Advertisement
Browse Subject Areas
?

Click through the PLOS taxonomy to find articles in your field.

For more information about PLOS Subject Areas, click here.

  • Loading metrics

The Proportion of Regulatory T Cells in Patients with Rheumatoid Arthritis: A Meta-Analysis

  • Takayoshi Morita,

    Affiliations Department of Respiratory Medicine, Allergy and Rheumatic diseases, Osaka University Graduate School of Medicine, Osaka, Japan, Department of Experimental Immunology, Immunology Frontier Research Center, Osaka University, Osaka, Japan

  • Yoshihito Shima,

    Affiliation Department of Respiratory Medicine, Allergy and Rheumatic diseases, Osaka University Graduate School of Medicine, Osaka, Japan

  • James Badger Wing,

    Affiliation Department of Experimental Immunology, Immunology Frontier Research Center, Osaka University, Osaka, Japan

  • Shimon Sakaguchi,

    Affiliations Department of Experimental Immunology, Immunology Frontier Research Center, Osaka University, Osaka, Japan, Department of Experimental Pathology, Institute for Frontier Medical Sciences, Kyoto University, Kyoto, Japan

  • Atsushi Ogata ,

    ogata@imed3.med.osaka-u.ac.jp

    Affiliations Department of Respiratory Medicine, Allergy and Rheumatic diseases, Osaka University Graduate School of Medicine, Osaka, Japan, Division of Allergy, Rheumatology and Connective tissue disease, NTT west Osaka Hospital, Osaka, Japan

  • Atsushi Kumanogoh

    Affiliation Department of Respiratory Medicine, Allergy and Rheumatic diseases, Osaka University Graduate School of Medicine, Osaka, Japan

The Proportion of Regulatory T Cells in Patients with Rheumatoid Arthritis: A Meta-Analysis

  • Takayoshi Morita, 
  • Yoshihito Shima, 
  • James Badger Wing, 
  • Shimon Sakaguchi, 
  • Atsushi Ogata, 
  • Atsushi Kumanogoh
PLOS
x

Abstract

Background

Regulatory T cells (Tregs) have important functions in peripheral immune tolerance. Dysfunction of Tregs is considered to be a pivotal cause of autoimmune diseases, including rheumatoid arthritis (RA). However, previous reports describing the proportion of Tregs among CD4+ T cells in RA patients were controversial because a range of markers are used to identify Tregs with little consensus. To clarify the status of Tregs in RA, we investigated the proportion of Tregs with focusing on the definitions of them.

Methods

We identified the studies reporting the proportion of Tregs in RA patients using PubMed and Google Scholar. We performed a systematic review of them and a meta-analysis to evaluate the proportion of Tregs (FOXP3-positive and/or CD25-positive) among CD4+ T cells in peripheral blood (PB) and synovial fluid (SF) of RA patients and control subjects.

Results

A total 31 studies were selected. The proportion of Tregs defined by all definitions among CD4+ T cells in PB was not significantly different between RA patients and control subjects (-0.65, [-1.30, 0.01]). Then we performed sub-analyses based on individual definitions. The proportion of Tregs defined by either CD25 or FOXP3 alone did not differ between RA patients and control subjects. The proportion of Tregs defined by both FOXP3 and CD25 was lower in RA patients than that in control subjects (-2.42 [-3.49, -1.34]). The proportion of Tregs defined by both FOXP3 and CD25 was higher in SF than that in PB among RA patients (3.27 [0.40, 6.14]).

Conclusion

The status of Tregs varied according to the definition system. The proportion of Tregs defined by stricter and functionally validated methods decreased in PB and increased in SF among RA patients. If the proportion of Tregs differs in RA, accurate and functionally relevant definitions of Tregs are necessary to elucidate their status in RA.

Introduction

Rheumatoid arthritis (RA) is a typical disease of chronic inflammation resulting from impaired immune homeostasis [1]. Clinical studies have shown that inflammatory cytokines such as tumor necrosis factor (TNF)-α and interleukin (IL)-6 are present at high levels in the sera of RA patients and play an important role in RA pathogenesis [2]. In addition, effector T cells such as the T helper17 (Th17) subset of CD4+ T cells producing IL-17 have been found to play important roles in the pathology of RA [3]. Furthermore, auto-reactive Th17 cells contribute to arthritogenesis in the SKG model of mouse arthritis [4]. In contrast, specialized regulatory cells suppress autoimmune response and maintain immune tolerance. Regulatory T cells (Tregs) are a distinct set of thymically produced T cells responsible for suppressing autoreactive deleterious activities of effector T cells [5]. The forkhead box P3 (FOXP3) transcription factor is considered the most reliable marker for Tregs and is essential to their regulatory function [6]. Dysfunction of FOXP3 is the cause of IPEX (immune dysregulation, polyendocrinopathy, enteropathy, and X-linked) syndrome, which is associated with refractory diarrhea, type1 diabetes, hypothyroidism, and glomerulonephritis [6]. IPEX is a fulminant autoimmune disease resulting from complete failure of self-tolerance. Failure of suppressive functions of Tregs is a possible cause of arthritogenesis as arthritis has also been observed in IPEX patients [7]. Consequently, the balance of Tregs and effector cells is critical to the maintenance of the self-tolerance.

Recently, the significance of Tregs in RA pathology has been a focus of much research. In the collagen induced arthritis (CIA) mouse model, reduction of Tregs and induction of Th17 cells has been reported [8]. In this model, a reduction in the proportion of Tregs following treatment with anti-CD25 antibody exacerbates arthritis [9], whereas injection of Tregs ameliorates the symptoms [10]. In the SKG model, arthritis is exacerbated by a reduction of FOXP3 function [4]. In human, FOXP3 genetic polymorphism and the FOXP3 protein level may be associated with susceptibility to RA [11]. Polymorphisms of cytotoxic T-lymphocyte associated antigen 4 (CTLA-4), one of the main suppressive molecules in Tregs [12], have been reported in RA patients and are associated with RA risk and reduced suppressive functions of Tregs [13]. Meanwhile, inflammatory cytokines such as TNF-α and IL-6 ameliorates the function of Tregs. TNF-α suppresses the expression of FOXP3 [14], while IL-6 suppresses the induction of Tregs from naïve T cells and promotes the induction of Th17 cells, thereby exacerbating RA pathology [15]. Although the results of basic studies strongly suggest that Tregs in RA play a role in the pathology of RA, the function of Tregs in peripheral immune tolerance of RA patients remains controversial.

The proportion of Tregs among CD4+ T cells in peripheral blood (PB) is considered a reliable method for the estimation of Tregs function in peripheral immune tolerance. Molecular markers are essential tools for identifying Tregs. Since Tregs express high levels of CD25, the alpha-chain of the IL-2 receptor, Tregs were first defined as CD25-positive cells [16]. Further to this, CTLA4, CD25, FOXP3, and loss of CD127 are widely used as markers for Tregs. However, these markers are not truly specific for Tregs as activated non-Tregs may also upregulate their expression, making it difficult to distinguish Tregs from activated effector T cells [17, 18]. As a result, reports of the status of Tregs in humans must be subjected to careful analysis.

In this study, we systemically reviewed original studies in which authors documented the proportion of Tregs among CD4+ T cells in PB or SF of RA patients, and performed a meta-analysis to evaluate the proportion of Tregs in RA patients with focusing on definition of Tregs.

Materials and Methods

Data sources and searches

We searched for studies documenting the proportion of Tregs among CD4+ T cells in PB of RA patients using two search websites, PubMed and Google Scholar. On PubMed, searches identified publications with articles containing the term “rheumatoid arthritis,” (“regulatory T,” or “Treg,”) and “human and patient”, but not containing the term “review or mice or rat” (Indeed, we performed searching by [“rheumatoid arthritis" AND ("regulatory T" OR Treg) AND (human OR patient) NOT (review OR mice OR rat)]. On Google Scholar, searches identified publications with titles containing the term “rheumatoid arthritis” and “regulatory T,” “Treg,” “CD25,” or “FOXP3” (Indeed, we performed searching by [allintitle:"rheumatoid arthritis" ("regulatory T" OR "Treg" OR "CD25" OR "Foxp3")]. Searches were aimed at identifying studies published between January 1, 2005 and May 31, 2016. We searched for studies four times using the same terms and the same websites. Final searches were performed on June 18, 2016.

Study selection

Inclusion criteria: 1) original studies (not reviews); 2) human research; 3) title or abstract including the terms “rheumatoid arthritis” and “regulatory T” (or “Treg”); 4) studies reporting the proportion of Tregs among CD4+ T cells in PB of RA patients; 5) available on the internet; the manuscript is linked from search site to the full text (PDF or website) of the manuscript. Simultaneously, we verified that all RA patients in the selected studies were diagnosed according to the American college of rheumatology (ACR) criteria from 1987 or the ACR/ The European League Against Rheumatism (EULAR) criteria from 2010.

Exclusion criteria: 1) studies that provided no raw data regarding the average and standard deviation (SD) of the proportion of Tregs among CD4+ T cells in PB of RA patients or control subjects; 2) studies that provided no raw data about the number of RA patients or control subjects. Redundancies between PubMed and Google Scholar search were eliminated, i.e., individual studies were counted only once in the analysis.

Validity and quality assessment

Two authors independently checked and selected all references. We assessed the evidence level of selected studies based on guidelines from the Oxford Center for Evidence-Based Medicine 2011 [19]. We validated and performed a quality assessment of our systematic review based on the Newcastle–Ottawa Quality Assessment Scale (Case Control Studies) [20]; the Newcastle-Ottawa Scale is a tool for assessing the quality of non-randomized studies in meta-analyses.

Data extraction

We extracted the following data from all assessed studies: author, country, year of publication, number of RA patients and control subjects (including healthy persons and patients with osteoarthritis [OA]), mean (or median) and SD of the proportion of Tregs among CD4+ T cells in PB of RA patients and control subjects, and definition of Tregs. When Tregs were defined by several patterns such as CD25-positive, CD25-high, FOXP3-positive, CD25+ CD127-low, and/or CD25+ FOXP3+ in a study, we extracted all types of data. When a study reported the proportion of Tregs among CD4+ T cells in PB of both active (acute) RA and remission (late phase) RA, we extracted both types of data. When a study reported the proportion of Tregs among CD4+ T cells in synovial fluid (SF), we extracted the SF data as well. In addition, we extracted the background data of RA patients (age, proportion of female subjects, disease duration, positive rate of rheumatoid factor [RF], disease activity score 28 [DAS28], C reactive protein [CRP] level, erythrocyte sedimentation rate [ESR], and treatment information).

Quantitative data synthesis

We performed quantitative data synthesis based on the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) Statement [21] (S1 File). We performed a meta-analysis to provide quantitative summary estimates of the standardized mean difference (SMD); the proportion of Tregs among CD4+ T cells in PB of RA patients minus that of control subjects, the proportion of Tregs among CD4+ T cells in PB of active RA patients minus that of remission RA patients, or the proportion of Tregs among CD4+ T cells in SF of RA patients minus that in PB of RA patients. Summary SMD was calculated using a random-effects model (REM) (DerSimonian and Laird method) [22]. REM was performed in our analysis because there were clear differences in the experimental methods or techniques among selected studies. When we evaluated publication bias in all selected studies by funnel plot, summary averages were also calculated using a fixed-effects model according to Peto and Mantel-Haenszel. Statistical significance was defined as P < 0.05. The I2 statistic was calculated to assess statistical heterogeneity across studies [23]. For I2 values >50%, we explored characteristics of individual studies and subgroups of the main body of evidence. All analyses were conducted using R version 3.1.2 (R project for Statistical Computing) and EZR version 1.29 [24].

Results

Study characteristics

We found 217 studies on PubMed by the search terms. Of these, 141 were not selected because they did not satisfy the inclusion criteria, and 51 were excluded because they satisfied the exclusion criteria. Ultimately, we selected 25 studies from PubMed. Similarly, we found 176 studies on Google Scholar by the search terms. Of these, 118 studies were not selected because they did not satisfy the inclusion criteria, and 36 were excluded because they satisfied the exclusion criteria. 16 were eliminated because they had already been selected by the PubMed search. Ultimately, we selected six studies from Google Scholar. In total, 31 studies were used for our analysis (Fig 1, Table 1) [2555]. Study (a), (d), (i), (o), (r), (s), (u), (v) and (y) were found on PubMed only, (b), (c), (l), (w), (aa), and (ad) were found on Google Scholar only, and the others were found on both PubMed and Google scholar.

thumbnail
Table 1. Study information about proportion of Tregs in RA patients or control subjects

https://doi.org/10.1371/journal.pone.0162306.t001

We provide the details of the selected studies in S1 Table. We extracted clinical data regarding RA patients; ranges were as follows: average age, 33–68 years; proportion of female subjects, 17–100%; disease duration, 0.8–16 years; average CRP value, 0.2–24.0 mg/dl; average ESR, 20–77 mm/hour; DAS28, 3.1–7.0; positive ratio of RF, 56–90%. Patients described in the studies were treated with disease-modifying anti-rheumatic-drugs (DMARDs), methotrexate (MTX), corticosteroids (CS), or biological agents such as TNF-α inhibitor (iTNF) and IL-6 receptor inhibitor (tocilizumab; TCZ). Most control subjects were healthy individuals, but OA patients were selected as controls in one study that evaluated SF. The definition of Tregs was various (Table 1) (25–55). The proportion of Tregs among CD4+ T cells was 0.4–23.0% in PB of RA patients and 0.20–26.4% in PB of control subjects. The studies reported data from 6–101 RA subjects and 8–44 control subjects. The selected studies were published by groups in USA, UK, France, Germany, Italy, Spain, Finland, Japan, Korea, China, Romania, Taiwan, Mexico, Oman, and Egypt. All studies were evidence level 4 (poor-quality case-control studies or case series). Although not all selected studies were case-control studies per se (e.g., some included bench research), we regarded them as such and scored them by The Newcastle Ottawa Quality Assessment Scale (NOQAS). The NOQAS score was 3–7 for all studies (S2 Table).

Publication bias was evaluated utilizing a funnel plot (S1 Fig), which indicated that each study was located apparently symmetrically.

The proportion of Tregs in peripheral blood of RA patients

At first, we performed meta-analysis of the proportion of Tregs between RA patients and control subjects regardless of the definition of Tregs (Fig 2). There was no significant difference in the proportion of Tregs among CD4+ T cells in PB between RA patients and control subjects in all studies (a)–(ae) (REM -0.65, [95% CI: (-1.30 to 0.01)]). Surprisingly, some studies showed quite different results. Some studies showed that the proportion of Tregs was higher in controls (b, h, j, m, n, p, q, u, w, x, ac). Other studies showed that the proportion of Tregs was higher in RA (l, r, z, aa, ab). Heterogeneity assessed by I2 statics was 96.3% (P < 0.0001) indicated extreme heterogeneity.

thumbnail
Fig 2. Forest plot for combined meta-analysis of the proportion of Tregs in peripheral blood.

Standardized mean difference (the proportion of Tregs among CD4+ T cells in PB of RA patients minus that of control subjects) was estimated by meta-analysis. All definitions of Tregs were selected for analysis. Studies distinguished by study identifier (a)–(ae) (Table 1).

https://doi.org/10.1371/journal.pone.0162306.g002

We hypothesized that the cause of significant dispersion may be different definitions of Tregs. Then we performed subgroup analysis based on individual definitions. First, we analyzed the studies (l, v, x, aa, ab, ac, ae) in which “CD25-positive alone” was used to define Tregs (S2A Fig). Because the definition of Tregs in study h was IL-10+, IL17-, and CD25+, study h was removed from this sub-analysis. There was no significant difference in the proportion of Tregs among CD4+ T cells in PB between RA patients and control subjects (REM 0.88, [-1.46 to 3.21]) (I2 = 97.4%, P < 0.0001). Next, we also analyzed the studies (f, m, w, ac, ad) in which “CD25-high” was used to define Tregs (Fig 3). Some studies (k, u, z, ab) were removed from this sub-analysis because CD25-high gating strategy was not clearly demonstrated. The cutoff line defining CD25-high was shown and either clearly higher than non-CD4 cells (f, w, ad), higher than CD45RA+CD25lo Naïve Tregs and CD45RA-CD25lo non-Tregs (m) or CD25 cells were demonstrated to be mostly FOXP3+ (ac). There was a significant reduction in the proportion of Tregs, when they were defined by “CD25-high” among CD4+ T cells in PB between RA patients and control subjects (REM -1.04, [-1.85 to -0.24]) (I2 = 84.2%, P < 0.0001).

thumbnail
Fig 3. Forest plot for sub-group meta-analysis of the proportion of Tregs defined by CD25-high.

Standardized mean difference (the proportion of Tregs among CD4+ T cells in PB of RA patients minus that of control subjects) was estimated by meta-analysis. Tregs, which were defined by “CD25-high”, were selected for sub-analysis.

https://doi.org/10.1371/journal.pone.0162306.g003

Tregs defined as “CD25-positive and CD127-negative” (e, f, y) showed no significant difference in the proportion of Tregs among CD4+ T cells in PB between RA patients and control subjects (REM -0.23, [-1.01 to 0.54]) (S2B Fig). Heterogeneity assessed by I2 statics was 76.4% (P = 0.0144), suggesting moderate heterogeneity.

Then, we analyzed the studies (i, r, s) in which Tregs were defined as “FOXP3-positive alone” (S2C Fig). There was no significant difference in the proportion of Tregs defined by “FOXP3-positive alone” in PB between RA patients and control subjects (REM 0.72, [-0.72 to 2.16]) (I2 = 86.5%, P = 0.0006). The proportion of Tregs defined by “CD25 and FOXP3 double positive” (a, c, j, n, o, p, q, t, ac) in PB of RA patients was lower than that of control subjects (REM -2.42, [-3.49 to -1.34]) (I2 = 96%, P < 0.0001) (Fig 4). Finally, when Tregs were defined by “CD25-high and FOXP3-positive” (b, m), the proportion of Tregs in PB of RA patients was significantly lower than that of control subjects (REM -1.03, [-1.42 to -0.63]) (S2D Fig). Heterogeneity assessed by I2 statics was 0% (P = 0.8443) suggesting very low heterogeneity.

thumbnail
Fig 4. Forest plot for sub-group meta-analysis of the proportion of Tregs defined by CD25 and FOXP3 double positive.

Standardized mean difference (the proportion of Tregs among CD4+ T cells in PB of RA patients minus that of control subjects) was estimated by meta-analysis. Tregs, which were defined by “CD25 and FOXP3 double positive”, were selected for sub-analysis.

https://doi.org/10.1371/journal.pone.0162306.g004

Disease activity of RA and the proportion of Tregs in PB of RA patients

Whether the proportion of Tregs among CD4+ T cells in PB differed between active RA patients and remission RA patients was analyzed (S3 Table). Five studies (c, j, l, n, w) evaluated the proportion of Tregs in active RA and remission RA patients. Studies (l, w) could not be used in sub-analysis because the definitions of Tregs in these two studies were CD25+ alone and CD25-bright respectively and only one study per group is insufficient for meta-analysis. Three of these studies (c, j, n) in which Tregs were defined by “CD25 and FOXP3 double positive” were selected to perform the meta-analysis, revealing that there was a significant reduction in the proportion of Tregs in PB of active RA patients compared to remission RA patients (REM -0.92, [-1.79 to -0.05]) (Fig 5A). Heterogeneity assessed by I2 statics was 86.4% (P = 0.0007).

thumbnail
Fig 5. Forest plots of the proportion of Tregs in peripheral blood of different diseases activity and synovial fluid.

(A) Standardized mean difference (the proportion of Tregs among CD4+ T cells in PB of active RA patients minus that of remission RA patients) was estimated by meta-analysis. Tregs, which were defined by “CD25 and FOXP3 double positive”, were selected for meta-analysis. (B) Standardized mean difference (the proportion of Tregs among CD4+ T cells in SF of RA patients minus that in PB of control subjects) was estimated by meta-analysis. Tregs, which were defined by “CD25 and FOXP3 double positive”, were selected for meta-analysis.

https://doi.org/10.1371/journal.pone.0162306.g005

The proportion of Tregs in synovial fluid of RA patients

The difference in the proportion of Tregs among CD4+ T cells of RA patients between SF and PB was evaluated (S4 Table). Five studies (a, f, k, ac, ae) evaluated the proportion of Tregs among CD4+ T cells in SF and PB of RA patients. Study (ac) provided data using three definitions of Tregs (CD25+, CD25-high, and CD25+ FOXP3+). To avoid the problem of variable definitions of Tregs, we did not use the data from two studies (ac, ae) in which Tregs were defined by CD25 positive alone in sub-analysis and instead focused on the definitions that previous research about Tregs and the results of our analysis of PBMC suggested are most reliable (CD25hi and CD25+ FOXP3+). Additionally, we avoided using studies (f, k, ac) in which Tregs were defined by CD25-high in sub-analysis because the CD25-high gating strategy in SF was not shown. Finally, two studies (a, ac) in which Tregs were defined as “CD25 and FOXP3 double positive” were selected for meta-analysis (REM 3.27, [0.40 to 6.14]) (Fig 5B); the results revealed that the proportion of Tregs among CD4+ T cells in SF was significantly higher than that in PB of RA patients. Heterogeneity assessed by I2 statics was 87% (P = 0.0056).

Discussion and Conclusions

Functional deficiency of Tregs has been considered a cause of autoimmune diseases including RA. However, the function of Tregs in peripheral immune tolerance of RA patients was not elucidated sufficiently in previous researches. Even in the proportion of Tregs in PB of RA patients, a common consensus about it was not provided from previous reports. Therefore, the status of Tregs in RA patients is currently under debate. To elucidate the status of Tregs in RA, we performed meta-analysis of the proportion of Tregs among CD4+ T cells in RA patients. Meta-analysis of all selected studies revealed that there was no significant difference in the proportion of Tregs between RA patients and control subjects. Surprisingly, some studies showed quite different results. Some studies showed that the proportion of Tregs was higher in control and other studies showed that the proportion of Tregs was higher in RA. The major problem was considered to be inconsistent definitions of Tregs. Our sub-analysis showed that stricter definitions reduced the heterogeneity of results and consistently came to the same conclusion. The proportion of Tregs, when defined as “CD25-high”, “CD25-high and Foxp3 double positive” or “CD25 and FOXP3 double positive”, among CD4+ T cells in the PB of RA patients was decreased. The definition of “CD25 high and FOXP3 double positive” showed minimum heterogeneity.

CD25 is a well-defined marker of Tregs. However, the proportion of Tregs defined by CD25 expression alone was not clearly associated with the status of proper Tregs with suppressive function. One reason for this may be that the CD25 positive population contains FOXP3 negative effecter T cells [56]. Although loss of CD127 expression is often utilized to exclude effector T cells, the proportion of Tregs, which were defined by CD25+ and CD127 negative, the results were varied among selected studies in our research. Sakaguchi et al. reported that only the 1–2% of cells with the highest CD25 expression exert suppressive functions, and should therefore be considered as true Tregs [57]. To address this, some studies assess the proportion of CD25-high cells, however the consensus of accurate cutoff to define CD25-high has been not established. Therefore, we selected studies in which the CD25-high CD4+ T cell population was clearly defined and the proportion of CD25-high CD4+ defined Tregs in RA patients was lower than that in control subjects with slightly lower dispersion. This suggested that excluding cells expressing low levels of CD25 (probable effecter CD4+ T cells) was important for evaluation of Tregs.

It is well known that the FOXP3 gene is a master regulator of Tregs and expression of FOXP3 is also utilized as a very important marker of Tregs. In our study, however, the proportion of Tregs, which were defined by FOXP3 positive alone, among CD4+ T cells in PB of RA patients was not significantly different to control subjects. The dispersion of results was extreme. Instead, the proportion of Tregs defined by “CD25 and FOXP3 double positive” was decreased (Fig 4) and the definition of Tregs by "CD25-high and FOXP3 positive" showed decreasing of the proportion of Tregs in PB and minimizing the heterogeneity of studies (S2D Fig) [26, 37]. From these results, more accurate definitions could give a uniform distribution of results.

Meanwhile, the proportion of Tregs in the SF of RA patients was increased in comparison to PB (Fig 5B). It seems likely that this is due to increased migration of Tregs to the SF in response to inflammation. The migration of Tregs is well known in carcinoma and infection [5860]. Moreover, intravenous injected Tregs inhibited arthritic symptoms in CIA mouse [10]. Some parts of TCR repertoires of Tregs in SF with juvenile idiopathic arthritis (JIA) patients corresponded those of Tregs in PB with JIA patients [61]. These results indicate that Tregs, which derive from the thymus and circulate in PB, may migrate from PB to inflammatory tissue such as the synovium and then the proportion of Tregs in PB decreases and that of Tregs in SF increases. In addition to migration, FOXP3+ T cells are induced from naïve CD4+ T cells under inflammatory conditions such as the synovial tissue during arthritis [57, 62]. Both migration of PB Tregs to SF and conversion of non-Tregs to Tregs in SF should have enhanced regulatory function of autoimmunity. However, increased number of Tregs in the SF of RA patients was clearly not sufficient to prevent the autoimmunity. The suppressive function of Tregs in SF of RA patients has been demonstrated to be weak [35], partly because the suppressive function of Tregs migrating from PB may be impaired by TNF-α produced in SF [35, 63].

This study has several limitations. First, disease activity in RA patients was not consistent across studies. However, it is postulated that most studies selected RA patients with high disease activity. Second, disease duration from onset of RA was not consistent across studies. A previous study indicated that the proportion of Tregs in early stage of RA might be high and that in late stage of RA might be low [64]. In many of the selected studies, disease duration ranged from 5 to 15 years. Finally, RA treatments were not consistent across studies. These treatments may influence on the proportion of Tregs in RA patients. It is difficult to exclude the influence of RA treatments on our results.

Previously, Miyara et al. demonstrated the sub-population of FOXP3+ (or CD25+) Tregs in PB by utilizing CD45RA and FOXP3 (or CD25). They showed that CD45RA+ FOXP3 (or CD25)-low cells were naïve Tregs (Fraction [Fr.] 1); CD45RA- FOXP3 (CD25)-high cells were effector Tregs (Fr. 2); and CD45RA- FOXP3 (or CD25)-low cells were non-Tregs (Fr. 3) [18]. Accurate definitions of Tregs such as the Miyara Fractions may provide us more reliable information of the status of Tregs in human. In our selected studies, 2 studies utilized the Miyara Fraction model to define Tregs. In these 2 studies, the proportion of effector Tregs (Fr. 2) in RA patients was lower than that in control subjects [31, 37]. Function related definitions of Tregs such as the Miyara Fractions might become important in the future.

In conclusion, the lack of clarity of previous studies as to the perturbation of the status of Tregs in RA was caused by the lack of a consistent identification method for Tregs. Here we demonstrated that differences in the definitions of Tregs could influence on the results of the proportion of Tregs. Evaluation of Tregs by stricter methods such as “CD25hi” and “CD25 and FOXP3 double positive” consistently identify that the proportion of Tregs may be decreased in PB of RA patients. It is beyond the scope of our meta-analysis to identify which definitions of Tregs are correct, rather we can see which definitions provide consistent results. However, since stricter definitions of Tregs that have been validated by gold standard methods such as by analysis of suppressive activity and demethylation of FOXP3 CNS2 [18, 57] consistently identify that the proportion of Tregs in PB are reduced in RA patients, this is a strong indication that this may be the case. As a result, it is likely that if the proportion of Tregs differs in RA patients, accurate and functionally relevant definitions of Tregs are necessary to elucidate their status in RA.

Supporting Information

S1 Fig. Funnel plot for combined meta-analysis of the proportion of Tregs among CD4+ T cells in peripheral blood.

Publication bias was evaluated by funnel plot. Each solid circle represents a study. The y-axis represents standard error that reflects the number of the samples, and the x-axis shows standard mean difference (the proportion of Tregs among CD4+ T cells in PB of RA patients minus that of control subjects) that reflects the effect size. Dotted line indicates random effect model estimate and dashed line indicates fixed effect model estimate. The outer dashed lines indicate the triangular region within which 95% of studies are expected to lie in the absence of both biases and heterogeneity.

https://doi.org/10.1371/journal.pone.0162306.s001

(TIFF)

S2 Fig. Forest plot for sub-group meta-analysis of the proportion of Tregs defined by specified definition.

Standardized mean difference (the proportion of Tregs among CD4+ T cells in PB of RA patients minus that of control subjects) was estimated by meta-analysis. (A) Tregs, which were defined by “CD25 positive”, were analyzed. (B) Tregs, which were defined by “CD25 positive and CD127 negative”, were analyzed. (C) Tregs, which were defined by “FOXP3 positive”, were analyzed. (D) Tregs, which were defined by “CD25-high and FOXP3 positive”, were analyzed.

https://doi.org/10.1371/journal.pone.0162306.s002

(TIFF)

S1 Table. Background of RA patients in each study.

Patient characteristics of RA patients in each study. Studies (c, j, l, n, r) were shown information of Remission RA patients, too. ABT = abatacept; CRP = C-reactive protein; DAS28 = disease activity score 28; DMARDs = disease modified anti-rheumatic-drugs; ESR = erythrocyte sedimentation rate; iTNF = tumor necrosis factor-α inhibitor; MTX = methotrexate, NA = not applicable; CS = Corticosteroids; RA = rheumatoid arthritis; RF = rheumatoid factor, TCZ = tocilizumab.

https://doi.org/10.1371/journal.pone.0162306.s004

(DOCX)

S2 Table. The Newcastle–Ottawa Quality Assessment.

The New Castle-Ottawa Assessment Scale. S1: case definition, S2: representativeness of the cases, S3: community controls, S4: no history of disease in controls, C1a: age-matched controls, C1b: controls for additional factor, E1: ascertainment of exposure (secure record), E1b: ascertainment of exposure (interview where blind to case/control status), E2: same method of ascertainment for case and controls, E3: non-response rate. (p) and (y) were not evaluated by The New Castle-Ottawa Assessment Scale in detail because these manuscripts were written in Chinese.

https://doi.org/10.1371/journal.pone.0162306.s005

(DOCX)

S3 Table. The proportion of Tregs in PB of active or remission RA patients.

Studies (c, j, l, n, w) were shown information of % Tregs in PB among active RA and remission RA patients.

https://doi.org/10.1371/journal.pone.0162306.s006

(DOCX)

S4 Table. The proportion of Tregs in SF and PB of RA patients.

Studies (a, f, k, ac, ae) were shown information of % Tregs in SF among RA patients.

https://doi.org/10.1371/journal.pone.0162306.s007

(DOCX)

Author Contributions

  1. Conceptualization: TM.
  2. Data curation: TM AO.
  3. Formal analysis: TM AO.
  4. Funding acquisition: AO AK.
  5. Investigation: TM.
  6. Methodology: TM.
  7. Project administration: TM AO SS AK.
  8. Resources: TM AO.
  9. Software: TM.
  10. Supervision: YS SS AK.
  11. Validation: TM.
  12. Visualization: TM AO.
  13. Writing – original draft: TM AO.
  14. Writing – review & editing: JBW YS AK.

References

  1. 1. Lee DM, Weinblatt ME. Rheumatoid arthritis. Lancet 2001;358:903–11. pmid:11567728
  2. 2. Manicourt DH, Triki R, Fukuda K, Devogelaer JP, Nagant de Deuxchaisnes C, Thonar EJ. Levels of circulating tumor necrosis factor-α and interleukin-6 in patients with rheumatoid arthritis. Arthritis Rheum 1993;36:490–9. pmid:8457224
  3. 3. van Hamburg JP, Asmawidjaja PS, Davelaar N, Mus AM, Colin EM, Hazes JM, et al. Th17 cells, but not Th1 cells, from patients with early rheumatoid arthritis are potent inducers of matrix metalloproteinases and proinflammatory cytokines upon synovial fibroblast interaction, including autocrine interleukin-17A production. Arthritis Rheum 2011;63:73–83. pmid:20954258
  4. 4. Ito Y, Hashimoto M, Hirota K, Ohkura N, Morikawa H, Nishikawa H, et. al. Detection of T cell responses to a ubiquitous cellular protein in autoimmune disease. Science 2014;346:363–8. pmid:25324392
  5. 5. Sakaguchi S, Yamaguchi T, Nomura T, Ono M. Regulatory T cells and immune tolerance. Cell 2008; 133:775–87. pmid:18510923
  6. 6. Hori S, Nomura T, Sakaguchi S. Control of regulatory T cell development by the transcription factor Foxp3. Science 2003;299:1057–61. pmid:12522256
  7. 7. Wildin R, Smyk-Pearson S, Filipovich A. Clinical and molecular features of the immunodysregulation, polyendocrinopathy, enteropathy, X linked (IPEX) syndrome. J Med Gene 2002;39:527–545. pmid:12161590
  8. 8. Zhu J, Chen XY, Li LB, Yu XT, Zhou Y, Yang WJ, et al. Electroacupuncture attenuates collagen-induced arthritis in rats through vasoactive intestinal peptide signalling-dependent re-establishment of the regulatory T cell/T-helper 17 cell balance. Acupunct Med 2015;33:305–11. pmid:25979865
  9. 9. Morgan ME, Sutmuller RP, Witteveen HJ, van Duivenvoorde LM, Zanelli E, Melief CJ, et al. CD25+ cell depletion hastens the onset of severe disease in collagen-induced arthritis. Arthritis Rheum 2003;48:1452–60. pmid:12746920
  10. 10. Morgan ME, Flierman R, van Duivenvoorde LM, Witteveen HJ, van Ewijk W, van Laar JM, et al. Effective treatment of collagen-induced arthritis by adoptive transfer of CD25+ regulatory T cells. Arthritis Rheum 2005;52:2212–21. pmid:15986351
  11. 11. Paradowska-Gorycka A, Jurkowska M, Felis-Giemza A, Romanowska-Próchnicka K, Manczak M, Maslinski S, et al. Genetic polymorphisms of Foxp3 in patients with rheumatoid arthritis. J Rheumatol. 2015;42:170–80. pmid:25448791
  12. 12. Wing K, Yamaguchi T, Sakaguchi S. Cell-autonomous and -non-autonomous roles of CTLA-4 in immune regulation. Trends Immunol 2011;32:428–433. pmid:21723783
  13. 13. Shi F, Liu J, Li Y. The effect of CTLA-4 A49G polymorphism on rheumatoid arthritis risk:a meta-analysis. Diagn Pathol 2014;9:157. pmid:25128482
  14. 14. Valencia X, Stephens G, Goldbach-mansky R, Wilson M, Shevach EM, Lipsky PE. TNF downmodulates the function of human CD4+ CD25hi T-regulatory cells. BLOOD 2015;108:253–61. pmid:16537805
  15. 15. Kimura A, Kishimoto T. IL-6: regulator of Treg/Th17 balance. Eur J Immunol 2010;40:1830–5. pmid:20583029
  16. 16. Sakaguchi S, Sakaguchi N, Asano M, Itoh M, Toda M. Immunologic self-tolerance maintained by activated T cells expressing IL-2 receptor alpha-chains (CD25). Breakdown of a single mechanism of self-tolerance causes various autoimmune diseases. J Immunol 1995;155:1151–64. pmid:7636184
  17. 17. Lastovicka J. The phenotypic markers of CD4+ CD25+ T regulatory lymphocytes. Res Immunol 2013;2013:119348.
  18. 18. Miyara M, Yoshioka Y, Kitoh A, Shima T, Wing K, Niwa A, et al. Functional delineation and differentiation dynamics of human CD4+ T cells expressing the FoxP3 transcription factor. Immunity 2009;30:899–911. pmid:19464196
  19. 19. OCEBM Levels of Evidence Working Group. The Oxford 2011 levels of evidence. Oxford: Oxford Centre for Evidence-Based Medicine. Available at http://www.cebm.net/ocebm-levels-of-evidence/
  20. 20. Wells GA, Shea B, O’Connell D, Peterson J, Welch V, Losos M, et al. The Newcastle-Ottawa Scale (NOS) for assessing the quality if nonrandomized studies in meta-analyses. Available at http://www.ohri.ca/programs/clinical_epidemiology/oxford.asp.
  21. 21. Moher D, Liberati A, Tetzlaff J, Altman DG; PRISMA Group. Preferred reporting items for systematic reviews and meta-analyses: the PRISMA statement. PLoS Med 2009;6:e1000097. pmid:19621072
  22. 22. DerSimonian R, Kacker R. Random-effects model for meta-analysis of clinical trials: an update. Contemp Clin Trials 2007;28:105–14. pmid:16807131
  23. 23. Higgins JPT, Thompson SG. Quantifying heterogeneity in a meta-analysis. Stat Med 2002;21:1539–58. pmid:12111919
  24. 24. Kanda Y. Investigation of the freely available easy-to-use software “EZR” for medical statistics. Bone Marrow Transplant 2013;48:452–8. pmid:23208313
  25. 25. Barbieri A, Dolcino M, Tinazzi E, Rigo A, Argentino G, Patuzzo G, et al. Characterization of CD30/CD30L+ cells in peripheral blood and synovial fuluid of patients with rheumatoid arthritis. J Immunol Res. 2015;2015:729654. pmid:26090498
  26. 26. Al-Zifzafa DS, El Bakryb SA, Mamdouhc R, Shawarbyc LA, Abdel Ghaffarc AY, Amer HA, et al. FoxP3+ T regulatory cells in Rheumatoid arthritis and the imbalance of the Treg/TH17 cytokine axis. Egypt Rheum, 2015;37:7–15.
  27. 27. Gaafar T, Farid R, Raafat H, Bayoumi F, Gerges B, Rasheed D. The TH17/Treg imbalance in rheumatoid arthritis and relation to disease activity. J Clin Cell Immunol, 2015;6:381.
  28. 28. Daïen CI, Gailhac S, Audo R, Mura T, Hahne M, Combe B, et al. High levels of natural killer cells are associated with response to tocilizumab in patients with sever rheumatolid arthritis. Rheumatology(Oxford) 2015;54:601–8. pmid:25231180
  29. 29. Cribbs AP, Kennedy A, Penn H, Read JE, Amjadi P, Green P, et al. Treg cell function in rheumatoid arthritis is compromised by CTLA-4 promoter methylation resulting in a failure to activate the indoleamine 2,3-dioxygenase pathway. Arthritis Rheum 2014;66:2344–54. pmid:24891289
  30. 30. Moradi B, Schnatzer P, Hagmann S, Rosshirt N, Gotterbarm T, Kretzer JP, et al. CD4+CD25+/highCD127low/- regulatory T cells are enriched in rheumatoid arthritis and osteoarthritis joints—analysis of frequency and phenotype in synovial membrane, synovial fluid and peripheral blood. Arthritis Res Ter 2014;16:R97. pmid:24742142
  31. 31. Matsuki F, Saegusa J, Miyamoto Y, Misaki K, Kumagai S, Morinobu A. CD45RA–Foxp3high activated/effector regulatory T cells in the CCR7+ CD45RACD27+ CD28+ central memory subset are decreased in peripheral blood from patients with rheumatoid arthritis. Biochem Biophys Res Commun 2013;438:778–83. pmid:23747721
  32. 32. Guggino G, Giardina A, Ferrante A, Giardina G, Schinocca C, Sireci G, et al. The in vitro addition of methotrexate and/or methylprednisolone determines peripheral reduction in Th17 and expansion of conventional Treg and of IL-10 producing Th17 lymphocytes in patients with early rheumatoid arthritis. Rheumatol Int. 2015;35:171–5. pmid:24792332
  33. 33. Ji L, Geng Y, Zhou W, Zhang Z. A study on relationship among apoptosis rates, number of peripheral T cell subtypes and disease activity in rheumatoid arthritis. Int J Rheum Dis 2013. pmid:24618356
  34. 34. Gao S, Hao B, Yang XF, Chen WQ. Decreased CD200R expression on monocyte-derived macrophages correlates with Th17/Treg imbalance and disease activity in rheumatoid arthritis patients. Inflamm Res 2014;63:441–50. pmid:24496593
  35. 35. Nie H, Zheng Y, Li R, Guo TB, He D, Fang L, et al. Phosphorylation of FOXP3 controls regulatory T cell function and is inhibited by TNF-α in rheumatoid arthritis. Nat Med 2013;19:322–30. pmid:23396208
  36. 36. Abazaa N, EL-kabarityb RH, Abo-Shadyb RA. Deficient or abundant but unable to fight? Estimation of circulating FoxP3+ T regulatory cells and their counteracting FoxP3- in rheumatoid arthritis and correlation with disease activity. Egypt Rheum 2013;35:185–92.
  37. 37. Kim JR, Chae JN, Kim SH, Ha JS. Subpopulations of regulatory T cells in rheumatoid arthritis, systemic lupus erythematosus, and Behcet’s disease. J Korean Med Sci 2012;27:1009–13. pmid:22969245
  38. 38. Niu Q, Cai B, Huang ZC, Shi YY, Wang LL. Disturbed Th17/Treg balance in patients with rheumatoid arthritis. Rheumatol Int 2012;32:2731–6. pmid:21809006
  39. 39. Chen J, Li J, Gao H, Wang C, Luo J, Lv Z, et al. Comprehensive evaluation of different T-helper cell subsets differentiation and function in rheumatoid arthritis. J Biomed Biotechnol 2012;2012:535361. pmid:23091349
  40. 40. Chen R, Tao Y, Qiu K, Huang W, Huang C, Li J. Association of circulating Treg cells with disease activity in patients with rheumatoid arthritis. J South Med Univ 2012;32:886–9. pmid:22699077
  41. 41. Xiao H, Wang S, Miao R, Kan W. TRAIL is associated with impaired regulation of CD4+CD25 T cells by regulatory T cells in patients with rheumatoid arthritis. J Clin Immunol 2011;31:1112–9. pmid:21732015
  42. 42. Furuzawa-Carballeda J, Lima G, Jakez-Ocampo J, Llorente L. Indoleamine 2,3-dioxygenase-expressing peripheral cells in rheumatoid arthritis and systemic lupus erythematosus: a cross-sectional study. Eur J Clin Invest 2011;41:1037–46. pmid:21366559
  43. 43. Loza MJ, Anderson AS, O'Rourke KS, Wood J, Khan IU. T–cell specific defect in expression of the NTPDase CD39 as a biomarker for lupus. Cell Immunol 2011;271:110–7. pmid:21763644
  44. 44. Tang TT, Song Y, Ding YJ, Liao YH, Yu X, Du R et al. Atorvastatin upregulates regulatory T cells and reduces clinical disease activity in patients with rheumatoid arthritis. J Lipid Res 2011;52:1023–32. pmid:21368320
  45. 45. Chen MH, Chen WS, Lee H, Tsai CY, Chou CT. Inverse correlation of programmed death 1 (PD-1) expression in T cells to the spinal radiologic changes in Taiwanese patients with ankylosing spondylitis. Clin Rheumatol 2011;20:1181–7. pmid:21547439
  46. 46. Ursaciuc C, Surcel M, Ciotaru D, Dobre M, Pirvu IR, Munteanu AN et al. Regulatory T cells and Th1/Th2 cytokines as immunodiagnosis keys in systemic autoimmune diseases. Roum Arch Microbiol Immunol 2010;69:79–84. pmid:21235134
  47. 47. Al-Shukaili A, AlKaabi J, Al-Gafri S, Hassan B, Al-Muneeri A. Quantification of CD4+ CD25+ regulatory T cells in peripheral blood of patients with systemic lupus erythematosus and rheumatoid arthritis. Open Autoimmun J 2009;1:5–9.
  48. 48. Sempere-Ortells JM, Perez-Garcia V, Marin-Alberca G, Peris-Pertusa A, Benito JM, Marco FM, et al. Quantification and phenotype of regulatory T cells in rheumatoid arthritis according to disease activity score-28. Autoimmunity 2009;42:636–45. pmid:19886735
  49. 49. Hung ZX, Cheng BR, Li TW, Lu YM, Gu JR, Han XY, et al. Study on expressions of CD25 and CD127 in CD4+ peripheral T lymphocytes of patients with rheumatoid arthritis. Nat Med J China 2009;89;2476–80. pmid:20137434
  50. 50. Han GM, O'Neil-Andersen NJ, Zurier RB, Lawrence DA. CD4+CD25high T cell numbers are enriched in the peripheral blood of patients with rheumatoid arthritis. Cell Immunol 2008;253:92–101. pmid:18649874
  51. 51. Yoon BY, Cho ML, Hong YS, Jhun JY, Park MK, Park KS, et al. Anti-CD3 Antibody induces IL-10-producing CD4+CD25+ regulatory T Cells, which suppress T cell response in rheumatoid arthritis patients. Immune Netw 2007;7:124–32.
  52. 52. Kao JK, Hsue YT, Lin CY. Role of new population of peripheral CD11c+CD8+ T cells and CD4+CD25+ regulatory T cells during acute and remission stages in rheumatoid arthritis patients. J Microbiol Immunol Infect 2007;40:419–27. pmid:17932602
  53. 53. Jiao Z, Wang W, Jia R, Li J, You H, Chen L, et al. Accumulation of FoxP3-expressing CD4+CD25+ T cells with distinct chemokine receptors in synovial fluid of patients with active rheumatoid arthritis. Scand J Rheumatol 2007;36:428–33. pmid:18092263
  54. 54. Minami R, Sakai K, Miyamura T, Yamamoto M, Suematsu E. The role of CD4+ CD25+ rgulatory T cells in patients with rheumatoid arthritis. Nihon Rinsho Meneki Gakkai Kaishi. 2006;29:37–42. [Published in Japanese] pmid:16505601
  55. 55. Möttönen M, Heikkinen J, Mustonen L, Isomäki P, Luukkainen R Lassila O. CD4+CD25+T cells with the phenotypic and functional characteristics of regulatory T cells are enriched in the synovial fluid of patients with rheumatoid arthritis. Clin Exp Immunol 2005;140:360–7. pmid:15807863
  56. 56. Seddiki N, Santner-Nanan B, Martinson J, Zaunders J, Sasson S, Landay A et al. Expression of interleukin (IL)-2 and IL-7 receptors discriminates between human regulatory and activated T cells. J Exp Med 2006;203:1693–700. pmid:16818676
  57. 57. Sakaguchi S, Miyara M, Costantino CM, Hafler DA. FOXP3+ regulatory T cells in the human immune system. Nat Rev Immunol 2010;10:490–500. pmid:20559327
  58. 58. Curiel TJ, Coukos G, Zou L, Alvarez X Cheng P, Mottram P, et al. Specific recruitment of regulatory T cells in ovarian carcinoma fosters immune privilege and predicts reduced survival. Nat. Med. 2004;10:942–949. pmid:15322536
  59. 59. Yurchenko E, Tritt M, Hay V, Shevach EM, Belkaid Y, Piccirillo CA. CCR5-dependent homing of naturally occurring CD4 regulatory T cells to sites of Leishmania major infection favors pathogen persistence. J. Exp. Med. 2006;203:2451–2460. pmid:17015634
  60. 60. Green AM, Mattila JT, Bigbee CL, Bongers KS, Lin PL, Flynn JL. CD4(+) regulatory T cells in a cynomolgus macaque model of Mycobacterium tuberculosis infection. J Infect Dis. 2010;202:533–41. pmid:20617900
  61. 61. David B, Eirini G, Hannah L, Lucy RW. Synovial regulatory T cells occupy adscrete TCR niche in human arthritis and require local signals to stabilize FOXP3 protein expression. J. Immunol, 2015;195:5616–24. pmid:26561546
  62. 62. Yadav M, Stephan S, Bluestone JA. Peripherally induced Tregs—role in immune homeostasis and autoimmunity. Front Immunol, 2013;4:232. pmid:23966994
  63. 63. Xavier V, Geoffrey S, Raphaela GM, Mildred W, Ethan MS, Peter EL. TNF downmodulates the function of human CD4+ CD25hi T-regulatory cells. Blood, 2007;108:253–261. pmid:16537805
  64. 64. Cao D, van Vollenhoven R, Klareskog L, Trollmo C, Malmstrom V. CD25 bright CD4+ regulatory T cells are enriched in inflamed joints of patients with chronic rheumatic disease. Arthritis Res Ther, 2004;6:R335–46. pmid:15225369