| Abstract: | Increases in snack consumption associated with Westernized lifestyles provide an opportunity to introduce nutritious foods into poor diets. We describe two 10-wk-long open label, single group assignment human studies that measured the effects of two snack prototypes containing fiber preparations from two sustainable and scalable sources; the byproducts remaining after isolation of protein from the endosperm of peas and the vesicular pulp remaining after processing oranges for the manufacture of juices. The normal diets of study participants were supplemented with either a pea- or orange fiber-containing snack. We focused our analysis on quantifying the abundances of genes encoding carbohydrate-active enzymes (CAZymes) (glycoside hydrolases and polysaccharide lyases) in the fecal microbiome, mass spectrometric measurements of glycan structures (glycosidic linkages) in feces, plus aptamer-based assessment of levels of 1,300 plasma proteins reflecting a broad range of physiological functions. Computational methods for feature selection identified treatment-discriminatory changes in CAZyme genes that correlated with alterations in levels of fiber-associated glycosidic linkages; these changes in turn correlated with levels of plasma proteins representing diverse biological functions, including transforming growth factor type β/bone morphogenetic protein-mediated fibrosis, vascular endothelial growth factor-related angiogenesis, P38/MAPK-associated immune cell signaling, and obesity-associated hormonal regulators. The approach used represents a way to connect changes in consumer microbiomes produced by specific fiber types with host responses in the context of varying background diets.Advances in our understanding of the role of the gut microbiome in regulating many aspects of human physiology hold the promise of evolving our view of human nutrition by establishing mechanistic connections between the foods we consume and how they affect health status. One manifestation of this effort is a series of studies, performed on well-phenotyped cohorts, that seek to relate features of gut microbial community composition (organisms, genes), dietary practices, and pre- and postprandial cardiometabolic responses to test meals (1–4). A key question raised by these initiatives relates to the nature of the “bioactive” components of foods. Specifically, what are the nutrients utilized by various gut community members or microbiome-encoded metabolic pathways? What products are produced by biotransformation of these nutrients? How are these products linked to specific host physiologic (or pathophysiologic) processes?Plant-derived dietary fibers represent a “poster child” for these efforts and illustrate the formidable challenges faced. The health benefits of dietary fibers are widely known, as is their inadequate representation in Western diets. However, natural fibers are structurally complex and highly diverse. They contain numerous, typically undefined polysaccharide structures and largely unspecified protein, lipid, and small molecule constituents. Their composition varies as a function of their origin (food staple and cultivar), the different methods employed to recover them from these sources, as well as the different techniques used to incorporate them into processed foods with acceptable organoleptic properties (5). Moreover, analyzing the host effects of metabolism of different fibers is confounded by the fact that there is substantial intra- and interpersonal variation in microbiome configuration (6, 7).Snacking is becoming an ever more dominant feature of daily life worldwide and thus provides an opportunity to introduce nutritious ingredients, such as fibers, into diets. However, obtaining structure-activity relationships for specific fiber types and their corresponding targets in the gut community is foundational for designing snack foods that evoke and/or reinforce microbiome responses that are beneficial to the host.Degradation of dietary polysaccharides is a function primarily performed by bacterial carbohydrate-active enzymes (CAZymes). The gut microbiome harbors tens of thousands of CAZyme genes belonging to at least 136 glycoside hydrolase (GH) and 29 polysaccharide lyase (PL) families [extrapolated and updated from El Kaoutari et al. (8)]. In contrast, the human genome only contains 98 GH and no PL genes (9), of which <20% contribute to the processing of dietary glycans.In the current study, we test the effects of dietary supplementation with two snack food prototypes, one containing pea fiber and the other orange fiber, in two pilot studies of overweight and obese individuals consuming their normal, unrestricted diets. Our strategy was to focus on fiber-associated changes in the abundances of microbial GH and PL genes to determine whether responses to the pea or orange fiber prototypes in the gut microbiome and host are decipherable against a background of varying dietary practices and starting microbiome configurations. Higher order singular value decomposition (10) was utilized as a feature selection tool to identify treatment-discriminating changes in GH and PL gene representation. Mass spectrometric assays of the levels of fecal glycan structures (glycosidic linkages) were subsequently performed and the results were correlated with changes in the abundances of treatment-discriminating GH and PL genes with known or predicted substrate specificities. Our analysis concluded by measuring changes in levels of 1,305 plasma proteins in each study participant as a function of fiber treatment and applying computational tools to identify links between these microbiome and plasma proteome changes in response to fiber consumption. Our results provide an approach, using pilot human studies, for selecting specific fiber preparations, plus informative microbiome and host biomarkers, that can be advanced to proof-of-concept clinical trials which assess their capacity for precise manipulation of microbiome and host features. |
