A total of 90 male 6-week-old specific pathogen-free (SPF) Kunming mice weighing 18-22 g were obtained from Jinan Pengyue Experimental Animal Breeding Co., Ltd. (Laboratory animal certificate: SCXK 2014-0007). The animals were housed in cages with an ambient temperature of 23-25°C and a humidity of 35-45%, under a 12-h light/dark cycle beginning at 08:00; standard mouse chow and water were provided ad libitum. The present study was performed in accordance with the recommendations of the Ministry of Science and Technology of China's Guidance for the Care and Use of Laboratory Animals. All procedures were approved by the Laboratory Animal Ethics Committee of Henan University of Chinese Medicine (the number of approval ethics code: DWLL201903026).

Using high-performance liquid chromatography, the purity of Andro provided by Shanghai Yuanye Bio-Technology Co., Ltd. (chemical structure presented in Fig. 1) was determined to be 98%. The maximum dose of Andro used in the present study (120 mg/kg; 12 mg/ml) was selected based on a previous study (15) and preliminary experiments (data not shown). A 200 mg/ml solution of Andro dissolved in dimethyl sulfoxide was prepared and diluted in distilled water to the appropriate concentration prior to use. The animals were randomly assigned to 5 groups as follows: The sham-operated group (n=10), repeated CIRI group (n=20) and 3 Andro groups (30, 60 and 120 mg/kg) (n=20). Following randomization, the corresponding drugs were administered orally, with the sham-operated group and CIRI group being treated with an equal volume of normal saline. At 1 h after the final administration on the 14th day, the mice in each group were subjected to repeated CIRI surgery.

Repeated CIRI was performed through bilateral common carotid artery occlusion (BCCAO) according to a previously published method (16). In brief, anesthesia was initialized with 4% isoflurane inhalation and then maintained with 1.5% isoflurane in 70% nitrous oxide (N₂O)/30% oxygen (O₂) using a laboratory anesthesia system (Matrix VIP 3000; Midmark Corp.). The bilateral common carotid arteries (BCCAs) were exposed, the carotid artery was lifted with 4-0 silk sutures, and the BCCAs were ligated with artery clamps for 10 min. The artery clamps were then removed and the BCCAs were visually inspected for reperfusion. After the blood flow had been restored for 10 min, the BCCAs were again ligated for 10 min. The sham-operated mice received the same procedure, but without carotid artery ligation. Wounds were sutured with a 4-0 silk suture. During the surgery, the temperature was maintained at 37±0.5°C. To ameliorate the suffering of animals observed throughout the experimental period, the animals were euthanized with 4% isoflurane inhalation followed by the dislocation of the cervical vertebra. Animal death was confirmed by respiratory and cardiac arrest, and no righting reflex. The brain tissues were removed (n=3), then fixed in 4% paraformaldehyde and embedded in paraffin prior to hematoxylin and eosin (H&E), Nissl and immunofluorescence staining. The other brain tissues of the mice were cryopreserved in liquid nitrogen for western blot analysis, reverse transcription-quantitative polymerase chain reaction (RT-qPCR) and biochemical detection.

Mouse brains were excised and fixed overnight in 4% paraformaldehyde at 4°C, embedded in paraffin and serial 5-µm-thick sections were prepared. The sections were stained with H&E (C0105, Beyotime Institute of Biotechnology) to assess brain histology. Briefly, the sections were stained with hematoxylin for 5 min, then washed with running water for 5 min and differentiated in 1% acid alcohol for 10 sec at room temperature. The sections were then stained with 1% eosin for 1 min, then washed with running water and dehydrated in a decreasing concentration of alcohol (70, 80, 90 and 100%) and cleared in xylene. Finally, the sections were mounted with neutral resin and observed under a microscope (Olympus, BX63; Olympus Corporation).

The brain tissue sections were stained with Nissl staining solution (C0117, Beyotime Institute of Biotechnology) for 30 min at 37°C. The brain sections were then washed with 95% ethyl alcohol and observed under a microscope (Olympus, BX63; Olympus Corporation). Large cell bodies with abundant cytoplasm and significant levels of Nissl bodies represent normal neurons. Cells with pyknosis or blurring of Nissl bodies represent damaged cells.

Serial 5-µm-thick paraffinembedded sections were baked at 60°C for 30 min, deparaffinized with xylene and rehydrated. The sections were then submerged into the citrate antigen retrieval buffer, heated in a microwave for antigen retrieval, treated with 3% hydrogen peroxide to quench endogenous peroxidase activity and incubated with 1% bovine serum albumin to block non-specific binding. The sections were incubated with primary antibody [glial fibrillary acidic protein (GFAP), 1:500, GB12096, Wuhan Servicebio Technology Co., Ltd.; neuronal nuclei (NeuN), 1:500, GB11138, Wuhan Servicebio Technology Co., Ltd.] overnight at 4°C. After washing, the tissue sections were treated with fluorescence-conjugated secondary antibody (G1213-100UL, Wuhan Servicebio Technology Co., Ltd.) in the dark (37°C) for 2 h. After washing with PBS, the sections were immediately examined under a fluorescence microscope (Olympus, BX63; Olympus Corporation).

The infarct sides of brain tissues following repeated CIRI surgery were used for assessing expression brain-derived neurotrophic factor (BDNF) and pro-inflammatory cytokines. The levels of BDNF and pro-inflammatory cytokines in the supernatant of the brain homogenates or serum were detected by enzyme-linked immunosorbent assay (ELISA).

The brain tissues of the mice were removed from the liquid nitrogen and placed in a pre-cooled mortar that had been cleaned and dried at high temperature in advance. The liquid nitrogen was poured into the mortar, and the tissues were ground into a powder in the liquid nitrogen. Approximately 100 mg of brain tissue powder were weighed and placed in a 5-ml EP tube, and 1 ml of RIPA lysis buffer containing 1% PMSF and 1% phosphatase inhibitor was added. Extraction with a syringe was performed until the tissue was completely fragmented completely and in full contact with the lysis buffer. The lysate was then placed on ice for 30 min. Lysates were centrifuged at 12,000 × g for 15 min in a high-speed centrifuge at 4°C, and the supernatant was collected to obtain total protein. The protein concentration was measured using a BCA protein assay kit (Beijing Solarbio Science & Technology Co., Ltd.). The samples were then separated on 10% SDS-PAGE gels and electro-transferred to polyvinylidene fluoride (PVDF) membranes. The membranes were blocked in Tris-buffered saline (TBS) with 5% non-fat milk for 2 h at room temperature, then incubated with primary antibodies [tropomyosin receptor kinase B (TrkB), 1:800, GB11295-1, Wuhan Servicebio Technology Co., Ltd.; GFAP, 1:1,000, GB12096, Wuhan Servicebio Technology Co., Ltd.; NeuN, 1:800, GB11138, Wuhan Servicebio Technology Co., Ltd.; p-PI3K, 1:1,000, 17366, Cell Signaling Technology, Inc.; PI3K, 1:800, 4257, Cell Signaling Technology, Inc.; Akt, 1:800, 4685, Cell Signaling Technology, Inc.; p-Akt, 1:1,000, 4058, Cell Signaling Technology, Inc.] at 4°C overnight. Subsequently, the membranes were washed with Tris-buffered saline/Tween-20 (TBST) and incubated with secondary antibodies, then washed with TBS. The secondary antibody used was IRDye800CW goat anti-rabbit antibody (1:20,000; LI-COR Biosciences). Finally, labeled proteins were detected using a near-infrared imaging system (LI-COR Biosciences) and analyzed using an image analyzer (Image Studio™ Software). The densitometric values were normalized to the corresponding GAPDH densitometry values.

Total RNA was extracted from the brain tissues using the TRIzol reagent (R0016, Beyotime Institute of Biotechnology). RNA concentration of the samples was determined by Natodrop 2000 (Thermo Fisher Scientific, Inc.). The total RNA reverse transcription was performed using the PrimeScript^® RT Master Mix Perfect Real-Time kit (RR036A, Takara Bio, Inc.). In the following, cDNA was taken as the template for real-time PCR with TB Green^® Premix Ex Taq (RR420A, Takara Bio, Inc.). The steps for qPCR included: Initial denaturation by incubating in 95°C for 30 sec; 40 qPCR cycles by 95°C for 5 sec and then 60°C for 30 sec in each cycle. Each sample was determined triply and relative expression levels were analyzed using the 2^−ΔΔCq method (17). β-actin was used for normalization. The sequences of the primers are presented in Table I.

Network pharmacology analyses included the following: i)

Establishment of the drug database. Traditional Chinese Medicine Systems Pharmacology (TCMSP) was used to screen, predict and collect the active ingredients of Chuanxinlian using oral bioavailability (OB) and drug-likeness (DL) as the limiting conditions.

All data are expressed as the means ± standard deviation and analyzed using SPSS 17.0 (SPSS, Inc.). One-way analysis of variance and the Bonferroni post hoc test were used to evaluate statistical significance. Results were considered statistically significant when P<0.05.

Through the TCMSP database and literature investigation, 49 chemical components of Chuanxinlian were obtained. Using OB ≥30% and DL ≥0.18 as the filter criteria, 13 active ingredients were found to meet the requirements. Basic information on the ingredients is presented in Table II.

All targets were obtained from the TCMSP, SEA and SIA databases, and the duplicates and false positives were removed. Subsequently, 428 targets of Chuanxinlian were integrated; 421 ischemic stroke symptom targets were obtained from the DisGeNET, Drugbank and TTD databases; in addition, 68 potential targets were obtained from the intersection of Chuanxinlian targets and ischemic stroke targets (Table II).

Cytoscape 3.2.1 was used to construct the Andrographis paniculata active ingredient-anti-ischemic target map (Fig. 2), which contained 81 nodes and 249 edges. The rose-red-colored circles represent the active ingredients of Chuanxinlian, while the blue-colored squares represent the targets of these active ingredients against cerebral ischemia, and the edge represents the interaction between the active ingredient and anti-cerebral ischemia targets.

Target interactions were obtained from the String database, using Cytoscape 3.2.1 to draw the network diagram (Fig. 3). The 'edge' represents the correlation between the targets, and the thickness of the edge corresponds to the combined score: The thicker the edge, the greater the value of the combined score and the greater the degree of association. The node represents the target, and the value of 'degree' represents the effect intensity. The larger the degree value, the larger the node, and the larger the degree value corresponding to the color ranged from blue to red. There were 98 nodes and 1,099 edges in the figure. The average node degree was 20.4 and the average local clustering coefficient was 0.624. Finally, nodes with degree ≥5 and betweenness ≥0.00262056 were selected as the main nodes; thus, there were a total of 22 main nodes, as listed in Table III.

Gene function and pathway analysis revealed the following results: i)

GO enrichment analysis. The GO enrichment function and the KEGG pathway in DAVID were used to analyze the 22 selected targets. GO enrichment analysis includes 3 branches: Biological processes, molecular function and cellular component. At the P<0.05 setting, 298 biological processes or pathways were obtained. Biological processes or pathways with a higher P-value were screened, and GraphPad Prism7.0 was used for mapping.

As shown in Fig. 4A, the main biological processes of the targets were the regulation of response to an external stimulus, regulation of apoptosis and the regulation of programmed cell death. As shown in Fig. 4B, the main molecular functions of the targets were drug binding, endopeptidase activity, lipid binding and peptidase activity. Plasminogen (PLG) is a single-chain glycoprotein that plays an important role in thrombolytic therapy for acute cardiogenic cerebral infarction. When stimulated, vascular endothelial cells secrete a large amount of tissue-type plasminogen activator, which converts PLG into fibrinase and this activates the fibrinolytic system in vivo. Some autosomal genes associated with ischemic heart disease, including the gene encoding PLG, have been identified (19). PLG is also a key molecule involved in the thrombolytic treatment of cerebral ischemia. As shown in Fig. 4C, the cell components of the targets included cell fraction, plasma membrane and neuron projection.

The brain sections of the mice were stained with H&E (Fig. 5), and the results revealed that neurons in the cerebral cortex and hippocampal CA1 region in the sham group were normal. In these mice, brain tissue was not damaged and no obvious vacuolar space was observed. The nerve cells were densely and evenly arranged, with normal structure and morphology, clear cell contour, complete normal morphology and clearly visible nucleoli. By contrast, in the CIRI group, ischemic necrosis occurred in the cerebral cortex, with the shrinkage of neuronal nuclei; some cells exhibited spot-like necrosis, with evident infiltration of microglial cells, neuronal disorder and tissue edema in the hippocampus.

The mouse brain sections were stained with Nissl (Fig. 6), and the results revealed that, compared with the sham group, the numbers of Nissl-positive cells in the CIRI group were markedly decreased. Following pre-treatment with Andro, the numbers of Nissl-positive cells were significantly increased compared to the CIRI group.

The results of immunofluorescence staining illustrated that, compared with the sham group, the expression of GFAP was significantly increased (Fig. 7A and C) and the expression of NeuN was significantly decreased (Fig. 7B and D) in the brains of the mice subjected to CIRI. Following pre-treatment with Andro, the expression of GFAP significantly decreased, while the expression of NeuN significantly increased.

BDNF is involved in the development of the nervous system, and CIRI promotes the release of BDNF. Compared with the sham group, the BDNF content in the serum of mice in the CIRI group was significantly increased (P<0.05), and the BDNF mRNA level in the brain tissues also exhibited an increasing tendency (Fig. 8A and B). This indicates that cerebral ischemia-reperfusion can stimulate the release of BDNF in a short period of time. Compared with the CIRI group, the serum BDNF levels were significantly increased in mice pre-treated with Andro (P<0.05), and the BDNF mRNA levels in the brain tissues of mice receiving CIRI also exhibited an increasing trend. In addition, Andro pre-treatment enhanced the transcription of BDNF in mice subjected to CIRI.

TrkB is a primary receptor of BDNF. Compared with the sham group, TrkB protein expression in the brain tissues of mice in the CIRI group was significantly decreased (P<0.01), and compared with the CIRI group, the mice that received Andro (120 mg/kg) pre-administration exhibited a significantly upregulated TrkB protein expression in the brain tissue (P<0.01; Fig. 8C and D).

The levels of proteins downstream of TrkB [including phosphorylated PI3K (p-PI3K) and phosphorylated Akt (p-Akt) were reduced in the mice in the CIRI group, while Andro pre-treatment markedly increased the phosphorylation levels of these proteins compared to the CIRI group (Fig. 8C, E and F).

GFAP is a specific marker of activated astrocytes, and NeuN is a specific marker of mature neurons. CIRI increased the expression of GFAP and inhibited the expression of NeuN, while Andro administration markedly reduced the expression of GFAP and enhanced the expression of NeuN (Fig. 8G-I).

ELISA was used to determine the effects of Andro on the levels of the inflammatory cytokines, IL-1β, IL-6 and TNF-α in serum. The serum levels of IL-1β and TNF-α were significantly increased in the CIRI group compared to the sham group. However, compared with the CIRI group, the serum levels of IL-1β, IL-6 and TNF-α were markedly decreased in the Andro groups (Fig. 9A-C).

In addition, the IL-1β mRNA, IL-6 mRNA and TNF-α mRNA levels in brain samples were analyzed by RT-qPCR (Fig. 9D-F). Compared with the sham group, the IL-1β mRNA, IL-6 mRNA and TNF-α mRNA levels were significantly increased in the CIRI group. In addition, pre-treatment with Andro significantly downregulated the IL-1β mRNA, IL-6 mRNA and TNF-α mRNA levels (Fig. 9D-F).