ToxSci Advance Access originally published online on January 12, 2006
Toxicological Sciences 2006 90(2):451-459; doi:10.1093/toxsci/kfj095
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Inhibition of Hepatobiliary Transport as a Predictive Method for Clinical Hepatotoxicity of Nefazodone
* Departments of Safety Science and Pharmacokinetics, Dynamics and Metabolism, Pfizer Global Research and Development, Ann Arbor, Michigan 48105 and Groton, Connecticut 06340; and the Department of Pathology, University of Pittsburgh Medical Center, Pittsburgh, Pennsylvania 15261
1 To whom correspondence should be addressed at Department of Safety Science, Pfizer, Ann Arbor, MI 48108. Fax: (734) 622-3478. E-mail: vsevolod.kostrubsky{at}pfizer.com.
Received September 29, 2005; accepted January 9, 2006
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Treatment with the antidepressant nefazodone has been associated with clinical idiosyncratic hepatotoxicty. Using membranes expressing human bile salt export pump (BSEP), human sandwich hepatocytes, and intact rats, we compared nefazodone and its marketed analogs, buspirone and trazodone. We found that nefazodone caused a strong inhibition of BSEP (IC50 = 9 µM), inhibition of taurocholate efflux in human hepatocytes (IC50 = 14 µM), and a transient increase in rat serum bile acids 1 h after oral drug administration. Buspirone or trazodone had no effect on biliary transport system. Nefazodone produced time- and concentration-dependent toxicity in human hepatocytes with IC50 = 18 µM and 30 µM measured by inhibition of protein synthesis after 6 h and 24 h incubation, respectively. Toxicity was correlated with the amount of unmetabolized nefazodone. Partial recovery in toxicity by 24 h has been associated with metabolism of nefazodone to sulfate and glucuronide conjugates. The saturation of nefazodone metabolism resulted in sustained decrease in protein synthesis and cell death at 50 µM. The toxicity was not observed with buspirone or trazodone. Addition of 1-aminobenzotriazole (ABT), an inhibitor of CYP450, resulted in enhancement of nefazodone toxicity at 10 µM and was associated with accumulation of unmetabolized nefazodone. In human liver microsomes, ABT also prevented metabolism of nefazodone and formation of glutathione conjugates. We suggest that inhibition of bile acid transport by nefazodone is an indicator of potential hepatotoxicity. Our findings are consistent with the clinical experience and suggest that described methodology can be applied in the selection of nonhepatotoxic drug candidates.
Key Words: BSEP; buspirone; trazodone; hepatocytes; metabolism.
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
With standard rodent and nonrodent test species, pharmaceutical industry scientists can only predict about 50% of clinical incidences of hepatotoxicity (Olson et al., 2000
). Hepatotoxicity includes unexpected increases in liver enzymes after drug enters clinical trials or idiosyncratic toxicity usually recognized after several weeks or months of treatment. The purpose of this work is to investigate alternative in vitro models which may predict clinical hepatotoxicity for drugs that are expected to have a significant biliary clearance component.
Nefazodone is an antidepressant introduced in 1994. In June of 2004, nefazodone (Serzone) was withdrawn from the U.S. market following its previous withdrawal from European and Canadian markets. Analysis of data on adverse drug reaction from the World Health Organization found a statistically significant number of reports (91) for nefazodone-mediated hepatic injury (Spigset et al., 2003). No significant hepatotoxicity was found for other serotonin reuptake inhibitors. Estimated incidence of hepatotoxicity based on reports collected by the Spanish Pharmacovigilance System was 29 cases per 100,000 patient years (Garcia-Pando et al., 2002). Finally, a black-box warning on the nefazodone label estimated the rate of liver failure as 1 case per 250,000–300,000 patients per year, a rate of about 3–4 times the background rate of liver failure. This was considered as an underestimated number because of underreporting (NDA, 2001). Most cases of liver failure were reported between 1 to 8 months after drug initiation (Aranda-Michel et al., 1999; Conway et al., 2004; Garcia-Pando et al., 2002; Lucena et al., 1999; Tzimas et al., 2003) The common clinical symptoms of hepatotoxicity included jaundice and increases in ALT (10x), AST (10x), total bilirubin (mostly conjugated), and prothrombin time. Dark urine and clay-colored stool also were reported. Available histological evaluations of livers demonstrated centrilobular necrosis, bile-duct proliferation with cholestasis (Aranda-Michel et al., 1999; Lucena et al., 1999). About 60% of radiolabeled nefazodone administered intravenously to dogs was recovered in feces, suggesting that bile represents a significant route of drug elimination (Shukla et al., 1993). In humans, after administration of a single 200-mg dose, the urinary excretion of radioactivity in the 24-h samples was 49%, suggesting that bile was the other major route of nefazodone clearance (Barbhaiya et al., 1996). Therapeutic doses of nefazodone associated with liver failure were in the range of 200–400 mg/day. The total drug plasma concentration was at or above 1 µg/ml (Barbhaiya et al., 1995; Dockens et al., 1996). In summary, clinical symptoms of hepatotoxicity, significant biliary drug clearance, high administered doses, and high plasma concentration of nefazodone were similar to those described previously for troglitazone, another drug withdrawn from the market where the role of biliary transport in hepatotoxicity was suggested (Funk et al., 2001; Kostrubsky et al., 2000; 2001). We hypothesized that the potential for clinical hepatoxicity of nefazodone and its analogs could be evaluated by the potency of inhibition of bile acid transport and associated in vitro toxicity. An additional biomarker of hepatic effect thought to be linked to compromised biliary elimination was reported to be a transient increase in serum bile acids in the absence of frank liver toxicity in rodents. We compared results from nefazodone and its marketed structural analogs, buspirone and trazodone (Fig. 1), in specialized in vitro and in vivo assays. Based on the results of these experiments we propose that: (1) clinical hepatoxicity of nefazodone is linked to the ability of drug to inhibit bile acid transport, and (2) hepatoxicity associated with nefazodone is based on the particular chemical structure, so that drugs with the same mechanism of action but slightly different structure will not be hepatotoxic. Therefore, when multiple compounds with a similar mechanism of action are available during pharmaceutical development, one can eliminate from the drug discovery process compounds with a higher potential for clinical hepatotoxicity.
|
|
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Chemicals.
HMM (modified William's E) culture medium and supplements were from BioWhittaker (Walkersville, MD). Bile salt export pump (BSEP) recombinant transporter was expressed using baculovirus-infected Sf-9 insect cells with a final concentration of 5 mg/ml and purchased from Solvo Biotechnology (Hungary). Nefazodone hydrochloride, buspirone hydrochloride, trazodone hydrochloride, and 1-aminobenzotriazole were purchased from Sigma-Aldrich (St. Louis, MO). 3H-Taurocholic acid (2 Ci/mmol) was obtained from NEN Life Science Products (Boston, MA). HPLC-grade water, methanol, and acetonitrile were purchased from Mallinckrodt Chemicals (Phillipsburg, NJ). Human liver microsomes were purchased from BD Gentest (Woburn, MA). All general solvents and reagents were of the highest grade commercially available.
Hepatocyte cultures and treatment protocol.
Human hepatocytes were isolated from livers not used for whole-organ transplant within 24 h of procurement. Hepatocytes were isolated by three-step collagenase perfusion as described previously (Strom et al., 1996). The viability of cells obtained, as measured by trypan blue exclusion test, ranged from 74 to 90%. Hepatocytes were plated in Williams E medium supplemented with 10–7 M dexamethasone, 10–7 M insulin, 100 Units/ml of penicillin G, 100 µg/ml of streptomycin, and 10% bovine calf serum. Hepatocytes (2 x 106/well) were plated on six-well culture plates previously coated with type I (rat-tail) collagen. Cells were allowed to attach for 4–6 h at 37°C, at which time the medium was replaced with serum-free medium with the supplements listed above and changed every 24 h thereafter. Following 24 h in culture, medium was removed, and the cells overlaid with a neutralized preparation of collagen at a final concentration of 1.5 mg/ml for transport studies, as described previously for cultured sandwich hepatocytes (Kostrubsky et al., 2003). After 72 h in culture, cells were treated with nefazodone, buspirone, or trazodone at concentrations indicated in the figure legend for 1 h, 6 h, or 24 h, and toxicity was determined by the measurement of total protein synthesis by pulse-labeling hepatocytes for 1 h with 14C-leucine, as described previously (Kostrubsky et al., 1997, 2000). This culture condition maintains a sufficient level of functional Phase I and Phase II drug, metabolizing activities in hepatocytes after 96 h in culture, as we demonstrated for several substrates including testosterone 6ß-hydroxylation (CYP3A), ethoxyresorufin- and methoxyresorufin O-dealkylation (CYP1A), glucuronidation and sulfation of acetaminophen, 4-methylumbelliferon, and troglitazone (Kostrubsky et al., 1999, 2000, 2005, Wen et al., 2002). At the same time, incubation for 96 h allows minimizing the difference in basal enzyme activities in control cells prepared from different donors (Kostrubsky et al., 1999).
Transport Assays
Cultured sandwich hepatocytes with developed canaliculi.
Inhibition of bile transport in cultured sandwich hepatocytes with developed canaliculi was conducted according to previously described protocol (Kostrubsky et al., 2003). Briefly, 1 µM 3H-taurocholic acid, with or without increasing concentrations of test compound, was added to hepatocytes for 15 min at 37°C. After stopping transport by removing buffer and washing cells, taurocholate egress from canalicular spaces was initiated by adding standard or Ca/Mg2+-free buffer for 10 min. Aliquots of media were harvested and counted in a liquid scintillation counter. The difference in amount of radioactivity between two buffer conditions in the absence of tested compound was defined as a 100% taurocholic acid efflux in canaliculi. In the presence of compound, this difference became smaller and was used to calculate the percent inhibition of bile acid efflux (Kostrubsky et al., 2003). All values were normalized per amount of total cellular protein.
Bile salt export pump.
Inhibition of bile acid transport via BSEP was studied in membrane vesicles prepared from recombinant baculovirus infected Sf9 cells expressing human BSEP. Membrane preparation had inside-out orientation vesicles transporting substrates via BSEP, in the presence of ATP, into the vesicles. Assay was performed according to manufacture's protocol (SOLVO Biotechnology) with some modifications using glyburide as a positive control. Specifically, test compounds were dissolved in dimethylsulfoxide as a 1000x stock. Vesicles (50 µg) were added to the assay mix containing 20% 50 mM HEPES-Tris (pH 7.4), 10% 1 M KNO3, 10% 0.1 M Mg(NO3)2 and 2 µM of 3H-taurocholate in the presence or absence of tested drugs and preincubated for 5 min at 37°C. Transport was initiated with 4 mM of ATP or AMP (background subtraction), and the mixtures were incubated for another 5 min at 37°C. Reactions were stopped with ice-cold wash buffer (10% 100 mM Tris–HCl (pH 7.4) and 10% 1 M KNO3). Membrane suspension was then filtered and washed twice with wash buffer to remove substrate on the outside of the vesicles. After rapid filtration to separate the vesicles from the incubation solution, filters were counted in scintillation counter. The results were expressed as percent inhibition of taurocholate transport in the presence of inhibitors relative to untreated control.
MRP2-mediated efflux of calcein AM.
Inhibition of multidrug resistant associated protein-2 (MRP2)-mediated efflux of calcein AM was studied in MDCK cells transfected with MRP2 obtained from Prof. Piet Borst, Netherlands Cancer Institute. Wild type MDCK cells and MDCK cells stably transfected with human MRP2 were grown and maintained in Minimum Essential Medium (MEM) supplemented with 10% fetal bovine serum (FBS). Cells were incubated at 37°C in 5% CO2 for 24 h before the study. Each cell line (50,000 cells per well) was plated into Costar 3904 black 96-well plates (Packard Instrument Co., Meridian, CT) with 100 ml medium supplemented with 1% FBS and allowed to become confluent overnight. Test compounds were added to monolayers in 10 µl of culture medium containing 1% DMSO as solvent. Plates were incubated at 37°C for 30 min. Calcein AM was added in 10 µl of Hank's Balanced Salt Solution (HBSS) to yield a final concentration of 0.1 µM, and plates were incubated for another 60 min. Cells were then washed three times with ice-cold phosphate buffered saline (PBS), and fluorescence was measured using a Victor fluorometer (Perkin Elmer, Downers Grove, IL) at excitation and emission wavelengths of 485 nm and 535 nm, respectively.
Inhibition of bile acid transport in intact rats.
In vivo experiments to measure the response of serum bile acids to drug administration were conducted in Sprague-Dawley rats weighing approximately 300 g. Animals received powdered rodent chow (Purina certified chow; Purina, St. Louis, MO) and tap water by bottle ad libitum. All procedures involving animals were conducted in accordance with Guide for the Care and Use of Laboratory Animals. Three rats were assigned to either nefazodone or buspirone experimental groups. A single oral dose of 75 mg/kg nefazodone or buspirone was given by gavage as a solution in sterile water at a dose volume of 10 ml/kg to nonfasted animals. Prior to each blood collection, animals were anesthetized with isoflurane. Blood was collected for evaluation of pretest serum bile acids 1 day before drug administration. Blood samples were collected from the jugular vein 1 h and 24 h after dosing, and serum was analyzed for the total bile acids using a Hitachi 911 analyzer.
Liquid chromatography tandem mass spectrometry (LC/MS/MS) analysis of nefazodone metabolites in media from cultured hepatocytes.
Aliquots (200 µl) of media were spiked with the internal standard (buspirone) and analyzed by LC/MS/MS without any further sample preparation. An aliquot (20 µl) from each sample was injected directly onto an HPLC column (Discovery C18, 150 x 2.1 mm, Supelco) coupled to an API 4000 mass spectrometer (Applied Biosystems/MDS SCIEX, Ontario, Canada), equipped with a TurboIonSpray® source held at 275°C. The electrospray needle was maintained at 4600 V with the declustering and exit potentials set at 46 and 10 V, respectively. Ultrapure nitrogen was used as the nebulizer and curtain gas. The mass spectrometer was operated in full scan (100–1000 amu) or multiple reaction monitoring (MRM) modes for qualitative and quantitative analyses, respectively. MRM analysis was carried out using nitrogen as the collision gas. The collision energy was kept at 25 eV. Other parameter settings for the MRM analyses included arbitrary values of 6, 25, 40, and 40 for collision (CAD), curtain, nebulizer, and turbo gases, respectively. The mass transitions for the metabolites and the internal standard were 386 
122 (buspirone, internal standard), 566 486 (sulfate conjugate), and 662 486 (glucuronide conjugate). The peak areas from each of these transitions were obtained, and ratios of the analytes to the internal standard were obtained for each sample. The metabolites of nefazodone were separated on the HPLC column using a gradient solvent system consisting of acetonitrile and 0.1% formic acid with the flow rate set at 0.3 ml/min. The initial conditions consisted of a mixture of acetonitrile and 0.1% formic acid (4:96 v/v) and were maintained for 3 min after the sample was injected. The percentage of acetonitrile was increased linearly from 3% to 90% in the next 17 min. After an additional 5 min at 90% acetonitrile, the column was reequilibrated with the initial mobile phase for 10 min before the next injection.
Metabolism of nefazodone by human liver microsomes.
The metabolism of nefazodone was investigated using human liver microsomes. Nefazodone (25 µM) was incubated in the presence of human liver microsomes (1 mg/ml), MgCl2 (3 mM), and NADPH (2 mM) in 0.1 M phosphate buffer (pH 7.4) at 37°C for 30 min. Where indicated, glutathione (2 mM) and 1-aminobenzotriazole (ABT, 1 mM) were added to the incubation mixtures. The samples were preincubated for 15 min prior to initiating the reaction by adding substrate. Incubations were also performed without the addition of NADPH as a control for possible nonenzymatic degradation of nefazodone. At the end of the incubation, the microsomal proteins were precipitated by adding 2 ml of ice-cold acetonitrile, followed by centrifugation at 3,000 rpm using an AllegraTM X-22R centrifuge equipped with SX4250 rotor. The supernatant was removed, dried under a stream of nitrogen, and reconstituted in 400 µl of ACN:0.1% formic acid (4:96 v/v). An aliquot (40 µl) was injected onto LC/MS using the conditions described earlier.
Statistical analysis and IC50 determination.
Results were analyzed by a two-factor ANOVA. A p < 0.05 was interpreted as the level of statistical significance.
The IC50 estimates were determined by nonlinear curve fitting using the WinNonLin software (Pharsight, Mountain View, CA) and were defined as the concentration of compound required to inhibit taurocholate transport or cause inhibition of protein synthesis by 50%.
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Effect of Nefazodone, Buspirone, and Trazodone on Bile Acid Efflux in Sandwich Human Hepatocytes, Membrane Vesicles, and Intact Rats
Incubation of cultured human hepatocytes (prepared from two different donors) with the nefazodone resulted in concentration-dependent inhibition of taurocholic acid efflux in canaliculi with IC50 = 14 µM (Fig. 2). In contrast, neither buspirone nor trazodone inhibited canalicular transport. Hepatocytes were fully capable of metabolizing nefazodone (see Figs. 8 and 9), and hence the inhibitory effect on bile transport may have resulted from either parent compound or metabolites. In addition, inhibition of transporter activity may be a consequence of toxicity. Therefore, we investigated the effect of drugs on BSEP-mediated transport activity in membrane vesicles. Incubation of nefazodone with BSEP-expressed vesicles resulted in a concentration-dependent inhibition of taurocholic acid transport with IC50 = 9 µM (Fig. 3), a result which was similar to that observed with human hepatocytes (Fig. 2), suggesting that the parent nefazodone directly inhibited BSEP-mediated bile acid transport. In the same experiment, glyburide was used as a positive control, which displayed an IC50 = 2 µM. We have previously shown the potent inhibitory effect of glyburide on taurocholate efflux in human hepatocytes, where glyburide at 10 µM inhibited taurocholate efflux by 70% (Kostrubsky et al., 2003
). In contrast, buspirone and trazodone produced no apparent inhibition of BSEP-mediated bile acid transport, when tested at concentrations as high as 100 µM (Fig. 3). Since increases in conjugated bilirubin were observed clinically, we tested whether nefazodone inhibited multidrug resistant associated protein-2 (MRP2), which is responsible for the canalicular excretion of conjugates including bilirubin, glutathione, bile salts, leukotrienes, and others. Nefazodone inhibited MRP2 with approximate IC50 = 113 µM. Buspirone and trazodone showed no inhibition up to the top concentration of 200 µM. Previously, we and others have shown a transient increase in rat serum bile acids following iv administration of clinically hepatotoxic drugs (troglitazone, bosentan, CI-1034) that also inhibited bile efflux transport (Fattinger et al., 2001; Funk et al., 2001; Kostrubsky et al., 2003). Often, rats are resistant to liver toxicity, which is characterized by traditional markers including an increase in serum hepatic enzymes and histological changes. Assuming that a drug has a high affinity for transporter-mediated biliary elimination, serum bile acids could be increased due to inhibition of hepatic bile acid efflux into canaliculi by the drug. This may occur when the drug reaches the maximum plasma concentration, which might parallel the peak liver concentration. This increase could be transient and diminish with drug clearance. However, it provides a potential biomarker for predicting hepatic effects in humans as well. In contrast to previous studies that monitored serum bile acids after iv administration of test drugs, in the present work, nefazodone and buspirone were administered orally to mimic the standard preclinical toxicity study. Blood was collected at the approximate Tmax time for both nefazodone and buspirone (Barbhaiya et al., 1995; Dockens et al., 1996; Salazar et al., 2001) and 24 h after dosing. As shown in Figure 4A, nefazodone increased serum bile acids by 60% after 1 h, and bile acids returned to control values by 24 h. No such response was detected for buspirone (Fig. 4B). Since there was a large animal-to-animal variation, as well as the difference in the pretest values, we repeated this experiment two more times. In the nefazodone group, the overall change in the pattern of serum bile acids characterized by an increase after 1 h followed by a decrease by 24 h was observed in all experiments. Specifically, the mean increases in serum bile acids after 1 h were 60%, 400%, and 40% in three separate experiments. This pattern was not found in the buspirone group. Insufficient amounts of trazodone prevented us from testing it along with nefazodone and buspirone. In these normal rats, we did not anticipate increase in serum conjugated bilirubin, since rats appear to have high transport clearance rate of MRP2 substrates compared to other species (Ishizuka et al., 1999) and do not always respond to treatment with cholestatic compounds. However, it is likely that MRP2-defficient (TR–) rats will experience a transient increase in bilirubin in response to nefazodone, similar to the increase we observed with troglitazone when it was administered by gavage as a single dose of 200 mg/kg (Kostrubsky et al., 2001).
|
|
|
|
|
Thus, all three systems, sandwich human hepatocytes, expressed BSEP, and rat serum bile acids, demonstrated that nefazodone inhibits bile acid transport.
Effect of Nefazodone, Buspirone, and Trazodone on Toxicity in Human Hepatocytes
Next, we compared the effect of three drugs on cellular toxicity in cultured human hepatocytes by measuring total protein synthesis. This sensitive method allowed the detection of toxicity as early as 1 h after drug treatment, preceding enzyme release and morphological changes as we reported previously (Kostrubsky et al., 2000). Hepatocytes were treated with nefazodone, buspirone, or trazodone at 10, 50, and 100 µM for 6 h, and total protein synthesis was determined (Fig. 5). Nefazodone produced a concentration-dependent decrease in protein synthesis with an IC50 = 18 µM. Concentrations equal to or exceeding 50 µM resulted in morphological cell death observed microscopically. Buspirone and trazodone did not decrease protein synthesis at the concentration range tested. We hypothesized that inhibition of protein synthesis may recover after a longer incubation with nefazodone, due to metabolism of the parent drug. As shown in Fig. 6, nefazodone at 10 µM produced 45%, 30%, and 5% decrease in protein synthesis after 1, 6, and 24 h incubation, respectively, indicating hepatocytes recovery after 24 h. In contrast, nefazodone at 50 and 100 µM produced sustained inhibition of protein synthesis, resulting in cell death.
|
|
Metabolism of Nefazodone in Human Hepatocytes
To confirm the effect of parent drug on toxicity, we analyzed culture media from cells treated with 10 µM nefazodone for 6 and 24 h. There was a 90% decrease in the concentration of parent nefazodone by 24 h compared to 6 h incubations (Fig. 7). In contrast, there were no significant differences for 50 and 100 µM nefazodone between 6 and 24 h, suggesting that cellular metabolism was compromised, resulting in the lack of nefazodone clearance at these toxic concentrations. Analysis of metabolites in the media indicated that nefazodone underwent oxidative metabolism followed by conjugation with sulfate and glucuronide, with only trace amounts of unconjugated hydroxylated metabolites detected in the culture media (Fig. 8). Collision-induced dissociation of the molecular ions at MH+ = 566 and MH+ = 662, which corresponded to the sulfate and glucuronide conjugates of mono-hydroxynefazodone, indicated that conjugation had occurred on the hydroxyl group on the 3-chlorophenyl ring in nefazodone. Incubation for 24 h with 10, 25, 50, and 100 µM nefazodone demonstrated the saturation of metabolism, as evident from the diminished capacity of the hepatocytes to produce conjugated metabolites at 25 µM. Furthermore, there was a decrease in the amount of conjugates produced at 50 and 100 µM, most likely due to cell toxicity resulting in compromised drug metabolizing enzyme activities (Figs. 5 and 8). Thus, metabolism of nefazodone to conjugated products at 10 µM resulted in recovery of protein synthesis. In contrast, insufficient metabolism at
25 µM resulted in a sustained decrease in protein synthesis (40% at 25 µM) and cell death at 50 and 100 µM by 24 h.
|
To further characterize the importance of metabolism in preventing nefazodone toxicity, we treated hepatocytes with the combination of 10 µM nefazodone and 1 mM ABT, a nonspecific inhibitor of P450 enzymes that has been shown previously to inhibit P450 activity in primary hepatocytes (Harrel et al., 1994
). Treatment with ABT and nefazodone (10 µM) for 24 h resulted in a 45% decrease in total protein synthesis (Fig. 9A). In contrast, there was almost full recovery in hepatocytes treated with 10 µM nefazodone alone (Fig. 9A). Analysis of culture media from cells treated with ABT showed an increase in the concentration of unmetabolized nefazodone associated with a decrease in formation of sulfate and glucuronide conjugates (Fig. 9B). Thus, inhibition of nefazodone metabolism enhanced toxicity and was associated with an accumulation of parent drug.
Metabolism of Nefazodone in Human Liver Microsomes
Incubation of human liver microsomes with nefazodone and ABT resulted in inhibition of nefazodone metabolism by 90% with a corresponding decrease in the formation of hydroxylated metabolites, thus confirming the role of ABT as a metabolic inhibitor observed in human hepatocytes. Since bioactivation of drugs to reactive metabolites is considered as one of the major risk factors for developing toxicity, and bioactivation can be modeled in vitro by the formation of drug–glutathione products, we also investigated whether nefazodone was capable of being metabolized to glutathione conjugates. Previously, bioactivation of nefazodone was demonstrated in human liver microsomes and recombinant CYP3A (Kalgutkar et al., 2005). Incubation of nefazodone (25 µM) in liver microsomes with glutathione resulted in the formation of a glutathione adduct of a hydroxylated nefazodone. This reaction was completely abolished by ABT. As we demonstrated with human hepatocytes, treatment with ABT and nefazodone potentiated toxicity due to accumulation of unmetabolized nefazodone, suggesting that metabolic bioactivation is not the mechanism responsible for nefazodone toxicity.
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
The critical link between biliary elimination and hepatotoxicity has been demonstrated in the bile duct–ligated rodent models, where the toxicity of xenobiotics was significantly enhanced in animals with obstructed bile flow (Klaassen, 1973
). In this study, our hypothesis was based on the major principle of toxicology stating that "the right dose differentiates a poison from a remedy" as was postulated by Paracelsus in 1541 AD. The initiating event for nefazodone hepatotoxicity may be an accumulation of drug and bile in the liver of certain individuals. Previously, we have shown that drugs that have strong inhibitory effects on bile acid transport also demonstrate clinical hepatotoxicity which is frequently not recognized in animal studies (Kostrubsky et al., 2003). Now, with the greater number of compounds we analyzed, it is becoming clear that a combination of three factors are associated with clinical hepatotoxicity: high-molecular-weight compounds with expected biliary excretion of parent and metabolites, affinity for biliary transporters as shown by a concentration-dependent inhibition of bile acid transport in vitro, and projected human drug plasma concentration at or above 1 µg/ml. Nefazodone meets all these criteria. In contrast, buspirone and trazodone did not have an inhibitory effect on bile acid efflux, suggesting a structure-specific interaction with transporters. The inhibitory effect of nefazodone on bile transport was mediated via BSEP in expressed system. However, in cultured hepatocytes and in humans other transporters may also contribute. In membrane vesicles, human MRP2 was inhibited by parent nefazodone, but at concentrations that are not pharmacologically relevant. Since nefazodone is metabolized to sulfate and glucuronide conjugates, the effect on MRP2 via metabolites could be much stronger compared to that found in membrane vesicles. It should be pointed out that the purpose of the screen proposed in this work is not to identify the individual transporters but to demonstrate the functional assay which is able to discriminate the effect of drugs on bile acid efflux. By screening the ability of drugs to inhibit bile transport at concentrations that will saturate active transporters (1–100 µM), we can potentially identify hepatotoxic compounds including those that could cause an unexpected increase in liver transaminases or an idiosyncratic response. As we described here and as was previously reported for troglitazone (Funk et al., 2001; Kostrubsky et al., 2001) it appears that the common initiating event for these drugs is the inhibition of bile acid transport. Much work, however, needs yet to be done, particularly in conducting structure-toxicity studies to validate the predictive power of this methodology.
Concentrations of 2–4 µM (1–2 µg/ml) were reported as the maximum plasma concentrations in humans taking nefazodone at doses of 200–400 mg/day (Barbhaiya et al., 1995; Dockens et al., 1996). These concentrations could go disproportionately higher with increasing dose (Barbhaiya et al., 1995). There is no information on plasma:liver ratio of nefazodone. However, it is likely that level of total drug including metabolites in the liver could be higher compared to plasma. In BSEP-expressed membranes and human hepatocytes, we found inhibition of bile acid transport at IC50 = 9 and 14 µM, respectively, suggesting that nefazodone, if it accumulates in the liver, could inhibit bile acid transport and its own elimination, resulting in progressively increasing drug and bile acid concentrations in hepatocytes of certain individuals. When administered for only 2 days, nefazodone did not cause hepatic changes in most patients, even at high doses with reported plasma concentration up to 9.6 µg/ml (20.4 µM) (Barbhaiya et al., 1995), suggesting that unique patient-specific factors including hepatic drug and bile acid accumulation over a longer time, genetic, and individual tissue repair response may play a role in a progression of toxicity in idiosyncratic cases. We also found that inhibition of bile transport occurred at concentrations lower than those where sustained toxicity in human hepatocytes was observed (IC50 = 30 µM). In addition, parent nefazodone inhibited bile acid transport in BSEP-expressed membranes at IC50 = 9 µM. These data would suggest that inhibition of bile transport occurs prior to cytotoxicity.
We found that nefazodone was metabolized to sulfate and glucuronide conjugates of mono-hydroxylated nefazodone. Conjugates were previously identified in human plasma but have not been characterized (Mayol et al., 1994). Nefazodone conjugation was saturated at 25 µM and showed apparent decrease in metabolites produced at 50 and 100 µM (Fig. 8), concentrations that caused cell death. This suggests that the dose at which saturation of conjugation occurs represents a threshold of toxicity and perhaps could be monitored in humans by the level of conjugates.
Toxicity was associated with accumulation of unmetabolized nefazodone (Fig. 7). Inhibition of nefazodone metabolism by ABT also resulted in toxicity (Fig. 9), supporting the critical role of parent drug. Furthermore, incubation of human liver microsomes with nefazodone and glutathione produced an adduct, suggesting that nefazodone can be metabolically bioactivated to form reactive metabolites, as previously reported (Kalgutkar et al., 2005). The formation of this adduct was fully prevented by ABT. However, this adduct was not observed in cultured human hepatocytes. This is not surprising, because the precursor (hydroxylated nefazodone) to the reactive intermediates appears to be rapidly glucuronidated or sulfated, hence preventing its further bioactivation. In addition, there was no evidence of reactive metabolites (GSH-related products) at 10–100 µM in either cell lysates or culture media, suggesting that neither recovery in protein synthesis nor toxicity at 50 µM and 100 µM was due to the formation of reactive metabolites. The total absence of any glutathione-related adduct of nefazodone in human hepatocytes, combined with the results from ABT study showing decreased metabolism, high parent concentration and enhanced toxicity, suggests that unmetabolized nefazodone, and not the products of bioactivation, is responsible for hepatotoxicity.
Serum bile acids were increased at Tmax time in rats administered a single oral dose of nefazodone. Subsequently, the bile acids returned to control levels after 24 h. This transient increase in serum total bile acids can be used as another signal for a compound's potential to cause human hepatotoxicity.
In conclusion, we ranked three drugs according to their ability to inhibit bile acid transport and cause in vitro toxicity. Nefazodone was a potent inhibitor of bile transport. Nefazodone also caused toxicity in human hepatocytes associated with saturation of conjugation and accumulation of parent drug. We propose that the testing approach described in this work, for compounds with a substantial portion of biliary clearance, can be used to predict the propensity of a drug to cause human hepatotoxicity. Such an approach should also be beneficial in rank-ordering potential new drug candidates for further advancement in development.
|
ACKNOWLEDGMENTS |
|---|
This work was supported in part by the Liver Tissue Procurement and Distribution Program NIH/NIDDK #N01-DK-9–2310 (SCS).
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Aranda-Michel, J., Koehler, A., Bejarano, P. A., Poulos, J. E., Luxon, B. A., Khan, C. M., and Ee, L. C. (1999). Nefazodone-induced liver failure: Report of three cases. Ann. Intern. Med. 130, 285–288.
Barbhaiya, R., Dandekar, K., and Green, D. (1996). Pharmacokinetics, absolute bioavailability, and disposition of [14C]nefazodone in humans. Drug Metab. Dispos. 24, 91–95.[Abstract]
Barbhaiya, R., Marathe, P., Greene, D., Mayol, R., Shukla, U., Gammans, R., and Pittman, K. (1995). Safety, tolerance, and preliminary pharmacokinetics of nefazodone after administration of single and multiple oral doses to healthy adult male volunteers: A double-blind, Phase I study. J. Clin. Pharmacol. 35, 974–984.[Abstract]
Conway, C., McGuire, M. J., and Baram, V. (2004). Nefazodone-induced liver failure. Clin. Phychopharmacol. 24, 353–354.
Dockens, R., Greene, D., and Barbhaiya, R. (1996). Assessment of pharmacokinetic and pharmacodinamic drug interactions between nefazodone and digoxin in healphy male volunteers. J. Clin. Pharmacol. 36, 160–167.[Abstract]
Fattinger, K., Funk, C., Pantze, M., Weber, C., Reichen, J., Stieger, B., and Meier, P. (2001). The endothelin antagonist bosentan inhibits the canalicular bile salt export pump: A potential mechanism for hepatic adverse reactions. Clin. Pharmacol. Ther. 69, 223–231.[CrossRef][ISI][Medline]
Funk, C., Ponelle, C., Scheuermann, G., and Pantze, M. (2001). Cholestatic potential of troglitazone as a possible factor contributing to troglitazone-induced hepatotoxicity: In vivo and in vitro interaction at the canalicular bile salt export pumb (Bsep) in the rat. Mol. Pharmacol. 59, 627–635.
Garcia-Pando, A. C., Poso, J. G., Sanchez, A. S., Martin, A. V., Castro, A. M., and Lucena, I. (2002). Hepatotoxicity associated with new antidepressants. J. Clin. Psychiatry 63, 135–137.[Medline]
Harrel, A. W., Wheeler, S. M., East, P., Clarke, S. E., and Chenery, R. J. (1994). Use of rat and human in vitro systems to assess the effectiveness and enzymology of deoxyguanine analogs as prodrugs of an antiviral agent. Drug Metab. Dispos. 22, 189–193.[Abstract]
Ishizuka, I., Konno, K., Shiina, T., Naganuma, H., Nishimura, K., Ito, K., Suzuki, H., and Sugiyama, Y. (1999). Species differences in the transport activity for organic anions across the bile canalicular membrane. J. Pharmacol. Exp. Ther. 290, 1324–1330.
Kalgutkar, A. S., Vaz, A., Lame, M. E., Henne, K. R., Soglia, J., Zhao, S. X., Abramov, Y. A., Lombardo, F., Collin, C., Hendsch, Z. C., et al. (2005). Bioactivation of the nontricyclic antidepressant nefazodone to a reactive quinone-imine species in human liver microsomes and recombinant cytochrome P4503A. Drug Metab. Dispos. 33, 243–253.
Klaassen, C. D. (1973). Comparison of the toxicity of chemicals in newborn rats to bile duct-ligated and sham-operated rats and mice. Toxicol. Appl. Pharmacol. 24, 37–44.[CrossRef][Medline]
Kostrubsky, V. E., Lewis, L. D., Wood, S. G., Sinclair, P. R., Wrighton, S. A., and Sinclair, J. F. (1997). Effect of taxol on cytochrome P4503A and acetaminophen toxicity in cultured rat hepatocytes: Comparison to dexamethsone Toxicol. Appl. Pharmacol. 142, 79–86.[CrossRef][ISI][Medline]
Kostrubsky, V. E., Ramachandran, V., Venkataramanan, R., Dorrko, K., Esplen, J., Zhang, S., Sincalir, J. F., Wrighton, S., and Strom, S. C. (1999). The use of human hepatocytes cultures to study the induction of cytochrome P-450. Drug Metab. Dispos. 27, 887–894.
Kostrubsky, V. E., Sinclair, J. F., Ramachandran, V., Venkataramana, R., Wen, Y., Kindt, E., and Galchev., V. (2000). The role of conjugation on hepatotoxicity of troglitazone in human and porcine hepatocyte cultures. Drug Metab. Dispos. 28, 1192–1197.
Kostrubsky, V. E., Sinclair, J. F., Strom, S. C., Wood, S., Urda, E., Wen, Y. H., Kulkarni, S., and Mutlib, A. (2005). Phenobarbital and phenytoin induced acetaminophen hepatotoxicity due to inhibition of UDP-glucuronosyltransferases in cultured human hepatocytes. Toxicol. Sci. 87, 146–155.
Kostrubsky, V. E., Strom, S. C., Hanson, J., Urda, E., Rose, K., Burliegh, J., Zocharski, P., Sinclai, J. F., and Sahi, J. (2003). Evaluation of cholestatic potential of drugs by inhibition of bile acid transport in cultured primary human hepatocytes and intact rats. Predictability of drug-induced cholestatis. Toxicol. Sci. 76, 220–228.
Kostrubsky, V. E., Vore, M., Kindt, E., Burliegh, J., Rogers, K., Peter, G., Altrogge, D., and Sinz, M. (2001). The effect of troglitazone biliary excretion on metabolite distribution and cholestasis in transporter-deficient rats. Drug Metab. Dispos. 29, 1561–1566.
Lucena, M. I., Andrade, R. J., Gomez-outes, A., Rubio, M., and Cabello, M. R. (1999). Acute liver failure after treatment with nefazodone. Dig. Dis. Sci. 44, 2577–2579.[CrossRef][Medline]
Mayol, R. F., Cole, C. A., Luke, G. M., Colson, K. L., and Kerns, E. H. (1994). Characterization of the metabolites of the antidepressant drug nefazodone in human urine and plasma. Drug Metab. Dispos. 22, 304–311.[Abstract]
NDA (2001). Hepatotoxicity boxed warning. New Drug Application (NDA) 20-152/S-028, supplemental application S-128, 12/4/2001.
Olson, H., Betton, G., Robinson, D., Thomas, K., Monro, A., Kolja, G., Lilly, P., Sanders, J., Sipes, G., Bracken, W., et al. (2000). Concordance of the toxicity of pharmaceuticals in humans and in animals. Regul. Toxicol. 32, 56–57.
Salazar, D., Frackiewicz, E., Dockens, R., Kollia, G., Fulmor, E., Tigel, P., and Uderman, H. (2001). Pharmacokinetics and tolerability of buspirone during oral administration to children and adolescents with anxiety disorder and normal healthy adults. J. Clin. Pharmacol. 41, 1351–1358.[Abstract]
Shukla, U., Marathe, P., Pittman, K., and Barbhaiya, R. (1993). Pharmacokinetics, absolute bioavailability, and disposition of [14C]nefazodone in the dogs. Drug Metab. Dispos. 21, 502–507.[Abstract]
Spigset, O., Hagg, S., and Bate, A. (2003). Hepatic injury and pancreatitis during treatment with serotonin reuptake inhibitors: Data from the World Health Organization (WHO) database of adverse drug reactions. Clin. Phychoparmacol. 18, 157–161.
Strom, S. C., Dorko, K., Thompson, M. T., Schuetz, J. D., and Schutz, E. G. (1996). Use of human hepatocytes to study P450 gene induction. Methods Enzymol. 272, 388–401.[ISI][Medline]
Tzimas, G. N., Dion, B., and Deschenes, M. (2003). Early onset, nefazodone-induced fulminant hepatic failure. Am. J. Gastroenterol. 98, 1664–1664.[Medline]
Wen, Y. H., Sahi, J., Urda, E., Kulkarni, S., Rose, K., Zheng, X., Sinclair, J. F., Cai, H., Strom, S. C., and Kostrubsky,V. E. (2002). Effects of bergamottin on human and monkey drug-metabolizing enzymes in primary cultured hepatocytes. Drug Metab. Dispos. 30, 977–984.
CiteULike 
Connotea 
Del.icio.us What's this?
This article has been cited by other articles:
|
|
 |
|
  J. J. Xu, P. V. Henstock, M. C. Dunn, A. R. Smith, J. R. Chabot, and D. de Graaf Cellular Imaging Predictions of Clinical Drug-Induced Liver Injury Toxicol. Sci., September 1, 2008; 105(1): 97 - 105. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
J. A. Dykens, J. D. Jamieson, L. D. Marroquin, S. Nadanaciva, J. J. Xu, M. C. Dunn, A. R. Smith, and Y. Will In Vitro Assessment of Mitochondrial Dysfunction and Cytotoxicity of Nefazodone, Trazodone, and Buspirone Toxicol. Sci., June 1, 2008; 103(2): 335 - 345. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 K. P DeSanty and C. M Amabile Antidepressant-Induced Liver Injury Ann. Pharmacother., July 1, 2007; 41(7): 1201 - 1211. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||