Unidirectional synergy between ferroptosis and apoptosis reveals a novel therapeutic strategy for refractory AML

Abstract

Abstract Background: Apoptosis resistance drives treatment failure in relapsed/refractory (R/R) acute myeloid leukemia (AML). Targeting non-apoptotic forms of regulated cell death, such as ferroptosis, presents a promising strategy to circumvent this resistance. Ferroptosis is characterized by iron-dependent lipid peroxidation (LP), with glutathione peroxidase 4 (GPX4) known as a major negative regulator among other factors. In AML, we previously reported that ferroptosis uniquely relies on mitochondrial LP, termed “mitochondrial ferroptosis” (Leukemia, 2024). Given that both apoptosis and ferroptosis are regulated by mitochondria in AML cells, we sought to explore the molecular crosstalk between these two distinct forms of cell death. Results: To test whether ferroptosis bypasses apoptotic resistance, we first treated venetoclax (VEN)-resistant AML cells including BAX/BAK double-knockout (DKO) AML cells with the selective GPX4 inhibitor ML210. ML210 effectively induced cell death in these models, confirming that ferroptosis involves a distinct, apoptosis-independent pathway. However, combining ML210 with VEN triggered synergistic cell death in both VEN-sensitive MOLM13 and VEN-resistant OCI-AML3 cells, with combination indices of 0.53 and 0.71, respectively, suggesting hitherto unknown molecular crosstalk between the two cell death pathways. The synergy was also observed in CD34⁺CD38⁻ leukemic stem/progenitor cells from VEN-resistant AML patients (combination indices < 0.5). Consistently, combined treatment with GPX4 inhibition and VEN significantly reduced peripheral leukemic burden in a patient-derived xenograft mouse model established from an R/R AML case previously treated with decitabine and VEN. To investigate the synergistic mechanisms between apoptosis and ferroptosis, we investigated whether ferroptosis induction supports apoptosis activation. The co-treatment with ML210 and VEN (ML210/VEN) reduced caspase 3 activation (cleaved caspase 3) compared to VEN alone. Notably, this suppression was reversed by the ferroptosis inhibitor ferrostatin-1 (Fer-1), indicating that ferroptosis inhibits apoptosis induction. In contrast, VEN significantly enhanced LP when combined with ML210. Remarkably, the synergistic effects of ML210/VEN were completely abrogated by Fer-1, indicating that the ferroptotic component is the dominant driver of the observed synergy. These findings suggest a unidirectional interaction: BCL2 inhibition promotes ferroptosis, whereas ferroptosis suppresses apoptosis. Interestingly, the mitochondria-targeted antioxidant MitoTEMPO completely blocked both LP and cell death induced by ML210/VEN, suggesting that mitochondrial ROS is a key trigger for the synergistic effects. Although VEN did not affect the transcriptional or protein expression of canonical ferroptosis regulators (e.g., GPX4, FSP1, ACSL4), metabolomics analysis revealed reduced glutathione levels. This suggests that BCL2 inhibition metabolically primes AML cells into a pro-ferroptotic state. Based on these findings, we optimized treatment using a sequential approach with VEN followed by ML210, which further enhanced LP and cell death while preserving caspase activation, compared to concomitant treatment. In vivo, VEN followed by doxycycline-inducible GPX4 knockdown significantly reduced leukemic burden and prolonged the survival in a xenograft mouse model. Strikingly, this sequential combinatory effect was completely abolished in BAX/BAK-DKO AML cells, suggesting a novel non-apoptotic role of BAX/BAK in ferroptosis regulation. Finally, given the current lack of clinically translatable GPX4 inhibitors, we tested FDA-approved repurposing agents as ferroptosis inducer and enhancer, demonstrating that this strategy is also operational by targeting alternative ferroptosis regulators besides GPX4. This finding further supports the clinical translatability of our therapeutic concept. Conclusion: Synergistic AML cell death caused by dual induction of mitochondrial ferroptosis and apoptosis involves paradoxical molecular interactions between the two cell death pathways: BCL2 inhibition enhances ferroptosis, which ultimately dominates the synergy despite the apoptosis-suppressive effects of GPX4 inhibition. Sequential induction of apoptosis followed by ferroptosis maximizes this synergy and may provide a promising future therapeutic strategy for R/R AML.