Alzheimer's disease (AD) is a progressive neurodegenerative disorder which has become a worldwide public health problem (1). The pathogenesis of AD is complex, and it may be contributed by genetic and environmental factors. The external environment can affect DNA methylation to change phenotype and gene expression (2). DNA methylation is an important epigenetic mechanism regulating the expression of aging genes in brain (3). The expression and closure of methylation regulatory genes are closely related to the human nervous system (4) and cognitive function (5).

The major challenge of AD is to identify new therapeutic targets and to develop new therapies for this disease (6). AD is closely related to tau hyperphosphorylation, oxidative stress, amyloid-β (Aβ) production, neuronal apoptosis, gene mutation, apolipoprotein E (APOE). Bridging integrator 1 (BIN1) is an important gene in the modulation of tau pathology, and BIN1 knockdown was shown to significantly suppress tau-mediated neurotoxicity (7). Sortilin-related receptor 1 (SORL1) is a member of the low-density lipoprotein receptor family that reduces amyloid-β (Aβ) production by regulating the intracellular transport and processing of APP (8). Nerve growth factor (NGF) contributes to the survival, regeneration and death of neurons during aging and in neurodegenerative diseases (9). PSEN2 is a transmembrane protein and AD-related presenilin mutations can alter intracellular calcium signaling, which leads to Aβ aggregation to form brain plaques and neuronal cell death (10). Genetic variation within these genes is associated with an increased risk of AD (9,11–14). The 8-oxoguanine DNA glycosylase 1 (OGG1) is a bifunctional enzyme with both glycosylase and AP lyase activities (15). Decreased OGG1 activity occurs early in the progression of AD (16). OGG1 was largely hypomethylated in LOAD and control blood DNA, and they do not support an increased promoter methylation of OGG1 in blood DNA of AD patients (17). Dihydrolipoamide succinyltransferase (DLST) is a subunit enzyme of the a-ketoglutarate dehydrogenase complex in the Krebs cycle. Polymorphisms of DLST were associated with AD in both Japanese and Caucasian populations (18–20).

In the present study, we aimed to validate the association of the five AD-associated variants (OGG1 rs1052133, BIN1 rs744373, SORL1 rs1133174, PSEN2 rs8383, and NGF rs6330) with AD in Xinjiang population. We also tested the association of OGG1 and DLST promoter methylation with AD.

Epidemiological investigation was carried out in Xinjiang province of China between 2014 and 2015. A total of 17 AD patients (75.65±5.86 years) and 34 well-matched controls (77.59±7.41 years) were selected for the present study (Table I). This study was approved by the First Affiliated Hospital of Xinjiang Medical University Ethics Committee. All the patients gave their written informed consent forms for the current study. The clinical diagnosis of AD was done according to the criteria of the Diagnostic and Statistical Manual-IV (DSM-IV). The details were the same as previously described (21). Whole blood was stored in EDTA tube at −80°C. Genomic DNA was extracted and dissolved in TE buffer, and then it was stored at −20°C. Polymerase chain reaction (PCR) was carried out in 40 µl volume containing 2 µl of each primer, 4 µl genomic DNA, 12 µl ddH2O and 20 µl 2X HotTaq Master Mix. PCR was performed in a Veriti 96-well thermal cycler (Applied Biosystems; Thermo Fisher Scientific, Inc., Waltham, MA, USA). Genotyping was done using the Sanger sequencing. The primer sequences were TTACTTCTCCACGGACAAC and CAAGATTCTAACAGGACTCATC for the forward and the reverse primers of PSEN2 rs8383 genotyping, GCCAGTCCATCTTCTTCT and ACCACATCTTAGCCACAG for the forward and the reverse primers of BIN1 rs744373 genotyping, CATCCATACTGCCTGAGTC and CCTGTGAGTCCTGTTGAAG for the forward and the reverse primers of NGF rs6330 genotyping, GTGGATTCTCATTGCCTTC and AAACTGACTGCTTGATTTGG for the forward and the reverse primers of OGG1 rs1052133 genotyping, and TGTGACTTGTGCTGTATGAT and ACGCTAGAAGAAGGCTTATC for the forward and the reverse primers of SORL1 rs1133174 genotyping. PCR consisted of an initial melting step at 95°C for 10 min, 35 cycles (NGF, BIN1, and OGG1) or 37 cycles (PSEN2) or 40 cycles (SORL1), and a final extension step at 72°C for 2 min. The cycling program was 95°C for 30 sec, 58°C (NGF and BIN1) or 54°C (OGG1) or 57°C (PSEN2) or 53°C (SORL1) for 45 sec for annealing, and 72°C for 30 sec. DNA bisulphite conversion was done using the EZ DNA Methylation-Gold™ Kit (Zymo Research Corp., Irvine, CA, USA). The details of bisulphite conversion were the same as previously described (22). Promoter methylation status of OGG1 and DLST were examined utilizing quantitative methylation-specific PCR (qMSP). The primer sequences were CGGTGGTTGAGTTTTATTTTC and CTCCTTACGACTTATCTTCTC for the upstream and the downstream primers of OGG1, respectively. And the upstream and the downstream primer sequences of DLST were GTTGTAGTCGGGATATTGG and CGAAACGAACCACTAACA, respectively.

The characteristics of AD and control groups were presented in Table I. Our results showed the two groups were well paired according to the facts that there were no significant difference on gender, age, hypertension, diabetes, lipid levels, smoking and drinking status between the AD group and control group (P>0.05). As shown in Table II, there were no associations of the five genetic polymorphisms with AD. Further APOE ε4 based subgroup analysis indicated there were no significant interaction of APOE ε4 with the five genetic variants (Table III, P>0.05).

The promoter regions of OGG1 and DLST were selected for the current methylation study (Fig. 1). In this study, we investigated the association of the methylation levels of OGG1 and DLST genes with AD (Fig. 2). Although OGG1 methylation was not associated with AD, our results showed that OGG1 methylation was significantly lower in AD patients with APOE ε4 allele than AD patients with APOE non-ε4 allele (P=0.027). Among AD patients older than 75 years old, the levels of OGG1 methylation were significantly lower in AD patients carrying APOE ε4 allele than AD patients who did not carry APOE ε4 allele (P=0.046).

As shown in Fig. 2, DLST methylation levels were significantly lower in AD patients (P=0.027). Further subgroup analysis by gender showed that the association of DLST methylation with AD was specific in females (P=0.025). Further subgroup analysis by APOE ε4 locus showed that DLST methylation was associated with AD in the APOE non-ε4 individuals (P=0.029). In the control group, the level of DLST methylation was positively correlated with TC (r=0.401, P=0.019; Table IV, Fig. 3). Further stratification by gender showed age and OGG1 methylation levels were significantly correlated in AD group (male: r=0.762, P=0.046; female: r=−0.753, P=0.012; Table V, Fig. 4). In control group, the level of DLST methylation was inversely correlated with age (r=−0.414, P=0.015), and further stratified by gender showed that there was an inverse correlation between age and DLST methylation in males (r=−0.607, P=0.010).

Previous studies have revealed the association of five variants with AD, including OGG1 rs1052133, BIN1 rs744373, SORL1 rs1133174, PSEN2 rs8383, and NGF rs6330 (15,23–27). And BIN1 rs744373 was found to have no interaction with APOE ε4 genotype (28). In the present study, we were unable to repeat the association of the above five variants with AD. And further APOE ε4 based subgroup analysis indicated that APOE ε4 did not have significant effects on five genetic polymorphisms. This might be explained by the moderate power and different ethnic background in the present pilot study. Future validation is needed in cohort with more samples.

DLST is a core component of KGDHC which is essential in the citric acid cycle (29). Deficiency of DLST will increase production of free radicals thereby inducing mitochondrial damage (29), which leads to an increase in the generation of reactive oxygen species (ROS). ROS damage various molecules, including DNA, protein and lipid, and induce apoptosis (30), eventually leading to the occurrence of AD (31). The results of this study suggest that DLST hypomethylation may contribute to the pathogenesis of AD in females. Women are more likely to have AD than men because women tend live longer than men (32–34). This finding might also help explain the sex differences in the risk of AD (35).

There were several limitations in the current study. Firstly, our pilot study only involved a moderate number of subjects (17 AD cases and 34 controls). This was due to the incidence rate of AD being low in Xinjiang. However, we chose a total of 51 well preserved samples, for which the transport process was reasonable and the basic informations were completed and matched. We were unable to validate the association of five gene polymorphisms (OGG1 rs1052133, BIN1 rs744373, SORL1 rs1133174, PSEN2 rs8383, and NGF rs6330) with AD in the Xinjiang population. This might be due to the limited number of samples in this study. Secondly, we only selected a fragment in the promoter CpG rich region to represent the methylation of OGG1 and DLST. The methylation of other regions of the two genes might be explored in the future. Thirdly, Xinjiang Uygur Autonomous Region is a multi-ethnic area. Future research with larger sample sets and more ethnic populations are required to confirm the present findings.

In summary, we found that the levels of DLST methylation were decreased in AD patients, especially in female AD patients. The results showed that the level of OGG1 promoter methylation might be interacted with APOE ε4 genotype.