Abstract
Liver cancer is one of the deadliest cancers worldwide. There is a growing need for natural therapeutic options due to the rising global morbidity incidence of liver cancer. Due to its crucial function in angiogenesis, vascular endothelial growth factor (VEGF) signaling is considered as an ideal target for therapeutic intervention. In this in silico study, we have screened 16 microRNAs (miRNAs) that target the KDR gene, which encode VEGFR-2, and 113 natural compounds derived from Persicaria hydropiper for their anti-angiogenic properties. MicroRNA, hsa-miR-17-3p was identified with potentials of downregulating KDR gene, using several in silico tools. Two possible compounds- 6-Hydroxyluteolin and Isorhamnetin were also identified after performing ADMET profiling and molecular docking. Furthermore, 100 ns molecular dynamics simulations were run to evaluate the stability and conformational changes of protein-ligand complexes. This study identified one microRNA and two natural compounds that showed strong interaction with KDR and its encoded VEGFR-2 receptor, respectively. To validate their effectiveness as therapeutic agents against hepatocellular carcinoma, additional wet lab studies using in vitro and in vivo techniques are necessary.
Introduction
Hepatocellular carcinoma (HCC) is the most prevalent form of liver cancer worldwide. Over 750,000 individuals die from HCC annually, making it the third most common cause of cancer-related deaths worldwide1. The World Health Organization predicts that more than one million individuals would die from HCC by 2030 2. This outlook shows the severe future consequences of liver cancer, which would further aggravate by its poor prognosis and limited treatment options. It highlights the urgent need for searching and investigation of therapeutic options.
Angiogenesis is the process by which new blood vessels are created from the existing vasculature3. It plays a crucial role in the development and progression of HCC by supplying more nutrients and oxygen with effective elimination of waste products4. As the tumor grows, it generates hypoxic conditions due to increased oxygen demand, which causes the activation of hypoxia inducible factors5. These substances increase the secretion of pro-angiogenic factors such as vascular endothelial growth factors (VEGF), epidermal growth factor, basic fibroblast growth factor, platelet-derived growth factor, and insulin-like growth factor6. Among these pro-angiogenic factors, VEGFs are secreted proteins with potent mitogenic properties. It plays a crucial role in pathological angiogenesis by regulating blood vessel growth, permeability, and survival1. VEGF is found in many isoforms, such as VEGF-A, VEGF-B, VEGF-C, VEGF-D, and placental growth factors. VEGF-A is one of the most critical isoforms of VEGF in this process, and it binds with a family of receptors, known as vascular endothelial growth factor receptors (VEGFR) present on almost all endothelial cells7. VEGFRs are of 3 types: VEGFR-1 (Flt-1), VEGFR-2 (KDR/Flk-1), and VEGFR-3 (Flt-4)8. Overexpression of VEGFR-2 was found to promote tumor by inducing angiogenesis, poor prognosis and tumor aggressiveness9. Upon binding to VEGFR-2, VEGF triggers a series of cellular signals like PI3K-Akt pathway, Raf-Mek-Erk pathway, MAPK pathway, phospholipase C-γ (PLCγ)-PKC-eNOS-NO pathway to initiates angiogenesis10. It causes endothelial cells (EC) to proliferate and migrate, forming new blood vessels necessary for rapid tumor growth and metastasis (Fig. 1). It was reported that blocking the VEGFR-2 receptor and its associated VEGF signaling pathways can inhibit abnormal blood vessel formation and related diseases11. Therefore, blocking of VEGF signaling by targeting VEGFR-2 either at the gene level or at the protein level could be the probable therapeutic options to inhibit new angiogenesis in tumor microenvironment. The main framework for development of novel VEGFR-2 inhibitors is to inhibit receptor dimerization and its autophosphorylation12. It was reported that the ATP binding domain of VEGFR-2 is highly conserved in nature13. The protein VEGFR-2 contains the C terminal lobe and the N terminal lobe. The catalytic region of the protein contains ATP binding site and is located at the junction between the C terminal lobe and the N terminal lobe. Modi and Kulkarni suggested that four key regions define the catalytic region14. The first is the hinge region, which is essential because it directly interacts with the adenine ring of ATP, helping it to position in the active site. The crucial amino acid residues in this region include phe918, Cys919, Leu 840 and Glu917. The hydrophobic pocket I region contains lys868 and Val916, connecting the DFG motif and hinge regions. The DFG motif region contains the amino acids glu885 and asp1086, which are essential for gamma phosphorylation.
VEGF mediated signaling pathways in angiogenesis. EC represents endothelial cell.
VEGFR-2 can adopt either “DFG out” or “DFG in” conformations. These conformations are determined by its Aspartate-Phenylalanine-Glycine (DFG) motif. The “DFG-in” conformation represents the active state, whereas the “DFG-out” conformation corresponds to the inactive state15. Three categories of VEGFR-2 inhibitors were reported16 among which type 1 inhibitors (viz. vandetinib, axitinib, and sunitinib) target the “DFG in” conformation by forming a hydrogen bond with cys-919 residue and involving in hydrophobic interaction with ATP binding regions of VEGFR-217. Type 2 inhibitors like sorafenib, regorafenib, and levitanib block the “DFG out” conformation by binding directly to the allosteric region next to the ATP binding pocket of VEGFR-2. On the other hand, type 3 inhibitors or covalent inhibitors are compounds (viz. vatalanib) that bind permanently to the cysteine residues in hydrophobic pocket II in the “DFG out” configuration. These tyrosine kinase inhibitor exhibits potent anti-angiogenic effects by binding with VEGFR-2 but also have serious side effects include diarrhea, hypothyroidism, and exhaustion17. Therefore, inhibition of VEGFR-2 expression by microRNA (miRNA) based KDR gene (which encode VEGFR-2) silencing may be an effective way to avoid this difficulty. On the other hand, exploring plant based natural compounds with low or no toxicity and great efficacy to inhibit VEGFR-2 may be another alternative to overcoming these constraints18. It was reported that that Persicaria hydropiper Linn, a member of the Polygonaceae family is rich in bioactive substances that exhibit anticancer, antioxidant, anxiolytic, antinociceptive, antileukemic, antibacterial, and tyrosine kinase inhibitory qualities19. Persicaria hydropiper Linn is widely found plant in Bangladesh, often referred to as Bishkatali; grows well near rivers, canals, lakes, and roadside areas in tropical and subtropical climates20. It was shown that injecting intraperitoneally into male Swiss Webster albino mice of 99% methanol extract of the aerial portions of Piper hydropiper has antiproliferative effects against Ehrlich ascites cancer cells21. In this backdrop, the present study was undertaken to investigate novel and potentially less toxic agents capable of inhibiting VEGF mediated angiogenesis in HCC. To achieve this objective, a dual approach was adopted: (i) exploration of microRNA (miRNA)-mediated silencing of the KDR gene, which encodes VEGFR-2, and (ii) identification of natural bioactive compounds from Persicaria hydropiper Linn., a medicinal plant with reported pharmacological properties, capable of inhibiting VEGFR-2 because simultaneous targeting of both the KDR gene and VEGFR-2 protein is hypothesized to enhance therapeutic efficacy while minimizing adverse effects. In this study, we report for the first time the integration of miRNA-mediated KDR silencing with plant-derived VEGFR-2 inhibitors from Persicaria hydropiper, introducing a novel dual-targeted anti-angiogenic strategy for the treatment of HCC. We have employed multi-tier in silico-based approaches such as Molecular docking and molecular dynamic simulation in this study due to its time saving and cost effectiveness than traditional laboratory experiments.
Materials and methods
Retrieval of miRNA related information
We have identified the miRNAs targeting KDR gene, sequences of mature miRNAs, target sites of miRNAs in 3′ UTR of KDR gene and the minimum free energy (MFE) of these miRNA-mRNA (KDR) interactions from miRTarBase 2025 database (https://mirtarbase.cuhk.edu.cn/miRTarBase/miRTarBase_2025)22.
Prediction of duplex miRNA-mRNA secondary structure
The binding affinity of miRNA-target interactions (MTIs) were measured through prediction of optimal miRNA-mRNA secondary structure using RNAfold web server (http://rna.tbi.univie.ac.at/cgi-bin/RNAfold.cgi).
Prediction of duplex miRNA-mRNA tertiary structure
We have used RNA COMPOSER web server (http://rnacomposer.cs.putpoznan.pl/) for the prediction of duplex miRNA-mRNA tertiary structure.
Preparation of receptor proteins
The three-dimensional crystal structure of the human VEGFR-2 kinase domain was retrieved from protein data bank (https://www.rcsb.org/) with attached novel pyrrolopyrimidine inhibitor (PDB ID: 3VHE). This protein has a resolution of 1.55 Å, 359 amino acids, and an R-value of 0.186 for work and 0.209 for free. The protein was prepared for analysis by eliminating all heteroatoms including pyrrolopyrimidine inhibitor and water molecules using Discovery Studio (Studio, 2020). Swiss-PdbViewer (version 4.1.0) was used to minimize the energy of protein. Pyrrolopyrimidine was also separately prepared from this co-crystal structure (PDB ID: 3VHE), added hydrogen atoms and saved in SDF format using PyMOL.
Preparation of ligands
From the Indian Medicinal Plants, Phytochemistry, and Therapeutics (IMMPAT) database, a total of 113 phytochemicals of Persicaria hydropiper were obtained23. Sunitinib, a multi-targeted receptor tyrosine kinase (RTK) inhibitor with well-established anti-angiogenic and anti-proliferative properties, was employed as the reference compound24. The chemical structures of sunitinib and the phytochemicals were retrieved from the PubChem database in SDF format25. Using PyRx, the 3D structures of the molecules were optimized and converted into the PDBQT format for docking studies; ChemDraw Pro-8 was used to build 2D structures shown in Table 1.
Molecular docking and virtual screening
We used molecular docking for evaluating the binding affinities and interaction types between ligands and protein26. PyRx software assisted virtual screening and structure-based molecular docking was performed using AutoDock Vina program27. Autodock vina employ a semi flexible docking approach where the protein structure was kept fixed and the ligand was treated as flexible, allowing for torsional rotation around rotatable bonds. It uses gradient based algorithms to explore multiple ligand binding poses within a defined grid and ranks them using an empirical scoring function. This scoring function estimates the binding free energy of each pose based on key molecular interactions, such as steric complementarity, hydrogen binding potential, hydrophobic interactions, torsional entropy. Its multithreading feature uses multiple CPU cores to do calculations in parallel to make sure that docking mode predictions are correct27. The pre prepared 113 naturally occurring compounds were used as ligands and VEGFR-2 as receptor. In the study, we used site-specific docking process and targeted the ATP-binding site of VEGFR-2, which is located within its intracellular tyrosine kinase domain. This ATP-binding site was selected for docking process as market-approved drug Sunitinib inhibits VEGFR-2 by binding to the same binding site. The binding pockets of VEGFR-2, which are critical for its functional activity, were identified based on previously published literature (Supplementary Fig. S1A)28 and the key residues of binding pocket include Glu885, Cys919, Asp1044, Ala866, Phe918, Phe1047, Lys868, Val848, Val916, Leu840, Leu1035, His1026, Ile888, Ile1044, Leu1019, Val898, Ile892, Cys1045, Gly841, Gly922, Val914, Glu917, Leu889, Asp104, and Val89928. No flexible residues were defined for the receptor. The grid box coordinates were carefully determined and the amino acid residues were selected to ensure comprehensive coverage of the VEGFR-2 binding pocket, for accurate docking and effective inhibition of its kinase activity (Supplementary Fig. S1B). The grid box was centered at coordinates x = − 26.87, y = 1.10, z = − 8.12, with dimensions of X = 36, Y = 20.69, and Z = 15. AutoDock Vina generated up to nine binding poses per ligand, each ranked based on predicted binding affinity (ΔG, Kcal/mol). The pose with the lowest (most negative) binding energy was selected as the best binding conformation for further analysis. The docked complexes were then analyzed using Discovery Studio and PyMOL (v2.5.2). Further pose validation was performed using PyMOL for hydrogen bonding and hydrophobic interaction analysis. We analyzed the ligand-receptor interactions of our top 10 compounds having the better negative energy.
Prediction of drug-likeness and ADMET properties
SwissADME, ADMET SAR, Protox III and pkCSM tools were used to investigate the drug likeness and pharmacokinetic (PK) parameters, such as absorption, distribution, metabolism, excretion, and toxicity (ADMET) of the selected compounds. The molecular structures of the ligands were retrieved in SMILES (Simplified Molecular Input Line Entry System) format from the PubChem database. Lipinski’s Rule of Five was applied to evaluate the drug-likeness of the ligands, ensuring that all relevant physicochemical properties were within the accepted thresholds. Key parameters like water solubility, molecular weight, hydrogen bond donors/acceptors, and lipophilicity of the compounds were also assessed, in addition to distribution characteristics like permeability of the blood-brain barrier (BBB) and metabolic interactions involving CYP2D6 and CYP3A4. By utilizing pkCSM and ADMET SAR, the ADMET profiles of the ligands were comprehensively analyzed and compared to that of the reference compound, Sunitinib, and further evaluated against established benchmarks to determine their potential as drug candidates. Using pkCSM and ADMET SAR intestinal absorption, AMES toxicity, and skin sensitization predictions were performed to further evaluate the toxicity of the compounds29. Furthermore, toxicity characteristics such as cytotoxicity, hepatotoxicity, mutagenicity, and immunogenicity of the selected compounds were predicted using the Protox-III web server30.
Molecular dynamics simulation
The Molecular Dynamics (MD) simulation was performed to assess the binding stability of the protein-ligand complexes within the entire system. Yasara softwere version 23.9.29 was applied to study the stability of the drug-VEGFR-2 complex over different time scales. The grid size of (96.9654 × 96.9654 × 96.9654) Å was set to define a cubic box with periodic boundary conditions. The AMBER14 force field31 was employed within this cubic box to simulate the drug-VEGFR-2 complex in an artificial water environment. To neutralize the system, sodium (Na+) and chloride (Cl−) ions were introduced using the TIP3P (transferable intermolecular interaction potential 3 points) model. The steepest descent method was applied to minimize the energy of the drug-VEGFR2 complex. The van der Waals and short-range Coulomb interactions were calculated with a radius cut-off of 8 Å, while long-range Coulomb interactions were computed using the PME (Particle Mesh Ewald) method32. All calculations were conducted under physiological conditions at 298 K, pH 7.4, and 0.9% NaCl and pressure P = 1 bar using standard NVT and NPT ensemble protocols to ensure system stability. Molecular dynamics (MD) simulations were then performed for 100 nanoseconds (ns) with a time step of 2.5 femtoseconds (fs). After the 100 ns simulation, trajectory files were analyzed to determine the RMSD (root mean square deviation), RMSF (root mean square fluctuation), Rg (radius of gyration), SASA (solvent-accessible surface area), and hydrogen bond interactions.
MM-PBSA evaluation
YASARA was used to estimate the binding free energy of the protein-ligand complexes by applying the Molecular Mechanics/Poisson-Boltzmann Surface Area (MMGBSA) approach. The following formula was used to calculate the binding energy:
∆GBinding = GComplex − (GProtein + GLigand).
Analysis of principal components (PCA)
We have assessed multiple multivariate energy variables using principal component analysis (PCA) to evaluate the quality of the protein structure alteration upon ligand binding during MD simulation33. According to structural and energetic factors such as bond lengths, bond angles, dihedral angles, planarity, Van der Waals interactions, and electrostatic energies, PCA revealed variations between groups34. The final 100 ns of the MD trajectory were analyzed for protein-ligand complexes. Plots were created using the factoextra package, and calculations were performed using Minitab 18 (https://www.minitab.com/en-us/) and internal scripts35.
Results
Angiogenesis in tumor microenvironment can be prevented by blocking VEGF signaling pathway which is possible either by blocking VEGF secretion/activity or by suppressing the activity of VEGFR-2 receptor36. As we have focused on targeting the VEGFR-2 receptor, this can be achieved either by silencing of VEGFR-2 encoding KDR gene or through inhibition of VEGFR-2 receptor activity by any specific known drugs/natural compounds. Many gene silencing was reported in literature through binding of specific miRNA with the 3′UTR of gene37. Therefore, we have divided the present study into two parts. In first part, we focused on miRNA mediated silencing of VEGFR-2 encoding KDR gene and in second part; the study was focused on blocking the VEGFR-2 receptor activity by plant based natural compounds.
Analyses of miRNA mediated silencing of KDR gene expression
Gene expression regulation through inhibition of translation or degradation of mRNA by miRNA was reported in many previous studies38. Therefore, in present study we have tried to find out potential miRNAs that can be used for KDR gene silencing in order to suppress VEGFR-2 receptor expression. Accordingly, we have identified 16 miRNAs that can target (experimentally validated) the 3′UTR of KDR gene from miRTarBase 2025 database (Supplementary Table S1). We have also obtained the sequences of these 16 mature miRNAs, their target sites in 3′ UTR of KDR mRNA and MFE for these MTIs from miRTarBase (Supplementary Table S1) by selecting the corresponding accession IDs of each miRNA in the database. Among these 16 miRNAs, 7 miRNAs had strong experimentally validated MTIs (stored in miRTarBase database) and were taken for further study. Among these 7 miRNAs, hsa-miR-19b-1-5p showed non-functional MTI (miRTarBase database) and hsa-miR-1236-3p was known to control VEGFR-339, which is not our study of interest. Therefore, these 2 miRNAs were excluded from our study and rests of the 5 miRNAs were considered for further study (Table 2). There were 3 predicted target sites for each miRNA stored in the miRTarBase database. We have considered taking only 1 out of 3 target sites for each of these 5 miRNAs on the basis of lowest MFE (strong MTI) from Supplementary Table S1 for further analyses. Among these 5 miRNAs, hsa-miR-17-3p showed same MFE (-13.9 Kcal/mol) for 2 target sites (nucleotide region:119–141 and 1470–1492) (Supplementary Table S1). Now to get insights into the MTIs of these 5 miRNAs and their selected target sites, we have predicted optimal secondary structures of these 5 miRNA-mRNA (KDR) duplexes, their corresponding MFEs and dot-bracket notations using RNAfold web server by giving the sequences of miRNA-mRNA duplex (obtained from miRTarBase database) as an input and the predicted results are summarized in Table 2.
Among these 5 miRNAs, hsa-miR-17-3p was considered as best miRNA for miRNA-mRNA duplex formation for both the target sites in terms of its lower MFEs (-9.20 Kcal/mol and − 11.50 Kcal/mol for target-1 and-2, respectively) than other 4 miRNAs. The 3D structure of all these 5 miRNA-mRNA duplexes were predicted using RNA COMPOSER web server by giving dot-bracket notations as an input. The 3D structures of these 5 miRNA-mRNA duplexes were visualized using BIOVIA Discovery Studio (Fig. 2).
Illustrated MTIs of KDR-mRNA with (A) hsa-miR-200b-3p (B) hsa-miR-16-5p (C) hsa-miR-17-3p for both Target-1 and Target-2 (D) hsa-miR-15b-5p and (E) hsa-miR-200c-3p, generated by VIOVIA Discovery Studio Visualizer v20.1.0.19295. Nucleotide positions of KDR-mRNAs and miRNAs are represented by black and orange color, respectively.
Molecular docking analyses
We have docked 113 ligands against VEGFR-2 receptor (Supplementary Table S2). From docking result, it was observed that 6 natural compounds had the binding affinity of − 9.2 to − 10.2 Kcal/mol, whereas control compound, Sunitinib showed the binding affinity of − 9.2 Kcal/mol (Table 1). But top two compounds (6-Hydroxyluteolin, Isorhamnetin) were selected in terms of good binding energy for further study. The binding site of the two lead compounds was validated through molecular docking using pyrrolo[3,2-d] pyrimidine, a co-crystallized ligand from the VEGFR-2 structure (PDB ID: 3VHE), which exhibited a binding affinity of − 10.2 Kcal/mol (Table 1).
Analyses of protein-ligand interactions
Protein-ligand interactions are essential for predicting the binding affinities of ligands with proteins40. The selected two compounds showed various non-bond interaction types with multiple residues of protein whereas, the control compound Sunitinib exhibited 1 conventional hydrogen bond with ASP A:1046 ( Distance: 2.0 Å) and 19 hydrophobic interactions with VAL A:848, LEU A:1035, LEU A:1035, LEU A:1035, CYS A: 1045, ILE A:888, ILE A: 892, VAL A:898, VAL A: 899, ILE A:1044, PHE A: 1047, VAL A:848, LYS A: 868, VAL A:899, VAL A: 916, ALA A: 866, LEU A: 840, VAL A: 848, ALA A: 866 ( Distance: 2.9 Å − 5.5 Å) (Fig. 3). 1 Halogen and 1 Electrostatic interaction with LEU A: 840 LYS A: 868 (Distance; 2.8 Å and 4.9 Å, respectively) were also observed along with others interactions (Table 3).
Depiction of 2D and 3D interactions between human VEGFR-2 and the 6-Hydroxyluteolin, Isorhamnetin, Pyrrolopyrimidine, and the reference inhibitor Sunitinib, as visualized using BIOVIA Discovery Studio Visualizer v20.1.0.19295.
6-Hydroxyluteolin exhibited 2 standard hydrogen bonds with CYS A: 919, ASP A: 1046 (Distance: 1.8 Å and 2.2 Å, respectively) and 10 hydrophobic interactions with LEU A: 840, VAL A: 848, LEU A: 1035, PHE A: 1047, LEU A: 840, VAL A: 848, ALA A: 866, LEU A: 1035, LYS A: 868, CYS A: 1045 (Distance: 3.6 Å − 5.4 Å). On the other hand, Isorhamnetin exhibited 5 conventional hydrogen bonds with LYS A:868, CYS A: 919, GLU A:885, CYS A:919, LEU A: 840 (Distance: 2.1 Å – 2.9 Å) and 12 hydrophobic interactions with VAL A:848, LEU A:1035, LEU A:1035, PHE A:1047, LEU A:840, LEU A:840. VAL A: 848, ALA A: 866. LEU A: 1035. LYS A: 868, LEUA: 1035, CYS A: 1045 (Distance: 3.6 Å – 5.4 Å). However, several forms of hydrophobic interactions (π-σ, π-alkyl, π-π) and standard hydrogen bonds of the top two compounds, control compound (Sunitinib) with various protein residues of VEGFR-2 are given in Table 3; Fig. 3. These two selected compounds interact with an active site residue in the protein. All the protein residues that interact with top two compounds are CYS A :919, ASP A:1046, LEU A :840, VAL A: 848, LEU A: 1035, PHE A: 1047, LEU A: 840, VAL A:848, ALA A: 866, LYS A:868, CYS A:1045, CYS A: 919, GLU A:885, CYS A:919, LEU A:1035, PHE A:1047, LEU A:840, LYS A: 868, LEUA:1035. Conversely, The pyrrolo[3,2-d] pyrimidine inhibitor exhibited 1 conventional hydrogen bond with HIS A:1026 and another with ASP A:1046, at distance of 2.15 Å and 2.77 Å, respectively. Additionally, 1 carbon-hydrogen bond was observed with PHE A: 1047 (Distance: 3.24 Å), along with a π–Cation electrostatic interaction involving LYS A: 868 (Distance: 4.55 Å), and a π–anion electrostatic interaction with GLU A: 885 (Distance: 3.99 Å). The inhibitor also engaged in 6 hydrophobic interactions with residues VAL A: 848, VAL A: 899, VAL A: 916, PHE A: 1047, LYS A: 868 and CYS A: 1045, with interaction distances ranging from 3.60 Å-5.20 Å. Moreover, a halogen interaction involving fluorine was detected with ASP A: 1046 at a distance of 3.49 Å (Table 3; Fig. 3).
Prediction of drug-like properties
Drug-like properties of the selected compounds were evaluated according to Lipinski’s Rule of Five (RO5) and the results of which were summarized in Table 4. The selected compounds exhibited Mw ≤ 500 HBA ≤ 10, HBD ≤ 5, and ilogp ≤ 5. Furthermore, the compounds follow Veber’s principle, with TPSA ≤ 140 Å2 and several rotatable bonds nRb ≤ 5. The selected compounds showed a high gastrointestinal absorption rate and optimal bioavailability of 0.55. Therefore, the selected compound fulfilled all the essential drug-like characteristics.
A drug-likeness diagram in the form of radar chart was obtained from SwissADME (Fig. 4). Each peak on the graph represents parameters of different drug properties that altogether determine the extent of drug likeness of the selected compounds. The pink region in the chart denotes the optimum range for each parameter of different drug properties. Accordingly, it was observed from Fig. 4 that except insaturation, all other parameters of two selected compounds were qualified to be considered for potent drug candidates.
Illustration of drug likeness in the form of radar chart that represents lipophilicity (LIPO), molecular weight (SIZE), polarity (POLAR), insolubility (INSOLU), saturation (INSATU), and rotatable bond flexibility (FLIX) for 6-Hydroxyluteolin (5 A), Isorhamnetin (5B), and the reference drug Sunitinib (5 C). Parameters within the pink zone were regarded as good indicator to be considered for potent drug candidates.
Analyses of ADME properties
To investigate the behavior of two selected compounds within the body we analyzed their ADME properties using ADMET SAR and pkCSM tool (Table 5). We also analyzed the ADME properties of the control compound using the same tools (Table 5). It was observed from Table 5 that both compounds had high GI absorption and suitable skin permeability, but don’t cross the blood-brain barrier in contrast to sunitinib, minimizing the risk of CNS toxicity. Excellent distribution rates (log VDss > 0.45) were also observed for both the compounds compared to Sunitinib suggesting greater tissue distribution. These two compounds didn’t inhibit major CYP450 enzymes such as CYP2D6 and CYP3A4, indicating a low likelihood of drug-drug interactions. In contrast to sunitinib, 6-hydroxyluteolin had superior water solubility and balanced lipophilicity making it more favorable oral formulations. 6-hydroxyluteolin exhibited longest half-life and MRT, indicating prolonged systemic activity and reduced dosing frequency compared to sunitinib with lower half life and Mean residence time. Compared to Sunitinib, both the compounds lack interactions with P-glycoproteins and BSEP minimizing the risk of hepatotoxicity and efflux related complications. Moreover, both the compounds have good drug clearance and do not inhibit Organic Cation Transporter 2 (OCT2), a critical renal transporter involved in drug excretion. Therefore, both 6-Hydroxyluteolin and Isorhamnetin exhibited suitable ADME profile compared to control compound (Sunitinib).
Toxicity prediction
To minimize the risk of adverse drug reactions and potential toxic effects on human body, we have predicted toxicity profiles of two selected lead compounds along with control compound using pkCSM, ADMET SAR and Protox-III web servers. The predicted outcomes obtained from pkCSM and ADMET SAR were summarized in Table 6, while the toxicity predictions generated by ProTox-III were presented in Table 7. It was observed from Table 6 that the selected compounds were non-carcinogenic and exhibited low acute toxicity in rat. The lethal dose (LD50) for 6-Hydroxyluteolin was 3.0200 mol/kg, classifying it under ‘Class II acute oral toxicity’ (LD50 < 500 mg/kg). On the other hand, isorhamnetin showed an LD50 of 2.7192 mol/kg, categorizing it as ‘Class III acute oral toxicity’ (500 mg/kg < LD50 < 5000 mg/kg). Furthermore, both compounds were predicted to be weak inhibitor of human ether-a-go-go-related genes (hERG I and hERG II) and did not show AMES toxicity. Additionally, analyses of Tables 6 and 7 also indicated the absence of hepatoxicity, cytotoxicity, or skin sensitization effects for both the compounds. Notably, the predicted probabilities for mutagenicity, carcinogenicity, and immunotoxicity were all below 80% for both the compounds, suggesting a relatively favorable safety profile. In contrast, the control compound Sunitinib demonstrated predicted probabilities exceeding 80% for both immunotoxicity and cytotoxicity, suggesting a greater potential for adverse toxicological effects. Overall, the toxicity assessments suggest that the two lead compounds, in comparison to the control compound, possess acceptable safety characteristics and are thus considered suitable candidates for further evaluation, including stability analysis through molecular dynamics (MD) simulations.
Molecular dynamics simulation
We have performed Molecular Dynamics (MD) simulations for 100 ns to get the insights of dynamic behavior of protein-ligand interactions. The simulations were conducted to assess the persistence of the protein-ligand complexes, increase the understanding of protein-ligand interactions, and examine the changing dynamics of the target protein. The MD simulations included the analysis of RMSD, RMSF, Rg, and hydrogen bond interactions, as presented in Table 8.
Root mean square deviation (RMSD) analysis
The RMSD of the protein-ligand complex was measured to determine stability of the complex41. Therefore, analyses of RMSD for 6-Hydroxyluteolin-VEGFR-2, Isorhamnetin-VEGFR2, and Sunitinib-VEGFR-2 complexes from Table 8, it was observed that 6-Hydroxyluteolin-VEGFR-2 complex showed a rise in structural deviation between 81 ns and 89 ns and then remain stable until 90 ns but most significant structural deviation was noted at 89.5 ns with an RMSD of 3.296 Å. On the other hand, Isorhamnetin-VEGFR-2 complex maintained nearly consistent structural deviation throughout the period. Moreover, though both complexes exhibited good stability, Isorhamnetin showed better stability than 6-Hydroxyluteolin when complexed with VEGFR-2 (Fig. 5).
RMSD (Å) of protein-ligand complexes over 100 ns of simulation, where complexes of 6-Hydroxyluteolin-VEGFR-2, Isorhamnetin-VEGFR-2, and Sunitinib-VEGFR-2 were denoted by, blue, orange, and green, respectively.
Root-mean-square fluctuations (RMSF) analysis
RMSF determines the flexibility of protein-ligand complexes42. Greater RMSF values signify increased flexibility and decreased stability of the complex, whereas lower RMSF values indicate reduced flexibility and enhanced stability43. Therefore, we have calculated RMSF values of protein-ligand complexes to investigate the dynamics of flexibility and stability of complexes (Table 8; Fig. 6). In the Sunitinib-VEGFR-2 complex, the highest fluctuation was observed at Gly1169 with RMSF of 4.98 Å, whereas 6-Hydroxyluteolin–VEGFR-2 and Isorhamnetin-VEGFR-2 complexes showed the highest fluctuation at Lys 939 and TYR 996, respectively, with RMSF of 9.97 Å and 6.69 Å. On the other hand, the lowest fluctuations were found at the amino acid residues SER 1090 of both Sunitinib-VEGFR-2 (0.30 Å) and 6-Hydroxyluteolin-VEGFR-2 (0.4 Å) complexes, as well as Gly 1092 in Isorhamnetin-VEGFR-2 (0.41 Å) complex. We have also calculated the RMSF values for the active sites identified from the literature28, located between amino acids 840–922 and 1019–1047 of -VEGFR-2, complexed with ligands. The average RMSF values for the control complex were 1.10042 Å for the 840–922 region and 0.96 Å for the 1019–1047 region. In Isohaematin-VEGFR-2 complex, the average RMSF values were 1.131 Å for region 840–922 and 0.964138 Å for region 1019–1047. For 6-hydroxyluteolin-VEGFR-2 complex, the average RMSF values were 1.118193 Å for region 840–922 and 0.90069 Å for region 1019–1047. We have also observed fluctuations within the amino acids 940–1001, but this region did not occur within any of the identified active sites.
RMSF (Å) of protein-ligand complexes over 100 ns simulation, where complexes of 6-Hydroxyluteolin-VEGFR-2, Isorhamnetin-VEGFR-2, and Sunitinib-VEGFR-2 were denoted by, blue, orange, and green, respectively.
Radius of gyration analysis
The radius of gyration (Rg) indicates the spatial distribution of atoms along a plane in a protein and measures the protein compactness with folding during MD simulations. Reduced Rg values signify enhanced compactness and rigidity, whereas higher Rg values indicate a less compact structure. The Rg analysis result is shown in Table 8; Fig. 7. The average Rg value (20.1 Å) of Isorhamnetin-VEGFR-2 complex is the same as control Sunitinib-VEGFR-2 complex and is very close to 6-Hydroxyluteolin-VEGFR-2 complex (20.13 Å). The maximum fluctuation observed for Sunitinib-VEGFR-2, 6-Hydroxyluteolin-VEGFR-2, and Isorhamnetin-VEGFR-2 complexes were 20.4 Å, 20.41 Å, and 20.44 Å, respectively. These results indicate that both compounds have nearly similar level of compactness and rigidity as Sunitinib when complexed with VEGFR-2.
Radius of gyration (Å) values of protein-ligand complexes over 100 ns simulation, where complexes of 6-Hydroxyluteolin-VEGFR-2, Isorhamnetin-VEGFR-2, and Sunitinib-VEGFR-2 were denoted by, blue, orange, and green, respectively.
Solvent accessible surface area (SASA) analysis
We have performed SASA analyses to assess the expansion of surface area of the protein-ligand complexes during simulation and the result of SASA analyses is shown in Table 8; Fig. 8. A higher SASA value suggests that the protein has expanded in volume, and minimum fluctuation of the protein. Significant changes in SASA can occur when small molecules bind, leading to noticeable alterations in the structure of protein44. It was observed from Table 8 that the average SASA value for the complexes of VEGFR-2 with Sunitinib, 6-Hydroxyluteolin, and Isorhamnetin were 14887.17Å2, 15015.28Å2, and 14900.47Å2, respectively. In complexes of VEGFR-2 with Sunitinib, 6-Hydroxyluteolin, and Isorhamnetin, the highest surface areas were observed 15,397.728 Å2 at 16 ns, 15,621.12 Å2 at 89.25 ns, and 15,468.117 Å2 at 97.75 ns, respectively. The fluctuations in surface area were minimum for both compound-VEGFR-2 complexes when compared to control Sunitinib-VEGFR-2 complex. Even though Isorhamnetin-VEGFR-2 complex showed a higher surface area, its fluctuation level was remained lower than the control Sunitinib-VEGFR-2 complex.
SASA (Å2) values of protein-ligand complexes over 100 ns simulation, where complexes of 6-Hydroxyluteolin-VEGFR-2, Isorhamnetin-VEGFR-2, and Sunitinib-VEGFR-2 were denoted by, blue, orange, and green, respectively.
Hydrogen bond interactions analyses
We have analyzed the hydrogen bond interactions for the protein-ligand complexes because hydrogen bonds play essential part in binding of the ligand to the target protein, thereby influencing drug selectivity, metabolism, and absorption45. The results of hydrogen bond interaction analysis are shown in Table 8; Fig. 9. During a 100 ns simulation, the average number of hydrogen bonds formed between the solute and solvent for Sunitinib-VEGFR-2, 6-Hydroxyluteolin-VEGFR-2, and Isorhamnetin-VEGFR-2 complexes were 530, 527, and 534, respectively. Isorhamnetin-VEGFR-2 complex exhibited the highest number of hydrogen bonds, surpassing Sunitinib-VEGFR-2 and 6-Hydroxyluteolin-VEGFR-2 complexes. The maximum hydrogen bonds for complexes of VEGFR-2 with Sunitinib, 6-Hydroxyluteolin, and Isorhamnetin were 567 at 3.5 ns, 566 at 5.75 ns, and 567 at 32.75 ns, respectively. The lowest number of hydrogen bonds for complexes of VEGFR-2 with Sunitinib, 6-Hydroxyluteolin, and Isorhamnetin were 498 at 22.25 ns, 498 at 13.75 ns, and 500 at 71.25 ns, respectively. No significant loss of hydrogen bonds was detected for both the protein-ligand complexes when compared to the control complex. This suggests that both compounds can achieve the desired effects of drug specificity, metabolism, and absorption.
Illustration of hydrogen bonds between the solvent and the solute in protein-ligand complexes during a 100 ns simulation, where complexes of 6-Hydroxyluteolin-VEGFR-2, Isorhamnetin-VEGFR-2, and Sunitinib-VEGFR-2 were denoted by, blue, orange, and green, respectively.
MM-PBSA analysis
The binding free energy of protein-ligand complexes were assessed via the MM-PBSA (Molecular Mechanics Poisson-Boltzmann Surface Area) method to investigate the structural alterations in the VEGFR-2 protein upon ligand binding throughout the simulation. The MM-PBSA analysis results are shown in Table 8; Fig. 10. The average binding free energies for complexes of VEGFR-2 with Sunitinib, 6-Hydroxyluteolin, and Isorhamnetin were found to be 3.456 KJ/mol, 14.733 KJ/mol, and − 3.457 KJ/mol, respectively. Among these, 6-Hydroxyluteolin-VEGFR-2 complex exhibited the highest binding free energy, while Isorhamnetin-VEGFR-2 complex showed the lowest.
Illustration of binding energy (KJ/mol) for protein-ligand complexes over 100 ns simulation, where complexes of 6-Hydroxyluteolin-VEGFR-2, Isorhamnetin-VEGFR-2, and Sunitinib-VEGFR-2 were denoted by, blue, orange, and green, respectively.
PCA analysis
Principal component analysis (PCA) measures the alterations in structural and energetic characteristics of protein-ligand complexes that take place during the MD simulation. Each point in the score plot of result output (Fig. 11A) denotes a distinctive conformer. In the PCA model shown in Fig. 11A, PC1 and PC2 cover 82.7% of the total variance, where PC1 contribute 66.3% and PC2 contribute 16.4%. The score plot exhibit 3 different clusters for 6-Hydroxyluteolin-VEGFR-2 complex (blue), Isorhamnetin-VEGFR-2 complex (red), and Sunitinib-VEGFR-2 complex (green). The points are densely clustered, indicating that the different protein complexes have similar structural and energy characteristics. There is no clear separation between the groups, suggesting that the protein does not change dramatically when VEGFR-2 bound to two test compounds (6-Hydroxyluteolin and Isorhamnetin) compared to standard compound (VEGFR-2-Sunitinib complex). The PCA loading plot (Fig. 11B) indicates that the bond, angle, and Van der Waals (VdW) variables positively correlate with the protein-ligand complexes. Ligand binding significantly impacts on protein structure and the two factors bond and angle may play a crucial role in this structural modification.
(A) A score plot illustrating four groups of data with distinct colors, where each dot indicating a time point. (B) Loading representation derived from Principal Component Analysis (PCA) of energy and structural data.
Discussion
Hepatocellular carcinoma (HCC) is one of the most lethal and prevalent cancer, with cases increasing yearly46. HCC presents significant therapeutic challenges due to its frequent late-stage diagnosis, high recurrence rate, adverse effects associated with current treatments, and resistance to conventional therapies47. Among the existing treatments, tyrosine kinase inhibitors (TKIs), including sorafenib, lenvatinib, regorafenib, and sunitinib, are used in the treatment of HCC. These agents target critical regulators of angiogenesis, particularly the VEGFR signaling pathway48. Despite the demonstrated clinical efficacy of tyrosine kinase inhibitors (TKIs), their therapeutic application is constrained by several limitations, including systemic toxicity, suboptimal target specificity, and the development of resistance mechanisms49. In light of these challenges, the development of safer and more selective alternative therapeutic strategies targeting VEGFR-2 represents a promising avenue for the treatment of HCC50,51. In this in silico study, we have identified potential miRNAs capable of down regulating the KDR gene, which encodes VEGFR-2, as well as natural compounds derived from Persicaria hydropiper with predicted inhibitory activity against VEGFR-2, for potential application in HCC therapy.
We have identified total 16 miRNAs targeting the VEGFR-2 encoding KDR gene. From these16 miRNAs, strong experimental evidences such as reporter assay, western blot and qPCR (stored in miRTarBase database) have identified 5 miRNAs with significant and functional MTIs with 3′UTR of KDR mRNA52,53,54,55,56. RNAfold analysis predicted optimal secondary structures with strong binding energy of -9.20 Kcal/mol and − 11.50 Kcal/mol for the interaction of miR-17-3p with its target sites located at nucleotide positions 119–141 and 1470–1492, respectively, within the 3′UTR of KDR mRNA. For regulation of gene expression, miRNAs form complementary base pairing with the target mRNA, a process typically associated with strong MTIs, as indicated by favorable MFE values38,57. The diagnostic and therapeutic potential of human miRNA has been reported in previous study58. Specifically, miR-17-3p has been identified as a negative regulator of endothelial cells angiogenesis of by down regulating Flk-1 expression56,59. Accordingly, miR-17-3p may serve as a potential therapeutic agent for the treatment of hepatocellular carcinoma by suppressing VEGFR-2 expression through silencing of the KDR/Flk-1 gene.
As an alternative approach to hepatocellular carcinoma treatment, we investigated natural compounds from Persicaria hydropiper targeting VEGFR-2 using Computer-Aided Drug Design (CADD). Many natural compounds like curcumin from curcuma long and Resvatratol from Polygonum cuspidatum were also used to inhibit angiogenesis by targeting VEGFR-2 in previous studies60,61. But many of these compounds exhibit limitation such as low bioavailability, limited specificity or incomplete pathway inhibition62. Persicaria hydropiper, a member of the Polygonaceae family, is recognized for its medicinal properties, particularly its anticancer, antioxidant, anti-inflammatory, and tyrosine kinase inhibitory activities63. Several plants from this family, including Polygonum cuspidatum and Polygonum multiflorum, have exhibited anti-angiogenic and antiproliferative effects by modulating the MAPK, PI3K-AKT, and apoptotic signaling pathways64. Although Persicaria hydropiper has demonstrated anticancer activity in Ehrlich ascites carcinoma and leukemia models65, its anti-angiogenic properties and targeted effects against hepatocellular carcinoma (HCC) remain largely unexplored. Therefore, we have screened 113 natural compounds derived from Persicaria hydropiper against VEGFR-2 using Computer-Aided Drug Design (CADD) methodologies. CADD has become a vital tool in modern drug development, significantly reducing costs, time, and ethical concerns associated with traditional methods66,67. We have prioritized 6 compounds from initial 113 natural compounds derived from Persicaria hydropiper based on their strong binding affinity (docking scores > − 9 Kcal/mol) to VEGFR-2 and potential anti-angiogenic activity. Among them, 6-Hydroxyluteolin (PubChem CID: 5281642) and Isorhamnetin (PubChem CID: 5281654) exhibited highest binding affinity (-10.2 Kcal/mol for each compound) than the control compound Sunitinib (-9.2 Kcal/mol) with VEGFR-2. Our findings are consistent with a recent virtual screening study involving 314 flavonoids from the NPACT database, which identified four compounds exhibiting strong VEGFR-2 inhibitory potential, surpassing axitinib in molecular docking scores68. Both 6-Hydroxyluteolin and Isorhamnetin exhibited strong binding interaction with the ATP binding allosteric site of the VEGFR-2 protein that involves key amino acid interaction of CYS A:919, PHE A:1047, ASP A:1046, LEU A:840, VAL A:848, LEU A:1035, CYS A:1045, LYS A:868, and ALA A:866. These interactions likely contribute to enhanced binding stability and specificity14. Furthermore, the binding profiles of these two lead compounds were consistent with that of a well-characterized pyrrolopyrimidine inhibitor of VEGFR-2, exhibiting similar binding affinities (− 10.2 Kcal/mol), thereby reinforcing the reliability of the observed interactions69.
To be considered viable drug candidates, the PK properties of the potential compounds must be thoroughly evaluated70. The two selected compounds exhibited favorable pharmacokinetic profiles, including optimal lipophilicity, molecular weight, Topological Polar Surface Area (TPSA), and hydrogen bonding characteristics, the key parameters associated with efficient absorption70. Additionally, both compounds demonstrated high gastrointestinal absorption, comparable to structurally related flavonoids such as luteolin and quercetin derivatives71. Furthermore, ADME analyses revealed optimal VDss score for both compounds, indicating their potential for effective distribute across various tissues similar to other flavonoids known to accumulate in lungs and liver72. Penetration of blood-brain barrier by natural compounds may cause central nervous system (CNS) toxicity73. Therefore, lack of blood-brain barrier penetration by both natural compounds represents a notable advantage over Sunitinib in reducing the risk of CNS toxicity. Metabolites must be water-soluble to facilitate their excretion from the body74. In our study, both lead compounds were found to have good water solubility compared to sunitinib, made them easily excretable from the body. It has been reported that CYP3A4 and CYP2D6 enzymes play important roles in drug metabolism75. Therefore, it is essential that candidate drugs do not inhibit these enzymes, and the two selected compounds were found to satisfy this criterion. Moreover, both compounds followed Lipinski’s Rule of Five and Veber’s Rules, confirming their drug-like properties.
The toxicity test is another major challenge in the drug development process, with many potential drug candidates fail in this stage76. In silico toxicity study offers clear advantages over traditional in vivo methods, eliminating issues such as ethical concerns with animal testing, high costs, and lengthy timelines77. hERG testing is crucial in the development of drugs, as the inhibition of the hERG potassium channel in cardiac tissue can lead to cardiac arrhythmias, a prevalent cause for the failure and withdrawal of many drug candidates78. Furthermore, an AMES test was conducted to assess mutagenic potential of the compounds via reverse mutations67. In toxicity analyses, both the selected compounds exhibited weak inhibition of the hERG channel and did not show AMES toxicity. In addition, the immediate acute oral toxicity profiles, as indicated by their LD50 values were within acceptable limits, supporting the suitability of the compounds for potential therapeutic use in humans. Both 6-hydroxyluteolin and isorhamnetin demonstrated low predicted probabilities (< 80%) for immunotoxicity and cytotoxicity, whereas sunitinib exceeded this threshold, consistent with its established immunosuppressive and cytotoxic effects observed in clinical settings79. The favorable toxicity profile of 6-hydroxyluteolin aligns with previous findings on structurally related flavonoids such as luteolin, which have exhibited low toxicity in both in vitro and in vivo models80. The slightly elevated predicted toxicity of isorhamnetin relative to 6-hydroxyluteolin may be attributed to the effects of methylation on flavonoid metabolism and potential tissue accumulation81. Overall, the more favorable safety profiles of these natural compounds, particularly 6-hydroxyluteolin, underscore their potential therapeutic advantages over sunitinib, including reduced adverse effects and improved tolerability.
It was reported that hydrogen bonds play vital roles in enzyme catalysis by stabilizing a ligand within the binding site of protein82. The compact interaction between the ligand and the amino acid residues of the protein suggests the involvement of multiple hydrogen bonds, contributing to increased stability of the protein-ligand complex82. In a recent study, three natural VEGFR-2 inhibitors were identified through the virtual screening of 13,313 African natural compounds using a comprehensive in silico pipeline comprising molecular docking and MD simulations. Among them, the most potent compound, Naringenin-7-p-coumaroylglucoside, demonstrated strong binding affinity to VEGFR-2, forming four hydrogen bonds with key active site residues83. Consistent with these observations, the selected natural compounds in the present study exhibited extensive hydrogen bond interactions with VEGFR-2. Specifically, Isorhamnetin formed five hydrogen bonds with Glu885, Lys868, Leu840, and two with Cys919, while 6-hydroxyluteolin established two hydrogen bonds with Asp1046 and Cys919. The reference compound, Sunitinib, binds to the active site of the VEGFR-2 receptor as a competitive ATP inhibitor by forming a hydrogen bond with Asp1046, a conserved residue critical for competitive binding of ATP. The greater number of hydrogen bonds formed by these natural compounds suggests stronger and more stable binding within the VEGFR-2 active site compared to Sunitinib. The hydrogen bond interaction with Cys919 is particularly important, as this residue has been identified as a potential hotspot for enhancing binding affinity in VEGFR-2 inhibitors84. Additionally, the hydrogen bond formed with Leu840 appears to be unique to these natural compounds, as this interaction is rarely reported among conventional VEGFR-2 inhibitors85.
Following static docking approaches, MD simulations capture the natural dynamics of the system, revealing how hydrogen bonds and other interactions evolve over time under realistic physiological conditions86. It was found from previous studies that a protein-ligand complex is considered steady with a lower value of RMSD and RMSF87,88.Therefore, low average RMSD value (< 3Å), and RMSF value (< 1.5 Å), indicate good structural stability and less fluctuation of the protein-ligand complexes in our study. Previous study reported that a greater number of hydrogen bonds in protein-ligand complexes correlate with enhanced binding affinity89. In the present study, both selected compounds formed almost similar number of hydrogen bonds with the VEGFR-2 before and after the simulation. This observation suggests that the binding pocket is structurally compact and undergoes minimal conformational changes, indicating enhanced stability of the drug candidates within the binding site. A higher binding free energy value is indicative of a more stable protein-ligand interaction90. In this context, 6-Hydroxyluteolin-VEGFR-2 complex exhibits the highest binding free energy, suggesting the formation of a highly stable protein–ligand complex. Both protein-ligand complexes exhibited comparable higher rigidity and compactness throughout the simulation, as indicated by the Rg values consistently ranging from 19 to 20 Å, closely aligning with their initial conformations. Additionally, the solvent-accessible surface area of both complexes also exhibited negligible variation from the initial state. Therefore, over the course of our 100 ns simulations, the ligand–VEGFR-2 complexes maintained stable binding. The ligands accommodated the intrinsic dynamic motions of the protein while preserving key interactions, indicating a strong yet flexible fit within the binding site.
Although this in silico study offers encouraging results and serves as a primary platform for identifying potential therapeutic candidates, it is limited by the constraints of computational methods. The ability of hsa-miR-17-3p to regulate the KDR gene depends on sequence complementarity and thermodynamic stability that cannot fully replicate the complexity of biological systems, often neglect post-transcriptional modifications, tissue-specific expression profiles and competitive endogenous RNA (ceRNA) networks which can significantly influence microRNA functionality in vivo91,92. Additionally, the interactions between hsa-miR-17-3p, the natural compounds (6-Hydroxyluteolin and Isorhamnetin), and the KDR/VEGFR-2 target have not been validated experimentally through in vitro assays like luciferase reporter or western blot analyses, nor through in vivo studies using animal models93. The study also focuses on a single molecular target (KDR/VEGFR-2), while hepatocellular carcinoma (HCC) involves multiple oncogenic pathways, suggesting that a multi-targeted approach may be necessary for effective therapy94. Moreover, ADMET (Absorption, Distribution, Metabolism, Excretion, and Toxicity) predictions rely on established datasets and may not accurately reflect the complex pharmacokinetic behaviors observed in living organisms95. The predictions from these models necessitate empirical validation through laboratory-based pharmacokinetic and toxicity experiments96. Molecular docking methods assess the interaction of natural compounds using static structural models (receptor) and scoring functions that estimate binding affinities. Because of conformational changes, receptor and target molecules exhibit significant flexibility in solution97. As a result, designing an inhibitor-based solely on the search for a single, rigid structure might lead to incorrect results96. While MD simulations offer insights into the conformational stability of protein-ligand complexes over time, their accuracy depends on the selected force fields, which are inherently simplified representations of molecular forces98.
Therefore, despite the promising identification of a regulatory miRNA and two bioactive compounds with potential anti-angiogenic effects against VEGFR-2, the conclusions drawn from this study remain hypothetical until confirmed through rigorous in vitro and in vivo experiments.
The findings of this study pave the way for the development of novel, natural therapeutic strategies targeting liver cancer, specifically through the inhibition of VEGF signaling via the KDR gene. The identification of hsa-miR-17-3p and the natural compounds 6-Hydroxyluteolin and Isorhamnetin present promising candidates for further exploration in treating hepatocellular carcinoma. Future research should focus on validating these interactions through in vitro and in vivo experiments, as well as evaluating the efficacy, bioavailability, and safety of these agents in clinical environments. Ultimately, these insights may contribute to the development of miRNA-based therapies or phytochemical-derived drugs that are less toxic and more effective in controlling the progression of liver cancer.
Conclusions
The current study focuses on identification of potential therapeutic agents targeting VEGFR-2 and its encoding gene KDR/Flk-1 for the treatment of hepatocellular carcinoma. The work draws correlation of strong functional MTIs of miR-17-3p with KDR mRNA and inhibition of expression through downregulation of KDR gene. Subsequent evaluation using molecular docking, toxicity test, PK analyses and dynamic MD simulations also revealed two natural compounds namely, 6-Hydroxyluteolin and Isorhamnetin that have strong inhibitory potential against the angiogenic VEGFR-2 receptor. Therefore, these two compounds and miR-17-3p can function as promising therapeutic agents for prevention and treatment of hepatocellular carcinoma. However, wet lab validation by in vivo study is necessary before considering them as effective therapeutic agents for the treatment of hepatocellular carcinoma.
Data availability
All data generated or analysed during this study are included in this published article (and its Supplementary Information files).
References
Aspriţoiu, V. M., Stoica, I., Bleotu, C. & Diaconu, C. C. Epigenetic regulation of angiogenesis in development and tumors progression: Potential implications for cancer treatment. Front. Cell. Dev. Biol. 9, 689962 (2021).
Hsu, P. et al. Regorafenib for Taiwanese patients with unresectable hepatocellular carcinoma after Sorafenib failure: Impact of alpha-fetoprotein levels. Cancer Med. 11 (1), 104–116 (2022).
Tsuruda, T. et al. Adventitial mast cells contribute to pathogenesis in the progression of abdominal aortic aneurysm. Circ. Res. 102 (11), 1368–1377 (2008).
Pinto, E., Pelizzaro, F., Farinati, F. & Russo, F. P. Angiogenesis and hepatocellular carcinoma: From molecular mechanisms to systemic therapies. Med. (B Aires). 59 (6), 1115 (2023).
Muppala, S. Growth Factor-Induced angiogenesis in hepatocellular carcinoma. Crit. Rev. Oncog. 26 (1), 61–68 (2021).
Chu, J. S. et al. Expression and prognostic value of VEGFR-2, PDGFR-β, and c-Met in advanced hepatocellular carcinoma. J. Exp. Clin. Cancer Res. 32 (1), 16 (2013).
Mabrouk, R. et al. Design, synthesis, and biological evaluation of new potential unusual modified anticancer immunomodulators for possible non-teratogenic quinazoline-based thalidomide analogs. Int. J. Mol. Sci. 24 (15), 12416 (2023).
Takahashi, M. ERK/MAPK-dependent PI3K/Akt phosphorylation through VEGFR-1 after VEGF stimulation in activated hepatic stellate cells. Hepatol. Res. 26 (3), 232–236 (2003).
Apte, R. S., Chen, D. S. & Ferrara, N. VEGF in signaling and disease: Beyond discovery and development. Cell 176 (6), 1248–1264 (2019).
Koch, S. & Claesson-Welsh, L. Signal transduction by vascular endothelial growth factor receptors. Cold Spring Harb Perspect. Med. 2 (7), a006502–a006502 (2012).
Niu, G. & Chen, X. Vascular endothelial growth factor as an anti-angiogenic target for cancer therapy. Curr. Drug Targets. 11 (8), 1000–1017 (2010).
Huang, L. et al. Development and strategies of VEGFR-2/KDR inhibitors. Future Med. Chem. 4 (14), 1839–1852 (2012).
Kornev, A. P., Haste, N. M., Taylor, S. S. & Ten Eyck, L. F. Surface comparison of active and inactive protein kinases identifies a conserved activation mechanism. Proc. Natl. Acad. Sci. 103 (47), 17783–17788. (2006).
Modi, S. J. & Kulkarni, V. M. Vascular endothelial growth factor receptor (VEGFR-2)/KDR inhibitors: Medicinal chemistry perspective. Med. Drug Discov. 2, 100009 (2019).
Zhu, Y. & Hu, X. Molecular recognition of FDA-Approved small molecule protein kinase drugs in protein kinases. Molecules 27 (20), 7124 (2022).
Elkaeed, E. B. et al. Discovery of new VEGFR-2 inhibitors: Design, synthesis, anti-proliferative evaluation, docking, and MD simulation studies. Molecules 27 (19), 6203 (2022).
Liu, Y. & Gray, N. S. Rational design of inhibitors that bind to inactive kinase conformations. Nat. Chem. Biol. 2 (7), 358–364 (2006).
Nisar, B., Sultan, A. & Rubab, S. L. Comparison of medicinally important natural products versus synthetic Drugs-A short commentary. Nat. Prod. Chem. Res. 06 (02), 308 (2018).
Khatun, A., Imam, M. Z. & Rana, M. S. Antinociceptive effect of methanol extract of leaves of persicaria hydropiper in mice. BMC Complement. Altern. Med. 15 (1), 63 (2015).
Ayaz, M. et al. Persicaria hydropiper (L.) delarbre: A review on traditional uses, bioactive chemical constituents and pharmacological and toxicological activities. J. Ethnopharmacol. 251, 112516 (2020).
Huq, A. K. M. M., Jamal, J. A. & Stanslas, J. Ethnobotanical, phytochemical, pharmacological, and toxicological aspects of persicaria hydropipe. Evid.-Based Complement. Altern. Med. 2014, 782830 (2014).
Cui, S. et al. MiRTarBase 2025: Updates to the collection of experimentally validated microRNA–target interactions. Nucleic Acids Res. 53, D147 (2024).
Mohanraj, K. et al. IMPPAT: A curated database of Indian medicinal Plants, phytochemistry and therapeutics. Sci. Rep. 8 (1), 4329 (2018).
Li, Y. et al. Treatment with a VEGFR-2 antibody results in intra-tumor immune modulation and enhances anti-tumor efficacy of PD-L1 Blockade in syngeneic murine tumor models. PLoS One. 17 (7), e0268244 (2022).
Kim, S. et al. PubChem 2023 update. Nucleic Acids Res. 51 (D1), D1373–D1380 (2023).
Ahammad, F. et al. Pharmacoinformatics and molecular dynamics simulation-based phytochemical screening of Neem plant (Azadiractha indica) against human cancer by targeting MCM7 protein. Brief. Bioinform 22 (5), bbab098 (2021).
Trott, O. & Olson, A. J. AutoDock vina: Improving the speed and accuracy of docking with a new scoring function, efficient optimization, and multithreading. J. Comput. Chem. 31 (2), 455–461 (2010).
Gaikwad, N. M. et al. Albendazole repurposing on VEGFR-2 for possible anticancer application: In-silico analysis. PLoS One. 18 (8), e0287198 (2023).
Pradeepkiran, J. A., konidala, K. & Yellapu, N. Bhaskar. Modeling, molecular dynamics, and Docking assessment of transcription factor rho: A potential drug target in Brucella melitensis 16M. Drug Des. Devel Ther. 9, 1897. (2015).
Banerjee, P., Kemmler, E., Dunkel, M. & Preissner, R. ProTox 3.0: A webserver for the prediction of toxicity of chemicals. Nucleic Acids Res. 52 (W1), W513–W520 (2024).
Krieger, E., Nielsen, J. E., Spronk, C. A. E. M. & Vriend, G. Fast empirical pKa prediction by Ewald summation. J. Mol. Graph Model. 25 (4), 481–486 (2006).
Darden, T., York, D. & Pedersen, L. Particle mesh ewald: An N ⋅log(N) method for Ewald sums in large systems. J. Chem. Phys. 98 (12), 10089–10092 (1993).
David, C. C. & Jacobs, D. J. Principal component analysis: A method for determining the essential dynamics of proteins. In Protein Dynamics 193–226. (2014).
Kitao, A. Principal component analysis and related methods for investigating the dynamics of biological macromolecules. J. (Basel). 5 (2), 298–317 (2022).
Lê, S., Josse, J. & Husson, F. FactoMineR: An R package for multivariate analysis. J. Stat. Softw. 25 (1), 1–18 (2008).
Wang, L., Liu, W. Q., Broussy, S., Han, B. & Fang, H. Recent advances of anti-angiogenic inhibitors targeting VEGF/VEGFR axis. Front. Pharmacol. 14, 1307860. (2024).
Kosek, D. M., Banijamali, E., Becker, W., Petzold, K. & Andersson, E. R. Efficient 3′-pairing renders MicroRNA targeting less sensitive to mRNA seed accessibility. Nucleic Acids Res. 51 (20), 11162–11177 (2023).
Nakanishi, K. Anatomy of four human argonaute proteins. Nucleic Acids Res. 50 (12), 6618–6638 (2022).
Jones, D. et al. Mirtron MicroRNA-1236 inhibits VEGFR-3 signaling during inflammatory lymphangiogenesis. Arterioscler. Thromb. Vasc Biol. 32 (3), 633–642 (2012).
Wang, L. et al. Accurate and reliable prediction of relative ligand binding potency in prospective drug discovery by way of a modern Free-Energy calculation protocol and force field. J. Am. Chem. Soc. 137 (7), 2695–2703 (2015).
Boakye, A., Gasu, E. N., Mensah, J. O. & Borquaye, L. S. Computational studies on potential small molecule inhibitors of Leishmania pteridine reductase 1. J. Biomol. Struct. Dyn. 41 (21), 12128–12141 (2023).
Bornot, A., Etchebest, C. & de Brevern, A. G. Predicting protein flexibility through the prediction of local structures. Proteins Struct. Funct. Bioinform. 79 (3), 839–852 (2011).
Liu, K., Watanabe, E. & Kokubo, H. Exploring the stability of ligand binding modes to proteins by molecular dynamics simulations. J. Comput. Aided Mol. Des. 31 (2), 201–211 (2017).
Tarek, M. & Tobias, D. J. The dynamics of protein hydration water: A quantitative comparison of molecular dynamics simulations and Neutron-scattering experiments. Biophys. J. 79 (6), 3244–3257 (2000).
Chen, D. et al. Regulation of protein-ligand binding affinity by hydrogen bond pairing. Sci. Adv. 2(3), e1501240 (2016).
Siegel, R. L., Miller, K. D. & Jemal, A. Cancer statistics, 2019. CA Cancer J. Clin. 69 (1), 7–34 (2019).
Vilarinho, S. & Taddei, T. Therapeutic strategies for hepatocellular carcinoma: New advances and challenges. Curr. Treat. Options Gastroenterol. 13 (2), 219–234 (2015).
Chen, R. et al. Modulation of the tumour microenvironment in hepatocellular carcinoma by tyrosine kinase inhibitors: From modulation to combination therapy targeting the microenvironment. Cancer Cell. Int. 22 (1), 73 (2022).
Wing Tung Ho, V., Yue Tan, H., Wang, N. & Feng, Y. Cancer management by tyrosine kinase inhibitors: Efficacy, limitation, and future strategies. In Tyrosine Kinases as Druggable Targets in Cancer. (IntechOpen, 2019).
Shibuya, M. Vascular endothelial growth factor (VEGF) and its receptor (VEGFR) signaling in angiogenesis: A crucial target for Anti- and Pro-Angiogenic therapies. Genes Cancer. 2 (12), 1097–1105 (2011).
Roskoski, R. Vascular endothelial growth factor (VEGF) signaling in tumor progression. Crit. Rev. Oncol. Hematol. 62 (3), 179–213 (2007).
Choi, Y. C., Yoon, S., Jeong, Y., Yoon, J. & Baek, K. Regulation of vascular endothelial growth factor signaling by miR-200b. Mol. Cells. 32 (1), 77–82 (2011).
Shi, L. et al. MiR-200c increases the radiosensitivity of non-small-cell lung cancer cell line A549 by targeting VEGF-VEGFR2 pathway. PLoS One. 8 (10), e78344 (2013).
Chan, L. S., Yue, P. Y. K., Wong, Y. Y. & Wong, R. N. S. MicroRNA-15b contributes to ginsenoside-Rg1-induced angiogenesis through increased expression of VEGFR-2. Biochem. Pharmacol. 86 (3), 392–400 (2013).
Chamorro-Jorganes, A. et al. MicroRNA-16 and MicroRNA-424 regulate cell-autonomous angiogenic functions in endothelial cells via targeting vascular endothelial growth factor receptor-2 and fibroblast growth factor receptor-1. Arterioscler. Thromb. Vasc Biol. 31 (11), 2595–2606 (2011).
Yin, R., Wang, R., Guo, L., Zhang, W. & Lu, Y. MiR-17-3p inhibits angiogenesis by downregulating Flk-1 in the cell growth signal pathway. J. Vasc Res. 50 (2), 157–166 (2013).
O’Brien, J., Hayder, H., Zayed, Y. & Peng, C. Overview of MicroRNA biogenesis, mechanisms of actions, and circulation. Front. Endocrinol. (Lausanne) 9, 402 (2018).
Goud, V. R. et al. A bioinformatic approach of targeting SARS-CoV-2 replication by silencing a conserved alternative reserve of the orf8 gene using host MiRNAs. Comput. Biol. Med. 145, 105436 (2022).
Sun, X., Chen, J., Wang, L., Li, G. & Wang, A. A gene chip study suggests that < scp > miR -17‐3p is associated with diabetic foot ulcers. Int. Wound J. 20 (5), 1525–1533 (2023).
Saberi-Karimian, M. et al. Vascular endothelial growth factor: An important molecular target of < u > curcumin. Crit. Rev. Food Sci. Nutr. 59 (2), 299–312 (2019).
Ab Halim, N. S. et al. Anti-Angiogenic effect of polygonum species: A comprehensive review of literature. Sains Malays. 53 (2), 369–381 (2024).
Shanmugam, M. K., Warrier, S., Kumar, A. P., Sethi, G. & Arfuso, F. Potential role of natural compounds as anti-angiogenic agents in cancer. Curr. Vasc. Pharmacol. 15 (6), 503–519 (2017).
Idoudi, S., Tourrette, A., Bouajila, J., Romdhane, M. & Elfalleh, W. The genus polygonum: An updated comprehensive review of its ethnomedicinal, phytochemical, pharmacological activities, toxicology, and phytopharmaceutical formulation. Heliyon 10 (8), e28947 (2024).
Wei, Q. & Zhang, Y. H. Flavonoids with anti-angiogenesis function in cancer. Molecules. 29 (7), 1570 (2024).
Kabidul Azam, M. N., Rahman, M. M., Biswas, S. & Ahmed, M. N. Appraisals of Bangladeshi medicinal plants used by folk medicine practitioners in the prevention and management of malignant neoplastic diseases. Int. Sch. Res. Notices. 2016, 1–12 (2016).
Madhavi Sastry, G., Adzhigirey, M., Day, T., Annabhimoju, R. & Sherman, W. Protein and ligand preparation: Parameters, protocols, and influence on virtual screening enrichments. J. Comput. Aided Mol. Des. 27 (3), 221–234 (2013).
Roy, A. S. et al. In Silico identification of potential inhibitors with higher potency than bumetanide targeting NKCC1: An important ion co-transporter to treat neurological disorders. Inf. Med. Unlocked 27, 100777 (2021).
Shah, A., Parmar, G., Shah, U. & Perumal, S. Virtual screening, molecular docking studies and DFT calculations of novel anticancer flavonoids as potential VEGFR-2 inhibitors. Chem. Afr. 6 (4), 1847–1861 (2023).
Oguro, Y. et al. Design, synthesis, and evaluation of 5-methyl-4-phenoxy-5H-pyrrolo[3,2-d]pyrimidine derivatives: Novel VEGFR2 kinase inhibitors binding to inactive kinase conformation. Bioorg. Med. Chem. 18 (20), 7260–7273 (2010).
Yang, M. M. et al. Lack of association of C3 gene with uveitis: Additional insights into the genetic profile of uveitis regarding complement pathway genes. Sci. Rep. 7 (1), 879 (2017).
Hu, L., Luo, Y., Yang, J. & Cheng, C. Botanical flavonoids: Efficacy, absorption, metabolism and advanced pharmaceutical technology for improving bioavailability. Molecules 30 (5), 1184 (2025).
Zhivkova, Z. D., Mandova, T. & Doytchinova, I. Quantitative Structure—Pharmacokinetics relationships analysis of basic drugs: Volume of distribution. J. Pharm. Pharm. Sci. 18 (3), 515 (2015).
Fricker, G. Drug interactions with natural products at the blood brain barrier. Curr. Drug Metab. 9 (10), 1019–1026 (2008).
Chen, C. H. Metabolic conversion of foreign compounds. In Activation and Detoxification Enzymes. 33–43. (Springer, 2024).
Finch, A. & Pillans, P. P-glycoprotein and its role in drug-drug interactions. Aust Prescr [Internet]. Sep [cited 2024 Sep 12];37. (2014). Available from: https://www.australianprescriber.com
Biala, G. et al. Research in the field of drug design and development. Pharmaceuticals. 16, 1283 (2023).
Madden, J. C., Enoch, S. J., Paini, A. & Cronin, M. T. D. A review of in silico tools as alternatives to animal testing: Principles, resources and applications. Altern. Lab. Anim. 48, 146–172. (2020).
Garrido, A., Lepailleur, A., Mignani, S. M., Dallemagne, P. & Rochais, C. hERG toxicity assessment: Useful guidelines for drug design. Eur. J. Med. Chem. 195, 112290 (2020).
Papaetis, G. S. & Syrigos, K. N. Sunitinib. BioDrugs 23 (6), 377–389. (2009).
Imran, M. et al. Luteolin, a flavonoid, as an anticancer agent: A review. Biomed. Pharmacother. 112, 108612 (2019).
Koirala, N., Thuan, N. H., Ghimire, G. P., Thang, D. & Van, Sohng, J. K. Methylation of flavonoids: Chemical structures, bioactivities, progress and perspectives for biotechnological production. Enzyme Microb. Technol. 86, 103–116 (2016).
Hubbard, R. E. & Kamran Haider, M. Hydrogen bonds in proteins: Role and strength. In Encyclopedia of Life Sciences. (Wiley, 2010).
Baammi, S., El Allali, A. & Daoud, R. Unleashing nature’s potential: A computational approach to discovering novel VEGFR-2 inhibitors from African natural compound using virtual screening, ADMET analysis, molecular dynamics, and MMPBSA calculations. Front. Mol. Biosci. 10, 1227643 (2023).
McTigue, M. et al. Molecular conformations, interactions, and properties associated with drug efficiency and clinical performance among VEGFR TK inhibitors. Proc. Natl. Acad. Sci. 109 (45), 18281–18289 (2012).
Wang, J. et al. Discovery of vascular endothelial growth factor receptor tyrosine kinase inhibitors by quantitative structure–activity relationships, molecular dynamics simulation and free energy calculation. RSC Adv. 6 (42), 35402–35415 (2016).
Salo-Ahen, O. M. H. et al. Molecular dynamics simulations in drug discovery and pharmaceutical development. Processes 9 (1), 71 (2020).
Ghosh, R., Chakraborty, A., Biswas, A. & Chowdhuri, S. Identification of polyphenols from broussonetia papyrifera as SARS CoV-2 main protease inhibitors using in Silico Docking and molecular dynamics simulation approaches. J. Biomol. Struct. Dyn. 39 (17), 6747–6760 (2021).
Singh, M. et al. Computational and biophysical characterization of heterocyclic derivatives of anthraquinone against human Aurora kinase A. ACS Omega. 7 (44), 39603–39618 (2022).
Yasmeen, N. et al. Anti-angiogenic potential of phytochemicals from clerodendrum inerme (L.) Gaertn investigated through in Silico and quantum computational methods. Mol. Divers. 29, 215 (2024).
Chen, D. E., Willick, D. L., Ruckel, J. B. & Floriano, W. B. Principal component analysis of binding energies for single-point mutants of hT2R16 bound to an agonist correlate with experimental mutant cell response. J. Comput. Biol. 22 (1), 37–53 (2015).
Dweep, H. & Gretz, N. miRWalk2.0: A comprehensive atlas of microRNA-target interactions. Nat. Methods. 12 (8), 697–697 (2015).
Bartel, D. P. MicroRNAs: Target recognition and regulatory functions. Cell 136 (2), 215–233 (2009).
Moffat, J. G., Vincent, F., Lee, J. A., Eder, J. & Prunotto, M. Opportunities and challenges in phenotypic drug discovery: An industry perspective. Nat. Rev. Drug Discov. 16 (8), 531–543 (2017).
Llovet, J. M. et al. Hepatocellular carcinoma. Nat. Rev. Dis. Primers. 7 (1), 6 (2021).
Schyman, P., Liu, R., Desai, V. & Wallqvist, A. vNN web server for ADMET predictions. Front. Pharmacol. 8, 889 (2017).
Shah, A. & Jain, M. Limitations and future challenges of computer-aided drug design methods. In Computer Aided Drug Design (CADD): From Ligand-Based Methods To Structure-Based Approaches. 283–297 (Elsevier, 2022).
Kokh, D. B., Wade, R. C. & Wenzel, W. Receptor flexibility in small-molecule Docking calculations. WIREs Comput. Mol. Sci. 1 (2), 298–314 (2011).
Karplus, M. & McCammon, J. A. Molecular dynamics simulations of biomolecules. Nat. Struct. Biol. 9 (9), 646–652 (2002).
Author information
Authors and Affiliations
Contributions
S.F.A. conceived the idea; S.F.A., S.S., M.F.I., and A.M. designed, performed the experiments, analyzed the results and drafted the manuscript; H.A.L., and A.A.E. performed the experiments; H.A.L., A.A.E. T.M, and S.F.A. reviewed the results and edited the manuscript; All authors have read and approved the manuscript.
Corresponding authors
Ethics declarations
Competing interests
The authors declare no competing interests.
Additional information
Publisher’s note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Supplementary Information
Below is the link to the electronic supplementary material.
Rights and permissions
Open Access This article is licensed under a Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International License, which permits any non-commercial use, sharing, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if you modified the licensed material. You do not have permission under this licence to share adapted material derived from this article or parts of it. The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by-nc-nd/4.0/.
About this article
Cite this article
Ahmed, S.F., Samin, S., Mondal, A. et al. Computational analysis of miRNA mediated KDR gene regulation and natural VEGFR2 inhibitors from Persicaria hydropiper in hepatocellular carcinoma. Sci Rep 15, 38071 (2025). https://doi.org/10.1038/s41598-025-21944-0
Received:
Accepted:
Published:
Version of record:
DOI: https://doi.org/10.1038/s41598-025-21944-0
