Combination of Helicobacter pylori infection and the interleukin 8 –251 T > A polymorphism, but not the mannose-binding lectin 2 codon 54 G > A polymorphism, might be a risk factor of gastric cancer

Background Mannose-binding lectin (MBL) acts in the innate immune response to Helicobacter pylori. Interleukin 8 (IL-8) is a potent cytokine produced by gastric epithelial cells in response to H. pylori. We aimed to investigate whether polymorphisms in MBL2 and IL-8 influence susceptibility to H. pylori infection, and the associations of these polymorphisms with the risk of gastroduodenal diseases in a Korean population. Methods We consecutively enrolled 176 H. pylori-negative control subjects, 221 subjects with H. pylori-positive non-atrophic gastritis, 52 mild atrophic gastritis (AG), 61 severe AG, 175 duodenal ulcer, and 283 gastric cancer (GC). Allele-specific PCR-RFLP was conducted for polymorphisms in MBL2 exon 1 (codon 52, 54, and 57) and IL-8 -251 T > A. IL-8 levels in gastric mucosal tissues and serum MBL levels were measured by enzyme-linked immunosorbent assay. Results MBL2 exon 1 polymorphic variants were found only in codon 54, and the allele frequencies did not differ significantly between the control and disease groups. Although serum MBL levels in codon 54 A/A mutants were markedly low, it did not influence susceptibility to H. pylori infection or the risk of gastroduodenal diseases. IL-8 levels were significantly different between T/T wild type, T/A heterozygote, and A/A mutant genotypes. IL-8 -251 A allele carriers (A/A + T/A) showed increased IL-8 levels, and were significantly associated with the risk of severe AG and GC. Conclusions We suggest that a combination of H. pylori infection and the IL-8 -251 T > A polymorphism might increase the risk of severe AG and GC in a Korean population.


Background
The Helicobacter pylori infection rate is about 50% among the worldwide adult population [1]. In Korea, the adult H. pylori infection rate was 66.9% in 1998, 59.6% in 2005, and dropped to 54.4% in 2011 [2]. The main cause of this decrease in infection rate is the improvement in unsanitary environmental conditions. Besides environmental factors, bacterial and host factors are involved in the pathogenesis of H. pylori infection. With regard to bacterial factors, H. pylori strains possessing the virulence factors cagA, vacA s1a/m1, and iceA1 are known to be particularly virulent, and are frequently associated with severe gastric epithelial damage [3]. In contrast to Western populations, the cagA protein is commonly found in Korean patients with gastric cancer (GC) and duodenal ulcer (DU) [4]. However, there have been no associations reported between different H. pylori genotypes and clinical outcome in Korean patients [5,6].
With regard to host factors, host genetic variants may influence susceptibility to H. pylori and the pathogenesis of gastroduodenal diseases. Host factors are mainly related to two processes: recognition of H. pylori by the innate immune system, and the level of the cytokine response [7,8]. Polymorphisms in pro-and antiinflammatory cytokines are associated with the risk of atrophic gastritis (AG) and GC. Interleukin 1 beta (IL-1β), tumor necrosis factor-alpha, IL-6, and IL-8 are upregulated in response to H. pylori infection [9]. Several anti-inflammatory cytokines such as IL-Rα and IL-10 are also related to H. pylori infection [10,11].
The innate immune response to H. pylori infection is a further candidate host factor. Toll-like receptors (TLRs) recognize conserved pathogen-associated molecular patterns expressed by many pathogens, including H. pylori [25]. Mannose-binding lectin (MBL), a pattern recognition receptor encoded by the MBL2 gene, recognizes lipopolysaccharide in the cell wall of gram-negative bacteria such as H. pylori [26,27]. H. pylori activates MBL in vitro, resulting in complement deposition [28,29]. Some studies have found a possible association of MBL2 haplotype with susceptibility to H. pylori infection, as well as with risk of GC [30,31]. However, other studies did not find any significant association between MBL genotype and H. pylori infection prevalence or GC risk [32,33].
Serum MBL levels vary widely between healthy individuals, mainly due to genetic variation [34][35][36]. The variation in serum MBL levels is correlated with point mutations in the coding and promoter regions of MBL2. Three mutations within exon 1 (in codons 52, 54, and 57) interfere with MBL function and are associated with low serum levels of MBL. In African populations, point mutations at codons 52 and 57 occur frequently [36,37]. In Caucasians, mutations at codons 52 and 54 are common [38]. In Chinese, Japanese, and Korean populations, mutations are predominantly common in codon 54, but not in codons 52 or 57 [39][40][41]. Polymorphisms within the promoter and 5′untranslated regions of MBL2 also affect serum levels of MBL, but the effects were found to be lower than those of the exon 1 polymorphisms [41].
The aims of this study were: 1) to examine the influence of the polymorphisms in codons 52, 54, and 57 of MBL2 (related to innate immunity) on susceptibility to H. pylori infection; 2) to evaluate the association of the IL-8 -251 T > A polymorphism with the risk of gastroduodenal diseases in a Korean population; and 3) to analyze our and other investigators' large-scale data regarding the IL-8 -251 T > A polymorphism and GC risk in Korean, Japanese, Chinese, and Caucasian populations.
All participants (n = 855) underwent upper gastrointestinal endoscopy and routine laboratory tests. The controls were asymptomatic subjects who visited the Health Screening Center for a health status check-up, and their endoscopic findings were normal. Exclusion criteria were H. pylori eradication history; use of antibiotics, proton pump inhibitors, nonsteroidal antiinflammatory drugs, or anticoagulant drugs; and severe systemic illnesses. Age, sex, alcohol consumption (current or never), smoking habits (current or never), salt intake (high, low-moderate), and family history of GC (first-degree relatives) were recorded. Informed consent was obtained from all included subjects. The Institutional Review Board of the Kyung Hee University Hospital approved the study protocol (KMC IRB 1523-04).

Diagnosis of H. pylori infection
The rapid urease test (or urea breath test) and serum anti-H. pylori immunoglobulin G antibody test were performed. A subject was defined as H. pylori infectionpositive if both tests were positive. A subject was defined as H. pylori infection-negative if both tests were negative. Subjects with only one positive test were excluded from this study.

Histologic examination of chronic gastritis
One pathologist histologically evaluated chronic gastritis status in biopsy specimens. AG was graded based on the presence and proportion of glandular loss (mild, moderate, and severe) according to the updated Sydney system [42].
Genotyping of MBL2 exon 1 codons 52, 54 and 57, and of IL-8 -251 Genomic DNA was extracted from peripheral venous blood using a genomic DNA purification method. Polymerase chain reaction (PCR) amplification, restriction fragment length polymorphism (RFLP) analysis, and gel electrophoresis were performed for MBL2 (codons 52, 54, and 57 in exon 1) and IL-8 (−251 promoter region) as described previously [7,34]. The PCR product involving codon 52 was digested by incubation with MluI at 37°C for 3 h, resulting in two bands of 204 and 94 bp for the T/T genotype (mutant), three bands of 298, 204, and 94 bp for the A/T genotype (heterozygote), and one band of 298 bp for the A/A genotype (wild type). The PCR product involving codon 54 was digested by BanI at 50°C for 3 h, resulting in two bands of 195 and 103 bp for the G/G genotype (wild type), three bands of 298, 195, and 103 bp for the G/A genotype (heterozygote), and one band of 298 bp for the A/A genotype (mutant). The PCR product involving codon 57 was digested with MboI at 37°C for 3 h, resulting in two bands of 190 and 108 bp for the A/A genotype (mutant), three bands of 298, 190, and 108 bp for the G/A genotype (heterozygote), and one band of 298 bp for the G/G genotype (wild type). For genotyping of the IL-8 -251 T > A polymorphism, PCR products were digested with MfeI at 37°C for 3 h, resulting in two bands of 449 and 92 bp for the A/A genotype (mutant), three bands of 541, 449, and 92 bp for the T/A genotype (heterozygote), and one band of 541 bp for the T/T genotype (wild type).

Measurement of serum MBL levels
MBL is a serum protein produced mainly by hepatocytes, and expressed in immune cells, but not in epithelial cells [43]. Circulatory MBL levels were taken as an indicator of the functional activity of MBL protein. Serum MBL levels were measured after overnight fasting by enzyme-linked immunosorbent assay (ELISA; MBL Oligomer ELISA kit; BioProto Diagnostics, Denmark).

Measurement of IL-8 levels in gastric mucosal tissues
Although measurement of serum IL-8 levels is straightforward, serum IL-8 levels do not reflect the severity of H. pylori-associated gastritis [44]. Therefore, we measured IL-8 levels in gastric mucosal tissues rather than serum IL-8 levels.
Three biopsy specimens were taken from the greater curvature side of the proximal antrum during endoscopic procedures. The specimens were put into a tube with 2.0 mL phosphate-buffered saline (pH 7.4), frozen on dry ice, and stored at −70°C. Samples were homogenized and centrifuged, and the supernatants were aliquoted. Total protein was measured using the bicinchoninic acid assay (Thermo Scientific, Rockford, IL, USA). Gastric mucosal IL-8 levels were measured by ELISA (R&D Systems Inc., Minneapolis, MN, USA). The mucosal level of IL-8 was expressed as picograms per milligram of gastric biopsy protein.

Statistical analysis
Data are expressed as mean values ± standard deviations or as frequencies and percentages. Chi-squared and Kruskal-Wallis tests were performed to compare clinical parameters between the control and disease groups. Hardy-Weinberg equilibrium for polymorphisms in MBL2 and IL-8 was tested using R version 3.1.0 (R Development Core Team). Biases caused by differences in clinical parameters were adjusted using the chisquared and Kruskal-Wallis tests. Multiple logistic regression analysis was performed to evaluate the associations of the genetic polymorphisms with susceptibility to H. pylori infection and the risk of gastroduodenal diseases using the SAS statistical software package version 9.4 (SAS Institute Inc.). All clinical parameters with a p value <0.20 in the univariate analysis were included in the full logistic regression model. The odds ratios (ORs) and their 95% confidence intervals (CIs) were used to compare the risks between the control and disease groups. P values <0.05 were considered statistically significant. Table 1 shows clinical features of the control and disease groups. Age, sex ratio, and alcohol consumption were similar among all groups. Risk factors for GC, such as smoking, high salt intake, and family history of GC, were more frequently observed in the H. pylori-positive GC group than in the control group. The differences were statistically significant.

Results
The frequencies of the MBL2 codon 54 and IL-8 -251 polymorphisms in the control group did not deviate significantly from those expected under Hardy-Weinberg equilibrium (p = 1.000 and p = 0.184, respectively). In this study population, MBL exon 1 polymorphic variants were found only in codon 54. There were no variants at codons 52 or 57; only the wild type was observed.
The frequencies of MBL2 codon 54 and IL-8 -251 genotypes in the control and disease groups The frequencies of MBL2 codon 54 genotypes were similar among the control and disease groups ( Table 2). The frequency of IL-8 -251 A allele carriers was higher in the H. pylori-positive severe AG and H. pylori-positive GC groups than in the control group, but the differences did not reach statistical significance (Table 2).

Association between MBL2 codon 54 G > A polymorphism and the risk of gastroduodenal diseases
We examined the association between the MBL2 codon 54 G > A polymorphism and the risk of gastroduodenal disease using univariate and multivariate logistic regression analysis. We regarded the control group as the reference subject group, and considered G/G (wild type) as the reference genotype. The MBL2 codon 54 G > A polymorphism did not increase susceptibility to H. pylori-positive NAG, mild AG, or severe AG, and also was not associated with the risk of DU and GC (Table 3).

Serum MBL levels
Serum levels of MBL were high in carriers of the G/G (wild type) genotype, intermediate in those with the G/A heterozygous genotype, and low in those with the A/A (mutant) genotype in all subjects (n = 855, Fig. 1). The differences between the three genotypes were highly statistically significant (p < 0.0001). However, there were no significant differences in serum MBL levels between the control (139.9 ± 83.2 ng/mL), H. pylori-positive NAG (149.3 ± 81.2 ng/mL), mild AG (146.9 ± 81.8 ng/mL), severe AG (140.2 ± 87.3 ng/mL), DU (143.8 ± 82.5 ng/mL), and GC (149.8 ± 82.6 ng/mL) groups.

Association between IL-8 -251 T > A polymorphism and the risk of disease development
Because the IL-8 cytokine response is mainly dependent on the H. pylori-associated inflammatory severity, we sub-classified the H. pylori-positive chronic gastritis group (n = 221) into H. pylori-positive NAG, mild AG, and severe AG. We regarded the control group as the reference subject group, and considered T/T (wild type) as the reference genotype. The IL-8 -251 A allele significantly increased the risk of severe AG and GC, as determined by both univariate and multivariate logistic regression analysis ( Table 4).
Comparison of IL-8 levels according to disease phenotypes and IL-8 -251 each genotype IL-8 levels were low in subjects with the T/T (wild type) genotype, intermediate in those with the T/A heterozygous genotype, and high in those with the A/A (mutant) genotype. The differences between the three genotypes were statistically significant (p = 0.0262, Fig. 2).
IL-8 levels were markedly low in H. pylori-negative control (n = 176, 4.43 ± 3.30 pg/mg protein) and H. pylorinegative subjects regardless of any disease phenotypes (n = 83, 5.02 ± 3.80 pg/mg protein) compared with H. pylori-positive subjects (n = 596, 154.05 ± 56.26 pg/mg protein; p < 0.0001). IL-8 levels were very low in the  (Fig. 3).  The control was regarded as the reference subject group, and the G/G wild type was considered as the reference genotype

Analysis of global results of IL-8 -251 T > A polymorphism and GC risk
The Korean population, including the subjects of this study, showed a significant positive association between the IL-8 -251 T > A polymorphism and GC risk. On the contrary, the Chinese and Caucasian populations showed a negative association. The Japanese population was similar to the Korean population. The combined Korean and Japanese populations showed significantly increased GC risk for the IL-8 -251 T/A and A/A genotypes compared to the T/T genotype, and for A allele carriers compared to non-carriers (Table 5).

Discussion
The innate immune response is the first line of defense against H. pylori infection in the human stomach. TLR and MBL are recognized as important proteins in innate immunity. Several studies have demonstrated that TLR4 and TLR2 polymorphisms are associated with the risk of GC [45][46][47]. However, some of the associations are controversial, and there are discrepancies between the results for Asian and Western populations [48]. A recent study in the Netherlands found that only the TLR1 polymorphism is associated with the prevalence of H. pylori seropositivity [49].
Further studies are needed in other populations worldwide to confirm these associations. MBL binds to bacteria, yeasts, and viruses via specific repeated oligosaccharide moieties on the cell surface. MBL activates the complement-lectin pathway, facilitates opsonization and phagocytosis, and induces direct cellular lysis. MBL deficiency or a low serum MBL level has been associated with several infectious and autoimmune diseases, including meningococcal meningitis, pneumonia, arterial thrombosis, systemic lupus erythematosus, and celiac disease [50,51].
At the time of its discovery, H. pylori was considered an extracellular bacterium that mainly colonized the gastric mucus layer or attached to gastric epithelial cells. However, it has since been demonstrated that H. pylori invades the lamina propria and gastric epithelial cells [52]. Therefore, H. pylori might be a target of phagocytosis by MBL activation. There have been few clinical studies regarding the role of MBL in H. pylori infection. Various microorganisms such as H. pylori, Neisseria meningitidis groups B and C, Nocardia farcinica, and Legionella pneumophila induce MBL activity in vitro [28]. Activated complements are found in the epithelium of patients with H. pylori-associated gastritis [29]. One pediatric study reported that MBL2 mRNA expression in gastric biopsy specimens was higher in H. pyloripositive chronic gastritis than in H. pylori-negative chronic gastritis patients [53]. However, the study had two weaknesses in terms of its ability to reach conclusions regarding the role of MBL2 expression in the development of H. pylori-infected chronic gastritis. The first weakness is the small number of biopsy specimens that were obtained, with only five H. pylori-positive children and four control children included. The second weakness is that they could not find any association between MBL2 genotype and the risk of H. pylori-infected chronic gastritis.
The association between the MBL2 haplotype and the risk of GC has been studied previously [30,31]. A study conducted in Southern Italy found that the HYP + D haplotype (H/Y promoter region mutation + P untranslated region mutation + codon 52 mutation) may be a genetic marker for H. pylori-positive GC risk [30]. Another study performed in Warsaw, Poland found that the HY + D haplotype (H/Y promoter region mutation + codon 52 mutation) was related to an increased risk of GC compared with the HY+ A haplotype (H/Y mutation + codon 52 wild type) [31]. Therefore, the codon 52 D variant (cysteine > arginine) was specifically related to the risk of GC in two populations. In contrast to the above studies, which reported positive associations, Australian researchers evaluated healthy individuals for H. pylori infection, MBL2 genotype, mannan binding level, and complement 4 level in plasma, and found that MBL deficiency, defined by either genotype or plasma activity, was not associated with higher susceptibility to H. pylori infection [33]. In a Japanese study, they could no significant differences were found in MBL2 genotypes between GC patients and healthy controls [32]. Instead, the investigators found that the MBL2 codon 54 polymorphism was weakly associated with severe AG and advanced GC [32,54]. In the present study, we first demonstrated that the codon 54 polymorphism did not increase susceptibility to H. pylori infection in a Korean population. Secondly, we did not find any evidence of a role for MBL2 in the development of gastroduodenal diseases. Thirdly, we did not find a higher risk of advanced GC or severe AG The control was regarded as the reference subject group, and the T/T wild type was considered as the reference genotype compared to early GC or mild AG, respectively, associated with MBL2 genotype. With regard to interracial differences, the Korean population differs from the European (Italian and Polish) and African populations reported previously. However, the results for the Korean population are very similar to those reported for the Chinese and Japanese populations [35,[38][39][40]. The frequencies of point mutations in European populations are in between those of the East Asian and African populations.
In the present study, serum levels of MBL, an indicator of the functional activity of MBL, differed significantly according to the genotype. However, serum MBL levels were not significantly different between the control and disease groups, because the frequency of each genotype was similar in these groups.
H. pylori infection stimulates IL-8 gene expression and increases the IL-8 cytokine level in gastric epithelial cells. A significant correlation between a high level of IL-8 in the gastric mucosa and the risk of GC has been reported [13]. Our previous study found that the IL-8 level in gastric mucosal tissues was significantly higher in H. pylori-infected subjects compared with that in H. pylori non-infected subjects, irrespective of their gastroduodenal disease phenotype. After H. pylori eradication, the IL-8 level decreased dramatically, to the same level observed in non-infected subjects [55]. In this study, we confirmed once again that the IL-8 level in gastric mucosal tissues is mainly dependent on H. pylori-positive status.
It has been reported that the IL-8 -251 T > A polymorphism is related to higher levels of IL-8 and to an increased risk of AG, gastric ulcer, and GC [13,14]. In this study, we also demonstrated that the IL-8 -251 T > A polymorphism increased IL-8 production, and was significantly associated with the risk of GC and severe AG. However, many other epidemiological studies have reported negative associations between the IL-8 -251 polymorphism and GC risk (18)(19)(20)(21)(22)(23), and a meta-analysis revealed no overall association (24). In this study, we analyzed large-scale raw data from controls and GC patients from Korean, Japanese, Chinese, and Caucasian (Poland, Finland, and Portugal) populations (13)(14)(15)(16)(17)(18)(19)(20)(21)(22)(23). Korean results, including ours, were consistent with Japanese results, but not with Chinese or Caucasian results. The concordance between the Korean and Japanese results might be explained by genetic similarities. In a large study of single nucleotide polymorphism (SNP) maps covering the human genome performed in African Americans, Asians (Japanese-Chinese-Korean), and European Americans (Caucasians) [56], SNP differences in autosomes were only 5.86% between Korean and Japanese populations. Therefore, the Korean population is very similar to  the Japanese population with respect to the pattern of SNPs [56].

Conclusions
The MBL2 codon 54 G > A polymorphism does not influence susceptibility to H. pylori infection and does not increase the risk of gastroduodenal diseases. We suggest that a combination of the IL-8 -251 T > A polymorphism and increased IL-8 production in response to H. pylori infection may be a risk factor for severe AG and GC development in a Korean population.