Introduction

Abstract Introduction Results Discussion References

Hepatic N- and ring-oxidations are important activation and inactivation steps in aromatic amine-induced toxicities. NADPH-dependent oxidation of N-acetylbenzidine occurs with rat and mouse liver microsomes (1). In mouse, the formation of NHA¹ was 4 times greater than NHA. In rat, the formation of both products was about equal. Ring-oxidation products were considered minor and were not quantitated. The N-oxidation of N,N-diacetylbenzidine to NHDA was much less than N-acetylbenzidine in both species. The specific oxidative pathways involved in the metabolism of these compounds were not determined. In both humans and rodents, benzidine is rapidly acetylated to N-acetylbenzidine and N,N-diacetylbenzidine (2-4). Workers exposed to high levels of benzidine have as much as a 100-fold increased risk for bladder cancer (5). Benzidine causes bladder cancer in dogs and liver cancer in rats (6, 7).

Rats, mice, or hamsters administered either benzidine or N-acetylbenzidine exhibit only a single adduct in liver, N-(deoxyguanosin-8-yl)-N-acetylbenzidine (8, 9). This is also the major adduct detected in rat liver after administration of N,N-diacetylbenzidine (9) and in exfoliated bladder cells from workers exposed to benzidine (10). One possible mechanism of N-(deoxyguanosin-8-yl)-N-acetylbenzidine formation may involve cytochrome P450 oxidation of N-acetylbenzidine to NHA with subsequent O-acetylation to its unstable N-acetoxy ester that reacts with DNA to form the adduct. Alternatively, N,N-diacetylbenzidine may be oxidized to NHDA and undergo intramolecular N,O-acetyltransfer to form this adduct (11, 12).

Understanding specific enzymes and the sequential steps involved in metabolism of chemicals are important for assessing chemical risk and preventing carcinogenicity. For example, by using specific cytochrome P450 inducers and inhibitors, aflatoxin B₁ was demonstrated to undergo metabolic activation in rat by cytochrome P4503A2 and P4502C11 to aflatoxin-8,9-oxide (13, 14). Because pretreatment of rats with -naphthoflavone, a specific P4501A1/1A2 inducer, dramatically increases oxidative metabolism of 4-aminobiphenyl (15), it was hypothesized that similarly treated rats might also exhibit increased oxidation of N-acetylbenzidine and N,N-diacetylbenzidine. The present study used specific inhibitors of cytochrome P450s and the specific inducer -naphthoflavone to elucidate rat oxidative metabolism of N-acetylbenzidine and N,N-diacetylbenzidine.

Materials and Methods

[³H]Benzidine (189 mCi/mmol) was purchased from Chemsyn (Lenexa, KS). Benzidine free base and hydrochloride salt, NADPH, -naphthoflavone, diethylenetriaminepentaacetic acid, ascorbic acid, and EDTA were bought from Sigma Chemical Co. (St. Louis, MO). DPEA was a gift from Eli Lilly Laboratories (Indianapolis, IN). Furafylline was purchased from Gentest Corp. (Woburn, MA), whereas the other cytochrome P450 inhibitors were bought from Sigma Chemical Co. N-Acetylbenzidine was synthesized by acetylation of benzidine in glacial acetic acid and N,N-diacetylbenzidine was synthesized by acetylation of benzidine with acetic anhydride as previously described (16). NHA, NHA, and NHDA were synthesized by Dr. Shu Wen Li (Department of Biochemistry, St. Louis University Medical School, St. Louis, MO). 4-Amino-4-nitrobiphenyl (TCI America, Portland, OR) was acetylated to 4-acetamido-4-nitrobiphenyl. Conversion of this product to NHA or NHDA was accomplished using a modification of a published procedure (17). NHA was synthesized from 4-trifluroacetamido-4-nitrobiphenyl that was prepared from 4-amino-4-nitrobiphenyl by trifluroacetylation. After conversion of the nitro group of 4-trifluroacetamido-4-nitrobiphenyl to N-hydroxy-N-acetamido, the latter compound was hydrolyzed to remove the trifluroacetyl group forming NHA. The identity and purity (>95%) of these synthetic compounds were established by TLC, NMR, and MS. Ring-oxidation products were prepared from 3-OH-benzidine, which was characterized by NMR and MS (18). To prepare 3-OH-N,N-diacetylbenzidine, 3-OH-benzidine was acetylated in pyridine:acetic anhydride, purified, and identified by MS (11). Ring-oxidation products of N-acetylbenzidine were prepared by treatment of 3-OH-N,N-diacetylbenzidine with porcine liver carboxylesterase (Sigma Chemical Co.). Alternatively, ring-oxidation products of N-acetylbenzidine were prepared by direct chemical synthesis from 4-amino-4-nitrobiphenyl after reaction with potassium persulfate. Male Fischer 344 rats (150-200 g) were purchased from Harlan Industries (Indianapolis, IN).

NADPH-Dependent Oxidation. Microsomes were prepared from control or -naphthoflavone-treated rats (19). Treated rats received 40 mg/kg -naphthoflavone in corn oil ip, once a day for three consecutive days (20). Control rats received only corn oil. Animals were euthanized by CO₂ asphyxiation and open thoracotomy. Microsomes were preincubated for 5 min at 37°C in 100 mM sodium phosphate buffer (pH 7.4) containing 5 mM MgCl₂, 1 mM diethylenetriaminepentaacetic acid, 1 mM NADPH, and 3 mM ascorbic acid. Reactions were started with the addition of [³H]arylamine (0.06 mM final concentration) and incubated (final volume of 0.1 ml) for 30 min at 37°C. Reactions were linear with respect to protein concentration (0.5-1.5 mg/ml for control and 0.25-1 mg/ml for -naphthoflavone-treated) and time (15-45 min). Incubation was stopped with an equal volume of methanol containing 2 mM ascorbic acid, kept at 4°C for 30 min, and centrifuged to remove precipitated protein. Supernatants were analyzed by HPLC. Test agents were dissolved in dimethylsulfoxide, with the final concentration of 2% in the reaction mixture. Inclusion of paraoxon, a deacetylase inhibitor, in the incubation mixture did not effect metabolism.

Identification of Metabolites. The characteristic response of each of these compounds to pH was used in solvent extraction purification and in their identification by HPLC. To a 0.8 ml incubation mixture, cold standards (0.125-0.25 mM) and 0.08 ml of 1.0 N NaOH were added, and immediately extracted twice with 1.6 ml of CHCl₃ saturated with 0.1 N NaOH. The aqueous fraction was quickly adjusted to pH 7.4 with 0.08 ml of 1.0 N HCl, and extracted twice with 1.6 ml of ethyl acetate saturated with pH 7.4, 100 mM phosphate buffer. NHA was extracted into CHCl₃, whereas NHDA, NHA, and ring-oxidation products were extracted into ethyl acetate. Recovery of each compound was judged to be at least 75% by analysis of the UV peak of synthetic standard on HPLC. For N-acetylbenzidine, further HPLC analysis was performed using the following solvent system: mobile phase contained 32% methanol:68% of 50 mM citrate/phosphate buffer (pH 4.0), 0-9 min; 32-34%, 9-11 min; 34-60%, 14-19 min; and flow rate, 1 ml/min. Elution times of the ring-oxidation products, NHA, N-acetylbenzidine, and NHA, were 6.8 and 7.4, 8.8, 10.5, and 12.7 min, respectively. For N,N-diacetylbenzidine, further HPLC analysis was performed using the following solvent system: mobile phase contained 20% methanol:80% of 20 mM phosphate buffer (pH 8.2), 0-2 min; 20-38%, 2-15 min; 38-80%, 25-32 min; and flow rate, 1 ml/min. Elution times of the ring-oxidation product, NHDA, and N,N-diacetylbenzidine were 20, 23.7, and 24.6 min, respectively. Because the UV peaks of the cold synthetic standards corresponded to the radioactive peaks of the proposed metabolites on at least two different HPLC systems, the radiolabeled microsomal incubation products are assumed to be the assigned compounds.

Analysis of Metabolite Binding to DNA and DNA Adducts. Samples were treated with RNase followed by sodium dodecyl sulfate and proteinase K. After extraction with phenol and then chloroform:isoamyl alcohol (24:1) (21), the aqueous fraction was adjusted to a final concentration of 0.25 M NaCl, cold ethanol added, and nucleic acid precipitated at 20°C overnight. Pellets were washed with 70% ethanol and dissolved in H₂O. Purity of DNA was determined by absorbance at 260 and 280 nm, with a ratio of A₂₆₀/A₂₈₀ of ~1.7 achieved for each sample. The radioactivity bound to DNA was determined.

	Results

Abstract Introduction Results Discussion References

-Naphthoflavone treatment resulted in a substantial increase in [³H]N-acetylbenzidine metabolism (table 1). The rate of NHA formation was increased 8-fold, and the formation of ring oxidation products was demonstrated. In contrast, NHA formation was at the limit of detection (1% of total HPLC recovered radioactivity) with -naphthoflavone-treated microsomes, but was a major metabolite with control microsomes. For control, more NHA was produced (49 ± 6 pmol/mg protein/min) than NHA (38 ± 5). Ring-oxidation products were not detected with control microsomes. These results are illustrated in fig. 1. In fig. 1A, with -naphthoflavone-treated microsomes, NHA eluted after N-acetylbenzidine with the ring-oxidation products eluting in a single early peak. NHA was at the limit of detection. NHA was observed as a major oxidative product with control microsomes along with NHA (fig. 1B). In each experiment, blank incubations without NADPH yielded no products of metabolism (fig. 1C). Cold NHA and NHA (~10 nmol of each) were coinjected on the HPLC to allow recovery of the radioactive NHA and NHA peaks. The UV peak derived from the cold synthetic standards corresponded to the radioactive peaks. The identity of these metabolites was verified by comparison with synthetic standards after partial purification from the reaction mixture and HPLC on at least two different solvent systems (see details in Materials and Methods).

View larger version (12K): [in this window] [in a new window]   Fig. 1.   Rat liver microsomal NADPH-dependent oxidation of [3H]N-acetylbenzidine (ABZ). (A-C) [3H]N-acetylbenzidine metabolism by -naphthoflavone-treated microsomes, control microsomes, and -naphthoflavone-treated microsomes without NADPH, respectively. Ring Ox., N-acetylbenzidine ring-oxidation products.

Although the metabolism of [³H]N,N-diacetylbenzidine was not detected with control microsomes, -naphthoflavone treatment resulted in significant metabolism (table 1). The formation of NHDA was more than twice that observed for the ring-oxidation product. The HPLC profile observed with -naphthoflavone-treated microsomes is illustrated in fig. 2A. Blank incubations without NADPH yielded no products of metabolism (fig. 2B). Cold NHDA (~10 nmol) was coinjected on the HPLC to allow recovery of the radioactive NHDA peak. The UV peak derived from the cold synthetic standards corresponded to the radioactive peaks. The identity of the proposed radioactive metabolites was verified by comparison to synthetic standards after HPLC on at least two different solvent systems of partially purified material from reaction mixtures (see details in Materials and Methods).

View larger version (14K): [in this window] [in a new window]   Fig. 2.   Rat liver microsomal NADPH-dependent oxidation of [3H]N,N-diacetylbenzidine (DABZ). (A and B) [3H]N,N-diacetylbenzidine metabolism by -naphthoflavone-treated microsomes with and without NADPH, respectively. Ring Ox., N-acetylbenzidine ring-oxidation products.

To assess the specific oxidative pathways metabolizing [³H]N-acetylbenzidine and [³H]N,N-diacetylbenzidine in -naphthoflavone-treated microsomes, specific cytochrome P450 inhibitors were used (25, 26) (table 2). Because -naphthoflavone is known to induce rat hepatic cytochrome P4501A1/1A2, a specific inhibitor of this family was examined for possible dose-response inhibition. Ellipticine elicited a dose-response inhibition of NHA and ring-oxidation product formation with an IC₅₀ of ~0.3 µM and complete inhibition at 1 µM (fig. 3). A similar dose-response inhibition was observed with ellipticine for [³H]N,N-diacetylbenzidine metabolism. For this reason, all inhibitors were tested at 10 µM (table 2). At this concentration, both -naphthoflavone and ellipticine, specific cytochrome P4501A1/1A2 inhibitors, elicited complete inhibition of metabolism of [³H]N-acetylbenzidine and [³H]N,N-diacetylbenzidine. A small amount of inhibition might have been detected with the cytochrome P4502E1 inhibitor, disulfiram (26). Inhibitors of cytochrome P4503A1/3A2 (troleandomycin) and P4502C6 (sulfinpyrazone) did not alter metabolism (26). In addition, yohimbine and ajmaline, inhibitors of human cytochrome P4502D6, did not alter metabolism.

View larger version (15K): [in this window] [in a new window]   Fig. 3.   Ellipticine dose-response inhibition of [3H]N-acetylbenzidine oxidation by -naphthoflavone-treated microsomes. Ring Ox., N-acetylbenzidine ring-oxidation products.

Additional test agents and conditions were used to evaluate [³H]N-acetylbenzidine metabolism by control microsomes (table 3). Consistent with metabolism by -naphthoflavone-treated microsomes, -naphthoflavone and ellipticine, specific cytochrome P4501A1/1A2 inhibitors (25, 26), completely prevented formation of NHA. In contrast, NHA formation was not prevented by these compounds or any of the other cytochrome P450 inhibitors tested in table 2. Furafylline, a specific inhibitor of cytochrome P4501A2 (27), was also tested and reduced NHA formation to 53% of the complete reaction mixture, while not affecting NHA formation. Because previous studies have demonstrated flavin-dependent monooxygenase metabolism of 2-aminofluorene (28), this possibility was assessed. Methimazole (1 mM) and heat treatment (37°C for 60 min), which inhibit this monooxygenase, were evaluated. Neither of these conditions effected NHA formation to a greater extent than NHA (table 3). The nonspecific cytochrome P450 inhibitor SKF-525A was shown to inhibit 2-aminofluorene metabolism (28) and was shown to exhibit partial inhibition (41% of complete reaction mixture) of NHA formation at 10 µM. Whereas higher concentrations of SKF-525A (100 µM) did not cause more inhibition, 1 µM was much less effective (82% of complete reaction mixture). Thus, 10 µM SKF-525A elicited a partial dose-response inhibition of NHA formation, but did not inhibit NHA. The nonspecific cytochrome P450 inhibitor DPEA at 10 µM inhibited both NHA (49% of complete reaction mixture) and NHA (29%) formation by control microsomes (table 3). At 100 µM, DPEA completely inhibited both NHA and NHA formation.

To assess direct activation of [³H]N-acetylbenzidine and [³H]N,N-diacetylbenzidine by rat liver microsomes, their binding to DNA during the incubation was assessed (table 4). For [³H]N-acetylbenzidine, a 24-fold increase in binding was observed with -naphthoflavone-treated microsomes. No increase in binding to DNA was observed for N,N-diacetylbenzidine. At 10 µM ellipticine, binding to DNA by both compounds was prevented with 30-100% and 60-100% inhibition observed using control and -naphthoflavone-treated microsomes, respectively.

To determine if N-(3-monophospho-deoxyguanosin-8-yl)-N-acetylbenzidine was formed, ³²P-postlabeling was performed on the samples illustrated in table 4. The amount of adduct detected in control and -naphthoflavone-treated microsomes incubated with [³H]N-acetylbenzidine was 190 and 14 fmol/mg DNA, respectively. This adduct was not detected in either control and -naphthoflavone-treated microsomes incubated with [³H]N,N-diacetylbenzidine.

	Discussion

Abstract Introduction Results Discussion References

This is the first study to assess the cytochrome P450s that oxidize N-acetylbenzidine and N,N-diacetylbenzidine in rat control and -naphthoflavone-treated liver microsomes. A previous study had assessed metabolism of these compounds by control Sprague-Dawley rat liver microsomes (1). In that study, the rates of NHA and NHA formation were 0.13 ± 0.03 and 0.11 ± 0.01 nmol/mg/min, respectively, with NHDA formation over 20-fold less (0.005 ± 0.001). These previous values for NHA and NHA are about 2-3 times more than that reported for control microsomes in our study (table 1). In both studies, the relative rates of formation are NHA > NHA  NHDA, with ring-oxidation at the limit of detection. Thus, the results were qualitatively similar in both studies, with quantitative differences attributed to strain differences in the rats used.

The cytochrome P4501A family has been shown to oxidize a variety of carcinogenic aromatic and heterocyclic amines, including 4-aminobiphenyl (29-31). -Naphthoflavone is a specific inducer for this family (20) and increased the oxidation of both N-acetylbenzidine and N,N-diacetylbenzidine (table 1). In addition, both -naphthoflavone and ellipticine, specific cytochrome P4501A family inhibitors (25, 26), elicited complete inhibition of oxidation of these acetylated benzidine analogs. Thus, the P4501A family seems to be responsible for oxidation by -naphthoflavone-treated liver microsomes. Although -naphthoflavone and ellipticine were also effective in preventing NHA formation by control microsomes, neither compound inhibited formation of NHA. Although these chemicals inhibit both cytochrome P4501A1 and P4501A2, furafylline, a relatively selective, irreversible/mechanism-based inhibitor of P4501A2 (26, 27, 32), was also tested for inhibition of NHA formation. Furafylline was ineffective in inhibiting NHA at 10 µM and only caused a partial inhibition of NHA formation by control microsomes, thus suggesting greater involvement of cytochrome P4501A1 than P4501A2 in NHA formation. The nonspecific cytochrome P450 inhibitors SKF-525A and DPEA have been previously shown to prevent aromatic amine oxidation (28, 29, 33) and elicited partial inhibition of NHA formation by control microsomes at 10 µM (table 1). At 100 µM, DPEA completely inhibited NHA formation. In control microsomes, different cytochrome P450s seem to be eliciting the formation of NHA and NHA. Conditions that inhibit flavin-dependent monooxygenase metabolism did not seem to alter significantly NHA or NHA formation. Therefore, whereas the P4501A family contributes to the oxidation of NHA by control microsomes, the cytochrome P450 responsible for NHA formation by control microsomes was not determined.

Numerous studies have suggested that activation of aromatic amines to bind DNA involves a number of metabolic and distribution steps (34, 35). After hepatic N-oxidation, the N-OH analogs are thought to undergo N-glucuronidation, excretion, and accumulation in the lumen of the bladder. The acid lability of these N-glucuronides results in their hydrolysis to the corresponding N-OH arylamine that can form DNA adducts in the bladder resulting in tumor formation (35, 36). Because the amount of N-OH products formed in liver may have a direct impact on bladder carcinogenicity, the proportion of ring vs. N-oxidized products is an important oxidative parameter. For control microsomes, only N-oxidation products are formed with N-acetylbenzidine. For -naphthoflavone-treated liver microsomes, N-oxidation is 3.6-fold and 2.3-fold greater than ring-oxidation for N-acetylbenzidine and N,N-diacetylbenzidine, respectively. Thus, under all conditions, N-oxidation exceeds ring-oxidation. Similar results were observed for 4-aminobiphenyl with 10-fold more N-oxidation than ring-oxidation products (37). In contrast, ring-oxidation of 2-amino--carboline exceeds N-oxidation by 5.7-fold (38).

DNA binding was used as an index of reactivity of microsomal oxidation products. Although -naphthoflavone treatment increased N-acetylbenzidine binding to DNA, no increase was detected for N,N-diacetylbenzidine (table 4). Binding was attributed to the N-OH products of metabolism (34, 39). Binding data are consistent with the greater N-oxidation, compared with ring-oxidation observed for N-acetylbenzidine than N,N-diacetylbenzidine, and with the greater stability of the N-OH analog of the latter compared with the former. That is, N-acetoxy-N,N-diacetylbenzidine reacts only slowly with 2-deoxyguanosine at pH 7.0, requiring 6 hr for significant adduct formation (24).

N-(3-monophospho-deoxyguanosin-8-yl)-N-Acetylbenzidine is observed in livers from rats, mice, and hamsters administered benzidine, N-acetylbenzidine, or N,N-diacetylbenzidine (8, 9, 24), and in exfoliated bladder cells from workers exposed to benzidine (10). The formation of this adduct was assessed with incubation conditions used to assess metabolism. Control microsomes incubated with N-acetylbenzidine exhibited a 14-fold higher content of N-(3-monophospho-deoxyguanosin-8-yl)-N-acetylbenzidine than microsomes from -naphthoflavone-treated rats. This result is consistent with metabolism studies, because NHA was only detected during N-acetylbenzidine metabolism by control microsomes, and NHA can react with DNA to form this adduct directly. Using the same incubation conditions, neither NHA nor NHDA would be expected to form this adduct.

N-Acetylation of aromatic amines makes them harder to oxidize and is considered a detoxification step (40). For the diamine benzidine, N-acetylbenzidine is considered a more active metabolite of benzidine than N,N-diacetylbenzidine, which has been considered a more detoxified metabolite (8, 9, 41). Results with control microsomes are consistent with this hypothesis (table 1). That is, substantial metabolism of N-acetylbenzidine is observed whereas metabolism of N,N-diacetylbenzidine is not detected. In addition, the NHA produced by control microsomes (table 1) is probably responsible for the N-(3-monophospho-deoxyguanosin-8-yl)-N-acetylbenzidine detected with this preparation (8-10). However, because control animals administered benzidine exhibit rapid acetylation to N,N-diacetylbenzidine with little accumulation of benzidine or N-acetylbenzidine (2-4, 24), N-acetylation would seem to result in detoxification in control animals. After -naphthoflavone treatment, NHA is not formed, and the metabolism of N,N-diacetylbenzidine exceeds N-acetylbenzidine (table 1). N,N-Diacetylbenzidine seems to be a better substrate than N-acetylbenzidine for the oxidizing enzyme(s) induced by -naphthoflavone treatment. In these induced rats, N-acetylation may not be a detoxification step.

A recent study has demonstrated 3-OH-benzidine as the major product of cytochrome P450 oxidation of benzidine (18). The rate of 3-OH-benzidine formation by control microsomes (68 ± 16 pmol/mg/min) is increased 12.3-fold after -naphthoflavone treatment (835 ± 81). Similar rates were reported for N-acetylbenzidine (table 1). Inhibitor studies demonstrated that cytochrome P4501A1/1A2 was responsible for benzidine metabolism. Thus, the oxidation of benzidine by rat liver microsomes is comparable with that reported herein for N-acetylbenzidine and N,N-diacetylbenzidine.

The constitutive level of cytochrome P4501A2 in control rat liver microsomes is quite low and that for cytochrome P4501A1 even lower (42, 43). -Naphthoflavone treatment can increase the activity of rat liver cytochrome P4501A1 and P4501A2 by 88-fold and 11-fold, respectively (44). This would be consistent with the low level of metabolism observed for benzidine and its acetylated analogs by control microsomes and the increased metabolism with -naphthoflavone treatment. Treatment by -naphthoflavone dramatically lowers hepatic cytochrome P4502C11 content (45). The latter effect on this or another cytochrome P450 could explain the lack of formation of NHA by -naphthoflavone-treated microsomes.

In conclusion, the cytochrome P4501A family is responsible for oxidative metabolism of benzidine and its N-acetylated analogs in control and -naphthoflavone-treated rat liver microsomes. Relative rates of N-acetylbenzidine and N,N-diacetylbenzidine oxidation favored activation (N-OH) rather than detoxification (ring-oxidation). A summary of structures and putative pathways is illustrated in fig. 4. Included in this summary are recent results demonstrating benzidine oxidation by cytochrome P4501A1/1A2 to 3-OH-benzidine (18). NHA, a potential carcinogenic metabolite, formation was observed with control, but not -naphthoflavone-treated microsomes. The cytochrome P450(s) producing NHA by control microsomes was not determined.

View larger version (21K): [in this window] [in a new window]   Fig. 4.   Cytochrome P450-dependent oxidation of benzidine, N-acetylbenzidine, and N,N-diacetylbenzidine by rat liver microsomes. Summary of results from this study with N-acetylbenzidine and N,N-diacetylbenzidine, and from a previous study assessing benzidine metabolism (18). AcCoA, acetyl coenzyme A.