https://orcid.org/0000-0003-1448-2517
Figures
Abstract
Triple-negative breast cancer (TNBC) is the most aggressive among the breast cancer subtypes and poses unique therapeutic challenges due to its distinct characteristics like lack of specific therapeutic targets. TNBC demonstrates poor survival rate enhanced immunogenic characteristics and a more favorable tumor microenvironment than other breast cancer variants. Also, TNBC patients show elevated levels of programmed death ligand-1 (PD-L1) expression in contrast to non-TNBC patients. Binding of PD-L1 with PD-1 produces an inhibitory signal, resulting in suppression of T-cell. Therapeutic approaches utilizing immunotherapies against PD-L1 exhibit promising outcomes in the treatment of TNBC. Limitations like suboptimal efficacy, inadequate oral bioavailability, and associated immune-related adverse effects of antibody-mediated anti PD-1/PD-L1 therapies have necessitated the exploration of alternative therapeutic approaches. Thus, small molecules become an alternate option for PD-1/PD-L1 inhibition. In the present study, we have used virtual screening to identify potential phytochemicals from selected Indian medicinal plants as PD-L1 inhibitors. A total of 953 phytochemicals derived from eleven selected medicinal plants were initially screened through molecular docking using the PyRx tool. Among the 953 identified phytochemicals, the top 20 compounds exhibiting the highest binding affinities in docking study were selected for further analysis. Following comprehensive ADMET analyses, 2 compounds were ultimately identified as suitable candidates for a molecular dynamics (MD) simulation study. The study identified 4-hydroxychalcone and flavylium from Glycyrrhiza glabra and Catharanthus roseus, respectively as potential PD-L1 inhibitors with enhanced stability relative to the reference molecule. Both compounds also showed enhanced gastrointestinal absorption with no predicted cytotoxic and immunotoxic effects. Consequently, these compounds present promising candidates for novel PD-L1 inhibitor development in TNBC therapy. Further experimental investigations are necessary to facilitate their clinical translation.
Citation: Ahmed SF, Hossain MS, Mondal A, Shahariar M, Sumya SA, Rizu NS, et al. (2025) In silico identification of promising PD-L1 inhibitors from selected indian medicinal plants for treatment of triple negative breast cancer. PLoS One 20(7): e0327475. https://doi.org/10.1371/journal.pone.0327475
Editor: Yusuf Oloruntoyin Ayipo, Kwara State University, NIGERIA
Received: December 15, 2024; Accepted: June 14, 2025; Published: July 10, 2025
Copyright: © 2025 Ahmed et al. This is an open access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.
Data Availability: All relevant data are within the paper and its Supporting Information files.
Funding: The author(s) received no specific funding for this work.
Competing interests: The authors have declared that no competing interests exist.
1. Introduction
Triple-negative breast cancer (TNBC) represents the most aggressive phenotypic subtype within the spectrum of breast cancer (BC) [1]. In TNBC, the minimal cellular expression of progesterone and estrogen receptors is typically at or below 1%, while human growth factor receptor 2 expression ranges from 0 to 1 + . TNBC accounts for 15–20% of all BC cases and has the least favorable five-year overall survival (OS) rate compared to other BC subtypes [2]. TNBC is unique due to its aggressive behavior, molecular complexity in the metastatic process, and lack of effective targeted therapies [1]. Therefore, chemotherapy remains the primary choice for its treatment [3]. While chemotherapy has served as the fundamental therapeutic option for TNBC in recent decades, the treatment is linked to an elevated likelihood of distant metastasis, increased rates of metastasis, and an enhanced propensity for disease relapse, contributing to a rise in chemoresistant individuals. Consequently, extensive research endeavors have been undertaken in the last few years to explore novel therapeutic approaches for TNBC [4,5]. In contrast to other subtypes of BC, TNBC demonstrates increased immunogenicity with a notable abundance of tumor-infiltrating lymphocytes (TILs). The infiltration of TILs establishes a favorable tumor microenvironment (TME) in TNBC, the condition chooses therapeutic approaches involving immune checkpoint inhibitors (ICIs) [6,7]. Amidst the burgeoning realm of checkpoint inhibitors that block the programmed cell death protein 1/programmed death-ligand 1 (PD-1/PD-L1) interaction, these have become a promising therapeutic interest nowadays, as PD-L1 is upregulated in TNBC patients compared to those without TNBC [8]. PD-L1 attenuates the host immune response to malignant tumor cells [9]. Therefore, PD-L1 has become a promising therapeutic target for treating TNBC, particularly in metastatic conditions [1,6,10]. Most ongoing trials emphasize the integration of anti-PD-1/PD-L1 therapy with neoadjuvant chemotherapy, adjuvant therapy, or targeted therapeutic approaches [11,12]. Compared to traditional anti-tumor therapies, the use of monoclonal antibodies (mAbs) has shown superior advantages by extending progression-free survival (PFS) and eliciting strong antitumor immune responses when combined with chemotherapy, radiation, and targeted therapies [1]. In March 2019, the FDA approved the use of Atezolizumab (mAb) in combination with nab-paclitaxel as the initial ICI for BC, specifically for the treatment of adult patients with metastatic TNBC exhibiting high PD-L1 expression [11]. In 2020, pembrolizumab was approved for the treatment of metastatic TNBC patients with PD-L1 expression when used in combination with chemotherapy [13]. The phase I/II clinical trial KEYNOTE-162 assessed the safety and effectiveness of the Poly ADP-ribose polymerase (PARP) inhibitor niraparib in combination with pembrolizumab for treating patients with unresectable locally advanced TNBC or mTNBC. The results indicated that patients with PD-L1-positive tumors responded more favorably than those with PD-L1-negative tumors, with response rates of 33% and 8%, respectively [14]. In July 2021, the drug received approval for an additional use: high-risk, early-stage TNBC. This decision was based on findings from the phase-III KEYNOTE-522 trial (NCT 03036488), which showed that incorporating pembrolizumab into neoadjuvant chemotherapy enhanced pathological complete response rates and event-free survival in early-stage TNBC [15,16]. Besides that, two lncRNAs, such as UCA1 and HCP5, have not yet been recognized concerning the tumor immune response in BC but have the potential to serve as valuable biomarkers for identifying patients suitable for anti-PD-1 antibody therapy [15]. Despite their advantages over traditional methods, the use of mAbs has faced several limitations. The most significant drawback of using mAbs for immune checkpoint blockade is the moderately low response rate observed in the majority of cancer cases, typically ranging from 10% to 30% [17]. This limitation led to the FDA’s revocation of approval for Atezolizumab due to the lack of apparent clinical benefits [18]. Other limitations of ICIs includes lack of oral bioavailability, elevated manufacturing expenses, poor stability, immune-related side effects, minimal tumor penetration, and possible immunogenicity [10]. Furthermore, TNBC imposes a substantial economic burden on both healthcare systems and patients. Patients with TNBC are typically diagnosed at more advanced stages, face a worse prognosis, have a greater likelihood of recurrence, and require more hospital resources and incur higher healthcare costs than those with non-TNBC subtypes. The average annual direct medical costs per patient varied from approximately $20,000 to over $100,000 for stage I–III TNBC and from $100,000 to $300,000 for stage IV TNBC in US. Cancer recurrence resulted in a notable decline in productivity and an increased likelihood of individuals leaving the workforce. The estimated indirect costs from productivity loss varied between $207 and $1573 per patient each month [19]. On the other hand, ineffective chemotherapy and chemotherapy-induced toxicity increase the burden of treatment by requiring greater resource utilization and extended hospital stays. They often lead to undesirable side effects and long-term adverse health consequences, negatively impacting the patient’s quality of life. [1]. Those with TNBC who were young and who received chemotherapy were 10% more likely to experience financial hardship [1,20]. The average monthly costs for patients were more than $1000 in mTNBC and over $2000 in eTNBC during both the neoadjuvant and adjuvant therapy periods in the US. These financial adverse events impact treatment choice, compliance, well-being, and cancer outcomes [21,22]. Therefore, researchers have focused on exploring the potential of small-molecule inhibitors, which are characterized by an enhanced therapeutic index and oral bioavailability.
Immunotherapy based on small molecules can function in the same mechanism as mAbs do without facing the limitations [23]. Furthermore, due to their shorter pharmacokinetic half-life, small molecules have the potential to offer a superior therapeutic index, thereby improving the management of any unforeseen adverse events [24]. On the other hand, the synthetic manufacturing of small molecules is more cost-effective for patients. Small molecule inhibitors designed to target PD-L1 are categorized into two main groups: small molecules inspired by amino acids that replicate the receptor-ligand interface, and the second group consists of compounds structured on the biphenyl scaffold. The pioneering amino acid-inspired interface mimic, CA-170, emerged as the foremost orally administrable small molecule inhibitor of PD-L1 that underwent clinical trials in 2016 and is now progressing through phase-II clinical trials [24]. Researchers from Bristol-Myers Squibb (BMS) have formulated a collection of biphenyl derivatives with substitutions based on their efficacy in impeding the interaction between PD-1 and PD-L1 [24]. Notably, BMS-1001 and BMS-1166 have shown the best safety profiles and remarkable potency in inhibiting the PD-1/PD-L1 interaction [23]. But none of the small-molecule inhibitors of PD-1/PD-L1 pathway have received approval from the FDA [10,23,24]. Throughout history, plant-based bioactive compounds have been used to treat numerous diseases, including cancer. The discovery of natural phytochemicals that inhibit PD-L1 could pave the way for the development of cancer immunotherapeutics [25–27]. In this backdrop, the present study was designed to identify promising natural compounds targeting PD-L1 through virtual screening of a compound database, with the aim of advancing therapeutic strategies for TNBC.
2. Methodology
2.1. Protein preparation
The crystal structure of human PD-L1 protein (PDB ID: 5J89) was retrieved from the RCSB Protein Data Bank (https://www.rcsb.org/) with a resolution of 2.20 Å. The protein was prepared using the Discovery Studio 2024 (https://discover.3ds.com/discovery-studio-visualizer-download) by eliminating the cofactors, ligands, water molecules, and metal ions bound to the protein. Energy minimization was performed using SWISS PDB Viewer (https://spdbv.vital-it.ch/) to stabilize the protein structure [28].
2.2. Ligand retrieval and preparation
In this study, 11 Indian medicinal plants (Table 1) were chosen for anti-PD-L1 drug development based on a literature study [29–35]. The PubChem Compound Identities (CIDs) of all the bioactive metabolites of these 11 plants were retrieved from the IMPPAT database (https://cb.imsc.res.in/imppat/). Following the removal of duplicates, 953 compounds (S1 Table) from these 11 plants were selected for subsequent investigation. The IMPPAT database (https://cb.imsc.res.in/imppat/) is a repository of phytochemical data on Indian medicinal plants [36]. The database utilizes cheminformatics methodologies to evaluate its physicochemical and drug-like properties, employing various scoring methodologies. The 3D structures of the ligands were retrieved from the PubChem database (https://pubchem.ncbi.nlm.nih.gov/) in SDF format [37,38]. We have used Open Babel (https://sourceforge.net/projects/openbabel/) to generate a ligand library [38]. Energy minimization of the ligand compounds was performed using the Universal Force Field (UFF) and the Conjugate Gradients algorithm of PyRx [39].
2.3. Virtual screening and molecular docking
The molecular docking method finds more potent, selective, and efficient drug candidates [44]. Accordingly, virtual screening and blind molecular docking techniques were performed to select a limited number of ligand compounds from the ligand library with desired biological functions, capable of interacting with the binding pocket of the target protein (receptor) [39,45] The ligand compounds were docked against the target protein PD-L1 using PyRx, an open-access virtual screening tool [46]. A molecular docking program, AutoDockVina under the PyRx, was employed for this docking purpose and to estimate the binding affinities of docked complexes. With the help of the calculation of the value of energy minimization and binding energy, molecular docking predicts possible drug-target interactions [47]. During the docking process, all PyRx configurations and parameters were kept at default. The 3D coordinates of the grid box were set as per the co-crystal ligand (active binding site of the reference drug) for structure-based virtual screening. The docked complexes were visualized and analyzed using the BIOVIA visualizer of Discovery Studio 2024 [44].
2.4. ADMET analysis
The top twenty ligand compounds were subjected to pharmacokinetic and drug-likeness evaluation using the SwissADME (www.swissadme.ch) web server, a widely used tool for in silico prediction of ADME (Absorption, Distribution, Metabolism, and Excretion) parameters [47]. Pharmacokinetic properties (PKs) like physicochemical properties, lipophilicity, water solubility Log S (ESOL), pharmacokinetics, drug-likeness rules (Lipinski), and medicinal chemistry (PAINS, Synthetic accessibility) of drugs hold immense significance in pharmacological research for understanding how drugs behave within the body [48]. The Canonical SMILES of each compound were utilized as input to retrieve data on physicochemical properties (e.g., molecular weight, hydrogen bond donors and acceptors), lipophilicity (iLOGP), water solubility (Log S), pharmacokinetic parameters (e.g., gastrointestinal absorption), drug-likeness criteria (e.g., Lipinski’s Rule of Five, bioavailability score), and medicinal chemistry filters (e.g., synthetic accessibility). These properties are critical for assessing the oral bioavailability and therapeutic potential of small molecules [37,47,48].
To ensure the safety profile of the selected compounds, toxicity predictions were conducted using admetSAR 2.0 (http://lmmd.ecust.edu.cn/admetsar2) and Protox-III (https://tox-new.charite.de/) web servers [49,50] web servers. Both servers utilize canonical SMILES as input to predict key toxicological endpoints. The admetSAR 2.0 provides insights into parameters such as AMES mutagenicity, acute oral toxicity, LD50 of Rat, and hERG channel inhibition, while Protox-III evaluates potential carcinogenicity, cytotoxicity, and immunotoxicity of selected ligand compounds.
2.5. Molecular dynamics simulation
Molecular dynamics (MD) simulation is one of the most efficient and frequently used computer techniques for studying the dynamics of macromolecules in biological systems [51]. Therefore, MD simulation was performed in our study to assess the stability of ligand-receptor complexes obtained from molecular docking. We have used YASARA Structure (Product name: YASARA Dynamics) (http://www.yasara.org/products) for MD simulation. MD run was conducted for the ligand and control molecules for 100 ns with 401 snapshots [52]. The homology modeling was carried out using YASARA in many stages. YASARA first determines the modeling parameters and homology modeling target that the macro specifies [53]. The scene mode was then subjected to MD simulations using the default settings of the YASARA Structure macro for MD run (http://www.yasara.org/md_run.mcr) [54]. The MD simulation was conducted using the AMBER (Assisted Model Building with Energy Refinement) force field under the following conditions: a temperature of 298 K, a pressure of 1 bar, Coulomb electrostatics with a cut-off of 7.86, 0.9% NaCl (to neutralize charges), solvent density of 0.997, pH 7.0, 1-fs time phases, periodic boundaries, and all mobile atoms [52,53]. MD simulation was performed using the same methodology as in previous studies [28]. The interaction energy, the root mean square deviation (RMSD), the root mean square fluctuation (RMSF), the radius of gyration (RG), solvent accessible surface area (SASA) and the total number of hydrogen bonds were evaluated from trajectory files.
2.6. MM-PBSA analysis/energy of binding analysis method
Considering the strong correlation of structural changes of the receptor upon ligand binding, the Molecular Mechanics Poisson-Boltzmann Surface Area (MM-PBSA) method was used to calculate the binding affinity of protein-ligand interaction, where a more positive energy value indicates a more favorable binding interaction [55]. Binding free energy assessment methodologies serve as a valuable tool for rescoring docked ligand-protein complexes, providing more accurate estimates of ligand-protein binding affinities compared to basic scoring functions. Different ligand orientations generated by docking software were considered as the initial ligand-receptor coordinates for MD simulations to achieve this goal. The MD simulation was further used for MM-PBSA assessment. The most favorable ligand orientation was subsequently determined by comparing the best binding free energy that was obtained [56,57]. The stable area of the three PD-L1 complexes was used to generate a 100 ns MD trajectory for MM-PBSA calculations. The binding energy components were measured using the MM-PBSA technique in the YASARA simulator (YASARA Biosciences). The g_mmpbsa tool calculates the binding energy of the protein-ligand complex using the following equation.
Where ∆GBinding denotes total binding energy, GComplex denotes the total free energy of binding complex, GProtein and GLigand symbolize the total free energy of the three molecules that are attached to PD-L1, correspondingly [52].
2.7. Principle component analysis
A multivariate statistical method, principal component analysis (PCA), was used to analyze collective motion of the protein-bound ligands during the MD simulation over a 100 ns covariance matrix of backbones atoms was created to conduct the PCA. Various structural and energy information, such as bond lengths, bond angles, dihedral angles, planarity, Van der Waals energies, and electrostatic energies, can be utilized to elucidate the distinctions among various categories [58].Minitab 18 statistical software was used for analysis and plot generation. This tool mitigates complexities linked to elucidating extensive data sets by decreasing their dimensionality and minimizing data loss [59]. The global motion of receptor-ligand complexes was observed by analyzing two principal components, PC1 (projection on eigenvector 1) and PC2 (projection of eigenvector 2).
3. Results
3.1. 3D grid box preparation
PD-L1 is upregulated in TNBC, especially in metastatic conditions. Therefore, it is necessary to inhibit the expression of PD-L1 through the development of suitable drug candidates. For this purpose, we have defined a grid box around the active residues within the binding pocket of PD-L1. The grid box dimensions (Å) were set as center (x = 16.6413, _y = −11.5197, _z = 181.8531), with size x = 39.8515712214, y = 56.774720726, and z = 52.5295007324.
3.2. Virtual screening and molecular docking analyses
We have screened 953 potential bioactive phytochemicals from the ligand library using PyRx. The top 20 compounds were selected for further analysis based on binding affinities between receptor and ligands in terms of binding energy (docking score) in kcal/mol (S1 Table). We have also docked BMS-1166 (PubChem CID-118434635), a control inhibitor of PD-L1, against the target receptor to compare the binding affinities of the selected compounds and observed to have docking score of −7.9 kcal/mol. The selected compounds exhibited higher binding affinities than the control compound, with docking scores ranging from –10.4 to –8.8 kcal/mol, indicating their potential as PD-L1 inhibitors (Table 2). The docked models in PDB format of these 20 protein-ligand complexes are given S1 File.
3.3. Interaction analyses of receptor-ligand and receptor-control complexes
The interactions analysis of the selected docked complexes was carried out using Discover Studio 2024. The various types of interactions mainly observed in these complexes were conventional hydrogen bond, carbon-hydrogen bond, Pi-Sigma, Pi-Pi Stacked, Pi-Pi T-shaped, Pi-Alkyl, Pi-Sulfur, Amide-Pi Stacked, and Pi-Cation (Table 3). These interactions are responsible for the extent of binding affinities and docking scores of docked complexes [60].
3.4. ADMET analysis
The ADME properties of the top 20 ligand compounds, along with the control compound BMS-1166, were predicted using the SwissADME web server (https://www.swissadme.ch). Drug-likeness was assessed based on Lipinski’s Rule of Five, which includes criteria such as molecular weight < 500 Da, hydrogen bond acceptors < 10, hydrogen bond donors < 5, and iLogP < 5 [45]. As presented in Table 4, five compounds (CID-452707, CID-637247, CID-14106343, CID-11597485 and CID-5320092) that violated lipnski rule of 5, with compound Gallotannin (CID 452707) exhibiting the highest number of violations (three). Nonetheless, the majority of compounds, including those with violations, met the oral bioavailability threshold (log Po/w < 5), indicating their potential for drug-like behavior.
The top 20 ligand compounds were then subjected to toxicological profiling using the admetSAR 2.0 web server. As summarized in Table 5, all selected ligands, similar to the control compound, were predicted to be weak inhibitors of the hERG channel and tested negative for AMES mutagenicity. Most compounds were classified as toxicity class III, indicating low toxicity [61]. Further evaluation of endpoints such as carcinogenicity, immunotoxicity, and cytotoxicity was conducted using the ProTox-III web server [47]. While none of the compounds, including the control, were predicted to be carcinogenic, several exhibited immunotoxic and cytotoxic potential. Based on comprehensive ADME and toxicity assessments, two lead compounds- CID 5282361 and CID 145858 were selected for further analysis.
Furthermore, Swiss-ADME was employed to visualize the drug-likeness of final two lead compounds using radar charts. Swiss-ADME represents drug-likeness of lead compounds as a radar chart that determines whether a certain molecule is a viable candidate for a drug. In the radar chart, each peak defines different drug properties, and the pink zone denotes the optimal range for each property of the drug. In this study, almost all the drug properties of final two selected lead compounds floated within the pink zone, indicating their potency as drugs (Fig 1) [48]. The drug-likeness of the control compound was also represented as a radar chart in Fig 1.
3.5 Molecular docking pose assessment
Subsequent to the molecular interaction analysis, the potential binding pose of the two selected compounds were assessed and similar kind of binding pose was demonstrated by these two compounds comparing with the binding pose of control compound, suggesting a conserved binding orientation (Fig 2A, 2C and 2E). The molecular interactions of PD-L1 with these two lead compounds and with control compound were shown in Fig 2B, 2D and 2F, respectively. Subsequently, molecular dynamics (MD) simulations were performed to further validate the stability of these predicted binding conformations (48).
Panels (A), (C), and (E) depict the selected binding poses of PD-L1 in complex with CID-5282361, CID-145858, and the control compound, respectively. Panels (B), (D), and (F) illustrate the corresponding 2D interaction diagrams of PD-L1 with CID-5282361, CID-145858, and the control compound, respectively.
3.6. Dynamic behavior analyses of ligand-receptor complexes by molecular dynamics simulation
For investigation of the inhibitory mechanism of PD-L1, its alteration of dynamic behavior upon ligand binding was assessed through MD simulation [62]. For this purpose, a 100 ns simulation was run for both the receptor-ligand complexes and compared the outcome to that of the receptor-control compound (BMS-1166) complex.
3.6.1. Root means square deviation (RMSD) analyses.
We used the RMSD to check the stability of the ligand protein complexes [63]. The average solute RMSD of PD-L1 was calculated (Fig. 4). The RMSD of average ligand movement was determined after superposing the two ligands (CID145858 and CID-5282361) and control compound (BMS-1166) separately on the PD-L1 receptor and found the values of 1.663 Å, 1.575 Å, and 4.776 Å, respectively. The 100 ns simulation showed that the compounds were stable, not exceeding the 4 Å limit. In contrast, the ligands, CID-145858 and CID-5282361, exhibited optimal range behavior, and their values were likewise lower than those of the control compound, indicating that the two ligand-receptor complexes were more stable (Fig 3).
3.6.2. Root Means Square Fluctuation (RMSF) analyses.
The thermodynamic motions of the protein residues were measured by calculating the Root Means Square Fluctuation (RMSF). This ascertains the amount of residual vibration in PD-L1 upon ligand binding. High RMSF values reveal greater flexibility, whereas low RMSF values show limitations on residue displacement throughout the 100 ns MD simulation, leading to a decrease in flexibility [64]. By analyzing the impact of ligand binding on protein flexibility, the highest fluctuations were observed with time within residues 42–54 and 132–144. On the other hand, complexes of two ligands (CID-145858 and 5282361) and control (BMS-1166) with the receptor showed average fluctuations of 1.8056 Å, 2.006 Å, and 2.169 Å, respectively.. Conversely, the CID-145858 and CID-5282361 molecules exhibited ideal range behavior, and their values were also lower than those of the control compound, suggesting a greater binding probability (Fig 4).
3.6.3. Radius of Gyration (RG) analyses.
RG determines a protein’s compactness and folding behavior based on its tertiary structure and overall conformational state [62]. During simulation, when achieving a stable protein structure, the fluctuation rate becomes a laser [51,62,63]. We have assessed the stability of PD-L1-CID-145858, PD-L1-CID-5282361, and PD-L1-BMS-1166 complexes by calculating the RG for each. The average RG values for the PD-L1-CID-145858 and PD-L1-CID-5282361 complexes were 20.18 Å and 20.184Å, respectively, with higher fluctuations of 21.064 Å and 21.221 Å. The control complex, PD-L1-BMS-1166, exhibited an average RG of 20.72 Å, with a greater fluctuation of 21.868Å (Fig 5). Minimal structural deviations were observed in the PD-L1-CID-145858 and PD-L1-CID-5282361 complexes compared to the control, suggesting overall stability throughout the simulation trajectory.
3.6.4. Solvent-accessible surface area (SASA) analyses.
SASA is a computational approach used to analyze the expansion of the ligand-protein complex’s surface area, which may undergo modifications due to biological interactions between the adsorbate and the solid surface. Alterations in the atomic structure of material surfaces can lead to surface reconstruction, significantly influencing protein conformation [65]. The control compound (BMS-1166) exhibits an average SASA value of 13703.144 Å, whereas the average values for the two selected compounds are 13326.344 Å (CID-145858) and 13407.143Å (CID-5282361), respectively. This indicates that the top two compounds have lower SASA values compared to the control. As no significant fluctuations were observed throughout the simulation period, protein expansion remains within the optimal range following binding of the top two compounds (Fig 6).
3.6.5. Hydrogen bond analyses.
Analyses of hydrogen (H) bonds are important because they bind the ligand to the target protein and control drug selectivity, metabolic processes, and adsorption [66]. In the present study, we calculated the total number of hydrogen bonds of three complexes (CID-145858, 5282361, and BMS-1166) in the solute and between the solute and the solvent over a 100-ns simulation. We observed 164–208, 163–209, and 163–205 numbers of H-bonds of the respective complexes in the solute. Additionally, the number of H-bonds between the solute and the solvent was 439–521, 448–532, and 456–532 for the respective complexes (Fig 7).
3.7. Principal component analysis (PCA) analyses
Principal component analysis often identifies a limited number of principal modes (components) that account for the majority of conformational variability within a dataset. It is widely used to analyze molecular dynamics (MD) trajectories to assess structural similarities and differences among sampled conformations [67]. In this study, PCA was applied to explore long-timescale dynamics of PD-L1 in complex with different ligands. PCA clustering (PC1 and PC2) results were visualized using scatter plots for each replicate. Each circle in the figure (Fig 8A) represents a conformer, with the distribution of circles indicating the conformational changes in the PD-L1 structure upon ligand binding. It was observed from Fig 8A that PCA-1 and PCA-2 collectively account for 80.6% of the variance, where contributions of PCA-1 and PCA-2 were 64.1% and 16.5%, respectively of the fluctuation. The score displayed on the PCA model indicates an overlap between the protein-CID-145858 (red circles) and protein-CID-5282361 (green circles) complexes. Along the positive direction of PCA1-PCA2, the red and green circles form distinct clusters separated from the protein-control cluster, suggesting that ligand binding induces unique and stabilized conformational states. In contrast, the control complex exhibited a more dispersed distribution, indicating higher conformational variability and reduced structural stability. Complementarily, the PCA loading plot (Fig 8B) derived from MD energy profiles and structural data revealed a positive correlation of bond, angle, and Van der Waals (VdW) variables with the ligand-induced conformational shifts, suggesting tighter clusters of PD-L1 complexes bound to CID-145858 and CID-5282361. Furthermore, the clustering in the upper right quadrant of the loading plot represents minor variations in dihedral energy, consistent with the stabilization of the protein-ligand complexes. Additionally, the dPCA (dihedral PCA) results aligned well with the original torsional angle distributions, reinforcing the method’s accuracy in capturing key conformational changes [68].
3.8. MM-PBSA/energy of binding analysis
The YASARA Stimulator’s MM-PBSA/binding energy analysis and Boundary Quick techniques were used to calculate the binding free energy of the top two ligand-receptor complexes and the control compound-PD-L1 complex. For the MM-PBSA calculations, a brief stable region of 100 ns was extracted from the simulated trajectory of the docked complexes using polar and apolar solvation parameters. PD-L1-CID-145858, PD-L1-CID-5282361, and PD-L1-BMS-1166 complexes were found to have average binding energies of 204.892, 117.003, and 69.364 kJ/mol, respectively. The top two compounds were found to form a stable complex with PD-L1 with substantial binding affinity as determined by the MM-PBSA analysis (Fig 9).
4. Discussion
The molecular characteristics and immunogenic nature of TNBC have paved the way for immunotherapy to be considered a promising therapeutic strategy in both advanced and pre-surgical settings. The expression of PD-L1 is observed in approximately 40−65% of TNBC cases and serves as a key predictive biomarker for clinical benefit from anti-PD-L1 monoclonal antibody therapies [69]. However, the comparative efficacy of these small-molecule agents relative to established monoclonal antibodies has yet to be fully characterized, highlighting the necessity for continued rigorous investigation in this area [69–71]. Natural compounds have a longstanding history in the treatment of various human diseases. Notably, approximately 47% of currently available anti-tumor agents are derived from natural sources. Research findings have also indicated that natural compounds have effectively broadened the scope of immunotherapy in treating cold tumors such as TNBC [72]. Therefore, phytochemicals derived from plants are widely considered as viable candidates for cancer treatment [73]. Although the anticancer potential of phytochemicals has been well-documented, their specific effects on the PD-1/PD-L1 interaction remain largely unexplored. In this context, the current study presents a novel investigation into eleven Indian medicinal plants as underexplored sources of PD-L1 inhibitors, highlighting their potential therapeutic role in the treatment of TNBC. While prior research has primarily focused on natural PD-L1 inhibitors, this work uniquely integrates the vast repository of traditional Indian medicinal knowledge with contemporary computational drug discovery approaches, thereby bridging ethnopharmacology and advanced scientific methodologies. Historically, these plants have been predominantly employed worldwide for the management of inflammation and oxidative stress-related disorders [73–80]. Both of these pathological conditions have been associated with the emergence of multiple malignancies, including TNBC. In particular, inflammation has been associated with the increased expression of PD-L1 in TNBC, subsequently restricting the efficacy of ICIs. Furthermore, oxidative stresses may upregulate PD-L1 in tumors via enhanced generation of reactive oxygen species (ROS). Elevated levels of ROS have also been reported in TNBC [81–84].
This study used molecular docking to screen 953 compounds from selected medicinal plants targeting human PD-L1 (PDB ID: 5J89) and BMS-1166 as the reference molecule. Initially, 20 compounds exhibiting higher binding affinities than the control were selected for pharmacokinetic and toxicity analyses. Among these, two compounds - CID-5282361 (Docking score: −9.9 kcal/mol) and CID-145858 (Docking score: −9.5 kcal/mol) demonstrated strong binding affinities toward PD-L1 in molecular docking studies, along with favourable ADMET profiles. Hotspot residues on protein surfaces are crucial for protein-ligand complex formation and serve as key targets in drug design to inhibit protein-protein interactions [85]. In the aim of formulating inhibitors designed to disrupt the PD-1/PD-L1 interaction, several residues of PD-L1, as documented in previous studies, include Tyr 56, Met 115, Ala 121, Asp122-Arg125, Tyr 123, and Ile 54 played significant roles [86–88]. In particular, TYR56 plays a critical role in binding of both antibodies and small molecular inhibitors with PD-L1, whereas MET115 is important for small molecules binding [89]. Investigations have shown that MET115 significantly impacts the conformational state of PD-L1 during its binding to diverse ligands [90]. Analyses of non-bond interactions revealed the presence of both hydrogen and hydrophobic bonds between the protein and the two most promising ligands and the control molecule. Each ligand established hydrophobic bonds with Tyr 56, Met 115, and Ala 121 residues of PD-L1, whereas the control drug interacted with Asp 122, Tyr 123, and Arg 125 residues. Hydrophobic interactions are also reported to amplify the binding affinity between the target and ligand [89]. Structural analysis of human PD-1/PD-L1 demonstrates that both proteins utilize extensive hydrophobic surfaces in their Ig-like V-type domains for ligand interaction. Furthermore, an in-depth analysis of the PD-L1 binding pocket indicates that the tunnel-shaped hydrophobic pocket formed by Tyr56, Met115, Ala121, and Asp122 is significant in ligand binding [86]. PD-L1 mutants like A121Q and Y56A/M115A showed reduced binding to effector cells, highlighting the importance of these residues [91]. Consequently, compounds that establish hydrophobic interactions within this specific region may enhance the inhibitory efficacy of PD-L1 by blocking PD-1/PD-L1 interactions [92]. Hydrophobic interactions are also reported to amplify the binding affinity between the target and ligand [89]. Therefore, our result is in agreement with the previous studies. On the other hand, both compounds exhibited benzene rings, as evidenced by their two-dimensional structural representations. Aromatic rings of compounds were found to enhance the affinity and specificity of drug-like entities by participating in protein-ligand interactions, facilitating non-covalent interactions with hydrophobic residues located within binding sites of the target protein [93,94]. This structural feature is critically important in developing protein kinase inhibitors as aromatic rings correlate positively with binding affinity, balancing the loss of hydrogen bond interactions [94]. In 2023, seven small-molecule protein kinase inhibitors got FDA approval for various cancer types [95]. A study conducted in 2019 further suggested that the occurrence of aromatic rings containing compounds could be anticipated as innovative PD-L1 inhibitors [96].Evaluation of pharmacokinetic and toxicological profiles of the top two lead compounds provided essential insights into their potential efficacy and safety, aiding in the efficient utilization of time and resources [97]. Both selected compounds exhibited lower molecular weight, moderate solubility, and improved gastrointestinal absorption, suggesting their potential suitability as oral drug candidates [89,90]. The toxicity profiles of the lead compounds demonstrate the absence of cytotoxicity and immunotoxicity, both of which were critical factors in the selection of ICI [98,99]. Both these compounds were subjected to 100 ns molecular dynamics simulations to evaluate their stability when bound to their target protein. Although extended simulations may reveal additional insights into conformational changes, we observed stabilization of the system within the 100 ns timeframe of the MD run and successfully fulfilled our objective of identifying the most promising candidates. Furthermore, the 100 ns MD simulation was extensively utilized to evaluate the inhibitory potential of phytochemicals [45,100,101]. Therefore, the MD run was limited to 100 ns. Both complexes exhibited RMSD values converging below 4.0 Å compared to the control complex (approximately 5 Å) after duration of 50 ns. On the other hand, RMSF analysis further demonstrated reduced flexibility (<1.5 Å) for these compounds compared to the control within the hotspot region of PD-L1 (residues 56–66), a finding of particular significance for antibodies and small molecule inhibitors in general [89]. Therefore, RMSD and RMSF from MD simulation further corroborated the stability of both ligand-protein complexes, affirming the inhibitory capabilities against PD-L1. Moreover, both selected compound complexes exhibited a lower mean Rg and reduced flexibility relative to the control complex, suggesting that they promote protein stabilization in a more compact conformation [102]. Consistent with the observed reductions in Rg values, SASA analysis confirmed that the protein expansion upon ligand binding remains within the optimal level. Notably, protein flavylium (CID-145858) complex exhibited the lowest SASA value, indicating minimal solvent exposure. It also demonstrated the highest stability of solute hydrogen bonds and the fewest solvent interactions, suggesting strong and specific binding affinity. This result is consistent with its MM-PBSA binding energy of −204.892 kJ/mol. In contrast, 4-Hydroxychalcone (CID-5282361) showed a slightly higher number of solvent hydrogen bonds, reflecting moderate binding strength with MM-PBSA of −117.003 kJ/mol. The control complex displayed the greatest number of solvent hydrogen bonds, suggesting weak binding specificity, corroborated by its MM-PBSA binding energy of −69.364 kJ/mol. Furthermore, principal component analysis (PCA) revealed more compact conformational clustering for both ligand-bound complexes compared to the control, indicating enhanced structural stability of these two PD-L1 complexes. Based on these findings, the selected two compounds have the potential to serve as lead compounds for inhibiting PD-L1’s biological activity.
5. Conclusions
Targeting Programmed death ligand-1 (PD-L1) represents a potential therapeutic strategy to address the challenges associated with TNBC management. Developing a novel small-molecule inhibitor may offer advantages in overcoming the limitations of antibody-based anti-PD-L1 drugs. Therefore, in the present study, we screened 953 new phytochemicals derived from selected Indian medicinal plants to identify anti-PD-L1 compounds for the treatment of TNBC using computational methodologies. The results of molecular docking analysis identified two compounds, 4-Hydroxychalcone (CID-5282361) and flavylium (CID-145858), exhibiting superior binding affinity to the target protein. Subsequent ADME and toxicity evaluations qualified these two compounds as promising drug candidates. Molecular dynamics simulations demonstrated the stable interaction of these 2 compounds with PD-L1 throughout the simulation period. Therefore, our findings suggest that 4-Hydroxychalcon and flavylium derivatives exhibit potential as PD-L1 inhibitors. Nonetheless, further experimental validation is essential to confirm their efficacy in PD-L1 inhibition and to substantiate their suitability as prospective drug candidates for the treatment of TNBC in the future.
Although these in silico studies offer a cost-effective and efficient approach to screen potential PD-L1 inhibitors, the precision of outcomes can be constrained by the quality of data and the effectiveness of the computational algorithms employed. Furthermore, the reliability of these predictions can be compromised by the inherent flexibility and flatness of the PD-L1 interacting surfaces [88]. Due to unforeseen biological factors, virtual screening may generate false positives when predicted active compounds fail to demonstrate efficacy in experimental validation [45]. Therefore, to increase the reliability of in silico results and support the transition of computational predictions into therapeutic use for TNBC, it is essential to validate these two compounds through wet lab experiments. Techniques such as Western Blot analysis and Flow Cytometry can be employed to assess changes in PD-L1 protein levels and quantify PD-L1 expression on TNBC cells following treatment with two compounds. MTT assay using the MDA-MB-231 cell line can also be used to assess the cytotoxic effects of these identified compounds. Furthermore, using TNBC xenograft models, the anti-tumor efficacy of compounds can be evaluated by monitoring tumor growth and metastasis. Lastly, transcriptomics analyses can be used to assess gene expression changes associated with immune evasion mechanisms.
These findings underscore the importance of integrating computational and experimental approaches in drug discovery. While computational methods offer valuable insights and can accelerate the initial stages of lead compound identification, they must be complemented by rigorous experimental validation to ensure the reliability and efficacy of potential drug candidates. This synergistic approach not only enhances the accuracy of predictions but also addresses the inherent limitations of computational techniques.
It is important to note that despite the advantages of using small molecules for targeted therapy, there are also several challenges. Phytochemicals can engage with various molecular targets, potentially causing unintended side effects that make their development as selective PD-L1 inhibitors more challenging [26]. Even if a phytochemical appears promising, establishing cost-efficient and scalable methods for its extraction or synthesis is crucial for clinical use. On the other hand, patient responses to phytochemical treatments can be affected by genetic and environmental factors, highlighting the need for personalized therapeutic approaches. Most targeted anti-cancer drugs develop resistance after some time. Targeted anticancer drugs also face low efficiency challenges [103]. With a comprehensive grasp of tumor pathology and advancements in new drug research and development technologies, we anticipate that more innovative small-molecule anti-cancer drugs targeting novel genes or mechanisms of action will emerge soon. Additionally, it is predicted that new fields, such as the integration of small-molecule targeted drugs with tumor immunotherapy, antibody-drug conjugate (ADC), and Proteolysis targeting chimeric (PROTAC) technology, will experience significant growth over the next decade.
Supporting information
S1 Table. List of natural compounds and their plant sources showing binding affinity in Kcal/mol with the receptor.
https://doi.org/10.1371/journal.pone.0327475.s001
(XLSX)
S1 File. PDB of top 20 protein-ligand docked complexes along with protein-control docked complex.
https://doi.org/10.1371/journal.pone.0327475.s002
(ZIP)
References
- 1. Gupta GK, Collier AL, Lee D, Hoefer RA, Zheleva V, Siewertsz van Reesema LL, et al. Perspectives on triple-negative breast cancer: current treatment strategies, unmet needs, and potential targets for future therapies. Cancers (Basel). 2020;12(9):2392. pmid:32846967
- 2. Almansour NM. triple-negative breast cancer: a brief review about epidemiology, risk factors, signaling pathways, treatment and role of artificial intelligence. Front Mol Biosci. 2022;9:836417. pmid:35145999
- 3. Noor ZS, Master A. Updates on targeted therapy for triple-negative breast cancer (TNBC). Curr Breast Cancer Rep. 2018;10(4):282–8.
- 4. Obidiro O, Battogtokh G, Akala EO. Triple negative breast cancer treatment options and limitations: future outlook. Pharmaceutics. 2023;15(7):1796. pmid:37513983
- 5. Neophytou C, Boutsikos P, Papageorgis P. Molecular mechanisms and emerging therapeutic targets of triple-negative breast cancer metastasis. Front Oncol. 2018;8:31. pmid:29520340
- 6. Loizides S, Constantinidou A. Triple negative breast cancer: immunogenicity, tumor microenvironment, and immunotherapy. Front Genet. 2023;13:1095839. pmid:36712858
- 7. Liu Y, Hu Y, Xue J, Li J, Yi J, Bu J, et al. Advances in immunotherapy for triple-negative breast cancer. Mol Cancer. 2023;22(1):145. pmid:37660039
- 8. Krishnamoorthy HR, Karuppasamy R. A multitier virtual screening of antagonists targeting PD-1/PD-L1 interface for the management of triple-negative breast cancer. J Cancer Edu. 2023;40.
- 9. Han Y, Liu D, Li L. PD-1/PD-L1 pathway: current researches in cancer. 2020. www.ajcr.us/
- 10. Marra A, Trapani D, Viale G, Criscitiello C, Curigliano G. Practical classification of triple-negative breast cancer: intratumoral heterogeneity, mechanisms of drug resistance, and novel therapies. NPJ Breast Cancer. 2020;6:54. pmid:33088912
- 11. Twomey JD, Zhang B. Cancer immunotherapy update: FDA-approved checkpoint inhibitors and companion diagnostics. AAPS J. 2021;23(2):39. pmid:33677681
- 12. Yang T, Li W, Huang T, Zhou J. Immunotherapy targeting PD-1/PD-L1 in early-stage triple-negative breast cancer. J Pers Med. 2023;13(3):526. pmid:36983708
- 13. Heeke AL, Tan AR. Checkpoint inhibitor therapy for metastatic triple-negative breast cancer. Cancer Immunol Immunother. 2021.
- 14. Ren Y, Song J, Li X, Luo N. Rationale and clinical research progress on PD-1/PD-L1-based immunotherapy for metastatic triple-negative breast cancer. Int J Mol Sci. 2022;23(16):8878. pmid:36012144
- 15. Mathias C, Kozak VN, Magno JM, Baal SCS, Dos Santos VHA, Ribeiro EM de SF, et al. PD-1/PD-L1 inhibitors response in triple-negative breast cancer: can long noncoding RNAs be associated?. Cancers (Basel). 2023;15(19):4682. pmid:37835376
- 16. Schmid P, Cortes J, Pusztai L, McArthur H, Kümmel S, Bergh J, et al. Pembrolizumab for early triple-negative breast cancer. N Engl J Med. 2020;382(9):810–21. pmid:32101663
- 17. He X, Xu C. Immune checkpoint signaling and cancer immunotherapy. Cell Res. 2020;30(8):660–9. pmid:32467592
- 18. Debien V, De Caluwé A, Wang X, Piccart-Gebhart M, Tuohy VK, Romano E, et al. Immunotherapy in breast cancer: an overview of current strategies and perspectives. NPJ Breast Cancer. 2023;9(1):7. pmid:36781869
- 19. Huang M, Haiderali A, Fox GE, Frederickson A, Cortes J, Fasching PA, et al. Economic and humanistic burden of triple-negative breast cancer: a systematic literature review. Pharmacoeconomics. 2022;40(5):519–58. pmid:35112331
- 20. Zagami P, Carey LA. Triple negative breast cancer: Pitfalls and progress. NPJ Breast Cancer. 2022;8(1):95. pmid:35987766
- 21. Sieluk J, Haiderali A, Huang M, Hirshfield K, Signorovitch J, Song Y, et al. PCN240 healthcare resource utilization associated with disease recurrence among surgically-treated patients with triple-negative breast cancer. Value in Health. 2021;24:S65.
- 22. ’Sieluk J, ’Yang L, ’Haiderali A, ’Huang M, ’Hirshfield KM. Systemic therapy, survival and end-of-life costs for metastatic triple-negative breast cancer: retrospective SEER-Medicare study of women age ≥65 years. Future Oncol. 2021;17(20).
- 23. Chen S, Song Z, Zhang A. Small-molecule immuno-oncology therapy: advances, challenges and new directions. Curr Top Med Chem. 2019;19(3):180–5. pmid:30854972
- 24. Sasikumar PG, Ramachandra M. Small molecule agents targeting PD-1 checkpoint pathway for cancer immunotherapy: mechanisms of action and other considerations for their advanced development. Front Immunol. 2022;13:752065. pmid:35585982
- 25. Wu T-N, Chen H-M, Shyur L-F. Current advancements of plant-derived agents for triple-negative breast cancer therapy through deregulating cancer cell functions and reprogramming tumor microenvironment. Int J Mol Sci. 2021;22(24):13571. pmid:34948368
- 26. Nakhjavani M, Shigdar S. Natural blockers of PD-1/PD-L1 interaction for the immunotherapy of triple-negative breast cancer-brain metastasis. Cancers (Basel). 2022;14(24):6258. pmid:36551742
- 27. ’Alharthia NS, et al. In silico assessment of a natural small molecule as an inhibitor of programmed death ligand 1 for cancer immunotherapy: a computational approach. J Biomol Struct Dynam. 2024.
- 28. Guex N, Peitsch MC, Schwede T. Automated comparative protein structure modeling with SWISS-MODEL and Swiss-PdbViewer: a historical perspective. Electrophoresis. 2009;30 Suppl 1:S162-73. pmid:19517507
- 29. Wahab S, Annadurai S, Abullais SS, Das G, Ahmad W, Ahmad MF, et al. Glycyrrhiza glabra (Licorice): a comprehensive review on its phytochemistry, biological activities, clinical evidence and toxicology. Plants (Basel). 2021;10(12):2751. pmid:34961221
- 30. Ahmad B, Hafeez N, Rauf A, Bashir S, Linfang H, Rehman M, et al. Phyllanthus emblica: a comprehensive review of its therapeutic benefits. South African J Botany. 2021;138:278–310.
- 31. Ilango S, Sahoo DK, Paital B, Kathirvel K, Gabriel JI, Subramaniam K, et al. A review on annona muricata and its anticancer activity. Cancers (Basel). 2022;14(18):4539. pmid:36139697
- 32. Saleem S, Muhammad G, Hussain MA, Altaf M, Bukhari SNA. Withania somnifera L.: insights into the phytochemical profile, therapeutic potential, clinical trials, and future prospective. Iran J Basic Med Sci. 2020;23(12):1501–26. pmid:33489024
- 33. Pham HNT, Vuong QV, Bowyer MC, Scarlett CJ. Phytochemicals derived from catharanthus roseus and their health benefits. Technologies. 2020;8(4):80.
- 34. Mwangi RW, Macharia JM, Wagara IN, Bence RL. The medicinal properties of Cassia fistula L: a review. Biomed Pharmacother. 2021;144:112240. pmid:34601194
- 35. Missoum A. An update review on Hibiscus rosa sinensis phytochemistry and medicinal uses. J Ayurvedic Herbal Med. 2018;4(3):135–46.
- 36. Mohanraj K, Karthikeyan BS, Vivek-Ananth RP, Chand RPB, Aparna SR, Mangalapandi P, et al. IMPPAT: a curated database of Indian medicinal plants, phytochemistry and therapeutics. Sci Rep. 2018;8(1):4329. pmid:29531263
- 37. Kim S, Thiessen PA, Bolton EE, Chen J, Fu G, Gindulyte A, et al. PubChem substance and compound databases. Nucleic Acids Res. 2016;44(D1):D1202-13. pmid:26400175
- 38. O’Boyle NM, Banck M, James CA, Morley C, Vandermeersch T, Hutchison GR. Open Babel: an open chemical toolbox. J Cheminform. 2011;3:33. pmid:21982300
- 39. Dallakyan S, Olson AJ. Small-molecule library screening by docking with PyRx. Methods Mol Biol. 2015;1263:243–50. pmid:25618350
- 40. Padayachee B, Baijnath H. An updated comprehensive review of the medicinal, phytochemical and pharmacological properties of Moringa oleifera. South African J Bot. 2020;129:304–16.
- 41.
Shehzad A, Qayyum A, Rehman R, Nadeem F, Shehzad MR. A review of bioactivity guided medicinal uses and therapeutic potentials of noxious weed (Alternanthera sessilis). 2018.
- 42. ’Chander M. Rauwolfia serpentina: a comprehensive phytochemical study of its bioactive metabolites. STM J. 2024;01(02):25–50.
- 43. Okoro BC, Dokunmu TM, Okafor E, Sokoya IA, Israel EN, Olusegun DO, et al. The ethnobotanical, bioactive compounds, pharmacological activities and toxicological evaluation of garlic (Allium sativum): a review. Pharmacol Res Mod Chinese Med. 2023;8:100273.
- 44. Chowdhury P. In silico investigation of phytoconstituents from Indian medicinal herb “Tinospora cordifolia (giloy)” against SARS-CoV-2 (COVID-19) by molecular dynamics approach. J Biomol Struct Dyn. 2021;39(17):6792–809. pmid:32762511
- 45. Alanzi A, Moussa AY, Mothana RA, Abbas M, Ali I. In silico exploration of PD-L1 binding compounds: structure-based virtual screening, molecular docking, and MD simulation. PLoS One. 2024;19(8):e0306804. pmid:39121024
- 46.
Tangyuenyongwatana P, Sapjaroen P. Study of steric factor on blind docking of phenylbutanoid dimers with PyRx 0.8 virtual screening tool. In: The 35 th International annual meeting in pharmaceutical sciences & CU-MPU International collaborative research conference study of steric factor on blind docking of phenylbutanoid dimers with PyRx 0.8 virtual screening tool. 2019. http://www.pharm.chula.ac.th/am2019/
- 47. Roy AS, Sawrav MSS, Hossain MS, Johura FT, Ahmed SkF, Hami I, et al. In silico identification of potential inhibitors with higher potency than bumetanide targeting NKCC1: an important ion co-transporter to treat neurological disorders. Inform Med Unlocked. 2021;27:100777.
- 48. Luo L, Zhong A, Wang Q, Zheng T. Structure-based pharmacophore modeling, virtual screening, molecular docking, ADMET, and molecular dynamics (MD) simulation of potential inhibitors of PD-L1 from the library of marine natural products. Mar Drugs. 2021;20(1):29. pmid:35049884
- 49. Yang H, Lou C, Sun L, Li J, Cai Y, Wang Z, et al. admetSAR 2.0: web-service for prediction and optimization of chemical ADMET properties. Bioinformatics. 2019;35(6):1067–9. pmid:30165565
- 50. Waring MJ. Lipophilicity in drug discovery. Expert Opin Drug Discov. 2010;5(3):235–48. pmid:22823020
- 51. Alonso H, Bliznyuk AA, Gready JE. Combining docking and molecular dynamic simulations in drug design. Med Res Rev. 2006;26(5):531–68. pmid:16758486
- 52. Verma AK, Majid A, Hossain MS, Ahmed SF, Ashid M, Bhojiya AA, et al. Identification of 1, 2, 4-Triazine and Its derivatives against lanosterol 14-Demethylase (CYP51) property of candida albicans: influence on the development of new antifungal therapeutic strategies. Front Med Technol. 2022;4:845322. pmid:35419560
- 53. Land H, Humble MS. YASARA: a tool to obtain structural guidance in biocatalytic investigations. Methods Mol Biol. 2018;1685:43–67. pmid:29086303
- 54. Prasasty VD, Istyastono EP. Structure-based design and molecular dynamics simulations of pentapeptide AEYTR as a potential acetylcholinesterase inhibitor. Indones J Chem. 2020;20(4):953.
- 55. Wang C, Nguyen PH, Pham K, Huynh D, Le T-BN, Wang H, et al. Calculating protein-ligand binding affinities with MMPBSA: method and error analysis. J Comput Chem. 2016;37(27):2436–46. pmid:27510546
- 56. Tuccinardi T. What is the current value of MM/PBSA and MM/GBSA methods in drug discovery?. Expert Opin Drug Discov. 2021;16(11):1233–7. pmid:34165011
- 57. Poli G, Granchi C, Rizzolio F, Tuccinardi T. Application of MM-PBSA methods in virtual screening. Molecules. 2020;25(8):1971. pmid:32340232
- 58. Verma AK, Ahmed SF, Hossain MS, Bhojiya AA, Upadhyay SK, Srivastava AK, et al. Unlocking SGK1 inhibitor potential of bis-[1-N,7-N, pyrazolo tetraethoxyphthalimido{-4-(3,5-Dimethyl-4-(spiro-3-methylpyazolo)-1,7-dihydro-1H-dipyrazolo[3,4-b;4’,3’-e]pyridin-8-yl)}]p-disubstituted phenyl compounds: a computational study. J Biomol Struct Dyn. 2022;40(24):13412–31. pmid:34696688
- 59. Nag A, Dhull N, Gupta A. Evaluation of tea (Camellia sinensis L.) phytochemicals as multi-disease modulators, a multidimensional in silico strategy with the combinations of network pharmacology, pharmacophore analysis, statistics and molecular docking. Mol Divers. 2023;27(1):487–509. pmid:35536529
- 60. Thillainayagam M, Malathi K, Ramaiah S. In-Silico molecular docking and simulation studies on novel chalcone and flavone hybrid derivatives with 1, 2, 3-triazole linkage as vital inhibitors of Plasmodium falciparum dihydroorotate dehydrogenase. J Biomol Struct Dyn. 2018;36(15):3993–4009. pmid:29132266
- 61. Yadav R, Imran M, Dhamija P, Chaurasia DK, Handu S. Virtual screening, ADMET prediction and dynamics simulation of potential compounds targeting the main protease of SARS-CoV-2. J Biomol Struct Dyn. 2021;39(17):6617–32. pmid:32715956
- 62. Verma AK, Ahmed SF, Hossain MS, Bhojiya AA, Mathur A, Upadhyay SK, et al. Molecular docking and simulation studies of flavonoid compounds against PBP-2a of methicillin-resistant Staphylococcusaureus. J Biomol Struct Dyn. 2022;40(21):10561–77. pmid:34243699
- 63. Koushki EH, Abolghasemi S, Mollica A, Aghaeepoor M, Moosavi SS, Farshadfar C, et al. Structure-based virtual screening, molecular docking and dynamics studies of natural product and classical inhibitors against human dihydrofolate reductase. Netw Model Anal Health Inform Bioinforma. 2020;9(1).
- 64. Zarezade V, Abolghasemi M, Rahim F, Veisi A, Behbahani M. In silico assessment of new progesterone receptor inhibitors using molecular dynamics: a new insight into breast cancer treatment. J Mol Model. 2018;24(12):337. pmid:30415281
- 65. Ozboyaci M, Kokh DB, Corni S, Wade RC. Modeling and simulation of protein-surface interactions: achievements and challenges. Q Rev Biophys. 2016;49:e4. pmid:26821792
- 66. Luzar A. Resolving the hydrogen bond dynamics conundrum. The J Chem Phy. 2000;113(23):10663–75.
- 67. Groth D, Hartmann S, Klie S, Selbig J. Principal components analysis. Methods Mol Biol. 2013;930:527–47. pmid:23086856
- 68. Altis A, Nguyen PH, Hegger R, Stock G. Dihedral angle principal component analysis of molecular dynamics simulations. J Chem Phys. 2007;126(24):244111. pmid:17614541
- 69. Valencia GA, Rioja P, Morante Z, Ruiz R, Fuentes H, Castaneda CA, et al. Immunotherapy in triple-negative breast cancer: a literature review and new advances. World J Clin Oncol. 2022;13(3):219–36. pmid:35433291
- 70. Guzik K, Tomala M, Muszak D, Konieczny M, Hec A, Błaszkiewicz U, et al. Development of the inhibitors that target the PD-1/PD-L1 interaction-a brief look at progress on small molecules, peptides and macrocycles. Molecules. 2019;24(11):2071. pmid:31151293
- 71. Deng J, Cheng Z, Long J, Dömling A, Tortorella M, Wang Y. Small molecule inhibitors of programmed cell death ligand 1 (PD-L1): a patent review (2019-2021). Expert Opin Ther Pat. 2022;32(5):575–89. pmid:35272536
- 72. Dong S, Guo X, Han F, He Z, Wang Y. Emerging role of natural products in cancer immunotherapy. Acta Pharm Sin B. 2022;12(3):1163–85. pmid:35530162
- 73. Dehelean CA, Marcovici I, Soica C, Mioc M, Coricovac D, Iurciuc S, et al. Plant-derived anticancer compounds as new perspectives in drug discovery and alternative therapy. Molecules. 2021;26(4):1109. pmid:33669817
- 74. Bao F, Bai H-Y, Wu Z-R, Yang Z-G. Phenolic compounds from cultivated Glycyrrhiza uralensis and their PD-1/PD-L1 inhibitory activities. Nat Prod Res. 2021;35(4):562–9. pmid:30908097
- 75. Aumeeruddy MZ, Mahomoodally MF. Global documentation of traditionally used medicinal plants in cancer management: a systematic review. South African J Bot. 2021;138:424–94.
- 76. Rahman MR, Bashar SB, Rifat RH, Poran MdS, Rahman MdA, Islam F, et al. Medicinal plants with anticancer effects available in Bangladesh: a review. J Pharmacogn Phytochem. 2021;10(3):41–9.
- 77. Kabidul Azam MN, Rahman MM, Biswas S, Ahmed MN. Appraisals of Bangladeshi medicinal plants used by folk medicine practitioners in the prevention and management of malignant neoplastic diseases. Int Sch Res Notices. 2016;2016:7832120. pmid:27382642
- 78. Mondal H, Saha S, Awang K, Hossain H, Ablat A, Islam MK, et al. Central-stimulating and analgesic activity of the ethanolic extract of Alternanthera sessilis in mice. BMC Complement Altern Med. 2014;14:398. pmid:25315440
- 79. Hasan MK, Ara I, Mondal MSA, Kabir Y. Phytochemistry, pharmacological activity, and potential health benefits of Glycyrrhiza glabra. Heliyon. 2021;7(6):e07240. pmid:34189299
- 80. Sharma A, Kumar A, Jaitak V. Pharmacological and chemical potential of Cassia fistula L- a critical review. J Herbal Med. 2021;26:100407.
- 81. Waldum H, Fossmark R. Inflammation and digestive cancer. Int J Mol Sci. 2023;24(17):13503. pmid:37686307
- 82. Baram T, Oren N, Erlichman N, Meshel T, Ben-Baruch A. Inflammation-driven regulation of PD-L1 and PD-L2, and their cross-interactions with protective soluble TNFα receptors in human triple-negative breast cancer. Cancers (Basel). 2022;14(14):3513. pmid:35884574
- 83. Bailly C. Regulation of PD-L1 expression on cancer cells with ROS-modulating drugs. Life Sci. 2020;246:117403. pmid:32035131
- 84. Al Haq AT, Tseng H-Y, Chen L-M, Wang C-C, Hsu H-L. Targeting prooxidant MnSOD effect inhibits triple-negative breast cancer (TNBC) progression and M2 macrophage functions under the oncogenic stress. Cell Death Dis. 2022;13(1):49. pmid:35017469
- 85.
Zhang F, et al. Structural basis of the therapeutic anti-PD-L1 antibody atezolizumab. 2019.
- 86. Sasmal P, Kumar Babasahib S, Prashantha Kumar BR, Manjunathaiah Raghavendra N. Biphenyl-based small molecule inhibitors: Novel cancer immunotherapeutic agents targeting PD-1/PD-L1 interaction. Bioorg Med Chem. 2022;73:117001. pmid:36126447
- 87. Guo Y, Jin Y, Wang B, Liu B. Molecular mechanism of small-molecule inhibitors in blocking the PD-1/PD-L1 pathway through PD-L1 dimerization. Int J Mol Sci. 2021;22(9):4766. pmid:33946261
- 88. Fantacuzzi M, Paciotti R, Agamennone M. A Comprehensive computational insight into the PD-L1 Binding to PD-1 and small molecules. Pharmaceuticals (Basel). 2024;17(3):316. pmid:38543102
- 89. Lim H, Chun J, Jin X, Kim J, Yoon J, No KT. Investigation of protein-protein interactions and hot spot region between PD-1 and PD-L1 by fragment molecular orbital method. Sci Rep. 2019;9(1):16727. pmid:31723178
- 90. Shi D, Zhou S, Liu X, Zhao C, Liu H, Yao X. Understanding the structural and energetic basis of PD-1 and monoclonal antibodies bound to PD-L1: a molecular modeling perspective. Biochim Biophys Acta Gen Subj. 2018;1862(3):576–88. pmid:29203283
- 91.
Skalniak L, et al. Small-molecule inhibitors of PD-1/PD-L1 immune checkpoint alleviate the PD-L1-induced exhaustion of T-cells. 2017.
- 92. Zak KM, Grudnik P, Magiera K, Dömling A, Dubin G, Holak TA. Structural biology of the immune checkpoint receptor PD-1 and its ligands PD-L1/PD-L2. Structure. 2017;25(8):1163–74. pmid:28768162
- 93. Lanzarotti E, Defelipe LA, Marti MA, Turjanski AG. Aromatic clusters in protein-protein and protein-drug complexes. J Cheminform. 2020;12(1):30. pmid:33431014
- 94. Zhu Y, Alqahtani S, Hu X. Aromatic rings as molecular determinants for the molecular recognition of protein kinase inhibitors. Molecules. 2021;26(6):1776. pmid:33810025
- 95. Roskoski R Jr. Properties of FDA-approved small molecule protein kinase inhibitors: a 2024 update. Pharmacol Res. 2024;200:107059. pmid:38216005
- 96. Almahmoud S, Zhong HA. Molecular modeling studies on the binding mode of the PD-1/PD-L1 complex inhibitors. Int J Mol Sci. 2019;20(18):4654. pmid:31546905
- 97. Jia C-Y, Li J-Y, Hao G-F, Yang G-F. A drug-likeness toolbox facilitates ADMET study in drug discovery. Drug Discov Today. 2020;25(1):248–58. pmid:31705979
- 98. Donati G, Viviano M, D’Amore VM, Cipriano A, Diakogiannaki I, Amato J, et al. A combined approach of structure-based virtual screening and NMR to interrupt the PD-1/PD-L1 axis: Biphenyl-benzimidazole containing compounds as novel PD-L1 inhibitors. Arch Pharm (Weinheim). 2024;357(3):e2300583. pmid:38110703
- 99. Xu J. Progress in small-molecule inhibitors targeting PD-L1. RSC Med Chem. 2023;:1–15.
- 100. Mittal L, Tonk RK, Awasthi A, Asthana S. Targeting cryptic-orthosteric site of PD-L1 for inhibitor identification using structure-guided approach. Arch Biochem Biophys. 2021;713.
- 101. Binshaya AS, Alkahtani OS, Aldakheel FM, Hjazi A, Almasoudi HH. Structure-based multitargeted docking screening, pharmacokinetics, DFT, and dynamics simulation studies reveal mitoglitazone as a potent inhibitor of cellular survival and stress response proteins of lung cancer. Med Oncol. 2024;41(5):101. pmid:38546811
- 102. Rosilan NF, Jamali MAM, Sufira SA, Waiho K, Fazhan H, Ismail N, et al. Molecular docking and dynamics simulation studies uncover the host-pathogen protein-protein interactions in Penaeus vannamei and Vibrio parahaemolyticus. PLoS One. 2024;19(1):e0297759. pmid:38266027
- 103. Zhong L, Li Y, Xiong L, Wang W, Wu M, Yuan T, et al. Small molecules in targeted cancer therapy: advances, challenges, and future perspectives. Signal Transduct Target Ther. 2021;6(1):201. pmid:34054126