The first 2 authors contributed equally to this work.
Early Detection and Diagnosis
Alcohol dehydrogenase 1C*1 allele is a genetic marker for alcohol-associated cancer in heavy drinkers†
Version of Record online: 14 NOV 2005
Copyright © 2005 Wiley-Liss, Inc.
International Journal of Cancer
Volume 118, Issue 8, pages 1998–2002, 15 April 2006
How to Cite
Homann, N., Stickel, F., König, I. R., Jacobs, A., Junghanns, K., Benesova, M., Schuppan, D., Himsel, S., Zuber-Jerger, I., Hellerbrand, C., Ludwig, D., Caselmann, W. H. and Seitz, H. K. (2006), Alcohol dehydrogenase 1C*1 allele is a genetic marker for alcohol-associated cancer in heavy drinkers. Int. J. Cancer, 118: 1998–2002. doi: 10.1002/ijc.21583
Part of this study was presented as Poster of Distinction at the Digestive Disease Week 2003 in Orlando, USA (Gastroenterology 2003;124:A546).
- Issue online: 15 FEB 2006
- Version of Record online: 14 NOV 2005
- Manuscript Accepted: 28 JUL 2005
- Manuscript Received: 12 MAY 2005
- University of Erlangen-Nürnberg. Grant Number: 01.03.14.1.
- alcohol dehydrogenase;
- genetic risk;
Chronic alcohol consumption is associated with an increased risk for upper aerodigestive tract cancer and hepatocellular carcinoma. Increased acetaldehyde production via alcohol dehydrogenase (ADH) has been implicated in the pathogenesis. The allele ADH1C*1 of ADH1C encodes for an enzyme with a high capacity to generate acetaldehyde. So far, the association between the ADH1C*1 allele and alcohol-related cancers among heavy drinkers is controversial. ADH1C genotypes were determined by polymerase chain reaction and restriction fragment length polymorphism in a total of 818 patients with alcohol-associated esophageal (n = 123), head and neck (n = 84) and hepatocellular cancer (n = 86) as well as in patients with alcoholic pancreatitis (n = 117), alcoholic liver cirrhosis (n = 217), combined liver cirrhosis and pancreatitis (n = 17) and in alcoholics without gastrointestinal organ damage (n = 174). The ADH1C*1 allele and genotype ADH1C*1/1 were significantly more frequent in patients with alcohol-related cancers than that in individuals with nonmalignant alcohol-related organ damage. Using multivariate analysis, ADH1C*1 allele frequency and rate of homozygosity were significantly associated with an increased risk for alcohol-related cancers (p<0.001 in all instances). The odds ratio for genotype ADH1C*1/1 regarding the development of esophageal, hepatocellular and head and neck cancer were 2.93 (CI, 1.84–4.67), 3.56 (CI, 1.33–9.53) and 2.2 (CI, 1.11–4.36), respectively. The data identify genotype ADH1C*1/1 as an independent risk factor for the development of alcohol-associated tumors among heavy drinkers, indicating a genetic predisposition of individuals carrying this genotype. © 2005 Wiley-Liss, Inc.
Chronic alcohol consumption is associated with the development of certain malignancies.1 It has been convincingly shown that excessive alcohol intake contributes to carcinogenesis in the oropharynx and larynx,2 esophagus,3 liver,4 female breast,5 and colon.6 Alcohol itself is not a carcinogen but may act as a tumor promoter.7 However, it remains unclear why only a minority of heavy drinkers develop alcohol-related tumors and severe alcohol-related diseases, while the vast majority of alcoholics do not. Recently, host genetic factors that modulate the individual risk have been suggested, after it was shown from twin studies that the susceptibility for alcohol-related organ injury is genetically determined in a substantial proportion.8 Since the inherited differences in alcohol metabolism could provide an attractive explanation for the individual susceptibility, the functional genetic polymorphisms of alcohol-metabolizing enzymes have been proposed as possible candidates. Alcohol is predominantly metabolized by the action of alcohol dehydrogenase (ADH) and the inducible microsomal cytochrome P450 2E1 (CYP2E1). Regarding ADH, several distinct isoenzymes have been cloned and characterized, which can be categorized into 5 different classes.9 The classification has recently been revised and the former isoenzymes ADH2 and ADH3 are now referred to as ADH1B and ADH1C. Two polymorphisms have been identified for ADH1C (ADH1C*1 and ADH1C*2), which encode the γ1 and γ2 subunits. The γ1γ1-isoenzyme (kcat, 87 min−1) reveals an enzymatic activity that is 2.5-fold higher than that of the γ2γ2-isoenzyme (kcat, 35 min−1), and therefore produces more acetaldehyde (AA).10 Thus, it has been hypothesized that individuals with homozygosity for the allele ADH1C*1 carry an increased risk for alcohol-induced organ damage, such as liver cirrhosis, than patients with heterozygosity or those homozygous for ADH1C*2.11 However, a large European cohort study involving 876 alcoholic patients could not detect such association.12 Nonetheless, a contribution of the functional ADH polymorphisms to the formation of alcohol-associated malignancies could be important. The established different kinetics of polymorphic ADH enzymes may modulate the development of cancers since the metabolic product of their enzymatic activity is AA, a highly toxic compound that has been identified as a carcinogen in animal experiments because of its mutagenic and carcinogenic properties that lead to genomic alterations, increased cell regeneration and metaplasia.13 However, when data from 7 population-based studies with respect to the possible role of ADH1C in the development of upper aerodigestive tract cancers including a total of 1,325 cases and 1,760 controls were recently analyzed, it was concluded that the ADH1C*1 allele does not confer an increased risk for head and neck cancers.14 This metaanalysis reevaluated pooled retrospective data originating from various regions with ethnic differences and largely minor to moderate alcohol consumption with limited data regarding smoking and, in some instances, drinking habits. The role of ADH1C*1/1 genotype in the development of hepatocellular and esophageal cancer was not investigated at all. A study from Puerto Rico demonstrated, that ADH1C*1/1 increases alcohol-associated cancer risk mainly in patients with high alcohol intake, showing evidence that the catalytic differences of ADH1C*1/1 may play a major role in these patients. Thus, the aim of the present study was to investigate a possible association between ADH1C*1/1genotype and alcohol-associated cancers among heavy drinkers.
Patients and methods
Altogether, 818 Caucasian patients from 5 German university centers were enrolled between 1999 and 2003: University of Lübeck (n = 570), University of Erlangen-Nürnberg (n = 71), University of Freiburg (n = 38), University of Regensburg (n = 22), and University of Heidelberg (n = 117). The study protocol was approved by the institutional Ethics Committees of all 5 participating centers, and all patients gave written informed consent to be included.
The present and past daily alcohol intake and tobacco consumption of all recruited patients was recorded prospectively during a face-to-face interview, and individuals were included if their alcohol consumption exceeded a minimum of 40 g/day for more than 10 years, and assignment to different groups was carried out according to the criteria as outlined below. Patients were not enrolled if alcohol consumption was <40 g/day. In addition, to analyze the genotype and allele frequencies of ADH1C in a geographically representative population, 204 community-based (laboratory and hospital staff) healthy individuals (99 men, 105 women, mean age 45 ± 20 years) with an alcohol consumption of 14 ± 8 g/day were also examined.
Heavy drinkers without alcohol associated gastrointestinal disease (n = 174)
Active heavy drinkers were recruited from patients admitted to the hospital for alcohol detoxification and/or for other health-related problems such as infections and accident injuries. Absence of alcoholic pancreatitis or liver cirrhosis was confirmed by the absence of clinical, biochemical, radiological or endoscopical evidence.
Chronic alcoholic pancreatitis (n = 117)
Diagnosis of chronic alcoholic pancreatitis was made if typical clinical symptoms and one of the following diagnostic signs were present: elevation of serum lipase of at least 3-fold the upper limit of normal; calcification and/or pancreatic duct changes as assessed by abdominal ultrasound and/or endosonography, computerized tomography and/or endoscopic retrograde cholangiopancreatography.
Alcoholic liver cirrhosis (n = 217)
Diagnosis of alcoholic liver cirrhosis was based on either histologic examination (n = 68) or indirect evidence such as portal hypertension, ascites, irregular liver surface, and/or splenomegaly as diagnosed by ultrasound, computerized tomography and/or endoscopy together with a history of heavy ongoing or previous daily alcohol intake, respectively. Chronic viral hepatitis, hemochromatosis, Wilson's disease, autoimmune hepatitis and primary biliary cirrhosis were excluded by testing for second generation antibodies against hepatitis C (anti-HCV), hepatitis B surface antigen (HBsAg), 24-hr urine copper excretion, ferritin and transferrin saturation, antinuclear autoantibodies, liver-kidney microsomal antibodies (LKM-1) and antimitochondrial antibodies, respectively. In all cirrhotics, hepatocellular carcinoma was excluded by repeated abdominal ultrasound examinations and α-fetoprotein measurements over a period of at least 1 year. Liver biopsy was not routinely performed in cirrhotic patients with Child-Pugh score B or C. In case of Child A cirrhosis, liver biopsy was performed only if no clinical or sonographical signs typical for cirrhosis and/or endoscopically proven esophageal varices could be determined.
Coincidental alcoholic liver cirrhosis and chronic pancreatitis (n = 17)
Patients who met both criteria for alcoholic liver cirrhosis and alcoholic pancreatitis were included for the statistical comparison of patients with alcohol-associated carcinoma versus alcohol-associated benign disorders.
Alcohol-associated hepatocellular carcinoma (n = 86)
The presence of HCC was confirmed either histologically (n = 71) or if abdominal ultrasound, and/or computerized tomography revealed a hepatic lesion together with elevated α-fetoprotein levels >1,000 ng/dl (100-times the upper limit of normal). None of the patients with HCC revealed other malignancies detectable by diagnostic procedures during tumor staging, which included either computed tomography or nuclear magnetic resonance tomography, chest X-ray, abdominal ultrasound, colonoscopy and laboratory screening.
Alcohol-associated esophageal carcinoma (n = 123)
Esophageal carcinoma was assessed through histology obtained during upper gastrointestinal endoscopy. Of the 123 patients, 85 had histologically confirmed squamous cell carcinoma, and 38 were diagnosed with adenocarcinoma. With regard to esophageal adenocarcinoma, the following criteria were applied: tumors were classified as a primary esophageal cancer, if the tumor mass was mainly located in the esophagus, or if there was endoscopically and/or histologically confirmed Barrett's syndrome. If histology was compatible with or suggestive of stomach cancer (i.e. oat cell carcinoma, mucin producing tumor), the patient was excluded.
Alcohol-associated head and neck cancer (n = 84)
All patients with head and neck squamous cell carcinoma underwent endoscopy, and the diagnosis was confirmed histologically. Tumors originated from the pharynx (n = 26), larynx (n = 37), and oral cavity (n = 23).
Genotyping was performed using genomic DNA extracted from whole blood, and trace amounts of DNA were amplified by PCR, as described previously.15 The nucleotide substitution at position 349 of the ADH1C gene produces a SspI (New England Biolabs, Boston, USA) restriction site and subsequent genotyping is based on restriction fragment length polymorphism. Generated fragments were separated by high-voltage electrophoresis on a 4% agarose gel, stained with ethidium bromide. All samples were tested in duplicates and interpretation of results was carried out by 2 independent investigators (N.H. and M.B.) who were unaware of the patients' medical history. To assure the reliability of the genotyping method, 25% of DNA samples were genotyped twice with no discordant results.
Allele and genotype frequencies were obtained by direct counting. To evaluate deviation from Hardy-Weinberg equilibrium, observed and expected genotype frequencies were compared by a Monte-Carlo goodness-of-fit test in different diagnostic groups as well as in origin groups. To determine the effect of ADH1C independently from clinical confounding factors, the following strategy was applied: In the first step, multivariate logistic regression analyses were performed to determine the impact of the clinical confounding factors age, gender, alcohol consumption (as numeric value in grams per day), as well as smoking status and consumption (as numeric value in number of cigarettes per day). To describe the coherence of the continuous variables age and alcohol with the diagnostic group, fractional polynomials were used.16 All variables and two-way interaction between variables were considered by stepwise backwards selection and remained in the model if p ≤ 0.01.
In the second step, the variable ADH1C genotype was included and added to the resulting models. ADH1C analyses were carried out as follows: Calculations of the impact of ADH1C on cancer risk were performed for allele frequencies of ADH1C*1 and ADH1C*2, respectively (with ADH1C as linear variable), and for the variable ADH1C*1 homozygosity vs. heterozygosity (ADH1C*2/1) vs. ADH1C*2 homozygosity. Odds ratios and 95% confidence intervals were calculated from the logistic model.
With regard to ADH1C genotype distributions and allele frequencies, no deviation from Hardy-Weinberg equilibrium was detected in any group. The genotypes of the 204 healthy nonalcoholics were as follows: ADH1C*1/1 (22.1%), ADH1C*1/2 (58.3%), ADH1C*2/2 (19.6%). The allele frequencies for the ADH1C*1 allele and ADH1C*2 were 51.2 and 48.8 %, respectively. The relevant demographic and clinical characteristics of the different patient groups are presented in Table I. Overall, patients with tumors were older than patients with nonmalignant alcohol-related organ damage. Gender distributions within the different groups showed more males (n = 641, 78.4%) than females (n = 177, 21.6%), with the highest proportion of men in patients with alcohol-associated HCC. The highest percentage of women was seen in the group with alcoholic liver cirrhosis. By definition, alcohol consumption was generally high with highest self-reported amounts in heavy drinkers without any established alcohol-induced organ damage. The lowest amounts were consumed in those with alcohol-associated HCC and those with liver cirrhosis. In all groups, the majority of patients were either active or former smokers with the highest percentage of ongoing nicotine consumption in those with alcohol-associated head and neck cancers and those with chronic pancreatitis were 95.2 and 79.5%, respectively. There were less smokers with malignant tumors, there being a high proportion of nonsmokers in patients with alcohol-associated hepatocellular cancer.
|Groups||Age (Years)||Gender||Alcohol (g/day)||Smoking status|
|Heavy drinkers(n = 174)||52.5 ± 11.7||135 (77.6)1||39 (22.4)||164 ± 130||118 (67.8)||8 (4.6)||48 (27.6)|
|Alcoholic pancreatitis (n = 117)||49.1 ± 10.6||98 (83.8)||19 (16.2)||135 ± 133||89 (76.1)||4 (3.4)||24 (20.5)|
|Alcoholic liver cirrhosis (n = 217)||57 ± 11.6||143 (65.9)||74 (34.1)||119 ± 93||116 (53.5)||14 (6.5)||87 (40)|
|Alcoholic pancreatitis and liver cirrhosis (n = 17)||52.7 ± 12.2||13 (76.5)||4 (23.5)||150 ± 128||7 (41.2)||3 (17.6)||7 (41.2)|
|Alcohol-associated HCC (n = 86)||65.9 ± 7.5||79 (91.9)||7 (8.1)||96 ± 58||34 (39.5)||5 (5.8)||47 (54.7)|
|Alcohol-associated esophageal cancer (n = 123)||63 ± 9.7||100 (81.3)||23 (18.7)||119 ± 93||89 (72.3)||12 (9.8)||22 (17.9)|
|Alcohol-associated head and neck cancer (n = 84)||57.2 ± 9.4||73 (86.9)||11 (13.1)||115 ± 62||76 (90.4)||4 (4.8)||4 (4.8)|
ADH1C genotypes and allele frequencies are displayed in Table II. Overall, patients with alcohol-related tumors had a higher ADH1C*1 allele frequency of 60.7–62.2% than patients with nonmalignant alcohol-related organ damage (47.0–52.9%). In addition, ADH1C*1 homozygosity was more frequent in patients with tumors (34.5–39.8%) than in those with nonmalignant alcohol-related organ injuries with the ADH1C*1/1 genotype well over 30% in all 3 tumor groups when compared to 17.6–21.8% in nontumor patients. The highest rate of ADH1C*1 homozygosity (39.8%) was detected in patients with esophageal carcinoma. Accordingly, the low-activity genotype ADH1C*2/2 was more frequent in drinkers without tumors than in patients with alcohol-related malignancy.
|Genotype||Allele frequency (%)|
|ADH 1C*1/1||ADH 1C*1/2||ADH 1C*2/2||ADH 1C*1||ADH 1C*2|
|Heavy drinkers (n = 174)||38 (21.8)1||92 (52.9)||44 (25.3)||48.3||51.7|
|Alcoholic pancreatitis (n = 117)||19 (16.3)||72 (61.5)||26 (22.2)||47.0||53.0|
|Alcoholic liver cirrhosis (n = 217)||43 (19.8)||122 (56.2)||52 (24.0)||47.9||52.1|
|Alcoholic pancreatitis and liver cirrhosis (n = 17)||3 (17.7)||12 (70.6)||2 (11.7)||52.9||47.1|
|Alcohol-associated HCC (n = 86)||32 (37.2)||43 (50.0)||11 (12.8)||62.2||37.8|
|Alcohol-associated esophageal cancer (n = 123)||49 (39.8)||54 (43.9)||20 (16.3)||61.8||38.2|
|Alcohol-associated head and neck cancer (n = 84)||29 (34.5)||44 (52.4)||11 (13.1)||60.7||39.3|
Using multivariate analysis age, gender and smoking were identified as independent risk factors for the development of alcohol-related head and neck cancer, esophageal carcinoma and HCC. After adjustment for these confounding factors, the ADH1C*1 allele and ADH1C*1/1 genotype remained risk factors for all 3 tumor entities under investigation as shown in table III. This positive association could also be observed for the comparison between malignant vs. benign alcohol-associated diseases when patients with alcoholic cirrhosis or pancreatitis and heavy drinkers without gastrointestinal diseases were combined, as ADH1C genotyping showed no differences between these patient groups.
|ADH1C*1/1 genotype||ADH1C*1 allele frequency|
|p-value||Odds ratio||p-value||Odds ratio|
|Malignant vs. benign tumor (n = 295 vs. n = 525)||<0.0001||2.77 (1.89–4.07)1||<0.0001||1.91 (1.47–2.48)|
|Alcohol-associated esophageal carcinoma vs. benign tumor (n = 123 vs. n = 525)||<0.0001||2.93 (1.84–4.67)||<0.0001||1.9 (1.36–2.64)|
|Alcohol-associated HCC vs. benign disease (n = 86 vs. n = 525)||0.011||3.56 (1.33–9.53)||0.016||2.14 (1.14–4)|
|Alcohol-associated head and neck cancer vs. benign tumor (n = 84 vs. n = 525)||0.024||2.2 (1.11–4.36)||0.023||1.75 (1.08–2.89)|
|Alcohol-associated HCC vs. liver cirrhosis (n = 86 vs. n = 217)||0.012||3.53 (1.31–9.47)||0.017||2.13 (1.14–3.98)|
Furthermore, odds ratios for ADH1C*1 genotypes were calculated showing the highest risk for the development of HCC if genotype ADH1C*1/1 was present. Odds ratios calculated for ADH1C alleles revealed that the carriage of one ADH1C*1 allele was also associated with an increased risk for all 3 tested malignancies.
The presented data clearly identify the ADH1C*1 allele and genotype ADH1C*1/1 as independent risk factors for the development of alcohol-associated carcinomas in heavy drinkers, a risk that remained after adjusting for potential confounders. Therefore, our data provide strong evidence that the heterozygosity, in particular homozygosity, for the ADH1C*1 allele represents a genetic marker for alcohol-related tumors in subjects with high alcohol intake.
Although a recent pooled analysis of individual patient data from the available studies on the role of ADH1C genotypes in alcohol-related head and neck cancers concluded that ADH1C did not seem to be a genetic risk factor in Caucasians,14 findings from the single studies were contradictory. In a study from Puerto Rico, a country with a high incidence of head and neck cancers, patients homozygous for ADH1C*1 revealed a 5-fold higher risk for developing oropharyngeal carcinoma than those with ADH1C*1 heterozygosity or ADH1C*2 homozygosity.17 The authors found a remarkable odds ratio of 40.1 for homozygosity of ADH1C*1 and of 7.0 for those with genotype ADH1C*1/2 with regard to upper gastrointestinal cancers in alcoholics who drank more than 57 drinks/week. A second case-control study from France on 39 patients with oropharyngeal and laryngeal carcinoma demonstrated a 2.6-fold and 6.1-fold risk for developing these tumors in patients with genotype ADH1C*1/1.18 However, 6 subsequent studies could not detect such an association.19, 20, 21, 22, 23, 24 Several aspects may account for this discrepancy. First, differences in the geographic distribution of ADH1C genotypes in Europe have been demonstrated,12 with particularly high frequencies of the ADH1C*1 allele in Southern France where the study by Bouchardy and coworkers was performed.20 Indeed, in this study, the ADH1C*1 allele was more frequent in controls than what we have observed and almost as high as in our cancer population. Thus, differences in the overall allele frequencies in this region may have masked some of the effect of this ADH1C variant on tumor development. In addition, this study excluded patients with severe hepatic disease, suggesting a possible selection bias. Second, ethnic confounders may have counteracted the detection of an association between the ADH1C*1 allele and alcohol-related cancers, which has to be accounted for in the study by Olshan et al.19 performed in North Carolina and that of Schwartz et al.21 carried out in Seattle, Washington. In both studies, the screened populations included a high proportion of Afro-Americans which could be an important aspect, given the high prevalence of genotype ADH1C*1/1 in up to 75% in Black Africans.25 Third, although Zavras and coworkers have included only Caucasians in their study, the number of cases and controls was rather low with an unexpectedly high ADH1C*1 frequency of 71% in cases and 69% in controls, which contrasts with all other studied Caucasian populations.23 Moreover, the assessed genotype frequencies did not fit the Hardy-Weinberg equilibrium, indicating uneven distribution of genotypes possibly due to migration effects.
However, the main difference between previous studies and our data is that the control groups mainly consisted of patients with minor to moderate alcohol consumption. Bouchardy et al. included only a small number of nondrinking but smoking controls from hospital staff, while the control group in the study by Olshan and coworkers consisted of hospital staff without significant alcohol consumption. Schwartz et al.21 included mainly nondrinking control subjects of which only 15.3% consumed more than 26 g/day, while Sturgis et al.22 gave no information on alcohol consumption at all. Thus, no conclusion concerning the possible link of ADH 1C*1/1 genotype on ethanol-associated cancer among heavy drinkers can be drawn from these studies. The approach of our study was different in this respect since it is the first attempt to investigate the effect of ADH1C genotypes in a homogenous population with a high alcohol intake of at least 40 g/day and an average consumption of more than 100 g/day for more than 10 years, ensuring a similar degree of alcohol exposure.
The enzyme coded by ADH1C*1/1 results in a 2.5 times higher capacity to generate AA when compared to the enzymes coded by ADH1C*1/2 and ADH1C*2/2. Thus, the AA accumulation is certainly much lower as observed in Asians with ALDH-2 deficiency, where relatively low levels of alcohol intake are associated already with a striking risk of cancer.26 Thus, the functional hypothesis for linking ADH1C*1/1 gene to alcohol-associated tumors via AA accumulation might be observed only in case of high alcohol intake such as reported here. This could explain, at least in part, the negative or contradictory findings in some of the previous studies.19, 20, 21, 22, 23, 24 A recent study has raised evidence that there could exist a significant interaction between alcohol use and genotype ADH1C.27 In this study on patients with head and neck cancer a higher increase in cancer risk was observed for heavy drinkers with the ADH1C*2/2 genotype when compared to patients heterozygous or homozygous for ADH1C*1, which contradicts our results. In this study only 29.9% of cases and only 11.5% of control patients (n = 69) had more than 26 drinks/week. A potential bias could appear in selecting only patients with heavy alcohol intake. However, in our study there was no independent effect of alcohol in modifying ADH1C*1 associated cancer risk.
Moreover, the “controls” as well as the “cases” consisted of very different gastrointestinal diseases such as pancreatitis, liver cirrhosis, patients without gastrointestinal diseases, hepatocellular cancer and esophageal cancer, and the observed differences regarding ADH1C genotype were remarkably stable in the different subgroups with a constant trend towards higher ADH1C1 frequencies among cancer patients.
This is supported by the fact that our observed genotypes in benign diseases are well in line with all other studies. In addition, we analyzed the genotype and allele frequencies of ADH1C in a geographically representative population of our study group. We found that the genotypes of our community-based healthy individuals were comparable to patients with benign diseases. It is noteworthy that ADH1C*1 allele frequencies in alcoholic patients without cancer are comparable to those of a healthy population studied in a similar geographical area12, 15 Thus a selection bias in our study population is unlikely.
Apart from its toxicity, AA is well-known for generating stable DNA adducts.28 It has been shown in cell culture that AA induces point mutations, sister chromatide exchanges and gross chromosomal aberrations.29, 30, 31 In addition, AA can cause upper gastrointestinal tract cancer in rodents.32 Blood AA levels in humans observed after ethanol intake usually do not exceed concentrations of 10 μM but AA exerts cell-toxic effects in a dose-dependent manner and concentrations of 10–50 μM already may be carcinogenic.33 In the upper gastrointestinal tract, AA can be derived from blood or saliva, or can be produced via mucosal or bacterial ethanol metabolism. Recently, it was demonstrated that substantial amounts of AA are generated by the oral microflora that are capable of using alcohol efficiently as an energy source.34 The ADH isoforms found in oral and esophageal epithelia differ from the ADH isoforms in other tissues. Instead of class I ADH (ADH1B and 1C), the predominant ADH isoform in the upper gastrointestinal tract is mainly class IV ADH (ADH4), which is characterized by a higher Vmax and Km than that of class I, leading to a substantial mucosal AA production.11 Since no genetic variants are described for ADH4, genetic variability in local AA levels are most likely related to ADH1C variants in Caucasians. The results of the present study are supported by findings recently demonstrated in a smaller cohort of alcoholic patients diagnosed with upper aerodigestive tract cancers from our group.15 In this study, patients homozygous for ADH1C*1 had a higher risk for developing cancers of the oral cavity, oropharynx, hypopharynx and esophagus most likely because of higher salivary AA. However, as we have not measured salivary or cellular AA levels in this study, we can only assume that the increased cancer risk for ADH1C*1 patients is caused by increased AA levels.
In summary, our data provide evidence that the ADH1C*1 allele frequency is associated with alcohol-associated malignancies in the upper gastrointestinal tract and the liver, and identifies genotype ADH1C*1/1 in Caucasians as the first known genetic marker for alcohol-related cancers in heavy drinkers. It further emphasizes the role of AA in alcohol-associated carcinogenesis.
F.S. is a recipient of a research fellowship by the Interdisciplinary Center for Clinical Research of the University of Erlangen-Nürnberg (IZKF). We are furthermore indebted to V. Benes, from the European Molecular Biology Laboratory, Heidelberg, Germany, and H. Schlichting, University of Lübeck, Germany, for skillful technical assistance. In addition, we are grateful to Dr C. Trautwein for his valuable support. Finally, the authors wish to thank Mrs. Hillesheim for writing the manuscript and Erich Wasserfallen for technical assistance.
- 13International Agency for Research on Cancer. Working Group on the valuation of the carcinogenic risk of chemicals to humans: acetaldehyde. IARC Monographs 1985; 36: 101–31.
- 24Laryngeal cancer risk in Caucasians is associated with alcohol and tobacco consumpriton but not modified by genetic polymorphisms in class I alcohol-dehydrogenases ADH1B and ADH1C and glutathione-S-transferases GSTM1 and GSTT. Pharmacogenetics 2003; 13225–30., , , , , , , .
- 32IARC monographs on the evaluation of the carcinogenic risk of chemicals to humans, Allyl Components, Aldehydes, Epoxides and Peroxides, Vol. 36. Lyon, France: International Agency for Research on Cancer, 1985. 101–32.