Abstract Acute myeloid leukemia (AML) is a common and aggressive hematopoietic malignancy. Leukemia stem cells (LSCs), which reside in specialized bone marrow (BM) niches, are recognized as the root cause of AML development and relapse. Recent studies from us and others have revealed critical roles of RNA modifications in leukemogenesis and LSC maintenance. However, although over 170 RNA modifications have been characterized, the biological functions of most have not yet been defined in AML. Here, we show that methyltransferase-like protein 1 (METTL1), a major enzyme depositing tRNA 7-Methylguanosine (m7G) modification, is a key regulator of LSC self-renewal and BM homing, and represents a druggable epitranscriptomic vulnerability in AML. Through an integrated analysis of genome-wide CRISPR/Cas9 screen data, we identified that, among all currently known RNA modification regulators (155 in total), METTL1 stood out as the one with the most promising clinical significance in AML for three reasons. (i) CRISPR/Cas9 screen showed that the survival of AML cells is more dependent on METTL1 than non-AML cancer cells (P = 0.0016); (ii) METTL1 is significantly elevated expression in AML patients (P = 1.56 × 10-6); and (iii) High levels of METTL1 are significantly associated with poor survival in AML patients (P = 0.015). METTL1 depletion in AMLs including primary patient-derived xenografts (PDX) (with MLL-r, FLT3ITD, or DNMT3A mutation) significantly suppressed cell growth, reduced LSC frequency in vitro, and delayed leukemia initiation and progression in vivo. These phenotypes could be fully rescued by wild-type METTL1 but not a catalytically inactive mutant (amino acids 107-109 EIR/AAA), demonstrating that METTL1's methyltransferase activity is required for its functions. Notably, scRNA-seq of hematopoietic cells from Mettl1 conditional knockout (cKO) mice showed minimal impact on normal hematopoiesis, suggesting a favorable therapeutic window. Mechanistically, transcriptomic and codon usage analysis revealed that METTL1 KO preferentially downregulated transcripts enriched in m7G-tRNA-dependent codons. Single-base resolution mapping of tRNA m⁷G showed that METTL1 KO led to decreased m⁷G abundance and reduced levels of tRNAPheGAA. This led to translation suppression and degradation of tRNAPheGAA-dependent transcripts, such as HCK, atyrosine-protein kinase that activates CXCR4 signaling. Disruption of HCK/CXCR4 signaling suppressed LSC self-renewal and BM homing. Thus, METTL1 specifically deposits m7G on tRNAPheGAA to increase its stability, which is needed for efficient HCK expression. This, in turn, amplifies CXCR4 signaling responsible for LSC retention in the BM niche and self-renewal. To translate these basic identifications into a potential clinical application, we screened small-molecule inhibitors targeting METTL1 with 264,086 compounds. The top 400 candidates were tested for their anti-leukemic efficacy in AML cell line (Mono-mac-6 with MLL-r). Then, the top 20 compounds with prominent inhibitory effects on AML cell viability were chosen for further validations:(i)In vitro anti-leukemia efficacy testing in multiple AML models including primary PDX blasts;(ii) Cell-free assays to measure inhibition of METTL1 methyltransferase activity;(iii) Cellular assays to assess METTL1 enzymatic activity suppression in AML cells. The top 'hit’ METTL1 inhibitor (M1i) potently inhibited METTL1 m7G methyltransferase activity both in cell-free system and in AML cells. Moreover, it directly interacted with METTL1 protein and showed low nanomolar IC50values in killing AML cells. Meanwhile, M1i treatment reduced tRNAm7G levels and disrupted the tRNAPheGAA/HCK/CXCR4 cascade. Notably, pharmacological inhibition of METTL1 with M1i significantly reduced LSC frequency and homing, delayed leukemogenesis, and prolonged survival in multiple myeloid leukemia models (e.g., MLL-r AML median survival: Ctrl 29 days vs M1i 46 days; P = 0.001; in vivo dosing: 0.2 mg/kg every other day ×10). Our study establishes the crucial role of the METTL1/tRNAPheGAA axis in LSC maintenance and leukemogenesis, and provides compelling proof-of-concept evidence that METTL1 is a druggable epitranscriptomic target for anti-leukemia therapy.
Ren et al. (Mon,) studied this question.