Association of Polygenic Variants Involved in Immunity and Inflammation with Duodenal Ulcer Risk and Their Interaction with Irregular Eating Habits

Genetic and environmental factors are associated with developing and progressing duodenal ulcer (DU) risk. However, the exact nature of the disease pathophysiology and the single nucleotide polymorphism (SNP)—lifestyle interaction has yet to be determined. The purpose of the present study was to examine the SNPs linked to DU risk and their interaction with lifestyles and diets in a large hospital-based cohort of Asians. Based on an earlier diagnosis, the participants were divided into the DU (case; n = 1088) and non-DU (control, n = 56,713) groups. The SNP associated with DU risk were obtained from a genome-wide association study (GWAS), and those promoted genetic impact with SNP–SNP interactions were identified with generalized multifactor dimensionality reduction analysis. The interaction between polygenic risk score (PRS) calculated from the selected genetic variants and nutrient were examined. They were related to actin modification, immune response, and cell migration by modulating leucine-rich repeats (LRR) domain binding, Shaffer interferon regulatory factor 4 (IRF4) targets in myeloma vs. mature B lymphocyte, and Reactome runt-related transcription factor 3 (RUNX3). Among the selected SNPs, rs11230563 (R225W) showed missense mutation and low binding affinity with different food components in the autodock analysis. Glycyrrhizin, physalin B, janthitrem F, and casuarinin lowered it in only wild CD6 protein but not in mutated CD6. Plastoquinone 8, solamargine, saponin D, and matesaponin 2 decreased energy binding affinity in mutated CD6 proteins. The PRS of the 5-SNP and 6-SNP models exhibited a positive association with DU risk (OR = 3.14). The PRS of the 5-SNP PRS model interacted with irregular eating habits and smoking status. In participants with irregular eating habits or smokers, DU incidence was much higher in the participants with high PRS than in those with low PRS. In conclusion, the genetic impact of DU risk was mainly in regulating immunity, inflammation, and actin modification. Adults who are genetically susceptible to DU need to eat regularly and to be non-smokers. The results could be applied to personalize nutrition.


Introduction
A peptic ulcer is a lesion found in the stomach and duodenum. It shows a denuded mucosa that extends through the muscularis mucosa into the deeper layers of the wall, such as the submucosa or muscularis propria. According to the anatomical location, they may be termed gastric (GU) or duodenal ulcers (DU) [1]. The most common type of peptic ulcer is DU. Their prevalence is reported to be 0.1-0.19% worldwide and is higher among the Asian population [1]. Although the figures vary in different studies, the prevalence of GU and DU is 3.3% and 2.1%, respectively, in Korea [1]. A peptic ulcer can occur at all ages, but it is more common in middle-aged adults [1]. The symptoms of peptic ulcers are varied, with mild to severe epigastric pain usually felt in the upper abdomen just below  The polygenic risk score (PRS) for the optimal genetic model was generated by adding the number of the risk alleles in each genetic variant with the selected genetic variantgenetic variant interaction model. If the risk allele of a genetic variant was G, the genetic scores of "GG", "AG", and "AA" were 2, 1, and 0, respectively. Among the models to meet the sign test of TEBA and CVC, the optimal model was selected by comparing the adjusted odds ratio (OR). The PRS of the optimal model was stratified into three categories viz. Low-PRS (PRS < 4), Middle-PRS (PRS = 4 and 5), and High-PRS (PRS > 5). Adjusted OR and 95% CI of DU risk were calculated with the PRS groups, and the best model was selected, which had a smaller number of SNPs and a bigger adjusted OR.

Molecular Docking of Food Compounds and Targets of Genes Related to DU
Among the corresponding genes of the selected 10 genetic variants, genes with missense genetic variants were used for changing binding affinity with food components. The changing binding affinity indicated that the catalytic activity of the genes would be changed. The wild-type protein structures of the genes with missense genetic variants were generated in the Protein Data Bank (PDB) format from the Iterative Threading Assembly Refinement (I-TASSER) website. The mutated gene was made by changing the amino acid corresponding to the genetic variants in the Swiss-PBD Viewer (SPDBV) program. Primary ligands were isolated from water molecules and hetero-macromolecules, such as whole protein, using the pleomorphic analysis methodology (PyMOL) software (DeLano Scientific LLC, USA) [27]. The PDB structure was converted into a PBD, partial charge (Q), and atom type (T; PDBQT; AutoDock format) lattice format using AutoDock Tools 1.5.6 (Molecular Graphics Laboratory, The Scripps Research Institute, Jupiter, FL, USA) [27]. The active sites in the protein were searched using the proteins plus website (https://proteins.plus/, accessed on 3 March 2022), and the active functional pockets with mutated sites were selected for molecular docking.
Food compounds (n = 20,000) were downloaded from the fooDB website (https: //foodb.ca/, accessed on 4 January 2022) and made into PDBQT files using AutoDock Tools 1.5.6. The active pockets, including the mutated site, acted as receptors for molecular docking with the food components. After each compound was docked, the compounds Nutrients 2023, 15, 296 6 of 18 with an energy binding affinity of less than −10 were chosen [28]. The lower binding affinity indicated better binding affinity with the active site.

Molecular Dynamics Simulation (MDS)
MDS was used to examine the conformational changes of the protein structure with a variant relative to the native conformation. It also detected changes in protein phenotypes to confirm the devastating consequences of computationally predicted disease-associated mutations. After the top docking poses of all investigated compounds were added, the simulations were performed on pure protein receptors and docked complexes separately by the DS software. The Chemistry at Harvard Macromolecular Mechanics (CHARMM) force field was added to each molecular structure prepared by "Simulation", and the protein was solvated by "Solvation".
The "Standard Dynamics Cascade" set the molecular dynamics simulation parameters for the protein added to the solvent system. The parameters were set to default values, except that the ramp-up time, equilibration time, simulation sampling time, and simulation step size were set to 40 ps, 400 ps, 10,000 ps, and 2 fs, respectively. The root mean square deviation (RMSD), root mean square fluctuations (RMSF), and hydrogen bond values were analyzed after the 10 ns simulation.

Statistical Analysis
The statistical analysis was conducted using SAS (version 9.3; SAS Institute, Cary, NC, USA) [29]. The 58,701 participants were the appropriate sample size to examine the significant differences of the genetic variants for the disease with about 2% prevalence at a statistical significance at α = 0.05, β = 0.99, and an odds ratio of 1.2 in the logistic regression analysis using a G-power calculator. The results of categorical variables were analyzed to show the frequency distributions, and their statistical differences between the DU and the non-DU groups were assessed using the chi-square test. The adjusted means with standard deviations of the continuous variables were calculated with the adjustment of covariates. The statistical differences between the genders and the DU/non-DU groups were determined using a two-way analysis of covariance (ANCOVA) [29]. If ANCOVA was significant at p < 0.05, multiple comparisons were conducted using Tukey's test.
The adjusted odds ratios (ORs) and 95% confidence intervals (CI) of DU were calculated in each metabolic parameter using logistic regression analysis after adjustment for covariates. Two different models with different covariates were applied to calculate adjusted ORs: covariates of model 1 included age, gender, survey year, residence area, body mass index (BMI), income, and education, and those of model 2 were the covariates of model 1 plus energy intake, alcohol and coffee consumption, physical activity, and smoking status.
Two-way ANCOVA with main effects and interaction terms was applied to analyze the interactions between DU and lifestyles after each lifestyle was categorized into the low or high groups according to the 30th percentiles of each variable or the dietary reference intake [30]. Adjusted ORs and 95% CI of DU with PRS were also determined using adjusted logistic regression analysis with covariates in the low and high groups of each lifestylerelated parameter. The proportion of DU patients was calculated according to the PRS groups using the χ 2 test in the low and high groups of each lifestyle-related parameter.

General Characteristics According to Their Gender and DU
The participants with DU were older than those without DU, and men had more DU than women (p < 0.0001; Table 1). The BMI and waist circumferences were lower in the DU group than in the non-ulcer group in both genders and were inversely associated with DU risk (p < 0.001). Serum glucose concentrations were lower in the DU than in the non-ulcer groups in both genders but not the HbA1c levels ( Table 1). The WBC count was higher in men and the DU group and was inversely associated with DU (Table 1). However, there was no significant difference in serum hs-CRP concentrations between the non-DU and DU groups. The incidence of metabolic syndrome did not differ between the groups. The incidence of immune-related diseases such as bronchitis, asthma, arthritis, allergy, gastritis, and periodontitis was higher in the DU group than in the non-ulcer group, regardless of gender. The risk of DU was 2.06-3.34 times higher in patients with these immune-related diseases. Interestingly, the risk of osteoporosis is 1.93 times higher in those with DU.  1 Values represented adjusted means ± standard error for continuous variables and the number and percentages of participants for categorical variables. 2 Adjusted odds ratio (OR) and 95% confidence intervals (CI). Covariates included age, sex, education, income, energy intake (percentage of estimated energy requirement), residence areas, daily activity, alcohol intake, and smoking status. The cutoff points of the reference for logistic regression are as follows: 3 <55 years; 4 <25 kg/m 2 for body mass index (BMI); 5 <90 cm for men and 85 cm for women waist circumferences; 6 <126 mL/dL fasting serum glucose plus diabetic drug intake; 7 <6.5% hbA1c plus diabetic drug intake; 8 <4.0 × 10 9 /L white blood cells (WBC); 9 <1.0 mg/dL high-sensitive C-reactive protein (hs-CRP). *** Significant differences by genders at p < 0.001. ++ Significant differences by DU at p < 0.01, +++ p < 0.001. # Significant interaction between genders and DU at p < 0.05. ⁑⁑ Significantly different from the control group in χ 2 test in each gender at p < 0.01, ⁑⁑⁑ at p < 0.001.a,b,c,d Different superscripts on bars indicate significant differences among the groups at p < 0.05.
63.9 50.9 Adjusted ORs (95% CI) Non-Ulcer (n = 19,725) DU (n = 540) Non-Ulcer (n = 37,810) DU (n = 548) Age (years) 3 57.1 ± 0.09 b1 58.3 ± 0.42 a 52.3 ± 0.06 c 53.4 ± 0.45 c***+++ 1. 40  Blood HbA1c (%) 7 5.68 ± 0.01 b 5.56 ± 0.05 b 5.72 ± 0.01 a 5.78 ± 0.05 a***# 0.720 (0.506-1.025) WBC (10 9 /L) 8 5.76 ± 0.02 b 5.84 ± 0.09 a 5.67 ± 0.01 c 5.47 ± 0.09 d*** 0.753 (0.657-0.863) Serum hs-CRP (mg/dL) 9 0.14 ± 0.004 0.13 ± 0.02 0.14 ± 0.002 0. 13  Values represented adjusted means ± standard error for continuous variables and the number and percentages of participants for categorical variables. 2 Adjusted odds ratio (OR) and 95% confidence intervals (CI). Covariates included age, sex, education, income, energy intake (percentage of estimated energy requirement), residence areas, daily activity, alcohol intake, and smoking status. The cutoff points of the reference for logistic regression are as follows: 3 <55 years; 4 <25 kg/m 2 for body mass index (BMI); 5 <90 cm for men and 85 cm for women waist circumferences; 6 <126 mL/dL fasting serum glucose plus diabetic drug intake; 7 <6.5% hbA1c plus diabetic drug intake; 8 <4.0 × 10 9 /L white blood cells (WBC); 9 <1.0 mg/dL high-sensitive C-reactive protein (hs-CRP). *** Significant differences by genders at p < 0.001. ++ Significant differences by DU at p < 0.01, +++ p < 0.001. # Significant interaction between genders and DU at p < 0.05. ⁑⁑ Significantly different from the control group in χ 2 test in each gender at p < 0.01, ⁑⁑⁑ at p < 0.001.a,b,c,d Different superscripts on bars indicate significant differences among the groups at p < 0.05. than women (p < 0.0001; Table 1). The BMI and waist circumferences were lower in the DU group than in the non-ulcer group in both genders and were inversely associated with DU risk (p < 0.001). Serum glucose concentrations were lower in the DU than in the nonulcer groups in both genders but not the HbA1c levels (Table 1). The WBC count was higher in men and the DU group and was inversely associated with DU (Table 1). However, there was no significant difference in serum hs-CRP concentrations between the non-DU and DU groups. The incidence of metabolic syndrome did not differ between the groups. The incidence of immune-related diseases such as bronchitis, asthma, arthritis, allergy, gastritis, and periodontitis was higher in the DU group than in the non-ulcer group, regardless of gender. The risk of DU was 2.06-3.34 times higher in patients with these immune-related diseases. Interestingly, the risk of osteoporosis is 1.93 times higher in those with DU. than women (p < 0.0001; Table 1). The BMI and waist circumferences were lower in the DU group than in the non-ulcer group in both genders and were inversely associated with DU risk (p < 0.001). Serum glucose concentrations were lower in the DU than in the nonulcer groups in both genders but not the HbA1c levels (Table 1). The WBC count was higher in men and the DU group and was inversely associated with DU (Table 1). However, there was no significant difference in serum hs-CRP concentrations between the non-DU and DU groups. The incidence of metabolic syndrome did not differ between the groups. The incidence of immune-related diseases such as bronchitis, asthma, arthritis, allergy, gastritis, and periodontitis was higher in the DU group than in the non-ulcer group, regardless of gender. The risk of DU was 2.06-3.34 times higher in patients with these immune-related diseases. Interestingly, the risk of osteoporosis is 1.93 times higher in those with DU. Values represented adjusted means ± standard error for continuous variables and the number and percentages of participants for categorical variables. 2 Adjusted odds ratio (OR) and 95% confidence intervals (CI). Covariates included age, sex, education, income, energy intake (percentage of estimated energy requirement), residence areas, daily activity, alcohol intake, and smoking status. The cutoff points of the reference for logistic regression are as follows: 3 <55 years; 4 <25 kg/m 2 for body mass index (BMI); 5 <90 cm for men and 85 cm for women waist circumferences; 6 <126 mL/dL fasting serum glucose plus diabetic drug intake; 7 <6.5% hbA1c plus diabetic drug intake; 8 <4.0 × 10 9 /L white blood cells (WBC); 9 <1.0 mg/dL high-sensitive C-reactive protein (hs-CRP). *** Significant differences by genders at p < 0.001. ++ Significant differences by DU at p < 0.01, +++ p < 0.001. # Significant interaction between genders and DU at p < 0.05. ⁑⁑ Significantly different from the control group in χ 2 test in each gender at p < 0.01, ⁑⁑⁑ at p < 0.001.a,b,c,d Different superscripts on bars indicate significant differences among the groups at p < 0.05. The participants with DU were older than those without DU, and men had more DU than women (p < 0.0001; Table 1). The BMI and waist circumferences were lower in the DU group than in the non-ulcer group in both genders and were inversely associated with DU risk (p < 0.001). Serum glucose concentrations were lower in the DU than in the nonulcer groups in both genders but not the HbA1c levels (Table 1). The WBC count was higher in men and the DU group and was inversely associated with DU (Table 1). However, there was no significant difference in serum hs-CRP concentrations between the non-DU and DU groups. The incidence of metabolic syndrome did not differ between the groups. The incidence of immune-related diseases such as bronchitis, asthma, arthritis, allergy, gastritis, and periodontitis was higher in the DU group than in the non-ulcer group, regardless of gender. The risk of DU was 2.06-3.34 times higher in patients with these immune-related diseases. Interestingly, the risk of osteoporosis is 1.93 times higher in those with DU. Values represented adjusted means ± standard error for continuous variables and the number and percentages of participants for categorical variables. 2 Adjusted odds ratio (OR) and 95% confidence intervals (CI). Covariates included age, sex, education, income, energy intake (percentage of estimated energy requirement), residence areas, daily activity, alcohol intake, and smoking status. The cutoff points of the reference for logistic regression are as follows: 3 <55 years; 4 <25 kg/m 2 for body mass index (BMI); 5 <90 cm for men and 85 cm for women waist circumferences; 6 <126 mL/dL fasting serum glucose plus diabetic drug intake; 7 <6.5% hbA1c plus diabetic drug intake; 8 <4.0 × 10 9 /L white blood cells (WBC); 9 <1.0 mg/dL high-sensitive C-reactive protein (hs-CRP). *** Significant differences by genders at p < 0.001. ++ Significant differences by DU at p < 0.01, +++ p < 0.001. # Significant interaction between genders and DU at p < 0.05. ⁑⁑ Significantly different from the control group in χ 2 test in each gender at p < 0.01, ⁑⁑⁑ at p < 0.001.a,b,c,d Different superscripts on bars indicate significant differences among the groups at p < 0.05. The participants with DU were older than those without DU, and men had more DU than women (p < 0.0001; Table 1). The BMI and waist circumferences were lower in the DU group than in the non-ulcer group in both genders and were inversely associated with DU risk (p < 0.001). Serum glucose concentrations were lower in the DU than in the nonulcer groups in both genders but not the HbA1c levels ( Table 1). The WBC count was higher in men and the DU group and was inversely associated with DU (Table 1). However, there was no significant difference in serum hs-CRP concentrations between the non-DU and DU groups. The incidence of metabolic syndrome did not differ between the groups. The incidence of immune-related diseases such as bronchitis, asthma, arthritis, allergy, gastritis, and periodontitis was higher in the DU group than in the non-ulcer group, regardless of gender. The risk of DU was 2.06-3.34 times higher in patients with these immune-related diseases. Interestingly, the risk of osteoporosis is 1.93 times higher in those with DU. Values represented adjusted means ± standard error for continuous variables and the number and percentages of participants for categorical variables. 2 Adjusted odds ratio (OR) and 95% confidence intervals (CI). Covariates included age, sex, education, income, energy intake (percentage of estimated energy requirement), residence areas, daily activity, alcohol intake, and smoking status. The cutoff points of the reference for logistic regression are as follows: 3 <55 years; 4 <25 kg/m 2 for body mass index (BMI); 5 <90 cm for men and 85 cm for women waist circumferences; 6 <126 mL/dL fasting serum glucose plus diabetic drug intake; 7 <6.5% hbA1c plus diabetic drug intake; 8 <4.0 × 10 9 /L white blood cells (WBC); 9 <1.0 mg/dL high-sensitive C-reactive protein (hs-CRP). *** Significant differences by genders at p < 0.001. ++ Significant differences by DU at p < 0.01, +++ p < 0.001. # Significant interaction between genders and DU at p < 0.05. ⁑⁑ Significantly different from the control group in χ 2 test in each gender at p < 0.01, ⁑⁑⁑ at p < 0.001.a,b,c,d Different superscripts on bars indicate significant differences among the groups at p < 0.05. higher in men and the DU group and was inversely associated with DU (Table 1). However, there was no significant difference in serum hs-CRP concentrations between the non-DU and DU groups. The incidence of metabolic syndrome did not differ between the groups. The incidence of immune-related diseases such as bronchitis, asthma, arthritis, allergy, gastritis, and periodontitis was higher in the DU group than in the non-ulcer group, regardless of gender. The risk of DU was 2.06-3.34 times higher in patients with these immune-related diseases. Interestingly, the risk of osteoporosis is 1.93 times higher in those with DU. Values represented adjusted means ± standard error for continuous variables and the number and percentages of participants for categorical variables. 2 Adjusted odds ratio (OR) and 95% confidence intervals (CI). Covariates included age, sex, education, income, energy intake (percentage of estimated energy requirement), residence areas, daily activity, alcohol intake, and smoking status. The cutoff points of the reference for logistic regression are as follows: 3 <55 years; 4 <25 kg/m 2 for body mass index (BMI); 5 <90 cm for men and 85 cm for women waist circumferences; 6 <126 mL/dL fasting serum glucose plus diabetic drug intake; 7 <6.5% hbA1c plus diabetic drug intake; 8 <4.0 × 10 9 /L white blood cells (WBC); 9 <1.0 mg/dL high-sensitive C-reactive protein (hs-CRP). *** Significant differences by genders at p < 0.001. ++ Significant differences by DU at p < 0.01, +++ p < 0.001. # Significant interaction between genders and DU at p < 0.05. ⁑⁑ Significantly different from the control group in χ 2 test in each gender at p < 0.01, ⁑⁑⁑ at p < 0.001.a,b,c,d Different superscripts on bars indicate significant differences among the groups at p < 0.05. 3886 (10.9) 110 (21.5) higher in men and the DU group and was inversely associated with DU (Table 1). However, there was no significant difference in serum hs-CRP concentrations between the non-DU and DU groups. The incidence of metabolic syndrome did not differ between the groups. The incidence of immune-related diseases such as bronchitis, asthma, arthritis, allergy, gastritis, and periodontitis was higher in the DU group than in the non-ulcer group, regardless of gender. The risk of DU was 2.06-3.34 times higher in patients with these immune-related diseases. Interestingly, the risk of osteoporosis is 1.93 times higher in those with DU. Values represented adjusted means ± standard error for continuous variables and the number and percentages of participants for categorical variables. 2 Adjusted odds ratio (OR) and 95% confidence intervals (CI). Covariates included age, sex, education, income, energy intake (percentage of estimated energy requirement), residence areas, daily activity, alcohol intake, and smoking status. The cutoff points of the reference for logistic regression are as follows: 3 <55 years; 4 <25 kg/m 2 for body mass index (BMI); 5 <90 cm for men and 85 cm for women waist circumferences; 6 <126 mL/dL fasting serum glucose plus diabetic drug intake; 7 <6.5% hbA1c plus diabetic drug intake; 8 <4.0 × 10 9 /L white blood cells (WBC); 9 <1.0 mg/dL high-sensitive C-reactive protein (hs-CRP). *** Significant differences by genders at p < 0.001. ++ Significant differences by DU at p < 0.01, +++ p < 0.001. # Significant interaction between genders and DU at p < 0.05. ⁑⁑ Significantly different from the control group in χ 2 test in each gender at p < 0.01, ⁑⁑⁑ at p < 0.001.a,b,c,d Different superscripts on bars indicate significant differences among the groups at p < 0.05. ulcer groups in both genders but not the HbA1c levels ( Table 1). The WBC count was higher in men and the DU group and was inversely associated with DU (Table 1). However, there was no significant difference in serum hs-CRP concentrations between the non-DU and DU groups. The incidence of metabolic syndrome did not differ between the groups. The incidence of immune-related diseases such as bronchitis, asthma, arthritis, allergy, gastritis, and periodontitis was higher in the DU group than in the non-ulcer group, regardless of gender. The risk of DU was 2.06-3.34 times higher in patients with these immune-related diseases. Interestingly, the risk of osteoporosis is 1.93 times higher in those with DU. Values represented adjusted means ± standard error for continuous variables and the number and percentages of participants for categorical variables. 2 Adjusted odds ratio (OR) and 95% confidence intervals (CI). Covariates included age, sex, education, income, energy intake (percentage of estimated energy requirement), residence areas, daily activity, alcohol intake, and smoking status. The cutoff points of the reference for logistic regression are as follows: 3 <55 years; 4 <25 kg/m 2 for body mass index (BMI); 5 <90 cm for men and 85 cm for women waist circumferences; 6 <126 mL/dL fasting serum glucose plus diabetic drug intake; 7 <6.5% hbA1c plus diabetic drug intake; 8 <4.0 × 10 9 /L white blood cells (WBC); 9 <1.0 mg/dL high-sensitive C-reactive protein (hs-CRP). *** Significant differences by genders at p < 0.001. ++ Significant differences by DU at p < 0.01, +++ p < 0.001. # Significant interaction between genders and DU at p < 0.05. ⁑⁑ Significantly different from the control group in χ 2 test in each gender at p < 0.01, ⁑⁑⁑ at p < 0.001.a,b,c,d Different superscripts on bars indicate significant differences among the groups at p < 0.05. ulcer groups in both genders but not the HbA1c levels ( Table 1). The WBC count was higher in men and the DU group and was inversely associated with DU (Table 1). However, there was no significant difference in serum hs-CRP concentrations between the non-DU and DU groups. The incidence of metabolic syndrome did not differ between the groups. The incidence of immune-related diseases such as bronchitis, asthma, arthritis, allergy, gastritis, and periodontitis was higher in the DU group than in the non-ulcer group, regardless of gender. The risk of DU was 2.06-3.34 times higher in patients with these immune-related diseases. Interestingly, the risk of osteoporosis is 1.93 times higher in those with DU. Values represented adjusted means ± standard error for continuous variables and the number and percentages of participants for categorical variables. 2 Adjusted odds ratio (OR) and 95% confidence intervals (CI). Covariates included age, sex, education, income, energy intake (percentage of estimated energy requirement), residence areas, daily activity, alcohol intake, and smoking status. The cutoff points of the reference for logistic regression are as follows: 3 <55 years; 4 <25 kg/m 2 for body mass index (BMI); 5 <90 cm for men and 85 cm for women waist circumferences; 6 <126 mL/dL fasting serum glucose plus diabetic drug intake; 7 <6.5% hbA1c plus diabetic drug intake; 8 <4.0 × 10 9 /L white blood cells (WBC); 9 <1.0 mg/dL high-sensitive C-reactive protein (hs-CRP). *** Significant differences by genders at p < 0.001. ++ Significant differences by DU at p < 0.01, +++ p < 0.001. # Significant interaction between genders and DU at p < 0.05. ⁑⁑ Significantly different from the control group in χ 2 test in each gender at p < 0.01, ⁑⁑⁑ at p < 0.001.a,b,c,d Different superscripts on bars indicate significant differences among the groups at p < 0.05. DU risk (p < 0.001). Serum glucose concentrations were lower in the DU than in the nonulcer groups in both genders but not the HbA1c levels ( Table 1). The WBC count was higher in men and the DU group and was inversely associated with DU (Table 1). However, there was no significant difference in serum hs-CRP concentrations between the non-DU and DU groups. The incidence of metabolic syndrome did not differ between the groups. The incidence of immune-related diseases such as bronchitis, asthma, arthritis, allergy, gastritis, and periodontitis was higher in the DU group than in the non-ulcer group, regardless of gender. The risk of DU was 2.06-3.34 times higher in patients with these immune-related diseases. Interestingly, the risk of osteoporosis is 1.93 times higher in those with DU. Values represented adjusted means ± standard error for continuous variables and the number and percentages of participants for categorical variables. 2 Adjusted odds ratio (OR) and 95% confidence intervals (CI). Covariates included age, sex, education, income, energy intake (percentage of estimated energy requirement), residence areas, daily activity, alcohol intake, and smoking status. The cutoff points of the reference for logistic regression are as follows: 3 <55 years; 4 <25 kg/m 2 for body mass index (BMI); 5 <90 cm for men and 85 cm for women waist circumferences; 6 <126 mL/dL fasting serum glucose plus diabetic drug intake; 7 <6.5% hbA1c plus diabetic drug intake; 8 <4.0 × 10 9 /L white blood cells (WBC); 9 <1.0 mg/dL high-sensitive C-reactive protein (hs-CRP). *** Significant differences by genders at p < 0.001. ++ Significant differences by DU at p < 0.01, +++ p < 0.001. # Significant interaction between genders and DU at p < 0.05. ⁑⁑ Significantly different from the control group in χ 2 test in each gender at p < 0.01, ⁑⁑⁑ at p < 0.001.a,b,c,d Different superscripts on bars indicate significant differences among the groups at p < 0.05. 3720 (10.5) 152 (29.8) DU risk (p < 0.001). Serum glucose concentrations were lower in the DU than in the nonulcer groups in both genders but not the HbA1c levels ( Table 1). The WBC count was higher in men and the DU group and was inversely associated with DU (Table 1). However, there was no significant difference in serum hs-CRP concentrations between the non-DU and DU groups. The incidence of metabolic syndrome did not differ between the groups. The incidence of immune-related diseases such as bronchitis, asthma, arthritis, allergy, gastritis, and periodontitis was higher in the DU group than in the non-ulcer group, regardless of gender. The risk of DU was 2.06-3.34 times higher in patients with these immune-related diseases. Interestingly, the risk of osteoporosis is 1.93 times higher in those with DU. Values represented adjusted means ± standard error for continuous variables and the number and percentages of participants for categorical variables. 2 Adjusted odds ratio (OR) and 95% confidence intervals (CI). Covariates included age, sex, education, income, energy intake (percentage of estimated energy requirement), residence areas, daily activity, alcohol intake, and smoking status. The cutoff points of the reference for logistic regression are as follows: 3 <55 years; 4 <25 kg/m 2 for body mass index (BMI); 5 <90 cm for men and 85 cm for women waist circumferences; 6 <126 mL/dL fasting serum glucose plus diabetic drug intake; 7 <6.5% hbA1c plus diabetic drug intake; 8 <4.0 × 10 9 /L white blood cells (WBC); 9 <1.0 mg/dL high-sensitive C-reactive protein (hs-CRP). *** Significant differences by genders at p < 0.001. ++ Significant differences by DU at p < 0.01, +++ p < 0.001. # Significant interaction between genders and DU at p < 0.05. ⁑⁑ Significantly different from the control group in χ 2 test in each gender at p < 0.01, ⁑⁑⁑ at p < 0.001.a,b,c,d Different superscripts on bars indicate significant differences among the groups at p < 0.05. DU group than in the non-ulcer group in both genders and were inversely associated with DU risk (p < 0.001). Serum glucose concentrations were lower in the DU than in the nonulcer groups in both genders but not the HbA1c levels ( Table 1). The WBC count was higher in men and the DU group and was inversely associated with DU (Table 1). However, there was no significant difference in serum hs-CRP concentrations between the non-DU and DU groups. The incidence of metabolic syndrome did not differ between the groups. The incidence of immune-related diseases such as bronchitis, asthma, arthritis, allergy, gastritis, and periodontitis was higher in the DU group than in the non-ulcer group, regardless of gender. The risk of DU was 2.06-3.34 times higher in patients with these immune-related diseases. Interestingly, the risk of osteoporosis is 1.93 times higher in those with DU. Values represented adjusted means ± standard error for continuous variables and the number and percentages of participants for categorical variables. 2 Adjusted odds ratio (OR) and 95% confidence intervals (CI). Covariates included age, sex, education, income, energy intake (percentage of estimated energy requirement), residence areas, daily activity, alcohol intake, and smoking status. The cutoff points of the reference for logistic regression are as follows: 3 <55 years; 4 <25 kg/m 2 for body mass index (BMI); 5 <90 cm for men and 85 cm for women waist circumferences; 6 <126 mL/dL fasting serum glucose plus diabetic drug intake; 7 <6.5% hbA1c plus diabetic drug intake; 8 <4.0 × 10 9 /L white blood cells (WBC); 9 <1.0 mg/dL high-sensitive C-reactive protein (hs-CRP). *** Significant differences by genders at p < 0.001. ++ Significant differences by DU at p < 0.01, +++ p < 0.001. # Significant interaction between genders and DU at p < 0.05. ⁑⁑ Significantly different from the control group in χ 2 test in each gender at p < 0.01, ⁑⁑⁑ at p < 0.001.a,b,c,d Different superscripts on bars indicate significant differences among the groups at p < 0.05. 2221 (6. 25) 68 (13.3) DU group than in the non-ulcer group in both genders and were inversely associated with DU risk (p < 0.001). Serum glucose concentrations were lower in the DU than in the nonulcer groups in both genders but not the HbA1c levels ( Table 1). The WBC count was higher in men and the DU group and was inversely associated with DU (Table 1). However, there was no significant difference in serum hs-CRP concentrations between the non-DU and DU groups. The incidence of metabolic syndrome did not differ between the groups. The incidence of immune-related diseases such as bronchitis, asthma, arthritis, allergy, gastritis, and periodontitis was higher in the DU group than in the non-ulcer group, regardless of gender. The risk of DU was 2.06-3.34 times higher in patients with these immune-related diseases. Interestingly, the risk of osteoporosis is 1.93 times higher in those with DU. Values represented adjusted means ± standard error for continuous variables and the number and percentages of participants for categorical variables. 2 Adjusted odds ratio (OR) and 95% confidence intervals (CI). Covariates included age, sex, education, income, energy intake (percentage of estimated energy requirement), residence areas, daily activity, alcohol intake, and smoking status. The cutoff points of the reference for logistic regression are as follows: 3 <55 years; 4 <25 kg/m 2 for body mass index (BMI); 5 <90 cm for men and 85 cm for women waist circumferences; 6 <126 mL/dL fasting serum glucose plus diabetic drug intake; 7 <6.5% hbA1c plus diabetic drug intake; 8 <4.0 × 10 9 /L white blood cells (WBC); 9 <1.0 mg/dL high-sensitive C-reactive protein (hs-CRP). *** Significant differences by genders at p < 0.001. ++ Significant differences by DU at p < 0.01, +++ p < 0.001. # Significant interaction between genders and DU at p < 0.05. ⁑⁑ Significantly different from the control group in χ 2 test in each gender at p < 0.01, ⁑⁑⁑ at p < 0.001.a,b,c,d Different superscripts on bars indicate significant differences among the groups at p < 0.05.

General Characteristics According to Their Gender and DU
The participants with DU were older than those without DU, and men had more DU than women (p < 0.0001; Table 1). The BMI and waist circumferences were lower in the DU group than in the non-ulcer group in both genders and were inversely associated with DU risk (p < 0.001). Serum glucose concentrations were lower in the DU than in the nonulcer groups in both genders but not the HbA1c levels (Table 1). The WBC count was higher in men and the DU group and was inversely associated with DU (Table 1). However, there was no significant difference in serum hs-CRP concentrations between the non-DU and DU groups. The incidence of metabolic syndrome did not differ between the groups. The incidence of immune-related diseases such as bronchitis, asthma, arthritis, allergy, gastritis, and periodontitis was higher in the DU group than in the non-ulcer group, regardless of gender. The risk of DU was 2.06-3.34 times higher in patients with these immune-related diseases. Interestingly, the risk of osteoporosis is 1.93 times higher in those with DU. Values represented adjusted means ± standard error for continuous variables and the number and percentages of participants for categorical variables. 2 Adjusted odds ratio (OR) and 95% confidence intervals (CI). Covariates included age, sex, education, income, energy intake (percentage of estimated energy requirement), residence areas, daily activity, alcohol intake, and smoking status. The cutoff points of the reference for logistic regression are as follows: 3 <55 years; 4 <25 kg/m 2 for body mass index (BMI); 5 <90 cm for men and 85 cm for women waist circumferences; 6 <126 mL/dL fasting serum glucose plus diabetic drug intake; 7 <6.5% hbA1c plus diabetic drug intake; 8 <4.0 × 10 9 /L white blood cells (WBC); 9 <1.0 mg/dL high-sensitive C-reactive protein (hs-CRP). *** Significant differences by genders at p < 0.001. ++ Significant differences by DU at p < 0.01, +++ p < 0.001. # Significant interaction between genders and DU at p < 0.05. ⁑⁑ Significantly different from the control group in χ 2 test in each gender at p < 0.01, ⁑⁑⁑ at p < 0.001.a,b,c,d Different superscripts on bars indicate significant differences among the groups at p < 0.05. than women (p < 0.0001; Table 1). The BMI and waist circumferences were lower in the DU group than in the non-ulcer group in both genders and were inversely associated with DU risk (p < 0.001). Serum glucose concentrations were lower in the DU than in the nonulcer groups in both genders but not the HbA1c levels ( Table 1). The WBC count was higher in men and the DU group and was inversely associated with DU (Table 1). However, there was no significant difference in serum hs-CRP concentrations between the non-DU and DU groups. The incidence of metabolic syndrome did not differ between the groups. The incidence of immune-related diseases such as bronchitis, asthma, arthritis, allergy, gastritis, and periodontitis was higher in the DU group than in the non-ulcer group, regardless of gender. The risk of DU was 2.06-3.34 times higher in patients with these immune-related diseases. Interestingly, the risk of osteoporosis is 1.93 times higher in those with DU. Values represented adjusted means ± standard error for continuous variables and the number and percentages of participants for categorical variables. 2 Adjusted odds ratio (OR) and 95% confidence intervals (CI). Covariates included age, sex, education, income, energy intake (percentage of estimated energy requirement), residence areas, daily activity, alcohol intake, and smoking status. The cutoff points of the reference for logistic regression are as follows: 3 <55 years; 4 <25 kg/m 2 for body mass index (BMI); 5 <90 cm for men and 85 cm for women waist circumferences; 6 <126 mL/dL fasting serum glucose plus diabetic drug intake; 7 <6.5% hbA1c plus diabetic drug intake; 8 <4.0 × 10 9 /L white blood cells (WBC); 9 <1.0 mg/dL high-sensitive C-reactive protein (hs-CRP). *** Significant differences by genders at p < 0.001. ++ Significant differences by DU at p < 0.01, +++ p < 0.001. # Significant interaction between genders and DU at p < 0.05. ⁑⁑ Significantly different from the control group in χ 2 test in each gender at p < 0.01, ⁑⁑⁑ at p < 0.001.a,b,c,d Different superscripts on bars indicate significant differences among the groups at p < 0.05.

General Characteristics According to Their Gender and DU
The participants with DU were older than those without DU, and men had more DU than women (p < 0.0001; Table 1). The BMI and waist circumferences were lower in the DU group than in the non-ulcer group in both genders and were inversely associated with DU risk (p < 0.001). Serum glucose concentrations were lower in the DU than in the nonulcer groups in both genders but not the HbA1c levels ( Table 1). The WBC count was higher in men and the DU group and was inversely associated with DU (Table 1). However, there was no significant difference in serum hs-CRP concentrations between the non-DU and DU groups. The incidence of metabolic syndrome did not differ between the groups. The incidence of immune-related diseases such as bronchitis, asthma, arthritis, allergy, gastritis, and periodontitis was higher in the DU group than in the non-ulcer group, regardless of gender. The risk of DU was 2.06-3.34 times higher in patients with these immune-related diseases. Interestingly, the risk of osteoporosis is 1.93 times higher in those with DU. Values represented adjusted means ± standard error for continuous variables and the number and percentages of participants for categorical variables. 2 Adjusted odds ratio (OR) and 95% confidence intervals (CI). Covariates included age, sex, education, income, energy intake (percentage of estimated energy requirement), residence areas, daily activity, alcohol intake, and smoking status. The cutoff points of the reference for logistic regression are as follows: 3 <55 years; 4 <25 kg/m 2 for body mass index (BMI); 5 <90 cm for men and 85 cm for women waist circumferences; 6 <126 mL/dL fasting serum glucose plus diabetic drug intake; 7 <6.5% hbA1c plus diabetic drug intake; 8 <4.0 × 10 9 /L white blood cells (WBC); 9 <1.0 mg/dL high-sensitive C-reactive protein (hs-CRP). *** Significant differences by genders at p < 0.001. ++ Significant differences by DU at p < 0.01, +++ p < 0.001. # Significant interaction between genders and DU at p < 0.05. ⁑⁑ Significantly different from the control group in χ 2 test in each gender at p < 0.01, ⁑⁑⁑ at p < 0.001.a,b,c,d Different superscripts on bars indicate significant differences among the groups at p < 0.05.

General Characteristics According to Their Gender and DU
The participants with DU were older than those without DU, and men had more DU than women (p < 0.0001; Table 1). The BMI and waist circumferences were lower in the DU group than in the non-ulcer group in both genders and were inversely associated with DU risk (p < 0.001). Serum glucose concentrations were lower in the DU than in the nonulcer groups in both genders but not the HbA1c levels ( Table 1). The WBC count was higher in men and the DU group and was inversely associated with DU (Table 1). However, there was no significant difference in serum hs-CRP concentrations between the non-DU and DU groups. The incidence of metabolic syndrome did not differ between the groups. The incidence of immune-related diseases such as bronchitis, asthma, arthritis, allergy, gastritis, and periodontitis was higher in the DU group than in the non-ulcer group, regardless of gender. The risk of DU was 2.06-3.34 times higher in patients with these immune-related diseases. Interestingly, the risk of osteoporosis is 1.93 times higher in those with DU. Values represented adjusted means ± standard error for continuous variables and the number and percentages of participants for categorical variables. 2 Adjusted odds ratio (OR) and 95% confidence intervals (CI). Covariates included age, sex, education, income, energy intake (percentage of estimated energy requirement), residence areas, daily activity, alcohol intake, and smoking status. The cutoff points of the reference for logistic regression are as follows: 3 <55 years; 4 <25 kg/m 2 for body mass index (BMI); 5 <90 cm for men and 85 cm for women waist circumferences; 6 <126 mL/dL fasting serum glucose plus diabetic drug intake; 7 <6.5% hbA1c plus diabetic drug intake; 8 <4.0 × 10 9 /L white blood cells (WBC); 9 <1.0 mg/dL high-sensitive C-reactive protein (hs-CRP). *** Significant differences by genders at p < 0.001. ++ Significant differences by DU at p < 0.01, +++ p < 0.001. # Significant interaction between genders and DU at p < 0.05. ⁑⁑ Significantly different from the control group in χ 2 test in each gender at p < 0.01, ⁑⁑⁑ at p < 0.001.a,b,c,d Different superscripts on bars indicate significant differences among the groups at p < 0.05. at p < 0.001. a, b, c, d Different superscripts on bars indicate significant differences among the groups at p < 0.05.

Dietary Intake and Lifestyles According to Gender and DU
Energy intake was not significantly different between the non-DU and DU groups ( Table 2). Among the four dietary patterns, the participants with DU had a lower proportion of WSD intake than those without ulcers, and the WSD was inversely associated with DU. The participants consuming meat cooked less showed a higher incidence of DU than those who did not. Those consuming lower amounts of cooked meats were at a 1.24 times higher risk of DU (Table 2). However, the intake of burnt meats and fried food was not associated with DU risk. The participants with DU had a lower coffee intake than those without DU, and coffee intake was inversely related to DU risk. However, there was no difference in the tea and alcohol intakes between the DU and non-DU groups (Table 2). Interestingly, the number of participants taking multivitamins was lower in the DU group than in the non-DU group, and taking multivitamin showed an inverse association with DU risk ( Table 2). Physical activity was not associated with DU risk. As expected, among men, non-smokers had a lower incidence of DU than former and current smokers, and smoking status exhibited a positive association with DU risk (Table 2). in those with DU.  2 Adjusted odds ratio (OR) and 95% confidence intervals (CI). Covariates included age, sex, education, income, energy intake (percentage of estimated energy requirement), residence areas, daily activity, alcohol intake, and smoking status. The cutoff points of the reference for logistic regression are as follows: 3 <55 years; 4 <25 kg/m 2 for body mass index (BMI); 5 <90 cm for men and 85 cm for women waist circumferences; 6 <126 mL/dL fasting serum glucose plus diabetic drug intake; 7 <6.5% hbA1c plus diabetic drug intake; 8 <4.0 × 10 9 /L white blood cells (WBC); 9 <1.0 mg/dL high-sensitive C-reactive protein (hs-CRP). *** Significant differences by genders at p < 0.001. ++ Significant differences by DU at p < 0.01, +++ p < 0.001. # Significant interaction between genders and DU at p < 0.05. ⁑⁑ Significantly different from the control group in χ 2 test in each gender at p < 0.01, ⁑⁑⁑ at p < 0.001.a,b,c,d Different superscripts on bars indicate significant differences among the groups at p < 0.05. higher risk of DU (Table 2). However, the intake of burnt meats and fried food was not associated with DU risk. The participants with DU had a lower coffee intake than those without DU, and coffee intake was inversely related to DU risk. However, there was no difference in the tea and alcohol intakes between the DU and non-DU groups (Table 2). Interestingly, the number of participants taking multivitamins was lower in the DU group than in the non-DU group, and taking multivitamin showed an inverse association with DU risk (Table 2). Physical activity was not associated with DU risk. As expected, among men, non-smokers had a lower incidence of DU than former and current smokers, and smoking status exhibited a positive association with DU risk (Table 2).  6 3.65 ± 0.03 a 3.28 ± 0.14 b 3.69 ± 0.02 a 3.03 ± 0.13 b+++ 0.648 (0.567-0.740) Tea (g/day) 7 45 43.9 ± 0.81 40.5 ± 3.89 42.9 ± 0.53 45.9 ± 3.84 1.115 (0.959-1.297) Alcohol (g/day) 8 30 DU. The participants consuming meat cooked less showed a higher incidence of DU than those who did not. Those consuming lower amounts of cooked meats were at a 1.24 times higher risk of DU (Table 2). However, the intake of burnt meats and fried food was not associated with DU risk. The participants with DU had a lower coffee intake than those without DU, and coffee intake was inversely related to DU risk. However, there was no difference in the tea and alcohol intakes between the DU and non-DU groups (Table 2). Interestingly, the number of participants taking multivitamins was lower in the DU group than in the non-DU group, and taking multivitamin showed an inverse association with DU risk (Table 2). Physical activity was not associated with DU risk. As expected, among men, non-smokers had a lower incidence of DU than former and current smokers, and smoking status exhibited a positive association with DU risk (Table 2). Values represented adjusted means ± standard error for continuous variables and the number and percentages of participants for categorical variables. 2 Adjusted odds ratio (OR) and 95% confidence intervals. Covariates included age, sex, education, income, energy intake (percentage of estimated energy requirement), residence areas, daily activity, alcohol intake, and smoking status. The cutoff points of the reference for logistic regression are as follows: 3 < Estimated energy requirement (EER); 4 <25th percentiles of each dietary pattern; 5 < twice/week; 6 <3 g/day; 7 <45 g/day; 8 <20 g/day. *** Significant differences by genders at p < 0.001. +++ Significant differences by DU at p < 0.001. ⁑ Significantly different from the control group in χ 2 test in each gender at p < 0.05, ⁑⁑ at p < 0.01, and ⁑⁑⁑ at p < 0.001. a,b Different superscripts on bars indicate significant differences among the groups at p < 0.05. ( Table 2). Among the four dietary patterns, the participants with DU had a lower proportion of WSD intake than those without ulcers, and the WSD was inversely associated with DU. The participants consuming meat cooked less showed a higher incidence of DU than those who did not. Those consuming lower amounts of cooked meats were at a 1.24 times higher risk of DU (Table 2). However, the intake of burnt meats and fried food was not associated with DU risk. The participants with DU had a lower coffee intake than those without DU, and coffee intake was inversely related to DU risk. However, there was no difference in the tea and alcohol intakes between the DU and non-DU groups (Table 2). Interestingly, the number of participants taking multivitamins was lower in the DU group than in the non-DU group, and taking multivitamin showed an inverse association with DU risk (Table 2). Physical activity was not associated with DU risk. As expected, among men, non-smokers had a lower incidence of DU than former and current smokers, and smoking status exhibited a positive association with DU risk (Table 2). Values represented adjusted means ± standard error for continuous variables and the number and percentages of participants for categorical variables. 2 Adjusted odds ratio (OR) and 95% confidence intervals. Covariates included age, sex, education, income, energy intake (percentage of estimated energy requirement), residence areas, daily activity, alcohol intake, and smoking status. The cutoff points of the reference for logistic regression are as follows: 3 < Estimated energy requirement (EER); 4 <25th percentiles of each dietary pattern; 5 < twice/week; 6 <3 g/day; 7 <45 g/day; 8 <20 g/day. *** Significant differences by genders at p < 0.001. +++ Significant differences by DU at p < 0.001. ⁑ Significantly different from the control group in χ 2 test in each gender at p < 0.05, ⁑⁑ at p < 0.01, and ⁑⁑⁑ at p < 0.001. a,b Different superscripts on bars indicate significant differences among the groups at p < 0.05. 18,788 (53.0) 302 (59.1) allergy, gastritis, and periodontitis was higher in the DU group than in the non-ulcer group, regardless of gender. The risk of DU was 2.06-3.34 times higher in patients with these immune-related diseases. Interestingly, the risk of osteoporosis is 1.93 times higher in those with DU. Values represented adjusted means ± standard error for continuous variables and the number and percentages of participants for categorical variables. 2 Adjusted odds ratio (OR) and 95% confidence intervals (CI). Covariates included age, sex, education, income, energy intake (percentage of estimated energy requirement), residence areas, daily activity, alcohol intake, and smoking status. The cutoff points of the reference for logistic regression are as follows: 3 <55 years; 4 <25 kg/m 2 for body mass index (BMI); 5 <90 cm for men and 85 cm for women waist circumferences; 6 <126 mL/dL fasting serum glucose plus diabetic drug intake; 7 <6.5% hbA1c plus diabetic drug intake; 8 <4.0 × 10 9 /L white blood cells (WBC); 9 <1.0 mg/dL high-sensitive C-reactive protein (hs-CRP). *** Significant differences by genders at p < 0.001. ++ Significant differences by DU at p < 0.01, +++ p < 0.001. # Significant interaction between genders and DU at p < 0.05. ⁑⁑ Significantly different from the control group in χ 2 test in each gender at p < 0.01, ⁑⁑⁑ at p < 0.001.a,b,c,d Different superscripts on bars indicate significant differences among the groups at p < 0.05.  6 3.65 ± 0.03 a 3.28 ± 0.14 b 3.69 ± 0.02 a 3.03 ± 0.13 b +++ 0.648 (0.567-0.740) Tea (g/day) 7 45 43 . DU risk (p < 0.001). Serum glucose concentrations were lower in the DU than in the nonulcer groups in both genders but not the HbA1c levels ( Table 1). The WBC count was higher in men and the DU group and was inversely associated with DU (Table 1). However, there was no significant difference in serum hs-CRP concentrations between the non-DU and DU groups. The incidence of metabolic syndrome did not differ between the groups. The incidence of immune-related diseases such as bronchitis, asthma, arthritis, allergy, gastritis, and periodontitis was higher in the DU group than in the non-ulcer group, regardless of gender. The risk of DU was 2.06-3.34 times higher in patients with these immune-related diseases. Interestingly, the risk of osteoporosis is 1.93 times higher in those with DU. Values represented adjusted means ± standard error for continuous variables and the number and percentages of participants for categorical variables. 2 Adjusted odds ratio (OR) and 95% confidence intervals (CI). Covariates included age, sex, education, income, energy intake (percentage of estimated energy requirement), residence areas, daily activity, alcohol intake, and smoking status. The cutoff points of the reference for logistic regression are as follows: 3 <55 years; 4 <25 kg/m 2 for body mass index (BMI); 5 <90 cm for men and 85 cm for women waist circumferences; 6 <126 mL/dL fasting serum glucose plus diabetic drug intake; 7 <6.5% hbA1c plus diabetic drug intake; 8 <4.0 × 10 9 /L white blood cells (WBC); 9 <1.0 mg/dL high-sensitive C-reactive protein (hs-CRP). *** Significant differences by genders at p < 0.001. ++ Significant differences by DU at p < 0.01, +++ p < 0.001. # Significant interaction between genders and DU at p < 0.05. ⁑⁑ Significantly different from the control group in χ 2 test in each gender at p < 0.01, ⁑⁑⁑ at p < 0.001.a,b,c,d Different superscripts on bars indicate significant differences among the groups at p < 0.05. 29 DU risk (p < 0.001). Serum glucose concentrations were lower in the DU than in the nonulcer groups in both genders but not the HbA1c levels ( Table 1). The WBC count was higher in men and the DU group and was inversely associated with DU (Table 1). However, there was no significant difference in serum hs-CRP concentrations between the non-DU and DU groups. The incidence of metabolic syndrome did not differ between the groups. The incidence of immune-related diseases such as bronchitis, asthma, arthritis, allergy, gastritis, and periodontitis was higher in the DU group than in the non-ulcer group, regardless of gender. The risk of DU was 2.06-3.34 times higher in patients with these immune-related diseases. Interestingly, the risk of osteoporosis is 1.93 times higher in those with DU. Values represented adjusted means ± standard error for continuous variables and the number and percentages of participants for categorical variables. 2 Adjusted odds ratio (OR) and 95% confidence intervals (CI). Covariates included age, sex, education, income, energy intake (percentage of estimated energy requirement), residence areas, daily activity, alcohol intake, and smoking status. The cutoff points of the reference for logistic regression are as follows: 3 <55 years; 4 <25 kg/m 2 for body mass index (BMI); 5 <90 cm for men and 85 cm for women waist circumferences; 6 <126 mL/dL fasting serum glucose plus diabetic drug intake; 7 <6.5% hbA1c plus diabetic drug intake; 8 <4.0 × 10 9 /L white blood cells (WBC); 9 <1.0 mg/dL high-sensitive C-reactive protein (hs-CRP). *** Significant differences by genders at p < 0.001. ++ Significant differences by DU at p < 0.01, +++ p < 0.001. # Significant interaction between genders and DU at p < 0.05. ⁑⁑ Significantly different from the control group in χ 2 test in each gender at p < 0.01, ⁑⁑⁑ at p < 0.001.a,b,c,d Different superscripts on bars indicate significant differences among the groups at p < 0.05. ever, there was no significant difference in serum hs-CRP concentrations between the non-DU and DU groups. The incidence of metabolic syndrome did not differ between the groups. The incidence of immune-related diseases such as bronchitis, asthma, arthritis, allergy, gastritis, and periodontitis was higher in the DU group than in the non-ulcer group, regardless of gender. The risk of DU was 2.06-3.34 times higher in patients with these immune-related diseases. Interestingly, the risk of osteoporosis is 1.93 times higher in those with DU. Values represented adjusted means ± standard error for continuous variables and the number and percentages of participants for categorical variables. 2 Adjusted odds ratio (OR) and 95% confidence intervals (CI). Covariates included age, sex, education, income, energy intake (percentage of estimated energy requirement), residence areas, daily activity, alcohol intake, and smoking status. The cutoff points of the reference for logistic regression are as follows: 3 <55 years; 4 <25 kg/m 2 for body mass index (BMI); 5 <90 cm for men and 85 cm for women waist circumferences; 6 <126 mL/dL fasting serum glucose plus diabetic drug intake; 7 <6.5% hbA1c plus diabetic drug intake; 8 <4.0 × 10 9 /L white blood cells (WBC); 9 <1.0 mg/dL high-sensitive C-reactive protein (hs-CRP). *** Significant differences by genders at p < 0.001. ++ Significant differences by DU at p < 0.01, +++ p < 0.001. # Significant interaction between genders and DU at p < 0.05. ⁑⁑ Significantly different from the control group in χ 2 test in each gender at p < 0.01, ⁑⁑⁑ at p < 0.001.a,b,c,d Different superscripts on bars indicate significant differences among the groups at p < 0.05. 1 Values represented adjusted means ± standard error for continuous variables and the number and percentages of participants for categorical variables. 2 Adjusted odds ratio (OR) and 95% confidence intervals. Covariates included age, sex, education, income, energy intake (percentage of estimated energy requirement), residence areas, daily activity, alcohol intake, and smoking status. The cutoff points of the reference for logistic regression are as follows: 3 <Estimated energy requirement (EER); 4 <25th percentiles of each dietary pattern; 5 <twice/week; 6 <3 g/day; 7 <45 g/day; 8 <20 g/day. *** Significant differences by genders at p < 0.001. +++ Significant differences by DU at p < 0.001.

of 19
Dietary Intake and Lifestyles According to Gender and DU Energy intake was not significantly different between the non-DU and DU groups le 2). Among the four dietary patterns, the participants with DU had a lower proporof WSD intake than those without ulcers, and the WSD was inversely associated with The participants consuming meat cooked less showed a higher incidence of DU than e who did not. Those consuming lower amounts of cooked meats were at a 1.24 times er risk of DU (Table 2). However, the intake of burnt meats and fried food was not ciated with DU risk. The participants with DU had a lower coffee intake than those out DU, and coffee intake was inversely related to DU risk. However, there was no rence in the tea and alcohol intakes between the DU and non-DU groups (Table 2). restingly, the number of participants taking multivitamins was lower in the DU group in the non-DU group, and taking multivitamin showed an inverse association with risk ( Table 2). Physical activity was not associated with DU risk. As expected, among , non-smokers had a lower incidence of DU than former and current smokers, and king status exhibited a positive association with DU risk (Table 2).  ues represented adjusted means ± standard error for continuous variables and the number and entages of participants for categorical variables. 2 Adjusted odds ratio (OR) and 95% confidence vals. Covariates included age, sex, education, income, energy intake (percentage of estimated gy requirement), residence areas, daily activity, alcohol intake, and smoking status. The cutoff ts of the reference for logistic regression are as follows: 3 < Estimated energy requirement (EER); th percentiles of each dietary pattern; 5 < twice/week; 6 <3 g/day; 7 <45 g/day; 8 <20 g/day. *** Sigant differences by genders at p < 0.001. +++ Significant differences by DU at p < 0.001. ⁑ Signifiy different from the control group in χ 2 test in each gender at p < 0.05, ⁑⁑ at p < 0.01, and ⁑⁑⁑ at p 01. a,b Different superscripts on bars indicate significant differences among the groups at p < Significantly different from the control group in χ 2 test in each gender at p < 0.05, Nutrients 2023, 15, x FOR PEER REVIEW

General Characteristics According to Their Gender and DU
The participants with DU were older than those without DU, and than women (p < 0.0001; Table 1). The BMI and waist circumferences DU group than in the non-ulcer group in both genders and were invers DU risk (p < 0.001). Serum glucose concentrations were lower in the D ulcer groups in both genders but not the HbA1c levels (Table 1). Th higher in men and the DU group and was inversely associated with D ever, there was no significant difference in serum hs-CRP concentration DU and DU groups. The incidence of metabolic syndrome did not groups. The incidence of immune-related diseases such as bronchitis allergy, gastritis, and periodontitis was higher in the DU group tha group, regardless of gender. The risk of DU was 2.06-3.34 times high these immune-related diseases. Interestingly, the risk of osteoporosis in those with DU. Values represented adjusted means ± standard error for continuous variables percentages of participants for categorical variables. 2 Adjusted odds ratio (OR intervals (CI). Covariates included age, sex, education, income, energy intak mated energy requirement), residence areas, daily activity, alcohol intake, and cutoff points of the reference for logistic regression are as follows: 3 <55 years mass index (BMI); 5 <90 cm for men and 85 cm for women waist circumferences serum glucose plus diabetic drug intake; 7 <6.5% hbA1c plus diabetic drug inta blood cells (WBC); 9 <1.0 mg/dL high-sensitive C-reactive protein (hs-CRP). *** S by genders at p < 0.001. ++ Significant differences by DU at p < 0.01, +++ p < 0.001 tion between genders and DU at p < 0.05. ⁑⁑ Significantly different from the c in each gender at p < 0.01, ⁑⁑⁑ at p < 0.001.a,b,c,d Different superscripts on ba differences among the groups at p < 0.05. at p < 0.01, and al Characteristics According to Their Gender and DU articipants with DU were older than those without DU, and men had more DU en (p < 0.0001; Table 1). The BMI and waist circumferences were lower in the than in the non-ulcer group in both genders and were inversely associated with < 0.001). Serum glucose concentrations were lower in the DU than in the nonps in both genders but not the HbA1c levels (Table 1). The WBC count was men and the DU group and was inversely associated with DU (Table 1). Howe was no significant difference in serum hs-CRP concentrations between the non-U groups. The incidence of metabolic syndrome did not differ between the he incidence of immune-related diseases such as bronchitis, asthma, arthritis, astritis, and periodontitis was higher in the DU group than in the non-ulcer gardless of gender. The risk of DU was 2.06-3.34 times higher in patients with une-related diseases. Interestingly, the risk of osteoporosis is 1.93 times higher ith DU. presented adjusted means ± standard error for continuous variables and the number and s of participants for categorical variables. 2 Adjusted odds ratio (OR) and 95% confidence I). Covariates included age, sex, education, income, energy intake (percentage of estirgy requirement), residence areas, daily activity, alcohol intake, and smoking status. The ts of the reference for logistic regression are as follows: 3 <55 years; 4 <25 kg/m 2 for body (BMI); 5 <90 cm for men and 85 cm for women waist circumferences; 6 <126 mL/dL fasting ose plus diabetic drug intake; 7 <6.5% hbA1c plus diabetic drug intake; 8 <4.0 × 10 9 /L white (WBC); 9 <1.0 mg/dL high-sensitive C-reactive protein (hs-CRP). *** Significant differences at p < 0.001. ++ Significant differences by DU at p < 0.01, +++ p < 0.001. # Significant interacen genders and DU at p < 0.05. ⁑⁑ Significantly different from the control group in χ 2 test der at p < 0.01, ⁑⁑⁑ at p < 0.001.a,b,c,d Different superscripts on bars indicate significant among the groups at p < 0.05. at p < 0.001. a, b Different superscripts on bars indicate significant differences among the groups at p < 0.05.

Characteristics of Polygenic Variants Involved in DU Risk
The Manhattan plot showed the statistical significance of each polymorphism with DU risk in the GWAS. The red lines indicated p < 5 × 10 −8 , which was applied with the Bonferroni correction. However, we used a statistical significance of p < 5 × 10 −6 since p < 5 × 10 −8 was conservative, and many genetic variants showed LD relations. Significant genetic variants existed in chromosomes 2, 5, 6, 8, 11, 13, and 15 ( Figure 2A). The Q-Q plot showed the deviation of the observed p values for genetic variants from their expected p values for DU risk. The lambda value representing the genome inflation factor was 1.034, indicating no inflation of the genetic variants associated with DU risk ( Figure 2B).

Characteristics of Polygenic Variants Involved in DU Risk
The Manhattan plot showed the statistical significance of each polymorphism with DU risk in the GWAS. The red lines indicated p < 5 × 10 −8 , which was applied with the Bonferroni correction. However, we used a statistical significance of p < 5 × 10 −6 since p < 5 × 10 −8 was conservative, and many genetic variants showed LD relations. Significant genetic variants existed in chromosomes 2, 5, 6, 8, 11, 13, and 15 (Figure 2A). The Q-Q plot showed the deviation of the observed p values for genetic variants from their expected p values for DU risk. The lambda value representing the genome inflation factor was 1.034, indicating no inflation of the genetic variants associated with DU risk ( Figure 2B). The genetic characteristics of DU-related genetic variants with SNP-SNP interaction are presented in Table 3. They were satisfied with p < 5 × 10 −6 for GWAS with DU, ≥0.01 in MAF, D' < 0.2 in LD, and p ≥ 0.05 in HWE. Ten genetic variants were listed in Table 3 as follows: rs576376935_C-X-C Motif Chemokine Receptor 2 (CXCR2), rs77063016_fragile histidine triad diadenosine triphosphatase (FHIT), rs10055925_tetratricopeptide repeat domain 33 (TTC33), rs2978977_PSCA, rs6584283_LINC01475, rs11230563_cluster of differentiation 6 (CD6) (R225W), rs7309887_inositol 1,4,5-trisphosphate receptor type 2 (ITPR2), rs78141015_ DNAJ heat shock protein family (DNAJC15), rs111690253_lipopolysaccharide-induced TNF factor (LITAF), and rs796980537_fucosyltransferase 2 (FUT2). The rs10055925_TTC33, rs2978977_PSCA, rs11230563_CD6(R225W), and rs7309887_ITPR2 (OR = 0.637-0.813) variants were inversely associated with DU risk, while the other six genetic variants (OR = 1.17 − 1.95) were positively associated with DU risk. The rs11230563_CD6 (R225W) variant was located in the gene transcript, while others were in the intron.

Pathways of DU Risk-Related Genetic Variants
The gene ontology (GO) terms and curated gene sets related to the genetic variants associated with DU risk were searched with a MAGMA gene-set analysis. Table 4 displays the pathways involved in the genetic variants associated with DU risk. The biological process in the GO, the actin modification pathway, was associated with the genetic variants associated with DU risk. The potential genes related to actin modification were CXCR2, FHIT, CD6, ITPR2, and DNAJC15. In the GO of molecular function, the leucine-rich repeats (LRR) domain binding pathway exhibited an association with the genetic variants linked to DU risk. Its related potential genes were CD6, FUT2, LITAF, and FHIT. The pathway of Shaffer interferon regulatory factor 4 (IRF4) targets in myeloma vs. mature B lymphocytes was associated with DU risk, potentially with the genetic variants of CXCR2, CD6, ITPR2, and FUT2. Furthermore, the pathway of the Reactome runt-related transcription factor 3 (runx3) regulates immune response and cell migration, which was linked to DU risk, possibly with the genetic variants of CXCR2, FHIT, CD6, ITPR2, and DNAJC15.
The PRS of models 5 and 6 exhibited a positive association with DU risk by about three times after adjusting with two sets of covariates ( Figure 3). The covariates of set 1 included age, gender, BMI, residence area, education, and income. Those of set 2 contained energy intake, alcohol intake, smoking status, and physical exercise plus covariate set 1. These results suggested that the five-SNP model accounted for DU risk in both covariate sets. However, PRS was not associated with other immune-related and metabolic parameters such as WBC, hs-CRP, BMI, and waist circumference and disease conditions such as metabolic syndrome, bronchitis, asthma, arthritis, allergy, osteoporosis, gastritis, and periodontitis (Supplementary Table S2). such as metabolic syndrome, bronchitis, asthma, arthritis, allergy, osteoporosis, gastritis, and periodontitis (Supplementary Table S2).

Energy Binding Affinity with Food Components and the Foods Containing the Food Components
Food components with lower binding affinity improved the wild and mutated CD6 protein activity. Foods containing high levels of these food components are presented in

Energy Binding Affinity with Food Components and the Foods Containing the Food Components
Food components with lower binding affinity improved the wild and mutated CD6 protein activity. Foods containing high levels of these food components are presented in Table 5. Azaspiracid 2 showed low binding affinity (−13.2 kcal/mol) in both wild and mutated types of CD6 protein (Table 5). However, glycyrrhizin, physalin B, janthitrem F, and casuarinin lowered binding energy only in the wild CD6 protein but not in the mutated CD6 at R225W (Table 5). Plastoquinone 8, solamargine, saponin D, and matesaponin 2 decreased energy binding affinity in mutated CD6 proteins at R225W (Table 5). Glycyrrhizin decreased the energy for binding affinity (−11.7 kcal/mol) in wild CD6 proteins at R225W (rs11230563) but not the mutated one ( Figure 4A-F).    RMSD for glycyrrhizin was lower in the wild type of CD6 than in the mutated type ( Figure 5A), although it was about 3 nm and within 2.9-3.5 nm over a period of 100 s in both wild and mutated types of CD6. RMSF for glycyrrhizin fluctuated within 0.5-4.3 nm in the wild type and 0.6-3.7 nm in the mutated type ( Figure 5B). The results showed that glycyrrhizin had a better binding affinity to the wild type of CD6 than the mutated one. Therefore, people with wild or mutated CD6 proteins should consume appropriate foods to decrease energy-binding affinity. The RMSD and RMSF of azaspiracid 2 in wild type and mutated CD6 rs11230563 (R225W) during molecular dynamic simulation ( Figure S1).

Interaction of PRS with Lifestyle Factors Influences DU Risk
DU was associated with lifestyle factors, such as irregular meals, eating raw or burnt meat, WSD, taking multivitamins, and smoking. The interaction between PRS and the lifestyles associated with DU risk was examined. Irregular meals were statistically correlated with DU risk (p = 0.0047; Table 6). The DU incidence showed significant differences in participants with irregular eating habits according to the PRS groups. Participants with irregular meal eating habits in the High-PRS exhibited a positive association with DU risk but not those with regular eating habits ( Figure 6A). The smoking status also interacted with PRS (p = 0.0015; Table 6). DU incidence in the Low-PRS group was lower in current smokers than in non-smokers and former smokers. The PRS showed a positive relation

Interaction of PRS with Lifestyle Factors Influences DU Risk
DU was associated with lifestyle factors, such as irregular meals, eating raw or burnt meat, WSD, taking multivitamins, and smoking. The interaction between PRS and the lifestyles associated with DU risk was examined. Irregular meals were statistically correlated with DU risk (p = 0.0047; Table 6). The DU incidence showed significant differences in participants with irregular eating habits according to the PRS groups. Participants with irregular meal eating habits in the High-PRS exhibited a positive association with DU risk but not those with regular eating habits ( Figure 6A). The smoking status also interacted with PRS (p = 0.0015; Table 6). DU incidence in the Low-PRS group was lower in current smokers than in non-smokers and former smokers. The PRS showed a positive relation with DU risk in smokers and non-smokers, while the positive association was greater in smokers than in non-smokers (Table 6). In participants taking no multivitamins, the DU incidence was much higher in those with High-PRS than those with Low-PRS and Middle-PRS ( Figure 6B). However, the increased DU incidence in the High-PRS group was lower in the participants taking multivitamin (Table 6). WSD (p = 0.550), eating raw meat (p = 0.623), and eating burnt meat (p = 0.780) did not show an interaction with PRS ( Figure 6C).  The polygenic risk score (PRS) for the best model was generated by summing the number of the risk alleles in each selected genetic variant with the gene-gene interaction model. The PRS of the model was divided into three categories as Low-PRS (PRS < 4), Middle-PRS (PRS = 4 and 5), and High-PRS (PRS > 5). The adjustment with covariates of age, gender, education, income, residence area, energy intake, alcohol intake, regular exercise, and smoking status.

Discussion
The stomach secretes acids and enzymes for digestion and to destroy microorganisms. The gastric mucus secreted by the cells in the stomach wall is a barrier to protect the wall from the acid and digestive enzymes within the stomach lumen. A similar mucosal lining protects the duodenal wall [2]. The mucus itself is made up of macromolecules called mucins. There are several well-known causes of DU. The most common causes of DU include inflammation of the duodenum lining due to H. pylori infection and prostaglandin inhibition by NSAIDs affecting the regeneration and function of the duodenal mucosal cell layer. Therefore, a decrease in immune function and insufficient mucin are believed to be the key inducers of DU [3]. Genetics and the environment regulate immunity and mucin content in the duodenum. Gastric acid secretion increases with smoking, excessive alcohol consumption, irregular food intake, and mental stress [31]. In the present study, we found that genetic variants associated with DU risk were related to regulating immunity and actin modification [31]. The 5-SNP model included TTC33_rs10055925, FUT2_rs796980537, LINC01475_rs6584283, ITPR2_ rs7309887, and PSCA_rs2978977. The PRS of the 5-SNP model was positively related to DU risk. Irregular eating, taking multivitamins, and smoking interacted with PRS to influence DU risk. The rs11230563 (R225W) showed a low binding affinity with different food components in the Autodock analysis. Glycyrrhizin, physalin B, janthitrem F, and casuarinin lowered it only in wild CD6 proteins but not in mutated CD6 proteins. Plastoquinone 8, solamargine, sap- The polygenic risk score (PRS) for the best model was generated by summing the number of the risk alleles in each selected genetic variant with the gene-gene interaction model. The PRS of the model was divided into three categories as Low-PRS (PRS < 4), Middle-PRS (PRS = 4 and 5), and High-PRS (PRS > 5). The adjustment with covariates of age, gender, education, income, residence area, energy intake, alcohol intake, regular exercise, and smoking status.

Discussion
The stomach secretes acids and enzymes for digestion and to destroy microorganisms. The gastric mucus secreted by the cells in the stomach wall is a barrier to protect the wall from the acid and digestive enzymes within the stomach lumen. A similar mucosal lining protects the duodenal wall [2]. The mucus itself is made up of macromolecules called mucins. There are several well-known causes of DU. The most common causes of DU include inflammation of the duodenum lining due to H. pylori infection and prostaglandin inhibition by NSAIDs affecting the regeneration and function of the duodenal mucosal cell layer. Therefore, a decrease in immune function and insufficient mucin are believed to be the key inducers of DU [3]. Genetics and the environment regulate immunity and mucin content in the duodenum. Gastric acid secretion increases with smoking, excessive alcohol consumption, irregular food intake, and mental stress [31]. In the present study, we found that genetic variants associated with DU risk were related to regulating immunity and actin modification [31]. The 5-SNP model included TTC33_rs10055925, FUT2_rs796980537, LINC01475_rs6584283, ITPR2_ rs7309887, and PSCA_rs2978977. The PRS of the 5-SNP model was positively related to DU risk. Irregular eating, taking multivitamins, and smoking interacted with PRS to influence DU risk. The rs11230563 (R225W) showed a low binding affinity with different food components in the Autodock analysis. Glycyrrhizin, physalin B, janthitrem F, and casuarinin lowered it only in wild CD6 proteins but not in mutated CD6 proteins. Plastoquinone 8, solamargine, saponin D, and matesaponin 2 decreased energy binding affinity in mutated CD6 proteins. These results can be used to devise personalized nutrition for DU prevention.
The genetic polymorphisms involved in DU risk were associated with actin modification genes for the development and differentiation of the intestinal cells and host immunity potentially against H. pylori infections [10]. Many studies have implicated H. pylori as a significant etiologic factor in DU. Some studies also link H. pylori infections with gastric cancer. Thus, in adults, a high host immunity can protect against these disease conditions [32]. In the UK Biobank, ABO, CDX2, CCKBR MUC1, MUC6, FUT2, PSCA, and GAST genes have been identified and found to be associated with peptic ulcer disease. These genes are believed to be involved in the susceptibility to H. pylori infection, protection against infection, gastrointestinal motility, or gastric acid secretion [10]. Genetic polymorphisms, including FUT2 and PSCA, are also associated with DU, consistent with the present study. The genetic predisposition of disintegrin and metalloproteases (ADAMs) and Th17-related cytokines, such as IL-17A, IL-17F, IL-33, IL-23, and IL-23R, was associated with DU risk [10,33]. Similar to earlier studies, the genetic variants of IL1RN, IL23R, and IL1R2 were found in the present study, but their statistical significance did not reach p < 5 × 10 −8 . Therefore, poor immunity, potentially to H. pylori, is linked to DU risk.
DU is associated with damage to the duodenal wall due to vigorous immune responses or severe cellular inflammatory responses. The genetic polymorphisms of proinflammatory cytokines, such as TNF-α [34], IL-1β, IL-4, and IL-8, and related to H. pylori infection, have been seen in different ethnicities [35,36]. The present study selected a genetic variant of lipopolysaccharide-induced TNF factor (LITAF) for DU risk among the genetic variants related to proinflammatory cytokines. While some genetic polymorphisms of IL were associated with DU risk, they did not meet the statistical significance of p < 5 × 10 −6 . A few studies have shown the association of IL genetic polymorphisms with DU risk in Koreans [37], which are reported to be associated with gastric cancer in Asian subjects [38,39]. LITAF acts as a mediator of local and systemic inflammatory responses to modulate the increment of TNF-α secretion in inflammatory diseases, including irritable bowel disease [40,41]. Lipopolysaccharides (LPS) are the major outer surface membrane components of H. pylori and act as potent stimulators of the host immune response [41]. An H. pylori infection increases TNF-α expression with LITAF. Therefore, H. pylori infection is associated with DU risk by modulating immunity and LPS-related inflammation.
Prostate stem cell antigen (PSCA), a glycosylphosphatidylinositol-anchored cell membrane glycoprotein, is involved in not only prostate cancer but also gastric cancer and DU in different ethnicities [2,11]. PSCA s2978977 is not linked to GU in Japanese, consistent with the present study [11]. However, the T allele of PSCA rs2978977 has been shown to be associated with gastric cancer (OR = 1.109, p = 8.11 × 10 −10 ) and an inverse association with DU (OR = 0.648, p = 6.86 × 10 −19 ) in the present study, consistent with previous studies [42,43]. The Cox regression analysis has revealed the T allele of rs2294008 as a prognosis factor in patients with diffuse-type gastric cancer (hazard ratio: 1.85; 95% CI: 1.12-3.06). The fragile histidine triad (FHIT) is also linked to protecting against developing H. pylori-related gastric cancer. However, the association between FHIT expression and DU risk may occur through the prevention or eradication of H. pylori infections [44]. ITPR2 also acts as a second messenger to modulate intracellular calcium levels by releasing it from the endoplasmic reticulum [45]. Intracellular calcium plays a critical role in gastric and duodenal wound repair, and its polymorphism influences DU risk [46]. Therefore, ITPR2 polymorphism can alter DU risk, and the A allele of rs7309887 might suppress duodenal cell repair in DU.
Although many studies have studied the relationship between genetic variants and diseases, their direct impacts are small and difficult to observe in small epidemiological studies. Genetic polymorphisms alter gene expressions, protein production, and binding affinity to molecular components such as medications and food. Many genetic polymorphisms are in the intron and promoter regions and alter gene expression. However, the mechanisms by which genetic variants in the intron affect the corresponding gene expression have yet to be revealed in the present study. Missense genetic variants can influence gene function by changing binding affinity to components, including food and drugs [47]. The ligand binding site of rs11230563_CD6 (R225W) missense was selected to find the food components to decrease their binding affinity, which may act as the regulators of the CD6 activity. CD6 is a well-known target protein for regulating immune responses [48]. CD6 negatively and positively influences the initiation and maintenance of T cell function, respectively [48]. Therefore, the mutation in the CD6 can affect binding affinity to food components by changing free energy. The present study showed that the rs11230563 nucleotide site changed binding affinity to a particular molecular site of the CD6 gene. These results indicated that rs11230563 affects the domains 1 and 3 areas to alter binding affinity to food components. People with the C allele of rs11230563 CD6 had a higher chance of inducing DU risk, and the food components to lower the binding affinity of the C allele in rs11230563 in CD6 might improve CD6 activity. Therefore, glycyrrhizin, physalin B, janthitrem F, and casuarinin might improve CD6 activity due to the lower binding energy in the CD6 with the C allele in rs11230563. However, the relationship between binding affinity and CD6 function needs to be validated.
The primary risk factor for DU is an H. pylori infection, and additional ones are being male, family history, blood type, skipping breakfast or more than one meal, coffee intake, consumption of NSAIDs, aspirin, alcohol, coffee, and cigarette smoking [49]. The present study showed consistent results for all of the above except for coffee intake. Among dietary patterns, DU was negatively associated with WSD but not KBD, PBD, and RMD. Irregular meal intake and eating uncooked and burnt meats were positively linked to DU risk. Exercise, alcohol intake, and mental stress have been reported to be related to DU risk [31], but the association was not seen in the present study. Coffee drinking, taking multivitamin, and smoking status were inversely associated with DU risk in the present study. No study to date has reported that lifestyle modifications interact with PRS. In the present study, the PRS associated with DU risk interacted with irregular meal intake, smoking, and multivitamin intake. The results suggested that people with high-PRS better to consume regular meals, take multivitamins, and smoke less to reduce DU risk.
A few studies have studied the genetic polymorphism of DU risk, although most studies have been conducted on peptic ulcers, including both gastric and duodenal ulcers. To the best of our knowledge, it is novel to determine the interaction between duodenal ulcers and lifestyles in a large hospital-based cohort (n = 58,701). The limitations of the study were (1) the cause-and-effect relationship could not be explained since it was a case-control study. (2) The patient history of a physician-diagnosed duodenal ulcer was brought up by recollection from the memory of the participant, not by the medical record.
(3) The usual food intake for the previous six months was assessed using an SQFFQ based on memory. However, the SQFFQ contained 106 food items related to Korean diets and was validated by 3-day food records for four seasons [19].
In conclusion, the PRS of the selected genetic variants for DU risk was mainly associated with actin modification, LRR domain binding, Shaffer IRF4 targets in myeloma vs. mature B lymphocytes, and Reactome runx3 regulated immune response and cell migration. The results suggest that the DU risk was related to modulating immunity, inflammation, and intestinal cell repair. Among the selected genetic variants, rs11230563_CD6 (R225W), missense mutation, was shown to reduce binding affinity with glycyrrhizin, physalin B, janthitrem F, and casuarinin in the T allele, and with plastoquinone 8, solamargine, saponin D, and matesaponin 2 in the C allele of rs11230563 CD6 in the Autodock analysis. The High-PRS of the 5-SNP and 6-SNP models showed a positive association with DU risk by about three times. In participants with irregular eating and rare meat-eating habits, DU incidence was much higher in the participants with High-PRS than those with Low-PRS, as with smokers. These results suggest that the genetic impact of DU risk was primarily involved with mediating immunity, inflammation, and actin modification. DU risk was linked to rare meat, irregular eating habits, and smoking status. The interaction of eating habits with PRS for DU risk and the missense mutations of genetic variants can be applied to devise personalized nutrition.