Computer Modeling to Help Predict Drug Failure

A team of pharmacologists and computer specialists has designed a program to analyze the failure of the cholesterol-lowering drug torcetrapib, which was withdrawn from clinical trials after causing fatal cardiovascular disease.

Torcetrapib acts by inhibiting cholesteryl ester transfer protein (CETP), which normally transfers cholesterol from HDL cholesterol to very low density or low-density lipoproteins (VLDL or LDL). Inhibition of this process results in higher HDL levels. Its development was halted in 2006, after more than 15 years of development at a cost of nearly $850 million, when phase III studies showed excessive all cause mortality in the treatment group receiving a combination of atorvastatin (Lipitor) and torcetrapib.

Investigators at the University of California, San Diego (USA) developed a novel computer program to study torcetrapib's protein-ligand binding profiles on a genome-wide scale. They reported in the May 15, 2008, online edition of the journal PLoS Computational Biology that torcetrapib binding was not limited to a specific receptor. Instead, torcetrapib actually acted on a dozen different receptors, resulting in unanticipated side effects. Binding to each receptor triggered changes in the activity of a molecular pathway. A combination of changes in many different pathways led to the overall physiological effect of the drug.

"This work extends the scope of chemogenomics - the study of genomic responses to chemical compounds - and exemplifies the role that systems biology has in the future of drug discovery," explained senior author Dr. Philip E. Bourne, professor of bioinformatics at the University of California, San Diego. "Torcetrapib actually acted on a dozen different receptors, resulting in an unanticipated side effect. This multi-inhibitor binding pattern may not be at all unusual. At this time we do not have a complete structural proteome to analyze, one that maps all the protein structures in the genome - either experimental or model - to which drugs could bind. So though we still may not have a complete understanding of off-target binding, this strategy is already useful."

Related Links:
University of California, San Diego