Int. J. Cancer: 85, 68–72 (2000) r 2000 Wiley-Liss, Inc. Publication of the International Union Against Cancer Publication de l’Union Internationale Contre le Cancer DETERMINANTS OF O 6-ALKYLGUANINE-DNA ALKYLTRANSFERASE ACTIVITY IN NORMAL AND TUMOUR TISSUE FROM HUMAN COLON AND RECTUM A.C. POVEY1,2*, C.N. HALL3, D.P. COOPER1, P.J. O’CONNOR1 and G.P. MARGISON1 1Cancer Research Campaign, Section of Genome Damage and Repair, Paterson Institute for Cancer Research, Manchester, UK 2School of Epidemiology and Health Sciences, Medical School, University of Manchester, Manchester, UK 3Department of Surgery, Wythenshawe Hospital, Wythenshawe, Manchester, UK O6-Alkylguanine-DNA-alkyltransferase (ATase) is an important modulator of alkylating agent–induced toxicity and carcinogenicity, but those factors which influence the expression of this repair protein in human tissues are poorly characterised. In this study, we have determined ATase levels in macroscopically normal and tumour tissues from 76 individuals with benign or malignant colorectal disease. All tissue samples had detectable ATase activity, with values ranging from 35 to 451 fmol/mg protein. ATase activity in normal rectal tissue was significantly higher than that in normal tissue from the sigmoid colon (148 ⴞ 76 vs. 100 ⴞ 40 fmol/mg protein, p ⴝ 0.01), whereas ATase levels within different regions of the colon (proximal vs. sigmoid colon) were similar. In normal tissue, inter-individual variation in ATase activity was 4-fold in the colon and 6-fold in the rectum, whereas in tumour tissue the corresponding figures were approx. 13.0and 7-fold, respectively. There was no detectable difference in normal tissue ATase activity between individuals with benign or malignant disease of the colon. Normal and tumour tissue ATase activities were strongly correlated in the sigmoid colon (r ⴝ 0.80) and rectum (r ⴝ 0.59) but not the caecum (r ⴝ ⴚ0.03). In a multivariate analysis, ATase activity in normal colon tissue increased with age (p ⴝ 0.01) and current smoking (p ⴝ 0.06), whereas tumour ATase activity increased only with use of anti-histamines (p ⴝ 0.05). In rectal tumour tissue, activity decreased with age (p ⴝ 0.05) and use of anti-muscarinic medications (p ⴝ 0.01): in normal rectal tissue, no modulating factors were identified. Int. J. Cancer 85:68–72, 2000. r 2000 Wiley-Liss, Inc. The presence of O 6-alkylguanine in DNA can have a number of deleterious cellular effects as this adduct is one of the principal cytotoxic, mutagenic and carcinogenic lesions induced by exposure to alkylating agents. By repairing O 6-alkylguanine in DNA, O 6-alkylguanine-DNA-alkyltransferase (ATase, also known as MGMT and AGT) provides protection against these effects. The alkyl group from the O 6 position is transferred to a cysteine in the protein in a process that regenerates the guanine base but inactivates the repair protein (Pegg et al., 1995). The cellular capacity for repairing O 6-alkylguanine lesions is thus dependent on the number of ATase molecules present within cells. Depletion of cellular ATase levels will render cells more susceptible to alkylating agents, whereas increased ATase expression will increase resistance. ATase expression in human tumour and normal cells is thus an important determinant of the effectiveness and toxicity of alkylating agent– induced chemotherapy, and in modulating the effects of the unavoidable exposure to methylating agents, ATase expression in normal cells may influence susceptibility to alkylating agent– induced carcinogenesis. ATase is generally expressed constitutively in normal cells, but a small number of individuals have no detectable expression of ATase in either normal or tumour cells (e.g., Gerson et al., 1995; White et al., 1997). The absence of ATase expression in these individuals arises from a lack of gene transcription, probably as a result of methylation of CpG islands within the promoter region of the ATase gene (Qian and Brent, 1997). Factors which affect ATase expression are poorly characterised, but tissue and cellular heterogeneities in ATase expression indicate the presence of tissue- and cell-specific enhancers or inhibitors. DNA strand breaks can act as inducing signals, and ionising radiation can induce ATase expres- sion in murine tissues and cell lines; this induction is dependent on wild-type p53 (Rafferty et al., 1996). ATase expression is also increased in both human and rodent tissues and cells by DNAdamaging agents such as cigarette smoke (Drin et al., 1994). ATase expression in the rat liver is reduced following a single ethanol dose, but after chronic administration ATase expression increases (data not shown). In humans, ATase RNA expression is similar in the liver of alcoholics and non-alcoholics with chronic hepatitis or cirrhosis (Miyakawa et al., 1996). ATase activity can also be inhibited by the use of alkylating agents and pseudo-substrates (e.g., Pegg et al., 1995). Polymorphisms within the ATase gene have been reported, but whether they are associated with changes in expression or altered substrate specificity remains to be determined (Otsuka et al., 1996). As with other human tissues, ATase expression in normal and tumour tissue from the colon is highly variable. Expression in normal colon tissue has been reported to vary up to 18-fold (Gerson et al., 1992; Zaidi et al., 1996) and in tumour tissue by 25-fold (Zaidi et al., 1996). There is increasing evidence that such variations in ATase expression within the colon may be important in determining cancer susceptibility. The RER phenotype, present in approx. 10% of sporadic colorectal tumours, is induced in vitro by continuous exposure to methylating agents in the presence of low ATase expression (Karran and Bignami, 1994). Colorectal DNA contains O 6-methylguanine, clearly indicating that colorectal tissue is exposed to methylating agents (Jackson et al., 1996). We have also observed that individuals with low ATase activity in normal colorectal tissue were more likely to have a K-ras GC = AT transition mutation than individuals with high ATase activity (Jackson et al., 1997). It thus becomes increasingly important to characterise those factors (whether genetic or environmental) which may influence ATase expression within the large bowel. In this population, we have examined whether ATase activity is associated with any demographic variable or exposure to cigarette smoke, alcohol, or medications. MATERIAL AND METHODS The study population consisted of 34 men and 42 women with a median age of 72 years: 22 were current smokers and 48 current drinkers. Of these 76 people, 14 had benign colon disease and 62 tumours of the colon or rectum. Paired tumour and macroscopically normal tissue, sufficient for ATase analysis, was obtained from 55 individuals. Tissue samples were obtained at surgery, snap-frozen and then stored at ⫺70°C until processed. Between 1 and 10 mg Grant sponsor: Cancer Research Campaign. D.P. Cooper’s current address is: Micromass UK Ltd., Floats Road, Wythenshawe, Manchester, M23 9LZ, UK. *Correspondence to: School of Epidemiology and Health Sciences, Medical School, University of Manchester, Oxford Rd., Manchester, M13 9PT, UK. Fax: ⫹44 161 275 5595. E-mail: [email protected] Received 11 May 1999; Revised 19 July 1999 O 6-ALKYLGUANINE-DNA ALKYLTRANSFERASE IN COLORECTAL CANCER 69 TABLE I – ATase LEVELS IN TUMOUR AND MACROSCOPICALLY NORMAL TISSUE FROM PATIENTS WITH DISEASES OF THE COLON AND RECTUM Normal tissue Sample set Colon disease Benign Malignant Colorectal tumours Proximal colon3 Sigmoid colon Rectum Tumour tissue Mean ⫾ SD (n)1 Range Mean ⫾ SD (n)1 Range 98 ⫾ 35 (14) 107 ⫾ 44 (36) 53–162 52–201 — 146 ⫾ 104 (34)2 — 35–451 118 ⫾ 47 (15) 100 ⫾ 40 (21) 148 ⫾ 76 (26)5 52–201 56–194 57–342 150 ⫾ 90 (14) 144 ⫾ 115 (20)4 170 ⫾ 90 (21) 41–320 35–451 48–348 1Results expressed in fmol/mg protein (n ⫽ number of samples).–2Activity significantly higher than in normal tissue ( p ⫽ 0.03 using t-test for paired samples).–3Includes individuals with tumours of the hepatic flexure or transverse colon or caecum.–4Activity significantly higher than in normal tissue ( p ⫽ 0.04 using t-test for paired samples).–5Activity significantly higher than in normal tissue from the sigmoid colon ( p ⫽ 0.01) using Mann-Whitney U-test. FIGURE 1 – Relationship between ATase activity in paired normal and tumour tissue from individuals with tumours of the proximal colon (a), sigmoid colon (b) and rectum (c). Proximal colon includes tumours of the caecum, transverse colon and hepatic flexure. tissue were analysed for ATase activity using calf thymus DNA methylated in vitro with N-nitroso-N-( 3H)-methylurea (approx. 20 Ci/mmol) as the substrate. Details of the assay procedure are described in full elsewhere (Watson and Margison, 1999). Associations were evaluated between ATase activity in either normal or tumour tissue and exposure variables (age, sex, tissue site, smoking, alcohol consumption and medication). Cigarette smoking and alcohol consumption were examined both as dichotomous (current vs. never/ex) and as polychotomous variables (non-smokers/exsmokers, 1 to 19 and ⱖ20 cigarettes per day, and non-drinkers, ⬍10 units/week, ⬎10 units/week). Univariate analysis was carried out using parametric techniques. Linear regression analysis was then performed using variables identified in the initial univariate analysis (at a significance level of p ⬍ 0.1), adjusted for age and gender. RESULTS ATase activity was detected in all colorectal normal and tumour tissue samples analysed, with levels ranging from 35 to 451 fmol/mg protein (Table I). In normal tissues, ATase activity varied by about 4-fold in the colon and 6-fold in the rectum, whereas in tumour tissue the corresponding figures were 13- and 7-fold, respectively. Differences in mean ATase activity within the large bowel were also observed in individuals with colorectal tumours (Table I). In particular, ATase activity in normal rectal tissue (but not tumour tissue) was significantly higher than that in normal tissue from the sigmoid colon (148 ⫾ 76 vs. 100 ⫾ 40 fmol/mg protein, p ⫽ 0.01), whereas ATase levels within different regions of the colon (proximal vs. sigmoid colon) were similar. There was no detectable difference in normal tissue ATase activity between individuals with benign or malignant colon disease (Table I). In the sigmoid colon but not the proximal colon or rectum, tumour ATase activity was significantly higher than normal tissue activity (144 ⫾ 115 vs. 100 ⫾ 40 fmol/mg protein, p ⫽ 0.04). There was no association between ATase levels in normal and tumour tissue from the same individual when the samples were located in the proximal colon (Fig. 1a), but there was a strong positive association in the sigmoid colon (r ⫽ 0.80, Fig. 1b) and rectum (r ⫽ 0.59, Fig. 1c). ATase levels in both normal and tumour tissue were slightly higher in men (between 1.2- and 1.4-fold) than in women, but these differences were not significant (data not shown). ATase activity in normal colon tissue increased significantly with age (Fig. 2a): a small but statistically significant positive association between ATase activity and age was detected in men ( p ⫽ 0.029) but not in women ( p ⫽ 0.19). Activity in colon tumour tissue varied little with age (Fig. 2b), whereas ATase activity in both normal and tumour tissue from the rectum decreased with age but not significantly (Fig. 2c,d). ATase activity in normal and tumour tissue from individuals with colon disease or rectal tumours was similar in drinkers and non-drinkers: in addition, ATase activity was not related to the 70 POVEY ET AL. FIGURE 2 – Relationship between ATase activity and age in samples from (a) normal colon tissue, (b) tumour tissue from the colon, (c) normal rectal tissue and (d) tumour tissue from the rectum. amount of alcohol consumed (data not shown). In the colon, ATase activity was higher in smokers than in non/ex-smokers in both normal [mean ⫾ SD 119 ⫾ 49 (n ⫽ 14) vs. 99 ⫾ 37 (n ⫽ 36) fmol/mg protein, p ⫽ 0.12] and tumour tissue [210 ⫾ 173 (n ⫽ 8) vs. 127 ⫾ 65 (n ⫽ 26) fmol/mg protein, p ⫽ 0.04]. However, there did not appear to be a dose-response relationship between the extent of smoking and ATase activity (Fig. 3a): highest ATase levels were obtained in individuals who smoked 1 to 19 cigarettes per day. In the rectum, there were no significant differences in the level of ATase activity between smokers and non/ex-smokers in either normal [150 ⫾ 68 (n ⫽ 8) vs. 147 ⫾ 81 (n ⫽ 18) fmol/mg protein, p ⫽ 0.93] or tumour [200 ⫾ 78 (n ⫽ 8) vs. 151 ⫾ 94 (n ⫽ 13) fmol/mg protein, p ⫽ 0.24] tissue, and increasing consumption of cigarettes also did not alter ATase activity significantly (Fig. 3b). The association between ATase activity and the use of medications was then examined. Evidence that medications influenced ATase activity was found only for anti-histamines, anti-muscarinics and benzodiazepines (Table II). ATase activity in normal and tumour colon tissue (but not rectal tissue) was higher in individuals receiving anti-histamine medication, but these results were of borderline significance. In normal colon tissue, ATase activity was 137 ⫾ 40 (n ⫽ 5) in users vs. 101 ⫾ 40 fmol/mg protein (n ⫽ 35) in non-users ( p ⫽ 0.07); in tumour tissue from the colon, the corresponding figures were 220 ⫾ 67 (n ⫽ 5) vs. 134 ⫾ 108 (n ⫽ 23, p ⫽ 0.09). Individuals using anti-muscarinic medication had significantly lower ATase activity in tumour tissue from the rectum, 143 ⫾ 91 (n ⫽ 14) in users vs. 223 ⫾ 61 fmol/mg protein in non-users (n ⫽ 7, p ⫽ 0.05), but there were no detectable effects in normal tissue from the rectum or normal and tumour tissue from the colon. ATase activity in normal colon tissue was lower in individuals receiving benzodiazepines [89 ⫾ 37 (n ⫽ 13) vs. 111 ⫾ 41 (n ⫽ 39), p ⫽ 0.08], but no effects were observed in colon tumour tissue or normal and tumour tissue from the rectum. FIGURE 3 – Effect of cigarette smoking in ATase activity in normal and tumour tissue from (a) colon and (b) rectum. In the colon, ATase activity in smokers who smoked ⬍20 cigarettes per day was significantly higher than that in non-smokers in both normal ( p ⫽ 0.03) and tumour ( p ⫽ 0.006) tissue. No significant difference in activity was observed between non-smokers and smokers who smoked ⱖ20 cigarettes per day ( p ⬎ 0.38) and between smokers who smoked ⬍20 or ⱖ20 cigarettes per day ( p ⬎ 0.2) in either normal or tumour tissue. In the rectum, there was no association between ATase activity and the numbers of cigarettes smoked. In a multivariate analysis, ATase activity was associated with age and current smoking in normal colon tissue but only with use of anti-histamine medication in tumour tissue (Table III). In the rectum, tumour ATase activity was associated with both age and use O 6-ALKYLGUANINE-DNA ALKYLTRANSFERASE IN COLORECTAL CANCER 71 TABLE II – ATase ACTIVITY IN COLON AND TUMOUR TISSUE AS A FUNCTION OF MEDICATION ATase activity (fmol/mg protein)1 Medication Colon tissue Anti-histamine Use Non-use Anti-muscarinic Use Non-use Benzodiazepines Use Non-use Rectal tissue Normal Tumour Normal Tumour 137 ⫾ 40 (5) 101 ⫾ 40 (45) 220 ⫾ 67 (5) 134 ⫾ 108 (23) 163 ⫾ 84 (5) 145 ⫾ 76 (21) 195 ⫾ 106 (4) 164 ⫾ 88 (7) 101 ⫾ 41 (29) 111 ⫾ 41 (21) 126 ⫾ 95 (17) 169 ⫾ 112 (17) 146 ⫾ 88 (18) 153 ⫾ 43 (8) 143 ⫾ 91 (14) 223 ⫾ 61 (7)2 89 ⫾ 37 (13) 111 ⫾ 41 (39) 156 ⫾ 145 (8) 143 ⫾ 91 (26) 150 ⫾ 68 (13) 147 ⫾ 86 (13) 187 ⫾ 96 (11) 151 ⫾ 82 (10) 1Results are expressed as mean ⫾ SD (n).–2Activity significantly higher in non-users than users of the medication ( p ⫽ 0.05). TABLE III – REGRESSION ANALYSIS FOR THE EFFECT OF AGE, GENDER, CURRENT SMOKING AND USE OF MEDICATIONS ON ATase ACTIVITY IN COLORECTAL TISSUES Site Normal tissue Variable b (95% CI) Colon Age1 Rectum Current smoker —3 0.79 (0.19, 1.39) 24.1 (⫺1.3, 49.7) — Tumour tissue p 0.01 0.06 — Variable Use of anti-histamines2 Age4 Use of anti-muscarinics b (95% CI) 95.2 (0.4 to 190.8) ⫺2.6 (⫺5.1 to ⫺0.02) ⫺99.6 (⫺26.8 to ⫺172.5) p 0.05 0.05 0.01 Variables excluded from the models: 1gender ( p ⫽ 0.22), use of benzodiazepines ( p ⫽ 0.19) and use of anti-histamines ( p ⫽ 0.11); 2age ( p ⫽ 0.46), gender ( p ⫽ 0.27) and current smoking ( p ⫽ 0.17); 3age ( p ⫽ 0.54) and gender ( p ⫽ 0.55); 4gender ( p ⫽ 0.38). of anti-muscarinic medication, but there was no detectable association between normal tissue activity and any study variable (Table III). We have reported that ATase activity in normal colorectal tissue was lower in individuals with a K-ras GC = AT transition mutation in the colorectal tumour than in individuals with other mutations or no mutation (Jackson et al., 1997), suggesting that high levels of ATase activity provide protection against the formation of an alkylating agent–induced K-ras GC = AT transition mutation. As there is a difference between ATase levels in the colon and rectum, we re-analysed the earlier mutation data for each tissue site (i.e., colon and rectum). In the colon, 6 of 17 individuals with normal tissue ATase activity below the median value (93.5 fmol/mg protein) had this GC = AT transition mutation compared to 1 of 19 individuals whose ATase activity was above the median ( p ⫽ 0.04, Fisher’s exact test). In the rectum (median normal ATase activity ⫽ 128 fmol/mg protein), 6 of 13 individuals with ATase activity below the median had the mutation compared to 2 of 12 whose ATase activity was greater than the median, a difference that did not reach statistical significance ( p ⫽ 0.20, Fisher’s exact test). DISCUSSION Studies in different populations have reported that mean ATase activities in normal colon tissue range from approx. 100 to 200 fmol/mg protein (this study; Myrnes et al., 1983, 1984; Gerson et al., 1986; Souliotis et al., 1992; White et al., 1997), whereas in tumour tissues, activity has ranged from approx. 145 to 990 fmol/mg protein (this study; Myrnes et al., 1983, 1984; Gerson et al., 1995; White et al., 1997). Inter-individual variations in activity from both normal and tumour tissue from the colon are large, being between 2- and 18-fold and 2- and 33-fold, respectively (Myrnes et al., 1983, 1984; Grafstrom et al., 1984; Gerson et al., 1986, 1992, 1995; Redmond et al., 1991; Chen et al., 1992; Souliotis et al., 1992; Zaidi et al., 1996; White et al., 1997); inter-individual variations in tumour activities were greater than those observed in normal tissue in all but one study (Souliotis et al., 1992). Two studies have also reported the absence of detectable ATase activity in a small number of individual normal and tumorous colon tissue specimens (Gerson et al., 1995; White et al., 1997). In contrast with other tissues, this phenotype is relatively rare, being present in ⬍1% of normal ATase tissue samples (2/352) and 3% of tumour samples (7/286), but this difference in frequency was not significant ( p ⫽ 0.09, Fisher’s exact test) (Myrnes et al., 1983, 1984; Grafstrom et al., 1984; Gerson et al., 1986, 1992, 1995; Redmond et al., 1991; Chen et al., 1992; Souliotis et al., 1992; Zaidi et al., 1996; White et al., 1997). The causes of the variation in activity of normal tissues and the larger variation in tumour tissues remains to be determined, but it is likely to reflect the fact that tumour samples are more heterogenous and contain different cells with varying ATase levels (e.g., Zaidi et al., 1996). It has also been reported in one study (Gerson et al., 1995) but not confirmed (Gerson et al., 1992) that ATase activity decreases along with advanced stages of disease. Previous studies have indicated that ATase activity does not vary within the human large bowel (Redmond et al., 1991; Gerson et al., 1995), and we also observed no significant variation in ATase activity within the mouse colon (data not shown). Our own re-analysis of the data of Redmond et al. (1991), however, suggests that ATase activity in normal caecal tissue (2.6 ⫾ 0.6 fmol/µg) was significantly lower than that in normal tissue from either the sigmoid colon (4.9 ⫾ 0.8 fmol/µg, p ⫽ 0.03, Mann-Whitney Utest) or rectum (4.8 ⫾ 1.3 fmol/µg, p ⫽ 0.01, Mann-Whitney U-test). In this study, we found that ATase activity in normal rectal tissue was much higher than that in the sigmoid colon, but there were no differences between the sigmoid colon and the proximal colon (i.e., caecum, transverse colon and hepatic flexure). Comparison of ATase activity at different locations is generally carried out using tissue samples obtained from different individuals, and the pronounced inter-individual variation observed in ATase activity may obscure these differences. For example, differences between normal and tumour ATase activities in the colon are generally detected only in studies using paired normal and tumour tissue from the same individual (this study; Redmond et al., 1991) rather than non-paired samples [i.e., from different individuals (Gerson et al., 1992; Zaidi et al., 1996)]. Differences in ATase activity within the colon may thus be detected only with multiple sampling from the same individual. 72 POVEY ET AL. In the present study, we found that different factors affected ATase expression in normal and tumour tissue and in tissue from the colon and rectum. In normal colon tissue, ATase activity was positively associated with age and, to a lesser extent, current smoking (Table III), but no study variable was associated with ATase activity in normal rectal tissue. Previous reports found no association between age and ATase activity in either the lung (Drin et al., 1994) or the colon, though the location of the tissue samples within the colon was not stated (Gerson et al., 1995; White et al., 1997). Cigarette smoking was also associated with a 1.4-fold increase in ATase expression in peripheral lung and in ex-smokers, was shown to decrease with time since that person had stopped smoking (Drin et al., 1994). Tumour ATase activity, but not normal tissue activity, was modified by the use of anti-histamines and anti-muscarinic medication in the colon and rectum, respectively. Although this association may reflect differences in disease between users and non-users of the medications or other confounding factors, it is also possible that it represents a direct effect of the medications concerned. If the latter is true, then it may be possible to identify other agents that may selectively modify ATase activity in normal and tumour tissue (i.e., increased activity in normal tissue but decreased activity in tumour tissue), which would be of value in chemotherapy employing relevant alkylating agents. In addition, it may be possible to identify those individuals whose tumours may respond to alkylating agent chemotherapy by measuring ATase activity within a biopsy of normal colon tissue as there is a strong correlation between normal and tumour tissue ATase activity in the sigmoid colon and rectum. As the assay is relatively simple and results can be obtained within a few hours, this may also be of therapeutic value. Whether the phenotypic variation in ATase activity in normal human tissue results in increased susceptibility to alkylating agent–induced cancer remains to be determined. We have found that ATase activity in normal colorectal tissue was lower in individuals with a K-ras GC = AT transition mutation in the colorectal tumour than in individuals with other mutations or no mutation, suggesting that high levels of ATase activity provide protection against the formation of an alkylating agent–induced K-ras GC = AT transition mutation (Jackson et al., 1997). Re-analysis of these data suggests that the effect may be confined to the colon, and further work in other populations is required to confirm this hypothesis. It is also of interest both for cancer treatment and chemoprevention to further establish whether medications and dietary components can modulate ATase and other DNA repair enzyme levels (e.g., mismatch repair enzymes). ACKNOWLEDGEMENTS We thank Mr J. Davis for technical support and Dr J. So for advice in classifying the medications. REFERENCES CHEN, J.-M., ZHANG, Y.-P., WANG, C., SUN, Y., FUJIMOTO, J. and KENAGA, M., O 6-Methylguanine-DNA methyltransferase activity in human tumors. Carcinogenesis, 13, 1503–1597 (1992). DRIN, I., SCHOKET, B., KOSTIC, S. and VINCZE, I., Smoking-related increase in O 6-alkylguanine-DNA alkyltransferase activity in human lung. Carcinogenesis, 15, 1535–1539 (1994). GERSON, S.L., ALLAY, E., VITANTONIO, K. and DUMENCO, L.L., Determinants of O 6-alkylguanine-DNA alkyltransferase activity in human colon cancer. Clin. Cancer Res., 1, 519–525 (1995). GERSON, S.L., BERGER, N.A., ARCE, C., PETZOLD, S.J. and WILLSON, J.K.V., Modulation of nitrosurea resistance in human colon cancer by O 6methylguanine. Biochem. Pharmacol., 43, 1101–1107 (1992). GERSON, S.L., TREY, J.E., MILLER, K. and BERGER, N.A., Comparison of O 6-alkylguanine-DNA alkyltransferase activity based on cellular DNA content in human, rat and mouse tissues. Carcinogenesis, 7, 745–749 (1986). GRAFSTROM, R.C., PEGG, A.E., TRUMP, B.F. and HARRIS, C.C., O 6Alkylguanine-DNA alkyltransferase activity in normal human tissues and cells. Cancer Res., 44, 2855–2857 (1984). JACKSON, P.E., HALL, C.N., BADAWI, A.F., O’CONNOR, P.J., COOPER, D.P. and POVEY, A.C., The frequency of K-ras mutations and DNA alkylation in colorectal tissue from individuals living in Manchester. Mol. Carcinogenesis, 16, 12–19 (1996). JACKSON, P.E., HALL, C.N., O’CONNOR, P.J., COOPER, D.P., MARGISON, G.P. and POVEY, A.C., Low O 6-alkylguanine DNA-alkyltransferase activity in normal colorectal tissue is associated with colorectal tumours containing a GC = AT transition in the K-ras oncogene. Carcinogenesis, 18, 1299–1302 (1997). KARRAN, P. and BIGNAMI, M., DNA damage tolerance, mismatch repair and genome instability. Bioessays, 16, 833–839 (1994). MIYAKAWA, H., LIU, J., NOGUCHI, O., MARUMO, F. and SATO, C., Effect of alcohol drinking on gene expression of hepatic O 6-methylguanine DNA methyltransferase in chronic liver diseases. Alcohol clin. exp. Res., 20 (Suppl. 9), 297A–300A (1996). MYRNES, B., GIERCKSKY, K.-E. and KROKAN, H., Interindividual variation in the activity of O 6-methylguanine-DNA methyltransferase and uracil-DNA glycosylase in human organs. Carcinogenesis, 4, 1565–1568 (1983). MYRNES, B., NORSTRAND, K., GIERCKSKY, K.-E., SJUNNERSKOG, C. and KROKAN, H., A simplified assay for O 6-methylguanine-DNA methyltransferase activity and its application to human neoplastic and non-neoplastic tissues. Carcinogenesis, 6, 1061–1064 (1984). OTSUKA, M., ABE, M., NAKABEPPU, Y., SEKIGUCHI, M. and SUZUKI, T., Polymorphism in the human O 6-methylguanine-DNA methyltransferase gene detected by PCR-SSCP analysis. Pharmacogenetics, 6, 361–363 (1996). PEGG, A.E., DOLAN, M.E. and MOSCHEL, R.C., Structure, function and inhibition of O 6-alkylguanine DNA-alkyltransferase. Progr. Nucleic Acid Res. mol. Biol., 51, 167–223 (1995). QIAN, X.C. and BRENT, T.P., Methylation hot spots in the 58 flanking region denote silencing of the O 6-methylguanine-DNA methyltransferase gene. Cancer Res., 57, 3672–3677 (1997). RAFFERTY, J.A., CLARKE, A.R., SELLAPPAN, D., KOREF, M.S., FRAYLING, I.M. and MARGISON, G.P., Induction of murine O 6-alkylguanine-DNAalkyltransferase in response to ionising radiation is p53 gene dosedependent. Oncogene, 12, 693–697 (1996). REDMOND, S.M.S., JONCOURT, F., BUSER, K., ZIEMIECKI, A., ALTERMATT, H.-J., FEY, M., MARGISON, G. and CERNY, T., Assessment of P-glycoprotein, glutathione-based detoxifying enzymes and O 6-alkylguanine-DNA alkyltransferase as potential indicators of constitutive drug resistance in human colorectal tumors. Cancer Res., 51, 2092–2097 (1991). SOULIOTIS, V.L., ZONGZA, V., NIKOLOPOULOU, V. and DIMITRIADIS, G.J., Measurement of O 6-methylguanine-type adducts in DNA and O 6alkylguanine-DNA-alkyltransferase repair activity in normal and neoplastic human tissues. Comp. biochem. Physiol., 101B, 269–275 (1992). WATSON, A.J. and MARGISON, G.P., O 6-Alkylguanine-DNA alkyltransferase assay. In: R. BROWN and U. BÖGER-BROWN (eds.), Cytotoxic drug resistance mechanisms, pp. 167–178, Humana Press, Clifton, NJ (1999). WHITE, A., KIBITEL, J., GARCIA, Y., BELANICH, M., YAROSH, D., WIDEROFF, J., LEVIN, L., HELD, D., FUCHS, A. and CITRON, M., O 6-Alkylguanine-DNA alkyltransferase in normal colon tissue and colon cancer. Oncol. Res., 9, 149–153 (1997). ZAIDI, N.H., LIU, L. and GERSON, S.L., Quantitative immunohistochemical estimates of O 6-alkylguanine-DNA-alkyltransferase expression in normal and malignant human colon. Clin. Cancer Res., 2, 577–584 (1996).