Capsulized Fecal Microbiota Transplantation Induces Remission in Patients with Ulcerative Colitis by Gut Microbial Colonization and Metabolite Regulation

Abstract

Fecal microbiota transplantation (FMT) can induce clinical remission in ulcerative colitis (UC) patients. Enemas, nasoduodenal tubes, and colonoscopies are the most common routes for FMT administration. However, there is a lack of definitive evidence regarding the effectiveness of capsulized FMT treatment in UC patients. In this study, we administered capsulized FMT to 22 patients with active UC to assess the efficiency of capsulized FMT and determine the specific bacteria and metabolite factors associated with the response to clinical remission. Our results showed that the use of capsulized FMT was successful in the treatment of UC patients. Capsulized FMT induced clinical remission and clinical response in 57.1% (12 of 21) and 76.2% (16 of 21) of UC patients, respectively. Gut bacterial richness was increased after FMT in patients who achieved remission. Patients in remission after FMT exhibited enrichment of Alistipes sp. and Odoribacter splanchnicus, along with increased levels of indolelactic acid. Patients who did not achieve remission exhibited enrichment of Escherichia coli and Klebsiella and increased levels of biosynthesis of 12,13-DiHOME (12,13-dihydroxy-9Z-octadecenoic acid) and lipopolysaccharides. Furthermore, we identified a relationship between specific bacteria and metabolites and the induction of remission in patients. These findings may provide new insights into FMT in UC treatment and provide reference information about therapeutic microbial manipulation of FMT to enhance its effects. (This study has been registered at ClinicalTrails.gov under registration no. NCT03426683). IMPORTANCE Fecal microbiota transplantation has been successfully used in patients. Recently, capsulized FMT was reported to induce a response in patients with UC. However, limited patients were enrolled in such studies, and the functional factors of capsulized FMT have not been reported in the remission of patients with UC. In this study, we prospectively recruited patients with UC to receive capsulized FMT. First, we found that capsulized FMT could induce clinical remission in 57.1% of patients and clinical response in 76.2% after 12 weeks, which was more acceptable. Second, we found a relationship between the decrease of opportunistic pathogen and lipopolysaccharide synthesis in patients in remission after capsulized FMT. We also identified an association between specific bacteria and metabolites and remission induction in patients after capsulized FMT. These findings put forward a possibility for patients to receive FMT at home and provide reference information about therapeutic microbial manipulation of FMT to enhance its effects.

Keywords: capsules; fecal microbiota transplantation; gut microbiome; metabolism; microbiome; ulcerative colitis.

FIG 1

Effect of fecal microbiota transplantation on 12-week remission in patients with ulcerative colitis. (A) Trial profile of fecal microbiota transplantation in patients with ulcerative colitis; (B) time points for obtaining fecal samples (triangles), clinical evaluation by Mayo scoring (diamonds), and colonoscopy (circles) during this clinical trial. Shown are changes in endoscopy and hematoxylin-eosin staining (C) and Mayo scores (D) for UC patients before and after treatment with FMT. The parallel line shows the change in Mayo score for individual patients, with a solid line indicative of values for FMT. Each line started at the baseline (W0 [circle]) and finished at its endpoint (W12 [triangle]). The achievement of clinical remission is indicated in green, and nonclinical remission is shown in red. Asterisks in in box plots represent the mean Mayo scores of each group. The significance was determined by paired Wilcoxon rank sum test. FMT, oral capsulized fecal microbiota transplantation.

FIG 2

Alterations in gut bacterial communities of UC patients after FMT. (A) Estimating the α diversity by the richness (observed OTUs and Chao1) and Shannon and Pielou’s evenness. The dashed line indicates significant (P < 0.05) differences between the two groups covered by the line. (B) Differences in gut bacterial community structures among UC patients at the baseline, at other time points during follow-up visits, and donors assessed by principal-component analysis of the Euclidean distance of OTU abundances. Data were subjected to a Hellinger transformation. The top 10 genera were fitted to PCA with a significance of P < 0.05. (C) Relative abundances of the top 15 families in samples obtained from patients in the FMT group. (D) Relative abundances of genera were significantly different between samples at the baseline and after FMT treatment. (E) Heat map of the genera showing significant differences at different times in the FMT group, based on the Z-score-transformed relative abundance of genera at all time points.

FIG 3

Changes in gut bacterial communities associated with Rm and NRm after FMT. (A) α diversity of Rm, NRm, and donor samples, estimated by richness (observed OTUs) and Shannon and Pielou’s evenness. The dashed line indicated significant (P < 0.05) differences between the two groups covered by the line. (B) Differences in gut bacterial community structures of Rm, NRm, and donors between the baseline and other time points during follow-up visits after FMT were assessed by PCA based on the Euclidean distance of OTU abundance. Data were subjected to a Hellinger transformation. (C) The relative abundance of the bacterial genera was uniquely associated with clinical remission. The genera were identified by analyzing significantly different genera between the baseline and other time points in Rm using the paired rank sum test and then removing the genera that were simultaneously significantly different between the baseline and other time points in NRm (W1 and W4) and vice versa to determine the right insert (W12).

FIG 4

Microbial signatures are specifically associated with clinical remission at week 4. Shown are (A) species, (B) KEGG modules, and (C) MetaCyc pathways whose abundance significantly decreased or increased uniquely in W4_Rm compared to the W0_Rm group (namely, not significantly differentiated in W4_NRm versus W0_NRm) during the shotgun metagenomic profiling of fecal samples. The significance was tested using the paired Wilcoxon test. The top 20 signatures of each category indicative of the relative abundance were visualized. Asterisks indicate the mean values of each signature.

FIG 5

Metabolic profiles are associated with clinical remission at the baseline and time points after FMT. (A) PLS_DA of metabolomic profiles in the positive-ion model with regard to time points (left panel) and Euclidean distances of samples within the same time points (right panel). Data were subjected to a Hellinger transformation. Letters indicate groupings with significant differences. (B) Metabolites whose relative abundances were significantly different between the baseline and samples after FMT were filtered by variable importance if the variable importance projection (VIP) score was >1 and were structurally identified in databases. (C) PLS_DA of metabolomic profiles in the positive-ion model at weeks 0 and 4 after FMT after achieving clinical remission or not achieving clinical remission (left panel), as well as Euclidean distances of patients between week 0 and week 4 (right panel). Data were subjected to a Hellinger transformation. (D) Metabolites whose levels were significantly decreased or increased uniquely in W4_Rm, compared with the W0_Rm group (namely, not significantly differentiated in W4_NRm versus W0_NRm), were filtered using a VIP score of >1 and structurally identified in databases. (E) Relative abundance of partial metabolites that changed over time after FMT and were further stratified by clinical remission.

FIG 6

Interomic correlations between gut microbes and metabolites. (A) Procrustes analysis of gut microbial species and serum metabolites; (B) proportion of the total variations in serum metabolomes for different gut microbial species; (C) Pearson correlation-based network analysis of the top 20 gut microbial species, MetaCyc pathways, KEGG modules, and serum metabolites that were uniquely significantly different between W0_Rm and W4_Rm. Species with significant correlations (P < 0.05) were retained. (D) W4_Rm enriched and decreased gut microbiomes (species, genes, and pathways) associated with serum metabolites that contain genes involved in LPS biosynthesis (left panel) and tryptophan metabolism (right panel).