Meta-Analysis of TNF 308 G/A Polymorphism and Type 2 Diabetes Mellitus

Background and Objectives Many investigations have focused the association between TNF 308 G/A polymorphism and risk for type 2 diabetes mellitus (T2DM). However, the sample sizes of most of the studies were small. The aim of this study is to evaluate the precise association between this variant and risk for T2DM in a large-scale meta-analysis. Methods All publications were searched on the association between TNF 308 G/A polymorphism and T2DM. The key words were as follows: diabetes, tumor necrosis factor and polymorphism/variant/genotype. This meta-analysis was assessed by Review manager 5.0. Results There were 18 studies identified. The odd ritos (ORs) and 95% confidence intervals (CI) for GA+AA versus GG genotype of TNF 308 G/A polymorphism were 1.03 (0.95–1.12), 1.03 (0.94–1.13) and 1.03 (0.78–1.36) in overall, Caucasian and Asian populations, respectively. The sensitivity analysis further strengthened the validity of this association. No publication bias or heterogeneity was observed in this study. Conclusion In summary, there was no significant association detected between the TNF 308 G/A polymorphism and risk for T2DM.


Introduction
The prevalence of diabetes is high, with high rates of morbidity and mortality. World Health Organization (WHO) estimations predicted that by the year 2030, 350 million individuals worldwide would suffer diabetes [1]. Recent decades, chronic inflammation has been found in obesity and diabetes [2]. Tumor necrosis factor alpha (TNF-a) was found to be a crucial component of the pro-inflammatory cytokines, mainly produced by macrophages and adipocytes [3], and can induce insulin resistance in T2DM [4].
As genetic variations in the promoter region may regulate TNFa production, single nucleotide polymorphisms in the promoter region of the TNF gene, such as 2238, 2308, 2857, 21031 in human [5], have been investigated for their roles on gene transcription as well as for their possible relationships with inflammatory related diseases. There were studies focused on TNF 308 G/A polymorphism found guanine (G) replaced by adenine (A) in TNF 308 position [6], which led to a higher rate of TNF gene transcription than that of wild-type in vitro expression studies [7,8]. For diabetes, our previous study found that TNF 308 G/A polymorphism as a risk factor for type 1 diabetes mellitus (T1DM) [9]. However, the association between this variation and risk for T2DM still remained unclear. Since studies were limited to modest sample size and different ethnicity, every single study may be underpowered to achieve a comprehensive and reliable conclusion, and a meta-analysis was required, which has been recently suggested by Boraska et al. [10].
In this study, we have therefore conducted a meta-analysis from all eligible studies to confirm whether TNF 308 G/A polymorphism would be associated with the risk for T2DM.

Study Selection
We searched several databases (Medline, PubMed and EMBase) through November, 2010 for all publications on the association between TNF 308 G/A polymorphism and T2DM. The key words were as follows: diabetes, tumor necrosis factor and polymorphism or variant or genotype. In addition, we also searched references of retrieved articles. Studies should meet the following criteria: (1) case-control study; (2) only diabetes as outcome, and (3) at least two comparison groups (diabetes vs. control groups) involved in a single study. Exclusive criteria: no report about the genotype frequency, or insufficient information for data extraction. The MOOSE Checklist and the flow chart for the studies were shown as Checklist S1 and Figure S1 in the supporting information.

Data Extraction
Two authors extracted data independently and in duplicate, and reached on all items, including: author's last name, journal and year of publication, country of origin, selection and characteristics of diabetes cases and controls, ethnicity of the study population, genotypes (rs1800629) and numbers of cases and controls. The results were compared and disagreements were discussed and resolved with consensus.

Statistical analysis
The meta-analysis was performed by using Review manager 5.0. We pooled the odds Ratios (ORs) for TNF 308 GA+AA versus GG genotype and further conducted subgroup analyses by ethnicity. Heterogeneity among studies was examined with I 2 statistics that was interpreted as the proportion of total variation contributed by between-study variation. If there was no statistical heterogeneity among studies (I 2 ,50% and P.0.05), the ORs and 95% CI would be estimated for each study in a fixed-effects model. Otherwise, a random-effect model should be employed. In sensitivity analysis, relative influence of each study on the pooled estimate was assessed by omitting one study at a time. Funnel plots were used to evaluate publication bias. All P-values were two-tailed.

Association between TNF 308 G/A polymorphism and the risk for T2DM
It has been shown in Figure 1 that the risk for T2DM conferred by TNF 308 G/A polymorphism in the overall 18 studies and subgroups did not reach the significant difference. The ORs of GA+AA versus GG for T2DM were 1.03 (0.95-1.12) for the overall, 1.03 (0.94-1.13) for Caucasian and 1.03 (0.78-1.36) for Asian, respectively. There was no evidence of heterogeneity among the overall 18 studies or subgroups (I 2 = 0 in overall populations and subgroups).
To further strengthen the confidence for the results, we conducted a sensitivity analysis. This analysis confirmed the stability of the null association between TNF 308 G/A polymorphism and T2DM ( Table 2). Exclusion of individual studies did not modify the estimates much, with pooled ORs ranging from 1.02 to 1.04.
The shape of the funnel plots was symmetrical, suggesting there was no evidence of publication bias among the studies ( Figure 2). Together with the above results, this meta-analysis showed no significant association between TNF 308 G/A polymorphism and the risk for T2DM.

Discussion
Inflammation has been widely known as an important feature of T2DM, with high levels of pro-inflammation cytokines, including IL-1, IL-6 and TNF-a. As TNF-a can impair insulin signal pathways and lead to B-cell destruction, elevated TNF-a is considered playing a central role in the development of T2DM. The TNF gene is located within the HLA III region in chromosome 6p21, involved in inflammatory responses [28]. Additionally, TNF 308 G/A single nucleotide polymorphism (SNP) in the promoter region of TNF was found to increase the expression of this proinflammatory cytokine in culture cells and positive associated with risk for T1DM [9,29]. So far, many studies had focused on association between TNF 308 G/A polymorphism and T2DM, but the results were still unclear.
In this large-scale meta-analysis, the combined evidence suggested that TNF 308 G/A polymorphism did not contribute to the development of T2DM. However, T2DM is a complex disease, and both environmental and genetic factors are involved in the development of T2DM. There are some points should be concerned for the inconsistent results in early reports. Firstly, ethnic differences may attribute to these different results, since the distributions of the TNF 308 G/A polymorphism were different between various ethnic populations. For instance, the frequencies of TNF 308 G/A polymorphism allele differs from 9% in Chinese population [30], 16% in French and Scandinavian populations [31,32], 18% in German [33], to 24% in Australians [34]. On the other hand, study design or small sample size or some environmental factors may affect the results. Most of these studies did not consider most of the important environmental factors. It is possible that variation at this locus has modest effects on T2DM, but environmental factors may predominate in the progress of T2DM, and mask the effects of this variation. Specific environmental factors like lifestyle and obesity that have been already well studied in recent decades [35]. Higher free fatty acids are considered as a major risk for T2DM, which can stimulate TNF-a secretion. It is still unknown whether the lifestyle characteristics of different populations influence the association between genotype and T2DM and whether genetic factors influence the ages of onset of T2DM. The unconsidered factors mixed together may cover the role of TNF 308 G/A polymorphism. Thus, Figure 1. Meta-analysis of the association between TNF 308 G/A polymorphism and the risk for type 2 diabetes mellitus. doi:10.1371/journal.pone.0018480.g001 even if the variation has a causal effect on T2DM, it may take a long time to be observed. Supporting this deduction, Ishii et al. reported that FPG in older men was higher and a trend for triglycerides to be higher and HDL cholesterol to be lower in the group with TNF 308 G/A polymorphism, but not in the young [14]. Insulin resistance may aggravate gradually in those Japanese subjects together with metabolic changes, such as hyperglycemia and dyslipidemia, on the presence of the TNF 308 G/A polymorphism. The effects of the TNF 308 G/A polymorphism are so small that in the short period the difference can not be observed. In addition, interaction of the TNF gene with other pro-and anti-inflammatory cytokine genes plays an integrated role in destruction of pancreatic beta cells [36]. And cytokines in the circulation interact with each other in the pathogenesis of diabetes. Therefore, it is not surprising that no influence of TNF 308 G/A polymorphism was found in susceptibility to T2DM.
However, this variant was reported to be related to some metabolic disorders and T1DM. Furta et al. found that TNF-a levels were not significantly different between Japanese T2DM patients with and without this variation [11] and TNF-a levels were affected by visceral fat area, consisting with Morris et al. [20]. Moreover, a meta-analysis on relation between TNF 308 G/A polymorphism and metabolic syndrome had indicated that individuals carrying TNFA allele had significantly higher fasting insulin level, systolic arterial blood pressure, higher risk of developing obesity and maybe HOMA-IR, but no significant association with BMI. WHR, fasting glucose and plasma leptin levels, which suggested TNFA allele would increase the risk of metabolic syndrome [37]. Also, It was reported that TNF 308 G/A polymorphism was more common in T2DM patients with than without macrovascular disease [18]. And our previous study confirmed that a higher frequency of TNF 308 A allele conferred a significant risk for T1DM [9]. All the available findings indicated that TNF 308 G/A polymorphism might not increase the development of T2DM, but it can enhance the risk for human metabolic disorder and some autoimmunity diseases.
In conclusion, we did not find any evidence of association between TNF 308 G/A polymorphism and T2DM in this large-scale meta-  analysis. Further prospective research, with larger numbers of participants and fully confounding risk factors considered, such as age, sex, ethnicity and life style, is warrant to examine the possible effects of this variation on T2DM to confirm our conclusion.

Supporting Information
Checklist S1 (DOC) Figure S1 The flow chart of the included studies. (DOC)

Author Contributions
Conceived and designed the experiments: C-HS YL. Performed the experiments: R-NF CZ. Analyzed the data: R-NF CZ. Contributed reagents/materials/analysis tools: R-NF CZ. Wrote the paper: R-NF.