What this is
- () lacks effective treatments, necessitating new therapeutic strategies.
- Withaferin A (WA), derived from Withania Somnifera, shows hepatoprotective effects.
- This study evaluates WA's impact on induced by D-galactosamine (GalN) and lipopolysaccharide (LPS) in mice.
Essence
- Withaferin A significantly alleviates in mice by targeting macrophages and inhibiting activation, independent of NRF2 signaling.
Key takeaways
- WA effectively prevents GalN/LPS-induced in wild-type mice, indicating its potential as a therapeutic agent.
- The hepatoprotective effect of WA is primarily mediated through macrophage targeting and is partially dependent on NLRP3 antagonism.
- WA's protective effects are independent of NRF2 signaling, AMPKα activation, and autophagy induction, suggesting alternative mechanisms at play.
Caveats
- The study's findings are based on animal models, which may not fully translate to human conditions.
- Further research is required to clarify the specific pathways through which WA exerts its effects on macrophages.
Definitions
- Fulminant hepatitis (FH): A severe form of liver failure characterized by rapid onset and high mortality, often requiring liver transplantation.
- NLRP3 inflammasome: A multi-protein complex that activates inflammatory responses, playing a key role in various diseases, including liver injury.
AI simplified
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 imperative1,2. In the clinic, lipopolysaccharide (LPS) released from intestinal bacterial stimulates macrophage to release tumor necrosis factor (TNF)-α3–5. While LPS alone only induces systematic inflammation and LPS pretreatment can even ameliorate hepatotoxicity3,6,7, 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 alcohol1,2,8. These hepatic insults could deplete hepatic glutathione levels9,10 and/or inhibit the protein and RNA transcription, both of which could sensitize hepatocytes to TNF-α-induced apoptosis11–13. To mimic the clinical FH, the D-galactosamine (GalN) and LPS-induced acute liver injury model has been established in mice7,14,15. GalN/LPS induces FH by GalN-induced inhibition of protein and RNA synthesis that sensitizes hepatocytes to TNF-α-induced apoptosis16, followed by TNF-α release through LPS-stimulated macrophage14. Therefore, FH progresses through two processes, sensitizing livers by GalN, and LPS activation of macrophage to release TNF-α, which together contribute to extensive TNF-α-induced hepatocyte apoptosis17,18.
Many targets in the liver and macrophage, which could contribute to hepatocellular stress including oxidative stress, inflammation and hepatocyte death, mediate GalN/LPS-induced FH4,19–25. Nuclear factor-erythroid 2-related factor 2 (NRF2) activation by various drugs protects against hepatic oxidative stress26,27 and the related liver injury28–31. Proinflammatory cytokines released from macrophage usually induce hepatic inflammation, and TNF-α among these cytokines, could induce liver apoptosis in GalN-pre-sensitized hepatocytes9,10,16. Protective autophagy, which could be regulated by AMP-activated protein kinase (AMPK) activation, protects against GalN/LPS-induced liver injury by alleviating apoptosis22,32. On the other hand, macrophage activation-induced inflammation also mediates the progression of hepatotoxicity33,34, among which NLRP3 inflammasome activation is one most well-documented mechanisms for activating inflammation30,35. Accordingly, strategies were proposed to target macrophage/NLRP3 activation for treating liver disease36,37.
Traditional herbs have been extensively investigated for treating liver diseases both preclinically and clinically37–40. Withaferin A (WA), a natural steroidal lactone isolated from the traditional herb Withania Somnifera, possesses various pharmacological activities, such as antitumor, anti-inflammation41, and also acts as a leptin sensitizer42. WA potently improves acetaminophen-induced hepatotoxicity43,44 and diet-induced nonalcoholic steatohepatitis45 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 hepatotoxicity44. WA also inhibits inflammation by directly inhibiting IκκB activity46,47 or NLRP3 inflammasome activation in vitro in immune cells48,49, and WA was suggested to AMPKα-dependently protect against cardiovascular disease50. 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/LPS-induced FH were investigated, particularly concentrating on the regulation of NLRP3, NRF2, AMPKα, and autophagy signaling.
Results
WA attenuated GalN/LPS-induced hepatotoxicity and the inflammatory response
WA protected against GalN/LPS-induced FH in WT mice. The chemical structure of WA.Dosing scheme of animal experiments.Serum ALT and AST levels at 3 h and 6 h after GalN/LSP treatment.Representative liver images and liver histological H&E staining, scale bar 50 μm.Survival curve (= 10). Data are presented as means ± SD (= 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.< 0.05,< 0.01, and< 0.005 versus Control group; *< 0.05, **< 0.01, and ***< 0.005 versus V + GalN/LPS group. a b c d e n n p p p p p p # ## ###
WA suppressed GalN/LPS-induced inflammatory response. Serum IL-1β levels (), Serum IL-6 levels (), and Serum TNF-α levels (). mRNA levels of hepatic()(), and(). Data are presented as means ± SD (= 5 unless otherwise indicated). Groups were same as described in Fig.legend. One-way ANOVA was used for statistical analyses.< 0.005 versus Control group; **< 0.01 and ***< 0.005 versus VGalN/LPS group. a b c d e f Il1b , Il6 Tnfa n p p p 1 ### +
WA suppressed hepatic apoptosis in vivo, but not TNF-α-induced hepatocyte death in vitro
WA suppressed hepatic apoptosis and alleviated hepatic oxidative stress. Representative TUNEL staining of liver sections.Statistical analysis of TUNEL staining (= 3).Western blot analysis of CASP3, cleaved PARP1 and cleaved CASP3 expression.quantitation analyses of cleaved PARP1, CASP3 and cleaved CASP3 (= 3).Hepatic GSH levels.Hepatic GSSG levels.Hepatic GSH/GSSG ratio. Data are presented as means ± SD (= 5 unless otherwise indicated). Groups were the same as described in Fig.legend. One-way ANOVA or unpaired two-tailedtest was used for statistical analyses.< 0.05,< 0.01,< 0.005 versus Control group; *< 0.05, **< 0.01, ***< 0.005 versus V + GalN/LPS group. a b c d e f g n n n t p p p p p p 1 # ## ###
WA alleviated hepatic oxidative stress and rescued the expression of NRF2 target genes in GalN/LPS-treated mice
The effect of WA on oxidative damage was explored. Hepatic glutathione (GSH) levels were significantly depleted by ~50% 6 h after GalN/LPS administration and 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 expression20. 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 target genes in both WT mice (Supplementary Fig. S2b) and Nrf2−/− mice (Supplementary Fig. S3b). Given that NRF2 target gene mRNAs were sharply decreased by GalN/LPS treatment (Supplementary Fig. S2b), WA likely upregulates the mRNA levels of NRF2 target genes in GalN/LPS-treated Nrf2−/− mice as a result of its hepatoprotective effect.
NRF2 was not required for the hepatoprotective effect of WA
WA is known to induce NRF2 signaling and NRF2-dependently alleviates acetaminophen-induced liver injury43,44. 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 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
WA induced hepatic autophagy, while 3-MA failed to abolish the hepatoprotective effect of WA. Western blot analysis for ATG3, LC3I, and LC3II (= 3).Quantitation analysis of relative protein levels (= 3).Experimental time scheme.Effect of 3-MA in the expression of ATG3, LC3I, and LC3II.Serum ALT and AST levels in control saline-pretreated mice treated with or without WA.Serum ALT levels and AST levels in 3-MA-pretreated mice treated with or without WA.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 (= 5 unless otherwise indicated). One-way ANOVA or unpaired two-tailedtest was used for statistical analyses.< 0.01,< 0.005 versus Control; *< 0.05, **< 0.01 and ***< 0.005 versus VGalN/LPS. a b c d e f g n n n t p p p p p ## ### +
Autophagy, hepatocyte AMPKα, or IB were not required for the protective effect of WA κκ
3-MA, as a widely-used autophagy inhibitor51, 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 activation50, while hepatic AMPKα1 activation protects against GalN/LPS-induced liver injury22,39. 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). Ampka1ΔHep mice were next employed and analyses of serum ALT levels revealed that the hepatoprotective effect of WA in Ampka1ΔHep mice was as potent as in the Ampka1fl/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 Ampka1ΔHep mice (Supplementary Fig. S4d, e). While WA directly targets IκκB to produce its anti-inflammatory effects46,47, WA still showed a potent effect in alleviating GalN/LPS-induced liver injury in IkkbΔHep 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
Hepatoprotective effect of WA was abolished in macrophage-depleted mice. Experimental time scheme.Serum ALT and AST levels in control liposomes-pretreated mice.Serum ALT and AST levels clodronate liposomes-pretreated mice.HepaticmRNA levels of control or clodronate alone-treated mice.Hepatic mRNA levels of proinflammatory cytokines including,,, andin control liposomes-pretreated mice.Hepatic mRNA levels of inflammation pathway including,,, andin clodronate liposomes-pretreated mice.Representative images of liver H&E staining, scale bar 50 µm. Data are presented as means ± SD,= 5. Groups were the same as described in Fig.legend. One-way ANOVA was used for statistical analyses.< 0.05,< 0.01, and< 0.005 versus Control group; ***< 0.005 versus V + GalN/LPS group. a b c d e f g F4/80 Tnfa Il6 Il1b Ccl2 Tnfa Il6 Il1b Ccl2 n p p p p 1 # ## ###
WA prevented GalN/LPS-induced FH partially by inhibiting activation of the NLRP3 inflammasome
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 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 NLRP3-independent 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 attenuated the hepatic inflammasome activation in WT mice, while the hepatoprotective effect of WA was partially compromised inmice. Nlrp3 −/− Representative images of western blot analyses for CASP1, ASC and IL-1β in WT livers (= 3).Quantitation for CASP1, ASC and IL-1β (= 3).Serum ALT and AST levels in GalN/LPS-treated WT ormice treated with or without WA.Representative hepatic H&E staining of V or WA-treated WT ormice subjected to GalN/LPS challenge, scale bar 50 µm. Effect of WA in mRNA levels of(),(),(), and(), in LPS-treated WT andmacrophage. One-way ANOVA or two-tailed t test was used for statistical analyses. Data were presented as mean ± SD (= 5 unless otherwise indicated). For the animal study,< 0.05 versus Control; *< 0.05 or ***< 0.005 versus V+GalN/LPS. For the macrophage culture study,< 0.05,< 0.01, or< 0.005 versus WT Control orControl, respectively; *< 0.05, **< 0.01, or ***< 0.005 versus WT LPS orLPS, respectively.< 0.005 represents comparison between WT LPS andLPS group. a b c d e f g h n n Nlrp3 Nlrp3 Tnfa Il6 Il1b Ccl2 Nlrp3 n p p p p p p Nlrp3 p p p Nlrp3 p Nlrp3 −/− −/− −/− # # ## ### −/− −/− &&& −/−
WA therapeutically improved GalN/LPS-induced hepatotoxicity
WA therapeutically alleviated GalN/LPS-induced liver injury. Experimental time scheme.Serum ALT and AST levels in GalN/LPS-treated mice followed by vehicle (V) or WA treatment.Representative hepatic H&E staining of V or WA-treated WT mice subjected to GalN/LPS challenge, scale bar 50 µm.graphic abstract for major findings and proposed mechanism. Unpaired two-tailed Student’stest was used for statistical analyses. GalN/LPS + V, mice treated with GalN/LPS followed by vehicle treatment; GalN/LPS + WA, mice treated with GalN/LPS followed by WA treatment. Data were presented as mean ± SD (= 5). **< 0.01 versus GalN/LPS + V group. a b c d t n p
Discussion
FH, also known as acute liver failure, could lead to lethality with liver transplantation as the only curable treatment1,2. Thus, pharmacotherapeutic discoveries are needed to treat this disease. WA was reported as a hepatoprotective agent in acetaminophen-induced hepatotoxicity mainly due to hepatocyte necrosis43,44, and diet-induced nonalcoholic steatohepatitis45 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, IkkbΔHep mice, Ampka1ΔHep 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 activity46,47. While the possibility that hepatocyte IκκB plays a role in the hepatoprotective effect of WA on GalN/LPS-induced FH was excluded by using IkkbΔHep 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 acetaminophen-induced liver injury dependent on NRF2 induction44. 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 necrosis56 and GalN/LPS mainly initiates hepatocyte apoptosis15. More importantly, the roles of NRF2 deficiency in these two models are different. Nrf2−/− mice were protected from GalN/LPS-induced FH31, but showed an enhanced hepatotoxicity in acetaminophen-treated mice57. Consistent with these findings, another herbal NRF2 activator, licochalcone A, also showed a potent hepatoprotective effect in GalN/LPS-treated Nrf2−/− mice31. This same study demonstrated that the autophagy inhibitor 3-MA did not abolish the hepatoprotective effect of licochalcone A31. 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 injury23,33,34,36,37,58–60 and that pharmacotherapies alleviate GalN/LPS-induced liver injury via inhibiting NLRP3 inflammasome activation30,61. 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 inflammation62. NLRP3 is physiologically beneficial in maintaining metabolic and immune homeostasis but could be detrimental upon overactivation63,64. 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 strain-matched WT mice, as well as between macrophage-depleted 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 steatohepatitis65, and NLRP3 inflammasome overactivation is required for fibrosis development and thus antagonism of NLRP3 inflammasome activation reduces nonalcoholic fatty liver disease66,67.
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 disease36 and support the potential of WA to be repurposed as a hepatoprotective agent against the clinical FH, particularly featured with TNF-α-related apoptosis.
Methods and Materials
Reagents
WA was purchased from ChromaDex (Irvine, CA, USA). ACTD, GalN, LPS, 3-methyladenine (3-MA), dimethyl sulfoxide, N-benzyloxycarbonyl-Val-Ala-Asp(O-Me) fluoromethyl ketone (ZVAD) and other reagents, if not otherwise indicated, were purchased from Sigma Aldrich (St. Louis, MO, USA). Recombinant mouse TNF-α was from Peprotech (Rocky Hill, NJ, USA). Cell-counting kit 8 was from Dojindo Molecular Technologies Inc (Rockville, MD, USA). ALT and AST kits were from Catachem (Oxford, CT, USA). Clodronate liposomes (neutral) were obtained from FormuMax (Sunnyvale, CA, USA). Antibodies against CASP3, cleaved PARP1, LC3I/II, ATG3, ASC, IL-1β, AMPKα, p-AMPKα, NRF2 and β-actin (ACTB) were from Cell Signaling Technology (Danvers, MA, USA). CASP1 antibody was from Santa Cruz Biotechnology (Dallas, Texas, USA). Details for the antibodies are described in Supplementary Table. S1
Animal Experiments
Wild-type (WT) mice, Nrf2−/−, Nlrp3−/−, or Ampka1fl/fl mice on a C57BL/6J background were purchased from the Jackson Laboratory, while IkkbΔHep mice on the C57BL/6J background were described previously68,69. 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 previously69 were bred with Ampka1fl/fl mice to generate Ampka1ΔHep 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 previously70, 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 modifications54. 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.