Genetic Variability in eIF2α Gene Is Associated with Islet β-Cell Function in the Development of Diabetes in a Chinese Han Population

Aims Protein kinase-like endoplasmic reticulum kinase (PERK)/eukaryotic translation initiation factor 2 alpha (eIF2α) pathway mutations lead to failure of β-cell function. The aim of this article was to assess the association between eIF2α and the risk of glucose metabolism abnormalities. Methods Two eIF2α SNPs (rs9840992 T>C and rs13072593 A>G) were selected based on CHB data from HapMap, and 1466 unrelated nondiabetes individuals were genotyped. All subjects were examined by the 75 g oral glucose tolerance test, and 733 participated in a subsequent insulin release test. Various indicators of insulin resistance and islet β-cell function were examined. Results There were no significant differences in genotype distribution and allele frequency between the prediabetes and controls. CC genotype carriers at rs9840992 showed higher insulin levels at 120 min after a 75 g glucose load than noncarriers. Also, CC homozygotes had higher ΔI30/ΔG30 and ΔI120/ΔG120 than noncarriers, even after adjusting for insulin resistance. CC homozygotes had greater AUCi values than noncarriers. Subjects aged ≥ 65 yrs, those with a BMI ≥ 24 kg/m2 and those carrying the rs9840992 risk allele, had a 2.5-fold higher risk of glucose abnormalities than subjects who had none of these risk factors. Conclusion The eIF2α polymorphism is associated with islet β-cell function in a Chinese population.


Introduction
Pancreatic β-cell dysfunction is a key factor in the development and progression of type 2 diabetes. Prediabetes, which is characterized by impaired fasting glucose and/or impaired glucose tolerance [1], pancreatic β-cell dysfunction, and insulin resistance are already in the early stages. Several genetic and environmental factors are likely to contribute to the progressive abnormality or even failure of pancreatic β-cell function, but the precise mechanism is still unknown. However, endoplasmic reticulum (ER) stress is a likely contributing factor.
Many studies have demonstrated that the unfolded protein response (UPR), induced by the accumulation of unfolded or misfolded proteins in the ER, plays an important role in chronic metabolic diseases, such as insulin resistance and type 2 diabetes. EIF2α is part of the complex that recruits the initiator methionyl-tRNA to the ribosome and is phosphorylated by protein kinase-like ER kinase (PERK) [2]. The PERK/eIF2α pathway plays a key role in transcriptional control [3,4]. We inferred that genetic variation in eIF2α might contribute to changes in pancreatic β-cell function or insulin sensitivity. To evaluate this hypothesis, we studied genetic variation in the eIF2α gene in a Chinese population with prediabetes. World Health Organization criteria based on the oral glucose tolerance test (75 g OGTT), which stipulates fasting plasma glucose (FBG) values of between 6.1 mmol/L and 7.0 mmol/L, 2 h post-OGTT (PBG) levels of between 7.8 mmol/L and 11.1 mmol/L, or both. A total of 1466 unrelated Han Chinese subjects were included in the study, of which 845 were prediabetes and 621 were included as controls. Controls were ≥55 years old and had no family history of diabetes.

Statistical Analysis
All statistical analyses were performed using the SPSS statistical package (version 13.0; SPSS, Chicago, IL, USA). Genotype distributions were used to assess departures from Hardy-Weinberg equilibrium at each polymorphic locus. Clinical and laboratory data are expressed as means ± standard deviation or as medians (lower-upper quartiles). Comparisons of the clinical and laboratory parameters between the prediabetes and control groups as well as between genotype groups were performed using unpaired Student's t-tests or χ 2 tests. Associations between prediabetes and genotypes were analyzed using multiple logistic regressions, with adjustments for the following potential confounders: gender, age, BMI, and waist-hip ratio. As a descriptive measure of the association between genotypes and outcomes, p < 0 05 was considered statistically significant and odds ratios (ORs) were calculated with 95% confidence intervals (CIs). Linkage disequilibrium, haplotype block structure (using a CI algorithm [10]), and haplotype analyses were performed using HaploView 4.1 (http://www.broadinstitute.org/haploview/haploview) [11]. Haplotype frequencies were estimated using an accelerated expectation-maximization algorithm [12]. Bonferroni correction was used to correct for multiple comparisons. Power and sample size calculation (version 3.1.2, 2014) was used for power calculation [13]. Statistical power was 0.826.

Results
The genotype distributions were in Hardy-Weinberg equilibrium at both loci (p > 0 01). People with prediabetes were younger and had higher BMIs and waist-hip ratios, higher TG levels, and lower HDL-C levels than control individuals ( Table 1).
The genotype distribution and allele frequencies were not significantly different between the prediabetes and control groups ( Table 2). SNPs rs9840992 and rs13072593 were distributed in one linkage disequilibrium (LD) block. We also compared the haplotype distributions between prediabetes and controls. Haplotype (AT, GC, and AC) distributions between prediabetes and controls showed no statistical significant difference.
We measured plasma insulin following 75 g OGTT in 748 of the 1466 total individuals. The carriers of genotype CC at rs9840992 had higher insulin levels at 120 min after the 75 g glucose load (INS120) than noncarriers (49.86 (30.38-79.61) versus 44.32 (26.42-70.75), p = 0 038). We  (Table 3). When we included age ≥ 65 yr, BMI ≥ 24 kg/m 2 , and carrying the risk allele T for the SNP rs9840992 as risk factors of prediabetes, we observed that the more factors an individual carried, the higher the risk for prediabetes (p ≤ 0 001). Individuals who had two risk factors had an 80% higher risk (OR = 1.882, p = 0 0 037) than individuals without risk factors, and subjects who had all three risk factors had an approximately 2.5-fold greater risk (OR = 2.428, p = 0 005) compared with individuals who had none of the three factors (Figure 1).

Discussion
The ER is a complex membranous network in eukaryotic cells and is the site for the synthesis, folding, assembly, and posttranslational modification of proteins. It includes a highly conserved system of proteins that facilitates protein folding and processing and protects cells from the toxic effects of unfolded protein accumulation (i.e., ER stress). When these functions fail, the UPR is activated by the accumulation of unfolded proteins, and the apoptosis process is initiated [14][15][16]. The UPR pathway is particularly important in adipocytes, hepatocytes, and pancreatic β-cells [17]. There are three ER transmembrane sensors of unfolded proteins in murine cells: activating transcription factor 6α, inositol-requiring 1α, and the PERK/eIF2α pathway. Many studies have demonstrated that the PERK/eIF2α pathway is related to pancreatic β-cell function [16,18]. EIF2α can also upregulate the expression of ATF4 and CHOP, leading to the apoptosis of islet β-cells [17].   In this study, we observed subjects with prediabetes to determine whether genetic variation in eIF2α contributes to the very early stage of pancreatic β-cell dysfunction. Interestingly, we observed an association between the SNP rs9840992 and islet β-cell function, but not prediabetes, in the Chinese Han population. To confirm this association, we used various parameters to evaluate islet β-cell function and obtained consistent results. Carriers of the major allele T had lower insulin levels, as estimated by ΔI30/ ΔG30 (p = 0.008) and ΔI120/ΔG120 values (p = 0 017), than noncarriers, even after adjusting for insulin resistance (p′ = 0 004 and p′ = 0 006, resp.). Additionally, individuals carrying the T allele had lower AUCi values than noncarriers (p = 0 023), and this difference was highly significant, even after adjusting for insulin resistance (p′= 3.7 × 10 −4 ).
When we analyzed age ≥ 65 yr, BMI ≥ 24 kg/m 2 , and carrying the risk allele T of SNP rs9840992 as risk factors of prediabetes in a cumulative risk analysis, we found that the odds of glucose homeostasis disturbance were related to the number of the risk factors; that is, a higher number of factors was associated with an increased risk for prediabetes. Subjects who had all three risk factors had a risk that was 1.5 times higher than that of subjects who had none of the factors (p = 0 005). Several studies have confirmed that aging and obesity are risk factors for diabetes owing to their associations with β-cell dysfunction. Therefore, age, BMI, and genetic variation at rs9840992 each play a role in pancreatic β-cell function and thereby contribute to the cumulative susceptibility to diabetes. Our results suggest that individuals who carry the risk allele at this locus may benefit from early intervention to reduce body weight, which can significantly reduce the incidence of prediabetes and even diabetes.
The SNP rs9840992 is located in an intronic region. The mechanisms by which genetic variation in introns affects β-cell function are unknown. However, Zheng et al. [19] showed that intronic SNPs influence gene expression. Therefore, we postulate that rs9840992 modifies the mRNA and protein expression levels of eIF2α, and these changes may contribute to the modulation of islet β-cell apoptosis and insulin secretion. However, it is more likely that the SNP is in strong linkage disequilibrium with other SNPs that have biological effects.
No studies have examined eIF2α in prediabetes or type 2 diabetes patients. However, this study had some limitations. The relatively small sample size might bias the results. In addition, functional studies are necessary. Our findings should be replicated in other populations, and the SNPs should be confirmed as predictive genetic markers in prospective studies.

Conclusion
Our findings suggest that genetic variation at the eIF2α locus is associated with islet β-cell function in a Chinese population.

Conflicts of Interest
There are no potential conflicts of interest relevant to this article.