Fisetin inhibits lipopolysaccharide-induced inflammatory response by activating β-catenin, leading to a decrease in endotoxic shock

Fisetin is a naturally occurring flavonoid that possesses several pharmacological benefits including anti-inflammatory activity. However, its precise anti-inflammatory mechanism is not clear. In the present study, we found that fisetin significantly inhibited the expression of proinflammatory mediators, such as nitric oxide (NO) and prostaglandin E2 (PGE2), and cytokines, such as interleukin-6 (IL-6) and tumor necrosis factor-alpha (TNF-α), in lipopolysaccharide (LPS)-stimulated RAW 264.7 macrophages. Additionally, fisetin attenuated LPS-induced mortality and abnormalities in zebrafish larvae and normalized the heart rate. Fisetin decreased the recruitment of macrophages and neutrophils to the LPS-microinjected inflammatory site in zebrafish larvae, concomitant with a significant downregulation of proinflammatory genes, such as inducible NO synthase (iNOS), cyclooxygenase-2a (COX-2a), IL-6, and TNF-α. Fisetin inhibited the nuclear localization of nuclear factor-kappa B (NF-κB), which reduced the expression of pro-inflammatory genes. Further, fisetin inactivated glycogen synthase kinase 3β (GSK-3β) via phosphorylation at Ser9, and inhibited the degradation of β-catenin, which consequently promoted the localization of β-catenin into the nucleus. The pharmacological inhibition of β-catenin with FH535 reversed the fisetin-induced anti-inflammatory activity and restored NF-κB activity, which indicated that fisetin-mediated activation of β-catenin results in the inhibition of LPS-induced NF-κB activity. In LPS-microinjected zebrafish larvae, FH535 promoted the migration of macrophages to the yolk sac and decreased resident neutrophil counts in the posterior blood island and induced high expression of iNOS and COX-2a, which was accompanied by the inhibition of fisetin-induced anti-inflammatory activity. Altogether, the current study confirmed that the dietary flavonoid, fisetin, inhibited LPS-induced inflammation and endotoxic shock through crosstalk between GSK-3β/β-catenin and the NF-κB signaling pathways.


Results
High concentrations of fisetin decrease the viability of RAW 264.7 macrophages. To investigate the effect of fisetin on the viability of RAW 264.7 macrophages, the cells were treated with the indicated concentrations of fisetin for 24 h in the presence or absence of LPS. No significant change in cell viability was observed at concentrations of up to 8 µM fisetin compared to that of the untreated cells (97.8 ± 1.6%, 97.3 ± 1.7%, 97.5 ± 1.5%, 96.2 ± 1.9% at 1, 2, 4, and 8 µM fisetin); however, higher concentrations of fisetin significantly decreased the viability of RAW 264.7 macrophages (89.5 ± 1.3% and 78.8 ± 5.7% at 10 and 20 µM, respectively) ( Fig. 1A). Additionally, LPS caused a decrease in cell viability (75.7 ± 2.4%) by potently inducing the differentiation of RAW 264.7 macrophages. Additionally, 10 µM and 20 µM fisetin further reduced cell viability in the presence of LPS (66.0 ± 0.4% and 55.2 ± 0.5%, respectively). However, cytotoxic hallmarks such as apoptotic bodies, floating cells, and cell debris were not visible at all concentrations of fisetin tested in this study under microscopy ( Fig. 1B). To further confirm the fisetin-induced decrease in cell viability, the total percentages of viable cells and dead cells were measured using flow cytometry (Fig. 1C,D). These results aligned with those of the MTT assay, the percentage of total viable cells was 78.2 ± 1.5%, while that of dead cells was 21.4 ± 1.2%, at 20 µM fisetin (Fig. 1C), which are comparable to those of the H 2 O 2 -treated cells (positive control; H 2 O 2 induces cell death). Furthermore, we found that LPS alone decreased the percentage of viable cell populations (75.0 ± 0.4%) and increased dead cell populations (25.0 ± 0.4%, Fig. 1D). Both 10 μM and 20 μM fisetin decreased viability to 62.28 ± 0.77% and 55.05 ± 0.98%, and increased dead cell populations up to 37.71 ± 0.77% and 44.95 ± 0.97%. Collectively, these results indicate that high concentrations of fisetin exhibit cytotoxicity in RAW 264.7 macrophages; however, cytotoxicity was not observed below a concentration of 8 µM.

Fisetin inhibits LPS-induced proinflammatory gene expression and concomitantly decreases macrophage and neutrophil recruitment to the inflammatory sites in zebrafish larvae.
To confirm whether fisetin inhibits the LPS-induced inflammatory response, 3 dpf zebrafish larvae were microinjected with LPS for 24 h. Thereafter, the time-dependent expression of proinflammatory mediators such as iNOS and COX-2a and cytokines such as IL-6 and TNF-α was evaluated. In the LPS-microinjected condition, all genes tested in this study were expressed and reached maximal levels at 18 h, with a slightly different expression patterns (Fig. 4A). iNOS and TNF-α were significantly expressed at 6 h and their expression lasted for 24 h, whereas COX-2a and IL-6 were highly expressed at 18 h. To evaluate the anti-inflammatory effect of fisetin in vivo, LPSmicroinjected zebrafish larvae were grown in the presence of the indicated concentrations of fisetin for 18 h, and the expression level of the proinflammatory genes was detected. RT-PCR showed that fisetin concentrationdependently decreased the expression of iNOS, COX-2a, IL-6, and TNF-α in LPS-microinjected zebrafish larvae (Fig. 4B). In particular, a concentration of 400 µM fisetin was the potent at downregulating the expression of all proinflammatory genes tested in this study (i.e., the levels reached those in the untreated larvae). Furthermore, we sought to determine whether fisetin prevents the recruitment of macrophages and neutrophils to the inflammatory site in LPS-microinjected zebrafish larvae. Neutral red staining revealed that LPS injection significantly increased the macrophage counts at the site where LPS was injected (inflammatory site) in the yolk sac (red dot in the red box) at 24 h; however, immersion in fisetin resulted in a gradual decrease in the accumulation of macrophages in the yolk sac (Fig. 4C), indicating that fisetin inhibits the recruitment of macrophages from the circulating blood to the yolk sac, leading to the generation of anti-inflammatory responses. In alignment with the inhibition of macrophage recruitment, LPS-microinjection significantly decreased the large and clear cytolymph lipid droplets (accumulation of neutrophils, yellow dot box) in the posterior blood island (PBI) as neutrophils infiltrated the inflammatory site, i.e., the yolk sac (Fig. 4D). We also found that fisetin impaired the migration of neutrophils to the inflammatory site in a concentration-dependent manner, which indicates that Figure 2. Fisetin decreases LPS-induced inflammatory mediator and cytokine levels in RAW 264.7 macrophages. RAW 264.7 macrophages (1 × 10 5 cells/mL) were treated with fisetin (0-8 µM) 2 h before treatment with 500 ng/mL LPS. (A) Total mRNA was isolated at 6 h after 500 ng/mL LPS treatment, and RT-PCR was performed. GAPDH was used as an internal control. (B) Total proteins were isolated at 24 h and western blotting was performed. β-Actin was used as an internal control. (C) The amount of NO production in the culture medium was determined using the Griess Reagent Assay. (D) The amount of PGE 2 was determined at 24 h using an ELISA according to the manufacturer's instructions. (E) Total mRNA was isolated at 6 h and subjected to RT-PCR for IL-6 and TNF-α. The amount of (F) IL-6 and (G) TNF-α was measured at 24 h by an ELISA. Each value indicates the mean ± SEM from three independent experiments. Significant differences among the groups were determined using the Student's t-test ( ### p < 0.001 vs. untreated cells) and one-way ANOVA with Bonferroni correction (***p < 0.001, **p < 0.05, and *p < 0.01 vs. LPS-treated cells).  (D) Heart rates were measured to assess toxicity. Each value indicates the mean ± standard error median (SEM), and is representative of the results obtained from 20 fish for each group. Significant differences among the groups were determined using Student's t-test ( # p < 0.01 vs. untreated zebrafish larvae) and one-way ANOVA with Bonferroni correction (*p < 0.01 vs. LPS-treated zebrafish larvae). www.nature.com/scientificreports/ fisetin attenuates the recruitment of neutrophils to the LPS-microinjected site. These results indicate that fisetin inhibits the LPS-induced inflammatory response by suppressing the expression of proinflammatory genes and reducing macrophage and neutrophil recruitment to the inflammatory sites.

Fisetin inhibits LPS-induced NF-κB activity in RAW 264.7 macrophages.
As NF-κB is considered as a major transcription factor in LPS-induced inflammatory response, we investigated whether fisetin negatively regulates the activation of the NF-κB pathway. Western blotting using nuclear extracts from RAW 264.7 macrophages confirmed that fisetin decreased LPS-induced nuclear localization of NF-κB p50 and p65 in a concentration-dependent manner (Fig. 5A). Additionally, immunohistochemistry confirmed that LPS rapidly stimulated the translocation of NF-κB p65 into the nucleus; however, fisetin inhibited the nuclear translocation of NF-κB p65 in the presence of LPS (Fig. 5B), indicating that fisetin inhibits LPS-induced NF-κB activation, resulting in the inhibition of inflammatory responses.
Fisetin enhances phosphorylation of GSK-3β at Ser9 and subsequent activation of β-catenin in RAW 264.7 macrophages. Recently, our research team revealed that fisetin binds to GSK-3β at non-ATPbinding site-through molecular docking prediction-and consequently activates β-catenin in B16F10 melanoma cells 20 . Deng et al. reported that β-catenin negatively regulates the inflammatory responses by inhibiting the expression of proinflammatory mediators and cytokines 14 . These data indicate that fisetin inhibits GSK-3β and subsequently stabilizes β-catenin, which attenuates LPS-induced inflammation. Therefore, we determined whether fisetin affects the expression of GSK-3β and β-catenin as well as the nuclear translocation of β-catenin. Fisetin concentration-dependently increased the phosphorylation of GSK-3β at Ser9-an inactive form-and enhanced the level of total β-catenin (Fig. 6A) and its nuclear translocation (Fig. 6B). Immunohistochemistry also revealed that fisetin enhanced the nuclear translocation of β-catenin regardless of the presence of LPS; however, an abundance of β-catenin was found in the cytosol after LPS treatment (Fig. 6C). Altogether, these results indicate that fisetin blocks the degradation of β-catenin by inhibiting GSK-3β, and consequently enhances the nuclear translocation of β-catenin.

Fisetin-induced anti-inflammatory response is related to activation of β-catenin in an endotoxic shock model of zebrafish larvae.
To confirm the significance of the β-catenin signaling pathway in fisetin-induced anti-inflammatory responses, we pharmacologically blocked the canonical β-catenin signaling pathway with FH535 in zebrafish larvae. We found that FH535 at 10 µM and 20 µM induced 20% and 100% mortality after 24-h of treatment. Further, 10 µM of FH535 increased the mortality up to 40% in LPSmicroinjected larvae (data not shown). Interestingly, FH535-induced mortality was completely blocked in the presence of fisetin (data not shown), indicating that fisetin could reduce the mortality by activating the β-catenin signaling pathway. Thus, we investigated whether, in the presence of FH535, fisetin influences the recruitment of macrophages and neutrophils in LPS-microinjected zebrafish larvae. Interestingly, we found that FH535 by itself significantly increased the neutral red intensity (macrophages) in the yolk sac (Fig. 8A) and increased sudan black-stained spots (neutrophils) in the whole body, with neutrophils being retained at the PBI (Fig. 8B). As expected, in FH535-treated conditions, fisetin inhibited the recruitment of macrophages and neutrophils to the inflammatory site in LPS-microinjected zebrafish larvae, which indicated that fisetin-induced β-catenin activation hindered macrophage and neutrophil recruitment to the inflammatory sites. Subsequently, we investigated the expression of iNOS and COX-2a under the same experimental conditions. We found that both inflammatory genes were highly expressed in the presence of FH535 or LPS. Additionally, fisetin significantly inhibited FH535or/and LPS-induced iNOS and COX-2a expression (Fig. 8C). These data imply that fisetin negatively regulates LPS-induced inflammation and endotoxic shock via activating the β-catenin signaling pathway.

Discussion
Endotoxic shock is a systemic inflammation accompanied by the excessive release of inflammatory mediators and cytokines, resulting in high cardiac output and mortality 22 . Non-steroidal anti-inflammatory drugs (NSAIDs), including aspirin, celecoxib, and diclofenac, are commonly used to combat systemic inflammation; however, they are associated with digestive problems such as upset stomach, heartburn, ulcers, and kidney injuries 23 . Therefore, small compounds, such as flavonoids, may provide a good alternative to prevent and reduce the risk of inflammation and endotoxic shock with relatively low side effects. Fisetin is a bioactive diphenyl propane flavone that is abundant in various plants. Its pharmacological properties including anti-cancer, antioxidant, and anti-inflammatory activity have been reported 16,[18][19][20] . Nevertheless, the molecular mechanisms underlying the anti-inflammatory properties of fisetin are unclear. In the current study, we evaluated the anti-inflammatory effect of fisetin on LPS-induced inflammation and endotoxic shock in RAW 264.7 macrophages and zebrafish larvae. Our findings suggest that fisetin inhibits LPS-induced inflammation and endotoxic shock by suppressing β-catenin-mediated NF-κB activity, which subsequently attenuates the expression of proinflammatory mediators, such as NO and PGE 2 , and cytokines, such as IL-6 and TNF-α (Fig. 9). Systemic inflammation is characterized by increase in proinflammatory cytokine levels, including IL-6, IL-12, and TNF-α, and other proinflammatory mediators, including NO and PGE 2 , via the activation of NF-κB [24][25][26][27][28] . Accumulating evidence on LPS-induced inflammatory disorders has revealed that LPS triggers the expression of inflammatory genes via the TLR4-dependent signaling pathway 3,5 . Binding of LPS to TLR4 leads to the phosphorylation of the IκB kinase (IKK) complex through the recruitment and activation of MyD88 and IRAK4, and in turn, phosphorylates IκB, which is degraded by the proteasome and allows the nuclear translocation of free NF-κB 6,7 . Therefore, targeting the NF-κB signaling pathway is thought to be a pivotal therapeutic strategy in the pathology of LPS-induced inflammatory disorders. Two research groups have previously reported that fisetin reduces ovalbumin-induced airway inflammation by inhibiting MyD88-mediated NF-κB activation 19,29 . In the present study, we reconfirmed that fisetin alleviated LPS-induced inflammatory mediator and cytokine levels in RAW 264.7 macrophages by inhibiting the NF-κB signaling pathway. Fisetin is also effective at protecting against metabolic dysfunction 30 , UVB 31 , cardiac ischemic injury 32 , and brain disorders 33 . Furthermore, fisetin inhibited LPS-induced macrophage activation and functional maturation of dendritic cells 34 as well as LPSinduced acute lung injury 35 by suppressing TLR4-mediated NF-κB signaling pathway. Nevertheless, whether the precious mechanism of fisetin-mediated endotoxic shock has not been elucidated. In the current study, using a zebrafish larva model, we proved that fisetin attenuated LPS-induced mortality and abnormality and caused a significant decrease in macrophage and neutrophil recruitment at the inflammatory site. Meanwhile, fisetin restored the heart rate up to the normal level along with the downregulation of proinflammatory genes during LPS stimulation, indicating that fisetin is a promising pharmacological candidate against LPS-induced endotoxic shock. Nevertheless, whether fisetin protects against endotoxic shock in preclinical studies is yet to be elucidated.
As the canonical Wnt/β-catenin signaling pathway plays a crucial role in the development of cancer through chronic inflammation, non-steroidal anti-inflammatory drugs, such as PPARγ agonists, might serve as potent therapeutic agents to treat cancers as they inhibit the Wnt/β-catenin pathway 36,37 . Alternatively, Swafford et al. reported that the deletion of Wnt receptors reduced histopathological severity and inflammation in the colon, along with high expression of anti-inflammatory cytokines such as IL-10, through the conditional activation of β-catenin 38 . Such findings suggest that the canonical Wnt/β-catenin pathway negatively regulates inflammation. Despite the dual regulation of Wnt/β-catenin in inflammation, in this study, fisetin increased total β-catenin and promoted its nuclear translocation in RAW 264.7 macrophages. Further, an increased expression of inactive GSK-3β (phosphorylated at Ser9) was observed. The inhibition of the Wnt/β-catenin signaling pathway by inhibitors of both Wnt/β-catenin and PPAR exacerbated the mortality and abnormality in zebrafish larvae. This was accompanied by a further accumulation of macrophages and neutrophils at the LPS-microinjected inflammatory site and a significant expression of proinflammatory genes. Interestingly, Deng et al. reported that β-catenin physically interacts with NF-κB to form a complex, and thereby reduces the DNA-binding ability of NF-κB 14 . Jun et al. also reported that gram-negative Salmonella infection constitutively activated β-catenin and thereby stabilized IκBα, which subsequently repressed the activity of NF-κB in HCT116 colon cancer cells 39 , indicating that the crosstalk between the Wnt/β-catenin and NF-κB pathways is linked to the regulation of inflammation. In agreement with the above studies, we found that fisetin-induced β-catenin attenuated LPS-induced inflammation and endotoxic shock by inhibiting NF-κB activation. Nevertheless, the discrepancy between the positive and negative regulatory role of the Wnt/β-catenin pathway in inflammatory disorders may be investigated in the context of NF-κB, even though NF-κB is a key activator of inflammation. . Fisetin inhibits LPS-induced inflammatory response in zebrafish larvae. Zebrafish larvae at 1 day post fertilization (dpf) were cultured in 0.003% PTU containing E3 embryo media. Briefly, 2 nL of 0.5 mg/mL LPS was microinjected into the yolk at 3 dpf. Zebrafish larvae were immediately immersed in E3 embryo media containing different concentrations of fisetin. (A) In LPS-microinjected conditions, 10 zebrafish were euthanized at the indicated time points and subjected to RT-PCR for evaluating the expression of iNOS, COX-2a, IL-6, and TNF-α. (B) At 18 h post injection (hpi), 20 zebrafish larvae from each treatment were euthanized and the expression of iNOS, COX-2a, IL-6, and TNF-α was measured by RT-PCR. (C) Neutral red staining of macrophages and (D) sudan black staining of the neutrophils were performed at 24 hpi. Each value indicates the mean ± standard error median (SEM) and is representative of the results obtained from three independent experiments. Significant differences among the groups were determined using the Student's t-test ( ### p < 0.001, ## p < 0.01, and # p < 0.05 vs. untreated zebrafish larvae) and one-way ANOVA with Bonferroni correction (***p < 0.001, **p < 0.01, and *p < 0.05 vs. LPS-treated zebrafish larvae). www.nature.com/scientificreports/ Our previous research predicted that fisetin directly binds to GSK-3β at the noncompetitive ATP-binding site and promotes β-catenin stabilization and nuclear translocation 20 . GSK-3β is a ubiquitous serine/threonine kinase involved in the molecular pathogenesis of severe disorders in humans, including inflammation, tumorigenesis, and neurological disorders 40 . In particular, Medunjanin et al. identified that GSK-3β directly phosphorylates NEMO, which is an essential activator of NF-κB and consequently activates NF-κB, indicating that GSK-3β stimulates a non-canonical NF-κB signaling pathway 41 . Ougolkov et al. also revealed that the inhibition or genetic depletion of GSK-3β inhibits NF-κB-induced gene transcription and subsequently leads to pancreatic cancer cell proliferation and survival by activating the NF-κB signaling pathway 42 . Based on our data, we deduced that the targeting of GSK-3β by fisetin inactivates the non-canonical NF-κB pathway and stabilizes β-catenin to inhibit LPS-induced inflammation and endotoxic shock. However, a recent study proposed the dual effect of GSK-3β on the anti-inflammatory and inflammatory response depending on the virulence factors, cell types, and physiological state of cells 43 . Therefore, the significance of the crosstalk between the GSK-3β/β-catenin and NF-κB signaling pathways needs to be persistently investigated in the pathogenesis of inflammatory disorders such as septic shock.
In conclusion, our findings suggest that fisetin attenuates LPS-induced inflammation and endotoxic shock by suppressing the β-catenin-mediated NF-κB signaling pathway. Fisetin can thus be considered as a potential anti-inflammatory drug for systemic inflammation.   , and is representative of the results obtained from three independent experiments. Significant differences among the groups were determined using one-way ANOVA with Bonferroni correction (***p < 0.001, **p < 0.005, and *p < 0.01). Western blot analysis 44 . Total cellular protein extracts were prepared by RIPA Lysis Buffer (iNtRON Biotechnology). The total protein lysates were centrifuged at 16,000g at 4 °C for 20 min. In a parallel experiment, was detected by RT-PCR. GAPDH was used as an internal control. Each value indicates the mean ± standard error median (SEM) and is representative of the results obtained from three independent experiments. Significant differences among the groups were determined using one-way ANOVA with Bonferroni correction (***p < 0.001, **p < 0.005, and *p < 0.01). Maintenance of zebrafish embryo and larvae. AB strain zebrafish were handled as previously described 49 . The zebrafish study was approved by Animal Care and Use Committee of Jeju National University (Jeju Special Self-governing Province, Republic of Korea; approval No.: 2020-0013). All methods were carried out in accordance with relevant guidelines and regulations. Additionally, all the methods were carried out in accordance with the ARRIVE guidelines 50 . Zebrafish were raised at 28.5℃ with a 14:10-h light:dark cycle in a water-recirculating tank system (pH 7.4 and 0.03% salinity). Fertilized embryos were collected after natural spawning and cultured at 28.5 °C in E3 embryo media containing 2 mg/L methylene blue. To inhibit melanin formation, 0.003% PTU was added to the egg water throughout the experimental period.

Figure 9.
Fisetin inhibits GSK-3β-mediated NF-κB activation in the presence of β-catenin, leading to the inhibition of inflammation-induced septic shock. Once macrophages are exposed to high concentrations of the bacterial endotoxin, LPS, they initiate an inflammatory response and endotoxic shock by upregulating the expression of NF-κB-induced inflammatory genes, such as iNOS, COX-2, IL-12, IL-6, and TNF-α. Fisetin binds to the noncompetitive ATP-binding sites of GSK-3β and phosphorylates GSK-3β at Ser9, resulting in the inactivation GSK-3β and release of β-catenin from the destruction complex. The released β-catenin inhibits the transcriptional activity of NF-κB, thereby alleviating LPS-induced inflammation and endotoxic shock. www.nature.com/scientificreports/ LPS microinjection and cardiac toxicity evaluation. Three dpf zebrafish larvae were anesthetized using 0.04% tricaine and LPS (0.5 mg/mL, 2 nL in each larva) was microinjected into the yolk sac using Drummond NANOJECT III Injector (Drummond Scientific, Broomall, PA, USA). The negative control group was microinjected with PBS. After microinjection of LPS, the larvae were immediately placed in E3 media containing the indicated concentrations of fisetin. Dead larvae were removed within 0.5 hpi. Each group of larvae (n = 20) was cultured at 28.5 °C and observed for signs of phenotypic abnormality and mortality. The heart rate of the larvae was manually counted for one minute and used as an indicator for the cardiac toxicity evaluations. All mentioned parameters were observed using Olympus SZ2-ILST Stereomicroscopy (Tokyo, Japan).
Neutral red staining 51 . Neutral red is a vital dye that accumulates in the lysosomes through endocytosis.
As macrophage cells undergo efficient endocytosis, neutral red more robustly labels macrophages than any other cell types. Optimal staining of macrophages in live embryos was achieved by incubating embryos in 2.5 μg/mL neutral read solution containing 0.003% PTU at 28.5℃ in the dark for 6-8 h. After staining, macrophage migration was observed using Olympus SZ2-ILST Stereomicroscopy.
Sudan black staining 51 . Sudan black is an azo stain that detects the presence of lipids with dark stains representing neutrophils. A stock solution of sudan black was prepared from sudan black powder (0.6 g) dissolved in pure ethanol (200 mL). A buffer solution was made from phenol (16 g) dissolved in pure ethanol (30 mL) plus Na 2 HPO 4 ·12H 2 O (0.3 g) dissolved in distilled water (100 mL). A working staining solution was made by mixing stock solution (30 mL) with buffer (20 mL). Whole larvae were fixed with 4% methanol-free PFA in PBS for 2 h at room temperature and rinsed in PBS. The larvae were incubated in sudan black solution for 40 min, washed extensively in 70% ethanol, and then progressively rehydrated with PBS plus 0.1% Tween-20. The stained neutrophils were observed using Olympus SZ2-ILST stereomicroscopy.