Riboflavin Supplementation Promotes Butyrate Production in the Absence of Gross Compositional Changes in the Gut Microbiota

Abstract

Aims: We performed a randomized, placebo-controlled trial, RIBOGUT, to study the effect of 2 weeks supplementation with either 50 or 100 mg/d of riboflavin on (i) Faecalibacterium prausnitzii abundance, (ii) gut microbiota composition, (iii) short-chain fatty acid (SCFA) profiles, and (iv) the satiety and gut hormones. Results: Neither dose of riboflavin, analyzed separately, impacted the abundance of F. prausnitzii, and only minor differences in SCFA concentrations were observed. However, combining the results of the 50 and 100 mg/d groups showed a significant increase in butyrate production. While the gut bacterial diversity was not affected by riboflavin supplementation, the complexity and stability of the bacterial network were enhanced. Oral glucose tolerance tests showed a trend of increased plasma insulin concentration and GLP-1 after 100 mg/d supplementation. Innovation: Dietary supplements, such as vitamins, promote health by either directly targeting host physiology or indirectly via gut microbiota modulation. Here, we show for the first time that riboflavin intervention changes the activity of the microbiota. The butyrate production increased after intervention and although the composition did not change significantly, the network of microbial interactions was enforced. Conclusion: This RIBOGUT study suggests that oral riboflavin supplementation promotes butyrate production in the absence of major shifts in gut microbiota composition. ClinicalTrials.gov Identifier: NCT02929459.

Keywords: Faecalibacterium prausnitzii; bacterial networks; butyrate (short-chain fatty acids); insulin; microbiota; riboflavin.

Conflict of interest statement

This clinical RIBOGUT trial was sponsored by DSM Nutritional Products AG. H.J.M.H. and G.D. received research funding from DSM.

Figures

FIG. 1.

Study design for the riboflavin intervention for N = 105 participants (A), the oGTT substudy in 36 participants (B), and flowchart for the participant inclusion (C). oGTT, oral glucose tolerance test.

FIG. 2.

The results of Faecalibacterium quantification based on FISH with probe Fprau645 and 16S rRNA gene sequencing. (A–C) The abundances of F. prausnitzii from FISH (A), the relative abundances (B), and 16S rRNA gene sequencing (C) were not changed over all time points among three groups (p > 0.05 for all groups, Kruskal–Wallis test). (D) The relative abundances of ASVs from 16S rRNA gene sequencing, which corresponded to the probe Fprau645, were not changed. ASV, Amplicon Sequence Variant; FISH, fluorescence in situ hybridization.

FIG. 3.

PCoA of the distances between taxonomical composition based on the Bray-Curtis (A) and weighted-Unifrac (B) over all time points of all three groups. Each dot represents a sample separated by color code and shape as indicated in the legend. PCoA, principal coordinates analysis.

FIG. 4.

PCoA of individual's taxonomical composition based on the Bray-Curtis distance between T2 and T3. (A–C), samples' taxonomical composition in each group compared between T2 and T3; (D–F), paired samples' taxonomical composition in each group compared between T2 and T3. Each dot represents a sample, and the dashed lines indicate the fecal samples from the same individual.

FIG. 5.

Boxplots of SCFAs concentrations of paired fecal samples significantly increased after 2 weeks of riboflavin supplementation in pairwise comparison. (A) Butyrate concentration increased significantly in Ribo100 group (p = 0.05, Wilcoxon signed-rank test); (B) acetate concentration increased significantly in Ribo50 group (p = 0.02, Wilcoxon signed-rank test); (C) no significant change was observed with propionate concentration. Each dot represents a sample, and horizontal bar indicates significant differences. SCFA, short-chain fatty acids.

FIG. 6.

Fecal butyrate concentration was increased in combining data of two doses riboflavin intervention groups. (A) Comparison within placebo and RiboCom groups (Mann–Whitney U test). (B) Pairwise comparison of samples T2 and T3 within placebo and RiboCom groups (Wilcoxon signed-rank test). Horizontal bar indicates adjusted p-values.

FIG. 7.

Gut bacterial networks over all time points. (A) Visualization of constructed MENs of placebo (top) and RiboCom groups (bottom) from T1 to T4. Top 5 large modules are shown in different colors, and smaller modules are shown in gray. Each network is shown based on the Pearson correlations (RMT-threshold 0.67, FDR adjusted p < 0.05) between the abundances of bacterial ASVs of 36 placebo samples and 69 RiboCom samples. N, nodes; L, links. (B–E), links (B), nodes (C), average degree (avgK) (D), and average path distance (GD) (E) are topological features of gut microbial networks. The baselines of T1 and T2 were calculated as means ± SE, and the values of T3 and T4 networks shown as dots. MEN, molecular ecological network; RMT, random matrix theory.

FIG. 8.

Boxplots of riboflavin concentrations in feces and plasma samples. (A) Riboflavin concentration in fecal samples was significantly increased after intervention at T3 (p < 0.05, Kruskal–Wallis test). (B) Riboflavin plasma levels of the oGTT groups significantly increased at T3 in both riboflavin supplementation groups (p < 0.05, Mann–Whitney U test). No differences in the placebo group were found in either fecal or plasma samples.

FIG. 9.

Oral glucose tolerance test results of plasma glucose (A, B) and insulin concentration (C, D) at T2 before riboflavin supplementation (A, C) and T3 after riboflavin supplementation (B, D).