An untargeted metabolomics study of blood pressure: findings from the Bogalusa Heart Study

Abstract

Objective: To identify novel and confirm previously reported metabolites associated with SBP, DBP, and hypertension in a biracial sample of Bogalusa Heart Study (BHS) participants.

Methods: We employed untargeted, ultra-high performance liquid chromatography tandem mass spectroscopy metabolomics profiling among 1249 BHS participants (427 African-Americans and 822 whites) with BP and covariable data collected during the 2013 to 2016 visit cycle. A total of 1202 metabolites were tested for associations with continuous and binary BP phenotypes using multiple linear and logistic regression models, respectively, in overall and race-stratified analyses.

Results: A total of 24 novel metabolites robustly associated with BP, achieving Bonferroni-corrected P less than 4.16 × 10 in the overall analysis and consistent effect sizes across race groups. The identified metabolites included three amino acid and nucleotide metabolites from histidine, pyrimidine, or tryptophan metabolism sub-pathways, seven cofactor and vitamin or xenobiotic metabolites from the ascorbate and aldarate metabolism, bacterial/fungal, chemical, and food component sub-pathways, 10 lipid metabolites from the eicosanoid, phosphatidylcholine, phosphatidylethanolamine, and sphingolipid metabolism sub-pathways, and four still unnamed metabolites. Six previously described metabolites were robustly confirmed by our study (Bonferroni-corrected P < 4.95 × 10 and consistent effect directions across studies). Furthermore, previously reported metabolites for SBP, DBP, and hypertension demonstrated 5.92-fold, 4.77-fold, and 4.54-fold enrichment for nominally significant signals in the BHS (P = 3.08 × 10, 5.93 × 10, and 2.30 × 10, respectively).

Conclusion: In aggregate, our study provides new information about potential molecular mechanisms underlying BP regulation. We also demonstrate reproducibility of findings across studies despite differences in study populations and metabolite profiling methods.

FIGURE 1

Volcano plots depicting the effect sizes and corresponding −log₁₀ P values for the associations of each metabolites with SBP (a), DBP (b), and hypertension (c).

FIGURE 2

Correlations of standardized effect sizes reported by previous studies with those observed in the Bogalusa Heart Study for (a) SBP and (b) DBP (points in red were robustly identified in the Bogalusa Heart Study), as well as (c) enrichment of previously reported SBP, DBP, and hypertension metabolites for nominal significance in the Bogalusa Heart Study.

FIGURE 3

Heat map showing pair-wise Pearson correlation coefficients (r) for novel and previously reported metabolites significant in the Bogalusa Heart Study.