Manganese Superoxide Dismutase Alanine-to-Valine Polymorphism at Codon 16 and Lung Cancer Risk

Manganese superoxide dismutase (MnSOD) catalyzes the dismutation of a specific type of reactive oxygen species, superoxide radicals, into hydrogen peroxide and oxygen (1). Accumulation of reactive oxygen species can damage DNA, proteins, and lipids, leading to the initiation or promotion of cancer (2,3). MnSOD, the only known superoxide scavenger in mitochondria, may be particularly important for antioxidant defense because mitochondria are the major sites for cellular metabolism and hence production of reactive oxygen species (4). The signal sequence is essential for correct transport and processing of proteins by mitochondria (5). Indirect evidence suggests that the alanine-to-valine polymorphism at codon 16 (Ala16Val) in the signal sequence of MnSOD (5), also described as the –9 position (6), produces a conformational change in the helical structure of the protein. This change may decrease the efficiency of transport into mitochondria for the Val isoform of the protein (6,7), although other studies (8,9) support alternate and opposing functional effects of this polymorphism. The Ala16Val MnSOD polymorphism is common among Caucasians [with a frequency of 41%–55% (8–11)]. In the present study, we evaluated the association of the common Ala16Val MnSOD polymorphism with the overall risk of lung cancer in a large Caucasian population. We hypothesized that individuals with the Val allele, associated with less efficient MnSOD enzyme transport into the mitochondria, have an increased susceptibility to lung cancer because of the more rapid accumulation of cancer-causing reactive oxygen species. As part of an ongoing, hospital-based, lung cancer case–control study initiated in 1992 at the Massachusetts General Hospital, Boston, MA, 2783 participants were enrolled as of September 2000. The study was approved by the Institutional Review Board at both the Massachusetts General Hospital and the Harvard School of Public Health, Boston, MA.1 This study population has been described in detail previously (12–16). DNA was available, and the MnSOD polymorphism was genotyped successfully for 2560 (92%) participants.2 Demographic characteristics for those with genotype data were similar to those of the entire study population. A modified version of the standardized American Thoracic Society respiratory questionnaire (17) was administered by a trained research nurse, who obtained information on demographics, medical and family cancer histories, and detailed smoking and work histories (12–16). Peripheral blood specimens were collected for all participants. To reduce potential variation in allele frequency by ethnicity, we analyzed only Caucasians (93% of our genotyped study population) with complete information on sex, age, and smoking variables (smoking status, pack-years, and years after smoking cessation). The study population consisted of 1101 case patients and 1239 control subjects. Formal tests of statistical interaction were conducted by use of the likelihood ratio test comparing nested models with and without the interaction terms of interest. All statistical tests were two-sided. Population differences between case patients and control subjects reflect typical risk factors for lung cancer (Table 1). Compared with control subjects, case patients were more likely to be older, male, and current smokers and to have a higher cumulative exposure to tobacco smoke and a family history of lung cancer but were less likely to be college graduates. Among control subjects, the frequency distribution of the MnSOD genotype was in Hardy–Weinberg equilibrium (χ2 goodness of fit, P = .89). This variant Val allele frequency of 48.6% was comparable to the frequency reported in another U.S. Caucasian population (8). The genotype distribution in our study was statistically significantly different between case patients and control subjects, where Val/Val was observed in more case patients and Ala/Ala was observed in more control subjects (Pearson χ2, P = .02). Table 2 presents the association of MnSOD genotype with lung cancer risk. In the crude analysis, we observed a statistically borderline increased lung cancer risk for heterozygotes (odds ratio [OR] = 1.16; 95% confidence interval [CI] = 0.95 to 1.42) compared with individuals homozygous for the wild-type Ala allele, whereas the Val/Val genotype was associated with a statistically significantly increased risk (OR = 1.40; 95% CI = 1.11 to 1.76). After adjustment for potential confounders, the association between MnSOD genotype and lung cancer was strengthened. Both Ala/Val and Val/Val genotypes were associated with statistically significantly increased risks (OR = 1.34 [95% CI = 1.05 to 1.70] and OR = 1.67 [95% CI = 1.27 to 2.20], respectively) and exhibited a statistically significant linear trend (P<.001). The magnitude and direction of risk associated with the Ala/Val or Val/Val genotypes, when stratified by sex, histologic type, or disease stage, were similar to those observed in our overall population (Table 2). Although younger individuals (aged ≤55 years) appeared to have higher risks with the Ala/Val and Val/Val genotypes compared with risks in older individuals (aged >55 years), these differences were not statistically significant (P = .15 for heterozygotes; P = .92 for Val/Val). Among nonsmokers, there was a statistically nonsignificant increased risk of cancer for heterozygotes (adjusted OR [ORadj] = 1.55; 95% CI = 0.81 to 2.95) and no association with the Val/Val genotype (ORadj = 0.95; 95% CI = 0.43 to 2.08). Variability in these estimates, however, was relatively large because of the small numbers of nonsmoking case subjects. In ever smokers, although the MnSOD genotype association in heterozygotes (ORadj = 1.31; 95% CI = 1.01 to 1.70) was similar to that in nonsmokers (P = .06), the Val/Val (versus Ala/Ala) genotype association (ORadj = 1.84; 95% CI = 1.36 to 2.49) was statistically significantly different from the OR in nonsmokers (P = .01). We observed no obvious pattern for the genotype effect among increasing tertiles of pack-years in ever smokers (data not shown). Among our control subjects, we observed a statistically significant but weak inverse correlation (Spearman rank correlation coefficient = –.06; P = .03) between increasing quartiles of pack-years (including nonsmokers) and increasing number of Val alleles. Given that multiple comparisons in subgroups may have led to chance findings, further evaluation of possible effect modification by smoking is needed. Our findings of an increased risk of lung cancer associated with the Val allele (corresponding to the lower efficiency of MnSOD transport into mitochondria) may be explained by decreased removal of reactive oxygen species. This increased risk results from the accumulation of cell damage. MnSOD plays a key role in cell survival by providing resistance to oxidative injury after exposure to adverse conditions, such as exposure to radiation (18,19). Studies (20–22) have found lower MnSOD expression in various transformed cancer-like cell lines compared with untransformed cells, suggesting that a lower underlying MnSOD activity may increase the likelihood of carcinogenesis. In addition, MnSOD may have tumor suppressor activity. Studies that induced overexpression of MnSOD in a variety of tumor or transformed cell lines (23–27) found suppressed growth rates and plating efficiency, increased morphologic differentiation, and decreased tumor formation in mice. Less efficient underlying MnSOD activity conferred by the Val allele may be associated with decreased tumor suppressor activity and increased cancer risk. However, MnSOD expression may also be negatively regulated by tumor suppressor p53 (18), and further evaluation of the biologic mechanisms for MnSOD and the Val isoform in carcinogenesis is needed. In contrast to our results, the two previously published cancer studies evaluating this polymorphism (8,9) found a statistically significantly increased risk of breast cancer associated with the Ala allele. A U.S. study (8) found that risk in premenopausal women with the Ala/Ala genotype was four times greater than that in women with the Val/Val genotype. A Finnish study (9), however, observed a statistically significant association for the Ala allele mainly in postmenopausal women (OR = 1.7; 95% CI = 1.2 to 2.5). These differences between lung and breast cancer associations with MnSOD genotype suggest that the role of MnSOD in carcinogenesis may vary for different populations and different tumors (8,28–33). Although we did not use population-based control subjects, allele frequencies were representative of those in the general Caucasian population. As discussed in our previous reports (12–16), it is unlikely that our selection of control subjects (i.e., spouses or friends) is related to genotype. These results, however, may not be applicable to non-Caucasian populations. Although control subjects were not matched at enrollment on risk factors (e.g., age and sex), we controlled for important risk factors in our analysis by including them in adjusted models or stratifying on these covariates. In conclusion, we found statistically significantly increased risks of lung cancer for individuals heterozygous (ORadj = 1.34; 95% CI = 1.05 to 1.70) or homozygous (ORadj = 1.67; 95% CI = 1.27 to 2.20) for the MnSOD Val allele as well as a statistically significant gene dose–response effect with increasing risk for each additional Val allele. Distribution of demographic characteristics and potential risk factors for lung cancer among case patients and control subjects Distribution of demographic characteristics and potential risk factors for lung cancer among case patients and control subjects Distribution of manganese superoxide dismutase (MnSOD) genotype (case patients/control subjects) and odds ratios (ORs) and 95% cofidence intervals (95% CI)* for lung cancer (overall and stratified by risk factors for lung cancer) Distribution of manganese superoxide dismutase (MnSOD) genotype (case patients/control subjects) and odds ratios (ORs) and 95% cofidence intervals (95% CI)* for lung cancer (overall and stratified by risk factors for lung cancer) We thank the following staff members of the Lung Cancer Susceptibility Group, Harvard School of Public Health, Boston, MA: Linda Lineback, Barbara Bean, Jeanne Jackson, Andrea Solomon, Lucy Ann Principe, Salvatore Mucci, Richard Rivera-Massa, Stephanie Shih, Dr. Wei Zhou, and Dr. Lilian Xu for patient recruitment, data collection, and laboratory assistance.