Associations between cytotoxic T-lymphocyte-associated antigen 4 gene polymorphisms and diabetes mellitus: a meta-analysis of 76 case–control studies

Background: Several genetic association studies already investigated potential roles of cytotoxic T-lymphocyte-associated antigen 4 (CTLA-4) gene polymorphisms in diabetes mellitus (DM), with inconsistent results. Therefore, we performed this meta-analysis to better assess the relationship between CTLA-4 gene polymorphisms and DM in a larger pooled population. Methods: PubMed, Embase, Web of Science, and CNKI were systematically searched for eligible studies. Pooled odds ratios (ORs) with 95% confidence intervals (CIs) were calculated to estimate the strength of associations between CTLA-4 gene polymorphisms and DM in all possible genetic models. Results: A total of 76 studies were finally included in our analyses. Significant associations with susceptibility to type 1 diabetes mellitus (T1DM) were detected for rs231775 (dominant model: P=0.008, OR = 0.83, 95%CI 0.73–0.95; recessive model: P=0.003, OR = 1.27, 95%CI 1.09–1.50; allele model: P=0.004, OR = 0.85, 95%CI 0.77–0.95) and rs5742909 (recessive model: P=0.02, OR = 1.50, 95%CI 1.05–2.13) polymorphisms in overall population. Further subgroup analyses revealed that rs231775 polymorphism was significantly associated with susceptibility to T1DM in Caucasians and South Asians, and rs5742909 polymorphism was significantly associated with susceptibility to T1DM in South Asians. Moreover, rs231775 polymorphism was also found to be significantly associated with susceptibility to type 2 diabetes mellitus (T2DM) in East Asians and South Asians. Conclusions: Our findings indicated that rs231775 and rs5742909 polymorphisms may serve as genetic biomarkers of T1DM, and rs231775 polymorphism may also serve as a genetic biomarker of T2DM.


Introduction
Diabetes mellitus (DM), characterized by chronic hyperglycemia caused by deficiency in insulin secretion or resistance against insulin, is the most prevalent metabolic disorder worldwide, and it currently affects over 350 million people globally [1,2]. So far, the exact underlying pathogenic mechanism of DM is still not fully understood. Nevertheless, the fact that over 100 genetic loci were already found to be correlated with an increased susceptibility to DM by past genome-wide association studies suggested that genetic factors were crucial for the occurrence and development of DM [3,4].
Cytotoxic T-lymphocyte-associated antigen 4 (CTLA-4) is mainly expressed on activated T cells, and it serves a negative regulator of T cell activation and proliferation [5]. Previous studies showed that CTLA-4 could induce T cell tolerance and attenuate T cell mediated immune responses by binding with co-stimulating molecules, B7-1 (CD80) and B7-2 (CD86) [6], and dysfunction of CTLA-4 was demonstrated to be implicated in various autoimmune diseases including type 1 diabetes mellitus (T1DM) [7,8]. Consequently, CTLA-4 gene polymorphisms were intensively studied with regard to their associations with T1DM [9][10][11][12]. Recently, some pilot studies also analyzed potential associations between CTLA-4 gene polymorphisms and the much more prevalent type 2 diabetes mellitus (T2DM) [13,14]. Nevertheless, whether CTLA-4 gene polymorphisms were associated with T1DM and T2DM or not remain controversial, especially when they were conducted in different populations. Therefore, we performed the present meta-analysis to pool the data of all relevant studies, and obtain more conclusive results on associations of CTLA-4 gene polymorphisms with T1DM and T2DM.

Literature search and inclusion criteria
The current meta-analysis was complied with the Preferred Reporting Items for Systematic Reviews and Meta-analyses (PRISMA) statement [15]. Potentially relevant articles were searched in PubMed, Medline, Web of Science, and CNKI using the following key words: 'Cytotoxic T-lymphocyte antigen 4' , 'CTLA-4' , 'polymorphism' , 'variant' , 'mutation' , 'genotype' , 'allele' , 'diabetes mellitus' , 'diabetes' , and 'DM' . The initial literature search was conducted in October 2018 and the latest update was performed in January 2019. We also screened the reference lists of all retrieved articles to identify other potentially relevant studies.
Included studies should met all the following criteria: (1) case-control study on associations between CTLA-4 gene polymorphisms and individual susceptibility to DM; (2) provide adequate data to calculate odds ratios (ORs) and 95% confidence intervals (CIs); (3) full text in English or Chinese available. For duplicate reports, only the most complete one was included. Family-based association studies, case reports, case series, reviews, comments, letters, and conference presentations were excluded.

Data extraction and quality assessment
The following data were extracted from included studies: (1) name of first author; (2) year of publication; (3) country and ethnicity of participants; (4) type of disease; (5) the number of cases and controls; and (6) genotypic distributions of CTLA-4 gene polymorphisms in cases and controls. The probability value (p value) of Hardy-Weinberg equilibrium (HWE) test was also calculated.
The Newcastle-Ottawa scale (NOS) was used to assess the quality of eligible studies from three aspects: (1) selection of cases and controls; (2) comparability between cases and controls; and (3) exposure in cases and controls [16]. The NOS has a score range of 0-9, and studies with a score of more than 7 were assumed to be of high quality.
Two reviewers conducted data extraction and quality assessment independently. When necessary, the reviewers wrote to the corresponding authors for extra information. Any disagreement between two reviewers was solved by discussion until a consensus was reached.

Statistical analyses
All statistical analyses in the present study were conducted with Review Manager Version 5.3.3 (The Cochrane Collaboration, Software Update, Oxford, United Kingdom). ORs and 95% CIs were used to assess potential associations of CTLA-4 gene polymorphisms with the susceptibility to DM in dominant, recessive, over-dominant, and allele models, and a P value of 0.05 or less was considered to be statistically significant. Between-study heterogeneity was evaluated by I 2 statistic. If I 2 was greater than 50%, random-effect models (REMs) would be used for analyses due to the existence of significant heterogeneities. Otherwise, fixed-effect models (FEMs) would be employed for analyses. Subgroup analyses by ethnicity of participants were subsequently performed. Sensitivity analyses were carried out to test the stability of the results. Funnel plots were applied to evaluate possible publication biases.

Characteristics of included studies
Our systematic literature search yielded 842 results. After exclusion of irrelevant and duplicate articles by reading titles and abstracts, 135 potentially relevant articles were retrieved for further evaluation. Another 59 articles were subsequently excluded after reading the full text. Finally, a total of 76 studies that met the inclusion criteria of our meta-analysis were included (see Figure 1). Characteristics of included studies are shown in Table 1.    Further subgroup analyses revealed that rs231775 polymorphism was significantly associated with susceptibility to T1DM in Caucasians (dominant, recessive, and allele models) and South Asians (dominant, recessive, over-dominant, and allele models), but not in East Asians. Moreover, rs231775 polymorphism was also significantly associated with susceptibility to T2DM in East Asians (over-dominant model) and South Asians (recessive and allele models), but not in Caucasians. Additionally, we also found that rs5742909 polymorphism was significantly associated with susceptibility to T1DM in South Asians (dominant, recessive, over-dominant, and allele models), but not in East Asians and Caucasians (see Table 2).

Sensitivity analyses
Sensitivity analyses were carried out to test the stability of meta-analysis results by eliminating studies that deviated from HWE. No changes of results were detected for investigated CTLA-4 gene polymorphisms in any comparisons, which indicated that our findings were quite statistically reliable.

Publication biases
Potential publication biases in the present study were evaluated with funnel plots. No obvious asymmetry of funnel plots was observed in any comparisons, which suggested that our findings were unlikely to be impacted by severe publication biases.

Discussion
Despite enormous advancements in pharmacotherapy over the past few decades, DM and its associated vascular complications are still leading causes of death and disability all over the world [17,18]. To date, the exact cause of DM is still largely unclear in spite of extensive investigations. However, the obvious familial aggregation tendency of DM indicated that genetic factors may significantly contribute to its occurrence and development [19]. Thus, identify potential genetic biomarkers is of particularly importance for an early diagnosis and a better prognosis of DM patients. Previous studies showed that interferon α and its associated pathways could induce autoantigen presentation, active autoreactive monocytes, cytotoxic T-lymphocytes and NK cells, elicit endoplasmic reticulum stress of human islet B cells, and impair insulin production [20,21]. These results indicated that autoimmunity might result in destruction of islet B cells, contribute to less insulin production, and give rise to the development of DM. As far as we know, this is so far the most comprehensive meta-analysis about CTLA-4 gene polymorphisms and DM, and our pooled analyses revealed that rs231775 and rs5742909 polymorphisms may serve as genetic biomarkers of T1DM, and rs231775 polymorphism may also serve as a genetic biomarker of T2DM. The stabilities of synthetic results were evaluated by sensitivity analyses, and no alterations of results were observed in any comparisons, which suggested that our findings were statistically stable. As for evaluation of heterogeneities, significant heterogeneities were detected for rs231775 polymorphism in every comparison of overall analyses for T1DM, and thus all analyses were performed with REMs. But in further subgroup analyses, a reduction tendency of heterogeneity was found in South Asians, which suggested that differences in ethnicity could partially explain observed heterogeneities between studies.
There are several points that need to be addressed about the present study. First, our findings indicated that rs231775 and rs5742909 polymorphisms could be used to identify individuals at higher risk of developing T1DM, and rs231775 polymorphism could also be used to identify individuals at higher risk of developing T2DM. There are two possible explanations for our positive findings. First, rs231775 and rs5742909 polymorphisms of the CTLA-4 gene may lead to alternations in gene expression or changes in CTLA-4 protein structure, which may subsequently affect biological functions of CTLA-4, result in immune dysfunction and ultimately impact individual susceptibility to DM, especially T1DM. Second, it is noteworthy that several analyses were still based on limited number of studies, and therefore, further replication studies, especially in T2DM are still warranted to confirm these findings. Third, the pathogenic mechanism of DM is extremely complex, and hence despite our positive findings, it is unlikely that a single genetic polymorphism could significantly contribute to its development [22,23]. Fourth, due to lack of raw data, we failed to explore possible interactions of investigated CTLA-4 gene polymorphisms. But to better illustrate the potential associations of CTLA-4 gene polymorphisms with DM, we strongly recommend further studies to perform haplotype analyses and explore potential gene-gene interactions.
Our meta-analysis certainly has some limitations. First, although the general methodology qualities of included studies were good, it should be noted that we did not have access to genotypic distributions of investigated polymorphisms according to base characteristics of study subjects. Therefore, our results were derived from unadjusted estimations , and failure to conduct further adjusted analyses for baseline characteristics of participants such as age, gender, and co-morbidity conditions may influence the authenticity of our findings [24]. Second, significant heterogeneities were detected in certain subgroup comparisons, which indicated that the inconsistent results of included studies could not be fully explained by differences in ethnic background, and other unmeasured characteristics of participants may also partially attribute to between-study heterogeneities [25]. Third, associations between CTLA-4 gene polymorphisms and DM may also be influenced by gene-environmental interactions. However, the majority of studies did not consider these potential interactions, which impeded us to perform relevant analyses accordingly [26]. Fourth, since only published articles were eligible for analyses, although funnel plots revealed no obvious publication biases, we still could not rule out the possibility of potential publication biases. Taken these limitations into consideration, the results of the present study should be interpreted with caution.
In conclusion, our findings indicated that rs231775 and rs5742909 polymorphisms may serve as genetic biomarkers of T1DM, and rs231775 polymorphism may also serve as a genetic biomarker of T2DM. Further well-designed studies, especially in T2DM are still warranted to confirm our findings, and future investigations also need to explore possible roles of other CTLA-4 gene polymorphisms in DM.