P. Geeleher,1 A. Loboda,2 D. Lenkala,1 F. Wang,1 J. Wang,1 M. Nebozhyn,2 M. Chisamore,2 J. Hardwick,2 M. L. Maitland,1 R. Huang1; 1University of Chicago, Chicago, IL, 2Merck Research Laboratories, North Wales, PA
BACKGROUND: Histone deacetylase inhibitors (HDIs) have diverse effects on the tumor and host tissues. Although approved for some indications in oncology, HDIs have had highly variable effects in clinic. The mechanisms-of-resistance and mechanisms-of-action of this class of drugs remain unclear and disparate biological processes have been implicated across numerous studies.
METHODS: Many candidate biomarkers of vorinostat (a HDI) have been proposed but none has been qualified for clinical use. We developed a polygenic gene expression model to generate robust predictive biomarkers of sensitivity to HDI.
RESULTS: Using cell lines, we show that sensitivity to vorinostat can be predicted using a polygenic model that accounts for the additive effect of many different genes. In fact, the model prediction accuracy is close to that attained when the same cell lines are re-screened using a cell viability assay. Our analyses recapitulate and contextualize many previous findings and suggest an important role for processes such as chromatin remodeling, autophagy and apoptosis. As a proof-of-concept, we have also discovered a novel causative role for CHD4, a helicase involved in the histone deacetylase complex that is very commonly mutated or amplified in clinical tumors. Finally, we have used a mouse model to measure changes in tumor expression following vorinostat exposure in a high resolution time-course, thus elucidating the molecular dynamics of drug response. We show that these results can be integrated with the results from the cell line analysis to uncover specific mechanisms of drug response.
CONCLUSION: Our work suggests a paradigm shift from the traditional single gene/pathway evaluation to simultaneously evaluating multiple independent high-throughput gene expression datasets.