Abstract: Abstract 3544 Background: In pediatric de novo AML, cytarabine (Ara-C)-based induction therapy results in above 80% complete response (CR) rates but nearly half of those who achieve an initial remission relapse die of their disease. Accurate prediction of initial response to chemotherapy at the time of diagnosis as well as identification of those at high risk of relapse despite initial remission would allow for patient specific therapy and improved clinical outcome. SCNP is a multiparametric flow cytometry-based assay that provides simultaneously quantitative measurements of extracellular surface markers as well as changes in intracellular signaling proteins in response to extracellular modulators at the single cell level (Kornblau et al. Clin Cancer Res 2010). In a previous study, we used this assay to define two distinct classifiers associated with response to standard induction therapy and risk of relapse in diagnostic bone marrow mononuclear cells from pediatric patients (pts) with non-M3 AML (ASH 2010;116:Abstract 954). This study is intended to confirm the validity of the pre-specified response to induction therapy classifier in an independent set of AML pediatric pts. Methods: The SCNP-based response classifier developed using 53 AML cryopreserved samples from patients enrolled on POG (now COG) trial 9421 was comprised of a combination of three SCNP readouts that measure apoptosis, MAPK signaling, and PI3K signaling and had a bootstrapped out-of-bag estimated Area Under the Receiver Operating Characteristic Curve (AUROC) of 0.84 (95% CI 0.67– 0.96). The classifier was tested on 68 cryopreserved samples (20 non–responders (NR) and 48 CRs) from patients enrolled on COG trials AAML0531 (samples from patients randomized to Ara-C, Daunomycin and Etoposide [ADE] induction therapy) and AAML03P1 (samples from patients treated with ADE plus Gemtuzumab Ozogamicin induction therapy). The primary hypothesis was that the prediction of induction response by the continuous score from the pre-specified classifier would yield an AUROC significantly greater than 0.5. Results: The primary objective of the study was met with an AUROC of 0.66 (n=68) p=0.042 (see table). The primary analysis used an NR classification that combined patients with either induction failure (n=14) or induction death (n=6). A pre-specified analysis in which induction deaths were removed resulted in an AUROC of 0.70 (n=62) p=0.021, suggesting that the underlying disease biology may be different for induction death vs. induction failure. In this study, WBC and cytogenetics risk groups were associated with induction response while age, gender and FLT3-ITD status were not. In a multivariate analysis of induction response that included WBC, cytogenetics and the pre-specified continuous SCNP classifier score, only cytogenetic risk group (p=0.001) and SCNP score (p=0.017) remained significant. Exploratory analyses excluding induction deaths suggest that the relationship between the SCNP score and induction response is strong among patients with an intermediate cytogenetic classification (n=23) (AUROC=0.88, p=0.002), while no relationship (AUC=0.48, p=0.959) is seen in those patients with a poor cytogenetic classification (n=17). Among the three SCNP signaling nodes contributing to the score, the node measuring drug-induced apoptosis performs most consistently across the training and validation sets. Conclusion: This study is the first validation of a SCNP-based classifier that predicts response to induction Rx. It shows that performing quantitative SCNP under modulated conditions can serve as the basis for developing biologically based tests in leukemia that offer new insights into the individual patient's disease biology. Additionally, this technology could prove useful in guiding alternative therapeutic strategies. Disclosures: Gayko: Nodality Inc.: Employment, Equity Ownership. Westfall:Nodality Inc.: Employment, Equity Ownership. Purvis:Nodality Inc.: Employment, Equity Ownership. Putta:Nodality Inc.: Employment, Equity Ownership. Hackett:Nodality Inc.: Employment, Equity Ownership. Cleary Cohen:Nodality Inc.: Consultancy, Equity Ownership.

Abstract: KIT receptor tyrosine kinase mutations are implicated as a prognostic factor in adults with core binding factor (CBF) acute myeloid leukemia (AML). However, their prevalence and prognostic significance in pediatric CBF AML is not well established. We performed KIT mutational analysis (exon 8 and exon 17) on diagnostic specimens from 203 pediatric patients with CBF AML enrolled on 4 pediatric AML protocols. KIT mutations were detected in 38 (19%) of 203 (95% CI, 14%-25%) patient samples of which 20 (52.5%) of 38 (95% CI, 36%-69%) involved exon 8, 17 (45%) of 38 (95% CI, 29%-62%) involved exon 17, and 1 (2.5%; 95% CI, 0%-14%) involved both locations. Patients with KIT mutations had a 5-year event-free survival of 55% (± 17%) compared with 59% (± 9%) for patients with wild-type KIT (P = .86). Rates of complete remission, overall survival, disease-free survival, or relapse were not significantly different for patients with or without KIT mutations. Location of the KIT mutation and analysis by cytogenetic subtype [t(8;21) vs inv(16)] also lacked prognostic significance. Our study shows that KIT mutations lack prognostic significance in a large series of pediatric patients with CBF AML. This finding, which differs from adult series and a previously published pediatric study, may reflect variations in therapeutic approaches and/or biologic heterogeneity within CBF AML. Two of 4 studies included in this analysis are registered at http://clinicaltrials.gov as NCT00002798 (CCG-2961) and NCT00070174 (COG AAML03P1).

Abstract: Acute myeloid leukemia (AML) is a heterogeneous disease. Risk factors such as karyotype, FAB subtype, FLT3 status and response to induction therapy are determinants of outcome with current therapies. We hypothesize that array comparative genomic hybridization (CGH) will identify gene copy number changes that are determinants of outcome. Array CGH was performed on diagnostic bone marrow samples from patients on the COG study POG #9421. In order to determine regions of altered gene copy number, labeled genomic DNA samples were hybridized together with sex-matching normal human reference DNA to cDNA microarrays with 41,751 features (corresponding to 24,473 unique Unigene cluster IDs), arrays were obtained from the Stanford University Microarray Core Facility. Control hybridizations were performed to assess intra- and inter-experimental variability. We studied 70 samples with adequate high-quality DNA. Circular binary segmentation was used to distinguish discrete gene copy number transition points from chance noise events and to transform primary clone-by-clone data into genomic regions of equal copy number. Using gain/loss threshold, based on two-standard deviation range of control self-to-self distribution, novel gene amplifications and deletions were found in profiled samples. The highest alteration recurrence was observed for gains of chromosome 8 (21%) and losses of chromosome 6 (29%). The area of chromosome 8 which was found to be gained is notable for the presence of potential oncogenes such as ERK8. The deleted area of chromosome 6 is notable for the presence of potential regulators of oncogenesis: MDC1, DDR1, NFKBIL1, TNF, and BRD2. In summary, array CGH has identified novel areas of gene copy number gain and loss in this population of pediatric de novo AML patients. Further studies are needed to assess whether these genes are associated with outcome, known risk factors and whether they will provide insight into the heterogeneity of de novo AML.

Abstract: The karyotype of the leukemia cell at diagnosis is of prognostic importance. The presence of t(8;21), inv(16) and t(15;17) are currently used to make therapeutic decisions. Additionally, specific mutations such as those involving the FLT3 gene are of prognostic importance as they indicate patients likely to relapse early or fail initial induction therapy. Approximately 20% of children with AML have normal karyotypes at diagnosis and no identifiable chromosomal abnormality using standard methods of analysis. In this subgroup there is an increased incidence of FLT3 mutations (internal tandem duplications or point mutations). We observed that patients with normal karyotypes who were enrolled in the Pediatric Oncology Group (POG) study #9421 had two significantly different clinical outcomes that were associated with the expression of FLT3 mutations. We hypothesized that gene expression profiles would identify genes that cooperate with FLT3 mutations in conferring poor clinical outcome. Bone marrow and/or blood samples from an initial 103 of 622 eligible patients registered on the POG study #9421 were studied. Cytogenetic testing was carried out by clinical laboratories at institutions where AML was diagnosed; all cytogenetic reports underwent centralized review. Samples were analyzed on a 43,760-element spotted array (containing 41,751 unique genes and ESTs) from the Stanford University Microarray Core Facility. Of the 103 diagnostic samples, 20 had normal karyotypes. The FLT3-mutation status was determined by RT-PCR analysis on RNA samples (exon 14 for ITD’s and exon 20 for point mutations). Twelve patients expressed FLT3-mutations and 8 expressed wild-type FLT3 (FLT3-WT). Both groups had a mean WBC of 103x109/L at diagnosis. Probability of event free survival (EFS) was 75% for the FLT3-WT subgroup and 9% for the FLT3-mutant subgroup (P=0.0006). We used Prediction Analysis for Microarrays (PAM) to find the minimum number of genes that could identify samples with FLT3 mutations. PAM identified a 7-gene cluster that differentiated normal cytogenetic cases by clinical outcome. A cluster of 5 overexpressed genes (i.e., HIST1H2AC, HIST1H2AL, HIST1H2BD, HIST1H2BG, and HIST2H2BE) and 2 underexpressed genes (GASP and IVNS1ABP) were associated with poor outcome. In summary, children with AML, normal karyotype, and absence of FLT3 mutations have a good outcome. In contrast, when these same patients express a FLT3 mutation their outcome is poor despite intensive therapy. This poor outcome is associated with increased expression of several histone genes, which may be downstream targets of the FLT3 signal transduction pathway. We will corroborate these findings in a larger set of POG 9421 diagnostic samples with normal karyotypes.

Abstract: BCR/ABL is associated with an unfavorable prognosis in adults with acute lymphoblastic leukemia (ALL). We used DNA microarrays to identify gene expression profiles and molecular interaction networks related to BCR/ABL status and clinical outcome in a set of 54 adult ALL specimens from the MRC UKALL XII/ECOG E2993 intergroup study (21 p185BCR/ABL- and 16 p210BCR/ABL-positive and 17 BCR/ABL negative). In order to avoid biases associated with commonly used sample amplification procedures, we have implemented an indirect two-step labeling protocol based on signal amplification by use of dendrimer technology. Using a two-class, non-parametric t-test and a false discovery rate cutoff of 5%, we identified 271 genes (including GAB1, CIITA, XBP1, CD83, SERPINB9, PTP4A3, NOV, LOX, CTNND1, BAALC and RAB21) as differentially expressed in BCR/ABL positive ALL compared with BCR/ABL negative ALL. They separate these two classes of adult ALL with an overall accuracy of 93% and are enriched for three highly relevant biological functions: Cellular Growth and Proliferation (57 genes, p = 0.004–0.044), Cell Death (49 genes, p = 0.0007–0.049), and Hematological System Development and Function (40 genes, p = 0.00004–0.049). Network analysis demonstrated complex interaction patterns of these genes, and identified FYN and IL15 as the hubs of the top-scoring network. We confirmed these findings by both qRT-PCR analysis of the initial set of samples and by cross-platform validation in an independent cohort of 128 adult ALL specimens. In addition, within the BCR/ABL positive subgroup, we identified 14 clones found to be over-expressed (TSPAN16, ADAMTSL4) or under-expressed (PILRB, STS-1, SPRY1) in p185BCR-ABL- relative to p210BCR-ABL-ALL. In a nearest-centroid classification, these clones correctly predict the BCR/ABL subtype with a cross-validated prediction accuracy of 95%. No differential gene expression was detected among Rho family GTPases and their known interaction partners. Finally, we constructed a gene expression- and interaction-based outcome predictor consisting of 27 genes (including GRB2, GAB1, GLI1, IRS1, RUNX2 and SPP1), which strongly correlated with overall survival in BCR/ABL positive adult ALL (p=0.0001), independently of other clinical parameters such as age (p=0.25) and white blood cell count at presentation (p = 0.003). These findings may be useful for developing novel therapeutic targets and prognostic markers in adult ALL.

Abstract: Abstract 482 Background: Acute megakaryoblastic leukemia (AMKL) in children without Down syndrome represents approximately 10% of pediatric acute myeloid leukemia (AML). Published reports regarding prognosis of non-Down syndrome children with AMKL vary; hence, some groups treat these patients as high-risk while others as standard-risk. As a result, the optimal treatment strategy, including the role of hematopoeitic stem cell transplant (HSCT), remains unclear. Methods: We reviewed the treatment and outcomes of patients with AMKL treated on two pediatric AML protocols. Pediatric Oncology Group (POG) 9421 (1995-99) included two cycles of induction chemotherapy with randomization to standard or high-dose DAT (daunorubicin, cytarabine, 6-thioguanine) followed by HSCT for children in remission with matched sibling donors or three cycles of consolidation chemotherapy with or without cyclosporine for those without suitable donors. St. Jude protocol AML02 (2002-2008) randomized patients during induction to receive high- or low-dose cytarabine with daunorubicin and etoposide (ADE). Patients without complete remission (CR) received gemtuzumab ozogamicin (GO) with induction 2 chemotherapy (ADE). After induction, AMKL patients with availabel donor proceeded to HSCT while those without donors received three cycles of consolidation chemotherapy including mitoxantrone and cytarabine. Comparisons of event-free survival (EFS) and overall survival (OS) were performed by Mantel-Haenszel log-rank test. Independent effect of HSCT on EFS or OS were analyzed by Cox proportional hazard model, with HSCT as time-dependent variable. Results: The 49 patients with AMKL treated on POG 9421 had a median age at diagnosis of 1.8 years (range 0.13-16.2 years), and a median leukocyte count of 13.5 × 109/L (range 0.3-98.4). The patient with the t(1;22) translocation remains in CR. 39 patients had CR after induction, 5 partial response and 4 no response with no difference between standard- and high-dose DAT. The 5-year EFS for all AMKL patients was 34.7% ± 7.5%. Of the 39 patients in CR, 6 underwent matched sibling HSCT, of whom 4 remain in CR, and the other 33 patients received consolidation chemotherapy, of whom 15 remain in continuous remission. The 5-year OS for patients undergoing HSCT was 66.7% ± 19.3% compared to 36.1% ± 8.4% for those receiving consolidation chemotherapy (p = 0.2). The 26 patients with AMKL treated on SJCRH AML02 had a median age at diagnosis of 1.2 years (range 0.21-11.2 years), and a median leukocyte count of 11.8 × 109/L (range 2.3-72.9). All 5 patients with the t(1;22) translocation achieved CR after one induction cycle, with negative minimal residual disease, and remain in CR after treatment with chemotherapy only. The 3-year EFS was 35.9% ± 12.9% for the other AMKL patients, significantly lower than the 61.5% ± 4.8% for non-AMKL patients (p = 0.026). The 3-years OS estimates are 100% for patients with t(1;22), 48% ± 14.1% for other AMKL patients, and 71.8% ± 4.4% for non-AMKL patients (p = 0.03). Of the 21 patients without t(1;22), 14 received HSCT with 7 survivors (3-year EFS 46.3 ± 17%) while 7 received chemotherapy with only one long-term survivor (3-year EFS 14.3 ± 9.4%) (p=0.55). Conclusions: AMKL patients with the t(1;22) translocation have excellent treatment outcome with chemotherapy alone. In contrast, AMKL patients without t(1;22) fared poorly, especially when they were treated with only chemotherapy. Additional studies of larger number of patients are needed to determine if they would benefit from HSCT. Disclosures: No relevant conflicts of interest to declare.

Abstract: Daunorubicin and cytarabine are the most effective chemotherapeutic agents for AML. Using MTT (3,[4,5-dimethylthiazol-2-yl]-2,5-diphenyltetrazolium bromide) assays as an in vitro test for sensitivity to these agents, investigators have found that leukemia cells from pediatric AML patients with Down Syndrome, which are more sensitive to cytarabine than blasts from non-Down Syndrome AML patients, have been associated with an increased ratio of deoxycytidine kinase to cytidine deaminase and an increased level of cystathionine-beta-synthetase [Ge, Jensen, Stout et al. Cancer Res2004; 64, 728–35] . Looking for related findings in non-Down syndrome AML, leukemic blasts from non-Down syndrome children with AML were examined utilizing gene expression profiling and MTT assays. Expression signatures from cDNA arrays were related to the drug sensitivity results to daunorubicin and cytarabine by supervised clustering. These results will then be analyzed to look for relationships to the patient’s clinical outcomes and identify genes that may be potential therapeutic targets for drug resistance reversal. Bone marrow or blood specimens from 103 patients registered on the Pediatric Oncology Group (POG) protocol 9421 were studied. The diagnostic specimens were tested using MTT assays to measure the IC50 (drug concentration that causes 50% of the cells to die) for both daunorubicin and cytarabine. Samples from ninety-three de novo childhood AML patients were analyzed with a 43,760 element spotted array (containing 41,751 unique genes and expression sequence tags [ESTs]) from the Stanford University Microarray Core Facility. We selected the gene expression profile from 10 samples with the highest and lowest IC50’s for each drug. Then, we used Significance Analysis of Microarrays (SAM) to find significant differences in the gene expression between samples sensitive and resistant to each drug. A cluster of 118 overexpressed genes was associated with resistance to daunorubicin (e.g., BCL6, BAPX1, BCL2A1, MBD2, RAB31 and CDKN1A). A cluster of 24 overexpressed genes was associated with resistance to cytarabine (e.g., RIT1, SH3BP2, BCL2A1, NFKBIA, and PTPRE). In summary, MTT assays and gene expression profiling have identified genes that correlate with daunorubicin and cytarabine resistance. Study of the relationship among drug resistance patterns, gene expression profiles and clinical outcome may allow us to better predict patients’ prognoses at diagnosis and also can be used to identify novel targets for drug resistance reversal.

Abstract: Background: Patients with relapsed or refractory (R/R) acute lymphoblastic leukemia (ALL) and lymphoblastic lymphoma have an unmet need for new treatment strategies. Navitoclax (Nav), a BCL-2/BCL-XL/BCL-W inhibitor, has shown promising effects in hematologic malignancies, however its usage is limited by prolonged thrombocytopenia mediated by inhibition of BCL-XL in platelets (J Clin Oncol. 2012;30:488). Venetoclax (Ven) is a highly selective BCL-2 inhibitor with similar anti-tumor efficacy. The addition of Nav to Ven has shown synergistic effects in preclinical models and might mitigate the dose-limiting thrombocytopenia associated with Nav monotherapy (Blood. 2016;128:1382). We report updated outcomes for ALL patients treated with Ven, Nav, and chemotherapy. Methods: R/R ALL patients aged ≥4 years were enrolled in this phase 1, multicenter, open-label, dose escalation study (NCT03181126). Patients received the weight-adjusted equivalent of 200 mg Ven on day 1, and a 400 mg equivalent daily thereafter. Daily oral Nav was administered on day 3 onward with up to three dose levels for patients ≥45 kg (25, 50, 100 mg) and up to two dose levels for patients 〈 45 kg (25, 50 mg). Patients could also receive chemotherapy from day 9 onward, consisting of peg-asparaginase (1250 IU/m2 intravenous [IV] on days 9 and 22), vincristine (1.5 mg/m2 IV on days 9, 15, 22, and 29), and dexamethasone (20 mg/m2/day orally on days 9-13 and 22-26) at the investigators' discretion. Disease assessment was made by flow cytometry on day 8 and day 36 and as clinically indicated. Minimal residual disease (MRD; 〈 10-4 cutoff for MRD-negativity) evaluation was performed at time of disease assessment, if clinically indicated. Results: Nine adult patients, ≥45 kg (5 with B-cell ALL and 4 with T-cell ALL) have been enrolled as of the data cutoff of June 1, 2018. All patients treated thus far have received 400 mg QD Ven and 25 mg QD Nav (Dose Level 1). Chemotherapy was started on day 9 for 7 patients; patient A began chemotherapy on day 169 and patient H began on day 10. Baseline characteristics of the first 9 patients are shown in Table 1. The most common adverse events (AEs) of any grade are shown in Table 2. Serious AEs included nausea, vomiting, febrile neutropenia (n=2 each), abdominal pain upper, pseudomonal sepsis, somnolence, septic shock, acute pancreatitis, and pulmonary embolism (n=1 each). Of the serious AEs, only febrile neutropenia was considered possibly related to Ven + Nav. AEs led to dose interruptions in 4 patients. No dose-limiting toxicities (DLTs) have occurred. As of the data cutoff, 2 patients have died. Patient B died due to disease progression on day 148 and Patient E died due to an event of sudden death on day 38 not related to study drug (event was associated with not wearing a LifeVest personal defibrillator). Preliminary efficacy for these 9 patients is shown in Table 3. All 9 patients were evaluated at day 8 after Ven + Nav therapy alone. One patient had a complete response with incomplete marrow recovery (CRi) and 1 patient had a partial response (PR); the remaining patients had stable disease (SD). Five patients had assessments at day 36 or later, including 4 patients who had completed 1 cycle of chemotherapy combined with Ven + Nav, and 1 who had received Ven + Nav only; 3 patients achieved CR, and 1 patient each achieved CRi and CRp as their best response. Patient A achieved CRi after Ven + Nav therapy without chemotherapy on day 8, had a duration of response of 5.1 months before developing progressive disease and is now also receiving chemotherapy. Patient C achieved CR on day 36, stopped all therapy on day 38, and maintains ongoing response at 3.5 months. Overall, in this heavily pretreated group, 5 patients achieved a complete response (CR/CRi/CRp). Two patients with CR had no detectable MRD. The remaining 4 patients have not reached the day 36 assessment as of the data cutoff; of these patients, 1 patient had PR and 3 patients had SD. Conclusions: Ven + Nav in combination with chemotherapy is well tolerated, without any unexpected side effects and no DLTs observed to date. Preliminary data suggest that Ven + Nav with chemotherapy is efficacious in some patients with R/R ALL who have had multiple lines of therapy, including prior stem cell transplant. Enrollment, clinical follow-up, and correlative biology studies are ongoing. Disclosures Alexander: Abbvie: Other: travel expenses. Jabbour:Bristol-Myers Squibb: Consultancy, Research Funding; Abbvie: Research Funding; Takeda: Consultancy, Research Funding; Novartis: Research Funding; Pfizer: Consultancy, Research Funding. Khaw:Abbvie: Research Funding; Amgen: Other: travel expenses, Research Funding; Bristol Myers Squibb: Research Funding; Jazz Pharmaceuticals: Research Funding; Novartis: Other: travel expenses. Mullighan:Cancer Prevention and Research Institute of Texas: Consultancy; Abbvie: Research Funding; Amgen: Honoraria, Speakers Bureau; Loxo Oncology: Research Funding; Pfizer: Honoraria, Research Funding, Speakers Bureau. Leonard:Amgen: Research Funding. Schmidt:Abbvie: Employment, Equity Ownership. Tong:Abbvie: Employment, Equity Ownership. Zhou:Abbvie: Employment, Equity Ownership. Ross:AbbVie, Inc: Employment, Equity Ownership. Rosenwinkel:Abbvie: Employment, Equity Ownership. Jacobson:Abbvie: Employment, Equity Ownership. Kim:Abbvie: Employment, Equity Ownership. Stock:Jazz Pharmaceuticals: Consultancy.

Abstract: Introduction: In children diagnosed with leukemia, relapse and its associated morbidity and mortality remain the most dreaded consequences of the disease. Therefore, the discovery and implementation of novel and broadly applicable therapeutic strategies for these patients are urgently needed. Currently, a number of precision therapeutic approaches have been formulated where molecular analyses of the malignant cells have been used to inform, often multiple, probable targets and potential therapeutic agents. However, a common drawback of this approach has been the uncertainty involved in selecting the drug with the most and clinically relevant cytotoxic potential. In vitro xenograft approaches, although can provide key information on drug activity and side effects, are time consuming and impractical and cumbersome in most cases. We have recently demonstrated the ability of a bone marrow stromal derived cell line to sustain the growth and survival of patient leukemic cells in culture that has allowed in vitro evaluations of drug response.This methodology was combined with a previously validated molecular pathway analysis program to identify effective agents or combinations for a subsequent informed precision clinical trial. Methods: Gene expression profiles from refractory pediatric leukemic cells were analyzed against similar data from more than 12, 000 tumors and outlier analyses were carried out to generate a list of overexpressed genes. This information was computed to identify hypothetically activated pathways, druggable targets and potential agents from a panel of FDA approved drugs. A bone marrow stromal cell line was established and characterized that has been shown to support leukemia cell proliferation in vitro. Briefly, stromal cells were co-cultured with leukemic cells at pre-determined ratios with and without the drugs identified in the genomic analysis. After four days in culture, leukemic cells were re-suspended and analyzed for proliferation. Target modulation and activated cell death pathways were queried by Western blot analyses. Results: Multiple targets and potential agents for effective therapeutics were identified against an initial set of relapsed leukemia specimens. For example, in patient # P700491 (pre B-ALL) gene expression data sets revealed clustering within the area of ALL and AML in the reference cancer genomics data. Comparative tumor RNA seq outlier analysis showed molecular abnormalities in BTK, JAK3 and PIK3CD, corresponding to molecular categories of RTK, JAK-SAT and PI3K-AKT-mTOR pathways targetable by the drugs Ibrutinib, WHI-P131 and Idelalisib, respectively. However, in vitro studies showed significant cell killing with Idelalisib and not with the other two agents. Target modulation assays showed effective induction of apoptosis including PARP cleavage in Idelalisib treated leukemia cells compared to controls, indicating the feasibility of this approach to effectively identify potentially applicable agents for this individual patient. Conclusions: We demonstrate the ability of a newly cloned bone marrow stromal derived cell line to sustain the growth and survival of patient leukemic cells in culture that has allowed in vitro target modulation and target validation analyses for cytotoxicity. This methodology was combined with a previously demonstrated molecular pathway interrogation program to ascertain effective agents or combinations for an experimentally informed precision clinical trial. Importantly, our data showed that genomically identified actionable targets are not universally predictive of tumor response and an in vitro cytotoxicity analysis step may enhance the accuracy of this approach. We describe the practical advantages and versatility of this work-flow to inform the selection of agents in future umbrella trials. It is anticipated that the information obtained will lead to an applicable clinical trial the near future. Disclosures Narendran: Bayer: Honoraria, Other: CANTRK Advisory Board .

Abstract: Introduction: There is strong experimental and clinical data to indicate the critical involvement of immune evasion in relapsed leukemia in children. A well-defined characteristic of refractory leukemia is the accumulation of genetic aberrations and mutations that may act as drivers or passengers in the process of tumor recurrence. Many of these mutations get translated into proteins that contain tumor-specific immune-stimulatory epitopes (neoantigens) that can elicit host antitumor immune responses. Although, in general, the mutation rate is lower in pediatric tumors, recent studies have shown that almost 90% of pediatric leukemias carry potentially actionable neoepitopes. In this study, we describe the results from a comprehensive experimental approach of neoantigen prediction coupled with antigen processing and HLA-binding prediction algorithms with in vitro validation assays for the generation of neoantigen vaccines against high-risk leukemias in children. Methods: DNA and RNA from leukemia cells and matched fibroblasts were obtained. Raw reads were aligned to human reference genome and somatic variants (SNVs) were called using Strelka v1.0.1441. RNA-seq data from leukemic cells were used to predict neoantigen expression levels resulting from SNVs using STAR (2.4.1)12 and Cufflinks v2.2.1. Normalized expression data were then cross-referenced with the list of SNVs to identify leukemia-specific mutant proteins. HLA typing for each sample was carried out from RNA-seq data using seq2HLA v2.2. Using the patient's HLA phenotype, we then used NetMHCons v1.1 to predict short peptides derived from leukemia-specific mutant proteins that will bind to autologous HLA Class I molecules. These 8/9-mers were filtered to predict a high likelihood of proteasomal or immune-proteasomal processing and transporter associated with antigen processing (TAP) using NetChop v3.1 and the immune epitope database (IEDB), respectively. The peptides identified were rank-ordered based on the composite immunogenicity score derived from MHC class I binding affinities, proteasomal processing and TCR binding predictions and synthesized accordingly. Peripheral blood derived dendritic cells (DCs) and CD8+ T-cells were isolated and expanded in culture with relevant cytokines. The DCs were pulsed with peptides and then co-cultured with CD8+ T-cells. After five days, the primed CD8+ T-Cells were separated, washed and exposed to the patient's leukemic cells at varying ratios and the leukemia specific CD8+ T-cell activation was quantified by IFN gamma secretion using ELISpot assays. Results: In the leukemia specimen studied, approximately 5% of all on-target germline mutations were found only in leukemic cells. Tumor mutational burden was, on average, 0.34 mut/Mb. Analysis of the highest ranking synthetic peptides (approximately 10 per leukemia sample) showed leukemia-specific activation of patient's T-cells as measured by the mean number of spots observed in ELISpot assays. For example, in patient one (15 year old male, high-risk ALL, one year off therapy), 14 individual short sequences were identified and corresponding peptides were synthesized. Among these, three peptides were not soluble and three peptides showed significant activity above controls. Maximum leukemia specific T-cell activation was noted with peptide #7 QQSALVLL (mean 135 ELISpots compared to 72 in controls, p & lt;0.05, triplicate) indicating a strong nonantigenic potential in this region. Furthermore, this activity was significantly diminished when an extra amino acid was added to this peptide (LQQSALVLL, mean 79 spots) showing the specificity of the approach. A number of other peptides and combinations in non-overlapping regions gave intermediate activities. Discussion: Completed data, including the vaccine peptide sequences and corresponding activities showed the feasibility of identifying pediatric leukemia neoantigen sequences in personalized mutational landscapes of these patients. In addition, we have provided an in vitro experimental approach to validate the potential of such vaccines in future clinical studies and this methodology can also be used to identify agents for effective combinations such as immune checkpoint inhibitors. A clinical trial using these strategies is in development for the treatment of high-risk leukemia in children. Disclosures No relevant conflicts of interest to declare.