Tobacco Documents Online

Relationship between acetylator status, smoking, diet and colorectal cancer risk in the north-east of England Mark R.Welfare~,2J, Jane Cooper2, Margaret F.BassendineI and Ann K.Dalyz Departments of IMedicin¢ and 2Pharmacological Science, University of Newcastle, Ne~vcastle-upon-Tyn¢ HE2 4HH, UK ~To whom correspondence should be addressed at: Academic Department, North Tyneside General Hospital, Rake Lane, North Shield¢ NE30 8NH, UK Some previous studies have suggested that the fast pheno- type of the N-acetyltransferase NAT2 may confer suscepti- bility to colorectal cancer because of greater activation of dietary heterocycHc amines, particularly in individuals who also consume well-done red meat, but other studies have not supported this. We describe a large case-control study examining the interaction between dietary, smoking and drinking habits, and acetylation genotype in relation to susceptibility to colorectal cancer. One-hundred-and- seventy-four incident cases and 174 matched controls were recruited. Genotyping for polymorphisms in NAT2 was performed using a method that detects >95% of slow alleles and data on personal habits were collected using a standardized questionnaire. We found no difference in the frequency of the fast acetylator genotype bet~veen cases and controls [odds ratio = 0.95 (95% CI 0.61-1.49)], and analysis by sex, age and site also revealed no difference in acetylator genotype. There was, however, considerable heterogeneity in dietary risk factors between fast and slow acetylators. Analysis by acetylator type shows that recent smoking was more frequent in slow acetylator cases than matched controls [OR = 2.31 (1.16-4~6)] and that heavy alcohol consumption was also more frequent in the slow acetylator cases than controls [OR = 2.5 (1.02-7.29)]. In contrast, frequent fried meat intake was seen more frequently in fast acetylator cases than matched controls [OR = 6.0 (1.34-55)]. The odds ratio for the combination of fast acetylator status and frequent fried meat consump- tion in cases was 6.04 (1.6-26). Our study suggests that there may be different risk factors for colorectal cancer in slow and fast acetylators, and reveals a new observation that slow acetylators may be at risk of colon cancer from smoking. In our community, the overall effect of acetylator status on colorectal cancer risk is neutral. Introduction Colorectal cancer is a major cause of death in many Western countries including the UK and USA. While up to 5% of cases are due to inherited forms of the disease, in particular familial adenomatous polyposis (FAP*) and hereditary non-polyposis colon cancer (HNPCC), the aetiology of the sporadic form of *Abbreviations: FAP, familial adenomatou$ polyposis; }-INPCC, hereditary non-polyposis colon cancer; MelQX, 2-amino-3,8-dimethylimidazo[4,5- .t]quinoxaline; Phi:P, 2-amino-l-methyl-6-phenylimid,azo[4,5-b]pyfidine; ABP, 4-aminobiphenyl. Oxford University Press the disease has proved difficult to determine although some evidence for increased relative risk for certain dietary, environ- mental and genetic factors have been reported (for review see reference 1). Of these factors, diet has been the most widely studied. Red meat intake has been shown to be associated with increased risk of the disease in a number of different studies, including case-control, cohort and migration (2-4). In addition, method and extent of cooking may affect risk of the disease with well-caoked, barbecued meat intake associated with the highest risk (5). An association between alcohol consumption and susceptibility to rectal cancer has also been reported in some, but not all studies (6), and the relationship between smo "king and colorectal cancer is also not clear (7). The increased risk of colorectaI cancer development associ- ated with eating well-done meat may reflect an important role for heteroeyclic amines as chemical carcinogens in this disease. Heterocyclic amines are formed as pyrolysis products during coo'king at high temperature and have been demonstrated to be both mutagenic and carcinogenic (8). I-Iigh concentrations of hetarocyclic amines such as 2-amino-3,8-dimethylimidazo[4,5- fJquinoxaline (MeIQX) and 2-amino-l-methyl-6-phenylimid- azo[4,5-b]pyridine (PhIP) have been demonstrated in grilled meat and fish and levels in the drippings and pan scrapings, which are often used for preparation of gravy, are particularly high (9-11). In the average US diet, the heterocyclie amine consumed at the highest concentration appears to be PhIP which has recently been demonstrated to give a specific signature mutation in the APC turnout suppressor gene when fed to rats (i2,I3). It is not -known whether the particular mutation also occurs in humans exposed to PhlP, but the APC gene is mutated in ~60% of human coloreetal tumours and mutation of this gene appears to be an important step in the carcinogenic process (14). There is also evidence that polymorphisms in genes encoding xenobiotic activating and metabolizing enzymes, particularly the cytochrome P450 CYPIA2, the N-acetyltransferases and the glutathione-S-transferases modulate risk of coIorectal cancer development. This relationship may be due to the role of these enzymes in heterocyclic amine activation and detoxication. In particular, it is known that heterocyclic amines undergo activation to mutageas in a series of reactions initiated by N- oxidation by CYP1A2 (15). The oxidized metabolite undergoes O-acetylation, either by the hepatic N-acetyltransferases NAT1 and NAT2 directly or it may first be glucuronidated, excreted in the bile, deconjugated by colonic bacteria and then finally undergo O-acetylation by colonic NAT's (16). Olutathione-S- transferases have been shown to prevent the formation of DNA-adduets in vitro, but whether this is relevant to the in vivo situation is not yet known (17). On present evidence, the polymorphism in N-acetyltrans- ferase 2 (NAT2) is the best candidate for a metabolic poly- morphism likely to modulate risk of colorectaI cancer developmem. At least seven different alleles resulting from single base-changes cause almost complete loss of enzyme 1351 This article is for individual use only and may not be further reproduced or stored electronically without wdtten permission from the copyright holder. Unauthorized reproduction may result in financial and other penalties. (c) OXFORD UNIV PRESS ENGLAND

M.R.Welfare et aL activity and possession of two mutant alleles restflts in the slow acetylator phenotype whereas possession of one or two wild-type alleles results in the fast acetylator type (18). The fast acetylator type is seen in between I0 and 90% of individuals, depending on ethnic origin. In Caucasians, -:45% of individuals are fast acetylators. Unlike some of the other polymorphisms suggested to be relevant to colorectal cancer, for which there is no mechanistic explanation, it is clear that NAT2 has a role in heterocyclic amine metabolism. The other human N-acetyitransferase NAT1 has also now been demonstrated to exhibit polymorphism (19) with the NAT!*IO allele encoding rapid metabolism. The one study reported so far suggests that the NATI*IO allele confers susceptibility to colorectal cancer with an odds ratio of 1.9 (20). There have been a number of previous studies on the NAT2 polymorphism in relation to colorectal cancer or polyp sasceptibiIity which have yielded conflicting results (21-30). Part of this inconsistency in results may have arisen from heterogeneity in risk between individuals or populations. For example, some of the negative studies have been carded out in countries with a high rate of fast acetylators and/or a low incidence of colorectal cancer, where factors other than meat may be important, such as Japan (21). Several of the positive studies in areas of high incidence have used phenotyping methods to determine acetylator status and there is some evidence that these methods may be influenced by disease or surgery (31). It seems noteworthy that all of the studies showing a positive association between fast acetylator status and colorectal cancer (22,23,25) have used phenotyping methods, whereas the four genotyping studies (21,24,27,28) and the three studies on polyps (25,29.30) have revealed no association. The study of interactions between dietary intakes and acetylator slams provides a more powerful method of analysis and allows dissection of heterogeneity in risk factors between individuals, Two recent phenotyping studies from the US and Australia (25,30) have indicated that the combination of the fast acetylator phenotype and high red meat intake is a particularly strong risk factor with an odds ratio of 3 (I.0- 13.1) found for consumption of meat in the highest fertile and fast acetylator status in one study (30). We have investigated the interaction between diet and acetylator status in a British population which has a similar genetic background to the US and Australian patients, but is I/kely to consume a diet with a lower heterocyctic amine content. We carried out a matched case-control study relating NAT2 genotype to diet, lifestyle and other factors, using larger numbers of cases and controls than in the previous studies. Materials and methods Cases We conducted a pmspectlv¢ case-conu'ol study in the Newcastle and Norda T~eside health disWicts over a 9 month perind. Two-hundred-and-thi~y-four incident cases were identified and 33 wer~ excluded because of concurrent colitis t4). inability to give consent of death before recrakment (9) or umvilling~ess to pa~icipate (20), 1~aving a total of 201 subjects. Age range was 38-91 years (median 69) and 58.5% were male. Of the total, 31.5% of the mmours were in the proximal coina. 37.9% ~ver~ in the distal colon and 30.5% were ia the r~ctum. There was no difference in the median a~ for any of the three sites or for eider sex. Amongst the female cases a higher propo~on wer~ proximal (44%) than in the ra~e cases (22%) log ffi 2.78 ¢ I.gl-5.53), P = 0.002]. A prior power calculatinn had detemdaed that ia a matched study design 170 cases were required to detect a 20% difference ia the NAT2 potymo~phism at the P < 0.0,5 ie~d with 90% confidence. I352 Controls Controls (n ffi 174) were individually matched to each ease lbr age and sex, and were recruited from the eonwaudity by choosing from the patient-llst of the general practitioner which their matebed ease attended. This method of recmitment was used to control fur social class and differences in diet between different sub-popuIadon~ in our area. Ninety-nine per earn of cases and controls were of Caucasian origin. Dietary and personal factors Diet in the year before interview was assessed by a food fmquancy questinn- nalre which focused on focd~ known to have a Idgh hetetocyelic atone content, The principaI food items assessed were the frequency of consump~on of red meal bacon, sausages, chicken, fish, brown gravy, fried, gtilIed or masted meat of my kind, and whether animal or vegetable fat was used for cooking. Questions on other food item~ were ineinded to obscure the hypothesis being t~sted, but the questionnaire was not subjected to_fofmal validation procedures. Cigarette smok.e~ were classified as ex-smokers of more or less than 5 years or as current smokers in the categories. 1-t0, tl-20 and >20 per day. Pipe and cigar smokers were classtfiad separately. NA~2 geno~Ttng Venous blood was taken t~om each subject and DNA was prepared by standard methodology. Geno~ype for HAT2 was detera~ed by a PCR-restric~on digest method that detec~ ~ 97% of slo~v acetylators in Caucasia~ pol:~Iations. PC~ was performed using primers NatHnI# and Haft-In 16 as described by t-Iickmaa and Sire (32). The a~eles NAT2*SA a~d NAT2*$B were detected ~y digest of this PCR product with the ms~ctio~ enzyme KpnL NAT*6A was detected by digest with TaqL and NAT~TB was detected ~vith BamI~. Digas~on was performed for >3 h at 37~C for Kpl~rl and BamHI, and at 65~C for ~-qI. ~'he prodnc~s of the three separate digests were run on 2.5% agaruse gels.., stained with ethidi~m bron~de and visnaJ.ized on a UV tra~illua~aator. Stattsttcol analysis l~ast acetylators are defined as havin~ one or two wild-type "alleles, and slow ecetylators as having none. All dietary and genotype data were diehtomlzad. and the proportions between cases and eontm/s were compared by McNemers test for comparison of matched p~2rs. Comparison~ between sub-groups of c~es uti~ chl-squa~d testing. Results NAT2 results Analysis by McNemars test showed that there was no signific- ant difference between" eases and matched controls in the proportion of individuals with the fast acetylator status. The fast acetylator genotype was present in 42% of cases and 42.5% of controls [OR = 0.95 (95% CI 0.61-1.49)3 (Table I). No significant differences in the proportions of fast acetylators was found when cases and individually matched controls were compared by McNemars test for each of the three age groups, when each sex was examined alone or when the location of the tumours was divided into proximal and distal colon and rectum. Data are given in TabIe I. There was no difference in the proportion of fast aeetylators between patients with Dukes stage A or B (43.5%) and patients with Dukes C and "D' (35.7%) [OR = 1.38 (0.72-2.66)]. Dietary data Dietary data were converted to binary variables in two categor- ies to enable testing by McNemars test. The categories were consumption more than twice a week or tess than twice a 3yeek, and consumption once or more a week or less than once a week. Seventeen variables were examined, but overall the only statistically significant variables W_e_re_ the consumption • of fried meat more than twice a week [OR = 3.0 (1.37-~7.25)], the regular use of animal fat products in cooking [OR. -- 1.68 (1.0-2.88)] and smoking cigarettes within the Iast 5 years [OR = 1.77 (1.03-3.I4)]. Comparison between eases at each site with their individually matched controls revealed some heterogeneity in risk factors for each site (see TabIe II). For ....... the proximal colon consumption of roast meat [OR --" 3.0 This article is for individual use only and may not be further reproduced or stored electronically without written permission from the copyright holder. Unauthorized reproduction may result in financial and other penalties. (c) OXFORD UNIV PRESS ENGLAND

Susceptibility to colorectal cancer Table L Distribution of the fast acetylator status in cases and thdr individually matched controls overall, in three age groups, by gender and by site Age Cases Controls Odds ratio Overall 731174 (42%) 75/174 (42.5%) 0.95 (0.61-1.~9) <65 years 21/45 (46.7%) 19/45 (42.35) 1.22 (0,46-334) 6~-75 3~82 (41..5%) 40/82 (48.8%) 0.78 (0.42-1.43) >75 18/47 (38.3%) 16/47 (34%) 1.25 (0.44-3.6) Men 41/102 (40.2%) 41/I02 (41.2%) I (0.54-1.7) Women 3~72 (44A%) 33/72 (45.8%) I (0.43-2.02) Proximal 21151 (41,2%) 27151 (52.9%) 0.66 (0.29-1.47) Distal 25162 (40.3%) 27162 (43.6%) 0.85 (0.36-2.0) Rectal 20/48 (41.(5%) 16/48 (33%) 1.4 (0.58-3.5) Odds ratios and confidence intervals calculated by McNemar~ test. The site w~ not available for all cases. (1.14-9.23)] or gravy at least once a week [OR = 2.83 (1.07- 8.78)] and gravy at least twice a week [OR -- 3.2 (1.11- 1t.2)], and regular use of animal fat in cooking [OR = 3.0 (1.14-10.5)] were all significant risk factors, but there were no significant differences in any of the dietary factors, alcohol intake or smoking status between the distal cases or rectal cases and their individually matched controls. Comparison of each sex with their individually matched controls gave similar results to the overall dietary analysis. Comparison of dietary data by acetylator status in cases against controls There was considerable heterogeneity in risk factors between fast acetylator cases and slow acetylator cases (see Table H) when analysed by McNemars test. Cigarette smoking within the last 5 years was found more frequently in the slow acetylator cases (31/98, 31,8%) than in their individually matched controls (17/98, 17.3%) [OR -- 2.31 (1.I6-4.6)] as was drinking >20 units of alcohol a week [OR ~ 2.5 (1.02- 7.29)], but no dietary variables reached significance. In contrast, fried meat consumption more than twice a week was found more in the fast acetylator cases (13/72, 18.3%) than in theLr matched controls (3/72, 4%) [OR -- 6.0 (1.34-55)]. There was no relationship in fast ac¢tylators with recent smoking. Because of the small numbers in some of "these sub-groups the confid- ence intervals are wide. We also compared the proportions of eases and controls with combinations of risk factors using chi-squared tested, as opposed to the use of McNemars test above. This showed that the combination of fast acetyIator status and fried meat consumption more than twice a week was found in 7.4% of cases and 1.7% of controls giving an odds ratio of 6.04 (1.6- 26, P = 0.001) and that the combination of recent smok.ing and slow acetylator status was found in 17.8% of cases and 7.9% of controls giving an odds ratio of 1.86 (1.0-3.49). Discussion In our study, acetylator status alone was not found to be a risk factor for colorectal cancer. Overall, the most important environmental factors were frequent consumption of fried meat and smoking within the last 5 years. Analysis of acetyIator type revealed heterogeneity in risk factors between slow acetylators, who appear to be at risk from smoking cigarettes and high alcohol consumption, and fast acetylators, who we confirmed to be at risk from frequent meat consumption, particularly fried. 'lhble II. Dietary risk factors in various sub-groups of cases compared to individually matched controls using MeNem~rs test Sub-group Cases Controls Odds ratio (CI) Proximal colon cancer Gr~.vy weekly or more 34148 26/48 2.83 (L07-8.8) Gravy > twice a week 19148 8/48 3.2 (1.11-11.2) Roast meat weekly or more 38/48 26/48 3.0 (1.14-9.23) Slow acetylators Cigarette smoking in past 5 years 31/98 17/98 2,31 (1,16-4.6) >20 units of alcohol a week 25192 15/92 2.5 (1.02-7,29) Fast acetylators Fried meat > twice a week 13/72 3/72 6,0 (1.34-55) The first two studies which examined the relationship between fast acetylator status and colorectal cancer found .highly significant associations (22,23), but both of these studies were very small, and in both the frequency of the fast acetylator type in the control group was below that expected in Caucasian communities, so they are perhaps best regarded as pilot studies. Two larger studies from America and Australia (25,30), involving both patients w~th colorectal adenomas and cancers, have shown a significant interaction between meat intake and the fast acetylator phenotype as risk factors for disease susceptibility. However, several studies that have only analysed acetylator type have proved negative, both in countries with a high incidence of colorectal adenomas and cancer (27,29) and those with a low incidence (21,26). Our study reveals some other possible confounding factors in the analysis of dietary/moleeular-epidemiological studies of colorectal cancer. In particular, our finding that smoking was a risk factor in slow acetylators is a new and potentially important finding as there is a possible mechanistic explanation and because a recent study has shown that slow acetylators who smoke are at increased risk of breast cancer (33). Arylamines in tobacco smoke, such as 4-aminobiphenyl (ABP), have been recognized as occupational bladder carcinogens for many years. Unlike heterocyclic amines, which undergo O- acet~lation to a more active carcinogen, ABP is detoxicated by N-acetylation, and it is known that slow acetylators are at greater risk of both occupational and smoking-related bladder cancer (34). There has been conflicting evidence for the role of smoking as a risk factor for colorectal cancer previously (7) and our study may help to explain some of the heterogeneity in risk factors between populations. In addition, recent cohort studies have suggested that there is increased risk of small adenomatous poIyps after 20 years, of large polyps after 20- 30 ~/6ars and of cancer after only 35 years of smoking (35). There is, therefore, increasing evidence for the aetiological role of smoking in some cases of eolorectal cancer. Our study suggests that some of the population may be at risk of colorectal cancer if they smoke or drink heavily, and conffmas that the rest are at an even greater risk from eating fried meat. Most importantiy, however, our study has shown that in the population of the north-east of England the overall effect of acetylator status is neutral. Acknowledgement This work was supported by a grant from the North of E~gland Cancer Research Campaign. 1353 This article is for individual use only and may not be further reproduced or stored electronically without written permission from the copyright holder..Unauthorized reproduction may result in financial and other penalties. (c) OXFORD UNIV PRESS ENGLAND

M.R.'~Vell'are et al. References l.Potterj.D.. Slattery, M., Bosfick,R.M. and Gapsmr, S.M. 0993) Colon c~ncer: a review of the epidemiology. Ep[demial. Re~:. 15, 499-545. 2. Haesnzel,W., Berg,L, Segi,M. and Lee,tLK.C. (1973) Large bowel cancer in Hawaiian Japanese. 3". Nail Cancer Inst.. 51, 1765-1779. 3.Armstrong,B. and Dofi,R. (1975) Env/roamentol factors and cancer incidence and mortafity in different countries, with speclal reference to dietary practices. Int. J. Cancer. l& 617-63I. 4.Willet't,W.C.. Stamp(oriel J., Colditz.G.A.o RosnanB.A. and Sperzer, F.E. (1990) Relation of meat and fiber intake to the risk of colon cancer in a prospective study among women. N. EngL J. Med., 323, 1664-1672. 5.Gerhardason de Verdier.M., I.-Iagman.U., Peters,R.K., Steineck,G. and Overik,E. (1991) Meat, cooking methods and coloreetai cancer:, a case- referent study in Stockbolm. Int. Z Cancer. 49, 520-525. 6.Longaeeker.M.P. 0990) A case-control study of alcoholic beverage consumption in relation to risk of colon cancer and the recv.mL Cancer Ceases Control, 1, 5-14. 7.Newcomb,P.A., Storer, B.E. and Marcus.P.M. ¢ t995) Cigarette smoking in relation to risk of large bowel cancer in women. Cancer Res., 55, 4906-4909. 8.Gross,G.A., Taresky, RJ., Fay,LB., Stlilweli,W.G., Skipper.P2_. and Tannenbaum.S.R. (I993) Heterocyclie aromatic amine formation in grilled bacon, beef and fish and in grill scrapings. Cawinogenesis, 14, 2313-2318. 9.Sugimu.ra,T., Sato,S. and Wakabasl~,K. (1988) Mutageas/carcinogens in pyrolysates of amino acids and proteins in cooked foods: heterocyclic amines. In Woo, Y.T.. Lai,D.Y, Areosd.C. and Arg~s,M.F. (eds), Chemical Acetylalion phenotype in colorectai carcinoma. Cancer Res., 47, 1466- t469. 2.3.Langdq.P.. Chu.D.Z.J., Hunter, C.F., Kendall.D.C.. Flammang,TJ. and Kadinbar, F.F. (1986) Role of aromatic amine acetyl~'ansferase in human colorectai cancer. Arch. Sm~.. 121, I259-'[26I. 24.Rodriguez,LW., K.h-IIn0W.G., Ferguson,RJ., DolI.M.A., Gray,K, Rnstan,T.D., Lee.M.E., Kemp.K., Urso,P. and HeimD.W. (1993) He.man acetylator genotype: relationship to eolorectai qance.r_, incidence and atylamine N-acetyltransfemse expression in colon eytosol. Arcl~ ToxicoL, 67, 445~.5 I. 25. Robexts-Thomson.L C.. RyamP.. Khoo,K.K., Harl.Wj., MeMiehael,AJ. and Bntler, R.N. (1996) Die~. acetylator phenotype, and risk of eolorectal neoplasia. Lancet. 347. i1372-1374. 26, Laderoj.M., GonzaiezJ.F., Beaiter.~., Vargas,F-, Femandez,M.2., Baki,W. and Diaz-Rubio,M. O991) Acetylator polymorphism in human coloreeml carcinoma. Cancer Res.,.~l, 2098-2100. 27.Spm'r2,LK.. Gough,A.C., Chinegwundoh,EL and Dale-Smith,C.A. (1995) PoIymorphisms in drug.metabolizing enzymes as modifiers of ~mcer risk. Clin. Chem.. 41. 1864-1869. 28. Oda,~ Tanaka,M. and Nakm'~hi,L (/994) Relation between the occa~rence of K-ras gone point mutations and genotypes of polymorpb2c N- acetyltransferase in human coloreetal carcinomas. Carcinogotes'.ts, 15. 1365-1369. 29.Probst-Hensch,N.M., Haile,R.W., Ingles,S.A. et al. (I995) A.eetylation pelymoxphisra and prevalence of colorectal adenomas. Cancer Res., 55. 2017-2020. 30.Lang,N.P., Butier, M.A.. MassengiIl,L, Lawson,M, Stotts~R.C., Honer- Induction of Cahoot; StT~wtaral Basis attd Biological Mechcmisms. Jensen,M. and Kadluber, EE gI994) Rapid metabolic phenotypes for Aeaderaic Press, New York, N'Y, pp. 681-710 ............ aeetyltransferase and cytocbxome P450IA2 and putative exposure ~o food- 10.Ito, N., Hasegawa,.R., Sano,M.0 Tamano,S., Esumi,H~, Takff~ama,S. and Sugimara,T. (1991) A new eolon and mammaxy earoinogea in cooked food. ~ CarchmgeneMs, 12, I503-1506. ll.Skog.K., SteinecLG., Angnstssonor~L and Jagerstad,M. (1995) Effect of cooking temperatures on the formation ot" heteroq~elic amines in fried meat products and pan residues. Carctno&enesis, 16, 861-867. 12.Gooderham,NJ., Murray, S., Ly~ch,A. et aL (1996) Heteroeyelie amines: evaluation of theh" role in diet associated human cancer. Bn J. Cl~n. PhmvnacoL. -12. 91-98. 13. Kalduchi,H.. War ansbe,M., UsMjima,T., Toyota,M, Imai,I,L. WeisburgerJ.H., Sugimara,T. and Nagao,M. (1995) Spe¢i~c 5'-GGGA-3' to 5'..C, GA-3' mutation of the Ape gone in rat colon tumors induced by 2-amino-l-methyl-6-phenylimldazo[4,5-b]pyridine. Prec. Natl Aca~ Sol. USA, 94, 910-914. 14.PowelLS.M., Zilz,N.. Beaz~'-Barelay, Y., Bryan,TdVl., H.amiltan,$.R., Thibodeau,S.N., Vogelstdn,B. and Kinzler, K.W. (1992) APC mutations occur early daring eoloreetal carcinogenesis. Nature. 3S9, 235-237. IS.Turesky, RJ., Lang24.P., Bufler, M.A., T~itol,C.I-I. and Kadinbar,F.F. (199I) Metabolic activation of eareinoganic heteroeyclic aromatic amines by human liver and colon. Carcinogenesis, 13, I307-1314. I6.MinehinoR.F., Reeves,P.T., Teitel,Cd-L, Memanns,M.E., Mojarrabi,B., lllet,K.E and Kadlubat, EE (1992) N- and O-acetylation of aromatio and heteroeyelte amine carcinogens by haman monomorphie and ~Iymorpttic acetyltransferases expressed in COS-I ceils. Blochem. Biophys. Res. Comnnm., 18~, 839-844. 17. LimD., Meyer, DJ, Ketterer, B., Lang,N.P. and Kadlubar, F.F. (1994) Effects of httman and rat glutati~one S-transferases on the covalent binding of the N-acetoxy derivatives of heteroeyeHc araine careinogans in vitro: a possible meohantsm of organ specificity ia their carcinogenesis. Cancer Res,. ~4, 4920-4926. 18.Vetsisd(.R, ~,hrtetl,K2. and Weber, W.W. (1991) Diverse point mutations tn the ~uman gone fox polymoxphio N-acetyltransferase. Proc. Natl Acad. SoL USA, 88, 6333-6337. lg.BelLD.A.. Badawi,A.F., Lang,N.P.o Ilett.K.F., Kadhtbar~EE and Hirvonan,A. (1995) Polymorphism in the N-acetylwansferase 1 (NAT1) polyadanylation signal: assoaiation of NAT*t0 allele ¢.~.th higl~er N- acetylation activity in bladder and colon tissue. Cancer Res., ~o 59,26- 5229. 20.BelI,D.A., Stepliens,lE.A., Castriano,T., Umbach,D.M., Watson~M.o D~lkindV[., Elder, J,, Hen&ickse,C., Dunean0H. and Strang~R.C. (1995) Polyadenylation polyraorphism in the acetyltransferase i gone (NATI) increases risk of eoloreetai cancer. Cancer Res., 5& 3537-3542. 21.Shibu~a,K.. Nakashlma,T., Abe,M., Mashimo,M.. Mori,M., Ueo,H., Akiyoshi,T., Sugimachi,K. and Sttzuki,T. (I994) Molecular genotyping for NAg polymoxphism in Japanese patients with colore~tal cancer. Cancer, 74. 3108-3112. 22.tlett,K.E, Beverley, M.D., Detehon,P., Cestieden,W.M. and Kwa, R. (1987) borne heterncyclic araines increase the dsk of coiorectal,canc~ ox polyps, Cancer Epldemlol. Biami~r~brs Prevent., 3, 6752-682. 31 .Welfare.M.R., Bassendine,M.l::. and Daly, A.K. (1996) Caffeine p/~anotyping -- in patients with familial adenonmtous polyposis may shed l~ght on sporadic colon cancer and aeetylator slams. Gut. 38, 153. 32.Hiekman,D. and Sim.E. (1990) N-ac~tyl~ransfemse polymo~phism: comparison of phenotype and genot'Yl~ in humans. Biochem. PharmacoL, 42, 1007-1014. 33.Ambrosons,C.B.. Frenderheimd.L., Graham,S" et al. (1996) C~arette smoking, N-acetyltransferase 2 genetic polymmpMsm~ and breast cancer ris'L J. Am. Med. Assoc, 276, 1494--I501. 34.Risch,A., Waiinoe.D.M..g.~ "Bather~,S. and Sim.E. (t995) Slow N- acetylation is a risk factor in occapaflonal and smoking-related btadder cancer. Hton. MoL Goner., 4, 231-236. 35.Giovannucei,E., Coldtiz,O.A.. Stampfer, M.l., I~anter.D.. Rosne'~B.A.. W'dlett,W.C. and Speizer,F.E. (I994) A prospective study of cigarette smoking and risk of colorectal adenoma and cancer in U.S. women. J. Natt Cancer Inst., 86, 192-199. Receh,ed January 16. 1997: revised on March 3, 1997. accepted on March 14. 1997 1354 This article is for individual use only and may not be further reproduced or stored electronically without wdtten permission from the copyright holder. Unauthorized reproduction may result in financial and other penalties. (c) OXFORD UNIV PRESS ENGLAND