0598: An Arthritogenic Strain of Subdoligranulum Specifically Detectable in the Feces of Individuals At-risk for and with Rheumatoid Arthritis Is Bound by ACPA and Stimulates Th17 Cell Activation in Those with Rheumatoid Arthritis

Meagan Chriswell¹, Cliff Rims², Megan Maerz², Alex Hsu³, Jennifer Seifert⁴, Marie Feser⁵, Michelle Bloom³, Elizabeth Bemis⁶, Kristen Demoruelle⁵, Kevin D Deane⁷, William Robinson⁸, Eddie James⁹, Jane Buckner⁹, V. Michael Holers¹⁰ and Kristi Kuhn⁵, ¹UC Denver SOM, Aurora, CO, ²Benaroya Research Institute, Seattle, WA, ³Stanford University, Stanford, CA, ⁴University of Colorado, Littleton, CO, ⁵University of Colorado Anschutz Medical Campus, Aurora, CO, ⁶Colorado School of Public Health Anschutz Medical Campus, Aurora, CO, ⁷University of Colorado Denver Anschutz Medical Campus, Denver, CO, ⁸Stanford University School of Medicine, Palo Alto, CA, ⁹Benaroya Research Institute at Virginia Mason, Seattle, WA, ¹⁰University of Colorado, Denver, CO

Background/Purpose: We previously demonstrated that intestinal bacteria within the families Lachnospiraceae and Ruminococcaceae are preferentially targeted by ACPA-reactive plasmablast-derived monoclonal antibodies (PB-mAbs) from dual IgA/IgG isotype families derived from individuals at-risk for RA. Furthermore, intestinal colonization of mice with a PB-mAb reactive primary Ruminococcaceae genus Subdoligranulum strain, designated strain 7, from an at-risk individual drove Th17 cell responses as well as serum RA-related autoantibodies and inflammatory joint swelling in monocolonized mice in the absence of other arthritogenic factors. Development of arthritis required B and T cells in mice, and could be transferred with serum to naïve mice. Given these findings, we hypothesized that the presence of this arthritogenic Subdoligranulum strain 7 is specific to those at risk for and with RA and that a specific anti-strain 7 immune response is found in human RA.

Methods: Feces from individuals at-risk for (n=12, defined as being serum anti-CCP3 positive) and with early RA (n=12, < 1 year from diagnosis), as well as healthy controls (n=12), were screened for the presence of Subdoligranulum strain 7 by a strain-specific qPCR with a sensitivity of 0.0002% abundance in feces. Human PBMCs from individuals with early RA (n=11) and healthy controls (n=12) were exposed to strains 1 and 7 as well as influenza, and T cell activation based on CD69 and CD154 upregulation were evaluated by flow cytometry. CD45RA-/CXCR3+/CCR4-/CCR6- cell markers were used as to define Th1 cells while CD45RA-/CXCR3-/CCR4+/CCR6+ were used for Th17 cells.

Results: Subdoligranulum strain 7 was detectable in the feces at a mean abundance 1.637 x 10⁷± 3.059 x 10⁷ CFUs in ~16% of individuals at-risk for and with early RA, but not detected in healthy individuals (Table 1). ACPA-reactive PB-mAbs bound strains 1 and 7, and digestion of cell wall associated proteins with 100 µg/mL Proteinase K decreased binding by 48%. In addition, informative PB-mAbs could immunoprecipitate a molecule of 10 kD from a cell wall-enriched fraction in strain 7 but not strain 1. Finally, strain 7 but not strain 1 induced T cell activation in human PBMCs from individuals with RA, but not controls (Fig. 1, P< 0.01). This response was HLA class II dependent and skewed towards a Th17 phenotype, as compared to influenza that induced a Th1-like response (P=0.0078, Fig. 2).

Conclusion: The preferential presence of, and immune responses against, a primary human-derived strain of Subdoligranulum in individuals at-risk for or with early RA, but not in controls, supports the conclusion that this is an arthritogenic bacterium of relevance in a subset of RA risk and patient populations. The finding of PB-mAb-reactive antigen associated with the cell wall fraction of Subdoligranulum, suggests that a cross-reactive protein may be stimulating systemic adaptive immunity in RA through molecular mimicry. Taken alongside our murine findings, these data implicate Subdoligranulum strain 7 as a candidate mucosal driver of RA development in a subset of individuals.


Fig. 1: PBMCs from RA cases (n=11) or controls (n=12) were stimulated with 50ng/ml strain 7. Fold change of the CD4 T cell activation based on CD154 and CD69 surface expression relative to DMSO is displayed. (p=0.0080, nonparametric Mann-Whitney test)

Fig 2.: Using CD45RA-/CXCR3+/CCR4-/CCR6- (Th1) or CD45RA-/CXCR3-/CCR4+/CCR6+ (Th17) as a defining marker combinations, we compared the relative proportion of Th1- and Th17-like cells for strain 7 specific (circles) or influenza specific (inverted triangles) memory T cell responses in T cells from RA cases (p=0.0078, nonparametric Mann-Whitney test).
Disclosures: M. Chriswell, None; C. Rims, None; M. Maerz, None; A. Hsu, None; J. Seifert, None; M. Feser, None; M. Bloom, None; E. Bemis, None; K. Demoruelle, Boehringer-Ingelheim, Pfizer; K. Deane, Werfen; W. Robinson, None; E. James, Janssen, Provention Bio, Bristol-Myers Squibb(BMS), Novartis; J. Buckner, Amgen, Bristol Myers Squibb, Gentiobio, Hot Spot Therapeutics, Janssen, Pfizer, Novo Nordisk, Allen Institute for Immunology, Type 1 Diabetes TrialNet Study Group, La Jolla Institute, Oklahoma Medical Research Foundation, Bristol Myers Squibb Immunology, Colton Center for Autoimmunity at Penn, Board of Scientific Counsellors; V. Holers, Janssen; K. Kuhn, None.