Skip to main content
NCBI home page
British Journal of Pharmacology logoLink to British Journal of Pharmacology
. 2000 Jun;130(3):655–663. doi: 10.1038/sj.bjp.0703348

[3H]-Trimetazidine mitochondrial binding sites: regulation by cations, effect of trimetazidine derivatives and other agents and interaction with an endogenous substance

Didier Morin 1,2,*, Rosa Sapena 1, Aziz Elimadi 1, Bernard Testa 3, Serge Labidalle 4, Alain Le Ridant 5, Jean-Paul Tillement 1
PMCID: PMC1572106  PMID: 10821795

Abstract

  1. Trimetazidine, an antiischaemic drug, has been shown to restore impaired mitochondrial functions. Specific binding sites for [3H]-trimetazidine have been previously detected in liver mitochondria. In the present study we confirm this observation and provide additional evidence for the involvement of these sites in the pharmacological effects of the drug.

  2. Inhibition experiments using a series of trimetazidine derivatives revealed the presence of three classes of binding sites. An N-benzyl substituted analogue of trimetazidine exhibited a very high affinity (Ki=7 nM) for one of these classes of sites.

  3. Compounds from different pharmacological classes were evaluated for their ability to inhibit [3H]-trimetazidine binding. Among the drugs tested pentazocine, ifenprodil, opipramol, perphenazine, haloperidol, and to a lower extent prenylamine, carbetapentane and dextromethorphan competed with high affinity, suggesting a similarity of high affinity [3H]-trimetazidine sites with sigma receptors.

  4. [3H]-Trimetazidine binding was modulated by pH. Neutral trimetazidine had about 10 fold higher affinity than protonated trimetazidine for its mitochondrial binding sites. Various cations also affected [3H]-trimetazidine binding. Ca2+ was the most potent inhibitor and totally suppressed the binding of [3H]-trimetazidine to the sites of medium affinity.

  5. An endogenous cytosolic ligand was able to displace [3H]-trimetazidine from its binding sites. Its activity was not affected by boiling for 15 min, suggesting a non-protein compound.

  6. These data suggest that mitochondrial [3H]-trimetazidine binding sites could have a physiological relevance and be involved in the antiischaemic effects of the drug.

Keywords: Trimetazidine, specific binding sites, rat liver mitochondria, brain mitochondria

Introduction

Since the demonstration of the antiischaemic effect of trimetazidine both by clinical (Detry, 1993) and experimental studies (for a review see Albengres et al., 1998), many efforts have been made to identify a cellular target responsible for these effects. During the last years pharmacological studies have focused on mitochondria. Indeed, at least four experimental findings indicate that trimetazidine could act selectively in protecting mitochondria: (1) it improves mitochondrial function during ischaemic damage (Guarnieri & Muscari, 1993); (2) it reverses impairment of oxidative phosphorylation induced by cyclosporine A (Salducci et al., 1996); (3) it inhibits mitochondrial swelling generated by high Ca2+ concentrations in the presence of a prooxidant (Elimadi et al., 1997); and (4) it counteracts the hepatic injury associated with ischaemia-reperfusion by preserving mitochondrial functions (Elimadi et al., 1998). Taken together, these results suggest that the mitochondrion is the major target responsible for the antiischaemic effects of trimetazidine and imply the probable existence of specific mitochondrial target sites. We recently identified such binding sites in mitochondria purified from rat liver using [3H]-trimetazidine as a ligand (Morin et al., 1998), and showed that [3H]-trimetazidine labels two classes of binding sites as revealed by equilibrium and displacement experiments yielding low Hill-coefficients. The nature of these sites was not established but we have shown that they are not related to other previously described mitochondrial sites. In addition we observed a relationship between the ability of several amphiphilic cations to displace [3H]-trimetazidine and their ability to inhibit mitochondrial swelling.

In the present study we have progressed in the characterization of these sites. First, we have demonstrated that they are not restricted to liver mitochondria but are also present in purified brain mitochondria. Furthermore, we have studied the effect of pH and of monovalent and divalent cations on trimetazidine binding as well as the interaction with trimetazidine derivatives and drugs of different pharmacological classes. Finally, we provide evidence for the existence of a possible endogenous ligand able to modulate trimetazidine binding.

Methods

Isolation of liver and brain mitochondria

Male Wistar rats weighing approximately 250–300 g were decapitated. Their livers were excised rapidly and placed in a medium containing 250 mM sucrose, 50 mM 2-amino-2-hydroxymethyl-propan-1,3-diol (Tris) and 5 mM ethylene glycol-bis(β-aminoethyl ether)-tetraacetic acid (EGTA), pH 7.8 at 4°C. Tissue samples (28 g) were minced and homogenized (10 g tissue in 60 ml medium) on ice using a Teflon Potter homogenizer. The homogenates were centrifuged at 600×g for 10 min (Sorvall® RC 28 S) and the supernatants were centrifuged for 10 min at 3300×g to obtain mitochondrial pellets. The resulting supernatants were kept on ice to prepare the cytosolic fraction. Mitochondrial pellets were washed twice with the same medium and centrifuged at 3300×g for 10 min.

The resulting pellets were washed with medium from which EGTA was omitted and centrifuged for 10 min at 3300×g. The final pellets were suspended in 36 ml of Tris-sucrose buffer (without EGTA) and mitochondria were purified on a discontinuous sucrose gradient according to Morin et al. (1998). The purified mitochondrial suspensions were stored on ice before use. Protein content was determined by the method of Lowry et al. (1951). The supernatants kept on ice were centrifuged at 125,000×g for 90 min. The pellets were discarded and the supernatants centrifuged at 226,000×g for 90 min. The final supernatants were considered as the purified cytosolic fraction.

Rat brain mitochondria were isolated by differential centrifugation and purified on a percoll gradient according to Sims (1990). Briefly, brains were removed, sliced and washed in a buffer containing 320 mM sucrose, 10 mM Tris and 1 mM ethylenediamine tetraacetic acid (EDTA), pH 7.4 at 4°C. The brains were then homogenized in 10 ml per g of tissue of the same buffer and centrifuged 3 min at 1330×g (Sorvall® RC 28 S). The supernatants (S1) were kept on ice and the pellets were suspended in 5 ml per g of tissue of the same buffer and centrifuged at 1330×g for 3 min. The pellets were discarded, the supernatants S2 pooled to S1 and centrifuged at 21,200×g for 10 min. The final pellets were suspended in 15% (v/v) percoll (10 ml g−1 of tissue originally homogenized) and layered on a discontinuous density gradient of 3.5 ml each of 23 and 40% percoll. The tubes were centrifuged at 30,700×g for 5 min. The purified mitochondrial rings were recovered at the interface of the 23–40% fractions using a Pasteur pipette and rinsed in the isolation buffer.

All animal procedures used in this study are in strict accordance with the French Agency's policy on animal experimentation (Ministère de l'Agriculture et de la Forêt, authorization No. 00768). Agreement for publication of our results has been given by all companies concerned.

Binding experiments

For inhibition experiments, purified mitochondria were suspended in a Tris buffer (50 mM Tris, 250 mM sucrose, pH 7.4) (2 mg ml−1 protein), and incubated with 7–10 nM [3H]-trimetazidine (83 Ci mmole−1) and 24 to 28 different concentrations of competing drug or Tris buffer for 45 min at 25°C in a total volume of 400 μl. Specific binding was defined as the difference between total binding and binding in the presence of 1 mM of the trimetazidine derivative S-00226 (Figure 1). For saturation analysis [14C]-trimetazidine (0.5–550 μM; specific activity 54.9 mCi mmol−1) was used and the non-specific binding was defined in the presence of 50 mM S-00226. The protein concentration was 2 mg ml−1 protein (liver) or 0.7 mg ml−1 protein (brain).

Figure 1.

Figure 1

Open in a new tab

Chemical structures of trimetazidine (1-(2,3,4-trimethoxybenzyl)piperazine) and trimetazidine derivatives.

At the end of each incubation, bound and free ligands were separated by rapid filtration through Whatman GF/B glass fibre filters (presoaked in 0.1% polyethylenimine). Each filter was washed twice with an additional 5 ml of ice-cold phosphate buffer (25 mM) and counted in a liquid scintillation counter Packard 1600 TR with an efficiency of 45%. The filtration process was rapid enough to avoid the dissociation of radiolabelled trimetazidine from its binding sites. In addition, in all cases radiolabelled trimetazidine binding to glass fibre filters was very low, less than 0.05% of total ligand concentration.

Drugs

Labelled (3H and 14C) and unlabelled trimetazidine (Figure 1) were kindly provided by Servier Laboratories (Neuilly-sur-Seine, France). Trimetazidine derivatives (Figure 1) either were provided by Servier Laboratories or synthesized by Synthe-Pharma (Seysses, France) at the request of Servier Laboratories. Other drugs and chemicals were obtained from Sigma (St Quentin Fallavier, France) or Merck (Nogent-sur-Marne, France) and were of the highest purity available.

Data analysis

The binding data for equilibrium saturation experiments were fitted to equation (1) describing the interaction of a selective ligand with two classes of high and low affinity binding sites. Bmax, F and Kd denote the site density, the free ligand and the dissociation constant of each class, respectively.

graphic file with name 130-0703348e1.jpg

In inhibition experiments, the displacement curves were fitted to the general binding isotherm (Hill model):

graphic file with name 130-0703348e2.jpg

where B is the number of binding sites observed in the presence of a particular inhibitor concentration (C), Beq the bound ligand at equilibrium, IC50 the concentration that inhibits 50% of the maximal binding and nH the pseudo-Hill coefficient (Weiland & Molinoff, 1981).

A nH value of 1 corresponds to a competitive interaction and thus indicates the presence of one class of binding sites. When nH value was statistically lower than 1 (according to the Aikake criterion), a competitive model considering a competitive interaction including two classes or more of binding sites was applied. The following equation was used:

graphic file with name 130-0703348e3.jpg

Data from all binding experiments were analysed by means of non-linear regression using commercially available software (Micropharm, INSERM 1990; Urien, 1995). All data are presented as the mean±s.e.mean of three or more individual experiments.

Results

Evidence for the existence of trimetazidine binding sites on liver and brain mitochondria

Binding studies of [14C]-trimetazidine were performed with purified rat liver and brain mitochondria. As described previously, the binding of [14C]-trimetazidine to liver mitochondria was saturable. Scatchard analysis of the data confirmed the presence of two populations of binding sites, the high affinity sites representing only 5% of the total (Table 1). An accumulation of the drug inside mitochondria without binding was also considered. Indeed, we recently showed that timetazidine can cross cellular membranes transferring protons from an acidified cellular compartment to the extracellular space (Reymond et al., 1999).

Table 1.

Equilibrium binding parameters of [14C]-trimetazidine for mitochondrial binding sites in rat brain and liver

graphic file with name 130-0703348t1.jpg

Open in a new tab

The same binding parameters were found when mitochondria were broken, subjected to hypo-osmotic shock and sonication, or frozen. These procedures eliminate the possibility of a trimetazidine uptake without binding. [14C]-Trimetazidine binding sites were also present in purified rat brain mitochondria. The Scatchard plot was linear and the Hill coefficient was close to unity indicating the presence of only one class of binding sites (Figure 2). The total binding capacity was 507 pmol mg−1 protein but no high affinity sites could be detected. However, displacement experiments allowed us to detect them (results not shown). This discrepancy between displacement and equilibrium experiments is due to the low selectivity of [14C]-trimetazidine between the two sites, and especially to the low proportion of high affinity binding sites in rat brain yielding their detection impossible by equilibrium experiments. Clearly, the study of these sites requires a better ligand.

Figure 2.

Figure 2

Open in a new tab

Equilibrium binding of [14C]-trimetazidine to rat brain mitochondria. [14C]-Trimetazidine (0.5–550 μM) was incubated with mitochondria (0.7 mg ml−1) for 45 min at 25°C. Non-specific binding was determined in presence of 50 mM S-00226. Bmax and Kd values were 0.355 and 223 μM, respectively. Values are means for three independent experiments performed in duplicate. Equilibrium binding parameters were estimated by a non-linear regression analysis. (A) Direct plot. (B) Scatchard plot.

Effect of trimetazidine derivatives on [3H]-trimetazidine binding

The need for a better ligand to label trimetazidine sites prompted us to investigate the effect of different trimetazidine derivatives on [3H]-trimetazidine binding. Thus, we examined a series of compounds substituted either on the ortho-oxygen, the piperazine ring, or both (Table 2). We generated inhibition binding curves by displacement of a low concentration of [3H]-trimetazidine (8 nM). This protocol permitted detection of the presence of several binding sites, and under these conditions IC50 values of the competing drugs were close to their Ki values (Swillens et al., 1995). The displacement of [3H]-trimetazidine by trimetazidine indicated that 7–10 nM [3H]-trimetazidine labelled both families of sites to a comparable extent (Table 2), confirming our previous observation (Morin et al., 1998). The same results were obtained with some (S-16858, S-00226, S-00700) but not all derivatives. Indeed, according to the displacing agent used, the ratio of high to low affinity sites varied from 25/75 to 65/35, and four drugs displaced all [3H]-trimetazidine binding sites with the same affinity. Some of these drugs showed a high affinity for site1, especially S-64329 (IC50=0.02 μM) which was also highly selective for this site with an IC50 ratio of 1710 (Table 2; Figure 3).

Table 2.

Inhibition concentrations and proportions of high and low affinity binding sites of trimetazidine derivatives for [3H]-trimetazidine binding sites in rat liver mitochondria

graphic file with name 130-0703348t2.jpg

Open in a new tab

Figure 3.

Figure 3

Open in a new tab

Inhibition of specific [3H]-trimetazidine binding to rat liver mitochondria (2 mg ml−1) by increasing concentrations of S-16950 and S-64329. Data are plotted as percentages of the control [3H]-trimetazidine binding value (in the absence of inhibitor : [3H]-trimetazidine bound =100%). [3H]-Trimetazidine concentration was 8 nM. Control binding values varied between 20 to 30 fmol mg−1 of protein and non-specific binding between 27 to 40%. The pseudo-Hill coefficient nH being lower than 1 (S-16950, nH=0.34±0.02; S-64329, nH=0.34±0.04), the curves were fitted to a two-site model. This model was not satisfactory for S-16950, but fitting to a three-site model significantly improved the goodness of fit (P<0.01).

These puzzling results are difficult to interpret. The only compatible assumption is the presence of more than two families of binding sites. Another trimetazidine derivative, S-16950, suggested a solution to the problem. Indeed, S-16950 inhibition of [3H]-trimetazidine binding resulted in very shallow competition curves with a nH value of 0.34±0.02 and a mean IC50 value of 0.32±0.14 μM. When these data were subjected to computer modelling, it was found that the inhibition curve was better described by a three-site rather than a two-site model (P<0.01; Figure 3). The one-site model was unacceptable. At this particular [3H]-trimetazidine concentration (7–10 nM), S-16950 IC50 values were 7,320 and 45,000 nM corresponding to binding site proportions of 41, 25 and 34% of the total, respectively (Figure 3). Interestingly, the combination of these percentages accounts approximately for all binding ratios (expressed as the percentage of the total number of binding sites) obtained in Table 2. For example, S-15739 showed a ratio of binding sites (site1/site2) of 65/35 (65% corresponding to the sum 41+25%), whereas S-00218 showed the reverse (35/65). For most drugs, the ratio was within the range 40/60–60/40 indicating that these drugs discriminate between the high (41%) and the two lower (25+34%) families of sites.

Dependence of [3H]-trimetazidine binding on pH and ions

Since trimetazidine is a dibasic compound with two pKa values, we investigated the effect of pH and cations on [3H]-trimetazidine binding. The pH-dependent binding curve of trimetazidine is shown in Figure 4A with specific binding increasing up to pH=8.5 and a mid-point being reached close to pH=7.5. In fact, this binding curve resembles an ionization curve, and indeed such a protonation curve can be shown to mimic approximately the binding curve. In Figure 4A, a protonation curve for a base of pKa=7.6 is compared with the binding curve, revealing a partial fit.

Figure 4.

Figure 4

Open in a new tab

Effect of pH on [3H]-trimetazidine binding to rat liver mitochondria. (A) [3H]-Trimetazidine binding to mitochondria (full curve), compared with the protonation curve (broken line running from 99 to 1% ionization) of a base with pKa=7.6. (B) Inhibition of [3H]-trimetazidine binding by increasing concentrations of S-16950 at different pH. [3H]-Trimetazidine concentration was 8 nM. The curves were fitted to a three-site model. IC50 values were 0.008, 1.1 and 64.4 μM at pH=9, 0.005, 0.34 and 28.9 μM at the pH=7.4 and 0.011, 0.19 and 216 μM at pH=6.

This pH-dependent binding profile prompted us to determine which binding sites were affected by pH. We carried out competition studies in the presence of S-16950 at three pH values: 6, 7.4 and 9 (Figure 4B). All [3H]-trimetazidine binding sites decreased with pH. The high affinity sites were more greatly affected, decreasing from 40 to 2.2 fmol mg−1 protein (−94.5%) when pH varied from 9 to 6.

The ability of [3H]-trimetazidine to bind to its sites was also found to be dependent on some ions present in the medium. As shown in Figure 5A, [3H]-trimetazidine specific binding was inhibited in a concentration-dependent manner by NaCl, KCl, MgCl2 and CaCl2, the latter being the most potent. The weak inhibition effect of NaCl was assigned to the Cl− ion because Na+-acetate did not significantly modify binding at concentrations up to 150 mM. Hence, Na+ was without effect. In contrast to Na+-acetate, K+-, Mg+- and Ca2+-acetate decreased binding activity (data not shown) but less than their chloride salts. These results confirm the inhibition of [3H]-trimetazidine binding by Cl−, K+, Mg2+ and Ca2+.

Figure 5.

Figure 5

Open in a new tab

Effect of cations on specific [3H]-trimetazidine binding to rat liver mitochondria. (A) Inhibition of [3H]-trimetazidine binding (8 nM) by increasing concentrations of various cations used as acetate and chloride salts. Data are plotted as percentages of the control [3H]-trimetazidine binding value (in the absence of cation : [3H]-trimetazidine bound=100%). (B) Inhibition of [3H]-trimetazidine binding by increasing concentrations of S-16950 in the absence or in the presence of 60 mM Ca2+.

In order to determine which family of [3H]-trimetazidine binding sites was affected by these ions, we performed competition experiments with S-16950 in the presence of 60 mM CaCl2. As expected, Ca2+ reduced the binding of [3H]-trimetazidine (Figure 5B). The higher and lower affinity sites were not affected, whereas [3H]-trimetazidine binding to the medium affinity sites was almost completely suppressed (Figure 5B).

Pharmacological profile of [3H]-trimetazidine binding sites

In a previous study (Morin et al., 1998), we observed that [3H]-trimetazidine binding sites differed from known mitochondrial sites and interacted with a number of drugs. Here we continue the pharmacological characterization of these sites, investigating the effects of several classes of drugs.

No antiischaemic agent (Table 3) competed with very high affinity, prenylamine and flunarizine being the most effective. Four drugs displaced total specific [3H]-trimetazidine binding in a monophasic manner, demonstrating a relatively high affinity for all [3H]-trimetazidine binding sites. Among the other classes of drugs, the antidepressant opipramol, the antipsychotics perphenazine and haloperidol, and the local anaesthetic butacaine were the most effective with IC50 values for the high affinity sites of 0.02, 0.04, 0.07 and 0.04 μM, respectively (Table 4); (−)-lobeline, proadifen and to a lesser extent dextromethorphan were also effective. Given that the most effective drugs belong to different pharmacological classes, we looked for another common property and found that they are all known to interact with sigma receptors. This prompted us to study the effect of more selective sigma receptor ligands, i.e. pentazocine, ifenprodil, 1,3-di(2-tolyl)guanidine (DTG), carbetapentane and 3-(3-hydroxyphenyl)-N-(1-propyl)piperidine (3-PPP) (Quirion et al., 1992). Indeed, these compounds inhibited [3H]-trimetazidine binding with IC50 values close to their sigma receptor affinity (Dehaven-Hudkins et al., 1994; Matsuno et al., 1996).

Table 3.

Inhibition concentrations and proportions of high and low affinity binding sites of antiischaemic drugs for [3H]-trimetazidine binding sites in rat liver mitochondria

graphic file with name 130-0703348t3.jpg

Open in a new tab

Table 4.

Inhibition concentrations and proportions of high and low affinity binding sites of several classes of antiischaemic drugs for [3H]-trimetazidine binding sites in rat liver mitochondria

graphic file with name 130-0703348t4.jpg

Open in a new tab

Evidence for an endogenous trimetazidine-displacing substance

Because [3H]-trimetazidine binding sites appeared specific and some analogues showed high affinity, it was tempting to investigate the possible presence of an endogenous ligand which was able to recognize these binding sites. To this end we isolated a cytosolic fraction devoid of organelles and membranes and evaluated increasing concentrations of this fraction on [3H]-trimetazidine binding (Figure 6). Complete inhibition of binding was seen at high concentrations. This persisted when the cytosol was boiled for 15 min and centrifuged to eliminate precipitated proteins. The presence of 5 mM EGTA in the isolation medium ruled out a possible action of Ca2+, and EGTA alone was without effect.

Figure 6.

Figure 6

Open in a new tab

Effect of increasing concentration of cytosol on specific [3H]-trimetazidine binding to rat liver mitochondria. The plot is representative of nine experiments performed with five different preparations of cytosol.

Discussion

Trimetazidine is an antiischaemic drug effective in various experimental models, but its mechanism of action is not fully understood. Numerous data indicate that mitochondria could be its main target, and the recent demonstration of specific trimetazidine binding sites located on liver mitochondria has reinforced this hypothesis (Morin et al., 1998). Although we have provided evidence that the low-affinity trimetazidine binding sites may be involved in the regulation of the mitochondrial permeability transition pore, the implication of these sites in the mechanism of action of trimetazidine needed to be established.

The physiological relevance of these sites is strengthened, using the cytosol of liver cells, by the present demonstration of the presence of an endogenous substance which is able to displace [3H]-trimetazidine from its binding sites. The nature of this compound is not yet established, but it does not seem to be a protein since it was not affected by boiling for 15 min. The demonstration of the selectivity of this substance towards trimetazidine sites, and its purification, will be one of the main objectives of future studies.

The pH strongly influenced [3H]-trimetazidine binding. This drug has two pKa values and its binding was closely related to its protonation state. [3H]-Trimetazidine binding increased with pH. The pKa of the more basic group in trimetazidine is 9.14±0.02, which seems incompatible with the apparent pKa of 7.6 giving the best concordance between the binding and protonation curves in Figure 4A. However, it must be remembered that the apparent aqueous pKa of a base partitioning in an organic solvent experiences a shift to lower values (pKa shift) which is related to the partition coefficient of the base. The well-known equation describing this phenomenon is:

graphic file with name 130-0703348e4.jpg

where P is the solvent/water partition coefficient of the base, pKoa is its aqueous pKa in the biphasic system, and r is the ratio of volumes of the organic and aqueous phases (Kaufman et al., 1975; Clarke & Cahoon, 1987; Avdeef, 1992).

Assuming that the protonation curve in Figure 4A is indeed that of trimetazidine in the mitochondria/water system, we have a pKa – pKoa≈1.5, i.e. P≈(101.5−1)/r≈33/r. Assuming r values of 1/100 or 1/1000, we obtain log P=3.5 or 4.5, respectively, for the mitochondria/water partitioning of neutral trimetazidine. This suggests that the binding curve of trimetazidine can indeed be interpreted to indicate that neutral trimetazidine has about 10 fold higher affinity than protonated trimetazidine for its binding sites on mitochondria. The fact that the fit between the two curves is far from perfect is suggestive of additional equilibria and the drop in specific binding beyond pH=8.5 could well be due to ionic changes in the mitochondrial membranes.

Divalent cations were more potent in inhibiting [3H]-trimetazidine binding than monovalent cations. However, the high concentrations needed to inhibit [3H]-trimetazidine binding suggest that only minimal displacement would occur in the extra- or intra-mitochondrial environment under normal conditions. Only under pathological conditions could this effect be of relevance. Indeed, it is generally thought that during ischaemia-reperfusion the binding of Ca2+ to an internal site of the mitochondrial membrane is a key event in the sequence of events leading to the opening of the mitochondrial transition pore (MTP; Bernardi et al., 1993; Zoratti & Szabo', 1995). The opening of the MTP results in mitochondrial and thus, cellular injury (Bernardi, 1996). We recently observed that trimetazidine is able to inhibit MTP opening and that a relation exists between the ability of different drugs to inhibit MTP opening induced by Ca2+ in the presence of a prooxidant and the displacement of [3H]-trimetazidine bound to mitochondria (Morin et al., 1998). This effect may result from an interaction between Ca2+ and the drug, the former being displaced by the drug. Indeed, our present data show that the regulation of [3H]-trimetazidine binding by Ca2+ is characterized by a competitive inhibition reflected by (1) a monophasic inhibition (nH=1) observed in the presence of increasing concentrations of Ca2+ and (2) a decrease in the density of the medium affinity sites without any modification either of the number of high affinity sites or of the affinity of S-16950 for these sites. This interaction could explain the inhibition of MTP opening by trimetazidine.

A major obstacle in understanding the role for these binding sites was the low affinity of the labelling drug and the diversity of the sites. In order to facilitate their identification and molecular characterization, we examined trimetazidine derivatives as molecular probes for [3H]-trimetazidine binding sites. This changed the binding profile of trimetazidine to become more heterogeneous since up to three binding sites were now detected in the presence of the N-benzyl derivative S-16950. It should be noted that the binding is highly dependent on the chemical structure of the drugs since a small substitution can induce a large variation in affinity. Indeed, replacement of a 2-methoxy group (S-64329) by a phenolic group (S-16950) confered a higher affinity and shifted the inhibition curve from a two-site to a three-site interaction (see Figure 3). S-16950 is the only drug tested up to now permitting the identification of three families of binding sites.

From a pharmacological point of view, it is interesting to note that the most effective inhibitors of [3H]-trimetazidine binding are drugs known to interact with sigma receptors, especially sigma1 receptors which are abundant in the liver (Dehaven-Hudkins et al., 1994). Moreover, the rank order of affinity pentazocine≈thinsp;4 opipramol≈thinsp;4 perphenazine≈thinsp;4 haloperidol>DTG>prenylamine>dextromethorphan etc . . . is consistent with binding to sigma sites (Klein & Musacchio, 1989; Basile et al., 1992; Dehaven-Hudkins et al., 1994). In addition, S-16950 is structurally related to N-substituted 4-benzylpiperidines, such as ifenprodil and eliprodil, which exhibit high affinity for sigma1 receptors (Karbon et al., 1990; Whittemore et al., 1997), and its chemical structure is close to that of SA4503 which has recently been shown to be a potent and selective sigma1 receptor agonist (Matsuno & Mita, 1998). It should be pointed out that ifenprodil also displaced [3H]-trimetazidine binding with a high affinity.

Collectively, these data suggest that one of the mitochondrial sites labelled by [3H]-trimetazidine could be a sigma receptor. This raises the question whether these receptors are located on mitochondria. Little is available to answer this question. Most studies were performed with brain preparations, showing an enrichment of sigma receptors in the microsomal and myelin fractions (McCann & Su, 1990). McCann et al. (1994) concluded that localization of these receptors in mitochondria is highly unlikely but Itzhak et al. (1991) and Dehaven-Hudkins et al. (1994) demonstrated a significant binding of sigma ligands to the mitochondrial fraction. Such results seem to depend on the radioligand used but cross-contamination is also a problem. However, our mitochondrial preparations were highly purified and the activities of representative markers of microsomes and plasma membranes were undetectable. Taken together, these data suggest that sigma receptors are indeed present on mitochondria.

The possible interaction of trimetazidine with sigma receptors would be of particular interest as several ligands of these receptors have been reported to have neuroprotective effects in in vitro and in vivo models of ischaemia (Matsuno & Mita, 1998). Although no clear pharmacological definition of a sigma agonist or antagonist and no clear correlation between sigma receptor affinity and neuroprotective potency have yet been established (Maurice & Lockhart, 1997), our hypothesis could contribute to a better understanding of the mechanism of action of trimetazidine.

In conclusion, the data presented in this study confirm the heterogeneity of [3H]-trimetazidine binding sites. The sites of medium affinity are regulated by Ca2+ and could explain the inhibition of MTP opening by trimetazidine. A trimetazidine derivative, S-16950, showed a very high affinity for trimetazidine binding sites. Various data indicate that these sites located on mitochondria could represent a sigma receptor; a detailed binding analysis of S-16950 is in progress to evaluate its possible interaction with these receptors. The demonstration of the possible existence of an endogenous ligand able to modulate trimetazidine binding reinforced the hypothesis that these binding sites could have physiological relevance.

Acknowledgments

This work was supported by the Institut de Recherches Internationales Servier, the Ministère de l'Education Nationale (DRED EA 427), and the Réseau de Pharmacologie Clinique.

Abbreviations

DTG

1,3-di(2-tolyl)guanidine

EDTA

ethylenediamine,tetraacetic acid

EGTA

ethylene glycol-bis(β-aminoethyl ether)-tetraacetic acid

MTP

mitochondrial transition pore

3-PPP

3-(3-hydroxyphenyl)-N-(1-propyl)piperidine

Tris

2-amino-2-hydroxymethyl-propan-1,3-diol

References

  1. ALBENGRES E., TILLEMENT J.P., LE LOUET H., MORIN D. Trimetazidine: experimental and clinical update review. Cardiovasc. Drug Rev. 1998;16:359–390. [Google Scholar]
  2. AVDEEF A. pH-metric log P. Part 1. Difference plots for determining ion-pair octanol-water partition coefficients of multiprotic substances. Quant. Struct-Act Relat. 1992;11:510–517. [Google Scholar]
  3. BASILE A.S., PAUL I.A., MIRCHEVICH A., KUIJPERS G., DE COSTA B. Modulation of (+)-[3H]pentazocine binding to guinea pig cerebellum by divalent cations. Mol. Pharmacol. 1992;42:882–889. [PubMed] [Google Scholar]
  4. BERNARDI P. The permeability transition pore. Control points of a cyclosporin A-sensitive mitochondrial channel involved in cell death. Biochim. Biophys. Acta. 1996;1275:5–9. doi: 10.1016/0005-2728(96)00041-2. [DOI] [PubMed] [Google Scholar]
  5. BERNARDI P., VERONESE P., PETRONILLI V. Modulation of the mitochondrial cyclosporin A-sensitive permeability transition pore. J. Biol. Chem. 1993;268:1005–1010. [PubMed] [Google Scholar]
  6. CLARKE F.H., CAHOON N.M. Ionization constants by curve fitting: determination of partition and distribution coefficients of acids and bases and their ions. J. Pharm. Sci. 1987;76:611–620. doi: 10.1002/jps.2600760806. [DOI] [PubMed] [Google Scholar]
  7. DEHAVEN-HUDKINS D.L., LANYON L.F., FORD-RICE F.Y., ATOR M.A. σ recognition sites in brain and peripheral tissues; Characterization and effects of cytochrome P450 inhibitors. Biochem. Pharmacol. 1994;47:1231–1239. doi: 10.1016/0006-2952(94)90395-6. [DOI] [PubMed] [Google Scholar]
  8. DETRY J.M. Clinical features of an anti-anginal drug in angina pectoris. Eur. Heart J. 1993;14 suppl G:18–24. doi: 10.1093/eurheartj/14.suppl_g.18. [DOI] [PubMed] [Google Scholar]
  9. ELIMADI A., MORIN D., SAPENA R., CHAUVET-MONGES A.M., CREVAT A., TILLEMENT J.P. Comparison of the effects of cyclosporine A and trimetazidine on Ca2+-dependent mitochondrial swelling. Fundam. Clin. Pharmacol. 1997;11:440–447. doi: 10.1111/j.1472-8206.1997.tb00206.x. [DOI] [PubMed] [Google Scholar]
  10. ELIMADI A., SETTAF A., MORIN D., SAPENA R., LAMCHOURI F., CHERRAH Y., TILLEMENT J.P. Trimetazidine counteracts the hepatic injury associated with ischemia-reperfusion by preserving mitochondrial function. J. Pharmacol. Exp. Ther. 1998;286:23–28. [PubMed] [Google Scholar]
  11. GUARNIERI C., MUSCARI C. Effect of trimetazidine on mitochondrial functions and oxidative damage during reperfusion of ischemic hypertrophied rat myocardium. Pharmacology. 1993;46:324–331. doi: 10.1159/000139070. [DOI] [PubMed] [Google Scholar]
  12. ITZHAK Y., STEIN I., ZHANG S.H., KASSIM C.O., CRISTANTE D. Binding of σ-ligands to C57BL/6 mouse brain membranes: effects of monoamine oxidase inhibitors and subcellular distribution studies suggest the existence of σ-receptor subtypes. J. Pharmacol. Exp. Ther. 1991;257:141–148. [PubMed] [Google Scholar]
  13. KARBON E.W., PATCH R.J., PONTECORVO M.J., FERKANY J.W. Ifenprodil potently interacts with [3H]-(+)-3-PPP σ binding sites in guinea pig brain membranes. Eur. J. Pharmacol. 1990;176:247–248. doi: 10.1016/0014-2999(90)90538-h. [DOI] [PubMed] [Google Scholar]
  14. KAUFMAN J.J., SEMO N.M., KOSKI W.S. Microelectrometric titration measurement of the pka's and partition and drug distribution coefficients of narcotics and narcotic antagonists and their pH and temperature dependence. J. Med. Chem. 1975;18:647–655. doi: 10.1021/jm00241a001. [DOI] [PubMed] [Google Scholar]
  15. KLEIN M., MUSACCHIO J.M. High affinity dextromethorphan binding sites in guinea pig brain. Effect of sigma ligands and other agents. J. Pharmacol. Exp. Ther. 1989;251:207–215. [PubMed] [Google Scholar]
  16. LOWRY O.H., ROSEBROUGH N.J., FARR A.L., RANDALL R.J. Protein measurement with the folin phenol reagent. J. Biol. Chem. 1951;93:265–275. [PubMed] [Google Scholar]
  17. MATSUNO K., MITA S. SA4503: a novel sigma1 receptor agonist. CNS Drugs Reviews. 1998;4:1–24. [Google Scholar]
  18. MATSUNO K., NAKAZAWA M., OKAMOTO K., KAWASHIMA Y., MITA S. Binding properties of SA4503 a novel and selective sigma1 receptor agonist. Eur. J. Pharmacol. 1996;306:271–279. doi: 10.1016/0014-2999(96)00201-4. [DOI] [PubMed] [Google Scholar]
  19. MAURICE T., LOCKHART B.P. Neuroprotective and anti-amnesic potentials of sigma (σ) receptors ligands. Prog. Neuropsychopharmacol. Biol. Psychiat. 1997;21:69–102. doi: 10.1016/s0278-5846(96)00160-1. [DOI] [PubMed] [Google Scholar]
  20. MCCANN D.J., SU T.P. Haloperidol-sensitive (+)[3H]SKF-10,047 binding sites (sigma sites) exhibit a unique distribution in rat brain subcellular fractions. Eur. J. Pharmacol. 1990;188:211–218. doi: 10.1016/0922-4106(90)90004-h. [DOI] [PubMed] [Google Scholar]
  21. MCCANN D.J., WEISSMAN A.D., SU T.P. Sigma-1 and Sigma-2 sites in rat brain: comparison of regional, ontogenic, and subcellular patterns. Synapse. 1994;17:182–189. doi: 10.1002/syn.890170307. [DOI] [PubMed] [Google Scholar]
  22. MORIN D., ELIMADI A., SAPENA R., CREVAT A., CARRUPT P.A., TESTA B., TILLEMENT J.P. Evidence for the existence of [3H]-trimetazidine binding sites involved in the regulation of the mitochondrial permeability transition pore. Brit. J. Pharmacol. 1998;123:1385–1394. doi: 10.1038/sj.bjp.0701755. [DOI] [PMC free article] [PubMed] [Google Scholar]
  23. QUIRION R., BOWEN W.D., ITZHAK Y., JUNIEN J.L., MUSACCHIO J.M., ROTHMAN R.B., SU T.P., TAM S.W., TAYLOR D.C. A proposal for the classification of sigma binding sites. Trends Pharmacol. Sci. 1992;13:85–86. doi: 10.1016/0165-6147(92)90030-a. [DOI] [PubMed] [Google Scholar]
  24. REYMOND F., STEYAERT G., CARRUPT P.A., MORIN D., TILLEMENT J.P., GIRAULT H.H., TESTA B. The pH-partition profile of the antiischemic drug trimetazidine may explain its reduction of intracellular acidosis. Pharm. Res. 1999;16:616–624. doi: 10.1023/a:1018899802836. [DOI] [PubMed] [Google Scholar]
  25. SALDUCCI D., CHAUVET-MONGES A.M., TILLEMENT J.P., ALBENGRES E., TESTA B., CARRUPT P.A., CREVAT A. Trimetazidine reverses calcium accumulation and impairment of phosphorylation induced by cyclosporine A in isolated rat liver mitochondria. J. Pharmacol. Exp. Ther. 1996;277:417–422. [PubMed] [Google Scholar]
  26. SIMS N.R. Rapid isolation of metabolically active mitochondria from rat brain and subregions using Percoll density gradient centrifugation. J. Neurochem. 1990;55:698–707. doi: 10.1111/j.1471-4159.1990.tb04189.x. [DOI] [PubMed] [Google Scholar]
  27. SWILLENS S., WAELBROECK M., CHAMPEIL P. Does a radioligand bind to a homogeneous population of non-interacting receptor sites. Trends Pharmacol. Sci. 1995;16:151–155. doi: 10.1016/s0165-6147(00)89007-0. [DOI] [PubMed] [Google Scholar]
  28. URIEN S. Micropharm-K, a microcomputer interactive program for the analysis and simulation of pharmacokinetic processes. Pharm. Res. 1995;12:1225–1230. doi: 10.1023/a:1016280430580. [DOI] [PubMed] [Google Scholar]
  29. WEILAND G.A., MOLINOFF P.B. Minireview: quantitative analysis of drug-receptor interactions: I. Determination of kinetic and equilibrium properties. Life Sci. 1981;29:313–330. doi: 10.1016/0024-3205(81)90324-6. [DOI] [PubMed] [Google Scholar]
  30. WHITTEMORE E.R., ILYIN V.I., WOODWARD R.M. Antagonism of N-methyl-D-aspartate receptors by σ site ligands: potency, subtype-selectivity and mechanisms of inhibition. J. Pharmacol. Exp. Ther. 1997;282:326–338. [PubMed] [Google Scholar]
  31. ZORATTI M., SZABO' I. The mitochondrial permeability transition. Biochim. Biophys. Acta. 1995;1241:139–176. doi: 10.1016/0304-4157(95)00003-a. [DOI] [PubMed] [Google Scholar]

Articles from British Journal of Pharmacology are provided here courtesy of The British Pharmacological Society

RESOURCES