ALBERT

Description: Abstract 3123 Although cure rates are high for pediatric Burkitt lymphoma (BL), a subset of patients relapses and succumbs to the disease. BL is characterized by translocation of the MYC gene with an immunoglobulin gene, but secondary changes including gain of 1q, gain of 13q, loss of 13q, loss of 17p, and others have been described by both conventional cytogenetic and oligo array CGH approaches. Secondary changes may contribute to the clinical heterogeneity of BL as evidenced by the fact that loss of 13q is associated with a worse prognosis in pediatric BL (Poirel et al., Leukemia 23:323, 2009). However, high resolution, genome-wide copy number analysis has not yet been reported in pediatric BL. The objective of this study was to identify copy number alterations (CNAs) in pediatric BL using a genome-wide approach, and to examine the relationship between CNAs and clinical parameters including outcome. After institutional review board approval, we identified 30 pediatric BL patients treated at Primary Children's Medical Center (n=25, Salt Lake City, UT) and Penn State Hershey Medical Center (n=5, Hershey, PA) with available formalin-fixed, paraffin-embedded (FFPE) diagnostic biopsy specimens. Age, site, and gender data were available for all specimens, and 22/24 of the Utah patients were treated according to the COG 5961 protocol with full clinical information and follow-up available. DNA was isolated from FFPE biopsies containing at least 80% tumor. In addition, germline DNA was isolated from negative staging bone marrow clot sections on Utah patients (n=25) to serve as a pooled normal reference and to provide germline copy number variation data on individual patients. Tumor and paired normal DNA was submitted for Molecular Inversion Probe (MIP) assay (330K Cancer Panel, Affymetrix, Santa Clara, CA). The Nexus Copy Number (BioDiscovery, El Segundo, CA) software package was used to analyze the MIP data with the following stringent call criteria: SNPRank segmentation, 5.0E-4 significance threshold, 1000 kb maximum contiguous probe spacing, minimum of 5 probes per segment, gains ≥ 2.7 copy number value, and loss ≤ 1.3 copy number value. Patients included 23 males and 7 females with a mean age of 8.0 years (range 2 – 18). At presentation, lactate dehydrogenase (LDH) levels ranged from 443 – 13851 U/L and uric acid levels ranged from 1.5 – 13 mg/dL. Patients were Murphy stage I (n=1), II (n=9), III (n=13), and IV (n=2). A total of three patients relapsed (one stage II patient and two stage III patients) and three died (the two stage III patients that relapsed and a different stage II patient who did not have a complete clinical response). 27 of the 30 tumor samples and 22/25 paired normals had adequate DNA for MIP analysis. The 3 tumor samples without adequate DNA were clinically similar to the others. We identified a total of 103 CNAs (defined as change seen in the same cytoband in 1 or more patients), which included 63 gains and 40 copy number losses. 23/27 cases (85%) had at least one gain or loss. We identified 21 recurrent CNAs (same cytoband affected in 2 or more patients), which included 14 gains and 7 losses. We found gains of 1q in 10/27 patients (37%), gains of 13q in 4/27 patients (15%) and losses of 17p in 3/27 patients (11%). Despite a relatively small sample size, deletion of 17p13 was significantly associated with relapse (p=0.041). To our knowledge, this is the first report of high-resolution genome-wide copy number analysis of pediatric BL. Furthermore, we show for the first time that FFPE archived materials can be used for high resolution gene copy number analysis in lymphomas. We identified CNAs in 85% of the pediatric BL cases, which include previously described changes of 1q, 13q, and 17p. Although sample size limited statistical power, deletion of the 17p13 locus was associated with relapse. We plan to extend these studies in a larger sample of patients to evaluate the potential prognostic significance of both the 17p13 locus deletion as well as additional recurrent CNAs. Disclosures: No relevant conflicts of interest to declare.

Description: Abstract 273 VPREB1 deletions occur in 1/3 of pediatric ALL (pALL) patients, but have been considered a normal consequence of λ light chain rearrangement due to VPREB1's location among λ chain variable region genes. Recently, VPREB1 deletions were reported in high risk pALL samples with BCR-ABL1-like gene expression (Den Boer 2009) and VPREB1 under-expression was found in a 38-gene signature associated with relapse (Kang 2010). We characterized VPREB1 deletions to determine if they are part of normal λ light chain rearrangement or represent a bona fide copy number alteration (CNA) that predicts worse outcome in pALL. We first used the genome-wide Molecular Inversion Probe (MIP) 330K Cancer Panel (Affymetrix) to identify VPREB1 deletions in a cohort of 52 pALL patients from the University of Utah (n=47 Precursor B-cell (Pre-B), 4 Precursor T-cell (Pre-T), 1 relapse Pre-B ALL). We found focal VPREB1 deletions in 25% (n=12/48) of Pre-B ALL and no (n=0/4) Pre-T ALL patients. We also identified a distinct and often homozygous deletion (14 kb in length) located ≂f 78 kb upstream of VPREB1 that could represent a potential enhancer region; this possible enhancer deletion was seen in 16.7% (n=8/48) of Pre-B and none of the Pre-T ALL patients. 50% of Pre-B ALL samples had at least one deletion of B-cell developmental gene (SPI1, IKZF1, BCL11A, TCF3, EBF1, PAX5, FOXP1, LEF1, VPREB1, BLNK), and of these, half (n=12/24) had deletion only of VPREB1 or its possible enhancer. We next sorted blood leukocytes from 10 healthy volunteers for quantitative PCR (qPCR) analysis on κ V-J junction (deletion=attempted κ rearrangement), λ V-J junction (deletion=attempted λ rearrangement), LOC96610 and PRAME (distal to VPREB1, among λ-light chain variable genes), VPREB1, and VPREB1 enhancer (proximal to VPREB1, among λ-light chain variable genes). As predicted for normal light chain rearrangement, we observed no deletions in monocytes; 100% of κ-sorted B-cells had κ V-J deletions without lambda deletions; and 100% of λ-sorted B-cells had both κ V-J and λ V-J deletions. We also analyzed 10 mature B-cell (Burkitt) lymphoma samples and observed VPREB1 deletions only in lambda-expressing samples with both κ and λ V-J deletions, and deletions always extended contiguously downstream to λ chain V-J region. We re-analyzed 30 Pre-B ALL samples (Utah cohort) and 11 Pre-B ALL cell lines by qPCR. Results validated MIP CNA patterns and differed from normal sorted B-cell and mature B-cell (Burkitt) results, displaying several types of rearrangements: 1) No light chain rearrangements with no VPREB1/enhancer deletion (n=5/30 [17%] clinical samples and 2/11 [18%] Pre-B cell lines), 2) Normal light chain (κ or κ/λ) rearrangement with no VPREB1/enhancer deletion (n=5/30 [17%] samples and 6/11 [55%] cell lines), 3) Normal light chain (κ or κ/λ) rearrangements with contiguous VPREB1/enhancer deletion (n=1/30 [3%] samples and no cell lines), 4) λ-light chain rearrangement with non-contiguous VPREB1/enhancer deletion (n=13/30 [43%] samples and 2/11 [18%] cell lines), and 5) No λ light chain rearrangement and focal VPREB1/enhancer deletion (n=6/30 [20%] samples and 1/11 [9%] cell lines). In the Utah Pre-B ALL cohort (n=47), VPREB1/enhancer deletions were significantly associated (Fisher's Exact Test, 2-tailed) with Hispanic ethnicity (p= 0.013), relapse (p= 0.027), death (p=0.026), and Day 14 M2 Marrow (5-25% blasts; p=0.038). Finally, we examined the publicly available St. Jude Children's Research Hospital ALL dataset (NEJM 351:533) and found VPREB1 under-expression associated with higher LC50 (resistance) for DNR (p=0.04) and VCR (p=0.04), but not PRED or ASP. In summary, we have shown VPREB1 deletions occur outside the context of normal light chain rearrangement. Unlike physiologic λ rearrangements, VPREB1 deletions do not always extend to the λ V-J junction in pALL. We also identified a focal, homozygous deletion just proximal to VPREB1 that may affect a possible enhancer region for this B-cell developmental gene. VPREB1/enhancer deletions were the only B-cell developmental gene lost in 25% of our cohort, perhaps defining a new Pre-B ALL subtype. Finally, deletion of VPREB1 (and its possible enhancer) predict worse clinical outcome and VPREB1 down-regulation correlates with in vitro drug resistance. Further studies of VPREB1's function in pediatric ALL may improve our understanding of leukemogenesis and could further refine clinical risk stratification. Disclosures: No relevant conflicts of interest to declare.

Description: Abstract 275 T-cell receptor (TCR) gene rearrangements occur in 50–80% of childhood precursor B-cell ALL and have been used for minimal residual disease (MRD) detection. However, TCR rearrangement has never been examined for its contribution to disease classification and outcome. We used Molecular Inversion Probe technology (Affymetrix) to analyze precursor B-cell ALL samples (n=48) from the University of Utah for genome-wide copy number changes, and evaluated the relationship between focal TCR deletions (indicating rearrangement) and clinical features. Copy number deletions were detected by Nexus Copy Number (BioDiscovery, Inc.) and were based on 5 consecutive probes having a value of

Description: Abstract 741 Identifying prognostic biomarkers in childhood acute lymphoblastic leukemia (ALL) is imperative for risk-stratification and intensifying therapy for children at high risk of remission induction failure or relapse. IKZF1 deletions/mutations were recently shown to correlate with poor outcome in ALL, highlighting genetic alterations as prognostic markers (NEJM 360:470, 2009). We recently described focal deletions in VPREB1 located within the lambda variable chain region which are not part of normal light chain rearrangement in precursor B-cell ALL in a local Utah cohort and the TARGET cohort of high-risk ALL patients treated on the Children's Oncology Group P9906 trial. Upon further inspection, we identified a second focal deletion in chromosome 22q11.22, nearly 80 kilobases (Kb) upstream from VPREB1 in the same lambda region. We characterized and correlated this deletion with outcome in childhood ALL using the Utah Cohort (N=60), TARGET P9906 cohort (N=221), and St. Jude Children's Research Hospital cohort (SJCRH, N=298). Microarray data was analyzed (Utah = Molecular Inversion Probe 330K [Affymetrix]; TARGET = SNP 500K & U133A [Affymetrix]; SJCRH = SNP 500K/6.0 & U133A [Affymetrix]) by Nexus Copy Number and Nexus Expression (BioDiscovery, Inc.). Each cohort contains the focal 22q11.22 loss (Hemizygous: Utah = 21%, TARGET = 16%, SJCRH = 18%; and Homozygous: Utah = 7%, TARGET = 16%, SJCRH = 10%). It spans 142 Kb, with the most common recurring region just under 10 Kb in length. The deleted segment encodes no known genes. In the SJCRH cohort, 22q11.22 loss differs by subtype: Down's syndrome 93%, BCR-ABL1 47%, ETV6-RUNX1 46%, Pseudodiploid 38%, Hyperdiploid 30%, Other 33%, MLL 12%, TCF3-PBX1 12%, T-ALL 7%, Hypodiploid 0%. In the TARGET cohort, 22q11.22 loss confers a trend for worse event-free survival (EFS; P=0.17) and overall survival (OS; P=0.06). Compared to normal 22q11.22/IKZF1 (77% EFS) or IKZF1 alternations alone (45% EFS), combined 22q11.22 loss plusIKZF1 alterations drastically reduces EFS to 15% (P

Description: Abstract 1589 Poster Board I-615 Genome-wide, high-resolution analyses of copy number alterations (CNAs) now play an increasingly important role in identifying new genomic loci associated with leukemia biology and prognosis. The most powerful of these studies include large numbers of patients with associated clinical features and outcome data. Molecular Inversion Probes (MIPs) analyze genetic target sequences in parallel at the highest genomic resolution and can detect both gene copy number and allelic imbalance in clinical samples, and have been demonstrated to work on archived formalin-fixed paraffin-embedded (FFPE) samples as old as 20 years. In this pilot study, we report for the first time the successful interrogation of high-resolution CNAs in archived FFPE samples in childhood leukemia. We first extracted genomic DNA from FFPE bone marrow aspirate clots from 18 pediatric patients diagnosed with precursor B-cell acute lymphoblastic leukemia (pre-B ALL) diagnosed between 2006-2008 at Primary Children's Medical Center at University of Utah. DNA from paired remission samples was also extracted for each patient, again using archived FFPE bone marrow aspirate clots. Blast percentages on pre-B ALL marrow clots ranged from 39-99% (Mean 88%, Median 94%). Genomic DNA was isolated using RecoverAll” Total Nucleic Acid Isolation Kit for FFPE Tissues (Ambion®, Applied Biosystems). Clinical features and outcome data were readily available and abstracted from the medical record. The 18 patients in the FFPE cohort included: ages 2-21 years old (Median 5.5 years old), 7 females, presenting WBC 1-75 × 103/uL (Median 3.9), CNS negative disease (n=18), no reported cytogenetic abnormalities (n=8), t(12;21) [n=6], t(9;22) [n=1], and MLL rearrangement (n=1). 7 patients were designated “High Risk” by NCI-Rome Criteria and 1 patient relapsed. The MIP assay was run using the customized 330K Cancer Panel (Affymetrix®, Santa Clara, CA), which includes both cancer-specific SNPs and genome-wide coverage with a median probe distance of 4,207 basepairs (bp). Copy number was calculated by comparing leukemia samples to pooled normal control signal intensity for each probe. CNA calls were based on 5 consecutive probes with 〉90% call rate, standard deviation 〈 20%, and copy number ' 1.2 or ≥ 2.8. MIPs revealed remarkably high-quality CNA data for each of the 18 FFPE samples, including the cytogenetically “normal” patients. Both known and novel recurring CNA loci were identified in this cohort. Deletions included: 14q11.2 (n=11 [61%], 51569 bp, no known genes), 22q11.2 (n=10 [56%], 185944 bp, VPREB1), 14q32.33 (n=10 [56%], 631377 bp, no known genes), 7q34 (n=9 [50%], 292586 bp, PRSS1, TRY6, PRSS2), and 12p13.2 (n=6 [33%], 292586 bp, ETV6). Gains included: 10p15.2 (n=10 [56%], 26481 bp, PFKP), 10q26.3 (n=10 [56%], 69691 bp, MGMT), 10p11.21 (n=8 [44%], 922257 bp, FZD8, CCNY, GJD4), and 8p23.3 (n=7 [39%], 90307 bp, ARHGEF10). Interestingly, of 52 amplified segments recurring in 35% or greater of samples, all but one were located in chromosome 10, 14, 17, 18, or 21 suggesting genetic amplification hotspots. Additionally, 100% of IKZF1-deleted samples (n=2/2) compared to 37.5% of IKZF1-normal samples (n=6/16) had M3 marrows (〉25% blasts) at Day 7. These same two IKZF1-deleted FFPE samples belonged to the patients with the two highest WBC values at presentation (75.3, 22.9 × 103/uL), in addition to containing two of the six highest bone marrow blast percentages at diagnosis. This is consistent with known findings that IKZF1 deletions are associated with high-risk ALL. In this pilot study of 18 patients, we have shown that archived FFPE bone marrow aspirate specimens (standard of care for all bone marrow procedures) can be used successfully for known and novel CNA analysis in childhood leukemia. We believe this is the first time that high-resolution, genome-wide CNA data from FFPE samples in any type of leukemia have been reported. This is an important development in CNA studies of hematological disease because now it is possible to investigate an unlimited number of archived FFPE childhood leukemia samples from around the world, or to explore rarer and more difficult to find specimens such as relapse or concordant identical twins. Disclosures No relevant conflicts of interest to declare.

Description: Abstract 1192 T cell acute lymphoblastic leukemia (T-ALL) is a challenging clinical entity. Over half of adult, and about 20% of pediatric, T-ALL patients either relapse or fail to achieve remission, and despite salvage attempts outcomes are poor. To discover new acquired genetic changes that occur in T-ALL, as well as those that contribute to disease relapse, we studied zebrafish (Danio rerio) T-ALL samples using array comparative genomic hybridization (aCGH). Five different D. rerio T-ALL models have been described, and all 5 develop neoplasias that clinically and molecularly resemble human T-ALL. We performed aCGH on 14 de novo T-ALLs representing 4 of these models. To evaluate possible similarities between human and zebrafish copy number aberrations (CNAs) in T-ALL, we compared all D. rerio genes found in any CNA in our 14 zebrafish cancers to a cohort comprised of 61 published primary human T-ALLs analyzed by aCGH. In these D. rerio CNAs, we identified 764 genes with human homologues. Of these, we found significant overlap (62%) with genes in CNAs from the human T-ALL dataset. In addition, 10 genes recurrently altered (〉3 samples) in zebrafish T-ALL were also seen in CNAs from 5 or more human T-ALL cases, suggesting a conserved role for these loci in T-ALL transformation across vertebrate species. In addition to studying primary T-ALL, we also conducted iterative allo-transplantations with 3 zebrafish malignancies. This technique selects for particularly aggressive disease, increasing engraftment rates and also resulting in shorter time to death of recipient animals in successive transplant rounds. Because these passaged cancers show more malignant behavior in vivo, this procedure models refractory and relapsed T-ALL. In these 3 serially-passaged cancers, 55% of the original CNAs were preserved after iterative transplantation, demonstrating the clonal relationship between the primary and passaged leukemia. In addition, 101 CNAs were acquired during passaging of the 3 T-ALLs. Genes in these loci may underlie the enhanced malignant behavior of these neoplasias. We also compared genes found in CNAs from passaged zebrafish malignancies to human aCGH results from 50 T-ALL patients who failed induction, went on to relapse, or had already relapsed. Again, many genes (n=76) were present in both datasets. Four genes altered in 2 of 3 D. rerio samples were also found in 〉4 cases from the human dataset. Collectively, these results suggest that zebrafish and human T-ALL are similar at the genomic level, with several homologous genes commonly gained or lost by cancers from both organisms. Genes from recurring CNAs in disease samples from both species may have particular relevance, as these candidates likely have conserved roles in T cell oncogenesis or T-ALL disease progression. As the comprehensive list of genes in CNAs from human T-ALL cases is vast and heterogeneous (comprising 15,724 unique genes from just 75 clinical samples), zebrafish T-ALL models may permit a more expedient prioritization of these candidates for further investigation. Disclosures: No relevant conflicts of interest to declare.