m⁶a Regulates Differentiation State and mRNA Translation in Myeloid Leukemia

N⁶-methyladenosine (m⁶A) is a nucleotide modification in mRNA that is required for the acquisition of cell fate in embryonic stem cells. Recent studies have indicated that methylation writers can act as both oncogenes and tumor suppressor genes. Here we show that m⁶A is a critical regulator of myeloid differentiation of human hematopoietic stem and progenitors and myeloid leukemia cells. Depletion of the m⁶A-forming enzyme METTL3 with shRNAs in human cord blood derived CD34+ hematopoietic stem/progenitor cells (HSPCs) decreased global m⁶A levels, promoted differentiation, and reduced cell proliferation and colony formation. Conversely, overexpression of wild type METTL3-WT, but not the catalytic dead form METTL3-CD, in HSPCs increased global m⁶A levels, inhibited differentiation, and promoted cell growth.

We surveyed METTL3 expression across different cancers and found that METTL3 is expressed at the highest level in acute myeloid leukemia (AML) compared to a broad range of solid tumors. METTL3 was highly expressed in wide variety of human AML cell lines (11/11) and in primary AML patients (3/3). Utilizing 2D-Thin layer chromatography (TLC), we found that m⁶A levels in poly(A) mRNAs were significantly increased in the MOLM13 leukemia cells and primary AML patient cells compared to normal HSPCs. To directly address the role of m⁶A in human myeloid leukemia cells, we demonstrated that METTL3 depletion with shRNAs resulted in reduced cell proliferation and the induction of apoptosis in human AML cell lines (MOLM13, NOMO-1 and KASUMI-1), and delayed leukemia in vivo (MOLM13 cells - median survival of 23 days control versus 35 and 36 days shRNA#9 and shRNA#12 respectively - in vivo knockdown was negatively selected as determined by immunoblotting). Validating an on-target effect, expression of shRNA-resistant form of METTL3 rescued both the apoptotic and differentiation phenotypes in METTL3-depleted cells and CRISPR/Cas9 mediate depletion of METTL3 (using two independent guide RNAs) resulted in similar phenotypes. Additionally, we looked into the rank of m⁶A "writers" and "erasers" in the genome-wide CRISPR-based screen (Wang et al. Cell 2017) for genes essential for survival in 14 AML cell lines. Interestingly, while all members of the "writer" complexes, METTL3, METTL14, WTAP and KIAA1429, were scored highly (at the top 10%), the eraser ALKBH5 showed no essentiality and FTO is only important for survival of EOL-1 cells. This data suggests that the m⁶A writer complex is required for leukemia.

To gain insights into the mechanism of why m⁶A and METTL3 are required in leukemia, we performed single-nucleotide resolution mapping of m⁶A in MOLM13 cells using miCLIP, gene expression analysis, ribosome profiling and reverse phase protein arrays in MOLM13 cells depleted for METTL3. We profiled mRNAs targeted for m⁶A modifications and found that transcripts with m⁶A were more stable but were translated with less efficiency. Gene Set Enrichment Analysis revealed negative enrichment of MYC and ESCs gene sets. We further validated c-MYC, BCL2 and PTEN as targets for m⁶A modifications. We performed meRIP-qPCR and showed that knockdown of METTL3 specifically reduced enrichment of m⁶A at mapped sites of these transcripts. We demonstrated that c-MYC, BCL2 and PTEN protein expression were reduced despite a 2-5 log₂ fold increase in mRNA expression after METTL3 depletion. The effects were reversed in cells overexpressing METTL3-WT, but not the METTL3-CD. Importantly, we observed a robust activation of p-AKT upon METTL3 knockdown, and treatment with inhibitors of PI3K and AKT partially rescued the differentiation effects in METTL3 depleted cells. Overall, m⁶A is critical for maintaining the differentiation program in the hematopoietic system and that this process is dysregulated in myeloid leukemia. Our data provides a rationale for targeting the mRNA methylation program in myeloid leukemia.