BV-2 cells were purchased from the Cell Bank of the Chinese Academy of Sciences. Lipopolysaccharide (LPS) was purchased from Sigma−Aldrich (St. Louis, MO, USA). Chlorogenic acid was purchased from Shanghai Yuanye Biotechnology Co. Ltd. (Shanghai, China). Anti-iNOS (1:1000, ab178945) and anti-MMP9 (1:1000, ab228402) were purchased from Abcam. Anti-CD86 (1:1000, BM4121), anti-Arg-1 (1:1000, M01106-4), anti-IL-10 (1:1000, RP1015), anti-CD206 (1:1000, A02285-2), anti-CXCL12 (1:1000, BA1389), anti-CXCR4 (1:1000, A00031-4), anti-PTGS2 (1:1000, A00084), and anti-TNF (1:1000, BA0131) were purchased from Boster Biological Technology Co. Ltd. (Shanghai, China). Cleaved caspase-3 (1:500, GB11532), IBA-1 (1:500, GB12105), and CD206 (1:400, GB113497) were purchased from Wuhan Servicebio Biotechnology Co. Ltd. (Wuhan, China). p-Akt1 (1:1000, 9018S), Akt1 (1:1000, 75692S), p-NF-κB (1:1000, 3033S), NF-κB (1:1000, 8242S), p-ERK1/2 (1:1000, 4370T), ERK1/2 (1:1000, 4696S), p-P38 (1:1000, 4511T), and P38 (1:1000, 8690T) antibodies were purchased from CST.

Male C57BL/6 mice (18-22 g, 8-10 w) were obtained from Henan Speckles Biotechnology Co. Ltd. Housing conditions for all mice included a room temperature of 23°C, 50% humidity, and a 12-hour light-dark cycle. As shown in Figure 1A , 48 mice were randomly numbered and then randomly divided into four groups of 12 mice each: Sham group (DMSO-treated group), LPS group (lipopolysaccharide-treated group), LPS+CGA group (chlorogenic acid pretreatment + lipopolysaccharide-treated group), and CGA group (chlorogenic acid-treated group). The grouping of mice and the sample size of each group were determined based on previous studies(19, 20). The LPS+CGA group and the CGA group were injected intraperitoneally with CGA (40 mg·kg⁻¹·d⁻¹) daily for 11 d. The dose of chlorogenic acid is based on previous investigations and our pre-experiments (21). From the 5th day, LPS (750 μg·kg^-1·d^-1 (22)) was given intraperitoneally for another 7 days in the LPS and LPS+CGA groups, and the daily LPS injection time was second to the CGA injection for 3 h. In addition, DMSO (5%) was injected intraperitoneally daily in the Sham and LPS groups as a solvent control. Behavioral tests were performed on days 6-10, and mice were euthanized after 11 days by administering an overdose of sodium pentobarbital. The study protocol was approved by the Experimental Animal Ethics Committee of the First Affiliated Hospital of Shihezi University School of Medicine (licence number A2018-052-01).

The neurobehavioral score assesses four components: auricular reflex, flip-right reflex, corneal reflex, caudal flexion and escape response (23). A score of “0” means the mouse has no reflex, “1” means the mouse’s reflex time is more than 1 s, and “2” means the mouse’s reflex time is 1 s. The lower the score obtained by the mice, the more severe the damage to the nervous system of the mouse brain. Neurobehavioral scores were recorded and calculated by two independent and single-blind researchers.

Three hours before the experiment, all mice were placed in the room for behavioral testing to adapt to the environment. A white experimental chamber with a size of 50 cm×50 cm×40 cm was placed for the mice to move freely in it. After setting the corresponding parameters in the software, the experimenter records the date, the number of animals and other information. Remove the mice from the cage, put them into the central area of the experimental box and leave them immediately. The trajectory of the mice in the experimental box for 10 min was monitored by the animal behavior analysis software, and the total distance the mice moved, the number of times they entered the central area, the percentage of time they were in the central area, and the percentage of time the mice moved at different speeds were recorded. The experimental chamber was cleaned with alcohol, and the next mouse was tested after an interval of 5-10 min.

Place a round platform in a round basin with a diameter of 1 m, and then add tap water of approximately 25°C to the basin, with the water surface 2-3 centimeters above the platform (22). Prior to LPS injection, we trained four groups of mice twice over a four-day period. Groups of mice were trained to find a hidden platform in the water maze within 60 s. If the mouse could not find the platform within 60 s, it was guided to the platform and allowed to stay on the platform for 10 s. The locus navigation experiment was performed on the fifth day after LPS injection, and the escape latency, i.e., the time from when the experimenter let go of the mouse to 3 s after the mouse remained on the platform, was recorded. On the next day, the platform was removed from the water maze, and the mice were subjected to spatial exploration experiments. The mice swam freely in the maze for 60 s, the trajectory of the mice was monitored, and data such as the average swimming speed, the distance swum and the percentage of total time spent in the target quadrant were recorded. The data were analyzed using Tracking Master V3.0 software (Shanghai Vanbee Intelligent Technology Co., Ltd.).

Mice were euthanized, perfused with saline via the heart, and then perfused with 4% paraformaldehyde solution. The brain tissue was carefully peeled out and fixed in 4% paraformaldehyde solution overnight. The tissue was dehydrated in ethanol and then embedded in paraffin to make 4 µm thick coronal sections of the brain. The paraffin sections were dewaxed in water, stained sequentially with hematoxylin and eosin, dehydrated and sealed, and neuronal damage in the hippocampal region of the mouse brain was observed by microscopy.

Prepared paraffin sections were dewaxed in water for antigen repair. The sections were covered uniformly with 3% BSA and blocked at room temperature for 30 min. Cleaved- caspase-3 primary antibody was added dropwise to the sections, which were incubated flat in a wet box at 4°C overnight. Then, the corresponding secondary antibody was incubated at room temperature. DAB color development solution was added dropwise, the color development time was controlled under the microscope, and the positive color was brownish yellow. Then, hematoxylin was restained for approximately 3 min, washed, and then hematoxylin differentiation solution was used to differentiate for a few seconds, washed, hematoxylin return blue solution returned to blue, and rinsed with running water. After dehydration and sealing, the sections were placed under a light microscope to observe neuronal damage in the hippocampal region of the mouse brain.

Paraffin sections were dewaxed in water and then subjected to antigen repair. The sections were blocked with 3% BSA at room temperature. A mixed primary antibody of IBA-1 and CD86 or CD206 was added dropwise to the section and incubated overnight at 4°C, with the corresponding secondary antibody incubated at room temperature. DAPI was used to restain the cell nucleus. After washing, the slices were sealed. Sections were imaged with a fluorescence microscope, and pictures were collected to observe the polarization of microglia in the hippocampal region of the mouse brain.

BV-2 cells were cultured in MEM containing 10% serum, 0.5% penicillin and 0.5% streptomycin. The cell culture conditions were 37°C and 5% CO₂. BV-2 cells were divided into the control group, the LPS group, and the LPS+CGA (50, 100, and 200 μM) group. After preintervention with CGA (50, 100, 200 μM) for 24 h, the cells were stimulated with LPS (1 μg·ml^-1) for 24 h (24).

Protein samples were electrophoresed on 10% or 12% SDS polyacrylamide gels (35-50 μg) and later transferred to PVDF membranes, which were blocked with 5% skim milk or 5% bovine serum albumin at room temperature for 2 h. Primary antibodies were added dropwise and incubated overnight at 4°C. The membranes were then incubated with the corresponding secondary antibodies at room temperature for 2 h. The PVDF membranes were incubated with enhanced chemiluminescence (ECL) reagents and exposed to a fully automated chemiluminescence image analysis system. The relative expression levels of target proteins were analyzed using ImageJ.

The treated BV-2 cells were transferred to centrifuge tubes, washed with PBS, counted, and blocked with 3% BSA for 30 min. The corresponding primary antibody was added and incubated at 37°C for 2 h. The corresponding fluorescent secondary antibody was added and incubated at 37°C for 2 h in the dark. After washing with PBS, the percentage of positive cells was detected by flow cytometry (BD, USA).

BV-2 cells were inoculated in 6-well plates and cultured until the cell density was approximately 90%. The cell layer was scratched with the tip of a 1 ml sterile pipette to form a straight line. After washing away floating cells and cell debris with PBS, interventions were performed in the above manner for different groupings. At 0 and 48 h after scratching, cell migration was observed by light microscopy, and images were collected. The number of migrating cells or the area of the trabeculae was counted using ImageJ.

BV-2 cells (4×10⁵ cells·ml^-1) were inoculated in the upper chamber of Transwell culture plates. The upper chamber was filled with MEM containing 2% serum, and the lower chamber was filled with MEM containing 10% serum. After intervention as described above, the plates were washed with PBS. After gently wiping off the unperforated cells in the upper chamber with a cotton swab, the perforated cells were fixed with 4% paraformaldehyde and stained with 0.1% crystal violet for 30 min. Four to five fields of view were randomly selected with a light microscope to count the number of perforated cells.

The expression levels of TNF-α and IL-6 in mouse brain tissue homogenates were detected using ELISA kits. Within 5 min, the optical density values at 450 nm were detected by enzyme standardization, and the standard curve was plotted. The measurements were repeated three times for both standards and samples.

In this study, the TCMSP database (http://tcmspw.com/tcmsp.php), SWISS target prediction database (http://swisstargetprediction.ch/), and QSAR database (Quantitative Structure Activity Relationship) were used to predict potential chlorogenic acid targets. The UniProt database was converted to UniPort ID (https://www.Unitprot.org/), and duplicate values were removed after merging the chlorogenic acid targets. The targets of neuroinflammation and cognitive dysfunction were obtained by searching “neuroinflammation” and “cognitive dysfunction”, and the databases included DisGeNET (https://www.disgenet.org/), G enecards database (https://www.genecards.org/), and OMIM database (https://www.omim.org/). The targets were converted to UniProt IDs using the UniProt database. Venn diagrams of common targets for chlorogenic acid, neuroinflammation and cognitive dysfunction were created using the OmicShare tool (https://www.omicshare.com/). The common targets were transformed into corresponding gene names using the UniProt database, and then protein interaction (PPI)-related information was obtained through the String database (https://string-db.org/). Visualization was performed, and the degree values of proteins and key proteins in the PPI network were obtained by Cytoscape 3.7.0 software.

In this research, molecular docking was used to evaluate the interaction between chlorogenic acid and core targets. Three-dimensional (3D) crystal structures of the targets Akt1, TNF, MMP9, PTGS2, MAPK1, MAPK14, and RELA were obtained from the Protein Data Bank (PDB), and the two-dimensional (2D) SDF format of chlorogenic acid was downloaded from the PubChem database. The docking of chlorogenic acid to each target was performed using Autodock. The specific docking process is described as follows: First, the crystal structures were retrieved and downloaded from the PDB protein database. Secondly, the receptor macromolecule was deleted with ligands and excess water molecules as the receptor structure for molecular docking. Then the rotatable bonds of the small molecule drug are identified and set to add hydrogen atoms to the protein structure. Finally, the blind docking approach we used, the docking box was selected to dock the whole protein. In particular, the docking method we chose was semi-flexible docking with 100 docking times.

The experimental results were expressed as the mean ± standard deviation. Statistical analysis was performed using SPSS software using one-way ANOVA. P values <0.05 were considered statistically significant.

We used a neurobehavioral scoring method at different periods to assess the effects of lipopolysaccharide-induced neuroinflammation on the brain. On the 3rd, 5th, and 7th days after LPS injection, the neurobehavioral scores of mice in the LPS group were lower than those in the Sham group, whereas the neurobehavioral scores of mice pretreated with chlorogenic acid were higher than those of mice in the LPS group ( Figure 1B ). Two days after LPS injection, we observed the voluntary locomotor ability of mice in each group by the open field experiment. Mouse activity trajectory plots showed that the locomotor complexity of mice in the LPS group was obviously reduced compared with that of mice in the sham group, and chlorogenic acid pretreatment was able to reverse the LPS-induced reduction in mouse activity complexity ( Figure 1C ). In addition, our data showed that the total distance moved, the number of times entering the central area, the percentage of time staying in the central area, and the duration of rapid locomotion were all reduced in the LPS group of mice compared to the Sham group, and pretreatment with chlorogenic acid improved these performances ( Figures 1D-H ). On the 6th and 7th days after LPS injection, we examined the learning and memory abilities of each group of mice by the Morris water maze test. The results of the positioning navigation test showed that the escape latency of mice in the LPS group was observably longer than that of mice in the sham group, while the escape latency of mice in the LPS+CGA group was shorter than that of mice in the LPS group ( Figures 1I, J ). The results of the space exploration test showed that LPS mice had lower swimming distance, average speed, and duration in the target quadrant than sham group mice, and chlorogenic acid-treated mice were able to attenuate the abovementioned performance caused by LPS ( Figures 1K-M ). These results suggest that chlorogenic acid can improve the cognitive dysfunction induced by LPS in mice.

The expression of inflammatory factors in mouse brain tissue homogenates was measured by ELISA to assess the inflammatory response in the brain. The results showed that LPS induced increased levels of TNF-α secretion in brain tissue, while chlorogenic acid attenuated this inflammatory response ( Figure 2A ).

The normal survival of neurons in the brain is closely related to cognitive and memory functions in mice. To demonstrate the effect of chlorogenic acid on the survival of neurons in the mouse brain, we performed HE staining on brain tissue sections of mice. The results showed that neurons in the hippocampal region of the sham group were abundant and closely arranged, with normal neuronal morphology, clear nucleus-cytoplasm demarcation, obvious nucleoli, and no obvious pathological changes. The brain tissues of mice in the LPS group showed a decrease in the number of neurons in the CA1 region of the hippocampus, a disorganized arrangement of neurons in the CA2 region, deep staining of neuronal consolidation in the CA3 region, poorly delineated nuclei and cytoplasm, and increased basophilia. The neuronal damage in the brain tissues of mice pretreated with CGA was significantly improved ( Figures 2B, C ). We further examined the apoptosis of neurons in the mouse brains by immunohistochemistry. The results showed that there was an obvious increase in brownish-yellow particles in the cytoplasm of neuronal cells in the cerebral cortex of LPS mice compared with sham mice, while there was a decrease in brownish-yellow particles in the cytoplasm of neuronal cells in the cerebral cortex of LPS+CGA mice ( Figures 2D, E ). These results suggest that chlorogenic acid can improve LPS-induced neuronal damage in the mouse brain.

Microglia, as intrinsic immune cells in the brain, are inextricably linked to central nervous system inflammation. Abnormal activation or dysregulation of the polarization ratio of microglia can cause severe inflammatory responses in the brain. We analyzed the polarization levels of microglia in the brain by immunohistofluorescence. The results showed that IBA-1⁺ expression was increased in the brains of mice in the LPS group compared to the sham group, the expression of IBA-1⁺ and CD86⁺ was increased, and the expression of IBA-1⁺ and CD206⁺ was decreased. This indicates that LPS can induce microglial activation and polarization toward the M1 phenotype. Compared with the LPS group, the expression of IBA-1⁺ in the brains of mice in the LPS+CGA group was decreased, and the expression of IBA-1⁺ and CD86⁺ was decreased, while the expression of IBA-1⁺ and CD206⁺ was increased. The above results indicated that CGA pretreatment inhibited LPS-induced microglial activation, inhibited microglial polarization toward the M1 phenotype and promoted microglial polarization toward the M2 phenotype ( Figures 3A–F ).

Next, we analyzed the morphology of microglia by Fiji, including the number of branches, junctions, endpoints, average branch length, and maximum branch length. The results showed that compared with the sham group, the microglia in the LPS group were in an activated state, the number of branches, junctions and endpoints of the cells were reduced, and their average branch length and maximum branch length were also decreased. Pretreatment with chlorogenic acid effectively inhibited microglial activation( Figures 4A–F ).

To study the effect of CGA on the polarization level of BV-2 cells, we detected the expression of M1/M2 polarization-related proteins by Western blotting. The results showed that the expression of iNOS and CD86 was elevated in the LPS group compared with the control group, i.e., LPS stimulation polarized the cells toward the M1. While the expression of iNOS and CD86 was decreased after pretreatment with different concentrations of CGA ( Figures 5A–C ), the expression of Arg-1, IL-10 and CD206 was increased ( Figures 5A, D–F ). We further detected the expression of M1/M2 polarization markers by flow cytometry. Consistent with the Western blot results, the proportion of CD86⁺ cells was increased in the LPS group ( Figures 5G, H ), and the proportion of CD206⁺ cells was decreased ( Figures 5I, J ). Compared with the LPS group, the percentage of CD86⁺ cells decreased and the percentage of CD206⁺ cells increased after pretreatment with CGA. The above results indicated that chlorogenic acid could inhibit LPS-induced M1 polarization of microglia and promote M2 polarization of microglia.

Activated microglia are capable of producing large amounts of inflammatory mediators that cause damage to the surrounding environment. The increased migratory capacity of activated microglia leads to further amplification of the inflammatory response. To investigate the effect of CGA on the migratory capacity of BV-2 cells, we examined the lateral and longitudinal migration levels of BV-2 cells by scratch and Transwell assays. The results showed that LPS stimulation upregulated the lateral and longitudinal migration levels of BV-2 cells, while CGA pretreatment effectively attenuated the migration ability of BV-2 cells ( Figures 6A–D ). We further examined the expression of migration-associated proteins by Western blotting. The results showed that LPS induced the upregulation of MMP9 expression, while CGA pretreatment was able to reduce the expression of MMP9 ( Figures 6E, F ). All these results indicated that microglial migration levels were increased in the inflammatory environment and that CGA was able to inhibit microglial migration.

To further investigate the potential mechanisms by which CGA improves neuroinflammation-induced cognitive dysfunction, we analyzed this using web-based pharmacology-related databases. A total of 650 chlorogenic acid-related targets were obtained from the TCMSP database, Swiss target prediction database, and TargetNet database. A total of 902 targets related to neuroinflammation and 650 targets related to cognitive dysfunction were obtained from the GeneCards and OMIM databases, and 117 common targets for chlorogenic acid, neuroinflammation and cognitive dysfunction were identified ( Figure 7A ). The PPI network of these 117 potential targets was constructed from the String database. The 117 targets were visualized using Cytoscape 3.8.0 according to the degree value in descending order ( Figure 7B ). The top 20 proteins were Akt1, TNF, CASP3, EGFR, MMP9, PTGS2, MTOR, PPARG, SIRT1, MAPK1, NOS3, APP, ICAM1, CXCR4, MAPK14, VCAM1, CXCL12, RELA, MMP2, and JAK.

We further performed GO and KEGG enrichment analyses on the top 20 targets. The results of GO enrichment analysis suggested that the top 20 targets were mainly involved in immune system regulation and intracellular signaling ( Figure 7C ). The results of KEGG enrichment analysis showed that the top 20 targets were enriched in the AGE-RAGE signaling pathway, TNF signaling pathway, relaxin signaling pathway, and IL-17 signaling pathways, among which the TNF signaling pathway was closely related to inflammation ( Figure 7D ). The above results suggest that the TNF signaling pathway may be a key pathway by which CGA improves cognitive dysfunction caused by neuroinflammation ( Figure 7E ). Moreover, Akt1, TNF, MMP9, PTGS2, MAPK1, MAPK14, and RELA are closely related to the development of inflammation, and they may become key targets for chlorogenic acid to exert its protective effects.

We evaluated the interaction between chlorogenic acid and Akt1, TNF, MMP9, PTGS2, MAPK1, MAPK14, and RELA by molecular docking. CGA docked with Akt1, RELA, MAPK14, MAPK1, TNFα, MMP9, and PTGS2 ( Figures 8A–G ). A small molecule-protein docking fraction <-1.2 indicates good binding between the two, so chlorogenic acid binds well to the above target molecules in docking and corroborates that chlorogenic acid may act through the above targets ( Table 1 ).

Then, we further examined the effects of chlorogenic acid on Akt1, TNF, MMP9, PTGS2, MAPK1, MAPK14, and RELA targets by Western blotting. The results revealed that LPS stimulation increased the protein expression levels of p-Akt1, p-NF-κB, p-p38, p-ERK1/2, MMP9, PTGS2, and TNFα in BV-2 cells, and pretreatment with chlorogenic acid inhibited the upregulation of these proteins ( Figures 9A–G , Figures 6E, F ). The above results indicated that chlorogenic acid was able to inhibit LPS-induced activation of Akt1, TNF, MMP9, PTGS2, MAPK1, MAPK14, and RELA targets in the TNF signaling pathway.