key: cord-0016496-92bcw6m2 authors: Zhang, Mingxu; Yang, Jiawei; Zhao, Xiulan; Zhao, Ying; Zhu, Siquan title: Network pharmacology and molecular docking study on the active ingredients of qidengmingmu capsule for the treatment of diabetic retinopathy date: 2021-04-01 journal: Sci Rep DOI: 10.1038/s41598-021-86914-8 sha: 2b54c1a1dbe5fa9446b45ae8dc4aa0e4934f2166 doc_id: 16496 cord_uid: 92bcw6m2 Diabetic retinopathy (DR) is a leading cause of irreversible blindness globally. Qidengmingmu Capsule (QC) is a Chinese patent medicine used to treat DR, but the molecular mechanism of the treatment remains unknown. In this study, we identified and validated potential molecular mechanisms involved in the treatment of DR with QC via network pharmacology and molecular docking methods. The results of Ingredient-DR Target Network showed that 134 common targets and 20 active ingredients of QC were involved. According to the results of enrichment analysis, 2307 biological processes and 40 pathways were related to the treatment effects. Most of these processes and pathways were important for cell survival and were associated with many key factors in DR, such as vascular endothelial growth factor-A (VEGFA), hypoxia-inducible factor-1A (HIF-1Α), and tumor necrosis factor-α (TNFα). Based on the results of the PPI network and KEGG enrichment analyses, we selected AKT1, HIF-1α, VEGFA, TNFα and their corresponding active ingredients for molecular docking. According to the molecular docking results, several key targets of DR (including AKT1, HIF-1α, VEGFA, and TNFα) can form stable bonds with the corresponding active ingredients of QC. In conclusion, through network pharmacology methods, we found that potential biological mechanisms involved in the alleviation of DR by QC are related to multiple biological processes and signaling pathways. The molecular docking results also provide us with sound directions for further experiments. One of the microvascular complications of diabetic mellitus (DM), diabetic retinopathy (DR), has become the main cause of irreversible visual impairment worldwide and is affecting a growing proportion of DM patients 1 . For diabetic patients with mild or imperceptible ocular symptoms who do not undergo regular fundus examinations, early diagnosis and effective control are difficult in the initial phase of DR. The traditional DR treatment of laser photocoagulation combined with anti-vascular endothelial growth factor (anti-VEGF) administration has seemed to be a considerably effective strategy for nonproliferative diabetic retinopathy (NPDR, diabetic retinopathy without neovascularization) patients 2 . Other treatments for DR include oral calcium dobesilate treatment, insulin administration, weight loss, control of blood pressure and lipids, and even surgery for proliferative diabetic retinopathy (PDR, diabetic retinopathy with neovascularization) patients with retinal detachment 3 . Recently, some new drugs have been used in in vivo or in clinical trials (such as canakinumab and doxycycline) 4 . With regard to molecular biology, neovascularization induced by vascular endothelial growth factor (VEGF) has been proposed to play a leading role in the development of PDR. Diverse herbs and natural products are used in China as complementary treatments for DM. However, the effects and mechanisms of the ameliorative effects of traditional Chinese medicines (TCMs) on DR remain controversial and unclear, and proper methods are needed to explore and fully explain them in detail. A Chinese www.nature.com/scientificreports/ patent medicine used to treat DR, Qidengmingmu Capsule (QC), contains 3 herbs: Erigeron breviscapus, Astragalus membranaceus and Radix Puerariae. Erigeron breviscapus has been proven to exert anti-insulin resistance and anti-oxidative stress effects in the treatment of DM 5 . The extract of Astragalus membranaceus, Astragalus polysaccharide (APS), also exerts validated therapeutic effects on DM by increasing insulin sensitivity through the AMPK pathway 6 . In a clinical trial performed by Xie, patients with nonproliferative DR who did not meet the indications of retinal photocoagulation or anti-VEGF treatment received QC therapy for 12 weeks. The results showed significant improvements in visual acuity, symptom complaints and retinal lesions. The safety of QC was also verified by the biochemical stability of blood and secretions during the whole treatment period and even 3 months later 7 . TCMs have been developed for thousands of years, and abundant formulae have been created by TCM practitioners to alleviate patients' suffering. However, it is difficult to modernize TCM through conventional pharmacological and biochemical methods due to the complex ingredients and unclear mechanisms. Network pharmacology, a revolutionary field based on systematic biology and network analysis, provides feasible and reliable ways to explore potential molecular mechanisms beyond the barriers associated with the current one drug-one target-one pathway research mode 8 . Based on the interactions between active herbal ingredients and potential targets of disease, network pharmacological methods can identify the potential mechanisms and key targets from network topological analysis. These methods can also be used to identify synergistic effects of herbal ingredients, to understand combinatorial rules of TCM formulae, and ultimately to help clinical practitioners design herbal formulae rationally 9 . The molecular docking technique plays an important role in drug-target binding prediction and synthetic drug design, enabling prediction of docking patterns and binding affinities (indicating the stability of docking patterns) of drug ligands and protein receptors. To explore and validate the potential mechanism of QC in the treatment of DR, we performed this study based on the methods mentioned above. Active ingredients of QC and QC-related target screening. Information about the ingredients of QC was obtained from the Traditional Chinese Medicine Systems Pharmacology online platform (https:// tcmspw. com/ tcmsp. php, TCMSP, Version 2.3) by searching the keywords "Erigeron Breviscapus", "Astragalus membranaceus" and "Radix Puerariae" 10 . The current formula used by TCM practitioners contains multiple and complicated ingredients, but not all of them have therapeutic effects on disease. Oral bioavailability (OB) represents the ability of a drug to enter the circulation. Druglikeness (DL), a valuable parameter influenced by molecular physical and biochemical properties, represents the similarity between a molecule and known drugs. A desirable OB and DL suggest that a drug candidate has properties that make it suitable for further study 11, 12 . The current computationally oriented drug discovery strategy integrates multiple assessments, such as quantitative estimate of druglikeness (QED) and Lipinski's rule-of-five (Ro5, a rule used to predict drug-likeness) assessments 13, 14 . In the present study, we selected ingredients with criteria of an OB ≥ 30% and a DL ≥ 0.18 based on similar previous studies [15] [16] [17] . Details about the corresponding targets of the active ingredients above were acquired from the TCMSP platform and validated from existing studies. The UniProt database (https:// www. unipr ot. org, updated on August 12, 2020) was used to obtain gene symbols and related information about the QC targets acquired from TCMSP. DR-related target screening. The keyword "Diabetic Retinopathy" was imported to acquire gene symbols of DR-related targets from the GeneCards (Version 5.0) and OMIM (updated on August 31, 2020) databases. We also obtained a BeadChip gene expression dataset (GSE60436) and a platform file (GPL6884) on 3 healthy retina samples and 6 retinal fibrovascular membrane (FVM, pathological tissue of PDR) samples of DR patients from the NCBI-GEO database. Differential gene expression analysis was performed with the Limma packages in R software (version 3.6.0) under the screening condition of an absolute logarithm of the fold change (log FC) value > 0.5. Heatmaps and volcano plots of the differentially expressed genes were created with the Pheatmap package of R software. Finally, we summarized the differentially expressed genes acquired from GEO and the gene symbols acquired from GeneCards and OMIM as the total DR-related targets. Active ingredient-DR target and PPI network construction. The QC targets intersecting with DRrelated targets were taken as QC-DR common targets. With the assistance of the VennDiagram package of R, we visualized the common targets with a Venn diagram. Cytoscape software (version 3.7.2) was used to construct an active ingredient-DR target network based on the active ingredients of QC and the QC-DR common targets. The different nodes represent DR targets, common targets and QC active ingredient targets, and the nodes are connected by edges (lines), which denote interactions between the nodes. The degree value of a node is denoted by the number of linked edges, which is consistent with the importance of the node. We imported the common targets into the STRING platform (version 11.0) and set the species to Homo sapiens and the confidence to 0.950 to obtained the concise protein-protein interaction (PPI) information for further analysis. The PPI network was also visualized with Cytoscape software. Using the ClusterProfiler package in R software 18 , we performed Gene Ontology (GO) biological processes and Kyoto Encyclopedia of Genes and Genomes (KEGG) metabolic pathway enrichment analyses on the QC-DR common targets [19] [20] [21] . According to the gene ratios, we chose the top 20 biological processes and pathways (P < 0.05) and show them as bubble or column charts that contain the gene counts, gene ratios, adjusted P values and other detailed information. The GO enrichment analysis included 3 categories: the biological process, cellular component and molecular function categories. The results of KEGG enrichment analysis were used to construct a KEGG key pathway network to discover important target proteins www.nature.com/scientificreports/ involved in the treatment effects of QC. Details about the involved pathways and key targets were downloaded from online platforms to explore the potential mechanisms of the treatment effects on DR. drugs, can simulate intermolecular combinational patterns between drug ligands and target proteins in 3-dimensional (3D) structures to predict possible docking modes and binding affinities. According to the degrees of common targets in the PPI network and their importance in DR etiology, we confirmed key targets and their corresponding active ingredients to perform molecular docking. Molecular 3D structures of active ingredients and key proteins were obtained from TCMSP databases and the RCSB PDB (http:// www. rcsb. org/), respectively, and then docking validation was performed after hydrogenation and dehydration through AutoDockTools (version 1.5.6) and AutoDock Vina (version 1.1.2) 22 . Because the functional binding sites of DR-related targets with ingredients of QC remain unknown, all binding sites were limited within the areas of docking pockets (the possible combinational areas of target protein) predicted by DeepSite software (https:// www. playm olecu le. com) 23 . The molecular docking patterns were visualized with PyMOL (version 2.4). Active ingredient and common target screening. We obtained 28 ingredients and 193 QC-related targets that satisfied the criteria of an OB ≥ 30% and a DL ≥ 0.18 by searching the keywords "Erigeron Breviscapus", "Astragalus membranaceus" and "Radix Puerariae" in the TCMSP database. We also obtained 3497, 213 and 4150 DR-related targets from the GeneCards, OMIM and NCBI-GEO databases, respectively. After removal of repetitive gene symbols, we acquired 7034 DR-related targets from the databases above. We generated a heatmap ( Fig. 1A ) and volcano plot (Fig. 1B ) of differentially expressed genes from a BeadChip gene expression dataset (GSE60436) from the NCBI-GEO database. After the DR-related targets and QC-related targets were intersected, we acquired 134 QC-DR common targets (shown in Fig. 1C ). By tracing the common targets, we confirmed that 20 active ingredients of QC are involved in the treatment of DR. Details about these active ingredients (chemical name, CAS number, oral bioavailability and druglikeness) are listed in Fig. 1D . The highest OB% was 101.06%, belonging to 1-hydroxy-2,3,5-trimethoxy-xanthone, and the highest DL was 0.75, belonging to beta-sitosterol and hederagenin. The CAS numbers of 3,9-di-O-methylnissolin and 7-O-methylisomucronulatol are not available. We imported the active ingredients and common targets into Cytoscape 3.7.2 to construct an active ingredient-DR target network. In the network ( Fig. 2A ), 20 active ingredients (shown as green rectangular nodes), 134 common targets (shown as blue oval nodes) and the edges among them are clearly exhibited. The network reflects the multiple and complicated effects of active ingredients of QC on DR. The PPI information from the STRING platform was imported into Cytoscape software and used to construct a PPI network based on common targets. The protein nodes are ordered by degree according to color from deep violet (high degree) to deep blue (low degree). The median degree was five and is set as the neutral color, gray (Fig. 2B ). According to the degree, the top 20 common targets are shown in a column chart (Fig. 2C ). We found that several key DR-related proteins, such as AKT1, HIF1A, VEGFA and TNF α, are involved in the treatment effects of QC on DR. Through GO enrichment analysis based on common QC-DR targets, we obtained a total of 2307 terms related to the treatment effects of QC on DR, and these terms could be divided into 3 categories (2108 terms in the biological process category, 50 terms in the cellular component category and 149 terms in the molecular function category). The top 20 terms in the 3 categories above are shown as bubble charts in Fig. 3A -C. According to the results of GO enrichment analysis, the active ingredients in QC primarily target the response to oxidative stress, the cytokine system, the activity of transcription factors and the regulation of apoptosis, which have been verified to play important roles in the incidence of DR 24, 25 . The 10 biological process terms with the lowest adjusted P values in each category (a total of 30 terms) are shown in the column chart in Fig. 3D . Through the results of KEGG pathway enrichment analysis, we acquired 39 signaling pathways involved in the possible mechanism by which QC treatment affects DR and show the top 20 pathways in Fig. 4A , B according to the gene ratios. The results suggested that QC-DR common targets are concentrated mainly in VEGFA signaling, HIF-1 signaling, PI3K-Akt signaling, NF-κB signaling and apoptosis-related signaling pathways, which heavily participate in the etiology of DR. To analyze the significance and importance of key targets in the pathways involved in the treatment effects of QC, we chose 10 key pathways according to the gene ratios and adjusted P values from the results of KEGG enrichment analysis and related targets (details listed in Fig. 4C ) to construct a KEGG key pathway network (Fig. 4D) . Analysis of the KEGG key pathway network revealed 14 target factors involved in more than 4 pathways, which included AKT1, VEGFA, BCL2, TP53, BCL2L1, MAPK1, RAF1, PRKCA, PRKCB, ERBB2, CCND1, CDKN1A, EGF and EGFR. Based on the results of the enrichment analysis, we propose that QC could target multiple functional and biological factors in DR, but the effects and profound influence still need further research for validation. Key target-ingredient molecular docking analysis. On the basis of the results of KEGG enrichment analysis, we examined several pathways that may be highly related to the incidence of DR, including the PI3K-AKT signaling pathway (which had the highest degree of gene enrichment), the HIF-1 signaling pathway (the initial pathway for retinal neovascularization), and apoptosis-related and NF-κB signaling pathways (which participate in the corruption of the blood-retinal barrier in DR) 26, 27 . The PPI network suggested that AKT1, HIF1A, TNFα and VEGFA represent key targets in the pathways mentioned above, and these nodes showed . The crystal structures of key target proteins were obtained from the RCSB PDB; these proteins were AKT1 (1unq), HIF1A (4h6j), TNFα (5yoy) and VEGFA (4zff). It is difficult for TCM researchers to identify the accurate binding sites of target proteins for herbal ingredient molecules, and the functional binding sites of DR-related targets for ingredients of QC remain unknown. To ensure the accuracy of the docking prediction results, we used DeepSite to predict the possible docking pockets of the target proteins 23 . The docking analysis of the ingredients and proteins above was performed with the assistance of AutoDock Vina (version 1.1.2), and the binding areas were limited within the docking pockets predicted by DeepSite. The docking patterns and binding affinities are shown in Fig. 5A , B. The docking results are shown as molecular surface representations, which can effectively reflect the topical details of binding sites. The binding sites are shown in different colors on the protein surface, and hydrogen bonds are shown as dotted lines (Fig. 5A) . The absolute values of binding affinities (kcal/mol) of all docking patterns were greater than 5 kcal/mol, indicating stable combination, and the details of the binding affinities are shown in Fig. 5B . DR has become the leading threat to global vision health with the increasing incidence of DM and the decreasing age of DM patients. Microvascular lesions are the main pathological change in DR, and they often arise long before the occurrence of appreciable symptoms 4 . Clinical scientists have found that the foveal avascular zone (FAZ, a physiological avascular zone located in retinal macular area) in DM patients without DR is enlarged compared with that in healthy people by optical coherence tomography angiography (OCTA, a widely used noninvasive optical examination for retinal vascular structure) examination, which means that retinal degeneration in DM patients is initiated earlier than we previously thought 28, 29 . QC is a traditional Chinese patent medicine (TCPM) that contains multiple active ingredients extracted from 3 herbs: Erigeron breviscapus, Astragalus membranaceus and Radix Puerariae. In in vivo experiments, QC has been proven to alleviate blood-retinal barrier (BRB, functional barrier to sustain the retinal fluid balance and to prevent leakage from vessel Diabetic retinopathy: Battling the global epidemic The early treatment diabetic retinopathy study historical review and relevance to today's management of diabetic macular edema Emerging insights and interventions for diabetic retinopathy The anti-insulin resistance effect of scutellarin may be related to antioxidant stress and AMPKα activation in diabetic mice Astragalus polysaccharide improves insulin sensitivity via AMPK activation in 3T3-L1 adipocytes Clinical Study of Qidengmingmu Capsule in the Treatment of Non-Proliferative Diabetic Retinopathy Network-based technologies for early drug discovery Traditional chinese medicine network pharmacology: Theory, methodology and application TCMSP: A database of systems pharmacology for drug discovery from herbal medicines A novel chemometric method for the prediction of human oral bioavailability The application of in silico drug-likeness predictions in pharmaceutical research The evolving druggability and developability space: Chemically modified new modalities and emerging small molecules Quantifying the chemical beauty of drugs Integrated meta-analysis, network pharmacology, and molecular docking to investigate the efficacy and potential pharmacological mechanism of Kai-Xin-San on Alzheimer's disease Systematically deciphering the pharmacological mechanism of fructus aurantii via network pharmacology. Evid. Based Complement Altern Analysis of the mechanism of shufeng jiedu capsule prevention and treatment for COVID-19 by network pharmacology tools ClusterProfiler: An R package for comparing biological themes among gene clusters Kyoto encyclopedia of genes and genomes Toward understanding the origin and evolution of cellular organisms KEGG: Integrating viruses and cellular organisms AutoDock vina: Improving the speed and accuracy of docking with a new scoring function, efficient optimization, and multithreading DeepSite: Protein-binding site predictor using 3D-convolutional neural networks Diabetic retinopathy: Hyperglycaemia, oxidative stress and beyond Perturbed biochemical pathways and associated oxidative stress lead to vascular dysfunctions in diabetic retinopathy Effects of HIF-1α on diabetic retinopathy angiogenesis and VEGF expression Akt1/protein kinase Balpha is critical for ischemic and VEGF-mediated angiogenesis Optical coherence tomography angiography in primary eye care Optical coherence tomography angiography in diabetic patients without diabetic retinopathy Effect of chinese medicine Qidengmingmu capsule on the STZ induced hyperglycemia rats blood-retinal barrier The effect of Qidengmingmu capsule on vascular endothelial growth factor in vitreous in spontaneous diabetes rats Kaempferol protects retinal ganglion ceils from high-glucose-induced injury by regulating vasohibin-1 Quercetin attenuates high glucose-induced injury in human retinal pigment epithelial cell line ARPE-19 by up-regulation of miR-29b Upregulation of TLR4 via PKC activation contributes to impaired wound healing in high-glucose-treated kidney proximal tubular cells Effect of Qidengmingmu capsule on protein kinase C in retina of spontaneous diabetic rats. Rec Prevention of hyperglycemia-induced myocardial apoptosis by gene silencing of toll-like receptor-4 Diabetes mellitus/poststroke hyperglycemia: A detrimental factor for tPA thrombolytic stroke therapy Oxidative stress-related mechanisms and antioxidant therapy in diabetic retinopathy Luteolin protects microglia against rotenone-induced toxicity in a hormetic manner through targeting oxidative stress response, genes associated with Parkinson's disease and inflammatory pathways Baicalein protects rats with diabetic cardiomyopathy against oxidative stress and inflammation injury via phosphatidylinositol 3-kinase (PI3K)/AKT pathway Hypoxia-inducible factor-1 (HIF-1) Distinct downstream signaling and the roles of VEGF and PlGF in high glucose-mediated injuries of human retinal endothelial cells in culture Endothelial cell autonomous role of Akt1: regulation of vascular tone and ischemia-induced arteriogenesis. Arterioscler Nitric oxide in the pathophysiology of retinopathy: Evidences from preclinical and clinical researches Adiponectin inhibits high glucose-induced angiogenesis via inhibiting autophagy in RF/6A cells Cellular death, reactive oxygen species (ROS) and diabetic complications Egr1 mediates retinal vascular dysfunction in diabetes mellitus via promoting P53 transcription Catechin weakens diabetic retinopathy by inhibiting the expression of NF-κB signaling pathway-mediated inflammatory factors Apocynin ameliorates diabetic retinopathy in rats: Involvement of TLR4/NF-κB signaling pathway A network pharmacology approach to determine active compounds and action mechanisms of Ge-Gen-Qin-Lian decoction for treatment of type 2 diabetes A comprehensive application: Molecular Docking And Network Pharmacology For The Prediction Of Bioactive Constituents And Elucidation Of Mechanisms Of Action In Component-Based Chinese medicine All drawings in Fig. 6 of the article were drawn by author Mingxu Zhang, and there is no copyright conflict with existing compositions. Z.M.X. wrote the manuscript, drew the figures and performed data analysis; Y.J.W. and Z.X.L. collected data about herbal ingredients and targets of both the drug and the disease, and improved the manuscription; Z.Y. complied with molecular docking and explained the results; Z.S.Q. directed the research and proposed changes on the manuscript; all authors reviewed the manuscript and approved the final version of the manuscript. This study was supported by National Natural Science Foundation of China (No. 81704125) as well as Open Funding Project of National Key Laboratory of Human Factors Engineering (No. SYFD180051808K). The authors declare no competing interests. Correspondence and requests for materials should be addressed to S.Z.Reprints and permissions information is available at www.nature.com/reprints.Publisher's note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http:// creat iveco mmons. org/ licen ses/ by/4. 0/.