Abstract

Fatty acids are an important source of energy. Excessive energy intake results in elevated levels of free fatty acids that are thought to be the pathogenic factors causing metabolic disorders such as dyslipidemia, obesity, insulin resistance, diabetes, and fatty liver. Underlying metabolic disorders have been suggested to be a predisposing factor for drug-induced liver injury. The steadily expanding population with metabolic disease may pose a higher risk for drug-induced toxicity. In order to understand the interaction of free fatty acids and drug-induced toxicity at the cellular level, we explored whether the saturated free fatty acid palmitate could modulate drug-induced cytotoxicity in HepG2 cells. A number of drugs known to induce hepatotoxicity in humans were selected to test this hypothesis. Drugs without reported hepatotoxicity were also tested to evaluate the specificity of the palmitate-induced effects. We demonstrate that palmitate, at sublethal concentrations, was able to potentiate the cytotoxicity and/or apoptosis induced by some but not all drugs tested. The palmitate and drug coincubation potentiated toxicity, which when combined with the plasma maximum concentration (Cmax), allowed us to identify idiosyncratic toxic drugs that were not flagged in previously deployed cytotoxicity assays. Our data suggest that treatment of cells with palmitate improves the sensitivity to detect compounds with risk of inducing idiosyncratic liver toxicity. Furthermore, this assay may be used to identify compounds that have higher safety risks in a population with metabolic syndrome.

Drug-induced toxicity has become known as a major cause for the termination of drug candidates and postmarket withdrawal of medications. Drug-induced liver injury (DILI) is the most common cause for safety related attrition, and accounts for more than 50% of the cases of acute liver failure in the United States (Lee, 2003). Most approved drugs cause liver injury infrequently in humans, and this type of toxicity is regarded as idiosyncratic, and thus patient dependent. Idiosyncratic hepato toxicity is one of the major reasons for drug withdrawal or assignment of “black box” warning due to the inability to predict this type of toxicity in animal models and even clinical trials in smaller populations. A number of factors, such as age, gender, drug metabolism/distribution, coexisting diseases, inflammation, nutritional status, drug coexposure, and environmental factors, have been hypothesized to contribute to the individual sensitivity to DILI (Roth and Ganey, 2010; Ulrich, 2007). Therefore, it is highly desirable to develop in vitro assays to flag such toxicity in the earlier phases of drug development, when much chemical diversity is available.

Idiosyncratic liver injury is thought of being a convergence of multiple mechanisms that lead to toxicity, including but not limited to generation of reactive metabolites, inhibition of mitochondrial functions, induction of oxidative stress, and inhibition of biliary efflux transporters (such as the bile salt export pump) and the multidrug resistance-associated protein 2 (Begriche et al., 2011; Morgan et al., 2010; Stepan et al., 2011). The pharmaceutical industry has recently put substantial resources into building assay platforms that can predict idiosyncratic liver toxicity. One such platform utilizes human hepatocytes and measures endpoints, such as glutathione content, mitochondrial membrane potential, and reactive oxygen species (ROS) generation, using an imaging approach (Xu et al., 2008). This platform was shown to have a 50% predictive value with a false-positive rate of less than 5%. However, this platform fails to detect about half of the toxic drugs. Assessment of reactive metabolite formation has become industry standard (Stepan et al., 2011). Impairment of mitochondrial function appears to be a frequent mechanism for DILI. Mitochondrial dysfunction may lead to oxidative stress, lipid accumulation (steatosis), apoptosis, and cell death. The aforementioned predisposing factors for idiosyncratic DILI could enhance drug-induced mitochondrial dysfunction (Begriche et al., 2011). A cell-based assay, based on the differential LC50 (concentration causing 50% lethality) for cells cultured in glucose versus galactose medium, has been developed that can detect drug toxicity primarily caused by mitochondrial impairment (Marroquin et al., 2007). Most of the currently deployed cell-based assays have not put risk factors, which make patients susceptible to idiosyncratic toxicity, into consideration. Only until recently, a cell model employing cytokine mix/lipopolysaccharide (LPS) has been evaluated to predict inflammation-associated idiosyncratic DILI. Cytokine mix/LPS synergize with a number of idiosyncratic drugs to induce cytotoxicity in HepG2 and primary hepatocytes, with the concentrations that cause cell death falling within the therapeutic exposure levels (Cosgrove et al., 2009).

Excessive energy intake leads to an epidemic of metabolic diseases, including obesity, insulin resistance, type II diabetes, and hyperlipidemia (Naser et al., 2006; Zimmet et al., 2001). A plethora of evidence suggests the association of mitochondrial dysfunction with metabolic syndromes (Leem and Koh, 2012). This underlying dysfunction could sensitize patients to drug-induced stress or toxicity. Some drugs, such as troglitazone, methotrexate, halothane, and cyclosporine A (CsA), have been reported to have increased incidences of drug-induced toxicity or worsen the preexisting diseases in patients with metabolic disease (Baum, 2001; Boelsterli, 2003; Wang et al., 2007). It is possible that metabolic syndrome may be a predisposing factor for some drug-induced toxicity, which may manifest itself in an idiosyncratic manner. The expanding population with metabolic disorders would potentially increase the incidence rate of drug-induced toxicity.

Free fatty acids are elevated in patients with obesity, diabetes, and dyslipidemia, and are considered as pathogenic factors for metabolic syndrome (Eitel et al., 2002; Feldstein et al., 2004; Malhi et al., 2006; Scaglione et al., 2010). Free fatty acids, especially saturated fatty acids such as palmitate, induce endoplasmic reticulum (ER) stress and apoptosis in various cells, including pancreatic β cells, hepatocytes, and cardiomyocytes (Eitel et al., 2002; Feldstein et al., 2004; Kong and Rabkin, 2000). Palmitate may cause stress in cells by increasing ROS, activating the stress-responsive c-Jun NH(2)-terminal kinase (JNK) pathway, modulating mitochondrial function, or stimulating proinflammatory cytokine secretion (Eitel et al., 2002; Feldstein et al., 2004; Kong and Rabkin, 2000; Malhi et al., 2006). Palmitate-treated cells have been used as models to evaluate lipotoxicity, insulin sensitivity, and β-cell function (Gao et al., 2010; Koshkin et al., 2008). We hypothesize that synergistic or additive toxicity may occur if a drug induces stress by mechanisms similar to those affected by palmitate. We have previously demonstrated that CsA and palmitate synergistically induced cytotoxicity in HepG2 cells, supporting the clinical observation that organ transplant patients with metabolic syndromes have poor outcome and higher incidence of adverse events (Grady et al., 1996). It appears that induction of oxidative stress and JNK activation play important roles in mediating the synergistic toxicity induced by palmitate/cyclosporine cotreatment (Luo et al., 2012).

Here, to test our hypothesis, we screened a large number of drugs in a cell model treated with palmitate. We used HepG2 cells as a model to examine the effects of palmitate on drug-induced cytotoxicity and apoptosis. Due to the pleiotropic effects of palmitate in cells, such as ER stress, mitochondrial function, and JNK activation etc., it is not known which mechanism might play an important role in potentiating drug-induced toxicity. Therefore we decided to test compounds with a variety of toxicity mechanisms instead of focusing on only particular mechanism. A number of drugs that are known to induce idiosyncratic liver toxicities, including mitochondrial toxicants, drugs showing synergy with cytokines, and drugs that were not previously identified by cytotoxicity assays (ATP depletion) in HepG2 cells, were selected to test in our model. Drugs not known to cause hepatotoxicity or any toxicity in humans were also selected to evaluate the specificity of the assay. We demonstrated that palmitate, at sublethal concentrations, was able to potentiate the cytotoxicity and/or apoptosis induced by some but not all drugs. We also explored the utility of this cell model, when combined with the plasma maximum concentration (Cmax), in identifying drug-induced toxicity or revealing idiosyncratic drugs that remain unidentified using currently deployed assays. We demonstrate that palmitate cotreatment potentiated the cytotoxicity of some idiosyncratic DILI drugs, but not that of nontoxic drugs, at therapeutically relevant dosing concentrations (Cmax). We conclude that this assay can be utilized to identify drugs that may be more toxic to patients with underlying metabolic disease.

MATERIALS AND METHODS

Materials. All chemicals were purchased from Sigma Aldrich (St Louis, MO). All tissue culture reagents were purchased from either Sigma or Invitrogen (Carlsbad, CA). The CellTiterGlo and Caspase-Glo 3/7 kits were purchased from Promega (Madison, WI). The 96-well white opaque bottom plates were purchased from Becton Dickinson Labware (Franklin Lakes, NJ).

Cell culture conditions for HepG2 cells. HepG2 cells were purchased from the American Type Culture Collection (Manassas, VA). Cells were grown in a culture medium containing 5mM glucose, Dulbecco’s modified Eagle’s medium (DMEM) supplemented with 10% fetal bovine serum (FBS), 5mM 4-(2-hydroxyethyl)piperazine-1-ethanesulfonic acid, and 100 units/ml penicillin-streptomycin in a 37°C, 5% CO2 humidified atmosphere. HepG2 cells were harvested in the exponential growth phase for the experiments and subcultured every 3 days up to passage 20. Cells were maintained in 175-cm2 flasks and seeded onto 96-well plates. To assess the mitochondrial liability, HepG2 cells were also maintained in DMEM supplemented with 10mM galactose instead of glucose as described (Swiss and Will, 2011).

Preparation of palmitate. Palmitic acid (Sigma, P5585) was conjugated to fatty acid–free (FFA) bovine serum albumin (BSA) (Sigma, A0281). Palmitic acid (51.2mg) was dissolved in 100% ethanol (1ml) to make 200mM of stock solution. About 10% FFA low-endotoxin BSA was prepared in DMEM. About 0.04ml of 200mM palmitate and 1.96ml of 10% FFA BSA were mixed for at least 2h to generate 4mM palmitate stock solution.

Compound treatment. HepG2 cells were plated at 10,000 cells/100 µl complete culture medium (containing 10% FBS) on 96-well plates and incubated at 37°C, 5% CO2, and 95% humidity for 18–24h to allow for cell attachment. About 4mM palmitate-BSA conjugate in 10% BSA or 10% BSA was diluted 20× in complete medium containing 10% FBS to prepare 200µM palmitate treatment or control medium. The total BSA concentrations in control and palmitate containing medium are kept the same to avoid differential protein binding effect on compounds. The following day, medium was aspirated from plates, and 100 µl of palmitate treatment or control medium containing compounds was transferred from the compound plates to the cell culture plates followed by 24-h incubation at 37°C.

Measurement of lactate dehydrogenase. To assess cell viability, following 24-h compound treatment, 40 µl media from each well were collected and lactate dehydrogenase (LDH) activities were measured using the Cytotoxicity Detection Kitplus (LDH) (Roche Applied Science, Mannheim, Germany) as described in manufacture’s manual. Absorbance of samples was measured with a SpectraMax spectrometer at the wavelength of 490 and 690nm, and “∆A = A490 nmA690 nm” was used to determine the percentage of LDH release. LDH release % = (∆A experiment − ∆A low control)/(∆A high control − ∆A low control) × 100. Low control determines LDH activity release from untreated cells and high control determines the maximum LDH activity in cells. The relative fold change in LDH release was calculated by normalizing to the control.

Measurement of caspase-3/7 as an indicator of apoptosis. Compound-treated plates were prepared as described above. To determine caspase-3/7 activity, 100 µl Caspase-Glo 3/7 reagent was added to each well. Plates were agitated for 30min in the dark at room temperature before luminescence was measured using an EnVision plate reader (PerkinElmer, Waltham, MA).

Statistical analysis. A two-sample unequal variance, two-tailed Student’s t-test was used to evaluate the statistical significance between the samples treated with and without palmitate at corresponding drug concentrations.

Selection of compounds for testing. To test the hypothesis that drug-induced toxicity could be modulated by palmitate, we selected 89 drugs to test in HepG2 cells (Table 1). The drug-induced toxicity was classified based on the information reported (Cosgrove et al., 2009), literature search, or drug labels (http://dailymed.nlm.nih.gov/dailymed). DILI information on compounds derived from literature or drug labels were classified according to the categories defined by Cosgrove et al. Drugs tested in this study were 14 nontoxic drugs (N1), 11 nonhepatotoxic drugs that showed toxicity in other organs (N2), 7 nonidiosyncratic hepatotoxic drugs (O1: toxic in animal and not tested in human; P1: toxic in animal/human), 51 drugs associated with idiosyncratic hepatotoxicity in humans (P2), and 6 drugs associated with rare liver toxicity (N3). The compounds were selected based on the following information: (1) Compounds reported to interfere with lipid metabolism, such as those known to induce steatosis. (2) Compounds that impair mitochondrial functions that may interfere with fatty acid β-oxidation. The effects of the compounds on mitochondrial function were evaluated in the glucose/galactose assay in HepG2 cells as described (Marroquin et al., 2007). A ratio of LD50glucose/LD50galactose ≥ 2 suggests that the compounds primarily impair mitochondrial function. Twenty compounds with LD50glucose/LD50galactose ≥ 2 were selected and labeled as “mito tox” (Table 1). (3) Inflammatory cytokines have been shown to potentiate the cytotoxicity induced by a number of drugs (Cosgrove et al., 2009), and palmitate and cytokines induce stress responses via similar pathways (Feldstein et al., 2004; Malhi et al., 2006). Therefore we also included a number of drugs that induce synergistic toxicity with the cytokine mix. (4) Drugs known to induce hepatotoxicity, which were either cytotoxic (LD50 ≤ 100μM in glucose culture condition) or noncytotoxic (LD50 > 100μM in glucose culture condition) in HepG2 cells, were selected to understand if cotreatment of palmitate could increase the sensitivity to identify these compounds. (5) Compounds without reported toxicity were selected as negative controls to evaluate the specificity. Therapeutically relevant drug exposure levels were defined by Cmax values reported in humans at commonly recommended therapeutic doses. Cmax values were obtained from a combination of literature searches and available databases. A concentration of 100-fold Cmax, encompassing a scaling factor to account for human population pharmacokinetic and toxicodynamic variability, was considered as a therapeutically relevant dosing limit for each drug, as previously described (Xu et al., 2008).

TABLE 1

Compounds Selected to Test in Palmitate Drug Cotreatment Model

Convention name Glu LD50 (μM) Gal LD50 (μM) DILI category Compound info Cmax (μM) 
Bezafibrate 300 300 N1 Noncytotoxic 16.58 
Buspirone 300 300 N1 Noncytotoxic 0.005 
Diphenhydramine 300 300 N1 Noncytotoxic 0.34 
Edrophonium 300 300 N1 Noncytotoxic   
Guanethidine 300 300 N1 Noncytotoxic 0.25 
Isoproterenol 300 300 N1 Noncytotoxic   
Loperamide 31 38 N1 Cytotoxic/cytokine potentiated 0.01 
Oxybutynin 170 78 N1 Mito tox 0.02 
Penbutolol 33 34 N1 Cytotoxic 0.96 
Phenoxybenzamine 300 160 N1 Noncytotoxic 0.2 
Probucol 300 300 N1 Noncytotoxic 7.55 
Propafenone 47 N1 Mito tox 0.001 
Propranolol 150 200 N1 Noncytotoxic/cytokine potentiated 0.2 
Theophylline 300 300 N1 Noncytotoxic 44.96 
Alendronate 300 300 N2 Noncytotoxic 0.14 
Cholecalciferol 300 138 N2 Noncytotoxic 0.1 
Droperidol 300 80 N2 Mito tox 0.01 
Ergocalciferol 300 186 N2 Noncytotoxic 0.05 
Foscarnet 300 300 N2 Noncytotoxic 186 
Loratadine 190 180 N2 Noncytotoxic 0.01 
Methysergide 300 300 N2 Noncytotoxic 
Pentamidine 300 N2 Mito tox 1.8 
Rifabutin 89 63 N2 Cytotoxic 0.68 
Sorafenib 20 N2 Mito tox 13.36 
Zoniporide 300 65 N2 Mito tox   
Benazepril 300 300 N3 Noncytotoxic 0.35 
Flufenamic acid 300 300 N3 Noncytotoxic 46.23 
Fluoxetine 18 19 N3 Cytotoxic/cytokine potentiated 0.04 
Ibuprofen 300 300 N3 Noncytotoxic 148.34 
Primaquine 190 190 N3 Noncytotoxic 0.23 
Rosiglitazone 300 290 N3 Noncytotoxic 1.04 
Cyproterone 300 144 O1 Noncytotoxic 0.65 
l-Ethionine 300 300 O1 Noncytotoxic/steatosis   
Amiodarone 110 59 P1 Noncytotoxic/steatosis/cytokine potentiated 1.86 
Benzbromarone 130 90 P1 Noncytotoxic/cytokine potentiated 3.44 
Hexachlorophene 71 16 P1 Mito tox 5.41 
Mercaptopurine 300 300 P1 Noncytotoxic 0.48 
Valproic acid 300 300 P1 Noncytotoxic/steatosis 153.94 
Mitoxantrone 21 29 P2 Cytotoxic 0.36 
Thioguanine 58 46 P2 Cytotoxic 0.06 
Tamoxifen 58 56 P2 Cytotoxic 0.16 
Clomipramine 27 48 P2 Cytotoxic/cytokine potentiated 0.19 
Fluvoxamine 38 38 P2 Cytotoxic/cytokine potentiated 0.12 
Nortriptyline 20 48 P2 Cytotoxic/cytokine potentiated 0.12 
Alpidem 229 68 P2 Mito tox 0.24 
Ethynodiol diacetate 136 34 P2 Mito tox 0.03 
Fipexide 300 115 P2 Mito tox   
Flecainide 300 120 P2 Mito tox 1.05 
Flutamide 300 73 P2 Mito tox 0.36 
Ketoconazole 170 53 P2 Mito tox 12.51 
Nefazodone 61 12 P2 Mito tox 1.32 
Troglitazone 300 150 P2 Mito tox 6.39 
Risperidone 300 130 P2 Mito tox 0.03 
Spironolactone 300 150 P2 Mito tox 0.19 
Nicardipine 300 68 P2 Mito tox 0.16 
Bicalutamide 300 137 P2 Mito tox 1.97 
Nitrofurantoin 300 97 P2 Mito tox 4.2 
Acenocoumarol 300 300 P2 Noncytotoxic 1823.33 
Acetazolamide 300 300 P2 Noncytotoxic 134.99 
Amrinone 300 300 P2 Noncytotoxic 13.78 
Betahistine 300 300 P2 Noncytotoxic 0.004 
Captopril 300 300 P2 Noncytotoxic 4.28 
Carbidopa 300 300 P2 Noncytotoxic 0.66 
Chloramphenicol 300 300 P2 Noncytotoxic 19.95 
Chlorpheniramine 300 211 P2 Noncytotoxic 0.52 
Cinchophen 300 300 P2 Noncytotoxic   
Clozapine 130 144 P2 Noncytotoxic 0.95 
Diclofenac 300 300 P2 Noncytotoxic 7.99 
Dipyridamole 300 300 P2 Noncytotoxic 3.65 
Estrone 300 300 P2 Noncytotoxic 0.02 
Furosemide 300 300 P2 Noncytotoxic 3.27 
Losartan 300 300 P2 Noncytotoxic 7.09 
Methimazole 300 300 P2 Noncytotoxic 1.87 
Nimesulide 300 300 P2 Noncytotoxic 21.08 
Nimodipine 103 78 P2 Noncytotoxic 0.11 
Norgestrel 300 300 P2 Noncytotoxic 0.01 
Phentolamine 210 300 P2 Noncytotoxic 0.09 
Piroxicam 300 300 P2 Noncytotoxic 5.13 
Pyrimethamine 300 300 P2 Noncytotoxic 3.29 
Quinapril 300 300 P2 Noncytotoxic 1.32 
Sitaxsentan 260 200 P2 Noncytotoxic 30.6 
Ticlopidine 300 300 P2 Noncytotoxic 1.9 
Zileuton 300 300 P2 Noncytotoxic 13.12 
Imipramine 210 180 P2 Noncytotoxic/cytokine potentiated 0.09 
Quinine 300 300 P2 Noncytotoxic/cytokine potentiated 9.25 
Riluzole 300 300 P2 Noncytotoxic/cytokine potentiated 1.65 
Trovafloxacin 300 300 P2 Noncytotoxic/cytokine potentiated 5.02 
Cyclosporine 300 300 P2 Noncytotoxic/steatosis 0.58 
Methotrexate 300 300 P2 Noncytotoxic/steatosis 0.77 
Tetracycline 300 300 P2 Noncytotoxic/steatosis 20.96 
Convention name Glu LD50 (μM) Gal LD50 (μM) DILI category Compound info Cmax (μM) 
Bezafibrate 300 300 N1 Noncytotoxic 16.58 
Buspirone 300 300 N1 Noncytotoxic 0.005 
Diphenhydramine 300 300 N1 Noncytotoxic 0.34 
Edrophonium 300 300 N1 Noncytotoxic   
Guanethidine 300 300 N1 Noncytotoxic 0.25 
Isoproterenol 300 300 N1 Noncytotoxic   
Loperamide 31 38 N1 Cytotoxic/cytokine potentiated 0.01 
Oxybutynin 170 78 N1 Mito tox 0.02 
Penbutolol 33 34 N1 Cytotoxic 0.96 
Phenoxybenzamine 300 160 N1 Noncytotoxic 0.2 
Probucol 300 300 N1 Noncytotoxic 7.55 
Propafenone 47 N1 Mito tox 0.001 
Propranolol 150 200 N1 Noncytotoxic/cytokine potentiated 0.2 
Theophylline 300 300 N1 Noncytotoxic 44.96 
Alendronate 300 300 N2 Noncytotoxic 0.14 
Cholecalciferol 300 138 N2 Noncytotoxic 0.1 
Droperidol 300 80 N2 Mito tox 0.01 
Ergocalciferol 300 186 N2 Noncytotoxic 0.05 
Foscarnet 300 300 N2 Noncytotoxic 186 
Loratadine 190 180 N2 Noncytotoxic 0.01 
Methysergide 300 300 N2 Noncytotoxic 
Pentamidine 300 N2 Mito tox 1.8 
Rifabutin 89 63 N2 Cytotoxic 0.68 
Sorafenib 20 N2 Mito tox 13.36 
Zoniporide 300 65 N2 Mito tox   
Benazepril 300 300 N3 Noncytotoxic 0.35 
Flufenamic acid 300 300 N3 Noncytotoxic 46.23 
Fluoxetine 18 19 N3 Cytotoxic/cytokine potentiated 0.04 
Ibuprofen 300 300 N3 Noncytotoxic 148.34 
Primaquine 190 190 N3 Noncytotoxic 0.23 
Rosiglitazone 300 290 N3 Noncytotoxic 1.04 
Cyproterone 300 144 O1 Noncytotoxic 0.65 
l-Ethionine 300 300 O1 Noncytotoxic/steatosis   
Amiodarone 110 59 P1 Noncytotoxic/steatosis/cytokine potentiated 1.86 
Benzbromarone 130 90 P1 Noncytotoxic/cytokine potentiated 3.44 
Hexachlorophene 71 16 P1 Mito tox 5.41 
Mercaptopurine 300 300 P1 Noncytotoxic 0.48 
Valproic acid 300 300 P1 Noncytotoxic/steatosis 153.94 
Mitoxantrone 21 29 P2 Cytotoxic 0.36 
Thioguanine 58 46 P2 Cytotoxic 0.06 
Tamoxifen 58 56 P2 Cytotoxic 0.16 
Clomipramine 27 48 P2 Cytotoxic/cytokine potentiated 0.19 
Fluvoxamine 38 38 P2 Cytotoxic/cytokine potentiated 0.12 
Nortriptyline 20 48 P2 Cytotoxic/cytokine potentiated 0.12 
Alpidem 229 68 P2 Mito tox 0.24 
Ethynodiol diacetate 136 34 P2 Mito tox 0.03 
Fipexide 300 115 P2 Mito tox   
Flecainide 300 120 P2 Mito tox 1.05 
Flutamide 300 73 P2 Mito tox 0.36 
Ketoconazole 170 53 P2 Mito tox 12.51 
Nefazodone 61 12 P2 Mito tox 1.32 
Troglitazone 300 150 P2 Mito tox 6.39 
Risperidone 300 130 P2 Mito tox 0.03 
Spironolactone 300 150 P2 Mito tox 0.19 
Nicardipine 300 68 P2 Mito tox 0.16 
Bicalutamide 300 137 P2 Mito tox 1.97 
Nitrofurantoin 300 97 P2 Mito tox 4.2 
Acenocoumarol 300 300 P2 Noncytotoxic 1823.33 
Acetazolamide 300 300 P2 Noncytotoxic 134.99 
Amrinone 300 300 P2 Noncytotoxic 13.78 
Betahistine 300 300 P2 Noncytotoxic 0.004 
Captopril 300 300 P2 Noncytotoxic 4.28 
Carbidopa 300 300 P2 Noncytotoxic 0.66 
Chloramphenicol 300 300 P2 Noncytotoxic 19.95 
Chlorpheniramine 300 211 P2 Noncytotoxic 0.52 
Cinchophen 300 300 P2 Noncytotoxic   
Clozapine 130 144 P2 Noncytotoxic 0.95 
Diclofenac 300 300 P2 Noncytotoxic 7.99 
Dipyridamole 300 300 P2 Noncytotoxic 3.65 
Estrone 300 300 P2 Noncytotoxic 0.02 
Furosemide 300 300 P2 Noncytotoxic 3.27 
Losartan 300 300 P2 Noncytotoxic 7.09 
Methimazole 300 300 P2 Noncytotoxic 1.87 
Nimesulide 300 300 P2 Noncytotoxic 21.08 
Nimodipine 103 78 P2 Noncytotoxic 0.11 
Norgestrel 300 300 P2 Noncytotoxic 0.01 
Phentolamine 210 300 P2 Noncytotoxic 0.09 
Piroxicam 300 300 P2 Noncytotoxic 5.13 
Pyrimethamine 300 300 P2 Noncytotoxic 3.29 
Quinapril 300 300 P2 Noncytotoxic 1.32 
Sitaxsentan 260 200 P2 Noncytotoxic 30.6 
Ticlopidine 300 300 P2 Noncytotoxic 1.9 
Zileuton 300 300 P2 Noncytotoxic 13.12 
Imipramine 210 180 P2 Noncytotoxic/cytokine potentiated 0.09 
Quinine 300 300 P2 Noncytotoxic/cytokine potentiated 9.25 
Riluzole 300 300 P2 Noncytotoxic/cytokine potentiated 1.65 
Trovafloxacin 300 300 P2 Noncytotoxic/cytokine potentiated 5.02 
Cyclosporine 300 300 P2 Noncytotoxic/steatosis 0.58 
Methotrexate 300 300 P2 Noncytotoxic/steatosis 0.77 
Tetracycline 300 300 P2 Noncytotoxic/steatosis 20.96 

Note. Mitochondrial toxicity of the compounds was assessed by the differential LD50 in HepG2 cells cultured in medium supplemented with glucose or galactose. LD50 was determined after 24-h incubation of compounds and LD50 ratio (glucose LD50/galactose LD50) of > 2 is considered as mitochondrial toxicants (mito tox). LD50 of < 100μM in glucose culture condition is considered as cytotoxic.

RESULTS

Identification of Optimal Palmitate Concentration

The goal of this initial study was to identify the optimal concentrations of palmitate that would induce minimal cytotoxicity in order to be able to study the effect of palmitate on drug-induced cytotoxicity. HepG2 cells were treated for 24h with 100–800μM of palmitic acid conjugated to BSA. Cytotoxicity and apoptosis levels were evaluated by measuring LDH released in medium and cellular caspase-3/7 activity, respectively as described in Materials and Methods.

Palmitic acid dose-dependently induced LDH release and caspase activity (Fig. 1). At 200μM, palmitic acid minimally increased LDH release but significantly increase apoptosis, determined by measuring caspase-3/7 activity. Marked increase of apoptosis and cytotoxicity were observed with palmitic acid at concentrations of 300μM or more. Based on these results, 200μM palmitic acid, which cause minimal stress in cells, was selected to test the effect of palmitate on drug-induced toxicity.

FIG. 1.

Determination of optimal palmitate concentrations that induce stress in HepG2 cells. Cells were treated with various concentrations of palmitic acid for 24h followed by measurement of LDH release for cell viability (panel A) and cellular caspase-3/7 activity for apoptosis (panel B). Relative LDH release was calculated as described in Materials and Methods. Fold change of caspase activity was determined by normalizing to control (no palmitic acid). Data are presented as mean ± SE from six biological samples from two separate experiments. #p < 0.05 and *p < 0.01 compared with control.

FIG. 1.

Determination of optimal palmitate concentrations that induce stress in HepG2 cells. Cells were treated with various concentrations of palmitic acid for 24h followed by measurement of LDH release for cell viability (panel A) and cellular caspase-3/7 activity for apoptosis (panel B). Relative LDH release was calculated as described in Materials and Methods. Fold change of caspase activity was determined by normalizing to control (no palmitic acid). Data are presented as mean ± SE from six biological samples from two separate experiments. #p < 0.05 and *p < 0.01 compared with control.

Identification of Drugs That Demonstrate Synergistic Toxicity With Palmitate Cotreatment

The drugs were tested at 300, 100, 30, 10, 3, 1, and 0.3μM concentrations in the presence or absence of 200μM palmitate for 24h as described in the Materials and Methods. Cytotoxicity and apoptosis were assessed by measuring LDH released into the medium and cellular caspase-3/7 activity. The relative fold change in LDH release or caspase activity was calculated by normalizing to HepG2 cells incubated without compounds and palmitate. Synergistic induction of cell death or caspase activity was assessed by a supra-additive synergy criterion that compares the relative fold change in LDH or caspase activity induced by drug-palmitate cotreatment to the additive projection of relative fold change in LDH or caspase activity obtained from drug-only and palmitate-only treatments (additive projection = LDH or caspasedrug-only + increased LDH or caspasepalmitate-only). Figure 2a shows the examples, nortriptyline and dipyridamole, which displayed synergistic toxicity in both cell death (Fig. 2a, panel A) and apoptosis (Fig. 2a, panel B) with palmitate/drug cotreatment. The summary of the results for LDH release are shown in Figure 2b in the form of a heatmap. The drug-palmitate synergy in LDH release was observed for 27/89 (30.3%) compounds tested: 1/7 (14.2%) O1+P1 hepatotoxicants, 20/51 (39.2%) P2 idiosyncratic hepatotoxicants, 1/6 (16.6%) hepatotoxicants with rare cases (N3), and 5/25 (20%) nonhepatotoxicants (N1, N2) (Supplementary table 1). The drugs that showed synergistic cytotoxicity and apoptosis with palmitate cotreatment are summarized in Figures 2c and 2d, respectively. Drug-palmitate induced synergy in apoptosis was observed in none of the overt hepatotoxicants (O1 and P1), 20/51 (39.2%) P2 idiosyncratic hepatotoxicants, 0/6 drugs with rare hepatotoxicity (N3), and 2/24 (8.3%) nonhepatotoxicants (N1/N2).

FIG. 2a.

Identification of drugs that demonstrate synergistic toxicity with palmitate cotreatment. (a) HepG2 cells were treated with 200μM palmitate (PA) and compounds for 24h followed by assays for LDH release and cellular caspase-3/7 activity as described in Materials and Methods. LDH release and caspase activity are presented as relative fold changes normalized to DMSO vehicle/no PA control samples. Additive projection was calculated by adding LDH or caspasedrug-only and increased LDH or caspasepalmitate-only. Data are presented as mean ± SD of triplicates. #p < 0.05 and *p < 0.01 comparing PA cotreatment samples to the corresponding additive projection at various drug concentrations. Two examples, nortriptyline (panel A) and dipyridamole (panel B), which exhibited synergistic toxicity are shown. (b) Summary results of the large-scale screening for LDH release. All drugs were dosed at 0, 0.3, 1, 3, 10, 30, 100, and 300μM concentrations in presence or absence of 200μM palmitate. The differential between fold change of LDHdrug+palmitate and LDH fold change of additive projection (calculated as described in 2a) was calculated and plotted in the heatmaps. Most of the drugs that showed synergistic toxicity are in the category of P2 idiosyncratic hepatotoxicity. (c) Summary of hits that showed synergistic toxicity with palmitate. Drugs that showed synergistic toxicity at more than one concentrations (p < 0.05 comparing fold change of LDHdrug+palmitate and additive projections at a corresponding drug concentration) are shown and the heatmap is plotted as in 2b. The 100 × Cmax values are marked in dots for each drug. Drugs that exhibited synergistic toxicity at concentrations less than their 100 × Cmax are highlighted in red. The data and 100 × Cmax for two representative drugs exhibiting synergistic toxicity are shown. Pyrimethamine, but not loratadine, displayed synergistic toxicity at concentration less than 100 × Cmax. #p < 0.05 comparing fold change of LDHdrug+palmitate and corresponding additive projection at various drug concentrations. +/− palmitate differentials are illustrated as double arrows. (d) Summary of drugs that demonstrate synergistic increase of apoptosis with palmitate cotreatment. The differential between relative fold change of caspasedrug+palmitate and caspase fold change of additive projection (calculated as described in 2a) was calculated and is plotted in the heatmaps. Drugs that showed synergistic toxicity at more than one concentration (p < 0.05 comparing fold change of caspasedrug+palmitate and corresponding additive projections) are shown. The 100 × Cmax values are marked in dots for each drug. The data and 100 × Cmax for two representative drugs exhibiting synergistic toxicity are shown. Pyrimethamine, but not loratadine, displayed synergistic toxicity at concentrations less than 100 × Cmax. #p < 0.05 comparing fold change of caspasedrug+palmitate and corresponding additive projections at various drug concentrations.

FIG. 2a.

Identification of drugs that demonstrate synergistic toxicity with palmitate cotreatment. (a) HepG2 cells were treated with 200μM palmitate (PA) and compounds for 24h followed by assays for LDH release and cellular caspase-3/7 activity as described in Materials and Methods. LDH release and caspase activity are presented as relative fold changes normalized to DMSO vehicle/no PA control samples. Additive projection was calculated by adding LDH or caspasedrug-only and increased LDH or caspasepalmitate-only. Data are presented as mean ± SD of triplicates. #p < 0.05 and *p < 0.01 comparing PA cotreatment samples to the corresponding additive projection at various drug concentrations. Two examples, nortriptyline (panel A) and dipyridamole (panel B), which exhibited synergistic toxicity are shown. (b) Summary results of the large-scale screening for LDH release. All drugs were dosed at 0, 0.3, 1, 3, 10, 30, 100, and 300μM concentrations in presence or absence of 200μM palmitate. The differential between fold change of LDHdrug+palmitate and LDH fold change of additive projection (calculated as described in 2a) was calculated and plotted in the heatmaps. Most of the drugs that showed synergistic toxicity are in the category of P2 idiosyncratic hepatotoxicity. (c) Summary of hits that showed synergistic toxicity with palmitate. Drugs that showed synergistic toxicity at more than one concentrations (p < 0.05 comparing fold change of LDHdrug+palmitate and additive projections at a corresponding drug concentration) are shown and the heatmap is plotted as in 2b. The 100 × Cmax values are marked in dots for each drug. Drugs that exhibited synergistic toxicity at concentrations less than their 100 × Cmax are highlighted in red. The data and 100 × Cmax for two representative drugs exhibiting synergistic toxicity are shown. Pyrimethamine, but not loratadine, displayed synergistic toxicity at concentration less than 100 × Cmax. #p < 0.05 comparing fold change of LDHdrug+palmitate and corresponding additive projection at various drug concentrations. +/− palmitate differentials are illustrated as double arrows. (d) Summary of drugs that demonstrate synergistic increase of apoptosis with palmitate cotreatment. The differential between relative fold change of caspasedrug+palmitate and caspase fold change of additive projection (calculated as described in 2a) was calculated and is plotted in the heatmaps. Drugs that showed synergistic toxicity at more than one concentration (p < 0.05 comparing fold change of caspasedrug+palmitate and corresponding additive projections) are shown. The 100 × Cmax values are marked in dots for each drug. The data and 100 × Cmax for two representative drugs exhibiting synergistic toxicity are shown. Pyrimethamine, but not loratadine, displayed synergistic toxicity at concentrations less than 100 × Cmax. #p < 0.05 comparing fold change of caspasedrug+palmitate and corresponding additive projections at various drug concentrations.

The majority of drugs that showed synergy in inducing LDH release (cell death) also induced apoptosis, measured by the increase of caspase-3/7 activity (Figs. 2c and 2d). However, some drugs induced synergistic increase of caspase activity but not LDH release, such as propafenone (N1) and the hepatotoxic drugs spironolactone, ticlopidine, and zileuton. On the other hand, some drugs showed synergy with palmitate in inducing LDH release but not apoptosis, such as the hepatotoxicants ketoconazole, quinine, amiodarone, nicardipine, and nefazadone; and the nonhepatotoxicants (N1 and N2) droperidol, loperamide, and methysergide.

To evaluate the clinical significance of the synergistic toxicity induced by palmitate and drug cotreatment, 100× of each drug’s own Cmax (peak plasma drug concentration) was used to examine whether palmitate would enhance drug-induced toxicity at therapeutically relevant exposure. The 100 × Cmax value was applied to evaluate idiosyncratic hepatotoxic drugs (Xu et al., 2008). The 100 × Cmax resulted from a scaling of several factors: a scaling factor of sixfold that accounts for population Cmax variability due to patient genetic or epigenetic factors that affect drug clearances, another sixfold uncertainty factor to account for higher drug exposure to the liver via liver portal vein for orally dosed drugs (Ito et al., 2002), and a final threefold uncertainty factor to account for drug-drug or drug-diet interactions and potential for increased drug exposure upon multiple days of dosing compared with the single-dose Cmax values used in in vitro studies.

As shown in Figure 2c, drugs that displayed drug-palmitate synergy in inducing cell death at concentrations less than 100 × Cmax were highlighted in red. None of the N1/N2 drugs showed drug-palmitate synergy at less than 100 × Cmax. In contrast, 15/20 (75%) of the idiosyncratic P2 drugs were potentiated by palmitate at concentrations at or less than its 100 × Cmax.

Of the six drugs that were reported to induce steatosis, only CsA and amiodarone showed synergy with palmitate in inducing LDH release and apoptosis, suggesting that disturbance of the lipid metabolism is not necessarily associated with the synergistic toxicity. Of the 20 drugs that were suggested to impair mitochondrial function by differential cytotoxicity in HepG2 cells cultured in glucose and galactose medium (Table 1), 9/20 (45%) showed synergy with palmitate in inducing LDH release. The ability of these compounds to synergize with palmitate did not correlate with the ratio of LD50glucose/LD50galactose or the LD50. Ten out of the 12 (83%) drugs that were reported to synergistically induce cytotoxicity with the cytokine mix (Table 1) demonstrated synergy with palmitate.

Characterization of Compounds That Synergized With Palmitate in Inducing Toxicity

A number of tricyclic antipsychotics/antidepressants (TCA), such as nortriptyline, clomipramine, imipramine, and clozapin (Fig. 3a for structures), showed synergistic toxicity with palmitate. Therefore we tested three additional TCA-like drugs, namely amitriptyline, maprotiline, and amoxapine. We found that the toxicities induced by all of these TCA compounds were enhanced in the presence of palmitate (Fig. 3b). The inductions of caspase activity by the TCA were also enhanced by palmitate (Fig. 3c). Loratadine, a H1 histamine antagonist, with a tricyclic-like structure, also synergistically induced cytotoxicity and caspase-3/7 activity in the presence of palmitate (Figs. 2b and 2c). The low incidence of reported toxicity for loratadine in humans is most likely due to its low exposure with a 100 × Cmax of approximately 1μM. The significant induction of toxicity was only observed at ≥ 3μM.

FIG. 3.

TCA synergize with palmitate to induce cell death and apoptosis. (a) Structures of selected tricyclic drugs. (b) TCA and palmitate cotreatment synergistically induce cell death and apoptosis. Cytotoxicity and apoptosis data are shown for clomipramine (panel A), amitriptyline (panel B), maprotiline (panel C), imipramine (panel D), amoxapine (panel E), and clozapine (panel F). The experimental procedures are the same as described in Figure 2. #p < 0.05 and *p < 0.01 comparing palmitate cotreatment samples to the corresponding additive projection at various drug concentrations.

FIG. 3.

TCA synergize with palmitate to induce cell death and apoptosis. (a) Structures of selected tricyclic drugs. (b) TCA and palmitate cotreatment synergistically induce cell death and apoptosis. Cytotoxicity and apoptosis data are shown for clomipramine (panel A), amitriptyline (panel B), maprotiline (panel C), imipramine (panel D), amoxapine (panel E), and clozapine (panel F). The experimental procedures are the same as described in Figure 2. #p < 0.05 and *p < 0.01 comparing palmitate cotreatment samples to the corresponding additive projection at various drug concentrations.

A number of structurally diverse antidepressants, such as the selective serotonin reuptake inhibitors (SSRI) fluoxetine and fluvoxamine (Fig. 4a for structures) also exhibited synergistic toxicity in HepG2 cells (Fig. 4b, panel A and B). We further tested the SSRI paroxetine and confirmed that it showed similar activity to other SSRIs (Fig. 4b, panel C). Chlorpheniramine (also called chlorphenamine), an antihistamine used for the prevention of allergic symptoms, robustly synergized toxicity with palmitate (Fig. 2c). Chlorpheniramine also has been reported to act as a SSRI and norepinephrine uptake inhibitor. The serotonin receptor antagonists, nefazoadone and risperidone also displayed synergistic toxicity with palmitate in HepG2 cells. With the exception of the antidepressants that target serotonin receptors or serotonin transporters, all other compounds that displayed synergistic toxicity were structurally diverse and had various primary pharmacological targets.

FIG. 4.

SSRI synergize with palmitate to induce cell death and apoptosis. (a) Structures of selected SSRI. (b) SSRI and palmitate cotreatment synergistically induce cell death and apoptosis. Cytotoxicity and apoptosis data are shown for fluoxetine (panel A), fluvoxamine (panel B), and paroxetine (panel C). #p < 0.05 and *p < 0.01 comparing palmitate cotreatment samples to the corresponding additive projection at various drug concentrations.

FIG. 4.

SSRI synergize with palmitate to induce cell death and apoptosis. (a) Structures of selected SSRI. (b) SSRI and palmitate cotreatment synergistically induce cell death and apoptosis. Cytotoxicity and apoptosis data are shown for fluoxetine (panel A), fluvoxamine (panel B), and paroxetine (panel C). #p < 0.05 and *p < 0.01 comparing palmitate cotreatment samples to the corresponding additive projection at various drug concentrations.

DISCUSSION

In this study, we aspired to develop an in vitro assay that would mimic a metabolic disease state, and to explore if such a model could potentiate drug-induced toxicity or discover idiosyncratic drugs that have not been identified using previously deployed cytotoxicity assays. We did so by testing if the free fatty acid palmitate would affect drug-induced LDH release (cytotoxicity) and apoptosis in cultured HepG2 cells. A number of drugs reported to induce hepatotoxicity, including steatosis, in humans were selected to test the hypothesis. Drugs without reported hepatotoxicity were also tested to evaluate the specificity of the assay. We demonstrated that palmitate, at sublethal concentrations, was able to potentiate the cytotoxicity and/or apoptosis induced by a variety of drugs. The palmitate and drug synergized toxicity, combined with plasma maximum drug concentration (Cmax), allowed us to identify idiosyncratic toxic drugs that were not flagged in previously reported cytotoxicity assays. The data suggest that treatment of palmitate could increase the sensitivity of cells to cytotoxicity induced by certain compounds. Furthermore, this assay may be used to identify compounds that have higher safety risks in patients with metabolic syndrome.

Of the drugs known to cause steatosis, CsA and amiodarone exhibited synergistic toxicity with palmitate. Both drugs have been shown to inhibit fatty acid β-oxidation and impair mitochondrial respiration (Begriche et al., 2011; Lemmi et al., 1990). It is well documented that CsA induces oxidative stress in cells (Wolf et al., 1997). Our laboratory has observed that CsA increases the GSSG (oxidized glutathione)/GSH (reduced glutathione) ratio in HepG2, indicating an elevated oxidative state (Luo et al., 2012). Furthermore, it has been shown that amiodarone induces oxidative stress in cultured cells including HepG2 (Golli-Bennour et al., 2010). Tetracycline-induced steatosis is due to either the increase of lipogenesis and/or the inhibition of fatty acid oxidation (Yin et al., 2006). The steatogenic mechanism for methotrexate is not clear. Although methotrexate-induced depletion of glutathione and subsequent oxidative stress has been shown to mediate its toxicity in animals (Tabassum et al., 2010), it did not induce oxidative stress in HepG2 cells or hepatocytes (Cordero et al., 2010). In addition, we did not observe any synergistic toxicity with methotrexate/palmitate in rat primary hepatocytes (data not shown). It appears that oxidative stress may be one of the elements contributing to the synergistic toxicity.

Of the drugs classified as mitochondrial toxicants based on LD50glucose/LD50galactose ratio, only 9/20 (45%) compounds displayed synergy with palmitate in inducing LDH release. Spironolactone and propafenone did not synergize LDH release, but did show synergistic induction of caspase-3/7 activity. Compounds not classified as mitochondrial toxicants based on the glucose/galactose differential LD50, showed robust synergy with palmitate in inducing cell death and/or caspase-3/7. These compounds were CsA, fluoxetine, and a number of TCA drugs. Therefore, it appears that impairment of mitochondrial function may contribute to, but may not be an essential factor for, the synergistic toxicity.

A number of drugs, either cytotoxic or noncytotoxic, have been reported to synergize cytotoxicity when cytokines are added to the cells (Cosgrove et al., 2009). We observed a high concordance of compounds that demonstrated synergy with palmitate or cytokine mix. Ten out of the 12 cytokine-potentiated compounds (83%), except benzbromarone and trovafloxacin, displayed enhanced toxicity in the presence of palmitate, which may reflect a common stress response signaling pathway induced by palmitate and cytokines. The JNK has been demonstrated to be activated by both cytokines and palmitate (Malhi et al., 2006). JNK has been suggested to be a likely component for drug-inflammatory cytokine-induced toxicity (Cosgrove et al., 2009). We have previously demonstrated that palmitate and cyclosporine activate JNK, and that the JNK inhibitor SP600125 was able to attenuate CsA/palmitate synergized toxicity (Luo et al., 2012), suggesting a critical role of JNK in mediating palmitate/drug toxicity. Benzbromarone and trovafloxacin demonstrated synergistic toxicity with palmitate in rat primary hepatocytes but not in HepG2 cells in Cosgrove’s study (Cosgrove et al., 2009). The discrepancy could be due to different cells used in these studies. Reactive metabolites of benzbromarone and trovafloxacin have been reported to be associated with their toxicity (McDonald and Rettie, 2007; Sun et al., 2008). The lack of synergy of these two compounds in our study could be due to the absence of bioactivation in HepG2 cells. There are several drugs, such as troglitazone, primethamine, and nimodipine, which showed synergy with palmitate but not with cytokines (Cosgrove et al., 2009), indicating the differential response to palmitate and cytokine treatment. Primethamine and nimodipine are hepatotoxicants (P2) (Table 1), and their toxicities were revealed only in the presence of palmitate cotreatment. Primethamine (100 × Cmax of 400μM) showed significant toxicity at 30–100μM in the presence of palmitate (Fig. 2d). Although nimodipine/palmitate cotreatment induced cytotoxicity at about 100μM, which is way above the 100 × Cmax (13μM), it is possible that longer treatment could increase toxicity at lower concentration. Therefore, the palmitate cotreatment model has its value in identifying cytotoxic compounds that were not previously flagged by cytokine cotreatment assays. However, we must keep in mind that the cultured HepG2 cells have limitations, such as the lack of mechanisms for bioactivation, detoxification, therefore, the assay could either over- or underestimate the toxicity of drugs. Testing of drugs in primary hepatocytes may provide extra information, but the cultured primary hepatocytes in many aspects are far from close to the hepatocytes in liver. Nevertheless, this assay has its value to flag certain mechanisms of toxicity in the early phases of compound selection and to choose the best compounds regarding safety to advance in drug development.

A class of antidepressants, including the TCA and SSRI, showed robust synergy with palmitate in inducing toxicity in HepG2 cells. Both TCAs and SSRIs have been shown to inhibit serotonin and/or norepinephrine transporters (Schloss and Williams, 1998). The antipsychotic drugs, nefazoadone and risperidone, which target serotonin receptors also displayed synergistic toxicity. We were unable to find any literature evidence suggesting the expression of serotonin transporter in hepatocytes or in HepG2, however we cannot exclude the possibility that the transporter is expressed in the cells. Because these drugs have very potent effect on serotonin transporter or the receptors with Ki ranging from < 1nM to 100nM (Gillman, 2007) and the enhanced toxicity was observed in HepG2 cells at > 1μM, therefore it is unlikely that the synergized toxicity we observed with this class of compounds is due to the primary pharmacology. Many of the compounds in this class are lipophilic amines and have been reported to induce oxidative stress in various cell types (El-Demerdash and Mohamadin, 2004). JNK activation appears to be the merging point for this class of drugs, palmitate and the cytokines (Gao et al., 2010; Malhi et al., 2006). We speculate that the induction of oxidative stress–activated JNK may be one of the underlying mechanisms that contribute to the synergistic toxicity with palmitate and cytokine mix.

The synergistic toxicity induced by drugs and palmitate indicate that elevation of free fatty acid levels, such as in patients with metabolic syndrome and fatty liver, could increase the susceptibility to drug-induced toxicity. Many of the hepatotoxic drugs tested did not show cytotoxicity at therapeutic exposure (at or less than 100 × Cmax); but some of them did significantly induce both cell death and/or apoptosis in the presence of palmitate at ≤ 100 × Cmax. There may be a relatively large proportion of drug recipients have dyslipidemia, whereas the incidence of DILI is low. This discrepancy is supported by our observation that many of these drugs showed enhanced toxicity at 10–100 × Cmax, only very few, such as ketoconazole, showed palmitate/drug synergized toxicity at 1 × Cmax. The low incidence of idiosyncratic DILI may be due to the variations of drug exposure and disposition in the human population, which have up to 100× variations (Xu et al., 2008). In addition, we cannot exclude the possibility that some mild DILI could be overlooked in clinical practice, thus underestimate the incidence rate. Clinical studies correlating adverse events and plasma lipid levels are scarce. There are several reports associating elevated lipid levels with CsA-induced toxicity in posttransplant patients. Preoperative obesity was reported to be associated with decreased survival after heart transplantation (Armstrong et al., 2005; Grady et al., 1996). Our study warrants further clinical or preclinical in vivo studies to investigate the correlation of free fatty acid levels and adverse events induced by drugs shown enhanced toxicity with palmitate. It should be noted that the 200μM palmitate concentration used in the assay is within the physiological range, and palmitate levels increase from approximately 100μM in healthy subjects to about 200μM in obese patients. It has been shown that several antidepressants activate sterol regulatory element-binding protein, which leads to induction of lipogenesis (Raeder et al., 2006). Lipogenesis would increase free fatty acid levels, which could in turn exacerbate the drug-induced toxicity. Here we only evaluated drug-induced cytotoxicity, it is likely that free fatty acids could also synergize with drugs to induce sublethal or nonlethal toxicity. For example, elevated fatty acid may increase methotrexate-induced steatosis (Aarsaether et al., 1988).

On the other hand, the synergistic toxicity we observed with drugs and palmitate, could also imply that these drugs may enhance lipotoxicity and induce metabolic disease, or aggravate preexisting metabolic disorders. Indeed, many antidepressants, antipsychotics, and CsA have increased the risk of metabolic syndromes, including obesity, dyslipidemia, and diabetes (Baum, 2001; Pramyothin and Khaodhiar, 2010). Although we cannot exclude the possibility that perturbation of serotonin signaling and/or reuptake may contribute to the metabolic disorders in humans, our results indicate that the metabolic adverse effect caused by these drugs could be, at least partly, attributable to the elevated lipotoxicity mediated by fatty acids, such as palmitate, and drugs. The exaggerated lipotoxicity would lead to insulin resistance and β-cell apoptosis (Eitel et al., 2002; Lee et al., 2010; Lemmi et al., 1990). Many TCA, SSRIs, and serotonin/norepinephrine reuptake inhibitors are amphiphilic cationic compounds (Figs. 3a and 4a), and have been shown to disturb lipid metabolism and induce phospholipidosis (Xia et al., 2000). Paroxetine has been shown to induce insulin resistance in cells (Levkovitz et al., 2007). It would be worthwhile to investigate whether these drugs worsen the insulin resistance and β-cell dysfunction induced by palmitate.

In summary, this study provided novel evidence that the saturated free fatty acid, palmitate, and certain drugs act synergistically to induce toxicity in cultured cells. It appears that multiple mechanisms, such as oxidative stress, JNK activation, and impairment of mitochondrial function, may be involved in mediating the enhanced toxicity. These observations warrant further studies to understand the mechanisms of potentiation for each drug or drug class identified here. The synergized toxicity would reduce therapeutic index, which could lead to the manifestation of idiosyncratic toxicity at therapeutic exposure in subjects that have elevated free fatty acids. Furthermore, these drugs could enhance lipotoxicity and aggravate preexisting metabolic disorders or induce the onset of adverse metabolic effects. Our findings suggest that patients with underlying metabolic syndrome may be predisposed to some drug-induced toxicity, including worsening preexisting metabolic disorders. This assay may be useful to identify compounds that have higher safety risks in a population with metabolic syndrome.

SUPPLEMENTARY DATA

Supplementary data are available online at http://toxsci.oxfordjournals.org/.

Acknowledgments

We thank our colleagues Sashikala Nadanaciva and William Pennie for insightful discussion suggestion. We appreciate the technique support from Rachel Swiss.

References

Aarsaether
N.,
Berge
R. K.,
Aarsland
A.,
Svardal
A.,
Ueland
P. M.
(
1988
)
Effect of methotrexate on long-chain fatty acid metabolism in liver of rats fed a standard or a defined, choline-deficient diet
.
Biochim. Biophys. Acta
 
958,
70
80
Armstrong
K. A.,
Hiremagalur
B.,
Haluska
B. A.,
Campbell
S. B.,
Hawley
C. M.,
Marks
L.,
Prins
J.,
Johnson
D. W.,
Isbel
N. M.
(
2005
)
Free fatty acids are associated with obesity, insulin resistance, and atherosclerosis in renal transplant recipients
.
Transplantation
 
80,
937
944
Baum
C. L.
(
2001
)
Weight gain and cardiovascular risk after organ transplantation
.
JPEN
 
25,
114
119
Begriche
K.,
Massart
J.,
Robin
M. A.,
Borgne-Sanchez
A.,
Fromenty
B.
(
2011
)
Drug-induced toxicity on mitochondria and lipid metabolism: Mechanistic diversity and deleterious consequences for the liver
.
J. Hepatol.
 
54,
773
794
Boelsterli
U. A.
(
2003
)
Disease-related determinants of susceptibility to drug-induced idiosyncratic hepatotoxicity
.
Curr. Opin. Drug Discov. Devel.
 
6,
81
91
Cordero
M. D.,
Sanchez-Alcazar
J. A.,
Bautista-Ferrufino
M. R.,
Carmona-Lopez
M. I.,
Illanes
M.,
Rios
M. J.,
Garrido-Maraver
J.,
Alcudia
A.,
Navas
P.,
de Miguel
M.
(
2010
)
Acute oxidant damage promoted on cancer cells by amitriptyline in comparison with some common chemotherapeutic drugs
.
Anticancer Drugs
 
21,
932
944
Cosgrove
B. D.,
King
B. M.,
Hasan
M. A.,
Alexopoulos
L. G.,
Farazi
P. A.,
Hendriks
B. S.,
Griffith
L. G.,
Sorger
P. K.,
Tidor
B.,
Xu
J. J.,
et al
(
2009
)
Synergistic drug-cytokine induction of hepatocellular death as an in vitro approach for the study of inflammation-associated idiosyncratic drug hepatotoxicity
.
Toxicol. Appl. Pharmacol.
 
237,
317
330
Eitel
K.,
Staiger
H.,
Brendel
M. D.,
Brandhorst
D.,
Bretzel
R. G.,
Haring
H. U.,
Kellerer
M.
(
2002
)
Different role of saturated and unsaturated fatty acids in beta-cell apoptosis
.
Biochem. Biophys. Res. Commun.
 
299,
853
856
El-Demerdash
E.,
Mohamadin
A. M.
(
2004
)
Does oxidative stress contribute in tricyclic antidepressants-induced cardiotoxicity?
Toxicol. Lett.
 
152,
159
166
Feldstein
A. E.,
Werneburg
N. W.,
Canbay
A.,
Guicciardi
M. E.,
Bronk
S. F.,
Rydzewski
R.,
Burgart
L. J.,
Gores
G. J.
(
2004
)
Free fatty acids promote hepatic lipotoxicity by stimulating TNF-alpha expression via a lysosomal pathway
.
Hepatology
 
40,
185
194
Gao
D.,
Nong
S.,
Huang
X.,
Lu
Y.,
Zhao
H.,
Lin
Y.,
Man
Y.,
Wang
S.,
Yang
J.,
Li
J.
(
2010
)
The effects of palmitate on hepatic insulin resistance are mediated by NADPH oxidase 3-derived reactive oxygen species through JNK and p38MAPK pathways
.
J. Biol. Chem.
 
285,
29965
29973
Gillman
P. K.
(
2007
)
Tricyclic antidepressant pharmacology and therapeutic drug interactions updated
.
Br. J. Pharmacol.
 
151,
737
748
Golli-Bennour
E. E.,
Bouslimi
A.,
Zouaoui
O.,
Nouira
S.,
Achour
A.,
Bacha
H.
(
2012
)
Cytotoxicity effects of amiodarone on cultured cells
.
Exp. Toxicol. Pathol. 64, 425–430.
 
Grady
K. L.,
Costanzo
M. R.,
Fisher
S.,
Koch
D.
(
1996
)
Preoperative obesity is associated with decreased survival after heart transplantation
.
J. Heart Lung Transplant.
 
15,
863
871
Ito
K.,
Chiba
K.,
Horikawa
M.,
Ishigami
M.,
Mizuno
N.,
Aoki
J.,
Gotoh
Y.,
Iwatsubo
T.,
Kanamitsu
S.,
Kato
M.,
et al
(
2002
)
Which concentration of the inhibitor should be used to predict in vivo drug interactions from in vitro data?
AAPS PharmSci.
 
4,
E25
Kong
J. Y.,
Rabkin
S. W.
(
2000
)
Palmitate-induced apoptosis in cardiomyocytes is mediated through alterations in mitochondria: Prevention by cyclosporin A
.
Biochim. Biophys. Acta
 
1485,
45
55
Koshkin
V.,
Dai
F. F.,
Robson-Doucette
C. A.,
Chan
C. B.,
Wheeler
M. B.
(
2008
)
Limited mitochondrial permeabilization is an early manifestation of palmitate-induced lipotoxicity in pancreatic beta-cells
.
J. Biol. Chem.
 
283,
7936
7948
Lee
J. Y.,
Cho
H. K.,
Kwon
Y. H.
(
2010
)
Palmitate induces insulin resistance without significant intracellular triglyceride accumulation in HepG2 cells
.
Metab. Clin. Exp.
 
59,
927
934
Lee
W. M.
(
2003
)
Drug-induced hepatotoxicity
.
N. Engl. J. Med.
 
349,
474
485
Leem
J.,
Koh
E. H.
(
2012
)
Interaction between mitochondria and the endoplasmic reticulum: Implications for the pathogenesis of type 2 diabetes mellitus
.
Exp. Diabetes Res.
 
2012,
242984
Lemmi
C. A.,
Miller
R. L.,
Rajfer
J.
(
1990
)
Inhibition of fatty acid-supported mitochondrial respiration by cyclosporine
.
Biochem. Med. Metab. Biol.
 
44,
266
270
Levkovitz
Y.,
Ben-Shushan
G.,
Hershkovitz
A.,
Isaac
R.,
Gil-Ad
I.,
Shvartsman
D.,
Ronen
D.,
Weizman
A.,
Zick
Y.
(
2007
)
Antidepressants induce cellular insulin resistance by activation of IRS-1 kinases
.
Mol. Cell Neurosci.
 
36,
305
312
Luo
Y.,
Rana
P.,
Will
Y.
(
2012
)
Cyclosporine A and palmitic acid treatment synergistically induce cytotoxicity in HepG2 cells
.
Toxicol. Appl. Pharmacol.
 
261,
172
180
Malhi
H.,
Bronk
S. F.,
Werneburg
N. W.,
Gores
G. J.
(
2006
)
Free fatty acids induce JNK-dependent hepatocyte lipoapoptosis
.
J. Biol. Chem.
 
281,
12093
12101
Marroquin
L. D.,
Hynes
J.,
Dykens
J. A.,
Jamieson
J. D.,
Will
Y.
(
2007
)
Circumventing the Crabtree effect: Replacing media glucose with galactose increases susceptibility of HepG2 cells to mitochondrial toxicants
.
Toxicol. Sci.
 
97,
539
547
McDonald
M. G.,
Rettie
A. E.
(
2007
)
Sequential metabolism and bioactivation of the hepatotoxin benzbromarone: Formation of glutathione adducts from a catechol intermediate
.
Chem. Res. Toxicol.
 
20,
1833
1842
Morgan
R. E.,
Trauner
M.,
van Staden
C. J.,
Lee
P. H.,
Ramachandran
B.,
Eschenberg
M.,
Afshari
C. A.,
Qualls
C. W.,
Jr
Lightfoot-Dunn
R.,
Hamadeh
H. K.
(
2010
)
Interference with bile salt export pump function is a susceptibility factor for human liver injury in drug development
.
Toxicol. Sci.
 
118,
485
500
Naser
K. A.,
Gruber
A.,
Thomson
G. A.
(
2006
)
The emerging pandemic of obesity and diabetes: Are we doing enough to prevent a disaster?
Int. J. Clin. Pract.
 
60,
1093
1097
Pramyothin
P.,
Khaodhiar
L.
(
2010
)
Metabolic syndrome with the atypical antipsychotics
.
Curr. Opin. Endocrinol. Diabetes Obes.
 
17,
460
466
Raeder
M. B.,
Fernø
J.,
Vik-Mo
A. O.,
Steen
V. M.
(
2006
)
SREBP activation by antipsychotic- and antidepressant-drugs in cultured human liver cells: Relevance for metabolic side-effects?
Mol. Cell Biochem.
 
289,
167
173
Roth
R. A.,
Ganey
P. E.
(
2010
)
Intrinsic versus idiosyncratic drug-induced hepatotoxicity--two villains or one?
J. Pharmacol. Exp. Ther.
 
332,
692
697
Scaglione
R.,
Di Chiara
T.,
Cariello
T.,
Licata
G.
(
2010
)
Visceral obesity and metabolic syndrome: Two faces of the same medal?
Intern. Emerg. Med.
 
5,
111
119
Schloss
P.,
Williams
D. C.
(
1998
)
The serotonin transporter: A primary target for antidepressant drugs
.
J. Psychopharmacol.
 
12,
115
121
Stepan
A. F.,
Walker
D. P.,
Bauman
J.,
Price
D. A.,
Baillie
T. A.,
Kalgutkar
A. S.,
Aleo
M. D.
(
2011
)
Structural alert/reactive metabolite concept as applied in medicinal chemistry to mitigate the risk of idiosyncratic drug toxicity: A perspective based on the critical examination of trends in the top 200 drugs marketed in the United States
.
Chem. Res. Toxicol.
 
24,
1345
1410
Sun
Q.,
Zhu
R.,
Foss
F. W.,
Jr
Macdonald
T. L.
(
2008
)
In vitro metabolism of a model cyclopropylamine to reactive intermediate: Insights into trovafloxacin-induced hepatotoxicity
.
Chem. Res. Toxicol.
 
21,
711
719
Swiss
R.,
Will
Y.
(
2011
)
Assessment of mitochondrial toxicity in HepG2 cells cultured in high-glucose- or galactose-containing media
.
Curr. Protoc. Toxicol.
 
Chapter 2,
Unit2.20
Tabassum
H.,
Parvez
S.,
Pasha
S. T.,
Banerjee
B. D.,
Raisuddin
S.
(
2010
)
Protective effect of lipoic acid against methotrexate-induced oxidative stress in liver mitochondria
.
Food Chem. Toxicol.
 
48,
1973
1979
Ulrich
R. G.
(
2007
)
Idiosyncratic toxicity: A convergence of risk factors
.
Annu. Rev. Med.
 
58,
17
34
Wang
T.,
Shankar
K.,
Ronis
M. J.,
Mehendale
H. M.
(
2007
)
Mechanisms and outcomes of drug- and toxicant-induced liver toxicity in diabetes
.
Crit. Rev. Toxicol.
 
37,
413
459
Wolf
A.,
Trendelenburg
C. F.,
Diez-Fernandez
C.,
Prieto
P.,
Houy
S.,
Trommer
W. E.,
Cordier
A.
(
1997
)
Cyclosporine A-induced oxidative stress in rat hepatocytes
.
J. Pharmacol. Exp. Ther.
 
280,
1328
1334
Xia
Z.,
Ying
G.,
Hansson
A. L.,
Karlsson
H.,
Xie
Y.,
Bergstrand
A.,
DePierre
J. W.,
Nässberger
L.
(
2000
)
Antidepressant-induced lipidosis with special reference to tricyclic compounds
.
Prog. Neurobiol.
 
60,
501
512
Xu
J. J.,
Henstock
P. V.,
Dunn
M. C.,
Smith
A. R.,
Chabot
J. R.,
de Graaf
D.
(
2008
)
Cellular imaging predictions of clinical drug-induced liver injury
.
Toxicol. Sci.
 
105,
97
105
Yin
H. Q.,
Kim
M.,
Kim
J. H.,
Kong
G.,
Lee
M. O.,
Kang
K. S.,
Yoon
B. I.,
Kim
H. L.,
Lee
B. H.
(
2006
)
Hepatic gene expression profiling and lipid homeostasis in mice exposed to steatogenic drug, tetracycline
.
Toxicol. Sci.
 
94,
206
216
Zimmet
P.,
Alberti
K. G.,
Shaw
J.
(
2001
)
Global and societal implications of the diabetes epidemic
.
Nature
 
414
782
787