Keywords: breast neoplasms; cytochrome P-450 CYP1A1; epidemiology; polymorphism, genetic

Abbreviations:
CI CYP1A1 CYP1A2 GSTM1 GSTT1 NAT NAT1 NAT2 NATP RR	confidence interval cytochrome P4501A1 cytochrome P4501A2 glutathione S -transferase class M1 glutathione S -transferase class T1 N -acetyltransferase N -acetyltransferase type 1 N -acetyltransferase type 2 N -acetyltransferase pseudogene relative risk

In humans, there are three N-acetyltransferase (NAT) loci: two expressed genes, NAT1 and NAT2 , and a pseudogene, N-acetyltransferase pseudogene (NATP) . Both expressed genes are 870 base-pair intron-less protein coding regions encoding 290 amino acid proteins (1) and are located on chromosome 8 (2) at 8p21.3-23.1 (3). 

The two isozymes use acetyl coenzyme A as a cofactor and function as phase II conjugating enzymes (4); they are capable of N-acetylation and O-acetylation and N, O-acetylation (5). N -acetylation is a detoxification pathway.
O-acetylation and N, O-acetylation occur in alternative metabolic pathways following activation by N -hydroxylation. The isozymes differ in their substrate specificities: isoniazid and sulphamethazine are NAT2-specific substrates; p-aminobenzoic acid and p-aminosalicylic acid are NAT1-specific substrates. Amongst the enzyme substrates are several carcinogenic compounds, many of which are present in cooked food and tobacco smoke (6). This has prompted speculation that the NAT enzymes and the genes encoding them may be involved in susceptibility to cancer including colorectal cancer because of the presence of carcinogenic heterocyclic amines in some cooked foods (7). 

NAT2 is primarily expressed in the liver, whereas NAT1 is primarily expressed at other sites, including the colon (8). In colon tissue removed from cadavers, the ratio of NAT1:NAT2 activity was found to change along the length of the intestine (9). Differences between the relative levels of isozyme activity were most marked in the distal colon in one individual; 50-70 fold higher NAT1 than NAT2 activity was observed.

The polymorphic nature of human NAT was first described in 1953 (10); a proportion of individuals receiving isoniazid therapy suffered adverse neurologic side effects due to an accumulation of unmetabolized drug. Family pedigree studies confirmed the genetic basis of the variation (11). Specific single base-pair substitutions responsible for altered enzyme activity were first reported in 1990 (2).  

NAT2 allele classification and nomenclature  

Three NAT2 phenotypes have been described. The fast acetylation phenotype results from possession of two copies of the wild-type allele. If only one allele is wild-type, an intermediate phenotype is observed. Persons with the slow acetylators phenotype possess two mutated alleles. Many early studies did not distinguish between fast and intermediate acetylators, categorising both types of subjects as “fast acetylators”. 

The first NAT2 alleles described were termed M1, M2 and M3. M1 consisted of a transition at nucleotide 481 (C⁴⁸¹T) together with T³⁴²C; M2 consisted of a transition of C²⁸²T and G⁵⁹⁰A; and M3 consisted of a transition of G⁸⁵⁷A. Alleles M1 and M2 accounted for 90% of the slow acetylators in the original study (12), which included 18 subjects phenotyped in vivo and 26 liver samples phenotyped in vitro . M3 was first described in Japan (13).  

The identification and characterisation of new allelic variants by many different laboratories gave rise to conflicting allele designations, which complicating the interpretation of earlier studies. The scheme suggested by Vatsis et al. (5),
(Table 1), provides a nomenclature of currently recognised NAT alleles and facilitates the inclusion of any further alleles. The use of this scheme has simplified the interpretation of more recent studies. According to this nomenclature (5), M1, M2 and M3 should now be termed NAT2 *5A, NAT2 *6A and NAT2 *7A respectively. 

click to view

Table 1  Human NAT2 allele designations (after Vatsis et al., 1995 (5), updated by  personal communication, 1998)

Ten point mutations have been reported in NAT2 (5), each a single base-pair substitution. Many published reports have investigated only single mutations and have based allele designations on this. However, recent improvements in techniques for detecting individual polymorphic sites have shown that isolated single substitutions are uncommon; combinations of mutations are more common. Within certain populations, some substitutions have been consistently observed to co-segregate (e.g. C⁴⁸¹T rarely occurs without T³⁴¹C (14)) . 

Functional significance of NAT2 mutations  

The NAT2 alleles described so far may contain up to four of the ten reported mutations (5). The functional significance of most combinations is unknown. However, it is plausible that each combination might result in a different phenotype.  

The functional significance of the ten mutations is summarised in Table 2 (12, 13, 15-19). Some of the mutations change the amino acid sequence of the resultant enzyme, but not all of these have been observed to alter phenotype (e.g. A⁸⁰³G (19)). However, some mutations, however, have been consistently observed to reduce acetylation activity (e.g. T³⁴¹C) (20). The functional effect on phenotype is due to impairment of the protein translation or stability; messenger ribonucleic acid (mRNA) levels are not altered (12). For several of the mutations, their designation as “fast” or “slow” is not yet definitive. 

click to view

Table 2  Functional significance of NAT2 mutations

The C⁴⁸¹T transition does not change the amino acid sequence; however, it has always been found with other mutations. In populations of European origin, C⁴⁸¹T most often occurs with the T³⁴¹C and A⁸⁰³G ; when C⁴⁸¹T is combined with T³⁴¹C, enzyme activity is reduced. G¹⁹¹A was discovered due to phenotype/genotype discordance in African-American subjects (16). Site directed mutagenesis (serial replacement of nucleotides within the gene coding region) showed codon 64 to be highly conserved between species and is implicated as the active site for acetyl transfer (17). Disruption of this region has been demonstrated to abolish enzyme activity in vitro . 

In European populations, relatively high concordance between acetylator phenotype (for NAT2 specific substrates) and NAT2 gene mutations has been demonstrated (12, 21).  

NAT1 allele nomenclature  

The study by Weber et al. (22) revealed several allelic variants, NAT1 was assumed to be monomorphic. The original alleles, designated V 1 (wild-type), V 2 (T¹⁰⁸⁸A and C¹⁰⁹⁵A) and V 3 (C ^-344T, A^-40T, a 9 base pair deletion between nucleotides 1065 and 1090, and C¹⁰⁹⁵A), have now been incorporated into the Vatsis nomenclature (5) and redesignated NAT1 *4, NAT1 *10, and NAT1 *11 respectively. 

Functional significance of NAT1 alleles  

Inter-individual variation in NAT1 activity has been reported. In addition, a two-fold intra-individual variation in activity (in 75 peripheral blood samples) was observed over a ten week period (23). If this finding is confirmed, it might suggest that NAT1 is inducible by exposure to endogenous or exogenous factors. 

The functional significance of the NAT1 allelic variants has not been fully established. Mutations can both increase and decrease the acetylation capacity relative to the wild-type allele. On the basis of the studies published to date (24-26), NAT1 *4 is considered the wild-type. The alleles that are believed to increase NAT1 acetylation capacity are NAT1 *10, *21, *24, *25. Alleles NAT1*14, *15, *17, *19 and *22 give rise to enzymes with reduced or no detectable activity. Alleles NAT1 *11, *20 and *23 produce enzymes with acetylation capacity similar to that of NAT1 *4. The current allele designations refer mainly to single substitutions, insertions or deletions. It is possible that combinations of these substitutions will be found to occur; these combinations may have functional consequences. 

Population frequencies  

We searched Medline and EMBASE using the Medical Subject Heading “arylamine N -acetyltransferase” and the textwords “NAT”, “NAT1 ”, “NAT2 ” and “N -acetyltransferase”. We also used the Medical Literature Search procedure in the Office of Genomics and Disease Prevention at the Centres for Disease Control and Prevention
( Atlanta , Georgia ). In addition, we reviewed reference lists in published articles. We identified and critically appraised relevant articles. This section includes studies reporting phenotype and genotype references 7, 14, 16, 18, 21, 23, 27-95; Table 3, frequencies in a variety of individuals without cancer. Studies reporting only the frequency of individual mutations or alleles were excluded since phenotype cannot be imputed. 

click to view

Table 3  Frequency of predicted NAT2 acetylator status in different groups of subject

In many published articles the studies subject selection criteria are not stated. When they were stated, the criteria were diverse. Some studies have included disease-free subjects matched to diseased patients on characteristics such as age and sex, others were based on hospitalised subjects, specific occupational groups or volunteers for whom recruitment procedure was not described. This made it difficult to determine the extent to which apparent geographic or ethnic variation reflected biological differences or methodological factors. For example, the lowest frequency of fast/intermediate NAT2 acetylation genotypes reported in Europe (12%) was based on specimens obtained from a cell bank in France (52); this frequency is substantially lower than the frequencies of 39% and 47% observed in other studies carried out in France (53, 54) and elsewhere in Europe. 

In most NAT2 genotype studies only a limited number of “indicator” mutations, thought to be tightly linked with other mutations and predictive of acetylator status, have been investigated. This is likely to have led to underestimation of the proportion of NAT2 slow acetylators. The designation of alleles according to the presence/absence of “indicator” mutations assumes particular patterns of linkage, which may not be tenable in other populations or ethnic groups. For example, the genotype-phenotype discordance observed for African-Americans and USA Hispanics may result from compound alleles that are different to those observed in other populations (14). 

The frequency of NAT2 genotypes associated with fast or intermediate acetylation varies markedly between, and to some extent within, continents. The highest frequency occurs in Asia , particularly in Japan (approximately 90%). The frequencies reported in other South-East Asian populations are 73% in Hong Kong (14), 72% in Malaysia (46) and 58% in Singapore (47). Studies carried out in other parts of Asia have reported lower frequencies: 32% in India (39), 37% in the United Arab Emirates (49) and 43% in Turkey (48). In most European populations, approximately 40% of study subjects have genotypes associated with fast or intermediate acetylation. Genotype frequencies within the USA vary by ethnic group: for white subjects frequencies are similar to European populations and for Asians these are similar to those populations in South-East Asia . The lowest frequencies which have been reported in two small African studies in which subject selection was not described (27). Interestingly, higher frequencies have been reported in African-Americans (14).  

All but three of the studies (83, 84, 88) of NAT phenotype used NAT2 substrates. The geographical variation in the frequency of the fast/intermediate NAT phenotype is generally consistent with that observed for NAT2 genotype. 

Few studies to date have investigated NAT1 genotype. So far, NAT1 *10 is the most common variant of those investigated. The reported frequency of wild-type homozygosity ranges from 24%-96%. In the UK alone, variation of 29%-96% was observed. The frequency of NAT1 wild-type homozygosity within a study depends upon the alleles investigated. Hubbard et al. (95) considered NAT1 *14 and NAT1 *15, whereas Bell et al. (25) examined NAT1 *3, NAT1 *10 and NAT1 *11. In the USA , the wild-type homozygote frequencies are 44%-62% among whites (35, 38) and, in the one reported study which published results, 24% in African-Americans (35). Studies carried out in Australia (23) and Japan (42) reported frequencies of 92% and 38% respectively.

Worldwide in 1996, an estimated 876,000 new cases of colorectal cancer occurred - 445,000 in males and 431,000 in females (96).  Less than one third of colorectal cancer cases occur in developing countries (97).  In developed countries colorectal cancer is the second most common cancer in both sexes (97). 

World age-standardised incidence rates are lowest, around 10 per 100,000 population per annum, in Africa, India, Thailand and in some Chinese populations (98).  The highest rates, more 40 per 100,000 in men and more 30 per 100,000 in women, are observed in North America, Northern Europe, Australia and New Zealand.  In many populations colorectal cancer incidence rates have been rising (99) with the greatest increases being observed in Japan. For example, in Miyagi, Japan, the rate among males increased from 19.7 per 100,000 in 1978-1981 to 41.5 per 100,000 in 1988-92 and the rate among females from 16.8 per 100,000 to 24.8 per 100,000 (98, 100). 

Less than 10% of colorectal cancers are believed to be due to recognised genetic syndromes (familial adenomatous polyposis and hereditary non-polyposis colorectal cancer) (101).  After exclusion of these syndromes, however, familial aggregation has been observed (102, 103), suggesting a role of genetic susceptibility in disease aetiology.  Results of migrant studies indicate that environmental factors have an important influence (104, 105).  Thus, this evidence suggests that the majority of cases are probably to be due to a combination of environmental or lifestyle exposures and genetic susceptibility. 

A high intake of vegetables is inversely associated with the risk of colorectal cancer and it is possible that increased intakes of fibre, starch and carotenoids are protective (106).  There is consistent evidence that the most physically active groups in the population are at lower risk (107).  On the basis of evidence from over 20 observational studies, it has been concluded that regular use of aspirin reduces risk (108). 

Increased risk has been associated with diets high in sugar, total and saturated fat, eggs and processed meat, although the evidence is inconsistent (106).  More consistent evidence exists for a positive association with red meat consumption (106), with increased risk possibly being due to exposure to heterocyclic aromatic amines formed when meat is cooked to pyrolytic temperatures (6), rather than to consumption of meat per se (109).  A recent large case-control study found that although “usual” dietary intake of heterocyclic amines was not associated with increased colon or rectal cancer risk, very high daily intake was (110).  A role for the N-acetyltransferase enzymes in the activation of heterocyclic aromatic amines has been proposed (109).  The N-acetyltransferase enzymes are also involved in the activation of aromatic amines (7) found in tobacco smoke (111).  While tobacco smoking has consistently been associated with adenomatous polyps, the evidence with regard to colorectal cancer is less strong (112).  While some recent large cohort studies have suggested that smoking may increase risk after a long latent period (113-116), this has not been a consistent finding (117)

The studies discussed in this section, and the following section, were identified using the search strategy described above with the addition of Medical Search Headings and text words relevant to colorectal cancer or polyps. These studies are summarised in Tables 4 and 5 (7, 8, 24, 25, 30, 31, 37, 38, 40, 66, 78, 81, 90, 95, 118-126). 

click to view

Table 4  Summary of studies of colorectal neoplasia and acetylator phenotype

click to view

Table 5  Summary of studies of colorectal neoplasia and NAT1 and NAT2 genotype

Until genotyping techniques were developed (2), investigations of the relation between acetylator status and colorectal cancer relied on phenotyping. Probe drugs such as isoniazid, sulphamethazine and caffeine were administered and the metabolites were measured by high performance liquid chromatography. If the probe drug was NAT2-specific (as it is for most studies), it would fail to account for NAT1 activity. NAT1 phenotyping has been done to validate genotyping or to investigate the effects of individual alleles but it has not been used in assessing the association between colorectal cancer and acetylator status. In case-control studies, phenotypic assessment of acetylation may be influenced by the disease status.  

Five studies have investigated the association between acetylator phenotype and colorectal cancer (Table 4) (7, 90, 119-121); two of them simultaneously investigated phenotype and colorectal adenomatous polyps (120, 121). Four of the five cancer studies (7, 90, 119, 120) and one of the polyp studies (120) used NAT2 specific probe drugs. Table 5 summarises data from eleven studies of the relation of NAT2 genotype to colorectal cancer (8, 25, 37, 38, 40, 66, 78, 81, 122, 123, 125, 126) and three studies of polyps. (24, 30, 31, 124)  The participants investigated by Lin et al.(24) overlap with those investigated by Probst-Hensch et al. (30, 124) Five cancer studies (25, 38, 95, 123, 125) and two polyp studies (24, 30) also investigated NAT1 genotype (Table 5). 

Colorectal neoplasia and acetylator phenotype  

The results of three of the four studies using NAT2 specific probe drugs suggested a positive association between fast acetylator phenotype and colorectal cancer (7, 119, 120). In the fourth study, no association with colorectal cancer was found (90). Roberts-Thompson et al. (120) also analysed a series of polyps; no association was found with the acetylator phenotype. Lang et al. (121) used caffeine, which is not an NAT2 specific substrate, to determine acetylator status; no association with cancer and polyps was found (relative risk (RR)=1.3 [95% confidence interval (CI) 0.8-2.3]).  

Colorectal cancer and NAT2 genotype  

In ten of the eleven studies of invasive colorectal cancer and NAT2 acetylator genotype (Table 5), no association with fast/intermediate acetylator genotype was observed. The remaining study reported a statistically significant positive association with fast acetylation genotypes (RR=2.0 95% CI [1.3-3.2]) (66). 

Several issues affect the interpretation of these studies. Two studies were based on tissue samples, obtained either at surgery (122) or from a tissue sample bank (8) . No association was apparent in either study. However, there was little information on the subjects from whom the samples were obtained and the number of samples was small (<50) which limited statistical power. The other studies were all based on at least 100 cases. However, the study in which a significant association was found was the smallest of these (66). In this study, controls were statistically significantly younger than cases; if NAT2 genotype were associated with survival, this may have biased the relative risk. In addition, the areas of residence of cases and controls differed (66). 

Three studies included controls who would be expected to have been representative of the population-at-risk of developing the disease (37, 38, 81). In the other studies, either the controls may not have been representative of the population-at-risk (8, 25, 66, 78, 122, 125) or the methods of control selection were not clearly described (40, 123).

The methods of case selecting varied between studies, possibly affecting their comparability. For example, Hubbard et al. (78) included operable cases and Bell et al. (25) included a “sample” of incident tumours. If the reasons for exclusion were related to disease aetiology or progression, this might have influenced the observed result. The study by Slattery et al. (37) related only to colon cancer, whereas the others also included rectal cancer (but would have had lower statistical power to investigate subsite-specific associations). 

Although one study (66) observed an association between colorectal cancer and NAT2 genotype, other studies have detected associations within subgroups. Hubbard et al. (78) reported a positive association between colorectal cancer and slow acetylation amongst subjects under seventy years (RR=1.7 95% CI [1.1-2.6]). In contrast, Slattery et al. (37) reported a positive association between colon cancer and fast/intermediate acetylator status among older women
( >67 years) (RR=1.4 95% CI [1.0-1.8]). In one study, data on specific NAT2 alleles were presented (125). There was a positive association between NAT2 *7A allele and colorectal cancer (RR=2.4 95% CI [1.5-3.9]). However, the interpretation of this finding is limited by the potential selection bias noted above. 

Colorectal cancer and NAT1 genotype  

Investigations of NAT1 and colorectal cancer were prompted by observations that NAT1 is expressed to a greater extent than NAT2 in the colon. It would therefore be expected that localised activation of the heterocyclic or aromatic amines within the colon would be predominantly due to NAT1 (8, 9). The low frequency of some NAT1 allelic variants could result in limited statistical power to detect any effect. 

An association between NAT1 and colorectal cancer was observed in only one (25) of five studies (Table 5). That study investigated a limited selection of the most common alleles and found a statistically significant increased risk associated with NAT1 *10 (RR=1.9 95% CI [1.2-3.1]). These investigators had previously demonstrated that NAT1 *10 was associated with higher acetylation activity in colon tissue (127). In the studies by Chen et al. (38) and Lee et al. (125), genotypes were assigned on the basis of the alleles investigated by Bell et al. (25). Jenkins et al. (123) did not specify the alleles investigated and Hubbard et al. (95) only investigated two relatively uncommon alleles. 

Colorectal cancer and combined NAT1 and NAT2 genotypes  

Five studies that determined NAT1 and NAT2 genotypes also investigated the effect of combinations of these genotypes(25, 38, 95, 123, 125). None reported any increased risk associated with any combination of NAT1/NAT2 genotypes. However, the statistical power to investigate some of these combinations would have been low. In addition, the limitations regarding study design of alleles detected discussed above also apply. 

Colorectal polyps and acetylator genotype  

Three studies have investigated the relationship between colorectal polyps and aspects of NAT genotype (Table 5;24, 30, 31, 124). No association was found between NAT1 or NAT2 genotype and adenomas. However, an association between NAT1 *10 and risk was found when the cases were restricted to “incident” adenomas (i.e. those with negative sigmoidoscopy results within the previous five years) (30). Subjects were recruited from persons undergoing sigmoidoscopy. It is possible that some of the controls may have harboured adenomatous polyps out of reach of the sigmoidoscope. This would have biased any association with NAT genotypes towards the null. Probst-Hensch et al. (30) suggested that the lack of an overall association might reflect the presence of undetected NAT1 mutations. However, the results of subsequent re-analysis (24) in which the NAT1 gene was screened for mutations but no association between genotype and disease was found, make this unlikely. One study analysed separately individuals with hyperplastic polyps only (31). No association was found between NAT2 genotype and disease.  

Comment on inconsistency between phenotype and genotype studies  

Overall the studies involving assessment of genotype provide little evidence of an association between acetylator status and risk of colorectal lesions. However, the studies of phenotype suggest a positive association between fast acetylation and disease risk. This inconsistency could be due to discordance between genotype and phenotype. As suggested above, the designation of the NAT2 and NAT1 genotypes to imputed phenotype is not yet definitive. Discrepancies between genotype and phenotype have been observed in approximately 5-7% of subjects assessed in studies of European populations (128). However, fast acetylation status per se would not be expected to raise risk in the absence of exposure to NAT substrates, therefore it is likely that the genotyping studies more accurately reflect risk attributable to acetylation status alone. The explanation for the increased risk observed in the phenotype studies is not clear, but possible contributing factors include: alteration of acetylation phenotype by the presence of disease; selection or participation bias of cases and/or controls; confounding of phenotype by exposures which cause colorectal cancer and chance.

It would not necessarily be expected that NAT genotype would be independently associated with risk of colorectal neoplasia. If the NAT genes have a role in the aetiology of colorectal neoplasia, it is likely that it is as a modifier of the relationships between particular environmental exposures and disease. Mechanisms by which environmental exposures might lead to malignancy, involving NATs, other enzymes and concomitant exposure to their substrates, have been proposed (129).  

The NAT substrates are also substrates for other enzymes in the putative detoxification/activation pathways of aromatic amines (129). A substrate may be either hydroxylated or N-acetylated by cytochrome P450 or NAT respectively. The hydroxylated substrate may undergo O-acetylation catalysed by NAT. The metabolic fate of each substrate depends upon the relative activity, specificity and affinity of the enzymes in these competing pathways towards that substrate. It is not clear at present how the metabolic fate of specific substrates may be affected by the various allelic forms of the NAT genes. 

Of the eleven studies of colorectal cancer and NAT2 genotypes, only three assessed possible exposure of both cases and controls to NAT substrates and analysed this together with genotype (37, 38, 81); one other study provided genotype-exposure data for cases only (25). The three studies investigating colorectal adenomas all collected exposure data (24, 30, 124); however, only one analysed this with genotype (124). Most of the studies which have investigated interactions have had relatively small numbers of subjects in the acetylator status-environmental exposure subgroups, which limited statistical power to detect an interaction should one have been present. 

NAT genotype and dietary exposures

The joint effects of dietary exposure and NAT genotype were investigated in three studies (38, 81, 126). Among NAT2 fast acetylators, Welfare et al. (81) found a significantly raised risk of colorectal cancer associated with consumption of fried meat more than twice weekly, as compared with less frequent consumption (RR=6.0 95% CI [1.3-55.0]). This was not found among slow acetylators. In the study by Chen et al. (38), among fast acetylators (based on combined NAT1*10 and fast/intermediate NAT2 genotypes) the relative risks for >0.5-1 and >1 daily servings of red meat versus £ 0.5 daily servings were modestly, but non-significantly, raised (>0.5-1 daily servings: RR=2.1 95% CI [0.81-5.65]; >1 daily servings: RR=2.4 95% CI [0.77-7.12]). This was not seen in non-fast acetylators. This pattern was more pronounced among subjects aged >60 years. However, the test for interaction was not statistically significant either for subjects of all ages (p=0.16) or for older subjects (p=0.25). The method by which the red meat was cooked was not reported nor was the preference for well-done meat. 

Kampman et al. (126), in further analysis of the subjects included in the study by Slattery et al. (37), investigated associations between various measures of meat consumption, NAT2 genotype and colon cancer. Among persons with the intermediate/rapid genotype, there were modestly raised risks associated with (1) higher consumption of red meat, (2) preference for well done red meat, (3) high levels of a red meat mutagen index, (4) higher consumption of processed meat and (5) higher levels of a total meat mutagen index. Unexpectedly, consumption of white meat was more strongly associated with risk among slow acetylators, but the relative risks were only modestly raised. 

NAT phenotype and dietary exposures

Two studies have investigated interactions between acetylator phenotype and diet and risk of colorectal lesions, with inconsistent results (118, 120). Wohlleb et al. (118) modelled various aspects of diet, phenotype and colorectal cancer risk. While consumption of luncheon meat and pork were each significantly associated with disease risk; the introduction of acetylator status into the model had no significant effect.  

Roberts-Thomson et al. (120) stratified subjects into slow and fast acetylator phenotype groups and assessed the linear trend in risk of (1) adenoma and (2) cancer across three categories of meat intake (low, medium and high) in the two strata. It is not clear how these categories of meat intake were determined. Among slow acetylators, there was no association between meat intake and risk of adenoma. Among fast acetylators, adenoma risk increased with increasing meat consumption (continuous variable RR=2.1 95% CI [0.9-4.7]; p for linear trend=0.08). For colorectal cancer, among slow acetylators the relative risk did not differ significantly from 1 across the categories of meat intake. Among fast acetylators, cancer risk increased with meat intake (continuous variable RR=1.7 95% CI [0.9-3.5]; p for linear trend=0.13). 

NAT genotype and smoking

The possibility of an interaction between NAT genotype and smoking was assessed in four studies (31, 37, 81, 124), with inconsistent results. Slattery et al. (37) found a modest association between several measures of tobacco exposure and colon cancer in men and women but this effect was modified only slightly by NAT2 genotype. 

In their study of hyperplastic polyps and adenomas, Potter et al. (31) also found significant association between smoking status and pack years of smoking and both types of polyps, but these associations were not modified by NAT2 genotype. Welfare et al. (81) reported that cigarette smoking in the past five years was not associated with cancer risk among NAT2 fast acetylators, but it was associated with significantly raised risk among slow acetylators (RR=2.3 95% CI [1.2-4.6]). 

In their study of adenomas, Probst-Hensch et al. (124) observed a raised risk in current smokers who were NAT2 fast acetylators, in comparison with persons who had never smoked and were slow acetylators (RR=2.3 95% CI [1.0-5.2]). Consistent with this, in an analysis of colorectal cancer cases only, NAT1 *10 was found to occur more frequently among smokers (52%) than among non-smokers (41%) (25). NAT2 fast acetylator genotypes were also more common among smokers (52%) than non-smokers (45%).

NAT genotype and other genes

Other enzymes in the detoxification/activation pathway are also polymorphic. For example, the glutathione S-transfearse class M1 (GSTM1) and glutathione S-transfearse class T1 (GSTT1) genes, involved in detoxification, are polymorphic (130), functionally significant alleles of cytochrome P4501A1 (CYP1A1) have been reported (131) and genetic polymorphism in cytochrome P4501A2 (CYP1A2) has recently been demonstrated (132, 133). This raises the possibility that these genes may interact to affect disease risk. Gene-gene-environment interactions have been investigated in three studies (37, 121, 124).  

Slattery et al. (37) stratified their subjects by joint GSTM1 and NAT2 genotype and investigated colon cancer risk associated with smoking within each stratum. Among men who were NAT2 -slow and GSTM1 -null, the relative risks associated with smoking <20 and >20 cigarettes a day, compared with not smoking, were 1.4 95% CI [0.8-2.3] and 1.7 95% CI [1.2-2.6] respectively. This trend was not observed in other genotype strata, or among women. Probst-Hensch et al. (124) reported a significantly raised risk of adenomas for fast acetylators who were GSTM1-null in comparison to slow acetylators who were GSTM1 -non-null, among current smokers (RR=10.3 95% CI [1.94-55.0]). This genotype combination was not associated with raised risk among never smokers (RR=1.0 95% CI [0.5-2.2]). 

Lang et al. (121) considered the joint effects of NAT2 phenotype, CYP1A2 phenotype (rapid or slow) and meat cooking preference (rare/medium or well-done) on the risk of colorectal cancer and polyps combined. The reference category comprised subjects who were NAT2 -slow, CYP1A2 -slow and preferred rare/medium meat. The relative risk for subjects who were NAT2 -rapid/CYP1A2 -rapid and preferred rare/medium meat was 3.1 and the relative risk for subjects who were NAT2 -rapid/CYP1A2 -rapid and preferred well-done meat was 6.5. Confidence intervals were not reported. The result for the test for interaction was not significant. It is likely that each stratum contained few subjects.

Early case-control studies on acetylation and colorectal cancer used phenotyping methodologies (Table 4). The NAT genotyping techniques currently used in epidemiological studies rely on initial amplification of the region of the gene in which the polymorphisms are found. Following amplification, restriction fragment length polymorphism (RFLP) analysis and allele-specific polymerase chain reaction (PCR) are most commonly used. An oligo-ligation assay has recently been developed and it is particularly suitable for automated studies (134).

Because of the marked inter-ethnic differences in NAT genotypes, it is important that the appropriate mutations be investigated so that alleles can be assigned correctly. Failure to fully define alleles may bias the estimate of imputed acetylator phenotype. For example, Lin et al. (24) re-analysed samples from the study by Probst-Hensch et al. (124) and re-categorised twenty subjects, who had been classified as “fast” in the original study, as slow acetylators. In addition, if additional substitutions are not explicitly detected by the techniques employed, but are assumed to be present due to previously observed linkage patterns, it is important that this be explicitly stated in the characterisation of the alleles.

There is currently insufficient evidence implicating polymorphic NAT genes in the aetiology of colorectal cancer or adenomatous polyps to justify population testing.

List of References

• Table 1 - Human NAT2 allele designations (after Vatsis et al., 1995 (5), updated by personal communication, 1998)
• Table 2 - Functional significance of NAT2 mutations
• Table 3 - Frequency of predicted NAT2 acetylator status in different groups of subject
• Table 4 - Summary of studies of colorectal neoplasia and acetylator phenotype
• Table 5 - Summary of studies of colorectal neoplasia and NAT1 and NAT2 genotype

Data on Disease Frequency

• IARC – Cancer Mondial (last accessed 5/2007)
• SEER (last accessed 5/2007)

General Information on Cancer

• Cancer Research UK (last accessed 5/2007)
• American Association of Cancer Research (last accessed 5/2007)
• National Cancer Institute (last accessed 5/2007)
• International Union against Cancer (last accessed 5/2007)

Gene Specific Information

• CDC Office of Genomics and Disease Prevention – Medical literature search (last accessed 5/2007)
• Public Health Genetics Unit (last accessed 5/2007)
• Human Gene Mutation Database (last accessed 5/2007)
• OMIM – Online Mendelian Inheritance in Man (last accessed 5/2007)
• GenAtlas (last accessed 5/2007)
• UniGene (last accessed 5/2007)
• GeneCards (last accessed 5/2007)
• The National Center for Biotechnology Information (last accessed 5/2007)
• Links to Chromosome Specific Databases and Other Site (last accessed 5/2007)