ACT001 was kindly provided by Tianjin Shangde Pharmaceutical Margin Technology Co., Ltd. Microglial BV-2 cells were obtained from China Center for Type Culture Collection. Ultrapure lipopolysaccharide (LPS), HEK Blue TLR4 293 cells, and HEK-Blue Selection were obtained from Invivogen. Phospha-Light™ SEAP Reporter Gene Assay System was purchased from Applied Biosystems. Dual-Glo Luciferase Assay System was purchased from Promega. Crystal violet, 2,3-diaminonaphthalene, protease inhibitor cocktails, and anti-β-actin antibodies were purchased from Sigma-Aldrich. Dulbecco’s Modified Eagle Medium (DMEM), TRIzol, RIPA buffer were purchased from Thermo Fisher Scientific. Fetal bovine serum was purchased from PAN-Seratech. RNeasy Mini Kit, RT² Easy First Strand cDNA Synthesis Kit, PCR primers and SYBR Green PCR Master Mix were obtained from Qiagen. Primary MD2 antibody, anti-Iba1 antibody, anti-TLR4 antibody and anti-GFAP antibody were purchased from Abcam. Primary antibodies targeting MyD88, p38 MAPK, NF-κB p65, ERK (1/2), IKK-β, SAPK/JNK, phospho-NF-κB p65, phospho-ERK(1/2), phospho-SAPK/JNK, phospho-IKK-α/β, and phospho-p38 MAPK antibodies were obtained from Cell Signaling Technology.

MD2 expression and purification were performed as described previously (22, 23). Fluorescence titrations of MD2 with ACT001 were performed on a Cary Eclipse spectrofluorometer. All measurements were carried out at room temperature using a 2×10 mm quartz cell with MD2. The fluorescence titration was carried out at a wavelength of 280 nm to excite the Tyr and Trp residues in MD2. Emission at 310-450 nm was measured. 0.5 μM MD2 was titrated with different concentrations of ACT001 and fluorescence intensity at 337 nm was plotted against ACT001 concentration.

MD2 was prepared in a phosphate buffer in D₂O (75 mM potassium phosphate, 150 mM sodium chloride, pH 7.5)and ACT001 was dissolved in DMSO-d6 (50 mM) as stock solution. Saturation transfer difference NMR spectroscopy experiments were performed to investigate ligand-protein interactions. NMR spectra were acquired at 25 °C in a Bruker Avance III-600 MHz (proton frequency) spectrometer with a conventional inverse 5 mm probe head with z-gradients using standard Bruker pulse programs. Samples containing 400 μM ACT001 in the absence or presence of MD2 (4 μM) in D₂O buffer were used for NMR spectra data acquisition.

Cellular thermal shift assay (CETSA) was performed as described (21).

BV-2 cells were cultured in supplemented DMEM (10% FBS, 50 U/mL penicillin, and 50 μg/mL streptomycin) and seeded at a density of 4×10⁴ cells per well in 96-well plates. After overnight incubation, media was aspirated and changed to DMEM media without FBS. Cells were then treated with LPS (200 ng/mL) and the indicated concentrations of ACT001. The NO concentration in the culture supernatant was determined by the 2,3-diaminonaphthalene-based fluorescent method as described (23).

BV-2 cells were cultured as above and treated with LPS (200 ng/mL) and 100 μM ACT001 for 6 h. Cell morphology images were collected by a Nikon microscope.

Cellular viability was determined by the crystal violet staining method and CCK-8 Kit as described (23, 40).

SEAP assay was performed as described (40).

Dual-luciferase NF-κB reporter assay was performed as described (40).

BV-2 cells were seeded at 4×10⁶ cells/dish in 100 mm culture dishes. After 24 h incubation, cells were stimulated by LPS (200 ng/mL) and indicated concentrations of ACT001 for 1 h. Cells were washed twice with ice-cold PBS and lysed in 1 mL Co-IP lysis buffer (25 mM Tris pH 8.0, 150 mM KCl, 5 mM EDTA, 0.5% NP-40) with a complete protease inhibitor cocktail, 1 mM DTT, and 1 mM PMSF by incubating on ice for 30 min. Cell supernatant was collected via centrifugation at 12,000 g at 4 °C for 12 min and incubated with corresponding primary antibody at 4 °C overnight. Washed magnetic beads were then incubated with the samples at room temperature for 1 h. The magnetic beads were washed twice with PBS and boiled with 50 μL 2 × SDS sample buffer at 100 °C for 8 min for immunoblotting.

Immunoblotting was performed as described (23).

BV-2 cells were seeded at a density of 4×10⁵ cells/well in 6-well plates. After overnight incubation, BV-2 cells were treated with LPS (200 ng/mL) and indicated concentrations of ACT001 for 6 h. Total RNA was isolated from BV-2 cells using TRIzol reagent and cDNA was generated with an oligo (dT) primer. Primer sequences are shown in Table 1 . The ribosomal protein L27 gene RPL27 was used as the internal control. qPCR was performed on a TOptical Real-Time qPCR Thermal Cycler (Analytik Jena, Thuringia, Germany) using the SYBR Green method. The data were analyzed by the 2^−ΔΔCT method and were normalized to RPL27.

Origin 8 was used for the plotting of the data and statistical analysis. Non-linear Logistic regression was used to plot and analyze concentration-response curves and to obtain IC₅₀. For the analyses of qRT-PCR data, immunoblotting data, von Frey test and quantification of immunofluorescence, an unpaired Student t-test was used for comparisons between two groups. Data are presented as mean ± SEM. P-value summary is mentioned on the bar of each figure. # P< 0.05; ## P< 0.01; ### P< 0.001 versus the control/sham group; *P < 0.05; **P < 0.01; ***P < 0.001 versus the LPS/CCI group. ns, not significant. P< 0.05 was considered statistically significant in all analyses.

Besides its anticancer action, ACT001 ( Figure 1A ) has also been reported to alleviate neuroinflammatory responses in the CNS (16). However, the molecular target responsible for the immunosuppressive effects of ACT001 is not known. Herein the acting of ACT001 is hypothesized to be, at least in part, mediated by TLR4, which plays a fundamental role in regulating innate immunity. MD2, a co-receptor of TLR4, is responsible for ligand recognition (45). Fluorescence quenching titration of MD2 was first performed to explore the possible interaction of ACT001 with MD2 as the potential target for the inhibition of innate immune signaling. ACT001 caused the quenching of MD2 intrinsic fluorescence ( Figure 1B ). A dissociation constant K_d of 2.8 ± 0.3 μM was derived by the nonlinear least-square fitting of the titration curve of MD2-ACT001 interaction. Moreover, saturation transfer difference (STD) nuclear magnetic resonance (NMR) was employed to characterize transient receptor-ligand interaction. Only ligand protons that are in close contact with the receptor-binding site and receive magnetization transfer will appear in the difference spectrum (46, 47). As the difference spectra shown in Figure 1C , hydrogens of methyl at positions 8, 19, 20 and 21 exhibited the most favorable binding characteristics, which confirm the direct interaction between ACT001 and MD2. To explore whether MD2 is the endogenous target of ACT001, cellular thermal shift assay (CETSA) was performed. CETSA is based on the principle that drug binding leads to the thermal stability change of the target protein as reflected by the shift of its melting temperature (T_m). ACT001 binding decreased the T_m value of MD2 by 6.4 ± 0.9 °C ( Figures 1D, E ), which indicates ACT001 directly binds to MD2 in the cellular context. Taken together, these biophysical binding characterizations show MD2 is a direct target of ACT001.

In order to investigate how ACT001 interacts with MD2, molecular docking and molecular dynamics simulation were conducted. ACT001 was found to dock into the conserved hydrophobic cavity and overlap with the space of R2’, R3 and R2’’ chains of Lipid A in MD2, therefore hindering the binding of LPS to MD2 ( Figure 2A ). The best docking pose was refined using molecular dynamics simulations. The root-mean-square deviation (RMSD) analysis of backbone atoms of apo-MD2 and MD2 bound with ACT001 showed that both systems reached stable states during 400 ns simulations ( Figure 2B ). To investigate the flexibility changes caused by ACT001, the root-mean-square fluctuation (RMSF) analysis was conducted with the last 100 ns equilibrated trajectories. The binding of ACT001 rendered most regions of MD2 to be more flexible ( Figure 2C ), indicating that ACT001 destabilizes MD2. This result is consistent with the experimental CETSA data. The exposed solvent-accessible surface areas (SASA) of MD2 ( Figure 2D ) did not change upon interacting with ACT001. Interestingly, further analysis showed that ACT001 binding decreased the percentage of hydrophobic area in the buried SASA of MD2 ( Figure 2E ). It should be noted that the hydrophobic residues prefer to be buried inside owing to the hydrophobic interactions to stabilize the apo-MD2 (48). These in silico simulation results explicitly explain that ACT001 binding decreases MD2 stability.

The detailed binding mode of ACT001 with MD2 was subsequently investigated. Figure 2F shows the representative pose of ACT001 binding to MD2 after the molecular dynamic equilibration. ACT001 formed hydrogen bonds with surrounding water molecules and formed hydrophobic interactions with residues of MD2. Specifically, methyl at position 8 of ACT001 interacted with Ile97 and Pro98; methyl groups at positions 19 and 20 of ACT001 formed interactions with Val73; methyl at position 21 of ACT001 interacted with Phe101 and Cys113. These are consistent with STD-NMR results. In addition, Phe56, Leu58, Glu72, Phe99 and Phe131 were also found to form hydrophobic interactions with ACT001. All these in silico simulation results confirmed that ACT001 interacted with MD2 and decreased the stability of MD2.

TLR4 activation leads to the recruitment of myeloid differentiation primary response protein 88 (MyD88) to activate NF-κB and MAPKs. Immunoprecipitation and immunoblotting were employed to measure the effect of ACT001 on TLR4 downstream signaling. As shown in Figures 3A-C , ACT001 inhibited LPS-induced MyD88 recruitment of TLR4 and significantly suppressed the formation of TLR4/MD2/MyD88 complex. LPS induced the phosphorylation of IKKβ, p65, JNK, ERK as well as p38 and ACT001 significantly inhibited LPS induced phosphorylation of these TLR4 signaling factors in a concentration-dependent manner ( Figures 3D-I ).

To further quantitatively investigate the effect of ACT001 on TLR4 signaling NF-κB activity, HEK TLR4 cell line with a SEAP reporter gene, under the control of NF-κB responsive element, was used. ACT001 was found to inhibit LPS-induced NF-κB activation in a dose-dependent manner, with an IC₅₀ of 22.4 ± 0.3 μM while no apparent cellular toxicity of ACT001 was observed within 200 μM ( Figure 4A ). In addition to HEK based NF-κB reporter cell, the effect of ACT001 on NF-κB activity in BV-2 cells, which reproduces many of the responses of immunocompetent primary microglia with high fidelity (49),was also examined. ACT001 inhibited LPS-induced NF-κB activation in BV-2 cells in a dose-dependent manner with an IC₅₀ of 24.1 ± 1.3 μM without apparent cellular toxicity within 200 μM ( Figure 4B ). These data clearly show that ACT001 inhibits TLR4 signaling NF-κB activation.

The pro-inflammatory mediators are downstream effectors of TLR4 innate immune responses. ACT001 inhibited LPS induced nitric oxide (NO) overproduction in BV-2 cells in a concentration-dependent manner with an IC₅₀ of 16.0 ± 1.2 μM ( Figure 5A ). No apparent cellular toxicity of ACT001 was observed, even at the concentration of 100 μM, which eliminates the possibility of the observed inhibition of TLR4 signaling by ACT001 was owing to cell death ( Figure 5A ). qRT-PCR was performed to measure the effect of ACT001 on LPS-induced pro-inflammatory factors iNOS, IL-1β, IL-6 and TNF-α mRNAs expression. ACT001 suppressed LPS-induced iNOS ( Figure 5B ), IL-1β ( Figure 5C ), IL-6 ( Figure 5D ) and TNF-α ( Figure 5E ) mRNA expression in a concentration-dependent manner. The LPS-induced cell body was enlarged, and the percentage of amoeboid-like microglia increased significantly, while ACT001 markedly reversed the LPS induced morphological changes in BV-2 cells ( Supplementary Figure 1 ). Together, these cellular signaling characterizations demonstrate that ACT001 inhibits the formation of TLR4/MD2/MyD88 complex and LPS-induced activation of NF-κB and MAPKs, therefore inhibiting TLR4 signaling downstream pro-inflammatory factors.

Chronic constriction injury (CCI)-induced allodynia is associated with up-regulated TLR4 expression in spinal cord and TLR4 antagonism has been shown to attenuate neuropathic pain (41).The mechanical threshold of ipsilateral hind paw of the CCI group decreased significantly compared to the sham group ( Figure 6 ). Repeated systemic administration of ACT001 resulted in significant attenuation of allodynia since the 21^stday after CCI surgery without interrupting weight gain in animals ( Figure 6 and Supplementary Figure 2 ). To further analyze whether the attenuation of allodynia by ACT001 was associated with a decrease in the expression of glial activation markers, the lumbar spinal cords (L4-L6) of three group animals were collected following the final behavioral testing. These tissues were stained for microglia and astrocyte activation markers Iba-1 and GFAP, respectively. The expression of Iba-1 and GFAP in the ipsilateral spinal dorsal horn segments in the CCI group significantly increased when compared to the sham group ( Figures 7 , 8 ). Moreover, the number and size of glia were also elevated in CCI group compared to sham group. ACT001 significantly inhibited the CCI-induced overexpression of glial markers as well as the Iba1- and GFAP-positive area and the cell number ( Figures 7 , 8 ). These results show ACT001 attenuates neuropathic pain and glial activation.