Exploring the mechanism of Radix Rhei Et Rhizome intervention in intracerebral hemorrhage based on systematic pharmacology and proteomics strategy

Abstract Objective: To explore the mechanism of Radix Rhei Et Rhizome (Dahuang, DH) intervention in intracerebral hemorrhage (ICH) based on systematic pharmacology and proteomics strategy. Methods: The systematic pharmacological strategies were utilized to find the bioactive compounds of Radix Rhei Et Rhizome, predict its potential targets, and collect ICH’s disease genes; then, the Cytoscape 3.7.1 software was applied for network construction and network topology analysis. After that, in-depth analysis of the proteomics data of Radix Rhei Et Rhizome intervention in ICH was performed to complement and validate the results of systematic pharmacological predictions. Results: A total of three major networks were constructed in the present study: (1) compound–compound target network of Radix Rhei Et Rhizome, (2) DH-ICH PPI network, (3) proteomics proteins’ PPI network. These three major networks have been analyzed by network topology, and several small networks derived (such as signaling pathway networks). The enrichment analysis showed that Radix Rhei Et Rhizome can intervene in several biological process (such as inflammation, smooth muscle proliferation, platelet activation, blood pressure regulation, angiogenesis, hypoxia, and inflammatory response of leukocytes), signaling pathway (such as FoxO signaling pathway, complement and coagulation cascades, cGMP-PKG signaling pathway, and Rap1 signaling pathway), and reactome pathway (such as signaling by interleukins, interleukin-4 and interleukin-13 signaling, nuclear receptor transcription pathway, and platelet activation). Conclusion: Radix Rhei Et Rhizome may intervene in ICH-related biological process, signaling pathway, and reactome pathway found in this research so as to achieve the effect of treating ICH related injuries.


Introduction
Intracerebral hemorrhage (ICH) refers to bleeding result from non-traumatic rupture of blood vessels in the brain parenchyma, accounting for 25-30% of all strokes. However, its acute mortality rate is 30-40%, and the survivor's invalidism rate is as high as 70%, which can leave severe complications such as paralysis, aphasia, epilepsy, and dementia, affecting the quality of life of patients [1,2]. The cause of ICH is mainly related to cerebrovascular diseases, such as hypertension with arteriolar sclerosis, microaneurysms or microhemangiomas, and cerebral vascular malformations. Its clinical manifestations vary depending 2241)", and the remaining parameters are default values. The reverse docking prediction results of each compound are downloaded, and the Z scores of the docking scores were arranged in descending order. The top 300 targets of each compound were selected for subsequent research. The UniProtKB (http://www.uniprot.org/), a database contains the accurate annotation of proteins and so on, was used for the correction of protein's names and the collection of official symbols with the species limited to (for potential targets) (Supplementary Table S1) or "Rattus norvegicus" (for proteomics data) (Supplementary Table S2). The ICH genes were collected from GeneCards database (http://www. genecards.org/) [30] and the OMIM database (http://www.ncbi.nlm.nih.gov/omim) [31]. A total of 689 ICH-related genes were obtained. The genes with relevance score> 3 were selected for sequence research (Supplementary Table  S3).
The potential targets of Radix Rhei Et Rhizome and CI genes were imported into String 11.0 (https://string-db. org/), and the species was restricted to "Homo sapiens" (for potential targets) or "Rattus norvegicus" (for proteomics data) to obtain protein-protein interaction (PPI) data [32].

Network construction and analysis methods
The active potential compounds, potential targets, ICH genes, and proteomics data were introduced into Cytoscape 3.7.1 (http://www.cytoscape.org/) [33] software to build compound-compound target network of Radix Rhei Et Rhizome, DH-ICH PPI network, proteomics proteins' PPI network, and other small networks derived from these major networks. In the network, nodes represent genes, proteins, or molecules; the connections between nodes are represented by edges, which stands for the interactions among these biological molecules [33]. Degree indicates the number of connections between nodes, while betweenness represents the number of shortest paths through a node [33]. In the DH-ICH PPI network and proteomics proteins' PPI network, the closely connected parts of the nodes are considered to be the functional area where Radix Rhei Et Rhizome plays the role of regulating the biological network, namely Clusters [33]. The Cytoscape's plug-in MCODE is used for cluster analysis of the network [15][16][17][18][19]33].

Enrichment analysis methods
The Gene Ontology (GO) enrichment analysis and KEGG signaling pathway enrichment analysis were performed by the Database for Annotation, Visualization and Integrated Discovery (DAVID) v6.8 (https://david-d.ncifcrf.gov) [34]. The reactome pathway enrichment analysis were performed by Reactome Pathway Database (https://reactome.org/) [35]. The biological processes, signaling pathways, and reactome pathways with P value <0.05 were collected for analysis.

Potential targets of Radix Rhei Et Rhizome and ICH genes
After introducing 14 potential compounds into pharmmapper for prediction, 425 potential targets were obtained. (-)-Catechin gets 295 potential targets; Aloe-emodin gets 294 potential targets; beta-sitosterol has 217 potential targets; Chrysophanol has 249 potential targets; Danthron has 172 potential targets; Daucosterol has 216 potential targets; Emodin has 293 potential targets; Eupatin has 293 potential targets; Mutatochrome has 251 potential targets; Palmidin A gets 297 potential targets; Physcion gets 294 potential targets; Rhein gets 296 potential targets; Sennoside A gets 297 potential targets; Toralactone gets 296 potential targets. Meanwhile, 423 ICH-related genes with relevance score> 3 were selected for research. There is overlap between the potential target set and the ICH gene set ( Figure 1).
The potential targets and potential compounds of Radix Rhei Et Rhizome were input into Cytoscape 3.7.1 to construct compound-compound target network of Radix Rhei Et Rhizome. This network consists of 14 compound nodes, 425 compound target nodes, and 3760 edges. In this network, some targets can be regulated by most compounds, such as:

Biological processes of DH-ICH PPI network
The DH-ICH PPI network was analyzed by the MCODE, and 20 clusters were returned (Table 1 and Figure 4). The genes and targets in the top 10 clusters were input into DAVID for GO enrichment analysis, and a lot of biological processes were obtained.
Cluster 1 is related to inflammation, smooth muscle proliferation, platelet activation, blood pressure regulation, angiogenesis, hypoxia, inflammatory response of leukocytes, vascular endothelial cells, vasodilation, vascular remodeling, and neuronal apoptosis. Cluster 2 is associated with the positive regulation of cytosolic calcium ion concentration, hypoxia, angiogenesis, vasoconstriction, blood pressure regulation, leukocyte activation and migration, platelet activation, axonal injury, synaptic transmission, neuronal apoptosis, and iron metabolism. Cluster 3 is mainly involved in platelet activation, blood pressure regulation, and glucose homeostasis. Cluster 4 is associated with platelet degranulation, hypoxia, blood pressure regulation, synaptic transmission, vasodilation, redox, and endothelial apoptosis. Cluster 5 is involved in angiogenesis. Cluster 10 is associated with coagulation, fibrinolysis, blood pressure regulation, redox, and hypoxia. Clusters 6, 7, 8, and 9 failed to return any ICH-related biological processes. The details were shown in Supplementary Table S4.
As the biological process of cluster 1 is more typical, it is shown as an example in Figure 5.

Signaling pathways of DH-ICH PPI network
The potential targets and ICH genes in DH-ICH PPI network were input into DAVID for signaling pathway enrichment analysis, and 28 ICH-related signaling pathways were obtained ( Figure 6). The P-value, fold enrichment, and count of those signaling pathways were shown in Figure 7. Meanwhile, the number of targets regulated by different compounds is different (for the detail information, see Supplementary Table S5). The compound nodes were sorted in descending order according to their degree, as follows:

Figure 2. Compound-compound target network of Radix Rhei Et Rhizome
Yellow and blue circles stand for potential compounds and potential targets, respectively. The larger the node size, the higher the degree of the node. The thicker the line, the greater the edge betweenness of the node.
Recent research also confirmed some of the findings of the present study. In terms of inhibiting inflammation, emodin can promote microglia apoptosis by inhibiting the levels of IL-1 β, TNF-α, and increasing caspase 3 and 7 [41]. In terms of oxidative stress, emodin can increase the production of reactive oxygen species (ROS) and induce apoptosis of inflammatory microglia via Akt/FOXO3 [36]. Rhein can reduce oxidative stress by inhibiting the extracellular regulated kinase (ERK)/matrix metalloproteinase-9 (MMP-9) pathway [42]. Rhein also improved the superoxide dismutase (SOD) and catalase (CAT) activities, increased glutathione (GSH) levels and the glutathione/glutathione disulfide (GSSG) ratio, and reduced levels of malondialdehyde (MDA) and GSSG in rat with traumatic brain injury [43]. In terms of protecting the blood-brain barrier, Radix Rhei Et Rhizome or its active compound (emodin, rhein, chrysophanol) can attenuate the destruction of the blood-brain barrier by increasing the expression of zonal closure protein-1 in rats with ICH [44]. Rhubarb can also maintain the integrity of the blood-brain barrier and reduce the swelling of astrocyte foot processes by inhibiting the expression of the AQP-4 gene [45]. Radix Rhei Et Rhizome can down-regulate MMP-9 and up-regulate ZO-1 by inhibiting the ERK signal pathway [46]. In terms of vasodilation, emodin attenuates the production of NO in mice after explosive-induced traumatic brain injury by inhibiting the expression and activity of inducible nitric oxide synthase (iNOS), thereby reducing brain damage and improving behavioral scores [47]. Interestingly, some compounds of Radix Rhei Et Rhizome have similar effects, and the target sets between those compounds also overlap, which may be related to the main active compound being anthraquinones. For example, chrysophanol and rhein have almost the same molecular structure, except that one methyl group of  chrysophanol is replaced by a carboxyl group; they can reduce the expression of caspase-3 and increase the activity of SOD in cerebral ischemic stroke models [48,49]. Same as rhein and chrysophanol, emodin removes one hydroxyl group and becomes chrysophanol; in cerebral ischemic stroke, they can reduce TNF-α, IL-1, and other inflammatory factors [50][51][52]. The present study also revealed the neuroprotective activity of anthraquinones from the perspective of chemoinformatics, and theoretically analyzed the mechanism of monomer compound interactions for treating ICH. In the future, further research is needed to confirm the mechanism of anthraquinone interactions in Radix Rhei Et Rhizome to treat ICH, and to find the best combination of anthraquinone monomers.
The mechanism of Radix Rhei Et Rhizome's intervention in ICH has been predicted above using network pharmacology strategies. In order to verify the above results and further explore the molecular mechanism of Radix Rhei Et Rhizome treatment of ICH, the previous proteomics data will be analyzed in depth below. The proteomics data come from reference [53].

Proteomics proteins' PPI network construction
The proteomics protein of reference [48] were shown in Supplementary Table S2 Figure 9). This network was analyzed by MCODE, and 10 clusters returned ( Figure 11B).

Enrichment analysis of proteomics proteins' PPI network
All proteomics proteins were input into DAVID and metascape (http://metascape.org/gp/index.html#/main/step1) for enrichment analysis. The biological processes, signaling pathways, and reactome pathways were shown in Supplementary Table S7 and Figure 12 Experimental studies have also shown that Radix Rhei Et Rhizome can regulate nerve-related modules, oxidative stress modules, and extracellular matrix-related modules [54][55][56][57].
Compared with the predicted results, it can be found that the two have common biological processes with high enrichment, such as coagulation module, fibrinolytic module, neuronal synaptic plasticity, inflammatory factors and inflammatory cells, calcium ion module, oxidative stress, and iron metabolism. The common signaling pathways are: complement and coagulation cascades, Rap1 signaling pathway, estrogen signaling pathway, cAMP signaling pathway, and calcium signaling pathway. The common reactome pathways are: platelet activation, signaling and aggregation, signaling by receptor tyrosine kinases, neutrophil degranulation, hemostasis, platelet degranulation, response to elevated platelet cytosolic Ca 2+ , formation of fibrin clot (clotting cascade), intracellular signaling by second messengers, innate immune system, FLT3 Signaling, signaling by VEGF, post-translational protein phosphorylation, MAPK1/MAPK3 signaling, MAPK family signaling cascades, and RAF/MAP kinase cascade. In addition, proteomics enrichment analysis also revealed more biological processes, signaling pathways, and reactome pathways, see Supplementary Table S7.

Conclusion
Radix Rhei Et Rhizome may intervene in biological process (such as inflammation, smooth muscle proliferation, platelet activation, blood pressure regulation, angiogenesis, hypoxia, and inflammatory response of leukocytes), signaling pathway (such as FoxO signaling pathway, complement and coagulation cascades, cGMP-PKG signaling pathway, and Rap1 signaling pathway), and reactome pathway (such as signaling by interleukins, interleukin-4 and interleukin-13 signaling, nuclear receptor transcription pathway, and platelet activation), so as to achieve the effect of treating ICH-related injuries.

Data Availability
The data that support the findings of this study are openly available in Supplementary Materials.