Using guided differentiation to predict cancer treatment related cardiotoxicity

ABSTRACT The premise of our proposal is that by characterizing the in vitro response of different types of cardiac single cells to different cancer drug treatments, we will be able to identify biomarkers that can help us classify patients at risk for cardiotoxicity. By applying single-cell sequencing to a novel system of cardiac guided differentiation cultures, we are proposing to study the effects of the anticancer drugs doxorubicin (DOX), 5-fluorouracil (5-FU), and bevacizumab (BVC) in multiple relevant cell types from a genetically diverse panel of 70 individuals. Our goal is to build a genetic-based classifier that can stratify cancer patients by their susceptibility to drug-induced cardiovascular toxicity (CT). To achieve this goal, we need to first identify candidate loci that underlie the functional differences between individuals. This is required, because whole-genome approaches to predict complex risk (such as polygenic risk scores) have generally shown not to be effective beyond the sample in which they were developed. A classifier based on relevant functional loci (the response eQTLs, in our case) is more likely to be generally effective and transferable. The premise of our study rests on two assumptions. The first is that inter-individual variability in response to chemotherapeutic drugs is at least partially mediated by genetic variants that affect gene regulation in a drug-dependent manner. In other words, genetic variants respond to chemotherapeutic drugs by regulating the activity of specific genes. The second assumption is that different cell types vary in their response to chemotherapeutic drugs. That is, a regulatory variant that affects gene expression in cardiomyocytes, for example, may have a different effect (or none at all) on gene expression in endothelial cells. To achieve our goals, we propose to collect single cell RNA-seq data from a panel of 100 cardiac guided differentiation cultures in control and drug-treated conditions (aim 1); to identify genetic variants that regulate the transcriptomic response of cardiac culture cell types to each drug (aim 2); and to test whether response QTLs identified in cardiac in vitro cultures can be used to retrospectively classify patients as resistant or sensitive to drug-induced cardiotoxicity (aim 3).