Inhibition of O6-methylguanine-DNA methyltransferase increases ...

Carcinogenesis vol.20 no.12 pp.2355–2360, 1999

SHORT COMMUNICATION

Inhibition of O6-methylguanine-DNA methyltransferase increases azoxymethane-induced colonic tumors in rats

Ramesh K.Wali, Susan Skarosi, John Hart1, Yingchun Zhang1, M.Eileen Dolan, Robert C.Moschel2, Lan Nguyen, Reba Mustafi, Thomas A.Brasitus and Marc Bissonnette3 Department of Medicine and 1Department of Pathology, University of Chicago, Chicago, IL 60637 and 2Carcinogen-Modified Nucleic Acid Chemistry, ABL-Basic Research Program, NCI–FCRDC, Frederick, MD, USA 3To

whom correspondence should be addressed at: Department of Medicine, MC 4076, University of Chicago Hospitals and Clinics, 5841 South Maryland Avenue, Chicago, IL 60637, USA Email: [email protected]

Azoxymethane (AOM) causes O6-methylguanine adduct formation which leads to G→A transitions. Their repair is carried out by O6-methylguanine-DNA methyltransferase (MGMT). To evaluate the importance of this repair event in AOM-induced carcinogenesis, we examined the effect of O6-benzylguanine (BG), a potent inhibitor of MGMT, on colonic tumor development. Rats were treated weekly for 2 weeks at 0 and 24 h with BG (60 mg/kg body wt i.p.) or vehicle (40% polyethylene glycol, PEG-400), followed 2 h after the first dose of BG with AOM (15 mg/kg body wt) or vehicle (saline) i.p. Rats were killed 35 weeks later and tumors harvested and DNA extracted. In the AOM-treated groups, BG caused a significant increase in tumor incidence with tumors in 65.9%, versus 30.8% in the AOM/PEGtreated group (P < 0.05). In the BG/AOM group there was also a significant increase in tumor multiplicity, with 2.3 tumors/tumor-bearing rat, versus 1.6 tumors/tumorbearing rat in the AOM/PEG group (P < 0.05). Since O6methylguanine adducts can cause activating mutations in the K-ras and β-catenin genes, we examined the effects of BG on these mutations. In the BG group there were seven mutations in codon 12 or 13 of exon 1 of the K-ras gene in 51 tumors examined, compared with no K-ras mutations in 17 tumors analyzed in the AOM/PEG group (P ⍧ 0.12). In the BG/AOM group there were 10 mutations in exon 3 of the β-catenin gene among 48 tumors evaluated, compared with six mutations in 16 tumors analyzed in the PEG/AOM group (P ⍧ 0.16). In summary, MGMT inhibition increases AOM-induced colonic tumor incidence and multiplicity in rats.

Colonic carcinogenesis is a multistep process involving the progressive accumulation of activating mutations in protooncogenes and inactivating mutations in tumor suppressor genes. Elucidation of genetic and epigenetic events are important both to enhance our understanding of malignant colonic Abbreviations: AOM, azoxymethane; ASOH, allele-specific oligonucleotide hybridization; BG, O6-benzylguanine; MGMT, O6-methylguanine-DNA methyltransferase; PEG, polyethylene glycol; PM-RFLP, primer-mediated restriction fragment length polymorphism; SSCP, single strand conformational polymorphism. © Oxford University Press

transformation and to identify potential targets for chemopreventive strategies. The azoxymethane (AOM) experimental model of colon cancer has been widely used to investigate these events. AOM is a colonic procarcinogen that is ultimately metabolized in vivo to an active methylating agent. The principal target for AOM is thought to be DNA guanine, leading to the formation of O6-methylguanine adducts (1). If unrepaired, these adducts can mispair with thymine and, during DNA replication, cause G→A transitions. O6-Methylguanine DNA methyltransferase (MGMT) is the principal protein involved in the repair of O6-methylguanine adducts (2). MGMT transfers the methyl group from the O6-methylguanine to a cysteine acceptor site in MGMT. This methylation of the MGMT cysteine residue prevents any further repair by the transferase and likely induces a conformational change leading to rapid ubiquitin-mediated degradation (3). Since this reaction irreversibly inactivates the repair protein, the capacity for removal of methyl groups from these adducts depends on the number of active methyltransferase molecules. There is abundant evidence that MGMT regulates the toxic, mutagenic and carcinogenic potential of many alkylating agents. The direct toxic effect of these agents appears to require ineffective DNA repair since this toxicity is inhibited, while tumorigenicity is enhanced, in mice lacking DNA mismatch repair enzymes (4). That MGMT is required to prevent O6-methylguanine-induced mutations has been directly demonstrated by in vitro studies using CHO cells deficient in this repair protein (5). The role of MGMT in AOM-induced experimental colon cancer in rats has not previously been examined. Earlier studies, involving other alkylating agents, were unable to demonstrate an effect of O6-benzylguanine (BG), a potent inhibitor of MGMT, on hepatic tumor incidence in rats despite a reduction in hepatic MGMT activity (6). The present studies were, therefore, undertaken to address the potential role of MGMT in AOM-induced colon cancer in rats. Since O6-methylguanine adducts induced by AOM have previously been shown to cause G→A transitions, and ultimately activating mutations in β-catenin and K-ras genes, it was also of interest to study the effect of BG on the mutational status of these genes. In initial studies we determined the kinetics of MGMT inhibition in rat colonic tissue by BG, which was synthesized as described (7). Our goal was to significantly inhibit MGMT activity for at least 48 h, the duration of the rat colonic epithelial cell cycle, to prevent repair of an AOM-induced mutational event (8). The dose of BG (60 mg/kg body wt) was chosen to achieve maximum inhibition of MGMT without compromising animal viability. Twenty-four Fisher 344 rats received BG solubilized in polyethylene glycol (PEG) buffer (40% PEG in 10 mM sodium phosphate buffer, pH 7.4). PEG buffer is the standard diluent for BG (9). A second injection of BG was administered 24 h later. Six rats were killed at each time point 0, 24, 48 and 72 h after the first injection. The colons and livers were resected, homogenates prepared and the MGMT assay was performed as described previously (7). 2355

R.K.Wali et al.

Table I. Effect of BG on MGMT activity in rat colon and liver Sample time (h)

Colonocyte MGMT (fmol/mg ⫾ SD)

Colonocytes (% control)

Liver MGMT (fmol/mg ⫾ SD)

0 24 48 72

8.8 ⫾ 2.9 NDa,b NDb 10.5 ⫾ 4.3

100 0 0 119

141.5 27.1 42.2 42.0

⫾ ⫾ ⫾ ⫾

Liver (% control)

56.3 23.3b 21.0b 33.1b

100 19 30 30

Rats were given BG, 60 mg/kg body wt i.p., at 0 and 24 h. Animals were killed at the indicated times. MGMT activities were measured in liver and colonocyte homogenates and expressed as means ⫾ SD. aND, not detectable. bP ⬍ 0.05, compared with untreated (0 time).

Table II. Tumor incidence and K-ras mutational status Group

Total

No. with tumors

Tumor incidence

Tumor/TBRa

K-ras wild-type

K-ras mutants

AOM⫹PEG AOM⫹BG

39 44

12 29

30.8 65.9b

1.6 2.3b

17 45

0 7

aTBR, tumor-bearing bP ⬍ 0.05 compared

rat. with PEG ⫹ AOM.

Table III. PEG/AOM tumorsa Tumor

Tumor grade

Ras

β-catenin

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17

TA w/CIS Ad CA MD Ad CA WD Ad CA TA w/CIS TA WD Ad CA WD Ad CA Ad CA WD Ad CA TA WD Ad CA Md Ad CA TA w/CIS TA w/CIS TA w/CIS WD Ad CA

WT WT WT WT WT WT WT WT WT WT WT WT WT WT WT WT WT

WT WT WT WT WT Mutant WT Mutant Mutant b

WT WT WT Mutant WT Mutant Mutant

TA, tubular adenoma; CIS, carcinoma in situ; Ad Ca, adenocarcinoma; MD Ad Ca, moderately differentiated adenocarcinoma; WD Ad Ca, well-differentiated adenocarcinoma; WT, wild type. aIn the PEG-AOM group, 17 of 19 tumors were evaluated for β-catenin and/or K-ras mutations. bNot evaluated.

To examine the effect of BG on O6-methylguanine adduct formation, 12 male Fisher 344 rats were divided into a control and two treated groups of four rats each. For the treated groups, at time 0 four rats received BG (60 mg/kg body wt) and four rats received the standard diluent, PEG buffer. Two hours later each rat in the treated groups received AOM (15 mg/kg body wt i.p.). Control and AOM ⫹ BG- or AOM ⫹ PEG-treated rats were killed 24 h after the AOM injection. Colonocyte DNA was extracted and the O6- and N7-methylguanine adducts measured as described (10). Based on the time course for MGMT inhibition by BG, the following experimental design was used to assess the effect of MGMT inhibition on the development of AOM-induced colonic tumors. One hundred and twenty male Fisher 344 rats, initially weighing 80–100 g, were randomized into four 2356

equal groups and given the following treatments at weeks 1 and 2. At the beginning of week 1, groups 1 and 2 were given 60 mg/kg body wt BG i.p. at time 0, followed 2 h later by AOM i.p. (15 mg/kg body wt) in group 1 and saline in group 2. To prolong the MGMT-deficient state in these rats one additional dose of BG was administered 24 h later. At the beginning of week 2, rats in groups 1 and 2 were treated a second time with the same regimen of BG, followed by AOM or saline, respectively. Groups 3 and 4 received PEG buffer, followed by AOM (group 3) or vehicle (group 4) treatments, as described for those in groups 1 and 2, respectively. Rats were maintained on standard rat chow (no. 5001; Purina Mills, Richmond, IN) and all animal procedures followed the guidelines approved by the University of Chicago Animal Care Committee. All rats were killed in the non-fasted state 35 weeks after the second carcinogen injection. Colons were removed and examined macroscopically for the presence of tumors. Tumors were rapidly excised, weighed and washed with ice-cold phosphate-buffered saline. A small portion from each tumor was fixed overnight at 4°C in 10% buffered formalin for microscopic examination, while the remainder was snap frozen in liquid nitrogen and stored at –80°C for later analyses. After formalin fixation, tissue specimens were paraffin embedded, sectioned and stained with hematoxylin and eosin, as described previously (11). All specimens were evaluated by a pathologist (J.H.) who was unaware of the treatment groups. Macroscopic lesions were classified as either benign (adenoma) or malignant (carcinoma in situ or adenocarcinoma) (12). To investigate K-ras mutations in the first or second position of codons 12, 13 and 59, a sequence from K-ras exon 1 (codons 12 and 13) or exon 2 (codon 59) was amplified by PCR and probed for K-ras mutations by allele-specific oligonucleotide hybridization (ASOH) as described (13). To confirm the ASOH results, we employed primer-mediated restriction fragment length polymorphism (PM-RFLP) for the detection of K-ras mutations in codons 12 and 13 (14). To assess mutations in CTNNB1, the gene coding for β-catenin, we employed single strand conformational polymorphism (SSCP) analysis after PCR amplification of exon 3

Role of DNA repair in AOM-induced colon cancer

Table IV. BG/AOM tumorsa Tumor

Tumor grade

Ras

β-catenin

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51

WD Aden CA TA w/CIS TA w/CIS TA w/CIS TA w/CIS WD Aden CA WD Aden CA TA w/CIS TA w/CIS WD Aden CA WD Aden CA TA w/CIS TA w/CIS WD Aden CA TA w/CIS TA w/CIS TA w/CIS WD Aden CA WD Aden CA WD Aden CA TA w/CIS TA MD Aden CA TA w/CIS TA w/CIS WD Aden CA TA WD Aden CA TA w/CIS MD Aden CA MD Aden CA TA w/CIS TA w/CIS WD Aden CA WD Aden CA Aden CA WD Aden CA WD Aden CA WD Aden CA WD Aden CA Aden CA TA w/CIS TA TA w/CIS MD Aden CA WD Aden CA WD Aden CA WD Aden CA WD Aden CA Aden CA TA w/CIS

WT WT WT WT WT WT WT WT WT WT K12 mutant WT WT WT WT WT WT WT WT WT WT WT WT WT WT K12 mutant WT WT K13 mutant K12 mutant WT WT WT WT WT WT WT K12 mutant WT WT WT WT WT K12/K13 mutant WT WT WT WT WT WT K13 mutant

WT WT WT b

WT b

WT WT WT Mutant WT WT WT WT WT WT WT b

WT WT WT WT WT WT WT Mutant WT WT WT Mutant WT WT Mutant WT WT WT WT WT Mutant Mutant Mutant Mutant Mutant Mutant WT WT WT WT WT WT WT

For abbreviations of histological classification, see Table III. aIn the BG-AOM group, 51 of 66 tumors were evaluated for β-catenin and/ or K-ras mutations. bNot evaluated.

of this gene exactly as described (15). K-ras and β-catenin mutations were confirmed by direct sequencing of amplified DNA using automated fluorescent DNA sequencing. Values were expressed as means ⫾ SEM. Differences between groups were compared by Fisher’s exact test, with P ⬍ 0.05 considered statistically significant. As shown in Table I, administration of BG abolished colonic MGMT activity for at least 48 h. By 72 h MGMT activity in the colon had returned to pretreatment levels. In contrast, BG treatment failed to completely suppress liver MGMT activity, which at baseline was ⬎10-fold higher than that of colon. Based

Fig. 1. K-ras mutational status determined by allele-specific oligonucleotide hybridization. DNA was extracted from flash frozen tumors in Tri-Reagent and K-ras exon 1 amplified by PCR. Amplified DNA was denatured and immobilized on nitrocellulose and subsequently hybridized with 32P-labeled oligonucleotides flanking K-ras codons 12 and 13. Shown is an autoradiogram of representative blots probed with the indicated wild-type or mutant 32P-labeled oligonucleotides followed by high stringency washes. Tumors in lanes 1–4 express codon 12 GGT→GAT mutations; lanes 4–5 are tumors with codon 13 GGC→GAC mutations; lanes 6–10 are tumors with wild-type K-ras, with codons 12 and 13 GGT and GGC, respectively.

Fig. 2. K-ras mutational status determined by PM-RFLP. DNA from AOM-induced tumors was amplified by PCR. PCR products were incubated overnight in buffer alone or with the indicated restriction enzymes and then resolved on a 4% agarose-1000 gel. The uncut amplified DNA runs at 116 bp, whereas fragments cut with BstN1 or Bgl1 migrate at ~87 bp. The four tumors in the upper panel are wild-type for K-ras. The first two tumors in the lower panel have mutations in codon 12, with incomplete cutting by BstN1, and the last two tumors in the lower panel have mutations in codon 13 with incomplete cutting by Bgl1.

on the kinetics of BG inhibition of MGMT, we administered BG 2 h prior to AOM and 24 h after carcinogen injection to maintain low levels of colonic MGMT activity during one cycle of cell division (8). Thirty-five weeks after the second AOM treatment, control and carcinogen-treated rats were killed. Control colonocytes and tumors were harvested. There were no significant differences in weight gain among the rats treated with AOM ⫹ BG or AOM ⫹ PEG, compared with the control groups treated with PEG or BG alone (data not shown). There were no tumors that developed in the rats receiving PEG or BG without carcinogen. In the carcinogentreated groups, however, BG significantly increased the tumor incidence, with tumors in 29 of 44 rats (65.9%), compared with only 12 of 39 rats with tumors (30.8%, P ⬍ 0.05) in the group receiving PEG (Table II). In addition, BG was associated with a significant increase in tumor multiplicity, with 66 tumors in 29 rats, compared with 19 tumors in 12 rats (P ⬍ 0.05), i.e. 2.3 and 1.6 tumors/tumor-bearing rat, respectively (Table II). Histological examination of the tumors revealed no significant differences in the incidence of adenocarcinomas, 2357

R.K.Wali et al.

Fig. 3. Detection of β-catenin mutations by SSCP. DNA was extracted in TRI-Reagent from AOM-induced tumors and exon-3 of the β-catenin gene amplified by PCR in the presence of [α-32P]dCTP. After denaturation, the PCR products were resolved on a 6% polyacrylamide gel and an autoradiogram prepared. C, control sample of colonic DNA from a PEG/vehicle-treated rat. Lanes marked with d indicate tumors with β-catenin mutations. In several lanes without detectable bands, DNA samples failed to amplify.

carcinoma in situ or adenomas between the two carcinogentreated groups (Tables III and IV). AOM is known to cause O6-methylguanine adduct formation in this experimental model which, if unrepaired by MGMT, leads to G→A transitions and ultimately K-ras and β-catenin mutations (15–17). Since BG inhibited colonic MGMT during tumor initiation, we speculated that this treatment might be associated with increased AOM-induced K-ras and/or βcatenin mutations. As shown in Figure 1 and Tables II–IV, and as predicted, BG treatment was associated with a numerical increase in K-ras mutations, which were observed in seven of 51 tumors examined (13.7% positive), compared with no mutations found in 17 tumors from the carcinogen-treated group receiving PEG (P ⫽ 0.12). Of the K-ras mutations, four occurred in codon 12, two in codon 13 and one tumor had mutations in both codons 12 and 13 (Table IV). As shown in Figure 2, the mutations identified by ASOH were all confirmed by PM-RFLP (17), as well as by direct sequencing of the PCR products (data not shown). As in the case of tumors with K-ras mutations, there was complete agreement between the ASOH and PM-RFLP techniques in identifying tumors expressing wild-type K-ras (Tables II–IV and Figures 1 and 2). The absence of K-ras mutations in codons 12 and 13 in the AOM ⫹ PEG-treated rats prompted us to screen for rarer mutations involving K-ras codon 59 (13), but none were found (data not shown). We reasoned that, in the absence of adequate MGMT activity during mutagenesis, G→A transitions involving the first position of codons 12 or 13, causing activating Gly→Ser substitutions, might also occur, although these have not been reported in the AOM model of colonic carcinogenesis. As in the case of codon 59, however, no such mutations were identified (data not shown). The frequency of K-ras mutations in this study is similar to that of Vivona et al. (18), but lower than that reported in other studies (16,17) using comparable AOM treatment regimens. The explanation for these differences is not apparent. We postulated that the increase in BG-associated, AOMinduced tumors, compared with the AOM ⫹ PEG-treated group, reflected an increase in O6-methylguanine adduct formation. To address this question we directly compared the effect of PEG versus BG on AOM-induced formation of O6methylguanine adducts. At 24 h, the mean of two independent determinations in duplicate of colonic O6-methylguanine adduct formation was 97 ⫾ 16 µmol/mol guanine for the 2358

PEG ⫹ AOM group, versus 248 ⫾ 49 µmol/mol guanine for the BG ⫹ AOM group (P ⬍ 0.05). These results confirmed that BG did, indeed, increase AOM-induced O6-methylguanine adduct formation in rat colon. As shown in Figure 3, and summarized in Tables II–IV, AOM was also found to induce mutations in β-catenin. The numbers of these mutations in the PEG/AOM (37.5%) and BG/AOM (20.8%) groups were not significantly different (P ⫽ 0.16). Surprisingly, however, in contrast to K-ras mutations, there were numerically fewer β-catenin mutations in the BG/AOM group. The incidences of β-catenin mutations in adenomas and carcinomas/carcinomas in situ were comparable, occurring in 40 versus 27.1%, respectively (P ⫽ 0.43). Of the total number of tumors assessed for both β-catenin and K-ras mutations, 44 were found to be wild-type for both, 13 had β-catenin but not K-ras mutations, four had K-ras mutations but were wild-type for β-catenin and in three tumors both mutations were present. There were no significant statistical associations between these mutations (P ⫽ 0.23). While MGMT has been demonstrated to play an important role in both the toxicity and tumorigenicity of alkylating agents in other organs, using transgenic mice overexpressing this protein (19–21), as well as knockout mice with MGMT gene interruption (22–24), this is the first study to demonstrate the role of this repair protein in the development of colon cancer. Previous studies with AOM in transgenic mice have only examined the development of aberrant crypt foci, putative precursors of malignancy (25). Inhibition of colonic MGMT during the initiation phase of AOM colonic carcinogenesis was associated with more than a doubling of tumor incidence, as well as increased tumor multiplicity. The increased O6methylguanine adduct formation caused by AOM in the presence of BG, compared with the AOM/PEG group, most likely accounts for the observed differences in tumorigenicity induced by these respective protocols (26). The specificity of AOM for the colon may, at least in part, reflect the relatively low abundance of MGMT in this organ (27) compared with liver (Table I), for example, as well as the presence of alcohol dehydrogenase in the colonic epithelium (28). Interestingly, low MGMT activity in human colonic mucosa from individuals with colon cancer was associated with increased mutations in K-ras involving G→A transitions, but not G→C/T transversions, implying a role for derangements

Role of DNA repair in AOM-induced colon cancer

in MGMT-induced O6-methylguanine adduct repair in a subset of human sporadic colon cancers (29). K-ras mutations have been observed in both human (30,31) and experimental animal models of colon cancer (16,17,32). In the AOM model, O6-methylguanine adducts in the K-ras gene cause G→A transitions involving predominantly guanine bases in the second position of codon 12 or 13. In this model of colon cancer ras mutations are early events found in aberrant crypt foci, putative premalignant precursors (16,32). Previous studies have found that manipulations which promote or inhibit carcinogen-induced colon tumors cause parallel alterations with respect to K-ras mutations (17,18,33). Using a transgenic mouse model, Gerson and associates found that overexpression of human MGMT resulted in a significant reduction in both AOM-induced aberrant crypt foci formation and K-ras mutations (25). In the present study, we have examined the effect of BG on the development of AOM-induced K-ras mutations. While BG treatment was associated with both increased AOM-induced tumorigenicity and K-ras mutations at the expected codon positions, the latter could potentially account only in part for the increased tumor development, similar to the findings of others in the murine model (25). We, therefore, also investigated other potential K-ras mutations involving G→A transitions in codon 59, as well as first position nucleotide changes in codons 12 and 13, but none were found. Since mutations in the K-ras gene could potentially only account for a fraction of the increased tumors observed in the BG ⫹ AOM group, we also examined these tumors for mutations of CTNNB1, the gene coding for β-catenin. This gene has recently been found to be mutated in the AOM model in exon 3, which codes for the regulatory domain of β-catenin (15). Mutations of this gene have also been identified in sporadic human colon cancer (34). Surprisingly, MGMT inhibition did not influence the frequency of β-catenin mutations. Studies are in progress to further explore the role of β-catenin mutations and to identify other molecular targets of AOM involved in colonic tumorigenesis in this model. Recent studies of MGMT have indicated that the repair protein co-localizes with genes undergoing active transcription (35). In addition, a number of studies have found that the DNA context adjacent to the O6-methylguanine adduct is a major determinant of DNA repair efficiency by this protein (36,37). These factors may have contributed to differences in the sensitivity of AOM-induced K-ras and β-catenin mutations to MGMT inhibition by BG. In summary, these studies firmly establish the importance of MGMT to limit the colonic tumorigenesis caused by this carcinogen in this model. Acknowledgements The studies were funded, in part, by USPHS grants CA69532 (M.B.), CA36745 (T.A.B. and M.B.) and DK42086 (T.A.B.) (Digestive Disease Research Core Center) and the Samual Freedman Research Laboratories for Gastrointestinal Cancer Research (T.A.B.). T.A.B. is the recipient of a MERIT Award from the National Cancer Institute, NIH. NCI, DHHS, under contract with ABL (R.C.M.).

References 1. Suaeyun,R., Kinouchi,T., Arimochi,H., Vinitketkumnuen,U. and Ohnishi,Y. (1997) Inhibitory effects of lemon grass (Cymbopogon citratus Stapf) on formation of azoxymethane-induced DNA adducts and aberrant crypt foci in the rat colon. Carcinogenesis, 18, 949–955. 2. Pegg,A.E. and Byers,T.L. (1992) Repair of DNA containing O6-alkylguanine. FASEB J., 6, 2302–2310.

3. Srivenugopal,K.S., Yuan,X.H., Friedman,H.S. and Ali-Osman,F. (1996) Ubiquitination-dependent proteolysis of O6-methylguanine-DNA methyltransferase in human and murine tumor cells following inactivation with O6-benzylguanine or 1,3-bis(2-chloroethyl)-1-nitrosourea. Biochemistry, 35, 1328–1334. 4. Kawate,H., Sakumi,K., Tsuzuki,T., Nakatsuru,Y., Ishikawa,T., Takahashi,S., Takano,H., Noda,T. and Sekiguchi,M. (1998) Separation of killing and tumorigenic effects of an alkylating agent in mice defective in two of the DNA repair genes. Proc. Natl Acad. Sci. USA, 95, 5116–5120. 5. Ellison,K.S., Dogliotti,E., Connors,T.D., Basu,A.K. and Essigmann,J.M. (1989) Site-specific mutagenesis by O6-alkylguanines located in the chromosomes of mammalian cells: influence of the mammalian O6alkylguanine-DNA alkyltransferase. Proc. Natl Acad. Sci. USA, 86, 8620–8624. 6. Lijinsky,W., Pegg,A.E., Anver,M.R. and Moschel,R.C. (1994) Effects of inhibition of O6-alkylguanine-DNA alkyltransferase in rats on carcinogenesis by methylnitrosourea and ethylnitrosourea. Jpn J. Cancer Res., 85, 226–230. 7. Dolan,M.E., Moschel,R.C. and Pegg,A.E. (1990) Depletion of mammalian O6-alkylguanine-DNA alkyltransferase activity by O6-benzylguanine provides a means to evaluate the role of this protein in protection against carcinogenic and therapeutic alkylating agents. Proc. Natl Acad. Sci. USA, 87, 5368–5372. 8. Rijke,R.P., Plaisier,H.M. and Langendoen,N.J. (1979) Epithelial cell kinetics in the descending colon of the rat. Virchows Arch. B Cell Pathol. Incl. Mol. Pathol., 30, 85–94. 9. Dolan,M.E. and Pegg,A.E. (1997) O6-benzylguanine and its role in chemotherapy. Clin. Cancer Res., 3, 837–847. 10. Jacoby,R.F., Bolt,M.J., Dolan,M.E., Otto,G., Dudeja,P., Sitrin,M.D. and Brasitus,T.A. (1993) Supplemental dietary calcium fails to alter the acute effects of 1,2-dimethylhydrazine on O6-methylguanine, O6-alkylguanineDNA alkyltransferase and cellular proliferation in the rat colon. Carcinogenesis, 14, 1175–1179. 11. Sitrin,M.D., Halline,A.G., Abrahams,C. and Brasitus,T.A. (1991) Dietary calcium and vitamin D modulate 1,2-dimethylhydrazine-induced colonic carcinogenesis in the rat. Cancer Res., 51, 5608–5613. 12. Earnest,D.L., Holubec,H., Wali,R.K., Jolley,C.S., Bissonnette,M., Bhattacharyya,A.K., Roy,H., Khare,S. and Brasitus,T.A. (1994) Chemoprevention of azoxymethane-induced colonic carcinogenesis by supplemental dietary ursodeoxycholic acid. Cancer Res., 54, 5071–5074. 13. Jacoby,R.F., Llor,X., Teng,B.-B., Davidson,N.O. and Brasitus,T.A. (1991) Mutations in the K-ras oncogene induced by 1,2 dimethylhydrazine in preneoplastic and neoplastic rat colonic mucosa. J. Clin. Invest., 87, 624–630. 14. Bolt,M.J.G., Mailloux,R.J., Rasenick,M.M., Wali,R.K., Skarosi,S., Bissonnette,M., Brasitus,T.A. and Sitrin,M.D. (1998) Expression of G protein α subunits in normal rat colon and in azoxymethane-induced colonic neoplasms. Gastroenterology, 115, 1494–1503. 15. Takahashi,M., Fukuda,K., Sugimura,T. and Wakabayashi,K. (1998) βcatenin is frequently mutated and demonstrates altered cellular location in azoxymethane-induced rat colon tumors. Cancer Res., 58, 42–46. 16. Shivapurkar,N., Tang,Z., Ferreira,A., Nasim,S., Garett,C. and Alabaster,O. (1994) Sequential analysis of K-ras mutations in aberrant crypt foci and colonic tumors induced by azoxymethane in Fischer-344 rats on high-risk diet. Carcinogenesis, 15, 775–778. 17. Singh,J., Kulkarni,N., Kelloff,G. and Reddy,B.S. (1994) Modulation of azoxymethane-induced mutational activation of ras protooncogenes by chemopreventive agents in colon carcinogenesis. Carcinogenesis, 15, 1317–1323. 18. Vivona,A.A., Shpitz,B., Medline,A., Bruce,W.R., Hay,K., Ward,M.A., Stern,H.S. and Gallinger,S. (1993) K-ras mutations in aberrant crypt foci, adenomas and adenocarcinomas during azoxymethane-induced colon carcinogenesis. Carcinogenesis, 14, 1777–1781. 19. Dumenco,L.L., Allay,E., Norton,K. and Gerson,S.L. (1993) The prevention of thymic lymphomas in transgenic mice by human O6-alkylguanine-DNA alkyltransferase. Science, 259, 219–222. 20. Nakatsuru,Y., Matsukuma,S., Nemoto,N., Sugano,H., Sekiguchi,M. and Ishikawa,T. (1993) O6-methylguanine-DNA methyltransferase protects against nitrosamine-induced hepatocarcinogenesis. Proc. Natl Acad. Sci. USA, 90, 6468–6472. 21. Becker,K., Dosch,J., Gregel,C.M., Martin,B.A. and Kaina,B. (1996) Targeted expression of human O6-methylguanine-DNA methyltransferase (MGMT) in transgenic mice protects against tumor initiation in two-stage skin carcinogenesis. Cancer Res., 56, 3244–3249. 22. Tsuzuki,T., Sakumi,K., Shiraishi,A., Kawate,H., Igarashi,H., Iwakuma,T., Tominaga,Y., Zhang,S., Shimizu,S., Ishikawa,T., Nakamura,K., Nakao,K., Katsuki,M. and Sekiguchi,M. (1996) Targeted disruption of the DNA

2359

R.K.Wali et al. repair methyltransferase gene renders mice hypersensitive to alkylating agent. Carcinogenesis, 17, 1215–1220. 23. Iwakuma,T., Sakumi,K., Nakatsuru,Y., Kawate,H., Igarashi,H., Shiraishi,A., Tsuzuki,T., Ishikawa,T. and Sekiguchi,M. (1997) High incidence of nitrosamine-induced tumorigenesis in mice lacking DNA repair methyltransferase. Carcinogenesis, 18, 1631–1635. 24. Sakumi,K., Shiraishi,A., Shimizu,S., Tsuzuki,T., Ishikawa,T. and Sekiguchi,M. (1997) Methylnitrosourea-induced tumorigenesis in MGMT gene knockout mice. Cancer Res., 57, 2415–2418. 25. Zaidi,N.H., Pretlow,T.P., O’Riordan,M.A., Dumenco,L.L., Allay,E. and Gerson,S.L. (1995) Transgenic expression of human MGMT protects against azoxymethane-induced aberrant crypt foci and G to A mutations in the K-ras oncogene of mouse colon. Carcinogenesis, 16, 451–456. 26. Pegg,A.E. (1984) Methylation of the O6 position of guanine in DNA is the most likely initiating event in carcinogenesis by methylating agents. Cancer Invest., 2, 223–231. 27. Gerson,S.L., Trey,J.E., Miller,K. and Berger,N.A. (1986) Comparison of O6-alkylguanine-DNA alkyltransferase activity based on cellular DNA content in human, rat and mouse tissues. Carcinogenesis, 7, 745–749. 28. Sohn,O.S., Puz,C., Caswell,N. and Fiala,E.S. (1985) Differential susceptibility of rat and guinea pig colon mucosa DNA to methylation by methylazoxymethyl acetate in vivo. Cancer Lett., 29, 293–300. 29. Jackson,P.E., Hall,C.N., O’Connor,P.J., Cooper,D.P., Margison,G.P. and Povey,A.C. (1997) Low O6-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. 30. Bos,J.L., Fearon,E.R., Hamilton,S.R., Verlaan-de Vries,M., van Boom,J.H., van der Eb,A.J. and Vogelstein,B. (1987) Prevalence of ras gene mutations in human colorectal cancers. Nature, 327, 293–297.

2360

31. Forrester,K., Almoguera,C., Han,K., Grizzle,W.E. and Perucho,M. (1987) Detection of high incidence of K-ras oncogenes during human colon tumorigenesis. Nature, 327, 298–303. 32. Stopera,S.A., Murphy,L.C. and Bird,R.P. (1992) Evidence for a ras gene mutation in azoxymethane-induced colonic aberrant crypts in Sprague– Dawley rats: earliest recognizable precursor lesions of experimental colon cancer. Carcinogenesis, 13, 2081–2085. 33. Llor,X., Jacoby,R.F., Teng,B.B., Davidson,N.O., Sitrin,M.D. and Brasitus,T.A. (1991) K-ras mutations in 1,2-dimethylhydrazine-induced colonic tumors: effects of supplemental dietary calcium and vitamin D deficiency. Cancer Res., 51, 4305–4309. 34. Sparks,A.B., Morin,P.J., Vogelstein,B. and Kinzler,K.W. (1998) Mutational analysis of the APC/β-catenin/Tcf pathway in colorectal cancer. Cancer Res., 58, 1130–1134. 35. Ali,R.B., Teo,A.K., Oh,H.K., Chuang,L.S., Ayi,T.C. and Li,B.F. (1998) Implication of localization of human DNA repair enzyme O6methylguanine-DNA methyltransferase at active transcription sites in transcription-repair coupling of the mutagenic O6-methylguanine lesion. Mol. Cell. Biol., 18, 1660–1669. 36. Georgiadis,P., Smith,C.A. and Swann,P.F. (1991) Nitrosamine-induced cancer: selective repair and conformational differences between O6methylguanine residues in different positions in and around codon 12 of rat H-ras. Cancer Res., 51, 5843–5850. 37. Mironov,N.M., Bleicher,F., Martel-Planche,G. and Montesano,R. (1993) Nonrandom distribution of O6-methylguanine in H-ras gene sequence from DNA modified with N-methyl-N-nitrosourea. Mutat. Res., 288, 197–205. Received June 8, 1999; revised August 11, 1999; accepted August 19, 1999