Microbiota-Derived Metabolites Suppress Arthritis by Amplifying Aryl-Hydrocarbon Receptor Activation in Regulatory B Cells

Abstract

The differentiation of IL-10-producing regulatory B cells (Bregs) in response to gut-microbiota-derived signals supports the maintenance of tolerance. However, whether microbiota-derived metabolites can modulate Breg suppressive function remains unknown. Here, we demonstrate that rheumatoid arthritis (RA) patients and arthritic mice have a reduction in microbial-derived short-chain fatty acids (SCFAs) compared to healthy controls and that in mice, supplementation with the SCFA butyrate reduces arthritis severity. Butyrate supplementation suppresses arthritis in a Breg-dependent manner by increasing the level of the serotonin-derived metabolite 5-Hydroxyindole-3-acetic acid (5-HIAA), which activates the aryl-hydrocarbon receptor (AhR), a newly discovered transcriptional marker for Breg function. Thus, butyrate supplementation via AhR activation controls a molecular program that supports Breg function while inhibiting germinal center (GC) B cell and plasmablast differentiation. Our study demonstrates that butyrate supplementation may serve as a viable therapy for the amelioration of systemic autoimmune disorders.

Keywords: 5-Hydroxyindole-3-acetic acid; B cells; aryl-hydrocarbon receptor; autoimmunity; butyrate; regulatory B cells; rheumatoid arthritis; serotonin; short chain fatty acid; tryptophan metabolism.

Graphical abstract

Figure 1

Butyrate Levels Correlate with Immature B Cells in Patients with Inactive Rheumatoid Arthritis (A) Representative histograms show stool acetate, propionate, and butyrate levels in healthy controls (HC, n = 20) and RA patients (n = 19) as measured by high-performance liquid chromatography. (B–D) Scatterplots show correlation between stool (B) acetate, (C) propionate, and (D) butyrate levels and CD19⁺CD24^hiCD38^hiB cell frequency in the peripheral blood of RA patients (n = 19). (E–G) Scatterplots show correlation between stool (E) acetate, (F) propionate, and (G) butyrate levels and IL-10⁺B cell frequency in the peripheral blood of RA patients (n = 19). Data represent mean ± SE (A, Mann-Whitney test; B–G, Spearmen correlation). See also Figure S1.

Figure 2

Butyrate Supplementation Suppresses Arthritis by Skewing the B Cell Compartment in Favor of a Regulatory Phenotype (A) Stool butyrate levels in WT mice pre-arthritis (n = 23), with acute arthritis (n = 8), and in remission from arthritis (n = 18) as measured by high-performance liquid chromatography (cumulative data are shown). (B) Mean clinical score of control (cumulative n = 25) and butyrate-supplemented B-WT chimeric mice or B-IL-10^−/−chimeric mice (n = 8 per group) (one representative experiment of two experiments is shown); y axis shows percentage swelling in antigen-injected knee compared to control knee. (C) Mean clinical score of control (cumulative n = 15) and butyrate-supplemented IL-10eGFP reporter mice (cumulative n = 13); y axis shows percentage swelling in antigen-injected knee compared to control knee (one representative experiment of two experiments is shown). (D) Representative H&E staining of knee joints from control and butyrate-supplemented IL-10eGFP reporter mice (left) and blinded histology scores (right) of joint damage. (E) Representative flow cytometry plots (left) and bar charts (right) showing CD19⁺CD21^hiCD24^hiIL-10eGFP⁺Breg frequency and number in control (cumulative n = 15) and butyrate-supplemented mice (cumulative n = 13) (one representative experiment of three experiments is shown). (F) Representative flow cytometry plots (left) and bar charts (right) showing CD19⁺CD138⁺Blimp-1⁺plasmablast frequency and number in control and butyrate-supplemented mice (cumulative n = 11 per group, one representative experiment of two experiments is shown). (G) Bar charts show ratio of CD19⁺CD21^hiCD24^hiIL-10eGFP⁺Bregs to plasmablast in control and butyrate-supplemented mice (cumulative n = 11 per group, one representative experiment of two experiments is shown). (H) Representative flow cytometry plots (left) and bar chart (right) shows the percentage and number of CD19⁺CD95⁺GL7⁺ germinal center (GC) B cells in control and butyrate-supplemented mice (cumulative n = 11 per group, one representative experiment of three experiments is shown). (I) Bar chart shows ratio of CD19⁺CD21^hiCD24^hiIL-10eGFP⁺Bregs to GC B cells in control and butyrate-supplemented mice (cumulative n = 11, one representative experiment of two experiments is shown). (J) Representative immunofluorescence blinded histological analysis of the number and size of GC control and butyrate-supplemented mice (original magnification 20×, n = 3). (K) Mean clinical score following transfer of CD19⁺CD21^hiCD24^hiIL-10eGFP⁺Bregs from control (cumulative n = 6) or butyrate-supplemented mice (cumulative n = 6), a control group that did not receive a transfer; y axis shows percentage swelling in antigen-injected knee compared to control knee (cumulative n = 8) (one representative experiment of two experiments is shown). Cells were isolated at day 7 post-disease onset. Data represent mean ± SE (A, one-way ANOVA; B, C, and K, two-way ANOVA; D–J, Student’s t test). See also Figures S2–S4.

Figure 3

Suppression of Arthritis by Butyrate Supplementation Depends upon AhR Activation and IL-10 Expression in B Cells (A) Bar chart shows expression of Cyp1a1 relative to β-actin in splenic B cells isolated from control or butyrate-supplemented mice (cumulative n = 5, one representative experiment of two experiments is shown). (B) Mean clinical score of control and butyrate-supplemented Mb1^cre/+ mice or Ahr^fl/−Mb1^cre/+ mice; y axis shows percentage swelling in antigen-injected knee compared to control knee (cumulative n = 15 per group, one representative experiment of five experiments is shown). (C) IL-10 production by splenic B cells isolated from control Mb1^cre/+ mice, butyrate-supplemented Mb1^cre/+ mice, control Ahr^fl/−Mb1^cre/+ mice, and butyrate-supplemented Ahr^fl/−Mb1^cre/+ mice at day 7 post-disease onset as measured by ELISA (cumulative n = 3 per group). (D) Representative flow cytometry plots and bar charts showing the frequency and number of CD19⁺CD21^hiCD24^hiB cells in control Mb1^cre/+ mice (cumulative n = 8), butyrate-supplemented Mb1^cre/+ mice (cumulative n = 5), control Ahr^fl/−Mb1^cre/+ mice (n = 7), and butyrate-supplemented Ahr^fl/−Mb1^cre/+ mice (cumulative n = 6) at day 7 post-disease onset (cumulative data are shown). (E) Representative flow cytometry plots and bar charts showing the frequency and number of CD19⁺CD138⁺Blimp-1⁺B cells in control Mb1^cre/+ mice (cumulative n = 8), butyrate-supplemented Mb1^cre/+ mice (cumulative n = 5), control Ahr^fl/−Mb1^cre/+ mice (cumulative n = 7), and butyrate-supplemented Ahr^fl/−Mb1^cre/+ mice (cumulative n = 6) (cumulative data are shown). (F) Representative flow cytometry plots and bar charts showing the frequency and number of CD19⁺CD95⁺GL7⁺B cells in control Mb1^cre/+ mice (cumulative n = 8), butyrate-supplemented Mb1^cre/+ mice (cumulative n = 5), control Ahr^fl/−Mb1^cre/+ mice (cumulative n = 7), and butyrate-supplemented Ahr^fl/−Mb1^cre/+ mice (cumulative n = 6) (cumulative data are shown). Cells were isolated at day 7 post-disease onset. Data represent mean ± SE (A, Student’s t test; C, E, and F, one-way ANOVA; B, two-way ANOVA). See also Figures S5 and S6.

Figure 4

Butyrate Supplementation Modulates the Transcriptional and Epigenetic Landscape of CD19⁺CD21^hiCD24^hiB Cells in an AhR-Dependent Manner (A) Volcano plots shows log₂ fold change (FC) in gene expression between CD19⁺CD21^hiCD24^hiB cells isolated from butyrate-supplemented Mb1^cre/+ mice compared to control Mb1^cre/+ mice (top plot) and between butyrate supplemented Ahr^fl/−Mb1^cre/+ compared to control Ahr^fl/−Mb1^cre/+ mice (bottom plot). Red dots represent significant DEGs, with the red line denoting a cut off p value of <0.05. (B) Signaling pathway impact analysis (SPIA) ranked on significance (pG) comparing the over-represented (red) and under-represented (blue) pathways in butyrate-supplemented compared to control CD19⁺CD21^hiCD24^hiB cells from Mb1^cre/+ mice (top graph) and Ahr^fl/−Mb1^cre/+ mice (bottom graph). The total perturbation accumulation (tA) score is listed for the “protein processing in endoplasmic reticulum” pathway. (C) Heatmap shows the expression of B cell differentiation genes in CD19⁺CD21^hiCD24^hiB cells isolated from control Mb1^cre/+ mice, butyrate-supplemented Mb1^cre/+ mice, control Ahr^fl/−Mb1^cre/+ mice, and butyrate-supplemented Ahr^fl/−Mb1^cre/+ mice. Mean z scores were calculated from log CPM values. Samples highlighted in red are significantly differentially expressed between CD19⁺CD21^hiCD24^hiB cells isolated from butyrate-supplemented Mb1^cre/+ mice compared to butyrate-supplemented Ahr^fl/−Mb1^cre/+ mice. Samples highlighted in bold are significantly differentially expressed between CD19⁺CD21^hiCD24^hiB cells isolated from butyrate-supplemented Mb1^cre/+ mice compared to control Mb1^cre/+ mice. (D) Representative ATAC-seq tracks for the Bcl6 and Gpr183 loci in CD19⁺CD21^hiCD24^hiB cells from butyrate-supplemented or control Mb1^cre/+ and Ahr^fl/−Mb1^cre/+ mice (n = 3). Track heights between samples are normalized through group autoscaling. For RNA-seq data, n = 3 per condition and genotype. (E) Heatmap shows inferred transcription factor activity scores based on accessibility at transcription factor binding motifs in CD19⁺CD21^hiCD24^hiB cells isolated from control Mb1^cre/+ mice, butyrate-supplemented Mb1^cre/+ mice, control Ahr^fl/−Mb1^cre/+ mice, and butyrate-supplemented Ahr^fl/−Mb1^cre/+ mice as measured by ATAC-seq. AhR co-factors are highlighted in red. For ATAC-seq data, n = 3 for Mb1^cre/+ mice and n = 2 for Ahr^fl/−Mb1^cre/+ mice. For RNA-seq data, n = 3 per group. Cells were isolated at day 7 post-disease onset. See also Figure S7.

Figure 5

CD45.2⁺CD19⁺CD21^hiCD24^hiB Cells from Butyrate-Supplemented WT but Not AhR^−/− Mice Retain Their Phenotype and Differentiate in IL-10⁺Bregs upon Adoptive Transfer (A and B) Representative flow cytometry plots show (A) CD45.2⁺CD19⁺B cell and (B) CD45.2⁺CD19⁺CD21^hiCD24^hi B cell frequency in CD45.1 congenic WT mice that had received a transfer of CD19⁺CD21^hiCD24^hiB cells isolated from control or butyrate-supplemented WT or AhR^−/− mice. (C) Bar chart shows number of CD45.2⁺CD19⁺CD21^hiCD24^hiB cells recovered post-transfer from CD45.1 congenic WT mice that had received a transfer of CD19⁺CD21^hiCD24^hiB cells isolated from control or butyrate-supplemented WT or AhR^−/− mice. (D) Representative flow cytometry plots and bar charts show CD45.2⁺CD19⁺IL-10⁺B cell frequency in CD45.1 congenic WT mice that had received a transfer of CD19⁺CD21^hiCD24^hiB cells isolated from control or butyrate-supplemented WT or AhR^−/− mice. (E) Bar chart shows number of CD45.2⁺CD19⁺IL-10⁺B cells recovered post-transfer from CD45.1 congenic WT mice that had received a transfer of control or butyrate-supplemented WT or AhR^−/− mice. Cells were isolated at 48 h post-transfer (cumulative n = 3 per group, cumulative data are shown). Data represent mean ± SE (C and E, one-way ANOVA).

Figure 6

Butyrate Supplementation Increases the Availability of AhR Ligands (A) Mean clinical score of control and butyrate-supplemented ABX-treated or untreated mice; y axis shows percentage swelling in antigen-injected knee compared to control knee (cumulative n = 8 per group, one representative experiment of two experiments is shown). (B) Bar chart shows expression of Il10 relative to β-actin in splenic B cells isolated from ABX-treated WT or untreated mice (cumulative n = 3 per group). (C) Bar chart shows relative abundance of bacterial phyla in the stool of naive, control arthritic, or butyrate-supplemented arthritic mice (n = 4 per group). (D) XY graph shows operational taxonomic units (OTUs) of bacterial genera in butyrate-supplemented and control arthritic mice (n = 4 per group). (E–G) Bar charts shows levels of tryptophan, tryptamine, indole (E), L-Kynurenine, Kynurenic Acid (KYNA) (F), and 5-HIAA (G) in the stool of control arthritic WT and butyrate-supplemented arthritic mice (cumulative n = 5 per group). Data represent mean ± SE (A, two-way ANOVA; B, one-way ANOVA; E–G, Student’s t test). See also Figure S7.

Figure 7

5-Hydroxyindole-3-Acetic Acid Increases Il10 Transcription by B cells In Vivo and In Vitro by Acting as a Ligand for AhR (A) Relative expression of Cyp1a1 and Il10 in total splenic B cells following 6-h culture with 5-HIAA or kynurenic acid (KYNA) compared to vehicle alone (n = 3, one of two experiments is shown). (B) Mean clinical score of control, 5-HIAA-gavaged, or KYNA-gavaged mice; y axis shows percentage swelling in antigen-injected knee compared to control knee (cumulative n = 8 per group, one representative experiment of two experiments is shown). (C) Bar charts show expression of Cyp1a1 and Il10 relative to β-actin in splenic B cells isolated from control, 5-HIAA-gavaged, or KYNA-gavaged mice. (D) Mean clinical score of control or 5-HIAA-gavaged Mb1^cre/+ mice or Ahr^fl/−Mb1^cre/+ mice; y axis shows percentage swelling in antigen-injected knee compared to control knee (cumulative n = 8 per group, one representative experiment of two experiments is shown). (E) Mean clinical score of control and butyrate-supplemented L-para-chlorophenylalanine (PCPA)-treated (tryptophanase inhibitor, TPH) or vehicle-treated mice; y axis shows percentage swelling in antigen-injected knee compared to control knee (cumulative n = 10 per group, one representative experiment of two experiments is shown). Data represent mean ± SE (A, Student’s t test; B, two-way ANOVA; C, one-way ANOVA; D and E, two-way ANOVA).