FerroEVs drive serial propagation of pharmacologically induced ferroptosis in acute leukemia cells

Abstract

Abstract Pharmacologically induced cell death not only eliminates target cells but also modulates neighboring cells through bioactive molecules. While extracellular vesicles (EVs) derived from apoptotic cells, commonly referred to as apoptotic bodies, have been well characterized, those originating from ferroptotic cells remain poorly understood. Here, we provide proof of concept that ferroptosis-associated extracellular vesicles (ferroEVs) mediate death propagation in leukemia cells. Erastin treatment of U937 leukemia cells (IC₅₀: 19.6 µM) induced ferroptosis, as evidenced by increased reactive oxygen species (ROS; 40.1% vs. 8.1%), elevated fatty acyl-CoA synthetase long-chain family member 4 (FACL4/ACSL4), reduced glutathione peroxidase 4 (GPX4) expression, increased lipid peroxidation (12.3 vs. 7.2), and mitochondrial shrinkage; all effects were reversed by ferrostatin-1, confirming ferroptosis as the predominant mode of cell death. Similar ferroptotic changes were also observed in Jurkat cells, supporting the generality of these findings across leukemia subtypes. FerroEVs were purified from ferroptotic U937 cells using a standard ultracentrifugation protocol. Scanning electron microscopy confirmed round vesicles approximately 150 nm in diameter, and immunoblotting demonstrated the expression of CD63 (a tetraspanin) and tumor susceptibility gene 101 (TSG101, an ESCRT component). Morphological and surface marker profiles were comparable between ferroEVs and EVs derived from untreated cells, indicating similar phenotypic properties. Notably, these ferroEVs are unlikely to represent apoptotic bodies but rather bona fide exosomes, given their small size and the absence of apoptotic markers such as Annexin V and cleaved caspase-3. To investigate functional intercellular propagation, U937-derived ferroEVs were added to naïve U937 cultures. FerroEVs increased cell death by approximately 2.1-fold (~108% increase; 22.0% vs. 10.6% for control EVs, p<0.01) and elevated ROS levels to 31.1% ± 2.3% compared with 25.5% ± 1.6% (p<0.05), without increasing Annexin V positivity, indicating ferroptosis induction. Furthermore, ferroEVs purified from these secondary cultures were introduced into a third batch of U937 cells, which similarly underwent ferroptosis, demonstrating serial propagation of ferroptotic cell death. A similar phenomenon was also observed in Jurkat cells, underscoring the generalizability of this ferroEV-mediated death cascade across leukemia subtypes. Mechanistically, comprehensive microRNA-array profiling of ferroEV cargo identified a distinct set of redox-regulating miRNAs enriched versus EVs from untreated cells. Pathway enrichment analysis (KEGG : Kyoto Encyclopedia of Genes and Genomes) mapped these miRNAs to glutathione metabolism, lipid peroxidation, and iron-handling pathways, converging on the GPX4/FACL4 axes. In line with this signature, recipient cells exposed to ferroEVs exhibited sustained ROS accumulation and lipid peroxidation—hallmarks of ferroptotic stress—supporting a model in which ferroEV cargo rewires antioxidant defenses and propagates ferroptotic cell death across neighboring leukemia cells. Collectively, these findings demonstrate that ferroptotic leukemia cells actively release ferroEVs that non–cell-autonomously propagate ferroptosis-like cytotoxicity to neighboring cells. This ferroEV-mediated oxidative stress transmission provides proof of concept for amplifying ferroptotic pressure within leukemic populations, offering a potential strategy to eradicate therapy-resistant clones.