Mitochondrial permeability transition as a potential determinant of hepatotoxicity of antidiabetic thiazolidinediones

Abstract

Troglitazone, a thiazolidinedione class of antidiabetic agent, causes serious idiosyncratic hepatotoxicity. Troglitazone is metabolized to a reactive metabolite that covalently binds to cellular macromolecules, but the role of the covalent adduct in the hepatotoxicity is controversial. Because troglitazone has been found to cause cytotoxicity to hepatocytes along with mitochondrial dysfunction, we investigated the effects of troglitazone and other thiazolidinediones on mitochondrial function by using liver mitochondria fraction isolated from male CD-1 mice. Incubation of energized mitochondria with succinate in the presence of Ca²⁺ and troglitazone induced mitochondrial swelling, and the swelling was partially inhibited by cyclosporin A. Troglitazone also induced decreases in mitochondrial membrane potential and mitochondrial Ca²⁺ accumulation. These results demonstrate that troglitazone induces mitochondrial permeability transition (MPT). Similar results were obtained for ciglitazone, whereas rosiglitazone and pioglitazone, which are less hepatotoxic than troglitazone, had little effect on these mitochondria functions. It is therefore possible that the troglitazone-induced opening of MPT pore, which is not induced by rosiglitazone or pioglitazone, may contribute to the hepatotoxicity induced specifically by troglitazone.

Introduction

Troglitazone, a thiazolidinedione class of antidiabetic agent, has been found to cause serious idiosyncratic hepatotoxicity and was withdrawn from the market. It has been proposed that troglitazone is metabolized to a reactive metabolite that covalently binds to cellular macromolecules according to identification of GSH-adducts of the troglitazone metabolites from microsomal incubation system with GSH and bile from troglitazone-treated animals (Kassahun et al., 2001, Tettey et al., 2001, Prabhu et al., 2002, He et al., 2004). One of the GSH-adducts involved oxidative cleavage of the thiazolidinedione ring, potentially generating highly electrophilic α-ketoisocyanate and sulfenic acid intermediates, which was catalyzed by CYP3A4 (Kassahun et al., 2001). Another involved oxidation of the substituted chromane ring system to a reactive o-quinone methide derivative, which was mediated by unidentified P450 enzymes (Kassahun et al., 2001). In relation to the covalent binding of reactive metabolite to cellular macromolecules, cytotoxicity of troglitazone and its metabolites have been observed in various cell systems including rat hepatocytes, HepG2 cells and human hepatocytes (Haskins et al., 2001, Tettey et al., 2001, Yamamoto et al., 2001, Hewitt et al., 2002, Tirmenstein et al., 2002). However, reactive metabolites may not contribute to the cytotoxicity according to the following observations: (1) cells possessing drug-metabolizing enzymes, e.g. rat hepatocytes were less sensitive than cells expressing lower P450s, e.g. HepG2 cells (Tettey et al., 2001); (2) human hepatocytes with higher CYP3A4 content were less susceptible to the cytotoxicity (Hewitt et al., 2002); (3) inhibition of P450 did not attenuate toxicity (Yamamoto et al., 2001, Tirmenstein et al., 2002); (4) quinone metabolite, a postulated precursor of the reactive metabolite, was less cytotoxic than troglitazone (Haskins et al., 2001, Tettey et al., 2001).

In contrast to troglitazone, relatively newer thiazolidinedione antidiabetics such as rosiglitazone and pioglitazone are reported to be less hepatotoxic (Isley, 2003, Tolman and Chandramouli, 2003). Because the in vitro studies elucidated similar differences in their cytotoxicities, e.g. troglitazone > rosiglitazone and pioglitazone (Haskins et al., 2001, Yamamoto et al., 2001, Bae et al., 2003, Narayanan et al., 2003, Shishido et al., 2003), the in vitro cytotoxicity tests may provide some mechanistic information relevant to the clinically observed hepatotoxicity. During the development of toxicities, mitochondrial dysfunction was often observed (Haskins et al., 2001, Tirmenstein et al., 2002, Narayanan et al., 2003, Shishido et al., 2003, Bova et al., 2005). Although the mitochondrial dysfunction is helpful to explain the toxicity mechanism independent of a reactive metabolite, the precise mechanism has remained unclear, because data for direct effects of troglitazone or other thiazolidinediones on the function of isolated mitochondria have not been available. Recent studies have suggested a pathogenetic role of mitochondrial permeability transition (MPT) in the mitochondria-mediated hepatocyte injury by chemical agents (Bernardi, 1996, Lemasters et al., 1998, Masubuchi et al., 2002, Masubuchi et al., 2005). MPT is characterized by a progressive permeabilization of the inner mitochondrial membrane dependent on the excessive amount of intramitochondrial Ca²⁺ and results in mitochondrial swelling, decrease in mitochondrial ΔΨ and release of accumulated Ca²⁺ (Bernardi, 1996, Lemasters et al., 1998). In the present study, we investigated potencies of thiazolidinediones to induce MPT by using mitochondria isolated from mouse liver as a determinant of their cytotoxicities, which may be relevant to the clinically observed hepatotoxicity.