Botulinum toxin type A promotes microglial M2 polarization and suppresses chronic constriction injury-induced neuropathic pain through the P2X7 receptor

Background Switching microglial polarization from the pro-inflammatory M1 phenotype to the anti-inflammatory M2 phenotype represents a novel therapeutic strategy for neuropathic pain (NP). This study aims to investigate whether botulinum toxin type A (BTX-A) regulates microglial M1/M2 polarization by inhibiting P2X7 expression in a rat model of NP. Results The BTX-A administration elevated pain threshold, induced microglial polarization toward the M2 phenotype, and decreased P2X7 protein level in a rat model of NP induced by chronic compression injury (CCI). Lipopolysaccharide (LPS) was used to activate HAPI rat microglial cells as an in vitro inflammatory model and we demonstrated that BTX-A promoted microglial M2 polarization in LPS-stimulated HAPI microglial cells through suppressing P2X7. Conclusions Our results indicate that BTX-A promotes microglial M2 polarization and suppresses CCI-induced NP through inhibiting P2X7 receptor. These findings provide new insights into the mechanism of BTX-A in relieving NP.


Background
Neuropathic pain (NP), characterized by allodynia, hyperalgesia and spontaneous pain, is a chronic condition that can develop after a lesion or disease affecting the somatosensory system [1,2]. NP seriously affects the health and life quality of patients. So far, there is still a lack of effective treatment for NP [3,4]. Microglia, the resident macrophage cells of the central nervous system, has been shown to mediate neuroinflammation [5]. Microglia activation and subsequent pro-inflammatory responses play an important role in the development of NP [6].
Microglia can be activated in a polarizing manner into a classical phenotype (pro-inflammatory, M1) or an alternative phenotype (anti-inflammatory, M2). The M1 phenotype is characterized by increased expression of several proteins including inducible NO synthase (iNOS) and CD68, as well as increased secretion of several proinflammatory factors such as interleukin (IL)-1β, IL-6, and tumor necrosis factor (TNF)-α [7,8]. The M2 phenotype is characterized by increased expression of several proteins such as arginase-1 (Arg-1) and mannose receptor (MR/CD206), as well as increased production of antiinflammatory cytokines such as IL-4 and IL-10 [5,9]. The microglial polarization toward the pro-inflammatory M1 phenotype often occurs during NP. Convincing evidence has indicated that modulation of inflammation by inhibition of M1 polarization can be a strategy for treatment of NP [10][11][12][13].
Botulinum toxin type A (BTX-A) is an exotoxin released by Gram-positive anaerobic Clostridium botulinum and has been widely used in the treatment of dystonia and in aesthetic field [14,15]. Clinical studies have confirmed that BTX-A can effectively relieve NP with mild adverse reactions [16,17]. However, the exact mechanism of BTX-A action in NP is still undefined. Increasing studies have suggested that the analgesic effect of BTX-A is mediated through neurons and glial cells, especially microglia [18]. However, no literature has reported the effect of BTX-A on microglial polarization.
P2X7 receptor (also called P2X7) is a member of the purinergic ion channel receptor (P2X) family (P2X1-P2X7), which is predominantly present on microglia, astrocytes and neurons. P2X7 has been suggested to be involved in the pain transmission and the occurrence of NP [19]. Studies have reported that BTX-A can inhibit the expression of P2X3 receptor in dorsal root ganglion (DRG) in rats with NP [20], but the effect of BTX-A on P2X7 receptor has not been reported. Recently, P2X7 receptor has been proposed as a marker of M1 microglia [21]. The involvement of P2X7 receptor in microglial M1 polarization is further supported by in vitro results that microglial M1 polarization could be avoided by inhibition of P2X7 receptor in ischaemic conditions [22]. Thus, in the present study, we attempt to investigate whether microglial polarization plays a role in the analgesic effect of BTX-A. We also explored whether BTX-A regulates microglial polarization via P2X7 receptor.

Animals
Specific pathogen-free (SPF) Sprague-Dawley male rats (weight 200-250 g) were used in our study. The animals were kept at controlled temperature (22 ± 2 °C) with a 12 h light-dark cycle. All animal procedures were in compliance with the National Institutes of Health Guidelines for the Care and Use of Laboratory Animals. This study was approved by the Ethics Committee of the Second Affiliated Hospital of Nanchang University.

Establishment of a rat model of NP
A rat model of NP was induced by chronic compression injury (CCI). Briefly, each rat was anesthetized by intraperitoneal injection of 2% pentobarbital sodium (50 mg/ kg) before surgical procedures. After skin preparation, approximately 1-cm incision was made along the middle of the lower margin of the right femur. The sciatic nerve was exposed by blunt separation of muscles. Then 4 loose knots at 1 mm-intervals were made with 4-0 chromium catgut. The epineurium was slightly compressed, and the tightness was to the slight twitching of the toes. Subsequently, the muscle fascia and skin were routinely sutured. If the rats after CCI showed limping, the right hindlimb showed slight valgus, sometimes licking, hanging and other hindlimb protection behaviors, it indicated that the NP model was successfully established. In Shamoperated group, the sciatic nerve was exposed for 2-3 min without knotting, and then the muscle fascia and skin were sutured.

Animal groups
Sprague-Dawley rats were randomly divided into the following four groups (n = 10/group): Sham, NP, BTX-A-10, and BTX-A-20 group. Rats in the BTX-A-10 and BTX-A-20 groups were given a subcutaneous injection of 10 U/ kg BTX-A (Allergan Pharmaceuticals Ireland) and 20 U/ kg BTX-A respectively into the metatarsal surface 3 d following model establishment of CCI. Rats in the Sham and NP were injected with normal saline instead of BTX-A.

Pain threshold determination
Measurement of mechanical withdrawal threshold (MWT) and thermal withdrawal latency (TWL) was performed at 0, 3, 5, 7, 10, 12, 14 days after induction of CCI. The MWT value of the injured side of rats was measured by using an electronic von Frey device (series 2390; IITC-Life Science Instruments, Woodland Hills, CA, USA). The plantar of the right hind limb of the quiet rats was gradually pressurized until a withdrawal response of evasive leg lifting occurred. The maximum pressure was recorded and expressed in grams. The experiment was repeated for 6 times at an interval of 5 min. The MWT was calculated as the mean of 4 values after removing the maximum and minimum values.
The TWL value of the injured side of rats was measured using an automatic Plantar Test (No. 37370; Ugo Basile, Varese, Italy). The instrument emitted infrared light to irradiate the plantar of the right hind limb of rats. The instrument automatically cut off the supply of heat and recorded the time from the beginning of irradiation to the occurrence of foot shrinkage escape. The experiment was repeated for 5 times at an interval of 10 min. The TWL was calculated as the mean of 3 values after removing the maximum and minimum values.

Tissue samples
Fourteen days after CCI modeling, the rats were sacrificed under anesthesia with pentobarbital sodium (50 mg/kg). The limbs were fixed, the thoracic cavity was fully exposed, and the aorta was intubated through the left ventricle. After the liver was cut off, the heart was perfused with 0.9% NaCl solution (200 mL), followed by infusion with 4% paraformaldehyde in PBS buffer (200 mL). The L4-L6 segments of the spinal cord of rats were isolated and fixed in 4% paraformaldehyde, dehydrated in gradient ethanol, and embedded in paraffin for immunofluorescent detection. The remaining L4-L6 spinal cord segments were stored at − 80 °C and used for western blot analysis.

Immunofluorescent staining
The sections were dewaxed, rehydrated, and then immersed into citrate buffer for antigen retrieval. After that, the sections were blocked with 2% bovine serum albumin and then incubated overnight at 4 °C with the following primary antibodies: anti-ionized calcium binding adaptor molecule-1 (Iba-1)

Flow cytometry
The L4-L6 spinal cord segments were collected and made into single cell suspension. The mononuclear cells were isolated by density gradient centrifugation using Percoll and stained with anti-CD11b. The cells were then stained with the following fluorochrome-labeled antibodies: anti-CD68-PE (Invitrogen) and anti-CD206-Alexa Fluor 647 (Abcam). The cells were then fixed and analyzed with an Accuri C6 flow cytometer (Becton Dickinson, Franklin Lakes, NJ, USA).

Enzyme-linked immunosorbent assay (ELISA)
The levels of various cytokines including IL-6, IL-10, and TNF-α in cell supernatant were measured with their commercial ELISA kits (R&D Systems, Minneapolis, MN, USA) according to the manufacturer's instructions.

Statistical analysis
The data are presented as the mean ± standard deviation (SD) from three independent experiments. Oneway analysis of variance was used to analyze differences among multiple groups. Statistical analyses were performed using SPSS version 20.0 (IBM, Chicago, IL, USA). P-value < 0.05 was considered significant in all the tests.

BTX-A elevated pain threshold and promoted microglial polarization toward the M2 phenotype in NP rats
Compared with the sham-operated group, the MWT and TWL values were significantly decreased in the NP group. However, MWT and TWL values were notably elevated in both BTX-A-10 group and BTX-A-20 group when compared with the NP group, indicating that BTX-A administration elevated rat pain threshold post-CCI (Fig. 1a, b). To explore whether BTX-A administration alters microglia polarization, we used antibodies specific to Iba-1 (a marker of microglial activation) and CD68 (a M1 marker of microglia) to identify M1, and antibodies specific to Iba-1 and CD206 (a M2 marker of microglia) to identify M2. As shown in Fig. 2a-c, a significant increase in the number of Iba-1 + CD68 + cells (M1) and a slight increase in the number of Iba-1 + CD206 + cells (M2) were observed in the NP group when compared with the Sham group, indicating microglial activation in the rat model of NP. Furthermore, BTX-A administration resulted in a considerable increase in the number of Iba-1 + CD206 + cells but a notable decrease in that of Iba-1 + CD68 + cells. Moreover, data from flow cytometry analysis further confirmed that BTX-A treatment promoted microglial polarization toward the M2 phenotype in NP rats (Fig. 3).

BTX-A promoted microglial M2 polarization in LPS-stimulated HAPI microglial cells
To corroborate the histological findings, HAPI cells were treated with BTX-A at different concentrations in the absence and presence of LPS, a known microglial activator. CCK-8 assay showed that both with or without LPS, BTX-A had no significant effect on cell viability at doses of 0.01 nM and 0.1 nM, whereas caused a significant decrease in cell viability at doses not less than 1 nM (Fig. 4a, b). Thus, 0.1 nM BTX-A was selected four subsequent experiments. We further verified whether BTX-A regulates microglial polarization. Fig. 3 Effect of BTX-A on M1 and M2 microglia in NP rats. The proportions of both M1 and M2 microglia in L4-L6 spinal cord segments were examined by flow cytometry. **P < 0.01, vs. Sham; # P < 0.05, ## P < 0.01, vs. NP LPS treatment resulted in a marked increase in expression of iNOS (a M1 marker of microglia) and M1-related cytokines (TNF-α and IL-6). Importantly, the effect of LPS was effectively attenuated by BTX-A, indicating BTX-A inhibited the LPS-induced microglial M1 polarization (Figs. 4c, e, f ). Furthermore, expression of Arg-1 (a M2 marker of microglia) and M2-associated IL-10 were greatly upregulated by BTX-A treatment, further suggesting that BTX-A promoted microglial polarization to the M2 phenotype (Fig. 4c, d).

Discussion
In the present study, we found that the MWT and TWL values in NP model rats were lower than those in the sham-operated rats, whereas BTX-A administration elevated the MWT and TWL in NP model rats. These findings confirmed the relief of NP by BTX-A. The analgesic effect of BTX-A in NP in the current study was supported by previous studies [16,17].
Activated microglia in the spinal cord in response to nerve injury can enhance the synaptic transmission of spinal dorsal horn neurons through various cell surface receptors and pro-inflammatory factors, and thus participate in the occurrence and development of NP [6]. In the present study, we observed a significant increase in Iba-1 + cells (microglial marker) in rat L4-L6 spinal cord segments following CCI, indicating microglial activation following NP. Our results were consistent with previous studies [23,24].
Then we explored whether the analgesic effect of BTX-A was associated with its regulation of microglial polarization. Evidence suggests that microglial M1 polarization and the subsequent neuroinflammation play a vital role in the development of NP. Piotrowska et al. [11] reported that chronic intrathecal administration of maraviroc diminished NP symptoms and nociceptive threshold as well as decreased levels of pro-inflammatory cytokines in Wistar rats post-CCI. Convincing researches have indicated that switching microglial polarization from the pro-inflammatory M1 phenotype to the anti-inflammatory M2 phenotype represents a novel therapeutic strategy for NP [10][11][12][13]. For example, Willemen et al. [13] proposed that persistent hyperalgesia in GRK2 (G protein-coupled receptor kinase)-deficient mice was associated with an increased ratio of M1/M2 type markers in spinal cord microglia/macrophages. Intrathecal miR-124 treatment could alleviate NP by restoring spinal M1/M2 markers, indicating that the phenotypic transition of M1/M2 microglia is critical for the relief of NP. To our knowledge, this study is the first to demonstrate that BTX-A administration induces microglial polarization toward the M2 phenotype. Further in vitro assays in this investigation confirmed that BTX-A treatment promoted microglial M2 polarization in LPS-stimulated rat HAPI microglial cells. P2X7 receptor is highly expressed in the DRG of NP rats and is implicated in the pain transmission and the occurrence of NP [19,25]. Silencing of long non-coding RNA NON-RATT021972 has been shown to reduce mechanical and thermal hyperalgesia in diabetic rats through decreasing the P2X7 receptor in DRG [19]. It was reported that P2X7 receptor plays a role in regulating microglial M1/ M2 polarization. Inhibition of P2X7 receptor could inhibit microglial M1 polarization in ischaemic conditions [22]. To our knowledge, the current study provided the first evidence that BTX-A promotes microglial M2 polarization by decreasing P2X7 expression in LPS-stimulated HAPI cells.

Conclusions
In conclusion, this study demonstrates that BTX-A promotes microglial M2 polarization and suppresses CCIinduced NP through suppressing the P2X7 receptor.
These findings provide new insights into the mechanism of BTX-A in relieving NP.