Use of Network Pharmacology and Molecular Docking to Investigate the Mechanism by Which Ginseng Ameliorates Hypoxia

Hypoxia is a common pathological process in various clinical diseases and is characterized by abnormal changes in metabolism, function, and morphological structure of tissues resulting from insufficient oxygen supply or oxygen barriers in tissues. In particular, hypoxia in vital organs such as the brain and heart is an important cause of death^[1]. The prevention of tissue hypoxia and the treatment of hypoxia-induced tissue damage are urgent issues.

Ginseng (Panax ginseng C. A. Mey) is commonly used as a nutritional dietary supplement and additive. In recent years, there has been increasing interest and research into the pharmacological and active ingredients of ginseng. It is necessary to identify the active ingredients in ginseng before developing new drugs that target hypoxia.

In the present study, we attempted to identify the main effective components of ginseng and elucidate the biological mechanisms by which they act using a combination of network pharmacology and molecular docking, with the aim of ameliorating tissue hypoxia. Moreover, the structural docking of related proteins and compounds provides a theoretical basis for the development of new bioactive components of traditional Chinese medicine.

We identified the active ingredients of ginseng by querying phytochemical databases and screening the relevant literature. We queried the following phytochemical databases: the traditional Chinese medicine systems pharmacology database and analysis platform (TCMSP; http://ibts.hkbu.edu.hk/LSP/tcmsp.php) and the TCM Database @ Taiwan (http://tcm.cmu.edutw/). We also used computer simulations to integrate absorption, distribution, metabolism, and excretion (ADME) models to screen pharmaceutically active ingredients. The ADME model used in the present study was mainly based on oral bioavailability prediction (PreOB). Oral bioavailability (OB) is one of the most important pharmacokinetic properties of oral drugs because it embodies the efficacy of oral drug delivery into the body's circulation^[2]. We used a computer screening model (OBioavail 1.1) to calculate the OB values of the active ingredients of ginseng^[3]. Finally, compounds with OB values > 30% were considered active ingredients that required further investigation^[4]. However, it should be noted that the ginsenosides Re, Rg1, Rg2, Rb1, Rb2, and Rc are all widely regarded as active ingredients of ginseng despite having OB values < 30%. Therefore, the present study also included the ginsenosides mentioned above for further analysis of biological activity. We downloaded the molecular structure files for all the candidate active ingredients from the ChemSpider database (http://www.chemspider.com), and saved them in mol format.

The hypoxia-related potential therapeutic targets included in the present study were derived from two databases: DrugBank (http://www.drugbank.ca/) and the Online Mendelian Inheritance in Man (OMIM) database (http://www.omim.org/). We used the following search keywords: 'hypoxia', 'brain hypoxia', and 'myocardial hypoxia'. We also searched the literature to identify potential therapeutic targets, and converted all the incorporated proteins to the UniProt ID representation. We obtained X-ray crystal structures for all the candidate therapeutic targets from the RCSB protein database (http://www.pdb.org/). The protein structures were pretreated: hydrogen atoms were fixed and cocrystallized ligands and water molecules in protein-ligand complexes were removed. Finally, we used Cytoscape 3.2.0 to build a 'component-target-disease' interaction network.

To clarify the in vivo pathways involved in the treatment of hypoxia with ginseng, we used database for annotation, visualization, and integrated discovery (DAVID) software (http://david.abcc.ncifcrf.gov/home.jsp) based on the Kyoto Encyclopedia of Genes and Genomes database (KEGG, http://www.genome.jp/kegg/) for pathway enrichment analysis. We also used gene ontology (GO) analysis to conduct in-depth analysis of enrichment pathways. Finally, we found key pathways and potential therapeutic targets that would help further analog docking of target proteins and ginseng active ingredients. We used AutoDock 4.2 software to simulate molecular docking to determine the binding affinity of candidate targets with the active ingredients of ginseng.

During the database search, we used oral bioavailability (OB ≥ 30%) as the active ingredient screening standard. We identified a total of 52 potential active constituents and 713 potential targets associated with the disease, and constructed an initial interaction network (Figure 1A).

Figure 1.  'Component-target-disease' interaction network of ginseng.

Subsequently, we analyzed the three main topological parameters (Closeness Centrality, Betweenness Centrality, and Degree) of the nodes in the network; the medians of the three topological parameters were 0.25, 0.00, and 1.00, respectively. We re-screened the nodes with topological parameters greater than two times the degree median and greater than the medians of closeness centrality and betweenness centrality as the most relevant nodes. Ultimately, we obtained 52 potential targets that are usually considered targets of higher disease relevance, and constructed a new interaction network (Figure 1B). Figure 1B shows 19 yellow squares that represent the central targets for the direct influence of drugs and diseases. They are generally considered the most valuable potential targets (Supplementary Table S1, available in www.besjournal.com). The 33 circular nodes in the figure represent potential targets that drugs and diseases may affect (Supplementary Table S2, available in www.besjournal.com). Trigonal nodes represent the most relevant active ingredients, which are considered the most effective ingredients in the therapeutic treatment of hypoxia with ginseng. As shown in Figure 1B, a total of 18 active ingredients were considered to be the main effective components (Supplementary Table S3, available in www.besjournal.com).

Based on the findings described above, we performed KEGG pathway enrichment for these 52 potential targets. We found that a total of nine pathways are involved in the therapeutic effects of ginseng on human hypoxia (Supplementary Table S4, available in www.besjournal.com). Supplementary Table S4 clearly shows that the P-values of those nine pathways are all less than 0.001, indicating that they are highly correlated with the therapeutic mechanism of ginseng in the treatment of hypoxia. However, the biological effects specifically controlled by these pathways during hypoxia remain unclear. To further clarify the biological effects of the major regulatory pathways involved in the treatment of hypoxia with ginseng, we performed GO analysis of drugs and diseases (Figure 2A). Ultimately, we discovered that ginseng exerts a therapeutic effect on human hypoxia, primarily through three biological pathways: smooth muscle adaptation, positive regulation of blood vessel diameters, and positive regulation of macroautophagy. Five targets directly regulate the three pathways listed above: interleukin-1 beta, heme oxygenase 1, vascular endothelial growth factor receptor 2, epidermal growth factor receptor, and nitric oxide synthase. When used as the main research subjects, these five the 18 potential targets that directly influence targets were found to be involved in the regulation of five other pathways, including the nitric oxide metabolic pathway, nitric oxide biosynthetic pathway, reactive nitrogen species metabolic pathway, nitric oxide synthase pathway, and reactive oxygen species biosynthetic pathway (Figure 2B).

Figure 2.  Results of the GO analysis of the regulatory pathway pertaining to ginseng hypoxia treatment.

We simulated the docking of 19 active ingredients, which were predicted by network pharmacology and five candidate targets-interleukin-1 beta; heme oxygenase 1; vascular endothelial growth factor receptor 2; epidermal growth factor receptor; and nitric oxide synthase-to investigate the docking activity. We selected the components with higher docking activity to provide a theoretical basis for the development of more targeted drugs. The molecular docking information pertaining to the components with strong binding power and the candidate targets is provided in Supplementary Table S5 (available in www.besjournal.com). The mechanism by which ginseng treatment ameliorates hypoxia can be explained more profoundly, as follows. Ki is often used as an indicator of joint strength; it represents the minimum concentration at which spontaneous reactions can occur, with smaller values indicating stronger binding forces. The combination of free energy can also characterize the strength of bonding from the perspective of total energy. A negative value indicates that the reaction is favored. The greater the absolute value, the stronger the bond. We used Ki < 1 μmol/L as a condition for further component screening. We displayed images of the docking processes with strong binding forces. We also demonstrated docking for potential targets with only one dockable component.

We found that only Rg2 and aposiopolamine were capable of binding to the potential targets of the epidermal growth factor receptor and vascular endothelial growth factor receptor 2, respectively (Supplementary Figure S1, available in www.besjournal.com). However, aposiopolamine, kaempferol, and suchilactone exhibited strong binding to heme oxygenase 1 (Supplementary Figure S2, available in www.besjournal.com). Furthermore, interleukin-1 beta bound strongly to various compounds such as frutinone A and suchilactone (Supplementary Figure S3, available in www.besjournal.com). Finally, we found that nitric oxide synthase also bound strongly to four active compounds: aposiopolamine, beta-sitosterol, inermin, and suchilactone (Supplementary Figure S4, available in www.besjournal.com).

Figure Supplementary Figure S1.  Simulated molecular docking, epidermal growth factor receptor and Rg2 (A), vascular endothelial growth factor receptor and aposiopolamine (B).

Figure Supplementary Figure S2.  Simultaneous docking of analog molecules, aposiopolamine (A), suchilactone (B), and kaempferol (C) docking with Heme oxygenase 1.

Figure Supplementary Figure S3.  Simulated molecular docking, interleukin-1 beta with frutinone A (A) and suchilactone (B).

Figure Supplementary Figure S4.  Simulation analysis docking, nitric oxide synthase with aposiopolamine (A), beta-sitosterol (B), inermin (C), and suchilactone (D), respectively.

Recent studies have shown that inadequate oxygen supply to tissue can trigger the activation of multiple inflammatory pathways, exacerbating tissue damage^[5-8]. From a clinical perspective, neural hypoxia is more common than myocardial hypoxia, and neurotrophic factors, which are key regulators of neuronal survival and death, play an important role in neuron survival during hypoxia^[9]. However, the present study revealed that ginseng has a multi-component effect, a multi-target effect, and multi-path control characteristics with regard to the treatment of human hypoxia.

The biological GO analysis revealed that five potential targets were closely related to the regulation of the biological pathways. Our results suggested that ginseng has a regulatory effect on these key proteins. Our molecular docking study revealed that Rg2, aposiopolamine, kaempferol, suchilactone, frutinone A, beta-sitosterol, and inermin-all of which are found in Ginseng-bind strongly to the five candidate targets mentioned above, and may be the most important components in ginseng for hypoxia treatment.