Similar to DNA methylation, m⁶A is a reversible and dynamic process regulated by three categories of enzymes, namely, “writers,” “erasers,” and “readers” ( Figure 1 ). At present, FTO and ALKB homolog 5 (ALKBH5) are the only two identified “erasers” that are responsible for reversing m⁶A (5, 7) ( Figure 1 ). FTO, the first m⁶A demethylase identified in 2011, has strongly promoted the development of research on m⁶A. The demethylation activity of ALKBH5 significantly impacts mRNA export, RNA metabolism, and mRNA processing factor assembly (8). The final biological function of m⁶A is mainly associated with m⁶A “readers” that recognize sites of m⁶A and induce it to bind to the target sites to perform different functions involving mRNA degradation, translation, splicing, stability, and export (9–15) ( Figure 1 ). m⁶A can be recognized by a set of RNA-binding proteins, including YT521-B homology (YTH) domain family proteins (YTHDF1/2/3, YTHDC1/2) (10–13, 16), insulin-like growth factor 2 mRNA-binding proteins (IGF2BPs, including IGF2BP1/2/3) (15), heterogeneous nuclear ribonucleoproteins (including HNRNPA2B1, HNRNPG, and HNRNPC), and eukaryotic translation initiation factor 3 (eIF3) (9, 17, 18). m⁶A is installed by methyltransferases (writers), which comprise several different “writer” proteins, including methyltransferase-like 3/5/14/16 (METTL3/5/14/16), Wilms tumor 1-associated protein (WTAP), Vir-like m⁶A methyltransferase associated (VIRMA, also called KIAA1429), RNA binding motif protein 15/15B (RBM15/15B), zinc finger CCHC-type containing 4 (ZCCHC4), and zinc finger CCCH-type containing 13 (ZC3H13) (19) ( Figure 1 ). Among the methyltransferases, METTL3 is the only one that has the S-adenosyl methionine (SAM)-binding protein in the catalytic pocket and composes a stable methyltransferase complex (MTC) heterodimer with METTL14 at 1:1 to exert methylation activity (20). Moreover, the activity of the METTL3/METTL14 core complex is assisted by an additional regulatory complex (known as MACOM, a m⁶A-METTL-associated complex) composed of WTAP, VIRMA, RBM15/15B, and ZC3H13. WTAP contributes to the heterodimer being located in nuclear speckles to complete m⁶A modification (21, 22). VIRMA recruits the catalytic core complex METTL3/METTL14/WTAP in the 3′UTR and is near the stop codon to methylate (23). RBM15/15B, as an X-inactive specific transcript-binding protein (XIST-binding protein), is associated with XIST-mediated gene silencing and regulates the m⁶A modification in XIST (24). ZC3H13 interacts with the m⁶A machinery and contributes the MTC to the mRNA-specific sites by binding factor Nito (25). ZCCHC4 acts on 28S rRNA by m⁶A modification and impacts mRNA translation (26, 27). METTL5 is the 18S rRNA m⁶A methyltransferase (26). MTTL16, a novel methyltransferase, is responsible for modifying the m⁶A modification of A43 in U6 small nuclear RNA and catalyzing m⁶A within a hairpin in MAT2A (28) ( Figure 1 ).

Because METTL3 plays a critical role in catalytic activity, many studies regarding its biological function in cancers have been widely reported, including lymphoma (29, 30), leukemia (31–33), breast cancer progression (34), liver cancer (35), glioblastoma (36, 37), bladder cancer (38, 39), gastric cancer (40), and lung cancer (41, 42). However, studies on m⁶A modification in hematology have not been systematically summarized thus far. Therefore, we will mainly focus on the role of METTL3 in normal hematopoiesis and hematological neoplasms in the present review and discuss directions for future research and potential clinical applications of METTL3 in hematological diseases.

METTL3 (also known as MT-A70), a 70-kDa protein, was first identified as a m⁶A “writer” and is highly conserved in eukaryotes from yeast to human (43). Although METTL3 forms a stable 1:1 heterodimer structure with METTL14 to exert higher methylation activity, it has been identified as the core catalytic enzyme of m⁶A methylation, and METTL14 mainly plays a role in the structure of MTC stabilization and recognizes target RNAs (20, 24, 44, 45). In contrast to METTL14, METTL3 contains S-adenosyl methionine (SAM)-binding protein and its product S-adenosyl homocysteine (SAH) in the catalytic pocket, which were not observed in METTL14 (20, 44–46) ( Figure 2 ). In addition, the catalytic site of METTL3 contains a more conserved DPPW motif involved in coordinating the adenine of the acceptor substrate, while METTL14 has a more divergent EPPL sequence (20, 46, 47). After replacing the DPPW motif with APPW(D395) in METTL3, the methylation of the METTL3–METLL14 complex was significantly destroyed but was very lightly affected after changing EPPL to APPL in METTL14 (44). Moreover, due to the collision between the adenine moiety and the side chain (Trp211 and Pro362) residues in METTL14, the binding of SAM would be prevented in the METTL14 catalytic site (46, 48). Furthermore, METTL3 also contains two CYS-CYS-HIS (CCCH)-type zinc binding motifs, which are critical for RNA methylation in vitro (44, 49). The methyltransferase domain of METTL3 (MTD3) presents a classic α–β–α fold, including a mixed eight-strand β-sheet, four α-helices, and three 310 helices, which makes a special catalytic cavity for METTL3, while the catalytic site of METTL14 is relatively occluded (20, 44). The methyltransferase domain of METTL14 (MTD14) contains residues 165–378, which is near the N-terminal α-helical motif (NHM, residues 116–163) and the C-terminal motif (CTM, residues 380–402) (20). MTD3 mainly contains residues 369–570, making three loops to fence the METTL3 catalytic cavity: gate loop 1 (residues 396–410), gate loop 2 (residues 507–515), and interface loop (residues 462–479) (20, 44). The two gate loops are adjacent to the SAM binding site and are associated with adenosine recognition, and the interface loop with the longer sequence allows METTL3 and METTL14 to bind each other tightly (20, 44, 50). Meanwhile, 11 residues of METTL3 are involved in SAM coordination, including D377, I378, Q550, N549, R536, D395, K513, H538, N539, E532, and L533 (20). Wang et al. further found that mutations of these residues completely abrogated methyltransferase activity (D377A, D395A, N539A, and E532A) or moderately weakened enzyme activity (R536, H538, N549, or Q550), while corresponding mutations of METTL14 have little effect on catalytic activity (20, 44). Between the interface of METTL3 and METTL14, a highly conserved groove comprises Arg465, Arg468, His474, and His478 of METTL3 and Arg245, Arg249, Arg254, and Arg255 of METTL14, which contributes to internal RNA binding (20, 46).

Furthermore, METTL3 has also been reported to promote translation independently of its methyltransferase activity or downstream m⁶A reader proteins (51, 52). The function of METTL3 in the cytoplasm promoting translation is to recruit the initiation factor eIF3 h to the translation initiation complex (51, 52). It can enhance epidermal growth factor receptor (EGFR) and TAZ protein expression independent of YTHDF1.

Normal hematopoiesis plays a vital role in hematology, presenting multipotent hematopoietic stem cells (HSCs) being differentiated into various mature cells of the blood system. The process is complex and multistep and is regulated by many factors, such as the transcription factor PU.1 for myeloid cells and C/EBPα for granulocyte/macrophage progenitors.

The first populations of hematopoietic stem cells (HSCs) are mainly produced by hemogenic endothelial cells (ECs), which later acquire a cell morphology and gene expression consistent with hematopoietic identity in a process called endothelial-to-hematopoietic transition (EHT) (64). In zebrafish, m⁶A has been reported to determine cell fate when EHT progresses to the earliest hematopoietic stem/progenitor cells (HSPCs) during embryogenesis (14). METTL3 was found to be enriched in sorted endothelial cells and hemogenic endothelium, correspondingly affecting hematopoiesis (14). The deletion of METTL3 results in impaired HSPC differentiation by activating Notch1 signaling. Increasing Notch signaling can abrogate the generation of hematopoietic cells by maintaining endothelial identity in EHT (14, 64). Notch1a m⁶A enrichment is significantly decreased in METTL3 morphants, but the expression of Notch1a is increased in endothelial cells, resulting in a decrease in HSPC generation (14). Conversely, overexpression of METTL3 inhibiting Notch1 activity could rescue this phenomenon in zebrafish (14). Additionally, the same phenotype was also observed in mice with METTL3 knockdown. In 2018, this team reported that METTL3 promotes HSPC generation by inhibiting Notch1 signaling in endothelial cells of the Vec-Cre; METTL3^fl/fl mouse aorta-gonad-mesonephros (AGM) region, consistent with the phenomenon in zebrafish (53). In 2019, Heather Lee et al. performed a study on Mx1-cre; METTL3^fl/fl mice and discovered that METTL3 deletion has little impact on HSC self-renewal and quiescence but significantly affects HSC differentiation (54). The deletion of METTL3 has resulted in a blocking of HSCs and an accumulation of HSCs by reducing Myc mRNA translation (54). Deleting METTL3 in myeloid cells from Lysm-cre; METTL3^fl/fl mice, they found that METTL3 is not indispensable for mature myeloid cell maintenance or function (54). Via Mx1-cre; METTL3^fl/fl mice, Cheng et al. reported that METTL3 depletion in normal murine HSCs results in a decrease in Myc mRNA and protein levels. Furthermore, METTL3 is required for normal hematopoiesis and maintains HSC symmetric commitment and identity by controlling Myc abundance in differentiating HSCs and Myc mRNA stability (55). Metll3 ablation can impair the differentiation of myeloid, megakaryocytes, and erythroid lineages, leading to an additional population that molecularly and functionally resembles multipotent progenitors (55). m⁶A loss by deleting Metll3 in mice blocked HSC transition to myeloid progenitors, notably presenting as decreases in common myeloid progenitors (CMPs) and granulocyte myeloid progenitors (GMPs) (55). Meanwhile, they also found a cell-intrinsic role of Metll3 in regulating HSC number and function in a bone marrow competitive transplantation trial (55). In our previous study, we discovered that the METTL3-mettl14 methyltransferase complex plays a vital role in regulating HSC self-renewal in adult mouse bone marrow, and METTL3 is mainly responsible for HSCs being in a quiescent state in mice (56). After conditional knockout of METTL3, Metll14, or both in mice, we found that the depletion of METTL3 is in charge of expanding phenotypic HSCs in adult mouse bone marrow and promotes the HSC cell cycle by regulating HSC self-renewal genes such as Nr4a2, p21, Bmi-1, and Prdm16 (56). In human HSPCs from cord blood, the depletion of METTL3 has been discovered to enhance cell differentiation, inhibit cell proliferation with fewer colony-forming units (CFUs) in all lineages, and hardly affect the apoptosis of HSPCs (31). Additionally, METTL3 absence contributes to myeloid differentiation, and METTL3 mRNA is expressed at lower levels in mature differentiated myeloid cells in both mouse HSCs and HSPCs (31). Taking human erythroleukemia (HEL) cells as a surrogate model for studying erythropoiesis, Kupper et al. also found that METTL3 blocked erythropoiesis by impacting the stage-specific gene expression of erythroid progenitors, such as the erythroid transcription factors GYPA, HBA1, SPTB, EPOR, and ALAS2 (57).

As METTL3 plays a significant role in normal hematopoiesis, an increasing number of studies on hematology malignancies have been reported in recent years, including AML, acute lymphocytic leukemia (ALL), CML, and lymphomas.

AML is a common hematological malignancy that is mainly caused by gene mutations and chromosomal aberrations resulting in changes in gene expression and, sequentially, aberrant growth and differentiation of hematopoietic stem cells (HSCs) (65). In 2017, Ly Vu et al. discovered that METTL3 has higher expression in AML cells than in normal HSPCs (31). In addition, they found that METTL3 disruption promotes the differentiation and apoptosis of AML cells both in vitro (MOLM-13 cells) and in vivo, which indicated that METTL3 affects the undifferentiated state and growth of leukemia cell lines (31). By connecting the single-nucleotide-resolution mapping of m⁶A and ribosome profiling, they revealed that MELLT3 deletion reduced the translation efficiency of c-MYC, BCL2, and PTEN in MOLM-13, resulting in phosphatidylinositol 3-kinase-AKT (PI3K/AKT) pathway activation (31). Meanwhile, Isaia Barbieri reported that METTL3 is necessary for leukemia cell growth and in maintaining an undifferentiated state (32). They further found that METTL3 could be recruited by CEBPZ to promoters, which led to m⁶A methylation of the respective mRNAs and increased translation. Among the promoters, translation of SP1 was significantly reduced after deleting METTL3 (32). Wang et al. also reported that METTL3 was more highly expressed in immature cells than in mature monocytes, and its depletion significantly inhibited cell proliferation and decreased MYC expression and m⁶A levels on MYC mRNA (66). Recently, a group of researchers reported that METTL3 plays a role in inhibiting adipogenesis of bone marrow mesenchymal stem cells (MSCs) and blocking the chemoresistance of acute myeloid leukemia cells by regulating the PI3K/AKT signaling pathway (58). However, METTL3 expression was significantly decreased in AML-MSCs, which enhanced the adipogenesis and chemoresistance of AML cells (58). They found that METTL3 impacted the m⁶A modification of AKT1 mRNA, resulting in a decrease in the protein level of AKT1 and an increase in adipogenesis. Correspondingly, activation of the PI3K/AKT signaling pathway contributes to adipogenesis and AML chemoresistance in MSCs (58). In our recent study, we found that METTL3 is highly expressed in AML patients, which results in poorer prognosis than in AML patients without METTL3 expression (p=0.017). Furthermore, knockdown of METTL3 in AML cells (K562 and Kasumi-1) inhibited proliferation and increased apoptosis and differentiation by regulating the p53 signaling pathway. METTL3 deletion led to decreased MDM2 expression and MDM2 mRNA transcript stability, which activated the p53 signaling pathway (59).

CML is caused by the oncogenic BCR-ABL1 fusion gene, which is mainly treated with tyrosine kinase inhibitors (TKIs) (67). However, TKI resistance is still a challenge for CML patients and increases the risk of transfer to AML (68). In 2018, Zaira Ianniello et al. discovered that METTL3 is a novel oncogene in CML and potentially a therapeutic target for TKI-resistant CML. The m6A methyltransferase complex METTL3/METTL14 and METTL3 is upregulated in primary CML patients, and its downregulation significantly impairs the proliferation of both primary CML cells and TKI-sensitive and TKI-resistant CML cells (60). Silencing METTL3 in K562 cells and the TKI imatinib mesylate-resistant K562 cell line (K562r), they found that METTL3 affects the growth and viability of CML cells directly and indirectly. MYC, as a transcriptional activator, is notably affected by METTL3 in CML cells, including the protein, mRNA, and premRNA levels. METTL3 knockdown strongly reduced MYC expression at multiple levels in CML, which consequently regulated the genes associated with RNA metabolism (60). Moreover, they found that the PES1 protein was potentially involved in blocking the cell cycle in G1 phase after METTL3 knockdown in CML cells (60). They showed that METTL3 both regulates PES1 by methyltransferase activity in the nucleus and directly promotes PES1 translation in the cytoplasm independently of its catalytic activity (60). Lai et al. recently reported that LINC00470 and METTL3 played a role in chemoresistance and autophagy in CML by regulating phosphatase and tensin homologue (PTEN) (61). PTEN, a well-known tumor suppressor, suppresses the activation of PI3K/AKT signaling and subsequently inhibits AKT activity and its downstream pathways (69). In the study, they disclosed that PTEN expression was obviously lower in chemoresistant CML cells than in K562 parental cells, which was negatively associated with LINC00470 and METTL3 (61). More interestingly, they found that overexpression of LINC00470 shortened the half-life of PTEN mRNA and enhanced the binding of METTL3 to PTEN mRNA (61). The depletion of METTL3 in K562 cells reversed the downregulation and degradation of PTEN mRNA and protein induced by LINC00470 and recovered the normal level of m⁶A modification in PTEN (61). Accordingly, overexpression of METTL3/LINC00470 promoted chemoresistance and reduced autophagy in CML cells by regulating PTEN stability and activating AKT. Fang-Yi Yao et al. reported that METTL3 was downregulated in CML cells, resulting in a decrease in the protein level of nuclear enriched abundant transcript 1 (NEAT1) (70). Furthermore, METTL3 downregulation in CMLs reduced its ability to modify NETA1 m⁶A, subsequently enhancing CML cell viability and inhibiting CML cell apoptosis. NEAT1, a lncRNA, is crucial for composing the subnuclear structure paraspeckle and is associated with the progression of hematological malignancies (62).

Lymphocytic neoplasm comprises lymphoblastic leukemia and lymphoma. Similar to AML, acute lymphoblastic leukemia (ALL) is also a severe hematology malignancy and is the most common form of cancer in children (71). However, studies on m⁶A modification in ALL are significantly fewer than those in AML. Congcong Sun et al. reported that the expression of METTL3 was lower in children with ETV6-RUNX1-positive ALL (63). Meanwhile, they also found that the METTL3 expression level was reduced in ALL relapse patients compared with non-relapse patients (63). However, they did not find any correlation between METTL3 expression and some basic clinical characteristics, including age, sex, initial white blood cell count, blast percentage, and LDH level (63). In 2021, a five-center case–control study reported that METLL3 gene polymorphisms were strongly associated with pediatric ALL, mainly including rs1263801 C>G, rs1139130 A>G, and rs1061027 A>C polymorphisms (33). All three polymorphisms were reported to remarkably increase the risk of common B-type and MLL fusion-type ALL in southern Chinese children (33). Additionally, all three polymorphisms were also related to primitive/naive lymphocytes and MRD after chemotherapy. The study showed that patients carrying rs1263801 CC and rs1139130 AA would have a better response to South China Children Leukemia Group chemotherapeutics (SCCLG) chemotherapeutics, and Chinese Children Cancer Group Chemotherapeutics (CCCG) chemotherapeutics are more efficient for patients with rs1061027 (33).

Lymphoma is a well-known hematological neoplasm that mainly includes Hodgkin and non-Hodgkin lymphomas. Diffuse large B-cell lymphoma (DLBCL) is the most common neoplasm in non-Hodgkin lymphoma, which is an aggressive lymphoma with a median survival of <1 year in untreated patients (72). A study regarding the correlation of m⁶A modifications with DLBCL reported that the expression level of METTL3 is higher in DLBCL tissues/cell lines than in normal lymph nodes/cells (29). Additionally, higher expression of METTL3 facilitates the proliferation of DLBCL cell lines and viability by regulating the m⁶A mRNA modification of pigment epithelium-derived factor (PEDF), which was usually regarded as inhibitor of canonical Wnt signaling in previous studies (73, 74). Subsequently, they found that overexpression of PEDF can disable the inhibitory effects of METTL3 silencing on DLBCL cell proliferation (29). These results suggest that the METTL3/PEDF axis may have therapeutic potential for DLBCL, but more specific studies are needed for verification.