The immunomodulatory activity of Orthosiphon aristatus against atopic dermatitis: Evidence-based on network pharmacology and molecular simulations

Abstract

Objectives

To explore the potential activity of Orthosiphon aristatus (OA) against atopic dermatitis (AD).

Methods

Phytocompounds from OA were identified through chromatography analysis, then continued to target identification and functional annotation to explore the potential target of OA. Then, network pharmacology from annotated proteins determined protein targets for OA phytocompounds. Protein with highest rank according to the betweenness and closeness algorithm then continued to molecular docking and validated through molecular dynamics analysis.

Results

Chromatography data analysis revealed thirty-six compounds, predominantly classified as carboxylic acid, fatty acyls, and polyphenols. Upon identifying these compounds, network biology-based target identification revealed their potential bioactivity in modulating inflammation in AD. Tumour Necrosis Factor-alpha (TNF-α) and Prostaglandin G/H synthase 2 (PTGS2) emerged as the most probable targets based on hub centrality in the protein–protein interaction network. Later, molecular docking analyses highlighted sixteen compounds with good inhibitory activity against these two proteins. Notably, molecular dynamics simulation revealed that three compounds out of the previous sixteen potential compounds were more likely to act as the TNF-α and PTGS2 inhibitor as well as their native inhibitor. Those compounds are (1R,9R)-5-Cyclohexyl-11- (propylsulfonyl)-7,11- diazatricyclo[7.3.1.02,7]trideca- 2,4-dien-6-one, also known as ZINC8297940, as the best TNF-α inhibitor along with dl-Leucineamide and Benazol P as the potential inhibitor of PTGS2.

Conclusions

These findings suggest that OA may exert therapeutic effects against AD by controlling inflammation through TNF-α and PTGS2 signalling pathways.

Introduction

Atopic dermatitis (AD) is a chronic inflammatory skin disorder that affects millions of patients on a global scale.¹ AD is characterized by pruritic, erythematous, and eczematous skin lesions and is often associated with a compromised skin barrier function.² The pathogenesis of AD is complex and involves various factors, including genetic predisposition, immune dysregulation, and environmental triggers.³ Amidst these multifaceted causes, certain links between AD and autoimmune factors have emerged.4, 5, 6 Therefore, inflammation appears to be central to the development of AD, and plays a crucial role in initiating and perpetuating the disease process.⁷

Inflammation in AD is mediated by a complex interplay of pro-inflammatory cytokines, chemokines, and immune cells, leading to disruption of the skin barrier, the hyperproliferation of keratinocytes, and the infiltration of immune cells into the skin.^7,8 The standard treatment for AD includes topical corticosteroids, calcineurin inhibitors, and moisturiz-ers, although these treatments may have certain limitations, including potential side effects and preliminary efficacy.⁹ Therefore, there is a growing interest in identifying novel therapeutic approaches for AD that can effectively modulate inflammation and promote integrity of the skin barrier.¹⁰ Herbal medicines (HM) stand out among various potential substitutes as a monotherapy or adjuvant therapy for AD.¹¹ A diverse range of phytocompounds, including alkaloids, flavonoids, terpenes, glycosides, and others, have proved to exert excellent effects in modulating AD, particularly as immunomodulators.12, 13, 14 Hence, investiga-ting naturally occurring phytocompounds may provide good candidates as an alternative or complementary remedy against AD.

Orthosiphon aristatus (OA), commonly known as Java tea or Cat's whiskers, is a traditional medicinal plant that is widely used in Southeast Asia for its anti-inflammatory and antioxidant properties.¹⁵ OA has been reported to possess various pharmacological activities, including anti-inflammatory and anti-oxidative effects.¹⁶ Moreover, OA is also known to exert wound healing activity¹⁷ that helps to recover skin barrier function. Furthermore, clinical studies have also reported the safety of OA as one of the components of Chinese HM.¹⁸ Unfortunately, to our knowledge, there has yet to be a study investigating the potential of OA in modulating inflammation in AD and explaining the underlying mechanism of its bioactivity. Thus, this study aimed to explore the potential activity of OA against AD by regulating immunoactivity.

Materials and Methods

Material

Dried OA leaf powder was obtained from Balai Materia Media, Department of Health, Batu City, East Java, Indonesia (code number: 210826.KKL.F.004). Reagents for compound extraction and phytochemical identification were purchased from Merck.

Extraction of plant compounds

One hundred grams of dried OA leaves were mixed with absolute ethanol (1:10 w/v) and shaken in a rotary shaker for 30 min, followed by overnight maceration. The soluble fraction was then evaporated using a rotary evaporator at 90 °C to remove the solvent. The remaining solvent was evaporated using a heated incubator until a stable weight was obtained. The resulting extract was frozen for future use.

Phytochemical identification

Chromatography procedures for phytochemical screening were performed as described in a previous study using a Thermo Scientific Dionex Ultimate 3000 RSLCnano coupled with a Thermo Scientific Q Exactive Mass Spectrometer.¹⁹ In brief, the extract was dissolved in an aqueous solution prior to injection. The mobile phase consisted of 0.1% formic acid in water as solvent A and 0.1% formic acid in acetonitrile as solvent B, with a Hypersil GOLD aQ 50 mm 1.9 m particle size column as the stationary phase. The chromatography was run for 40 min with a 40 μL/min flow rate. The peaks were then analysed using Compound Discoverer and the mzCloud MS/MS library as the reference database. Compounds with a similarity score of ≥80% were selected for further analysis.

Target identification and biological network construction

Compounds identified by chromatography analysis were entered into SwissTargetPrediction to predict potential protein targets for each compound.²⁰ Targets with a probability score higher than one were considered as potential targets. Proteins related to atopic dermatitis were also identified in the GeneCard database²¹ using keywords such as “atopic dermatitis”, “atopic dermatitis inhibitor”, and “anti-atopic dermatitis”. The proteins obtained from SwissTargetPrediction and the GeneCard database were then analysed using Venny 2.1 to identify overlapping proteins. The overlapping proteins were annotated using the Database for Annotation, Visualization, and Integrated Discovery (DAVID) to analyse biological processes.²² The biological process with the highest number of targets was selected for protein–protein interaction (PPI) analysis.

The PPI analysis was performed using Cytoscape 3.9.0 with the STRING protein database at a confidence score of 0.8.^23,24 The Cytohubba plugin²⁵ was used to analyse essential hubs in the constructed protein network. The important hubs were determined based on betweenness and closeness scoring and ranking. The highest-ranking hub for each scoring and ranking function was selected as the target for molecular docking and molecular dynamics simulations.

Three-dimensional structure retrievals

The three-dimensional (3D) structure of compounds was obtained from the PubChem database. The 3D structure was built by Avogadro²⁶ according to OPSIN: Open Parser for Systematic IUPAC nomenclature for compounds not available in the PubChem database. The Compound ID (CID) and SMILES code are tabulated in Supplementary Table 1. The 3D structures of Tumour Necrosis Factor (TNF)-α and Prostaglandin G/H synthase 2 (PTGS2) were retrieved from the RCSB Protein Data Bank (PDB) with PDB identities 2AZ5 and 5IKR, respectively.^27,28

Molecular docking

Prior to the docking process, the 3D structures of compounds were energy minimized using OpenBabel²⁹ in the PyRx package.³⁰ Furthermore, TNF-α and PTGS2 structures were prepared by removing water and native ligand atoms using Discovery Studio 2019. The native ligand was chosen as the control molecule, or native inhibitor for each protein, and the structure was extracted from the PDB structure using PyMOL. The compounds were considered as flexible ligands, while the proteins were considered as rigid receptors³¹ during the docking process using AutoDock Vina 1.2.3 with eight exhaustiveness in the PyRx interface.³² The grid setting is described in Supplementary Table 2. Compounds with a binding energy of ≤7 kcal/mol were selected for protein-ligand analysis by comparing amino acid interactions between the compounds and proteins using Discovery Studio.³³ The 3D structures of protein-ligand complexes were visualized using PyMOL. Protein-ligand complexes exhibiting both interaction similarity to protein-native ligand interactions and a greater number of hydrogen bonds were selected for further molecular dynamics simulations.

Molecular dynamics simulation

YASARA version 21 was used to evaluate structural stability under molecular mechanics simulation.³⁴ The simulation was run under an AMBER14 forcefield³⁵ follow-ed by a number of physiological parameters, including a pH of 7.4, an NaCl concentration of 0.9%, a water density of 0.997, a pressure or 1 bar, and a temperature of 310 K.³⁶ The complex was simulated in a cubic grid shape for a simulation time of 20 ns.

Result and discussion

Phytochemical contents of Orthosiphon leaf extract

Chromatography data analysis revealed thirty-six compounds, predominantly classified as carboxylic acid and fatty acyls, with some carboxylic acids originating from amino acids (Table 1). The compounds with the most significant peak area were Choline and Caffeic Acid, consistent with earlier findings that identified Caffeic Acid and its derivatives as the significant phytochemicals in OA leaves.³⁷ The presence of Caffeic Acid will enhance OA's anti-inflammatory properties and suppress the AD-like phenotype.^38,39 Moreover, fatty acyl compounds and polyphenols are known to exhibit antioxidant and anti-inflammatory activity for the management of AD.⁴⁰ Furthermore, some prenol lipids were identified by chromatography analysis (Table 1). This group of compounds are known to exhibit anti-inflammatory activity by modulating nitric oxide synthase (iNOS) and PTGS2.¹⁵ While most of the compounds identified in this study were not previously reported, they can be developed as anti-inflammatory regulators in addition to those reported previously.

Table 1.

Phytocompounds identified from the ethanolic extract of Orthosiphon aristatus.

Compound No.	Compound name	Formula	Calculated molecular weight	Retention time (min)	Area (Max)
1	Pyrogallol	C6H6O3	126.03172	0.827	61,466,799.74
2	4-Acetamidobutanoic acid	C6H11NO3	145.07369	0.866	63,371,441.24
3	Caprolactam	C6H11NO	113.08413	0.867	40,228,542.43
4	(2S,5aS,8aR)-1-Methyl-6-(4- methylbenzyl)-2-[3-(4-morpholinyl)-3- oxopropyl]octahydropyrrolo[3,2-E][1,4]diazepin-5(2H)-one	C23H34N4O3	436.24443	0.898	7,969,479.05
5	Isoleucine	C6H13NO2	131.0945	0.904	161,188,329.71
6	l-Phenylalanine	C9H11NO2	165.07618	0.931	56,704,526.78
7	Ingenol-3-angelate	C25H34O6	452.21831	0.933	121,277,667.50
8	Thromboxane B2	C20H34O6	408.18953	0.933	102,503,656.38
9	(1R,9R)-5-Cyclohexyl-11- (propylsulfonyl)-7,11-diazatricyclo[7.3.1.02,7]trideca- 2,4-dien-6-one	C20H30N2O3S	378.19808	0.934	48,100,046.04
10	Choline	C5H13NO	103.09989	0.937	2,447,886,975.37
11	dl-Leucineamide	C6H14N2O	89.0844	0.94	83,291,610.14
12	d-(+)-Proline	C5H9NO2	115.06342	0.94	66,613,509.10
13	Betaine	C5H11NO2	117.07894	0.94	262,338,080.66
14	6-formyl-10-(hydroxymethyl)-5- methoxy-3-methylidene-2-oxo- 2H,3H,3aH,4H,5H,8H,9H,11aHcyclodeca[b]furan-4-yl 2-methylbutanoate	C21H28O7	414.15782	0.943	29,861,968.78
15	(7E,13E)-9,15-dihydroxy-4,10,16- trimethyl-1,5,11-trioxacyclohexadeca-7,13-diene- 2,6,12-trione	C16H22O8	364.10951	0.956	106,265,368.39
16	l-Tyrosine	C9H11NO3	181.07367	0.958	12,043,022.71
17	[(2R,3S,4S,5R,6R)-3,4,5-trihydroxy- 6-{[(2S,3S,4R,5R)-4-hydroxy-2,5- bis(hydroxymethyl)-3-{[(2E)-3- (3,4,5-trimethoxyphenyl)prop-2- enoyl]oxy}oxolan-2-yl]oxy}oxan-2-yl]methyl 4-hydroxybenzoate	C31H38O17	704.1964	0.97	46,376,012.13
18	N-{[(1S,4S,6S)-4-{[5-(2- Fluorophenyl)-1,3,4-oxadiazol-2-yl]methyl}-6-isopropyl-3-methyl-2- cyclohexen-1-yl]methyl}isonicotinamide	C26H29 FN4O2	470.20784	6.74	15,745,373.79
19	Dibenzylamine	C14H15N	197.12012	7.432	42,155,207.53
20	Caffeic acid	C9H8O4	180.04197	8.174	2,036,691,177.66
21	Diazepam	C16H13 ClN2O	284.07674	12.863	33,592,588.02
22	Sulfaphenazole	C15H14N4O2S	314.08553	12.937	146,744,065.70
23	6-Gingerol	C17H26O4	276.17748	13.07	33,640,454.81
24	Palmitic Acid	C16H32O2	273.27071	14.689	27,822,083.53
25	Shogaol	C17H24O3	276.17748	16.072	10,847,005.17
26	α-Eleostearic acid	C18H30O2	278.22933	17.016	14,425,731.00
27	Benazol P	C13H11N3O	225.09122	18.099	18,037,978.82
28	Dibutyl phthalate	C16H22O4	278.15735	18.117	42,907,860.63
29	12-Oxo phytodienoic acid	C18H28O3	274.19802	18.871	12,857,695.52
30	9(Z),11(E),13(E)-Octadecatrienoic Acid methyl ester	C19H32O2	292.2452	19.462	281,211,957.49
31	17α-Hydroxyprogesterone	C21H30O3	308.2429	19.991	65,742,938.12
32	1,2,3,4-Tetramethyl-1,3- cyclopentadiene	C9H14	122.10944	20.547	11,496,788.87
33	Ethyl palmitoleate	C18H34O2	282.26058	20.549	51,830,907.75
34	Pinolenic acid	C18H30O2	278.22933	20.558	164,185,936.60
35	4-Methoxycinnamic acid	C10H10O3	178.06271	20.727	19,139,415.22
36	9,11-Dideoxy-9alpha,11alpha-epoxymethanoprostaglandin F2alpha	C21H34O4	310.25861	20.811	168,009,533.13

Proteins and biological targets of Orthosiphon leaf extract

SwissTargetPrediction identified approximately 700 potential targets for OA phytocompounds, while the GeneCard database reported over 1000 proteins involved in the pathogenesis of AD. However, only 245 proteins were common to both the SwissTargetPrediction and GeneCard results, thus indicating that these proteins may be the most likely targets for OA (Figure 1A). These findings suggest that these proteins may play a crucial role in the therapeutic effects of OA against AD. Among these proteins, 50 were annotated as being involved in the inflammatory response (as shown in Figure 1B); these proteins are listed in Figure 1C. The identification of 50 proteins involved in the inflammatory response among the standard set of proteins further supports the hypothesis that OA may exert its therapeutic effects in AD by regulating inflammatory events. The inflammatory response has been implicated in the development and progression of AD^41,42; thus, targeting inflammatory pathways could represent a promising approach for the treatment of AD.^42,43 However, identifying the critical regulatory protein is essential if we are to better modulate inflammatory events in AD.

Figure 1.

Target identification of OA phytocompounds through structure–activity relationship analysis using SwissTargetPrediction accompanied by the GeneCard database revealed potential regulatory activity via an inflammatory pathway. (A) Venn diagram of the common proteins found as potential targets and proteins involved in the pathogenesis of AD, (B) annotation summary from the common proteins mainly classified as inflammatory regulators, and (C) a list of genes involved in the inflammatory response from previous annotation results.

Network biology or protein–protein interaction (PPI) analysis was conducted to identify the most promising target based on previously identified inflammatory regulatory proteins. However, only 30 proteins were found to form an interconnected PPI network (Figure 2A). Subsequently, the centrality of network topology was analysed based on closeness and betweenness. Nodes with high closeness are typically more central to the network and may act as important hubs of information flow.⁴⁴ In the present analysis, TNF-α was identified as the centre point of the network (Figure 2B).

Figure 2.

Biological network of the proteins involved in the inflammatory response. (A) Initial network topology of the proteins involved in the inflammatory response, (B) network topology built by the closeness centrality algorithm, (C) network topology of the proteins visualized according to the betweenness centrality algorithm, and (D) the network topology of the proteins visualized according to degree algorithm. The rank and score of the proteins are listed accordingly, followed by different colour intensities of the top ten highest-ranked proteins from each centrality algorithm; the higher the intensity (red), the higher the rank.

On the other hand, PTGS2 was ranked highest among all proteins in the network based on betweenness centrality (Figure 2C), thus indicating that this protein can connect between two or more virtual nodes in a particular pathway.⁴⁴ A previous study reported that the inhibition of TNF-α and PTGS2 has a beneficial impact on achieving a better prognosis in AD.⁴⁵ Furthermore, since cytokines such as TNF-α are known to be crucial for advancing AD, blocking TNF-α will improve the balance to alleviate the AD.⁴⁶ This result suggests that TNF-α and PTGS2 are promising targets for the management of AD.

Molecular interaction of the compounds of Orthosiphon with TNF-α and PTGS2

The inhibitory potential of phytocompounds from OA against TNF-α and PTGS2 was investigated using molecular docking to understand the intermolecular chemical interactions involved. Of the compounds identified by chromatography analysis, nine compounds exhibited binding to TNF-α with a binding energy lower than −7.0 kcal/mol, while ten compounds showed similar binding to PTGS2. Compound number 9 and number 18 exhibited lower energy requirements than the native inhibitor of TNF-α (−9.143 and −9.442 against −8.032, respectively), and none of the compounds exhibited a binding energy lower than the native inhibitor of PTGS2 (Table 2 and Supplementary Table 3). Furthermore, most of these compounds interacted with similar residues as the native inhibitors of TNF-α and PTGS2 (Supplementary Table 4). However, only four compounds were selected for further analysis due to their resemblance with the native inhibitor in terms of residue–ligand interaction.

Table 2.

The calculated binding energy of phytocompounds screened against TNF-α and PTGS2.

TNF		PTGS
Ligand	Binding Energy (kcal/mol)	Ligand	Binding Energy (kcal/mol)
Inhibitor	−8.432	Inhibitor	−9.391
Compound No. 4	−8.485	Compound No. 11	−8.154
Compound No. 9	−9.143	Compound No. 16	−7.687
Compound No. 7	−8.199	Compound No. 19	−8.342
Compound No. 15	−8.178	Compound No. 22	−7.237
Compound No. 18	−9.442	Compound No. 23	−7.710
Compound No. 21	−7.545	Compound No. 25	−8.209
Compound No. 22	−7.096	Compound No. 26	−7.392
Compound No. 27	−7.626	Compound No. 27	−8.386
Compound No. 31	−8.032	Compound No. 29	−7.496
–		Compound No. 31	−7.201

Compounds number 4, number 9, and number 18 showed the most similar interacting residues as the native inhibitor of TNF-α. These compounds also exhibited low binding affinities that were comparable to the native inhibitor. These formed a more significant number of hydrogen bonds compared to other compounds, indicating good stability for inhibitory activity.⁴⁷ At least one hydrogen bond, several hydrophobic bonds, and van der Waals interactions, were formed. In contrast, the native inhibitor did not form any hydrogen bonds but formed two halogen bonds between the C and O atoms of GLY121 and the F atom of the native inhibitor (Figure 3). This finding provides preliminary evidence for the inhibitory potential of these three compounds being as good as the native inhibitor.

Figure 3.

Structural visualization of TNF-α interacting with a native inhibitor and selected OA phytocompounds in both 3D and 2D. In the 3D visualization, the protein is coloured in magenta ribbons, while the structures of the compounds are represented in red spheres or sticks. The 2D interaction scheme describes the interactions between residues and ligands for each complex.

Similar patterns were observed when considering the PTGS2 complexes. The native inhibitor exhibited the least number of hydrogen bonds when compared to the OA compounds. The native inhibitor formed a carbon–hydrogen bond and several hydrophobic bonds. Three OA compounds, namely compound number 11, number 25, and number 27 exhibited the most similar interacting residues as the native inhibitor. Compound number 11 had the most significant number of hydrogen bonds, while compound number 27 formed an electrostatic bond at ARG120 (Figure 4). As mentioned earlier, these three compounds from OA potentially serve as regulators of PTGS2, in manner similar to the native inhibitor. However, further molecular dynamics simulation is required to understand the structural and binding stability of these compounds when exerting inhibitory activity against their respective receptors. Therefore, these compounds, along with the native inhibitors, will be subjected to molecular dynamics simulation in our future analysis.

Figure 4.

Structural visualization of PTGS2 interacting with a native inhibitor and selected OA phytocompounds in both 3D and 2D. In 3D visualization, the protein is coloured in blue ribbons, while the structure of the compounds is represented by red spheres or sticks. The 2D interaction scheme describes the interactions between residues and ligand atoms for each complex.

Next, the stability of the receptors, ligands, and their binding was assessed by molecular dynamics simulations. This study used several predictors to determine protein structural stability and integrity, including the Root-Mean-Square Deviation (RMSD) of backbone atoms, the Root-Mean-Square Fluctuation (RMSF) of residues, and the Solvent Accessible Surface Area (SASA).^33,36,48 The RMSD of ligand conformation evaluates ligand conformational stability. In contrast, ligand movement and the number of hydrogen bonds were used to assess the binding stability of the ligands to their respective receptors.^33,36

The molecular dynamics results revealed that the binding of the native inhibitor and selected OA compounds had minimal impact on the stability of TNF-α, as indicated by slight fluctuations in the RMSD of the backbone atoms, RMSF, and SASA (Figure 5A–C). Some residues (GLU25:B, HIS73:B, and ARG103:B) appeared to be less stable, particularly from the TNF-α-native inhibitor complex (Figure 5B). However, minor differences were observed in PTGS2-compound number 11, where SASA peaked around 10th ns and then dropped towards the end of the simulations (Figure 6C). Nonetheless, the RMSD of backbone atoms and RMSF showed minor differences between native ligands and OA compounds (Figure 6A–B). Less fluctuating residues were exhibited by PTGS2-compound complexes than TNF-α-compound complexes, with LYS83 and TYR134 observed to have RMSF values beyond 4 Å (Figure 6B). Since the stability of the protein structure, particularly in the vicinity of the binding site, can determine the strength and duration of these interactions,^49,50 a stable protein structure can provide a favourable environment for ligand binding and facilitates the formation of stable protein-ligand complexes.^49,50 Accordin-gly, all simulated complexes were able to exhibit robust inhibitory activity during the simulation time, but not with PTGS2-compound number 11.

Figure 5.

Structural stability evaluation of TNF-α complexes showed that compound number 9 represented a stable inhibitor across different simulation times. (A) RMSD of backbone atoms, (B) RMSF of each residue of TNF-α, (C) SASA, (D) RMSD of ligand conformation, (E) RMSD of ligand movement, and (F) the number of hydrogen bonds in a solute or complex.

Figure 6.

Structural and binding stability evaluation of PTGS2 complexes using molecular dynamics simulation over 20 ns. (A) the RMSD of backbone atoms, (B) the RMSF of each residue of TNF-α, (C) the SASA, (D) the RMSD of ligand conformation, (E) the RMSD of ligand movement, and (F) the number of hydrogen bonds in a solute or complex.

The conformational stability of a ligand during binding is also crucial. A ligand with a stable conformation can adopt and maintain an optimal binding position, thus allowing for stronger interactions with the protein and increased inhibitory activity.⁵¹ Thus, compound number 9 might may have exhibited the most stable inhibitory activity to TNF-α throughout the simulation time (Figure 5D). Furthermore, all ligands of PTGS2 showed excellent conformational stability to support more decisive inhibitory action (Figure 6D). This event is also associated with the structure of a compound, as most of the selected ligands of TNF-α had more rotatable bonds than the ligands of PTGS2.

Furthermore, the binding stability of the ligands from TNF-α complexes showed variable results, with compound number 9 out-performing the native ligand and other compounds with regards to binding stability, regardless of the number of hydrogen bonds (Figure 5E–F). This data also highlights the importance of conformational stability in terms of the stability of a ligand binding to its receptor.⁵¹ On the other hand, almost all of the selected compounds, except for compound number 25, exhibited stable binding stability towards PTGS2. However, there was a logarithmic escalation in the RMSD of ligand movement in compound number 25 from the beginning of the simulations (Figure 6E). Nevertheless, the impact of solute hydrogen bonds on the binding stability among the simulated complexes was indifferent (Figure 6F). Therefore, this study also demonstrates that the importance of hydrogen bonds in ligand binding stability is conditional and depends on the interaction of donors and acceptors of hydrogen bonds with the water as a solvent.⁴⁷

Referring to the simulation results, it is clear that compound number 9 is suspected as the promising TNF inhibitor. With the sole exception of compound number 25, the group of PTGS2 ligands demonstrated favourable inhibitory potential and were distinguished by conformational solid equilibrium and binding consistency. Although there have been few in-depth studies into the specific bioactivity of compound 9, this has been correctly identified as an alkaloid entity. Despite this, the larger family of alkaloid compounds has attracted significant attention for its admirable anti-AD properties.^12,52 On the other hand, compound number 11 (dl-Leucineamide) is a leucine derivative reported to inhibit mTOR signalling.⁵³ Such targeted inhibition can improve abnormal autoreactive T-cell responses and reduce the inflammatory environment.^54,55 Notably, compound number 27, also known as Benazol P or Drometrizole, is a cosmetic ingredient recognized for its ability to absorb ultraviolet (UV) radiation.⁵⁶ Introducing photoprotective attributes augments the therapeutic armamentarium for treating AD and provides a sophisticated approach to improve the efficacy of treatment.⁵⁷ In summary, these results highlight the potential of compound number 9 as a TNF-α inhibitor and suggests its promising inhibitory properties for most PTGS2 ligands, while compound number 11 shows potential to inhibit mTOR signalling and the UV-absorbing characteristics of compound number 27 could enhance AD treatments.

Conclusion

Our chromatography analysis predominantly detected carboxylic acids, fatty acyls, and polyphenols as the primary phytoconstituents in the ethanolic extract of OA. Upon identifying these phytocompounds, network biology-based target identification revealed their potential bioactivity for modulating the inflammation of AD. Notably, TNF-α and PTGS2 emerged as the most probable targets based on hub centrality in the protein–protein interaction network. Subsequently, molecular docking analyses highlighted sixteen compounds with good inhibitory activity against TNF-α and PTGS2. However, only six of these compounds exhibited binding site similarities that were comparable to those of native inhibitors for each target. Further analysis, using molecular dynamics simulations, identified compound number 9 as the most promising inhibitor of TNF-α, while compounds 11 and 27 showed the highest potential as PTGS2 inhibitors. Nonetheless, additional studies are now warranted to validate these findings and explore other potential targets for controlling inflammation in AD.

Source of funding

This research received no specific grant from funding agencies in the public, commercial, or not-for-profit sectors.

Conflict of interest

The authors have no conflict of interest to declare.

Ethical approval

This study did not require ethical approval.

Authors contributions

TAP: Conceptualization; Investigation; Writing - original draft. KH: Writing - review & editing; Supervision. DW: Writing - review & editing; Supervision. HS: Writing - review & editing; Supervision. All authors have critically reviewed and approved the final draft and are responsible for the content and similarity index of the manuscript.

Appendix A. Supplementary data

The following is the Supplementary data to this article: