Neuroprotective Effect of Aurantio-Obtusin, a Putative Vasopressin V1A Receptor Antagonist, on Transient Forebrain Ischemia Mice Model

Traditional Chinese medicines (TCMs) have been a rich source of novel drug discovery, and Cassia seed is one of the common TCMs with numerous biological effects. Based on the existing reports on neuroprotection by Cassia seed extract, the present study aims to search possible pharmacological targets behind the neuroprotective effects of the Cassia seeds by evaluating the functional effect of specific Cassia compounds on various G-protein-coupled receptors. Among the four test compounds (cassiaside, rubrofusarin gentiobioside, aurantio-obtusin, and 2-hydroxyemodin 1-methylether), only aurantio-obtusin demonstrated a specific V1AR antagonist effect (71.80 ± 6.0% inhibition at 100 µM) and yielded an IC50 value of 67.70 ± 2.41 μM. A molecular docking study predicted an additional interaction of the hydroxyl group at C6 and a methoxy group at C7 of aurantio-obtusin with the Ser341 residue as functional for the observed antagonist effect. In the transient brain ischemia/reperfusion injury C57BL/6 mice model, aurantio-obtusin attenuated the latency time that was reduced in the bilateral common carotid artery occlusion (BCCAO) groups. Likewise, compared to neuronal damage in the BCCAO groups, treatment with aurantio-obtusin (10 mg/kg, p.o.) significantly reduced the severity of damage in medial cornu ammonis 1 (mCA1), dorsal CA1, and cortex regions. Overall, the findings of this study highlight V1AR as a possible target of aurantio-obtusin for neuroprotection.


Introduction
Vasopressin (AVP) and oxytocin have been implicated in the etiology of psychiatric disorders, such as schizophrenia [1], autism [2][3][4], and depression [5]. AVP acts centrally within the central nervous system (CNS) where it modulates a range of behaviors from learning and memory and responses to stressors to social behaviors [6]. The vasopressin 1A receptor (V 1A R) has been contemplated to play the dominant role in regulating behavior and until recently, among vasopressin subtypes (V 1A , V 1B and V 2 ); the V 1A R is thought to be the only subtype expressed widely in the brain [6][7][8]. A recent study showed a marked reduction in anxiety-like behavior and a profound impairment in social recognition in V 1A R knock-out (V 1A RKO) mice [9]. Similarly, a study by Ferris et al. [10] suggested that V 1A receptor antagonists may be used to treat interpersonal violence co-occurring with an illness such as attention-deficit/hyperactivity disorder, autism, bipolar disorder, and substance abuse. Therefore, it is hypothesized that antagonistic action on the V 1A receptor might attribute to the treatment approach in anxiety-like behavior and recently, discovery of a potent, selective, and brain penetrant V 1A receptor antagonist is emerging [4,11]. Additionally, AVP had been reported to mediate brain edema formation and cerebral ischemia by regulating water permeability in astrocytes [12]. Similarly, the peripheral role of V 1A R in the inflammatory process of inflammatory bowel disease (IBD) mediated by prostaglandin release has been reported recently [13]. In the published report, V 1A R promoted COX-2dependent prostaglandin release from a mucosal mast in 2,4,6-trinitrobenzene sulfonic acid (TNBS)-induced colitis in mice, which was attenuated by conivaptan (a V 1A R antagonist).
Besides, AVP has numerous peripheral roles. Linas and colleagues [14] had previously reported an increased AVP level along with impaired renal water excretion and abnormal renal hemodynamics in a mouse model of CCl 4 -induced liver cirrhosis. Similarly, a recent study on the ischemia-reperfusion injury mouse model [15] identified upregulated V 1 R expression in hepatocytes and highlighted the importance of the hepatocyte V 1 R/Wnt/βcatenin/FoxO3a/Akt pathway in hepatoprotection. Additionally, V 1A R in vascular smooth muscles is responsible for vasoconstriction, myocardial contractility, platelet aggregation, and uterine contraction [16,17] and regulates blood pressure in vascular walls [18,19]. It has been reported previously that AVP is released in response to peripheral inflammation [20,21] which is deleterious to various immune-mediated diseases. A recent report on the Phase II clinical trial of a selective V 1A R antagonist balovaptan suggested it as a potential treatment for the socialization and communication deficits in autism spectrum disorder [4,22]. Likewise, investigations on the functional role of V 1A R in cardiovascular homeostasis using gene targeting demonstrated lower basal blood pressure in mutant mice lacking the V 1A R gene (V 1A −/− ) compared to the wild-type mice (V 1A +/+ ) [18,23]. Cassia seed has a long history of use in brewing tea in South Korea. Additionally, in traditional Chinese medicine, Cassia seeds have been used as vision improving, aperient, diuretic, antiasthenic, and an effective agent in lowering cholesterol and reducing blood pressure along with anthraquinones and naphthopyrones derivatives as predominant constituents, particularly the glycosides (cassiaside and rubrofusarin gentiobioside) [24]. Likewise, a comparison of HPLC chromatograms of various anthraquinones from Cassia seeds, revealed aurantio-obtusin as a predominant aglycon [25]. The major biological activities of aurantio-obtusin reported so far include antihypertensive [26], hepatoprotective [27], antimutagenic [28], osteogenic [29], and estrogenic activities [30]. Cassia seeds extract and its constituents have been reported for the management of various diseases including Alzheimer's disease [31][32][33][34], Parkinson's disease [35], diabetes and diabetic complications [36,37], hepatotoxicity [38,39], inflammation [34], oxidative stress [40,41], and many more [42]. Similarly, seeds extract has been reported as a therapy for neurodegenerative disorders in recent years [43]. However, there exist only a few reports on the effects of specific compounds present in Cassia seeds. More recently, we have discovered anthraquinones as promising human monoamine oxidase-A (hMAO-A) inhibitors [44], and emodin and alaternin (=7-hydroxyemodin) as potent vasopressin V 1A R antagonists and dopamine D 3 R agonists [45]. In this study, we evaluated the functional effect of cassiaside, rubrofusarin gentiobioside, and aurantio-obtusin ( Figure 1) on vasopressin V 1A receptor, 5-HT 1A receptor, neurokinin receptor, and dopamine D3 receptor because these receptors were predicted as top protein targets of cassia-derived compounds in neurodegenerative diseases [45]. Additionally, the effect of aurantio-obtusin has been compared with 2-hydroxyemodin 1-methylether for the elucidation of structure-activity relationship and the probable mechanism of ligand-receptor interaction was assessed via molecular docking simulation. Besides, we report herein the neuroprotective effect of aurantio-obtusin in a transient forebrain ischemia mice model.

Functional Effect of Test Compounds via cAMP Modulation
The functional role (agonist and antagonist effect) of test compounds ( Figure 1) on dopamine receptors was assessed by measuring their effect on cAMP modulation in transfected Chinese hamster ovary (CHO) cells using homogeneous time-resolved fluorescence (HTRF) detection. As shown in Table 1, only aurantio-obtusin showed mild agonist effect by stimulating the dopamine effect by 33.00 ± 1.84% (p < 0.05). However, cassiaside and rubrofusarin gentiobioside demonstrated a negligible agonist effect on hD 3 R. The compounds stimulated the effect of dopamine (300 nM) by 14.75 ± 1.20 and 7.05 ± 1.63%, respectively. The reference control agonist, dopamine exhibited an EC 50 value of 2.7 nM. However, none of the test compounds showed an antagonist effect. The percentage inhibition of agonist response of dopamine (10 nM) by the test compounds was negative for hD 3 R. Reference antagonist (+)-butaclamol exhibited an IC 50 value of 25 nM.

Functional Effect of Test Compounds via Intracellular Ca 2+ Ion Modulation
The functional effect (agonist and antagonist effect) of test compounds on the h5-HT 1A R, hNK 1 R, and hV 1A R was evaluated by measuring the intracellular Ca 2+ concentra-tion. As shown in Table 1, the agonist response of test compounds on the tested receptors was negligible at 100 µM (% stimulation of control agonist effect by the test compounds was negligible). Likewise, the antagonist effect of cassiaside and rubrofusarin gentiobioside was also negligible. Only aurantio-obtusin demonstrated antagonist effect on hNK 1 R (47.60 ± 3.11% at 100 µM; p < 0.01) and V 1A R (71.80 ± 6.08% at 100 µM; p < 0.001). Since the V 1A R antagonist effect of aurantio-obtusin was greater than 70% at 100 µM, the concentration-dependent antagonist effect was evaluated at various concentrations up to 100 µM and the 50% inhibition concentration (IC 50 value) was determined. Additionally, we tested the V 1A R antagonist effect of 2-hydroxyemodin 1-methylether which is a substructure of aurantio-obtusin. As shown in Figure 2  . Reference drugs and test samples were tested at the indicated concentration for antagonist effect by determining the percentage inhibition of control response to 10 nM AVP. (B) Represents a comparative inhibition pattern of aurantio-obtusin with its substructure, 2-hydroxyemodin 1-methylether. Values are expressed as mean ± SD of triplicate experiment.

Molecular Docking Simulation
Aurantio-obtusin-V 1A R interaction was analyzed via molecular docking simulation using AutoDock 4.2 and the mechanism was compared with its substructure, 2hydroxyemodin 1-methylether along with the reference ligands. Overall docking results (binding score and interacting residues) are depicted in Figure 3 and tabulated in Table 2.

Drug-Likeness and ADME Prediction
Drug-likeness was predicted for aurantio-obtusin and the prediction results are tabulated in Table 3. As shown in the table, aurantio-obtusin exhibited good drug-like properties.
It was predicted mid-structure according to the MDDR-like rule [46] and considered as a suitable drug candidate molecule based on the Lipinski's rule [47]. Table 3. Drug-likeness and absorption, distribution, metabolism, and excretion (ADME) characteristics as predicted by PreADMET.

Compounds
Drug-Likeness g ADME Characteristics

MDDR-like rule
Lipinski's rule According to absorption, distribution, metabolism, and excretion (ADME) prediction, moderate plasma protein binding (86.98%), good human intestinal absorption (84.66%), and good lipophilicity (2.53) were predicted. As reviewed earlier [48], lipophilicity (Log Po/w) value in the range of 1.5-2.5 indicates the suitability for CNS delivery. Likewise, the blood-brain barrier (BBB) penetration value ([brain]/[blood]) was 0.48% indicating moderate absorption by the CNS. All these predicted results could be utilized for optimizing drug-like properties. Additionally, permeability across Madin-Darby Canine Kidney (MDCK) and human epithelial colorectal adenocarcinoma (Caco2) cells was predicted to be 113.20 and 19.19 nm/s, respectively.

Discussion
In this study, we tested functional effect of cassiaside, rubrofusarin gentiobioside, and aurantio-obtusin on V 1A R, D 3 R, NK 1 R, and 5-HT 1A R which were the top protein targets for Cassia compounds in neurodegenerative diseases predicted via proteocheminformatics modeling [45]. Among the tested compounds, only an anthraquinone aurantioobtusin showed a good V 1A R antagonist effect. Naphthopyrone glycosides cassiaside and rubrofusarin gentiobioside remained inactive on these receptors. Quinone derivatives were previously predicted as reactive and pan assay interference compounds (PAINS) through high throughput screening that could show false biological activities [49,50]. If so, then all quinones should show activity. Therefore, we compared the V 1A R antagonist effect of aurantio-obtusin with its substructure 2-hydroxyemodin 1-methylether ( Figure 2). Both compounds are anthraquinones but only aurantio-obtusin showed good antago-nist effect while 2-hydroxyemodin 1-methylether showed mild V 1A R antagonist effect (42.3% inhibition at 100 µM concentration). Likewise, in our recent report on G-proteincoupled receptors modulation by anthraquinones from Cassia seed [45], only emodin and 7-hydroxyemodin showed activity on dopamine and vasopressin receptors. Other anthraquinones aloe-emodin and questin remained inactive at the tested concentrations. Therefore, the SAR would be evidence that all quinones are not PAINS [49].
Comparison of the V 1A R antagonist effect of anthraquinones emodin and alaternin (=7-hydroxy emodin) with other anthraquinnones from our recent study [45] revealed that the hydroxyl group at C1, C3 and C8 and a methyl group at C6 of anthraquinone structure are essential for the hV 1A R antagonist effect. In this study, we tested aurantioobtusin (1,3,7-trihydroxy-2,8-dimethoxy-6-methylanthracene-9,10-dione) which differs slightly from emodin and alaternin for its functional effect on V 1A R and compared with its sustructure 2-hydroxyemodin 1-methylether. The only difference between the test compounds is the presence of a methoxy group at the C2 position. Aurantio-obtusin has a methoxy group at C2 position while 2-hydroxyemodin 1-methylether lacks it. Compared to their functional effect on V 1A R, only aurantio-obtusin showed an antagonist effect with an IC 50 value of 67.70 ± 2.41 µM. The antagonist effect of 2-hydroxyemodin 1-methylether on V 1A R was 42.30 ± 9.89% at 100 µM concentration. A small change in the substituent at the C2 position showed a marked difference in the activity. Therefore, to further demonstrate the mechanism and clarify the reason behind the difference in activity, molecular docking simulation was conducted. As shown in Table 2, both the test compounds had a similar binding score and the same interacting residues at the active site of V 1A R. However, the hydroxyl group at C6 and a methoxy group at C7 of aurantio-obtusin showed two additional H-bond interactions with the Ser341 residue. The same functional groups of aurantio-obtusin were responsible for H-bond interaction with key amino acid residues surrounding the catalytic cavity of human thrombin for inhibition [51]. Interaction with Ser341 was not observed in the case of 2-hydroxyemodin 1-methylether-V 1A R binding. Consequently, it remains unclear whether Ser341 is responsible for the observed functional effect of aurantio-obtusin on V 1A R.
Water extract from Cassia obtusifolia seeds reduced blood pressure in cold-induced hypertensive mice, modulated blood lipid contents, and improved pathological changes in renal structure [52]. Furthermore, the Cassia component, gluco-aurantio-obtusin exhibited good inhibition of angiotensin-converting enzyme (ACE) activity with an IC 50 value of 30.24 ± 0.20 µM revealing its blood pressure regulating property while, its acid-hydrolyzed product aurantio-obtusin exhibited no activity [26].
A study on the effect of aurantio-obtusin on immunoglobulin E (IgE)-mediated allergic responses and LPS-induced RAW264.7 cells demonstrated suppression of degranulation, histamine production and ROS generation, inhibition of mRNA expression of TNF-α and IL-4, suppression of PGE2 production, and expression of COX-2 [25,53]. This demonstrates the benefits of aurantio-obtusin in treating allergy-related diseases. Likewise, aurantioobtusin was reported for its larvicidal effect in Anopheles gambiae [54], inhibitory effect on IL-6 production in IL-1β-treated lung epithelial cells, A549, and attenuation of lung inflammatory responses in a mouse model of LPS-induced acute lung injury in male ICR mice [55], thereby revealing the therapeutic potential for treating inflammatory diseases [53]. Similarly, aurantio-obtusin stimulated chemotactic migration of MC3T3-E1 osteoblast cells and osteoblast differentiation and mineralization which are the therapeutic strategies to prevent osteoporosis and other metabolic bone diseases [29]. Additionally, aurantioobtusin showed concentration-dependent vasorelaxation in phenylephrine precontracted rat mesenteric arteries rings via endothelial PI3K/Akt/eNOS pathway [56].
Since AVP levels in patients with heart failure and left ventricular (LV) dysfunction are often elevated [57][58][59], it is hypothesized that AVP might contribute to circulatory response in patients with heart failure and play a role in the development and progression of heart failure [60]. Moreover, the V 1A R antagonist effect of aurantio-obtusin might be a promising approach for the treatment. Transient brain ischemia/reperfusion injury occurs due to a temporary blockage of blood supply to the brain, and triggers selective neuronal loss/death in the most vulnerable brain region, especially the cornu ammonis 1 (CA1) field in the hippocampus [61][62][63]. Therefore, we further evaluated the neuroprotective effect of aurantio-obtusin in the transient brain ischemia/reperfusion injury C57BL/6 mice model. In the passive avoidance test, the BCCAO groups showed a significant reduction in latency time, however, treatment with aurantio-obtusin attenuated that reduction significantly (Figure 4). Likewise, compared to neuronal damage in the BCCAO groups, aurantio-obtusin significantly reduced the severity of damage in mCA1, dCA1, and cortex regions ( Figure 6). However, treatment with 10 mg/kg aurantio-obtusin alone showed no toxicity which was comparable to the sham group. Overall, in vivo data depicts the neuroprotective effect of aurantio-obtusin. Additionally, the drug-likeness and ADME characteristics of aurantio-obtusin further support the possibility of drug development and optimization. However, whether the neuroprotective effect of aurantio-obtusin is regulated via V 1A R remains unclear. This necessitates in-depth pharmacology of aurantio-obtusin using the V 1A R deficit (V 1A −/− ) mice model along with detailed molecular dynamic studies.
This study evaluated the functional effect of major components from Cassia seedscassiaside and rubrofusarin gentiobioside along with the anthraquinone aurantio-obtusin and its substructure 2-hydroxyemodin 1-methylether. Based on the structure-activity relationship, additional interaction of the hydroxyl group at C6 and a methoxy group at C7 of aurantio-obtusin with the Ser341 residue was predicted functional for the observed V 1A R antagonist effect. According to the previous study [64], aurantio-obtusin can cause hepatoxicity at a higher dose than 40 mg/kg in the rat. However, the dose applied for the rat needs verification through clinical trial, and demonstrating the optimal dose requires several clinical trial phases. In this study, a dose of 10 mg/kg was enough to exhibit a neuroprotective effect in C57BL/6 mice model. In addition, it was reported that aurantio-obtusin exist mainly in the form of metabolites such as sulfonation products and glucuronidation products in the body [65]. Therefore, further experiment should be conducted to find new active metabolites of auratio-obtusin, but also to unravel their pharmacokinetics and hepatotoxicity.
Altogether, this result highlights aurantio-obtusin as a V 1A R antagonist and V 1A R as a possible target for neuroprotection. However, in-depth in vivo studies on the V 1A R deficit (V 1A −/− ) mice model is warranted to demonstrate the V 1A R-regulated neuroprotection mechanism.

Plant Material
The raw seeds of Cassia obtusifolia Linn were purchased from Omni Herb Co.

GPCR Functional Assay
Cell based functional GPCR assays were conducted at Eurofins Cerep (Le Bois I'Eveque, France) using transfected cells expressing human cloned receptors namely dopamine (D 3 R), serotonin (5-HT 1A R), tachykinin (NK 1 R), and vasopressin (V 1A R). Agonist/antagonist effect of test compounds in each receptor was evaluated by measuring the level of secondary messengers.

Measurement of cAMP Level
The effect of test compounds on hD 3 receptor expressed in CHO cells was evaluated by measuring their effect on cAMP modulation using HTRF detection. In brief, a plasmid containing the GPCR gene of interest (dopamine D3) was transfected into Chinese hamster ovary (CHO) cells. The resulting stable transfectants (CHO-GPCR cells line) were suspended in HBSS buffer (Invitrogen, Carlsbad, CA, USA) supplemented with 20 mM HEPES buffer and 500 µM IBMX. The solutions were distributed into microplates at a density of 5 × 10 3 cells/well and incubated for 30 min at room temperature (RT) in the absence (control) or presence of aurantio-obtusin (100 µM) or reference agonist. Cells were lysed and a fluorescence acceptor (D3-labeled cAMP) and fluorescence donor (anti-cAMP antibody with europium cryptate) were added following the incubation. The fluorescence transfer was measured (λex = 337 nm and λem = 620 or 665 nm) using a microplate reader (Envision, PerkinElmer, Waltham, MA, USA) after 60 min at RT. Results are expressed as a percentage of the control response to dopamine for the agonist effect and as percent inhibition of the control response to dopamine. The standard reference control was dopamine [67]. Cellular agonist effect was calculated as the percentage of the control response to 300 nM dopamine for D3R, and cellular antagonist effect was calculated as the percentage inhibition of agonist response of 10 nM dopamine. To validate the result, reference antagonist (+)-butaclamol was used for D 3 R.

Measurement of Intracellular Ca 2+ Ion Concentration
Functional effect of test compounds on the h5-HT 1A R, hNK 1 R, and hV 1A R was evaluated fluorimetrically by measuring intracellular Ca 2+ concentration. Agonist activity of test compounds on the h5-HT 1A R expressed in Ba/F3 cells, hNK 1 R expressed in U373MG cells, and V 1A R expressed in transfected CHO cells was determined by measuring their effect on cytosolic Ca 2+ ion mobilization using a fluorimetric detection method described in our previous reports [68,69]. For antagonist activity, the effect on agonist-induced cytosolic Ca 2+ ion mobilization was measured. Cellular agonist effect at h5-HT 1A R was calculated as the percentage of the control response to serotonin (2.5 µM), and antagonist effect was calculated as the percentage inhibition of the control response to 30 nM serotonin. To validate the result, reference antagonist (S)-WAY-100635 was employed. Likewise, for the cellular agonist effect at hNK 1 R, the percentage of the control response to 30 nM [Sar9, Met(O2)11]-SP was determined and for antagonist effect, percentage inhibition of control response to 1 nM [Sar9, Met(O2)11]-SP was recorded. The standard reference antagonist L 733,060 was used to validate the result. Additionally, for the cellular agonist at hV 1A R, the percentage of the control response to 1 µM AVP was determined and for antagonist effect, percentage inhibition of control response to 10 nM AVP was recorded. The standard reference antagonist [d(CH 2 )5 1 ,Tyr(Me) 2 ]-AVP was used to validate the result.

Homology Modeling of V 1A R
The primary sequence of the human V 1A R was obtained from UniProt (ID: P37288). µ-Opioid receptor obtained from RCSB protein data bank (PDB) with ID of 4DKL was used as a template for homology modeling of V 1A R. Modeling was conducted using SWISS-MODEL and refined in ModRefiner server (RMSD = 0.645 Å) [70].

Molecular Docking
To get insight of reciprocal interactions between compounds and the target, docking simulation was conducted using AutoDock 4.2. program [71]. 3D structure of aurantioobtusin and 2-hydroxyemodin 1-methylether were constructed using Chem3D Pro v12.0 and refined using Discovery Studio (v17.2, Accelrys, San Diego, CA, USA). To assess the appropriate binding conformation of the ligands with protein target, AutoDockTools (ADT) was used to conduct docking simulation. For the docking calculations, Gasteiger charges were added by default, the rotatable bonds were set by ADT and all torsions were allowed to rotate. The grid maps were generated using AutoGrid. The docking protocol for rigid and flexible ligand docking included 10 independent genetic algorithms. The results were analyzed and visualized using Discovery Studio.

Drug-Likeness and ADME Prediction
Drug-likeness prediction was carried out with PreADMET (v2.0, YONSEI University, Seoul, Korea). This web-based server can be used to predict absorption, distribution, metabolism, and excretion (ADME) data and build a drug-likeness library in silico.

Animal
Male C57BL/6 mice (22-26 g, 7 weeks) were purchased from the Orient Co. Ltd., a branch of Charles River Laboratories (Seoul, Korea), and kept in the University Animal Care Unit for 1 week prior to the experiments. The animals were housed five per cage, allowed access to water and food ad libitum; the environment was maintained at a constant temperature (23 ± 1 • C) and humidity (60 ± 10%) under a 12-h light/dark cycle (the lights were on from 07:30 to 19:30 h). Forty mice were divided equally into four groups (sham + vehicle, n = 10; sham + drug, n = 10; bilateral common carotid artery occluded ischemia + vehicle, n = 10; bilateral common carotid artery occluded ischemia + drug, n = 10) for experiment. The treatment and maintenance of the animals were carried out in accordance with the Animal Care and Use Guidelines of Dong-A University, Korea. All in vivo experiments were performed according to the protocols approved by the Institutional Animal Care and Use Committee of Dong-A University (approved protocol numbers: DIACUC-approved- [17][18][19][20] and were in accordance with the National Institutes of Health guidelines.

Transient Forebrain Ischemia Surgery
C57BL/6 mice were anesthetized with 2.0% isoflurane and 70% nitrous oxide in oxygen and subjected to transient forebrain ischemia. The transient forebrain ischemia was induced by bilateral common carotid artery occlusion (BCCAO) with aneurysm clips for 20 min, and the circulation was restored by removing the clips. Mice that received the same surgical operation without carotid artery clipping served as sham-operated controls. During the surgical procedure, the rectal temperature was maintained at 37 ± 0.5 • C with a heating pad (Biomed S.L., Barcelona, Spain). The regional cerebral blood flow (rCBF) was monitored using laser Doppler flowmetry (Perimed, PF5010, JarFalla, Sweden). The mice that showed between 80% and 95% reduction of rCBF were used in the study. After reperfusion, the animals were placed in a warm incubator (32-33 • C) and returned to their home cages. Aurantio-obtusin, which was dissolved in 10% Tween 80 solution, was administered from 1 h to 7 days after BCCAO (10 mg/kg, p.o., once daily).

Passive Avoidance Test
Passive avoidance test was conducted 1 h after the last drug administration. The animals underwent two separated trials, an initial training trial and a test trial 24 h later. For the training trial, a mouse was initially placed in the light compartment, and the door between the two compartments was opened 10 s later. When the mouse entered the dark compartment, the guillotine door automatically closed and an electrical foot shock (0.5 mA, 3 s) was delivered through the floor. For the retention trial, the mouse was again placed in the light compartment and the time required to enter the dark compartment was recorded.

Slices Preparation and Nissl Staining
One day after the test trial of passive avoidance test, mice were anesthetized with Zoletil 50 ® (10 mg/kg, i.m.) and then perfused transcardially with a 100 mM phosphate buffer (pH 7.4) followed by ice-cold 4% paraformaldehyde. The brains of the mice were removed and post-fixed in a phosphate buffer (50 mM, pH 7.4) containing 4% paraformaldehyde overnight, then immersed in a 30% sucrose solution (in 50 mM phosphate-buffered saline, PBS), and stored at 4 • C until sectioned. The frozen brains were coronally sectioned on a cryostat at 30 µm and then stored in a storage solution (30% ethylene glycol, 30% glycerin, and 20 mM phosphate buffer) at 4 • C. Hippocampal sections were collected based on the mouse brain atlas.
After the sections were mounted onto gelatin-coated slides, they were stained with 0.5% cresyl violet, dehydrated through graded alcohols (70,80,90, and 100% × 2), placed in xylene, and covered with a coverslip after the addition of Histomount media. The number of cells in selected regions (medial CA1, mCA1; dorsal CA1, dCA1; CA2; cortex) were determined using a computerized image analysis system (Leica Microsystems AG, Wetzlar, Germany). The cells were counted in six sections by every eight sections interval (total 48 sections) per animal by a person blind to the treatment group, and the average cell count per section was computed. The degree of damage by the Nissl staining after ischemia was semiquantitatively scored from 0 to 3 (0, normal; 1, <30% of the neurons were irreversibly damaged; 2, 30-60% of the neurons were irreversibly damaged; 3, 60-100% of the neurons were irreversibly damaged).

Statistical Analysis
Statistical analysis was performed by Student's t-test using Microsoft Excel 2016 (Microsoft Corporation, Redmond, WA, USA). All experiments were carried out in triplicate on three individual days and are expressed as the mean ± standard deviation (SD). Results of Nissl staining and passive avoidance test were analyzed using one-way ANOVA (GraphPad Prism ver. 9). Data are expressed as the mean ± SD with raw data. * p < 0.05.

Institutional Review Board Statement:
The study was conducted according to the guidelines of the Declaration of Helsinki, and approved by the Institutional Animal Care and Use Committee of Dong-A University (approved protocol numbers: DIACUC-approved- [17][18][19][20] and were in accordance with the National Institutes of Health guidelines.