Children with Down syndrome (DS) have an increased risk of developing acute lymphoblastic leukemia (ALL), and consistently demonstrate poorer outcomes due to higher rates of both relapse and treatment-related mortality compared to other children with ALL. The biology of ALL in DS is unique, with lower frequency of the classic cytogenetic lesions generally observed in childhood ALL, and increased frequency of JAK2 mutations and CRLF2 rearrangements, which are not clearly associated with adverse prognosis in DS-ALL. In order to improve risk stratification and identify potential novel therapeutic targets in this vulnerable population, we analyzed 90 DS-ALL cases for prognostically significant copy number abnormalities (CNAs) in a collaborative cohort from the Children's Oncology Group (n=30), St. Jude Children's Research Hospital (n=22), AIEOP (n=16), Texas Children's Cancer Center (n=10), UKALL 2003 (n=6) and an archival UK sample (n=1), and Utah's Primary Children's Medical Center (n=5). Copy number profiling was performed using 500K, 6.0, CytoScan HD, and OncoScan FFPE Express arrays (Affymetrix), and Human CNV370-Duo arrays (Illumina). Gene expression profiling was performed using U133 Plus2.0 arrays (Affymetrix). Copy number was analyzed with Nexus Copy Number (BioDiscovery, Inc.) and gene expression was analyzed with Partek Genomics Suite (Partek, Inc.) and Gene Set Enrichment Analysis (Broad Institute). Deletions of a focal region on 22q11.22 (present in 28.9% of cases, similar to the incidence previously reported in a non-DS cohort [Mangum et al, ASH 2011:741]), and deletions of IKZF1 (present in 20.0% of cases), were significantly associated with poor event-free survival (EFS) (5-year EFS was 88.6 ± 6.3% in wild-type cases [n=31], 68.1 ± 13.3% in cases with deletion of 22q11.22 only [n=15], and 60.0 ± 21.9% in those with deletion of IKZF1 only [n=5]; combined deletion [n=6] was associated with an even poorer EFS (33.3 ± 19.3%, p<0.0001, Figure 1A). Patients were observed to have focal deletions on 22q11.22 spanning up to 235 kb; the most common recurring shared region is just under 10 kb in length, and does not contain known coding genes. Increased number of copies of a 14 kb region in the platelet-derived growth factor receptor alpha gene (PDGFRA), located at 4q12, occurred in 35.5% of cases performed on microarrays containing evaluable probes in this region, and were significantly associated with poor outcome (4-year EFS 87.5 ± 8.3% in wild-type cases [n=20] and 33.9 ± 17.6% in PDGFRA gain cases [n=11], p = 0.004, Figure 1B). Five cases contained 3 copies of PDGFRA, and six cases had at least 4 copies, with no correlation between outcome and number of copies gained. PDGFRA rearrangements and point mutations have been reported in other malignancies, but copy number gain is a novel mechanism of alteration. Cases with a focal 22q11.22 deletion were associated with an increased frequency of CRLF2 rearrangement (75% versus 47%, Fisher's exact p=0.028). No other differences in age, initial WBC, or CRLF2 rearrangement in cases with versus without the focal 22q11.22 deletion, IKZF1 deletion, or PDGFRA gain were present, nor were there significant differences in incidence of focal PDGFRA gains in cases with versus without IKZF1 or 22q11.22 deletions. Gene expression profiling in 27 evaluable cases demonstrated upregulation of genes with kinase activity in 12 cases with either focal 22q11.22 or IKZF1 deletions. This study represents one of the largest collaborative cohorts of DS-ALL for genomic profiling, confirms the poor prognosis associated with IKZF1 deletion (Buitenkamp et al., Leukemia 2012), and identifies two additional genomic loci strongly associated with poor prognosis. If validated in additional DS-ALL cohorts, these findings suggest key lesions that contribute to the poor outcomes observed in this population. These specific genomic changes may improve risk stratification of treatment in children with DS and ALL, and may lead to enhanced sensitivity to tyrosine kinase inhibitors in the estimated 60% of cases bearing one or more of these three lesions.
No relevant conflicts of interest to declare.
This icon denotes a clinically relevant abstract
Asterisk with author names denotes non-ASH members.