Genome-Wide Association Study Identifies Germline Polymorphisms Associated with Relapse of Childhood Acute Lymphoblastic Leukemia

Treatment outcome of childhood acute lymphoblastic leukemia (ALL) has improved dramatically in the last 40 years thanks to risk-directed therapy. However, a substantial proportion of patients still experience relapse, many of whom have no known risk factors. Prior efforts to improve risk stratification have primarily focused on genetic variations of the tumor (e.g., cytogenetic abnormalities) or on assessment of early antileukemic response (e.g., upfront prednisone response, minimal residual disease [MRD] after remission induction). The role of inherited genetic variation on treatment response in children with ALL is poorly characterized.

We performed a genome-wide association study to evaluate the association of genotypes at 444,044 germline SNPs with the risk of relapse in 2,535 children with newly diagnosed ALL enrolled on 5 frontline clinical trials: St. Jude Total Therapy XIIIB, XV, the Children's Oncology Group (COG) P9904, P9905, and P9906 protocols. The associations between SNP genotypes and ALL relapse were evaluated by Gray's test and by Fine and Gray's hazard regression model, adjusting for genetic ancestry and treatment regimens. To identify SNPs reproducibly associated with relapse across treatment regimens, we performed 100 rounds of discovery and replication, each round dividing patients into 2 cohorts at a 1:1 ratio. Each time, we used the discovery cohort to perform a genome-wide screen, and then filtered SNPs based on their association in the replication cohort. Finally, SNPs were prioritized on the basis of the number of times their association with relapse was replicated.

A total of 134 SNPs, representing 88 genomic loci, were replicated in at least 10 rounds of discovery-replication tests. Across the genome, the strongest association with relapse risk was observed at 14q22.1 in the PYGL gene (rs7142143). Each copy of the C allele at this PYGL intronic SNP (rs7142143) conferred a 3.6-fold increase in the hazard rate of relapse (P=6.7×10⁻⁹) and the association of this SNP with relapse was replicated in 79 of 100 rounds of discovery-replication tests. Of 134 relapse SNPs, 73 were associated with one or more known prognostic clinical features in childhood ALL: 32 SNPs were related to leukocyte count ≥50,000/μl at diagnosis; 19 were enriched in children older than 10 years of age; 16 were associated with hyperdiploid ALL (DNA index ≥ 1.16); and 61 were associated with ALL molecular subtype and/or lineage (MLL rearrangements, ETV6-RUNX1, TCF3-PBX1, BCR-ABL1, or T-cell ALL). Interestingly, 110 of the 134 relapse SNPs (82%) were prognostic even among MRD-negative patients, and 133 (99%) remained significantly associated with relapse after adjusting for all known risk factors, strongly indicating the potential value of germline genetic variations in ALL risk classification. To explore the mechanisms by which SNPs might influence treatment outcome of ALL, we examined the association of the 134 relapse SNPs with four pharmacokinetic and pharmacodynamic endophenotypes in the St. Jude Total XIIIB and XV cohorts: methotrexate plasma clearance, intracellular accumulation of polyglutamated (active) methotrexate, dexamethasone plasma clearance, and asparaginase antibody levels. Fourteen of the 134 relapse SNPs were significantly associated with at least one of the four pharmacologic phenotypes in a manner consistent with a pharmacokinetically intuitive association with relapse (i.e., lower drug exposure translated into a higher risk of relapse).

In this genome-wide association study, we systematically identified host genetic variations related to treatment outcome of childhood ALL, most of which were prognostic independent of known risk factors for relapse, and some also influenced outcome by affecting host disposition of antileukemic drugs.

No relevant conflicts of interest to declare.