Search

Supplementary Excel Tables

Supplementary Tables and Figures

Genomic characterization of pediatric patients with acute myeloid leukemia (AML) has led to the discovery of somatic mutations with prognostic implications. Although gene-expression profiling can differentiate subsets of pediatric AML, its clinical utility in risk stratification remains limited. Here, we evaluate gene expression, pathogenic somatic mutations, and outcome in a cohort of 435 pediatric patients with a spectrum of pediatric myeloid-related acute leukemias for biological subtype discovery. This analysis revealed 63 patients with varying immunophenotypes that span a T-lineage and myeloid continuum designated as acute myeloid/T-lymphoblastic leukemia (AMTL). Within AMTL, two patient subgroups distinguished by FLT3-ITD and PRC2 mutations have different outcomes, demonstrating the impact of mutational composition on survival. Across the cohort, variability in outcomes of patients within isomutational subsets is influenced by transcriptional identity and the presence of a stem cell–like gene-expression signature. Integration of gene expression and somatic mutations leads to improved risk stratification.Significance:Immunophenotype and somatic mutations play a significant role in treatment approach and risk stratification of acute leukemia. We conducted an integrated genomic analysis of pediatric myeloid malignancies and found that a combination of genetic and transcriptional readouts was superior to immunophenotype and genomic mutations in identifying biological subtypes and predicting outcomes.This article is highlighted in the In This Issue feature, p. 549

Integrating somatic mutation analysis and gene expression profiling distinguishes pediatric AML subtypes with differential prognoses and clinical risks., Genomic characterization of pediatric patients with acute myeloid leukemia (AML) has led to the discovery of somatic mutations with prognostic implications. Although gene-expression profiling can differentiate subsets of pediatric AML, its clinical utility in risk stratification remains limited. Here, we evaluate gene expression, pathogenic somatic mutations, and outcome in a cohort of 435 pediatric patients with a spectrum of pediatric myeloid-related acute leukemias for biological subtype discovery. This analysis revealed 63 patients with varying immunophenotypes that span a T-lineage and myeloid continuum designated as acute myeloid/T-lymphoblastic leukemia (AMTL). Within AMTL, two patient subgroups distinguished by FLT3-ITD and PRC2 mutations have different outcomes, demonstrating the impact of mutational composition on survival. Across the cohort, variability in outcomes of patients within isomutational subsets is influenced by transcriptional identity and the presence of a stem cell–like gene-expression signature. Integration of gene expression and somatic mutations leads to improved risk stratification. Significance: Immunophenotype and somatic mutations play a significant role in treatment approach and risk stratification of acute leukemia. We conducted an integrated genomic analysis of pediatric myeloid malignancies and found that a combination of genetic and transcriptional readouts was superior to immunophenotype and genomic mutations in identifying biological subtypes and predicting outcomes. This article is highlighted in the In This Issue feature, p. 549

The genomic events responsible for the pathogenesis of relapsed adult B-lymphoblastic leukemia (B-ALL) are not yet clear. We performed integrative analysis of whole-genome, whole-exome, custom capture, whole-transcriptome (RNA-seq), and locus-specific genomic assays across nine time points from a patient with primary de novo B-ALL. Comprehensive genome and transcriptome characterization revealed a dramatic tumor evolution during progression, yielding a tumor with complex clonal architecture at second relapse. We observed and validated point mutations in EP300 and NF1, a highly expressed EP300-ZNF384 gene fusion, a microdeletion in IKZF1, a focal deletion affecting SETD2, and large deletions affecting RB1, PAX5, NF1, and ETV6. Although the genome analysis revealed events of potential biological relevance, no clinically actionable treatment options were evident at the time of the second relapse. However, transcriptome analysis identified aberrant overexpression of the targetable protein kinase encoded by the FLT3 gene. Although the patient had refractory disease after salvage therapy for the second relapse, treatment with the FLT3 inhibitor sunitinib rapidly induced a near complete molecular response, permitting the patient to proceed to a matched-unrelated donor stem cell transplantation. The patient remains in complete remission more than 4 years later. Analysis of this patient's relapse genome revealed an unexpected, actionable therapeutic target that led to a specific therapy associated with a rapid clinical response. For some patients with relapsed or refractory cancers, this approach may indicate a novel therapeutic intervention that could alter outcome.

Therapy-related myeloid neoplasms (tMN) occur in children as a consequence of cytotoxic therapies used to treat childhood malignancies, are typically resistant to conventional chemotherapies, require hematopoietic cell transplantation as the only curative option, and typically have a dismal prognosis. While the genomic alterations that drive tMN in children have yet to be comprehensively described, alterations involving the MECOM locus have been described in some myeloid neoplasms like tMN. Overexpression of MECOM is associated with poor prognosis in myeloid malignancies, is present in ~10% of adult cases, and occurs at an increased frequency in pediatric AML with KMT2A rearrangements (KMT2Ar). RNA sequencing from a cohort of 56 pediatric tMN cases collected at St. Jude shows that 24 (43%) cases within our cohort express high levels of MECOM (FPKM value >5). In these 24 cases, we identified one with a canonical fusion (RUNX1-MECOM), one with a NUP98 fusion (NUP98-HHEX), and 18 with KMT2Ar. Whole-genome sequencing demonstrated that two of the remaining MECOMHigh cases have a t(2;3)(p21;q26.2) involving MECOM on chromosome 3 and noncoding regions of chromosome 2 adjacent to ZFP36L2, a gene highly expressed in hematopoietic cells. Further, ENCODE data support that this region of the genome is an active enhancer, suggesting a proximity effect in which this enhancer has been hijacked to drive high levels of MECOM expression. The two remaining cases with aberrant and unexplained MECOM expression are currently under evaluation. In our cohort, MECOM expression levels are predictive of a worse outcome (overall survival at 2 years: MECOMHigh=12.5% vs. MECOMLow=40.6%; log rank p Citation Format: Tamara Lamprecht, Jason R. Schwartz, Jing Ma, Michael P. Walsh, Jeffery M. Klco. MECOM dysregulation is associated with poor outcome in pediatric therapy-related myeloid neoplasms [abstract]. In: Proceedings of the AACR Special Conference on the Advances in Pediatric Cancer Research; 2019 Sep 17-20; Montreal, QC, Canada. Philadelphia (PA): AACR; Cancer Res 2020;80(14 Suppl):Abstract nr B16.

Germline SAMD9 and SAMD9L mutations cause a spectrum of multisystem disorders that carry a markedly increased risk of developing myeloid malignancies with somatic monosomy 7. Here, we describe 16 siblings, the majority of which were phenotypically normal, from 5 families diagnosed with myelodysplasia and leukemia syndrome with monosomy 7 (MLSM7; OMIM 252270) who primarily had onset of hematologic abnormalities during the first decade of life. Molecular analyses uncovered germline SAMD9L (n = 4) or SAMD9 (n = 1) mutations in these families. Affected individuals had a highly variable clinical course that ranged from mild and transient dyspoietic changes in the bone marrow to a rapid progression of myelodysplastic syndrome (MDS) or acute myeloid leukemia (AML) with monosomy 7. Expression of these gain-of-function SAMD9 and SAMD9L mutations reduces cell cycle progression, and deep sequencing demonstrated selective pressure favoring the outgrowth of clones that have either lost the mutant allele or acquired revertant mutations. The myeloid malignancies of affected siblings acquired cooperating mutations in genes that are also altered in sporadic cases of AML characterized by monosomy 7. These data have implications for understanding how SAMD9 and SAMD9L mutations contribute to myeloid transformation and for recognizing, counseling, and treating affected families.

Donor-derived hematologic malignancies are rare complications of hematopoietic cell transplantation (HCT). Although these are commonly either a myelodysplastic syndrome (MDS) or acute myeloid leukemia (AML), in general, they are a heterogeneous group of diseases, and a unified mechanism for their development has remained elusive. Here we report next-generation sequencing, including whole-exome sequencing (WES), whole-genome sequencing (WGS), and targeted sequencing, of a case of donor-derived MDS (dMDS) following HCT for high-risk B-lymphoblastic leukemia (B-ALL) in an adolescent. Through interrogation of single-nucleotide polymorphisms (SNPs) in the WGS data, we unequivocally prove that the MDS is donor-derived. Additionally, we sequenced 15 samples from 12 time points, including the initial B-ALL diagnostic sample through several post-HCT remission samples, the dMDS, and representative germline samples from both patient and donor, to show that the MDS-related pathologic mutations, including a canonical ASXL1 (p.Y700*) mutation, were detectable nearly 3 yr prior to the morphological detection of MDS. Furthermore, these MDS mutations were not detectable immediately following, and for >1 yr post-, HCT. These data support the clinical utility of comprehensive sequencing following HCT to detect donor-derived malignancies, while providing insights into the clonal progression of dMDS over a 4-yr period.

Myelodysplastic syndromes (MDS) are uncommon in children and have a poor prognosis. In contrast to adult MDS, little is known about the genomic landscape of pediatric MDS. Here, we describe the somatic and germline changes of pediatric MDS using whole exome sequencing, targeted amplicon sequencing, and/or RNA-sequencing of 46 pediatric primary MDS patients. Our data show that, in contrast to adult MDS, Ras/MAPK pathway mutations are common in pediatric MDS (45% of primary cohort), while mutations in RNA splicing genes are rare (2% of primary cohort). Surprisingly, germline variants in SAMD9 or SAMD9L were present in 17% of primary MDS patients, and these variants were routinely lost in the tumor cells by chromosomal deletions (e.g., monosomy 7) or copy number neutral loss of heterozygosity (CN-LOH). Our data confirm that adult and pediatric MDS are separate diseases with disparate mechanisms, and that SAMD9/SAMD9L mutations represent a new class of MDS predisposition., Myelodysplastic syndromes (MDS) are uncommon in children and have poor prognosis. Here, the authors interrogate the genomic landscape of MDS, confirming adult and paediatric MDS are separate diseases with disparate mechanisms, and highlighting that SAMD9/SAMD9L mutations represent a new class of MDS predisposition.

DNMT3A mutations occur in ~25% of acute myeloid leukemia (AML) patients. The most common mutation, DNMT3AR882H, has dominant negative activity that reduces DNA methylation activity by ~80% in vitro. To understand the contribution of DNMT3A-dependent methylation to leukemogenesis, we performed whole-genome bisulfite sequencing of primary leukemic and non-leukemic cells in patients with or without DNMT3AR882 mutations. Non-leukemic hematopoietic cells with DNMT3AR882H displayed focal methylation loss, suggesting that hypomethylation antedates AML. Although virtually all AMLs with wild-type DNMT3A displayed CpG island hypermethylation, this change was not associated with gene silencing, and was essentially absent in AMLs with DNMT3AR882 mutations. Primary hematopoietic stem cells expanded with cytokines were hypermethylated in a DNMT3A-dependent manner, suggesting that hypermethylation may be a response to, rather than a cause of, cellular proliferation. Our findings suggest that hypomethylation is an initiating phenotype in AMLs with DNMT3AR882, while DNMT3A-dependent CpG island hypermethylation is a consequence of AML progression.

Therapy-related acute myeloid leukaemia (t-AML) and therapy-related myelodysplastic syndrome (t-MDS) are well-recognized complications of cytotoxic chemotherapy and/or radiotherapy. There are several features that distinguish t-AML from de novo AML, including a higher incidence of TP53 mutations, abnormalities of chromosomes 5 or 7, complex cytogenetics and a reduced response to chemotherapy. However, it is not clear how prior exposure to cytotoxic therapy influences leukaemogenesis. In particular, the mechanism by which TP53 mutations are selectively enriched in t-AML/t-MDS is unknown. Here, by sequencing the genomes of 22 patients with t-AML, we show that the total number of somatic single-nucleotide variants and the percentage of chemotherapy-related transversions are similar in t-AML and de novo AML, indicating that previous chemotherapy does not induce genome-wide DNA damage. We identified four cases of t-AML/t-MDS in which the exact TP53 mutation found at diagnosis was also present at low frequencies (0.003-0.7%) in mobilized blood leukocytes or bone marrow 3-6 years before the development of t-AML/t-MDS, including two cases in which the relevant TP53 mutation was detected before any chemotherapy. Moreover, functional TP53 mutations were identified in small populations of peripheral blood cells of healthy chemotherapy-naive elderly individuals. Finally, in mouse bone marrow chimaeras containing both wild-type and Tp53(+/-) haematopoietic stem/progenitor cells (HSPCs), the Tp53(+/-) HSPCs preferentially expanded after exposure to chemotherapy. These data suggest that cytotoxic therapy does not directly induce TP53 mutations. Rather, they support a model in which rare HSPCs carrying age-related TP53 mutations are resistant to chemotherapy and expand preferentially after treatment. The early acquisition of TP53 mutations in the founding HSPC clone probably contributes to the frequent cytogenetic abnormalities and poor responses to chemotherapy that are typical of patients with t-AML/t-MDS.

SummarySomatic mutations in DNMT3A, which encodes a de novo DNA methyltransferase, are found in ∼30% of normal karyotype acute myeloid leukemia (AML) cases. Most mutations are heterozygous and alter R882 within the catalytic domain (most commonly R882H), suggesting the possibility of dominant-negative consequences. The methyltransferase activity of R882H DNMT3A is reduced by ∼80% compared with the WT enzyme. In vitro mixing of WT and R882H DNMT3A does not affect the WT activity, but coexpression of the two proteins in cells profoundly inhibits the WT enzyme by disrupting its ability to homotetramerize. AML cells with the R882H mutation have severely reduced de novo methyltransferase activity and focal hypomethylation at specific CpGs throughout AML cell genomes.

SummaryThe relationships between clonal architecture and functional heterogeneity in acute myeloid leukemia (AML) samples are not yet clear. We used targeted sequencing to track AML subclones identified by whole-genome sequencing using a variety of experimental approaches. We found that virtually all AML subclones trafficked from the marrow to the peripheral blood, but some were enriched in specific cell populations. Subclones showed variable engraftment potential in immunodeficient mice. Xenografts were predominantly comprised of a single genetically defined subclone, but there was no predictable relationship between the engrafting subclone and the evolutionary hierarchy of the leukemia. These data demonstrate the importance of integrating genetic and functional data in studies of primary cancer samples, both in xenograft models and in patients.

Therapy-related myeloid neoplasms (tMN) occur in children secondary to cytotoxic therapies used to treat pediatric malignancies, are typically resistant to conventional chemotherapy, require hematopoietic cell transplantation as the only curative option, and have a dismal prognosis. The genomic alterations that drive tMN in children have yet to be comprehensively described, and it is unclear if particular genomic lesions hold prognostic value. We have characterized the genomic profile of 62 pediatric tMN cases (tMDS: n=23, tAML: n=39) obtained from the St. Jude Children's Research Hospital Tissue Bank from patients diagnosed between 1987 and 2018. These cases arose following treatment for a variety of primary tumors (hematological (74%), bone and soft tissue (23%), and brain (3%)). Acute lymphoblastic leukemia was the most frequent primary tumor (n=39, 63%). Conventional cytogenetics (n=60) showed a complex karyotype (≥3 structural alterations) in 19 (32%) cases, and 7 of these cases contained a deletion involving chromosome 7 (del(7)). Eleven (18%) other cases without complex karyotypes had del(7). Deletions of chromosome 5 were present in 9 (15%) cases, but only in the context of a complex karyotype. We hypothesized that the patients' younger age and the different spectrum of primary tumor types and chemotherapy would give rise to a mutational spectrum distinct from adult tMN. We used whole exome (WES), whole genome (WGS), and RNA sequencing (RNA Seq) to describe the mutational profile of our pediatric tMN cohort. WES was completed for 58 tumor/normal pairs using Nextera Rapid Capture Expanded Exome (Illumina). Fifteen cases were analyzed by WGS (11 also had WES). Normal comparator genomic DNA was obtained from flow-sorted lymphocytes. An average of 21 coding variants/patient (range: 1-131) was observed from the gene-coding region, and these include synonymous, non-synonymous, and splice site variants. Ras/MAPK pathway mutations were present in 44% of the cases (43 mutations in 27 cases). Canonical KRAS (n = 16), NF1 (n = 10), and NRAS (n = 7) mutations were the most frequent coding mutations. Eleven (18%) patients had either heterozygous deletion or a copy neutral loss of heterozygosity event involving chromosome 17p and the TP53 locus; 5 of these cases had concurrent TP53 missense mutations identified at allele frequencies near 100%. Unlike tMN in adults, mutations in PPM1D were not identified. RNA-Seq completed on 56 evaluable cases identified 28 (50%) cases with KMT2A rearrangement (KMT2Ar). MLLT3 was the most common fusion partner (n=13, 46%). In addition to KMT2A rearrangements, RNA-Seq also identified a RUNX1-MECOM fusion. Alterations involving the MECOM locus have been described in some myeloid neoplasms like tMN, and its overexpression is associated with a poor prognosis and some AMLs with KMT2Ar. MECOM expression levels were variable in this cohort (FPKM range: 0.004 - 38.4) with 24 cases (43%) having an FPKM>5 (MECOMHigh). In addition to the RUNX1-MECOM event, these 24 MECOMHigh cases included 18 with KMT2Ar (64% of KMT2Ar group) and 1 with a NUP98 fusion (NUP98-HHEX). The remaining 4 MECOMHigh cases demonstrate allele-specific MECOM expression, suggesting a cis-regulatory element is driving this expression. Two of these 4 cases have WGS and were found to contain a t(2;3)(p21;q26.2) involving MECOM on chromosome 3 and noncoding regions of chromosome 2 adjacent to ZFP36L2, a gene highly expressed in hematopoietic cells. ENCODE data supports that this region of the genome is an active enhancer in hematopoietic cells, suggesting a proximity effect in which this enhancer has been hijacked to drive high levels of MECOM expression. In our cohort, MECOM expression levels are predictive of a worse outcome (overall survival (OS) at 2 years: High=14.6% vs. Low=46.3%; log rank p In conclusion, we report the genomic profile of a large cohort of pediatric tMN cases and show that high levels of MECOM expression, a portion of which is driven by enhancer hijacking, predicts a worse outcome. Disclosures Gruber: Bristol-Myers Squibb: Consultancy.

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.

Chromosomal translocations are common in acute myeloid leukemia (AML), causing gene fusions that encode oncogenic proteins. The NUP98 gene participates in chromosomal rearrangements with over 30 fusion partners and comprises 6-10% of de novo cases of pediatric AML. These fusions are associated with poor prognosis in children and adults. Among them, NUP98-KDM5A is further enriched in specific subpopulations, found in 15% of non-Down syndrome acute megakaryoblastic leukemia cases and 20% of pediatric acute erythroleukemia. Despite these associations, the direct impact of NUP98-KDM5A on the growth and differentiation of human hematopoietic cells has not been systemically studied. In this study, we transduced cord blood CD34+ hematopoietic stem and progenitor cells (HSPCs) with NUP98-KDM5A or control lentiviral vectors and examined cell proliferation and differentiation. Exposure to cytokines (SCF, TPO, IL-6, FLT3L, SR-1) selected for CD34+ cell growth, and expression of NUP98-KDM5A increased proliferation rate in liquid culture by 19.4-fold (p To further elucidate NUP98-KDM5A effects on differentiation, cytokine supplements known to drive myeloid differentiation (SCF, TPO, G-CSF, and GM-CSF) were added to culture media, which demonstrated the same disruption in myelo-erythropoiesis based on a 4.0-fold decrease (p To evaluate the leukemic potential of NUP98-KDM5A, cells were injected into NRG-SGM3 mice with NUP98-KDM5A inducing at 12 weeks a rapidly lethal myeloid disease in vivo (97.8% human CD45+ cells vs. 2.3% in control, p In conclusion, the expression of NUP98-KDM5A in HSPCs drives proliferation and alters differentiation to maintain stem cell markers while driving an erythroid phenotype in vitro. Furthermore, this fusion is sufficient to engraft in the bone marrow and spleen of NRG-SGM3 mice in vivo and cause marked splenomegaly. Taken together, these data suggest this model properly recapitulates the human disease phenotype seen in patients expressing NUP98-KDM5A and future functional assays and drug screens may provide important insights into understanding the pathophysiology and pharmacologic vulnerabilities of this fusion class. Disclosures Mullighan: Loxo Oncology: Research Funding; Amgen: Honoraria, Other: speaker, sponsored travel; Illumina: Honoraria, Membership on an entity's Board of Directors or advisory committees, Other: sponsored travel; AbbVie: Research Funding; Pfizer: Honoraria, Other: speaker, sponsored travel, Research Funding.

BACKGROUND The genetic alterations responsible for an adverse outcome in most patients with acute myeloid leukemia (AML) are unknown. METHODS Using massively parallel DNA sequencing, we identified a somatic mutation in DNMT3A, encoding a DNA methyltransferase, in the genome of cells from a patient with AML with a normal karyotype. We sequenced the exons of DNMT3A in 280 additional patients with de novo AML to define recurring mutations. RESULTS A total of 62 of 281 patients (22.1%) had mutations in DNMT3A that were predicted to affect translation. We identified 18 different missense mutations, the most common of which was predicted to affect amino acid R882 (in 37 patients). We also identified six frameshift, six nonsense, and three splice-site mutations and a 1.5-Mbp deletion encompassing DNMT3A. These mutations were highly enriched in the group of patients with an intermediate-risk cytogenetic profile (56 of 166 patients, or 33.7%) but were absent in all 79 patients with a favorable-risk cytogenetic profile (P

We and others recently showed that the mutational spectrum of de novo pediatric myelodysplastic syndrome (MDS) and acute myeloid leukemia (AML) is different than those in adults. MDS and AML also occur in children as a consequence of cytotoxic therapies used to treat childhood malignancies and are collectively referred to as therapy-related myeloid neoplasms (tMN). The incidence of pediatric tMN is ~1% in the pediatric cancer population. These secondary malignancies are usually resistant to conventional chemotherapy and managed with hematopoietic cell transplantation (HCT). These patients have a dismal prognosis. TP53 mutations and somatic alterations in chromatin modifiers predominate in adults with tMN, yet whether children with tMN have a similar constellation of genetic alterations remains unclear since comprehensive genomic profiling has not been completed in a large pediatric tMN cohort. We hypothesize that the mutational profile of pediatric tMN will be different than adult tMN given the patients' younger age and the different spectrum of primary tumor types and chemotherapies. Here we describe the somatic mutational profile of pediatric tMN (including tMDS & tAML) using whole exome (WES) and RNA-sequencing. We evaluated 65 diagnostic bone marrow samples from 61 unique patients, obtained from the St. Jude Children's Research Hospital Tissue Bank from patients diagnosed between 1987 & 2018. The cohort contains 26 tMDS and 39 tAML cases; in 4 patients both tMDS and tAML samples were included. Primary tumors included hematological malignancies (n=45), bone and soft tissue solid tumors (n=14), and brain tumors (n=2); acute lymphoblastic leukemia (ALL) was the most common primary tumor (n = 38, 62%). WES was completed for 61 tumor/normal pairs using Nextera Rapid Capture Expanded Exome (Illumina), while WGS was completed on 4 pairs. Normal comparator genomic DNA was obtained from flow-sorted lymphocytes. Median sequencing coverage for the tumor and normal samples were 107x and 95x, respectively. An average of 49 variants/patient (range: 6-217) was observed in the tMN cohort, including coding, non-coding, silent, and splice site variants, which is significantly different than our previously reported 5 variants/patient in pediatric primary MDS (p = 1x10-6). There was not a significant difference in the number of mutations/patient when tMDS was compared to tAML. Mutational signature analysis (https://cancer.sanger.ac.uk/cosmic/signatures) identified 3 major signatures, the most predominant was characterized by a strong bias for C>A mutations (Signature 24), followed by a signature with strong transcriptional strand bias for T>A mutations (Signature 27) and then a smaller subset resembling MDS and AML (Signature 1). Interestingly, patients with Signature 1 had an inferior 2-year overall survival than the other mutational signatures, with a median survival of 0.3 years (p = 0.0005). WES data and conventional karyotyping showed that chromosome 7 deletions (del(7)) were frequent (n=21, 32%), followed by deletions involving chromosome 5 (del(5)) (n=10, 15%). All of the cases with del(5) had complex cytogenetics and 6 of the 10 cases also had del(7). Ras/MAPK pathway mutations were present in 37% of the cases (40 total mutations in 27 cases). Canonical KRAS (n = 14), NF1 (n = 8), and NRAS (n = 7) mutations were the most frequent coding mutations present overall. Only 5 patients had somatic TP53 mutations, all of which had complex karyotypes. RNA sequencing was performed on 55 samples with suitable RNA. KMT2A rearrangements (KMT2Ar) were common (n = 29, 53%), 4 of which were cytogenetically cryptic. KMT2A rearrangements were more common in tAML (n = 25) but were present in tMDS (n = 4). Nearly half of these KMT2Ar cases also harbored an additional Ras/MAPK mutation. Fusions involving NUP98, RUNX1, MECOM, and ETV6 were also detected. In conclusion, we show that the mutational profile of pediatric tMN has fewer TP53 mutations and more KMT2Ar than adults, as well as a unique set of mutational signatures. These differences are likely a reflection of age-specific chemotherapeutic strategies and fewer pre-existing TP53 mutant hematopoietic clones in children. Future studies understanding the clonal evolution of pediatric tMN development will be helpful in describing pediatric tMN further. Disclosures No relevant conflicts of interest to declare.

The purpose of this study was to define the mutations causing familial monosomy 7 with myelodysplastic syndrome (MDS) and/or acute myeloid leukemia (AML). We have previously reported frequent germline mutations in the genes SAMD9 and SAMD9L in pediatric MDS. We obtained samples on 32 individuals in 7 families with multiple siblings affected with MDS/AML and monosomy 7. We conducted targeted sequencing of genes frequently mutated in MDS/AML and amplicon sequencing of SAMD9 and SAMD9L. We performed cell cycle analysis assays on SAMD9/9L mutations to study their effects on cell proliferation. Our data show that 5 of the 7 families have germline pathologic mutations in SAMD9 (n=1) and SAMD9L (n=4), with significant clinical variability in the phenotypes displayed, ranging from transient monosomy 7 to overt AML with monosomy 7. Consistent with previous data, the mutant SAMD9 or SAMD9L allele was underrepresented in the bone marrow. This decrease in the mutant allele paralleled the extent of monosomy 7, suggesting that only the wild-type allele was retained in cells with monosomy 7. Functional analysis revealed that all germline SAMD9/9L mutations displayed a gain-of-function ability to suppress cell cycle progression. The patients with more aggressive disease were found to harbor somatic mutations in SETBP1, ETV6 or RUNX1. In many individuals we found additional somatic revertant mutations in cis with the germline mutation, which overcame the growth-suppressive effects of the germline mutation. Our data show that germline mutations in SAMD9/9L represent a new class of predisposition genes for familial monosomy 7 with MDS/AML and should be screened for in all children with MDS. Citation Format: Victoria Bryant, Jasmine Wong, Jason Schwartz, Tamara Lamprecht, Jing Ma, Charles Mullighan, Mignon Loh, Kevin Shannon, Jeffery Klco. SAMD9/SAMD9L mutations in familial monosomy 7 [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2018; 2018 Apr 14-18; Chicago, IL. Philadelphia (PA): AACR; Cancer Res 2018;78(13 Suppl):Abstract nr 2063.

The DNA methyltransferases DNMT3A and DNMT3B are primarily responsible for de novo methylation of specific cytosine residues in CpG dinucleotides during mammalian development. While loss-of-function mutations in DNMT3A are highly recurrent in acute myeloid leukemia (AML), DNMT3A mutations are almost never found in AML patients with translocations that create oncogenic fusion genes such as PML-RARA, RUNX1-RUNX1T1, and MLL-AF9. Here, we explored how DNMT3A is involved in the function of these fusion genes. We used retroviral vectors to express PML-RARA, RUNX1-RUNX1T1, or MLL-AF9 in bone marrow cells derived from WT or DNMT3A-deficient mice. Additionally, we examined the phenotypes of hematopoietic cells from Ctsg-PML-RARA mice, which express PML-RARA in early hematopoietic progenitors and myeloid precursors, with or without DNMT3A. We determined that the methyltransferase activity of DNMT3A, but not DNMT3B, is required for aberrant PML-RARA–driven self-renewal ex vivo and that DNMT3A is dispensable for RUNX1-RUNX1T1– and MLL-AF9–driven self-renewal. Furthermore, both the PML-RARA–driven competitive transplantation advantage and development of acute promyelocytic leukemia (APL) required DNMT3A. Together, these findings suggest that PML-RARA requires DNMT3A to initiate APL in mice.

Importance Tests that predict outcomes for patients with acute myeloid leukemia (AML) are imprecise, especially for those with intermediate risk AML. Objectives To determine whether genomic approaches can provide novel prognostic information for adult patients with de novo AML. Design, Setting, and Participants Whole-genome or exome sequencing was performed on samples obtained at disease presentation from 71 patients with AML (mean age, 50.8 years) treated with standard induction chemotherapy at a single site starting in March 2002, with follow-up through January 2015. In addition, deep digital sequencing was performed on paired diagnosis and remission samples from 50 patients (including 32 with intermediate-risk AML), approximately 30 days after successful induction therapy. Twenty-five of the 50 were from the cohort of 71 patients, and 25 were new, additional cases. Exposures Whole-genome or exome sequencing and targeted deep sequencing. Risk of identification based on genetic data. Main Outcomes and Measures Mutation patterns (including clearance of leukemia-associated variants after chemotherapy) and their association with event-free survival and overall survival. Results Analysis of comprehensive genomic data from the 71 patients did not improve outcome assessment over current standard-of-care metrics. In an analysis of 50 patients with both presentation and documented remission samples, 24 (48%) had persistent leukemia-associated mutations in at least 5% of bone marrow cells at remission. The 24 with persistent mutations had significantly reduced event-free survival vs the 26 who cleared all mutations (median [95% CI]: 6.0 months [95% CI, 3.7-9.6] for persistent mutations vs 17.9 months [95% CI, 11.3-40.4] for cleared mutations, log-rank P P P = .003; HR, 2.86 [95% CI, 1.39-5.88], P = .004). Among the 32 patients with intermediate cytogenetic risk, the 14 patients with persistent mutations had reduced event-free survival compared with the 18 patients who cleared all mutations (median [95% CI]: 8.8 months [95% CI, 3.7-14.6] for persistent mutations vs 25.6 months [95% CI, 11.4-not estimable] for cleared mutations, log-rank P = .003; HR, 3.32 [95% CI, 1.44-7.67], P = .005) and reduced overall survival (median [95% CI]: 19.3 months [95% CI, 7.5-42.3] for persistent mutations vs 46.8 months [95% CI, 22.6-not estimable] for cleared mutations, log-rank P = .02; HR, 2.88 [95% CI, 1.11-7.45], P = .03). Conclusions and Relevance The detection of persistent leukemia-associated mutations in at least 5% of bone marrow cells in day 30 remission samples was associated with a significantly increased risk of relapse, and reduced overall survival. These data suggest that this genomic approach may improve risk stratification for patients with AML.

HOX genes are highly expressed in many acute myeloid leukemia (AML) samples, but the patterns of expression and associated regulatory mechanisms are not clearly understood. We analyzed RNA sequencing data from 179 primary AML samples and normal hematopoietic cells to understand the range of expression patterns in normal versus leukemic cells. HOX expression in AML was restricted to specific genes in the HOXA or HOXB loci, and was highly correlated with recurrent cytogenetic abnormalities. However, the majority of samples expressed a canonical set of HOXA and HOXB genes that was nearly identical to the expression signature of normal hematopoietic stem/progenitor cells (HSPCs). Transcriptional profiles at the HOX loci were similar between normal cells and AML samples, and involved bidirectional transcription at the center of each gene cluster. Epigenetic analysis of a subset of AML samples also identified common regions of chromatin accessibility in AML samples and normal CD34+ cells that displayed differences in methylation depending on HOX expression patterns. These data provide an integrated epigenetic view of the HOX gene loci in primary AML samples, and suggest that HOX expression in most AML samples represents a normal stem cell program that is controlled by epigenetic mechanisms at specific regulatory elements.

Acute myeloid leukemia (AML) is characterized by dysregulated gene expression and abnormal patterns of DNA methylation; the relationship between these events is unclear. Many AML patients are now being treated with hypomethylating agents, such as decitabine (DAC), although the mechanisms by which it induces remissions remain unknown. The goal of this study was to use a novel stromal coculture assay that can expand primary AML cells to identify the immediate changes induced by DAC with a dose (100nM) that decreases total 5-methylcytosine content and reactivates imprinted genes (without causing myeloid differentiation, which would confound downstream genomic analyses). Using array-based technologies, we found that DAC treatment caused global hypomethylation in all samples (with a preference for regions with higher levels of baseline methylation), yet there was limited correlation between changes in methylation and gene expression. Moreover, the patterns of methylation and gene expression across the samples were primarily determined by the intrinsic properties of the primary cells, rather than DAC treatment. Although DAC induces hypomethylation, we could not identify canonical target genes that are altered by DAC in primary AML cells, suggesting that the mechanism of action of DAC is more complex than previously recognized.

Altered DNA methylation is a well-known feature of acute myeloid leukemia (AML) genomes, but the mechanisms underlying these changes and their relevance for AML pathogenesis are unclear. We previously showed that DNMT3A is the predominant de novo methyltransferase expressed in AML cells, and that the DNMT3AR882H mutation in AML creates a dominant negative protein that reduces in vitro DNA methylation activity by ~80%. Since DNMT3A provides themajority of the methylation activity in AML cells, we hypothesized that AML samples with and without DNMT3AR882H could reveal novel insights about the role of this enzyme in AML initiation and progression. We performed whole-genome bisulfite sequencing (WGBS) of 38 primary human AML samples and 17 normal human hematopoietic cell samples, as well as a remission sample from a patient with a persistent DNMT3AR882H mutation, and blood samples from a non-leukemic patient with a constitutional DNMT3AR882H mutation. We first identified 3,848 differentially methylated regions ('DMRs') between DNMT3AR882H and DNMT3AWT AMLs, virtually all of which were hypomethylated in the DNMT3AR882H AMLs. Further, 28% (1,087/3,848) of these DMRs were also hypomethylated when compared to CD34 cells, implying that these regions are truly hypomethylated in the AML cells with the R882H mutation. In contrast, 72% (2,759/3,848) of the DMRs were unmethylated in bothDNMT3AR882H AMLs and CD34 cells, but were hypermethylated in the DNMT3AWT AML samples. These loci were associated with CpG dense regions, suggesting that they represent abnormal CpG island hypermethylation that occurs only in AML samples with wild-type DNMT3A. Analysis of 21 additional primary AML samples with wild-type DNMT3A identified 4,912 hypermethylated regions compared to CD34 cells, of which 4,544 (92%) were significantly less methylated in DNMT3AR882H AMLs, implying that functional DNMT3A mediates abnormal CpG island hypermethylation in AML. WGBS analysis of two non-leukemic hematopoietic samples with DNMT3AR882H mutations was also performed to understand the direct effects of DNMT3AR882H in non-transformed myeloid cells. These samples included peripheral blood (PB) neutrophils and monocytes from a newly identified 9-year old patient with an overgrowth syndrome and developmental delay (Tatton-Brown et. al., Nature Genetics 2014), who was found to have a heterozygous DNMT3AR882H mutation in all skin and peripheral blood cells. His CBC was normal, and he had no evidence of clonal hematopoiesis by exome sequencing. We identified 2,051 DMRs in his PB myeloid cells, all of which were hypomethylated compared to control PB myeloid cells from his healthy 13-year old brother (and also normal CD34 cells), demonstrating that DNMT3AR882H directly causes focal methylation loss. We also performed WGBS on cells expanded from single stem/progenitor cells from an AML patient with a persistent DNMT3AR882H mutation during remission. Expanded cells with DNMT3AR882H were hypomethylated relative to wild-type DNMT3A cells expanded from the same sample. The majority of the hypomethylated regions were also present in the patient's AML cells, implying that DNMT3AR882H-associated hypomethylation in pre-leukemic cells is maintained during AML progression. These findings demonstrate that DNMT3AR882H-associated hypomethylation precedes leukemia development, and may therefore represent an important initiating phenotype for AML. Our data also suggest that the abnormal hypermethylation of CpG islands in AML cells is DNMT3A-dependent, and must occur during disease progression. This hypermethylation is absent in AMLs with DNMT3AR882H, revealing that it is not required for leukemia progression. We therefore propose a model where DNMT3A-dependent DNA methylation in AML cells acts as a 'brake' that prevents abnormal self-renewal; the abnormal CpG island hypermethylation in DNMT3AWT AMLs may be an adaptive response that is ultimately overcome during leukemia progression. The absence of this 'braking' activity in AMLs with DNMT3AR882H may contribute directly to leukemia initiation. The restoration of DNMT3A activity in AML cells with the DNMT3AR882H mutation is therefore a therapeutic goal. Disclosures Spencer: Cofactor Genomics: Consultancy.

Myelodysplastic syndromes are uncommon in children (incidence of ~2 cases/million) and have a poor prognosis. Despite the wealth of knowledge about the genomic landscape of adult MDS, much less is understood about pediatric MDS, and many recurrent mutations found in adults are not common in children (Hirabayashi, Blood 2012). Furthermore, the clinical presentation, bone marrow morphology, and cytogenetic abnormalities are also different when comparing adult and pediatric MDS, suggesting disparate underlying mechanisms. Here we describe the somatic and germline genomic landscape of pediatric MDS using whole exome sequencing (WES) and RNA-sequencing. We evaluated 88 diagnostic bone marrow samples obtained from the St. Jude Children's Research Hospital Tissue Bank from patients diagnosed between 1988 and 2015. This cohort contains 34 primary MDS, including Refractory Cytopenia of Childhood/RCC (n=19) and Refractory Anemia with Excess Blasts/RAEB (n=15). For comparison, we also included 32 treatment-related (tMDS), 14 MDS/MPN (including 10 Juvenile Myelomonocytic Leukemia/JMML), and 8 cases of AML with Myelodysplasia-Related Changes/AML-MRC (including 5 cases that previously would have been classified as RAEB in transformation/RAEB-T). WES was completed for 87 tumor/normal pairs (tumor only, n=1) using the Nextera Rapid Capture Expanded Exome (Illumina). Normal comparator gDNA was obtained from flow-sorted lymphocytes and variants were classified as germline if present at greater than 30% variant allele frequency (VAF) in the lymphocyte sample; thus, bone marrow mosaicism cannot be excluded. Mean sequencing coverage for the tumor and normal samples were 102x and 105x, respectively. An average of 7.9 variants were observed per patient in the primary MDS cohort (RCC=6.3, RAEB=10.2), compared to 25.5/patient in the tMDS cohort (p=0.001). Copy number information, obtained from WES data, determined that deletions involving chromosome 7 were frequent (n=28, 32%). Approximately 50% of RCC cases had deletions involving chromosome 7 (9 of 19), compared to only 20% of RAEB cases (3 of 15). Amplification of chromosomes 8 (n=7, 8%) and 21 (n=6, 7%), and deletions of 17 (n=5, 6%) were present at low frequency. In total, we detected 43 additional copy number abnormalities (including 9 cryptic chromosome 7 abnormalities) with WES compared to standard conventional karyotyping. RAS/MAPK pathway mutations were present in 42% of the patients (49 total mutations in 37 cases, including 4 germline variants). Fourteen of the 34 primary MDS cases (41%) had at least one RAS/MAPK mutation, including 13 somatic and 2 germline variants. Mutations in RNA splicing genes (germline, n=0; somatic, n=7; 8% of cohort) were less common, unlike what is observed in adult MDS. As expected, 2 patients with JMML harbored germline variants in PTPN11 and NF1. Surprisingly, presumed germline variants were detected in RRAS and NF1 in patients with primary MDS. Germline variants in transcription factors seen in familial MDS/AML (e.g., RUNX1, CEBPA, ETV6, GATA2) were uncommon, although a germline GATA2 variant was found in a single AML-MRC case. RNA-seq using the TruSeq (Illumina) Stranded RNA protocol was performed on 70 samples with suitable RNA. Fusion transcripts were uncommon in primary MDS, while fusions involving KMT2A, NUP98, RUNX1, MECOM, and ETV6were detected in the tMDS and AML-MRC cohorts. Although many of the mutations affecting the RAS/MAPK pathway were in common genes (NRAS, PTPN11, NF1, or CBL), many other mutations were in genes less frequently reported to be mutated in myeloid neoplasms, such as BRAF, SOS1, RIT1 and RRAS. We demonstrated that the mutations in BRAF (G469A, D594N, N581I) and SOS1 (E433K, G328R, S548R) found in our cohort activate the RAS/MAPK pathway to variable levels, as measured by ERK phosphorylation. In addition, expression of the BRAFvariants conferred IL3-independence in Ba/F3 cells. In conclusion, we show that the genomic landscapes of pediatric and adult MDS are different, namely in patterns of RAS/MAPK pathway and RNA splicing gene mutations, thus supporting the notion that MDS in adults and children comprise unique biological entities. The enrichment of RAS/MAPK mutations in pediatric MDS suggests biologic overlap with JMML and may provide direction for future therapeutic options. Disclosures No relevant conflicts of interest to declare.

SummaryMost mutations in cancer genomes are thought to be acquired after the initiating event, which may cause genomic instability and drive clonal evolution. However, for acute myeloid leukemia (AML), normal karyotypes are common, and genomic instability is unusual. To better understand clonal evolution in AML, we sequenced the genomes of M3-AML samples with a known initiating event (PML-RARA) versus the genomes of normal karyotype M1-AML samples and the exomes of hematopoietic stem/progenitor cells (HSPCs) from healthy people. Collectively, the data suggest that most of the mutations found in AML genomes are actually random events that occurred in HSPCs before they acquired the initiating mutation; the mutational history of that cell is “captured” as the clone expands. In many cases, only one or two additional, cooperating mutations are needed to generate the malignant founding clone. Cells from the founding clone can acquire additional cooperating mutations, yielding subclones that can contribute to disease progression and/or relapse.

Most patients with acute myeloid leukaemia (AML) die from progressive disease after relapse, which is associated with clonal evolution at the cytogenetic level. To determine the mutational spectrum associated with relapse, we sequenced the primary tumour and relapse genomes from eight AML patients, and validated hundreds of somatic mutations using deep sequencing; this allowed us to define clonality and clonal evolution patterns precisely at relapse. In addition to discovering novel, recurrently mutated genes (for example, WAC, SMC3, DIS3, DDX41 and DAXX) in AML, we also found two major clonal evolution patterns during AML relapse: (1) the founding clone in the primary tumour gained mutations and evolved into the relapse clone, or (2) a subclone of the founding clone survived initial therapy, gained additional mutations and expanded at relapse. In all cases, chemotherapy failed to eradicate the founding clone. The comparison of relapse-specific versus primary tumour mutations in all eight cases revealed an increase in transversions, probably due to DNA damage caused by cytotoxic chemotherapy. These data demonstrate that AML relapse is associated with the addition of new mutations and clonal evolution, which is shaped, in part, by the chemotherapy that the patients receive to establish and maintain remissions.

Metagenomic characterization of complex biomes remains challenging. Here we describe a modification of digital karyotyping—biome representational in silico karyotyping (BRISK)—as a general technique for analyzing a defined representation of all DNA present in a sample. BRISK utilizes a Type IIB DNA restriction enzyme to create a defined representation of 27-mer DNAs in a sample. Massively parallel sequencing of this representation allows for construction of high-resolution karyotypes and identification of multiple species within a biome. Application to normal human tissue demonstrated linear recovery of tags by chromosome. We apply this technique to the biome of the oral mucosa and find that greater than 25% of recovered DNA is nonhuman. DNA from 41 microbial species could be identified from oral mucosa of two subjects. Of recovered nonhuman sequences, fewer than 30% are currently annotated. We characterized seven prevalent unknown sequences by chromosome walking and find these represent novel microbial sequences including two likely derived from novel phage genomes. Application of BRISK to archival tissue from a nasopharyngeal carcinoma resulted in identification of Epstein-Barr virus infection. These results suggest that BRISK is a powerful technique for the analysis of complex microbiomes and potentially for pathogen discovery.

Acute promyelocytic leukemia (APL) is a subtype of acute myeloid leukemia (AML). It is characterized by the t(15;17)(q22;q11.2) chromosomal translocation that creates the promyelocytic leukemia-retinoic acid receptor α (PML-RARA) fusion oncogene. Although this fusion oncogene is known to initiate APL in mice, other cooperating mutations, as yet ill defined, are important for disease pathogenesis. To identify these, we used a mouse model of APL, whereby PML-RARA expressed in myeloid cells leads to a myeloproliferative disease that ultimately evolves into APL. Sequencing of a mouse APL genome revealed 3 somatic, nonsynonymous mutations relevant to APL pathogenesis, of which 1 (Jak1 V657F) was found to be recurrent in other affected mice. This mutation was identical to the JAK1 V658F mutation previously found in human APL and acute lymphoblastic leukemia samples. Further analysis showed that JAK1 V658F cooperated in vivo with PML-RARA, causing a rapidly fatal leukemia in mice. We also discovered a somatic 150-kb deletion involving the lysine (K)-specific demethylase 6A (Kdm6a, also known as Utx) gene, in the mouse APL genome. Similar deletions were observed in 3 out of 14 additional mouse APL samples and 1 out of 150 human AML samples. In conclusion, whole genome sequencing of mouse cancer genomes can provide an unbiased and comprehensive approach for discovering functionally relevant mutations that are also present in human leukemias.

Acute myeloid leukemia is heterogeneous with respect to clinical outcome and molecular pathogenesis. Approximately 20% of AML cases are refractory to induction chemotherapy, and about 50% of patients ultimately relapse within a time interval that ranges from months to years. At the molecular level, diverse chromosomal abnormalities and genetic mutations have been observed across patients1. Although several clinical factors (age, white blood cell count), cytogenetic aberrations (t[15;17] translocation, loss of chromosome 5) 2-4, and genetic mutations (DNMT3A, FLT3) have been associated with differences in survival 5,6, these factors are of limited prognostic utility. Moreover, few studies have integrated sequence data with clinical and cytogentic factors to build predictive models of patient outcome. Here, we sought to identify genomic predictors of refractory disease or early relapse. We used whole genome and exome sequencing to analyze the genomes of 71 adult de novo AML patients treated with anthracycline and cytarabine-based induction chemotherapy. Of these, 34 had refractory disease or relapsed within 6 months, 12 relapsed in 6-12 months, and 25 had a long first remission (>12 months). We also developed an enhanced exome sequencing (EES) approach to identify and follow leukemia-associated variants over time. In 12 additional patients that achieved morphologic remission after induction chemotherapy, we used EES to identify and track variants at time of diagnosis, time of morphologic remission (roughly 30 days later), and a final time point corresponding to eventual relapse (n=8) or extended remission (n=4). No novel coding or non-coding variants present at the time of diagnosis were found to be predictive of refractory disease or early relapse. Using EES, however, we were able to detect leukemia-associated variants in the initial remission bone marrow in all eight patients who eventually relapsed. One persistent leukemia-associated variant was also detected in one patient still in remission, but all other variants in that patient were eliminated. We also detected 64 somatic variants that became enriched following chemotherapy, but were not detected in the original leukemic cells. These may represent relapse-specific variants or oligoclonal hematopoiesis after bone marrow recovery. Overall, our data suggest that the persistence of leukemia-associated variants after bone marrow recovery from cytotoxic therapy is strongly correlated with relapse, and may be used to complement more traditional, morphologic measures of leukemic cell clearance. 1. Cancer Genome Atlas Research N. Genomic and epigenomic landscapes of adult de novo acute myeloid leukemia. The New England Journal of Medicine 2013;368:2059-74. 2. Byrd JC, Mrozek K, Dodge RK, et al. Pretreatment cytogenetic abnormalities are predictive of induction success, cumulative incidence of relapse, and overall survival in adult patients with de novo acute myeloid leukemia: results from Cancer and Leukemia Group B (CALGB 8461). Blood 2002;100:4325-36. 3. Grimwade D, Hills RK, Moorman AV, et al. Refinement of cytogenetic classification in acute myeloid leukemia: determination of prognostic significance of rare recurring chromosomal abnormalities among 5876 younger adult patients treated in the United Kingdom Medical Research Council trials. Blood 2010;116:354-65. 4. Schlenk RF, Dohner K, Krauter J, et al. Mutations and treatment outcome in cytogenetically normal acute myeloid leukemia. The New England Journal of Medicine 2008;358:1909-18. 5. Kihara R, Nagata Y, Kiyoi H, et al. Comprehensive analysis of genetic alterations and their prognostic impacts in adult acute myeloid leukemia patients. Leukemia 2014;28:1586-95. 6. Ley TJ, Ding L, Walter MJ, et al. DNMT3A mutations in acute myeloid leukemia. The New England Journal of Medicine 2010;363:2424-33. Citation Format: Jeffery M. Klco, Christopher A. Miller, Malachi Griffith, Allegra Petti, David H. Spencer, Shamika Ketkar-Kulkarni, Lukas D. Wartman, Matthew Christopher, Tamara L. Lamprecht, Jacqueline E. Payton, Jack Baty, Sharon E. Heath, Obi L. Griffith, Dong Shen, Jasreet Hundal, Gue Su Chang, Robert S. Fulton, Michelle O'laughlin, Catrina Fronick, Vincent Magrini, Ryan Demeter, David E. Larson, Shashikant Kulkarni, Bradley A. Ozenberger, John S. Welch, Matthew J. Walker, Timothy A. Graubert, Peter Westervelt, Jerald P. Radich, Daniel C. Link, Elaine R. Mardis, John F. DiPersio, Richard K. Wilson. Genomic approaches for risk assessment in acute myeloid leukemia. [abstract]. In: Proceedings of the AACR Special Conference on Translation of the Cancer Genome; Feb 7-9, 2015; San Francisco, CA. Philadelphia (PA): AACR; Cancer Res 2015;75(22 Suppl 1):Abstract nr PR11.

Acute myeloid leukemia is heterogeneous with respect to clinical outcome and molecular pathogenesis. Approximately 20% of AML cases are refractory to induction chemotherapy, and about 50% of patients ultimately relapse within a time interval that ranges from months to years. At the molecular level, diverse chromosomal abnormalities and genetic mutations have been observed across patients1. Although several clinical factors (age, white blood cell count), cytogenetic aberrations (t[15;17] translocation, loss of chromosome 5) 2-4, and genetic mutations (DNMT3A, FLT3) have been associated with differences in survival 5,6, these factors are of limited prognostic utility. Moreover, few studies have integrated sequence data with clinical and cytogentic factors to build predictive models of patient outcome. Here, we sought to identify genomic predictors of refractory disease or early relapse. We used whole genome and exome sequencing to analyze the genomes of 71 adult de novo AML patients treated with anthracycline and cytarabine-based induction chemotherapy. Of these, 34 had refractory disease or relapsed within 6 months, 12 relapsed in 6-12 months, and 25 had a long first remission (>12 months). We also developed an enhanced exome sequencing (EES) approach to identify and follow leukemia-associated variants over time. In 12 additional patients that achieved morphologic remission after induction chemotherapy, we used EES to identify and track variants at time of diagnosis, time of morphologic remission (roughly 30 days later), and a final time point corresponding to eventual relapse (n=8) or extended remission (n=4). No novel coding or non-coding variants present at the time of diagnosis were found to be predictive of refractory disease or early relapse. Using EES, however, we were able to detect leukemia-associated variants in the initial remission bone marrow in all eight patients who eventually relapsed. One persistent leukemia-associated variant was also detected in one patient still in remission, but all other variants in that patient were eliminated. We also detected 64 somatic variants that became enriched following chemotherapy, but were not detected in the original leukemic cells. These may represent relapse-specific variants or oligoclonal hematopoiesis after bone marrow recovery. Overall, our data suggest that the persistence of leukemia-associated variants after bone marrow recovery from cytotoxic therapy is strongly correlated with relapse, and may be used to complement more traditional, morphologic measures of leukemic cell clearance. 1. Cancer Genome Atlas Research N. Genomic and epigenomic landscapes of adult de novo acute myeloid leukemia. The New England Journal of Medicine 2013;368:2059-74. 2. Byrd JC, Mrozek K, Dodge RK, et al. Pretreatment cytogenetic abnormalities are predictive of induction success, cumulative incidence of relapse, and overall survival in adult patients with de novo acute myeloid leukemia: results from Cancer and Leukemia Group B (CALGB 8461). Blood 2002;100:4325-36. 3. Grimwade D, Hills RK, Moorman AV, et al. Refinement of cytogenetic classification in acute myeloid leukemia: determination of prognostic significance of rare recurring chromosomal abnormalities among 5876 younger adult patients treated in the United Kingdom Medical Research Council trials. Blood 2010;116:354-65. 4. Schlenk RF, Dohner K, Krauter J, et al. Mutations and treatment outcome in cytogenetically normal acute myeloid leukemia. The New England Journal of Medicine 2008;358:1909-18. 5. Kihara R, Nagata Y, Kiyoi H, et al. Comprehensive analysis of genetic alterations and their prognostic impacts in adult acute myeloid leukemia patients. Leukemia 2014;28:1586-95. 6. Ley TJ, Ding L, Walter MJ, et al. DNMT3A mutations in acute myeloid leukemia. The New England Journal of Medicine 2010;363:2424-33. This abstract is also presented as a poster at the Translation of the Cancer Genome conference. Citation Format: Jeffery M. Klco, Christopher A. Miller, Malachi Griffith, Allegra Petti, David H. Spencer, Shamika Ketkar-Kulkarni, Lukas D. Wartman, Matthew Christopher, Tamara L. Lamprecht, Jacqueline E. Payton, Jack Baty, Sharon E. Heath, Obi L. Griffith, Dong Shen, Jasreet Hundal, Gue Su Chang, Robert S. Fulton, Michelle O'laughlin, Catrina Fronick, Vincent Magrini, Ryan Demeter, David E. Larson, Shashikant Kulkarni, Bradley A. Ozenberger, John S. Welch, Matthew J. Walker, Timothy A. Graubert, Peter Westervelt, Jerald P. Radich, Daniel C. Link, Elaine R. Mardis, John F. DiPersio, Richard K. Wilson. Genomic approaches for risk assessment in acute myeloid leukemia. [abstract]. In: Proceedings of the AACR Special Conference on Computational and Systems Biology of Cancer; Feb 8-11 2015; San Francisco, CA. Philadelphia (PA): AACR; Cancer Res 2015;75(22 Suppl 2):Abstract nr PR03.

Mutations in the de novo DNA methyltransferase DNMT3A are found in ~25% of patients with acute myeloid leukemia (AML) and most commonly affect codon 882 within the catalytic domain of the protein. We have previously shown that this mutation has dominant negative activity in vitro and is associated with hypomethylation at specific CpG dinucleotides in primary AML samples using array-based methylation data. However, the genome-wide extent and patterns of DNA methylation associated with this hypomethylation are currently unknown. In addition, it is unclear if the methylation differences caused by this mutation result in RNA expression changes at specific targets across the genome, or whether they are associated with altered chromatin structure. To explore the genome-wide consequences of the DNMT3A R882H mutation on DNA methylation and chromatin structure, we carried out whole-genome bisulfite sequencing (WGBS) and transposase-mediated chromatin accessibility profiling (ATAC-seq) on 3 primary normal karyotype AML samples with the DNMT3A R882H mutation and 4 matched AML samples without a DNMT3A mutation. All 7 had the NPMc mutation but lacked mutations in other genes involved in DNA methylation, including IDH1, IDH2, and TET2. WGBS produced methylation data on >93% of the CpGs in the human reference sequence with a median coverage of 7-13x. The overall mean methylation was not statistically different in the samples with R882H mutations, although there was a small but statistically significant difference in the methylation at CpGs in CpG islands (DNMT3A R882H mean: 18.1%, DNMT3A wild-type mean: 21.4%; P=0.02). Differential methylation analysis was performed on ~5 million CpG clusters (median of 5 CpGs per cluster; median cluster size of 202 bp) and identified 95,845 differentially methylated clusters with a mean difference >25% and a q-value < 0.01, the majority of which (88,512; 93%) were hypomethylated in the DNMT3A R882H samples. Using more strict criteria (>50% mean difference) and merging differentially methylated clusters within 50 bp, we identified 2,782 differentially methylated regions (DMRs) with a mean size of 255 bp (median of 11 CpGs), of which 97% were hypomethylated. These DMRs were distributed across the genome and were statistically associated with CpG dense regions, including annotated CpG islands and shores (islands: 1,104 of 2,782; 29.9%; shores: 1,118 of 2,782; 30.3%; P Analysis of chromatin accessibility data from 6 samples (3 DNMT3A R882H and 3 DNMT3A wild-type) showed that a subset of the DNMT3A R882H-associated hypomethylated DMRs (366 of 2,704; 13.5%) were located within 100 bp of an ATAC-seq peak unique to DNMT3A R882H AML samples. Further analysis of all DMRs showed ATAC-seq signal enrichment in the R882H samples specifically at hypomethylated loci (Figure 1). Similar enrichment was not observed in the DNMT3A wild-type AMLs at hypomethylated DMRs (N=78), suggesting that hypomethylation caused by the DNMT3A R882H mutation is specifically associated with changes in chromatin structure. Initial analysis of existing PolyA+ RNA-seq data for these AMLs did not reveal canonical expression changes in annotated genes located near the DMRs, implying that methylation and other epigenetic changes might affect distant genes or previously unannotated RNA species that were not present in our dataset. Efforts to sequence all RNA species present in these samples are therefore underway. In summary, we have conducted an initial analysis of genome-wide, CpG-resolution DNA methylation data from primary AML samples with the DNMT3A R882H mutation. This mutation is associated with a genome-wide, focal hypomethylation phenotype that occurs at small, CpG-dense loci across the genome. We also found that many hypomethylated loci are associated with changes in chromatin structure. These findings represent the first evidence that the methylation changes caused by this mutation can have functional consequences on the epigenetic state of specific loci in AML cells, and set the stage for defining the specific events that are responsible for AML pathogenesis in patients who have this mutation. Figure 1 WGBS (bottom tracks) and chromatin accessibility (ATAC-seq, top tracks) from 3 primary AML samples with the DNMT3A R882H mutation (in red) and 3 with no DNMT3A mutation (in blue) at a hypomethylated locus within the HS3ST3B1 gene. Figure 1. WGBS (bottom tracks) and chromatin accessibility (ATAC-seq, top tracks) from 3 primary AML samples with the DNMT3A R882H mutation (in red) and 3 with no DNMT3A mutation (in blue) at a hypomethylated locus within the HS3ST3B1 gene. Disclosures No relevant conflicts of interest to declare.

Acute promyelocytic leukemia (APL) is an AML subtype that is characterized by aberrant expansion of immature myeloid progenitors and precursors that are arrested at the promyelocyte stage. Almost all APL cases are characterized by the t(15;17)(q22;q11.2) translocation that creates the PML-RARA fusion oncogene. Human APL cells are known to have a canonical expression signature and a specific methylation phenotype that is unique to this form of AML. Our laboratory previously created a mouse model of APL by expressing a human PML-RARA cDNA from the mouse Cathepsin G (Ctsg) locus (Ctsg-PML-RARA), which activates human PML- RARA expression in early myeloid progenitor cells, with peak expression in promyelocytes. After a long latent period (6-12 months), ~60% of these mice develop a clonal, APL-like myeloid malignancy. The long latent period is probably due to the requirement for cooperating mutations that synergize with PML-RARA to accelerate the disease. Human APL samples have a unique gene expression signature that distinguishes them from all other subtypes of AML. We evaluated RNA-Seq data derived from Poly A+ enriched cDNAs obtained from purified promyelocytes derived from 3 young (6 week old) WT and 3 Ctsg-PML-RARA mice. We identified 779 annotated genes that are significantly dysregulated in murine promyelocytes expressing PML-RARA with a log2 fold change >= 2 and P= 2. Differential expression analysis yielded 56 dysregulated lncRNA regions in PML-RARA expressing promyelocytes. To explore the association between gene dysregulation and DNA methylation in promyelocytes, we carried out whole-genome bisulfite sequencing using DNA derived from the purified promyelocytes of a 6 week old Ctsg-PML-RARA mouse, and a WT littermate. We generated a total of approximately 800 million sequencing reads, of which 78% mapped uniquely to the reference genome (mm9); we were able to map ~19 million CpGs with at least 10x coverage. Differential methylation analysis performed on ~4.5 million 1 Kb windows spanning the entire genome identified 17,633 differentially methylated regions with a mean difference of >= 25% and a q-value of < 0.01, the vast majority of which (17,264, 98%) were hypomethylated in the Ctsg-PML-RARA promyelocytes. These windows overlap several known genes, including Runx1, Jak2, Dnmt3a, Gata2, and the Hoxa and Hoxb gene clusters. Using more strict criteria (> 50% mean methylation difference), we identified 87 differentially methylated regions of at least 2 Kb in size. Of these 87 distinct regions, 74 (85%) were hypomethylated in PML-RARA promyelocytes, and 13 were hypermethylated; examples of both as shown in Figure 1. These data strongly suggest that PML-RARA has at least two distinct mechanisms by which it can modify DNA methylation. In regions where CpGs are hypomethylated, PML-RARA may be blocking the normal methylation of CpGs by the de novo DNA methyltransferases Dnmt3a and/or Dnmt3b. In contrast, PML-RARA may be directing de novo methyltransferases to act on the hypermethylated regions. Regardless, these data, when coupled with comprehensive chromatin accessibility mapping and complete RNA sequencing data, should provide new insights into the mechanisms used by PML-RARA to alter gene expression and initiate APL. Figure1. Examples of differentially methylated regions. Black=WT cells. Red=PML-RARA expressing cells. Each CpG in the region is represented as a dot. Scale is 0-100% methylated at each position. Top panel: a region on chromosome 8 that is hypomethylated in PML-RARA expressing promyelocytes. Bottom panel: a region on chromosome 4 that is hypermethylated in PML-RARA expressing promyelocytes. Figure1. Examples of differentially methylated regions. Black=WT cells. Red=PML-RARA expressing cells. Each CpG in the region is represented as a dot. Scale is 0-100% methylated at each position. Top panel: a region on chromosome 8 that is hypomethylated in PML-RARA expressing promyelocytes. Bottom panel: a region on chromosome 4 that is hypermethylated in PML-RARA expressing promyelocytes. Disclosures No relevant conflicts of interest to declare.

Somatic mutations in the DNA methyltransferase, DNMT3A, have been identified in >30% of de novo AML cases with a normal karyotype, and in >10% of patients with MDS and T-ALL. To understand whether mutations in DNMT3A alter hematopoietic development, we generated a transgenic mouse model capable of overexpressing either wild type human DNMT3A or the most common DNMT3A mutation found in AML cases (R882H, a hypomorphic variant that acts as a potent dominant negative inhibitor of WT DNMT3A, D. Germain et al., Cancer Cell 2014). Full-length human cDNAs encoding WT or R882H DNMT3A were cloned into a mammalian expression vector directly downstream from a tetracycline responsive element. This allows for the inducible expression of DNMT3A upon the expression of the rtTA coactivator, and the presence of Doxycycline (Dox). A single founder line for the WT DNMT3A allele, and two founder lines for the R882H DNMT3A allele, were established in the C57Bl6/J background. The WT DNMT3A transgene overexpressed 3.5x more human DNMT3A than endogenous murine DNMT3A in bone marrow cells; R882H DNMT3A transgenic line 1 expressed at a 4.5 fold excess, and R882H line 2 at a 16 fold excess. To determine whether overexpression of the R882H allele was associated with focal DNA hypomethylation in the bone marrow cells of mice (similar to that observed in human AML samples), we used a novel CpG capture approach with bisulfite sequencing to assess 200,000 genomic regions containing ~3 million CpGs in the bone marrow cells of 3 WT C57Bl6/J mice, 3 Dnmt3a null mice, and healthy transgenic mice noted above that had been on Dox chow for either 6 months or 1 year (transgenic mice do not develop hematopoietic malignancies even after one year of transgene induction). We were able to assess 1.6 million CpGs with 10X or greater coverage in all 14 samples. The Dnmt3a null marrow samples contained 188,367 differentially methylated CpGs (average of >25% difference compared to WT bone marrow, q value=99%); the hypomethylated CpGs were nearly identical in all three samples. Marrow cells from the two mice overexpressing the WT DNMT3A gene had only 338 differentially methylated CpGs compared to two matched rtTA control mice; of these, 337 were hypermethylated (>99%). For the two mice overexpressing the R882H allele in line 2 (16x overexpression), bone marrow cells had 2,356 differentially methylated CpGs, of which 2,316 were hypomethylated (98%). Of these CpGs, 1,745 (73%) overlapped with hypomethylated CpGs in the Dnmt3a null marrow samples, indicating that R882H overexpression causes hypomethylation in a subset of CpGs whose methylation in bone marrow cells is Dnmt3a dependent. Because none of our mice developed hematologic malignancies even after one year, but had shown significant hypomethylation in the bone marrow, we hypothesized that cooperating mutations were necessary to produce malignancy. We transduced whole bone marrow cells from four transgenic mice: WT DNMT3A Tg x rtTA; R882H-1 Tg x rtTA; R882H-2 Tg x rtTA; and rtTA only (the same samples analyzed for methylation changes in the previous paragraph) with an MSCV-derived virus containing a human FLT3-ITD cDNA, and transplanted the transduced cells into 8-10 lethally irradiated recipients. Mice of all genotypes succumbed to myeloproliferative disease, T-cell lymphoma, T-lymphoma/ALL, or T-ALL. Overall median latencies were: rtTA=155 days, WT DNMT3A Tg x rtTA=164 days, R882H Tg-1 x rtTA=108.5 days, R882H-2 Tg x rtTA=135.5 days. The average latency for T cell malignancies demonstrated even greater differences among the four genotypes: rtTA n=4, 160.8 +/- 12.49 days (SEM), WT DNMT3A Tg x rtTA n=5, 167.3 +/- 4.854, R882H Tg-1 x rtTA n=3, 124.7 +/- 17.7, R882H-2 Tg x rtTA n=4 124.5 days +/- 22.14. T malignancies derived from R882H expressing cells were especially homogeneous compared to other groups; these tumors were CD4/CD8 double positive in all hematopoietic compartments. Despite the small sample size, these results demonstrate a trend towards a decreased latency for T malignancies in R882H expressing marrow cells, using a FLT3-ITD viral transduction model. We are confirming these data with additional mice. Taken together, our results demonstrate a clear focal hypomethyation phenotype in the bone marrow cells of DNMT3A R882H overexpressing mice, which may lead to increased susceptibility to neoplastic transformation. Disclosures No relevant conflicts of interest to declare.

Therapy-related acute myeloid leukemia (t-AML) and therapy-related myelodysplastic syndrome (t-MDS) are well-recognized complications of cytotoxic chemotherapy and/or radiotherapy. Compared to de novo AML, there is a higher incidence of TP53 mutations, abnormalities of chromosomes 5 or 7, and complex cytogenetics in t-AML. The response to chemotherapy is also reduced compared to de novo AML, and long-term remissions are rare. From a pathogenesis perspective, the key difference between de novo AML and t-AML is the prior exposure of t-AML patients to cytotoxic therapy during treatment of their primary cancer. We previously reported the initial findings of our study to sequence the genome of 22 cases of t-AML (Ramsingh et al, Abstract #784, 2012). These data were compared to the genomic sequence of 49 cases of de novo AML (TCGA, NEJM, 2013). Surprisingly, the data showed that the total number of somatic single nucleotide variants (SNVs) and the percentage of transversions (which are associated with chemotherapy-induced DNA damage) were similar in t-AML and de novo AML. Thus, there is no evidence that chemotherapy induces genome-wide DNA damage in t-AML. We extended this data set by sequencing 149 genes of interest in an additional 89 patients with t-AML/t-MDS. Mutations of TP53 were the most frequent (31.5%), and TP53 was the only gene mutated at a higher frequency in t-AML or t-MDS compared with de novo AML or MDS respectively. The mechanism by which TP53 mutations are selectively enriched in t-AML/t-MDS is unclear. Since t-AML/t-MDS has a similar mutation burden compared to de novo AML, it is not likely that chemotherapy directly induces TP53 mutations. We recently reported that individual HSCs accumulate somatic mutations as a function of age. By age 50, there are on average 5 coding gene mutations per HSC clone (Welch et al, Cell, 2012). As it is estimated that there are 10,000 HSCs in humans, it is possible that HSC clones harboring aging-related TP53 mutations are present in a subset of healthy individuals. Based on these observations, we propose a model in which rare TP53 mutant-bearing HSC clones have a selective growth advantage in patients undergoing chemotherapy. This results in their clonal expansion, with additional mutations in the HSC clone undoubtedly required for full leukemic transformation. This model predicts that TP53 mutations are present in many patients years prior to the development of t-AML/t-MDS. To test this prediction, we developed a next generation sequencing technique able to detect rare subclones harboring TP53 mutations to as low as 1 in 1,000 cells. We identified 7 cases of t-AML/t-MDS with specific TP53 mutations for whom we had blood or bone marrow banked 3-10 years prior to the development of t-AML/t-MDS. The specific TP53 mutations were detected in two cases. In patient UPN 530447, bi-allelic mutations of TP53 were identified in mobilized peripheral blood leukocytes 6 years prior to the development of t-AML at frequencies of 0.5 and 0.7%. In patient UPN 341666, a heterozygous TP53 mutation was identified in mobilized peripheral blood leukocytes 3 years prior to t-MDS diagnosis at a frequency of 0.1%. In at least some cases, we have clear evidence that the TP53 mutations occurred prior to the acquisition of other driver mutations or the development of cytogenetic abnormalities. To directly test the hypothesis that TP53 mutations confer a clonal advantage after chemotherapy, we generated mixed bone marrow chimeras containing wild type and Tp53+/- cells. No clonal advantage of Tp53+/- cells was identified in untreated chimeras. However, upon treatment with N-ethyl-N-nitrosourea, Tp53+/- HSCs had a significant growth advantage. Collectively, these data suggest that HSCs that acquire heterozygous TP53 mutations as a function of normal aging may be selected for in the presence of cytotoxic therapy. This provides a potential mechanism for the high incidence of TP53 mutations in t-AML/t-MDS. The early acquisition of TP53 mutations in the founding clone likely contributes to the frequent cytogenetic abnormalities and poor response to chemotherapy that are typical of t-AML/t-MDS. Disclosures: Welch: Eisai: Research Funding.

Mutations in DNMT3A (encoding one of two mammalian de novo DNA methyltransferases) are found in >30% of normal karyotype AML cases and correlate with poor clinical outcomes. Most DNMT3A mutations occur at position R882 within the catalytic domain (most commonly R882H) and are virtually always heterozygous. This over-representation suggests that mutations at R882 may result in gain-of-function or dominant-negative activity that contributes to leukemogenesis. However, how DNA methylation might be altered in DNMT3A-mutant cases of AML remains unclear, and no published study to date has addressed the effects of mixing wild-type (WT) and R882H DNMT3A. Importantly, mouse HSPCs deficient in Dnmt3a dramatically expand over time and have a concurrent defect in differentiation (Challen, GA et al. Nat Genet, 2011). Mice haploinsufficient for Dnmt3a, on the other hand, do not have a measurable defect in hematopoiesis. Collectively, these data suggest that the heterozygous R882 mutations probably cause more than a simple loss-of-function phenotype. We purified full-length, human WT and R882H DNMT3A using a mammalian tissue culture system to produce recombinant proteins for biochemical modeling of the de novo methylation potential of a DNMT3A-mutant AML cell. rhR882H DNMT3A exhibits roughly 10-20% of the de novo DNA methyltransferase activity of rhWT DNMT3A, similar to observations by other groups. We added increasing amounts of R882H DNMT3A to a fixed amount of WT DNMT3A and observed a linear increase in the net enzymatic activity, reflecting the summed activity of the two forms of DNMT3A in these 4-hour in vitro reactions. In contrast, 12-hour in vitro DNA methylation assays with mixed WT and R882H DNMT3A demonstrated net methylation less than the predicted summed activity of the two enzymes, suggesting that a dominant-negative effect of R882H DNMT3A may occur with a long equilibration time. To better simulate an AML cell with a heterozygous R882H mutation, we co-transfected HEK293T cells with equal amounts of poly-His-tagged WT and R882H DNMT3A expression vectors. Subsequently co-purified (i.e. in vivo-mixed) WT and R882H DNMT3A exhibited a striking reduction in methyltransferase activity, with total activity similar to R882H DNMT3A alone (Figure 1A). TSQ mass spectrometry allowed us to verify the presence and quantify the relative concentration of WT and R882H DNMT3A in our co-purified samples. We exploited a novel tryptic cleavage site in DNMT3A produced by the R882H mutation to generate standard concentration curves using recombinant peptides distinguishing the two protein forms. Our co-purified enzyme preparations had WT:R882H ratios ranging from 0.79 to 1.60; all demonstrated the dominant-negative effect of R882H. DNMT3A is a processive enzyme, catalyzing multiple methyl-group transfers before dissociating from target DNA. This is dependent on the ability of WT DNMT3A to form homo-oligomers (tetramers and larger), which was recently shown to be disrupted by the R882H mutation using the catalytic domain of DNMT3A produced in E.coli (Holz-Schietinger, C et al. JBC, 2012). We therefore postulated that the dominant-negative effect of R882H may be due to the disruption of WT DNMT3A oligomerization. Using a Superose 6 size exclusion column, we confirmed the tetramerization defect of R882H DNMT3A relative to WT DNMT3A. Notably, in vivo-mixed (co-purified) WT and R882H DNMT3A complexes exhibited a pattern of oligomerization identical to R882H DNMT3A alone. However, WT and R882H DNMT3A mixed in vitro exhibited a distribution of oligomers corresponding to the expected average of the WT and R882H curves (Figure 1B). These data demonstrate that production of equal amounts of WT and R882H DNMT3A within the same cell provides an environment where R882H DNMT3A can exert a potent dominant-negative effect on WT DNMT3A. Furthermore, our data suggest that this effect is associated with diminished formation of tetramers when WT and R882H DNMT3A are complexed together. Thus, the R882H mutation has two distinct consequences that affect DNMT3A activity in AML cells: 1) it severely reduces its own de novo methyltransferase activity, and 2) it disrupts the ability of WT DNMT3A to form functional tetramers. These two effects severely reduce total DNMT3A activity in AML cells, and may explain why this mutation is virtually always heterozygous in AML samples, since homozygosity would not further reduce DNMT3A activity. Disclosures: No relevant conflicts of interest to declare.

Most recurrent somatic mutations in acute myeloid leukemia (AML) have been identified. However, it is now clear that many mutations are not always present in all the cells within an AML sample. Our previous studies have shown that all AML samples are comprised of a founding clone and usually one or more subclones that are derived from the founding clone. The clonal architecture of an AML sample can be identified using single nucleotide variants (SNVs) that cluster according to discrete variant allele fractions (VAFs); the accurate identification of these clusters in AML samples generally requires deep digital sequencing of all the variants identified by whole genome sequencing (WGS), since AML samples have so few mutations in coding sequences. However, it is not yet clear whether human AML samples engrafted in immunodeficient mice accurately recapitulate the subclonal architecture of the injected sample. Similarly, it not yet clear whether the presence of human cytokines, such as those expressed in the NSG-SGM3 strain (NSG mice with transgenes expressing human IL-3, SCF, and GM-CSF) would alter the engrafting potential of individual subclones within an injected sample. We injected 1 million bulk (unsorted and unmanipulated) cells from 9 different oligoclonal, de novo AML samples into 73 mice (31 NSG and 42 NSG-SGM3) via lateral tail vein injections. Five of the 9 samples were concurrently injected into both NSG and NSG-SGM3 mice. Engraftment was then assessed at 14 weeks (or the first sign of illness) using flow cytometry for human CD45, CD33 and/or CD34. Eight of the 9 samples engrafted, and 58 of the 73 mice had detectable leukemia at 14 weeks (4-11 mice per engrafted AML sample). Consistent with previous studies, engraftment was more efficient and robust in the NSG-SGM3 strain: 39/42 (92.9%) NSG-SGM3 mice had greater than 1% human AML in the bone marrow, compared to 19/31 NSG (61.3%). NSG-SGM3 mice had an average 47.5% human AML cells in the bone marrow, compared to 23.9% in NSG mice (P We performed targeted deep sequencing on human AML cells purified from the engrafted mice to define the VAFs of all known somatic variants previously identified by whole genome sequencing for these cases. We have evaluated 36 xenografts from 5 different AML samples; 3 of these samples were injected into both NSG strains. Despite multiple subclones present per sample at the time of injection, 31 of 36 xenografts only contained cells from a single subclone. 3/5 samples consistently engrafted the same subclone, while multiple different subclones engrafted from the other two samples. Preferential engraftment and outgrowth of subclones that comprised In summary, engraftment of AML samples in NSG mice (and NSG-SGM3 mice) typically results in skewing of the sample’s original clonal architecture with a dramatic restriction in the number of subclones. Further, the expression of human hematopoietic cytokines can influence the engraftment and/or outgrowth of specific subclones. These findings suggest that rigorous analyses of samples before and after xenotransplantation are necessary to define the subclones (and the mutations contained therein) for the testing of novel targeted therapies. Disclosures: No relevant conflicts of interest to declare.

Abstract 407 Acute myeloid leukemia (AML) is a biologically heterogenous malignancy of hematopoietic cells. All AML samples are comprised of a founding clone and usually one or more subclones that are derived from the founding clone; subclones can gain or lose mutations as they evolve from the founding clone, and often become dominant at relapse1. The clonal architecture of an AML sample can be identified using single nucleotide variants (SNVs) that cluster according to discrete variant allele frequencies (VAFs). To accurately identify clusters with common VAFs, deep digital sequencing must be performed using all of the SNVs present in each genome (hundreds of events). In this report, we studied the subclonal architecture of AML samples from UPN 452198, collected from a 55-year old woman with normal karyotype acute monocytic leukemia (FAB M5) with a high peripheral WBC count (72,700/mm3) at presentation. This bone marrow sample contained a founding clone and 3 subclones at presentation. The SNVs of the founding clone had a mean VAF of 46.4% (i.e. heterozygous mutations found in 92.8% of the cells in the bone marrow sample), including DNMT3A R882H and NPM1 W288 frameshift mutations. The mean VAFs of Subclones 1, 2 and 3 were 31.2%, 12.0% and 2.4%, respectively, and they contained all of the variants in the founding clone, along with additional variants, most notably FLT3 D835H and IDH1 R132H mutations in Subclone 1. The tumor at relapse consisted entirely of Subclone 3, which also contained 42 relapse-specific variants. We designed an oligonucleotide capture reagent to track all 118 de novo and relapse-specific variants, and obtained deep read counts (mean coverage per site: 412 reads/SNV) on the de novo AML sample under different experimental and biological conditions, as follows: 1) We showed that peripheral blood and bone marrow leukemia samples obtained at the same time had nearly identical clonal architectures. We verified this correlation using 4 additional AML samples, suggesting that clonal architecture is preserved in the peripheral blood for many AML samples. 2) We flow-sorted the leukemic peripheral blood sample into blasts, monocytes, and lymphocytes based on side-scatter characteristics and expression of CD45 and CD33. The founding clone and all three subclones were detected in the monocyte population, which was the predominant leukemic cell population in the peripheral blood. By flow cytometry, blasts comprised only 3.3% of the cells, but were strongly enriched for variants in Subclone 3 (mean VAF in sorted blasts 33.9% versus 3.0% in unsorted peripheral blood, p Disclosures: Ley: Washington University: Patents & Royalties.

Abstract 1329 De novo CpG methylation is catalyzed by two enzymes (DNMT3A and DNMT3B), while DNMT1 is responsible for maintenance methylation during cell replication. DNMT3L, a catalytically inactive protein, interacts with and influences DNMT3A and DNMT3B target preference and methylation kinetics. Recurrent mutations in DNMT3A have been found in over 20% of patients with acute myeloid leukemia (AML) and have been associated with poor clinical outcomes (Ley, TJ et al. NEJM, 2010). Greater than 50% of DNMT3A mutations are found at position R882 within the catalytic domain. Because R882H mutations in AML are nearly always heterozygous, because the mutant allele is expressed at the same level as the corresponding WT allele (Ley, TJ et al. NEJM, 2010), and because the mutant enzyme has reduced methyltransferase activity (Yamashita, Y et al. Oncogene, 2010; Holz-Schietinger, C et al. JBC, 2012), it has been suggested that the R882H mutation contributes to leukemogenesis by leading to haploinsufficiency for DNMT3A. However, mice haploinsufficient for Dnmt3a exhibit normal hematopoiesis, while HSPCs lacking Dnmt3a exhibit increased self-renewal and decreased differentiation after serial transplantation (Challen, GA et al. Nat Genet, 2011). To address this conundrum, we have studied the R882H mutation in a setting that mimics the intrinsic de novo methylation capacity of a typical AML cell. Using expression array and RNA-Seq data from 178 AML patients, we discovered that DNMT3L is not expressed in AML cells, and that DNMT3A is expressed on average 2.3-fold higher than DNMT3B. Interestingly, 92% of AML patients predominantly express inactive splice variants of DNMT3B, regardless of FAB or mutational profile (median ratio of inactive to active DNMT3B transcripts is 3.1:1). Given that the inactive splice variant DNMT3B3 is the most highly expressed isoform in most patients in our cohort, we explored the functional interactions between WT DNMT3A, R882H DNMT3A, and DNMT3B3 using recombinant enzymes made in eukaryotic cells. In vitro methylation of plasmid DNA (pcDNA3.1) with 3H-SAM using purified recombinant full-length human DNMT3A protein confirmed that the R882H mutation severely reduces the catalytic activity of DNMT3A, resulting in an enzyme with ∼10% of the activity of the WT enzyme. These results were verified by independent in vitro methylation experiments analyzed by bisulfite sequencing, which also revealed that the CpG-flanking sequence preferences of WT and R882H DNMT3A are identical and consistent with the expected “TNCGCY” motif previously described (Wienholz, BL et al. PLoS Genet, 2010). Mixing WT and R882H DNMT3A at equimolar ratios resulted in no significant changes in CpG-flanking sequence preference (compared to WT or R882H enzyme alone; Spearman correlation between WT DNMT3A and WT+R882H DNMT3A = 0.99). In contrast, mixing WT and R882H DNMT3A at equimolar ratios in a 12-hour methylation assay demonstrated that R882H DNMT3A exerts an inhibitory effect on the catalytic activity of WT DNMT3A in vitro. Instead of increasing net methylation activity by a predicted 10% (summing the activity of the two individual enzymes), R882H DNMT3A led to a 20% reduction in the measured methylation. Similarly, the addition of catalytically inactive DNMT3B3 to WT DNMT3A resulted in a mean decrease in methylation of 38%. Combining equimolar amounts of WT DNMT3A, R882H DNMT3A, and DNMT3B3 led to an additive inhibition of methylation compared to WT DNMT3A alone (62% decrease; p < 0.001; Figure 1). This scenario closely mimics the ratio of these enzymes in AML cells, and our data therefore suggest that the additive inhibitory effects of R882H DNMT3A and DNMT3B3 could severely reduce the total de novo methylation activity of DNMT3A in AML cells. The reduction of enzyme activity below haploinsufficient levels may be important for AML pathogenesis, and these findings provide a mechanism to achieve these levels. Figure 1: The de novo methyltransferase activity of WT DNMT3A is inhibited by R882H DNMT3A and DNMT3B3. Mixing equimolar amounts of WT DNMT3A, R882H DNMT3A, and DNMT3B3 leads to additive inhibition of methylation by 62% (p < 0.001). Figure 1:. The de novo methyltransferase activity of WT DNMT3A is inhibited by R882H DNMT3A and DNMT3B3. Mixing equimolar amounts of WT DNMT3A, R882H DNMT3A, and DNMT3B3 leads to additive inhibition of methylation by 62% (p < 0.001). Disclosures: Ley: Washington University: Patents & Royalties.

Abstract 2527 Acute myeloid leukemia (AML) is a hematopoietic neoplasm with high mortality that is typically treated with daunorubicin/cytarabine induction chemotherapy. Alternative therapies with cytosine analogs such as decitabine are also used in some cases with a variable clinical response that some have estimated to be as high as 25%. The mechanism of these agents is unclear, but at low doses they produce passive DNA hypomethylation by inhibiting DNMT1. Although the impact of these drugs on cell growth and DNA methylation in AML cell lines has been evaluated1, studies using primary cells are limited; importantly, most have involved extended drug treatments that may be confounded by the differentiation of the treated cells2. In addition, some evidence suggests that decitabine has a differential effect on methylation in patients who respond to treatment2, but the utility of this phenotype as an in vitro biomarker for decitabine responsiveness is unknown. In this study, we used a novel in vitro culture system for primary leukemia cells to explore the initial genomic effects of short-term low dose decitabine on primary samples from 22 AML patients. Primary bone marrow or blood samples from these patients were cultured on HS27 stromal cells in DMEM supplemented with beta-mercaptoethanol and 15% FBS along with hSCF, hIL3, hIL-6, hTPO and hFLT3L for an initial 4-day period prior to daily treatment for 3 days with either 100 nM decitabine, 100 nM cytarabine, or vehicle controls. Cells were then evaluated for growth, cell cycle effects, and differentiation (by flow cytometry and morphologic evaluation). DNA was prepared from all samples for 5-methylcytosine content measurements by mass spectrometry, and 8 samples were selected for genome-wide methylation and gene expression profiling with the Illumina Human Methylation 450 and Affymetrix Human Exon 1.0ST array platforms. Mass spectrometry revealed a mean decrease in 5-mdC of 29% (range: 13% to 62%) in the decitabine-treated samples; in comparison, cytarabine treatment resulted in a mean increase in 5-mdC of 5% (range: −10% to 37%). Methylation arrays also showed a modest shift toward lower methylation values, but unsupervised hierarchical clustering demonstrated that methylation patterns were driven by sample-specific differences and not drug treatment. Analysis of methylation changes showed the most pronounced hypomethylation at CpGs with high baseline methylation levels, irrespective of CpG island and gene-based annotation, suggesting that the initial methylation status of each CpG is responsible for preferential effects of decitabine, rather than its genomic context. Methylation at promoter-associated CpGs showed a small but statistically significant negative correlation with change in gene expression, but expression changes at individual genes were not consistent across the samples, including genes previously shown to be regulated by methylation-dependent mechanisms (eg. CDKN2B and CDx H1). In addition to these findings, we observed that a sample from a long-term decitabine responder had an exaggerated in vitro response to decitabine (58% decrease in 5-mdC after 6 days of treatment), compared to a cohort of decitabine non-responders; a sample from a second patient also showed marked hypomethylation by both mass spectrometry and methylation array, although this patient was not treated with decitabine. While more investigation is needed, this observation might suggest that extreme in vitro hypomethylation in response to decitabine could serve as a biomarker for a clinical response. In summary, our study showed that short-term low dose decitabine treatment has modest but detectable effects on DNA methylation and gene expression, but these changes did not result in activation of any canonical gene expression pathway at this early time point. We found that the baseline methylation status of a CpG appears to be the best predictor of decitabine-induced hypomethylation, with highly methylated CpGs showing the greatest change. We also observed that hypomethylation is highly variable across primary samples and at specific genes, implying that single gene approaches for measuring decitabine effect may be problematic. Finally, extreme in vitro decitabine-induced hypomethylation should be further investigated as a biomarker for decitabine responsiveness. Disclosures: No relevant conflicts of interest to declare.

Because PML-RARA-induced acute promyelocytic leukemia (APL) is a morphologically differentiated leukemia, many groups have speculated about whether its leukemic cell of origin is a committed myeloid precursor (e.g. a promyelocyte) versus an hematopoietic stem/progenitor cell (HSPC). We originally targeted PML-RARA expression with CTSG regulatory elements, based on the early observation that this gene was maximally expressed in cells with promyelocyte morphology. Here, we show that both Ctsg, and PML-RARA targeted to the Ctsg locus (in Ctsg-PML-RARA mice), are expressed in the purified KLS cells of these mice (KLS = Kit(+)Lin(-)Sca(+), which are highly enriched for HSPCs), and this expression results in biological effects in multi-lineage competitive repopulation assays. Further, we demonstrate the transcriptional consequences of PML-RARA expression in Ctsg-PML-RARA mice in early myeloid development in other myeloid progenitor compartments [common myeloid progenitors (CMPs) and granulocyte/monocyte progenitors (GMPs)], which have a distinct gene expression signature compared to wild-type (WT) mice. Although PML-RARA is indeed expressed at high levels in the promyelocytes of Ctsg-PML-RARA mice and alters the transcriptional signature of these cells, it does not induce their self-renewal. In sum, these results demonstrate that in the Ctsg-PML-RARA mouse model of APL, PML-RARA is expressed in and affects the function of multipotent progenitor cells. Finally, since PML/Pml is normally expressed in the HSPCs of both humans and mice, and since some human APL samples contain TCR rearrangements and express T lineage genes, we suggest that the very early hematopoietic expression of PML-RARA in this mouse model may closely mimic the physiologic expression pattern of PML-RARA in human APL patients.

Mice lacking rods and cones retain pupillary light reflexes that are mediated by intrinsically photosensitive retinal ganglion cells (ipRGCs). Melanopsin is necessary and sufficient for this nonvisual photoreception. The mammalian inner retina also expresses the potential blue light photopigments cryptochromes 1 and 2. Previous studies have shown that outer retinal degenerate mice lacking cryptochromes have lower nonvisual photic sensitivity than retinal degenerate mice, suggesting a role for cryptochrome in inner retinal photoreception.Nonvisual photoreception (pupillary light responses, circadian entrainment, and in vitro sensitivity of intrinsically photosensitive retinal ganglion cells) were studied in wild-type, rd/rd, and circadian clock-mutant mice with and without rd/rd mutation.Loss of cryptochrome in retinal degenerate mice reduces the sensitivity of the pupillary light response at all wavelengths but does not alter the form of the action spectrum, suggesting that cryptochrome does not function as a photopigment in the inner retina. The authors compounded the rd/rd retinal degeneration mutation with mutations in other essential circadian clock genes, mPeriod and Bmal1. Both mPeriod1⁻/⁻; mPeriod2⁻/⁻;rd/rd and Bmal1⁻/⁻;rd/rd mice showed significantly lower pupillary light sensitivity than rd/rd mice alone. A moderate amplitude (0.5 log) circadian rhythm of pupillary light responsiveness was observed in rd/rd mice. Multielectrode array recordings of ipRGC responses of mCryptochrome1⁻/⁻;mCryptochrome2⁻/⁻ and mPeriod1⁻/⁻;mPeriod2⁻/⁻ mice showed minimal sensitivity decrement compared with wild-type animals. mCryptochrome1⁻/⁻;mCryptochrome2⁻/⁻;rd/rd, mPeriod1⁻/⁻;mPeriod2⁻/⁻;rd/rd and Bmal1⁻/⁻;rd/rd mice all showed comparable weak behavioral synchronization to a 12-hour light/12-hour dark cycle.The effect of cryptochrome loss on nonvisual photoreception is due to loss of the circadian clock nonspecifically. The circadian clock modulates the sensitivity of nonvisual photoreception.

Abstract 404 To characterize the genomic events associated with distinct subtypes of AML, we used whole genome sequencing to compare 24 tumor/normal sample pairs from patients with normal karyotype (NK) M1-AML (12 cases) and t(15;17)-positive M3-AML (12 cases). All single nucleotide variants (SNVs), small insertions and deletions (indels), and cryptic structural variants (SVs) identified by whole genome sequencing (average coverage 28x) were validated using sample-specific custom Nimblegen capture arrays, followed by Illumina sequencing; an average coverage of 972 reads per somatic variant yielded 10,597 validated somatic variants (average 421/genome). Of these somatic mutations, 308 occurred in 286 unique genes; on average, 9.4 somatic mutations per genome had translational consequences. Several important themes emerged: 1) AML genomes contain a diverse range of recurrent mutations. We assessed the 286 mutated genes for recurrency in an additional 34 NK M1-AML cases and 9 M3-AML cases. We identified 51 recurrently mutated genes, including 37 that had not previously been described in AML; on average, each genome had 3 recurrently mutated genes (M1 = 3.2; M3 = 2.8, p = 0.32). 2) Many recurring mutations cluster in mutually exclusive pathways, suggesting pathophysiologic importance. The most commonly mutated genes were: FLT3 (36%), NPM1 (25%), DNMT3A (21%), IDH1 (18%), IDH2 (10%), TET2 (10%), ASXL1 (6%), NRAS (6%), TTN (6%), and WT1 (6%). In total, 3 genes (excluding PML-RARA) were mutated exclusively in M3 cases. 22 genes were found only in M1 cases (suggestive of alternative initiating mutations which occurred in methylation, signal transduction, and cohesin complex genes). 25 genes were mutated in both M1 and M3 genomes (suggestive of common progression mutations relevant for both subtypes). A single mutation in a cell growth/signaling gene occurred in 38 of 67 cases (FLT3, NRAS, RUNX1, KIT, CACNA1E, CADM2, CSMD1); these mutations were mutually exclusive of one another, and many of them occurred in genomes with PML-RARA, suggesting that they are progression mutations. We also identified a new leukemic pathway: mutations were observed in all four genes that encode members of the cohesin complex (STAG2, SMC1A, SMC3, RAD21), which is involved in mitotic checkpoints and chromatid separation. The cohesin mutations were mutually exclusive of each other, and collectively occur in 10% of non-M3 AML patients. 3) AML genomes also contain hundreds of benign “passenger” mutations. On average 412 somatic mutations per genome were translationally silent or occurred outside of annotated genes. Both M1 and M3 cases had similar total numbers of mutations per genome, similar mutation types (which favored C>T/G>A transitions), and a similar random distribution of variants throughout the genome (which was affected neither by coding regions nor expression levels). This is consistent with our recent observations of random “passenger” mutations in hematopoietic stem cell (HSC) clones derived from normal patients (Ley et al manuscript in preparation), and suggests that most AML-associated mutations are not pathologic, but pre-existed in the HSC at the time of initial transformation. In both studies, the total number of SNVs per genome correlated positively with the age of the patient (R2 = 0.48, p = 0.001), providing a possible explanation for the increasing incidence of AML in elderly patients. 4) NK M1 and M3 AML samples are mono- or oligo-clonal. By comparing the frequency of all somatic mutations within each sample, we could identify clusters of mutations with similar frequencies (leukemic clones) and determined that the average number of clones per genome was 1.8 (M1 = 1.5; M3 = 2.2; p = 0.04). 5) t(15;17) is resolved by a non-homologous end-joining repair pathway, since nucleotide resolution of all 12 t(15;17) breakpoints revealed inconsistent micro-homologies (0 – 7 bp). Summary: These data provide a genome-wide overview of NK and t(15;17) AML and provide important new insights into AML pathogenesis. AML genomes typically contain hundreds of random, non-genic mutations, but only a handful of recurring mutated genes that are likely to be pathogenic because they cluster in mutually exclusive pathways; specific combinations of recurring mutations, as well as rare and private mutations, shape the leukemia phenotype in an individual patient, and help to explain the clinical heterogeneity of this disease. Disclosures: Westervelt: Novartis: Speakers Bureau.

Abstract 99 Whole genome sequencing with next generation technologies represents a new, unbiased approach for discovering somatic variations in cancer genomes. Our group recently reported the DNA sequence and analysis of the genomes of two patients with normal karyotype acute myeloid leukemia (AML). Improvements in next generation sequencing technologies (principally, paired-end sequencing) led us to reevaluate the first case (Ley et al, Nature 456:66–72, 2008) with deeper sequence coverage. We discovered a novel frameshift mutation in DNMT3A, one of the three genes in humans (DNMT1, DNMT3A, and DNMT3B) that encodes a DNA methyltransferase that catalyzes the addition of methyl groups to cytosine within CpG dinucleotides. We then sequenced all the coding exons of this gene in 280 additional de novo cases of AML to define recurring mutations. 62/281 de novo AML cases (22%) had mutations with translational effects in the DNMT3A gene. 18 different missense mutations were identified, the most common of which was at amino acid R882 (37 cases). Frameshifts (n=6), nonsense mutations (n=6), splice site mutations (n=3), and a 1.5 Mbp deletion that included the DNMT3A gene were also identified. DNMT3A mutations were highly enriched in cases with intermediate risk cytogenetics (56/166=33.7%; p Disclosures: Westervelt: Novartis: Honoraria; Celgene: Honoraria, Speakers Bureau. DiPersio:Genzyme: Honoraria.