Functional Profiling Demonstrates That a Sulfide-Reducing Diet Achieves Microenvironmental Targets in Ulcerative Colitis

Abstract

Background: As a dietary approach to reducing inflammation in ulcerative colitis, the 4-SURE (4 Strategies to Sulfide Reduction) diet was designed to correct pathogenic alterations of excessive protein fermentation and hydrogen sulfide (H2S) production in the distal colon. We aimed to perform a deep functional analysis (microbial and metabolomic) of the feces of 28 adults with mild-moderately active ulcerative colitis who adhered to the 4-SURE diet over 8 weeks to explore whether the 4-SURE diet could modulate the intraluminal environment as intended.

Methods: Fecal samples were collected at week 0 and 8 of dietary intervention, processed and aliquoted. Metagenomic sequencing was undertaken to identify changes in H2S-metabolizing genes, while gas chromatography-mass spectrometry was used to analyze fecal volatile organic compounds and H2S production.

Results: The 4-SURE diet significantly increased alpha diversity between weeks 0 and 8. By random forest plot classifier, the abundance of taxonomic groups comprising known H2S-producing genera were markedly lower at week 8, specifically Odoribacter and Peptostreptococcaceae, and were of highest importance in discriminating between before- and after-diet samples. The capacity for bacterial H2S metabolism was altered with diet, with differences in 12 of 67 analyzed sulfur-metabolizing genes identified. H2S production and indole, a specific marker of protein fermentation, were significantly decreased due to the diet.

Conclusions: Here, we demonstrate that the objectives of the 4-SURE diet were fulfilled. This application of deep functional analysis to a dietary intervention study is novel and highlights an exemplar framework for including microbial and metabolomic biomarkers of pathogenic relevance in the analysis of therapeutic diet strategies. (Australian New Zealand Clinical Trials Registry, Number: ACTRN12619000063112).

Graphical Abstract

Figure 1.

Relative abundance of phylum-level diversity at baseline and end of study. Data shown for 19 participants.

Figure 2.

(A) Alpha-diversity (Shannon diversity index) at weeks 0 and 8, performed at genus taxonomic rank. A significant difference was observed between weeks 0 and 8 (P = .012; Wilcoxon signed rank test). (B) Beta-diversity plot of the microbiome community composition (Centered log-ratio (clr)-transformed) between weeks 0 and 8, performed at genus taxonomic rank. The first and second principal components (PCs) (explaining 13% and 11% of the variance, respectively) are shown.

Figure 3.

Random forest variable importance highlighting the microbial genera that contributed most for microbiome differences before and after the 8-week 4-SURE (4 Strategies to Sulfide Reduction) diet intervention for 19 participants.

Figure 4.

Change in gene abundance from weeks 0 to 8 for the 12 genes with significantly different abundances after the 4-SURE (4 Strategies to Sulfide Reduction) diet. Significance determined with Wilcoxon signed rank test (P < .05) after adjusting for multiple comparisons with Benjamini-Hochberg method. Data shown as difference between week 8 and week 0 counts per million (CPM) for each gene for 19 participants.

Figure 5.

(A-F) Volatile organic compounds that altered in abundance in 28 adults with mild-to-moderately active ulcerative colitis who followed the 4-SURE (4 Strategies to Sulfide Reduction) diet for 8-weeks presented as a log2 fold change (Wilcoxon signed rank test without false discovery rate correction). (A) Indole, (B) undecan-2-one, (C) beta-caryophyllene, (D) alpha-cubebene, (E) 3-methylpentan-2-one, and (F) 3-methyl-1H-indole. A log2 value of below −1 or above 1 represents a 2-fold difference in these volatile organic compounds.

Figure 6.

Change in ex vivo production of fecal hydrogen sulfide in the fecal samples of 25 adults with mild-moderately active ulcerative colitis before and after dietary intervention (Wilcoxon signed rank test; P = .002).

Figure 7.

(A) Gene abundances for genes significantly different between responders (n = 5) and nonresponders (n = 14) for the sseA gene. (B) Gene abundances for genes significantly different between responders (n = 5) and nonresponders (n = 14) for the sqr gene. *Significance determined with Mann-Whitney U test, P = .034 (sseA) and P = .034 (sqr).