A genome-wide association study of susceptibility to acute lymphoblastic leukemia in adolescents and young adults

Abstract

Acute lymphoblastic leukemia (ALL) in adolescents and young adults (AYA) is characterized by distinct presenting features and inferior prognosis compared with pediatric ALL. We performed a genome-wide association study (GWAS) to comprehensively identify inherited genetic variants associated with susceptibility to AYA ALL. In the discovery GWAS, we compared genotype frequency at 635 297 single nucleotide polymorphisms (SNPs) in 308 AYA ALL cases and 6,661 non-ALL controls by using a logistic regression model with genetic ancestry as a covariate. SNPs that reached P ≤ 5 × 10(-8) in GWAS were tested in an independent cohort of 162 AYA ALL cases and 5,755 non-ALL controls. We identified a single genome-wide significant susceptibility locus in GATA3: rs3824662, odds ratio (OR), 1.77 (P = 2.8 × 10(-10)) and rs3781093, OR, 1.73 (P = 3.2 × 10(-9)). These findings were validated in the replication cohort. The risk allele at rs3824662 was most frequent in Philadelphia chromosome (Ph)-like ALL but also conferred susceptibility to non-Ph-like ALL in AYAs. In 1,827 non-selected ALL cases, the risk allele frequency at this SNP was positively correlated with age at diagnosis (P = 6.29 × 10(-11)). Our results from this first GWAS of AYA ALL susceptibility point to unique biology underlying leukemogenesis and potentially distinct disease etiology by age group.

Figure 1

GWAS study design. ALL susceptibility variants were identified by comparing SNP genotype frequency in AYA ALL cases compared with non-ALL controls in the discovery GWAS, followed by replication. CALGB: The Alliance-Cancer and Leukemia Group B; COG: Children’s Oncology Group; ECOG: Eastern Cooperative Oncology Group; MDACC: MD Anderson Cancer Center; SJ: St. Jude Children’s Research Hospital.

Figure 2

Genome-wide association of SNP genotype with ALL susceptibility in AYAs. The association between genotype and ALL susceptibility was evaluated by using a logistic regression model for 635 297 SNPs in 308 AYA ALL cases and 6661 non-ALL controls. P values (y-axis) were plotted against respective chromosomal position of each SNP (x-axis). Points above the blue horizontal line indicate SNPs achieving the genome-wide significant threshold (P < 5 × 10⁻⁸). Gene symbol was indicated for the GATA3 locus at 10p14.

Figure 3

Association of GATA3 SNP rs3824662 with ALL in AYAs by race/ethnicity. In the discovery group (A) the A allele at rs3824662 was overrepresented in AYA ALL cases relative to non-ALL controls. This association was true within the European Americans (>95% European genetic ancestry), African Americans (>70% African ancestry), or Hispanic Americans (>10% Native American genetic ancestry, and Native American ancestry > African genetic ancestry). (B) Similar association was confirmed in the replication group (1-tailed test). Genetic ancestry was determined by using STRUCTURE (version 2.2.3) with HapMap CEU, YRI, CHB/JPT, and indigenous Native Americans as reference populations.

Figure 4

GATA3 SNP genotype and ALL genetic subtypes in AYAs. The allele frequency of rs3824662 varied substantially by ALL somatic genomic abnormalities, with the ALL risk allele underrepresented in hyperdiploid cases and more common in the Ph-like subtype. Numbers are based on the ALL cases included in the discovery GWAS (N = 308). The frequency of A allele at rs3824662 was 20% among unrelated non-ALL controls (MESA).

Figure 5

GATA3 SNP rs3824662 and age at ALL diagnosis. In a largely unselected cohort of ALL cases enrolled on the Children’s Oncology Group (COG) P9900 trials (N = 1,827), the frequency of ALL risk allele at rs3824662 was positively correlated with patient age at diagnosis consistently across race/ethnicity. Inset figure: the relative risk of ALL (odds ratio) conferred by each copy of the A allele at rs3824662 increased progressively with age, as estimated by logistic regression after adjusting for genetic ancestry. Horizontal dotted line is odds ratio of 1. Unrelated participants from the Multi-ethnic Study of Atherosclerosis (MESA) were considered as non-ALL controls.