Department of Pharmacology and Toxicology (G.J.S., C.A.M.),
University of Arizona, Tucson, Arizona; and Department of
Pharmacology (M.P., E.S.), University of Oxford, Oxford, United
Kingdom
 |
Introduction |
N-Acetyltransferases
(NATs)2 catalyze the acetyl coenzyme A-dependent
N-acetylation of arylamines and hydrazines. The human genes
involved in the N-acetylation of these compounds have been cloned, sequenced, and designated NAT1 and NAT2
(Grant et al., 1991
). It is the latter locus that is responsible for
the acetylation polymorphism observed in humans with drugs such as
isoniazid or sulfamethazine (SMZ). Recent studies show that the slow
acetylator phenotype can be explained by several mutations in the
NAT2 gene (Vatsis et al., 1991; Deguchi, 1992), which result
in decreased intrinsic stability of NAT2 (Blum et al., 1991). NAT1
demonstrates a polymorphism in the N-acetylation of
p-aminosalicylic and p-aminobenzoic acids (PABA;
Vatsis and Weber, 1993; Hughes et al., 1998). The genotype/phenotype
correlation for NAT1 has been subject to controversy in that
NAT1*10 has been associated with higher NAT1 activity in
tissue cytosols (Bell et al., 1995b). However, this has appeared less
marked in other studies (Payton and Sim, 1998; Smelt et al., 1998). In
contrast, NAT1*14 has clearly been associated with low activity (Grant et al., 1997; Doll et al., 1997; Deitz et al., 1997;
Hughes et al., 1998; Butcher et al., 1998; Payton and Sim, 1998).
The heterocyclic arylamine, batracylin (BAT;
8-aminoisoindolo[1,2-b]quinazolin-12(10H)-one),
was shown to be active against a number of solid tumor cell lines. BAT
exhibited antineoplastic activity against early and advanced stage
adenocarcinoma 38 in mice (Plowman et al., 1988), as well as murine
leukemia P-388 cell lines resistant to cisplatin, doxorubicin
(Adriamycin), and methotrexate (Waud et al., 1991). Differences
in the toxicity of BAT were observed in mice, rats, and dogs. Rats were
extremely sensitive to the drug compared with mice, whereas dogs were
relatively insensitive (El-Hawari et al., 1989). Oral administration of
BAT to male rats at a dose equivalent to one-tenth the mouse LD10 was
lethal to all rats treated. Kidney, testicular, and gastrointestinal toxicity was observed in rats receiving one-eightieth the mouse LD10.
Dogs, lacking NAT, could tolerate twice the LD10 of mice. A similar
species difference was seen with BAT genotoxicity where rat hepatocytes
were more sensitive to the genotoxic effects of BAT than mouse cells
(Stevens and McQueen, 1994). These species variations in the toxicity
of BAT may be due to differences in N-acetylation of BAT
because greater susceptibility of rats to the systemic toxicity of BAT
was associated with high plasma concentrations of
N-acetylbatracylin (ABAT; Ames et al., 1991). After
administration of equivalent doses of BAT, the plasma concentration of
ABAT in rats was nine times greater than that of mice (Ames et al.,
1991). A significant correlation was demonstrated between BAT
N-acetylation and mutagenicity in bacteria (Stevens et al.,
1996b). These data suggested that N-acetylation was involved
in the adverse effects of BAT.
The role of NAT in the potential human toxicity of BAT has not been
determined. A number of animal models have been established to study
the relationship between the human NAT polymorphism and toxicity. These
include mice (Glowinski and Weber, 1982), rats (Juberg et al.,
1991), hamsters (Hein et al., 1982a), and rabbits (Hein et al., 1982b).
With the exception of rabbits, the substrate specificity in these
species differed from humans. Substrates such as PABA and
para-aminosalicylic acid, which were
N-acetylated by NAT2 in rodents, were acetylated by NAT1
in humans and the arylamines used to determine human NAT2 phenotype
were often poor substrates in rodents. It appears that the NAT2 isozyme
in mice is orthologous to human NAT1 (Stanley et al., 1997;
Estrada-Rodgers et al., 1998). Species variation in NAT activities was
also noted with heterocyclic amines such as
2-amino-3,8-dimethylimidazo[4,5-f]quinoxaline (Kato, 1986).
Therefore, extrapolating BAT N-acetylation characteristics from rodents to humans must be done cautiously. In the present study
the kinetics of BAT N-acetylation were evaluated in nine human liver samples.
| |
Materials and Methods |
Chemicals.
BAT was obtained from the Pharmaceutical Resources Branch, Division of
Cancer Treatment, National Cancer Institute (Bethesda, MD) or provided
by Dr. Edmond LaVoie (Rutgers University, Piscataway, NJ)
(Meegalla et al., 1994). ABAT was synthesized by the Synthetic Core of the Southwest Environmental Health Sciences Center. HPLC grade
solvents were obtained from Fisher Scientific (Springfield, NJ).
Carnitine acetyltransferase (pigeon breast muscle), acetylcarnitine, PABA, and SMZ were purchased from Sigma Chemical Co. (St. Louis, MO).
Acetylsulfamethazine was obtained from ICN (Costa Mesa, CA) and
4-acetamidobenzoic acid was purchased from Aldrich (Milwaukee, WI).
Acetyl-Coenzyme A was purchased from Boehringer Mannheim (Indianapolis,
IN). All other chemicals used were reagent grade.
Cytosolic Preparation.
Human liver samples, obtained from pathologically normal human donor
tissue, were provided by the Human Cell Culture Center (Folkston, GA).
The liver sections were immediately frozen in liquid nitrogen and
stored at 
80°C until cytosols were prepared. The liver samples were
homogenized in 3 volumes of buffer consisting of 20 mM Tris-HCl, pH
7.5, 1 mM dithiothreitol, 1 mM EDTA, 50 µM phenylmethylsulfonyl
fluoride, and 10 µM leupeptin (Mattano and Weber, 1987). Homogenates
were centrifuged for 10 min at 9000g and the resulting
supernatant was centrifuged at 100,000g for 1 h. The
cytosolic fractions (supernatant) were stored at 80°C before use.
Protein was determined using a Coomassie Plus Reagent from Pierce
(Rockford, IL) with BSA as standard.
NAT Assays.
NAT activity was assayed using an acetylcoenzyme A recycling system
with BAT, SMZ, or PABA as the acceptor amine (Mattano and Weber, 1987).
Samples were incubated in a total volume of 90 µl, containing 15 mM
acetylcarnitine, 2 U/ml carnitine acetyltransferase, 2 mM EDTA, 2 mM
dithioerythritol, 50 mM Tris-HCl pH 7.5, diluted protein, an
appropriate concentration of substrate (dissolved in dimethyl sulfoxide
at a final concentration of 2%) and acetylcoenzyme A (0.5 mM), which
was used to initiate the reaction. Reactions were terminated with the
addition of 90 µl methanol, samples were centrifuged at 14,000 rpm
for 1 min, and a 50-µl aliquot of the supernatant was analyzed by
HPLC. The acetylated products were separated using a Microsorb-MV C18
(4.6 × 250 mm, 5 µM) column (Rainin, Woburn, MA). Separation of
SMZ from its acetylated derivative was performed using a linear
gradient of 98% 20 mM ammonium acetate and 2% methanol ramped to
100% methanol over 30 min at a flow of 1 ml/min. The absorption was
monitored at 254 nm with retention times of 12 and 14 min for SMZ and
acetylsulfamethazine, respectively. PABA samples were separated
under isocratic conditions utilizing 90% 50 mM acetic acid and 10%
acetonitrile at a flow rate of 1.0 ml/min and absorbance monitored at
270 nm. Retention time for PABA was 7.0 and 13 min for acetyl
PABA. The HPLC conditions for the separation of BAT from ABAT
were as described previously (Stevens et al., 1996b). Formation
of acetylated products were measured with reference to standard curves
constructed from authentic acetylarylamines.
Kinetic constants for human liver NAT were determined using substrate
concentrations ranging from 0 to 25 µM BAT. Concentrations greater
than 25 µM were inhibitory. SMZ concentrations ranged from 0 to 1000 µM and PABA concentrations ranged from 0 to 220 µM. PABA
concentrations greater than 220 µM proved to be inhibitory. For each
substrate and individual donor samples were incubated for 10 min at
protein concentrations that resulted in less than 30% conversion of
substrate to acetylated product. Data were evaluated by Eadie-Hofstee
plots (velocity/substrate versus velocity) and Hanes-Woolf plots
(substrate/velocity versus substrate). Kinetic constants were
calculated from the Hanes-Woolf plots. The equation of the best fit
line was determined by linear regression and used to calculate
Km (x-intercept) and
Vmax (reciprocal of the slope).
NAT Genotyping, Cloning, and Sequencing.
Genotyping of NAT1 and NAT2 was as described
previously by Payton and Sim (1998) and by Risch et al. (1995),
respectively. To confirm the identity of the NAT1*11 allele,
which is identical with the NAT1*17 allele (see Table
2, footnote b) (Deitz et al., 1997), cDNA
was amplified from genomic DNA (DiLella and Woo, 1987) using high
fidelity proof reading PFU DNA polymerase (Stratagene, Inc., La Jolla,
CA). Amplification was performed using the primers N-376(5'-TAT TGC ATG
ATT CTC CTG CCT A-3') and N-1117 (5' GGA ATT CAA CAA TAA ACC AAC AT-3')
(Payton and Sim, 1998). Primers were annealed at 58°C for 30 s
followed by elongation at 72°C for 3 min to yield a product of the
expected size 1543bp. Amplified DNA was tagged at the 3' end with dATP
and cloned into the vector pGEMT (Promega, Madison, WI). Four
representative clones containing the entire cDNA product were analyzed
on an ABI 377 automated sequencer in both the forward and reverse
direction (Department of Biochemistry DNA sequencing service,
University of Oxford) using primers to M13. The sequence was confirmed
by primer walking.
Statistics.
Student's t tests were performed to determine significance
between phenotypes in NAT activity. Correlation of BAT and SMZ acetylation rate was determined using linear regression. Data were
considered significantly different at p < .05.
| |
Results |
Liver samples obtained from nine Caucasian donors were analyzed to
determine NAT1 and NAT2 genotypes (Table
2). Five samples were homozygous for
NAT1*4. The remaining four samples were heterozygous with
each containing one NAT1*10 allele. One individual with
NAT1*10 also had the NAT1*11 (*17)
allele (Deitz et al., 1997). The assignment of these alleles was
confirmed by sequencing from nucleotide 376 to 1117. Only one sample
(HL-3) was homozygous for the NAT2*4 allele, which is
associated with the rapid acetylator phenotype. Samples HL-1 and -5 were heterozygous for this allele. Six samples had some combination of
the NAT2*5A, B, C, or
6A alleles, which results in decreased acetylation (Vatsis
et al., 1995).
NAT activities were determined using BAT, SMZ, or PABA as the arylamine
substrate (Table 3). The NAT activity for
SMZ and NAT2 genotype was used to classify the donor samples
as rapid or slow acetylator phenotype. Assays were performed with BAT, SMZ, and PABA to assess the effects of time and protein on NAT activity
(data not shown). Reactions were linear up to 60 min and protein
concentrations as high as 2 mg/ml. Assays for determination of kinetic
parameters were performed for 10 min at protein concentrations that
resulted in less than a 30% conversion of substrate to the acetylamine.
N-Acetylation of BAT by human cytosol was measured at substrate
concentrations ranging from 0 to 22 µM. The hyperbolic curves were
linearized to create Eadie-Hofstee plots. Samples HL1-3 are shown in
Fig. 1. Plots constructed from all nine
samples were linear. Linear Eadie-Hofstee plots were also seen with SMZ
and PABA (data not shown). Hanes-Woolf transformations were performed and used to calculate kinetic parameters and representative samples are
illustrated in Fig. 2. A 3- to 5-fold
variation in apparent Km values was
observed but there were no significant differences in apparent
Km values for any of the substrates between
rapid and slow acetylators (Table 3). No significant substrate
correlations (p > .05) were observed based on
apparent Km (SMZ and BAT,
r2 = 0.20; PABA and BAT,
r2 = 0.45). As previously noted with human
liver (Kilbane et al., 1991), the apparent
Vmax of SMZ was significantly greater in
the rapid phenotype compared with slow (Table 3), with a similar difference being observed with BAT. Variability in the apparent Vmax for PABA acetylation was not
SMZ-phenotype-dependent. The apparent intrinsic clearance
(Vmax/Km) of
BAT was compared to the substrates preferentially acetylated by either
NAT1 (PABA) or NAT2 (SMZ) (Fig. 3). A
correlation was found between BAT and SMZ (r2 = 0.97, p < .001), but not between PABA and BAT
(r2 = 0.02). The
Vmax correlation between BAT and SMZ was
r2 = 0.80 (p < .005) and r2 = 0.34 for PABA and BAT.
View larger version (15K):
[in this window]
[in a new window]
|
Fig. 1.
Eadie-Hofstee transformation of BAT
acetylation in HL1-3.
The velocity of the reaction was determined at various concentrations
of BAT. Velocity divided by substrate concentration was plotted against
velocity. Data represent the mean of duplicate samples.
|
|
View larger version (12K):
[in this window]
[in a new window]
|
Fig. 3.
The intrinsic clearance for BAT, SMZ, and
PABA.
Using linear regression analysis a significant correlation
(r2 = 0.97) was observed between BAT and SMZ
(p < .001). No significant correlation was
observed between BAT and PABA.
|
|
Inhibition of BAT N-acetylation was monitored in the
presence of either 250 µM SMZ or 250 µM PABA (Fig.
4). A greater than 65% inhibition of BAT
N-acetylation was observed with SMZ in the rapid acetylators
(HL-1 and HL-3), whereas only a 20 to 30% inhibition was observed in
the slow acetylator samples (HL-2 and HL-4). PABA had less of an effect
on BAT N-acetylation in the four samples.
View larger version (23K):
[in this window]
[in a new window]
|
Fig. 4.
Inhibition of BAT acetylation by SMZ and
PABA.
Assays were performed at 0.25 mg/ml cytosolic protein for 10 min in
HL1-4 using 44 µM BAT and 250 µM SMZ or PABA. The percent
inhibition is in parentheses. Data represent the mean of duplicate
samples.
|
|
NAT1 or NAT2 genotypes were compared to NAT
activities with either PABA or SMZ. Classification of NAT2
phenotype based on SMZ NAT activity correlated with genotype (Table
4). Mean NAT activities for individuals
containing at least one NAT2*4 allele were 4.21 ± 2.05 and 8.81 ± 0.69 nmol/min/mg for SMZ and BAT, respectively.
Samples with one of the two alleles previously associated with
decreased NAT2 activity had values of 0.54 ± 0.47 nmol/min/mg (SMZ) and 1.40 ± 0.81 nmol/min/mg (BAT). NAT1 activity
was measured by N-acetylation of PABA. Sample HL-7 was
higher than those having either one or no NAT1*10 alleles
(Table 5). Genotyping revealed that HL-7
was NAT1*10/*11.
| |
Discussion |
The role of N-acetylation in the bioactivation of
aromatic and heterocyclic amines is species- and substrate-dependent.
An early report demonstrated that N-acetylation of
2-aminofluorene (2-AF) and other aromatic amines was greatest in
hamsters, followed by guinea pig, mouse, and rat liver cytosol (Lower
and Bryan, 1973). A similar species difference was observed with
heterocyclic aromatic amines derived from protein pyrolysates
(Shinohara et al., 1986; Kato, 1986). The lowest NAT activities were
found in rat liver with either 2-AF or
3-amino-1,4-dimethyl-5H-pyrido[4,3-b]indole. In
other species, the N-acetylation of
3-amino-1,4-dimethyl-5H-pyrido[4,3-b]indole and
other food-derived heterocyclic amines was also less than that of 2-AF
although activities varied depending on substrate. Although
N-acetylation of heterocyclic amines, including
2-amino-3,8-dimethylimidazo[4,5-f]quinoxaline or
2-amino-1-methyl-6-phenylimidazo[4,5-b]pyridine can occur, formation of the N-acetylated derivative did not appear to be a major
route of bioactivation, but rather O-acetylation
predominated (Davis et al., 1993; Kato and Yamazoe, 1994; Malfatti et
al., 1994). Both of these acetylation reactions as well as the
intramolecular N,O acetyltransfer are catalyzed by the
proteins coded by NAT1 and NAT2 (Glowinski et
al., 1980; Mattano et al., 1989; Hein et al., 1994). Although NAT
activity for many substrates such as 2-AF is higher in mouse than rat,
this is not the case with BAT. In vivo, rats had higher plasma
concentrations of ABAT than mouse (Ames et al., 1991). Hepatic
N-acetylation of BAT was also greater in rat than mouse
(Stevens et al., 1996a).
There are few studies evaluating the N-acetylation of
heterocyclic amines in humans. Given that N-acetylation of
many of the heterocyclic amine protein pyrolysates is dependent on
substrate and species, BAT N-acetylation was measured in
human liver of known genotypes and compared to known substrates of NAT1
and NAT2. Phenotypically, there were two distinct groups based on
differences in the apparent maximum velocities of the human NAT2
selective substrate, SMZ (Table 4). The rapid phenotype had a mean
apparent Vmax that was almost 8-fold
greater than the slow phenotype. The phenotypic designation of samples
was confirmed by analysis of NAT2 genotype. All samples
considered to have a rapid phenotype had at least one copy of the
wild-type NAT2*4 allele. Samples classified as slow acetylators contain
two copies of alleles associated with decreased NAT2 activity (Vatsis
et al., 1995).
NAT activity for PABA, an NAT1 selective substrate, was also determined
for each sample. The apparent Vmax ranged from a
low of 0.25 to 11.7 nmol/min/mg (Table 5). Genotypic analysis of the
samples revealed that five samples were homozygous for the wild-type
NAT1*4. The remaining four samples were heterozygous for the
NAT1*10 allele. This allele was associated with an
approximately 2-fold increase in PABA NAT activity in bladder and colon
(Bell et al., 1995a), but others have reported less of an association (Payton and Sim, 1998; Smelt et al., 1998). Although a trend toward higher NAT1 activity was observed in whole blood lysates when NAT1*4/NAT1*10 heterozygotes were compared with
NAT1*4 homozygotes (Payton and Sim, 1998), there was no
significant correlation between the number of NAT1*10
alleles and NAT1 specific activity in a study of placental homogenates
(Smelt et al., 1998). It is unclear whether these differences lie in
the genotyping methodologies used in these studies (Bell et al.,
1995a,b compared with Payton and Sim, 1998) or whether the sources of
the human tissues cannot be compared directly. In the present study
using liver cytosols, there was no statistically significant difference
when the samples with one NAT1*10 allele were compared with
the samples with no NAT1*10 allele (Table 5). However, one
sample, HL-7, had a much higher activity than any other sample. This
individual HL-7 was genotyped as
NAT1*10/*11(*17). The association of a
high NAT1 activity with this allele has been suggested from in vitro
expression studies (Doll et al., 1997). The high level of NAT1 activity
in HL-7 may be a result of the combination of alleles in this
individual. A previous study (Risch et al., 1996) found that the
NAT1*11 allele, identified by genotyping, not by sequencing,
was present with NAT1*4 in an individual with a low level of
NAT1 activity in erythrocytes. The relationship between genotype and
phenotype of NAT1 requires further investigation before it can be fully
understood. The levels of NAT1 and NAT2 activities are similar in
liver. This is in contrast to gut where the NAT1 activity is much
higher than the NAT2 activity (Hickman et al., 1998).
BAT had high affinity for human NAT with an apparent
Km 50-fold lower than that for SMZ and
5-fold lower than that for PABA. The kinetic parameters for these later
substrates were similar to previous reports from human tissue (Grant et
al., 1991; Kilbane et al., 1991; Derewlany et al., 1994). Although
estimated apparent Km values for SMZ and
PABA did not differ between phenotypes (Table 3), a significant
difference in the apparent Vmax of SMZ
acetylation was observed. These results were similar to previous
reports attributing the phenotypic variation in the rate of SMZ
acetylation to differences in apparent Vmax
of the reaction (Kilbane et al., 1991). There was an 8-fold difference
in the mean Vmax for rapid and slow
phenotypes for BAT and SMZ (Table 3), with a significant correlation
between their apparent maximum velocities (r2 = 0.80) or apparent intrinsic clearance
(Vmax/Km)
(r2 = 0.97) values. The mean apparent intrinsic
activity for BAT was 1.72 ± 1.20 ml/min/mg for rapid acetylators
and 0.27 ± 0.10 ml/min/mg for slow acetylators (Table 3) and are
100-fold higher than those calculated for SMZ. Thus, the apparent
clearance of BAT by human NAT was much faster than SMZ.
A previous study in bacteria-expressing human NAT suggested
that N-acetylation of BAT was catalyzed by both NAT1 and
NAT2 (Stevens et al., 1996b). Because the level of expression of these genes in Salmonella typhimurium was not evaluated, the
isoform specificity of BAT was unclear. However, when human liver
cytosol was examined, BAT NAT activity correlated only with SMZ NAT
activity, a selective substrate of NAT2. In addition, BAT
N-acetylation was preferentially inhibited by SMZ (Fig. 4)
and the Eadie-Hofstee plots were consistent with the involvement of a
single enzyme. Consequently, it appears that BAT was predominately
acetylated by NAT2. A more extensive population study of BAT
N-acetylation may reveal a greater range of activities than
presented here. Nevertheless, these results are consistent with the
hypothesis that NAT activity contributes to the toxicity of BAT and
suggest that patients who are rapid acetylators would be at greater
risk of developing drug-induced toxicity than slow acetylators.
Human liver samples were generously provided by Brent Bardsley,
Human Cell Culture Center, Folkston, Georgia. G.J.S. was supported by
National Institutes of Health Training Grant ES07091 and a Procter and
Gamble predoctoral fellowship. Studies were supported by Grant ES-05174
(C.A.M.), Center Grant P30 ES06694 and the Wellcome Trust (E.S. and
M.P.). Presented in part at the 35th Annual Meeting of the Society of
Toxicology, Anaheim, California, March 1996.
![[Feedback]](/icons/banner/feedbackACT.gif){kind=icon}
![[For Subscribers]](/icons/banner/subscriberACT.gif){kind=icon}
![[Archive]](/icons/banner/archiveACT.gif){kind=icon}
![[Search]](/icons/banner/searchACT.gif){kind=icon}
![[Contents]](/icons/banner/tocACT.gif){kind=icon}
Top
Abstract
Introduction
), HL-2 (
), and HL-3 (
) were constructed
using linear regression and used to calculate
Km (x-intercept) and
Vmax (1/slope).