Antiproliferative Activity Predictor: A New Reliable In Silico Tool for Drug Response Prediction against NCI60 Panel

Area of Science:

• Computational Biology
• Medicinal Chemistry
• Drug Discovery

Background:

• In vitro antiproliferative assays are crucial for anticancer drug discovery and understanding mechanisms of action.
• The National Cancer Institute's Developmental Therapeutics Program (NCI-DTP) extensively tests compounds against the NCI60 tumor cell line panel.
• Vast biological data necessitates computational tools for predicting anticancer properties of novel agents.

Purpose of the Study:

• To develop a novel in silico tool, the Antiproliferative Activity Predictor (AAP), for predicting GI50 values against the NCI60 panel.
• To provide a reliable and robust ligand-based protocol for early-stage anticancer drug discovery.
• To aid medicinal chemists in identifying and optimizing potential anticancer small molecules.

Main Methods:

• Utilized antiproliferative data from the NCI and molecular descriptors to build the AAP tool.
• Developed a ligand-based computational protocol for GI50 value prediction.
• Validated the AAP tool using internal and external compound sets, including curcumin analogues and NCI-selected molecules.

Main Results:

• The AAP tool demonstrated high reliability and robustness, achieving an error of less than ±1 unit for GI50 values in a test set of 99 structures.
• The tool showed strong correlation with experimental data, particularly for GI50 values in the 4-6 range.
• AAP successfully identified potentially active compounds within an in-house curcumin analogue database and confirmed its potential with NCI-evaluated molecules.

Conclusions:

• The AAP tool is a powerful and reliable in silico method for predicting anticancer activity and aiding in the discovery and optimization of novel anticancer small molecules.
• Integration with the DRUDIT web service makes AAP accessible for the medicinal chemistry community.
• Future development includes expanding the training set and adapting the protocol for predicting TGI and LC50 values to assess toxicity.