Communication Dans Un Congrès Année : 2019

Untargeted metabolomics in a prospective cohort to identify diet-related metabolites associated with age-related cognitive decline

Dorrain Y. Low
  • Fonction : Auteur
  • PersonId : 1130981
Sophie Lefèvre-Arbogast
Raúl González-Domínguez
Mireia Urpi-Sarda
Pierre Micheau
Aniko Korosi
Paul J Lucassen
  • Fonction : Auteur
Ludwig. Aigner
  • Fonction : Auteur
  • PersonId : 1049976
Cécile Proust-Lima
Boris P. Hejblum
Catherine Helmer
Cristina Andres-Lacueva
Sandrine Thuret
Cécilia Samieri

Résumé

With a global rise in ageing population and age-associated diseases, understanding how diet modifies cognitive ageing represents key revenues for prevention. In this discovery (D-CogPlast#) study, we aimed to identify a combination of diet-derived metabolites associated with accelerated cognitive decline using untargeted metabolomics. We leveraged the French Three-City cohort of elderly people and using an exploratory approach, designed a nested case-control study contrasting the metabolic profiles of 209 cases of cognitive decline over 13 years against 209 controls (matched for age, gender and educational level) over 12 years following baseline blood draw. Serum samples were profiled using high-resolution UHPLC-QToF. Validated PLS-DA of the baseline serum profiles clearly distinguished between case and control populations. A signature of 22 serum metabolites were associated with cognitive decline, using bootstrap-enhanced least absolute shrinkage and selection operator (LASSO) regression. Seven metabolites derived from the food metabolome included coffee-derived metabolites (atractyligenin glucuronide, cyclo(leucyl-prolyl) and caffeine), a biomarker of citrus intake (proline betaine), 3-carboxy-4-methyl-5-pentyl-2-furanpropionic acid (CMPFP), a cocoa-derived metabolite (cyclo(prolyl-valyl)) and an unknown compound putatively linked to wine intake. The other endogenous metabolites included three acylcarnitines, glycodeoxycholic acid, a glycerophospholipid, trimethyllysine, glucose, cortisol, creatinine and arginine. The 22 metabolite-signature increased the predictive performance of cognitive decline from a cross-validated Area Under the Receiver Operating Curve of 62% (95% CI 56-67%) to 75% (95% CI 70-80%). Our untargeted metabolomics study supports a role for the food metabolome (e.g., metabolites from coffee, citrus fruits, cocoa and possibly fish and wine) and various alterations in endogenous metabolism in cognitive aging

Fichier non déposé

Dates et versions

hal-04796907 , version 1 (21-11-2024)

Identifiants

  • HAL Id : hal-04796907 , version 1

Citer

Dorrain Y. Low, Sophie Lefèvre-Arbogast, Raúl González-Domínguez, Mireia Urpi-Sarda, Pierre Micheau, et al.. Untargeted metabolomics in a prospective cohort to identify diet-related metabolites associated with age-related cognitive decline. 15th annual conference of the metabolomics society. Metabolomics 2019, Jun 2019, La Hague, Netherlands. ⟨hal-04796907⟩
118 Consultations
0 Téléchargements

Partager

  • More