Effect of RAGE gene polymorphisms and circulating sRAGE levels on susceptibility to gastric cancer: a case–control study

This was a hospital-based case–control study. The cases were inpatients newly diagnosed with histologically confirmed gastric cancer, consecutively recruited from the First Affiliated Hospital of Guangxi Medical University, Guangxi, China between April 2015 and December 2015. Patients were excluded if they had any of the following: (1) concomitant malignant neoplasias, (2) acquired immunodeficiency syndrome, and (3) acute or chronic inflammatory diseases. Controls were randomly recruited from a pool of healthy volunteers without clinical evidence of any cancer who visited the general health check-up centers at the same hospital during the same period of the study all the participators provided written informed consent. This case–control study was approved by the ethics committee of The First Affiliated Hospital of Guangxi Medical University.

For each participant, demographic features, laboratory data, and pathological findings were obtained from electronic medical records. The demographic data included age, gender, ethnicity, body mass index (BMI), family history of cancer, smoking status, and drinking status. The laboratory data included information on carcinoembryonic antigen and carbohydrate antigen 199 levels. The pathological findings recorded were the tumor location (upper, middle, lower, and whole), tumor node metastasis stage, and differentiation.

SNP genotype information was retrieved from the dbSNP database (http://​www.​ncbi.​nlm.​nih.​gov/​projects/​SNP/​) and HapMap database (http://​hapmap.​ncbi.​nlm.​nih.​gov/​). The selection of the SNPs was based on them being well-characterized, well-known, and common functional SNPs, as well as on an extensive literature study, their population frequency based on a minor allele frequency of >5%, and a previously described association with cancer. Based on the aforementioned criteria, four SNPs were selected: rs2070600 (Gly82Ser) in the third exon, rs184003 (G1704A) in the seventh intron, rs1800624 (T-374A) in the promoter region, and rs1800625 (T-429C) in the promoter region.

The sample size was estimated using Quanto software (version 1.2.4) based on a probability of α = 0.05 and β = 0.10. The case–control design used an approximately 1:1 ratio. The assumed prevalence of the SNP rs2070600GG genotype in the control group was 0.488 (HapMap Project dbSNP database: http://​www.​ncbi.​nlm.​nih.​gov/​SNP/​). The inheritance model was recessive. The estimated marginal genetic effect was 2.0. The type I error rate was 0.05 (two-sided). According to the above parameters, the power analysis showed that a sample size of 134 would have more than an 80% power to detect a genotype-related risk.

DNA was isolated from peripheral leukocytes using the phenol–chloroform protocol, as described in our previous study [23]. The concentration and purity of the DNA were determined spectrophotometrically. The obtained DNA was stored at −20 °C until it was analyzed using the polymerase chain reaction (PCR). The RAGE rs2070600, rs184003, rs1800624, and rs1800625 polymorphisms were screened using a thermocycler PCR system, followed by a restriction fragment length polymorphism (RFLP) assay. For each SNP, the PCR was conducted in a reaction volume of 25 μl, consisting of 1 μl of each specific primer, 2 μl of genomic DNA, 12.5 μl of Green PCR Master Mix (Shanghai Sangon Biotech Co., Ltd., China), and 8.5 μl of nuclease-free water. The specific primers for rs2070600 were: forward, 5′-GAAGGTCCTGTCTCCCCAG-3′; reverse, 5′-GTAAGAGGGAGGCCTTGGAG-3′. For the rs184003, the primers used were: forward, 5′-GAGACAGGGCTCTTCACACT-3′; reverse, 5′-TTTCCCTCGTTAGCCCTCTG-3′. For the rs1800624 and rs1800625, the primers used were: forward, 5′-GGGCAGTTCTCTCCTCACTT-3′; reverse, 5′-CGTCTTGTCACAGGGAATGC-3′. The PCR conditions for these four SNPs were all as follows: initial denaturation at 95 °C for 5 min, 30 cycles of denaturation at 95 °C for 30 s, annealing at 61 °C for 30 s, extension at 72 °C for 45 s, and a final extension at 72 °C for 10 min.

Restriction enzymatic digestion of all the PCR products was performed using 1 μl of restriction enzymes (Thermo Scientific): AluI for rs2070600 and rs1800625, FspBI for rs184003, and MunI for rs1800624.The digested fragments were directly separated by electrophoresis in 3% agarose gel and visualized in ultraviolet light after GoldView I staining. For rs2070600, the sizes of the fragments were 149 + 63 bp for the G allele and 212 bp for the A allele. For rs184003, the sizes of the fragments were 35 + 467 bp for the G allele and 502 bp for the T allele. For rs1800624, the sizes of the fragments were 212 + 289 bp for the T allele and 501 bp for the A allele. For rs1800625, the sizes of the fragments were 342 + 159 bp for the C allele and 501 bp for the T allele.

To determine the accuracy of the PCR–RFLP method, about 10% of the samples were randomly selected, and the genotypes were verified by the direct sequencing method with an ABI Prism 3100 (Applied Biosystems, Shanghai Sangon Biological Engineering Technology & Services Co., Ltd., China). The resultant genotypes showed no differences.

The sRAGE levels in 90 sera samples from the gastric cancer patients and 90 sera samples from the controls were measured by a double-sandwich ELISA (Cusabio Biotech Co., Ltd., Wuhan, China), following the manufacturer’s instructions and using a standard range of 0.0–5000 pg/ml. The readers of the laboratory assay were blinded to the clinical data.

Continuous variables were expressed as means ± standard deviations (SD) or medians ± interquartile ranges. Group differences and normally distributed data were analyzed by the Student’s T test, and the Mann–Whitney U test was applied for data not normally distributed. Categorical variables were presented as frequencies and were compared between patients and controls by the χ² test and Fisher’s exact test, when appropriate. The χ² test was used to test the Hardy–Weinberg equilibrium at each locus using a contingency table of observed-versus-expected genotypic frequencies.

The genotypes of the four examined SNPs in the RAGE gene were explored by binary logistical regression analyses to obtain the ORs and their 95% CIs under the assumptions of codominant, dominant, and recessive models of inheritance after adjusting for age, gender, BMI, ethnicity, family history of cancer, smoking status, and drinking status. To evaluate the joint effects of the four SNPs in the RAGE gene, SHEsis software (http://​analysis.​bio-x.​cn/​myAnalysis.​php) [24] was employed to construct haplotypes between the patients and controls. To avoid chance findings, only common haplotypes with frequencies ≥0.03 in all the study participants were analyzed. To investigate the effect of other potential confounding variables on the association between the RAGE polymorphisms and gastric cancer risk, the population was stratified according to gender, smoking status, and drinking status. SPSS version 16.0 for Windows (SPSS Inc., IL, USA) software was used for all the statistical analyses. A two-sided P value of <0.05 was accepted as statistically significant.

Table 1 shows the demographic and clinical characteristics of all the subjects in the study. In total, 200 gastric cancer patients and 207 controls were enrolled in the study. There were no differences between the two groups in terms of mean age, smoking status, drinking status, and ethnicity. However, there was a higher proportion of males and a family history of cancer among the patients with gastric cancer than among the controls. The patient groups had a significantly lower BMI and higher serum levels of carcinoembryonic antigen and carbohydrate antigen 199 when compared to those of the healthy controls. Among the 200 gastric cancer cases, 20 (0.010), 145 (0.725), 30 (0.150), and 5 (0.025) patients had undifferentiated, well-differentiated, moderately differentiated, and poorly differentiated carcinomas, respectively; 26 (0.130), 35 (0.175), 85 (0.425), and 54 (0.270) patients had stage I, II, III, and IV tumors, respectively.

The observed genotype distributions of the SNPs rs2070600, rs184003, rs1800624, and rs1800625 in the RAGE gene were consistent with the Hardy–Weinberg equilibrium in the controls (P > 0.05). The genotype and allele frequencies of the four polymorphisms in the patients and controls are shown in Table 2. In the logistic regression analyses, a moderately higher risk of gastric cancer was observed in the AG carriers of rs2070600, with GG used as a reference (OR 1.62, 95% CI 1.03–2.58; P = 0.038), after adjustment for gender, age, BMI, family history of cancer, ethnicity, smoking status, and drinking status. Moreover, subjects carrying at least one copy of the A allele for the rs2070600 SNP (dominant model: AG + AA vs. GG) were 1.56 times more likely to develop gastric cancer (OR 1.56, 95% CI 1.01–2.39, P = 0.044). In contrast, after adjustment for the above-mentioned variables, the rs184003 GT genotype was associated with a significantly reduced risk of gastric cancer (OR 0.62, 95% CI 0.39–0.99; P = 0.048). However, there were no significant differences in the genotype and allele distributions of the other two polymorphisms (rs1800624 and rs1800625) between the cases and controls (P > 0.05).

To investigate the effect of other potential confounders on the association between the RAGE polymorphisms and gastric cancer risk, the population was stratified according to gender, smoking status, and drinking status. In addition, to better understand the prognosis of gastric cancer, the relationship between the RAGE genotype polymorphisms and cancer stage was assessed. Additional files 1, 2 and 3 present the results of the subgroup analysis by smoking status, drinking status, and gender. Nonsmokers with the rs2070600 AG genotype but not smokers with this genotype showed a significantly elevated risk of gastric cancer (adjusted OR 1.71, 95% CI 1.01–2.91; P = 0.043). As regards drinking status, in nondrinkers, the presence of the rs2070600 variant genotypes was associated with a significantly increased risk of gastric cancer (AG vs. GG: OR 1.75, 95% CI 1.03–2.97; dominant model AG + AA vs. GG: OR 1.66, 95% CI 1.01–2.74), whereas the association was not statistically significant among drinkers. No effect of gender on the association between the RAGE polymorphism and susceptibility to gastric cancer was observed. Additional file 4 presents the results of the stratified analyses by the cancer stage. As shown by the binary analysis, AG carriers of rs2070600 had a markedly higher risk of stage III gastric cancer (OR 2.00 95% CI 1.13–3.56; P = 0.018), using GG as a reference, after adjustment for age, gender, BMI, family history of cancer, ethnicity, drinking status, and smoking status. The gene polymorphisms of rs184003, rs1800624, and rs1800625 did not affect the risk of gastric cancer at different stages.

Haplotypes of the four polymorphisms in the RAGE gene were derived to detect haplotypes specifically correlated with gastric cancer. Their frequencies (≥3%) are summarized in Table 3. Four haplotypes were identified in the order rs2070600, rs184003, rs1800624, and rs1800625, with the GGTT haplotype the most prevalent in both gastric cancer patients and controls. However, no statistically significant association was found between the haplotypes and gastric cancer risk.

The average sRAGE levels in the gastric cancer patients were 56.86 ± 147.74 pg/ml, and they were 108.31 ± 132.97 pg/ml in the controls (Table 4; Fig. 1). The serum sRAGE levels in the controls were significantly higher than those in the gastric cancer patients (P = 0.015).

As shown in Table 4, decreased levels of sRAGE were found in the gastric cancer subjects carrying the rs2070600 AG genotype (49.13 ± 104.88 pg/ml) compared with those carrying the AA genotype (133.72 ± 137.73 pg/ml; P = 0.01). In addition, the serum sRAGE level of the gastric cancer patients with the rs184003 TT genotype was significantly higher (379.64 ± 0.00 pg/ml) than that of the subjects with the wild-type GG genotype (89.68 ± 114.24 pg/ml; P = 0.014).