Tissue-specific activities of methylated dibenzo[c,g]carbazoles in mice: carcinogenicity, DNA adduct formation, and CYP1A induction in liver and skin.

The potent multitissue carcinogen 7H-dibenzo[c,g]carbazole and nine methylated derivatives, synthesized on the basis of the positions in the parent compound that are involved in metabolism, were tested for their ability to induce sarcomas and hepatic tumors in XVIInc/Z mice. In addition, the capacity of these compounds to induce DNA adducts in skin and liver was investigated by (32)P-postlabeling analysis after their topical administration. Induction by these compounds of cytochromes P450 of the 1A family in liver and skin was investigated and correlated to their carcinogenic potential. A clear correlation was found between the tissue specificity of DNA binding and the capacity of each compound to induce skin or liver tumors. In contrast, no direct relationship was observed between the capacity of the compounds to induce cytochromes 1A1/1A2 and the tissue specificity of carcinogenesis or DNA binding in liver or skin. The results are discussed with respect to the positions of methyl groups in the 7H-dibenzo[c,g]carbazole molecule.

Inversion of genetic predisposition to carcinogenesis in liver of two lines of mice selected for resistance (Car-R) or susceptibility (Car-S) to skin carcinogenesis.

Two lines of mice, one resistant (Car-R) and one susceptible (Car-S) to skin carcinogenesis, were produced by bi-directional selective breeding. To see whether the characteristics of susceptibility or resistance to tumorigenesis were also expressed in the liver and lung, the two lines were submitted comparatively to treatment with 5,9-dimethyl dibenzo[c,g]carbazole (DiMeDBC), a potent hepatocarcinogenic derivative of the ubiquitous heterocyclic carcinogenic pollutant, 7H-dibenzo[c,g]-carbazole (DBC). An inversion of genetic predisposition to carcinogenesis in liver was observed. Car-R animals displayed rapid tumorigenesis in 100% of cases while Car-S mice were remarkably less sensitive, showing a 4-fold lower mean tumor multiplicity and a 4-month longer latency time. In parallel adduct formation by DiMeDBC and DBC in liver DNA was analyzed by the 32P-postlabeling method, showing a remarkably higher level in Car-R mice than in Car-S animals. These data indicate that tissue-specific sensibility in carcinogenesis may involve gene expression at various levels.

Tissue-specific differences in adduct formation by hepatocarcinogenic and sarcomatogenic derivatives of 7H-dibenzo[c,g]carbazole in mouse parenchymal and nonparenchymal liver cells.

Parenchymal (PC) and nonparenchymal (NPC) liver cells have different tissue-specific, procarcinogen activation enzymes. NPC appear to be protected against the mutagenic effects of lipotropic bulky adducts induced by carcinogens by a unknown mechanism. Most studies of activation have been conducted with whole liver. The purpose of this study was to differentiate adduct formation in mouse PC and in NPC, isolated after in vivo administration of 7H-dibenzo(c,g)carbazole (DBC), the most efficient liver carcinogen in mice, which also has potent sarcomagenic and epitheliomagenic activities. The very sensitive 32P-postlabeling method was used to detect adducts. Two tissue-specific DBC derivatives, 6-methoxy-DBC (6MeODBC), which is exclusively sarcomagenic, and 5,9-dimethyl-DBC (DiMeDBC), which is exclusively hepatocarcinogenic, were analyzed in parallel. Both PC and NPC generated the ultimate metabolites of DBC, but NPC were substantially less efficient. Clear-cut tissue-specific differences in adduct formation were established: the sarcomagenic 6MeODBC gave rise only to NPC-DNA adducts, and the hepatocarcinogenic DiMeDBC only to PC-DNA adducts. The chromatograms of the adducts were compared with those of mouse embryonic cells in culture and mouse epidermal cells. The results are discussed in connection with animal experiments with DBC, 6MeODBC, and dimethylbenzanthracene and with published data on PC and NPC activating enzymes.

Interaction of 7H-dibenzo[c,g]carbazole and its organspecific derivatives with hepatic mitochondrial and nuclear DNA in the mouse.

The recent observation of a high level of adducts in mitochondrial DNA (mtDNA) of cells exposed to chemical carcinogens aroused new interest in the hypothesis that carcinogen-induced damage in mitochondria plays a role in one or more stages of carcinogenesis. In order to investigate whether differences in the metabolic activation of carcinogens have qualitative and quantitative effects on ml- and nuclear DNA (nuDNA) adduct formation, mice were exposed to the potent hepatocarcinogenic and sarcomagenic polycyclic hydrocarbon 7H-dibenzo[c,g]carbazole (DBC) and to three of its derivatives that show large differences in enzymatic activation: N-acetyl-DBC (N-AcDBC), which is carcinogenic for several tissues; 5,9-dimethyl-DBC (DiMeDBC), which is exclusively hepatocarcinogenic; and N-methyl-DBC (N-MeDBC), which is exclusively sarcomagenic. Adduct formation and toxic effects were measured over 48 hr. With a moderate 5 mumol/kg dose of DBC, the adduct level in liver 24 hr after treatment was always higher in nuDNA than in mtDNA; after 48 hr a substantial increase in the level of adducts in mtDNA was observed, with a parallel decrease in the level in nuDNA. With DiMeDBC, a 4.9-fold increase in mtDNA was seen at 48 hr, whereas, at the same dose, the non-hepatocarcinogenic N-MeDBC induced a very small number of adducts. In order to obtain a nearly identical level of adducts in nu- and mtDNA at 24 hr, the dose of DBC must be three times higher (15 mumol/kg); this and higher dose levels had a strong cytotoxic effect in liver cells. Qualitative differences in adduct distribution were observed on chromatograms of mtDNA and nuDNA, showing that the access to mtDNA is a complex process. Our results confirm that mouse liver mtDNA is a major target for DBC and its hepatocarcinogenic derivatives. The possible interference of genotoxic alterations in mtDNA with carcinogenic mechanisms is discussed.

Lack of promoting activity of 7-methoxy-2-nitro-naphtho-[2,1-b]furan (R7000) in a medium-term rat liver bioassay.

Among the nitro-naphthofurans, 7-methoxy-2-nitro-naphtho[2,1-b]furan (R7000) has proved to be a very potent mutagen that causes sarcomas at the subcutaneous injection site and carcinomas in the forestomach after p.o. administration. In the present study, possible promoting activity for diethylnitrosamine (200 mg/kg, i.p.)-initiated liver carcinogenesis was assessed in rats. R7000 dissolved in 5% ethanol solution at doses of 10 or 50 mg/l given as the drinking water for 6 weeks did not enhance development of pre-neoplastic glutathione S-transferase placental form positive foci. Thus, it was concluded that R7000 may not be carcinogenic for the rat liver under the present experimental conditions.

32P-postlabeling analysis of inhibition by norharman of formation of dibenzo[a,e]fluoranthene--DNA adducts in mouse embryo fibroblasts.

Quantitative and qualitative changes in the inhibition of DNA adduct formation in the presence of increasing concentrations of norharman (NH) were investigated in vivo in mouse fibroblasts treated with dibenzo[a,e]fluoranthene (DBF), a potent carcinogen in mice. The nuclease P1 modification of the 32P-postlabeling technique was used to identify adducts. A dose-dependent reduction in DBF-DNA adduct formation was observed: an 80% reduction with 0.06 mM NH and 90% with 0.12 mM NH. At 0.12 mM NH, all of the spots coming from hydroxylated DBF vicinal dihydrodiol (DHD) epoxides were missing; the only clear spot was that of the major DBF adduct produced by the ultimate DBF metabolite, DBF-3,4-DHD-1,2 oxide. Spots representing other DBF-DHD epoxide adducts appeared only in trace amounts. These results can be interpreted as a dose-dependent competition or inhibition of some secondary metabolic step, most probably secondary epoxidation; however, a direct protective effect of NH during adduct formation cannot be excluded. NH is a strong inhibitor of DBF-DNA adduct formation in vivo.

Differences in metabolic activation of dibenzo[a,e]fluoranthene characterized by 32P-postlabeling in two mouse fibroblast models.

The formation of DNA adducts was investigated in mouse fibroblasts from two different tissues--embryos and adult lung--after incubation with dibenzo[a,e]fluoranthene (DBF) or its major proximate metabolites. The nuclease P1 modification of the 32P-postlabeling method was adapted for detection of DBF-DNA adducts. Quantitative and qualitative differences were observed in the metabolic activation mediated by the two cell types. DBF-DNA adducts generated three major spots reproducibly, and more than ten spots of medium or weak importance. The highest level of DNA binding occurred via the DBF-bay region vicinal dihydrodiol epoxide but with significant differences in the quantitative distribution of adducts. Striking qualitative differences were observed when lung fibroblasts were incubated with the DBF-pseudo bay region dihydrodiol (DBF-12,13-DHD). The spots representing adducts induced in embryo fibroblasts by DBF-3OH-12,13-DHD, a further metabolite of DBF-12,13-DHD, were totally absent from chromatograms of lung cells. These results show that both embryo and lung fibroblasts can activate DBF but that different cytochrome P-450 forms and substrate affinities are involved. The finding that different activation systems may be present in subcategories of the same tissue, may provide a partial explanation for the wide variations in sensitivity to carcinogens among species, organs and tissues.

Isolation and partial characterization of a cytochrome P-450 isoenzyme (cytochrome P-450tu) from mouse liver tumors.

Experimental hepatomas induced with 5,9-dimethyldibenzo[c,g]carbazole in female XVIInc/Z mice display a strong microsomal steroid 15 alpha-hydroxylation activity. A cytochrome P-450 isoenzyme (cytochrome P-450tu), specific for this activity, has been isolated by an HPLC derived method using various Fractogel TSK and hydroxyapatite supports. On SDS polyacrylamide gel electrophoresis the purified protein appeared as one major band with an apparent Mr of 50,000. Its specific cytochrome P-450 content was 7.55 nmol/mg protein. As deduced from the visible spectrum, the heme iron of the isolated P-450tu was to 72% in the high-spin state. The CO-bound reduced form showed an absorption maximum at 450 nm. In addition to the stereospecific 15 alpha-hydroxylation of progesterone (2.3 min-1) and testosterone (2.5 min-1), the enzyme catalyzed also 7-ethoxycoumarin O-deethylation, benzphetamine N-demethylation and aniline 4-hydroxylation. Its N-terminal amino-acid sequence (21 residues) was identical to that of cytochrome P-450(15) alpha, isolated by Harada and Negishi from liver microsomes of 129/J mice. P-450tu differed from P-450(15) alpha by its higher molecular weight, its 40-times lower steroid 15 alpha-hydroxylation and its 4-times higher benzphetamine N-demethylation.

32P-post-labeling analysis of DNA adducts in mouse embryo fibroblasts treated with dibenzo[a,e]fluoranthene and its major metabolites.

The formation of DNA adducts was investigated in mouse fibroblasts treated with dibenzo[a,e]fluoranthene (DBF), using the nuclease P1 modification of the 32P-post-labeling method. In order to separate the poorly soluble, bulky DNA adducts of this potent sarcomogenic, six-ring polycyclic aromatic hydrocarbon, several modifications of the method were introduced. Chromatographic spots were identified by incubating fibroblasts with the four major proximate metabolites of DBF and observing the co-migration of adducts with those of DBF. DNA-DBF adducts chromatographed very reproducibly in three major spots and in greater than 10 spots of medium or low importance. The most prominent spots, 2 and 3, were present characteristically after incubation of cells with the DBF-bay region dihydrodiol (+/- -trans-3,4-dihydro-3,4-dihydroxyDBF; DBF-3,4-DHD). Incubation with the DBF pseudo-bay region dihydrodiol (+/- -trans-12,13-dihydro-12,13-dihydroxyDBF; DBF-12,13-DHD) gave rise to a more complex pattern of nine spots, two of which, spots 4 and 5, were prominent. Direct in vitro reaction between DNA and the synthetic anti-isomer of the DBF-bay region DHD epoxide yielded adducts in spots 2 and 3, while the DBF-anti-pseudo-bay region DHD epoxide yielded adducts in spots 4 and 5. Peripheral, fast-migrating spots present in the DBF chromatogram were identified as adducts of DBF-7OH-3,4-DHD and DBF-3OH-12,13-DHD. Major spot 1 was present in all DBF chromatograms but not after incubation with the DBF bay and pseudo-bay region proximate metabolites. Its probable origin as a non-bay region epoxide reaction is discussed. In previous experiments, the physicochemically very similar DBF-bay region and pseudo-bay region tritium-labeled adducts co-eluted in HPLC as a single peak. 32P-Post-labeling analysis allowed reproducible separation of DBF-DNA adducts and showed in addition the existence of several new adducts models of DBF. Quantification of DBF adducts made it possible to identify the DBF-bay region DHD epoxide and the metabolites responsible for spot 1 adducts as the major ultimate DBF metabolites in fibroblasts.

Particular cytochrome P-450-dependent steroid metabolism: a new class of mouse liver tumor markers.

An altered pattern of cytochrome P-450-dependent microsomal steroid metabolism was identified in female mouse liver tumors induced by 5,9-dimethyldibenzo[c,g]carbazole, a potent organo-specific liver carcinogen. These tumor tissues were compared to extratumoral liver parenchyme, to normal, fetal and neonatal livers and to spontaneous liver tumors, the frequency of which is very low in the highly hybridized mouse strain (XVIInc/Z) used for liver tumorigenesis. Cytochrome P-450-dependent steroid hydroxylase activities were measured by the identification and quantification of four monohydroxyprogesterone and eight monohydroxytestosterone metabolites. In contrast to a general decrease (50%) of total P-450 in tumor microsomes, the individual steroid hydroxylases were regulated differently. Progesterone 16 alpha- and testosterone 6 alpha-, 6 beta-, 7 alpha- and 16 alpha-hydroxylase activities were decreased 50%, and more, whereas progesterone and testosterone 15 alpha-hydroxylase activities were raised 3-4 times with regard to microsomal protein content and 6-7 times with regard to total P-450. Consequently the most prominent feature of the steroid metabolism by tumor-borne microsomes is the hydroxylation at the 15 alpha-position. Furthermore, minor testosterone 2- and 15 beta-hydroxylase activities showed equally an increase of approximately 4 times (8 times with regard to total P-450). The observed new tumoral pattern of P-450-dependent microsomal steroid metabolism appearing characteristically in spontaneous and chemically induced liver tumors indicates that particular P-450 enzymes are strongly expressed in mouse liver tumors. These enzymes may be used as markers for early stages in liver tumorigenesis.

Structure and carcinogenicity of dibenzo(c,g)carbazole derivatives.

The carcinogenicity of several groups of carcinogens is evoked with particular reference to Dibenzo(c,g)carbazole derivatives. The activity of these derivatives is discussed with respect to their species and organ specificity. The enzymatic equipment is decisive as to whether the compounds formed can react with DNA or are simply detoxified and eliminated. All these carcinogens are complete carcinogens, i.e. they have the property of both initiation and promotion.

Non-random distribution of dibenzo[a,e]fluoranthene-induced DNA adducts in DNA loops in mouse fibroblast nuclei.

The three-dimensional distribution of nuclear DNA damage induced by dibenzo(a,e)fluoranthene (DBF), a potent carcinogen for mouse fibroblasts, has been examined. The intact supercoiled nuclear DNA obtained from nucleoids of mouse fibroblasts incubated with DBF was fractionated into loop DNA attached to the matrix (10%) and bulk loop DNA (90%). Preferential binding of DBF to the DNA of the extremities of loops, which are rich in regulatory sequences, was observed in all experiments. An increase of the preferential DBF binding was seen when fibroblasts were incubated with both DBF and novobiocin or hydroxyurea. The excess damage seen in loop DNA attached to the cage may be due to the kinetics of diffusion to the interior of the nucleus of hydrophobic DBF metabolites accumulated in lipid-rich nuclear membrane.

Organ-specific, carcinogenic dibenzo[c,g]carbazole derivatives: discriminative response in S. typhimurium TA100 mutagenesis modulated by subcellular fractions of mouse liver.

7H-Dibenzo[c,g]carbazole (DBC) has carcinogenic effects on mouse subcutaneous fibroblasts and liver; the N-methyl derivative (N-MeDBC) induces only sarcomas; 3-methyl- and 5,9-dimethyldibenzo[c,g]carbazole (3-MeDBC and 5,9-DMeDBC) are specific, potent hepatocarcinogens in mice. The mutagenicity in S. typhimurium TA100 of these 4 compounds was evaluated in relation to the concentration of mouse liver 9000 X g supernatant (S9) and to the proportions of microsomes and cytosol in the medium. Optimal mutagenicity of N-MeDBC was seen with a low concentration of S9 or microsomes; a 5-10 times higher concentration of the subcellular fraction was necessary to induce optimal mutagenicity of the hepatocarcinogens 3-MeDBC and 5,9-DMeDBC. Intermediate quantities were needed in the case of DBC, which is carcinogenic in both cell types. Whereas the presence of cytosol had an inhibitory effect on the mutagenicity of the sarcomagenic N-MeDBC, the cytosolic fraction was essential for optimal mutagenic expression by the 2 hepatocarcinogenic derivatives. The activating cytosolic fraction is not inducible. These experiments show that programmed modulation of the metabolic activation system in the Ames test can be used to study organ-specific chemical carcinogenesis. The results suggest that differences in the enzymatic composition of target tissues are a determining factor in the organ specificity of carcinogens such as DBC.

Formation and persistence of DNA-protein cross-links induced in mouse fibroblasts by dibenzo[a,e]fluoranthene, and modulation by stimulators and inhibitors of polycyclic aromatic hydrocarbons (PAH) metabolism.

The production by dibenzo[a,e]fluoranthene (DBF) of DNA-protein cross-links in cultured mouse fibroblasts is probably mediated by the activation of proximate metabolites of DBF and not by the DBF molecule itself. In order to test this hypothesis, several agents that enhance or reduce production of the DBF metabolite putatively involved in cross-linking were tested. Increasing NADPH concentrations in the medium enhanced cross-link production; 1,2-epoxy-3,3,3-trichloropropane (TCPO), an inhibitor of epoxide hydrolases, slightly reduced DNA-protein cross-link formation at high concentrations; norharman (NH), an inhibitor of certain steps in the metabolism of DBF, totally blocked cross-linking. The possible involvement of DBF-bisdihydrodiol, a bifunctional metabolite identified in vitro, is discussed. Postincubation in DBF-free medium did not induce a significant reduction in cross-links, indicating that repair did not take place.

Marked differences between mutagenicity in Salmonella and tumour-initiating activities of dibenzo[a,e]fluoranthene proximate metabolites; initiation inhibiting activity of norharman.

Dibenzofluoranthene-12,13-dihydrodiol (DBF-12,13-DHD) is six times more mutagenic in Salmonella TA100 than dibenzofluoranthene-3,4-dihydrodiol (DBF-3,4-DHD). However, these two major dibenzo[a,e]fluoranthene (DBF) proximate metabolites, which are immediate precursors of the corresponding diolepoxides, showed on an equimolar basis nearly identical initiation activities on mouse skin; they induced three times more papillomas than the parent hydrocarbon. On the other hand the epithelioma initiation capacities, i.e. the number of papillomas progressing to malignant tumours, of DBF or the two dibenzofluoranthene dihydrodiols were equivalent. Norharman, a putative vicinal diolepoxidation inhibitor in DBF metabolism when administered topically together with the initiation dose (100 nmol), strongly inhibited the induction of tumours by DBF-3,4-DHD and DBF. The relationship between in vitro mutagenic activity in Salmonella and the carcinogenicity of DBF metabolites in mice appears to be qualitative rather than quantitative.

Formation and removal of dibenzo[a,e]fluoranthene-DNA adducts in mouse embryo fibroblasts.

In vivo binding of dibenzo[a,e]fluoranthene (DBF) to mouse embryo fibroblast DNA was compared with that observed previously in vitro on calf thymus DNA incubated with mouse liver microsomes. The h.p.l.c. elution patterns of the adducts formed by DBF metabolites with DNA and obtained in vivo at the optimal exposure time of 42-48 h were qualitatively very similar to the patterns obtained in vitro, but their amplitude was quantitatively reduced. There are two striking differences between the in vivo and in vitro results. Firstly, the most polar peak A, very abundant in vitro, was absent in vivo. Secondly, the reactivity of the two major proximate metabolites of DBF, the bay and pseudo-bay region dihydrodiols, was very different in intact cells compared with the results in vitro. When incubated in vitro, pseudo-bay region dihydrodiol DBF was twice as reactive as bay region dihydrodiol DBF. The opposite reactivities were observed in vivo. The major DBF-DNA adducts formed in vivo were collected in the peaks E, B and C. The predominant peak E contained DNA adducts of both bay and pseudo-bay region dihydrodiolepoxides which are the major ultimate metabolites of DBF in vivo and in vitro. The other two prominent peaks B and C contained DNA adducts of 3-hydroxy DBF pseudo-bay region dihydrodiolepoxide and the 7-hydroxy DBF bay region dihydrodiolepoxide, respectively. After adduct formation, post incubation of fibroblasts for a further 48 h, in the absence of DBF, eliminated half the amount of adducts present. Peak B adducts were repaired more efficiently than those of peaks E, C D and F. The carcinogenic initiating activity of DBF appears to be a complex process in which several DNA adducts play a role.

Sexual differences in the expression of gamma-glutamyl transpeptidase during 5,9-dimethyldibenzo[c,g]carbazole-induced hepatocarcinogenesis in mice.

5,9-Dimethyl-7H-dibenzo[c,g]carbazole (5,9-diMeDBC) is an organ-specific carcinogen for mouse liver. It induces malignant hepatomas in 100% of XVIInc/Z and C57BL mice with a cumulative dose of only 12 mumol administered subcutaneously. Large sex-related differences in sensitivity towards this carcinogen were observed: entire males were much less sensitive than females. Castration of males nearly completely restored the sensitivity observed for females. The kinetics of appearance of gamma-glutamyl transpeptidase (gamma GT)-positive foci was enhanced in females and castrated males and was correlated with the tumour outbreak. On the contrary the expression of gamma GT in males was very reduced. The sex-dependent sensitivity seems to be regulated by a complex interplay of endogenous factors.

Possible involvement of a vicinal, non-bay-region dihydrodiol-epoxide in the activation of dibenzo[a,e]fluoranthene into bacterial mutagens.

Dibenzo[a,e]fluoranthene, its 7-hydroxy, 3,4- and 12,13-dihydrodiol metabolic derivatives as well as three synthetic, structurally related hydrocarbons, were tested for mutagenicity towards Salmonella typhimurium TA100 strain in the presence of 3-methylcholanthrene-treated rat and mouse liver post-mitochondrial supernatants. Of these compounds, the 12,13-dihydrodiol showed the highest activity, being 6-10 times more mutagenic than the parent compound. Our data, in conjunction with those of previous studies on the liver microsomal metabolism and DNA binding of dibenzo[a,e]fluoranthene and its dihydrodiols, indicate that activation of dibenzo[a,e]fluoranthene to bacterial mutagens may occur predominantly through a vicinal, non-bay-region 12,13-dihydrodiol epoxide.

The metabolic activation of dibenzo[a,e]fluoranthene in vitro. Evidence that its bay-region and pseudo-bay-region diol-epoxides react preferentially with guanosine.

Dibenzo[a,e]fluoranthene ( DBF ), a non- alternant carcinogenic polycyclic aromatic hydrocarbon (PAH), binds covalently to DNA. The main adducts were characterized as covalent additions of its bay-region and pseudo-bay-region diol-epoxides. The structure of these 2 adducts was analyzed by mass spectrometry using their persilyl derivatives. 3,4-Dihydroxy-1,2-epoxy-1,2,3,4-tetrahydro- DBF (3,4-diol-1,2-epoxy- DBF ) and 12,13-dihydroxy-10,11-epoxy-10,11,12,13-tetrahydro- DBF (12, 13-diol-10,11-epoxy- DBF ) obtained by synthesis were allowed to react in vitro with calf thymus DNA or with poly(G). The comigration of DNA and poly(G) adducts isolated after acid hydrolysis of DNA and poly(G) was in good agreement with mass spectroscopic results: both bay-region and pseudo-bay-region DBF diol-epoxides reacted with guanine residues.

In vitro metabolism of N-methyl-dibenzo [c,g]carbazole a potent sarcomatogen devoid of hepatotoxic and hepatocarcinogenic properties.

The metabolism of N-methyl substituted 7H-dibenzo[c,g]carbazole (N-Me DBC) was investigated in vitro using liver microsomes from 3-methylcholanthrene (MC)-, benzo[c]carbazole (BC) and Arochlor-pretreated mice and rats. N-Me DBC is a potent sarcomatogen devoid of hepatotoxicity and liver carcinogenic activity. The ethyl acetate-extractable metabolites were separated by high performance liquid chromatography (HPLC) and most of them were identified by proton magnetic resonance (PMR), mass spectrometry (MS) and comparison with synthetically prepared specimens. Mouse and rat microsomes gave rise to the same metabolites. The major metabolites were 5-OH-N-Me DBC (50%), N-hydroxymethyl (HMe) DBC (25-30%) and 3-OH-N-Me DBC (10%). Addition of 1,1,1-trichloropropene-2,3-oxide (TCPO) to the standard incubation medium permitted the identification of two dihydrodiols among the minor metabolites. No metabolite of DBC was observed after incubation of N-Me DBC, or its major metabolite N-HMe DBC, with either mouse or rat microsomes, but the possibility of a slight demethylation cannot be totally excluded. The lack of biotransformation at the nitrogen atom site may explain the lack of hepatotoxicity and liver carcinogenic activity of N-Me DBC. The modulation of metabolism by epoxide hydrolase, cytosol and glutathione was also investigated. The results are discussed in the light of data previously obtained with hepatotoxic and hepatocarcinogenic DBC.