Genome-wide analysis links NFATC2 with asparaginase hypersensitivity

Abstract

Asparaginase is used to treat acute lymphoblastic leukemia (ALL); however, hypersensitivity reactions can lead to suboptimal asparaginase exposure. Our objective was to use a genome-wide approach to identify loci associated with asparaginase hypersensitivity in children with ALL enrolled on St. Jude Children's Research Hospital (SJCRH) protocols Total XIIIA (n = 154), Total XV (n = 498), and Total XVI (n = 271), or Children's Oncology Group protocols POG 9906 (n = 222) and AALL0232 (n = 2163). Germline DNA was genotyped using the Affymetrix 500K, Affymetrix 6.0, or the Illumina Exome BeadChip array. In multivariate logistic regression, the intronic rs6021191 variant in nuclear factor of activated T cells 2 (NFATC2) had the strongest association with hypersensitivity (P = 4.1 × 10(-8); odds ratio [OR] = 3.11). RNA-seq data available from 65 SJCRH ALL tumor samples and 52 Yoruba HapMap samples showed that samples carrying the rs6021191 variant had higher NFATC2 expression compared with noncarriers (P = 1.1 × 10(-3) and 0.03, respectively). The top ranked nonsynonymous polymorphism was rs17885382 in HLA-DRB1 (P = 3.2 × 10(-6); OR = 1.63), which is in near complete linkage disequilibrium with the HLA-DRB1*07:01 allele we previously observed in a candidate gene study. The strongest risk factors for asparaginase allergy are variants within genes regulating the immune response.

Figure 1

Manhattan plot of P values for genome-wide SNP association with asparaginase hypersensitivity. (A) The Manhattan plot shows the negative log10 of the P values for association with asparaginase hypersensitivity vs each chromosome on the x-axis. The open circles indicate SNPs genotyped using the Affymetrix 6.0 array, whereas the open squares indicate SNPs genotyped by the Illumina Exome Beadchip array. The dashed line denotes the genome-wide threshold (P = 5 × 10⁻⁸). (B) SNPs genotyped using the Affymetrix SNP 6.0 array identified the NFATC2 rs6021191 variant associated with asparaginase hypersensitivity (P = 4.1 × 10⁻⁸, OR = 3.11, CI = 2.07-4.66). AA, subjects homozygous for the reference allele (A); TT, subjects homozygous for the variant allele (T); AT, heterozygotes. (C) Illumina Exome array genotyping identified the strongest association with the HLA-DRB1 rs17885382 variant (P = 3.2 × 10⁻⁶, OR = 1.63, CI = 1.33-2.00). CC, subjects homozygous for the reference allele (C); TT, subjects homozygous for the variant allele (T); CT, heterozygotes. Associations were determined using a general linear model adjusted for treatment, ALL immunophenotype, gender, age group, and ancestry.

Figure 2

The NFATC2 rs6021191 variant is an eQTL. RNA-seq data were available from the ALL tumor samples of 65 SJCRH patients and from 52 unrelated YRI HapMap samples to investigate the association between rs6021191 and NFATC2 gene expression. We found higher NFATC2 gene expression in samples with the rs6021191 variant compared with samples without the variant in both (A) ALL tumor samples (P = 1.1 × 10⁻³) and (B) YRI HapMap samples (P = .03). Allele abbreviations are explained in Figure 1.

Figure 3

Patients carrying both the HLA-DRB1 rs17885382 and NFATC2 rs6021191 variant have a higher risk of developing asparaginase hypersensitivity compared with patients with no risk variant. The risk of developing hypersensitivity was determined for patients carrying a single-risk variant (HLA-DRB1 rs17885382 or NFATC2 rs6021191) or for patients carrying both risk variants (NFATC2 rs6021191 and HLA-DRB1 rs17885382). The risk of hypersensitivity was higher in patients carrying a single variant (OR_rs6021191 = 3.07, CI_rs6021191 = 1.87-4.94, P_rs6021191 = 5.4 × 10⁻⁶; OR_rs17885382 = 1.66, CI_rs17885382 = 1.3-2.1, P_rs17885382 = 3.5 × 10⁻⁵) or both variants (OR = 8.58, CI = 3.14-22.92, P = 1.7 × 10⁻⁵) compared with patients with no risk alleles. The associations were determined using a general linear model adjusted for treatment, ALL immunophenotype, gender, age group, and ancestry.

Figure 4

Genes involved in T-cell function may contribute to the risk of developing asparaginase hypersensitivity reactions. IPA was used to determine if genes most associated with asparaginase hypersensitivity were overrepresented in specific biological pathways. A total of 334 genes invoked by 499 SNPs associated with asparaginase hypersensitivity annotated to 234 distinct IPA pathways. (A) The top 20 pathways identified, all with P < 3 × 10⁻³, were enriched for genes involved in T-cell apoptosis, T-cell signaling, T-cell activation, or T-cell disorders. (B) A total of 33 genes containing SNPs associated with hypersensitivity were included within the top 20 canonical pathways (shown on x-axis). Many of these genes were included in multiple pathways. The number of pathways (out of 20) in which each of these 33 genes is involved is plotted on the y-axis. For example, HLA-DRB1 is involved in 14 out of the 20 top canonical pathways. GPCR, G-protein coupled receptor; nNOS, neuronal NOS; IL, interleukin; iCOS, inducible T-cell costimulator and its ligand, iCOSL.