Study on the molecular mechanism of anti-liver cancer effect of Evodiae fructus by network pharmacology and QSAR model

Abstract

Introduction: Evodiae Fructus (EF) is the dried, near ripe fruit of Euodia rutaecarpa (Juss.) Benth in Rutaceae. Numerous studies have demonstrated its anti-liver cancer properties. However, the molecular mechanism of Evodiae fructus against liver cancer and its structure-activity connection still require clarification. Methods: We utilized network pharmacology and a QSAR (2- and 3-dimensional) model to study the anti-liver cancer effect of Evodiae fructus. First, by using network pharmacology to screen the active substances and targets of Evodiae fructus, we investigated the signaling pathways involved in the anti-liver cancer actions of Evodiae fructus. The 2D-QSAR pharmacophore model was then used to predict the pIC50 values of compounds. The hiphop method was used to create an ideal 3D-QSAR pharmacophore model for the prediction of Evodiae fructus compounds. Finally, molecular docking was used to validate the rationality of the pharmacophore, and molecular dynamics was used to disclose the stability of the compounds by assessing the trajectories in 10 ns using RMSD, RMSF, Rg, and hydrogen bonding metrics. Results: In total, 27 compounds were acquired from the TCMSP and TCM-ID databases, and 45 intersection targets were compiled using Venn diagrams. Network integration analysis was used in this study to identify SRC as a primary target. Key pathways were discovered by KEGG pathway analysis, including PD-L1 expression and PD-1 checkpoint pathway, EGFR tyrosine kinase inhibitor resistance, and ErbB signaling pathway. Using a 2D-QSAR pharmacophore model and the MLR approach to predict chemical activity, ten highly active compounds were found. Two hydrophobic features and one hydrogen bond acceptor feature in the 3D-QSAR pharmacophore model were validated by training set chemicals. The results of molecular docking revealed that 10 active compounds had better docking scores with SRC and were linked to residues via hydrogen and hydrophobic bonds. Molecular dynamics was used to show the structural stability of obacunone, beta-sitosterol, and sitosterol. Conclusion:Pharmacophore 01 has high selectivity and the ability to distinguish active and inactive compounds, which is the optimal model for this study. Obacunone has the optimal binding ability with SRC. The pharmacophore model proposed in this study provides theoretical support for further screening effective anti-cancer Chinese herbal compounds and optimizing the compound structure.

Keywords: Euodiae fructus; QSAR model; liver cancer; molecular docking; molecular dynamics simulation; network pharmacology.

FIGURE 1

GO and KEGG analysis. (A) Top ten item of GO enrichment analysis of biological processes (BP), cell components (CC), and molecular functions (MF). (B) Bubble diagram of KEGG analysis.

FIGURE 2

Network construction. (A) C-T-P network. (B) Degree value distribution histogram of EF compounds in C-T-P network.

FIGURE 3

Key targets analysis. (A) PPI network. (B) Top six key targets derived from cytoHubba’s MCC algorithm. (C) Scores for the top six key targets.

FIGURE 4

Structural of top 10 highly active compounds in EF.

FIGURE 5

The three-dimensional structure of “24-methyl-31-norlanost-9(11)-enol” was used as an example to display “pharmacophore 01” (blue indicates hydrophobic features, green indicates hydrogen bond accepter features).

FIGURE 6

3D-QSAR models with energy grid points as descriptors. (A) Small molecule matching to electrostatic field coefficient isoelectric maps in the model; (B) Small molecule matching to stereo field coefficients isoplot in model.

FIGURE 7

Heat map of the training set compounds predicted by the nine pharmacophore models.

FIGURE 8

Molecular docking results of Obacunone with SRC (1BYG). Gray dashed lines represent hydrophobic interactions, blue lines represent hydrogen bonding interactions, and yellow dashed lines represent salt bridges.

FIGURE 9

The results of Molecular dynamics simulation. (A) RMSD curve of protein-ligand complexes; (B) RMSF curve of protein-ligand complexes; (C) Radius of gyration of complexes; (D) The number of hydrogen bonds formed between the active compounds and 1BYG; (E) The number of hydrogen bonds formed between Ponatinib and 1BYG. In (A–C), 1BYG-ponatinib, 1BYG-Obacunone, 1BYG-Beta-sitosterol, and 1BYG-Sitosterol are represented by black, red, green, and blue curves, respectively.