Plasma lipidome variation during the second half of the human lifespan is associated with age and sex but minimally with BMI

Abstract

Recent advances in mass spectrometry-based techniques have inspired research into lipidomics, a subfield of '-omics', which aims to identify and quantify large numbers of lipids in biological extracts. Although lipidomics is becoming increasingly popular as a screening tool for understanding disease mechanisms, it is largely unknown how the lipidome naturally varies by age and sex in healthy individuals. We aimed to identify cross-sectional associations of the human lipidome with 'physiological' ageing, using plasma from 100 subjects with an apolipoprotein E (APOE) E3/E3 genotype, and aged between 56 to 100 years. Untargeted analysis was performed by liquid chromatography coupled-mass spectrometry (LC-MS/MS) and data processing using LipidSearch software. Regression analyses confirmed a strong negative association of age with the levels of various lipid, which was stronger in males than females. Sex-related differences include higher LDL-C, HDL-C, total cholesterol, particular sphingomyelins (SM), and docosahexaenoic acid (DHA)-containing phospholipid levels in females. Surprisingly, we found a minimal relationship between lipid levels and body mass index (BMI). In conclusion, our results suggest substantial age and sex-related variation in the plasma lipidome of healthy individuals during the second half of the human lifespan. In particular, globally low levels of blood lipids in the 'oldest old' subjects over 95 years could signify a unique lipidome associated with extreme longevity.

Fig 1. Lipid-lowering medications and plasma lipids.

Effect of lipid-lowering medication usage on a) concentrations (mmol) of cholesterol, LDL-C and HDL-C, triglycerides and HLR (*p<0.05, Mann-Whitney U test) and (b) on normalised lipid abundances for lipid classes.

Fig 2. Correlation matrices of traditional lipid measures and lipid classes.

(a) Correlation matrix of cholesterol, LDL-C, HDL-C and triglyceride levels with LC-MS measured lipid classes; numbers show 2-digit rounded correlation values; dendrograms represent the hierarchical clustering of lipid classes according to their correlation measures. All correlations above r = 0.30 are considered significant at the p = 0.05 level. (b) Correlation matrix of lipid classes with each other ordered by hierarchical clustering to group together the correlated lipid classes. Heatmap scale represents correlation strength, with red and blue for positive and negative correlations respectively. All correlations above r = 0.30 or below -0.30 are considered significant at the p = 0.05 level.

Fig 3. Boxplots of lipid class normalised abundance across age groups.

The 95+ group had significantly lower abundance compared to younger age groups (Kruskall Wallis test and pairwise Mann Whitney U-tests, p<0.05) for all lipid classes with the exception of DG lipids and subclasses.

Fig 4. Boxplots of lipid normalised abundances for specific sphingomyelins and phospholipids.

Note phospholipids presented were taken from subjects aged over 75 years by sex; p-values derived based on Mann-Whitney U test.