Effects of medium‐ and long‐chain fatty acids on acetaminophen‐ or rifampicin‐induced hepatocellular injury

Abstract Drug‐induced liver injury (DILI) is one of the common adverse effects of drug therapy, which is closely associated with oxidative stress, apoptosis, and inflammation response. Medium‐chain fatty acids (MCFA) were reported to relieve inflammation and attenuate oxidative stress. However, little has been known about the hepatoprotective effects of MCFA in DILI. In the present study, acetaminophen (AP) and rifampicin (RFP) were used to establish DILI models in LO2 cells, and the cytoprotective effects of MCFA on hepatocellular injury were investigated. Results showed that the optimal condition for the DILI model was treatment with 10 mM AP or 600 µM RFP for 24 hr. LCFA treatment markedly reduced the cell viability and increased the activities of alanine aminotransferase, aspartate aminotransferase, and lactate dehydrogenase. Meanwhile, LCFA treatment aggravated cell apoptosis, mitochondrial dysfunction, and oxidative stress. The mRNA and protein expression levels of inflammatory cytokines (IL‐1β and TNF‐α) were significantly elevated by LCFA. In contrast, MCFA treatment did not significantly affect cell viability, apoptosis, oxidative, stress and inflammation, and it did not produce the detrimental effects on DILI models. Therefore, we proposed that MCFA may be more safe and suitable than LCFA as nutrition support or the selection of daily dietary oil and fat for the patients with DILI.

algesic and antipyretic drug, may easily induce DILI when it was used overdose (Ramachandran & Jaeschke, 2019). Rifampicin (RFP), known as the first-line antituberculosis drugs, caused DILI when given separately or given in combination with others (Benson et al., 2016). Therefore, AP or RFP (always combined with isoniazid) was a classical drug for establishing the DILI model in vivo or in vitro. In this study, we first determined the conditions of AP and RFP to establish a DILI model in LO2 cells and then investigated the effects of fatty acid (FA) on the hepatocellular injury.
Furthermore, another study found that MCFA attenuated lipopolysaccharide-induced liver injury through down-regulating necroptotic and inflammatory signaling pathways . However, these studies were hardly focusing on the relationship between FA and DILI, and many patients with DILI consumed LCFA (especially oleic acid) in daily oil and fat diets due to lack of scientific references.
In this study, we found that LCFA aggravated the hepatocellular injury. Therefore, this study may contribute to the development of functional oils and provide knowledge on the selection of oil and fat for patients with DILI.

| Materials
LO2 cells (normal human liver cell line HL7702) were obtained from the Shanghai Institute of Biochemistry and Cell Biology, Chinese Academy of Sciences. RPMI-1640 medium was purchased from Hyclone Laboratories Inc. AP, RFP, and fetal bovine serum were from Sigma-Aldrich Co. Sodium caprylate, sodium caprate, sodium laurate, and sodium oleate were from Aladdin Industrial Corp.

| Cellular model of drug-induced liver injury
LO2 cells were grown in RPIM 1640 supplemented with 10% fetal bovine serum and maintained in a humidified CO 2 -regulated incubator with a 5% CO 2 atmosphere at 37°C. The AP stock solution was dissolved in PBS to 200 mM, and the RFP stock solution was dissolved in DMSO to 1,000 µM and diluted by the serum-free medium. LO2 cells were seeded into 96-well plates at a density of 1 × 10 4 cells per well and grew overnight at 37°C in 5% CO 2 . Then, the cells were treated with AP at 5-20 mM or RFP at 200-800 µM. Control cells were treated with vehicle alone. Plates were incubated at 37°C for 24, 48, and 72 hr; then, the cell viability was determined using MTT assay.
In another set of experiments, cells were seeded into 24-well plates at a density of 5 × 10 4 cells per well. After treated as above, plates were incubated at 37°C for 24 hr. Then, the media were collected to measure the activities of AST and ALT using activity determination kits, according to the manufacturer's instructions. Then, the apoptotic cells were measured by double staining with Hoechst 33342/ PI assay. Briefly, the treated cells were washed three times with cold PBS and stained with Hoechst 33342/PI at 4°C for 20 min in the dark, then washed twice with cold PBS and observed under a fluorescence microscope. AST and ALT assay, and apoptosis observation were conducted as confirmation experiments to support the MTT assay.

| Measurement of cell viability
Fatty acid (including octanoic acid, decanoic acid, lauric acid, and oleic acid) were dissolved in PBS to form the stock solution at 8 mM and diluted by the serum-free medium. For the viability assay, LO2 cells were seeded into 96-well plates at a density of 1 × 10 4 cells per well and cultured overnight. Then, the cells were treated with different concentrations of FA (0,50,100,200,400, and 800 µM) for 24 hr. For further studies, LO2 cells were also treated with 20 mM AP and 600 µM RFP for 24 hr and then incubated with 200 µM FA for an additional 24 hr. Then, 10 µl of MTT solution was added to each well and incubated for 4 hr at 37°C. Next, the supernatant was carefully removed, the purple formazan crystals were dissolved in 150 µl DMSO, and the absorbance value at 490 nm was measured by a SpectraMax ® absorbance reader (Molecular Devices).

| Lactate dehydrogenase (LDH) assay
Lactate dehydrogenase assay was conducted as a confirmation experiment to support the MTT assay. The level of LDH released from injured or dead cells was measured as an indicator of cytotoxicity. The LO2 cells were seeded at 1 × 10 4 cells/well onto a 96-well plate for 12 hr and treated as the procedure for MTT assay. The supernatant was collected to measure the activity of LDH using an activity determination kit, according to the manufacturer's instructions. The absorbance at 450 nm was then determined by a SpectraMax ® absorbance reader.

| Apoptosis assay
LO2 cells were seeded at a density of 1 × 10 6 cells per well onto 6-well plates and cultured overnight. Then, the wells were divided into the following groups: The cells were treated with 20 mM AP or 600 µM RFP for 24 hr and then incubated with 200 µM octanoic acid (C8:0 group), decanoic acid (C10:0 group), lauric acid (C12:0 group), and oleic acid (C18:1 group) for an additional 24 hr, the cells treated with 20 mM AP or 600 µM RFP for 24 hr and then incubated with serum-free medium containing PBS or 0.1% DMSO for an additional 24 hr (naturally restoring group, NR group), and the cells treated with serum-free medium (normal control group, NC group).
After the treated cells were rinsed three times with cold PBS and harvested by trypsinization, they were washed twice with cold PBS followed by centrifugation at 112 g for 10 min. Then, the cell pellet was resuspended in 500 µl binding buffer followed by incubation with Annexin V-FITC (5 µl) and PI (5 µl) for 15 min at room temperature in the dark. The excitation wavelength was 488 nm, and the emission wavelength was 525 nm. A flow cytometer with BD Accuri C6 software (FCM, Becton-Dickinson) was used to determine the apoptosis rate after acquiring 10,000 cells. Furthermore, apoptotic hepatocytes were identified by morphology, after the treated cells were washed three times with cold PBS and stained with Hoechst 33342/PI at 4°C for 20 min in the dark, then washed twice with cold PBS and observed under a fluorescence microscope.

| Measurement of mitochondrial membrane potential (MMP)
The treated cells were evaluated for the change in MMP using Rhodamine 123 (Rh-123) cationic dye and a flow cytometer. Briefly, the cells were harvested and resuspended in complete RPMI 1640 media (with 10% FBS) containing Rh-123 (5 µg/ml) and kept at 37°C for 30 min. Next, the cells were washed twice with PBS, followed by centrifugation at 447 g for 5 min to remove the unreacted Rh-123 dye. Finally, the cell pellet was resuspended in 500 µl PBS and detected by flow cytometry. Ten thousand cells were acquired and analyzed by BD Accuri C6 software.

| Determination of intracellular antioxidant indexes
Intracellular antioxidant indexes, including total antioxidant capacity (T-AOC), levels of malondialdehyde (MDA), and glutathione (GSH), and activities of superoxide dismutase (SOD) and catalase (CAT) were determined with assay kits according to the instructions of the manufacturers. Briefly, the cells were harvested and incubated in a chilled cell lysis buffer for 30 min. The cells were centrifuged at 16,099 g for 10 min at 4°C, and the supernatant was collected to measure antioxidant indexes. Total protein content in the supernatant was determined by BCA protein assay kit (Beyotime Biotechnology), using bovine serum albumin as a standard. All assays were conducted at least in six replicates, and values given are average of these replicates.

| Quantitative real-time polymerase chain reaction (q-PCR)
The expression levels of the IL-1β, IL-6, IL-8, MCP-1, and TNF-α mRNAs were analyzed by q-PCR. The total RNA was isolated from a cell sample

| Western blot analyses
The cells were lysed in RIPA buffer (Beyotime Biotechnology) with a protease inhibitor PMSF (Beyotime Biotechnology) and centrifuged Tris-buffered saline Tween-20 (TBST) containing 5% skimmed milk powder for 2 hr at room temperature, the membranes were incubated with different primary antibodies: anti-IL-1β (1:4,000, mouse monoclonal antibody), anti-TNF-α (1:1,500, mouse monoclonal antibody), and anti-β-actin (1:2,000, rabbit polyclonal antibody) at 4°C overnight. Then, the cells were incubated with secondary HRPconjugated goat anti-rabbit/mouse IgG (1:3,000) for 2 hr at room temperature after washed by TBST three times. At last, the protein bands on the membranes were visualized with the enhanced ECL detection system (Bio-Rad) after washed by TBST three times.
Protein bands were quantitatively analyzed using AlphaEaseFC software and normalized to β-actin.

| Statistical analysis
Statistical Package for the Social Sciences 19.0 software (SPSS, Inc.) was used to process all data. Data were presented as the mean ± standard deviation (SD). The differences between groups were analyzed by one-way analysis of variance (ANOVA), followed by the independent t test. Values at p < .05 and p < .01 were considered to be statistically significant and very significant.

| Effects of FA on the cell viability and toxicity of LO2 cells and AP-or RFP-treated LO2 cells
The Results showed that the activities of ALT and AST in C18:1 group were significantly increased compared with MCFA groups (p < .01).
These results showed that less than 200 µM FA did not cause obvious cytotoxicity. Compared with the NR group, treatment with 200 µM LCFA worsen liver cell death but treatment with 200 µM MCFA did not aggravate cell injury in AP or RFP model.

| Effects of MCFA and LCFA on apoptosis in AP-or RFP-treated LO2 cells
In order to determine whether MCFA and LCFA may influence apoptosis in AP-or RFP-induced liver injury cell model, double staining with Hoechst 33342/PI assay was used to identify morphological features of apoptotic cells and double staining with Annexin V-FITC/ PI was used to determine apoptosis rate. As shown in Figure 3a,

| Effect of MCFA and LCFA on mitochondrial membrane potential in AP-or RFP-treated LO2 cells
Mitochondrial membrane potential (MMP) serves as a central regulator of cell health, which is associated with oxidative stress and apoptosis in the cell. In order to explore the effects of MCFA and LCFA on oxidative stress and apoptotic pathway in AP or RFP model, Rhodamine 123, a mitochondrion-specific fluorescent dye, was used to determine the change in MMP by flow cytometry. As shown in Figure 4a,b, the fluorescence intensities were not significantly differ-  oxidative stress, but when the cells were exposed to LCFA, oxidative stress may increase compared to the NR group in AP-or RFP-treated LO2 cells.

| Effects of MCFA and LCFA on the expression of inflammatory cytokines in AP-or RFPtreated LO2 cells
As shown in Figure 5a,b, in the AP or RFP model, the mRNA expression levels of IL-1β, IL-6, IL-8, MCP-1, and TNF-α were significantly increased in the C18:1 group compared with the NR group (p < .01).
Additionally, in the AP model, the mRNA expression level of TNF-α
It was reported that the annual incidence rates of DILI were 19.1 and 13.9 cases per 100,000 inhabitants in Iceland and France (Björnsson, Bergmann, Björnsson, Kvaran, & Sigurdur, 2013;Sgro et al., 2002). In recent years, over 1,000 drugs have been associated with DILI and the list is continuing to grow, which is also leading to the common abortion of new medicines approval and withdrawal of postmarketing drugs (Kaplowitz, Win, Than, Liu, & Dara, 2015;Senior, 2007;Stirnimann, Kessebohm, & Lauterburg, 2010;Teschke & Andrade, 2015). Both AP and RFP have been used to establish animal or cell liver injury model for screening antihepatotoxic and/ or hepatoprotective activities of drugs Darvin et al., 2018). As the concentration of AP or RFP increased and the treatment time was prolonged, the cell viability decreased in a dose-and time-dependent manner (Figure 1a,b). The IC50 values for RFP were reported to be 548.91 and 234.96 µM in 24 and 48 hr, respectively, in LO2 cells (Zhang & Xu, 2018). Other studies found that the IC50 value for RFP in 24 hr was 530.33 µM and established cell injury model by treatment with 10 mM AP for 24 hr (Wu et al., 2016; Yuan-Jing, Wei, Jian-Ping, Yu-Xia, & Zi-Ling, 2016). These reports were similar to our results in this study. Our results showed that the optimal damage condition of the DILI cell model was treatment with 10 mM AP or 600 µM RFP in LO2 cells for 24 hr, and the treated cells had low cell viability, an increased cell apoptosis rate, and elevated leakage levels of ALT, AST, and LDH ( Figure 1). LO2 cells treated with AP or RFP also exhibited increased oxidative stress and lipid peroxidation as assessed by MDA production (Tables 2 and 3).
Medium-chain fatty acids are absorbed easily as they are transported to the liver directly, which are not incorporated into chylomicrons. And they can supply energy rapidly because they could readily penetrate the mitochondrial membrane and do not need a carnitine shuttle system and carnitine palmitoyltransferase . In contrast, LCFA are slowly absorbed, because they must be incorporated into chylomicrons and rely on the lymphatic and vascular system to transport (Poppitt et al., 2010).
Therefore, LCFA are preferentially stored as body fat even liver fat and may easily induce hepatocyte injury as they have different pharmacokinetics and utilize different metabolic pathways compared with MCFA (Kim, Choe, et al., 2017;Kim, Nam, et al., 2017;Lemarié, Beauchamp, Legrand, & Rioux, 2016). In addition, a diet containing high MCFA could significantly reduce body fat accumulation, lower  insulin resistance, as well as ameliorate apoptosis, oxidative stress, and inflammatory responses as compared to LCFA diet in rats Fu, Zeng, Zeng, Wang, Wen, et al., 2016;Hu, Shen, Xiong, Zhu, & Deng, 2018;Zhou, Wang, Jiang, Zhang, et al., 2017). These factors (apoptosis, oxidative stress, and inflammatory responses) may be the causes of DILI. Thus, the effects of MCFA on the hepatotoxicity induced by drugs in human liver cells compared with LCFA (oleic acid) are worth to be studied.
It has been reported that treatment with less than 200 µM FA (octanoic acid, decanoic acid, lauric acid, and oleic acid) did not cause obvious toxicity in LO2 cells . It was found that MCFA did not cause apoptosis, oxidative stress, and inflammation in the hepatic cells with steatosis compared with LCFA .
Data from the present study ( Figure 2) have shown that the proliferation rates of LO2 cells were reduced when the concentrations of FA The overproduction of NAPQI led to GSH depletion and presumably altered protein functions by covalently binding to cellular or F I G U R E 5 Effect of FA on mRNA expression levels of inflammatory cytokines in AP-and RFP-induced LO2 cells as determined by q-PCR. The mRNA expression levels of inflammatory cytokines in AP model (a) and RFP model (b). The relative expression ratio for each gene was presented as the ratio to the NC group. The housekeeping gene β-actin was used as an internal control. Data are expressed as mean ± SD (n = 3). *p < .05, **p < .01 compared with the C18:1 group; # p < .05, ## p < .01 compared with the NR group F I G U R E 6 Protein expression of IL-1β and TNF-α in AP-and RFP-induced LO2 cells treated with FA for 24 hr. Protein expression levels of IL-1β and TNF-α in AP model (a) and RFP model (b). Relative protein expression levels of IL-1β and TNF-α in AP model (c) and RFP model (d) shown as a percentage of β-actin were presented. Data are expressed as mean ± SD (n = 3). *p < .05, **p < .01 compared with the C18:1 group; # p < .05, ## p < .01 compared with the NR group mitochondrial proteins, eventually leading to mitochondrial dysfunction, oxidative stress, and even hepatocyte necrosis (Huo et al., 2017;Lv et al., 2018;Zhang et al., 2017).
It was reported that AP-induced LO2 cells injury may be dominated by necrosis or programmed necrosis and may not involve relevant apoptosis because the fundamental difference between primary hepatocytes and hepatoma cells is the dramatically lower expression of cytochrome P450 enzymes, which are responsible for the toxicity (Jaeschke, Duan, Akakpo, Farhood, & Ramachandran, 2018).
Oxidative stress, inflammation, and cell apoptosis were the extensively accepted causes for AP-or RFP-induced liver injury. Hence, we investigated the effects of medium-and long-chain fatty acids on AP-or RFP-induced hepatocellular injury by mitochondrial dysfunction, oxidative stress, inflammatory response, and apoptotic pathways. Our data confirmed that the administration of LCFA to AP or RFP model deteriorated oxidative stress, as shown by the elevated activities of SOD and CAT and levels of T-AOC and GSH as well as a reduced level of MDA compared to the NR group (Tables 2   and 3). Additionally, the administration of LCFA to AP or RFP model significantly aggravated mitochondrial dysfunction concurrent with increased apoptosis rate as well as enhanced expression of inflammatory cytokines compared to the NR group (Figures 4-6), which supported reduced cell proliferation and the increased leakage of ALT, AST, and LDH. However, our results also showed that the apoptosis rate and the parameters of oxidative stress and inflammation remained unchanged in the MCFA groups compared with the NR group in AP or RFP model. Therefore, we proposed that MCFA may be safer than LCFA as nutrition support or the selection of daily oil and fat for the patients with DILI. Surely, to further confirm whether the effects of MCFA were mediated by oxidative stress, inflammatory response, and apoptotic pathways, we need to determine the expression of related signaling pathways such as sirtuin 1 (SIRT1), c-Jun N-terminal kinase (JNK), and nuclear factor erythroid 2-related factor 2/Kelch-like ech-associated protein 1 (Nrf2/Keap1) pathways in subsequent experiments Rada et al., 2017;Yuan-Jing et al., 2016).
In conclusion, our data have shown that treatment with 10 mM AP or 600 µM RFP for 24 hr was the optimum condition for the establishment of the DILI model in LO2 cells. Our results demonstrated that treatment with LCFA may deteriorate DILI by reducing activities of antioxidant enzymes, enhancing lipid peroxidation, inducing apoptosis, and up-regulating inflammatory cytokine expression. On the contrary, treatment with MCFA did not produce the detrimental effects and may attenuate inflammation or slightly improve oxidative stress on DILI. Therefore, this work revealed the effects of different FA on DILI and also meant it is important to consider the effect of FA on human health in further studies.

CO N FLI C T O F I NTE R E S T
The authors declare no financial or commercial conflict of interest.

E TH I C A L A PPROVA L
This study does not involve any human or animal testing.

DATA AVA I L A B I L I T Y S TAT E M E N T
The data that support the findings of this study are available from the corresponding author upon reasonable request.