Background
Pediatric acute lymphoblastic leukemia (ALL) is the most common childhood cancer globally. However, patients experience wide inter-individual variability in treatment outcomes and toxicity. Pharmacogenetic studies have identified some genetic variants as biomarkers for treatment dosing and outcomes in pediatric ALL patients; yet under-studied regions including the Middle East have distinct genetic profiles warranting investigation.

Objectives
This study aims to investigate candidate and genome-wide genetic variants associated with variability in 6-mercaptopurine (6-MP) dosing in pediatric ALL patients using two cohorts from the Middle East, as well as their potential impact on gene expression.

Methods
We collected clinical data and performed whole-genome sequencing (30×) on blood samples of 255 pediatric ALL patients from Sidra Medicine and American University of Beirut Medical Center (AUBMC). Haplotypes of 13 candidate genes (NUDT15, TPMT, ABCC4, DROSHA, FOLH1, FTO, GATA3, GNMT, GSTP1, ITPA, MED4, PACSIN2, SLCO1B1) were constructed as 10–40 kb regions, consistent with LD structure, tagging regulatory and coding variants. These haplotypes were evaluated for association with 6-MP dose percentage (%, calculated as administered versus protocol-required dose and adjusted based on absolute neutrophil count) using mixed linear models adjusting for age, sex, ALL subtype and top ten principal components of genome data (accounting for population structure). Meta-analysis was performed using METAL software. Haplotypes with FDR < 0.2 and mean effect size ≥ 10% were considered significant. Functional impact of variants within 6-MP-related haplotypes was further screened for RNA expression quantitative trait loci (e-QTL) associations in 800 whole blood and 326 lymphocyte samples. Moreover, genome-wide analysis of rare missense/loss-of-function variants in association with 6-MP dosing was performed using burden test. Results with FDR < 0.05 were considered significant.
 
Results
In total, 54 haplotypes in Sidra cohort and 12 haplotypes in AUBMC cohort were significantly associated with 6-MP dose %. In Sidra cohort, significant haplotypes included four missense, eight synonymous, and 1290 intronic variants, as well as four known 6-MP-related variants, including rs1800460 and rs12201199 (TPMT), rs3765534 (ABCC4), rs16952570 (FTO). In AUBMC cohort, significant haplotypes included two missense, one synonymous, and 32 intronic variants, as well as a known 6-MP-related variant rs116855232 (NUDT15). Meta-analysis showed two significant haplotypes including two synonymous and 41 intronic variants, and a known 6-MP-related variant rs12201199 (TPMT). Haplotypes including previously known variants had the same direction of change in 6-MP dose as reported in the literature. Interestingly, rare haplotypes were enriched in patients requiring at least 40% 6-MP dose change in both cohorts (Sidra P=1.06x10-6; AUBMC P=0.01), suggesting their role in 6-MP sensitivity/toxicity. Functional characterization of variants within 6-MP-related haplotypes revealed 79 variants with cis-expression alterations in whole blood; among them, 34 were associated with cis-expression alterations in lymphocytes too (FDR<0.05). Genome-wide analysis showed genes related to cell cycle regulation marginally associated with higher 6-MP dose in both cohorts, though underlying mechanisms warrant investigation.

Conclusions
This study identified common and rare haplotypes impacting 6-MP dosing and toxicity in pediatric ALL patients from the Middle East, advancing the potential for personalized therapy in under-studied populations.

Study design