ARID5B genetic polymorphisms contribute to racial disparities in the incidence and treatment outcome of childhood acute lymphoblastic leukemia

Abstract

Purpose: Recent genome-wide screens have identified genetic variations in ARID5B associated with susceptibility to childhood acute lymphoblastic leukemia (ALL). We sought to determine the contribution of ARID5B single nucleotide polymorphisms (SNPs) to racial disparities in ALL susceptibility and treatment outcome.

Patients and methods: We compared the association between ARID5B SNP genotype and ALL susceptibility in whites (> 95% European genetic ancestry; 978 cases and 1,046 controls) versus in Hispanics (> 10% Native American ancestry; 330 cases and 541 controls). We determined the relationships between ARID5B SNP genotype and ALL relapse risk in 1,605 children treated on the Children's Oncology Group (COG) P9904/9905 clinical trials.

Results: Among 49 ARID5B SNPs interrogated, 10 were significantly associated with ALL susceptibility in both whites and Hispanics (P < .05), with risk alleles consistently more frequent in Hispanics than in whites. rs10821936 exhibited the most significant association in both races (P = 8.4 × 10(-20) in whites; P = 1 × 10(-6) in Hispanics), and genotype at this SNP was highly correlated with local Native American genetic ancestry (P = 1.8 × 10(-8)). Multivariate analyses in Hispanics identified an additional SNP associated with ALL susceptibility independent of rs10821936. Eight ARID5B SNPs were associated with both ALL susceptibility and relapse hazard; the alleles related to higher ALL incidence were always linked to poorer treatment outcome and were more frequent in Hispanics.

Conclusion: ARID5B polymorphisms are important determinants of childhood ALL susceptibility and treatment outcome, and they contribute to racial disparities in this disease.

Fig 1.

rs10821936 genotype frequency by acute lymphoblastic leukemia (ALL) subtype, sex, and age at diagnosis. The risk allele (allele C) at rs10821936 was significantly over-represented in the hyperdiploid subtype of ALL in both whites (A) and Hispanics (B), although no difference was observed by sex or age groups. P values were determined by χ² tests. Ctrls, controls.

Fig 2.

ARID5B single nucleotide polymorphisms are associated (CC, CT, and TT) with both acute lymphoblastic leukemia (ALL) susceptibility and relapse. (A) Relationship between genotype at ARID5B single nucleotide polymorphism rs6479778 and ALL relapse in the Children's Oncology Group P9904/9905 clinical trials. P value was determined by Fine and Gray's regression model after adjusting for treatment arm and ancestry. (B) Allele T at rs6479778 was more frequent in ALL cases than in controls and more common in Hispanics than in whites. P values were estimated by logistic regression.

Fig 3.

Racial differences in the risk allele frequency at rs10821936 and in acute lymphoblastic leukemia (ALL) incidence. (A) Genotype of rs10821936 is associated with ALL susceptibility in all three race groups. (B) Frequency of the risk allele (allele C) increases in order for blacks, whites, and Hispanics, consistent with the racial differences in ALL incidence. *Association of rs10821936 with ALL in blacks is based on a previous report by Yang et al. †ALL incidence by race is based on the report by Linabery et al.