MATERIALS AND METHODS

Patients with histologically confirmed, earlier untreated adenocarcinoma of the stomach were recruited at the institution from March, 2003 to Dec., 2007. Cancer-free control subjects were selected from individuals presenting with either gastric ulcer or chronic gastritis in the gastroenterology clinic during the same time period. All subjects were nationals of Iran and were living in Tehran. Subjects were excluded if they had a known history of immune compromise (HIV positive, used steroids in the previous month, or received blood transfusion within the prior 6 months) or anti-H. pylori antibiotic therapy. After informed consent, each participant provided a blood sample.

The amount of immunoglobulin G (IgG) antibody to the high-molecular-weight cell-associated proteins of H. pylori was determined by using a well-characterized enzyme-linked immunosorbent assay (ELISA) (CaptiaJ H. pylori IgG ELISA; Trinity Biotech, Jamestown, NY) according to the manufacturer=s instructions. The manufacturer-recommended cutoff points are values #0.9 ELISA units (EU) for negative, 0.9 to 1.1 EU for indeterminate, and $1.1 EU for positive. A pilot study (unpublished data) using the same ELISA kit on serum samples from 75 biopsy-positive cases and 48 biopsy-negative controls achieved a sensitivity of 100% and a specificity of 75%.

For the assays of p53 protein, the sections were mounted on organosilane-coated glass slides, dewaxed and rehydrated. One slide of each case was transferred into citrate buffer and boiled twice for 5 min in a microwave oven. The sections were then incubated overnight at +4°C with monoclonal anti-p53 antibody D07 (Dako, Glistrup, Denmark) diluted to 1:1000, washed and treated for 30 min with biotinylated secondary antibody (ABC Vectastain Elite Kit, Vector). The slides were incubated in preformed avidin-biotin-peroxidase complex, and peroxidase was made visible by immersing the slides in diaminobenzine tetrahydrochloride containing hydrogen peroxide (Sigma, Paisley, UK). Finally, the sections were counterstained with Mayer`s haematoxylin. A sample of gastric cancer, previously shown to express p53, was used as a positive control.

For Ki-67, the sections were boiled in a household microwave oven six times, 5 min each, in citrate buffer. The Ki-67 antigen was stained by applying a polyclonal antibody (Dako, Glistrup, Denmark) overnight as a 1:300 dilution and visualized as described above. Tonsil tissue was used as a positive control.

The analysis was performed in 2 steps. First, Chi-square and Student=s t-test were used to compare the groups in univariate analysis. Statistically and clinically significant variables were extracted and included in multivariate binary logistic regression, in which Odds Ratios (OR) and 95% confidence intervals (95% CI) were determined and the effect of potential confounders was assessed. All analysis were done using SPSS 15.0 and the level of significance was always set at 0.05.

RESULTS

One case subject and one control subject had indeterminate results when tested for presence of H. pylori-specific IgG and were excluded from the analysis. Sixty-nine cases and 96 controls were eventually included. There was no significant difference between the groups for age (p = 0.17). Although male sex was more common in both case and control subjects but the groups were significantly different regarding sex (p = 0.004) (Table 1). The frequency of H. pylori seropositivity and presence of p53 mutation were more common in case group (p<0.001 and p<0.001) while the difference between groups was not statistically significant for presence of Ki-67 mutation (p = 0.61).

Table 1:	Demographic data, p53, Ki-67, and H. pylori status for the groups
*p<0.05

Table 2:	Multivariate regression on variables extracted from univariate analysis (Table 1)
*p<0.05

Sex, H. pylori status, and presence of p53 and Ki-67 were selected for further analysis (Table 2). In multivariate binary logistic regression, sex, H. pylori status, and presence of p53 were selected as covariates because of their statistically significance while presence of Ki-67 was selected as covariate because of its clinically significance. The effect of p53 mutation (p<0.001, adjusted OR = 126.56, 95% CI: 28.58-560.40) and Ki-67 mutation (p = 0.02, adjusted OR = 0.21, 95% CI: 0.05-0.84) remained significant.

DISCUSSION

By multivariate analysis, which is the appropriate method when potentially confounding factors exist, we adjusted for sex and H. pylori seropositivity between cases and controls. They were already matched in age. The results showed that p53 and Ki-67 are independent significant affecting factors for gastric cancer. Mutations of the p53 gene are the most common genetic changes found in human cancers thus for, occurring in over 60% of all cancers (Sipponen et al., 1998; Ranzani et al., 1995). According to the earlier studies, the expression rate of p53 protein in gastric cancer varies from 25 to 61% (Joypaul et al., 1993). The p53 database International Agency for Research on Cancer (IARC), listed 378 gastric cancer cases with point mutation in the p53 gene (Hainaut et al., 1998). These mutation leads to error in the expression of cell cycle inhibitory and the final process lead to dysregulation of the cell growth.

H. pylori infects approximately half the world population (Malaty, 2007). The prevalence of infection is particularly high in developing countries, with nearly all adults being infected (Malaty, 2007). The prevalence of infection in adult Iranians has been estimated to be approximately 90% (Massarrat et al., 1995; Zendehdel et al., 2005). The prevalence, however, largely varies one region to another. A recent study on 300 randomly selected blood donors in Tehran revealed a prevalence of 51.7% (Khodarahmi et al., 2008), which is rather close to the one obtained in this study and might reflect a higher socioeconomic status of the participants compared with those in the study by Massarrat et al. (1995). In Japan, it has been reported that each year gastric cancer develops in 300,000 (0.5%) of 60 million people who are H. pylori positive, which means gastric cancer develops in 5% of H. pylori positive persons over 10 years (Uemura et al., 2001). Uemura et al. (2001) reported that gastric cancers developed in 2.6% H. pylori-positive Japanese patients and none of H. pylori-negative patients during a mean follow up period of 7.8 years.

The expression rates of Ki-67 antigen in present study are a little higher than those previously reported (27-45%) and few reports concerning Ki-67 and its association to stage and survival in gastric cancer are controversial (Kakeji et al., 1991; Kimura et al., 1992; Victorzon et al., 1996). In the present study, Ki-67 expression had no impact on developing gastric cancer in univariate analysis but it found its protective role in multivariate analysis. A number of limitations accompanied present study. First, the groups were not originally matched in sex and H. pylori seropositivity status. Therefore, we had to adjust the variables by use of logistic regression. Second, we used only serology to diagnose H. pylori infection. Using direct methods to diagnose infection can significantly add to the accuracy of our results. Finally, the study was cross-sectional, over a period of 5 years. Carefully designed cohort studies with larger sample size can better demonstrate if p53 and Ki-67 expressions in relation with H. pylori seropositivity are associated with gastric cancer.

ACKNOWLEDGMENT

The study was sponsored by Otorhinolaryngology Research Center, Tehran University of Medical Sciences, Tehran, Iran.