A Genome-Wide Association Study of Serum Choline and Related-Metabolites

Objectives: Common single nucleotide polymorphisms (SNPs) in one-carbon metabolism have been shown to influence the dietary requirement for choline in humans. The aim of the study was to identify SNPs associated with serum choline levels, as well as circulating choline-to-betaine and betaine-to-dimethylglycine ratios. Methods: Data from the Collaborative Study of Genes, Nutrients and Metabolites (CSGNM, N=2,402) was used to model the individual association of serum choline (μmol/L) with 680,975 SNPs, using linear regression under an additive genetic model. Extracted DNA was genotyped using the Illumina HumanOmni1-Quad Beadchip. Additive models adjusting for age, sex, and selected principal components (PC), as well as age, sex, selected PC, vitamins B12, B9, and B6, and holotranscobalamin were constructed. A p-value of < 5.0 X 10-5 was considered significant. Affected genes were identified on the NCBI dbSNP website. Results: Twenty-four SNPs were associated with serum choline concentration in the age, sex, and PC adjusted model; SNPs were predominantly intronic variants and included nearby genes with functional links to choline, including the recently described mitochondrial choline transporter, SLC25A48. Several nearby genes without current links to choline metabolism included lysophospholipase like 1 (LYPLAL1) and phosphotyrosine interaction domain containing 1 (PID1). Adjustment for additional one-carbon metabolic covariates yielded 6 additional affected genes. Notably, many of the significant SNPs lie in gene regions thought to play a role in aspects of neurological development and neurodegenerative disease. Conclusions: Our results identify potentially novel genetic determinants of serum choline in need of replication in a follow-up cohort, as well as targets for follow-up analyses to determine their mechanistic relationship to choline homeostasis.