to 10 months) of AML development despite the persistent engraftment of MA9-expressing cells. All three groups of mice maintained MA9 EGFP+ cells in peripheral blood over the entire experiment, and eventually gave rise to similar end stage AML with a Gr-1+/Mac-1+/Kit+/B220-/CD3- phenotype and myelomonocytic blast morphology. Combined with our observation in human CD34+ cord blood cells transduced with MA9, these in vitro and in vivo data indicate that MA9-mediated transformation and survival requires Rac and their downstream effectors. Rac2 signaling appears to be particularly important in the murine MA9 AML model. Therapeutic targeting of Rac could be a unique and important approach to treating MLL leukemia.