Characterization of diet-dependent metabolic serotypes: proof of principle in female and male rats.

Our research seeks to identify a serum profile, or serotype, that reflects substantial changes in food intake. Earlier studies demonstrated that a number of low-molecular-weight, redox-active compounds of metabolome were sufficiently stable analytically and biologically to identify biomarkers of dietary restriction (DR, restriction of total food intake) in rats. A second initial requirement is to demonstrate feasibility, i.e., that concentration changes in selected serum metabolites can contain sufficient information to classify rats by diet. The current study distinguished 101 (female) and 112 (male) chromatographically identifiable compounds that differ between ad libitum (AL) consumption and DR 6-mo-old rats. In a cohort of female rats, both hierarchical cluster analysis (HCA) and principal component analyses (PCA) could distinguish dietary groups with 100% efficiency (101 metabolites). Repeating the classification studies using the 63 biologically and analytically most robust metabolites decreased noise without affecting categorical separation. In a cohort of male rats, PCA, but not HCA, distinguished the original dietary groups with 100% accuracy (112 metabolites). A subset of 52 of the 112 metabolites enabled both HCA and PCA to group the male rats with 100% accuracy. These data demonstrate that quantitative analysis of selected serum metabolites can yield sufficient information by which to classify the dietary intake of a group of rats, identify such markers chromatographically and set the stage for validation of these metabolic serotypes in independent datasets.