Bortezomib was purchased from Biovision (Mountain View, CA, USA). APAP, CCl₄, TAA, MG132, sodium 4-phenylbutyrate (4-PBA), and 2-aminoethyl diphenylborinate (2-APB) were purchased from Sigma-Aldrich (St. Louis, MO, USA). The primary antibodies used in this study were anti-CYP2E1 (AB1252; Millipore, Bedford, MA, USA), anti-connexin 32 (CX32; 35-8900; Invitrogen, Carlsbad, CA, USA), anti-glyceraldehyde-3-phosphate dehydrogenase (GAPDH; MAB374) (Millipore, Billerica, MA, USA), anti-binding immunoglobulin protein (BiP; 3177; Cell Signaling Technology, Beverly, MA, USA), anti-CCAAT-enhancer-binding protein homologous protein (CHOP; 5554; Cell Signaling Technology) and anti-β-actin (A5316; Sigma-Aldrich).

Male C57BL/6J mice, which were 6–8 weeks of age, were purchased from Orient Bio, Inc. (Seoul, Korea) and housed under special pathogen-free conditions. All animals were treated in accordance with the Animal Care Guidelines of Ewha Womans University. To induce hepatotoxicity, mice were injected intraperitoneally with the following: APAP (350 or 500 mg/kg), TAA (200 mg/kg), or CCl₄ (2 ml/kg), as described previously (10). Before being injected, mice were fasted overnight. To inhibit the proteasome, bortezomib was injected twice: first at 12 h (1 mg/kg) and then at 1 h (1 mg/kg) prior to the injection of APAP, as previously described (11). The inhibitor MG132 was injected (5 mg/kg) twice: first at 12 h and then at 1 h prior to the administration of APAP, as previously described (12,13). In some cases, different doses of bortezomib (0-1 mg/kg) were injected twice: first at 12 h and then at 1 h prior to the administration of APAP, and 1 mg/kg bortezomib was injected at different times prior to the administration of APAP. For the inhibition of gap junctions, 2-APB (20 mg/kg) was injected 2 h prior to the administration of APAP. DMSO-treated mice were used as the relevant control group. Each group consisted of 4–6 mice. For liver extraction, the mice were sacrificed at 0, 2, 4 6 h after the APAP injection and, at 24 h after the TAA or CCl₄ injection. The livers were then perfused with PBS to remove the blood via portal vein.

Human hepatocarcinoma Hep3B cells were cultured in Dulbecco's modified Eagle's medium (DMEM) supplemented with 10% (v/v) heat-inactivated foetal bovine serum, penicillin, and streptomycin (Gibco, Carlsbad, CA, USA). Cells were maintained at 37°C in a humidified atmosphere containing 5% CO₂. Prior to RNA extraction for expression studies or luciferase assays, cells were treated with bortezomib (10-250 nM) (14,15) in serum-free medium for 15 h. To reduce ER stress, 4-PBA (1–5 mM) was added 1 h before bortezomib treatment.

Serum aspartate aminotransferase (AST) and alanine aminotransferase (ALT) levels were measured by the Korean Animal Clinical Research Center (Guri, Korea) using a Hitachi 7020 automatic biochemical analyser (Hitachi, Tokyo, Japan).

Total mRNA from the liver tissues was extracted using an RNeasy Mini kit (Qiagen, Valencia, CA, USA), and cDNA was prepared from the mRNA using a Verso cDNA synthesis kit (Thermo Fisher Scientific, Waltham, MA, USA) according to the manufacturer's instructions. Primer sets are described in Table I. Relative gene expression was calculated as 2^−ΔΔCt by quantitative PCR, using the SYBR Green PCR Master Mix (Applied Biosystems, Warrington, UK) and an ABI PRISM 7500 Sequence Detection system (Applied Biosystems) (16,17).

Liver tissues were lysed in radioimmu-noprecipitation assay buffer [50 mM Tris-HCl, pH 7.5, 150 mM NaCl, 1% (v/v) Nonidet P-40, 0.5% (w/v) sodium deoxycholate, 0.1% (w/v) SDS] containing 50 mM NaF, 2 mM Na₃VO₄, and protease and phosphatase inhibitors (Sigma-Aldrich). Protein concentration was quantified using Bradford assay reagent (Bio-Rad Laboratories, Hercules, CA, USA). Protein samples (40 µg) were subjected to SDS-polyacrylamide gel [8% (w/v)] electrophoresis and then transferred to nitrocellulose membranes. Membranes were incubated with primary antibodies followed by peroxidase-conjugated secondary antibodies (111-035-003 and 115-035-003; Jackson ImmunoResearch, West Grove, PA, USA) after blocking with 5% (w/v) BSA in 20 mM Tris-HCl at pH 7.5, 500 mM NaCl, and 0.1% (v/v) Tween-20. The immunocomplexes were detected by chemiluminescence using SuperSignal West Pico chemiluminescent substrate (Thermo Fisher Scientific) and detected with a Bio-Imaging Analyzer (LAS-4000; Fuji, Tokyo, Japan).

Liver tissues were fixed in 4% (w/v) paraformaldehyde, embedded in paraffin, sectioned at 4-µm thickness, and then stained with H&E using standard methods. Briefly, following deparafinization, the sections were stained with hematoxylin (Sigma-Aldrich) and then destained with acid ethanol (1 ml hydrochloric acid + 99 ml 70% ethanol). After washing with tap water, sections were stained with eosin (Sigma-Aldrich).

Levels of GSH in the fresh liver were measured using a GSH assay kit (Biovision) according to the manufacturer's instructions. Briefly, 40 mg of liver tissues were homogenized with cold glutathione assay buffer, and 6 N perchloric acid was added. The samples were then precipitated with 3 N potassium hydroxide, and then centrifuged for 2 min at 13,000 x g. The neutralized samples were incubated with the o-phthalaldehyde probe and glutathione assay buffer for 40 min to detect reduced GSH. Fluorescence was deteced using a microplate fluorescence reader (BioTek Synergy H1; BioTek, Winooski, VT, USA).

In order to measure CYP2E1 activity, the rate of transformation of p-nitrophenol to p-nitrocatechol was analysed with isolated hepatic microsomes, as described previously (18,19). Briefly, the liver microsomes were incubated at 37°C in assay buffer containing potassium phosphate buffer (50 mM, pH 7.4), NADPH (1 mM) and p-nitrophenol (0.1 mM), at the final concentrations indicated. After 20 min, the reactions were terminated by the addition of 0.6 N perchloric acid, and centrifuged at 10,000 × g for 5 min. Subsequently, 6 N sodium hydroxide was added to the supernatant, and the absorbance was measured at 546 nM using a microplate reader (BioTek Synergy H1; BioTek).

In the present study, a 540-bp human CYP2E1 gene promoter using 5′-TAGGTACCCAGAAGTGAGATTCCTGTTCT-3′ and 5′-CCCAAGCTTTGCCGATGGGGCTCCACTCT-3′ as primers was subcloned into the corresponding restriction sites of the luciferase reporter pGL3-basic vector (Promega, San Luis Obispo, CA, USA), as described in a previous study (20). The underlined letters denote the corresponding restriction sites, KpnI and HindIII The subcloned sequences were verified by DNA sequence analyses.

The luciferase assay was performed with the Dual-Luciferase Reporter assay system (Promega) as previously described (21). Reporter plasmid containing the CYP2E1 promoter region was transfected into Hep3B cells using Metafectene transfection reagent (Biontex, San Diego, CA, USA USA), and pRL-cytomegalovirus promoter (Renilla luciferase vector; Promega) was co-transfected to normalise the transfection efficiency. If necessary, pcDNA5/FRT/HNF-1α, obtained from Addgene (Addgene plasmid 31104), or its control vector pcDNA5/FRT/TO was transfected. After 24 h, each well was treated with bortezomib or 4-PBA, as indicated in the figure legends. Cells were lysed using Promega lysis buffer (Promega). Firefly and Renilla luciferase activities were measured sequentially in the same sample using a GloMax™ 20/20 Luminometer (Promega). Luciferase activities were normalised by dividing the firefly luciferase activity by the Renilla luciferase activity.

Values are expressed as the means ± SEM. Statistical significance was determined using the Student's t-test, and a p-value <0.05 was considered to indicate a statistically significant difference.

In order to determine the effect of bortezomib on DILI induced by APAP, we injected bortezomib intraperitoneally twice: first at different time points (1 mg/kg) (as indicated in Fig. 1A) and then at 1 h (1 mg/kg) prior to the administration of APAP. Injecting 500 mg/kg APAP markedly elevated serum AST and ALT (Fig. 1). When bortezomib was injected 2 h before APAP administration, serum AST and ALT levels were greatly elevated, similar to the DMSO-treated group (Fig. 1A). Treatment with bortezomib for more than 3 h was needed to diminish the APAP-induced hepatotoxicity, as indicated by decreased serum AST and ALT levels and H&E staining (Fig. 1A and C). In addition, bortezomib decreased APAP-induced hepatotoxicity in a dose-dependent manner, and this protective effect was noted at concentrations greater than 0.063 mg bortezomib/kg (Fig. 1B and D).

Since the toxic mechanism differs between various drugs and chemicals, we investigated whether bortezomib treatment also protects against liver damage caused by other chemicals. Surprisingly, bortezomib treatment also alleviated CCl₄- and TAA-induced hepatotoxicity (Fig. 2). Thus, we posit that the common toxic mechanism of all three chemicals involves CYP2E1 (22–24) or gap junction function (25).

GSH is an important antioxidant that detoxifies NAPQI by conjugation, and GSH depletion leads to toxic cascades of NAPQI (3,4). Basal levels of GSH did not differ significantly between DMSO- and bortezomib-treated mice (Fig. 3A). Although GSH was markedly decreased in both DMSO- and bortezomib-treated groups following APAP administration, GSH levels remained higher at all time-points in bortezomib-versus DMSO-treated animals (Fig. 3A). Lower GSH depletion in bortezomib-treated mice implies the decreased formation of NAPQI. Since NAPQI is transformed by CYP2E1, we subsequently measured CYP2E1 expression by both RT-qPCR (Fig. 3B) and western blot analysis (Fig. 3C). Bortezomib treatment significantly decreased CYP2E1 transcription (Fig. 3B), which resulted in lower CYP2E1 protein levels (Fig. 3C). In addition, CYP2E1 enzymatic activity was significantly lower in the livers of bortezomib-treated versus control mice (Fig. 3D). Therefore, we conclude that the decreased CYP2E1 enzyme activity caused by bortezomib administration alleviates liver damage induced by APAP, CCl₄, and TAA exposure. This is in agreement with results of previous studies, in which a CYP2E1 deficiency prevented hepatotoxicity induced by these three chemicals (22–24).

In order to confirm whether bortezomib regulates CYP2E1 transcription, we transfected a hepatocyte cell line (Hep3B cells) with the reporter plasmid containing the CYP2E1 gene promoter region, and performed luciferase assays. Bortezomib treatment significantly decreased CYP2E1 promoter activity by itself (Fig. 4A) as well as HNF-1α-induced CYP2E1 promoter activity (Fig. 4B). HNF-1α is the main transcription factor which positively regulates CYP2E1 gene transcription (26–28).

Since bortezomib can lead to ER stress in vitro (29,30), which can induce decreased CYP2E1 mRNA levels (31), we next investigated whether bortezomib administration also causes ER stress in the liver in vivo. We examined altered levels of several proteins that were reported to be induced during ER stress. BiP, also referred to as glucose-regulated protein 78, is a central regulator of ER stress (32), and CHOP is one of the most highly induced transcription factors during ER stress (33). Bortezomib injections markedly elevated hepatic mRNA and protein levels of BiP and CHOP, indicating the in vivo induction of ER stress (Fig. 5). IRE-1α protein expression was also increased 12 h after bortezomib treatment (Fig. 5C). Subsequently, we used 4-PBA to confirm whether ER stress is directly involved in bortezomib-induced CYP2E1 decrement. Treatment with 4-PBA (5 mM), which is reported to decrease ER stress (34), partially recovered CYP2E1 luciferase activity decreased by bortezomib treatment (Fig. 6). Therefore, bortezomib-induced ER stress partially contributes to decreased CYP2E1 levels.

In the present study, in order to investigate whether this effect of bortezomib is due to proteasome inhibition, we injected MG132, another well-known proteasome inhibitor, before APAP administration. By contrast to bortezomib treatment (Fig. 7A and B), MG132 aggravated APAP-induced hepatotoxicity, as demonstrated by elevated serum AST and ALT levels (Fig. 7C) and by enhanced H&E staining (Fig. 7D). These results suggest that the protective effect of bortezomib on APAP-induced hepatotoxicity is not derived from direct inhibition of the proteasome.

Since we noted that MG132, a second proteasome inhibitor, aggravated APAP-induced liver injury, by contrast with bortezomib (Fig. 1), we subsequently examined the effect of MG132 on CYP2E1 expression and ER stress. Unlike treatment with bortezomib, hepatic mRNA levels of CYP2E1 were not diminished upon MG132 administration (Fig. 8A), and we also noted relatively unaltered CYP2E1 enzyme activity upon administration of MG132 (Fig. 8B). In addition, it was demonstrated that MG132 administration did not induce ER stress, as indicated by unaltered mRNA levels of BiP and CHOP (Fig. 8C and D).

In addition to CYP2E1 enzyme activity, it has previously been suggested that the common toxic mechanism of all three chemicals (APAP, TAA, and CCl₄) may also involve gap junctions (10,25,35,36), since gap junctions have been reported to play a critical role in the propagation of hepatotoxicity of all three chemicals used in the present study; ablation of CX32 (a key protein in hepatic gap junctions) significantly protects against DILI (10,25,35,36). Thus, we examined CX32 levels. Notably, MG132 administration significantly increased CX32 protein expression (Fig. 9A), possibly by inhibiting its degradation, and 2-APB, a gap junction inhibitor, abolished MG132-aggravated APAP-induced liver damage (Fig. 9B). Unexpectedly, bortezomib decreased CX32 expression, as well as CX43 mRNA levels (Fig. 9C and D). Therefore, the opposite effects of bortezomib and MG132 on CX32 levels may contribute to the different impacts of bortezomib and MG132 on DILI.