HT22 cells derived from mouse hippocampal neuronal cell line (cat. no. CC-Y2137; Shanghai Enzyme Research Biotechnology Co., Ltd.) were utilized. Ηigh sugar medium DMEM (cat. no. 01-043-1A; Biological Industries) with 1% penicillin-streptomycin liquid (cat. no. 03-031-5B; Biological Industries) and 10% fetal bovine serum (cat. no. 504090618; Shanghai Yeasen Biotechnology Co., Ltd.) were used for cell culture at 37˚C with 5% CO₂ in a HF90 incubator (Shanghai Lishen Scientific Equipment Co., Ltd.). When the cells proliferated to the logarithmic growth period, they were transferred into 6-well plates at a concentration of 1x10⁵/ml for the subsequent experiments. G-Rb₃ (cat. no. CCPE900218; Henan Wanjia Standard Material R&D Center Co., Ltd.; https://cdn.bzwzzx.com/product/search.html?keywords=G-Rb3&pageindex=1; purity ≥99.86%) was dissolved in phosphate-buffered saline to a mother liquor concentration of 1 mmol/l. The cells were randomly divided into 4 groups: Control group, OGD/R group, G-Rb₃ high dose group (10 µmol/l) and G-Rb₃ low dose group (5 µmol/l). Since relevant cell safety experiments were already conducted in the previous study, it was found that the effective concentrations of G-Rb₃ to improve OGD/R were 2.5, 5 and 10 µmol/l (12); For the present study, it was observed that the most effective drug concentrations were 5 and 10 µmol/l and were therefore selected for the experiments.

Na₂S₂O₄ is an effective oxygen scavenger without harming the cells, thus Na₂S₂O₄ (cat. no. S817915; Shanghai Macklin Biochemical Co., Ltd.) was chosen to simulate the OGD/R model in this experiment. When the cell density increased to 80-90%, except for the control group which was left untreated, the remaining groups were incubated with 10 mmol/l Na₂S₂O₄ in sugar-free DMEM (cat. no. 01-042-1A; Biological Industries) at 37˚C for 2 h and then replaced with serum-free high-sugar DMEM at 37˚C for 2 h in order to replicate the in vitro OGD/R model. Meanwhile, in the G-Rb₃ group, the drug started being added 24 h before modelling, and continued until the end of re-glycation and reoxygenation.

HT22 cells were used to inoculate 6-well plates at a uniform inoculum density of 1x10⁵/ml (2 ml cell suspension per well). When cells proliferated to logarithmic growth phase, a total of 1 ml EDTA-free trypsin (cat. no. 15050-065; Thermo Fisher Scientific, Inc.) was added to each well at 37˚C for 1 min, followed by the addition of trypan blue (cat. no. G1019; Wuhan Servicebio Technology Co., Ltd.), which was mixed with the cell suspension at a ratio of 1:9 and was then left to stand at room temperature for 2 min. A 20 µl aliquot of the cell suspension was aspirated to a hemocytometer for observation under an inverted microscope (DMI1; Leica Microsystems Ltd.). In total, four large squares in the field of vision of the hemocytometer were selected for counting cells. Subsequently, the cell viability was calculated.

HT22 cells were used to inoculate 6-well plates at a uniform inoculum density of 1x10⁵/ml (2 ml cell suspension per well). The cells were digested with 1 ml of EDTA-free trypsin at 37˚C for 1 min and were subsequently collected. Annexin V-AbFluor^™ 488/PI apoptosis detection kit (cat. no. KTA0002; Abbkine Scientific Co., Ltd.) was utilized. The cells were first resuspended in 100 µl AnnexinV Binding Buffer diluted in deionized water to which, 5 µl of AnnexinV-AbFluor^TM488 was added, and then incubated on ice for 15 min. A total of 2 µl propidium iodide dye was pre-added. Detected was performed by flow cytometry (FACSCelesta; BD Biosciences) within 30 min. BD FACSDiva^™ software (v8.0.1.1;) was used to analyze data.

The HT22 cells of each group in the logarithmic growth phase were collected and cultured in 6-well plates at a density of 6.5x10⁵/ml for 24 h. The cells were lysed on the ice with RIPA lysis buffer (cat. no. P0013C; Beyotime Institute of Biotechnology). The supernatant was centrifuged at 12,000 x g at 4˚C for 5 min. Determination of protein concentration by BCA method (cat. no. P0010; Beyotime Institute of Biotechnology). According to the molecular weight, 10% separation gel and 5% compression gel were prepared, 20 µg protein samples were added to each lane, and separated by SDS-PAGE electrophoresis for 30 min; the electrophoresis was terminated when the desired target band reached the appropriate position. The cut PVDF membrane (0.45 µm; cat. no. IPFL85R; MilliporeSigma) was soaked in methanol for 1 min. After the transfer, the membrane was blocked with 5% skim milk (cat. no. P0216; Beyotime Institute of Biotechnology) powder for 1 h at room temperature. Subsequently, the membranes were incubated at 4˚C overnight with the following primary antibodies: Bax (1:2,000; cat. no. 50599-2-Ig; Proteintech Group, Inc.), Bcl-2 (1:1,000; cat. no. sc-7382; Santa Cruz Biotechnology, Inc.), caspase-3 (1:1,000; cat. no. 9662; Cell Signaling Technology, Inc.) and β-actin (1:1,000; cat. no. ab8226; Abcam). Then, the membranes were incubated for 1 h at room temperature with the following secondary antibodies: Goat anti-rabbit IgG (1:10,000; cat. no. ab6721; Abcam) and rabbit anti-mouse IgG (1:10,000; cat. no. ab6728; Abcam). The aforementioned membrane washing procedure was repeated. Next, the luminescent reagent liquid A and liquid B were mixed in equal volume and were applied in the membrane. After 5 min, the protein bands were detected with Tanon-6600 luminescence imaging workstation (Shanghai Tianneng Life Science Co., Ltd.). Protein expression was analyzed by using the ImageJ v2 software (National Institutes of Health) to analyze the optical density values.

HT22 cells were inoculated into 6-well plates at a concentration of 1x10⁵/ml. The cells were divided into an OGD/R group and a G-Rb₃ group, using one sample per two wells, with a total of six samples collected for analysis. Glass beads (cat. no. G8772; Shanghai Lianshuo Biotechnology Co., Ltd.) and 1,000 µl acetonitrile (cat. no. AS1121; Shanghai Yaokan Chemical Industry Co., Ltd.) were added, and the mixture was vortexed for 30 sec and then ground for 2 min at 60 Hz in a tissue grinder. The mixture was transferred to a centrifuge tube, spun at 14,000 x g for 10 min at 4˚C, the supernatant was isolated, and 300 µl of a 2-chlorophenylalanine (4 ppm) solution were accurately prepared using acetonitrile with 0.1% formic acid (cat. no. 28905; Thermo Fisher Scientific, Inc.) (1:9, v/v). The solution was used to re-dissolve the samples. The supernatant was filtered through a 0.22 µm membrane and was subsequently added to the detection vials for LC-MS analysis.

Chromatographic conditions: Waters ACQUITY (Waters Corporation) and ACQUITY UPLC^® HSS T3 (2.1x150 mm, 1.8 µm) (Waters Corporation) column were used with a flow rate of 0.25 ml/min at 40˚C. The injection volume was 2 µl. The mobile phases were: 0.1% formic acid in acetonitrile (B1) and 0.1% formic acid in water (A1) in positive ion mode, and the gradient elution program was as follows: 0-1 min, 2% B1; 1-9 min, 2-50% B1; 9-12 min, 50-98% B1; 12-13.5 min, 98% B1; 13.5-14 min, 98-2% B1; 14-12 min, 98-2% B1; 14-20 min, 2% B1. In negative ion mode, the mobile phases were acetonitrile (B2) and ammonium formate water (A2), and the gradient elution procedures were as follows: 0-1 min, 2% B2; 1-9 min, 2-50% B2; 9-12 min, 50-98% B2; 12-13.5 min, 98% B2; 13.5-14 min, 98-2% B2; 14-17 min, 2% B2.

A Thermo Q Exactive MS detector (Thermo Fisher Scientific, Inc.) with an electrospray ioniation source was used to collect data separately in positive and negative ion modes, the settings were as following: Positive ion spray voltage, 3.50 kV; negative ion spray voltage, -2.50 kV; sheath gas, 30 arbs; auxiliary gas, 10 arbs and capillary temperature, 325˚C. The primary full scan was performed at a resolution of 70,000 m/z, with a primary ion scanning range of 100-1,000 m/z. Higher-energy collisional dissociation was used for the secondary cleavage, with a collision energy of 30 eV and a secondary resolution of 17,500 m/z. The first 10 ions of the acquired signal were fragmented, with dynamic exclusion used to remove unnecessary tandem MS information.

The raw files were converted to mzXML file format using MSConvert in the ProteoWizard package (v3.0.8789), with parameters set to ‘bw=2’, ‘ppm=15’, ‘peakwidth=c’ (parameters, 5,30), ‘mzwid=0.015’, ‘mzdiff=0.01’ and ‘method=centWave’ (19-21). After obtaining the quantitative list of substances, the following databases: HMDB (22), massbank (23), LipidMaps (24), mzcloud (25) and Kyoto Encyclopedia of Genes and Genomes (KEGG) (26) were used to identify the substances, followed by data rectification. The quality control samples with relative standard deviation >30% of the substances were removed. The R language Ropls package (v1.15.0; Autodesx) was used to perform principal component analysis (PCA), partial least-square discriminant analysis (PLS-DA). The sample data were analyzed by orthogonal partial least-square discriminant analysis (OPLS-DA) to reduce dimensions. The final screening was performed via calculating the P-value and variable projection importance (VIP) value, and metabolite molecules were considered statistically significant when the P-value was <0.05 and the VIP value was >1. Pathway analysis was mainly carried out by MetaboAnalyst software (v6.0) developed by Xia-lab of McGill University in Canada, which is used for enrichment analysis of differential metabolites obtained from screening, and for browsing differential metabolites by using pathway maps in KEGG Mapper tool (27).

Firstly, according to the RNA extraction kit (cat. no. DP430; Tiangen Biotech Co., Ltd.), the RNA of OGD/R group and G-Rb₃ group was extracted, its concentration and purity were detected by a spectrophotometer and it was subsequently reversed into cDNA. The substance was diluted and mixed to a total volume of 100 µl, and was then placed on ice. According to the PCR array kit for detection of gene expression. (cat. no. wc-mRNA0263-M; Shanghai WcGene Technology Co., Ltd.; www.wcgene.cn) plates were centrifuged at 100 x g for 20 sec before use and the sealing membrane was carefully torn off at the end of centrifugation. A total of 920 µl of cDNA were prepared and mixed the components are revealed in Table I, then add the prepared mixed sample into a 96-well plate (9 µl per well), the plate was then sealed with a transparent sealing membrane, centrifuged at 100 x g for 20 sec and was finally assessed on the PCR instrument (Veriti^™ 96-Well Fast Thermal Cycler; Thermo Fisher Scientific, Inc.). The reaction system was set at 10 µl, and the conditions are shown in Table II. The real time PCR reaction was started on the machine, and the results were exported for data analysis. The PCR primer sequences were not provided by the company, although the primer sequences of the differential genes are included in Table SI.

Data are expressed as the mean ± standard deviation (SD) (n=6). Data were statistically analyzed using GraphPad Prism 8.0 software (Dotmatics) and were normally distributed. If the variances were equal, Bonferroni's multiple comparison test was performed using one-way analysis of variance (ANOVA). If the conflicts were not equal, Dunnett's multiple comparison test was used in Welch's ANOVA test. P<0.05 was considered to indicate a statistically significant difference.

The cell morphology after G-Rb₃ intervention indicated that G-Rb₃ effectively improved OGD/R injury (Fig. 1A). The results of trypan blue staining showed that cell viability was significantly higher in the drug treatment group than the model group (P<0.001), at a drug concentration of 10 µmol/l, indicating that G-Rb₃ effectively improved the survival rate of HT22 cells in the OGD/R model (Fig. 1B and E). Similarly, the apoptosis rate showed that G-Rb₃ effectively inhibited apoptosis (Fig. 1C and F). When the dose of G-Rb₃ reached 10 µmol/l, there was significant difference compared with the model group (P<0.0001). The chemical structural formula of G-Rb₃ is illustrated in Fig. 1D.

Protein expression levels of Bax, Bcl-2, caspase-3 and cleaved caspase-3 are shown in Fig. 2A. There was no significant difference in the expression of caspase-3 protein compared with either the control or model groups (Fig. 2D). The OGD/R group showed increased levels of Bax and cleaved caspase-3 protein expression (Fig. 2B and E), and decreased levels of Bcl-2 protein expression (Fig. 2C), compared with the control group. The levels of Bax and cleaved caspase-3 protein expression were decreased (Fig. 2B and E) and the levels of Bcl-2 protein expression were increased, compared with the G-Rb₃ group (Fig. 2C). The original blots of Fig. 2 are shown in Fig. S1.

The screening methods for differential metabolites mainly include PCA, PLS-DA and OPLS-DA. PCA is an unsupervised discriminant analysis method, which reveals clustering of the samples within the groups and dispersion of the samples between the groups. The results were reliable (Fig. 3A and B). The advantage of PLS-DA over PCA is that it is a supervised discriminant analysis method, in which the samples have to be specified and grouped, and the separation of groups was improved compared with the PCA analysis (Fig. 3C and D). The OPLS-DA is based on PLS-DA, with orthogonal transformation correction, which improves the model's resolving ability and validity. The samples were further resolved with OPLS-DA analysis to characterize the true differences between groups and identify biomarkers from them. In both positive and negative ion modes, intra-group sample clustering and inter-group sample dispersion could be observed in the OGD/R group and the G-Rb₃ group (Fig. 3E and F), suggesting that the results can be used for further analysis.

In this method, P-value <0.05 and VIP value >1 were used to screen for differential metabolites between groups, and a total of 31 metabolites showed significant differences (Table III). In total, 12 metabolites were upregulated: 3-indoleacetonitrile, 4-pyridoxic acid, enalaprilat, (R) 2,3-dihydroxy-3-methylvalerate, sorbitol, yohimbine, 4-fumarylacetoacetate; pimelic acid, oleamide, dihydrouracil, linoleic acid and cis-4-hydroxy-D-proline. In addition, 19 metabolites were downregulated: geranyl diphosphate, cis-aconitate acid, D-arabitol, D-lyxose, D-galactose; alpha-D-glucose, adipic acid, isocitric acid, 6-phosphogluconic acid, 3-hydroxymethylglutaric acid, linoleic acid, thiamine, aminoethylphosphate, L-glutamate-gamma-semialdehyde, geranyl diphosphate, procollagen 5-hydroxy-L-lysine, phosphoglycolic acid, guanosine and suberic acid. The screening results are presented in Table SII.

Heatmaps provide relative quantitative hierarchical clustering of differential metabolites, where clustered metabolites have similar expression patterns and may have similar functions or participate in the same metabolic processes or cellular pathways. Red color indicates the higher expression level and blue colour indicates lower expression level (Fig. 3G). The distribution of differential metabolites between the two groups of samples can be visualized using a volcano plot, where the horizontal coordinates represent the multiplicity of differences and the vertical coordinates represent the significance. Red represents metabolites with significant upregulation, blue represents metabolites with significant downregulation and gray represents metabolites with no significant differences (Fig. 3H). To observe the overall changes in metabolism, this assessment captured the average and overall changes in all metabolites in the pathway based on differential abundance scores. The differential metabolites between the G-Rb₃ and OGD/R groups mainly interacted through ABC transporters, galactose metabolism, glyoxylate and dicarboxylate metabolism, citrate cycle, linoleic acid metabolism and other different pathways affect nerves, energy metabolism and other systems together. These data illustrated that the higher the contribution of a detected metabolite under the ABC transporters and galactose metabolism pathways, the more significant is the effect of the differential metabolite on these pathways (Fig. 3I).

A total of 87 target genes in the apoptosis PCR array (Fig. 4A) were investigated, according to the ‘log₂FoldChange’ value as a reference. In total, the results revealed that 30 target genes were upregulated and 57 target genes were downregulated (Table SIII). The differential genes screened according to |log₂FoldChange |≥1 were the following: Trp63, Trp73, Dapk1, Casp14 and Cd70, all of which were downregulated in expression by the action of G-Rb₃ (Fig. 4B).