The mechanism of grape seed oligomeric procyanidins in the treatment of experimental autoimmune encephalomyelitis mice

Objective This study aims to investigate the mechanism of grape seed oligomeric procyanidins (GPC) in the treatment of experimental autoimmune encephalomyelitis (EAE) mice, providing pharmacodynamic materials for drug development in the treatment of multiple sclerosis (MS). Methods The constituents which have nerve protective effect in GPC were collected through literature retrieval. We used PharmMapper and STITCH database to predict drug targets, GeneCards and OMIM database to predict MS-related genes. Targets of GPC treating MS were obtained from intersected targets between drug and disease. The GO functional enrichment and KEGG pathway enrichment analysis were performed by DAVID database. EAE mouse model was used to study the therapeutic mechanism of GPC. Results Forty-two targets were discovered to be related to the process of GPC treating MS. KEGG enriched a total of 32 pathways. The pharmacological experiment showed that GPC improved the clinical symptoms of EAE mice, inhibited the expression of indicators of oxidative stress and inflammatory response in CNS, and decreased the expression of P-Akt, P-ERK, and P-JNK. Conclusion The therapeutic effect of GPC in EAE mice is associated with the suppression of MAPK and PI3K-Akt signaling pathways, providing a theoretical basis for the application of GPC inMS.


Introduction
Multiple sclerosis (MS) is an inflammatory demyelinating disease of the central nervous system (CNS), affecting over 2 million people worldwide, the main pathological features of which include inflammation, demyelination, axonal loss, and glial cell proliferation (Reich et al. 2018).
Experimental autoimmune encephalomyelitis (EAE) is the classical animal model for MS's relevant research and exhibits similar characteristics of MS (Robinson et al. 2014). At present, the drugs to treat MS are mainly hormones and immunosuppressants, with many side effects and are of high prices (Filippini et al. 2017). Therefore, many scholars are committed to find active ingredients from natural compounds and explore new therapeutic drugs (Sanadgol et al. 2017). However, there are few experiments on the treatment of MS. Also, network pharmacology can systematically elucidate the mechanism of drug's actions from multiple levels, such as targets and pathways (Liu and Sun 2012). So this study was designed to analyze the mechanism of GPC in the treatment of EAE mice by the combination of network pharmacology and experimental verification to provide basis for the clinical application of GPC in the treatment of MS. The specific process is shown in Fig. 1.

The screening of active ingredients and the prediction of targets for GPC
We used the PubMed database (https://pubmed.ncbi.nlm.nih.gov/) to retrieve the references and collect active compounds that protect the nerves in GPC. Then, target proteins were obtained from 4 databases of PharmMapper (http://www.lilab-ecust.cn/pharmmapper/) and STITCH (http://stitch.embl.de/). The targets from the PharmMapper database were filtered by a score>0.9.
Targets are standardized in the UniProt database (http://www.uniprot.org/). Cytoscape 3.6.1 software was used to map the 'drug-target' network.

Acquisition of MS-related targetgenes
GeneCards database (https://www.genecards.org/) and OMIM database (https://omim.org/) were employed to detecte MS-related target genes by searching 'multiple sclerosis'. The targets from the GeneCards database were filtered by a relevance score>6.705.

Construction of network and topological analysis
We used the Excel function to take the intersection between drug and disease targets to obtain therapeutic targets of GPC against MS, and drew Venn diagrams in the bioinformatics platform (https://www.bioinformatics.com.cn). We introduced these common genes into STRING 11.0 database (https://string-db.org/) to collect the functional relationship between the targets. Resultant data were introduced into Cytoscape 3.6.1 software to establish the PPI network among the common targets. Cytoscape 3.6.1 software was used to map the 'drug-target-disease' network.
Network analyzer in Cytoscape was utilized in analyzing topological parameters of mean and maximum degrees of freedom in PPI network among the common targets.

Biological function and pathway enrichment analyses of the core targets
The core targets obtained in 2.3 were imported into DAVID 6.8 (http://david.nifcrf.gov/)database in the format of GeneSymbol for GO analysis and KEGG pathway enrichment analysis. The important biological functions and core pathways were screened out (P<0.05, count ≥7) to draw the bubble diagram.

Administration of GPC and the evaluation of clinical score
Eighteen mice were divided into the EAE control group (n=12) and the GPC treatment group (n=6). The international standard 5-point method was adopted to observe and evaluate the symptoms of the experimental mice at regular intervals from day 0 p.i.. The evaluation criteria were as follows: 0 point were scored for the absence of any clinical symptoms;1 point for tottering gait or loss of tension with tail; 2 points for gait with partial paralysis of hindlimbs and the tail (ataxia); 3 points for total paralysis of unilateral hind limb and tail scored; 4 points for total paralysis of both hind limbs and tail scored; 5 points for near death or death. Symptoms ranged between the two criteria were measured as ± 0.5 points.

The collection of samples and analysis using different methods
On the 28 p.i., 10% chloral hydrate (0.2 ml/piece) was injected intraperitoneally for anesthesia in the sterile condition. Then, half of the mice of two groups were perfused intracardially with NS followed by 4% paraformaldehyde (PFA) in PBS (0.01 M, pH = 7.4). The spinal cord of mice was 6 quickly extracted and dehydrated by soaking in sucrose solution of 15%, 25%, and 30% for 24 hours, respectively. The coronal sections of the spinal cord (10 μm) were obtained using a cryostat microtome (Leica CM1850, Germany) and stored at 4 ℃ for Haematoxylin/eosin (H&E), Luxol fast blue (LFB) and Immunofluorescence staining. The rest of the mice were perfused intracardially with NS only. The spinal cords were removed for ELISA and western blot.

Histological staining
Pathological changes in the spinal cords were detected by H&E and LFB staining. For H&E staining, the slides were immersed in haematoxylin solution for 2 min, in water for 1 min, and then in eosin solution for 10 min. Then, differentiation was performed in 1% hydrochloric acid ethanol for 10 s.
For LFB staining, the slides were immersed in LFB at 56°C for 16 h. The excessive blue stain was removed in 95% ethanol and distilled water, and the slides were differentiated in lithium carbonate solution for 15 s. After being washed with distilled water and 80% alcohol, the slides were dehydrated with gradient ethanol and finally mounted with neutral balsam. A set of matched serial sections (3 sections/animal) were imaged under a light microscope (DM4000B, Leica, Germany).
The number of infiltrating cells (>20 mononuclear cells/field) was observed by H&E staining, and the integrated optical density of LFB staining in the spinal cord was measured in the lesion area/field using ImagePro Plus 6.0.

Immunofluorescence staining
Sections of the spinal cord were sealed with 1% BSA at room temperature for 30 min. Sections of the two groups were added with 1%BSA primary anti-MBP (1:500), respectively, and incubated overnight at 4℃. The next day, PBST was rinsed and sectioned for 3 times, then the corresponding secondary antibody was added and incubated at room temperature for 1 hour. Rinse and slice PBST for 3 times and incubate DAPI for 5 min. Glycerin was used to seal the slides, and the slides were observed by blind fluorescence microscopy. The optical density was measured by Image-Pro Plus 6.0.

ELISA
The levels of oxidative stress and inflammatory response indicators, including superoxide dismutase

Statistical method
The GraphPadPrism 7.0 statistical analysis software (Cabinet Information Technology Co., Ltd. Shanghai, China) was used for data analysis. All the data were first checked to see if it is normally distributed. One-way analysis of variance (ANOVA) was performed and followed by Tukey's posttest for multiple comparisons or Student's t-test for comparisons of 2 groups. Results are presented as the mean ± SEM. P<0.05 was considered to be statistically significant. databases, and potential targets were standardized in the Uniprot database. Cytoscape 3.6.1 software was used to build the 'compound-target' network, and the results were shown in Fig. 2. The network consisted of 85 nodes and 215 edges.

The prediction of disease target genes and the building of networks
We collected 1701 MS-related targets by retrieving 'Multiple sclerosis' in GeneCards database and OMIM database, respectively (duplicates were removed). The drug's target proteins and the disease's target genes were intersected to obtain the common targets, and the Venn diagram was drawn using

Enrichment analysis
GO belongs to the classification system of gene function, which has three branches, including

GPC alleviated clinical symptoms and central inflammation and protected myelin sheath in EAE mice
On day 11 p.i., the mice developed the disease. On day 28 p.i., samples were collected. The pattern diagram is shown in Fig 5A. Compared with the EAE control group, the clinical scores of mice in the GPC treatment group were lower (Fig. 5B). By histopathological analysis (Fig. 5C), H&E staining revealed reduced infiltration of inflammatory cells in mice of the GPC treatment group compared with that of the EAE control group (P<0.05). LFB staining showed that compared with EAE control mice, the demyelination of the GPC treatment group was reduced (P<0.01).
Immunofluorescence staining (Fig. 5D) showed that MBP mean fluorescence IOD of the GPC treatment group was higher than the EAE control group (P<0.05). Western blotting (Fig. 5E) showed that MBP relative grey value of the GPC treatment group was lower than the EAE control group (P<0.05).

Effects of GPC on markers of oxidative stress and inflammatory response
GPC has good antioxidant stress and anti-inflammatory effects. Therefore, ELISA detected the expression of related indexes in spinal cord tissue of mice. As shown in Fig. 6A, there was no significant difference in SOD expression between the two groups.

Inhibition of inflammatory signaling pathways by GPC
In order to verify the anti-inflammatory mechanism of GPC in EAE mice, we analyzed the regulatory effect of GPC on MAPK and PI3K-Akt signaling pathways according to the results of KEGG enrichment analysis. Western blot was performed to detect Akt, JNK, and ERK and their phosphorylation levels, and it is found that GPC could inhibit the phosphorylation of Akt, JNK, and ERK (Fig. 7).

Discussion
At present, the pathogenesis of MS is still unclear, and inflammatory response is a critical process in its pathogenesis (Lassmann 2018  In addition, the blood-brain barrier (BBB) between blood and brain tissue, which is composed of brain capillary endothelial cells, basal membrane, and the footplate of astrocyte protuberance (Liebner et al. 2018 Therefore, in the present study, we investigated the therapeutic effect of GPC on EAE, and the results proved that GPC has therapeutic effects. However, because the specific pathogenesis of EAE and MS are different, and EAE is a central inflammatory injury caused by artificially inducing peripheral immunity, and whether targets of GPC is in the center or the periphery needs to be further studied.

Summary
In this study, the active components, targets and action pathways of GPC were analyzed and