C57BL/6 J mice, aged 7–8 weeeks, were obtained from Chengdu Dashuo Animal Co., Ltd. All animal care and experimental techniques in this work were authorized by the Animal Ethics Committee of Chengdu University of Traditional Chinese Medicine, in accordance with the National Institutes of Health guidelines for the care and use of laboratory animals.

In short, mice were given at least 30 min to acclimate in Plexiglas chambers (15 cm × 12 cm × 10 cm) before formalin tests. Ten percent formalin (Solarbio, China) was diluted to 2% in PBS (Solarbio,China) and stored at room temperature. Twenty microliters of 2% formalin was administered to the plantar side of the left hind paw. The formalin-induced pain behavior-flinching was kept under observation in separate cages for 60 min, with 5-min intervals. For evaluation, pain grades were combined and grouped into two phases: phase I (0–10 min) and phase II (10–60 min). All tests were performed between 11:00 a.m. and 3:00 p.m. in a quiet room.

For the treatment of electroacupuncture (EA), mice were restrained without anesthesia. A pair of needles (0.25 × 25 mm, Hwato-Med. Co., Jiangsu, China) were delivered into the ipsilateral “Yongquan” acupoint (Fig. 1b). Additionally, an assist needle was positioned at a depth of 3 mm in the tail root, which does not correspond to an acupoint. The two needles were connected to the output terminals of an EA device (HANS-200A Acupoint Nerve Stimulator, Nanjing Jisheng Medical Technology Co., Ltd., Jiangsu, China). The frequency of the electrical stimulation was set at 2/200 Hz, 0.5 mA. Formalin was injected after 30 min of EA treatment. The sham EA group received the same intervention as EA group without electric stimulation. In brief, following skin cleaning, the ipsilateral acupoint was treated for 30 min utilizing a pair of identical needles employed in the EA treatment, placed at exactly the same depth as in the EA group. The EA device did not connect to the needles, yet.

Analgesic effect of EA on formalin-induced pain behavior in mice. Mice’s left hind paws were injected with either 2% formalin or PBS alone at 0 min. In an additional set of trials, mice’s ipsilateral Yongquan acupoint (EI1) was treated with EA or sham EA for 30 min prior to the injection of 2% formalin. The needle was positioned identically to that of the real EA for the sham EA, but without any electrical stimulation. A total of 12 times, starting at the 0-min time point and continuing every 5 min until the 60-min time point, the number of flinches was recorded. For comparison, the number of flinching was counted in the phase I (0 to 10 min) and in the phase II (10 to 60 min), respectively. In Figs. 1, 2, 3, 4, and 5, this experimental approach was used. Thirty two mice were used to generate the data presented below (c–d). a Schematic of the experimental design. b Location of Yongquan acupoint (EI1). c Flinching number in phase I (0–10 min), phase II (10–60 min). In phase I (0–10 min), PBS vs formalin P < 0.0001; formalin vs EA + formalin P < 0.001; formalin vs sham EA + formalin P > 0.05. In phase II (10–60 min), PBS vs formalin P < 0.0001; formalin vs EA + formalin P < 0.0001; formalin vs sham EA + formalin P > 0.05. d Flinching number in total. PBS vs formalin P < 0.0001; formalin vs EA + formalin P < 0.05; formalin vs sham EA + formalin P > 0.05. Significance was assessed by two-way ANOVA followed by Tukey’s multiple comparisons test in c and Kruskal–Wallis one‐way ANOVA in d. All data are presented as the mean ± SEM. n = 8, *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001

C57BL/6 J mice were anesthetized with isoflurane (1% for maintenance and 4% for induction; RWD Life Sciences, Shenzhen, China) then fixed in a stereotactic frame (RWD Life Sciences, Shenzhen, China). 0.5 μL adeno-associated virus (AAV) using a micro syringe pump (RWD Life Sciences, Shenzhen, China) at a flow rate of 0.05 μL/min was injected into the RSCs (AP: − 0.4 mm; ML: ± 1.0 mm; DV: − 1.0 mm). Diffusion and reflux control were allowed for additional 10 min. In another round of tests, drug delivery catheter (RWD Life Sciences, Shenzhen, China) was implanted into the RSCs (AP: − 0.4 mm; ML: ± 1.0 mm; DV: − 1.0 mm) and secured with dental cement. In order to avoid postoperative infection, animals received an intravenous injection of 5 mg/kg enrofloxacin (RWD Life Sciences, Shenzhen, China) at the end of the procedure. Additionally, all animals were placed on a heating pad (37 °C) to keep their body temperature stable during the procedure.

The astrocyte chemogenetic modification in this investigation involved the application of the following viruses: rAAV2/5-GfaABC1D-hM4D(Gi)-EGFP (titer: 2.02 × 1012 VG/mL), rAAV2/5-GfaABC1D-hM3D(Gq)-EGFP (titer: 2.10 × 1012VG/mL, all from Brain Case, China). The full expression of virus was allowed for 3 weeks. The chemogenetic receptors hM3Dq (Gq) and hM4Di (Gi) were used to manipulate astrocyte activity. Subsequently, mice received an intraperitoneal injection of 1 mg/kg clozapine-N-oxide (CNO; Sigma-Aldrich, Saint Louis, MO, USA) to activate Gq and Gi receptors. hM3Dq (Gq) + CNO activates astrocytes, while hM4Di (Gi) + CNO suppresses astrocytes. One hour after CNO injection, 20 ul of 2% formalin was injected into the left paw and the ensuing pain behavior was observed and recorded every 5 min for 60 min (Fig. 2a). In a different series of tests, EA was administered for 30 min following a 1-h interval post CNO injection. Following EA treatment, 2% formalin 20 ul was administered into the left paw, and the formalin-induced flinching was monitored every 5 min for 60 min (Fig. 2a). The following virus was applied to suppress P2X7R on astrocytes: ssAAV-GFAP-mir30shRNA(mP2X7)-WPRE-SV40pA (titer: 1 × 1012GC/mL, AAV8, all from PackGene Biotech, China). ssAAV-GFAP-mir30shRNA(NC)-WPRE-SV40pA (titer: 1 × 1012GC/mL, AAV8) was used as control virus. After 2 weeks of virus exprssion, 2% formalin 20 ul was administered into the left paw and then measured every 5 min until the 60 min time point (Fig. 5a). In another set of experiments, after 2 weeks of virus expression, EA was performed for 30 min, and then, 20 µl of 2% formalin was delivered into the left paw, and the animal was monitored every 5 min for 60 min (Fig. 5a).

Suppression of astrocytes in RSCs resulted in increased formalin-induced acute pain behavior, which was reversed by EA. All mice were given an injection of 2% formalin into their left hind paw at 0 min following 3 weeks of virus administration. In a different series of tests, mice were given EA injections to the ipsilateral Yongquan acupoint (EI1) for 30 min prior to receiving an injection of 2% formalin. The flinching number was computed as Fig. 1 illustrates. The findings shown below (b–e) were produced using 48 mice. a Schematic of the experimental design. b, c Effects of chemogenetic activation of astrocytes on the number of phase I, II, or total flinches in post-formalin-induced mice. b In phase I (0–10 min), hM3D (Gq) vs hM3D (Gq) + EA P > 0.05; NC vs hM3D (Gq) P > 0.05. In phase II(10–60 min), hM3D (Gq) vs hM3D(Gq) + EA P < 0.05; NC vs hM3D (Gq) P > 0.05; NC vs NC + EA P < 0.001. c Flinching number in total. NC vs NC + EA P < 0.01; there was no difference between NC and hM3D (Gq). d, e Effects of chemogenetic inhibition of astrocytes on the number of phase I, II, or total flinches in post-formalin-induced mice. d There was no difference in Phase I. In Phase II(10–60 min), hM4D(Gi) significantly increased formalin-induced pain (P < 0.05), which was reversed by EA (P < 0.0001). e In the total number of flinches, hM4D(Gi) vs hM4D(Gi) + EA P < 0.001; NC vs hM4D(Gi) #: P > 0.999. Significance was assessed by two-way ANOVA followed by Tukey’s multiple comparisons test in b, d and Kruskal–Wallis one‐way ANOVA in c, e. All data are presented as the mean ± SEM. n = 8, * P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001

The drugs used as following: amiloride hydrochloride, 2′(3′)‐O‐(4-benzoylbenzoyl)adenosine‐5′‐triphosphate tri(triethylammonium) salt (benzoylbenzoyl ATP, Bz-ATP); 3-((5-(2,3-dichlorophenyl)-1H-tetrazol-1-yl)methyl)pyridine hydrochloride (A438079); l-alpha-aminoadipate (l-AA); (all from Sigma-Aldrich, Saint Louis, MO, USA); and PBS (Solarbio,China). Bz-ATP and A438079 utilized as P2X7R agonists and antagonists, respectively. l-AA is a specific toxic drug for astrocytes, which leads to the inactivation of astrocytes. Stock solutions of all drugs (A438079, 1 uM; Bz-ATP, 10 uM; LAA, 40 mg/kg) were diluted in PBS. At least 1 week after surgery, mice were anesthetized and fixed in a stereotactic frame. 0.5 µL of drug was injected into the RSCs at a rate of 0.1 µL/min, with an additional 5 min to prevent drug leakage. After 30-min injection, 2% formalin 20 µl was injected into the left paw of the hindlimb and measured every 5 min until the 60 min time point (Fig. 3a). In another set of experiments, after 30 min of drug injection, EA was performed for 30 min, and then, 2% formalin 20 ul was injected into the left hind paw and observed every 5 min until 60 min (Fig. 3a).

LAA reverses the analgesic effect of EA on formalin-induced nociception. Mice were treated with LAA into the brain one week after surgery, whereas controls received either PBS injections or no medication at all. Thirty minutes later, all mice received an injection of 2% formalin into their left hind paw. In a different series of tests, mice were given EA injections to the ipsilateral Yongquan acupoint (EI1) for 30 min prior to receiving an injection of 2% formalin. The flinching number was computed as Fig. 1 illustrates. a Schematic of the experimental design. b Schematic diagram of intracerebral drug delivery in mice. The information shown below was produced using an additional 40 mice (e, f). c Immunofluorescence of GFAP expression of astrocytes after LAA or PBS in RSC region. Red fluorescence: GFAP; blue fluorescence: DAPI. d Summary data for the number of astrocytes colocalized with GFAP immunofluorescence in RSC, n = 9 sections from 3 mice. Scale bar, 40 μm. Effect of LAA and EA on the amount of phase I, phase II (e formalin vs formalin + EA P < 0.0001; LAA + EA + formalin vs EA + formalin P < 0.01; LAA + formalin vs LAA + EA + formalin P > 0.05; PBS + formalin vs LAA + formalin P > 0.05 in phase II), or total flinching (f EA + formalin vs LAA + EA + formalin &: P = 0.1309; formalin vs formalin + EA P < 0.01) in formalin-induced mice. Significance was assessed by two-way ANOVA followed by Tukey’s multiple comparisons test in e and Kruskal–Wallis one‐way ANOVA in f. d was assessed by unpaired t-tests. All data are presented as the mean ± SEM. n = 8,*P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001

Two percent pentobarbital (15 ml/kg; Sigma-Aldrich, Saint Louis, MO, USA) was used to profoundly sedate the mice. Following sedation, 0.9% saline and 4% paraformaldehyde were intracardially perfused. Brain were removed and fixed for 24 h at 4℃ with 4% paraformaldehyde (Sigma-Aldrich, USA) and dehydrated for another 48 h with 30% sucrose. Then they were embedded in Tissue-Tek OCT compound (Sakura Finetek, Umkirch, Germany). For immunofluorescence, the embedded tissues were sliced 25-μm thick using a cryostat microtome (Sakura Finetek, Umkirch, Germany). The sections were sequentially incubated with 5% goat serum (1 h, 37 °C), 0.5% Triton X-100 (Biosharp, China), Mouse anti-GFAP (GA5) antibody (1:200, 24 h, 4 °C, Cell Signaling Technology, USA), Mouse anti-P2X7 antibody (1: 100, 48 h, 4 °C, Santa Cruz Biotechnology, sc-514962, USA). Subsequently, the sections were washed with PBS three times (10 min each) and incubated with second antibodies (2 h, 37 °C; Alexa Fluor 594 1:200, Zenbio, China). Nuclei staining was performed using DAPI combined with an anti-fluorescent bursting agent (RT, Solarbio, China) for 5 min. A fluorescence microscope (Olympus, Japan) was used to obtain the fluorescence signal.

Data were analyzed offline using GraphPad Prism software, version 9.0 for Windows (GraphPad Software, USA). Significance was assessed by two-way ANOVA followed by Tukey’s multiple comparison test for phase I, phase II, and Kruskal–Wallis one-way ANOVA for total number of flinches; immunofluorescence data were analyzed by unpaired Student’s t-test. All data are presented as mean ± standard error of the mean (SEM), and P < 0.05 was considered significant.