Protective function of sclerosing cholangitis on IBD

Abstract

Objective: There is a strong clinical association between IBD and primary sclerosing cholangitis (PSC), a chronic disease of the liver characterised by biliary inflammation that leads to strictures and fibrosis. Approximately 60%-80% of people with PSC will also develop IBD (PSC-IBD). One hypothesis explaining this association would be that PSC drives IBD. Therefore, our aim was to test this hypothesis and to decipher the underlying mechanism.

Design: Colitis severity was analysed in experimental mouse models of colitis and sclerosing cholangitis, and people with IBD and PSC-IBD. Foxp3⁺ Treg-cell infiltration was assessed by qPCR and flow cytometry. Microbiota profiling was carried out from faecal samples of people with IBD, PSC-IBD and mouse models recapitulating these diseases. Faecal microbiota samples collected from people with IBD and PSC-IBD were transplanted into germ-free mice followed by colitis induction.

Results: We show that sclerosing cholangitis attenuated IBD in mouse models. Mechanistically, sclerosing cholangitis causes an altered intestinal microbiota composition, which promotes Foxp3⁺ Treg-cell expansion, and thereby protects against IBD. Accordingly, sclerosing cholangitis promotes IBD in the absence of Foxp3⁺ Treg cells. Furthermore, people with PSC-IBD have an increased Foxp3⁺ expression in the colon and an overall milder IBD severity. Finally, by transplanting faecal microbiota into gnotobiotic mice, we showed that the intestinal microbiota of people with PSC protects against colitis.

Conclusion: This study shows that PSC attenuates IBD and provides a comprehensive insight into the mechanisms involved in this effect.

Keywords: COLONIC MICROFLORA; Cholangitis; INFLAMMATORY BOWEL DISEASE; PRIMARY SCLEROSING CHOLANGITIS.

Figure 1

Spontaneous colitis is reduced in mice with concomitant experimental primary sclerosing cholangitis in Il10 ^−/− Mdr2 ^−/− mice. (A) Graphical breeding scheme for generation of Il10 ^−/− and Il10 ^−/− Mdr2 ^−/− littermates. Mice were bred under specific pathogen-free (SPF) conditions in the local mouse facility (MB1). After weening, litters were separated with respect to their genotype. At an age of 12 weeks, (B) body weight (n=22 Il10 ^−/− , n=16 Il10 ^−/− Mdr2 ^−/− ) and (C) colon inflammation was assessed by mouse colonoscopy (n=25 Il10^−/− , n=13 Il10^−/−Mdr2^−/− ), as described in material and methods. (D, E) Flow cytometry analysis of colon infiltrating CD4⁺ T-cell (n=17 Il10^−/− , n=13 Il10^−/−Mdr2^−/− ) and Foxp3⁺ Treg-cell frequencies of 12 weeks old mice (n=12 Il10^−/− , n=10 Il10^−/−Mdr2^−/− ). (F) Graphical breeding scheme for generation of Il10^−/− and Il10^−/−Mdr2^−/− littermates bred in the presence of a colitogenic SPF microbiome (MB2) containing Helicobacter hepaticus, that was transferred to the founding animals. After weening, litters were separated with respect to their genotype. At the age of 12 weeks (G) body weight (n=8 Il10^−/− , n=13 Il10^−/−Mdr2^−/− ), (H) colonoscopy (n=8 Il10^−/− , n=12 Il10^−/−Mdr2^−/− ) and (I, J) frequencies of colon infiltrating CD4⁺ T cells and Foxp3⁺ Treg cells (n=6 Il10^−/− , n=11 Il10^−/−Mdr2^−/− ) were analysed. For statistical analysis, Mann-Whitney U test was performed.

Figure 2

Spontaneous colitis is reduced in Il10^−/− mice with concomitant 3,5-diethoxycarbonyl-1,4-dihydrocollidine (DDC)-mediated liver cholestasis. (A) Graphical scheme of the experimental setup. At an age of 6–8 weeks Il10^−/− mice were gavaged with MB2. Four weeks after reconstitution, liver cholestasis was induced by 0.1% DDC feeding supplemented into the normal chow diet. After 8 days, (B) colonic inflammation was analysed by mouse colonoscopy (n=22 mice per group). (C, D) On day 9, mice were sacrificed and frequencies of colon infiltrating CD4⁺ T cells and Foxp3⁺ Treg cells were analysed using flow cytometry (8=mice per group). For statistical analysis Mann-Whitney U test was performed.

Figure 3

Increased colitis manifestation in Rag1^−/−Mdr2^−/− mice after Foxp3⁻CD45RB^high T-cell transfer. (A) Graphical scheme of the experimental setup. (B) At an age of 8–10 weeks Rag1^−/− and Rag1^−/−Mdr2^−/− mice were gavaged with MB2. After 4 weeks of reconstitution, colitis was induced on transfer of flow cytometry sorted Foxp3⁻CD45RB^high CD4⁺ T cells, isolated from Foxp3-RFP reporter mice. After 13 days of T-cell reconstitution, (B) weight loss and (C) colonic inflammation by colonoscopy were analysed (n=13 Rag1^−/− , n=12 Rag1^−/−Mdr2^−/− ). (D, E) At day 14, mice were sacrificed and frequencies of colon infiltrating CD4⁺ T cells and Foxp3⁺ Treg cells were analysed by flow cytometry in one of three experiments (n=4 Rag1^−/− n=4 Rag1^−/−Mdr2^−/− ). For statistical analysis Mann-Whitney U test was performed.

Figure 4

Reduced colitis severity in germ-free wild-type mice after transfer of Mdr2^−/− microbiota. (A) Graphical scheme of the experimental procedure. In brief, faecal microbiota obtained from wild-type and Mdr2^−/− mice, harbouring MB2 microbiome, was gavaged into germ-free wild-type mice. One day later, colitis was induced in these mice by intraperitoneal injection of 0.25 mg anti-IL10Rα antibody per mouse two times a week. After 13 days of colitis induction, (B) weight loss was determined and (C) colonic inflammation was analysed by colonoscopy (n=9 WT-FMT, n=10 Mdr2^−/− -FMT). FMT, faecal microbiota transplantation; WT, wild-type.

Figure 5

FOXP3 mRNA expression and endoscopic IBD scoring reveal reduced clinical manifestation of IBD in people with primary sclerosing cholangitis (PSC-IBD). Description of a cohort including 29 people with Crohn’s disease (CD), 22 with Ulcerative colitis (UC) and 41 with PSC-IBD. (A) FOXP3 mRNA expression levels were analysed from intestinal biopsies taken from the terminal ileum, ascending and descending colon and sigma/rectum from every person. (B) IBD severity was determined based on CDAI (persons with CD) and Mayo score (all other persons). Both scores were merged into a unified IBD score (healthy/remission: 0, mild: 1, moderate: 2, severe: 3 points). (C) FOXP3 mRNA expression levels were analysed from intestinal biopsies taken from the terminal ileum, ascending and descending colon and sigma/rectum from every person with clinically active disease. To test for significance MLEM, post hoc Dunnett test was used for (A and C). Fisher’s exact test was used for (B).

Figure 6

Colitis severity in germ-free mice is attenuated after FMT from people with primary sclerosing cholangitis and colitis (PSC-IBD), enriched for genera of the Lachnospiraceae family. Microbiota profiling was performed on mucosal tissue samples of our IBD and PSC-IBD cohort, as described in the material and methods. (A) PCoA of Bray-Curtis dissimilarities shows beta diversity across people with IBD and PSC-IBD. (B) Genera with significantly different abundance between people with IBD and PSC-IBD. (C) Graphical scheme of the protocol for faecal microbiota transplantation of stool derived from IBD or PSC-IBD patients into germ-free wild-type mice, and subsequent DSS colitis induction. After 9 days of colitis induction, (D) weight loss was determined and (E and F) colonic inflammation was analysed by colonoscopy (each dot represents one mouse). IBD activity of the donor is shown as: remission (black), mild (green), moderate (blue) and severe (red). (G and H) Microbiota profiling from stool samples collected from mice after reconstitution with stool samples from our IBD and PSC-IBD cohort. (D–H) n=21 mice transplanted with IBD stool; n=21 mice transplanted with PSC-IBD stool were used in four independent experiments.