Network Pharmacology and Computational Approach to Unveiling the Mechanism of Berberine in Depression

Introduction: Depression is a prevalent and often underdiagnosed neuropsychiatric disorder. Natural herbal medicinal products are receiving more attention as potential antidepressants due to their efficacy and low side effects. Berberine, as a primary ingredient in the Indian medicinal plant Berberis vulgaris, has been shown to possess a promising antidepressant effect, although the mechanism of action remains unclear. materials and methods: 2.1 Construction of phytocompounds and corresponding target selection Approximately 52 phytocompounds were identified from B. vulgaris using IMPPAT (https://cb.imsc.res.in/imppat/basicsearch/phytochemical)[33], and the KnapSAcK (http://www.knapsackfamily.com/knapsack_core/top.php) database [34]. The 3D structures (SDF-formatted) of phytocompound was obtain from the PubChem (https://pubchem.ncbi.nlm.nih.gov) database. The retrieved phytocompounds were subject to additional screening for their pharmacokinetic characteristics, which encompassed the ADME property. Various parameters were consider such as oral bioavailability (OB), druglikeness (DL), and Lipinski's rule of five. The requirement that the substances not be mutagenic or allergic was also taken into account. 26 Phytocompounds that met all of the aforementioned requirements were chosen to build the network. Molsoft and SwissADME compute OB and DL, which are used to weed out substances that are unlikely to be pharmaceuticals [35]. Protox-2 was used to examine the toxicity of the chosen active chemicals in order to determine their active and inactive characteristics [36]. Protein targets that interacted with those putative active chemicals were predicted using Swiss Target Prediction (http://www.swisstargetprediction.ch/). The target's probability was fixed at > 0.5. 2.2 Procurement of targets and disease-linked genes The Swiss Target Prediction provided the target genes for the drugs that meet the pharmacokinetics requirements. With the use of a suspected bioactive molecule, this web application allows to determine the macromolecular targets. Furthermore, it was critical to identify the protein linked to depression and stress in order to construct the phytocompound-disease network. The known targets linked to the aetiology of depression were searched using keywords like "depression" or "stress" from two significant databases namely, GeneCards (https://www.genecards.org/), and OMIM (https://www.omim.org/) [37]. Using “Homo sapiens” as a filter only the carefully chosen targets were linked to the disease, and the repeating targets were eliminated to lower the number of false positives. Finally. Potential targets (15421) were identified by combining the proteins/genes obtained from GeneCard and OMIM through a Venn diagram. 2.3 Compound (C)-target (T) network and gene ontology The designed PPI network was constructed by Cytoscape 3.10.2 [38] for visualization of molecular interaction networks and integrated gene expression profiles. The clustered gene was acquired in order to identify the extremely significant genes using the CytoHubba plugin. Subsequently, the most crucial parameters—Degree Centrality (DC), measures the nodes connected to multiple nodes, Closeness Centrality (CC), measures the sum of points between two points, and Betweenness Centrality (BC), measures the network's shortest path were used to choose the core targets grounded on the topological features of the built networks. GO and KEGG pathway analysis was used to perform the biological elucidation of core gene targets, resulting in a functionally ordered network. DAVID was then utilised to determine the cellular location of the analysed core genes as well as potential biomarkers associated with the pathway [39]. 2.3 Protein-Protein Interaction (PPI) network construct Using the maximum confidence of 0.9, a PPI network was built using the target proteins importing to the STRING [40]. The nodes in the network represent proteins, and the related proteins are represented by their edges [41]. The PPI network was pictured with Cytoscape (v3.10.2) [42]. The important genes were determined by calculating two topological features: betweenness and degree. The degree of a node indicated the number of connections whereas the shortest pathways in between two nodes were denoted by betweenness [43]. 2.5 Molecular docking simulation The crystal molecular structures of antidepressant protein namely AKT Serine/Threonine Kinase 1 (AKT1; PDB ID- 6HHG), Proto-oncogene tyrosine-protein kinase Src (SRC; PDB ID- 1BKL), Signal transducer and activator of transcription 3 (STAT3; PDB ID- 6NJS) and HSP990AA1 (Heat shock protein 90α; PDB ID- 3O01) were retrieved from the RCSB database (https://www.rcsb.org/). The proteins were purified using AutoDock tools 1.5.6 by removing water and heteroatoms followed by addition of polar hydrogens for identification of atom types for scoring [44]. The Kollmann charges were added to consider the molecules’ chemical environment. After energy reduction, the protein PDB structure is converted to PDBQT for molecular docking. The ligand was prepared by converting SDF to PDB followed by PDBQT by adding Kollmann charges using AutoDock Tools 1.5.6, using same directory as the protein [45]. The AutoDock Vina 1.2.5, command-line programme, was used to finish the phytocompounds virtual screening [46]. Through the docking simulation, ligands were regarded as flexible while proteins were found to be rigid due to the many torsions generated during ligand synthesis. The first binding pose to be generated showed best docking score having root-mean square deviation (RMSD) zero, hence regarded as being exceptionally genuine out of 10. Additionally, it possesses the greatest binding affinity of all the positions, indicating a more efficient binding. Molecular docking visualised with Biovia Discovery Studio 2024. The total number of Hb, total count of intermolecular bonds, and binding affinity were used to calculate the extent of ligand interaction [47]. 2.6 Molecular Dynamics (MD) Simulation The C-T complexes were simulated through MD by using iMOD (http://imods.chaconlab.org) and CABS-flex v2.0 (http://biocomp.chem.uw.edu.pl/CABSflex2) [48,49]. MD simulations utilized the iMOD server that was used to determine the molecular stability and mobility of the bound C-T complexes where a docked PDB file was uploaded to the server, setting all parameters at default. For simulation all parameters were set default while the time was changed to 10 ns. Using the default settings, the RMSF (root-mean-square fluctuations) were obtained based on the NMR ensemble/MD trajectory. The B-factor, eigenvalues, deformability, covariance and variance map, and elastic network were used to examine the stability of C-T complexes. Methods: To elucidate the multifaceted mechanism of berberine in depression, an integrated computational approach combining network pharmacology, molecular docking, gene ontology (GO) analysis, and molecular dynamics (MD) simulation was employed. The interaction between berberine and potential protein targets was evaluated through molecular docking and binding affinity analyses. results: In this study, the antidepressant effects of berberine and an associated process was investigated. This study explores additional therapy targets for depression and provides the most recent scientific basis for evaluating the effectiveness of multi-component, multi-target chemical formulae. Based on these findings, berberine has the potential to mitigate symptoms of depression, lower inflammatory state, prevent cell apoptosis, and lessen cell damage that may be associated with the AKT/STAT3/HSP9OAA1 signalling pathway. This study does however, have certain limitations because more pharmacological and clinical research is required to corroborate our results. For example, due to scarcity of databases, their moderation and range of information sources may cause contradictions and it is possible to miss the identification of one or two phytochemical during chromatographic analysis. Experimental validation is required to evaluate berberine and its pharmacokinetic characteristics in complex with identified target proteins. This method, however, lays the foundation for future investigations into B. vulgaris antidepressant-protective mechanisms as well as the uses of network pharmacology in drug discovery. Our results will serve as a guide for the clinical use of berberine and will generate fresh ideas for berberine research for depression. Results: The network analysis has identified the neuroactive ligand-receptor interaction-signalling pathway as the primary target, showing the highest binding affinity for the AKT Serine/Threonine Kinase 1 (AKT1) protein. The molecular docking predicts that AKT1 (-10.3 kcal/mol), HSP90AA1 (-8.8 kcal/mol), and STAT3 (-7.5 kcal/mol) may have the most significant binding affinity for berberine. MD simulations confirmed the stability of berberine within the inhibitor binding pockets of these proteins, primarily through hydrogen bonding and π-alkyl interactions. Discussion: Berberine's antidepressant effects appear to be caused by a synergistic, polypharmacological mecha