Identification Of Dietary Biomarkers For Meat And Fish Intake Through Metabolomics

William CHEUNG, International Agency for Research on Cancer, France
ASSI N. 1 , ROTHWELL J. 1 , KESKI-RAKHONEN P. 1 , BRENNAN L. 3 , SCALBERT A. 1 , FROST G. 2 , SLIMANI N. 1 , FERRARI P. 1 , GIBBONS H. 3

1 International Agency for Research on Cancer, Nutrition and Metabolism Section, Lyon, France.
2 Division of Endocrinology and Metabolism, Nutrition and Dietetic Research Group, Imperial College London, London, United Kingdom
3 Institute of Food and Health, University College Dublin, Dublin, Republic of Ireland

Several biomarkers of meat or fish intake have been proposed but their specificity for different types of meat or fish has seldom been considered. In this work a metabolomic approach was used to identify dietary biomarkers for three different types of meat and fish in a dietary intervention study and in the European Prospective Investigation on Cancer and nutrition (EPIC) cohort study.
24-Hour urine samples were first collected from four groups of subjects (n=10 in each group) after consumption of meals containing three different doses of either chicken, red meat (beef), processed meat (cooked ham) or fish (hake). Urine samples were analysed by high resolution mass spectrometry (MS) coupled with Ultra-High Performance Liquid Chromatography. After correction for multiple testing, 230 MS features were found to be significantly associated with the intake of either one of these four foods and showed a monotonic increase with the dose. The same MS features were analysed in 24-hour urine samples collected in EPIC subjects. Exclusive consumers of one of the same 4 foods (based on 24-hour dietary recalls) were compared to subjects who did not consume any meat or fish during the 24 hours of urine collection (n=10 per group). Sixty-five MS features out of the 230 previously identified were able to distinguish consumers of either of the 3 meats or fish when assessed by Receiver Operating Curve analysis with permutation testing. The identities of 8 of them were finally confirmed based on their mass fragmentation spectra and comparison with authentic standards: anserine, carnosine, acetylcarnitine, propionylcarnitine, 1-methylhistidine, 3-methylhistidine, trimethylamine-N-oxide, and 2-methylbutyroylcarnitine. These 8 biomarkers used alone or in combination, may provide accurate measurements of chicken, red meat, ham and fish intake, and will be further validated in a larger set of EPIC samples.