Background: Quantum chemical methods and molecular mechanics approaches face a
lot of challenges in drug metabolism study because of either insufficient accuracy, huge computational
cost, or lack of clear molecular level pictures for building computational models. Low-cost
QSAR methods can often be carried out, even though molecular level pictures are not well defined;
however, they show difficulty in identifying the mechanisms of drug metabolism and delineating
the effects of chemical structures on drug toxicity because a certain amount of molecular descriptors
are difficult to be interpreted.
Objective: In order to make a breakthrough of QSAR, mechanistically interpretable molecular descriptors
were used to correlate with biological activity to establish structure-activity plots. The biological
activity is the lethality of anthracycline anticancer antibiotics denoted as log LD50. The
mechanistically interpretable molecular descriptors include electrophilicity and the mathematical
function in the London formula for dispersion interaction.
Method: The descriptors were calculated using quantum chemical methods.
Results: The plots for electrophilicity, which is interpreted as redox reactivity of anthracyclines, can
describe oxidative degradation for detoxification and reductive bioactivation for toxicity induction.
The plots for the dispersion interaction function, which represents the attraction between anthracyclines
and biomolecules, can describe efflux from and influx into the target cells of toxicity. The
plots can also identify three structural scaffolds of anthracyclines that have different metabolic
pathways, resulting in their different toxicity behavior.
Conclusion: This structure-dependent toxicity behavior revealed in the plots can provide perspectives
on drug design and drug metabolism study.