Serum Metabolomics of Activity Energy Expenditure and its Relation to Metabolic Syndrome and Obesity

Abstract

Modifiable lifestyle factors, including exercise and activity energy expenditure (AEE), may attenuate the unfavorable health effects of obesity, such as risk factors of metabolic syndrome (MetS). However, the underlying mechanisms are not clear. In this study we sought to investigate whether the metabolite profiles of MetS and adiposity assessed by body mass index (BMI) and central obesity are inversely correlated with AEE and physical activity. We studied 35 men and 47 women, aged 30-60 years, using doubly labeled water to derive AEE and the Sedentary Time and Activity Reporting Questionnaire (STAR-Q) to determine the time spent in moderate and vigorous physical activity. Proton nuclear magnetic resonance spectroscopy was used for serum metabolomics analysis. Serine and glycine were found in lower concentrations in participants with more MetS risk factors and greater adiposity. However, serine and glycine concentrations were higher with increasing activity measures. Metabolic pathway analysis and recent literature suggests that the lower serine and glycine concentrations in the overweight/obese state could be a consequence of serine entering de novo sphingolipid synthesis. Taken together, higher levels of AEE and physical activity may play a crucial part in improving metabolic health in men and women with and without MetS risk factors.

Figure 1

Supervised OPLS-DA score scatter plots and loadings plots showing the separation between MetS_any (circles) and MetS_zero (dots) for (A)women and (B) men. Every dot/circle represents one participant. The score scatter plot (left) and loading plot (right) are superimposable and indicate which (VIP > 1) metabolites associate with which MetS group in women (R² = 0.42, Q² = 0.30) and in men (R² = 0.31, Q² = 0.12), respectively.

Figure 2

Supervised OPLS-DA score scatter plots and loadings plots for men, showing AEE_DLW and physical activity variables stratified for MetS_any (A–D) and MetS_zero (E–F). Every dot/circle represents one participant, with circles indicating participants with the highest 50% (circles) and the lowest 50% (dots) of each variable. The score scatter plots (left) and loading plots (right) are superimposable and show (VIP > 1) metabolites. Only variables that resulted in models are shown. (A) AEE_DLW (R² = 0.67, Q² = 0.36), (B) AEE/kg_DLW (R² = 0.34, Q² = 0.24), (C) PAL_DLW (R² = 0.86, Q² = 0.46), (D) moderate physical activity (R² = 0.38, Q² = 0.21) for MetS_any and (E) moderate physical activity (R² = 0.62, Q² = 0.51) and (F) vigorous physical activity (R² = 0.57, Q² = 0.25) for MetS_zero.

Figure 3

Supervised OPLS-DA score scatter plots and loadings plots for women, showing AEE_DLW and physical activity variables stratified for MetS_any (A–D) and MetS_zero (E–I). Every dot/circle represents one participant, with circles indicating participants with the highest 50% (circles) and the lowest 50% (dots) of each variable. The score scatter plots (left) and loading plots (right) are superimposable and show VIP > 1 metabolites. Only variables that resulted in models could be shown. The variables presented include (A) AEE_DLW (R² = 0.45, Q² = 0.27), (B) AEE/kg_DLW (R² = 0.33, Q² = 0.11), (C) PAL_DLW (R² = 0.44, Q² = 0.26), (D) vigorous physical activity (R² = 0.43, Q² = 0.22), for MetS_any and (E) AEE_DLW (R² = 0.43, Q² = 0.37), (F) AEE/kg_DLW (R² = 0.32, Q² = 0.21), (G) PAL_DLW (R² = 0.40, Q² = 0.33), (H) moderate physical activity (R² = 0.28, Q² = 0.14) and (I) vigorous physical activity (R² = 0.28, Q² = 0.16) for MetS_zero.

Figure 4

Serum serine and arginine concentrations for MetS_any (black bars) and MetS_zero women (white bars) with high (H) or low (L) levels of AEE/kg_DLW and PAL_DLW, and AEE_DLW respectively. Significant difference compared to MetS_zero-H (**p < 0.01 ***p < 0.001), MetS_zero-L (†p < 0.05) and MetS_any-H (‡p < 0.05) is indicated. Serum concentrations were normalized to the total sum for each sample to assure normal distribution and comparability across samples.