Background: The mucosal origins hypothesis suggests rheumatoid arthritis (RA) is triggered at mucosal sites in genetically predisposed hosts[1]. Murine models of inflammatory arthritis support that intestinal microbiota-induced T helper 17 (Th17) cells are involved in the pathogenesis of RA[2] and other rheumatic diseases[3,4]. Dietary interventions are known to modulate human gut microbial community structures and host immunity. Specifically, a ketogenic diet in patients with metabolic disease was shown to decrease gut Th17 cells in mice by reducing bifidobacteria via ketone bodies[5]. However, which host-microbiota interactions are involved in fasting of patients with RA, that improves disease activity clinically[6], are currently unknown.

Objectives: We set out to characterize the gut microbial communities, key metabolites, and host-microbiota interactions with a focus on type 17 immunity in a cohort of patients with RA who underwent fasting.

Methods: Stool and serum samples were obtained from a subset of patients with RA enrolled in the NutriFast study[6]. 10 patients fasting (~300 Kcal) for 7 days and 8 patients under an anti-inflammatory diet were analyzed with regards to microbiomes, metabolomes, and clinical parameters. Fecal microbiota composition was defined by full-length 16S rDNA sequencing on the PacBio platform. Diet-sensitive taxa were profiled longitudinally by quantitative PCR. A cytokine-bead array assay was used to measure cytokine levels in the serum. Beta-hydroxybutyrate (BHB) was determined via dried blood spot metabolomics. Candidate taxa were isolated from the stool of patients with RA by culture (Bifidus Selective Medium Agar) and co-cultured with PBMCs from healthy individuals to determine human Th17 induction in vitro in the presence of CD3/CD28 beads or IL-2. Uncorrected Fisher´s LSD was used to compare baseline and intervention among both groups and Mann Whitney U test to compare BHB or bacterial abundances between groups.

Results: One week of fasting significantly reduced the HAQ score (p = 0.019) and circulating lymphocyte counts (p = 0.011) of patients with RA, while increasing ketone bodies (BHB; p < 0.0001) compared to an anti-inflammatory diet. The gut microbiota composition was also different between groups. In particular, Bifidobacterium adolescentis (BA) was the most strongly reduced species after seven days of fasting compared to the control group, which was confirmed by a species-specific qPCR analysis (p = 0.030). The absence of BA after fasting was modestly associated with higher circulating levels of BHB (p = 0.095). Independently of the dietary regimens, BA+ patients with RA displayed higher serum levels of IL-17 compared to patients not colonized with this species (p = 0.0219). In vitro studies on human PBMCs revealed that a BA isolate from an RA patient prior to fasting and the type strain of BA both induce IL-17/IFNg cytokine co-production in human CD3+ T cells as typical for pathogenic Th17 cells.

Conclusion: A fasting intervention for seven days in patients with RA altered the gut microbiota compared to an anti-inflammatory diet. Fasting decreased BA, a species which was able to induce human IL-17-producing T lymphocytes in vitro . BA was previously shown to induce intestinal Th17 in mice[5], exacerbate experimental arthritis[7], and disrupt epithelial barrier function[8]. Altogether, our results put forth a diet-sensitive candidate pathobiont in human RA that warrants further study. They also delineate a host-microbiota interaction that may partly explain the rapid, beneficial effects of fasting in patients with RA.

REFERENCES: [1] Holers et al., 2018 Nat Rev Rheumatol 14(9), 542-557.

[2] Evans-Marin et al., 2018 Arthritis Rheumatol 70(12), 1971-1983.

[3] Manfredo Vieira et al., 2018 Science 359(6380), 1156-1161.

[4] Gronke et al., 2023 bioRxiv.

[5] Ang et al., 2020 Cell, 181(6), 1263-1275 e1216.

[6] Hartmann et al, 2022 Front Nutr, 9, 1030380.

[7] Tan et al, 2016 Proc Natl Acad Sci U S A, 113(50), E8141-E8150.

[8] Bootz-Maoz et al., 2022, Cell Reports 41, 111657.

Acknowledgements: Mengwei Niu for stool sample preparations.

Disclosure of Interests: Márcia S. Pereira: None declared, Katja Stuhlträger: None declared, Natalie Effelsberg: None declared, Anika Rajput Khokhar: None declared, Sylvio Redanz: None declared, Hebah Ebid: None declared, Bérénice Hansen: None declared, Cédric C. Laczny: None declared, Ulrike Löschberger: None declared, Stefan Bletz: None declared, Jochen G. Schneider: None declared, Paul Wilmes: None declared, Christan S. Kessler: None declared, Andreas Michalsen: None declared, Alexander Mellmann: None declared, Martin A. Kriegel Novartis, GSK, Roche, Genentech, MSD, AstraZeneca, Roche, Roche, BMS, Eligo Biosciences, Enterome, Maat Pharma, Roche, AbbVie, BiomX.