Abstract P452: Pathway Analysis Of One-carbon Metabolism In Association With Acute Myocardial Infarction In Patients With Stable Angina Pectoris

Background: One-carbon metabolism (OCM) is involved in RNA, DNA synthesis and in epigenetic regulation. Pathway gene analysis is more likely to capture potential gene-disease associations comparing to single gene analysis. We investigated whether genetic variants in folate network as well as their interactions would predict subsequent acute myocardial infarction (AMI) in patients undergoing elective coronary angiography for stable angina pectoris (SAP). Methods: A total of 2381 subjects (98% Caucasian) recruited to the Western Norway B Vitamin Intervention Trial (WENBIT) in 2000-2004 were studied. Patients were followed until a subsequent AMI or end of 2006. We genotyped 18 commonly investigated single nucleotide polymorphisms (SNPs) in 13 genes in the OCM network. Predictable variants were selected based on both out-of-bag importance scores from random survival forests and their incremental effects from recursive nested analysis. We then assessed gene-gene interactions among those predictive variants with generalized logistic model and calculated hazard ratio (HR), confidence interval (CI) for each predictor in Cox proportional hazard models. Results: There are 204 (8.57%) patients who experienced an AMI throughout 4.9 years (median) follow-up. We observed no violation from Hardy-Weinberg equilibrium in 15 SNPs in our population. Four SNPs (rs1076991, rs1021737, rs1801133 and rs3733890) in folate network were identified as promising candidates for predicting AMI risk. Interestingly, only rs1076991 associated significantly with AMI (HR=1.30, 95% CI: 1.10-1.52), indicating the other SNPs primarily mediate their risk bygene-gene interactions. There was a strong and stable interaction between the dominant model of rs3733890 and dominant model of rs34489327 (P for interaction: 0.003; HR =2.23, 95% CI: 1.27-3.92, adjusted for age, gender, smoking, diabetes mellitus, hypertension, extent of CAD by angiography (0-3), apolipoprotein (Apo) A1 and B, left ventricular ejection fraction and estimated glomerular filtration rate). Further adjustment for vitamins involved in OCM (folate, B2, B6, B12), the substrates (betaine and homocysteine) and products (methionine and dimethylglycine) of the betaine--homocysteine S-methyltransferase (BHMT) had minimal influence on the estimate. Conclusions: Prior studies show that BHMT transcription is associated with co-transcription of Apo B and that the rs3733890 of this gene is associated with higher dimethylglycine levels. We have recently shown that elevated plasma dimethylglycine is associated with increased risk of AMI and improves risk prediction in patients with SAP independent of Apo B levels. The current investigation extends previous knowledge, suggesting that a genetic variant in BHMT may modulate the risk of AMI by interfering with thymidylate synthesis independent of plasma dimethylglycine.