Dl-3-n-Butylphthalide Reduces Cognitive Impairment Induced by Chronic Cerebral Hypoperfusion Through GDNF/GFRα1/Ret Signaling Preventing Hippocampal Neuron Apoptosis

Hippocampal neuron death is a key factor in vascular dementia (VD) induced by chronic cerebral hypoperfusion (CCH). Dl-3-n-butylphthalide (NBP) is a multiple-effects drug. Therefore, the potential molecular mechanisms underlying CCH and its feasible treatment should be investigated. This study had two main purposes: first, to identify a potential biomarker in a rat model of CCH induced VD using antibody microarrays; and second, to explore the neuroprotective role of NBP at targeting the potential biomarker. Glial cell line-derived neurotrophic factor (GDNF)/GDNF family receptor alpha-1 (GFRα1)/receptor tyrosine kinase (Ret) signaling is altered in the hippocampus of CCH rats; however, NBP treatment improved cognitive function, protected against hippocampal neuron apoptosis via regulation of GDNF/GFRα1/Ret, and activated the phosphorylation AKT (p-AKT) and ERK1/2 (p-ERK1/2) signaling. We also found that 1 h oxygen-glucose deprivation (OGD) followed by 48 h reperfusion (R) in cultured hippocampal neurons led to downregulation of GDNF/GFRα1/Ret. NBP upregulated the signaling and increased neuronal survival. Ret inhibitor (NVP-AST487) inhibits Ret and downstream effectors, including p-AKT and p-ERK1/2. Additionally, both GDNF and GFRα1 expression are markedly inhibited in hippocampal neurons by coincubation with NVP-AST487, particularly under conditions of OGD/R. GDNF/GFRα1/Ret signaling and neuronal viability can be maintained by NBP, which activates p-AKT and p-ERK1/2, increases expression of Bcl-2, and decreases expression of Bax and cleaved caspase-3. The current study showed that GDNF/GFRα1/Ret signaling plays an essential role in the CCH induced VD. NBP was protective against hippocampal neuron apoptosis, and this was associated with regulation of GDNF/GFRα1/Ret and AKT/ERK1/2 signaling pathways, thus reducing cognitive impairment.


INTRODUCTION
Vascular dementia is widely recognized as the second most common type of dementia. VD is primarily related to diverse cerebrovascular diseases, such as CCH (Du et al., 2017;Duncombe et al., 2017;Yin et al., 2018). CCH is induced by chronic, moderate and persistent deficit of CBF, increasing evidence from laboratorial and clinical researches has indicated CCH is a robustly common factor in pathogenesis of cerebrovascular diseases and neurodegenerative disorders, results in the development and progression of cognitive impairment, such as VD (Du et al., 2017). But to date, it has been difficult to modulate the complex pathological changes caused by CCH . Model animals of BCCAO are often used to study VD resulting from hypoperfusion (Du et al., 2017). The vascular hypothesis suggests that moderate and persistent cerebral hypoperfusion leads to vasculotoxic and neurotoxic effects due to diminished CBF and prolonged hypoxia, which promotes neurodegeneration and cognitive impairment (Raz et al., 2016). In the neurovascular units of the brain, neurons are the electrically functional cells, and demand a continuous supply of glucose and oxygen (Mao et al., 2019). Neuronal death, a main cause of neuronal loss, is a key hallmark in CCH-related VD . The hippocampus has critical roles in cognitive processes of learning, memory consolidation, and information retrieval (Wittenberg et al., 2002). Therefore, there may be a correlation between hippocampal neuron vulnerability to CCH induced cognitive damage. The restoration of CBF and rescue of dying neurons represent two primary therapeutic measures. There is uncertainty regarding the exact time of clinical occurrence in hypoperfusion, the irreversibility of neuron damage after reperfusion; therefore, understanding the molecular mechanisms underlying hippocampal neuron death in CCH would be more invaluable for the development of new therapeutic approaches to alleviate vascular cognitive impairment.
The quest for suitable drug candidates can be daunting when using traditional screening methods. Identifying a few biologically active compounds from huge libraries of biomolecules is similar to searching for a needle in a haystack. High-efficiency protein microarray technology can identify disease-or drug-related biomarkers more effectively than these traditional methods (Sun et al., 2013). In this study, using an antibody-microarray method, we propose a proteomic strategy that can robustly identify DEPs associated with NBP treatment in a CCH rat model of VD. The functions of the DEPs were further assessed by constructing a PPI network and by biological function enrichment analysis. In addition, the potential neuroprotective targets of NBP will be explored. We aimed to obtain better insight into the mechanisms-of-action of NBP and clarify the potential therapeutic target in improving cognitive function in CCH.

MATERIALS AND METHODS
Sprague Dawley (SD) rats (200-250 g), adult male, aged 8 weeks, were housed at 25 ± 2 • C and 40-70% humidity with a cycle of 12h light/dark for 4 weeks of adaptive feeding. Food and water were provided ad libitum. All animal protocols were approved by the institutional animal care committee of the experimental animal management center of Jinan University (Guangzhou, China), and conformed to internationally ethical standards (Guide for the Care and Use of Laboratory Animals. United States NIH Publication 86-23, revised 1985). A total of 97 rats were used. A complete flow chart for this study and the experimental groupings used are shown in Figures 1A,B.

Animal Surgical Procedures and Drug Administration
Bilateral common carotid artery occlusion operation has been proposed to reproduce the effects of CCH rat model Mao et al., 2019). Each rat was anesthetized with 3% sodium pentobarbital (0.2 ml/kg, intraperitoneal injection, Sigma-Aldrich, United States). A minimally invasive median incision was made in the neck carefully. The bilateral common carotid arteries and the vagus nerves were separated and isolated. For the BCCAO group (total n = 70; NBP-treated group, n = 28; CCH-treated group, n = 42), the bilateral carotid arteries were FIGURE 1 | (A) The experimental flowchart in vivo. Eight-week-old adult male SD rats were obtained and kept for 4 weeks of adaptive feeding. The day before 12 weeks of age (12w-1d), MRI scanning was performed, and the CBF was obtained before the operation (pre-BCCAO). At 12 weeks of age, the operation was performed and the CBF was detected using MRI (15 min after operation). Rats were randomly divided into three groups: Sham (control operation), the bilateral carotid arteries were separated as in the BCCAO group, without being ligated; CCH (BCCAO operation), the bilateral carotid arteries were double-ligated; CCH + NBP: BCCAO operation combined with consecutive NBP treatment for 3 weeks. Eight weeks after the BCCAO operation (20 weeks old), behavioral function was evaluated in the MWM test, and CBF was detected using MRI scanning. Rats were sacrificed for further experiments (21 weeks old). (B) Animal groupings used in this study. Rats were randomly divided into three groups: Sham, without ligated; CCH, 8 weeks after double-ligated operation, injection with solvent; CCH + NBP: NBP injection for early 3 weeks with 8 weeks after BCCAO operation. (C) Changes in CBF in the cortex and hippocampus were showed by MRI images. Color images are from 3D ASL, gray images from T2WI. Red and green circles indicate detected ROIs of CBF in the hippocampus and cortex, respectively. Histogram showing quantitative results in the different groups after BCCAO operation. (D) CBF in the bilateral cortex and hippocampus decreased immediately following BCCAO and 8 weeks after BCCAO, compared to pre-occlusion CBF (P < 0.01). CBF increased in the NBP treatment group compared to the BCCAO group. ( * P < 0.05, * * P < 0.01, compared to the pre-BCCAO; # P < 0.05, ## P < 0.01, the NBP treatment compared to the CCH). (E) Diagram of the MWM. The four colors divide the round swimming pool (the large circle) into four quadrants (Yellow, 1st quadrant; Red, 2nd quadrant; Green, third quadrant; Blue, fourth quadrant). The empty square represents the platform in the 2nd quadrant. (F) Escape latency changes in the different groups from day 1 to day 5 ( * P < 0.05, * * P < 0.01, CCH group vs. sham group; ## P < 0.01, CCH + NBP group vs. sham group; & P < 0.05, && P < 0.01, CCH vs. CCH + NBP group). (G) Swimming path of rats at day 6. The swimming path distance in the platform quadrant was reduced in the CCH. NBP treatment increased the frequency at the platform. (H) Quantified changes in the frequency at the platform between the three groups ( * P < 0.05, CCH group vs. the sham group; # P < 0.05, CCH vs. CCH + NBP group). CCH: CCH 8w; CCH + NBP: CCH 8w + NBP.
double-ligated with 2-0 sutures. For the sham group (Sham, n = 27), the bilateral carotid arteries were separated, without being ligated. The body temperature of the rats was maintained during surgery, as well as during recovery from anesthesia.
Following anesthesia, animals were placed in a prone position before scanning. All imaging parameters for the 3D ASL series, and CBF calculation methods were automatically recorded, as described in previous studies . The regions of interests (ROIs) (area: 4 mm 2 ) were measured in the bilateral hippocampus and cortex; three values were selected on each side to calculate an average value.
A blind test was performed prior to the experimental task to exclude blind rats. The utilized maze was a round tank, divided into four quadrants, 150 cm in diameter and 100 cm deep, filled up to a depth of 40 cm with tepid water (25 ± 1 • C). A movable square platform, 10 cm in diameter, was located 5 cm below the water surface. The maze was surrounded by white paper, on which black visual stimuli of four different shapes and sizes were placed in the four quadrants. Every trial started at a different point, alternating between the four quadrants. A camera was located above the center of the maze that relayed images to a videocassette recorder and the Image Analysis Computer System (Ethovison XT, Noldus Information Technology Co., Hague, Netherlands). Each test consisted of four trials from the four quadrants per day, and was conducted on five consecutive days. The hidden platform was always located in the 2nd quadrant of the water maze. Rats were submerged gently into the water, facing toward the inside wall of the tank. In the maze, rats were allowed to swim for a maximum time of 60 s. Rats were allowed to remain on the platform for 20 s at the end of each trial. Performance was evaluated for EL time for all trials. The platform was withdrawn at the sixth day of training. Swimming time spent (60 s) in the target quadrant with the retracted platform, and the frequency of time in the target quadrant, were used as parameters for retention of spatial memory.
We prepared a protease inhibitor cocktail and a cell lysis buffer at a ratio of 1:99, which was added to the hippocampal tissue on the ice. After 30 min, samples were spun down in a refrigerated centrifuge, at 13000 rpm, for 20 min. We then extracted the supernatant. The minimum protein concentration of the samples was 15000 µg/ml, as determined by comparison to a bovine serum antigen (BSA) standard. Next, all the samples were taken at 2 mg/ml, 50 µL, in the dialysis tubes. The tubes were then placed in 1 × phosphate buffered saline (PBS; pH = 8, 4000 ml) at 4 • C, with stirring, and the dialysis fluid was changed at intervals of 3 h.
During biotin labeling, reagent contamination was prevented by addition of amines or sodium azide (e.g., Tris, glycine). After the glass assay chips were left at 20-25 • C for 20-30 min, we removed the sealing strips and placed the chips in a vacuum dryer or at room temperature for 1-2 h. In each well on the chip, 400 µL of 1 × sealing solution was added and incubated at room temperature for 1 h to avoid bubbles. Then, we removed the fluid and added 400 µL of sample to each well, with one array per sample, and incubated the arrays at 4 • C, overnight, with constant shaking. Then, the samples were removed, and 1 × washing solution was added to each well and incubated at room temperature with constant shaking. The chips were washed four times for 5 min each. After repeated washing, the washing solution was removed, and 400 µL of a diluted fluorescent agent, streptavidin with Cy3, was added to each well. The arrays were covered with an aluminum foil seal, and samples were incubated, in the dark, at room temperature, for 2 h. After incubation, the arrays were again washed with washing solution.
Using a laser scanner (InnoScan 300 Microarray Scanner; Innopsys; Parc d'Activités Activestre; 31 390 Carbonne, France), fluorescence was measured in the Cy3, or green, channel (excitation frequency = 532 nm; resolution: 10 µm). Data were extracted by the original analysis software of the instrument. To determine spot intensities, the mean pixel intensity per spot was calculated. To determine background intensities, the median pixel intensity per background "doughnut" was calculated. Individual array spots were background subtracted locally (by subtracting the median background across spot replicates in each sample).

Protein-Protein Interaction Network Construction by STRING
Through inputting the ID number of DEP, DEPs were analyzed by the online tool, STRING (a database of known and predicted protein interactions) 1 , to predict interactions. A combined score of >0.9 (a high confidence score) was considered significant.

Functional Analysis of DEPs
Conservation analysis of the DEPs in rats and humans was conducted using BLAST R 2 and NCBI-Gene 3 .
Biological functional enrichment analyses of the DEPs, including GO functional analysis and KEGG pathway analysis, were conducted. In the GO analysis, the categories used included CC, BP, and MF terms. P < 0.05 was considered a statistically significant difference. In the KEGG pathways analysis, enriched pathways were identified according to the hypergeometric distribution with P < 0.05.

Real-Time Quantitative PCR
For each sample, cDNA (2 µL) was quantified and duplicated using Light Cycler 480 SYBR Green I on a Light Cycler 480 II as per manufacturer instructions. Cycling conditions: 10 min at 95 • C, 40 cycles of 15 s at 95 • C, 60 s at 60 • C. Melt curve cycles were immediately performed and the cycling conditions were as follows: 5 s at 95 • C, 60 s at 60 • C, then a gradual temperature rise to 95 • C at a rate of 0.3 • C/s, followed by 15 s at 60 • C. Melt curve analysis was performed to verify primer specificity. Data are displayed as a fold change above the proliferative condition mRNA levels using 2 ∧( Ct) values. Primer

Western Blot
Fresh hippocampal tissue samples were a random subset picked blindly. All tissues were lysed in RIPA buffer cocktail (containing protease and phosphatase inhibitors) (Cell Signaling Technology, United States). Total protein concentrations were determined with a BCA Protein Assay Kit (Thermo Fisher Scientific, United States). Total protein (10-20 µg) was loaded for each sample into pre-cast 4-12% bis-tris gels and run with MOPS buffer (Invitrogen, United States). The proteins were transferred onto polyvinylidene fluoride membranes (Millipore, United States). The membranes were incubated overnight at 4 • C with antigen-specific primary antibodies and detected with species-specific horseradish-peroxidase-labeled secondary antibodies. An ECL western blotting detection kit (GE Healthcare) was used to obtain a chemiluminescence signal, which was detected using Amersham Hyperfilm ECL (GE Healthcare

Immunostaining
Seven days after the behavioral experiments, rats were perfused transcardially with a 0.9% physiological saline solution and, subsequently, put under deep anesthesia with 4% paraformaldehyde. After the process of perfusion, the brains were removed and immersed in 4% paraformaldehyde immediately for 48 h, then embedded in paraffin. To ensure matching of hippocampal sections between groups, we used anatomical landmarks provided by the brain atlas.
Coronal brain sections (selected areas, Supplementary Figure S1b), 5 µm in thickness, were stained with hematoxylineosin (HE) and Nissl's staining. Sections were labeled with CD34 (for endothelial cells of microvessels), NeuN (for neurons), glial fibrillary acidic protein (GFAP) (for astrocytes), Iba1 (for microglial cell), and 4 ,6-diamidino-2-phenylindole dihydrochloride (DAPI) (for the cell nucleus) antibodies. First, sections were immersed in a blocking solution at room temperature for 2 h, and then incubated overnight at 4 • C with either CD34 antibody ( After three washes with PBS, sections were then incubated with the appropriate secondary antibody (1:500) for 2 h at room temperature. After three washes in PBS, sections were mounted with a fluorescence antifade mounting medium. For each type of labeling, n = 3 rats from each of the sham, BCCAO 8w, and NBP groups were used. Digital images were captured from the CA1 region at 40 × 10 magnification, and from the CA3 region and dentate gyrus (DG) at 20 × 10 magnification. Three images were captured from both sides of each region in each group, separately. The number of positive cells was counted with Image J and Image-Pro Plus 6.0 For each photograph, the exposure time was consistent. Two experienced pathology experts, blinded to all experimental information, independently measured and interpreted the final results.

TUNEL Staining
The TdT-mediated dUTP Nick-End Labeling (TUNEL) assay was performed according to the manufacturer's protocol (DeadEnd TM Fluorometric TUNEL System, Promega, Madison, WI, United States). The cerebral sections or cell specimens were incubated with a proteinase K solution for 15 min at 37 • C to enhance permeability. The sections were then incubated with the TUNEL reaction mixture for 1 h at 37 • C. After being rinsed with PBS, the sections were mounted using an antifade mounting media containing DAPI. Images were acquired using a fluorescent microscope (Nikon, Japan), the number of positive TUNEL cells were counted with Image J and Image-Pro Plus 6.0.

In situ Hybridization
To detect the expression and location of GFRα1 and GDNF, three rats were randomly chosen from the NBP group for in situ hybridization and immunohistochemistry. The hippocampus of the rats were harvested and cut into 20-µm thick sections. The sections were fixed in 4% paraformaldehyde and acetylated with 0.25% acetic anhydride. Prior to hybridization with the probes for 12-16 h (37 • C), the sections were prehybridized in a hybridization solution without probes. Subsequently, the sections were blocked at 37 • C for 1 h, immersed in a second antibody (1: 1000) at 4 • C overnight, and finally visualized with a fluorescent microscope (Nikon, Japan).

Transmission Electron Microscopy
Fresh hippocampal tissues (n = 5 per group) were cut into pieces of 1 mm 3 , and fixed in a 2.5% glutaraldehyde solution overnight at 4 • C. Specimens were rinsed with PBS, then soaked in osmium tetroxide. After dehydration in acetone, specimens were embedded in epoxide resin, and 70-nm in thickness for sections. Specimens were then stained with uranyl acetate followed by lead citrate. Lastly, ultrastructural changes of neurons were imaged using a TEM (HT7700, Hitachi, Japan).

Culture of Primary Hippocampal Neurons
Cultures of primary hippocampal neurons were extracted from E18-20 SD rat embryos as described previously, with some adjustments and improvements . In brief, the hippocampus of the embryos (n > 20/every time) were dissected in a Dulbecco's modified Eagle medium (DMEM) (high glucose) solution supplemented with a 10% fetal bovine serum (FBS) (Thermo Fisher Scientific, Gibco, United States) and 1% penicillin-streptomycin (Biological Industries, Israel), at 0-4 • C. It was then digested with 0.125% trypsin-EDTA (Thermo Fisher Scientific, Gibco, United States) and 0.4 mg/ml deoxyribonuclease I (Worthington Biochemical Corporation, United States) for 15 min, at 37 • C. After centrifugation at 1000 rpm for 5 min, samples were resuspended in a DMEM/F12 medium (Thermo Fisher Scientific, Gibco, United States) supplemented with 10% FBS and 1% penicillin-streptomycin. Viable cells were counted using 0.4% trypan blue (G-clone, Beijing, China) in a hemocytometer (QIUJING, Shanghai, China) and plated at a density of 2 × 10 5 /mL into a 96well plate pre-coated with poly-L-lysine (Sigma-Aldrich) (8-12 h). After 4 h, the culture medium was replaced with a serum-free neurobasal medium (Thermo Fisher Scientific, Gibco, United States) supplemented with 2% B-27 (Thermo Fisher Scientific, Gibco, United States) and 1% penicillin-streptomycin (Thermo Fisher Scientific, Gibco, United States). Half of the culture medium was replaced with fresh medium every 3 days.

Oxygen-Glucose Deprivation/Reperfusion and Drug Treatment in vitro
Existing studies have adopted OGD/R to simulate CCH model in vitro Mao et al., 2019). On day 7, the medium was replaced with DMEM without glucose (Thermo Fisher Scientific, Gibco, United States) and the neurons were incubated in an AnaeroPack R -Anaero (MGC AnaeroPack Series, Japan) to induce OGD injury at 37 • C for 1 h. The medium was then replaced with serum-free neurobasal medium supplemented with 2% B-27 to induce reperfusion (R). For concurrent treatment, NBP (60 µM) (Supplementary Figures S4f,g) Figures S4h,i). NVP-AST487 is a Ret kinase inhibitor, inhibiting Ret autophosphorylation and activation of downstream effectors. In addition, expression of both GDNF and GFRα1 are markedly inhibited by coincubation with NVP-AST487.

Cell Survival Assays
Neuron viability was assessed by Cell Counting Kit-8 (CCK-8) detection (450 nm) at 48 h after OGD, and before measurement cells were incubated with 10 µL CCK-8 (Dojindo, Japan) in 90 µL fresh serum-free neurobasal complete medium for 4 h at 37 • C. All samples were assayed in five replicates and each experiment was repeated at least five times.

Statistical Analyses
The raw data obtained from the scanning of the antibody microarray were acquired by Raybiotech R software and normalized between the microarrays. For cell quantification (TUNEL and immunofluorescence), three rats in each group were used in vivo, three slices per animal were stained, three images were analyzed per slice (CA1/CA3/DG section); in vitro model, three to five images per well were taken, and three repetitions in each experiment. Cells were counted (automatically or manually) and statistically analyzed. All other statistical analyses were performed using SPSS 19.0 (Abacus Concepts Inc., Chicago, IL, United States). Data were represented as the mean ± standard deviation (SD). Statistical analyses for differences between groups were performed using the two independent samples t test. Statistical differences among the groups were assessed by one-way ANOVA, when homogeneity of variance was determined, Fisher's least significant difference (LSD) test was used; when homogeneity of variance was not determined, Tamhane's T2 test was used. P < 0.05 was considered statistically significant.

NBP Promoted Recovery of CBF After BCCAO
In this study, we used the 3D ASL to dynamic observation. CBF was measured in the cortex and hippocampus of rats at six timepoints, including pre-occlusion, BCCAO, and the 1st, 2nd, 4th, and 8th week after BCCAO. In cortical areas, CBF was consistently lower following BCCAO than in the pre-BCCAO state, while CBF was restored in the hippocampus at the BCCAO 8w (Supplementary Figures S1c,d). These results indicated successful establishment of the BCCAO-induced CCH model.
Cerebral blood flow in the hippocampus and cortex of NBPtreated rats was assessed at pre-occlusion, BCCAO operation, and BCCAO 8w, and compared with that in the corresponding CCH-treated rats ( Figure 1C). Following BCCAO operation, the CBF decreased immediately after surgery in both the cortex and hippocampal areas (left cortex: CBF decreased to 50.91% ± 15.97 in the CCH-treated and 62.84% ± 21.64 in the NBP-treated groups. Right cortex: CBF decreased to 45.07% ± 11.11 in the CCH-treated and 59.14% ± 20.85 in the NBP-treated groups. Left hippocampus: CBF decreased to 67.52% ± 16.31 in the CCH-treated and 74.13% ± 27.09 in the NBP-treated groups. Right hippocampus: CBF decreased to 64.88% ± 13.32 in the CCH-treated and 68.64% ± 26.91 in the NBP-treated groups). The cortex in the CCH-treated (BCCAO 8w) rats had lower CBF at 8th week after BCCAO, while the cortex in NBP-treated rats showed recovery of CBF [Left cortex: CBF decreased to 70.18% ± 15.25 in the CCH-treated group, compared to pre-BCCAO (P < 0.01); CBF was increased to 96.49% ± 10.67 in the NBP-treated group compared BCCAO 8w (P < 0.01). Right cortex: CBF decreased to 67.48% ± 19.19 in the CCH-treated compared with pre-BCCAO (P < 0.01); CBF was increased to 90.69 ± 11.64% in the NBP-treated group compared to BCCAO 8w (P < 0.05)] (Figure 1D). In the hippocampus, CBF was restored to the preoperative level with or without NBP treatment [Left hippocampus: CBF increased to 86.61% ± 26.26 in the CCH-treated and 104.54% ± 13.28 in the NBP-treated groups compared to pre-BCCAO (P > 0.05). Right hippocampus: CBF increased to 96.34% ± 28.76 in the CCH-treated and 105.37% ± 8.6 in the NBP-treated groups compared to pre-BCCAO (P > 0.05)] (Figure 1D).
These results suggest that NBP treatment promotes recovery of cerebral CBF in CCH rats.

A Greater Degree of Cognitive Impairment Is Observed in CCH-Treated Than in NBP-Treated Group
The different experimental groups were tested for learning and memory deficits as well as EL in the MWM test. The frequency of crossing the platform was analyzed.
After the surgery, EL was significantly prolonged at CCH 2w, CCH 4w and CCH 8w (Supplementary Figure S1e). The frequency of crossing the original platform was decreased in the CCH 2w, CCH 4w, and CCH 8w (Supplementary Figure S1f) compared to sham groups.
These above results demonstrated the success establishment of the CCH-induced VD model at BCCAO 8w/CCH 8w.
Escape latency was notably reduced in the NBP-treated group (CCH 8w + NBP) compared to the CCH group (CCH 8w) (day 1, P < 0.05; day 2 to day 4, P < 0.01; day 5, P < 0.05). Rats in the CCH group had significantly increased EL compared to that in the sham rats from day 2 to day 5 (all P < 0.05); From day 2 to day 5, EL in NBP-treated rats was comparable to that in sham rats (P > 0.05) (Figures 1F,G). In addition, the frequency of crossing the platform at day 6 was significantly decreased in the CCH-treated group compared to both the sham rats and the NBP-treated group (Figures 1G,H).
These above results demonstrated the NBP reduce cognitive impairment in CCH rats model.

Angiogenesis in the Hippocampus of Rats With CCH
CD34 immunofluorescence was used to show changes in angiogenesis; we found cortical CD34 positive cells were increased in CCH 8w and NBP-treated groups. However, we did not find any statistical significance in CD34 positive cells between the three groups in the hippocampus (CA1, CA3, and DG areas) (Supplementary Figure S2). Quantitative analysis of data is shown in Supplementary Figures S2a-d.

NBP Reduced Hippocampus Neuronal Apoptosis in CCH Model
Hematoxylin-eosin and Nissl's staining showed neuronal damage in the CA1 and CA3 areas in CCH 4w and CCH 8w rats. This change was not observed in the CCH 2w rats and was most pronounced in the CCH 8w rats (Supplementary  Figures S1g,h), suggesting that the process of CCH had led to delayed neuronal death.
TUNEL staining is a marker for cell apoptosis. TUNEL analysis showed TUNEL-positive cells were widely present in the hippocampus (CA1 and CA3 areas) of the CCH-treated rats, while NBP-treated could significantly reverse the phenomenon (Figures 3A-C).
Western blotting showed that cleaved caspase-3 expression was decreased, however, Bcl-2 expression and Bcl-2/Bax ratio were increased in the sham/NBP-treated group compared to the CCH-treated group (Figures 3D,F). Quantitative analysis of data is shown in Figures 3E,G. FIGURE 2 | Dl-3-n-butylphthalide improves the pathological changes of the hippocampus. (A) Representative photomicrographs of hippocampal HE staining in the three groups. HE staining showed morphological changes in the rat hippocampus. Neuronal damage was detected in the SP in both the CA1 and CA3 areas in the CCH group. NBP treatment significantly increased the number of viable neurons, compared with the CCH group. (B) Representative photomicrographs of hippocampal Nissl staining in the three groups. Neuronal damage was detected in both the CA1 and CA3 area, and a partial region of the GL of the DG area in the CCH group. Normal neurons were arranged in an orderly manner, with typical morphology, and an evident nucleus and nucleolus. Abnormal neurons had lost their morphology, shrunken and deeply stained. NBP treatment reduced the damage, both in number of neurons and the morphological structure of the hippocampus. CCH: CCH 8w; CCH + NBP: CCH 8w + NBP. CA, cornus ammonis; DG, dentate gyrus; SO, stratum oriens; SP, stratum pyramidale; SR, stratum radiatum; ML, molecular layer; GL, granule cell layer; PL, polymorphic layer. CA1 area, magnification 200×, scale bar = 50 µm, CA3 and DG areas, magnification 100×, scale bar = 100 µm.

Alterations in Hippocampus Neuronal Morphology and Ultrastructure
Normal hippocampal neurons had large oval nuclei with clear nuclear membranes. In the CCH 8w group, neurons had irregular nuclear membranes, chromatin condensation, and many vacuoles. These changes were alleviated in the NBP-treated group ( Figure 3H). These results indicate that NBP treatment alleviates hippocampal neuron damage in CCH rats.

Networks and Function of DEPs in CCH and NBP Treatment Rats
Identification of DEPs between CCH-treated, NBP-treated, and sham rats was based on the results of protein arrays.
At first, there were 6 DEPs and GFRα1 served as a target protein in CCH 8w and sham groups, because GFRα1 expression was significantly downregulated in the CCH 8w, with a high degree of conservation (Supplementary Figures S3a,b). GO (Supplementary Figures S3c-e) and KEGG (Supplementary Figure S3f) analysis showed these DEPs were involving with some pathways.
Then tissues of three groups were compared; the top three DEPs [GFRα1 (P < 0.05), TIMP-1 (P < 0.05) and ACTH (P < 0.05)] were selected for further analysis (Figures 4A,B). The top three pathways revealed by KEGG analysis were the adipocytokine signaling pathway, the melanogenesis signaling pathway, and the HIF-1 signaling pathway ( Figure 4C). The GO analysis also showed that the DEPs were associated with the following BP: response to aging, steroid hormone secretion, and regulation of appetite; CC: the microbody lumen and peroxisomal matrix; and MF: neuropeptide hormone activity, neuropeptide receptor binding (Figures 4D-F). Next, using Protein BLAST R , the conservation of the three DEPs between Homo sapiens and Rattus norvegicus was analyzed. The results showed that GFRα1 was the most conserved DEP (Query cover = 100%; E-value = 0.0; Align identities score = 90.81%) ( Figure 4H).
STRING was used to predict the protein-protein interactions (PPIs) of the DEP. Three node proteins, GDNF, Ret, and NCAM, showed a strong association with GFRα1, indicating that these four hub proteins might play crucial roles in our animal model ( Figure 4G). Using the above methodology, we identified highly interconnected clusters of receptors and ligands, GDNF/GFRα1 binding to Ret or NCAM, which could be potential NBP therapy targets in CCH induced VD.

Hippocampal GDNF/GFRα1/Ret Expression in Different Experimental Groups
Rt-qPCR ( Figure 4I) and western blotting ( Figure 4J) showed GFRα1, GDNF, and Ret mRNA and protein levels were significantly increased in NBP-treated rats compared to CCHtreated rats. However, NCAM levels were not different between the three groups. A quantitative analysis of western blotting is shown in Figure 4K.
The number of NeuN-positive cells in the CA1 and CA3 areas in CCH 8w rats was significantly decreased compared to that in The expression of Bcl-2, Bax and Bcl-2/Bax ratio were quantified and represented as a histogram. β-actin was used as an internal control. n = 3 experiments. * * P < 0.01, CCH group vs. sham group; ## P < 0.01, CCH vs. CCH + NBP group. (H) NBP treatment protects against morphological and ultrastructural changes of neurons in the hippocampus. Sham group: the neurons in the cortex had large round nuclei, the double nuclear membranes were clear and complete. CCH group: neurons were irregular and exhibited chromatin condensation, cytoplasm dissolution, and vacuole formation. The nuclear membranes and organelles of neurons were dissolved or absent. NBP treatment group: the damage to the neurons was alleviated (n = 5 per group). Scale bar = 2 µm. CCH: CCH 8w; CCH + NBP: CCH 8w + NBP.
sham rats, but significantly increased by NBP. In addition, there was no difference between the NBP-treated and the sham rats (P > 0.05) (Figures 5A,H). Quantitative analysis of data is shown in Figure 5D.
To determine the cellular localization of the GDNF/GFRα1/Ret protein in the hippocampus, coimmunofluorescence staining for GDNF/GFRα1/Ret and three neuron cells markers (NeuN; GFAP; Iba1) in brain sections of the three groups was performed. GFRα1/GDNF/Ret protein expression was localized within NeuN-positive cells (Figures 5A,H,L), but it was not observed in GFAP-positive astrocytes (Figures 5B,I) or Iba1-positive microglia (Figures 5C,J).
In addition, GFRα1 + /GDNF + /Ret + -NeuN positive cells were downregulated in the hippocampus of CCH 8w rats but upregulated in those treated with NBP (Figures 5A,H,L).
Quantitative analysis of data is shown in Figures 5G,K,M. In situ hybridization results showed that GDNF/GFRα1 were located in the neurons (Supplementary Figure S3g). These findings suggest that upregulation of GDNF/GFRα1 binding to Ret in the hippocampus might be the mechanismof-action of NBP treatment and is associated with neuronal survival in CCH rats.

NBP Reduced Hippocampus Glial Cell Activation in CCH Rats
The changes in GFAP and Iba1 immunofluorescent labeling are shown in Figure 5. Quantitative analysis demonstrated that the number of GFAP positive cells was increased in the CCH 8w group compared to the NBP-treated and sham groups (Figure 5E), as were the number of Iba1 positive cells (Figure 5F).

NBP Treatment Increased Primary Hippocampus Neuronal Viability and Reduced Apoptosis
After 7 days in vitro, the purity of the primary hippocampal neurons was tested (Supplementary Figure S4a). HIF-1α was increased, without severity of neuron death in OGD 1 h/R 48 h (Supplementary Figures S4c-e).
Oxygen-glucose deprivation 1 h/R 48 h decreased the survival rate of neurons to 47.8%, while treatment with NBP (60 µM) increased neuron survival to 74.9% (Figure 6A). OGD 1 h/R 48 h treatment increased the number TUNEL-positive neurons (Figures 6B,C) compared to those in control and NBP-treated (60 µM) neurons.
Western blotting results showed that OGD 1 h/R 48 h led to increased expression of Bax and cleaved caspase-3, as well as an increased cleaved caspase-3/pro-caspase-3 ratio. It also led to decreased expression of Bcl-2 and a decreased Bcl-2/Bax ratio. Treatment with NBP reversed these effects (Figure 6D). A quantitative analysis of the data is shown in Figures 6E-J. Expression Levels of GDNF/GFRα1/Ret and p-AKT/p-ERK1/2 in Hippocampal Neurons Following OGD/R or NBP Treatment Compared with control or NBP treatment, OGD 1 h/R 48 h significantly decreased positive GDNF + /GFRα1 + /Ret + neurons (Figures 7A-C). Quantitative analysis of data is shown in Figures 7D-F. Western blotting results showed that NBP given alone in vitro model did not significantly affect the expression of GDNF/GFRα1/Ret and p-AKT/p-ERK1/2 ( Supplementary  Figures S4J,k). But OGD 1 h/R 48 h led to downregulation of the expression of GDNF/GFRα1/Ret and p-AKT/p-ERK1/2; NBP treatment can significantly increase expression of GDNF/GFRα1/Ret and p-AKT/p-ERK1/2 ( Figure 7G).

Inhibiting Ret Increases Apoptosis, While NBP Regulates Ret Expression and Reduces Neuronal Apoptosis in the OGD/R Model
To probe for a direct link between GDNF/GFRα1/Ret and neuronal apoptosis in CCH, we used an OGD 1 h/R 48 h model and Ret inhibition in hippocampal neurons. Neurons were treated with normal control, OGD 1 h/R 48 h (OGD/R), OGD 1 h/R 48 h + NBP (NBP), Ret inhibitor (Ret i), OGD 1 h/R 48 h + Ret inhibitor (OGD/R + Ret i), and OGD 1 h/R 48 h + Ret inhibitor + NBP (OGD/R + Ret i + NBP).
When the expression of Ret was inhibited, neuronal death was significantly increased. CCK-8 results showed that neuron survival decreased (Figure 6A), the number of TUNEL positive cells increased (Figures 6B,C), and the expression of apoptotic proteins increased (Figure 6D). Under conditions of OGD/R, NBP improved neuron survival, even in combination with Ret i (Figure 6A). Expression of apoptotic proteins was also relatively decreased in NBP-treated rats ( Figure 6D). Quantitative analysis of the data is shown in Figures 6E-J.

DISCUSSION
A comprehensive understanding of the status of a disease cannot be obtained from genomic studies alone. This study used a new approach, using an antibody microarray, to identify potential therapeutic targets associated with CCH induced hippocampal neuron apoptosis and NBP treatment. Antibody microarrays are high-throughput tools, with lower sample volume and antibody concentration requirements and higher format versatility. Its applications include disease biomarker discovery for drug response, diagnosis, characterization of protein pathways (Sanchez-Carbayo, 2006;Jaeger et al., 2016;Li et al., 2019). Through antibody microarray detection, we identified the expression of GDNF, GFRα1, and Ret was decreased in the hippocampal tissue of the CCH 8w; however, NBP produced an inverse effect. At present, there are no researches assessing the correlation between NBP and the GDNF/GFRα1/Ret signaling axis. This current study provides insight into the potential protective effects of NBP treatment against CCH induced hippocampal neuron apoptosis by regulation of GDNF/GFRα1/Ret signaling, resulting in improved cognitive function.

BCCAO Lead to Hypoperfusion and Cognitive Impairment, Accompany With Hippocampal Neuron Apoptosis
Lower CBF is predictive of a higher future risk of preclinical dementia (Hays et al., 2016;Raz et al., 2016). Permanent BCCAO is a classical experimental animal model for studying mechanisms of cognitive impairment underlying CCH (Du et al., 2017;Xiong et al., 2017;. At 8 weeks after BCCAO, concomitant cortical hypoperfusion and cognitive impairment were observed. HIF-1α is a primary transcriptional mediator of the hypoxic response and a master regulator of O 2 homeostasis (Yang et al., 2017). The levels of HIF-1α were significantly increased at BCCAO 8w in our study, indicating that a hypoxic state remained. Therefore, BCCAO 8w/CCH 8w was selected as the successful establishment of the CCH induced VD model for further study. Many researches claimed that cognitive damage and pathological changes that occur due to CCH can be attenuated by improvement of the CBF. However, although CBF gradually returned to its pre-occlusion level, neuropathological changes did not improve and actually deteriorated with the passage of time. A cascade effect may occur once an individual has experienced CCH for a period of time and that the methods used to recover CBF are effective only at earlier time points (Zou et al., 2018). CCH induces a compensatory mechanism to maintain optimal CBF, but this mechanism is limited and cannot prevent neuronal loss or cognitive impairment (Jing et al., 2015). In addition, the effect of NBP in improving CBF and is also limited .
A long-period hypoxia and HIF-1α expression in adult organisms may contribute to angiogenesis (Semenza, 2003;Greijer et al., 2005). Compared to the sham rats, the number of CD34 positive cells were increased in the cortex of CCH 8w and NBP-treated groups, indicating that angiogenesis occurred with effect of NBP, and at a later time point in CCH. Further, the number of CD34 positive cells was the same in the three groups and hippocampal CBF completely recovered to the pre-operative levels, the hippocampus is largely supplied by the posterior cerebral artery, and receives only a minor contribution from the anterior choroidal artery from the internal carotid artery (Kirschen et al., 2018). In addition, thickened vertebral arteries may provide a better blood supply Li et al., 2019). Although CBF recovered in the later stage of CCH to a certain degree, hippocampal neurons apoptosis in CA1 and CA3 areas was irreversible. At the same time point, learning and memory was significantly impaired. Further elucidation of the basic molecular mechanisms underlying the hippocampus neuronal apoptosis in CCH will be invaluable to the development of new therapeutic approaches.

GDNF/GFRα1/Ret Signaling and Neuron Survival
Glial cell line-derived neurotrophic factor is one of the members of the four ligands in the GDNF family, which belong to the transforming growth factor-β superfamily (Drinkut et al., 2016). GDNF signaling is mediated by two-component receptor consisting of GFRα1 and the transmembrane receptor tyrosine kinase Ret. GDNF dimers bind preferentially to GFRα1 with high affinity. In addition, GDNF may use the α1-NCAM dependent signaling pathway instead of the Retdependent pathway. Once activated, the resulting complex recruits Ret and sends signals through the brain by forming a FIGURE 6 | (A) CCK-8 results showed that OGD/R significantly increased neuron mortality ( * * P < 0.01, compared to the control), and NBP treatment significantly increased cell viability ( ## P < 0.01, OGD/R + NBP (60 µM) vs. OGD/R). When the expression of Ret was inhibited, neuron viability was significantly decreased ( * * P < 0.01, compared to the control), NBP treatment significantly increased cell viability ( ## P < 0.01, OGD/R + Ret i + NBP (60 µM) vs. OGD/R + Ret i). n = 5 replicates per group. Ret i, Ret inhibitor; OGD/R, OGD 1 h/R 48 h. (B) TUNEL staining showed that OGD/R significantly increased neuronal apoptosis. NBP treatment significantly decreased the number of TUNEL-positive neurons. When the expression of Ret was inhibited, the number of TUNEL-positive neurons was significantly increased, and NBP treatment significantly decreased this number. Imaged at 400× magnification, scale bar = 25 µm. Area for quantitative analysis, 200 µm 2 . (C) Histogram showing quantitative analysis of TUNEL staining in different groups of primary hippocampal neurons. n = 3 texts per group. * * P < 0.01, compared to the control; ## P < 0.01, OGD/R + NBP vs. OGD/R; ## P < 0.01, OGD/R + Ret i + NBP vs. OGD/R + Ret i. Ret i, Ret inhibition. (D) NBP treatment reduced apoptosis of primary hippocampal neurons under OGD/R conditions by regulating Ret signaling. Western blotting showed OGD/R significantly increased neuron apoptosis (increased expression of Bax, cleaved caspase-3 and cleaved caspase-3/pro-caspase-3 ratio, decreased expression of Bcl-2 and a decreased Bcl-2/Bax ratio). NBP treatment significantly decreased apoptosis. When expression of Ret was inhibited, apoptosis was significantly increased. NBP treatment significantly decreased neuron apoptosis.   multicomponent receptor complex, composed of the glycosylphosphatidyl inositol-anchored receptor GFRα1 and Ret. This leads to its activation at specific cytoplasmic tyrosine residues, thus initiating a number of downstream intracellular pathways (Sariola and Saarma, 2003;Curcio et al., 2015;Kramer and Liss, 2015). Ret can activate various signaling pathways, such as ERK, PI3K/AKT, p38 MAPK, and JNK pathways (Ichihara et al., 2004;Yue et al., 2017). These pathways have been demonstrated to play an important role in regulating cell proliferation, apoptosis, and survival in various systems.
Deficiency of GDNF receptor GFRα1 or Ret in neurons, results in neuronal death in AD and PD (Konishi et al., 2014;Kramer and Liss, 2015;Requejo et al., 2018). The absence of Ret completely abolished the neuroprotective and regenerative effects of GDNF on the midbrain dopaminergic system. This establishes Ret signaling as absolutely required for GDNF to prevent and compensate for dopaminergic system degeneration, which suggests Ret activation is the primary target of GDNF therapy in PD (Drinkut et al., 2016). In cultured hippocampal neurons, brain ischemia downregulates the neuroprotective GDNF-Ret signaling by a calpain-dependent mechanism. Preserving Ret receptors may be a good strategy to increase the endogenous neuroprotective mechanisms (Curcio et al., 2015). It suggests that the main goal of clinical trials using GDNF and related substances should be to activate the Ret receptor. Although transient ischemic injury upregulates GDNF expression in the injured region of the brain, this is less likely to result in an increase in neuroprotection (Curcio et al., 2015). CCH is a process in the chronic ischemic category; therefore, under conditions of CCH, the downregulation of Ret affects GFRα1 and the GDNF/GFRα1 complex, which also affects downstream effectors of Ret, including p-AKT and p-ERK1/2. This eventually leads to delayed neuron death. Based on what has been discussed above, activating GDNF/GFRα1/Ret signaling may be a good strategy to increase endogenous neuroprotective mechanisms in CCH-induced VD (Curcio et al., 2015;Ibanez and Andressoo, 2017).

Potential Molecular Contribution of NBP in CCH Induced VD
The complex etiology of CCH means that multifunctional agents may be beneficial for the treatment of this disease. In a previous study where NBP was tested in CCH animals, NBP could reduce cognitive impairment induced by CCH, but the molecular mechanisms underlying the therapeutic effect were different as the study investigated the therapeutic effects of NBP on downregulation of the amyloid precursor protein Aβ40, MMP-2 and MMP-9 proteins in cortex and hippocampus (Wei et al., 2012). An amelioration in delayed neuronal death occurring after CCH might be a vital therapeutic target to improve the long-term outcome of VD. In this current study, CBF in the bilateral cortex in NBP-treated rats recovered compared to the pre-operation stage with a decrease in HIF-1α levels, with cognitive improvement. However, the neuroprotective effect of NBP ameliorates hippocampal neuron apoptosis is also noteworthy to study in CCH models.
Neurotrophic factors are good therapeutic candidates for neurodegenerative diseases. GDNF is a diffusible peptide, involved in neuronal differentiation and survival, and has been identified as potential biomarker in AD in an unbiased proteomic assay (Ray et al., 2007), in amyotrophic lateral sclerosis (Stanga et al., 2018) and is the most potent dopaminergic factor described for the treatment of PD (Garbayo et al., 2016). In early acute hypoxia, GDNF counteracts acute hypoxic damage to hippocampal neural network function in vitro (Shishkina et al., 2018).
In the course of a long-term hypoxia attack, GDNF/GFRα1/Ret signaling was downregulated and not the NCAM dependent signaling pathway, which may inform the development of potential therapeutic strategies for CCH induced VD through upregulation of GDNF/GFRα1/Ret expression. However, previous attempts to move GDNF into clinical practice have been only moderately successful. One of the factors limiting the effectiveness of GDNF therapies is its short biological half-life due to its labile nature. Furthermore, the factor does not cross the blood brain barrier (BBB) and often has serious side effects when administered systemically, necessitating an effective drug delivery strategy to reach the brain. Therefore, in order to use GDNF effectively as a therapeutic agent, it is essential to develop a safe and effective brain delivery system (Garbayo et al., 2016). Gene and cell therapy have experienced moderate success, but are also complex and expensive. Therefore, the development of small-molecule drugs that can pass the BBB is a promising prospect.
Racemic NBP is a multi-target neuroprotective agent and is a chiral compound containing L-and D-isomers. It is also a small molecule, and, with HP-β-CD as the carrier, can easily pass through the BBB (Diao et al., 2015). Although it has previously been studied, new molecular targets of NBP are still worth exploring.
With a combination of in vivo and in vitro experiments, we show that NBP promotes the upregulation of Ret. It then upregulated GDNF/GFRα1 and caused downstream effector pathways, specifically AKT and ERK1/2 pathways, to become activated and reduce hippocampal apoptosis, and improve cognitive function in CCH (Figure 8). In addition, activation of astrocytes and microglia was significantly reduced in the NBP rats, consistent with our previous study . GDNF family ligands are able to regulate microglial function and may play a role in the suppression of microglial activation, as microglia are the target cells of members of the GDNF family (Rickert et al., 2014). Reduction of activated glia may be associated with improved collateral circulation, or the reduced delayed neuronal apoptosis, by NBP. Some researchers indicated that BCCAO might not effectively represent a clinical VD model, due to the limitation of CCH process induced by BCCAO. Therefore, further experiments including more experimental models of VD/more behavioral tests, are needed to support, and to determine the deeper molecular mechanism of CCH and targets of NBP.

CONCLUSION
The, GDNF/GFRα1/Ret signaling might be a potential therapeutic target in CCH. NBP treatment may be a good strategy to mediate cognitive improvement. It may also help reduce hippocampal neuron apoptosis by regulating neuroprotective mechanisms of GDNF/GFRα1/Ret signaling.

DATA AVAILABILITY
The raw data supporting the conclusions of this manuscript will be made available by the authors, without undue reservation, to any qualified researcher (Supplementary Table S1).

ETHICS STATEMENT
All animal protocols were approved by the institutional animal care committee of the experimental animal management center of Jinan University (Guangzhou, China) and conformed to internationally accepted ethical standards (Guide for the Care and Use of Laboratory Animals. United States NIH Publication 86-23, revised 1985), ensuring humane and proper care of research animals.

AUTHOR CONTRIBUTIONS
WL, LH, and DW designed the research and prepared the manuscript. WL and XX performed the BCCAO rat model and measured the water maze experiments. WL, JYL, and JXL performed the experiments with magnetic resonance imaging. WL and DW performed the primary neuronal cell culture and related experiments and performed all the experiments related to histology, immunohistochemistry, and Western blot analyses. WL, KS, and JXL collected and analyzed the data.

SUPPLEMENTARY MATERIAL
The Supplementary Material for this article can be found online at: https://www.frontiersin.org/articles/10.3389/fncel.2019. 00351/full#supplementary-material FIGURE S1 | (a) Structure of NBP. Its chemical name is (±) 3-butyl-3H-2-benzofuran-1-one (chemical formula: C 12 H 14 O 2 ; molecular weight: 190.24). (b) A representative HE stained section showing selected areas in the cortex and hippocampus (CA1, CA3 and DG area) for evaluation. (c,d) Changes in CBF after BCCAO CBF was measured in the cortex and hippocampus of rats pre-occlusion, immediately following BCCAO, and 1, 2, 4, and 8 weeks after BCCAO using ASL. In cortical areas, CBF was consistently lower after BCCAO than before, while CBF was restored in the hippocampus at the 8th week after BCCAO. Histograms show quantitative results of the CBF at different time points after BCCAO. CBF in the bilateral cortex was significantly decreased upon BCCAO, as well as the 1st, 2nd, 4th, and 8th weeks after BCCAO, compared to CBF pre-occlusion (c). CBF in the bilateral hippocampus was significantly decreased upon BCCAO, as well as the 1st, 2nd, and 4th weeks after BCCAO, compared to CBF pre-occlusion, but was increased in the 8th week (P > 0.05) (d). * P < 0.05, * * P < 0.01, compared to pre-BCCAO; # P < 0.05, ## P < 0.01, BCCAO 8w vs. BCCAO 4w. L, left side; R, right side. Changes in learning and memory after BCCAO. (e)We investigated whether the BCCAO model used in the study would induce cognitive impairment. Rats were trained to perform the MWM test. After surgery, the EL was significantly prolonged at the 2nd, 4th, and 8th week after BCCAO ( * P < 0.05, * * P < 0.01, CCH 2w vs. sham; # P < 0.05, ## P < 0.01, CCH 4w vs. sham; && P < 0.01, CCH 8w vs. sham). (f) On the 6th day, rats were allowed to navigate the water for 60 s. Quantitative data showed that the frequency of crossing the original platform was decreased 2, 4, and 8 weeks after BCCAO ( * P < 0.05 compared to sham). Morphological changes in pyramidal cells and delayed neuronal death in BCCAO groups. Morphological changes in pyramidal cells were found in the CA1 and CA3 areas of the CCH 8w group. They also occurred in the CA1 area of CCH 4w group, but were not found in the CCH 2w or sham groups (g). Delayed neuronal death was observed in the CA1 and CA3 areas of the CCH 8w group, as well as in the CA1 area of the CCH 4w group, but it has not been observed in the CCH 2w or sham groups. Neurons in those areas denoted hyperchromatism, shown with pyknosis and shaped into diamond and triangle (h). CA1, CA3, and DG areas, magnification 200×, scale bar = 50 µm.