Leukemic Cell-Intrinsic Innate Nucleic Acid Receptor Signaling Governs Efficacy of Hypomethylating Agents and Immunosurveillance in AML

Introduction: Hypomethylating agents (HMA), such as Azaciditine (AZA), are a long-established treatment option for acute myeloid leukemia (AML). Besides epigenetic reprogramming of malignant cell clones, HMAs are recognized to facilitate antineoplastic immunity. In vitro assessments in solid tumor models suggest that a major mechanism underlying AZA-triggered immune responses may be the activation of innate nucleic acid sensing pathways. As such, the RNA- and DNA-sensing receptor systems, RIG-I/MDA-5 and cGAS/STING – and subsequent type I interferon (IFN-I) signaling seem to be triggered by the transcriptional upregulation of otherwise DNA-hypermethylated and thus silenced endogenous retroviruses (ERVs). In AML, it is unknown whether similar processes are involved and/or by which molecular mechanisms, HMAs modulate anti-leukemic immunity against AML blasts, and how these processes are affected during acquired HMA resistance.

Methods: Murine (c1498 and Wehi-3B) and human (MOLM-13) AML cell lines were exposed to AZA in vitro and analyzed for expression of ERVs. Nucleic acid receptor sensing pathways and IFN-I genes are not usually DNA methylated by AZA at promoter regions. However, the re-expression of endogenous nucleic acids or ERVs could trigger their activation. Protein and transcriptional expression of major components of nucleic acid receptor sensing pathways and IFN-I genes were, therefore, analysed. To understand the importance of AML cell-intrinsic nucleic acid receptor signaling in HMA-induced immunosurveillance, syngeneic c1498 and Wehi-3B AML models were employed. Mice received 2 cycles of 2.5mg/kg AZA (6 days on, 5 days off, 6 days on) intraperitoneally to mimic the principal treatment course used in patients. To address the contribution of the nucleic acid-sensing pathways to AZA treatment efficacy in vitroand in vivo, AML cells with acquired AZA resistance (AZA-R) or gene engineered AML cells deficient in RNA receptors (RIG-I/MDA-5), or the RNA- and DNA-sensing adapter proteins (MAVS and STING) were employed.

Results: A targeted ERV screen revealed upregulation of various ERV transcripts in c1498 and Wehi-3B cells and was similarly observed in MOLM-13 AML cells following AZA exposure in vitro. Increased RNA transcript and protein expression of RIG-I/MAVS and (to a lesser degree) STING was observed, resulting in activation of downstream constituents including higher phosphorylation of the transcription factors IRF3/7 and NF-kB, IFN-I, and other ISGs. This was abrogated in RIG-I and MAVS, but not STING, deficient murine and human AML cells. Interestingly, expression of RIG-I/MDA-5, STING and all other downstream constituents, but not MAVS, were diminished in AZA-R AML cells. In mice harboring MAVS deficient AML (c1498), HMA treatment failed to show anti-leukemic activity resulting in rapid disease progress and death of recipient animals. Loss of STING in leukemic cells also impacted AZA treatment efficacy, although to a much less pronounced degree. These findings were similarly reproduced with Wehi-3B AML-engrafted mice. Following antibody-mediated depletion of T cells but not NK cells, mice rapidly succumbed to AML disease and demonstrated no anti-leukemic effect of AZA treatment. CyTOF and flow cytometric analysis also revealed increased bone marrow and spleen infiltration as well as activation of CD4⁺/CD8⁺ T cells, NK cells, and MHC-II⁺ APCs/Myeloid cells in AZA treated mice bearing wildtype AML, compared to untreated. This was not observed in AZA-treated mice bearing MAVS-deficient or AZA-R AML.

Conclusion: Our findings indicate that nucleic acid receptor signaling (specifically RIG-I/MAVS) within leukemic cells is imperative for AZA efficacy and immunosurveillance in AML. Whether these RNA-sensing receptors are triggered by ERVs or alternate (endogenous) RNAs remains to be determined. Elucidating the molecular mechanisms of how nucleic acid sensors within leukemic blasts shape treatment efficacy of AZA may guide development of systematic immunotherapeutic strategies for AZA-R AML by targeted combinatorial approaches.