ALBERT

Description: Abstract 182 Chromosomal alterations are a hallmark of acute lymphoblastic leukemia (ALL), but many cases lack a recurring cytogenetic abnormality. To identify novel alterations contributing to leukemogenesis, we previously performed genome-wide profiling of genetic alterations in pediatric ALL using single nucleotide polymorphism (SNP) microarrays. This identified a novel focal deletion involving the pseudoautosomal region (PAR1) of Xp/Yp in 15 B-progenitor ALL cases lacking sentinel chromosomal abnormalities, including six of eight cases of ALL associated with Down syndrome (DS-ALL). The deletion involved hematopoietic cytokine receptor genes, including IL3RA and CSF2RA, but due to poor array coverage, it was not possible to define the limits of deletion using SNP array data alone. To characterize this abnormality, we examined an expanded cohort of 329 B-ALL cases, including 22 B-progenitor DS-ALL cases. Strikingly, 12 (55%) DS-ALL cases harbored the PAR1 deletion. Mapping using high density CGH arrays showed the deletion to be identical in each case, and involved a 320kb region extending from intron 1 of the purinergic receptor gene P2RY8 to the promoter of CRLF2 (encoding cytokine receptor like factor 2, or thymic stromal lymphopoietin receptor). The deletion resulted in a novel fusion of the first, non-coding exon of P2RY8 to the entire coding region of CRLF2 in each case. The P2RY8-CRLF2 fusion resulted in elevated expression of CRLF2 detectable by quantitative RT-PCR, and flow cytometric analysis of leukemic cells. One DS-ALL case with elevated CRLF2 expression lacked the PAR1 deletion, but had an IGH@-CRLF2 translocation detected by fluorescence in situ hybridization (FISH). CRLF2 alteration was associated with gain of chromosome X (which was shown by FISH to result in duplication of the PAR1 deletion), deletion of 9p, and the presence of Janus kinase (JAK1 and JAK2) mutations. Ten (53%) of patients with CRLF2 alteration had JAK mutations, compared with two patients lacking CRLF2 abnormalities (P

Description: 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

Description: Previous genomic analyses identified a high frequency of genetic alterations targeting B-lymphoid development (〉40%) and cell cycle/tumor suppression (CDKN2A/B, 35%) in pediatric B-progenitor acute lymphoblastic leukemia (B-ALL). The frequency and nature of copy number alterations (CNAs) differs significantly across ALL subtype with less than one CNA/case in MLL ALL and over 8 lesions/case in BCR-ABL1 and ETV6-RUNX1 ALL; and the near obligate presence of IKZF1 (Ikaros) deletion in BCR-ABL1 ALL, but never in ETV6-RUNX1 ALL. Although important insights have been obtained from these genome-wide CNAs, it is important to realize that the SNP array platforms that have been used are limited by uneven probe spacing and poor coverage of some genes (including IKZF1 and CDKN2A/B). Moreover, additional CNAs are detected as array resolution increases (e.g. 6.5 lesions per BCR-ABL1 case using 315,000 markers, and 8.8 for 615,000 marker data). Thus, the true frequency of CNAs in ALL remains unknown. To address this question, we now report ultra-high resolution CNA analysis using 2.17 million feature oligonucleotide arrays (Roche Nimblegen) of 20 MLL, 20 BCR-ABL1, and 4 miscellaneous karyotype B-ALL cases. All had previously been examined using Affymetrix 500K SNP arrays. We identified a mean of 6.2 deletions (range 0–38) and 1.4 gains (0–13) per case. There were more lesions in BCR-ABL1 than MLL ALL cases (mean 11.2 v 1.15 deletions, P

Description: In genome-wide analyses of DNA copy number abnormalities (CNAs) and loss of heterozygosity (LOH) using single nucleotide polymorphism (SNP) microarrays, we previously reported a high frequency of genetic alterations of regulators of B lymphoid development and cell cycle in B-progenitor acute lymphoblastic leukemia (B-ALL), and a near-obligate deletion of the transcription factor Ikaros (IKZF1) in BCR-ABL1 ALL (Nature2007;446:758 and Nature2008;453:110). To assess the prognostic significance of genetic alterations and their effects on the leukemic cell transcriptome, we have now extended these studies to a cohort of 221 NCI high risk pediatric ALL cases treated on the Children’s Oncology Group (COG) P9906 trial. The majority of these cases (N=170) lacked known cytogenetic abnormalities, and none were BCR-ABL1 positive. CNA and LOH data were generated using Affymetrix 500K SNP arrays, and gene expression data was available for 207 cases. Genomic resequencing was performed for PAX5 and IKZF1. These analyses revealed mutations in genes regulating B lymphoid development in 57.5% of cases, including PAX5 CNAs (31.7%), PAX5 sequence mutations (8.1%), IKZF1 deletions (24.9%) and IKZF1 sequence mutations (2.6%). In addition, recurring CNAs were detected in a number of other genes known to play roles in transformation and drug response including CDKN2A/B (45.7%), RB1 (11.3%), BTG1 (10.4%), IL3RA (6.8%), KRAS (6.3%), NRAS (2%), NR3C1 (2%), NR3C2 (3%), and ERG (6.3%). To examine associations between genetic alterations and outcome in a genome-wide fashion, we used a semi-supervised principal components approach and identified a predictor of outcome driven by deletion/mutation of IKZF1, EBF, and BTLA (hazard ratio 3.74, P=8.1×10−5). IKZF1 lesions were most strongly associated with poor outcome (cumulative incidence of relapse 68.8% in IKZF1-mutated ALL vs. 29.1% for IKZF1 wild type cases, P=0.002). This copy number predictor also predicted outcome in an independent cohort of 258 B-progenitor ALL cases treated at St Jude Children’s Research Hospital (HR 3.76, P=0.00023). Importantly, the St Jude cohort had a much higher proportion of cases with conventional recurring cytogenetic abnormalities including hyperdiploidy and translocations than the COG P9906 cohort (60.4% versus 23%, respectively), including BCR-ABL1 (N=21; 8.1%). Notably, the association of IKZF1 mutation and poor outcome was also observed in the cases lacking BCR-ABL1. The presence of an IKZF1 mutation was also associated with a high-level of minimal residual disease following induction therapy in both cohorts. We next defined a gene expression profile of high risk (IKZF1-mutated) ALL in each cohort. Using gene set enrichment analysis (GSEA), we found the high risk gene expression signature to be positively enriched for hematopoietic stem cell genes and genes associated with minimal residual disease, and negatively enriched for genes regulating cell cycle progression and genes upregulated during B cell maturation. Moreover, GSEA also demonstrated significant similarity of the expression signatures of high risk COG and St Jude ALL, a striking finding given the marked differences in sample composition of the two groups. As IKZF1 mutation is a near obligate lesion in BCR-ABL1 lymphoid leukemia (Nature2008;453:110) and in this study is a strong independent predictor of poor outcome in BCR-ABL1 negative ALL, we next examined similarity of the gene expression signatures of high risk (IKZF1-mutated) BCR-ABL1 negative ALL and BCR-ABL1 positive ALL. Using GSEA and direct comparison of differentially expressed genes, there was highly significant similarity between the two signatures. These findings indicate that IKZF1 mutation is a key determinant of the gene expression signature of both BCR-ABL1 positive and negative ALL, and is thus likely to be central to the pathogenesis and poor outcome of both leukemia subtypes. Thus, IKZF1 deletion is a new prognostic factor in ALL, and is associated with a gene expression profile indicative of impaired maturation of the leukemic blasts, an observation compatible with the known role of IKZF1 in regulating early lymphoid development. These findings have important implications for risk stratification of ALL.

Description: Abstract 537 Acute lymphoblastic leukemia (ALL), the commonest childhood malignancy, is characterized by recurring gross and submicroscopic structural genetic alterations that contribute to leukemogenesis. Disordered epigenetic regulation is a hallmark of many tumors, and while analysis of DNA methylation of limited numbers of genes or ALL samples suggests epigenetic alterations may also be important, a large-scale integrative genome-wide analysis evaluating DNA methylation in ALL has not been performed. Here, we report an integrated epigenomic, transcriptional and genetic analysis of 167 childhood ALL cases, comprising B-progenitor ALL with hyperdiploidy (N=26), ETV6-RUNX1 (N=27), TCF3-PBX1 (N=9), BCR-ABL1 (N=19), rearrangement of MLL (MLLr) (N=20), rearrangement of CRLF2 (N=11, CRLF2r), deletion of ERG (N=11), miscellaneous or normal karyotype (N=14), and T-lineage ALL (N=30), including 4 MLLr cases and 8 cases with early T-cell precursor immunophenotype. Genome-wide profiling of structural DNA alterations was performed for all cases using Affymetrix 500K and SNP 6.0 arrays. Affymetrix U133A gene expression profiling data was available for 154 cases. Genome-wide methylation profiling was performed using the HELP microarray assay, which measures methylation at approximately 50,000 CpGs distributed among 22,722 Refseq promoters. Methylation data was compared to that of normal pro-B (CD34+CD19+sIg-), pre-B (CD34-CD19+sIg-) and mature B (CD34-CD19+sIg+) cells FACS-sorted from bone marrow of 6 healthy individuals. Unsupervised hierarchical clustering of the top 4043 most variable methylation probesets identified 9 B-ALL clusters with significant correlation to specific genetic lesions including ETV6-RUNX1, MLLr, BCR-ABL1, CRLF2r, TCF3-PBX1 and ERG deletion. T-ALLs and hyperdiploid B-ALLs also defined specific DNA methylation clusters. Supervised analysis including limma and ANOVA identified distinct DNA methylation signatures for each subtype. Notably, the strength of these signatures was subtype dependent, with more differentially methylated genes observed in ALL cases with genetic alterations targeting transcriptional regulators (e.g. ETV6-RUNX1 and MLLr) and fewer genes in cases with alterations deregulating cytokine receptor signaling (e.g. CRLF2r). Aberrant DNA methylation affected specific and distinct biological processes in the various leukemia subtypes implicating epigenetic regulation of these pathways in the pathogenesis of these different forms of ALL (e.g. TGFB and TNF in ERG deleted leukemias; telomere and centriole regulation in BCR-ABL1 ALL). Aberrantly methylated genes were also enriched for binding sites of known or suspected oncogenic transcription factors that might represent cooperative influences in establishing the phenotype of the various B-ALL subtypes. Most importantly, an integrated analysis of methylation and gene expression of these ALL subtypes demonstrated striking inversely correlated expression of the corresponding gene transcripts. The methylation signatures of each subtype exhibited only partial overlap with those of normal B cells, indicating that the signatures do not simply reflect stage of lymphoid maturation. In a separate approach, we discovered that 81 genes showed consistent aberrant methylation across all ALL subtypes, including the tumor suppressor PDZD2, HOXA5, HOXA6 and MSH2. Inverse correlation with expression was confirmed in 66% of these genes. These data suggest the existence of a common epigenetic pathway underlying the malignant transformation of lymphoid precursor cells. Integrative genetic and epigenetic analysis revealed hypermethylation of genes on trisomic chromosomes that do not show increased expression, suggesting that epigenetic silencing may control genes within amplified regions and explain why only selected genes are overexpressed. Finally, analysis of individual genes targeted by recurring copy number alterations in ALL revealed a subset of genes also targeted by abnormal methylation, with corresponding changes in gene expression (e.g. ERG, GAB1), suggesting that such genes are inactivated far more frequently than suggested by genetic analyses alone. Collectively, the data support a key role of epigenetic gene regulation in the pathogenesis of ALL, and point towards a scenario where genetic and epigenetic lesions cooperatively determine disease phenotype. Disclosures: No relevant conflicts of interest to declare.

Description: Using high-resolution SNP arrays and genomic resequencing, we recently reported deletions, translocations, and mutations involving regulators of B cell development in 40% of pediatric B-progenitor ALL (Nature2007;446:758). The most frequently involved genes were PAX5, EBF1, IKZF1 (Ikaros), IKZF3 (Aiolos) and LEF1. Ikaros is a transcription factor required for normal lymphoid development and acts as a tumor suppressor in mice. Focal deletions of IKZF1 were observed in 7 B-progenitor ALL cases, suggesting that the previously reported expression of dominant-negative, non-DNA binding Ikaros isoforms may be caused by genomic IKZF1 abnormalities. We have now extended the analysis to 283 pediatric ALL cases, including B-progenitor ALL with high hyperdiploidy (N=39), hypodiploidy (N=10), rearrangement of MLL (N=23), TCF3-PBX1 (N=17), ETV6-RUNX1 (N=48) and BCR-ABL1 (N=21), as well as cases with low hyperdiploid, normal or miscellaneous cytogenetics (N=75), and T-lineage ALL (N=50). We also examined 22 adult BCR-ABL1 positive ALL cases, 37 acute leukemia cell lines and 49 samples from 23 chronic myeloid leukemia (CML) cases at various stages of disease, including 15 with matched blast crisis samples (12 myeloid, 3 lymphoid). All samples were examined with 500,000 feature Affymetrix SNP arrays (250k Sty and Nsp). The 50k Hind and Xba arrays were also used in 252 ALL cases. Sixty-two (20.3%) ALL cases harboured IKZF1 deletions, including 36 of 43 (83.7%) BCR-ABL1 positive B-ALL cases (76.2% of 21 childhood cases, and 90.9% of 22 adult cases). The deletions were limited to IKZF1 in 25 BCR-ABL1 ALL cases, and in 19 cases deleted an internal subset of IKZF1 exons, most commonly 3–6 (Δ3–6). Remarkably, chronic phase CML samples lacked evidence of IKZF1 deletion, whereas four of 15 matched CML blast crisis samples (66% of lymphoid and 17% of myeloid) had acquired an IKZF1 deletion. The IKZF1 Δ3–6 deletion was also detected in the BCR-ABL1 B-progenitor cell lines BV173, OP1 and SUP-B15, the Δ1–6 deletion in the MYC-IGH/BCL2-IGH B-progenitor cell line 380, and Δ1–7 in the BCR-ABL1 B-progenitor cell line TOM-1. RT-PCR analysis for IKZF1 transcripts demonstrated complete concordance between the extent of IKZF1 deletion and the expression of aberrant Ikaros transcripts lacking internal exons. Importantly, on quantitative RT-PCR analysis and western blotting, expression of the dominant-negative Ik6 transcript and protein, which lacks exons 3–6, was exclusively observed in those cases with IKZF1 Δ3–6, demonstrating that the Ik6 transcript is the result of a specific genetic lesions and not alternative splicing of wild-type IKZF1. Lastly, sequence analysis of the IKZF1 Δ3–6 breakpoints indicated that the deletions arise from aberrant activity of RAG-mediated V(D)J recombination. Taken together, these data demonstrate that deletion of IKZF1, resulting in either haploinsufficiency or the expression of a dominant negative form of the transcription factor, is a central event in the pathogenesis of both pediatric and adult BCR-ABL1 B-progenitor ALL.

Description: Abstract 703 Relapse occurs across the spectrum of cytogenetic subtypes of acute lymphoblastic leukemia (ALL), and the biologic factors influencing risk of relapse are poorly understood. Previous studies have demonstrated substantial evolution in the complement of DNA copy number alterations from diagnosis to relapse, with the majority of cases acquiring new lesions at relapse that are not present at diagnosis, and also losing lesions present in the predominant clone at diagnosis. Moreover, fingerprinting and backtracking of deletions indicate a common ancestral origin of the diagnosis and relapse clones in most cases, suggesting that multiple, genetically distinct clones are present at diagnosis, and that specific genetic alterations influence risk of relapse. At present, detailed examination of DNA sequence variations in relapsed ALL has not been performed. In this study, we have performed genomic resequencing of 238 genes in leukemic samples obtained at diagnosis and relapse in 23 childhood ALL cases. The cohort comprised cases with high hyperdiploidy (N=3), TCF3-PBX1 (N=1), ETV6-RUNX1 (N=3), MLL-rearrangement (N=3), BCR-ABL1 (N=3), and cases with low hyperdiploid, pseudodiploid, normal, and miscellaneous karyotypes (N=10). All samples had over 80% blasts or were flow sorted to high purity prior to DNA extraction. Whole genome amplification of DNA was performed prior to sequencing. We selected genes targeted by recurring copy number alterations in ALL, genes in key pathways targeted by genetic alterations in ALL (e.g. lymphoid development, tumor suppression, cell cycle regulation and apoptosis), known cancer genes and tyrosine kinases. The complete coding region of each gene was sequenced in both diagnosis and relapse sample in all cases. Validation of putative variants was performed by sequencing of matched normal DNA. 248 putative protein changing variations were identified. After removal of variants also identified in matched normal DNA, 55 variants in 32 genes were identified in 20 cases (mean 2.5 variants per case, range 0-5). Eleven genes were mutated in multiple patients (ASMTL, CREBBP, ERG, FLT3, KRAS, NF1, NRAS, PAX5, PTPN11, TP53 and TUSC3). We identified tumor-acquired variations in genes previously known to be mutated in acute leukemia, including ETV6 (1 case) JAK1 (1), NRAS (5), KRAS (2), NF1 (3) PTPN11 (2), PAX5 (2), FBXW7 (1), and TP53 (2). In addition, we identified recurring mutations in genes not previously known to be mutated in cancer, including the transcriptional regulators CREBBP (N=4), NCOR1 (N=2), the tumor suppressor candidate gene TUSC3 (N=2), and the ETS family transcription factor ERG (N=2). Single mutations were also identified in transcriptional regulators (THADA, SPI1 (PU.1), TCF4, TCF7L2), the histone gene HIST1H2BG, and additional genes also targeted by copy number alterations in ALL (ARMC2, ATP10A, PLEKHG1, STIM2). The patterns of evolution of sequence variations between diagnosis and relapse were similar to those previously reported for DNA copy number alterations. Four patients had identical sequence mutations at diagnosis and relapse. Twelve cases had the some sequence variations identified at both diagnosis and relapse, but either acquired additional mutations at relapse (N=9), lost mutations present at diagnosis (N=2), or both acquired new lesions at relapse and lost variants at diagnosis (N=1). An additional three cases had variants detected at either diagnosis or relapse with no commonality between the two samples. Notably, eight (35%) cases had lesions resulting in constitutive RAS activation at diagnosis or relapse (NRAS in 3 cases, FLT3 in 2, PTPN11 in 2, NF1 in 2, and KRAS in 1), with 5 cases harboring mutations at diagnosis only (NF1 in 2 cases, NRAS in 2, and FLT3 in 1), and three cases harboring mutations at relapse only (FLT1, PTPN11 and NRAS 1 case each). In several cases, relapse-acquired mutations were identified as minor subclones at diagnosis. These data have identified novel targets of somatic mutation in ALL, and suggest that sequence variation is important determinant of risk of relapse. Identification of mutations in multiple cases suggests that several of these variants are driver mutations. These findings also indicate that resequencing of the entire coding genome of relapsed ALL will be essential to identify all lesions influencing response to therapy. Disclosures: No relevant conflicts of interest to declare.