Network pharmacology and experimental validation to identify the potential mechanism of Hedyotis diffusa Willd against rheumatoid arthritis

Rheumatoid arthritis (RA) is a chronic, systemic, autoimmune disease that may lead to joint damage, deformity, and disability, if not treated effectively. Hedyotis diffusa Willd (HDW) and its main components have been widely used to treat a variety of tumors and inflammatory diseases. The present study utilized a network pharmacology approach, microarray data analysis and molecular docking to predict the key active ingredients and mechanisms of HDW against RA. Eleven active ingredients in HDW and 180 potential anti-RA targets were identified. The ingredients-targets-RA network showed that stigmasterol, beta-sitosterol, quercetin, kaempferol, and 2-methoxy-3-methyl-9,10-anthraquinone were key components for RA treatment. KEGG pathway results revealed that the 180 potential targets were inflammatory-related pathways with predominant enrichment of the AGE-RAGE, TNF, IL17, and PI3K-Akt signaling pathways. Screened through the PPI network and with Cytoscape software, RELA, TNF, IL6, TP53, MAPK1, AKT1, IL10, and ESR1 were identified as the hub targets in the HDW for RA treatment. Molecular docking was used to identify the binding of 5 key components and the 8 related-RA hub targets. Moreover, the results of network pharmacology were verified by vitro experiments. HDW inhibits cell proliferation in MH7A cells in a dose and time-dependent manner. RT-qPCR and WB results suggest that HDW may affect hub targets through PI3K/AKT signaling pathway, thereby exerting anti-RA effect. This study provides evidence for a clinical effect of HDW on RA and a research basis for further investigation into the active ingredients and mechanisms of HDW against RA.

www.nature.com/scientificreports/ acteristics (OB ≥ 30% and DL ≥ 0.18) and ADME information, 11 active components were selected from 142 ingredients of HDW. The TCMSP and Swiss Target Prediction databases were used to determine the pharmacological targets of the HDW components. Table 1 shows active components and the number of the corresponding potential targets of HDW. Detailed information of these components and targets is listed in Supplementary  Table 1. Eventually, 180 potential targets were identified (after removing duplicates) using the Uniprot database.
Identification of the potential targets of RA. Using "Rheumatoid Arthritis" as the search term, 42,192,141,174, and 623 disease-targets were obtained from the OMIM, DrugBank, TTD, GeneCards, and Dis-GeNET databases, respectively. Merging all results from the five databases and removing duplicates, 942 related-RA potential targets were finally collected.
Construction of the active components-common targets-RA network. Applying a Venn diagram, 85 common targets were found overlapped between HDW compound targets and RA-related targets ( Fig. 2A). We imported 11 active components and 85 common targets into Cytoscape 3.9.0 software to construct a components-targets-RA network. Among these, the active ingredients with the highest degree value were stigmasterol, β-sitosterol, quercetin, kaempferol and 2-methoxy-3-methyl-9,10-anthraquinone. However, poriferasterol and scopoletin were removed since they lacked common targets in the network. The active ingredients- www.nature.com/scientificreports/ targets-RA network is shown in Fig. 2B. Results indicate that 5 components may provide the key to successful treatment of RA.
GO and KEGG enrichment analysis. In total, GO analysis identified 1542 significantly enriched GO terms (P.adjusted < 0.01 adjusted with Benjamini-Hochberg), consisting mainly of 1459 biological processes, 18 cellular components, and 65 molecular functions. We screened the top 10 ranked GO terms shown in Fig. 3A. In the biological process (GO:BP) category, the top terms were involved in responses to lipopolysaccharides, molecules of bacterial origin, and reactive oxygen species metabolic processes. In the cellular component (GO:CC) category, the top terms included membrane rafts, membrane microdomains, and membrane regions. In the molecular function (GO:MF) category, the top terms consisted of nuclear receptor activity, transcription factor activity, and cytokine receptor binding. To further identify underlying signaling pathways, we analyzed KEGG  Network visualization and identification of hub targets. Next, we analyzed 85 potential therapeutic targets by using the STRING database to obtain a PPI network to explore the relationship between RA-related targets. The PPI relationship network, with a total of 85 nodes, 238 edges and an average node degree of 5.6 was generated with a confidence of 0.9 ( Supplementary Fig. 1). PPI network diagrams were imported into Cytoscape 3.9.0 software for visualization (Fig. 4A). We further identified the subnetwork and hub targets from the PPI network using the CytoNCA plug-in (Fig. 4B). As shown in Fig. 4C, a subnetwork was identified, including 8 nodes and 27 edges. Moreover, RELA, TNF, IL6, TP53, MAPK1, AKT1, IL10, and ESR1 were identified as the hub targets in the HDW for RA treatment (Supplementary Table 3).  www.nature.com/scientificreports/ Molecular docking. Candidate compounds stigmasterol, β-sitosterol, quercetin and kaempferol, and 2-methoxy-3-methyl-9,10-anthraquinone, are the top 5 (ranked by degree) in the compounds-targets-RA network. The hub targets, RELA, TNF, IL6, TP53, MAPK1, AKT1, IL10, and ESR1, play a significant role in the action of HDW against RA. Molecular docking of the 5 compounds and 8 hub genes revealed binding energies shown in Fig. 5. Five components of HDW exhibited strong binding to the 8 core targets with β-sitosterol showing the highest binding energy. these results imply that treatment with HDW may affect all the Figure targets in RA patients. The target proteins and the small molecules with strong binding affinity were visualized by PyMoL software (Fig. 6).
Cell experiments. Overproliferation of fibroblast-like synoviocytes (FLS) are important pathogenesis of RA. Therefore, we investigated the effect of HDW at different concentrations (0, 0.5, 1, 2 mg/mL) on the proliferation of MH7A cells after 48 h. CCK-8 assay showed that HDW inhibited the proliferation of MH7A cells in a dose-dependent (Fig. 7A). According to the results, we chose to perform subsequent experiments with the dose of 0.5, 1, and 2 mg/mL. The WB result showed that HDW could inhibit the PI3K/AKT pathway by reducing phosphorylation of AKT in MH7A cells (P < 0.05, Fig. 7B). Hub targets in our work were verified by RT-qPCR and results showed that RELA, TNF, and IL6 were up-regulated while IL10 was down-regulated (Fig. 7C). These results validated our network pharmacology analysis, suggesting that HDW can play a role in treating RA by regulating PI3K/AKT signaling pathway and RA-related targets.

Discussion
Given their potent anti-inflammatory, anti-fibroblastic, and immunomodulatory actions, HDW and its components have been widely used in animal models over the last several years as an intervention for RA 12,13,[19][20][21][22][23] . Unfortunately, the potential targets and molecular mechanism of HDW against RA are still inadequately understood. TCM network pharmacology emerging recently has become a flourishing field in TCM modern studies along with the rapid progress of bioinformatics. The study of TCM and biological network appeared for the first time in 2007 28 , and some applications of traditional medicine network pharmacology for herbs or herbal formulae in RA 29 . In the present network pharmacological analysis, a total of 142 compounds of HDW were identified from TCMSP and published literature, and 11 compounds were selected by TCMSP and ADME criteria screening. A total of 180 targets related to potential compounds and 942 targets associated with RA were identified, and 85 common target genes were obtained from the overlapping part of identified compounds and RA. The components-targets-RA network analysis visualized the interaction of multi-components and multitargets about HDW on RA. The compounds targets network analysis indicated that the 5 compounds, including stigmasterol, β-sitosterol, quercetin, kaempferol, and 2-methoxy-3-methyl-9,10-anthraquinone, were linked to ≥ 10 target genes, and the 8 target genes (RELA, TNF, IL6, TP53, MAPK1, AKT1, IL10, and ESR1) were core target genes in the network. GO enrichment analysis indicated that numerous targets are involved in response to lipopolysaccharides and molecules of bacterial origin in BP, are localized to membrane rafts and membrane microdomains in CC, and are associated with nuclear receptor and transcription factor activities in MF. KEGG pathway analysis indicated that numerous targets are associated with certain inflammatory events and cancer. Molecular docking showed that stigmasterol, β-sitosterol, quercetin, kaempferol, and 2-methoxy-3-methyl-9,10-anthraquinone have good binding activity with RELA, TNF, IL6, TP53, MAPK1, AKT1, IL10, and ESR1 targets. Finally, the molecular mechanisms of HDW predicted by network pharmacology approach against RA were validated by in vitro experiments. www.nature.com/scientificreports/ A potential components-targets-RA target network indicated that stigmasterol, β-sitosterol, quercetin, kaempferol, and 2-methoxy-3-methyl-9,10-anthraquinone, are likely to play vital roles in the process of RA treatment (Table 2). Indeed, apart from 2-methoxy-3-methyl-9,10-anthraquinone, all these components have previously been reported to exhibit potential antirheumatic therapeutic activity. For example, stigmasterol has been shown to protect CIA rats by suppressing proinflammatory mediators (TNF-α, IL-6, IL-1β, iNOS and COX-2) and increasing anti-inflammatory cytokine IL-10 30 . β-sitosterol exerts an inhibitory influence on synovial angiogenesis by suppressing endothelial cell proliferation and migration, thereby alleviating joint swelling and bone destruction in CIA mice 31 , and quercetin inhibits the release of proinflammatory cytokines (IL-6, TNF-α, IL-1β, IL-8, IL-13, IL-17) by activating SIRT1, thereby becoming a potential effector of RA 32,33 . Kaempferol inhibits the proliferation and migration of RA-FLSs and the release of activated T-cell-mediated inflammatory cytokines by suppressing fibroblast growth factor receptor 3-ribosomal S6 kinase 2 (FGFR3-RSK2) signaling 34 . Collectively, these active components exhibit antirheumatic activity by various mechanisms, including anti-inflammatory, immunoregulatory, and reduction of bone destruction. Notably, however, there have been few previous studies on the treatment of RA with stigmasterol and β-sitosterol.
KEGG pathway analysis indicated that the 85 therapeutic targets were enriched in viral infection and cancer, such as Kaposi's sarcoma-associated herpesvirus infection, human cytomegalovirus infection, human papillomavirus infection, and prostate cancer. The evidence that viral infection such as human cytomegalovirus infection 35 contributes to RA is strong, and that RA is associated with an increased risk of cancer 36 . Additionally, the KEGG analysis results also indicated that the AGE-RAGE, TNF, IL-17, and PI3K-Akt signaling pathways   www.nature.com/scientificreports/ may be critical in the network pharmacology. It has been shown that the PI3K/AKT signaling pathway, the key molecular mechanism of the occurrence and development of RA, affects synovial inflammation 37 , synovial angiogenesis 38 , chondrocyte proliferation, apoptosis and autophagy 39 , and migration and invasion of fibroblastlike synoviocytes in RA 40 . The WB result showed that HDW could inhibit the PI3K/AKT pathway by reducing phosphorylation of AKT in MH7A cells. Construction of the common target-related PPI network permitted screening of the top 8 core genes: RELA (transcription factor p65), TNF (tumor necrosis factor), IL6 (interleukin-6), TP53 (cellular tumor antigen p53), MAPK1 (mitogen-activated protein kinase 1), AKT1(RAC-α serine/threonine-protein kinase), IL10 (interleukin-10), and ESR1(estrogen receptor). RELA, ranking first with the highest connection in the PPI network, is a subunit of nuclear factor (NF)-κB and is regarded as an important member of the NF-κB pathway 41 . RELA participates in essential life activities, including cell proliferation, transformation, ap-optosis, inflammation and immune response and has been confirmed as an important component of the pathogenesis of RA 42 . TNF is an important therapeutic target for RA and TNF agents were the first molecular targeting drugs developed for the treatment of RA 43,44 . IL6 and IL 10 are critical for the development and progression of RA by affecting the inflammatory process, osteoclast-mediated bone resorption and pannus development 45,46 . RT-qPCR results showed that RELA, TNF, and IL6 were up-regulated while IL10 was down-regulated with HDW treatment. However, there were no significant differences in other four targets, including TP53, MAPK1, AKT1, and ESR1. Studies with TP53, MAPK1, AKT1, and ESR1 have focused on cancer, whereas those on RA have focused on deficiency. Of course, it is possible that the level of mRNA observed may not accurately reflect the level of protein present. As such, whether HDW can directly influence these hub genes requires further study. The antirheumatic effect of compounds such as quercetin and kaempferol is partially associated with these target genes. For example, quercetin, a natural substance, can induce mitochondrial apoptosis of RA-FLSs through the p53 mechanism 47 . These results therefore illustrate the range of interactions between the multi-components and multi-targets of HDW in the treatment of RA.
As noted above, RELA, TNF, IL6, TP53, MAPK1, AKT1, IL10, and ESR1 are the top 8 hub genes in the PPI network. Additionally, however, potential targets were found to be involved in multiple pathways in the KEGG enrichment analysis. This indicates that HDW may exert its anti-RA effects by the combined interaction of multipathways and multi-targets. Also, the molecular docking analysis was carried out to investigate the interaction of 5 compounds and 8 targets. For example, the absolute value of binding energy about β-sitosterol and ESR1 is the highest in each group, indicating that β-sitosterol has a higher binding affinity than other target genes. Although there have been relatively few studies with stigmasterol and RA, the docking results indicated that stigmasterol performed good binding activity with TP53, ESR1, and IL10. In brief, the higher the binding affinity of components and targets, the more likely HDW treatment of RA will be achieved by modulating several related targets. However, further experiments are needed to clarify active compounds and mechanisms of action analysis of HDW for RA.

Methods
Active components and potential targets of HDW. Active components and potential targets of HDW were selected from the Traditional Chinese Medicine Systems Pharmacology Database and Analysis Platform (TCMSP, https:// old. tcmsp-e. com/ tcmsp. php) 55 . As recommended by TCMSP, the compounds with OB ≥ 30% have good absorption and slow metabolism after oral administration, while the compounds with DL ≥ 0.18 were chemically suitable for drug development. We therefore used OB ≥ 30% and DL ≥ 0.18 as filtering for thresholds 56 . An exhaustive search for components was performed using PubMed (https:// pubmed. ncbi. nlm. nih. gov/) and CNKI (https:// www. cnki. net/). The active components of HDW were determined by SwissADME database (http:// www. swiss adme. ch/). Afterward, the targets (probability ≥ 0.9) corresponding to the components were screened from the SwissTargetPrediction database (http:// swiss targe tpred iction. ch/) 57 . Duplicates were deleted to obtain potential targets for further analysis.
Common targets and construction of the compound-target-RA network. Targets related to the active components of HDW were mapped to RA-related targets to identify ones common to both. These were considered as targets of HDW on treating RA. Following compilation of these, the "compound-target-RA" network was established using Cytoscape 3.9.0 software to determine the interaction relationships 63 and to further screen for key active components by calculating the extent of these.
Gene ontology and pathway enrichment analysis. Gene Ontology (GO) analysis and KEGG pathway analysis were performed using the R 3.6.3 and related R packages (clusterProfiler, Org.Hs.eg.db, and ggplot2) 64 . After adjusting the p-value using the Benjamini-Hochberg (BH) method, P.adjusted values < 0.01 were considered to be statistically significant.

Network visualization and identification of hub targets.
To further identify interactive relationships among the common target proteins, the protein list was mapped to the STRING database (https:// cn. string-db. org/) 65 . To ensure reliability, we used a cutoff ≥ 0.9 (high-confidence interaction score) to obtain the Western blotting (WB). MH7A cells were cultured in T75 cell culture flask. After 48 h of treatment, cells were lysed with 1 × RIPA lysis buffer containing 1% of PMSF and 1% of phosphotransferase inhibitor. Protein concentration was detected by the BCA Protein Quantitative Kit (Thermo Fisher Scientific, USA). Subsequently, proteins (25 μg) were denatured by employing heating followed by electrophoresis with 10% SDS-PAGE, and were transferred to polyvinylidene fluoride membranes (Millipore, USA). The membranes were sealed with 5% skimmed milk (or 5% BSA) for 2 h at 37 °C, and incubated at 4 °C overnight with the following primary antibodies: protein kinase B (AKT), phospho-Akt (p-AKT) and β-actin. After incubation, the membrane was further incubated with horseradish peroxidase (HRP)-conjugated secondary antibodies for 1.5 h at 37 °C. Lastly, specific bands were detected using enhanced chemiluminescence reagent (Thermo Fisher Scientific) and calculated by Image J software for the quantitative analysis of proteins.

Conclusions
In summary, this study explored the therapeutic effect and mechanism of HDW on RA through a strategy combining network pharmacology and in vivo experiment verification. Five active components and 8 core targets of HDW were finally screened. PI3K/AKT signaling pathways were enriched by KEGG analysis in addition to virus and cancer-related pathways. Molecular docking showed that the active ingredients successfully docked with related targets. Through experimental validation, we found that HDW may affect target genes RELA, TNF, IL6, and IL10 through multiple signaling pathways PI3K/AKT signaling pathways, thereby affecting the pathological process of RA and ultimately inhibiting the occurrence and development of RA. Our findings provide a theoretical basis for the use of HDW as a therapeutic for RA.