Search

To study the mechanisms of relapse in acute lymphoblastic leukemia (ALL), we performed whole-genome sequencing of 103 diagnosis-relapse-germline trios and ultra-deep sequencing of 208 serial samples in 16 patients. Relapse-specific somatic alterations were enriched in 12 genes (NR3C1, NR3C2, TP53, NT5C2, FPGS, CREBBP, MSH2, MSH6, PMS2, WHSC1, PRPS1, and PRPS2) involved in drug response. Their prevalence was 17% in very early relapse (36 months) groups. Convergent evolution, in which multiple subclones harbor mutations in the same drug resistance gene, was observed in 6 relapses and confirmed by single-cell sequencing in 1 case. Mathematical modeling and mutational signature analysis indicated that early relapse resistance acquisition was frequently a 2-step process in which a persistent clone survived initial therapy and later acquired bona fide resistance mutations during therapy. In contrast, very early relapses arose from preexisting resistant clone(s). Two novel relapse-specific mutational signatures, one of which was caused by thiopurine treatment based on in vitro drug exposure experiments, were identified in early and late relapses but were absent from 2540 pan-cancer diagnosis samples and 129 non-ALL relapses. The novel signatures were detected in 27% of relapsed ALLs and were responsible for 46% of acquired resistance mutations in NT5C2, PRPS1, NR3C1, and TP53. These results suggest that chemotherapy-induced drug resistance mutations facilitate a subset of pediatric ALL relapses.

Children with acute myeloid leukemia (AML) have a dismal prognosis due to a high relapse rate; however, the molecular basis leading to relapsed pediatric AML has not yet been fully characterized. To define the spectrum of alterations common at relapse, we performed integrated profiling of 136 relapsed pediatric AML cases with RNA sequencing (RNA-seq), whole-genome sequencing, and target-capture sequencing. In addition to well-characterized fusion oncoproteins, such as those involving KMT2A (n=36, 26.5%) or NUP98 (n=18, 13.2%), we also identified somatic mutations in UBTF (upstream binding transcription factor) in 12 of 136 cases (8.8%) of this relapsed cohort. Somatic alterations of the UBTF gene, which encodes a nucleolar protein that is a component of the RNA Pol I pre-initiation complex to ribosomal DNA promoters, have rarely been observed in AML. In our cohort, all alterations can be described as heterozygous in-frame exon 13 tandem duplications (UBTF-TD), either at the 3' end of exon 13 of UBTF or of the entire exon 13 (Fig. A). As we noticed limited detection in our pipeline as a result of complex secondary indels alongside the duplications, we established a soft-clipped read-based screening method to detect UBTF-TD more efficiently. Applying the screening to RNA-seq data of 417 additional pediatric AMLs from previous studies and our clinical service, we identified 15 additional UBTF-TDs, many of which have not been previously reported. At the amino acid level, UBTF-TDs caused amino acid insertions of variable sizes (15-181 amino acids), duplicating a portion of high mobility group domain 4 (HMG4), which includes short leucine-rich sequences. UBTF-TD AMLs commonly occurred in early adolescence (median age: 12.6, range: 2.4-19.6), and 19 of the total 27 cases had either normal karyotype (n=12) or trisomy 8 (n=7). UBTF-TD is mutually exclusive from other recurrent fusion oncoproteins, such as NUP98 and KMT2A rearrangements (Fig. B), but frequently occurred with FLT3-ITD (44.4%) or WT1 mutations (40.7%). The median variant allele fraction (VAF) of the UBTF-TD was 48.0% (range: 9.7-66.7%). In four cases with data at multiple disease time points, the identical UBTF-TDs were present at high allele fractions at all time points, suggesting that UBTF-TD is a clonal alteration. tSNE analysis of the transcriptome dataset showed that UBTF-TD AMLs share a similar expression pattern with NPM1 mutant and NUP98-NSD1 AML subtypes, including NKX2-3 and HOXB cluster genes (Fig. C) . Altogether, these findings suggest that UBTF-TD is a unique subtype of pediatric AML. To address the impact of UBTF-TD expression in primary hematopoietic cells, we introduced UBTF-TD and UBTF wildtype expression vectors into cord blood CD34+ cells via lentiviral transduction. UBTF-TD expression promotes colony-forming activity and cell growth, yielding cells with a persistent blast-like morphology (Fig. D). Further, transcriptional profiling of these cells demonstrated expression of HOXB genes and NKX2-3, similar to UBTF-TD AMLs in patients, indicating that UBTF-TD is sufficient to induce the leukemic phenotype. To investigate the prevalence of UBTF-TDs in larger de novo AML cohorts, we applied the above UBTF-TD screening method to the available de novo AML cohorts of TCGA (n=151, adult), BeatAML (n=220, pediatric and adult), and AAML1031 (n=1035, pediatric). We identified UBTF-TDs in 4.3% (45/1035) of the pediatric AAML1031 cohort, while the alteration is less common (0.9%: 3/329, p=0.002) in the adult AML cohorts (Fig. E). In the AAML1031 cohort, UBTF-TDs remain mutually exclusive with known molecular subtypes of AML and commonly occur with FLT3-ITD (66.7%) and WT1 (40.0%) mutations and either normal karyotype or trisomy 8. The presence of UBTF-TDs in the AAML1031 cohort is associated with a poor outcome (Fig. F, median overall survival, 2.3 years) and MRD positivity; multivariate analysis revealed that UBTF-TD and WT1 are independent risk factors for overall survival within FLT3-ITD+ AMLs. In conclusion, we demonstrate UBTF-TD defines a unique subtype of AMLs that previously lacked a clear oncogenic driver. While independent of subtype-defining oncogenic fusions, UBTF-TD AMLs are associated with FLT3-ITD and WT1 mutations, adolescent age, and poor outcomes. These alterations have been under-recognized by standard bioinformatic approaches yet will be critical for future risk-stratification of pediatric AML. Figure 1 Figure 1. Disclosures Iacobucci: Amgen: Honoraria; Mission Bio: Honoraria. Miller: Johnson & Johnson's Janssen: Current Employment. Mullighan: Pfizer: Research Funding; Illumina: Membership on an entity's Board of Directors or advisory committees; AbbVie: Research Funding; Amgen: Current equity holder in publicly-traded company.

Genetic rearrangements involving the KMT2A gene (KMT2A-R) are seen in around 10% of acute leukemia overall. KMT2A-R occurs in all ages and usually correlates with high-risk clinical features, in particular in infants aged 0-12 months of age with acute lymphoblastic leukemia (ALL). To uncover age- and leukemia-subtype specific molecular patterns in KMT2A-R ALL and acute myeloid leukemia (AML), we performed whole genome (WGS), whole exome (WES), and RNA-sequencing on a well-annotated Nordic KMT2A-R cohort of 104 patients, including infant ALL (n=33), childhood ALL (n=18), adult ALL (n=15), and pediatric AML (n=38) patients. For 77 patients, we performed WGS (40x) at diagnosis and remission as well as WES (140x) on the diagnostic sample, and remaining patients underwent WES only (n=27). RNA-sequencing was performed on 58 cases with available RNA. Twenty-two genes were recurrently altered and remarkably, NRAS, KRAS, FLT3, PAX5, TP53, CDKN2A/B and IKZF1 accounted for 70% of mutations. The landscape of mutations suggested the presence of leukemia and age-specific associations with MYST4, PTPN11, and SETD2 uniquely altered in AML and PIK3CD, DNAH11, NOTCH1, CSMD3 and CDKN2A/B in ALL. Some genes were mutated in both KMT2A-R ALL and AML, but were more common in one disease, such as FLT3 and KRAS in AML and PAX5, TP53 and IKZF1 in ALL. Moreover, age-associated patterns were seen in ALL with NRAS more frequently mutated than KRAS in infant ALL (26% vs 15%), and KRAS more frequently mutated than NRAS in childhood ALL (24% vs 18%), with adult ALL having fewer such mutations (NRAS 13%; KRAS 7%). Alterations of CDKN2A/B and TP53 were absent in infant ALL, detected in childhood and adult ALL only. PAX5 alterations were primarily detected in childhood ALL (22%, 9% infant ALL, 7% adult ALL), with all three PAX5-altered infant cases having the KMT2A-MLLT3 fusion gene. Finally, KMT2A-R pediatric AML had the highest fraction of FLT3 mutations (24%, 9% infant ALL, 11% childhood ALL, 0% adult ALL) and all but one mutation occurred in KMT2A-MLLT3 rearranged cases and most were kinase domain point mutations. We next expanded our analysis to include non-recurrent alterations. PI3K/RAS pathway alterations were detected across ages and subtypes with the highest fraction in pediatric AML (63%) and the lowest in adult ALL (27%, 43% infant ALL, 41% childhood ALL). Further, cell cycle related genes were primarily mutated in childhood (39%) and adult ALL cases (33%) and rarer in infant ALL (12%) and pediatric AML (16%) and genes within the B-cell pathway were more commonly altered in childhood ALL (29%) than in infant ALL (9%). Finally, in line with our previous study (Andersson et al, Nat Genet 2015) epigenetic mutations were absent in infant ALL, but present in 20-35% of the other patients. RNA-sequencing identified the KMT2A-fusion in 56/58 cases, with low exonic coverage preventing detection of the fusion in two cases. The reciprocal KMT2A fusion was only expressed in 13/39 cases where it was predicted to be expressed based on karyotype or whole genome sequencing data with 11/13 cases having the KMT2A-AFF1 fusion gene. In addition, RNA-sequencing identified 6 in-frame and 12 out-of-frame fusion genes that had formed either as part of the KMT2A-R itself or that were independent genetic events. Further, a novel in-frame KMT2A-ACIN1 fusion was identified in a child aged 1 year with B-precursor ALL. ACIN1 encodes Apoptotic Chromatin Condensation Inducer 1 and the fusion was formed through an insertion of 14q11 into 11q23. ACIN1 is also rearranged as part of the ACIN1-NUTM1 that we identified in an infant with ins(15;14)(q22;q11.2q32.1) (Andersson et al Nat Genet 2015). To study the ability of KMT2A-ACIN1 to induce leukemia in mice, we injected retrovirally transduced mouse bone marrow cells containing the fusion into syngeneic mice and KMT2A-MLLT3 was used as control. All mice succumbed to disease at an average of 112 days for KMT2A-ACIN1 (n=12) and 63 days for KMT2A-MLLT3 (n=5) and mice displayed splenomegaly and leukocytosis with an immunophenotype indicative of AML. Primary leukemia cells isolated from moribund mice gave rise to leukemia in sublethally irradiated recipients with reduced disease latency. In conclusion, these results highlight the differential molecular patterns in KMT2A-R leukemia across infancy to adulthood thereby providing novel pathogenetic insight. Disclosures No relevant conflicts of interest to declare.

Alterations of genes encoding transcriptional regulators of lymphoid development are a hallmark of B-progenitor acute lymphoblastic leukemia (B-ALL) and most commonly involve PAX5, encoding the DNA-binding transcription factor paired-box 5. The majority of PAX5 alterations in ALL are heterozygous, and key PAX5 target genes are expressed in leukemic cells, suggesting that PAX5 may be a haploinsufficient tumor suppressor. To examine the role of PAX5 alterations in leukemogenesis, we performed mutagenesis screens of mice heterozygous for a loss-of-function Pax5 allele. Both chemical and retroviral mutagenesis resulted in a significantly increased penetrance and reduced latency of leukemia, with a shift to B-lymphoid lineage. Genomic profiling identified a high frequency of secondary genomic mutations, deletions, and retroviral insertions targeting B-lymphoid development, including Pax5, and additional genes and pathways mutated in ALL, including tumor suppressors, Ras, and Janus kinase-signal transducer and activator of transcription signaling. These results show that in contrast to simple Pax5 haploinsufficiency, multiple sequential alterations targeting lymphoid development are central to leukemogenesis and contribute to the arrest in lymphoid maturation characteristic of ALL. This cross-species analysis also validates the importance of concomitant alterations of multiple cellular growth, signaling, and tumor suppression pathways in the pathogenesis of B-ALL.

RUNX1 is an important transcription factor for hematopoiesis. There are multiple alternatively spliced isoforms of RUNX1. The best known isoforms are RUNX1a from use of exon 7A and RUNX1b and c from use of exon 7B. RUNX1a has unique functions due to its lack of C-terminal regions common to RUNX1b and c. Here, we report that the ortholog of human RUNX1a was only found in primates. Furthermore, we characterized 3 Runx1 isoforms generated by exon 6 alternative splicing. Runx1bEx6(-) (Runx1b without exon 6) and a unique mouse Runx1bEx6e showed higher colony-forming activity than the full-length Runx1b (Runx1bEx6(+)). They also facilitated the transactivation of Runx1bEx6(+). To gain insight into in vivo functions, we analyzed a knock-in (KI) mouse model that lacks isoforms Runx1b/cEx6(-) and Runx1bEx6e. KI mice had significantly fewer lineage-Sca1(+)c-Kit(+) cells, short-term hematopoietic stem cells (HSCs) and multipotent progenitors than controls. In vivo competitive repopulation assays demonstrated a sevenfold difference of functional HSCs between wild-type and KI mice. Together, our results show that Runx1 isoforms involving exon 6 support high self-renewal capacity in vitro, and their loss results in reduction of the HSC pool in vivo, which underscore the importance of fine-tuning RNA splicing in hematopoiesis.

Introduction: Cytarabine, daunorubicin and etoposide (ADE) are commonly used for remission and intensification of pediatric acute myeloid leukemia (AML). However, development of drug resistance is a major cause of treatment failure. In this study, we performed a comprehensive evaluation of expression levels of genes of pharmacological significance (pharmacokinetic /pharmacodynamic) to ADE and derived a drug response score predictive of treatment outcomes in pediatric AML patients. Methods: This study included 163 cases (median age=8.79 year, range= (0.013-21.1)) with AML enrolled in the multicenter AML02 clinical trial (ClinicalTrials.gov Identifier: NCT00136084) with Affymetrix U133A microarray gene expression and clinical data available. We used a penalized LASSO regression algorithm (glmnet R-package) to fit a cox regression model on diagnostic leukemic cell gene expression levels of 66 genes of pharmacological significance to ADE. We performed 1000 bootstraps of LASSO regression with event free survival (EFS) as the outcome variable and the five genes represented in at least 95% of the models were included to build an ADE-Response Score (ADE-RS) equation. Patients were classified into low or high score groups using recursive portioning implemented in Rpart-R package and evaluated for association with minimal residual disease after induction I (MRD1), EFS and overall survival (OS). ADE response score equation was further validated using RNA-Seq gene-expression data obtained from diagnostic samples of 432 pediatric AML patients enrolled in Children's Oncology Group (COG) AAML0531 and AAML03P1 treatment protocols. Results: After applying LASSO regression, we defined the equation: ADE-RS = (0.128 x DCTD) - (0.0993 x TOP2A) + (0.212 x ABCC1) - (0.113 x MPO) - (0.126 x CBR1) to develop ADE-response score (ADE-RS), followed by classifying patients into low (60%; 98 patients) or high (40%; 65 patients) score groups. Patients in the high ADE-RS group had significantly worse EFS (HR=4.07(2.43-6.84), P < 0.0001; Figure 1A) and OS (HR= 4.54(2.42-8.49), P In a multivariable cox-regression models that included pLSC6-ADE response score groups, MRD1 status, risk groups, WBC at diagnosis and age in AML02 cohort, high pLSC6-ADE score group was found significantly associated with poor EFS (HR=6.02(2.71-13.2), P Discussion: In this study, we defined a pharmacological response score focused on key genes of PK/PD significance to ADE. We further integrated LSC score with the ADE response score to improve our ability to predict treatment outcome in AML patients across different clinical trials. ADE-RS was composed of five genes: DCTD, which is a deaminase, involved in ara-C inactivation; CBR1, a carbonyl reductase involved in inactivation of daunorubicin (DNR); MPO, myeloperoxidase, an etoposide activator; ABCC1, an efflux transporter of DNR and etoposide; and TOP2A, DNA topoisomerase II alpha, which is a target for DNR and etoposide. Integrated pLSC6 and ADE-RS has a potential to predict treatment outcomes using diagnostic gene expression levels and accordingly develop treatment strategies to improve treatment outcome. Figure 1 Disclosures Gruber: Bristol-Myers Squibb: Consultancy. Rubnitz:AbbVie: Research Funding.

Acute myeloid leukemia (AML) comprises a heterogeneous group of malignancies that are linked by the presence of blasts displaying morphologic and immunophenotypic features of myeloid cell differentiation. With the development of genome-wide gene expression profiling (GEP), array-base comparative genomic hybridization methodologies, and next generation sequencing technologies, the field has gained a greater understanding of the molecular features of this malignancy. Several pathologic lesions have been found to have prognostic implications contributing to a continuous refinement of risk stratification in the context of modern therapy. Recently, the Children's Oncology Group (COG)-National Cancer Institute (NCI) TARGET AML initiative molecularly characterized 993 pediatric AML cases including 197 specimens that underwent comprehensive whole genome sequencing. Of these, 94 carried one of three oncogenic fusions known to be strong drivers of leukemogenesis: RUNX1-RUNX1T1, CBFB-MYH11 and KMT2A rearrangements (KMT2Ar). Among all the alterations detected only ten occurred in more than 5% of subjects, all of which had been previously described. This suggested that low-frequency molecular subsets may exist that require larger cohorts to fully elucidate. To address this limitation, we selected 122 pediatric AML specimens that lacked RUNX1-RUNX1T1, CBFB-MYH11 and KMT2Ar by clinical testing for whole genome (WGS), exome (WES) and RNA (RNAseq) sequencing to enrich for cases that carry low-frequency events. GEP coupled with somatic mutation calls and outcome data were utilized to identify distinct molecular subtypes with prognostic implications. Structural variations, copy number alterations, single nucleotide variations and indels were determined by our established pipelines, as well as an evaluation for regulatory rearrangements driving oncogene overexpression through enhancer hijacking. In addition to known AML somatic mutations and rearrangements in genes such as CEBPA, GATA2, NPM1, WT1, FLT3, NRAS, KRAS, ETV6, Cohesin, NUP98 and KAT6A, we identified rare novel events in known oncogenic drivers. These include a GATA2-ITD as well as the repositioning of a distal MYC enhancer to ectopically activate either the MECOM or BCL11B loci. Interestingly, several AML cases carrying loss of function mutations in polycomb repressive complex 2 (PRC2) genes were found to resemble an early T-cell precursor acute lymphoblastic leukemia (ETP-ALL) GEP by gene set enrichment analysis. ETP-ALL exhibits aberrant expression of stem cell and myeloid markers and has been shown to have a GEP consistent with transformation of a stem cell progenitor. Further, mixed phenotype acute leukemias (MPAL) with T and myeloid lineage characteristics have been previously suggested to be in this spectrum of immature leukemias. We therefore hypothesized that these PRC2-mutated AML cases represented the myeloid end of this continuum. To provide global transcriptional context to these ETP-like AMLs and evaluate a comprehensive cohort encompassing a range of pediatric myeloid malignancies, we integrated results from previously published AML, MPAL, acute megakaryoblastic Leukemia (AMKL), and ETP-ALL datasets that had RNAseq and either WES or WGS available for a total of 436 cases. t-SNE visualization using a 381 gene list derived from the top 100 most variably expressed transcripts within each cohort revealed a clear molecular classifier identifying groups that had consistent mutational compositions and disease outcomes but were agnostic of immunophenotype. This approach allowed the distinction of 63 ETP-like cases comprising a mixture of AML, MPAL, and ETP-ALL leukemias which fell into two subgroups distinguished by FLT3-ITD and PRC2 alterations. Irrespective of treatment approach, FLT3-ITD positive ETP-like leukemias enjoyed a favorable outcome whereas those with PRC2 mutations had a poor prognosis. Our data support a refined classification of pediatric myeloid malignancies based on molecular determinants that can be used for risk stratification in therapeutic trials. Disclosures Gruber: Bristol-Myers Squibb: Consultancy. Rubnitz:AbbVie: Research Funding. Reinhardt:Jazz: Other: Participation in Advisory Boards, Research Funding; CSL Behring: Research Funding; Novartis: Other: Participation in Advisory Boards; Roche: Research Funding. Locatelli:bluebird bio: Consultancy; Bellicum: Consultancy, Membership on an entity's Board of Directors or advisory committees; Amgen: Honoraria, Membership on an entity's Board of Directors or advisory committees; Miltenyi: Honoraria; Novartis: Consultancy, Membership on an entity's Board of Directors or advisory committees. Zwaan:Roche: Consultancy; Servier: Consultancy; Daiichi Sankyo: Consultancy; Novartis: Consultancy; Sanofi: Consultancy; Pfizer: Consultancy, Research Funding; BMS: Research Funding; Celgene: Consultancy, Research Funding; Jazz pharmaceuticals: Other: Travel support; Janssen: Consultancy; Incyte: Consultancy.

Introduction: Although recent studies have refined the classification of B-progenitor and T-lineage acute lymphoblastic leukemia into gene-expression based subgroups, a comprehensive integration of significantly mutated genes and pathways for each subgroup is needed to understand disease etiology. Methods: We studied 2789 children, adolescents and young adults (AYA) with newly diagnosed B-ALL (n=2,322 cases) or T-ALL (n=467) treated on Children's Oncology Group (n=1,872) and St. Jude Children's Research Hospital trials (n=917). The cohort comprised childhood NCI standard-risk (41.8%; age range 1-9.99 yrs, WBC ≤ 50,000/ml), childhood NCI high-risk (44.5%; age range ≥10 to 15.99 yrs) and AYA (9.9%; age range 16-30.7 yrs). Genomic analysis was performed on tumor and matched-remission samples using whole transcriptome sequencing (RNA-seq; tumor only; n=1,922), whole exome sequencing (n=1,659), whole genome sequencing (n=757), and single nucleotide polymorphism array (n=1,909). Results: For B-ALL, 2104 cases (90.6%) were classified into 26 subgroups based on RNA-seq gene expression data and aneuploidy or other gross chromosomal abnormalities (iAMP21, Down syndrome, dicentric), deregulation of known transcription factors by rearrangement or mutation (PAX5 P80R, IKZF1 N159Y), or activation of kinase alterations (Ph+, Ph-like). For T-ALL, cases were classified into 9 previously described subtypes based on dysregulation of transcription factor genes and gene expression. In 1,659 cases subject to exome sequencing (1259 B-ALL, 405 T-ALL) we identified 18,954 nonsynonymous single nucleotide variants (SNV) and 2,329 insertion-deletion mutations (indels) in 8,985 genes. Overall, 161 potential driver genes were identified by the mutation-significance detection tool MutSigCV or by presence of pathogenic variants in known cancer genes. Integration of sequence mutations and DNA copy number alteration data in B-ALL identified 7 recurrently mutated pathways: transcriptional regulation (40.6%), cell cycle and tumor suppression (38.0%), B-cell development (34.5%), epigenetic regulation (24.7%), Ras signaling (33.0%), JAK-STAT signaling (12.0%) and protein modification (ubiquitination or SUMOylation, 5.0%). The top 10 genes altered by deletion or mutation in B-ALL were CDKN2A/B (30.1%), ETV6 (27.0%), PAX5 (24.6%), CDKN1B (20.3%), IKZF1 (17.6%), KRAS (16.5%), NRAS (14.6%), BTG1 (7.5%) histone genes on chromosome 6 (6.9%) and FLT3 (6.1%), and for T-ALL, CDKN2A/B (74.7%), NOTCH1 (68.2%), FBXW7 (21.3%), PTEN (20.5%) and PHF6 (18.2%) (Figure 1A). We identified 17 putative novel driver genes involved in ubiquitination (UBE2D3, UBE2A, UHRF1, and USP1), SUMOylation (SAE1, UBE2I), transcriptional regulation (ZMYM2, HMGB1), immune function (B2M), migration (CXCR4), epigenetic regulation (DOT1L) and mitochondrial function (LETM1). We also observed variation in the frequency of genes and pathways altered across B-ALL subtypes (Figure 1B). Interestingly, alteration of SAE1 and UBA2, novel genes that form a heterodimeric complex important for SUMOylation, and UHRF1 were enriched in ETV6-RUNX1 cases. Deletions of LETM1, ZMYM2 and CHD4 were associated with near haploid and low hypodiploid cases. Deletion of histone genes on chromosome 6 and alterations of HDAC7 were enriched in Ph+ and Ph-like ALL. Mutations in the RNA-binding protein ZFP36L2 were observed in PAX5alt, DUX4 and MEF2D subgroups. Genomic subtypes were prognostic. ETV6-RUNX1, hyperdiploid, DUX4 and ZNF384 ALL were associated with good outcome (5-yr EFS 91.1%, 87.2%, 91.9% and 85.7%, respectively), ETV6-RUNX1-like, iAMP21, low hyperdiploid, PAX5 P80R and PAX5alt were associated with intermediate outcome (5-yr EFS 68.6%, 72.2%, 70.8%, 77.0% and 70.9%, respectively), whilst KMT2A, MEF2D, Ph-like CRLF2 and Ph-like other conferred a poor prognosis (55.5%, 67.1%, 51.5% and 62.1%, respectively). TCF3-HLF and near haploid had the worst outcome with 5-yr EFS rates of 27.3% and 47.2%, respectively. Conclusions: These findings provide a comprehensive landscape of genomic alterations in childhood ALL. The associations of mutations with ALL subtypes highlights the need for specific patterns of cooperating mutations in the development of leukemia, which may help identify vulnerabilities for therapy intervention. Disclosures Gastier-Foster: Bristol Myers Squibb (BMS): Other: Commercial Research; Incyte Corporation: Other: Commercial Research. Willman:to come: Patents & Royalties; to come: Membership on an entity's Board of Directors or advisory committees; to come: Research Funding. Raetz:Pfizer: Research Funding. Borowitz:Beckman Coulter: Honoraria. Zweidler-McKay:ImmunoGen: Employment. Angiolillo:Servier Pharmaceuticals: Consultancy. Relling:Servier Pharmaceuticals: Research Funding. Hunger:Jazz: Honoraria; Amgen: Consultancy, Equity Ownership; Bristol Myers Squibb: Consultancy; Novartis: Consultancy. Loh:Medisix Therapeutics, Inc.: Membership on an entity's Board of Directors or advisory committees. Mullighan:Amgen: Honoraria, Other: speaker, sponsored travel; Loxo Oncology: Research Funding; AbbVie: Research Funding; Pfizer: Honoraria, Other: speaker, sponsored travel, Research Funding; Illumina: Honoraria, Membership on an entity's Board of Directors or advisory committees, Other: sponsored travel.

Reversing the aberrant biochemical output of oncogenic Ras proteins is one of the great challenges in cancer therapeutics; however, it is uncertain which Ras effectors are required for tumor initiation and maintenance. To address this question, we expressed oncogenic K-Ras(D12) proteins with "second site" amino acid substitutions that impair PI3 kinase/Akt or Raf/MEK/ERK activation in bone marrow cells and transplanted them into recipient mice. In spite of attenuated signaling properties, defective K-Ras oncoproteins initiated aggressive clonal T-lineage acute lymphoblastic leukemia (T-ALL). Murine T-ALLs expressing second site mutant proteins restored full oncogenic Ras activity through diverse mechanisms, which included acquiring novel somatic third site Kras(D12) mutations and silencing PTEN. T-ALL cell lines lacking PTEN had elevated levels of phosphorylated Akt, a gene expression pattern similar to human early T-cell precursor ALL, and were resistant to the potent and selective MEK inhibitor PD0325901. Our data, which demonstrate strong selective pressure to overcome the defective activation of PI3 kinase/Akt and Raf/MEK/ERK, implicate both Ras effector pathways as drivers of aberrant growth in T-ALL and further suggest that leukemia cells will deploy multiple mechanisms to develop resistance to targeted inhibitors in vivo.

One recently identified subtype of pediatric B-precursor acute lymphoblastic leukemia (ALL) has been termed BCR-ABL1–like or Ph-like because of similarity of the gene expression profile to BCR-ABL1 positive ALL suggesting the presence of lesions activating tyrosine kinases, frequent alteration of IKZF1, and poor outcome. Prior studies demonstrated that approximately half of these patients had genomic lesions leading to CRLF2 overexpression, with half of such cases harboring somatic mutations in the Janus kinases JAK1 and JAK2. To determine whether mutations in other tyrosine kinases might also occur in ALL, we sequenced the tyrosine kinome and downstream signaling genes in 45 high-risk pediatric ALL cases with either a Ph-like gene expression profile or other alterations suggestive of activated kinase signaling. Aside from JAK mutations and 1 FLT3 mutation, no somatic mutations were found in any other tyrosine kinases, suggesting that alternative mechanisms are responsible for activated kinase signaling in high-risk ALL.

Background As new technologies allow investigators to collect multiple forms of molecular data (genomic, epigenomic, transcriptomic, etc) and multiple endpoints on a clinical trial cohort, it will become necessary to effectively integrate all these data in a way that reliably identifies biologically important genes. Methods We introduce CC-PROMISE as an integrated data analysis method that combines components of canonical correlation (CC) and projection onto the most interesting evidence (PROMISE). For each gene, CC-PROMISE first uses CC to compute scores that represent the association of two forms of molecular data with each other. Next, these scores are substituted into PROMISE to evaluate the statistical evidence that the molecular data show a biologically meaningful relationship with the endpoints. Results CC-PROMISE shows outstanding performance in simulation studies and an example application involving pediatric leukemia. In simulation studies, CC-PROMISE controls the type I error (misleading significance) rate very near the nominal level across 100 distinct null settings in which no molecular-endpoint association exists. Also, CC-PROMISE has better statistical power than three other methods that control type I error in 396 of 400 (99 %) alternative settings for which a molecular-endpoint association is present; the power advantage of CC-PROMISE exceeds 30 % in 127 of the 400 (32 %) alternative settings. These advantages of CC-PROMISE are also observed in an example application. Conclusion CC-PROMISE very effectively identifies genes for which some form of molecular data shows a biologically meaningful association with multiple related endpoints. Availability The R package CCPROMISE is currently available from www.stjuderesearch.org/site/depts/biostats/software. Electronic supplementary material The online version of this article (doi:10.1186/s12859-016-1217-0) contains supplementary material, which is available to authorized users.

We sequenced 120 candidate genes in 187 high-risk childhood B-precursor acute lymphoblastic leukemias, the largest pediatric cancer genome sequencing effort reported to date. Integrated analysis of 179 validated somatic sequence mutations with genome-wide copy number alterations and gene expression profiles revealed a high frequency of recurrent somatic alterations in key signaling pathways, including B-cell development/differentiation (68% of cases), the TP53/RB tumor suppressor pathway (54%), Ras signaling (50%), and Janus kinases (11%). Recurrent mutations were also found in ETV6 (6 cases), TBL1XR1 (3), CREBBP (3), MUC4 (2), ASMTL (2), and ADARB2 (2). The frequency of mutations within the 4 major pathways varied markedly across genetic subtypes. Among 23 leukemias expressing a BCR-ABL1-like gene expression profile, 96% had somatic alterations in B-cell development/differentiation, 57% in JAK, and 52% in both pathways, whereas only 9% had Ras pathway mutations. In contrast, 21 cases defined by a distinct gene expression profile coupled with focal ERG deletion rarely had B-cell development/differentiation or JAK kinase alterations but had a high frequency (62%) of Ras signaling pathway mutations. These data extend the range of genes that are recurrently mutated in high-risk childhood B-precursor acute lymphoblastic leukemia and highlight important new therapeutic targets for selected patient subsets.

To identify new markers for minimal residual disease (MRD) detection in acute lymphoblastic leukemia (ALL), we compared genome-wide gene expression of lymphoblasts from 270 patients with newly diagnosed childhood ALL to that of normal CD19+CD10+ B-cell progenitors (n = 4). Expression of 30 genes differentially expressed by ≥ 3-fold in at least 25% of cases of ALL (or 40% of ALL subtypes) was tested by flow cytometry in 200 B-lineage ALL and 61 nonleukemic BM samples, including samples containing hematogones. Of the 30 markers, 22 (CD44, BCL2, HSPB1, CD73, CD24, CD123, CD72, CD86, CD200, CD79b, CD164, CD304, CD97, CD102, CD99, CD300a, CD130, PBX1, CTNNA1, ITGB7, CD69, CD49f) were differentially expressed in up to 81.4% of ALL cases; expression of some markers was associated with the presence of genetic abnormalities. Results of MRD detection by flow cytometry with these markers correlated well with those of molecular testing (52 follow-up samples from 18 patients); sequential studies during treatment and diagnosis-relapse comparisons documented their stability. When incorporated in 6-marker combinations, the new markers afforded the detection of 1 leukemic cell among 105 BM cells. These new markers should allow MRD studies in all B-lineage ALL patients, and substantially improve their sensitivity.

Introduction: Resistance and relapse remain major obstacles in the treatment of acute myeloid leukemia (AML). Pre-existence and persistence of drug resistant leukemia stem cells (LSCs) is considered one of the major causes of relapse. A previous study (Ng et al., 2016) reported a prognostic signature of 17 genes (LSC17 score) differentially expressed in LSC+ compared to LSC- cell fractions that predicted outcome in patients with AML thereby classifying patients into high and low risk groups. The goal of this study is to determine the validity of LSC17 score in pediatric AML patients and to enhance its clinical utility by exploring a new score with limited number of stem cell genes. Methods: 150 pediatric patients with AML enrolled in the multicenter AML02 clinical trial (ClinicalTrials.gov Identifier: NCT00136084) with Affymetrix U133A microarray gene expression data and clinical data were included in the study. Since only 14 of the 17 genes were represented on the Affymetrix U133A gene chip we tested the validity of the LSC14 score using the previously defined equation (Ng et al, 2016) with multiple clinical endpoints such as minimum residual disease (MRD), event free survival (EFS) and overall survival (OS). To reduce the model complexity, we applied a penalized regression algorithm called the least absolute shrinkage and selection operator (LASSO) implemented in the glmnet R-package using event free survival (EFS) as an outcome variable. Score of the new equation, which included three genes, was designated as pediatric-LSC3 (pLSC3). pLSC3 was tested in the AML02 cohort for association of high or low pLSC3 (based on the median value) with clinical endpoints mentioned above. pLSC3 score equation was validated using publically available gene-expression data from 117 pediatric relapse enriched AML patient cohort enrolled in Children's Oncology Group (COG) protocol (TARGET database). COX-proportional hazard models and Log rank test were used for survival data analysis. Results: AML02 cohort: Patients with high LSC14 scores (greater than median), had significantly worse MRD (p In a multivariate COX regression model, pLSC3 score groups was the only significant covariate (table 1). It explained 13.1% of variability in EFS and 11.6% of variability in OS, while other prognostic factors such as risk groups, FLT3 status, treatment arm and age collectively explained 15.1 and 12.1 % of variability. Discussion: In summary, our results show validity of a previously defined LSC14 score in a pediatric AML population from the multicenter AML02 clinical trial. To enhance the clinical utility, score equation was further simplified and the final score (pLSC3) was derived from three genes: DNMT3B, which encodes for DNA methyltransferase; CD34, an important cell surface marker for early-undifferentiated LSCs; and GPR56, a G protein coupled receptor of significance in AML. Given that there is need to refine classification of a highly heterogeneous group of patients with standard risk AML, we show that differentiating standard risk patients based on pLSC3 score should be considered in the future. We show the relevance of pLSC3 in two independent cohorts, opening up opportunities to improve treatment outcomes of pediatric patients with AML. Disclosures No relevant conflicts of interest to declare.

With the increase in availability of high depth whole genome sequencing (WGS) data of individuals with sickle cell disease (SCD), easy access to the raw sequencing data remains an issue due to technical and regulatory challenges. A compliant system that can provide facile data access would accelerate scientific discovery of genetic variants associated with clinical phenotypes. Cloud storage and computing provide an ultimate solution to this data access, which we have shown through the St. Jude Cloud (https://stjude.cloud) where over 5000 whole genome sequences for pediatric cancer patients are being shared in collaboration with DNANexus and Microsoft. Here we expand the St. Jude Cloud to sickle cell disease data through the Sickle Genome Project (SGP) Data Portal (https://pecan.stjude.org/permalink/sgp) to allow instantaneous raw data access (following data access committee approval), as well as visualization of genotype calls at individual level in a novel genome. The SGP WGS data was generated from 871 patients from St. Jude Children's Research Hospital (St. Jude) through the Sickle Cell Clinical Research and Intervention Program (SCCRIP, Pediatr Blood Cancer. 2018 May 24: e27228) and from Baylor College of Medicine (BCM). All study participants provided informed consent for genomic study and data sharing on IRB-approved research protocols. The SGP data portal will have multi-tiered access. All users will have access to a general heat map view which shows anonymized patient clinical values (e.g., fetal hemoglobin (HbF), mean corpuscular volume (MCV), hemoglobin concentration (Hb)) and relevant SCD modifying variants (e.g., Beta-globin locus, MYB, BCL11A, HBA). The GenomePaint bowser allows for viewing coding and noncoding variants. Displayed with each variant will be a visual indication of the median fetal hemoglobin values for patients homozygous for the reference allele, heterozygous, or homozygous for the alternative allele. The browser also displays erythroid specific DNA-accessibility and epigenetic marks and indicates variant that may disrupt erythroid specific transcription factor binding sites (GATA1 and BCL11A). For anonymization purposes, within the genome browser and heat map views, clinical values and the patients age will be binned into ranges when displayed as single or low count values. Lastly, the ProteinPaint tool (Zhang and Zhou, Nature Gen, Dec 29, 2015) will enable visualization and filtering of variants with reference to protein domain and amino acid sequence. To access processed data such as BAM and VCF files for downstream analyses, a user will be required to apply for access which will be adjudicated by a data access committee. Verified researchers will be granted access to clinical data in a manner consistent with the protocol specific informed consent documentation and protocol under which the sequencing was performed. This may include coded clinical and demographic data when specified by the research protocol and informed consent The SGP data set will be one of the first WGS datasets from primarily African American Sickle cell patients to be made available to clinicians and researchers worldwide. In addition, no SCD-centric data portal exists that contains controlled access to data and provides graphical tools for visual analysis. The combination of the visual tools and ability to download tools provides the scientific community an invaluable resource for studying sickle cell disease. Figure. Figure. Disclosures Estepp: Daiichi Sankyo: Consultancy; NHLBI: Research Funding; Global Blood Therapeutics: Consultancy, Research Funding; ASH Scholar: Research Funding. Hankins:Global Blood Therapeutics: Research Funding; bluebird bio: Consultancy; NCQA: Consultancy; Novartis: Research Funding.

Introduction B progenitor acute lymphoblastic leukemia (B-ALL) is a leading cause of childhood cancer death. Many chimeric genes have been identified and led to a refined classification of B-ALL and tailored therapies. Still, up to 30% of B-ALL cannot be classified into established subtypes, and the outcome for many is poor. Methods To identify novel subtypes of B-ALL, we performed integrative genomic analysis including transcriptome sequencing (RNA-seq) of 1,988 cases from St. Jude, Children's Oncology Group and adult cooperative group studies and analyzed chromosomal rearrangements, gene-expression profiles (GEP), somatic mutations and chromosome-level copy-number alterations. Cases lacking known or putative subtype-defining alterations underwent whole genome sequencing. Effects on proliferation and transformation of novel lesions were assessed by retroviral expression in cell lines and point-mutation knock-in mice using CRISPR/Cas9 genome editing. Results Using integrated genetic alterations and gene expression profiling, we classified 23 B-ALL subtypes (Table and Figure). Three groups included cases with similar GEP as canonical subtypes (ETV6-RUNX1, KMT2A-rearranged, and ZNF384-rearranged), but lacking the expected drivers (e.g., ETV6-RUNX1-like, n=42). Eighteen cases (0.9%) had rearrangements of BCL2, MYC and/or BCL6 showing a distinct GEP; they were mostly adults (n=16) with very poor outcome. These alterations, rarely seen in ALL, are identical to those observed in "double/triple hit" lymphoma, and are of pre-B immunophenotype. Eight cases with tightly clustered GEP comprised a novel subtype defined by IKZF1 N159Y missense mutation. N159Y is in the DNA-binding domain of IKZF1, and is known to perturb IKZF1 function, with distinct nuclear mislocalization and induction of aberrant intercellular adhesion. We identified two subtypes with distinct GEP characterized by PAX5 alterations. One, herein termed PAX5 altered (PAX5alt), comprised 148 (7.4%) cases, was characterized by diverse PAX5 alterations including rearrangements (n=57), sequence mutations (n=46) and/or focal intragenic amplifications (n=8). These PAX5 alterations were found in 73.6% of PAX5alt cases and different alteration types were mutually exclusive. Other PAX5 alterations, including deletions and large-scale amplifications were also assessed using SNP array, but were not enriched in the PAX5alt group. Clinically, PAX5alt pediatric and adult patients had favorable (96.8±3.2%) and intermediate (42.1±10.2%) 5-year overall survival (OS), respectively. The other GEP distinct subtype comprised 44 cases, all with PAX5 P80R missense mutations. In 30 of these cases, PAX5 P80R was homozygous due to deletion of the wild-type (WT) PAX5 allele or copy-neutral loss of heterozygosity. Of the other 14 cases with heterozygous PAX5 P80R mutations, 13 had a second frameshift (n=7), nonsense (n=2) or deleterious missense (n=4) PAX5 mutation. Four of the remaining 1,944 cases also had the PAX5 P80R mutation, but all were heterozygous with preservation of a WT PAX5 allele, consistent with the notion that homozygous or compound heterozygous PAX5 P80R mutation is the hallmark of this subtype. Adult PAX5 P80R cases (n=14) showed better 5-year OS (61.9±13.4%) than those in PAX5alt subtype (42.1±10.2%). To examine the effects of PAX5 P80R on B-cell maturation, WT PAX5, PAX5 P80R, V26G and P34Q were expressed in Pax5-/- lineage-depleted bone marrow cells. Expression of WT PAX5, PAX5 V26G and P34Q resulted in near complete rescue of B cell differentiation; however, expression of PAX5 P80R blocked the differentiation at the pre-pro-B stage of B-cell maturation. Further, Pax5 P80R heterozygous or homozygous mice developed pre-B-ALL with a median latency of 166 and 87 days, respectively, with heterozygous mice acquiring alterations on the second allele. In contrast, Pax5+/- mice, and those harboring G183S mutation observed in familial leukemia, do not spontaneously develop B-ALL. Conclusions These results show the utility of transcriptome sequencing in defining subtypes and founding genetic alterations in B-ALL, provide a revised taxonomy of the disease across the age spectrum, and reinforce the central role of PAX5 as a checkpoint in B lymphoid maturation and leukemogenesis. Disclosures McKay: ImmunoGen Inc.: Employment. Tallman:Orsenix: Other: Advisory board; AROG: Research Funding; BioSight: Other: Advisory board; Cellerant: Research Funding; AbbVie: Research Funding; Daiichi-Sankyo: Other: Advisory board; ADC Therapeutics: Research Funding. Stock:Jazz Pharmaceuticals: Consultancy. Konopleva:Stemline Therapeutics: Research Funding. Relling:Shire Pharmaceuticals: Research Funding. Mullighan:Cancer Prevention and Research Institute of Texas: Consultancy; Amgen: Honoraria, Speakers Bureau; Abbvie: Research Funding; Loxo Oncology: Research Funding; Pfizer: Honoraria, Research Funding, Speakers Bureau.

Although sickle cell disease (SCD) is a monogenic disorder, the severity and specific organ dysfunction and failure are strongly influenced by genetic modifiers. Rapid identification of all modifiers in patients and well-phenotyped cohorts will better define the impact of relevant variants on clinical status, inform disease biology, and identify new therapeutic strategies. We created the Sickle Genome Project (SGP), a whole genome sequencing (WGS) strategy, to define genomic variation and modifiers of SCD. We performed WGS on 871 African American SCD patients from St. Jude Children's Research Hospital who participated in the Sickle Cell Clinical Research and Intervention Program (SCCRIP, Hankins et al. Pediatr Blood Cancer, 2018) and Texas Children's Hospital Hematology Center (TCHC). We developed robust pipelines for accurate detection of single nucleotide polymorphisms (SNPs), identification of structural variants and data retrieval/sharing via the St. Jude Cloud platform (to be described elsewhere). Notable findings include: 1) Confirmed associations of common genetic modifiers with SCD phenotypes, including levels of fetal hemoglobin (BCL11A, HBS1L-MYB, HBB), bilirubin (UGT1A1), and microalbuminuria (APOL1). Additional associations approaching genome-wide significance require further investigation, including replication in independent samples. 2) Improved determination of the SCD modifier α-thalassemia. The most common α-thalassemia mutations in SCD are 3.7 kb or 4.2 kb deletions (-α3.7 and -α4.2 alleles), which arose from recombination between homologous HBA1 and HBA2 genes and are difficult to map using standard WGS reads. Three independent crossover events are described for -α3.7 and one for -α4.2 in SCD cohorts. We developed a novel approach to identify α-globin gene deletions by local de novo assembly of WGS data and coverage depth analysis. We identified 5 -α3.7 alleles (frequencies 0.77-32.12%) and 7 -α4.2 alleles (frequencies 0.19-5.77%). Collectively, the frequency of all -α alleles was 57%, reflecting at least 12 distinct recombination events, greatly exceeding previously published counts. These findings better define the evolution of α-globin genes to allow improved understanding of their regulation and influence on SCD. 3) Characterization of β0-thalassemia alleles. Mutations in the extended β-globin locus influence SCD phenotypes. Five SGP patients had large β-globin (HBB) deletions associated with elevated fetal hemoglobin, which ameliorates symptoms of SCD. Twenty-three patients had HbSβ0-thalassemia, which reduces the severity of some SCD phenotypes. Overall, 48.6% (18/37) of patients clinically designated as HbSβ0 -thalassemia had no identified β-thalassemia mutation. Moreover, 4/680 patients (0.6%) designated HbSS were identified to be β0-thalassemia heterozygotes. The MCV, RBC and %HbA2 distributions overlapped substantially in correct vs. incorrect genotype assignments. Improved discrimination of HbSβ0 vs HbSS genotypes by WGS will better define associated phenotype differences to impact clinical care. 4) Determination of a genetic variant linked to vaso-occlusive crisis (VOC). Previously, a single GWAS study linked rs3115229, located 63.7 kb 5′ upstream of the KIAA1109 gene, with VOC at borderline significance (P = 5.63 × 10−8) (Chaturvedi et al, Blood 130, 2017). Using WGS data for 327 SGP participants (HbSS or HbSβ0-thalassemia) enrolled in the SCCRIPP study, we found strong association (p = 7 x 10-5) between the onset of VOC and a 4-SNP diplotype within an adjacent LD block of the KIAA1109-TENR-IL2-IL21 region (chr4: 122.8Mb - 123.8Mb) which has been previously associated with numerous inflammatory disorders. We validated this association using imputed genome-wide array data in an independent group of SCD patients (Sleep and Asthma Cohort, n= 181 patients, p = 0.05) (Cohen et al, Ann Am Thorac Soc, 2016). This works provides confirmation that the region surrounding KIAA1109 is associated with pain crisis in SCD. Our studies provide new information on the genomic architecture of SCD patients and delineate a consolidated approach for future applications of precision medicine. Disclosures Hankins: Novartis: Research Funding; Global Blood Therapeutics: Research Funding; NCQA: Consultancy; bluebird bio: Consultancy. Estepp:Global Blood Therapeutics: Consultancy, Research Funding; ASH Scholar: Research Funding; NHLBI: Research Funding; Daiichi Sankyo: Consultancy.

Introduction Relapsed childhood acute lymphoblastic leukemia (ALL) is a leading cause of cancer-related death in children and has poor prognosis due to acquired drug resistance. However, the clonal evolution leading to drug resistance at ALL relapse is incompletely understood. Methods We performed whole-genome sequencing (WGS) of samples at diagnosis and relapse from 103 Chinese patients, most of whom were enrolled on the Shanghai Children's Medical Center (SCMC) ALL2005 frontline treatment protocol. We also performed ultra-deep sequencing at 5,000-50,000X coverage of 211 serial bone-marrow samples from 17 of these patients (3-23 per case) collected during ALL therapy. Fifteen ALL subtypes were identified based on copy number variation, structural re-arrangement and gene fusion by WGS and RNA-Seq analysis. Cases were selected based on sample availability with no bias on age, gender, cytogenetics, or immunophenotype compared to the characteristics of all relapsed cases at SCMC; however only one CRLF2-rearranged case was observed suggesting that CRLF2-rearrangement could be a rare event in Asian patients. Functional impact of novel mutations was assessed by ectopic expression in ALL cell lines. Results Relapse-specific somatic alterations were significantly enriched in 12 genes (NT5C2, NR3C1/2, PRPS1/2, TP53, CREBBP, MSH2/6, PMS2, WHSC1, and FPGS) predominantly involved in drug metabolism and response pathways. These somatic alterations were present in over 50% (56/103) of relapsed ALLs and were enriched in patients with early relapse (9-36 months from diagnosis) compared to very early (36 months) patients. NR3C1 mutations resulted in loss of glucocorticoid receptor transcriptional activity and severely impaired glucocorticoid response in vitro, while FPGS mutations led to reduced enzymatic activation of methotrexate. Genome-wide analysis identified 9 mutational signatures, including two novel ones exclusive to relapse. Novel signature 1 was characterized by C>G mutations and was present in 15% of the relapsed cases with an enrichment for hyperdiploid ALL, while novel signature 2, present in 14% of the relapsed cases, was characterized by C mutations (C>T in particular) followed by a G. These novel signatures gave rise to relapse-specific drug resistance mutations, including PRPS1, TP53, NT5C2, and KRAS mutations. The majority (59%) of the cases underwent a selective sweep as the relapse clone arose from a subclone detectable at diagnosis; however, very early relapse cases were less likely to experience a clonal sweep and had multiple lineages present at relapse. In patients tracked serially through multiple relapses, the clonal composition of drug resistance variants evolved substantially in response to chemotherapy. For example, in one patient, two independent NT5C2-mutant clones appeared at relapse, manifesting convergent evolution (Fig. 1a). One of the NT5C2-mutant clones expanded after subsequent treatment while the other diminished in prevalence. Sequential acquisition of multiple drug resistance mutations was also evident, with one patient sequentially acquiring KRAS, FBXW7, NR3C1 and FPGS mutations over a period of 5 years through multiple relapses (Fig. 1b). Indeed, 17% (18/103) of patients acquired multiple drug resistance mutations at relapse. The median time from detecting resistant clones to relapse was 41 days, suggesting that ultra-deep sequencing offers a means for early detection of genetic lesions that could serve as an indicator of relapse to enable earlier therapeutic intervention. Conclusions Very early relapses likely have different resistance mechanisms than early or late relapses, perhaps due to increased intrinsic resistance or the presence of multiple drug-resistant clones. At least a subset of relapse-specific drug resistance mutations may be acquired de novo, rather than pre-existing, as evidenced by resistance mutations bearing the novel relapse-specific mutational signatures which are likely to be therapy-induced. Relapsed ALL can acquire multiple drug resistance mutations targeting different drug classes. Frequent monitoring of disease, such as via cell-free DNA sequencing, may enable earlier detection of relapse and facilitate timely treatment to avoid selection for drug-resistant clones. Disclosures No relevant conflicts of interest to declare.

To determine whether gene expression profiling could improve outcome prediction in children with acute lymphoblastic leukemia (ALL) at high risk for relapse, we profiled pretreatment leukemic cells in 207 uniformly treated children with high-risk B-precursor ALL. A 38-gene expression classifier predictive of relapse-free survival (RFS) could distinguish 2 groups with differing relapse risks: low (4-year RFS, 81%, n = 109) versus high (4-year RFS, 50%, n = 98; P < .001). In multivariate analysis, the gene expression classifier (P = .001) and flow cytometric measures of minimal residual disease (MRD; P = .001) each provided independent prognostic information. Together, they could be used to classify children with high-risk ALL into low- (87% RFS), intermediate- (62% RFS), or high- (29% RFS) risk groups (P < .001). A 21-gene expression classifier predictive of end-induction MRD effectively substituted for flow MRD, yielding a combined classifier that could distinguish these 3 risk groups at diagnosis (P < .001). These classifiers were further validated on an independent high-risk ALL cohort (P = .006) and retainedindependent prognostic significance (P < .001) in the presence of other recently described poor prognostic factors (IKAROS/IKZF1 deletions, JAK mutations, and kinase expression signatures). Thus, gene expression classifiers improve ALL risk classification and allow prospective identification of children who respond or fail current treatment regimens. These trials were registered at http://clinicaltrials.gov under NCT00005603.

Motivation: In some applications, prior biological knowledge can be used to define a specific pattern of association of multiple endpoint variables with a genomic variable that is biologically most interesting. However, to our knowledge, there is no statistical procedure designed to detect specific patterns of association with multiple endpoint variables. Results: Projection onto the most interesting statistical evidence (PROMISE) is proposed as a general procedure to identify genomic variables that exhibit a specific biologically interesting pattern of association with multiple endpoint variables. Biological knowledge of the endpoint variables is used to define a vector that represents the biologically most interesting values for statistics that characterize the associations of the endpoint variables with a genomic variable. A test statistic is defined as the dot-product of the vector of the observed association statistics and the vector of the most interesting values of the association statistics. By definition, this test statistic is proportional to the length of the projection of the observed vector of correlations onto the vector of most interesting associations. Statistical significance is determined via permutation. In simulation studies and an example application, PROMISE shows greater statistical power to identify genes with the interesting pattern of associations than classical multivariate procedures, individual endpoint analyses or listing genes that have the pattern of interest and are significant in more than one individual endpoint analysis. Availability: Documented R routines are freely available from www.stjuderesearch.org/depts/biostats and will soon be available as a Bioconductor package from www.bioconductor.org. Contact: stanley.pounds@stjude.org Supplementary information: Supplementary data are available at Bioinformatics online.

To characterize the biology and optimal therapy of acute mixed-lineage leukemia in children, we reviewed the pathologic and clinical features, including response to therapy, of 35 patients with mixed-lineage leukemia. The majority of cases (91%) had blasts cells that simultaneously expressed either T-lineage plus myeloid markers (T/myeloid, n = 20) or B-lineage plus myeloid markers (B/myeloid, n = 12). Overall survival rates for the B/myeloid and T/myeloid subgroups were not significantly different from each other or from the rate for acute myeloid leukemia (AML) but were inferior to the outcome in children with acute lymphoblastic leukemia (ALL). Patients who failed to achieve complete remission with AML-directed therapy could often be induced with a regimen of prednisone, vincristine, and L-asparaginase. Analysis of gene-expression patterns identified a subset of biphenotypic leukemias that did not cluster with T-cell ALL, B-progenitor ALL, or AML. We propose that treatment for biphenotypic leukemia begin with one course of AML-type induction therapy, with a provision for a switch to lymphoid-type induction therapy with a glucocorticoid, vincristine, and L-asparaginase if the patient responds poorly. We also suggest that hematopoietic stem cell transplantation is often not required for cure of these patients.

Methotrexate polyglutamates (MTXPGs) determine in vivo efficacy in acute lymphoblastic leukemia (ALL). MTXPG accumulation differs by leukemic subtypes, but genomic determinants of MTXPG variation in ALL remain unclear. We analyzed 3 types of whole genome variation: leukemia cell gene expression and somatic copy number variation, and inherited single nucleotide polymorphism (SNP) genotypes and determined their association with MTXPGs in leukemia cells. Seven genes (FHOD3, IMPA2, ME2, RASSF4, SLC39A6, SMAD2, and SMAD4) displayed all 3 types of genomic variation associated with MTXPGs (P < .05 for gene expression, P < .01 for copy number variation and SNPs): 6 on chromosome 18 and 1 on chromosome 10. Increased chromosome 18 (P = .002) or 10 (P = .036) copy number was associated with MTXPGs even after adjusting for ALL subtype. The expression of the top 7 genes in leukemia cells accounted for more variation in MTXPGs (46%) than did the expression of the top 7 genes in normal HapMap cell lines (20%). The top 7 inherited SNPs in patients accounted for approximately the same degree of variation (17%) in MTXPGs as did the top 7 SNP genotypes in HapMap cell lines (20%). We conclude that acquired genetic variation in leukemia cells has a stronger influence on MTXPG accumulation than inherited genetic variation.

To the editor: Wnt signaling in hematopoietic cells and the bone marrow microenvironment plays a critical role in maintaining the pool of hematopoietic stem cells (HSCs) and in regulating differentiation.[1][1],[2][2] Wnt signaling is tightly regulated by the interplay of multiple cytoplasmic

To the editor: Hypodiploid acute lymphoblastic leukemia (ALL) with

New prognostic factors may result in better risk classification and improved treatment of children with acute lymphoblastic leukemia (ALL). Recently, high expression of a gene named OPAL1 (outcome predictor in acute leukemia) was reported to be associated with favorable prognosis in ALL. Therefore, we investigated whether OPAL1 expression was of prognostic importance in 2 independent cohorts of children with ALL treated on Cooperative Study Group for Childhood Acute Lymphoblastic Leukemia (COALL)-92/97 (n = 180) and St Jude Total 13 protocols (n = 257). We observed a consistently higher (2.8-fold) expression of OPAL1 in TEL-AML1-positive ALL compared with TEL-AML1-negative ALL in both cohorts, but higher OPAL1 expression was not consistently associated with other favorable prognostic indicators such as age and white blood cell count, or ALL genetic subtype. Lower OPAL1 expression was also not associated with increased in vitro drug resistance. Multivariate analyses including known risk factors showed that OPAL1 expression was not independently related to prognosis in either the COALL or St Jude cohorts. In conclusion, OPAL1 expression may not be an independent prognostic feature in childhood ALL, and its previously reported prognostic impact appears to be treatment dependent.

FMS-like tyrosine kinase 3 (FLT3) is almost universally expressed in B-precursor childhood acute lymphoblastic leukemia (ALL). Cases of ALL with MLL gene rearrangements and those with high hyperdiploidy (> 50 chromosomes) express the highest levels of FLT3, and activating mutations of FLT3 occur in 18% of MLL-rearranged and 28% of hyperdiploid ALL cases. We determined the antileukemic activity of CEP-701, a potent and selective FLT3 inhibitor, in 8 ALL cell lines and 39 bone marrow samples obtained at diagnosis from infants and children with various subtypes of ALL. CEP-701 induced pronounced apoptotic responses in a higher percentage of samples that expressed high levels of FLT3 (74%, n = 23) compared with samples with low levels of expression (8%, n = 13; P = .0003). Sensitivity to FLT3 inhibition was particularly high in samples with MLL gene rearrangements (82%, n = 11; P = .0005), high hyperdiploidy (100%, n = 5; P = .0007), and/or FLT3 mutations (100%, n = 4; P = .0021). Seven of 7 sensitive samples examined by immunoblotting demonstrated constitutively phosphorylated FLT3 that was potently inhibited by CEP-701, whereas 0 of 6 resistant samples expressed constitutively phosphorylated FLT3. We conclude that the FLT3 inhibitor CEP-701 effectively suppresses FLT3-driven leukemic cell survival. Clinical testing of CEP-701 as a novel molecularly targeted agent for the treatment of childhood ALL is warranted.

Contemporary treatment of pediatric acute myeloid leukemia (AML) requires the assignment of patients to specific risk groups. To explore whether expression profiling of leukemic blasts could accurately distinguish between the known risk groups of AML, we analyzed 130 pediatric and 20 adult AML diagnostic bone marrow or peripheral blood samples using the Affymetrix U133A microarray. Class discriminating genes were identified for each of the major prognostic subtypes of pediatric AML, including t(15;17)[PML-RARα], t(8;21)[AML1-ETO], inv16 [CBFβ-MYH11], MLL chimeric fusion genes, and cases classified as FAB-M7. When subsets of these genes were used in supervised learning algorithms, an overall classification accuracy of more than 93% was achieved. Moreover, we were able to use the expression signatures generated from the pediatric samples to accurately classify adult de novo AMLs with the same genetic lesions. The class discriminating genes also provided novel insights into the molecular pathobiology of these leukemias. Finally, using a combined pediatric data set of 130 AMLs and 137 acute lymphoblastic leukemias, we identified an expression signature for cases with MLL chimeric fusion genes irrespective of lineage. Surprisingly, AMLs containing partial tandem duplications of MLL failed to cluster with MLL chimeric fusion gene cases, suggesting a significant difference in their underlying mechanism of transformation.

The AML1/CBFβ transcriptional complex is essential for the formation of definitive hematopoietic stem cells (HSCs). Moreover, development of the hematopoietic system is exquisitely sensitive to the level of this complex. To investigate the effect of AML1 dosage on adult hematopoiesis, we compared the hematopoietic systems of AML1+/– and AML1+/+ mice. Surprisingly, loss of a single AML1 allele resulted in a 50% reduction in long-term repopulating hematopoietic stem cells (LTR-HSCs). This decrease did not, however, extend to the next level of hematopoietic differentiation. Instead, AML1+/– mice had an increase in multilineage progenitors, an expansion that resulted in enhanced engraftment following transplantation. The expanded pool of AML1+/– progenitors remained responsive to homeostatic mechanisms and thus the number of mature cells in most lineages remained within normal limits. Two notable exceptions were a decrease in CD4+ T cells, leading to an inversion of the CD4+ to CD8+ T-cell ratio and a decrease in circulating platelets. These data demonstrate a dosage-dependent role for AML1/CBFβ in regulating the quantity of HSCs and their downstream committed progenitors, as well as a more restricted role in T cells and platelets. The latter defect mimics one of the key abnormalities in human patients with the familial platelet disorder resulting from AML1 haploinsufficiency.

St Jude Total Therapy Study XIIIB for childhood acute lymphoblastic leukemia (ALL) incorporated more stringent risk classification, early intensification of intrathecal chemotherapy, reinduction treatment, and the addition of dexamethasone to postremission therapy to increase the proportion of event-free survivors without jeopardizing their quality of life. Cranial irradiation was reserved for the 12% of patients who had T-cell ALL and a presenting leukocyte count of 100 × 109/L or more, or CNS-3 (5 or more leukocytes/μL with identifiable blast cells in an atraumatic sample or the presence of cranial nerve palsy) status. Among the 247 consecutive patients enrolled in the study, 117 were classified as having lower-risk leukemia and received mainly antimetabolite-based continuation therapy; the 130 cases with higher-risk leukemia received more intensive continuation chemotherapy with multiple drug pairs administered in weekly rotation. The 5-year event-free survival estimate was 80.8% ± 2.6% (SE); the 8-year rate was 78.6% ± 5.8%. The 5-year cumulative risk of an isolated central nervous system (CNS) relapse was 1.7% ± 0.8%, and that of isolated plus combined CNS relapse was 3.0% ± 1.1%. The 5-year cumulative risks of etoposide-related myeloid malignancies were 1.8% ± 1.3% in the lower-risk patients who received a cumulative dose of 1.2 g/m2 and 5.0% ± 2.0% in the higher-risk patients who received a cumulative dose of up to 14.4 g/m2 (P = .18). Independent adverse prognostic features included the presence of MLL-AF4 or BCR-ABL fusion gene and minimal residual leukemia of 0.01% or more at the end of the 6-week remission induction phase. Our results suggest the efficacy of early intensification of intrathecal chemotherapy and provide the basis for studies omitting cranial irradiation altogether. (Blood. 2004;104:2690-2696)

Because de novo purine synthesis (DNPS) is a target of widely used antileukemic agents (eg, methotrexate, mercaptopurine), we determined the rate of DNPS and the expression of genes involved in purine metabolism in different subtypes of acute lymphoblastic leukemia (ALL). Among 113 children with newly diagnosed ALL, lymphoblasts with the TEL-AML1 translocation had significantly lower DNPS than all other genetic subtypes of B-lineage ALL or T-lineage ALL (352 ± 57 versus 1001 ± 31 or versus 1315 ± 76 fmol/nmol/h, P < .0001). By assessing the expression of 82 genes involved in purine metabolism (KEGG pathway database) in ALL blasts from 38 patients with B-lineage ALL (14 with TEL-AML1, 24 without), we identified 16 genes that were differentially expressed in TEL-AML1–positive and TEL-AML1–negative ALL (P < .001, false discovery rate [FDR] = 5%). The pattern of expression of these 16 genes discriminated TEL-AML1–positive ALL with a true accuracy of 84% in an independent test set (n = 17, confidence interval 70% to 94%, P < .001). Western blots of selected genes documented corresponding levels of the proteins encoded. Differentially expressed genes included HPRT, IMPDH, PAICS, and GART, all of which were expressed at a significantly lower level in TEL-AML1 ALL. These findings have established that TEL-AML1 ALL has significantly lower de novo purine synthesis and differential expression of genes involved in purine metabolism.

The Runx1/core binding factor-β (CBFβ) transcriptional complex is required for the establishment of hematopoiesis during development. Despite its critical role during development, a detailed analysis of Runx1 expression within specific lineages and developmental stages of the adult hematopoietic system is lacking. To address this, we have developed a Runx1—green fluorescent protein (GFP) knock-in mouse. We show that Runx1 is expressed in several hematopoietic lineages, including myeloid, B-lymphoid, and T-lymphoid cells. By contrast, Runx1 is weakly expressed in early erythroid cells, and its expression is rapidly extinguished during later stages of erythropoiesis. Runx1 expression is induced during early B-cell development and is expressed at a uniform level during all subsequent stages of B-cell development. Within the thymus, Runx1 is expressed at the highest level in CD4-CD8- double-negative thymocytes. In peripheral T cells, Runx1 is differentially expressed, with CD4+ T cells expressing 2- to 3-fold higher levels of Runx1 than CD8+ cells. Taken together, these findings indicate that although widely expressed in the hematopoietic system, the expression of Runx1 is regulated in a cell type— and maturation stage—specific manner. In addition, the Runx1-IRES-GFP knock-in mouse strain should prove valuable for investigation of Runx1 function in adult hematopoiesis.

Contemporary treatment of pediatric acute lymphoblastic leukemia (ALL) requires the assignment of patients to specific risk groups. We have recently demonstrated that expression profiling of leukemic blasts can accurately identify the known prognostic subtypes of ALL, including T-cell lineage ALL (T-ALL), E2A-PBX1, TEL-AML1, MLL rearrangements, BCR-ABL, and hyperdiploid karyotypes with more than 50 chromosomes. As the next step toward developing this methodology into a frontline diagnostic tool, we have now analyzed leukemic blasts from 132 diagnostic samples using higher density oligonucleotide arrays that allow the interrogation of most of the identified genes in the human genome. Nearly 60% of the newly identified subtype discriminating genes are novel markers not identified in our previous study, and thus should provide new insights into the altered biology underlying these leukemias. Moreover, a proportion of the newly selected genes are highly ranked as class discriminators, and when incorporated into class-predicting algorithms resulted in an overall diagnostic accuracy of 97%. The performance of an array containing the identified discriminating genes should now be assessed in frontline clinical trials in order to determine the accuracy, practicality, and cost effectiveness of this methodology in the clinical setting.

The t(8;21) translocation, which encodes the AML1-ETO fusion protein (now known as RUNX1-CBF2T1), is one of the most frequent translocations in acute myeloid leukemia, although its role in leukemogenesis is unclear. Here, we report that exogenous expression of AML1-ETO in human CD34+ cells severely disrupts normal erythropoiesis, resulting in virtual abrogation of erythroid colony formation. In contrast, in bulk liquid culture of purified erythroid cells, we found that while AML1-ETO initially inhibited proliferation during early (erythropoietin [EPO]–independent) erythropoiesis, growth inhibition gave way to a sustained EPO-independent expansion of early erythroid cells that continued for more than 60 days, whereas control cultures became growth arrested after 10 to 13 days (at the EPO-dependent stage of development). Phenotypic analysis showed that although these cells were CD13− and CD34−, unlike control cultures, these cells failed to up-regulate CD36 or to down-regulate CD33, suggesting that expression of AML1-ETO suppressed the differentiation of these cells and allowed extensive self-renewal to occur. In the early stages of this expansion, addition of EPO was able to promote both phenotypic (CD36+, CD33−, glycophorin A+) and morphologic differentiation of these cells, almost as effectively as in control cultures. However, with extended culture, cells expressing AML1-ETO became refractory to addition of this cytokine, suggesting that a block in differentiation had been established. These data demonstrate the capacity of AML1-ETO to promote the self-renewal of human hematopoietic cells and therefore support a causal role for t(8;21) translocations in leukemogenesis.

We developed 'clipping reveals structure' (CREST), an algorithm that uses next-generation sequencing reads with partial alignments to a reference genome to directly map structural variations at the nucleotide level of resolution. Application of CREST to whole-genome sequencing data from five pediatric T-lineage acute lymphoblastic leukemias (T-ALLs) and a human melanoma cell line, COLO-829, identified 160 somatic structural variations. Experimental validation exceeded 80%, demonstrating that CREST had a high predictive accuracy.

We developed Copy Number Segmentation by Regression Tree in Next Generation Sequencing (CONSERTING), an algorithm for detecting somatic copy-number alteration (CNA) using whole-genome sequencing (WGS) data. CONSERTING performs iterative analysis of segmentation on the basis of changes in read depth and the detection of localized structural variations, with high accuracy and sensitivity. Analysis of 43 cancer genomes from both pediatric and adult patients revealed novel oncogenic CNAs, complex rearrangements and subclonal CNAs missed by alternative approaches.

Acute lymphoblastic leukemia (ALL) is the most common childhood cancer, with T lineage ALL (T-ALL) accounting for approximately 20% of cases. Many T-ALL patients present with high risk clinical features, and receive aggressive multi-agent genotoxic therapies that pose a substantial risk of long-term adverse health effects. Glucocorticoids (GCs) are an integral part of current treatment strategies for T-ALL and other lymphoid cancers. Despite decades of clinical use, the molecular mechanisms underlying GC efficacy and resistance are incompletely understood. Limited initial response to GCs correlates with poor outcome, and secondary GC resistance is frequently observed in relapsed patients. Thus, identifying drugs that either enhance initial responses or prevent resistance to GCs could significantly improve the treatment of T-ALL and other lymphoid malignancies. Our laboratory performed retroviral insertional mutagenesis (RIM) in wild-type (KrasWT) and KrasG12D "knock-in" mutant mice to generate a panel of primary transplantable T-ALLs that recapitulate the genetic heterogeneity found in human cancers. We used these reagents to assess the efficacy of treatment with the GC dexamethasone (DEX) alone and in combination with the PI3 kinase (PI3K) inhibitor GDC-0941. We observed a robust and significant overall response to 15 mg/kg/day DEX in cohorts of recipient mice transplanted with 10 independent parental KrasWT (n=5) and KrasG12D (n=5) T-ALLs, which was modestly enhanced by combined treatment with 125 mg/kg/day GDC-0941. Prolonged in vivo treatment resulted in outgrowth of 65 independent relapsed T-ALLs, many of which harbor novel retroviral integrations. We verified intrinsic drug resistance in a number of these relapsed leukemias by transplanting them into secondary recipients and treating these mice with 15 mg/kg/day DEX. Intriguingly, 23 (35%) of the T-ALLs that relapsed after an initial response to treatment exhibited markedly reduced GC receptor (GR) protein expression, suggesting a novel and common mechanism for evading GC-induced cell death. Similarly, analysis of T-ALL patient samples showed that low GR expression is rare in diagnostic specimens, but enriched in relapsed/refractory cases. We have identified two likely mechanisms of GR down-regulation in relapsed mouse T-ALLs including a nonsense mutation in the gene encoding GR and a retroviral integration in a putative distal GR promoter region that greatly reduces mRNA expression. We also performed transcriptome (RNA-seq) analysis in one sensitive parental and corresponding resistant KrasWT T-ALL after short-term in vivo DEX treatment, and observed a dramatic reduction in the number of differentially expressed genes in the resistant versus parental leukemia. We recently generated transcriptomes from this panel of 10 primary RIM-induced T-ALLs with the goal of identifying a GC response signature, and also to assess the ability of resistant cells that have lost or retained GR expression to activate this program. Finally, in contrast to previous studies in this genetically accurate in vivo preclinical model (Dail et al. Nature 2014), very few relapsed T-ALLs showed reduced Notch intracellular domain (NICD) expression relative to the corresponding parental leukemia suggesting that loss of Notch1 activation is not a major mechanism contributing to DEX resistance. Taken together, these data reveal an unexpected and common putative mechanism of GC resistance in T-ALL that can inform the development of treatment strategies for relapsed/refractory patients. We are currently focusing on analyzing transcriptome data and performing whole exome sequencing in order to uncover additional mechanisms of resistance to DEX, and then using these data to identify and investigate targeted inhibitors that might overcome GC resistance in vivo. Disclosures Dail: Genentech, Inc.: Employment. Sampath:Genentech: Employment.

Acquired activating mutations in kinase/PI3K/RAS signaling pathways occur in about half of pediatric acute leukemia cases with Mixed Lineage Leukemiagene rearrangements (MLL-R; MLL, also known as KMT2A). Mutations resulting in activated signaling cooperate with the MLL-R in mouse leukemia models, however, detailed insight on the effect on the transcriptional and proteomic landscape is lacking. In infant MLL-R acute lymphoblastic leukemia (ALL), a subtype with a very poor prognosis, a majority of the activating mutations are subclonal, but the biological mechanisms by which subclonal mutations affect leukemogenesis remain unclear. Here we show that NRASG12D, FLT3internal tandem duplication(ITD)and FLT3N676K,cooperates with MLL-MLLT3 in myeloid leukemogenesis using a competitive murine retroviral bone marrow transplantation model. The addition of an activating mutation remodels the gene expression patterns of the MLL-R leukemia with distinct profiles for each mutation as determined by RNA-sequencing. Gene set enrichment analysis revealed enrichment of genes involved in chromatin assembly, transcription and stemness in leukemia induced by MLL-MLLT3 and NRASG12D, FLT3ITD or FLT3N676K. Leukemia induced by only MLL-MLLT3 displayed upregulation of genes involved in signal transduction suggesting activation of such pathways by alternative mechanisms. Upon secondary transplantation, mice receiving MLL-MLLT3-only leukemic cells succumbed to disease at a similar latency to those receiving MLL-MLLT3 and an activating mutation, supporting that fully transformed MLL-MLLT3 leukemias have sustained active signaling via high expression of genes involved in signal transduction. Using the same model as above but with significantly reduced numbers of transplanted cells containing an activating mutation, we could further show that a subclonal mutation, exemplified by FLT3N676K, causes significantly reduced disease latency as compared to mice receiving MLL-MLLT3 only (34 vs 50 days). The size of the double positive (MLL-MLLT3+FLT3N676K) and single positive (MLL-MLLT3) clones within each mouse was determined by flow cytometry revealing that 8/24 mice had a double positive subclone (≤50%). The clonal evolution of the double positive cells from 22/24 mice was assessed in secondary recipients showing three distinct patterns 1) increase in size, 19/22 mice 2) maintained, 2/22 mice or 3) decreased in size, 1/22 mice. Targeted gene re-sequencing of the latter leukemia that lost its MLLT3+FLT3N676K subclone, identified a de-novo CblA308T in the SH2-like domain, in the MLL-MLLT3 onlycells that had gained clonal dominance in the secondary recipient. The decreased disease latency in mice with subclonal activating mutations raises the possibility that cells with an activating mutation support the growth of other leukemic cells by direct cell-cell contact or through secreted factors. Transcriptome and proteomic analyses identified a high expression of the macrophage inhibitory factor (MIF), a pro-inflammatory cytokine, in mice with MLL-MLLT3 and NRASG12D, FLT3ITD or FLT3N676K. Addition of rMIF increased the survival of MLL-MLLT3 murine leukemic cells in vitro. Although additional factors likely mediate pro-leukemic effects in vivo, our data suggest that MIF could be one of these factors. In summary, our data demonstrate that activating mutations cooperate with the MLL-R in murine leukemogenesis and cause widespread changes in the transcriptional landscape. In addition, our results suggest that cells containing an activating mutation in addition to an MLL-fusion positively influence the survival and likely also the growth of other leukemic cells, suggesting a pro-leukemic effect mediated by interclonal cooperation between clones carrying distinct mutational set-ups in leukemogenesis. Disclosures No relevant conflicts of interest to declare.

Although abnormalities involving the short arm of chromosome 12 (12p) are one of the most frequently observed rearrangements in childhood acute lymphoblastic leukemia (ALL), little is known about the frequency of different structural abnormalities and their relationship to the status of the ETV6 (also named TEL) gene in this region. Of 815 children with newly diagnosed ALL, 94 (11.5%) had a total of 104 cytogenetic 12p abnormalities. Loss of genetic material was observed in 67 (64%) of these abnormalities. Cases with 12p alterations had a much lower frequency of hyperdiploidy greater than 50 (7%) than did the ALL population in general, but these cases had a similar distribution of immunophenotype and similar 5-year event-free survival (70% ± 5% SE v 64% ± 2%, P = .64). Rearrangement of the ETV6 gene was identified in 36 (56%) of 64 cases evaluated. The ETV6-CBFA2 (TEL-AML1) fusion transcript was found in 25 (66%) of 38 cases evaluated, and all but one of these showed ETV6 rearrangement. Importantly, ETV6 rearrangement was associated with a favorable prognosis (5-year event-free survival: 89% ± 6% v 60% ± 1%, P < .01). We conclude that most but not all 12p cytogenetic abnormalities in childhood ALL involve ETV6, and that rearrangement of ETV6 is associated with a favorable treatment outcome.

TEL gene rearrangement is the most common genetic lesion in pediatric acute lymphoblastic leukemia (ALL), occurring in about 25% of B-lineage cases. We previously showed that, among patients treated on St Jude protocols, TEL rearrangement independently conferred an excellent prognosis. To extend these results to patients treated with antimetabolite-based therapy, we performed Southern blot analysis to determine the TEL gene status of 104 cases of B-lineage ALL treated on Pediatric Oncology Group 8602, matched on age, gender, and leukocyte count. There were 52 failures among the 77 patients with germline TEL, compared with only 8 failures among 27 patients in the rearranged group. Based on a two-sided logistic regression analysis, stratified for age (subdivided at 10 years), leukocyte count (subdivided at 50,000), and gender, the estimated odds of failing by 4 years in the germline TEL group is 5.4 times that of the rearranged TEL group, with 95% confidence from 1.9 to 15.6, two-sided P = .0009. Thus, the presence of a rearranged TEL gene is also associated with an improved survival among patients treated with antimetabolite-based therapy. Our results indicate that all newly diagnosed ALL patients should be screened for TEL gene rearrangements and suggest that these patients are candidates for less intensive therapy.

MLL is fused to ENL or ELL in acute leukemias that contain t(ll;19)(q23;p13). Although ENL and ELL localize to chromosome 19, bands p13.3 and p13.1, respectively, these breakpoints are not always readily distinguished by standard cytogenetics. We therefore used reverse transcriptase-polymerase chain reaction (RT-PCR) assays to analyze 26 cases of childhood acute leukemia containing t(11;19) to determine the frequencies of ENL and ELL involvement. All 17 cases of acute lymphoblastic leukemia (ALL) had MLL/ENL fusion transcripts. By contrast, of the 9 cases of acute myeloid leukemia (AML) analyzed, 6 had MLL/ENL fusions, 2 had MLL/ELL fusions, and 1 case had no RT-PCR- detectable MLL fusion mRNA. These data suggest that the majority of 11;19 translocations involve ENL, whereas involvement of ELL is relatively uncommon in childhood acute leukemia and may be restricted to AML.

MLL gene rearrangements are associated with an extremely poor prognosis in infants with acute lymphoblastic leukemia (ALL), but little is known about their clinical significance in older children. Therefore, we studied 45 cases of childhood ALL with abnormalities of chromosome 11q23 for rearrangement of the MLL gene to determine if this feature confers a uniformly poor prognosis. MLL gene rearrangements were detected in all 18 cases with the common t(4;11), t(9;11) or t(11;19) translocations, whereas only 5 of 12 patients with either unbalanced or uncommon balanced translocations demonstrated a rearrangement. Abnormalities of the MLL gene were not detected in any of the 15 cases with a deletion or inversion of the chromosomes 11q23 region. The presence of an MLL rearrangement was significantly associated with age less than 1 year (P < .001), leukocyte count –>50 x 10(9)/L (P = .003), and the absence of leukemic cell CD10 expression (P < .001). In a stratified statistical analysis adjusted for age and treatment protocol, MLL gene rearrangement was correlated with an inferior treatment outcome (P = .028). The 4-year event-free survival estimate (+/-SE) was 10% +/-6.5% for cases with a rearranged MLL gene and 64% +/-19.2% for other cases. When infants were excluded from the analysis, MLL rearrangement was still significantly associated with a poor outcome (P = .02), and remained so with the exclusion of t(4;11)-positive cases (P = .05). Thus, regardless of presenting age, MLL gene rearrangement identifies a high-risk subgroup of patients who are not likely to be cured with conventional treatment.

Balanced translocations affecting the 11q23 region are among the most frequent chromosomal abnormalities in childhood acute lymphoblastic leukemia (ALL), comprising 5% to 6%. These cases consistently have a rearranged MLL gene and are associated with high-risk presenting features, hyperleukocytosis and younger age, and a poor treatment outcome. To assess the clinical and biologic significance of 11q23- associated structural chromosomal abnormalities other than translocations, we studied 17 cases of childhood ALL [14 with del(11)(q23) and 3 with inv(11)(p12q23)] that were identified among 785 cases with successful chromosome analysis. In contrast to reported cases with 11q23 and MLL gene rearrangement, our series was characterized by relatively low leukocyte counts (median, 15.1 x 10(9)/L), expression of CD10 antigen but not myeloid-associated CD15 and CDw65 antigens, a relatively high frequency of T-cell immunophenotypes, and a generally favorable prognosis. All 13 cases with interpretable molecular analysis lacked MLL gene rearrangements. We suggest that most cases with deletions or inversions affecting the 11q23 region represent clinically and biologically different entities as compared with those defined by 11q23 translocation.

The t(2;5)(p23;q35) translocation was initially identified in cases of anaplastic large-cell lymphoma (ALCL) that expressed the Ki-1 (CD30) antigen. We have recently cloned this translocation and shown it to encode a chimeric product consisting of the N-terminal portion of a nonribosomal nucleolar phosphoprotein, nucleophosmin (NPM), from chromosome 5, fused to the kinase domain of a novel transmembrane tyrosine-specific protein kinase, anaplastic lymphoma kinase (ALK), from chromosome 2. To better define the spectrum of lymphomas that contain this translocation, we have analyzed 70 cases of non-Hodgkin's lymphoma (NHL) for expression of the t(2;5)-derived NPM/ALK chimeric message by reverse transcriptase-polymerase chain reaction (RT-PCR). Using a previously described set of oligonucleotide primers, NPM/ALK chimeric transcripts were detected in 21 of 22 cases that contained the t(2;5) by cytogenetic analysis and in 10 of 48 cases that either lacked evidence of the t(2;5) or had unsuccessful cytogenetics. In all but 1 case, the NPM/ALK PCR products were of identical size and sequence, suggesting that the genomic chromosome breaks are clustered in a single intron in both NPM and ALK. The NPM/ALK-expressing cases were not confined to NHLs with anaplastic morphology and included 15 ALCLs, 6 immunoblastic lymphomas, and 10 diffuse large-cell lymphomas. Moreover, only slightly greater than half of the cases with anaplastic morphology and 59% of CD30-expressing cases were NPM/ALK positive. Thus, neither anaplastic morphology nor the expression of CD30 accurately predicted the presence of this molecular genetic subtype of lymphoma.

To identify genomic alterations in chronic lymphocytic leukemia (CLL), we performed single-nucleotide polymorphism–array analysis using Affymetrix Version 6.0 on 353 samples from untreated patients entered in the CLL8 treatment trial. Based on paired-sample analysis (n = 144), a mean of 1.8 copy number alterations per patient were identified; approximately 60% of patients carried no copy number alterations other than those detected by fluorescence in situ hybridization analysis. Copy-neutral loss-of-heterozygosity was detected in 6% of CLL patients and was found most frequently on 13q, 17p, and 11q. Minimally deleted regions were refined on 13q14 (deleted in 61% of patients) to the DLEU1 and DLEU2 genes, on 11q22.3 (27% of patients) to ATM, on 2p16.1-2p15 (gained in 7% of patients) to a 1.9-Mb fragment containing 9 genes, and on 8q24.21 (5% of patients) to a segment 486 kb proximal to the MYC locus. 13q deletions exhibited proximal and distal breakpoint cluster regions. Among the most common novel lesions were deletions at 15q15.1 (4% of patients), with the smallest deletion (70.48 kb) found in the MGA locus. Sequence analysis of MGA in 59 samples revealed a truncating mutation in one CLL patient lacking a 15q deletion. MNT at 17p13.3, which in addition to MGA and MYC encodes for the network of MAX-interacting proteins, was also deleted recurrently.

To identify cooperating lesions in core-binding factor acute myeloid leukemia, we performed single-nucleotide polymorphism-array analysis on 300 diagnostic and 41 relapse adult and pediatric leukemia samples. We identified a mean of 1.28 copy number alterations per case at diagnosis in both patient populations. Recurrent minimally deleted regions (MDRs) were identified at 7q36.1 (7.7%), 9q21.32 (5%), 11p13 (2.3%), and 17q11.2 (2%). Approximately one-half of the 7q deletions were detectable only by single-nucleotide polymorphism-array analysis because of their limited size. Sequence analysis of MLL3, contained within the 7q36.1 MDR, in 46 diagnostic samples revealed one truncating mutation in a leukemia lacking a 7q deletion. Recurrent focal gains were identified at 8q24.21 (4.7%) and 11q25 (1.7%), both containing a single noncoding RNA. Recurrent regions of copy-neutral loss-of-heterozygosity were identified at 1p (1%), 4q (0.7%), and 19p (0.7%), with known mutated cancer genes present in the minimally altered region of 1p (NRAS) and 4q (TET2). Analysis of relapse samples identified recurrent MDRs at 3q13.31 (12.2%), 5q (4.9%), and 17p (4.9%), with the 3q13.31 region containing only LSAMP, a putative tumor suppressor. Determining the role of these lesions in leukemogenesis and drug resistance should provide important insights into core-binding factor acute myeloid leukemia.

Introduction. Prior studies have described a subset of B-progenitor ALL cases with a distinct gene expression profile and/or deletions involving ERG (encoding the ETS family member v-ets avian erythroblastosis virus E26 oncogene), however the relationship of these alterations and their role in leukemogenesis are poorly understood. We performed integrated genomic and epigenetic analyses, biochemical studies and leukemogenesis assays to define the genetic basis of this form of ALL. Methods. We studied 1674 childhood, adolescent and young adult B-progenitor ALL cases with microarray gene expression profiling and/or RNA-sequencing data to enable the identification of ERG ALL by unsupervised clustering and predictive analysis of microarrays. Detailed genomic analysis was performed for 144 ERG ALL cases, including whole genome (N=38), exome (n=46) and/or RNA-sequencing (n=57) cases, and single nucleotide polymorphism array analysis. Epigenetic profiling, including whole genome bisulfite sequencing, chromatin immunoprecipitation and sequencing for ERG and histone modifications and ATAC-sequencing were performed for a subset of 8 xenografted ERG tumors and reference cell lines. ERG transcript expression was measured by analysis of RNA-seq analysis and quantitative RT-PCR assays, and by interrogation of TCGA and PCGP RNA-seq data. The function of ERG isoforms was evaluated by EMSA and transcriptional reporter assays, immunofluoresence, colony forming assays and retroviral bone marrow transplant assays. Results. One hundred and forty four cases (8.6%) of B-ALL cases exhibited a distinct gene expression profile and lacked known chromosomal rearrangements (ERG ALL). Such cases had favorable outcome. Eighty cases (55.6%) had focal deletions of ERG with no evidence of oncogenic or chimeric ERG fusions. The deletions were most commonly heterozygous and involving exons 3-7 (n=27) or 3-9 (n=22) of 10 coding exons, and less commonly involving exon 1, or a larger region of the gene. No ERG deletions were identified in non-ERG ALL. Two cases harbored missense mutations in the ETS domain. Analysis of whole genome and exome sequencing data of 71 cases identified a high frequency of alterations of lymphoid transcription factors (46.5%; IKZF1 36.7%, PAX5 11.3%); mutation of transcription factors otherwise uncommon in ALL (21%; MYC, MYCBP2, MGA, ZEB2, GATA3); activation of signaling pathways, most commonly NRAS or KRAS (35.2%); cell cycle regulation (22.5%); and epigenetic modifiers (56.3%), most commonly KMT2D, SETD2, ARID2 and NCOR1. Notably, the five year event-free survival of ERG ALL cases with IKZF1 alterations exceeded 85% in both St Jude and Children's Oncology Group cohorts. We observed striking transcriptional deregulation at the ERG locus. Most (51/56) ERG- deleted cases expressed an ERG isoform encoded by a novel exon in intron 6 that splices in frame to distal exons, resulting in expression of a truncated C-terminal ERG protein that lacks the pointed and central regulatory domains, but retains the ETS and transactivation domain (ERGalt). ERGalt was also present in most (36/44) cases lacking an ERG deletion, and was strongly associated with presence of ERGalt protein in leukemic cells. We also identified expression of an Antisense Long non-coding RNA associated with the ERG locus (ALE) in ERG ALL. ERGalt and ALE were absent, or uncommonly expressed at very low levels in non-ERG ALL. ERGalt was absent, and ALE rarely expressed in non-ALL PCGP and TCGA samples. ERGalt and point mutant ERG were retained in the nucleus, bound DNA targets and acted as competitive inhibitors of wild type (WT) ERG in transcriptional reporter assays. Lineage-negative Arf -null bone marrow cells transduced with ERG WT induced an aggressive erythro-megakaryoblastic leukemia; in contrast ERGalt induced an immature lymphoid progenitor leukemia. Conclusions. Genomic alterations drive aberrant transcription of ERG, resulting on expression of a truncated, C-terminal oncogenic ERG protein. This represents a novel mechanism of transcription factor deregulation in leukemia. As a subset of ERG ALL cases lack ERG deletion, and as IKZF1 alterations are not associated with inferior outcome in this form of ALL, diagnostic approaches must incorporate gene expression profiling in addition to identification of ERG and IKZF1 alterations to accurately identify this form of leukemia. Disclosures Evans: Prometheus Labs: Patents & Royalties: Royalties from licensing TPMT genotyping. Stock:Gilead: Membership on an entity's Board of Directors or advisory committees. Voorhees:Oncopeptides: Consultancy; Onyx Pharmaceuticals: Research Funding; GSK: Consultancy; Oncopeptides: Research Funding; Janssen: Research Funding; A Takeda Oncology Company: Consultancy, Research Funding; Celgene: Consultancy; Millennium Pharmaceuticals: Consultancy, Research Funding; Acetylon Pharmaceuticals, Inc.: Research Funding; Novartis: Consultancy; Array BioPharma: Consultancy; GSK: Research Funding; Celgene: Research Funding. Hunger:Spectrum Pharmaceuticals: Consultancy; Jazz Pharmaceuticals: Consultancy; Sigma Tau: Consultancy; Merck: Equity Ownership. Mullighan:Incyte: Consultancy; Amgen: Honoraria.

Comprehensive studies examining the genomic landscape of T-lineage ALL are lacking, but are important to identify all oncogenic drivers. Here we report sequencing of 264 T-ALL consecutive cases treated on the Children's Oncology Group AALL0434 clinical trial. Whole exome sequencing, copy number analysis using exome and single nucleotide polymorphism array analysis of tumor and remission DNA, and RNA-sequencing of tumor RNA were performed. Cases with immunophenotypic data (N=189) included 19 early T-cell precursor (ETP) cases, 24 near-ETP (with normal CD5 expression) and 146 Non-ETP cases. Median exomic coverage was 89% (72%-96%) of exons with at least 20-fold coverage. We identified 4657 non-synonymous clonal and subclonal somatic mutations (3926 single nucleotide variants (SNV) and 731 insertion-deletion mutations; indels) in 3030 genes, with a mean of 17.6 per case (range 1-50). 176 potential driver genes were identified statistical analysis or by known pathogenic role in cancer. These included NOTCH1 (n=194, 73%), FBXW7 (n=64, 24%), PHF6 (n=50, 19%), PTEN (n=37, 14%), USP7 (n=32, 12%), DNM2 (n=29, 11%) and BCL11B (n=27, 10%). New mutations in T-ALL included CCND3 (n=15, 6%), MYB (n=13, 5%), CTCF (n=13, 5%), MED12 (n=7, 3%), USP9X (n=7, 3%), SMARCA4 (n=7, 3%) and CREBBP (n=6, 2%). In addition to MYB amplification, we identified missense mutations and in-frame protein insertions at the N-terminus of MYB, with a hotspot at codon 14 in a region of six acidic residues in an otherwise hydrophilic N-terminal tail. These mutations resulted in a disordered region that is predicted to affect nuclear localization. The MYB mutations detected were enriched in non-ETP cases (n=13; 8 non-ETP, 1 near-ETP, 4 unknown). Other genes enriched in non-ETP cases included RPL10, CNOT3, MYCN and DDX3X. MED12 mutations were more common in ETP ALL. Sub-clonal mutations (mutant allele fraction of less than 30%) were identified in 111 of 176 driver genes in 198 (75%) cases including NOTCH1 (n=94), FBXW7 (n=29) and PTEN (n=17) indicating that sub-clonal evolution is a hallmark of T-ALL. In addition, multiple mutations in individual genes were commonly observed in single cases. For example, up to 3 different somatic NOTCH1 mutations were detected in each of 9 patients, with 2 different NOTCH1 mutations in 49 cases. Integration of sequence mutations with copy number aberration data showed the following pathways to be most frequently mutated: cell cycle/tumor suppression (N=225; CDKN2A/B (n=206), CDKN1B (n=35), RB1 (n=28)); NOTCH1/FBXW7 (n=212), PI3K-AKT (n=130), JAK-STAT (n=99) and Ras (n=51). Mutations in the PI3K-AKT, JAK-STAT and Ras signaling pathways were mutually exclusive. We identified a high frequency of mutations in transcriptional regulators in 222 cases, including 108 cases with mutations in a core regulatory complex comprising TAL1 (n=51), MYB (n=45) RUNX1 (n=18) and GATA3 (n=13). In 90 of these 108 cases (83%), only a single mutation was present in any of the four genes, consistent with a central role of this complex in leukemogenesis. Epigenetic alterations were identified in 178 cases, including PHF6 (n=63), SMARCA4 (n=23), KDM6A (n=22) and EZH2 (n=18), and new deletions and mutations in KMT2A (MLL; n=11). Interim analysis of transcriptome sequencing data of 126 T-ALL cases detected fusions in 61 (48%) samples, which could be separated into two categories. One weres in-frame fusions resulting in a chimeric protein. The most frequent included MLLT10 fusions (PICALM-MLLT10 (n=3), DDX3X-MLLT10 (n=2) and NAP1L1-MLLT10 (n=1)), KMT2A fusions (KMT2A-MLLT1 (n=4), MLLT6-KMT2A (n=1) and MKT2A-MLLT4 (n=1)), as well as internal tandem duplication mutations involving FLT3 (n=6; 3 ETP, 1 near-ETP, 1 non-ETP, 1 unknown) and NOTCH1 (n=2). We also identified novel fusions including ETV6-CTNNB1 and STMN1-SPI1 (n=1 each). The other category contains rearrangement-driven aberrant expression, including rearrangements in TLX1 (n=11), TLX3 (n=4), TAL1 (n=2), and TAL2 (n=3). Moreover, we found a novel TAL2 transcript in all the 3 cases with TAL2 rearrangement, hijacking a new exon 6kb upstream of the canonical TAL2 transcription start site and extending approximate 3.5kb downstream. These findings provide the first comprehensive landscape of genomic alterations in T-ALL and have provided new insights into the genes and pathways mutated in this disease, their interaction, and the nature of clonal heterogeneity in T-ALL. Disclosures Hunger: Spectrum Pharmaceuticals: Consultancy; Jazz Pharmaceuticals: Consultancy; Sigma Tau: Consultancy; Merck: Equity Ownership. Mullighan:Cancer Science Institute: Membership on an entity's Board of Directors or advisory committees; Amgen: Honoraria, Speakers Bureau; Incyte: Consultancy, Honoraria; Loxo Oncology: Research Funding.

Introduction: Despite risk stratification according to presenting clinical and genetic features, 10-25% of children with acute lymphoblastic leukemia (ALL) relapse, which is associated with a poor prognosis. Here, we sought to provide a comprehensive overview of the genetic alterations associated with relapse in ALL. Methods: We studied 93 children (27 female, 43 male) diagnosed with ALL (62 B-progenitor, 25 T-lineage) between 1987 and 2008 and treated on total therapy studies XI-XVI who experienced relapse and/or a second tumor. Age at diagnosis ranged from 3 months to 18 years. Median time to relapse was 3 years (range 3 months to 10 years). Seventy patients had a single relapse, 15 cases had 2 relapses, and 8 cases developed a second tumor of different lineage (B-cell lymphoma, chronic myeloid leukemia (n=1 each) and acute myeloid leukemia (n=6)). Diagnosis, relapse and matched normal samples (n=299) were studied using Affymetrix SNP 6.0 microarrays and whole genome or whole exome sequencing. Results: We found 2692 copy number aberrations (CNAs) with a median of 9 (range 0-109) in the diagnosis samples (n=91) compared to a median of 10 (range 0-112) in the relapse samples (n=89) and 12 (range 0-70) in subsequent samples (n=20). The number of CNAs did not differ significantly between diagnosis, relapse or subsequent samples. We identified a 7286 non-silent single nucleotide variants (SNVs) and small insertions or deletions (indels) in 5002 genes, 1392 of which were recurrent. The median number of variants was 12 (range 0-70) at diagnosis (n=91), 21 (range 0-858) at relapse (n=91; P=0.0029 v. diagnosis) and 60 (range 10-650) in subsequent samples (n=20; P The most frequently mutated genes were NOTCH1 (n=33), NRAS (n=24), CREBBP (n=20) and KRAS (n=16). Of the recurrently altered genes, only 87 genes were known to be affected in cancer (Cancer Gene Census, COSMIC database), of which 59 were affected in leukemia and lymphoma tissues, indicating that we have identified 1306 novel recurrently affected genes, most commonly C13orf40 and MKI67. Mutations in epigenetic regulators were particularly frequent, with genes mutated in at least 3 cases altered in over 60% of the cohort (e.g. CREBBP, EP300, MLL2, MLL3, KDM6A/B, CTCF, SETD2, TET2/3, and EZH2). Clonal evolution analyses showed multiple patterns of evolution, with relapses sharing either few or many variants with the diagnosis sample in a frequency that reflects both predominant clones and minor subclones propagating relapse. Variants in NOTCH1, NRAS, and CREBBP were preserved from a major clone at diagnosis in 4, 6, and 5 cases respectively, but acquired at relapse or grown out from a minor subclone at diagnosis in 3, 5, and 8 cases respectively. In contrast, variants in USH2A (n=4), FOXA1 (n=3), and purine/pyrimidine synthesis pathway genes NT5C2 (n=3), PRPS1 (n=3) and NT5C1B (n=1) were exclusively found in relapse samples. Notably, the NT5C2 mutations, which are thought to confer resistance to thiopurines, were subclonal at relapse in the majority of cases. We identified 13 cases (10 B-lineage, 3 T-lineage) in which the diagnosis and relapse were fully discordant for all CNAs and sequence mutations, only 4 of which showed a prolonged remission time (>5 years). This suggests that these patients developed a second primary malignancy and may be predisposed to leukemia development. Indeed, one case revealed focal amplifications on chromosome 1q21.1 encompassing the neuroblastoma breakpoint family genes, which are implicated in cancer development. Comprehensive germline analyses are underway. Conclusion: This study has provided detailed insight into the genetic basis of relapse, implicating multiple new genes and pathways involved in treatment resistance, demonstrating multiple patterns of clonal evolution, and revealing an unexpectedly high frequency of genetically discordant second malignancy in relapse in ALL. Disclosures Evans: Prometheus Labs: Patents & Royalties: Royalties from licensing TPMT genotyping. Mullighan:Amgen: Honoraria, Speakers Bureau; Cancer Science Institute: Membership on an entity's Board of Directors or advisory committees; Incyte: Consultancy, Honoraria; Loxo Oncology: Research Funding.

Acute Megakaryoblastic Leukemia (AMKL) is a subtype of acute myeloid leukemia (AML) that morphologically resembles abnormal megakaryoblasts. While extremely rare in adults, pediatric cases comprise 4-15% of newly diagnosed AML patients. Clinical outcomes for Down syndrome (DS) patients with AMKL are uniformly excellent, whereas studies on non-DS patients (non-DS-AMKL) are more variable with the majority reporting inferior survival rates compared to other AML subtypes. Furthermore, the recommendation for stem cell transplant (SCT) in first remission for non-DS-AMKL patients is not uniform among pediatric cooperative groups. Previous efforts have identified chimeric oncogenes in non-DS-AMKL cases, including RBM15-MKL1, CBFA2T3-GLIS2, MLL gene rearrangements and NUP98-KDM5A. The etiology of 30-40% of cases, however, remains unknown. To better understand the genomic landscape of non-DS-AMKL and its contribution to clinical outcomes, we performed RNA and exome sequencing on specimens from 115 patients compiled from eight institutions and three cooperative groups including 90 pediatric and 25 adult cases. Of the 104 patients for whom RNA was available, 27.8% (5/18) adult and 72% (62/86) pediatric cases carried a high confidence fusion event by RNAseq. The most frequent fusions in the pediatric cohort when combining RNAseq data, cytogenetics and RT-PCR include CBFA2T3-GLIS2 (17/90), MLL r (13/90), NUP98-KDM5A (9/90), and RBM15-MKL1 (9/90). Previously described low frequency fusions identified in this expanded cohort, include a case of NIPBL-HOXB9 and a novel but functionally analogous NIPBL-HOXA9 fusion. Similarly, a case carrying GATA2-HOXA10 was identified, which is functionally equivalent to the GATA2-HOXA9 fusion that has been reported in a single case. Chimeric transcripts not previously described include several fusions involving genes within the HOX cluster (HOTAIRM1-HOXA3, HOXA_AS3-HOXA9, EWSR1-HOXB8, PLEK-HOXA11-AS, and BMP2K-HOXD10 each in a single case). Collectively, fusions involving a HOX cluster gene (HOX r) occurred in 11% of the pediatric cohort. Single Nucleotide Variation (SNV) analysis of exome and RNAseq data on the cohort revealed the presence of truncating GATA1 mutations in one adult and 10 pediatric specimens lacking fusion genes. Patients carrying GATA1 mutations did not have stigmata of DS or evidence of mutant reads in germline DNA, suggesting they are not mosaics. To determine if these fusion events contribute significantly to gene expression patterns, samples with greater than 60% purity were subjected to unsupervised clustering. Confirming the strength of the fusions in altering gene expression signatures, samples clustered according to fusion subtype and were distinct from those carrying GATA1 mutations. Specifically MLL r, HOX r, NUP98-KDM5A, and CBFA2T3-GLIS2 cases formed distinct clusters. When analyzing differentially upregulated genes within these subgroups, HOX r cases demonstrated upregulation of a HOX gene signature. Combined with MLL r and NUP98-KDM5A, chimeric oncogenes also known to upregulate HOX cluster genes, roughly one-third of pediatric non-DS-AMKL patients carry a HOX gene expression program. These cases were distinct from those carrying the CBFA2T3-GLIS2 inversion. HOX genes play a significant role in normal hematopoietic development and data suggests that deregulated expression has a central role in the etiology of several subtypes of acute leukemia, in part through the acquisition of enhanced self-renewal. We evaluated our identified HOXr for their ability to serially replate in murine colony formation assays as a surrogate marker of this characteristic. Confirming their pathogenicity, chimeric transcripts conferred an enhanced ability to replate. We conclude that chimeric transcripts involving HOX cluster genes comprise a distinct subset of pediatric AMKL. Clinical outcome analyses between genomic subgroups of this heterogeneous malignancy may allow us to more effectively risk stratify these patients and determine those that may benefit from SCT in first remission. JdR and CB contributed equally FL, DR, MH-E, MF, CMZ, and TAG co-corresponding authors on behalf of AIEOP, BFM, DCOG, and SJCRH study groups Disclosures Shih: Novartis: Research Funding.

Mice with severe combined immunodeficiency (SCID) were injected intravenously with primary bone marrow blasts from 12 children with newly diagnosed t(4;11)(q21;q23) acute lymphoblastic leukemia (ALL). Blasts from eight patients caused overt disseminated leukemia, whereas blasts from the other four patients produced occult leukemia that was detectable only by the polymerase chain reaction (PCR) technique. Only one patient among eight whose blasts caused disseminated leukemia in SCID mice remains alive and disease-free at 48.4 months postdiagnosis. In contrast, three of the other four patients whose blasts did not cause overt leukemia in SCID mice remain alive and disease-free at 6.1, 23.6, and 35.9 months, respectively. Thus, the occurrence of overt leukemia in SCID mice may be a predictor of patients' disease-free survival. The described SCID mouse model system may prove useful for designing more effective treatment strategies against therapy- refractory t(4;11) ALL.