Micheliolide inhibits LPS-induced inflammatory response and protects mice from LPS challenge

Sepsis is the principal cause of fatality in the intensive care units worldwide. It involves uncontrolled inflammatory response resulting in multi-organ failure and even death. Micheliolide (MCL), a sesquiterpene lactone, was reported to inhibit dextran sodium sulphate (DSS)-induced inflammatory intestinal disease, colitis-associated cancer and rheumatic arthritis. Nevertheless, the role of MCL in microbial infection and sepsis is unclear. We demonstrated that MCL decreased lipopolysaccharide (LPS, the main cell wall component of Gram-negative bacteria)-mediated production of cytokines (IL-6, TNF-α, MCP-1, etc) in Raw264.7 cells, primary macrophages, dendritic cells and human monocytes. MCL plays an anti-inflammatory role by inhibiting LPS-induced activation of NF-κB and PI3K/Akt/p70S6K pathways. It has negligible impact on the activation of mitogen-activated protein kinase (MAPK) pathways. In the acute peritonitis mouse model, MCL reduced the secretion of IL-6, TNF-α, IL-1β, MCP-1, IFN-β and IL-10 in sera, and ameliorated lung and liver damage. MCL down-regulated the high mortality rate caused by lethal LPS challenge. Collectively, our data illustrated that MCL enabled maintenance of immune equilibrium may represent a potentially new anti-inflammatory and immunosuppressive drug candidate in the treatment of sepsis and septic shock.

Natural products play a significant role not only in the design, synthesis and discovery of new drugs, but also as the most promising source of bioactive substances and innovative drugs 14 . Sesquiterpene lactones (SLs) are a class of naturally occurring plant terpenoids belonging to Asteraceae family, many of which exhibit diverse biological activities (e.g. antimalarial, anticancer, antiviral, antibacterial, antifungal and anti-inflammatory) [15][16][17] . Micheliolide (MCL) is a sesquiterpene lactone (Fig. 1a), isolated from Michelia compressa (Magnoliaceae) 18 .
In the current study, we demonstrated that the natural product MCL decreased lipopolysaccharide (LPS)-mediated production of IL-6, TNF-α , monocyte chemotactic protein 1 (MCP-1), interferon β (IFN-β ) and IL-10 in Raw264.7, primary peritoneal macrophages, dendritic cells, human monocytic cell THP-1 and human CD14 + monocytes. MCL was found to attenuate NF-κ B and PI3K/Akt/p70S6K activation, especially inhibiting the phosphorylation of Iκ Bα (Ser32/36), Akt (Ser473) and p70S6K (Thr389). The anti-inflammatory effect of MCL was similarly observed in vivo. MCL also protects mice from LPS-induced organ damage and high mortality. Therefore, MCL maintains immune equilibrium by down-regulating proinflammatory cytokines, chemokines, and type I interferon secretion in TLR4 signaling, and thus attenuate host damage and reduce mortality. Furthermore, MCL is a drug candidate for the development of novel potential immunosuppressive and anti-inflammatory agents for the treatment of septic shock triggered by lethal LPS challenge.

Results
MCL does not affect the growth and apoptosis of Raw264.7. In order to determine the cytotoxicity of MCL, we first determined the apoptotic sensitivity of Raw264.7 to MCL. MCL at different concentrations (1 μM to 10 μM) was added with or without LPS (100 ng/mL) to the cell culture supernatant, and incubated for further 18 h. The FACS results showed that none of the examined concentrations of MCL (up to 10 μM) induced any apoptosis in resting or LPS-activated Raw264.7 at 18 h, as evidenced by Annexin V and 7-AAD labeling (Fig. 1b). Further, we evaluated whether MCL played an inhibitory role in the growth of Raw264.7 using the CCK8 assay.
The absorbance values showed that MCL (5 μM or 10 μM) did not reduce the cell viability of Raw264.7 within 3 days (Fig. 1c). No cytotoxic effects of MCL were observed up to 10 μM in rat mesangial cells using the MTT method 19 .

MCL inhibits LPS-triggered inflammatory responses in mouse macrophages and dendritic cells (DCs).
Macrophages and DCs are antigen-presenting cells and inflammatory cytokine-producing cells that play a crucial role in the regulation of inflammatory diseases. We investigated whether MCL regulated the inflammatory response of macrophages and dendritic cells to LPS. Different concentrations of MCL (1 μM to 10 μM) and LPS (100 ng/mL) were added simultaneously to the cell culture supernatant of mouse macrophage cell line Raw264.7. Cytokines in the supernatants were measured by ELISA, and IL-1β mRNA expression was examined by qRT-PCR. The results showed that MCL treatment inhibited LPS-induced IL-6, TNF-α , IL-1β , MCP-1, IFN-β and IL-10 production in Raw264.7 (Fig. 2). The anti-inflammatory role of MCL was similarly observed in mouse primary peritoneal macrophages (Fig. 3). Further, we examined whether MCL regulated the inflammatory response to LPS in bone marrow-derived DCs (BMDCs). We found that MCL also inhibited the secretion of IL-6, TNF-α and MCP-1 after LPS stimulation in BMDCs ( Supplementary Fig. S1). Thus, MCL down-regulated the LPS-induced inflammatory response by decreasing the production of proinflammatory cytokines, chemokines, interferons, and even anti-inflammatory cytokines.
MCL inhibits LPS-triggered proinflammatory cytokines, chemokines and type I interferon production in human monocytes. Human cells are reliable indicators of the pharmacological activity of drug candidates in vivo 7 . Therefore, we confirmed the pharmacological effect of MCL in human acute monocytic leukemia cell line THP-1 and human primary CD14 + monocytes derived from healthy donors. The results showed similar and significant inhibition of LPS-induced IL-6, TNF-α , MCP-1, IFN-β and even anti-inflammatory IL-10 production by MCL in THP-1 (Fig. 4a-e). A decreased expression of TNF-α was also observed in human CD14 + monocytes following treatment with MCL (Fig. 4f).
We found that MCL did not affect the activation of LPS-induced ERK, JNK and p38 MAPKs (Fig. 5a). MCL treatment inhibited LPS-induced phosphorylation of Iκ Bα (Ser32/36) in Raw264.7 (Fig. 5b). In order to elucidate the regulatory role of MCL in NF-κ B signaling pathway, Raw264.7 cells were cotransfected with NF-κ B luciferase reporter plasmid and pRL-TK-Renilla-luciferase plasmid in a reporter gene assay. As shown in Fig. 5c, MCL attenuated the activity of NF-κ B luciferase reporter gene. Therefore, MCL inhibits NF-κ B activation following LPS stimulation.
Phosphorylation of Akt at Ser473 represents PI3K/Akt pathway activation. Results showed that MCL treatment inhibited LPS-evoked phosphorylation of Akt (Ser473) (Fig. 6a) in Raw264.7. A similar regulatory role of MCL was observed in mouse primary peritoneal macrophages and human monocytic cell line THP-1 ( Supplementary Fig. S2). The isoforms of p70 S6 kinase (p70S6K) and p85S6K are mitogen-activated Ser/Thr protein kinases downstream of PI3K/Akt, and p70S6K phosphorylates the S6 protein of the 40S ribosomal subunit and controls the translation of 5' oligopyrimidine tract mRNAs 21 . Phosphorylation of Thr389 is correlated with p70S6K activity in vivo most directly 22 . As shown in Fig. 6a, MCL impaired LPS-induced phosphorylation of p70S6K at Thr389. LY294002 is a generally accepted inhibitor of PI3K. Raw264.7 was pretreated with LY294002 or MCL for 30 min, and then stimulated with LPS (100 ng/mL) for the indicated time periods. Compared with simultaneous MCL treatment, pretreatment with MCL further reduced the phosphorylation of Akt (Ser473) at a relatively earlier time point (Fig. 6b). Therefore, MCL inhibits TLR4-triggered activation of PI3K/Akt/ p70S6K signaling pathway. Accurately, PI3K/Akt, NF-κ B and p38 MAPK pathways attribute to the production of LPS-induced IL-10 in macrophages 13 . To explore the role of MCL on PI3K/Akt pathway, Raw264.7 cell line was transfected with Akt expressing plasmid (Myr-Akt-HA) or empty plasmid (mock). It was then stimulated with LPS and different concentrations of MCL. Phosphorylation of Akt (Ser473) was enhanced by Akt overexpression (Fig. 6c). Results showed that Akt overexpression induced the increase of IL-10 secretion at 6 h or 18 h, but MCL abrogated the enhanced IL-10 expression following Akt overexpression and phosphorylation (Fig. 6d).

MCL attenuates the secretion of serum cytokines in LPS-challenged mice.
We used the acute peritonitis mouse model following LPS (10 mg/kg) injection intraperitoneally. MCL suppressed LPS-induced production of proinflammatory cytokines, chemokines and IFN in macrophages and DCs. Therefore, we focused on MCL-mediacted attenuation of the secretion of these cytokines in vivo. The glucocorticoid dexamethasone (DXM) is regularly used to cure sepsis or septic shock clinically 23 . In the current study, DXM was selected as a positive control in LPS-challenged mouse models. We selected MCL doses of 20 mg/kg or 10 mg/kg (combined with DXM) in serum cytokine evaluation, which are almost equivalent to the doses of 10 μM and 5 μM in cellular experiments in vitro, respectively. Similar to the results in vitro, MCL-treated mice generated decreased proinflammatory cytokines IL-6, TNF-α , IL-1β , IFN-β , and anti-inflammatory IL-10 in the sera compared with the levels in mouse models ( Fig. 7a-c,e,f). However, MCL did not affect MCP-1 secretion at 20 mg/kg in vivo (Fig. 7d). Obviously, a clinical dose of DXM (7 mg/kg) decreased cytokine secretion dramatically. Further, the production of IL-6 and IFN-β was lower in the combined DXM and MCL treatment group than that in DXM group, showing the combinational role of MCL.
MCL protects against lung and liver tissue damage. In sepsis, the overwhelming production of proinflammatory cytokines and chemokines induces edema, vascular leakage, vasodilatation, multiple organ failure (acute lung and liver injury, etc.), shock and death 24 . In order to examine organ damage, mice were treated as shown in Fig. 7a-f. A histologic examination of lung and liver was conducted. H&E-stained lungs in the phosphate-buffered saline (PBS) treatment group exhibited no pathological changes (Fig. 7g). However, the LPS treatment group showed increased thickness of alveolar wall, and significant inflammatory cell infiltration. The pathological changes observed in the mice treated with MCL, DXM or their combination were weaker than in the LPS-treated mice, showing clearer alveolar wall and less inflammatory cell infiltration (Fig. 7g). Similar to the pathological changes in lung, only the livers in the LPS group showed significant necrosis. The livers in the treated groups (MCL, DXM, and the combination treatment group) showed less severe pathological changes. The present results showed that MCL as well as DXM attenuated the pathological changes of lung and liver in acute peritonitis, corresponding to serum cytokine secretion.
MCL enhanced the survival rate in mice following lethal septic shock. To determine the protective role of MCL against septic shock, a survival analysis was carried out. The protective role of different doses was tested in septic shock mouse model, including low, medium, and high doses of MCL, DXM, or their combination and a medium dose of MCL. PBS or MCL treatment alone failed to induce any mortality within 6 days (data not shown) and caused no significant adverse effects. At the end of observation period (6 days), 10% of model mice (LPS group mice) survived. Ninety percent or 70% of mice exposed to high or medium dose of MCL survived at the end of observation (compared with LPS group, p < 0.01). Although 50% of mice in the low-dose MCL treatment group survived, they did not differ significantly from mice in the LPS treated group (p = 0.051 > 0.05). In fact, no mortality was seen in the DXM treatment group or the combined treatment group (LPS + DXM + MCL group) in our experiments. Mice in the combined treatment group exhibited significant recovery with ameliorated clinical symptoms (such as, lethargy, hunched posture and piloerection) compared with those in the DXM group. The survival data revealed that MCL at a medium or high dose (10 mg/kg or 20 mg/kg) protected mice against septic shock following lethal LPS challenge.

Discussion
Sepsis is a life-threatening syndrome with a 13% increase in annual incidence worldwide. It is triggered by an overwhelming inflammation induced by microbial components or toxins during infection 1 . Lipopolysaccharide (LPS), the cell wall constituent of Gram-negative bacteria, is the most well-known sepsis-inducing factor and recognized by Toll-like receptor 4 (a member of pattern recognition receptors) 5 . Systemic inflammatory response syndrome (SIRS) in septic patients is characterized by an exacerbation of inflammation, with increased levels of pro-inflammatory cytokines (IL-6, TNF-α , IL-1β , etc.), as well as anti-inflammatory cytokines (IL-10, TGF-β , IL-1Ra) in the bloodstream 25 . Neutralization of bacterial LPS or inhibition of its recognition by host cell receptors is a promising therapeutic approach. Eritoran tetrasodium (E5564) is one of the TLR4-antagonists, which cleared phase 2 trial 26 but failed in phase 3 clinical trial 2 . Improving the design of clinical studies and the standard of patient enrollment may increase its treatment efficacy under realistic conditions 27 . Anti-microbial peptide (e.g., Pep19-2.5) 6,7 and lipopolyamine (e.g., DS96) 8 are promising drug candidates acting at the receptor level. Another promising strategy includes inhibition of the signal transduction downstream of TLR4 and down-regulation of endotoxin-induced cytokine storm 5 . Ephedrine hydrochloride (EH) is a promising drug candidate in inflammatory diseases. It balances the production of pro-inflammatory and anti-inflammatory cytokines in TLR4 or TLR2 signaling 28,29 . In this study, MCL belongs to the second type of anti-septic drug candidates. MCL was recently reported to inhibit DSS-induced inflammatory intestinal diseases, colitis-associated cancer 30 and rheumatic arthritis 31 . In addition, the selective toxicity of MCL against acute myelogenous leukemia (AML) stem cells rather than CD34 + mesenchymal stem cells was also confirmed 32 . However, there was no report investigating the role of MCL in microbial infection or sepsis.
In this study, we provided systemic and substantial evidence showing that MCL inhibits LPS-induced production of inflammatory cytokines IL-6, TNF-α , MCP-1, IFN-β , and even anti-inflammatory cytokine IL-10 in Raw264.7, primary peritoneal macrophages and bone marrow-derived dendritic cells. Fink and Warren 7 proposed that additional emphasis should be placed on the design of cell-based assays for the development of new anti-septic pharmacological agents. Human cells, rather than mouse cells, are probably more reliable in estimating the potential benefit or harm of drug candidates in patients. In our experiments, we proved a similar and significant anti-inflammatory role of MCL in human monocytic cell line THP-1 and human primary CD14 + monocytes. Both of IL-6 and TNF-α are the primary mediators of sepsis 33 . A positive correlation has been found between serum IL-6 and TNF-α level and multiple organ failure 34,35 . Further, both TNF-α and IL-1β induce the production of additional inflammatory mediators from target cells 36 . Thus, we monitored the secretion of these cytokines and the histologic changes in LPS-induced peritonitis mouse model to confirm the protective role of MCL in vivo. The decrease in serum cytokine secretion and the corresponding amelioration in lung and liver pathology were observed following MCL treatment. Most importantly, a high or medium dose of MCL significantly decreased the mortality in mice with acute peritonitis.
The role of anti-inflammatory IL-10 in acute peritonitis or experimental sepsis is controversial. IL-10 balances pro-inflammatory cytokines (IL-6, TNF-α , etc.). An extremely high level of IL-10 may harm the host. Statistical analysis from 165 adult patients with severe abdominal sepsis showed that the serum TNF-α levels were 4.7-fold and IL-10 levels were 3.3-fold higher in the blood samples of non-survivors compared with those of the survivors. IL-10 levels were higher in the early stages (within 24 h) of patients who died 37 . As Kalechman reported, AS101, a nontoxic immunomodulator, increased the proportion of mice surviving cecal ligation and puncture (CLP) following inhibition of IL-10, as co-treatment with murine recombinant IL-10 abolished the protective activity of AS101 38 . In this study, MCL decreased the secretion of both pro-and anti-inflammatory (IL-10) cytokines to extinguish the excessive inflammation, and thus protected mice from septic shock.
After TLR4 ligation, MyD88 and TRIF were recruited to phosphorylate IRAKs. The TRAF6 interacts with a pre-assembled kinase complex containing TAK1 and TAB1/2/3, resulting in the activation of NF-κ B and MAPKs, culminating in the expression of pro-inflammatory cytokines and chemokines 39 . Screening for the activation of different signaling pathways demonstrated that the phosphorylation of Iκ Bα (Ser32/36) (Fig. 5b) was down-regulated by MCL treatment. The inhibitory role of MCL on NF-κ B pathway in Raw264.7 was also observed by Viennois and his colleagues 30 . In the human THP-1 monocytes, phosphorylation of Iκ Bα was also decreased by MCL (Fig. 5c). The down-regulation of MCL on NF-κ B activation was confirmed by NF-κ B luciferase reporter gene assay (Fig. 5d), which accounted for the down-regulation of the LPS-induced pro-inflammatory cytokines and chemokines.
In addition to MAPKs and NF-κ B, PI3K/Akt signaling also contributed to the signal transduction of TLR4, regulating the expression of inflammatory and anti-inflammatory cytokines, e.g., IL-10 11 . MCL treatment also  inhibited the phosphorylation of Akt (Ser473) and the downstream p70S6K (Thr389) significantly in Raw264.7 (Fig. 6a). Similar down-regulation of PI3K/Akt pathway was also observed in mouse primary macrophages and THP-1. MCL pretreatment further decreased the phosphorylation of Akt (Ser473) in Raw264.7. IL-10 expression was reduced by inhibition of PI3K/Akt activation in LPS-activated macrophages, DCs and human monocytes 13 . Therefore, enhanced Akt phosphorylation accounted for IL-10 production. In our trial, the elevated IL-10 expression due to Akt activation was abolished by MCL treatment. Thus, MCL inhibits LPS-induced PI3K/Akt/p70S6K activation, at least partially accounting for the decreased IL-10 expression.
In conclusion, MCL inhibits LPS-induced inflammatory response and protects mice from LPS challenge via NF-κ B and PI3K/Akt pathways (Fig. 8). Further studies are needed to investigate the precise molecular targets of MCL. Our results provide new insights into the regulation of LPS-triggered inflammatory responses and indicate a promising anti-septic role of MCL in acute peritonitis mouse model. MCL represents an ideal natural compound and drug candidate to treat endotoxic shock, sepsis or other severe inflammation.

Materials and Methods
Mice and reagents. Female C57BL/6J mice (4-8 weeks old, weighing 20 ± 3 g) were obtained from Joint  with medium before use. Recombinant mouse granulocyte-macrophage colony stimulating factor (GM-CSF) and IL-4 were purchased from R&D Systems (Minneapolis, MN).
Cell culture. Mouse macrophage-like cell line Raw264.7 was obtained from ATCC (Manassas, VA) and cultured as described previously 40 . Human acute monocytic leukemia cell line THP-1 was obtained from ATCC (Manassas, VA) and was maintained in RPMI-1640 supplemented with 10% heat-inactivated FBS. Thioglycolate-elicited mouse primary peritoneal macrophages were prepared from female C57BL/6J mice (6-8 weeks of age) as described previously 28 . After 2 h, non-adherent cells were removed and the adherent cells were used as peritoneal macrophages. Bone marrow-derived dendritic cells (DCs) from C57BL/6J mice (4 weeks of age) were generated as described 41 . Peripheral blood mononuclear cells (PBMCs) of healthy human donors were prepared by density gradient isolation using Ficoll-Paque (Sigma-Aldrich). To purify CD14 + monocyte from PBMCs, cells were separated with CD14 + cell isolation kit (Miltenyi Biotech, Germany).

Cell viability assay.
Western blot analysis. Cells were lysed with M-PER TM Protein Extraction Reagent (Pierce, Rockford, IL) supplemented with protease inhibitor cocktail (Calbiochem, San Diego, CA). The protein concentration of each sample was measured with BCA assay (Pierce, Rockford, IL). Equal amounts of extracts were separated by 10% SDS-PAGE, transferred onto a PVDF membrane, and then blotted as described previously 28,29 . Actin was used as an internal control.
Plasmid constructs and transfection. Akt expressing plasmid (Myr-Akt-HA) and its corresponding empty vector plasmid were kind gifts from Prof. Chaofeng Han (National Key Laboratory of Medical Immunology, Second Military Medical University). Mouse NF-κ B luciferase reporter gene plasmid and pRL-TK-Renilla-luciferase plasmid were described previously 23 . Transfection of Raw264.7 macrophages with jet-PEI TM (PolyPlus-Transfection, Illkirch, France) was performed according to manufacturer's instructions.
Assay of luciferase reporter gene expression. Raw264.7 was cotransfected with the mixture of 100 ng NF-κ B luciferase reporter plasmid and 20 ng pRL-TK-Renilla-luciferase plasmid using jetPEI TM (Polyplus). After 30 h, cells were stimulated with LPS for 24 h, and luciferase activities were measured with Dual-Luciferase Reporter Assay System (Promega) according to the manufacturer's instructions. To exclude the influence of transfection efficiency, data were normalized by division of Firefly luciferase activity with that of Renilla luciferase. The relative values were presented as fold increase 23 . In vivo LPS challenge and serum cytokine detection. Female C57BL/6J mice (6-8 weeks old) were injected intraperitoneally with LPS (10 mg/kg; Sigma) with or without MCL or DXM. LPS was prepared in sterile phosphate-buffered saline (PBS) before use. Mice were killed 2 h after injection and plasma samples were clotted and collected for cytokine analysis by ELISA. PBS or dexamethasone (DXM) (7 mg/kg, equivalent to the clinical dose) were selected as the negative or positive control, separately 28 . The synergetic role of MCL and DXM was verified in vivo.
Histopathology. Female C57BL/6J mice (6-8 weeks old) were injected intraperitoneally with LPS (10 mg/kg), MCL or DXM as indicated. After 12 h, mice were sacrificed and lung or liver tissues were collected to be fixed with 4% formaldehyde and paraffin-embedded. The tissues were sliced and stained with hematoxylin and eosin (H&E). The histopathological changes were observed using a Zeiss Imager M2 microscope (Carl Zeiss MicroImaging) equipped with an AxioCam HRc CCD camera (Carl Zeiss).
Scientific RepoRts | 6:23240 | DOI: 10.1038/srep23240 LPS-induced septic shock mouse model. The septic shock mouse model was established by intraperitoneal injection of LPS (20 mg/kg; Sigma) as described 42 . The experiments were carried out in the following groups: PBS treatment group; MCL (20 mg/kg) treatment group; LPS treatment group; LPS + MCL (5 mg/kg) group; LPS + MCL (10 mg/kg) group; LPS + MCL (20 mg/kg) group; LPS + dexamethasone (DXM) (7 mg/kg) group and the combination treatment LPS + DXM + MCL (10 mg/kg) group. The survival statuses of different groups were recorded at different intervals as described previously 28 . Statistical analysis. Results were given as means ± standard deviation (SD) or means ± standard error (SE).
Comparisons between 2 groups were done using Student's t test analysis. Survival analysis between multiple groups were done using Log-Rank test. The survival curve was drawn by Sigmaplot software. Statistical significance was determined as p < 0.05 or p < 0.01.