1218
Chem. Res. Toxicol. 2007, 20, 1218–1224
In Vitro Bioactivation of 3-(N-Phenylamino)propane-1,2-diol by Human and Rat Liver Microsomes and Recombinant P450 Enzymes. Implications for Toxic Oil Syndrome Anna Martínez-Cabot,† Anna Morató,† Jan N. M. Commandeur,‡ Nico P. E. Vermeulen,‡ and Angel Messeguer*,† Department of Biological Organic Chemistry, IIQAB, CSIC, J. Girona, 18-26, 08034 Barcelona, Spain, and Leiden/Amsterdam Centre for Drug Research, DiVision of Molecular Toxicology, Department of Pharmacochemistry, Vrije UniVersiteit, De Boelelaan 1083, 1081 HV Amsterdam, The Netherlands ReceiVed June 8, 2007
Toxic oil syndrome (TOS) was a massive food-borne intoxication that occurred in Spain in 1981. Epidemiological studies imputed 3-(N-phenylamino)propane-1,2-diol (PAP) derivatives as the toxic agents. The in vitro bioactivation of PAP by rat and human liver microsomes was studied. In both cases, 3-[N(4′-hydroxyphenyl)amino]propane-1,2-diol (1) was detected as the main metabolite. Inhibition studies with pooled human liver microsomes in the presence and absence of P450-specific inhibitors suggest that 2C8 and 2E1 are the main enzymes involved in PAP bioactivation, followed by 3A4/5, 1A1/2, and 2C9. Incubations of PAP with 10 different recombinant P450 enzymes showed that 2C8, 2C9, 2C18, 2D6, and 2E1 catalyzed PAP 4′-hydroxylation. Incubations of phenol 1 with rat and human liver microsomes in the presence of GSH resulted in the formation of a glutathione conjugate of a quinoneimine metabolite derived from 1. In rat liver microsomes, P450 enzymes play a key role in the bioactivation of 1, whereas in human liver microsomes, autoxidation appears to be the major mechanism. The implications of these results for toxic oil syndrome are discussed. Introduction 1
Toxic oil syndrome (TOS) was a serious food intoxication episode that occurred in 1981 in central and northern Spain, causing more than 400 deaths and affecting more than 20 000 people (1, 2). TOS was attributed to the ingestion of rapeseed oil that had been adulterated with aniline and then illegally refined for human consumption (3). Epidemiological investigation established 3-(N-phenylamino)propane-1,2-diol (PAP) derivatives as markers of the toxic oil, since they were found in only oil batches that caused intoxication (4). In particular, the diester derivatives of PAP (dPAPs) were imputed and are thought to have been generated during the refining process. These PAP fatty acid esters are converted into PAP by human pancreatic lipase (5) and have been shown in rats to be absorbed by the gastrointestinal tract and then metabolized similarly to phospholipids (6). Nevertheless, the causal agent of TOS remains unknown, principally because no animal model reproducing the full symptomatic spectrum of the syndrome has yet been developed. This suggests that TOS could exhibit species-specific toxicity for humans. Moreover, family members affected by TOS differed in their responses to the toxic oil (7, 8), suggesting a * To whom correspondence should be addressed: Institut d’Investigacions Químiques i Ambientals de Barcelona (CSIC), Jordi Girona, 18-26, 08034 Barcelona, Spain. Telephone: +34-93-400 61 21. Fax: +34-93-204 59 04. E-mail:
[email protected]. † IIQAB, CSIC. ‡ Vrije Universiteit. 1 Abbreviations: DDC, diethyl dithiocarbamate; ESI-HRMS, electrospray high-resolution mass spectrometry; FDA, Food and Drug Administration; HLM, human liver microsomes; PAP, 3-N-(phenylamino)propane-1,2-diol; dPAP, diacyl derivatives of PAP; P(1)PAP, monopalmitoyl ester at C-1 of PAP; P(2)PAP, monopalmitoyl ester at C-2 of PAP; RLM, rat liver microsomes; TOS, toxic oil syndrome.
dose factor and/or susceptibility trait. With respect to the latter, an immunological mechanism was initially proposed and extensively investigated (9–11). Thus, differences in xenobiotic metabolism as well as in inherited genes among exposed subjects may have accounted for differences in bioactivation and/or inactivation of the toxic species and, consequently, in different concentrations of toxic metabolites. Ladona et al. assessed genetic polymorphisms related to xenobiotic metabolism among TOS patients. They found that related patients had higher levels of arylamine N-acetyltransferase-2 (NAT2) than did unrelated patients, which suggested a correlation between impaired acetylation and susceptibility to intoxication by the oil (12). In vivo metabolism studies performed in our laboratory, in collaboration with the Ladona group, showed extensive bioactivation of racemic [14C]PAP after intraperitoneal administration in two different mice strains. The major metabolite that was detected was (2-hydroxy-3-phenylamino)propanoic acid, indicating that the main metabolic pathway for PAP was oxidation of the primary alcohol. Other metabolites resulting from oxidation of the aromatic ring, such as 3-[N-(4′-hydroxyphenyl)amino]propane-1,2-diol (1) and quinoneimine 2, were also detected. Furthermore, 4-aminophenol and paracetamol were found in the mice’s urine after treatment with glucoronidase and sulfatase. These two compounds could be generated from quinoneimine 2 after prototropic rearrangement and further hydrolysis to yield 4-aminophenol, which after in vivo acetylation would give rise to paracetamol (Scheme 1) (13, 14). These findings may have important implications for the toxicology of TOS, by providing evidence not only that PAP metabolism can generate well-known toxic products, such as 4-aminophenol and paracetamol, but also that potentially toxic intermediates, such as quinoneimine 2, could be involved in these bioactivations. Moreover, we recently reported that glutathione and N-acetyl-
10.1021/tx700209p CCC: $37.00 2007 American Chemical Society Published on Web 08/03/2007
In Vitro BioactiVation of PAP by P450 Scheme 1. Proposed Pathway for the in Vivo Bioactivation of PAP into 4-Aminophenol and Paracetamol (13, 14)
cysteine adducts formed with quinoneimine 2 are also unstable and can be easily reoxidized to the corresponding quinoneimine conjugate. This conjugate, in the absence of thiol reagents, can isomerize to the imine, which in turn can be hydrolyzed to yield the corresponding 4-aminophenol adduct and glyceraldehyde (15). These results suggest that this family of compounds may be toxic even when trapped as the respective thiol conjugates (16–20). The enzymes involved in the oxidative metabolism of PAP in TOS as yet have not been identified. Therefore, the aim of this study is to identify the enzymes by studying the oxidative bioactivation of PAP using rat and human liver microsomes and recombinant human P450 enzymes and determine their kinetic parameters. On the basis of the results, the possibility of P450-derived pharmacogenetic factors that could account for the differences in susceptibility to TOS among the exposed population will be discussed.
Materials and Methods Caution: As PAP deriVatiVes haVe been implicated in TOS, special precautions should be taken when handling these substances to aVoid potential risks (gloVes, masks, and Ventilated hood cabinets should be used when handling solids or solutions). Chemicals Reagents and Suppliers. Glycidol (Merck, 98%) was first distilled and then stored at -20 °C. Aniline, 4-aminophenol, DDC, tryptamine, zinc sulfate, and palmitic acid were obtained from Aldrich Chemical Co. (Milwaukee, WI). Ammonium acetate and activated palladium/carbon (10%) were obtained from Merck (Darmstadt, Germany). R-Naphthoflavone, quercetin, ticlopidine, sulfaphenazole, ketoconazole, NADP(H), glutathione, and taurodeoxycholic acid were obtained from Sigma Chemical Co. (St. Louis, MO). Quinidine was obtained from Alfa Aesar GmbH & CoKG (Karlsruhe, Germany). Human and Sprague-Dawley rat liver microsomes were obtained from Advancell (Barcelona, Spain). Recombinant human P450 enzymes were obtained from In Vitro Technologies (Baltimore, MD). Acetonitrile was of HPLC grade and purchased from VWR Prolabo (Leuven, Belgium), and isopropyl alcohol and n-hexane were also of HPLC grade and purchased from Merck. PAP, phenol 1, and glutathione conjugate 3 were synthesized as previously described (15, 21). General Methods and Instruments. HPLC analyses were performed on a Hewlett-Packard 1100 system equipped with a DAD detector and employing a Kromasil 100 C18 (25 cm × 0.4 cm, 5 µm) column (Scharlab). The HPLC eluents were 10 mM NH4OAc in water (A) and acetonitrile (B). The gradient consisted of an isocratic step (from 0 to 1 min) at 1 mL/min with 100% A, followed by a linear ramp to reach 95% A (from 1 to 15 min), then a second ramp to reach 60% A (from 15 to 25 min), and, finally, a return to the initial conditions. The detection was set at 300 nm. For direct phase conditions, a Luna CN 100A (15 cm × 4.6 cm, 5 µm) column (Phenomenex) was used. Incubation samples were centrifuged in a Biofuge pico centrifuge (Heraeus instruments, Osterode, Ger-
Chem. Res. Toxicol., Vol. 20, No. 8, 2007 1219 many). ESI-HRMS data were obtained with an Acquity UPLC system (Waters, Milford, CT) coupled to a model LCT Premier XE mass spectrometer (Waters) provided with a TOF analyzer in positive mode. An Acquity UPLC BEH C18 column (1.7 µm, 2.1 mm × 50 mm, Waters) was employed for the UPLC analyses. 1H NMR (400 MHz) and 13C NMR (100 MHz) spectra were recorded on a Varian Mercury 400 spectrometer. Spectra were recorded in neutralized CDCl3, and chemical shifts are given in parts per million related to tetramethylsilane for 1H. In Vitro Identification of Phase I Metabolites of PAP. Incubations with rat liver microsomes (total volume of 800 µL) were performed in 50 mM sodium phosphate buffer containing 5 mM MgCl2 (pH 7.4) in the presence of 64 µL of pooled RLM (20 mg of protein/mL) and 160 µL of 25 mM PAP (final concentration of 5 mM). The reactions were started by addition of 160 µL of 5 mM NADP(H) (final concentration of 1 mM). The samples were then incubated at 37 °C. Aliquots of 150 µL were withdrawn at different times, and the incubations were stopped by cooling the mixtures to 0 °C and adding 50 µL of 125 mM zinc sulfate. Denatured proteins were removed by centrifugation (10 min at 10 000 rpm), and the supernatants (100 µL) were analyzed by HPLC. The same procedure was followed for the incubations with human liver microsomes (HLM) except for the use of 32 µL of pooled HLM (38.4 mg of protein/mL) instead of RLM. In both cases, a new peak (compound 1) in the HPLC profile was observed (see Figure 1). 1: ESI-HRMS m/z (RLM) 184.0974 and m/z (HLM) 184.0974; for (M + H)+ C9H14NO3 requires m/z 184.0974. Kinetic Studies. The enzyme kinetic parameters Km and kcat were determined for the 4′-hydroxylation of PAP by rat and human liver microsomes following the same procedure described above, but using different PAP concentrations (0.125, 0.25, 0.5, 1, 2, and 4 mM) and withdrawing incubation aliquots at 3, 5, 7, and 10 min. All the experiments were performed in triplicate (one experiment per day on three different days). The data were fitted to the standard Michaelis–Menten equation, V ) Vmax/(1 + Km/[S]), where [S] is the substrate concentration, using GraphPad Prism 3.03 (GraphPad Software Inc., San Diego, CA). Inhibition Studies. Microsomal incubations (total volume of 250 µL) using pooled HLM (10 µL, 38.4 mg of protein/mL) were performed in 50 mM sodium phosphate buffer containing 5 mM MgCl2 (pH 7.4), 50 µL of 5 mM NADP(H) (final concentration of 1 mM), and 5 µL of each specific inhibitor dissolved in DMSO. The final concentrations used for the inhibitors were 20 times the corresponding Ki, according to FDA criteria (22), and were as follows: 0.2 µM R-naphthoflavone, 8 µM quinidine, 34 µM tryptamine, 4 µM ticlopidine, 242 µM DDC, 0.3 µM ketoconazole, 22 µM quercetin, and 6 µM sulfaphenazole. After preincubation for 5 min, 50 µL of 15 mM PAP (final concentration of 3 mM) was added. After incubation for 30 min at 37 °C, the incubations were stopped by adding 50 µL of an aliquot to 150 µL of 125 mM zinc sulfate. Denatured proteins were removed by centrifugation (10 min at 10 000 rpm), and the supernatants (100 µL) were analyzed by HPLC. All incubations were performed in triplicate, and control incubations (i.e., with no chemical inhibitor added) were performed. Incubations with Recombinant Human P450 Enzymes. Incubations of 5 mM PAP with 10 different recombinant human P450 enzymes were performed following the In Vitro Technologies protocol (23). The reactions were terminated after incubation for 20 min at 37 °C by adding 50 µL of an aliquot to 150 µL of 125 mM zinc sulfate. Denatured proteins were removed by centrifugation (10 min at 10 000 rpm), and the supernatants (100 µL) were analyzed by HPLC. All incubations were performed in triplicate. Kinetic Studies with Recombinant Human P450 Enzymes. Preincubations (total volume of 400 µL) of the recombinant human enzymes 2C9, 2C18, 2D6, 2C8, and 2E1 (final concentration of 50 nM) in 500 nM Tris buffer containing 2 mM NaCl and 10 mM EDTA were performed at different PAP concentrations (0.25, 0.5, 1, 2, 4, and 8 mM for 2C9, 2C18, and 2C8; 0.5, 1, 2, 4, 8, and 16 mM for 2D6; and 1, 2, 4, 6, and 8 mM for 2E1). After preincubation for 5 min at 37 °C, the reactions were started by addition of 80 µL of NADP(H) (final concentration of 1.2 mM), and 85 µL aliquots
1220 Chem. Res. Toxicol., Vol. 20, No. 8, 2007
Martínez-Cabot et al.
Figure 1. HPLC profiles of microsomes incubated for 20 min at 37 °C. Panels A and D show incubations of RLM and HLM, respectively, in the presence of 1 mM NADP(H) but without substrate. Panels B and E show incubations of RLM and HLM, respectively, in the presence of 5 mM PAP and 1 mM NADP(H), and panels C and F show the standard phenol 1. The UV spectra of peaks b and e are also shown in the insets of panels B and E and compared with the UV spectra of the standard phenol 1 (insets in panels C and F). The ESI-HRMS spectra of peaks b and e are also shown.
were withdrawn at 3, 5, 7, and 10 min. The reactions were stopped by adding 28 µL of 125 mM zinc sulfate at 0 °C. Denatured proteins were removed by centrifugation (10 min at 10 000 rpm), and the supernatants (95 µL) were analyzed by HPLC. All the experiments were performed in triplicate on three different days. Enzyme kinetic parameters, including Km and kcat, for the 4′-hydroxylation of PAP in the presence of human recombinant enzymes 2C9, 2C18, 2D6, 2C8, and 2E1 were determined as described above. Kinetic Studies of in Vitro Glutathione Formation. Microsomal incubations (total volume of 1 mL) of a solution of phenol 1 (final concentration of 125 µM) containing GSH (final concentration of 250 µM) in 50 mM sodium phosphate buffer containing 5 mM MgCl2 (pH 7.4) and 80 µL of pooled RLM (20 mg of protein/mL) were performed at 37 °C in the presence and absence of NADP(H) (final concentration of 1 mM). Incubations of phenol 1 and GSH were also carried out under the same conditions but in the absence of microsomes and using boiled microsomes. The same procedure was followed for the case of HLM, except that 40 µL of pooled HLM (38.4 mg of protein/mL) was used instead of RLM. In both cases, 150 µL aliquots were withdrawn at 0, 2, 5, 13, 20, and 30 min, and the reactions were stopped by adding 50 µL of 125 mM zinc sulfate at 0 °C. Denatured proteins were removed by
centrifugation (10 min at 10 000 rpm), and the supernatants (100 µL) were analyzed by HPLC. Unfortunately, the concentration of conjugate 3 could not be determined due to the instability of the corresponding standard, which could not be isolated and wholly characterized as it is easily transformed into the corresponding 4-aminophenol adduct (15). The rate of formation of 3 was therefore measured by the HPLC peak area and confirmed by ESI-HRMS. 3: ESI-HRMS m/z (RLM) 489.1664 and m/z (HLM) 489.1646; for (M + H)+ C19H29N4O9S requires m/z 489.1655. All incubations were performed in triplicate (one experiment per day on three different days). Synthesis of P(1)PAP. A mixture of PAP (1.23 g, 7.3 mmol), palmitic acid (1.44 g, 5.6 mmol), N,N-dicyclohexylcarbodiimide (1.98 g, 9.6 mmol), and N,N-(dimethylamino)pyridine (0.09 g, 0.73 mmol) in CH2Cl2 was stirred for 3 h at 25 °C leading to a concomitant formation of P(1)PAP and P(2)PAP. The crude mixture was purified by Biotage flash chromatography, and P(1)PAP was obtained with a purity of 95% (HPLC): 1H NMR (CDCl3) δ 7.19 (t, 2 H, J ) 8 Hz, Har), 6.75 (t, 1 H, J ) 7.2 Hz, Har), 6.67 (d, 2 H, J ) 8 Hz, Har), 4.25–3.97 (ac, 3 H, CH2-O, CH-OH), 3.31 (dd, 1 H, J1 ) 12.8 Hz, J2 ) 4 Hz, CH2-N), 3.16, (dd, 1 H, J1 ) 12.8 Hz, J2 ) 7.2 Hz, CH2-N), 2.35 (t, 2 H, J ) 7.2 Hz, CR-H2), 1.63
In Vitro BioactiVation of PAP by P450 (m, 2 H, Cβ-H2), 1.24 (s, 24 H, CH2), 0.87 (t, 3 H, J ) 6.4 Hz, CH3); 13C NMR (CDCl3) δ 174.4 (CO), 147.9 (Car-1), 129.6 (Car3, Car-5), 118.5 (Car-4), 113.6 (Car-2, Car-6), 68.7 (CH2-O), 66.65 (CH-OH), 46.9 (CH2-N), 34.4 (CR-H2), 31.8 (CH2-CH2CH3), 29.9 (10 CH2), 25.6 (Cβ-H2), 22.9 (CH2-CH3), 14.3 (CH3); ESI-HRMS m/z 406.3258; for (M + H)+ C25H44NO3 requires m/z 406.3243. Incubations of P(1)PAP with HLM. Incubations (total volume of 200 µL) with pooled HLM (8 µL, 38.4 mg of protein/mL) were performed in 50 mM sodium phosphate buffer containing 5 mM MgCl2 and 4 mM taurodeoxycholic acid (pH 7.4) in the presence of 80 µL of 25 mM P(1)PAP (final concentration of 5 mM). A blank sample without P(1)PAP was also carried out. The reactions were started by addition of 40 µL of 5 mM NADP(H) (final concentration of 1 mM). The samples were then incubated at 37 °C. Aliquots of 75 µL were withdrawn at different times, and the incubations were stopped by cooling the mixture to 0 °C and adding 20 µL of 125 mM zinc sufate and 150 µL of CH2Cl2. After centrifugation (10 min at 10 000 rpm), the organic phase was collected and solvent evaporated. The sample was redissolved with a mixture of n-hexane and isopropyl alcohol (99:1) and analyzed by HPLC at 245 nm. The gradient elution conditions consisted of an initial step at 99% A (n-hexane) and 1% B (isopropyl alcohol) for 2 min, then a linear ramp from 99 to 70% A from 2 to 17 min, which was maintained for 3 min under these conditions, and then a return to the initial conditions. In this experiment, a new peak at 13.1 min was detected and identified as PAP by ESI-HRMS and by comparison with the corresponding standard (see Figure 6 of the Supporting Information): ESI-HRMS m/z 168.1017; for (M + H)+ C9H14NO2 requires m/z 168.1025.
Chem. Res. Toxicol., Vol. 20, No. 8, 2007 1221 Table 1. Inhibition of Human Liver Microsome-Induced PAP 4′-Hydroxylation by Enzyme-Specific Inhibitors inhibitor
P450 enzyme
percentage of inhibition ( SD
R-naphthoflavone quinidine tryptamine ticlopidine DCC ketoconazole quercetin sulfaphenzole
1A1/2 2D6 2A6 2B6/2C19 2E1 3A4/5 2C8 2C9
52.6 ( 2.7 4.9 ( 2.7 38.6 ( 3.9 29.1 ( 9.1 65.0 ( 3.2 56.9 ( 6.6 71.8 ( 4.7 42.5 ( 2.7
Table 2. Kinetic Parameters for the Main P450 Enzymes Involved in PAP 4′-Hydroxylation P450 enzyme
kcat ( SD (min-1)
Km ( SD (mM)
kcat/Km ( SD (min-1 mM-1)
% HLM
2C9 2C18 2D6 2C8 E1
138 ( 20 118 ( 18 144 ( 8 20 ( 1 19 ( 3
5.7 ( 1.0 6.7 ( 1.8 4.9 ( 0.7 1.0 ( 0.2 4.7 ( 1.4
24 ( 20 18 ( 10 29 ( 11 19 ( 1 4(2
20a NRb 2c
