A total of 25 male rats (Rattus norvegicus), aged between 7 and 9 weeks and weighing at least 200 g from the Animal House Faculty of Medicine Universitas Brawijaya (Malang, Indonesia), were used in the present study. The duration of the study was 4 months. These rats were divided into five different groups, with each group consisting of 5 rats, as demonstrated in Table I. The initial total number of experimental animals used was 30 rats, with 6 rats per group. At the final stage of the experiment, there were 2 rats that died naturally and 3 rats that died following middle cerebral artery occlusion (MCAO) due to complications associated with this procedure. A total of 25 rats were used for parameter analysis. The selection of these groups was made to evaluate the efficacy of 7,8-DHF treatment compared to the placebo and to normal rats. The rats were placed in a suitable housing habitat and allowed to acclimatize at room temperature (22-28˚C). The humidity in the room was maintained at 45-50%, with a 12-h light-dark cycle. Food and water were provided ad libitum. Each week, the experimental animals were examined and monitored concerning overall health, neurological functions, animal behavior and physical parameters, such as body weight and body condition score. The indication for euthanasia was determined based on various factors, including self-mutilation, the amputation or crushing of limbs and tail, generalized dermatitis, infection and severe dehydration. Death was determined based on the absence of respiration, absence of heartbeat, dilated pupils, the absence of reflexes, poor pigmentation of the extremities and mucous membranes, and the presence of dry and opaque corneas. The research was carried out following internationally recognized ethical principles and the guidelines outlined in the National Institutes of Health's Guide for the Care and Use of Laboratory Animals of the Medical Faculty, Brawijaya University (No. 160/EC/KEPK/09/2020).

The experimental procedure involved the induction of ischemia-reperfusion brain injury in the rats by performing MCAO. In the present study, the rats were injected with ketamine (40-100 mg/kg) and xylazine (5-13 mg/kg) to induce and maintain anesthesia. The experimental setup included the use of a heat lamp and a heating pad to maintain body temperature at a constant level of 37˚C. A longitudinal incision of ~2 cm was made along the midline of the neck. The right common carotid artery (CCA) was carefully dissected from the vagus nerve and briefly occluded with cotton thread. The left external carotid artery was ligated at its bifurcation into the internal and external carotid arteries. Following a delicate ligation procedure on the right internal carotid artery (ICA), a 3-0 monofilament nylon was carefully inserted into the ICA through the CCA incision. The nylon material exhibited signs of detachment over a 90-min occlusion period. Upon the completion of the experimental procedure, the rats were transferred to the incubator for specialized and attentive medical care (21,22). In the sham-operated (sham) group, the rats underwent an identical surgical technique but without the induction of MCAO. Following 14 days of treatment, the rats were euthanized with intraperitoneal overdoses of ketamine (40-100 mg/kg body weight) and xylazine (5-13 mg/kg body weight), followed by decapitation. Blood and brain tissue samples were collected. Brain tissue samples were fixed overnight by immersion in a 10% formalin solution.

7,8-DHF was obtained from the manufacturer, Tokyo Chemical Industry Co., Ltd. (product no. D1916; lot no. OQ5VB). The 7,8-DHF isolate has a purity of 99.5%, as shown by high performance liquid chromatography and 100% by neutralization titration (data not shown). This isolate was dissolved using a concentration of 10% dimethyl sulfoxide (Cas No. 67-68-5, MilliporeSigma) for the dissolution (23). Isolate preparations of 7,8-DHF were prepared every 5 days to prevent oxidation and damage. The administration of 7,8-DHF was performed by a professional assistant using a sonde/feeding tube (Thermo Fisher Scientific, Inc.) as doses, as demonstrated in Table I.

Immunohistochemistry (IHC) was used to measure the amount of Nrf2 and GPX-1 protein in the brain tissues of rats in the model of ischemic stroke. The kit used was from Bioss with the following antibodies: [Nrf2, cat. no. bs-2013R; (1:75), and GPX-1, cat.no. bs-3882R; (1:75)]. The calculation was based on the amount of Nrf2, and GPX-1 protein found in the cytoplasm and nucleus of brain tissue in the cortical lesion area. This included neurons, and glia and Schwann cells observed under a microscope at x400 magnification. Briefly, the brain tissue samples were deparaffinized and stored at room temperature for 24 h prior to use. Following deparaffinization, the slides were immersed in a sodium citrate buffer solution (pH 6.0) and then washed with phosphate-buffered saline (PBS) (P-5119, MilliporeSigma) three times for 5 min. Blocking was performed with a blocking buffer (skimmed milk) (Blocker BLOTTO in Tris-buffered saline; cat. no 37530, Thermo Fisher Scientific, Inc.) followed by incubation for 60 min at room temperature. Nrf2 polyclonal antibody, unconjugated (bs-2013R), GPX-1 polyclonal antibody, unconjugated (bs-3882R) at a ratio of 1:75 were applied overnight at 4˚C. Secondary antibodies (rabbit anti-rat IgM antibody (H+L), biotin-conjugated; 1:200; bs-0346R-Biotin; Bioss) were used and incubated with the samples for 60 min at room temperature, then washed three times for 5 min with PBS. Streptavidin HRP was then incubated with the samples for 40 min at room temperature. The chromogen diaminobenzidine (DAB; AMF080, ScyTek Laboratories, Inc.) ratio was 1:40 (DAB chromogen: DAB buffer) and the samples were incubated for 5 min at room temperature. Mayer's hematoxylin (cat. no. 109249, MilliporeSigma) (Mayer's hematoxylin: distilled water, 1:3) was added, incubated for 2 min at room temperature, then washed with distilled water and mounted with Entella. The final result was observed under a confocal microscope (BX51, Olympus Corporation) at x400 magnification. To minimize examination bias, IHC calculations were conducted by two observers under the supervision of a pathologist. The counting of positive cells was performed in a single-blind manner, with codes known only to one individual to confirm the results.

To assess the extent of oxidative damage in the brain tissue, the concentration of MDA, a reliable indicator of lipid peroxidation, was quantified in the brain samples. This was performed using a commercially available ELISA kit (MDA Elabscience^® ELISA kit, cat. no. E-EL-0060, Elabscience Biotechnology, Inc.) according to the manufacturer's instructions.

The ADMET of 7,8-DHF was evaluated using SwissADME (http://www.swissadme.ch/) by inserting the SMILES formula of 7,8-DHF. Lipinski's rule of 5 (LR5) was used to evaluate the pharmacokinetic potential of the compounds. A chemical can be classified as a drug if it meets the LR5 criteria. Finally, the OSIRIS software tool (https://www.organic-chemistry.org/prog/peo/) was used to predict the human safety of a compound based on its toxicity analysis (24,25).

To derermine the potential activity of 7,8-DHF, the biological activity was measured using the prediction of activity spectra for substances (PASS) online web server (http://www.way2drug.com/index.php). The purpose of the web server is to predict the biological activity spectrum of organic compounds by analyzing their structural formulas. It covers a wide range of >4,000 biological activities and achieves an average accuracy rate of >95%. The predictions are obtained by studying structure-activity relationships in a training set consisting of structure and biological activity data for >300,000 organic molecules (26,27). Briefly, a canonical molecule from PubChem was entered into the PASS online web server. The biological activity prediction of a 7,8-DHF was then calculated. Graph visualization was performed using GraphPad Prism version 10 software (https://www.graphpad.com/features), while the potential molecular electrostatics were measured using Gaussian 09W and Gaussian View 6 software (https://gaussian.com/gaussview6/). The molecular electrostatic potential (MESP) was also predicted using Chimera (https://www.cgl.ucsf.edu/chimerax/).

In the present study, molecular docking simulation was performed using the Keap/Nrf2 complex structure (PDB ID: 2FLU). 7,8-DHF was downloaded from PubChem (https://pubchem.ncbi.nlm.nih.gov/compound/7_8-Dihydroxyflavone) with code CID: 1880. Molecular docking was performed using PyRx (https://pyrx.sourceforge.io/) on a personal computer running Windows. Protein stabilization was performed using PyMOL by removing water atoms and introducing hydrogen atoms. To verify the accuracy and reliability of the molecular docking process, a redocking procedure was performed. The original ligand was repositioned into the target binding site using the PyRx with accurate grid coordinates. After the redocking process, an evaluation was performed to determine the position of the ligand. Ligands were required to have a root mean square deviation (RMSD) of <2.0. The lattice and docking parameters were modified according to the docking validation results. The interaction visualization was performed using the Protein ChimeraX (https://www.cgl.ucsf.edu/chimera/) and Discovery Studio (https://discover.3ds.com/discovery-studio-visualizer-download) (28,29).

MD simulations were used to determine how 7,8-DHF and Keap1 interact. The OpenMM engine and the AMBER force field were used for both the protein and ligand systems. The simulations were performed within the Google Collaboratory platform (https://colab.research.google.com/?utm_source=scs-index) as described in a previous study (30). In the present study, the model was subjected to thorough solvation using the TIP3P water model with 108 water molecules contained in a cubic water box of 10x10x10 units. A solution of sodium chloride was introduced at a concentration of 150 mM. The simulation was performed using the ff19SB force field. The system underwent 1,000 steps of minimization followed by 20 nsec of equilibration in the isothermal-isobaric (NPT) ensemble using a 2-femtosecond time step. Temperatures are set at 310 Kelvin, and pressures are in the range of 1 atmosphere. The present study focuses on the investigation of the RMSD.

Statistical analysis in the present study was conducted using GraphPad Prism 9 software (Dotmatics). Data are presented as mean ± SD. Normality testing was performed using the Shapiro-Wilk test due to the sample size being <50 subjects. The analysis of the Nrf2, GPX-1 and MDA parameters was carried out using a one-way ANOVA parametric test to assess the overall efficacy of administering the 7,8-DHF treatment, followed by post-hoc Tukey analysis to examine the differences within each group. P-values <0.05 were considered to indicate a statistically significant difference.

To determine the antioxidant effects of 7,8-DHF, Nrf2 expression was examined. 7,8-DHF was found to increase the expression of Nrf2 in dose-dependent manner. The data demonstrated that 7,8-DHF at 20 and 50 mg/kg body weight (MCAO20 and MCAO50, respectively) significantly increased Nrf2 expression compared with the control, as shown in Fig. 1A and C.

GPX-1 expression was also examined using IHC. The results revealed a significant increase in GPX-1 expression beginning at the dose of 20 mg/kg body weight compared with the control. The most significant increase in GPX-1 expression was observed in the MCAO50 group, as illustrated in Fig. 1B and D.

Furthermore, the level of MDA was lower in the MCAO20 group than in the control (MCAOP group). The most notable decrease level in the MDA level was found in the MCAO20 and MCAO50 groups compared with the other groups. There was no significant difference in the MDA level between the MCAO20 and MCAO50 groups, as shown in Fig. 1E.

To examine the characteristics of 7,8-DHF as a drug, ADMET, drug-likeness and toxicity analyses were performed. In drug absorption, lipophilicity and solubility are critical molecular characteristics. This property is critical to the drug feasibility requirements, particularly for orally administered drugs (31). This drug feasibility test is also known as the drug likeliness criteria. Numerous criteria have been established, including the Lipinski criteria. A drug-like compound that can be taken orally should have no more than one violation of the criteria outlined by Lipinski's rule: No more than 5 hydrogen bond donors, no more than 10 hydrogen bond acceptors, a molecular weight of no more than 500 Da, and an octanol-water partition coefficient (log P) of no more than 5(28). The results revealed that 7,8-DHF fulfilled the LR5 and had a good ADMET, as presented in Table II. This indicates that the 7,8-DHF compound has sufficient oral bioavailability to be administered orally.

7,8-DHF has a BBB score of -0.057, indicating low permeability. BBB permeability is interpreted as logBBB >0.3, indicating that molecules can easily pass through the BBB, whereas molecules with logBB-1 are poorly distributed in the brain. The difference between BBB permeability and central nervous system (CNS) permeability lies in their respective properties. For CNS permeability, logPS units accurately predict the compound administered into the carotid artery. Compounds with logPS >-2 can penetrate the CNS, whereas compounds with logPS-3 cannot penetrate the CNS. The 7,8-DHF score logPS was-1.927, indicating that it can penetrate the CNS (Table II). Finally, these data were confirmed with BOILED-Egg analysis prediction using SwissADME (http://www.swissadme.ch/), demonstrating that 7,8-DHF can penetrate the BBB, as shown in Fig. S1. As shown by the results, the 7,8-DHF compound has moderate potency based on its distribution profile (32).

The toxicity of 7,8-DHF was also predicted using the biocomputational approach, an important parameter for safety in the human body. Several parameters are evaluated to determine toxicity. The Ames test is a widely used technique to determine the mutagenic potential of substances using bacteria (33). The Ames test used in the present study was a prediction test using OSIRIS (https://www.organic-chemistry.org/prog/peo/). The OSIRIS toxicity prediction model uses structural features and substructures associated with toxic effects to determine the potential toxicity of a chemical molecule. OSIRIS evaluates the submitted chemical structure by comparing it to a database of recognized toxicophores (substructures associated with toxic effects) and structural alerts. Toxicophores refer to distinct chemical properties that have been associated with various forms of toxicity. OSIRIS provides toxicity predictions for several domains, including mutagenicity, tumorigenicity, irritation, reproductive effects and others. Each category can have different warnings associated with it (28). The data demonstrated that 7,8-DHF was mostly safe, although there was a moderate risk of mutagenicity, as shown in Table II.

7,8-DHF has a molecular structure consisting of two aromatic rings connected by a single rotatable bond. It has four acceptor sites and two donor hydrogen bonds. The flavone skeleton contains two hydroxyl groups at positions 7 and 8 with a trans configuration. The biological activity of the molecule is influenced by several functional groups, with the hydroxyl groups being critical for protein-receptor interaction. The planarity and stability of the flavone backbone are due to the presence of a carbonyl group at position 4 and a double bond between positions 2 and 3. A representative 3D structure of 7,8-DHF is presented in Fig. 2A.

To determine the characteristics of the structure, MESP was performed. The MESP surface is a tool used to identify electrophilic and nucleophilic regions of a molecule. Different shades on the surface represent different levels of electrostatic potential. The ability of a molecule to bind to biomolecules can be determined by calculating its electrophilicity index (34). Different colors represent the electrostatic potential levels at the surface in the following order: red-orange-yellow-green-blue. The color coding for these maps in the identified compound varies from -6.457 (deepest red) to +6.457 (deepest blue) (blue indicates the most electronegative and red is the most electropositive). The most electronegative region of 7,8-DHF in a hydrogen atom is shown in Fig. 2B.

To determine the biological activity of 7,8-DHF based on chemical structure, the biological activity was predicted using a PASS online web server. PASS is a computer program designed to predict the biological activities of chemical substances. The PASS online tool allows users to predict the pharmacological, biological and toxicological properties of organic substances by analyzing their chemical structures. This can be beneficial in the drug discovery and development process (35). The biological activity prediction of 7,8-DHF shows that it has the probability of anti-inflammatory and antioxidant activity using PASS online. The highest predicted score is antioxidant activity with a score of 0.741, followed by anti-inflammatory activity with a score of 0.640, indicating that 7,8-DHF has promising anti-inflammatory and antioxidant activity. Notably, the prediction also specified 7,8-DHF as an inhibitor of tumor necrosis factor (TNF)-α expression with a score of 0.502, as shown in Fig. 2C. This suggests a possible mechanism of anti-inflammatory activity through the inhibition of TNF-α expression. Finally, this prediction also consists of a neurotrophic factor enhancer with a score of 0.248. Neurotrophic factor enhancer activity is important in the pathogenesis of stroke, which is reduced in stroke conditions. It is important to recognize that while PASS and other similar predictive methods are valuable for screening large databases of chemical compounds, they have limitations. Predictions are derived from the current information contained in the training data set, and the accuracy of these predictions depends on the quality and variety of the data used to develop the model. Experimental validation is critical to verifying and improving the expected biological activity of compounds.

To determine the antioxidant effect related to the molecular prediction mechanism, molecular docking was performed by targeting Keap1. The interaction of the Keap1/Nrf2 complex is illustrated in Fig. 3A. Nrf2 appears to interact with the binding pocket of Keap1 in the middle of the protein. There are several key amino acid residues that play critical roles in the complex interaction between Keap1 and Nrf2. These are tyrosine (TYR)334, arginine (ARG)415, ARG483, serine (SER)508, glutamic acid (GLU)30, SER555 and SER602, as shown in Fig. 3B. To clarify whether the binding of 7,8-DHF can inhibit Keap1, a small-molecule inhibitor [(3~{S})-3-(4-chlorophenyl)-3-(2-oxidanylidene-1~{H}-pyridin-4-yl)propanoic acid] of Keap1 was used as a control in the present study (36). It was found that this small-molecule inhibitor can bind to the binding pocket site of the Keap1-Nrf2 interaction, thereby directly disrupting the Keap1/Nrf2 interaction, as shown in Fig. 3C. Before beginning molecular docking, redocking was performed to clarify the molecular procedure. The redocking result was good as it had the same configuration, which has an RMSD value <2 Å, as shown in Fig. 3D.

Following successful validation, molecular docking simulations were performed between 7,8-DHF and Keap1 proteins, whose value data, or ΔG (kcal/mol), are presented in Table III. The results of molecular docking revealed that 7,8-DHF can bind to Keap1 with a binding affinity of -7.4 Kcal/mol. This binding affinity is lower than that of the control (-7.8 Kcal/mol). To confirm whether the binding is similar to the control at important residues, visualization was performed using ChimeraX and Discovery Studio. It was found that 7,8-DHF has a bond at the same site as the control ligand, as shown in Fig. 4A. This interaction appears to have an interaction with an important amino acid in Keap1, as shown in Fig. 4B. The control was found to be able bind to SER508, ARG483 and glutamine (GLN)530 in a hydrogen bond manner; isoleucine (ILE)461, phenylalanine (PHE)478, GLY462, glycine (GLY)509, GLY364, GLY603, SER555 and TYR572 with van der Waals; and TYR525, alanine (ALA)556, and ARG415 in a hydrophobic bond manner. While 7,8-DHF can bind to the Keap1 protein via GLY364, ILE461, GLY462, ARG483, GLY509, TYR525, SER555, TYR572, SER602 and GLY603 in a hydrogen bond manner, while ARG415, GLN530 and ALA556 have hydrophobic bond modes. Almost all bonds formed between 7,8-DHF and Keap1 are similar to the control, except for SER602, as shown in Fig. 4C.

Based on the molecular dynamic analysis of 7,8-DHF and control against Keap1. It was found that 7,8-DHF was slightly more unstable than the control. The data revealed that the Keap1 protein complex against the control had an average RMSD fluctuation of 0.8 angstrom (Å) with an equilibrium of 5 nsec, while 7.8 DHF has an RMSD fluctuation from 2 to 6 ns and an equilibrium from 6 to 9, with an average value of 0.81 Å, as shown in Fig. 5. This indicates that there is no notable difference in the stability interaction between 7,8-DHF and the control.