Targeting Mitochondrial Metabolism to Promote Viral Mimicry and Type I Interferon Response in Acute Myeloid Leukemia

Acute myeloid leukemia (AML) cells possess different mechanisms to escape immune recognition. Type-I interferon (IFN) is primarily involved in combating viral infections. However, emerging preclinical and clinical evidence suggests that type-I IFN production within the tumor microenvironment can enhance anti-cancer immune activity and treatment outcomes. In line with this, several clinical trials have investigated the use of modified and recombinant type-I IFNs in acute lymphoblastic and myeloid leukemia patients to prevent disease recurrence following hematopoietic stem cell transplantation or post-remission therapy. Accumulation of cytosolic DNA can trigger the type-I IFN response via the cGAS-STING intracellular DNA-sensing pathway. While the use of IFN-alpha and STING agonists can elicit an anti-tumor immune response, their administration poses clinical challenges due to significant off-target toxicity and bioavailability issues. The release of mitochondrial DNA in the cytosol can activate the type-I IFN response, but this aspect has not been explored in AML. Given that the leukemia-initiating cells in AML are known to have high mitochondrial activity and critically depend on oxidative phosphorylation, we hypothesized that targeting mitochondrial metabolism could activate STING signaling and the type-I IFN response in AML cells.

Transcriptomic analysis of persisting AML cells post-induction therapy (cytarabine and doxorubicin) in an MLL-AF9 immune-competent C57BL/6 mouse model revealed that the persisting AML cells were negatively enriched for type-I IFN response and positively enriched for myc, oxidative phosphorylation, unfolded protein response and fatty acid metabolism. The bone marrow of mice with AML was enriched in exhausted CD8 and CD4 T cells and showed changes in NK cell subpopulations, indicating a dysfunctional immune microenvironment

To understand the impact of mitochondrial metabolism on the type-I IFN response, we generated an ISRE (interferon stimulatory response element) luciferase-based reporter AML cell line (human MonoMac 6) to screen for mitochondria-targeting metabolic compounds that would potentially activate type-I IFN signaling. We discovered that Artesunate (ART; an anti-parasitic drug used to treat malaria), and Niclosamide (NA; an anti-helminthic drug used to manage tapeworm infections), elicited ISRE activity in a dose-dependent manner. Phosphorylation of STING, STAT1, and STAT2 (indicators of active type-I IFN signaling) was observed in the cells treated with ART but not with NA. To confirm that ISRE activation occurred through STING signaling, we used inhibitors H151 and GSK8612 in combination with ART and found that ISRE activation was prevented.

A MitoStress Test confirmed that ART and NA significantly affect mitochondrial metabolism in AML cells. Treatment of AML cells with ART, NA, or the BH3 mimetic Venetoclax (mitochondrial inhibitor used in the management of AML) promoted the leakage of mitochondrial DNA in the cytosol; however, only NA and ART increased the activation of IFN-stimulated genes, highlighting that not all mitochondria-targeting agents act in the same way.

In vivo, ART (60mg/kg) effectively reduced the leukemic burden in an MLL-AF9 AML mouse model as a single agent, while the sequential administration of conventional iCT followed by ART further reduced AML progression and increased mouse survival.

These findings show that altering the mitochondrial metabolism of AML cells can activate the type-I IFN response via the cytosolic DNA sensing pathway. We propose that ART can be repurposed as a potential STING agonist. Further studies are in progress to identify the effect of ART on immune cells in leukemic mice and test the use of ART as a STING agonist in different tumor models.