Abstract

Cigarette smoking has been inconsistently associated with colon cancer risk. To evaluate the hypothesis that smoking is primarily linked to a specific colon tumor subgroup(s), we assessed associations between smoking and the occurrence of mutations in the APC , K- ras and p53 genes, p53 overexpression, and microsatellite instability (MSI) in a Dutch population-based case–control study on sporadic colon carcinomas. The study population consisted of 176 cases and 249 controls. Smoking status (never, ever), number of cigarettes smoked per day (never, <15, ≥15), total years of smoking (never, ≤30, >30), and years since first started smoking (never, ≤35, >35) were all evaluated. Cigarette smoking status was significantly differently related to p53 overexpression-positive (p53 pos ) tumors compared with p53 overexpression-negative (p53 neg ) tumors (p53 pos versus p53 neg , OR 0.4, 95% CI 0.2–0.9), as well as to tumors with transversion mutations in APC , K- ras or p53 (transv + ) compared with tumors without transversion mutations in one of these genes (transv ) (transv + versus transv , OR 2.5, 95% CI 1.0–5.9). Positive associations were observed with p53 neg tumors and transv + tumors when compared with the population-based controls (ever versus transv , OR 1.5, 95% CI 0.9–2.8 and OR 2.2, 95% CI 0.9–5.6, respectively), inverse associations with p53 pos tumors and transv tumors (ever versus never, OR 0.5, 95% CI 0.3–1.0 and OR 0.8, 95% CI 0.5–1.3, respectively). Similar patterns of association were observed for the other smoking variables evaluated. In addition, although statistically non-significant, smoking was more notably positively associated with tumors that exhibit K- ras mutations, especially K- ras transversion mutations, than with tumors without K- ras mutations. An inverse relationship between smoking and the occurrence of APC mutations was suggested, whereas no clear associations were observed with MSI. Our data suggest that smoking-related colon cancers develop through a p53 neg pathway and that smoking particularly results in colon carcinomas with transversion mutations.

Introduction

Cigarette smoking has been consistently associated with colorectal adenomas, the precursor lesions of most colon carcinomas. Most studies observed a 2- to 3-fold increased risk of adenomas for long-term, heavy cigarette smoking ( 1 ). The association of smoking with colon cancer risk has, however, not been consistent ( 1 ). A possible explanation for this discrepancy is that cigarette smoking is only involved in the production of certain types of mutations in specific genes and, thus, primarily affects, and results in, the development of a particular subset(s) of colon carcinomas.

Mutations in the adenomatous polyposis coli ( APC ), K- ras and p53 genes as well as microsatellite instability (MSI) are commonly observed genetic alterations in colon cancer. Mutations in APC (i.e. those mutations resulting in loss of APC function) are believed to be a key initiating event in colon cancer ( 24 ). K- ras mutations are thought to accompany the conversion of small to larger adenomas and have been reported to occur in 30–40% of sporadic colon tumors ( 5 , 6 ), whereas mutation of p53 seems to be especially important in the later stages of colon tumorigenesis ( 7 ). MSI occurs in most colon tumors associated with the hereditary non-polyposis colorectal cancer syndrome and in ∼10–20% of sporadic colon tumors ( 811 ).

Interestingly, significant inverse relationships have been observed between the presence of MSI and mutations in APC , K -ras and p53 in sporadic colon tumors ( 12 , 13 ). This suggests different molecular pathways to colon cancer which in turn may reflect different environmental exposures. Supporting this idea, Bardelli et al . ( 14 ) recently demonstrated that exposure to specific carcinogens can indeed select for colon tumor cells with distinct forms of genetic alterations. Regarding smoking, K- ras ( 15 , 16 ) and p53 mutations ( 17 , 18 ), G→T transversions in particular, are significantly more prevalent in lung cancers from smokers than in lung cancers from non-smokers, suggesting an etiological link between exposure to tobacco smoke carcinogens and these genetic alterations.

Few studies have examined associations between smoking and genetic alterations in colon cancer. Those that have ( 6 , 1921 ) generally limited their analyses to alterations in one specific gene. Intriguingly, Freedman et al . ( 19 ), who used p53 overexpression as an indicator of p53 mutations, observed an increased risk of p53 overexpression-negative (p53 neg ) colon tumors for smokers. Slattery et al . ( 20 ) and Yang et al . ( 21 ) both reported a positive association between cigarette smoking and sporadic colon tumors with MSI. To (further) explore the hypothesis that cigarette smoking is primarily associated with a specific colon tumor subgroup(s), in this study we assess the associations between smoking and the occurrence of (specific) mutations in the APC , K -ras and p53 genes, p53 overexpression and MSI in sporadic colon carcinomas.

Material and methods

Study population

A population-based case–control study on diet and colon cancer was conducted in The Netherlands between 1989 and 1993. Details were described previously ( 22 , 23 ). Briefly, cases ( n = 204) were women and men newly diagnosed with histologically confirmed, first primary incident colon carcinoma. They were recruited in regional hospitals located in the eastern and central regions of The Netherlands and invited to participate by their medical specialists within 3 months of diagnosis. Cancer registries were used to check for completeness. Of all eligible cases diagnosed in the cooperating hospitals, 47% were invited to participate. Sixty percent of those invited agreed to participate. Controls ( n = 259), frequency matched to the cases by age (5 year intervals), sex, region and degree of urbanization, were randomly recruited by the general practitioners of the cases. Of the controls invited, 57% agreed to participate. All subjects were Dutch speaking, Caucasian, up to 75 years old at time of diagnosis, mentally competent to complete the interview and had no known personal history of cancer, familial adenomatous polyposis, hereditary non-polyposis colorectal cancer, ulcerative colitis or Crohn’s disease. Except for a more favorable Dukes’ stage among cases, participants did not differ importantly from non-participants. Cigar and/or pipe only smokers ( n = 18) and two subjects with missing data on smoking status were excluded from the analyses. Colon tumor tissue could not be obtained from 18 cases due to administrative reasons. In total, 176 cases and 249 controls were included in the analyses here presented.

Data collection

Participants were interviewed in their own homes by trained dieticians using a structured questionnaire. The interval between diagnosis and the interview was, for cases, 3–6 months. Cigarette smoking status (never, ever, ex, current) was determined. Participants who had stopped smoking 1 year or more prior to the date of the interview were classified as ex-smokers; participants who had stopped <1 year prior to the interview date or were still smoking were classified as current smokers. Information about the number of years smoked and the number of cigarettes usually smoked per day (categorized in four categories: 1–<5, 5–<15, 15–<25 and ≥25 cigarettes) was obtained. Years since first started smoking was calculated from information on duration of smoking and, if applicable, time since stopped smoking. The interview also consisted of a dietary history part in which information on the frequency and amounts of foods consumed in the year prior to the interview (for cases, the year preceding diagnosis or complaints) was collected. Information on aspirin and non-steroidal anti-inflammatory drug use, family history of colorectal cancer and medical history was also obtained during the interview.

DNA extraction

Both tumor and normal DNA were extracted from formalin-fixed, paraffin-embedded colon tissue, collected before chemo- or radiotherapy started, as described elsewhere ( 24 ). Microdissection was performed and for tumor DNA only those areas containing >60% tumor cells were used. Corresponding normal DNA was isolated from tumor-free colon tissue.

APC mutation detection

Single-strand conformation polymorphism (SSCP) analysis was used to screen the APC gene for mutations. The majority of the somatic mutations in APC seem to cluster within a small region in exon 15 (codons 1286–1513), the so-called mutation cluster region (MCR) ( 2527 ). Our analysis covered codons 1286–1585 (extended MCR) of APC . The region was divided into five ∼220 bp long overlapping fragments (codons 1286–1358, 1337–1404, 1387–1455, 1437–1526 and 1509–1585, respectively), which were separately amplified in two consecutive PCRs using the following primer sets (primer sequence 5′→3′). Fragment 1: 1.1 forward CAGACTTATTGTGTAGAAG, reverse CGCTCCTGAAGAAAATTCAAG (codons 1260–1358); 1.2 forward GAAATAGGATGTAATCAGACG, reverse CGCTCCTGAAGAAAATTCAAC (codons 1286–1358). Fragment 2: 2.1 forward ACTGCAGGGTTCTAGTTTATC, reverse TCTGCTTGGTGGCATGGTTT (codons 1337–1436); 2.2 forward ACTGCAGGGTTCTAGTTTATC, reverse GAGCTGGCAATCGAACGACT (codons 1337–1404). Fragment 3: 3.1 forward CTCAGACACCCAAAAGTCC, reverse ATTTTTAGGTACTTCTCGCTTG (codons 1366–1455); 3.2 forward TACTTCTGTCAGTTCACTTGATA, reverse ATTTTTAGGTACTTCTCGCTTG (codons 1387–1455). Fragment 4: 4.1 forward AAACACCTCCACCACCTCC, reverse TCATTCCCATTGTCATTTTCC (codons 1437–1536); 4.2 forward AAACACCTCCACCACCTCC, reverse GCATTATTCTTAATTCCACATC (codons 1437–1526). Fragment 5: 5.1 forward ACTCCAGATGGATTTTTCTTG, reverse GGCTGGCTTTTTGCTTTAC (codons 1497–1596); 5.2 forward GAGCCTCGATGAGCCATTTA, reverse TGTTGGCATGGCAGAAATAA (codons 1509–1585).

PCR reaction mixtures (total volume 50 μl) contained 50 ng DNA (or 2 μl of the 1:100 diluted product of the first PCR), 0.2 μM both primers, 0.2 mM dNTPs, 10 mM Tris–HCl, pH 9.0, 1.5–2.5 mM MgCl 2 , 50 mM KCl, 0.01% Tween, 10% glycerol and 0.3 U Taq DNA polymerase. Reaction conditions were: first PCR, 25 cycles of 30 s at 94°C, 45 s at 55°C (53°C for primer set 1.1; 57°C for 4.1), 1 min at 72°C, followed by 5 min at 72°C; second PCR, 30 cycles of 30 s at 94°C, 45 s at 52°C (56°C for primer sets 3.2 and 4.2; 57°C for 5.2), 1 min at 72°C, followed by 5 min at 72°C. Products were checked using an ethidium bromide stained 2% agarose gel. SSCP was performed as described earlier ( 24 ) with electrophoresis at 10 and 18°C. The original PCR products from the samples that displayed an abnormal pattern in the SSCP were subjected to sequencing in both directions using the same primers as in the second PCR. Sequencing was performed as described previously ( 24 ). Mutation analysis started in all samples with fragment 1 and only if no truncating mutations (i.e. nonsense or frameshift mutations) were detected was fragment 2 screened for mutations, and so on. We focused on truncating mutations because these mutations indisputably result in function loss of the APC protein whereas the biological significance of missense mutations in APC is uncertain. Carcinomas were classified as APC + (with a truncating mutation in the extended MCR of APC ) or APC (without a truncating mutation in the extended MCR of APC and all five fragments completely analyzed for mutations).

K-ras mutation detection

Codons 12 and 13 of the K -ras gene were examined for mutations (i.e. those resulting in an amino acid change) by mutant allele-specific amplification as described earlier ( 28 ).

p53 mutation detection

To enable the evaluation of (specific) p53 mutations, SSCP analysis was used to screen exons 5–8 of the p53 gene for mutations as described previously ( 24 ). We focused on exons 5–8 as it has been observed that most p53 mutations occur in this region of the gene ( 29 ). The original PCR products from the samples that displayed an abnormal pattern in the SSCP were subjected to sequencing in both directions. Samples in which a mutation (i.e. one that resulted in an amino acid change or truncation of the protein) was detected were excluded from analysis of the subsequent exons. The exons were screened in the following order: 7, 8, 5 and 6. Carcinomas were classified as p53 + (with a mutation in codons 5–8 of p53 ) or p53 (without a mutation in codons 5–8 of p53 and all four codons completely analyzed for mutations).

p53 immunohistochemistry

Overexpression of p53 was determined using a mixture of two antibodies (DO-7, which recognizes both mutant and wild-type forms of p53, and PAb 240, which recognizes only mutant forms of p53) as published earlier ( 24 ). Stained sections were scored independently by two investigators (A.A.van Kraats and G.N.P.van Muijen). Tumors were scored as p53 neg if <20% of the cells displayed nuclear positivity and as p53 overexpression-positive (p53 pos ) if otherwise, as in Freedman et al . ( 19 ).

Microsatellite instability

For MSI analysis, paired tumor and normal DNA were investigated with the five Bethesda reference panel markers ( 30 ): BAT25 , BAT26 , D5S346 , D2S123 and D17S250 . PCR reaction mixtures (total volume 25 μl) contained 100 ng DNA, 10 pmol forward (fluorescent labeled) and reverse primers, 2.0 mM MgCl 2 (2.5 mM MgCl 2 for BAT26 ), 0.2 mM dNTPs, 75 mM Tris–HCl, pH 9.0, 20 mM (NH 4 ) 2 SO 4 , 0.01% Tween, 0.3 U Thermoperfect DNA polymerase (Integro). PCR reaction conditions were: 35 cycles of 30 s at 92°C, 45 s at 50°C, 1 min at 72°C, followed by 30 min at 72°C. Products were checked using an ethidium bromide stained 2% agarose gel. An aliquot of 1 μl of (diluted) PCR product was added to 10 μl of formamide and 0.5 μl of ROX-500 (size standard), denatured at 95°C for 5 min, chilled on ice and loaded on an ABI PRISM ® 3100 Genetic Analyzer (Applied Biosystems). Genotyper ® ABI PRISM version 3.5 NT (Perkin Elmer) was used to analyze the data. MSI at a specific marker was defined as the presence of a novel length allele in tumor tissue when compared with corresponding normal tissue. When matching normal DNA was not available ( n = 22), only the mononucleotide repeat markers BAT25 and BAT26 were checked for instability. Tumors were classified as MSI-H if two or more markers showed instability, MSI-L if one marker showed instability and MSS if none of the markers examined showed instability ( 30 ).

Data analysis

The distribution of APC , K- ras and p53 mutations, p53 overexpression and MSI in the colon carcinomas was determined. Differences in (tumor) characteristics between ‘never smoked’ and ‘ever smoked’ cases and between MSI-H and MSS tumors were assessed using t -tests for continuous and χ 2 tests for categorical variables; P values <0.05 were considered significant. Case–case comparisons were conducted to evaluate heterogeneity in risk factors for the different tumor subsets. In addition, case–control comparisons, separately comparing cases with and cases without specific alterations with the population-based controls, were conducted to estimate the relative risk of developing carcinomas respectively with and without this particular status. The risk factors evaluated were: cigarette smoking status (never, ever), number of cigarettes usually smoked per day (never, <15, ≥15), total years of smoking (never, ≤30, >30) and years since first started smoking (never, ≤35, >35). Odds ratios (ORs) and the corresponding 95% confidence intervals (95% CIs) were calculated using multiple logistic regression models. ‘Never’ was always the referent category. All analyses were adjusted for age, sex, total energy and alcohol intake. Additional adjustment for Dukes’ stage, tumor location, body mass index and the consumption of vegetables, fruit and red meat did not change the estimates importantly (i.e. not more than 10%). All analyses were performed with the use of the SAS ® statistical software package (SAS v.6.12; SAS Institute, Cary, NC).

Results

Characteristics of the cases in this study population, categorized by cigarette smoking status, are given in Table I . There were significantly more women among the never smokers. Total energy and alcohol intake were higher among the ever smokers whereas age and body mass index did not differ between the two groups. Current smokers smoked more cigarettes per day and had smoked for more years than ex-smokers (data not shown).

Tumor characteristics, categorized by cigarette smoking status of the cases, are described in Table II . Although Dukes’ stage CD and proximal tumors (cecum, ascending colon, hepatic flexure and transverse colon) were both more common among the never smokers, the difference from ever smokers was not statistically significant. There were significantly more carcinomas with truncating APC mutations among the never smokers. p53 overexpression was observed at a higher frequency in the carcinomas of never smokers, whereas K- ras mutations, and especially codon 12 transversion mutations, were more common in the carcinomas of ever smokers than of never smokers (both not statistically significant). Transversion mutations, G→T in particular, were, overall, more common among ever smokers. Tumor MSI status did not differ between never and ever smokers. Regarding MSI analysis of the tumors for which no matching normal DNA was available ( n = 22), in 10 of these tumors both BAT25 and BAT26 were unstable, whereas in the other 12 tumors both markers were stable. Fifty-five percent of all p53 pos tumors exhibited a p53 mutation. Regarding the genetic alterations examined, frequencies observed among women were comparable with those observed among men (data not shown).

MSI-H tumors exhibited a significantly lower frequency of APC , K- ras and p53 mutations and were less often p53 pos than MSS tumors (MSI-H tumors, 21% APC , 21% K- ras , 8% p53 , 15% p53 pos ; MSS tumors, 38% APC , 41% K- ras , 36% p53 , 51% p53 pos ; not in Table II ). They were also significantly more often located in the proximal part of the colon. In most (54%) MSI-H tumors none of the examined genes were mutated and p53 was not overexpressed (data not shown).

Ever smoking was not associated with increased overall colon cancer risk in this study population (Table III ). Table III also shows the adjusted ORs and 95% CIs of the case–case and case–control comparisons for cigarette smoking status and the occurrence of the various genetic alterations examined in this study. No significant associations were observed between cigarette smoking status and tumor APC mutation status. Of the other cigarette smoking variables evaluated (i.e. usual number of cigarettes smoked per day, total years of smoking and years since first started), only first starting smoking ≤35 years ago was significantly differently associated with APC + tumors compared to APC tumors (APC + versus APC , OR 0.2, 95% CI 0.1–0.7; not in Table III ). In the case–control comparisons, first starting smoking ≤35 years ago was found to be significantly inversely associated with APC + tumors only (APC + versus controls, OR 0.2, 95% CI 0.1–0.8; APC versus controls, OR 1.2, 95% CI 0.6–2.3; not in Table III ). No associations were observed between first starting smoking >35 years ago and tumor APC mutation status.

Ever smoking, though not statistically significantly, seems more notably positively associated with tumors that exhibit K- ras mutations, especially K- ras transversion mutations, than with tumors without K- ras mutations (Table III ). The other smoking variables evaluated were, although again not statistically significantly, also positively associated with tumors with K- ras mutations, more pronounced with K- ras transversion mutations, and negatively with tumors without K- ras mutations (data not shown).

Ever smoking was significantly differently associated with p53 pos tumors compared to p53 neg tumors (Table III ). The case–control comparisons showed that ever smoking was inversely associated with p53 pos tumors and positively, although not statistically significant, associated with p53 neg tumors. Evaluation of the other cigarette smoking variables provided additional support that, in this population, cigarette smoking is inversely associated with p53 pos tumors. Smoking ≥15 cigarettes/day, smoking >30 years and first starting smoking >35 years ago were significantly differently associated with p53 pos tumors compared to p53 neg tumors (p53 pos versus p53 neg , OR 0.4, 95% CI 0.2–1.0, OR 0.3, 95% CI 0.1–0.7 and OR 0.4, 95% CI 0.2–1.0, respectively; not in Table III ). Additionally, all were inversely associated with p53 pos tumors and positively with p53 neg tumors when compared to the population-based controls (p53 pos versus controls, OR 0.5, 95% CI 0.2–1.0, OR 0.4, 95% CI 0.2–0.9 and OR 0.6, 95% CI 0.3–1.3, respectively; p53 neg versus controls, OR, 1.4, 95% CI 0.7–2.8, OR 1.6, 95% CI 0.8–3.1 and OR 1.7, 95% CI 0.9–3.2, respectively; not in Table III ). Interestingly, no clear associations were observed between cigarette smoking and tumor p53 mutation status.

To evaluate whether cigarette smoking was specifically associated with the presence of transversion mutations, we additionally assessed the associations between smoking and carcinomas with transversion mutations in APC , K- ras or p53 (transv + tumors). The majority of the transv + tumors harbored, at least, a transversion mutation in K- ras . Ever smoking was significantly differently associated with tumors with transversion mutations compared to tumors without transversion mutations (Table III ). The case–control comparisons showed that ever smoking was positively, although not statistically significant, associated with tumors with transversion mutations and negatively, again not statistically significantly, with tumors without transversion mutations. Similar patterns were observed for the other smoking variables evaluated (data not shown). No clear associations were observed between smoking and tumors with transition mutations in APC , K- ras or p53 (Table III ).

Regarding MSI, no clear associations were observed between the smoking variables examined and MSI status (Table III ). In addition, no clear associations were observed when the analyses were repeated with the subset MSI-L/MSS instead of MSS (data not shown).

Discussion

In this study, associations between cigarette smoking and various genetic alterations in sporadic colon carcinomas were assessed to evaluate the hypothesis that cigarette smoking is primarily linked to a specific colon tumor subset(s). Cigarette smoking was significantly differently related to p53 pos tumors than to p53 neg tumors. Consistent inverse associations were observed with p53 pos tumors while positive associations were found with p53 neg tumors. Smoking was also significantly differently related to tumors with transversion mutations in APC , K- ras or p53 than to tumors without transversion mutations in one of these genes. Positive associations were observed with tumors with transversion mutations but not with tumors without transversion mutations. In addition, smoking seemed to especially increase the risk of developing tumors with K- ras mutations, in particular K- ras transversion mutations. No clear associations were observed with MSI status.

Truncating APC mutations were identified in 34.1%, K- ras mutations in 36.4%, p53 mutations in 30.7% and p53 overexpression in 44.3% of the carcinomas in this study. Twenty-two percent of the carcinomas were MSI-H and the occurrence of MSI was significantly inversely related to the presence of genetic alterations in the APC , K- ras and p53 genes and to p53 overexpression. Microdissection was performed and the observed frequencies and characteristics of the mutations identified were consistent with those in comparable populations previously reported by others ( APC database, http://perso.curie.fr/Thierry.Soussi/APC/html ; IARC p53 database, http://www.iarc.fr/p53/index.html ; 4–6,13,19,20,25,26). However, it remains possible that, due to contaminating normal DNA, alterations were missed, eventually resulting in misclassifications, which in turn may have attenuated some of our results.

In this study the MSI status of 22 tumors was determined using BAT25 and BAT26 tumor results only as for these tumors matching normal DNA was not available. The markers were either both unstable or both stable. Recently, polymorphisms have been identified in BAT25 as well as in BAT26 , disputing the earlier suggested quasimonomorphic allelic profile of these two loci and warranting caution in the interpretation of MSI data based on BAT25 and BAT26 tumor results only ( 31 , 32 ). However, the polymorphisms appear to be population dependent and to occur significantly more frequently in African-Americans than in Caucasians ( 31 , 32 ). In the latter, they truly seem to be uncommon, and being polymorphic at both loci will most likely be even more uncommon. Therefore, we do not expect that our decision to use BAT25 and BAT26 tumor results only, when matching normal DNA was not available, has resulted in extensive misclassification or led to serious misinterpretation of our data.

As in any retrospective study, an important concern is the possibility of information and selection bias. The smoking habits of our controls were comparable to those of the general Dutch population at the time of interview ( 33 ). Since the cases were unaware of the molecular profile of their tumors, systematic errors in recall are less likely to bias results from case–case comparisons. Recall of (smoking) habits, however, can also be influenced by tumor stage or treatments. Our cases were relatively healthy, i.e. the frequency of Dukes’ A and B tumors among the cases was relatively high, 63%, compared with the 51% reported by the Dutch Cancer Registry ( 34 ). Adjusting the case–case comparisons for Dukes’ stage did not change the estimates significantly. A long time lag between smoking exposure and occurrence of colon carcinomas has previously been suggested as a possible explanation for smoking being a risk factor for adenomas but not for colon cancer ( 35 , 36 ). In our study population 72% of the ever smokers among the cases first started smoking ≥35 years ago.

This is, to our knowledge, the first study that has evaluated associations between cigarette smoking and alterations in the APC gene in sporadic colon carcinomas. If anything, our data suggest a slight inverse association between the two; it seems that most sporadic colon cancers related to smoking are not initiated via alterations in the APC gene. This is not entirely unexpected considering that the frequency of APC mutations observed in sporadic adenomas is similar to that in sporadic carcinomas ( 4 ) and that APC appears to also play a role in the later stages of tumor development ( 37 , 38 ).

A few studies have previously looked at cigarette smoking and other genetic alterations in colon tumors. In a large population-based case–control study on sporadic colon cancer, Slattery et al . ( 6 ) observed a slight increased risk of K-ras tumors when smoking ≥20 cigarettes/day (K-ras versus controls, OR 1.3, 95% CI 1.1–1.6). However, the risk of K-ras + tumors increased as well (K-ras + versus controls, OR 1.2, 95% CI 0.9–1.5). Additionally, smoking >20 cigarettes/day was associated with an increased risk of overall colon cancer in their study population ( 39 ). In line with our results for K- ras , Martinez et al . ( 40 ) reported a slight positive, but non-significant, association with K- ras mutations in 0.5 cm or larger sporadic colorectal adenomas for current versus never smokers.

In the study population of Slattery et al ., cigarette smoking was found to be positively associated with microsatellite unstable tumors ( 20 ). Yang et al . ( 21 ), who purposely enriched their study for MSI-H cases, also reported a positive association between cigarette smoking and MSI-H tumors. We observed no clear associations between smoking and MSI-H tumors and, hence, did not confirm their results. Both Slattery et al . and Yang et al . did not use the Bethesda reference panel ( 30 ) to assess MSI, which may explain the difference in results. Additionally, it is possible that we observed no associations due to the size of our study population and/or because the frequency of heavy smokers among our smokers was lower. Similarly to our results, Slattery et al . reported an inverse relation between K- ras mutations and MSI ( 13 , 20 ).

We observed no clear associations with p53 mutations but our results for p53 overexpression suggest that most colon cancers related to cigarette smoking develop through a p53 neg pathway (which is not necessarily also p53 mutation-negative; see below). Consistent with our findings for p53 overexpression, Freedman et al . ( 19 ) observed an increased risk of p53 neg colon tumors for current and former smokers. They did not evaluate p53 mutations. Although p53 overexpression is often used as an indicator of p53 mutations, not all p53 mutations (e.g. nonsense and frameshift mutations) result in the accumulation of inactive p53 protein and can be detected by immunohistochemical analysis ( 41 ). In our population for instance, 20% of the p53 mutations resulted in p53 neg tumors. Under normal circumstances, wild-type p53 is activated and accumulates in response to DNA damage and various other types of stress, resulting in either growth arrest or apoptosis ( 42 ). Cancer cells need to somehow circumvent this checkpoint to be able to proliferate. A possible explanation for the observed association with a p53 neg pathway is that cigarette smoking can somehow (e.g. by preventing post-translational modifications to occur) suppress the induction or stabilization of wild-type p53 (resulting in p53 neg cells), allowing cells to proliferate in conditions where cells with intact p53 function are eliminated.

Interestingly, we observed a positive association between cigarette smoking and tumors with transversion mutations in APC , K- ras or p53 . This suggests that cigarette smoking is particularly involved in the production of transversion mutations in colon cells. However, it should be noted that genetic alterations in tumors not only represent the interactions of carcinogens with DNA repair processes but also reflect the, possibly tissue-specific, selection of those mutations that provide pre-malignant and malignant cells with a clonal growth advantage. Consistent with our results, transversion mutations in p53 and K- ras are also commonly found in smoking-related lung cancers ( 1518 ). Additionally, Conway et al . ( 43 ) reported recently that p53 transversion mutations, and especially G→T transversions, were significantly more prevalent in breast tumors from current smokers than in breast tumors from never smokers.

To conclude, our data suggest that smoking-related colon cancers develop through a p53 neg pathway and that cigarette smoking particularly results in colon tumor cells with transversion mutations. Regarding the latter, it appears that cigarette smoking especially results in colon tumors with K- ras transversion mutations. This may be due to hypersensitivity of codons 12 and 13 of K- ras for exposure to tobacco smoke carcinogens or to a higher selective advantage for colon tumor formation exerted by these mutations in K- ras than in one of the other genes examined. Our results, if confirmed in other studies, provide support for the hypothesis that cigarette smoking is primarily associated with specific colon tumor subgroups.

Table I.

Characteristics of the colon cancer cases by cigarette smoking status

  All cases ( n = 176)   Never smokers ( n = 55)   Ever smokers ( n = 121)  
a Never smokers versus ever smokers P < 0.05, t -test for continuous variables and χ 2 test for categorical variables.  
b Adjusted for total energy intake by regression analysis, for women and men separately.  
Age (years, mean ± SD)  61.7 ± 10.3  62.6 ± 9.9  61.3 ± 10.2 
Sex (% women)  47.2  69.1  37.2 a 
Body mass index (kg/m 2 , mean ± SD)   25.9 ± 4.4  25.0 ± 4.4  26.3 ± 4.4 
Smoking variables (%)    
    Current smokers    26.4 
    ≥15 cigarettes/day    51.2 
    ≥25 cigarettes/day    32.3 
    >30 years of smoking    49.2 
    >35 years since first smoked    72.2 
Dietary factors (mean ± SD)    
    Total energy intake (kJ/day) 10373.7 ± 3241.6 9491.4 ± 2767.7  10774.8 ± 3370.2 a 
    Vegetables (g/day)  170.5 ± 81.7  153.8 ± 74.9  178.5 ± 83.7 
    Fruit (g/day)  213.5 ± 155.7  242.6 ± 166.8  200.2 ± 149.3 
    Red meat (g/day)  79.7 ± 35.4  73.9 ± 34.9  82.3 ± 35.5 
    Alcohol b (g/day)   17.1 ± 25.4  9.8 ± 17.6  20.4 ± 27.6 a 
  All cases ( n = 176)   Never smokers ( n = 55)   Ever smokers ( n = 121)  
a Never smokers versus ever smokers P < 0.05, t -test for continuous variables and χ 2 test for categorical variables.  
b Adjusted for total energy intake by regression analysis, for women and men separately.  
Age (years, mean ± SD)  61.7 ± 10.3  62.6 ± 9.9  61.3 ± 10.2 
Sex (% women)  47.2  69.1  37.2 a 
Body mass index (kg/m 2 , mean ± SD)   25.9 ± 4.4  25.0 ± 4.4  26.3 ± 4.4 
Smoking variables (%)    
    Current smokers    26.4 
    ≥15 cigarettes/day    51.2 
    ≥25 cigarettes/day    32.3 
    >30 years of smoking    49.2 
    >35 years since first smoked    72.2 
Dietary factors (mean ± SD)    
    Total energy intake (kJ/day) 10373.7 ± 3241.6 9491.4 ± 2767.7  10774.8 ± 3370.2 a 
    Vegetables (g/day)  170.5 ± 81.7  153.8 ± 74.9  178.5 ± 83.7 
    Fruit (g/day)  213.5 ± 155.7  242.6 ± 166.8  200.2 ± 149.3 
    Red meat (g/day)  79.7 ± 35.4  73.9 ± 34.9  82.3 ± 35.5 
    Alcohol b (g/day)   17.1 ± 25.4  9.8 ± 17.6  20.4 ± 27.6 a 
Table II.

Tumor characteristics, categorized by cigarette smoking status of the cases

  All cases ( n = 176)   Never smokers ( n = 55)   Ever smokers ( n = 121)  
a χ 2 test, never smokers versus ever smokers, P < 0.05.  
b One tumor exhibited a missense mutation in codon 12 as well as in codon 13.  
c MSI-H, ≥2 markers unstable; MSI-L, 1 marker unstable; MSS, 0 markers unstable.  
Dukes’ stage (% CD)  36.8 44.4 33.6 
Tumor location (% proximal)  45.7 54.2 42.2 
Mutation distribution [ n (%)]     
APC    
    Tumors with mutation  60 (34.1) 25 (45.5)  35 (28.9) a 
        Transversions (see p53 )   8 (13.3) [4× G→T]  1 (4.0) [0× G→T]  7 (20.0) 
        Transitions  9 (15.0) [all C→T]  3 (12.0)  6 (17.1) 
        Insertions/deletions  43 (71.7) 21 (84.0) 22 (62.9) 
K- ras    
    Tumors with mutation(s) b  64 (36.4) 16 (29.1) 48 (39.7) 
        Codon 12 mutations  53 (81.5) 13 (81.3) 40 (81.6) 
            Transversions  30 (56.6) [21× G→T]  6 (46.2) [5× G→T] 24 (60.0) 
            Transitions  23 (43.4) [all G→A]  7 (53.8) 16 (40.0) 
        Codon 13 mutations  12 (18.5)  3 (18.7)  9 (18.4) 
            Transversions  1 (8.3) [G→T]  0 (0)  1 (11.1) 
            Transitions  11 (91.7) [all G→A]  3 (100)  8 (88.9) 
p53    
    Overexpression-positive  78 (44.3) 29 (52.7) 49 (40.5) 
    Tumors with mutation  54 (30.7) 16 (29.1) 38 (31.4) 
        Transversions  13 (24.1) [5× G→T]  2 (12.5) [0× G→T] 11 (28.9) 
        Transitions  35 (64.8) [20× G→A] 13 (81.3) [7× G→A] 22 (57.9) 
        Insertions/deletions  6 (11.1)  1 (6.3)  5 (13.2) 
MSI c    
    MSI-H  39 (22.2) 13 (23.6) 26 (21.5) 
    MSI-L  20 (11.4)  7 (12.7) 13 (10.7) 
        MSS 117 (66.5) 35 (63.6) 82 (67.8) 
  All cases ( n = 176)   Never smokers ( n = 55)   Ever smokers ( n = 121)  
a χ 2 test, never smokers versus ever smokers, P < 0.05.  
b One tumor exhibited a missense mutation in codon 12 as well as in codon 13.  
c MSI-H, ≥2 markers unstable; MSI-L, 1 marker unstable; MSS, 0 markers unstable.  
Dukes’ stage (% CD)  36.8 44.4 33.6 
Tumor location (% proximal)  45.7 54.2 42.2 
Mutation distribution [ n (%)]     
APC    
    Tumors with mutation  60 (34.1) 25 (45.5)  35 (28.9) a 
        Transversions (see p53 )   8 (13.3) [4× G→T]  1 (4.0) [0× G→T]  7 (20.0) 
        Transitions  9 (15.0) [all C→T]  3 (12.0)  6 (17.1) 
        Insertions/deletions  43 (71.7) 21 (84.0) 22 (62.9) 
K- ras    
    Tumors with mutation(s) b  64 (36.4) 16 (29.1) 48 (39.7) 
        Codon 12 mutations  53 (81.5) 13 (81.3) 40 (81.6) 
            Transversions  30 (56.6) [21× G→T]  6 (46.2) [5× G→T] 24 (60.0) 
            Transitions  23 (43.4) [all G→A]  7 (53.8) 16 (40.0) 
        Codon 13 mutations  12 (18.5)  3 (18.7)  9 (18.4) 
            Transversions  1 (8.3) [G→T]  0 (0)  1 (11.1) 
            Transitions  11 (91.7) [all G→A]  3 (100)  8 (88.9) 
p53    
    Overexpression-positive  78 (44.3) 29 (52.7) 49 (40.5) 
    Tumors with mutation  54 (30.7) 16 (29.1) 38 (31.4) 
        Transversions  13 (24.1) [5× G→T]  2 (12.5) [0× G→T] 11 (28.9) 
        Transitions  35 (64.8) [20× G→A] 13 (81.3) [7× G→A] 22 (57.9) 
        Insertions/deletions  6 (11.1)  1 (6.3)  5 (13.2) 
MSI c    
    MSI-H  39 (22.2) 13 (23.6) 26 (21.5) 
    MSI-L  20 (11.4)  7 (12.7) 13 (10.7) 
        MSS 117 (66.5) 35 (63.6) 82 (67.8) 
Table III.

Cigarette smoking status and mutations in sporadic colon carcinomas: case–case and case–control comparisons

  Odds ratios (95% confidence intervals) a 
 Never smoked Ever smoked 
a Adjusted for age, sex, total energy and alcohol intake.  
b APC + , tumors with truncating APC mutation; APC , tumors without truncating APC mutation.  
c K-ras + , tumors with mutation in K- ras ; K-ras , tumors without mutation in K- ras ; K-ras transv , tumors with transversion mutation in K- ras ; K-ras trans , tumors with transition mutation in K- ras .  
d p53 + , tumors with mutation in p53 ; p53 , tumors without mutation in p53 .  
e transv + , tumors with transversion mutations in APC , K- ras or p53 ; transv , tumors without transversion mutations in one of these genes.  
f trans + , tumors with transition mutations in APC , K- ras or p53 ; trans , tumors without transition mutations in one of these genes.  
g p53 pos , tumors in which ≥20% of the cells displayed nuclear positivity; p53 neg , tumors in which <20% of the cells displayed nuclear positivity.  
h MSI-H, tumors in which two or more markers showed instability; MSS, tumors with no unstable markers.  
Overall   
No. cases/controls 55/79 121/170 
    Cases versus controls 1.0 1.0 (0.6–1.5) 
APCb   
    No. APC + /APC /controls  25/30/79 35/86/170 
        APC + versus APC  1.0 0.6 (0.3–1.2) 
        APC + versus controls  1.0 0.7 (0.4–1.4) 
        APC versus controls  1.0 1.2 (0.7–2.1) 
K- rasc   
    No. K-ras + /K-ras /controls  16/39/79 48/73/170 
        K-ras + versus K-ras  1.0 1.7 (0.8–3.4) 
        K-ras + versus controls  1.0 1.4 (0.7–2.8) 
        K-ras versus controls  1.0 0.8 (0.5–1.4) 
    No. K-ras transv /K-ras trans /K-ras /controls  6/10/39/79 25/23/73/170 
        K-ras transv versus K-ras  1.0 2.2 (0.8–6.2) 
        K-ras transv versus controls  1.0 2.2 (0.7–6.3) 
        K-ras trans versus K-ras  1.0 1.4 (0.6–3.4) 
        K-ras trans versus controls  1.0 1.0 (0.4–2.5) 
p53 (mutations) d   
    No. p53 + /p53 /controls  16/39/79 38/83/170 
        p53 + versus p53  1.0 1.1 (0.5–2.2) 
        p53 + versus controls  1.0 0.9 (0.4–1.9) 
        p53 versus controls  1.0 1.0 (0.6–1.7) 
Transversion mutations e   
    No. transv + /transv /controls  8/47/79 38/83/170 
        transv + versus transv  1.0 2.5 (1.0–5.9) 
        transv + versus controls  1.0 2.2 (0.9–5.6) 
        transv versus controls  1.0 0.8 (0.5–1.3) 
Transition mutations f   
    No. trans + /trans /controls  21/34/79 46/75/170 
        trans + versus trans  1.0 1.0 (0.5–2.0) 
        trans + versus controls  1.0 0.9 (0.5–1.7) 
        trans versus controls  1.0 1.0 (0.6–1.8) 
p53 (overexpression) g   
    No. p53 pos /p53 neg /controls  29/26/79 49/72/170 
        p53 pos versus p53 neg 1.0 0.4 (0.2–0.9) 
        p53 pos versus controls  1.0 0.5 (0.3–1.0) 
        p53 neg versus controls  1.0 1.5 (0.9–2.8) 
MSI h   
    No. MSI-H/MSS/controls 13/35/79 26/82/170 
        MSI-H versus MSS 1.0 0.8 (0.3–1.7) 
        MSI-H versus controls 1.0 0.8 (0.3–1.8) 
        MSS versus controls 1.0 1.1 (0.6–1.8) 
  Odds ratios (95% confidence intervals) a 
 Never smoked Ever smoked 
a Adjusted for age, sex, total energy and alcohol intake.  
b APC + , tumors with truncating APC mutation; APC , tumors without truncating APC mutation.  
c K-ras + , tumors with mutation in K- ras ; K-ras , tumors without mutation in K- ras ; K-ras transv , tumors with transversion mutation in K- ras ; K-ras trans , tumors with transition mutation in K- ras .  
d p53 + , tumors with mutation in p53 ; p53 , tumors without mutation in p53 .  
e transv + , tumors with transversion mutations in APC , K- ras or p53 ; transv , tumors without transversion mutations in one of these genes.  
f trans + , tumors with transition mutations in APC , K- ras or p53 ; trans , tumors without transition mutations in one of these genes.  
g p53 pos , tumors in which ≥20% of the cells displayed nuclear positivity; p53 neg , tumors in which <20% of the cells displayed nuclear positivity.  
h MSI-H, tumors in which two or more markers showed instability; MSS, tumors with no unstable markers.  
Overall   
No. cases/controls 55/79 121/170 
    Cases versus controls 1.0 1.0 (0.6–1.5) 
APCb   
    No. APC + /APC /controls  25/30/79 35/86/170 
        APC + versus APC  1.0 0.6 (0.3–1.2) 
        APC + versus controls  1.0 0.7 (0.4–1.4) 
        APC versus controls  1.0 1.2 (0.7–2.1) 
K- rasc   
    No. K-ras + /K-ras /controls  16/39/79 48/73/170 
        K-ras + versus K-ras  1.0 1.7 (0.8–3.4) 
        K-ras + versus controls  1.0 1.4 (0.7–2.8) 
        K-ras versus controls  1.0 0.8 (0.5–1.4) 
    No. K-ras transv /K-ras trans /K-ras /controls  6/10/39/79 25/23/73/170 
        K-ras transv versus K-ras  1.0 2.2 (0.8–6.2) 
        K-ras transv versus controls  1.0 2.2 (0.7–6.3) 
        K-ras trans versus K-ras  1.0 1.4 (0.6–3.4) 
        K-ras trans versus controls  1.0 1.0 (0.4–2.5) 
p53 (mutations) d   
    No. p53 + /p53 /controls  16/39/79 38/83/170 
        p53 + versus p53  1.0 1.1 (0.5–2.2) 
        p53 + versus controls  1.0 0.9 (0.4–1.9) 
        p53 versus controls  1.0 1.0 (0.6–1.7) 
Transversion mutations e   
    No. transv + /transv /controls  8/47/79 38/83/170 
        transv + versus transv  1.0 2.5 (1.0–5.9) 
        transv + versus controls  1.0 2.2 (0.9–5.6) 
        transv versus controls  1.0 0.8 (0.5–1.3) 
Transition mutations f   
    No. trans + /trans /controls  21/34/79 46/75/170 
        trans + versus trans  1.0 1.0 (0.5–2.0) 
        trans + versus controls  1.0 0.9 (0.5–1.7) 
        trans versus controls  1.0 1.0 (0.6–1.8) 
p53 (overexpression) g   
    No. p53 pos /p53 neg /controls  29/26/79 49/72/170 
        p53 pos versus p53 neg 1.0 0.4 (0.2–0.9) 
        p53 pos versus controls  1.0 0.5 (0.3–1.0) 
        p53 neg versus controls  1.0 1.5 (0.9–2.8) 
MSI h   
    No. MSI-H/MSS/controls 13/35/79 26/82/170 
        MSI-H versus MSS 1.0 0.8 (0.3–1.7) 
        MSI-H versus controls 1.0 0.8 (0.3–1.8) 
        MSS versus controls 1.0 1.1 (0.6–1.8) 
2
To whom correspondence should be addressed Email: ellen.kampman@wur.nl

We thank Winny van Geloof and Hanneke Braam for excellent technical assistance and the Pathology Departments of the following hospitals for providing tumor tissue: Canisius-Wilhelmina Hospital, Nijmegen; UMC Utrecht, Utrecht; St Antonius Hospital, Nieuwegein; Eemland Hospital, Amersfoort. The initial case–control study was conducted in collaboration with regional hospitals and cancer registries and with financial support from the Dutch Dairy Foundation on Nutrition and Health. The current study was funded by the Dutch Cancer Society (grants LUW 96-1374 and LUW 98-1687).

References

1
Giovannucci,E. (
2001
) An updated review of the epidemiological evidence that cigarette smoking increases risk of colorectal cancer.
Cancer Epidemiol. Biomarkers Prev.
  ,
10
,
725
–731.
2
Vogelstein,B., Fearon,E.R., Hamilton,S.R., Kern,S.E., Preisinger,A.C., Leppert,M., Nakamura,Y., White,R., Smits,A.M. and Bos,J.L. (
1988
) Genetic alterations during colorectal-tumor development.
N. Engl. J. Med.
  ,
319
,
525
–532.
3
Fearon,E.R. and Vogelstein,B. (
1990
) A genetic model for colorectal tumorigenesis.
Cell
  ,
61
,
759
–767.
4
Powell,S.M., Zilz,N., Beazer-Barclay,Y., Bryan,T.M., Hamilton,S.R., Thibodeau,S.N., Vogelstein,B. and Kinzler,K.W. (
1992
) APC mutations occur early during colorectal tumorigenesis.
Nature
  ,
359
,
235
–237.
5
Bos,J.L., Fearon,E.R., Hamilton,S.R., Verlaan-de Vries,M., Van Boom,J.H., Van der Eb,J.A. and Vogelstein,B. (
1987
) Prevalence of ras gene mutations in human colorectal cancers.
Nature
  ,
327
,
293
–297.
6
Slattery,M.L., Anderson,K., Curtin,K., Ma,K.-N., Schaffer,D., Edwards,S. and Samowitz,W. (
2001
) Lifestyle factors and Ki- ras mutations in colon cancer tumors.
Mutat. Res.
  ,
483
,
73
–81.
7
Kikuchi-Yanoshita,R., Konishi,M., Ito,S., Seki,M., Tanaka,K., Maeda,Y., Lino,H., Fukayama,M., Koike,M. and Mori,T. (
1992
) Genetic changes of both p53 alleles associated with the conversion from colorectal adenoma to early carcinoma in familial adenomatous polyposis and non-familial adenomatous polyposis patients.
Cancer Res.
  ,
52
,
3965
–3971.
8
Aaltonen,L.A., Peltomäki,P., Leach,F. et al . (
1993
) Clues to the pathogenesis of familial colorectal cancer.
Science
  ,
260
,
812
–816.
9
Thibodeau,S.N., Bren,G. and Schaid,D. (
1993
) Microsatellite instability in cancer of the proximal colon.
Science
  ,
260
,
816
–819.
10
Ionov,Y., Peinado,M.A., Malkhosyan,S., Shibata,D. and Perucho,M. (
1993
) Ubiquitous somatic mutations in simple repeated sequences reveal a new mechanism for colonic carcinogenesis.
Nature
  ,
363
,
558
–561.
11
Aaltonen,L.A., Peltomäki,P., Mecklin,J.-P. et al . (
1994
) Replication errors in benign and malignant tumors from hereditary nonpolyposis colorectal cancer patients.
Cancer Res.
  ,
54
,
1645
–1648.
12
Salahshor,S., Kressner,U., Pahlman,L., Glimelius,B., Lindmark,G. and Lindblom,A. (
1999
) Colorectal cancer with and without microsatellite instability involves different genes.
Genes Chromosomes Cancer
  ,
26
,
247
–252.
13
Samowitz,W.S., Holden,J.A., Curtin,K. et al . (
2001
) Inverse relationship between microsatellite instability and K- ras and p53 gene alterations in colon cancer.
Am. J. Pathol.
  ,
158
,
1517
–1524.
14
Bardelli,A., Cahill,D.P., Lederer,G., Speicher,M.R., Kinzler,K.W., Vogelstein,B. and Lengauer,C. (
2001
) Carcinogen-specific induction of genetic instability.
Proc. Natl Acad. Sci. USA
  ,
98
,
5770
–5775.
15
Mills,N.E., Fishman,C.L., Rom,W.N., Dubin,N. and Jacobson,D.R. (
1995
) Increased prevalence of K- ras oncogene mutations in lung adenocarcinoma.
Cancer Res.
  ,
55
,
1444
–1447.
16
Ahrendt,S.A., Decker,P.A., Alawi,E.A., Zhu,Y., Sanchez-Cespedes,M., Yang,S.C., Haasler,G.B., Kajdacsy-Balla,A., Demeure,M.J. and Sidransky,D. (
2001
) Cigarette smoking is strongly associated with mutation of the K- ras gene in patients with primary adenocarcinoma of the lung.
Cancer
  ,
92
,
1525
–1530.
17
Hainaut,P. and Pfeifer,G.P. (
2001
) Patterns of p53 G→T transversions in lung cancer reflect the primary mutagenic signature of DNA-damage by tobacco smoke.
Carcinogenesis
  ,
22
,
367
–374.
18
Hainaut,P., Olivier,M. and Pfeifer,G.P. (
2001
) TP53 mutation spectrum in lung cancers and mutagenic signature of components of tobacco smoke: lessons from the IARC TP53 mutation database.
Mutagenesis
  ,
16
,
551
–553.
19
Freedman,A.N., Michalek,A.M., Marshall,J.R., Mettlin,C.J., Petrelli,N.J., Zhang,Z.-F., Black,J.D., Satscidanand,S. and Asirwatham,J.E. (
1996
) The relationship between smoking exposure and p53 overexpression in colorectal cancer.
Br. J. Cancer
  ,
73
,
902
–908.
20
Slattery,M.L., Curtin,K., Anderson,K., Ma,K.-N, Ballard,L., Edwards,S., Schaffer,D., Potter,J., Leppert,M. and Samowitz,W.S. (
2000
) Associations between cigarette smoking, lifestyle factors and microsatellite instability in colon tumors.
J. Natl Cancer Inst.
  ,
92
,
1831
–1836.
21
Yang,P., Cunningham,J.M., Halling,K.C., Lesnick,T.G., Burgart,L.J., Wiegert,E.M., Christensen,E.R., Lindor,N.M., Katzmann,J.A. and Thibodeau,S.N. (
2000
) Higher risk of mismatch repair-deficient colorectal cancer in alpha(1)-antitrypsin deficiency carriers and cigarette smokers.
Mol. Genet. Metab.
  ,
71
,
639
–645.
22
Kampman,E., Van’t Veer,P., Hiddink,G.J., Van Aken-Schneijder,P., Kok,F.J. and Hermus,R.J.J. (
1994
) Fermented dairy products, dietary calcium and colon cancer: a case–control study in the Netherlands.
Int. J. Cancer
  ,
59
,
170
–176.
23
Kampman,E., Verhoeven,D., Sloots,L. and Van’t Veer,P. (
1995
) Vegetable and animal products as determinants of colon cancer risk in Dutch men and women.
Cancer Causes Control
  ,
6
,
225
–234.
24
Voskuil,D.W., Kampman,E., Van Kraats,A.A., Balder,H.F., Van Muijen,G.N.P., Goldbohm,R.A. and Van’t Veer,P. (
1999
) P53 overexpression and p53 mutations in colon carcinomas: relation to dietary risk factors.
Int. J. Cancer
  ,
81
,
675
–681.
25
Miyoshi,Y., Nagase,H., Ando,H., Horii,A., Ichii,S., Nakatsuru,S., Aoki,T., Miki,Y., Mori,T. and Nakamura,Y. (
1992
) Somatic mutations of the APC gene in colorectal tumors: mutation cluster region in the APC gene.
Hum. Mol. Genet.
  ,
1
,
229
–233.
26
Miyaki,M., Konishi,M., Kikuchi-Yanoshita,R. et al . (
1994
) Characteristics of somatic mutation of the adenomatous polyposis coli gene in colorectal tumors.
Cancer Res.
  ,
54
,
3011
–3020.
27
Rowan,A.J., Lamlum,H., Ilyas,M., Straub,J., Papadopoulou,A., Bicknell,D., Bodmer,W.F. and Tomlinson,I.P.M. (
2000
) APC mutations in sporadic colorectal tumors: a mutational ‘hotspot’ and interdependence of the ‘two hits’.
Proc. Natl Acad. Sci. USA
  ,
97
,
3352
–3357
28
Kampman,E., Voskuil,D.W., Van Kraats,A.A., Balder,H.F., Van Muijen,G.N.P., Goldbohm,R.A. and Van’t Veer,P. (
2000
) Animal products and K- ras codon 12 and 13 mutations in colon carcinomas.
Carcinogenesis
  ,
21
,
307
–309.
29
Greenblatt,M.S., Bennett,W.P., Hollstein,M. and Harris,C.C. (
1994
) Mutations in the p53 tumor suppressor gene: clues to cancer etiology and molecular pathogenesis.
Cancer Res.
  ,
54
,
4855
–4878.
30
Boland,C.R., Thibodeau,S.N., Hamilton,S.R. et al . (
1998
) A National Cancer Institute workshop on microsatellite instability for cancer detection and familial predisposition: development of international criteria for the determination of microsatellite instability in colorectal cancer.
Cancer Res.
  ,
58
,
5248
–5257.
31
Samowitz,W.S., Slattery,M.L., Potter,J.D. and Leppert,M.F. (
1999
) BAT-26 and BAT-40 instability in colorectal adenomas and carcinomas and germline polymorphisms.
Am. J. Pathol.
  ,
154
,
1637
–1641.
32
Pyatt,R., Chadwick,R.B., Johnson,C.K., Adebamowo,C., De la Chapelle,A. and Prior,T.W. (
1999
) Polymorphic variation and the BAT-25 and BAT-26 loci in individuals of African origin. Implications for microsatellite instability testing.
Am. J. Pathol.
  ,
155
,
349
–353.
33
Nicholaides-Bouman,A., Wald,N., Forey,B. and Lee,P. (eds) (1993) International Smoking Statistics . Oxford University Press, London, UK.
34
Coebergh,J.W.W., Van der Heijden,L.H. and Janssen-Heijnen,M.L.G. (1995) Cancer Incidence and Survival in the Southeast of the Netherlands . Comprehensive Cancer Centre South, Eindhoven, The Netherlands.
35
Giovanucci,E., Rimm,E.B., Stampfer,M.J., Colditz,G.A., Ascherio,A., Kearny,J. and Willett,W.C. (
1994
) A prospective study of cigarette smoking and risk of colorectal adenoma and colorectal cancer in U.S. men.
J. Natl Cancer Inst.
  ,
86
,
183
–191.
36
Giovanucci,E., Colditz,G.A., Stampfer,M.J., Hunter,D., Rosner,B.A., Willett,W.C. and Speizer,F.E. (
1994
) A prospective study of cigarette smoking and risk of colorectal adenoma and colorectal cancer in U.S. women.
J. Natl Cancer Inst.
  ,
86
,
192
–199.
37
Fodde,R., Kuipers,J., Rosenberg,C. et al . (
2001
) Mutations in the APC tumour suppressor gene cause chromosomal instability.
Nature Cell Biol.
  ,
3
,
433
–438.
38
Kaplan,K.B., Burds,A.A., Swedlow,J.R., Bekir,S.S., Sorger,P.K. and Näthke,I.S. (
2001
) A role for the adenomatous polyposis coli protein in chromosome segregation.
Nature Cell Biol.
  ,
3
,
429
–432.
39
Slattery,M.L., Potter,J.D., Friedman,G.D., Ma,K.-N. and Edwards,S. (
1997
) Tobacco use and colon cancer.
Int. J. Cancer
  ,
70
,
259
–264.
40
Martinez,M.E., Maltzman,T., Marshall,J.R., Einsphar,J., Reid,M.E., Sampliner,R., Ahnen,D.J., Hamilton,S.R. and Alberts,D.S. (
1999
) Risk factors for Ki- ras protooncogene mutation in sporadic colorectal adenomas.
Cancer Res.
  ,
59
,
5181
–5185.
41
Soussi,T. and Béroud,C. (
2001
) Assessing TP53 status in human tumours to evaluate clinical outcome.
Nature Rev.
  ,
1
,
233
–240.
42
Pluquet,O. and Hainaut,P. (
2001
) Genotoxic and non-genotoxic pathways of p53 induction.
Cancer Lett.
  ,
174
,
1
–15.
43
Conway,K., Edmiston,S.N., Cui,L. et al . (
2002
) Prevalence and spectrum of p53 mutations associated with smoking in breast cancer.
Cancer Res.
  ,
62
,
1987
–1995.