Chemometric approach to estimate kinetic properties of paclitaxel prodrugs and their substructures for solubility prediction through molecular modelling and simulation studies

Abstract

AbstractPaclitaxel drug is administered in the treatment of ovarian and breast cancer and also in Kaposi sarcoma. In spite of being nanomolar active, use of this drug is confined because of its low aqueous solubility, hence many prodrugs for increasing paclitaxel's solubility were formed, but the formation process was not rational. In the current study, quantitative structure property relationship (QSPR) models were formed for the solubility prediction of paclitaxel prodrugs. Structures of all molecules were optimized at the parameterization method 6 (PM6) and Austin Model 1 (AM1) levels, after which Dragon‐based 5250 descriptors and quasi‐mixture descriptors were calculated. Independent descriptors were selected in multiple steps, and QSPR models having 12 and 10 descriptors with R2 and Q2 values of 0.78 and 0.60 and 0.80 and 0.69 for AM1‐ and PM6‐optimized geometry datasets, respectively, were formed. Also, for substituent group dataset, QSPR models with 8 and 9 descriptors having R2 and Q2 values of 0.82 and 0.76 and 0.93 and 0.83 were determined for AM1‐ and PM6‐optimized geometry datasets, respectively. Quasi‐mixture descriptors, which were calculated for substituent group datasets, gave the QSPR model with R2 and Q2 values of 0.70 and 0.58 and 0.69 and 0.52 respectively for AM1‐ and PM6‐optimized geometries. After the models' development, the substituent group dataset was employed for the formation of docking and molecular dynamics simulation–based models for the metabolic study with CYP1A2 enzyme. It is anticipated that the proposed QSPR models will serve as a base for the designing of new paclitaxel prodrugs with improved aqueous solubility.