Investigation on the active ingredient and mechanism of Cannabis sativa L. for treating epilepsy based on network pharmacology

Abstract

Cannabis sativa L. (cannabis) is a medicinal plant and has been used for many years for the treatment of epilepsy (EP), which is a common neurological disease. This study aimed to investigate the mechanism of cannabis action in EP, with emphasis on the leading compounds, targets and pathways. In this study, systematic pharmacology and bioinformatics approaches were employed to identify the active ingredients and potential targets of cannabis for treating EP. Furthermore, network construction, Gene Ontology (GO) and Kyoto Encyclopedia of Genes and Genomes (KEGG) enrichment analysis, and molecular docking were used to elucidate the mechanism of cannabis against EP. A total of 360 compounds were collected in this work. Among them, 226 active compounds and 116 predicted targets were obtained based on absorption, distribution, metabolism and excretion (ADME) screening and databases, respectively. Among the 226 active compounds, most were cannabinoids. The topological analysis showed that cannabinoid receptor 1, albumin and glycogen synthase kinase-3 beta (CNR1, ALB and GSK3B) were the key targets with intense interaction. The GO and KEGG enrichment analysis suggested cannabis might produce the antiepileptic effects by regulating many pathways, including calcium signalling pathway, MAPK signalling pathway, GABAergic synapse, etc. Additionally, cannabinol methyl ether (M54) might be the leading compound based on molecular docking. Consequently, this study holistically illuminates the active constituents and mechanism of cannabis based on network pharmacology, which contributes to searching for leading compounds and development of new drugs in the treatment of EP.

Supplemental data for this article is available online at https://doi.org/10.1080/13102818.2021.1942208 .

Keywords:

• Cannabis sativa L
• epilepsy
• network pharmacology
• molecular docking
• antiepileptic activity

Introduction

Epilepsy (EP) is a nervous system disease with transient, stereotyped, recurrent and repetitive characteristics, which is caused by sudden abnormal discharge of brain neurons. According to the meta-analysis of international studies, the prevalence of EP is 6.4 per 1,000 people, and the annual incidence is 67.8 per 100,000 people [1]. There are more than 50 million EP patients in the world, and the mortality rate is 3 times that of the general population [2]. Due to the causes above, EP has increasingly attracted people’s attention recently. The pathogenesis of EP has been intensively studied, as relating signalling pathways include AK signalling pathway, BDNF signalling pathway, mTOR signalling pathway, TGF signalling pathway, calcium signalling pathway and MAPK signalling pathway [3].

Cannabis sativa L. (cannabis), an annual dioecious herb of the genus Cannabis of the Cannabaceae family, has been extensively used as a traditional herbal medicine for the treatment of EP and neuropathic pain. It mainly contains various chemical components such as cannabinoids, flavonoids, terpenoids and so on, leading to a mass of pharmacological effects with a wide range. Most importantly, the cannabidiol (CBD) component of cannabis has been reported to treat EP [4, 5]. However, other active components and mechanisms of cannabis therapy for EP have not been fully elucidated.

Network pharmacology, an attractive concept put forward by Hopkins [6], integrates the ideas of systems biology and multi-directional pharmacology. It analyzes the action mechanism of drugs by constructing a complex network relationship of compound–target–disease. Recently, pharmacological research has shifted from the traditional search for a single target to multiple targets multi-level comprehensive network research [7, 8]. It promotes further research on Traditional Chinese Medicine (TCM) and expands the development of modern drugs. In addition, molecular docking is a theoretical simulation method based on receptor characteristics and molecular interaction [9]. It can predict the binding sites and affinity between drug molecules and target proteins, so as to verify the experimental results of network pharmacology [10].

To explore the pharmacological and molecular mechanism of antiepileptic activity of cannabis at the overall level, the network pharmacology analysis was employed (Figure 1). Firstly, the chemical database of cannabis was constructed, and the active compounds were screened according to SwissADME. Second, the targets of compounds and EP were predicted from five databases, involving Swiss TargetPrediction, PharmMapper, GeneCards, Therapeutic Target Database (TTD) and Online Mendelian Inheritance in Man (OMIM). The common targets were considered as the candidate antiepileptic targets of cannabis. Furthermore, networks were constructed by Cytoscape, and Gene Ontology (GO) and Kyoto Encyclopedia of Genes and Genomes (KEGG) were performed to clarify the mechanism of the cannabis compounds against EP. Finally, the interaction between key compounds and targets was analyzed using molecular docking.

Figure 1. Workflow of the network pharmacological investigation on the potential use of cannabis in EP treatment.

Materials and methods

Chemical ingredients database building

All chemical ingredients from cannabis were collected from Traditional Chinese Medicines for Systems Pharmacology database and analysis platform (TCMSP, available online: http://lsp.nwsuaf.edu.cn/) [11] and related literatures.

Active ingredients screening

The molecular properties of compounds were determined using SwissADME database (available online: http://www.swissadme.ch/index.php) [12]. The selection for candidate compounds involved gastrointestinal absorption (GI absorption) and drug likeness rules (Lipinski, Ghose, Veber, Egan and Muegge). For the screening indicator, the GI absorption was considered as high, and two or more indices of drug likeness rules had to be Yes.

Target prediction and validation

Target prediction is helpful to clarify the underlying mechanism of action of cannabis. Swiss TargetPrediction (available online: http://www.swisstargetprediction.ch/) and PharmMapper (available online: http://www.lilab-ecust.cn/pharmmapper/) databases were used to retrieve and validate related targets of compounds. Then, the target name–gene name standard conversion was performed using the UniProt database (available online: https://www.uniprot.org/), so that the names of each target were corrected to the official names. After that, the candidate targets for the treatment of EP were screened by searching the GeneCards (available online: https://www.genecards.org/), TTD (available online: http://db.idrblab.net/ttd/) and OMIM (available online: https://omim.org/) databases [13]. The Venny 2.1 platform was used to draw a Veen diagram showing the overlapping of cannabis targets and EP targets. The overlapping was considered for potential therapeutic targets of cannabis against EP.

Compound-target (C-T) network construction

For a scientific explanation of the pharmacological mechanism of cannabis, two networks were established as follows: (1) the C–T network was established by linking active ingredients with their targets that were related to EP; (2) the C-T network of cannabinoids and their related targets was constructed.

GO and KEGG enrichment

In organisms, genes cannot perform their functions independently. Different genes coordinate with each other to finish a series of biochemical reactions to perform their biological functions. Therefore, GO and KEGG pathway enrichment analysis were used to provide a more systematic and comprehensive understanding of the mechanism of action. The Metascape platform [14] was applied to thoroughly analyze enrichment information. GO/KEGG terms with p < 0.05 were considered significantly enriched. The top 20 pathways containing the number of targets were selected. The Target–Pathway (T–P) network was constructed to link compounds and pathways with related targets by Cytoscape [15], and the network topological properties were analyzed by Network Analyzer. In addition, the KEGG Mapper tool (available online: https://www.kegg.jp/kegg/mapper.html) was used for the construction of the pathway map.

Molecular docking

Active ingredients that were in the top ten and the top five targets in terms of degree value were screened based on the C–T network. The structures of the target proteins and the compounds were obtained from the PDB (available online: http://www.rcsb.org/) and PubChem (available online: http://www.ncbi.nlm.gov/pccompond) databases, respectively. Both of them were introduced into Schrodinger for molecular docking, and the docking score was analyzed to predict and evaluate the interaction between compounds and targets (the lower the docking score, the more negative number, and the better the binding affinity).

Results

Classification of active compounds and predicted targets

In this research, a total of 360 compounds were found in cannabis from the literatures and TCMSP database. Among them, 235 active compounds were selected through SwissADME database. The screening criteria were GI absorption shown as ‘high’, and at least two rules being ‘yes’ in the drug likeness rules (Table 1). In addition, a total of 1117 targets were predicted through Swiss TargetPrediction and PharmMapper databases using these active compounds, and 1403 EP-related targets were retrieved from the GeneCards, TTD and OMIM databases. By comparing these targets with the predicted cannabis targets, 116 common targets were filtered as the key targets for researching on the antiepileptic activity of the cannabis compounds (Figure 2). Nine ingredients that did not have any relevant targets were removed. The detailed information of the 116 candidate targets is displayed in Table 2.

Figure 2. Venn diagram between cannabis targets and epileptic targets.

Table 1. 235 potential active ingredients and ADME parameters of Cannabis sativa L.

No.	Name	GI absorption	Druglikeness
			Lipinski	Ghose	Veber	Egan	Muegge
M1	(-)-Δ^9-trans-Tetrahydrocannabinol	High	Yes	No	Yes	Yes	No
M2	(-)-Δ^9-trans-Tetrahydrocannabinolic acid A	High	Yes	Yes	Yes	Yes	No
M3	(-)-Δ^9-trans-Tetrahydrocannabinolic acid B	High	Yes	Yes	Yes	Yes	No
M4	(-)-Δ^9-trans-Tetrahydrocannabinol -C4	High	Yes	Yes	Yes	Yes	No
M5	(-)-Δ^9-trans-Tetrahydrocannabinolic acid A-C4	High	Yes	Yes	Yes	Yes	No
M6	(-)-Δ^9-trans-Tetrahydrocannabivarin	High	Yes	Yes	Yes	Yes	No
M7	(-)-Δ^9-trans-Tetrahydrocannabivarinic acid	High	Yes	Yes	Yes	Yes	No
M8	(-)-Δ^9-trans-Tetrahydrocannabiorcol	High	Yes	Yes	Yes	Yes	Yes
M9	(-)-Δ^9-trans-Tetrahydrocannabiorcolic acid	High	Yes	Yes	Yes	Yes	Yes
M10	8a-Hydroxy-Δ^9-trans-tetrahydrocannabinolate	High	Yes	Yes	Yes	Yes	No
M11	8b-Hydroxy-Δ^9-trans-tetrahydrocannabinolate	High	Yes	Yes	Yes	Yes	No
M12	11-Acetoxy-Δ^9- trans-tetrahydrocannabinolic acid A	High	Yes	Yes	Yes	Yes	No
M13	Δ^9-THC aldehyde	High	Yes	Yes	Yes	Yes	No
M14	8-Oxo-Δ^9-trans -tetrahydrocannabinol	High	Yes	Yes	Yes	Yes	No
M15	β-Fenchyl-Δ^9-trans-tetrahydrocannabinolate	High	Yes	No	Yes	No	No
M16	α-Fenchyl-Δ^9-trans-tetrahydrocannabinolate	High	Yes	No	Yes	No	No
M19	α-Terpenyl-Δ^9-trans-tetrahydrocannabinolate	High	Yes	No	Yes	No	No
M24	(-)-D^9-cis-(6aS, 10aR)-Tetrahydrocannabinol	High	Yes	No	Yes	Yes	No
M25	Cannabitetrol	High	Yes	Yes	Yes	Yes	Yes
M26	Δ^8-Tetrahydrocannabinol	High	Yes	No	Yes	Yes	No
M27	Δ^8-Tetrahydrocannabinolic acid A	High	Yes	Yes	Yes	Yes	No
M28	10a-Hydroxy-D^8-tetra-hydrocannabinol	High	Yes	Yes	Yes	Yes	No
M29	10b-Hydroxy-D^8-tetrahydrocannabinol	High	Yes	Yes	Yes	Yes	No
M30	10aa-Hydroxy-10-oxo-D^8-tetrahydrocannabinol	High	Yes	Yes	Yes	Yes	Yes
M31	(-)-trans-Cannabitriol	High	Yes	Yes	Yes	Yes	Yes
M32	(+)-trans-Cannabitriol	High	Yes	Yes	Yes	Yes	Yes
M33	(±)-cis-Cannabitriol	High	Yes	Yes	Yes	Yes	Yes
M34	(-)-trans-10-Ethoxy-9-hydroxy-D6a(10a)-tetrahydrocannabinol	High	Yes	Yes	Yes	Yes	Yes
M35	(±)-trans-Cannabitriol-C3	High	Yes	Yes	Yes	Yes	Yes
M36	Cannabitriol-C3- homologue	High	Yes	Yes	Yes	Yes	Yes
M37	(-)-trans-10-Ethoxy-9-hydroxy-D^6a(10a)-tetrahydrocannabivarin-C3	High	Yes	Yes	Yes	Yes	Yes
M38	8,9-Dihydroxy-D^6a(10a)-tetrahydrocannabinol	High	Yes	Yes	Yes	Yes	Yes
M40	10-Oxo-D^6a(10a)-tetrahydrocannabinol	High	Yes	Yes	Yes	Yes	Yes
M41	Cannabiripsol	High	Yes	Yes	Yes	Yes	Yes
M42	9β,10β-Epoxyhexahydrocannabinol	High	Yes	Yes	Yes	Yes	No
M43	9α-Hydroxyhexahydrocannabinol	High	Yes	Yes	Yes	Yes	No
M44	7-Oxo-9α-hydroxyhexahydrocannabinol	High	Yes	Yes	Yes	Yes	Yes
M45	10α-Hydroxyhexahydrocannabinol	High	Yes	Yes	Yes	Yes	No
M46	10aα-Hydroxyhexahydrocannabinol	High	Yes	Yes	Yes	Yes	No
M47	9α-Hydroxy-10-oxo-D^6a,10a-THC	High	Yes	Yes	Yes	Yes	Yes
M48	10α-Hydroxy-D^9,11-hexahydrocannabinol	High	Yes	Yes	Yes	Yes	No
M49	(±)-D^7-cis-Isotetrahydrocannabivarin-C3	High	Yes	No	Yes	Yes	No
M50	(-)-D^7trans-(1R,3R,6R)-Isotetrahydrocannabivarin-C3	High	Yes	No	Yes	Yes	No
M51	(-)-D^7-trans-(1R,3R,6R)-Isotetrahydrocannabinol-C5	High	Yes	No	Yes	Yes	No
M52	Cannabinol	High	Yes	No	Yes	Yes	No
M53	Cannabinolic acid	High	Yes	Yes	Yes	Yes	No
M54	Cannabinol methyl ether	High	Yes	No	Yes	No	No
M55	Cannabinol-C4	High	Yes	Yes	Yes	Yes	No
M56	Cannabivarin	High	Yes	Yes	Yes	Yes	No
M57	Cannabinol-C2	High	Yes	Yes	Yes	Yes	Yes
M58	Cannabiorcol-C1	High	Yes	Yes	Yes	Yes	Yes
M59	8-Hydroxy cannabinolic acid A	High	Yes	No	Yes	Yes	No
M60	8-Hydroxycannabinol	High	Yes	No	Yes	Yes	No
M61	4-Terpenyl cannabinolate	High	Yes	Yes	Yes	Yes	Yes
M62	1’S-Hydroxy-cannabinol	High	Yes	Yes	Yes	Yes	Yes
M63	Cannabinodiol	High	Yes	No	Yes	Yes	No
M64	Cannabinodivarin	High	Yes	Yes	Yes	Yes	No
M65	Cannabidiol	High	Yes	No	Yes	Yes	No
M66	Cannabidiolic acid	High	Yes	Yes	Yes	Yes	No
M67	Cannabidiol monomethyl ether	High	Yes	No	Yes	No	No
M68	Cannabidiol-C4	High	Yes	Yes	Yes	Yes	No
M69	Cannabidivarin	High	Yes	Yes	Yes	Yes	No
M70	Cannabidivarinic acid	High	Yes	Yes	Yes	Yes	No
M71	Cannabidiorcol	High	Yes	Yes	Yes	Yes	Yes
M72	(±)-Cannabichromene	High	Yes	No	Yes	No	No
M73	(±)-Cannabichromenic acid	High	Yes	No	Yes	Yes	No
M74	(±)-Cannabivarichromene	High	Yes	Yes	Yes	Yes	No
M75	(±)-Cannabivarichromevarinic acid	High	Yes	Yes	Yes	Yes	No
M76	(+)-Cannabichromevarin	High	Yes	Yes	Yes	Yes	No
M77	(±)-4-Acetoxycannabichromene	High	Yes	No	Yes	Yes	No
M78	2-Methyl-2-(4-methyl-2-pentyl)-7-propyl-2H-1-benzopyran-5-ol	High	Yes	Yes	Yes	Yes	No
M79	(±)-3″-Hydroxy-D^4″-cannabichromene	High	Yes	Yes	Yes	Yes	No
M80	(-)-7-Hydroxycannabichromane	High	Yes	Yes	Yes	Yes	No
M81	Cannabichromanone-C5	High	Yes	Yes	Yes	Yes	Yes
M82	Cannabichromanone-C3	High	Yes	Yes	Yes	Yes	Yes
M83	Cannabichromanone B	High	Yes	Yes	Yes	Yes	No
M84	Cannabichromanone C	High	Yes	Yes	Yes	Yes	Yes
M85	Cannabichromanone D	High	Yes	Yes	Yes	Yes	No
M86	Cannabigerol	High	Yes	No	Yes	No	No
M87	Cannabigerolic acid	High	Yes	No	Yes	No	No
M88	Cannabigerol monomethylether	High	Yes	No	Yes	No	No
M90	Cannabigerovarinic acid A	High	Yes	Yes	Yes	Yes	No
M91	Cannabigerovarin	High	Yes	Yes	Yes	Yes	No
M95	Cannabinerolic acid A	High	Yes	Yes	Yes	Yes	No
M96	Camagerol	High	Yes	Yes	Yes	Yes	No
M98	(±)-6,7-trans-Epoxycannabigerolic acid	High	Yes	Yes	Yes	Yes	No
M99	(±)-6,7-cis-Epoxycannabigerolic acid	High	Yes	Yes	Yes	Yes	No
M100	(±)-6,7-trans-Epoxycannabigerolic	High	Yes	Yes	Yes	Yes	No
M101	(±)-6,7-cis-Epoxycannabigerolic	High	Yes	Yes	Yes	Yes	No
M102	Cannabielsoic acid A	High	Yes	Yes	Yes	Yes	Yes
M103	Cannabielsoin	High	Yes	Yes	Yes	Yes	Yes
M104	Cannabielsoic acid B	High	Yes	Yes	Yes	Yes	Yes
M105	C3-Cannabielsoic acid B	High	Yes	Yes	Yes	Yes	Yes
M106	C3-Cannabielsoin	High	Yes	Yes	Yes	Yes	Yes
M107	Dehydrocannabifuran	High	Yes	No	Yes	No	No
M108	Cannabifuran	High	Yes	No	Yes	No	No
M109	8-Hydroxy-isohexahydrocannabivirin	High	Yes	Yes	Yes	Yes	Yes
M110	Cannabicyclol	High	Yes	No	Yes	Yes	No
M111	Cannabicyclolic acid	High	Yes	No	Yes	Yes	No
M112	Cannabicyclovarin	High	Yes	No	Yes	Yes	No
M114	(-)-(7R)-Cannabicoumarononic acid	High	Yes	No	Yes	No	No
M115	2-Geranyl-5-hydroxy-3-npentyl-1,4-benzoquinone	High	Yes	Yes	Yes	Yes	No
M116	5-Acetoxy-6-geranyl-3-npentyl-1,4-benzoquinone	High	Yes	Yes	No	Yes	No
M117	Cannabicitran	High	Yes	No	Yes	Yes	No
M118	4-Actoxy-2-geranyl-5-hydroxy-3-n-pentylphenol	High	Yes	No	No	Yes	No
M119	Cannabimovone	High	Yes	Yes	Yes	Yes	Yes
M120	Cannabioxepane	High	Yes	No	Yes	No	No
M121	Vitexin	High	Yes	Yes	Yes	Yes	Yes
M122	Isovitexin	High	Yes	Yes	Yes	Yes	Yes
M123	apigenin-6,8-di-C-β-D-glucopyranoside	High	Yes	Yes	Yes	Yes	Yes
M125	Cannflavin B	High	Yes	Yes	Yes	Yes	No
M129	Orientin	High	Yes	Yes	Yes	Yes	Yes
M130	Orientin-7-O-rhamnoglucosid	High	Yes	Yes	Yes	Yes	Yes
M131	6-Geranylapigenin	High	Yes	Yes	Yes	Yes	Yes
M132	Apigenin	High	Yes	Yes	Yes	Yes	Yes
M133	Chrysoeriol	High	Yes	Yes	Yes	Yes	Yes
M134	Apigenin-7-O-glucoside	High	Yes	Yes	Yes	Yes	Yes
M135	Luteolin	High	Yes	Yes	Yes	Yes	Yes
M136	Luteolin-7-O-α-D-glucoside	High	Yes	Yes	Yes	Yes	Yes
M137	2″-O-Glucopyranosylvitexin	High	Yes	Yes	Yes	Yes	Yes
M138	2″- O-Glucopyranosylorientin	High	Yes	Yes	Yes	Yes	Yes
M139	Kaempferol	High	Yes	Yes	Yes	Yes	Yes
M140	Kaempferol 3-O-sophoroside	High	Yes	Yes	Yes	Yes	Yes
M142	Quercetin	High	Yes	Yes	Yes	Yes	Yes
M143	Quercetin-O-glucoside(s)	High	Yes	Yes	Yes	Yes	Yes
M144	Quercetin-3-O-sophoroside	High	Yes	Yes	Yes	Yes	Yes
M145	trans-Anethol	High	Yes	No	Yes	Yes	No
M146	cis-Anethol	High	Yes	No	Yes	Yes	No
M147	iso-Eugenol	High	Yes	Yes	Yes	Yes	No
M148	Eugenol	High	Yes	Yes	Yes	Yes	No
M149	Methyleugenol	High	Yes	Yes	Yes	Yes	No
M150	N-p-coumaroyltyramine	High	Yes	Yes	Yes	Yes	Yes
M151	N-trans-caffeoyltyramine	High	Yes	Yes	Yes	Yes	Yes
M152	N-trans-feruloyltyramine	High	Yes	Yes	Yes	Yes	Yes
M174	Coumaroylaminobutanol glucopyranoside	High	Yes	Yes	Yes	Yes	Yes
M175	Cannithrene 1	High	Yes	Yes	Yes	Yes	Yes
M176	Cannithrene 2	High	Yes	Yes	Yes	Yes	Yes
M177	4,5-Dihydroxy-2,3,6-trimethoxy-9,10-dihydropenanthrene	High	Yes	Yes	Yes	Yes	Yes
M178	4-Hydroxy-2,3,6,7-tetramethoxy-9,10-dihydropenanthrene	High	Yes	Yes	Yes	Yes	Yes
M179	Dihydroresveratrol	High	Yes	Yes	Yes	Yes	Yes
M180	3,4″-Dihydroxy-5-methoxy bibenzyl	High	Yes	Yes	Yes	Yes	Yes
M181	3,3″-Dihydroxy-5,4″-dimethoxy bibenzyl	High	Yes	Yes	Yes	Yes	Yes
M182	Cannabistilbene llb	High	Yes	Yes	Yes	Yes	Yes
M183	Cannabistilbene lla	High	Yes	Yes	Yes	Yes	Yes
M184	Cannabistilbene l	High	Yes	Yes	Yes	Yes	No
M185	3,4″-Dihydroxy-5,3″-dimethoxy-5″-isoprenyl	High	Yes	Yes	Yes	Yes	No
M186	Canniprene	High	Yes	Yes	Yes	Yes	No
M187	Cannabispirone
M188	iso-Cannabispirone	High	Yes	Yes	Yes	Yes	Yes
M189	Cannabispirol	High	Yes	Yes	Yes	Yes	Yes
M190	Acetylcannabispirol	High	Yes	Yes	Yes	Yes	Yes
M191	7-Hydroxy-5-methoxyindan-1-spiro-cyclohexane	High	Yes	Yes	Yes	Yes	Yes
M192	5-Hydroxy-7-methoxyindan-1-spiro-cyclohexane	High	Yes	Yes	Yes	Yes	Yes
M193	5,7-Dihydroxyindan-1-spiro-cyclohexane	High	Yes	Yes	Yes	Yes	Yes
M194	α-Cannabispiranol 4′-O-β-D-glucopyranose	High	Yes	Yes	Yes	Yes	Yes
M195	Cannabispirenone-A	High	Yes	Yes	Yes	Yes	Yes
M196	Cannabispirenone-B	High	Yes	Yes	Yes	Yes	Yes
M197	Cannabispiradienone	High	Yes	Yes	Yes	Yes	Yes
M200	Cannabispirketal	High	Yes	Yes	Yes	Yes	Yes
M222	1,4-Cineol	High	Yes	No	Yes	Yes	No
M223	1,8-Cineol	High	Yes	No	Yes	Yes	No
M226	Borneol	High	Yes	No	Yes	Yes	No
M227	Bornyl acetate	High	Yes	Yes	Yes	Yes	No
M229	Camphenehydrate	High	Yes	No	Yes	Yes	No
M230	Camphor	High	Yes	No	Yes	Yes	No
M231	Carvacrol	High	Yes	No	Yes	Yes	No
M232	Carvone	High	Yes	No	Yes	Yes	No
M233	cis-Linalool oxide	High	Yes	No	Yes	Yes	No
M234	cis-α-Ocimene	High	Yes	No	Yes	Yes	No
M235	Citral B	High	Yes	No	Yes	Yes	No
M236	Citronellol	High	Yes	No	Yes	Yes	No
M237	Dihydrocarveyl acetate	High	Yes	Yes	Yes	Yes	No
M238	Dihydrocarvone	High	Yes	No	Yes	Yes	No
M239	Fenchone	High	Yes	No	Yes	Yes	No
M240	Fenchyl alcohol	High	Yes	No	Yes	Yes	No
M241	Geraniol	High	Yes	No	Yes	Yes	No
M242	Geranyl acetone	High	Yes	Yes	Yes	Yes	No
M243	Ipsdienol	High	Yes	No	Yes	Yes	No
M245	Linalool	High	Yes	No	Yes	Yes	No
M247	Nerol	High	Yes	No	Yes	Yes	No
M248	Nerolidol	High	Yes	Yes	Yes	Yes	No
M250	p-Cymene-8-ol	High	Yes	No	Yes	Yes	No
M251	Perillene	High	Yes	No	Yes	Yes	No
M252	Pinocarveol	High	Yes	No	Yes	Yes	No
M253	Pinocarvone	High	Yes	No	Yes	Yes	No
M254	Piperitenone oxide	High	Yes	Yes	Yes	Yes	No
M255	Piperitenone	High	Yes	No	Yes	Yes	No
M256	Piperitone oxide	High	Yes	Yes	Yes	Yes	No
M257	Pulegone	High	Yes	No	Yes	Yes	No
M259	Sabinol	High	Yes	No	Yes	Yes	No
M260	Safranal	High	Yes	No	Yes	Yes	No
M261	Thujyl alcohol	High	Yes	No	Yes	Yes	No
M262	trans-Linalool oxide	High	Yes	Yes	Yes	Yes	No
M263	trans-Sabinene hydrate	High	Yes	No	Yes	Yes	No
M265	α-Pinene oxide	High	Yes	No	Yes	Yes	No
M268	α-Terpinene-4-ol	High	Yes	No	Yes	Yes	No
M269	α-Terpineol	High	Yes	No	Yes	Yes	No
M272	β-Cyclocitral	High	Yes	No	Yes	Yes	No
M273	β-Fenchol	High	Yes	No	Yes	Yes	No
M276	β-Terpineol	High	Yes	No	Yes	Yes	No
M278	m-Mentha-1.8(9)-dien-5-01-ol	High	Yes	No	Yes	Yes	No
M282	Eucalyptol	High	Yes	No	Yes	Yes	No
M283	4-Allylanisole	High	Yes	No	Yes	Yes	No
M284	(-)-Verbenone	High	Yes	No	Yes	Yes	No
M285	α-Pulegone	High	Yes	No	Yes	Yes	No
M286	(-)-Isoulegone	High	Yes	No	Yes	Yes	No
M287	Linalylacetate	High	Yes	Yes	Yes	Yes	No
M288	Cis-Sabinene hydrate	High	Yes	Yes	Yes	Yes	No
M289	DL-Citronellyl acetate	High	Yes	Yes	Yes	Yes	No
M290	Carvacryl acetate	High	Yes	Yes	Yes	Yes	No
M294	Caryophyllene oxide	High	Yes	Yes	Yes	Yes	No
M297	Famesol	High	Yes	Yes	Yes	Yes	No
M298	Famesyl Acetone	High	Yes	No	Yes	Yes	No
M299	Humulene epoxide I	High	Yes	Yes	Yes	Yes	No
M300	Humulene epoxide II	High	Yes	Yes	Yes	Yes	No
M302	Ledol	High	Yes	Yes	Yes	Yes	No
M308	α-Bisabolol	High	Yes	Yes	Yes	Yes	No
M309	α-Caryophyllene alcohol	High	Yes	Yes	Yes	Yes	No
M314	α-Eudesmol	High	Yes	Yes	Yes	Yes	No
M324	β-Eudesmol	High	Yes	Yes	Yes	Yes	No
M331	γ-Eudesmol	High	Yes	Yes	Yes	Yes	No
M334	trans-Nerolidol	High	Yes	Yes	Yes	Yes	No
M335	Guaiol	High	Yes	Yes	Yes	Yes	No
M339	(+)-Cedrol	High	Yes	Yes	Yes	Yes	No
M344	β-Bisabolol	High	Yes	Yes	Yes	Yes	No
M348	N-( p-hydroxy--phenylethyl)- p-hydroxy-trans-cinnamamide	High	Yes	Yes	Yes	Yes	No
M349	Dihydroactinidiolide	High	Yes	Yes	Yes	Yes	No
M350	Dihydrovomifoliol	High	Yes	Yes	Yes	Yes	Yes
M351	Vomifoliol	High	Yes	Yes	Yes	Yes	Yes
M352	β-Ionone	High	Yes	Yes	Yes	Yes	No
M354	Muscarine	High	Yes	Yes	Yes	Yes	No
M356	Trigonelline	High	Yes	No	Yes	Yes	No
M357	L-(+)-isoleucine betaine	High	Yes	No	Yes	Yes	No
M359	Anhydrocannabisativine	High	Yes	Yes	Yes	Yes	Yes
M360	Cannabisativine	High	Yes	Yes	Yes	Yes	Yes

Table 2. Information on 116 potential targets of Cannabis sativa L.

No.	Gene names	Protein names	Uniprot ID
1	CASR	Extracellular calcium-sensing receptor	P41180
2	CNR1	Cannabinoid receptor 1	P21554
3	CTSD	Cathepsin D	P07339
4	GLRA1	Glycine receptor subunit alpha-1	P23415
5	GPR55	G-protein coupled receptor 55	Q9Y2T6
6	GRM5	Metabotropic glutamate receptor 5	P41594
7	TACR3	Neuromedin-K receptor	P29371
8	TSPO	Translocator protein	P30536
9	DUSP3	Dual specificity protein phosphatase 3	P51452
10	GRIN1	Glutamate receptor ionotropic, NMDA 1	Q05586
11	MTOR	Serine/threonine-protein kinase mTOR	P42345
12	GPHN	Gephyrin	Q9NQX3
13	KCNMA1	Calcium-activated potassium channel subunit alpha-1	Q12791
14	SCN2A	Sodium channel protein type 2 subunit alpha	Q99250
15	FOLH1	Glutamate carboxypeptidase 2	Q04609
16	SCN5A	Sodium channel protein type 5 subunit alpha	Q14524
17	GSK3B	Glycogen synthase kinase-3 beta	P49841
18	TP53	Cellular tumor antigen p53	P04637
19	GSR	Glutathione reductase, mitochondrial	P00390
20	SOD2	Superoxide dismutase [Mn], mitochondrial	P04179
21	ADORA1	Adenosine receptor A1	P30542
22	APP	Amyloid-beta precursor protein	P05067
23	CACNA2D2	Voltage-dependent calcium channel subunit alpha-2/delta-2	Q9NY47
24	SCN9A	Sodium channel protein type 9 subunit alpha	Q15858
25	GCK	Hexokinase-4	P35557
26	GABRA1	Gamma-aminobutyric acid receptor subunit alpha-1	P14867
27	GABRA2	Gamma-aminobutyric acid receptor subunit alpha-2	P47869
28	CYP2C9	Cytochrome P450 2C9	P11712
29	CYP3A4	Cytochrome P450 3A4	P08684
30	CYP2C19	Cytochrome P450 2C19	P33261
31	PIK3CA	Phosphatidylinositol 4,5-bisphosphate 3-kinase catalytic subunit alpha isoform	P42336
32	DAB1	Disabled homolog 1	O75553
33	ME2	NAD-dependent malic enzyme, mitochondrial	P23368
34	SLC6A3	Sodium-dependent dopamine transporter	Q01959
35	GABRB3	Gamma-aminobutyric acid receptor subunit beta-3	P28472
36	NTRK2	BDNF/NT-3 growth factors receptor	Q16620
37	MAPK10	Mitogen-activated protein kinase 10	P53779
38	SLC2A1	Solute carrier family 2, facilitated glucose transporter member 1	P11166
39	TRPV1	Transient receptor potential cation channel subfamily V member 1	Q8NER1
40	GRM1	Metabotropic glutamate receptor 1	Q13255
41	AR	Androgen receptor	P10275
42	HTR2A	5-hydroxytryptamine receptor 2A	P28223
43	MRI1	Methylthioribose-1-phosphate isomerase	Q9BV20
44	MT-CO3	Cytochrome c oxidase subunit 3	P00414
45	ALB	Albumin	P02768
46	RORA	Nuclear receptor ROR-alpha	P35398
47	SNX27	Sorting nexin-27	Q96L92
48	VRK2	Serine/threonine-protein kinase VRK2	Q86Y07
49	CACNA1B	Voltage-dependent N-type calcium channel subunit alpha-1B	Q00975
50	MECP2	Methyl-CpG-binding protein 2	P51608
51	PPOX	Protoporphyrinogen oxidase	P50336
52	CASP1	Caspase-1	P29466
53	DHFR	Dihydrofolate reductase	P00374
54	GABRG2	Gamma-aminobutyric acid receptor subunit gamma-2	P18507
55	MTNR1B	Melatonin receptor type 1B	P49286
56	GRM2	Metabotropic glutamate receptor 2	Q14416
57	FGFR1	Fibroblast growth factor receptor 1	P11362
58	CTSA	Lysosomal protective protein	P10619
59	BRD2	Bromodomain-containing protein 2	P25440
60	ABCB1	ATP-dependent translocase ABCB1	P08183
61	GABRA5	Gamma-aminobutyric acid receptor subunit alpha-5	P31644
62	GYS1	Glycogen [starch] synthase, muscle	P13807
63	ITGA4	Integrin alpha-4	P13612
64	GABBR2	Gamma-aminobutyric acid type B receptor subunit 2	O75899
65	CTSB	Cathepsin B	P07858
66	OPRM1	Mu-type opioid receptor	P35372
67	ABCG2	Broad substrate specificity ATP-binding cassette transporter ABCG2	Q9UNQ0
68	IL1B	Interleukin-1 beta	P01584
69	CACNA1C	Voltage-dependent L-type calcium channel subunit alpha-1C	Q13936
70	SHBG	Sex hormone-binding globulin	P04278
71	HTR1A	5-hydroxytryptamine receptor 1A	P08908
72	KCNQ3	Potassium voltage-gated channel subfamily KQT member 3	O43525
73	GBA	Lysosomal acid glucosylceramidase	P04062
74	CHRNA7	Neuronal acetylcholine receptor subunit alpha-7	P36544
75	GRIN2A	Glutamate receptor ionotropic, NMDA 2A	Q12879
76	CHRNA4	Neuronal acetylcholine receptor subunit alpha-4	P43681
77	AKT3	RAC-gamma serine/threonine-protein kinase	Q9Y243
78	ADK	Adenosine kinase	P55263
79	IRAK1	Interleukin-1 receptor-associated kinase 1	P51617
80	HTR1D	5-hydroxytryptamine receptor 1D	P28221
81	TK2	Thymidine kinase 2, mitochondrial	O00142
82	SMARCA2	Probable global transcription activator SNF2L2	P51531
83	GRM4	Metabotropic glutamate receptor 4	Q14833
84	CASK	Peripheral plasma membrane protein CASK	O14936
85	CSNK1G1	Casein kinase I isoform gamma-1	Q9HCP0
86	SCN10A	Sodium channel protein type 10 subunit alpha	Q9Y5Y9
87	GABRA6	Gamma-aminobutyric acid receptor subunit alpha-6	Q16445
88	ADRA2B	Alpha-2B adrenergic receptor	P18089
89	ICK	Serine/threonine-protein kinase ICK	Q9UPZ9
90	FGFR3	Fibroblast growth factor receptor 3	P22607
91	RORB	Nuclear receptor ROR-beta	Q92753
92	FOS	Proto-oncogene c-Fos	P01100
93	ASAH1	Acid ceramidase	Q13510
94	MAPT	Microtubule-associated protein tau	P10636
95	CA5A	Carbonic anhydrase 5A, mitochondrial	P35218
96	GABBR1	Gamma-aminobutyric acid type B receptor subunit 1	Q9UBS5
97	KCNJ11	ATP-sensitive inward rectifier potassium channel 11	Q14654
98	PLA2G6	85/88 kDa calcium-independent phospholipase A2	O60733
99	GLUL	Glutamine synthetase	P15104
100	ACP1	Low molecular weight phosphotyrosine protein phosphatase	P24666
101	ALDH5A1	Succinate-semialdehyde dehydrogenase, mitochondrial	P51649
102	ABAT	4-aminobutyrate aminotransferase, mitochondrial	P80404
103	SNCA	Alpha-synuclein	P37840
104	TNF	Tumor necrosis factor	P01375
105	IL6	Interleukin-6	P05231
106	CTSC	Dipeptidyl peptidase 1	P53634
107	PRNP	Major prion protein	P04156
108	HTR3A	5-hydroxytryptamine receptor 3A	P46098
109	HNMT	Histamine N-methyltransferase	P50135
110	CACNA1G	Voltage-dependent T-type calcium channel subunit alpha-1G	O43497
111	SLC1A1	Excitatory amino acid transporter 3	P43005
112	GRIA4	Glutamate receptor 4	P48058
113	DNM1	Dynamin-1	Q05193
114	TGFB1	Transforming growth factor beta-1 proprotein	P01137
115	CSNK1E	Casein kinase I isoform epsilon	P49674
116	CTSF	Cathepsin F	Q9UBX1

C–T Network construction and analysis

To understand the interaction between compounds and targets, the C–T network (Figure 3(A)) was constructed by mapping 226 candidate ingredients and their related 116 targets, which consisted of 342 nodes and 1668 edges. The possible interactions between natural products and target proteins were assessed by degree, an important topological parameter. The results showed that cannabinoids displayed more intimate association to most targets, when docking with Cannabinoid receptor 1 (CNR1), Albumin (ALB) and Glutamate receptor ionotropic, NMDA 2 A (GRIN2A), etc. To further identify the active ingredients of cannabis for the treatment of EP, the C-T network (Figure 3(B)) of cannabinoids was also constructed, which embodied 199 nodes (106 cannabinoid-compounds and their related 93 targets) and 996 edges. A network analysis showed that cannabicitran (M117, degree = 20), cannabichromanone D (M85, degree = 20), (±)-6,7-trans-epoxycannabigerolic (M100, degree = 19) and cannabinol methyl ether (M54, degree = 18) had the highest number of connections to different, which might play a key role in the treatment of EP. What is more, it could be observed that CNR1 (degree = 113), Androgen receptor (AR, degree = 97), Glycogen synthase kinase-3 beta (GSK3B, degree = 59), ALB (degree = 53) and Mitogen-activated protein kinase 10 (MAPK10, degree = 51), which corresponded to multiple compounds, might be the key targets of the network.

Figure 3. C-T network. (A) The compound – target network of all compounds. (B) The compound – target network of cannabinoids.

GO and KEGG enrichment analysis

The Metascape database was used to analyze the 116 potential targets of cannabis for EP by using GO and KEGG enrichment analysis. The threshold was set at p < 0.05, and the previous GO annotation results and KEGG pathway results were filtered out, as shown in Figure 4. GO enrichment results showed that the main biological processes involved in the active ingredients of cannabis were chemical synaptic transmission, anterograde trans-synaptic signalling and so on. The main cellular components of cannabis were synaptic membrane, post-synapse, postsynaptic membrane and so on. The major molecular functions were neurotransmitter receptor activity, postsynaptic neurotransmitter receptor activity and so on. KEGG pathway analysis displayed that cannabis could play an overall regulatory role through multiple pathways, including neuroactive ligand-receptor interaction, nicotine addiction and so on.

Figure 4. Enrichment analysis of potential targets from main active ingredients of cannabis. (A) Gene Ontology terms for molecular function. (B) Gene Ontology terms for biological process. (C) Gene Ontology terms for cellular component. (D) KEGG pathways.

Pathway construction and analysis

Based on the prediction results of KEGG enrichment, the pathways and targets were shown in Supplemental Table S1. The results showed an average degree of 9.71 per target and 5.57 per pathway, and several target proteins (82 out of 98) mapped to multiple pathways. The T–P network of the top 20 pathways with 73 targets mapped to 20 pathways was constructed as shown in Figure 5. The results displayed that the candidate ingredients were intensively associated with the pathways as follows: neuroactive ligand-receptor interaction, MAPK signalling pathway, taste transduction, GABAergic synapse, retrograde endocannabinoid signalling and cAMP signalling pathway. An integrated ‘EP pathway’ was constructed using 15 EP-related signalling pathways extracted from KEGG pathway as shown in Figure 6(A). The targets of the integrated ‘EP pathway’ exhibited a close functional relationship with those related to EP. The calcium signalling pathway, MAPK signalling pathway and synaptic connections were the main antiepileptic pathways in cannabis, which were displayed in Figure 6(B)–(D), respectively.

Figure 5. The T–P network.

Figure 6. ‘EP’ pathway. (A) Distribution of targets of cannabis on compressed ‘EP pathway’. (B) Calcium signalling pathway. (C) MAPK signalling pathway. (D) Synaptic connections.

Molecular docking analysis

In order to further explore the active mechanism, the interactions between the potential active compounds (M10, M11, M30, M34, M46, M54, M85, M100, M116 and M117) and the targets proteins were elucidated using Schrodinger, the docking scores of targets with the active compounds are listed in Table 3. The interaction between the active site residues and the target protein is shown in Figure 7. The results indicated that M10 and M46 produced hydrogen bonds with GSK3B (LYS-85, ASP-200 and GLC-185) and MAPK10 (H₂O), respectively (Figure 7(A) and (B)). In addition, M54 showed a PI–PI stacking interaction with CNR1 (PHE-379 residue) and AR (TYR–210 and TRP–21), respectively (Figure 7(C) and (D)). According to the docking result of ALB, there was not only conventional PI-PI stacking interaction (TYR-138) but also a Pi–anion interaction with ABG-117 residue (Figure 7(E)).

Figure 7. Schematic 3 D representation that the molecular docking model. (A) Binding mode of M10 to GSK3B. (B) Binding mode of M46 to MAPK10. (C) Binding mode of M54 to CNR1. (D) Binding mode of M54 to AR. (E) Binding mode of M34 to ALB.

Table 3. Molecule docking scores for potential active ingredients with targets.

Compound	CNR1	AR	GSK3B	MAPK10	ALB
M10	−7.749	−6.566	−6.439	−5.902	−5.503
M11	−8.039	−6.798	−5.876	−5.862	−5.095
M30	−8.52	−6.621	−5.458	−5.573	N/A
M34	−7.904	−7.629	−5.547	−5.235	−6.429
M46	−8.55	−6.325	−5.623	−6.307	N/A
M54	−8.761	−8.706	−5.187	−4.242	−6.306
M85	−8.446	−3.851	−4.633	−5.167	N/A
M100	−8.336	−7.362	−6.183	−4.964	−6.057
M116	−8.571	−8.262	−4.966	−3.996	−5.745
M117	−8.574	−8.015	−4.182	−4.106	N/A

Discussion

Network pharmacology has caught more and more attention and becomes an effective tool in identifying alternative targets for traditional medicines and developing new drugs. In the present study, the action mechanism of cannabis against EP was analyzed at the overall level based on the network pharmacology. As a result, a total of 116 potential targets and 226 active compounds in cannabis were obtained, suggesting a potential comprehensive treatment strategy based on TCM featured by multiple compounds, targets and pathways by applying a variety of methods, including C-T and T-P network construction, GO and KEGG pathway enrichment analysis and molecule docking.

Firstly, we predicted the active compounds in cannabis which might be the leading compounds for further drug development, indicating that cannabinoids have better performance corresponding to the potential targets, such as CNR1, GSK3B and MAPK10. As shown in Figure 3(B), a total of 106 active cannabinoids and 93 potential targets of cannabinoids against EP were obtained. It is interesting to note that in addition to CBD and tetrahydrocannabinol (THC) [16], other cannabinoids might have the antiepileptic effect that has not been confirmed, such as M117, M85, M100 and M54. In these cannabinoids, the oxygen of the hydroxyl and keto groups prefer to interact with the targets by forming hydrogen bonds with the active site residues, while the benzene rings and hexatomic rings prefer to form PI–PI stacking interaction or Pi–anion interaction with the active site residues, which might have better performance in the treatment of EP. CNR1, AR, GSK3B, ALB and MAPK10 acted on most cannabinoids, which might be the key targets for the treatment of EP. Already accumulated evidence indicated that CNR1 is abundant in the hippocampus of the central nervous system, and activation of CNR1 inhibits nerve transmission, reduces nerve excitability and regulates the intrinsic excitability of neurons, and most cannabinoids act on CNR1 [17]. GSK3B plays a key role in maintaining intracellular homeostasis. When GSK3B is activated, the expression of Potassium/sodium hyperpolarization-activated cyclic nucleotide-gated channel 4 (HCN4) increases and the phosphorylation of AMPA receptor subunit GluA1 at Serine 831 is inhibited, which protects the central nervous system and achieves antiepileptic effect [18]. Therefore, active cannabinoids might play a key role in the process of EP treatment.

Secondly, GO, the main bioinformatics approach, consists of biological process, cellular component and molecular function. KEGG, a database about pathways, was accessed to obtain not only a gene set but also to define the complex relationship between genes. Based on the results of GO and KEGG pathway enrichment analysis, the targets were associated with various biological processes and pathways, including calcium signalling pathway, MAPK signalling pathway, Synaptic connections, nicotine addiction and morphine addiction, which were interacting.

Calcium signalling pathway

The calcium signalling pathway (Figure 6(B)) plays an important role in epileptic seizures, and Voltage-dependent N-type calcium channel subunit alpha-1B (CACNA1B), Voltage-dependent L-type calcium channel subunit alpha-1C (CACNA1C), Voltage-dependent T-type calcium channel subunit alpha-1G (CACNA1G) and Neuronal acetylcholine receptor subunit alpha-7 (CHRNA7) in the pathway are all related to EP. These targets change the concentration of Ca²⁺ in and out of cells, which in turn affects the cell activities such as transmitter release and cell excitation, and prevents the occurrence and seizure of EP. The mutation of CACNA1B affected the calcium ion transmission and leads to the damage of the synaptic nerve, thus causing EP-related diseases [19]. Clinical studies had shown that the mutation of CACNA1C encoding L-type calcium channel was associated with Timothy and Brugada syndrome [20]. At the same time, CACNA1G mutation was closely related to the onset of cerebellar atrophy in childhood [21]. These provide evidence for the active compounds in cannabis to inhibit EP.

MAPK signalling pathway

The MAPK signalling pathway (Figure 6(C)), which is involved in signal transduction after activation of various growth factors, cytokines, mitogens and hormone receptors [22, 23], is closely related to the pathogenesis of EP. Growth factor receptor-bound protein 2 (GRB2), which binds Fibroblast growth factor receptor 1 (FGFR1) and Fibroblast growth factor receptor 3 (FGFR3) to Son of sevenless (SOS), Transduced rat sarcoma (Ras), which in turn activate Mitogen-activated protein kinase (Mek) and Mitogen-activated protein kinase kinase kinase 1 (MEKK1) and phosphorylate MAPK10. P38 MAPK is activated and its phosphorylation level is increased in the model of epileptic seizure in rats. Inhibition of P38 MAPK reduces neuronal damage and thus reduces the epileptic seizures [24]. This underscores the importance of the MAPK signalling pathway in preventing seizures.

Synaptic connections

During seizures, the synaptic connections (Figure 6(D)) in the brain are abnormal, which in turn increases brain excitability. Cannabis might be able to treat EP through GABAergic Synapse, glutamatergic synapse, dopaminergic synapse, cholinergic synapse, and so on. Studies have shown that GABA_A receptor mutation causes GABAergic synapses to be impaired, the function of GABA_A receptor to be reduced and the binding force of GABA to be reduced; and GABRA5, the subunit of GABA_A receptor, has been proposed as the pathogenic gene of epileptic encephalopathy with early onset [25, 26]. When the content of GABA is increased, excitatory neurotransmitter is inhibited, to control seizures [27]. Glutamatergic Synapse is directly related to an abnormal discharge of neurons and has always been thought to be closely related to the occurrence of EP. Hyperactivity of GLS and other targets in this pathway leads to an increase in glutamate content, which acts on ion channels and increases excitability [28–30]. Similarly, dopaminergic synapse and cholinergic synapse also facilitate epileptic seizures [31–33].

Other signalling pathways

The study also predicted that nicotine addiction and morphine addiction, which was also effective in EP, reduced the duration and frequency of epileptic seizures by stimulating neurocytokines and suppressing neurotransmission [34, 35]. Retrograde endocannabinoid could promote seizures by inhibiting the release of GABA at inhibitory synapses [36].

In summary, these pathways play a significant role in the treatment of EP and also provide a piece of powerful evidence for multi-target treatment. Besides, molecular docking was used as the target and molecular interaction widely. The good molecular docking results showed that M54 (cannabinol methyl ether) had a high binding score with its corresponding targets, which could possibly become the lead compound for further research.

Conclusions

To clarify the mechanism of action of cannabis in the EP treatment systematically, the network pharmacology and molecular docking were employed to explore the active compounds, the targets of cannabis against EP, and the related signalling pathways. A total of 226 potentially active compounds in cannabis and 1117 of their targets were predicted by multiple databases. At the same time, 1403 EP-related targets were screened by multiple databases. A total of 116 common targets were obtained by statistical intersection. The GO and KEGG enrichment analysis indicated that there were multiple interactions. The results showed that compounds in cannabis, especially cannabinoids, interact with the targets (CNR1, ALB, GSK3BA, AR and MAPK10) to ameliorate EP, and the effect is due to the joint action of the calcium signalling pathway, MAPK signalling pathway, synaptic connections and so on. Moreover, molecular docking analysis confirmed the antiepileptic effect of cannabis. However, in vitro and in vivo experiments should be carried out based on this study. In sum, the mechanism of cannabis in the EP treatment was analyzed systematically based on network pharmacology. This study provides a practicable strategy for the development of new cannabis drugs for EP.

Abbreviations
EP	=	epilepsy
Cannabis	=	Cannabis sativa L.
CBD	=	cannabidiol
C-T	=	compound-target
T-P	=	target-pathway
TCM	=	traditional Chinese medicine
ADME	=	absorption, distribution, metabolism and excretion
OMIM	=	online Mendelian inheritance in man
TTD	=	therapeutic target database
GO	=	gene ontology
KEGG	=	Kyoto encyclopedia of genes and genomes
TCMSP	=	traditional Chinese medicines for systems pharmacology database and analysis platform
GI	=	gastrointestinal
TCM	=	traditional Chinese medicine
THC	=	tetrahydrocannabinol
HCN4	=	potassium/sodium hyperpolarization-activated cyclic nucleotide-gated channel 4
CACNA1B	=	voltage-dependent N-type calcium channel subunit alpha-1B
CACNA1C	=	voltage-dependent L-type calcium channel subunit alpha-1C
CACNA1G	=	voltage-dependent T-type calcium channel subunit alpha-1G
CHRNA7	=	neuronal acetylcholine receptor subunit alpha-7
GRB2	=	growth factor receptor-bound protein 2
FGFR1	=	fibroblast growth factor receptor 1
FGFR3	=	fibroblast growth factor receptor 3
SOS	=	son of sevenless
Mek	=	mitogen-activated protein kinase
MEKK1	=	mitogen-activated protein kinase kinase kinase 1.

Data availability

The data supporting the findings reported in this study are available from the corresponding author upon reasonable request.

Disclosure statement

W.X. is affiliated with Kanion Pharmaceutical Co. Ltd. There is no conflict of interest to declare regarding the content of interest of this article.

Additional information

Funding

This work was supported by High-level talents project of Dalian City (2016RQ064); Research Foundation of Education Bureau of Liaoning Province (2016J008); and National Natural Science Foundation (U1603285) of China.