Molecular allelokaryotyping of pediatric acute lymphoblastic leukemias by high-resolution single nucleotide polymorphism oligonucleotide genomic microarray

Abstract

Pediatric acute lymphoblastic leukemia (ALL) is a malignant disease resulting from accumulation of genetic alterations. A robust technology, single nucleotide polymorphism oligonucleotide genomic microarray (SNP-chip) in concert with bioinformatics offers the opportunity to discover the genetic lesions associated with ALL. We examined 399 pediatric ALL samples and their matched remission marrows at 50,000/250,000 SNP sites using an SNP-chip platform. Correlations between genetic abnormalities and clinical features were examined. Three common genetic alterations were found: deletion of ETV6, deletion of p16INK4A, and hyperdiploidy, as well as a number of novel recurrent genetic alterations. Uniparental disomy (UPD) was a frequent event, especially affecting chromosome 9. A cohort of children with hyperdiploid ALL without gain of chromosomes 17 and 18 had a poor prognosis. Molecular allelokaryotyping is a robust tool to define small genetic abnormalities including UPD, which is usually overlooked by standard methods. This technique was able to detect subgroups with a poor prognosis based on their genetic status.

Figure 1

Result of SNP-chip analysis. (A) Results of normal and abnormal chromosomes visualized by CNAG software. Blue lines above each chromosome show total gene dosage; level 2 indicates diploid (2N) amount of DNA, which is normal. Green bars under each chromosome indicate the SNP sites showing heterozygosity in leukemic cells. When heterozygosity is not detected in the leukemic cells but is detected in matched normal controls, the result suggests that the leukemic cells have allelic imbalance (AI) in that region. Pink bars that replaced green ones suggest AI. The bottom lines (green and red lines) in each panel show allele-specific gene dosage levels (one indicates the gene dosage of paternal allele, the other indicates the gene dosage of maternal allele). Level 1 is normal for each gene dosage. (i) Pattern of normal chromosome 9. Blue line is at level 2 (2N DNA). Large number of SNP sites shows normal heterozygosity (green bars under the chromosomes), and no pink bars are detected. Allele-specific gene dosage is at level 1. Panel ii shows the pattern of whole chromosome 9 uniparental disomy (UPD) detected by SNP-chip. Total gene dosage (blue line) is normal (level 2). A number of pink bars (AI) are detected. Allele-specific gene dosage data show that one allele is deleted (level 0) and the other allele is duplicated (level 2). Panel iii shows the pattern of partial UPD. Left half shows the pattern of UPD as described above. Right half shows the pattern of a normal chromosome as described above. This case also has homozygous deletion of p16INK4A gene (see that both allele-specific dosage lines [green and red lines] and total gene dosage line [blue line] are at zero). Panel iv shows nondominant UPD. Total gene dosage (blue line) indicates 2N. Allele-specific gene dosage lines (arrows, green and red lines) on left half show that one allele (green line) is lower than normal, and the other allele (red line) is higher than normal. Allele-specific gene dosage on right half show that each allele has same level. (B-D) Validation of SNP-chip data by direct nucleotide sequencing of SNP sites and qPCR. Top panels: direct nucleotide sequencing of SNP sites in ALL samples with matched controls. ALL indicates leukemic samples; N, matched control samples. Heterozygous SNP sites are indicated by arrows. Middle panels: results of SNP-chip data (see Figure 1 legend). Bottom panels: qPCR at each chromosome location. Gene dosage levels were examined using qPCR at indicated chromosomal region. Gene dosage was determined relative to the levels at the 2p21 region. Gene dosage in leukemic cells (ALL) was compared with the matched normal control DNA (N). (B) ALL with 9p hemizygous deletion; homozygous deletion of 9p21 is also detected. (C) ALL with whole chromosome UPD. (D) ALL with 9p UPD.

Figure 2

Allelokaryotyping of pediatric ALL. (A) Genetic clustering of 393 ALL samples. Genetic status of each chromosomal region is visualized. Vertical axis: chromosomes, p: short arms, q: long arms. Horizontal axis: individual cases. CN: copy number of alleles. UPD: uniparental disomy. Locations of PBX1, INK4A/ARF(p16), and ETV6 genes are indicated. A rectangle indicates cases having 9p and 20q deletions. (B) Three common genomic abnormalities detected in ALL by SNP-chip analysis. Left panel: Deletion of 9p is frequently detected; the arrow indicates the commonly deleted region (CDR) where the p16INK4A gene is located. Right panel: Deletion of 12p often occurs. The arrow indicates that the CDR is where the ETV6 gene is localized. Green lines under the chromosome indicate the deleted regions in individual cases. Brown lines above the chromosome indicate the duplicated regions. Only representative cases are shown. Green and red bands on idiograms indicate centromeres and noncoding regions, respectively. Bottom panel: Venue diagram of 3 common abnormalities detected in this study. Numbers of respective cases in each category are indicated. HD: ALL with hyperdiploid (chromosomes > 50). ETV6Del: ALL with deletion of ETV6 gene. p16Del: ALL with deletion of p16INK4A gene.

Figure 3

Numerical chromosomal changes and uniparental disomy in pediatric ALL. (A) Frequency of pentasomy/tetrasomy/trisomy affecting each chromosome. For X chromosome, trisomy (105 cases) contains trisosomy X in male pateints (67 cases) and disomy X in female patients (38 cases). All tetrasomy X were female patients. (B) Frequency of uniparental disomy (UPD). Whole: cases with whole chromosome UPD; p arm: cases with UPD of short arm; and q arm: cases with UPD of long arm. Chr: chromosome. UPD involving X chromosome was detected only in female cases. (C) Distribution of whole and partial chromosome UPD in HD and non–HD-ALL. Whole chromosome UPD is frequently detected in HD-ALL. Thirty-four cases with whole chromosome UPD were HD-ALL. Partial UPD is frequently detected in non–HD-ALL. Fifteen of 63 cases with partial UPD were HD-ALL. (D) Frequency of deletion of p16INK4A gene in whole chromosome 9 UPD and 9p UPD. Twenty-three cases showed deletion of p16INK4A, out of a total of 30 cases with 9p UPD. One case had p16INK4A deletion from a total 18 ALL samples with whole chromosome 9 UPD.

Figure 4

Prognostic impact of chromosomes 17 and 18. Absence of a gain of either chromosome 17 (left panel) or chromosome 18 (middle panel) is associated with poor prognosis in patients with HD-ALL; concurrent absence of a gain of both chromosomes 17 and 18 (right panel) is associated with very poor prognosis. Tri/tetra 18 and 17: HD-ALL with trisomy or tetrasomy 18 or 17. Others: HD-ALL without gain of chromosomes 17 and/or 18.