Regulation of Sirt1/Nrf2/TNF-α signaling pathway by luteolin is critical to attenuate acute mercuric chloride exposure induced hepatotoxicity

Inorganic mercury, though a key component of pediatric vaccines, is an environmental toxicant threatening human health via accumulating oxidative stress in part. Luteolin has been of great interest because of its antiinflammatory, anticarcinogenic and antioxidative effects. Here we hypothesized that luteolin would attenuate hepatotoxicity induced by acute inorganic mercury exposure. Kunming mice were treated with luteolin (100 mg/kg) 24 h after administration of 4 mg/kg mercuric chloride (HgCl2). The results showed that luteolin ameliorated HgCl2 induced anemia and hepatotoxicity, regulating radical oxygen species (ROS) production and hepatocyte viability in vitro and oxidative stress and apoptosis in vivo. Furthermore, luteolin reversed the changes in levels of inflammation- and apoptosis-related proteins involving NF-κB, TNF-α, Sirt1, mTOR, Bax, p53, and Bcl-2, and inhibited p38 MAPK activation. Luteolin enhanced antioxidant defense system based on Keap1, Nrf2, HO-1, NQO1, and KLF9. Moreover, luteolin did not affect miRNA-146a expression. Collectively, our findings, for the first time, elucidate a precise mechanism for attenuation of HgCl2-induced liver dysfunction by dietary luteolin via regulating Sirt1/Nrf2/TNF-α signaling pathway, and provide a foundation for further study of luteolin as a novel therapeutic agent against inorganic mercury poisoning.

Inorganic mercury is a well-known environmental toxicant and normally occurs in rocks, soil, water, atmosphere, and organisms in trace amounts. Researches have shown that, in mammals, mercury can induce a wide range of adverse effects on systems and tissues [1][2][3][4][5] . Aplastic anemia is another potential consequence of inorganic mercury exposure 6 . As a critical organ for drug metabolism, the liver is primary target of toxic chemicals. In chronic poisoning experiments, inorganic mercury induces severe liver injury as shown by hepatic morphological changes and apoptosis, as well as negative effect on hepatic function 5 .
The toxicity of inorganic mercury primarily involves undermining antioxidant defense systems through reactions with cellular thiols 7 . Moreover, the toxicity and therapeutic effects on some diseases of inorganic mercury have been, in part, attributed to increased oxidative stress. Mercury is, in addition, a potent apoptosis inducer, through cytochrome c release 8 , and can upregulate nuclear factor kappa-light-chain-enhancer of activated B cells (NF-κ B) level 9 and activate p38 mitogen-activated protein kinases (MAPK) 10 .
Thimerosal is regarded as an irreplaceable ingredient in some pediatric vaccines but, though it is not conclusive, is also thought to contribute to adverse neurobehavioral effects 11 . Inorganic mercury is also used in cosmetics for skin whitening 12 . In addition, daily mercury intake can occur by eating certain foods, especially fish contaminated with inorganic mercury 13 and inhaling air which contains vapor phase and particulate mercury 14 .
Overall, human health is being seriously threatened by mercury exposure.
In the clinic, using sodium 2,3-dimercapto-1-propanesulfonate as a chelating agent is an effective current treatment for removing mercuric ion (Hg 2+ ) from organs 15 . Combination therapies with chelating agents, plasma exchange, hemodialysis, and plasmapheresis are used for effectively treating severe inorganic mercury poisoning 16 . Nonetheless, these therapies require long-term treatment and meticulous supportive care. Unfortunately, these treatments currently applied in the clinic are ineffective at repairing tissue damage. Considering that inorganic mercury depresses antioxidant defense system, antioxidant compounds have been proposed as potential treatments which are nontoxic and have low side effects.
Luteolin (3′ ,4′ ,5,7-tetrahydroxyflavone), a natural flavone derived from many traditional Chinese medicinal plants, has numerous health benefits. This molecule has received extensive attention because of its antiinflammatory 17 , antioxidative 18 , and anticarcinogenic 19 activities. Increasing evidence has indicated that luteolin might modulate the homeostasis between oxidants and antioxidants, and reduce reactive oxygen species (ROS) production and apoptosis 18 . Liver, intestine, and kidney are vital target organs for luteolin. Researches have demonstrated protective effects of luteolin against liver injury induced by acetaminophen 20 or tetrachloromethane 21 , through mechanisms involving restoring antioxidant enzyme activities and attenuating proinflammatory factors expression. Systemic administration of luteolin suppressed tumor cell growth in cancers 22 . It was proposed that luteolin ameliorated diet-induced obesity and related complications through interactions between liver and adipose tissue 23 .
The molecular mechanisms of the anticancer effects of luteolin have been well described, primarily involving cell cycle block and apoptosis inhibition 24 . However, existing literature has not supported antioxidant effect of luteolin. Nuclear factor (erythroid-derived 2)-like 2 (Nrf2) was described as being important for the antioxidant response 25 , antiinflammatory response 26 , and cytoprotection of hepatocytes 27 . A recent study showed that miRNA-146a regulated Nrf2 translation through binding to Nrf2 mRNA in aging 28 . Sirtuin type 1 (Sirt1), a NAD + -dependent histone deacetylase strongly expressed in the liver, is intimately related to cell proliferation, differentiation, apoptosis, and metabolism 29 . Mammalian target of rapamycin (mTOR) is also involved in drug metabolism in the liver 30 . Modulation of these factors may explain mechanisms of liver injury and hepatoprotection.
There has been increasing attention on using natural products to prevent and cure diseases induced by environmental toxicants 19,[31][32][33] . Luteolin was reported to inhibit vascular endothelial growth factor release from human mast cells exposed to mercuric chloride (HgCl 2 ) 34 . However, whether luteolin could affect hepatotoxicity induced by acute inorganic mercury exposure has not yet been elucidated. Thus, we hypothesized that luteolin would attenuate hepatotoxicity induced by acute HgCl 2 poisoning. To address this problem, we investigated effects of dietary luteolin on HgCl 2 induced changes in proinflammatory factors, the antioxidant defense system, and apoptotic signaling pathway as well as potential mechanisms for luteolin-mediated protection against HgCl 2 -induced hepatotoxicity.

Results
Protection by luteolin against HgCl 2 -induced changes in the blood. Amounts for white blood cells (WBC) and neutrophils were decreased significantly in the HgCl 2 -treated group (P < 0.05), whereas these were restored by luteolin (Table 1). Table 1 also showed the decreases in red blood cell (RBC) amount and hemoglobin concentration (HGB) with HgCl 2 administration, and luteolin reversed these effects. Noticeable decreases in erythrocyte mean corpuscular volume (MCV), mean corpuscular hemoglobin (MCH), and blood platelets (PLT) amount were also observed (P < 0.05), while mean corpuscular hemoglobin concentration (MCHC) and red blood cell distribution width (RDW) were increased in the HgCl 2 -treated group. These findings suggested normal pigment positive cell anemia in mice. Combined with the clinical manifestations and decreased neutrophils, aplastic anemia was identified in the HgCl 2 group. Based on the comprehensive analysis of MCV, MCH, MCHC, RDW and PLT, luteolin alleviated HgCl 2 -induced aplastic anemia.

Luteolin attenuated HgCl 2 -induced oxidative stress in liver tissue and liver dysfunction.
Alanine aminotransferase (ALT) and aspartate aminotransferase (AST) activities were used to assess liver dysfunction. ALT and AST activities in serum from different groups were shown in Fig. 1a,b. In the HgCl 2 -treated group, ALT (Fig. 1a) and AST (Fig. 1b) activities were significantly increased compared with in the control group (P < 0.05). However, post treatment with luteolin significantly (P < 0.05) reversed the effects of HgCl 2 on serum ALT and AST activities (Fig. 1a,b). Malondialdehyde (MDA) is a biomarker for oxidative stress and reduced glutathione (GSH) is an antioxidant, preventing damage caused by free radicals, lipid peroxide, and heavy metals. Treatment with HgCl 2 increased MDA concentrations in liver tissue, but luteolin reversed this effect (P < 0.05) (Fig. 1c). GSH concentrations were clearly decreased in hepatic homogenates of the HgCl 2 -treated group, compared with the control group (P < 0.05), while post treatment with luteolin caused a significant increase in GSH (P < 0.05) compared with the levels observed in the group receiving only HgCl 2 (Fig. 1d).
Luteolin ameliorated HgCl 2 -induced liver injury and apoptosis. Histopathological assessments of liver sections from the mice were shown in Fig. 2a. Congestion of the central vein was observed, with severe erythrocyte infiltration, in the HgCl 2 -treated group, along with plasmolysis of the hepatocytes and broadening of the hepatocellular gap around the central vein. In the HgCl 2 + luteolin group, there was swelling of hepatocytes and slight erythrocyte infiltration. However, there were no obvious histopathological changes in livers from the other groups.
As shown by the TUNEL assay ( Fig. 2c), in the HgCl 2 -treated group, the level of apoptotic hepatocytes ( Fig. 2b) was significantly higher than in the control group. However, there was no significant difference between luteolin and control group. The stimulatory effect by HgCl 2 was attenuated in the HgCl 2 + luteolin-treated group, indicating that luteolin significantly prevented HgCl 2 -induced apoptosis (P < 0.05).

Luteolin reversed the changes in protein levels regulated by HgCl 2 . The complex of Nrf2 and
Kelch-like ECH-associated protein 1 (Keap1) plays major role in modulating cellular oxidative stress. We found that Nrf2 and Keap1 were involved in protection against HgCl 2 -stimulated oxidative stress by luteolin. HgCl 2 induced a significant (P < 0.05) decrease in translation of Nrf2 and Keap1, and luteolin significantly (P < 0.05) reversed these effects. Levels of heme oxygenase 1 (HO-1) and NAD(P)H: quinone oxidoreductase 1 (NQO1), target proteins of Nrf2, were significantly (P < 0.05) lower in the HgCl 2 -treated than in the control group. However, luteolin post treatment significantly attenuated this decrease (P < 0.05). In addition, it was reported that oxidative stress increased Kruppel-like factor 9 (KLF9) level 35 . In our study, level of KLF9 was significantly (P < 0.05)

Figure 1. Effects of luteolin on the liver function indicators activities and oxidative stress indicators levels regulated by HgCl 2 . (a) ALT and (b)
AST activities in serum of all samples from luteolin group, control group, HgCl 2 group, and HgCl 2 + luteolin group were detected with a Uni Cel DxC Synchron chemistry system. Values are mean ± SEM (n = 7). (c) MDA and (d) GSH concentrations in mice liver of all samples from luteolin group, control group, HgCl 2 group, and HgCl 2 + luteolin group were determined by commercial assay kits. Values are mean ± SEM (n = 7). *Significantly different from the corresponding control group, P < 0.05; # Significantly different from the corresponding HgCl 2 group, P < 0.05. decreased in the HgCl 2 -treated group, compared with control group, and luteolin significantly (P < 0.05) reversed this effect (Fig. 3a,b).
Levels of NF-κ B and tumor necrosis factor alpha (TNF-α) were significantly higher in the HgCl 2 -treated than in the control group (P < 0.05). In Fig. 3c-e, luteolin alone decreased NF-κ B and TNF-α levels in the liver, and it also significantly (P < 0.05) prevented NF-κ B expression and TNF-α production induced by HgCl 2 . Treatment with HgCl 2 significantly enhanced phosphorylation of p38 MAPK (P < 0.05), but luteolin produced a significant (P < 0.05) inactivation of p38 MAPK against active action of HgCl 2 .
Sirt1 and mTOR are proteins related to metabolism. Sirt1 and mTOR levels were significantly (P < 0.05) suppressed in the HgCl 2 -treated group, but luteolin reversed this effect (Fig. 3h,i). HgCl 2 and luteolin had no effect on miRNA-146a expression. MiRNA-146a is a potential regulator of Nrf2 28 . As shown in Fig. 4, no significant differences in miRNA-146a levels were observed among the 4 groups. Neither HgCl 2 nor luteolin affected miRNA-146a transcription in this experiment.

Discussion
Inorganic mercury is an important environmental pollutant causing systemic toxicity and threatens human health. Results of WBC count, a biomarker of chemical intake, in our study, luteolin maintains WBC amount. This suggests that luteolin attenuates HgCl 2 -induced injury possibly through attenuating total mercury accumulation in mice. Complete blood analysis indicates that there is aplastic anemia in mice treated with HgCl 2 , which is consistent with other reports 6,36 . Therefore, we hypothesize that HgCl 2 might induce hemopoietic stem cell injury, while luteolin may serve a protective role in its progression. Strong evidence has suggested that inorganic mercury potently inhibits uroporphyrinogen decarboxylase 5 , an important enzyme catalyzing conversion of uroporphyrinogen to coproporphyrinogen. This can prevent heme synthesis and ultimately arrest production of hemoglobin. Luteolin most likely protects against inorganic mercury and restores hemoglobin levels by attenuating the obstruction of heme biosynthesis induced by HgCl 2 . All the above results suggest that luteolin may be useful for reducing the toxic effects of Hg 2+ on uroporphyrinogen decarboxylase, ameliorating anemia, inhibiting Hg 2+ accumulation, and attenuating injury to the organism. The serum ALT and AST activities indicate that luteolin protects the mice from HgCl 2 -induced liver injury. Its protection against liver injury is consistent with the liver histological observations. Moreover, luteolin mitigates HgCl 2 -induced apoptosis and maintains hepatocyte viability. Together, all these results indicate that luteolin inhibits HgCl 2 -induced hepatic inflammation, apoptosis, and cytotoxicity.
The mechanism of liver injury induced by HgCl 2 is believed to involve ROS production and free radical mediated damage. Measurements of MDA and GSH of HgCl 2 -treated mice indicate ROS production and free radical damage, in good agreement with the cellular ROS levels in vitro. Hg 2+ complexed tightly with hydrosulphonyl moieties after entering the body, causing depletion of intracellular hydrosulphonyl moieties and release of reactive oxygen free radicals. It resulted, either indirectly or directly, in oxidative stress and lipid peroxidation 37 . However, luteolin attenuates oxidative stress and free radical damage, and enhances the antioxidant system including superoxide dismutase (SOD) and GSH, indicating that luteolin provides protection against HgCl 2 -induced lipid peroxidation and oxidative stress.
Findings above confirm that luteolin arrests ROS production and decreases oxidative stress to prevent HgCl 2 -induced hepatotoxicity. Excessive oxidative stress consumes a large amount of Nrf2 and Keap1, disrupting the homeostasis between expression and degradation of these two factors. Our protein expression data indicate that luteolin promotes Nrf2 expression, and reverses the depletion of Nrf2 caused by acute HgCl 2 exposure, thus improves the ability to resist oxidative stress.
Levels of downstream proteins of Nrf2 such as NQO1, HO-1 and SOD were upregulated in HgCl 2 + luteolin group in good agreement with effect of luteolin on Nrf2 levels. This indicates that luteolin activates Nrf2 signaling pathway to benefit detoxification and antioxidant defense system. KLF9 modulated cell death and oxidative injury under conditions of excessive oxidative stress, and was positively regulated by Nrf2 35 . Therefore, it can be concluded that luteolin attenuates HgCl 2 -induced oxidative stress via alleviating depletion of Nrf2 and activating Nrf2 signaling pathway to upregulate KLF9 and enhance antioxidant defense system. TNF-α is a cytokine involved in systemic inflammation and a component of the acute phase reaction 38 . In this study, luteolin suppresses TNF-α production in the liver in the presence of HgCl 2 . Luteolin also was reported to inhibit TNF-α release by inhibiting extracellular regulated protein kinases, p38 MAPK, and casein kinase 2 activation from macrophages 39 . TNF-α , when binding to tumor necrosis factor receptors (TNFR), binds to the TNFR type 1-associated death domain protein (TRADD) and then activates p38 MAPK and NF-κ B 40 . P38 MAPK represents a class of MAPKs that can also activate NF-κ B 41 . Our results suggest that luteolin reduces NF-κ B and phosphorylation of p38 which occurs in the presence of HgCl 2 . This implies that luteolin inhibits TNF-α to inactivate p38 MAPK and inhibit NF-κ B, to reverse the HgCl 2 -induced inflammatory response. Nrf2 has a negative effect on TNF-α expression 42 , which suggests that upregulation of Nrf2 by luteolin may target inactivation of HgCl 2 -induced inflammatory signaling pathways. Luteolin attenuates HgCl 2 -induced excessive oxidative stress to ameliorate inflammation thereby preventing liver injury. Therefore, we conclude that Nrf2 is a key regulatory factor in antioxidant and antiinflammatory defense systems, and plays a critical role in the protection against HgCl 2 exposure by luteolin.
Apoptosis signaling pathways involves p53 and the Bcl-2 protein family 43 , including proapoptotic and prosurvival proteins 44 . The tumor suppressor protein p53 influences apoptosis and can modulate levels of the Bcl-2 protein family 43 . In our study, luteolin suppresses p53, thereby increases Bcl-2 level and decreases Bax level, and finally protects hepatocytes against HgCl 2 -induced apoptosis.
NF-κ B activation plays a dual role in regulating apoptosis in various tissues and cells 45,46 . The relative protein levels of Bcl-2, Bax, and NF-κ B show that luteolin suppresses NF-κ B, thereby inhibits apoptosis 45 . NF-κ B and p53 could be upregulated by p38 MAPK 41,47 . Activation of the p38 pathway significantly stimulated p53 function 47 . Moreover, p38 MAPK also affected NF-κ B levels by promoting phosphorylation of Iκ B, resulting in the dissociation and degradation of NF-κ B and Iκ B complexes 41 . The levels of p53, NF-κ B, and p38 demonstrates that luteolin inhibited p38-activated NF-κ B and p53 pathways, which then contributes to the protection of luteolin against HgCl 2 -induced inflammation and apoptosis.
Sirt1, a NAD + -dependent protein deacetylase, regulates such cellular processes as stress response and longevity 48 . mTOR is a serine/threonine protein kinase that regulates cell survival, protein synthesis, and translation 49 . Our data, for the first time, show that luteolin activates Sirt1 and mTOR, which are inhibited by HgCl 2 . Sirt1 directly suppresses NF-κ B and p53 activation, because its N-terminal domain promotes deacetylation of NF-κ B p65 48 and p53 50 . This suggests that, in our experiments, luteolin protects hepatocytes and inhibits the inflammation and apoptosis via promoting Sirt1 expression to suppress NF-κ B and p53 induced by HgCl 2 exposure. Moreover, there is reliable evidence that mTOR can regulate Bcl-2 activation, by a positive feedback mechanism, to inhibit apoptosis 51 . In addition, it was reported that Sirt1 activated the Nrf2 pathway to decrease ROS production induced by advanced glycation end products in glomerular mesangial cells 52 . Together, all these demonstrate that Sirt1 is a key factor in regulating inflammation, apoptosis and antioxidant defense systems, contributing to prevent the hepatotoxity of HgCl 2 by luteolin.
Interestingly, miRNA-146a was reported to inhibit Nrf2 protein synthesis, but maintaining Nrf2 mRNA levels, in aging rats 28 . Luteolin was also reported to inhibit procarcinogenic miRNAs 53 . Regretfully, in our study, neither HgCl 2 nor luteolin has any effect on miRNA-146a transcription, arguing against a role for miRNA-146a in the effects we observed with both HgCl 2 and luteolin. Therefore, we infer that luteolin may maintain Nrf2 production in mice liver by activating the Sirt1 signaling pathway or by directly affecting the Nrf2-Keap1 complex.
In conclusion, luteolin protects hepatocytes from oxidative stress, inflammation, and apoptosis induced by HgCl 2 in the liver via modulating the Sirt1/Nrf2/TNF-α signaling pathway (summarized in Fig. 6). Moreover, luteolin also have attenuated HgCl 2 -induced blood toxicity by modulating hemoglobin synthesis and reducing mercury accumulation, though the detailed mechanism still requires further study. Therefore, we insist luteolin, in combination with inorganic mercury, may improve the safety of pediatric vaccines with mercury. In addition, dietary intake of luteolin may offer a novel and safe method to protect human health against inorganic mercury exposure.

Materials and Methods
Animals and treatments. All animal protocols were approved by the Ethical Committee for Animal Experiments (Northeast Agricultural University, Harbin, China). Twenty-eight adult healthy male Kunming mice (25 ± 5 g body weight) were obtained from Harbin Veterinary Research Institute (Harbin, China). All animals were acclimated for 1 w under the same laboratory conditions with a 12 h interval light/dark cycle, a minimum of 40% relative humidity, a room temperature of 21 ± 4 °C, standard food, and water ad libitum. Housing The 28 mice were randomly and equally divided into 4 groups of 7 animals each. The groups were: control, luteolin, HgCl 2 , and HgCl 2 + luteolin. Total HgCl 2 (4 mg/kg) (Beijing Chemical Plant, Beijing, China) was administered by intraperitoneal injection, as a suspension in 0.9% (w/v) physiological saline. Luteolin (100 mg/kg) (Xi'an Weiao Biological Technology Company Ltd., Xi'an, China) was administered intragastrically as a suspension in 1% (v/v) dimethyl sulfoxide (DMSO). In the control group, equal amount of 0.9% physiological saline and 1% DMSO were given as vehicles orderly. In the luteolin group, mice received a single dose of luteolin (100 mg/ kg) only. In the HgCl 2 -treated group, mice received HgCl 2 (4 mg/kg) only, and in the HgCl 2 + luteolin-treatment group mice received luteolin (100 mg/kg) 24 h after HgCl 2 administration.
All mice were killed by given ether anesthesia 24 h after the last treatment. Blood samples were collected from the abdominal vein into vacuum tubes containing heparin sodium anticoagulant. Liver tissues were rapidly excised and homogenized in phosphate-buffered saline (PBS) pH 7.4 using an Ultra-Turrax T25 Homogenizer. After centrifugation at 10,000 × g for 10 min at 4 °C, the supernatant was used for biochemical determinations.
Complete blood count and biochemical analysis. Some of the blood samples were used for complete blood count, which were obtained with an automated Auto Hematology Analyzer BC-2600Vet (Mindray, Shenzhen, China). Other blood samples were centrifuged at 3,000 × g for 10 min. Activities of ALT and AST were detected in the serum with a Uni Cel DxC Synchron chemistry system (Beckman Coulter Inc., Fulton, CA, USA). Western blot analysis. The Bicinchoninic Acid Kit (Beyotime Institute of Biotechnology) was used to determine protein content of liver samples to ensure gel loading for western blots. Equal aliquot (8 μ g) of the protein samples were separated by SDS-PAGE gel electrophoresis using a BioRadÓ Mini-PROTEANÒ 3 electrophoresis cell (BioRad, Hercules, CA, USA) and electrophoretically transferred to polyvinylidene fluoride membrane (Immobilon ® -P Transfer Membrane, EMD Millipore, Billerica, MA, USA), and then the membranes were probed with appropriate combination of primary and horseradish peroxidase-conjugated secondary antibodies from Santa Cruz Biotechnology (Dallas, TX, USA). Proteins in the membranes were visualized by enhanced chemiluminescence kits. The protein bands were quantified by the average ratios of integral optic density following normalization to the levels of internal control GAPDH, and the results were further normalized to control.

Measurement of oxidative stress indicators in liver tissues
MiRNA-146a isolation and absolute quantitative real-time PCR analysis.  was extracted using SanPrep Column miRNA Mini-Preps Kit (Sangon Biotech, Shanghai, China) according to the manufacturer's instructions. MiRNA-146a detection by real-time analysis involved reverse transcription of cDNA using a small RNA specific stem-loop RT primer (mmu-miR-146a-5p; 5′ -CTCAACTGGTGTCGTGGAGTCGGCAATTCAGTTGAGAACCCATGG-3′ ). Once specific cDNA had been generated, individual miRNA was detected using SYBR Green RNA assay real time PCR analysis (mmu-miR-146a-5p; 5′ -ACACTCCAGCTGGGTGAGAACTGAATTC-3′ ). Real-time PCR was conducted using Roche LightCycler480 (Roche, Basel, Switzerland). The thermal cycling included 3 min of denaturation at 95 °C followed by 45 PCR cycles, including 15 s at 95 °C, 20 s at 57 °C, and 30 s at 72 °C. Linearized plasmid was quantified using a spectrophotometer and copy numbers were calculated.
Establishment of the absolute quantitative standard curve. In order to examine the miRNA-146a copy number, generation of the absolute quantitative standard curve was necessary. Six different concentrations of standard samples were prepared respectively, by mixing cDNA obtained by reverse transcription with plasmid XM709-2. The parameters of the standard curve was: log N = − 3.176Δ Ct + 36.91 (R 2 = 0.9981, P < 0.01). The standard curve was shown in Supplementary Figure 1. Hepatocyte culture and treatment. Adult male Kunming mouse was injected intraperitoneally with pentobarbital and heparin. Mouse hepatocytes were prepared as described previously 54 . Briefly, the liver was perfused in situ with collagenase (Sigma, St. Louis, MO, USA) through the hepatic portal vein. The total cells released were centrifuged 3 times, for 3 min at 100 × g, 50 × g, 50 × g. Hepatocytes were suspended at a density of 5 × 10 5 cells/mL in adherent culture medium. Dulbecco modified eagle medium (DMEM, Invitrogen, Grand Island, NY, USA) was supplemented with 2 g/L HEPES (Gibco, NY, USA), 6 mg/L insulin (Sigma), 1 mg/L dexamethasone (Sigma), 1% (v/v) penicillin/streptomycin (Thermo Fisher Scientific), and 10% (v/v) fetal bovine serum (Hyclone, Logan, UT, USA). Next, 2.5 mL cell suspension was plated into 6-well plates containing collagen-coated glass cover slips. After culturing at 37 °C under 5% CO 2 for 24 h, medium and nonadherent hepatocytes were aspirated and replaced with culture medium containing 5% (v/v) fetal bovine serum.
Determination of hepatocyte viability. Hepatocyte viability was determined by using WST tetrazolium salt (CCK-8, Dojindo, Kumamoto, Japan) following the manufacturer's instructions. Briefly, 10 4 cells/well were seeded in 96-well plates with DMEM media. After culture overnight, hepatocytes were treated with 5 μ M HgCl 2 for 24 h, with or without pretreatment with 20 μ M luteolin for 2 h. The medium was discarded and hepatocytes were incubated in 100 μ L medium with 10 μ L CCK-8 solution at 37 °C for 4 h. The optical density was measured at 450 nm on a Bio-Tek Epoch microplate reader (Bio-Tek, Winooski, VT, USA).

Measurement of ROS generation. Generation of intracellular ROS was determined by the Reactive
Oxygen Species Assay Kit (Beyotime Institute of Biotechnology) according to the manufacturer's instructions. Briefly, 10 4 cells/well were seeded onto square glass coverslips (24 × 24 mm) in 6-well plates. After overnight culture, the hepatocytes were treated with 5 μ M HgCl 2 for 24 h with or without pretreatment with 20 μ M luteolin for 2 h. After treatments, cells were incubated with DFCH-DA at a final concentration of 10 μ M at 37 °C for 20 min. The hepatocytes were observed by fluorescence microscopy (Olympus IX51, Nikon, Tokyo, Japan) with an excitation wavelength of 488 nm and emission wavelength of 525 nm.
Statistical analysis. Data are presented as mean ± standard error of the mean (SEM). Statistical analyses were performed with SPSS 19.0 software (SPSS, Chicago, IL, USA). Shapiro-Wilk was performed to assess the normality of the data, and Levene's Test for equality of variances was performed. One-way analysis of variance was used to determine differences among 4 groups. Tukey Test for post hoc multiple comparison was used to determine differences between means. A two-tailed P < 0.05 was considered as being significant.