Introduction

Human N-acetyltransferase (NAT) polymorphisms are important for explaining interindividual differences in response to drugs (Meisel 2002) and industrial chemicals (Golka et al. 2002). Human NAT polymorphisms are primarily due to single-nucleotide polymorphisms (SNP) in the coding region of NAT genes (Pompeo et al. 2002). Variant alleles of NAT2 have been identified and assigned to slow and rapid acetylator phenotypes (Grant et al. 1997). A unified gene nomenclature as a consensus has been established (Vatsis et al. 1995; Hein et al. 2000a) (http://www.louisville.edu/medschool/pharmacology/NAT.html).

The concordance and/or discordance of NAT2 genotypes and phenotypes has been a matter of discussion (Butler et al. 1992; Hickman et al. 1992; Bell et al. 1993; Cascorbi et al. 1995; Golka et al. 1996; Gross et al. 1999), and a potential for mis-classification has been noted (Butcher et al. 2002). Specifically, the classifications of the NAT2*12A and NAT2*13 alleles have been challenged. Both alleles are generally addressed as determining a rapid acetylator phenotype (Hein et al. 2000b). However, Gross et al. (1999) have pointed to inconsistencies and concluded that these assignments require further re-evaluation. In this context, they emphasized the need of further investigations of the coherence of NAT2 phenotype and genotype assignments.

In principle, assignment problems may arise from both NAT2 genotyping (Cascorbi and Roots 1999) and phenotyping (see Discussion section). In order to comply with the need for validation of methods in this field, the Deutsche Forschungsgemeinschaft (DFG) has assessed current methodologies and issued a recent volume that contains established and validated methods on genotyping and phenotyping of polymorphic genes/enzymes of xenobiotic metabolism, among others of NAT2 (DFG 2004a, 2004b). On the basis of these validated methods for routine use, we have re-investigated the coherence of NAT2 genotypes and phenotypes to elucidate the apparent differences. The genotyping procedure (DFG 2004a) was based on the methods of Cascorbi et al. (Cascorbi et al. 1995; Cascorbi and Roots 1999). For phenotyping, the caffeine test of Grant et al. (1984) was used, as validated by the DFG (2004b).

Materials and methods

This study of comparative NAT2 genotyping and phenotyping was performed on 211 persons attending the Unit of Clinical Occupational Medicine of the Institute of Occupational Physiology at the University of Dortmund. Most of these persons suffered from transitional cell carcinoma of the urinary bladder and there was suspicion of an occupational disease (in Germany: BK 1301) caused by aromatic amines. In these persons, NAT2 genotyping and phenotyping was part of the medical expertise on the individuals that had to be provided by the institute. This pre-selection caused an over-representation of slow acetylators, compared with the normal central European population (“Caucasians”).

N-Acetyltransferase 2 genotyping

NAT2 genotyping by PCR was performed as described and validated by the DFG (2004a). Whole-blood EDTA samples were stored until processing at –20°C. DNA isolation used commercially available kits, following the manufacturer’s instructions.

The method of the DFG is based on the work of Cascorbi et al. (1995) which originally produced three amplificates where one initially amplified NAT2 DNA fragment comprising 1211 bp served as starting DNA for two further amplificates with sizes of 442 and 420 bp, respectively. Instead, the modified method of the DFG (2004a) uses only two PCR runs, while still applying the usual seven restriction enzymes. The resulting general workflow is as follows: DNA isolation, first PCR with check of success, three restriction enzymes (MSP I, Fok I, Dde I) with gel electrophoresis and second PCR with check of success, three restriction enzymes (Kpn I, Dde I, BamH I) with gel electrophoresis and Taq I restriction overnight, gel electrophoresis of the Taq I restriction cleavage and evaluation. The evaluation followed the current assignment of NAT2 alleles (Hein et al. 2000b), as described in detail by the DFG (2004a).

N-Acetyltransferase phenotyping (caffeine test)

The caffeine test method of Grant et al. (1984) was used for NAT2 phenotyping. The procedure has been validated by the DFG (2004b). The subjects drank 1–2 cups of coffee (sometimes tea or a caffeine-containing soft drink), and spot urine was taken 2 h after caffeine intake. Most of the cases provided a second urine sample after 4 h that was used as an analytical duplicate. No differences were observed in phenotypes based on the duplicate samples, and for duplicate analyses the mean AFMU/1X ratio was taken for further assessment.

The urine was immediately adjusted to pH 3.5 by addition of the required small amounts of 1 mol L−1 hydrochloric acid and stored in 10 mL plastic tubes at –20°C until processing. The chromatographic standard 1-methylxanthine (1X) was a commercial product of Sigma (Deisenhofen, Germany) and the standard 5-acetylamino-6-formylamino-3-methyluracil (AFMU) had been synthesised in our group (Röhrkasten et al. 1997). The HPLC procedure followed the method of Grant et al. (1984), as validated by the DFG (2004b). The urinary AFMU/1X molar ratio served as the final quantitative determinant. As outlined previously (Golka et al. 1996), the determination of the cut-off point (antimode) between slow and rapid acetylator phenotypes is crucial. This is addressed in the Results and Discussion sections (v.i.).

Results

Study group, AFMU/1X ratios and genotypic acetylator assignments

The caffeine test for NAT2 phenotyping was performed on the study group of 211 persons, and the resulting urinary AFMU/1X ratios are compiled in Table 1. Genotypic assignments of slow (S/S homozygotes) and rapid acetylators (R/R homozygotes and R/S heterozygotes) were done according to current procedures and nomenclature (Butcher et al. 2002; Hein et al. 2000b; DFG 2004a, 2004b). The maximum individual AFMU/1X ratio of the homozygous S/S genotypes and the minimum of the homozygous R/R genotypes was practically identical (0.85 and 0.84, respectively). Hence, a AFMU/1X ratio of 0.85 was subsequently taken as cut-off point (antimode) between slow and rapid acetylators. By doing this, there was some overlap in the group assigned to the R/S genotype (v.i.).

Table 1 AFMU/1X ratios vs genotypes, based on the current allele assignments according to Hein et al. (2000a) and DFG (2004a). Genotypic acetylator state: S/S homozygously slow, R/S heterozygotes, R/R homozygously rapid
Full size table

Coincidence of NAT2 phenotyping and genotyping

The basic results of the genotypings and the corresponding phenotypings (AFMU/1X ratios) are given in Table 2 (median, minimum, maximum values) and in Fig. 1 (graphical presentation of individual values). As mentioned above, the present data are consistent with an AFMU/1X ratio of 0.85 that had been described by Braz Vieira de Silva Pontes et al. (1992) as cut-off point (antimode) between phenotypically slow (AFMU/1X<0.85) and rapid (AFMU/1X>0.85) acetylators. If this is accepted, several R/S allele combinations do not comply, either fully or partly, with their assigned phenotypes.

Table 2 AFMU/1X molar ratios in subgroups assigned to NAT2 allele combinations
Full size table
Fig. 1
figure 1

Individual NAT2 phenotyping data (ratios of urinary AFMU/1X), allocated to genotypic groups of NAT2 alleles and allele combinations

Full size image

The current view is that “rapid acetylator alleles” NAT2*4 (“wild type”), NAT2*12A, NAT2*12B, NAT2*12C, and NAT2*13 dispose to a rapid acetylator phenotype (Grant et al. 1997; Butcher et al. 2002; Hein et al. 2000b). If one disregards the only minor overlap of the “rapid” allele combination designated in Table 2 and Fig. 1 as #10 (NAT2*4/NAT2*5A), the data of some heterozygous allele combinations deserve attention. All four persons with the heterozygous allele combination #14 (comprising combinations NAT2*12B/*5E; NAT2*5C/*6A; NAT2*5D/*6C) were phenotypically slow acetylators. This might be ascribed to the alternative allele combinations NAT2*5C/*6A and/or NAT2*5D/*6C in which only slow alleles are present, and is therefore in line with the current concept.

More difficult to explain is the wide phenotypic overlap of the heterozygous allele combinations #11 (NAT2*12A/*5A; NAT2*12C/*5D; NAT2*4/*5B), #15 (NAT2*13/*6B / NAT2*4/*6A) and #16 (NAT2*13/*7A; NAT2*4/*7B). According to the current assignment system, these should be rapid acetylators. However, these groups obviously contain both phenotypically rapid and phenotypically slow acetylators. If one assumes that, in any case, the presence of one “wild type” allele NAT2*4 defines a rapid acetylator, the assignment of the alleles NAT2*12A and/or NAT2*12C (#11), and NAT*13 (#15, #16) as determinants of a rapid acetylator phenotype must be questioned. This refers to the nucleotide changes A803G (NAT2*12A, NAT2*12C) and C282T (NAT2*13). Because of the larger number of persons (n=35) in group #11, this seems especially true for the change A803G (lys268→arg) that is associated with both NAT2*12A and NAT2*12C.

Discussion

This study was based on established standard methods of NAT2 phenotyping and genotyping, as evaluated by the DFG (2004a, 2004b). As far as NAT2 phenotyping by the “caffeine test” method is concerned, it has previously been discussed that the exact positioning of the cut-off point (antimode) between phenotypically slow and rapid acetylators seems laboratory-dependent (Golka et al. 1996), as discriminating AFMU/1X ratios vary between 0.48 (Grant et al. 1984) and 1.0 (Golka et al. 1996). These discrepancies may be attributed to differences in chromatographic conditions. The Deutsche Forschungsgemeinschaft (DFG 2004b) has pointed out that the stability of AFMU depends on pH, temperature, and storage period, and that AFMU may be degraded to 5-acetylamino-6-amino-3-methyluracil (AAMU) by deformylation. The procedure used for synthesis of the standard used in our laboratory had been described by Röhrkasten et al. (1997), to which reference has been made by the DFG (2004b).

As far as the current procedures of NAT2 genotyping are concerned, there is a clear discrepancy in the phenotypic assignment of the NAT2*12A (A803G) and NAT2*13 (C282T) alleles as determinants of a slow or rapid acetylator state. Originally, it had been suggested that NAT2*13 was associated with the slow acetylator phenotype (Cascorbi et al. 1995), but later this was ascribed to a genotyping artefact (Cascorbi and Roots 1999). Clearly, different mechanisms for slow acetylator phenotypes must be considered (Leff et al. 1999), and there is a potential of mis-classification of genotypes and deduced genotypes (Butcher et al. 2002).

On one hand, rapid acetylator phenotypes are currently deduced for the NAT2*12 alleles A–C (Butcher et al. 2002). This is based on the argumentation of Hein et al. (2000b) that recombinantly expressed enzymes of the human NAT2*12 and NAT2*13 clusters reach N-, O-, and N,O-acetyltransferase activities comparable to those of the rapid acetylator NAT2*4 enzyme (Hein et al. 1994, 1995). However, this is not in line with a comparative study of NAT2 genotyping and phenotyping by O’Neil et al. (2000) that points, for persons with the NAT2*12A, NAT2*12B or NAT2*13 alleles, mostly to either a slow acetylator phenotype or a borderline result.

On the other hand, the specific assignments of NAT2*12A and NAT2*13 as determinants of the slow acetylator phenotype have been challenged by the group of Kadlubar, based on their own data (Gross et al. 1999). In their discussion, they pointed out that the assignment of NAT2*12A as leading to a rapid acetylator had been, in first instance, based on data from just two subjects from a single family with NAT2*5B/*12A genotypes (Cascorbi et al. 1996). In addition, the recombinant human NAT2 activity of the NAT2*13 is similar to that of the rapid NAT2*4 allele (Hein et al. 1995). This is explainable because the SNP leading to the NAT2*13 allele does not implicate an amino acid exchange in the enzyme. Therefore, it is being considered that the expression of NAT2*12A and NAT2*13 could be influenced by additional factors (Gross et al. 1999).

In essence, there is now accumulating evidence that the assignments of the NAT2*12A and NAT2*13 alleles as determinants of a rapid acetylator state should be revisited.