High-intensity swimming alleviates nociception and neuroinflammation in a mouse model of chronic post-ischemia pain by activating the resolvin E1-chemerin receptor 23 axis in the spinal cord

Abstract

Physical exercise effectively alleviates chronic pain associated with complex regional pain syndrome type-I. However, the mechanism of exercise-induced analgesia has not been clarified. Recent studies have shown that the specialized pro-resolving lipid mediator resolvin E1 promotes relief of pathologic pain by binding to chemerin receptor 23 in the nervous system. However, whether the resolvin E1-chemerin receptor 23 axis is involved in exercise-induced analgesia in complex regional pain syndrome type-I has not been demonstrated. In the present study, a mouse model of chronic post-ischemia pain was established to mimic complex regional pain syndrome type-I and subjected to an intervention involving swimming at different intensities. Chronic pain was reduced only in mice that engaged in high-intensity swimming. The resolvin E1-chemerin receptor 23 axis was clearly downregulated in the spinal cord of mice with chronic pain, while high-intensity swimming restored expression of resolvin E1 and chemerin receptor 23. Finally, shRNA-mediated silencing of chemerin receptor 23 in the spinal cord reversed the analgesic effect of high-intensity swimming exercise on chronic post-ischemic pain and the anti-inflammatory polarization of microglia in the dorsal horn of the spinal cord. These findings suggest that high-intensity swimming can decrease chronic pain via the endogenous resolvin E1-chemerin receptor 23 axis in the spinal cord.

Introduction

Chronic pain is a significant and costly problem worldwide. Chronic complex regional pain syndrome (CRPS) is a debilitating and chronic condition affecting the limbs that may be induced by clinical surgery or clinical trauma (Goh et al., 2017; Shim et al., 2019). CRPS is usually divided into type 1 (CRPS-1) and type 2 (CRPS-2) based on whether nerve injury is present. CRPS-1, previously known as reflex sympathetic dystrophy, is more common than CRPS-2 (Shim et al., 2019). The ambiguity and complexity of CRPS-1 pathogenesis largely hinder its clinical treatment. Although nonsteroidal anti-inflammatory drugs are often used in clinics to treat patients with CRPS-1, the analgesic effect is lower than ideal, and long-term use of nonsteroidal anti-inflammatory drugs can have substantial side effects (Harden et al., 2022). Therefore, there is an urgent need to explore new analgesic strategies that are more effective and have minimal side effects.

Nonpharmacological therapies such as physical exercise are commonly and highly recommended in the clinical management of a variety of types of chronic pain, such as fibromyalgia, osteoarthritis, and low back pain (Busch et al., 2011; Brosseau et al., 2017; Qaseem et al., 2017). Exercise plays important roles in reducing pain, promoting functional recovery, and normalizing proprioception in patients with CRPS-1 (Staal et al., 2019; Harden et al., 2022). Exercise is also an effective treatment for childhood CRPS-1 to decrease the incidence of long-term symptoms or dysfunction (Sherry et al., 1999). Furthermore, aerobic exercise, like walking and swimming, is recommended by clinical guidelines for the relief of pain associated with CRPS-1 (Stanton-Hicks et al., 1998; Harden et al., 2022). In addition, many laboratory studies have reported the analgesic effect of swimming, treadmill use, and voluntary wheel running in animal models of CRPS-1 (Martins et al., 2013; Belmonte et al., 2018; Shi et al., 2018). However, research on the cellular and molecular mechanisms of exercise-induced analgesia (EIA) in subjects with CRPS-1 is still lacking, which substantially limits the application of this type of therapy to the rehabilitation of patients with CRPS-1.

Specialized pro-resolving lipid mediators (SPMs), endogenous anti-inflammatory lipid mediators, have been recently shown to be involved in inflammation resolution and pain relief (Xu et al., 2010; Fattori et al., 2019; Chiang and Serhan, 2020). Resolvin E1 (RvE1), the first identified resolvin derived from omega-3 polyunsaturated fatty acid, has been reported to relieve chronic pain by acting on its receptor chemerin receptor 23 (ChemR23) (Xu et al., 2010; Tao et al., 2020). Interestingly, exercise has been reported to increase endogenous SPM levels in urine or peritoneal lavage exudates from mice (Gangemi et al., 2003; Zheng et al., 2019). However, whether the endogenous spinal RvE1-ChemR23 axis plays an important role in EIA has not been demonstrated. In the present study, we established a mouse model of chronic post-ischemia pain (CPIP) to mimic CRPS-1 and subjected these mice to a swimming intervention to investigate the mechanism of EIA.

Methods

Animals

Male adult C57BL/6J mice weighing 25–30 g (8 weeks old) were ordered from Shanghai SLAC Laboratory Animal Company (Shanghai, China, license No. SCXK (Hu) 20220004) and were housed under standard environmental conditions in an animal facility with a 12/12-hour light/dark cycle, 24 ± 2°C room temperature, and 60–80% relative humidity. Food and water were provided ad libitum. All experiments were performed in accordance with Tongji University guidelines for animals use in biological studies. The protocol was approved by the Animal Study Committee of the Tongji University School of Medicine, Shanghai, China on February 10, 2022 (permission number TJAA08522101). All mice were randomly assigned to the control or experimental group (Figure 1).

Figure 1.

Study flow chart.

ChemR23: Chemerin receptor 23; CPIP: chronic postischemia pain; ELISA: enzyme-linked immunosorbent assay; HIS: high-intensity swimming; IF: immunofluerescence; LIS: low-intensity swimming; RT-PCR: reverse transcription-polymerase chain reaction; RvE1: Resolvin E1.

Mouse model of CPIP and behavioral tests

Ischemia/reperfusion (I/R) was induced by placing a tight-fitting O-ring around the ankle joint to establish a mouse model of CPIP that resembles the occurrence and development of CRPS-1 in patients (Chen et al., 2020a; De Logu et al., 2020). The mice were anesthetized with 1% sodium pentobarbital (Sigma-Aldrich, St. Louis, MO, USA) injected intraperitoneally at 50 mg/kg body weight, and then an elastic nitrile O-ring (Susong Mingwei Seal Technology Co., Ltd., Anqing, Anhui, China) with a 1.2-mm internal diameter was placed tightly around the right hindlimb just proximal to the ankle joint to completely block the blood flow for 3 hours. Finally, the O-ring was cut to restore blood flow to the limb. For the sham group, a cut O-ring that did not block blood flow was placed around the ankle.

After establishment of the CPIP model, cardinal features of inflammation, namely edema, redness, heat, and pain, were measured in the hindpaws. Digital calipers (Baigong, Shanghai, China) were used to measure paw thickness, and a thermometer (MicoCare, Shenzhen, Guangdong, China) was used to measure paw temperature. Paw skin redness was assessed visually and compared to the ipsilateral and contralateral paws in CPIP and control mice.

All behavioral experiments were conducted in the daytime, from 9:00 to 17:00. For the punctate mechanical hyperalgesia test, mice were placed in clear Plexiglas boxes (9 cm × 7 cm × 11 cm; Shengzhuang Industry and Trade Co., Ltd. Jinhua, Zhejiang, China) on an elevated wire mesh platform for more than 30 minutes for habituation before the test. Then, a series of von Frey filaments (0.16, 0.4, 0.6, 1, 1.4, and 2 g, Danmic Global, LLC, San Jose, CA, USA) was applied vertically to stimulate the hindpaw. The measurement started with a 0.16 g filament, and the paw was stimulated with each filament five times, with an interval of at least 15 seconds between each application, using the “up and down” method (Chaplan et al., 1994).

Heat hyperalgesia was assessed using the Hargreaves test (Model 400, IITC Life Science, Woodland Hills, CA, USA). The ventral midplantar surface of the mouse hindpaw was exposed to a radiant heat source from a thermal stimulator (Model 400, IITC Life Science) through a glass floor. The latency of an abrupt withdrawal of the paw and licking and vigorous shaking in response to stimulation was recorded. Latency was measured five times at 10-minute intervals for each mouse. The average value of the five responses was recorded as the paw withdrawal latency. A cutoff time of 20 seconds was set to prevent tissue damage.

Mouse model of spared nerve injury

As described previously (Decosterd and Woolf, 2000), the mice were anesthetized with 1% sodium pentobarbital intraperitoneally at 50 mg/kg body weight. The sciatic nerve near the thigh was separated to expose its branches tibial nerve, common peroneal nerve and sural nerve. The tibial nerve and the common peroneal nerve were ligated with 4-0 silk (Shanghai Pudong Jinhuan Medical Products Co., Ltd. Shanghai, China) and cut off. The sural nerve was preserved during the surgery.

HIS and LIS exercise protocols

The swimming protocol was described earlier (Mazzardo-Martins et al., 2010; Martins et al., 2013). Mice were placed in a plastic box (17 cm × 10 cm × 25 cm) filled with water at a depth of 20 cm at 37°C. On days 5 and 6 after CPIP model establishment, the mice were habituated to the swimming environment twice a day for 2 minutes at 2-hour intervals. Then, the mice began the HIS or LIS protocol. The maximal lactate steady state level, which represents the upper boundary of constant load endurance training, has been used to guide interval training sessions in different endurance sports (Faude et al., 2009). HIS is an exercise protocol with an intensity level above the maximal lactate steady state and consists of a 30-minute swimming exercise that has been shown to constitute high-intensity exercise (Mazzardo-Martins et al., 2010; Martins et al., 2013). As in a previous study (Mazzardo-Martins et al., 2010), the HIS protocol consisted of swimming sessions of 30 minutes each day for 5 consecutive days. For LIS, the mice were allowed to swim for 5, 10, and 15 minutes for 3 days. Mice that did not engage in any exercise were used as the control group. When observing the time course of the analgesic effects of exercise, nociceptive behaviors before exercise were used as a control.

Open field test

The open field test was performed as described previously (Tajerian et al., 2014). The experiment was performed in a quiet environment. Mice were placed in the center of a 40 cm × 40 cm × 40 cm white box and allowed to explore freely for 10 minutes. All movements were recorded with a video camera (JLBehv CCD, Shanghai, China), and the distance traveled, velocity, and time spent in the center of the open field were determined using an animal behavior analysis system (Shanghai Jiliang Software Technology Co., Ltd., Shanghai, China). The inner and bottom surfaces of the box were cleaned with 75% alcohol after each test.

Tail suspension test

The tail suspension test was performed as described previously (Zhou et al., 2019). The experiment was performed in a quiet environment. Each mouse was fixed by the tail tip (the distal third of the tail) with paper tape to the inside of the suspension box. The head of the mouse was facing the camera approximately 10 cm from the bottom of the box. A video analysis system (Shanghai Jiliang Software Technology Co., Ltd., Shanghai, China) was used to record each for a total of 6 minutes after suspension, and the immobility time in the last 4 minutes was analyzed.

Drug administration

RvE1 was purchased from Cayman Chemical (Cat# 10007848). Stock solutions of RvE1 were prepared in ethanol (50 μg in 500 μL of 100% ethanol) and stored at –80°C. Clonidine hydrochloride was purchased from Sigma-Aldrich (Cat# C7897). Morphine hydrochloride and clonidine hydrochloride were stored at 4°C. RvE1 (10 or 20 ng) and morphine (500 ng) were diluted with 0.9% normal saline. Clonidine (10 µg) was dissolved in 0.9% normal saline. We used 1% ethanol as a vehicle for RvE1 and normal saline as a vehicle for morphine and clonidine. These drugs were administered intrathecally to mice.

Conditional place preference test

The conditional place preference (CPP) test was performed as described previously (Donnelly et al., 2021). The CPP apparatus consisted of two chambers with different visual and tactile cues. The CPP test comprised three distinct phases: pre-pairing, pairing, and post-pairing. During the pre-pairing test, each mouse was allowed to explore the two chambers freely for 15 minutes. Pre-pairing recordings revealed that mice had a slight but not significant preference for one of the chambers. The day after the pre-pairing test, the mice were subjected to a pairing test. After injection of vehicle (normal saline or 1% ethanol, intrathecally 10 μL), the mice were placed into the preferred chamber for 50 minutes. Four hours later, the mice were injected intrathecally with vehicle (10 µL), morphine (500 ng), clonidine (10 μg), or RvE1 (10 or 20 ng) and were immediately placed into the nonpreferred chamber for 50 minutes. The day after pairing, the mice were subjected to a post-pairing test, in which each mouse was allowed to explore the chambers freely for 15 minutes. The pre-pairing time spent in morphine-, clonidine-, or RvE1-paired chambers was subtracted from that of the post-pairing time to calculate the CPP score as the preference criterion (Donnelly et al., 2021). The movements of all mice were recorded by a video camera (JLBehv CCD) and assessed using an animal behavior analysis system (Shanghai Jiliang Software Technology Co., Ltd.).

Western blot assay

After the animals were anesthetized with 1% sodium pentobarbital intraperitoneally at 50 mg/kg body weight and transcardially perfused with phosphate buffer solution, the L4–L6 spinal cord was immediately removed and stored at –80°C. Frozen samples were directly homogenized in a buffer containing protease inhibitors and phosphatase inhibitors. After centrifugation at 21,000 × g for 40 minutes, the supernatant was extracted to detect the protein concentration using a bicin-choninic acid protein assay kit (Epizyme, Shanghai, China). Next, loading buffer was added to the protein samples, which were then boiled for 7 minutes at 100°C. Total protein (40 μg) was loaded onto a 10% sodium dodecyl sulfate-polyacrylamide gel, separated, and transferred onto polyvinylidene difluoride membranes. Membranes were blocked with 5% nonfat milk for 2 hours and incubated with primary antibodies (mouse monoclonal anti-ChemR23, 1:1000, Santa Cruz Biotechnology, Santa Cruz, CA, USA, Cat# sc-398769; rabbit polyclonal anti-glyceraldehyde-3-phosphate dehydrogenase (GAPDH), 1:5000, Proteintech, Chicago, IL, USA, Cat# 10494-1-AP, RRID: AB_2263076) overnight at 4°C. Membranes were then washed with Tris-buffered saline containing 0.1% Tween and incubated with horseradish peroxidase-conjugated secondary antibodies (goat anti-mouse IgG, 1:3000, Cat# LF101, Epizyme, Shanghai, China; goat anti-rabbit IgG, 1:3000, Cat# LF102, Epizyme, Shanghai, China) for 1 hour at room temperature. The quantitative analysis was performed using ImageJ software v1.8.0 (National Institutes of Health, Bethesda, MD, USA) (Schneider et al., 2012). The expression level of the target proteins was normalized to GAPDH.

Immunofluorescence staining

After anesthetization with sodium pentobarbital (50 mg/kg), the mice were perfused through the ascending aorta with phosphate buffer solution, followed by 4% paraformaldehyde in 0.1 M phosphate buffer (pH 7.4). The L4–L6 spinal cord and L4–L6 dorsal root ganglia (DRG) were collected in 4% paraformaldehyde, fixed, and dehydrated with 20% sucrose solutions. The spinal tissues and DRG were then cut into thin sections (30 and 14 μm) with a cryostat (CM1950, Leica, Wetzlar, Germany). The sections were incubated with the following primary antibodies at 4°C overnight: rabbit polyclonal anti-activating transcription factor 3 (ATF3; 1:1000, Sigma-Aldrich, Cat# HPA001562, RRID: AB_1078233), goat polyclonal anti-ionized calcium binding adaptor molecule 1 (Iba1; 1:1000; Abcam, Waltham, MA, USA, Cat# ab5076, RRID: AB_2224402), rabbit polyclonal anti-ChemR23 (1:1000; Invitrogen, Carlsbad, CA, USA, Cat# PA5-85322, RRID: AB_2792464), mouse monoclonal anti-neuronal nuclei (NeuN; 1:1000; Millipore, Birrika, MA, USA, Cat# MAB377, RRID: AB_2298772), and mouse monoclonal anti-glial fibrillary acidic protein (GFAP; 1:1000; Sigma-Aldrich, Cat# G6171, RRID: AB_1840893). The sections were then incubated at room temperature for 2 hours in the dark with corresponding fluorescent polyclonal Alexa Fluor 488-conjugated donkey anti-rabbit secondary antibody (1:1000, Invitrogen, Cat# A21206, RRID: AB_253579), Alexa Fluor 488-conjugated donkey anti-mouse secondary antibody (1:1000, Invitrogen, A21202, RRID: AB_141607), Alexa Fluor 488-conjugated donkey anti-goat secondary antibody (1:1000, Invitrogen, SA5-10086, RRID: AB_2556666), or Alexa Fluor 555-conjugated donkey anti-rabbit secondary antibody (1:1000, Invitrogen, Cat# A31572, RRID: AB_162543). Images of the DRG and spinal cord slice sections were captured under a fluorescence microscope (Axio Imager. M2; Carl Zeiss Jena, Oberkochen, Germany) with a 10× or a 20× objective. ImageJ was used to measure Iba-1 fluorescence intensity. Adobe Photoshop CC 2018 (Adobe, San Jose, CA, USA) was used to measure ATF3 and Iba-1 expression and colocalization of ChemR23 with NeuN, Iba-1, and GFAP.

Collection of cerebrospinal fluid and serum

Sodium pentobarbital (50 mg/kg body weight) was injected intraperitoneally to anesthetize the mice. The muscles from the occipital bone to the atlas were separated with tweezers to expose the white dura mater. A glass electrode was inserted into the fourth ventricle to absorb the cerebrospinal fluid (CSF). Approximately 20 µL CSF were collected from each mouse, and the CSF from six mice was mixed to prepare one sample that met the sample volume requirements of the enzyme-linked immunosorbent assay (ELISA) kit. For serum collection, the mice were anesthetized with 1% sodium pentobarbital through an intraperitoneal injection, blood was collected at the cardiac apex, and the serum was separated by centrifugation. CSF and serum were used to detect the RvE1 levels using a mouse RvE1 ELISA Kit and serum was also used to detect the corticosterone levels using a mouse corticosterone ELISA Kit.

ELISA

A mouse RvE1 ELISA kit (detection sensitivity: 1 pg/mL) was purchased from MyBiosource (San Diego, CA, USA, Cat# MBS755469). The spinal cord samples were directly homogenized in a protein lysis buffer containing protease inhibitors and phosphatase inhibitors. After centrifugation at 21,000 × g for 40 min, the supernatant was extracted, and the protein concentration was determined using the bicin-choninic acid assay. Then, 100 µL samples of the supernatant were added to a 96-well plate to detect RvE1 levels. A mouse corticosterone ELISA Kit (detection sensitivity: 4.98 ng/mL) was purchased from Elabscience Biotechnology Co., Ltd. (Cat# E-OSEL-M0001, Wuhan, Hubei, China), and 50-µL serum samples were added to a 96-well plate to detect corticosterone levels. The RvE1 and corticosterone ELISAs were performed according to the manufacturers’ protocols. The optical densities of RvE1 or corticosterone in the samples were measured at a wavelength of 450 nm using a microplate reader (SpectraMax M5, Molecular Devices, Sunnyvale, CA, USA). The RvE1 or corticosterone concentration was determined based on a standard curve generated using an appropriate set of internal standards.

Quantitative reverse transcription-polymerase chain reaction

The lumbar enlargement of the spinal cord was collected and ground. Total RNA was extracted, and complementary DNA was synthesized using a reverse transcription kit purchased from Tiangen Biochemical Technology (Cat# KR106-02, Beijing, China). Reverse transcription-polymerase chain reaction was performed using a Roche real-time fluorescence quantitative PCR instrument (Roche, Basel, Switzerland). The mRNA levels of ChemR23, interleukin-1 beta (IL-1β), IL-6, tumor necrosis factor-alpha (TNF-α), IL-4, IL-10, G protein-coupled receptor 18 (GPR18), GPR37, lipoxin A4 receptor/formyl peptide receptor 2 (ALX/FPR2), and GPR32 were determined by real-time PCR. The relative mRNA expression levels of the target genes were normalized to GAPDH and determined using the 2^–ΔΔCt method (Schmittgen and Livak, 2008). The primer sequences for ChemR23, IL-4, IL-10, TNF-α, IL-1β, IL-6, GPR18, GPR37, FPR2, GPR32, and GAPDH are listed in Additional Table 1. The primers were synthesized by Shanghai Huajin Biotechnology Co., Ltd. (Shanghai, China).

Additional Table 1.

Primers sequence for quantitative reverse transcription-polymerase chain reaction

Primer	Forward (5’-3’)	Reverse (5’-3’)
ChemR23	ATG GAG TAC GAC GCT TAC AAC G	GGT GGC GAT GAC AAT CAC CA
IL-1β	GCC CAT CCT CTG TGA CTC AT	CTC ATA TGG GTC CGA CAG CA
IL-6	TCT ATA CCA CTT CAC AAG TCG GA	GAA TTG CCA TTG CAC AAC TCT
		TT
TNF-α	AAT GGC CTC CCT CTC ATC AG	AGC CTT GTC CCT TGA AGA GA
IL-4	CCA TGA ATG AGT CCA AGT CC	TGA TGC TCT TTA GGC TTT CC
IL-10	GGG AAG AGA AAC CAG GGA GA	GGG GAT GAC AGT AGG GGA AC
GPR18	GTG GTG TTT TAC CCA AGC CTC	TGG TCA GGG TCA TTA CCC AGA
GPR37	ACC GGA CAC AAT CTA TGT TTT GG	TCT TCC GAG CAG TCA CTA GAG
FPR2	CCG TCC TTT ACG AGT CCT TAC A	CAG GAG GTG AAG TAG AAC TGG
		T
GPR32	AGA CAG CAC TAG CCA TCA CC	GAC TTC ACC AAG AGC ACC AC
GAPDH	CCA ATG TGT CCG TCG TGG ATC	GTT GAA GTC GCA GGA GAC AAC

ChemR23: Chemerin receptor 23; FPR2: formyl peptide receptor 2; GAPDH: glyceraldehyde-3-phosphate dehydrogenase; GPR: G protein-coupled receptor; IL: interleukin; TNF-α: tumor necrosis factor-α.

Virus injection

Mice were anesthetized with sodium pentobarbital (50 mg/kg body weight). An incision was made over the lumbar enlargement of the spinal cord, and the mice were then fixed on a stereotaxic apparatus (Reward Life Technology, Shenzhen, China). Next, the fascia and muscles covering the vertebrae were separated with tweezers to expose the spinal cord. One-microliter pulled glass micropipettes were connected to a microinjection system (World Precision Instruments, Sarasota, FL, USA) that was attached to the stereotaxic apparatus. After drawing in 300 nL of virus, the glass micropipettes were inserted into the L4–L6 spinal cord (300 μm lateral to the dorsal artery, 300 μm deep) vertically, and the virus was injected at a rate of 10 nL/min. The micropipette was left in place for 10 minutes and then slowly withdrawn to prevent backflow. Each mouse was injected with the virus bilaterally, at two sites on each side. The ChemR23 shRNA-expressing lentivirus and its nontargeting control shRNA-expressing lentivirus were synthesized by Shanghai Genechem Co., Ltd. (Shanghai, China) according to published sequences (Xu et al., 2010). Mice injected with nontargeting control shRNA-expressing lentivirus and naive mice were used as controls.

In vivo multichannel electrophysiology

Mice were fixed on a stereotaxic apparatus after being anesthetized with 10% urethane (1.5 g/kg body weight, Cat# 94300, Sigma-Aldrich). A heating plate was placed below the apparatus to maintain the animal’s body temperature at approximately 37°C. A skull nail was drilled into the posterior fontanelle of the skull to connect the ground wire of the homemade multichannel electrode. The skin was incised longitudinally through the midline of the back, and a spinal cord clamp was used to fix lumbar vertebra and expose the L4–L6 spinal cord. The electrode connected to the holding rod was inserted into the spinal cord ipsilateral to the ischemic side using a stereotactic apparatus (300 μm lateral to the dorsal artery; 200–500 μm deep). The recording was started and continued until stable neuron firing was observed. Electrode insertion was stopped when the spinal dorsal horn neurons were observed to respond to innocuous stimuli (brush) and nocuous stimuli (pinch) on the plantar surface of the hindpaw ipsilateral to the ischemic side. A multichannel data acquisition system (Zeus, Bio-Signal Technologies, McKinney, TX, USA) was employed to collect data at a sampling frequency of 30 kHz. The raw data were preprocessed using Offline Sorter V2.8 (Plexon, Dallas, TX, USA). Individual neurons were isolated using K-means and valley-seeking algorithms following principal component analysis. NeuroExplorer V5.2 (Plexon) was employed to further analyze well-isolated neurons.

Statistical analysis

Sample size selection was guided by prior experience. All experiments were performed in a double-blind manner. No mice were excluded. The electrophysiological data are presented as the median ± interquartile ranges (IQRs) and were analyzed using Mann-Whitney U test or Wilcoxon signed-rank test. Other data are presented as the mean ± standard error of the mean (SEM) and were analyzed using two-tailed unpaired Student’s t-test or one-way or two-way analysis of variance followed by Bonferroni post hoc test. The criterion for statistical significance was P < 0.05. GraphPad Prism version 7.0.4 for Windows (GraphPad Software, San Diego, CA, USA, www.graphpad.com) was used to perform the statistical analyses.

Results

CPIP mice exhibit cardinal features of inflammation, but no mood disruption or nerve injury

We first established the CPIP mouse model by inducing prolonged hindpaw ischemia-reperfusion to mimic clinical CRPS-1 (Chen et al., 2020a; De Logu et al., 2020). The CPIP mice exhibited cardinal features of inflammation, including redness, swelling, heat, and nociceptive behavior. The ipsilateral hindpaws of CPIP mice showed transient but obvious redness, edema, and an increased temperature 30 minutes after reperfusion, and these effects lasted for 1 day (Figure 2A–C). Furthermore, the CPIP mice showed clear, persistent heat hyperalgesia and punctate mechanical hyperalgesia in the bilateral hindpaws that was maintained throughout the 4-week experimental period (Figure 2D–G). We also performed a CPP assay, which is widely used to measure ongoing or spontaneous pain (Tajerian et al., 2015; Donnelly et al., 2021), to detect nonreflexive nociception in the CPIP mice. A single injection of morphine or clonidine induced obvious CPP in the CPIP mice (Figure 2H–K), suggesting the existence of both reflexive and nonreflexive pain in CPIP mice.

Figure 2.

Cardinal features of inflammation in CPIP mice.

(A) The hindpaws showed obvious but transient swelling at 30 minutes and 1 day after establishment of the CPIP model. The arrow indicates the ipsilateral side. (B) Statistical analysis showing that the ipsilateral paw thickness was transiently increased after CPIP establishment (n = 8 mice/group). **P < 0.01 and ****P < 0.0001, vs. control ipsilateral hindpaw. (C) The temperature of the ipsilateral hindlimbs was transiently increased after CPIP establishment (n = 6 mice/group). **P < 0.01, vs. control ipsilateral hindpaw. (D–G) Punctate mechanical hyperalgesia and heat hyperalgesia were observed in the bilateral hindpaws of CPIP mice (n = 7 mice/group). *P < 0.05, **P < 0.01, ***P < 0.001 and ****P < 0.0001, vs. control group. (H) Schematic of the CPP assay. (I) Schematic of the CPP protocol for CPIP mice. (J) Representative CPP traces from mice injected with vehicle, morphine, or clonidine. CPIP mouse activity was clearly increased in morphine- and clonidine-paired chambers compared with vehicle-paired chambers. (K) CPP results in CPIP mice treated with morphine or clonidine (n = 5 mice/group). *P < 0.05 and **P < 0.01, vs. vehicle group. (L–N) The distance traveled (L) and velocity of mice (M) in the open field test were reduced at the late stage of CPIP. The time spent (N) in the center in the open field did not change in CPIP mice (n = 6 mice for Control group; n = 5 mice for CPIP group). **P < 0.01, vs. control group. (O) The immobile time during the tail suspension did not change in CPIP mice (n = 6 mice for Control group; n = 7 mice for CPIP group). (P) The expression of ATF3 (green, Alexa Fluor 488) was increased in the ipsilateral DRG neurons of SNI mice on days 3, 7, and 14, but not in control and CPIP mice (n = 3 mice/group). Scale bar: 100 μm. All data are presented as the mean ± SEM, and were analyzed by two-way analysis of variance followed by Bonferroni post hoc test (B–G and L–O) or one-way analysis of variance followed by Bonferroni post hoc test (K). Detailed statistical information is shown in Additional Table 2 ^{(125.3KB, pdf)} and Additional file 1 ^{(136.8KB, pdf)} . ATF3: Activating transcription factor 3; BL: baseline; Control: the ankles of mice were surrounded by a cut-O-ring; CPIP: mice were submitted to chronic postischemia pain; CPP: conditional place preference; Vehicle: mice were submitted to chronic postischemia pain and injected with normal saline; Morphine: mice were submitted to chronic postischemia pain and injected with morphine; Clonidine: mice were submitted to chronic postischemia pain and injected with clonidine; DRG: dorsal root ganglion; SNI: mice were submitted to spared nerve injury.

Next, an open field test and tail suspension test were performed to observe motor function and mood, respectively. We found that the distance traveled and velocity of movement in the open field test were slightly decreased at the late stage (day 12) but not at the early stage (day 3) in CPIP mice compared with sham control mice (Figure 2L and M). However, no differences between CPIP and control mice in terms of the time spent in the center of the open field or in the immobility time during the tail suspension test were observed, suggesting that CPIP mice did not experience anxiety or depression (Figure 2N and O). Furthermore, expression of ATF3, an indicator of peripheral neuron injury (Wang et al., 2021), did not differ between CPIP and control mice (Figure 2P). In contrast, ATF3 expression was increased in the DRG of a mouse model of spared nerve injury (SNI) (Figure 2P). Thus, CPIP mice exhibited obvious peripheral inflammation but no nerve injury, consistent with patients with CRPS-1 (Shim et al., 2019).

Neuroinflammation and central sensitization in the spinal cord of CPIP mice

Next, we examined microglia activation in the spinal cord using the microglial marker Iba1 (Chen et al., 2020b) to determine whether CPIP causes neuroinflammation. Iba1 fluorescence intensity and the number of Iba1-positive cells were significantly increased in the spinal cord dorsal horn of CPIP mice compared with control mice on days 3, 7, and 14 (Figure 3A–C). Moreover, the mRNA levels of the proinflammatory cytokines TNF-α, IL-1β, and IL-6 were higher in the spinal cord dorsal horn of CPIP mice at both the early stage (day 3) and late stage (day 12) compared with control mice (Figure 3D–G). In contrast, the mRNA level of IL-4, an anti-inflammatory cytokine (Kedong et al., 2020), was lower at both the early and late stages of CPIP, while the expression of another anti-inflammatory cytokine, IL-10 (Han et al., 2022), was lower at only the later stage of CPIP, compared with control mice (Figure 3D–G). Moreover, the spontaneous firing rate of spinal wide dynamic range (WDR) neurons in CPIP mice was higher than that in control mice (Figure 3H–J). These results suggest that CPIP mice exhibit neuroinflammation and central sensitization in the spinal cord, consistent with previous studies (Tang et al., 2018; Shim et al., 2019).

Figure 3.

Neuroinflammation and central sensitization in the spinal cord of CPIP mice.

(A) Representative images of Iba1 (green, Alexa Fluor 488) immunofluorescence staining in the bilateral spinal dorsal horn in CPIP and control mice. The Iba1 fluorescence signal was increased in CPIP mice compared with control mice. The righthand images (bar: 100 μm) are high magnifications of the boxes in the lefthand images (bar: 200 μm). (B, C) Iba1 fluorescence intensity and numbers of Iba1-positive microglia in the bilateral spinal dorsal horn on days 3, 7, and 14 in CPIP and control mice (n = 3 mice/group). **P < 0.01, ***P < 0.001 and ****P < 0.0001, vs. control group. (D, E) mRNA expression levels of the proinflammatory cytokines IL-1β, IL-6, and TNF-α were increased, and mRNA expression of the anti-inflammatory cytokine IL-4 was decreased, in the bilateral spinal lumbar enlargement on day 3 after CPIP establishment (n = 6 mice/group). *P < 0.05, **P < 0.01, ***P < 0.001 and ****P < 0.0001, vs. control group. (F, G) mRNA expression levels of the proinflammatory cytokines IL-1β, IL-6, and TNF-α were increased, and the mRNA expression levels of the anti-inflammatory cytokines IL-4 and IL-10 were decreased, in the bilateral spinal lumbar enlargement on day 12 after CPIP establishment (n = 6 mice/group). *P < 0.05, **P < 0.01, ***P < 0.001 and ****P < 0.0001, vs. control group. (H–J) The firing rate of spinal WDR neurons in CPIP mice on days 3 and 12 was significantly higher than that in control mice. Representative trace of a spinal cord WDR neuron discharge (H); representative recordings of spontaneous spikes from spinal WDR neurons on days 3 and 12 in CPIP mice and control mice (I). Statistical analysis of WDR neuron spontaneous firing rate on days 3 and 12 in CPIP mice and control mice (J) (n = 5 mice/group (day 3); n = 6 mice/group (day 12), n = 18–22 cells). *P < 0.05 and ***P < 0.001, vs. control group. All data in B-G are presented as the mean ± SEM, and the data in J are presented as the median ± IQR, and were analyzed by two-way analysis of variance followed by Bonferroni post hoc test (B, C), two-tailed unpaired Student’s t-test (D–G), or Mann-Whitney U test (J). Detailed statistical information is shown in Additional Table 2 ^{(125.3KB, pdf)} and Additional file 1 ^{(136.8KB, pdf)} . Contra: Contralateral; Control: the ankles of mice were surrounded by a cut-O-ring; CPIP: mice were submitted to chronic postischemia pain; Iba1: ionized calcium binding adaptor molecule 1; IL: interleukin; IOD: integrated optical density; Ipsi: ipsilateral; TNF-α: tumor necrosis factor-α; WDR: wide dynamic range.

HIS, but not LIS, alleviates nociception, neuroinflammation, and central sensitization in CPIP mice

To test whether swimming alleviates nociceptive behavior in CPIP mice, we subjected mice to HIS and LIS sessions (Figure 4A and B) and measured nociceptive behaviors 24 hours after the last session. We found that punctate mechanical hyperalgesia and heat hyperalgesia were significantly lower in the HIS group compared with the LIS group and the no exercise group in CPIP mice (Figure 4C–F), and the analgesic effect lasted for 44 hours (Additional Figure 1A ^{(371.4KB, tif)} and B ^{(371.4KB, tif)} ). Swimming is usually considered to cause stress in mice (Mazzardo-Martins et al., 2010), so we also assessed emotional reactions 24 hours after the last swimming session by performing open field and tail suspension behavioral tests to determine whether the swimming-induced analgesia was caused by stress or negative emotions. We found that the distance traveled, velocity, time spent in the center of the open field, and immobility time after tail suspension did not differ among the HIS or no exercise control mice or CPIP mice (Figure 4G–J). In addition, there was no difference in serum corticosterone content between the CPIP and control mice 24 hours after the last swimming session (Figure 4K). Thus, our results showed that HIS-induced analgesia in CPIP mice is not secondary to stress, depression, or anxiety. In addition, the CPIP and control mice swam for a comparable percentage of the HIS period (approximately 80%) (Figure 4L), suggesting that nociception was reduced by swimming.

Figure 4.

HIS, but not LIS, alleviates nociceptive behavior in CPIP mice, but there is no difference in serum corticosterone content among control and CPIP mice subjected to exercise or no exercise.

(A, B) HIS (A) and LIS (B) protocols. (C–F) Compared with the no exercise and LIS groups, bilateral mechanical and thermal pain thresholds were significantly increased in the HIS group in CPIP mice (n = 6 mice/group in the no exercise group; n = 6 mice/group in the LIS group; n = 6 control mice; n = 7 CPIP mice in the HIS group). **P < 0.01, vs. no exercise CPIP mice; #P < 0.05, ##P < 0.01 and ###P < 0.001, vs. LIS CPIP mice. (G–J) The distance traveled (G), velocity (H), time spent in the center of the open field test (I), and immobility time during the tail suspension test (J) did not differ among control and CPIP mice subjected to exercise or no exercise (n = 6 control mice and n = 5 CPIP mice with no exercise; n = 5 control mice and n = 7 CPIP mice with exercise (G–I); n = 7 mice/group (J)). (K) HIS does not affect serum corticosterone levels in control and CPIP mice 24 hours after exercise (n = 3 control mice and n = 6 CPIP mice). (L) The percentage of swimming time during the HIS protocol did not differ between control and CPIP mice (n = 5 mice/group). All data are presented as the mean ± SEM, and were analyzed by one-way analysis of variance followed by Bonferroni post hoc test (C–F), two-tailed unpaired Student’s t-test (G–K), or two-way analysis of variance followed by Bonferroni post hoc test (L). Detailed statistical information is shown in Additional Table 2 ^{(125.3KB, pdf)} and Additional file 1 ^{(136.8KB, pdf)} . Control: The ankles of mice were surrounded by a cut-O-ring; CPIP: mice were submitted to chronic postischemia pain; HIS or Exercised: control or CPIP mice were performed high-intensity swimming; LIS: control or CPIP mice were performed low-intensity swimming; Non-exercised: control or CPIP mice were not performed swimming.

Next, we asked whether HIS attenuated CPIP-induced neuroinflammation and central sensitization. To assess this, we examined spinal Iba1 immunofluorescence staining in CPIP mice that engaged in HIS or no exercise. Iba1 fluorescence intensity and the number of Iba1-positive cells were reduced in the HIS group compared with the no exercise group (Figure 5A–C). In addition, TNF-α, IL-1β, and IL-6 levels were lower, and IL-4 and IL-10 levels were higher, in the bilateral spinal cord in the HIS CPIP group compared with the no exercise CPIP group (Figure 5D and E). We also observed lower spontaneous firing of WDR neurons in the HIS CPIP group compared with the no exercise CPIP group (Figure 5F and G). Based on these results, HIS exercise attenuates nociception by ameliorating neuroinflammation and decreasing the hyperactivity of spinal dorsal horn neurons.

Figure 5.

HIS abolishes neuroinflammation and central sensitization in the spinal cord of CPIP mice.

(A) Representative images of Iba1 (green, Alexa Fluor 488) immunofluorescence staining in the bilateral spinal dorsal horn in CPIP mice subjected to exercise or no exercise. Iba1 fluorescence was reduced in CPIP mice subjected to exercise. The righthand images (bar: 100 μm) are high magnifications of the boxes in the lefthand images (bar: 200 μm). (B, C) Iba1 fluorescence intensity and numbers of Iba1-positive microglia (n = 3 mice/group). **P < 0.01, ***P < 0.001 and ****P < 0.0001, vs. no exercise mice in the CPIP group. (D, E) After HIS exercise, bilateral IL-1β, IL-6, and TNF-α cytokine levels were decreased, and IL-4 and IL-10 levels were increased in CPIP mice (n = 9 mice/group). ****P < 0.0001, vs. no exercise mice in the CPIP group. (F) Representative recordings of spontaneous spikes from spinal WDR neurons in CPIP mice subjected to exercise or no exercise. Spontaneous WDR neuron spikes were reduced in CPIP mice subjected to exercise. (G) The spontaneous firing rate of WDR neurons in CPIP mice subjected to exercise was significantly lower than that of CPIP mice that did not engage in exercise (n = 8 mice in the no exercise group and n = 12 mice in the exercise group, n = 29–46 cells). **P < 0.01, vs. no exercise CPIP mice. All data in B–E are presented as the mean ± SEM, and the data in G are presented as the median ± IQR, and were analyzed by two-tailed unpaired Student’s t-test (B-E) or Mann-Whitney U test (G). Detailed statistical information is shown in Additional Table 2 and Additional file 1 ^{(136.8KB, pdf)} . Contra: Contralateral; Exercised: mice were submitted to chronic postischemia pain and performed high-intensity swimming; Iba1: ionized calcium binding adaptor molecule 1; IL: interleukin; IOD: integrated optical density; Ipsi: ipsilateral; Non-exercised: mice were submitted to chronic postischemia pain and were not performed swimming; TNF-α: tumor necrosis factor-α; WDR: wide dynamic range.

HIS upregulates expression of RvE1-ChemR23 axis in the spinal cords of CPIP mice

To investigate whether HIS relieves nociception in CPIP mice by regulating endogenous RvE1, we assessed RvE1 levels in the spinal cord by ELISA and found that RvE1 levels were obviously reduced in the spinal cords of CPIP mice (Figure 6A). Then, we measured the mRNA expression of SPM receptors, including G protein-coupled receptor 18 (GPR18; receptor for RvD2) (Chiang et al., 2017), GPR37 (receptor for NPD1) (Bang et al., 2018), ChemR23 (receptor for RvE1 and RvE2) (Arita et al., 2005b; Deyama et al., 2018), lipoxin A4 receptor/formyl peptide receptor 2 (ALX/FPR2; receptor for lipoxin A4 (LXA4), RvD1 and RvD3) (Bisicchia et al., 2018; Lee et al., 2020), and GPR32 (receptor for RvD1, RvD3 and RvD5) (Chen et al., 2018; Arnardottir et al., 2021), in the spinal cords of CPIP mice and control mice. Spinal mRNA levels of GPR37, FPR2, and ChemR23 were lower in CPIP mice (Figure 6B). Among these mRNAs, ChemR23 exhibited the most robust reduction in CPIP mice (Figure 6B), and ChemR23 protein levels were also downregulated in CPIP mice (Figure 6C and D). HIS increased the RvE1 concentration in the spinal cord and CSF, but not in the serum, of CPIP mice (Figure 6E–G), implying that RvE1 functions in the central nervous system in CPIP. ChemR23 was also up-regulated at the mRNA and protein level after HIS (Figure 6H–J). In addition, ChemR23 co-localized with NeuN and Iba1 but not GFAP in the spinal cord of mice (Additional Figure 2A ^{(766.9KB, tif)} –L ^{(766.9KB, tif)} ). These results suggest that HIS upregulates expression of spinal RvE1-ChemR23 axis in CPIP mice.

Figure 6.

HIS upregulates the expression of RvE1-ChemR23 axis in the spinal cord and increases RvE1 levels in the CSF but not the serum of CPIP mice.

(A) RvE1 levels in the spinal lumbar enlargement were significantly decreased after CPIP model establishment (n = 6 mice/group). ****P < 0.0001, vs. control group. (B) The mRNA levels of GPR37, ChemR23, and FPR2 in the bilateral lumbar enlargement of the spinal cord were decreased significantly in CPIP mice compared with control mice (n = 12 mice/group). ***P < 0.001 and ****P < 0.0001, vs. control group. (C, D) ChemR23 protein levels in the lumbar enlargement of the spinal cord were significantly decreased in CPIP mice compared with control mice (n = 12 mice/group). **P < 0.01, vs. control group. (E) HIS increased RvE1 levels in spinal lumbar enlargement of CPIP mice (n = 12 mice/group). *P < 0.05, vs. no exercise mice in the CPIP group. (F, G) HIS exercise increased RvE1 levels in the CSF (F) but not in the serum (G) in CPIP mice (n = 3 samples/group, n = 18 mice/group (F); n = 7 mice/group (G)). *P < 0.05, vs. no exercise mice in the CPIP group. (H) ChemR23 and FPR2 mRNA levels in the lumbar enlargement of the spinal cord were significantly increased in CPIP mice after HIS exercise (n = 6 mice/group). *P < 0.05 and ***P < 0.001, vs. no exercise mice in the CPIP group. (I, J) The ChemR23 protein level in the lumbar enlargement of the spinal cord of CPIP mice was significantly increased after HIS (n = 9 mice/group). **P < 0.01, vs. no exercise mice in the CPIP group. All data are presented as the mean ± SEM, and were analyzed by two-tailed unpaired Student’s t-test (A, B, D–H, J). Detailed statistical information is shown in Additional Table 2 ^{(125.3KB, pdf)} and Additional file 1 ^{(136.8KB, pdf)} . ChemR23: Chemerin receptor 23; Control: the ankles of mice were surrounded by a cut-O-ring; CPIP Exercised: mice were submitted to chronic postischemia pain and performed high-intensity swimming; CPIP Non-exercised: mice were submitted to chronic postischemia pain and were not performed swimming; CPIP: mice were submitted to chronic postischemia pain; CSF: cerebrospinal fluid; FPR2: formyl peptide receptor 2; GAPDH: glyceraldehyde-3-phosphate dehydrogenase; GPR: G protein-coupled receptor; RvE1: Resolvin E1.

The RvE1-ChemR23 axis in the spinal cord mediates HIS-induced EIA and inhibits neuroinflammation

To explore the role of the RvE1-ChemR23 axis in HIS-induced analgesia, we next injected mice with an RNA interference virus to knock down ChemR23 expression in the spinal cord. As expected, ChemR23 protein expression was lower in the sh-ChemR23 group than in the control virus group, verifying that ChemR23 was downregulated in the spinal cord (Figure 7A and B). After successfully knocking down ChemR23 expression in the spinal cord, we found that the analgesic effect of HIS was substantially inhibited in the sh-ChemR23 group compared with the naïve and sh-Neg control groups (Figure 7C–F). In addition, the HIS-induced inhibition of microglial activation in CPIP mice and the decreased levels of IL-1β, IL-6, and TNF-α in the HIS CPIP group were all blocked in the sh-ChemR23 group compared with the sh-Neg control group (Figure 7G–K). Thus, the RvE1-ChemR23 axis may mediate HIS-induced analgesia and inhibit neuroinflammation in CPIP mice.

Figure 7.

The RvE1-ChemR23 axis is involved in HIS-induced analgesia.

(A, B) ChemR23 expression was significantly lower in the sh-ChemR23 group than in the sh-Neg control group and naïve group (n = 6 mice/group). *P < 0.05, **P < 0.01 and #P < 0.05, vs. naïve or sh-Neg control group. (C–F) The bilateral mechanical and heat pain thresholds in the sh-ChemR23 group were significantly decreased compared with those in the sh-Neg control and naïve groups (n = 7 naïve mice, n = 13 sh-Neg control mice, and n = 13 sh-ChemR23 mice (C); n = 9 naïve mice, n = 13 sh-Neg control mice, and n = 13 sh-ChemR23 mice (D); n = 6 naïve mice, n = 10 sh-Neg control mice, and n = 10 sh-ChemR23 mice (E and F)). *P < 0.05, **P < 0.01 and ***P < 0.001, vs. naïve group; ##P < 0.01 and ###P < 0.001, vs. sh-Neg control group; $$P < 0.01 and $$$P < 0.001, vs. before exercise. (G) Representative images of Iba1 (green, Alexa Fluor 488) immunofluorescence staining in the bilateral spinal dorsal horn in CPIP mice in the sh-Neg control exercise group and the sh-ChemR23 exercise group. The Iba1 fluorescence signal in CPIP sh-ChemR23 mice subjected to exercise was greater than that observed in CPIP sh-Neg control mice subjected to exercise. The righthand images (bar: 100 μm) are high magnifications of the boxes in the lefthand images (bar: 200 μm). (H, I) Iba1 fluorescence intensity and numbers of Iba1-positive microglia (n = 3 mice/group). **P < 0.01 and ****P < 0.0001, vs. sh-Neg control group. (J, K) The CPIP-induced increases in proinflammatory cytokine mRNA levels were not decreased after HIS in the sh-ChemR23 group (n = 3 mice/group). *P < 0.05, **P < 0.01 and ***P < 0.001, vs. sh-Neg control group. All data are presented as the mean ± SEM, and were analyzed by two-way analysis of variance followed by Bonferroni post hoc test (B–F) or two-tailed unpaired Student’s t-test (H-K). Detailed statistical information is shown in Additional Table 2 ^{(125.3KB, pdf)} and Additional file 1 ^{(136.8KB, pdf)} . ChemR23: Chemerin receptor 23; Contra: contralateral; CPIP sh-ChemR23: mice were injected ChemR23 shRNA-expressing lentivirus and submitted to chronic postischemia pain; CPIP sh-Neg con: mice were injected were injected nontargeting control shRNA-expressing lentivirus and submitted to chronic postischemia pain; GAPDH: glyceraldehyde-3-phosphate dehydrogenase; Iba1: ionized calcium binding adaptor molecule 1; IL: interleukin; IOD: integrated optical density; Ipsi: ipsilateral; RvE1: Resolvin E1; sh-ChemR23: mice were injected with ChemR23 shRNA-expressing lentivirus; sh-Neg con: mice were injected with nontargeting control shRNA-expressing lentivirus; TNF-α: tumor necrosis factor-α.

Exogenous RvE1 exerts analgesic, anti-neuroinflammatory, and central sensitization effects in CPIP mice

RvE1 was intrathecally injected to test the analgesic effect of exogenous RvE1 on nociception in CPIP mice. Intrathecal injection of RvE1 (10 ng) reduced the response to mechanical and heat-induced pain in the bilateral hindpaws of CPIP mice (Figure 8A–D). In addition, RvE1 (20 ng) induced CPP in CPIP mice, indicating that RvE1 suppresses ongoing pain (Figure 8E–H). RvE1 also exerted marked anti-neuroinflammatory effects, as indicated by reduced microglial activation and lower levels of the proinflammatory cytokines TNF-α, IL-1β, and IL-6 in the bilateral spinal cord (Figure 8I–M). However, RvE1 had no effect on the expression of the anti-inflammatory cytokines IL-4 and IL-10 (Figure 8L and M). Compared with control mice, the inhibitory effect of RvE1 on WDR neuron spontaneous firing was more robust in CPIP mice (Figure 8N–Q). These data suggested that spinal cord supplementation with exogenous RvE1 simulates the effect of HIS on decreasing nociceptive behavior, neuroinflammation, and central sensitization in CPIP mice.

Figure 8.

Exogenous RvE1 inhibits CPIP-induced nociception, neuroinflammation, and central sensitization.

(A–D) Intrathecal injection of RvE1 (10 ng) relieved bilateral punctate mechanical hyperalgesia and heat hyperalgesia on day 12 after CPIP model establishment (n = 6 mice/group). **P < 0.01, ***P < 0.001 and ****P < 0.0001, vs. CPIP RvE1 group. (E) Schematic of the CPP assay. (F) Schematic of the CPP protocol for CPIP mice. (G) Representative CPP traces from mice injected with vehicle, 10 ng of RvE1, or 20 ng of RvE1. CPIP mice activity was clearly increased in 20 ng of RvE1-paired chambers compared with vehicle- and 10 ng of RvE1-paired chambers. (H) Intrathecal injection of RvE1 (20 ng) induced CPP in CPIP mice (n = 5 mice/group). **P < 0.01, vs. vehicle. (I) Representative images of Iba1 (green, Alexa Fluor 488) immunofluorescence staining in the bilateral spinal dorsal horn of CPIP mice intrathecally injected with vehicle or RvE1 (10 ng). Iba1 fluorescence in CPIP mice was clearly reduced after RvE1 (10 ng) injection. The righthand images (bar: 100 μm) are high magnifications of the boxes in the lefthand images (bar: 200 μm). (J, K) Iba1 fluorescence intensity and numbers of Iba1-positive microglia (n = 3 mice/group). **P < 0.01, ***P < 0.001 and ****P < 0.0001, vs. CPIP vehicle group. (L, M) Intrathecal injection of RvE1 (10 ng) reduced the increase in mRNA levels of the proinflammatory cytokines IL-1β, IL-6, and TNF-α in the bilateral spinal lumbar enlargement induced by CPIP (n = 5 mice/group for TNF-α and n = 6 mice/group for the other cytokines). **P < 0.01 and ****P < 0.0001, vs. CPIP vehicle group. (N) Perfusion of the spinal cord with 10 ng of RvE1 but not 1 ng of RvE1 decreased the spontaneous firing rate of spinal cord WDR neurons in control mice (n = 3 mice/group, n = 11 cells). *P < 0.05, vs. baseline. (O) WDR neuron firing rate ratio to baseline in control mice injected with 1 ng and 10 ng of RvE1 respectisely. (P) Perfusion of the spinal cord with 1 ng or 10 ng of RvE1 decreased the spontaneous firing rate of spinal cord WDR neurons in CPIP mice (n = 3 mice/group, n = 11 cells). ***P < 0.001, vs. baseline. (Q) WDR neuron firing rate ratio to baseline in CPIP mice injected with 1 ng or 10 ng of RvE1 respectively. All data in A–D, H, J–M, O and Q are presented as the mean ± SEM, and the data in N and P are presented as the median ± IQR, and were analyzed by two-way analysis of variance followed by Bonferroni post hoc test (A-D), one-way analysis of variance followed by Bonferroni post hoc test (H), two-tailed unpaired Student’s t-test (J–M, O, Q), or Wilcoxon signed-rank test (N, P). Detailed statistical information is shown in Additional Table 2 ^{(125.3KB, pdf)} and Additional file 1 ^{(136.8KB, pdf)} . Contra: Contralateral; Control RvE1: the ankles of mice were surrounded by a cut-O-ring and were injected with RvE1; Control Vehicle: the ankles of mice were surrounded by a cut-O-ring and were injected with 1% ethanol; CPIP RvE1: mice were submitted to chronic postischemia pain and injected with RvE1; CPIP Vehicle: mice were submitted to chronic postischemia pain and injected with 1% ethanol; CPP: conditional place preference; Iba1: ionized calcium binding adaptor molecule 1; IOD: integrated optical density; IL: interleukin; Ipsi: ipsilateral; RvE1: Resolvin E1; TNF-α: tumor necrosis factor-α; WDR: wide dynamic range.

Discussion

In the present study, we found that nociceptive behavior, neuroinflammation, and central sensitization in CPIP mice are significantly alleviated by HIS but not LIS. Furthermore, we found that the RvE1-ChemR23 axis in the spinal cord mediates HIS-induced analgesia, inhibition of neuroinflammation, and central sensitization in CPIP mice.

The CPIP rodent model has been intensively used in recent studies of pain related to CRPS-1 (Millecamps et al., 2010; Martins et al., 2013; Chen et al., 2020a; De Logu et al., 2020; Zhang et al., 2022). Ischemia could be the main cause of CRPS-1 (Birklein et al., 2000; Koban et al., 2003). Thus, in the present study, we established a CPIP mouse model to mimic CRPS-1. In addition to the CPIP model, the tibia fracture model is also frequently used as a CRPS-1 model (Shi et al., 2018; Guo et al., 2019). Comparatively, CPIP takes less time to establish than the tibia fracture and mimics the clinical signs of CRPS-1 in many ways, with skin warming and swelling in the acute phase, accompanied by neuropathic-like pain in the chronic phase (Coderre et al., 2004). In addition, the symptoms exhibited by the CPIP mouse model are consistent with most of the Budapest Clinical Diagnostic Criteria for CRPS adopted by the International Association for the Study of Pain (Harden et al., 2022), such as continuous pain, hyperalgesia and allodynia, changes in skin temperature and color, edema, and motor dysfunction (Additional Table 3). However, there are some inconsistencies between this CPIP model and clinical CRPS. Pain in CRPS-1 is usually unilateral, although CRPS patients sometimes exhibit mirror pain (van Rooijen et al., 2013; Terkelsen et al., 2014). However, the CPIP mice in our study and other studies showed bilateral pain (Coderre et al., 2004; Han et al., 2012; Chen et al., 2020a), suggesting that the CPIP model does not completely represent the spatial aspects of human CRPS.

Additional Table 3.

The comparison of Clinical Diagnostic Criteria for CRPS adopted by the IASP in 2012 and the characteristics observed in our CPIP model mice

Budapest clinical diagnostic criteria for CRPS adopted by the IASP in 2012	Characteristics observed in our CPIP mice
1. Continuing pain, which is disproportionate to any inciting event	CPIP model mice showed persistent pain.
2. Must report at least one symptom in three of the four following categories:	Not applicable (The symptoms can not be measured in CPIP mice)
Sensory: Reports of hyperalgesia and/or allodynia	Not applicable
Vasomotor: Reports of temperature asymmetry and/or skin color changes and/or skin color asymmetry	Not applicable
Sudomotor/Edema: Reports of edema and/or sweating changes and/or sweating asymmetry	Not applicable
Motor/Trophic: Reports of decreased range of motion and/or motor dysfunction (weakness, tremor, dystonia) and/or trophic changes (hair, nail, skin)	Not applicable
3. Must display at least one sign* at time of evaluation in two or more of the following categories:	The signs of CPIP mice were consistent with most criteria
Sensory: Evidence of hyperalgesia (to pinprick) and/or allodynia (to light touch and/or deep somatic pressure and/or joint movement)	The CPIP mice showed hyperalgesia to von Frey filament stimulation and radiant heat stimulation, accompanied by spontaneous pain (Figure 1D-K)
Vasomotor: Evidence of temperature asymmetry and/or skin color changes and/or asymmetry	After ischemia/reperfusion, the temperature of ipsilateral hind paw of our CPIP mice was obvious increased in a short time (Figure 1C)
Sudomotor/Edema: Evidence of edema and/or sweating changes and/or sweating asymmetry	After ischemia/reperfusion, the thickness of ipsilateral hind paw of our CPIP mice was obviously increased in a short time (Figure 1A and B)
Motor/Trophic: Evidence of decreased range of motion and/or motor dysfunction (weakness, tremor, dystonia) and/or trophic changes (hair, nail, skin)	The distance traveled and velocity of CPIP mice in the open field test were decreased at the late stage of CPIP generation (Figure 1L and M)
4. There is no other diagnosis that better explains the signs and symptoms	Not applicable

CPIP: Chronic postischemia pain; CRPS: complex regional pain syndrome; IASP: International Association for the Study of Pain.

Neuroinflammation in the spinal cord contributes to the transition from acute to chronic pain in CRPS-1 and plays an important role in central sensitization in the CPIP rodent model (Chen et al., 2020a). Furthermore, neuroinflammation is positively correlated with the degree of pain in patients with CRPS-1 (Jung et al., 2019). Microglia, as intrinsic immune cells of the spinal cord and brain, play an important role in pathological pain, including inflammatory pain, neuropathic pain, and cancer-related pain (Ji et al., 2013; Chen et al., 2018; Inoue and Tsuda, 2018; Apryani et al., 2020). Microglia are activated in the spinal cord of patients with CRPS-1 (Del Valle et al., 2009), which accelerates the neuroinflammatory response (Ji et al., 2018). In our CPIP mice, microglia in the spinal cord were activated at the early stage and persisted during the development of chronic pain. Consistently, the levels of the proinflammatory cytokines IL-1β, IL-6, and TNF-α were increased in the spinal cord of CPIP mice, which is supported by previous studies reporting that microglia are the major source of cytokines, including IL-1β, IL-6, and TNF-α (Hanisch, 2002; Zhang et al., 2018). Release of these proinflammatory cytokines by activated microglia in the spinal cord after injury could promote central sensitization by increasing the excitability of excitatory neurons and inhibiting the excitability of inhibitory neurons (Chen et al., 2018; Ji et al., 2018). In addition, the spontaneous firing rate of WDR neurons was increased not only at the early stage but also at the late stage of CPIP, and the pain persisted after peripheral inflammation, as indicated by redness and swelling, was relieved. Thus, we believed that CPIP mice presented central sensitization.

Exercise, a nonpharmacological treatment, has been suggested to be an effective method and an integral part of rehabilitation for the treatment of patients suffering from a variety of chronic pain conditions. Exercise can effectively inhibit microglial activation, reduce the levels of pro-inflammatory cytokines, increase the levels of anti-inflammatory cytokines, and decrease oxidative stress (Shi et al., 2018; Mee-Inta et al., 2019). To explore the mechanism of EIA, we subjected CPIP mice to a swimming protocol, as swimming is a common aerobic exercise that is suitable for animals with limb nociception due to the lighter load borne by the affected limb (Hutchinson et al., 2004). We adopted a previously reported HIS protocol and found that HIS not only substantially relieved punctate mechanical hyperalgesia and heat hyperalgesia, but also decreased neuroinflammation and central sensitization in the spinal cord of CPIP mice. However, when we reduced the swimming training from 5 days to 3 days and decreased the swimming time every day with no more than 15 minutes, there was no effect on the nociceptive behavior threshold of the CPIP mice compared with the no exercise group, which supported the findings from a previous report that intense exercise is important for altering sensory hypersensitivity (Stagg et al., 2011).

Furthermore, the analgesic effect mediated by swimming was not associated with stress response or negative emotional reactions, because there was no difference in serum corticosterone levels or open field and tail suspension behaviors between the HIS group and the no exercise group. We also found that the percentage of time spent swimming during the HIS protocol was comparable in CPIP and control mice (at approximately 80%), further supporting the conclusion that nociceptive behavior of mice was relieved by swimming, and not due to stress. Several factors might explain why the effects are attributable to HIS rather than to stress-induced analgesia. First, the mice were habituated to the swimming environment for 2 days before the swimming sessions started. Second, the mice swam in warm water (37°C) and not in cold water; the latter can easily induce stress. Third, the emotional behavior test was conducted 24 hours after the last swimming exercise, consistent with the time at which we measured nociceptive behavior; thus, even if stress was present, its influence would have been minimized at this time point (Mazzardo-Martins et al., 2010). Fourth, the swimming protocols were performed more than once, which may have helped the mice adapt to the experimental conditions.

SPMs, including the omega-3-derived resolvin, protectin, and maresin families, as well as arachidonic acid-derived lipoxins, are endogenous mediators that promote the inflammation resolution, microbe clearance, pain reduction, and tissue regeneration (Arita et al., 2005a; Svensson et al., 2007; Serhan et al., 2009, 2014, 2018; Sekheri et al., 2020). It has been reported that SPMs are synthesized during the acute phase of inflammation, and that failure of SPM biosynthesis leads to a transition from acute to chronic inflammation (Serhan et al., 2008; Ji et al., 2011). According to recent research, SPMs suppress microglial activation and inhibit pain through receptor-mediated actions in the central nervous system (Xu et al., 2010, 2013; Tao et al., 2020). RvE1, the first identified omega-3 polyunsaturated fatty acid–derived resolvin, has been reported to suppress inflammatory responses, relieve chronic pain, and suppress Lipopolysaccharides-induced depression-like behaviors by acting on its receptor ChemR23 (Arita et al., 2007; Xu et al., 2010; Deyama et al., 2018). SPM levels were found to be increased and exert anti-inflammatory effects after exercise (Zheng et al., 2019), which prompted us to explore the role of the RvE1-ChemR23 axis in mediating EIA in CPIP mice. Our findings suggest that HIS may attenuate nociception by activating the RvE1-ChemR23 axis in the spinal cord. Firstly, CPIP mice presented lower levels of RvE1 in the spinal cord compared to control mice. Secondly, ChemR23 expression in the spinal cord was significantly decreased after injury. Thirdly, RvE1 and ChemR23 levels were significantly increased after HIS. Fourth, HIS-induced analgesia, decreased neuroinflammation, and decreased central sensitization were substantially inhibited by ChemR23 silencing. Fifth, exogenous supplementation with RvE1 via intrathecal injection exerted the same effects as HIS, namely analgesia, inhibition of neuroinflammation, and central sensitization, in CPIP mice. However, we cannot exclude the possibility that HIS induced the synthesis and release of other endogenous substances. HIS-induced analgesia was weakened but not blocked by ChemR23 silencing. This effect could be explained by the expression of other SPM receptors, such as FPR2, which was also increased after HIS.

There are some limitations to the present study. Firstly, although we demonstrated that HIS increased RvE1 expression, the cell type that biosynthesizes and releases RvE1 in the spinal cord was not investigated. Secondly, since ChemR23 is expressed by both microglia and neurons, microglia- or neuron-specific knockout of ChemR23 should be performed to further clarify the role of the RvE1-ChemR23 axis in mediating HIS-induced analgesia. Thirdly, we only used male mice in this study. Emerging evidence has demonstrated sex differences in microglia in the central nervous system that may play a vital role in many physiological and pathological processes. Whether the effect and mechanism of HIS-induced analgesia are the same in male and female mice needs further study.

In summary, the RvE1-ChemR23 axis in the spinal cord mediates HIS-induced analgesia by inhibiting neuroinflammation and central sensitization in CPIP mice. Our study not only provides a new mechanism for EIA induced by swimming but also supports the utility of rehabilitation treatment with physical exercise for patients with chronic pain, particularly patients with CRPS-1.

Additional files:

Additional Figure 1 ^{(371.4KB, tif)} : The time course of HIS-induced analgesia in CPIP mice.

Additional Figure 1

The time course of HIS-induced analgesia in CPIP mice.

(A, B) Compared with pre-exercise, the bilateral mechanical allodynia (A) and heat hyperalgesia thresholds (B) after HIS in CPIP male mice were significantly increased, which lasted for more than 40 hours. All data in A and B are presented as the mean ± SEM (n = 6 mice). ***P < 0.001 and ****P < 0.0001, vs. before exercise in contralateral; ####P < 0.0001, vs. before exercise in ipsilateral (two-way analysis of variance followed by a Bonferroni post hoc test (A, B)). CPIP: Chronic postischemia pain; HIS: high-intensity swimming. Detailed statistical information was shown in Additional Table 2 ^{(125.3KB, pdf)} and Additional file 1 ^{(136.8KB, pdf)} .

Additional Figure 2 ^{(766.9KB, tif)} : Colocalization of ChemR23 with NeuN (neuronal marker), Iba1 (microglial marker), and GFAP (astrocyte marker) in the spinal cord dorsal horn of mice.

Additional Figure 2

Colocalization of ChemR23 with NeuN (neuronal marker), Iba1 (microglial marker), and GFAP (astrocyte marker) in the spinal cord dorsal horn of mice.

(A to L) Colocalization of ChemR23 (red, Alexa Fluor 555) with NeuN (green, Alexa Fluor 488) (A-D), Iba1 (green, Alexa Fluor 488) (E-H) or GFAP (green, Alexa Fluor 488) (I-L). The D, H, and L are high magnifications of boxes in C, G, and K. The arrowhead indicates the colocalization of ChemR23 with NeuN and Iba1. Scale bars: 100 µm. ChemR23: Chemerin receptor 23; GFAP: glial fibrillary acidic protein; Iba1: ionized calcium binding adaptor molecule 1; NeuN: neuronal nuclei.

Additional Table 1: Primers sequence for quantitative reverse transcription-polymerase chain reaction.

Additional Table 2 ^{(125.3KB, pdf)} : Statistical data of Figures 2–8 and Additional Figure 1 ^{(371.4KB, tif)} .

Additional Table 2

Statistical data of Figures 2-8 and Additional Figure 1 ^{(371.4KB, tif)}

Additional Table 3: The comparison of Clinical Diagnostic Criteria for CRPS adopted by the IASP in 2012 and the characteristics observed in our CPIP model mice.

Additional file 1 ^{(136.8KB, pdf)} : The mean, standard deviation or median, interquartile range of all data in figures.

Additional file 1

The mean, standard deviation or median, interquartile range of all data in figures.