2.1 Pre-processing & Docking results
Obtained crystal coordinates of an extracellular region of EGFR and PubChem compounds were pre-processed for molecular docking studies. The best-established compound was Lapatinib (PubChem ID: 208908) based on docking tests of complete pre-established 16 medicines [Table 2]. Lapatinib has a high-affinity score for our target protein, with a molecular weight of 581.058 g/mol, a hydrogen bond donor count of 2 and a hydrogen bond acceptor count of 9, a topological polar surface area of 115A2, and a logP value of 5.9. In Glioblastoma, the chemical Lapatinib exhibits a higher inhibitory affinity on the protein EGFR.
Table 1
List of inhibitors by survey of literature.
Inhibitor | Pub ID | MW | HBD | HBA | LogP | ref no. |
Gefitinib | 123631 | 446.9 g/mol | 1 | 8 | 4.1 | 33 |
Erlotinib | 176870 | 393.4 g/mol | 1 | 7 | 3.3 | 34 |
Lapatinib | 208908 | 581.1 g/mol | 2 | 9 | 5.9 | 35 |
capecitabine | 60953 | 359.35 g/mol | 3 | 7 | 0.6 | 36 |
cyclopamine | 442972 | 411.6 g/mol | 2 | 3 | 3.5 | 38 |
Afatinib | 10184653 | 485.9 g/mol | 2 | 8 | 3.6 | 39 |
Temozolomide | 5394 | 194.15 g/mol | 1 | 5 | -1.1 | 40 |
carmustine | 2578 | 214.05 g/mol | 1 | 3 | 1.5 | 41 |
Cetuximab | 91820602 | 354.3 g/mol | 2 | 3 | 4 | 42 |
Dacomitinib | 11511120 | 469.9 g/mol | 2 | 7 | 4.4 | 43 |
AEE788 | 10297043 | 440.6 g/mol | 2 | 5 | 4.6 | 45 |
Tarceva (OSI-774) | 176870 | 393.4 g/mol | 1 | 7 | 3.3 | 46 |
Temodar | 5394 | 194.15 g/mol | 1 | 5 | -1.1 | 47 |
carboplatin | 426756 | 371.25 g/mol | 4 | 6 | 0.4 | 48 |
procarbazine | 4915 | 221.3 g/mol | 3 | 3 | 0.1 | 49 |
vandetinib | 3081361 | 475.4 g/mol | 1 | 7 | 4.9 | 50 |
Table 2
Docking results of established compounds against EGFR.
PubChem ID | MolDock Score | Rerank Score | HBond | MW |
208908 | -151.764 | -113.707 | -4.15386 | 581.058 |
176870 | -145.666 | -110.522 | -4.31581 | 393.436 |
176870 | -139.421 | -104.214 | -4.62263 | 393.436 |
11511120 | -127.929 | -100.121 | -2.0006 | 469.939 |
123631 | -131.007 | -98.9755 | -3.45071 | 446.902 |
60953 | -130.411 | -98.9504 | -6.47584 | 359.35 |
10297043 | -126.492 | -97.9906 | -2.30242 | 440.583 |
3081361 | -143.034 | -96.9528 | -2.74 | 475.354 |
3081361 | -129.968 | -96.8904 | -2.40655 | 475.354 |
442972 | -121.699 | -96.8892 | -1.69628 | 411.62 |
176870 | -141.357 | -96.4829 | -2.04013 | 393.436 |
2.2 Virtual screened results
Virtual screening uses high-performance computation to screen vast chemical libraries for potent lead compounds. For the chemical Lapatinib, an advanced similarity search yielded 407 results. The top 10 docking results of the full 407 virtual screened compounds are shown in [Table 3]. The molecule PubChem ID: 59671768, SCHEMBL1427033 is shown in [Fig. 2(II)] as a high-affinity compound with the lowest re-rank score. This molecule has a molecular weight of 581.058 g/mol, a topological surface area of 116A2, and a logP value of 6.1. It has two hydrogen bond donors and ten hydrogen bond acceptors. Among these 407 compounds, the medication with PubChem CID: 59671768 has a lot of potential as a Glioblastoma inhibitor over the EGFR target protein.
Table 3
Molecular Docking results of virtual screened compounds.
Ligand PubChem ID | MolDock Score | Rerank Score | HBond | MW |
59671768 | -198.554 | -137.713 | -3.60625 | 581.058 |
118717618 | -209.355 | -136.801 | -8.206 | 667.147 |
73713599 | -189.563 | -134.892 | -3.6824 | 645.143 |
22017830 | -170.591 | -134.607 | 0 | 609.068 |
51002442 | -179.238 | -134.565 | -5.44535 | 581.058 |
68178766 | -189.388 | -133.297 | -6.46064 | 581.058 |
123656076 | -199.356 | -132.452 | -6.42519 | 597.1 |
142038799 | -185.028 | -132.077 | -3.80955 | 613.144 |
117807540 | -185.92 | -131.928 | -5.51334 | 583.073 |
145585441 | -178.826 | -131.251 | -6.64474 | 580.05 |
44532710 | -184.249 | -130.587 | -5.35679 | 595.084 |
143782896 | -190.146 | -130.19 | -2.40091 | 581.124 |
44199881 | -188.898 | -130.106 | -7.01001 | 581.058 |
58729910 | -193.099 | -129.403 | -5.12112 | 567.031 |
149028802 | -169.044 | -129.125 | -4.56433 | 595.084 |
143924483 | -176.155 | -129.104 | -9.4083 | 627.083 |
22017830 | -197.361 | -128.844 | -6.58476 | 609.068 |
57347591 | -168.052 | -127.905 | -2.36981 | 581.058 |
2.3 Molecular Dynamics Simulation
We have simulated ligand-protein complexes with high affinity and low binding energy for 100 ns to compute the RMSD and RMSF with the targeted protein (PDB ID: 5XWD)[41] [Fig. 3, (I-IV)]. The protein and ligand root mean square deviations, which are commonly employed to measure the scalar distance between the protein (C-alpha backbone) and the ligand throughput trajectory, were utilized to calculate the MD of protein-ligand complexes. According to Fig. 3(I), the protein's root mean square deviation remained steady in the presence of an established ligand (PubChem ID: 208908) at 20 ns, later it has shown the highest variations around 14.35 Å between 20 to 80 ns. But at the end of the simulation, the RMSD was noted with the lowest value of ~ 12 Å, & less fluctuations of stability. RMSF (root mean square fluctuation) analysis is to reveal the flexibility of the movement of the atom while the interaction of the ligand, graphical representation shown in Fig. 3. Overall average RMSF of established compounds with protein was obtained at 4.5–5.0 Å, observed with high differences, shown in Fig. 3(II). In the presence of a virtual screened ligand with targeted protein (PubChem ID: 59671768), the RMSD of the protein complex was noted a sudden high peak of fluctuations from the beginning to 20 to 25ns. But it was started to stabilize with very less fluctuations value of 7.5 to 10.5 Å during the entire simulation period, shown in Fig. 3 (III). Where, the RMSF graph was investigated with a less fluctuation value relative to the established compound with an average value of 3.0–4.0 Å, given in Fig. 3 (IV).
2.4 Protein-Ligand interaction
Protein-ligand interactions indicated the protein's conformational stability and correlation, which allows for a better understanding of simulations. Thus, the detailed investigation of both established and screened compounds were described below.
Lapatinib (PubChem ID: 208908) – The most effective established compound -
Figure 4 shows the results of post-dynamics protein-ligand interaction studies on Lapatinib with EGFR, which indicates that the residues mainly ARG353, GLU388 (from A Chain), SER95, ASP96, SER93 (from B Chain), and TYR61 (from H Chain) revealed a mixture of H-bond, hydrophobic, Salt & Water bridges interactions during the MD simulation. As a result, it has been discovered that most of the interactions are continuously kept following docking. According to a literature review, Lapatinib (PubChem ID: 208908) is also known as Lapatinib Ditosylate, an orally active drug used to treat solid malignancies such as glioblastoma and breast cancer. It's a dual tyrosine kinase inhibitor that works by blocking both the HER2/neu and epidermal growth factor receptor pathways.
SCHEMBL14272033 (PubChem ID: 59671768) – The effective Virtual Screened compound –
The results of post-dynamics protein-ligand interaction studies on SCHEMBL14272033 with EGFR during MD simulation are shown in Fig. 5. Residues like PHE352, ARG353, LEU363, PRO365, TRP386, GLU388, ASN420 (from A Chain) and SER93, SER93 (from D Chain) have been identified with strong H-bond, hydrophobic, Salt & Water bridges interactions with the ligand, accounting for 75 and 80 percent interaction with the protein residues, respectively. Similarly, the virtually screened molecule found with most of the interactions is continually kept following docking investigations, just as established compounds. We believe that SCHEMBL14272033 (PubChem ID: 59671768) can be repurposed as an anticancer medication to treat glioblastoma and breast cancer based on these findings. As a result, it is possible to recommend it for further screening and optimization using In-vitro experiments.
2.5 Ligand property
The properties of established and screened compounds are estimated and compared in terms of RMSD, Gyration, molecular surface area (MolSA), solvent accessible surface area (SASA), and polar surface area (PSA), shown in Fig. 6 & Fig. 7.
RMSD - The RMSD of the ligand varies at first, then gradually approaches stabilization towards the conclusion of the simulation time. In the case of the established ligand, RMSD has been noted with a range between 0.8–2.4 Å with an equilibrium value of approximately 1.6Å [Fig. 6]; whereas the best virtual screened ligand exhibits a range of RMSD values between 1.5–3.0 Å with an equilibrium value of around 3.0Å[Fig. 7].
rGyr - If the body's (ligands) total mass were concentrated, the rGyr (gyration) value is the radial distance to a place with the same moment of inertia as the body's (ligands) true distribution of mass. The ligands' rGyr fluctuates dramatically up to 20 ns simulation and then gradually returns to equilibrium. In the case of the established ligand,rGyrexhibits a range between 6–6.8Å with an equilibrium value of ~ 6.4Å [Fig. 6], while the screened ligand exhibits a range of rGyr values between 7.5–8.5 Å with an equilibrium value of ~ 8.0 Å[Fig. 7].
MolSA - MolSA is a method for surface calculation that uses a 1.4 probe radius. The surface area of a van der Waals is equal to this number. In the case of the established ligand, MolSA remains constant throughout the simulation except at 40–65 ns, noted with a range of 510 to 522 Å2with an equilibrium value of 516.Å2[Fig. 6]; whereas the virtual screened ligand remains constant throughout the simulation except at 80–100 ns, and noted with a range of 568 to 592 Å2with an equilibrium value of 584 Å2[Fig. 7].
SASA - The surface area of a molecule that can be reached by a water molecule is referred to as SASA. The established ligand's SASA is constant and shows no notable fluctuations until gradually nearing equilibrium, but the screened compound showed higher stability practically throughout the simulation, with only a minor fluctuation near the conclusion. The established compound identified the SASA range of 240–420Å2with an equilibrium value of 300Å2[Fig. 6]; whereas the virtual screened ligand exhibits the SASA range of around 300 to 750 Å2with an equilibrium value of approximately 450 Å2[Fig. 7].
PSA -The solvent-accessible surface area of a molecule made solely of oxygen and nitrogen atoms is referred to as PSA. The established ligand exhibits a PSA range of 120–165 Å2 with an equilibrium value of 135Å2[Fig. 6]; whereas the screened compound’s PSA exhibits an early variation of up to 35 ns, which has a PSA range of around 135 to 180 Å2with an equilibrium value of approximately 165Å2[Fig. 7]. Overall, it is concluded that all ligand characteristics gradually stabilize and prove the ligand's stability at the protein's active site.
2.6 Drug - Drug comparison studies
Table 4 displays the re-rank scores of the compounds against the Glioblastoma target protein EGFR. The total energy of the newly discovered inhibitor PubChem ID- 59671768 was the lowest among the complete virtual screened molecules, indicating its higher affinity. Surprisingly, according to the steric energy of PLP (Piecewise Linear Potential), the other interaction of both compound’s displaying the virtual screened compound has less affinity interaction properties than the pre-established Lapatinib. Because the virtual screened compound's hydrogen bond stability is similar to that of the established inhibitor Lapatinib, which makes the newly screened compound a novel therapeutic for EGFR in Glioblastoma.
Table 4
The Drug-Drug comparison studies between established and screened compounds.
| Established Compound PubChem ID:208908 | Virtual Screened Compound PubChem ID: 59671768 |
Energy overview Descriptors | MolDock Score | Rerank Score | MolDock Score | Rerank Score |
Total Energy | -154.117 | -115.565 | -198.554 | -137.713 |
External Ligand interactions | -187.094 | -155.135 | -220.187 | -184.185 |
Protein - Ligand interactions | -187.094 | -155.135 | -220.187 | -184.185 |
Steric (by PLP) | -180.583 | -123.88 | -218.025 | -149.565 |
Hydrogen bonds | -6.511 | -5.156 | -4.162 | -3.712 |
Internal Ligand interactions | 32.977 | 39.57 | 43.472 | 44.473 |
Torsional strain | 16.463 | 15.442 | 18.632 | 17.477 |
Hydrogen bonds | 0 | | 0 |
Steric (by PLP) | 16.514 | 2.84 | 24.84 | 4.273 |
2.8 Pharmacophore mapping images
Pharmacophore mapping, in addition to molecular docking, provides a three-dimensional basic systemic topography of molecular interaction with complicated target receptors. Pharmacophore studies introduce a specific phenomenon about the best interaction mode for a certain target annotation and describe the molecule's aligned poses, which helps us figure out how the target protein and the novel drug interact. Even with Lapatinib's excellent affinity and good interaction profile (PubChem ID: 208908) shows hydrogen bonding between the drug and residues LYS31, GLU388, TYR3 [Fig. 8(A)]. The interaction of the practically screened chemical compound SCHEMBL14272033 (PubChem ID: 59671768) with the cavity of the target protein EGFR reveals that the residues ASN27, ASN420, TRP386 found with hydrogen bonding with ligands [Fig. 8(B)].
2.9 ADME/T studies
Table 5 displays the estimated ADME/T value for the best virtual screened compound (PubChem CID: 59671768) and established inhibitor (PubChem CID: 208908). A virtual screened compound has a greater absorptive value than an established compound in every way; the established compound's BBB + value is 0.9738, while the virtually screened molecule's value is 0.9755. The bioavailability indication for the two top docking findings derived from the SwissADME online tool demonstrates the compounds' activity potential. The established molecule has a higher P-glycoprotein probability value than the virtually screened compound, implying that it is more lipophilic. Both chemicals can be present in the cell, although in distinct places. Although the practically screened compound (PubChem CID: 59671768) exhibits high CYP inhibitor promiscuity and might be employed as a CYP3A4 inducer inhibitor, and other way it has a lower possibility of acting as a substrate for CYP450 2C9 and CYP450 2D6 than the established chemical. AMES toxicity is absent in both compounds, demonstrating that they are not mutagenic. The regression value for absorption and toxicity (in terms of aqueous solubility and rat acute toxicity of parecoxib) has been noted with a higher value in the established compounds than in virtual screening compounds, shown in Table 6. R programming was used to create a graphical depiction of the comparative research between the two best virtual screened substances and the two best-established compounds [Fig. 9]. It demonstrates that the virtual screening chemical (PubChem CID 59671768) is significantly less toxic than the established drug (PubChem CID 208908), and that its absorption and BBB values are comparable to the established compound.
Table 5
ADME/T profile studies performed with admetSAR tool.
ADMET predicted profile | Virtual Screened Compound PubChem ID: 59671768 | Established Compound PubChem ID:208908 |
Value | Probability | Value | Probability |
Human Intestinal Absorption | + | 0.9774 | + | 0.9814 |
Caco-2 | - | 0.8460 | - | 0.8511 |
Blood Brain Barrier | + | 0.9755 | + | 0.9738 |
Human oral bioavailability | - | 0.5827 | - | 0.8429 |
Subcellular localzation | Mitochondria | 0.3474 | Lysosomes | 0.4124 |
OATP2B1 inhibitior | + | 0.7163 | + | 0.5751 |
OATP1B1 inhibitior | + | 0.8420 | + | 0.8609 |
OATP1B3 inhibitior | + | 0.9394 | + | 0.9386 |
MATE1 inhibitior | - | 0.7400 | - | 0.7200 |
OCT2 inhibitior | + | 0.5750 | - | 0.5500 |
BSEP inhibitior | + | 0.9882 | + | 0.9971 |
P-glycoprotein inhibitior | + | 0.8933 | + | 0.8639 |
P-glycoprotein substrate | + | 0.8717 | + | 0.9343 |
CYP3A4 substrate | + | 0.7109 | + | 0.6943 |
CYP2C9 substrate | - | 0.8087 | - | 1.0000 |
CYP2D6 substrate | - | 0.7664 | - | 0.7463 |
CYP3A4 inhibition | + | 0.8224 | + | 0.8065 |
CYP2C9 inhibition | + | 0.5000 | - | 0.5544 |
CYP2C19 inhibition | + | 0.6282 | + | 0.5274 |
CYP2D6 inhibition | - | 0.7210 | - | 0.7861 |
CYP1A2 inhibition | + | 0.6637 | + | 0.5386 |
CYP inhibitory promiscuity | + | 0.8989 | + | 0.9104 |
UGT catelyzed | - | 0.0000 | - | 0.0000 |
Carcinogenicity (binary) | - | 0.9286 | - | 0.9286 |
Carcinogenicity (trinary) | Non-required | 0.5442 | Non-required | 0.5703 |
Eye corrosion | - | 0.9836 | - | 0.9886 |
Eye irritation | - | 0.9531 | - | 0.9541 |
Ames mutagenesis | - | 0.5700 | - | 0.5400 |
Human either-a-go-go inhibition | + | 0.9488 | + | 0.9566 |
micronuclear | + | 0.8700 | + | 0.9100 |
Hepatotoxicity | + | 0.6000 | + | 0.7250 |
Acute Oral Toxicity (c) | III | 0.5778 | III | 0.5949 |
Estrogen receptor binding | + | 0.8483 | + | 0.8218 |
Androgen receptor binding | + | 0.8934 | + | 0.8909 |
Thyroid receptor binding | + | 0.6652 | + | 0.6466 |
Glucocorticoid receptor binding | + | 0.7627 | + | 0.6743 |
Aromatase binding | + | 0.5838 | + | 0.5872 |
PPAR gamma | + | 0.7981 | + | 0.7119 |
Honey bee toxicity | - | 0.7272 | - | 0.7526 |
Biodegradation | - | 0.9000 | - | 0.8250 |
crustacea aquatic toxicity | + | 0.6800 | + | 0.6200 |
Fish aquatic toxicity | + | 0.6826 | + | 0.9246 |
Table 6
ADMET predicted profile – Regressions.
| Virtual Screened Compound PubChem ID: 59671768 | Established Compound PubChem ID:208908 |
| Value | Unit | Value | Unit |
Water solubility | -3.501 | logS | -3.647 | logS |
Plasma protein binding | 1.245 | 100% | 1.201 | 100% |
Acute Oral Toxicity | 2.894 | kg/mol | 3.089 | kg/mol |
Tetrahymena pyriformis | 1.89 | pIGC50 (ug/L) | 1.124 | pIGC50 (ug/L) |
2.10 Boiled egg Plot
The Boiled Egg plot attempts to forecast the drugs' gastrointestinal absorption and blood-brain barrier characteristics. For the same objective, the best pre-established inhibitor lapatinib (PubChem ID: 208908, PubChem ID: 176870) and the best virtual screening inhibitors (Pub CID: 59671768, Pub ID: 118717618) were chosen. [Fig. 10]. The virtual screened compounds (Pub CID: 59671768, Pub ID: 118717618) are present in the yolk region of the egg plot, indicating that the virtually screened compound is capable of crossing the blood-brain barrier, which is required for the treatment of glioblastoma. When compared to the best-established compounds, both are present in the grey region, indicating their lower gastrointestinal absorption and inability to cross the blood-brain barrier.
The Bioavailability Radar efficiently analyzes a molecule's drug-likeness. Each property has a pink area that represents the optimal range (lipophilicity: XLOGP3 between 0.7 and + 5.0, size: MW between 150 and 500 g/mol, polarity: TPSA between 20 and 1302, solubility: logS significantly less than 6, saturation: percentage of carbons in the sp3 hybridization not more than 25%, and flexibility: no more than 9 rotatable bonds). Figure 11 depicts the bioavailability map for both well-established and virtually screened substances.