Can In Vitro Metabolism-Dependent Covalent Binding Data

332

Chem. Res. Toxicol. 2009, 22, 332–340

Can In Vitro Metabolism-Dependent Covalent Binding Data Distinguish Hepatotoxic from Nonhepatotoxic Drugs? An Analysis Using Human Hepatocytes and Liver S-9 Fraction Jonathon N. Bauman, Joan M. Kelly, Sakambari Tripathy, Sabrina X. Zhao, Wing W. Lam, Amit S. Kalgutkar, and R. Scott Obach* Pharmacokinetics, Dynamics and Metabolism Department, Pfizer Global Research and DeVelopment, Groton, Connecticut 06340, and Pfizer Research Technology Center, Cambridge, Massachusetts 02139 ReceiVed October 31, 2008

In vitro covalent binding studies in which xenobiotics are shown to undergo metabolism-dependent covalent binding to macromolecules have been commonly used to shed light on the biochemical mechanisms of xenobiotic-induced toxicity. In this paper, 18 drugs (nine hepatotoxins and nine nonhepatotoxins) were tested for their proclivity to demonstrate metabolism-dependent covalent binding to macromolecules in human liver S-9 fraction (9000g supernatant) or human hepatocytes, as an extension to previous work that used human liver microsomes published in this journal [Obach et al. (2008) Chem. Res. Toxicol. 21, 1814-1822]. In the S-9 fraction, seven out of the nine drugs in each category demonstrated some level of metabolism-dependent covalent binding. Inclusion of reduced glutathione, cofactors needed by conjugating enzymes, and other parameters (total daily dose and fraction of total intrinsic clearance comprised by covalent binding) improved the ability of the system to separate hepatotoxins from nonhepatotoxins to a limited extent. Covalent binding in human hepatocytes showed that six out of the nine hepatotoxins and four out of eight nonhepatotoxins demonstrated covalent binding. Taking into account estimates of total daily body burden of covalent binding from the hepatocyte data showed an improvement over other in vitro systems for distinguishing hepatotoxins from nonhepatotoxins; however, this metabolism system still displayed some false positives. Combined with the previous study using liver microsomes, these findings identify the limitations of in vitro covalent binding data for prospective prediction of hepatotoxicity for new drug candidates and highlight the need for a better understanding of the link between drug bioactivation, covalent adduct formation, and toxicity outcomes. Directly relating covalent binding to hepatotoxicity is likely an oversimplification of the process whereby adduct formation ultimately leads to toxicity. Understanding underlying complexities (e.g., which macromolecules are important covalent binding targets, interindividual differences in susceptibility, etc.) will be essential to any understanding of the problem of metabolism-dependent hepatotoxicity and predicting toxicity from in vitro experiments. Introduction 1

Idiosyncratic adverse drug reactions (IADRs) continue to plague the pharmaceutical industry as a major cause of drug withdrawal from the marketplace. IADRs reflect the inability of a very small segment of patients to tolerate the drug in a manner similar to the majority of the population. Among lifethreatening IADRs, hepatotoxicity is the most common cause for drug withdrawal and accounts for 50% of the cases of acute liver failure mimicking all forms of acute and chronic liver disease (1). IADRs are not associated with known drug pharmacology, and although they are dose-dependent in susceptible individuals, they can occur at any dose within the usual therapeutic range (2). Because serious IADRs are rare (1 in 10000 to 1 in 100000 patients), they are often difficult to detect even in large clinical trials conducted prior to registration of a * To whom correspondence should be addressed. Tel: 860-441-6122. E-mail: [email protected]. 1 Abbreviations: IADRs, idiosyncratic adverse drug reactions; CLint, intrinsic clearance; PAPS, 3′-phosphoadenosine-5′-phosphosulfate; SAM, S-adenosylmethionine; UDPGA, uridine-5′-diphosphoglucuronic acid; UGT, uridine 5′-diphosphoglucuronosyl transferase; LC-MS/MS, liquid chromatography-tandem mass spectrometry; NSAID, nonsteroidal antiinflammatory drug; S-9, supernatant of liver homogenate at 9000g.

new drug. Likewise, conventional animal models of toxicity are poor predictors of IADRs in humans (3). Overall, this issue creates an urgent need for new approaches for identifying drug candidates with IADR potential. Considering that a detailed understanding of the mechanisms of IADRs remains poorly understood (4, 5), it is currently impossible to accurately predict which new drugs will be associated with a significant incidence of IADRs, and this poses a significant challenge in drug discovery/development. Despite the disadvantage, there are a few assays in place that could likely reduce the probability of occurrence of IADRs with new drugs. Because it is widely appreciated that reactive metabolites, as opposed to the parent molecules from which they are derived, are responsible for the pathogenesis of some IADRs (6-9), most pharmaceutical research organizations have implemented procedures to evaluate the potential of a new drug candidate to undergo metabolic activation (bioactivation) with the goal of eliminating or minimizing reactive metabolite formation by rational structural modification of the lead chemical class. To this end, various in vitro approaches have been proposed as screens for reactive metabolites, such as assays in which nucleophilic trapping agents (e.g., GSH, cyanide, and/or amines) are included in metabolic incubations with human hepatic tissue

10.1021/tx800407w CCC: $40.75  2009 American Chemical Society Published on Web 01/22/2009

Metabolism-Dependent CoValent Binding of 18 Drugs

and the trapped electrophilic metabolites are detected (10, 11). Likewise, it has been advocated that in vitro covalent binding to human liver microsomal proteins in the presence and the absence of NADPH be used routinely to evaluate drug candidates, with cutoff values for when concerns should be raised and included with other important factors termed “qualifying considerations” (6, 11). While reactive metabolite trapping and/or covalent binding studies have provided retrospective mechanistic insights into the possible association between drug metabolism and a resulting toxicological response (12-14), extending these approaches as predictive tools for toxicity can be challenged based upon our recently published findings on the side-by-side comparison of microsomal covalent binding by well-established hepatotoxic drugs and nonhepatotoxic drugs (15). As expected, many of the hepatotoxic drugs included in the analysis (e.g., acetaminophen, carbamazepine, diclofenac, indomethacin, nefazodone, sudoxicam, and tienilic acid) demonstrated covalent binding to microsomal protein. What was unexpected were our observations that several nonhepatotoxic drugs (e.g., buspirone, diphenhydramine, meloxicam, paroxetine, propranolol, and raloxifene) also demonstrated irreversible incorporation of radioactivity in human liver microsomes. A quantitative comparison of covalent binding in vitro intrinsic clearance (CLint) did not separate the two groups of compounds, and in fact, the highly prescribed antidepressant paroxetine (from the nonhepatotoxin group) showed the greatest amount of covalent binding in microsomes. Including factors such as the fraction of total metabolism comprised by covalent binding and the total daily dose of each drug improved the discrimination between hepatotoxic and nonhepatotoxic drugs based on in vitro covalent binding data. Nevertheless, overinterpretation of the results of such assays in drug discovery would lead to misidentification of some agents as potentially hepatotoxic. In this paper, we extend our microsomal covalent binding studies with the same set of 18 drugs to include super natant of human liver homogenate at 9000g (S-9) fractions and cryopreserved human hepatocytes. A key driver for these studies was to evaluate whether switching the biological system from liver microsomes to a “more complete” metabolizing system would improve the relationship between covalent binding and hepatotoxicity.

Experimental Procedures Chemicals. Radiolabeled materials were obtained from four sources: [14C]-Buspirone, [14C]-benoxaprofen, [14C]-carbamazepine, [3H]-diphenhydramine, [3H]-felbamate, [14C]-meloxicam, [14C]nefazodone, [14C]-propranolol, [14C]-raloxifene, [14C]-simvastatin, and [14C]-sudoxicam were custom synthesized by Nerviano Medical Sciences (Nerviano, Italy). [14C]-Acetaminophen, [14C]-diclofenac, and [14C]-ibuprofen were purchased from GE-Healthcare Biosciences Corp. (Piscataway, NJ). [3H]-Paroxetine was purchased from Perkin-Elmer (Boston, MA). [3H]-Indomethacin and [14C]theophylline were obtained from American Radiochemicals (Columbia, MO). Positions of the radionuclides in the drugs were the same as previously described (15). Acetaminophen, alamethicin, carbamazepine, diclofenac, diphenhydramine, reduced GSH, ibuprofen, indomethacin, NADPH, 3′-phosphoadenosine-5′-phosphosulfate (PAPS), paroxetine, propranolol, raloxifene, saccharolactone, S-adenosylmethionine (SAM), theophylline, and uridine-5′-diphosphoglucuronic acid (UDPGA) were obtained from Sigma (St. Louis, MO). Meloxicam, felbamate, nefazodone, simvastatin, and buspirone were purchased from Sequoia (Oxford, United Kingdom). Benoxaprofen, tienilic acid, and sudoxicam were available in the chemical bank at Pfizer Global Research and Development (Groton). Pooled human liver S-9 fraction was obtained from BD-Gentest

Chem. Res. Toxicol., Vol. 22, No. 2, 2009 333 Table 1. In Vitro Incubation Conditions Used for Covalent Binding Intrinsic Clearance Determinationsa incubation time (min)

substrate concentration range (µM)

drug

hepatocytes

S-9

hepatocytes

S-9

acetaminophen benoxaprofen buspirone diclofenac diphenhydramine indomethacin meloxicam nefazodone paroxetine propranolol raloxifene simvastatin sudoxicam tienilic acid

15 15 15 15 15 15 15 15 15 15 15 15 15 15

20 15 20 10 30 20 30 10 10 30 10 30 20 10

1–1000 1–200 1–200 1–200 1–200 1–1000 1–200 1–200 1–200 1–200 1–200 1–200 1–200 1–200

2–200 3–200 2–200 2–200 2–200 1–200 4–200 2–200 1–200 3–200 3–200 3–171 2–200 2–200

a Carbamazepine, felbamate, and theophylline are not listed since binding was not observed. Ibuprofen did not show binding in S-9, and CLint for covalent binding was not determined in hepatocytes.

(Woburn, MA). Cryopreserved human hepatocytes from three donors were purchased from In Vitro Technologies (Baltimore, MD), and Leibovitz L-15 media was from Invitrogen Corp. (Carlsbad, CA). Intrinsic Clearance Determinations for Covalent Binding to Liver S-9 Fraction. Initial determinations were made to define the incubation time that provided linearity for covalent binding (Table 1). Radiolabeled substrates at varying concentrations (Table 1) were incubated with pooled human liver S-9 fraction (5.0 mg/mL) in 1.0 mL of 0.1 M potassium phosphate buffer, pH 7.45, containing 3.3 mM MgCl2 and 1.3 mM NADPH. Incubations were commenced with the addition of S-9 fraction and conducted in a shaking water bath at 37 °C open to air. All incubations were conducted in duplicate. Incubations were terminated by addition of 5 mL of methanol/acetonitrile (4:1) to precipitate protein and salts. The mixtures were spun in a centrifuge (2100g) for 5 min, the supernatants were decanted, and the pellets were resuspended in methanol/acetonitrile by mixing on a vortex mixer and using metal wire to aid in disruption of the pellets. This cycle of spinning in a centrifuge, discarding the supernatant, and resuspension of the pellet was repeated six more times. The final pellet was dissolved in 1 mL of NaOH (1M) overnight, and the radioactivity was determined by mixing the alkaline digest with 1 mL of water and 18 mL of Optiphase scintillation fluid (Perkin-Elmer). In some experiments, the effects of various cofactors and additives on covalent binding were also assessed. Thus, in human liver S-9 fraction, covalent binding was determined in the presence of 0.1 mM SAM, 0.1 mM PAPS, 5 mM GSH, 1 mM semicarbazide, 0.1 mM potassium cyanide, or conditions to favor activity of uridine glucuronosyl transferase (UGT) enzymes (3 mM UDPGA, 0.05 mg/ mL alamethicin, and 1 mM saccharolactone). Total Intrinsic Clearance Determinations in Liver S-9 Fraction. In vitro metabolic lability determinations, in duplicate, were done using the same conditions as the covalent binding incubations. Incubations at a substrate concentration of 1 µM were commenced with the addition of NADPH, and aliquots of the incubation mixture were removed at t ) 0 and various time points thereafter, up to 1 h, and added to acetonitrile to terminate the reactions. The precipitated protein was removed by spinning in a centrifuge, and the supernatant was analyzed by liquid chromatography-tandem mass spectrometry (LC-MS/MS) for remaining parent drug to yield in vitro t1/2 values used in the calculation of CLint. The system consisted of two pumps (Shimadzu Inc., Columbia, MD), a Shimadzu SIL-5000 autoinjector (Shimadzu Inc.), and a Sciex API4000 Q-trap mass spectrometer (PE-Sciex, Ontario, Canada). Samples were injected onto a Optilynx C18 guard column (1 mm × 15 mm, 13 µm; Optimize Technologies, Oregon City, OR) equilibrated in 98% 2 mM ammonium acetate containing

334

Chem. Res. Toxicol., Vol. 22, No. 2, 2009

2% CH3OH at a flow rate of 1.5 mL/min. Elution was done by increasing the organic solvent composition to 90% 50/50 CH3OH/ CH3CN. The analytes were detected using selected reaction monitoring. For some compounds, the rate of turnover was too slow to permit a reliable determination of in vitro t1/2 in this manner. In these instances, radiolabeled drug was incubated as before, and analysis was conducted by HPLC with radiometric detection. The system consisted of a Surveyor HPLC quaternary pump (Thermo, Billerica, MA) and a Polaris C18 column (4.6 mm × 250 mm, 5 µm; Varian, Lake Forest, CA). The mobile phase was comprised of 0.1% formic acid in water and CH3CN at a flow rate of 0.8 mL/min. Mobile phase gradients were customized to optimize separation of parent drug from metabolites for each compound within a 30 min run time. The eluent was collected into 20 s fractions into 96 well Scintiplates (Wallac, Turku, Finland), the solvent was evaporated in a vacuum centrifuge (Genevac, Valley Cottage, NY), and the plates were counted in a Wallac Microbeta scintillation counter. The percentages represented by metabolite peaks were summed, and the data were used to describe the first-order decline in parent drug with time. Covalent Binding in Human Hepatocytes. Pooled cryopreserved human hepatocytes from three donors (In Vitro Technologies) were suspended in Leibovitz L-15 media (Invitrogen Corp.) at a final concentration of 0.5 × 106 cells/mL. Cell viability based on trypan blue exclusion was >80%. The cell suspensions (4 mL) were incubated at 37 °C for up to 4 h with radiolabeled substrate (1 µM). Aliquots were removed at different time points, centrifuged at 3500 rpm for 15 min, and exhaustively extracted (up to N ) 7) with organic solvent (CH3CN:CH3OH, 4:1). Supernatant fractions were monitored until no radioactivity remained. Sodium hydroxide (1 M) was added to the remaining protein pellet and placed in a water bath overnight to dissolve the pellet, and the total radioactivity was determined by liquid scintillation counting. To determine covalent binding intrinsic clearance, the cell suspensions (1.25 mL, 0.5 × 106 cells/mL) were incubated for 15 min with radiolabeled substrate at a final concentration range of 1-200 or 1-1000 µM. The incubations were terminated and analyzed as above. To determine total intrinsic clearance, the cell suspensions (1 mL, 0.5 × 106 cells/mL) were incubated for 4 h with substrate (1 µM). Acetaminophen, benoxaprofen, carbamazepine, diphenhydramine, felbamate, ibuprofen, indomethacin, meloxicam, and theophylline were performed at 1 × 106 cells/mL. Aliquots were removed at different time points, quenched with acetonitrile containing internal standard, and centrifuged at 2000 rpm for 15 min. The LC system consisted of Shimadzu LC20AD pumps, DGU20A5 degasser and VP Option box (Columbia, MD), an HTC PAL autosampler (Leap Technologies, Cary, NC), and a Phenomenex Gemini HPLC column (2.0 mm × 50 mm, 5 µm; Phenomenex, Torrance, CA). HPLC mobile phase A was 10 mM ammonium formate:IPA:formic acid in water (98.9:1.0:0.1, %v/v), and mobile phase B was formic acid in acetonitrile (0.1%). The flow rate was 0.5 mL/min. The HPLC gradient started at 10% B for 0.5 min, was ramped linearly to 100% B over 0.3 min, was held at 100% B over 0.7 min, and then was returned to the initial conditions over 0.25 min. The LC system was interfaced to an API 4000 Q-trap mass spectrometer (Sciex, Toronto, Canada) equipped with the Turboionspray source. Supernatants were injected onto LC/MS, area ratios determined by MRM of substrate and internal standard. Data Analysis. The following equation was used to calculate radiolabeled covalent binding (pmol/mg protein):

radioactivity in pellet (dpm) × substrate concentration (nmol/mL) × 1000 total radioactivity added to incubation (dpm) × protein concentration (mg/mL) (1) The covalent binding data were transformed to reaction velocity values by correcting for incubation time and protein concentration. Duplicate measurements were averaged, and individual values did

Bauman et al. not differ by more than 15% of each other. The velocity vs substrate concentration values were fit to the Michaelis-Menten equation or the Michaelis-Menten equation with an additional nonsaturable term (CLint,cb,2):

V)

Vmax × [S] Vmax × [S] or V ) + CLint,cb,2 × [S] KM + [S] KM + [S]

(2)

Intrinsic clearance of covalent binding (CLint,cb,1) was calculated as Vmax/KM, with total covalent binding intrinsic clearance being the sum of CLint,cb,1 and CLint,cb,2 when the second nonsaturable term was obtained from the velocity vs substrate concentration curve in hepatocytes. The fraction of metabolism comprised by covalent binding (fCL,cb) was calculated as:

fCL,cb )

CLint,cb CLint + CLint,cb

(3)

in which CLint,cb is the covalent binding intrinsic clearance and CLint is the intrinsic clearance for S-9 metabolism as determined by:

CLint )

0.693 × incubation volume t1⁄2 × mg protein

(4)

or for hepatocyte metabolism by:

CLint )

[S] V × AUC number of cells

(5)

in which [S] is the initial substrate concentration, V is the incubation volume, and AUC is the area under the [S] vs time curve. A hypothetical estimate of “total daily body burden of covalent binding” (Dcb) was calculated as:

Dcb)fCL,cb × daily dose

(6)

This value was also modified by the inclusion of other cofactors in S-9 incubations that could reduce covalent binding in vivo (Dcb,corr), by multiplying the total daily body burden of covalent binding by the fraction of covalent binding observed in the presence of these cofactors (fCB,cofactor n) in which n is GSH, SULT, MT, or the set of cofactors needed to support UGT activity:

Dcb,corr)Dcb × fCB,GSH × fCB,UGT × fCB,SULT × fCB,MT

(7)

The analysis does not include other clearance pathways (e.g., extrahepatic metabolism, renal secretion, biliary secretion) and, as such, provides upper-limit estimates.

Results Covalent Binding in Human Liver S-9 Fraction. A summary of the covalent binding intrinsic clearance data is presented in Table 2. With the exception of felbamate and carbamazepine, all of the remaining hepatotoxins showed NADPH-dependent measurable covalent binding. In the nonhepatotoxin set, all compounds with the exception of ibuprofen and theophylline displayed some form of covalent binding to S-9 in a NADPH-dependent fashion. Nefazodone, raloxifene, and tienilic acid showed the greatest rate of covalent binding. While nefazodone showed the greatest rate of covalent binding, it also showed a very high rate of overall metabolism, and the fraction of nefazodone CLint comprised by covalent binding is lower than many of the other compounds tested. There was no apparent trend dissociating the hepatotoxins from nonhepatotoxins with regard to total covalent binding CLint (Figure 1). Effect of Conjugative Enzyme Activity and Nucleophiles on Covalent Binding in S-9 Fraction. Unlike liver microsomes, which were utilized in the previous investigation (15), liver S-9 fraction possesses a richer complement of xenobiotic-metabolizing enzymes that can detoxicate reactive metabolites and/or their


Chem. Res. Toxicol., Vol. 22, No. 2, 2009 335

Table 2. Determination of the Enzyme Kinetics for Covalent Binding and Total Metabolic Intrinsic Clearance in Human Liver S-9 Fractiona drug acetaminophen benoxaprofen carbamazepine diclofenac felbamate indomethacin nefazodone sudoxicam tienilic acid

KM

Vmax

CLint,cb,1 metabolism CLint

hepatotoxins 1.0 0.0048 1.1 0.0092 NB 7.5 0.090 0.012 NB 77 1.2 0.015 18 15 0.83 10 0.055 0.0055 21 4.6 0.22

210 120

nonhepatotoxins buspirone 22 1.6 0.072 diphenhydramine 1.2 0.066 0.055 ibuprofen NB meloxicam 130 0.98 0.0075 paroxetine 2.6 0.28 0.11 propranolol 3.6 0.53 0.15 raloxifene 12 9.2 0.77 simvastatin 99 10 0.10 theophylline NB

Table 3. Effect of Other Enzyme Activities and Nucleophile Trapping Agents on Covalent Binding of 14 Drugs in Human Liver S-9 Fractiona

fCL,cb

covalent binding (% of control) drug

0.075 0.45 0.40 17 NM 0.54 680 2.8 4.3 8.9 2.3 9.5 5.9 1.9 3.5 2.5 260 0.006

0.060 0.020 ND 0.00071 ND 0.027 0.0012 0.0020 0.049 0.0080 0.023 0 0.0013 0.055 0.041 0.24 0.00038 ND

a Units are KM (µM), Vmax (pmol/min/mg), and CLint (µL/min/mg). NB, no binding observed; ND, parameter not determined; and NM, no metabolism observed.

Figure 1. Comparison of hepatotoxic and nonhepatotoxic drugs with regard to covalent binding intrinsic clearance in human liver S-9 fraction. Abbreviations: APAP, acetaminophen; BEN, benoxaprofen; BUS, buspirone; CAR, carbamazepine; DIC, diclofenac; DIPH, diphenhydramine; FEL, felbamate; IBU, ibuprofen; INDO, indomethacin; MEL, meloxicam; NEF, nefazodone; PAR, paroxetine; PRO, propranolol; RAL, raloxifene; SIM, simvastatin; SUD, sudoxicam; TA, tienilic acid; and THEO, theophylline.

precursors. The impact of supplementing various cofactors and conditions needed for non-P450 drug-metabolizing enzymes on covalent binding is included in Table 3, along with the impact of various nucleophilic trapping agents. In addition to the effect

UGT

SULT

MT

GSH

KCN

SCZ

Inc. 64 37 15 32 39 57

60 87 115 111 93 105 103

60 44 107 81 102 109 108

67 97 36 10 32 28 60

69 62 93 73 102 101 96

111 53 104 83 99 92 85

acetaminophen benoxaprofen diclofenac indomethacin nefazodone sudoxicam tienilic acid

88 Inc. 176 68 88 98 90

hepatotoxins 78 53 64 61 91 111 75 112 94 92 92 99 103 115

buspirone diphenhydramine meloxicam paroxetine propranolol raloxifene simvastatin

70 106 78 72 48 65 195

nonhepatotoxins 118 100 21 91 83 116 119 33 108 77 83 64 104 105

a UGT, uridine glucuronosyl transferase: inclusion of UDPGA (3 mM), alamethicin (20 µg/mL), and saccharolactone (1 mM); SULT, sulfotransferase: inclusion of PAPS (0.1 mM); MT, methyl transferase: inclusion of SAM (0.1 mM); GSH, inclusion of glutathione (5 mM); KCN, inclusion of potassium cyanide (1 mM); and SCZ, inclusion of semicarbazide (1 mM). Felbamate, carbamazepine, ibuprofen, and theophylline were not tested since covalent binding was not observed. Inc., inconclusive results. For acetaminophen and benoxaprofen, covalent binding was below the time zero control in the presence of GSH and UGT, respectively. An explanation for this observation is not apparent.

on covalent binding, this information can also yield insight into potential bioactvation mechanisms, leading to electrophilic intermediates. To varying degrees, GSH caused a decrease in covalent binding for all of those drugs that demonstrated covalent binding. For acetaminophen, covalent binding was completely abolished with inclusion of GSH. Inclusion of SAM to enable the catalytic activity of methyltransferases had a large effect on the covalent binding of paroxetine (67% decrease) and acetaminophen (47% decrease). Inclusion of PAPS to support sulfotransferase activity caused a 79% decrease in diphenhyrdramine covalent binding. Glucuronidation appeared to show a decrease in covalent binding of propranolol (52% decrease) and completely blocked covalent binding of benoxaprofen to S-9 fraction. Addition of UDPGA and other cofactors necessary for UGT activity led to an increase in covalent binding of diclofenac and simvastatin. Hard nucleophiles cyanide anion and semicarbazide caused decreases in the covalent binding of diphenhydramine, benoxaprofen, paroxetine, and acetaminophen. Covalent Binding in Human Hepatocytes. In pooled cryopreserved human hepatocytes, covalent binding was observable for six of nine hepatotoxins and for four of eight nonhepatotoxins (Table 3), and trends for the drugs are shown in Figure 2. (Data for covalent binding of ibuprofen in hepatocytes showed an unusual behavior indicating the possibility of covalent binding; however, the covalent adduct(s) between protein and the small molecule appeared to be unstable. Further investigation of this drug is ongoing, and it is not included in these analyses.) Among the known hepatotoxins, tienilic acid showed the greatest rate of covalent binding. Sudoxicam, felbamate, and carbamazepine did not show measurable covalent binding. Carbamazepine also did not show metabolism in hepatocytes, making interpretation of covalent binding results difficult (i.e., turnover may be too slow to detect covalent binding). However, acetaminophen, benoxaprofen, and indomethacin also did not show measurable consumption or

336


Bauman et al. Table 4. Determination of the Enzyme Kinetics for Covalent Binding and Total Metabolic Intrinsic Clearance in Human Hepatocytesa drug acetaminophen benoxaprofen carbamazepine diclofenac felbamate indomethacin nefazodone sudoxicam tienilic acid buspirone diphenhydramine ibuprofen meloxicam paroxetine propranolol raloxifene simvastatin theophylline

metabolism CLint

fCL,cb

hepatotoxins ND ND 0.47 0.47 3.7 0.29 0.032 0.11 NB 1.4 0.30 0.046 0.26 NB 880 180 ND 0.21 99 10 ND 0.10 NB 0.76 0.88 0.082 1.2

NM NM NM 16 NM NM 17 2.0 1.9

1.0 1.0 0 0.016 0 1.0 0.0059 0 0.63

nonhepatotoxins 0.021 0.32 NB ND ND ND NB NB ND 0.13 0.13 0.99 0.027 0.13 4.2 ND 0.11 NB

5.0 0.83 1.8 1.2 1.4 5.3 18 27 NM

0.064 0 ND 0 0 0.025 0.0072 0.0041 0

KM

11

Vmax CLint,cb,2 CLint,cb,total

3.2

ND ND 9.4 38

a Units are KM (µM), Vmax (pmol/min/million cells), and CLint (µL/ min/million cells). NB, no binding observed; ND, parameter not determined; and NM, no metabolism observed.

Table 5. Estimates of Daily Dose Equivalents of Covalent Binding Based on the In Vitro Covalent Binding Intrinsic Clearance, Fraction of Metabolism Comprised by Covalent Binding, and the Total Daily Dosea Figure 2. Comparison of hepatotoxic and nonhepatotoxic drugs with regard to covalent binding intrinsic clearance in human hepatocytes. Abbreviations: APAP, acetaminophen; BEN, benoxaprofen; BUS, buspirone; CAR, carbamazepine; DIC, diclofenac; DIPH, diphenhydramine; FEL, felbamate; IBU, ibuprofen; INDO, indomethacin; MEL, meloxicam; NEF, nefazodone; PAR, paroxetine; PRO, propranolol; RAL, raloxifene; SIM, simvastatin; SUD, sudoxicam; TA, tienilic acid; and THEO, theophylline.

formation of metabolites, despite showing measurable covalent binding. The four nonhepatotoxins that demonstrated covalent binding in human hepatocytes were propranolol, buspirone, raloxifene, and simvastatin. Among the nonhepatotoxins that did not demonstrate covalent binding, consumption of the substrate was observed for three (meloxicam, diphenhydramine, and paroxetine). Estimations of Total “Body Burden” of Covalent Binding. Consideration of the total covalent binding CLint, the fraction of total consumption that is comprised by covalent binding, and correcting this for the total daily dose yield a metric that, with acceptance of some assumptions and simplifications, should reflect a hypothetical estimate of total body burden of covalent binding. This parameter is listed in Table 5 for each of the 18 drugs in this analysis, using the S-9 and hepatocyte data. For S-9 covalent binding data, a slight trend can be observed when comparing the hepatotoxins and nonhepatotoxins (Figure 3; left side); however, the fidelity is not high, and there is considerable overlap between the two groups of compounds. Addition of cofactors to support a more complete complement of drug-metabolizing enzymes caused a decrease in the estimated covalent binding body burden values for almost all of the compounds, and in case of two compounds, these cofactors reduced covalent binding to an undetectable level. However, inclusion of the cofactors did not change the overall comparison of hepatotoxins and nonhepatotoxins (Figure 3, right side). Rather, this merely decreased the magnitude of the estimates.

estimated daily body burden of covalently bound material (mg)c calculated from S-9 data daily dose (mg)

no cofactors

with cofactors

calculated from hepatocytes

acetaminophen benoxaprofen carbamazepine diclofenac felbamate indomethacin nefazodone sudoxicam tienilic acid

4000 600 1200 200 3600 200 600 50 1000

hepatotoxins 240 12 none 0.14 none 5.4 0.72 0.10 49

none none none 0.093 none 0.46 0.18 0.035 30

4000 600 none 3.2 none 200 3.6 none 630

buspirone diphenhydramine ibuprofen meloxicam paroxetine propranolol raloxifene simvastatin theophylline

60 150 3200 15 50 320 60 80 600

nonhepatotoxinsb 0.48 3.5 none 0.020 2.8 13 14 0.030 none

0.27 0.68 none 0.0054 0.078 1.7 1.4 0.038 none

3.8 none NDa none none 8.0 0.43 0.33 none

drug

a ND, not determined. b Simvastatin is not considered as a bioactivated hepatotoxin; however, it can cause effects on the liver presumably via its pharmacological mechanism. c The estimate of daily body burden to covalent binding is made by multiplying the total daily dose by the fraction of total intrinsic clearance comprised of the covalent binding intrinsic clearance.

For hepatocytes, there is a trend that the highest values for “estimated covalent binding body burden” are observed for some of the hepatotoxins (acetaminophen, tienilic acid, benoxaprofen, and indomethacin). However, the drugs do not sort in a dichotomous fashion, and there is considerable overlap between the hepatotoxins and the nonhepatotoxins. Furthermore, three hepatotoxins did not show appreciable covalent binding (su-


Figure 3. Categorization of hepatotoxins and nonhepatotoxins based on estimated total daily body burden of covalent binding in human liver S-9 fraction with (right) and without (left) cofactors. Abbreviations: APAP, acetaminophen; BEN, benoxaprofen; BUS, buspirone; CAR, carbamazepine; DIC, diclofenac; DIPH, diphenhydramine; FEL, felbamate; IBU, ibuprofen; IND, indomethacin; MEL, meloxicam; NEF, nefazodone; PAR, paroxetine; PRO, propranolol; RAL, raloxifene; SIM, simvastatin; SUD, sudoxicam; TA, tienilic acid; and THEO, theophylline.

doxicam, felbamate, and carbamazepine). If a 1 mg/day cutoff value were used, there are eight compounds above that value (six hepatotoxins and two nonhepatotoxins) and nine compounds below that value (three hepatotoxins and six nonhepatotoxins). This would reflect a sensitivity of 67% (identified six of nine hepatotoxins) and a selectivity of 71% (correctly categorized 12 of the 17 compounds). Clearly, the performance of a hepatocyte-based covalent binding assay to distinguish hepatotoxins from nonhepatotoxins would require correction for fraction of metabolism yielding covalent binding and the daily dose. Attempting to do this by comparison of covalent binding rate alone (Figure 2) or by using S-9 fraction (Figure 3) would not yield an ability to classify compounds. This was the case for the use of liver microsomes as well (15).

Discussion Seven of the nine hepatotoxic drugs displayed some degree of NADPH-dependent covalent binding to human liver S-9 fraction, which, in all cases (e.g., acetaminophen, diclofenac, indomethacin, nefazodone, and tienilic acid), was attenuated upon the inclusion of GSH. These findings are generally consistent with the covalent binding data generated in liver microsomes (15) and the well-established P450-mediated bioactivation pathways for these agents (10). NADPH-dependent covalent binding to liver S-9 by the nonsteroidal anti-inflam-


Figure 4. Categorization of hepatotoxins and nonhepatotoxins based on estimated total daily body burden of covalent binding from human hepatocyte data. Abbreviations: APAP, acetaminophen; BEN, benoxaprofen; BUS, buspirone; CAR, carbamazepine; DIC, diclofenac; DIPH, diphenhydramine; FEL, felbamate; IND, indomethacin; MEL, meloxicam; NEF, nefazodone; PAR, paroxetine; PRO, propranolol; RAL, raloxifene; SIM, simvastatin; SUD, sudoxicam; TA, tienilic acid; and THEO, theophylline. Ibuprofen was not included.

matory drug (NSAID) benoxaprofen was an interesting finding, considering that the NSAID did not reveal a similar characteristic in NADPH-supplemented human liver microsomes (15). Furthermore, inclusion of nucleophilic trapping agents (semicarbazide and GSH) and cofactors for non-P450 enzymes (SAM and PAPS) appeared to reduce covalent binding of benoxaprofen, suggesting a possibility that a reactive metabolite(s) or its precursor(s) generated in the presence of NADPH is detoxicated in the S-9 system. Our findings suggest that glucuronidation decreases bioactivation of this NSAID rather than increasing it (presumably through acyl glucuronide formation), as was observed in the case of diclofenac and simvastatin. While acyl glucuronidation has been proposed as a bioactivation mechanism responsible for benoxaprofen toxicity (16-18), it is also noteworthy to point out that acyl glucuronidation was a major route of metabolic clearance in dog, a species used in toxicity evaluation (32), suggesting that this pathway may not be the one important for the hepatotoxicity associated with this NSAID. Consequently, providing an insight into the observed covalent binding of benoxaprofen to human hepatocytes is difficult; it is possible that irreversible binding to hepatocytes occurs via a combination of more than one electrophilic species derived from phase I (NADPH-dependent) metabolic processes, since supplementation with GSH or cofactors needed for sulfation and methylation reduced covalent binding. Also of some interest were the observations that the antiepileptic drugs felbamate and carbamazepine were devoid of covalent binding attributes in both S-9 and hepatocytes. Lack of covalent binding of these agents to cryopreserved human

338


hepatocytes was also noted in a previously published study by Leone et al. (19). While the observations on the lack of covalent binding of felbamate to S-9 and hepatocytes are consistent with previous results in human liver microsomes (15, 19), the absence of covalent binding to S-9 and hepatocytes with carbamazepine contrasts previous observations in liver microsomes from human and mice wherein the antiepileptic agent exhibited NADPHdependent binding (15, 20) in a manner consistent with its known bioactivation pathways involving cytochrome P450 (20, 21). The specific reason(s) for differences in carbamazepine covalent binding in microsomes vs S-9 and/or hepatocytes is difficult to explain at this point. In the case of felbamate, the bioactivation pathway leading to the reactive metabolite (2phenylpropenal) involves non-P450 enzymes including esterase(s) and alcohol dehydrogenase(s). Consequently, we anticipated that switching the metabolism system from liver microsomes to liver S-9 and/or hepatocytes may capture felbamate bioactivation and subsequent covalent binding. However, as judged from the lack of metabolic turnover (and covalent binding), this does not seem to be the case. It is also interesting to note that there are no studies to date that have characterized the GSH adduct of 2-phenylpropenal in in vitro incubations with human hepatic tissue (liver microsomes, liver S-9, and/or hepatocytes). Some differences in S-9 covalent binding characteristics were also discerned with the carboxylic acid-containing NSAIDs and known hepatotoxins indomethacin and diclofenac. Consistent with the known oxidative bioactivation mechanisms involving conversion to quinone-imine intermediates (22, 23), both NSAIDs exhibited NADPH-dependent covalent binding to S-9 that was dramatically reduced upon inclusion of GSH in the incubations. However, inclusion of UDPGA enhanced covalent binding to S-9 by diclofenac but not indomethacin. A similar behavior was noted in liver microsomal incubations with the two NSAIDs in the presence of UDPGA. The reason(s) for this anomaly remains unclear, especially since the carboxylic acid moiety in both NSAIDs is subject to acyl glucuronidation and both of these electrophilic metabolites have been shown to bind to hepatic tissue (24, 25). With the exception of ibuprofen and theophylline, all remaining nonhepatotoxins displayed NADPH-dependent covalent binding to S-9 in a similar fashion as that noted previously in the liver microsome study (15). Attenuation of covalent binding upon inclusion of GSH in the S-9 incubations for paroxetine, meloxicam, propranolol, and raloxifene is consistent with the role of the thiol nucleophile in scavenging electrophilic intermediates generated via bioactivation by P450 (26-29). Attenuation caused by addition of SAM to support methyltransferase activity was observed for four drugs (paroxetine, acetaminophen, benoxaprofen, and raloxifene). The findings with paroxetine and raloxifene can be rationalized by the potential formation of catechols that can be methylated before being further oxidized to reactive ortho-quinones. For acetaminophen and benoxaprofen, generation of catechol intermediates is not known, so the effect of SAM in covalent binding attenuation requires further investigation for these compounds. Addition of PAPS to support sulfation caused a decrease in binding for diphenhydramine and benoxaprofen. The findings showing considerable binding of diphenhydramine (in S-9 or in microsomes; ref 15) was unexpected since this drug has been used safely for decades. While some information on the metabolism of diphenhydramine is available (31), it is limited, and it is possible that multiple oxidations occur on the phenyl groups to yield phenols, catechols, and/or hydroquinones that are sulfated or otherwise

Bauman et al.

Figure 5. Categorization of hepatotoxins and nonhepatotoxins based on the observation of covalent binding and substrate turnover in hepatocytes.

further oxidized to quinones that could react with GSH. Further delineation of the metabolism of diphenhydramine to better understand these findings is underway. Assessment of covalent binding in cryopreserved human hepatocytes revealed no improvement in distinguishing hepatotoxins from nonhepatotoxins using CLint of covalent binding (Figure 3). Covalent binding was observed for nontoxins buspirone, propranolol, raloxifene, and simvastatin, while no binding was observed for hepatotoxins felbamate, carbamazepine, or sudoxicam. However, when corrected for a fraction of total CLint comprised by covalent binding and the daily dose, hepatotoxins and nonhepatotoxins are more readily separated by hepatocytes than by S-9 or microsomes (15). If a cutoff of 1 mg of estimated daily body burden is used, the overall sensitivity and selectivity for distinguishing hepatotoxins from nonhepatotoxins would be 67 and 71%, respectively. Nevertheless, a challenge for in vitro covalent binding assays is the potential for very low overall rates of metabolism, confounding the results. For the hepatocyte data, a quaternary classification system can be proposed (Figure 5) in which drugs are categorized by whether there was covalent binding and whether substrate consumption was observed. The worst cases would be those compounds exhibiting detectable covalent binding with no measureable turnover; only three drugs met this criteria, and they were all hepatotoxins (acetaminophen, benoxaprofen, and indomethacin). The best cases would be drugs for which turnover is observed but covalent binding is not detected. Four drugs met this criteria, but one of them was a hepatotoxin (sudoxicam). However, diphenhydramine, paroxetine, and meloxicam also fell in this category, revealing a different outcome for these drugs from what was observed using S-9 or microsome systems. For compounds that show no binding but also show no turnover, the interpretation would be ambiguous, since the lack of observation of binding could be due to no turnover (carbamazepine, felbamate, and theophylline). Finally, another set of drugs (including both hepatotoxins and nonhepatotoxins)


showed covalent binding along with overall substrate consumption. Categorization of drugs within this set would be challenging, and considerations of the fraction of metabolism comprised by covalent binding and the total daily dose would likely be needed to provide the needed context to the in vitro findings. Furthermore, the success of such a categorization would be confined to the limitations of the in vitro system; for example, slowly metabolized compounds could show detectable covalent binding if the activities of drug-metabolizing enzymes could be sustained for longer to permit longer incubation times. There are several reports that describe the use of liver microsomes to assess cytochrome P450-mediated reactive metabolite formation and covalent binding potential of drugs (6, 14, 15, 30). Earlier reports focused only on demonstrating covalent binding for known toxic compounds, while more recent reports have compared toxic and nontoxic compounds for covalent binding. In work in our laboratory, liver microsomes failed to separate hepatotoxins from nonhepatotoxins, and several nontoxins showed considerable covalent binding relative to the toxins (15). Generating covalent binding CLint data, as opposed to covalent binding rates at single substrate concentrations, and coupling that parameter with the contribution to overall clearance and total daily dose improved the ability of in vitro covalent binding data to distinguish between hepatotoxins and nonhepatotoxins; however, there was still considerable overlap between the two groups of compounds. We used a metric “estimated daily body burden of covalent binding” as a means to account for the fraction of total metabolism represented by covalent binding and the total daily dose in attempting to understand the potential relationship between in vitro covalent binding and in vivo hepatotoxicity. As such, this is merely a contrived index and not an attempt to actually compute in vivo covalent binding. It does sort hepatotoxic and nonhepatotoxic drugs better than just considering in vitro covalent binding rates alone. Recent work by Takakusa et al. (30) examined covalent binding in liver microsomes and included four nontoxins along with 16 toxins at a single concentration and incubation time. While many of the toxins showed greater covalent binding than the four nontoxins, there were still several that showed low covalent binding comparable to the nontoxins (e.g., phenytoin, valproic acid, zomepirac, and others). These investigators extended their examination into assessing covalent binding in vivo using a rat model and suggested that covalent binding observed in vivo correlated to that generated in rat liver microsomes. The fact that covalent binding measurements alone do not yield an approach whereby toxic and nontoxic drugs can be delineated, nor can a correlation be described between binding and toxicity, provides evidence that there are other factors besides the amount of covalent binding contributing to hepatotoxicity. Hepatotoxicity is a complex phenomenon caused by multiple possible mechanisms. For example, specific targets to which drugs form adducts are likely important in eliciting toxicity rather than gross covalent binding. Furthermore, some hepatotoxins that demonstrate bioactivation and covalent binding in vitro may actually cause their hepatotoxic effects by depletion of GSH or causing a redox imbalance in hepatocytes rather than the covalent binding itself. While the detection of a bioactivation phenomenon using reactive metabolite trapping and/or covalent binding studies is relatively straightforward, the utility of these assays as screening tools to predict toxicity comes into question since several drugs form reactive metabolites that covalently bind to proteins but only a fraction thereof cause toxicity. However, in drug discovery, these assays can be very useful from a mechanistic


perspective to assess bioactivation pathways for new chemical entities associated with safety liabilities (e.g., S-9/NADPHdependent mutagenicity in the Salmonella Ames assay, mechanism-based inactivation of P450 enzymes, or even preclinical species hepatotoxicity); this in turn can influence the rational design of successor drug candidates devoid of the bioactivation liability, thus eliminating potential safety concerns. In addition, in early stages of drug discovery (hit to lead or lead optimization), high-throughput reactive metabolite trapping studies provide a convenient means to assess bioactivation potential of new compounds that possess obvious structural alerts/toxicophores (e.g., furans, anilines, etc.), information that can influence subsequent high-speed chemistry design wherein the toxicophores are replaced with appropriate bioisosteres devoid of the bioactivation liability. To this point, liver S-9 fraction and hepatocytes have received much less attention as matrices to study bioactivation phenomenon as compared to liver microsomes. However, there may be advantages since these systems possess a more complete complement of drug-metabolizing enzymes, whereas for liver microsomal covalent binding assays, only the enzymes that can bioactivate drugs to reactive intermediates are active (i.e., cytochromes P450), which may give artifactual results for some compounds. For example, raloxifene appeared to be one of the worst drugs when assessed in liver microsomes; however, when hepatocytes were used, covalent binding for this drug was one of the lowest. Among the systems that we have tested, pooled human hepatocytes may give the best results for such an assessment; however, even in this system, there is an imperfect fidelity in distinguishing hepatotoxins from nonhepatotoxins. Consideration of dose and fractional contribution of covalent binding to overall metabolism is likely important and attempts to address the extent to which binding could occur in vivo rather than merely comparing rates of genesis of covalent binding. Additionally, if in vitro covalent binding (or generation of trapped reactive intermediate) is not observed, it is important to ensure that adequate turnover of the test compound occurred, as illustrated by the examples of felbamate and carbamazepine in which binding was not observed in hepatocytes but neither was substrate consumption. Nevertheless, it is clear that directly relating covalent binding to toxicity is an oversimplification of the process whereby adduct formation ultimately leads to toxicity. Understanding these complexities (e.g., which macromolecules are important covalent binding targets and why, interindividual differences in susceptibility, etc.) will be essential to any understanding of the problem of metabolism-dependent hepatotoxicity and predicting toxicity from in vitro experiments.

References (1) Abboud, G., and Kaplowitz, N. (2007) Drug-induced liver injury. Drug Saf. 30, 277–294. (2) Uetrecht, J. P. (2000) Is it possible to more accurately predict which drug candidates will cause idiosyncratic drug reactions? Curr. Drug Metab. 1, 133–141. (3) Shenton, J. M., Chen, J., and Uetrecht, J. P. (2004) Animal models of idiosyncratic drug reactions. Chem.-Biol. Interact. 150, 53–70. (4) Liquori, M. J., and Waring, J. F. (2006) Investigations toward enhanced understanding of hepatic idiosyncratic drug reactions. Expert Opin. Drug Metab. Toxicol. 2, 835–846. (5) Ju, C., and Uetrecht, J. P. (2002) Mechanism of idiosyncratic drug reactions: Reactive metabolite formation, protein binding and the regulation of the immune system. Curr. Drug. Metab. 3, 367–377. (6) Evans, D. C., Watt, A. P., Nicoll-Griffith, D. A., and Baillie, T. A. (2004) Drug-protein adducts: An industry perspective on minimizing the potential for drug bioactivation in drug discovery and development. Chem. Res. Toxicol. 17, 3–16. (7) Kalgutkar, A. S., Gardner, I., Obach, R. S., Shaffer, C. L., Callegari, E., Henne, K. R., Mutlib, A. E., Dalvie, D. K., Lee, J. S., Nakai, Y.,

340

(8) (9) (10) (11)

(12)

(13)

(14)

(15)

(16)

(17)

(18) (19)

(20)

Chem. Res. Toxicol., Vol. 22, No. 2, 2009 O’Donnell, J. P., Boer, J., and Harriman, S. P. (2005) A comprehensive listing of bioactivation pathways of organic functional groups. Curr. Drug Metab. 6, 161–225. Erve, J. C. (2006) Chemical toxicology: Reactive intermediates and their role in pharmacology and toxicology. Expert Opin. Drug Metab. Toxicol. 2, 923–946. Guengerich, F. P., and MacDonald, J. S. (2007) Applying mechanisms of chemical toxicity to predict drug safety. Chem. Res. Toxicol. 20, 344–369. Kalgutkar, A. S., and Soglia, J. R. (2005) Minimising the potential for metabolic activation in drug discovery. Expert Opin. Drug Metab. Toxicol. 1, 91–142. Kumar, S., Kassahun, K., Tschirret-Guth, R. A., Mitra, K., and Baillie, T. A. (2008) Minimizing metabolic activation during pharmaceutical lead optimization: Progress, knowledge gaps and future directions. Curr. Opin. Drug DiscoVery DeV. 11, 43–52. Mitchell, J. R., Jollow, D. J., Potter, W. Z., Davis, D. C., Gillette, J. R., and Brodie, B. B. (1973) Acetaminophen-induced hepatic necrosis. I. Role of drug metabolism. J. Pharmacol. Exp. Ther. 187, 185–194. Nelson, S. D., Mitchell, J. R., Timbrell, J. A., Snodgrass, W. R., and Corcoran, G. B. (1976) Isoniazid and iproniazid: Activation of metabolites to toxic intermediates in man and rat. Science 193, 901– 903. Masubuchi, N., Makino, C., and Murayama, N. (2007) Prediction of in vivo potential for metabolic activation of drugs into chemically reactive intermediate: Correlation of in vitro and in vivo generation of reactive intermediates and in vitro glutathione conjugate formation in rats and humans. Chem. Res. Toxicol. 20, 455–464. Obach, R. S., Kalgutkar, A. S., Soglia, J. R., and Zhao, S. X. (1814) (2008) Can in vitro metabolism-dependent covalent binding data in liver microsomes distinguish hepatotoxic from nonhepatotoxic drugs? An analysis of 18 drugs with consideration of intrinsic clearance and daily dose. Chem. Res. Toxicol. 21, 1822. El Mouelhi, M., Ruelius, H. W., Fenselau, C., and Dulik, D. M. (1987) Species-dependent enantioselective glucuronidation of three 2-arylpropionic acids. Naproxen, ibuprofen, and benoxaprofen. Drug Metab. Dispos. 15, 767–772. Qiu, Y., Burlingame, A. L., and Benet, L. Z. (1998) Mechanisms for covalent binding of benoxaprofen glucuronide to human serum albumin. Studies By tandem mass spectrometry. Drug Metab. Dispos. 26, 246–256. Dong, J. Q., Liu, J., and Smith, P. C. (2005) Role of benoxaprofen and flunoxaprofen acyl glucuronides in covalent binding to rat plasma and liver proteins in vivo. Biochem. Pharmacol. 70, 937–948. Leone, A. M., Kao, L. M., McMillan, M. K., Nye, A. Y., Parker, J. B., Kelley, M. F., Suzuki, E., Parkinson, A., Lord, P. G., and Johnson, M. D. (2007) Evaluation of felbamate and other antiepileptic drug toxicity potential based on hepatic protein covalent binding and gene expression. Chem. Res. Toxicol. 20, 600–608. Permeated, M., Kittteringham, N. R., Guenthner, T. M., Breckenridge, A. M., and Park, B. K. (1992) An investigation of the formation of

Bauman et al.

(21)

(22)

(23)

(24) (25)

(26)

(27)

(28)

(29)

(30)

(31) (32)

cytotoxic, protein-reactive and stable metabolites from carbamazepine in vitro. Biochem. Pharmacol. 43, 1675–1682. Bu, H. Z., Zhao, P., Dalvie, D. K., and Pool, W. F. (2007) Identification of primary and sequential bioactivation pathways of carbamazepine in human liver microsomes using liquid chromatography/tandem mass spectrometry. Rapid Commun. Mass Spectrom. 21, 3317–3322. Tang, W., Stearns, R. A., Wang, R. W., Chiu, S. H., and Baillie, T. A. (1999) Roles of human hepatic cytochrome P450s 2C9 and 3A4 in the metabolic activation of diclofenac. Chem. Res. Toxicol. 12, 192– 199. Ju, C., and Uetrecht, J. P. (1998) Oxidation of a metabolite of indomethacin (desmethyldeschlorobenzoylindomethacin) to reactive intermediates by activated neutrophils, hypochlorous acid, and the myeloperoxidase system. Drug Metab. Dispos. 26, 676–680. Tang, W. (2003) The metabolism of diclofenacsEnzymology and toxicology perspectives. Curr. Drug Metab. 4, 319–329. Wang, J., Davis, M., Li, F., Azam, F., Scatina, J., and Talaat, R. (2004) A novel approach for predicting acyl glucuronide reactivity via Schiff base formation: Development of rapidly formed peptide adducts for LC/MS/MS measurements. Chem. Res. Toxicol. 17, 1206–1216. Zhao, S. X., Dalvie, D. K., Kelly, J. M., Soglia, J. R., Frederick, K. S., Smith, E. B., Obach, R. S., and Kalgutkar, A. S. (2007) NADPHdependent covalent binding of [3H]paroxetine to human liver microsomes and S-9 fractions: Identification of an electrophilic quinone metabolite of paroxetine. Chem. Res. Toxicol. 20, 1649–1657. Obach, R. S., Kalgutkar, A. S., Ryder, T. F., and Walker, G. S. (2008) In vitro metabolism and covalent binding of enol-carboxamide derivatives and anti-inflammatory agents sudoxicam and meloxicam: Insights into the hepatotoxicity of sudoxicam. Chem. Res. Toxicol. 21, 1890–1899. Narimatsu, S., Arai, T., Watanabe, T., Masubuchi, Y., Horie, T., Suzuki, T., Ishikawa, T., Tsutsui, M., Kumagai, Y., and Cho, A. K. (1997) Covalent binding of a reactive metabolite derived from propranolol and its active metabolite 4-hydroxypropranolol to hepatic microsomal proteins of the rat. Chem. Res. Toxicol. 10, 289–295. Chen, Q., Ngui, J. S., Doss, G. A., Wang, R. W., Cai, X., DiNinno, F. P., Blizzard, T. A., Hammond, M. L., Stearns, R. A., Evans, D. C., Baillie, T. A., and Tang, W. (2002) Cytochrome P450 3A4-mediated bioactivation of raloxifene: Irreversible enzyme inhibition and thiol adduct formation. Chem. Res. Toxicol. 15, 907–914. Takakusa, H., Masumoto, H., Yukinaga, H., Makino, C., Nakayama, S., Okazaki, O., and Sudo, K. (2008) Covalent binding and tissue distribution/retention assessment of drugs associated with idiosyncratic drug toxicity. Drug Metab. Dispos. 36, 1770–1779. Chang, T, Okerholm, R. A., and Glazko, A. J. (1974) Identification of diphenhydramine (Benadryl) metabolites in human subjects. Res. Commun. Clin. Pathol. Pharmacol. 9, 391–404. Chatfield, D. H., and Green, J. N. (1978) Disposition and metabolism of benoxaprofen in laboratory animals and man. Xenobiotica 8, 133–144.

TX800407W

Can In Vitro Metabolism-Dependent Covalent Binding Data

Recommend Documents