Mechanism of total flavonoids in the treatment of rheumatoid arthritis based on network pharmacology

LIU Xiao-yu, CHEN Guang-yao, WU Zi-yu, JIN Qi, TAO Qing-wen

1. Beijing University of Chinese Medicine, Beijing 100029, China

2. Department of TCM Rheumatology of China-Japan Friendship Hospital, Beijing Key Lab for Immune-Mediated Inflammatory Diseases, China-Japan Friendship Hospital, Beijing 100029, China

Keywords:Total flavonoids of Rhizoma Drynariae Rheumatoid arthritis Network pharmacology Luteolin

ABSTRACT Objective: To explore the mechanism of total flavonoids of Rhizoma Drynariae (TFRD) in the treatment of rheumatoid arthritis (RA) based on network pharmacology. Methods: The active components of TFRD were identified from the Traditional Chinese Medicines for Systems Pharmacology Database and Analysis Platform (TCMSP). Relevant targets of TFRD were predicted by TCMSP, PubChem, SwissTargetPrediction and STITCH database, and standard target names were obtained from Uniprot database. RA-related targets were retrieved from GenBank, GeneCards, DisGeNet and OMIM databases. The intersection genes were imported into STRING database to construct a protein-protein interaction (PPI) network, in order to predict the core proteins. GO functional annotation analysis and KEGG pathway enrichment analysis were then performed. A “drug-component-target-disease” regulatory network was constructed using Cytoscape 3.8.2, and the active components of TFRD were molecularly docked with the predicted core targets. Results: A total of 10 active compounds,including luteolin, naringenin and kaempferol, were screened out from TCMSP database, and 210 relevant targets were predicted. A total of 2 009 RA-related targets were screened out,and 123 targets of TFRD-RA intersection targets were obtained. STAT3, MAPK1, MAPK3,AKT1, MAPK8, IL-6, TNF, MAPK14, IL-4, IL-2, VEGFA, IL-1β and MAPK9 may be the key targets of TFRD in the treatment of RA. GO functional enrichment analysis suggested that transcriptional regulation and cytokine activity regulation may play an important role in TFRD in RA treatment. KEGG pathway enrichment analysis suggested that PI3K/AKT, IL-17, TNF-α, T cell receptor transduction and regulation of Th17 cell differentiation may play an important role in the treatment of RA by TFRD. The regulatory network suggested that luteolin and naringenin may be the key components of TFRD in the treatment of RA, which also had good affinity with the core targets AKT1, PI3K and STAT3 by molecular docking,which may further affect their phosphorylation by changing the molecular conformation.Conclusion: TFRD may inhibit the inflammatory response of RA by regulating inflammationrelated cytokines and related conduction pathways, and PI3K/AKT pathway may be an important pathway of TFRD in the treatment of RA. Luteolin is the key component of TFRD in the treatment of RA, and may play a vital role in the downstream pathway by inhibiting the conformation of related core proteins and thereby regulating phosphorylation.

1. Introduction

Rheumatoid arthritis (RA) is a chronic autoimmune inflammatory disease mainly characterized by symmetrically accumulated facet joints, with a worldwide incidence of about 0.5%-1.0%[1] . The main pathological manifestations of RA are abnormal proliferation of synovial membrane caused by inflammatory response, which eventually leads to the formation of new vessels in synovial membrane and further bone erosion [2]. Uncontrolled RA often leads to joint damage and eventually joint deformity and loss of function [3]. The disease-modifying anti-rheumatic drugs(DMARDs), represented by methotrexate, are the most commonly used by modern physicians to treat RA. However, a series of adverse reactions limit its clinical use, such as gastrointestinal reactions, bone marrow suppression and liver function damage [4].Traditional Chinese medicine (TCM) treatment of RA has few side effects and definite curative effect, and can be targeted according to the clinical characteristics of patients [5]. Tripterygium wilfordii polyglycosides extracted from TCM have been proved by multiple RCT experiments to have a good improving effect on RA, and have been recommended in the clinical treatment of RA by relevant guidelines [6]. The extraction of effective components from TCM and its application in the clinical treatment of RA has become a hot spot in current research.

Rhizoma Drynariae is the dry rhizome of Drynaria fortunei(Kunze) and has the function of healing wounds, relieving pain,tonifying kidney and strengthening bone, and is often used in the treatment of rheumatism related diseases and fractures. Flavonoids are the main active ingredients in Rhizoma Drynariae, and Total flavonoids of Rhizoma Drynariae (TFRD) is approved to be used in the clinical treatment of osteoporosis. Relevant studies have shown that TFRD have a good intervention effect on fracture,osteoarthritis, streptomycin ototoxicity and atherosclerosis [7,8]. Relevant basic studies have shown that TFRD have a good inhibitory effect on osteoclast activity, inflammatory cytokines and matrix metalloproteinases expression [9,10]. Osteoporosis is often accompanied by RA. In clinical treatment of patients with RA accompanied by osteoporosis, it was found that TFRD could relieve the symptoms of RA, such as joint swelling and pain. At the same time, the occurrence of bone erosion in RA is closely related to abnormal activation of osteoclasts, overexpression of inflammatory cytokines, and increased secretion of matrix metalloproteinases.Therefore, it is speculated that TFRD have a certain intervention effect on RA. Therefore, this study intends to explore the mechanism of its treatment of RA through network pharmacology, to provide a basis for further experimental research.

2. Materials and methods

2.1 Screening of effective flavonoids from Rhizoma Drynariae

Taking "Drynaria" as the key word, the chemical constituents were screened in Traditional Chinese Medicine Systems Pharmacology Database and Analysis Platform (TCMSP, https://tcmspw.com/tcmsp.ph). All the components satisfied the following rules were selected: OB 30% and DL 0.18, and the active TFRD were further screened through molecular structure.

2.2 Target prediction of effective flavonoids in Rhizoma Drynariae

The related targets of active TFRD were included in TCMSP database (https://www.Uniprot.org/). The abbreviation of the targets was standardized. At the same time, PubChem database (https://pubchem.ncbi.nlm.nih.gov/) was used to search the SMILE standard structure of active ingredients and import into SwissTargetPrediction database (http://swisstargetprediction.ch/) and STITCH database(http://stitch.embl.de/) to carry out target prediction, and the targets with confidence ＞ 0.4 were included in the following study.

2.3 Target prediction of RA related diseases

GenBank database (https://www.ncbi.nlm.nih.gov/), Genecards database (https://www.genecards.org/), DisGeNET database (https://www.disgenet.org/) and Online Human Mendelian Genetic database(OMIM, https://omim.org/) was used to search for RA-related targets. Using "rheumatoid arthritis" as the key word, the diseaserelated genes were searched in the above databases and the duplicate genes were deleted. The Venn map of target- disease gene was drawn.

2.4 Construction of protein-protein interaction network and screening of key proteins

The protein interaction network of drug-disease common targets was obtained on STRING database (https://string-db.org/). The above intersection protein targets were imported into STRING database, and the species was set as "Homo sapiens". The proteinprotein interaction (PPI) scores higher than 0.9 were screened out, and the PPI network was constructed by importing them into Cytoscape 3.8.2. The plug-in “Analyze Network” in Cytoscape 3.8.2 was used to analyze the topological feature of the PPI network, and the protein with degree greater than the average value was selected as the key protein, and the network of key protein was created.

2.5 Construction of regulatory network of Drynaria total flavonoids against RA

The active components, target components and RA intersection genes of TFRD were imported into Cytoscape 3.8.2 to construct the regulatory network of TFRD-RA.

2.6 GO enrichment analysis

The targets were imported into DAVID enrichment database(https://david.ncifcrf.gov/). H. sapiens was selected to analyze the target genes and predict the functional distribution of genes.According to the p-value from small to large, the biological process,molecular function and cellular component were sorted, and the top 10 biological processes were selected respectively. The number of enriched genes was counted, and the bar graph was drawn with GraphPad Prism 8.

2.7 KEGG pathway enrichment analysis

The KEGG pathway of drug-disease genes was enriched and analyzed, and the important pathways that may be regulated by active components were obtained. The top 20 pathways with P value＜ 0.05 were screened out and visualized.

2.8 Molecular docking of active ingredients with core targets

The key components of TFRD and the selected key proteins were selected for molecular docking. The key proteins’ structure were downloaded and saved it in RCSB PDB database (http://www1.rcsb.org/) and in PDB format. The structures of key components of TFRD was downloaded from PubChem database, and the structure of key components was converted into mol2 format through Open Babel GUI. Autodock tools 1.5.6 was used to conduct the docking between key components of Rhizoma Drynaria flavonoids and key protein,and PyMOL was used to visualize the docking results.

3. Results

3.1 Screening of active TFRD

Based on the keywords of "Gusuibu", a total of 71 active components of Gusuibu were found in the TCMSP database, among which 10 flavonoids with OB 30% and OL 0.18 were found,including luteolin, kaempferol, and naringenin. Specific information of active TFRD is shown in Table 1.

Tab 1 Components of effective flavonoids of Rhizoma Drynariae

3.2 Target prediction of active TFRD

71 relevant targets of active TFRD were identified by TCMSP.With Confidence ＞ 0.4 as the condition, 61 and 108 related targets were retrieved from SwissTargetPrediction database and STITCH database, respectively. Repeated targets were deleted, and a total of 210 targets corresponding to active components were obtained.

3.3 Prediction of genes associated with RA disease

Based on "Rheumatoid arthritis", 1 263 related genes were found in GenBank database, and 898 related genes (Relevance≥4) were found in GeneCard database. 867 related genes were found in DisGeNET database (Score≥0.05), and 42 related genes were found in OMIM database. After deleting the duplicated genes, 2 009 genes were related to RA. 123 intersection targets of active TFRD were found,and the Venn diagram of the intersection genes is shown in Figure 1.

Fig 1 Venn diagram of intersection genes of total flavonoids of Rhizoma Drynariae (TFRD) and rheumatoid arthritis (RA)

3.4 PPI network and core protein of RA treated by total flavonoids of Rhizoma Rhizoma

In this study, the STRING database was used to screen out the protein relationship with an interaction score greater than 0.9 to construct the PPI network of TFRD in the treatment of RA, and it was imported into Cytoscape 3.8.2 for visualization (Figure 2A).There were 108 functional proteins and 541 interaction relationships in the network. The average node degree was 8.8, the local clustering coefficient was 0.478, and the expected edge value was 112. The number of linking targets (Degree) in the PPI network is positively correlated with the Degree of protein core. Proteins with Degree greater than the average value are selected as core proteins and the network map of core proteins is generated. The larger the node is and the darker the color is, the more linking targets of this protein are(Figure 2B). The number of core protein binding targets in the PPI network is shown in Figure 3. Proteins closely related to RA include STAT3, MAPK1, MAPK3, AKT1, MAPK8, IL-6, TNF, MAPK14,IL-4, IL-2, VEGFA, IL-1β, MAPK9, etc. It mainly involves inflammatory cytokines (tumor necrosis factors, interleukin),cytokine signal transduction pathways (STAT and MAPK signaling pathways), neovascularization and other pathways.

Fig 2 PPI network and core protein network

Fig 3 Bar graph of the number of core protein junction targets in the PPI network

3.5 Regulation network of total flavonoids in the treatment of RA

In this study, 10 active components of TFRD were screened out,among which 7 active components, including luteolin, naringenin and kaempferol, were involved in the RA regulatory network.Cytoscape 3.8.2 was used to construct a drug-active ingredientsaction target-disease visual regulatory network (Figure 4). There were 132 nodes (123 action targets, 7 active ingredients, 1 disease name, 1 drug name) and 253 connections in the network.

In the figure, the blue rectangle represents TFRD, the red hexagon represents rheumatoid arthritis, the green oval represents the active components of TFRD, the pink "V" shape represents the common drug-disease targets, and the line represents the relationship among drugs, active components, action targets and diseases. The size of nodes of active components of TFRD reflects the number of node connecting targets, and the larger the nodes are, the more core the nodes are in the network. The results showed that the active components of TFRD could regulate multiple action targets, among which luteolin and naringin had the largest number of connections with targets, 125 and 29, respectively, suggesting that these two substances may be the key components of TFRD in the treatment of RA.

3.6 GO gene enrichment analysis

The GO gene enrichment analysis of RA targets treated by TFRD was carried out using the David Enrichment database, and the top 10 biological processes were screened out from small to large according to the p-value (Figure 5). GO enrichment analysis showed that in terms of biological processes, targets were related with positive regulation of transcription from RNA polymerase II promoter, positive regulation of transcription, positive regulation of gene expression, negative regulation of apoptotic process, extrinsic apoptotic signaling pathway in absence of ligand and inflammatory response, etc. In terms of cellular composition, the targets involved cytosol, extracellular region, protein complex and nucleoplasm.In terms of molecular function, targets were related with enzyme binding, protein binding, transcription factor binding, transcriptional activator activity, RNA polymerase II transcription factor activity,protein kinase activity and cytokine activity.

Fig 4 Regulatory network of drug-active ingredient-target-disease

Fig 5 Top 10 GO enrichment analysis results of biological process (BP), molecular function (MF) and cellular component (CC)

3.7 KEGG pathway enrichment analysis

KEGG pathway enrichment analysis was performed on the above TFRD-RA intersection genes, and 164 significantly enriched related pathways were obtained. The top 20 KEGG pathways were screened out according to the degree of enrichment significance from small to large (Figure 6). The horizontal axis in the figure represents the ratio of genes enriched in the pathway to the total number of genes in the pathway, the bubble size represents the number of genes enriched,and the bubble color represents the p-value. Combined with relevant literature analysis, it is suggested that the PI3K/AKT pathway [11],the IL-17 pathway [12], the TNF-α pathway [13], the T cell receptor transduction pathway [14] and the Th17 cell differentiation regulation pathway [15] may play an important role in the treatment of RA by TFRD [15]. The above pathways were combined with the intersection targets of TFRD-RA to construct a network (Figure 7).

Fig 6 KEGG pathway enrichment

Fig 7 Target-key pathway diagram

3.8 Molecular docking verification

In order to evaluate the binding affinities between key components and key targets of TFRD, AutodockTools were used for molecular docking. Luteolin and naringenin, which had the largest number of target connections in the RA regulatory network treated by TFRD,were selected as molecular docking ligands, and key proteins STAT3 in the protein interaction network and important regulatory proteins AKT1 and PI3K in the PI3K/AKT pathway were selected as receptors for molecular docking. The binding affinities of luteolin with AKT1, PI3K and STAT3 were -5.45, -6.95 and -6.85 kcal/mol,respectively. The binding affinities of naringenin with AKT1, PI3K and STAT3 were -3.37, -4.38 and -4.13 kcal/mol, respectively (Table 2). Compared with naringenin, luteolin has stronger binding ability to key targets AKT1, PI3K and STAT3, suggesting that luteolin is more likely to affect its structure and thus affect its phosphorylation,thereby regulating the STAT3 and PI3K/AKT signaling pathway.The molecular docking schematic diagrams of AKT1, PI3K, STAT3 and key components of TFRD are shown in Figure 8.

Tab 2 Binding energy of key components and key targets of total flavonoids of Rhizoma Drynariae(kcal/mol)

Fig 8 Molecular docking patterns of key components of total flavonoids of Rhizoma Drynariae and AKT1, PI3K and STAT3

4. Discussion

Previous studies have shown that flavonoids are the key effective components in the treatment of RA in TCM, and may be closely related to the anti-inflammatory and antioxidant mechanisms [16,17].The results showed that luteolin and naringenin might be the key components in the treatment of RA. In previous studies, luteolin has been proved to play a therapeutic role in RA by inhibiting the expression of inflammatory cytokines, matrix hydrolase and inflammatory factor related pathways in synovium. Naringenin has not only strong anti-inflammatory effect, but also can effectively relieve liver damage caused by methotrexate treatment of RA [19]. It is suggested that the TFRD has not only potential therapeutic effect on RA, but also the mechanism of reducing toxicity and efficiency by combining with western medicine, which is worth further investigation.

Cytokines are a kind of small molecular white matter which has a wide range of biological activities synthesized and secreted by immune cells (monocytes, macrophages, T cells, B cells, NK cells,etc.) and some non immune cells (endothelial cells, epidermal cells,fibroblasts, etc.) through stimulation. Cytokines bind to cell surface receptors and further initiate complex interactions among molecules in cells by regulating transcription. Rheumatoid arthritis is a kind of autoimmune inflammatory disease, and many cytokines participate in the process of its related diseases. The results show that RA plays an important role in the development of inflammation by activating the expression of inflammatory related genes. Mitogen-activated protein kinase (MAPK) is one of the important signaling proteins in organism. It participates in the process of regulating transcription by binding of cytokines and receptors, and plays an important role in regulating biological growth, differentiation, stress adaptation and inflammatory response. [21]. Tumor necrosis factor-α (Tumornecrosis factor-α, TNF-α), Interleukin-6 (IL-6), IL-1β are important cytokines involved in inflammatory and immune responses, which can stimulate the inflammatory response of synovium and further cause abnormal proliferation of synovium [22]. The monoclonal antibodies of TNF-α, IL-6, IL-1β were used in the treatment of RA and juvenile idiopathic arthritis, and showed good clinical effect.IL-2 and IL-4 are the characteristic cytokines secreted by Th1 and Th2 cells respectively, which play an important role in regulating the immune system [22]. The results showed that low dose IL-2 can play a therapeutic role in RA by regulating TH17/Treg balance. PPI results suggest that cytokines TNF-α, IL-1β, IL-2, IL-4 and IL-6 are the key targets of TFRD in RA intervention. MAPK1, MAPK3,MAPK8, MAPK14 and STAT3 play an important role in the process of total flavonoids intervention in RA, STAT3 is the most important target of regulation. Caspase3 and VEGF suggested that TFRD might be involved in the apoptosis of RA and the process of synovial neovascularization [25,26].

The results of KEGG enrichment analysis showed that PI3K/AKT pathway, IL-17 pathway and TNF pathway-α, T-cell receptor pathway and Th17 cell differentiation pathway may play an important role in the treatment of rheumatoid arthritis with total flavonoids of RA. 35 TFRD-RA common genes participate in the regulation of PI3K/AKT pathway, and abnormal activation of PI3K/AKT can activate NF-κB. And then promote TNF-α, IL-17 and IL-2 cytokines are transcribed, which activate IL-17 pathway and TNF-α pathway, T-cell receptor pathway and Th17 cell differentiation regulation pathway. It is suggested that PI3K/AKT pathway may be the most important regulatory pathway of total flavonoids in the treatment of RA. In the further molecular docking experiment, luteolin and naringin, the most important components of RA treated by total flavonoids, were used as the key target for molecular docking with the key proteins STAT3 in PPI network and PI3K, AKT1 in PI3K/AKT pathway. The results of molecular docking showed that luteolin could be combined with these targets better, suggesting that luteolin might regulate the phosphorylation process of STAT3, PI3K, AKT and so on by changing the molecular structure, and further affect its downstream function.

In conclusion, TFRD may inhibit the inflammatory response of RA by regulating inflammatory cytokines and related conduction pathways, and PI3K/AKT pathway may be an important way for the treatment of RA. Luteolin and naringenin are the key molecules in the treatment of RA by TFRD. It may play a downstream role by inhibiting the conformation of related key proteins and then regulating phosphorylation. In the future, experimental research will provide relevant basis for the treatment of RA and provide the basis for clinical application.

Author’s contribution:

Liu Xiao-yu: Design and implement the research plan, collect and analyze data, draw charts, write the paper, and revise the final version. Chen Guang-yao: proposed research ideas, collected and analyzed data, wrote the paper, and revised the final version. Wu Ziyu, Jin Qi: Analyzing data, drawing charts, and participating in paper revision. Tao Qing-wen: supervision and guidance, the final version of the review.

All authors declare no conflict of interest.