Changes of Regulatory T Cells and of Proinflammatory and Immunosuppressive Cytokines in Patients with Type 2 Diabetes Mellitus: A Systematic Review and Meta-Analysis

Objective. The aim of this study was to investigate the changes of regulatory T cells (Treg), interleukin-6 (IL-6), IL-10, transforming growth factor-β (TGF-β), and tumor necrosis factor-alpha (TNF-α) in patients with type 2 diabetes mellitus (T2DM). Methods. We performed a comprehensive search up to July 2016 for all clinical studies about the changes of Treg, IL-6, IL-10, IL-17, TGF-β, and TNF-α in T2DM patients versus healthy controls. Results. A total of 91 articles (5642 cases and 7378 controls) were included for this meta-analysis. Compared with the controls (all p < 0.001), the patients had increased serum levels of IL-6, TGF-β, and TNF-α but decreased the percentage of peripheral CD4+CD25+Foxp3+Treg and serum IL-10 level. Furthermore, the percentage of peripheral CD4+CD25+Foxp3+Treg (p < 0.001) and serum IL-10 level (p = 0.033) were significantly lower in the patients with complication and in the patients without complication, respectively. No significant changes about the percentage of CD4+CD25+Treg (p = 0.360) and serum IL-17 level (p = 0.459) were found in T2DM patients. Conclusions. T2DM patients have decreased the percentage of peripheral CD4+CD25+Foxp3+Treg and levels of serum IL-10 but elevated serum levels of IL-6, TGF-β, and TNF-α. Presence of diabetic complications further lowers the peripheral CD4+CD25+Foxp3+Treg number.


Introduction
Type 2 diabetes mellitus (T2DM) is one of the most common noncommunicable diseases characterized by insulin resistance and impaired insulin secretion [1,2]. Metabolic proinflammatory disorder including chronic hyperglycemia and increased levels of circulating cytokines suggests immunological disturbances [3][4][5][6][7], which seriously affects the quality of life of the patients and imposes a large economic burden on the national health care system [8]. Genetic and environmental factors are blamed for T2DM and up to 25% of first-degree relatives of T2DM patients may develop this disease [9]. The origin and development of T2DM were involved in multiple risk factors [10]. Regulatory T cells (Treg) and cytokines play important roles in the development of T2DM.
Treg is a subset of CD4 + T cells that maintain peripheral tolerance and suppress antigen specific immune responses by secreting transforming growth factor-(TGF-), interleukin-10 (IL-10), and IL-4 to inhibit autoimmunity [11]. It was found that the ratios of CD4 + CD25 hi Treg/Th17 cells and CD4 + CD25 hi Treg/Th1 cells were significantly decreased in T2DM patients [12]. Expression of Foxp3, a key player for the development and function of Treg, correlates well with regulatory activity and number of Treg. Indeed, Foxp3 is exclusively expressed in CD4 + CD25 + Treg [13][14][15][16]. A positive correlation between CD4 + CD25 + Foxp3 + Treg and the enhanced expression of IL-6 on CD4 + T cells was observed in T2DM patients [17]. IL-10, as a multifunctional cytokine and secretion of Treg, plays a key role in the inflammatory response that is associated with insulin resistant states and T2DM [18]. Increased levels of IL-17 were found to protect against autoimmune mediated T1DM in nonobese diabetic mice [19]. On the other hand, loss of IL-17 has been associated with disease susceptibility in part because it has been suggested that the absence of IL-17 results in enhanced production of other proinflammatory cytokines [20]. TGF-is also a multifunctional cytokine circulating as a biologically inactive form in human plasma [21,22]. The TGF-family includes multifunctional molecules that exert specific effects on cell proliferation, differentiation, migration, development, tissue remodeling, and repair [23]. TNF-inhibits the insulin signaling cascade through regulating several pivotal regulatory proteins, such as the insulin receptor substrate (IRS) and Akt substrate 160 in human skeletal muscle in vitro [24] and in vivo [25]. It has reported that polymorphism of immune genes such as TNF- [26] and TGF- [27] was associated with the development of T2DM. Intriguingly, increased renal production of TGFwas a distinct feature of diabetes [28][29][30][31].
Within the past few years, many clinical studies have been focusing on the association of Treg with proinflammatory and immunosuppressive cytokines in T2DM. Despite intensive research efforts, results of these studies have been inconsistent. Therefore, we performed this meta-analysis synthesizing the data from case-control studies to evaluate changes of Treg, IL-6, IL-10, IL-17, TGF-, and TNF-in T2DM patients.

Inclusion Criteria.
We reviewed all relevant articles using the following inclusion criteria: (1) the study should evaluate the relationship of CD4 + CD25 + Foxp3 + Treg, CD4 + CD25 + Treg, IL-6, IL-10, IL-17, TGF-, or TNF-with T2DM patients; (2) the design had to be a case-control study; (3) original data were displayed or could be converted to as mean ± SD; and (4) original report showed no duplicated data.

Quality Assessment and Data
Extraction. The data were extracted independently by two reviewers (Yong-chao Qiao and Jian Shen) by using predefined data extraction forms and the quality of all eligible studies was evaluated according to the Newcastle-Ottawa Scale (NOS) [33]. The following information was extracted: (1) name of the first author; (2) date of publication; (3) country of the study; (4) study design; (5) sample size of patients and controls; (6) mean age of the sample; and (7) mean ± SD of patients and controls. In case of disagreement, a third investigator (Hai-lu Zhao) was invited to assess such articles and the disagreements were resolved through discussion.

Statistical Analysis.
We presented the data (sample size, mean ± SD) to illustrate the changes of Treg, IL-6, IL-10, IL-17, TGF-, and TNF-in T2DM patients versus healthy controls, and Chi-squared test and 2 statistics were used to assess heterogeneity. When < 0.1 or 2 > 50%, the heterogeneity was considered significant and a random effect model was used; otherwise, a fixed-effect model was used. Considering the influence of diabetic complications, patients were divided into two groups (T2DM with complication and T2DM without complication) for subgroup analysis. Regression analysis is also an important method for exploring sources of heterogeneity. We performed sensitivity analysis by limiting the studies of NOS score ≥ 7 or excluding studies with a high risk of bias. Publication bias was examined graphically by constructing Egger's test and < 0.05 was considered to be representative of statistically significant publication bias. Stata 12.0 software was performed in this meta-analysis.

Author
Year        Journal of Diabetes Research    = 0.586; 95% CI, −4.84 to 2.86). Therefore, diabetic complication was a key influencing factor for the high heterogeneity in the metaanalysis of TGF-but not the others.

Sensitivity Analysis.
Sensitivity analysis was used to assess the stability of the results by excluding studies with high risk of bias and no significant changes in the results were found. We further conducted sensitivity analysis by including studies with high NOS score (≥7) and found that all the results remained consistent.

Publication Bias.
Egger's test showed significant publication bias in the meta-analysis of TNF-but not the others (Figure 9).

Discussion
In this study, we found that the patients with T2DM had increased serum levels of IL-6, TGF-, and TNF-but decreased percentage of peripheral CD4 + CD25 + Foxp3 + Treg and serum IL-10 level. Furthermore, the percentage of peripheral CD4 + CD25 + Foxp3 + Treg and serum IL-10 level were influenced by diabetic complication. The expression of inflammatory and proinflammatory cytokines from peripheral blood T lymphocyte plays an important role in the development of diabetes and diabetic complications [17]. Many studies have proved the maintenance of immunological self-tolerance by CD4 + CD25 + Treg and CD4 + CD25 + Foxp3 + Treg [115]. Treg could suppress inflammatory response through contact inhibition [116]. In this study, the finding of decreased percentage of peripheral CD4 + CD25 + Foxp3 + Treg in T2DM patients indicates that Foxp3 might be a key player for the development and function of Treg. CD4 + CD25 + Foxp3 + Treg differs from CD4 + CD25 + Treg. CD4 + CD25 + Treg might not sufficiently represent the negative regulatory Treg. Some researchers also considered that the differentiation and function maintenance of Treg were dependent on the expression of the Foxp3, and, consequently, Foxp3 is considered as the key transcriptional factor in Treg cells [117][118][119].
IL-10 and TGF-secreted by Treg [116,120] are the biomarkers in T2DM patients [2,116]. Previous studies suggested that IL-10 could suppress the proliferation of T leukomonocyte and the secretion of cytokines [121], whereas TGF-may sustain the expression of Foxp3 in CD4 + CD25 + Treg to enhance immunosuppressive function [122,123]. Consistent with our findings, several studies have shown a significantly decreased level of serum IL-10 in T2DM patients [88,124]. Correlation of T2DM with Treg cells and TGF-is generally negative [116].
IL-6 can be released from macrophages and adipocytes in adipose tissue [125][126][127]. Adipose tissue also produces TNF-to stimulate IL-6 gene expression [128]. A recent investigation has showed that IL-6 could enhance Treg in mice [129]. In the present meta-analysis of T2DM patients,  increased levels of serum IL-6, TGF-, and TNF-coexisted with decreased levels of IL-10 and decreased percentage of CD4 + CD25 + Foxp3 + Treg. This finding highlights that the cytokines and growth factors may originate from multiple sources such as macrophages, T cells, and other tissue cells rather than Treg alone. Furthermore, chronic persistent activation of innate immunity and IL-6 secretion occurring in T2DM might inhibit the development of inducible Treg cells.

Journal of Diabetes Research
Th17 cells could produce IL-17, TNF-, and IL-6 and induce inflammation in the pathogenesis of autoimmune diseases [130]. Th17 cells are a major T cell subset implicated in the pathogenesis of multiple sclerosis, rheumatoid arthritis, and psoriasis [131]. A previous study has revealed that not only Th1/Th2 imbalance but also Th17/Treg imbalance can contribute to the pathogenesis of autoimmune diseases such as T1DM as well as proinflammatory disorders and such as T2DM [2]. T2DM patients have elevated serum levels of IL-6, IL-1 , and TGF-, the cytokines known to induce Th17 differentiation [131]. Enhanced production of IL-6 and TNF-and decreased levels of serum IL-10 that occurred in T2DM patients may suppress Treg cells and ratios of Treg to Th17 and Th1 cells [132,133]. The immunocompromised effects on macrophages and lymphocytes likely drive an inflammatory state to contribute to the occurrence of diabetic complications [12]. Here, in this study, no significant changes of Foxp3 + Treg cells and serum IL-17 levels were found in T2DM subjects without complication. In contrast, decreased Foxp3 + Treg cells were evident in T2DM subjects with complication. These findings indicate a close correlation of CD4 + CD25 + Foxp3 + Treg and diabetic complication in T2DM.
There is an intimate relationship of the differentiation of Th17 cells with the relative abundance of peripheral CD4 + CD25 + Foxp3 + Treg cells and the serum levels of IL-6, IL-10, and TGF-. Although changes of serum levels of IL-17 were not significant in this meta-analysis of T2DM patients versus controls, IL-17 may be a clue to the possible involvement of Th17 cells in T2DM pathogenesis. Firstly, a decrease of Treg cells might be accompanied by an increase of Th17 cells. The study by Guan et al. has indicated the existence of a developmental switch between Th1/Th17 cells, on one hand, and Th2/Treg cells, on the other hand [134]. Secondly, in the presence of high serum levels of IL-6 and  TGF-, as we reported here, differentiation of Th17 cells might be favoured. Lastly, Th17 cells might be, together with innate cells, a primary source of the increased IL-6 levels and might be actively orchestrating the immunitydriven, chronic inflammation of target tissues and organs in T2DM. In this systematic review, the studies examining the number of Th17 cells in T2DM were too scarce for being included in the meta-analysis. Future studies are required to focus on the role of Th17/Treg and products of the Th17 cells in the pathogenesis of T2DM and associated complications.
Diabetic complications such as retinopathy, nephropathy, and cardiovascular disease affect immune cells and cytokines in type 2 diabetes [135,136]. Actually, urinary TGFlevels are elevated in the presence of microalbuminuria and overt proteinuria [137]. Additionally, elevated plasma TGF-may reflect the state of hyperglycemia in T2DM patients [48]. Systemic inflammation in T2DM is linked to the development of diabetic complications [138,139]. Yet, the mechanism of immune alteration in T2DM and diabetic complication remains unclear. In this meta-analysis, diabetic complication indeed has an impact on the percentage of peripheral CD4 + CD25 + Foxp3 + Treg and level of serum IL-10. The percentage of Treg cells and levels of cytokines in T2DM may also depend on ethnicity, sex, weight, age, and disease duration. Publication bias might influence the interpretation of our final results. The results of Egger's tests explain that no publication bias existed in all comparisons except for TNF-. The publication bias in this meta-analysis might be attributed    to studies of small samples and positive results published more easily than negative reports.
There are some limitations in this meta-analysis when interpreting the findings. Firstly, we have selected random effect model to synthesize SMD because of the high heterogeneity existing in some comparisons, but this selection may affect the accuracy of outcome. Secondly, we could not conduct further subgroup analysis of gender, weight, and disease duration because most of the included studies lack sufficient original data. Thirdly, articles published in Chinese or English are included, while unpublished data and papers published in other languages are unknown.

Conclusions
In summary, T2DM patients and the patients with diabetic complication have decreased immunosuppressive CD4 + CD25 + Foxp3 + Treg cells and increased proinflammato-