Study on the Multitarget Mechanism of Sanmiao Pill on Gouty Arthritis Based on Network Pharmacology

Sanmiao pill (SMP), a Chinese traditional formula, had been used to treat gouty arthritis (GA). However, the active compounds and underlying mechanism remained unclear. Hence, network pharmacology and molecular docking were utilized to explore bioactive compounds and potential mechanism of action of SMP in treating GA. In the study, the compounds of SMP, corresponding targets, and GA-related targets were mined from various pharmacological databases. Then, herb-compound-target, compound-target, PPI, and target-pathway networks were constructed. Ultimately, molecular docking was carried out to verify the predicted results. The results indicated that 47 active compounds, 338 targets, and 144 disease targets were collected. Network analysis implied that Phellodendron chinense Schneid. played a vital role in the whole formula. Moreover, 7 compounds (quercetin, kaempferol, wogonin, rutaecarpine, baicalein, beta-sitosterol, and stigmasterol) and 4 targets (NFKB1, RELA, MAPK1, and TNF) might be the kernel compounds and targets of SMP against GA. According to GOBP and KEGG pathway enrichment analysis and target-pathway network, SMP might exert a therapeutic role in GA by regulating numerous biological processes and pathways, including lipopolysaccharide-mediated signaling pathway, positive regulation of transcription, Toll-like receptor signaling pathway, JAK-STAT signaling pathway, NOD-like receptor signaling pathway, and MAPK signaling pathway. The results of molecular docking showcased that 11 pairs of compound with targets had tight binding strength. Thereinto, 4 compounds of MAPK1 and 5 compounds of NFKB1 possessed a better combination, suggesting that MAPK1 and NFKB1 might be considered as therapeutic targets in treatment of GA. This study verified that SMP had synergistic effect on GA by multicomponents, multitargets, and multipathways.


Introduction
Gouty arthritis (GA) was a common inflammatory arthritis, which affected at least 1% of the population in developed countries and increased obviously over the last few decades [1]. Due to reduction of uric acid excretion occur and/or purine metabolism disorders, the deposition of monosodium urate (MSU) crystals in synovial fluid and other tissues in the presence of elevated urate concentrations gave rise to GA [2]. e main symptoms of GA included severe pain, swelling, and tenderness resulting in difficulty moving the affected joint(s) [3], which significantly lowered the quality of life and caused a substantial economic burden [4].
In addition, GA could also bring in a series of comorbidities including obesity, hypertension, type 2 diabetes mellitus, coronary heart disease, metabolic syndrome, and renal disease [5,6]. e remedies for GA were mainly nonsteroidal anti-inflammatory drugs, colchicine, or glucocorticoids which led to serious side reactions [1,7]. Accordingly, hunting for therapeutic drugs in treatment of GA was urgently required.
Recently, traditional Chinese medicine (TCM) had been regarded as alternative therapies or dietary supplements for treatment of GA. Sanmiao pill (SMP), a basic traditional prescription for treatment of GA since the Ming Dynasty, was composed of three TCMs: Phellodendron chinense Schneid. (Huangbo, HB), Achyranthes bidentata Blume. (Niuxi, NX), and Atractylodes lancea ( unb.) DC. (Cangzhu, CZ) [8]. One study had reported that Achyranthes bidentata Blume. accumulated berberine in the knee joint and enhanced the anti-inflammatory effect under its lowerguiding drug in SMP [9]. In addition, SMP intervened in the expression of matrix metalloproteinases (MMPs)-3 and aggrecanases (ADAMTSs)-4 and enhanced the expression of GA-reduced tissue inhibitors (metalloproteinases (TIMPs)-1 and -3), resulting in inhibition of cartilage matrix degradation effectively in GA [10]. Although recent pharmacological studies had demonstrated that SMP might exert protective effect on GA, the underlying molecular mechanism of SMP acting on GA had not been fully elucidated.
Network pharmacology was emerging discipline which integrated network biology and polypharmacology [11]. Recently, a growing number of network pharmacology had been applied to comprehensively and systematically elucidate active compounds and potential mechanisms of action of TCM formulas by integrating multiple networks between compounds, targets, diseases, and pathways [12]. It provided a new sight into evaluating the efficacy mechanisms of drugs. e study integrated network pharmacology and molecular docking to firstly uncover underlying mechanism of SMP in treatment of GA. e detailed flowchart in the study is depicted in Figure 1.

Identification of Bioactive SMP Compounds.
e compounds in SMP formula were harvested from the Traditional Chinese Medicine Systems Pharmacology Database (TCMSP, http://ibts.hkbu.edu.hk/LSP/tcmsp.php). en, all compounds in SMP were subjected to ADME screening, which was evaluated by the possibility of each compounds in SMP as drug and carried out by the TCMSP database. In the integrative ADME model, oral bioavailability (OB) and drug-likeness (DL) were significantly pharmacokinetic parameters and regarded as screening criteria. erefore, active compounds were remained which were satisfactory with criteria of OB ≥ 30% and DL ≥ 0.18 for further study [13]. e simplified molecular-input line-entry specification (SMILES) of active compounds was ultimately downloaded from PubChem (https://pubchem.ncbi.nlm.nih.gov/) and OSRA (https://cactus.nci.nih.gov/osra/) databases.

Prediction
Targets of SMP. SMP targets were retrieved from three public online databases, namely, TCMSP, Swiss TargetPrediction, and DrugBank databases. ereinto, Swiss TargetPrediction and DrugBank databases predicted targets per active compound based on the similarity of chemical structure. erefore, in order to ensure the credibility of the targets, the targets with probability ≥0.5 were retained from the Swiss TargetPrediction database. Meanwhile, the targets which met the criteria of similar drugs approved by FDA with similarity threshold ≥0.7 were selected from the DrugBank database. In addition, the corresponding targets of active ingredients were supplemented through the TCMSP database. Taken together, the SMP targets were imported from UniProt (http://www.uniprot.org/) by restricting the species to "Homo sapiens" to standardize the gene name for the following data processing.

Construction of Herb-Compound-Target
Network. An herb-compound-target network was established by Cytoscape (3.6.0) software, so as to explore the relationship between herbs in SMP formula, compounds, and corresponding targets. Degree, one of topological parameters, which was calculated by network analysis, was applied to preliminarily identify the mainstream herb, compound, and target.
e compound-related targets and GA-related targets were, respectively, put into the Venn software which was from the OmicShare platform (https:// www.omicshare.com/). Ultimately, we could acquire SMPrelated targets on GA for the following study.

Construction of Compound-Target Network.
To accurately elicit interactions between active compounds and its corresponding GA-related targets of SMP, we constructed a compound-target network. Furthermore, topological parameters were applied eventually to screen key active compounds and GA-related targets of SMP.

Construction of Protein-Protein Interaction (PPI)
Network. GA-related targets of SMP were then processed by the String database (https://string-db.org/cgi/input.pl) to acquire protein-protein interactions (PPIs). "Organism" was set option to "Homo sapiens," and only PPI with a combined score more than 0.4 was selected in the specific operation process. e PPI network of GA-related targets of SMP was built through Cytoscape (3.6.0) software. As previously reported [14], we selected the targets whose degree value was greater than the average of all the other nodes in the network as hub targets of SMP-related GA.

GO and KEGG Pathway
Analysis. In order to systematically analyze Gene Ontology (GOBP) and the KEGG pathway of SMP targets related to GA, we imported hub targets into DAVID Bioinformatics Resources 6.8 (https:// david.ncifcrf.gov/). e GOBP terms and KEGG pathways with P ≤ 0.05 were screened for further analysis.

Construction of Target-Pathway Network.
e targetpathway network was built by the linkage with hub targets and KEGG pathways so as to detect the mechanism of SMP acting on GA. Subsequently, we selected the pathway with degree more than or equal to median value and associated with GA as a major pathway for further study.

Molecular Docking Verification.
To validate the significant compounds and targets which were selected from analysis of compound-target and target-pathway networks, the AutoDock tool was applied to perform a molecular docking. First of all, we prepared 3D structure of targets and compounds which were mined from PDB (https://www. rcsb.org/) and PubChem (https://pubchem.ncbi.nlm.nih. gov/) databases before performing the docking progress, respectively. en, the target proteins were imported into the AutoDock tool for a series of processes including removing H 2 O, adding hydrogen atoms, and calculating charge and atom type. e grid box was the docking area of the protein and compound which covered as much the protein and ligand as possible. Molecular docking employed the Lamarckian genetic algorithm (LGA). Meanwhile, we ignored all bond rotations of ligand. Ultimately, the docking binding energy of every compound-protein was calculated via the AutoDock tool to assess binding interactions.

Selection of Compounds in Herbal Medicines
Using ADME Evaluation. 364 compounds of the whole formula were mined from the TCMSP database in total, comprising 49 in CZ, 139 in HB, and 176 in NX. Among which, 19 compounds were excluded owing to duplicate (Supplementary Table 1). 56 active compounds of SMP were initially selected which met screening criteria of OB ≥ 30% and DL ≥ 0.18. However, owing to 9 bioactive compounds had no corresponding targets for the subsequent putative targets and thus excluded, we ultimately obtained 47 active compounds (Table 1).

Herb-Compound-Target Network.
e herb-compoundtarget network was built so as to preliminarily understand the interactions between herbs, compounds, and targets. e network embodied 388 nodes (3 herb nodes, 47 compound nodes, and 338 compound target nodes) and 1255 edges ( Figure 2). Degree, one of topological parameters, which referred to the number of nodes connected directly with it, was applied to assess the importance of herbs and compounds in the network. Degree values were calculated by network analysis (Supplementary Table 3). e degree values of HB, CZ, and NX were, respectively, 29, 8, and 9, suggesting that one herb was capable to correlated closely numerous active compounds. Meanwhile, HB displayed the highest degree value, indicating that HB potentially played a crucial role in SMP. A previous study had vindicated the anti-inflammation effect of HB in various animal inflammatory models by adjusting release of nitric oxide (NO), inducing nitric oxide synthase (iNOS) production, attenuating the release of tumor necrosis factor-α (TNF-α) and interleukin 1β (IL-1β) [15]. Moreover, 47 active compounds had at least linkage with two targets, showcasing that a multitude of targets portioned mutual compounds. Besides, the mean degree value of targets was 3.55, respectively, signifying that each targets correlated with an average of 3.55 compounds, which further confirmed that a compound Evidence-Based Complementary and Alternative Medicine could not only affect one target, but it also could act on multiple targets and exert potential synergistic effects.

SMP Targets Associated with GA.
We obtained a total of 144 disease targets from the two pharmacology databases, including 138 targets in the GeneCards database, 37 targets in the DisGeNet database, and 31 redundancies which were excluded (Supplementary Table 4). Figure 3 displays that 33 different GA-related targets were acquired by matching 144 disease targets to 338 compound targets.

Compound-Target Network.
To disclose the intimate interactions of compounds and GA-related targets of SMP, the compound-target (C-T) network was constructed. C-T network embodied 71 nodes (38 compounds and 33 SMP targets related to GA) and 130 edges ( Figure 4). Based on the network analysis, average degree value of each compound was 3.42 (Supplementary Table 5). Nodes with degrees only greater than the average degree could be significant compounds. Quercetin (degree = 27), kaempferol (degree = 14), wogonin (degree = 10), rutaecarpine (degree = 6), baicalein (degree = 5), beta-sitosterol (degree = 4), and stigmasterol (degree = 4) were  [16]. Kaempferol exhibited therapeutic effect on rheumatoid arthritis and intervened the migration and invasion of fibroblast-like synoviocytes by obstructing MAPK pathway activation [17]. Wogonin could restrain effectively extrinsic and intrinsic apoptotic pathways and induce antiapoptotic proteins in treatment of osteoarthritis chondrocytes [18]. Additionally, the average degree value of per target was 3.94. 10 targets (PTGS2, PTGS1, VDR, NOS2, F2, HTR3A, TNF, MMP9, PPARG, and RELA) were regulated by more than or equal to 4 bioactive compounds, which might play a critical role in treating GA.

PPI Network.
To shed light on the hub GA-relevant targets of SMP for the following GO and KEGG pathway enrichment, the PPI network was visualized by Cytoscape (3.6.0) software with 32 nodes and 264 edges based on the protein-protein interactions from the String database ( Figure 5). e degree value of each target was calculated. e targets with degree value greater than average (16.5) were designated as the hub GA-associated targets of SMP (Supplementary Table 6). We eventually selected 18 hub targets, which were potential therapeutic effects on GA, for instance, tumor necrosis factor (TNF), interleukin-6 (IL6), interleukin-8 (CXCL8), C-C motif chemokine 2 (CCL2), interleukin-1 beta (IL1B), tyrosine-protein kinase SRC (SRC), prostaglandin-endoperoxide synthase 2 (PTGS2), mitogen-activated protein kinase 1 (MAPK1), peroxisome NEK6

33 305
Disease target Compound target It was reported that the pathogenesis of GA was closely related to the NLRP3 inflammasome and inflammatory cytokines [19]. MSU crystals could be recognized by pattern recognition receptors and bind to NALP3 inflammasome. In turn, that leaded to the increasing number of mononuclear macrophages to engulf MSU crystals. Eventually, histamines and chemokines were released, resulting in inflammatory reactions such as local vasodilation and leukocyte accumulation. In addition, inflammatory factors were produced and secreted, such as TNF, IL1B, IL6, CXCL8, and CCL2, which aggravated the inflammatory reaction of MSU crystal deposition area and further amplified the inflammatory reaction of GA [20]. MAPK1 and MAPK14, the serinethreonine protein kinase family, modulated multiple cellular activities, such as cell proliferation, apoptosis, inflammation, and innate immunity. e MAPK signaling pathway was significant regulators of inflammatory cytokine which was activated by MAPK phosphorylation [21]. Oxidative stress played an important role in NLRP3 inflammasome activation. When oxidative stress was suppressed, the generation and secretion of IL-1β induced by NLRP3 activation would be blocked, whereas HMOX1 modulated the balance of  Evidence-Based Complementary and Alternative Medicine redox signaling. Consequently, HMOX1 also played an indirect role in the development of GA [22]. MMP2 and MMP9 were the members of the matrix metalloproteinase (MMP) family, which increased cartilage degradation, further resulting in the development of GA [23]. SERPINE1, one of the serine protease inhibitors, played a pathophysiological role in the synovial fluid of gouty arthritis [24]. CRP, an acute reactant protein, had a primitive defense function against the exogenous stimulus, which was significant part of inflammation [25]. COX2 was encoded by the PTGS2 gene, which participated in pathological processes such as inflammation and fever [26]. RELA and NFKB1, the transcription factor protein family, were the central mediator of the inflammatory process [27]. In addition, SRC, PPARG, and TGFB1 also regulated the inflammatory response [28][29][30].

GO and KEGG Pathway Analysis.
To clarify the molecular mechanisms of GA-related targets of SMP, GOBP and KEGG pathway enrichment of 18 hub GA-related targets was performed. A total of 44 GOBP terms were enriched with P ≤ 0.05 (Supplementary Table 7). e top significant GOBP terms are shown in Figure 6. e results indicated that GA-related targets of SMP could participate in a majority of biological process in treatment of GA, such as lipopolysaccharide-mediated signaling pathway, positive regulation of transcription, DNA-templated, positive regulation of NF-kappa B transcription factor activity, cellular response to interleukin-1, positive regulation of transcription from RNA polymerase II promoter, protein kinase B signaling, positive regulation of gene expression, positive regulation of apoptotic process, positive regulation of fever generation, and response to drug.
According to KEGG pathway enrichment analysis, 17 of 18 SMP targets related to GA enriched 57 pathways with P ≤ 0.05 (Supplementary Table 8), which were divided into six categories: immunity, inflammation, endocrine system, nervous system, disease and cell proliferation, and differentiation and migration. ereinto, 32 disease-related pathways not only included infectious diseases such as Chagas disease, influenza A, hepatitis B, and malaria but also included inflammatory processes of chronic diseases such as rheumatoid arthritis, insulin resistance, and cancer. In addition, 2 inflammation-related pathways were cytokine-cytokine receptor interaction and adipocytokine signaling pathway. 7 immunity-related pathways included the T-cell receptor signaling pathway, cytosolic DNA-sensing pathway, hematopoietic cell lineage, etc.. 8 pathways associated with inflammation and immunity were VEGF signaling pathway, NF-kappa B signaling pathway, TNF signaling pathway, NOD-like receptor signaling pathway, Toll-like receptor signaling pathway, etc. Ultimately, pathways associated with endocrine system, nervous system, and cell proliferation, differentiation, and migration were, respectively, 3, 1, and 4. is suggested that SMP could treat GA by acting on multiple pathways.

Construction of Target-Pathway Network.
To systematically and comprehensively disclose interactions of GA-related targets and pathways, target-pathway network was built by Cytoscape (3.6.0) software. e target-pathway network contained 74 nodes (17 hub targets and 57 pathways) and 311 edges (Figure 7). After network analysis, the topological parameters were calculated (Supplementary Table 9). Based on the degree values of pathways, 27 pathawys were identified as major pathways with degree more than the mean degree (5.46). Notably, the top 6 signaling pathways were Chagas disease (American trypanosomiasis) (hsa05142), TNF signaling pathway (hsa04668), NOD-like receptor signaling pathway (hsa04621), tuberculosis (hsa05152), hepatitis B (hsa05161), and influenza A (hsa05164).
Multiple studies focused that the approaches of treating GA lowered levels of uric acid and reduced inflammation [31]. Although Chagas disease (American trypanosomiasis), tuberculosis, hepatitis B, and influenza A were pathways associated with infection disease, the pathways had complex interactions. e four pathways related to infection disease had Evidence-Based Complementary and Alternative Medicine common subpathways, namely, the Toll-like receptor signaling pathway. e Toll-like receptor signaling pathway was implicated in MSU crystal-induced inflammatory cytokine release in monocytes/macrophages [32]. In addition, the JAK-STAT signaling pathway, NOD-like receptor signaling pathway, and MAPK signaling pathway in four pathways related to disease arose in high frequency. It was reported that MSU crystals could directly or indirectly activate the Toll-like receptor (TLRs) which recruited myeloid differentiation factor 88 (MyD88) in turn. en, interleukin-1 receptor-related kinase (IRAK) and transformed growth factor-beta activated kinase (TAK1) were activated. ereby, the IkB kinase cascade was triggered and NF-κB was activated. Meanwhile, the MAPK signaling pathway was activated to increase the expression of proinflammatory molecules, for instance, tumor necrosis factor (TNF)-α, IL-6, and IL-8, resulting in inflammatory reaction. From the above analysis, SMP in treatment of GA might inhibit the occurrence of inflammatory reaction via the Toll-like receptor/NF-κB/MAPK pathway [33]. In addition, recent studies had shown that the JAK-STAT signaling pathway [34], NODlike receptor signaling pathway [35,36], and TNF signaling pathway [37] were activated, causing secretion of proinflammatory and autoimmunity in the occurrence and development of GA. Pathogenesis of GA is associated with inflammation, immunity, and metabolism. Accordingly, SMP acted on anti-inflammation and immunity in treatment of GA via the Toll-like receptor signaling pathway, JAK-STAT signaling pathway, NOD-like receptor signaling pathway, MAPK signaling pathway, and so on. ese pathways associated with GA-relevant targets of SMP might become multiple inflammatory and immune-related pathways in the development of GA. Besides, we also found that some key targets acted on multiple pathways (e.g., NFKB1, RELA, MAPK1, and TNF acted on more than or equal to 35 pathways). NFKB1 and RELA, members of the NF-κB family, were involved in the NF-κB signaling pathway and TLR4-NFκB-IL1b signaling pathway, respectively, which regulated the expressing pro-inflammatory cytokines and played a central role in the pathogenesis of acute inflammation in arthritis [38][39][40]. MAPK1 (also known as ERK2) and TNF were proinflammatory cytokines which contributed to inflammation of acute GA [41][42][43]. Taken together, NFKB1, RELA, MAPK1, and TNF might be considered as potential therapeutic targets in treating GA, but they also needed molecular docking for further verification.  Figure 7: Target-pathway network. e blue ellipses display pathways. e purple diamonds display SMP targets related to GA. baicalein, beta-sitosterol, and stigmasterol) and targets (NFKB1, RELA, MAPK1, and TNF) were obtained from the result of compound-target and target-pathway network analysis for the binding validation experiment. e binding energy was applied to assess the effects of compound-target interactions. e binding energy was lower, indicating that the molecular docking was more effective. In the molecular docking process, the compound-target interaction with docking energy less than 5 kcal/mol would deem the effective docking [44]. e result of molecular docking is shown in Supplementary  MAPK1-rutaecarpine MAPK1-beta-sitosterol MAPK1-stigmasterol NFKB1-stigmasterol NFKB1-beta-sitosterol NFKB1-rutaecarpine e heat map of molecular docking is portrayed in Figure 8. From the above results of molecular docking, we concluded that 5 compounds (beta-sitosterol, rutaecarpine, stigmasterol, kaempferol, and wogonin) possessed a better combination with NFKB1. Additionally, 4 compounds (rutaecarpine, stigmasterol,quercetin, and beta-sitosterol) had a close interaction with MAPK1, suggesting that NFKB1 and MAPK1 might be considered as therapeutic targets in treatment of GA. Partial molecular docking patterns are depicted in Figure 9.

Conclusions
In the present study, network pharmacology was proposed to dissect the molecular mechanism of SMP in treatment of GA. Our study found that HB potentially played a crucial role in the whole SMP formula. We also successfully identified that quercetin, kaempferol, wogonin, rutaecarpine, baicalein, beta-sitosterol, and stigmasterol could exert synergistic or antagonistic effects in treatment of GA through acting on multiple targets. Moreover, SMP took up significant position in anti-inflammatory by regulating multiple pathways. In addition, the results of molecular docking displayed the reliability and accuracy of the prediction results. In summary, the study had provided systematic and comprehensive sight into the mechanism of SMP in treating GA, which might contribute to the further application for SMP development.

Data Availability
e original data used to support the findings of this study are included within the supplementary information file.

Conflicts of Interest
e authors declare no conflicts of interest.

Supplementary Materials
Supplementary