Environmental signals, such as dietary, microbial, or xenobiotic factors, are sensed in the intestinal tissue by the aryl hydrocarbon receptor (AhR), which is an important regulator of metabolism, but also influences immune cell homeostasis and immune activation in the intestine [1
]. AhR expression is especially high in the liver and in barrier tissues such as skin, lung, and gut [2
]. Additionally, it has been shown that the AhR can be induced by immunological stimuli, such as pathogen-associated molecular patterns [1
AhR activation plays an important role in intestinal immunity, contributing to intestinal homeostasis, inflammation, and host defense [3
]. The AhR controls interleukin (IL)-22 production by innate lymphoid cells (ILCs), and thus confers host defense against Citrobacter rodentium
infection in mice [3
]. Besides IL-22 production, AhR activation through high affinity AhR ligands has been shown to stimulate the production of antimicrobial peptides and regulate tissue regeneration [5
]. AhR-deficient mice are highly susceptible to dextran sodium sulfate (DSS)-induced colitis, which at least in part results from a reduction of intraepithelial lymphocytes (IELs) [6
] and impaired homeostasis of intestinal epithelial cells (IEL) [8
]. Furthermore, the AhR has been shown to be an important regulator of T cell immunity in intestinal inflammation, regulating IL-17, Foxp3, IL-10, and IL-22 expression, and altogether ameliorating colitis symptoms and maintaining intestinal homeostasis [9
]. In addition, Monteleone et al. could show that the AhR is down-regulated in intestinal tissue of patients with inflammatory bowel disease (IBD) and that AhR signalling is able to inhibit inflammation in colitis of the gastrointestinal tract of mice [12
]. This indicates a major role of the AhR in resolving intestinal inflammation and makes the AhR an interesting pharmacological target in IBD.
One of the key target genes activated by the genomic AhR pathway is the AhR repressor (AhRR). The AhRR is highly homologous to the AhR, but lacks the Period/ARNT/Single-minded (PAS)-B and the transactivation domain. It competes with AhR for AhR nuclear translocator (ARNT) binding, but cannot initiate transcription, and hence suppresses the AhR signalling pathway [13
]. Additionally, it has been demonstrated that the AhRR may interfere with non-canonical AhR-mediated signalling by interacting with RelB, and thereby inhibiting inflammatory responses [17
expression is restricted to some cell populations in a given tissue and does not strictly correlate with expression of Cyp1a1
, and Cyp1b1
, other major target genes of the AhR encoding important cytochrome p450 enzymes, which are responsible for degradation of AhR ligands [2
]. Specifically, using AhRR/Enhanced Green Fluorescent Protein (EGFP) reporter mice, we could show that Ahrr
expression is restricted to immune cells in barrier organs, such as skin and intestine [21
]. We could further observe that AhRR activation is indeed largely AhR-dependent and that cell types that highly express the Ahrr
display only mild Cyp1a1
expression and vice versa, indicating the importance of different feedback inhibition mechanisms in different cell types. Interestingly both, AhR and AhRR-deficient mice are highly susceptible to DSS-induced colitis, which is likely due to the highly cell type specific expression of the AhRR and the resulting cell type-specific differences in AhR/AhRR signalling [21
The mucosal surface area of the gut represents an enormous area, which is in direct contact with the environment. In addition to occasional pathogen encounters, the intestinal immune system is constantly exposed to antigens from the diet or the microbiota. Therefore, it is essential that gut-associated immune cells maintain a balance between protection against harmful infections and tolerating harmless food-derived antigens and commensals. Constant availability of dietary substances and the microbiota in the intestinal lumen lead to continuous stimulation of the AhR signalling pathway. With regard to food, cruciferous vegetables are an important source of AhR ligands in the intestine as they contain high concentrations of glucobrassicin. Enzymatic processing of this substance leads to the formation of indole-3-carbinol (I3C). Stomach acid-catalysed condensation of I3C then generates a number of biologically active substances, such as the AhR ligands 3,3′-diindolylmethane (DIM) and indolo[3,2-b]carbazole (ICZ) [6
]. Dietary supplementation of rodents with I3C [6
] or broccoli extracts containing glucobrassicin [23
] led to profound changes in the microbiome and conferred protection from DSS-induced colitis. Next to dietary components, microbial factors are able to activate AhR-signalling. The AhR senses bacterial pigments, leading to induction of canonical detoxifying genes as well as regulation of cytokines and chemokines, thereby fighting bacterial infections [24
]. Furthermore, tryptophan derivatives derived from the diet or bacterial metabolites have been shown to tune the intestinal immune system by signalling through the AhR [5
]. The availability of microbiota produced AhR ligands is critically dependent on the local microbiota composition and the presence of certain bacterial strains. Lactobacillus reuteri
, which has the ability to metabolize tryptophan to AhR activating indoles, and the probiotic strain Lactobacillus bulgaricus
OLL1181, have shown AhR activating potential [26
In this study, we analyzed the influence of dietary AhR ligands on AhRR expression, colitis pathology, and changes in the microbiome. Using AhRR-reporter mice, we could show that AhR activation in intestinal immune cells is modulated by dietary AhR ligands, but not by the absence of commensal microbes. The application of several AhR-ligands in normal mouse chow or supplementation of a ligand-reduced diet (LRD) with I3C protected from DSS-induced colitis and significantly changed the microbial community in the gut compared with LRD. We provide insight into the interplay of gut microbiota and AhR signalling and demonstrate that dietary intervention with I3C acts in an AhR-dependent as well as AhR-independent manner.
The uptake of dietary AhR ligands exerts important regulatory functions on the integrity of the mucosal barrier, cellular metabolism, and the proper function of the intestinal immune system. Whereas uptake of environmental pollutants such as dioxins may cause liver and immune toxicity [38
], natural AhR ligands contained in vegetables and fruit represent important constituents of a healthy diet [1
]. Using different AhR ligand-deprived experimental mouse diets and the well-established mouse model of dietary I3C supplementation [6
], mimicking the consumption of cruciferous vegetables, we here investigated the effect of dietary AhR ligands on AhR activation in various intestinal immune cell subsets. We show that the AhRR/EGFP reporter mouse model is a valuable tool to precisely quantify AhR target gene expression in immune cells in the gut-associated lymphoid tissue (GALT). Interestingly, Ahrr
expression was predominantly regulated by the presence of dietary AhR ligands, but not by the intestinal microbiota and their metabolites. Although dietary I3C supplementation induced strong expression of the AhRR reporter in myeloid cells and T cells of the lamina propria, as well as IELs and DCs located in the mLNs, many of the profound changes in microbiome composition caused by I3C supplementation also occurred in AhR-deficient mice, and were thus shown to be independent of AhR signalling. Using thorough bioinformatic analyses of these I3C-dependent alterations in the microbiome and direct comparison of WT, AhRR-deficient, and AhR-deficient mice, we were able to identify the microbial communities that responded to I3C stimulation in an AhR-dependent or -independent manner.
Up to now, activation of the AhR through oral ingestion of environmental pollutants or phytochemicals in the intestine has been mainly assessed at the level of AhR target gene expression, with a focus on expression of Cyp1a1
, the most strongly inducible AhR target gene encoding the cytochrome p450 oxygenase CYP1A1, an important xenobiotic metabolizing enzyme. For this, either RT-qPCR analysis [41
] or fluorescent protein reporter mice have been utilized [8
]. In the intestine, however, Cyp1a1
is only expressed in intestinal epithelial cells, but not in intestinal immune cells, and its expression is barely detectable under homeostatic conditions [8
]. In contrast, the AhRR/EGFP reporter mouse model offers the possibility to easily quantify the presence of dietary AhR ligands by means of EGFP reporter expression in intestinal immune cells. Ahrr
expression is readily detectable in mice fed a normal chow diet [21
], containing phytochemicals derived from grains and vegetable oils. Remarkably, in mice fed the purified diet AIN-93G (here called LRD), which lacks phytochemicals [31
], AhRR expression was reduced by about 5–10-fold based on EGFP reporter expression, indicating a major contribution of food-derived AhR ligands to intestinal AhR activation. Further, we also observed reduced AhRR/EGFP expression in the skin of LRD fed mice, indicating that dietary AhR ligands may also affect distal organ functions (data not shown). As shown in this study, Ahrr
expression is strongly induced in intestinal immune cells by feeding I3C, which is metabolized to the AhR ligands DIM and ICZ during the gastric passage [42
]. As demonstrated earlier by other groups and in line with the high susceptibility of AhR−/−
mice to development of colitis [4
], we also demonstrated an impaired barrier integrity, low IL-22 levels, and increased colitis susceptibility in mice fed LRD. A similar reduction in intestinal Ahrr
expression as shown for LRD was observed when mice were fed a commercially available HFD or matching control diet. Both HFD and HFD ctrl are purified diets similar to LRD, but differ from LRD regarding carbohydrate composition, and by the addition of a relatively undefined source of fat in the form of lard. Therefore, we would like to point out that commonly used models of diet-induced obesity may be confounded by alterations in intestinal AhR signaling, a problem that is even more pronounced when feeding of HFD is compared with NC rather than HFD ctrl.
In contrast to the dominant role of dietary AhR ligands in regulating Ahrr
expression in immune cells, we could not detect differences in Ahrr
reporter expression after depletion of the intestinal microbiota by broad spectrum antibiotics. This does, however, not absolutely exclude a possible participation of microbiota derived AhR ligands in AhR activation in mice fed NC. As the strong reduction in Ahrr
expression after feeding LRD was accompanied by a major change in microbiota composition, the production of potential AhR ligands by the microbiota may have simultaneously been altered. As an alternative explanation, it is very possible that the composition of the microbiota in our specific-pathogen-free mouse facility does not comprise commensal strains such as Lactobacillus reuteri
, which have previously been described to be potent producers of AhR activating indoles [5
The most unexpected finding of the experiments described in this study was that the I3C induced changes in microbiome composition were to a large extent not AhR-dependent, as they also occurred in AhR-deficient mice and were similar in the AhRR-deficient background. Obviously, I3C stimulation strongly activated the AhR, leading to upregulation of Ahrr
expression, but this accounted only for a fraction of the alterations in the representation of microbial species. LRD + I3C prevented the outgrowth of potentially colitogenic Enterobacteriaceae
, which were almost down to the level observed after NC feeding. In addition, the Clostridiales
were strongly expanded. Clostridia
are known as major producers of short-chain fatty acids (SCFAs), which are important energy sources for enterocytes and also exert immunoregulatory functions [45
]. They additionally support IL-22 production [46
], which was markedly increased after I3C supplementation in our colitis experiments. Clostridia
-produced SCFAs further support the expansion of regulatory T cells, and thus inhibit intestinal inflammation [44
]. This is a valid explanation for the protection from colitis through I3C, although it appears to be at least partially AhR-independent. In a study very similar to ours, in which the effect of I3C on 2,4,6-trinitrobenzenesulfonic acid-induced colitis was analyzed, an expansion of SCFA-producing Roseburia spp.
belonging to the Lachnospiraceae
family was described and correlated with enhanced production of IL-22 and protection form colitis [22
]. In this study, however, the dependence of this effect on AhR stimulation was not assessed. Moreover, LRD + I3C did not fully restore the alpha diversity of the microbiota to the complexity seen in NC-fed WT and AhRRE/E
mice, which may additionally account for the fact that LRD + I3C fed mice still had a higher susceptibility for DSS colitis than mice fed NC.
In line with our findings, Julliard et al. reported that an I3C-supplemented diet protected both WT and AhR−/−
mice from an infection with Clostridium difficile
, indicating that I3C can act in both an AhR-dependent and -independent manner [48
]. It may be important in this context that I3C has been described to possess direct broad-spectrum anti-microbial activity against gram-positive and gram-negative bacteria, including pathogenic Escherichia coli
at a dose range of 20–80 μg/mL [49
]. This concentration could potentially have been reached in the intestine by our feeding regime, although it is unclear to which extent and with which kinetics I3C is metabolized to other compounds during gastric passage. Thus, an anti-microbial activity of I3C may contribute to the loss of Enterobacteriaceae
observed after feeding LRD + I3C as compared with LRD alone. Another possible mode of action of I3C is via its metabolite DIM, which acts as a ligand not only of the AhR, but also of the orphan G-protein-coupled receptor GPR84 [50
]. GPR84 is expressed on immune cells and has been shown to trigger pro-inflammatory signaling pathways and phagocytosis in macrophages [51
Besides such AhR-independent effects of I3C, our bioinformatic analysis also revealed distinct alterations of the microbiome, which exhibited a clear AhR- or AhRR-dependency. Alloprevotella
, for example, specifically increased after feeding of LRD + I3C in WT, but not AhR−/−
mice. This genus was also detected at a higher frequency in AhRRE/E
mice fed LRD compared with WT and AhR−/−
mice, possibly indicating a higher AhR activation by residual AhR ligands in LRD-fed mice in the absence of the AhRR. The genus Mucispirillum
) and Faecalibaculum
), in turn, showed a significant outgrowth in both WT and AhRRE/E
mice, but not in AhR−/−
mice after feeding LRD + I3C as compared with LRD. Changes in the frequency of Erysipelotrichaceae
have been associated with AhR signaling in at least two other studies. Brawner et al. showed, however, that Erysipelotrichaceae
were enriched in the feces of mice fed the LRD AIN-76A, which is slightly different from the LRD used in this study [37
]. In addition, feeding of mice with a diet containing broccoli also decreased the abundance of Erysilelotrichaceae
]. In line with our results, on the other hand, a strong reduction of Erysipelotrichaceae
was associated with a higher susceptibility to inflammatory bowel disease in humans [52
In conclusion, our findings are in agreement with earlier reports that dietary I3C supplementation restores AhR activation in the intestinal mucosa under conditions of malnutrition and deprivation of natural AhR ligands. In humans, such malnutrition may result from a severely reduced consumption of vegetables and fruit in favor of a carbohydrate rich, high fat diet typical of a Western diet, as opposed to a Mediterranean diet. Moreover, in experimental research, it should be taken into consideration that the commonly used HFD products and matching control diets can lead to an impairment of AhR signaling in intestinal immune cells and epithelial cells. On the other hand, we also demonstrate that feeding of LRD with or without I3C supplementation not only affects AhR activation, but may also cause major changes in the composition of the microbiota in an AhR-independent manner, for example, through recognition of dietary constituents by other chemosensing receptors of the host or by direct action on the microbiota itself.
4. Materials and Methods
(AhRRE/E), heterozygous AhRRE/+
mice, and AhR−/−
] were bred at the animal facility of the LIMES, Bonn, Germany. Homozygous AhRRE/E
mice are deficient in AhRR expression, and thus represent AhRR-knockout mice. WT littermate mice served as controls. Six- to twelve-week-old male and female mice were used for the experiments and were bred according to German guidelines for animal care. All experiments were performed according to German and Institutional guidelines for animal experimentation (permits: AZ 84-02.04.2011.A186 (approval date: 24 February 2011) and 84-02.04.2016.A210 (approval date: 16 February 2017) and were approved by the government of North Rhine-Westphalia (Germany).
4.2. Feeding of Experimental Diets
Mice were fed for four weeks from weaning onwards with different diets. Normal chow diet (NC) was purchased from LASvendi (Soest, Germany), while all other diets were purchased from ssniff Spezialdiäten GmbH (Soest, Germany). AIN 93G1, termed ligand-reduced diet (LRD), was used as purified diet or supplemented with 2 g I3C per kg (LRD + I3C). As an experimental diet with high fat content (HFD), we used ssniff® EF acc. D12492 (I) mod. and ssniff® EF acc. D12450B (I) mod.* was used as control diet, suggested by the vendor. After four weeks of feeding, mice were either analyzed directly or DSS-induced colitis was performed.
4.3. Oral Antibiotic Treatment
Mice were treated orally with a broad-sprectrum antibiotic cocktail via the drinking water for four weeks (Ampicillin 1 g/L, Vancomycin 500 mg/L, Ciprofloxacin 200 mg/L, Imipenem 250 mg/L, and Metronidazole 1 g/L in drinking water) in order to deplete the gut microbiota. Depletion was controlled by incubating fecal pellets for 24 h in thioglycolate bouillon, plating them on Columbia blood agar plates and incubating these for another 24 h before analysis. AhRR expression was analyzed in isolated intestinal immune cells by flow cytometry and histologically in colon and mLNs.4.4. DSS-Induced Colitis
Dextran sodium sulfate (DSS) colitis was induced in aged-matched female mice by adding 3% DSS (w/vol) to the drinking water. Mice were kept on DSS-containing water for four days. The general state of health, behavior, posture, stool consistency, rectal bleeding, and body weight were monitored every other day. At the end of the experiment, serum was obtained for analysis of inflammatory cytokines or markers. Small intestine, colon, and mLNs were removed and either transferred into 4% paraformaldehyde (PFA) or stored in phosphate-buffered saline (PBS) on ice until further use. In addition, colon length was determined.
4.4. Cytokine Determination in Colon Tissue
After five days of DSS colitis, a small piece of colon tissue was transferred to 500 µL radioimmunoprecipitation assay buffer in a reaction vessel with glass beads and homogenized using the Precellys® 24 tissue homogenizer. The lysate was then transferred to a 1.5 mL Eppendorf tube and centrifuged at 15,000× g for 15 min at 4 °C. The supernatant was transferred to a new tube and cytokines were measured by enzyme-linked immunosorbent assay (bio-techne, Wiesbaden, Germany) according to the manufacturer’s instructions.
4.5. Determination of Intestinal Permeability
Intestinal barrier function and intestinal barrier permeability were determined after oral application of FITC-coupled Dextran. Mice were fasted 4 h prior to FITC-Dextran administration to facilitate absorption from the intestinal lumen. Mice were orally gavaged with FITC-Dextran in PBS (600 mg per kg body weight). Mice were left without food for a further four hours. Thereafter, serum was sampled and FITC fluorescence was determined with a Tecan infinite M200 plate spectrophotometer at an excitation wavelength of 492 nm and an emission wavelength of 525 nm.
Tissue samples were fixed for 3 h in 4% PFA at 4 °C and saturated in a sucrose gradient from 5% to 20% sucrose. Samples were embedded in cryomedium and sectioned at a thickness of 10 μm. Sections were counterstained with 0.5 μg/mL DAPI (4,6-diamidino-2-phenylindole). Images were acquired with a Keyence B2900 digital microscope (Keyence Corporation, Osaka, Japan) and analyzed with BZII Analyzer software (Keyence Cooperation).
4.7. Isolation of Intestinal Immune Cell Subsets
For isolation of IEL or lamina propria lymphocytes (LPL), the colon and SI were opened longitudinal and cut into 1–2 cm long pieces. To isolate IEL, the tissue was incubated in 15 mM 4-(2-hydroxyethyl)-1-piperazineethanesulfonic acid (HEPES), 5 mM ethylenediaminetetraacetic acid (EDTA), and 10% fetal calf serum (FCS) in PBS for 45 min at 37 °C while shaking. Cells were afterwards filtered through a 70 μM cell strainer. For LPL isolation, mucus removal of intestinal tissue pieces was performed in 5 mM dithiothreitol, 2% FCS, 100 U/mL Penicillin, and 100 ug/mL streptomycin in Hank’s balanced salt solution (HBSS) for 20 min at 37 °C. Epithelial cells were removed by incubation in 5 mM EDTA, 2% FCS, 100 U/mL penicillin, and 100 μg/mL streptomycin in HBSS three times for 15 min at 37 °C and washed with 10 mM HEPES, 100 U/mL penicillin, and 100 μg/mL streptomycin in HBSS for 10 min at 37 °C. Tissue was digested with 4 U/mL liberase and 4000 U/mL DNaseI in 10 mM HEPES, 100 U/mL penicillin, and 100 µg/mL streptomycin in HBSS for 45 min at 37 °C, and afterwards filtered through a 70 μm cell strainer. Cell suspension was centrifuged and stained for analysis by flow cytometry. For preparation of mLN cell suspensions, tissues were meshed through a 100 and 70 μm cell strainer.
4.8. Flow Cytometry
Cell suspensions were stained with antibodies against CD3 (145-2C11), CD4 (RM4-5), CD8α (53-6.7), CD8β (YTS156.7.7), MHCII (M5/114.15.2), CD11c (HL-3), F4/80 (CI:A3-1, Abd Serotec, Oxford, UK), CD11b (M1/70), CD64 (X54-5/7.1), TCR β chain (H57-597), TCRγδ (GL3), and Foxp3 (FJK16s). If not indicated otherwise, antibodies were purchased from eBioscience. For intracellular cytokine and transcription factor analysis, cells were fixed with 2% PFA for 20 min, permeabilized with 0.5% saponin in PBS/BSA, and stained for 60 min at room temperature (RT) in the dark. For anti-GFP staining (purified anti-GFP and anti-Rabbit IgG AF488 from Life Technologies), incubation was performed overnight at 4 °C. Cell populations were analyzed with a LSRII Cytometer (BD Biosciences) or a BD FACSymphony™; data were analyzed with FlowJo software (Tree star, Ashland, OR, USA).
4.9. Real-Time PCR Analysis
SI tissue was taken from LRD or LRD + I3C fed mice, tissue lysis and homogenization was performed with the Precellys®24 homogenizer, and RNA was isolated using the Zymo Research Direct-zol MiniPrep kit according to the manufacturer’s instructions. First-strand cDNA was synthesized from 1 μg of total RNA using Revert Aid reverse transcriptase (Thermo Fisher Scientific, Bonn, Germany). Real-time PCR was performed on a BioRad CFX96 Touch™ Real-Time PCR Detection System using absolute SYBR-green ROX master mix (Thermo Fisher Scientific). Primers were designed using the Universal Probe Library (Roche Applied Science, Mannheim, Germany): cyp1a1 fwd: 5′-CCTCATGTACCTGGTAACCA-3′, cyp1a1 rev: 5′-AAGGATGAATGCCGGAAGGT-3′, GAPDH fwd: 5′-GAGCCAAACGGGTCATCA-3′, GAPDH rev: 5′-CATATTTCTCGTGGTTCACACC-3′.
4.10. 16S rRNA Gene Data Collection and Sequencing
Fecal DNA was isolated and purified using the QIAamp DNA Stool Mini Kit (QIAGEN Co., Germany). The V3–V4 hypervariable region of 16S rRNA genes was analyzed according to the Illumina protocol for 16S metagenomic sequencing library preparation with minor modifications. The fecal DNA was amplified with 341F and 806R primers:
PCR amplicons were purified using AMPure XP (Beckman Coulter Co., USA) according to the manufacturer’s instructions. Adapters and barcodes (Nextera XT, Illumina Co., San Diego, CA, USA) were attached to the amplicons to conduct multiplex sequencing. Barcoded amplicons were sequenced using the Illumina MiSeq 2 × 300 bp platform with the MiSeq reagent kit v3 (Illumina Co.) according to the manufacturer’s instructions. The 3′ region of each sequence read with a quality score less than 30 was trimmed using BaseSpace (Illumina Co.).
4.11. Microbiome Sequencing Analysis
Raw reads were processed using QIIME2 v.2019.1 with a DADA2 plugin to denoise quality filter reads and call amplicon sequence variants (ASVs), and a feature table of ASV counts was generated. In the quality filtering step, the datasets were truncated to a read length of 270 to 250 base pairs for forward and reverse reads (all other parameters were set to default values). After quality filtering, bacterial taxonomies were assigned to the ASV feature table using the Naïve Bayesian Q2 feature classifier as implemented in QIIME2. Data were compared against a SILVA reference database trained on the V3–V4 region of the 16S rRNA gene. LEfSe analysis was performed to identify taxa displaying the largest differences in abundant microbiota between groups. Only taxa with LDA scores > 2.0 and p < 0.05, as determined by Wilcoxon signed-rank test, are shown. All data analyses were performed using R software v3.6.3.
4.12. Statistical Analysis
Statistical analysis was performed using the unpaired Student’s t-test, if just two independent groups or two conditions on one experimental group were compared. For the determination of any statistical difference between three or more independent experimental groups, one-way analysis of variance (ANOVA) was used. Two-way ANOVA was used when analyzing the mean differences between groups that have been split on two independent variables. All data are presented as mean plus standard error of the mean (SEM) if not stated otherwise in the figure legend. Significance was defined by reaching certain p-values in the statistical tests (* p < 0.05, ** p < 0.01, *** p < 0.001, **** p < 0.0001). All statistical analyses were conducted with the software GraphPad Prism.
4.13. Data Availability
The sequence data for microbiome sequencing were deposited in National Center for Biotechnology Information (NCBI) under the accession number PRJNA615700.