Genome-Wide Association Study of Susceptibility Loci for T-Cell Acute Lymphoblastic Leukemia in Children

Abstract

Background: Acute lymphoblastic leukemia (ALL) is the most common cancer in children and can arise in B or T lymphoid lineages. Although risk loci have been identified for B-ALL, the inherited basis of T-ALL is mostly unknown, with a particular paucity of genome-wide investigation of susceptibility variants in large patient cohorts.

Methods: We performed a genome-wide association study (GWAS) in 1191 children with T-ALL and 12 178 controls, with independent replication using 117 cases and 5518 controls. The associations were tested using an additive logistic regression model. Top risk variants were tested for effects on enhancer activity using luciferase assay. All statistical tests were two sided.

Results: A novel risk locus in the USP7 gene (rs74010351, odds ratio [OR] = 1.44, 95% confidence interval [CI] = 1.27 to 1.65, P = 4.51 × 10-8) reached genome-wide significance in the discovery cohort, with independent validation (OR = 1.51, 95% CI = 1.03 to 2.22, P = .04). The USP7 risk allele was overrepresented in individuals of African descent, thus contributing to the higher incidence of T-ALL in this race/ethnic group. Genetic changes in USP7 (germline variants or somatic mutations) were observed in 56.4% of T-ALL with TAL1 overexpression, statistically significantly higher than in any other subtypes. Functional analyses suggested this T-ALL risk allele is located in a putative cis-regulatory DNA element with negative effects on USP7 transcription. Finally, comprehensive comparison of 14 susceptibility loci in T- vs B-ALL pointed to distinctive etiology of these leukemias.

Conclusions: These findings indicate strong associations between inherited genetic variation and T-ALL susceptibility in children and shed new light on the molecular etiology of ALL, particularly commonalities and differences in the biology of the two major subtypes (B- vs T-ALL).

Figure 1.

Genome-wide association study of childhood T-ALL. The association between genotype and ALL was evaluated by using a logistic regression model for 7 967 910 genotyped/imputed single-nucleotide polymorphisms (SNPs) in 1191 T-ALL cases and 12 178 unrelated non-ALL controls. P values (−log₁₀P) were plotted against the respective chromosomal position of each SNP. Gene symbols are indicated for two loci achieving genome-wide significance (two-sided P < 5 × 10^–8, dashed black horizontal line): CDKN2A (9p21) and USP7 (16p13.3). T-ALL = T-cell acute lymphoblastic leukemia.

Figure 2.

Functional annotation and activity screen of regulatory DNA at the USP7 locus. A) Functional annotation of regulatory DNA at the USP7 locus. Genomic positions and scale for the human genome assembly February 2009 (GRCh37/hg19) are shown on the top. Genome-wide statistically significant T-ALL risk variants in USP7 (P < 5 × 10^–8) are marked in the middle panel, and the log-transformed P values (for association with T-ALL) are shown in the bed graph. The gene structure, ChIP-seq signals for histone modifications (ie, H3K4me3, H3K27ac, and H3K79me2) in T-ALL cell lines (ie, DND41 and Jurkat) (28), and ATAC-seq signals of hematopoietic cells (29) are also included. B) Functional activity screen of cis-elements with different genotypes/haplotypes by luciferase reporter assay in Jurkat cells. Jurkat cells were transiently transfected with pGL4.23 [luc2/minP] construct (luciferase gene with USP7 intronic sequences spanning GWAS hit SNPs for each allele) and pGL-TK (Renilla luciferase). Firefly luciferase activity was measured 24 hours post-transfection and normalized to Renilla luciferase activity. Relative luciferase activity indicates the ratio over the value from pGL4.23 vector alone. All experiments were performed in triplicate, and repeated three times for rs74010351. Statistical significance was evaluated by using two-sided Student's t test. Error bars indicate SD. SNP = single-nucleotide polymorphisms; T-ALL = T-cell acute lymphoblastic leukemia; WT = wild-type.

Figure 3.

Relationship of USP7 genotype and/or somatic mutation with T-ALL subgroups and clinical features. A) This analysis was restricted to 190 T-ALL children in the COG AALL0434 cohort with both germline and somatic genomic data available. Proportion of T-ALL patients with USP7 risk genotype at rs74010351 and/or USP7 somatic mutation was calculated within each T-ALL subgroup with at least 10 cases (TLX3, TLX1, LMO2/LYL1, HOXA, TAL1, and T-other) (5). The association of USP7 status with T-ALL subgroups was evaluated by Fisher exact test. B and C) The analysis was restricted to patients with T-ALL in the COG AALL0434, representing a largely unselected and nationwide patient population. The correlation between USP7 genotype and genetic ancestry was tested by generalized linear regression test. The association of USP7 genotype at rs74010351 with age at diagnosis was evaluated by logistic regression test (ie, age < 10y vs age ≥ 10y) adjusting for genetic ancestry. All statistical tests were two sided. T-ALL = T-cell acute lymphoblastic leukemia.

Figure 4.

Similarities and differences in genetic predisposition to ALL between B and T lineage. The association between genotype and ALL was evaluated by using a logistic regression model for GWAS hits reported previously in B-ALL (ie, ARID5B [rs10821936], IKZF1 [rs11978267], CDKN2A/B [rs11978267], IKZF3 [rs17607816], PIP4K2A [rs7088318], 8q24.21 [rs28665337], TP63 [rs17505102], SP4 [rs2390536], GATA3 [rs3824662], LHPP [rs35837782], ELK3 [rs4762284], CEPBE [rs2239633], and 2q22.3 [rs17481869]) or newly identified in T-ALL (ie, USP7 [rs74010351]) in 1191 T-ALL cases (COG AALL0434 and St. Jude Total Therapy XVI protocols) vs 12 178 unrelated non-ALL controls (HRS), and 1824 B-ALL cases (COG P9904/9905/9906) vs 5518 unrelated non-ALL controls (MESA). The locus specifically statistically significant (P < .05) in T-ALL is marked in red, with B-ALL marked in blue, and common marked in green. ALL = acute lymphoblastic leukemia; COG = Children’s Oncology Group; HRS = Health and Retire Study; MESA = Multi-Ethnic Study of Atherosclerosis.