Paeonol alleviates neuropathic pain by modulating microglial M1 and M2 polarization via the RhoA/p38MAPK signaling pathway

Abstract Background This study aimed to investigate the potential mechanism of paeonol in the treatment of neuropathic pain. Methods Relevant mechanisms were explored through microglial pseudotime analysis and the use of specific inhibitors in cell experiments. In animal experiments, 32 SD rats were randomly divided into the sham operation group, the chronic constrictive injury (CCI) group, the ibuprofen group, and the paeonol group. We performed behavioral testing, ELISA, PCR, Western blotting, immunohistochemistry, and immunofluorescence analysis. Results The pseudotime analysis of microglia found that RhoA, Rock1, and p38MAPK were highly expressed in activated microglia, and the expression patterns of these genes were consistent with the expression trends of the M1 markers CD32 and CD86. Paeonol decreased the levels of M1 markers (IL1β, iNOS, CD32, IL6) and increased the levels of M2 markers (IL10, CD206, ARG‐1) in LPS‐induced microglia. The expression of iNOS, IL1β, RhoA, and Rock1 was significantly increased in LPS‐treated microglia, while paeonol decreased the expression of these proteins. Thermal hyperalgesia occurred after CCI surgery, and paeonol provided relief. In addition, paeonol decreased the levels of IL1β and IL8 and increased the levels of IL4 and TGF‐β in the serum of CCI rats. Paeonol decreased expression levels of M1 markers and increased expression levels of M2 markers in the spinal cord. Paeonol decreased IBA‐1, IL1β, RhoA, RhoA‐GTP, COX2, Rock1, and p‐p38MAPK levels in the spinal dorsal horn. Conclusion Paeonol relieves neuropathic pain by modulating microglial M1 and M2 phenotypes through the RhoA/p38 MAPK pathway.


| INTRODUC TI ON
Neuropathic pain (NP) is pain caused by injury or disease of the somatosensory nervous system. 1 A total of 6.9%-10% of the world's population suffers from NP. 2 Patients experience NP even after the initial stimulus has been removed. 3 The National Institute for Health and Clinical Excellence (NICE) guidelines recommend NSAIDs for neuropathic pain relief, 4 but NSAID resistance and long-term side effects continue to plague patients. [5][6][7] Therefore, there is an urgent need for novel medicine for the treatment of NP.
Microglia are resident macrophage-like immune cells in the central nervous system thought to play an important role in NP. 8 Numerous studies have shown that inhibition of microglial activation can alleviate NP. [9][10][11][12] Microglia can be activated into the M1 phenotype and the M2 phenotype, 13 in which the M1 phenotype is mainly involved in pro-inflammatory responses and the M2 phenotype is mainly involved in anti-inflammatory responses. 14 Inflammatory factors such as IL1β and TNFα secreted by M1 phenotype microglia mediate the abnormal excitability of spinal neurons leading to hyperalgesia, 8 while M2 phenotype microglia produce opioid peptides to reduce pain. 15 Therefore, converting the M1 phenotype to the M2 phenotype may be a strategy for treating NP.
RhoA and Rho-related kinase (Rho kinase; ROCK) play key roles in neuropathic pain. 16,17 Studies have shown that ROCK inhibition reduces thermal hyperalgesia induced by TRP channels. 18 The ROCK pathway may lower the spinal cord nociceptor threshold and be involved in the development and maintenance of neuropathic pain. 19 RhoA/p38MAPK activates microglia in the spinal cord and increases neural excitability through P2Y12 receptors. 20 Paeonol is an anti-inflammatory and antioxidant compound widely distributed in Paeonia lactiflora. Several studies have shown that paeonol can relieve pain in mouse and rat models. 21,22 Paeonol can inhibit the expression of inflammatory factors in microglia, which indicates that paeonol can relieve NP through microglia. However, whether paeonol can regulate microglial M1 and M2 polarization and the underlying mechanism in NP remains unknown.
The purpose of this study was to determine the therapeutic effect of Paeonol on neuralgia and related mechanisms. We investigated the underlying mechanisms at the cellular level through single-cell Pseudotime analysis and in vitro experiments. The therapeutic effect of paeonol on the CCI model was evaluated by behavioral examination, pathological examination, and immunohistochemistry.

| NP model
Based on previous studies, we chose to generate the chronic contractile injury (CCI) pattern of the sciatic nerve as a disease model for sciatica. The right sciatic nerve of rats was exposed following an intraperitoneal injection of sodium pentobarbital (3%; 40 mg/ kg), and the right sciatic nerve was ligated under a microscope using 4.0 sutures (repeated 4 times at ~1 mm intervals). Rats in the sham-operated group were not subjected to nerve ligation.
To prevent infection, gentamicin (10 mg/mL) was injected into the right biceps femoris muscle. A total of 18 rats were randomized into two groups: the CCI and sham operation groups. After 21 days, the rats were sacrificed with an overdose of sodium pentobarbital (200 mg/kg).

| Single-cell RNA-seq data processing and quality control
Single-cell sequencing data of mouse spinal cord microglia 14 days after peripheral nerve injury were obtained from the sequencing data of Mouse microglia 207-R, Mouse microglia 210, and Mouse microglia 212 in the GEO database GSE162807 cohort. It contains transcriptome sequencing data for 9055 cells. To ensure the reliability of the data analysis, we removed low-quality cells and retained cells with more than 200 genes but fewer than 4000 genes and fewer than 20% mitochondrial genes. We then applied the Seurat R package for quality control. Gene expression measurements per cell were normalized by total expression using the global scaling normalization method "LogNormalize," and then the results were multiplied by a scaling factor (default 10,000) and log-transformed in each dataset. We annotated the data using the built-in dataset "ImmGenData" in the "SingleR" package, which re-  We used PyMol software to remove all water molecules from the 3D structure of the protein and added polar hydrogens and Gasteiger charges to the protein and ligands using AutoDockTools 1.5.6. The AutoDock tool is used to match compounds and genes, where AutoDock Vina is used to find optimal docking conditions. Docking energies were used to evaluate the results of molecular docking. PyMol enables analysis and visualization of docking results.

| Cell grouping and drug concentration
The cells were divided into 6 groups: the control group, the LPS group (10 μg/mL LPS), the CCG group (10 μg/mL LPS + 10 μM CCG-1423), the SB203580 group (10 μg/mL LPS + 10 μM SB203580), 18 the LPS + paeonol group (10 μg/mL LPS + 4 μM paeonol), and the paeonol group (4 μM paeonol). GMI-R1 cells were cultured in 6well plates for 24 h. The medium was then discarded, and the cells were washed with PBS. The corresponding drugs and LPS were added to the medium at the same time, and then the cells were cultured for 24 h.

| Flow cytometry analysis
Cells were seeded in 6-well dishes at a density of 1 × 10 6 per well and treated with drugs and LPS for 24 h. Cells were then digested in the dishes by trypsinization, washed, and resuspended in cold PBS.
The membrane protein CD32 was detected by direct immunofluorescence staining, and the cells were incubated with PE-conjugated CD32 antibody for 30 min at room temperature in the dark. For

| Immunofluorescence
Glass slides coated with poly-L-lysine (0.1 mg/mL) were placed in a 24-well plate. GMI-R1 cells were cultured and processed as instructed, fixed with 4% paraformaldehyde for 30 min at room temperature, and then permeabilized with 0.3% Triton X-100 for 15 min.
Cells were blocked with PBS containing 10% donkey serum for 1 h.
Cells were then incubated with rabbit anti-CD32 and CD206 primary antibodies (1:200) in a humidified chamber at 4°C overnight. The next day, the cells were washed with PBS and incubated with goat anti-rabbit secondary antibody (1:400) for 50 min at room temperature. After that, the slides were washed with PBS, and DAPI staining solution was added. After incubation in the dark at room temperature for 10 min, the slides were mounted with anti-fluorescence quenching mounting medium. All images were captured with a fluorescence microscope (OLYMPUS, BX53).

| Animal treatment
Thirty-two rats were randomly divided into the sham operation group, the CCI group, the ibuprofen group and the paeonol group, with 8 rats in each group. Rats in the sham operation group and CCI group were given normal saline (0.9%, 6 mL/kg/bid), the ibuprofen group was given ibuprofen (31.5 mg/kg/bid) and the paeonol group was given paeonol (100 mg/kg/bid) by intragastric administration, 23 starting on the first day after surgery and continuing for 21 days.

| Hot plate experiment
Thermal hyperalgesia was evaluated by the hot plate test (YLS 6B; Jinan Yiyan Technology Development Co., Ltd.). To measure the paw withdrawal threshold (PWT), the lateral plantar surface of the right paw was placed on a heating plate (50°C), and the time required for the rat to lift the right foot was recorded. PTW values were recorded on the first day before surgery and on days 1, 4, 7, 14, and 21 after surgery.

| H&E staining
We fixed ligated sciatic nerve tissues in 4% paraformaldehyde for 1 day, embedded them in paraffin, and sectioned them at 3 μm thickness. Sections were deparaffinized in xylene and rehydrated by 100, 90, 80, and 70% ethanol. We then rinsed the slices in PBS for 5 min. After staining, two pathologists blinded to the experimental design viewed the stained images and assessed tissue damage.

| ELISA
Rat serum frozen in a −80°C refrigerator was taken out to detect

| Immunohistochemistry
The right sciatic nerve was fixed in 4% paraformaldehyde for 24 h, immersed in 20% (w/v) sucrose solution, sectioned, and then slices from individual animals were incubated in sodium citrate antigen retrieval solution (1:1000 dilution; pH = 6). Next, sections were incubated with primary and secondary antibodies. Then, DAB was added. Finally, the slides were washed in distilled water, dehydrated, made transparent, and sealed with xylene. The slices were examined under an OLYMPUS fluorescence microscope (BX53). The mean density was calculated using Image-Pro Plus software (Media Cybernetics, Inc.) and statistical analysis was performed.

| Quantitative real-time PCR (qRT-PCR)
RNA was extracted using RNAiso Plus and reverse-transcribed to cDNA using an RT-qPCR kit according to the manufacturer's instructions. The amplification parameters were 95°C for 1 min, followed by 40 cycles of 95°C for 10 sec and 60°C for 30 s. The relative gene expression data were analyzed by RT-qPCR using the 2 −ΔΔCT method. The primers used are listed in Table 1.

| Western blotting
The right spinal dorsal horn of the rat was homogenized in RIPA buffer containing 1 mM of phenylmethylsulfonyl fluoride (PMSF) and centrifuged at 14,000× for 20 min. Total protein was extracted from GMI-R1 cells in RIPA lysis buffer containing 1 mM of PMSF. We acquired the membrane fraction using a cell membrane protein and cytoplasmic protein extraction kit. The proteins (30 μg) were separated by 12% SDS-PAGE and transferred onto PVDF membranes;

| Expression of the RhoA/p38MAPK pathway in microglia after peripheral nerve injury
To gain insight into the role of M1/M2 polarization of macrophages  Figure 1J). Mapk14 was highly expressed in the other 7 clusters except the 7th and 8th clusters ( Figure 1K).

| Differentiation trajectories of microglia after peripheral nerve injury
Since the RhoA/p38MAPK pathway is highly expressed in microglia, to further study the biological function of this pathway in the differentiation of microglia, we used "Monocle2" to analyze the differentiation trajectory. We found that microglia could be classified into seven states as the pseudotime course progressed (Figure 2A-C).
We examined changes in the RhoA/p38MAPK pathway in microglial differentiation trajectories. We found that Rhoa, Rock1, and Mapk14 were highly expressed in state 1, then the expression levels of Rhoa and Rock1 began to decrease in states 2, 3, and 4, increased in states 5 and 6, and finally in state 7. Mapk14 with Pseudotime changes slowly decrease. Rock 2 shows a slow upward trend from states 1 to 7 ( Figure 2D). Then, in order to understand whether the expression trends of proteins in the RhoA/p38 pathway are consistent with the expression trends of M1 and M2 markers. We performed a "DifferentialGeneTest" on the expression pattern of the RhoA/p38 pathway using markers for M1 and M2. We found that the expression pattern of Rhoa, Rock1, and Mapk14 was consistent with that of the M1 markers Fcgr2b (CD32) and CD86, whereas that of Rock2 was consistent with the M2 markers Mrc1 and Clec10a ( Figure 2E). This result suggests that the RhoA/p38MAPK pathway may be involved in the activation of microglia with the M1 phenotype.

| Validation of the interaction between paeonol and RhoA by molecular docking
To verify the binding ability of paeonol to the RhoA/p38 MAPK pathway, we performed molecular docking analysis and evaluated the binding energy of the four proteins for paeonol. The binding energy is less than −1.2 kcal/mol, indicating good binding activity.
The results showed that paeonol and RhoA proteins interacted stably through three hydrogen bonds ( Figure 3A). The results of this analysis are shown in Table 2. The 3D map of the interaction of paeonol with Rock1, Rock2, and p38 MAPK is shown in Figure 3B-D.

| RhoA, p38MAPK inhibitor, and paeonol convert M1 microglia to M2 microglia in vitro
According to the results of the CCK8 assay, 4 μM of paeonol did not affect the viability of GMI-R1 cells, but above 8 μM, the cell viability  Paeonol also promoted the transformation of LPS-treated GMI-R1 microglia from the M1 phenotype to the M2 phenotype.

| Effects of RhoA, p38MAPK inhibitor, and paeonol on inflammatory cytokines produced by LPSinduced GMI-R1 cells
To observe the effects of RhoA, p38MAPK inhibitor, and paeonol

| Paeonol alleviates thermal hyperalgesia in CCI model rats
Rats had the most severe thermal hyperalgesia on the 4th day after CCI surgery. The paeonol group and ibuprofen group relieved the thermal hyperalgesia from the 7th day after the treatment but failed to return to normality on day 21. No significant differences were observed between paeonol and ibuprofen ( Figure 5A).

| Paeonol decreases serum inflammatory factor levels in CCI model rats
Serum concentrations of IL1β, IL8, IFNγ, IL4, and TGFβ were measured. IFNγ and IL4 polarized microglia into the M1 and M2 phenotypes. IL1β and IL8 are produced by polarized M1 microglia, while TGFβ is produced by polarized M2 microglia. The results showed that the serum levels of IL1β, IL8, and TGFβ in the CCI group were higher than those in the sham group, while the levels of IL4 were not different from those in the sham group. Compared with the CCI group, ibuprofen and paeonol decreased the levels of IL1β and IL8 in serum ( Figure 5B,C, p < 0.05) and increased the levels of IL4 and TGFβ ( Figure 5E,F, p < 0.05). However, there was no difference in IFNγ among the four groups ( Figure 5D). This finding indicates that paeonol can reduce the level of M1 phenotype proinflammatory factors and increase the level of M2 phenotype anti-inflammatory factors in the serum of CCI model rats.

RhoA and inflammatory factors after CCI surgery
We observed sciatic nerve structures by H&E staining. The myelin sheath of the sciatic nerve was destroyed and infiltrated by multiple inflammatory cells after CCI surgery. Myelin structure was restored in the paeonol group and ibuprofen group, and the level of inflammatory cell infiltration was significantly reduced compared with that in the CCI group ( Figure 5G). The expression levels of RhoA, COX2, nNOS, and IL1β were observed using immunohistochemistry.
COX2 is one of the markers of inflammation and pain, and nNOS is a marker of nervous injury. We found that the expression levels of RhoA, COX2, nNOS, and IL1β increased substantially after CCI surgery. Paeonol and ibuprofen reduced the expression levels of RhoA, COX2, nNOS, and IL1β ( Figure 5J). All these findings suggest that paeonol can improve the pathological structure of the sciatic nerve and inhibit the expression of RhoA and inflammatory factors in CCI rats. The liver and kidney structures were normal in all groups ( Figure 5H,I), indicating that the administration and CCI operation had no adverse effects on the liver and kidney of rats.

| Paeonol reduces the expression of M1 markers and increases the expression of M2 markers in the spinal cord
To further explore whether paeonol can regulate the switch of microglial cells from M1 to M2 phenotype in the spinal cord in vivo, we examined the levels of M1 marker CD32 and M2 marker CD206 in the spinal cord by immunofluorescence. The results showed that the fluorescence intensity of spinal cord CD32 in the CCI model group was higher than that in the sham group, and the fluorescence intensity of spinal cord CD32 in the CCI rats was significantly reduced after ibuprofen and paeonol treatment, but not reduced to the same level as the Sham group. Compared with the sham group, the CD206 fluorescence intensity of the spinal cord of the rats in the CCI model group was slightly increased, while the CD206 fluorescence intensity of the CCI rats was significantly increased after the treatment of ibuprofen and paeonol ( Figure 6A). We detected the levels of M1 marker and M2 marker in the spinal cord by qPCR and Western blot. The results showed that compared with the sham group, the expression levels of CD32 and CD206 in the CCI group increased, while paeonol and ibuprofen could reduce the expression levels of CD32 and iNOS. For the M2 markers CD206 and ARG-1, there was no difference in the expression levels of CD206 and ARG-1 in the CCI group compared with the Sham group, while ibuprofen and paeonol could significantly increase the expression levels of CD206 and ARG-1( Figure 6B). This shows that paeonol can switch M1 phenotype microglia to M2 phenotype in the spinal cord of CCI model rats.

| Paeonol inhibits the RhoA/p38MAPK pathway and the activation of microglia in the spinal dorsal horn of CCI rats
To understand the expression of the RhoA/p38MAPK pathway in microglia in the spinal cord after sciatic nerve injury and to determine whether microglia are activated, we examined the RhoA/ p38MAPK pathway and the microglial activation marker IBA-1 in the spinal dorsal horn by Western blotting. The results showed that the protein levels of RhoA, Rock1, p-p38MAPK, COX2, IL1β, and IBA-1 increased in the CCI group, while paeonol and ibuprofen decreased the expression levels of these proteins. There was no significant difference in p38MAPK and Rock2 among the 4 groups. Paeonol reduced RhoA, Rock1, p-p38MAPK, COX2, and IL1β levels to nearnormal levels but failed to restore IBA-1 levels to normal ( Figure 6C).
This finding indicates that paeonol can reduce neuroinflammation by inhibiting the activation of spinal microglia through the RhoA/ p38MAPK pathway.

| DISCUSS ION
This study provides evidence that paeonol may relieve sciatica by Microglia are resident brain immune cells involved in innate immunity. 24 Growing evidence suggests that microglia play a crucial role in the development and maintenance of neuropathic pain. 25 Activation of microglia and neuron-glia interactions are considered important mechanisms mediating neuroinflammation and central sensitization. 26 Central sensitization and neuroinflammation are the main mechanisms leading to persistent hyperalgesia. 27  Inflammation is an important factor leading to M1 polarization.
LPS and IFNγ can induce M1 polarization, and the M1 phenotype secretes proinflammatory factors and aggravates neuroinflammation, forming a vicious cycle. 33 We found by serum ELISA that paeonol can reduce the serum proinflammatory cytokines IL1β and IL8 and increase the anti-inflammatory cytokines IL4 and TGFβ. Through HE staining and immunohistochemical analysis of the sciatic nerve, paeonol was found to reduce the infiltration level of inflammatory cells in the sciatic nerve and the levels of COX2, nNOS, Ilβ, and iNOS in sciatic nerve tissue. These results indicate that paeonol can reduce neuroinflammation and thus relieve pain.
Several studies have shown that the RhoA/p38MAPK pathway promotes the activation of microglia. 18,34 In this study, microglial trajectory analysis and GMI-R1 cell experiments confirmed that the RhoA/p38MAPK pathway can lead to the M1 polarization of microglial cells. Studies have shown that Rock activity may be involved in neuropathic, inflammatory, and pain mechanisms. [35][36][37] Rock modulates neuronal activity by affecting nitric oxide production and plays a key role in microglial activation through p38MAPK. 38 In this study, we found that the expression of the RhoA/p38MAPK pathway con-

| CON CLUS ION
Our study shows that paeonol attenuates central sensitization by modulating the M1 and M2 polarization of microglia, thereby suppressing the resulting neuropathic pain.

AUTH O R CO NTR I B UTI O N S
Xin Li, Huimei Shi, Di Zhang, and Guoping Zhao contributed substantially to the experimental design, data analysis, and experimental procedures. Bei Jing, Zhenni Chen, YaChun Zheng, and Shiquian Chang assisted with the English writing. Guoping Zhao is the corresponding author.

FU N D I N G I N FO R M ATI O N
This study was supported by the National Natural Science Foundation of China (grant No. 81874404).

CO N FLI C T O F I NTER E S T S TATEM ENT
The authors declare no conflicts of interest.

DATA AVA I L A B I L I T Y S TAT E M E N T
The data used to support the findings of this study are available from the corresponding author upon request.