Network Pharmacology- and Molecular Docking-Based Identification of Potential Phytocompounds from Argyreia capitiformis in the Treatment of Inflammation

The methanolic extract of Argyreia capitiformis stem was examined for anti-inflammatory activities following network pharmacology analysis and molecular docking study. Based on gas chromatography-mass spectrometry (GC-MS) analysis, 49 compounds were identified from the methanolic extract of A. capitiformis stem. A network pharmacology analysis was conducted against the identified compounds, and Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway analysis and Gene Ontology analysis of biological processes and molecular functions were performed. Six proteins (IL1R1, IRAK4, MYD88, TIRAP, TLR4, and TRAF6) were identified from the KEGG pathway analysis and subjected to molecular docking study. Additionally, six best ligand efficiency compounds and positive control (aspirin) from each protein were evaluated for their stability using the molecular dynamics simulation study. Our study suggested that IL1R1, IRAK4, MYD88, TIRAP, TLR4, and TRAF6 proteins may be targeted by compounds in the methanolic extract of A. capitiformis stem to provide anti-inflammatory effects.


Introduction
Inflammation describes a biological process that occurs in tissues to protect the host against harmful stimuli, such as microorganisms and abnormal or damaged cells. Inflammation stimulates the immune system and regulates protective responses via immune cells, blood vessels, and molecular biological agents [1,2]. Many chronic diseases, including cardiovascular and gastrointestinal illnesses, diabetes, rheumatism, and cancer, are associated with upregulated inflammation [3]. Chronic diseases represent a major human health concern according to the World Health Organization (WHO). e incidence of chronic inflammation-related disorders is expected to steadily increase in the United States (US) over the next 30 years. Approximately 125 million people in the US were diagnosed with chronic diseases in 2000, with 61 million (21%) having multiple conditions [4][5][6]. Typically, cellular and molecular mechanisms and interactions among various factors can efficiently limit the potential damage and prevent further infection during an acute inflammatory response, resulting in the eventual repair of cellular homeostasis following the resolution of acute inflammation. However, uncontrolled acute inflammation can develop into chronic inflammation, causing a range of chronic inflammatory diseases [7]. ree out of five people worldwide die from chronic inflammatory conditions, such as stroke, respiratory infections, cardiovascular diseases, cancers, diabetes, and obesity [4][5][6]. A growing interest in the use of medicinal plants has developed for the treatment and management of diseases in an effort to identify safer and more efficient anti-inflammatory agents for the prevention of inflammatory conditions rather than using synthetic anti-inflammatory drugs [8].
Argyreia capitiformis (Poir.) Ooststr. is a member of the Convolvulaceae family of the Argyreia genus, which is not toxic and has medicinal and ornamental uses [9,10]. Traditionally, a paste made from the leaves of A. capitiformis has been used as an effective treatment of bruising on the legs. A. capitiformis has also been used in traditional medicine as a purgative and to treat sexual debility and ear pain [10][11][12]. Several studies have been reported for the Argyreia species, including antioxidant, anti-inflammatory, immunomodulatory, and CNS activities with several bioactive compounds [13][14][15]. However, no such study has been evaluated for A. capitiformis except the recent study on the methanolic extract of A. capitiformis leaves that suppressed the nuclear factorkappa B (NF-κB) pathway and inhibited the lipopolysaccharide-induced production of nitric oxide and inducible nitric oxide synthase in RAW 264.7 cells, demonstrating anti-inflammatory activities [9]. However, the chemical compounds found in A. capitiformis that are responsible for these anti-inflammatory effects and the underlying mechanisms remain unknown. Further research is necessary to identify the potential lead compounds responsible for these biological anti-inflammatory outcomes.
Network pharmacology has become a widely accessible analysis method following the increased availability of biomedical data sets during the postgenomic period, supporting the growth of the fields of systems biology and polypharmacology [16]. Complex compound-gene and compound-protein interactions can be evaluated systematically to develop a prototype for efficient therapy. New therapeutic mechanisms may be discovered by network pharmacology analysis, which is oriented toward a "multigoals, multi-disease" paradigm rather than "one target, one drug" [17][18][19]. Network pharmacology represents an effective method for selecting and elucidating the synergistic effects among bioactive chemicals through the mechanistic exploration of effects on multiple disease pathways [19,20]. Additionally, spectrometric and chromatographic technologies used in the initial evaluation of medicinal plants provide valuable information on bioactive activities that aid in the selection of biologically active species. Alkaloids, phenolic compounds, organic acids, esters, and amino acids are among the chemicals that GC-MS can detect quickly and accurately. us, in this investigation, GC-MS was used to detect and identify phytochemical constituents in A. capitiformis [21][22][23][24]. e network pharmacology approach connects targeted genes with the effects of bioactive compounds; thus, the present study was designed to elucidate the anti-inflammatory effects of the methanolic extract of A. capitiformis stem using a network pharmacology approach. Bioactive compounds in the methanolic extract of A. capitiformis stem were identified for this study using gas chromatographymass spectrometry (GC-MS) analysis, followed by a molecular docking assay to investigate potential ligand-receptor interactions, including the assessment of binding affinity and stability.

Plant Extraction.
e stems of A. capitiformis were collected from the Sitakunda Eco-park, Chittagong, Bangladesh, in March 2020 and later identified by a taxonomist. e stems were subjected to air-drying and ground to a coarse powder. e powder (200 g) was soaked in methanol (1 L) for 7 days [25,26]. Subsequently, the extract was filtered through Whatman filter paper and evaporated at 45°C. After the evaporation, 2.67 g of the black methanol extract yield (1.34%) was collected in an amber glass vial and refrigerated at 4°C until further use.

GC-MS
Analysis. An Agilent GC 7890A (Agilent Technologies Inc., Wilmington, DE, USA), combined with a triple-axis detector 5975 C single-quadrupole mass spectrometer, was used for GC-MS analysis. e chromatographic column was an Agilent HP 5MS column (30 m × 0.25 mm × 0.25 µm film thickness), using high-purity helium as the gas carrier at a flow rate of 1 mL per min. e injector temperature was 230°C, and the sample was injected using a splitless injector at 20 : 1. e temperature was set primarily to 40°C (held for 1 min), raised to 150°C at a rate of 5°C per min (held for 2 min), before being increased to 300°C at a rate of 5°C per min (held for 10 min). e temperature of the MS ion source was set to 150°C, and the temperature of the inlet line was set to 280°C. e scan range was set between 50 and 550 mass, with 70 eV electron energy and a 4-min solvent delay. Finally, by comparing the spectra against the NIST 2008 database (National Institute of Standard and Technology library), tentative compounds were identified. e total analysis time required for the sample was 65 min [27].

Network Pharmacology.
e network pharmacology analysis was performed using the STITCH platform (http:// stitch.embl.de/) to identify putative associations between the identified compounds and target genes. Multiple compound targets were identified using the Homo sapiens genome [27][28][29]. Multiple functional nodes and edges were identified in the network. Kyoto Encyclopedia of Genes and Genomes (KEGG) analysis was performed on the components identified in the network to obtain a biological interpretation of 2 Evidence-Based Complementary and Alternative Medicine the vast list of potential targets and to identify potential antiinflammatory pathways that are targeted. e STRING (search tool for retrieval of interacting genes) database (https://string-db.org), which includes predicted protein-protein interactions (PPIs), was used to predict functional protein interactions [30].

Ligand Preparation.
We selected 47 compounds identified from A. capitiformis, according to the qualitative GC-MS analysis, which we submitted to the molecular docking study. e selected compounds were retrieved in .sdf format from the PubChem database. In addition, aspirin (CID: 2244) was utilized in this study as a positive antiinflammatory control. e three-dimensional (3D) structures of the selected compounds were constructed in Schrödinger using LigPrep (Maestro v11.1), utilizing the force field OPLS3.

Grid Generation and Molecular Docking.
To create receptor grids and execute a molecular docking analysis, Glide (Schrödinger, Maestro v11.1) was used. e grids were generated in Glide with the default settings and the OPLS3 force field. A cubic box with a boundary box (14Å × 14Å × 14Å) was specified for the receptors. All docking studies were conducted using Glide's standard precision (SP) and flexible docking modes, and the lowest docking score for each ligand was recorded.

MM-GBSA and Ligand Efficiency Analysis.
e free energies of binding (ΔG; kcal/mol) for each ligand and the target receptors were calculated using the Schrödinger software package Prime/MM-GBSA module (OPLS3) [37,38]. e ligand efficiency (LE) was assessed for each ligand by obtaining the ratio of ΔG to the number of heavy atoms (NHA): LE � −(ΔG)/NHA [39].

Molecular Dynamics Simulation.
e molecular dynamics simulation study was conducted in YASARA dynamics by the aid of AMBER14 force field [40,41]. e docked complexes were optimized and cleaned, and hydrogen bond network system was oriented. e cubic simulation cell was created where the TIP3P solvation model was used with periodic boundary conditions [42]. e simulation system was neutralized at 310 K temperature, pH 7.4, and 0.9% NaCl. e initial energy minimization was conducted by steepest grained algorithms by simulating annealing methods. e long-range electrostatic interactions were calculated by the particle mesh Ewald (PME) methods with a cutoff radius of 8.0Å [43,44]. e simulation time step was set as 2.0 fs. e simulation trajectories were saved after 100 ps and finally run for 20 ns by following the constant pressure and Berendsen thermostat [45]. e simulation trajectories were used to calculate the root-meansquare deviation, solvent accessible surface area, radius of gyration, and hydrogen bond [46][47][48][49][50][51][52][53][54].

Molecular Docking and Simulation.
A total of 47 compounds docking results are presented in Tables 3-8, which show the findings. Aspirin has been used as a positive control for this study.
is study's findings reveal that binding energies of most of the ligands to receptors are negative, as later validated by MM-GBSA analysis. e proteins and ligands' molecular interaction is presented in Supplementary Materials (Tables S4-S9). To better understand the docking score, we have studied the ligand efficiency, which demonstrated excellent support for molecular docking scores. e molecular dynamics simulation of the targeted receptors (IL1R1, IRAK4, MYD88, TIRAP, TLR4, and TRAF6) against the best stable compounds is presented in Figures 1-6.

Discussion
Plants have been used throughout the history of traditional medicine to induce a variety of biological effects, and extensive pharmaceutical resources have recently been devoted to the identification and investigation of new remedies, including those derived from plants. A critical issue encountered by researchers who perform phytoscience is that a single plant can harbor a wide range of bioactive chemicals [55,56]. e pharmaceutical industry relies on phytochemicals to develop new drugs and therapeutic agents. Finding natural bioactive components is the first step in developing novel drugs. Screening plant extracts for therapeutically active chemicals is a novel strategy. It is important to know that plants have a lot of different types of phytochemicals, which have a lot of different biological properties.
Determining which compounds are responsible for the biological activities associated with plant materials can help understand toxicities, determine suitable doses, and identify ideal methods for compound extraction. e successful acquisition of components from plant materials depends primarily on the solvent used during the extraction process [60,61]. e stem methanolic extract was analyzed by GC-MS analysis, and the results indicated 49 different chemicals with varying retention times and peak areas.
e network pharmacology analysis was performed to evaluate potential interactions between the identified chemical compounds and proteins, followed by multiple comparisons to determine the number of genes responsible for the anti-inflammatory effects, which was represented by the STITCH platform ( Figure 7). Table 2 shows the results of KEGG pathway analysis performed on potential target Evidence-Based Complementary and Alternative Medicine genes, which identified signaling pathways associated with anti-inflammatory actions. Analyses of biological processes and molecular functions, which identified proteins and pathways with significant values, were also performed (Table S1). A comparison of the compound-gene interaction network and the protein-protein interaction network was performed to reveal biological and molecular functions (Table S2). From a molecular and functional perspective, these findings can assist in understanding the computational rules of compounds that have the potential to treat diseases. For example, protein kinase C (PKC) binding is known to treat inflammatory diseases [62], and the present results showed that binding with TIRAP, UGT1A10, and UGT1A7 was significant (p � 0.0447) and similar to that for PKC. Also, retinoic acid binding has a great impact on the immune system and the inflammatory response. Our findings identified the genes UGT1A1, UGT1A3, UGT1A7, UGT1A8, and UGT1A9 as being the most significant interactions for the compounds in our extract (p � 2.69E − 06).        Although a remarkable amount of functional and structural data has been compiled for each identified protein, our knowledge regarding protein-protein relationships remain scattered. e purpose of the STRING database is the collection, scoring, integration, and complementation with computational predictions for all public sources of PPIs [63][64][65][66]. In the present study, 10 proteins (IL1R1, IRAK4,  IRAK2, MYD88, TIRAP, TRAF6, TLR3, TLR4, TLR5, and TLR6) were used to analyze a PPI network ( Figure S2), which   downstream. TAK1 activates the IKKβ-to-IκB-α-to-NF-κB pathway, inducing the transcription of proinflammatory genes. TAK1 also influences gene expression by activating mitogen-activated protein kinase (MAPK) cascades. TRAF6, βRIP1, and TAK1 are activated by the binding of TLR with TRIF, resulting in the activation of MAPK, interferon regulatory factor 3 (IRF3), NF-κB, and interferon-transcription activation. e TRIF pathway also promotes the release of proinflammatory cytokines, although to a lesser degree than the MYD88 pathway. TLRs are a class of specific receptors that are key players in mounting an effective innate immune response to infection [67][68][69]. Subsequently, gene co-occurrence can be used to identify gene families whose patterns exhibit similarity across genomes. ree types of analyses have been used to examine genomes (neighborhood, fusion, gene co-occurrence) based on the systemic comparison of all-against-all genomes to evaluate the impacts of historical genome restructurings, genetic gains and losses, and gene fusion [70,71]. In the present study, 100% sequence conservation was observed for the ten proteins (IL1R1, IRAK4, IRAK2, MYD88, TIRAP,  TRAF6, TLR3, TLR4, TLR5, and TLR6) selected for the PPI network analysis, as shown in Figure S3. Additionally, gene coexpression was also studied in the present study, as summarized in Table S3 and Figure S4. e coexpression pathway is predicated by performing gene-by-gene correlation testing across many gene expression databases.   [63,72]. e present findings suggested co-expression among the ten selected proteins (IL1R1,  IRAK4, IRAK2, MYD88, TIRAP, TRAF6, TLR3, TLR4,  TLR5, and TLR6) in Homo sapiens. In addition, coexpression (transferred) of three more genes was observed in Gallus gallus, Mus musculus, and Rattus norvegicus. Structure-based drug discovery is gradually becoming a key technique for facilitating the rapid and cost-effective discovery and optimization of lead compounds. e use of a rational, structure-based drug design strategy is more efficient than conventional drug development techniques because this approach seeks to understand the molecular basis of diseases and incorporates information regarding the biological target's 3D structure during the drug design process [37]. A molecular docking study was incorporated into the present study to predict the complex structure formed by ligand-protein binding and analyze the ligand's conformational space within the protein-binding site. A score function  is then used for each docking analysis to assess the free energy of the interaction between the protein and ligand [37,73,74]. Additionally, the LE is calculated, which can enrich docking functions and allow for the coordination between docking outcomes and experimental results. Critical information regarding a molecule's properties, such as the NHA, can then be combined into a single table [39]. IL-1 controls a range of innate immune pathways, making it a key regulator of inflammation [75]. Two IL-1 cell surface receptors and a decoy receptor have been identified, including IL1R1 and IL1R2. First, IL-1 binds with IL1R1, inducing the formation of a heterodimer between IL1R1 and either IL-1RAcP or IL1R3, followed by IL-1 receptor-associated kinase (IRAK) and MyD88. e inflammatory response induced by IL1R1 occurs when IL1R1 binds with either the IL-1α or IL-1β ligands, whereas the T-lymphocytes, fibroblastic cells, epithelial cells, and endothelial cells have been indicated [76][77][78]. In the present study, selected six target proteins based on the network pharmacology analysis for the molecular docking study: IL1R1 (PDB: 1ITB), IRAK4 (PDB: 6EGA), and MYD88 (PDB: 4EO7) was used for the docking study. For IL1R1 (PDB: 1ITB), the majority of the 47 tested compounds demonstrated good docking scores, except for 12 compounds (Table 3). Among the 33 compounds with good docking scores, catechol (3.02), phenol (2.94), 4(1H)-pyrimidinone, 6-hydroxy-(2.79), hydroquinone (2.70), 3-furaldehyde (2.51), and sulcatone (2.51) showed the best ligand efficiencies, compared with the LE value of 1.63 for the positive control aspirin (Table S4 and Figure S5). e compounds with the best ligand efficiencies were found to interact with LYS-93 by H-bond (distances >4Å), and catechol interacted via two H-bonds. Similar findings were also observed for the positive control aspirin. According to an earlier study, the noncontiguous binding epitope containing LYS-93 was identified for IL-1β [79].
Finally, TRAF6 (PDB: 3HCT) was assessed against the 45 chemicals found in the methanolic extract of A. capitiformis stems, of which 34 were identified as interacting (Table 8) Table S9 and Figure S10. ARG-6 formed H-bond interactions with coniferol, 2-(2-hydroxy-2-phenylethyl)-3,5,6trimethylpyrazine, and aspirin, whereas hydrophobic interactions were identified for 2-(2-hydroxy-2-phenylethyl)-3,5,6-trimethylpyrazine and sulcatone. GLN-54 also formed H-bond interactions with multiple compounds, including 2-(2-hydroxy-2-phenylethyl)-3,5,6-trimethylpyrazine and 3-furaldehyde. e root-mean-square deviations from the C-alpha atoms from the docked complexes IL1R1 (PDB: 1ITB) proteins are illustrated in Figure 1. e complexes' RMSD values were calculated to find out the deviations among the simulation complexes and structural stability. Figure 1(a) demonstrates that the complexes had similar RMSD profiles and did not fluctuate much in the simulation trajectories. e RMSD profile of the complexes reached the steady state after 5 ns and maintained the structural stability till the last periods of the simulations, which defines the stability of the complexes. e SASA of the complexes was analyzed to find out the change in the surface area. e higher SASA represents the expansion of the surface area of the protein, whereas the lower SASA indicates the truncated nature of the complexes. Figure 1(b) shows that the complexes were in a stable state in SASA. e radius of gyration defines the labiality and mobility of the complexes, where Figure 1(c) indicates the lower deviations. e hydrogen bond patterning follows a similar stable profile (Figure 1(d)). e molecular dynamics simulation of IRAK4 (PDB: 6EGA) is presented in Figure 2. e RMSD of the complexes had a stable trend in the RMSD profile for all the complexes except catechol. e higher RMSD of these complexes defines the higher flexibility of these compounds in the simulating environments (Figure 2(a)). e SASA profile of the complexes was stable, did not fluctuate much, and had a steady trend in SASA (Figure 2(b)).
is SASA profile correlates with the stable and rigid profile of the complexes (Figure 2(c)). e radius of gyration and hydrogen bond  Evidence-Based Complementary and Alternative Medicine pattern systems were similar and did not change too much in the simulations (Figure 2(d)). e RMSD of the MYD88 (4EO7) protein complexes had a lower level of fluctuations, and lower deviations were observed across the compounds. All compounds had a lower RMSD than 2.5Å in whole simulation periods (Figure 3(a)). e SASA of the MYD88 (4EO7) complexes were stable, but the complex of trans-13docosenamide had a lower SASA than the other complexes.
is SASA profile of trans-13-docosenamide defines the MYD88 (4EO7) experienced the condensed conformation upon binding with the corresponding ligands (Figure 3(b)). e 2-(2-Hydroxy-2-phenylethyl)-3,5,6-trimethylpyrazine complexes had a higher Rg than other complexes, which defines the complexes' flexible nature than other complexes (Figure 3(c)). ese complexes also had a higher SASA in Figure 3(b), which depicts the changes in the confirmations than other complexes in simulated environments. e hydrogenbonding pattern of the complexes for MYD88 (4EO7) protein was found stable and did not change too much in simulations (Figure 3(d)).
e RMSD profile of TIRAP (PDB: 4FZ5) complexes is illustrated in Figure 4. All complexes from TIRAP (PDB: 4FZ5) protein had an initial rise of RMSD due to a higher degree of flexibility but stabilized subsequently after 5 ns times. e complex 4-ethylresorcinol had a higher RMSD than the other complexes, which might be responsible for the more remarkable conformational changes and the flexibility of the compounds (Figure 4(a)). e SASA of the complexes had a stable and similar trend for all the compounds. But the complex phenol had a lower SASA profile at the last phase of SASA, which defines the complexes' truncated nature in simulations (Figure 4(b)). e radius of gyration profile of the complexes had a lower trend, which correlates with the less flexibility of the complexes (Figure 4(c)). e hydrogen bond patterning of the complexes had a stable profile in Figure 4(d).
e molecular dynamics simulation study of the TLR4 (PDB: 3FXI) and complexes was done to analyze the structural deviation in the docked structure. e root-meansquare deviations of all complexes are illustrated in Figure 5(a). e RMSD value of the complexes initially followed the upper trend from the beginning. is might be happening due to the higher flexibility level. But all the complexes from TLR4 (PDB: 3FXI) had a stable profile after 10 ns times and followed a similar trend until the last periods, demonstrating structural stability. e SASA of the TLR4 (PDB: 3FXI) complexes had lowered the degree of the deviations from the beginning and followed lower fluctuations, which define no changes in the surface area of the complexes ( Figure 5(b)). e Rg and hydrogen bond patterns of the simulation systems were stable and did not change too much, which correlates with the structural stability (Figures 5(c) and 5(d)). e docked complexes from the TRAF6 (PDB: 3HCT) protein and their simulation descriptors are illustrated in Figure 6. e RMSD profile of the TRAF6 (PDB: 3HCT) complexes defines that the phenol and aspirin had a higher level of RMSD than other complexes, which correlates with the comparative higher degree of the deviations of the complexes. All complexes had reached a stable state after 5 ns of the simulation times. Moreover, the complexes exhibit RMSD lower than 2.5Å, which defines the complexes with a higher rigidity degree (Figure 6(a)). e SASA profile of the complexes had a lower deviation, as illustrated in Figure 6(b). e phenol had a lower SASA value than all the compounds, indicating the truncated nature of the protein complexes compared with the others. Moreover, the radius of gyration from Figure 6(c) demonstrates that the complexes had a lower degree of deviations, and no significant higher fluctuations were observed.
is Rg profile indicates the complexes had lower mobility and flexibility during the simulation times. e hydrogen bond pattern of the complexes was stable during the simulations (Figure 6(d)).

Conclusions
e network pharmacology analysis revealed key pathways involved in the anti-inflammatory activities induced by the chemical compounds found in the methanolic extract of A. capitiformis stem. Six inflammation pathways were obtained from the KEGG pathway analysis (IL1R1, IRAK4, MYD88, TIRAP, TLR4, and TRAF6), and molecular docking studies of these pathways revealed that the identified chemical compounds had strong binding affinities with these pathway components.
e current study discovered that 3-furaldehyde, phenol, catechol, and hydroquinone were effective anti-inflammatory compounds found within the methanolic extract of A. capitiformis and played an important part in the inflammation pathway by targeting these six proteins. Additional in vitro and in vivo experiments will be helpful to validate and optimize the findings of this study.

Abbreviations
GC-MS: Gas chromatography-mass spectrometry KEGG: Kyoto Encyclopedia of Genes and Genomes WHO: World Health Organization NF-κB: Nuclear factor-kappa B STRING: Search Tool for Retrieval of Interacting Genes PPIs: Protein-protein interactions IL: Interleukins PKC: Protein kinase C MAPK: Mitogen-activated protein kinase IRF: Interferon regulatory factor IRAK: Receptor-associated kinase TLR: Toll-like receptor 3D: ree-dimensional PDB: Protein Data Bank RMSD: Root-mean-square deviation of the complexes SASA: Solvent accessible surface area Rg: Radius of gyration.
Data Availability e data are available within the manuscript and also accessible from the corresponding author upon request.

Conflicts of Interest
e authors declare no conflicts of interest.