Metabolomics applications to nutritional intervention studies

Abstract

Plant phytochemicals might not be essential throughout life, but they may promote health and well-being by preventing, for example, certain non-communicable diseases. Among phytochemicals, flavan-3-ols are particularly relevant, as they are extensively consumed within diet, being characteristic compounds of tea, cocoa, wine, pome fruits, berries, and nuts. The existing literature shows a clear evidence that flavan-3-ols exert beneficial effects on health, however, discrepancies of results have been highlighted when encouraging results obtained using observational or animal studies are compared with more wary results from in vivo human intervention studies. The reason for this inconsistency may lay in the profound structural modifications that these compounds undergo once in contact with the human body, and in the inter-individual variability described in the capacity of our body to transform and absorb them. So, assessment of flavan-3-ol bioavailability and associated inter-individual variability are crucial factors to understand to what extent and in which forms they are available for the individual internal compartments, a fundamental information to unravel the biological effects of these compounds in human health. This information is then completely lacking for cranberry flavan-3-ols. Among dietary sources rich in phytochemicals, coffee is particularly relevant as it is one of the most consumed drinks worldwide. In many epidemiological studies, regular coffee consumption has been associated with a reduced risk of several non-communicable diseases, but no association has yet been found between circulating and excreted coffee bioactives and derived metabolites and physiological responses, making the mechanisms through which coffee exerts its potential preventive effects still widely undisclosed. In this frame, metabolomics may aid in achieving insights into such open questions, thanks to its several applications in the field of nutrition. The global objective of this doctoral thesis was to apply metabolomics to nutritional intervention studies to pursue several aims. Firstly, to assess the bioavailability of flavan-3-ols from cranberry through the analysis of their main colonic metabolites (phenyl-γ-valerolactones and phenyl-valeric acids), also considering their potential as biomarkers of flavan-3-ol intake and inter-individual variability in their appearance in plasma and urine. Secondly, to elucidate the presence of different metabotypes (metabolic phenotypes) in the urinary excretion of flavan-3-ol colonic metabolites as a strategy to manage the associated inter-individual variability, also assessing the impact of the statistical technique used in the process. Lastly, to study the changes occurring at metabolomic level following consumption of different patterns of coffee and the impact on known metabolic pathways. Plasma and/or urine samples collected in previously conducted human intervention trials were analysed trough LC-MS based metabolomics approaches, either targeted or untargeted. Specifically, a targeted analysis of all the potential conjugated forms of phenyl-γ-valerolactones and phenylvaleric acid revealed that sulfate and glucuronide conjugates of 5-(3′,4′-dihydroxyphenyl)-γ-valerolactone are the main circulating and excreted metabolites after cranberry flavan-3-ols intake, reaching maximum plasma concentration at about 4–6 h post consumption. These metabolites show a clear dose-response relationship with the amount of cranberry flavan-3-ols consumed, supporting their suitability as biomarkers of cranberry flavan-3-ol intake. Glucuronidation seems to be the most favored phase II conjugation after cranberry juice intake, despite wide differences among individuals. A high inter-individual variability is also reported in circulating and urinary metabolite levels, with some subjects displaying a greater efficiency in metabolizing flavan-3-ols and producing phenyl-γ-valerolactones. On the basis of this evidence and based on the results of a preliminary set of experiments, where a different urinary production of phenyl-γ-valerolactones and 3-(hydroxyphenyl)propanoic acids upon green tea flavan-3-ol consumption led to the elucidation of three putative metabotypes, a targeted analysis of these metabolites revealed different profiles in their urinary excretion upon cranberry consumption in two diverse experimental settings. A methodological workflow for metabotype definition and validation, based on broadly accessible multivariate and univariate tools, has been proposed, highlighting the importance of data pre-treatment and clustering methods on the final outcomes when analysing datasets of flavan-3-ol metabolites. Non-transformed, centered, and UV-scaled data has been key to unravel metabolic patterns based on colonic metabolism. Cluster analysis based on k-means and a final consensus algorithm led to quantitative-based models, whereby the distribution of the clusters was due to the amount of metabolites excreted (high vs. low). Differently, the expectation-maximization algorithm and clustering according to principal component scores yield metabotypes characterized by quali-quantitative differences in the excretion of colonic metabolites. PLS-DA, together with univariate analyses, served to validate the urinary metabotypes in the production of flavan-3-ol metabolites and to confirm the robustness of the methodological approach. This metabotyping strategy may be key to manage the inter-individual variability reported in the colonic metabolism of flavan-3-ols and to further investigate its consequences in the impact on the observed health effects attributed to this class of compounds. Finally, the use of an untargeted approach has allowed the detection of changes in the metabolome due to the intake of different patterns of coffee. Indeed, besides metabolites specifically derived from coffee intake, endogenous metabolites have been detected whose levels were modulated by the different patterns of coffee consumed. New metabolic pathways have therefore been identified to be modulated by the intake of different patterns of coffee, such as the metabolism and biosynthesis of specific amino acids, that may in turn potentially influence human health. In conclusion, the application of metabolomics approaches to nutrition intervention studies previously conducted has offered the possibility to gain new and further insights in the research field of dietary bioactives, namely the bioavailability of flavan-3-ols and the associated inter-individual variability, the elucidation of a metabotyping strategy to manage it, and the metabolic routes by which coffee and coffee-related metabolites may exert effects on human health, opening the door to new hypotheses about the health effects of coffee consumption and the underlying mechanisms.