A Limited Association of OGG1 Ser326Cys Polymorphism for Adenocarcinoma of the Lung

The 8-oxoguanine DNA glycosylase (OGG1) repairs DNA by removing 8-hydroxyguanine, a highly mutagenic oxidative DNA adduct. Recently, the gene for OGG1 was cloned and several polymorphisms have been reported. Because environmental carcinogens produce 8-hydroxyguanine residues that potentially cause oncogenic mutations by mismatching to this modified base, the capacity to repair these lesions can be involved in cancer susceptibility. This study investigated the association between OGG1 Ser326Cys polymorphism and risk of the lung adenocarcinoma for Japanese by a prevalent case-control study in Japan. The subjects comprised 138 cases and 241 non-cancer outpatients as controls. OGG1 gene polymorphism was genotyped by a PCR-CTPP (polymerase chain reaction with confronting two-pair primers) method. The distribution of OGG1 Ser326Cys genotype among controls (Ser/Ser, 28.3%; Ser/Cys, 49.2%; and Cys/Cys, 22.5%) was not different from that among cases (Ser/Ser, 29.0%; Ser/Cys, 51.4%; and Cys/Cys, 24.0%). The sex-age adjusted odds ratio (OR) was 1.06 with 95% confidence interval (CI) 0.64-1.76 for Ser/Cys genotype and 0.81 with 0.44-1.52 for Cys/Cys genotype. The ORs according to the interval between diagnosis and study enrollment were also examined because the polymorphism was a potential prognostic factor of lung cancer. The ORs of Ser/Cys and Cys/Cys genotypes in the cases less than 3 years after diagnosis were higher than overall ORs; 1.86 (95%CI, 0.91 -3.77), and 1.46 (0.64-3.35), respectively. The OR for smoking was not statistically different among genotype, though the sample size was too small to detect even a moderate interaction. This study supported the first study by Sugimura et al (Cancer Epidemiol Biomarkers Prev, 1999; 8: 669-674), that the association of OGG1 Ser326Cys polymorphism was limited for the risk of lung adenocarcinoma.


INTRODUCTION
Oxidative DNA damage is thought to cause mutations, which can activate oncogenes or inactivate tumor suppresser genes finally leading to cancer 1, 2). To protect the damage, organisms have developed highly efficient DNA repair machineries 3,4). Interindividual variation in the repair capacity has been implicated as a cancer susceptibility factor. 8-hydroxyguanine is one of the major forms of oxidative DNA damage produced by reactive oxygen species (ROS), and causes mutagenic / carcinogenic DNA misreading through G:C to T:A transversion 5,6). G:C to T:A transversion are widely seen in tumors e. g. in lung tumors where these transversion are very frequently found in the p53 gene and compared with other tumors such as those of colon or breast 7,8). Several kinds of human cancer tissues, including lung cancer, showed higher levels of 8-hydroxyguanine compared with their non-cancerous counterparts 9-11). These findings indicate that this kind of oxidative DNA damage causes human cancer development.
The human 8-oxoguanine DNA glycosylase I is one of 8hydroxyguanine repair enzymes 12-16). OGGI is the gene encoding the glycosylase, which belongs to the base excision repair gene family 17). Studies on genetic structure have revealed the presence of several polymorphisms within the OGG1 locus 18,19) Among them, Ser326Cys polymorphism has been shown to be very common in Japanese and Chinese populations 1321) but less common in Caucasian populations 20,22). The Cys326 protein has been shown to have about 7 times weaker 8-hydroxyguanine-repair capacity than Ser326 protein in complementation assay using Escherichia coli mutant strain deficient in 8hydroxyguanine 19). However, Ser326Cys polymorphism did not correlate with the mean 8-hydroxyguanine levels in peripheral lymphocytes 19) and nor with the adduct levels in human lung samples 23). In addition, recent studies reported no difference in the catalytic activities between Ser326 and Cys326 alleles 24,25). Therefore, the difference in the function in vivo between the two alleles remains unclear.
The association between OGG1 Ser326Cys polymorphism and the risk of lung cancer has been investigated in 3 studies 19, 20,22). In all studies, the frequencies of Cys/Cys genotype were slightly higher among cases than among controls, but not significant. For other cancers, a significant association with the risk of esophageal cancer risk has been reported in China 21) In Japan, lung cancer is the leading cause of deaths among malignancies 26).The cumulative incidence of lung adenocarcinoma has been steadily increasing for both males and females, while that for lung squamous cell carcinoma was almost constant from 1970s 27). The reason is not clear, but the change in the kind of cigarettes from high-tar no-filtered to low-tar filtered is suspected to be one of the factors 27). Although the association of old type cigarettes with the adenocarcinoma is weaker than with the squamous or small cell carcinoma, the adenocarcinoma may have a stronger association with low-tar filtered cigarettes. In order to detect high-risk individuals, studies on the interactions between genetic traits and smoking should be conducted for many tobacco-related polymorphisms. In this paper, we examined the association between Ser326Cys polymorphism and adenocarcinoma of the lung by a prevalent case-control study with hospital controls. Furthermore, the effect modification by this polymorphism was examined for smoking habit.

Study subjects
Cases were lung cancer patients who visited Aichi Cancer center Hospital during 1999-2000. At the first stage, they were asked by their doctors in charge whether to participate in this study or not (not counted how many outpatients were asked by the doctors). At the second stage, the assentients at first stage were enrolled after written informed consent by staffs of  Table 2). As shown in Table 3, the crude and adjusted ORs for Ser/Cys or Cys/Cys genotypes compared with Ser/Ser were not significant. We also analyzed the ORs by the interval from diagnosis, because there was a significant difference in OGG1 genotype distribution between the recent cases and the longer survivors. The adjusted ORs compared with Ser/Ser genotype were 1.86 (95% CI, 0.83-3.24) for Ser/Cys and 1.53 (0.69-3.38) for Cys/Cys in the recent cases, and 0.69 (0.38-1.25) and 0.48 (0.22-1.09) in the longer survivors, respectively. The adjusted ORs for Ser/Cys and Cys/Cys genotypes combined were 1.72 (0.88-3.39) in the recent cases and 0.64 (0.36-1.14) in the longer survivors. The OR of the genotype among current smokers did not differ markedly from that among never smokers, though former smokers had a lower OR ( Table 4).
The age-sex-adjusted OR relative to never smokers was

DISCUSSION
This was the first study that OGG1 Ser326Cys polymorphism was genotyped by the PCR-CTPP method 29), whose genotyping was quite clear as demonstrated in Figure 1. The method is easy, timesaving and inexpensive compared with PCR-SSCP. The genotype distribution was in accordance with Hardy-Weinberg law of equilibrium among controls (p=0.873), which partly supported appropriate control selection and correct genotyping. The genotype distribution (28.3% for Ser/Ser, 49.2% for Ser/Cys, and 22.5% for Cys/Cys) was similar to those previously reported by Sugimura  In our study, compared with Ser/Ser genotype , the Ser/Cys and Cys/Cys genotypes were not significantly associated with the risk of adenocarcinoma of the lung . Because this was a prevalent case-control study, the estimated ORs indicated not only with disease occurrence but also with disease prognosis . In prevalent case-control studies, longer survivors are more likely to be sampled as cases 31). Accordingly, the factors relat-ed to shorter survival are less frequent among prevalent cases than among incident cases, which include those with poor prognosis as well as long survivors. It produces a lower OR for the prevalent cases than that estimated for the incidence cases. We divided the case subjects into 2 groups (less than 3 years after diagnosis versus 3 years or longer) and estimated the OR according to the interval from diagnosis. Although the difference was not statistically significant, the ORs of Ser/Cys and Cys/Cys genotypes compared with Ser/Ser genotype were larger in the recent cases than in the longer survivors. This finding implies that OGG1 may influence prognosis of lung adenocarcinoma. To date, there have been no studies on the association between this polymorphism and cancer prognosis. The suspected prognostic effect should be examined prospectively in survival analysis adjusting for other prognostic factors.
In this study population, 70 (50.7%) of 138 patients with adenocarcinoma were women, and this proportion was similar to that for the whole patients with lung adenocarcinoma diagnosed at Aichi Cancer Center Hospital (46.8% of 828 patients between 1984 and 2000). The corresponding proportion was higher at Aichi Cancer Center than that reported in Japan V. Therefore, the high proportion of women with adenocarcinoma in this study was reflected in the sex distribution at Aichi Cancer Center Hospital.
In the study by Sugimura et al.22), the significant association between the Cys/Cys genotype of OGG1 and lung cancer risk was observed only for squamous cell carcinoma (n=118), and a slightly elevated insignificant OR for adenocarcinoma of the lung (n=78) was also documented. Confirming the findings for the adenocarcinoma was required. Our finding for the recent cases was very similar in the size of OR to their results. Our study was also too short to confirm the association, but added information to the finding observed by the previous study. Although different histological types of lung adenocarcinoma may be associated differently with this polymorphism, neither this study nor previous studies were large enough to examine the possible heterogeneous association.
Biological evidence supporting the association with lung carcinoma risk is still limited. Previous reports showed that loss of hetrozygosity (LOH) of 3p, which includes OGG1 gene, was very common in lung cancer tissues 21, 24, 25) Although mean 8-hydroxyguanine levels in lung tissue was higher with LOH than without LOH, the genotype of OGG1 Ser326Cys was not associated with the level, but a polymorphism of glutathione peroxidase I gene also located in 3p was correlated 33). Recently an association between lung adenocarcinoma risk and new OGGI polymorphisms at exon 1 has been reported for Japanese 20) Further biological studies on these OGG1 polymorphisms, especially in relation to environmental exposures, will be required to elucidate the biological roles in the carcinogenesis.
Currently, adenocarcinoma of the lung is considered to be associated with smoking, though the relative risk is lower than of squamous cell and small cell carcinoma [34][35][36]. Since high levels of 8-hydroxyguanine have been detected in smokers' lung tissue and leukocytes 37), the interaction with the OGG1 genotype was to be examined. In our study, the OR of the genotype was not different between never smokers and current smokers, and the difference in the OR of smoking was not large among the genotypes; 1.89 for Ser/Ser, 0.92 for Ser/Cys, and 2.47 for Cys/Cys. This suggests that OGG1 Ser326Cys polymorphism-smoking interaction was not marked in adenocarcinoma of the lung enough to be detected in our study. In the past studies, the possible interaction with smoking was reported for CYP1A1 and GSTM1 38,39). Our previous study found the interaction between L-myc L/S polymorphism and smoking for esophageal cancer 40).
In conclusion, this study suggested that the association between OGGI Ser326Cys polymorphism and the risk of lung adenocarcinoma was limited, if any, but that the Cys/Cys genotype might be associated with the poor prognosis. In this study, no difference in the OR of smoking among the genotypes was observed. The sample size was not enough for detailed analysis, but this was the largest study on the association with lung adenocarcinoma so far.