(C) PLOS One
This story was originally published by PLOS One and is unaltered.
. . . . . . . . . .



The activity of the aryl hydrocarbon receptor in T cells tunes the gut microenvironment to sustain autoimmunity and neuroinflammation [1]

['Andrea R. Merchak', 'Department Of Neuroscience', 'University Of Virginia', 'Charlottesville', 'Virginia', 'United States Of America', 'Neuroscience Graduate Program', 'Charlottesville Virginia', 'Center For Brain Immunology', 'Glia']

Date: 2023-02

Multiple sclerosis (MS) is a T cell-driven autoimmune disease that attacks the myelin of the central nervous system (CNS) and currently has no cure. MS etiology is linked to both the gut flora and external environmental factors but this connection is not well understood. One immune system regulator responsive to nonpathogenic external stimuli is the aryl hydrocarbon receptor (AHR). The AHR, which binds diverse molecules present in the environment in barrier tissues, is a therapeutic target for MS. However, AHR’s precise function in T lymphocytes, the orchestrators of MS, has not been described. Here, we show that in a mouse model of MS, T cell-specific Ahr knockout leads to recovery driven by a decrease in T cell fitness. At the mechanistic level, we demonstrate that the absence of AHR changes the gut microenvironment composition to generate metabolites that impact T cell viability, such as bile salts and short chain fatty acids. Our study demonstrates a newly emerging role for AHR in mediating the interdependence between T lymphocytes and the microbiota, while simultaneously identifying new potential molecular targets for the treatment of MS and other autoimmune diseases.

Funding: This work received funding from the National Institutes of Health (R33 MH108156 to A.G., T32 NS115657 to A.R.M., T32 GM008136 to R.M.B.), from the UVA Wagner fellowship (R.M.B.), from the Owens Family Foundation (A.G.) from the UVA Trans University Microbiome Initiative pilot grant (A.G. and A.R.M.) and from the UVA Presidential Fellowship in Neuroscience (C.R.N.). The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.

Here, we explore the role of AHR activity in CD4+ T cells using EAE. We have demonstrated that deleting Ahr from CD4+ T cells increases recovery in EAE in a microbiome-dependent manner. This improved recovery was the result of increased T cell apoptosis after activation in the central nervous system (CNS). While AHR deficiency did not grossly alter the composition of the microbiome, it significantly impacted the production of microbiome-mediated metabolites. In particular, AHR deficiency in T cells led to an increase in bile acids (BAs) and a subset of short-chain fatty acids (SCFAs) that ultimately impacted T cell viability. Our study aimed to understand how a microbiota response element can act in the inverse in the context of autoimmunity. This is the first demonstration, to our knowledge, of a role for AHR in T lymphocytes as a regulator of the microbiome activity that ultimately influences the outcome of CNS autoimmunity. Our discovery builds a foundation for further work that could lead to a microbiome centric approach to dampen the overactive immune system in MS and related autoimmune disorders.

The aryl hydrocarbon receptor (AHR) is a prime candidate for understanding the microbiome-immune interface. The AHR is a cytoplasmic receptor that, when activated, traffics to the nucleus to execute downstream transcriptional programing [ 13 ]. While canonical immune sensors have been primarily characterized based on their response to pathogenic material, the AHR is a homeostatic regulator that is activated by a variety of nonpathogenic exogenous ligands present in barrier tissues. These ligands include indoles [ 14 ], kynurenines [ 15 ], and other small molecules [ 16 ]. Downstream immune effects are dependent on the cell type and ligand. The AHR has also been directly tied to MS. AHR activation is the suspected mechanism by which the novel MS therapeutic, laquinimod, acts [ 17 , 18 ]. It is thought that AHR activation by laquinimod in antigen presenting cells reduces the ratio between inflammatory T cells and regulatory T cells (Treg). Microbiome changes have been identified in AHR null mice and in mice treated with a potent AHR agonist 2,3,7,8-tetrachlorodibenzo-p-dioxin (TCDD), further supporting the connection between AHR, the microbiome, and MS [ 19 – 23 ]. While AHR activity in response to microbiome derived metabolites is well described, a gap of knowledge remains about the impact of AHR expression on the microbiota composition and function.

Multiple sclerosis (MS) is a chronic inflammatory demyelinating disease that affects 2.3 million people worldwide. MS etiology and pathology are linked to both genetic and environmental factors [ 1 ]. A primary environmental factor impacting disease pathology lies with the gut microbiota [ 2 – 5 ]. Indeed, the connection between the microbiota and the immune system in autoimmune disorders such as MS is well established [ 6 ]. The gut flora, composed of bacteria, fungi, and viruses, aids food digestion and maintains pathogen control, but recent studies have also highlighted the integral role of the microbiome in modulation of the homeostatic immune state both locally and systemically [ 7 ]. Gut dysbiosis is a hallmark of disease in both patients with MS and in experimental animal models of MS [ 2 , 5 , 8 – 10 ]. The dysbiotic microbiome can cause earlier onset and increase disease severity in mice [ 10 ]. Furthermore, germ-free mice lacking a microbiome are resistant to experimental autoimmune encephalomyelitis (EAE), an animal model of MS [ 11 ]. These findings spurred the examination of various bacterial supplements in mouse models of MS, and numerous potential candidates have been identified [ 6 , 12 ]. However, the mechanism(s) of action is not well understood. The importance of the microbiome in MS is well described, but a better understanding of the cross-talk between the immune system and the microbiome is required to translate these findings to the clinic. Further, discoveries in the context of MS provide a foundation for understanding environmental factors impacting many other autoimmune disorders.

( A ) Schematic representation of experiments in B and C. In vitro derived T H 17 cells from C57BL6/J mice were treated with a metabolite in addition to mild anti-CD3 stimulation for 24 h. ( B ) Taurocholic acid (10 μm and 100 μm) was administered during anti-CD3 stimulation and apoptosis of T H 17 cells isolated from C57BL6/J mice was measured using Annexin V staining by flow cytometry. (n = 5–6 mice/group; N = 2 experiments; mixed effects analysis with Geisser–Greenhouse correction [p = 0.0070]; Dunnett’s multiple comparison test [Veh vs. 10 μm p = 0.8178, Veh vs. 100 μm p = 0.0233]). ( C ) Administration of isovaleric acid and hexanoic acid (SCFAs that are increased in mice that recover) are sufficient to drive apoptosis in T H 17 cells isolated from C57BL6/J mice. (n = 8 mice/group; N = 2 experiments; RM one-way ANOVA with the Geisser–Greenhouse correction [p = 0.0399]; Dunnett’s multiple comparisons test [Veh vs. 1 mM hexanoic p = 0.0764, Veh vs. 10 mM hexanoic p = 0.0352, Veh vs. 1 mM isovaleric p = 0.0026, Veh vs. 10 mM isovaleric p = 0.0254]). ( D ) Schematic depicting approach to FMT. ( E ) EAE clinical scores of C57BL6/J mice receiving FMT from separately housed Ahr fl/fl or Cd4 cre Ahr fl/fl mice. (Females; representative plot includes n = 11 mice/group; total replicates of N = 2 experiments; Mann–Whitney U test on total scores reported in legend [p = 0.0096]). ( F ) Incidence of clinical sign development in mice treated with FMT. ( G ) Schematic depicting approach to oral supplement of 12.5 mg taurocholic acid/day in C57BL6/J mice immunized for EAE. ( H ) EAE clinical scores of C57BL6/J mice receiving 7 days of oral taurocholic acid (Females; n = 12 mice/group; total replicates of N = 1 experiment; Mann–Whitney U test on single days reported on plot). ( I ) Incidence of clinical sign development in mice treated with taurocholic acid. Error bars represent standard error from the mean. Raw data can be found in Supporting information ( S1 Data ). EAE, experimental autoimmune encephalomyelitis; FMT, fecal material transfer; SCFA, short-chain fatty acid.

The chemical composition of the cecal microenvironment between separately housed Cd4 cre Ahr fl/fl and Ahr fl/fl animals is significantly changed. We next aimed to determine what functional differences arose from the chemical differences both in vitro and in vivo. We first examined the most highly increased bile acid in Cd4 cre Ahr fl/fl mice, taurocholic acid. When administered to C57BL6/J-derived T H 17 cells in vitro, taurocholic acid was sufficient to drive apoptosis similarly to cecal content isolates ( Fig 6A and 6B ). We also examined isovaleric acid, the SCFA, elevated in the absence of AHR. Similarly to taurocholic acid, isovaleric acid could induce apoptosis of in vitro generated T H 17 cells ( Fig 6C ). Similar results were obtained with hexanoic acid, which was also trending higher in Cd4 cre Ahr fl/fl mice ( Fig 6C ). Collectively, our data suggest that the combined effects of the cecal milieu of Cd4 cre Ahr fl/fl mice could induce changes in T cell fitness, ultimately resulting in recovery from paralysis in EAE. As such, we aimed to determine whether this held true in vivo by conducting a fecal material transfer (FMT) to C57BL/6 wild-type mice. The mice were given a cocktail of oral antibiotics for 2 weeks followed by a series of 3 oral gavages of the flushed contents of the intestinal tract from separately housed Cd4 cre Ahr fl/fl and Ahr fl/fl animals. During this time, the animals were also cohoused with an untreated member of the respective genotype to promote lasting engraftment of the microbiome ( Fig 6D ). When these mice underwent EAE, the wild-type mice given FMT from Cd4 cre Ahr fl/fl animals had a similar, but less pronounced reduction in disease severity as the donor mice ( Fig 6E and 6F ). Because the FMT resulted in an intermediate phenotype, we next aimed to determine whether an individual component of the gut microenvironment could act as a therapeutic for EAE. As taurocholic acid was the most efficient driver of apoptosis in vitro ( Fig 6B ), we chose this as our prime candidate. We administered taurocholic acid orally for 7 days after EAE induction ( Fig 6G ). Compared to mice given saline, those receiving taurocholic acid had decreased peak EAE scores ( Fig 6H ) and delayed onset ( Fig 6I ). These data are in support of recent work showing that bile acid receptor agonists and bile acid cocktails can reduce T cell activation and suppress EAE [ 36 , 37 ].

( A ) Partial least-squares plot of LC-MS untargeted metabolomics of the <3 kDa fraction of the cecal contents from Cd4 cre Ahr fl/fl and Ahr fl/fl mice. ( B ) Metaboanalyst mouse KEGG pathways significantly changed by genotype. ( C ) MetaMapp network analysis visualizing the chemically or biologically related significantly changed molecules. ( D ) qPCR of bile acid transporters (mrp3, ostb) and bile acid response elements (ezh2, fxr) in bulk tissue of the duodenum and jejunum of separately housed animals illustrated as relative quantity (R.Q.) from Ahr fl/fl tissue. (n = 6 mice/group; N = 2 experiments; multiple t tests with Welch correction [duodenum p = 0.0104, 0.3591, 0.0071, 0.3828; jejunum p = 0.2230, 0.0623, 0.1487, 0.1004]). ( E ) Quantitative targeted metabolomics for SCFAs show a significant increase in isovaleric acid and a trending increase in valeric and hexanoic acid in Cd4 cre Ahr fl/f mice. (n = 6 mice/group; N = 1 experiment; unpaired t tests [p = 0.0427, 0.1934, 0.1083]) Error bars represent standard error from the mean. Raw data can be found in Supporting information ( S1 , S2 , and S3 Data files). AHR, aryl hydrocarbon receptor; SCFA, short-chain fatty acid.

To identify the impact of CD4+ cell AHR loss on the gut microenvironment, we began by performing 16S sequencing of DNA isolated from fecal pellets and cecal contents of Ahr fl/fl or Cd4 cre Ahr fl/fl mice. Neither cohoused nor separately housed animals saw differences in fecal or cecal microbial signatures ( S4A–S4D Fig ). Given the marked effect of the metabolites prepared from the cecums of Cd4 cre Ahr fl/fl and Ahr fl/fl on T cell apoptosis ( Fig 4F ), we next conducted untargeted metabolomics by LC-MS on these preparations (<3 kDa). Two-dimensional clustering of the annotated products revealed distinct clustering of genotypes, indicating different chemical profiles ( Fig 5A ). This difference occurred despite there being no significant difference in the 16S microbial signatures of the cecal contents suggesting that AHR activity in T cells modify microbial metabolism without altering the microbe composition to a degree that can be captured by 16S sequencing ( S4E and S4F Fig ). Impact analysis of biological pathways identified primary bile acid biosynthesis as the most significantly up-regulated pathway with 6 primary (cholic, taurocholic) or secondary (glycocholic, sulfocholic, dehydrocholic, lawsonic) bile acids significantly increased in the cecal microenvironment of Cd4 cre Ahr fl/fl mice ( Fig 5B and 5C ). Bile acids are dysregulated systemically in patients with MS and adding back primary bile acid is sufficient to reduce the severity of EAE [ 31 ]. Bile acid biosynthesis is known to be dependent on the gut microbiota composition [ 32 , 33 ]. Of the changed bile acids, taurocholic acid was the most highly up-regulated with over a 10-fold increase in Cd4 cre Ahr fl/fl compared to control mice ( Fig 5C ). Increased bile acid accumulation in these mice may be due to the down-regulation of bile acid transporters and response elements in the small intestine of the Cd4 cre Ahr fl/fl mice ( Fig 5D ). In addition to changes in bile acids, alanine, aspartate, and glutamate metabolism were also up-regulated, whereas purine metabolism and tryptophan metabolism were down-regulated in Cd4 cre Ahr fl/fl mice. As SCFAs have been reported to impact the viability of inflammatory T cell and cannot be captured through untargeted metabolomics, we next performed targeted SCFA metabolomics [ 34 , 35 ]. Isovaleric acid was the only SCFA with a significantly higher abundance in the Cd4 cre Ahr fl/fl cecums compared to controls ( Fig 5E ).

( A ) Representative image of spinal cords stained with TUNEL assay and T cell marker (CD3) at the peak of disease. Closed triangles indicate TUNEL+CD3+ cells and arrows indicate TUNEL+CD3- cells. Scale bar = 25 μm. ( B ) Quantification of the number of TUNEL+ T cells (2–6 lesion sites averaged per animal; n = 5–6 mice; N = 2 EAE experiments; unpaired t test [p = 0.0357]). ( C ) T cells isolated from the spinal cord at the peak of disease stained with viability dye and measured by flow cytometry. Cells gated on singlets, CD45+, gate shown on x-axis, then dead cells (n = 5 mice/group; two-way ANOVA [p = 0.0192] with Sidak’s multiple comparison tests [p = 0.0465, 0.2862, 0.6800]). ( D ) T cells skewed to T H 17 in vitro from separately housed Cd4 cre Ahr fl/fl mice stained for apoptotic marker Annexin V 24 h after stimulation with anti-CD3 antibody. (n = 6 mice/group; N = 2 experiments; unpaired t test [p = 0.0133]). ( E ) T H 17 cells from cohoused Cd4 cre Ahr fl/fl and Ahr fl/fl mice show no differences in the number of Annexin V positive cells 24 h after stimulation with anti-CD3 antibody. (n = 5–6 mice/group; N = 2 experiments; unpaired t test [p = 0.5851]). ( F ) In vitro differentiated T H 17 cells from C57BL6/J mice were exposed to the <3 kDa fraction of cecal contents from Ahr fl/fl and Cd4 cre Ahr fl/fl mice for 24 h with anti-CD3 stimulation. (n = 6–9 mice/group; N = 3 experiments; mixed-effects analysis, with Geisser–Greenhouse correction [p = 0.0013]; Tukey’s post hoc analysis [Ctr vs. Fl p = 0.7729; Ctr vs. Cre p = 0.0333, Fl vs. Cre p < 0.0001]). Error bars represent standard error from the mean. Raw data can be found in Supporting information ( S1 Data ). EAE, experimental autoimmune encephalomyelitis.

We next investigated the mechanism behind the reduced T cell numbers in the spinal cords of Cd4 cre Ahr fl/fl mice during chronic EAE. We performed a gene expression analysis on whole spinal cord tissues isolated from two Cd4 cre Ahr fl/fl and two Ahr fl/fl animals at the peak of EAE using a qPCR array composed of 384 genes associated with neuroinflammation. We found that the expression of many genes involved in apoptosis were modulated in the spinal cord tissue of Cd4 cre Ahr fl/fl mice; in particular, we noted increased expression of Fas, GrB, Casp3, and Casp9 and decreased expression of the antiapoptotic transcript of Xiap, indicating an increase in apoptosis in CD4-specific Ahr-deficient mice ( S3A and S3B Fig ). To test if T cells lacking AHR were undergoing apoptosis in EAE, we stained spinal cords sections for CD3 expression by immunofluorescence and TUNEL assay ( Fig 4A ). While we could detect some apoptotic T cells in the spinal cords of control mice subjected to EAE, CD3+TUNEL+ cell numbers were significantly increased in the spinal cords of Ahr-deficient animals (closed triangles; Fig 4B ). We confirmed this observation by quantifying cellular death of isolated immune cells from the spinal cord at the peak of disease using flow cytometry. We observed a significant and specific increase in the percentage of dead CD4+ T cells without change in the percentage of dead CD8+ and or total TCRΒ+ cells between genotypes ( Fig 4C ). To further understand the mechanisms behind these observations, we focused examination on the primary pathogenic subset of CD4+ cells in EAE, the type 17 T helper cell (T H 17) [ 30 ]. To obtain T H 17 cells, we isolated naïve CD4+ cells from the lymph nodes and spleen, then cultured the cells in T H 17 differentiation conditions. Similar to the in vivo observations, Cd4 cre Ahr fl/fl T H 17 cells have a significant increase in apoptosis after mild in vitro stimulation by anti-CD3 as measured by Annexin V staining compared to littermate controls ( Fig 4D ). Increased T H 17 cell death was only observed in cells isolated from Cd4 cre Ahr fl/fl animals separated at weaning, as T cells isolated from cohoused Cd4 cre Ahr fl/fl mice did not display elevated apoptosis ( Fig 4E ). Taken together, these results support the hypothesis that gut microenvironment may be implicated in the Ahr-dependent T cell fitness phenotype. To directly test this hypothesis, we isolated metabolites (<3 kDa) from cecums of separately housed Cd4 cre Ahr fl/fl and Ahr fl/fl mice, and added them to in vitro skewed T H 17 cells prepared from C57BL6/J mice during the 24-h stimulation period. We next analyzed apoptosis by flow cytometry. T H 17 cells exposed to metabolites derived from the gut of Cd4 cre Ahr fl/fl mice had elevated apoptosis compared to cells exposed to metabolites from Ahr fl/fl microbiome ( Fig 4F ). Collectively, these data demonstrate that exposure to the microbial metabolites from Cd4 cre Ahr fl/fl mice is sufficient to induce early T cell apoptosis which likely contributes to recovery in EAE.

( A ) Spinal cords of separately housed Cd4 cre Ahr fl/fl and Ahr fl/fl littermate controls were stained for total immune cells (CD45+) by immunohistochemistry at 4 levels at the peak of disease (day 16 after immunization) (Each dot represents a mouse n = 7 mice/group; N = 2 experiments; two-way ANOVA [p = 0.5336]) ( B) or at the chronic phase of disease (day 31 after immunization). (Two-way ANOVA [p = 0.0419]; Sidak’s multiple comparisons test [p = 0.9896, 0.9209, 0.7898, 0.1753]). ( C ) Spinal cords of separately housed Cd4 cre Ahr fl/fl and Ahr fl/fl littermate controls were stained for T cells (CD3+) by immunohistochemistry at 4 levels at the peak of disease (day 16 after immunization) (n = 7 mice/group; N = 2 experiments, two-way ANOVA [p = 0.0830]) ( D ) or at chronic phase of disease (day 31 after immunization). (n = 7 mice/group; N = 2 experiments, two-way ANOVA [p = 0.0054]; Sidak’s multiple comparisons test [p = 0.9578, 0.9029, 0.7095, 0.0070]). ( E ) Spinal cords of separately housed Cd4 cre Ahr fl/fl and Ahr fl/fl littermate controls were stained for macrophages (CD11b+) by immunohistochemistry at 4 levels at the peak of disease (day 16 after immunization) (n = 4–5 mice/group; N = 2 experiments; two-way ANOVA [p = 0.1271]) ( F ) or at chronic phase of disease (day 31 after immunization). (n = 5 mice/group; N = 2 experiments, two-way ANOVA [p = 0.2269]). ( G ) Gating strategy for flow cytometry analysis of immune cells isolated from dissociated spinal cord tissue ( H ) at peak of disease (day 16 after immunization) (n = 5–6/group; multiple unpaired t tests with Benjamini and Yekutieli corrections [p values listed in the Supporting information ( S1 Data )]), ( I ) or chronic phase of disease (day 31 after immunization) (n = 5 mice/group; multiple unpaired t tests with Benjamini and Yekutieli corrections [p values listed in the Supporting information ( S1 Data )]). Scale bars represent 400 μm. Error bars represent standard error from the mean. Raw data can be found in Supporting information ( S1 Data ).

To determine whether EAE recovery in Cd4 cre Ahr fl/fl was linked to a difference in immune cell number during pathology, we quantified CD45+ immune cells in the spinal cord tissue using immunohistochemistry at the peak and chronic stages of EAE progression ( Fig 3A and 3B ). While the quantity of CD45+ cells was equivalent at the peak of the disease, we observed a decrease in the number of CD45+ cells in the Cd4 cre Ahr fl/fl spinal cord compared to controls at the chronic phase ( Fig 3B ). We next aimed to determine whether this reduction was a result of changes in the number of infiltrating T cells or macrophages. To accomplish this, we quantified expression of CD3, a pan-T cell marker, by immunohistochemistry. We saw no difference in the CD3+ cells in the spinal cord at the peak of disease ( Fig 3C ); however, at the chronic phase, significantly lower CD3+ T cell coverage was observed in the Cd4 cre Ahr fl/fl mice compared to the separately housed littermate controls ( Fig 3D ). Meanwhile, monocytes and macrophages, the other primary effector cells in EAE, showed no differences at either peak or chronic phase as shown by CD11b expression ( Fig 3E and 3F ). We confirmed this outcome using flow cytometry on Percoll-isolated immune cells from the spinal cord. At the peak of disease, we again observed no differences in infiltrating T cells (CD4+ or CD8+), T cell subsets (RORgt+, Tbet+ GATA3+, or FoxP3+), or macrophages (CD11b+) ( Fig 3G and 3H ). At the chronic phase of disease, we found that the loss of T cells was limited to the CD4+ compartment and not the CD8+ compartment. Of the classic CD4 helper cell subtypes, all trended down in Cd4 cre Ahr fl/fl mice with no one subtype constituting a significant decrease, indicating a pan-CD4+ phenotype, not associated with a specific subtype of CD4+ T cells Fig 3I . Taken together, our data suggest that in separately housed mice, EAE recovery observed in Cd4 cre Ahr fl/fl mice is not due to a difference in cytokine expression or T cell differentiation. Instead, recovery correlates with an overall reduction in CD4+ T cell numbers at the chronic phase of disease and suggests an impact for AHR and the microbiome on CD4+ T cell fitness.

( A ) Naïve CD4+ T cells were isolated from cohoused Cd4 cre Ahr fl/fl and littermate control animals and treated to promote differentiation. Flow cytometry gating strategy to measure percent of cells expressing the transcription factors specific to each cell type is shown. ( B ) Representative histogram and quantification of Tbet+ cells after in vitro differentiation of naïve CD4+ cells to T H 1 cell type, (n = 6 biological replicates/group; N = 2 experiments; unpaired t test [p = 0.8715]) ( C ) GATA3 staining of T H 2 differentiated cells, (n = 6 biological replicates/group; N = 2 experiments; unpaired t test [p = 0.9356]), ( D ) RORγt staining of T H 17 cells (n = 8 biological replicates/group; N = 3 experiments; unpaired t test [p = 0.3989]) ( E ), and FoxP3 staining for T regs (n = 8 biological replicates/group; N = 3 experiments; unpaired t test [p = 0.4078]). ( F ) ELISA analysis of culture supernatant after 24 h of anti-CD3 stimulation in T H 1cells (n = 9 biological replicates/group; N = 3 experiments; unpaired t test [p = 0.3924]) ( G ) T H 2 cells (n = 8 biological replicates/group; N = 3 experiments; unpaired t test [p = 0.5917]) ( H ) T H 17 cells (n = 7–8 biological replicates/group; N = 3 experiments; unpaired t test [p = 0.4013]) ( I ) or Treg cells. (n = 8 biological replicates/group; N = 3 experiments; unpaired t test [p = 0.2278]) ( J ) Gating strategy for ( K ) flow cytometry of single-cell suspensions of Peyer’s patches from cohoused Cd4 cre Ahr fl/fl and Ahr fl/fl mice. (Multiple t tests with Welch correction [p = 0.070748, 0.081520, 0.325104, 0.080798]) ( L ) Analysis of cell compositions from mesenteric lymph nodes (multiple t tests with Welch correction [p = 0.127038, 0.179526, 0.444670, 0.118472]) ( M ) and the inguinal lymph nodes from cohoused Cd4 cre Ahr fl/fl and Ahr fl/fl mice (multiple t tests with Welch correction [p = 0.381964, 0.224164, 0.664161, 0.465348]). ( N–S ) Luminex assay on whole spinal cord homogenate of separately housed Cd4 cre Ahr fl/fl and Ahr fl/fl mice at the peak of disease (day 16 after immunization; n = 8 mice/group; N = 2 experiments; unpaired t test), (N) [p = 0.3087], (O) [p = 0.4580], (P) [p = 0.8313], (Q) [p = 0.4941], (R) [p = 0.0215], (S) [p = 0.9333]. Error bars represent standard error from the mean. Raw data can be found in Supporting information ( S1 Data ).

CD4+ T cells are responsible for orchestrating EAE pathology, and defects in these cells have been shown to impact disease outcomes [ 25 ]. Additionally, AHR agonists can modulate T cell skewing in a ligand dependent manner in vivo [ 26 ]. We aimed to explore the role of AHR in T cell differentiation and cytokines production. First, we examined T cell differentiation and the production of cytokines in cohoused Cd4 cre Ahr fl/fl mice. We found that the deletion of AHR in T cells did not impact their capacity for in vitro differentiation as determined by expression of lineage transcription factors by flow cytometry ( Fig 2A–2E ) or production of the signature cytokine for each lineage by ELISA ( Fig 2F–2I ). This was mirrored in separately housed animals ( S2A–S2G Fig ). Furthermore, there were no differences in the immune cell composition between the knockout and control animals at baseline in gut associated lymphoid tissue ( Fig 2J–2M ). We next measured several cytokines important to EAE induction in the spinal cords of mice at the peak of disease. We expected to see significant reduction in the pathological cytokines in the separately housed Cd4 cre Ahr fl/fl mice that ultimately recover, but surprisingly we found minimal differences ( Fig 2N–2S ). The only change was a small reduction in TNFα that may warrant further investigation. These findings suggest that CD4 cell intrinsic AHR activity is not necessary for normal T cell development and cytokine expression supporting published work indicating that AHR activation of the antigen presenting cell may be more important to T cell subset generation in vivo [ 27 – 29 ]. Further, these data reinforce the hypothesis that the T cell-intrinsic reduction of AHR activity in Cd4 cre Ahr fl/fl mice is not responsible for EAE recovery described in separately housed mice.

( A ) Clinical score of Cd4 cre Ahr fl/fl and Ahr fl/fl mice cohoused (females; representative plot includes n = 9 mice/group; total replicates of N = 2 experiments). Spinal cords sections of Cd4 cre Ahr fl/fl and Ahr fl/fl were stained with Luxol fast blue and hematoxylin/eosin stain at day 31 post EAE induction ( B ) Representative images and ( C ) quantification of myelin stain. Images from 4 equally spaced spinal cord levels were averaged for each mouse. (n = 5 mice/group; unpaired t test p = 0.0706) ( D ) Clinical score of Cd4 cre Ahr fl/fl and Ahr fl/fl littermate controls separated at weaning (3 weeks of age). (Females; representative plot includes n = 8–9 mice/group; total replicates of N = 2 experiments; Mann–Whitney U test on total scores reported in legend [p = 0.0096] and on single days reported on plot) ( E ) Luxol fast blue with hematoxylin/eosin stain at the peak stage of EAE (day 16) and at chronic phase (day 31). ( F ) Quantification of myelin stain by Luxol fast blue alone in Ahr fl/fl mice and ( G ) Cd4 cre Ahr fl/fl mice. (n = 7–8 mice/group; N = 2 experiments; unpaired t tests [p = 0.8343, 0.0322]) Scale bars represent 400 μm. Error bars represent standard error from the mean. Raw data can be found in Supporting information ( S1 Data ). EAE, experimental autoimmune encephalomyelitis.

Given the emerging role of the microbiome and the well-accepted contribution of T cells to MS pathology, AHR is ideally positioned to influence CNS autoimmunity. Here, we explored the role of AHR in CD4+ T cells in EAE. For this study, we used the previously validated Cd4 cre Ahr fl/fl and littermate control Ahr fl/fl mouse strains [ 24 ]. Cohoused adult littermate mice were immunized with myelin oligodendrocyte glycoprotein (MOG) 35-55 to induce active EAE. The cohoused mice showed no difference in EAE clinical score when Ahr expression was removed from CD4+ cells ( Figs 1A and S1A , females; S1B and S1C Fig , males). Further, we observed no difference in myelin content in the spinal cord, as revealed by Luxol fast blue staining at the chronic phase of disease ( Fig 1B and 1C ). To assess the role of AHR-driven microbiome changes in EAE, we separated mice by genotype at weaning (3 weeks) and induced EAE between 8 and 16 weeks of age. In this condition, while the time of onset and the peak clinical scores of EAE were not impacted, mice lacking Ahr expression in T cells presented with a significant recovery when compared to wild-type animals ( Figs 1D and S1D , females). Furthermore, the change in recovery does not appear to be driven by sex as it was conserved in both sexes ( S1E and S1F Fig , males). Supporting the clinical scores, the demyelination of the spinal cords was not different between the 2 groups at the peak of disease (day 16); however, we noted a significant increase in myelin staining at the chronic phase (day 31) in the Cd4 cre Ahr fl/fl animals when compared to the control ( Fig 1E–1G ). Together, these data indicate that the lack of Ahr expression in CD4+ cells promotes EAE recovery in a microbiome-dependent manner that can be reversed by cohousing.

Discussion

Here, we demonstrate that the deletion of the AHR in CD4+ cells can promote recovery in chronic EAE while having no impact on the onset nor the initial magnitude of the disease. We further show that this phenotype is dependent on the microbiome, as cohousing littermate Cd4creAhrfl/fl and Ahrfl/fl mice abrogates recovery in EAE. These data are in support of previous work in similar models has shown no difference in EAE recovery after T cell-specific Ahr deletion [38]. At the mechanistic level, our data suggest that specific metabolites elevated in the cecum of Cd4creAhrfl/fl mice reduce T cell fitness and viability. In particular, the cecal environment of Cd4creAhrfl/fl animals is characterized by high levels of isovaleric acid and taurocholic acid, which induce T cell apoptosis in vitro, suggesting a potential mechanism for EAE recovery.

AHR activity can regulate autoimmunity via natural killer (NK) cells [39], macrophages [40], dendritic cells [29], and T cells [26,29,41]—all contributors to EAE pathogenesis. By utilizing CD4 cre mice, we will be targeting all cell types that are affected by CD4 expression at any time-including CD8 T cells that express CD4 during thymic development. AHR activation is a known regulator of T cell differentiation and inflammatory to regulatory T cell ratios. Importantly, T cell activity is dependent on the specific ligand activation with some promoting a proinflammatory environment and others having the opposite effect [26]. It has been postulated that T cell-intrinsic AHR activity is necessary for differentiation but several groups have shown that AHR antagonists do not have a cell-intrinsic effect on T cell differentiation leaving remaining questions in the field [42]. Here, we demonstrate that T cell-specific AHR expression is not necessary for the differentiation of T cells in vitro supporting the hypothesis that the AHR activity in antigen presenting cells is the necessary factor influencing T cell differentiation in vivo [27–29]. Beyond T cells, AHR activation in glial cells regulates the pathogenic effects of microglia and astrocytes in EAE via changed cytokine production [43,44]. AHR’s ubiquitous expression in barrier tissues coupled with its complex ligand-binding profile warrants further study of AHR function in autoimmunity. We have identified that the changes of the luminal microenvironment of the intestine resulting from Ahr knockout in CD4+ cells lead to decreased post-activation lifespan in T cells; however, the impact of these changes on other cell types and other models of autoimmunity have yet to be determined. Though we aimed to target T cells using the CD4 cre line, we acknowledge that other CD4 expressing cells including some gut macrophages may also be contributing to the observed phenotype.

Bile acids, SCFAs, and other small molecules we identified can move into circulation and some can gain entry into the CNS by crossing the blood brain barrier [45]. There are likely systemic effects that have yet to be described. For example, our data suggest that significant remyelination is taking place in the chronic phase of EAE in CD4-specific Ahr knockout animals. Given the described impact of the microbiome and SCFA on myelination [46,47], it would be critical to test if the metabolic signature generated by the lack of Ahr in T cells can influence oligodendrocyte generation.

The data presented here support a bidirectional communication between T cells and the microbiome via AHR. However, it remains to be understood how T cells can modulate the metabolic signature of the cecum. We hypothesize that there are many mechanisms that could be mediating the change. For example, we show that the expression of bile acid transporters is reduced in the intestines of mice with Ahr deletion. This could lead to heightened levels of bile acids that have an impact on microbial viability, metabolism, and community structure. Additionally, it is possible that Ahr deletion in T cells may lead to modulation of IgA secretion by B cells [48,49]. IgA is the primary mechanism by which the adaptive immune system can directly modulate mucosal microbial communities [48]. An alternative hypothesis is that immune cell activity is modulating antimicrobial peptide release indirectly through communication with Paneth cells [50,51]. Most work on these systems thus far has focused on models of severe microbiome dysbiosis. Less is known about the subtler changes in the microbiome subject to homeostatic regulation. Therefore, the mechanism by which the adaptive immune system can interface with the microbiota is of great interest for understanding the basis of autoimmune disorders.

Ahr knockout in T cells results in a change in both primary and conjugated bile acids in the cecum. Bile acids are produced by the host for digestion and are normally reabsorbed by transporters (primary) or passively (primary conjugated) in the ileum while secondary bile acids are not as easily absorbed [52]. There is an increase in conjugated bile acids in Ahr knockout mice, but lower levels of secondary bile acids. This is the inverse of patients with MS indicating that there may be a causative connection in both patients and mice with EAE [31]. Further studies are needed to understand which step of this pathway: production, conversion, or absorption are influenced by the gut flora and/or the T cells. Our results support the existing work in synthetic bile acid receptor agonists showing that T cell fitness and activity can be mediated by bile acid receptor activation [36,37]. As previous groups have shown that T cells traffic to the gut before the CNS during EAE, we hypothesize that exposure to these factors during priming in the lymph node may lead to the early apoptosis after trafficking to the CNS [53]. This evidence that bile acids can have a direct influence on T cell function warrants further study in the context of autoimmune diseases.

Microbiome-modulated immune responses become particularly important in patients with chronic and relapsing autoimmune conditions like MS [54–58]. We have described a novel function for the environmental sensor AHR in T cells as a regulator of the gut microenvironment. As we further understand the complexities of harnessing the gut microbiome, we predict tools for controlling host-derived microbe modulation will be a new frontier for therapeutics. Our work is an important step in the discovery of a targetable endogenous modulator of the microbiome and opens a potential new therapeutic avenue for MS patients.

[END]
---
[1] Url: https://journals.plos.org/plosbiology/article?id=10.1371/journal.pbio.3002000

Published and (C) by PLOS One
Content appears here under this condition or license: Creative Commons - Attribution BY 4.0.

via Magical.Fish Gopher News Feeds:
gopher://magical.fish/1/feeds/news/plosone/

You are viewing proxied material from magical.fish. The copyright of proxied material belongs to its original authors. Any comments or complaints in relation to proxied material should be directed to the original authors of the content concerned. Please see the disclaimer for more details.