Protective role of colitis in inflammatory arthritis via propionate-producing Bacteroides in the gut

Abstract

Objectives: To investigate whether and how inflammatory disease in the intestine influences the development of arthritis, considering that organ-to-organ communication is associated with many physiological and pathological events.

Methods: First, mice were given drinking water containing dextran sodium sulfate (DSS) and then subjected to inflammatory arthritis. We compared the phenotypic symptoms between the cohoused and separately-housed mice. Next, donor mice were divided into DSS-treated and untreated groups and then cohoused with recipient mice. Arthritis was then induced in the recipients. The fecal microbiome was analyzed by 16S rRNA amplicon sequencing. We obtained type strains of the candidate bacteria and generated propionate-deficient mutant bacteria. Short-chain fatty acids were measured in the bacterial culture supernatant, serum, feces, and cecum contents using gas chromatography-mass spectrometry. Mice fed with candidate and mutant bacteria were subjected to inflammatory arthritis.

Results: Contrary to expectations, the mice treated with DSS exhibited fewer symptoms of inflammatory arthritis. Intriguingly, the gut microbiota contributes, at least in part, to the improvement of colitis-mediated arthritis. Among the altered microorganisms, Bacteroides vulgatus and its higher taxonomic ranks were enriched in the DSS-treated mice. B. vulgatus, B. caccae, and B. thetaiotaomicron exerted anti-arthritic effects. Propionate production deficiency further prevented the protective effect of B. thetaiotaomicron on arthritis.

Conclusions: We suggest a novel relationship between the gut and joints and an important role of the gut microbiota as communicators. Moreover, the propionate-producing Bacteroides species examined in this study may be a potential candidate for developing effective treatments for inflammatory arthritis.

Keywords: Bacteroides; arthritis; colitis; gut microbiota; propionate.

Figure 1

Protective effects of colitis on CFA-induced arthritis. (A–C) Mice (n = 5 or 4 per group) were provided with drinking water with or without DSS for 5 days. CFA and IFA (20 μL) were then subcutaneously injected into each footpad. The swelling at the footpads and ankles was measured at indicated time points after the injections (C) The correlation between the DAI score and swelling at footpads and ankles of the mice on day 9 among data in the (B). (D) Mice (n = 5 per group) were subcutaneously treated with CFA and IFA, and 2 days later, given drinking water containing DSS or fresh water for 5 days (marked as a bar). The concentration of DSS was 2.5% (A, D) and Low DSS, Int. DSS, and High DSS indicated 2, 2.5, and 3% concentration, respectively (B). The results are representative of at least two independent experiments. The data are shown as means ± SEM. *P < 0.05 and **P < 0.01 with the Mann–Whitney test (A, D). *P < 0.05 (between the vehicle and DSS groups or high DSS groups) and #P < 0.05 (between the vehicle and intermediate DSS groups) with the Mann–Whitney test (B).

Figure 2

Gut microbiota contributes to colitis-mediated protection from arthritis. (A) Mice (n = 5 per group) were provided with fresh water or 2.5% DSS water bottles for 5 days. CFA and IFA were then subcutaneously injected into the footpads of all the mice. The control and colitis groups were mixed in the same cages or raised separately in the divided cages. The swelling at the footpads and ankles was then assessed. (B–E) Recipient mice (n = 6 per group) were given water bottles containing an antibiotic cocktail for 2 weeks before cohousing. Donor mice (n = 8 per group) were provided with fresh water or 2.5% DSS water bottles for 5 days before cohousing. The donor mice were cohoused with the recipient mice and, 2 days later, CFA and IFA were subcutaneously injected into each footpad of the recipient mice. (B) Experimental scheme. (C) Beta-diversity (non-metric multidimensional scaling [NMDS] plot) of the microbial communities on day 4. Each dot represents an individual mouse. (D) The swelling at the footpads and ankles of the recipient mice. (E) Representative macroscopic images of the CFA-injected limbs. The results are representative of at least two independent experiments. The data are shown as means ± SEM. *P < 0.05 and **P < 0.01 (between separately-housed groups) or #P < 0.05 and ##P < 0.01 (between cohoused groups) with the Mann–Whitney test (A). *P < 0.05 and **P < 0.01 with the Mann–Whitney test (D).

Figure 3

Gut microbial candidates involved in the colitis-mediated protection from arthritis. The gut microbiome obtained from the recipients in Figure 2B was analyzed. (A) Beta-diversity (NMDS plot) of the microbial communities of the recipient mice. Each dot represents an individual mouse. (B) Relative abundance of bacterial taxa at the genus level. The top seven taxa are listed in the right legend. (C) LEfSe analysis of the recipients cohoused with the vehicle and DSS-induced colitis groups (logarithmic LDA score > 4). Significant differences in taxa were obtained using the Kruskal–Wallis test (α < 0.05). (D) Among the bacterial taxa selected from the LEfSe analysis, the relative abundance of the family Bacteroidaceae, genus Bacteroides, and species Bacteroides vulgatus in each taxonomic rank are shown. Each dot represents an individual mouse and the means are displayed as a line. *P < 0.05 and **P < 0.01 with the Mann-Whitney test.

Figure 4

B. vulgatus protects from CFA-induced arthritis. Mice (n = 9 or 10 per group) were given drinking water containing an antibiotic cocktail for 14 days. After the cessation of the antibiotics, the mice were repeatedly administered B. vulgatus at indicated time points. Four days after the cessation of the antibiotics, CFA and IFA were injected into each footpad. (A) Experimental scheme. (B) The swelling at the footpads and ankles. The data are shown as means ± SEM. Representative macroscopic images (C) and H&E staining image (D) of CFA-injected limbs. Scale bar, 2 mm. The results are representative of at least two independent experiments. *P < 0.05 and **P< 0.01 with the Mann–Whitney test.

Figure 5

Production of short-chain fatty acids. (A) The amount of SCFAs secreted by B. vulgatus (2 × 10⁹ CFU/mL) during cultivation in BHIS broth. **P < 0.01 by an unpaired t-test. (B, C) Mice (n = 9) were provided with drinking water with DSS for 5 days, which was then replaced with fresh water. (B) The amount of SCFAs in the feces collected at indicated time points. The data are shown as means ± SEM. *P < 0.05, ***P < 0.001 and ****P < 0.0001 by a one-way RM ANOVA. (C) The amount of SCFAs in the serum collected before DSS-treatment and after 12 days of DSS-treatment. Each dot represents an individual mouse. *P < 0.05 and **P < 0.01 with a paired t-test.

Figure 6

Other propionate-producing Bacteroides and mutant form and their protective effects on CFA-induced arthritis. The amount of SCFAs secreted by B. caccae (A) and B. thetaiotaomicron (WT and mutant Δ1686-1689) (B) (2 × 10⁹ CFU/mL) during cultivation in BHIS broth. (C) Mice (n = 5 per group) were intragastrically administered B. caccae and B. thetaiotaomicron (WT and mutant Δ1686-1689) and treated with CFA/IFA according to the same scheme in Figure 4B . The data are shown as means ± SEM. *P < 0.05 and **P < 0.01 (B. caccae to vehicle) or #P < 0.05 and ##P < 0.01 (B. thetaiotaomicron WT to vehicle) with the Mann–Whitney test.