Carcinogenesis vol.19 no.8 pp.1495–1501, 1998
Induction of hepatic CYP1A in channel catfish increases binding of 2-aminoanthracene to DNA in vitro and in vivo
David E.Watson1,3,4, William Reichert2 and Richard T.Di Giulio1,4 1Ecotoxicology
Laboratory, Nicholas School of the Environment, Duke University, Durham, NC 27708–0328, 2Environmental Conservation Division, Northwest Fisheries Science Center, National Marine Fisheries Service, 2725 Montlake Blvd East, Seattle, WA 98112–2097, USA
3Present
address: NIEHS, Mail Drop F1–08, PO Box 12233, RTP, NC 27709, USA
4To
whom correspondence should be addressed Email:
[email protected]
Data are presented from in vitro and in vivo studies that indicate cytochrome P4501A (CYP1A) in channel catfish (Ictalurus punctatus) hepatic tissue activates 2-aminoanthracene (AA) to a reactive metabolite that binds to DNA. Channel catfish were injected i.p. with vehicle or 10 mg/kg β-naphthoflavone (βNF) on two consecutive days. Two days after the final injection of vehicle or βNF, vehicle or [3H]AA was injected i.p. at 10 mg/kg, creating four different treatments: vehicle only, βNF only, [3H]AA only, and βNF/[3H]AA. Hepatic tissue was examined for CYP1Aassociated ethoxyresorufin-O-de-ethylase (EROD) activity, and for DNA adducts at 1, 2, 4 and 7 days following administration of vehicle or [3H]AA. Hepatic EROD activity in βNF-treated fish was 17-fold higher at day 0 and remained significantly greater than untreated animals for the 7-day experiment. Hepatic DNA adducts, as measured by tritium-associated DNA, ranged from 4.8 to 8.6 pmol/ mg DNA in vehicle-pretreated fish injected with [3H]AA, but ranged from 12.6 to 22.7 pmol/mg DNA in βNFpretreated fish injected with [3H]AA. Thus, pretreatment with βNF significantly increased binding of [3H]AA to hepatic DNA in vivo at all four times. Analysis by 32P-postlabeling and thin layer chromatography of hepatic DNA from channel catfish treated with AA revealed two major and several minor spots, which are indicative of DNA adduct formation. Hepatic microsomes from βNF-pretreated fish were more effective at catalysing the binding of [3H]AA to DNA in vitro than were microsomes from non-treated fish. In addition, binding was decreased by the CYP1A inhibitor 3,39,4,49-tetrachlorobiphenyl. Collectively, these data demonstrate that CYP1A is involved in the activation of AA in channel catfish. Introduction Arylamines are used as synthetic dyes (1) and occur as contaminants in processed fuels (2,3). Several arylamines have been studied in depth and much is known about their genotoxic and carcinogenic properties, particularly in mammals (4). The
arylamines 2-aminofluorene, 2-aminonaphthalene, 4-aminobiphenyl, and their acetyl derivatives, have received the bulk of attention by researchers. The arylamine that is the subject of this paper, 2-aminoanthracene (AA*), has been the subject of comparatively few mechanistic toxicological studies even though it is widely used as a positive control in the Ames test and has long been known to be a mammalian carcinogen (5,6). In addition, aminoanthracenes are significant mutagenic constitutents of the waste materials from coal-derived liquids (7). These and other arylamines are thought to be activated in vitro and in vivo by enzymatic oxidation, primarily of the amino group. In mammals this oxidation can be catalysed by cytochromes P450 (CYP) (8), flavin-containing monooxygenases (FMO) (9) and peroxidases (10,11). Purified enzyme systems, subcellular fractions and genetically engineered cell lines (8,12–14) have been used to demonstrate that these enzymes, and CYP1A2 in particular, catalyse the covalent binding of arylamines to proteins and nucleic acids, and increase arylamine mutagenicity and genotoxicity in bacterial tests. Two related goals of our laboratory are to improve understanding of the etiology of neoplasia in animals inhabiting polluted environments (15) and to understand how xenobiotic metabolism by different aquatic species may affect their susceptibility to chemical-induced neoplasia (16). An arylamine was chosen for these studies because the occurrence of arylamines in soil, sediments and fish tissues (17,18) indicates that they are of environmental concern, and because our understanding of the biochemical toxicology of arylamines in aquatic vertebrates is limited to a few publications. In comparison with mammals, there is substantial evidence that channel catfish (Ictalurus punctatus) and other fish species contain one, but not two, CYP1A proteins (19,20). In addition, and in contrast with mammals, channel catfish hepatic tissue lacks both immunological and enzymatic properties that are characteristic of FMO (21). Based on the apparent lack of FMO activity and the presence of only one, highly inducible CYP1A protein in channel catfish, we investigated the role of CYP1A in the oxidative bioactivation of AA in vivo and in vitro. Three related hypotheses were tested: (i) binding of AA to hepatic DNA in vivo would be greater in animals pretreated with the CYP1A inducer β-naphthoflavone (βNF); (ii) binding of AA to DNA in vitro would be more effectively catalysed by hepatic microsomes from animals pretreated with βNF than by microsomes from control animals; and (iii) microsome-mediated binding of AA to DNA in vitro would be inhibited by the CYP1A inhibitor 3,39,4,49-tetrachlorobiphenyl (TCB). The results support these three hypotheses. Materials and methods
*Abbreviations: AA, 2-aminoanthracene; CYP1A, cytochrome P4501A; DMA-N-oxide, N,N-dimethylaniline-N-oxide; EROD, ethoxyresorufin-O-deethylase; FMO, flavin-containing monooxygenase; NA, 2-nitroanthracene; PPL, 32P-post-labeling; TCB, 3,39,4,49-tetrachlorobiphenyl. © Oxford University Press
Purchased chemicals AA (.98%) and 3,39,4,49-tetrachlorobiphenyl (TCB, .99%) were obtained from Accustandard (New Haven, CT). All other reagents were of the highest
1495
D.E.Watson, W.Reichert and R.T.Di Giulio purity available. Enzymes were obtained from Sigma (St Louis, MO). Desferal mesylate was a generous gift from CIBA (Suffern, NY). Synthesis of [ring-3H]-2-nitroanthracene and [ring-3H]AA Nitroanthracene (NA) was synthesized by the method of Scribner and Miller (45), as modified by Bodine et al. (22), and tritiated by Amersham (2.7 Ci/mmol; Arlington Heights, IL) without detectable degradation of NA. This ring-tritiated-NA was reduced with hydrogen gas to form [ring-3H]-2aminoanthracene ([3H]AA), as previously reported (20), and purified by chromatography on silica using 10% ethyl acetate (EA)/90% hexanes (23). Specific activity, and chemical and radiochemical purity of [3H]AA were 2.5 Ci/mmol and .97% (by HPLC), respectively. Microsome preparation and analysis for monooxygenase activity Hepatic microsomes for the in vitro experiments were prepared from the following: a single Sprague–Dawley rat (200 g), a group of 20 untreated juvenile channel catfish (livers pooled), and a group of 20 βNF-pretreated juvenile channel catfish (livers pooled). Hepatic microsomes were prepared from freshly isolated liver tissue using a method similar to that of Erikkson et al. as described previously (20). Microsomal protein concentration was determined using the Bio-Rad Bradford protein kit (Hercules, CA) and bovine serum albumin as the standard. Determination of ethoxyresorufin-O-deethylase (EROD) activity was by the method of Burke and Mayer (25) using 0.1 M sodium phosphate (pH 7.8) and a jacketed Perkin Elmer LS50 fluorometer (Norwalk, CT) held at 26°C. Reactions were linear for .1 min. Fluorescence of resorufin standards were measured in the presence of buffer and microsomes, but not NADPH. FMO activity of hepatic microsomes was assayed by N,N-dimethylanilineN-oxide (DMA-N-oxide) production using the method of Ziegler and Pettit (26). Hepatic microsomes from a single, untreated Sprague–Dawley rat were used as the positive control for this assay. The assay temperature was 37°C for rat microsomes and 20°C for fish microsomes (27,28). Incubations contained hepatic microsomal protein (1 mg/ml), N,N-dimethylaniline (DMA, 1 mM), and NADPH (2 mM), and were pre-incubated for 3 min prior to addition of DMA. DMA-N-oxide production was determined using a millimolar extinction coefficient of 8.2; the production of DMA-N-oxide by rat microsomes was linear for at least 12.5 min using these conditions. The limit of detection of this method was ~0.3 nmol N,N-DMA-N-oxide/min/mg microsomal protein. Microsome-mediated binding of AA to DNA in vitro Microsome-mediated binding of AA to calf thymus DNA was determined by a previously reported method (16). Borosilicate glass tubes contained 0.5 ml of 0.1 M sodium phosphate (pH 7.4), microsomal protein (1 mg/ml), salmon testes DNA (1 mg/ml; Sigma), 5 µM [3H]AA (211 mCi/mmol), and 2 mM cofactor (NADP1 for controls and NADPH for the other three treatments). The CYP1A inhibitor 3,39,4,4’-tetrachlorobiphenyl (TCB; 29) was used at 50 µM. [3H]AA (211 mCi/mmol; .95% chemical and radiochemical purity by HPLC) and TCB stocks were prepared in DMSO. Final concentration of DMSO was 2% (v/v) in all incubations. All components except [3H]AA were pre-incubated at 26°C for 3 min. Incubations were then started by the addition of [3H]AA, and were continued for up to 10 min. Reactions were stopped by the addition of 0.75 ml buffer-saturated ethyl acetate, immediately vortexed for 10 s, and centrifuged at maximum speed for 30 s using a tabletop microfuge. The aqueous layer was transferred to a new microfuge tube. The extraction of the aqueous material was continued using equal volumes of the following water-saturated organic solvents: butanol (three times), phenol (once), chloroform/isoamyl alcohol (CIA, 24:1, twice). After each extraction, the aqueous layer was transferred to a new microfuge tube. DNA was then precipitated by the addition of 0.1 vol 5 M NaCl and 1.3 vol chilled ethanol. The DNA was resuspended in 0.5 ml of water and extracted again with equal volumes of phenol followed by CIA (twice), and precipitated by the addition of ethanol (2 vols). Following this exhaustive extraction procedure, samples did not contain ethyl acetate-extractable tritium, indicating that non-covalently bound [3H]AA had been successfully extracted. DNA was resuspended in 0.25 ml water, quantitated by UV (260/280 nm ratios were 1.8 or greater) mixed at 50 µg DNA to 20 ml Scintiverse LC (Fisher; Pittsburgh, PA) and counted for tritium on a Beckman LS6000C scintillation counter (Beckman, Fullerton, CA). No precipitate was observed in the scintillation fluid. Animals and in vivo exposures Channel catfish (12–16 g; 8–10 months old) were obtained from Blue Ridge Fish Hatchery (Kernersville, NC). Fish were maintained under flow-through conditions and fed to satiation every other day for 2 weeks prior to use. Water quality parameters were as follows: ammonia and nitrite were not detectable; temperature was 16°C; total hardness was 22 mg/l; pH 7.6. All toxicants were administered as i.p. injections using 97% propylene glycol/3% DMSO as a vehicle (5 ml/kg).
1496
Each animal received a total of three injections. The first two injections, βNF (10 mg/kg each injection) or vehicle, were made 24 h apart at days –3 and –2. A third injection, [3H]AA (10 mg/kg; 2.5 Ci/mmol) or vehicle, was made at day 0. Thus, four groups were generated: (i) vehicle/vehicle; (ii) βNF/vehicle; (iii) vehicle/[3H]AA; and (iv) βNF/[3H]AA. Animals were killed at days 0, 1, 2, 4 and 7. Five animals were used for each time point and treatment (n 5 5). Binding of AA to DNA in vivo DNA was isolated from the 9000 g pellet produced during microsome isolation (see above) as described elsewhere (16). The pellet was resuspended in 1% SDS/0.1 mM Na2EDTA (pH 7.0, 10 ml/g liver tissue) and digested with 1.5 mg Pronase E/g liver tissue for 1.5 h at 37°C. Tris-HCl (1 M, pH 7.8, 50 µl added per ml digestion solution) was added and the solution was extracted with phenol then CIA (twice) followed by precipitation of nucleic acids with 5 M NaCl (0.1 vol) and ethanol (2 vols). The precipitate was resuspended in 2 ml 150 mM sodium citrate/15 mM NaCl (pH 7.0), and digested with RNase A (200 µg) and RNase T1 (100 units) for 30 min at 37°C. The solution was again extracted with phenol then CIA (twice) and the DNA was precipitated by addition of NaCl and ethanol as described above. DNA was resuspended in the citrate/NaCl buffer, analysed by UV (260/280 nm ratios were 1.8 or greater) and counted for tritium by mixing 100 µg DNA with 20 ml Scintiverse LC. No EA-extractable tritium was detected in the DNA solutions, indicating that non-covalently bound [3H]AA was not present with the DNA. Chromatographic analysis of hepatic DNA from AA-exposed catfish Hepatic AA-DNA adduct levels in channel catfish were analysed 1 day after i.p. injection with 100 mg AA/kg body wt by 32P-post-labeling (PPL). Hepatic DNA was isolated by the method of Reddy and Randerath (30), and the PPL assay was conducted essentially according to Gupta and Randerath (31). Salmon sperm DNA from Atlantic salmon (Salmo salar) was used as a negative control in each set of analyses. DNA (10 µg) was enzymatically hydrolysed using micrococcal endonuclease and spleen phosphodiesterase. Separate aliquots of the hydrolysed DNA were processed by either butanol enhancement (32) or nuclease P1 enhancement (33) methods to remove normal nucleotides, thereby enriching the mixture in adducted 39-monophosphates. The two enhancement methods were used because they may have a differential response when DNA adducts derived from aromatic amines are present in the enzyme hydrolysate (34,35). Samples were then post-labeled using [γ-32P]ATP synthesized according to Gupta and Randerath (31). The 32P-labeled adducts were chromatographed on polyethyleneiminecellulose thin-layer chromatography (TLC) sheets prepared in the laboratory (31). The solvent systems used in the multi-directional chromatography were as follows: D1, 1.0 M sodium phosphate, pH 6.0; D2 was omitted; D3, 7.65 M urea and 4.32 M lithium formate, pH 3.5; and D4, 7.65 M urea, 1.44 M lithium chloride and 0.45 M Tris, pH 8.0. Elution of the chromatograms in D5 was not done. The 32P-labeled DNA adducts on the chromatograms were located and quantitated using storage phosphor imaging technology (36). Total nucleotides were determined by one-dimensional TLC of 59-labeled nucleotides using 0.24 M ammonium sulfate in 8 mM sodium phosphate, pH 7.4, as the solvent, followed by quantitation of the deoxyguanosine-39,59-bisphosphate spot which was assumed to represent 21% of the total nucleotides. Statistics and calculation of apparent half-life of EROD activity Two-way ANOVA procedures were employed to assess the significance of differences among the treatment groups, time points and time3treatment interactions in the data presented in Table I and Figures 1 and 4. Pairwise comparisons were made using Fisher’s least significance difference (LSD) test (Table I; Figures 1 and 4). Variance-stabilizing logarithmic (Table I; Figure 4) and square root (Figure 1) transformations were employed in these analyses. In Figure 1, the data at time 0 min were not included in the statistical analyses because these values were below the detection limit. One-way ANOVA combined with the Student–Newman–Keuls method of pairwise multiple comparisons was used to test for statistically significant differences among sample treatments at a given time (Figures 2 and 3). Results were considered statistically significant when P , 0.05. The apparent half-life of EROD activity in βNF-treated animals was calculated according to the formula t12 5 0.693/k. In this equation k represents the slope of the linear regression line obtained from a plot of the natural logarithm of EROD activity for βNF-treated animals versus experimental time (Figure 6).
Results In vivo hepatic microsomal EROD activities Hepatic microsomal EROD activities for the in vivo experiment are presented in Table I. There was a highly significant (P
CYP1A induction enhances binding of AA to DNA
Table I. Ethoxyresorufin-O-de-ethylase (EROD) activitya of channel catfish hepatic microsomes from the in vivo experimentb Time (days)
Vehicle only
[3H]AA only
βNF only
βNF/[3H]AA
0 1 2 4 7
48.5 6 10.6 86.7 6 25.5 67.5 6 13.4 66.5 6 5.41 25.8 6 7.16
— 38.9 6 6.31 52.1 6 14.8 37.1 6 8.50 53.1 6 13.5
809 6 143 824 6 60.8 533 6 46.5 328 6 57.7 110 6 17.9
— 276 6 19.2 285 6 27.4 305 6 30.6 149 6 20.3
aEROD
activity expressed as pmol resorufin produced per minute per mg microsomal protein 6 SEM (five animals per treatment per time point). bAll animals received three i.p. injections. Fish were injected with either 10 mg βNF per kg or vehicle on days –3 and –2. At day 0 fish were injected with either vehicle or 10 mg AA per kg in vehicle. The vehicle was 3% DMSO/97% propylene glycol and was delivered at 5 ml per kg per injection. These data were reported previously as a graph (20). Two-way ANOVA procedures were employed to assess the significance of AA, βNF, time and the various interactions among these three factors on EROD activity. The variance stabilizing logarithmic transformation was employed in these analyses. Pairwise comparisons were made using Fisher’s least significant difference (LSD) test. The important results from the 153 pairwise comparisons are summarized in the Results section.
Fig. 1. Time course of binding of [3H]AA to salmon sperm DNA mediated by hepatic microsomes from βNF-induced channel catfish. Incubations (26°C) contained 0.125 mg hepatic microsomal protein from βNF-induced channel catfish, 0.5 mg salmon sperm DNA, 2 mM NADPH or NADP1 and 5 µM [3H]AA (211 mCi/mmol) in 0.5 ml of 0.1 M sodium phosphate buffer (pH 7.4). Prior to counting DNA-associated tritium, DNA samples were extracted exhaustively (see Materials and methods) to remove noncovalently bound [3H]AA from the DNA. Error bars represent SEM for n 5 3 determinations per treatment per timepoint. Two-way ANOVA revealed a highly significant (P , 0.01) time3treatment interaction. Treatments not sharing a common letter are significantly different at P , 0.05. Pairwise comparisons were made using Fisher’s LSD test.
, 0.001) overall effect of βNF. Indeed, all nine pairwise comparisons of βNF groups versus the corresponding control were highly significant (P , 0.001), regardless of time and regardless of whether or not AA was present. There was also a significant (P , 0.01) time3βNF interaction. The time3βNF interaction was further investigated by plotting the natural logarithm of EROD activity for βNF-treated animals versus time (Figure 6). The correlation coefficient (r2) and slope for this semi-log plot were 0.975 and 0.298, respectively. Thus, EROD activity in βNF-treated catfish decreased over the 7 days in a log-linear fashion with an apparent half-life of 2.33
Fig. 2. Hepatic microsomes from βNF-induced catfish catalyse the in vitro binding of [3H]AA to salmon sperm DNA more effectively than do those from non-induced fish. Assay conditions were identical to those for Figure 1 except the final concentration of microsomal protein from control or βNFinduced channel catfish was 0.5 mg per ml. Incubations were for 10 min. Bars represent means 6 SEM for three determinations per treatment. Treatments not sharing a common letter are significantly different at P , 0.05. One-way ANOVA with Student–Newman–Keuls method of pairwise multiple comparisons was used to test for statistically significant differences among sample treatments.
Fig. 3. Microsome-mediated binding of [3H]AA to DNA is inhibited by the CYP1A inhibitor 3,39,4,4’-tetrachlorobiphenyl (TCB). Assay conditions were identical to those for Figure 1 except for the presence of 50 µM TCB (or equivalent volume of solvent). Microsomes were from βNF-treated fish. Incubations were for 10 min. Values reported are means 6 SEM for three determinations per treatment. Treatments not sharing a common letter are significantly different at P , 0.05. Statistical procedures were the same as in Figure 2.
days. It is also noteworthy that at the time of injection of [3H]AA, βNF-treated animals had a 17-fold greater mean hepatic microsomal EROD activity than did non-βNF-treated animals. The effects of AA on EROD activity were less consistent than the effects of βNF. AA caused a significant decrease in hepatic EROD activity in βNF-treated fish 1 and 2 days after injection of [3H]AA, but there were no differences at 4 and 7 days. We believe this AA-dependent decrease in hepatic EROD activity of βNF-treated fish is due to the suicide inhibition (mechanism-based inactivation) of CYP1A by AA (20). The effect of AA on EROD activities in non-βNF-treated fish is more complex. Compared with animals treated only with vehicle, those treated with AA had EROD activities that were significantly less at days 1 and 4, not significantly different at day 2 and significantly greater at day 7 (Table I). 1497
D.E.Watson, W.Reichert and R.T.Di Giulio
Fig. 4. Time course of hepatic AA/DNA adduct concentration in channel catfish injected with [3H]AA (Vehicle/AA) or βNF and [3H]AA (βNF/AA). Injections of vehicle or βNF (10 mg/kg) were made at experimental days –3 and –2. Injections of [3H]AA (10 mg/kg) were made at experimental day 0. Hepatic DNA was isolated from catfish at days 1, 2, 4, and 7 following injection with [3H]AA. Error bars represent SEM for n 5 5 animals. Twoway ANOVA revealed a highly significant (P , 0.01) treatment effect and a highly significant (P , 0.01) time effect, but no significant time3treatment interaction. Treatments not sharing a common letter are significantly different at P , 0.05. Pairwise comparisons were made using Fisher’s LSD test.
Table II. Ethoxyresorufin-O-de-ethylase (EROD) activity and N,Ndimethylaniline-N-oxide production (FMO activity) of channel catfish hepatic microsomes used for in vitro DNA binding study Source of hepatic microsomes
EROD activitya
FMO activityb
Channel catfish, uninducedc Channel catfish βNF-inducede Sprague–Dawley rat, uninducedf
91.7 1370 NDg
,0.3d ,0.3 2.96
apmol bnmol
resorufin produced per min per mg microsomal protein. N,N-DMA-N-oxide produced per min per mg microsomal protein. cMicrosomes from pooled livers of 20 untreated channel catfish. dLimit of detection of assay was 0.3 nmol of N,N-dimethylaniline-N-oxide produced per min per mg microsomal protein. eMicrosomes from pooled livers of 20 βNF-treated channel catfish. fMicrosomes from a single, untreated Sprague–Dawley rat. gNot determined.
In vitro hepatic microsomal enzyme activities EROD and FMO activities of channel catfish microsomes used for the in vitro DNA binding experiments are presented in Table II. EROD activity of hepatic microsomes from βNFpretreated channel catfish was 14.9-fold greater than that of microsomes from non-pretreated animals. However, FMO activity was not detectable (,0.3 nmol/min/mg microsomal protein) in microsomes from either the βNF-pretreated nor the control channel catfish microsomes, as determined by the DMA-N-oxide production assay. Hepatic microsomes from an untreated Sprague–Dawley rat (2.96 pmol/min/mg microsomal protein) served as the positive control for the FMO assay and catalysed the production of DMA-N-oxide at a rapid, linear rate for at least 12.5 min (Table II). This value for FMO activity is similar to values reported by others for rat hepatic microsomes (37,38). Microsome-mediated binding of [3H]AA to DNA in vitro Hepatic microsomes (0.25 mg protein/ml) from βNF-pretreated channel catfish catalysed the binding of [3H]AA to DNA in a time-dependent manner and reached a maximum binding 1498
Fig. 5. TLC chromatograms of 32P-post-labelled hepatic DNA from channel catfish 1 day after i.p. injection with: vehicle control (A) and (C); or 100 mg AA/kg body wt (B) and (D). DNA adducts were enriched by either the nuclease P1 method (A and B) or butanol extraction (C and D)
after 5 min (Figure 1). Binding was significantly greater in incubations that contained NADPH compared with those that contained NADP1. Using 10 min incubations and a higher microsomal protein concentration (0.5 mg protein/ml), hepatic microsomes from fish pretreated with βNF catalysed a significantly greater level of binding of [3H]AA to DNA (7.16 6 1.14 pmol/mg DNA) than did microsomes from animals not pretreated with βNF (1.46 6 0.41 pmol [3H]AA/mg DNA; Figure 2). Also, when incubations contained the CYP1A inhibitor TCB, significantly lower concentrations of [3H]AA were bound to DNA (1.61 6 0.18 pmol/mg DNA) than those observed in incubations without TCB (5.8 6 0.61 pmol/mg DNA) (Figure 3). In vivo binding of [3H]AA to hepatic DNA in channel catfish DNA-associated tritium was used to quantitate the concentration of AA bound to DNA (DNA adducts) in fish exposed in vivo. In fish treated only with [3H]AA, the range in concentration of hepatic DNA adducts was 4.8–8.6 pmol/mg DNA (Figure 4). In fish pretreated with βNF then injected with [3H]AA, the range in concentration of binding of [3H]AA to hepatic DNA was 12.6 to 22.7 pmol/mg DNA (Figure 4). Statistical analyses revealed a highly significant (P , 0.01) overall treatment effect (AA only versus βNF/AA) and a highly significant (P , 0.01) time effect. However, there was no significant interaction, implying that the treatment effect was relatively constant over time. Pairwise comparisons indicated a highly significant (P , 0.01) treatment effect (AA only versus βNF/AA) at all four time points (Figure 4). Additional comparisons indicated that for the animals treated only with AA, the reduction in response relative to day 1 was significant at days 4 and 7, but not at day 2. Conversely, for the βNF/AA group the reduction was significant at day 2, but not at days 4 and 7. Hepatic DNA from channel catfish one day after exposure to 100 mg AA/kg was analysed by the PPL assay using both the nuclease P1 and butanol extraction procedures (Figure 5).
CYP1A induction enhances binding of AA to DNA
Fig. 6. Semi-log plot showing the log-linear decrease of hepatic microsomal EROD activity in βNF-treated catfish versus time. Fish were injected i.p. with 10 mg/kg β-naphthoflavone at days –3 and –2, and were killed at days 0, 1, 2, 4 or 7. Data points are the average EROD activity for n 5 5 fish per time-point.
Two primary spots and several of lesser intensity were detected in DNA using the nuclease P1 version of the PPL assay. Compared with the nuclease P1 version, the butanol version of the PPL method showed a much weaker recovery of the two major AA-DNA adducts and a comparable recovery of the minor adducts. The values obtained were ~5% of the values obtained by the nuclease P1 version. Recovery of the BaPDE-dG-39p adduct standard was in the normal range for the butanol procedure indicating that the extraction step was effective. Discussion Hepatic CYP1A activates AA in channel catfish Two major conclusions are drawn from this work: (i) hepatic CYP1A from channel catfish activates AA to a metabolite which binds to DNA in vitro and in vivo; and (ii) hepatic concentrations of AA/DNA adducts in vivo are greater in catfish with elevated CYP1A activity. In vivo concentrations of hepatic AA/DNA adducts were significantly greater in CYP1A-induced animals compared with control animals at 1, 2, 4 and 7 days following a single i.p. injection of [3H]AA (10 mg/kg). Similar observations were made with the related Ictalurid catfish, the brown bullhead (Ameriurus nebulosus), in which CYP1A-induced brown bullhead had greater AA/ DNA adduct concentrations than did non-induced animals using the same experimental design (16). Additionally, the hepatic microsome-mediated binding of AA to DNA in vitro was greater using microsomes from CYP1A-induced compared with control channel catfish and AA/DNA adduct formation was significantly decreased by coincubation with the CYP1A inhibitor TCB. The conclusion that hepatic CYP1A from channel catfish activates AA to a metabolite that binds to DNA is consistent with Ames test data from experiments using channel catfish subcellular fractions. Hepatic S9 from channel catfish pretreated with the CYP1A inducers βNF or 3-methylcholanthrene catalysed the mutagenicity of AA and 2-aminofluorene more
effectively than did hepatic S9 from uninduced channel catfish (39). Furthermore, addition of the P450 inhibitors piperonyl butoxide and alpha-naphthoflavone to the catfish S9 fractions used in these Ames tests resulted in a reduction in the number of revertants caused by AA and 2-aminofluorene (39). These data in teleosts are similar to those in mammals in which there is considerable evidence that mammalian CYP1A proteins are important in the activation of arylamines. Cells engineered to express either rat or human CYP1A2 cDNA have greater capacity to catalyse the mutagenicity and Noxidation of arylamines than do control cells (12–14). Also, both CYP1A1 and CYP1A2 protein reconstituted in vitro can catalyse the N-oxidation of 2-aminofluorene (8). It is not known whether AA is activated by mammalian CYP1A1, CYP1A2 or both. CYP1A2 is likely to be an activation enzyme in hepatic tissue based on the fact that structurally-related arylamines are activated by 1A2. However, 1A1 may also be involved. Mice exposed intratracheally to benzo[a]pyrene (B[a]P), Aroclor-1254 or coal gas condensate, followed 1 day later by a single intratracheal dose of AA had induced lung B[a]P-3-hydroxylase activity and AA/DNA adducts than did animals exposed only to AA (40). CYP1A1 activity (EROD) is inducible in mouse lung (41) and may be responsible for the metabolic activation of AA observed by Mitchell (40). It is also possible that oxidative enzymes other than CYP1A1 in the lungs of these mice were involved in the activation of AA, or that a reactive metabolite of AA was transported to the lung from another organ. The lack of detectable FMO activity in channel catfish hepatic microsomes is consistent with results of other investigators who found neither FMO activity nor immunodetectable protein in channel catfish hepatic tissue (21). These data suggest that FMO is not involved in the metabolism of toxicants in hepatic tissue of channel catfish. However, other aquatic species have been reported to contain FMO activity and immunodetectable protein important in the activation of arylamines (43) and pesticides (28). For example, in several species of freshwater and marine mussels, the activation of the arylamines AA and 2-aminofluorene to mutagenic metabolites in the Ames test has been attributed entirely to FMO activity (43). In addition, FMO activity in rainbow trout (Oncorhynchus mykiss) is partly responsible for activation of the herbicide aldicarb (28). These findings suggest that the contribution made by FMO to the activation of toxicants varies greatly between aquatic species. Finally, it is possible that enzymes other than CYP1A contribute to the activation of AA in the channel catfish, but CYP1A appears to be the principal hepatic enzyme. Chromatographic analysis of AA/DNA adducts PPL analysis of hepatic DNA from channel catfish exposed to AA in vivo revealed that AA formed two major and several minor AA/DNA adducts (Figure 5). The nuclease P1 enhancement method was clearly more effective at detecting the two major AA/DNA adducts than was the butanol enhancement method. This observation was confirmed by a second in vivo experiment that used an experimental design that was identical to the first. However, Whong et al. (42) reported detection of 32–45% greater DNA adduct concentrations by butanol versus nuclease P1 enrichment of AA/DNA adducts in rat lung exposed to AA via intratracheal installation. It is not known why the AA/DNA adducts observed in catfish, but not rats, are more effectively detected by the 1499
D.E.Watson, W.Reichert and R.T.Di Giulio
nuclease P1 enrichment procedure. It is possible that the structures of the AA/DNA adducts differ significantly between the species. This explanation is consistent with the fact that detection of arylamine/DNA adducts by any version of the PPL assay is dramatically affected by the structure of the arylamine/DNA adducts themselves (34,35). The majority of arylamine/DNA adducts are comprised by C8-substituted purines (4). These C8 adducts are thought to be more susceptible to the phosphatase action of nuclease P1 than are other bulky adducts (35). As a result, C8 adducts are less readily detected by PPL using nuclease P1 versus butanol enrichment (34,35). Based on these facts we conclude that the two major DNA adducts being detected in these catfish by nuclease P1, and to a much lesser extent by butanol enrichment, are nonC8 adducts (Figure 5). Non-C8 DNA adducts are known to be formed by arylamine metabolites in vitro and in animals exposed in vivo (44). Unfortunately, structural characterization of the DNA adducts formed by AA has yet to be reported in any system. Similarly, it is not known whether AA is activated by the classic Noxidation pathway or whether ring-oxidation may be significant. In both rats (42) and catfish (Figure 5), the success of the nuclease P1 enrichment procedure suggests a higher level of non-C8 adduct formation by AA than occurs for other arylamines. This suggests that hydroxylamine rearrangement or ring oxidation may be more significant in the activation of AA than it is for other arylamines. Summary Data are presented that indicate that CYP1A in hepatic tissue of channel catfish is responsible for the activation of AA in vitro and in vivo to a metabolite that binds to DNA. Binding of AA to hepatic DNA in vivo was greater in animals with induced levels of CYP1A than in controls. TLC analysis of 32P-post-labeled hepatic DNA from channel catfish exposed to AA revealed two major and several minor spots that are indicative of DNA adduct formation. The identity of these DNA adducts is not known, but appears to differ from the C8-purine adducts commonly formed in mammals by other arylamines. Acknowledgements Thanks are due to Drs Greg Kedderis, George Dubay, and Fred Kadlubar, and Mr Thomas Burns Jr for technical advice on many aspects of this research, and to Ms Barbara French for excellent technical assistance. Thanks are also due to Dr Joseph Saugier and ChemSyn Laboratories for advice and for the carboxylic acid intermediate used in the synthesis of 2-nitroanthracene, and to CIBA Corp. for the desferal mesylate. This work was supported by the Society of Environmental Toxicology and Chemistry Fellowship sponsored by The Procter & Gamble Co. (David Watson is a SETAC/P&G Predoctoral Fellow). Support also came from the US Environmental Protection Agency (Grant no. R822509), the Exxon Corporation, and from National Institute of Environmental Health Sciences Training Grant No. T32ES07031-15S1.
References 1. Ellis,D.D., Jone,C.M., Larson,R.A. and Schaeffer,D.J. (1982) Organic constituents of mutagenic secondary effluents from wastewater treatment plants. Arch. Environ. Contam. Toxicol., 11, 373–382. 2. Ho,C.-H., Ma,C.Y., Clark,B.R., Guerin,M.R., Rao,T.K. and Epler,J.L. (1980) Separation of neutral nitrogen compounds from synthetic crude oils for biological testing. Environ. Res., 22, 412–422. 3. Wilson,B.W., Pelroy,R. and Cresto,J.T. (1980) Identification of primary aromatic amines in mutagenically active subfractions from coal liquefaction materials. Mutat. Res., 79, 193–202. 4. Beland,F.A. and Kadlubar,F.F. (1990) Metabolic activation and DNA
1500
adducts of aromatic amines and nitroaromatic hydrocarbons. In Cooper,C.S. and Grover,P.L. (eds) Handbook of Experimental Pharmacology, vol 94: Chemical Carcinogenesis and Mutagenesis. Springer-Verlag, Berlin, Heidelberg, pp. 267–325. 5. Shear,M.J. (1938) Carcinogenic activity of some anthracene derivatives. J. Biol. Chem., 123, cviii–cix (Abstract). 6. Bielschowsky,F. (1946) Tumours produced by 2-anthramine. Br. J Exp. Pathol., 27, 54–61. 7. Pelroy,R.A. and Wilson,B.W. (1981) Relative concentrations of polyaromatic primary amines and azaarenes in mutagenically active nitrogen fractions from a coal liquid. Mutat. Res., 90, 321–335. 8. Hammons,G.J., Guengerich,F.P., Weis,C.C., Beland,F.A. and Kadlubar,F.F. (1985) Metabolic oxidation of carcinogenic arylamines by rat, dog and human hepatic microsomes and by purified flavin-containing and cytochrome CYP mono-oxygenases. Cancer Res., 45, 3578–3585. 9. Kadlubar,F.F. and Hammons,G.J. (1987) The role of cytochrome P-450 in the metabolism of chemical carcinogens. In Guengerich,F.P. (ed.) Mammalian Cytochromes P-450. CRC Press, Boca Raton, FL, vol. II, pp. 81–130. 10. Tsuruta,Y., Subrahmanyam,V.V., Marshall,W. and O’Brien,P.J. (1985) Peroxidase-mediated irreversible binding of arylamine carcinogens to DNA in intact polymorphonuclear leukocytes activated by a tumor promoter. Chem. Biol. Interact., 53, 25–35. 11. Josephy,P.D. (1989) Activation of aromatic amines by prostaglandin H synthase. Free Radic Biol. Med., 6, 533–540. 12. Aoyama,T., Gonzalez,F.J. and Gelboin,H.V. (1989) Human cDNAexpressed cytochrome P450 IA2: mutagen activation and substrate specificity. Mol. Carcinog., 2, 192–198. 13. Wolfel,C., Platt,K.-L., Dogra,S., Glatt,H., Wachter,F. and Doehmer,J. (1991) Stable expression of rat cytochrome CYP1A2 cDNA and hydroxylation of 17β-estradiol and 2-aminofluorene in V79 chinese hamster cells. Mol. Carcinog., 4, 489–498. 14. Trottier,Y., Waithe,W.I. and Anderson,A. (1992) The detection of promutagen activation by extracts expressing cytochrome CYP1A2 cDNA: preincubation dramatically increases revertant yield in the Ames test. Mutat. Res., 281, 39–45. 15. Eufemia,N.A., Collier,T.K., Stein,J.E. Watson,D.E. and Di Giulio,R.T. (1997) Biochemical responses to sediment-associated contaminants in brown bullhead from the Niagara River ecosystem. Ecotoxicol., 6, 13–34. 16. Watson,D.E. and Di Giulio,R.T. (1997) Hepatic CYP1A in brown bullhead catalyzes the binding of 2-aminoanthracene to DNA in vivo and in vitro. Aquatic Toxicol., 37, 21–36. 17. Diachenko,G.W. (1979) Determination of several industrial aromatic amines in fish. Environ. Sci. Tech., 13, 329–333. 18. Nelson,C.R. and Hites,R.A. (1980) Aromatic amines in and near the Buffalo River. Environ. Sci. Tech., 14, 1147–1149. 19. Stegeman,J.J. and Hahn,M.E. (1994) Biochemistry and molecular biology of mono-oxygenases: current perspectives on forms, functions and regualtion of cytochrome CYP in aquatic species. In Malins,D.C. and Ostrander,G.K. (eds) Aquatic Toxicology: Molecular, Biochemical and Cellular Perspectives. CRC Press, Boca Raton, FL, pp. 87–206. 20. Watson,D.E., Me´nard,L., Stegeman,J.J. and Di Giulio,R.T. (1995) Aminoanthracene is a mechanism-based inactivator of CYP1A in channel catfish hepatic tissue. Toxicol. Appl. Pharmacol., 135, 207–215. 21. Schlenk,D., Ronis,M.J.J., Miranda,C.L. and Buhler,D.R. (1993) Channel catfish liver mono-oxygenases. Immunological characterization of constitutive cytochromes P450 and the absence of active flavin-containing mono-oxygenases. Biochem. Pharmacol., 45, 217–221. 22. Bodine,R.S., Ruehle,P.H., Roth,R.W., Bosch,G., Bosch,L., Opperman,G. and Saugier,J.,H. (1981) Synthesis of selected nitro polcyclic aromatic hydrocarbons. In Cooke,M. and Dennis,A.J. (eds) Seventh International Symposium on Polynuclear Aromatic Hydrocarbons. Battelle Press, Columbus, pp. 135–145. 23. Still,W.C., Kahn,M. and Mitra,A. (1978) Rapid chromatographic technique for preparative separations with moderate resolution. J. Org. Chem., 43, 2923–2925. 24. Erikkson,L.C., De Pierre,J.W. and Pallner,G. (1978) Preparation and properties of microsomal fractions. Pharmacol. Ther., 2, 281–317. 25. Burke,D.M. and Mayer,R.T. (1974) Ethoxyresorufin: direct fluorometric assay of a microsomal O-dealkyaltion which is preferentially inducible by 3-methylcholanthrene. Drug Metab. Dispos., 2, 583–588. 26. Ziegler,D.M. and Pettit,F.H. (1964) Formation of an intermediate N-oxide in the oxidative demethylation of N,N-dimethylaniline catalysed by liver microsomes. Biochem. Biophys. Res. Commun., 15, 188–193. 27. Gold,M.S. and Ziegler,D.M. (1973) Dimethylaniline N-oxidase and aminopyrine N-demethylase activities of human liver tissue. Xenobiotica, 3, 179–189.
CYP1A induction enhances binding of AA to DNA 28. Schlenk,D. and Buhler,D.R. (1991) Role of flavin-containing monooxygenase in the in vitro biotransformation of aldicarb in rainbow trout (Onchorhynchus mykiss). Xenobiotica, 21, 1583–1589. 29. Gooch,J.W., Elskus,A.A., Kloepper-Sams,P.J., Hahn,M.E. and Stegeman,J.J. (1989) Effects of ortho- and non-ortho-substituted polychlorinated biphenyl congeners on the hepatic mono-oxygenase system in scup (Stenotomus chrysops). Toxicol. Appl. Pharmacol., 98, 422–433. 30. Reddy,M.V. and Randerath,K. (1987) 32P-analysis of DNA adducts in somatic and reproductive tissues of rats treated with the anticancer antibiotic, mitomycin C. Mutat. Res., 179, 75–88. 31. Gupta,R.C. and Randerath,K. (1988) Analysis of DNA adducts by 32Plabeling and thin layer chromatography. In Friedberg,E.C. and Hanawalt,P.H. (eds) DNA Repair. Marcel Dekker, New York, vol. 3, pp. 399–418. 32. Gupta,R.C. (1985) Enhanced sensitivity of 32P-post-labeling analysis or aromatic carcinogen DNA adducts. Cancer Res., 45, 5656–5662. 33. Reddy,M.V. and Randerath,K. (1986) Nuclease P1-mediated enhancement of sensitivity of 32P-post-labeling test for structurally diverse DNA adducts. Carcinogenesis, 7, 1543–1551. 34. Gupta,R.C. and Earley,K. (1988) 32P-adduct assay: comparative recoveries of structurally diverse DNA adducts in the various enhancement procedures. Carcinogenesis, 9, 1687–1693. 35. Gallagher,J.E., Jackson,M.A., George,M.H., Lewtas,J.E. and Robertson,I.G.C. (1989) Differences in detection of DNA adducts in the 32P-post-labeling assay after either 1-butanol extraction or nuclease P1 treatment. Cancer Lett., 45, 7–12. 36. Reichert,W.L., Stein,J.E., French,B., Goodwin,P. and Varanasi,U. (1992) Storage phosphor imaging technique for detection and quantitation of DNA adducts measured by the 32P-post-labeling assay. Carcinogenesis, 13, 1475–1479. 37. Kedderis,G.L. and Rickert,D.E. (1985) Loss of rat liver microsomal cytochrome P-450 during methimazole metabolism. Drug Metab. Dispos., 13, 58–61. 38. Ziegler,D.M. (1980) Microsomal flavin-containing mono-oxygenase: oxygenation of nucleophilic nitrogen and sulfur compounds. In Jakoby,W.B. (ed.) Enzymatic Basis of Detoxication. Academic Press, New York, pp. 201–227. 39. Johnson,B.T. (1993) Genotoxicity testing with fish hepatic S9 for evaluation of complex mixtures in the aquatic environment: the use of channel catfish as a model. Aquatic Toxicol., 7, 93–314. 40. Mitchell,C.E. (1985) Effect of aryl hydrocarbon hydroxylase induction on the in vivo covalent binding of 1-nitropyrene, benzo[a]pyrene, 2aminoanthracene, and phenanthridone to mouse lung deoxyribonucleic acid. Biochem. Pharmacol., 34, 545–551. 41. Diliberto,J.J., Akubue,P.I., Luebke,R.W. and Birnbaum,L.S. (1995) Doseresponse relationships of tissue distribution and induction of CYP1A1 and CYP1A2 enzymatic activities following acute exposure to 2,3,7,8tetrachloro-p-dioxin (TCDD) in mice. Toxicol. Appl. Pharmacol., 130, 197–208. 42. Whong,W.-Z., Stewart,J.D. and Ong,T. (1992) Comparison of DNA adduct detection between two enhancement methods of the 32P-post-labeling assay in rat lung cells. Mutat. Res., 283, 1–6. 43. Kurelec,B. (1985) Exclusive activation of aromatic amines in the marine mussel Mytilus edulis by FAD-containing mono-oxygenase. Biochem. Biophys. Res. Commun., 127, 773–778. 44. Yamazoe,Y., Miller,D.W., Weis,C.C., Dooley,K.L., Zenser,T.V., Beland,F.A. and Kadlubar, F.F. (1985) DNA adducts formed by ring-oxidation of the carcinogen 2-naphthylamine with prostaglandin H synthase in vitro and in the dog urothelium in vivo. Carcinogenesis, 6, 1379–1387. 45. Scribner,J.D. and Miller,J.A. (1965) Synthesis of 2-nitroanthracene and Nhydroxy-2-anthrylamine. J. Chem. Soc., pp. 5377–5380. Received on September 4, 1997; revised on February 19, 1998; accepted on April 2, 1998
1501
