Abstract
Human metabolic health may be defined as the organism´s capability to maintain or regain homeostasis in an everchanging environment, and thus also in response to dietary conditions. To maintain long-term homeostasis, a proper `buffering capacity` with the ability to counteract quickly on metabolic disturbances is needed. I define a metabotype as the observable (and quantifiable) property of the organism to respond to such metabolic disturbance (challenge) in a distinct and reproducible manner. I shall use data from EU-project NutriTech to demonstrate the existence of metabotypes and how robust they are. In NutriTech, 72 healthy volunteers (aged 50-65 years with BMI of 25-35kg/m2) underwent a baseline characterization followed by the assessment of their metabolic responses to a standardized oral glucose tolerance test (OGTT), a mixed-meal tolerance test (MMTT) and a mixed-meal tolerance test with built-in exercise (MMTT+PA). Time-dependent sampling of venous blood before and during the challenges allowed clinical chemistry, and plasma metabolite and protein profiling. Volunteers were then randomized into a weight loss (n=40) or weight maintenance group (n=32) and their metabolic responses to the same challenges were followed a second time after 12 weeks. In addition to blood analysis of around 600 entities (metabolites, hormones, cytokines, etc.), body composition was assessed using MRI at the beginning and the end of the intervention period. Exome sequencing, as well as fecal microbiome analysis by 16S sequencing, were performed as well. Metabolomics employed NMR, GC- and LC-MS/MS to cover a wide range of metabolites. Metabotypes were identified based on the time-courses in a subset of around 20 plasma metabolites as representative indicators of lipolysis, fatty acid ß-oxidation and hepatic ketogenesis that all respond drastically to rising insulin levels during the MMTT or OGTT. Volunteers were clearly separated into 2 clusters with almost equal distribution of men and women within each cluster. Most interestingly, when the effects of the weight loss intervention (mean weight loss: 5.6 kg) were assessed, only one of the clusters revealed a significant improvement, for example, of the glucose profile during the OGTT. This suggests that the two subgroups display different `sensitivity` to metabolic improvements upon a modest weight loss. Amongst the metabolite panel, bile acids (BA) were the most responsive metabolites in the postprandial state and most interesting also when only glucose was provided in the OGTT. However, they also displayed a high level of variability. In analyzing the causes of the diversity of the responses we explored genetic heterogeneity in a subset of around 30 genes/proteins involved in BA synthesis and enterohepatic circulation and also in the gut microbiome sequences. Although the latter did not contribute too much to plasma BA diversity, SNPs in genes were affiliated with BA metabolism and, in particular, absorption in the small intestine showed significant effects. NutriTech provides a wealth of information on human metabolic diversity in response to standardized challenges which asks for more efforts in exploitation. I shall however also address current limitations in the approaches to do better metabolic phenotyping and show why nutrition science needs better standardization.