QSAR studies to predict acridinones’ ability to interstrand DNA crosslinks formation were
performed. The study is based on experimental, as well as predicted retention data (log k) and was
conducted by connected QSRR and QSAR strategy. For this purpose, chromatography analysis of
acridinone derivatives was utilized. Moreover, computer modeling of the above-mentioned compounds
was performed. Afterwards, statistical analysis of the obtained results was performed by two
different HPLC stationary phases: phosphatidylcholine (IAM) and α1-glycoprotein (AGP). This approach
allowed determining retention parameter log k, which characterizes binding affinity of acridinones
to phospholipids or proteins. Moreover, molecular modeling was performed based on the chemical structure of considered
acridinones using HyperChem 8.0 program (HyperCube Inc., Gainesville, FL, USA). Structural descriptors were
obtained from Dragon 6.0 software (Talete, Italy).
Those data were used to create general QSAR equations. Derived QSAR models described acridinones’ ability to interstrand
DNA crosslinks formation (C0) depending on HPLC retention parameters, whereas log k parameter obtained by
HPLC analysis was mostly dependent on molecular descriptors calculated.
Additionally, the predictive performance of obtained QSARs and QSRRs models allowed us to predict the ability to interstrand
DNA crosslinks formation by acridinones derivatives. It also enabled us to predict their chromatographic retention
parameters. Proposed connected QSRR and QSAR approach could be a useful tool for in silico experiments, which verifies
acridinones’ activity, without any in vivo biological test.