Wogonin alleviates liver injury in sepsis through Nrf2‐mediated NF‐κB signalling suppression

Abstract Sepsis is a life‐threatening organ dysfunction syndrome, and liver is a susceptible target organ in sepsis, because the activation of inflammatory pathways contributes to septic liver injury. Oxidative stress has been documented to participate in septic liver injury, because it not only directly induces oxidative genotoxicity, but also exacerbates inflammatory pathways to potentiate damage of liver. Therefore, to ameliorate oxidative stress is promising for protecting liver in sepsis. Wogonin is the compound extracted from the medicinal plant Scutellaria baicalensis Geogi and was found to exert therapeutic effects in multiple inflammatory diseases via alleviation of oxidative stress. However, whether wogonin is able to mitigate septic liver injury remains unknown. Herein, we firstly proved that wogonin treatment could improve survival of mice with lipopolysaccharide (LPS)‐ or caecal ligation and puncture (CLP)‐induced sepsis, together with restoration of reduced body temperature and respiratory rate, and suppression of several pro‐inflammatory cytokines in circulation. Then, we found that wogonin effectively alleviated liver injury via potentiation of the anti‐oxidative capacity. To be specific, wogonin activated Nrf2 thereby promoting expressions of anti‐oxidative enzymes including NQO‐1, GST, HO‐1, SOD1 and SOD2 in hepatocytes. Moreover, wogonin‐induced Nrf2 activation could suppress NF‐κB‐regulated up‐regulation of pro‐inflammatory cytokines. Ultimately, we provided in vivo evidence that wogonin activated Nrf2 signalling, potentiated anti‐oxidative enzymes and inhibited NF‐κB‐regulated pro‐inflammatory signalling. Taken together, this study demonstrates that wogonin can be the potential therapeutic agent for alleviating liver injury in sepsis by simultaneously ameliorating oxidative stress and inflammatory response through the activation of Nrf2.


| INTRODUC TI ON
Sepsis is a life-threatening syndrome characterized by the dysfunction of multiple organs due to dysregulated host immune response to infection. The incidence of sepsis is about 0.3%-1.03% in highincome countries, and its mortality is between 30% and 50% as the leading cause of death in the intensive care units (ICUs). [1][2][3][4] Of note, septic liver injury substantially contributes to the poor prognosis of patients and is regarded as an independent indicator of the mortality.
Recent studies have demonstrated that damaged hepatocytes would release damage-associated molecular patterns (DAMPs) to trigger more pronounced regional and systemic inflammatory responses, exacerbating the dysfunction of liver and other organs. 5,6 In addition, septic liver injury is closely associated with cholestasis, hypoxic hepatitis and severe coagulopathy, which would progress to liver failure and worsen the outcome of patients with sepsis. Therefore, to effectively ameliorate liver injury is crucial for the treatment of sepsis and the improvement of patients' prognosis.
During the progression of sepsis, hypotension-induced ischaemia, severe infection and dysregulated immune response are all greatly implicated in inducing septic liver injury. The generation of reactive oxygen species (ROS) is tightly linked to these crucial pathogenic insults and mediates their effects. On one hand, oxidative stress with excessive accumulation of ROS can directly cause damage to hepatocytes by promoting the formation of oxidative protein adducts and lipid peroxides. 7 On the other hand, continuous oxidative stress is a potent trigger of inflammatory signalling to amplify the release of cytokines like tumour necrosis factorα (TNFα) and interleukin-6 (IL-6) via the activation of nuclear factor-κB (NF-κB) pathway in hepatocytes and Kupffer cells, thus forming a positive feedback loop to recruit more neutrophils and lymphocytes and augment the damage of liver. Recently, some studies have revealed that the agents with anti-oxidative capacity could effectively suppress the systemic immune response and improve the outcome of patients with sepsis. 8,9 However, whether septic liver injury could be relieved by the scavenging of ROS with some specific anti-oxidative agents and thus bring benefits to sepsis treatment remains far from understood.
Wogonin (5,7-dihydroxy-8-methoxyflavone) is the natural component isolated from the root of Scutellaria baicalensis Georgi, the crude drug commonly used for allergies and cancer as complementary treatment. 10,11 Intriguingly, its anti-oxidative effect has been extensively proved in multiple diseases, in particular those with the involvement of inflammatory responses. To be specific, wogonin was capable of attenuating oxidative stress via the activation of peroxisome proliferator-activated receptorγ (PPARγ)/adiponectin receptor 2 (AdipoR2) pathway and exerted the therapeutic effect on non-alcoholic steatohepatitis (NASH) in mice. 12 Moreover, in drugor heavy metal-induced nephritis, wogonin exerted its anti-oxidative capacity through PPARγ and NF-κB signalling so as to alleviate nephrotoxicity. 13,14 What's more, wogonin played a neuroprotective role by restraining oxidative stress-induced chronic inflammation in primary cultured cortical neurons. 15 Nevertheless, it still remains to be elicited whether wogonin can suppress septic liver injury via the prevention of oxidative stress.
Nuclear erythroid 2-related factor (Nrf2) is a crucial regulator of redox balance with its transcriptional activity in various tissues and cells. After the translocation to nucleus, Nrf2 binds to the antioxidant response elements (AREs) in DNA of downstream genes and initiates expression of antioxidant enzymes such as haeme oxygenase-1 (HO-1) and superoxide dismutases (SODs). 16 In vivo studies demonstrate that knockout of Kelch-like ECH-associated protein 1 (Keap1) can greatly improve the treatment outcome of septic mice via the potentiation of the transcriptional activity and anti-oxidative capacity of Nrf2. 17 Besides, in hyperlipidaemia-and hyperglycaemia-induced liver injury model, hydrogen sulphide alleviates liver injury via Ssulfhydrated-Keap1/Nrf2 signalling, whereas Nrf2 deficiency can abolish the protective effects, indicating the indispensable role of Nrf2 signalling for maintaining liver homeostasis and preventing liver injury. Recently, the activation of Nrf2 signalling has been proved to be responsible for the protective role of wogonin in different tissues and cells; thus, we suggest that wogonin exerts protective effects towards liver injury via activation of Nrf2 signalling in sepsis. Blood and liver tissue were harvested 8 hours after LPS injection.

| Sepsis mouse model
To establish caecal ligation and puncture (CLP)-induced sepsis mouse model, C57BL/6J mice were completely anesthetized with 1% pentobarbital solution and a midline abdominal incision was performed. The cecum was ligated at the 1/2 of the distal end and was perforated by sterile needles (7 # ) to induce polymicrobial peritonitis and sepsis. The abdominal wall was sutured in two layers followed by injection of 0.1 mL saline subcutaneously for fluid resuscitation.
The animals in the sham group underwent laparotomy and bowel manipulation without ligation and perforation. Wogonin treatment was performed 2 hours before CLP operation. Sham and CLP groups were treated with vehicle at the same time. Body temperature and respiratory rate were monitored 2 hours before and 12 hours after operations. Blood, liver and kidney tissue were harvested 12 hours after CLP operation.

| Detection of cytokines and hepatic transaminases by ELISA
Detection of TNFα, IL-1β, interferonγ (IFNγ) and IL-6 in serum was performed using the commercially available ELISA kits (Dakewe Bio-engineering Co, Ltd, Cat. No: 1217202 for TNFα, 1210122 for IL-1β, 1210002 for IFNγ, 1210602 for IL-6). Detection of aspartate transaminase (AST) and alanine transaminase (ALT) in serum was performed using the pre-coated ELISA kits (mlbio, Cat. No: ml002169-1 for ALT, ml058577-1 for AST). Blood samples were collected 8 hours after LPS i.p. whereas 12 hours after CLP operation, centrifuged at 1200 g for 20 minutes at 4°C, and the supernatant was collected to detect TNFα, IL-1β, IL-6, IFNγ, AST and ALT levels in serum. All the procedures were done according to the manufacturers' instructions.
The plates were measured with an ELx 808 plate reader (Bio-Tek Instruments Inc) at 450 nm.

| Haematoxylin and Eosin (H&E) staining
Liver and kidney tissue of the sepsis mice were harvested, immersed in 4% paraformaldehyde fixative and then embedded in paraffin. were observed from one slide for scoring. The liver histopathological scores were calculated as previously reported. 18 The renal tissue damage scores were calculated as previously reported. 19 Briefly, the percentage of damaged renal tubules was used to assess the score of tissue damage: 0, no damage; 1, <25% damage; 2, 25%-50% damage; 3, 51%-75% damage; 4, >75% damage, and the criteria of tubular damage included loss of brush border, tubular dilation, cast formation, inflammatory cell infiltration and cell lysis. to inactivate endogenous peroxidases, blocked with goat serum at room temperature for 20 minutes, respectively. The slides were incubated with the corresponding primary antibodies at 4°C overnight, which were washed with PBS, followed by the incubation with biotinylated goat anti-rabbit IgG and subsequent HRP-conjugated streptomycin. Diaminobenzidine (ZSGB-Bio) was added to the slides for chromogenic reaction. The slides were mounted and observed with optical microscope (Olympus).

| TUNEL staining
Liver tissue of the sepsis mice for TUNEL staining was harvested, im-  Table S1.

| RNA interference
Small interfering RNA (siRNA) specifically targeting Nrf2 (siNrf2) and negative control siRNA (siNC) were designed and synthesized by GenePharm. Sequences of the siRNA were listed in Table S2.
The siRNA was transfected into cells using Lipofectamine™ 3000 (Invitrogen) and Opti-MEM (Gibco). The transfection procedure was performed according to the manufacturer's instructions.

| Statistical analysis
The results were analysed by two-tailed Student's t test or oneway analysis of variance (ANOVA). The survival curves were analysed by log-rank (Mantel-Cox) test. The results were presented as mean ± SD through at least 3 independent experiments, with P < .05 considered to be statistically significant. Statistical analyses were accomplished using GraphPad Prism 6.0 (GraphPad Software).

| Wogonin treatment improves the prognosis of mice with LPS-or CLP-induced sepsis
In order to study whether wogonin exerts its protective role in the treatment of sepsis, we first established the mouse sepsis model via the intraperitoneal injection of LPS or the surgery of caecal ligation and puncture (CLP) as previously described 23 In addition, the concentrations of pro-inflammatory cytokines TNFα, IL-1β, IFNγ and IL-6 in circulation were prominently repressed after wogonin treatment in these two types of sepsis mice ( Figure 1G), with the levels of these cytokines gradually decreased as the dose of wogonin increased ( Figure S1C,D). Of note, the lethality is reported to occur in an acute manner in LPS-induced sepsis mouse model, whereas begins relatively later after the surgery of CLP, 24,28 as was revealed in previous reports and our Kaplan-Meier survival analysis ( Figure 1H). We further illustrated the survival rate of LPS-or CLP-induced sepsis mice 48 hours or 7 days after wogonin treatment, respectively, which showed that wogonin treatment prominently improved the survival of these two types of sepsis mice ( Figure 1H), in particular in a dose-dependent manner ( Figure S1E,F). Taken together, our results demonstrate that wogonin was able to effectively improve the manifestations, outcome of mice with LPS-or CLP-induced sepsis and suppress the levels of circulating pro-inflammatory cytokines.

| Wogonin mitigates liver injury in mice with sepsis
Sepsis is commonly characterized by the dysfunction of multiple organs, among which the damage of liver occurs frequently and can develop to fulminant hepatitis or a lethal syndrome of hepatocellular, metabolic and hemodynamic disorder as liver failure. 27,29 Therefore, septic liver injury has been documented as a crucial and independent indicator of the mortality of patients with sepsis. 29 We therefore went on to investigating the effects of wogonin on liver in sepsis mice.
Previous studies have demonstrated the histological malformation of liver tissue in sepsis mice, with different characteristics indicating distinct degrees of liver injury, which were used as histological scoring standard to evaluate the severity of liver injury. To be specific, swollen hepatocytes, disappearance of sinus hepaticus and slightly inflammatory cell infiltration suggest lower degree of liver injury, whereas obscure nucleus, vacuolar degeneration, ballooning degeneration of hepatocytes and severe inflammatory cell infiltration indicate moderate degree of liver injury. In addition, the above-mentioned alterations of moderate degree together with focal necrosis indicate higher degree of liver injury. 30,31 Through the H&E staining analysis, we found that liver tissue from the two sepsis models showed obscure nucleus (indicated by red arrows) and vacuolar degeneration (indicated by yellow arrows) in hepatocytes (Figure 2A,B). Excessive periportal neutrophil and macrophage infiltration (shown by blue arrows) were also found in liver tissue of LPS-and CLP-induced septic mice (Figure 2A,B). In contrast, the liver tissue isolated from wogonin-treated sepsis mice displayed mild histological malformation including swollen cells, obscure nucleus and vacuolar degeneration, as well as less infiltration of inflammatory cells (Figure 2A,B). In line with the histological appearance, the staining scores were significantly higher in liver tissue of either LPS-or CLP-induced sepsis mice compared with the control, whereas wogonin treatment could re-suppress the increased histological scores, indicating ameliorated septic liver injury ( Figure 2C,D). In line with this, our TUNEL staining showed that wogonin treatment prominently reduced the apoptotic rate of hepatocytes in liver tissue from these two types of sepsis mice ( Figure 2E,F and Figure S2A,B). The biomarkers ALT and AST in serum for reflecting liver injury were both prominently increased in sepsis mice, whereas significantly re-suppressed after wogonin treatment ( Figure 2G,H), forwardly supporting that wogonin could reduce the severity of liver injury in sepsis mice.
Considering the protective role of wogonin reported in drug-or heavy metal-induced kidney injury, 13,14 we also investigated the effect of wogonin on CLP-induced kidney injury. To evaluate the effect of wogonin on renal dysfunction, we examined levels of BUN and serum creatinine 12 hours after CLP operation ( Figure S2C,D). The results showed that CLP operation caused significant increase of BUN and creatinine, and wogonin treatment efficiently reduced both of the BUN and creatinine levels, indicating the therapeutic effect of wogonin on CLP-induced septic kidney dysfunction. Besides, we conducted H&E staining analysis of renal tissues to examine the histological malformation of renal tissue in CLP-induced sepsis mice. According to the previous study, 19 the percentage of damaged renal tubules was used to assess the score of tissue damage. The criteria of tubular damage included loss of brush border, tubular dilation, cast formation, inflammatory cell infiltration and cell lysis. Through H&E staining analysis, we found that renal tissue from CLP-induced sepsis mice showed more severe characteristics of renal histological damage, including cell necrosis or lysis (indicated by red arrows), cast formation (indicated by yellow arrows) and inflammatory cell infiltration (indicated by blue arrows) than the sham group ( Figure S2E). In contrast, the renal tissue from wogonin-treated sepsis mice displayed mild histological malformation ( Figure S2E). In line with the histological observation, the staining scores were significantly higher in renal tissues in CLP-induced sepsis mice compared with the sham group, whereas wogonin treatment could reduce the score of renal tissue damage, indicating ameliorated CLP-caused renal injury induced by wogonin treatment ( Figure S2F).

| Wogonin activates Nrf2 pathway and alleviates oxidative stress in hepatocytes
Accumulative evidence has revealed that oxidative stress plays a crucial role in the pathogenesis of septic liver injury. The excessive generation of ROS could not only directly cause the cell death of hepatocytes by augmenting the formation of lipid peroxides and oxidative protein adducts, but also amplify the pro-inflammatory signalling cascades to induce more severe systematic immune and inflammatory responses. [32][33][34][35] Therefore, we turned to study whether wogonin exerted its therapeutic effect on septic liver injury via the prevention of oxidative stress. As predicted, LPS injection or CLP surgery promoted the formation of lipid peroxidation product MDA in liver tissue of sepsis mice. Meanwhile, the activity of the crucial anti-oxidative enzyme SOD in liver was prominently repressed ( Figure 3A,B). After the treatment with wogonin, the accumulation of MDA in liver was significantly ameliorated, together with the increased activity of SOD ( Figure 3A,B), indicating the potent anti-oxidative capacity of wogonin in vivo. Previous reports have F I G U R E 1 Wogonin treatment improves the prognosis of mice with LPS-or CLP-induced sepsis. A, Schematic presentation of the establishment of LPS-and CLP-induced sepsis mouse model. B and C, Body temperature and respiratory rate changes 1 h before and 8 h after LPS injection or 2 h before and 12 h after CLP operation in sepsis mouse model (n = 6, respectively). D, Detection of serum TNFα, IL-1β, IFNγ and IL-6 in LPS-and CLP-induced sepsis mice (n = 6). E and F, Body temperature and respiratory rate changes after wogonin treatment in LPS-and CLP-induced sepsis mice (n = 6). G, Detection of serum TNFα, IL-1β, IFNγ and IL-6 after wogonin treatment in LPSand CLP-induced sepsis mice (n = 6). H, Survival analysis of LPS-and CLP-induced sepsis mice (n = 7) with or without wogonin treatment. The concentration of wogonin was 50.0 mg/kg. The results are presented as mean ± SD through at least 3 independent experiments and were analysed by one-way analysis of variance (ANOVA). The results of survival comparison were analysed by Kaplan-Meier survival curves and log-rank test (Mantel-Cox method). *P < .05, **P < .01, ***P < .001, ****P < .0001  Figure 3G). Furthermore, through the immunofluorescence staining analysis, we found that wogonin treatment promoted the nuclear translocation of Nrf2 ( Figure 3H), which is essential for its transcriptional activity that promotes the expressions of downstream anti-oxidative enzymes. Taken together, these results demonstrate that wogonin could alleviate oxidative stress in hepatocytes and effectively activate Nrf2 pathway.

| Wogonin prevents hepatocytes from LPS/ TNFα-induced apoptosis
It has been unveiled that Nrf2 possesses the anti-inflammatory capacity that is highly associated with the interplay between Nrf2 signalling and NF-κB signalling. The genetic deletion of Nrf2 can potentiate inflammation whereas its up-regulation suppresses proinflammatory responses controlled by NF-κB, thus contributing to the progression of chronic obstructive pulmonary disease (COPD), traumatic brain injury and some other diseases. [44][45][46][47] As we discovered that wogonin was able to activate Nrf2 and alleviate oxidative stress, we wondered whether it could restrain pro-inflammatory response in sepsis liver. Our qRT-PCR analysis in AML12 cells showed that while LPS combined with TNFα elevated the levels of TNFα, IL-6 and IL-1β, wogonin treatment prominently suppressed the expressions of these pro-inflammatory cytokines ( Figure 4A). In addition, our immunoblotting analysis revealed that in response to the stimulation with LPS combined with TNFα, the phosphorylation of NF-κB was significantly increased, whereas wogonin treatment could restrain the activation of NF-κB ( Figure 4B). These results indicated the great potential of wogonin in the prevention of inflammatory response in sepsis liver.

Our in vitro experiments in AML12 cells proved that wogonin
could simultaneously potentiate the activity of Nrf2-dependent antioxidative machinery and inhibit NF-κB-dependent pro-inflammatory pathway. Thereafter, we examined whether wogonin could protect hepatocytes against septic liver injury in vitro as well. The stimulation of LPS combined with TNFα resulted in robust increase of apoptosis of AML12 cells, which was significantly suppressed after the co-treatment with wogonin, especially in a dose-dependent manner ( Figure 4C). Moreover, we employed immunoblotting analysis to testify the expressions of apoptosis-related molecules. The expression of anti-apoptotic Bcl2 was markedly increased, whereas the expressions of pro-apoptotic Bax, Cleaved-Caspase3 and Cleaved-PARP were significantly decreased after the co-treatment with wogonin ( Figure 4D,E). Combined with the results in sepsis mice, our data proved that wogonin could ameliorate septic liver injury both in vitro and in vivo.

| Wogonin ameliorates liver injury in a Nrf2dependent manner in sepsis
Forwardly, we went on to investigating whether wogonin exerted its protective effects on septic liver injury via the activation of Nrf2. We transfected AML12 cells with siRNA against Nrf2 to obtain the knock-down of Nrf2 expression, followed by the treatment with LPS combined with TNFα and wogonin. As was shown, ( Figure 5A). In addition, the activity of anti-oxidative SOD potentiated by wogonin was also re-suppressed after the knock-down of Nrf2 ( Figure 5B), whereas the accumulation of intracellular ROS was potentiated by the knockdown of Nrf2 ( Figure 5C). Subsequent

| Wogonin activates Nrf2 signalling in liver tissue of sepsis mice
Finally, we employed IHC analysis to clarify the role of wogonin in Nrf2 signalling in the liver of sepsis mice. As was revealed, wogonin treatment was able to potentiate the staining scores of Nrf2 and its downstream HO-1, SOD1 and SOD2 in liver tissue from both LPSand CLP-induced sepsis mouse models ( Figure 6A-D and Figure S3A-D). In contrast, the staining score of phosphorylated-p65 (Ser536) was repressed ( Figure 6E and Figure S3E), which was in line with previous in vitro results. Therefore, we provided in vivo evidence that wogonin facilitated the activation of Nrf2 to enhance the antioxidative capacity and restrain the pro-inflammatory signalling, thus effectively ameliorating septic liver injury.

| D ISCUSS I ON
In the present study, we initially proved that wogonin treatment  51,52 Wogonin was found to exert protective role in endothelial cells through inhibition of high mobility group box 1 (HMGB1) in LPS-and CLP-induced sepsis and reduced sepsis-related mortality in mice. 53 Therefore, apart from the direct effect on hepatocyte reported in the present study, wogonin can also exert protective effect on septic liver injury by maintaining intrahepatic endothelial barrier, raising the notion that both of the localized hepatocyte and hepatocyte-associated microenvironment should be taken into consideration in case of the treatment of septic liver injury.
Abnormal activation of NF-κB signalling plays the pivotal role in inflammatory diseases such as asthma, rheumatoid arthritis and inflammatory bowel disease so that the antagonization of NF-κB signalling can be an effective therapeutic strategy for inflammatory diseases. 54 The crosstalk between Nrf2 and NF-κB signalling pathways has been frequently observed that anti-oxidative Nrf2 signalling pathway was able to inhibit NF-κB signalling activation by reducing DNA-binding activity of NF-κB. 55 In respiratory syncytial virus infection model, Nrf2 depletion strengthened DNA-binding activity of NF-κB in Nrf2 −/− mice and exacerbated inflammation together with more protein/lipid oxidation. 56 In acute liver injury induced by LPS/D-GalN injection, Nrf2 inhibited NF-κB DNA binding by reducing MafK expression, the transcription factor which mediated p65 acetylation and its DNA-binding affinity. 57 What's more, as the redox-sensitive transcriptional factor, 58 NF-κB has been proved to be activated by oxidative stress since that intracellular ROS induced the serine phosphorylation of the NF-κB subunit p65 and thereby facilitated its nuclear translocation and downstream gene expression like IL-1β. 59 In Nrf2 −/− septic shock mice, NF-κB signalling was magnified in lung and peritoneal macrophages whereas antioxidants N-acetyl cysteine and GSH-monoethyl ester suppressed NF-κB activation, which implied that anti-oxidative agents could inhibit NF-κB signalling by alleviating oxidative stress and re-suppressed Nrf2 depletion-induced NF-κB activation. 60 The present study consistently showed that suppression of oxidative stress induced by Nrf2 signalling activation restrained p65 phosphorylation and proinflammatory cytokine production, reflecting that anti-oxidative capacity of wogonin could also attenuate liver injury by suppressing NF-κB signalling pathway, which suggests that the manipulation of anti-oxidative stress machinery could be referred as alternative strategy for sepsis and other inflammation-associated disorders. In conclusion, the present study showed that wogonin could ameliorate F I G U R E 6 Wogonin activates Nrf2 signalling in liver tissue of LPS-induced sepsis mice. A-E, Representative IHC images and IHC score analysis of Nrf2, HO-1, SOD1, SOD2 and p-p65 (Ser536) in liver tissue from LPS-induced septic mice (n = 5). The number of scored fields = 10. Magnification, 400×. Scale bar = 50 μm. The concentration of wogonin was 50.0 mg/kg. The results are presented as mean ± SD through at least 3 independent experiments and were analysed by one-way analysis of variance (ANOVA). ns, no statistical significance, *P < .05, **P < .01, ***P < .001 F I G U R E 7 Schematic presentation of the mechanism of wogonin to alleviate septic liver injury. Wogonin treatment activated Nrf2 signalling, promoted transcription and expression of endogenous antioxidants including NQO-1, GST, HO-1, SOD which suppressed sepsisinduced oxidative stress in hepatocytes. Moreover, wogonin-promoted Nrf2 signalling inhibited NF-κB activation and decreased production of pro-inflammatory cytokines, which mitigated inflammatory damage of hepatocytes in sepsis Wang and Yan Qiu for technical assistance.

CO N FLI C T O F I NTE R E S T
The authors confirm that there are no conflicts of interest.

DATA AVA I L A B I L I T Y S TAT E M E N T
The data generated and analysed during the current study are available from the corresponding author on reasonable request.