A total of 48 male, pathogen-free, Sprague-Dawley rats (15) [8–12 weeks old, 180–220 g in weight, provided by Changzhou Cavens Experimental Animal Co., Ltd.; laboratory animal certificate no. SCXK (Su) 2021–0013] were used in the present study. Rats were raised in a specific pathogen-free environment maintained at a temperature of 21–25°C and relative humidity of 40–60%, and rats were provided ad libitum access to food and water. All animal experiments were approved by the Animal Ethics Committee of Yunnan University of Traditional Chinese Medicine (approval no. R-062019039). All rats were handled according to the Guidelines for the Care and Use of Laboratory Animals of the National Institute of Health (16). Before the experiments, the rats were randomly divided into a sham group (Sham), a model group (MCAO/R), or a PHBA group (MCAO/R+PHBA) with 18 rats per group. PHBA (≥98% purity) was purchased from Chengdu Alpha Biotechnology Co., Ltd. According to our previous study (17) the effective dose for the treatment group was 20 mg/kg PHBA and via gavage for 7 days, while the model and Sham groups were administered an equivalent volume of distilled water.

Although the risk factors for stroke in older rats are similar to those in humans (15), aged rats were excluded due to anatomical/pathological changes such as middle cerebral artery and common carotid artery aberrations and luminal stenosis (15). Based on a previous preparation method (10), emboli were prepared using a 0.26 mm in diameter nylon thread (Ruibo Biotechnology Co., Ltd.), and then placed in heparin sodium (cat. no. 152104037A; Qianhong Biopharma Co., Ltd.). The rats were anesthetized with a small animal anesthesia machine (ZS-MV–IV; Beijing Zhongshidichuang Science and Technology Development Co., Ltd.) using 5% isoflurane (RWD Life Science Co., Ltd.) and maintained with 3% isoflurane, and fixed in the supine position on the experimental table.

The left common carotid artery (CCA) was isolated under a stereomicroscope through a median cervical incision. The left external carotid artery (ECA) and internal carotid artery (ICA) were separated upward, and the two external carotid artery branches of the superior thyroid artery and occipital artery were ligated to reduce the error of thread insertion. The ECA was doubly ligated at 5–8 mm from the near CCA bifurcation and occluded with a micro artery clip at the proximal ends of the ICA and CCA. A 0.2 mm diameter V-shaped microincision was made at the proximal ligation of the ECA. After the nylon thread bolt was gently inserted, the nylon thread was tightened, the artery clamp was loosened, and the nylon thread was moved into the brain along the ECA via the ICA. The insertion depth was 18–20 mm and insertion was halted if slight resistance was felt. The head end of the nylon thread was used to block the blood flow of the MCA. Then, the incision was sutured, and the ischemia time was recorded. After 2 h, the slow and light pulling of the bolt caused the head end to return the CCA, at which point the cerebral I/R was considered to be realized. Samples were obtained from the rats 24 h after reperfusion. In the Sham group: the carotid artery was exposed and isolated without I/R. The body temperatures of all rats were maintained at 37°C throughout the procedure. During this time, none of the animals exhibited symptoms that would require termination, such as not eating, breathing difficulties, convulsions, or hypothermia, nor did any die prematurely. All rats in the study were euthanized before the end of the study.

After 24 h, the degree of neurological deficit was evaluated using a score of 0–4 (18) as follows. 0, no neurological deficit; 1, left forelimb flexion when lifting the tail in the air; 2, rats walked in circles; 3, hemiplegia causing rats to crawl; and 4, rats were unable to walk spontaneously and had a decreased level of consciousness. A higher score indicated a more serious behavioral disorder.

A total of 24 h after brain I/R, rats were injected intraperitoneally with 2% pentobarbital sodium (150 mg/kg) to induce deep anesthesia. After no response to a tail clip, the rats were euthanized by decapitation. and death was confirmed by a lack of response to tail clamping. The brain was frozen at −20°C for 20 min. The brain was sectioned from front to back with 2 mm coronal sections on ice, stained with 2% TTC staining solution at 37°C for 20 min, and then fixed in 4% paraformaldehyde at room temperature for 24 h. The experimental results were analyzed using ImageJ version 1.52a (National Institutes of Health) software to obtain the total infarcted area and the total area of brain slices. The infarct rate (%) was calculated as the total infarct area/total area of brain slices.

H&E staining was used to show the components of cytoplasm, muscle fibers, and the general morphological structure of lesions. Nissl is an important structure involved in protein synthesis in neurons. When neurons are stimulated, the Nissl bodies in the cell become disordered and crumpled. Therefore, detection using H&E staining and Nissl staining provides mutual verification of the pathological injury status of brain tissue (19). Brain tissue sections embedded in paraffin (thickness 5 µm, n=3) were heated at 60°C for 3 h, stained with a hematoxylin (using an H&E staining kit; cat. no. KGA224; Nanjing KeyGen Biotech Co., Ltd.) for 2 min, and then washed. Eosin staining was performed for 1 min at room temperature. The samples were dehydrated using a gradient of alcohol solutions, made transparent using xylene, and then sealed with neutral gum. An inverted bright field phase contrast microscope (IX83, Olympus Corporation) was used to obtain the hippocampal images. Fields of view were randomly selected and images were taken at a magnification of ×400, and the images were analyzed using ImagePro Plus version 6.0 (Media Cybernetics, Inc.).

Brain tissue sections embedded in paraffin (thickness 5 µm, n=3) were dewaxed, incubated in Nissl staining solution (cat. no. G1430; Beijing Solarbio Science & Technology Co., Ltd.) at 50–60°C for 40 min, washed with deionized water and differentiated using the included differentiation solution. Then, anhydrous ethanol was used for dehydration, xylene was used to clear the tissue, and neutral gum was used for sealing. Images of randomly selected fields of view of the hippocampus were obtained using a phase contrast microscope at a magnification of ×400 and were analyzed using ImagePro Plus 6.0 (Media Cybernetics, Inc.).

Whole blood was collected with a heparin sodium anticoagulant tube and centrifuged at room temperature at 3,000 × g for 10 min. After the blood cells settled entirely to the bottom of the tube, the upper plasma was taken. Plasma (50 µl) from the model group and PHBA group was added to a methanol solution containing an internal standard (4-chlorophenyl alanine, 1 µg/ml) and was then shaken, mixed for 5 min, and centrifuged at 4°C (14,200 × g for 10 min). The supernatant (200 µl) was transferred to a centrifuge tube and dried using a vacuum drier. The supernatant was then redissolved in 100 µl ultra-pure water and methanol (1:1) and centrifuged at 4°C (14,200 × g for 10 min). The supernatant volume was 80 µl, and the sample volume injected was 5 µl.

The LC-QTOF/MS conditions were: Waters HSS T3 1.8 µm, 2.1×100 mm column; flow rate, 0.3 ml/min; column temperature box, 40°C; mobile phase, water phase (ultra-pure water + 0.1% formic acid); B organic phase (acetonitrile). The mass spectrometry conditions were: Mass detection was performed in the positive ion (4,000 V) and negative ion (−4,500 V) mode, respectively, and scanned using a Turbo V electrospray ionization (ESI). The parameters were set as follows: Ion spray voltage, 7 kV; turbine spray temperature, 500°C; declustering potential, 70 V; collision energy, 30 eV; atomizer gas, 55 psi; heater gas, 55 psi; curtain gas, 35 psi. The atomizer and auxiliary gas are maintained by nitrogen. The scanning range of the TOF MS was mCompz 100–1,200.

Taking the mixed solution as the QC sample, in the process of instrumental analysis, a quality control sample was inserted every 10 tests and analysis samples. Through the overlapping display and analysis of the TIC chromatogram of the essential spectrum detection and analysis of the same quality control sample, the stability of the instrument during the testing was judged.

The LC-MS data were processed using MS-DIAL (version 5.1.230719) (20). The ‘mixOmics’ (11) package in R (21) (www.r-project, version 4.2.1) was used for partial least squares discriminant analysis (PLS-DA). The data were unit variance scaled before analysis. The results of the systematic cluster analysis of samples and metabolites are presented as a heat map and tree map generated using the R package ‘pheatmap’ (11). The variable importance projection (VIP) value was extracted from the PLS-DA results. An unpaired Student's t-test was used to determine the significance. Metabolites that were significantly differentially regulated between groups were determined using the VIP and absolute log₂FC. The results are presented as a volcanic map (‘ggplot2’ package) (22). The KEGG database was used to annotate metabolites, which were then mapped to the KEGG pathway database, and the Metaboanalyst website (http://www.metaboanalyst.ca/) was used for metabolic pathway analysis. Then, the pathways to which the metabolites with significant regulatory effects were mapped and metabolite concentration were analyzed. The P-value of the hypergeometric test and the size of the pathway influence factors were used to determine the importance of the metabolites.

First, the hippocampal tissues were lysed using RIPA cleavage buffer (PSMF:RIPA lysate=1:100; cat. no. 051021210825; Beyotime Institute of Biotechnology) on ice for 20 min, after which the lysate was centrifuged at 4°C at 13,017 × g for 5 min. The protein concentration was quantified using a bicinchoninic acid assay. Samples of 80 µg protein were loaded on an 8% SDS gel, resolved using SDS-PAGE at 80 V for ~30 min, followed by 120 V until the target band reached a suitable position. The membrane was then transferred to a PVDF membrane, which was subsequently blocked for 1 h with 5% skimmed milk in TBST at room temperature. Membranes were and then incubated with the primary antibody diluted in TBST solution overnight at 4°C. The primary antibodies used were anti-PSD-95 (1:1,000; cat. no. ab238135; Abcam), anti-Bcl-2 (1:1,000; cat. no. sc-7832; SantaCruz Biotechnology, Inc.), anti-GluA1 (1:1,000; cat. no. 13185; Cell Signaling Technology, Inc.), anti-caspase 3 (1:1,000; cat. no. 9662; Cell Signaling Technology, Inc.), light chain-3 protein, LC3 II/I (1:1,000; cat. no. 4108; Cell Signaling Technology, Inc.), Bax (1:2,000; cat. no. 50599-2-Ig; ProteinTech Group, Inc.), sequestosome-1 (SQSTM1/P62) (1:5,000; cat. no. 18420-1-AP; ProteinTech Group, Inc.), and autophagy effector protein 1 (Beclin1) (1:2,000; 66665-1-Ig; ProteinTech Group, Inc.). The membranes were washed using TBST (0.05% Tween) buffer then incubated with the secondary Goat Anti-Rabbit IgG antibody (cat. no. ab6721) or Rabbit Anti-Mouse IgG (cat. no. ab6728) at room temperature for 1 h (1:10,000; Abcam). Signals were visualized using an ECL kit according to the manufacturer's protocol (Thermo Fisher Scientific, Inc.) and a Tanon 6600 luminous imaging workstation (Tanon Science & Technology Co., Ltd.). Image ProPlus was used for densitometry analysis.

TUNEL staining was performed using a TUNEL detection kit (cat. no. C1090, Beyotime Institute of Biotechnology). After heating at 60°C for 60 min, xylene was used to dewax the brain slices twice, after which samples were hydrated in a series of ethanol solutions (100, 95, 80, and 75%), with 5 min per solution. A total of 20 µg/ml protease K without DNA enzymes was added to each tissue section and allowed to react at 37°C for 30 min. Subsequently, 50 µl TUNEL fluorescence detection solution was added to each sample, which was then incubated at 37°C for 60 min, and DAPI double staining was performed at room temperature for 5 min. The hippocampus was observed under a laser confocal microscope (Zeiss LSM). Randomly selected fields of view were captured at ×400 magnification.

Using GraphPad Prism Version 9.0.0 (GraphPad Software, Inc.) for analysis and processing, if the data were normally distributed and the variance (ANOVA F-test) was uniform, a Bonferroni's multiple comparison test after a one-way ANOVA was used. Data are presented as the mean ± the standard error. P<0.05 was considered to indicate a statistically significant difference.

The results of TTC staining showed that there were notable white infarcted areas following brain I/R. Compared with the model group, PHBA significantly reduced the size of infarct sizes [F (2,6) 1.822, P=0.413] (Fig. 1A and B) and improved the neurological function of rats [F (2,12) 1.333, P=0.300] (Fig. 1C). It has been shown that PHBA can effectively reduce brain injury in MCAO/R rats (10). To further explore the effects of PHBA on the pathological changes of the brain following I/R injury, HE and Nissl staining were used in this study. The results showed that the number of cells [F (2,15) 1.273, P=0.529] in the sham group was higher, the nuclei and cell bodies of neurons in the sham group were clear, and the Nissl bodies [F (2,15) 0.121, P=0.887] were abundant. Conversely, in the model group, hippocampal ischemic neuronal injury, nuclear pyknosis, a disorderly arrangement of neurons, and sparsity of Nissl bodies were observed. PHBA reversed these changes (Fig. 1D-G). The above data suggest that PHBA had a protective effect on hippocampal neuronal death after I/R.

The high stability of the LC-QTOF/MS instrument provides an essential guarantee for the repeatability and reliability of data. The curve overlaps of the total ion currents of the QC samples were high, suggesting that the experimental data were stable (Fig. 2). To ascertain the changes of metabolites following PHBA treatment, the plasma of rats in the model and PHBA groups (n=6) were collected. In addition, based on the LC-QTOF/MS detection platform and database, 856 metabolites were detected under positive and negative ions, including purines, pyrimidines, amino acids, and other metabolites. To further determine the differences between groups, the multivariate statistical analysis method PLS-DA was used to analyze the two groups in pairs. The results showed there was a significant separation between the model and the PHBA group, and the internal correlation between each group was very high (Fig. 3A and B). The screening results are shown as a volcano map (Fig. 3C and D). In addition, the stratified cluster heat map was used to show the changes in metabolites more directly. The heat map showed that the model and PHBA group could be divided into two parts according to the identified metabolites (Fig. 3E and F). Finally, 13 differential metabolites were identified, of which 6 metabolites were upregulated: benzamide, pyroglutamic acid, fumaric acid, d-Xylulose, cholic acid, and cis-Aconitic acid, and 7 metabolites were downregulated: 2′-Deoxyguanosine 5′-monophosphate (dGMP), hexadecanedioic acid, deoxycholic acid, deferoxamine, 4-Nitrophenyl phosphate, bilirubin, leukotriene C4 (Table I). Among these metabolites, the P-value of dGMP was the lowest (0.018).

To comprehensively observe the changes in metabolic pathways, pathway enrichment analysis of all differential metabolites was performed. In positive ion mode, the primary ways that the model group and the PHBA group were affected by the differential metabolites were: The Tricarboxylic acid (TCA) cycle', ‘glutathione metabolism’, and ‘mutual transformation of pentose and glucuronates’ (Fig. 4A). This may be related to the improvement of energy metabolism and the antioxidant effect of glutathione metabolic pathway by PHBA, which has been reported in the pathogenesis of glucose metabolic diseases (23). In negative ion mode, the primary pathways affected by the differential metabolites were: The ‘TCA cycle’ and ‘arachidonic acid metabolism’, amongst others (Fig. 4B). This suggests that MCAO/R inhibits energy production and induces inflammation. These pathways interact with each other and are closely related to I/R. From the above results, it can be inferred that PHBA components are absorbed into the blood and achieve their neuroprotective effects by regulating various metabolic processes in the body, especially those related to the TCA cycle (Fig. 4C).

The most significantly different metabolite induced by PHBA was dGMP. Guanosine kinase (GK) can phosphorylate dGMP to dGDP as a central enzyme in the guanine rescue pathway (24). The changes in PSD-95 levels in the membrane-associated guanosine kinase (MAGUK) family are related to the content of synaptic AMPAR (GluA1), which regulates the intensity of hippocampal synaptic activity and is associated with neuroplasticity (25). Here, the protein expression levels of PSD-95 and GluA1 were determined. Western blotting results showed that the expression of PSD-95 [F (2,6) 0.035, P=0.966] and GluA1 [F (2,6) 1.684, P=0.263] increased significantly compared with the model group following PHBA treatment (Fig. 5A-C). This suggested that PHBA increased the levels of dGMP in the brain following IS by promoting the interaction between PSD-95 and AMPAR.

To verify whether PHBA ameliorates I/R by activating autophagy, western blotting was used to detect the effects of PHBA on the expression of the autophagy marker LC3 and essential autophagy proteins p62 and Beclin1 (Fig. 6A). The results showed that compared with the model group, PHBA significantly increased the expression of LC3-II/LC3I [F (2,6) 0.463, P=0.650] and Beclin1 [F (2,6) 1.730, P=0.255] (Fig. 6B and C), and decreased the protein expression levels of p62 [F (2,6) 0.103, P=0.903] (Fig. 6D). This suggests that autophagy in the hippocampus is significantly activated following PHBA treatment, which may be related to the interaction between PSD-95 and AMPAR.

It is well-established that brain I/R is often associated with neuronal injury and apoptosis (18,19). TUNEL staining showed that the number of positive cells in the hippocampus of the model group was significantly higher than that of the sham group. After treatment with PHBA, the number of positive cells in the hippocampus decreased significantly [F (2,6) 1.169, P=0.373] (Fig. 7A and B). In addition, the protein expression levels of apoptosis-related genes were analyzed by western blotting (Fig. 7C-G). MCAO/R resulted in the upregulation of Bax [F (2,6) 1.207, P=0.363] and cleaved-caspase-3 [F (2,6) 0.349, P=0.719] expression and downregulation of Bcl-2 [F (2,6) 0.734, P=0.519] expression. However, PHBA partially reversed the MCAO/R-induced apoptosis. These results suggested that PHBA inhibited I/R-induced apoptosis of hippocampal cells and reduced the loss of neurons.