Quantitative Structure-activity Relationship (QSAR) Modelling of Indomethacin Derivatives using Regression Analysis
Neerja Shukla,Bechan Sharma
DOI: https://doi.org/10.2174/0109298673245890231004152136
Abstract:Background: Indomethacin is a non-steroidal anti-inflammatory drug (NSAID) used for medication to reduce fever, spondylitis, or shoulder pain. It mainly works by the inhibition of prostaglandins, the endogenous signaling molecules. Methods: Fifteen indomethacin derivatives have been analyzed in relation to their physicochemical and molecular properties. Two-dimensional (2D) structures of fifteen indomethacin derivatives were drawn using the ACD Lab Chem Sketch version. Most of the topological parameters, such as wiener index (W), mean wiener index (Wa), Balaban indices (J), Balaban centric index (BAC), and molecular connectivity (χ), were calculated by using E Dragon software. The most common molecular file formats accepted in EDragon software were SMILES notations created online by Babel software and 2D structures of various derivatives, which were converted into 3D optimized structures using online CORINA, provided by Molecular Networks GMBH. 3D structures of compounds were also drawn on Gauss View software for calculations of various density functional theory (DFT) based quantum chemical descriptors, such as total energy (TE), softness (S), hardness (η), chemical potential (μ), highest occupied molecular orbital energy (HOMO), and lowest unoccupied molecular orbital energy (LUMO). All species were fully optimized in the gas phase with a 6-31+G* basis set. The harmonic vibrational frequency calculations were used to confirm that the optimized structures were minima, as characterized by positive vibrational frequencies. Results: Combinations of various descriptors, such as D, ID, IOR, Log P, Mr, Mv, Mw, Pc, BAC, Pz, St, W, Wa, 0χ, 1χ, 2χ,3χ,4χ, 5χ, and Xeq have been found to be significant for modeling of activity. QSAR model no. 2: pIC50= -20.605 (±6.600) IOR - 0.747 (±0.454) I1 -5.083 (±3.478) Xeq + 51.647 optimized with empirical parameters with high statistical quality (R= 0.921, R2=0.848) was found to be the best model obtained. Conclusion: The QSAR model obtained suggests that substituents with a lesser value of the index of refraction and less electronegative groups were favourable for the activity, whereas indomethacin derivatives with a CH2CH2NHCONH(CH2)3ONO2 group at R1 position were unfavourable for the activity. The results were critically discussed based on regression data and cross-validation techniques. Pogliani factor Q and the results of the LOO (leave-one-out) method confirmed the reliability and predictability of the proposed models that could be highly beneficial for the future designing of new analogues with higher potency.