Interleukin-8 Polymorphisms Are Not Associated with Gastric Cancer Risk in a Polish Population

Inflammation seems to play a critical role in the development of many types of cancer, including gastric, colorectal, and bladder cancers (1-6). The presence of chronic inflammation has been well described in gastric cancer, the fourth most common cancer diagnosis and second most common cause of cancer death worldwide (4, 7-9). Genetic variants of several pathways critical for the inflammatory response have been studied and a number of single nucleotide polymorphisms (SNPs) in several genes from different pathways have been associated with gastric cancer risk, including IL1, IL1RN, IL1B, TNF, IL6, IL8, and IL10 (10-17).

Interleukin-8 (IL8), a potent chemokine, may play a role in gastric cancer pathogenesis. Gastric cancer specimens have increased IL-8 protein levels, and many gastric cancer cell lines express high levels of IL8 mRNA and protein (16, 18, 19). SNPs in the IL8 promoter (−251 A>T, rs4073) and a linked SNP in intron 1 (IVS +230 G>T, rs2227307) were associated with increased risk for gastric cardia adenocarcinoma in a high-risk Chinese population (14). The A allele of the IL8 −251 A>T SNP (rs4073) was associated with increased risk for gastric ulcer and gastric cancer in Helicobacter pylori–infected Japanese individuals (13). Increased IL-8 protein production was seen in individuals with the IL8 −251 A allele after lipopolysaccharide stimulation of whole blood (20). The A allele also seems to increase transcriptional activity in vitro in response to IL1-1β and tumor necrosis factor-α stimulation (13).

Central and Eastern European populations, including Poland, show a high incidence and mortality of stomach cancer among Caucasians (21). Because IL-8–mediated inflammation and IL8 SNPs may play a role in gastric cancer etiology, we studied genetic variation in IL8 in a population-based case-control study of gastric cancer in individuals from Warsaw, Poland (22).

As previously described (22), cases consisted of residents of Warsaw, Poland, ages 21 to 79 years, newly diagnosed with gastric cancer between March 1994 and April 1996. All cases were gastric adenocarcinoma and pathologic slides were reviewed in a standardized fashion. Controls were frequency matched to cases by sex and 5-year age groups and randomly selected from a computerized registry of Warsaw residents. Written informed consent was obtained from all participants, and the Institutional Review Boards at the U.S. National Cancer Institute, Bethesda, MD and the Cancer Center and M. Sklodowska-Curie Institute of Oncology, Warsaw, Poland approved the study.

Genomic DNA from 288 cases and 430 controls was extracted from buffy coats by standard methods. Of the cases, 211 were from the distal stomach, 32 were from the cardia only, and 35 were distal and cardia, and location within the stomach was unknown in 10 cases. Four SNPs in IL8 were genotyped, IL8 −251 A>T (also annotated as IL8 −351, rs4073), IL8 IVS1 +230 G>T (rs2227307), IL8 IVS1 −240 C>T (rs2227306), and IL8 Ex1 −65 C>T (rs2227538), by either Taqman Assays (Applied Biosystems, Foster City, CA) or MGB Eclipse Assays (Epoch Biosciences, Bothell, WA) at the National Cancer Institute's Core Genotyping Facility. Details on assay design and conditions are available at http://snp500cancer.nci.nih.gov/ (23).

Analysis of the case-control genotype data was conducted with contingency tables using additive, dominant (one or two copies of variant allele necessary for risk), and recessive (two copies of variant allele necessary for risk) genetic models in SAS v8.02 software. We also examined whether the association between the SNPs and cancer risk was modified by other risk factors, including age, smoking, and site of tumor origin. Haplotypes were constructed, and a case-control permutation test was done using PHASE v2.1 (24). HaploStats v1.1.1 (25) was used to construct haplotypes and determine the global score P, haplotype frequencies, and odds ratios. Hardy-Weinberg equilibrium was determined for all loci in controls by χ².

Allele frequencies and P values using codominant, dominant, and recessive models are shown in Table 1. IL8 Ex1 −65 C>T was monoallelic in this study population. IL8 −251 A>T, IL8 IVS1 +230 G>T, and IL8 IVS1 −240 C>T were in Hardy-Weinberg equilibrium in controls. These three SNPs in IL8 do not seem to be associated with risk for gastric cancer in this study of individuals from Warsaw, Poland. When only distal stomach cases were analyzed, there was no association between genotype and risk for disease. The low frequency of gastric cardia cases precluded individual analyses. Analyses for association based on age, smoking status, or H. pylori infection was not statistically significant. Haplotypes of IL8 −251 A>T, IL8 IVS1 +230 T>G, and IL8 IVS1 −240 C>T constructed with PHASEv2.1 and HaploStats v1.1.1 did not show a statistically significant differences between controls and total cases, tumor site, or other risk factors.

Because the IL8 −251 A>T SNP seems to effect promoter function (13, 20), it is an important candidate for studies of genetic variation, inflammation, and disease risk. Genetic variation in the three IL8 SNPs studied (IL8 −251 A>T, rs4073; IL8 IVS1 +230 T>G, rs2227307; and IL8 IVS1 −240 C>T, rs2227306) does not seem to be a risk factor for gastric cancer in this study of individuals from Poland. This study of 288 cases and 430 controls is somewhat larger than previous studies of gastric cancer and IL8 polymorphisms (13, 14). Ohyauchi et al. found an association between the A allele of the IL8 −251 A>T SNP in 212 H. pylori–positive gastric cancer (majority noncardia) patients from Japan and two control groups (n = 244 and n = 102; ref. 13). Increased risk for gastric cardia cancer was seen in Chinese individuals (90 cases and 454 controls) who carried the A allele of the IL8 −251 A>T SNP and the G allele of the linked IL8 IVS1 +230T>G (14) but noncardia gastric cancer, and H. pylori were not evaluated. IL8 −251 A>T SNP has also been extensively studied in relation to other cancers. Table 2 shows a summary of association studies that investigated the IL8 −251 A>T SNP in a variety of cancers in different parts of the world. Although IL8 −251 A>T has not been consistently associated with cancer risk, it may contribute to cancer risk in certain populations.

Due to the lack of consistent association of the IL8 −251 A>T allele and cancer risk, it is important to take into consideration environmental factors and population characteristics. The allele frequencies of the IL8 −251 A>T SNP differ between ethnic groups, perhaps in response to differential selective pressure from geographic infectious diseases. The dbSNP database (build 124, http://www.ncbi.nlm.nih.gov/SNP/snp_ref.cgi?rs=4073) illustrates the differences in frequencies of the A and T alleles (Table 3). Differences between individuals of African descent (African/African American) and those of European descent (Caucasian) are most notable from the PGA-UW-FHCRC and SNP500Cancer panels. However, there also seem to be significant differences in the Kyugen data of between the CEPH parents (European descent) and those from East Asia, with an A allele frequency of 42% and 34%, respectively. In addition, there are other cytokine genes with significant genetic variation between different ethnic groups (30). It is possible that population-specific genetic variation in different genes in the same or related pathways could also contribute to cancer risk.

The lack of association of IL8 SNPs in our study of gastric cancer risk in Poland may be due to differences in allele frequencies between ethnic groups, disease etiology, and/or sample size limitations. Like the other studies described (13, 14), our study of genetic variation in IL8 in gastric cancer in Poland is somewhat limited by sample size. Significantly larger case-control studies of genetic variation in IL8 and its role as a risk factor in inflammation and cancer are needed.