Rat liver endothelial and Kupffer cell-mediated mutagenicity of polycyclic aromatic hydrocarbons and aflatoxin B1.

The ability of isolated rat liver endothelial and Kupffer cells to activate benzo(a)pyrene (BP), trans-7,8-dihydroxy-7,8-dihydrobenzo(a)pyrene (DDBP), trans-1,2-dihydroxy-1,2-dihydrochrysene (DDCH), and aflatoxin B1 (AFB1) to mutagenic metabolites was assessed by means of a cell-mediated bacterial mutagenicity assay and compared with the ability of parenchymal cells to activate these compounds. Endothelial and Kupffer cells from untreated rats were able to activate AFB1 and DDBP; DDBP was activated even in the absence of an NADPH-generating system. Pretreating the animals with Aroclor 1254 strongly enhanced the mutagenicity of the dihydrodiol, whereas the mutagenicity of AFB1 showed a slight increase. BP and DDCH were only activated by endothelial and Kupffer cells isolated from Aroclor 1254-pretreated rats. Parenchymal cells from untreated animals activated all four carcinogens tested; Aroclor 1254 enhanced the parenchymal cell-mediated mutagenicity of BP and DDCH but did not affect that of DDBP and clearly reduced that of AFB1. The reduced mutagenicity of AFB1 correlates with the decrease in the amount of 2 alpha-hydroxytestosterone formed when testosterone was incubated with parenchymal cell microsomes from Aroclor 1254-pretreated rats (compared with microsomes from untreated animals): the formation of 2 alpha-hydroxytestosterone is specifically catalyzed by cytochrome P-450h, a hemoprotein thought to be involved in the activation of AFB1. These results show that not only rat liver parenchymal cells, but also endothelial and Kupffer cells, activate several carcinogens to mutagenic metabolites.


Introduction
In the mammalian liver, parenchymal cells account for more than 90% of the mass and 65% of the total number of cells. The nonparenchymal cells of the liver are primarily endothelial and Kupffer cells, which occupy the sinusoidal lining of the intact liver. All three of these cell types are heterogeneous in the way they respond to insult from xenobiotics. The basis of the heterogeneous responses to hepatotoxins is poorly understood, but several factors, including the distribution of the xenobiotics within the liver, differences in the DNA repair mechanisms among the cell populations, and differences in the ability to activate and detoxify these compounds, might be involved. It is the cell's ability to activate and detoxify hepatotoxins that we wish to address in the present study.
Recent studies have shown that in the sinusoidal lining cells oxidative enzyme activities (aminopyrine Ndemethylase and ethoxyresorufin O-deethylase) are low, while postoxidative enzyme activities (glutathione S-transferase and epoxide hydrolase) are relatively high (1,2). The aim of this study was to analyze the capacity of sinusoidal lining cells, with their low cytochrome P-450-dependent enzyme activities, to activate several carcinogenic compounds to mutagenic metabolites and compare this capacity with that of parenchymal cells. The activating potential of the isolated liver cell subpopulations was estimated by means of a cell-mediated mutagenicity assay (3). To characterize the cytochrome P-450 profile of rat liver parenchymal cells, as well as nonparenchymal cells, from untreated and Aroclor 1254pretreated rats, the oxidative metabolism of testosterone with microsomes from parenchymal, endothelial, and Kupffer cells was investigated. It has been shown recently that various forms of rat liver microsomal cytochrome P-450 catalyze the hydroxylation of testosterone with a high degree of regio-and stereoselectivity (4,5).

Materials and Methods
Male Sprague-Dawley rats (200-240 g body weight) were used. Aroclor 1254 in corn oil was administered as a single IP dose (500 mg/kg body weight) 5 days before killing. Control rats received appropriate volumes of corn oil (2.5 mL/kg body weight).
Parenchymal, endothelial, and Kupffer cells were isolated and characterized according to methods previously described in the literature (1). For each mutagenicity assay, pools of Kupffer cells and of endothelial cells isolated from three untreated or Aroclor 1254-pretreated rats were prepared. All the experiments were performed in cell homogenates obtained by sonicating the cells for 30 sec at 60% duty cycle on a Branson cell disruptor (model B-15).
Reversion of Salmonella typhimurium his-, a system developed by Ames et al. (6), was used for the estimation of mutagenicity. The incubation mixture consisted of 1 mL complete Krebs-Henseleit buffer containing previouslAT sonicated cells (1 x 106 parenchymal cells or 10 x 10 nonparenchymal cells), bacteria (1.7 x 10' cells), and the test compounds dissolved in 10 ,uL dimethyl formamide. Incubations were supplemented with 2.5 mM NADP+ and 2.0 mM glucose-6-phosphate. Samples were incubated in a shaking water bath at 37°C in the dark for 2 hr. Two milliliters of 45°C warm top agar, which contained 0.55% w/v agar, 0.55% w/v NaCl, 50 ,uM histidine, and 50 ,uM biotin in 25 mM sodium phosphate buffer (pH 7.4), was then added, and the mixture was poured onto a Petri dish containing 22 mL of minimal agar (1.5% agar in Vogel-Bonner E medium with 2% glucose). After incubation at 37°C in the dark for 3 days, colonies of his' revertants were counted.
Testosterone hydroxylation assays were performed by incubating microsomes (1 mg protein) from parenchymal, endothelial, and Kupffer cells for 30 min with 1 mM testosterone in the presence of 0.6 mM NADP+, 8 mM glucose-6-phosphate, 1.4 units of glucose-6-phosphate dehydrogenase, and 5 mM MgCl2. Testosterone metabolites were extracted with dichloromethane and quantified by HPLC as previously described (7). Endothelial and Kupffer cells from three animals were pooled for the preparation of microsomes.
Pretreatment of the animals with Aroclor 1254 (a polychlorinated biphenyl mixture that exhibits both phenobarbital and 3-methylcholanthrene-inducing properties) did not affect the yield, viability, and purity of the isolated parenchymal, endothelial, and Kupffer cell fractions, but it significantly increased the protein con-centrations of parenchymal and nonparenchymal cells by about 80 and 30%, respectively.

Mutagenicity Studies
The cell-mediated mutagenicity assays were performed with roughly equal protein concentrations of parenchymal and nonparenchymal cells, i.e., 1 x 106 parenchymal cells and 10 x 10 nonparenchymal cells; only in the case of trans-7,8-dihydroxy-7,8-dihydrobenzo(a)pyrene (DDBP), 0.5 x 106 parenchymal cells and 5 x 106 nonparenchymal cells were used. None of the compounds investigated was mutagenic in the absence of cell homogenates.
Parenchymal cells isolated from untreated animals showed a limited capacity to activate benzo(a)py-rene (BP) (Fig. 1A), whereas endothelial and Kupffer cells were unable to activate this compound (Fig. 1B); this was still the case if the cell number and/or the incubation time were doubled (data not shown). After administration of Aroclor 1254, a clear mutagenic effect of BP, mediated by parenchymal as well as nonparenchymal cells, was observed (Figs. 1A and B).
DDBP was activated by parenchymal, endothelial, and Kupffer cells isolated from untreated rats (Figs. 1C and D). Interestingly, this also occurred in all three cell types in the absence of the NADPH-generating system. Pretreatment of the animals with Aroclor 1254 enhanced the nonparenchymal cell-mediated mutagenicity of the dihydrodiol, while it did not alter the mutagenic effect mediated by the parenchymal cells (Figs. 1C and D).
Aflatoxin B1 (AFB1) was strongly mutagenic after incubation with parenchymal and nonparenchymal cells isolated from untreated animals (Figs. 1G and H). Pretreatment of the rats with Aroclor 1254 slightly enhanced the mutagenicity of the mycotoxin mediated by endothelial and Kupffer cells, whereas it significantly decreased the activating potential of the parenchymal cells (Figs. 1G and H).

Testosterone Metabolism
The oxidation products identified after incubating testosterone with parenchymal cell microsomes from untreated rats were 6a-, 7a-, 61-, 16a-, 161-, 2a-, and 21hydroxytestosterone and androstenedione (Fig. 2). The hydroxylation of testosterone occurred at a significantly lower rate in microsomes of endothelial and Kupffer cells; furthermore, neither 161nor 21-hydroxytestosterone could be detected with these microsomes. If incubations were performed with microsomes of parenchymal and nonparenchymal cells from Aroclor 1254pretreated rats, the formation rate of all the testoster-

Discussion
Bay-region diol epoxides are now well established as principal ultimate carcinogenic metabolites of the polycyclic aromatic hydrocarbon class of carcinogens. Three metabolic steps are involved in the conversion of a polycyclic aromatic hydrocarbon into its bay-region diol epoxides: formation of an arene oxide, which is catalyzed by cytochrome P-450; hydrolysis of the arene oxide to the trans-dihydrodiol, a step that requires epoxide hydrolase; a second oxygenation by cytochrome P-450 to yield the diol epoxide. In parenchymal, endothelial, and Kupffer cells, no significant activation of BP was observed, a fact that can be attributed to the very low levels of cytochrome P-450 forms catalyzing the metabolic steps in these cells (8). After administration of Aroclor 1254, BP was activated by all three cell types, this effect being accompanied by a strong increase in the cytochrome P-450 content of parenchymal and nonparenchymal cells (8). The striking ability of parenchymal, endothelial, and Kupffer cells from untreated rats to activate DDBP so effectively (compared to BP) may be due to the fact that the dihydrodiol is the immediate precursor of the highly mutagenic diol epoxide, the amount of cytochrome P-450 present in untreated cells being sufficient to activate this compound with a high degree of efficiency. Alternatively, the conversion of the dihydrodiol to the diol epoxide can be catalyzed by prostaglandin endoperoxide synthetase during the oxidation of arachidonic acid by prostaglandins (9,10). Present studies are aimed at elucidating the possible role of prostaglandin endoperoxide synthetase in the activation of dihydrodiols derived from polycyclic aromatic hydrocarbons by isolated rat liver cells. DDCH was activated by parenchymal cells from untreated and Aroclor 1254-pretreated rats, while only endothelial and Kupffer cells from induced animals were able to activate the dihydrodiol. DDCH is a rather poor substrate for the cytochrome P-450 monooxygenase system: it is metabolized by liver microsomes from 3-methylcholanthrene-pretreated rats at about 8% of the rate at which DDBP is metabolized (11). Thus, the observation that endothelial and Kupffer cells from untreated animals were unable to activate DDCH might again be ascribed to the very low amount of cytochrome P-450 in these cells. AFB1 was activated by parenchymal and nonparenchymal cells isolated from untreated rats; this finding suggests that a constitutive form(s) of cytochrome P-450 is involved in the activation of the mycotoxin. In accordance with this proposal, it has recently been shown that four cytochrome P-450 isoenzymes purified from untreated animals were able to activate AFB1 in a reconstituted monooxygenase system (12); among them was cytochrome P-450h, a male-specific form of cytochrome P-450. Pretreatment of the rats with Aroclor 1254 significantly decreased the parenchymal cellmediated mutagenicity ofAFB1, whereas the mutagenic potential of the hepatotoxin in the presence of induced endothelial and Kupffer cells was slightly enhanced when compared to control cells. Furthermore, the formation rate of 2a-hydroxytestosterone, which reflects the levels of cytochrome P-450h, was strongly reduced in parenchymal cells of induced rats, while it remained unchanged in nonparenchymal cells. Thus, cytochrome P-450h seems to be involved in the activation of AFB1.
However, as mentioned earlier, other constitutive cytochrome P-450 forms might also mediate the activation of the mycotoxin. Hepatic S-9 preparations from newborn male rats were able to do so (data not shown), although cytochrome P-450h was not detected in these livers. On the other hand, the strong reduction in the parenchymal cell-mediated mutagenicity of AFB1 after administration of Aroclor 1254 might be due (at least in part) to the induction of AFB1-4-hydroxylase (13). This enzyme catalyzes the conversion of AFB1 to aflatoxin M1, the latter compound being with and without further metabolism less than 5% as active as AFB1.
In the present study liver cells were homogenized and a NADPH-generating system was added to the incubation medium. The homogenization leads to the dilution of several cofactors (e.g., glutathione), whereas the addition of NADP+ and glucose-6-phosphate fortifies the cytochrome P-450-dependent monooxygenase system. Taking into account that the major detoxification route of AFB1-8,9-epoxide is its conjugation with glutathione, catalyzed by glutathione S-transferase (14), the result of mutagenicity assays might be different if whole cells are used. Present experiments are aimed at clarifying this issue.
In conclusion, this study shows that rat liver parenchymal as well as endothelial and Kupffer cells are able to activate several carcinogens to mutagenic metabolites. Interestingly, DDBP is activated by all three cell types even in the absence of an NADPH-generating system, while constitutive cytochrome P-450 forms seem to be involved in the activation of AFB1.