Metabolism of Benzo(a)pyrene with Isolated Hepatocytes and the Formation and Degradation of DNA-binding Derivatives*

The metabolism of [Wlbenzo(a)pyrene with isolated hepatocytes from 3-methylcholanthrene-treated rats was exam-ined with the aid of high pressure liquid chromatography. Covalent binding of [SHlbenzo(a)pyrene metabolites to the intracellular DNA was investigated. The effects of (Y-naphthoflavone, salicylamide, trichloropropene oxide, and diethylmaleate, individually or combined, on the metabolism and covalent DNA binding of benzo(a)pyrene were determined. The results indicate that the initial organic-solu-ble metabolites were arene oxides, phenols, quinones, and dihydrodiols and that these were subsequently converted to relatively polar, organic-soluble nonconjugated and sulfate-conjugated metabolites and to aqueous-soluble nonconjugated and glucuronide-and glutathione-conjugated metabolites. ru-Naphthoflavone inhibited the formation of benzo(a)pyrene metabolites that covalently bound to hepa- tocyte DNA, while the binding was stimulated by salicylamide, or

The metabolism of [Wlbenzo(a)pyrene with isolated hepatocytes from 3-methylcholanthrene-treated rats was examined with the aid of high pressure liquid chromatography. Covalent binding of [SHlbenzo(a)pyrene metabolites to the intracellular DNA was investigated.
The effects of (Ynaphthoflavone, salicylamide, trichloropropene oxide, and diethylmaleate, individually or combined, on the metabolism and covalent DNA binding of benzo(a)pyrene were determined. The results indicate that the initial organic-soluble metabolites were arene oxides, phenols, quinones, and dihydrodiols and that these were subsequently converted to relatively polar, organic-soluble nonconjugated and sulfateconjugated metabolites and to aqueous-soluble nonconjugated and glucuronide-and glutathione-conjugated metabolites.
ru-Naphthoflavone inhibited the formation of benzo(a)pyrene metabolites that covalently bound to hepatocyte DNA, while the binding was stimulated by salicylamide, trichloropropene oxide, or diethylmaleate. Our observations indicate that benzo(a)pyrene-oxide hydration and glutathione conjugation, and glucuronide and sulfate conjugation of hydroxylated benzo(a)pyrene metabolites operate in concert to detoxify electrophilic DNA-binding benzo(a)pyrene metabolites in isolated hepatocytes. The degree of covalent binding of benzo(a)pyrene to the nuclear DNA of isolated hepatocytes seems thus to be correlated with the production of electrophilic benzo(a)pyrene metabolites and the rate of their disposal by epoxide hydratase and conjugation reactions.
It now appears well established that carcinogenic polycyclic aromatic hydrocarbons such as benzo(a)pyrene (l-3), are not inherently harmful, but are activated during their biotransformation to electrophilic metabolites that are toxic (4,5), are mutagenic , and covalently bind to tissue macromolecules (1)(2)(3). Although the covalent interaction of these metabo-lites with cellular macromolecules may involve other nucleophiles in addition to DNA (111, there are good correlations between carcinogenic potency and DNA binding (12-X).
It is believed that the reactive metabolites are the initially formed, or recycled, products of the cytochrome P-450-linked monooxygenase and that subsequent conjugations with glucuronic acid, sulfate, or glutathione are true detoxication reactions (26). The most potent of the metabolites in regard to cell transformation and covalent binding to DNA appear to be benzo(a)pyrene 7,8-dihydrodiol-9,10-oxide and benzo(a)pyrene 4,5-oxide (6,7,10,27). However, benzo(a)pyrene phenols are also mutagenic (27)(28)(29) and capable of being metabolically activated to . Studies on the activation of benzo(a)pyrene to DNA-binding metabolites have mainly been performed in experimental systems containing liver microsomes and pure DNA or isolated cell nuclei (30,(33)(34)(35). These models do, however, lack the conjugative detoxication mechanisms and are therefore not suited for investigating the "balance" between toxifying and detoxifying reactions which may in fact be decisive for the toxic effect produced. As an experimental model for such studies we have chosen isolated hepatocytes which actively catalyze the cytochrome P-450-and epoxide hydratase-linked, as well as the conjugative, reactions (36, 37) and which, when isolated from 3-methylcholanthrene-treated rats, have been found to convert benzo(a)pyrene to product(s) that bind to endogenous DNA (38,39).
We have previously reported some characteristics of benzo(a)pyrene metabolism in isolated hepatocytes, including a rough estimate of the pattern of metabolites, using thin layer chromatography for separation (40). In the present study, we have used HPLC for a detailed analysis of the benzo(a)pyrene metabolite pattern in the hepatocytes and in and eluted with a choice of two alcohol/water gradients. These HPLC systems were based on those described by Selkirk et al. (44,45) and Holder et al. (46). System A was a linear gradient of changing solvent composition, changing from 45% water, 55% methanol:ethanol (2:l) to 20% water, 80% methanol:ethanol (2:l). System B was also a linear gradient, from 50% water, 50% methanol:ethanol (15) to 25% water, 75% methanol:ethanol (1:5). System B was used in order to separate benzo(a)pyrene 4,5-oxide. It failed to separate benzo(a)pyrene 4,5and 7,8-dihydrodiols, but this was achieved with System A (Fig. 1). Both gradients were run for 1 h with a solvent flow rate of 1 ml/min. At the end of 1 h with either system the solvent was changed to 100% methanol:ethanol mixture in order to elute unmetabolized benzo-(a)pyrene.
The effluent stream was monitored for absorbance at 254 nm and collected as fractions (100 x 0.7 ml) for scintillation counting.
Fractions were counted in 10 ml of Aquasol containing 1% acetic acid using a Beckman LS-150 scintillation counter. Counting efficiency for all fractions was 84%. Better than 98% of the radioactivity applied to the HPLC column was recovered while the overall recovery of radioactivity added to the hepatocyte incubations was more than 85%. All estimations of benzo(a)pyrene metabolites have been corrected to 100% recovery for each experiment. The added butylated hydroxytoluene was extracted into the ethyl acetate and eluted on the HPLC with benzo(a)pyrene, after the benzo(a)pyrene metabolites.
Throughout this report the identifications of radioactive benzo(a)pyrene metabolites are based on their co-chromatography with authentic compounds. In the light of other studies using similar HPLC conditions (46, 471, it is possible that the benzo(a)pyrene 4,5-dihydrodiol was contaminated with 11,12-dihydrodiol, that any benzo(a)pyrene 4,5-or 11,12-quinone metabolites eluted with the identified quinones, and that the metabolite identified as 3-hydroxybenzo(a)pyrene included some l-and 7-hydroxybenzo(alpyrene DNA Extraction and Assay of p7HIBenzo(a)pyrene Covalently Bound to DNA-The DNA isolated from hepatocytes was exhaustively purified of protein and RNA. The centrifugal pellet of hepatocytes from 5 ml of incubation mixture was solubilized for 30 min at room temperature with 5 ml of 1% sodium dodecyl sulfate containing 10 mM EDTA, using slow inversion-rotation.
This was then digested for 60 min at 37" with 5 ml of 0.2 M Tris buffer, pH 7.5, containing 10 mM EDTA and 2.5 mg of protease.
Next, the sample was extracted for 30 min at 2" with 10 ml of phenol/chloroform (l:l), the emulsion was broken by centrifugation at 10,000 x g for 10 min, and the organic phase was aspirated off. The aqueous phase was twice extracted with 5 ml of ether, then solid NaCl was added to 1 M concentration and 20 ml of cold (-40") ethanol were slowly added. At this stage the sample was stored overnight at -40". The ethanolprecipitated nucleic acids and glycogen were redissolved in 3 ml of 0.1 x SSC, 1.5 ml of a 9 M LiCl solution was added (final concentration, 3 M), and the sample was allowed to stand for 24 h at 2". Ribosomal RNA was precipitated by centrifugation at 79,000 x g for 30 min. The supernatant was collected and 9 ml of cold ethanol C-40") were slowly added. The precipitate was redissolved in 3 ml of 0.1 x SSC, as above, and 1 M Tris buffer, pH 7.5, and EDTA were added to a final concentration of 100 and 10 mrq respectively. The data in Table  I typify  the metabolite  pattern  both  during the initial phase of incubation (5 min), when the concentration of benzo(a)pyrene was not rate-limiting, and at a time (30 min) when more than 94% of the benzo ( (19) 11.5 i 3.1 (63)   Benzo(a)pyrene Metabolism and DNA Binding with Hepatocytes 9,10-dihydrodiol was found outside the cells, while the majority of 9-hydroxybenzo(a)pyrene, 3-hydroxybenzo(a)pyrene, and the Fraction I metabolites remained inside the hepatocytes. The 4,[5][6][7] and the benzo(a)pyrene quinones were distributed approximately equally inside and outside the cells. An almost identical distribution pattern was observed after incubating benzo(a)pyrene with the hepatocytes for 30 min in the presence of salicylamide, which allowed accumulation of the phenols and dihydrodiols by inhibiting their conjugation (see later).
Effects of a-Naphthoflavone, Salicylamide, TCPO, or Diethylmaleate on Benzo(a)pyrene Metabolism-The effects of selective enzyme inhibitors on benzo(a)pyrene metabolism in hepatocytes isolated from 3-methylcholanthrene-treated rats were investigated. The results of these experiments are illustrated in Fig. 3 and Table III. For each inhibitor the results of one experiment are shown, together with results of the appropriate control incubation (hepatocytes from the same preparation but with inhibitor absent). The experiments were repeated three times, with virtually identical results. The proportion of leaking hepatocytes was determined and was not significantly altered by benzo(a)pyrene and TCPO or diethylmaleate.
a-Naphthoflavone inhibits the 3-methylcholanthrene-induced microsomal monooxygenase, probably at cytochrome P-448 (50). Accordingly, a-naphthoflavone decreased the amounts of all the organic-and aqueous-soluble metabolites formed from benzo(a)pyrene with isolated hepatocytes of 3methylcholanthrene-treated rats ( Fig. 3 and Table III). We have shown in Fig. 3 the results of a-naphthoflavone inhibition at both 5 and 30 min since the effect was more pronounced at the earlier time, when there were still relatively large quantities of benzo(a)pyrene phenols and dihydrodiols in the absence of a-naphthoflavone.
Salicylamide inhibits glucuronide and sulfate conjugation (51, 52) and in our experiments 2 mM salicylamide accordingly lowered the amounts of the Fraction I metabolites and the water-soluble metabolites, whereas the amounts of phenols, quinones, and dihydrodiols were increased (Table III and Fig.  3). The total metabolism of benzo(a)pyrene was not affected by addition of salicylamide. /3-Glucuronidase and aryl sulfatase hydrolysis of the metabolites obtained after incubation of benzo(a)pyrene with hepatocytes for 30 min confirmed the effects of salicylamide (Table  IV). p-Glucuronidase and aryl sulfatase hydrolyzed 48 and 20%, respectively, of the water-soluble metabolites into organic-soluble benzo(a)pyrene metabolites such as dihydrodiols, quinones, and phenols. The total amount of Fraction I metabolites was not affected by p-glucuronidase treatment, whereas with aryl sulfatase the amount was decreased, indicating that the Fraction I consisted partly of sulfate conjugates. Free quinones were probably formed during spontaneous oxidation of the benzo( a)pyrene phenols or from hydrolyzed dihydroxybenzo(a)pyrene glucuronides (20,21) and sulfates. The data in Table IV are from one experiment and were supported by the results of a second investigation (not shown). TCPO inhibits epoxide hydratase, thereby leading to increased amounts of arene oxides and their spontaneous isomerization product, phenols, and to decreased dihydrodiols (53). This effect of TCPO was also seen in the metabolism of benzo(a)pyrene with isolated hepatocytes (Fig. 3, Table III). It has to be noted that TCPO added to the hepatocyte suspensions also reduced the intracellular level of GSH by approximately 98%. With TCPO the amount of Fraction I metabolites VIII, benzo(a)pyrene 4,5-oxide; IX, 9-hydroxybenzo(a)pyrene; X, 3-hydroxybenzo(a)pyrene. For or-naphthofla-"one, the results for metabolism of benzo(a)pyrene during 5 min, following 5 min of preincubation, are also given (ilzset). Metabolites were analyzed by HPLC and scintillation counting as described under "Materials and Methods." was decreased. Of the possible arene oxides, only the stable K region oxide, benzo(a)pyrene 4,5-oxide, was identified. TCPO slightly inhibited total benzo(a)pyrene metabolism over a period of 30 min and it increased the ratio of organic-to aqueoussoluble metabolites. The presence of salicylamide, diethylmaleate, or TCPO caused an increase in a minor unidentified metabolite, which eluted from the HPLC in Fractions 29 to 32 between the benzo(a)pyrene 9,10-and 4,5-dihydrodiols (data not shown). TCPO was most potent in this respect and also caused a novel peak of metabolite radioactivity to elute immediately after 3-hydroxybenzo(a)pyrene (Fig. 1). The level of GSH in freshly prepared hepatocytes isolated from 3-methylcholanthrene-treated rats was 6.2 ~g/lO" cells. After a 35-min incubation at 37", a 15% decrease in the GSH level was observed. The presence of benzo(a)pyrene, added after 5 min of incubation, decreased the GSH level after 35 min   (Table  III). The main effect of diethylmaleate on the organic-soluble metabolites was to increase the benzo(a)pyrene phenols and 4,5-oxide (Fig. 3).
The effects on benzo(a)pyrene metabolism of combining two inhibitors were cumulative but not strictly additive (Fig. 3, Table III). The combinations were slightly more inhibitory to total benzo(a)pyrene metabolism than were the individual inhibitors, but this increased inhibition was manifested solely in the amount of aqueous-soluble metabolites, whereas the organic-soluble metabolites were enhanced (Table III). Dual inhibitor combinations inhibited the production of aqueoussoluble metabolites more effectively than did single inhibitors. There was a remarkable increase in 9-hydroxybenzo(a)pyrene relative to the major metabolite 3-hydroxybenzo(a)pyrene when TCPO was combined with salicylamide during a 30-min incubation (Fig. 3). leate on benzo(a)pyrene metabolism, we investigated their influence on the covalent binding of l:sHlbenzo(a)pyrene to the DNA of hepatocytes isolated from 3-methylcholanthrenetreated rats (Fig. 4). As described earlier and discussed later, we interpret this covalent binding as being mainly due to lYHlbenzo(a)pyrene metabolites. The data presented in Fig. 4

DISCUSSION
With suspensions of hepatocytes isolated from 3-methylcholanthrene-treated rats the pattern of organic-soluble benzo(a)pyrene metabolites produced after 5 min metabolism was similar to the patterns reported for rat liver microsomes (Table I; Refs. 46,47,541,pyrene and benzo(a)pyrene 7,8-and 9,10-dihydrodiols constituting the major metabolites. But unique to initial benzo(a)pyrene metabolism with isolated hepatocytes was the major fraction of relatively polar Fraction I metabolites that eluted from the HPLC before benzo(a)pyrene 9,10-dihydrodiol. This fraction, which increased steadily during a 30-min incubation, was probably closely related to the very polar organic-soluble benzo(a)pyrene metabolites that accumulate with liver microsomes during prolonged incubations and are considered to be polyhydroxylated benzo(a)pyrene derivatives recycled through the monooxygenase (22,23).
It is clear from our results that the pattern of organicsoluble benzo(u)pyrene metabolites changes with the duration of the reaction and that eventually all the added benzo(a)pyrene becomes metabolized to water-soluble and comparatively polar organic-soluble metabolites (cf. Fig. 2). The observed effects of p-glucuronidase confirm that glucuronide conjugation was a major reaction of benzo(a)pyrene phenols, as has been found in uivo (55) and using liver microsomes (24). Aryl sulfatase treatment, in turn, revealed that the inhibitory effect of salicylamide on the formation of the most polar of the organic-soluble benzo(u)pyrene metabolites was due to inhibition of sulfate conjugation (cf. Fig. 3; Table  IV). With rodent lung tissue, sulfate conjugates of 3-, 7-, and 9hydroxybenzo(u)pyrene are formed (25,561 and similar results were obtained with the isolated hepatocytes (cf . Table IV) where aryl sulfatase treatment produced an increase in phenols identified as 3-and 9-hydroxybenzo(a)pyrene; the metabolite identified as 3-hydroxybenzo(u)pyrene may have included some l-and 7-hydroxybenzo(a)pyrene (46, 47). The organic-soluble metabolites were distributed between the hepatocytes and the extracellular medium (Table II). This distribution was asymmetric for some metabolites and was probably not simply related to the relative polarities of the metabolites since the most polar Fraction I metabolites were preferentially retained, whereas the less polar benzo(a)pyrene 9,10-dihydrodiol was preferentially excluded from the cells. The possible hepatic excretion of benzo(a)pyrene metabolites such as phenols and dihydrodiols, which can be further metabolized to active derivatives (6,29,32,57), may be of importance in viva. Thus it is tempting to speculate that lung and skin, for example; which form much lower quantities of the initial benzo(u)pyrene metabolites than does liver, will be able to reactivate primary metabolites that have been formed in the liver and transferred to the target tissues.
Hepatocytes isolated from 3-methylcholanthrene-treated rats were able to activate benzo(u)pyrene to metabolite(s1 that bound to endogenous DNA, whereas no binding could be detected in hepatocytes from control rats. In the absence of any added inhibitors, the maximal binding was approximately 1 pmol of benzo(a)pyrene metabolites/50 pg of DNA (equivalent to 5 x 10" hepatocytes (48)) which represented 0.001% of the metabolized benzo(u)pyrene.
Since the binding of benzo-(ujpyrene metabolites to DNA continued to increase for several minutes after the disappearance of nonmetabolized benzo(a)pyrene, it is likely that the binding also involved further activated or recycled metabolites. Although the "initial" benzo(a)pyrene 4,5-oxide and reactivated benzo(a)pyrene phenols also show marked DNA-binding capacity, the recycled metabolite benzo(a)pyrene 7,8-dihydrodiol-9,10-oxide, is currently considered one of the most potent DNA-binding benzo(a)pyrene derivatives (30)(31)(32)58). Isolated rat liver nuclei have been shown to metabolize benzo(a)pyrene as well as its microsomal metabolites to DNAbinding derivatives (33-35) and it has been suggested that the primary benzo(a)pyrene metabolites produced by the endoplasmic reticulum are further activated in the close vicinity of the nucleus, possibly in the nuclear envelope, before reacting with nuclear constituents (34).
Salicylamide inhibits glucuronide (51) and sulfate (52) conjugation of benzo(u)pyrene metabolites, and, as could be expected, addition of salicylamide to the cellular incubations decreased the level of aqueous-soluble metabolites, whereas the quantity of organic-soluble metabolites, such as phenols and dihydrodiols, was enhanced (cf. Fig. 3). Since the inhibitor also caused an increase in the binding of benzo(a)pyrene metabolites to DNA, these conjugation reactions obviously play a role in the detoxication of the initial benzo(u)pyrene metabolites. The increased DNA binding resulting from salicylamide inhibition of conjugation was thus probably related to an accumulation of benzo(u)pyrene phenols and dihydrodiols which could be further metabolized to electrophilic products (32,34,57).
Several benzo(u)pyrene oxides are conjugated with glutathione both spontaneously and enzymatically by glutathione transferases (18,59,60), and glutathione has been shown to decrease the binding of benzo(ulpyrene metabolites to nucleic acids of isolated liver cell nuclei (33, 611. In the intact hepatocytes, conjugation of benzo(u)pyrene metabolites with glutathione can be inhibited by addition of diethylmaleate, which decreases the intracellular GSH level (62) and is a glutathione transferase substrate (63). Diethylmaleate added to the cellular incubations lowered the formation of aqueous-soluble metabolites and elevated the amount of benzo(u)pyrene 4,5-oxide and benzo(u)pyrene phenols (cf. Fig. 3). The increased levels of benzo(u)pyrene 4,5-oxide and probably also of other "initially" formed benzo(a)pyrene oxides and benzo(u)pyrene phenols, were manifested in an enhanced binding to cellular DNA (cfi Figs. 3 and 4). Thus, a direct interaction of primary benzo(u)pyrene oxides with nuclear constituents has also to be considered when interpreting the observed results. Interpretation of the action of TCPO in the present experiment is complicated since the agent had dual effects on isolated hepatocytes: it inhibited epoxide hydratase (53) and thereby the formation of dihydrodiols (cfi Fig. 31 and it also depleted GSH by reacting both spontaneously (53) and with glutathione transferase catalysis (64). Its effects on benzo(u)pyrene metabolism with the hepatocytes were, however, characteristic of its inhibition of epoxide hydratase and different from those of diethylmaleate (cfi Fig. 3). Thus, TCPO inhibited the formation of dihydrodiols and increased the levels of benzo(u)pyrene phenols and benzo(u)pyrene 4,5-oxide. The reduced amount of the most polar Fraction I metabolites and of aqueous-soluble metabolites indicates that some of these were further metabolites of benzo(u)pyrene dihydrodiols recycled through the cytochrome P-450-epoxide hydratase system, for example benzo(u)pyrene triols and tetrols (65