Carboxylic Acid Drug-Induced DNA Nicking in HEK293 Cells

Sep 20, 2007 - Department of Cardiology and Clinical Pharmacology, The Queen Elizabeth ... in the bioactivation of carboxylic acid drugs to genotoxic ...
1 downloads 0 Views 740KB Size
Download PDF
1520

Chem. Res. Toxicol. 2007, 20, 1520–1527

Carboxylic Acid Drug-Induced DNA Nicking in HEK293 Cells Expressing Human UDP-Glucuronosyltransferases: Role of Acyl Glucuronide Metabolites and Glycation Pathways Hamish T. Southwood,†,‡ Yvette C. DeGraaf,†,‡ Peter I. Mackenzie,§ John O. Miners,§ Philip C. Burcham,| and Benedetta C. Sallustio*,†,‡ Department of Cardiology and Clinical Pharmacology, The Queen Elizabeth Hospital, Adelaide, SA 5011, Australia, Discipline of Pharmacology, School of Medical Sciences, UniVersity of Adelaide, Adelaide, SA 5000, Australia, Department of Clinical Pharmacology, Flinders UniVersity, Adelaide, SA 5042 Australia, and Pharmacology Unit, School of Medicine and Pharmacology, UniVersity of Western Australia, Perth, WA 6009, Australia ReceiVed May 29, 2007

Glucuronidation of carboxylic-acid-containing drugs can yield reactive acyl (ester-linked) glucuronide metabolites that are able to modify endogenous macromolecules. Previous research has shown that several carboxylic acid drugs are genotoxic in isolated mouse hepatocytes, and that DNA damage is prevented by the glucuronidation inhibitor, borneol. Whether these species induce comparable genetic damage in human cells is unknown. In this study, we investigated the mechanisms of clofibric acid-induced genotoxicity in HEK293 cells expressing the human UDP-glucuronosyltransferases UGT1A3, UGT1A9, or UGT2B7, and screened three other carboxylic acid drugs for UGT-dependent genotoxicity. DNA damage was detected using the alkaline version of the comet assay. HEK293 cells were incubated for 18 h with vehicle (2.5 mM NaOH), 0.1–2.5 mM clofibric acid or 0.1–1.0 mM benoxaprofen, bezafibrate, or probenecid. To identify mechanisms underlying any observed genotoxicity, we treated UGT2B7 transfectants with 10 mM aminoguanidine, 1 mM borneol, or 2 mM desferrioxamine mesylate prior to co-incubation with 1 mM clofibric acid for 18 h. Compared to vehicle, clofibric acid, benoxaprofen, and probenecid produced significant DNA damage in all three UGT-transfected HEK293 cell lines, detectable from the lowest concentration tested. Bezafibrate caused DNA damage only at higher concentrations (1.0 mM) in UGT2B7- and UGT1A9-, but not UGT1A3-transfected cells. No drug-induced DNA damage was detected in untransfected cells, consistent with the limited glucuronidation capacity of these cells. The glycation/glycoxidation inhibitor aminoguanidine and the glucuronidation inhibitor borneol significantly decreased clofibric-acid-mediated DNA damage in UGT2B7 transfected cells by 73.5 and 94.8%, respectively. The inhibitor of transition -metal-catalyzed oxidation, desferrioxamine mesylate, had no significant effect on DNA damage. This study demonstrates the substrate-dependent role of human UGTs in the bioactivation of carboxylic acid drugs to genotoxic acyl glucuronide metabolites that are able to damage nuclear DNA via glycation and/or glycoxidation mechanisms. Introduction Glucuronidation of clinically useful carboxyl-containing xenobiotic substrates, including fibrate hypolipidemic agents (clofibric acid and gemfibrozil), anticonvulsants (valproic acid), and non-steroidal anti-inflammatory drugs (diclofenac), can generate reactive electrophilic acyl (ester-linked) glucuronide metabolites (1, 2). These electrophilic intermediates have the ability to generate protein adducts and have been implicated in a wide range of adverse drug effects such as drug hypersensitivity and cellular toxicity (1, 2). The reactivity of acyl glucuronides is due to the intrinsic electrophilicity of the ester group and the inherent reactivity of the glucuronic acid monosaccharide moiety. Electrophilicity of the ester group gives rise to an ability to undergo nonenzymatic reactions, including hydrolysis, transacylation reactions with nucleophilic groups on proteins and peptides, and intramolecular rearrangement (3). Intramolecular rearrangement occurs when the ester participates in transacylation reactions with the * Corresponding author. Dr Benedetta C Sallustio, Clinical Pharmacology Laboratory, The Queen Elizabeth Hospital, 28 Woodville Road, Woodville SA 5011, Australia. Phone: 61 8 8222 6510. Fax: 61 8 8222 6033. E-mail: [email protected]. † The Queen Elizabeth Hospital. ‡ University of Adelaide. § Flinders University. | University of Western Australia.

hydroxyl groups on the glucuronic acid moiety, resulting in migration of the aglycone to various carbon positions on the sugar (4). This rearrangement exposes the hemiacetal function of the glucuronic acid moiety, allowing a series of reactions typical of monosaccharides, including protein glycation (5), as well as the formation of advanced glycation end-products (AGEs1 ) (6). Monosaccharide-derived protein and DNA AGEs have genotoxic potential (7, 8). Additionally, assuming that acyl glucuronides possess the same capacity for autoxidation reactions as glucose, they may also induce oxidative DNA damage, which is catalyzed by transition metals (9). We have recently reported that a number of carboxylic acid drugs cause DNA damage in isolated mouse hepatocytes, with clofibric acid (CA) showing the highest DNA damaging potency (10). In humans, between 60 and 95% of a CA dose is conjugated to form CA glucuronide (11), which makes the drug highly suitable for exploring the biological properties of acyl glucuronides. We have also previously (12) reported that CA glucuronide produces significant concentration-dependent nicking of double-stranded plasmids. In addition, direct exposure of bacteriophage DNA to CA glucuronide or gemfibrozil glucuronide, but not the parent aglycones, resulted in significant concentration-dependent decreases in transfection efficiencies in E.coli (12). Moreover, CA-induced DNA damage in mouse hepatocytes occurred at CA concentrations within the range of those attained clinically and was dependent on intracellular

10.1021/tx700188x CCC: $37.00  2007 American Chemical Society Published on Web 09/20/2007

UGT-Dependent DNA Damage by Carboxylic Acid Drugs

formation of CA glucuronide (13). Interestingly, in the mouse hepatocyte studies, 18 h exposure to 0.1 mM CA produced an amount of DNA damage comparable to that of 2 h exposure to 10 mM concentrations of the known genotoxicant styrene (13). Collectively, these studies reveal the genotoxic potential of acyl glucuronides formed from fibrate hypolipidemic drugs and highlight the importance of designing genotoxicity studies that detect a contribution by glucuronidation products of carboxylcontaining drugs. Although evidence of acyl-glucuronide-mediated DNA damage has been documented using both a viral (12) and an animal (10, 13) model, the bioactivation of carboxylic acid compounds to DNA-damaging intermediates in human cells has not been addressed. The human embryonic kidney (HEK293) cell line is a suitable human model for investigating the involvement of UDP-glucuronosyltransferases (UGTs), as untransfected cells have negligible UGT activity and cells transfected with UGTs display good expression of UGT proteins with localization to the endoplasmic reticulum, similar to constitutively expressing cells such as hepatocytes (14). The UGT enzymes shown to catalyze the glucuronidation of carboxylic acid moieties include UGT 2B7 (15), 1A3 (16), 1A9 (17), and 1A6 (18). CA is a substrate for UGTs 2B7, 1A3, and 1A9 but not 1A6 (18). The aims of the present study were to confirm in situ xenobiotic glucuronidation in HEK293 cells transfected with UGT1A3, UGT1A9, or UGT2B7 and to demonstrate the UGTdependent bioactivation of CA. It was hypothesized that CA would form acyl glucuronide metabolites and cause significant DNA damage only in transfectants, not native/untransfected cells. The mechanisms of CA-induced DNA damage were further investigated using the glucuronidation inhibitor, borneol, as well as reagents known to interfere with possible genotoxic pathways of acyl glucuronides, namely, aminoguanidine (AG), which inhibits glycation/glycoxidation chemistry and desferrioxamine mesylate (DFOM), which modulates responses to oxidative stress by chelating transition metals. In addition, control and transfected HEK293 cells were screened for UGTdependent bioactivation of three other carboxylic acid drugs, benoxaprofen, bezafibrate, and probenecid.

Experimental Procedures Reagents. 4-Methylumbelliferone (4-MU), 4-MU β-D-glucuronide, CA, bezafibrate, probenecid, and AG were purchased from Sigma Chemical Co. (St. Louis, MO). DFOM was obtained from Ciba-Geigy (Pendle Hill, NSW, Australia), benoxaprofen from Lilly (Indianapolis, IN), and (–)-borneol from Aldrich Chemical Co. (Milwaukee, WI). SYBR Green I was obtained from Molecular Probes, Inc. (Eugene, OR). Low melt agarose (LMA) was purchased from Bio-Rad Laboratories (Hercules, CA). Fetal bovine serum (FBS) was obtained from JRH Biosciences (Brooklyn, VIC, Australia) and puromycin was obtained from A.G. Scientific Inc. (San Diego, CA). High-glucose Dulbecco’s modified eagle’s media (DMEM) and Gluta MAX-1 were obtained from Invitrogen (Carlsbad, CA). Acetonitrile, obtained from Crown Scientific (Beverly, SA, Australia), was of HPLC grade. All other reagents used were of analytical grade. 1 Abbreviations: AGE, advanced glycation end-product; CA, clofibric acid; HEK293, human embryonic kidney 293 T; UGT, UDP-glucuronosyltransferase; AG, aminoguanidine; DFOM, desferrioxamine mesylate; 4-MU, 4-methylumbelliferone; LMA, low melt agarose; FBS, fetal bovine serum; DMEM, Dulbecco’s modified Eagle’s media; LDH, lactate dehydrogenase; MTT, 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyl tetrazolium bromide; LC10, concentration causing 10% cell death; LC30, concentration causing 30% cell death; IQR, interquartile range.

Chem. Res. Toxicol., Vol. 20, No. 10, 2007 1521

UGT Transfected Cells. UGT1A3, UGT1A9, and UGT2B7 were stably expressed in HEK293 cells (ATCC). Cells were transfected separately with UGT1A3, UGT1A9, or UGT2B7 cDNA cloned into the expression vector pEF-IRES-puro-6 and selected with puromycin (19). Cell lines were stored in liquid nitrogen and thawed for use by the addition of prewarmed DMEM, supplemented with 10% (v/v) FBS, 1% (v/v) Gluta MAX-1, 100 mg/L sodium pyruvate, and 0.5% (v/v) penicillin/ streptomycin (penicillin 10000 IU/mL, streptomycin 10 mg/mL). The cell lines were incubated with 95% O2/5% CO2 at 37°C and split 1:10 every 2–3 days. Cytotoxicity Assays. The cytotoxicity of all test compounds added to cell cultures was assessed using the 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyl tetrazolium bromide (MTT) viability assay (13, 20) using 96-well microtiter plates layered with 25000 cells/well in supplemented DMEM culture medium as previously described. After an overnight attachment period, cultured cells were incubated with fresh 0.5% (v/v) FBS supplemented DMEM media containing increasing concentrations of CA (1 µM to 5 mM), benoxaprofen (1 µM to 1 mM), bezafibrate (1 µM to 10 mM), probenecid (1 µM to 1 mM), borneol (1 µM to 4 mM), AG (10 µM to 120 mM), or DFOM (0.1 µM to 50 mM) for 18 h. Media was then removed, replaced with FBS-free media containing 0.25 mg/mL MTT and incubated for a further 2 h. The medium was removed, cells were lysed by the addition of DMSO, and the optical density of the cell contents was measured at an absorbance of 570 nm using a microplate reader (Polarstar Galaxy, BMG Labtechnologies GmbH, Germany). Untransfected HEK cells were used as controls. Concentrations causing 10 and 30% cell death (LC10 and LC30) were approximated from concentration-response curves obtained using nonlinear regression analysis in GraphPad Prism (version 4.03, GraphPad Software Inc., San Diego, CA). Cell morphology was also closely monitored. The concentrations of all drugs used in subsequent genotoxicity assays were below the calculated LC30 values. HPLC Analysis of 4-MU glucuronide formation. The in situ glucuronidation capacity of control (untransfected) cells and cells transfected with the human UGT enzymes was assessed semiquantitatively by investigating the glucuronidation of a nonselective UGT substrate (19), 4-MU (0, 0.1, and 1.0 mM), during 18 h incubations at 37 °C. Concentrations of 4-MU glucuronide in HEK293 culture media were determined using reverse-phase HPLC as previously described (21). Briefly, 250 µL aliquots of culture media were centrifuged at 13000 g for 5 min, and 200 µL aliquots of the supernatant fractions were stored at -20 °C until analysis. Samples were thawed at room temperature and acidified by adding 2 µL of 70% (w/w) perchloric acid. After being cooled on ice and centrifuged at 13000 g for 3 min, 100 µL aliquots were analyzed using an Agilent 1100 series HPLC system (Agilent Technologies, Sydney, Australia), which consisted of a quaternary gradient pump, autosampler, and UV detector. The HPLC system was fitted with a NovaPak C18 column (5 µm, 150 × 3.9 mm, Waters, Milford, MA). The mobile phase was pumped at a flow rate of 1 mL/min and was a series of gradient steps consisting of solvent A (10% (v/v) acetonitrile in an aqueous solution of 10 mM triethylamine adjusted to pH 2.5 with 70% (w/w) perchloric acid) and solvent B (acetonitrile) in the (v/v) ratios of 96:4 from 0–3.0 min, 70:30 from 3.1–4.1 min, and 96:4 from 4.2–6 min. Under these conditions, 4-MU glucuronide, detected by UV absorbance at a wavelength of 316 nm, had a retention time of 3.74 min.

1522 Chem. Res. Toxicol., Vol. 20, No. 10, 2007

HPLC Analysis of CA Glucuronide. Because of the facile rearrangement and hydrolysis of CA glucuronide at physiological pH and temperature (22), the glucuronidation of CA was assessed qualitatively in control and UGT2B7 transfected cells cultured for 18 h at 37 °C with CA (1 mM) in the presence or absence of borneol (1 mM). Cell lysate was prepared by the addition of 100 µL of 5% (v/v) TritonX-100 to each plate followed by sonication. A 600 µL aliquot was acidified with 6 µL of 70% (w/w) perchloric acid, mixed and centrifuged, and the supernatant was stored frozen at -20 °C until analysis. Samples were thawed and centrifuged, and 30 µL was then injected directly into the HPLC. Resolution of CA glucuronide was carried out using an Alltima CN column (5 µm, 4.6 × 250 mm, Alltech) with a mobile phase (pH 3.5) consisting of 17% (v/v) acetonitrile in 5 mM tetrabutyl ammonium hydrogen sulfate, pumped at a flow rate of 1.0 mL/min, and ultraviolet detection at a wavelength of 235 nm. Comet Assay. The alkaline version of the comet assay (single-cell gel electrophoresis) is a sensitive genotoxicity test for the detection of DNA strand breaks, the most prominent effect observed in plasmids exposed to CA glucuronide (12). The comet assay is based on the principle that DNA fragments formed via DNA damage can be detected following agarose gel electrophoresis and fluorescent staining (23). Moreover, the use of different pH conditions during the cell lysis step allows the detection of different types of DNA damage, including single- and double-strand breaks, and alkali-labile sites (24). The comet assay was performed as previously described (13). Briefly, each HEK293 cell line was seeded at 2.5 × 105 cells/ mL in 35 mm dishes 24 h before drug treatment. Cells were treated at 37 °C for 18 h with 0.5% (v/v) FBS supplemented DMEM medium containing either vehicle (2.5 mM NaOH) or 0.1–2.5 mM CA alone, or pretreated with 10 mM AG, 1 mM borneol, or 2 mM DFOM for 30 min, followed by co-incubation with 1mM CA for 18 h. In a separate series of incubations, cells were treated with vehicle, benoxaprofen (0.1 mM), bezafibrate (0.1 and 1.0 mM), probenecid (0.1 and 1.0 mM), or CA (0.1 mM). Cell viability was assessed at the end of each incubation by measuring lactate dehydrogenase (LDH) activity in culture medium as previously described (25). Microscope slides were prepared by immersion in 1% (w/v) normal melting agarose and dried for 2 h on a 50 °C hot plate. One volume of cell suspension (100 µL, containing approximately 5 × 105 cells) was mixed with 9 volumes of 1% (w/v) LMA maintained at 42°C in a water bath, after which 100 µL of the diluted suspension was layered on a precoated slide. The slide was immediately covered with a coverslip and incubated at 4°C to solidify the agarose. After coating the slides with a third layer of 0.5% (w/v) LMA at 4 °C for 20 min, embedded cells were immersed for 3 h at 4 °C in cold lysis buffer (2.5 M NaCl, 1% (w/v) sodium N-lauroylsarcosinate, 100 mM disodium EDTA, 10 mM Tris base, pH 10) supplemented with 1% (v/v) Triton X-100. Slides were placed in a horizontal electrophoresis assembly containing fresh electrophoresis buffer (300 mM NaOH, 10 mM disodium EDTA, pH > 13) to a depth of 10 mm. To allow DNA unfolding and unwinding, we left the slides in the buffer for 30 min prior to electrophoresis. Following electrophoretic resolution (27 V for 13 min) using a recirculating horizontal tank (FB-SBR-2025, Fisher Scientific, PA), slides were immersed in neutralizing buffer (0.4 M TrisHCl, pH 7.5) for 25 min. They were then stained with SYBR Green I (diluted 1:1 × 104 in TE8 buffer (10 mM Tris-HCl, 1 mM EDTA, pH 8) for 20 min. Slides were photographed at 100× magnification under a fluorescent microscope (Olympus

Southwood et al.

Figure 1. Representative chromatograms of culture media from 18 h incubations of 4-MU (1 mM) with (A) HEK293 cells transfected with UGT1A3 and (B) untransfected HEK293 cells. Retention times for authentic 4-MU and 4-MU glucuronide were 5.8 and 3.7 min, respectively. The arrows indicate the retention time of authentic 4-MU glucuronide.

BX50, Tokyo, Japan) equipped with a 450–490 nm reflector and a 530–550 nm barrier filter set. Images of the nuclei were acquired in random, non-overlapping fields using a Peltier cooled CCD SPOT RT video camera (Diagnostic Instruments, MI). Resulting photographs of fluorescently labeled comets were scored on the basis of tail extent moment using Komet 5.5 (Kinetic Imaging, U.K.) and 30 comets were scored per slide, with 2 slides per treatment over 2–3 separate experimental days (i.e., 120–180 comets scored per treatment). Untransfected HEK293 cells were used as controls. Data are presented as medians ( interquartile range (IQR). Statistical analysis was carried out for each drug using Kruskal Wallis with Dunns post hoc tests (GraphPad Prism, GraphPad Software Inc., San Diego, CA).

Results Cytotoxicity Assays. Because the quality of data obtained using the Comet assay is undermined by endonuclease activation secondary to compound toxicity, all compounds used in this study were first evaluated for toxicity towards untransfected and UGT-expressing HEK293 cells. The concentrations of CA (0.1–2.5 mM) used in all cell cultures were below the LC30 values in control, UGT1A3, UGT1A9, and UGT2B7 transfected cells, which ranged from 3.7 to >5.0 mM. LC30 values were approximately 0.23 mM for benoxaprofen and >1.0 mM for bezafibrate and probenecid. The concentrations of borneol (1 mM), AG (10 mM), and DFOM (2 mM) were below their LC10 values in control (1.8, 18.9, and 9.0 mM, respectively) and UGT2B7 transfected cells (1.9, 20.6, and 4.5 mM, respectively). As a supplementary measure of cellular viability, LDH leakage into incubation medium was also assessed following 18 h incubations (i.e. prior to agarose embedment in the comet assay protocol). In each of the transfectants tested, none of the compounds used in these experiments (i.e. CA, benoxaprofen, bezafibrate, probenecid, borneol, AG, DFOM) induced LDH leakage that exceeded 10% of total cell LDH levels. Glucuronidation Capacity of Control and UGT-Transfected HEK293 Cells. The in situ formation of 4-MU glucuronide and CA glucuronide in UGT-transfected HEK293 cells

UGT-Dependent DNA Damage by Carboxylic Acid Drugs

Chem. Res. Toxicol., Vol. 20, No. 10, 2007 1523

Figure 2. Representative chromatograms of cell lysates from 18 h incubations of (A) UGT2B7-transfected cells with vehicle (2.5 mM NaOH); (B) UGT2B7-transfected cells with CA (1 mM); (C) UGT2B7-transfected cells with CA (1 mM) and borneol (1 mM); and (D) untransfected cells with CA (1 mM). The arrows indicate the retention time of the CA glucuronide. Under these conditions, the biosynthetic and intramolecular rearrangement isomers of CA glucuronide are not completely resolved.

Table 1. Concentration of 4-MU Glucuronide in Culture Medium following Single Incubations of 4-MU with Control HEK293 Cells and Those Transfected with UGT1A3, UGT1A9, and UGT2B7a 4-MU glucuronide (µM) 4-MU (µM)

untransfected

UGT1A3

UGT1A9

UGT2B7

0 100 1000

n.d. n.d. n.d.

n.d. 3.5 7.8

n.d. 31.8 18.9

n.d. 2.7 8.7

a

n.d. ) Below lower limit of detection (0.05 µM).

is shown in Figures 1 and 2, respectively. Untransfected control HEK293 cells displayed no glucuronidation capacity towards either the nonselective UGT substrate 4-MU (Figure 1B, Table 1) or CA (Figure 2D). 4-MU glucuronide, identified on the basis of coelution with authentic 4-MU glucuronide standard, was formed in UGT1A3, UGT1A9, and UGT2B7 transfected cells (Figure 1, Table 1). The CA glucuronide peak was present in chromatograms obtained from UGT2B7 transfected cells incubated with CA (Figure 2B), but absent in samples from control HEK293 cells incubated with CA (Figure 2D) and UGT2B7 transfectants co-incubated with CA and the glucuronidation inhibitor, borneol (Figure 2C). Comet Assay. Representative nuclei from vehicle or CAtreated UGT2B7-transfected cells are depicted in Figure 3. Cells exposed to 1 mM CA (Figure 3B) clearly show a high level of DNA strand breaks not evident in the controls (Figure 3A). The effects of 0.1–2.5 mM CA on DNA damage in control and UGT-transfectants are shown in Figure 4. There was no significant difference in median tail extent moment between the three UGT-transfectants exposed to vehicle only. However, exposure of all three UGT-transfectants to CA concentrations of 0.1 mM or higher caused significant and concentrationdependent increases in tail extent moment. In UGT1A3- and UGT1A9-transfectants exposed to 0.1 and 1.0 mM CA, median (IQR) tail extent moments were 15.8 (6.7–21.6) and 23.4 (17.6–30.8) (Figure 4B), and 4.8 (1.6–8.5) and 16.4 (10.9–22.1) (Figure 4C). In UGT2B7-transfectants exposed to 0.1, 0.5, 1.0, and 2.5 mM CA, median (IQR) tail extent moments were 2.7 (1.4–4.4), 10.2 (6.5–15.3), 13.8 (6.7–20.5), and 29.4 (19.9–37.4), respectively (Figure 4D). The median tail extent moments of UGT1A3-transfectants treated with 0.1 or 1.0 mM CA were significantly higher than those of the other transfectants. In the untransfected cells, DNA damage became evident only after treatment with 2.5 mM CA (Figure 4A).

The effect of the glucuronidation inhibitor borneol, the glycation inhibitor AG, and the metal chelator DFOM on CAmediated DNA damage was explored in UGT2B7-transfected cells (Figures 3 and 5). In untransfected HEK293 cells, no significant change in baseline tail extent moment occurred after any treatment (Figure 5A). In UGT2B7-transfected cells, borneol and AG significantly reduced CA-induced DNA damage by 94.8 and 73.5%, respectively (Figures 3 and 5B). DFOM reduced CA-mediated DNA damage by 22.2% (Figure 5B); however, this was not statistically significant. A comparison of the DNA damage caused by benoxaprofen, bezafibrate, probenecid, and CA in control and UGT-transfected HEK293 cells is shown in Table 2. Because of its cytotoxicty, benoxaprofen was tested only at the 0.1 mM concentration. No DNA damage was detected in control HEK293 cells exposed to carboxylic acid drugs, or in UGT-transfected cells exposed to vehicle only. Benoxaprofen and probenecid produced significant DNA damage in all three UGT-transfected cell lines. In contrast, bezafibrate caused significant DNA damage in UGT2B7- and UGT1A9-, but not UGT1A3-transfected cells (Table 2).

Discussion We have previously shown that clinically relevant concentrations of CA induce DNA damage in isolated mouse hepatocytes (13). The importance of UGT-dependent bioactivation of CA was demonstrated using the glucuronidation inhibitor borneol, which essentially abolished both CA glucuronide formation and DNA damage (13). In a later study, clinically relevant concentrations of probenecid were also shown to cause DNA damage in isolated mouse hepatocytes, which was essentially abolished by borneol (10). The aim of the present study was to explore whether this route to DNA damage was relevant in a human cell line, confirm the role of human UGTs in the formation of genotoxic glucuronide metabolites of CA, and determine whether other carboxylic acid drugs undergo bioactivation by human UGTs to DNA-damaging intermediates. Although UGT-transfected HEK293 cells are used mainly for the preparation of microsomal fractions or lysates for in vitro drug metabolism reactions (16, 19), significant glucuronidation has been reported by intact cultured UGT-transfected HEK293 cells (26, 27). The present study similarly demonstrates that intact UGT-transfected HEK293 cells, but not untransfected controls, are capable of intracellular metabolism of both the

1524 Chem. Res. Toxicol., Vol. 20, No. 10, 2007

Southwood et al.

Figure 3. Representative fluorescence microscopy photographs of nuclei from UGT2B7-transfected cells incubated for 18 h with (A) vehicle (2.5 mM NaOH), (B) 1 mM CA, (C) 1 mM CA plus 1 mM borneol, (D) 1 mM CA plus 10 mM AG, and (E) 1 mM CA plus 2 mM DFOM.

Figure 4. DNA damage produced by CA in HEK293 cells that were (A) untransfected, or transfected with (B) UGT1A3, (C) UGT1A9, and (D) UGT2B7. Results are shown as median ((IQR) tail extent moment calculated from 30 comets scored per slide, from 2 slides per treatment, over 2–3 separate experiments (total 120–180 comets). The 0.5 and 2.5 mM concentrations were not tested with the UGT1A3- and UGT1A9-transfected cells. ***P < 0.001 compared to control (NaOH vehicle) incubations.

nonselective UGT substrate 4-MU and the fibrate hypolipidemic drug CA. We have previously reported an in vitro intrinsic clearance of 4-MU by UGT1A9-transfected HEK293 cell lysates of 1.1 mL min–1 mg–1 protein (19), compared to 0.4 mL min–1 mg–1 protein in human liver microsomes (28), suggesting good UGT protein expression in the transfected cells used here. In the current study, the greater rate of 4-MU glucuronidation by UGT1A9-expressing cells compared to 1A3 or 2B7 (Table 1) is consistent with the in vitro glucuronidation of 4-MU by cell lysates of UGT-expressing HEK293 cells (19). In addition, lower formation of 4-MU glucuronide by UGT1A9-expressing cells incubated with 1000 µM 4-MU compared to cells incubated with 100 µM 4-MU (Table 1) is consistent with substrate

inhibition previously reported for the in vitro glucuronidation of 4-MU by cell lysates of UGT1A9-expressing HEK293 cells (with a substrate inhibitor constant of 226 µM) (19). The extensive DNA damage detected following exposure of HEK293 cells expressing human UGT enzymes to CA was clearly dependent on glucuronidation of CA, as little or no damage was detected in untransfected cells or in transfected cells in the presence of borneol. Notably, significant DNA strand breaks were induced by 0.1 mM CA, a concentration that is lower than peak total plasma concentrations in patients receiving 500–2000 mg clofibrate (11) and, assuming an unbound fraction of 3% (11), slightly above (approximately 3×) the range of unbound concentrations that may be achieved clinically. In

UGT-Dependent DNA Damage by Carboxylic Acid Drugs

Chem. Res. Toxicol., Vol. 20, No. 10, 2007 1525

Figure 5. Effects of AG (10 mM), borneol (1 mM), and DFOM (2 mM) on DNA damage induced by CA (1 mM) in (A) untransfected and (B) UGT2B7-transfected HEK293 cells. Data are shown as median ((IQR) tail extent moment scored from a total of 60 comets. ***P < 0.001 compared to 1 mM CA.

Table 2. Median (interquartile range) Comet Moments following 18 h Incubations of Carboxylic Acid Drugs with Untransfected or UGT-Transfected HEK293 Cellsa drug

untransfected

UGT2B7

UGT1A3

NaOH benoxaprofen 0.1 mM bezafibrate 0.1 mM 1.0 mM probenecid 0.1 mM 1.0 mM CA 1.0 mM

1.0 (0.3–2.3)

2.0 (0.7–4.7)

1.4 (0.4–3.0)

1.6 (0.5–3.5)

0.8 (0.2–2.1)

13.6 (6.8–22.9)b

6.2 (3.5–9.1)b

4.3 (1.4–7.6)b

n.t. 1.2 (0.3–2.9)

6.3 (1.5–31.2) 12.4 (6.0–30.2)b

n.t. 0.7 (0.2–2.1)

n.t. 4.5 (2.5–10.5)b

n.t. 1.3 (0.4–3.3)

12.7 (2.4–21.6)b 9.3 (3.4–18.9)b

n.t. 15.9 (9.8–22.7)b

n.t. 7.7 (5.1–14.3)b

1.0 (0.4–2.6)

11.2 (4.0–18.9)b

n.t.

n.t. ) not tested (N.B. 1.0 mM CA was tested in all cell groups in separate experiments, as shown in Figure 4). corresponding NaOH control. a

previous studies with isolated mouse hepatocytes, CA induced significant UGT-dependent DNA damage at concentrations as low as 0.01 mM (10), raising the possibility that DNA damage may occur in patients treated with CA. In these previous studies, incubation of mouse hepatocytes with 0.01–1.00 mM CA resulted in the generation of low micromolar concentrations of CA glucuronide in the medium (13), similar to the steady-state plasma concentration of CA glucuronide attained clinically (29). In contrast, in vivo intrahepatic and bile concentrations of acyl glucuronides are likely to be orders of magnitude greater, because of the concentrating action of hepatic membrane transporters (1). Interestingly, clinical use of clofibrate was associated with an increased incidence of gastrointestinal disorders, including malignancies in liver, intestine, and gall bladder (30, 31), organs that are potentially exposed to high concentrations of glucuronide metabolites (1). The DNA damage detected in UGT1A3-, UGT1A9-, and UGT2B7-transfected cells treated with CA is consistent with previous studies showing that CA is a substrate for all three UGTs (18). A small but significant increase in DNA strand breaks was also seen in untransfected cells treated with 2.5 mM CA. Although such damage was less extensive than that caused by the lowest CA concentration in any of the UGT-transfectants, it suggests that some background UGT expression may occur in untransfected HEK293 cells or, alternatively, that at high concentrations, the parent compound CA is able to directly induce DNA damage. However, UGT can not be detected by immunoblotting in untransfected cells (19), and in the current study, UGT activity was undetectable in untransfected cells using the nonselective substrate 4-MU. CA produced significantly greater DNA damage in UGT1A3-transfected cells

UGT1A9

n.t. b

p < 0.05 compared to

compared to the other enzymes, suggesting that CA may be a higher-affinity substrate for UGT1A3 and/or a higher degree of UGT1A3 expression in the HEK293 cells compared to the other enzymes. We have previously demonstrated similar expression of UGT1A3 and UGT1A9 in the HEK293 cells (19); however, it is not currently possible to compare UGT1A and UGT2B expression because of the unavailability of a suitable antibody. Nonetheless, although our study was not designed to quantitate the formation of CA glucuronide, the different levels of DNA damage in the three UGT transfectants, despite all three cell types being incubated with the same concentrations of the CA, strongly implicates the acyl glucuronide metabolite as the DNA-damaging species, rather than the parent CA. Similar to CA, UGT-dependent DNA damage was also demonstrated for the non-steroidal anti-inflammatory drug, benoxaprofen; the fibrate hypolipidemic agent, bezafibrate; and the uricosuric agent, probenecid. Interestingly, bezafibrate caused significant DNA damage in cells expressing UGT2B7 and UGT1A9, but not in those expressing UGT1A3, again consistent with a role for UGT-dependent bioactivation, determined in part by substrate specificity. However, although bezafibrate is metabolized to an acyl glucuronide in humans (32), the UGT enzymes involved have not been identified. Consistent with our previous observations in isolated mouse hepatocytes (10), benoxaprofen had the greatest cytotoxic potency of the four drugs tested, and at a concentration of 0.1 mM, caused significant DNA damage in all three UGT-transfected cell lines. Benoxaprofen is a substrate of UGT2B7 (15); however, its metabolism by UGT1A3 or UGT1A9 has not been investigated. Probenecid also produced significant DNA damage in isolated mouse hepatocytes (10). The extent of damage appeared to be

1526 Chem. Res. Toxicol., Vol. 20, No. 10, 2007

concentration-dependent, but plateaued at probenecid concentrations >0.05 mM, similar to the DNA damage observed in the current study with UGT2B7-trasfected cells. Probenecid is metabolized to an acyl glucuronide in humans (33, 34), but again, the UGT enzymes involved have not been identified. Although UGT-transfected HEK293 cells clearly demonstrate the direct role of UGTs in the bioactivation of carboxylic acid drugs, they lack many of the uptake and efflux transporters that are normally expressed on the hepatocyte cell membrane. For example, HEK293 cells do not express OATP2 or OATP8 (35), so that the uptake of parent carboxylic acid drug may be lower than that in hepatocytes. Similarly, HEK293 cells do not express MRP2 (36) or MRP3 (37), so that the efflux of intracellularlygenerated acyl glucuronides may be lower than that in heptaocytes. Thus, a direct comparison between the genotoxic potency of CA, bezafibrate, benoxaprofen, and probenecid in HEK293 cells and hepatocytes is not possible. Earlier work (38) has highlighted the importance of ensuring that compounds are tested at noncytotoxic concentrations during Comet assay-based evaluations of genotoxic potential. The current consensus is that optimal Comet assay data are obtained from cells exposed to concentrations of suspected genotoxicants that produce no greater than 30% cell death (24, 38). In this study, all carboxylic acid drug concentrations were below the mean LC30 in each transfectant, whereas borneol, AG, and DFOM concentrations were below their respective mean LC10. Additionally, incubations in all transfectants produced LDH leakages of no more than 10%. It thus seems unlikely that DNA damage detected in drug-treated HEK293 transfectants was due to necrosis or apoptosis-related DNA fragmentation. The mechanisms underlying acyl-glucuronide-induced DNA damage are currently not well-understood. However, on the basis of their known chemical properties and reactivity with protein nucleophiles, several genotoxic mechanisms may be possible. For example, as electrophilic species, the ester group possessed by acyl glucuronides may undergo transacylation reactions with nucleophilic centers within DNA, a reaction pathway that is well-documented with protein nucleophiles (5, 39). Alternatively, in reactions analogous to those described for endogenous sugars (7, 40), the reactive aldehyde possessed by open chain acyl glucuronide rearrangement isomers may facilitate glycation/ glycoxidation reactions, potentially resulting in the formation of AGEs within DNA. In addition, acyl glucuronides may induce DNA damage via transition-metal-accelerated autoxidative reactions (10), which form oxygen radicals in reactions similar to those described for aqueous glucose solutions (9). Furthermore, although monosaccharides can directly damage DNA via AGE-related cross-linking and strand breakage (7), AGEs arising from the reactions of monosaccharides with proteins also exhibit genotoxic properties either via receptor-mediated pathways that involve increased production of reactive oxygen species (8) or directly via further glycoxidation reactions (7). These mechanisms would not necessarily require nuclear access by acyl glucuronides. For example, the latter could involve reactive intermediates such as dicarbonyl compounds, formed as a result of glycoxidation reactions of acyl-glucuronide-modified proteins in extranuclear compartments. The nucleophilic hydrazine AG scavenges carbonyl compounds such as the open-chain aldehydic form of sugars and reactive dicarbonyls formed during glycoxidative reactions (41) and inhibits monosaccharide-induced DNA damage during the early phase of glycation/glycoxidation reactions (42). In the present study, AG suppressed DNA damage in UGT2B7transfected cells exposed to 1 mM CA, raising the possibility

Southwood et al.

that glycation/glycoxidation reactions are involved in the induction of DNA damage by CA glucuronide. The lack of effect of the metal chelating agent DFOM on CA-induced DNA damage implies that in UGT-transfected HEK293 cells, the primary route to DNA damage by these compounds was via nonoxidative glycation mechanisms, because this compound is expected to suppress DNA damage mediated by reactive oxygen species. Recent work in mouse hepatocytes has similarly suggested a glycation/glycoxidation mechanism for DNA damage caused by CA glucuronide (10). In conclusion, this work demonstrates that four carboxylic acid drugs generate DNA damage in HEK293 cells transfected with UGT 1A3, 1A9, or 2B7. Together with previous reports (10, 12, 13), our findings with CA and bezafibrate contradict the conventional view that fibrate hypolipidemics are nongenotoxic (43). Our observations that other carboxylic acid drugs (probenecid, benoxaprofen, and nafenopin) also cause DNA damage in UGT-transfected HEK293 cells and/or isolated mouse hepatocytes (10), and those of others demonstrating that nafenopin and ciprofibrate induce sister chromatid exchanges, chromosomal aberrations, and micronuclei in rat hepatocytes (44), further raise concern over the potential UGT-dependent bioactivation of carboxyl-containing drugs in general. Importantly, many current genotoxicity assays may be unsuitable for the detection of UGT-dependent genotoxicity. For example, the glucuronidation capacity of bacterial cells used in the Ames test, or of lymphocytes used for micronucleus and sister chromatic exchange assays, is likely to be minimal. Although S9 bioactivation systems are commonly employed to enhance the metabolic capacity of in vitro assays, the cofactors necessary for glucuronidation (UDP-glucuronic acid and detergents to activate membrane bound UGTs) are rarely included, and the expression of transporters able to facilitate cellular uptake of the hydrophilic glucuronide conjugates has not been demonstrated. One strategy to address such limitations of the Ames test has involved construction of Salmonella strains that express biotransformation enzymes (45, 46). However, development of recombinant Salmonella that express UGTs has proved challenging and to date only a single enzyme (UGT1A1) has been successfully expressed (47). Thus, in vitro models such as UGTtransfected HEK293 cell lines may provide an alternate approach for reassessing the role of UGT enzymes in drug genotoxicity. At present, the clinical significance of UGT-dependent bioactivation of carboxylic acid drugs is unclear. Further work is required to determine the relationship between intracellular acyl glucuronide concentrations and genotoxic potency, as well as to investigate the kinetics of both DNA damage and repair using a range of chemically diverse carboxylic acid compounds, in addition to expanding our observations using in vivo models with multiple measures of genotoxicity. Acknowledgment. This work was funded by a Project Grant from the National Health and Medical Research Council of Australia. Hamish Southwood was the recipient of a Queen Elizabeth Hospital Honours Research Scholarship.

References (1) Sallustio, B. C., Sabordo, L., Evans, A. M., and Nation, R. L. (2000) Hepatic disposition of electrophilic acyl glucuronide conjugates. Curr. Drug Metab. 1, 163–180. (2) Bailey, M. J., and Dickinson, R. G. (2003) Acyl glucuronide reactivity in perspective: biological consequences. Chem.-Biol. Interact. 145, 117–137. (3) Spahn-Langguth, H., and Benet, L. Z. (1992) Acyl glucuronides revisited: is the glucuronidation process a toxification as well as a detoxification mechanism. Drug Metab. ReV. 24, 5–48.

UGT-Dependent DNA Damage by Carboxylic Acid Drugs (4) Akira, K., Taira, T., Hasegawa, H., Sakuma, C., and Shinohara, Y. (1998) Studies on the stereoselective internal acyl migration of ketoprofen glucuronides using 13C labelling and nuclear magnetic resonance spectroscopy. Drug Metab. Dispos. 26, 457–464. (5) Qui, Y., Burlingame, A. L., and Benet, L. Z. (1998) Mechanisms for covalent binding of benoxaprofen glucuronide to human serum albumin. Drug Metab. Dispos. 26, 246–256. (6) Smith, P. C., and Wang, C. (1992) Nonenzymic glycation of albumin by acyl glucuronides in vitro: comparisons of reactions with reducing sugars. Biochem. Pharmacol. 44, 1661–1668. (7) Lee, A. T., and Cerami, A. (1990) In vitro and in vivo reactions of nucleic acids with reducing sugars. Mutat. Res. 238, 185–191. (8) Stopper, H., Schinzel, R., Sebekova, K., and Heidland, A. (2003) Genotoxicity of advanced glycation end products in mammalian cells. Cancer Lett. 190, 151–156. (9) Wolff, S. P., and Dean, R. T. (1987) Glucose autoxidation and protein modification. The potential role of “autoxidative glycosylation” in diabetes. Biochem. J. 245, 243–250. (10) Sallustio, B. C., DeGraaf, Y. C., Weekley, J. S., and Burcham, P. C. (2006) Bioactivation of carboxylic acid compounds by UDP-glucuronosyltransferases to DNA-damaging intermediates: role of glycoxidation and oxidative stress in genotoxicity. Chem. Res. Toxicol. 19, 683–691. (11) Cayen, M. N. (1985) Disposition, metabolism and pharmacokinetics of antihyperlipidemic agents in laboratory animals and man. Pharmacol. Ther. 29, 157–204. (12) Sallustio, B. C., Harkin, L. A., Mann, M. C., Krivickas, S. J., and Burcham, P. C. (1997) Genotoxicity of acyl glucuronide metabolites formed from clofibric acid and gemfibrozil: a novel role for phaseII-mediated bioactivation in the hepatocarcionogenicity of the parent aglycones. Toxicol. Appl. Pharmacol. 147, 459–464. (13) Ghaoui, R., Sallustio, B. C., Burcham, P. C., and Fontaine, F. R. (2003) UDP-glucuronosyltransferase-dependent bioactivation of clofibric acid to a DNA-damaging intermediate in mouse hepatocytes. Chem.-Biol. Interact. 145, 201–211. (14) Radominska-Pandya, A., Pokrovskaya, I. D., Xu, J., Little, J. M., Jude, A. R., Kurten, R. C., and Czernik, P. J. (2001) Nuclear UDPglucuronosyltransferases: identification of UGT2B7 and UGT1A6 in human liver nuclear membranes. Arch. Biochem. Biophys. 399, 37– 48. (15) Jin, C., Miners, J. O., Lillywhite, K. J., and Mackenzie, P. I. (1993) Complementary deoxyribonucleic acid cloning and expression of a human liver uridine diphosphate-glucuronosyltransferase glucuronidating carboxylic acid-containing drugs. J Pharmacol. Exp. Ther. 264, 475–479. (16) Green, M. D., King, C. D., Mojarrabi, B., Mackenzie, P. I., and Tephly, T. R. (1998) Glucuronidation of amines and other xenobiotics catalysed by expressed human UDP-glucuronosyltransferase 1A3. Drug Metab. Dispos. 26, 507–512. (17) Ebner, T., and Burchell, B. (1993) Substrate specificities of two stably expressed human liver UDP-glucuronosyltransferases of the UGT1 gene family. Drug Metab. Dispos. 21, 50–55. (18) Sakaguchi, K., Green, M., Stock, N., Reger, T. S., Zunic, J., and King, C. (2004) Glucuronidation of carboxylic acid containing compounds by UDP-glucuronosyltransferase isoforms. Arch. Biochem. Biophys. 424, 219–225. (19) Uchaipichat, V., Mackenzie, P. I., Guo, X.-H., Gardenr-Stephen, D., Galetin, A., Houston, J. B., and Miners, J. O. (2004) Human UDPglucuronosyltransferases: isoform selectivity and kinetics of 4-methylumbelliferone and 1-naphthol glucuronidation, effects of organic solvents, and inhibition by diclofenac and probenecid. Drug Metab. Dispos. 32, 413–423. (20) Park, J. G., Kramer, B. S., Steinberg, S. M., Carmichael, J., Collins, J. M., Minna, J. D., and Gazdar, A. F. (1987) Chemosensitivity testing of human colorectal carcinoma cell lines using a tetrazolium-based colorimetric assay. Cancer Res. 47, 5875–5879. (21) Hanioka, N., Jinno, H., Tanaka-Kagawa, T., Nishimura, T., and Ando, M. (2001) Determination of UDP-glucuronosyltransferase UGT1A6 activity in human and rat liver microsomes by HPLC with UV detection. J. Pharm. Biomed. Anal. 25, 65–75. (22) Bailey, M. J., and Dickinson, R. G. (1996) Chemical and immunochemical comparison of protein adduct formation of four carboxylate drugs in rat liver and plasma. Chem. Res. Toxicol. 9, 659–666. (23) Singh, N. P., McCoy, M., Tice, R. R., and Schneider, E. (1988) A simple technique for quantitation of low levels of DNA damage in individual cells. Exp. Cell Res. 175, 184–191. (24) Tice, R. R., Agurell, E., Anderson, D., Burinson, B., Hartman, A., Kobayashi, H., Miyamae, Y., Rojas, E., Ryu, J. C., and Sasaki, Y. F. (2000) Single-cell gel/comet assay: guidelines for in vitro and in vivo genetic toxicology testing. EnViron. Mol. Mutagen. 35, 206–221.

Chem. Res. Toxicol., Vol. 20, No. 10, 2007 1527 (25) Burcham, P. C., and Fontaine, F. R. (2001) Extensive protein carbonylation precedes acrolein-mediated cell death in mouse haptocytes. J. Biochem. Mol.Toxicol. 15, 309–316. (26) Barbier, O., Levesque, E., Belanger, A., and Hum, D. W. (1999) UGT2B23, a novel uridine diphosphate glucuronosyltransferase enzyme expressed in steroid target tissues that conjugate androgen and estrogen metabolites. Endocrinology 140, 5538–5548. (27) Barbier, O., Belanger, A., and Hum, D. W. (1999) Cloning and characterization of a simian UDP-glucuronosyltransferase enzyme UGT2B20, a novel C19 steroid-conjugating protein. Biochem. J. 337, 567–574. (28) Miners, J. O., Lillywhite, K. J., Matthews, A. P., Jones, M. E., and Birkett, D. J. (1988) Kinetic and inhibitor studies of 4-methylumbelliferone and 1-naphthol glucuronidation in human liver microsomes. Biochem. Pharmacol. 37, 665–671. (29) Faed, E. M., and McQueen, E. G. (1979) Measurement of clofibric acid (CPIB) metabolites in plasma of patients on clofibrate therapy. Clin. Exp. Pharmacol. Physiol. 6, 267–273. (30) Oliver, M. F., Heady, J. A., Morris, J. N., and Cooper, J. (1978) A co-operative trial in the primary prevention of ischaemic heart disease using clofibrate. Br. Heart J. 40, 1069–1118. (31) Oliver, M. F., Heady, J. A., Morris, J. N., and Cooper, J. (1980) W.H.O. cooperative trial on primary prevention of ischaemic heart disease using clofibrate to lower serum cholesterol: mortality follow-up. Lancet August 23, 379–385. (32) Abshagen, U., Bablok, W., Koch, K., Lang, P. D., Schmidt, H. A. E., Senn, M., and Stork, H. (1979) Disposition pharmacokinetics of bezafibrate in man. Eur. J. Clin. Pharmacol. 16, 31–38. (33) Cunningham, R. F., Israili, Z. H., and Dayton, P. G. (1981) Clinical pharmacokinetics of probenecid. Clin. Pharmacokinet. 6, 135–151. (34) Hansen-Moller, J., and Schmit, U. (1991) Rapid high-performance liquid chromatographic assay for the simultaneous determination of probenecid and its glucuronide in urine. Irreversible binding of probenecid to serum albumin. J. Pharm. Biomed. Anal. 9, 65–73. (35) Hirano, M., Maeda, K., Shitara, Y., and Sugiyama, Y. (2004) Contribution of OATP2 (OATP1B1) and OATP8 (OATP1B3) to the hepatic uptake of pitavastatin in humans. J. Pharmacol. Exp. Ther. 311, 139–146. (36) Hagmann, W., Nies, A., Konig, J., Frey, M., Zentgraf, H., and Keppler, D. (1999) Purification of the human apical conjugate export pump MRP2. Reconstitution and functional characterization as substratestimulated ATPase. Eur. J. Biochem. 265, 281–289. (37) Zeng, H., Liu, G., Rea, P. A., and Kruh, G. D. (2000) Transport of amphipathic anions by human multidrug resistance protien 3. Cancer Res. 60, 4779–4784. (38) Henderson, L., Wolfreys, A., Fedyk, J., Bourner, C., and Windebank, S. (1998) The ability of the comet assay to discriminate between genotoxins and cytotoxins. Mutagenesis 13, 89–94. (39) Ding, A., Zia-Amirhosseini, P., McDonagh, A. F., Burlingame, A. L., and Benet, L. Z. (1995) Reactivity of tolmetin glucuronide with human serum albumin. Identification of binding sites and mechanisms of reaction by tandem mass spectrometry. Drug Metab. Dispos. 23, 369– 376. (40) Baynes, J. W. (2002) The Maillard hypothesis on aging: time to focus on DNA. Ann. N.Y. Acad. Sci. 959, 360–367. (41) Khalifah, R. G., Baynes, J. W., and Hudson, B. G. (1999) Amadorins: novel post-amadori inhibitors of advanced glycation reactions. Biochem. Biophys. Res. Commun. 257, 251–258. (42) Lee, A.T., and Cerami, A. (1987) The formation of reactive intermediate(s) of glucose 6-phosphate and lysine capable of rapidly reacting with DNA. Mutat Res. 179, 151–158. (43) Ashby, J., Brady, A., Elcombe, C. R., Elliot, B. M., Ishamael, J., Odum, J., Tugwood, J. D., Kettle, S., and Purchase, I. F. H. (1994) Mechanistically-based human hazard assessment of peroxisome proliferator-induced hepatocarcinogenesis. Human Exp. Toxicol. 13, S1– S117. (44) Reisenbichler, H., and Eckl, P. (1993) Genotoxic effects of selected peroxisome proliferators. Mutat. Res. 286, 135–144. (45) Guengerich, F. P., Gillam, E. M. J., and Shimada, T. (1996) New applications of bacterial systems to problems in toxicology. Crit. ReV. Toxicol. 26, 551–583. (46) Kamataki, T., Suzuki, A., Kushida, H., Iwata, H., Watanabe, M., Nohmi, T., and Fujita, K. (1999) Establishment of a Salmonella tester strain highly sensitive to mutagenic heterocyclic amines. Cancer Lett. 143, 113–116. (47) Fujita, K., Mogami, A., Hayashi, A., and Kamataki, T. (2000) Establishment of a Salmonella strain expressing catalytically active human UDP-glucuronosyltransferase 1A1 (UGT1A1). Life Sci. 66, 1955–1967.

TX700188X