The Laryngoscope Lippincott-Raven Publishers, Philadelphia 0 1997 The American Larvngolorncal. Rhinological and Otological soci&y, Inc. Genetic Susceptibility to Head and Neck Cancer: Interaction Between Nutrition and Mutagen Sensitivity Stimson P. Schantz, MD; Zuo-Feng Zhang, MD, PhD; Margaret S. Spitz, MD, MPH; Ming Sun, MS; Tao C. Hsu, PhD The development of head and neck cancer may depend not only on exposure to environmental carcinogens but also on a genetically based susceptibility to carcinogen-induced damage. This thesis presents a case-control study that demonstrates the significance of mutagen sensitivity, a measure of an individual’s intrinsic DNA repair capacity against free radical damage, as a risk factor for the disease. As part of the case-control analysis, 167 previously untreated patients and 177 age- and sex-matched healthy controls were assessed for various lifestyle factors including tobacco and alcohol habits, occupational exposures, and diet. Mutagen sensitivity expressed by each individual was determined by quantifying bleomycin-induced chromosomal breaks within peripheral blood lymphocytes in vitro. Consistent with our initial observations and those of others, mutagen hypersensitivity was strongly associated with increased risk of head and neck cancer (odds ratio, 4.95;95% confidence interval, 2.67 to 9.17) after adjusting for age, sex, and race. Low intake of vitamins C and E was also associated with an increased risk of disease and was interactive with mutagen sensitivity in risk estimates. Individuals with both a low intake of various antioxidants and increased chromosomal sensitivity to oxidant-induced DNA damage were at greatest risk. This study supports the concept that the risk of head and neck cancer is determined by a balance of factors that either enhance or protect From the Head and Neck Surgery Service (S.P.S.) and the Epidemiology and Biostatistics Service (Z-F.Z.), Memorial Sloan-Kettering Cancer Center, New York, New York and the Departments of Epidemiology (M.S.S.) and Cell Biology (T.C.H.), M.D. Anderson Cancer Center, Houston, Texas. Presented a s a Candidate’s Thesis to the American Laryngological, Rhinological and Otological Society, Inc. This study was supported by grant no. R01 CA51845 from the National Cancer Institute. Send Reprint Requests to Stimson P. Schantz, MD, Memorial SloanKettering Cancer Center, 1275 York Avenue New York, NY 10021, U.S.A. Editor’s Note: This manuscript received the Harris P. Mosher Award. Laryngoscope 107: June 1997 against free radical oxygen damage, including innate capacities for DNA repair. Laryngoscope, 1OR765-781,1997 INTRODUCTION Head and neck carcinogenesis may depend not only on environmental exposures but also on an increased susceptibility to those exposures-an interaction between’both exogenous carcinogens and endogenous host influences. External factors include levels of carcinogen exposure and other potential coetiologic agents, such as alcohol and the human papilloma virus.1-3 External factors that serve to diminish cancer risk, such as various nutritional substances, may also be involved.4 The host factor that forms the principal focus of this report is the individual’s innate capacity for DNA repair.5 In 1983, Hsu6 postulated that chromosomal fragility, a measure of DNA repair capacity, was not an all-or-none phenomenon. Rather, a gradient exists within the general population. Furthermore, fragility would not be expressed spontaneously, but hidden. Its existence would become apparent only after clastogenic exposures. Hsu postulated that the individuals who expressed the greatest degree of mutagen-induced chromosome fragility would be at greatest risk for cancer development. Such a process would also account for the site of disease development. Only tissues exposed to mutagens would be at risk. Thus the concept of susceptibility to mutageninduced chromosomal breakage would account for diseases such as colon cancer, lung cancer, and head and neck cancer. It would not explain other neoplastic processes in which environmental exposures are not a principal feature, such as breast cancer, sarcomas, and central nervous system tumors. To establish the relevance of chromosome fragility (i.e., mutagen sensitivity) as a risk factor Schantz et al.: Genetic Susceptibility to Cancer 765 for disease development, Hsu et al.7 developed the bleomycin-sensitivity assay. Bleomycin had long been known as a clastogenic agent. The mechanism of its action had generally been considered to be radiomimetic (i.e., through the generation of free oxygen radicals), a fact relevant to tobacco carcinogenesis because numerous compounds within tobacco condensate may generate free oxygen radicals capable of producing DNA double-strand breaks and subsequent mutations.8-10 The assay they developed (which forms the basis of this report) involves the use of peripheral blood lymphocytes. In brief, lymphocytes are obtained from donor subjects and cultured for 5 days in the presence of the mitogen phytohemagglutinin. During the last 5 hours of the culture, cells are exposed to the clastogen bleomycin, and chromosomal breaks are generated. To quantitate the breaks, cells are arrested in mitosis using Colcemid. The number of chromosomal breaks per cell are then quantitated following analysis of 50 cellular metaphases. The mechanisms that account for differences in chromosomal sensitivity to bleomycin among individuals are undoubtedly multiple, including susceptibility to initial damage as well as repair. 11J2 The first study to evaluate mutagen sensitivity as a potential risk factor in head and neck carcinogenesis involved the assessment of 46 patients and 39 healthy controls.13 A statistically significant increase in bleomycin-induced breaks was noted in the head and neck cancer patients. Seventy-two percent of the patients were noted to be hypersensitive to bleomycin as compared with 28% of the controls. The greatest differences in sensitivity to bleomycin were observed when comparing individuals with greater than 40-pack-per-year cigarette exposure who remained disease free versus the head and neck cancer patients with similar levels of exposures. Similar findings regarding the resistance of elderly individuals without a history of tobacco-induced disease had been previously reported by Hsu14 and Hsu et al.15 The difference between a smoker without cancer and one with head and neck cancer is that the latter individual is abnormally sensitive to mutagens. Several reports followed that investigated mutagen sensitivity expressed by head and neck cancer patients.16-20 In each case similar findings were noted. Head and neck cancer patients expressed increased sensitivity to mutagens. Furthermore, relevant to the conduct of the present casecontrol study, an interaction between tobacco exposure, mutagen sensitivity, and head and neck cancer could be identified. Individuals who both smoked and were mutagen sensitive were at greatest risk. Further evidence supporting the biologic and clinical significance of mutagen sensitivity was provided by a prospective longitudinal follow-up study of assessed head and neck cancer patients in which 84 Laryngoscope 107: June 1997 766 patients were evaluated prior to the treatment of their index disease.21A wide range of mutagen-sensitivity values were noted, with patients expressing both resistance and sensitivity to mutagen-induced damage. It was postulated in this study that the risk of second primary malignancies would be greatest in the most sensitive individuals. The inherent inability to repair DNA damage would increase the carcinogenic impact of tobacco exposure. The probability that neoplastic transformation would involve multiple cells would be increased. In support of this concept it was noted that head and neck cancer patients who were hypersensitive to mutagens expressed a 4.4-fold higher risk for multiple cancers than more resistant individuals in this prospective analysis.21 Subsequent studies by Cloos et al.19 and Spitz et a1.22 have provided supportive evidence for this association. RATIONALE FOR PRESENT CASECONTROL ANALYSIS The purpose of the present study was to establish through a standard case-control design the significance of mutagen sensitivity as an independent risk factor for disease. It was to support original observations cited above, as well as t o extend those observations by looking for interactions with other lifestyle factors that may account for disease. A major focus of this presented case-control study is the interaction of diet, mutagen sensitivity, and head and neck disease. Numerous studies have indicated the role of diet in either promoting or preventing disease progression. The majority of these studies have pointed to the protective effect of fruits and vegetables (reviewed in Schottenfeldl and Wind). It is postulated that this latter beneficial influence is exerted through its antioxidant affect (i.e., the provision of free radical oxygen scavengers). High levels of such scavengers, including various vitamins such as vitamins C and E, would prevent electrophilic genotoxic-induced damage. The in uitro effect of dietary free radical scavengers on bleomycin-induced chromosomal damage had been previously rep0rted.23~24In these latter studies, lymphocytes from head and neck cancer patients were cultured in uitro and exposed to bleomycin. Varying concentrations of free radical scavengers, including vitamins A, C, and E, were individually tested for their ability to block chromosomal breakage. In each case, increasing levels of the respective scavengers inhibited bleomycin-induced damage. We hypothesized in this case-control analysis that a similar process would occur i n uiuo. Namely, fkee radical oxygen damage represents a basis for cancer initiation. The overall probability of developing disease would reflect a balance between environmental exposure, nutritional habits that may protect against that exposure, and an inherited susceptibility to DNA damage. The significance of such an epidemiologic study are twofold: to provide Schantz et al.: Genetic Susceptibilityto Cancer insight into further scientific investigations which provide a mechanistic basis for head and neck cancer development and to provide a basis for targeting prevention efforts to those at greatest risk. MATERIALS AND METHODS Cases All patients entered into this study were diagnosed with histologically confirmed squamous cell carcinoma of the upper aerodigestive tract and were seen a t Memorial Sloan-Kettering Cancer Center. Sites of disease were classified by the American Joint Committee on Cancer criteria and consisted of oral cavity, pharynx, and larynx. All patients had been previously untreated. A total of 167 patients were identified between the years 1992 and 1994. Controls One hundred seventy-seven age- (within 5 years) and sex-matched controls were identified for this study. Controls were without previous history of cancer and were identified within the Blood Bank Center of Memorial Sloan-Kettering Cancer Center. Data Collection Each patient and control was asked to complete a risk factor questionnaire. The questionnaire identified the following variables: 1. demographic information: gender, race, birth year, place of birth, religion, and education; 2. dietary factors: ingestion of fat, vegetables, fruits, and supplements such as vitamins C, E, and A; 3. occupational and environmental exposures; 4. family history of cancer; 5. sexual behavior; 6. personal habits: cigarette smoking, years of smoking, age a t initiation of smoking, alcohol consumption, types of alcohol, frequency and quantities of alcohol consumption, years of alcohol ingestion; and 7. medical and oral hygienic history. The validation of this self-administered, comprehensive cancer risk factor questionnaire has been previously reported.25 Mutagen-Sensitivity Assay The assay used in this study has been described in detail previously.15 A peripheral blood sample was collected from each donor in a heparinized tube prior to the initiation of lymphocyte culture. The standard lymphocyte culture procedure used RPMI-1640 medium, supplemented with 15% fetal calf serum and phytohemagglutinin in a ratio of blood to medium of 1:9.At 67 hours of incubation, one set of cultures is treated with bleomycin (0.03 U/mL) for 5 hours. Colcemid (0.04 mg/mL) is added in the last hour to induce mitotic arrest prior to harvesting. Conventional cell harvesting procedure follows: the cells are treated with hypotonic KC1 (0.06M KC1) solution for 15 to 20 minutes, fixed, washed with a freshly prepared mixture of methanol and acetic acid (3:l) and air dried on wet slides. The slides are stained with Giemsa solution without banding. Fifty well-spread metaphases are examined from coded slides. Chromatid aberrations are recorded as frank chromatid breaks or exchanges. Bleomycin tends to induce few chromatid exchanges (which if present, are considered as two breaks). Chromatid gaps or attenuated regions are Laryngoscope 107:June 1997 disregarded. The frequency of breakage was expressed as breaks per cell (b/c) for purposes of comparison. The reliability of cytogenetic scoring has previously been evaluated by comparing four separate blood samples from a respective donor with a minimum interval between samples of 1 week.15 Using a random-effect, one-way analysis of variance model, significance within group variation was noted, suggesting that sensitivity appeared to be stable and representative. Nutrient Estimates A modified National Cancer Institute Health Habits and History Questionnaire26-28 including food items, frequency of consumption, usual portion size, and a restaurant use section was employed to collect usual dietary history. The estimates of nutrients were calculated by the National Cancer Institute’s algorithm.26.27 In brief, nutrient indices were obtained from the Second National Health and Nutrition Examination Survey29 conducted from 1976 to 1980, which was based on U.S. Department of Agriculture food composition computer data tape,30 as well as industry and other sources. Nutrient composition values for vitamin A are based on Revised Handbook 8 values.31 The food database includes three parts: portion sizes, food in grams, and nutrients per 100 g for each food item. Each nutrient was calculated by the following equation: (reported food frequency x gram portion size x the nutrient contentI100 g x seasonality factor) +loo. Nutrients were then summed over all food and beverage items, including restaurant foods, on the questionnaire to obtain the average intake per day. During calculation, vegetables and fruits were adjusted for seasonal consumption by multiplying the computed yearly intake of each food item by the seasonal index (proportion of year during which seasonal consumption of food occurred). The restaurant food frequency was compared with the frequency of the corresponding food item in the main food list, and appropriate adjustment was made before the calculation of the estimates according to National Cancer Institute’s algorithm.26.27 Median values of each nutrient were used for subjects with missing information on the dietary section (a total of 21 subjects including seven cases and 14 controls were missing). A Statistical Application System program was developed to calculate nutrients in this study. Statistical Analysis Uniuariate analysis. The relationship between risk of head and neck cancer and putative risk factors such as mutagen hypersensitivity; intake of dietary vitamins C, E, and carotenoids; cigarette smoking; alcohol drinking; and body mass index (BMI) were measured using the odds ratios (OR) and their 95% confidence interval (95% CI), derived from logistic regression analysis.32 Variables of nutritional vitamin intakes were divided into quartiles according to their distribution. Dummy variables were employed t o estimate the OR for each category of exposure in logistic regression analysis. The presence of a trend for ordered variables was initially assessed by Mantel-Haenszel extension methods and then performed by assigning the score j to the jth exposure level of a categorical variable and treating it as a continuous variable in unconditional logistic regression.33 Schantz et al.: Genetic Susceptibilityto Cancer 767 Interaction assessment. Stratified analysis was used to assess possible additive interaction and, if no interaction was found, to control for potential confounding effects and to obtain a summary statistics by using the Mantel-Haenszel method.33 The test of significance for additive interaction between two risk factors was performed by methods described by Rothman34 and Daya1.35 The data were stratified and compared between various subgroups against one standard reference group. The standard reference group was usually selected to be the group with minimal level of exposure on all risk factors for head and neck cancer.36 For example, to assess the additive interaction between cigarette smoking and mutagen sensitivity on cancer risk, cases and controls were stratified into four groups according to smoking (yeslno) and mutagen sensitivity ( 4and 21):nonsmoking and mutagen sensitivity d, smoking and mutagen sensitivity d,nonsmoking and mutagen sensitivity 21, and smoking and mutagen sensitivity 21. For each of four combinations of smoking and mutagen sensitivity, the odds of disease can be calculated by dividing the number of cases over the number of controls. To calculate the OR for each category, a reference group with minimum exposure was chosen (nonsmoking and mutagen sensitivity <l); then the odds of disease of each group was compared with the odds of disease of the reference group. The synergy index (S) was calculated and a chi-squared test with one degree of freedom was employed to test the significance of synergy.34~35 Multiplicative interaction was assessed by unconditional logistic regression model for mutagen sensitivity and other risk factors that were significant in univariate analysis. A product term of mutagen sensitivity and each potential risk factor was created and introduced into the multivariate model with both independent variables. Multivariate analysis. All variables that were significant in univariate analysis were entered into a multivariate logistic model after controlling for potential confounding effects such as age, sex, race, and education. Odds ratios and their 95%CIS were estimated and trend tests were performed. RESULTS Demographic characteristics of both the cases and controls are shown in Table I. Overall, no differences in age, race, sex, or place of birth could be identified between cases and controls.The mean ages were 60.9 years (SD k12.37) for cases and 58.4 years (SD 43.63) for controls (Student’s t-test, P > 0.05). Similar distributions were found for cases and controls for gender and race. The cancer population was less likely to be college educated (P c 0.01). Likewise, the median income of the cancer patients was lower than the controls, but this did not reach statistical difference (P = 0.05). The above data suggests that cases and controls were comparable in terms of age, sex, and race in this study. The association between selected factors and the risk of head and neck cancer was analyzed after controlling for potential confounding effects such as age, sex, and race (Table 11).Current tobacco use was associated with a significantly increased risk of head Laryngoscope 107:June 1997 768 and neck cancer (OR = 6.22; 95% CI = 3.55 to 12.35) (P = 0.0001). A strong dose-response relationship was observed between cigarettes per day and the risk of head and neck cancer: the ORs were 2.19 (95% CI = 1.23 to 3.92) for those who smoked one to 20 cigarettes per day, and 4.62 (95% CI = 2.46 to 8.69) for those smoked more than 20 cigarette per day (trend test, P = 0.0001). Of those who had stopped using tobacco for at least 5 years prior to diagnosis, the OR of head and neck cancer was only marginally elevated (OR = 1.16; 95% CI = 0.62 to 2.17 (P > 0.05).The median duration of smoking cessation for cancer patients was 3 years (range, 1to 66 years) and that of smoking cessation for controls was 16.5 years (range, 1 to 53 years). Passive smoking was identified to be associated with an increased risk of head and neck cancer (OR = 3.14; 95% CI = 1.45 to 6.76).The use of alcohol was associated with increased risk of head and neck cancer in only those individuals with highest level of alcohol use (i.e., >lo0 drinks per month (OR = 6.22; 95% CI = 2.83 to 13.65) (trend test, P = 0.001). Increased body weight and BMI were significantly associated with a decreased risk of head and neck cancer (trend tests, P = 0.0005 and P = 0.0302, respectively) (Table 11). Dietary nutrients were examined after controlling for age, sex, race, and calorie intake. No obvious association was identified for total dietary protein, fat, or carbohydrates after controlling for age, sex, race, and total intake of calories. High intakes of vitamins C and E were related to a decreased risk (trend tests, P = 0.0049 and P = 0.0093, respectively). Although dietary beta-carotene was not associated, high intake of other specific carotenoids such as cryptoxanthin and lycopene were related to a decreased risk (trend tests, P = 0.0004 and P = 0.0379, respectively) (Table 111).No significant association was observed between vitamin supplement (A, C, and E) and risk of head and neck cancer, although the ORs for total intake of vitamins C and E were protective. Additional analysis on food items showed that risk of head and neck cancer was significantly increased with high consumption of the following food items: liverwurst, eggs, ice cream, beef, liver, pork, and fried chicken; the risk was decreased with high intake of fruits and vegetables such as apples, tomatoes, and green salad (data not shown). An analysis was performed of mutagen sensitivity as a risk factor for head and neck cancer development. Table IV provides risk estimates after adjusting for age, sex, and race. Various cutoff points were examined including the simple division of the control population into approximately four equal groups. In each instance the highest risk was associated with a breaks per cell (b/c) value 21 (OR = 5.99; 95% CI = 3 to 11.99). Previous case-control studies have used ~ 0 . 8b/c as a cutoff value between mutagen-sensitive and mutagen-nonsensitive indiSchantz et al.: Genetic Susceptibility to Cancer TABLE I. Demographic Characteristics of Cases and Controls. Factors Cases (n = 167) No. Yo Controls (n = 177) No. % P value Age (Y) c40 40-49 50-59 60-69 270 * Mean SD Sex Male Female Race White Non-white Education <High school College >College Place of birth United States Elsewhere Religion Catholic Jewish Baptist Other Protestant Seventh-Day Adventists Other religions None Family income ($/year) ~$13,000 $13,000-$22,999 $23,000-$32,999 $33,000-$42,999 $43,000-$52,999 $53,000-$62,999 1$63,000 Mutagen sensitivity test No Yes Site Oral cavity Pharynx Larynx Nasal cavity Unknown Laryngoscope 107: June 1997 7 22 48 50 43 60.9 i 12.37 4.1 12.9 28.2 29.4 25.3 5.84i 13.63 19 24 43 56 35 10.7 13.6 24.3 31.6 19.8 0.14 107 57 65.2 34.8 111 63 63.8 36.2 0.781 146 19 88.5 11.5 157 16 90.8 9.2 0.494 102 47 18 61.1 28.1 10.8 49 90 39 27.5 50.6 21.9 0 139 31 81.8 18.2 156 23 87.2 12.8 0.164 104 15 8 30 62.3 9 4.8 18 91 38 4 32 51.1 21.4 2.3 18 1 7 2 0.6 4.2 1.2 1 6 6 0.6 3.4 3.4 0.03 13 17 27 18 12 13 34 9.7 12.7 20.2 13.4 9 9.7 25.4 8 9 21 20 21 19 57 5.2 5.8 13.6 12.9 13.6 12.3 36.8 0.053 76 91 45.5 54.5 46 131 79 25 42 1 3 52.7 16.7 28 0.7 2 26 74 Schantz et al.: Genetic Susceptibility to Cancer 769 TABLE (I. Analysis of Selected Risk Variables. Variables Cigarette smoking No Yes Quit Current+ No./day 1-20 >20 Cases (n) Controls (n) Odds Ratios' 95% CI 27 138 41 97 62 112 72 40 1 2.9 1.16 6.22 1.7-4.95 0.62-2.17 3.55-12.35 69 67 72 35 2.19 4.62 P = 0.0001 1.23-3.92 2.46-8.69 20 68 42 46 40 15 1.05 3.89 6.72 P = 0.0001 0.51-2.16 2.1-7.19 3.06-1 4.7 10 156 29 145 27 41 33 61 36 63 51 16 1 0.88 0.99 6.22 P = 0.0001 0.45-1.72 0.49-1.99 2.83-1 3.65 52 67 26 18 40 41 42 40 1 1.17 0.43 0.35 P = 0.0005 0.61-2.25 0.2-0.91 0.15-0.77 56 41 40 26 40 40 39 41 1 0.66 0.68 0.46 P = 0.0302 0.35-1.24 0.36-1.28 0.23-0.89 Trend test Years of smoking 1-20 21-40 >40 Trend test Passive smoking No Yes Alcohol use (drinksho) No 1-30 31-100 >loo Trend test Usual body weight (kg) 165.8 65.9-80.8 80.9-93.1 293.2 Trend test Body mass index 523.2 23.3-25.8 25.9-29 229.1 Trend test 1 3.14 1.45-6.76 'Odds ratios were adjusted for age, sex, and race. +Current smokers included smokers who had quit less than 5 years previously. viduals. Using this latter cutoff point, the OR of head and neck cancer was 3.97-fold higher in the mutagen-sensitive population (95% CI = 2.11 to 7.47). We also divided cases and controls into three categories: ~ 0 . 8b/c, 0.8 to e l b/c, and 2 1 b/c. The ORs were 1.75 (95% CI = 0.75 to 4.08) for individuals with mutagen sensitivity between 0.8 and 1b/c and 5.99 (95% CI, 3 to 11.99) for those whose mutagen sensitivity was 1 b/c or higher (trend test, P=O.OOOl). We then compared individuals with a mutagen sensitivity value of 1 b/c or higher with those with a value lower than 1b/c. The OR was 4.95 (95% CI = 2.67 to 9.17). Laryngoscope 107:June 1997 770 The associations between mutagen sensitivity and the above-mentioned risk factors including cigarette smoking, alcohol drinking, and dietary vitamin C, E, cryptoxanthin, and lycopene intake, were assessed separately in both cases and controls, but no statistically significant association was observed (Table V). These results suggest that intake of various nutrients, tobacco, and alcohol did not influence individual mutagen sensitivity values. Table VI shows risk estimates for additive interactions between cigarette smoking; alcohol use; and dietary vitamin C, E, cryptoxanthin, lycopene intake; and mutagensensitivity We used >1b/c as the cutoff Schantz et al.: Genetic Susceptibilityto Cancer TABLE Ill. Analysis of Dietary Vitamin Intakes* Variables Odds Ratios" Cases (n) Controls (n) 95% CI 57 37 40 36 45 43 47 44 52 45 20 53 45 45 45 44 27 55 26 62 45 45 45 44 32 60 30 48 45 45 50 39 1 1.75 0.67 1.3 P = 0.7988 0.94-3.25 0.34-1.31 0.67-2.55 37 43 26 64 45 45 49 40 1 0.96 0.46 1.39 P = 0.6221 0.51-1.81 0.23-0.91 0.73-2.67 61 39 40 30 45 43 47 44 54 34 25 57 45 45 47 42 70 26 22 52 45 45 45 44 1 0.34 0.26 0.6 P = 0.0379 44 45 1 Vitamin C (mg) <91.5 915141.4 141.5-175.2 >175.2 Trend test Vitamin E (a-TE) <5.4 5.4-7.7 7.8-8.6 28.6 Trend test Vitamin A (IU) <5256 5256-8622 8622-10565 210565 Trend test Alpha carotene (pg) <181.9 181.9-426.2 426.3-609.5 2609.5 Trend test Beta carotene (pg) ~2064 2064-3503 3503-4421 24421 Trend test Cryptoxanthin (pg) <61.4 61.4-1 15.3 115.3-148 2148 Trend test Lutein (pg) 4678 1678-3333 3333-3732 >3732 Trend test 1 0.68 0.48 0.4 P = 0.0049 1 0.54 0.2 0.45 P = 0.0093 0.37-1.27 0.26-0.9 0.2-0.8 0.28-1.03 0.09-0.43 0.21-0.95 1 1.89 0.68 1.64 P = 0.6309 0.99-3.59 0.33-1.41 0.83-3.26 1 0.64 0.45 0.3 P = 0.0004 0.35-1.18 0.24-0.84 0.15-0.6 1 0.58 0.34 0.95 P = 0.5708 0.31-1.08 0.17-0.66 0.52-1.73 Lycopene (pg) <910.7 910.7-1509.4 1509.4-1898.4 21898.4 Trend test Retinol (pg) <323.9 0.18-0.64 0.13-0.51 0.32-1.I 1 (Continues) Laryngoscope 107: June 1997 Schantz et al.: Genetic Susceptibility to Cancer 771 TABLE 111. (Continued) Analysis of Dietary Vitamin Intakes Variables 323.9-506.9 506.9-600.4 >600.4 Cases (n) Controls (n) Odds Ratios* 95% CI 35 28 63 45 45 44 0.68 0.5 1.07 P = 0.9622 0.35-1.3 0.26-0.99 0.55-2.08 32 48 30 60 45 45 45 44 1 1.37 0.68 1.39 P = 0.6582 Trend test Pro-A carotene (pg) ~2320 2320-3920 3920-4959 s4959 Trend test 0.73-2.57 0.35-1.36 0.73-2.67 'Odds ratios were adjusted for age, sex, race, and caloric intake. value to define hypersensitivity for this analysis. In nearly every instance an interactive effect was suggested ( S >O). The risk estimate was markedly higher when assessing for the presence of combined risk factors than for each risk factor alone. A n additive interaction effect was identified when evaluating the combined effect of cigarette smoking and mutagen sensitivity. The ORs were 5.34 (95% CI = 1.04 to 27.26) for individuals who had a positive test result for mutagen sensitivity (21 b/c) but were nonsmokers, 5.05 for individuals who had a negative test result for mutagen sensitivity (<1 b/c) but were smokers, and 24.92 (95%CI = 6.59 to 94.24) for those with a positive test result who were smokers (S = 1.41;x2 = 3.86; P < 0.05). Individuals who both consumed large amounts of alcohol (>lo0 drinks per month) and were hypersensitive to mutagens had an OR of 45.33 (95% CI = 9.25 to 222).The joint effects were also identified for mutagen sensitivity and passive smoking (OR = 12.11; 95%CI = 3.19 to 46.031, lower dietary vitamin C intake (OR = 15.6;95%CI = 5.56 to 43.711, lower vitamin E intake (OR = 13.69;95%CI = 4.79to 39.Q lower cryptoxanthin (OR = 12.05; 95% CI = 4.66 to 31.12), lower lycopene intake (OR = 8.06; 95% CI = 3.11 to 20.85), and lower BMI (OR = 11.11; 95%CI = 4.19 to 29.46).None of these fadors were identifled to have a signiscant multiplicative interaction with mutagen sensitivity on the risk of head and neck cancer (Table VII). In the multivariate risk model after including variables that were significant in the univariate analysis, the ORs were 4.58 (95%CI = 1.37 to 15.32) for cigarette smoking, 7.31 (95%CI = 2.36 to 22.67)for alcohol use, and 7.75 (95%CI = 2.95 to 20.35) for mutagen sensitivity. Lower dietary intake of vitamins was associated with increased risk but was not statistically significantly related to the risk of head and neck cancer (Table VII). Laryngoscope 107: June 1997 772 DISCUSSION The present case-control study demonstrates the significance of a quantitative analysis of mutagen sensitivity within head and neck cancer patients. Specifically, our data reveal that the OR of head and neck cancer is markedly elevated in individuals who express sensitivity to the clastogen bleomycin. The case-control study here confirms a previously performed case-control study that used a convenience sample as the control population.16 In that initial study, a subsequent study by Spitz et al.,17 and this present study an enhanced interaction was noted between tobacco use and mutagen sensitivity. Indeed, results among the three studies were nearly identical with ORs of head and neck cancer of 5.05 in individuals who smoked but were not mutagen sensitive, 5.34 in those who did not use tobacco but were mutagen sensitive, and 24.9 in those who both smoked and were mutagen sensitive. "he O h for head and neck cancer in the original convenience sample and in the subsequent case-control study by Spitz et al. were 19.8 and 23, respectively.lGJ7 Unlike these previous two studies, our analyses also adjusted for age, sex, and race, and used a higher cutoff ratio for hypersensitivity to mutagens, namely 1 b/c as compared with 0.8 b/c. The present case-control analysis is the first such analysis to identify a relationship between diet, mutagen sensitivity, and head and neck cancer. Patients in this study could be characterized by their low fruit and vegetable consumption as compared with the age- and sex-matched controls. The OR of head and neck cancer increased in the group of patients with the least consumption of these food stuffs.Similar observations were noted when calculations for specific nutrients were derived from the food frequency questionnaire. An increased risk of head and neck cancer was noted in individuals who had low intake of vitSchantz et al.: Genetic Susceptibility to Cancer TABLE IV. Mutagen Sensitivity and Risk of Head and Neck Cancer. Variables Cases (n) Controls (n) Odds Ratios* 95% CI Mutagen sensitivity (breakskell) <1 21 <0.8 0.8-<1 21 35 56 22 13 56 101 30 72 29 30 1 4.95 1 1.75 5.99 P = 0.0001 11 10 14 56 34 35 32 30 1 0.89 1.55 5.62 P = 0.0001 22 69 72 59 11 10 14 56 34 35 30 32 1 0.89 1.66 5.21 P = 0.0001 0.32-2.44 0.63-4.38 2.19-12.38 10 11 14 56 23 40 38 30 1 . 0.63 0.9 3.99 P = 0.0001 0.22-1.78 0.33-2.47 1.56-1 0.16 Trend test Quartile (1 fixed) <0.58 0.58-0.76 0.77-0.99 tl Trend test Mutagen sensitivity (breakskell) <0.8 20.8 1 3.97 2.67-9.17 0.75-4.08 3-1 1.99 0.32-2.46 0.59-4.08 2.35-13.44 2.11-7.47 Quartile 10.56 >O .56-<0.76 >0.76-50.96 >0.96 Trend test Quartile <0.52 20.52-<0.76 20.76-<1 21 Trend test *Odds ratios were adjusted for age, sex, and race. amins C and E as well as low intake of other lessunderstood nutrients such as cryptoxanthin. This latter agent is a retinoid precursor that functions to inhibit head and neck cancer cells by mechanisms distinct from beta-carotene or retinol palmitate.37 Free Radical Oxygen Damage and Its Repair The findings of this case-control study, as well as previous in vitro and epidemiologic studies, point to the relevance of free radical oxygen (FRO) damage and its repair as important determinants of head and neck carcinogenesis.l,4,23,24,38 FRO is generated by a variety of mechanisms, including both normal endogenous metabolism and various carcinogenic exposures (Fig. 1L39-43 Tobacco, for instance, has been noted to contain over 5000 compounds, many of which have oxidant capacity.42.44 Three important classes of FRO are produced on a daily basis within an individual and have the capacity to induce nearly 20,000 potentially genotoxic lesions per day: superoxides ( 0 2 - ) , singlet oxygen Laryngoscope 107:June 1997 and hydroxyl radicals (OH1.40-44 The majority are generated during the sequential reduction of oxygen to water. Oxidative damage to DNA mediated by these molecules is complex and involves both direct and indirect mechanisms.38 The latter may include inactivation of target enzymes that are involved in the synthesis of DNA precursor molecules. Direct FRO-induced DNA damage includes single- and double-strand breaks and the induction of apyrimidinic sites. It also includes the generation of numerous types of DNA adducts and resultant base modifications such as thymine glycol, 5-hydroxymethyluracil, and 8-hydroxy-2-deoxyguanosine.38 The ability to effectively repair FRO-induced genetic damage would thus be critical to the prevention of cell death andor mutational events. (l02), Relevant to our understanding of head and neck carcinogenesis, as well as to the interpretation of our present case-control results, numerous reports have demonstrated the capacity for tobacco-combustion products to induce FRO damage (reviewed in Schantz et al.: Genetic Susceptibility to Cancer 773 TABLE V. Distribution of Mutagen Sensitivity by Selected Factors and Disease Status. BreaksKell Cases Controls Smoking status, n (%) Current c0.8 0.k1 21 Former Nonsmoker 2 (22.2) 6 (24) 1 (11.1) 3 (12) 16 (64) 6 (66.7) x 2 = 0.463; P = 0.977 13 (24.1) 9 (16.7) 32 (59.2) Current Former Nonsmoker 18 (58.1) 8 (25.8) 5 (16.1) 30 (55.6) 9 (16.7) 15 (27.7) x 2 = 2.99; P = 0.56 21 (48.8) 12 (27.9) 10 (23.3) 1-20 None 34 (61.8) 9 (16.4) 12 (21.8) x 2 = 2.61 ; P = 0.626 21 (48.8) 12 (27.9) lO(23.3) Cigarettedday, n (YO) 1-20 120 <0.8 0.8-4 21 None >20 13 (50) 7 (26.9) 6 (23.1) 8 (20.5) 2 (22.2) 1 (11.1) 6 (15.4) 25 (64.1) 6 (66.7) ~2 = 0.694; P = 0.952 11 (25.5) 6 (15) 23 (57.5) Alcohol drinking, n (Yo) Yes <0.8 0.&<1 <1 No 11 (31.4) 8 (22.9) 36 (45.7) x2 Yes No 8 (615) 58 (53.2) 3 (23.1) 24 (22) 2 (15.4) 27 (24.8) x 2 = 0.381; P = 0.826 11 (21.2) 5 (9.6) 36 (69.2) = 5.26; P = 0.072 Dietary vitamin C intake, n (“/.) Low c0.8 0.841 21 Low High 15 (26.8) 10 (17.9) 31 (55.4) x2 7 (20) 3 (8.6) 25 (71.4) = 2.61 ; P = 0.271 High 36 (49.3) 19 (26) 18 (24.7) = 2.31 ; P = 0.316 36 (62.1) 10 (17.2) 12 (20.7) x’ Dietary vitamin E intake, n (7.) Low <0.8 0.841 21 Low High 17 (29.3) 9 (15.5) 34 (55.2) x2 5 (15.2) 4 (12.1) 24 (72.7) = 2.97; P = 0.227 High 36 (56.2) 36 (53.7) 14 (21.9) 15 (22.4) 14 (21.9) 16 (23.9) x 2 = 0.099; P = 0.952 Dietary cryptoxanthin intake, n (Yo) Low c0.8 0.841 t l Low High 14 (24.6) 9 (15.8) 34 (59.7) x2 8 (23.5) 4 (11.8) 22 (64.7) = 0.339; P = 0.844 High 36 (60) 36 (50.7) 10 (16.7) 19 (26.8) 14 (23.3) 16 (22.5) x 2 = 2.02; P = 0.365 Dietary lycopene intake, n (YO) Low c0.8 0.8-4 21 High 15 (28.3) 8 (15.1) 30 (56.6) x2 Laryngoscope 107: June 1997 774 7 (18.4) 5 (13.2) 26 (68.4) = 1.45; P = 0.483 Low High 34 (52.3) 38 (57.6) 15 (22.7) 14 (215) 13 (19.7) 17 (26.2) x 2 = 0.78; P = 0.676 Schantz et al.: Genetic Susceptibility to Cancer TABLE VI. 21 )* Risk Estimates for Additive Interactions Between Cigarette Smoking, Alcohol Use, and Mutagen Hypersensitivity (4, Variables Cigarette smoking No No Yes Yes Odds Ratio? Mutagen Sensitivity Cases (n) Controls (n) 95% CI No Yes No Yes 3 6 31 48 S = 1.41 1 5.34 5.05 24.92 P < 0.05 1.04-27.26 1.4-1 8.24 6.59-94.24 'x 33 10 65 20 = 3.86 Passive smoking No No Yes Yes No Yes No Yes 3 3 32 51 s = 1.21 18 5 81 25 x2 = 3.21 1 3.67 2.55 12.11 P > 0.05 0.55-24.6 0.69-9.38 3.19-46.03 Alcohol use No No Yes Yes No Yes No Yes 16 36 19 16 S = 5.13 1 6.74 11.3 45.33 P > 0.05 3.06-14.84 4.23-30.17 9.25-222.1 'x 82 27 11 2 = 0.72 Vitamin C intake High High Low Low No Yes No Yes 10 25 25 31 S = 0.83 55 18 46 12 x2 = 0.71 1 8.33 4.47 15.6 P > 0.05 3.21-21.64 1.73-1 1.57 5.56-43.71 Vitamin E intake High High Low Low No Yes No Yes 9 24 26 32 S = 0.44 51 16 50 14 x2 = 0.25 1 8.44 3.89 13.69 P > 0.05 3.15-22.61 1.46-1 0.41 4.79-39.1 Cryptoxanthin intake High High Low Low No Yes No Yes 12 22 23 34 S = 0.77 55 16 46 14 x2 = 0.73 1 6.08 2.97 12.05 P > 0.05 2.38-15.53 1.22-7.23 4.66-31.12 Lycopene intake High High Low Low No Yes No Yes 12 26 23 30 S = 0.59 1 5.52 1.68 8.06 P > 0.05 2.21-13.77 0.71-3.96 3.11-20.85 'x 48 17 53 13 = 0.42 1.82-16.12 1.62-10.87 6.97-56.67 Usual body weight High High Low Low No Yes No Yes 8 12 22 44 S = 1.65 51 14 38 15 x2 = 3.01 1 5.43 4.2 19.87 P > 0.05 Body mass index High No 9 48 1 (Continues) Laryngoscope 107: June 1997 Schantz et al.: Genetic Susceptibility to Cancer 775 TABLE VI. (Continued) Risk Estimates for Additive Interactions Between Cigarette Smoking, Alcohol Use, and Mutagen Hypersensitivity ( 4 ,21): Variables Mutagen Sensitivity Cases (n) High Low Low Yes 21 21 35 No Yes Controls (n) Odds Ratiot 14 39 15 S = 0.48 'x 7.4 3.02 11.11 = 0.29 95% CI 2.70-20.27 1.22-7.49 4.19-29.46 P > 0.05 'The cut-point for low/high for vitamins C and E, usual body weights and body mass index was the median value of the control group. 'For smoking, passive smoking, and alcohol use, odds ratios were adjusted for age, sex, and race; for nutritional factors and body mass index, odds ratios were adjusted for age, sex, race, and total calorie intake. Pryor and Stone42). The active oxygen within cigarette smoke condensate is generated mostly from polyphenols such as catechols and catechol-methyl derivatives, as well as hydroquinone, each of which occurs in abundance in cigarette smoke. Concrete evidence for the ability of cigarette smoke condensate to induce DNA single-strand breaks was demonstrated in the mid-1980s. Nakayama et a1.45 used the alkaline elution technique to measure DNA fragments. Large amounts of FRO could be generated by trapping tobacco smoke within phosphatebuffered saline. Exposure of DNA to the FRO-containing solution led to a measurable increase in DNA fragmentation. In support of the role of FRO as the damaging agent, DNA fragmentation could be inhibited by exposing the DNA to naturally occurring enzymes that have as their sole function the conversion of FRO to inactive species. Weitberg and TABLE VII. Risk Estimates for Multiplicative Interactions Between Cigarette Smoking, Alcohol Use, and Mutagen Sensitivity (el,>l)." Variables Odds Ratios P value 95% CI Smoking (yes vs. no) Mutagen sensitivity Interaction 5.05 5.34 0.93 1.4-1 8.24 1.04-27.26 0.16-5.42 Passive smoking (yes vs. no) Mutagen sensitivity Interaction 2.55 3.67 1.3 0.69-9.38 0.55-24.6 0.18-9.59 0.1606 0.18 0.7995 Drinking (100/movs. others) Mutagen sensitivity Interaction 11.3 6.74 0.6 4.23-30.17 3.06-14.84 0.1-3.75 0.0001 0.0001 0.5808 Vitamin C intake (low vs. high) Mutagen sensitivity Interaction 4.47 8.33 0.42 1.73-11.57 3.21-21.64 0.12-1.5 0.002 0.0001 0.182 Vitamin E intake (low vs. high) Mutagen sensitivity Interaction 3.89 8.44 0.42 1.46-10.41 3.15-22.61 0.12-1.51 0.0001 0.1819 Cryptoxanthin intake (low vs. high) Mutagen sensitivity Interaction 2.97 6.08 0.67 1.22-7.23 2.38-1 5.53 0.19-2.33 0.0168 0.0002 0.5271 Lycopene intake (low vs. high) Mutagen sensitivity Interaction 1.68 5.52 0.87 0.71-3.96 2.21-13.77 0.25-2.99 0.2339 0.0003 0.8217 Usual body weight (low vs. high) Mutagen sensitivity Interaction 4.2 5.43 0.87 1.62-10.87 1.83-16.12 0.22-3.42 0.0031 0.0023 0.8442 Body mass index (low vs. high) Mutagen sensitivity Interaction 3.02 7.4 0.5 1.22-7.49 2.70-20.27 0.14-1.83 0.0172 0.0001 0.2925 , 0.0136 0.0442 0.9319 0.0068 *For dietary vitamin intake, age, sex, race, and total calories were adjusted. For all other factors, age, sex, and race were controlled. Laryngoscope 107: June 1997 776 Schantz et al.: Genetic Susceptibility to Cancer TABLE VIII. Multivariate Logistic Regression Analysis of Risk Factors for Upper Aerodigestive Tract Cancer.' Variable Cigarette smoking (yes vs. no) Passive smoking (yes vs. no) Alcohol use (>lo0drinks/mo vs. others) Vitamin C intake (low vs. high) Vitamin E intake (low vs. high) Cryptoxanthin intake (low vs. high) Lycopene intake (low vs. high) Body mass index (low vs. high) Mutagen sensitivity (21 vs. 4) Age (Y) Sex (M/F) Race (Nonwhite vs. white) Education* Family Incomes Odds Ratio* 95% CI 4.58 6.06 7.31 2.48 2.02 1.17 1.14 2.09 7.75 1.03 1.87 2.84 1.03 0.8 1.37-15.32 0.91-40.22 2.36-22.67 0.71-8.71 0.65-6.28 0.39-3.52 0.42-3.11 0.87-5.05 2.95-20.35 0.99-1.07 0.66-5.31 0.56-14.27 0.50-2.11 0.62-1.05 *Total calorie intake was adjusted in the analysis. *Odds ratios were estimated when all variables were included in a multiple logistic regression model. *Education was defined as grade school = 1; some high school = 2; high school graduate = 3;technicalhrocationalschool certificate = 4;some college = 5;college graduate = 6; postgraduate degree = 7. §Incomewas defined as c$13,000/y = 1; $13,000-22,999= 2;$23,000-32,999= 3; $33,00042,999= 4;$43,000-52,999= 5;$53,00042,999 = 6;>$63,000= 7. Corvese46 extended these observations by demonstrating the ability of tobacco-specific nitrosamines to potentiate FRO production and double-strand DNA breakage. Similar to findings of Nakayama et al., Weitberg and Corvese showed that free radical scavengers abrogated the nitrosamine-induced damage. The complexity of DNA repair of FRO damage is readily apparent. A multiplicity of genes and gene products are involved. In a recent review of the subject, Pryor and Stone42 point to two principle mechanisms in which aberrant repair may contribute to mutations following FRO-induced strand breakage. CELL MEMBRANE Error-prone DNA ligase activity may join DNA strands across gaps following excision of altered bases. When such repair occurs across strand breaks accompanied by base release, a deletional mutation would result. The second mechanism relates to base misincorporation by aberrant polymerase activity in the attempt to restore normal sequences in damaged genes. Bleomycin and Free Radical Oxygen Also pertinent to the interpretation of this casecontrol study as well as to the understanding of head and neck carcinogenesis, numerous studies CYTOPLASM NUCLEUS Selenium Fig. 1. The generation of free radical oxygen and its modulation by cellular constituents. Free radicals are generated from both endogenous and exogenous sources. Constituentswithin the extracelMar space, cell membrane, or cellular cytoplasm can either block (-) or enhance (+) free radical carcinogenic potential. BAP = benzo(a)pyrene; HQ = hydroquinones; CC = catechols; Fe = iron; TSN = tobacco-specific nitrosamines, GSTu = glutathione-S-transferase mu; SOD = superoxides dismutase; CAT = catalase. Laryngoscope 107: June 1997 Schantz et al.: Genetic Susceptibility to Cancer 777 have demonstrated that bleomycin induces chromosomal breakage by means of generation of free oxygen radicals.8-10,47-49On initial entry into the cell, bleomycin complexes with ferrous iron and molecular 0 2 . The complex intercalates with DNA, principally between G-T and G-C sequences, leading to hydrogen abstraction from deoxyribose and thereby leaving a free radical. The latter combines with 02, which generates a peroxyl radical high-energy state, Similar to free radical interactions from other sources such as radiation or tobacco exposure, energy induced in this manner is decomposed through strand breakage. It is also relevant that bleomycin can produce 8-hydroxyguanine in DNA with a yield approximating %OO that of strand cleavage.9 As reported by Sikic,g however, no other type of base damage was identified. Tobacco contains a broad array of mutagenic compounds, each capable of interacting with DNA in a characteristic manner. Bleomycin sensitivity represents only one of several mechanisms of tobacco-induced carcinogenesis. It is, however, reflective of a very critical process (i.e., FRO damage). Our study demonstrated also for the first time that specific antioxidant vitamin intake (i.e., vitamin C or E) was interactive with quantitative assessment of mutagen sensitivity in defining risk of head and neck cancer. Both vitamin intake and mutagen sensitivity together provided more information than either factor alone. Results were consistent with previous in uitro experimentation, which showed the protective effect of these agents on bleomycin-induced chromosomal damage.23224 These latter in uitro studies used peripheral blood lymphocytes harvested from head and neck cancer patients. Lymphocytes were cultured and exposed to bleomycin in a manner identical t o the assay used in our case-control study. Prior to exposure t o bleomycin, lymphocytes were preincubated for varying periods with differing concentrations of vitamins A, C, or E. A clear dose-response protective effect was identified.23124 Cells preincubated for more than 2 hours with the above vitamins showed less sensitivity to mutagens as reflected in lower levels of chromosomal breakage. The in uitro concentrations of vitamins C and E that were used in the laboratory studies by Trizna et al.24, which protected against bleomycin-induced damage, can be achieved within the peripheral blood and thus are biologically and clinically relevant.403e52 Alcohol and DNA Repair The present case-control study also confirmed a highly significant interaction between alcohol and mutagen sensitivity. The OR of head and neck cancer increased dramatically in those individuals who both drank alcohol and were mutagen sensitive (OR = Laryngoscope 107: June 1997 778 45.8; 95%CI = 9.25 to 222.1). Results are in contrast to a recent study of Spitz et al.17 in which the odds ratio for interaction between alcohol and mutagen sensitivity was 5.8. One explanation for the difference may be related to methods of categorical grouping. Spitz et al. used a cutoff value of 0.8 b/c as the definition for mutagen sensitivity as compared with 1 b/c in our study. Likewise, referent categories for alcohol in our study were heavy drinkers (>lo0 drinks per month) versus non-heavy drinkers. Spitz et al. used nondrinkers only as the referent group. Our choice of each categorical grouping was associated with an increased disease OR and would account for an increased OR when used in combination (Tables I11 and V). The potential interactive effect between mutagen sensitivity and alcohol demonstrated in the present case-control analysis suggests a mechanism by which alcohol may contribute to head and neck carcinogenesis, specifically the inhibition of DNA repair16,53-56 Hsu et a1.53, and Hsu and Furlong 54 tested this hypothesis directly using the bleomycin assay and immortalized lymphocytes obtained from head and neck cancer patients. Hsu et al.53 noted that coincubation of lymphocytes with alcohol increased bleomycin-induced chromosomal breakage in a dosedependent manner. Kumano and Kajii55 showed that ethanol-enhanced aphidocolin induced fragile sites in human lymphocyte cultures. Aphidocolin functions to inhibit DNA repair,57 Relevant to our case-control results, Lin et al.56 examined the effect of alcohol on bleomycin-induced damage in Chinese hamster ovary cells and noted enhanced levels of chromosomal breakage when alcohol was applied simultaneously with bleomycin as compared with bleomycin alone. Pretreatment with alcohol had no effect on chromosomal breakage, indicating that alcohol had no direct clastogenic effect. Lin et al.,56 however, noted that multiple mechanisms may account for the synergistic effect between bleomycin and alcohol, inducing alcohol-mediated increased cellular membrane permeability, alterations of bleomycin metabolizing enzymes, and the inhibition of DNA repair. Hsu and Furlong 54 provided supportive evidence for the capacity of alcohol t o inhibit DNA repair. The latter authors performed a kinetic analysis of repair of bleomycin-induced chromosomal damage and noted that by 5 hours after bleomycin exposure, breaks were repaired to a level approximating spontaneous breakage and repair. In a subsequent set of experiments, cells were washed following bleomycin exposure and then reincubated in bleomycin-free media containing varying concentrations of ethanol. Marked inhibition of repair occurred only in a 2%ethanol solution. The inhibition of repair could be reversed by removal of the ethanol. The data provide direct evidence for alcohol's inhibition of DNA repair and provides a basis for interpreting results of our case-control study and othSchantz et al.: Genetic Susceptibility to Cancer ers regarding the interaction between alcohol and mutagen sensitivity in contributing to head and neck cancer risk.54 As previously discussed by Hsu et a1.53 and Hsu and Furlong,54 several mechanisms could account for this potentiation effect, which are independent of DNA repair. Alcohol may increase the permeability of the cellular membrane as well as increase the rate of cellular movement through the G2 phase into mitosis. Caffeine has been shown to function through this latter mechanism. The normal response to DNA damage by cells in the G2 phase is cell cycle arrest so as to allow repair. Chromosomal damage in mitosis would be more apparent in cells that were pharmacologically prevented from the adaptive G2-phase arrest response to clastogens. Multiple mechanisms may account for mutagen sensitivity. Indeed, the expression of mutagen sensitivity in the head and neck cancer population is undoubtedly a reflection of a combination of events. CONCLUSION This thesis has attempted to extend our knowledge of genetic susceptibility to head and neck carcinogenesis by performing a case-control epidemiologic investigation that incorporated the mutagen-sensitivity assay, a measure of DNA repair capacities. It is apparent that our understanding of heritable factors contributing to head and neck carcinogenesis is in its infancy. Current perceptions will undoubtedly change in time. It is highly likely, however, that we are dealing with a polygenic model of inheritance involving numerous contributing DNA polymorphisms. The effects of modulating carcinogen metabolism, racial and gender differences, inborn alterations in nutrient metabolism such as iron storage, as well as DNA repair are most likely interactive. As a result, the relative risk of disease provided by any single genetic measure of head and neck cancer will be limited. We also conclude that a central feature of head and neck carcinogenesis involves FRO damage, including its generation, promotion, and dissolution. This may be fundamental to our understanding of disease not only in patients who smoke, but also in those who avoid such exposures, given that FRO is generated from multiple sources including normal endogenous metabolism. Some type of associated mutational event may be the same in these individuals, even though the source of FRO-induced damage is different. Foremost may be the understanding of factors that prevent or ameliorate FRO-DNA interactions. When speaking of heritable traits, the interactions will primarily involve the role of DNA repair. The basis of our thesis has been the bleomycin assay, which was used to quantitate sensitivity to Laryngoscope 107: June 1997 mutagen damage. It is concluded that the relevance of this assay is based on its associated FRO-damaging effect and the ability of the host to repair that damage. Thus the assay involves only one of several mechanisms by which tobacco and other environmental exposures can induce DNA damage, albeit, a central mechanism. Results for this study confirm the significance of quantitative measures of mutagen sensitivity as an independent risk factor for upper aerodigestive tract cancers. The risk interaction was suggested between mutagen sensitivity and tobacco and alcohol. Each of these factors considered jointly provides more information than any factor alone. The possible interaction was found between alcohol and mutagen sensitivity, which supports the contention that both factors work through a common mechanism (i.e., DNA repair). Reported for the first time are the results of combined measures of nutrient intake and bleomycin-induced chromosomal damage as a contributing factor to head and neck cancer. Nutrients that were significant protectors in the development of disease function principally as free radical scavengers. The test of the significance of any epidemiologic study is the degree to which it influences future efforts involving both the basic and clinical sciences. Several critical questions raised by this study must be addressed. The data regarding the heritable nature of mutagen sensitivity are still limited. More extended pedigree analysis is required. Likewise, it is likely that the quantitative measures of mutagen sensitivity within head and neck cancer patients are a summation event (i.e., the end result of a multiplicity of factors such as nutritional state of the cell, iron catabolism, stress response, and DNA repair). Thus sensitivity within any single individual may be due to any one of several factors that may distinguish that individual from another mutagen-sensitive head and neck cancer patient. Clinical approaches must now be considered, such as the incorporation of mutagen sensitivity into chemopreventive trials designed to prevent second primaries. Indeed, such efforts are currently ongoing.21 Likewise, as a result of better defining high-risk populations, new strategies of screening are required. Investigations into the definition of early disease, the role of behavior modification, the development of new screening methods designed to detect the field cancerization process, and the ability to provide health access to high risk populations are all required. BIBLIOGRAPHY 1. Schottenfeld D. The etiology and prevention of aerodigestive cancers. Adu Exp Med Biol 1992;320:1-19. 2. Wynder EL, Stellman SD. Comparative epidemiology of tobacco-related cancers. Cancer Res 1977;37:4608-22. 3. de Villiers EM, Weidauer H, Otto H, zur Hauzen H. Papillo- Schantz et al.: Genetic Susceptibilityto Cancer 779 mavirus DNA in human tongue carcinomas. Znt J Cancer 1985;36:575-8. 4. Winn DM. Diet and nutrition in the etiology of cancer. Am J Clin Nutr 1995;61:437%45S. 5. Strauss BS. Cellular aspects of DNA repair. Adv Cancer Res 1995;45:45-105. 6. Hsu TC. Genetic instability in the human population: a working hypothesis. Hereditas 1983;98:1-11. 7. Hsu TC, Cherry LM, Samann NA. Differential mutagen-susceptibility in cultured lymphocytes of normal individuals and cancer patients. Cancer Genet Cytogenet 1985;17: 307-13. 8. Iqbal ZM, Zohn KW, Ewig AG, Fornace AJ. Single-strand scission and repair of DNA in mammalian cells by bleomycin. Cancer Res 1976;36:3834-8. 9. Sikic BI. Biochemical and cellular determinants of bleomycin cytotoxicity. Cancer Surv 1986;5:82-91. 10. Vie BK, Lewis R. Genetic toxicology -- of bleomycin. Mutat Res i978$5:121-45. 11. Wang RY, Hsu TC, Brock WA, Liang J. DNA fragility and repair capability are separate genetic phenotypes: studies on in t u nick translation and chromosome breakage. Znt J Oncol 1995;6:51-4. 12. Kuo MT, Hsu TC. Biochemical and cytological studies of bleomycin actions on chromatin and chromosomes. Chromosome 1978;68:22940. 13. Schantz SP, Hsu TC. Mutagen-induced chromosome fragility within peripheral blood lymphocytes of head and neck cancer patients. Head Neck 1989;11:337-42. 14. Hsu TC. Genetic predisposition t o cancer with special reference to mutagen sensitivity. Zn Vitro Cell Developmental Biol 1987;23:591-603. 15. Hsu TC, Johnston DA, Cherry LM, et al. Sensitivity to genotoxic effects of bleomycin in humans: possible relationship to environmental carcinogenesis. Znt J Cancer 1989;43: 403-9. 16. Spitz MR, Fueger JJ, Beddingfield NA, et al. Chromosome sensitivity to bleomycin-induced mutagenesis in patients with upper aerodigestive cancers: a case-control analysis. Cancer Res 1989;49:4626-38. 17. Spitz MR, Fueger JJ, Halabi S, et al. Mutagen sensitivity in upper aerodigestive tract cancer. Cancer Epidemiol Biomarkers Prev 1993;2:329-33. 18. Schantz SP, Hsu TC, Ainslie N, Moser RP. Young adults with head and neck cancer express increased susceptibility to mutagen-induced chromosomal damage. JAMA 1989; 262~3313-5. 19. Cloos J, Braakhuis BJ, Steen I, et al. Increased mutagen sensitivity in head and neck squamous cell carcinoma patients, particularly those with multiple primary tumors. Znt J Cancer 1994;56:816-9. 20. Li AT, Wang TD, Yang RT. Pingyangomycin-induced chromosome damage in lymphocytes of laryngeal cancer patients and healthy control subjects. Head Neck 1994;16:510. 21. Schantz SP, Spitz MR, Hsu TC. Mutagen sensitivity in patients with head and neck cancers: a biologic marker for risk of multiple primary malignancies. J Natl Cancer Znst 1990;82:1773-5. 22. Spitz MR, Hoque A, Trizna Z, et al. Mutagen sensitivity as a risk factor for second malignant tumors following malignancies of the upper aerodigestive tract. J Natl Cancer Znst 1994;86:1681-4: 23. Trizna Z, Schantz SP, Hsu TC. Effects of N-acetyl-L-cysteine and ascorbic acid on mutagen-induced chromosomal sensitivity in head and cancer patients. Am J Surg 1991;162: 294-8. 24. Trizna Z, Schantz SP, Lee JJ, et al. In vitro protective effects of chemopreventive agents against bleomycin-induced genotoxicity in lymphoblastoid cell lines and peripheral blood lymphocytes of head and neck cancer patients. Cancer Detect Prevent 1993;17:575-83. 25. Spitz MR, Fueger JJ, Borrud LG, Newel1 GR. The development of a comprehensive institutional-based patient risk Laryngoscope 107: June 1997 780 evaluation program. Part I. Development, content, and data management. Am J Prev Med 1988;4:183-7. 26. Block G, Dresser CM, Hartman AM, et al. Nutritional source in the American diet: quantitative data from the NHANES I1 Survey. Part I. Vitamins and minerals. Am J Epidemiol 1985;122:13-26. 27. Block G. Dresser CM. Hartman AM. et al. Nutrient sources in the k e r i c a n diet:’quantitative data from the NHANES I1 Survey. Part 11. Macronutrients and fats. Am J Epidemiol 1985;122:27-40. 28. Block G. A review of validations of dietary assessment methods. Am J Epidemiol 1987;126(5):796-802. 29. National Health and Nutrition Examination Survey I1 (NHANES 11).Hyattsville, MD: National Center for Health Statistics, 1976-1980. 30. Public use d a t a tape documentation. Model gram and nutrient composition. Hyattsville, MD: National Center for Health Statistics, 1982 (computer tapes 5702 and 5703). 31. US Department of Agriculture. Nutrient database for standard reference: expansion of data published in Composition of Foods-Raw, Processed, Prepared. Revisions of Agriculture Handbook No. 8. Hyattsville, MD: US Department of Agriculture, May 1988 (computer tape). 32. Breslow NE, Day NE, eds. Statistical Methods in Cancer Research. Vol 1. The Analysis of Case-Control Studies. Lyon, France: International Agency for Research on Cancer, 1980 (LARC scientific publication no. 32). 33. Mantel N, Haenszel W. Statistical aspects of the analysis of data from retrospective studies of disease. J Natl Cancer Znst 1959;22:719-48. 34. Rothman KJ. Synergy and antagonism in cause-effect relationships. Am J Epidemiol 1974;99:385-8. 35. Dayal HH. Additive excess risk model for epidemiologic interaction in retrospective studies. J Chronic Dis 1980;33; 653-60. 36. Schlesselman JJ, ed. Case-Control Studies: Design, Conduct, Analysis. New York,.Oxford University Press, 1982:63-68. 37. Krinsky NI. Actions of carotenoids i n biological systems. Annu Rev Nutr 1993;13:561-87. 38. Joenje H. Genetic toxicology of oxygen. Mutat Res 1989;219: 193-208. 39. Buettner GR. The pecking order of free radicals and antioxidants. Arch Biochem Biophys 1993;300:535. 40. Halliwell B. Free radicals, antioxidants, and human disease: where are we now? J Lab Clin Med 1992;119: 598-620. 41. Mullaart E, Lohman PHM, Berendz F, Vijg J. DNA damage: metabolism and aging. Mutat Res 1990;237:189-210. 42. Pryor WA, Stone K. Oxidants in cigarette smoke: radicals, hydrogen peroxide, peroxynitrate, and peroxynitrite. Ann N Y Acad Sci 1993;686:12-27. 43. Cheng KC, Loeb LA. Genomic instability and tumor progression: mechanistic considerations. Adv Cancer Res 1993;60:121-56. 44. Hoffmann D, and Hecht SS. Advances in tobacco carcinogenesis. In: Grover P, ed. Handbook of Experimental Pharmacology. New York: Spinger Verlag, 1989. 45. Nakayama T, Kaneko M, Kodama M, Nagata C. Cigarette smoke induces DNA single-strand breaks in cells. Nature 1985;314:462-4. 46. Weitberg AB, Corvese D. Oxygen radicals potentiate the genetic toxicity of tobacco-specific nitrosamines. Clin Genet 1993;43:88-91. 47. Dorr RT. Bleomycin pharmacology: mechanism of action and resistance and clinical pharmacokinetics. Semin Oncol 1992;19(S5):3-8. 48. Vorechovsky I, Munzarova M, Lokaj J. Increased bleomycininduced chromosome damage in lymphocytes of patients with common variable immunodeficiency indicates an involvement of chromosomal instability in their cancer predisposition. Cancer Immunol Zmmunother 1989;29: 303-6. Schantz et al.: Genetic Susceptibility to Cancer 49. Povirk LF, Austin MJ. Genotoxicity of bleomycin. Mutat Res 1991;257:127-43. 50. Frei B et al. Reactive oxygen species and antioxidant vitamins: mechanisms of action. Am J Med 1994;97(3A): 5s-13s. 51. Krinsky NR. Mechanism of action of biological antioxidants. Proc Soc Exp Biol Med 1992;200:24&54. 52. Pohl H, Reidy J. Vitamin C intake influences the bleomycininduced chromosomal damage assay: implications for detection of cancer susceptibility and chromosome breakage syndromes. Mutat Res 1989;224:247-52. 53. Hsu TC, Furlong C, Spitz MR. Ethyl alcohol as a carcinogen with special reference to the aerodigestive tract: a cytogenetic study. Anticancer Res 199l;ll: 1097-102. 54. Hsu TC, Furlong C. The role of ethanol in oncogenesis of the upper aerodigestive tract: inhibition of DNA repair. Anticancer Res 1991;11:1995-8. 55. Kumano A, Kajii T Synergisticeffect of aphidmlin and ethanol on the indudion of common fiagde sites.Hum Genet 1987;75:7543. 56. Lin YC, Ho IC, Lee TC. Ethanol and acetaldehyde potentiate the clastogenicity of ultraviolet light, methyl methanesulfonate, mitomycin C, and bleomycin in Chinese hamster ovary cells. Mutat Res 1989;216:93-9. 57. Hsu TC, Ramkissoon D, Furlong C. Differential susceptibility to a mutagen among human individuals: synergistic effect on chromosome damage between bleomycin and aphidiColin. Anticancer Res 1986;6:1171-6. The Robert A Jahrsdoerfer Lectureship The University of Texas Health Science Center Department of Otolaryngology and Division of Plastic Surgery announce The Robert A. Jahrsdoerfer Lectureship, July 10-12,1997. The topic will be The Role of Unilateral Atresia Repair in Congenital Grade I11 Microtia. Guest Speakers: Burt Brent, Laryngoscope 107: June 1997 MD; F. Fred Aguilar, MD; Paul Lambert, MD; Robert Jahrsdoerfer, MD; Antonio De la Cruz, MD. CME credits will be offered. For more information contact The Craniofacial Foundation of Houston, 6410 Fannin, Suite 927, Houston, TX 77030. TEL: 713-7970085. Schantz et al.: Genetic Susceptibility to Cancer 781
© Copyright 2026 Paperzz