Chem. Res. Toxicol. 1998, 11, 925-936
925
Activation of Heterocyclic Aromatic Amines by Rat and Human Liver Microsomes and by Purified Rat and Human Cytochrome P450 1A2† Robert J. Turesky,*,‡ Anne Constable,‡ Janique Richoz,‡ Nathalia Varga,‡ Jovanka Markovic,‡ Martha V. Martin,§ and F. Peter Guengerich*,§ Nestle´ Research Center, Nestec Ltd., Vers-chez-les-Blanc, 1000 Lausanne 26, Switzerland, and Department of Biochemistry and Center in Molecular Toxicology, Vanderbilt University School of Medicine, Nashville, Tennessee 37232-0146 Received February 5, 1998
The dietary mutagens 2-amino-3,8-dimethylimidazo[4,5-f]quinoxaline (MeIQx) and 2-amino1-methyl-6-phenylimidazo[4,5-b]pyridine (PhIP) are activated to genotoxins by rat and human liver cytochrome P450 (P450) 1A1- and 1A2-mediated N-oxidation. Immunoquantitation of 51 human liver samples revealed a wide range in P450 1A2 expression (10-250 pmol/mg of microsomal protein, median 71 pmol/mg), with 39% of the livers containing >100 pmol/mg of protein. There was no evidence for expression of P450 1A1 (85%, there was still significant P450-mediated HAA N-oxidation, and rat P450 1A1 appears to significantly contribute to the N-oxidation of both MeIQx and PhIP. These data are consistent with previous reports where rat P450 1A1 was found to significantly contribute to N-oxidation of PhIP (38) and to the bacterial mutagenicity of MeIQx following activation with microsomes of rats pretreated with 3-MC, where P450 1A1 was estimated to account for 50% of the activity (53). Genotoxicity Assays. Genotoxicity assays were done with the umu test with MeIQx activation by several different human and rat liver microsomal preparations. Three human liver samples were chosen, which showed high (HL-G), intermediate (HL-136), and low (HL-100) rates of MeIQx N-oxidation, and compared to liver microsomes prepared from Sprague-Dawley rats pretreated with Aroclor 1254, Fischer-344 rats pretreated with 3-MC, and untreated Fischer-344 rats. HL-G and HL-136 microsomes were more effective in inducing the umu response than any of the rat microsomal preparations; activity induced by sample HL-100 was weak and comparable to the activity of liver microsomes prepared from untreated rats (Figure 5). Mutagenicity assays were also done using the Ames reversion test with MeIQx and PhIP, as this latter HAA shows very weak biological effects in the umu test. Four human liver microsomes (HL-G, HL-28, HL-31, and HL111), which were active in HAA N-oxidation, the weakly active HL-100 microsome containing 500 µM. On the basis of the P450 inhibition studies with furafylline and ANF, this enzyme appears to be P450 1A1 (15) (Table 2 and Figure 4).
932 Chem. Res. Toxicol., Vol. 11, No. 8, 1998
Turesky et al.
chromatographic properties and chemical instability of 5-HO-MeIQx, kinetic measurements of formation could not be made with purified P450s or microsomes. For comparison, the kinetic parameters of MROD, a substrate for both rat and human P450 1A2 (40, 46), were determined with purified rat P450 1A2 and recombinant human P450 1A2. The nonlinear regression analyses revealed that both enzymes display similar Km and kcat values for MROD activity (Table 3). Therefore, the catalytic efficiencies and differences in activities between rat and recombinant human P450 1A2 are highly dependent upon the chemical structure of the substrates.
Discussion
Figure 6. Mutagenicity of MeIQx (A) and PhIP (B) in the Ames reversion assay with four active human liver microsome samples containing 130-190 pmol of P450 1A2/mg of protein (HL-G, HL28, HL-31, HL-111), a weakly active human liver microsome containing 35 pmol/mg of microsomal protein). P450 1A1 was not detected in human liver, and it was present at levels 500 µM for accurate determination.
female rat livers, respectively. Oxidation of HAAs to mutagenic products by human liver microsomes was generally low and comparable to rat microsomes. These findings differ from our results, where activation of MeIQx was generally far greater in human liver microsomes than untreated rat liver microsomes. The low enzyme activities reported in their study (21) suggests that denatured P450 was present at signficant levels because of unsuitable isolation or storage conditions after death. Other investigations on microsomal activation of HAAs to genotoxins showed that the majority of human microsomes displayed activities greater than or equal to microsomes of untreated rats but were considerably less active than microsomes of rats pretreated with PCB or 3-MC. Several reports of high HAA activity in human microsomes also appear in the literature (56, 57). In one study, two of the four human liver microsomes examined were more active in inducing the umu response by IQ than microsomes of rats pretreated with ISF (57), suggesting that this analogue of MeIQx is also activated more efficiently by human P450 1A2 than rat P450s 1A1 and 1A2. Another study reported that several human liver microsomes transformed grilled beef extract containing HAAs to bacterial mutagens at levels comparable to microsomes of rats pretreated with PCB (56). The rat P450 1A1 and 1A2 enzymes are respectively 79% and 75% identical in amino acid sequence to their human counterparts (58, 59). Despite these strong similarities, the changes in amino acid sequence can have a profound effect on catalytic activities (13) and the regioselectivity of P450-mediated oxidation of MeIQx and PhIP, as seen with rat and human microsomes and purified P450s. Rat P450s 1A1 and 1A2 account for >95% of MeIQx and PhIP N-oxidation (activation) in liver microsomes of PCB-pretreated rats. On the basis of inhibition of product formation by ANF and furafylline (Table 2, Figure 4), combined with P450 data (Table 3), we conclude that both rat P450s 1A1 and 1A2 also catalyze ring oxidation (detoxication) of PhIP and MeIQx
to a lesser extent (Scheme 1). The P450-catalyzed detoxication of these HAAs in rat microsomes is significant and accounts for approximately 30-50% of the metabolism. In human microsomes, where P450 1A1 is not expressed (17, 18), N-oxidation accounts for >95% of the metabolism of these HAAs while ring oxidation of MeIQx at the C-5 position is not detected and 4′hydroxylation of PhIP accounts for
