Constitutional and Somatic Variants in ARID5B in Pediatric B-Cell Precursor Acute Lymphoblastic Leukemia

Abstract

INTRODUCTION Acute lymphoblastic leukemia (ALL) is the most common malignancy in children, with the high hyperdiploid (HeH) subtype accounting for approximately 25% of B-cell precursor (BCP) ALL cases. It has been shown that germline variants in the ARID5B gene in chromosome band 10q21.2 are associated with increased risk of BCP ALL, in particular HeH ALL, and that the risk alleles result in lower expression of ARID5B in hematopoietic cells. ARID5B codes for a protein involved in regulating gene expression and chromatin remodeling. We have previously reported somatic deletions in the ARID5B locus in two cases of HeH ALL, but the overall frequency of acquired copy number changes and rearrangements involving ARID5B in BCP ALL remains unknown. Here, we have investigated constitutional and somatic ARID5B variants in pediatric BCP ALL, with a particular focus on HeH cases. METHODS Constitutional variants We studied four known risk single nucleotide polymorphisms (SNPs) in the ARID5B locus (rs7090445, rs7089424, rs7073837 and rs10740055) in HeH ALL cases heterozygous for the risk SNP and with trisomy 10. These were investigated in three different cohorts, including a total of 92 cases informative for rs7090445, 92 cases for rs7089424, 119 cases for rs7073837 and 66 cases for rs10740055. The genotype and relative allele frequencies were ascertained from SNP array or whole genome sequencing (WGS) data. One-sided binomial tests were applied to investigate whether the risk allele was more often in the duplicated chromosome than the non-risk allele. Somatic variants For somatic variants, we ascertained copy number status based on SNP array and/or WGS analysis in the ARID5B region in a total of 466 pediatric HeH ALL cases and structural rearrangements based on WGS in 77 cases. We also studied, using SNP array analysis, somatic copy number variants in a separate cohort consisting of 590 non-HeH BCP ALL cases. To compare the proportions of deletions in the HeH cohort and in the other genetic subtypes, we used Fisher's Exact two-sided test. RESULTS Constitutional variants All four risk SNPs showed a significantly higher proportion of risk allele duplication (one-sided binominal test); rs7090445 ( P=0.009), rs7089424 ( P=0.005), rs7073837 ( P=0.03) and rs10740055 ( P=0.04). This validates that there is a clonal selection for HeH blast cells with gain of the chromosome 10 homologue carrying the risk allele as opposed to gain of the homologue that carries the non-risk allele. Somatic variants Somatic deletions targeting ARID5B were found in 9/466 cases (1.9%) of HeH cases. The deletions covered different parts of ARID5B, with no minimally deleted region, suggesting that the functional outcome was downregulation of ARID5B expression. In the cohort with non-HeH BCP ALL (other genetic subtypes), 4/590 somatic deletions were found (0.68%). There was no statistically significant difference between the frequency of copy number aberrations targeting ARID5B between HeH and the other genetic subtypes of ALL (P=0.09). WGS analysis of the HeH cases revealed one translocation and one missense mutation in 2/77 cases (2.6%). The translocation involved the PAN3 and ARID5B genes and was also present in RNA sequencing data from this case. CONCLUSIONS We show that our previous finding that HeH ALL constitutionally heterozygous for ARID5B risk alleles and with an acquired trisomy 10 more commonly duplicates the chromosome 10 homologue carrying the risk allele holds true in a much larger cohort. Furthermore, somatic deletions involving ARID5B are recurrent in pediatric BCP ALL.