Article Text

Download PDFPDF

Prevotella copri in individuals at risk for rheumatoid arthritis
  1. Deshire Alpizar-Rodriguez1,
  2. Till Robin Lesker2,
  3. Achim Gronow2,
  4. Benoît Gilbert1,
  5. Elena Raemy1,
  6. Celine Lamacchia1,
  7. Cem Gabay1,
  8. Axel Finckh1,
  9. Till Strowig2,3
  1. 1 Division of Rheumatology, Geneva University Hospital, Geneva, Switzerland
  2. 2 Department of Microbial Immune Regulation, Helmholtz Centre for Infection Research, Braunschweig, Germany
  3. 3 Hannover Medical School, Hannover, Germany
  1. Correspondence to Dr Axel Finckh, Division of Rheumatology, Geneva University Hospital, Geneva 1205, Switzerland; axel.finckh{at}; Dr Till Strowig, Department of Microbial Immune Regulation, Helmholtz Centre for Infection Rsearch, Braunschweig 38124, Germany; Till.Strowig{at}


Objectives Rheumatoid arthritis (RA) has been associated with a relative expansion of faecal Prevotellaceae. To determine the microbiome composition and prevalence of Prevotella spp. in a group of individuals at increased risk for RA, but prior to the development of the disease.

Methods In an ongoing cohort study of first-degree relatives (FDRs) of patients with RA, we identified ‘FDR controls’, asymptomatic and without autoantibodies, and individuals in pre-clinical RA stages, who had either developed anticitrullinated peptide antibodies or rheumatoid factor positivity and/or symptoms and signs associated with possible RA. Stool sampling and culture-independent microbiota analyses were performed followed by descriptive statistics and statistical analyses of community structures.

Results A total of 133 participants were included, of which 50 were categorised as ‘FDR controls’ and 83 in ‘pre-clinical RA stages’. The microbiota of individuals in ‘pre-clinical RA stages’ was significantly altered compared with FDR controls. We found a significant enrichment of the bacterial family Prevotellaceae, particularly Prevotella spp., in the ‘pre-clinical RA’ group (p=0.04).

Conclusions Prevotella spp. enrichment in individuals in pre-clinical stages of RA, before the onset of RA, suggests a role of intestinal dysbiosis in the development of RA.

  • microbiome
  • rheumatoid arthritis
  • epidemiology of ra
  • pre-clinical stages of ra

Statistics from

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.

Key messages

What is already known about this subject?

  • A high relative abundance of Prevotella copri has been identified in patients newly diagnosed with rheumatoid arthritis (RA), suggesting a role of gut microbiota dysbiosis in the aetiopathogenesis of the disease.

What does this study add?

  • This is the first study to describe a significantly altered microbiota, particularly a Prevotella spp. enrichment, already in individuals in pre-clinical stages of RA, compared with controls.

How might this impact on clinical practice or future developments?

  • Our results, together with previous studies in patients with early RA and recent mechanistic studies, support the mucosal origins hypothesis and the role of intestinal dysbiosis in the development of RA.

  • Intestinal dysbiosis could act as an early environmental modulator and may be a target of future preventive interventions in individuals at risk of RA, before the onset of the disease.


The aetiopathogenesis of rheumatoid arthritis (RA) is thought to result from a multistep process, where environmental factors induce a pathological activation of the immune system in susceptible individuals.1 Recent studies have suggested that the initial steps of the pathological autoimmune response originate in mucosal sites, rather than in the joints.2 Intestinal dysbiosis has been suggested to have a causal role in the pathogenesis of RA and has been shown to trigger arthritis development in genetically susceptible mice.3–6 Prevotella copri has been identified as highly enriched in the gut microbiota of patients newly diagnosed with RA and an increased immune response to this organism has been demonstrated in patients with RA suggesting a role of P. copri in the disease onset.7–9 Sequence homology between RA-specific autoantigens and epitopes from proteins of P. copri have been reported, supporting the molecular mimicry hypothesis, although exact mechanisms remain uncertain.8 Considering these observations, intestinal dysbiosis involving Prevotella spp. may be a risk factor for RA and a potential therapeutic target. However, to formally establish a causal role of intestinal dysbiosis in RA development, longitudinal studies prior to the onset of RA are required to demonstrate that the presence of Prevotella spp. precedes the development of RA. The aim of this study was thus to characterise the microbiota and determine the prevalence of Prevotella spp. in individuals during the pre-clinical phases of RA, before the development of clinically apparent RA.

Materials and methods

Study design and study population

First-degree relatives of patients with RA (RA-FDRs) have an increased risk of developing RA compared with the general population.10 11 The SCREEN-RA study is an ongoing cohort study of RA-FDRs, comprising subjects without a diagnosis of RA at enrolment, described in detail elsewhere (online supplementary text).12

We performed a nested case–control study within SCREEN-RA cohort to analyse the intestinal microbiota in individuals in pre-clinical phases of the disease. We identified participants in ‘pre-clinical RA’ stages based on the European League Against Rheumatism terminology for pre-clinical phases of RA.13 Operationally, we combined two pre-clinical RA stages for statistical power reasons: (1) ‘systemic autoimmunity associated with RA’ defined by anticitrullinated protein autoantibodies positivity and/or rheumatoid factor (RF) positivity,14 and/or (2) ‘symptoms and signs associated with possible RA’ as defined by the Connective Tissue Disease Screening Questionnaire with or without undifferentiated arthritis (see online supplementary text for details).15–17 We included a control group, namely ‘FDR controls’, namely RA-FDRs without any autoantibodies or symptoms associated with possible RA.

Participants were contacted by telephone to explain the objectives of the study and invited to provide stool samples for microbiome analysis. We included individuals with complete clinical information at the time of the stool sampling. We excluded participants who had undergone antibiotic therapy within the last 3 months, with a known history of inflammatory bowel disease and/or gastrointestinal tract surgery. The protocol was approved by the ethics committee and all participants signed an informed consent before providing a stool sample.

Sampling, DNA extraction and amplicon sequencing analysis to analyse the faecal microbiota

The DNA Genotek OMNIgene∙Gut Stool Microbiome Kit was used to collect, store and ship the samples.18 Stool samples processing and culture-independent analyses were performed. After DNA extraction, the variable region 4 (V4) region of the 16S rRNA gene was amplified using barcoded primers (F515/R806) and sequencing was performed on an Illumina MiSeq as previously described19 (details in the online supplementary text).

Statistical analysis

Controls and individuals in pre-clinical stages of RA were matched by sex, age and tobacco at the sampling stage. Based on our a priori hypothesis, the primary outcome of the study was the prevalence of bacteria from the family of Prevotellaceae, particularly Prevotella spp. Based on the mucosal origins hypothesis of RA,2 we postulated that the relative prevalence of Prevotellaceae in the stool of individuals in pre-clinical stages of RA would be increased compared with FDR controls. Statistical analyses of community structures were performed. We used linear discriminant analysis (LDA) effect size (LEfSe), an algorithm to compare the relative abundance of the different features between groups, as previously described.19 20 We performed subgroup analyses, dividing the group of ‘pre-clinical stages of RA’ into ‘systemic autoimmunity associated with RA’ and ‘symptoms and signs associated with possible RA’. We further explored the general characteristics association with Prevotellaceae abundance.


Study population

Among the 1067 RA-FDR participants in the SCREEN RA cohort, 183 (17%) were invited to provide stool samples, based on a priori inclusion criteria and the matching algorithm. A total of 133 RA-FDRs sent stool samples and could be analysed. General characteristics were balanced between the two groups (table 1).

Table 1

General characteristics at stool collection (133 participants)

Microbiota analysis

The comparison of microbial diversity in the faecal microbiota within individuals and between individuals, that is, alpha and beta diversity, respectively, of the FDR control and the pre-clinical RA groups did not reveal significant differences (see online supplementary figures S1–S3). We used the LEfSe method to analyse potentially more specific differences in microbiota composition between FDR controls and individuals in the ‘pre-clinical stages of RA’.20 Indeed, we found statistically significant differences in the relative abundances of bacterial taxonomic groups between the participants in pre-clinical stages of RA development and FDR controls (figure 1, LDA score >2, p<0.05). The family Prevotellaceae was the group of bacteria with the highest LDA score and was significantly enriched in individuals in ‘pre-clinical stages of RA’ (LEfSe p=0.040).

Figure 1

Linear discriminant analysis (LDA) effect size (LEfSe) evaluates the different relative abundance of bacteria. The faecal microbiota composition of a subset of participants of the SCREEN-RA cohort was compared using 16S rRNA gene sequencing. (A) Bacterial families identified using LEfSe (LDA >2, p<0.05). Red bars: bacterial taxa enriched in the preclinical RA group. Padjust: p values with Bonferroni adjustment. (B) Relative abundance (range 0 to 1) of the bacterial families Prevotellaceae (left panel) and Lactobacillaceae (right panel) in individual samples of the two groups. The thick horizontal dashed line in each graph shows median relative abundance and the solid line indicates mean relative abundance. FDR, first-degree relative; RA, rheumatoid arthritis.

In a subgroup analysis, the family Prevotellaceae was enriched particularly in participants with ‘systemic autoimmunity associated with RA’ compared with ‘FDR controls’ (online supplementary figure S4; LEFSe p=0.019), and no significant difference was found between individuals in the two groups of pre-clinical stages of RA (online supplementary figure S5), which allowed us to analyse them together.

We then specifically analysed the relative abundance of the family Prevotellaceae and associated taxa to evaluate whether all individuals of the pre-clinical RA phases display an enrichment of Prevotellaceae or whether an enrichment is observed only in some individuals (figure 2). This analysis confirmed that a larger proportion of individuals within the pre-clinical RA group compared with FDR controls (53% vs 30 %) had significant levels of Prevotellaceae (>1%), but Prevotellaceae are not present in all individuals. The general characteristics of individuals with high relative abundance (>1%) of Prevotellaceae were not different compared with individuals with no Prevotellaceae or lower relative abundance, but for a higher prevalence of RF positivity (online supplementary table S2). Furthermore, besides P. copri, other Prevotella spp. in other operational taxonomic units contribute to the Prevotellaceae enrichment in ‘pre-clinical RA’ (online supplementary figure S6).

Figure 2

Relative abundance of species belonging to the Prevotellaceae family in individual samples. The samples are ordered by decreasing cumulative relative abundance of operational taxonomic units (OTUs) assigned to the taxonomic level of Prevotella species. OTUs assigned only to the level of family or genus are not displayed. for each listed OTU, the closest related taxonomically described species is listed. ‘D’ indicates the sequence similarity between them. FDR, first-degree relative; RA, rheumatoid arthritis.


The present study focused on the prevalence of Prevotella spp. in the stool of individuals at risk for RA during pre-clinical phases of the disease. The microbiota of individuals in pre-clinical RA stages was significantly altered compared with FDR controls. In particular, the relative abundance of bacteria of the Prevotellaceae family and associated taxa were enriched among individuals in pre-clinical stages of RA and differed significantly from controls, in particular in individuals with ‘systemic autoimmunity associated with RA’, which is consistent with the mucosal origins hypothesis of RA development.2

A previous study analysed the microbiome of faecal samples of American patients with new-onset untreated RA and detected high abundance (>5%) of P. copri in 75% (33 of 44) compared with only 21.4% (6 of 28) of healthy individuals.9 This finding was not replicated in a study involving Chinese patients with RA.21 Cross-sectional studies in patients with RA do not allow making causal inferences, as this association could be due to differences in behaviours between patients and controls. Our study describes an increased relative abundance in Prevotella spp. in individuals in ‘pre-clinical RA stages’, using participants enrolled in a FDR-RA cohort. While this is still not a longitudinal study, the demonstration of a larger proportion of individuals in pre-clinical stages of RA with a significant abundance of Prevotellaceae strengthens the case for an involvement of Prevotella spp. in the RA aetiopathogenesis. However, longitudinal studies are needed to determine the specific role of intestinal dysbiosis and whether P. copri or other Prevotella spp. trigger systemic autoimmunity or drives the development of symptoms associated with RA.

Our study had limitations. The demonstration of a specific immune response against P. copri during pre-clinical stages would have strengthened our findings. In patients with RA, an increased humoral and Th1 cellular immune response against P. copri has been demonstrated.7 8 The microbiome study of the family members with RA and a replication of our results in a new-onset RA population would have further reinforced internal consistency. Our results, together with previous studies in patients with established RA and recent mechanistic studies, support the mucosal origins hypothesis and the role of Prevotella spp. dysbiosis in RA development.

In conclusion, we demonstrated that individuals at risk for RA with systemic autoimmunity and/or symptoms associated with RA have an enrichment of Prevotella spp. compared with FDR controls. Our findings support the mucosal origins hypothesis in the development of RA. Intestinal dysbiosis could act as an early environmental modulator and may be the target of future preventive interventions.


To participants of SCREEN-RA study. To Dr. Delphine S Courvoisier for providing methodological and statistical advice. DAR received an unrestricted research grant from AbbVie Switzerland for this project.



  • DA-R and TRL are joint first authors.

  • AF and TS contributed equally.

  • Handling editor Josef S Smolen

  • Correction notice This article has been corrected sinceit published Online First. The equal contributor statement has been added and the correspondence details updated.

  • Contributors DAR, TRL, AF and TS designed the study. DAR, ER, CL and AF were involved in patient recruitment, samples and data collecting. TRL, AG and TS were involved in samples processing and analysis. DAR, AF, TS and TRL were involved in statistical analyses and interpretation of data. All authors were involved in writing the manuscript and approved the final version. The first authors and corresponding authors had full access to all the data in the study and had final responsibility for the decision to submit for publication.

  • Funding DAR granted the Abbvie Grant 2016 and the Swiss Government Excellence Scholarship for PhD studies of The Federal Commission for Scholarships for Foreign Students. The work was supported by grants from the EU (StG337251) and Helmholtz Association (VH-NG-933) to TS. No other specific funding was received to carry out the work described in this article.

  • Competing interests None declared.

  • Patient consent for publication Obtained.

  • Ethics approval The protocol was approved by the ethics committee.

  • Provenance and peer review Not commissioned; externally peer reviewed.

  • Data sharing statement Raw data are available upon request from the corresponding author.