Healthy adult male-specific-pathogen-free (SPF) Sprague–Dawley rats (260 ± 20 g) were purchased from Huaxing Experimental Animal Center (License No.: SCXK (YU) 2019–0002, Zhengzhou, China). Rats were raised in the Central Laboratory of the First Affiliated Hospital of the Henan University of Chinese Medicine with free access to water and food. Rats were allowed to acclimatize to the new surroundings for 1 week. Our animal experiments were approved by the Animal Ethics Committee of the First Affiliated Hospital of the Henan University of Chinese Medicine (Approval No. YFYDW2019037).

To exclude rats with congenital dementia from our project, the spatial learning and memory abilities of rats after 1 week of acclimatization were assessed by MWM assay as previously described (23, 24). A circular water tank (diameter: 1.6 m; height: 0.5 m) was used in the MWM assay. The tank was filled with water (water depth: 30 cm; water temperature: 21–25°C) before training. The water tank was divided into four equal quadrants. A colorless transparent circular escape platform (diameter: 12 cm; height: 29 cm) was placed in the center of one quadrant and submerged ~1 cm below the water surface. Visual references were hung at different sites around the maze. Before animal experiments, rats with congenital dementia were screened out by place navigation tests after three rounds of training. For each round of training, rats were initially placed into the water in quadrant 1 and then sequentially placed into the other three quadrants in a clockwise direction. Rats were given 90 s to find the hidden platform. If rats could find the platform within 90 s and stay on the platform for over 3 s, the time that they spent finding the escape platform (i.e., escape latency time) was recorded. If rats could not find the platform within 90 s, they were guided to the escape platform and allowed to rest for 10–30 s before the subsequent trials, and the escape latency time was recorded to be 90 s. Rats that did not find the platform within 90 s were considered abnormal in intelligence and were excluded from our project. A total of 72 rats were used in the subsequent experiments. Among these 72 rats, 12 rats were assigned to the control group and 15 rats were allocated to the sham group. The other 45 rats were used for the construction of the MCAO/R model. These rats were randomly assigned.

Additionally, the effect of EA treatment on the spatial learning and memory abilities of MCAO/R rats was assessed by a spatial navigation test on days 9–13 after EA treatment. Rats were subjected to two rounds of training each day. On day 14 after EA treatment, the spatial memory abilities of rats were further examined by a spatial probe test. Briefly, the platform was removed after spatial navigation experiments. Next, the rats were placed in the water and the number of times the rats crossed the original platform position within 90 s was recorded.

After 12 h of fasting, rats in the sham (n = 15) and MCAO/R (n = 45) groups were anesthetized using anesthetic pentobarbital sodium (50 mg/kg body weight, Sigma-Aldrich, St. Louis, MO, USA) by intraperitoneal injection. Then, MCAO/R rat models were constructed as previously reported (25–27). Briefly, rats (n = 45) were fixed onto operating tables in the supine position after anaesthetization. Then, the left common carotid artery (CCA), the external carotid artery (ECA), and the internal carotid artery (ICA) were exposed through a neck longitudinal incision. The proximal parts of the ECA and the CCA were ligatured, and a slip knot was reserved at the bifurcation site of the CCA near the ICA and the ECA. The distal end of the ICA was clamped using the artery clip. The MCAO monofilament (type: 2636A4, Beijing Xinong Technology Co., LTD, China) was inserted into the ICA (insertion depth: 18–20 mm) through a CCA “V”-shaped incision ~1 cm away from the bifurcation of the ECA and the ICA until mild resistance was felt. Approximately, 1 cm of MCAO monofilament was reversed outside. The CCA and the MCAO monofilaments were ligated at the reversed slipknot to fix the MCAO monofilament. Next, the neck incision was sutured and the end of the MCAO monofilament was labeled using a black marker pen. After 2 h of ischemia, the MCAO monofilament was withdrawn to achieve reperfusion. Rats in the sham group (n = 15) only suffered from exposure to CCA, ECA, and ICA. After surgery, all rats were intraperitoneally injected with gentamicin (20,000 units/rat) for three consecutive days and intramuscularly injected with furosemide (0.1 mg/kg). The electric incandescent lamp was used to control the body temperature of rats during the MCAO procedure. After anesthetization and surgery, rats were placed in a large box containing an electric heating blanket to control their body temperature. Nine rats died after MCAO/R surgery. Of these nine rats, five died during monofilament insertion, and four died after the removal of the monofilaments. The survival rate of the MCAO/R rats was 80.0% (36/45).

At 2 h after surgery, the neurological function impairment patterns of rats were evaluated by the Zea-Longa score system as previously described (25). The neurological deficit scores were examined by a researcher who was blinded to our experimental designs. The details of the Zea-Longa score system were as follows: a score of 0 did not show neurological injury symptoms; a score of 1 signified a mild focal neurological deficit with the symptom of a failure to extend the surgery contralateral forepaw fully; a score of 2 represented a moderate focal neurological deficit with the symptom of circling to the contralateral side; a score of 3 indicated a severe focal neurological deficit with the symptom of falling to the contralateral side; and a score of 4 denoted a serious neurological deficit characterized by the inability to walk spontaneously and reduction/loss of consciousness. The MCAO/R model was regarded to be successfully established in rats with scores of 1–3. Rats with scores of 0 (n = 4) and 4 (n = 2) were excluded. Finally, 30 rats in the MCAO/R model construction group were enrolled in our subsequent experiments and randomly divided into the MCAO/R group (n = 15) and the electroacupuncture (EA) group (n = 15).

Day 2 after surgery, rats in the EA group were subjected to EA treatment at the GV20 and GV24 acupoints one time a day for a total of 14 days using the Huatuo electroacupuncture treatment device (Suzhou Medical Appliance Factory, Suzhou, China). The needles were inserted into the GV20 and GV24 acupoints at an angle of 45° and a depth of 2 mm. GV20 is in the middle of the parietal bone, while GV24 is on the anterior midline, in front of the frontal–parietal bone junction (Figure 1). The parameters of the Huatuo electroacupuncture treatment device were as follows: dilatational wave, 1–20 Hz of frequency, 1–3 mA of electric current intensity, and 30 min of duration time per day. On Day 7 or 14 after EA treatment, the neurological deficit patterns of rats were also assessed by the Zea-Longa score system.

Three rats per group were randomly selected from the control, sham, MCAO/R, or EA groups. Then, the whole brains of these rats were isolated for a subsequent triphenyl tetrazolium chloride (TTC) staining assay as previously described (28). Briefly, rat hearts were exposed after anesthesia and then perfused with 0.9% normal saline. Rats were sacrificed by cervical dislocation when the colorless liquid flowed from the right auricle incision and then the brains were quickly isolated. The other rats were decapitated after anesthesia. Next, the brains and hippocampi of these rats were quickly isolated on ice, snap-frozen in liquid nitrogen, and stored at −80°C.

After isolation, rat brains were frozen for ~20 min at −20°C and sectioned into 2 mm coronal slices. Next, the sections were stained with 2% TTC solution for 30 min at 37°C in the dark and then imaged. The areas of cerebral infarction were manifested by white staining, while the areas with no cerebral infarction were stained red. The volume of cerebral infarction was analyzed by Image-Pro Plus software (Media Cybernetics, Silver Spring, MD, USA).

Whole hippocampal tissues from the Sham (n = 9), MCAO/R (n = 9), and EA (n = 9) groups were used in the TMT-based proteomics analysis. Each group contained nine samples, and three samples per group were mixed into a proteomics specimen. In the proteomics analysis, the sham group samples were named C1 (TMT label no. 126), C2 (TMT label no. 127N), and C3 (TMT label no. 127C). The MCAO/R model group samples were named M1 (TMT label no. 128N), M2 (TMT label no. 128C), and M3 (TMT label no. 129N). The EA treatment group samples were termed T1 (TMT label no. 129C), T2 (TMT label no. 130N), and T3 (TMT label no. 130C). Proteins were regarded as significantly differentially expressed at an upregulated ratio of ≥ 1.2 or a downregulated ratio of ≤ 0.83 and an adjusted P-value of ≤ 0.05. The detailed experimental procedures of TMT-based proteomics analysis are shown in Supplementary file 1.

Venn diagram analysis was performed using the Jvenn website (http://jvenn.toulouse.inra.fr/app/example.html) (29). Protein–protein interaction (PPI) networks were established using STRING (version: 11.5) (https://cn.string-db.org/) (30, 31). K-means clustering analysis was performed using the MeV (MultiExperiment Viewer) program (https://mev.tm4.org/) (32).

Protein was extracted from hippocampal tissues of the sham (n = 3), MCAO/R (n = 3), and EA (n = 3) groups using RIPA buffer (Beijing Solarbio Science & Technology Co., Ltd., Beijing, China) supplemented with the protease inhibitor phenylmethanesulfonyl fluoride (PMSF). The protein concentration was determined using the BCA Protein Assay Kit (Solarbio Science & Technology Co., Ltd., Beijing, China). Next, an equal amount of protein (40 μg) was separated by 10% SDS-PAGE and transferred onto 0.22 μm polyvinylidene fluoride (Millipore, Bedford, MA, USA). After the blockade of nonspecific protein signals using 5% skimmed milk, the membranes were hybridized overnight with primary antibodies against Pak4 (1:1,000; ImmunoWay Biotechnology Company, Plano, TX, USA), Akt3 (1:1,000; ImmunoWay Biotechnology Company, Plano, TX, USA), Efnb2 (1:1,000; ImmunoWay Biotechnology Company, Plano, TX, USA), and Gapdh (1:10,000; ImmunoWay Biotechnology Company, Plano, TX, USA) at 4°C. On the next day, the membranes were probed for 1 h with goat–anti-rabbit or goat–anti-mouse (1:10,000; ImmunoWay Biotechnology Company, Plano, TX, USA) secondary antibodies conjugated with horseradish peroxidase at room temperature. Finally, protein bands were detected using the ECL Western Blotting Substrate (Beijing Solarbio Science & Technology Co., Beijing, China).

Statistical analysis was performed using GraphPad Prism software (version 7, La Jolla, CA, USA), and the results are shown as the mean ± standard deviation. The differences among groups were analyzed using one-way or two-way analysis of variance (ANOVA) and Tukey's test. The differences were defined as statistically significant when the p-value was < 0.05.

In this study, the neurological functions of rats in the control, sham, MCAO/R, and EA groups were assessed by the Zea-Longa score system. The results showed that there were no neurological deficit symptoms in the control and sham rats, while certain neurological deficits were observed in rats in the MCAO/R and EA groups at 2 h after surgery (Table 1). There was no significant difference (P > 0.05) in the neurological deficit scores between the MCAO/R and EA groups at 2 h after modeling. However, the number of rats with lower neurological deficit scores (scores 0 and 1) was notably increased in the EA group at 7 or 14 days after EA treatment compared to the MCAO/R model group (Table 1), suggesting that EA treatment could markedly alleviate MCAO/R-induced rat neurological injury.

The MWM assay was performed to examine the effect of EA treatment on the spatial learning and memory behaviors of MCAO/R rats. As seen in Figure 2A, MCAO/R model rats spent more time escaping latency than rats in the sham group. Moreover, the spatial probe assay showed that there was a marked decrease in the frequency of swimming across the platform within 90 s in the MCAO rats vs. the sham group (Figure 2B). These data showed that MCAO/R rats had impaired spatial learning and memory behaviors compared to the sham group. Additionally, EA treatment noticeably reduced the escape latency time of MCAO/R rats from Day 9 to Day 13 after stimulation (Figure 2A). After EA treatment, MCAO/R rats crossed the platform positions more times than untreated MCAO/R rats (Figure 2B). These outcomes suggested that EA treatment at the GV20 and GV24 acupoints notably ameliorated the spatial learning and memory impairments of MCAO/R rats.

TTC staining analysis further revealed that MCAO/R rats developed obvious cerebral infarction, while the volume of cerebral infarction was notably reduced in MCAO/R rats after EA treatment compared with untreated MCAO/R rats (Figure 3), suggesting that EA stimulation could markedly relieve cerebral infarction in MCAO/R rats.

TMT-based proteomics analysis showed that the expression levels of 75 proteins were notably downregulated and the expression levels of 120 proteins were markedly upregulated in the MCAO/R group vs. the sham group (Supplementary Table 1). The volcano plot of differentially expressed proteins in the MCAO/R vs. sham group is shown in Figure 4A. Moreover, a total of 218 differentially expressed proteins (50 downregulated, 168 upregulated) were identified in the EA treatment group compared with the MCAO/R group (Supplementary Table 2). The volcano plot of differentially expressed proteins in the EA vs. MCAO/R group is shown in Figure 4B.

Next, 26 proteins that were notably upregulated in the MCAO/R vs. sham group and markedly downregulated in the EA vs. MCAO/R group were screened out by jvenn analysis and are shown in Figure 5A and Supplementary Table 3. The K-means clustering plots of the expression profiles of these 26 proteins are shown in Figure 5B. Additionally, 36 proteins that were noticeably downregulated in the MCAO/R vs. sham group and upregulated in the EA vs. MCAO/R group were filtered out by jvenn analysis and are shown in Figure 5C and Supplementary Table 3. The K-means clustering plots of these 36 proteins are shown in Figure 5D. The heatmap of these 62 proteins with converse alteration trends in the MCAO/R vs. sham and EA vs. MCAO/R groups is shown in Figure 5E. As seen in Figure 5E, the repeatability of the samples in each group was favorable. That is, our proteomics data were relatively credible. We concluded that these 62 proteins might play vital roles in the EA-mediated protective effect against MCAO/R-induced cerebral ischemia/reperfusion injury.

Based on GO biological process enrichment data of differentially expressed proteins in the MCAO/R vs. sham group in Supplementary Table 4, GO biological processes related to brain and nerve development were screened out by searching for the terms containing the keywords brain, neurotransmitter, neuropeptide, neuromuscular, neurogenesis, neuron, nervous, nerve, axon, axonogenesis, synaptic, and synapse. These biological processes are shown in Supplementary Table 5. The differentially expressed proteins in the abovementioned biological processes associated with brain and nerve development are shown in Supplementary Table 6. These dysregulated proteins might be closely linked with MCAO/R-induced brain and nerve damage. Additionally, the potential interactions of proteins as shown in Supplementary Table 6 were predicted using the STRING database. As seen in Figure 6A, serine/threonine protein kinase Akt3 could interact with scribble planar cell polarity protein (Scrib) or serine/threonine protein kinase Pak4 (Figure 6A). Moreover, differentially expressed proteins in the EA vs. MCAO/R model group that might be implicated in brain and nerve development were filtered out based on GO enrichment analysis in Supplementary Table 7 using the abovementioned methods. The screened GO biological process terms and proteins are shown in Supplementary Tables 8, 9, respectively. PPI analysis of these proteins revealed that there were potential regulatory networks among eight proteins [i.e., Pak4, Akt3, phosphatidylinositol 3,4,5-trisphosphate 3-phosphatase and dual-specificity protein phosphatase Pten, DNA mismatch repair protein Msh2, histone deacetylase 2 (Hdac2), neuropilin-2 (Nrp2), ephrin-B2 (Efnb2), and semaphorin 4D (Sema4d)] (Figure 6B). Given the vital roles of the Akt pathway in multiple biological processes and the differential expressions of Akt pathway-related proteins, such as PAK4, AKT3, Pten, Nrp2, and Efnb2, in response to MCAO/R alone or in combination with EA treatment, we hypothesized that AKT pathway-related proteins might be closely linked with MCAO/R-induced brain damage and EA-mediated therapeutic effects against MCAO/R-induced brain damage. Proteomics expression profiles of these Pten/AKT pathway-related proteins in the sham, MCAO/R, and EA groups are shown in Figure 6C. Proteomics outcomes revealed that the Scrib expression level was markedly upregulated (upregulated ratio ≥ 1.2 and P ≤ 0.05) in the MCAO/R model group compared with the sham group (Figure 5C). Pak4 and Akt3 expression levels were notably downregulated (downregulated ratio ≤ 0.83 and P ≤ 0.05) in the MCAO/R model vs. sham group but markedly upregulated (upregulated ratio ≥ 1.2 and P ≤ 0.05) in the EA vs. MCAO/R model group (Figure 6C). Hdac2 was highly expressed (upregulated ratio ≥ 1.2 and P ≤ 0.05) in both the MCAO/R model vs. sham group and the EA vs. MCAO/R model group (Figure 6C). The Pten expression level was markedly reduced (downregulated ratio ≤ 0.83 and P ≤ 0.05) in the EA vs. MCAO/R model group but was not changed in the MCAO/R model group vs. the sham group (Figure 6C). Msh2, Nrp2, Efnb2, and Sema4d expression levels were markedly increased (upregulated ratio ≥ 1.2 and P ≤ 0.05) in hippocampal tissues of rats in the EA group compared with the MCAO/R group (Figure 6C). Jvenn analysis of differentially expressed proteins related to brain and nerve development in Supplementary Tables 7, 9 revealed that there were 20 common proteins in the MCAO/R vs. sham and EA vs. MCAO/R groups (Figure 6D). In addition, 19 or 20 proteins were specifically dysregulated in the MCAO/R vs. sham group or the EA vs. MCAO/R group, respectively (Figure 6D). The expression profiles of these proteins are shown in Supplementary Table 10. Interestingly, converse alteration trends of 15 proteins related to brain and nerve development were observed in the MCAO/R vs. sham and EA vs. MCAO/R groups among 20 common proteins between MCAO/R vs. sham group and EA vs. MCAO/R group (Supplementary Table 10). We supposed that these 15 proteins might play crucial roles in the EA-mediated protective effect against MCAO/R-induced brain injury.

The western blot assay demonstrated that the protein levels of Pak4, Akt3, and Efnb2 were markedly reduced in the hippocampal tissues of MCAO/R rats relative to the sham group, while EA treatment abrogated the inhibitory effect of MCAO/R on the expression of Pak4, Akt3, and Efnb2 (Figure 7). The western blot outcomes were consistent with the TMT proteomics results, suggesting that our proteomics data were relatively reliable.