Skip to main content
NCBI home page
NIHPA Author Manuscripts logoLink to NIHPA Author Manuscripts
. Author manuscript; available in PMC: 2026 Oct 6.
Published in final edited form as: Ann Rheum Dis. 2025 Sep 29;85(2):254–264. doi: 10.1016/j.ard.2025.08.029

Short-Chain Fatty Acids and Their Gut Microbial Pathways Distinguish Rheumatoid Arthritis in Discordant Monozygotic Twins

Rebecca B Blank 1,*, Kevin Bu 2,*, Xinyuan Zhang 3,*, Weixi Chen 4, Ian Cunningham 1, Jeremy Sokolove 5, Lauren Lahey 5, Adriana Heguy 6, Rhina Medina 1, Carles Ubeda 7,8, Renuka R Nayak 9,10, Jiyuan Hu 4, Adam Cantor 2, Jakleen Lee 2, Frances MK Williams 3,*, Jose C Clemente 2,*, Jose U Scher 1,11,*
PMCID: PMC13636000  NIHMSID: NIHMS2199892  PMID: 41027803

Abstract

Background

Although genetic risk factors, such as HLA-DRB1 alleles, contribute to the pathogenesis of rheumatoid arthritis (RA), the concordance rate in monozygotic (MZ) twins is low, suggesting that other factors are involved in disease development. Further, the relative contribution of non-genetic elements in identical twins has not been characterized. Here, we aimed to characterize host and microbial biomarkers of RA by studying MZ twins discordant for disease using a multi-omics approach.

Methods

Eight pairs of MZ twins discordant for RA (n=16) were enrolled in the U.S. Gut microbiome was assessed using shotgun metagenomic sequencing. Autoantibodies, cytokines, and plasma proteins were measured in both plasma and feces. Levels of short-chain fatty acids (SCFAs) from serum and feces were quantified using gas chromatography mass spectrometry (GC-MS). Metagenomic data from theTwinsUK registry (N=14) was used to validate findings from the U.S. population.

Results

While microbiome diversity and composition did not differ between twins, we observed a significant decrease in the SCFA-producing bacteria Blautia faecis and significantly lower concentrations of fecal butyrate and propionate in affected RA twins in the U.S. TwinsUK showed similar reduction in the SCFA-producers Gemmiger formicilis, and Faecalicatena fissicatena and bacterial SCFA metabolism pathways.

Conclusion

Multi-omics biomarkers differentiate MZ twins discordant for RA. Fecal butyrate and propionate, and SCFA-producing bacteria were decreased in affected twins. We found a similar decrease in SCFA-producing taxa in affected twins in a geographically distinct cohort in the UK. Our results suggest that, if further validated in larger cohorts, multi-omics approaches may improve our understanding of RA pathogenesis and, potentially, contribute to more accurate diagnostics and co-adjuvant therapies.

Introduction

Genetic factors, including susceptibility alleles, have been identified in the pathogenesis of rheumatoid arthritis (RA)1,2, and can contribute to up to 60% of the risk of developing RA3. However, the concordance rate in monozygotic (MZ) twins is only 12 to 15%4,5, suggesting that other factors must contribute to disease onset and perpetuation. Among those, autoantibodies, the microbiome, and the metabolome have all been associated with RA.

The gut microbiome has been shown to be associated with inflammatory arthritis in various murine models6–8, and different studies have implicated microbial taxa in the pathogenesis of human RA, including periodontal microbes Porphyromonas gingivalis9,10 or Lactobacillus salivarius11 (present in both saliva and gut) and gut microbes (e.g., Prevotella copri12,6,13 and Subdoligranulum14), among others15. In fact, some of these bacteria have been implicated in the induction of RA-associated pathogenic autoantibodies which can be observed in both pre-RA10 and symptomatic RA patients9,11,13,14. These microbes can subsequently activate innate, B- and T-cell responses that produce pro-inflammatory cytokines such as IL-176,7,13, IL-1, IL-6, TNF-a and others15. Conversely, gut microbial fermentation byproducts, such as short chain fatty acids (SCFAs), have been shown to induce a tolerogenic immune response16, ameliorate inflammatory arthritis in murine models17,18 and have been found to be decreased in serum concentrations in pre-RA individuals19. While previous studies have investigated the contribution of each of these factors to RA in isolation, few have integrated different assays to characterize the molecular landscape of immune-mediated inflammatory disease20–22, and no study to date has used a multi-omics approach while controlling for host genetics. Here, we combined microbiome, autoantibodies, metabolome and circulating plasma proteins data in a cohort of MZ twins discordant for RA to identify disease signatures that discriminate between affected and unaffected twins. In an independent, geographically distinct cohort of monozygotic RA discordant twins, we validated the relevance of specific bacterial functions as potential biomarkers of non-progression in RA.

Methods

Participants

Monozygotic (MZ) twin pairs discordant for RA were recruited from the population-based Mid-Atlantic Twin Registry (MATR). Eight pairs of MZ twins were enrolled in the study. Diagnosis of RA was obtained through patient medical records. Upon enrollment, demographics, detailed medical history, including review of past and present medication were collected. In addition, MD-HAQ and swollen joint counts (SJC) and tender joint counts (TJC) were estimated. Stool, serum, and plasma samples were collected at time of enrollment. An additional validation cohort of MZ twin pairs discordant for RA were recruited from a population-based UK twin registry (TwinsUK). Available data included fecal shotgun metagenomic data from time of enrollment.

Patient and public Involvement

Patients and/or the public were not involved in the design of our research.

Metagenomics

Fecal samples were collected, stored on freezer packs and shipped to the Scher lab within 24 hours of collection and immediately stored at −80°C as previously described23. Bacterial DNA extraction and shotgun metagenomic sequencing to characterize the gut microbiome was performed as previously described23. Briefly, DNA was extracted from fecal samples using the MoBio Powersoil DNA extraction kit (QIAGEN, USA). Metagenomic sequencing was performed using the NextSeq500 platform (Illumina, USA) with 2 × 150bp reads according to the manufacturer’s instructions. Analysis of metagenomic sequencing data was performed using Metaphlan 424 to estimate bacterial composition and Humann 325 to determine bacterial functions.

Fecal sample collection, DNA extraction, library preparation in TwinsUK cohort were performed as previously described26. Metagenomics sequencing was performed using an Illumina HiSeq 2500 with 2 × 125 bp reads by GenomeScan (https://genomescan.nl/). For TwinsUK, paired-end reads quality control was performed using the YAMP pipeline (v.0.9.5.0, mode “qc”)26 with “genomescan” parameter. Taxonomic profiling was performed with MetaPhlAn 4 and functional profiling was performed with HUMAnN 3.

Diversity metrics and visualizations were estimated as implemented in QIIME2 V2020.8.027. Shannon Diversity was used to measure alpha diversity, while Bray-Curtis was used for beta diversity. The Bray-Curtis distance matrix was used as input for principal coordinate analysis (PCoA). Comparison of microbiome diversity distances between three groups (intra-twin, inter-RA and inter-unaffected) was performed as follows. For alpha diversity, the intra-twin distances were obtained by computing the absolute value of the differences in Shannon Entropy for each of the twin pairs (|αUnaffected - αRA|). For the inter-RA distances, the absolute value of the difference in Shannon Entropy was computed for each pair of RA-affected twins, i.e. for all (8 choose 2) = 28 pairs. Analogous distances were computed for the inter-unaffected group.

Beta diversity distances were computed in a similar fashion. The intra-twin distances consist of the 8 Bray-Curtis distances between unaffected and affected twins, i.e. one per twin pair. The inter-RA differences consist of aggregating all (8 choose 2) = 28 distances between each pair of individuals with RA, and an analogous calculation was performed for the inter-unaffected differences. The distances of these three groups (intra-twin, inter-RA, and inter-unaffected) were then compared using the Kruskal-Wallis test. Correlations of bacterial diversity (Shannon entropy for alpha diversity or Bray-Curtis intra-twin for beta diversity) with swollen joint count (SJC), tender joint count (TJC), C-reactive protein (CRP), and Disease Activity Score-CRP (DAS28-CRP) were performed using Spearman correlation. Cohen’s D effect size was computed using the method previously described28.

Metabolomics

Fecal short- and medium-chain fatty acids (SCFA, MCFA) measurements were performed at the University of Michigan Metabolomics Core (Michigan Medicine Biomedical Research Facilities, Ann Arbor, MI, USA). Briefly, SCFA analysis was performed by GC-MS on an Agilent 69890N GC-5973 MS detector as previously described29. MCFA analysis was performed by RPLC-MS on an Agilent 1290 Infinity II / 6545 qTOF MS system with the JetStream Ionization (ESI) source (Agilent Technologies, Inc., Santa Clara, CA USA) using the Waters Acquity HSS T3 1.8 μm 50 mM column (Waters Corporation, Milford, MA). Each sample was analyzed twice, once in positive and once in negative ion mode. Mobile phase A was 100% water with 0.1% formic acid and mobile phase B was 100% methanol with 0.1% formic acid. The gradient for both positive and negative ion modes was as follows: 2% B (0 min), 75% B (20 min), 98%B (22 min), 98%B (30 min), 2% B (30.1 min) was used. The column was then reconditioned for 7 min with 2%B before moving to the next injection The flow rate was 0.46 mL/min and the column temperature was 40°C. The injection volume for positive and negative mode was 5 μL and 8 μL, respectively. Source parameters were: drying gas temperature 350°C, drying gas flow rate 10 L/min, nebulizer pressure 30 psig, sheath gas temp 350°C and flow 11 l/min, and capillary voltage 3500V, with internal reference mass correction. Data analysis of SCFA and MCFA were processed using MassHunter Quantitative analysis version B.07.00. SCFAs and MCFAs were normalized to the nearest isotope labeled internal standard and quantitated using 2 replicated injections of 6 standards to create a linear calibration curve with accuracy better than 80% for each standard. Other compounds in the analysis were normalized to the nearest internal standard, and the peak areas were used for differential analysis between groups. Samples were normalized to wet sample weight after quantification. Serum samples were collected, centrifuged at 1700g for 10 minutes and immediately stored at −80° C. Frozen serum sample aliquots were shipped on dry ice for SCFA metabolomics profiling by Metabolon Inc. (Morrisville, NC), as previously described30, using the quantitative SCFA panel. Quantitative analysis was performed by LC-MS/MS on an Agilent 1290 UHPLC/Sciex QTrap 5500 instrument.

Multiplex autoantibody assays

Autoantibody concentration was determined in stool and plasma using multiplex flow cytometry assay as previously described31,32. Briefly, a panel of 103 citrullinated and non-citrullinated self-protein antigens were conjugated to fluorescent beads using the Bio-Plex multiplex assay platform (BIO-Rad Laboratories, Hercules, CA, USA) and analyzed using the Luminex 200 (Luminex, Austin, TX, USA). Pooled beads were mixed with either plasma or stool samples and incubated at room temperature. After washing, the sample-dyed bead mixture was incubated with anti-human IgG antibody conjugated to phycoerythrin (PE) at room temperature. After second wash, the bead mixture was passed through the laser detector, Luminex 200, that identified the beads based on fluorescence of the dyes. Quantity of autoantibody bound to each bead was determined by PE fluorescence and expressed as mean fluorescence intensity (MFI). Serum anti-CCP antibody was measured by QuantaLite CCP 3.1 IgG/IgA ELISA (INOVA Diagnostics, Fairfax, VA, USA).

Analysis of plasma cytokines and proteins

Plasma analytes were measured by proximity-extension assay (PEA) (OLINK technology, Bioxpedia Aarhus, Denmark) using a 92 inflammation-associated plasma protein biomarker panel. PEA was performed at the Human Immune Monitoring Center, Icahn School of Medicine, NY, NY, USA. PEA allows for detection of very low concentrations of plasma analytes and the final concentration is given in Normalized Protein eXpression (NPX)33 units, which is an arbitrary unit on a log2 scale. Of the 92 analytes in the panel, six analytes were removed from analysis because at least 33% of the sample measurements were below the limit of detection.

Statistical Analyses

Association of clinical metrics (e.g. age, sex, ethnicity, swollen and tender joint counts) with disease status was assessed using Fisher’s Exact Test for dichotomous variables and paired t-tests for continuous variables. Analysis of inflammatory cytokines, fatty acids, and citrullinated peptides was performed using paired t-tests. Comparison of alpha diversities, taxonomic abundances, and pathway abundances in unaffected versus affected twins was done using paired Wilcoxon-Signed rank tests. Permutational Multivariate Analysis of Variance (PERMANOVA) using 999 permutations was used to assess differences in composition between affected and unaffected twins. Comparison of alpha diversity differences or beta diversity distances between subgroups (e.g. comparing differences between individuals of the same diagnosis versus differences between individuals with the same genetics) was done using Kruskal-Wallis tests. Differential taxa enriched or depleted in affected twins was done using Wilcoxon-Signed Rank tests. Correlations were computed using Spearman’s Correlation coefficient and were further inspected for false positives using CUTIE34.

Statistical analyses (paired and unpaired t-tests, Wilcoxon, Mann-Whitney U, Cohen’s D35), and plotting were performed using Python 3.7 and libraries (pandas36, seaborn37), and R 4.2.1 and libraries (stats, ggplot2)38,39. Unless otherwise specified, functions, plotting, and testing were performed with default parameters.

Results

Eight monozygotic twin pairs discordant for RA were enrolled in the United States (US), with a mean age of 56.3 years and mean disease duration of 7.0 years (Table 1). The majority were female (87.5%), and all were non-Hispanic white. The RA-affected twins had moderate disease activity with a mean disease activity score-C-reactive protein (DAS 28-CRP) of 4.02, their mean tender joint count (TJC) was 10.4 and mean swollen joint count (SJC) was 6.25. Both affected and unaffected twins had a wide range of serum anti-cyclic citrullinated peptide (CCP) antibody levels as quantified by CCP 3.1 IgG/IgA ELISA, with no significant difference between affected and unaffected twins. Most affected twins (75%) were currently or had previously taken methotrexate (MTX) and 62.5% were currently taking MTX at time of the study. 62.5% of affected twins were being treated with or had previous exposure to a biologic agent (Table 1); four affected twins had exposure to TNF inhibitors and one had had exposure to multiple TNF inhibitors and abatacept and was on rituximab at the time of enrollment. None had taken or were currently taking a JAK inhibitor. Potential confounders (smoking, antibiotics, NSAIDs, steroid usage) were not significantly associated with any of the categories tested (Table 1 and Supplementary Table 1).

Table 1:

Cohort demographics and clinical information.

Characteristic Unaffected (n = 8) RA (n = 8) p-value
Age- years, mean (median) 56.3 (58.0) 56.3 (58.0) 1
Female- n (%) 7 (87.5) 7 (87.5) 1
White/Caucasian- n (%) 8 (100) 8 (100) 1
Disease duration, years, mean (median) -- 7.0 (4.0) --
Disease activity parameters
 TJC-28, mean (median) -- 10.4 (9.5) --
 SJC-28, mean (median) -- 6.25 (4.0) --
 DAS 28-CRP, mean (median) -- 4.02 (4.0) --
Serum CCP antibody, units, mean (SD) 13314 (36841) 27203 (33538) 0.117
Ever smoked, n (%) 4 (50%) 3 (37.5%) >0.999
Current NSAID use, n (%) 4 (50%) 5 (62.5%) >0.999
Medication use ever
 Prednisone, n (%) -- 4 (50.0) --
 Methotrexate, n (%) -- 6 (75.0) --
 Biologic agent, n (%) -- 5 (62.5) --
Open in a new tab

TJC = tender joint count; SJC = swollen joint count; DAS28-CRP = Disease activity score-C reactive protein; CCP = cyclic citrullinated peptide

Gut microbiome of affected and unaffected twins

Alpha and beta diversity were not significantly different between affected and unaffected twins (Mann-Whitney U p=0.95 and PERMANOVA p=0.80, respectively; Figures 1A–B). We next compared the gut microbiome of individuals with a shared diagnosis versus individuals with a shared genetic background. There were no significant differences between intra-twin and intra-disease differences in overall species diversity using Shannon Entropy (Kruskal-Wallis p=0.67, Supplementary Figure 1A). Namely, the distance within twin pairs is significantly lower than the distance within unrelated RA subjects or the distance within unrelated healthy subjects. These trends were recapitulated to a lesser extent in analyses of beta diversities, where twin pairs tended to have a gut microbiome more similar to each other than unrelated individuals with a shared diagnosis, although not significantly (Kruskal-Wallis p=0.15, Supplementary Figure 1B). Regarding specific features, we observed a depletion of Blautia faecis in affected twins compared to unaffected twins (Wilcoxon p=0.04) (Figures 1C–D). Analysis of microbial functions also identified 7 pathways that were differentially abundant between affected and unaffected twins, including lysine metabolism, cyanoamino acid metabolism, and monobactam biosynthesis, which were more abundant in RA (Figure 1E, Supplementary Figure 2).

Figure 1.

Figure 1.

Open in a new tab

Gut bacterial composition and predicted functional pathways in affected and unaffected twins in US cohort. (A) Shannon alpha diversity in participants, with box plots representing median and interquartile range. Lines are drawn between MZ pairs. (B) Principal coordinate analysis (PCoA) of Bray-Curtis beta diversity between groups. (C) Linear discriminant analysis Effect Size (LEfSe) showing differential species in affected twins compared to unaffected twins. (D) Boxplot showing distribution of B. faecis abundance in unaffected and affected twins. (E) Barplot of Wilcoxon signed p-values for differential pathways. *p<0.05, **p<0.01.

Current treatment with MTX was not associated with significant differences in alpha (Mann-Whitney U p=0.39, Supplementary Figure 3A) or beta diversity (Mann-Whitney U p=0.57, Supplementary Figure 3B). Alpha diversity was also not significantly associated with disease activity, as measured by swollen joint count (SJC) (Spearman rho=0.11, p=0.8), TJC (Spearman rho=0.36, p=0.39), C-reactive protein (CRP) (Spearman rho=−0.10, p=0.82) or DAS28-CRP (Spearman rho=0.12, p=0.78) (Supplementary Figure 4A–D). Finally, we found no significant correlation of these four clinical measures with the intra-twin Bray-Curtis distances for the affected twins (p=n.s.), suggesting that disease severity is not associated with larger difference in microbial composition between affected and unaffected twins.

Affected twins have lower concentration of short-chain fatty acids

We next tested whether there were differences in fatty acid concentration between affected and unaffected twins. Compared to their unaffected siblings, affected twins had significantly lower concentrations of fecal butyrate (p=0.03), propionate (p=0.03), and acetate (p=0.05) and a significantly higher concentration of fecal valerate (p=0.04) (paired t-test, Figure 2A–D). There was no significant difference in serum SCFAs between the two groups, although unaffected twins had higher concentrations of SCFA (paired t-test; p=0.41 for 2-methylbutyric acid, p=0.31 for isobutyric acid, p=0.67 for acetic acid, p=0.15 for hexanoic acid in Supplementary Figure 5).

Figure 2.

Figure 2.

Open in a new tab

Comparison of fecal short chain fatty acid (SCFA) concentrations between unaffected and affected twins. (A) Affected twins have lower fecal concentration of acetate. (B) Butyrate fecal levels are depleted in affected twins. (C) Propionate fecal concentration is lower in affected twins. (D) Valerate fecal levels are higher in affected twins. Lines are drawn between MZ pairs. P-values calculated with paired t test; *p<0.05.

Integrative analysis of fecal autoantibodies, SCFA, and inflammatory phenotypes

We further explored whether autoantibodies considered to be pathogenic in RA were differentially expressed in twin pairs. Using multiplex bead assay, we measured 103 fecal autoantibodies (Supplementary Table 2) against either citrullinated (ACPA) or non-citrullinated native proteins. Only one fecal autoantibody was significantly elevated in U.S. affected twins compared to their unaffected siblings (Supplementary Figure 6 and Supplementary Table 3). We then conducted Spearman correlation analysis between taxa and functional pathways against cytokines, fatty acids, and autoantibody data in the U.S. affected and unaffected twins separately, focusing on the differential taxa (B. faecis) and 7 differential pathways. We found no significant correlations involving B. faecis and 34 involving differential pathways in the affected group (two of which also involved differential inflammatory markers), and 30 significant correlations involving B. faecis and 48 involving differential pathways in the unaffected twins, after filtering out false positive results. Out of the 30 significant correlations involving B. faecis in the unaffected twins, two were with cytokines and 28 were with fecal ACPAs. B. faecis was positively correlated with FMS-like tyrosine kinase 3 ligand (Flt3L) in unaffected twins (rho=0.90, p=0.002) and negatively correlated with stem cell factor (SCF) (rho=−0.98, p=0.00003; Supplementary Figure 7A–B). Additionally, all but one of the remaining 28 significant associations between B. faecis and fecal autoantibodies were negative correlations.. B. faecis was not correlated with fecal butyrate, propionate, valerate and acetate levels, although the correlation approached statistical significance for acetate (Spearman’s rho=0.45, p=0.09). We next performed correlation analyses in unaffected and affected twins separately, to assess the association of fecal SCFAs with fecal inflammatory cytokines. We found 70 correlations in the affected twins and 43 in the unaffected twins that were significant at p<0.05, though none passed FDR correction. Notably, all of the correlations were negative in the affected twins, including correlations between fecal butyrate and IL-8 (rho=−0.86, p=0.007), fecal butyrate and EN-RAGE (rho=−0.76, p=0.028), fecal propionate and IL-8 (rho=−0.93, p=0.001) or fecal propionate and EN-RAGE (rho=−0.83, p=0.01) (Supplementary Figure 8). Lastly, we compared the effect sizes of the differential features of each data type using the paired Cohen’s D statistic (Figure 3 and Supplementary Table 5). Among the metagenomic pathways, lysine biosynthesis and atrazine degradation exhibited the largest effect sizes. The SCFAs butyrate, propionate, and valerate also had large effect sizes (D>0.8), as did the fecal autoantibody Clusterin 231–250 sm-1 cyclic IgG and Blautia faecis.

Figure 3.

Figure 3.

Open in a new tab

Cohen’s D (paired) effect sizes for top differential features of each datatype along with directionality of enrichment.

Metagenomic analyses in a geographically distinct cohort of monozygotic twins discordant for RA

To validate findings observed in the U.S. cohort, we obtained shotgun metagenomic data from a geographically independent cohort of seven pairs of monozygotic twins discordant for RA from the TwinsUK registry in the United Kingdom. The cohort consisted of white females with a mean age of 63.4 years and mean disease duration of 19.7 years (Supplementary Table 4). Five affected twins had previously or were currently taking MTX (71.4%), and three were being treated with or had previous exposure to a biologic agent (42.9%) (Supplementary Table 4). We found no significant difference in either alpha diversity (Wilcoxon test, p = 0.71) or beta diversity (PERMANOVA, p=0.51) between affected and unaffected twins (Figure 4A–B). We identified different SCFA-producing bacterial taxa with decreased abundance in the affected twins, including Gemmiger formicilis and Faecalicatena fissicatena (Figure 4C). Clostridia were also significantly abundant in the affected twins, while E. coli was enriched in the unaffected twins (Figure 4C). In the U.S. cohort, although not significantly, Clostridia was also more abundant in affected twins (affected twins mean abundance and standard deviation: 1.5e-4 ± 3.0e-4; unaffected: 0 ± 0; p=0.18) and E. coli in the unaffected twins (affected: 1.2e-5 ± 2.5e-5, unaffected twins: 5.2e-3 ± 1.4e-2; p=0.14).

Figure 4.

Figure 4.

Open in a new tab

Gut bacterial composition and predicted functional pathways in affected and unaffected twins in UK cohort. (A) Shannon alpha diversity in participants, with box plots representing median and interquartile range. Lines are drawn between MZ pairs. (B) Principal coordinate analysis (PCoA) of Bray-Curtis beta diversity between groups. (C) Linear discriminant analysis Effect Size (LEfSe) showing differential species in affected twins compared to unaffected twins. (D) Barplot of Wilcoxon signed p-values for differential pathways. *p<0.05, **p<0.01.

Analysis of microbial functions identified six pathways that were differentially abundant between affected and unaffected twins in the UK cohort, including propionate metabolism, ascorbate and aldarate metabolism, and biosynthesis of ansamycins, which were more abundant in unaffected twins (Figure 4D.) Propionate metabolism and biosynthesis of ansamycins, which were significantly more abundant in unaffected twins in the UK cohort, exhibited the same trend in unaffected twins in the US cohort although the difference was not significant (Supplementary Table 6). We further investigated whether other SCFA metabolic pathways were differentially abundant in unaffected vs. affected twins. While not statistically significant in either of the US or UK cohorts, there was also higher abundance of butyric acid metabolism and pyruvate metabolism (an intermediate in the acetic acid pathway) in unaffected twins (Supplementary Table 6).

Discussion

While studies on twin pairs have been previously conducted to determine concordance rates in RA, ours is the first to perform an integrated multi-omics investigation of microbial, immune, and metabolic features that differentiate monozygotic twins discordant for RA. Two prior studies have used a multi-omics approach to study RA, although limited to studying gut microbiome and metabolomic differences between RA patients and unrelated (non-twins) healthy controls21,22. Another recent study assessed the differences in gut microbiome, colonic transcriptome and peripheral blood immune profiles between patients with psoriasis and healthy controls20. While we found no overall differences in microbial alpha or beta diversity between affected and their unaffected twins, we identified one species, Blautia faecis, that was differentially abundant in the US cohort. Blautia spp. are butyrate and propionate producers40, which may explain the increased levels of fecal SCFA observed in the unaffected twins. Additionally, Blautia faecis, which produces lactate and acetic acid and reduces expression of inflammatory cytokines such as TNFα and IL-840,41, was enriched in unaffected twins. Flt3L, which has been implicated in both inflammatory42 and regulatory43,44 processes in RA, was positively correlated with B. faecis abundance in unaffected twins. We hypothesize that B. faecis may modulate a regulatory process via its association with Flt3L that could contribute to protection against RA. This hypothesis is supported by the negative correlation between fecal autoantibodies and B. faecis in the unaffected twins. Furthermore, and unlike serum autoantibodies, which are broadly elevated in affected RA twins compared to their unaffected twins45, out of the 103 fecal autoantibodies tested, only one was elevated in affected twins. This suggests that fecal autoantibodies might not be a differentiating feature of monozygotic discordant twins per se, but might provide insights when paired with microbiome analyses.

The lack of overall differences in microbiome diversity between affected and unaffected twins was expected, given that affected twins had been treated or were undergoing treatment at time of sample collection. In fact, the gut microbiome of untreated RA and healthy controls are significantly different12 but those differences are partially reduced after treatment8,11,46,22. Effective RA treatment may also contribute to our observation that microbial composition was more similar between intra-pair MZ twins than between unrelated individuals of the same disease state, either unaffected or RA. Although P. copri has been previously reported to be enriched in early RA6,12,13, we found that its abundance was not significantly different between unaffected and affected twins in either the U.S. or UK cohorts. A recent study by Andreasson et al. showed that P. copri abundance decreased after three months of treatment with either MTX or TNF inhibition, and P. copri was negatively correlated with improved disease activity47. Similarly, treatment with MTX has been previously reported to be associated with a reduction in B. fragilis48. Nayak and colleagues have also found that P. copri decreased after methotrexate treatment in a gnotobiotic mouse model and in ex vivo cultures of gut microbial communities from RA patients8. We thus hypothesize that the low P. copri abundance in our twin cohorts might simply reflect the pathophysiology of established disease and the effect of treatment with MTX and other drugs.

Our study incorporates a validation cohort of a geographically distinct set of monozygotic twins from the UK, the TwinsUK registry, that were also discordant for RA. There were several commonalities between the cohorts, including comparable age and medication use, and a lack of significant differences in microbial alpha or beta diversity. Although the specific taxa found in the U.S. and UK cohorts were distinct, we observed that those depleted in the affected twins, Blautia spp. in the US, and Gemmiger and Faecalicatena in the UK, are all SCFA producers40,49,50. We also found that the enrichment of Clostridia and depletion of E. coli in affected twins was consistent across both cohorts, and in agreement with previous literature51–53. Importantly, we found that fecal butyrate and other SCFAs were elevated in unaffected U.S. twins, which was paralleled by an enrichment in SCFA-producing bacteria in the unaffected UK twins. Functional pathway analysis of the UK cohort found that both propionate metabolism and biosynthesis of ansamycins were more abundant in unaffected twins, a finding that was also observed in the unaffected U.S. twins, although differences were not statistically significant. Increased propionate metabolism in the unaffected UK twins suggests that SCFAs, such as propionate, are an important discriminator between unaffected and affected twins, and are in line with the observation that unaffected U.S. twins have increased fecal SCFAs levels and that SCFAs are negatively associated with RA, as discussed above and elsewhere17,18,22,54.

The anti-inflammatory properties and their protective role against disease progression of some gut microbial fermentation products, such as SCFAs, have been well-studied in both murine models of RA and humans17,18,54. We found negative correlations between fecal butyrate and propionate and inflammatory cytokines in the affected U.S. twins, suggesting that fecal SCFA concentration may regulate pro-inflammatory responses in affected twins. SCFAs such as propionate remain low in early RA patients, both before and after MTX treatment22. Although patients in our cohort have long-standing disease, we hypothesize that reduced SCFA in affected twins is more likely due to the disease state and not an effect of treatment regimen. Our finding that SCFAs are elevated in the feces of unaffected twins suggests that they may play a protective role in individuals that are genetically predisposed to developing RA. Further work is required to characterize the long-term dynamics of intestinal SCFA in response to various therapeutics.

Although our study is highly innovative in the multi-omic interrogation of monozygotic twins discordant for RA from two geographically distinct cohorts, we acknowledge that our work has various limitations.

Our sample size is limited due to the relatively low prevalence of RA in the general population and, particularly, of twins discordant for RA. One unaffected twin was eventually diagnosed with RA more than a year after conclusion of the study; therefore, it is unlikely to have confounded our results. Most affected twins in our study were being treated with DMARDs at time of enrollment, and the mean duration of their disease was 7 years in the US cohort and 19.7 years in the UK cohort. Further, our design was cross-sectional in nature, which prevented us from formulating interpretations on the causality of disease pathogenesis and the effect of treatment. Other potential limitations include the use of specific DMARDs in each of the cohorts and lack of detailed dietary information. Other medical and lifestyle factors that can influence the microbiome55–57 could also be seen as limitations, although we did not find significant differences between cohorts, limiting their potential impact. Further work on these and similar cohorts is ongoing to expand our multi-omic analyses, such as expanding our methodologic capabilities with single cell technologies.

Despite these limitations, our work identified several differentiating features between RA and their unaffected identical twins in two geographically distinct cohorts. Unaffected twins in both cohorts harbored an abundance of SCFAs producers such as Blautia, Gemmiger and Faecalicatena. U.S. unaffected twins had higher concentrations of fecal SCFA, propionate and butyrate, and UK unaffected twins had increased abundance of propionate metabolism pathways. These findings were significant even when considering potential confounders. Taken together, our results suggest that SCFAs may be a biomarker of protection against development of RA and highlight its potential as a therapeutic or preventive treatment approach. Our study is the first to use a multi-omics approach to investigate microbial, metabolomic and molecular features in twins discordant for rheumatoid arthritis. The unique setting of our cohorts thus provides a novel avenue to investigate RA pathogenesis while controlling for genetic predisposition and highlights the need to integrate metabolomics and host molecular features to microbiome studies.

Supplementary Material

1

Supplementary Analyses

A github containing additional results not highlighted in this manuscript and its supplementary figures was created at https://github.com/clemente-lab/twinsra_analyses.

Key messages.

What is already known on this topic

  • Genetic and environmental factors contribute to rheumatoid arthritis (RA) development and perpetuation.

  • Genetics plays only a limited role in RA disease development, as suggested by the low disease concordance rate in monozygotic twins.

  • Gut microbial taxa and their metabolites (e.g., short-chain fatty acids) are implicated in RA pathogenesis.

  • Few RA studies have utilized a multi-omics approach to understand disease pathogenesis, and none have controlled for the confounding effect of genetics.

What this study adds

  • This study characterizes, for the first time, the molecular landscape of RA using a multi-omics approach while controlling for host genetics.

  • Our results identify SCFA-producing gut microbes and intestinal SCFAs as potential biomarkers for non-progression in unaffected monozygotic discordant RA twins.

  • Analysis of a geographically distinct cohort of monozygotic twins finds similarities in microbial taxa and pathways of SCFA producers in unaffected twins.

How this study might affect research, practice or policy

  • The study of genetic and environmental factors in isolation limits our understanding of RA pathogenesis. A comprehensive multi-omics characterization of monozygotic twins discordant for RA, including microbiome, metabolome, and proteome, provides an integrated molecular landscape of disease pathogenesis and may help with biomarker discovery for earlier diagnosis and potential SCFA-based therapeutics.

Acknowledgements

The project was facilitated by the Mid-Atlantic Twin Registry (MATR) administered by the Cohort and Registry Administration (CARA) Core, a Virginia Commonwealth University shared resource. We would like to extend our gratitude to all of the MATR twins who took part in this study.

Funding

This work was supported by NIH/NIAMS UC2AR081034 (J.U.S., PI; J.C.C., MPI; R.B.B.) NIH/NIAMS R01AR074500 (J.U.S, PI; J.C.C; R.N.), R03 AR072182 (J.U.S., PI), National Center for Advancing Translational Sciences (NCATS), National Institutes of Health, through Grant Award Number UL1TR001445 (R.B. trainee); NIH/NIAMS T32AR069515 (J.U.S., MPI; R.B. trainee); NYU Judith and Stewart Colton Center for Autoimmunity (J.U.S.); R03DK138490 (J.H., PI); PID2023-150086OB-I00 from MICINN (Spain) (C.U., PI); NIH K08AR073930 (R.N.); VA I01CX002557 (R.N., PI).

TwinsUK is funded by the Wellcome Trust, Medical Research Council, Versus Arthritis (grant 21227), European Union Horizon 2020, Chronic Disease Research Foundation (CDRF), Zoe Ltd and the National Institute for Health Research (NIHR) Clinical Research Network (CRN) and Biomedical Research Centre based at Guy’s and St Thomas’ NHS Foundation Trust in partnership with King’s College London.

Footnotes

Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

Competing interests:

RBB, KB, WC, IC, AH, RH, RRN, JH, AC, JL, JS, LL, CU, JCC, XZ, FW have no conflicts to declare. JUS has served as a consultant for Janssen, Novartis, Pfizer, Sanofi, Amgen, UCB, BMS, and AbbVie; and has received funding for investigator-initiated studies from Janssen and Pfizer.

Declaration of interests

The authors declare the following financial interests/personal relationships which may be considered as potential competing interests:

Jose U. Scher reports a relationship with Janssen Pharmaceuticals Inc that includes: consulting or advisory and funding grants. Jose U. Scher reports a relationship with Novartis that includes: consulting or advisory. Jose U. Scher reports a relationship with Pfizer that includes: consulting or advisory and funding grants. Jose U. Scher reports a relationship with Sanofi that includes: consulting or advisory. Jose U. Scher reports a relationship with Amgen Inc that includes: consulting or advisory. Jose U. Scher reports a relationship with UCB that includes: consulting or advisory. Jose U. Scher reports a relationship with Bristol Myers Squibb Co that includes: consulting or advisory. Jose U. Scher reports a relationship with AbbVie Inc that includes: consulting or advisory. If there are other authors, they declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

References

  • 1.Okada Y, Wu D, Trynka G, et al. Genetics of rheumatoid arthritis contributes to biology and drug discovery. Nature 2014;506(7488):376–381. DOI: 10.1038/nature12873. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 2.Smolen JS, Aletaha D, Barton A, et al. Rheumatoid arthritis. Nature Reviews Disease Primers 2018;4(1):18001. DOI: 10.1038/nrdp.2018.1. [DOI] [PubMed] [Google Scholar]
  • 3.MacGregor AJ, Snieder H, Rigby AS, et al. Characterizing the quantitative genetic contribution to rheumatoid arthritis using data from twins. Arthritis Rheum 2000;43(1):30–7. DOI: 10.1002/1529-0131(200001)43:1<30::AID-ANR5>3.0.CO;2-B. [DOI] [PubMed] [Google Scholar]
  • 4.Silman AJ, MacGregor Aj Fau - Thomson W, Thomson W Fau - Holligan S, et al. Twin concordance rates for rheumatoid arthritis: results from a nationwide study. Br J Rheumatol 1993;32(10):903–7. (In eng). DOI: 10.1093/rheumatology/32.10.903. [DOI] [PubMed] [Google Scholar]
  • 5.Aho K, Koskenvuo M, Tuominen J, Kaprio J. Occurrence of rheumatoid arthritis in a nationwide series of twins. J Rheumatol 1986;13(5):899–902. (https://www.ncbi.nlm.nih.gov/pubmed/3820198). [PubMed] [Google Scholar]
  • 6.Maeda Y, Kurakawa T, Umemoto E, et al. Dysbiosis Contributes to Arthritis Development via Activation of Autoreactive T Cells in the Intestine. Arthritis Rheumatol 2016;68(11):2646–2661. DOI: 10.1002/art.39783. [DOI] [PubMed] [Google Scholar]
  • 7.Wu HJ, Ivanov II, Darce J, et al. Gut-residing segmented filamentous bacteria drive autoimmune arthritis via T helper 17 cells. Immunity 2010;32(6):815–27. DOI: 10.1016/j.immuni.2010.06.001. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 8.Nayak RR, Alexander M, Deshpande I, et al. Methotrexate impacts conserved pathways in diverse human gut bacteria leading to decreased host immune activation. Cell Host & Microbe 2021;29(3):362–377.e11. DOI: 10.1016/j.chom.2020.12.008. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 9.Hitchon CA, Chandad F, Ferucci ED, et al. Antibodies to Porphyromonas gingivalis Are Associated with Anticitrullinated Protein Antibodies in Patients with Rheumatoid Arthritis and Their Relatives. The Journal of Rheumatology 2010;37(6):1105–1112. DOI: 10.3899/jrheum.091323. [DOI] [PubMed] [Google Scholar]
  • 10.Mikuls TR, Thiele GM, Deane KD, et al. Porphyromonas gingivalis and disease-related autoantibodies in individuals at increased risk of rheumatoid arthritis. Arthritis Rheum 2012;64(11):3522–3530. DOI: 10.1002/art.34595. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 11.Zhang X, Zhang D, Jia H, et al. The oral and gut microbiomes are perturbed in rheumatoid arthritis and partly normalized after treatment. Nature Medicine 2015;21(8):895–905. DOI: 10.1038/nm.3914. [DOI] [PubMed] [Google Scholar]
  • 12.Scher JU, Sczesnak A, Longman RS, et al. Expansion of intestinal Prevotella copri correlates with enhanced susceptibility to arthritis. Elife 2013;2:e01202. DOI: 10.7554/eLife.01202. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 13.Pianta A, Arvikar S, Strle K, et al. Evidence of the Immune Relevance of Prevotella copri, a Gut Microbe, in Patients With Rheumatoid Arthritis. Arthritis Rheumatol 2017;69(5):964–975. (In eng). DOI: 10.1002/art.40003. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 14.Chriswell ME, Lefferts AR, Clay MR, et al. Clonal IgA and IgG autoantibodies from individuals at risk for rheumatoid arthritis identify an arthritogenic strain of Subdoligranulum. Sci Transl Med 2022;14(668):eabn5166. (In eng). DOI: 10.1126/scitranslmed.abn5166. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 15.Zaiss MM, Joyce Wu HJ, Mauro D, Schett G, Ciccia F. The gut-joint axis in rheumatoid arthritis. Nat Rev Rheumatol 2021;17(4):224–237. (In eng). DOI: 10.1038/s41584-021-00585-3. [DOI] [PubMed] [Google Scholar]
  • 16.Correa-Oliveira R, Fachi JL, Vieira A, Sato FT, Vinolo MA. Regulation of immune cell function by short-chain fatty acids. Clin Transl Immunology 2016;5(4):e73. DOI: 10.1038/cti.2016.17. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 17.Mizuno M, Noto D, Kaga N, Chiba A, Miyake S. The dual role of short fatty acid chains in the pathogenesis of autoimmune disease models. PLoS One 2017;12(2):e0173032. DOI: 10.1371/journal.pone.0173032. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 18.Kim DS, Kwon JE, Lee SH, et al. Attenuation of Rheumatoid Inflammation by Sodium Butyrate Through Reciprocal Targeting of HDAC2 in Osteoclasts and HDAC8 in T Cells. Front Immunol 2018;9:1525. DOI: 10.3389/fimmu.2018.01525. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 19.Martinsson K, Durholz K, Schett G, Zaiss MM, Kastbom A. Higher serum levels of short-chain fatty acids are associated with non-progression to arthritis in individuals at increased risk of RA. Ann Rheum Dis 2022;81(3):445–447. DOI: 10.1136/annrheumdis-2021-221386. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 20.Chang H-W, Yan D, Singh R, et al. Multiomic Analysis of the Gut Microbiome in Psoriasis Reveals Distinct Host-Microbe Associations. JID Innovations 2022;2(3):100115. DOI: 10.1016/j.xjidi.2022.100115. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 21.Yun H, Wang X, Wei C, et al. Alterations of the intestinal microbiome and metabolome in women with rheumatoid arthritis. Clinical and Experimental Medicine 2023;23(8):4695–4706. DOI: 10.1007/s10238-023-01161-7. [DOI] [PubMed] [Google Scholar]
  • 22.Marazzato M, Iannuccelli C, Guzzo MP, et al. Gut Microbiota Structure and Metabolites, Before and After Treatment in Early Rheumatoid Arthritis Patients: A Pilot Study. Frontiers in Medicine 2022;9. DOI: 10.3389/fmed.2022.921675. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 23.Artacho A, Isaac S, Nayak R, et al. The Pretreatment Gut Microbiome Is Associated With Lack of Response to Methotrexate in New-Onset Rheumatoid Arthritis. Arthritis Rheumatol 2021;73(6):931–942. (In eng). DOI: 10.1002/art.41622. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 24.Blanco-Míguez A, Beghini F, Cumbo F, et al. Extending and improving metagenomic taxonomic profiling with uncharacterized species using MetaPhlAn 4. Nature Biotechnology 2023;41(11):1633–1644. DOI: 10.1038/s41587-023-01688-w. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 25.Beghini F, McIver LJ, Blanco-Miguez A, et al. Integrating taxonomic, functional, and strain-level profiling of diverse microbial communities with bioBakery 3. Elife 2021;10. DOI: 10.7554/eLife.65088. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 26.Visconti A, Martin TC, Falchi M. YAMP: a containerized workflow enabling reproducibility in metagenomics research. Gigascience 2018;7(7). DOI: 10.1093/gigascience/giy072. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 27.Bolyen E, Rideout JR, Dillon MR, et al. Reproducible, interactive, scalable and extensible microbiome data science using QIIME 2. Nat Biotechnol 2019;37(8):852–857. DOI: 10.1038/s41587-019-0209-9. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 28.Lakens D. Calculating and reporting effect sizes to facilitate cumulative science: a practical primer for t-tests and ANOVAs. Front Psychol 2013;4:863. (In eng). DOI: 10.3389/fpsyg.2013.00863. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 29.Yue M, Kim JH, Evans CR, et al. Measurement of Short-Chain Fatty Acids in Respiratory Samples: Keep Your Assay above the Water Line. Am J Respir Crit Care Med 2020;202(4):610–612. DOI: 10.1164/rccm.201909-1840LE. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 30.Paine A, Brookes PS, Bhattacharya S, et al. Dysregulation of Bile Acids, Lipids, and Nucleotides in Psoriatic Arthritis Revealed by Unbiased Profiling of Serum Metabolites. Arthritis Rheumatol 2023;75(1):53–63. (In eng). DOI: 10.1002/art.42288. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 31.Sokolove J, Bromberg R, Deane KD, et al. Autoantibody Epitope Spreading in the Pre-Clinical Phase Predicts Progression to Rheumatoid Arthritis. PLoS ONE 2012;7(5):e35296. DOI: 10.1371/journal.pone.0035296. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 32.Kronzer VL, Hayashi K, Yoshida K, et al. Autoantibodies against citrullinated and native proteins and prediction of rheumatoid arthritis-associated interstitial lung disease: A nested case-control study. Lancet Rheumatol 2023;5(2):e77–e87. (In eng). DOI: 10.1016/s2665-9913(22)00380-0. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 33.Assarsson E, Lundberg M, Holmquist G, et al. Homogenous 96-Plex PEA Immunoassay Exhibiting High Sensitivity, Specificity, and Excellent Scalability. PLoS ONE 2014;9(4):e95192. DOI: 10.1371/journal.pone.0095192. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 34.Bu K, Wallach DS, Wilson Z, et al. Identifying correlations driven by influential observations in large datasets. Briefings in Bioinformatics 2022;23(1). DOI: 10.1093/bib/bbab482. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 35.Cohen J. Statistical power analysis for the behavioral sciences. 2nd ed. Hillsdale, N.J.: L. Erlbaum Associates, 1988. [Google Scholar]
  • 36.McKinney W. Data Structures for Statistical Computing in Python. SciPy:56–61. [Google Scholar]
  • 37.Waskom M. seaborn: statistical data visualization. Journal of Open Source Software 2021;6(60):3021. DOI: 10.21105/joss.03021. [DOI] [Google Scholar]
  • 38.Wickham H. ggplot2 : Elegant Graphics for Data Analysis. Use R!,. 2nd ed. Cham: Springer International Publishing : Imprint: Springer,; 2016:1 online resource (XVI, 260 pages 232 illustrations, 140 illustrations in color. [Google Scholar]
  • 39.Team RC. R: A language and environment for statistical computing. R Foundation for Statistical Computing. Vienna, Austria: 2021. [Google Scholar]
  • 40.Liu X, Mao B, Gu J, et al. Blautia—a new functional genus with potential probiotic properties? Gut Microbes 2021;13(1):1875796. DOI: 10.1080/19490976.2021.1875796. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 41.Verstraeten S, Sencio V, Raise A, et al. Description of a Newly Isolated Blautia faecis Strain and Its Benefit in Mouse Models of Post-Influenza Secondary Enteric and Pulmonary Infections. Nutrients 2022;14(7):1478. DOI: 10.3390/nu14071478. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 42.Ramos MI, Perez S, Aarrass S, et al. FMS-related tyrosine kinase 3 ligand (Flt3L)/CD135 axis in rheumatoid arthritis. Arthritis Research & Therapy 2013;15(6):R209. DOI: 10.1186/ar4403. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 43.Svensson MND, Erlandsson MC, Jonsson IM, Andersson KME, Bokarewa MI. Impaired signaling through the FMS-like tyrosine kinase 3 receptor increases osteoclast formation and bone damage in arthritis. Journal of Leukocyte Biology 2016;99(3):413–423. (In English). DOI: 10.1189/jlb.3HI1114-572RR. [DOI] [PubMed] [Google Scholar]
  • 44.Svensson MN, Andersson SE, Erlandsson MC, et al. Fms-like tyrosine kinase 3 ligand controls formation of regulatory T cells in autoimmune arthritis. PLoS One 2013;8(1):e54884. DOI: 10.1371/journal.pone.0054884. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 45.Hensvold AH, Frisell T, Magnusson PK, Holmdahl R, Askling J, Catrina AI. How well do ACPA discriminate and predict RA in the general population: a study based on 12 590 population-representative Swedish twins. Ann Rheum Dis 2017;76(1):119–125. DOI: 10.1136/annrheumdis-2015-208980. [DOI] [PubMed] [Google Scholar]
  • 46.Picchianti-Diamanti A, Panebianco C, Salemi S, et al. Analysis of Gut Microbiota in Rheumatoid Arthritis Patients: Disease-Related Dysbiosis and Modifications Induced by Etanercept. International Journal of Molecular Sciences 2018;19(10):2938. DOI: 10.3390/ijms19102938. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 47.Andréasson K, Olofsson T, Lagishetty V, et al. Treatment for Rheumatoid Arthritis Associated With Alterations in the Gastrointestinal Microbiota. ACR Open Rheumatology;n/a(n/a). DOI: 10.1002/acr2.11673. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 48.Zhou B, Dong C, Zhao B, et al. Bacteroides fragilis participates in the therapeutic effect of methotrexate on arthritis through metabolite regulation. Front Microbiol 2022;13(1664-302X (Print)) (In eng). DOI: 10.3389/fmicb.2022.1015130. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 49.Meng X, Shu Q. Novel primers to identify a wider diversity of butyrate-producing bacteria. World Journal of Microbiology and Biotechnology 2024;40(2):76. DOI: 10.1007/s11274-023-03872-1. [DOI] [PubMed] [Google Scholar]
  • 50.Zaplana T, Miele S, Tolonen AC. Lachnospiraceae are emerging industrial biocatalysts and biotherapeutics. Frontiers in Bioengineering and Biotechnology 2024;11. DOI: 10.3389/fbioe.2023.1324396. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 51.Muñiz Pedrogo DA, Chen J, Hillmann B, et al. An Increased Abundance of Clostridiaceae Characterizes Arthritis in Inflammatory Bowel Disease and Rheumatoid Arthritis: A Cross-sectional Study. Inflammatory Bowel Diseases 2019;25(5):902–913. DOI: 10.1093/ibd/izy318. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 52.Rogier R, Evans-Marin H, Manasson J, et al. Alteration of the intestinal microbiome characterizes preclinical inflammatory arthritis in mice and its modulation attenuates established arthritis. Scientific Reports 2017;7(1). DOI: 10.1038/s41598-017-15802-x. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 53.Gomez A, Luckey D, Yeoman CJ, et al. Loss of Sex and Age Driven Differences in the Gut Microbiome Characterize Arthritis-Susceptible *0401 Mice but Not Arthritis-Resistant *0402 Mice. PLoS ONE 2012;7(4):e36095. DOI: 10.1371/journal.pone.0036095. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 54.Clemente JC, Manasson J, Scher JU. The role of the gut microbiome in systemic inflammatory disease. BMJ 2018;360:j5145. DOI: 10.1136/bmj.j5145. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 55.Mäkivuokko H, Tiihonen K, Tynkkynen S, Paulin L, Rautonen N. The effect of age and non-steroidal anti-inflammatory drugs on human intestinal microbiota composition. British Journal of Nutrition 2010;103(2):227–234. DOI: 10.1017/s0007114509991553. [DOI] [PubMed] [Google Scholar]
  • 56.Rogers MAM, Aronoff DM. The influence of non-steroidal anti-inflammatory drugs on the gut microbiome. Clinical Microbiology and Infection 2016;22(2):178.e1–178.e9. DOI: 10.1016/j.cmi.2015.10.003. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 57.Prakash A, Peters BA, Cobbs E, et al. Tobacco Smoking and the Fecal Microbiome in a Large, Multi-ethnic Cohort. Cancer Epidemiol Biomarkers Prev 2021;30(7):1328–1335. DOI: 10.1158/1055-9965.EPI-20-1417. [DOI] [PMC free article] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

1
NIHMS2199892-supplement-1.docx (709.3KB, docx)

RESOURCES