ALBERT

Description: Background: Interaction between surface receptor CXCR4 (s-CXCR4) and chemokine SDF-1 (CXCL12) is critical in signaling between leukemic blasts and the bone marrow (BM) microenvironment. We previously demonstrated: 1) chemotherapy-induced upregulation of s-CXCR4 in acute myeloid leukemia (AML) and ALL enhances stromal protection from chemotherapy-induced apoptosis; 2) the FDA-approved CXCR4 inhibitor plerixafor reverses stromal protection and chemotherapy resistance both in vitro in stromal co-cultures of pre-B cell ALL cell lines and in vivo in xenografts of primary samples of infant MLL-rearranged (MLL-R) ALL; 3) the novel Protein Epitope Mimetic POL5551, a selective and potent antagonist of CXCR4, blocks the SDF-1-binding site of CXCR4, inhibits SDF-1-induced chemotaxis, and reverses stromal-mediated protection from chemotherapy in vitro in pre-B and T ALL cell lines. Here, we further characterize the effects of POL5551 (POL) on surface adhesion molecule expression in ALL, and its in vivo effects in a xenograft model of HR pediatric ALL. In Vitro Methods/Results: We have previously shown that POL inhibits 12G5 antibody binding to s-CXCR4 in ALL cell lines, suggesting an overlapping of the two binding sites. We sought to verify these results using primary samples of pediatric ALL. We treated 3 pre B and 3 T cell ALL primary samples with a dose range of POL and measured s-CXCR4 by FACS at multiple time points. POL inhibition of 12G5 binding was potent (average IC50 at 2 hours pre B 8.3 nM, T 1.4 nM), rapid (24 hours). Further, POL was significantly more potent at inhibiting 12G5 binding than plerixafor (average IC50 at 2 hours pre B 18.4 nM, T 8.4 nM). To further characterize POL’s effects in ALL, we treated 2 pre B and 2 T cell ALL cell lines with POL or vehicle control and then treated with SDF-1α or vehicle control. Treatment with POL inhibited SDF-1α-induced phosphorylation of ERK1/2 in a dose-dependent manner. In parallel, we measured POL-induced compensatory upregulation of the alternative surface adhesion molecules CXCR7 and VLA-4 (CD49d), and found that POL led to increased CXCR7 expression at early time points that began to decrease after 24 hours. We did not find a consistent effect of POL on CD49d surface expression. Xenograft Methods/Results: Infant MLL-R ALL primary samples (n=4) were transplanted into sublethally irradiated NSG mice. After 2 weeks, mice were treated on days 1-3 of 2 consecutive weeks with 1) vehicle control (C), 2) POL (5 mg/kg SC), 3) AraC (200 mg/kg IP), or 4) POL followed by AraC 4 hours later (POL+AraC). One week after treatment, cells were harvested from BM, spleen, and peripheral blood (PB). Leukemic blasts were defined as human CD19+ and CD45+. Overall leukemic burden (average % blasts in BM+spleen+PB) did not differ between mice treated with either C (56.2%) or POL (49.5%, p=0.12). However, treatment with AraC (36.7%, p=3E-07) or POL+AraC (26.3%, p=4E-15) significantly decreased total leukemic burden compared to C. Notably, POL+AraC significantly decreased total leukemic burden compared to AraC alone (p=0.001), demonstrating that POL increased overall sensitivity to AraC. When analyzed by organ-specific leukemic burden, POL+AraC resulted in decreased leukemic burden compared to AraC alone in BM (42.8 vs. 49.8%, p=0.27), spleen (16.9 vs. 30.8%, p=0.002), and PB (19.3 vs. 29.6%, p=0.008). Interestingly, AraC and POL+AraC led to significantly increased CXCR7 expression (blasts from BM p

Description: Abstract 2734 Background: CXCR4 is a transmembrane G-protein-coupled receptor expressed by hematopoietic stem cells (HSC) and leukemia cells. CXCR4's ligand is SDF-1 (CXCL12), a chemokine constitutively secreted in the bone marrow stromal microenvironment. CXCR4+ HSCs and leukemia cells migrate to marrow niches in response to SDF-1. High level CXCR4 expression may be a prognostic indicator in acute leukemias. We hypothesized that surface CXCR4 expression (s-CXCR4) by leukemias may change dynamically in response to chemotherapy (chemo), and that chemo-mediated upregulation of s-CXCR4 may represent a mechanism of acquired chemoresistance due to enhanced protective SDF-1/CXCR4 signaling. Methods and Results: ALL (697, HB1119, NALM-6, SEMK2) and AML (MOLM-14, MV4-11) cell lines were treated for 48 hrs with dose ranges of 6 chemo agents (dauno, araC, etop, vcr, dex, and mtx). Levels of s-CXCR4 (by FACS) varied at baseline, with NALM-6 [mean fluorescence index (MFI) 4444] and MOLM-14 (MFI 108) having the highest and lowest values. These differences were functional, as NALM-6 demonstrated striking SDF-1-induced chemotaxis, while MOLM-14 did not. Chemo exposure resulted in s-CXCR4 upregulation in 697, MOLM-14, and MV4-11 (“upregulators”), and downregulation in HB1119, NALM-6, and SEMK2 (“downregulators”). For example, MOLM-14 had an average 7-fold increase in MFI from baseline, while HB1119 had a 4-fold decrease. Each cell line responded to all chemo agents in a consistent manner (i.e., either up- or downregulation). To measure stromal protection from chemo-induced apoptosis, cells were treated for 48 hrs in the presence or absence of normal human bone marrow stroma feeder layers and analyzed by FACS, after staining with Annexin V and 7-AAD. IC10 through IC90 values were calculated using Calcusyn. Protective index (PI) was defined as the IC value on stroma divided by the IC value off stroma: therefore, PI 〉1 indicated stromal protection, PI

Description: Abstract 568 Background: Infant acute lymphoblastic leukemia (ALL) is clinically and biologically distinct from ALL in older children. About 80% of infant ALL cases harbor MLL rearrangements (MLLr). MLLr infant ALL is an aggressive disease with poor prognosis, particularly in cases diagnosed at

Description: Background: WT1 is a zinc finger transcriptional regulator and acts as a tumor suppressor gene in various cell types. WT1 mutations are reported in approximately 10% of both adult and pediatric patients with acute myeloid leukemia (AML), and at a lower frequency in patients with myelodysplastic syndome (MDS). Reported mutations consist of insertions, deletions or point mutations, and are thought to alter WT1 DNA-binding ability and result in a loss of function. WT1 mutations are associated with FLT3/ITD mutations in AML, suggesting possible leukemogenic cooperativity, and yet WT1 mutations have been independently associated with treatment failure and a poor prognosis. Recently, a physical interaction demonstrated between WT1 and TET2 suggests a common functional pathway, and explains the mutual exclusivity of these mutations in AML. Despite these observations, the functional contribution of WT1 mutations in hematologic malignancies is not entirely understood. To our knowledge, we are the first to describe here a hematologic phenotype in a WT1 mutant mouse model and in a novel WT1 mutant x FLT3/ITD crossbred mouse model. Methods: Knock-in WT1 mutant mice are heterozygous for missense mutation R394W in the DNA-binding domain, which has been described in cases of human AML. Mice with a heterozygous 18-bp ITD knocked into the FLT3 gene were crossbred with the WT1 mutant mice, and Kaplan-Meier survival analysis was performed across genotypes. CBCs and BM cytospin morphology from moribund mutant mice were compared to wild type controls. To create a transplant model, 2e6 whole BM cells from each genotype were injected into lethally irradiated congenic mice. Competitive transplants were performed by injecting a 1:1 ratio of CD45.1 wild type (control) cells with CD45.2 WT1 mutant or wild type (test) cells into lethally irradiated C45.1 recipients. Results: We noted an expansion of lineage negative cells and various progenitor cell compartments in WT1 mutant (WT1mut) BM relative to wild type (wt); including the megakaryocyte-erythroid progenitor (MEP) compartment. WT1mut BM cells from two-month old mice showed an increased ability to serially replate in methylcellulose culture compared to wt BM cells, demonstrating aberrantly enhanced self-renewal capacity. WT1mut mice demonstrated a trend towards an inferior late survival compared to wt in survival analysis, and several moribund WT1mut mice were found to have anemia and erythrodysplasia. Most ITD mice developed a fatal myeloproliferative neoplasm (MPN), as previously described. Interestingly, double mutant mice (WT1mut+ITD) had an inferior survival compared to ITD (p

Description: Background FLT3 is expressed in most human acute leukemias. When activated by FL, wild type (wt) FLT3 dimerizes and initiates downstream signals that result in proliferation and inhibition of apoptosis and differentiation. Activating FLT3 mutations (internal tandem duplications (ITDs) or point mutations) are common in AML and rare in ALL. ITD mutations confer a poor outcome in AML. In vitro, mutant FLT3 signaling can be further enhanced by binding of FL. Peripheral blood (PB) plasma FL levels rise in adults with AML, peaking about two weeks after initiation of chemotherapy. We sought to determine plasma levels of FL in pediatric patients after chemotherapy, and the functional effect of various levels of FL on both wt and mutant FLT3 leukemia cells. Methods FL levels were measured using FL ELISA on plasma samples (n=352) isolated from PB of children (n=75) enrolled on 4 multi-center acute leukemia clinical trials. Functional studies were performed on AML and ALL cell lines with wt FLT3 (HL60, RS4;11, SEMK2, and KOPN-8) and mutant FLT3 (MOLM14, MV4-11, and HB-1119). 72 hr etoposide IC50 was determined by WST-1 for each line. Cells were plated (250,000 cell/mL) for 72hr at etoposide IC50 in RPMI 1640 along with increasing concentrations of recombinant human FL (62.5 to 4,000 pg/ml). Cell cycle and apoptosis were analyzed using propidium iodide staining and annexin V/7-AAD binding, respectively. To explore the mechanism of FL effects, Ba/F3-ITD cells were incubated for 72hr in serum-free conditions with either 4,000 pg/mL (“high”), 62.5 pg/mL (“low”), or no FL. After washing, total and phosphorylated FLT3 protein levels were determined by Western blot. Results Pediatric patients receiving chemotherapy for the treatment of acute leukemia demonstrate a pattern of plasma FL rise with low levels at baseline (mean 41 pg/ml) and peak levels at day 11-14 following initiation of therapy (mean: 1,190 pg/mL; max: 5,783 pg/mL)(Fig 1A). Cell lines with FLT3 activating mutations selectively demonstrate resistance to etoposide-induced apoptosis (Fig 1B) and G2/M cell cycle arrest (Fig 1C) at low concentrations of FL (62.5 pg/mL). Dose-dependent reduction of etoposide resistance is seen with increasing concentrations of FL up to 4,000 pg/mL, suggesting that optimal etoposide-induced killing of FLT3-mutant leukemias may occur when FL plasma levels are at their peak. Ba/F3-ITD cells pre-incubated with peak concentrations of FL showed diminished baseline FLT3 phosphorylation, suggesting that the interaction of FL and FLT3/ITD exhibits substrate inhibition kinetics and results in a loss of FLT3/ITD-induced activation with high level FL exposure, thus providing a mechanistic basis for the observed loss of etoposide resistance. Conclusions Plasma FL rises to peak levels 11-14 days after initiation of chemotherapy. Through substrate inhibition of mutant FLT3 enzymatic activity, peak FL levels may reduce the etoposide resistance that characterizes FLT3-mutant leukemia cells exposed to pre-chemotherapy levels of FL. Thus, introduction of etoposide in a “time sequential” manner during periods of peak plasma FL levels may enhance killing of residual chemoresistant FLT3-mutant leukemia cells. Disclosures: No relevant conflicts of interest to declare.

Description: Background We have previously demonstrated that inhibition of CXCR4 in ALL decreases CXCR4 antibody binding, inhibits SDF-1α-(CXCL12)-induced chemotaxis, and overcomes chemotherapy resistance conferred by the bone marrow microenvironment. Specifically, we found that treatment with plerixafor and araC significantly decreased leukemic burden in a xenograft model of infant ALL, compared to treatment with araC alone. In those experiments, we treated mice on 3 consecutive days per week for 2 weeks with plerixafor and araC. However, the combination did not eradicate the leukemia in our model. We hypothesized that extended exposure to plerixafor may have led to increased interactions between surviving leukemic blasts and the bone marrow microenvironment. In our current experiments, we sought to characterize the effects of prolonged exposure to plerixafor in ALL. Methods/Results We treated pre-B (HB-1119, Nalm-6) and T (CCRF-CEM-1301, Jurkat) ALL cell lines with a dose range of plerixafor and harvested cells for FACS over an extended time course. We measured surface CXCR4 (s-CXCR4) expression using 3 antibodies: 12G5, which attaches to the SDF-1α and drug-binding site of CXCR4, and 1D9 and 2B11, which do not compete with SDF-1α or drug binding. 12G5 binding was decreased by plerixafor even at 1 hour and this effect was concentration-dependent. Interestingly, we found a time and dose-dependent increase in 1D9 and 2B11 antibody binding, suggesting that plerixafor caused an actual increase in s-CXCR4 over time. Increases in 1D9 and 2B11 binding were inversely proportional to decreases in 12G5 binding. We also measured surface expression of CD49d (VLA-4), which binds to fibronectin and VCAM-1; CXCR7, which binds to SDF-1α and CXCL11; and CXCR3, which binds to CXCL9, 10, and 11. We hypothesized that CXCR4 inhibition would lead to upregulation of parallel pathways of leukemia-stroma interactions. CD49d was highly expressed at baseline, while CXCR7 and CXCR3 were expressed to a lesser degree. Treatment with plerixafor led to dose-dependent increases in CXCR7 and variable changes in CD49d and CXCR3 surface expression, suggesting that plerixafor can modulate surface expression of adhesion molecules other than CXCR4. Next, we treated ALL cell lines with plerixafor (0, 10, 100 nM) for 72 hours, washed with PBS, and resuspended the cells in fresh medium to determine the effects of extended exposure to plerixafor and subsequent withdrawal. First, we measured surface expression of s-CXCR4 after 72 hours of treatment with plerixafor and found that 12G5 binding was decreased, while 1D9/2B11 binding was increased in an inversely proportional manner. After withdrawal, 12G5 binding increased to untreated levels between 4 and 24 hours, while 1D9/2B11 binding decreased to untreated levels between 4 and 72 hours. We also measured surface expression of CD49d, CXCR7, and CXCR3 and found that the effects of plerixafor treatment and withdrawal were variable by cell line. For example, after plerixafor treatment, surface expression of CD49d and CXCR7 was increased in Nalm-6 and surface expression of CXCR7 and CXCR3 was increased in CCRF-CEM-1301. Interestingly, 4 hours after plerixafor withdrawal, CD49d expression was increased in Jurkat and Nalm-6, and CXCR7 expression was increased in CCRF-CEM-1301, HB-1119, and Jurkat. Finally, we measured migration of washed cells from each treatment condition through a permeable membrane toward medium containing SDF-1α or medium alone. Despite CXCR4 inhibition for 72 hours, all plerixafor-treated cells migrated in response to SDF-1α. In addition, some plerixafor-treated cells exhibited significantly increased SDF-1α-induced chemotaxis compared to control-treated cells. These findings imply that increases in s-CXCR4 induced by 72 hours of treatment with plerixafor are functional. Conclusions Treatment of ALL cell lines with plerixafor led to a dose-dependent decrease in 12G5 antibody binding with a simultaneous overall increase in s-CXCR4 expression. Prolonged exposure to plerixafor led to increased s-CXCR4 expression that persisted for up to 72 hours after drug withdrawal, modulated surface expression of additional adhesion molecules, and enhanced SDF-1α-induced chemotaxis. Therefore, additional careful studies of CXCR4 inhibitors and other microenvironment-targeted agents must be performed in order to determine their optimal use in ALL. Disclosures: No relevant conflicts of interest to declare.

Description: Background The WT1 gene encodes for a zinc finger-containing transcription factor involved in differentiation, cell cycle regulation and apoptosis. WT1 expression is developmentally regulated and tissue-specific, with expression maintained in the kidney and in CD34+ hematopoietic progenitor cells. Inactivating mutations of this tumor suppressor gene are well-described in sporadic Wilms tumor and as germline mutations in Wilms tumor predisposition syndromes. WT1 mutations have been reported in approximately 10% of both adult and pediatric patients with cytogenetically-normal acute myeloid leukemia (CN-AML), and have been associated with treatment failure and a poor prognosis. These reported mutations consist of insertions, deletions or point mutations. Many are frameshift mutations in exon 7, can occur as biallelic double mutations, and result in truncated proteins which may alter DNA-binding ability. Missense mutations in exon 9 have also been identified, and reports suggest that these may act in a dominant-negative manner, resulting in a loss of function. Despite these observations, the functional contribution of WT1 mutations to leukemogenesis is still largely undetermined. Methods/Results We obtained a novel knock-in WT1 mutant mouse model, which is heterozygous for the missense mutation R394W in exon 9, and homologous to exon 9 mutations seen in human AML. We hypothesized that WT1 mutations may have an aberrant effect on hematopoiesis, and specifically, could alter progenitor cell differentiation or proliferation. To investigate this, we collected lineage-negative bone marrow (lin- BM) cells from two-month old WT1 mutant (WT1mut) and wild-type (wt) mice. We performed methylcellulose colony-forming assays, serially replating cells every 10-12 days. Strikingly, WT1mut progenitor cells showed higher in vitro colony-forming capacity and an increased ability to serially replate, suggesting aberrantly enhanced self-renewal capability. Furthermore, WT1mut colonies from secondary and tertiary passages were larger and more cohesive than wild-type colonies, demonstrating increased proliferation and morphology consistent with blast colony-forming units (CFU-blast). Flow cytometric analysis of these WT1mut cells at tertiary replating revealed an immature, largely c-Kit+ population. Next, in order to study the effects of WT1mut on HSCs in vivo, we performed serial competitive transplantation of HSC-enriched, lineage-depleted BM into lethally irradiated mice. At 14 weeks post-transplant, the donor bone marrow cells were harvested and analyzed by flow cytometry. We observed a significant expansion of the LT-HSC compartment in the WT1mut mice compared to wild-type mice. These data provide new insight into the biology and functional role of WT1 mutations in the aberrant regulation of hematopoietic stem and progenitor cell expansion. Conclusion Oncogenic WT1 mutations confer enhanced proliferation and renewal of myeloid progenitor cells in vitro and expansion of LT-HSCs in vivo. Our findings suggest that WT1 mutations enhance stem cell self-renewal, potentially priming these cells for leukemic transformation upon acquisition of cooperative events. Disclosures: No relevant conflicts of interest to declare.

Description: Introduction: Patients who harbor the Philadelphia (Ph+) chromosome t(9;22) translocation account for approximately 20-30% of adult ALL and 2-5% of pediatric ALL. Prior to approval and use of imatinib, a small molecule TKI which targets the Ph+ chromosome BCR-ABL1, these patients had poor survival & EFS - with long term survival rates in the 20% range. With the addition of imatinib and later generation TKIs to chemotherapy backbones and bone marrow transplant, EFS & survival rates have substantially improved - surpassing 50% in studies in adults and even higher in children. However, resistance to imatinib and other TKIs has become a significant problem in Ph+ ALL, especially in adults. ABL1 kinase domain mutations are the dominant form of TKI resistance, however other resistance mechanisms include upregulation of parallel pathways such as SRC family kinases, MAPK and BCL6 pathways. BCL6 is an oncogene that suppresses transcription of tumor suppressor genes such as p53 and CDNK1A. Interestingly, BCL6 has been shown to be upregulated and activated through deacetylation following imatinib treatment in Ph+ ALL, likely leading to its role in resistance. Histone deacetylase inhibitors (HDACi) have been shown to act synergistically with TKIs in imatinib sensitive and resistant Ph+ leukemia though multiple mechanisms including attenuation of BCR-ABL1 levels and other downstream proliferation promoting pathways. We have shown that HDACi treatment acetylates (and thus inactivates) BCL6 in Ph+ ALL, and that the combination of HDACis and TKIs leads to synergistic effects in vitro and in vivo (using xenograft models). Methods: In vitro WST-1 cell viability assays were carried out on TOM1 cells (non-ABL1 mutant, imatinib sensitive Ph+ ALL) and NALM1 cells (non-ABL1 mutant, imatinib resistant CML lymphoid blast crisis) with imatinib and entinostat (a HDACi). Synergy was assessed using Calcusyn software. Western blots were performed assessing BCL6 expression and acetylation, and expression of downstream effectors of apoptosis. Two separate in vivo xenograft mouse experiments were performed transplanting TOM1 and NALM1 cells into Nod SCID Gamma (NSG) mice. Cohorts of TOM1 mice were treated with imatinib 50mg/kg BID, entinostat 15mg/kg QD, imatinib plus entinostat combination, or vehicle control. In the NALM1 mice we added a higher dose imatinib cohort (100 mg/kg BID) due to known imatinib resistance. Results: In vitro, there was substantially more synergy of the imatinib/entinostat combination in imatinib-resistant NALM1 cells vs. the imatinib-sensitive TOM1 cells. Average Combination Index (CI) values in TOM1 cells across multiple entinostat and imatinib doses was 1.2 (CI: =1 suggest additive effect, 1 = antagonism), while the CI in NALM1 cells at the same dose combinations was 0.53. We noted BCL6 upregulation and decreased BCL6 acetylation - signs correlating with resistance - in Western blots of NALM1 and TOM1 cells treated with imatinib, while exposure to entinostat caused increased acetylation of BCL6 and increased expression of downstream tumor suppressors. In the imatinib-sensitive TOM1 xenograft trial, the combination displayed a significant reduction in bone marrow leukemic blast involvement versus control following 6 weeks of dosing as measured by flow cytometry (36.9% mean decrease, p=0.001). There was a trend toward decreased bone marrow involvement between the combination treatment and other active treatment arms. There was no difference in peripheral blood blast percentage between arms. In the imatinib-resistant NALM1 xenograft trial, the combination showed a significant decrease in peripheral blood blast percentage in the combination arms versus all other arms after only two weeks of therapy (p=0.0008). Conclusions: Upregulation of activated BCL6 is a known mechanism of resistance in Ph+ ALL that may be abrogated by acetylation of BCL6 with HDACi, as our in-vitro data suggests. Further, we have shown in xenograft models of Ph+ acute lymphoblastic leukemia that combination therapy with HDACi + imatinib, even in imatinib-resistant leukemia, has significant activity. Interestingly, the combination appears more active in resistant disease than in imatinib-sensitive disease. This combination could prove a viable strategy to attenuate imatinib- (and perhaps other TKI-) resistance in Ph+ ALL relapse, particularly in cases not driven by ABL1 kinase domain mutations. Disclosures No relevant conflicts of interest to declare.

Description: Abstract 3605 Background: AC220 is a novel class III receptor tyrosine kinase (RTK) inhibitor that is potent and highly selective for mutant and wild type (WT) FLT3 and other class III RTK's including KIT, CSF1R, RET and PDGFR. In childhood acute myeloid leukemia (AML), ∼18% of children have FLT3 internal tandem duplication mutations (FLT3-ITD), and ∼10% high WT FLT3 expression. FLT3-ITD is associated with poor prognosis. In childhood acute lymphoblastic leukemia (ALL), the highest levels of FLT3 mRNA expression occur in cases of infants (80%) and childhood ALL with MLL rearrangements (MLL-r) (5%), both conferring poor prognosis.1,2 Study Design: TACL 2009–004is a first-in-children study using AC220 in combination with cytarabine and etoposide. Children 〉 1 month and 〈 21 years of age with relapsed/refractory AML or MLL-rearranged ALL are eligible. A standard 3+3 dose escalation design is utilized. The three doses tested (25, 40 and 60 mg/m2/day) are significantly lower than those tested in adults. Dose escalation past 60 mg/m2 occurs only if adequate biologic activity as determined by a plasma inhibitory assay (PIA) is not achieved. Intravenous (IV) cytarabine (1 gm/m2/dose every 12 hours) and IV etoposide (150 mg/m2/dose daily) are given over 5 days. AC220 is administered once daily as an oral solution on days 7–28. Patients can receive up to 2 courses of therapy. PIA testing is performed at trough time points weekly during exposure to AC220 to determine biologic activity. Results: To date, 13 patients (pts) were enrolled and 12 are evaluable for toxicity and response. One pt died from infectious complications (not drug-related) after a single dose of AC220 and was replaced. Median age at study entry was 10.2 years (range 11 mo – 20 yrs), average number of prior regimens was 2.8 (range 1–5), and 5 pts had prior stem cell transplant. Nine pts had relapsed AML, 2 had relapsed MLL-r ALL, and 1 had secondary AML. Of patients with AML, 4 had FLT3-ITD mutations and one had a D835 mutation. Toxicities were consistent with intensive relapsed leukemia regimens. Across all dose levels, non-hematologic toxicities ≥ grade 3 attributed to AC220 included vomiting (n=1), elevated transaminases (n=1), anorexia (n=2), and infection (n=3). One pt experienced a dose-limiting toxicity (DLT) on dose level 2 (40 mg/m2/day) of recurrent grade 3 elevated lipase. Dose level 2 was expanded to 6 pts without additional DLTs. Of 3 pts treated at 60 mg/m2/day, there have been no DLTs. Near total (〉99%) inhibition of FLT3 phosphorylation by PIA is seen in every patient across all dose levels. Of 12 pts evaluable for response to date, 1 patient achieved a complete response (CR), 3 achieved complete response with incomplete neutrophil and platelet recovery (CRi), 5 had stable disease (SD), and 3 had progressive disease (PD). Responses in the 4 FLT3-ITD pts include 1 CR, 2 CRi and 1 SD. The FLT3-ITD patient with SD had reduction in marrow blasts without peripheral blood count recovery. An additional 6 pts will be enrolled at 60 mg/m2/day to complete safety evaluation and confirm biologic activity. Conclusions: AC220 plus intensive chemotherapy is well tolerated at doses up to 60 mg/m2/day with near complete inhibition of FLT3 phosphorylation in all pts tested to date. Response rates to date in pre-treated children with relapsed FLT3-ITD AML are encouraging. Disclosures: Off Label Use: AC220 in relapsed/refractory pediatric acute leukemia. Gammon:Ambit Biosciences: Employment.

Description: Background AC220 is a novel class III receptor tyrosine kinase (RTK) inhibitor that is potent and highly selective for mutant and wild type (WT) FLT3 with lower activity against other class III RTKs including KIT, CSF1R, RET and PDGFR. In childhood acute myeloid leukemia (AML), FLT3 internal tandem duplication mutations (FLT3-ITD) confer poor prognosis. In childhood acute lymphoblastic leukemia (ALL), high WT FLT3 expression is seen in the high risk MLL-rearranged subset. Study Design TACL 2009-004 was the first clinical trial using AC220 in children, and the first study overall in combination with intensive chemotherapy. Children 〉 1 month and ≤ 21 years of age with relapsed/refractory AML or MLL-rearranged ALL (MLL-r ALL) were eligible. A standard 3+3 dose escalation design was utilized to identify a safe and biologically active dose. The three doses tested (25, 40 and 60 mg/m2/day) are significantly lower than those tested in adults. All patients (pts) received intravenous (IV) cytarabine(1 gm/m2/dose every 12 hours) and IV etoposide (150 mg/m2/dose daily) on days 1-5. AC220 was administered once daily as an oral solution on days 7-28. Patients were eligible to receive up to 2 courses of therapy. Plasma inhibitory assay (PIA) testing was performed at trough time points weekly to determine biologic activity of AC220. Results Twenty-two pts were enrolled: 18 patients were evaluable for response and 20 were evaluable for toxicity. Four pts (1 ALL, 3 AML (2 FLT3-ITD)) not evaluable for response received 〈 75% of AC220 and 2 pts without DLTs were not evaluable for toxicity for the same reason. Four pts had relapsed MLL-r ALL, 17 relapsed AML and 1 secondary AML. Of pts with AML, 9 were FLT3-WT, 8 FLT3-ITD+, and one had unknown FLT3 status. The median number of prior regimens was 2 (range 1-10) for ALL and 3 (range 1-5) for AML. No pts with ALL had prior hematopoietic stem cell transplant (HSCT) while 10/18 with AML had prior HSCT. Toxicities were consistent with intensive relapsed leukemia regimens. Across all dose levels, non-hematologic toxicities ≥ grade 3 attributed to AC220 included vomiting (n=1), elevated transaminases (n=1), anorexia (n=2), rash (n=1), hypophosphatemia (n=1), hypoxia (n=1), headache (n=1), and infection (n=2). One of 6 pts experienced a dose-limiting toxicity (DLT) on dose level 2 (40 mg/m2/day) of grade 3 elevated lipase, and 1 of 9 pts experienced DLT of hyperbilirubinemia on dose level 3 (60 mg/m2/day). Near total (〉99%) inhibition of FLT3 phosphorylation by PIA was seen in every patient across all dose levels. Therefore, dose escalation past 60 mg/m2/day was not performed and a maximum tolerated dose (MTD) was not reached. Responses are listed in the table below. Of note, 4/6 (67%) FLT3-ITD patients achieved CR or CRi. Both FLT3-ITD pts with SD had significant reduction in marrow blasts without peripheral blood count recovery. Conclusions: AC220 plus intensive chemotherapy is well tolerated at 60 mg/m2/day with near complete inhibition of FLT3 phosphorylation in all pts tested to date. The favorable toxicity profile and encouraging response rates warrant further testing ofAC220 in pediatric pts with FLT3-ITD mutated AML. Disclosures: Off Label Use: AC220 is not approved for use in pediatric acute leukemia. Gammon:Ambit: Employment.