Skip to main content
ARS Home » Pacific West Area » Davis, California » Western Human Nutrition Research Center » Obesity and Metabolism Research » Research » Publications at this Location » Publication #302606

Title: Habitual physical activity and plasma metabolomics patterns distinguish individuals with low- versus high-weight loss during controlled energy restriction

item PICCOLO, BRIAN - University Of California
item Van Loan, Marta
item Gertz, Erik
item Woodhouse, Leslie
item SOUZA, ELAINE - University Of California
item Gustafson, Mary
item CAMPBELL, CAITLIN - University Of California
item FIEHN, OLIVER - University Of California
item Keim, Nancy
item Adams, Sean
item Newman, John

Submitted to: Journal of Nutrition
Publication Type: Peer Reviewed Journal
Publication Acceptance Date: 12/30/2014
Publication Date: 1/29/2015
Citation: Piccolo, B., Van Loan, M.D., Gertz, E.R., Woodhouse, L.R., Souza, E., Gustafson, M.B., Campbell, C., Fiehn, O., Keim, N.L., Adams, S.H., Newman, J.W. 2015. Habitual physical activity and plasma metabolomics patterns distinguish individuals with low- versus high-weight loss during controlled energy restriction. Journal of Nutrition. 145(4):681-690. doi: 10.3945/jn.114.201574.

Interpretive Summary: Weight loss caused by reducing energy intake is highly variable even in tightly-controlled conditions, and knowledge of factors that could predict a persons’ ability to respond to caloric restriction would be a valuable weight management tool in the clinic, and provide insight into the fundamental biology regulating body weight. To search for variables that co-vary with changes in body weight, blood was collected and analyzed from 67 overweight and obese adults enrolled in a 12-week controlled feeding intervention with a 500 kcal/day energy deficit. Multivariate statistics were used to generate predictive models of body weight changes using pre- and post-intervention measures, including the plasma metabolites, endocrine and inflammatory markers, clinical indices, body composition, diet, and physical activity measurements. Correlation networks were used to visualize inter-relationships among variables. Starting with 900 variables, a reduced model with only 127 variables was generated which accurately predicted weight loss (root mean error of prediction (RMSEP) = 1.85 kg; Q2 = 0.43) and explained 73% of the variability in the entire data set. Individuals with the greater weight loss had more light and moderate physical activity and less sedentary physical activity before the intervention started. High weight losers also had lower adiposity and respiratory exchange ratios at the end of the study, while fasting plasma fatty acids, 3-hydroxybuterate and citrate concentrations were increased, suggesting greater levels of fat metabolism. Alterations in catabolic intermediates of leucine and threonine also suggested enhanced amino acids oxidation in high weight losers. Unfortunately, models using only pre-intervention variables did not accurately predict total weight loss. In conclusion, we found that person-to-person differences in habitual physical activity are accompanied by distinct blood metabolite signatures that correlate with weight loss in overweight and obese individuals under controlled feeding conditions. The differential impacts of sedentary and light activity on metabolism may drive this association.

Technical Abstract: Weight loss (WL) induced by energy restriction is highly variable even in controlled clinical trials. An integrative analysis of the plasma metabolome coupled to traditional clinical variables may reveal a WL “responder” phenotype. Therfore, we predicted WL in overweight and obese individuals on an energy-restricted diet in a controlled feeding trial using the plasma metabolome in addition to clinical indices, body composition parameters, plasma endocrine and inflammatory markers, dietary parameters, and physical activity (PA) measurements. A total of 940 pre-intervention (PrI) and post-intervention (PoI) variables were assessed. Partial least squares regression accurately predicted weight loss along latent variable 1 (root mean error of prediction (RMSEP) = 1.81 kgs; Q2 = 0.43) and revealed that PrI light and moderate PA were associated with high weight loss, whereas PoI sedentary PA and 2-ketoisocaproic acid were associated with low weight loss. Network correlation analysis revealed a direct connection between lower PrI respiratory quotient, higher PoI 3-hydroxybutanoic acid and higher PoI citric acid (P < 0.00001) in high weight losers. Our results suggest that habitual light-to-moderate PA are accompanied by distinct blood amino acid and lipid metabolic signatures and correlated with WL in overweight and obese individuals under controlled conditions.