Establishment of a mouse model of troglitazone-induced liver injury and analysis of its hepatotoxic mechanism

概要

【Introduction】
　Drug-induced liver injury (DILI) is a major problem in drug development and clinical drug therapy. Troglitazone (TGZ), a thiazolidinedione antidiabetic drug for the treatment of type II diabetes, was found to induce rare idiosyncratic severe liver injury in patients, which led to its withdrawal in 2000. L-buthionine-(S,R)-sulfoximine (BSO), an inhibitor of glutathione (GSH) synthesis, is a compound that commonly caused glutathione depletion. In vivo studies have shown that there is enhanced susceptibility of hepatocytes to necrosis in DILI model, induced by the co-administration of BSO with carbamazepine or methimazole. To date, there is still no evidence of a successful model with normal mice in TGZ-induced liver injury. In this study, I attempted to establish a new model to investigate whether TGZ causes liver injury in normal mice and aimed to find the mechanisms of TGZ hepatotoxicity.

【Materials and methods】
　Female BALB/c and C57BL/6 mice (6-week-old) were fasted for 12 h before each experiment. Rosiglitazone (RGZ) was used as a negative drug of TGZ in this study. Mice were randomly divided into five groups: vehicle, TGZ, BSO+TGZ, RGZ, and BSO+RGZ. BSO (700 mg/kg, 10 mL/kg) was administered intraperitoneally to mice 1 h before TGZ (30, 100, and 300 mg/kg, 20 mL/kg) or RGZ (300 mg/kg, 20 mL/kg) administration. The control animals were given the vehicle (10 mL/kg of saline and 20 mL/kg of 5% DMSO and 0.1% CMC). At 6 h or 24 h after TGZ administration, animals were anesthetized with isoflurane and sacrificed.
　Plasma biochemical analyses were performed to measure alanine aminotransferase/ aspartate aminotransferase (ALT/AST) levels, and high-mobility group box 1 (HMGB1) contents. Hepatic GSH content was measured. Paraffin-embedded tissue sections were processed for hematoxylin and eosin. Microarray experiments were investigated for gene analysis. Hepatic mRNA and plasma miR-122-5p expression were determined through real-time RT-PCR. The expression of total and phosphorylated signal transducer and activator of transcription (STAT) 1/3 in the liver was detected by western blot.

【Results】
　The plasma ALT/AST levels showed dose-dependent increases 6 h after TGZ administration in BALB/c mice, and a significant increase at 300 mg/kg of TGZ compared with the vehicle control, which was back to normal again at 24 h. Moreover, BALB/c mice were more sensitive than C57BL/6 mice because of the higher ALT/AST levels (Figure 1). Therefore, female BALB/c mice were used for further experiments. Histopathological examination of the liver showed hepatocellular injury at 24 h, but not 6 h after TGZ dosing, which suggested that the process of the mouse model was irreversible (Figure 2).
　Significant increases in the expression level of the plasma miR-122-5p (liver specific miRNA) and HMGB1 (predictive biomarker for DILI) were observed in mice treated with TGZ alone (Figure 1). Significant decreases in hepatic total GSH, GSH, and GSH/disulfide GSH (GSSG) (oxidative stress marker) were observed in TGZ-treated mice (Figure 3).
　To investigate the mechanisms by which TGZ modified expression of genes, gene set enrichment analysis (GSEA) on microarray data using the gene ontology (GO) gene sets were carried out on the hepatic transcriptome profiles (Figures 4 and 5). This analysis showed that 48 and 14 gene sets were significantly upregulated by TGZ and RGZ groups, respectively. Meanwhile, 1 and 11 gene sets were significantly downregulated by TGZ and RGZ groups, respectively. Among the 48 gene sets upregulated in TGZ group, 44 gene sets were specific to the TGZ group. Interestingly, I found that there were 4 gene sets correlated with STAT cascade among the 44 gene sets specific to the TGZ group (STAT cascade, regulation of STAT cascade, positive regulation of STAT cascade, and regulation of tyrosine phosphorylation of STAT protein), which provided an important evidence that the STAT signaling pathway might be involved in TGZ-induced liver injury.
　The hepatic mRNA expression levels of inflammatory-, oxidative stress-, damage- associated molecular pattern molecules (DAMPs) and Janus kinase (JAK)/STAT3 signaling pathway-related factors were significantly enhanced at 6 h in the TGZ alone- and BSO+TGZ co-administered groups, but not in the RGZ alone- and BSO+RGZ co- administered groups, compared with the vehicle control (Figure 6). These results indicated the induction of inflammation, oxidative stress, DAMPs and STAT3 signaling pathway in TGZ-exposed liver. Western blot results showed the phosphorylation of STAT3 (not STAT1) in TGZ alone- and BSO+TGZ co-administered mice, which further proved that the activation of STAT3 signaling pathway in TGZ-induced liver injury (Figure 7).
　To investigate the unique genes related to toxicity of TGZ, the up- and down-regulated genes were picked up through TGZ vs. BSO+TGZ (Table 1). I found that the log2 fold change of ryanodine receptor (Ryr) 2 was much higher in TGZ (5.27) than that of BSO+TGZ (0.10). To determine the effect of Ryr in TGZ-induced liver injury, mice were pre-treated with dantrolene sodium (DAN, Ryr inhibitor) before TGZ administration. The plasma ALT level and hepatic mRNA expression of Ryr2 were significantly decreased in DAN+TGZ comparing with TGZ alone group (Figure 8). These results suggested that inhibition of Ryr activity alleviated the liver injury induced by TGZ.

【Discussion】
　The present study aimed to establish a new mouse model of TGZ-induced liver injury and analyze the mechanisms of TGZ hepatotoxicity. Strain differences as important factors also influence the evaluation of toxicity because of their different responsiveness or susceptibility of chemicals among mouse strains. Female BALB/c and C57BL/6 mice were chosen for evaluation because the strains are generally used in pre-clinical safety studies for drug development, and BALB/c mice particularly showed higher sensitivity to TGZ than C57BL/6 mice.
　To identify changes in functional gene group expression, GSEA, the powerful analytical method was used in the current paper. Our GSEA results showed that TGZ-induced liver injury had its correlation with the expression of gene sets associated with the STAT cascade. Activation of the JAK/STAT signaling pathway (especially STAT3 activation) might promote the liver damage induced by TGZ.
　In vivo studies have shown that BSO pretreatment could enhance DILI caused by carbamazepine or methimazole, while, an opposing effect was observed in this study. This phenomenon was also found in chimeric-mice with humanized liver model of TGZ-induced liver injury. However, there is no enough evidence to explain the mechanism of BSO in the hepatoprotection of TGZ -induced liver injury. Previous study showed that elevated GSSG enhanced sarcoplasmic reticulum Ca2+ release regulated by Ryr. Ryr2 as an isoform of Ryr regulates Ca2+ release from the sarcoplasmic reticulum and massive Ca2+ release induces apoptosis. The GSSG content, ALT level and Ryr2 mRNA expression were significantly decreased in BSO+TGZ vs. TGZ. In addition, Ryr inhibitor (DAN) pretreatment also decreased ALT levels and Ryr2 mRNA expression in DAN+TGZ vs. TGZ. These novel results might provide new insights into Ryr activity in the regulation of TGZ- induced liver injury (Figure 9).

【Conclusion】
　I established a mouse model of TGZ-induced liver injury with female BALB/c mice. TGZ caused inflammation, oxidative stress, DAMPs, HMGB1, and IL-6/JAK/STAT3 signaling pathway activation during the liver injury. Inhibition of Ryr activity alleviated TGZ-induced liver injury. These findings would provide important information to understand the mechanism of idiosyncratic DILI.