Exploiting convergent evolution to design biomarker extraction tools for the prediction of therapeutic response in cancer

点击翻译按钮获取中文摘要

PROJECT SUMMARY/ABSTRACT The effective treatment of drug resistant tumors represents one of the greatest unmet needs in oncology research. The evolution of therapeutic resistance in cancer is a dynamic process, shaped by many external forces, including selection pressures, microenvironment, and the timescales of clinical treatments. As tumors evolve under these heterogeneous settings, a variety of genotypes emerge and lead to large differences in drug response phenotypes between patients. By grouping tumors based on their response to treatment, we can exploit principles of convergent evolution, where similar phenotypes evolve independently between individuals. In doing so, this work aims to aid precision medicine by identifying commonalities between tumors with similar drug response phenotypes. Gene expression signatures are a powerful tool that can be used to predict convergent states of drug sensitiv- ity and resistance. Using vast open-source datasets, Aim 1 of this proposal will demonstrate a novel method for extracting and validating gene expression signatures to predict therapeutic response in cancer. Cell lines with the best and worst response to a given drug are pooled and compared using differential gene expression analysis. Genes with increased expression in a state of sensitivity or resistance become seed genes in a co-expression network based on gene expression from tumor samples. From there, only seed genes with strong co-expression within patient samples are extracted to form the final gene expression signature. This novel approach integrates clinical sample data to the signature extraction method in order to increase translational value compared to molec- ular signatures extracted using only cell line datasets. Next, Aim 2 of this proposal investigates the phenomenon of collateral sensitivity, where resistance to one drug aligns with sensitivity to another drug. Because the evo- lution of collateral resistance and sensitivity can be unpredictable, molecular signatures of convergent states of collateral sensitivity and resistance could greatly enhance treatment planning once resistance to first-line ther- apy has evolved. Using EGFR+ non-small cell lung cancer cell lines as a model system, this project aims to identify molecular signatures of evolutionarily convergent collateral sensitivity/resistance phenotypes during the experimental evolution of therapeutic resistance to targeted therapies.