A total of 63 male-specific pathogen-free SD rats (5–8 weeks-old; 250–280 g weight) were purchased from the Hunan Shrek Jingda Experimental Animal Co., Ltd. [Laboratory animal qualification certificate: scxk (Xiang) 2019-0004]. The rats had free access to food and water; the feeding environment was maintained in a 12-h light/dark cycle, at a temperature of 20–23°C and a relative humidity of 40–60%. All animal experiments were approved by the Animal Ethics Committee of Yunnan University of Traditional Chinese Medicine (approval no. R-062021088) and the care and use of experimental animals were in accordance with the guidelines of the National Institutes of Health. Prior the experiment, the rats were randomly categorized into sham (Sham), model (MCAO/R) and treat (MCAO/R + DBD 10 mg/kg) groups, with 21 rats in each group. DBD (≥98% purity) was purchased from Chengdu Alfa Biotechnology Co., Ltd. Based on the effective doses determined in our previous study (17), the treat group was continuously given 10 mg/kg by gavage for 7 days, whereas the model and sham groups were given the same amount of distilled water by gavage.

The MCAO/R model was replicated by the suture method as described previously (26) with slight modifications. In brief, rats were anesthetized by intraperitoneal injection of 2% sodium pentobarbital (40 mg/kg). The right common carotid artery (CCA) and vagus nerve of the rats were isolated and a 0.36-mm-diameter polynylon monofilament with a rounded tip (cat. no. 2636-50A4; Beijing Cinontech Co., Ltd.) was inserted into the middle cerebral artery from the CCA through the internal carotid artery; the insertion was stopped if a slight resistance was encountered (~18–20 mm). After 2 h of ischemia, the middle cerebral artery was reperfused by gently pulling the suture back to the CCA. The sham group underwent the same process except for suture blockage. Neurological scoring was performed at 24 h after reperfusion. During this period, none of the animals developed humane endpoint indications, such as non-feeding, dyspnea and hypothermia, or died prematurely. Neurological deficit was assessed on a scale of 0–4 points: 0 points: regular activity, no neurological deficit; 1 point, when the tail is lifted, the left forelimb is adducted and cannot be fully extended; 2 points, the body rotates to the left when crawling; 3 points, the body tilts to the left when crawling; 4 points, inability to walk spontaneously with a decreased level of consciousness (26). The inclusion criterion for this model was a neurological score of 1–3; those with scores of 0 or 4 were excluded.

After 24 h of I/R, the rats were injected intraperitoneally with 2% sodium pentobarbital (150 mg/kg) to induce deep anesthesia and no response to tail entrapment. Then the animals were rapidly sacrificed using a decapitation device. The whole brains were removed and frozen at −20°C for 20 min. Coronal sections from front to back (2 mm) were made on the brains on ice, stained with 2% triphenyl tetrazolium chloride (TTC) (cat. no. A610558; Sangon Biotech Co., Ltd.) in the dark for 20 min at 37°C and then fixed in 4% paraformaldehyde for 24 h at room temperature. The infarcted areas were analyzed with ImageJ 1.52 a (National Institutes of Health).

The prepared brain tissue paraffin sections were heated at 60°C for 3 h. Hematoxylin was administered for 2 min and then washed off; the sections were stained with eosin dye for 1 min at room temperature. The samples were dehydrated and dried with graduated alcohol series, cleared with xylene to make and sealed with neutral gum. Images were captured using an inverted phase contrast microscope (IXplore; OM Digital Solutions Corp.) under a field of view of ×400 magnification. A HE staining kit was used for histomorphological analysis (cat. no. KGA224; Nanjing KeyGen Biotech Co., Ltd.).

Brain sections were dewaxed, dried at 60°C for 60 min, dewaxed twice with xylene and hydrated with ethanol (100, 95, 80 and 75%) for 5 min each time. Proteinase K working solution (100 µl) was added dropwise to each tissue section and incubated at 37°C for 30 min. Next, 50 µl of TUNEL fluorescence detection solution was added to each sample. Then, the sample was incubated at 37°C for 60 min in the dark. Nuclei were counterstained with DAPI for 5 min at room temperature. Observed the cortex from the cerebral under a laser confocal microscope (Zeiss LSM; Carl Zeiss AG) and images captured at 400× magnification (n=3). The areas of interest were then cropped to approximately 50 µm for presentation. All analyses were performed in a blinded manner. The TUNEL detection kit (cat. no. C1090; Beyotime Institute of Biotechnology).

After 24 h of I/R, the rats were euthanized and ~1-mm³ of the brain tissue isolated from the ischemic cortex from the cerebral infarction area was collected, fixed with 4% glutaraldehyde and then post-fixed with 1% osmic acid at 4°C for 2 h. The samples were dehydrated using graduated alcohol series and acetone, embedded in epoxy resin at 60°C incubator for 48 h, sliced into 80 nm-thick sections and stained with 4% ethyl citrate uranyl for 15 min at room temperature. The ultrastructure of mitochondria was observed and randomly selected fields of view were captured at 400× magnification (n=3) using transmission electron microscopy (JEM-1400flash; JEOL Ltd.).

Based on the regions identified in a previous study (27), 200 mg of the cortical tissue of the ischemic hemisphere was collected following MCAO/R for 24 h. The blood was rinsed with normal saline, dried using a filter paper, cut into pieces and placed in a 2 ml glass homogenizer for homogenization. High-purity mitochondria were extracted using a commercially available mitochondrial extraction kit (cat. no. SM0020; Beijing Solarbio Science & Technology Co., Ltd.) according to the manufacturer's instructions and stored at −80°C. The specific operation steps were as follows: Centrifugation at 4°C, 1,000 × g for 5 min, transferring the supernatant to a new centrifuge tube at 4°C, 1,000 × g for 5 min, continuing to remove the supernatant and transferring to the new centrifuge tube, 4°C, 12,000 × g for 10 min. The supernatant, after centrifugation, contained cytoplasmic components, while the mitochondria precipitated at the bottom of the tube. After removing the supernatant, 0.5 ml Wash Buffer was added to the mitochondrial precipitation and centrifuged at 4°C 1,000 × g for 5 min. Then the supernatant was transferred to a new centrifuge tube for 10 min at 4°C and 12,000 × g. Finally, the supernatant was removed and the high-purity mitochondria were precipitated at the bottom of the tube. Subsequently, the mitochondrial purity was evaluated. Purified mitochondria and non-purified mitochondrial suspension (mitochondria not treated with wash buffer in mitochondrial extraction kit and containing other cell tissues) were mixed with 100 µl each of 0.3% Janus green B (mitochondria specific staining; cat. no. S19083, Shanghai Yuanye Bio-Technology Co., Ltd.) and 0.15% neutral red solution (lysosomal and Golgi staining; cat. no. DN300; Beijing Dingguo Changsheng Biotechnology Co., Ltd.). The purified and non-purified mitochondria samples from the cerebral cortex (n=3) were stained for 10 min at room temperature and then observed and randomly selected fields of view under an optical microscope at 400× magnification (DMi1; Leica Microsystems, Inc.).

A mitochondrial protein Extraction kit (cat. no. G008-1; Nanjing Jiancheng Bioengineering Institute) was used to extract total protein from the isolated mitochondria. Then, the protein content of each sample was determined using a BCA protein quantification kit (cat. no. PC0020; Beijing Solarbio Science & Technology Co., Ltd.). Additionally, ATP content in brain tissue was determined using an ATP assay kit (cat. no. A095-1-1; Nanjing Jiancheng Bioengineering Institute). The oxidation indexes of brain tissue and the level of reactive oxygen species (ROS) were detected by using DCFH-DA as the fluorescent probe (cat. no. E004-1-1; Nanjing Jiancheng Bioengineering Institute). Malondialdehyde (MDA) content was detected using the thiobarbituric acid method (cat. no. BC0025; Beijing Solarbio Science & Technology Co., Ltd.). For the detection of brain mitochondrial function indicators, the mitochondrial respiratory chain complex I–IV activity detection kit (cat. nos. BC0515, BC3235, BC3245 and BC0945; Beijing Solarbio Science & Technology Co., Ltd.) was used. The electron transport chain (ETC) contains four multisubunit enzyme complexes, including mitochondria complex I (nicotinamide adenine dinucleotide dehydrogenase, NADH dehydrogenase), mitochondria complex II (succinate dehydrogenase), mitochondria complex III (cytochrome c reductase) and mitochondria complex IV (cytochrome c oxidase). The mitochondrial swelling method was used to measure the opening degree of mPTP. The MMP level was measured by fluorescence spectrophotometry (cat. nos. GMS10101 and GMS10013.1, Shanghai Genmed Gene Medicine Technology Co., Ltd.). The content of cytochrome c (Cyt-c) in brain mitochondria was measured using ELISA kit (H190-1-1, NanJing JianCheng Bioengineering Inc.).

RIPA lysis buffer (PSMF:RIPA lysis buffer=1:100; cat. nos. 042121210730; 051021210825; Beyotime Institute of Biotechnology) was added to the treated brain tissue and lysed on ice for 20 min and the lysed product was centrifuged at 14,200 × g for 5 min at 4°C. Protein quantification was performed by the BCA method. Equal amounts of protein (80 µg) were separated by 8% SDS-PAGE and transferred to PVDF membranes. Electrophoresis was performed at 80 V for ~30 min. After the protein marker was separated and the sample entered the separation gel, the parameter was changed to a constant voltage of 120 V. After electrophoresis, the membrane was transferred until the target band reached the appropriate position. The membrane was blocked with 5% bovine serum albumin (SW3015, Beijing Solarbio Science & Technology Co., Ltd.) at room temperature for 1 h. After blocking, the membrane was washed twice with TBST buffer (0.1% Tween; cat. no. QN1236; Beijing Biolab Technology Co., Ltd.). Subsequently, the membrane was incubated with primary antibodies, including Bax (1:1,000; cat. no. 50599-2-Ig; ProteinTech Group), Bcl-2 (1:500; cat. no. 26593-1-AP; ProteinTech Group), Caspase-3 (1:1,000; cat. no. 9662; Cell Signaling Technology, Inc.), O-GlcNAc transferase (OGT; 1:500; cat. no. 61355, Active Motif, Inc.) and β-Actin (1:25,000; cat. no. 66009-1-Ig; ProteinTech Group, Inc.), overnight at 4°C. Then, it was incubated with goat anti-rabbit IgG and Rabbit Anti-Mouse IgG secondary antibodies (both at 1:5,000, cat. nos. ab6721 and ab97046, Abcam) for 1 h at room temperature. The immunoreactive bands were developed using enhanced chemiluminescence reagent (cat. no. A38555; Thermo Fisher Scientific, Inc.) and images were captured in the Bio-Rad ChemiDoc XRS gel imaging system (Bio-Rad Laboratories, Inc.). ImageJ Lab V4.0 software (Bio-Rad Laboratories, Inc.) was used for quantitative analysis.

After 2 h of ischemia and 24 h of reperfusion, the rats were sacrificed and 200 mg of the cortical tissue samples of the infarcted hemisphere was quickly collected and stored at −80°C for later use. A total of six brain tissue samples were collected from each group. After the sample was thawed and homogenized, 0.05 g of the sample was mixed with 500 µl of 70% methanol/water. The sample was vortexed for 3 min at of 2,500 rpm and centrifuged at 14,200 × g for 10 min at 4°C. Afterwards, 300 µl of supernatant was collected into a new centrifuge tube and placed in a −20°C refrigerator for 30 min, Then, the supernatant was centrifuged again at 14,200 × g for 10 min at 4°C. After centrifugation, 200 µl of supernatant was transferred through protein precipitation plate for further LC-MS analysis. Each group of 20 µl samples was mixed as quality control (QC) samples. Analyzing the coefficient of variation (CV) of QC samples, the CV value indicates the ratio of the standard deviation of the original data to the mean of the actual average data, reflecting the degree of data dispersion. Furthermore, indicating the stability of experimental data was tested to ensure QC.

Linear ion trap and triple quadrupole scans were acquired using a triple quadrupole-linear ion trap mass spectrometer (QTRAP), QTRAP 6500+ LC-MS/MS System (Sciex), operating in positive and negative ion mode. The ESI source operation parameters were as follows: ion source, ESI+/-; source temperature, 550°C; ion spray voltage 5,500 V (positive) and −4,500 V (negative); curtain gas was set at 35 psi. Tryptophan and its metabolites were analyzed using scheduled multiple reaction monitoring. Data acquisitions were performed using Analyst 1.6.3 (Sciex). Multiquant 3.0.3 (Sciex) was used to quantify all metabolites.

Unsupervised principal component analysis (PCA) was performed by statistics function prcomp V1.0.1 (r-project.org) within R (www.r-project.org). Data were unit variance scaled before unsupervised PCA. The hierarchical cluster analysis results of samples and metabolites were presented as heatmaps with dendrograms using R package pheatmap V1.2.1 (r-project.org). Significantly regulated metabolites among groups were determined by variable importance in projection (VIP) and absolute Log₂FC (fold change). VIP-values were extracted from orthogonal partial least squares discriminant analysis (OPLS-DA) result and generated using R package MetaboAnalystR V1.0.1 (28). P-value of the unpaired student's test determined their significance. Metabolites were annotated using the KEGG database and then mapped to the KEGG pathway database (http://www.kegg.jp/kegg/pathway.html). Then, the pathways to which the metabolites with significant modulation were mapped were fed into metabolite set enrichment analysis and the P-value of the hypergeometric test determined their significance.

GraphPad Prism 9.0.0 (GraphPad Software, Inc.) was used for statistical analyses. The data conform to the normal distribution. If the variance was homogeneous, then it was analyzed with Bonferroni's multiple comparisons test using one-way analysis of variance (ANOVA). If the variance is unequal, then it was analyzed Dunnett's T3 multiple comparisons test in Welch's ANOVA test. Non-normally distributed data were analyzed with Dunn's multiple comparisons test in the Kruskal-Wallis test. All values were presented as mean ± standard error of the mean. P<0.05 was considered to indicate a statistically significant difference.

TTC staining reflects cerebral infarction after injury and the area of ischemic necrosis is stained white (29). The results showed that obvious white infarcts in the model group after cerebral I/R and DBD could significantly reduce the infarct volume (Fig. 1B and C). Compared with the model group, the neurological function of the rats in the DBD group was significantly improved (Fig. 1D). Among them, the ATP content in the brain tissue of MCAO/R rats was significantly decreased, but this was reversed by DBD (Fig. 1E). In addition, the levels of oxidative stress-related indicators ROS and MDA increased following I/R, while DBD could significantly inhibit the generation of ROS and MDA (Fig. 1F and G). DBD effectively attenuated brain injury and oxidative stress in MCAO/R rats while restoring energy supply.

At the heart of mitochondrial dysfunction is ATP reduction and to determine the effect of DBD on mitochondrial damage, follow-up experiments were performed. Observation of mitochondrial ultrastructure in neurons from the cortex tissue was performed using transmission electron microscopy. It was shown that I/R resulted in severe swelling of mitochondria and partial disappearance of cristae, rupture of inner and outer membranes (the matrix is electron transparent). By contrast, DBD ameliorated this injury, leaving the inner and outer membranes and ridges largely intact with only mild swelling (Fig. 2A). Then, high-purity mitochondria were isolated by gradient centrifugation for detection of related indexes. Microscopic examination revealed a number of dark reds in non-purified mitochondria and a single blue-green color in purified mitochondria (Fig. S1). These results indicated that the purity of the extracted mitochondria was high. Following brain I/R, the activity of mitochondrial respiratory chain enzyme complexes I–IV was inhibited. Mitochondrial respiratory chain enzyme complexes I and II can oxidize NADH produced by the TCA cycle and the oxidation of succinate, respectively and donate electrons to ETC (30). Complex III reduces the mobile electron carrier Cyt-C, ferries single electrons that are subsequently transferred from complex III to complex IV, where molecular oxygen is bound and reduced to water (31). This electron transfer chain provides an electrochemical gradient across the inner mitochondrial membrane to drive ATP synthesis. In addition, ROS are mainly formed by premature leakage of electrons from complexes I, II and III (31). mPTP was abnormally opened and MMP level and mitochondrial Cyt-c content were decreased. However, BDB had a positive effect on mitochondrial health in MCAO/R rats, manifested by increased activity of mitochondrial respiratory chain enzyme complexes I–IV (Fig. 2B-E), restored mPTP levels (Fig. 2F), addition to decreased MMP loss (Fig. 2G) and increased mitochondrial Cyt-c content (Fig. 2H). Experiments showed that DBD alleviated cerebral I/R injury, which may restore cellular energy supply by improving mitochondrial dysfunction, thereby inhibiting the transfer of mitochondrial Cyt-c to the cytoplasm, reducing apoptosis and protecting neurons.

Evaluation of QC samples showed that the proportion of substances with CV values less than 0.5 in the test and QC samples were higher than 85%, indicating that the experimental data was stable (Fig. 3). To understand the changes on metabolites in I/R of DBD treatment more clearly, the sham group cortical tissue of rats and model and treat cerebral cortical tissue of the ischemic hemisphere (n=6) were collected for metabolic study. Analyst 1.6.3 software (Sciex) was used to process LC-MS data and to obtained the total ion chromatogram of all samples (Fig. S2, Fig. S3, Fig. S4). In addition, a total of 57 metabolites were detected based on the UPLC-MS/MS detection platform and database, including 17 nucleotides and their metabolites, 12 amino acids, 10 Organic acids and their derivatives, eight phosphate sugars, five phosphoric acids, three coenzymes and vitamins, one lysophosphatidylethanolamine and amino acid derivatives.

Principal component analysis and heatmap clustering was used to examine all metabolites in the data. The PCA results showed clear separation among the different treatments (Fig. 4A), indicating that the metabolites of samples changed significantly following MCAO/R, consistent with the I/R phenotype. The change was mainly dominated by the first principal component (PC1) of the X-axis, which can explain 45.82% of the characteristics of the original data set. From PC1, it can be found that the model group and the treatment group are separated and that there are differences between them.

The differences in the accumulation patterns of metabolites in the sham-operated group, the model group and the treatment group can be analyzed by clustering heatmaps (Fig. 4B). The results showed significant differences in substances in different groups, which were divided into two clusters. The metabolites in cluster 1 were the highest in the sham operation group, medium in the treatment group and the model group. The lowest content, including most nucleotides and their metabolites, such as adenosine monophosphate (AMP), ATP, adenosine diphosphate (ADP), guanosine and uracil, decreased following MCAO/R, which may result from inhibited energy production. The metabolites in cluster 2 were highest in the model group, moderate in the treatment group and lowest in the sham group. Among them, the increased expression of phosphoenolpyruvic acid in the gluconeogenesis pathway after IS was mainly found in astrocytes, which can promote acidosis and oxidative damage, resulting in disturbance of glucose metabolism during I/R (32). The same excess of 3-phenyllactic acid (PLA) has also been shown to induce ROS production in glial cell (33). In addition, ornithine, L-citrulline, L-leucine and other essential amino acids increased in MCAO/R, which may be the accumulation of protein synthesis disorders. The different biological replicates also clustered together, indicating good homogeneity between biological replicates and high data reliability.

OPLS-DA analysis is a multivariate statistical analysis method with supervised pattern recognition, which can effectively eliminate influences irrelevant to the present study to screen differential metabolites. OPLS-DA was used to perform a pairwise analysis of the three groups to draw a score map. In this model, R²X and R²Y represented the interpretation rate of the X and Y matrices of the built model, respectively, and Q² represented the predictive ability of the model. Q² were all higher than 0.9, P<0.05, indicating that the constructed model was suitable (Fig. 5A). The OPLS-DA score plot showed that there was a clear separation between the model group and the treatment group (Fig. 5B). Fold Change (FC) and VIP was used to screen differential metabolites, metabolites satisfying FC ≥2 and FC ≤0.5 and VIP ≥1 were considered to have significant differences and the screening results were displayed in a volcano plot (Fig. 5C). The results showed that in the comparison between the model group and the treatment group, there were a total of nine differential metabolites, of which eight metabolites were upregulated: Succinyl-CoA, UDP-GlcNAc, deoxythymidine monophosphate (dTMP), AMP, inosine monophosphate (IMP), uridine 5′-monophosphate (UMP), fructose 1,6-bisphosphate, 6-phosphogluconic acid; 1 metabolite was downregulated: 3-phenyllactic acid (Fig. 5D; Table I).

To comprehensively observe changes in metabolic pathways, the present study employed a pathway-based metabolic change analysis method, Differential Abundance (DA) Score, which captures the overall changes of all metabolites in a pathway (Fig. 5E). The differential metabolites between the model group and the control group were mainly annotated and enriched in fructose and mannose metabolism, tropane, piperidine and pyridine alkaloid biosynthesis, amino sugar and nucleotide sugar metabolism and pentose phosphate pathway. These pathways interact and are closely related to I/R. Based on the results described above, a metabolite overview of key pathways in brain tissue is shown in (Fig. 6).

The most significantly differential metabolite induced by DBD was UDP-GlcNAc and OGT is the key catalytic enzyme of UDP-GlcNAc. Subsequently, the content of UDP-GlcNAc in the brain detected by metabolomics was visualized and the protein expression of OGTase detected. The experimental results showed that compared with the sham-operated group, the range of UDP-GlcNAc in the model group was reduced, which could be improved by DBD (Fig. 7A). Western blotting results showed that the expression of OGT was significantly increased following DBD treatment compared with the model group (Fig. 7B). This indicated that DBD increased the level of UDP-GlcNAc in the brain after IS by promoting the expression of OGT.

Apoptosis is the final and primary determinant of brain I/R injury. In HE staining, the brain tissue cells of the rats in the Sham group were arranged neatly and their morphological structures were normal. In the model group, the gap of cell structure was widened, nuclear pyknosis appeared, and the fiber arrangement was disordered, while the treat group exhibited improved pathology of brain tissue (Fig. 8A). TUNEL staining showed that MCAO/R decreased the number of neurons and increased the number of positive apoptotic cells. Following DBD treatment, apoptotic cells were significantly reduced (Fig. 8B and C). In addition, the protein expression of genes related to apoptosis was analyzed by western blotting (Fig. 8D-H). MCAO/R induced the up-regulation of Bax and cleaved-Caspase-3 and the downregulation of Bcl-2. However, DBD partially reversed MCAO/R-induced apoptosis compared with the MCAO/R group. It was hypothesized that DBD inhibits MCAO/R-induced brain cell apoptosis and reduces neuronal loss caused by cerebral I/R injury.