Abstract 15156: Statistical Modeling of Cardiac-Derived Progenitor Cell Extracellular Vesicles to Predict Clinical Trial Outcomes

Abstract

Introduction: Cardiac-derived c-kit+ progenitor cells (CPCs) are under investigation in the CHILD phase I clinical trial (NCT03406884) for the treatment of hypoplastic left heart syndrome (HLHS). Therapeutic efficacy of CPCs can be attributed to the release of extracellular vesicles (EVs), carrying beneficial RNA cargo. Despite some successes with pediatric cardiac cell therapy, large variation in cell populations and patient outcomes remains a problem. To investigate these discrepancies and understand sources of cell therapy variability we took a machine learning approach: combining bulk CPC-derived EV (CPC-EV) RNA sequencing and cardiac-relevant in vitro experiments to build a predictive model. Methods: We isolated CPCs from cardiac biopsies of congenital heart disease patients (n=30) and the lead-in HLHS patients in the CHILD trial (n=7). We cultured CPCs, collected EVs from the media, and sequenced EV miRNA and total RNA. We treated cardiac endothelial cells and fibroblasts with CPC-EVs and measured inflammatory or fibrotic gene expression with qRT-PCR. We treated mesenchymal stromal cells with CPC-EVs and assessed migration in a Boyden chamber system. We treated cardiac endothelial cells seeded on Matrigel with CPC-EVs and measured tube formation. Finally, we put together the CPC-EV sequencing and experimental data in partial least squares and regularized regression models. Results: With our models we determined that the most important CPC-EV RNA signals involved in pro-reparative outcomes had significant fit to cardiac development and signaling pathways. Furthermore, using a model trained on previously collected CPC-EVs, we were able to predict in vitro outcomes for the CHILD clinical samples. Finally, we formed rank-based predictions for the CHILD lead-in patients, based on the in vitro assays. Conclusions: The CHILD phase I trial is set to conclude next year. We will validate our predictions with the release of these clinical data. With this work, we (1) investigated the EV RNA signals involved in pediatric cardiac repair and (2) will use our in vitro results to predict CHILD clinical trial outcomes. Our results highlight the importance of quantitative studies in determining the underlying variance in pediatric cardiac cell therapy.