Antioxidant-Rich Diet, GSTP1 rs1871042 Polymorphism, and Gastric Cancer Risk in a Hospital-Based Case-Control Study

Background Chronic gastritis along with Helicobacter pylori (H. pylori) infection has been implicated in inflammatory response-related genes linked to the causation of gastric cancer. Glutathione S-transferase Pi (GSTP1) plays a role in regulating oxidative stress and detoxification against carcinogenesis. In this study, we aimed to determine whether an antioxidant-rich diet is associated with gastric cancer risk and identify how this association could be altered by GSTP1 genetic variants. Methods This study included 1,245 participants (415 cases and 830 controls) matched for age and sex. The dietary antioxidant capacity was estimated based on the oxygen radical absorbance capacity (ORAC) incorporated with a semiquantitative food frequency questionnaire. Five single nucleotide polymorphisms (SNPs) of GSTP1 (rs1695, rs749174, rs1871042, rs4891, and rs947895) were selected among the exome array genotype data. Results High dietary ORAC was inversely associated with gastric cancer (hydrophilic ORAC OR T3 vs . T1, 95% CI = 0.57, 0.39–0.82, P = 0.004; lipophilic ORAC = 0.66, 0.45–0.95, P = 0.021; total phenolics = 0.57, 0.39–0.83, P = 0.005). The polymorphism rs1871042 increased the risk of gastric cancer (OR, 95% CI = 1.55, 1.10–2.16, P = 0.01, CT+TT vs. CC). A remarkably reduced risk of gastric cancer was observed among those who had a high dietary ORAC according to rs1871042 polymorphism (hydrophilic ORAC OR T3 vs . T1, 95% CI = 0.36, 0.17–0.78, P for trend = 0.013; lipophilic ORAC = 0.58, 0.37–0.93, P for trend = 0.021; total phenolics = 0.38, 0.17–0.83, P for trend = 0.019). Conclusions Our findings indicate that dietary ORAC intake may be inversely associated with the risk of gastric cancer altered by genetic variants of GSTP1, providing new intervention strategies for gastric cancer patients.


BACKGROUND
Gastric cancer (GC) was the leading cause of cancer death and the fifth most common cancer worldwide in 2018 (1). Although the global incidence rates of GC have declined, the incidence of GC in East Asia, including Korea, remains high (1,2). There are several major risk factors for the development of GC, including Helicobacter pylori (H. pylori) infection, smoking, alcohol consumption, obesity, and excess sodium intake (2). Generally, H. pylori infection is a known carcinogen and a strong risk factor for non-cardia GC by the classical histopathologic Correa cascade, consequently resulting in GC (3)(4)(5). Helicobacter pylori infection is particularly associated with an increased risk of not only non-cardia GC but also cardia GC according to several studies targeting East Asian countries, such as Korea (5,6). In addition, evidence suggests that gastritis derived from chronic inflammation of normal mucosa may be linked to various other dietary factors, such as a high intake of salted or preserved foods and grilled or processed meats and a low intake of fruits (7,8).
Dietary effects have been reported to mediate the risk of cancer by playing a role in either the prevention of cellular carcinoma or diet-induced carcinogenesis (9,10). Helicobacter pylori infection is required to consider the causes of GC derived from either direct or indirect inflammation in the gastric mucosa (11,12). Cumulative studies have reported that multiple H. pylori-induced inflammatory responses from reactive oxygen species (ROS) can be suppressed by bioactive compounds abundant in fruits and vegetables (13,14). Moreover, current studies have consistently demonstrated that a variety of bioactive compounds, such as phytochemicals, play pivotal roles in sympathetic activation ranging from the inhibition of cellular proliferation to the suppression of metastasis in gastric carcinoma cells (15)(16)(17). Given that the risk of GC is linked to dietary factors and the inflammatory response associated with H. pylori infection, this study focused on exploring the integrated and synergistic effects of antioxidants on GC using the oxygen radical absorbance capacity (ORAC) of a diet. The ORAC is an experimental value representing the total antioxidant capacity (TAC) and is used to indicate the capacity to scavenge free radicals from food components indicating hydrophilic ORAC (H-ORAC), lipophilic ORAC (L-ORAC), and total phenolics (TPs) (18). The use of ORAC to assess dietary effects on disease has the advantage of exploring the antioxidant activity of food rather than that of a specific nutrient (19). Based on the benefits of using ORAC, an examination of the antioxidant capacity of a diet using this metric in the context of GC risk with inflammation and H. pylori infection is needed.
Among the inflammation-and oxidative stress-related genes, glutathione S-transferase Pi (GSTP1) is a cytosolic detoxifying enzyme that encodes Pi-class glutathione S-transferases (GST) and is involved in phase II xenobiotic metabolism by conjugating glutathione with hydrophobic and electrophilic substrates (20)(21)(22). The deregulation of GSTP1 contributes to inducing oxidative stress by producing excessive ROS, leading to various types of tumors, including esophageal, stomach, lung, breast, and colorectal cancer (23)(24)(25)(26)(27)(28)(29)(30)(31). Regarding the risk of GC, studies exploring genetic polymorphisms of GSTP1 revealed the role of GSTP1 in and relevance of GSTP1 for GC susceptibility (26,27). Some epidemiological studies have reported a significant association between GSTP1 rs1695 and the risk of GC (32,33). One genetic polymorphism of GSTP1, namely, c.313 A > C (rs1695) in exon 5, results in an amino acid change (A to G) of isoleucine (Ile) to valine (Val), leading to impaired detoxification and catalytic activity (34). Furthermore, different genotypes of GSTP1 rs1695 have significant interaction effects with environmental factors, including H. pylori infection, smoking, and alcohol consumption on the risk of GC (35)(36)(37). However, evidence regarding the associations between dietary factors and GSTP1 polymorphisms in the context of GC risk based on the effects of antioxidants and imbalanced oxidative stress mechanisms is insufficient.
Given these points, we selected five single nucleotide polymorphisms (SNPs; rs1695, rs749174, rs1871042, rs4891, and rs947895) in GSTP1 that were found among the 713,348 SNPs assayed in a Korean population based on quality control (QC) criteria in genome-wide association studies (GWASs), determined the association between ORAC and GC, and examined whether this association was modified by GSTP1. The aim of this study was to identify how dietary ORAC intake is associated with GC risk alterations by GSTP1 genetic variants. We evaluated whether dietary ORAC intake affects the risk of GC. Additionally, we explored the associations between GC risk and ORAC intake according to GSTP1 genotypes.

Study Population
This case-control study was conducted at the Center for Gastric Cancer (CGC) and the Center for Cancer Prevention & Detection (CCPD) of the National Cancer Center (NCC) in Korea between March 2011 and December 2014. The cases were recruited among patients who were diagnosed with early GC within the preceding 3 months with confirmed invasive carcinoma in the CGC. Individuals with diabetes mellitus, severe systemic or mental disease, or a history of cancer and women who were pregnant or breastfeeding were excluded. The control group comprised individuals who underwent health screening check-ups at the CCPD at the same hospital. Participants in the control group who had diabetes mellitus, gastric or duodenal ulcers, a history of cancer, or previous H. pylori treatment were excluded. Of these initial 1,727 participants who were enrolled in this study, 56 individuals with an incomplete semiquantitative food frequency questionnaire (SQFFQ) and 15 individuals with an implausible total energy intake (< 500 or > 4,000 kcal/day) were excluded. Among the remaining 1,656 participants, the cases and controls were frequency-matched at a ratio of 1:2 (case: control) by 5-year age groups and sex. Regarding the genetic variants of GSTP1, we excluded low-quality samples and markers from the cases and controls as follows: rs1695 (n = 38 and n = 74), rs749174 (n = 50 and n = 113), rs1871042 (n = 50 and n = 113), rs4891 (n = 44 and n = 91), and rs947895 (n = 50 and n = 113). Consequently, the total population included the final analysis was as follows: rs1695 (n = 1,133), rs749174 (n = 1,082), rs1871042 (n = 1,082), rs4891 (n = 1,110), and rs947895 (n = 1,082) ( Figure 1). This study was approved by the Institutional Review Board (IRB) of NCC (IRB number: NCCNCS-11-438), and written informed consent was obtained from all participants.

Data Collection and Dietary Antioxidant Capacity Based on the ORAC Database
The sociodemographic characteristics were collected from each participant using a self-administered questionnaire. The status of H. pylori infection was assessed histologically or serologically from at least a positive result on a rapid urease test (Pronto Dry, Medical Instruments Corporation, Solothurn, Switzerland). The dietary intake data were obtained using a validated and reliable 106-item SQFFQ administered by a well-trained interviewer (38). The daily nutrient intake was calculated based on a combination of average intake frequency (never or rarely, 1 time per month, 2-3 times per month, 1-2 times per week, 3-4 times per week, 5-6 times per week, one time per day, two times per day, and three times per day) and the portion size (small, medium, and large) using CAN-PRO 4.0 (computer aided nutritional analysis program, Korean Nutrition Society, Seoul, Korea).
To estimate the values of the dietary antioxidant capacity, the ORAC database from the USDA release 2 was incorporated into our dietary intake data according to the food description (39). Our previous study reported the associations between dietary ORAC intake and interleukin-6 levels regarding the risk of colorectal cancer (40). Briefly, the ORAC database contains the antioxidant activity level of 326 food items. To calculate the dietary ORAC of each participant, the 106-item SQFFQ was integrated into the ORAC database by common food items. Of these food sources, 56 food items, including mostly fruits and vegetables, from the SQFFQ were considered to have antioxidant compounds and selected for the analysis in this study. It reported H-ORAC and L-ORAC as mmol of trolox equivalents per 100 g (mmol TE/100 g) and TPs as mg gallic acid equivalents per 100 g (mg GAE/100 g). We calculated each daily index of ORAC through the same process applied in the daily nutrient intake.

Genotyping and SNP Selection
In this study, the SNPs of GSTP1 were selected based on our previous study using exome array genotype data as described elsewhere (41). The genomic DNA samples were extracted from the peripheral blood leukocytes of all participants. The genotyping was performed using an Affymetrix Axiom ® Exome 319 Array containing 318,983 SNPs (Affymetrix Inc., Santa Clara, CA, USA). In the QC procedure applied to genotype data, the samples and genetic markers were excluded according to the call rate (< 95%), deviation from HWE, and MAF (< 0.01) (41,42). After genotype imputation with an Asian population of 1000 Genome haplotypes phase III, we selected five SNPs of GSTP1 (rs1695, rs749174, rs1871042, rs4891, and rs947895) for further analysis (Supplementary Table S1). The LD patterns of the SNPs were analyzed for the efficient selection of tag SNPs in GSTP1 according to the pairwise D' and r 2 using Haploview (43).

Statistical Analysis
The differences in the sociodemographic, anthropometric, lifestyle factors, and total energy intake between the cases and controls were assessed by using the c 2 test for categorical variables and Student's t-test for continuous variables. The dietary ORAC intake and foods contributing to each ORAC, covering up to 90% of the cumulative contribution of 56 food items, were compared between the cases and controls by using a Wilcoxon signed-rank test because of its distribution (Supplementary Tables S2-S4). Energy-adjusted ORAC values and their contributing foods adjusted by a residual method were used in the analyses (44). The values of H-ORAC, L-ORAC, and TPs were divided into three groups depending on the median value of the controls. To analyze the associations among dietary ORAC intake, GSTP1 polymorphisms and GC risk, unconditional logistic models were constructed to estimate the odds ratios (OR) and 95% confidence intervals (95% CI) of the risk of GC while considering potential confounding factors, such as age, BMI, education level, income, physical activity, smoking status, first-degree family history of GC, and total energy intake identified by the backward selection procedure in a stepwise regression analysis. The H. pylori infection status was additionally considered in the final statistical models. The analyses of the associations between dietary ORAC intake and GC risk were also stratified by GSTP1 SNPs, particularly in the dominant model. All tests were performed using the SAS package (SAS 9.4; SAS Institute Inc., Cary, NC, USA) with a two-sided Pvalue of 0.05 regarded as significant. Table 1 shows the distribution of the sociodemographic characteristics and dietary ORAC intake in the cases and controls. The GC patients had a higher prevalence of positive H. pylori infection, a first-degree family history of GC and current smoking status, and a lower prevalence of regular exercise, education level, and income than the control group (P < 0.05). However, no differences in age, sex, BMI, or alcohol consumption were observed.

General Characteristics
The cases had a higher intake of daily total energy than the controls (

Association Between Dietary ORAC Intake and GC Risk
The associations between each index of dietary ORAC intake and GC risk are presented in Table 2.

Haplotype of GSTP1 Polymorphisms and Association With GC Risk
The minor allele frequencies of the five SNPs (rs1695, rs749174, rs1871042, rs4891, and rs947895) were common (minor allele frequency, MAF > 5%), and the genotype frequencies of the SNPs were consistent with Hardy-Weinberg equilibrium (HWE) (Supplementary Table S1). We identified the linkage disequilibrium (LD) structure of the five SNPs using a pairwise LD test (Figure 2). We observed that all five SNPs of GSTP1 gene were in the same block with high LD, supporting the strong correlation among the five SNPs in this study. According to the LD patterns with r 2 > 0.8, tag SNPs (rs1695 and rs1871042) in GSTP1 were selected. Table 3 presents the associations between GSTP1 variants and GC risk. Apart from H. pylori infection, an       ORAC, oxygen radical absorbance capacity; H-ORAC, hydrophilic oxygen radical absorbance capacity; L-ORAC, lipophilic oxygen radical absorbance capacity; TPs, total phenolics; TE, trolox equivalents; GAE, gallic acid equivalents; T, tertile; OR, odds ratio; 95% CI, 95% confidence interval. Model I: crude OR; model II: age (continuous), BMI (<25 kg/m 2 or ≥25 kg/m 2 ), education level (less than college or college and higher), income (<200, 200-<400 or ≥400), physical activity (yes or no), smoking status (current, ex-or non-smoker), first-degree family history of GC (yes or no), and total energy intake; Model III: additionally adjusted for H. pylori infection (positive or negative).
rs1871042 polymorphism on gastric cancer risk, each dietary ORAC intake was divided into low and high groups based on the median level of the intake of controls (Supplementary Table  S7). Although there are no interaction effects between dietary ORAC intake and GSTP1 rs1871042 polymorphism on gastric cancer, a high intake of ORAC significantly reduced the risk of gastric cancer in patients homozygous for CC at rs1871042 after adjusting for potential confounding factors (OR, 95% CI:

DISCUSSION
The present study aimed to determine the association between dietary ORAC intake and GC risk according to GSTP1 genetic variants. A high dietary intake of ORAC was significantly associated with a decreased risk of GC in a Korean population.
Regarding the genetic variants of GSTP1 gene, dietary ORAC intake was inversely associated with GC risk according to GSTP1 rs1871042 genotypes. Gastric adenocarcinoma occurs when normal mucosa cells are continuously exposed to a variety of carcinogens that lead to uncontrolled cell proliferation in the gastric mucosa membrane (45). The following two major mechanisms are linked to the development of GC with H. pylori infection: (1) epigenetic alterations in gastric epithelial cells by H. pylori infection and (2) H. pylori-induced inflammation in the gastric mucosa (46). Many studies have shown that persistent inflammation, through cytokines, chemokines, growth factors, and oxygen-derived free radicals is responsible for GC risk associated with H. pylori infection (46,47). The role of oxidative stress from inflammation in GC has been determined, suggesting the importance of a balance between radical production and the antioxidant defense system (48). Numerous studies have reported that the intake of fruits and vegetables is inversely associated with GC risk, while some studies found no such associations (49)(50)(51)(52)(53). Specifically, a high intake of fruits by H. pylori-negative subjects decreased the risk of GC compared to a low intake of fruits by H. pylori-positive subjects, indicating that the intake of fruits and vegetables may play a role in preventing H. pylori-induced gastric carcinogenesis (52)(53)(54). In contrast, data regarding the effects of vitamin A, vitamin C, vitamin E, and carotenoids on GC risk were inconsistent or conflicting due to the different doses used (55)(56)(57). In this study, we examined the antioxidant capacity of food and determined the antioxidant effects of ORAC on gastric carcinogenesis. A recent meta-analysis reported inverse associations between cancer risk and dietary TAC by using multiple methods, including ORAC (58). Other previous studies found inverse associations between ORAC intake and risk of other cancers but not GC (59)(60)(61)(62). We observed similar findings between GC risk and three indices of dietary ORAC, namely, H-ORAC, L-ORAC, and TPs, after adjusting for H. pylori infection and other potential confounding factors. Furthermore, in the comparisons of the food items that highly contribute to the ORAC level, the food items with the highest ORAC were brewed green tea and fruits for H-ORAC, spicy red, or black pepper for L-ORAC, and canned tomato juice for TPs ( Supplementary Tables S2-S4).
The major function of GSTP1 is to detoxify exogenous or endogenous factors involved in carcinogenesis by regulating cell death and DNA damage (21,63). Additionally, GSTP1 plays a role as a modifier gene in the regulation of the molecular expression and activation of enzymes from other GST subfamilies and their effects on cancer, and GSTP1 expression regulates cellular redox homeostasis in carcinogenesis (20,64,65). Although many studies have shown associations between GSTP1 polymorphisms and various types of cancer, the results of a few studies investigating the associations between GC risk and GSTP1 genetic variants are inconsistent across geographic areas and diverse populations. In a Chinese population, the Val allele of GSTP1, namely, the Val/Val genotype, was significantly associated with an increased risk of GC (37,66,67). However, GSTP1 Ile105Val (rs1695) and GSTP1 Val114Ala (rs1138272) polymorphisms were not associated with the risk of GC in either a South European or an Indian population (68,69). In a Korean population, we observed that five GSTP1 polymorphisms (rs1695, rs749174, rs1871042, rs4891, and rs947895) located in the same block with a strong correlation with high LD had a tendency to increase GC risk, although the risk increase with rs1695 polymorphism was not statistically significant. These conflicting results suggest that ethnic differences in GSTP1 genetic susceptibility may affect the development of GC with epigenetic interactions of environmental factors and that the relevance of GSTP1 genetic variants to GC risk needs to be confirmed in future studies. Among five GSTP1 polymorphisms examined in this study, four polymorphisms (rs749174, rs1871042, rs4891, and rs947895) have been investigated in only a few studies in the context of lung cancer and asthma, and to date, their associations with GC risk have not been determined (70)(71)(72).
In this study, we observed an association between a high intake of dietary ORAC and a reduced GC risk according to GSTP1 rs1871042 polymorphism. Our findings can be explained by the interconnections between dietary TAC and the role of GSTP1 gene in the regulation of oxidative stress and detoxification of the immune response against gastric carcinogenesis-induced chronic inflammation by H. pylori infection. Imbalanced oxidative stress plays an obligatory role in gastric carcinogenesis by increasing the level of ROS induced by H. pylori infection, leading to DNA damage and tumor progression (4,73). A high intake of dietary ORAC is responsible for the scavenging substances produced by H. pyloriinfected gastric cells and, thus, may protect against the promotion of gastric carcinogenesis. More than half of H. pylori strains produce various cytotoxins, such as Cag-A, which can damage gastric mucosal cell membranes and trigger local immune responses (74). Previous studies have shown that vitamin C protects against H. pylori infection-related GC by neutralizing free radicals and directly modifying the anticancer immune response against malignant progression (75,76). In addition to the role of GSTP1 gene, the specific allele of GSTP1 is able to regulate oxidative stress and detoxification against carcinogenesis (21,65). Moreover, the impact of H. pylori infection on the relationship between GSTP1 genetic polymorphisms and GC risk varies, suggesting that H. pylori infection may have different oncogenic effects depending on GSTP1 genetic polymorphism, including controlling the activation of the detoxification system, thereby resulting in gastric carcinogenesis (36,77). A high intake of dietary ORAC may synergistically interact with GSTP1 rs1871042 polymorphism by detoxifying and eradicating excessive ROS, eventually leading to protection against the development of GC.
Nevertheless, some limitations should be noted. First, selection and recall bias should be considered; the controls were recruited among patients who visited the clinic for a health check-up and may have been more health conscious than the patients with GC. To reduce the selection bias, controls who were confirmed to be cancer-free by linking to the Korea Central Cancer Registry database were recruited. However, it may be that the individuals who chose to visit a health check-up program may have a healthier lifestyle than those who did not choose to undergo a check-up. Moreover, the participants provided the structured questionnaire and validated SQFFQ by a well-trained interviewer to reduce the recall bias. The SQFFQ includes the average intake frequency and the portion size during the year preceding the interview. Second, the food items included in our food database were insufficient to cover the entire United States Department of Agriculture (USDA) ORAC database. Additionally, the antioxidant capacity from ORAC is based on in vitro antioxidant assays, which are limited to measuring the absorption rate in the body. Third, the sample size in each tertile of the case group is relatively small. Further prospective studies are needed to confirm and extend our findings with a larger sample size.

CONCLUSIONS
In conclusion, this study examined whether the associations between dietary ORAC and GC risk were modified by GSTP1 polymorphisms. We found associations between the risk of GC and dietary ORAC intake, including fruits, vegetables, spices, and nuts, depending on the genetic variants of GSTP1. Considering the highest incidence rates of GC with H. pylori infection in East Asia, the associations among dietary ORAC intake, GSTP1 polymorphisms, and GC risk may provide an effective strategy for the primary prevention of GC in Asian populations.

DATA AVAILABILITY STATEMENT
The datasets presented in this study can be found in online repositories. The names of the repository/repositories and accession number(s) can be found in the article/Supplementary Material.

ETHICS STATEMENT
The studies involving human participants were reviewed and approved by the Institutional Review Board (IRB) NCC (IRB number: NCCNCS-11-438), and written informed consent was obtained from all participants. The patients/participants provided their written informed consent to participate in this study.