Main

Acetylation is an important biotransformation step of many drugs and arylamine xenobiotics(1,2). Acetylation reactions are catalyzed by two N-acetyltransferase isozymes, which are encoded at two loci. One locus, nat, encodes N-acetyltransferase NAT1 that is thought to be a "monomorphic" isozyme, although genetic variants have been recently reported. The second nat locus is polymorphic and encodes NAT2 (EC 2.3 1.5) that is markedly reduced in genetically slow acetylators(3,4). Nine different point mutations have been found in the NAT2 gene, seven of which lead to amino acid changes: G→A at nucleotide (nt) position 191 (Arg to Glu), T→C at nt position 341 (Ile to Thr), A→C at nt position 434 (Gln to Pro), G→A at nt position 590, A→G at nt position 803 (Lys to Arg), A→C at nt position 845 (Lys to Thr), and G→A at nt position 857 (Gly to Glu). The remaining two C→T point mutations at nt position 282 and 481 cause no amino acid changes but occur together with other mutations(5–7). Wide distribution of these mutations is responsible for the fact that >60% of the white population show decreased activity of NAT2 or lack of it(3–8). Lack of activity of the enzyme may lead to idiosyncratic reactions to drugs and to so-called environmental diseases induced by exposure to arylamine xenobiotics(1,9).

The activity of NAT2 in vivo is studied by the evaluation of acetylation phenotype. In recent years, caffeine (1,3,7-methylxanthine) has been used as a model drug in acetylation phenotype studies. Its main metabolite, formed in a reaction catalyzed by NAT2, is AFMU(10,11). Urinary metabolic ratio of AFMU to 1X is used for the evaluation of NAT2 activity(12–15). However, in infants, the evaluation of acetylation phenotype is not a standard procedure because at this age physiologic differences in body build and composition result in great variability of drug pharmacokinetics, which changes during development(16). Pariente-Khayat et al.(17) showed in prospective studies including 54 infants that during the first year of life in >60% of the studied infants, a change from slow to fast acetylation phenotype occurred. Therefore, to detect a genetically determined defect of acetylation in infants and young children, other supplementary methods must be used. In the present study, we have implemented the identification of the genotype coding for NAT2 as well as the evaluation of the phenotype for this purpose. We limited our analysis to the NAT2 missense mutations G590A, A803G, and G857AS, and silent mutation C 481T. Taking into account that the T341C mutation is linked with either 481T or 803G, the T341C mutation has not been determined in our study. Other mutations, rare in the white Caucasian population, were not identified.

METHODS

A group of 82 children (all Caucasians) aged from 1 mo to 17 y (57 boys and 25 girls) and including 37 (45%) children <1 y of age was studied. Children were hospitalized at the Institute of Paediatrics in Łódź because of respiratory tract infections of varying etiology (viral and bacterial). The development of all the children was normal, past medical history was uneventful, and there was no family history of genetic disorders. In all the children, the phenotype was studied and, during routine hospital diagnostics, 1 to 2 cm3 of blood was collected to identify genotype coding for NAT2. The study was approved by the Ethical Committee of the Medical University of Łódź-Poland.

Acetylation phenotype study protocol. The acetylation phenotype was evaluated on the basis of the metabolic ratio of AFMU to 1X, which was determined by the chromatographic analysis of urine from children after a single dose of drug. The procedure was done after therapy had been completed and the children were not receiving drugs, in particular theophylline, barbiturates, anticonvulsants, ketoconazole, cimetidine, etc., which could alter the results of the test. All children had normal kidney and liver function tests. Caffeine was given orally as a solution of coffeinum natrium benzoicum (Galena PL) on empty stomach in the dose of 2.5 mg/kg per body weight. Six to eight hours after ingestion of caffeine, urine was collected, acidified with ascorbic acid to pH 3.5, and frozen at -20°C for further analysis.

Chromatographic separation of caffeine metabolites. Caffeine metabolites were extracted as described by Grant et al. and as modified by Klebovich and Evans(12–14,18). Briefly, 100 µL of urine with 5 µL of N-acetyl-p-aminophenol (120 mg/L of chromatographic mixture) as internal standard was saturated with 60 mg of ammonium sulfate (Sigma Chemical Co., Aldrich), mixed with 6 mL of chloroform:isopropanol [85:15, vol/vol (Sigma Chemical Co., Aldrich)], shaken and vortexed vigorously, and then centrifuged for 5 min at 10 000 × g. The organic phase was dried under a steam of nitrogen, and the residue was resuspended in 95 µL of the mixture 10% methanol (Baker isocratic grade, HPLC) in water acidified with acetic acid (Sigma Chemical Co., Aldrich).

Metabolites were separated on a reversed-phase C18 Spherisorb ODS2 cartridge column (5 µm, 4.6 × 250 mm) on a Thermo-Separation HPLC System (P2000 Binary Gradient Pump, AS3000 Autosampler, UV1000 Detector, PC1000 Software v.2.5). Fifty-microliter samples were eluted with 10% methanol in water containing 0.05% acetic acid at a flow rate of 1.2 mL/min at 36°C and monitored by UV absorbance at 280 nm. Retention times of metabolites were achieved through analysis of blank samples spiked with a known amount of AFMU (gift of Dr. René Fumeaux, Nestle Research Centre, Lausanne, Switzerland), 1U, 1X, 1,7 X, and N-acetyl-p-aminophenol (all from Sigma Chemical Co., Aldrich); times were 3, 3.5, 7.2, 17.1, and 8.4 min, respectively. The linearity of detector response to different concentrations of each compound was determined at concentrations of 25, 50, 100, 200, and 400 µM. The limit of detection was 1 µM for AFMU and 4 µM for 1U, 1X, and 1,7 X. The interday repeatability of the method was evaluated by analysis of five urine samples spiked to five concentrations: 25, 50, 100, 200, and 400 µM. The coefficients of variation were 2-6% for all standard metabolites.

For each patient, the metabolic ratio of AFMU/1X was calculated. According to the literature data, fast acetylation phenotype was determined when metabolic index AFMU/1X was >0.4(12,17).

NAT2 genotype studies

DNA isolation. DNA was isolated from 1- to 2-mL samples of whole blood mixed with EDTA as follows. After isolation and the disruption of leukocytes with the lysing solution (10 mM Tris-HCl, 150 mM NaCl, 0.2% SDS), proteinase K (Sigma Chemical Co., Aldrich) was added, followed by phenol/chloroform extraction. DNA was then precipitated in isopropyl/alcohol, alcohol, and DNA samples were dissolved in distilled water and stored at -20°C for further analysis.

PCR amplification. Oligonucleotide primers NATPL6 (5′-GCC TCA GGT GCC TTG CAT TT) and NATPP6 (5′-CGT GAG GGT AGA GAG GAT AT) were synthesized in the Department of Bioorganic Chemistry, Center of Molecular and Macromolecular Studies, Polish Academy of Sciences, Łódź, Poland. Specific sequences of these primers were selected with "Oligo 4.0" software (designed by Dr. W. Rychlik) by use of the NAT2 gene sequence(19). The amplified DNA was 535 bp long, and the spanned sites of mutations were recognized by four restriction enzymes: BamHI (USB), DdeI (Pharmacia), KpnI (BioLabs), and TaqI (Peterfarm). Each PCR cycle was performed in 100 µL (total volume) of 10 mM Tris-HCl (pH 8.9), 50 mM KCl, 1.5 mM MgCl2, 1 µg of DNA, 200 µM of each dNTP, 0.1 µM of each primer, and 2 U of Taq DNA polymerase (Promega). To start the reaction, a "hot-start" procedure was used. Briefly, after initial denaturation (95°C, 5 min) and cooling (4°C, 1 min), 45 cycles of amplification (denaturation 95°C, 1 min; annealing 51.1°C, 1 min; extension 72°C, 2 min) were carried out, followed by the final extension for 7 min at 72°C. After amplification, the DNA product from each sample (10 µL) was analyzed by PAGE (6%) in 0.05 M TBE buffer.

Identification of mutant alleles by RFLP. RFLP methodology was used to identify mutations in the NAT2 gene. Briefly, digestion of the amplified DNA samples was carried out in a total volume of 20 µL by use, separately, of the following enzymes: BamHI (10 U, 37°C, 16 h), DdeI (10 U, 37°C, 16 h), KpnI (10 U, 37°C, 16 h), and TaqI (10 U, 65°C, 4 h). Each sample was then analyzed by PAGE (6%), and the digested DNA fragments were identified by comparison with the appropriate standards (pUC19 plasmid digested with HaeIII). The wild-type (wt) allele was recognized by the occurrence of restriction sites for the BamHI, DdeI, KpnI, and TaqI endonucleases. Loss of the restriction sites for BamHI, KpnI, and TaqI, and, in the case of the DdeI endonuclease, gain of restriction sites were identified on the basis of the number of electrophoretic gene fragments. The electrophoretic pattern of these fragments provided the basis for identification of the 481C→T, 590G→A, 803A→G, and 857G→A mutations.

Use of the four restriction enzymes allowed identification of the five alleles with differently combined mutations and one wt allele.

Genotypes were recorded according to the identified mutations (e.g. wt/481T allele contains one mutation in position 481 and the other allele is the wt). When two mutated alleles were found, their symbol included all identified mutations. This system of genotype terminology allows the showing of all possible mutation combinations we were able to detect but does not include new allele nomenclature, which can be applied if all known mutations are considered(20).

Statistics

The concentration of particular metabolites in urine served as the basis for calculation of metabolic ratio AFMU:1X, which is characteristic for NAT2 activity(12,14,18). Strength of the connection between age and AFMU/1X index value was determined by Pearson's linear correlation coefficient, with statistical significance at p < 0.05. Statistical significance of the differences in relation to the identified NAT2 genotype (wt/wt and mut/mut) and metabolic ratio AFMU:1X was checked with the Kolmogorow-Smirnov test. Statistical analysis was accomplished with Statistica for Windows v. 5.1.

RESULTS

In all children >20 wk of age, AFMU was detected in various proportions to 1X, which allowed calculation of the AFMU/1X index. The majority of children also eliminated other metabolites, but in 12 (14%) cases, no 1,7 X was detected, and in 22 (26%) cases, no 1U was detected. In 10 (12%) children <20 wk of age, AFMU was not detected, which made calculation of the metabolic index impossible. Thus, in this case, the value of the AFMU/1X was taken as 0.00. For this reason, it was possible to calculate the metabolic index characteristic for the activity of NAT2 in only 72 children. All infants in whom AFMU was not detected were qualified as slow acetylation phenotype.

In infants, the linear relationship between age and the value of metabolic index AFMU/1X was noted (Pearson's χ2 r = 0.39; p < 0.004). However, in children aged >1 y, we did not observed any correlation between coefficient AFMU/1X and age (Pearson's χ2 r = 0.11; p > 0.1).

DNA isolation from 1 mL of blood sample, as described in Methods, allowed us to obtain comparable products of amplification that, in turn, lent themselves to the restriction analysis identifying the mutations under study. Restriction analysis of the amplified fragments of NAT2 gene in the studied children led to the identification of five mutated and one wt alleles, which occurred both in homo- and heterozygotic combinations, forming 21 genotypes. Totally, in 82 children, 164 alleles were identified. There were 71 transitions of 481 (C→T), 51 mutations of 803 (A→G), 33 mutations of 590 (G→A), and 14 mutations of 857 (G→A) among them. Frequency of point mutations in the NAT2 gene, tested among 82 children, is given in Table 1.

Table 1 Frequency of point mutations in the NAT2 gene tested among 82 children
Full size table

The genotype distribution in the studied children are shown in Tables 2 and 3. Genotypes with two mutated alleles were present in 67% of the studied children (Table 3), whereas genotypes with at least one wt allele (genotypically fast acetylators) were identified in 33% (Table 2).

Table 2 Observed frequencies of NAT2 genotypes encoding normal activity of N-acetyltransferase
Full size table
Table 3 Observed frequencies of NAT2 genotypes encoding slow acetylation
Full size table

Statistical analysis did not show significant relation between the genotype of slow and fast acetylation and sex of the studied children (Pearson χ2 r = 0.18, p = 0.67). Among the investigated children, those genotypes contained one or two wt alleles (genotype wt/wt or wt/mutation); the average value of AFMU/1X was 0.83 ± 0.63. Children with genotype containing two mutated alleles (genotype mut/mut) had average value of coefficient AFMU/1X 0.34 ± 0.4. However, those differences were not significant statistically. The relation between genotype coding for slow (mut/mut) and fast acetylation (wt/wt or wt/mut) and AFMU/1X index value in studied children is presented in Figure 1.

Figure 1
figure 1

The relation between genotype coding for slow (mut/mut) and fast acetylation (wt/wt or wt/mut) and AFMU/1X index value (means, SE, SD) in all studied children.

Full size image

Determination of the relation between the acetylation phenotype and genotype revealed that 18 children (23%) containing at least one wt allele had AFMU/1X <0.4 (slow acetylation), and 7 (8%) of genotypically slow acetylators presented high (AFMU/1X >0.4) metabolic ratio (high acetylation activity). The values of AFMU/1X for children of different age when compared with identified genotype are presented in Figure 2. The disagreement between the acetylation phenotype and genotype is more often found in the group of children characterized by low AFMU/1X.

Figure 2
figure 2

Values of metabolic index AFMU/1X and NAT2 genotypes arranged according to age of children. Eighteen (23%) of genotypically fast acetylators had low metabolic ratio AFMU/1X, and 7 (8%) of genotypically slow acetylators presented high metabolic ratio (high acetylation activity).

Full size image

DISCUSSION

The results of the present study comparing the acetylation phenotype with NAT2 genotype reveal new possibilities of studying genetic metabolic defects in children of various ages, including infants.

Our results confirm earlier reports by Pariente-Khayat et al. on the domination of slow acetylation phenotype in children in the first few weeks of life(17). These authors showed that in 100% of children <55 d of age and in 22% of children <121 d of age, molar index was very low (<0.4), and, therefore, they were qualified as slow acetylation phenotype. In the present study, in 10 (12%) children <20 wk of age, AFMU was not detected. Although the metabolic index AFMU/1X could be calculated on the basis of two metabolites, the lack of the other two (1,7 X and 1U) shows that in children during the first year of life, metabolic index characterizing the activity of enzymes cannot be relied upon. The fact that some metabolites are missing in a single portion of urine may be connected with changes in urine flow and urinary pH in children. It should be kept in mind that the urinary elimination of both 1,7 X and caffeine is dependent on urine flow(15). Further, 1,7 X, like caffeine itself, is excreted as a minor metabolite because of its metabolic degradation(21). Moreover, some authors have indicated that xanthine oxidase, responsible for conversion of 1X to 1U, is deficient in approximately 4% of the population. The great variability of 1U excretion was reported in previous studies(11,22).

In the remaining 72 children, the metabolic index AFMU/1X could be calculated. However, in children <1 y of age, a linear dependence between age and the values of metabolic index was observed (r = 0.39, p < 0.004). It should be noted that our studies were not performed prospectively; therefore, we are not able to discuss the maturation of NAT2 activity.

The main purpose of the present study was evaluation of the concordance between phenotype acetylation and genotype encoding NAT2 in children of various ages including infants. We found that the genotype associated with fast acetylation occurs in 33% of children, whereas 67% of these subjects exhibit the slow acetylation genotype. These frequencies are generally in accord with the genotype distribution in the populations from other parts of Europe(5,6).

Our results indicate that in 23% of children with genotype coding for normal activity of NAT2, low values of AFMU/1X have been found. Discordance between the acetylation phenotype and genotype coding for NAT2 seems to be complicated for the following reasons:

  • Low values of AFMU/1X in children with genotypes containing the wt allele may be the result of incomplete expression of NAT2 genotype.

  • There is a faulty assumption that the methods of study of the acetylation phenotype used in adults can be applied in children. It is well known that differences in the pharmacokinetics of drugs in children are caused not only by disturbances in metabolism but also can result from physiologic differences in absorption, distribution, and elimination, which in turn depend on the activity of enzymes, different conditions of blood flow through organs (liver, kidneys), and excretion of drug and its metabolites via the kidneys.

Elimination of caffeine metabolites with urine in a child may depend on many factors and, therefore, should not become the basis for evaluation of the activity of one enzyme. In the first weeks of life, caffeine is eliminated mainly through renal excretion(16). Clearance of the majority of its metabolites, including methylxanthines, is significantly lower until adolescence(23). Therefore, studying the acetylation phenotype according to methods used in adults may give rise to false classification of phenotype. It seems that in children and particularly in infants, because of more frequent urination and limited ability of urine runoff, the identification of all caffeine metabolites in a single portion of urine may not be possible. Additionally, because the diet of infants is based on milk, their urine is often alkaline, and AFMU can be spontaneously deformylated in the bladder to the stable AAMU. According to some reports, the high pH and temperature of urine may cause low stability of 1X and AFMU and, thus, false classification of phenotype. Therefore, some patients may be unduly classified as slow acetylators(24). Tang et al.(25) showed that in case of low values of molar ratio, the probability of classifying the phenotype incorrectly is even greater. Because of this fact, use of the methodology applied in adults may result in a lower number of identified caffeine metabolites in the urine of children. Thus, it may be supposed that calculating metabolic indexes does not allow for evaluation of the genetically determined activity of enzymes.

Cases of high values of the AFMU/1X index in children with genotype including two mutated alleles of NAT2 gene (genotype mut/mut) should be discussed separately. It is known that some mutations are responsible only for decreased activity of the enzyme. Therefore, although the genotype includes two mutated alleles, production of acetyl metabolites of drugs and arylamines is possible(3). Additionally, Cribb et al.(26) showed that in persons with genotype containing two mutated alleles of NAT2 gene, there exists a possibility of producing AFMU with an enzyme of a different substrate specificity. In the present study, in seven children with genotype containing two mutated alleles, the values of AFMU/1X index were >0.4, which qualifies them as fast acetylators. The reason may be low concentrations of 1X that determine high values of AFMU/1X index regardless of AFMU concentration. It is known that in children the activity of the enzymes active in phase I metabolism is higher than in adults(27). When the activity of NAT2 is decreased and the metabolism of methylxanthines is increased, the proportions of these metabolites may differ regardless of the activity of enzymes metabolizing caffeine. It should also be stressed that the activity of cytochrome P4501A2, which is responsible for the production of 1,7 X and 1X, may be different. Polymorphism among genes encoding proteins that govern the expression of the CYP1A2 gene, like the Ah receptor and the Arnt proteins, are responsible for various expression levels(28,29). For this reason, individual differences in the concentration of methylxanthines may be caused by differences in the activity of cytochrome CYP1A2(11,21,22). It may be concluded that is children before adolescence, the acetylation phenotype does not reflect genetically encoded activity of enzymes, as confirmed by our studies.

In studies on adult population, incompatibility of the classification of acetylation phenotype and NAT2 genotype was 4 to 6%(6,30–32). Cascorbi et al.(6), for each case of incompatibility of genotype with phenotype, analyzed individual sequences of NAT2 gene, which confirmed the credibility of genotype studies. Similarly, Mrozikiewicz et al.(32), in their studies on NAT2 in children >3 y of age, analyzed gene sequences in case of incompatibility; their analyses always confirmed the correctness of genotype studies. The method used for the identification of genotypes in the present study can be compared to that used by Mrozikiewicz and Cascorbi. However, in our study, as many as 45% of the children were infants and, therefore, the disproportion between the identified genotype and phenotype of acetylation was observed more frequently.

Our results comparing the acetylation phenotype with NAT2 genotype show that in all children, including infants, evaluation of the genetic acetylation defect is possible. Genotype studies are superior to phenotype evaluation in that regardless of the age, diet, and disease of the patient, 1 or 2 cm3 of blood is sufficient to determine the alleles of several genes at the same time, without model-drugs administration and burden-some urine collection(33). On the basis of NAT2 genotype studies, it is possible to prognosticate hypersensitivity to drugs, the metabolism of which depends on the activity of NAT2. These include drugs such as metamizole, clonazepam, cotrimoxazole, and izoniazid, frequently administered to children, and also others less frequently administered such as procainamide, dapsone, aminofide, etc.(9,34–39). Studies of genotype coding for NAT2 are becoming more popular for the evaluation of individual predisposition to diseases that result from long exposure to environmental arylamines such as cigarette smoke, the metabolism of which depends on the activity of NAT2(40,41). In our earlier studies, we have shown that 91% of children with atopic allergy have genotypes coding for slow acetylation(42). The evaluation of individual NAT2 genotype may play a role in the prognosis of predisposition to allergic diseases.

In conclusion, the presented results confirm earlier reports on the domination of slow acetylation phenotype in infants. We have shown that in children, the in vivo method based on identification of caffeine metabolites could not be used for the evaluation of genetically encoded NAT2.

The differences observed in the pharmacokinetics of caffeine in children show that the only credible method of evaluating the genetically determined acetylation defect is the study of genotype coding for NAT2. In children, NAT2 genotype evaluation should become the basis for determination of individual predisposition to environmental diseases and hypersensitivity to drugs, the metabolism of which depends on the acetylation rate.