Withaferin A alleviates fulminant hepatitis by targeting macrophage and NLRP3

Fulminant hepatitis (FH) is an incurable clinical syndrome where novel therapeutics are warranted. Withaferin A (WA), isolated from herb Withania Somnifera, is a hepatoprotective agent. Whether and how WA improves D-galactosamine (GalN)/lipopolysaccharide (LPS)-induced FH is unknown. This study was to evaluate the hepatoprotective role and mechanism of WA in GalN/LPS-induced FH. To determine the preventive and therapeutic effects of WA, wild-type mice were dosed with WA 0.5 h before or 2 h after GalN treatment, followed by LPS 30 min later, and then killed 6 h after LPS treatment. To explore the mechanism of the protective effect, the macrophage scavenger clodronate, autophagy inhibitor 3-methyladenine, or gene knockout mouse lines NLR family pyrin domain containing 3 (Nlrp3)-null, nuclear factor-erythroid 2-related factor 2 (Nrf2)-null, liver-specific AMP-activated protein kinase (Ampk)a1 knockout (Ampka1ΔHep) and liver-specific inhibitor of KB kinase β (Ikkb) knockout (IkkbΔHep) mice were subjected to GalN/LPS-induced FH. In wild-type mice, WA potently prevented GalN/LPS-induced FH and inhibited hepatic NLRP3 inflammasome activation, and upregulated NRF2 and autophagy signaling. Studies with Nrf2-null, Ampka1ΔHep, and IkkbΔHep mice demonstrated that the hepatoprotective effect was independent of NRF2, hepatic AMPKα1, and IκκB. Similarly, 3-methyladenine cotreatment failed to abolish the hepatoprotective effect of WA. The hepatoprotective effect of WA against GalN/LPS-induced FH was abolished after macrophage depletion, and partially reduced in Nlrp3-null mice. Consistently, WA alleviated LPS-induced inflammation partially dependent on the presence of NLRP3 in primary macrophage in vitro. Notably, WA potently and therapeutically attenuated GalN/LPS-induced hepatotoxicity. In conclusion, WA improves GalN/LPS-induced hepatotoxicity by targeting macrophage partially dependent on NLRP3 antagonism, while largely independent of NRF2 signaling, autophagy induction, and hepatic AMPKα1 and IκκB. These results support the concept of treating FH by pharmacologically targeting macrophage and suggest that WA has the potential to be repurposed for clinically treating FH as an immunoregulator.


Introduction
Fulminant hepatitis (FH) is a global life-threatening clinical syndrome with no curable drugs and measures available other than liver transplantation, and thus novel drug discovery is imperative 1,2 . In the clinic, lipopolysaccharide (LPS) released from intestinal bacterial stimulates macrophage to release tumor necrosis factor (TNF)-α [3][4][5] . While LPS alone only induces systematic inflammation and LPS pretreatment can even ameliorate hepatotoxicity 3,6,7 , LPS-induced hepatotoxicity is usually enhanced when livers are pre-exposed to various other hepatic insults that are frequently involved with humans in the clinic, such as viruses, toxins and alcohol 1,2,8 . These hepatic insults could deplete hepatic glutathione levels 9,10 and/or inhibit the protein and RNA transcription, both of which could sensitize hepatocytes to TNF-α-induced apoptosis [11][12][13] . To mimic the clinical FH, the D-galactosamine (GalN) and LPS-induced acute liver injury model has been established in mice 7,14,15 . GalN/LPS induces FH by GalN-induced inhibition of protein and RNA synthesis that sensitizes hepatocytes to TNF-α-induced apoptosis 16 , followed by TNF-α release through LPS-stimulated macrophage 14 . Therefore, FH progresses through two processes, sensitizing livers by GalN, and LPS activation of macrophage to release TNFα, which together contribute to extensive TNF-α-induced hepatocyte apoptosis 17,18 .
Traditional herbs have been extensively investigated for treating liver diseases both preclinically and clinically [37][38][39][40] . Withaferin A (WA), a natural steroidal lactone isolated from the traditional herb Withania Somnifera, possesses various pharmacological activities, such as antitumor, antiinflammation 41 , and also acts as a leptin sensitizer 42 . WA potently improves acetaminophen-induced hepatotoxicity 43,44 and diet-induced nonalcoholic steatohepatitis 45 in rodents. However, whether and how WA improves GalN/ LPS-induced FH has not been explored. Mechanically, WA is an NRF2 inducer and WA NRF2-dependently protects against acetaminophen-induced hepatotoxicity 44 . WA also inhibits inflammation by directly inhibiting IκκB activity 46,47 or NLRP3 inflammasome activation in vitro in immune cells 48,49 , and WA was suggested to AMPKαdependently protect against cardiovascular disease 50 . Thus, WA probably protects against FH by targeting the macrophage and/or hepatocyte stress via activating NRF2, AMPKα, autophagy, inhibiting IκκB activity or antagonizing the NLRP3 inflammasome activation. In this study, the effects and mechanisms of WA in GalN/LPSinduced FH were investigated, particularly concentrating on the regulation of NLRP3, NRF2, AMPKα, and autophagy signaling.

WA attenuated GalN/LPS-induced hepatotoxicity and the inflammatory response
To test the preventive effect of WA (structure shown in Fig. 1a) in GalN/LPS-induced FH, mice were pretreated as schemed (Fig. 1b). Serum alanine aminotransferase (ALT) and aspartate aminotransferase (AST) levels were slightly increased at 3 h and markedly increased at 6 h, both of which were decreased by WA (Fig. 1c). WA also sharply attenuated the GalN/LPS-induced histological liver damage, consistent with the normalized gross liver appearance (Fig. 1d). Furthermore, WA significantly improved the survival rate of GalN/LPS-treated mice (Fig.  1e). While FH is accompanied by cytokine release and hepatic inflammation, WA significantly reduced the GalN/LPS-induced increase of serum IL-1β, IL-6, and TNF-α ( Fig. 2a-c). WA significantly attenuated the GalN/ LPS-induced increase of hepatic Il1b and Il6 mRNAs, with a tendency to decrease Tnfa mRNA ( Fig. 2d-f). Thus, WA potently protects against GalN/LPS-induced hepatotoxicity and inflammation.
WA suppressed hepatic apoptosis in vivo, but not TNF-α-induced hepatocyte death in vitro Typical apoptosis markers, terminal deoxynucleotidyl transferase dUTP nick end labeling (TUNEL) staining, poly-(ADP-ribose) polymerase (PARP1) and caspase 3 (CASP3) cleavage, were next examined to determine whether WA rescued GalN/LPS-induced hepatocyte apoptosis in vivo. WA was found to significantly decrease GalN/LPS-induced positive TUNEL staining (Fig. 3a, b) and attenuate the increase of cleaved CASP3 and cleaved PARP1 (Fig. 3c, d). To test whether WA has a direct antiapoptosis effect, the ACTD/TNF-α and GalN/TNF-αinduced apoptosis models were employed in primary mouse hepatocytes. As expected, either ACTD or GalN synergized with TNF-α to induce a decrease in cell viability. However, WA pretreatment at 0.01-0.5 μM had no significant effect in reducing ACTD/TNF-α or GalN/TNF-αinduced cell death, while ZVAD efficiently prevented cell death in both models as a positive control ( Supplementary  Fig. S1). The doses of WA were chosen based on a previous publication 43 and no toxicity of WA was found at 0.01-0.5 μM (Supplementary Fig. S1). Thus, WA alleviates GalN/LPS-induced apoptotic hepatocyte death in vivo, but shows no direct effect against TNF-α-induced hepatocyte apoptosis in vitro.  c Serum ALT and AST levels at 3 h and 6 h after GalN/LSP treatment. d Representative liver images and liver histological H&E staining, scale bar 50 μm. e Survival curve (n = 10). Data are presented as means ± SD (n = 5 unless otherwise indicated). Control, mice dosed with control vehicle and saline; WA only, mice dosed with WA and saline; V + GalN/LPS, mice dosed with control vehicle (V), and GalN/LPS; WA + GalN/LPS, mice dosed with WA and GalN/LPS. One-way ANOVA was used for statistical analyses. # p < 0.05, ## p < 0.01, and ### p < 0.005 versus Control group; *p < 0.05, **p < 0.01, and ***p < 0.005 versus V + GalN/LPS group.
were recovered to levels comparable with that of control mice by WA treatment (Fig. 3e). While the hepatic levels of oxidized glutathione (GSSG) remained unchanged between groups (Fig. 3f), the GSH/GSSG ratio was significantly decreased and rescued by WA (Fig. 3g), indicative of normalized oxidative stress. Further analyses for how WA affected the expression of NRF2 target gene mRNAs was next performed. In the absence of GalN/LPS treatment, WA significantly upregulated the mRNA levels of several NRF2 target genes in WT mice, and not in Nrf2 −/− mice ( Supplementary Fig. S3a), suggesting an NRF2-dependent effect. GalN/LPS was found to induce NRF2 protein levels when compared with the control group, consistent with a previous study showing a similar role for LPS in inducing NRF2 protein expression although decreasing Nrf2 mRNA expression 20 . After GalN/LPS administration, WA treatment was found to further increase hepatic NRF2 protein ( Supplementary  Fig. S2a), and to significantly increase the mRNA levels of NRF2 was not required for the hepatoprotective effect of WA WA is known to induce NRF2 signaling and NRF2dependently alleviates acetaminophen-induced liver injury 43,44 . To further determine if WA protects against FH depending on NRF2, Nrf2 −/− mice were employed. The results revealed that WA sharply decreased serum ALT levels in GalN/LPS-treated Nrf2 −/− mice (Supplementary Fig. S2c), and this hepatoprotective effect was confirmed by hematoxylin and eosin (H&E) staining ( Supplementary Fig. S2d). Analysis of hepatic Nrf2 mRNA levels validated the Nrf2 knockout efficiency of Nrf2 −/− mice ( Supplementary Fig. S3a). These data demonstrate , and Tnfa (f). Data are presented as means ± SD (n = 5 unless otherwise indicated). Groups were same as described in Fig. 1 legend. One-way ANOVA was used for statistical analyses. ### p < 0.005 versus Control group; **p < 0.01 and ***p < 0.005 versus V + GalN/ LPS group.
that the presence of NRF2 does not influence the hepatoprotective effect of WA in treating GalN/LPS-induced liver injury, at least under the present experiment conditions.

WA induced hepatic autophagy in GalN/LPS-treated mice
Previously, another herbal NRF2 activator, licochalcone A, potently alleviated GalN/LPS-induced liver injury in Nrf2 −/− mice by inducing autophagy 31 . Thus, whether WA alleviated the FH via induced autophagy similar to licochalcone A was examined. WA significantly rescued the GalN/LPS-induced decrease of autophagy related 3 (ATG3), microtubule-associated protein light chain 3A/B (LC3II), the ratio of LC3II/LC3I (Fig. 4a, b), indicating that WA upregulated hepatic autophagy signaling in GalN/LPS-treated mice. . c Western blot analysis of CASP3, cleaved PARP1 and cleaved CASP3 expression. d quantitation analyses of cleaved PARP1, CASP3 and cleaved CASP3 (n = 3). e Hepatic GSH levels. f Hepatic GSSG levels. g Hepatic GSH/GSSG ratio. Data are presented as means ± SD (n = 5 unless otherwise indicated). Groups were the same as described in Fig. 1 legend. One-way ANOVA or unpaired two-tailed t test was used for statistical analyses. # p < 0.05, ## p < 0.01, ### p < 0.005 versus Control group; *p < 0.05, **p < 0.01, ***p < 0.005 versus V + GalN/LPS group.  3). b Quantitation analysis of relative protein levels (n = 3). c Experimental time scheme. d Effect of 3-MA in the expression of ATG3, LC3I, and LC3II. e Serum ALT and AST levels in control saline-pretreated mice treated with or without WA. f Serum ALT levels and AST levels in 3-MApretreated mice treated with or without WA. g Representative liver histological H&E staining of control saline-pretreated and 3-MA-pretreated mice treated with or without WA, scale bar 50 μm. Data are presented as means ± SD (n = 5 unless otherwise indicated). One-way ANOVA or unpaired two-tailed t test was used for statistical analyses. ## p < 0.01, ### p < 0.005 versus Control; *p < 0.05, **p < 0.01 and ***p < 0.005 versus V + GalN/LPS. Autophagy, hepatocyte AMPKα, or IκκB were not required for the protective effect of WA 3-MA, as a widely-used autophagy inhibitor 51 , was then applied to block WA-induced autophagy. Mice were treated as designed based on previous publications (Fig.  4c) 20,22,31,52,53 . 3-MA at the tested dose, significantly inhibited the expression of ATG3 and decreased the LC3II/LC3I ratio, two typical autophagy markers (Fig. 4d), indicative of the efficient autophagy inhibition by 3-MA. In 3-MA-pretreated mice, WA significantly decreased the GalN/LPS-induced increase of the serum ALT and AST levels to a similar extent with control vehicle-pretreated mice (Fig. 4e, f), consistent with the H&E staining (Fig. 4g). Therefore, 3-MA does not abolish the hepatoprotective effect of WA. WA was found to protect against myocardial ischemia/ reperfusion injury via AMPKα activation 50 , while hepatic AMPKα1 activation protects against GalN/LPS-induced liver injury 22,39 . Whether WA protected against GalN/ LPS-induced FH by AMPKα activation was investigated. WA showed no significant effect on expression of AMPKα and p-AMPKα proteins ( Supplementary Fig. S4a,  b). Ampka1 ΔHep mice were next employed and analyses of serum ALT levels revealed that the hepatoprotective effect of WA in Ampka1 ΔHep mice was as potent as in the Ampka1 fl/fl littermates ( Supplementary Fig. S4c), supporting an AMPKα1-independent hepatoprotective effect of WA. Western blot analyses confirmed that the AMPKα1 expression was efficiently depleted in livers of Ampka1 ΔHep mice ( Supplementary Fig. S4d, e). While WA directly targets IκκB to produce its anti-inflammatory effects 46,47 , WA still showed a potent effect in alleviating GalN/LPS-induced liver injury in Ikkb ΔHep mice by analyses of serum ALT levels ( Supplementary Fig. S4f) and H&E staining (Supplementary Fig. S4g). Thus, the hepatoprotective effect of WA is independent of hepatocyte AMPKα1, hepatocyte IκκB and hepatic autophagy signaling.

Macrophages depletion abrogated the hepatoprotective effect of WA
Given that WA still elicited a potent effect in Nrf2 −/− mice, autophagy-inhibited WT mice, and Ampka1 ΔHep and Ikkb ΔHep mice, the hepatoprotective effect of WA was then suspected to not directly target hepatocytes, but target macrophage, another key determinant of hepatotoxicity. To answer this question, clodronate liposomes were used to deplete macrophage, and mice were treated as schemed (Fig. 5a). In control liposome-injected mice, WA alleviated the GalN/LPS-induced increase of serum ALT and AST levels (Fig. 5b), while upon clodronate injection, WA failed to significantly decrease serum ALT and AST levels (Fig. 5c). Liver F4/80 mRNA was markedly decreased in clodronate-treated mice compared with control liposome treatment (Fig. 5d), indicative of efficient macrophage depletion by clodronate. Further analyses of liver histology and hepatic proinflammatory cytokines mRNA expression confirmed that the hepatoprotective effect of WA was largely compromised in clodronateinjected mice compared with that in the control liposome-injected mice (Fig. 5e-g). Thus, WA targets macrophage to protect against GalN/LPS-induced FH.
WA prevented GalN/LPS-induced FH partially by inhibiting activation of the NLRP3 inflammasome NLRP3 inflammasome activation is a well-documented pathway for inflammation induction in macrophage in vitro and in various diseases models in rodents 30,35,54,55 , while WA directly antagonizes NLRP3 inflammasome activation in vitro 48,49 . Thus, the question arose whether WA attenuated GalN/LPS-induced FH by inhibiting NLRP3 inflammasome activation. The time course of NLRP3 inflammasome-mediated inflammation signaling activation was examined, and western blot analyses demonstrated that interleukin 1β (IL-1β) was significantly induced as early as 1 h, which peaked 3 h after LPS dosing, and the protein levels of both NLRP3 and ASC were significantly upregulated and peaked 6 h post LPS dosing ( Supplementary Fig. S5a, b). The time-course change of NLRP3 inflammasome signaling correlated well with the time-course change of serum ALT and AST levels, both of which were also sharply upregulated 6 h post LPS dosing, while only showing a slight increase 3 h post LPS dosing (Fig. 1c). Given that all tested proteins involved in NLRP3 inflammasome signaling were consistently upregulated 6 h post LPS dosing, how WA affected the hepatic NLRP3 inflammasome signaling activation in GalN/LPS-dosed mice was next investigated. WA significantly attenuated the GalN/LPS-induced increase of the proteins including ASC, cleaved-CASP1 and IL-1β in liver, all of which play a key role in NLRP3 inflammasome activation (Fig. 6a, b). To examine whether the inhibitory effect of WA in hepatic NLRP3 inflammasome activation is a cause or result of its hepatoprotective effect, Nlrp3 −/− mice were employed. Analyses of serum ALT and AST levels showed that Nlrp3 deficiency did not abrogate, but significantly reduced the hepatoprotective effect of WA by~20% percent (Fig. 6c), which were further supported by histology analyses (Fig. 6d). Thus, WA alleviated GalN/LPSinduced FH partially dependent on NLRP3.
To test whether the direct anti-inflammatory role of WA depends on NLRP3, primary macrophages were isolated from WT mice and matched Nlrp3 −/− mice. The knockout efficacy of Nlrp3 was validated ( Supplementary  Fig. S5c). In WT macrophage, WA, at 0.2 µM and 0.5 µM, dose-dependently attenuated LPS-induced Tnfa, Il6, Il1b mRNAs ( Fig. 5e-g) and Nlrp3 mRNA ( Supplementary  Fig. S5d). The inhibitory effects of WA in LPS-induced f Hepatic mRNA levels of inflammation pathway including Tnfa, Il6, Il1b, and Ccl2 in clodronate liposomes-pretreated mice. g Representative images of liver H&E staining, scale bar 50 µm. Data are presented as means ± SD, n = 5. Groups were the same as described in Fig. 1 legend. One-way ANOVA was used for statistical analyses. # p < 0.05, ## p < 0.01, and ### p < 0.005 versus Control group; ***p < 0.005 versus V + GalN/LPS group. upregulation of Tnfa, Il6, Il1b, were lost in Nlrp3 −/− macrophage when WA was used at the lower dose of 0.2 µM, while this effect was still found with WA at the higher dose of 0.5 µM in Nlrp3 −/− macrophage ( Fig.  6e-g), suggesting that WA has both NLRP3-dependent and NLRP3-independent anti-inflammatory effects. Consistent with the possibility that WA has an NLRP3independent anti-inflammatory effect, LPS-induced upregulation of chemokine (C-C motif) ligand 2 (Ccl2) mRNA, which does not act directly downstream of NLRP3, was dose-dependently decreased by WA at 0.2 µM and 0.5 µM in both WT macrophage and Nlrp3 −/− macrophage (Fig. 6h). Notably, LPS-induced upregulation of Tnfa, Il6, Il1b and Ccl2 mRNAs were all markedly decreased in Nlrp3 −/− macrophage compared with WT macrophage (Fig. 6e-h), indicating a key role for NLRP3 in mediating LPS-induced inflammation. WA at 0.2 µM and 0.5 µM was nontoxic to both WT and Nlrp3 −/− macrophage ( Supplementary Fig. S5e, f). These data suggest that WA produces a strong anti-inflammatory effect in macrophages, which is only partially dependent on NLRP3.

WA therapeutically improved GalN/LPS-induced hepatotoxicity
To further explore the therapeutic potential of WA in GalN/LPS-induced hepatotoxicity, a therapeutic dosing scheme was designed to treat WT mice (Fig. 7a). Notably, WA was found to markedly alleviate the GalN/LPSinduced increase of serum ALT and AST levels as well as histological damage even when WA was dosed 2 h after GalN dosing compared with control vehicle (Fig. 7b, c). These data support the view that WA has a potent therapeutic effect in GalN/LPS-induced liver injury in mice.

Discussion
FH, also known as acute liver failure, could lead to lethality with liver transplantation as the only curable treatment 1,2 . Thus, pharmacotherapeutic discoveries are needed to treat this disease. WA was reported as a hepatoprotective agent in acetaminophen-induced hepatotoxicity mainly due to hepatocyte necrosis 43,44 , and dietinduced nonalcoholic steatohepatitis 45 that is characterized by hepatic steatosis, inflammation and fibrosis, while little is known about its role and mechanism in alleviating GalN/LPS-induced FH that is associated with apoptotic hepatocyte death and inflammation.
In the present study, hepatocellular stress alleviation that could be mediated by rescuing expression of hepatic NRF2 target genes, hepatic IκκB inhibition, and protective autophagy induction via AMPKα signaling were first excluded as contributors to the hepatoprotective effects of WA by using Nrf2 −/− mice, Ikkb ΔHep mice, Ampka1 ΔHep mice and 3-MA-injected WT mice under the conditions used in this study. The hepatoprotective properties of WA is likely due to effects on macrophage rather than by directly protecting hepatocytes. The protective effects of WA was sharply deminished by macrophage depletion, supporting the view that WA targets the macrophage to treat experimental FH. Further studies demonstrated that the protective effect of WA against the GalN/LPS-induced FH was partially dependent on NLRP3 inflammasome antagonism in vivo and has a direct anti-inflammatory influence partially dependent on the presence of NLRP3 in macrophage in vitro, indicating that other targets, beyond NLRP3, mediate the anti-inflammatory effect of WA. WA inhibits inflammation by direct hyperphosphorylation of IκκB to inhibit the IκκB activity 46,47 . While the possibility that hepatocyte IκκB plays a role in the hepatoprotective effect of WA on GalN/LPS-induced FH was excluded by using Ikkb ΔHep mice, it is still possible that WA targets the IκκB present in macrophage to produce its anti-inflammatory effects. Thus, molecular targets present in macrophage other than NLRP3, such as IκκB, that mediate the anti-inflammatory effect of WA deserve further study.
Previously, WA was found to alleviate acetaminopheninduced liver injury dependent on NRF2 induction 44 . In contrast, the hepatoprotective effect of WA was currently found to be independent of NRF2 in the GalN/LPS model as revealed by using Nrf2 −/− mice. This may be due to the different pathological hepatotoxicity mechanisms for these two liver injury models, where acetaminophen mainly induces hepatocyte necrosis 56 and GalN/LPS mainly initiates hepatocyte apoptosis 15 . More importantly, the roles of NRF2 deficiency in these two models are different. Nrf2 −/− mice were protected from GalN/LPSinduced FH 31 , but showed an enhanced hepatotoxicity in acetaminophen-treated mice 57 . Consistent with these findings, another herbal NRF2 activator, licochalcone A, also showed a potent hepatoprotective effect in GalN/ LPS-treated Nrf2 −/− mice 31 . This same study demonstrated that the autophagy inhibitor 3-MA did not abolish the hepatoprotective effect of licochalcone A 31 . Similarly, the current study found that 3-MA did not abolish the hepatoprotective effect of WA. These results suggest that the NRF2 presence and autophagy induction play only minor roles in mediating the hepatoprotective effects of these herbs at the employed experiment conditions. Thus, WA targets macrophage instead of directly protecting hepatocyte stress pathways to alleviate GalN/LPS-induced FH, at least under the present experiment conditions. Evidence in the current study reveal that the macrophage-NLRP3 axis at least partially determines the hepatoprotective effect of WA. Indeed, previous studies also support the concept of targeting macrophage to alleviate liver injury 23,33,34,36,37,[58][59][60] and that pharmacotherapies alleviate GalN/LPS-induced liver injury via inhibiting NLRP3 inflammasome activation 30,61 . Both macrophage and NLRP3 have pleotropic effects in maintaining systemic homeostasis. For example, macrophage are divided into two polarized types, the M2-like macrophage that are hepatoprotective and the M1-like macrophage that stimulate inflammation 62 . NLRP3 is physiologically beneficial in maintaining metabolic and immune homeostasis but could be detrimental upon overactivation 63,64 . In the current study, the hepatoprotective effect of WA on GalN/LPS-induced FH depends on the presence of macrophage and partially dependent on NLRP3, which may be explained by the potent inhibitory effects of WA on the adverse events caused by overactivation of the macrophage-NLRP3 axis in the GalN/LPS model. However, no statistically significant difference in sensitivity to GalN/LPS-induced FH was found between Nlrp3 -/mice and the background strainmatched WT mice, as well as between macrophagedepleted mice and vehicle-treated control mice, which could result from a pooled effect of both the diminished beneficial effects and harmful effects upon loss of macrophage or the Nlrp3 gene, or due to the potent toxic effects of GalN/LPS that overwhelm the potential influence of macrophage and NLRP3 on the disease phenotype. These results suggest that macrophage could have different effects between modulating the onset of the disease and WA suppression of the disease. In line with this point, inflammasome-deficient mice have increased severity to induced nonalcoholic steatohepatitis 65 , and NLRP3 inflammasome overactivation is required for fibrosis development and thus antagonism of NLRP3 inflammasome activation reduces nonalcoholic fatty liver disease 66,67 .
In summary, the current study demonstrates a novel role for WA in both preventing and therapeutically alleviating GalN/LPS-induced FH via targeting the macrophage that is partially dependent on NLRP3 (Fig. 7d). WA, although having effects of inducing hepatic NRF2 and autophagy signaling, the dependence of its hepatoprotective effect in these two signaling pathways is minor under the experimental conditions used in this study. These results demonstrate that WA targets macrophages to treat GalN/LPS-induced FH, which further support the possibility of targeting hepatic macrophage to treat liver disease 36 and support the potential of WA to be repurposed as a hepatoprotective agent against the clinical FH, particularly featured with TNF-α-related apoptosis.

Animal Experiments
Wild-type (WT) mice, Nrf2 −/− , Nlrp3 −/− , or Ampka1 fl/fl mice on a C57BL/6J background were purchased from the Jackson Laboratory, while Ikkb ΔHep mice on the C57BL/6J background were described previously 68,69 . All mice were housed in the National Cancer Institute animal facility that was a pathogen-free environment controlled for temperature, light (25°C, 12-h light/dark cycle) and humidity (45-65%). Albumin-Cre mice as described previously 69 were bred with Ampka1 fl/fl mice to generate Ampka1 ΔHep mice. Age and body weight-matched 6-to 8-week-old males were randomized into groups (at least N = 5) and treated with 700 mg/kg of GalN and 50 µg/kg of LPS or other reagents as described in detail in the Supplementary materials. The National Cancer Institute Animal Care and Use Committee approved all animal experiments conducted in this study.

Cell culture
Primary hepatocytes, isolated as described previously 70 , were pretreated with WA at 0, 0.01, 0.1, and 0.5 μM or 10 μM ZVAD for 30 min, followed by 5 mM of GalN or 0.3 μM ACTD for 30 min, and then 25 ng/mL of recombinant mouse TNF-α for additional 13 h. Cell counting kit-8 reagent was used to measure cell viability at 13 h after TNF-α treatment. Dimethyl sulfoxide (0.1%) was used as control vehicle. Primary peritoneal macrophage were isolated as described previously with minor modifications 54 . The macrophage were pretreated with WA at 0, 0.2, and 0.5 µM for 2 h, followed by 100 ng/mL of LPS treatment for additional 18 h. Other details can be found in Supplementary Materials.

Blinding, randomization, statistical analysis, and reproducibility of experiments
The investigators were blind to the experiment treatments. The experiments were repeated at least twice. The animal experiments were randomized. Statistical analyses were determined by two-tailed unpaired student's t test between two groups or by one-way ANOVA followed by Dunnett's multiple comparisons test among multiple groups for independent samples using GraphPad Prism 7.0 (San Diego, CA, USA). To predetermine sample sizes, power analysis was performed using StatMate version 2.0 (GraphPad Software). All data were presented as mean ± SD. P values < 0.05 were considered statistically significant.