Interrogating early molecular events of doxorubicin-induced cardiotoxicity at single-cell level to identify new cardioprotective agents

Abstract Funding Acknowledgements Type of funding sources: Foundation. Main funding source(s): Leducq Foundation Background Doxorubicin (DOX) is a highly effective chemotherapeutic agent widely used for treating various types of malignancies. Unfortunately, its clinical implication is hampered by cardiotoxic side effects which are responsible for increased mortality among cancer survivors compared to the general population. Preclinical and clinical studies have highlighted potential mechanisms of anthracycline-induced cardiotoxicity (AIC), but the core molecular basis remains largely unknown. More importantly, only a few studies have focused on the characterization of the molecular events that distinguish the early reversible phase of AIC from an advanced and irreversible state of the disease. Aim We aim to provide an unbiased and in-depth profile of early transcriptional changes induced by DOX at single cell resolution, to identify new druggable targets for the development of cardioprotective agents to prevent AIC. Methods Based on our previously established murine model of AIC, BALB/c mice were injected with saline (Vehicle) or DOX (3 weekly injections of 4 mg/kg), and hearts were collected at either 3 days (acute cardiotoxicity) or 6 weeks after the first injection (chronic cardiotoxicity). Nuclei were isolated from frozen hearts for single-nuclei transcriptomic (snRNAseq; 10x Genomics, Chromium Single Cell 3' v3.1). Seurat pipeline was mainly used for downstream bioinformatic analysis. Results Echocardiography revealed DOX-induced systolic dysfunction at 6 weeks, confirming the establishment of AIC. SnRNAseq data from 12 hearts allowed to identify 8 major cell types including cardiomyocytes (CM), endothelial and mural cells, fibroblasts, myeloid cells, B and T lymphocytes and neuronal-like cells. To unravel changes in CM, we performed unbiased subclustering which defined 6 different subpopulations, including a CM_Stressed population, characterized by enrichment of typical cardiac stress markers, such as natriuretic peptide hormone (Nppb) and myosin heavy chain 7 (Myh7). Differential abundance analysis showed enrichment of the CM-Stressed population at both 3 days and 6 weeks, suggesting that DOX drives a transcriptional change toward stress status in CM both at early and late stages of cardiotoxicity. Differential gene expression analysis revealed Golgi associated kinase 1B (Gask1B) to be significantly upregulated at both 3 days and 6 weeks in CM_Stressed. In agreement, preliminary experiments showed that Gask1b gene silencing partially rescued DOX-induced heart dysfunction in zebrafish model and hiPSC-derived CM, identifying Gask1b as a potential new player in DOX cardiotoxicity. Conclusions We generated a single-nucleus dataset of DOX-treated mouse hearts and identified transcriptional changes driven by DOX in CM at early and late stages of the disease. Furthermore, we identified Gask1b gene, whose role in cardiac pathophysiology was previously unappreciated, as a new potential marker and determinant of DOX cardiotoxicity.