University Pediatric Cancer Genome Project Research Articles

The landscape of coding RNA editing events in pediatric cancer

BackgroundRNA editing leads to post-transcriptional variation in protein sequences and has important biological implications. We sought to elucidate the landscape of RNA editing events across pediatric cancers.MethodsUsing RNA-Seq data mapped by a pipeline designed to minimize mapping ambiguity, we investigated RNA editing in 711 pediatric cancers from the St. Jude/Washington University Pediatric Cancer Genome Project focusing on coding variants which can potentially increase protein sequence diversity. We combined de novo detection using paired tumor DNA-RNA data with analysis of known RNA editing sites.ResultsWe identified 722 unique RNA editing sites in coding regions across pediatric cancers, 70% of which were nonsynonymous recoding variants. Nearly all editing sites represented the canonical A-to-I (n = 706) or C-to-U sites (n = 14). RNA editing was enriched in brain tumors compared to other cancers, including editing of glutamate receptors and ion channels involved in neurotransmitter signaling. RNA editing profiles of each pediatric cancer subtype resembled those of the corresponding normal tissue profiled by the Genotype-Tissue Expression (GTEx) project.ConclusionsIn this first comprehensive analysis of RNA editing events in pediatric cancer, we found that the RNA editing profile of each cancer subtype is similar to its normal tissue of origin. Tumor-specific RNA editing events were not identified indicating that successful immunotherapeutic targeting of RNA-edited peptides in pediatric cancer should rely on increased antigen presentation on tumor cells compared to normal but not on tumor-specific RNA editing per se.

Abstract 5478: CICERO: An accurate method for detecting complex and diverse driver fusions using cancer transcriptome sequencing (RNA-seq) data

Abstract Gene fusions are important biomarkers for cancer diagnosis, subtype classification and therapeutic decision-making. While fusion detection using RNA-seq data has become a standard practice, existing computational methods primarily focus on identifying canonical exon-to-exon fusions. However, more complex events such as multi-partner fusions, truncations, enhancer hijacking and internal tandem duplications (ITD) can also lead to abnormal function or aberrant transcription of cancer driver genes. To aid discovery of complex and diverse driver fusions, we developed CICERO (CICERO Is Clipping Extended for RNA Optimization), a local assembly-based algorithm that integrates RNA-seq reads bearing aberrant mapping signatures with extensive annotation for ranking candidate fusions. Our benchmark data set, designed to support the main application of RNA-seq fusion analysis, consists of 184 driver fusions from 170 pediatric leukemia, solid tumor and brain tumor detected by paired tumor-normal WGS and orthogonally validated by capture sequencing, RT-PCR and/or FISH. CICERO detected 95% of these fusions with an average ranking of 1.9, whereas ChimeraScan, deFuse, FusionCatcher and STAR-Fusion detected only 63%, 66%, 77% and 63% with an average ranking of 37.0, 9.0, 18.1 and 4.4, respectively. Notably, events such as ITD and rearrangements involving the highly repetitive IGH locus were detected almost exclusively by CICERO. Our re-analysis of 167 RNA-seq data from the TCGA Glioblastoma Multiforme (GBM) cohort unveiled 158 fusions of cancer genes that were not reported previously. These include kinase fusions (KLHL7-BRAF), ITD of EGFR kinase domain and a 13% prevalence of EGFR C-terminal truncation compared to the 6% reported by the TCGA Network. CICERO has greatly improved our ability to discover non-canonical fusions which are overlooked by existing fusion detection methods, and has been used to analyze >2,000 RNA-seq samples generated by the two largest pediatric cancer genomics initiatives: the St. Jude/Washington University Pediatric Cancer Genome Project (PCGP) and the Therapeutically Applicable Research to Generate Effective Treatments (TARGET) project. We anticipate that CICERO will also improve fusion analysis for adult cancer RNA-seq data, as demonstrated through our re-analysis of TCGA-GBM and our recent discovery of MAP3K8 C-terminal truncation fusion in 2% of TCGA melanoma samples. CICERO is accessible via standard (https://github.com/stjude/Cicero) or cloud-based (https://platform.stjude.cloud/tools/rapid_rna-seq) implementation. To further improve accuracy, fusions predicted by CICERO can be curated by FusionEditor (https://proteinpaint.stjude.org/FusionEditor/), an interactive viewer allowing inspection of protein domains involved in the fusion and the gene expression status of fusion-positive samples. Citation Format: Liqing Tian, Yongjin Li, Michael N. Edmonson, Xin Zhou, Scott Newman, Clay McLeod, Yu Liu, Bo Tang, Michael C. Rusch, John Easton, Jing Ma, Austyn Trull, J. Robert Michael, Andrew Thrasher, Charles Mullighan, Suzanne J. Baker, James R. Downing, David W. Ellison, Jinghui Zhang. CICERO: An accurate method for detecting complex and diverse driver fusions using cancer transcriptome sequencing (RNA-seq) data [abstract]. In: Proceedings of the Annual Meeting of the American Association for Cancer Research 2020; 2020 Apr 27-28 and Jun 22-24. Philadelphia (PA): AACR; Cancer Res 2020;80(16 Suppl):Abstract nr 5478.

Abstract 3671: Visualize 10,000 whole-genomes from pediatric cancer patients on St. Jude Cloud

Abstract Whole-genome sequencing (WGS) is invaluable for investigating genetic abnormalities contributing to the initiation, progression and long-term clinical outcome of pediatric cancer. St. Jude Cloud (https://www.stjude.cloud/) hosts 10,000 (10K) harmonized WGS samples generated from: 1) St. Jude/Washington University Pediatric Cancer Genome Project, 2) the Genomes for Kids Clinical Trial, 3) the St. Jude Lifetime Cohort Study, and 4) the Childhood Cancer Survivor Study. To enable on-the-cloud discovery and eliminate the need for data download, we developed GenomePaint, an interactive genomics browser, to explore the somatic and germline variants of the 10K genomes with rich annotation. Germline variants in cancer predisposition genes were annotated for pathogenicity. Using GenomePaint, users can compare pathogenic variants from a locus of interest across multiple cancers or test for association of a germline variant with a specific cancer type on the fly. By matching germline variants to somatic mutation hotspots from www.cancerhotspots.org, we annotated potential germline mosaic mutations including IDH1 R132H, FBXW7 R465C, and KRAS A146T. For noncoding variants, we investigated overlap with ATAC and DNase peaks in 50 cancer cell lines along with transcription factor motif change predictions. These features will enable exploration of the functional impact of genetic variations with potential clinical status such as genetic risk for a specific cancer type, genetic association with age of onset, or development of subsequent malignancies for pediatric cancer survivors. GenomePaint also provides an integrated view of somatic SNV/indel, copy number variation, loss-of-heterozygosity, structural variation, and gene fusion. These are shown together with tumor gene expression at the single tumor level. GenomePaint also presents allele-specific expression (ASE) and outlier expression as an indicator for assessing dysfunction of regulatory regions caused by genomic variants. Cloud-based on-the-fly ASE analysis is also available for user’s samples with paired DNA and RNA sequencing results. Such gene expression integration will drive novel insights about the functional aspects of somatic coding and noncoding mutations in pediatric cancer. The innovative visualization of whole-genome sequencing data generated from 10K pediatric cancer patients on the St. Jude Cloud enables genomic discovery by scientists and clinicians through exploration of this unprecedented resource. Citation Format: Xin Zhou, Clay Mcleod, Scott Newman, Zhaoming Wang, Michael Rusch, Kirby Birch, Michael Macias, Jobin Sunny, Gang Wu, Jian Wang, Edgar Sioson, Shaohua Lei, Robert J. Michael, Aman Patel, Michael N. Edmonson, Stephen V. Rice, Andrew Frantz, Ed Suh, Keith Perry, Carmen Wilson, Leslie L. Robinson, Yutaka Yasui, Kim E. Nichols, Gregory T. Armstrong, James R. Downing, Jinghui Zhang. Visualize 10,000 whole-genomes from pediatric cancer patients on St. Jude Cloud [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2019; 2019 Mar 29-Apr 3; Atlanta, GA. Philadelphia (PA): AACR; Cancer Res 2019;79(13 Suppl):Abstract nr 3671.

Abstract 1287: Exploring somatic DNA structural alteration and aberrant genomic interactions in cancer through GenomePaint

Abstract Somatic structural variations (SV) play an important role in tumorigenesis as they may cause oncogenic gene fusions or transcriptional activation of oncogenes by introducing aberrant promoter-enhancer interactions. Evaluating the oncogenic implications of SVs is challenging, especially for SVs in noncoding regions, which requires integrating data from 1) whole-genome sequencing (WGS), 2) RNA-seq from the tumor sample and series patient samples with same tumor type to compare with, 3) epigenetic profiling (ChIP-seq), and 4) genome-wide chromosome conformation capture studies such as Hi-C. To facilitate the discovery of oncogenic SVs, we developed GenomePaint, an interactive browser to integrate and visualize somatic SVs and copy number alterations (CNAs) analyzed by WGS, and gene expression data analyzed by RNA-seq in >2,000 pediatric cancers, as well as ChIP-seq and Hi-C data from pediatric tumors and cell lines. The genomic profiling of patient samples was generated by the St Jude/Washington University Pediatric Cancer Genome Project and the NCI-TARGET project, and includes all major subtypes of pediatric leukemia, solid tumors and brain tumors. A global view of the entire patient cohort using GenomePaint revealed SV hotspots (e.g. within first intron of TP53 in osterosarcoma) and recurrent CNAs (e.g. TAL1 deletion in T-ALL) in each cancer type. Each SV was integrated with results on gene expression, ChIP-seq and Hi-C, on the rearranged chromosome. Our discovery of the mechanism underlying aberrant MYC overexpression, in a subset of neuroblastomas (NBL) lacking MYCN amplification, proves the power of this tool in facilitating the discovery of oncogenic SV drivers. The sample with the highest MYC expression in this cohort harbors an SV between chr 8 and 4, with breakpoint on chr 8 located 50 Kb downstream of MYC. Similar SVs were detected in multiple NBL cell lines by WGS, which also express high levels of MYC. By analyzing Hi-C data from these cell lines, a new topologically associating domain (TAD) extending beyond the SV breakpoint was observed along the chimeric chromosome. This new TAD shows an aberrant interaction of a super-enhancer with a broad H3K27ac peak associated with HAND2/FBXO8 on chr 4 and the MYC promoter, demonstrating the rewired regulatory architecture introduced by SV as the driver for MYC dysregulation. In conclusion, our new GenomePaint interactive browser facilitates the analysis of SVs, CNAs and gene expression in concert studies by WGS and Hi-C in pediatric tumors. Coding mutations are rarely identified in many types of childhood tumors, even in the presence of pronounced chromosomal SVs, indicating new approaches are needed to unveil oncogenic mechanisms. GenomePaint provides an integrative analysis tool to evaluate the pathogenic basis underlying the often complex SVs in noncoding genome and function as oncogenic drivers in a large fraction of pediatric malignancies. Citation Format: Yu Liu, Xin Zhou, Jian Wang, Ying Shao, John Easton, Mark W. Zimmerman, Brian J. Abraham, A. Thomas Look, Jinghui Zhang. Exploring somatic DNA structural alteration and aberrant genomic interactions in cancer through GenomePaint [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 1287.

Abstract 2628: Molecular diagnosis for pediatric cancer through integrative analysis of whole-genome, whole-exome and transcriptome sequencing

Abstract Next-generation sequencing (NGS) of the whole genome, whole exome, and transcriptome has enabled characterization of genetic landscapes of multiple cancers. By analyzing over 2,000 pediatric cancer patients, we have developed a comprehensive database for recurrent somatic alterations and pathogenic germline mutations as part of the St. Jude/Washington University Pediatric Cancer Genome Project. However, there is no systematic evaluation on whether NGS is able to identify germline and somatic lesions reported by existing molecular diagnostic assays and what combination of NGS platforms is best suited for clinical sequencing. Here we report the first comprehensive study that employs whole-genome sequencing at 30-45X coverage, whole-exome sequencing at 100X coverage and transcriptome sequencing using matched tumor/normal samples from cancer patients. A pilot study was carried out to perform NGS analysis on 78 children of leukemia, solid tumor or brain tumor with a total of 112 diagnostic or prognostic biomarkers previously characterized by multiple molecular diagnostic assays. We implemented an analysis pipeline that integrates the genetic lesions detected by all three NGS platforms to characterize somatic and germline single nucleotide variations (SNVs), short insertions and deletions (indels), structural variations including fusions, karyotypes, copy number alterations, loss of heterozygosity, tumor purity and tumor-in-normal contamination. The turn-around time for data analysis is 2 weeks with an overall sensitivity of 99% on detecting known biomarkers. Extensive validation of >3,000 somatic sequence mutations or structural variations from 38 cases shows that the specificity for somatic SNV, indel and structural variation is at 98%, 95% and 84% across the genome. We demonstrate that in addition to providing cross-validation, multi-platform NGS is required for detecting all genetic lesions of pathological significance including complex re-arrangements such as chromothripsis. In addition to known pathogenic or likely pathogenic mutations, our analysis has also unveiled novel pathogenic mutations (e.g. a germline deletion in TP53 in one patient with medulloblastoma) and identified multiple variants of unknown significance that may be worth further exploration (e.g. an in-frame deletion of exons 3-9 of DNMT3A in one neuroblastoma). Our study demonstrates that NGS is able to detect a wide range of genetic lesions currently characterized by multiple molecular diagnostic assays, providing critical insight into the design of clinical sequencing for ongoing studies. Citation Format: Jinghui Zhang, Michael Rusch, Joy Nakitandwe, Zhaojie Zhang, Michael N. Edmonson, Matthew Parker, Xiaotu Ma, Jared Becksfort, Andrew Thrasher, Jiali Gu, Yongjin Li, Erin Hedlund, Aman Patel, John Easton, Donald Yergeau, Bhavin Vadodaria, Xiang Chen, Tanja A. Gruber, Rose McGee, David Ellison, Sheila Shurtleff, James R. Downing. Molecular diagnosis for pediatric cancer through integrative analysis of whole-genome, whole-exome and transcriptome sequencing. [abstract]. In: Proceedings of the 107th Annual Meeting of the American Association for Cancer Research; 2016 Apr 16-20; New Orleans, LA. Philadelphia (PA): AACR; Cancer Res 2016;76(14 Suppl):Abstract nr 2628.

Abstract 2033: Germline mutations in ETV6 confer risk of thrombocytopenia and acute lymphocytic leukemia

Abstract Germline mutations of transcription factors (e.g. PAX5, CEBPA, GATA2, RUNX1) have been associated with an inherited susceptibility to acute leukemia. Here we report 2 unrelated kindreds harboring germline mutations in ETV6, the gene encoding the transcription factor ETS variants 6. These kindreds were primarily characterized by thrombocytopenia and acute lymphoblastic leukemia (ALL). The first kindred, identified at MSKCC and HMC, includes 9 individuals with thrombocytopenia, and 3 individuals with pre-B ALL. Sequencing a subset of common and somatically altered leukemia genes in this family identified a rare heterozygous non-synonymous missense variation (T>C) in 6 family members with thrombocytopenia and 2 with ALL. Notably, this variant did not segregate in 9 individuals in the kindred without these phenotypes. The amino acid alteration is predicted to lead to an L349P substitution within the DNA binding domain of ETV6 (L349P, NPP_001978). In silico analyses using SIFT and Polyphen assigned pathogenic variant classifications to the L349P variant. The proband of the second family presented to SJCRH with pre-B ALL, a history of thrombocytopenia, and demonstrated hypersensitivity to methotrexate (grade 3 bone marrow suppression). His leukemia transformated to myelodsyplastic syndrome and acute myeloid leukemia, which was treated by unrelated donor transplantation. The proband's mother, maternal aunt and maternal grandfather had thrombocytopenia, and the maternal great grandfather was diagnosed with chronic myeloid leukemia. The proband and his mother were found to carry a heterozygous 5 bp deletion in ETV6, N385fs (c.1153-5_1153_1delAACAG), which is predicted to produce a truncated protein of 389 amino acids with the last 4 residues different from the canonical ETV6 sequence. This variant has not been described in public sequencing data repositories. Finally, a review of 1,120 cases from the SJCRH/Washington University Pediatric Cancer Genome Project, including 588 cases of leukemia (among which 472 were ALL), revealed 3 rare germline ETV6 variants: L442P (n = 1), R181H (n = 1), and V37M (n = 1), with 2 of these variants L442P and R181H predicted to be pathogenic based on prediction and/or genomic location. Notably, these 2 variants each occurred in patients with ALL. Collectively, these findings suggest that germline ETV6 mutations are associated with a novel syndrome of thrombocytopenia with susceptibility to leukemia. Furthermore, these mutations may account for approximately 0.4% (2/472) of pediatric ALL cases. Further phenotypic characterization of affected patients and functional assessment of ETV6 germline variants, in progress, will be required to reveal the clinical effects of these mutations, their incidence in patients with ALL and their role in leukemogenesis. Citation Format: Vijai Joseph, Michael F. Walsh, Sabine Topka, Gang Wu, Rose B. McGee, Emily Quinn, Hiroto Inaba, Christine Hartford, Ching-Hon Pui, Alberto S. Pappo, Michael Edmonson, Lauren Jacobs, Villano Danylo, Kasmintan Schrader, Pragna Gaddam, Zsofia Stadler, Michael Zhang, Polina Stepensky, Peter Steinherz, James Bussel, M Harit, Michael Weintraub, Akiko Shimamura, Jinghui Zhang, James R. Downing, Kenneth Offit, Kim E. Nichols. Germline mutations in ETV6 confer risk of thrombocytopenia and acute lymphocytic leukemia. [abstract]. In: Proceedings of the 106th Annual Meeting of the American Association for Cancer Research; 2015 Apr 18-22; Philadelphia, PA. Philadelphia (PA): AACR; Cancer Res 2015;75(15 Suppl):Abstract nr 2033. doi:10.1158/1538-7445.AM2015-2033

Decoding the Cancer Genome: Insights from Bioinformatic Studies

The characterization of the landscape of genetic lesions that underlie cancer has been significantly advanced with the recent application of next-generation sequencing (NGS) technology. To define the genomic landscapes of 21 different pediatric cancer subtypes of brain tumors, solid tumors and leukemias, we analyzed >1,000 pediatric cancers and matched control tissue by whole-genome, whole-exome or transcriptome sequencing as part of the St Jude Children’s Research Hospital – Washington University Pediatric Cancer Genome Project (PCGP). Novel bioinformatics methods for integrative analysis of single-nucleotide variation (SNV), small insertion/deletion (indel), copy number alteration (CNA) and structural variation (SV) have been developed to ensure high sensitivity and accuracy, which is critical for the discovery of highly recurrent somatic lesions that have the potential for developing new cancer therapy. For example, our fusion transcript detection method CICERO has played a key role in identifying recurrent kinase fusions in 90% of Philadelphia chromosome-like acute lymphoblastic leukemia (Ph-like ALL), a subgroup characterized by a gene expression profile similar to BCR-ABL1-positive ALL. To evaluate whether NGS is able to identify germline and somatic lesions reported by molecular diagnostic assay, we carried out a pilot clinical sequencing study that employed whole-genome, whole-exome and transcriptome sequencing of matched tumor/normal samples from 78 pediatric cancer patients. We implemented an analysis pipeline that integrates the genetic lesions from all three NGS platforms as well as a data portal that supports classification of somatic and germline lesions. We present a comparison of the sensitivity and accuracy of single-platform analysis with that of integrative multi-platform analysis in identifying somatic and germline SNVs/indels, translocation, gene fusion, CNAs, karyotype, and loss of heterozygosity. Our experience provides informative insight for the design of clinical sequencing of pediatric cancer. Disclosures No relevant conflicts of interest to declare.

Incidence of Germline Mutations in Cancer-Predisposition Genes in Children with Hematologic Malignancies: a Report from the Pediatric Cancer Genome Project

▪Pathologic germ line mutations that predispose patients to cancer are estimated to occur in 4-30% of all pediatric oncology cases. In addition to leukemia specific familial predisposition syndromes, children with rare constitutional syndromes, heterogeneous dysmorphic syndromes, and multiple-cancer hereditary predisposition syndromes are all at an increased risk for hematologic malignancies. However, to date no genome-wide analysis has been done to define the range of germ line mutations that occur in pediatric patients with hematological malignancies.To determine the frequency of pediatric cancer patients that have germ line variants of pathological significance in genes that predisposed to cancer, we analyzed the germ line and tumor DNA from 1120 pediatric cancer patients that were enrolled in the St. Jude – Washington University Pediatric Cancer Genome Project (PCGP). Samples were analyzed by whole-genome sequencing (n = 595), whole-exome sequencing (n = 456), or both (n = 69). Single nucleotide variants (SNVs), insertions/deletions (indels), structural variations (SV) and copy number alterations (CNAs) were detected using our analytical pipeline and all single nucleotide polymorphisms (SNPs) previously identified in non-cancer populations were filtered out. Our analysis then focused on the 23 cancer predisposition genes recently recommended for germ line analysis by the American College of Genetics and Genomics, along with an additional 8 genes that have been previously shown to predispose to pediatric cancer at a high penetrance. All variants in these 31 genes were classified as pathologic, likely pathologic, uncertain significance, likely benign, and benign based on literature review and in-silico predictions on the effect of novel mutations. An expanded analysis including a total of 565 genes known to play a role in oncogenesis was also evaluated.Pathologic or likely pathologic germ line variants in one of the 31 genes were detected in 8% (90/1120) of patients, including: 16% (46/287) of patients with solid tumors, 8.6% (21/245) with brain tumors, and 3.9% (23/588) with leukemia. Expanding this analysis to 565 cancer gene resulted in only a slight increase, with a pathologic or likely pathologic variant being detected in 8.6% (97/1120) of patients. The most frequently effected genes included TP53 (n=48), APC (n=7) and BRCA2(n=6). Importantly, in >50% of these patients, analysis of their tumor DNA revealed the absence of a wild type allele for the cancer predisposition gene that was altered in the germ line.The 588 pediatric patients with leukemia included 116 acute myeloid leukemias (AMLs: FAB M7 n=20; Core Binding Factor leukemias n=86; MLL-R n=10) and 472 acute lymphoblastic leukemias (ALLs: E2A-PBX1 n=53; ERG-R n=39; TEL-AML1 n=53; Hyperdiploid n=69; Hypodiploid n=47; BCR-ABL1 n=40; T-ALL n=32; MLL-R n=40; BCR-ABL-like n=31; and Other n=68). Across this cohort, 3.9% (23/588) of leukemia patients harbored a pathologic germ line mutations in one of the 31 cancer pre-disposing genes. This number increased to 4.6% (27/588; 28 mutations) when the expanded gene list was evaluated. TP53 (n=10) was the most frequently altered germ line gene in pediatric leukemia patients and was found predominantly in low-hypodiploid ALL, as previously reported. Germ line pathologic variants were also identified in KRAS, RUNX1, APC, BRCA2, and RET (2 cases each), and NRAS, SH2B3, BRCA1, MUTYH, PTCH1, SDHA,VHL, and NF2 (1 case each). Although germ line mutations in RUNX1 and SH2B3are typically associated with myeloid neoplasms, we identified these lesions in 3 cases of B lineage ALL suggesting an association with a wider spectrum of leukemia.In conclusion, a small but significant proportion of pediatric patients with leukemia carry a germ line variant of pathologic significance in a cancer predisposition gene. These results suggest that these germ line lesions likely play a direct role in the pathogenesis of the patient’s presenting leukemia. Moreover, our results suggest that these patients would benefit from future clinical surveillance for the development of a second cancer. Lastly, these data demonstrate the power of comprehensive next generation DNA/RNA sequencing for the identification of pediatric patients who carry a germ line pathologic variant in a cancer predisposition gene. DisclosuresNo relevant conflicts of interest to declare.

Abstract IA11: The St. Jude Children's Research Hospital - Washington University Pediatric Cancer Genome Project: Lessons learned

Abstract In January of 2010, St. Jude Children's Research Hospital and The Genome Institute at the Washington University announced the launch of the Pediatric Cancer Genome Project (PCGP), a three year, $65 million privately funded initiative. The stated goal of this effort was to obtain 30-fold haploid coverage of the whole genome of 600 pediatric tumors and matched non-tumor germ line DNA samples (1200 total genomes), and to analyze these data in order to define the landscape of somatic mutations that underlie the major subtypes of pediatric leukemia, brain tumors, and solid malignancies. The initial phase of this project was completed in February of this year with over 700 pediatric cancers analyzed (1400 genomes). This effort has produced fundamental insights into each of the pediatric cancer subtypes sequenced. Not only has the data defined the spectrum of mutations that underlying specific pediatric cancers, but for some of the diseases analyzed, we have identified new prognostic markers and/or therapeutic targets. The effort has led to the generation of an incredibly rich data source that should serve the scientific community for years to come. Based on the success of PCGP, we have begun a $30 million second phase of the project (PCGP Phase II) that will be completed in a two year timeframe. In this second phase, we will focus not only on discovery science, but also on the establishing the technologies, analytical tools, and medical interface needed to move genome-wide sequence analyses into the diagnostic work up of pediatric cancer patients. In this talk, I will present an overview of this project and highlight some of the important results that have emerged. Citation Format: James R. Downing. The St. Jude Children's Research Hospital - Washington University Pediatric Cancer Genome Project: Lessons learned. [abstract]. In: Proceedings of the AACR Special Conference on Pediatric Cancer at the Crossroads: Translating Discovery into Improved Outcomes; Nov 3-6, 2013; San Diego, CA. Philadelphia (PA): AACR; Cancer Res 2013;74(20 Suppl):Abstract nr IA11.

Abstract A15: The landscape of somatic mutations in the core-binding factor acute myeloid leukemias

Abstract Acute myeloid leukemia (AML) results from mutations that promote aberrant self-renewal, proliferation and differentiation. The specific combination of mutations not only influences the phenotype of the leukemia but also its response to therapy. The core-binding factor (CBF) leukemias represent one of the most common forms of AML and are defined by genetic lesions that target the RUNX1(AML1)/CBFβ transcription factor complex. These include t(8;21)[RUNX1-ETO], inv(16) and t(16;16)[CBFβ-MYH11], point mutations in RUNX1, and several other rarer translocations that target RUNX1. The AML1/CBFβ transcription factor complex functions as a master transcriptional regulator that is essential for the generation of definitive hematopoietic stem cells during development, and plays an important role in the differentiation of a number of hematopoietic lineages including T cells and megakaryocytes. Using a conditional Runx1-Eto murine knock-in model, we previously demonstrated that the expression of Runx1-Eto was insufficient to induce leukemia. To define the landscape of mutation that cooperate with RUNX1-ETO in leukemogenesis, the St. Jude Children's Research Hospital - Washington University Pediatric Cancer Genome Project performed whole genome DNA Sequencing (WGS) on a cohort of 17 pediatric CBF leukemia [7 t(8;21) and 10 inv(16) cases. Somatic mutations were identified using the PCGP analytical pipeline and all identified mutations were experimentally validated. The frequency of the identified mutations were then assessed by performing exome sequencing on a validation cohort consisting of 151 pediatric and adult CBF leukemias [79 t(8;21) and 72 inv(16)]. The combined analyses identified very few copy number alterations (CNAs) or structural alterations per case, and an average of only 12.5 somatic non-synonymous single nucleotide variations or insertions/deletions per case. Despite the low burden of mutations, recurrent activating mutations in the Kinase/RAS signaling pathway (NRAS, KIT, FLT3, KRAS, and PTPN11, and inactivating mutations in NF1) were collectively identified in >67% of cases An even more striking result was the identification of a high frequency of loss-of-function mutations in genes that encode key epigenetic regulatory proteins (28% including ASXL2, EZH2, UTX, SETD2, MLL2/3), and missense and loss of function mutations in genes that encode components of the cohesin complex (10% including SMC1A, SMC2, RAD21). Moreover, the frequency of both classes of mutations were significantly higher in the t(8;21) versus inv(16) containing cases (epigenetic regulatory mutations 48% vs 6.25% and cohesin mutations 20% vs 0%). No significant difference was noted in the frequency of cohesin mutations between pediatric and adult cases. The identified mutations targeted genes encoding core components of the cohesion complex required to generate the ring shaped DNA binding structure, SMC1A, SMC3, and RAD21, as well as a single case with a mutation in the SMC3 acetyl transferase ESCO2. In addition, loss of the X chromosome is a frequent somatic event in female patients with t(8;21) leukemia and SMC1A is located on Chr. X and is not subjected to X inactivation, suggesting that loss of Chr. X leads to haploinsufficiency. The majority of leukemias with cohesion mutations also contain activating mutation in the RAS/Kinase signaling pathway, suggesting that these mutations cooperate with AML1-ETO to induce leukemia. Collectively, these data provide an unprecedented view of the range of cooperating mutations that occur in the CBF leukemias and highlight fundamental differences between t(8;21) and inv(16) containing cases. Citation Format: Amanda Larson Gedman, Jinjun Cheng, Xiang Chen, Ina Radtke, Anna Andersson, Heather Mulder, Kristy Boggs, Bhavin Vadodaria, Donald Yergeau, Joy Nakitandwe, Lars Bullinger, Michael Kuehn, Hartmut Doehner, Konstanze Doehner, Sheila Shurtleff, John Easton, Richard Wilson, Jinghui Zhang, James R. Downing. The landscape of somatic mutations in the core-binding factor acute myeloid leukemias. [abstract]. In: Proceedings of the AACR Special Conference on Pediatric Cancer at the Crossroads: Translating Discovery into Improved Outcomes; Nov 3-6, 2013; San Diego, CA. Philadelphia (PA): AACR; Cancer Res 2013;74(20 Suppl):Abstract nr A15.

Highlights from Recent Cancer Literature

Highlights from Recent Cancer Literature

Advances in childhood cancers

Advances in childhood cancers

Abstract 4873: Comprehensive analysis of 160 whole-genome sequences reveals striking telomere alteration patterns in 9 pediatric cancers

Abstract Although telomere shortening has long been recognized as a feature of cancer genomes, a comprehensive characterization of telomere alteration in pediatric cancer has not yet been performed. As part of the St. Jude Children's Research Hospital-Washington University Pediatric Cancer Genome Project (PCGP), we are performing whole genome sequence analysis on 600 pediatric cancers and matched germ line samples. This database will allow us to carry out a systematic investigation of telomere alteration across a broad range of pediatric cancers. To characterize the relative level of telomere length within a tumor, we have developed an approach that utilizes the WGS data and identifies all sequencing reads containing the telomeric repeat (TTAGGG)4. This read count is then normalized for genome-wide coverage and compared to the result from the matched germline sample to define the relative level of telomere amplification or loss in the tumor. To date we have analyzed 160 tumor/normal pairs collected from 9 distinct pediatric cancers types (Early T Progenitor Acute Lymphoblastic Leukemia (ALL), Infant ALL, Hypo-diploid ALL, Medulloblastoma, Retinoblastoma, Neuroblastoma, Osteosarcoma, Rhabdomyosarcoma, and Adrenocortical Carcinoma). Our analysis revealed normal telomere length or minimal loss in Early T Progenitor Acute Lymphoblastic Leukemia (ALL), Hypo-Diploid ALL, Infant-ALL, Medulloblastoma, Retinoblastoma, and Rhabdomyosarcoma. By contrast a subset of three distinct tumor types (Neurobalstoma, Oteosarcoma, and Adrenocortical Carcinoma) were characterized by marked increases in the number of telomere reads in the tumor consistent with telomere lengthening. These latter findings were validated in neuroblastoma by quantitative polymerase chain reaction (13/14) and fluorescence in situ hybridization (7/7). The presence of long telomeres was statistically associated with an increase in the number of somatic structural variations in tumor, with this association most evident in osteosarcoma. These data suggest that some pediatric cancers are associated with marked abnormalities in telomere maintenance leading to expansion in the telomere repeats. Defining the specific mutational processes that are responsible for these telomere defects and expanding the dataset to all 600 cancers in the project should help to define the mechanisms through which these alterations contribute to the transformed phenotypes. Citation Format: {Authors}. {Abstract title} [abstract]. In: Proceedings of the 103rd Annual Meeting of the American Association for Cancer Research; 2012 Mar 31-Apr 4; Chicago, IL. Philadelphia (PA): AACR; Cancer Res 2012;72(8 Suppl):Abstract nr 4873. doi:1538-7445.AM2012-4873

Abstract SY25-01: Analysis of next-generation sequencing data for cancer genomes: challenges and pitfalls

Abstract The characterization of the landscape of genetic lesions that underlie cancer has been significantly advanced with the recent application of next-generation sequencing (NGS) technology. This methodology can be used to sequence selected subsets of genes, the whole exome, the whole genome, or the expressed transcriptome within a cancer cell. By comparing the acquired sequences from both a cancer and matched normal tissue sample from the same patients, one should be able to identify almost all somatic lesions within the cancer. As part of the St Jude Children's Research Hospital - Washington University Pediatric Cancer Genome Project (PCGP), we have undertaken the approach of performing whole genome sequencing (WGS) on 600 pediatric cancers and matched control tissue (1200 total genomes). Although the acquisition of the primary sequence is a formidable challenge, the analysis of these data is where the real work begins. Unfortunately, the majority of published NGS analysis methods were developed to identify germ line variation and therefore perform sub-optimally when applied to the task of identifying somatic mutations in cancer genomes. This is in part a result of the distinct difference in logic that must be used to accurately identify all somatic lesions within a cancer. A cancer genome typically exists within a heterogenous DNA sample that is composed of normal cells admixed with an oligoclonal tumor sample. Moreover, the range of somatic lesions seen in cancer is broader than what is seen as part of germ line genetic variation, with some cancers having exceedingly complex genomes containing focal insertions, deletions, inversions, intra-chromosomal and inter-chromosomal rearrangements and large copy number abnormalities. The accurate identification of these lesions requires not only the presence of the lesions within the cancer DNA, but also their absence from the matched germ line sample. To approach these problems, we, as well as others, have recently developed new analytical approaches to enhance our ability to identify the somatic mutations in cancer. The starting point for these analyses is ≥75 bp paired-end sequencing reads from patient matched tumor and normal DNA samples. Our goal is to identify all somatic single-nucleotide variation (SNV), small insertion/deletion (indel), copy number alteration (CNA) and structural variation (SV) that occur within the cancer DNA sample. Paired tumor-normal NGS data were analyzed together to ensure sensitivity for detecting DNA alterations in tumor and for confirming their absence in the matched normal sample. Somatic lesions initially identified by mapped NGS reads were further analyzed using more accurate algorithms to correct errors cause by suboptimal NGS mapping. The sensitivity of the methods we have developed depends on the read depth, but with WGS at 30X haploid coverage we are able to detect mono-allelic mutation present in as low as ∼25% of the analyzed cellular populations. This sensitivity can be significantly enhanced with greater read densities. Key among the methods we have developed are two new algorithms focused on identifying gross DNA alterations: CREST (Clipping REveals STructure) for SV analysis and CONSERTING (COpy Number SEgmentation by Regression Tree) for CNA analysis. CREST uses sequencing reads with partial alignments to the reference human genome (so-called soft-clipped reads) to directly map the breakpoints of somatic SVs. CONSERTING integrates read depth analysis with SV detection and adjust for sequencing artifacts, coverage bias and germ line CNVs. Together, these methods identify somatic lesions with a high validation rate (92-98% of SNV and Indels, 80% for SVs). In this talk, I will highlight the NGS analytic pipeline we have developed and the recent discoveries that have emerged through its application to pediatric cancer genomes. In addition, I will point out some of the significant challenges that remain to be tacked in order for us to identify the full landscape and functional consequences of the somatic mutations in cancer. Citation Format: {Authors}. {Abstract title} [abstract]. In: Proceedings of the 103rd Annual Meeting of the American Association for Cancer Research; 2012 Mar 31-Apr 4; Chicago, IL. Philadelphia (PA): AACR; Cancer Res 2012;72(8 Suppl):Abstract nr SY25-01. doi:1538-7445.AM2012-SY25-01

Abstract 4869: Whole genome sequence analysis of MLL rearranged infant acute lymphoblastic leukemias reveals remarkably few somatic mutations: A Report From the St Jude Children's Research Hospital - Washington University Pediatric Cancer Genome Project

Abstract Infant acute lymphoblastic leukemia (ALL) is characterized by MLL rearrangements (MLLr) and poor prognosis. To determine the complement of somatic mutations in this high risk leukemia, we performed whole genome sequencing (WGS) on 22 infants with MLL rearranged ALL. An analysis of the structure of the MLLr revealed that over half had complex rearrangements involving either three or more chromosomes, carried cryptic rearrangements, or contained at the breakpoints deletions, amplifications, insertions, or inversion of sequences. In three cases, genetic rearrangements were predicted to generate in addition to the MLL-partner gene fusion, novel in-frame fusions including KRAS-MLL; RAD51B-MLL / AFF1-RAD51B, AFF1-RAD51B-MLL; MLLT10-ATP5L-YPEL4 / ATP5L-YPEL4. An analysis of the number of non-silent mutations revealed infant ALL to have the lowest frequency of somatic mutations of any cancer sequenced to date. After removal of SVs and CNAs associated with the MLLr, a mean of only 3.5 SVs and 2.2 SNVs affecting the coding region of annotated genes or regulatory RNAs were detected per case. Despite the paucity of mutations several pathways were recurrently targeted including PI3K/RAS pathway in 45% (KRAS (n=4), NRAS (n=2), and non-recurrent mutations in NF1, PTPN11, PIK3R1, and ARHGAP32 (p200Rho/GAP)), B cell differentiation in 23% as a result of mono-allelic deletion or gains of PAX5, 14% with deletions of the CDKN2A/B, and 2 cases with focal deletions of the non-coding RNA genes DLEU1/2. WGS of two infant ALL relapse samples and comparison with their matched diagnostic samples revealed a marked increase in the number of mutations at relapse. Moreover, an analysis of the allelic ratios of mutated genes revealed clonal heterogeneity at diagnosis with relapse appearing to arise from a minor diagnostic clone. Because of the exceedingly low number of mutations detected in infant ALL, we used exome sequencing to determine the frequency of non-silent SNVs in 20 MLLr leukemias (9 ALLs, 10 AMLs and 1 AUL) in older children (7-19 yrs). This analysis revealed that non-infant pediatric MLL leukemias harbor a significantly higher number of non-silent somatic SNVs than infant ALL (mean 7.4/case in older patients vs 2.2/case in infants, p<0.001). Although the higher frequency of mutations may be a reflection of age, the low number of cooperating mutations in infants raises the possibility that the target cell of transformation differs between infants and older children, with the cells present during early development requiring fewer cooperating mutations to induce leukemia. In summary our data demonstrated an exceedingly small number of mutations in infant leukemia. The number of detected somatic mutations may represent the lower limit required to transform a normal human cell into cancer. Citation Format: {Authors}. {Abstract title} [abstract]. In: Proceedings of the 103rd Annual Meeting of the American Association for Cancer Research; 2012 Mar 31-Apr 4; Chicago, IL. Philadelphia (PA): AACR; Cancer Res 2012;72(8 Suppl):Abstract nr 4869. doi:1538-7445.AM2012-4869

Abstract SY25-04: Sequencing the genome of acute lymphoblastic leukemia

Abstract Acute lymphoblastic leukemia (ALL) is the commonest childhood malignancy, and remains the leading cause of non-traumatic death in children and young adults. Current treatment of ALL utilizes combination systemic and intrathecal therapy with mostly non-targeted chemotherapeutic agents, and further escalation of treatment intensity is limited by unacceptable short- and long term toxicity. Targeted therapy with tyrosine kinase inhibitors in BCR-ABL1 positive leukemia has been dramatically successful, and this observation has stimulated intense interest in identifying rational targets for novel therapies in additional patients with ALL. Moreover, ALL has long been known to consist of a number of subtypes characterized by distinct constellations of genetic alterations defined by karyotypic analysis, microarray profiling and candidate gene sequencing, including structural chromosomal rearrangements, DNA copy number alterations and DNA sequence mutations. While these approaches have been informative in identifying novel somatic genetic alterations in ALL, they are inherently incapable of comprehensively identifying all and somatic genetic alterations driving leukemogenesis and influencing treatment outcome. In addition, these approaches have limited ability to fully characterize the nature of intratumoral clonal heterogeneity and genetic diversity, which is now recognized to be a hallmark of ALL, and an important determinant of treatment failure and relapse. Several groups are now utilizing next-generation sequencing to provide a comprehensive portrait of the nature and frequency of inherited and somatic genetic alterations in ALL. These projects, including the St Jude Children's Research Hospital - Washington University Pediatric Cancer Genome Project, and the National Cancer Institute - Children's Oncology Group TARGET (Therapeutically Applicable Research to Generate Effective Therapies) Initiative multi-institutional collaborative efforts that are using transcriptome sequencing (RNA-seq), exome sequencing and whole genome sequencing (WGS) of large numbers of ALL samples. Results and future directions for several of these projects will be described, and include: (1) Early T-cell precursor (ETP) ALL: WGS of this aggressive leukemia has shown that ETP is characterized by frequent mutations in multiple critical cellular pathways, including hematopoetic development, cytokine receptor and Ras signaling, and epigenetic regulation of gene expression. These findings have identified several potential new approaches for therapy in ETP ALL. (2) High risk “BCR-ABL1-like” B-progenitor ALL: This form of leukemia exhibits many features similar to BCR-ABL1 positive ALL, yet lacks expression of the BCR-ABL1 chimeric fusion protein. RNA-seq and WGS have identified a range of novel rearrangements, structural variations and sequence mutations that activate cytokine receptor and tyrosine kinase signaling. Importantly, these alterations are transforming, and potentially druggable with currently available tyrosine kinase inhibitors. (3) Infant leukemia. Infant ALL is highly aggressive and commonly accompanied by rearrangement of the mixed lineage leukemia (MLL) gene. WGS of MLLr infant leukemia extends the findings of pre-WGS genomic profiling studies by showing that very few additional somatic genetic alterations are present in this form of ALL. These studies have also stimulated the development of novel, sensitive and specific algorithms for the detection of structural genetic alterations described by Dr Zhang in this session, and notably, have shown that WGS and RNA-seq may provide complementary genetic information. Also, these studies have also shown that in addition to comprehensive characterization of “discovery” cohorts of leukemia genomes, detailed analysis of mutational frequency in extended recurrence cohorts, and complementary profiling of gene expression is required to accurately define the nature and frequency of genetic alterations across the range of ALL subtypes, and to assess their role in leukemogenesis. Importantly, these studies have largely focused on somatic genetic variation affecting coding genes, and a key goal of ongoing studies is the investigation of the role of inherited and somatic non-coding variation in leukemogenesis. Citation Format: {Authors}. {Abstract title} [abstract]. In: Proceedings of the 103rd Annual Meeting of the American Association for Cancer Research; 2012 Mar 31-Apr 4; Chicago, IL. Philadelphia (PA): AACR; Cancer Res 2012;72(8 Suppl):Abstract nr SY25-04. doi:1538-7445.AM2012-SY25-04

Whole Genome Sequence Analysis of 22 MLL Rearranged Infant Acute Lymphoblastic Leukemias Reveals Remarkably Few Somatic Mutations: A Report From the St Jude Children‘s Research Hospital - Washington University Pediatric Cancer Genome Project

Abstract 69Infant (< 1 year of age) acute lymphoblastic leukemia (ALL) is a rare disease characterized by rearrangements of the Mixed Lineage Leukemia (MLL) gene at 11q23 and a poor prognosis. In an effort to determine the total complement of somatic mutations occurring in this high risk leukemia, we performed paired-end whole genome sequencing (WGS) on diagnostic leukemia blasts and matched germ line samples from 22 infants with MLL rearranged ALL using the Illumina platform. In addition, we sequenced 2 paired relapse samples. Somatic alterations, including single nucleotide variations (SNV), and structural variations (SV) including insertions, deletions, inversion, and inter- and intra-chromosomal rearrangements were detected using complementary analysis pipelines including Bambino, CREST and CONSERTING. Validation of identified somatic mutations was performed using PCR amplification of the leukemia and germ line DNA followed by Sanger or 454-based sequencing, or by array-based capture followed by Illumina-based sequencing.Analysis of the structure of MLL rearrangements at the base pair level revealed that over half had complex rearrangements that involved either three or more chromosomes, or contained at the breakpoints deletions, amplifications, insertions, or inversion of sequences. In five of the complex cases, chromosomal rearrangements were predicted to generate not only a MLL-partner gene fusion, but also novel in-frame fusions including KRAS-MLL; RAD51B-MLL / AFF1-RAD51B; MLLT10-CTNNAP3B; MLLT10-ATP5L / ATP5L-YPEL4; and CRTAM-GNL3. An analysis of the sequence surrounding the breakpoints of MLL and its partner genes suggest that the predominant mechanism of rearrangement involved non-homologous end joining.An analysis of the total number of non-silent mutations revealed infant ALL to have the lowest frequency of non-silent somatic mutations of any cancer sequenced to date. After removal of SVs and CNAs associated with the MLL rearrangements, a mean of only 2 somatic SVs and 2 SNVs affecting the coding region of annotated genes or regulatory RNAs were detected per case, with a range of non-silent mutation of between 0 and 11 per case (0–7 SV and 0–5 SNV). Despite the paucity of mutations several pathways were recurrently targeted. Mutations leading to activation of signaling through the PI3K/RAS pathway was observed in 45% of the cases with mutation of individual components including KRAS (n=4), NRAS (n=2), and non-recurrent mutations in NF1, PTPN11, PIK3R1, and the GTPase activating protein ARHGAP32 (p200Rho/GAP), which mediates cross-talk between RAS and Rho signaling. Other pathways altered include B cell differentiation, with 23% of cases containing mono-allelic deletion or gains of PAX5, 14% with deletions of the CDKN2A/B, and 2 cases with focal deletions of the non-coding RNA genes DLEU1/2.WGS of two infant ALL relapse samples and comparison with the data from their matched diagnostic samples revealed a marked increase in the number of mutations at relapse with additional SVs, SNVs, and CNAs identified. Moreover, an analysis of the allelic ratios of mutated genes revealed clonal heterogeneity at diagnosis with relapse appearing to arise from a minor diagnostic clone.Because of the exceedingly low frequency of mutations detected in infant ALL, we decided to define the frequency of non-silent SNVs in MLL rearranged leukemia occurring in older children (7–19 years of age). Exome sequencing was performed on 13 MLL leukemias (8 ALLs and 5 AMLs). This analysis revealed that non-infant pediatric MLL rearranged leukemias harbor a significantly higher number of non-silent somatic SNVs than infant ALL (mean 8/case in older patients versus 2/case in infants, p<0.001). Although the increased frequency of mutations may be a reflection of the older age, the low number of cooperating mutations in infants raises the possibility that the target cell of transformation differs between infants and older children, with the cells present during early development requiring fewer cooperating mutations to induce leukemia.In summary our analysis demonstrated an exceedingly small number of mutations required to generate infant MLL rearranged leukemia. The number of detected somatic mutations may represent the lower limit of mutations required to transform a normal human cell into cancer. Disclosures:Fioretos:Cantargia AB: Equity Ownership, Membership on an entity's Board of Directors or advisory committees; Qlucore AB: Equity Ownership, Membership on an entity's Board of Directors or advisory committees.

Discovery of Novel Recurrent Mutations in Childhood Early T-Cell Precursor Acute Lymphoblastic Leukemia by Whole Genome Sequencing - a Report From the St Jude Children's Research Hospital - Washington University Pediatric Cancer Genome Project

Abstract 68Early T-cell precursor acute lymphoblastic leukemia (ETP ALL) is characterized by an immature T-lineage immunophenotype (cCD3+, CD1a-, CD8- and CD5dim) aberrant expression of myeloid and stem cell markers, a distinct gene expression profile and very poor outcome. The underlying genetic basis of this form of leukemia is unknown.Here we report results of whole genome sequencing (WGS) of tumor and normal DNA from 12 children with ETP ALL. Genomes were sequenced to 30-fold haploid coverage using the Illumina GAIIx platform, and all putative somatic sequence and structural variants were validated. The frequency of mutations in 43 genes was assessed in a recurrence cohort of 52 ETP and 42 non-ETP T-ALL samples from patients enrolled in St Jude, Children's Oncology Group and AEIOP trials. Transcriptomic resequencing was performed for two WGS cases, and whole exome sequencing for three ETP ALL cases in the recurrence cohort.We identified 44 interchromosomal translocations (mean 4 per patient, range 0–12), 32 intrachromosomal translocations (mean 3, 0–7), 53 deletions (mean 4, 0–10) and 16 insertions (mean 1, 0–5). Three cases exhibited a pattern of complex rearrangements suggestive of a single cellular catastrophe (“chromothripsis”), two of which had mutations targeting mismatch and DNA repair (MLH3 and DCLRE1C). While no single chromosomal alteration was present in all cases, 10 of 12 ETP ALLs harbored chromosomal rearrangements, several of which involved complex multichromosomal translocations and resulted in the expression of chimeric in-frame novel fusion genes disrupting hematopoietic regulators, including ETV6-INO80D, NAP1L1-MLLT10, RUNX1-EVX1 and NUP214-SQSTM1, each occurring in a single case. An additional ETP case with the ETV6-INO80D fusion was identified in the recurrence cohort. Additionally, 51% of structural variants had breakpoints in genes, including those with roles in hematopoiesis and leukemogenesis, and genes also targeted by mutation in other cases (MLH3, SUZ12, RUNX1).We identified a high frequency of activating mutations in genes regulating cytokine receptor and Ras signalling in ETP ALL (67.2% of ETP compared to 19% of non-ETP T-ALL) including NRAS (17%), FLT3 (14%), JAK3 (9%), SH2B3 (or LNK; 9%), IL7R (8%), JAK1 (8%), KRAS (3%), and BRAF (2%). Seven cases (5 ETP, 2 non-ETP) harbored in frame insertion mutations in the transmembrane domain of IL7R, which were transforming when expressed in the murine cell lines, and resulted in enhanced colony formation when expressed in primary murine hematopoietic cells. The IL7R mutations resulted in constitutive Jak-Stat activation in these cell lines and primary leukemic cells expressing these mutations.Fifty-eight percent of ETP cases (compared to 17% of non-ETP cases) harbored mutations known or predicted to disrupt hematopoietic and lymphoid development, including ETV6 (33%), RUNX1 (16%), IKZF1 (14%), GATA3 (10%), EP300 (5%) and GATA2 (2%). GATA3 regulates early T cell development, and mutations in this gene were observed exclusively in ETP ALL. The mutations were commonly biallelic, and were clustered at R276, a residue critical for binding of GATA3 to DNA. Strikingly, mutations disrupting chromatin modifying genes were also highly enriched in ETP ALL. Genes encoding the the polycomb repressor complex 2 (EZH2, SUZ12 and EED), that mediates histone 3 lysine 27 (H3K27) trimethylation were deleted or mutated in 42% of ETP ALL compared to 12% of non-ETP T-ALL. In addition, alterations of the H3K36 trimethylase SETD2 were observed in 5 ETP cases, but not in non-ETP ALL. We also identified recurrent mutations in genes that have not previously been implicated in hematopoietic malignancies including RELN, DNM2, ECT2L, HNRNPA1 and HNRNPR.Using gene set enrichment analysis we demonstrate that the gene expression profile of ETP ALL shares features not only with normal human hematopoietic stem cells, but also with leukemic initiating cells (LIC) purified from patients with acute myeloid leukemia (AML). These results indicate that mutations that drive proliferation, impair differentiation and disrupt histone modification cooperate to induce an aggressive leukemia with an aberrant immature phenotype. The similarity of the gene expression pattern with that observed in the LIC of AML raises the possibility that myeloid-directed therapies might improve the outcome of ETP ALL. Disclosures:Evans:St. Jude Children’s research Hospital: Employment, Patents & Royalties; NIH & NCI: Research Funding; Aldagen: Membership on an entity’s Board of Directors or advisory committees.

Abstract PL02-02: The molecular pathology of acute leukemia

Abstract PL02-02: The molecular pathology of acute leukemia