To investigate its hepatotoxic mechanism at the molecular level, RIF was orally administered to mice at doses of 177 and 442.5 mg/kg (LD₁₀ and LD₂₅) for 14 days (Figure 1). To examine the general toxic effect of RIF, changes in body and liver weights were determined during the 14-day experimental period. As shown in Table 1, body weight was not affected by RIF administration. However, liver weights increased significantly by 1.7- and 2.7-fold compared with those of the vehicle control groups after treatment for 14 days with 177 and 442.5 mg/kg RIF, respectively. The liver weights relative to body weights also increased significantly by 1.7- and 2.2-fold for 177 and 442.5 mg/kg RIF, respectively. The dose-dependent increases of liver weights with both 177 and 442.5 mg/kg RIF indicate the initiation of a liver abnormality like fat accumulation.

The hepatotoxicity induced by RIF was investigated based on blood parameters after oral administration for 14 days (Table 2 and Table S1). In response to 177 and 442.5 mg/kg RIF, alanine aminotransferase (ALT) activity, an indicating parameter of liver specific damage, in mice increased significantly by 2.1- and 3.4-fold compared to that of the vehicle group, respectively (Table 2). The other parameters for detecting liver damage, the activity levels of aspartate transaminase (AST) and lactate dehydrogenase (LDH), increased significantly by 2.9- and 3.1-fold, respectively, for 442.5 mg/kg RIF (Table 2 and Table S1). However, the activity of alkaline phosphatase (ALP), another parameter of liver damage derived from abnormal biliary excretion, did not increase after RIF administration for 14 days. Additionally, RIF resulted in hyperbilirubinemia; total bilirubin in the blood increased by 4.5- and 9-fold for doses of 177 and 442.5 mg/kg, respectively, while the triglyceride level with 442.5 mg/kg RIF was a half of that of the vehicle group. Total cholesterol increased significantly by 2.3- and 3.7-fold for 177 and 442.5 mg/kg RIF, respectively. Low-density lipoprotein (LDL) cholesterol increased by 4.4- and 5.2-fold, but HDL-C did not increase (Table S1). Following on Roussel Uclaf Causality Assessment Method (RUCAM) classification, the R value (R = ALT/ALP) at 177 and 442.5 mg/kg were calculated as 2.1 and 3.4, respectively. We defined the hepatotoxic mechanism of RIF as mixed liver injury because the R values were estimated at the range of 2 < R < 5 [14]. These patterns of parameter changes after RIF administration in mice were consistent with those of a previous study [15].

The largest and the second largest lobes were cut from the mouse livers and slides were prepared. They were H&E-stained and observed at 100× and 200× magnification (Figure 2). We observed extreme hepatocyte hypertrophy characterized by a notable increase in cell size accompanied by binucleate hepatocytes with enlarged hepatocyte nuclei.

RIF is an agonist of pregnane X receptor (PXR), and PXR protein overexpression and activation by RIF increases the metabolism of cytochrome P450 (CYP) 3A4 substrates [16]. Accordingly, mouse livers were homogenized to produce the S9 fraction and the activity levels of CYP1A, 2B, 2C, 2D, and 3A were determined by a cocktail probe assay [17]. As shown in Figure S1A, CYP2C-mediated omeprazole 5-hydroxylation increased significantly by 3.4-fold compared with that of the vehicle control group after 14 days of treatment with 442.5 mg/kg RIF. CYP3A-catalyzed midazolam hydroxylation increased slightly after RIF treatment in a dose-dependent manner.

To investigate the mechanism of toxicity in mouse livers after the administration of 177 and 442.5 mg/kg RIF, liver proteins were analyzed by global proteomics coupled with a chemical tagging method (Figure 1). The pooled protein homogenates were trypsin-digested, and the tryptic peptides were subjected to 3-plex isobaric TMT tag analyses by 2D-LC using a hybrid quadrupole-orbitrap mass spectrometer. All MS/MS spectra were analyzed using MaxQuant software, allowing a maximum FDR of 1% for the proteins and peptides. To obtain a high-quality protein list, we excluded contamination and results with less than one unique peptide/protein from the total protein list.

In total, 1101 proteins were identified, among which 1038 proteins were quantified. All data consisted of technical duplicates. Proteins were divided into two categories based on relative levels; those with a quantitative ratio of greater than 1.5 were considered up-regulated, while those with a quantitative ratio of less than 0.667 were considered down-regulated (Table S2).

A total of 29 and 40 proteins were up-regulated, and 27 and 118 proteins were down-regulated after RIF administration at doses of 177 and 442.5 mg/kg, respectively (Table S2). To further characterize the mechanism of RIF-induced hepatotoxicity, we performed Gene Ontology (GO) term and Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway enrichment analyses to identify the biological processes, cellular components, and molecular functions that were enriched in RIF-treated liver samples (Figure 3 and Figure 4, Figures S2 and S3). With respect to molecular function, glutathione transferase and protein dimerization activities were up-regulated in a dose-dependent manner, and steroid hydroxylase, arachidonic acid epoxygenase, and monooxygenase activities were down-regulated in a dose-dependent manner (Figure 3). In the biological process category, up-regulated proteins were involved in chromosome and chromatin organization and DNA conformational changes (Figure S2A). In addition, nuclear and chromosome-related proteins were up-regulated and cytosolic and ribosomal proteins were down-regulated in RIF-treated livers (Figure S2B).

Based on the KEGG pathway enrichment-based clustering analysis, the peroxisome proliferator-activated receptor-γ (PPARγ) signaling pathway, metabolism of xenobiotics by CYP, glutathione metabolism, chemical carcinogenesis, retinol metabolism, drug metabolism, and alcoholism-related proteins increased in a dose-dependent manner in RIF-treated livers (Figure S3). The up-regulated proteins list is shown in Table S3. In particular, significant protein enrichment was observed in the KEGG reference pathways, as shown in Figure 4. The PPARγ signaling pathway was enriched in the up-regulated group, including apolipoprotein C-III, acyl-CoA-binding protein, 3-ketoacyl-CoA thiolase A and B, and perilipin-2, which increased, but not dose-dependently. Pathways for drug metabolism, the glutathione pathway, and alcoholism were enriched for up-regulated genes and these increases were dose-dependent. The up-regulated proteins after RIF administration were CYP2A5, 2B10, 2C55, and 3A11 in the metabolism pathway and glutathione S-transferases (GST) Mu 1, 2, and 3, A1, and theta 2 in the GST activity pathway. GST are important phase II metabolic enzymes to protect the cells from oxidative stress and a super-gene family with detoxification function. Recent studies found that the content of GST was increased in diseased liver tissues and serums [18]. In addition, two pathways involved in systemic lupus erythematosus and alcoholism included the histones H2A, H2B, H3, and H4.

Pathogen-free male ICR mice were purchased from Orient Bio Inc. (Seongnam, Korea). Upon receipt, mice were randomly divided into groups of four to five in cages and were acclimated for one week in strictly controlled conditions at 23 ± 3 °C and 50 ± 10% relative humidity, with a 150–300 lux light source on a 12 h light/dark cycle. All animal procedures followed the guidelines recommended by the Society of Toxicology (Reston, VA, USA) in 1989 and were approved by the Institutional Review Board at Kyungpook National University (approval No. 2016-57-2, approval date 28 June 2016).

The 50% lethal dose (LD₅₀) of RIF is 885 mg/kg in mice; accordingly, RIF (Yuhan Co., Seoul, Korea) was administered orally to animals at 0, 177 (LD₁₀), and 442.5 mg/kg(LD₂₅) once a day for two weeks. RIFwas administered in the form of a mixture of solvents consisting of 5%(v/v) DMSO and 25%(v/v) polyethylene glycol (PEG) in water [43].

After 14 days of administration, mice were sacrificed to obtain blood and liver samples. Prior to necropsy, all animals were fasted overnight for approximately 18 h. The animals were anesthetized with isoflurane (Ifran liquid, Hana Pharm, Co., Ltd., Seoul, Korea) and blood samples of approximately 1 mL were collected from the abdominal aorta. The blood was placed in evacuated tubes containing K2-EDTA as anticoagulant (BDTM Vacutainer, Becton Dickinson, Franklin Lakes, NJ, USA) and used for the plasma biochemical tests.

To obtain histology slides of the liver, the entire liver was weighed and two 1-cm liver samples were taken from the two major lobes. They were put in vials with Formalin Solution, neutral buffered, 10% (Sigma Aldrich, St. Louis, MO, USA), and fixed at room temperature. Histology slides were produced at Histoire (Ansan, Korea). Slides were stained with H&E and observed at 100× and 200× magnification.

For proteomic profiling, the mouse livers were added to four volumes (w/v) of 0.1 M phosphate buffer (pH 7.4) with 10 mM phenylmethylsulfonylfluoride (PMSF) and manually homogenized using a glass homogenizer (Wheaton, #357538, Millville, NJ, USA) until completely broken. Liver homogenates were stored at −80 °C for the prevention of protein degradation until they were used. A Bradford assay (BioRadLaboratories, Hercules, CA, USA) was performed to measure the protein concentration of homogenates.

In total, 100 µg of protein samples were pooled with the same amount for biologically replication, according to Figure 1. For protein precipitation, the liver homogenates (200 µg of protein in 90 µL of 0.1 M phosphate buffer) were added 10 µL of trichloroacetic Acid (TCA) and incubated for 4 h at 4 °C. The precipitated samples were centrifuged 12,000× g for 10 min at 4 °C and withdraw the supernatant. The protein pellets were washed by 500 µL of ice-cold acetone and centrifuged again as described twice. Finally, protein pellets were dried by speed-vacuum (Lebconco, Kansas City, MO, USA) for 1 min to remove acetone and suspended in 100 mM triethyl ammonium bicarbonate (TEAB) to obtain a final volume of 100 μL. Then, 5 μL of 200 mM trisphosphine hydrochloride was added to each sample and incubated at 55 °C for 1 h for reduction. Next, 5 μL of 375 mM iodoacetamide was added to the sample, followed by incubation at room temperature for another 30 min with light protection. After incubation, the samples were mixed with pre-chilled (−20 °C) acetone and frozen at −20 °C until the formation of a precipitate. For the digestion of the acetone-precipitated protein pellet, the pellet was resuspended in 100 μL of TEAB, and re-measured protein amount using Bradford assay. After reduction/alkyation, 50 μg of proteins were transferred at new tubes and 2.5 μg of trypsin was added to each sample for overnight digestion at 37 °C.

According to procedure, 0.8 mg of 6-plex-TMT Label Reagent (Pierce Biotechnology, Rockford, IL, USA) in 41 μL of acetonitrile was added to trypsin-digested peptides for the chemical tagging of samples (vehicle control, TMT-126 and 129; low dose RIF, TMT-127 and 130; High dose RIF, TMT-128 and 131), and the reaction was processed for 1 h at room temperature. The reaction was stopped by adding 8 μL of 5% hydroxylamine. Each sample was collected and mixed in a new tube and fractionated using the Pierce™ High pH Reversed-Phase Peptide Fractionation Kit (Pierce Biotechnology). After elution, samples were dried by vacuum centrifugation and then desalted by C18 Ziptip (Millipore, Darmstadt, Germany) according to procedure. The desalted samples were dried by vacuum centrifugation and resuspended in 25 μL of 0.1% formic acid solution before they were analyzed by LC-MS/MS.

Chromatographic separation was performed using a custom-made capillary column (10 cm length, 75 μm internal diameter) packed with Jupiter C12 resin (4 μm particle size, 90 Å pore size; Phenomenex Inc., Torrance, CA, USA). The NanoLC-1D Plus HPLC System (Eksigent Technologies LLC, Dublin, CA, USA) was used for gradient elution at a constant flow of 300 nL/min. The LC solventswere as follows: A (water, 0.1% formic acid) and B (acetonitrile, 0.1% formic acid). The mobile phase was programmed as follows for gradient elution: (minutes, %B) = (0, 5); (12, 5); (40, 10); (50, 20); (74, 90); and (76, 90). The MS/MS analysis was performed using a LTQ Orbitrap Velos Mass Spectrometer (Thermo, Waltham, MA, USA) equipped with a nanospray ion source operating in positive ion mode with the following settings: nebulizer gas at 0 (arbitrary units), curtain gas at 0 (arbitrary units), auxiliary gas at 0, ion spray voltage 1.8 kV, capillary temperature 300 °C, collision energy at 40 with HCD mode. MS data were recorded from m/z 100 to 2000 with an accumulation time of 1 s and a pause between mass ranges of 0.5 milliseconds, operating in positive mode and MS resolution was set at 70,000. MS/MS was operated using a top-15 data-dependent method with 7500 of MS/MS resolution. For all experiments, the dynamic exclusion time was set to 5 s.

Peptides and proteins were identified using MaxQuant 1.5 [44] with a precursor mass error of 10 ppm, monoisotopic mass selected, and a fragment ion mass error of 20 ppm. The enzyme was digested with trypsin allowing two missed cleavages. Cysteine carbamidomethylation (57.021 Da) and TMT-3plex on peptide N-term and lysine (229.163 Da) was searched as a fixed modification. Methionine oxidation (15.995 Da) and protein N-term acetylation (42.011 Da) were searched as a variable modification. A decoy search was performed, and peptides were filtered using a false discovery rate (FDR) of 0.01 and remove the reverse and potential contaminant proteins from total identified proteins. Positive identification of proteins required a minimum of two unique peptides.

For quantification of proteins by TMT labeling, the reporter ions are calculated by TMT-3plex method at MaxQaunt. All the other parameters in MaxQuant were set to default values. In addition, quantified proteins were defined that up to 1.5 ratio is up-regulated proteins and ratio of below 0.666 is down-regulated proteins.

The mass spectrometry proteomics data have been deposited to the ProteomeXchange Consortium via the PRIDE [45] partner repository with the dataset identifier PXD006313.

The GO annotation proteome was obtained from the UniProt GO database (http://www.ebi.ac.uk/GOA/). The identified protein ID was converted to the UniProt ID and mapped to GO IDs. If identified proteins were not annotated in the UniProt GOA database, InterProScan was used to annotate GO functions using a protein sequence alignment method. Then, proteins were classified based on GO annotation into three categories: biological process; cellular component; and molecular function. For KEGG annotation, the KEGG database was used to annotate protein pathways. First, the KEGG online service tool KAAS was used to annotate proteins based on KEGG database descriptions. Then, the annotation results were mapped to the KEGG pathway database using the KEGG online service tool KEGG mapper.

For each category, a two-tailed Fisher’s exact test was employed to test the enrichment of each differentially expressed protein against all identified proteins. Correction for multiple hypothesis testing was performed using standard false discovery rate control methods. GO results with a corrected p-value of <0.05 were considered significant. The KEGG database was used to identify enriched pathways by a two-tailed Fisher’s exact test to determine the enrichment of each differentially expressed protein against all identified proteins. Correction for multiple hypothesis testing was performed using standard false discovery rate control methods. Pathways with a corrected p-value of <0.05 were considered significant. These pathways were classified into hierarchical categories according to the KEGG website.

The hepatotoxic parameters were analyzed using Statistical Package for Social Sciences software (SPSS 22.0 K, IBM, Seoul, Korea), and all values are presented as mean ± standard error (S.E). All statistical analyses were performed with an Independent t-test that was performed to compare the results at the termination of the experiment. Statistical significance was set at p < 0.05 and p < 0.01.