Mechanisms of benzarone and benzbromarone-induced hepatic toxicity


  • Conflict of interest: Nothing to report.


Treatment with benzarone or benzbromarone can be associated with hepatic injury. Both drugs share structural similarities with amiodarone, a well-known mitochondrial toxin. Therefore, we investigated the hepatotoxicity of benzarone and benzbromarone as well as the analogues benzofuran and 2-butylbenzofuran. In isolated rat hepatocytes, amiodarone, benzarone, and benzbromarone (20 μmol/L) decreased mitochondrial membrane potential by 23%, 54% or 81%, respectively. Benzofuran and 2-butylbenzofuran had no effect up to 100 μmol/L. In isolated rat liver mitochondria, amiodarone, benzarone, and benzbromarone, but not benzofuran, decreased state 3 oxidation and respiratory control ratios for L-glutamate (50% decrease of respiratory control ratio at [μmol/L]: amiodarone, 12.9; benzarone, 10.8; benzbromarone, <1). Amiodarone, benzarone, and benzbromarone, but not benzofuran, also uncoupled oxidative phosphorylation. Mitochondrial β-oxidation was decreased by 71%, 87%, and 58% with 100 μmol/L amiodarone or benzarone and 50 μmol/L benzbromarone, respectively, but was unaffected by benzofuran, whereas ketogenesis was not affected. 2-Butylbenzofuran weakly inhibited state 3 oxidation and β-oxidation only at 100 μmol/L. In the presence of 100 μmol/L amiodarone, benzarone or benzbromarone, reactive oxygen species production was increased, mitochondrial leakage of cytochrome c was induced in HepG2 cells, and permeability transition was induced in isolated rat liver mitochondria. At the same concentrations, amiodarone, benzarone, and benzbromarone induced apoptosis and necrosis of isolated rat hepatocytes. In conclusion, hepatotoxicity associated with amiodarone, benzarone, and benzbromarone can at least in part be explained by their mitochondrial toxicity and the subsequent induction of apoptosis and necrosis. Side chains attached to the furan moiety are necessary for rendering benzofuran hepatotoxic. (HEPATOLOGY 2005.)

Drug-induced hepatic injury is often regarded as a consequence of the formation and toxicity of reactive metabolites. In recent years, mitochondrial damage has been recognized as an additional important mechanism of drug-induced hepatotoxicity.1 Mitochondrial β-oxidation and oxidative phosphorylation are fundamental physiological processes. Acquired or inherited impairment of these processes can affect the function of many organs; in the liver, this typically leads to microvesicular steatosis, a potentially fatal disease.2, 3 Microvesicular steatosis has been observed in patients and/or animals being treated with the antiarrhythmic amiodarone.4 Amiodarone is composed of a di-iodobenzene ring carrying a diethylaminoethoxy side chain and a benzofuran ring carrying a C4H9 side chain (Fig. 1). Amiodarone is a well-characterized hepatic mitochondrial toxin. It inhibits enzyme complexes of the electron transport chain, impairs β-oxidation, and uncouples oxidative phosphorylation.4–6

Figure 1.

Chemical structures of the substances investigated.

Benzbromarone is another benzofuran derivative acting as an uricosuric agent by reducing the proximal tubular reabsorption of uric acid.7 Benzarone is a non–halogenated benzbromarone derivative used for the treatment of venous vascular disorders.8 As shown in Fig. 1, the structures of benzbromarone as well as benzarone are closely related to that of amiodarone. This structural relationship coupled with the fact that these two drugs can also cause hepatic injury,9 in some patients with fatal outcome,8, 10 led us to hypothesize that their mechanism of toxicity may be similar to that of amiodarone. In support of our hypothesis, it has been shown that benzbromarone can inhibit several enzyme complexes of the electron transport chain in isolated mouse liver mitochondria.11

Many recent studies have revealed that mitochondria not only provide energy for the cell, but can also unleash “machineries of death.”12 A variety of key events in cell death focus on mitochondria, in particular initiation of apoptosis and necrosis.12–14 Following these considerations, we decided to study the mechanisms of the suspected mitochondrial toxicity of benzarone and benzbromarone and to compare them with those of amiodarone. In particular, we were interested in mitochondrial mechanisms leading to cell damage or even death. We also investigated the structure–toxicity relationship by including the molecular analogues benzofuran and 2-butylbenzofuran along with amiodarone, benzarone, and benzbromarone in our studies.


DMSO, dimethylsulfoxide; RCR, respiratory control ratio; ATP, adenosine triphosphate; PPI, propidium iodide; ROS, reactive oxygen species; mpt, mitochondrial permeability transition.

Material and Methods


Benzarone was obtained from Norchim (Saint Leu d'Esserent, France), benzofuran was obtained from Merck (Darmstadt, Germany), and [1-14C]palmitic acid was obtained from Amersham (Dübendorf, Switzerland). 2-Butylbenzofuran was kindly provided by Dr. Riem Ha (University Hospital Zurich, Zurich, Switzerland). Fetal calf serum and all culture medium supplements were obtained from Gibco (Paisley, UK). Williams E media was obtained from Cambrex (Verviers, Belgium). The 96-well plates were purchased from Becton Dickinson (Franklin Lakes, NJ), the 8-chamber slides were obtained from Nalge Nunc (Rochester, NY), and unifilters and scintillation cocktail were obtained from Packard (Meriden, CT). The Vybrant Apoptosis Assay Kit #2 was purchased from Molecular Probes (Eugene, OR). All other chemicals used were of the best quality available from Sigma Aldrich (Schnelldorf, Germany).

Cell Lines

The HepG2 cell line was kindly provided by Professor Dietrich von Schweinitz (University Hospital Basel, Basel, Switzerland). The cell line was cultured in Dulbecco's Modified Eagle Medium supplemented with 10% (v/v) inactivated fetal calf serum, 10 mmol/L HEPES buffer (pH 7.4), 2 mmol/L GlutaMAX (Invitrogen, Basel, Switzerland), nonessential amino acids, and penicillin/streptomycin (100 U/mL). The culture conditions were 5% CO2 and 95% air atmosphere at 37°C.


Male Sprague-Dawley rats (Iffa Credo/Charles River, Les Onins, France) were used for all experiments. They were fed ad libitum and held on a 12-hour dark/light cycle. The study protocol was accepted by the Cantonal Animal Ethics Committee.

Isolation of Rat Liver Mitochondria

The rats weighed 314 ± 61 g at the time of sacrifice. The animals were anesthetized with carbon dioxide and decapitated. The liver (14.2 ± 1.9 g) was extirpated, rinsed, weighed, minced, and washed with ice-cold MSM buffer (220 mmol/L mannitol, 70 mmol/L sucrose, 5 mmol/L 3-[N-morpholino]propanesulfonic acid [pH 7.4]). All subsequent procedures were perfomed on ice. Mitochondria were isolated by differential centrifugation according to Hoppel et al.15 The final mitochondrial pellets were resuspended in MSM buffer and frozen at −70°C if not used fresh. The mitochondrial protein content was determined using the biuret method with bovine serum albumin as the standard.16

Isolation of Rat Hepatocytes

Hepatocytes were isolated from male rats weighing 200 to 250 g17 and suspended in Williams E medium (Cambrex, East Rutherford, NJ), supplemented with 10% heat-inactivated fetal bovine serum, 10 mmol/L HEPES buffer, 2 mmol/L GlutaMAX, 1,000 U/mL penicillin/streptomycin, 0.25 μg/mL amphotericin B, 4 μg/mL insulin, and 0.1 μmol/L dexamethasone. Culture conditions were an atmosphere of 5% CO2 and 95% air at 37°C. In all experiments, initial hepatocyte viability, determined by trypan blue exclusion, was more than 85%.

Mitochondrial Membrane Potential

The mitochondrial membrane potential was determined as described by Wan et al.18 Briefly, freshly isolated cells were washed twice with incubation buffer containing 137 mmol/L sodium chloride, 4.74 mmol/L potassium chloride, 2.56 mmol/L calcium chloride, 1.18 mmol/L potassium phosphate, 1.18 mmol/L magnesium chloride, 10 mmol/L HEPES, and 1 g/L glucose (pH 7.4). After labelling with 40 nmol/L [3H]-tetraphenylphosphonium bromide, cells were seeded in a 96-well plate and incubated with the test compounds for 1 hour at 37°C. The cells were harvested on a unifilter GF/B (Packard), which was mixed with 50 μL/well scintillation cocktail (Top Count; Packard), covered with a plate sealer, and counted for [3H]-radioactivity.

Oxygen Uptake

Oxygen uptake was monitored polarographically using a 1-mL chamber equipped with a Clark-type oxygen electrode (Yellow Springs Instruments, Yellow Springs, OH) at 30°C.15 The oxygen content of respiration buffer equilibrated with air was assumed to be 223 nmol O2/mL at 30°C.19 The final concentrations of the substrates used were 20 mmol/L for L-glutamate and succinate. The test compounds were dissolved in dimethylsulfoxide (DMSO) and added at concentrations given in Tables 1 and 2. Control experiments were performed in the presence of the solvent containing no inhibitor.

Table 1. Effects of the Test Compounds on Oxidative Metabolism by Isolated Rat Liver Mitochondria
 L-glutamate (20 mmol/L) [natoms O2 × min−1 × mg−1]Succinate (20 mmol/L) [natoms O2 × min−1 × mg−1]
State 3RCRState 3RCR
  • NOTE. Data are presented as the mean ± SEM of at least three experiments using different mitochondrial preparations.

  • Abbreviation: RCR, respiratory control ratio.

  • *

    P < .05 vs. 1% DMSO.

  • **

    P < .01 vs. 1% DMSO.

Control (no inhibitor)86 ± 36.5 ± 1.0163 ± 284.1 ± 0.4
Amiodarone (μmol/L)    
 1095 ± 223.0 ± 0.5**167 ± 442.4 ± 0.3**
 2081 ± 102.6 ± 0.0**170 ± 392.1 ± 0.3**
 5082 ± 91.7 ± 0.2**157 ± 491.6 ± 0.6**
 10034 ± 28**1.2 ± 0.3**135 ± 371.3 ± 0.5**
Benzarone (μmol/L)    
 1080 ± 343.6 ± 0.8167 ± 163.2 ± 0.1
 2065 ± 292.0 ± 0.5**161 ± 272.3 ± 0.5**
 5059 ± 281.0 ± 0.0**136 ± 181.0 ± 0.0**
 10030 ± 11**1.0 ± 0.1**64 ± 23**1.0 ± 0.0**
Benzbromarone (μmol/L)    
 0.1114 ± 126.3 ± 1.9190 ± 293.0 ± 1.2**
 1113 ± 192.0 ± 0.3**194 ± 21.9 ± 0.1**
 10105 ± 181.0 ± 0.0**121 ± 621.2 ± 0.3**
 10034 ± 34**1.0 ± 0.0**33 ± 7**1.0 ± 0.1**
Benzofuran (μmol/L)    
 1076 ± 115.9 ± 2.0161 ± 474.1 ± 0.5
 2078 ± 255.1 ± 1.0151 ± 443.7 ± 0.4
 5075 ± 155.9 ± 1.3138 ± 383.5 ± 0.7
 10072 ± 194.9 ± 1.0172 ± 163.5 ± 0.1
2-Butylbenzofuran (μmol/L)    
 10122 ± 535.4 ± 0.6200 ± 513.8 ± 0.6
 20114 ± 214.9 ± 1.0211 ± 74.4 ± 0.2
 50104 ± 144.7 ± 0.8*192 ± 104.4 ± 0.4
 10075 ± 53.3 ± 0.5**183 ± 102.2 ± 0.5**
Table 2. Effects of the Test Compounds on β-oxidation in Isolated Rat Liver Mitochondria
CompoundActivity [nmol × min−1 × mg−1]
  • NOTE. Data are presented as the mean ± SEM of at least three experiments using different mitochondrial preparations.

  • *

    P < .05 vs. 1% DMSO.

  • **

    P < .01 vs. 1% DMSO.

Amiodarone (μmol/L) 
 Control0.64 ± 0.10
100.48 ± 0.10
200.42 ± 0.06**
500.19 ± 0.04**
800.12 ± 0.02**
1000.15 ± 0.03**
Benzarone (μmol/L) 
 Control0.60 ± 0.10
50.70 ± 0.12
100.64 ± 0.11
200.34 ± 0.11**
500.24 ± 0.13**
1000.10 ± 0.05**
Benzbromarone (μmol/L) 
 Control0.61 ± 0.08
0.50.41 ± 0.10
20.30 ± 0.10*
50.24 ± 0.08**
100.19 ± 0.10**
500.26 ± 0.06**
Benzofuran (μmol/L) 
 Control0.75 ± 0.10
100.73 ± 0.10
200.70 ± 0.08
500.69 ± 0.09
800.73 ± 0.10
1000.72 ± 0.10
2-Butylbenzofuran (μmol/L) 
 Control0.62 ± 0.10
100.59 ± 0.10
200.55 ± 0.09
500.62 ± 0.11
800.53 ± 0.10
1000.45 ± 0.08**

In isolated mitochondria, state 3, state 4, and the respiratory control ratio (RCR) were determined as previously described20 and as defined by Estabrook.21 The test compounds were added to the mitochondrial incubations before the addition of the respective substrate.

For the determination of oxygen consumption by isolated hepatocytes, 1 × 106 cells were treated with oligomycin (final concentration 5 μg/mL) to inhibit mitochondrial ATPase. After 2 minutes, the test compounds were added to the incubation chamber and the oxygen consumption was determined as a marker for the uncoupling potential of the test compounds. Control experiments were performed with solvent only.

Mitochondrial β-oxidation and Formation of Ketone Bodies

β-Oxidation by freshly isolated liver mitochondria was assessed via the formation of 14C-acid–soluble β-oxidation products from [1-14C]palmitic acid in the presence of the test compounds. Experiments were performed as previously described22 with the modifications described by Spaniol et al.5

Ketone body formation by liver mitochondria was measured in the presence of an acetyl-CoA–generating system using freeze–thawed mitochondria according to Chapman et al.23 with the modifications described by Spaniol et al.5 The supernatants of the incubations were analyzed for acetoacetate according to Olsen24, using an enzyme-catalyzed reaction inducing changes in the concentration of β-nicotinamide adenine dinucleotide.

Activities of Mitochondrial β-oxidation Enzymes

All enzyme activities were determined using spectrophotometric assays at 37°C. Freeze-thawed mitochondria were treated 1:1 with 5% cholic acid to disrupt the mitochondrial membranes. The solution was then diluted 100 times with 50 mmol/L potassium phosphate buffer (pH 7.4). The effects on the enzymes of β-oxidation were only investigated with amiodarone, benzarone, benzbromarone, and 2-butylbenzofuran, because they had an inhibitory effect on β-oxidation. The activities of acyl-CoA dehydrogenase and β-ketothiolase were both determined spectrophotometrically as described by Hoppel et al.15

Reactive Oxygen Species

Confluent cultures of HepG2 cells, seeded in 96-well plates, were incubated with Dulbecco's Modified Eagle Medium without fetal bovine serum in the presence of 5 mmol/L 2,7-dichlorofluorescin diacetate. After incubation, the medium was replaced by phosphate-buffered saline and cellular fluorescence (λex = 485 nm, λem = 520 nm) was determined at room temperature using a microtiter plate reader (HTS 7000 Plus Bio Assay Reader; Perkin Elmer, Buckinghamshire, UK).

Mitochondrial Swelling

Measurements of mitochondrial swelling were performed using flow cytometry (FACScalibur, Becton Dickinson). The mitochondrial suspension (1 mg mitochondrial protein) was mixed with 200 μL of MSM buffer, and the test compounds were added subsequently. Because Ca2+ is a well-known inducer of swelling, the effect of 1 mmol/L CaCl2 was used as a positive control. Where indicated, 2 μL of cyclosporin A (final concentration: 2 μmol/L) was added as an inhibitor of mitochondrial swelling. For quantification purposes, mean forward scatter was determined.

Hepatocellular Adenosine Triphosphate Content

Freshly isolated hepatocytes (200,000 cells/well) were settled down in a 12-well plate (Becton Dickinson) and treated for 8 hours with test compounds. Following treatment, cells were washed with phosphate-buffered saline (pH 7.4), suspended in 1 mL of reagent grade water, snap-frozen in liquid nitrogen, and stored at −80°C. For the determination of adenosine triphosphate (ATP), cells were extracted with boiling water according to Yang et al.25 The concentration of ATP was determined in the supernatant of the cooled and centrifuged samples (12,000 rpm at 4°C for 5 min) using a commercial luciferin–luciferase reagent kit (FL-AA; Sigma, Deisenhofen, Germany). Values obtained were compared against an ATP standard curve, analyzed in duplicate at the same time as the unknown samples.

Apoptosis and Necrosis

Both assays were performed using freshly isolated rat hepatocytes cultured on poly-D-lysine coated (0.1 mg/mL, 30 min) glass slides. Soluble Fas-ligand (supernatant of Fas-ligand–transfected N2a cells) was used as a positive control and the supernatant of mock-transfected N2a cells as a negative control for apoptosis. Both supernatants were both a kind gift from Dr. Felix Bachmann (Aponetics AG, Witterswil, Switzerland). The detergent NP 40 was used as a positive control for necrosis (final concentration: 0.1%).

Hoechst 33342 Nuclear Staining.

Cells were treated for 8 hours with different concentrations of test compounds before they were incubated for 30 minutes at room temperature with Hoechst 33342 (50 μmol/L in phosphate-buffered saline) and visualized using fluorescence microscopy (Olympus IX 50, Hamburg, Germany).

Annexin V and Propidium Iodide Staining.

Cells were treated with the test compounds at the concentrations given in Tables 1 and 2. After incubation for 8 hours, cells were stained with 25 μL Annexin V-Alexa Fluor 488 and 2 μL propidium iodide (PPI) (final concentration: 1.5 μg/L) using an in situ Apoptosis Detection Kit (Vybrant Apoptosis Kit #2, Molecular Probes). After 15 minutes of incubation at room temperature, samples were analyzed via flow cytometry.

Mitochondrial Release of Cytochrome c

For immunocytochemistry, HepG2 cells were grown in an 8-chamber slide coated with poly-D-lysine for 12 hours at 37°C and then treated with the test compounds for 8 hours. Cytochrome c was visualized using an anti-cytochrome c antibody (Sigma, Buchs, Switzerland) and an anti-sheep immunoglobulin G antibody conjugated with Cy3 (Jackson Laboratories, West Grove, PA) according to the manufacturer's protocol.

Statistical Methods

Data are presented as the mean ± SEM. ANOVA was used for comparisons of more than 2 groups. A P value of .05 or less was considered significant. If ANOVA revealed significant differences, comparisons between the control and the other incubations were performed using Dunnett's posttest procedure. T tests (unpaired, two-tailed) were performed if only two groups were analyzed.


Mitochondrial Membrane Potential.

As shown in Fig. 2, increasing concentrations of amiodarone, benzbromarone, and benzarone resulted in a progressive decrease in the mitochondrial membrane potential. At 20 μmol/L, the mitochondrial membrane potential was 18% of the initial value for benzbromarone, 41% for benzarone, and 67% for amiodarone. In contrast, benzofuran and 2-butylbenzofuran did not affect the potential up to 100 μmol/L. Because these results confirmed that amiodarone, benzarone, and benzbromarone are mitochondrial toxins, their effect on mitochondria was characterized further.

Figure 2.

Effect on mitochondrial membrane potential. Rat hepatocytes were labeled with [3H]-tetraphenylphosphonium bromide, and mitochondrial accumulation of radioactivity was determined. 1% DMSO served as a control and was set at 100%. Data presented are the mean ± SEM of three individual preparations. *P < .05 versus control. **P < .01 versus control.

Oxidative Metabolism of Mitochondria.

The toxicity of these substances on the oxidative metabolism of isolated rat liver mitochondria was studied using L-glutamate or succinate as substrates (see Table 1). In the presence of L-glutamate or succinate, amiodarone, benzarone, and benzbromarone induced a progressive depression of the RCR. The corresponding concentrations associated with a 50% decrease in the RCR were 9.3 μmol/L for amiodarone and 11.2 μmol/L for benzarone (L-glutamate as substrate), and 23.9 μmol/L for amiodarone and 23.3 μmol/L for benzarone (succinate as substrate). For benzbromarone, this concentration was less than 1 μmol/L for both substrates, whereas 2-butylbenzofuran decreased the RCR less potently (50% decrease at 50–100 μmol/L) and benzofuran did not affect mitochondrial respiration at all. In contrast to the effects on the RCR, state 3 respiration, reflecting the activity of the respiratory chain, was decreased only at the highest concentration (100 μmol/L) of amiodarone, benzarone, and benzbromarone, but not by 2-butylbenzofuran or benzofuran. The RCR values were decreased mainly due to an increase in state 4, suggesting uncoupling of oxidative phosphorylation.

Uncoupling of oxidative phosphorylation was investigated directly using isolated hepatocytes in the presence of oligomycin to block mitochondrial conversion of ATP to adenosine diphosphate (Fig. 3). Benzbromarone increased oxygen consumption by hepatocytes starting at 10 μmol/L, showing that it is a potent uncoupler. For concentrations >20 μmol/L, the respiration rate was decreasing, reflecting progressive inhibition of the respiratory chain. Amiodarone stimulated respiration only at the highest concentration (100 μmol/L) and benzarone from 50 μmol/L upward, whereas benzofuran or 2-butylbenzofuran did not induce uncoupling up to 100 μmol/L. These results agree with the state 4 oxidation rates and demonstrate that amiodarone, benzarone, and benzbromarone are uncouplers.

Figure 3.

Uncoupling of oxidative phosphorylation using isolated, intact rat hepatocytes. In coupled mitochondria, blocking of the F0F1-ATPase by oligomycin results in a restricted electron transport similar to state 4. (A) Uncouplers increase oxygen consumption under these circumstances. Benzbromarone acts as an uncoupler at concentrations of 20 μmol/L or less. At higher concentrations, it also inhibits electron transport. (B) Respiration rates in the presence of amiodarone (1, 10, 20, 50, 80, 100 μmol/L), benzarone (1, 20, 50, 80, 100 μmol/L), benzbromarone (0.1, 1, 10, 20, 50, 80, 100 μmol/L), benzofuran (1, 10, 20, 50, 80, 100 μmol/L), and 2-butylbenzofuran (1, 10, 20, 50, 80, 100 μmol/L) are presented. 1% DMSO served as the control. Data are presented as the mean ± SEM of four individual mitochondrial preparations. **P < .01 versus 1% DMSO. ADP, adenosine diphosphate.

Mitochondrial β-oxidation and Formation of Ketone Bodies.

As has been shown for amiodarone,5 benzarone and benzbromarone also inhibited the formation of acid-soluble β-oxidation products in a dose-dependent manner (see Table 2). The corresponding IC50 values were 34 μmol/L for amiodarone, 34 μmol/L for benzarone, and 2 μmol/L for benzbromarone. In comparison, 2-butylbenzofuran was only a weak inhibitor (28% inhibition at 100 μmol/L), whereas benzofuran showed no inhibition up to 100 μmol/L. Because the formation of acid-soluble products reflects both β-oxidation and ketogenesis, ketogenesis was assessed directly. In contrast to the formation of acid-soluble products, ketogenesis was not affected by any of the substances investigated (data not shown).

To localize the inhibitory effect of β-oxidation more precisely, two enzymes of the β-oxidation cycle were investigated. In the presence of 100 μmol/L amiodarone, benzarone, benzbromarone, or 2-butylbenzofuran, acyl-CoA dehydrogenase activity was inhibited by 28%, 33%, 34%, or 22%, respectively, compared with control values. For β-ketothiolase, the last enzyme of the β-oxidation cycle, the average decrease in the presence of 100 μmol/L benzarone or benzbromarone was 11% and 25%, respectively, whereas amiodarone and 2-butylbenzofuran revealed no inhibitory effect. The discrepancy between inhibition of β-oxidation (determined using intact mitochondria) and individual enzymes of the β-oxidation cycle as well as ketogenesis (both determined using broken mitochondria) may be explained by an inhibition of the activation of palmitate and/or transport across the inner mitochondrial membrane. It is well established that CPT I, an enzyme involved in the import of long-chain fatty acids into the mitochondrial matrix, can be rate-limiting for hepatic fatty acid metabolism.26, 27

Production of Reactive Oxygen Species.

As a by-product of the formation of ATP via oxidative phosphorylation, mitochondria also produce reactive oxygen species (ROS).12, 28 Blockade of the electron flow through the electron transport chain stimulates ROS production, particularly in the presence of an uncoupler.29 In contrast, in the presence of an uncoupler only, the electron transport chain works more efficiently and leads to less leakage of electrons; therefore, lower levels of ROS are generated.30–32 ROS generation is claimed to play an important role in the induction of mitochondrial-mediated cell death.33–36 Because amiodarone, benzarone, or benzbromarone inhibit the electron transport chain and are uncouplers of oxidative phosphorylation (see Table 1), their effect on ROS generation was investigated in HepG2 cells using 2,7-dichlorofluorescin diacetate as an indicator.37 These investigations showed a concentration-dependent production of ROS in the presence of these substrates, which started to be detectable at concentrations of 0.1 nmol/L for benzbromarone, 10 nmol/L for benzarone, and 1 μmol/L for amiodarone (Fig. 4). In contrast, 2-butylbenzofuran and benzofuran were not associated with an increase in ROS production. Antimycin (1 μmol/L), an inhibitor of the electron transport from cytochrome b to ubiquinone, led to a significant increase in ROS production, whereas dinitrophenol (100 μmol/L), an uncoupler of oxidative phosphorylation, did not augment ROS generation. To ensure the specificity of the assay used, we blocked ROS generation by adding ascorbic acid at a concentration of 250 μmol/L38 to selected incubations (see Fig. 4).

Figure 4.

ROS formation by HepG2 cells. (A) Confluent monolayers of HepG2 cells were exposed to different compounds for 1 hour at 37°C (B) in the presence or absence of ascorbate (250 μmol/L). ROS formation was detected using 2,7-dichlorofluorescin diacetate (indicated as fluorescence units). Data are presented as the mean ± SEM of at least three independent experiments in quadruplicate. Asterisks indicate values significantly different from (A) control or (B) the incubation in the absence of ascorbate (*P < .05; **P < .01). FU, fluorescence units.

Mitochondrial Swelling.

Because ROS production can be associated with opening of the mitochondrial membrane pore,13 mitochondrial swelling was investigated as an indicator for mitochondrial permeability transition (mpt). As shown in Fig. 5, addition of Ca2+ ions (1 mmol/L) induced swelling that could be partially inhibited by the addition of cyclosporin A, indicating that swelling was dependent on mpt. A significant increase in mitochondrial size (at least partially inhibitable by cyclosporine A; results not shown) was also detectable for amiodarone, benzbromarone, and benzarone at 100 μmol/L (see Fig. 5).

Figure 5.

Mitochondria studied using flow cytometry. (A-D) Histograms (forward scatter vs. counts) of flow cytometry analyses of freshly isolated rat liver mitochondria. The forward scatter assesses the mitochondria's size and “counts” the number of mitochondria of this size. (E) Quantification of the mitochondrial volume as the mean forward scatter. Mitochondrial volume was measured in the (A) absence or presence of (B) Ca2+ (1 mmol/L) and (C) Ca2+/cyclosporin A (2 μmol/L) and with (D) test compounds (amiodarone 100 μmol/L). Ca2+, a known inducer of mitochondrial swelling, increases mitochondrial volume. Cyclosporine A partially blocks mitochondrial swelling by Ca2+. (E) Amiodarone, benzarone, and benzbromarone also induce swelling, but the swelling is less pronounced than that induced by Ca2+. Data are presented as the mean ± SEM of at least three independent experiments. *P < .05 versus 1% DMSO. **P < .01 versus 1% DMSO. FSC, forward scatter.

Apoptosis and Necrosis.

Swelling induced by mtp can lead to a disruption of the outer mitochondrial membrane with consecutive release of cytochrome c, which can activate caspases and induce apoptosis.13, 39 As shown in Fig. 6, 100 μmol/L amiodarone, benzarone, or benzbromarone were all associated with leakage of cytochrome c into the cytoplasm of HepG2 cells, whereas this was not the case for benzofuran or 2-butylbenzofuran.

Figure 6.

Mitochondrial cytochrome c release caused by treatment of HepG2 cells with test compounds (each 100 μmol/L) for 8 hours and detected via immunocytochemistry. In the presence of (A) 1% DMSO (control), (E) benzofuran, or (F) 2-butylbenzofuran, no leakage of cytochrome c into the cytoplasm occurred, as indicated by a diffuse, punctual pattern of immunoreactivity. In these incubations, immunoreactivity is linked to mitochondria (punctual pattern), and the nuclei are clearly visible in most cells. After treatment with (B) amiodarone, (C) benzbromarone, or (D) benzarone, cells may partially detach from the surface, and the immunoreactivity is spread in most cells over the entire cytoplasm. The homogenous, diffuse pattern of immunoreactivity, which may cover the nuclei, indicates release of cytochrome c from the mitochondrial intermembrane space followed by cytoplasmic spreading.

Chromatin fragmentation and/or condensation occurring during apoptosis can be visualized by fluorescence microscopy upon staining with dyes intercalating with DNA. As shown in Fig. 7, untreated cells showed an apoptosis rate of 3%. Apoptosis was induced by amiodarone, benzarone, and benzbromarone in a concentration-dependent manner (data not shown). At a concentration of 100 μmol/L, the proportion of apoptotic cells was 10% for amiodarone, 13% for benzarone, and 11% for benzbromarone. Fas-ligand, used as a positive control, induced DNA fragmentation in 9% of the cells, whereas 100 μmol/L benzofuran or 2-butylbenzofuran showed DNA fragmentation frequencies similar to control incubations.

Figure 7.

Illustration and quantification of apoptosis. Rat hepatocytes were treated with test compounds at a concentration of 100 μmol/L. After 8 hours of exposure, the cells were stained with Hoechst 33342 for 15 minutes and analyzed via fluorescence microscopy. Apoptotic cells are characterized by chromatin condensation (arrow a) and/or chromatin fragmentation (arrow b). (A) Untreated cells as control. (B) Benzarone 100 μmol/L. (C) Percentage of cells with apoptotic nuclei. Data are presented as the mean ± SEM of at least three independent experiments. *P < .05 versus 1% DMSO. **P < .01 versus 1% DMSO.

A hallmark of the early stages of apoptosis is the translocation of phosphatidylserine from the inner to the outer leaflet of the plasma membrane. Externalized phosphatidylserine can then be detected using Annexin V, a protein with a high affinity for this phospholipid. Cells were costained with PPI as a marker of cell membrane permeability, which increases during the later stages of apoptosis as well as necrosis. Because PPI only enters cells with already-disintegrated membranes, early apoptotic cells can be distinguished from late apoptotic and necrotic cells. Similar to staining with Hoechst 33342, flow cytometric analysis of the cells revealed a progressive decrease of viable, unstained cells with increasing concentrations of amiodarone, benzarone, or benzbromarone. At 100 μmol/L, only 38% of cells exposed to amiodarone, 53% of cells exposed to benzarone, and 11% of cells exposed to benzbromarone were viable (Fig. 8). These findings suggest that necrosis was initiated in addition to apoptosis, because the amount of cells undergoing cell death were considerably higher than the cells undergoing chromatin fragmentation and/or condensation.

Figure 8.

Annexin V binding and PPI uptake by rat hepatocytes. Hepatocytes were incubated for 8 hours at a concentration of 1 or 100 μmol/L of each test compound and subsequently stained with Annexin V and PPI. (A-D) The samples were analyzed via flow cytometry. Region R3 represents the unstained, thus viable cells. In region R4, the cells are undergoing early apoptosis and are stained only with Annexin V. Region R2 represents the double-stained cells undergoing necrosis or later stages of apoptosis. Under control conditions, the vast majority of cells are located in region R3 (A). (B) Fas-ligand induces apoptosis, whereas (C,D) benzbromarone increases the counts in region R2 in a concentration-dependent manner (C: 1 μmol/L; D: 100 μmol/L). (E) Percentage of cells undergoing necrosis or late stages of apoptosis relative to control values. NP 40 is a detergent used as a positive control for necrosis. Data are presented as the mean ± SEM of at least three independent experiments. *P < .05 versus 1% DMSO. **P < .01 versus 1% DMSO.

Because ATP is a prerequisite for apoptosis,40, 41 the cellular ATP content was determined in hepatocytes incubated for 8 hours with the compounds tested. The ATP content of control incubations was 4.0 ± 0.8 μmol/106 hepatocytes and remained stable in the presence of 1 or 100 μmol/L benzofurane or 2-butylbenzofuran. In the presence of 1 μmol/L amiodarone, benzarone, or benzbromarone, it was 93 ± 5%, 80 ± 20%, and 68 ± 15% of the control incubations, respectively. In the presence of 100 μmol/L amiodarone, benzarone, or benzbromarone, it dropped to 30 ± 13%, 68 ± 8%, and 18 ± 5% of the control incubations, respectively. These results give further support to the notion that apoptosis and necrosis occur under these conditions, in particular in the presence of benzbromarone.

If ROS generation were involved in the initiation of cell death, addition of an antioxidant, which was proven to reduce ROS production, should lead to a decrease of cell death. Therefore, the above-mentioned experiments (see Fig. 4) were repeated in the presence or absence of 250 μmol/L ascorbic acid. In the presence of ascorbate, the fraction of cells undergoing chromatin fragmentation and/or condensation was significantly reduced but not completely prevented (Fig. 9). Similar results were obtained with the Annexin V PPI staining (data not shown).

Figure 9.

Reduction of apoptosis in the presence of an antioxidant. In the presence of 250 μmol/L ascorbate, the fraction of cells undergoing chromatin condensation/fragmentation during incubation with test compounds is diminished. Data are presented as the mean ± SEM of at least three independent experiments. Asterisks indicate values significantly different from the incubations in absence of ascorbate (*P < .05; **P < .01). DMSO, dimethylsulfoxide.


Our investigations demonstrate that amiodarone, benzarone, and benzbromarone are inhibitors of the mitochondrial respiratory chain and β-oxidation and are uncouplers of oxidative phosphorylation. Furthermore, they can induce ROS production and mitochondrial swelling and lead to apoptosis and necrosis of cells.

The mitochondrial electron transport chain has been recognized as one of the major sources of reactive oxygen species, in particular complex I and complex III.42 Electrons passing across the electron transport chain can escape from this chain and can react with molecular oxygen to form superoxide (O2•−). Under normal conditions, superoxide is degraded by intramitochondrial, antioxidative systems such as glutathione and superoxide dismutase.33, 34 In the presence of uncouplers, protons are transported into the mitochondrial matrix, bypassing the F0F1-ATPase. During uncoupling, the respiratory chain works more efficiently to re-establish the proton gradient, and therefore, ROS production is usually not increased or can even be reduced.30 In the presence of inhibitors of the respiratory chain, generation of ROS is increased, because electrons may escape from the electron transport chain and react with molecular oxygen. Indeed, inhibitors of the electron transport chain (e.g., rotenone or antimycin) have been shown to increase ROS generation in mitochondria.30, 42, 43 In this context, it is interesting to note that amiodarone, benzarone, and benzbromarone are both uncouplers and inhibitors of the electron transport chain. Because inhibition of the electron transport chain may generate a higher amount of ROS in the presence of an uncoupler,42 such substances may be particularly toxic for mitochondria.

It is well known that ROS within the mitochondrial matrix can trigger mpt pore opening.13 An increase in mpt is an important route through which mitochondrial toxins can activate cell death in mammalian cells.13, 44 ROS generation and mpt induction have been identified as possible common effectors of cell death by apoptotic as well as by necrotic stimuli.45, 46 Pore opening induces release of cytochrome c, which is located on the outer surface of the inner mitochondrial membrane (see Fig. 6). After release, cytochrome c triggers the subsequent effector steps for apoptosis, in particular caspase activation. ATP is critical for the action of cytochrome c, because it is required for Apaf-1 oligomerization, which is followed by caspase activation. Low cellular ATP concentrations, as observed in hepatocytes incubated with 100 μmol/L amiodarone, benzbromarone, or benzarone, can therefore switch apoptotic stimuli toward cell necrosis. Depending on the intensity of the oxidative stimuli and the extent to which generation of ATP is impaired (e.g., by inhibition of the electron transport chain, uncoupling or impairment of β-oxidation), ROS generation can therefore be associated with both apoptosis and necrosis. Further evidence for the above-mentioned hypothesis that ROS generation works as a trigger for cell death induction was provided by demonstrating a reduction of cell death in the presence of ascorbate.

Besides their effects on oxidative phosphorylation and mitochondrial electron transport, amiodarone, benzarone, and benzbromarone are efficient inhibitors of mitochondrial β-oxidation. The inhibition of β-oxidation is independent from the inhibition of the electron transport chain, because the corresponding concentrations for 50% inhibition are clearly lower for β-oxidation than for state 3 oxidation. After depletion of glycogen (e.g., during starvation), the liver depends mainly on β-oxidation for the production of energy. It is therefore likely that under such circumstances, and in the presence of inhibitors of mitochondrial β-oxidation, the cellular ATP content may decrease, possibly leading to hepatocyte necrosis.

Concerning the structure–toxicity relationship, our investigations demonstrate clearly that the benzofuran structure per se, which was assumed to mediate mitochondrial toxicity,5 is not hepatotoxic. Our studies show that the hepatotoxicity of benzofuran is dependent on the presence of sidechains in position 1 and/or 2 of the furan ring. Bromide atoms in the p-hydroxybenzene moiety are not essential for hepatotoxicity associated with these compounds, but they clearly enhance it.

In conclusion, benzarone, benzbromarone, and amiodarone are toxic to isolated rat liver mitochondria as well as whole hepatocytes. The benzofuran structure alone is not responsible for the hepatotoxic effects—sidechains at the furan ring are also necessary. Hepatic injury associated with the ingestion of these drugs can be explained by mitochondrial damage with subsequent induction of apoptosis and necrosis.