Computational modelling and analysis of Pyrimidine analogues as EGFR inhibitor in search of anticancer agents

Introduction and Aim: Epidermal Growth Factor Receptor tyrosine kinase is a well-known and widely studied cancer therapeutic target protein. Based on the reported anticancer activity of pyrimidines, a series of 13 compounds are designed. In the present study the EGFR kinase domain is targeted with the designed 13 compounds. Materials and Methods: With missing residue in the kinase domain of EGFR crystallized structure, the domain is modelled using homology modelling, evaluated, energy-based optimization is carried out using OPLS in Gromacs. The default binding site was considered from the known EGFR kinase domain – Erlotinib complex crystallized structure. The molecular docking is carried out using Autodock Vina, Insilico toxicity profiling and enrichment analysis of pathway is studied using Swiss-ADME and Enrich R. Results: Compounds 7, 9, 10 and 12 revealed a binding energy of -8.8, -8.3, -8.3 and -8.4 Kcal/mol and makes two h-bonds with MET-769. All the 13 compounds are under the range of Lipnski drug likeness, with high GI-absorption rate. Considering the metabolic enzyme activity, the entire series of compounds are predicted to inhibit the metabolizing enzyme CYP1A2, CYP2D6 and CYP3A4. Compounds 2, 3, 7, 8 and 13 acts as a substrate to CYP2C19 and compound 1, 3,4, 5, 6, 7, 8, 9, 10, 11, 12 and 13 act as a substrate to CYP2C9. Conclusion: The inhibition of metabolizing enzyme may affect the poor metabolizing and slowing down the excretion time of molecules from the body. The current in-silico molecular docking, in-silico PKPD study of compounds suggesting that they can be developed as putative lead compounds for developing new anti-cancer drugs.


INTRODUCTION
yrimidines have a long and distinguished history extending from the days of their discovery as important constituents of nucleic acids to their current use in the chemotherapy of AIDS. The pyrimidine ring is found in vitamins like thiamine and folic acid during the last two decades, several pyrimidine derivatives have been developed as chemotherapeutic agents and have found wide clinical applications. Antineoplastic and anticancer agents. There are a large number of pyrimidine-based antimetabolites. They are usually structurally related to the endogenous substrates that they antagonize. The structural modification may be on the pyrimidine ring or on the pendant sugar groups. One of the early metabolites prepared was 5-fluorouracil, a pyrimidine derivative also exhibits some useful antineoplastic activities (1)(2)(3).
Pyrimidine is a useful structural motif for displaying potent antitumor activity. Several analogues of Pyrimidine are synthesized and tested for evaluating their anticancer activity. Due to promising anticancer activity exhibited by Pyrimidine, in the present manuscript attempt to design newer molecules which will exhibits useful anticancer activity. Attempt to design newer analogues by using Pyrimidine as pharmacophore with the hope that newer molecules will possess anticancer activity by inhibiting EGFR (Epidermal growth Factor Receptor) is described in the present manuscript. EGFR is kind of protein kinases, were proved to be a viable target for anticancer drug development (4)(5)(6)(7). In addition, EGFR-TK is utmost important kinases that plays a fundamental role in signal transduction pathways (8). EGFR and its ligands, Epidermal Growth Factor [EGF] and transforming growth factor-α (TGF-α) have been implicated in numerous tumors of epithelial origin (8,9). Therefore, the design of inhibitors that target EGFR-TK is an attractive approach for design of new therapeutic agents.

Receptor preparation
Receptor Epidermal Growth Factor Receptor tyrosine kinase domain was considered in the experiment. Due to missing residues in most of the 3D structure, we have adopted the homology modeling concept to model the 3D structure of EGFR kinase domain. Using P Swiss-model server, the EGFR domain sequence is subjected to Swiss-model server and optimum homolog is selected for modeling. The modeled structure evaluated using Ramchandran plot and ProSA analysis. Further the structure was energy wise minimized using steepest method in Gromacs. The protein is solvated in water body (TIP3P model) in OPLS method and neutralized the system by adding 08 Na + ions. Total 144973 molecules are subjected to energy minimization process, the potential energy, bonds, bond angles and prop-dihedral are evaluated for the stability purpose of the protein (10).

Ligand sketching and optimization
The selected 13 compound structures were sketched using Chemsketch tool version 12.0 and saved in .mol format. Using universal force field (UFF) using steepest descent algorithm all the 13 compounds are geometrically / energy wised optimized to attain their global lowest energy formed (11). Each compound is optimized up to it converged value by optimizing the bonds, angle, di-hedral and improper that depicted the lowest form of energy and finally saved in the pdb format

Molecular docking
We have autodock vina in autodock PyRx tool (12,13). Here we have selected the default binding site region of Erlotinib the indigenous inhibitor of EGFR kinase with following pharmacophoric amino acids included in it LEU-694, LYS-721, MET-769, LEU-820, THR-830. The binding site size are as follows center value, X: 59.7232; Y: 48.5137; 51.1831with a dimension of X: 25; Y: 25; Z: 25 A. all the 13 compounds were loaded into the autodock PyRx tool and finally target protein EGFR kinase and compounds were converted to.pdbqt file format as a default input file format for autodock operations.

Computational modeling of ADME properties
The Drug development process imbibes the evaluation of absorption, distribution, metabolism and excretion (ADME). The ADME process highly dependent on physiological activity based model that includes the desired target association and dissociation (formation of bonds/selective interactions), blood brain permeation, toxicity, poor efficacy, interaction with metabolizing enzyme for clearance and off-target interactions, leading to toxicity and undesirable effect by penetrating significant metabolic pathways. Current insilico based modeling and analysis helps us to segregate the poor performer versus optimum and leads to early estimation of ADME based failure in the clinical phases. In the present work, we have predicted the ADMET properties of the 13 compounds were through SwissADME database, freely available at http://www.swissadme.ch/ (14).
Here, we have assessed the lipophilicity (lipophilicity, XLOGP3 between −0.7 and + 5.0; size, MW between 150 and 500 g/mol; polarity, TPSA between 20 and 130 Å; solubility, log S (calculated with the ESOL model) not higher than6; saturation, fraction of carbons in the sp3 hybridization not less than 0.25; and flexibility, no more than 9 rotatable bonds); drug likeness is calculated using Lipinski rule of five.
Elimination/excretion and toxicity for molecules were tested against the cytochromes P450 (CYP) and P-gp. The overexpression of P-gp in some tumor cells is well reported and leads to multidrug-resistant cancers, including a major role of P-gp which is to protect the central nervous system from xenobiotics. It is a wellknown fact that about 40 to 90% of drugs are substrate of five main isoforms (CYP1A2, CYP2C19, CYP2C9, CYP2D6, and CYP3A4). CYP and P-gp can process small molecules. Inhibition of these isozymes causes drug-drug interactions, toxicity, or unwanted side effects due to the lower clearance and accumulation of the drugs or its metabolites. Considering the predicted outcome for inhibition of CYP family enzymes, here we have perform the enrichment analysis to these CYP family enzymes and reported the metabolic pathways that might be affected by inhibition of these enzymes using EnrichR online enrichment analysis server (15,16).

RESULTS
The target sequence exhibited a 100% homologous to 1M17.pdb a crystallized 3D structure with an inhibitor Erlotinib. The Ramchandran plot exhibited a 99.3% in most favored and favored region and 0.7 % is in disallowed region ( fig. 1A), ProSA analysis resulted the modeled structure in x-ray crystallized zone supporting the high degree of good prediction of model with a Z-score of -6.76 ( fig. 1 B).   The pyrimidine derivative was geometrical based optimized by converging the bonds, angle, dihedral, improper-torsion, Van-der Waals and Coulomb charges to their lowest energy form, results are summarized in table1 and final structures were subjected for further docking process. In comparison to our synthesized 13 compounds all have acquired the binding site region with better binding energy with respect to Erlotinib, the binding energy ranges from -7.5 Kcal/mol to -8.8 Kcal/mol.

Computational modeling / based drug likeness and toxicity
In order to reduce the cost, time and late failure of compound and gain early estimation of true positive results we have performed in-silico based ADME analysis. All the 13 compounds were subjected to Swiss-ADME online server for their drug likeness and toxicity prediction.  . 6).

DISCUSSION
Compounds well known as pyrimidine derivatives are well explored and studied for their anticancer activity by targeting the EGFR wild type and mutants (17,18). Numerous curcumin analogues containing pyrimidine ring are effective in inhibition of EGFR and evaluated by in-silico methodology (19). Pyrimidine analogues are also well studied for their dual inhibition roles like ATX and EGFR (20) as well as for EGFR/HER2 and well supported by molecular docking and interactions based study (21,22). Out of the 13 designed pyrimidine analogues, Compound 07, 09, 10 and 12 exhibited a good degree of binding interactions within the EGFR binding pocket (fig. 7). The in-silico based drug likeness and PKPD analysis supported high degree of GI-absorption rate and highly commendable to the Lipinski filter shows the drug-able activity. As all the designed compounds exhibited as an inhibitor to CYP1A2, CYP2D6 and CYP3A4 suggests they may affect some of the essential biological pathway. Whereas they also act as a substrate to the other metabolic enzyme reveals the signal of metabolizing of these compounds further and may excrete via appropriate mechanism. Some of the compounds holds a strong binding affinity, better outcome from insilico toxicity profiling and pathway analysis and can be further studied for biological evaluations.

CONCLUSION
In conclusion, the 3D structure of EGFR kinase domain resulted a 99.3% in most favored and favored region and 0.7% is in disallowed region of Ramachandran plot that shows a highly acceptable range of structure. All the designed compounds exhibited a good docking score in comparison to the indigenous inhibitor Erlotinib, including a highly similar pharmacophoric interactions pattern, where compound 7, 9, 10 and 12 are optimum. The outcomes from the insilico toxicity profiling and pathway analysis resulted few compounds to be further studied and evaluate for more parameters in terms of their biological evaluation in in-vivo and in-vitro condition. The present work suggests that the current studied compounds can be considered as putative lead compounds for developing new anti-cancer drugs.