Interleukin-1B (IL-1B) polymorphisms and gastric mucosal levels of IL-1β cytokine in Korean patients with gastric cancer

Authors


Abstract

Interleukin-1B and IL-1 receptor antagonist gene polymorphisms are associated with an increased risk of gastric cancer (GC) in Caucasian populations. However, recent studies could not find any association between IL-1B-511T polymorphism and the risk of GC in Asians. We tested for an association between IL-1 loci polymorphisms with increased gastric mucosal levels of IL-1β and an increased risk of developing GC in a Korean population. Polymorphisms of IL-1A-889, IL-1B-31, IL-1B-511 and IL-1RN were genotyped in 434 controls and 234 patients with GC. Mucosal IL-1β cytokine was measured using an ELISA. The frequencies of IL-1A, IL-1B-511, IL-1B-31 and IL-1RN were not statistically different between controls and all patients with GC. After subclassification of GC, only patients with intestinal-type GC showed a higher frequency of IL-1B-31T homozygotes (OR = 2.2; 95% CI = 1.1–4.3) compared with controls. Risk was also significantly increased in these patients for IL-1B-31T homozygotes compared with patients with diffuse-type GC (OR = 3.4; 95% CI = 1.5–7.7). As in Caucasian populations, linkage disequilibrium between IL-1B-31 and IL-1B-511 was nearly complete, but the pattern of haplotype related to the risk of GC (IL-1B-31T/IL-1B-511C) was opposite (IL-1B-511T/IL-1B-31C). Mucosal IL-1β levels in H. pylori-infected GC patients were higher in patients homozygous for IL-1B-31T compared with IL-1B-31C/T and IL-1B-31C/C. Thus, the combined effects of H. pylori infection and IL-1B-31T/IL-1B-511C polymorphisms with enhanced mucosal IL-1β production contributed to the development of intestinal-type GC in this Korean population. © 2004 Wiley-Liss, Inc.

In Korea, gastric cancer (GC) is the most frequently diagnosed malignancy, with an annual incidence rate of 37.1 per 100,000 population.1 About 60% of Korean adults are infected with Helicobacter pylori (H. pylori).2, 3 This is regarded as a leading cause of gastric carcinogenesis and is associated with a 2-fold increased risk of developing gastric adenocarcinomas.4, 5 However, H. pylori infection is associated with divergent clinical outcomes, ranging from simple asymptomatic gastritis to more serious conditions such as peptic ulcer disease and gastric neoplasia. The prevalence of H. pylori infection is high among Asian populations, but the incidence of patients developing GC differs greatly among northern (Korea, Japan and northern China) and southern (Thailand and Vietnam) Asian populations. H. pylori infection is common in Thai and Vietnamese populations, yet the incidence of GC is extremely low (the so-called Asian paradox). Differences in the incidence of GC raise questions regarding differences in environmental factors, such as H. pylori strains, diet and other host factors.

H. pylori strains possessing cagA, vacA s1a/m1 and iceA1 allelotype are infective virulent bacterial factors associated with increased gastric epithelial damage.6, 7 However, the cagA gene is commonly found in H. pylori isolates from Korean patients with GC and duodenal ulcers.8 Furthermore, there was no association between the vacA or iceA status and the clinical outcome in Korean patients.9, 10

In addition to H. pylori infection, host genetic factors may play important roles in the development of GC. Germline truncating mutations in the E-cadherin gene have been found in families with hereditary diffuse GC,11 but these are not specific for sporadic GC. Human leukocyte antigen (HLA)-DQA1, tumor necrosis factor (TNF) and cytochrome P450 2E (CYP2E) genes are also associated with gastric carcinogenesis.12, 13, 14 Polymorphisms of those genes may thus confer some protection against GC. Patients with glutathione S-transferase mu 1 (GSTM1) null and N-acetyltransferase 1 (NAT1)* 10 allele polymorphisms show an increased risk of developing GC because of their low ability to detoxify carcinogens.15, 16 However, studies on the role of genetic factors in GC risk remain inconclusive because of a number of limitations in study designs, and host factors that influence acid secretion are highly likely to influence the outcome of an H. pylori infection. Interleukin-1β (IL-1β, encoded by IL-1B), a proinflammatory cytokine17 and also a potent inhibitor of gastric acid secretion,18 is a prime candidate in this regard. The IL-1 receptor antagonist (IL-1ra, encoded by IL-1RN) is a naturally occurring antiinflammatory cytokine that modulates the potentially injurious effects of IL-1 proteins by competitively binding to the IL-1 receptor.19IL-1RN*2 is associated with enhanced IL-1β production in vitro. Synergistic interaction between IL-1B-511T carriers and IL-1RN*2 genotype increases the risk of developing GC, and production of IL-1β cytokine is increased in the gastric mucosa of patients infected with H. pylori. Two recent epidemiologic studies reported that IL-1B and IL-1RN polymorphisms (IL-1B-511T, IL-1B-31C and IL-1RN*2 alleles) were associated with hypochlorhydria, gastric atrophy and increased risk of developing GC in Caucasian populations.20, 21 In contrast, 2 recent studies could not find any association between the IL-1B-511 polymorphism and GC in a Japanese population and a Taiwanese population.22, 23

The aims of this study were to elucidate the association of IL-1A, IL-1B and IL-1RN loci polymorphisms as host genetic factors with the increased risk of developing GC in a Korean population and to examine the combined effect of IL-1 polymorphisms with subsequent enhanced gastric mucosal levels of IL-1β and H. pylori infection in gastric carcinogenesis.

Material and methods

Subjects

From October 2000 to March 2003, we recruited subjects (control group) who wished to receive a routine health checkup including gastroscopy, a common screening examination for GC in Korea, and evaluation of their H. pylori status. All participants were interviewed on their medical and family histories, followed by a brief physical examination. The control group comprised patients being investigated for dyspepsia and healthy asymptomatic subjects. Subjects with previous gastrectomy, peptic ulcer disease, gastric atrophy and premalignant intestinal metaplasia were excluded. Subjects with systemic diseases, such as diabetes, hypertension and heart disease, were also excluded. We registered both the control group and patients with GC (GC group) who visited the gastroenterology and endoscopy section of Kyung Hee University Hospital, Seoul, Korea. Because Koreans have not significantly married outside their ethnic group for over 2,000 years, we believe our subjects can be considered ethnically homogeneous. Before the examination, the purpose of this study was explained to the participants, and informed consent was obtained from all individuals. A questionnaire involved taking a family history of any previous GC and recording the individual's age, sex, alcohol consumption and smoking habits. GCs were diagnosed using surgical and/or gastroscopic biopsies (Table I). The Human Medical Research Committee of Kyung Hee University approved the protocol.

Table I. Baseline Clinical Characteristics of the Subjects
 Control group (n = 434)Gastric cancer group (n = 234)p
Mean age (range)47.5 (21–72)57.1 (23–75)< 0.001
Male/total (%)249/434 (57.4)149/234 (63.7)NS
Alcohol drinking (%)218 (50.2)92 (39.3)< 0.01
Smoking habits (%)116 (26.7)91 (38.9)< 0.05
Family history of GC (%)31 (7.1)29 (12.4)< 0.001
H. pylori-positive (%)271 (62.4)172 (73.5)< 0.05

Assessment of prevalence of H. pylori infection

During gastroscopic examination, biopsies (one each) from the antrum and body were obtained for rapid urease testing (CLOtest; Delta West, Bentley, Australia) and histologic examination (2 specimens each from the antrum and the body). The specimens were fixed in 10% formalin, embedded in paraffin wax, and 5 μm sections were stained with hematoxylin-eosin and a modified Giemsa method. Subjects in whom the rapid urease test and modified Giemsa staining results were positive were diagnosed as H. pylori-positive.

Genotyping of IL-1A, IL-1B and IL-1RN

Genomic DNA was isolated from peripheral blood using a standard phenol/chloroform extraction method. For the polymorphism at position 889 in the promoter region of the IL-1A, the polymerase chain reaction with confronting two-pair primers (PCR-CTPP) was used (Table II); this does not require a step to digest PCR products for single nucleotide polymorphism genotyping.24 This biallelic polymorphism produces a C allele of 428 and 279 base pairs (bp), and a T allele of 428 bp and 186 bp. A single bp polymorphism at position 31 in the promoter region of the IL-1B was analyzed by PCR of restriction fragment length polymorphisms (RFLPs). The PCR products were digested with AluI at 37°C for 3 hr and then separated on ethidium bromide-stained 3% agarose gels. The C allele was designated if 4 bands of 344, 79, 20 and 5 bp were obtained and the T allele was designated if 5 bands of 247, 97, 79, 20 and 5 bp were obtained.25 Polymorphism at position 511 in the promoter region of the IL-1B was analyzed by PCR-RFLP. The PCR products were digested with AvaI at 37°C for 18 hr and separated as above. The C allele was designated if 2 bands of 88 and 67 bp were obtained and the T allele was designated if a single band of 155 bp was obtained.22 The PCR products for tandem repeat polymorphisms of IL-1RN were used directly for electrophoresis. The IL-1RN gene has a pentaallelic 86 bp tandem repeat (VNTR) at intron 2,26 and the alleles were coded conventionally as follows: allele 1, 4 repeats; allele 2, 2 repeats; allele 3, 5 repeats; allele 4, 3 repeats; and allele 5, 6 repeats (Table II).

Table II. PCR Conditions For IL-1A, IL-1B and IL-1RN
PolymorphismPrimersPCR conditions
IL-1A, C to T at −889F1 5′-ATCACACCTAGTTCATTTCC-3′PCR-CTPP (confronting two-pair primers)
 R1 5′-TACATATGAGCCTTCAATGA-3′10 min at 95°C, 30 cycles of 1 min at 95°C, 48°C and 72°C and 5 min at 72°C
F2 5′-TAATAGTAACCAGGCAACAC-3′
 R2 5′-CAACACATTTCCTATAGAGG-3′ 
IL-1B, T to C at −31F 5′-AGAAGCTTCCACCAATACTC-3′PCR-RFLP (AluI)
 R 5′-ACTAAC TT TAGGGTGTCAG-3′10 min at 95°C, 36 cycles of 1 min at 94°C, 54°C and 72°C and 5 min at 72°C
IL-1B, C to T at −511F 5′-GCCTGAACCCTGCATACCGT-3′PCR-RFLP (AvaI)
 R 5′-GGAATCTTCCCACTTACAGATGG-3′10 min at 95°C, 40 cycles of 1 min at 95°C, 55°C and 72°C and 5 min at 72°C
IL-1RN, 86 bp VNTR at intron 2F 5′-CTCAGCAACACTCCTAT-3′10 min at 95°C, 35 cycles of 1 min at 95°C, 55°C and 72°C and 5 min at 72°C
R 5′-TCCTGGTCTGCAGGTAA-3′

IL-1β cytokine measurement in gastric mucosa

Four biopsy specimens were obtained from the greater curvature side of the lower stomach during gastroscopic examination. The biopsy specimens were immediately placed in 2.0 mL of PBS (pH 7.4), frozen on dry ice and stored at −70°C until used. Samples were homogenized and centrifuged (10,000g, 10 min). Total protein in supernatants was measured by a modified Lowry method. IL-1β levels were assayed in duplicate by enzyme-linked immunosorbent assay (ELISA; R&D Systems, Minneapolis, MN). The mucosal IL-1β levels were expressed as picograms of cytokine per milligram of biopsy protein (pg/mg protein).27

Statistical analysis

In the control group, alleles at the individual loci of IL-1A, IL-1B and IL-1RN were assessed by the exact test for Hardy-Weinberg equilibrium. Comparison of genotype frequencies among control group and GC group was assessed by chi-square statistics. The observed haplotype frequencies for pairs of alleles were estimated and compared with those expected under no association by maximum likelihood using the estimating haplotype frequencies (EH) software program (ftp://linkage.rockefeller.edu/software/eh). Odds ratios (ORs) with 95% confidence intervals (CIs) and multivariate logistic regression models controlling for the effects of possible confounders were computed using the statistics program SAS System, version 8.01 (SAS Institute, Cary, NC). The following adjusting variables were used: age (in years), gender, H. pylori infection, family history of GC, alcohol consumption and smoking habits. For analysis of mucosal IL-1β levels, Duncan's multiple comparison test was used for comparing the data for 3 genotypes of IL-1B-31. Differences were considered significant at p < 0.05.

Results

Clinical characteristics

The 668 subjects included 434 controls and 234 patients with GC (Table I). The mean age was higher in the GC group than in the control group. Smoking habits were more common but alcohol consumption was less common in the GC group compared with the control group. Family histories of GC and H. pylori infection were more common in the GC group compared with the controls (Table I). GCs were classified according to Lauren28 as intestinal (n = 117), diffuse (n = 116) and atypical (n = 1) type. Early GC (n = 83) was defined as stage 1A (T1 N0 M0) or stage 1B (T1 N1 M0) by TNM staging of American Joint Committee on Cancer and International Union Against Cancer.29, 30 Advanced GC (n = 151) was defined as stage 1B (T2 N0 M0) and stages II, III and IV. Cancer of the gastric cardia (n = 20) was defined as an adenocarcinoma with its center located within 1 cm proximal and 2 cm distal to the gastroesophageal junction.31 Adenocarcinoma of the antrum or body or fundus was defined as a noncardia cancer (n = 214).

Hardy-Weinberg equilibrium test

In the control group, the observed genotype frequencies of IL-1A-889C/T, IL-1B-31T/C, IL-1B-511C/T and IL-1RN polymorphisms did not deviate significantly from Hardy-Weinberg equilibrium (p = 0.074, 0.188, 0.154 and 0.141, respectively).

Genotype of IL-1A-889

Most subjects were found to have either the C/C (76.0%) or C/T (20.1%) genotypes; the remaining 3.9% had a T/T genotype. There were no significant differences in the genotype frequencies of IL-1A-889 between the GC group and the control group (chi-square = 0.336; p = 0.563; Table III). The IL-1A-889 (C-to-T) genotype was not associated with an increased risk of developing GC.

Table III. Comparison of IL-1 and IL-1RN Genotyping
GenotypeControl group, n = 434 (%)Gastric cancer group, n = 234 (%)Chi-square1p1
  • 1

    Comparison of chi-square and p value between the control group and the GC group. There were no significant differences in IL-1A, IL-1B-31, IL-1B-511 and IL-1RN genetic frequencies between the 2 groups.

IL-1A-889 C/C327 (75.3)181 (77.4)0.3360.563
 C/T90 (20.7)44 (18.8)  
 T/T17 (3.9)9 (3.8)  
IL-1B-31 C/C103 (23.7)52 (22.2)1.1470.564
 C/T246 (56.7)128 (54.7)  
 T/T85 (19.6)54 (23.1)  
 T carrier331 (76.3)182 (77.8)  
IL-1B-511 C/C87 (20.0)54 (23.1)0.8530.653
 C/T245 (56.5)128 (54.7)  
 T/T102 (23.5)52 (22.2)  
IL-1RN 1/1389 (89.6)204 (87.2)  
 1/229 (6.7)20 (8.5)  
 2/2001.6240.512
 1/46 (1.4)6 (2.6)  
 1/310 (2.3)4 (1.7)  

Genotype IL-1B-31

There was no significant difference in genotype frequencies of IL-1B-31 between the GC group and the control group (chi-square = 1.147; p = 0.564; Table III). No significant increase in GC risk in the GC group was observed for IL-1B-31T carriers (adjusted OR = 1.0; 95% CI = 0.7–1.5) or for those with IL-1B-31T homozygotes (adjusted OR = 1.3; 95% CI = 0.7–1.8) compared with the control group. When cases were subdivided according to the histologic type of the tumors, genotype frequencies were only significant in intestinal-type GC compared with controls (chi-square = 7.4; p = 0.025). In this subgroup of GC, risk was increased for T carriers (crude OR = 1.8; 95% CI = 1.0–3.2) and T homozygotes (crude OR = 2.4; 95% CI = 1.3–4.6). Multivariate logistic regression analysis after controlling for the effects of possible confounders showed that the risk of developing intestinal-type GC was significantly increased for the patients homozygous for T (adjusted OR = 2.2; 95% CI = 1.1–4.3), but not for T carriers (adjusted OR = 1.7; 95% CI = 0.9–3.0; Table IV). There was a significant difference in genotype frequency between patients with intestinal and diffuse-type GC (chi-square = 9.4; p = 0.009). Risk was increased in intestinal-type GC for both T carriers (adjusted OR = 2.4; 95% CI = 1.2–4.6) and T homozygotes (adjusted OR = 3.4; 95% CI = 1.5–7.7) compared with patients with diffuse-type GC. However, when the patients were divided according to the tumor stage (early or advanced) and subsite of the tumors (cardia or noncardia), comparisons of IL-1B-31 genotype frequencies showed no significant differences between any subgroup of GC patients and the control group. Further, a risk of developing GC was not observed for IL-1B-31T carriers or T homozygotes (Table IV).

Table IV. Comparison Between IL-1B-31 Genotypic Frequencies in the Control Group and the Gastric Cancer Group According to Histologic Type, Tumor Stage and Subsite of the Tumors
Genotype IL-1B-31ControlIntestinalDiffuseEarly GCAdvanced GCCardiaNoncardia
Cases (n = 434)Cases (n = 117)OR (95% CI)Cases (n = 116)OR (95% CI)Cases (n = 83)OR (95% CI)Cases (n = 51)OR (95% CI)Cases (n = 20)OR (95% CI)Cases (n = 214)OR (95% CI)
  1. In patients with intestinal-type GCs, risk was increased for IL-1B-31T homozygotes (adjusted OR = 2.2; 95% CI = 1.1–4.3) compared with the control group. However, subtypes of GC divided according to the tumor stage (early or advanced) and subsite of the tumors (cardia or noncardia) showed no significant differences in IL-1B-31 genotype frequencies compared with the control group.

C/C10317Referent34Referent19Referent33Referent3Referent49Referent
C/T246661.5 (0.8–2.8)620.7 (0.4–1.2)430.8 (0.5–1.6)851.0 (0.6–1.6)151.9 (0.5–7.0)1130.9 (0.6–1.3)
T/T85342.2 (1.1–4.3)200.7 (0.3–1.2)211.2 (0.6–2.3)331.1 (0.6–2.0)20.7 (0.1–4.6)521.1 (0.7–1.9)
T carriers3311001.7 (0.9–3.0)820.7 (0.4–1.1)640.9 (0.5–1.7)1181.0 (0.6–1.6)171.6 (0.5–5.7)1650.9 (0.6–1.4)

Genotype IL-1B-511

There was no significant difference in genotype frequencies of IL-1B-511 between the GC group and the control group (chi-square = 0.853; p = 0.653; Table III). No significant increase in GC risk for the GC patients was observed for IL-1B-511T carriers (adjusted OR = 0.9; 95% CI = 0.6–1.3) or for IL-1B-511T homozygotes (adjusted OR = 0.9; 95% CI = 0.6–1.5) compared with the control group. Because the alleles of IL-1B-511 are closely related to those of IL-1B-31 by strong linkage disequilibrium, a significant inverse risk of developing intestinal-type GC was observed in patients who were IL-1B-511T homozygotes (adjusted OR = 0.42; 95% CI = 0.2–0.9) compared with the control group.

Genotype IL-1RN

There were 4 genotypes (2/4, 3/4, 4/4, 4/5 repeats) of IL-1RN in our study. Allele 1 (4/4 repeats) was the most common genotype in the GC group (82.9%) as well as in the control group (85.9%). In contrast to Caucasian populations, allele 2* (2/2 repeats) was not found in all subjects. The heterozygote allele 2* (2/4 repeats) was found in 8.6% of the GC group and in 6.7% of the control group. In this study, there was no significant difference in genotype frequencies of IL-1RN between the GC group and the control group (chi-square = 1.624; p = 0.512; Table III).

Haplotypes IL-1B-31 and IL-1B-511

Estimation of observed haplotype frequencies and comparison with those of expected haplotype frequencies showed a near complete linkage disequilibrium among IL-1B-31 and IL-1B-511 in the GC group (D′ = 0.99; chi-square = 428.0; p < 0.001), the control group (D′ = 1.0; chi-square = 805.6; p < 0.001) and all subjects pooled (D′ = 1.0; chi-square = 1233.0; p < 0.001; Table V). The linkage disequilibrium coefficients were nearly complete, as seen in Caucasian patients.20 However, whereas the pattern of haplotype IL-1B-511T/-31C was associated with the overall risk of developing GC in Caucasian populations, we found that haplotype IL-1B-31T/-511C was associated with the risk of intestinal-type GC in this Korean population.

Table V. Observed and Expected IL-1B-31/IL-1B-511 Haplotype Frequencies and Linkage Disequilibrium Coefficients
PopulationHaplotypeFrequencyDisequilibrium
IL-1B-31IL-1B-511ExpectedObservedD′Chi-squarep
Control group (n = 434)CC0.2510.0051.0805.6< 0.001
 CT0.2690.516   
 TC0.2310.478   
 TT0.2480.001   
GC group (n = 234)CC0.2500.0040.99428.0< 0.001
 CT0.2460.491   
 TC0.2540.499   
 TT0.2500.004   
All subjects (n = 668)CC0.2510.0051.01233.0< 0.001
 CT0.2610.507   
 TC0.2390.486   
 TT0.2490.002   

Mucosal IL-1β levels and polymorphisms

IL-1β cytokine levels in gastric biopsy specimens were measured in subjects with variant genotypes of IL-1B-31. The subjects included 74 controls and 44 GC patients without H. pylori infection and 125 controls and 114 GC patients with H. pylori infection. There were no significant differences in mucosal IL-1β levels in H. pylori-noninfected subjects regardless of the genotypes of IL-1B-31. H. pylori-infected subjects had higher IL-1β levels than those of H. pylori-noninfected subjects. However, there were no significant differences in IL-1β levels among variant genotypes in the H. pylori-infected controls. In contrast, in the GC patients with H. pylori infection, those patients homozygous for IL-1B-31T showed a significantly higher mean IL-1β level (42.5 ± 25.8 pg/mg protein) compared with those having IL-1B-31C/T (28.5 ± 21.9 pg/mg protein; p < 0.05) and IL-1B-31C/C genotypes (29.4 ± 26.7 pg/mg protein; p < 0.05; Table VI). These differences were also found in the overall population (Table VI, Fig. 1).

Table VI. Mucosal IL-1β Level and IL-1B-31 Polymorphisms
GroupIL-1β Level (pg/mg protein)p
IL-1B-31 T/TIL-1B-31 C/TIL-1B-31 C/C
  1. Mucosal IL-1β levels of subjects carrying the IL-1B-31T/T allele compared with IL-1B-31C/T and IL-1B-31C/C genotypes analyzed using Duncan's multiple comparison test.

H. pylori-noninfected controls (n = 74)8.2 ± 5.7 (n = 19)6.1 ± 6.5 (n = 38)5.7 ± 5.1 (n = 17)NS
H. pylori-noninfected patients with GC (n = 44)13.1 ± 12.2 (n = 14)9.8 ± 8.0 (n = 18)12.0 ± 9.4 (n = 12)NS
H. pylori-infected controls (n = 125)24.9 ± 25.2 (n = 32)23.5 ± 17.3 (n = 60)20.4 ± 15.9 (n = 33)NS
H. pylori-infected patients with GC (n = 114)42.5 ± 25.8 (n = 34)28.5 ± 21.9 (n = 48)29.4 ± 26.7 (n = 32)< 0.05
H. pylori-infected controls plus GCs (n = 239)33.9 ± 26.9 (n = 66)25.7 ± 22.0 (n = 108)24.8 ± 19.3 (n = 65)< 0.05
Figure 1.

Gastric mucosal levels of IL-1β cytokine in relation to the genotypes at IL-1B-31 in the H. pylori-infected controls and patients with GC. (a) Mucosal IL-1β levels were not different among the variant genotypes in the controls. (b) In contrast, in the patients with GC, those carrying the IL-1B-31T/T allele showed significantly higher IL-1β levels compared with those of −31C/T and −31C/C genotypes (p < 0.05). (c) These differences were also found in all subjects pooled (controls plus GC patients; p < 0.05). The horizontal line indicates the mean value.

Discussion

It is widely accepted that chronic H. pylori infection induces hypochlorhydria and gastric atrophy, both of which are precursors of GC. Although the mechanism of H. pylori-induced carcinogenesis is not clear, a critical factor is gastric acid secretion, which both influences and is influenced by H. pylori-induced gastritis. Host genetic factors that influence acid secretion are thus highly likely to influence the outcome of an H. pylori infection. IL-1β is a potent proinflammatory cytokine that is upregulated in the presence of H. pylori and is important in initiating and amplifying the inflammatory response to this infection.17 IL-1ra, a naturally occurring antiinflammatory cytokine, is locally produced in various tissue in response to infection or inflammation. Both IL-1β and IL-1ra play a key role in modulating the inflammatory response in the gastrointestinal mucosa as well as in regulating gastric acid secretion.18, 19

The IL-1B gene has 3 diallelic polymorphisms at positions −511, −31 and +3954 base pairs from the transcription start site.32, 33 The polymorphism at 3954 is not associated with the risk of GC and H. pylori infection.24 The T allele at −31, but not −511, in the promoter region of the IL-1B forms a TATA box, suspected of enhancing gene expression and induction of IL-1β. A breakthrough in the investigation of H. pylori-associated diseases was the discovery that polymorphisms of IL-1B-511T/IL-1B-31C and IL-1RN*2 alleles are significantly associated with the occurrence of hypochlorhydria and GC among Caucasians.20 That study also found that there was nearly complete linkage disequilibrium between IL-1B-511 and IL-1B-31 and concluded that the IL-1B-31T allele is the wild type and the IL-1B-31C allele is a mutant. However, ethnicity might greatly affect genotype status, and 2 recent studies revealed no difference in IL-1B-511 genotypes between GC and controls in Japanese and Taiwanese populations.22, 23 Even within Asia, the IL-1B polymorphisms differ between northern (Japanese and Chinese) and southern (Thai and Vietnamese) populations. In subjects with severe gastric mucosal atrophy, the IL-1B-511C/C polymorphism was dominant in the Japanese population and the IL-1B-511T/T + T/C polymorphism was dominant in the Chinese population; however, no differences in C- and T-allele frequencies were found in the Thai and Vietnamese populations.34 Interestingly, a recent study found that the IL-1B genotypes differed between a high GC prevalence region (northern China) and a low GC prevalence region (southern China).35 In the high prevalence region, the frequency of IL-1B-31C/C was higher among GC patients than controls. However, in the low prevalence region, this genotype frequency was higher in controls than in patients with GC; moreover, there was not near complete linkage disequilibrium between the IL-1B-31 and IL-1B-511 loci. The study considered the 2 genotypes to be independent risk factors for gastric carcinogenesis. In Korea, 2 recent studies did not find an association between IL-1B-511T and IL-1RN polymorphisms and an increased risk of developing GC.36, 37

In our study, polymorphisms of IL-1B-511T were not associated with the risk of developing GC in a Korean population. We found that in this group, the risk was more closely related to polymorphisms of IL-1B-31T rather than 1B-511T. Although IL-1B-31T polymorphism was not frequently observed in the overall group of patients with GC, when we subclassified the GC cases according to their histologic type (intestinal, diffuse, or atypical), tumor stage (early or advanced) and anatomic subsite (cardia or noncardia), IL-1B-31T homozygotes were significantly associated with an increased risk of intestinal-type GC. Compared with diffuse-type GC, risk was increased in intestinal-type GC for both IL-1B-31T homozygotes and IL-1B-31T carriers. In contrast to a study on IL-1B-511T in Portugal,21 we found an association between the risk of developing intestinal-type GC and IL-1B-31T polymorphisms.

Western20 and Japanese reports24 and this study found strong linkage disequilibrium between the IL-1B-31 and IL-1B-511 alleles. El-Omar et al.20 reported that the alleles at position 31 were tightly linked with those at position 511: 0.005 for the C-C combination, 0.375 for C-T, 0.610 for T-C and 0.010 for T-T among the controls in Scotland, and 0.000 for C-C, 0.298 for C-T, 0.699 for T-C and 0.002 for T-T among the controls in Poland. Furthermore, they confirmed that IL-1B-31T/IL-1B-511C is the more common haplotype and IL-1B-31C/IL-1B-511T is the less common haplotype by forward and reverse sequencing of both of these loci. In the study on Japanese people, the corresponding values were 0.010, 0.438, 0.552 and 0.000 in H. pylori-infected patients without GC.24 In the present study, the corresponding values were 0.005, 0.516, 0.478 and 0.001 in the control group and 0.004, 0.491, 0.499 and 0.004 in the GC group. Although the linkage disequilibrium coefficients in this study were similar to those of the study on Caucasian subjects, the haplotype pattern of IL-1B-511/IL-1B-31 related to the risk of GC was in the opposite direction. Whereas the IL-1B-511T/IL-1B-31C haplotype in Caucasian populations was associated with the risk of developing GC, in this Korean population, the risk was with the IL-1B-31T/IL-1B-511C haplotype.

These results are very interesting and might surprise Western investigators. Ethnic origin is a key determinant of the frequency of genetic markers in a population. Our findings thus might reflect the influence of past selective pressures on genotypes of Korean populations over a long time. Korea has been invaded by foreign countries many times over the past 2,000 years, so maintaining our own lineage has traditionally been regarded as very important. As Koreans have long resisted interracial marriage, they tend to be ethnically homogeneous. However, we could not find another reverse haplotype to support our results in Korean genetic studies.

The hypotheses regarding IL-1 genetic polymorphism and gastric carcinogenesis are based on the assumption that carriers of these genotypes are associated with increased levels of IL-1β in the gastric mucosa in response to H. pylori infection. However, the effect of these polymorphisms remains unclear. Some investigators examined the issue by using in vitro cell culture study and have reported that these polymorphisms are related to IL-1β production, whereas others did not find an association.38, 39, 40 Recently, gastric mucosal IL-1β levels were measured using ELISA in H. pylori-infected patients with variant IL-1B-511 and IL-1RN genotypes; carriers of the IL-1B-511T/T genotype or IL-1RN*2 allele had higher mucosal IL-1β levels than noncarriers. Carriers of both IL-1B-511T/T genotype and IL-1RN*2 allele had the highest IL-1β levels.25 Furthermore, another recent study showed that an H. pylori-positive group with IL-1B-511T/T genotype had higher gastric secretion pH, lower pepsinogen I/pepsinogen II ratios, higher gastric atrophy and gastritis scores compared with those of C/T and C/C groups.27 In this study, we investigated the effects of IL-1 genetic polymorphism on IL-1β production in the gastric corpus mucosa. The subjects included H. pylori-infected and H. pylori-noninfected controls and GC patients. IL-1β levels in H. pylori-noninfected subjects were significantly lower than those in H. pylori-infected subjects, regardless of IL-1B genotype. IL-1β levels in H. pylori-infected controls were not different among variant genotypes. However, in H. pylori-infected patients with GC, IL-1β levels were significantly higher in subjects bearing the IL-1B-31T/T allele compared with IL-1B-31C/T and C/C genotypes. Therefore, we confirmed that the polymorphisms of IL-1B-31T homozygotes induced mucosal IL-1β production in response to H. pylori infection. These results were reverse to another report in which the highest IL-1β level was found in subjects with IL-1B-511TT genotypes.25 El-Omar et al.20 considered the IL-1B-31C allele to be a proinflammatory gene (mutant type), whereas the IL-1B-31T allele is considered a wild type that induces normal expression. However, mutation of T-to-C in the TATA box in the IL-1B gene promoter will result in downregulation of IL-1β expression. Therefore, we believe that the IL-1B-31T allele is a proinflammatory allele; this was also proposed in a Chinese study by Zeng et al.35 Our results are compatible with a study of Japanese people by Takagi et al.41 They found that IL-1B polymorphisms (IL-1B-31T/T and IL-1B-511C/C) enhanced not only IL-1β production but also IL-8 production in the gastric body.

IL-1A is reported to have 3 polymorphisms: diallelic polymorphisms at positions −889, 4845 and 46 bp VNTR polymorphisms at intron 6.17 The association of IL-1A polymorphisms with GC risk has not been elucidated. We examined the association between IL-1A polymorphisms at position −889 and the risk of developing GC. About 76% of all subjects were found to have the C/C genotype. There was no association of IL-1A-889 polymorphisms with the risk of GC. The IL-1RN gene has a pentaallelic 86 bp tandem repeat (VNTR) in intron 2, of which the less common allele 2 (IL-1RN*2) is associated with a wide range of chronic inflammatory and autoimmune conditions.19, 26IL-1RN*2 is associated with enhanced IL-1β production in vitro. IL-1RN*2 allele was associated with an increased risk of hypochlorhydria and development of GC in Caucasian populations.20, 21 These studies suggested that synergistic interaction between IL-1B-511T carriers and IL-1RN*2 genotype increased the risk of developing GC. In the present study, most subjects in the control group and the GC group had the IL-1RN*1 (4/4 repeats) allele. The IL-1RN*2 (2/4 repeats) heterozygous allele was found in 7.3% of all subjects. However, the IL-1RN*2 homozygous allele was not found. Our results are similar with other East Asian studies on the IL-1 polymorphism, with the frequency of the IL-1RN*2 allele 1% in Japanese24 and 2.5% in Taiwanese.42 We could not find any association between the IL-1RN genotype and the risk of developing GC in this Korean population.

In conclusion, although we did not measure gastric acid output, IL-1B-31 polymorphisms with high mucosal IL-1β levels might result in hypochlorhydria. These results support the hypothesis that the combined effect of H. pylori infection and host genetic factors such as IL-1B-31T/IL-1B-511C polymorphisms with subsequent enhanced mucosal IL-1β production contribute to the development of intestinal-type GC in a Korean population.

Ancillary