Background: The discovery of novel potent molecules for both cancer prevention and
treatment has been continuing over the past decade. In recent years, identification of new, potent, and
safe anticancer agents through drug repurposing has been regarded as an expeditious alternative to traditional
drug development. The cyclooxygenase-2 is known to be over-expressed in several types of
human cancer. For this reason cyclooxygenase-2 inhibition may be useful tool for cancer chemotherapy.
Objective: The first aim of the study was to develop a validated linear model to predict antitumor activity.
Subsequently, applicability of the model for repurposing these cyclooxygenase-2 inhibitors as
antitumor compounds to abridge drug development process.
Methods: We performed a quantitative structure-toxicity relationship (QSTR) study on a set of coumarin
derivatives using a large set of molecular descriptors. A linear model predicting growth inhibition
on leukemia CCRF cell lines was developed and consequently validated internally and externally.
Accordingly, the model was applied on a set of 143 cyclooxygenase-2 inhibitor coumarin derivatives to
explore their antitumor activity.
Results: The results indicated that the developed QSAR model would be useful for estimating inhibitory
activity of coumarin derivatives on leukemia cell lines. Electronegativity was found to be a prominent
property of the molecules in describing antitumor activity. The applicability domain of the developed
model highlighted the potential antitumor compounds.
Conclusion: The promising results revealed that applied integrated in silico approach for repurposing
by combining both the biological activity similarity and the molecular similarity via the computational
method could be efficiently used to screen potential antitumor compounds among cyclooxygenase-2