The role of the gut microbiome in osteoarthritis and joint pain

      Purpose: Osteoarthritis is a complex disease, caused by both genetic and environmental factors. One proposed environmental factor involved in OA risk and pathogenesis is the gastrointestinal microbiome. Obesity is known to affect the host gastrointestinal microbiome and vice versa. A proposed mechanism by which obesity could predispose to OA is through low grade systemic inflammation, mediated by bacterial endotoxins which migrate from the gut into the circulation. We aim to examine the gastrointestinal microbiome composition in relation to OA presence, severity and joint pain, using gastrointestinal microbiome data of 1,444 participants of the Rotterdam Study cohort.
      Methods: For 1,444 participants of the Rotterdam Study III (RS-III) population cohort we have determined the gut microbial composition using 16S ribosomal RNA-sequencing (V3–V4). Operational taxonomic units (OTUs) were generated and taxonomy was assigned using RDP classifier using SILVA (119) reference database. Radiographic OA phenotypes were determined by Kellgren-Lawrence scores (KL-score) for OA severity of the knee and hip and joint pain measures are based on the WOMAC score (Western Ontario and McMaster Universities Arthritis Index). Statistical analysis on inter-individual gastrointestinal microbial composition differences was performed by Bray-Curtis dissimilarly matrix and by Adonis permutation P-value calculation (VEGAN R package). Association analysis on the relative abundancy of individual species to each studies phenotype were assessed using a multivariate statistical linear regression analysis using the MaAsLin program. We adjusted for age, sex and technical covariates and BMI.
      Results: We found no significant association between the overall inter-individual variation of gastrointestinal microbiome composition and Hip OA (48 cases, 891 controls), Knee OA (124 cases, 826 controls) or with OA pain and severity (n = 1425). When we examine the relative abundance of single microbial taxonomies to OA, we find 35 nominal significant associations with Knee and Hip OA phenotypes of which seven remain significant after correction for multiple testing (FDR > 0.05). After additional correction for BMI, six of these remain microbiome-wide significant. We find two bacterial families within the order of Clostridiales to be significant associated with Hip OA WOMAC scores (Ruminococcaceae, P = 3*10-3 ) and Hip KL-scores (Uncertain Clostridiales family, Incertae Sedis XIV, P = 3.4810-3). For Knee WOMAC scores we find four significant associations (Table1.) on different taxonomic levels (Class – Order – Family – Genus) leading to the bacterial genus of Streptococcus (P = 1.6*10-4).
      Conclusions: We here present the first association results of the gut microbiome with osteoarthritic pain and severity in both the knee and hip joints. We identified highly significant associations between individual microbiome taxonomies and Knee and Hip OA WOMAC pain scores. These results seem independent of BMI, possibly suggesting a more obesity independent role for the microbiome in OA. We find a microbiome-wide association with Knee WOMAC and Streptococcus, group A. Streptococcus species are known to potentially cause rheumatic fever, an inflammatory joint disease affecting primarily the heart and joints. This may provide potential validation of the relationship between the gut microbiome, low grade inflammation in the joint, and OA.
      Table 1Knee WOMAC association results with individual bacterial Taxa.
      PhenotypeBacterial TaxonomyCoefficientP-valueFDR P-value
      Knee WOMACBacteria|Firmicutes|Bacilli0.0056.46E-075.82E-05
      Knee WOMACBacteria|Firmicutes|Bacilli|Lactobacillales0.0052.73E-073.44E-05
      Knee WOMACBacteria|Firmicutes|Bacilli|Lactobacillales|Streptococcaceae0.0046.73E-051.84E-03
      Knee WOMACBacteria|Firmicutes|Bacilli|Lactobacillales|Streptococcaceae|Streptococcus0.0041.60E-043.60E-03