m6A Maintains Hematopoietic Stem and Progenitor Cell Identity

Abstract N-6-methyladenosine (m6A) is one of the most abundant posttranscriptional modifications in eukaryotic mRNAs and long noncoding RNAs. We previously found a critical role for m6A in promoting human myeloid leukemia (Vu et al. Nature Medicine 2017). Targeting the RNA methylation program in leukemias has been suggested as a potential novel therapeutic strategy. However, it is unknown whether the m6A modification controls normal adult hematopoiesis and hematopoietic stem cells (HSC) function. To investigate the role of m6A in adult hematopoiesis, we crossed with the Mettl3 conditional knockout (cKO) mouse model with the interferon inducible Mx1-Cre system to abolish m6A in the hematopoietic compartment. Deletion of Mettl3 (3 weeks post pIpC injections) resulted in pancytopenia (white blood count 9.44 k/ul in WT versus 4.35 k/ul cKO) and a 55% reduction in red blood cell counts and nearly 70% loss in platelet counts. Most remarkably, METTL3 depletion resulted in a 5-fold increase in the number of overall hematopoietic stem and progenitors HSPCs (LSKs; Lin-c-kit+Sca1+). Within the HSPC compartment there was a about a 10-fold expansion in immunophenotypic long term hematopoietic stem cells LT-HSCs (Lin-c-kit+Sca1+CD150+CD48-), multipotent progenitors (MPP-2s), (Lin-c-kit+Sca1+CD150+CD48+) and MPP4s (Lin-c-kit+Sca1+CD150-CD48+). In contrast to this general increase in HSPCs, we observed a decrease in common myeloid progenitor (CMP), granulocyte-monocyte progenitor (GMP) populations by 70% and 60% respectively. Altogether, these results suggest that loss of METTL3 results in a partial blockage in hematopoietic stem and progenitor cell differentiation, and an accumulation of LT-HSC and MPPs. Despite this phenotypic expansion of LT-HSCs and MPPs, their function was impaired as demonstrated by a reduction in long-term chimerism in non-competitive transplants into congenic mice compared to the remaining wildtype host cells (66% in WT vs 26% in cKO). Interestingly, the relative differentiation block and accumulation of HSCs remained after transplantation. Furthermore, Mettl3 cKO HSCs are less quiescent (79% WT vs. 44% cKO) and more proliferative based on cell cycle profiling with pyronin Y/Hoechst or Ki67/Hoechst. RNA-seq in sorted LT-HSC, MPP1, MPP2 and MPP4s from WT and Mettl3 cKO mice demonstrated that Mettl3 cKO HSCs lose the HSC signature while MPP1 and MPP2 cells exhibited gene signatures resembling WT HSCs. Interestingly, genes uniquely upregulated in Mettl3 cKO LT-HSC showed a significant enrichment of the Mettl3 KO mouse embryonic stem cells (mESCs) expression signature. To decipher METTL3 regulation of transcriptional states in individual cells along the hematopoietic hierarchy, we performed single-cell RNA-seq using sorted WT and Mettl3 cKO cKit+ cells. Surprisingly, tSNE analysis of scRNA-seq data uncovered the loss of the normal HSC cluster and the emergence of three unique HSC-like populations in Mettl3 cKO mice. Single cell analysis identified a reduction in c-MYC levels in the HSPCs that was then confirmed by immunofluorescence in sorted HSCs. More importantly, when forced into division, Mettl3 cKO HSCs exhibited increased symmetric c-MYC low division pattern compared to control (60% to 48%) based on c-MYC IF staining in paired daughter cells. We further confirmed that sorted HSCs underwent increased symmetric renewal divisions in the Mettl3 cKO compared to controls (65% to 49%) without any effect on asymmetric division based on NUMB IF staining in paired daughter cells. These data suggest that without METTL3, HSCs are partially blocked in a self-renewing state albeit with reduced fitness compared to wildtype HSCs. Overall, our studies uncovered a novel role for METTL3 and RNA methylation to maintain normal HSC identity and progenitor differentiation. Also, these studies suggest that inhibiting METTL3 could result in significant hematopoietic defects. Disclosures No relevant conflicts of interest to declare.