Molecular interaction fields (MIFs) to predict lipophilicity and ADME profile of antitumor Pt(II) complexes

Purpose: To describe a computational tool to calculate molecular descriptors of potential application in ADME virtual screening of antitumor Pt(II) drug candidates. Methods: The multistep computational procedure consists in (a) building and optimization (PM3) of the 3D structures of the investigated complexes, (b) parametrization of Pt(II) and its implementation in GRID, (c) GRID calculations and extraction of the information content with VolSurf and BIOCUBE4mf, and (d) PLS analysis to look for the correlation between experimental data and the molecular descriptors. Results: The following results were obtained: (a) the calibration of the GRID force field to take into account the platinum di-cation, (b) solid PLS models between log k30 and log kw with VolSurf descriptors which highlight the main structural differences between the two chromatographic parameters, (c) the prediction of virtual (of each conformer) log k30 and log kw, and (d) the identification of the main descriptors governing VD _ss of drugs in clinical use. Conclusion: The study suggests a strategy to identify good Pt(II) complexes prior to their synthesis to eliminate as soon as possible drug candidates with unfavorable PK profile.