The intestinal tract harbors are the largest bacterial community within the human body. During development into adulthood, gut microbiota shapes the tissues, cells, and molecular profiles of the human intestinal immune system [1
]. Microbiome, the total amount of the genes of microbial cells, is closely related to the development of innate and adaptive immune responses [2
]. In healthy microbiota, there is an optimal proportion of pro- and anti-inflammatory organisms that provide signals to the developing immune system, resulting in a balance of type 17 helper T (Th17) and regulatory T (Treg) cells activities [4
]. It has been hypothesized that the interactions between microbes and host factors lead to mucosal inflammation and the breaking of immune tolerance [6
]. Many genes in the human microbiome generate proteins that can enter the circulation and affect pro- or anti-inflammatory responses [2
]. The disruption of the intestinal mucosal barrier may lead to increased proinflammatory cytokines such as tumor necrosis factor (TNF)-α and interleukin (IL)-17A, subsequently inducing a chronic inflammation [8
]. Recent studies reveal that Collinsella species may contribute to RA pathogenesis by increasing gut permeability, lowering the expression of tight junction proteins and influencing the epithelial production of IL-17A [9
]. Moreover, Zhang et al. demonstrated overexpression of genus Prevotella
and concurrent decrease in genus Bacteroides
in the feces from new-onset RA patients, suggesting the potential of using microbiome composition for disease diagnosis [11
]. In addition, increased abundance of Prevotella copri has been revealed in the gut microbiota of early RA patients, while Prevotella histicola
from human gut microbiota could suppress the development of arthritis [12
]. Given that proinflammatory cytokines, such as TNF-α, IL-6, and IL-17A are crucial inflammatory mediators in synovitis and subsequent tissue damage in rheumatoid arthritis (RA) [14
], gut dysbiosis has been implicated in the pathogenesis of RA [15
]. However, an association of gut microbiota with inflammatory parameters of RA remained yet to be explored.
The 16S ribosomal RNA (rRNA) sequences could represent taxonomic signatures due to their highly conserved homology for common probes. Next-generation sequencing (NGS) on 16S rRNA gene and the whole genome shotgun sequencing analysis have been used as a comprehensive analysis tool for assessing bacterial composition and diversity [12
This study aimed to investigate: (1) the differential expression of gut microbiota in RA patients compared with healthy controls (HC); (2) the associations of gut microbiota with inflammatory parameters including disease activity and serum levels of TNF-α, IL-6 and IL-17A; and (3) the associations of gut microbiota with the positivity of rheumatoid factor (RF) or anti-citrullinated peptides antibodies (ACPA), which is related to disease severity of RA [18
In the present study, we performed 16s rRNA deep sequencing and obtained large amounts of data concerning the microbial composition, diversity, and abundance in healthy subjects and RA patients with various clinical phenotypes. The results have demonstrated that a lower abundance and evenness of gut microbiota in RA patients compared with healthy subjects, with α-diversity lower in RF-positive or ACPA-positive RA patients than healthy subjects. Using the permutational multivariate analysis of variance (PERMANOVA), the β-diversity of gut microbiota in healthy subjects was significantly different from that in RA patients with different phenotypes. The phylum Verrucomicrobiae and genus Akkermansia were more abundant in RA patients compared with healthy subjects. Among RA patients, those with higher levels of TNF-α or IL-17A had increased the relative abundance of the phylum Gammaproteobacteria compared to those with lower cytokine levels. In addition, RA patients exhibited a different microbial composition which was related to RF/ACPA seropositivity.
Our RA patients in either active or inactive status had a lower α-diversity of gut microbial taxa than healthy subjects, consistent with the findings in Korean patients with early RA [41
]. Moreover, α-diversity of gut microbiota was lower in ACPA-positive patients compared to HC. Although Chen et al. revealed similar findings [15
], the significant association of α-diversity with ACPA positivity in their RA patients could not be reproduced in the present study. Our statistical data has suggested that there was no significant difference which may be due to the small sample size of RA subgroups with different phenotypes. In addition, the production of ACPA may be multifactorial, such as genetic factor, smoking, and the severity of periodontitis [42
]. After the pooling of RF-positive and ACPA-positive patients in a single goup, a significantly lower α-diversity was observed in seropositive patients compared to seronegative patients.
Regarding the difference in the β-diversity between healthy subjects and RA patients, the PERMANOVA analysis showed a significant difference in phenotypes characterized by different disease activity and RF/ACPA status (Table 2
). The weighted UniFrac distance, qualitative measurement of phylogenetic abundance difference, showed a statistical significance between RA patients and healthy subjects. There was a trend of difference finding in the β-diversity between patients with the lowest and the highest levels of TNF-α or IL-17A, indicating an association of gut microbiota composition with the inflammatory parameters in RA patients.
Using pre-trained Naive Bayes Greengenes classifier to explore the taxonomic composition of gut microbiota, we observed an increased abundance of phylum Verrucomicrobiae, which might be driven by the genus Akkermansia (Akkermansia muciniphila
), in RA patients compared with healthy subjects. Our findings support the results revealing that the implication of Akkermansia muciniphila
in the pathogenesis of murine model of arthritis [43
]. Akkermansia has also been associated with the proinflammatory pathways, including upregulation of B- and T-cell receptor signaling [44
]. These proinflammatory effects may be related to its ability to degrade mucus and thus increase the exposure of resident immune cells to gut microbial antigens [45
]. In contrast to previous studies revealing that the predominance of Prevotella in the gut microbiota was associated with the untreated or new-onset RA [1
] and phylum Bacteroidetes was enriched in early RA patients [41
], we did not observe significant increase in the abundance of family Prevotellaceae or
phylum Bacteroidetes. The differences in patients’ characteristics and treatment regimens of our enrolled patients may explain this discrepancy.
It is interesting that active RA patients have an increased abundance of Collinsella compared with inactive RA patients. The arthrogenic role of Collinsella has been confirmed both in vivo using a humanized murine model of arthritis and in vitro by a human intestinal epithelium cell-based study [10
]. Chen et al. further demonstrated that Collinsella could be transferred to germ-free mice with resultant exacerbation of arthritis in a murine model [10
]. These observations suggest a potential role of the Collinsella expansion in the pathogenesis of active RA.
Accumulating evidence indicates that gut dysbiosis plays an important role in T-cell polarization towards a pro-inflammatory phenotype [46
]. Schirmer et al. demonstrate that human gut microbiome-host interactions modulate the production of inflammatory cytokines [47
]. The results of a recent study also show that germ-free mice conventionalized with the gut microbiota from collagen-induced arthritis-susceptible mice, which have higher levels of serum IL-17, develop greater severity of arthritis [48
]. In addition, Maeda et al. reveal that the gut microbiota transplant from RA patients to arthritis-prone SKG mice induced an increased number of intestinal Th17 cells and severe arthritis [12
]. Therefore, we stratified RA patients based on serum cytokine levels. Our results showed a significantly higher proportion of the phylum Gammaproteobacteria in RA patients with higher levels of TNF-α or IL-17A compared with those with lower cytokines levels. Similarly, we revealed a positive correlation between the abundance of the phylum Euryarchaeota with serum levels of IL-6 or IL-17A. These findings support the findings of a recent study showing a direct correlation between the proinflammatory property of Gammaproteobacteria and RA disease activity [49
]. The differential abundance analysis with Gneiss also revealed that RA patients with higher levels of TNF-α or IL-17A had higher proportions of Enterobacteriaceae (belonging to Gammaproteobacteria) and Klebsiella than those with low cytokines levels. The lipopolysaccharide (LPS) of Enterobacteriaceae and Klebsiella could promote inflammation [50
] by increasing intestinal permeability [51
]. The disruption of the intestinal mucosal barrier may lead to increased proinflammatory cytokines such as TNF-α and IL-17A [8
] and alter the interactions between gut mucosa and environment [46
]. In contrast, patients with higher levels of TNF-α or IL-17A have less abundance of Bifidobacterium (Bifidobacterium adolescentis and Bifidobacterium longum) compared with those with lower cytokines levels. Previous studies have similarly demonstrated that decreased abundance of Bifidobacterium in active RA [52
]. Given that Bifidobacterium could reduce inflammation by inducing the production of immunosuppressive Treg cells [54
], the decreased abundance of Bifidobacterium may contribute to the inflammatory responses in RA patients. In addition, there is a discrepancy in the relative abundance of Klebsiella between overall RA patients and those with high TNF or high IL-6 levels. Our findings suggest the involvement of gut microbiota varies in the different subsets of RA patients.
Regarding the relation between seropositivity and gut microbiota, we revealed that RF-positive patients have higher proportion of Blautia and Collinsella than RF-negative patients. ACPA-positive patients also have higher proportions of Blautia, Akkermansia, and Clostridiales compared with ACPA-negative patients. The association of Blautia with seropositivity of RF/ACPA has not been reported previously. The arthrogenic role of Collinsella has been confirmed in both cell-based and humanized murine models of arthritis [15
], possibly by means of increasing gut mucosal permeability observed in RA [55
]. In addition, gut microbiota enriched with Akkermansia and Clostridiales has been found in RA patients or inflammatory bowel disease-associated arthropathy [43
]. However, the causal relationship between RF/ACPA seropositivity and the composition of gut microbiota needs to be elucidated in the future.
The understanding of the dysbiosis of gut microbiota in human RA has opened up a window to more therapeutic opportunities such as modulating microbes through diet and probiotics [57
]. Probiotic administration with immunomodulatory microbial species has shown some benefits [56
], and intervention targeting gut microbiota in RA patients has also shown the potential to revolutionize the modern therapeutics [59
Despite the novel findings presented here, some limitations should be considered. Firstly, feces contain bacterial colonies from both the lumen and the mucosa of intestinal tracts, and the fecal microbiota cannot represent the microbiota in the intestinal mucosa. Secondly, the proportions of various bacterial Gneiss in gut microbiota are low among RA patients with different phenotypes (Figure 4
and Figure 5
). The bacterial numbers using specific primer by qPCR will be needed in the future study. In addition, given that the enrolled RA patients had received different therapeutic agents before this investigation, the effects of the medications used on fecal microbiota should be considered [11
]. However, it is difficult to obviate such limitation in tertiary-care settings in Taiwan. Thirdly, the small sample size of RA patients with seropositivity for RF or ACPA may explain the lack of statistical significance regarding the diversity of gut microbiota. Therefore, our preliminary results require further confirmation by larger cohort studies.