Abstract
Head and neck cancers are causally related to alcohol consumption, but the underlying mechanisms are unclear. Ethanol is metabolized to acetaldehyde, an experimental carcinogen. Quantitation of the major DNA adduct of acetaldehyde, N2-ethylidenedeoxyguanosine, in human tissues could help to elucidate the mechanism of alcohol carcinogenicity. We applied a quantitative method for the analysis of this adduct, measured as the NaBH3CN reduction product N2-ethyldeoxyguanosine (N2-ethyl-dGuo) by liquid chromatography-electrospray ionization-tandem mass spectrometry-selected reaction monitoring, on DNA (0.04 ± 0.03 mg) isolated from blood collected from control subjects recruited from two studies conducted in different areas of Europe between 1999 and 2005. The group selected from the first study (n = 127) included alcohol drinkers and abstainers while the group from the second study (n = 50) included only heavy drinkers. N2-ethyl-dGuo was detected in all DNA samples. After adjusting for potential confounders, in the first study, drinkers showed a higher level of N2-ethyl-dGuo (5,270 ± 8,770 fmol/μmol dGuo) compared with nondrinkers (2,690 ± 3040 fmol/μmol dGuo; P = 0.04). A significant trend according to dose was observed in both studies (P = 0.02 and 0.04, respectively). Taking into account the amount of alcohol consumption, adduct levels were higher in younger compared with older subjects (P = 0.01), whereas no differences were observed comparing men with women. These results show the feasibility of quantifying N2-ethyl-dGuo in small-volume blood samples and are consistent with the hypothesis that ethanol contributes to carcinogenesis through DNA adducts formation. (Cancer Epidemiol Biomarkers Prev 2008;17(11):3026–32)
Introduction
Many epidemiologic studies have established the relationship between alcohol consumption and different types of cancers including those of the upper aerodigestive tract (1, 2). However, the mechanisms underlying alcohol drinking as a carcinogen are unclear; several hypotheses have been formulated, including an effect of alcohol metabolites (3). Ethanol is metabolized in the body through different pathways leading to formation of acetaldehyde, a genotoxic agent in animal and in vitro studies and experimental carcinogen (4).
Acetaldehyde is an electrophile that can interact with DNA to form adducts. The major DNA adduct produced is N2-ethylidenedeoxyguanosine (N2-ethylidene-dGuo; ref. 5). This adduct can undergo in vivo reduction to N2-ethyldeoxyguanosine (N2-ethyl-dGuo), which was found in the DNA of both ethanol-treated mice and human alcoholics (6-8). Structures of these adducts are illustrated in Fig. 1. Recently, a sensitive quantitative method was developed for the analysis of N2-ethylidene-dGuo as its NaBH3CN reduction product N2-ethyl-dGuo, which is more stable and thus easier to detect (5, 9). This method has been successfully applied for the quantification of the adduct in DNA extracted from human liver tissue and leukocytes (10, 11).
The development of the method provided the basis for investigating the influence of alcohol drinking on adduct levels and the potential use of N2-ethyl-dGuo as a biomarker for the study of the role of acetaldehyde in human cancers related to alcohol exposure. Furthermore, a relationship between this DNA adduct and alcohol consumption could lead to the identification of a potentially important tool for the assessment of alcohol exposure providing information on its biologically effective dose. Recall bias in self-reported estimates of alcohol consumption is well documented; patients may underreport their alcohol consumption, leading to measurement error and misclassification especially in those cultural settings where alcohol drinking is not socially acceptable (12). There is potential for the development of biomarkers of alcohol to address these limitations and improve exposure assessment. Using data and blood samples from controls recruited in two case-control studies conducted in different geographic areas of Europe, we pursued two objectives. First, we investigated the possibility of applying the newly developed method for the quantitation of N2-ethyl-dGuo on small-volume samples of whole blood, verifying its applicability in large epidemiologic studies. Secondly, we investigated the relationship between different levels of alcohol drinking and the formation of N2-ethyl-dGuo. Possible effect modifications due to gender and age were taken into account. Particular attention was given to the adjustment for cigarette smoking, because it is highly correlated to alcohol consumption and is another important source of acetaldehyde (11).
Materials and Methods
Study Subjects
The study subjects were selected from the controls of two case-control studies conducted in Europe to have blood samples from individuals with different levels of daily alcohol consumption ranging from null in abstainers to very high levels in alcohol abusers. One study was conducted at the National Institute of Oncology of Budapest from 2000 to 2005 and aimed at comparing cytogenetic biomarkers between healthy alcoholic controls and alcoholic patients with nonmalignant and malignant alcoholic liver disease (13). The other study was Alcohol-Related Cancers and Genetic Susceptibility in Europe (ARCAGE), a multicenter case-control study on head and neck cancers conducted from 2002 to 2005, in 14 centers from 10 European countries.10
The group of subjects selected from the Hungarian study comprised 50 volunteers with high alcohol intake, entering the hospital to start a program for treatment of alcohol abuse. Subjects with high alcohol intake were included if they consumed >36 g (men) or >20 g (women) of pure alcohol equivalents per day for at least 5 years and if they were free of known malignancies including human papillomavirus or EBV infections and alcoholic hepatitis. These subjects were interviewed with a structured questionnaire collecting information on residency and lifestyle factors, including cigarette smoking. From these subjects, 3 mL whole blood was obtained.A total of 127 subjects were selected from the controls from six centers of the ARCAGE study (Aviano, Barcelona, Padova, Prague, Turin, and Zagreb). These were subjects recruited from hospitals who had a recent diagnosis from a defined list of diseases unrelated to tobacco and alcohol consumption. The blood withdrawal was done within 20 days from admission to the hospital. All the subjects were interviewed with a structured questionnaire on residential and lifestyle history. The group included 58 subjects with no alcohol intake (abstainers) and 69 subjects with different levels of alcohol drinking. From these subjects, 1 mL whole blood was obtained.
Overall, 177 subjects were included in the present work and were all of European ethnicity. All blood samples were immediately stored at −30°C after withdrawal and kept at this temperature until DNA isolation. DNA was isolated from 49 samples from the Hungarian study and from 126 samples from the ARCAGE study, leaving 175 subjects for analyses. All the subjects included in the studies signed an informed consent for participation and ethical approval was obtained from the relevant committees.
High-Performance Liquid Chromatography-UV Analysis
Liquid Chromatography-Electrospray Ionization-Tandem Mass Spectrometry Analysis
This analysis was carried out, following the procedure reported in the literature (11), with an Agilent 1100 capillary flow HPLC (Agilent Technologies) with a 250 × 0.5 mm 5 μm particle size C18 column (Agilent Zorbax SB-C18) and a Discovery Max (Thermoelectron) triple quadrupole mass spectrometer. Adducts were quantified by tandem mass spectrometry with selected reaction monitoring at m/z 296 → m/z 180 ([M + H]+ → [BH]+ of N2-ethyl-dGuo) and the corresponding transition at m/z 301 → m/z 185 of [15N5]N2-ethyl-dGuo. A calibration curve was constructed before each analysis using a standard solution of N2-ethyl-dGuo and [15N5]N2-ethyl-dGuo. A constant amount of [15N5]N2-ethyl-dGuo (40 fmol) was mixed with differing amounts of N2-ethyl-dGuo (4, 8, 16, 32, 128, and 512 fmol) and analyzed by liquid chromatography-electrospray ionization-tandem mass spectrometry-selected reaction monitoring.
Chemicals and Enzymes
N2-ethyl-dGuo and [15N5]N2-ethyl-dGuo were prepared as described (10). Ethanol was obtained from AAPER Alcohol and Chemical. Isopropanol was purchased from Acros Organics. Puregene DNA purification solutions were obtained from Gentra Systems. Calf thymus DNA was purchased from Worthington Biochemical. Alkaline phosphatase (from calf intestine) was obtained from Roche Diagnostics. All other chemicals and enzymes were purchased from Sigma-Aldrich.
DNA Isolation from Human Whole Blood
DNA was isolated using the DNA purification from compromised whole blood protocol (Gentra Systems) with several modifications. Briefly, 9 mL RBC lysis solution were added to 3 mL whole blood. The WBC pellet was collected by centrifugation and treated with 3 mL cell lysis solution and 15 μL RNase A (4 mg/mL). To the cell lysate was added 1.4 mL protein precipitation solution and the mixture was centrifuged to remove protein. DNA was precipitated from the supernatant by addition of 4 mL ice-cold isopropanol. The DNA pellet was transferred in 70% ethanol in H2O, washed twice with the same solvent, and then finally washed with 100% ethanol. DNA was dried in a stream of N2 and stored at −20°C until use. The same protocol was used to isolate DNA from 1 mL samples, reducing the volumes of the solutions and the quantities of the reagents accordingly.
Sample Enrichment
For enzyme hydrolysis, the method reported previously in the literature was followed (11). Briefly, DNA was dissolved in 10 mmol/L Tris/5 mmol/L MgCl2 buffer containing [15N5]N2-ethyl dGuo and NaBH3CN. After the pH was adjusted to 7 with 0.1 N HCl, the DNA was initially digested overnight at room temperature with DNase I (type II, from bovine pancreas). Then to the resulting mixture were added additional DNase I, phosphodiesterase I (type II, from Crotalus adamanteus venom), and alkaline phosphatase. The mixture was incubated at 37°C for 70 min and then allowed to stand overnight at room temperature. Enzymes were removed by centrifugation using a centrifree MPS device (molecular weight cutoff of 30,000; Amicon). The hydrolysate was desalted and purified using a solid-phase extraction cartridge (Strata-X 33 μm, 30 mg/1 mL; Phenomenex). The 70% CH3OH fraction was collected and evaporated to dryness, dissolved in H2O, and purified using a mixed mode, anion exchange reverse-phase extraction cartridge (Oasis MAX, 30 mg/cartridge; Waters). The residue obtained was dissolved in H2O and analyzed by liquid chromatography-electrospray ionization-tandem mass spectrometry. Buffer blanks containing internal standard were processed as above and analyzed to check the mass spectrometry instrument baseline and possible contamination. Calf thymus DNA (0.2 mg) with internal standard added as above was used as a positive control to determine interday precision and accuracy. Each set of samples was run together with one buffer blank and three positive controls.
Statistical Analysis
Because the distribution of levels of N2-ethyl-dGuo was skewed toward the smaller values, we applied natural log transformation to the data for the analysis. Results are expressed as mean ± SD. Linear regression analysis was conducted to investigate the relationship between N2-ethyl-dGuo levels and the amount of alcohol drunk per day. The same method was used to investigate the effects of gender, age, and cigarette smoking on the DNA adduct levels. Each analysis was done with mutual adjustment for the other factors evaluated. The analyses were done considering the Hungarian subjects and the ARCAGE subjects separately. The analysis was repeated combining the results of both studies when considering the effects of gender, age, and cigarette smoking. In the case of the results from the samples selected from ARCAGE, each analysis was done on all the study subjects and then restricted to subjects whose blood samples were taken within 4 days from admission to the hospital. Trend tests for ordered variables were done by treating the categorical variable as a continuous predictor in the linear regression model. P values < 0.05 were considered statistically significant. All statistics were conducted using the STATA program (version 9.0).
Results
The demographic characteristics of the study subjects from the two studies are summarized in Tables 1 and 2. Detailed information on cigarette smoking and alcohol drinking habits is reported in Table 3.
Distribution of the samples from the National Institute of Oncology of Budapest according to age, sex, drinking, and smoking status
. | Drinkers . | . | ||
---|---|---|---|---|
. | Men . | Women . | ||
Total | 25 | 25 | ||
Age (y) | ||||
<35 | 3 | 1 | ||
35-44 | 10 | 8 | ||
45-55 | 9 | 11 | ||
>55 | 3 | 5 | ||
Smoking | ||||
Nonsmokers | 3 | 4 | ||
Current smokers | 22 | 21 |
. | Drinkers . | . | ||
---|---|---|---|---|
. | Men . | Women . | ||
Total | 25 | 25 | ||
Age (y) | ||||
<35 | 3 | 1 | ||
35-44 | 10 | 8 | ||
45-55 | 9 | 11 | ||
>55 | 3 | 5 | ||
Smoking | ||||
Nonsmokers | 3 | 4 | ||
Current smokers | 22 | 21 |
Distribution of the samples from the ARCAGE consortium according to age, sex, drinking, and smoking status
Total . | Nondrinkers . | . | Drinkers . | . | ||||
---|---|---|---|---|---|---|---|---|
. | Men (n = 30) . | Women (n = 28) . | Men (n = 43) . | Women (n = 26) . | ||||
Age (y) | ||||||||
<35 | 2 | 2 | 4 | — | ||||
35-44 | 4 | 1 | 6 | 4 | ||||
45-55 | 8 | 10 | 12 | 7 | ||||
>55 | 16 | 15 | 21 | 15 | ||||
Smoking | ||||||||
Nonsmokers | 13 | 16 | 6 | 11 | ||||
Current smokers | 17 | 12 | 37 | 15 |
Total . | Nondrinkers . | . | Drinkers . | . | ||||
---|---|---|---|---|---|---|---|---|
. | Men (n = 30) . | Women (n = 28) . | Men (n = 43) . | Women (n = 26) . | ||||
Age (y) | ||||||||
<35 | 2 | 2 | 4 | — | ||||
35-44 | 4 | 1 | 6 | 4 | ||||
45-55 | 8 | 10 | 12 | 7 | ||||
>55 | 16 | 15 | 21 | 15 | ||||
Smoking | ||||||||
Nonsmokers | 13 | 16 | 6 | 11 | ||||
Current smokers | 17 | 12 | 37 | 15 |
Alcohol and cigarette consumption among the subjects selected from the two studies
. | Hungary . | ARCAGE . | Total . | |||
---|---|---|---|---|---|---|
Alcohol consumption | ||||||
Nondrinkers | — | 58 | 58 | |||
Drinkers | 50 | 69 | 119 | |||
Drinkers (g/d alcohol) | ||||||
<10 | — | 9 | 9 | |||
10-99 | 10 | 38 | 48 | |||
100-199 | 21 | 15 | 36 | |||
200-300 | 13 | 1 | 14 | |||
>300 | 6 | — | 6 | |||
Missing | — | 6 | 6 | |||
Cigarette consumption | ||||||
Nonsmokers | 7 | 46 | 53 | |||
Current smokers | 43 | 81 | 124 | |||
Current smokers (no. cigarettes/d) | ||||||
<10 | 2 | 15 | 17 | |||
10-20 | 9 | 38 | 47 | |||
>20 | 32 | 20 | 52 | |||
Missing | — | 8 | 8 |
. | Hungary . | ARCAGE . | Total . | |||
---|---|---|---|---|---|---|
Alcohol consumption | ||||||
Nondrinkers | — | 58 | 58 | |||
Drinkers | 50 | 69 | 119 | |||
Drinkers (g/d alcohol) | ||||||
<10 | — | 9 | 9 | |||
10-99 | 10 | 38 | 48 | |||
100-199 | 21 | 15 | 36 | |||
200-300 | 13 | 1 | 14 | |||
>300 | 6 | — | 6 | |||
Missing | — | 6 | 6 | |||
Cigarette consumption | ||||||
Nonsmokers | 7 | 46 | 53 | |||
Current smokers | 43 | 81 | 124 | |||
Current smokers (no. cigarettes/d) | ||||||
<10 | 2 | 15 | 17 | |||
10-20 | 9 | 38 | 47 | |||
>20 | 32 | 20 | 52 | |||
Missing | — | 8 | 8 |
The mean age of the subjects was 50.9 ± 12.4 years for men (n = 98, 55.4%) and 53.9 ± 11.0 years for women (n = 79, 44.6%). All the subjects were interviewed with a structured questionnaire collecting specific information on number, type, and volume of drinks consumed per day. The grams of ethanol per day were calculated following the method reported in the literature (14). Subjects were categorized according to the dose of alcohol drunk per day. Daily alcohol consumption ranged from 40 to 416 g/d alcohol, with a mean of 174.1 ± 93.6 g, for the subjects from the Hungarian study and from 0 to 225 g/d alcohol for the subjects from the ARCAGE study, which included 58 abstainers and 69 alcohol drinkers with a mean of 61.7 ± 51.1 g/d alcohol. Information on cigarette smoking status was available for all subjects, and for 123 subjects, more detailed information on the number of cigarettes smoked per day was provided. Subjects were categorized considering smokers as current smokers and nonsmokers as never and former smokers.
The average quantities of DNA extracted from the two groups of samples of 3 and 1 mL whole blood were 44 ± 27 and 24 ± 20 μg, respectively. N2-ethyl-dGuo was detected in all samples analyzed (average, 4,910 ± 7,190 fmol/μmol dGuo; range, 410-43,402 fmol/μmol dGuo). A large interindividual variation in the levels of the adduct was observed in both groups of samples, but the levels in the samples from the Hungarian study were significantly higher (average, 7,780 ± 9,140 fmol/μmol dGuo) than those from the ARCAGE study (average, 3,790 ± 5,950 fmol/μmol dGuo; P = 0.02), whereas no significant differences were observed when comparing the levels in subjects from different centers within the ARCAGE study (P = 0.1). NaBH3CN-treated calf thymus DNA with internal standard was used as positive control in each set of samples processed. The reproducibility of the measurements was good, with an average relative SD of 19%.
Table 4 summarizes the effects of alcohol drinking on levels of N2-ethyl-dGuo in the samples from the Hungarian study. A trend was observed for increasing levels of N2-ethyl-dGuo with increasing alcohol consumed per day (P = 0.04) as shown in Fig. 2.
Mean levels of N2-ethyl-dGuo in different drinking categories in the samples from Hungary
. | n . | Mean ± SD (fmol/μmol dGuo) . | P . | |||
---|---|---|---|---|---|---|
Drinking categories (g/d) | ||||||
0-9 | — | |||||
10-99 | 9 | 5,820 ± 5,410 | ||||
100-199 | 21 | 6,890 ± 8,140 | ||||
200-300 | 13 | 8,310 ± 8,280 | ||||
>300 | 6 | 14,810 ± 12,850 | 0.04* |
. | n . | Mean ± SD (fmol/μmol dGuo) . | P . | |||
---|---|---|---|---|---|---|
Drinking categories (g/d) | ||||||
0-9 | — | |||||
10-99 | 9 | 5,820 ± 5,410 | ||||
100-199 | 21 | 6,890 ± 8,140 | ||||
200-300 | 13 | 8,310 ± 8,280 | ||||
>300 | 6 | 14,810 ± 12,850 | 0.04* |
Test for trend adjusted for sex, age, and number of cigarettes smoked.
Levels of N2-ethyl-dGuo (fmol/μmol dGuo) across categories of daily alcohol consumption (g/d alcohol) in subjects from the Hungarian study.
Levels of N2-ethyl-dGuo (fmol/μmol dGuo) across categories of daily alcohol consumption (g/d alcohol) in subjects from the Hungarian study.
Table 5 shows the effects of alcohol drinking on the adduct levels in the samples from ARCAGE. Overall, the mean level of the adduct was higher in the group of drinkers (4,190 ± 7,200 fmol/μmol dGuo) than in the group of nondrinkers (3,300 ± 3,980 fmol/μmol dGuo), but the difference was not significant (P = 0.2). The levels of N2-ethyl-dGuo increased by level of drinking, but the trend was not significant (P = 0.6). When considering the data from samples taken within 4 days from admission to the hospital (n = 79), the adduct level was 5,270 ± 8,770 fmol/μmol dGuo in drinkers and 2,690 ± 3,040 fmol/μmol dGuo in nondrinkers (P = 0.04) and the P value of the trend by level of drinking was 0.02. In the samples from the ARCAGE study, a decreasing trend was observed in the levels of the adduct according to length of stay in the hospital. For 72 subjects, information on the years of duration of drinking was also available, but no significant correlation with the level of N2-ethyl-dGuo was observed for this variable (P = 0.98).
Mean levels of N2-ethyl-dGuo in different drinking categories in the samples from ARCAGE
. | n . | Mean ± SD (fmol/μmol dGuo) . | P . | |||
---|---|---|---|---|---|---|
Drinking status | ||||||
Nondrinkers | 57 | 3,300 ± 3,980 | ||||
Drinkers | 69 | 4,190 ± 7,200 | 0.2* | |||
Drinking status† | ||||||
Nondrinkers | 38 | 2,690 ± 3,040 | ||||
Drinkers | 41 | 5,270 ± 8,770 | 0.04* | |||
Drinking categories (g/d) | ||||||
0-9 | 58 | 3,524 ± 4,224 | ||||
10-99 | 40 | 3,984 ± 7,237 | ||||
100-199 | 17 | 4,711 ± 8,494 | 0.6*‡ | |||
200-300 | — | |||||
>300 | — | |||||
Drinking categories (g/d)† | ||||||
0-9 | 39 | 3,010 ± 3,580 | ||||
10-99 | 28 | 4,630 ± 8,530 | ||||
100-199 | 7 | 9,090 ± 11,800 | 0.02*‡ | |||
200-300 | — | |||||
>300 | — |
. | n . | Mean ± SD (fmol/μmol dGuo) . | P . | |||
---|---|---|---|---|---|---|
Drinking status | ||||||
Nondrinkers | 57 | 3,300 ± 3,980 | ||||
Drinkers | 69 | 4,190 ± 7,200 | 0.2* | |||
Drinking status† | ||||||
Nondrinkers | 38 | 2,690 ± 3,040 | ||||
Drinkers | 41 | 5,270 ± 8,770 | 0.04* | |||
Drinking categories (g/d) | ||||||
0-9 | 58 | 3,524 ± 4,224 | ||||
10-99 | 40 | 3,984 ± 7,237 | ||||
100-199 | 17 | 4,711 ± 8,494 | 0.6*‡ | |||
200-300 | — | |||||
>300 | — | |||||
Drinking categories (g/d)† | ||||||
0-9 | 39 | 3,010 ± 3,580 | ||||
10-99 | 28 | 4,630 ± 8,530 | ||||
100-199 | 7 | 9,090 ± 11,800 | 0.02*‡ | |||
200-300 | — | |||||
>300 | — |
Adjusted for sex, age, and number of cigarettes smoked.
Considering only the subjects with samples taken within 4 d after hospitalization.
Test for trend.
The effects of gender, age, and cigarette smoking on the levels of N2-ethyl-dGuo are summarized in Table 6. When considering these variables, the same results were observed both in the study-specific and in the overall estimates. Higher levels of N2-ethyl-dGuo were found in men compared with women (Fig. 3), but the difference was not significant when adjusting for alcohol intake (P = 0.16). When considering the effects of age, the levels of the adduct were higher in subjects ages <35 years and declined in older age groups, showing a significant trend (P = 0.01) reported in Fig. 4. Smokers had a slightly higher level of the DNA adduct compared with nonsmokers. For 124 subjects, information on number of cigarettes smoked per day was available and levels of N2-ethyl-dGuo augmented with the increase in the number of cigarettes smoked per day (Fig. 5). However, none of the results associated with the smoking variables were statistically significant, even when restricting the analysis to subjects whose blood was taken within 4 days from hospitalization. In addition, for 72 subjects, information on duration of cigarette smoking was also provided; no association between N2-ethyl-dGuo levels and this variable was observed (data not shown).
Mean levels of N2-ethyl-dGuo by gender, age, and cigarette smoking in the samples from the two studies
. | n . | Mean ± SD (fmol/μmol dGuo) . | P . | |||
---|---|---|---|---|---|---|
Gender | ||||||
Men | 97 | 6,370 ± 8,820 | ||||
Women | 78 | 3,090 ± 3,720 | 0.16* | |||
Age (y) | ||||||
<35 | 12 | 10,700 ± 11,700 | ||||
35-44 | 33 | 6,210 ± 7,910 | ||||
45-55 | 57 | 4,920 ± 7,365 | ||||
>55 | 73 | 3,360 ± 5,030 | 0.01*† | |||
Cigarette smoking | ||||||
Nonsmokers | 52 | 4,000 ± 5,500 | ||||
Current smokers | 123 | 5,290 ± 7,790 | 0.65* | |||
Cigarette smoking categories (cigarettes/d) | ||||||
<10 | 17 | 4,330 ± 10,100 | ||||
10-20 | 47 | 4,560 ± 6,280 | ||||
>20 | 52 | 6,520 ± 8,420 | 0.17*† |
. | n . | Mean ± SD (fmol/μmol dGuo) . | P . | |||
---|---|---|---|---|---|---|
Gender | ||||||
Men | 97 | 6,370 ± 8,820 | ||||
Women | 78 | 3,090 ± 3,720 | 0.16* | |||
Age (y) | ||||||
<35 | 12 | 10,700 ± 11,700 | ||||
35-44 | 33 | 6,210 ± 7,910 | ||||
45-55 | 57 | 4,920 ± 7,365 | ||||
>55 | 73 | 3,360 ± 5,030 | 0.01*† | |||
Cigarette smoking | ||||||
Nonsmokers | 52 | 4,000 ± 5,500 | ||||
Current smokers | 123 | 5,290 ± 7,790 | 0.65* | |||
Cigarette smoking categories (cigarettes/d) | ||||||
<10 | 17 | 4,330 ± 10,100 | ||||
10-20 | 47 | 4,560 ± 6,280 | ||||
>20 | 52 | 6,520 ± 8,420 | 0.17*† |
Adjusted for alcohol intake and the other variables.
Test for trend.
Levels of N2-ethyl-dGuo (fmol/μmol dGuo) in men and women from the two studies.
Levels of N2-ethyl-dGuo (fmol/μmol dGuo) in men and women from the two studies.
Levels of N2-ethyl-dGuo (fmol/μmol dGuo) across age categories in subjects form the two studies.
Levels of N2-ethyl-dGuo (fmol/μmol dGuo) across age categories in subjects form the two studies.
Levels of N2-ethyl-dGuo (fmol/μmol dGuo) across categories of number of cigarettes smoked per day in subjects from the two studies.
Levels of N2-ethyl-dGuo (fmol/μmol dGuo) across categories of number of cigarettes smoked per day in subjects from the two studies.
Discussion
Our study had two purposes. One was to verify the applicability of the method for the quantitation of N2-ethylidene-dGuo, through the detection of its NaBH3CN reduction product N2-ethyl-dGuo, to small-volume samples collected in field studies, to use this technique on samples readily available from epidemiologic studies. The method was applied to DNA extracted from human whole blood samples (1-3 mL) and N2-ethyl-dGuo was detected in all samples analyzed. Data were consistent according to the reproducibility of the measurements observed in the analysis of the positive controls from each set of samples. These results show the high sensitivity of the method and suggest its applicability in large epidemiologic studies.
The other objective was to explore the relationship of N2-ethyl-dGuo with one of the major potential sources of acetaldehyde exposure: the metabolism of ethanol from alcohol drinking. Therefore, we applied the method to samples from individuals with different alcohol consumption habits. Levels of N2-ethyl-dGuo in the Hungarian subjects were particularly high (7,780 ± 9,140 fmol/μmol dGuo) compared with those found in previous studies and those detected in the samples from the ARCAGE study. These subjects were selected among heavy drinkers entering a program for the treatment of alcohol abuse. Hungarians are among the populations with the highest rate of cancer mortality in Europe, a phenomenon thought to be mainly related to lifestyle factors, including high alcohol intake, and the high prevalence of smoking (15, 16). Alcohol consumption in Hungary has been historically linked to heavy alcohol use, related mainly to homemade fruit-derived spirits for which the ethanol concentration might be very high (17). These factors could suggest a particularly high level of exposure to acetaldehyde in the study subjects. Among ARCAGE subjects, levels of the adduct were lower than those measured among Hungarian subjects (average, 3,850 ± 6,270 fmol/μmol dGuo) and similar to the ones reported in previous studies (11). Drinkers had a higher level of N2-ethyl-dGuo than nondrinkers and a dose-response relationship was observed across categories of alcohol intake. These results were statistically significant only when limiting the analysis to samples taken within 4 days from admission to the hospital. Indeed, the samples from the ARCAGE study were taken from hospitalized individuals and a decreasing trend, even if not significant, was observed in the levels of N2-ethyl-dGuo with increasing number of days between hospital admission and blood drawing. This could suggest a possible effect of the change of the subjects' lifestyle after entering the hospital, which could affect levels of exposure and consequently the level of the DNA adduct. Because the time between admission to the hospital and blood withdrawal was a modifying effect and not a confounder instead of adjusting for the duration of the stay in the hospital, the analysis was restricted to the categories of subjects who had the blood taken within 4 days from hospital admission. DNA adducts are generally considered to provide accurate estimates of exposure beyond several half-lives of the DNA adduct (18) and N2-ethylidene-dGuo in DNA was reported to have a half-life of ∼24 h at 37°C (11), suggesting to restrict the analysis to samples taken not after 4 days of hospitalization. This could be a limitation in particular when considering a possible use of N2-ethyl-dGuo as a biomarker of exposure in hospital-based epidemiologic studies. In these studies, blood is not always obtained in the first days of hospitalization, information on the time between admission to the hospital and blood withdrawal might not be always available, and this period could change between cases and controls. Further studies on the stability in humans of N2-ethylidene-dGuo should be conducted to investigate the half-life of the adduct in vivo and to clarify the type of exposure information provided by the measurement of this DNA adduct.
Thus far, N2-ethyl-dGuo has been found in the DNA of peripheral WBC of heavy drinkers in two other studies (7, 8). However, none of them investigated the dose-response relationship. Moreover, none of these studies detected the adduct in all the samples analyzed, and as the NaBH3CN reduction step was not used, it is unclear whether the measured N2-ethyl-dGuo originated from N2-ethylidene-dGuo.
Ethanol pharmacokinetics varies significantly between men and women. Studies have shown that the maximal ethanol concentration reached after exposure to the same amount of alcohol is lower in females and that the excretion of ethanol in women is faster than in men (19). We found a higher level of N2-ethyl-dGuo in men than in women. However, when adjusting for amount of alcohol intake, this result was no longer significant, showing that the difference observed is related to disparity in the levels of alcohol consumption rather than to metabolic differences between men and women. Similar results were observed both in the overall and in the study-specific estimates.
Possible effects of age were also investigated. Younger subjects showed higher levels of N2-ethyl-dGuo and a decreasing trend was observed when considering older age groups after adjusting for alcohol intake and cigarette smoking. This trend remained statistically significant when considering only samples taken within 4 days from hospitalization. Also, in this case, this trend was observed both in the overall and in the study-specific estimates.
Underreporting of alcohol intake could be higher among the youngest. Also, different drinking patterns (binge drinking and heavy episodic drinking) or changes in drinking habits over a lifetime together with possible changes of ethanol metabolisms with age are likely to lead to different types of exposure (20). Higher levels of N2-ethyl-dGuo in younger subjects could be related to these observations.
Another main source of exposure to acetaldehyde is tobacco smoke and the investigation of the possible relationship between N2-ethyl-dGuo and alcohol drinking would have been more accurate by measuring the adduct in nonsmokers, but few heavy drinkers are also nonsmokers. For this reason, the effects of cigarette smoking were taken into account by statistical adjustment. No significant differences were seen when comparing smokers and nonsmokers and no significant trend was observed in the increase of N2-ethyl-dGuo with the increase of number of cigarettes smoked per day both in the overall and in the study-specific estimates. Exposure to acetaldehyde from cigarette smoking might be low compared with the exposure due to ethanol metabolism because a typical drink is considered to contain around 10 g ethanol, which is then metabolized to acetaldehyde, whereas exposure from cigarette smoke is estimated to be in average ∼1.5 mg/cigarette (range, 0.6-2.1 mg/cigarette; ref. 21). In addition, alcohol consumption and cigarette smoking could provide acetaldehyde exposure through different routes, which could be of different relevance for the adduct formation in leukocytes DNA.
There are some limitations to this study. As mentioned above, the two groups of samples differ in the amounts of alcohol drunk per day and in the study design they originate from: the ARCAGE samples were selected among hospital-based controls, whereas the Hungarian samples were not. The assessment of alcohol intake through questionnaire aimed at identifying the usual alcohol intake of the study subjects. In light of the decrease of the adduct, seen after 4 days from hospitalization, a more detailed investigation of the amount of alcohol drunk in the days right before hospital admission could have added some useful information. Recent exposure might be more relevant for the study of the effects of alcohol consumption on N2-ethyl-dGuo levels than the estimation of usual alcohol intake. Also, no information was available on alcohol drinking patterns such as irregular heavy drinking or heavy drinking between meals, which could have an effect on exposure. Finally, due to the large interindividual variation in the levels of N2-ethyl-dGuo, the relatively small number of subjects included in the study might have reduced the possibility of observing a more subtle variation due to exposure to acetaldehyde from cigarette smoke. This could also explain why no significant difference is observed when comparing the levels of N2-ethyl-dGuo in subjects drinking a few glasses a day and in abstainers. This large variation in the levels of N2-ethyl-dGuo could be related to other sources of exposure contributing to the adduct formation. Endogenous formation of acetaldehyde from physiologic metabolism may represent a main source of exposure. Individual differences due to polymorphisms in genes encoding for enzymes involved in the metabolic processes leading to acetaldehyde formation and degradation, together with polymorphisms in genes involved in DNA repair, could contribute to explain this observation. The effects on levels of N2-ethyl-dGuo observed for carriers of a mutant allele encoding for an inactive subunit of aldehyde dehydrogenase-2 (8) support this hypothesis and underline the importance and the need of further investigations of the relationship of levels of the adduct with polymorphisms in genes functionally related to these enzymes.
In conclusion, these results show that alcohol drinking contributes to leukocyte levels of an acetaldehyde-DNA adduct in a dose-dependent manner giving strength to the hypothesis that N2-ethyl-dGuo is a potentially useful biomarker for the study of alcohol-related carcinogenicity. Further studies are needed to better characterize and to validate this potential biomarker for the investigation of alcohol-related carcinogenicity.
Disclosure of Potential Conflicts of Interest
No potential conflicts of interest were disclosed.
Grant support: European Commission's 5th Framework Program contract QLK1-2001-00182; Environmental Cancer Risk, Nutrition and Individual Susceptibility, a network of excellence operating within the European Union 6th Framework Program, Priority 5: “Food Quality and Safety” contract 513943; U.S. National Institute of Environmental Health Sciences grant ES-011297; and Hungarian National grant NKFP/1B/020/04. The Hungarian study was supported by grant NKFP/1B/020/04. The work of the collaborators from the University of Turin was supported by Compagnia San Paolo, Italian Association for Cancer Research, and Region Piedmont grants.
Acknowledgments
The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.