Integrating Expression Profiling Into Drug Discovery

Abstract

Toxicity and lack of efficacy account for about 60% of drug pipeline attrition, and early stage amelioration of these issues can increase the probability of success for new drug candidates. To achieve this effectively requires knowledge of target and off-target effects, identification of biomarkers, and an understanding of genotypic/phenotypic diversity within patient groups. Analysis of gene expression profiles obtained from target tissues reveals distinct exposure-related characteristics as defined by distinct transcriptional signatures. Using a compendium of expression profiles, compound-specific effects for new chemicals can be compared to responses for known and characterized compounds to help identify target effects and toxic liabilities even when gene annotations are incomplete or inconclusive. Further, where annotations exist discrete gene responses can provide specific mechanistic insights through the identification of response pathways, including those controlled by both the intended therapeutic target and untoward off-target effects. By examining human gene expression, discrete preclinical gene changes can also help identify putative transcriptional biomarkers that may translate into clinically accessible biomarkers. Thus, expression profiling may ultimately drive lower failure rates at all stages of drug development allowing faster pipeline progression, as well as producing successful therapies for patient subgroups.