Here we sequence 633 genes, encoding the majority of known epigenetic regulatory proteins, in over 1000 pediatric tumors to define the landscape of somatic mutations in epigenetic regulators in pediatric cancer. Our results demonstrate a marked variation in the frequency of gene mutations across 21 different pediatric cancer subtypes, with the highest frequency of mutations detected in high-grade gliomas, T-lineage acute lymphoblastic leukemia, medulloblastoma, and a paucity of mutations in low-grade glioma, and retinoblastoma. The most frequently mutated genes are H3F3A, PHF6, ATRX, KDM6A, SMARCA4, ASXL2, CREBBP, EZH2, MLL2, USP7, ASXL1, NSD2, SETD2, SMC1A, and ZMYM3. Importantly, we identify novel loss-of-function mutations in the ubiquitin-specific-processing protease 7 (USP7) in pediatric leukemia, which result in a decrease in deubiquitination activity. Collectively, our results help to define the landscape of mutations in epigenetic regulatory genes in pediatric cancer and yield a valuable new database for investigating the role of epigenetic dysregulations in cancer.

Acute myeloid leukemia (AML) comprises a heterogeneous group of leukemias frequently defined by recurrent cytogenetic abnormalities, including rearrangements involving subunits of the core-binding factor (CBF) transcriptional complex. To better understand the genomic landscape of CBF-AMLs, we analyzed both pediatric (n=87) and adult (n=78) samples, including cases with RUNX1-RUNX1T1 (n=85) or CBFB-MYH11 (n=80) rearrangements, by whole-genome or whole-exome sequencing. In addition to previously reported somatic mutations in the Ras signaling pathway, we identified recurrent stabilizing mutations in CCND2, suggesting a recurrent and previously unappreciated cooperating pathway in CBF-AML. Outside of signaling alterations, RUNX1-RUNX1T1 and CBFB-MYH11 AMLs demonstrated a remarkably different spectrum of cooperating mutations as RUNX1-RUNX1T1 cases harbored recurrent somatic mutations in DHX15 and ZBTB7A, as well as an enrichment of somatic mutations in epigenetic regulators, including ASXL2, and in components of the cohesin complex. This detailed analysis provides insights into the pathogenesis and development of CBF-AML, while highlighting dramatic differences in the landscape of cooperating mutations between these related AML subtypes.

The AML1:CBFbeta transcription factor complex is essential for definitive hematopoiesis. Null mutations in mouse AML1 result in midgestational lethality with a complete lack of fetal liver hematopoiesis. While the cell autonomous nature and expression pattern of AML1 suggest an intrinsic role for this transcription factor in the developing hematopoietic system, no direct link to a functional cell type has been made. Here, we examine the consequences of AML1 loss in hematopoietic stem cells (HSC) of the mouse embryo. We demonstrate an absolute requirement for AML1 in functional HSCs. Moreover, haploinsufficiency results in a dramatic change in the temporal and spatial distribution of HSCs, leading to their early appearance in the normal position in the aorta-gonad-mesonephros region and also in the yolk sac.

Pediatric de novo acute myeloid leukemia (AML) is an aggressive malignancy with current therapy resulting in cure rates of only 60%. To better understand the cause of the marked heterogeneity in therapeutic response and to identify new prognostic markers and therapeutic targets a comprehensive list of the genetic mutations that underlie the pathogenesis of AML is needed. To approach this goal, we examined diagnostic leukemic samples from a cohort of 111 children with de novo AML using single-nucleotide-polymorphism microarrays and candidate gene resequencing. Our data demonstrate that, in contrast to pediatric acute lymphoblastic leukemia (ALL), de novo AML is characterized by a very low burden of genomic alterations, with a mean of only 2.38 somatic copy-number alterations per leukemia, and less than 1 nonsynonymous point mutation per leukemia in the 25 genes analyzed. Even more surprising was the observation that 34% of the leukemias lacked any identifiable copynumber alterations, and 28% of the leukemias with recurrent translocations lacked any identifiable sequence or numerical abnormalities. The only exception to the presence of few mutations was acute megakaryocytic leukemias, with the majority of these leukemias being characterized by a high number of copy-number alterations but rare point mutations. Despite the low overall number of lesions across the patient cohort, novel recurring regions of genetic alteration were identified that harbor known, and potential new cancer genes. These data reflect a remarkably low burden of genomic alterations within pediatric de novo AML, which is in stark contrast to most other human malignancies.copy number alterations ͉ single-nucleotide-polymorphism (SNP) ͉ microarray ͉ candidate gene resequencing ͉ loss-of-heterozygosity (LOH)

To investigate the frequency of isocitrate dehydrogenase 1 (IDH1) and 2 (IDH2) mutations in pediatric acute myeloid leukemia (AML) and acute lymphoid leukemia (ALL), we sequenced these genes in diagnostic samples from 515 patients (227 AMLs and 288 ALLs). Somatic IDH1/IDH2 mutations were rare in ALL (N=1), but were more common in AML, occurring in 3.5% (IDH1 N=3 and IDH2 N=5), with the frequency higher in AMLs with a normal karyotype (9.8%). The identified IDH1 mutations occurred in codon 132 resulting in replacement of arginine with either cysteine (N=3) or histidine (N=1). By contrast, mutations in IDH2 did not affect the homologous residue but instead altered codon 140, resulting in replacement of arginine with either glutamine (N=4) or tryptophan (N=1). Structural modeling of IDH2 suggested that codon 140 mutations disrupt the enzyme's ability to bind its substrate isocitrate. Accordingly, recombinant IDH2 R140Q/W were unable to carry out the decarboxylation of isocitrate to α-ketoglutarate (α-KG), but instead gained the neomorphic activity to reduce α-KG to R(−)-2-hydroxyglutarete (2-HG). Analysis of primary leukemic blasts confirmed high levels of 2-HG in AMLs with IDH1/IDH2 mutations. Interestingly, 3/5 AMLs with IDH2 mutations had FLT3 activating mutations, raising the possibility that these mutations cooperate in leukemogenesis.

The t(8;21)-encoded AML1-ETO chimeric product is believed to be causally involved in up to 15% of acute myelogenous leukemias through an as yet unknown mechanism. To directly investigate the role of AML1-ETO in leukemogenesis, we used gene targeting to create anAML1-ETO “knock-in” allele that mimics the t(8;21). Unexpectedly, embryos heterozygous for AML1-ETO(AML1-ETO/+) died around E13.5 from a complete absence of normal fetal liver–derived definitive hematopoiesis and lethal hemorrhages. This phenotype was similar to that seen following homozygous disruption of either AML1 orCBFβ. However, in contrast to AML1- or CBFβ-deficient embryos, fetal livers from AML1-ETO/+ embryos contained dysplastic multilineage hematopoietic progenitors that had an abnormally high self-renewal capacity in vitro. To further document the role of AML1-ETO in these growth abnormalities, we used retroviral transduction to express AML1-ETO in murine adult bone marrow–derived hematopoietic progenitors. AML1-ETO–expressing cells were again found to have an increased self-renewal capacity and could be readily established into immortalized cell lines in vitro. Taken together, these studies suggest that AML1-ETO not only neutralizes the normal biologic activity of AML1 but also directly induces aberrant hematopoietic cell proliferation.

The t(8;21)-encoded AML1-ETO chimeric product is believed to be causally involved in up to 15% of acute myelogenous leukemias through an as yet unknown mechanism. To directly investigate the role of AML1-ETO in leukemogenesis, we used gene targeting to create anAML1-ETO “knock-in” allele that mimics the t(8;21). Unexpectedly, embryos heterozygous for AML1-ETO(AML1-ETO/+) died around E13.5 from a complete absence of normal fetal liver–derived definitive hematopoiesis and lethal hemorrhages. This phenotype was similar to that seen following homozygous disruption of either AML1 orCBFβ. However, in contrast to AML1- or CBFβ-deficient embryos, fetal livers from AML1-ETO/+ embryos contained dysplastic multilineage hematopoietic progenitors that had an abnormally high self-renewal capacity in vitro. To further document the role of AML1-ETO in these growth abnormalities, we used retroviral transduction to express AML1-ETO in murine adult bone marrow–derived hematopoietic progenitors. AML1-ETO–expressing cells were again found to have an increased self-renewal capacity and could be readily established into immortalized cell lines in vitro. Taken together, these studies suggest that AML1-ETO not only neutralizes the normal biologic activity of AML1 but also directly induces aberrant hematopoietic cell proliferation.

Expression of BCR-ABL1 is the hallmark of chronic myelogenous leukemia (CML) and a subset of de novo acute lymphoblastic leukemia (ALL), but the factors determining disease lineage, and progression of CML to myeloid or lymphoid blast crisis, are incompletely understood. We recently reported deletion of IKZF1 (encoding the lymphoid transcription factor Ikaros) in 85% of de novo pediatric and adult BCR-ABL1 ALL, and in lymphoid blast crisis in a small cohort of CML cases (Nature2008;453:110), suggesting that IKZF1 deletion is important in the pathogenesis of BCR-ABL1 lymphoid leukemia. To identify genetic determinants of disease stage and blast crisis lineage in CML, we have now performed high-resolution, genome wide analysis of DNA copy number abnormalities (CNA) and loss-of heterozygosity (LOH) and candidate gene resequencing in a cohort of 90 CML patients that included 64 samples obtained at chronic phase (CP), 15 samples at accelerated phase (AP), 9 lymphoid blast crisis (LBC) and 22 myeloid blast crisis (MBC) samples. Importantly, 25 patients had sequential samples (CP and/or AP, as well as blast crisis samples) enabling analysis of lesions acquired at progression to blast crisis. All blast crisis samples were flow sorted to at least 90% purity prior to DNA extraction. Germline samples for 28 cases obtained at remission or by flow sorting of blast crisis samples were also examined. Affymetrix SNP 6.0 arrays, interrogating over 1.87 million genomic loci, were used for 85 samples, and 500K arrays for the remainder. Identification of tumor-specific (somatic) copy number analysis was performed by directly comparing CML samples to matched germline samples were available, or by filtering results against databases of inherited copy number variants for samples lacking germline material. Genomic resequencing of IKZF1, PAX5 and TP53 was performed for all AP, LBC and MBC samples. There were few CNAs in CP-CML (mean 0.27 deletions and 0.07 gains per case), with no recurring lesions identified apart from deletions or gains at the chromosomal breakpoints of BCR and ABL1 (3 cases each). Notably, the size of these translocation associated deletions was highly variable, ranging from 6kb (one ABL1 deletion) and 15 kb (one BCR deletion) to deletions extending to the telomeres of chromosomes 9 and 22. No significant increase in lesion frequency was identified in AP cases (0.14 deletions and 0.9 gains per case), however the number and cumulative extent of genomic aberrations was significantly higher in both lymphoid and myeloid blast crisis samples. LBC cases had a mean of 8.1 deletions/case (P<0.0001v CP) and 2.8 gains/case (P=0.0024), where as MBC had fewer alterations with only an average of 2.8 deletions/ case (P=0.028 v CP) and 2.2 gains/case (P=0.0018). Similarly, the cumulative extent of DNA altered by CNAs was higher in both LBC (200 Mb/case) and MBC (257 Mb/case) than CP-CML (4.1 Mb/case). There were striking differences in the type of CNAs in MBC and LBC samples. Seven of 9 LBC cases had focal CNAs targeting genes regulating normal B-lymphoid development, including IKZF1 (6 cases, 2 homozygous), PAX5 (4 cases), and EBF1 (1 case with focal homozygous deletion restricted to the EBF1 locus). Thus, of these 7 cases, two had a single CNA in this pathway, three had two lesions, and two cases had three lesions. In contrast, only 4 of 22 MBC cases had lesions in this pathway, most commonly from whole or sub chromosomal deletions involving chromosomes 7 and 9. Deletion of the CDKN2A/B locus (encoding the tumor suppressors and cell cycle regulators INK4A, ARF and INK4B) was seen in 6 (67%) LBC samples, but only 2 (9%) MBC cases, and never in CP or AP CML. Other lesions commonly seen in de novo BCR-ABL1 ALL were also observed in LBC samples, including deletions of MEF2C, C20orf94, and the HBS1L gene immediately upstream of the oncogene MYB. Apart from acquisition of new or more complex abnormalities involving BCR and ABL1, the only recurring mutation observed in MBC was deletion (4 cases) or splice-site point mutations (2 cases) of TP53. These data demonstrate a lack of genomic instability with few genetic alterations in CP or AP CML. Lymphoid blast crisis samples have similar genetic alterations to those seen in de novo BCR-ABL1 ALL, whereas myeloid blast crisis displays completely distinct patterns of mutation, most commonly targeting P53. These results indicate that genomic abnormalities are important determinants of lineage and disease progression in BCR-ABL1 leukemia.