Inflammatory bowel disease is associated with extraintestinal diseases such as primary sclerosing cholangitis in the liver. Interestingly, it is known that an imbalance between Foxp3+ regulatory T cells (Treg) and Th17 cells is involved in inflammatory bowel disease and also in primary sclerosing cholangitis. To explain these associations, one hypothesis is that intestinal inflammation and barrier defects promote liver disease because of the influx of bacteria and inflammatory cells to the liver. However, whether and how this is linked to the Treg and Th17 cell imbalance is unclear. To address this, we used dextran sodium sulfate (DSS) and T cell transfer colitis mouse models. We analyzed the pathological conditions of the intestine and liver on histological, cellular, and molecular levels. We observed bacterial translocation and an influx of inflammatory cells, in particular Th17 cells, to the liver during colitis. In the DSS colitis model, in which Treg were concomitantly increased in the liver, we did not observe an overt pathological condition of the liver. In contrast, the T cell–mediated colitis model, in which Treg are not abundant, was associated with marked liver inflammation and a pathological condition. Of note, upon depletion of Treg in DEREG mice, DSS colitis promotes accumulation of Th17 cells and a pathological condition of the liver. Finally, we studied immune cell migration using KAEDE mice and found that some of these cells had migrated directly from the inflamed intestine into the liver. Overall, these data indicate that colitis can promote a pathological condition of the liver and highlight an important role of Treg in controlling colitis-associated liver inflammation.

This article is featured in In This Issue, p.3477

The gut–liver axis is of clinical relevance in several liver diseases. Recent publications suggest that gut microbes leak from the colon to the liver because of barrier defects during colitis. Accordingly, it was proposed that the liver may act as a functional vascular firewall that clears commensals that have penetrated the intestinal vascular circuits (1). Furthermore, bacterial translocation has been shown to promote immune cell activation in the liver.

In line with this observation, there is a strong association between inflammatory bowel disease (IBD) and primary sclerosing cholangitis (PSC), a cholestatic liver disease that is characterized by inflammation and fibrosis of the bile ducts (2). Thus, ∼75% of all PSC patients show concomitant IBD (3, 4). Moreover, there are similarities between IBD and PSC. Both diseases are characterized by a fecal dysbiosis marked by reduced bacterial diversity (5). Furthermore, the Th17–Foxp3+ regulatory T cell (Treg) balance is also affected in PSC patients (6, 7), and this is known to play a key role in the regulation of intestinal inflammation. Indeed, it was shown that single nucleotide polymorphisms in the IL2RA gene, which is a key gene for Foxp3+ Treg, is associated with both IBD and PSC (8, 9). Accordingly, PSC patients show a reduced number and impaired function of Foxp3+ Treg, which is again associated with IL2RA gene polymorphisms (7, 10). Likewise, PSC patients demonstrate more Th17 cells in the periductular areas of the liver. Moreover, PBMCs from PSC patients show an increased Th17 cell response upon pathogen stimulation (6).

Despite the known associations, whether inflammatory cells migrate to the liver directly from the intestine remains to be tested. More importantly, it is unclear whether and under which conditions colitis induces a pathological condition of the liver. The aim of our study was to address these questions. To this end, we used two complementary colitis models: 1) the acute and chronic dextran sodium sulfate (DSS) colitis model, in which the oral administration of DSS in the drinking water causes intestinal barrier defects and, consequently, colitis in immune competent mice (11); and 2) the transfer colitis model, in which the transfer of differentiated Th17 cells into lymphopenic hosts promotes severe T cell–driven colitis (12). Additionally, we used depletion of Treg (DEREG) mice (13), in which Foxp3+ Treg can be depleted to assess the role of Treg. Finally, we used the KAEDE transgenic mice (14), in which UV light administration in the colon lumen leads to photoconversion of the exposed cell, allowing us to track the migration of these cells into the liver.

We found that colitis leads to increased infiltration of Th17 cells into the liver and that they migrated in part from the colon into the liver. However, numbers of Foxp3+ Treg were also increasing in wild type mice with colitis, which did not develop an overt pathological condition of the liver. On the contrary, transfer of Th17 cells into Rag1−/− mice that lacked Foxp3+ Treg caused severe colitis, increased infiltration of Th17 cells into the liver, and a concomitant pathological condition of the liver. Accordingly, wild type mice with DSS colitis developed a pathological condition of the liver upon depletion of Foxp3+ Treg. Thus, our data highlight the importance of the Th17–Treg axis in an IBD-associated pathological condition of the liver.

C57BL/6, Rag1−/−, KAEDE (14), and Foxp3–red fluorescent protein (RFP) (15), Il17a-eGFP (16), Ifng-FP635 (12), Il17a-FP635 (12) reporter mice, and DEREG mice (13) were bred and housed under specific pathogen-free conditions at the University Hospital Hamburg-Eppendorf animal facility. Age- (12–14 wk) and sex-matched littermates were used for all experiments. Animal experiments were approved by the local ethics committee (G12/37, G13/035, and G17/12).

Mice received drinking water supplemented with 3% DSS for 7 d, followed by 2 d of pure drinking water in the absence of DSS to induce acute DSS colitis (DSS m.w.: 36.000–50.000; MP Biomedicals, Illkirch, France). Then, the mice were sacrificed and analyzed for pathological conditions of the intestine and liver. Mice received three cycles of drinking water supplemented with 2.5% DSS for 5 d, followed by 16 d of regular water to induce chronic DSS colitis. The third DSS cycle was 7 d long. Two days after each DSS cycle, colitis development was monitored by endoscopy. Three days after the third cycle, mice were sacrificed and analyzed.

Two hundred fifty nanograms per mouse of diphtheria toxin was injected i.p. every 3 d, starting at day −1 before DSS administration.

Lymphocytes were isolated from spleen and lymph nodes of Foxp3 RFP × Il17a eGFP × Ifng FP635 reporter mice. Naive CD4+CD25CD44 T cells were enriched by depletion of CD25+ and CD44+ cells, followed by enrichment of CD4+ T cells using MACS according to the manufacturer’s instructions (Miltenyi Biotec, Bergisch-Gladbach, Germany). The purity of CD4+ T cells obtained was ∼90% as determined by flow cytometry. Naive CD4+ T cells were cultured in the presence of plate-bound 10 μg/ml anti-CD3 mAb, 10 μg/ml soluble anti-CD28 mAb, 100 U/ml murine IL (mIL)-2, 10 ng/ml mIL-12 (both PeproTech, Hamburg, Germany), and 1 μg/ml anti–IL-4 mAb (clone: 11B11) to differentiate Th1 cells. Naive CD4+ T cells were cultured in the presence of 3 μg/ml soluble anti-CD3 mAb, 1 μg/ml soluble anti-CD28 mAb, 10 μg/ml anti–IL-4 mAb, 10 μg/ml anti–IFN-γ mAb (clone: XMG), 10 ng/ml mIL-6, 20 ng/ml mIL-23 (all BioLegend, Koblenz, Germany), 10 ng/ml mIL-1-β (R&D Systems, Abingdon, UK), and APCs at a ratio of 4:1 (APCs to T cells) to differentiate Th17 cells. All Abs used for differentiation were azide free. After 4 d, cells were harvested and FACS sorted for Foxp3 monomeric RFP (mRFP) Ifng FP635+ Th1 or Foxp3 mRFPIl17a eGFP+ Th17 cells, using a FACSAria III (BD Biosciences, Heidelberg, Germany).

To induce T cell transfer colitis, 2.5 × 104 Th1 or Th17 cells were injected i.p. into Rag1−/− mice. Mice were monitored by endoscopy weekly for the development of intestinal inflammation starting 3 wk after T cell transfer. To assess whether T cells migrate from the colon into the liver during transfer colitis, 1 × 105 FACS-sorted CD4+ CD44 CD25 T cells from spleens of untreated KAEDE mice were i.p. injected into Rag1−/− mice. Mice were monitored for colitis development weekly. Upon colitis development KAEDE green cells in the colon were photoconverted. Two days later, mice were sacrificed, and the accumulation of KAEDE red cells in the liver and in the colon was assessed.

Colonoscopy was performed at the indicated time points to monitor severity of intestinal inflammation as described before (17), using the Coloview System (Karl Storz, Germany). In brief, anesthetized mice were endoscopically scored based on five parameters: thickening of the colon, changes in vascular pattern, granularity of the mucosal surface, stool consistency, and visible fibrin, each graded 1–3, resulting in an overall score between 0 (healthy) and 15 (severe colitis).

Mice were anesthetized with 4% (vol/vol) isoflurane delivered in 5 L/min 100% O2. The Coloview endoscope was connected to a Dymax LED UV laser. For green-to-red photoconversion of colonic cells, the endoscope was inserted into the colon of the KAEDE mice, and the tissue was exposed to UV light for 1 min three times starting at the proximal site of the rectum. After exposure, the endoscope was immediately moved 0.5–1 cm distally and the exposure was started again.

Cells were preincubated with 2.4G2 mAb to block FcγR. Cells were stained with CD45.2 (clone: 104), CD11b (clone: M1/70), CD11c (clone: HL3), Ly6C (clone: HL1.4), Ly6G (clone: 1A8), NK1.1 (clone: PK136), or CD3 (clone: 17A2) for 20 min at 4°C to detect cell surface molecules. Unless otherwise specified, mAbs were purchased from BioLegend (London, England). Intracellular cytokine detection was performed by stimulating purified cells with 50 ng/ml PMA (Sigma-Aldrich), 1 mM ionomycin (Sigma-Aldrich) for 4 h at 37°C. Monensin was added for the last 2 h of incubation. Cells were stained for cell surface markers CD45.2 (clone: 104) and CD4 (clone: GK1.5) for 20 min at 4°C, fixed with 4% formalin for 30 min, and permeabilized with 0.1% Nonidet P-40 for 4 min both at room temperature. For intracellular detection of IFN-γ (clone: XMG1.2), IL-17A (clone: TC11-18H10.1), or Foxp3 (clone: JES5-16E3; eBioscience), cells were incubated with fluorochrome-conjugated mAb for 1 h at room temperature or overnight at 4°C, respectively. Cells were analyzed using an LSR II or Fortessa instrument (BD Biosciences) and FlowJo software (Tree Star, Ashland, OR). Gating strategies for innate immune cells and T cell subsets are shown in Supplemental Fig. 1.

RNA was extracted using TRIzol reagent, and cDNA was synthesized by the High Capacity cDNA Reverse Transcription Kit according to the manufacturer’s protocols (Life Technologies, Darmstadt, Germany). The real-time PCR system (Step One Plus; Life Technologies) was used for quantitative PCR. The primer probes for Il6 (Mm0044619_m1), Il17a (Mm00439618_m1), Tnfa (Mm00443260_g1), and Ifng (Mm01168134_m1) were purchased from Applied Biosystems. Hypoxanthine phosphoribosyltransferase (Hprt) (Mm01545399_m1) was used as an internal reference. Data are shown as fold change of mRNA expression using ΔΔ threshold cycle (Ct) method.

Organs were removed and fixed in 4% formalin solution for 24 h. Embedding of organs in paraffin, sectioning, and H&E staining were performed at the University Hospital Hamburg-Eppendorf Department of Pathology. Infiltration of inflammatory cells and tissue destruction of the colon were each graded 0–3 as described previously (18) to evaluate severity of colonic disease. To assess the pathological condition of the liver, the modified histological activation index (mHAI) was used and sections were graded as follows: interface hepatitis (0–4), confluent necrosis (0–6), focal lytic necrosis, apoptosis and focal inflammation (0–4), and portal inflammation (0–4) (19).

To monitor liver damage, aspartate aminotransferase (AST), alanine aminotransferase, and γ glutamyl transferase (GGT) were analyzed in blood serum by the University Medical Center Hamburg-Eppendorf, Institute of Clinical Chemistry, and Central Laboratories.

Liver tissue was removed under sterile conditions, dissected mechanically, and mixed with 3 ml of sterile PBS. One milliliter of cell suspension each was inoculated into aerobic and anaerobic blood culture vials (BD BACTEC PEDS Plus/BD BACTEC Plus anerobic). Blood cultures were monitored for bacterial growth for 5 d (BD BACTEC system, Heidelberg, Germany). To detect microbial 16S/18S rDNA, 1 ml of the suspension was analyzed using the Universal Microbe Detection-Kit according to the manufacturer’s protocols (Molzym, Bremen, Germany).

Bile duct ligation (BDL) and sham operations were carried out under anesthesia (isoflurane 1.5%) as previously described (20). Prior to operation, mice were injected i.p. with 5 mg/kg carprofen, followed by a midline abdominal incision and isolation and occlusion of the common bile duct. In sham-operated controls, the bile duct was mobilized but not ligated. On day 6 after BDL, operation mice were infected orally with 1010 CFU of bioluminescent Listeria monocytogenes (Xen32) by oral gavage.

Mice were injected with 0.1 mg/kg body weight of LPS i.p. every 24 h for up to 5 d as previously described (21). One day after the last LPS injection, mice were infected orally with 1010 CFU of bioluminescent L. monocytogenes (Xen32) by oral gavage.

Statistical analysis was performed using GraphPad Prism Software (GraphPad Software, San Diego, CA). The nonparametric two-sided Mann–Whitney U test was used to compare the groups. Bonferroni correction was used to counteract the problem in case of multiple comparisons. To gain time-dependent weight-loss data, a repeated-measures ANOVA was used to assess the significance of the main effects and an experimental group–time interaction. The significance level α was set to 0.05. The statistical test used is included in each figure legend.

To assess the role of intestinal inflammation on pathological conditions of the liver, we used the DSS-induced colitis model. As expected, mice developed severe colitis based on weight loss, endoscopic findings, colon length, and leukocyte infiltration upon oral administration of DSS in the drinking water (Supplemental Fig. 2A). The peak of disease was between days 9 and 11, followed by a recovery phase. At the disease peak, gene expression of the proinflammatory cytokines Il6 and Tnfa, Ifng, and Il17a was increased in the colon, as were Th1, Th17, and Foxp3+ Treg numbers and frequencies (Supplemental Figs. 1, 2B, 2C). Interestingly, following similar timing, we also found an upregulation of Il6, Tnfa, and Il17a gene expression in the liver accompanied with increased accumulation of Foxp3+ Treg and Th17 cells (Fig. 1A, 1B), albeit at lower levels compared with the colon. Because granulocytes and monocytes/macrophages play an important role as first-line defense against pathogens and promote T cell recruitment to inflammatory sides, we analyzed accumulation of these cells in the liver during the course of DSS treatment. We observed a significant increase in CD11b+ Ly6G+ granulocytes and CD11b+CD11cLy6C+ monocytes/macrophages in the livers of these mice (Fig. 1C, Supplemental Fig 1). These findings prompted us to analyze pathological conditions of the liver in wild type mice with established DSS colitis in more detail (Fig. 2). Ten days after induction of DSS colitis, infiltrates of CD45+ hematopoietic cells increased in the liver (Fig. 2A). However, we did not detect increased liver transaminase/GGT levels and histological signs of liver disease (Fig. 2B, 2C). Taken together, these data indicate that acute DSS colitis is associated with increased infiltration of Th17 cells, Foxp3+ Treg, monocytes, and granulocytes in the liver. However, this is not accompanied with pathological conditions of the liver based on serum transaminase levels and histological findings per se.

FIGURE 1.

Proinflammatory cytokine expression and cellular infiltration into the liver during the course of acute DSS colitis. Acute DSS colitis was induced in 12- to 14-wk-old C57BL/6 mice by addition of 3% DSS to the drinking water for 7 d, followed by pure drinking water until the day of analysis. Before the start of DSS treatment (day 0) and at each indicated time point, five mice were sacrificed and pathological conditions of the liver were analyzed. (A) Expression of Il6, Tnfa, Ifng, and Il17a mRNA extracted from liver tissues normalized to control gene Hprt. Fold change of mRNA levels, compared with day 0, were calculated for the indicated time points. At each time point, five mice were analyzed. (B) Frequencies and numbers per gram of liver infiltrating CD45+CD4+ Foxp3+ Treg, IFN-γ+ IL-17A (Th1) cells, IFN-γ IL-17A+ (Th17) cells, and IFN-γ+ IL-17A+ (Th1 + Th17) cells, determined by flow cytometry after restimulation with PMA/ionomycin. (C) Frequencies and numbers of CD45+CD3 Ly6G+ neutrophils and CD45+CD3Ly6GCD11c CD11b+Ly6C+ monocytes/macrophages. Data show mean of five mice ± SEM per time point. For statistical analysis, nonparametric one-way ANOVA was performed. Indicated significant values refer to day 0. *p ≤ 0.05, **p ≤ 0.01, ***p ≤ 0.001.

FIGURE 1.

Proinflammatory cytokine expression and cellular infiltration into the liver during the course of acute DSS colitis. Acute DSS colitis was induced in 12- to 14-wk-old C57BL/6 mice by addition of 3% DSS to the drinking water for 7 d, followed by pure drinking water until the day of analysis. Before the start of DSS treatment (day 0) and at each indicated time point, five mice were sacrificed and pathological conditions of the liver were analyzed. (A) Expression of Il6, Tnfa, Ifng, and Il17a mRNA extracted from liver tissues normalized to control gene Hprt. Fold change of mRNA levels, compared with day 0, were calculated for the indicated time points. At each time point, five mice were analyzed. (B) Frequencies and numbers per gram of liver infiltrating CD45+CD4+ Foxp3+ Treg, IFN-γ+ IL-17A (Th1) cells, IFN-γ IL-17A+ (Th17) cells, and IFN-γ+ IL-17A+ (Th1 + Th17) cells, determined by flow cytometry after restimulation with PMA/ionomycin. (C) Frequencies and numbers of CD45+CD3 Ly6G+ neutrophils and CD45+CD3Ly6GCD11c CD11b+Ly6C+ monocytes/macrophages. Data show mean of five mice ± SEM per time point. For statistical analysis, nonparametric one-way ANOVA was performed. Indicated significant values refer to day 0. *p ≤ 0.05, **p ≤ 0.01, ***p ≤ 0.001.

Close modal
FIGURE 2.

Absence of a pathological condition of the liver in acute DSS colitis. Acute DSS colitis was induced in 11- to 12-wk-old C57BL/6 mice. Ten days after colitis induction, mice were sacrificed, and pathological conditions of the liver were assessed by (A) CD45+ leukocyte infiltration, (B) serum AST, alanine aminotransferase, and GGT levels from five control mice and eight DSS-treated mice, (C) H&E staining and mean mHAI ± SEM from four mice per group (see 2Materials and Methods for a definition of mHAI). Representative histological liver sections are shown. Scale bars, 100 μm. Cutouts in the top left corner show an overview of the total histological slide. Arrows mark mild cellular infiltrations. Significant liver injury was neither detected in untreated mice nor in mice after DSS treatment. Data from two independent experiments are summarized. For statistical analysis, Mann–Whitney U test was performed. *p ≤ 0.05.

FIGURE 2.

Absence of a pathological condition of the liver in acute DSS colitis. Acute DSS colitis was induced in 11- to 12-wk-old C57BL/6 mice. Ten days after colitis induction, mice were sacrificed, and pathological conditions of the liver were assessed by (A) CD45+ leukocyte infiltration, (B) serum AST, alanine aminotransferase, and GGT levels from five control mice and eight DSS-treated mice, (C) H&E staining and mean mHAI ± SEM from four mice per group (see 2Materials and Methods for a definition of mHAI). Representative histological liver sections are shown. Scale bars, 100 μm. Cutouts in the top left corner show an overview of the total histological slide. Arrows mark mild cellular infiltrations. Significant liver injury was neither detected in untreated mice nor in mice after DSS treatment. Data from two independent experiments are summarized. For statistical analysis, Mann–Whitney U test was performed. *p ≤ 0.05.

Close modal

Thus, we next asked whether chronic inflammation of the colon might affect pathological conditions of the liver. To this end, we induced chronic DSS colitis by repetitive administration of DSS in the drinking water in wild type mice. Chronic colitis was characterized by weight loss, shortening of the colon, histological and endoscopic signs of colitis, and upregulation of proinflammatory gene expression, such as Il6 and Tnfa, in addition to accumulation of immune cells in the colon (Supplemental Fig. 3A–E). In line with the results obtained in the acute DSS colitis model, we observed CD45+ hematopoietic cell infiltrates in the liver (Fig. 3A). In addition, we observed increased Il6 and Tnfa gene expression in the liver of these mice (Fig. 3B). However, there were no signs of a significant pathological condition of the liver based on histological findings and transaminase/GGT levels (Fig. 3C). In some mice in the DSS-treated groups, mild cellular infiltration was observed. We thus performed further analyses of single-cell suspensions using flow cytometry. The overall infiltration of Th17 cells, Foxp3+ Treg (Fig. 3D), CD11b+Ly6G+ granulocytes, and CD11b+CD11cLy6C+ monocytes/macrophages (Fig. 3E) was significantly increased in the liver after induction of chronic DSS colitis. Additionally, we screened mice with or without colitis for the presence of bacteria and bacterial products in the liver using culture of liver homogenates and PCR analysis and found that almost 40% of mice in the acute and chronic DSS colitis models tested positive for the presence of bacteria, compared with ∼20% of mice without colitis (Table I). To further test increased bacterial translocation, we cultured portal vein blood from mice after treatment with DSS and LPS and after BDL. We could indeed detect bacteria in portal vein blood cultures but not in hepatic vein blood cultures from DSS- and LPS-treated mice (Fig. 4A). BDL, in contrast, resulted in bacterial accumulation in both portal vein and hepatic vein blood cultures. Furthermore, oral infection with bioluminescent Listeria (Xen32) prior to DSS treatment resulted in a specific accumulation of Listeria in the liver of DSS-treated but not untreated control mice, whereas accumulation of Listeria in spleen was not affected by DSS treatment (Fig. 4B), further suggesting that the liver may act as a functional vascular firewall that clears commensals leaking from the colon to the liver, as suggested by a recent publication (1).

FIGURE 3.

Increased infiltration of granulocytes and Th17 cells in the liver during chronic DSS colitis. Chronic DSS colitis was induced in 12- to 14-wk-old C57BL/6 mice by giving three cycles 2.5% DSS followed by pure drinking water as described in 2Materials and Methods. Three days after the third DSS cycle, pathological conditions of the liver were determined by (A) CD45+ leukocyte infiltration, (B) gene expression of proinflammatory cytokines Il6 and Tnfa in liver tissues normalized to control Hprt, (C) AST, GGT levels, and H&E staining from liver tissues of four control and eight DSS treated mice. Scale bars, 100 μm. (D) Frequencies and absolute cell numbers of liver infiltrating CD45+CD4+Foxp3+ Treg, IFN-γ+IL-17A (Th1) cells, IFN-γIL-17A+ (Th17) cells, and IFN-γ+IL-17A+ (Th1 + Th17) determined by flow cytometry after restimulation with PMA/ionomycin, and (E) frequencies and numbers of CD45+CD3 Ly6G+ neutrophils and CD45+CD3 CD11b+Ly6C+ monocytes/macrophages. Data show mean ± SEM from eight control mice (A and C–E) and 15 (A) or 18 (C–E) DSS-treated mice obtained in four independent experiments. For statistical analysis, Mann–Whitney U test was performed. *p ≤ 0.05, **p ≤ 0.01, ***p ≤ 0.001.

FIGURE 3.

Increased infiltration of granulocytes and Th17 cells in the liver during chronic DSS colitis. Chronic DSS colitis was induced in 12- to 14-wk-old C57BL/6 mice by giving three cycles 2.5% DSS followed by pure drinking water as described in 2Materials and Methods. Three days after the third DSS cycle, pathological conditions of the liver were determined by (A) CD45+ leukocyte infiltration, (B) gene expression of proinflammatory cytokines Il6 and Tnfa in liver tissues normalized to control Hprt, (C) AST, GGT levels, and H&E staining from liver tissues of four control and eight DSS treated mice. Scale bars, 100 μm. (D) Frequencies and absolute cell numbers of liver infiltrating CD45+CD4+Foxp3+ Treg, IFN-γ+IL-17A (Th1) cells, IFN-γIL-17A+ (Th17) cells, and IFN-γ+IL-17A+ (Th1 + Th17) determined by flow cytometry after restimulation with PMA/ionomycin, and (E) frequencies and numbers of CD45+CD3 Ly6G+ neutrophils and CD45+CD3 CD11b+Ly6C+ monocytes/macrophages. Data show mean ± SEM from eight control mice (A and C–E) and 15 (A) or 18 (C–E) DSS-treated mice obtained in four independent experiments. For statistical analysis, Mann–Whitney U test was performed. *p ≤ 0.05, **p ≤ 0.01, ***p ≤ 0.001.

Close modal
Table I.
Bacteriology in the liver after induction of DSS colitis
No. of Positive MiceAnerobic BacteriaAerobic Bacteria16S PCR < Ct3518S PCR < Ct35
Control 3/16 2/16 0/16 1/16 0/16 
Acute DSS 2/5 2/5 1/5 0/2 0/5 
Chronic DSS 9/24 4/24 3/24 6/24 0/14 
No. of Positive MiceAnerobic BacteriaAerobic Bacteria16S PCR < Ct3518S PCR < Ct35
Control 3/16 2/16 0/16 1/16 0/16 
Acute DSS 2/5 2/5 1/5 0/2 0/5 
Chronic DSS 9/24 4/24 3/24 6/24 0/14 
FIGURE 4.

Acute colitis and BDL induce translocation of gut microbiota into the liver. (A) Mice were infected orally with 1010 CFU of bioluminescent Listeria monocytogenes (Xen32) by oral gavage after BDL operation (left), DSS treatment (middle), and LPS treatment (right). After 24 h, bacterial counts (CFU) in portal and hepatic venous blood were analyzed. (B) The bacterial loads with Listeria (CFU/organ) in the liver (left) and spleen (right) of DSS treated and control mice are shown. Symbols represent individual mouse and mean ± SEM from at least four mice per group obtained in two independent experiments. *p ≤ 0.05.

FIGURE 4.

Acute colitis and BDL induce translocation of gut microbiota into the liver. (A) Mice were infected orally with 1010 CFU of bioluminescent Listeria monocytogenes (Xen32) by oral gavage after BDL operation (left), DSS treatment (middle), and LPS treatment (right). After 24 h, bacterial counts (CFU) in portal and hepatic venous blood were analyzed. (B) The bacterial loads with Listeria (CFU/organ) in the liver (left) and spleen (right) of DSS treated and control mice are shown. Symbols represent individual mouse and mean ± SEM from at least four mice per group obtained in two independent experiments. *p ≤ 0.05.

Close modal

Collectively, we found increased numbers of granulocytes, monocytes/macrophages, and Th17 cells in the liver of wild type mice during acute and chronic DSS colitis, which was accompanied by bacterial translocation. However, infiltration of these cells was not associated with an overt pathological condition of the liver. Concomitantly, Foxp3+ Treg numbers were also increased in the liver of these mice. We therefore hypothesized that Foxp3+ Treg may protect mice against pathological conditions of the liver during DSS colitis.

To test the hypothesis that Foxp3+ Treg protect mice against pathological conditions of the liver during DSS colitis, we induced T cell–mediated colitis in lymphopenic Rag1−/− mice (CD45.1) by transfer of differentiated Th17 cells (CD45.2) (12). To differentiate Th17 cells, we used naive CD4+ T cells isolated from Ifng FP635 × Il17a eGFP × Foxp3 mRFP triple reporter mice (15, 16), which allowed us to sort and transfer a pure Th17 cell population without contamination of Foxp3+ Treg. As control, we injected PBS or differentiated Th1 cells, which also repopulate recipient Rag1−/− mice but induced only mild colitis (Supplemental Fig. 4A). Transfer of Th17 cells but not Th1 cells caused moderate colitis as assessed by weight loss, endoscopic findings, and histological findings (Supplemental Fig. 4A, 4B). Furthermore, we found an upregulation of the proinflammatory cytokines, Il6 and Tnfa compared with PBS control mice (Supplemental Fig. 4C). We also found an accumulation of Th17 cells in the colon. In addition, we observed Th1 cells in the colon that had been differentiated from the injected Th17 cells (Supplemental Fig. 4D). Interestingly, upon colitis induction via transfer of Th17 cell into Rag1−/− mice, we detected increased AST levels, histological signs of hepatitis (Fig. 5A), upregulation of Il6 and Tnfa mRNA expression (Fig. 5B), and increased accumulation of leukocytes (Fig. 5C). We also found a significant accumulation of Th17 cells and Th1 cells in the liver after Th17 cell transfer (Fig. 5D). In contrast to our data from the DSS colitis model, we did not find significantly increased bacterial translocation to the liver in diseased mice in the transfer colitis model compared with healthy mice (data not shown). However, the colitis was less severe in this model compared with the DSS colitis model, which may explain this difference. Interestingly, transfer of Th1 cells caused mild colitis and not pathological conditions of the liver, although the Rag1−/− mice were also repopulated with T cells (Fig. 5A–D). These data indicate that transfer of effector T cells into Rag1−/− mice does not induce pathological conditions of the liver per se. However, further studies are warranted to clarify this lack of pathogenicity of Th1 cells. In conclusion, transfer of Foxp3-negative Th17 into Rag1−/− causes colitis and concomitant liver disease, suggesting that colitis and/or Th17 cells might be involved in the liver phenotype of these mice.

FIGURE 5.

Induction of hepatitis after induction of T cell–transfer colitis. Th1 and Th17 cells were differentiated from naive CD45.2+ CD4+ T cells isolated from Foxp3 mRFP × Il17a eGFP × Ifng FP635 reporter mice as described in 2Materials and Methods. After 5 d of differentiation, Foxp3 RFPIfng FP635+ (Th1) or Foxp3Il17a eGFP+ (Th17) cells were sorted and injected into 11- to 17-wk-old congenic CD45.1+Rag1−/− mice. Six weeks after T cell transfer, mice were sacrificed, and (A) pathological conditions of the liver were determined by AST levels and mHAI. H&E sections show representative staining for each group, with asterisks showing cellular infiltration (see 2Materials and Methods for a definition of mHAI). Scale bars, 200 μm. (B) Il6 and Tnfa gene expression in liver tissues normalized to Hprt. (C) Infiltration of total CD45+ leukocytes, CD45.1+ recipient leukocytes, and CD45.2+ donor cells. (D) FACS analyses of the frequencies and number of liver infiltrating CD45+CD4+ Foxp3+ Treg, IFN-γ+ IL-17A (Th1) cells, IFN-γ IL-17A+ (Th17) cells, and IFN-γ+ IL-17A+ (Th1 + Th17) cells after restimulation with PMA/ionomycin. Data show mean of 10 mice ± SEM per group obtained from two independent experiments. For statistical analysis, nonparametric one-way ANOVA was performed. Indicated significant values refer to day 0. *p ≤ 0.05, **p ≤ 0.01, ***p ≤ 0.001.

FIGURE 5.

Induction of hepatitis after induction of T cell–transfer colitis. Th1 and Th17 cells were differentiated from naive CD45.2+ CD4+ T cells isolated from Foxp3 mRFP × Il17a eGFP × Ifng FP635 reporter mice as described in 2Materials and Methods. After 5 d of differentiation, Foxp3 RFPIfng FP635+ (Th1) or Foxp3Il17a eGFP+ (Th17) cells were sorted and injected into 11- to 17-wk-old congenic CD45.1+Rag1−/− mice. Six weeks after T cell transfer, mice were sacrificed, and (A) pathological conditions of the liver were determined by AST levels and mHAI. H&E sections show representative staining for each group, with asterisks showing cellular infiltration (see 2Materials and Methods for a definition of mHAI). Scale bars, 200 μm. (B) Il6 and Tnfa gene expression in liver tissues normalized to Hprt. (C) Infiltration of total CD45+ leukocytes, CD45.1+ recipient leukocytes, and CD45.2+ donor cells. (D) FACS analyses of the frequencies and number of liver infiltrating CD45+CD4+ Foxp3+ Treg, IFN-γ+ IL-17A (Th1) cells, IFN-γ IL-17A+ (Th17) cells, and IFN-γ+ IL-17A+ (Th1 + Th17) cells after restimulation with PMA/ionomycin. Data show mean of 10 mice ± SEM per group obtained from two independent experiments. For statistical analysis, nonparametric one-way ANOVA was performed. Indicated significant values refer to day 0. *p ≤ 0.05, **p ≤ 0.01, ***p ≤ 0.001.

Close modal

To further strengthen these data, we used DEREG mice in which Foxp3+ Treg can be depleted by the injection of diphtheria toxin (Fig. 6A). Foxp3+ Treg have been shown to play a protective role in DSS colitis, particularly during the recovery phase (18, 22). Furthermore, long-term depletion of Foxp3+ Treg also promotes systemic and intestinal inflammation (13). Thus, we performed a short-term depletion of Foxp3+ Treg and analyzed diphtheria toxin–treated and control mice on day 9, at which point the DSS-treated mice had an overall similar colitis severity in the absence and presence of Foxp3+ Treg as determined by endoscopic score, colon length, and histological score (Fig. 6C). Interestingly, DSS colitis was also accompanied with pathological conditions of the liver, proven by increased histological score and by increased serum liver enzymes upon depletion of Foxp3+ Treg (Fig. 6B).

FIGURE 6.

Induction of hepatitis in DEREG mice after depletion of Foxp3+ Treg during acute colitis. Acute DSS colitis was induced in untreated DEREG males and after injection of diphtheria toxin (DTX) as described in 2Materials and Methods. (A) At day 10, frequencies of GFP+Foxp3+ Treg in the liver and colon of control mice, DSS-treated mice, and DSS + DTX–treated mice were determined. (B) A pathological condition of the liver was determined by AST serum levels and mHAI score of liver sections (see 2Materials and Methods for a definition of mHAI). (C) Colitis severity was monitored by endoscopic score at day 9, weight loss, colon shortening, and histological score. For statistical analysis, nonparametric one-way ANOVA was performed. Indicated significant values refer to day 0. *p ≤ 0.05, **p ≤ 0.01, ***p ≤ 0.001.

FIGURE 6.

Induction of hepatitis in DEREG mice after depletion of Foxp3+ Treg during acute colitis. Acute DSS colitis was induced in untreated DEREG males and after injection of diphtheria toxin (DTX) as described in 2Materials and Methods. (A) At day 10, frequencies of GFP+Foxp3+ Treg in the liver and colon of control mice, DSS-treated mice, and DSS + DTX–treated mice were determined. (B) A pathological condition of the liver was determined by AST serum levels and mHAI score of liver sections (see 2Materials and Methods for a definition of mHAI). (C) Colitis severity was monitored by endoscopic score at day 9, weight loss, colon shortening, and histological score. For statistical analysis, nonparametric one-way ANOVA was performed. Indicated significant values refer to day 0. *p ≤ 0.05, **p ≤ 0.01, ***p ≤ 0.001.

Close modal

Taken together, our data indicate that colitis does not induce pathological conditions of the liver per se, despite the accumulation of Th17 cells in the liver. However, Foxp3+ Treg are concomitantly recruited to the liver. Interestingly, we found that colitis can induce pathological conditions of the liver in mice with a defect in Foxp3+ Treg. Thus, our data support the notion that Foxp3+ Treg might protect mice from developing pathological conditions of the liver during colitis.

The next step was to test whether the inflammatory cell infiltration in the liver during colitis is due to direct migration of these cells from the colon to the liver. To this end, we used KAEDE transgenic mice (14), which express the green florescence KAEDE protein ubiquitous under the control of the CAG promotor. In these mice, all cells express a KAEDE green that can be photoconverted into KAEDE red upon UV light exposure. To convert cells in a specific location of the colon, we used a murine endoscopy system to administer the UV light in the distal colon. Using this system, we indeed converted immune cells in the distal colon but not liver or spleen of untreated KAEDE mice (data not shown). We next illuminated the colon using UV light at the peak of acute DSS colitis. Two days later, we isolated cells from the liver and the colon. We found KAEDE red CD45+ leukocytes in the liver (Fig. 7A), indicating that these cells had been present in the colon at the time of UV light exposure. Interestingly, we could not detect a significant amount of photoconverted cells in untreated mice, indicating that the presence of intestinal inflammation is essential (Fig. 7A). In the KAEDE red cell population, we found CD4+ T cells, a minor fraction of CD8+ T cells, and also B220+CD11c B cells (Fig. 7B). To further confirm that Th17 cells migrate from the colon to the liver, we used the transfer colitis model again and transferred T cells (CD45.2) form KAEDE mice into Rag1−/− (CD45.1+) mice. Upon colitis development, we illuminated the colon with UV light using the endoscopy system and analyzed the mice two days later. Indeed, we found KAEDE red–positive CD4+ CD45.2+ T cells in the liver of these mice, indicating that these CD4+ T cells had been in the colon during the time of UV light administration (Fig. 7C). In conclusion, we found that inflammatory cells, including CD4+ T cells, can migrate during colitis from the colon into the liver.

FIGURE 7.

Migration of B cells and CD4+ T cells from the colon to the liver during colitis. Acute DSS colitis was induced in KAEDE mice as described before. At day 9, colonic cells were photoconverted using a UV light connected to an endoscope. At the indicated time points, mice were sacrificed, and accumulation of KAEDE green (red−) and photoconverted KAEDE red+ cells in the colon and liver was analyzed by FACS. (A) Accumulation of KAEDE red+ CD45+ leukocytes in the liver and colon in control mice and in mice after induction of acute DSS colitis. At day 7, colonic cells were photoconverted, and 2 d later the accumulation of CD45+ cells was analyzed. (B) Dot plots show FACS staining of KAEDE red+ and KAEDE green+ CD4+ T cells, CD8+ T cells, and B220+ B cells in the liver at day 11. Diagrams summarize the frequencies of CD4+ T cells, CD8+ T cells, and B220+ B cells among KAEDE green+ and KAEDE red+ leukocytes in the colon at day 9 after photoconversion and in the liver at day 11. Data summarize mean ± SEM from four mice at day 9 and five mice at day 11. (C) T cell–transfer colitis was induced in Rag1−/− by reconstitution with naive KAEDE CD4+ T cells. After 6 wk, colonic cells were photoconverted. Two days after photoconversion, colon and liver were analyzed for the presence of photoconverted KAEDE red+ cells. Diagrams summarize mean percentage of KAEDE red+ CD4+ T cells in the colon (upper panel) and liver (lower panel) of four mice. Dot plots show representative data obtained from one mouse. *p ≤ 0.05.

FIGURE 7.

Migration of B cells and CD4+ T cells from the colon to the liver during colitis. Acute DSS colitis was induced in KAEDE mice as described before. At day 9, colonic cells were photoconverted using a UV light connected to an endoscope. At the indicated time points, mice were sacrificed, and accumulation of KAEDE green (red−) and photoconverted KAEDE red+ cells in the colon and liver was analyzed by FACS. (A) Accumulation of KAEDE red+ CD45+ leukocytes in the liver and colon in control mice and in mice after induction of acute DSS colitis. At day 7, colonic cells were photoconverted, and 2 d later the accumulation of CD45+ cells was analyzed. (B) Dot plots show FACS staining of KAEDE red+ and KAEDE green+ CD4+ T cells, CD8+ T cells, and B220+ B cells in the liver at day 11. Diagrams summarize the frequencies of CD4+ T cells, CD8+ T cells, and B220+ B cells among KAEDE green+ and KAEDE red+ leukocytes in the colon at day 9 after photoconversion and in the liver at day 11. Data summarize mean ± SEM from four mice at day 9 and five mice at day 11. (C) T cell–transfer colitis was induced in Rag1−/− by reconstitution with naive KAEDE CD4+ T cells. After 6 wk, colonic cells were photoconverted. Two days after photoconversion, colon and liver were analyzed for the presence of photoconverted KAEDE red+ cells. Diagrams summarize mean percentage of KAEDE red+ CD4+ T cells in the colon (upper panel) and liver (lower panel) of four mice. Dot plots show representative data obtained from one mouse. *p ≤ 0.05.

Close modal

Enterohepatic migration of T cells and bacteria from the intestine to the liver is thought to be involved in the pathogenesis of liver diseases. In addition, the Th17–Foxp3 Treg axis has been implicated in liver pathogenesis, as shown for PSC (6, 7). However, direct evidence sustaining these hypotheses is weak and mainly based on correlative studies in patients. We thus aimed to do a mechanistic study, using two complementary mouse colitis models. Furthermore, we wanted to test whether inflammatory cells might be able to migrate directly from the intestine to the liver. We found an increased bacterial burden in the liver in mice with colitis. Furthermore, we could show that adaptive immune cells, including CD4+ T cells, can migrate directly from the inflamed colon to the liver. Accordingly, Th17 cells increased significantly in the liver in mice with colitis. However, this increase of Th17 cells was accompanied by a concomitant increase of Foxp3+ Tregs. These seem to protect mice from the development of pathological conditions of the liver.

The DSS colitis model has been used before to study the role of colitis for liver disease. Interestingly, Seidel et al. (23) had shown that chronic colitis per se leads to improved costimulatory capacity of APCs and an increased activation status of CD8+ T cells in the liver. Furthermore, DSS colitis aggravated Ag-specific CD8+ T cell–driven cholangitis in this study (23). However, CD4+ T cells had not been investigated in that study. Yet CD4+ T cells seem to be particularly relevant in PSC pathogenesis because recent human studies suggest that the Th17 versus Foxp3+ Treg balance plays a key role. Indeed, single nucleotide polymorphisms in the IL-2RA gene, a key molecule for Foxp3+ Treg, has been shown to be associated with IBD and PSC (8). Accordingly, PSC patients show fewer Foxp3+ Treg in the liver (7). Finally, PSC patients have increased bacterial RNA in the liver and demonstrate an increased Th17 response to pathogen stimulation (6). Thus, one major aim of our study was to analyze CD4+ T cells in the liver during colitis. Interestingly, we found increased numbers of Th17 cells in the liver during acute and chronic DSS colitis. Unexpectedly, these inflammatory cells did not cause liver disease, and one possible explanation was that Foxp3+ Treg are likewise increased in the liver during DSS colitis. We therefore used a Th17 cell mediated colitis model, which is deficient in Foxp3+ Treg. Treg-depleted CD45RBhigh T cells differentiate into Th1 and Th17 effector T cells upon transfer into immune deficient Rag1−/− mice. Owing to the absence of Treg, they cause colitis and also systemic inflammation. Of note, the IL-23–Th17 cells axis seems to play a major role in intestinal, but not in systemic, inflammation (24). On the basis of this finding, we had previously modified this model and transferred purified Th17 cells into Rag1−/− mice, which also caused severe colitis (12, 16, 25). However, we had not characterized the liver phenotype in this model before. As control, we injected Th1 cells, which also repopulated the host. Interestingly, transfer of Th17 but not Th1 cells caused moderate to severe colitis. Moreover, the Th17 cell number was increased in the liver upon Th17 but not Th1 cell transfer. Finally, the presence of colitis and Th17 cell infiltration was associated with a significant pathological condition of the liver on the basis of serum transaminase levels and histological findings. These data support the notion that Th17 cells in the absence of Foxp3+ Treg are in principle able to drive pathological conditions of the liver. However, further studies will be essential to clarify the lack of pathogenicity of the Th1 cell transfer.

Microbes might leak from the colon to the liver because of barrier defects during colitis. Accordingly, it was proposed that the liver may act as a functional vascular firewall that clears commensals that have penetrated the intestinal vascular circuits (1). Based on these data, one hypothesis is that infectious agents, such as bacterial DNA derived from the gut microbiota, might drive the immune response in the liver and thus promote liver disease, as suggested in PSC (26). Indeed, we detected bacterial DNA in more mice during DSS colitis compared with healthy controls. These data are in line with a previous study that demonstrated increased LPS levels in the portal vein in DSS colitis (27). Interestingly, we found that the presence of Foxp3+ Treg prevents the development of liver inflammation. One possible explanation is that Foxp3+ Treg interfere with bacteria clearance, and future studies to evaluate the impact of Treg on antibacterial immune responses in colitis associated with a pathological condition of the liver are important.

T cells, which display a gut homing phenotype, have been identified in the liver of PSC patients (28, 29). On the basis of this finding, the enterohepatic migration of T cells from the colon to the liver has been proposed as one mechanism driving PSC. However, despite these studies, direct evidence demonstrating the migration of T cells from the colon into the liver was missing. We used the KAEDE transgenic mice which allowed us to mark immune cells specifically in the colon during colitis and thus track their migration to other organs (14). Indeed, we found lymphocytes in the liver, including CD4+ T cells, which had been in the colon during the peak of colitis. By contrast, we did not find a significant number of innate cells, which had been in the colon during colitis, in the liver. These data provide direct evidence for enterohepatic migration of adaptive immune cells from the colon to the liver during colitis.

In conclusion, our data indicate that colitis does not induce a pathological condition of the liver in the presence of Foxp3+ Treg per se. However, intestinal CD4+ T cells from mice with colitis are able to migrate to the liver. Furthermore, colitis promotes pathological conditions of the liver in the absence of Foxp3+ Treg. Thus, our results suggest that colon-derived CD4+ T cells, upon migration to the liver, might be stimulated by the same microbial Ags, which they have been primed to in the intestine. During this process, CD4+ T cells might drive inflammation in the liver if not controlled by Foxp3+ Treg. These data sustain previous human studies that link increased Th17 cell numbers and a deficiency in Foxp3+ Treg with PSC development (68). Taken together, we propose that IBD might promote liver disease in genetically susceptible patients. Therefore, a strict control of IBD activity in these patients would be warranted. However, it remains unclear whether colitis actually promotes bystander hepatitis or whether it drives hepatic disease development and progression in genetically susceptible patients. Clarifying this will require further studies combining colitis models, hepatitis models, and PSC models in mice.

We thank Cathleen Haueis and Sandra Wende for excellent technical assistance and Elaine Hussey for correcting the manuscript.

This work was supported by the Deutsche Forschungsgemeinschaft (Grant SFB841 to S.H. and C.S. and Grant KFO306 to S.H., J.H., and C.S.), the Stiftung Experimentelle Biomedizin (to S.H.), and a Heisenberg Professorship (to S.H.).

The online version of this article contains supplemental material.

Abbreviations used in this article:

     
  • AST

    aspartate aminotransferase

  •  
  • BDL

    bile duct ligation

  •  
  • Ct

    threshold cycle

  •  
  • DEREG

    depletion of Treg

  •  
  • DSS

    dextran sodium sulfate

  •  
  • GGT

    γ glutamyl transferase

  •  
  • Hprt

    hypoxanthine phosphoribosyltransferase

  •  
  • IBD

    inflammatory bowel disease

  •  
  • mHAI

    modified histological activation index

  •  
  • mIL

    murine IL

  •  
  • mRFP

    monomeric RFP

  •  
  • PSC

    primary sclerosing cholangitis

  •  
  • RFP

    red fluorescent protein

  •  
  • Treg

    regulatory T cell.

1
Balmer
,
M. L.
,
E.
Slack
,
A.
de Gottardi
,
M. A.
Lawson
,
S.
Hapfelmeier
,
L.
Miele
,
A.
Grieco
,
H.
Van Vlierberghe
,
R.
Fahrner
,
N.
Patuto
, et al
.
2014
.
The liver may act as a firewall mediating mutualism between the host and its gut commensal microbiota.
Sci. Transl. Med.
6
:
237ra66
.
2
Chapman
,
R. W.
,
B. A.
Arborgh
,
J. M.
Rhodes
,
J. A.
Summerfield
,
R.
Dick
,
P. J.
Scheuer
,
S.
Sherlock
.
1980
.
Primary sclerosing cholangitis: a review of its clinical features, cholangiography, and hepatic histology.
Gut
21
:
870
877
.
3
Loftus
,
E. V.
 Jr.
,
G. C.
Harewood
,
C. G.
Loftus
,
W. J.
Tremaine
,
W. S.
Harmsen
,
A. R.
Zinsmeister
,
D. A.
Jewell
,
W. J.
Sandborn
.
2005
.
PSC-IBD: a unique form of inflammatory bowel disease associated with primary sclerosing cholangitis.
Gut
54
:
91
96
.
4
de Vries
,
A. B.
,
M.
Janse
,
H.
Blokzijl
,
R. K.
Weersma
.
2015
.
Distinctive inflammatory bowel disease phenotype in primary sclerosing cholangitis.
World J. Gastroenterol.
21
:
1956
1971
.
5
Sabino
,
J.
,
S.
Vieira-Silva
,
K.
Machiels
,
M.
Joossens
,
G.
Falony
,
V.
Ballet
,
M.
Ferrante
,
G.
Van Assche
,
S.
Van der Merwe
,
S.
Vermeire
,
J.
Raes
.
2016
.
Primary sclerosing cholangitis is characterised by intestinal dysbiosis independent from IBD.
Gut
65
:
1681
1689
.
6
Katt
,
J.
,
D.
Schwinge
,
T.
Schoknecht
,
A.
Quaas
,
I.
Sobottka
,
E.
Burandt
,
C.
Becker
,
M. F.
Neurath
,
A. W.
Lohse
,
J.
Herkel
,
C.
Schramm
.
2013
.
Increased T helper type 17 response to pathogen stimulation in patients with primary sclerosing cholangitis.
Hepatology
58
:
1084
1093
.
7
Sebode
,
M.
,
M.
Peiseler
,
B.
Franke
,
D.
Schwinge
,
T.
Schoknecht
,
F.
Wortmann
,
A.
Quaas
,
B. S.
Petersen
,
E.
Ellinghaus
,
U.
Baron
, et al
.
2014
.
Reduced FOXP3(+) regulatory T cells in patients with primary sclerosing cholangitis are associated with IL2RA gene polymorphisms.
J. Hepatol.
60
:
1010
1016
.
8
Liu
,
J. Z.
,
J. R.
Hov
,
T.
Folseraas
,
E.
Ellinghaus
,
S. M.
Rushbrook
,
N. T.
Doncheva
,
O. A.
Andreassen
,
R. K.
Weersma
,
T. J.
Weismüller
,
B.
Eksteen
, et al
International IBD Genetics Consortium
.
2013
.
Dense genotyping of immune-related disease regions identifies nine new risk loci for primary sclerosing cholangitis.
Nat. Genet.
45
:
670
675
.
9
Ellinghaus
,
D.
,
L.
Jostins
,
S. L.
Spain
,
A.
Cortes
,
J.
Bethune
,
B.
Han
,
Y. R.
Park
,
S.
Raychaudhuri
,
J. G.
Pouget
,
M.
Hübenthal
, et al
.
2016
.
Analysis of five chronic inflammatory diseases identifies 27 new associations and highlights disease-specific patterns at shared loci.
Nat. Genet.
48
:
510
518
.
10
Liberal
,
R.
,
C. R.
Grant
,
M. S.
Longhi
,
G.
Mieli-Vergani
,
D.
Vergani
.
2015
.
Regulatory T cells: mechanisms of suppression and impairment in autoimmune liver disease.
IUBMB Life
67
:
88
97
.
11
Kiesler
,
P.
,
I. J.
Fuss
,
W.
Strober
.
2015
.
Experimental models of inflammatory bowel diseases.
Cell. Mol. Gastroenterol. Hepatol.
1
:
154
170
.
12
Gagliani
,
N.
,
M. C.
Amezcua Vesely
,
A.
Iseppon
,
L.
Brockmann
,
H.
Xu
,
N. W.
Palm
,
M. R.
de Zoete
,
P.
Licona-Limón
,
R. S.
Paiva
,
T.
Ching
, et al
.
2015
.
Th17 cells transdifferentiate into regulatory T cells during resolution of inflammation.
Nature
523
:
221
225
.
13
Lahl
,
K.
,
C.
Loddenkemper
,
C.
Drouin
,
J.
Freyer
,
J.
Arnason
,
G.
Eberl
,
A.
Hamann
,
H.
Wagner
,
J.
Huehn
,
T.
Sparwasser
.
2007
.
Selective depletion of Foxp3+ regulatory T cells induces a scurfy-like disease.
J. Exp. Med.
204
:
57
63
.
14
Tomura
,
M.
,
N.
Yoshida
,
J.
Tanaka
,
S.
Karasawa
,
Y.
Miwa
,
A.
Miyawaki
,
O.
Kanagawa
.
2008
.
Monitoring cellular movement in vivo with photoconvertible fluorescence protein “Kaede” transgenic mice.
Proc. Natl. Acad. Sci. USA
105
:
10871
10876
.
15
Wan
,
Y. Y.
,
R. A.
Flavell
.
2005
.
Identifying Foxp3-expressing suppressor T cells with a bicistronic reporter.
Proc. Natl. Acad. Sci. USA
102
:
5126
5131
.
16
Esplugues
,
E.
,
S.
Huber
,
N.
Gagliani
,
A. E.
Hauser
,
T.
Town
,
Y. Y.
Wan
,
W.
O’Connor
Jr.
,
A.
Rongvaux
,
N.
Van Rooijen
,
A. M.
Haberman
, et al
.
2011
.
Control of TH17 cells occurs in the small intestine.
Nature
475
:
514
518
.
17
Becker
,
C.
,
M. C.
Fantini
,
C.
Schramm
,
H. A.
Lehr
,
S.
Wirtz
,
A.
Nikolaev
,
J.
Burg
,
S.
Strand
,
R.
Kiesslich
,
S.
Huber
, et al
.
2004
.
TGF-beta suppresses tumor progression in colon cancer by inhibition of IL-6 trans-signaling.
Immunity
21
:
491
501
.
18
Huber
,
S.
,
C.
Schramm
,
H. A.
Lehr
,
A.
Mann
,
S.
Schmitt
,
C.
Becker
,
M.
Protschka
,
P. R.
Galle
,
M. F.
Neurath
,
M.
Blessing
.
2004
.
Cutting edge: TGF-beta signaling is required for the in vivo expansion and immunosuppressive capacity of regulatory CD4+CD25+ T cells.
J. Immunol.
173
:
6526
6531
.
19
Ishak
,
K.
,
A.
Baptista
,
L.
Bianchi
,
F.
Callea
,
J.
De Groote
,
F.
Gudat
,
H.
Denk
,
V.
Desmet
,
G.
Korb
,
R. N.
MacSween
, et al
.
1995
.
Histological grading and staging of chronic hepatitis.
J. Hepatol.
22
:
696
699
.
20
Constandinou
,
C.
,
N.
Henderson
,
J. P.
Iredale
.
2005
.
Modeling liver fibrosis in rodents.
Methods Mol. Med.
117
:
237
250
.
21
Guo
,
S.
,
R.
Al-Sadi
,
H. M.
Said
,
T. Y.
Ma
.
2013
.
Lipopolysaccharide causes an increase in intestinal tight junction permeability in vitro and in vivo by inducing enterocyte membrane expression and localization of TLR-4 and CD14.
Am. J. Pathol.
182
:
375
387
.
22
Schwinge
,
D.
,
F.
von Haxthausen
,
A.
Quaas
,
A.
Carambia
,
B.
Otto
,
F.
Glaser
,
B.
Höh
,
N.
Thiele
,
T.
Schoknecht
,
S.
Huber
, et al
.
2017
.
Dysfunction of hepatic regulatory T cells in experimental sclerosing cholangitis is related to IL-12 signaling.
J. Hepatol.
66
:
798
805
.
23
Seidel
,
D.
,
I.
Eickmeier
,
A. A.
Kühl
,
A.
Hamann
,
C.
Loddenkemper
,
E.
Schott
.
2014
.
CD8 T cells primed in the gut-associated lymphoid tissue induce immune-mediated cholangitis in mice.
Hepatology
59
:
601
611
.
24
Ahern
,
P. P.
,
C.
Schiering
,
S.
Buonocore
,
M. J.
McGeachy
,
D. J.
Cua
,
K. J.
Maloy
,
F.
Powrie
.
2010
.
Interleukin-23 drives intestinal inflammation through direct activity on T cells.
Immunity
33
:
279
288
.
25
Huber
,
S.
,
N.
Gagliani
,
E.
Esplugues
,
W.
O’Connor
Jr.
,
F. J.
Huber
,
A.
Chaudhry
,
M.
Kamanaka
,
Y.
Kobayashi
,
C. J.
Booth
,
A. Y.
Rudensky
, et al
.
2011
.
Th17 cells express interleukin-10 receptor and are controlled by Foxp3 and Foxp3+ regulatory CD4+ T cells in an interleukin-10-dependent manner.
Immunity
34
:
554
565
.
26
Terjung
,
B.
,
U.
Spengler
.
2009
.
Atypical p-ANCA in PSC and AIH: a hint toward a “leaky gut”?
Clin. Rev. Allergy Immunol.
36
:
40
51
.
27
Gäbele
,
E.
,
K.
Dostert
,
C.
Hofmann
,
R.
Wiest
,
J.
Schölmerich
,
C.
Hellerbrand
,
F.
Obermeier
.
2011
.
DSS induced colitis increases portal LPS levels and enhances hepatic inflammation and fibrogenesis in experimental NASH.
J. Hepatol.
55
:
1391
1399
.
28
Ponsioen
,
C. Y.
,
H.
Kuiper
,
F. J.
Ten Kate
,
M.
van Milligen de Wit
,
S. J.
van Deventer
,
G. N.
Tytgat
.
1999
.
Immunohistochemical analysis of inflammation in primary sclerosing cholangitis.
Eur. J. Gastroenterol. Hepatol.
11
:
769
774
.
29
Eksteen
,
B.
,
A. J.
Grant
,
A.
Miles
,
S. M.
Curbishley
,
P. F.
Lalor
,
S. G.
Hübscher
,
M.
Briskin
,
M.
Salmon
,
D. H.
Adams
.
2004
.
Hepatic endothelial CCL25 mediates the recruitment of CCR9+ gut-homing lymphocytes to the liver in primary sclerosing cholangitis.
J. Exp. Med.
200
:
1511
1517
.

The authors have no financial conflicts of interest.

Supplementary data