Genome-Wide Association Study of Metabolites in Pediatric Patients with Newly Diagnosed Acute Myeloid Leukemia Identified Novel Metabolite Quantitative Trait Loci

Acute myeloid leukemia (AML) presents significant treatment challenges due to high variability in patient responses to chemotherapy. Despite advances in treatments and supportive care, the 5-year survival rates for pediatric AML (pAML) remain lower than other childhood hematological cancers. Genetic lesions and cytogenetic features are important in determining prognosis and disease progression. However, the impact of common genetic variants on metabolic pathways and their association with clinical outcomes in pAML is not well understood. Previous metabolomics studies have identified metabolic alterations in hematological malignancies, but few have focused on pAML patients. To that end, we conducted a genome-wide association study (GWAS) on 67 newly diagnosed pAML patients to identify single nucleotide polymorphisms (SNPs) associated with 124 annotated metabolites in serum samples obtained at diagnosis. We also investigated metabolites for association with event-free survival (EFS) and overall survival (OS).

Serum samples were profiled using liquid chromatography/mass spectrometry (LC/MS). Whole-genome genotyping was performed using Illumina's Omni2.5Exome-8 BeadChip array from genomic DNA collected from leukemic blasts before treatment. After GWAS quality control procedures, approximately 1.1 million SNPs were analyzed for association with serum metabolomic profiles. All samples were from pAML patients treated in the multicenter AML02 clinical trial (n=67, NCT00136084). Linear regression was used to model metabolite profiles with SNPs as predictors. The genome-wide significance level was set at Bonferroni-corrected P<5x10^-8 and a suggestive significance level of P<1x10^-5. Cox proportional hazard models examined the association of metabolites with EFS and OS, with significance set at P<.05.

One of the top SNPs significantly associated with the Ursodeoxycholic Acid (UDCA) metabolite is located between long non-coding RNA (lncRNA) LINC02540 and the 5-Hydroxytryptamine Receptor 1B (HTR1B) gene on chromosome 6. The presence of the variant C allele was associated with significantly reduced UDCA (P=3.43x10^-8) and showed an allele-dosage effect, with each additional C allele reducing UDCA levels (P=7.7x10-5). Interestingly, when examining the UDCA metabolomic profile by cytogenetic groups, pAML patients with lysine-specific methyltransferase 2A (KMT2A) gene rearrangements had significantly reduced UDCA levels as compared to all other cytogenetic groups (P=.002). The HTR1B gene encodes the 5-HT1B receptor, a G-protein-coupled receptor implicated in leukemogenesis. Previous reports indicate that serotonin receptor type 1 is differentially expressed on AML cells and implicated in leukemogenesis (PMID:28193998). Another noteworthy SNP is near the importin 8 (IPO8) gene on chromosome 12, associated with reduced nicotinamide levels (P=4.20x10^-8). For each additional copy of the variant G allele, a pronounced reduction in nicotinamide levels was observed (P=1.2x10^-6). Further, when investigating nicotinamide with 10-year survival outcomes, individuals with lower levels of the metabolite experienced worse EFS and OS (EFS: HR=2.59, 95% CI (1.04-6.43), P=.040; OS: HR=4.02, 95% CI (1.11-14.61), P=.035). The IPO8 gene plays a role in nuclear transport, potentially influencing cancer-related cellular processes. Elevated nicotinamide levels are linked to resistance in AML treatments, particularly venetoclax and azacitidine regimens (PMID: 32822582).

This study is among the first GWAS in pAML to identify SNPs associated with metabolite profiles. Our findings indicate that genetic variations in biologically relevant genes impact clinically significant metabolites. Furthermore, top SNPs associated with nicotinamide are linked to EFS and OS. Future research will explore other clinical endpoints, such as risk of relapse and measurable residual disease, with validation of these findings in a larger pAML cohort.