Metabolomics analysis of human plasma metabolites reveals the age- and sex-specific associations

Intrapersonal Stability of Plasma Metabolomic Profiles over 10 Years among Women

In epidemiological studies, samples are often collected long before disease onset or outcome assessment. Understanding the long-term stability of biomarkers measured in these samples is crucial. We estimated within-person stability over 10 years of metabolites and metabolite features (n = 5938) in the Nurses’ Health Study (NHS): the primary dataset included 1880 women with 1184 repeated samples donated 10 years apart while the secondary dataset included 1456 women with 488 repeated samples donated 10 years apart. We quantified plasma metabolomics using two liquid chromatography mass spectrometry platforms (lipids and polar metabolites) at the Broad Institute (Cambridge, MA, USA). Intra-class correlations (ICC) were used to estimate long-term (10 years) within-person stability of metabolites and were calculated as the proportion of the total variability (within-person + between-person) attributable to between-person variability. Within-person variability was estimated among participants who donated two blood samples approximately 10 years apart while between-person variability was estimated among all participants. In the primary dataset, the median ICC was 0.43 (1st quartile (Q1): 0.36; 3rd quartile (Q3): 0.50) among known metabolites and 0.41 (Q1: 0.34; Q3: 0.48) among unknown metabolite features. The three most stable metabolites were N6,N6-dimethyllysine (ICC = 0.82), dimethylguanidino valerate (ICC = 0.72), and N-acetylornithine (ICC = 0.72). The three least stable metabolites were palmitoylethanolamide (ICC = 0.05), ectoine (ICC = 0.09), and trimethylamine-N-oxide (ICC = 0.16). Results in the secondary dataset were similar (Spearman correlation = 0.87) to corresponding results in the primary dataset. Within-person stability over 10 years is reasonable for lipid, lipid-related, and polar metabolites, and varies by metabolite class. Additional studies are required to estimate within-person stability over 10 years of other metabolites groups.

Statistical mediation of the relationships between chronological age and lipoproteins by nonessential amino acids in healthy men

Aging is a major risk factor for metabolic impairment that may lead to age-related diseases such as cardiovascular disease. Different mechanisms that may explain the interplay between aging and lipoproteins, and between aging and low-molecular-weight metabolites (LMWMs), in the metabolic dysregulation associated with age-related diseases have been described separately. Here, we statistically evaluated the possible mediation effects of LMWMs on the relationships between chronological age and lipoprotein concentrations in healthy men ranging from 19 to 75 years of age. Relative and absolute concentrations of LMWMs and lipoproteins, respectively, were assessed by nuclear magnetic resonance (NMR) spectroscopy. Multivariate linear regression and mediation analysis were conducted to explore the associations between age, lipoproteins and LMWMs. The statistical significance of the identified mediation effects was evaluated using the bootstrapping technique, and the identified mediation effects were validated on a publicly available dataset. Chronological age was statistically associated with five lipoprotein classes and subclasses. The mediation analysis showed that serine mediated 24.1% (95% CI: 22.9 – 24.7) of the effect of age on LDL-P, and glutamate mediated 17.9% (95% CI: 17.6 – 18.5) of the effect of age on large LDL-P. In the publicly available data, glutamate mediated the relationship between age and an NMR-derived surrogate of cholesterol. Our results suggest that the age-related increase in LDL particles may be mediated by a decrease in the nonessential amino acid glutamate. Future studies may contribute to a better understanding of the potential biological role of glutamate and LDL particles in aging mechanisms and age-related diseases.

Network Analysis in Systems Epidemiology

Traditional epidemiological studies have identified a number of risk factors for various diseases using regression-based methods that examine the association between an exposure and an outcome (i.e., one-to-one correspondences). One of the major limitations of this approach is the "black-box" aspect of the analysis, in the sense that this approach cannot fully explain complex relationships such as biological pathways. With high-throughput data in current epidemiology, comprehensive analyses are needed. The network approach can help to integrate multi-omics data, visualize their interactions or relationships, and make inferences in the context of biological mechanisms. This review aims to introduce network analysis for systems epidemiology, its procedures, and how to interpret its findings.