A network pharmacology-based study on the effect of Periploca forrestii Schltr.on hepatocellular carcinoma

Shuang Qiu,Peng Zhao,Xiaoxiao Huang ,Shaojiang Song

Key Laboratory of Computational Chemistry-Based Natural Antitumor Drug Research &Development,Liaoning Province,School of Traditional Chinese Materia Medica,Shenyang Pharmaceutical University,Shenyang 110016,China

Abstract Hepatocellular carcinoma (HCC) remains the sixth most common malignancy and the fourth leading cause of cancer death worldwide.Numerous Chinese herbs have been used to prevent and treat HCC.Periploca forrestii Schltr.,belonging to the family of Asclepiadaceae,has been reported to have strong inhibitory efficacy against various types of cancer cells,but the mechanism is unclear.Hence,we utilized the compound-disease target network based on the network pharmacology analysis to find out the active ingredients from P.forrestii and the potential target proteins involved in hepatocellular carcinoma.The results showed that the most effective compounds in this plant might be caffeoylquinic acids and the most potential target might be GMP synthase[glutamine-hydrolyzing] (PDB:2vxo).This study is of great significance to the further research of P.forrestii and provides a novel treatment strategy for HCC.

Keywords: Periploca forrestii Schltr.;hepatocellular carcinoma;network pharmacology

1 Introduction

Hepatocellular carcinoma (HCC) is a malignant tumor that occurs in intrahepatic bile duct epithelial cells or hepatocytes [1].It is the sixth most common malignant tumor and the fourth leading cause of cancer death.The incidence rate of HCC has increased year by year [2].Recently,traditional Chinese medicine (TCM) has been widely adopted to prevent and treat HCC in China,Japan and other Asian countries.A number of Chinese herbal compounds have been isolated,which have demonstrated significant effects on hepatocellular carcinoma [3].First-line therapy for late stage HCC includes sorafenib and lenvatinib.However,due to side effects and drug resistance,they do not always produce optimal therapeutic effects [4].Thus,the discovery of lead compounds from TCM for the treatment of HCC is urgently needed.

Periploca forrestii

Schltr.of Asclepiadaceae family is a toxic shrub mainly distributed in southwest China.The whole plant of

P.forrestii

,also known as“Heiguteng”,is a well-known Miao medicine.It was commonly used to treat rheumatoid arthritis,bruises,stomachache and dyspepsia with considerable therapeutic efficacy [5].Previous pharmacological studies have shown that the constituents of

P.forrestii

exhibit a broad range of biological activities,such as anti-inflammatory,antirheumatoid arthritis,anti-cancer,immunosuppressive and neuroprotective activities [6].It has been reported that the complex mixtures and some pure chemical components from the plants of genus

Periploca

have presented remarkable cytotoxicity activities against about ten different carcinoma cell lines,including hepatocellular carcinoma cells,colon carcinoma cells,lung carcinoma cells and breast carcinoma cells [7].Han et al.evaluated the cytotoxicity of periplogenin against S180,SGC7901,A549,LOVO,MCF7 and Hep G2 cell lines [8].They found that periplogenin had significant cytotoxic potency on all neoplasm cells and the ICvalue for all tested cells was in the range of 0.46-1.50

μ

g/mL.TCM is characterized by multiple targets,multiple pathways and multiple mechanisms,making it difficult to elucidate the pharmacological mechanisms and potential efficacy.With the rapid development of bioinformatics and computer technology,the concept of network pharmacology was coined by Hopkins [9].Based on the“diseaseprotein-drug”interaction network,network pharmacological analysis was used to observe the intervention and influence of candidate compounds and disease-associated proteins.Network pharmacology has emerged as a powerful tool to evaluate the molecular mechanisms of TCM [10].In this study,the active components and potential targets of

P.forrestii

in the treatment of HCC were screened using network pharmacology.A total of 153 natural products reported from

P.forrestii

were docked with 61 target proteins of HCC to explore the possible binding models against HCC.Results of the study may provide new ideas for the treatment of HCC and the further development of

P.forrestii.

2 Materials and methods

2.1 Compounds database building

A total of 153 reported compounds in

P.forrestii

were screened through literature survey [11-14].They were composed of steroids,terpenoids,flavonoids,quinones and phenylpropanoids.3D chemical structural formula of the ingredients was constructed with ChemBioDraw 3D,and the structure is optimized with the structural optimization plug-in included in the Sybyl-X software (version 2.0,TRIPOS Inc.).Finally,a catalog of chemical ingredients was generated with Sybyl-X (version 2.0,TRIPOS Inc.) software.

2.2 Screening and processing of target proteins

Protein targets were collected from Therapeutic Target Database (TTD) (http://db.idrblab.net/ttd/)and PubMed (https://www.ncbi.nlm.nih.gov/).The crystal structure of the target protein complex containing the original ligand was downloaded from the RCSB Protein Data Bank (RCSB PDB)database (https://www.rcsb.org/).Finally,61 target proteins were obtained from the PDB database.All the protein structures were embellished using the Sybyl-X (version 2.0,TRIPOS Inc.) software and the procedures included removing the ligands,adding hydrogen,removing water,optimizing and patching amino acids.

2.3 Molecular docking

Molecular docking is a structure-based method.It has been proved to be an important tool for predicting the interaction between molecules and biological targets in drug discovery and development [15].In this study,the Sybyl-X (version 2.0,TRIPOS Inc.) docking software was applied to molecular semi-flexible docking.Sixty-one proteins were selected as targets and the binding pockets of the ligands against the targets were established.In order to improve the accuracy of prediction results and make data processing convenient,we chose molecules with docking scores higher than 7.0.

2.4 Construction of compound-target network

The compound-target network was established using Cytoscape 3.7.1.In the graphical network,each component was described by nodes,and the edges represented the interaction between compounds and target proteins.The network was analyzed using Cytoscape plug-in CentiScaPe and the topological parameters were calculated,including the degree,betweenness,closeness and centroid.The active groups and key targets could be predicted on the basis of the obtained parameters.Key parameters include degree of network,betweenness and so on.Degree was defined as the number of links to nodes and betweenness was defined as the number of shortest paths from node to node.

2.5 Prediction of the ligand-binding mechanisms

The docking results showed that GMP synthase[glutamine-hydrolyzing] (PDB:2vxo) had the highest degree and might be the key protein.In order to confirm the binding affinity and binding site between the potential compounds and target protein,iGEMDOCK 2.1 was utilized to evaluate the binding affinity between the ligand and GMP synthase[glutamine-hydrolyzing] (PDB:2vxo) and Discovery Studio was used to investigate the intermolecular forces between candidate compounds and proteins.

3 Results

3.1 Prediction of the active compounds and potential targets

The compound-target network was constructed based on the molecular docking results to identify the active compounds and the possible key proteins(Fig.1).In the network,there were 150 nodes (92 compounds,58 target proteins) and 750 edges.The round and diamond nodes represented the compounds and the target proteins,respectively,and the size and color of the nodes indicated the degree of combination between the molecules and the targets.Edges represented the interaction between the compounds and the target proteins.Two significant topological parameters,degree and betweenness centrality,were used to evaluate the most influential nodes (compounds and target proteins).Among 92 small molecules involved in the network,nine compounds with higher value of degree and betweenness centrality were chosen,which might play strong and significant roles in targeting proteins of HCC,including (2

S

,3

S

,4

R

,10

E

)-2-[(2

R

)-2-hydroxytetracosanoylamino]-10-octadecene-1,3,4-triol (

HO135

),1-

O

-

β

-Dglucopyranosyl-(2

S

,3

S

,4

R

,10

E

)-2-[(2

R

)-2-hydroxytetracosanoylamino]-10-octadecene-3,4-diol (

HO134

),3,4-di-

O

-caffeoylquinic acid(

HP105

),6’-

O

-(

E

)-feruloyl sucrose (

HP94

),4,5-di-

O

-caffeoylquinic acid (

HP107

),(7

S

,8

R

)-dihydrodehydrodiconiferyl alcohol 9-

O

-

β

-Dapiofuranosyl-(1→6)-

O

-

β

-D-glucopyranoside(

HP117

),1,3-di-

O

-caffeoylquinic acid (

HP104

),(+)-syringaresinol-4-

O

-

β

-D-glucopyranoside(

HP115

),and 3,5-di-

O

-caffeoylquinic acid (

HP106

).The structures were shown in Fig.2 and the parameters were listed in Table 1.

Table 1 Degree and betweenness of top nine potential compounds

Fig.1 Compound-target network.The round nodes represent compounds,and the diamond nodes represent protein targets

Fig.2 Structures of the top nine potential compounds from Periploca forrestii

(to be continued)

Continued fig.2

3.2 Prediction of potential targets

The characteristics of protein nodes in molecule-target protein network map constructed with Cytoscape 3.7.1 indicated that 58 target proteins could interact with at least one molecule.Among 58 target proteins,GMP synthase [glutaminehydrolyzing] (2vxo),Protein cereblon (6boy),Nuclear receptor coactivator 2 (3up3),TGF-beta receptor type-1 (1b6c),Hepatocellular carcinomaassociated antigen 90 (6bjc) and RAC-alpha serine/threonine-protein kinase (4ejn) were selected as the candidate proteins,which were in closer contact with the molecules than other targets (Table 2).In order to better show the interaction between compounds and potential key proteins screened from the network in Fig.1,three small networks of the top three candidate proteins were constructed (Fig.3),which explicated the molecular mechanisms of active compounds.

Fig.3 Network of compounds targeting the top three candidate proteins

Table 2 Specific parameters of the six potential target proteins

3.3 Prediction of binding modes

Results showed that GMP synthase [glutaminehydrolyzing] (PDB:2vxo) was the key potential target with the highest degree in network map.To further explore the interaction between the potential compounds and GMPS,we chose three different types of high-score compounds (

HP107

,

HS45

and

HT70

) for further docking using iGEMDOCK 2.1 software.The results were shown in Table 3.Discovery Studio software was used to visualize the interaction between three different types of potential compounds and the target protein (PDB:2vxo).As shown in Fig.4,the hydrogen bonding,

π

-Alkyl and Alkyl interactions were observed between the active site residues and the potential compounds.For example,hydrogen bonds were formed between several phenolic hydroxyls of compound

HP107

and several amino acid residues,such as ALA385 and SER382.Meanwhile,the

π

-Alkyl interaction was observed between PRO441 and the benzene ring,and the

π

-

δ

interaction was observed between PRO439 and the benzene ring.In addition,two methyl groups were bound to the LFU379 and PHF645 residues by the Alkyl interaction in picture A2 (compound

HS 45

).

Table 3 Interaction energy (kcal/mol) of 2vxo and three potential compounds obtained from molecular docking

Fig.4 The binding modes between the active site residues and the top three compounds.A1,A2 and A3 represent the interactions between the amino acid residues and the top three compounds HP107,HS45 and HT70,respectively.B1,B2 and B3 represent H-bonds interaction between the active pocket and the top three compounds HP107,HS45 and HT70,respectively

(to be continued)

Continued fig.4

4 Discussion

In recent years,TCM has gradually attracted more attention thanks to its low toxicity and precise efficacy.Understanding the molecular basis of TCM is crucial for its modernization,which will provide sufficient theoretical support for TCM research [16].Network pharmacology provides a simple method to find potential bioactive compounds by mapping chemical compounds into the disease-protein network,and reveals the underlying complex relationship between multi-components and multi-targets.To date,it has been applied to study many traditional Chinese herbs and herbal prescriptions [17].

In the present study,153 compounds isolated from

P.forrestii

were docked with 61 HCC target proteins.Ninety-two compounds were found to possess good interaction at the active sites.We chose the top nine compounds with higher scores for further discussion,which can be divided into three categories according to their structure types,including ceramides derivatives (

HO134

,

HO135

),caffeoylquinic acids (

HP104

,

HP105

,

HP106

,

HP107

) and lignans.Kurata Rie et al.investigated the inhibitory activity of caffeic acid,di-and tricaffeoylquinic acids on selected human cancer cells,including a stomach cancer (Kato III),a colon cancer (DLD-1),and a promyelocytic leukemia cell (HL-60) [18].The results showed that caffeic acid,di-and tri-caffeoylquinic acids could depress the proliferation of cancer cells in a dosedependent manner.Our research showed that such ingredients might also have good anti-liver cancer activity.However,further cell level assay or

in

vivo

experimental research is still needed.

In addition,according to the network analysis,GMP synthase [glutamine-hydrolyzing] (PDB:2vxo) showed significant activity compared with other targets,which indicated that it might be the most potential key protein related to small molecules.Kerstin Holzer et al.Found that GMP synthase [glutamine-hydrolyzing],a key enzyme of de novo purine biosynthesis,was an important p53 repression target.As a major barrier to prevent malignant transformation and progression,the tumor-suppressor protein p53 plays an important role in HCC [19,20].In order to treat HCC better,GMP synthase [glutamine-hydrolyzing] enzyme inhibitors deserve further research and development.

5 Conclusion

In this paper,the biological network between the compounds and HCC associated proteins was constructed,and the active ingredients,potential targets and binding modes of the candidate compounds derived from

P.forrestii

related to HCC were predicted.Based on the compound-target network,we can infer that ceramides derivatives and caffeoylquinic acids may have potential antihepatoma activity.Meanwhile,GMP synthase[glutamine-hydrolyzing] (PDB:2vxo) may be the most potential key protein in HCC.The results of this study still need further pharmacological experiments to prove.