Molecular modeling studies on HIV-1 inhibitors and their potential use as anticancer agents.

Abstract

Acquired Immunodeficiency Syndrome (AIDS), currently regarded as one of the deadliest diseases, is a disease of the human immune system caused by the Human Immunodeficiency Virus (HIV). This dissertation addresses two classes of HIV-1 inhibitors: (i) integrase and (ii) protease inhibitors. With the first class, a 2D-QSAR study was carried out on compounds from a variety of structural classes; 40 diketo acid and carboxamide derivatives; possessing integrase inhibitory activity. This study investigated the relationship between molecular properties and HIV-1 integrase inhibitor activities and established accurate QSAR predictive model using the Genetic Function Algorithm (GFA) statistical model. The logarithmic inverse values of IC50 (μM) and physicochemical parameters represent the dependent variable and independent variable, respectively. Results demonstrated that the radius of gyration, Zagreb index, Wiener index and minimized energy are statistically significant with the correlation coefficient value of 0.820 and play an important role in HIV-1 integrase inhibition. With the second class, the binding affinities of some FDA-approved HIV-1 protease inhibitors, which were reported to possess anticancer activities, were estimated. The findings proposed here may alter perceptions about how NFV binds to the human Hsp90; the protein responsible for the overexpression of HER2+ breast cancer; since it has only been reported to inhibit NSCLC and a collection of yeast strains. A human Hsp90 homologue was built due to the lack of a full X-ray crystal structure of the human Hsp90 on protein data bank. The Ramachandran plot showed the validity of the human Hsp90 homologue where 98% of all residues, including the active site residues, were in the favoured region and 99.8% were in the allowed region. The NTD active acid residues were found to be Leu43, Asn46, Lys53, Ile91, Asp97, Met93, Asn101, Ser108, Gly109, Phe133 and Thr179. The obtained active site residues for the human Hsp90 homologue CTD were Gln523, Val534, Ser535, Lys538, Thr595, Tyr596, Gly597, Trp598 and Met602. The system stability and overall convergence of simulations were evaluated. The RMSD of all nine PIs did not exceed 2Å and the system stabilised after 1000 ps and 1800 ps MD

simulation at the NTD and CTD, respectively. The fluctuations of potential energies at the NTD were <2000 kcal/mol for 5 ns of MD simulation and CTD show that the fluctuations of the potential energy to be ≤8000 kcal/mol. The free binding energy of NFV was -83.03 kcal/mol at the NTD and -39.3 kcal/mol at the CTD. This value shows a significant difference (~43.73 kcal/mol) between the interaction energy at the NTD and CTD. Energy decomposition analysis at the NTD and CTD show that these two active sites have major energy contributions from their respective active site residues. This study is of great importance to medicine as it predicts the biological activity of some potent HIV-1 IN and investigates the potential use of the current HIV-1 PR drugs as anticancer agents.