The efficacy of METTL3 inhibition in pre-clinical models of MDS and AML.

Abstract

e18584 Background: Splicing factor (SF) mutations commonly occur in myelodysplastic syndrome (MDS) and portend a poor prognosis. N6-methyladenosine (m 6 A) is an RNA post-transcriptional modification that is influenced by splicing. METTL3, the enzyme responsible for depositing m 6 A on RNA, is upregulated in acute myeloid leukemia (AML) and effective inhibitors of this enzyme (METTL3i) that have preclinical efficacy in AML are in early-stage solid tumor clinical trials. The role of m 6 A and therapeutic potential of METTL3i in MDS pathogenesis is currently unknown. Our lab previously demonstrated that deletion of METTL3 during murine embryonic development resulted in production of double-stranded RNAs (dsRNAs), and subsequent induction of an innate immune response. This mechanism has recently been shown to drive preclinical anti-tumor efficacy in solid tumors and to enhance cure in mouse solid tumor models when combined with immune checkpoint inhibition. Methods: We utilized bioinformatic analyses combined with well validated preclinical models of splicing factor mutated MDS and AML to determine whether METTL3i can be leveraged in the treatment of MDS/AML and to elucidate the consequences of METTL3i in the hematopoietic context. Results: When evaluating m 6 A modifications across multiple cell lines, MDS driver genes are m 6 A modified. The most modified genes are KMT2A, NIPBL, ATRX, CUX1, and NF1. We utilized a highly specific METTL3i, STM3765, in vivo in transplantation models of MDS/AML. We show that METTL3i results in improved mouse overall survival. Additionally, leukemic grafts displayed macrophage infiltration with M1 polarized macrophages as determined via single cell RNA sequencing (scRNAseq) and histologic analysis. In our MDS competitive transplantation model we demonstrate a differential decrease in splicing factor mutated hematopoietic stem and progenitor cells (HSPCs) following METTL3i via analytical flow cytometry. Single cell RNAseq demonstrated induction of an innate immune signature in METTL3 treated mice, in particular in wildtype cells. In SF mutant cells, a hyperinflammatory signature was present at baseline, although within this immune signature, differences were noted in genes associated with tumor infiltration in innate immune cell subsets and viral recognition in HSPCs. Conclusions: We utilized multiple myeloid malignancy models to show that METTL3i has activity against murine models of leukemia as well as MDS. This is likely related to m 6 A modification of multiple MDS driver genes as well as induction of an anti-viral innate immune response secondary to aberrant dsRNA formation. Mutant, but not wildtype HSPCs and immune cells express inflammatory genes at baseline that are further modified by METTL3i. In future studies we will employ our sophisticated MDS patient derived xenograft model in cytokine humanized mice with the goal of bringing manipulation of the epitranscriptome to the clinic for MDS.