A strategy for predicting the chemosensitivity of human cancers and its application to drug discovery

Area of Science:

• Bioinformatics
• Computational Biology
• Cancer Research

Background:

• The National Cancer Institute's 60 human cancer cell lines (NCI-60) screen has limitations.
• The NCI-60 panel lacks representation for all cancer types.
• NCI-60 drug responses do not always predict clinical efficacy in patients.

Purpose of the Study:

• To develop a method for predicting drug activity in cell types not included in the NCI-60.
• To predict clinical responses of cancer patients based on drug screening data.
• To extrapolate drug activity from existing databases to new cell types and patient tumors.

Main Methods:

• Developed and applied a novel algorithm termed coexpression extrapolation (COXEN).
• Utilized gene expression microarray data as a Rosetta Stone for data translation.
• Applied COXEN to predict drug sensitivity in bladder cancer cell lines and clinical responses in breast cancer patients.

Main Results:

• COXEN accurately predicted drug sensitivity in bladder cancer cell lines.
• COXEN accurately predicted clinical responses in breast cancer patients treated with chemotherapeutic drugs.
• In silico screening of 45,545 compounds using COXEN identified a novel agent active against human bladder cancer.

Conclusions:

• Coexpression extrapolation (COXEN) is a viable method for predicting anticancer drug activity.
• COXEN can bridge the gap between cell line drug screening and clinical patient response.
• This approach facilitates the identification of novel therapeutic agents for various cancer types.