Systematic Deciphering of the Mechanisms of Embelin via Network Pharmacology Blood-Brain Barrier; GO: Gene Ontology; KEGG: Kyoto Encyclopedia of Genes and Genomes; CTD: Comparative Toxicogenomics Database

Background: Embelin is a major active compound of Embelia ribes , exerting a wide spectrum of pharmacological activities. However, the studies about molecular mechanisms induced by embelin have not been widely reported yet, most of which are not systematical and conclusive. Methods: First, we evaluated the druggability of embelin using the Traditional Chinese Medicine Systems Pharmacology Database (TCMSP) server. Next, the potential target genes were predicted by chemical-gene interaction analysis. Furtherly, gene ontology and pathway analyses were investigated with the target genes. Finally, we constructed a drug-target-pathway network to provide a systematic overview of the potential target genes and the mechanisms of action about embelin. Results: The pharmacokinetic properties of embelin meet all these requirements of Lipinski’s rule of five, meaning embelin could be considered as a preferable candidate for drug development. 31 target genes of embelin were identified and used to performed GO and KEGG analyses. GO, KEGG, and network analyses revealed that these targets were associated with cancers and hepatitis B. Notably, there were 9 pathways directly related to cancer among the top 10 predicted KEGG pathways, suggesting that embelin might exert dramatic effect on multiple cancers with diverse underlying mechanisms. Conclusion: Embelin could be a promising compound for the development of effective multi-targeted drug, especially for treatment of cancers. Large-scale and comprehensive research is needed to clarify our results.


Introduction
Natural products provide a resource of numerous active compounds for drug discovery. Identification and investigation of active constituent from natural plants are important for assessing their potential therapeutic effects. Embelia ribes has been widely used as an herb in traditional medicine in India and China for a long time [1]. Embelin (2,5-dihydroxy-3-undecyl-1,4-benzoquinone) is a major active compound of Embelia ribes, which belongs to the Myrsinaceae family. It could be also found in other plants, such as Radix Clematidis, Lysimachia punctata (Primulaceae) [2], and Erythrorhiza (Oxalidaceae) [3]. In recent decades, increasing evidence shows that embelin could exert multiple pharmacological activities, for example anticancer activity [4,5], anti-inflammatory activity [6], anthelmintic activity [7], antidiabetic property [8], neuroprotective effect [9,10], hepatoprotective effect [11] and so on. Embelin could be considered as a valuable compound with complex biological processes and pathways with diverse underlying mechanisms due to its effect in kinds of diseases [4].
However, the studies about molecular mechanisms induced by embelin and the resulting changes in cellular phenotypes have not been widely reported yet, most of which are not systematical and conclusive. Meanwhile, network pharmacology, which was firstly reported by Hopkins in 2007 [12], could provide a systematic overview of promising signaling pathways related to a certain compound [13]. Furthermore, target identification as well as the molecular mechanisms might be paid attention to in drug discovery and design processes. Therefore, we tried to reveal the effects of embelin by bioinformatics methods systematically. Firstly, we evaluated the druggability of embelin using TCMSP server [14].
Secondly, a chemical-gene interaction analysis was used to identify the potential target genes [15]. Furtherly, we performed gene ontology and pathway analyses with those target genes. Finally, we constructed a network including embelin, targets and pathways in order to provide an overview of the potential target genes and the mechanisms of function about embelin.

Target genes Identification
The Comparative Toxicogenomics Database (CTD, http:// ctdbase.org/), a publicly available database, could provide numerous information about chemical-gene-disease interactions relationships according to scientific literature [15]. Given a certain chemical, CTD would show the corresponding target genes immediately.

Analysis by GeneMania
GeneMANIA (http://www.genemania.org) could provide many bioinformatics services including predicting gene function, analyzing gene lists and prioritizing genes for functional assays [18]. With a gene list, GeneMANIA can prodict the proteins sharing properties or function similarly with the original content.

GO and KEGG Pathway Analyses
The web-based program Database for Annotation, Visualization and Integrated Discovery (DAVID, http:// david.abcc.ncifcrf.gov/) was used to perform functional enrichment analysis including Kyoto

Encyclopedia of Genes and Genomes (KEGG) pathway analysis and
Gene Ontology (GO) annotation for the target genes [19][20][21]. A count number larger than 2 and P-value < 0.05 were chosen as cutoff criteria. Visualization of the results of GO and KEGG analyses was done using ggplot2 package of R software [22].

Network Construction
To understand the complex relationships among embelin, target genes and pathways, we chose Cytoscape (v 3.7.1) to construct and analyze the three-layer network.

Targets Identification of Embelin
Totally, CTD identified 33 candidate target genes. Then, the 33 genes were filtered using the threshold chemical-gene interaction ≥1. Finally, 31 unique target genes for embelin remained (Table 2).
Then further investigation was based on these 31 target genes.   To study the 31 identified target genes furtherly, we performed GO and KEGG enrichment analyses using DAVID. As shown in Figure   3, the top seven functions by GO analysis were demonstrated as apoptotic process (11/31), negative regulation of apoptotic process (10/31), negative regulation of transcription from RNA polymerase II promoter (8/31), positive regulation of transcription from RNA polymerase II promoter (7/31), protein binding (31/31), identical protein binding (9/31), zinc ion binding (9/31). These functional terms were highly related to cell growth processes. As for KEGG analysis, the targets participated in top 10 pathways with gene counts adjusted P-value including "pathways in cancer", "hepatitis B", "apoptosis", "small cell lung cancer", "microRNAs in cancer", "PI3K-Akt signaling pathway", "proteoglycans in cancer", "bladder cancer", "colorectal cancer", and "p53 signaling pathway".

Network Analysis
Based on the complex relationships among embelin, target genes and pathways, we constructed an entire compound-targetpathway network by Cytoscape. As shown in Figure 4, this network has 60 nodes and 226 edges. The green oblong, red inverted triangles, and blue circles correspond to embelin, target genes and pathways, respectively.

Discussion
Development of new therapeutic agents from natural sources has great promise for diverse diseases, especially human cancers [23]. Network pharmacology approach to systematic and multitarget drug discovery could lead to a new generation of candidates with improved physicochemical and pharmacokinetics properties [24,25]. Today, Lipinski's rule of five is generally known as a criterion for drug optimization. The important properties involved in the rule should be considered for oral drugs discovery [26]. The rule of five means MW < 500 Da, ALogP < 5, as well as numbers of Hdom and Hacc less than 5 and 10, respectively. Moreover, among all pharmacokinetics properties, OB is the most important feature for orally administered drugs [14,16]. The pharmacokinetic properties of embelin met all the requirements of Lipinski's rule of five (Table 1), meaning that embelin could be considered as a preferable candidate for drug development. As is an important step in drug discovery, target identification investigates the mechanisms of effects about compounds by identifying their interacting proteins [27][28][29][30] (Graphic Abstract).
Graphic Abstract Kinds of approaches have been developed and applied widely. As listed in Table 2   The 9 pathways included "pathways in cancer", "apoptosis", "small cell lung cancer", "microRNAs in cancer", "PI3K-Akt signaling pathway", "proteoglycans in cancer", "bladder cancer", "colorectal cancer", and "p53 signaling pathway", suggesting that embelin might exert dramatic effect on multiple cancers with diverse underlying mechanisms. Previous studies have reported that embelin was effective against cancers. For instance, in lung cancer cells, activation of p53 was found to play a pivotal role in the apoptotic activity of embelin [32]. Moreover, treatment of lung cancer cells with embelin caused activation of p-p38 and p-JNK, which led to the activation of caspase-3 and cell death [33]. Additionally, embelin inhibited growth of bladder cancer cells by inducing apoptosis via PI3K/Akt pathway [34]. For the rest pathway related to hepatitis B, embelin has been explored as one of novel hepatitis B virus inhibitors [35]. These findings above closely coincide with our results from GO and KEGG analyses. The drug-target-pathway network ( Figure 4) also revealed that embelin had multiple targets and it could possess multiple pharmacological effects, consist with several well-known studies [1]. Embelin are more effective for the treatment of complex diseases, for instance, hepatitis B and tumors.
Hence, embelin could be a promising compound that might be an active candidate for future drug discovery in treatment of numbers of diseases, especially human cancers.
In conclusion, we would like to underline that embelin could be a promising compound for the development of a safe and effective multi-targeted drug. It is also hoped that this study would encourage further research on the pharmacological effects of embelin, both in vitro and in vivo, in order to provide a more in-depth insight into the mechanisms associated with the therapeutic effects of embelin.