Therapeutic Targeting of Ferroptosis Pathway in Combination with Mitochondrial Oxidative Stress Induction in Acute Myeloid Leukemia

Background: Targeting apoptosis pathways in cancer has been extensively studied including the recent breakthrough therapy using venetoclax in acute myeloid leukemia (AML). However, certain therapy-resistant subsets of AML still pose a clinical challenge, leading us to investigate therapeutic strategies to bypass apoptosis pathways. Recently, non-apoptotic regulated cell death modes have attracted attention. Ferroptosis is a form of regulated cell death characterized by reactive oxygen species (ROS) induction and iron-dependent lipid peroxidation. Leukemia cells exhibit increased oxidative stress, which is further enhanced by iron overload, but only limited studies have explored therapeutic potential of targeting ferroptosis in AML. With mitochondria being the center for ROS production and iron metabolism containing 20-50% of cellular iron, mitochondrial (mito-) regulation of ferroptosis remains unclear. Incorporating our recent knowledge on mito-protease ClpP, hyperactivation of which selectively kills cancer cells through mito-proteolysis and oxidative stress (Ishizawa et al. 2019 Cancer Cell), we here aim to study the therapeutic targeting of the ferroptosis pathway and potential combinatorial targeting of mitochondria in AML.

Results: We first utilized the TCGA dataset to analyze gene expression of key components in one of the major anti-ferroptosis pathways cysteine-glutathione-GPX4 axis. AML patients with higher mRNA expression of SLC7A11 (cystine importer), GCLM (a subunit of rate-limiting glutathione synthetase) or GPX4 have significantly shorter overall survival (p < 0.05), suggesting potential prognostic impact of this pathway. Here we focused on GPX4, the most downstream molecule that reduces lipid hydroperoxide to block ferroptosis. Although GPX4 is essential for embryonic development in mice, acquired depletion with conditional knock out was reported to have no significant effect on the number and function of hematopoietic stem cells, suggesting the tolerability of GPX4-targeted therapy.

We demonstrate that GPX4 inhibition with sub-micromolar concentrations of ML210, a covalent GPX4 inhibitor with proteome-wide specificity, or with shRNA-mediated knockdown induces prominent cell death in various AML cell lines. The anti-leukemia effects were associated with lipid peroxidation and were almost completely abrogated by a lipophilic antioxidant Liproxstatin-1 (Lip1). The effect of ML210 was also blocked by the iron chelator deferoxamine, indicating ferroptosis. We also demonstrate that GPX4 knockdown increases ROS prominently in mitochondria before cell death is induced, suggesting that mitochondria is involved in the anti-leukemia effects. Meanwhile, GPX4 is among the top 15 sensitization hits in a previously published genome-wide CRISPR screening of leukemia cells treated with potent ClpP agonists ONC201 and ONC212 (Jacques et al. 2020 Genetics), indicating the protective function of GPX4 against mito-oxidative stress. We then found that genetic or pharmacologic hyperactivation of ClpP upregulated GPX4 protein expression predominantly in mitochondria in AML cells, collectively suggesting that upregulation of mito-GPX4 is cell-protective against ClpP-mediated cell death. Indeed, combinatorial ClpP hyperactivation enhanced the anti-leukemia effects of GPX4 inhibition through increased induction of mito-ROS and lipid peroxidation. We further demonstrate that the synergistic cell killing is inhibited by Lip1, indicating that the dual targeting of ClpP and GPX4 induces lipid peroxidation-mediated cell death. Furthermore, the inhibition of lipid peroxidation simultaneously resulted in over 80% reduction of mito-ROS, unexpectedly. This suggests that most of the increased mito-ROS by this combination is associated with lipid peroxidation. Mechanisms of induction of mitochondrial lipid peroxides jointly engaged by ClpP and mito-GPX4, as well as its involvement in the induced oxidative cell death, have yet to be explored.

Conclusion: Our data suggests the potential involvement of mitochondrial lipid peroxidation in the anti-leukemia effects of GPX4 inhibition, along with its therapeutic potential in conjunction with mito-oxidative stress induction though instability in mito-proteome. Further investigations are in progress to assess the molecular mechanisms and the in vivo efficacy of the combinatorial treatment.