Genetic polymorphism in the conjugating enzyme UGT1A1 and the risk of head and neck cancer
UDP-glucuronosyltransferase 1A1 (UGT1A1) is an enzyme which catalyses not only the glucuronidation of tobacco smoke carcinogens like benzopyrene, but also of the endogenous substrate bilirubin. Bilirubin for a long time was considered to be only a toxic waste product of hemoglobin degradation, but recent findings have shown that bilirubin is a potent antioxidant, which may play a protective role against cancer. We investigated whether a genetic polymorphism in UGT1A1 (UGT1A1*28), associated with a reduced UGT1A1 enzyme activity, may have a risk-modifying effect on head and neck carcinogenesis. Blood samples from 421 patients with oral, pharyngeal or laryngeal carcinoma, and 417 healthy controls were investigated for the UGT1A1*28 polymorphism. On the basis of the occurrence of this polymorphism, patients and controls were divided according to predicted UGT1A1 enzyme activity (low, intermediate, high). Logistic regression analysis showed a significant increased distribution of predicted high activity UGT1A1*1 polymorphisms among the patients (OR: 1.37; 95% CI: 1.02–1.83). Stratified analyses demonstrated that predicted high activity UGT1A1 polymorphisms were present even more significantly in patients with laryngeal cancer, older patients, heavy smokers and heavy drinkers. In conclusion, the predicted high activity UGT1A1*1 polymorphism, which results in lower serum levels of the endogenous antioxidant bilirubin, was associated with an increased risk of head and neck cancer.
Squamous cell carcinoma of the head and neck (SCCHN) takes the 5th place in cancer incidence worldwide.1 Tobacco smoking and alcohol consumption are the most important causes in developing of SCCHN. Approximately 57% of men and 10% of women worldwide are tobacco smokers.2 Because of individual differences in susceptibility to develop a tobacco smoke–related cancer, only a small percentage of them will ultimately suffer from SCCHN. However, the mechanisms which influence this individual risk to develop SCCHN and other tobacco and alcohol consumption–related cancers are still not elucidated.
Smoking individuals may be daily exposed to a large variety of harmful or even carcinogenic compounds present in tobacco smoke.3–5 However, this threat of harmful compounds is encountered by the efficient and complex detoxification systems that exist in the epithelial cells lining the aerodigestive tract.6 This detoxification is the result of a complex interaction between Phase I and Phase II biotransformation enzymes. UDP-glucuronosyltransferase enzymes (UGTs) are an important class of Phase II conjugating enzymes, which catalyze the conjugation with UDP-glucuronic acid of many compounds, which subsequently can be excreted via bile or urine7, 8 Two main UGT families have been classified: UGT1A and UGT2B.7 UGTs generally are being considered as detoxification enzymes as their glucuronide products are more water soluble and less biologically active as compared to the nonglucuronidated parent compound. UGT1A1 is one of the important UGTs involved in the detoxification of tobacco smoke carcinogens, like benzopyrenes,8, 9 but UGT1A1 seems to be hardly expressed in the epithelial lining of the human aerodigestive tract.8
However, UGT1A1 is also the only enzyme which catalyzes the glucuronidation of bilirubin, as part of the hemoglobin catabolism, and therefore, it facilitates the excretion of bilirubin. For a long time, bilirubin was considered to be only a toxic waste product of hemoglobin degradation, but recent findings have revealed that bilirubin is a potent antioxidant, which may play a protective role against common diseases, like cardiovascular diseases and cancer.10–13 Because the concentration of serum bilirubin is inversely correlated with the UGT1A1 enzyme activity,14–18 genetic polymorphisms in UGT1A1 associated with a decreased enzyme activity might increase serum bilirubin levels and subsequently decrease the individual risk of developing SCCHN.
In the promoter region of the UGT1A1 gene, a thymine-adenosine (TA) repeat polymorphism exists.15 The presence of either 6 or 7 TA dinucleotides in the TATA region of the UGT1A1 gene promoter (UGT1A1*1 or UGT1A1*28 allele, respectively) influences the transcriptional activity of the UGT1A1 gene and may also subsequently influence the UGT1A1 enzyme activity. The transcriptional activity is inversely related to the number of TA repeats. Caucasians with the heterozygous variant TA6/TA7 genotype (UGT1A1*1/UGT1A1*28) showed an intermediate UGT1A1 enzyme activity, individuals with the TA7/TA7 genotype (UGT1A1*28/UGT1A1*28) demonstrated the lowest, ∼3-fold reduced enzyme activity as compared to individuals with the wild type TA6/TA6 genotype (UGT1A1*1/UGT1A1*1), which is associated with the highest enzyme activity.15–17 The UGT1A1*28 polymorphism is found in ∼55% of the Caucasians.19 Grant et al. have hypothesized that the UGT1A1*28 polymorphism may influence the susceptibility to oxidative damage and cancer development.20
In this study, we investigated the relation between the UGT1A1*28 polymorphism in the promoter region of UGT1A1 and the risk of SCCHN.
Material and Methods
Patients and controls
A total of 439 Caucasian patients with newly diagnosed and histological confirmed squamous cell carcinoma of the oral cavity, oropharynx, hypopharynx and larynx have been recruited in the period 1995–2005 for a study on the relation between genetic polymorphisms in detoxification enzymes and risk of SCCHN. All patients admitted to the Department of Otorhinolaryngology, Head and Neck Surgery of the Maastricht University Medical Center, to undergo diagnostic panendoscopy because of their malignancy were asked to participate in the study. The patients were referred to the Maastricht University Medical Center from the south-east region of The Netherlands, which is the referral region of this hospital. Because of failure in isolation of DNA of sufficient quality or failure in genotyping, 18 patients were not eligible for the evaluation and ultimately 421 patients (333 males, 88 females; 79 and 21%, respectively) were included in the study. This group consists of 82 patients (19.5%) with oral cavity carcinoma, 115 (27.3%) patients with oropharyngeal carcinoma, 174 patients (41.3%) with laryngeal carcinoma and 50 patients (11.9%) with hypopharyngeal carcinoma. Median age of the patient group was 61 years (range 23–91 years, see Table 1).
Table 1. General characteristics of the study populations
From the same referral region, a control group of 443 Caucasians was recruited. This group consists of healthy blood donors obtained through the blood bank situated in the referral region of our hospital. Only smokers and past-smokers were asked to participate in the control group. Because of failure in isolation of DNA of sufficient quality, failure in genotyping or due to lack of data on smoking history, ultimately 417 controls (323 males and 94 females; 78 and 22%, respectively) were included in the study. Median age of this group was 57 years (range 36–91 years, see Table 1). All participants from the control group underwent regular medical check-up before the blood donation. Controls did not suffer from any malignant disease and had no history of malignancy. The investigations were approved by the Medical Ethical Review Committee of the Maastricht University Medical Center and informed consent was obtained from all patients and controls.
Both patients and controls were asked to fill in a questionnaire with items on demographics, life-long smoking and alcohol consumption. According to the criteria described by Benhamou et al.,21 we categorized tobacco use into the amount of pack-years (py): for cigarette smokers, 1 py = 20 cigarettes per day for 1 year; for cigar smokers, 1 py = 4 cigars per day for 1 year and for pipe smokers, 1 py = 5 pipes per day for 1 year. No other form of tobacco use was found in our study population. According to the study of Elahi et al.,22 we considered 1 glass wine, 1 glass beer and 1 small-glass of hard liquor as roughly equivalent to each other, and alcohol consumption was calculated as the number of consumptions (units) per day. Participants were defined as not drinkers, if they had not consumed alcohol at all, moderate or “social” drinkers if they consumed 1–4 units per day (≤28 units per week) and heavy drinkers if they consumed more than 4 units per day (>28 units per week).
Blood sampling and assessment of genetic polymorphisms
Whole blood from patients and healthy controls was obtained by venapuncture in sterile vacutainer tubes, anticoagulated with EDTA and stored at −20°C until use. DNA was isolated from whole blood using the Pure Gene DNA isolation kit, according to the instructions of the manufacturer (Gentra Systems, Minneapolis, MN) and was stored at 4°C.
The number of TA-repeats in the promoter region of the UGT1A1 gene was analyzed using polymerase chain reaction (PCR) conditions and primers exactly as described by Monaghan et al.16 Amplification was confirmed by agarose electrophoresis before fragments were resolved on 12% polyacrylamide gels (19:1 acrylamide/bisacrylamide; Biorad Laboratories, Veenendaal, The Netherlands) in Tris-borate-EDTA buffer. Gels (20 × 20 × 0.075 cm3) were run at 400 V for 3 hr and were stained with ethidium bromide for 30 min.17 Fragments of 98 bp indicate the TA6 (UGT1A1*1) allele containing 6 TA repeats and fragments of 100 bp indicate the TA7 (UGT1A1*28) allele, containing 7 TA repeats.
Classification of predicted UGT1A1 enzyme activity as low, intermediate and high depending on combinations of different UGT1A1 genotypes15, 17 is shown in Table 2.
Table 2. Logistic regression analysis of predicted UGT1A1 activity in patients with SCCHN and controls based on different UGT1A1 genotypes
Unconditional logistic regression models were applied to estimate odds ratios (OR) and 95% confidence intervals (CI) for the genotypes with a predicted reduced enzyme activity, adjusting for age (continuous, per year increase), gender, alcohol consumption (0; 1–4 or >4 units per day) and smoking behavior (0; 1–19, 20–39, 40–59 or >59 py). Stratified regression analyses were performed, according to gender and smoking habits (less than 40 py, vs. 40 py or more). Separate regression analyses were also performed for patients with laryngeal cancer, oral/oropharyngeal cancer and those with hypopharyngeal cancer. In all analyses, a probability level of 0.05 was used as the criterion of significance. All analyses were performed with the software SPSS for Windows, version 13.0 (SPSS Inc., Chicago, IL).
The distribution of the UGT1A1 genotypes in our study population is given in Table 2. This distribution fitted the Hardy Weinberg equilibrium in the patient as well as in the control group (p = 0.92 and p = 0.64, respectively).
When the individuals with predicted high enzyme activity genotype were compared to the individuals with intermediate and low predicted activity genotypes taken together, a significant difference in the distribution of genotypes between patients with SCCHN and control subjects was found. The predicted high activity TA6/TA6 genotype was more frequent among the patients with SCCHN than among the controls (OR, 1.37; 95% CI, 1.02–1.83; see Table 3).
Table 3. Logistic regression analyses of predicted UGT1A1 activity in patients with SCCHN according to tumor site and controls
A stratified analysis according to tumor site showed that a high activity UGT1A1 allele was significantly more often present in patients with laryngeal cancer (OR, 1.68; 95% CI, 1.14–2.46) but not in patients with cancer of the oral cavity, oropharynx or hypopharynx.
Stratified logistic regression analyses according to age, sex, smoking behavior and alcohol consumption showed that the higher prevalence of the high activity UGT1A1 genotype was significantly more often present in older patients (>60 years) and patients classified as being heavy smokers (≥40 py) or heavy drinkers (>4 units per day), as compared to the corresponding control subjects (Table 4).
Table 4. Logistic regression analyses of predicted UGT1A1 activity in patients with SCCHN and controls, according to age, sex, smoking behavior and alcohol consumption
In this first study on the relation between the UGT1A1*28 polymorphism and risk of head and neck cancer, we found a significant association between the prevalence of the predicted high activity genotype and an increased risk of SCCHN. At first sight, this seems an unlogical finding as UGT1A1 is involved in the detoxification of tobacco smoke carcinogens and one would expect exactly the opposite, namely an increased risk in the predicted low activity genotype group. However, the UGT1A1 enzyme facilitates also the conjugation of bilirubin and the finding of this study might be explained by the protective effect of bilirubin against cancer, which has been postulated in several epidemiological and in vitro studies in the last years.10, 11, 23, 24 The exact mechanisms involved in the anticarcinogenic effects of bilirubin are not completely understood. However, Ollinger et al. have found that bilirubin can induce a cell cycle arrest in abnormally proliferating cells. They also stated that bilirubin may play a role in the defense against cancer by interfering with procarcinogenic signaling pathways.25
It is also not clear whether the serum concentration of total bilirubin (conjugated and unconjugated) or only the unconjugated fraction of bilirubin is important as an anticarcinogen. Novotny and Vitek in their meta-analysis on the inverse relationship between serum bilirubin and atherosclerosis in humans suggest that hyperbilirubinaemia due to a concomitant liver disease (increasing of conjugated fraction of bilirubin) may not exert the same protective affect as found by an increased concentration of unconjugated bilirubin26. If the same is true also for the anticarcinogenic effect of bilirubin is not clear yet.
In the present study, we did not measure the serum concentration of bilirubin. However, UGT1A1 is the only enzyme involved in the conjugation of bilirubin and clearance of bilirubin solely depends on the function of this enzyme. An inverse relationship between the UGT1A1 enzyme activity based on the UGT1A1*28 polymorphism, and the blood/serum/plasma concentrations of bilirubin in Caucasians has been firmly established.15–18, 27 Thus, the UGT1A1 enzyme has a permanent and long-term influence on the blood concentrations of bilirubin in humans.
Because the UGT1A1*28 polymorphism can be involved in the etiology of different cardiovascular and probably also other nonmalignant and malignant diseases, we have decided to avoid a hospital-linked selection of our control group to exclude the potential selection bias in our study population.27–29 Instead of a hospital-linked control group, we have chosen for a population of healthy smokers or past-smokers, whose health conditions are confirmed by regularly performed medical check-ups. The population of blood donors participating in our study fulfills these criteria.
Although the association between the UGT1A1*28 polymorphism and risk of SCCHN is significant for the whole patient group as compared to the controls, this association is even more pronounced for the subgroup of patients with larynx carcinoma and is not significant for the patients with cancer of the oral cavity or pharynx. It is unclear whether this is due to potentially different pathomechanisms for laryngeal versus pharyngeal carcinogenesis, as larynx carcinoma is mostly associated with tobacco smoking and alcohol consumption, whereas pharynx carcinoma is often based on infection with carcinogenic serotypes of the human papilloma virus (HPV), with or without additive exposure to tobacco smoke and alcohol. One can argue that the antioxidative and anticarcinogenic effect of bilirubin may influence the pathways in which tobacco smoke and alcohol metabolites are converted into procarcinogens and carcinogens, leading to DNA damage of the mucosa and subsequently to cancer. On the other hand, there are no reasons to believe that bilirubin can influence the incorporation of the carcinogenic HPV in DNA, which may result in unlocking the carcinogenic cascade of the infected cells. In this context, it is plausible that individuals with the highest predicted UGT1A1 enzyme activity and consequently with the lowest levels of the circulating antioxidant bilirubin in their blood may have the highest cancer risk; especially, when these individuals have a high consumption of cigarettes and alcohol as is the case in the larynx cancer subgroup. One can imagine that in these individuals, the balance between attack (by cigarette smoke carcinogens) and protection (by bilirubin) is disturbed and might lead to cancer development. This could explain why especially the heavy smokers (≥40 py) and heavy alcohol drinkers (>4 units per day) with the predicted high activity UGT1A1 genotype (with concomitant low bilirubin protection) are associated with a higher cancer risk, as compared to the moderate smokers and drinkers. For a better understanding, one has to realize that the UGT1A1 enzyme seems to be hardly expressed in the epithelial lining of the human aerodigestive tract8 so that a direct protecting effect of this enzyme may hardly be present in the mucosa of the aerodigestive tract, whereas an indirect effect of bilirubin (distributed by the blood) may be present at all sites and in all tissues.
Recently, we investigated the effect of the UGT1A7 polymorphism and the risk of SCCHN. UGT1A7 is another enzyme of the UGT1 family, which is involved in detoxification of tobacco smoke (pro)carcinogens.30 Surprisingly, we found an association between an increased risk of SCCHN and the high activity polymorphisms of UGT1A7, instead of the expected low activity polymorphisms. We postulated that one of the most likely explanations of this phenomenon could be the linkage between the UGT1A7 polymorphism and other functional genetic variants of the UGT1A locus, which could overrule the effect of the predicted high enzyme activity UGT1A7 polymorphisms. In our study population, we observed a strong relationship between the predicted enzyme activities of UGT1A1 and UGT1A7 polymorphisms (see Table 5). Eighty-two percent of the individuals with predicted low enzyme activity polymorphism of UGT1A1 were also associated with predicted low activity UGT1A7 polymorphism and 74% of the individuals with predicted high enzyme activity UGT1A1 polymorphism were also associated with predicted high activity UGT1A7 polymorphism. Similar high level of agreement was observed for cases and controls separately. Comparable linkage between the predicted high activity UGT1A1 and UGT1A7 polymorphisms was described in Caucasians and Egyptians by Kohle et al.31 Thus, the noticed low anticarcinogenic effect of the UGT1A1*1 allele with a high enzyme activity can probably overrule the expected strong anticarcinogenic effect of high activity UGT1A7 polymorphisms. The opposite (protective) effect on carcinogenesis of SCCHN due to presence of the UGT1A1*28 polymorphism can probably compensate the less protective effect of the low activity UGT1A7 polymorphisms.
Table 5. Predicted UGT1A1 versus UGT1A7 enzyme activity for the total group
In conclusion, in this study on the relation between the UGT1A1*28 polymorphism and risk of SCCHN, we found that the high activity UGT1A1*1/UGT1A1*1 genotype is significantly associated with an increased risk of SCCHN. The concomitant decreased blood level of bilirubin (as compared to genotypes with the UGT1A*28 allele), leading to decreased anticarcinogenic and antioxidant capacity of bilirubin might explain this phenomenon. Because this UGT1A1*28 polymorphism might also potentially modify the susceptibility for other malignancies, results of this study stresses the need for more research on the impact of UGT1A1 polymorphisms in relation to bilirubin blood levels and cancer risk, in general. If such a relation could be confirmed, the therapeutic strategies for reducing cancer risk by increasing the blood concentrations of bilirubin with therapeutics like probenecid, as proposed by Mc Carty12, could be established in the future.