Bezafibrate Induces a Mitochondrial Derangement in Human Cell Lines

Bezafibrate is a hypolipidemic drug that belongs to the group of peroxisome ... In this view, the human acute promyelocytic leukemia HL-60 cell line a...
0 downloads 0 Views 408KB Size
1440

Chem. Res. Toxicol. 2003, 16, 1440-1447

Bezafibrate Induces a Mitochondrial Derangement in Human Cell Lines: A PPAR-Independent Mechanism for a Peroxisome Proliferator R. Scatena,*,† P. Bottoni,† F. Vincenzoni,† I. Messana,‡ G. E. Martorana,† G. Nocca,† P. De Sole,† N. Maggiano,§ M. Castagnola,† and B. Giardina† Istituto di Biochimica e Biochimica Clinica, Universita’ Cattolica del Sacro Cuore, Rome, Italy, Dipartimento di Scienze Applicate ai Biosistemi, Universita’ di Cagliari, Cagliari, Italy, and Istituto di Anatomia Patologica, Universita’ Cattolica del Sacro Cuore, Rome, Italy Received May 28, 2003

Bezafibrate is a hypolipidemic drug that belongs to the group of peroxisome proliferators because it binds to peroxisome proliferator-activated receptors type R (PPARs). Peroxisome proliferators produce a myriad of extraperoxisomal effects, which are not necessarily dependent on their interaction with PPARs. An investigation on the peculiar activities of bezafibrate could clarify some of the molecular events and the relationship with the biochemical and pharmacological properties of this class of compounds. In this view, the human acute promyelocytic leukemia HL-60 cell line and human rabdomiosarcoma TE-671 cell line were cultured in media containing bezafibrate and a number of observations such as spectrophotometric analysis of mitochondrial respiratory chain enzymes, NMR metabolite determinations, phosphofructokinase enzymatic analysis, and differentiation assays were carried on. Bezafibrate induced a derangement of NADH cytochrome c reductase activity accompanied by metabolic alterations, mainly a shift to anaerobic glycolysis and an increase of fatty acid oxidation, as shown by NMR analysis of culture supernatants where acetate, lactate, and alanine levels increased. On the whole, the present results suggest a biochemical profile and a therapeutic role of this class of PPARs ligands more complex than those previously proposed.

Introduction Bezafibrate, an aryloxyalkanoic fibrate (like clofibric acid, fenofibrate, etc.), is a well-known antihyperlipoproteinemic drug that also affects glucose metabolism. The latter effect has been generally considered a metabolic consequence of the hypolipidemic action (1, 2). From a biochemical viewpoint, these therapeutic properties have been related to the liganding activity with peroxisome proliferator-activated receptors (PPARs) (3-5). These receptors are a group of lipid-activated nuclear receptors that regulate transcription of genes mainly involved in lipid metabolism and that are now attracting the attention for their different therapeutic potentials in cardiovascular, endocrine, immunological, and proliferative disorders (2, 4, 5). Some contradictory aspects concerning the pathophysiology of PPARs in general and the pharmacological role of their synthetic ligands are also coming to light. For example, pharmacological activities of fibrates cannot be regarded exclusively as derived directly from their PPAR liganding capacity since they possess also other intriguing (less known and/or considered) biochemical properties such as (i) a strong allosteric effect on human hemoglobin * To whom correspondence should be addressed. Tel: +390630154215. Fax: +39063053598. E-mail: [email protected]. † Istituto di Biochimica e Biochimica Clinica, Universita’ Cattolica del Sacro Cuore. ‡ Dipartimento di Scienze Applicate ai Biosistemi, Universita’ di Cagliari. § Istituto di Anatomia Patologica, Universita’ Cattolica del Sacro Cuore.

in vitro and in vivo (6, 7); (ii) an enhancement of the oxidative burst of leukocytes (8); (iii) an auxinic and herbicide-like activity (9); and (iv) a debated inhibitory activity of mitochondrial energetics in vitro and in vivo (10-12). Moreover, the following points have to be taken in consideration. (i) Synthetic PPARs ligands are a very heterogeneous class of amphipatic chemicals (2-4), which display very similar effects. (ii) PPAR-R and -γ induce the expression of different gene families related to lipid metabolism (catabolism or storage of fatty acid, respectively). Despite their different physiological roles, synthetic ligands for PPAR-R or -γ actually cause similar therapeutic effects in vivo (hypolipidemic and hypoglycemic effects) (4, 5, 10-12). (iii) A great number of studies demonstrated that synthetic ligands for PPAR-R and -γ display both a differentiating activity and a carcinogenic role on experimental models (13-17). Paradoxically, these activities, much alike as their metabolic effects are regarded, do not strictly depend on cell or tissue PPAR isotype expression (4, 5). These findings, together with the pharmacotoxicological profile of fibrates, clearly indicate the existence of other biochemical activities related to energy metabolism and, probably, independent from a direct PPAR activation. On this basis, we hypothesized that bezafibrate, causing a selective derangement of the mitochondrial respiratory chain, may induce a series of metabolic compensatory mechanisms, which, apart from PPAR activity, could be responsible of some pharmacological and toxicological properties of this class of molecules.

10.1021/tx0341052 CCC: $25.00 © 2003 American Chemical Society Published on Web 09/20/2003

Bezafibrate and Mitochondria

In this paper, to investigate fibrate-induced mitochondrial derangements and the related metabolic adaptations together with the differentiating and cancerogenic activities of this class of drugs, the effects of bezafibrate on different human cell lines have been studied.

Experimental Procedures Cells and Treatments. Human acute promyelocytic leukemia HL-60 cell line and human rabdomiosarcoma TE-671 cell lines were kindly supplied by Prof. A. Cittadini (Institute of General Pathology, Faculty of Medicine, Catholic University, Rome, Italy). Cells were maintained at 37 °C under a humidified atmosphere of 5% CO2 in RPMI 1640 Hepes-modified medium supplemented with 10% (v/v) heat-inactivated fetal calf serum, 2 mM glutamine, 100 IU/mL penicillin, and 100 µg/mL streptomycin. Cell number was determined using a Neubauer hemocytometer, and viability was assessed by their ability to exclude trypan blue. Cytotoxic activity determined by microscopic assay was confirmed in an alternative assay by monitoring LDH release in the medium, measured with a Hitachi 917 automated analyzer (Roche Diagnostics, Switzerland) and appropriate reagent kits (GOT-POD, Sigma Italy). Electron microscopy analysis was realized according to the procedure of Maunsbach (18). Materials and Reagents. Unless otherwise indicated, all chemicals and reagents (cell culture grade) were obtained from Sigma Chemical Co., Milan, Italy. Protein content was determined by Biorad Protein Assay (Biorad). Drug Treatment. The stock solutions were prepared immediately before use. Bezafibrate, gemfibrozil, clofibric acid, and ciglitizone were dissolved in DMSO. In metabolic experiments, HL-60 and TE-671 exponentially growing cells (3 × 105 cells/ mL and 3 × 104 cells/cm2, respectively) were incubated in media containing various concentrations of drugs for 96 h. The final concentration of the drugs were as follows: bezafibrate, 1 × 10-4, 2.5 × 10-4, 5 × 10-4, and 1 × 10-3 M; gemfibrozil, 5.6 × 10-5, 1.12 × 10-4, and 2.3 × 10-4 M; clofibric acid, 1.1 × 10-4, 2.23 × 10-4, and 7.00 × 10-4 M; and ciglitizone, 1 × 10-3, 5 × 10-3, and 5 × 10-2 mM, respectively. Considering bezafibrate concentrations, some literature data seem to indicate that the Cmax of the drug, at the therapeutic doses usually adopted, could be around 1 × 10-4 M (1). However, successive pharmacological studies clearly showed that in particular clinical conditions (i.e., elderly patients), maximum plasma concentrations can be much higher than those recorded in younger healthy subjects (7, 8, 16). For these considerations only, 1 × 10-3 M is well above the therapeutic range and was experimentally employed here to elicit a more evident effect. The final concentration of DMSO, used as vehicle, was the same in all samples during the experiments (0.1% v/v). Mitochondrial Respiratory Chain Enzyme Assay. NADHcytochrome c reductase (complex I-III), succinate-cytochrome c reductase (complex II-III), and cytochrome c oxidase (complex IV) were analyzed by spectrophotometric assay on digitoninpermeabilized whole cells using cytochrome c as electron donor/ acceptor, according to the method of O’Donnell et al. (19). In this set of experiments, bezafibrate was added just before the enzymatic analysis. These activities were measured as nmol of cytochrome c reduced/min/106 cells. Succinate dehydrogenase (complex II) activity was determined on whole cell extracts and measured as nmol of 2,6-dichlorophenol-indophenol (DCPIP) reduced/min/106 cells, using a modified protocol of Robinson et al. (20). These data were expressed as ratio of enzyme activity of treated samples vs controls. NMR Metabolite Determinations. At different incubation times, supernatant (3 mL) was mixed with an equal volume of cold 12% (v/v) HClO4 solution. Denatured material was centrifuged for 10 min at 2000g, and the upper solution was neutralized with potassium carbonate, lyophilized, and then dissolved in 0.7 mL of D2O containing 0.75% sodium-3-(trimethylsilyl)[2,2,3,3-2H4]-1-propionate (TSP, used as reference peak, 0.00

Chem. Res. Toxicol., Vol. 16, No. 11, 2003 1441 ppm) for 1H NMR analysis. In 1H NMR spectra, registered at 25 °C by a Gemini 300 spectrometer (Varian, Palo Alto, CA), signals of lactate, alanine, and acetate methyl groups were evident (1.33, 1.42, and 1.92 ppm, respectively). The concentration of the three metabolites was determined on the basis of the peak area of the corresponding methyl groups normalized with respect to the area of the signal of TSP, whose concentration was known. A set of NMR determinations was performed incubating cells in a medium containing 5 mmol/L [2-13C]glucose to follow the course of labeled carbon atom through metabolic pathways. This procedure permitted the time course monitoring of metabolite production in forms of the major isotopomer. In particular, the determination of labeled acetate with respect to unlabeled acetate allows us to evaluate glycolytic acetate with respect to nonglycolytic acetate. The same experiments permit us to evaluate the pentose phosphate pathway flux, utilizing the method of Schrader et al. (21) that is based on the experimental determination of the ratio between the NMR signals of 13C-3 and 13C-2 of lactate. Glucose consumption was measured on perchloric acid extracts by following the time course of 13C-MR signals of R and β anomers of labeled glucose. To normalize glucose consumption to different proliferation rates of cells, data were expressed as AUC (area under curve) vs AUC of cell growth during 96 h of culture. Determination of Phosphofructokinase (PFK) Enzyme Activity. Enzyme measurements were determined from whole cell extracts. Cell pellets obtained after centrifugation at 1200 rpm for 5 min were washed in PBS solution and suspended in a lysis buffer at a density of 2 × 107 cells/mL and sonicated twice for 30 s at 20 W with a sonifier model VC50 Vibracell (Sonics & Materials Inc., U.S.A.) according to the method of Brambilla et al. (22). PFK activity was determined according to Beutler (23) following NADH oxidation at 340 nm. Briefly, PFK analysis was carried out in 0.1 mol/L Tris-HCl, 0.5 mmol/L EDTA (pH 8.0), 10 mmol/L MgCl2, 0.2 mmol/L NADH, 2 mmol/L ATP, and fructose 6-phosphate at saturating concentrations (2 mmol/L). Determinations were performed on cellular lysates obtained after 96 h of drug treatment. The samples were incubated for 15 min at 37 °C in the dark before the enzymatic assay. This procedure was performed using a Cary 3E UVvisible spectrophotometer (Varian) at 37 °C. Values were expressed as µmol NADH oxidated/min/mg protein. Differentiation Assay. Phorbol-12-myristate-13-acetate (PMA) and zymosan-stimulated reactive oxygen species (ROS) metabolism, mostly due to the activity of NADPH oxidase system, were adopted as a marker of differentiation in human myeloid cell line HL-60. ROS metabolism was studied by a chemiluminescence (CL) assay as already described (24). Luminol, the chemiluminigenic probe, and PMA stock solutions were prepared at 50 and 3 mmol/L, respectively, in DMSO. The working solution, a modified Krebs Ringer phosphate (KRP) medium containing 119 mmol/L NaCl, 4.8 mmol/L KCl, 1.2 mmol/L MgSO4, 5.5 mmol/L glucose, 0.25 mmol/L CaCl2, and 16.6 mmol/L phosphate buffer (pH 7.4), was prepared immediately before use. Zymosan was prepared and opsonized as described previously (25). Briefly, 25 mg of zymosan was incubated for 20 min at 37 °C in 1.0 mL of fresh plasma obtained from a pool of at least 10 blood donors. After three washings, the opsonified zymosan was diluted to 5 mg/mL. Each sample was stimulated with PMA and/or opsonized zymosan. ROS production was measured by luminol amplified luminescence. Assays were performed in triplicate in an automatic luminometer (Autolumat LB 953, EG&G, Turku, Finland) at 25 °C for 120 min with cycles of 5 min each. The CL system contained 1 × 105 cells, without treatment (control) or treated with 0.1% DMSO (vehicle alone), or cells treated with different differentiating factors, 100 nmol of luminol, and 0.5 mg of opsonized zymosan or 150 pmol of PMA in a 1.0 mL final volume with KRP solution. Unstimulated activity was measured without addition of any stimulus.

1442 Chem. Res. Toxicol., Vol. 16, No. 11, 2003

Scatena et al.

Figure 1. Bezafibrate inhibited activity of NADH-cytochrome c reductase in digitonin-permeabilized HL-60 and TE-671. The inhibitory effect was greater with myeloid cells than with muscular cells. Results are expressed as the mean ( SEM, calculated from four experiments, each performed in duplicate. The group means were compared by ANOVA followed by a multiple comparison of means by Student-Newman-Keuls. p < 0.05 was considered significant; *** ) p < 0.001. Stimulated CL was evaluated by determination of a stimulation index:

areastimulated cells (expressed as counts/cell/120 min)/ areaunstimulated cells (expressed as counts/cell/120 min)

Results

Figure 2. NMR metabolite determinations in supernatant of HL-60 cells cultured for 96 h in the presence of bezafibrate. A dose-dependent increase of lactate, acetate, and alanin pointed out a compensatory shift to anaerobic metabolism. Results are expressed as the mean ( SEM, calculated from four experiments, each performed in duplicate. The group means were compared by ANOVA followed by a multiple comparison of means by Student-Newman-Keuls. p < 0.05 was considered significant; *** ) p < 0.001, * ) p < 0.05.

Mitochondrial Respiratory Chain Enzyme Activities. To determine whether fibrates influence mitochondrial respiration, experiments were performed in HL-60 (human leukemia cell line) and TE-671 (human rabdomyosarcoma cell line) using bezafibrate, which is one of the most used fibrates with PPARR-β binding activity (2, 4). Results from spectrophotometric measurements on mitochondrial respiratory chain enzyme activity in digitonin-permealized cells demonstrated that bezafibrate induced a significant dose-dependent reduction of NADH-cytochrome c reductase activity in both HL-60 and TE-671 cell lines. Interestingly, HL-60 cell lines seemed to show a greater sensitivity to bezafibrate than TE-671 cells. This different inhibitory activity could be explained, at least in part, by the lower number of mithocondria per cell, which characterizes the myeloid cell line (Figure 1). As shown by the figure, within the range of drug concentrations used, we did not achieve a total inhibition of enzymatic activity. Moreover, it could be worthwhile to mention that the activities of succinate dehydrogenase, succinate-cytochrome c reductase, and cytochrome oxidase did not show any significant modification induced by the drug in both cell lines (data not shown). NMR Metabolites Determinations. Drug-treated cell cultures do modify their metabolism, as showed by

NMR spectroscopy analysis of culture medium. In particular, HL-60 cell lines cultured with bezafibrate showed a significant dose-dependent increase of lactate, acetate, and alanine production with respect to controls during 96 h of culture. These increments could depend on the perturbation of mitochondrial oxidative metabolism, causing a compensatory shift to anaerobic metabolism. Interestingly, these experiments seemed to show that acetate metabolism is modified earlier by drug concentrations, which did not yet influence lactate and alanine levels (Figure 2). In other words, to obtain a significant perturbation of lactate and alanine metabolisms, higher concentrations of the drug are needed. These data were confirmed by time course NMR determinations in the supernatant of HL-60 cells and TE671 cells cultured for 96 h in the presence of bezafibrate (1 and 0.5 mmol/L, respectively). In both cell lines, this drug induced a time-dependent increase of lactate levels. Interestingly, in TE-671 cells, lactate production was already significantly increased after 48 h of culture (Figure 3). Moreover, a number of experiments have been performed in culture medium containing 5 mmol/L [2-13C]glucose in order to follow the time course of specifically labeled atoms. The results did not show any

Statistical Analysis. All results are expressed as means ( SEM. The group means were compared by ANOVA followed by a multiple comparison of means with the Newman-Keuls test. p < 0.05 was considered significant.

Bezafibrate and Mitochondria

Figure 3. Time course NMR lactate determinations in supernatant of HL-60 cells and TE-671 cells cultured for 96 h in the presence of bezafibrate. Cells were incubated with medium containing 5 mmol/L 13C-glucose, in the presence of bezafibrate (1 mmol/L for HL-60 and 0.5 mmol/L for TE-671) or vehicle, for 96 h. NMR spectra of supernatants allowed us to monitor production of lactate and its isotopomers. Total lactate production significantly increased in bezafibrate cells (9) with respect to control (0). This increment resulted earlier and higher in TE671 than in the HL-60 cell line. Data are expressed as the mean ( SEM, calculated from four experiments, each performed in duplicate. The group means were compared by ANOVA followed by a multiple comparison of means by Student-NewmanKeuls. p < 0.05 was considered significant; *** ) p < 0.001, ** ) p < 0.01.

significant modification of the ratio between the NMR signals of 13C-3 and 13C-2 of lactate, indicating that glucose partition between glycolysis and pentose phosphate pathway is not perturbed (data not shown). NMR analysis of acetate levels showed that this metabolite rapidly increased in bezafibrate-treated cultures in both HL-60 cells and TE-671 cells. According to its metabolic sources, which link acetate to acetyl-CoA hydrolysis (26, 27), acetate levels increased faster and greater than lactate in both HL-60 and TE-671 cell lines. Interestingly, NMR analysis of [2-13C]glucose showed that acetate derived from both glycolytic and nonglycolytic pathways increased in a similar way in drug-treated cultures (Figures 4 and 5). The alteration of mitochondrial respiratory activity with related compensatory metabolic mechanisms should determine an elevated glycolytic flux and a significant increase of glucose consumption per single cell. NMR analysis of glucose utilization in cell culture medium, related to cell growth kinetics, showed that bezafibratetreated cells did metabolize more glucose than control cells in both HL-60 and TE-671 cell lines. This result seemed to show that bezafibrate-treated cells are metabolically more active than control cells despite different cell growth kinetics, at least as glucose metabolism is regarded (Figure 6). This significant metabolic perturbation was further confirmed by the observation of surprising stimulation of PFK activity, a key regulatory enzyme of the glycolytic pathway. Both HL-60 and TE-671 cell lines cultured

Chem. Res. Toxicol., Vol. 16, No. 11, 2003 1443

Figure 4. Drug induced a time-dependent increment of acetate in culture medium of the HL-60 cell line. Cells were incubated with medium containing 5 mmol/L 13C-glucose, in the presence of bezafibrate (1 mmol/L, 9) or vehicle (0), for 96 h. NMR spectra of supernatants allowed us to monitor production of acetate from glycolytic and nonglycolytic pathways. Results are expressed as the mean ( SEM, calculated from four experiments, each performed in duplicate. The group means were compared by ANOVA followed by a multiple comparison of means by StudentNewman-Keuls. p < 0.05 was considered significant; *** ) p < 0.001.

in the presence of bezafibrate 1 mmol/L for 96 h showed a slight but significant rise of enzyme activity (Figure 7). At last, electron microscopy analysis of HL-60 cells, cultured for 96 h in the presence of bezafibrate (1 mmol/ L), confirmed fibrate-induced mitochondrial damage. Cells clearly showed some morphological alterations and, above all, a patchlike condensation and breaking of mitochondria with respect to controls (Figure 8). Other Synthetic PPAR Ligands and Related Compounds. The same set of experiments were realized in the HL-60 cell line using other PPAR ligands (clofibric acid, ciglitizone) and a related compound (gemfibrozil). Interestingly, all of these compounds, even though with different potency, afforded similar effects in terms of inhibition of cell growth, differentiation activity, and, above all, impairment of mitochondrial NADH-cytochrome c reductase activity. Importantly, once activities are expressed on a molar basis, it is evident that mitochondrial toxicity was positively correlated with both the differentiation capacity and the antiproliferative effect of these drugs (Table 1).

1444 Chem. Res. Toxicol., Vol. 16, No. 11, 2003

Scatena et al.

Figure 6. Cell cultured in the presence of bezafibrate consumed more glucose than controls although different cell growth kinetics. HL-60 and TE-671 cultured in the presence of bezafibrate, 1 and 0.5 mmol/L, respectively (9), were shown to significantly increase glucose utilization with respect to basal levels consumed (0). Results are expressed as the mean ( SEM, calculated from four experiments, each performed in duplicate. The group means were compared by ANOVA followed by a multiple comparison of means by Student-Newman-Keuls. p < 0.05 was considered significant; *** ) p < 0.001, ** ) p < 0.01. Figure 5. Bezafibrate induced a time-dependent increment of acetate in culture medium of the TE-671 cell line. Cells were incubated with medium containing 5 mmol/L 13C-glucose, in the presence of bezafibrate (0.5 mmol/L, 9) or vehicle (0), for 96 h. NMR spectra of supernatants allowed us to monitor production of acetate from glycolitic and nonglycolytic pathways. Results are expressed as the mean ( SEM, calculated from four experiments, each performed in duplicate. The group means were compared by ANOVA followed by a multiple comparison of means by Student-Newman-Keuls. p < 0.05 was considered significant; *** ) p < 0.001, ** ) p < 0.01, * ) p < 0.05.

Discussion Previous studies have demonstrated that synthetic peroxisome proliferators do not only interact with PPARs but also present other intriguing biological activities (4, 5, 12-17). At this regard, the interaction of fibrates with human hemoglobin is particularly indicative. It is wellknown in fact that bezafibrate, clofibrate, and related compounds such as gemfibrozil may act as allosteric effectors, binding the Hb tetramer in the cavity between the two R-chains (28, 29). This effect, among others, could be suggestive of the molecular mechanisms underlying the biochemical reactivity of this class of compounds. In other words, by analogy with what is known on human Hb, many other proteins may be functionally modified by these compounds. Moreover, it should be worthy of note that bezafibrate and other peroxisome proliferators seem also to affect mitochondrial function. Such a damage could be well at the basis of some of the pharmacotoxicological effects observed during in vivo treatment of PPARs ligands in general and of fibrates in particular. In this study, bezafibrate has been shown to inhibit mitochondrial respiratory chain at the level of NADHcytochrome c reductase activity in a dose-dependent

Figure 7. Cell cultured in the presence of bezafibrate showed an intriguing increment of PFK activity. HL-60 and TE-671 cultured in the presence of bezafibrate, 1 and 0.5 mmol/L, respectively (9), for 96 h showed a significant increase of enzymatic activity of PFK with respect to control cells (0). Results are expressed as the mean ( SEM, calculated from three experiments, each performed in duplicate. The group means were compared by ANOVA followed by a multiple comparison of means by Student-Newman-Keuls. p < 0.05 was considered significant; ** ) p < 0.01.

fashion. This effect, compromising the oxidative metabolism, pushes both HL-60 and TE-671 cell lines toward a series of complex compensatory adaptations, as shown by the rise of acetate and lactate levels. As far as acetate is concerned, the dramatic increment of this metabolite in the supernatant of bezafibrate-treated cells has been generally considered as a result of PPAR-mediated stimulation of peroxisomal and mitochondrial β-oxidation (4, 5). However, in our opinion, it could also represent a direct consequence of the energy metabolism derangement induced by fibrates. In particular, fibrate-induced inhibition of NADH oxidation, impairing Krebs cycle, causes both the increase in acetate production and the compensatory switch of the respiratory mitochondrial chain preferentially toward FADH oxidation (complex II). In this respect, it could be useful to consider that in

Bezafibrate and Mitochondria

Chem. Res. Toxicol., Vol. 16, No. 11, 2003 1445 Table 1. Quantitative Comparison, on a Molar Ratio Basis (Drug Activity/[Drug] × 100), between Different PPARs Ligands and Gemfibrozil in Terms of Inhibition of NADH Cytochrome c Reductase Activity, Differentiation Effect, and Antiproliferative Index on HL-60 Cell Linea antimitochondria proliferative damage index index bezafibrate clofibric acid gemfibrozil ciglitizone

58 ( 6 47 ( 2 164 ( 8 640 ( 83**

45 ( 4.3 27.1 ( 3.9 109 ( 6.7 940 ( 72** r2 ) 0.9993 p ) 0.0004

differentiation activity (PMA/zymosan) 14.8 ( 2.9/11.9 ( 2.1 2.8 ( 0.3/3.6 ( 0.9 12.7 ( 2.2/10.6 ( 0.9 164 ( 17*** - 238 ( 21*** r2 ) 0.9924-0.9912 p )0.0038-0.0044

a Results are expressed as the mean ( SEM, calculated from four experiments, each performed in duplicate. The group means were compared by ANOVA followed by a multiple comparison of means by Student-Newman-Keuls. p < 0.05 was considered significant; ** ) p < 0.01; *** ) p < 0.001.

Figure 8. Electron microscopy of HL-60 cell line treated with bezafibrate 1 mmol/L and cultured for 96 h. The electron microscopy analysis on the HL-60 cell line, cultured in the presence of bezafibrate (1 mmol/L) for 96 h, showed that the perturbation of mitochondrial respiratory activity induced by bezafibrate determines a serious morphological alteration of mitochondria; (A) vehicle and (B) bezafibrate.

normal conditions (i.e., in untreated cells) glucose oxidation is characterized by a NADH/FADH2 ratio of about 5:1. In bezafibrate-treated cells, the overall metabolism

is modified in such a way that this ratio tends to decrease much more than NADH dehydrogenase activity is hampered. On the basis of this hypothesis, PPAR-R activation could be due, at least in part, to a shift in metabolic energy state, which does preferentially use lipids through glycerol catabolism by mitochondrial FAD-dependent glycerol 3-phosphate dehydrogenase and fatty acids β-oxidation via electron transferring flavoprotein. In other words, according to Kersten et al. (5), fatty acids stimulate their own metabolism. The same pathogenic mechanism, i.e., a switch toward lipid metabolism secondary to mitochondrial complex I inhibition, could be at the basis of the reported stimulation of expression of genes encoding for the cytochrome P450 IV family, which are responsible for microsomal ω-oxidation of long chain and very long chain fatty acids (30). In conclusion, these data confirm previous studies on the mitochondrial derangement connection with the extraperoxisomal effects of peroxisome proliferators (10-12, 31). On the whole, these results suggest that fibrate-induced modifications of cell metabolism are linked to their action on mitochondria, independently from their binding to PPARs. In this way, the hypotriglyceridemic and hypoglycemic activities of fibrates find a common explanation, thus confirming the hypothesis of Kersten et al. (5) on some conflicting aspects of glucose disposal related to PPAR ligands that in our opinion too cannot be considered a direct consequence of their hypolipidaemic activity, as previously inferred (3, 4). Recently, Motojima reported an early and rapid induction of pyruvate dehydrogenase kinase 4 after fibrate administration in the mouse (32). The metabolic switching hypothesis forwarded is in good agreement with our observations, differing only by prospecting the inhibition of pyruvate dehydrogenase as the primary causative agent of the metabolic shift to increased fatty acid oxidation. Our data suggest an alternative molecular mechanism for the hypoglycemic action of some PPARs ligands that appear to be independent from PPARs tissue distribution. Yet, this particular mitochondrial derangement could also permit to better explain the peroxisomal proliferation of PPAR-R synthetic ligands in rodents. In fact, the lipidic component of the diet of these animals is particularly rich in polyunsaturated fatty acids. Hence, the peroxisomal

1446 Chem. Res. Toxicol., Vol. 16, No. 11, 2003

β-oxidation, normally more active in rodents with respect to humans and primates, could be further enhanced by the hampering of mitochondrial NADH oxidation (33, 34). Moreover, the increase of oxygen free radical species produced by this stimulated peroxisomal β-oxidation could have a role in the carcinogenic properties of PPAR ligands in rodents. The role of the mitochondrial derangement in the pharmacology of PPARs ligands is further confirmed by the observation that ligands of different isoforms of PPARs caused, although with a different efficiency, the same mitochondrial derangement at the level of NADHcytochrome c reductase (Table 1). As far as PPARγ pharmacotoxicology is concerned, this peculiar effect could contribute to the toxicity of troglitazone (the first pharmacological PPARγ ligand), a new antidiabetic therapeutic agent (35). In fact, this drug is such a powerful inhibitor of mitochondrial NADH oxidation that it can account for the heavier derangement of complex I and impairs NADH oxidation so strongly as to forbid the induction of β-oxidation as a compensatory energetic way, leaving cell energy metabolism almost exclusively dependent on glycolisis. The damage to the energy producing pathways can thus be at the basis of (i) the prevalent hypoglicemic activity of γ-ligands, even with minor or absent hypolipidemic effects; (ii) the weight gain typically recorded in patients treated with PPARγ ligands; (iii) the differentiation activity on adipose tissue exerted by these drugs; and (iv) the often dramatic cardiac and hepatic toxicity of troglitazone (35, 36). Furthermore, it is interesting to emphasize that other peroxisome proliferators and related compounds, which share hypotrygliceridemic and hypoglycemic therapeutic characteristics, display also mitochondrial inhibitory activity, antiproliferative properties, and differentiating functions (Table 1). Finally, the mitochondrial derangement could also help in explaining some intriguing and dangerous side effects (rhabdomyolysis, cardiac insufficiency, and acute liver insufficiency) observed upon treatment with fibrates (37). A similar toxic effect could be suspected for the rhabdomyolysis occurring when HMG-CoA reductase inhibitors, which reduce the synthesis of coenzyme Q (i.e, cerivastatin), are administered to patients in combined therapy with fibrate derivatives (38, 39). In our opinion, these data should push to further investigate the biochemical and pharmacotoxicological profiles of the so-called PPARs ligands and, at the same time, the signal transduction pathways that link mitochondrial metabolism to cell proliferation and/or cell death.

References (1) Witzum, J. L. (1996) Drugs used in the treatment of hyperlipoproteinemias. In The Pharmacological Basis of Therapeutics, 9th ed. (Hardman, J. G., Limbird, L. E., Molinoff, P. R., Ruddon, R. W., Goodman, G. A., and Rall, T.W., Eds.) pp 875-897, Pergamon Press, New York. (2) Vamecq, J., and Latruffe, N. (1999) Medical significance of peroxisome proliferator-activated receptors. Lancet 354, 141-147. (3) Issemann, I., and Green, S. (1990) Activation of a member of the steroid hormone receptor superfamily by peroxisome proliferators. Nature 347, 645-650. (4) Desvergne, B., and Wahli, W. (1999) Peroxisome proliferatoractivated receptors: Nuclear control of metabolism. Endocr. Rev. 20, 649-688. (5) Kersten, S., Desvergne, B., and Wahli, W. (2000) Roles of PPARs in health disease. Nature 405, 421-424.

Scatena et al. (6) Perutz, M. F., and Poyart, C. (1983) Bezafibrate lowers oxygen affinity of hemoglobin. Lancet 2, 881-882. (7) Scatena, R., Nocca, G., Messana, I., De Sole, P., Baroni, S., Zuppi, C., Castagnola, M., and Giardina, B. (1995) Effects of gemfibrozil on the oxygen transport properties of erythrocytes. Br. J. Clin. Pharmacol. 39, 322-327. (8) Scatena, R., Nocca, G., De Sole, P., Baroni, S., Zuppi, C., Castagnola, M., Giardina, B. (1997) The priming effect of gemfibrozil on reactive oxygen metabolism of phagocytic leukocytes. An intriguing side effect. Clin. Chim. Acta 266, 173-183. (9) Borkird, C., Choi, J. H., and Sung, Z. R. (1986) Effects of 2,4dichlorophenoxyacetic on the expression of embryogenic program in carrot. Plant Physiol. 81, 1143-1146. (10) Zhou, S., and Wallace, K. B. (1999) The effect of peroxisome proliferators on mitochondrial bioenergetics. Toxicol. Sci. 48, 8289. (11) Dzhekova-Stojkova, S., Bogdanska, J., and Stojkova, Z. (2001) Peroxisome proliferators: their biological and toxicological effects. Clin. Chem. Lab. Med. 39, 468-474. (12) Youssef, J., and Badr, M. (1998) Extraperoxisomal targets of peroxisome proliferators: mitochondrial, microsomal, and cytosolic effects. Implications for health and disease. Crit. Rev. Toxicol. 28, 1-33. (13) Lazar, M. A. (2001) Progress in cardiovascular biology: PPAR for the course. Nat. Med. 7, 23-24. (14) Sarraf, P., Mueller, E., Jones, D., King, F. J., DeAngelo, D. J., Partridge, J. B., Holden, S. A., Chen, L. B., Singer, S., Fletcher, C., and Spiegelman, B. M. (1998) Differentiation and reversal of malignant changes in colon cancer through PPAR γ. Nat. Med. 4 (9), 1046-1052. (15) Lefebvre, A. M., Chen, I., Desreumaux, P., Najib, J., Fruchart, J. C., Geboes, K., Briggs, M., Heyman, R., and Auwerx, J. (1998) Activation of the peroxisome proliferator-activated receptor γ promotes the development of colon tumors in C57BL/6JAPCmin/+ mice. Nat. Med. 4 (9), 1053-1057. (16) Scatena, R., Nocca, G., De Sole, P., Rumi, C., Puggioni, P., Remiddi, F., Bottoni, P., Ficarra, S., and Giardina, B. (1999) Bezafibrate as differentiating factor of human myeloid leukemia cells. Cell Death Differ. 6 (8), 781-787. (17) Park, B. H., Breyer, B., and He, T. C. (2001) Peroxisome proliferator activated receptors: roles in tumorigenesis and chemoprevention in human cancer. Curr. Opin. Oncol. 13, 7883. (18) Maunsbach, A. B. (1994) Cell Biology. A Laboratory Handbook (Celis, J. E., Ed.) Vol. 2, pp 103-125, Academic Press, San Diego, CA. (19) O’Donnel, V. B., Spycher, S., and Azzi, A. (1995) Involvement of oxidants and oxidant-generating enzyme(s) in tumour-necrosisfactor-alpha-mediated apoptosis: role for lipoxygenase pathway but not mitochondrial respiratory chain. Biochem. J. 310, 133141. (20) Robinson, J. B., Jr., Brent, L. G., Sumegi, B., and Sreve, P. A. (1987) An enzymatic approach to the study of the tricarboxylic acid cycle. In Mitochondria. A Pratical Approach (Barley-Usmar, B. M., Rickwood, D., and Wilson, M. T., Eds.) pp 160-161, IRL, Oxford, U.K. (21) Schrader, M. C., Eskey, C. J., Simplaceanu, V., and Ho, C. (1993) A carbon-13 nuclear magnetic resonance investigation of the metabolic fluxes associated with glucose metabolism in human erythrocytes. Biochim. Biophys. Acta 1182, 162-178. (22) Brambilla, L., Cairo, G., Sestili, P., O’Donnel, V., Azzi, A., and Cantoni, O. (1997) Mitochondrial respiratory chain deficiency leads to overexpression of antioxidant enzymes. FEBS Lett. 418 (3), 247-250. (23) Beutler, E. (1984) Red Cell Metabolism. A Manual of Biochemical Methods, 3rd ed., Grune e Stratton, NY. (24) De Baetselier, P., and Scrham, E. (1986) Luminescent bioassay based on macrophage cell lines. Methods Enzymol. 133, 507-530. (25) Allen, R. C. (1986) Phagocytic leukocyte oxygenation activities and chemiluminescence: a kinetic approach to analysis. Methods Enzymol. 133, 449-492. (26) Crabtree, B., Souter, M. J., and Anderson, S. E. (1989) Evidence that the production of acetate in rat hepatocytes is a predominantly cytoplasmic process. Biochem. J. 257, 673-678. (27) Leighton, F., Bergseth, S., Rortveit, T., Christiansen, E. N., and Bremer, J. (1989) Free acetate production by rat hepatocytes during peroxisomal fatty acid and dicarboxylic acid oxidation. J. Biol. Chem. 264, 10347-10350. (28) Abraham, D. J., Perutz, M. F., and Phillips, S. E. (1983) Physiological and X-ray studies of potential antisickling agents. Proc. Natl. Acad. Sci. U.S.A. 80, 324-328.

Bezafibrate and Mitochondria (29) Perutz, M. F., Fermi, G., Abraham, D. J., Poyart, C., and Bursaux, E. (1986) Hemoglobin as a receptor of drugs and peptides: X-ray studies of the steochemistry of binding. J. Am. Chem. Soc. 108, 1064-1078. (30) Reddy, J. K., and Hashimoto, T. (2001) Peroxisomal β-oxidation and peroxisome proliferator-activated receptor R: an adaptive metabolic system. Ann. Rev. Nutr. 21, 193-230. (31) Qu, B., Li, Q., Wong, K. P., Tan, T. M., and Halliwell, B. (2001) Mechanism of clofibrate hepatotoxicity: mitochondrial damage and oxidative stress in hepatocytes. Free Radical Biol. Med. 31, 659-669. (32) Motojima, K. (2002) A metabolic switching hypothesis for the first step in the hypolipidemic effects of fibrates. Biol. Pharm. Bull. 25, 1509-1511. (33) Halliwell, B., and Cutteridge, J. M. C. (1986) Free Radical in Biology and Medicine, Oxford University Press, NY.

Chem. Res. Toxicol., Vol. 16, No. 11, 2003 1447 (34) Lehninger, A. L., Nelson, D. L., and Cox, M. M. (2000) Lehninger Principles of Biochemistry, Worth Publisher, NY. (35) Scheen, A. J. (2001) Thiazolidinediones and liver toxicity. Diabetes Metab. 27, 305-313. (36) Shek, A., and Ferrill, M. J. (2001) Statin-fibrate combination therapy. Ann. Pharmacother. 35, 908-917. (37) Gale, E. A. (2001) Lessons from the glitazones: a story of drug development. Lancet 357, 1870-1875. (38) Hodel, C. (2002) Myopathy and rhabdomyolysis with lipidlowering drugs. Toxicol Lett. 128, 159-168. (39) Omar, M. A., and Wilson, J. P. (2002) FDA adverse event reports on statin-associated rhabdomyolysis. Ann. Pharmacother. 36, 288-295.

TX0341052