Wei-Long Jiang,Yu-Feng Zhang,Meng-Ying Liu,Su-Yan Zhang,Jun-Xian Xu,Min Li,Hong Xue,Li-Jun Tian＊,Xu-Dong Han＊

1Department of Respiratory Medicine, Jiangyin Hospital of Traditional Chinese Medicine, Jiangyin Hospital Affiliated to Nanjing University of Chinese Medicine, Jiangyin 214400, China. 2Department of Pulmonary and Critical Care Medicine, Nanjing Drum Tower Hospital Clinical College of Nanjing Medical University, Nanjing, 210008, China. 3Department of Critical Care Medicine,Nantong Third People's Hospital, Nantong University, Nantong 226001, China. 4Department of Traditional Chinese Medicine,Nantong Third People's Hospital, Nantong University, Nantong, Jiangsu 226001, China. 5Department of Infectious Disease,Nantong Third People's Hospital,Nantong University,Nantong,Jiangsu 226001,China

Abstract Objective: To explore the mechanism of Kangguan decoction in the treatment of coronavirus disease 2019(COVID-19) and then perform preliminary verification. Methods: The effective compounds and target genes of Kangguan decoction were obtained from Traditional Chinese Medicine Systems Pharmacology Database and Analysis Platform. COVID-19 related target genes were searched in the GeneCards database. The active target genes of Kangguan decoction acting on COVID-19 were identified to perform GO function enrichment and KEGG pathway enrichment analysis. The compound-target network and protein-protein interaction were constructed;Molecular docking simulations of macromolecular protein target receptors and their corresponding compounds were performed. The clinical data of COVID-19 patients were retrieved from their electronic medical records of Nantong Third People's Hospital.Results:We screened out 137 effective compounds and 274 effective target genes of Kangguan decoction from TCMSP. The active target genes of Kangguan decoction were compared with the COVID-19 related target genes, and 63 active target genes for Kangguan decoction acting on COVID-19 were identified.GO function enrichment and KEGG pathway enrichment analysis were performed.The compound-target network and PPI network were constructed and the key compounds and key targets were selected to construct a key compound-target network. Finally, the binding of the target and its corresponding components was verified by molecular docking and two clinical cases with obvious clinical efficacy after Kangguan decoction application were demonstrated. Conclusion: The pharmacological mechanism of Kangguan decoction acting on COVID-19 has been explored, and the active compounds and targets of Kangguan decoction acting on COVID-19 and clinical efficacy for Kangguan decoction treating COVID-19 patients have been preliminarily verified.

Key words: Kangguan decoction, Coronavirus disease 2019, Network pharmacology, Molecular docking, Clinical efficacy,Quercetin,PTGS2,Interleukin-6.

Background

In December 2019,COVID-19,caused by severe acute respiratory SARS-CoV-2, first broke out in Wuhan,China [1, 2].Soon the global epidemic broke out and no definitive conclusion has been drawn about its origin. Despite efforts have been made to control the global epidemic, some areas are still not optimistic.Update to November 11,2020, there were 50,676,072 confirmed cases,1,261,075 confirmed deaths,and 219 countries, areas or territories with cases, according to the WHO (https://www.who.int/emergencies/diseases/novel-coronavirus-2019).

There are no special drugs for COVID-19 treatment.How to form a rapid and effective COVID-19 treatment scheme is a major problem. The specific mechanism and efficacy of traditional Chinese medicine(TCM)in the treatment of COVID-19 are not very clear and clinical evidence is needed to evaluate the efficacy of TCM in the future[3,4].TCM has been playing an important role in the prevention and treatment of major outbreaks. Chinese herbal formula could be an alternative approach for prevention of COVID-19 in high-risk population and TCM could be considered as an adjunctive therapeutic option in the management of COVID-19[5−7].

TCM plays a key role in the treatment of viral diseases,especially relieving symptoms [8, 9].Yinqiao Power and Sangju Yin are two TCM prescriptions,long-term used for respiratory infections, pneumonia and other diseases. Studies have shown their role in anti-virus,anti-inflammatory and antipyretic[10−12].

Nantong Third People's Hospital is a specialized infectious diseases hospital, which is designated to admit and manage most COVID-19 patients in Nantong city, China [13]. Based on the plan of the State Administration of TCM and traditional experience, Kangguan decoction (KGD) were used to treat of COVID-19 patients in our hospital. KGD was came up based on Yinqiao Power and Sangju Yin,mainly composed of Bohe (Herba Menthae Haplocalycis), Dandouchi (Semen Sojae Preparatum),Gualou (Fructus Trichosanthis), Houpo (Cortex Magnoliae Officinalis),Huangqin(Radix Scutellariae),Jiegeng (Radix Platycodi), Jinyinhua (Flos Lonicerae),Xingren (Semen Armeniacae), Lianqiao (Fructus Forsythiae), Niubangzi (Fructus Arctii), and Sangye(Folium Mori). From January 23 to now, 34 COVID-19 patients have been diagnosed and all successfully treated by the combined therapy of KGD and Western medicine. Through the application of KGD, we observed certain curative effect in the treatment of KGD on COVID-19 patients.In this study,we used network pharmacology approach to explore the pharmacological mechanism of KGD on COVID-19 and performed preliminary verification using molecular docking and clinical cases. The flowchart of our study is depicted in Figure 1.

Materials and methods

Screening of effective compounds in KGD

The compounds in the herbs of KGD were obtained from the Traditional Chinese Medicine Systems Pharmacology Database and Analysis Platform(TCMSP) (http://tcmspw.com/tcmsp.php) [14].TCMSP is a unique systems pharmacology platform of Chinese herbal medicine that captures the relationships between drugs, targets and diseases. Oral bioavailability (OB) and drug-likeness (DL) were used to screen for effective compounds. OB is an important index to measure the effectiveness of drugs entering the human circulatory system and good OB is one of the basic prerequisites for the pharmacological activity of compounds. DL is the physical and chemical properties and biological characteristics associated with good clinical efficacy. OB ≥30% and DL ≥0.18 are frequently valued in network pharmacology[15].

Screening of the target genes of effective compounds

The target genes of effective compounds were also acquired from TCMSP.The target genes were imported into the UniProt knowledge base, a comprehensive resource for protein sequence and annotation data(http://www.uniprot.org/) [16], and selecting search format as "homo sapiens", the human official gene symbols were charactered, which were considered as the effective target genes of KGD.

Acquisition of COVID-19 related target genes and identification of active target genes of KGD acting on COVID-19

Using "novel coronavirus pneumonia" and"coronavirus disease" as the key words in the GeneCards database(https://www.genecards.org/)[17],which is a searchable, integrative database providing comprehensive, user-friendly information on all annotated and predicted human genes, COVID-19 related target genes were searched and acquired.Then,the effective target genes of KGD were compared with the COVID-19 related target genes, and the intersecting genes were characterized as the active target genes of KGD acting on COVID-19.

GO function enrichment and KEGG pathway enrichment analysis

RGUI4.0.3 and org.Hs.eg.db package were used to gain the entrezIDs of candidate target genes. Then,RGUI, the clusterProfiler package were applied to operate GO function enrichment analysis, which included biological process (BP), molecular function(MF), cellular component (CC) analysis, and KEGG pathway enrichment analysis[18].

Construction of compound-target network and selection of key compounds

Cytoscape3.6.0 software was used to construct a compound-target network and its NetworkAnalyzer tool function was used to analyze the network [19].Nodes performed compounds and target genes, and edges represented the relationships between them. On the basis of the degree of connection between the compound and the target gene, the key compounds in KGD acting on COVID-19 were elected.

Construction of PPI network and selection of key targets

A PPI network of KGD acting on COVID-19 was constructed after introducing the active target genes into the STRING database, supporting functional discovery in genome-wide experimental datasets(https://string-db.org/) [20]. Selecting the research species as "homo sapiens" and the lowest interaction score as 0.4,a PPI network was obtained.Then the PPI network was made topology analysis and the key targets of KGD acting on COVID-19 were selected according to the degree values of each target gene using Cytoscape3.6.0 software and its NetworkAnalyzer tool.

Verification of molecular docking

The binding of the target and its corresponding component was verified by molecular docking.The 2D and 3D structures of the small-molecule compounds were obtained from the PubChem Database(https://pubchem.ncbi.nlm.nih.gov/) and the macromolecular protein target receptors were obtained from the RCSB PDB database (http://www.rcsb.org/).Molecular docking simulations of macromolecular protein target receptors and their corresponding compounds were performed using AutoDockTool 1.5.6 and AutoDock Vina software [21, 22], and further demonstration was using the PyMOL Molecular Graphics System(Version 2.4.0).

Participants of Clinical cases

The clinical data of COVID-19 patients were retrieved from their electronic medical records, including their symptoms, clinical parameters, and treatment and outcomes. The study was approved by Nantong Third Hospital Ethics Committee (E2020003). Written informed consents were obtained from each of the involved patients.

Statistical analysis

Using the bioinformatic tools of the platforms and software mentioned above, some statistical analysis was performed automatically. In GO function enrichment and KEGG pathway enrichment, adjP<0.05 was considered as statistically significant.

Results

Screened effective compounds in KGD

We obtained 164 compounds inHerba Menthae Haplocalycis, 15 compounds inSemen Sojae Preparatum, 80 compounds inFructus Trichosanthis,139 compounds inCortex Magnoliae Officinalis, 143 compounds inRadix Scutellariae, 102 compounds inRadix Platycodi, 236 compounds inFlos Lonicerae,113 compounds inSemen Armeniacae,150 compounds inFructus Forsythiae, 144 compounds inFructus Arctii, and 269 compounds inFolium Morifrom TCMSP (Supplementary File 1). Setting filter criteria as OB ≥30% and DL ≥0.18, 10 effective compounds inHerba Menthae Haplocalycis, 2 effective compounds inSemen Sojae Preparatum, 11 effective compounds inFructus Trichosanthis, 2 effective compounds inCortex Magnoliae Officinalis, 36 effective compounds inRadix Scutellariae, 7 effective compounds inRadix Platycodi, 23 effective compounds inFlos Lonicerae,19 effective compounds inSemen Armeniacae, 23 effective compounds inFructus Forsythiae, 8 effective compounds inFructus Arctii, and 29 effective compounds inFolium Moriwere selected. Finally, 137 effective compounds in KGD were identified with exclusion of duplications.The basic information on the effective compounds in KGD is shown in Table 1.

Herb Effective compound Compound ID OB(%) DL HQ 5,8,2'-Trihydroxy-7-methoxyflavone MOL002908 37.01 0.27 HQ 5,7,2,5-tetrahydroxy-8,6-dimethoxyflavone MOL002909 33.82 0.45 HQ Carthamidin MOL002910 41.15 0.24 HQ 2,6,2',4'-tetrahydroxy-6'-methoxychaleone MOL002911 69.04 0.22 HQ Dihydrobaicalin_qt MOL002913 40.04 0.21 HQ JYH Eriodyctiol(flavanone) MOL002914 41.35 0.24 HQ Salvigenin MOL002915 49.07 0.33 HQ 5,2',6'-Trihydroxy-7,8-dimethoxyflavone MOL002917 45.05 0.33 HQ 5,7,2',6'-Tetrahydroxyflavone MOL002925 37.01 0.24 HQ dihydrooroxylin A MOL002926 38.72 0.23 HQ Skullcapflavone II MOL002927 69.51 0.44 HQ oroxylin a MOL002928 41.37 0.23 HQ Panicolin MOL002932 76.26 0.29 HQ 5,7,4'-Trihydroxy-8-methoxyflavone MOL002933 36.56 0.27 HQ NEOBAICALEIN MOL002934 104.34 0.44 HQ DIHYDROOROXYLIN MOL002937 66.06 0.23 HQ JYH LQ NBZ SY beta-sitosterol MOL000358 36.91 0.75 HQ Norwogonin MOL000525 39.40 0.21 HQ 5,2'-Dihydroxy-6,7,8-trimethoxyflavone MOL000552 31.71 0.35 HQ ent-Epicatechin MOL000073 48.96 0.24 HQ JYH XR SY Stigmasterol MOL000449 43.83 0.76 HQ coptisine MOL001458 30.67 0.86 HQ bis[(2S)-2-ethylhexyl]benzene-1,2-dicarboxylate MOL001490 43.59 0.35 HQ NBZ SY Supraene MOL001506 33.55 0.42 HQ Diop MOL002879 43.59 0.39 HQ epiberberine MOL002897 43.09 0.78

Herb Effective compound Compound ID OB(%) DL HQ Moslosooflavone MOL008206 44.09 0.25 HQ 11,13-Eicosadienoic acid,methyl ester MOL010415 39.28 0.23 HQ 5,7,4'-trihydroxy-6-methoxyflavanone MOL012245 36.63 0.27 HQ 5,7,4'-trihydroxy-8-methoxyflavanone MOL012246 74.24 0.26 HQ rivularin MOL012266 37.94 0.37 JG cis-Dihydroquercetin MOL004580 66.44 0.27 JG 2-O-methyl-3―O-β-D-glucopyranosyl platycogenate A MOL005996 45.15 0.25 JG dimethyl 2-O-methyl-3-O-a-D-glucopyranosyl platycogenate A MOL006026 39.21 0.25 JG robinin MOL006070 39.84 0.71 JYH Ethyl linolenate MOL001495 46.10 0.20 JYH phytofluene MOL002707 43.18 0.50 JYH(-)-(3R,8S,9R,9aS,10aS)-9-ethenyl-8-(beta-Dglucopyranosyloxy)-2,3,9,9a,10,10ahexahydro-5-oxo-5H,8Hpyrano[4,3-d]oxazolo[3,2-a]pyridine-3-carboxylic acid_qt MOL003006 87.47 0.23 JYH secologanic dibutylacetal_qt MOL003014 53.65 0.29 JYH NBZ SY beta-carotene MOL002773 37.18 0.58 JYH ZINC03978781 MOL003036 43.83 0.76 JYH Chryseriol MOL003044 35.85 0.27 JYH kryptoxanthin MOL003059 47.25 0.57 JYH 4,5'-Retro-.beta.,.beta.-Carotene-3,3'-dione,4',5'-didehydro-MOL003062 31.22 0.55 JYH 5-hydroxy-7-methoxy-2-(3,4,5-trimethoxyphenyl)chromone MOL003095 51.96 0.41 JYH 7-epi-Vogeloside MOL003101 46.13 0.58 JYH Caeruloside C MOL003108 55.64 0.73 JYH Centauroside_qt MOL003111 55.79 0.50 JYH Ioniceracetalides B_qt MOL003117 61.19 0.19 JYH XYLOSTOSIDINE MOL003124 43.17 0.64

Herb Effective compound Compound ID OB(%) DL JYH dinethylsecologanoside MOL003128 48.46 0.48 JYH LQ NBZ SY kaempferol MOL000422 41.88 0.24 JYH LQ SY quercetin MOL000098 46.43 0.28 XR estrone MOL010921 53.56 0.32 XR Diisooctyl succinate MOL010922 31.62 0.23 XR 11,14-eicosadienoic acid MOL002211 39.99 0.20 XR(6Z,10E,14E,18E)-2,6,10,15,19,23-hexamethyltetracosa-2,6,10,14,18,22-hexaene MOL002372 33.55 0.42 XR gondoic acid MOL005030 30.70 0.20 XR CLR MOL000953 37.87 0.68 XR LQ Mairin MOL000211 55.38 0.78 XR (+)-catechin MOL000492 54.83 0.24 XR Glycyrol MOL002311 90.78 0.67 XR Ziziphin_qt MOL003410 66.95 0.62 XR Licochalcone B MOL004841 76.76 0.19 XR liquiritin MOL004903 65.69 0.74 XR Glabridin MOL004908 53.25 0.47 XR Phaseol MOL005017 78.77 0.58 XR Machiline MOL007207 79.64 0.24 XR l-SPD MOL012922 87.35 0.54 LQ 20(S)-dammar-24-ene-3β,20-diol-3-acetate MOL003281 40.23 0.82 LQ(2R,3R,4S)-4-(4-hydroxy-3-methoxy-phenyl)-7-methoxy-2,3-dimethyloltetralin-6-ol MOL003283 66.51 0.39 LQ NBZ(3R,4R)-3,4-bis[(3,4-dimethoxyphenyl)methyl]oxolan-2-one MOL003290 52.30 0.48 LQ (+)-pinoresinol monomethyl ether MOL003295 53.08 0.57 LQ PHILLYRIN MOL003305 36.40 0.86 LQ ACon1_001697 MOL003306 85.12 0.57 LQ (+)-pinoresinol monomethyl ether-4-D-beta-glucoside_qt MOL003308 61.20 0.57 LQ 3beta-Acetyl-20,25- MOL003315 33.07 0.79

Abbreviations: BH, Herba Menthae Haplocalycis; HQ, Radix Scutellariae; JG, Radix Platycodi; GL, Fructus Trichosanthis; XR, Semen Armeniacae; JYH, Flos Lonicerae; LQ, Fructus Forsythiae; DDC, Semen Sojae Preparatum; SY, Folium Mori; HP, Cortex Magnoliae Officinalis; NBZ, Fructus Arctii; OB, oral bioavailability;DL,drug-likeness.

Screened effective target genes of KGD

We also obtained the corresponding target genes of the 137 effective compounds from TCMSP, in which 28 compoundsdid not have targets (Supplementary File 2A).

And then, the corresponding gene symbols were screened by setting the format as "homo sapiens" from the UniProt knowledge base, and 3 compounds(MOL001790, MOL007180, and MOL003305)had no gene symbol matching their corresponding target genes(Supplementary File 2B). Finally, 274 effective target genes of 106 effective compounds in KGD were identified(Supplementary File 2C).

Acquired COVID-19 related target genes and identified active target genes of KGD acting on COVID-19

We used "novel coronavirus pneumonia" and"coronavirus disease"as the key words to search in the GeneCards database and 795 COVID-19 related target genes were acquired(Supplementary File 3).

The 274 effective target genes of KGD were compared with the 795 COVID-19 related target genes,and 63 active target genes of KGD acting on COVID-19 were identified(Figure 2,Table 2).

Figure 2 Active target genes of Kangguan decoction acting on COVID-19. The 274 active target genes of Kangguan decoction were compared with the 795 COVID-19 related target genes, and 63 active target genes of Kangguan decoction acting on COVID-19 were identified.

Table 2 Gene symbol and entrezID of active target genes

GO function enrichment and KEGG pathway enrichment analysis

The entrezIDs of active target genes of KGD acting on COVID-19 were obtained using RGUI and org.Hs.eg.db (Table 2). Next, GO function enrichment and KEGG pathway enrichment analysis were operated using RGUI and clusterProfiler.

Active target genes of KGD acting on COVID-19 were significantly enriched in response to toxic substance, response to oxidative stress, response to metal ion, response to molecule of bacterial origin,response to lipopolysaccharide, cellular response to biotic stimulus, response to reactive oxygen species,cellular response to lipopolysaccharide, cellular response to molecule of bacterial origin, regulation of apoptotic signaling pathway,and so on(Supplementary File 4A).The top 20 GO BP enrichments ranked by adjPvalue are shown in Figure 3A.

Active target genes of KGD acting on COVID-19 were significantly enriched in cytokine receptor binding, protein serine/threonine kinase activity,cytokine activity, phosphatase binding, BH domain binding, cofactor binding, heme binding, tetrapyrrole binding, protein phosphatase binding, ubiquitin-like protein ligase binding, and so on (Supplementary File 4B). The top 20 GO MF enrichments ranked by adjPvalue are shown in Figure 3B.

Figure 3 GO function and KEGG pathway enrichments. A, Enriched biological process functions of active target genes. B, Enriched molecular function activities of active target genes. C, Enriched cellular component regions of active target genes.D,KEGG pathway enrichments.

Active target genes of KGD acting on COVID-19 were significantly enriched in membrane raft,membrane microdomain, membrane region,cyclin-dependent protein kinase holoenzyme complex,serine/threonine protein kinase complex,protein kinase complex,caveola,vesicle lumen,mitochondrial outermembrane, plasma membrane raft, and so on(Supplementary File 4C). The top 20 GO CC enrichments ranked by adjPvalue are shown in Figure 3C.

KEGG pathway enrichment analysis showed that active target genes of KGD acting on COVID-19 were significantly enriched in advanced glycation end products-(AGE-) receptor for the AGE signaling pathway in diabetic complications, interleukin-(IL-)17 signaling pathway, tumor necrosis factor (TNF)signaling pathway,and so on(Supplementary File 4D).The top 20 KEGG pathway enrichments ranked by adjPvalue are shown in Figure 3D.

Constructed compound-target network and selected key compounds

As MOL000359, MOL004355, MOL006756,MOL007165, MOL007171, MOL007172,MOL001490, MOL002879, MOL010415,MOL003036, MOL003128, MOL002211,MOL000953, MOL000211, MOL003315,MOL001771, MOL003851, MOL003856, and MOL003860 had no association with an overlapping active target gene, the 63 overlapping active target genes correlate with 87 active compounds(Table 3).

Table 3 Active compounds correlated with overlapping active target genes

Abbreviations: BH, Herba Menthae Haplocalycis;HQ, Radix Scutellariae; JG, Radix Platycodi ;GL, Fructus Trichosanthis; XR, Semen Armeniacae; JYH, Flos Lonicerae; LQ, Fructus Forsythiae; DDC, Semen Sojae Preparatum;SY,Folium Mori;HP,Cortex Magnoliae Officinalis;NBZ,Fructus Arctii;

Then, we constructed a network of compound-target using Cytoscape software and analyzed by the NetworkAnalyzer tool. There were 150 nodes (87 compound nodes and 63 target gene nodes) and 460 edges in the network(Supplementary File 5,Figure 4).The top 15 compounds listed by degree in the network were quercetin, luteolin, wogonin, kaempferol,arachidonic acid, naringenin, acacetin, baicalein,iristectorigenin A, glycitein,5,7,4'-Trihydroxy-8-methoxyflavone,moslosooflavone,chryseriol, licochalcone B, and tetramethoxyluteolin,which can be acknowledged the key compounds in KGD acting on COVID-19. The basic information of these key compounds is shown in Table 4.

Table 4 Key compounds in Kangguan decoction acting on COVID-19.

Abbreviations:JYH,Flos Lonicerae;LQ,Fructus Forsythiae;SY,Folium Mori;BH,Herba Menthae Haplocalycis;JG, Radix Platycodi; HQ, Radix Scutellariae; NBZ, Fructus Arctii; DDC, Semen Sojae Preparatum; XR, Semen Armeniacae;

Figure 4 Compound-target network. There were 150 nodes (87 compound nodes and 63 target gene nodes) and 460 edges in the network. Rectangles represent active compounds(different colors represent different compounds),Triangle represent the active target genes,and the edges represent links between the nodes.

Constructed PPI network and selected key targets

We mapped 63 overlapping active target genes into the STRING database and obtained the PPI network. In which, 62 target proteins had interactions (APOD had no interaction) and 791 edges represented the interactions between the proteins, when lowest interaction score was set to 0.40(Supplementary File 6,Figure 5).

Figure 5 The protein protein interaction network of Kangguan decoction acting on COVID-19.In the network,62 target proteins had interactions (APOD had no interaction) and 791 edges represented the interactions between the proteins,the lowest interaction score was set to 0.40.

The top 20 target genes ranked by degree in the network are shown in Table 5,which can be considered the key targets of KGD acting on COVID-19.

Table 5 Key targets of Kangguan decoction acting on COVID-19

Constructed key compound-target network

Introducing the key compounds (small-molecule compounds) and key targets (macromolecular protein target receptors) and their relationships into Cytoscape3.6.0 software, a key compound-target network was constructed (Supplementary File 7,Figure 6).

Figure 6 Key compound-target network.Quercetin,luteolin, wogonin,kaempferol, arachidonic acid,naringenin,acacetin,baicalein,iristectorigenin A,glycitein,5,7,4'-Trihydroxy-8-methoxyflavone,Moslosooflavone,Chryseriol,Licochalcone B and Tetramethoxyluteolin were the key compounds considered as small-molecule compounds,while VEGFA, MAPK3, MAPK1,CASP3, IL6,ALB, MAPK8, FOS, PTGS2, CCL2, CASP8, CCND1, MAPK14,IL1B, ICAM1, RELA, BCL2L1, EGFR, IL2 and CAT were the key targets considered as macromolecular protein target receptors.

Molecular docking

We obtained the 3D structures of the small-molecule compounds from the PubChem Database and the macromolecular protein target receptors from the RCSB PDB database. Then, molecular docking simulations of potential targets and their corresponding compounds were performed using AutoDockTool 1.5.6 and AutoDock Vina software. Finally, the binding of the target and its corresponding component was verified by molecular docking and demonstrated by the PyMOL Molecular Graphics System. We selected several representative to demonstrate. In the molecular docking simulations of PTGS2-quercetin, minimum affinity was -9.5 kcal/mol, grid center was 22.594,40.999 and 39.56, and distance from best mode was 0.000 rmsd l.b. and 0.000 rmsd u.b. (Figure 7). In the molecular docking simulations of IL6-luteolin,minimum affinity was affinity-7.3 kcal/mol,grid center was -0.585, 0.365 and 0.253, distance from best mode was 0.000 rmsd l.b.and 0.000 rmsd u.b.(Figure 8).

Figure 7 PTGS2-quercetin molecular docking. A, 3D structures of quercetin. B, 3D structures of PTGS2. C,Molecular docking simulation.D,Molecular docking simulation(display protein surface).

Figure 8 IL6-luteolin molecular docking. A, 3D structures of luteolin. B, 3D structures of IL6. C, Molecular docking simulation.D,Molecular docking simulation(display protein surface).

Clinical cases

Case 1.A 58-year-old male was admitted to our hospital on January 31, 2020 due to cough and fever for 4 days.He had an exposure history to SARS-CoV-2 infected persons on January 19, 2020. When he had a fever with body temperature (38.2°C) on January 27,2020, he received intravenous injection of cefmetazole and peramivir in local community hospital. Fever persisted as a body temperature ranging from 38.2°C to 39.8°C,even though there was a temporary decrease after ibuprofen used. On January 31, her chest computed tomography scan showed bilateral distribution of patchy shadows and diagnosis of COVID-19 was confirmed subsequently by positive results of a throat swab for SARS-CoV-2. Laboratory workup included white blood cell, lymphocyte and inflammatory markers C-reactive protein (CRP),erythrocyte sedimentation rate, procalcitonin,interleukin-6 (IL-6) and ferroprotein was shown in Supplementary File 8A. Aerosol inhalation of interferon α-2b (5 million units twice a day), lopinavir plus ritonavir (500 mg twice a day orally), and arbidol(200 mg three times a day orally) was given for antiviral therapy; moxifloxacin (0.4 g once a day,orally) for secondary infection prevention. On the second day after admission, the patient received KGD(one dose a day,take one dose for twice).His symptom of fever continued to improve quickly and his temperature returned to normal on hospital day 5(Figure 9). Finally, the patient was cured and discharged from our hospital on March 3,2020.

Figure 9 Body temperature of patient in case 1. Body temperature ranging from 38.2°C to 39.8°C, even though there was a temporary decrease after ibuprofen used. On the second day after admission, the patient received Kangguan decoction,body temperature returned to normal on hospital day 3.

Case 2.On January 29, 2020, a 45-year-old man was admitted to our hospital, with a 5-day history of fever,and a 2-day history of cough and dyspnea. He had returned to Nantong, China on January 18 after traveling to Wuhan, China. He had persistent severe fever with body temperature from 37.5°C to 39.2°C for 5 days. He started to receive intravenous injection of cefmetazole in local community hospital and took oseltamivir (75 mg, twice a day) for 3 days. On January 29, her chest computed tomography scan showed typical ground-glass lesions and diagnosis of COVID-19 was confirmed subsequently by positive results of a throat swab for SARS-CoV-2. His inflammatory markers such as CRP (84.4 mg/L), IL-6(218.82 pg/mL) and ferroprotein (1607.87 ng/mL) on admission were significantly higher than other COVID-19 patients in the same period. Aerosol inhalation of interferon α-2b (5 million units twice a day), lopinavir plus ritonavir (500 mg twice a day orally), and arbidol (200 mg three times a day orally)was given for antiviral therapy. On the second day after admission, the patient received KGD (one dose a day, take one dose for twice). With the above combined therapy of KGD and western medicine, his dyspnea was gradually improved, the temperature returned to normal on hospital day 2, and the inflammatory markers decrease gradually (Figure 10,Supplementary File 8B). Finally,the patient was cured and discharged from our hospital on February 12,2020.

Figure 10 Laboratory workup of patient in case 2.Inflammatory markers such as C-reactive protein(84.4 mg/L),IL-6 (218.82 pg/mL) and ferroprotein (1607.87 ng/mL) on admission were significantly higher than other COVID-19 patients were in the same period.With the above combined therapy of Kangguan decoction and western medicine,the inflammatory markers decrease gradually.

Discussion

COVID-19 belongs to the category of epidemic disease of TCM.As the disease is located in the lung,the basic pathogenesis is characterized by dampness, heat,poison and blood stasis, and Chinese medicine has unique advantages in syndrome differentiation and treatment.

Yinqiao Power and Sangju Yin are from the ancient Chinese medicine prescription "Febrile Disease Differentiation". Yinqiao Power has cold-pungent diaphoresis, clearing heat and detoxification effect,often used in the treatment of influenza, acute tonsillitis, measles from the beginning, as well as encephalitis B, meningitis, mumps, pharyngitis,pharynx isthmus herpes and other warm disease from the beginning. Sangju Yin has the effect of clearing away wind and clearing heat, promoting lung and relieving cough. It is often used to treat cold, acute bronchitis,upper respiratory tract infection,pneumonia,acute tonsillitis and so on. Based on the addition and subtraction of Yinqiao Power and Sangju Yin, KGD has been used in our hospital to treat the COVID-19 patients.

At present, network pharmacology is generally applied in the study of TCM. Network pharmacology promotes the research method of network goal and multi-component therapy, which promote the research of TCM from the current model single component and target to a new model of multi-component and network target.TCM network pharmacology can provide a new way for TCM to move from experience-based medicine to evidence-based medicine, and improve the current drug research strategy[23−25].

Compound-target network showed that the active target genes were related to all the 11 herbs in KGD,which indicates that the KGD prescription is reasonable and each single herb plays a role in acting COVID-19 (Figure 3). The key compounds and key targets were taken to construct a key compound-target network. Compound-target network showed that the key target genes were still related to most herbs in KGD (Figure 5). In particular, JYH, LQ, SY and HQ were closely related to key target genes to play a major role. In TCM theory, these herbs are the key herbs,which are known as monarch medicine and minister medicine.

We identified 63 active target genes of KGD acting on COVID-19. GO BP for KGD acting on COVID-19 were enriched in response to toxic substance,oxidative stress, metal ion, molecule of bacterial origin,lipopolysaccharide, reactive oxygen species, cellular response to lipopolysaccharide, biotic stimulus,molecule of bacterial origin and regulation of apoptotic signaling pathway. GO MF analysis indicated that KGD acting on COVID-19 involved in cytokine receptor binding, protein serine/threonine kinase activity, cytokine activity, phosphatase binding, BH domain binding, cofactor binding, heme binding,tetrapyrrole binding, protein phosphatase binding and ubiquitin-like protein ligase binding. GO CC analysis indicated that KGD in the treatment of COVID-19 were enriched in membrane raft, membrane microdomain, membrane region, cyclin-dependent protein kinase holoenzyme complex, serine/threonine protein kinase complex, protein kinase complex,caveola, vesicle lumen, mitochondrial outer membrane and plasma membrane raft. These function enrichments are intimately related to the pathological mechanisms of COVID-19[26−29].

KEGG enrichment pathway analysis showed that many pathways were closely related to the pathogenesis of COVID-19. The main pathways encompassed IL-17 signaling pathway, measles, TNF signaling pathway,hepatitis B,human cytomegalovirus infection, influenza A and pertussis. Most of these pathways are associated with viruses as well as inflammation. These results illustrated that KGD acts on COVID-19 through multiple pathways [30−34].These pathways and relevant target genes deserve priority for further study.

We constructed a compound-target network and identified the key compounds of KGD acting on COVID-19. The fundamental key compounds were quercetin, luteolin, wogonin, kaempferol, arachidonic acid, naringenin, acacetin, baicalein, iristectorigenin A and glycitein. Quercetin and vitamin C could be used in the prevention of high-risk populations or in the treatment of COVID-19 patients as adjunctive drugs for promising pharmacological agents such as remisivir or convalescent plasma [35]; a bioavailable form of quercetin should be considered a possible candidate to clinically face COVID-19 [36]. Luteolin could suppress systemic and neuroinflammatory responses in COVID-19 [37]; luteolin is a potential key to designing antiviral therapies for inhibiting viral proteases[38].Kaempferol 3-O-beta-rutinoside has the potential to be developed as a COVID-19 main protease inhibitor[39].Arachidonic acid may be useful both to prevent and manage COVID-19 [40].Naringenin may be a promising treatment strategy against COVID-19 [41]. Baicalein and baicalin have been characterized as the first noncovalent,nonpeptidomimetic inhibitors of SARS-CoV-2 3CLpro and exhibited potent antiviral activities in a cell-based system [42].These studies are related to the regulation of anti-virus, anti-inflammatory effects. KGD action on COVID-19 should be the result of the interaction of these multiple compounds.However,there are still few related studies about these compounds acting on COVID-19,which can be further studied.

PPI network showed that the main key targets were VEGFA, MAPK3, MAPK1, CASP3, IL6, ALB,MAPK8, FOS, PTGS2, CCL2, CASP8, CCND1,MAPK14, IL1B, ICAM1, RELA, BCL2L1, EGFR,IL2 and CAT. Previous research has indicated that levels of Levels of IP-10, HGF, IL-6, MCP-1,MIP-1alpha, IL-12p70, IL-18, VEGF-A, PDGF-BB and IL-1RA significantly correlate with SARS-CoV-2 infection severity [43]. SARS-CoV-2 ORF3a can efficiently induce apoptosis in cells, activated caspase-3 associate with AORF3a [44]. The serum levels of IL6 and CRP can effectively assess disease severity and predict outcome in COVID-19 patients[45]; blocking the signal transduction pathway of IL-6 is expected to become a new method for the treatment of severe COVID-19 patients [46]. Immune signatures of COVID-19 relate to shifts in neutrophil to T cell ratio, elevated serum IL-6 and modulation of CD14(+)monocyte phenotype and function and modified features of CD14(+) monocytes included poor induction of the prostaglandin-producing enzyme,COX-2, as well as enhanced expression of the cell cycle marker Ki-67 [47]. COVID-19 patients show profound and sustained T CD4+, CD8+ and B lymphopenia, higher HLA-DR expression on monocytes and higher serum concentrations of EGF,GM-CSF, IL-10, CCL2/MCP-1, CCL3/MIP-1a,CXCL10/IP-10, CCL5/RANTES, and CCL20/MIP-3a[48]. SARS-CoV-2 triggers inflammatory responses and cell death through caspase-8 activation [49].Critical cases of COVID-19 patients exhibit stronger interactions between epithelial and immune cells, as indicated by ligand-receptor expression profiles, and activated immune cells, including inflammatory macrophages expressing CCL2, CCL3, CCL20,CXCL1, CXCL3, CXCL10, IL8, IL1B and TNF [50].IL-1-beta, IL-6, TNF-alpha, the cell adhesion molecules ICAM-1, VCAM-1, and E-selectin, and a group of IFN-gamma-induced genes have been specifically identified in the"cytokine storm"observed in fatal cases of COVID-19 [51]. The SARS-CoV-2 infection strongly activates TNF and NF kappaB-signaling pathways through significant upregulation of the TNF, IL1B, IL6, IL8, NFKB1,NFKB2 and RELB genes [52]. The treatment of cytokine storm cytokine release syndrome,characterized by IL-6, IL-2, IL-7, IL-10 has been proposed as a critical part of rescuing severe COVID-19 [53]. The relationship between these genes and COVID-19, some of which already have relevant studies, can be further studied, and possible mechanisms can explore about those genes, which have not been studied before. All of these have explained that the action of KGD on COVID-19 is related to multiple targets.

We also performed molecular docking to verify specific interactions between key compounds and their predicted protein targets, which could improve the accuracy of the network [54]. Preliminary molecular docking results show the key active compounds in KGD have high binding activities with the key gene target proteins. These active compounds may be some important material basis for KGD treating COVID-19 through related signaling pathways.

We have been treating COVID-19 patients since January 23, 2020. KGD has been applied to the patients, and has achieved certain curative effect. We demonstrate two typical clinical cases with obvious efficacy after KGD application.

The antipyretic effect of KGD is especially obvious.The patient in case 1 was characterized of fever.Ibuprofen was used to antipyretic, but the body temperature rose after a period. The application of KGD also played the same rapid antipyretic effect as ibuprofen, while the patient's body temperature is not elevated after KGD continuous used one dose a day.First, we believe that KGD has the same target as ibuprofen and plays an antipyretic role. In fact,ibuprofen is a non-selective COX inhibitor and the inhibition of COX-2 activity decreases the synthesis of prostaglandins involved in mediating inflammation,pain,fever,and swelling; ibuprofen has a specific drug target PTGS2 to play a series of pharmacological effects [55]. Our network pharmacology explains that PTGS2 is one of the key targets of KGD acting on COVID-19, and molecular docking shows that the key active compounds in KGD have high binding activities with PTGS2.

KGD may be involved in anti-inflammatory,anti-virus processes. As discussed above, KGD acting on COVID-19 is through multiple target and multiple pathways, which are related to anti-inflammatory and anti-virus. Major pathways involve a variety of inflammatory factors. The patient in case 2 inflammatory factor elevated. With the combined therapy of KGD and western medicine,his temperature returned to normal rapid in the same way, and the inflammatory markers decrease gradually.Our network pharmacology explains that interleukin series, such as IL6, IL1B, and IL2 are the key targets of KGD acting on COVID-19, and molecular docking shows that the key active compounds in KGD have high binding activities with them.

These curative effects may be strongly associated with the results of our network pharmacology and molecular docking. Around these clinical outcomes,we will pay further attention to obtain a higher level of clinical evidence in order to facilitate the combination of network pharmacology and molecular docking results for further basic research.

The pharmacological mechanisms of KGD acting on COVID-19 have been explored through a network pharmacology approach and preliminary verification has been performed,but there are still some limitations in our study. First, we searched compounds and target genes of KGD from the TCMSP database, the screening criteria for effective compounds were fixed,and COVID-19 related target genes were achieved from the GeneCards database. Some compounds and target genes may still be omitted, although these databases are more comprehensive presently applicable.Other methods, such as literature, may contain more compounds or genes. Second, we performed GO function enrichment and KEGG pathway enrichment analysis and constructed a PPI network to investigate the target genes and pathways of KGD acting on COVID-19.These potential target genes and pathways need to be further studied by experimental analysis.Third, we only carry out preliminary molecular docking verification and clinical cases demonstration.The research on molecular docking of small-molecule compounds and macromolecular protein target receptors needs further study, while high evidence level of clinical research needs to be conducted.

Conclusion

In conclusion, this study explored the relationships between the compounds, target genes of KGD acting on COVID-19, and screened out the key compounds and key targets of KGD acting on COVID-19. In a word, the pharmacological mechanism of KGD acting on COVID-19 has been explored, and the active compounds and targets of KGD acting on COVID-19 and clinical effects for KGD treating COVID-19 patients have been preliminarily verified.