Contribution of bacterial pathogens to evoking serological disease markers and aggravating disease activity in rheumatoid arthritis

Abstract

Commensal bacteria and their pathogenic components in the gastrointestinal tract and oral cavity may play pathological roles in autoimmune diseases. To study the possible involvement of bacterial pathogens in autoimmune diseases, IgG and IgA antibodies against pathogenic components produced by three strains of commensal bacteria, Escherichia coli-lipopolysaccharide (E. coli-LPS), Porphyromonas gingivalis-LPS (Pg-LPS) and peptidoglycan polysaccharide (PG-PS) from Streptococcus pyogenes, were determined by an improved ELISA system for sera from two groups of patients with rheumatoid arthritis (RA), who met rapid radiographic progression (RRP) criteria and non-RRP, and compared to normal (NL) controls. Antibody responses to these bacterial pathogens are unique and consistent in individuals, and no fundamental difference was observed between RA and NL controls. Despite the similar antibody responses to pathogens, lower IgG or higher IgA and consequent higher IgA/IgG antibody ratio among the patients with RA related to disease marker levels and disease activity. Peculiarly, the IgA/IgG anti-Pg-LPS antibody ratio resulted from lower IgG and higher IgA antibody responses to Pg-LPS strongly correlated not only with rheumatoid factor (RF), but also correlated with erythrocyte sedimentation rate (ESR), C-reactive protein (CRP) and disease activity score of 28 joints with ESR (DAS28-ESR) in the RRP group. In contrast, the IgA/IgG anti-E. coli-LPS and anti-PG-PS antibody ratio correlated or tended to correlate with RF, ESR, CRP, and DAS28-ESR in the non-RRP group, whereas either the IgG or IgA anti-Pg-LPS antibody levels and consequent IgA/IgG anti-Pg-LPS antibody ratio did not correlate with any clinical marker levels in this group. Notably, anti-circular-citrullinated peptide (CCP) antibody levels, which did not correlate with either IgG or IgA antibody levels to any pathogens, did not correlate with severity of arthritis in both RRP and non-RRP. Taken together, we propose that multiple environmental pathogens, which overwhelm the host antibody defense function, contribute independently or concomitantly to evoking disease makers and aggravating disease activity, and affect disease outcomes.

Trial registration: UMIN CTR UMIN000012200.

Fig 1. Comparison of antibody responses to potential pathogenic environmental agents between RA and NL controls.

IgG and IgA antibody levels against E. coli-LPS, Pg-LPS and PG-PS were determined in sera from 38 NL controls, 54 patients with RRP and 101 patients with non-RRP (a). IgG and IgA antibody levels of individual patients were divided by the average values of NL controls, and shown as IgG and IgA index values (b). Data are shown as median and interquartile range (IQR). E. coli: Escherichia coli, Pg: Porphyromonas gingivalis, LPS: lipopolysaccharide, PG-PS: peptidoglycan polysaccharide from Streptococcus pyogenes, RRP: rapid radiographic progression, NOTE: Index 1: sum of anti-E. coli-LPS + anti-Pg-LPS, Index 2: sum of anti-E. coli-LPS + anti-PG-PS, Index 3: sum of anti-E. coli-LPS + anti-Pg-LPS + anti-PG-PS.

Fig 2. Linkage of RF with IgG and IgA antibody responses to bacterial pathogens in RA patients.

IgG and IgA antibody levels against individual pathogens and their IgA/IgG antibody ratio were analyzed for possible correlation with RF levels in 54 patients with RRP (a) and 101 patients with non-RRP (b) by Spearman’s rank correlation coefficient analysis. NOTE: Pink: significant correlation at p<0.05, Blue: trending toward correlation at 0.05≤p<0.15, No color: no correlation.

Fig 3. Linkage of ESR with IgG and IgA antibody responses to bacterial pathogens in RA patients.

IgG and IgA antibody levels against individual pathogens and their IgA/IgG antibody ratio were compared with ESR in 54 patients with RRP (a) and 101 patients with non-RRP (b) by Spearman’s rank correlation coefficient analysis. NOTE: Pink: significant correlation at p<0.05, Blue: trending toward correlation at 0.05≤p<0.15, No color: no correlation.

Fig 4. Linkage of CRP with IgG and IgA antibody responses to bacterial pathogens in RA patients.

IgG and IgA antibody levels against individual pathogens and their IgA/IgG antibody ratio were compared with CRP levels in 54 patients with RRP (a) and 101 patients with non-RRP (b) by Spearman’s rank correlation coefficient analysis. NOTE: Pink: significant correlation at p<0.05, Blue: trending toward correlation at 0.05≤p<0.15, No color: no correlation.

Fig 5. Linkage of DAS28-ESR with IgG and IgA antibody responses to bacterial pathogens in RA patients.

IgG and IgA antibody levels against individual pathogens and their IgA/IgG antibody ratio were compared with DAS28-ESR score values in 54 patients with RRP (a) and 101 patients with non-RRP (b), using Spearman’s rank correlation coefficient analysis. NOTE: Pink: significant correlation at p<0.05, Blue: trending toward correlation at 0.05≤p<0.15, No color: no correlation.

Fig 6. Relationship of RF with antibody response functions and clinical maker levels in patients with RRP and non-RRP.

RF levels were compared with IgG and IgA index values, severity of arthritis, disease marker levels, serum cytokine levels, and hematological values in 54 patients with RRP and 101 patients with non-RRP, using Spearman’s rank correlation coefficient analysis. NOTE: Plot: Visual display for positive and negative “ρ” value of Spearmen correlation coefficient. Cells highlighted with yellow indicate significant correlation at p<0.05. Index 1: sum of anti-E. coli-LPS + anti-Pg-LPS, Index 2: sum of anti-E. coli-LPS + anti-PG-PS, Index 3: sum of anti-E. coli-LPS + anti-Pg-LPS + anti-PG-PS.

Fig 7. Relationship of ESR with antibody response functions and clinical marker levels in patients with RRP and non-RRP.

ESR values were compared with IgG and IgA index values, severity of arthritis, disease marker levels, serum cytokine levels, and hematological values in 54 patients with RRP and 101 patients with non-RRP, using Spearman’s rank correlation coefficient analysis. NOTE: Plot: Visual display for positive and negative “ρ” value of Spearmen correlation coefficient. Cells highlighted with yellow indicate significant correlation at p<0.05. Index 1: sum of anti-E. coli-LPS + anti-Pg-LPS, Index 2: sum of anti-E. coli-LPS + anti-PG-PS, Index 3: sum of anti-E. coli-LPS + anti-Pg-LPS + anti-PG-PS.

Fig 8. Relationship of CRP with antibody response functions and clinical marker levels in patients with RRP and non-RRP.

CRP levels were compared with IgG and IgA index values, severity of arthritis, disease marker levels, serum cytokine levels, and hematological analytical values in 54 patients with RRP and 101 patients with non-RRP, using Spearman’s rank correlation coefficient analysis. NOTE: Plot: Visual display for positive and negative “ρ” value of Spearmen correlation coefficient. Cells highlighted with yellow indicate significant correlation at p<0.05. Index 1: sum of anti-E. coli-LPS + anti-Pg-LPS, Index 2: sum of anti-E. coli-LPS + anti-PG-PS, Index 3: sum of anti-E. coli-LPS + anti-Pg-LPS + anti-PG-PS.

Fig 9. Relationship of anti-CCP antibody levels with antibody response functions and clinical marker levels in patients with RRP and non-RRP.

Anti-CPP antibody levels were compared with IgG and IgA index values, severity of arthritis, disease marker levels, serum cytokine levels, and hematological analytical values in 54 patients with RRP and 101 patients with non-RRP, using Spearman’s rank correlation coefficient analysis. NOTE: Plot: Visual display for positive and negative “ρ” value of Spearmen correlation coefficient. Cells highlighted with yellow indicate significant correlation at p<0.05. Index 1: sum of anti-E. coli-LPS + anti-Pg-LPS, Index 2: sum of anti-E. coli-LPS + anti-PG-PS, Index 3: sum of anti-E. coli-LPS + anti-Pg-LPS + anti-PG-PS.

Fig 10. Differences in the relationships between individual disease markers in the RRP and non-RRP groups.

Potential correlations between individual disease marker levels in 54 patients with RRP and 101 patients with non-RRP were confirmed by Spearman’s rank correlation coefficient analysis. NOTE: Pink: significant correlation at p<0.01, No color: no correlation.

Fig 11. Multiple bacterial pathogens are implicated in evoking serological disease markers and consequently aggravating disease activity in the RRP and non-RRP groups.

Low IgG antibody responses to E. coli-LPS, Pg-LPS and PG-PS are linked to RF, ESR and CRP levels in RRP. Among these putative pathogens, Pg-LPS and related pathogens derived from oral bacteria may play critical pathological roles in the RRP group. On the other hand, E. coli-LPS, PG-PS and other related pathogenic components, instead of Pg-LPS, may contribute to increasing RF, ESR and CRP levels in non-RRP. Anti-CCP antibody levels clearly correlate with RF in both RRP and non-RRP groups, but are not linked to severity of arthritis. NOTE: Arrows with solid line: Linkage, Arrows with dotted line: Possible linkage, Dashed line with X: No apparent linkage.