2.1.1 m6A methyltransferases

m6A writers are a class of methyltransferases that catalyze the formation of m6A by assembling m6A onto target RNA through a complex. The m6A methyltransferase complex (MTC) consists of a heterodimer of methyltransferase-like protein 3 (METTL3) and METTL14, Wilms tumor 1-associated protein (WTAP), RNA binding motif protein 15 (RBM15), RBM15B, Cbl proto-oncogene like 1 (CBLL1, also known as HAKAI), Vir-like m6A methyltransferase associated (VIRMA, also known as KIAA1429), and zinc finger CCCH-type containing 13 (ZC3H13).^{2, 28, 29} METTL3 serves as the catalytic core of the m6A MTC, using S-adenosyl methionine (SAM) as the methyl donor. METTL14 provides structural support for METTL3 and forms a stable heterodimer core complex, which is crucial for substrate recognition.³⁰ The METTL3–METTL14 dimer mediates the deposition of m6A on mammalian RNA, functioning as both oncogenes and tumor suppressors in different tumors or even within the same tumor. Cytoplasmic METTL3 directly promotes the translation of several oncogenes, including epidermal growth factor receptor (EGFR) and Hippo pathway effector TAZ.³¹ In AML, high METTL3 expression plays a critical role in AML cell proliferation, maintaining the undifferentiated phenotype of AML cells, and is essential for AML development in mouse xenograft models.³² At the molecular level, METTL3 promotes translation of oncogenic target genes, such as modifying the mRNA encoding the oncogenic transcription factor MYC.³³ In immunodeficient mice, downregulation of METTL3 leads to cell cycle arrest, leukemia cell differentiation, and an inability to establish a leukemia model. Although METTL14 is not significantly expressed in leukemia, it also acts as an oncogene in AML, functioning in a MYC-dependent manner.^{34, 35} In liver cancer, METTL3 overexpression drives tumor growth by promoting the degradation of SOCS2 mRNA and upregulating SNAIL translation, leading to epithelial–mesenchymal transition (EMT) in cancer cells.³⁶ In contrast, low expression of METTL14 has been shown to have the opposite effect, reducing the metastatic potential of liver cancer cells by interacting with the microprocessor complex DGCR8 and promoting miRNA maturation ZC3H13.³⁷ METTL3 oncogenic role has also been reported in lung cancer, where it is overexpressed in human lung cancer tissues compared with normal tissues, and METTL14 protein and transcript abundance are also increased.^{38, 39}

WTAP, lacking methylation activity, acts as a regulatory subunit of the complex, recruiting the METTL3–METTL14 complex to form the catalytic core targeting RNA. Knockout and overexpression of METTL3 both lead to upregulation of WTAP protein expression, indicating that METTL3 levels play a critical role in WTAP protein homeostasis.⁴⁰ However, merely upregulating WTAP without functional METTL3 is insufficient to promote cell proliferation, suggesting that WTAP oncogenic function is closely related to the functional m6A methylation complex. WTAP influences MAPK, AKT, Wnt, and nuclear factor kappa-B (NF-κB) signaling pathways, promoting tumor progression by regulating downstream targets such as EGR3, HK2, ETS1, and CAV-1.⁴¹ RBM15, a member of the SPEN family located at chromosome 1p13.3, encodes homologs of the RNA-binding protein RBM15/RBM15B.⁴² RBM15/RBM15B, interacting partners of WTAP, recruit the complex to specific RNA sites by binding to U-rich RNA consensus sequences, mediating m6A formation in XIST and cellular mRNA.⁴² Knockdown of RBM15/15B significantly reduces overall m6A levels, indicating their crucial role in the MTC. In tumor progression, the role of RBM15 in the MTC has been reported only in leukemia, liver cancer, and laryngeal cancer.^{43, 44} KIAA1429 (VIRMA) is a key subunit of the MTC, functioning as a scaffold by recruiting the MTC to the 3′ UTR and near the stop codon. KIAA1429 selectively mediates methylation and promotes liver cancer progression and metastasis by inducing m6A methylation on the 3′ UTR of GATA3 precursor mRNA, leading to its degradation.⁴⁵ Additionally, KIAA1429 is significantly upregulated in liver cancer tissues and inhibits ID2 expression by increasing m6A levels in ID2 mRNA, thereby promoting liver cancer cell migration and invasion.⁴⁶ Recent studies have shown that ZC3H13, a transcription factor with a conserved zinc finger structure, plays a critical role in RNA m6A methylation by mediating the nuclear localization of the ZC3H13-WTAP complex.⁴⁷ This factor is highly expressed in various cancers and is associated with prognosis. ZC3H13 mainly promotes the binding of the MTC to RNA and has been shown to have tumor suppressor effects, inhibiting colorectal and breast cancer progression and metastasis through regulation of Ras–ERK and Wnt signaling pathways.^{48, 49} Furthermore, METTL16 has been identified as the methyltransferase for U6 small nuclear RNA (snRNA), regulating SAM homeostasis by catalyzing substrates with the “UACAGAGAA” sequence. Under conditions of SAM deficiency, METTL16 induces splicing of a retained intron, promoting MAT2A expression and increasing SAM levels. However, the role of METTL16 in cancer requires further investigation.⁵⁰ In summary, the m6A modification introduced by the m6A MTC plays a crucial role in regulating RNA metabolism, including RNA splicing, stability, transport, and translation efficiency. Dysregulation of m6A methylation has been closely associated with various diseases.

2.1.2 m6A demethylases

In 2011, Wei et al.⁵¹ first demonstrated in vivo that the fat mass and obesity-associated protein (FTO) can reverse m6A, making FTO the first identified m6A demethylase. In 2017, FTO's role in tumor progression was first reported. Studies have shown that FTO reduces m6A levels on ASB2 and RARA mRNA transcripts, regulating their expression and promoting AML progression.⁵² Additionally, FTO promotes tumor progression in liver, lung, breast, cervical, and colorectal cancers (CRCs) while acting as a tumor suppressor in renal, pancreatic, thyroid, and cholangiocarcinomas.⁴³ Following FTO, ALKBH5 was identified as the second m6A demethylase, involved in the biological progression of various cancers and exhibiting both oncogenic and tumor-suppressive functions.⁵³ In breast cancer, ALKBH5 expression is induced by hypoxia-inducible factor 1α (HIF1α) and HIF1β, and its overexpression under hypoxic conditions reduces NANOG mRNA methylation levels, increasing the number of breast cancer stem cells.⁵³ RBM33, a recently identified RNA-binding protein associated with substrate selectivity for ALKBH5, can act as an m6A recognition protein, binding to m6A-modified RNA substrates and recruiting ALKBH5 for demethylation. Moreover, SUMOylation significantly inhibits the m6A demethylase activity of ALKBH5, and the SUMOylation of ALKBH5 can be dynamically and reversibly regulated by the SUMO E3 ligase PIAS4 and the deSUMOylase SENP1.⁵⁴ Additionally, a concurrent study revealed that the acetylation modification on lysine 235 (K235) of ALKBH5, along with its regulatory subunit PSPC1, jointly determines the m6A demethylase activity and oncogenic function of ALKBH5. The K235 acetylation of ALKBH5 enhances its binding recognition to substrate RNA m6A, thus augmenting ALKBH5 removal of m6A modifications on RNA.⁵⁵

In summary, the dynamic regulation of m6A modifications by m6A demethylases plays a balancing role in RNA metabolism. Abnormal expression or dysfunction of m6A demethylases has been shown to be associated with various diseases, including cancer, neurological disorders, and metabolic diseases, highlighting their potential as therapeutic targets in pathological processes.

2.1.3 m6A reader proteins

The dynamic and reversible regulation of m6A modifications is crucially influenced by the interplay between m6A writers and erasers. Various types of m6A methyl readers enhance the diversity of m6A's biological functions. m6A readers include the YTH domain family proteins and the insulin-like growth factor 2 mRNA-binding protein (IGF2BP) family.^{56, 57} The YTH domain contains approximately 160 amino acids and can bind to single-stranded RNA. It is highly conserved across yeast, plants, and animals. In human cells, the YTH domain family includes five members: YTHDC1-2 and YTHDF1-3. YTHDC1, a nuclear m6A reader, has distinct biological functions compared with cytoplasmic readers. Studies have shown that YTHDC1 can interact with splicing factors SRSF3 and SRSF10,⁵⁸ and then bind to the m6A sites on mRNA, regulating selective splicing of nuclear mRNA. Additionally, the mRNA export factor NXF1 closely associates with SRSF3, promoting the export of m6A-modified mRNA from the nucleus.⁵⁹ Furthermore, research indicates that YTHDC2 is present in the cytoplasm and not only enhances the translation of m6A-modified mRNA but also promotes its degradation.⁶⁰ Another YTHDF family members, such as YTHDF1, YTHDF2, and YTHDF3, are located in the cytoplasm and exhibit distinct functions. YTHDF1 promotes mRNA translation, YTHDF2 facilitates mRNA degradation, and YTHDF3 assists in both translation and degradation, though the mechanisms underlying these different functions remain unclear.⁶¹ Studies have shown that YTHDF1 and YTHDF3 exhibit oncogenic roles in cancers.⁶¹ Analysis of The Cancer Genome Atlas database revealed that YTHDF1 expression is significantly upregulated in hepatocellular carcinoma (HCC) and positively correlates with pathological stage. Kaplan–Meier analysis indicated that lower levels of YTHDF1 are associated with better survival rates in HCC patients. GO and KEGG pathway analyses of YTHDF1 coexpressed genes suggest that YTHDF1 plays a crucial role in regulating the cell cycle and metabolism of HCC cells.⁶² In lung cancer, YTHDF2 is upregulated and promotes the growth of cancer cells by binding to the m6A-modified sites in the 3′-UTR of G6PD, enhancing its translation, and subsequently increasing the pentose phosphate pathway flux, thus driving tumor growth.^{61, 63} In CRC, long ncRNA (lncRNA)–GAS5 interacts directly with the WW domain of Yes-associated protein (YAP), promoting its phosphorylation and ubiquitin-mediated degradation, thereby weakening YAP-mediated transcription of YTHDF3. Research shows that YTHDF3 selectively and reversibly binds to m6A-methylated GAS5, triggering its degradation and forming a GAS5–YAP–YTHDF3 negative feedback loop, thereby promoting CRC progression.^{64, 65} Unlike the YTH domain family proteins, IGF2BP family proteins specifically recognize m6A-modified RNA via their KH domains, stabilizing mRNA rather than promoting its degradation.⁵⁶ The IGF2BP family includes IGF2BP1, IGF2BP2, and IGF2BP3, which are highly expressed in various cancer types and involved in numerous molecular mechanisms. IGF2BP1 and IGF2BP3 are oncofetal proteins produced by tumors and fetal tissues but downregulated in adult tissues.⁵⁶ Recent studies have found that IGF2BP1 binds to the m6A site in the 3′-UTR of SOX2 mRNA, inhibiting its degradation and thereby promoting the proliferation and metastasis of endometrial cancer cells.⁶⁶ In addition to the aforementioned factors, there is the HNRNPs family, including HNRNPC/G and HNRNPA2/B1. HNRNPA2/B1, as a nuclear m6A reader protein, directly recognizes the RGm6AC motif and regulates RNA selective splicing in a METTL3-mediated manner.⁶⁷ Furthermore, HNRNPA2/B1 plays a crucial role in promoting the processing of precursor miRNAs. Knocking down HNRNPA2/B1 leads to abnormal alternative splicing in cells, ultimately reducing the production of mature miRNAs.⁶⁸ In summary, m6A reader proteins participate in various aspects of RNA metabolism by recognizing and binding to m6A modifications, including posttranscriptional regulation and cellular responses. Dysregulation of m6A reader proteins is closely associated with the onset and progression of various diseases, indicating their significant role in disease mechanisms and potential as therapeutic targets.

3.3.1 Transcription factors

Transcription factors are proteins that bind to specific DNA sequences to regulate gene expression.¹¹³ Hematopoiesis is an extraordinarily complex process, with many transcription factors playing crucial regulatory roles in normal hematopoiesis. These transcription factors are expressed in a cell-type and developmental-stage-specific manner, controlling the differentiation of HSPCs and the fate determination of hematopoietic cells.¹¹⁴ Key transcription factors such as c-MYC, C/EBPa, and PU.1 are essential for hematopoiesis.¹¹⁵ c-MYC is a critical transcription factor regulating hematopoiesis; its deficiency severely impairs HSC function. C/EBPa is a major regulator of normal hematopoiesis; its knockout in mice leads to ineffective hematopoiesis in progenitor cells, resulting in reduced peripheral blood cells.¹¹⁶ PU.1 is involved in hematopoietic regulation, being essential for normal bone marrow and lymphoid development. Variations in PU.1 expression levels are crucial for the differentiation of hematopoietic lineages and the regulation of cell fate.¹¹⁷

3.3.2 DNA methylation modifications

Epigenetics refers to stable, heritable changes in gene expression that do not alter the DNA sequence.¹¹⁸ Key areas of epigenetics research include DNA methylation, histone modifications, and RNA methylation. DNA methylation and histone modifications primarily regulate gene expression at the transcriptional level, whereas RNA methylation regulates gene expression posttranscriptionally.¹¹⁹ Epigenetic modifications dynamically and reversibly regulate HSC self-renewal, multipotent differentiation, and lineage fate determination.¹²⁰ Comprehensive genome-wide analyses of epigenetic modifications (such as DNA methylation and histone modifications) during hematopoietic development have been conducted, revealing dynamic changes in epigenetic landscapes that regulate lineage-specific gene expression and, consequently, cell lineage fate. DNA methylation primarily occurs on cytosine–guanine dinucleotides (CpG sites) and is a reversible epigenetic modification.¹²¹ Approximately 80% of CpG sites in the human genome are methylated, with most unmethylated CpG sites concentrated in CpG islands located in promoter regions.¹²¹ DNA methyltransferases (DNMTs) catalyze the conversion of CpG to 5-methylcytosine (m5C) using S-adenosylmethionine (SAM) as a methyl donor. The known methyltransferases include DNMT3A, DNMT3B, and DNMT1. DNMT1 mediates maintenance methylation, whereas DNMT3A and DNMT3B are involved in de novo DNA methylation, modifying unmethylated DNA.¹²² The TET protein family (TET1, TET2, and TET3) functions as demethylases, catalyzing the conversion of 5-mC to hm5C, thus facilitating DNA demethylation. Research has shown that DNA methylation affects genetic stability, gene structure, and gene expression. Promoter CpG island methylation can regulate gene expression, with low promoter methylation being characteristic of gene activation, while high methylation is associated with gene silencing.¹²³ Studies have highlighted the significant role of DNA methylation in hematopoietic development. Abnormal activation or inactivation of DNMTs and demethylases can impair HSC development, indicating that DNA methylation levels are crucial in stem cell regulation.¹²⁴ DNMTs play important roles in regulating HSC self-renewal and differentiation. DNMT1-mediated methylation is essential for HSC self-renewal; its knockout leads to hypomethylation in HSCs, resulting in a pronounced bias toward myeloid differentiation and a marked impediment to lymphoid differentiation.¹²⁵ In contrast, DNMT3A knockout leads to increased HSC numbers, impaired hematopoietic reconstitution, and differentiation defects. DNMT3B knockout has minimal impact on HSC function, but combined knockout with DNMT3A causes more severe defects in HSC proliferation and differentiation. Loss of TET family members also promotes HSC self-renewal and leads to differentiation biases.¹²⁶ TET2, a major TET enzyme affecting hm5C levels in HSCs, is mutated or inactivated in certain cases, leading to increased proliferation of hematopoietic progenitor cells and enhanced self-renewal.¹²⁷ These findings indicate that HSC homeostasis and self-renewal require the regulation of DNA methylation. Normally, promoter CpG sites are unmethylated, with only a few dynamically changing. These dynamically changing DNA methylation sites often overlap with binding sites of lineage-specific transcription factors or enhancers. Methylation of promoters or enhancers can interfere with transcription factor binding and suppress downstream gene expression, thereby regulating HSC differentiation and self-renewal.

3.3.3 Histone acetylation

The fundamental structural unit of chromatin is the nucleosome, which consists of an octamer of histones wrapped around approximately 147 base pairs of DNA. This histone octamer comprises two copies. The fundamental structural unit of chromatin is the nucleosome, consisting of an octamer of histones wrapped around approximately 147 base pairs of DNA. This histone octamer is composed of two copies each of the core histones H2A, H2B, H3, and H4. The tails of histones extend out from the nucleosome, and posttranslational modifications, such as methylation and acetylation, occur on the tails of histones H3 and H4. The relative activities of histone modification writers and erasers determine the levels of histone modifications, which can be recognized by readers, including chromatin remodelers and transcription factors, to regulate gene transcription and expression.¹²⁸

Histone acetylation promotes chromatin relaxation, allowing transcription factors to bind specifically to DNA binding sites, thereby activating gene transcription and enhancing gene expression.¹²⁹ Conversely, histone deacetylation can lead to chromatin compaction and gene repression. The levels of histone acetylation also impact the biological functions of HSCs.¹³⁰ Abnormalities in histone acetyltransferases can impair the blood system. Knockout of histone deacetylases HDAC1 or HDAC2 affects hematopoiesis, while simultaneous knockout of HDAC1 and HDAC2 results in severe HSC defects, leading to anemia and reduced blood cell counts, indicating functional redundancy between HDAC1 and HDAC2 in hematopoiesis.¹³¹ SIRT1 (Sirtuin 1), a histone deacetylase, is also involved in the regulation of HSCs. The knockout of SIRT1 impairs the self-renewal and differentiation capabilities of HSCs.¹³² Similarly, the knockout of SIRT6 (Sirtuin 6) activates the Wnt signaling pathway, promoting HSC proliferation. However, HSCs lacking SIRT6 exhibit impaired self-renewal capacity in serial competitive transplantation experiments.¹³³

3.3.4 Histone methylation

Histone methylation occurs on lysine (K, Lys) and arginine (R, Arg) residues, with histone H3 being the most important, followed by H4. Both lysine and arginine can exist in multiple methylation states. Lysine residues on histones (such as H3K4, H3K9, H3K27, and H3K20) can be monomethylated (me), dimethylated (me2), or trimethylated (me3). Histone methylation influences chromatin conformation and the specific recruitment of downstream effector molecules, thereby regulating the transcription and expression of specific genes.¹³⁴ Histone methyltransferases are a class of enzymes containing the SET domain. The SET1/MLL family complex is the most common H3K4 methyltransferase in mammals. It consists of SET1A/B (KMT2F/G) and MLL1-4 (KMT2A-D) as catalytic subunits, along with WRAD (WDR5/RBBP5/ASH2L/DPY30) as the core subunits, which are essential for the methylation activity of the complex. ASH2L and RBBP5 are critical for maintaining the activity of MLL proteins. The SET1/MLL family complex plays a vital role in regulating HSC self-renewal and progenitor cell expansion. MLL fusion proteins downregulate RUNX1/CBFβ, thereby enhancing the self-renewal capacity of HSCs. Polycomb repressive complex 2 (PRC2) is a conserved multicomponent transcriptional repressive complex, with core components including EED, SUZ12, RBBP4, EZH1, and EZH2. EZH1 and EZH2 exhibit H3K27 methyltransferase activity, but require EED and SUZ12 for catalytic activity in vitro. While the partial loss of PRC2 caused by heterozygous deletions of EZH1, EZH2, or EED has minimal effects on HSCs, the simultaneous knockout of EZH1 and EZH2 or EED knockout, leading to complete PRC2 depletion, results in HSC exhaustion. This indicates that PRC2 regulates normal HSC function in a dose-dependent manner. EED knockout leads to HSC exhaustion by disrupting HSC self-renewal, differentiation, and apoptosis.¹³⁵ There are also reports that PRC2 can regulate NK cell differentiation and function through histone methyltransferase activity.

Histone demethylases, collectively known as the KDM family, all contain the JmjC demethylase domain, except for KDM1. LSD1 (also known as KDM1A) was the first demethylase identified and possesses demethylase activity for both H3K4 and H3K9, thereby repressing or activating its target genes. LSD1 plays a crucial role in hematopoiesis. Complete knockout of LSD1 results in pancytopenia and impairs the self-renewal and differentiation potential of HSCs. Conditional knockout of LSD1 restricts hematopoietic differentiation and promotes the expansion of HSPCs in the bone marrow by regulating methylation modifications on hematopoietic transcription factors Gfi1 and Gfi1b.¹³⁶ KDM4/JMJD2 is a specific demethylase for H3K9 and H3K36. Conditional knockout of Kdm4a/Kdm4b/Kdm4c in mice demonstrates that KDM4 activity is essential for maintaining HSCs in vivo, as the loss of KDM4 demethylase leads to the accumulation of H3K9me3 at transcription start sites, downregulating genes critical for hematopoiesis.¹³⁷ KDM6A (also known as UTX), a demethylase for H3K27, is part of the compass and SWI/SNF complexes, and plays a key role in regulating aging-related genes in the hematopoietic system. Histone methylation is dynamically regulated during hematopoietic development. Studies have shown that during HSC differentiation into erythroid progenitors, the levels of H3K4me1, H3K9me1, and H3K27me1 increase in enhancer regions of differentiation-related genes, suggesting that monomethylation of these histones is involved in the hematopoietic differentiation process.¹³⁸

Bivalent histone methylation refers to the presence of both activating H3K4me3 and repressive H3K27me3 methylation on the same gene, with LT-HSCs having the largest number of bivalent genes.¹³⁸ As HSCs differentiate, most genes are silenced postdifferentiation and lose H3K4me3, while a few are activated and lose H3K27me3. This suggests that bivalent marks can distinguish specific lineages and remain preserved in parallel differentiation pathways.¹³⁹ A study using iCHIP sequencing to investigate 16 populations during the hematopoietic process focused on four histone modifications: H3K4me1, H3K4me2, H3K4me3, and H3K27ac. Researchers identified 48,415 signal peaks in promoter regions, mapping their dynamic changes throughout differentiation and correlating them with transcriptional changes, illustrating that histone modifications determine the fate of cells.¹⁴⁰

3.3.5 Noncoding RNA

The Encyclopedia of DNA Elements project has shown that at least 80% of human genomic loci are transcribed into RNA, yet less than 2% of these transcribed regions encode proteins. Among these ncRNAs, those longer than 200 nucleotides are referred to as lncRNAs.¹⁴¹ In addition, ncRNAs include miRNAs and circRNAs.^{142, 143} miRNAs are small, naturally occurring ncRNAs, approximately 18–25 nucleotides in length, which regulate gene expression by binding to target RNAs, including mRNAs and other ncRNAs, through complementary base pairing.¹⁴² Hematopoiesis is a complex process regulated by multiple mechanisms, and miRNAs are involved in the directional differentiation of HSCs and the development of hematopoietic cells, including erythropoiesis. They have conserved sequences and a unique circular structure formed through back-splicing, enabling their stable expression and tissue specificity. With the advancement of transcriptome sequencing technologies, more and more ncRNAs are being discovered and confirmed to play important roles in regulating normal hematopoiesis and HSC self-renewal.

miRNAs are a class of ncRNAs, only 18–25 nucleotides in length, that regulate gene expression throughout biological evolution. They play a crucial role in the formation of nearly all mammalian tissues. O'Connell and colleagues reported that miR-125a and miR-125b are significantly upregulated in LT-HSCs, and overexpression of miR-125b significantly enhances bone marrow cell reconstitution capacity, indicating their importance in maintaining HSC self-renewal.¹⁴⁴ Additionally, miR-29a has been reported to regulate the Wnt signaling pathway, thereby maintaining HSC stemness.¹⁴⁵ Tenedini et al.¹⁴⁶ reported that miR-299 is specifically upregulated in megakaryocytes differentiated from human CD34+ hematopoietic progenitor cells in vitro, and overexpression of miR-299 promotes megakaryocyte and granulocyte differentiation. Ferrari et al.¹⁴⁶ found that overexpression of miR-221 and miR-222 inhibits CD34+ progenitor cell proliferation by downregulating c-Kit and promotes erythroid differentiation.

Research on the function of lncRNAs in hematopoietic differentiation indicates that they regulate HSC self-renewal and the differentiation and maturation of progenitor cells at multiple levels, including transcription, posttranscription, and translation. lncRNAs are key components of the gene expression regulatory network and represent important new members of this network. Hematopoiesis is the process by which HSCs gradually differentiate and mature into various functional blood cells. HSCs are a classic type of tissue stem cell with self-renewal and multipotent differentiation potential. During the classical process of hematopoietic differentiation, HSCs first differentiate into MPPs, which further differentiate into myeloid and lymphoid progenitors, ultimately forming mature cells of various lineages.¹⁴⁷ As regulatory factors, lncRNAs play essential roles in gene expression and cell fate determination throughout the stages of hematopoietic differentiation. Cabezas-Wallscheid et al.¹⁴⁸ constructed a transcriptional profile of HSCs and MPPs in mice and identified 682 lncRNAs specifically expressed in HSCs and MPPs, 79 of which were differentially expressed. However, this study did not specify the functions or mechanisms of these lncRNAs in HSC and MPP differentiation. Subsequently, Luo et al.¹⁴⁹ used flow sorting technology to isolate HSCs and conducted deep transcriptome sequencing, discovering 323 previously unreported lncRNAs. Comparison of their expression patterns during lineage differentiation revealed that 159 lncRNAs were highly expressed in HSCs, with some exhibiting HSC-specific expression. These lncRNA genes share epigenetic modification features similar to protein-coding genes, such as DNA methylation regulating their expression. Knockdown experiments on two lncRNAs specifically expressed in HSCs (lncHSC-1 and lncHSC-2) revealed that both play important roles in regulating HSC self-renewal and lineage differentiation. ChIRP-seq results of lncHSC-2 showed that it binds to the bHLH site of the hematopoietic transcription factor E2A and is enriched with other hematopoietic transcription factors (Erg, Fli1, Meis1, Pu.1, etc.) in the surrounding regions. The binding sites of lncHSC-2 are also enriched with H3K4me3/H3K27ac histone modifications, suggesting that lncHSC-2 may regulate target gene expression by recruiting key transcription factors through altering their epigenetic states.¹⁴⁹ These two studies, based on transcriptomics analysis, identified potential lncRNAs that regulate the quiescent and differentiated states of HSCs, providing an important resource for future research.

In addition to using transcriptomic data to screen for lncRNAs that potentially regulate HSC quiescence and differentiation, studies have also explored the roles of lncRNAs involved in other tissues and organs in hematopoietic differentiation. For example, the imprinted lncRNA gene H19, which regulates embryonic development, has also been shown to play a key role in maintaining HSC quiescence in mice. Conditional knockout of the imprinted control region upstream of maternal H19, also known as the H19 differentially methylated region (H19-DMR), activates the Igf2-Igf1r pathway, releases FoxO3 (Forkhead box O3)-mediated cell cycle arrest, and forces quiescent HSCs into the cell division cycle, leading to HSC exhaustion. Under physiological conditions, this pathway is suppressed by miR-675, which is encoded by an exon of the H19 gene.¹⁵⁰ The metastasis-associated lung adenocarcinoma transcription 1 (Malat1) is a highly conserved lncRNA across species. Malat1 is highly expressed in early HSPCs (Lin-Rhodamine^low Hoechst^low) in the mouse bone marrow and is downregulated in more differentiated progenitor cells (Lin-Rhodamine^bright Hoechst^low), suggesting that Malat1 may maintain the undifferentiated state of HSPCs. During the differentiation of erythroid myeloid lymphoid (EML) cells induced by all-trans retinoic acid (ATRA), Malat1 expression is downregulated by P53, which binds to its promoter, thereby inhibiting cell proliferation and indirectly promoting EML cell differentiation.¹⁵¹ However, the role of Malat1 in HSPCs and the mechanism by which it regulates their proliferation remains unclear. Overall, ncRNAs play a crucial role in regulating HSCs, contributing to their self-renewal, differentiation, and functional maintenance. Abnormal expression of ncRNAs has been associated with various hematological disorders, including leukemia and lymphoma, underscoring their importance in the regulation of HSC function and providing new insights for the diagnosis and treatment of related diseases.

3.3.6 Mitochondrial regulation

HSCs maintain blood system homeostasis by regulating cellular metabolism, with mitochondria being a major energy source for cellular metabolism.¹⁵² This highlights the crucial role of mitochondria in blood system homeostasis. HSCs in a quiescent state maintain a low metabolic rate, relying primarily on glycolysis for energy.¹⁵³ Quiescent HSCs exhibit low mitochondrial metabolic activity, low mitochondrial membrane potential, low oxidative phosphorylation (OXPHOS), high mitochondrial autophagy, and low mitochondrial reactive oxygen species (ROS) levels. Upon differentiation, HSCs exit the quiescent state and enter the cell cycle, shifting their energy source from glycolysis to mitochondrial OXPHOS, accompanied by mitochondrial activation.¹⁵⁴ This results in increased mitochondrial quality, membrane potential, and ROS levels, with reduced mitochondrial autophagy. There is a significant metabolic difference between quiescent and active HSCs.¹⁵⁵ Selective degradation of mitochondria through mitochondrial autophagy is critical for maintaining blood system homeostasis; under normal conditions, mitochondrial autophagy removes damaged mitochondria and reduces ROS production. HSCs that cannot undergo autophagy accumulate damaged mitochondria and increased ROS levels, leading to continuous activation of HSCs. When ROS levels exceed cellular antioxidant defenses, HSCs undergo aging or cell death. The deficiency of Atg7 (autophagy-related gene 7) leads to mitochondrial metabolic dysregulation, impairing HSC quiescence and resulting in HSC exhaustion.¹⁵⁶ Cellular autophagy maintains normal mitochondrial status to regulate the metabolic rate of HSCs and sustain their quiescence.¹⁵⁶ The interaction between mitochondria and lysosomes is also crucial for maintaining HSC homeostasis. Quiescent HSCs contain numerous active lysosomes, and lysosomal activity is essential for regulating the quiescent state of HSCs.¹⁵⁷

4.2.1 Self-renewal

Self-renewal is a fundamental characteristic of all stem cells and is one of the most notable features of LSCs.¹⁷⁰ The biological properties of LSCs share many similarities with HSCs. HSCs are renowned for their proliferation and self-renewal capabilities, responsible for maintaining the hematopoietic system and residing in the bone marrow.¹⁷¹ LSCs have a stronger self-renewal capacity compared with HSCs and exhibit enhanced cell expansion ability, contributing to the high relapse rate observed after acute AML chemotherapy.¹⁷² Various signaling pathways regulate the self-renewal of LSCs, including the Wnt/β-catenin and Notch signaling pathways.¹⁷³

4.2.2 Antiapoptotic properties

In normal cells, the complex interactions between proapoptotic and antiapoptotic proteins are tightly regulated, with their relative balance determining cellular responses to stress.¹⁷⁴ An important reason for the high relapse rate in leukemia is the upregulation of antiapoptotic proteins (such as BCL-2 and BCL-xL), which prevents apoptosis and thereby promotes the survival of LSCs during treatment.¹⁷⁵ NF-κB is a nuclear transcription factor that regulates tumor growth by inducing the expression of its target genes. The BCR–ABL fusion gene in chronic myelogenous leukemia (CML) acts as an antiapoptotic gene, activating various signaling pathways such as PI3K/AKT, JAK/STAT, RAS, and NF-κB.¹⁷⁶ The expression of the antiapoptotic protein myeloid cell leukemia 1 has also been linked to NF-κB activation.¹⁷⁷ These findings indicate that the expression of antiapoptotic genes in leukemia is related to NF-κB. Thus, theoretically, targeting NF-κB activity to reduce antiapoptotic gene expression and induce apoptosis in LSCs could achieve targeted eradication of LSCs.¹⁷⁸

4.2.3 Drug resistance

LSCs with the CD34+CD38− phenotype exhibit resistance to chemotherapeutic agents, which may contribute to minimal residual disease.¹⁷⁹ Although the rapid proliferation of differentiated leukemia cells results in adverse clinical outcomes, the LSCs themselves are largely quiescent.¹⁸⁰ Quiescent LSCs remain in the G0 phase of the cell cycle, making it challenging for conventional chemotherapy, which targets rapidly dividing cancer cells, to eradicate them. The drug resistance of LSCs results in the inability of current treatments to completely eliminate them, leading to a high relapse rate after AML therapy.¹⁸¹ Inducing LSCs to exit the quiescent state and enter the cell growth cycle could enhance their sensitivity to chemotherapy, representing an effective therapeutic strategy for eradicating LSCs.

4.2.4 Microenvironment of LSCs

The bone marrow microenvironment supports normal hematopoiesis through the secretion of various growth factors and physical interactions with HSCs and progenitor cells.¹⁸² Given that LSCs share stem cell-like characteristics with normal HSCs, they are also likely regulated within the bone marrow microenvironment.¹⁸³ Notably, during the formation of AML, the bone marrow microenvironment contributes to leukemia development through interactions with LSCs. LSCs, while residing in the bone marrow microenvironment, can reciprocally alter this environment, making it more conducive to their survival and less favorable for HSCs.¹⁸⁴ The bone marrow microenvironment thus acts as a sanctuary for LSCs, aiding their evasion of chemotherapy and immune escape. If drugs can be designed to accurately identify and eliminate LSCs within the bone marrow microenvironment or modify the environment to be unfavorable for LSC survival without affecting normal hematopoiesis, the treatment of AML and prevention of disease relapse could become more manageable.¹⁸⁵

4.2.5 Immunophenotypes associated with LSCs

The gene expression profile of LSCs, which includes genes specifically expressed by LSCs, is frequently used for risk stratification and prognostic assessment in AML.¹⁸⁶ During various stages of hematopoietic differentiation in AML, mutations in relevant genes can result in different immunophenotypes for LSCs, thus allowing for the diagnosis and differentiation of leukemia with unclear lineage by identifying these immunophenotypes.¹⁸⁷ LSCs exhibit stem cell-like properties and have similar immunophenotypes to normal HSCs, such as CD34+CD38−. However, LSCs also show higher expression of additional markers, such as CD123, CD96, CD47, CD25, G-protein coupled receptor 56 (GPR56), C-type lectin-like receptor-1 (CLL-1), interleukin 1 receptor accessory protein (IL1RAP), N-cadherin, and recombinant TEK tyrosine kinase (Tie2).¹⁸⁸ CD123, the alpha subunit of the interleukin-3 (IL-3) receptor, is highly expressed on LSCs but is not found on normal HSCs, making it a unique marker for leukemia. CD123 can distinguish between normal HSCs, which show almost no CD123 expression, and CD123-positive LSCs within the CD34+CD38− population. High levels of CD123+ LSCs are negatively correlated with chemotherapy outcomes and prognosis in AML patients, with high CD123 expression at diagnosis serving as an independent prognostic indicator.¹⁸⁹ Therefore, CD123 may be an important marker for assessing treatment response and prognosis in clinical leukemia management.

CD96, a member of the immunoglobulin superfamily and a novel immune checkpoint receptor, plays a crucial role in antitumor immune responses. CD96 is expressed on LSCs but at lower levels on normal CD34+CD38− cells. In some patients with poor prognosis, CD96 expression can be as high as 90%. Studies on surface molecules of LSCs in children with acute leukemia have shown that CD96 is predominantly expressed in AML, with lower expression in acute lymphoblastic leukemia (ALL).¹⁹⁰ CD96+ patients tend to have a lower complete remission rate after induction chemotherapy and higher infection and relapse rates compared with CD96− patients. This suggests that CD96+ leukemia patients are more prone to relapse postchemotherapy, indicating CD96 as a potential new prognostic marker.¹⁹¹ CD47, also known as integrin-associated protein, is a transmembrane protein widely expressed on various cell surfaces and is highly expressed on LSCs. Its primary function is to provide an antiphagocytic signal.¹⁹² High CD47 expression on LSCs enables them to evade phagocytosis and elimination.¹⁹³ Research by Irving L Weissman et al. in 2009 found that CD47 is overexpressed on LSCs derived from AML patients, and CD47 inhibits phagocytosis by interacting with inhibitory receptors on phagocytes, facilitating AML progression. Increased CD47 expression correlates with reduced overall survival in adult AML patients and is an independent factor for poor prognosis.¹⁹⁴ Additionally, studies using lentiviral vectors to deliver CD47-siRNA into AML-derived LSCs demonstrated effective inhibition of LSC proliferation and promotion of apoptosis, highlighting CD47 as a valuable target for leukemia treatment and prognostic prediction.¹⁹⁵ GPR56 is an adhesion molecule that interacts with the bone marrow niche through binding to type III collagen.¹⁹⁶ This suggests that GPR56 may play a role in the interaction between LSCs and the bone marrow microenvironment. A multicenter Phase III trial found that GPR56 is highly expressed in AML patients with NPM1 and FLT3 mutations, and its high expression is associated with poor prognosis in AML patients.¹⁹⁷

N-cadherin, a classic cadherin in the cadherin superfamily, and Tie2, a receptor tyrosine kinase primarily expressed on endothelial cells and significantly upregulated in tumor neovascularization, are also relevant.¹⁹⁸ A study by Zhi et al.¹⁹⁹ analyzed the proportion of CD34+CD38−CD123+ LSCs coexpressing N-cadherin or Tie2 before and after chemotherapy, finding that chemotherapy enriches the population of N-cadherin and Tie2 positive CD34+CD38−CD123+ LSCs. Therefore, N-cadherin and Tie2 may also serve as potential markers for identifying LSCs. CD25, the alpha chain of the IL-2 receptor, is primarily expressed on normal T cells and induces their proliferation and differentiation. It is also highly expressed on LSCs.²⁰⁰ CLL-1, a type II transmembrane glycoprotein, plays a key role in immune regulation as an inhibitory receptor and is expressed in over 90% of AML patients’ malignant cells but almost not in HSCs.²⁰¹ Similarly, IL1RAP, which shares similarities with CLL-1, is upregulated in the LSCs of most AML patients and is not expressed in normal HSCs.²⁰² Overall, the diverse immunophenotypes highly expressed on LSCs provide more avenues for basic research on LSCs and offer additional options for clinical targeted therapy in AML.

4.3.1 Wnt signaling pathway

The Wnt/β-catenin pathway is a highly conserved signal transduction cascade that plays crucial roles in both normal development and disease states in humans.²⁰⁵ Among the four types of Wnt pathways, the Wnt/β-catenin pathway is classical and primarily involved in gene transcription regulation and activation of nuclear target gene expression. This pathway is often found to be hyperactivated in various cancers and is particularly involved in the self-renewal and proliferation of LSCs.²⁰⁶ β-Catenin, a central effector and key molecule in Wnt signaling, is critical for LSC self-renewal, tumorigenesis, progression, relapse, and drug resistance. Theoretically, inhibiting this pathway to suppress β-catenin expression could potentially reduce or even eliminate the self-renewal capability of LSCs.²⁰⁷ Thus, β-catenin is considered a potential target for reducing LSC self-renewal and thereby eliminating LSCs.

4.3.2 Hedgehog signaling pathway

The Hedgehog signaling pathway plays a significant role in embryonic development, maintaining adult stem cells, and regulating cell proliferation and differentiation.²⁰⁸ In addition to its roles in normal embryogenesis and tissue homeostasis, abnormalities in the Hedgehog pathway also affect the survival of leukemia-associated cells. The Hedgehog pathway can be activated through two main routes, with a focus on the canonical pathway here.²⁰⁹ The canonical pathway involves three ligands—Indian Hedgehog, Desert Hedgehog, and Sonic Hedgehog—and two receptors—Smoothened (SMO) and Patched (PTCH). In the absence of ligands, PTCH inhibits SMO, whereas in the presence of ligands, ligand binding to PTCH relieves this inhibition, allowing SMO to activate glioma-associated oncogene homolog (GLI) through phosphorylation.²¹⁰ Activated GLI binds to target DNA in the nucleus, leading to the expression of specific genes involved in cell cycle induction, antiapoptosis, and cell differentiation. As a crucial node in the Hedgehog pathway, inhibition or reduction of SMO can decrease the number of LSCs and impair their self-renewal capability.²¹¹ Therefore, targeting SMO to reduce its expression could potentially improve the cure rate for leukemia.

4.3.3 NF-κB signaling pathway

The NF-κB signaling pathway is persistently activated in various leukemia cells, especially in LSCs.²¹² NF-κB is a vital transcription factor involved in cell survival, proliferation, and differentiation. It is expressed at lower levels in normal HSCs but is significantly overexpressed in LSCs.²¹³ The overactivation of NF-κB in LSCs induces the expression of antiapoptotic genes, blocking LSC apoptosis and contributing to the persistence of AML, resulting in decreased treatment efficacy and increased relapse rates. Inhibiting the abnormal expression of NF-κB during AML progression and downregulating antiapoptotic pathways to restore normal apoptosis could potentially impede the proliferation and differentiation of LSCs, leading to a reduction in their number.²¹³ In summary, these distinct signaling pathways regulate various downstream target genes and play crucial roles in the self-renewal of LSCs.

6.1.1 The role of m6A MTCs in AML

METTL3 plays a critical role in both normal and malignant hematopoiesis. In normal hematopoiesis, METTL3 maintains the homeostasis of the bone marrow microenvironment for HSCs and suppresses the myeloid differentiation of HSCs.^{227, 247} Silencing METTL3 or reducing m6A methylation levels can induce differentiation and apoptosis of bone marrow mesenchymal stem cells, leading to hematopoietic failure.²²⁷ In AML, METTL3 is overexpressed compared with normal HSCs. Through its m6A methyltransferase function, METTL3 increases the m6A methylation levels of MYC, BCL-2, and PTEN, thereby upregulating their expression. This action inhibits normal HSC differentiation and induces hematopoietic arrest at an earlier stage.^{248, 249} Sang et al.²⁵⁰ found that METTL3 knockout induces AML cell differentiation and promotes apoptosis. Furthermore, METTL3 expression is upregulated in AML patients at diagnosis and during relapse, but downregulated upon achieving complete remission, suggesting a potential role of METTL3 in AML drug resistance.²⁵⁰ Another study by Li et al.²⁴⁹ demonstrated through RNA sequencing of complete remission and relapsed AML patients that METTL3 plays a significant role in AML chemoresistance. Cell experiments confirmed that METTL3 enhances ITGA4 mRNA stability in an m6A-dependent manner, upregulating ITGA4 protein expression, which increases AML cell adhesion and migration, promoting the homing and engraftment of AML cells in the bone marrow microenvironment, and ultimately leading to chemoresistance.²⁴⁹ METTL14 is highly expressed in AML cells and regulates the stability of MYB and MYC mRNA through m6A methylation. This regulation enhances the self-renewal capacity of LSCs/leukemia-initiating cells (LSCs/LICs) and contributes to the oncogenic potential of AML.³⁵ Lei et al.²⁵¹ discovered that METTL14 oncogenic role in AML is positively regulated by lncRNA. Through in vivo and in vitro experiments, they found that lncRNA UCA1 binds to METTL14, promoting its stability and upregulating CXCR4 and CYP1B1 protein levels, which enhances AML cell proliferation and progression.²⁵¹ This study links METTL14 with lncRNA, suggesting a new molecular mechanism for AML development and providing new insights into AML cell proliferation and progression. Another study connected METTL14 with miRNA and validated that miR-1306-5p downregulates METTL14 expression, thereby decreasing MYB and MYC m6A methylation levels, reducing the stability of their mRNA transcripts, and promoting AML cell apoptosis.²⁵² This negative regulatory role of miRNA in AML offers new strategies for future AML treatments. These studies indicate that, beyond regulating mRNA, m6A methylation also involves the regulation of lncRNA and miRNA. METTL16 has been identified as essential for the survival of AML cells. It is overexpressed in human AML as well as LSCs and LICs. Knockout of METTL16 significantly inhibits AML initiation/development and maintenance and markedly impairs LSC/LIC self-renewal, with moderate effects on normal hematopoiesis in mice. Mechanistic studies show that METTL16 exerts its oncogenic effects by promoting the expression of branched-chain amino acid (BCAA) transaminase 1 (BCAT1) and BCAT2 in an m6A-dependent manner, thereby reprogramming BCAA metabolism in AML.²²⁶ In addition to its role in regulating ncRNAs, another study by Shao et al. revealed the role of the m6A methyltransferase WTAP in histone modification regulation.²⁵³ This study confirmed that the transcriptional activator HIF1α binds to WTAP to positively regulate the m6A methylation level of KDM4B mRNA transcripts, upregulating KDM4B protein expression. This upregulation inhibits H3K9me3 activity, thereby promoting AML cell proliferation.²⁵⁴ This finding offers new insights into AML development and progression from the perspective of WTAP interaction with histone methylation modifications (Figure 6). Overall, the above studies indicate that m6A methyltransferase play a crucial role in the progression of AML.

6.1.2 The role of m6A demethylases in AML

Studies have shown that FTO is overexpressed in AML with FLT3–ITD and NPM1 mutations. This overexpression downregulates m6A methylation levels of ASB2 and RARA mRNA, leading to decreased ASB2 and RARA protein expression. Consequently, this diminishes the sensitivity of AML cells to ATRA, thus promoting cell proliferation and AML progression.²⁵⁵ Therefore, targeting FTO with inhibitors in combination with ATRA holds substantial potential for AML treatment. Previous research has reported that FTO inhibitors enhance m6A methylation of CEBPA mRNA, reducing CEBPA protein levels and ultimately inhibiting AML cell proliferation.²⁵⁶ Thus, targeting the FTO/m6A/CEBPA pathway could also be a viable strategy for AML therapy. In AML subtypes with NPM1 mutations, which exhibit high autophagic activity, the specific mechanisms maintaining this high level of autophagy remain unclear. Huang et al.²⁵⁷ confirmed that FTO plays a critical role in maintaining high autophagic activity in NPM1+ AML. FTO lowers the m6A methylation level of TP53INP2 mRNA, thereby upregulating TP53INP2 protein expression. This enhancement promotes interactions with autophagy-related proteins such as microtubule-associated protein 1 light chain 3 (LC3) and ATG7, thus increasing autophagic activity in NPM1+ AML cells and inducing cell survival and proliferation.²⁵⁷ Hence, targeting and inhibiting FTO and TP53INP2 could improve therapeutic precision and may be particularly effective in NPM1 AML with high FTO expression, potentially translating into clinical applications. These studies suggest that FTO role in AML involves various cellular regulatory pathways, providing a theoretical basis for the development of FTO inhibitors. Wang et al.²⁵⁸ found that during the leukemogenesis of AML, ALKBH5 expression is regulated by chromatin state changes and is essential for maintaining LSC function. Mechanistic studies revealed that lysine-specific histone demethylase 4C (KDM4C) increases ALKBH5 gene chromatin accessibility by downregulating H3K9me3 expression, promoting MYB and Pol II recruitment, thereby sustaining high ALKBH5 expression. AXL is one of the targets of ALKBH5; ALKBH5 regulates AXL mRNA stability in an m6A-dependent manner through YTHDF2, thus promoting AML cell proliferation and maintaining AML LSC function. This provides a theoretical foundation for targeting ALKBH5 to eliminate LSCs and overcome AML resistance.²⁵⁸ In another study, Shen et al.²⁵⁹ identified TACC3 as another target of ALKBH5. ALKBH5 downregulates P21 and upregulates MYC protein levels by modulating TACC3 mRNA stability, thereby promoting AML cell proliferation and LSC self-renewal, which is associated with poor outcomes in AML patients. Therefore, targeting ALKBH5 and TACC3 holds promise as a therapeutic approach for AML (Figure 6). Overall, the above studies indicate that m6A demethylases play a crucial role in the progression of AML.

6.1.3 The role of m6A reader proteins in AML

Recent research has shown that YTHDF1 is overexpressed in human AML samples and significantly upregulated in LSCs, with no effect on normal mouse hematopoiesis. Knocking down YTHDF1 inhibits tumorigenesis of both human and mouse AML cells in vivo and in vitro. Mechanistically, YTHDF1 promotes the translation of cyclin E2, thereby advancing AML progression. Interestingly, research has found that decitabine can inhibit YTHDF1, thereby suppressing YTHDF1-regulated translation and leukemia development.²⁶⁰ Paris et al.^{45, 238} discovered that YTHDF2 inhibits TNF-α expression through an m6A-mediated mRNA decay mechanism, thereby suppressing apoptosis and promoting LSC self-renewal as well as AML initiation and progression. Additionally, another study demonstrated that YTHDF2 is upregulated in t(8;21) AML and specifically targets AML1/ETO-HIF1α.²⁶¹ Inhibition of AML1/ETO or HIF1α downregulates YTHDF2 expression, enhancing methylation levels of TNFRSF1b mRNA and upregulating its expression, which promotes AML cell apoptosis and suppresses LSC proliferation.²⁶¹ YTHDC1 has been reported to advance AML through different mechanisms. Analysis of CRISPR Screen data in AML cell lines revealed that the nuclear m6A reader protein YTHDC1 is crucial for AML cell survival. YTHDC1 is significantly overexpressed in various AML subtypes. Knockdown of YTHDC1 in AML cell lines or via CRISPR knockout slows AML cell proliferation, enhances myeloid differentiation, and increases apoptosis. In AML cell line and patient-derived xenograft mouse models, YTHDC1 knockdown significantly inhibits AML development, further proving its critical role in AML progression.¹¹⁷ Mechanistic studies show that YTHDC1 forms compartmentalized structures in the nucleus, binding to m6A-modified RNA and forming liquid-like condensates known as nuclear YTHDC1–m6A condensates (nYACs). This process regulates downstream target gene MYC expression, and MYC overexpression can partially compensate for the phenotype induced by YTHDC1 knockdown.¹¹⁷ Unlike cytoplasmic m6A reader YTHDF2, which promotes m6A-modified mRNA degradation, YTHDC1 binds m6A–mRNA to form nYACs, stabilizing its target mRNA and thus enhancing gene expression. Additionally, our laboratory found that YTHDC1 is significantly overexpressed in various karyotypic AML types. Knockdown of YTHDC1 significantly inhibits AML cell growth, increases apoptosis and differentiation. Using Ythdc1 knockout mouse models, we found that Ythdc1 knockout significantly suppresses AML cell growth in mice and impairs LSC function, extending the survival of AML mice. Mechanistic studies reveal that YTHDC1 stabilizes MCM4 mRNA by recognizing m6A modifications on MCM4.²⁴¹ Re-expression of MCM4 can significantly restore the growth suppression and increased apoptosis and differentiation caused by YTHDC1 loss. Another recent study reported that YTHDC1 interacts with RBM15, recruiting the PRC2 to RBFOX2 binding sites, thereby achieving chromatin silencing and transcriptional repression.²⁶² The RBFOX2/m6A/RBM15/YTHDC1/PRC2 axis plays a critical role in myeloid leukemia. Downregulation of RBFOX2 significantly inhibits AML cell survival/proliferation and promotes myeloid differentiation. RBFOX2 is also essential for the self-renewal of LSCs/LICs and the maintenance of AML.²⁶²

Another m6A reader protein, IGF2BP2, is also involved in regulating LSCs and AML progression. IGF2BP2 enhances the stability of MYC, SLC1A5, and GPT2 mRNA in an m6A-dependent manner, promoting their translation and maintaining LSC/LIC function, thus accelerating AML progression.²⁶³ A recent interesting study found that IGF2BP2 is highly expressed in AML patient samples and is associated with poor prognosis. IGF2BP2 deficiency promotes apoptosis, damages LSC function, and inhibits AML development. Notably, the research team identified protein arginine methyltransferase PRMT6 as a potential key gene in leukemia initiation. They confirmed that IGF2BP2 regulates PRMT6 mRNA stability, making PRMT6 a downstream functional target of IGF2BP2.²⁶⁴ PRMT6 is highly expressed in AML samples and maintains LSC function; PRMT6 knockout inhibits AML development. The study further confirmed that the PRMT6 inhibitor EPZ020411 effectively suppresses LSC function.²⁶⁴ Research has shown that PRMT6 mainly catalyzes the formation of the repressive mark H3R2me2a. Upregulating PRMT6 downstream key target gene lipid transporter MFSD2A regulates the transport of fatty acids such as docosahexaenoic acid (DHA), maintaining LSC function.²⁶⁴ Additionally, Y box-binding protein 1 (YBX1) regulates AML at the posttranscriptional level through interaction with IGF2BP2. Feng et al.^{48, 115} confirmed that YBX1 interacts with IGF2BP2, enhancing MYC and BCL-2 mRNA stability and upregulating their translation in an m6A-dependent manner, thereby promoting AML cell survival. This provides a theoretical basis for targeting YBX1 in AML treatment. Currently, the fundamental cause of AML resistance and relapse is the inability of chemotherapeutic agents to effectively eliminate LSCs. Both YTHDF2 and IGF2BP2 maintain LSC function and survival through m6A methylation modifications. Therefore, YTHDF2 and IGF2BP2 could serve as valuable targets for clearing LSCs, offering new possibilities for overcoming resistance and curing AML.

IGF2BP3 has been reported to be specifically overexpressed in AML and is essential for AML cell survival. Knocking down IGF2BP3 significantly inhibits cell apoptosis, reduces proliferation, and weakens leukemia ability of AML cells in vitro and in vivo. Mechanistic studies reveal that IGF2BP3 interacts with RCC2 mRNA and stabilizes the expression of m6A-modified RNA.²⁶⁵ Furthermore, a recent studies have identified significant upregulation of the m6A reader proteins IGF2BP2 and IGF2BP3 in AML, regulated by super-enhancers. The BRD4 inhibitor JQ1 has been shown to effectively inhibit the expression of IGF2BP2 and IGF2BP3. As target genes of IGF2BP2 and IGF2BP3, DDX21 significantly promotes AML cell proliferation, suppresses apoptosis, and induces cell cycle arrest.²⁶⁶ Additionally, it reduces tumor burden in mouse models and extends survival time. Mechanistic studies reveal that DDX21 interacts with the transcription factor YBX1 to enhance ULK1 transcription, thereby promoting AML progression.²⁶⁶ Similarly, research has reported upregulation of IGF2BP3 in high-risk AML patients, with a positive correlation to SENP1 expression. IGF2BP3 binds to the SENP1 3′-UTR in an m6A-dependent manner, enhancing SENP1 expression and activating the AKT signaling pathway, which contributes to AML malignancy.²⁶⁷ IGF2BP3 is also known to directly bind EPOR mRNA, regulating its stability in an m6A-dependent manner to facilitate AML progression²⁶⁸ (Figure 6). Overall, the above studies indicate that m6A readers protein play a crucial role in the progression of AML.

6.1.4 m6A influence on chronic myeloid leukemia cells

Previous research has demonstrated that PTEN, through its phosphatase activity, inhibits the PI3K-AKT-mTOR pathway and acts as a tumor suppressor in various cancer cells.²⁶⁹ In CML, PTEN has been shown to suppress the self-renewal and survival of LSCs) thus exhibiting tumor-suppressive effects. Recent studies have found that METTL3 is significantly enriched in the long intergenic ncRNA 470 (LINC00470) transcripts.²⁷⁰ LINC00470 positively regulates the m6A modification level of PTEN mRNA by binding with METTL3, promoting PTEN mRNA degradation and downregulating PTEN protein expression. This leads to inhibition of autophagy in CML cells and contributes to chemotherapy resistance (Figure 7). Overall, the above studies indicate that m6A demethylases play a crucial role in the progression of chronic myeloid leukemia.

6.1.5 m6A influence on diffuse large B-cell lymphoma cells

Diffuse large B-cell lymphoma (DLBCL) is the most common subtype of lymphoid malignancies, with over a hundred classifications and high heterogeneity. Zhang et al.²⁷¹ discovered that WTAP levels are persistently upregulated in DLBCL tissues, promoting cell proliferation, inhibiting apoptosis, and affecting chemotherapy efficacy. This upregulation depends on the stability provided by heat shock protein Hsp90, suggesting that Hsp90 inhibitors should be considered in chemotherapy regimens for WTAP-high DLBCL.^{271, 272} Additionally, WTAP can form a complex with BCL6 (a transcriptional repressor blocking B-cell differentiation), though the role of this complex in transcriptional activity or other functions requires further investigation. Literature reports also indicate that WTAP, as a specific target of piRNA-30473 (a Piwi-interacting RNA), is involved in regulating DLBCL cell proliferation and the cell cycle, closely related to patient prognosis. The mechanism involves piRNA-30473 upregulating WTAP levels, which enhances the methylation and expression of the key target gene HK2 mRNA, thereby promoting DLBCL cell survival and proliferation²⁷³ (Figure 7). Overall, the above studies indicate that m6A demethylases play a crucial role in the progression of DLBCL.

6.1.6 m6A influence on multiple myeloma cells

Jiang et al. demonstrated through cell proliferation assays, in vitro knockouts, and subcutaneous tumorigenesis experiments that HNRNPA2B1 promotes multiple myeloma (MM) cell proliferation and inhibits apoptosis in both in vivo and in vitro settings. Silencing HNRNPA2B1 reduces ILF3 m6A modification and transcription. Knocking down ILF3 reverses HNRNPA2B1-induced cell proliferation, indicating that ILF3 is a crucial m6A/HNRNPA2B1 target in MM cells.²⁷⁴ Further analysis showed that AKT3 mRNA is enriched at ILF3 binding sites, with its expression positively regulated by HNRNPA2B1. This study highlights that HNRNPA2B1 mediates increased AKT3 expression by enhancing m6A-dependent ILF3 mRNA stability, thus playing a critical role in MM progression and correlating with poor prognosis²⁷⁴ (Figure 7). Overall, the above studies indicate that m6A demethylases play a crucial role in the progression of multiple myeloma.

6.1.7 m6A modification in myelodysplastic syndromes

Research indicates that RNA m6A epigenetic dysregulation exists in the bone marrow of myelodysplastic syndromes (MDS) patients, with elevated m6A modification levels closely associated with high-risk MDS. METTL14 is a key regulator causing m6A modification disorders in MDS RNA. Silencing METTL14 significantly reduces RNA m6A levels in MDS cells, markedly suppresses cell proliferation and colony formation, and prolongs mouse survival. Mechanistic studies reveal that METTL14 enhances SETBP1 mRNA m6A modification by forming a heterodimer with METTL3, increasing SETBP1 mRNA stability, and activating the PI3K-AKT signaling pathway, thereby promoting MDS development.²⁷⁵ ALKBH5 is reported to regulate SF3B1 5′-UTR m6A demethylation, affecting SF3B1 translation, DNA repair, and epigenetic regulator splicing, driving MDS progression.²⁷⁶ Studies also show that DDX41 promotes YTHDC1 recruitment to the R loop by enhancing METTL3 and YTHDC1 binding, maintaining genomic stability in MDS and promoting disease progression²⁷⁷ (Figure 7). Overall, the above studies indicate that m6A demethylases play a crucial role in the progression of MDS.

7.1.1 Targeting the Wnt/β-catenin signaling pathway

Indomethacin, a reversible cyclooxygenase inhibitor, blocks the activation of the Wnt pathway by inhibiting β-catenin expression. Studies by Wang et al.³⁷ have shown that indomethacin treatment can suppress LICs crucial for AML development and affect established LSCs.²⁸⁸ In addition to single-agent inhibitors, combination therapy has also shown promising results in targeting LSCs. Lin et al.²⁸⁹ assessed the effects of sulfadiazine (SFN) on Imatinib (IM)-resistant LSCs and found that combining SFN with IM effectively eradicates CD34+/CD38− LSCs in CML. The efficacy of SFN in targeting LSCs and reducing resistance primarily involves downregulation of β-catenin expression.²⁸⁹ At low doses, doxorubicin (DXR) acts as an inhibitor of Akt–β-catenin interactions.²⁹⁰ The synergistic effect of Wnt/β-catenin and PI3K/AKT pathways can promote tumorigenesis and resistance to therapy. Results by Perry et al.²⁹⁰ indicated that combined chemotherapy and low-dose DXR significantly improved survival in leukemia model mice, with low-dose DXR exhibiting stronger specificity for targeting LSCs, including those resistant to chemotherapy.

7.1.2 Targeting the Hedgehog signaling pathway

Glasdegib, a selective small-molecule inhibitor of SMO, has demonstrated efficacy against AML cells both in vitro and in vivo using human xenograft mouse models.²⁹¹ Its shorter half-life helps reduce toxicity to normal stem cells and enhance its ability to target LSCs, particularly when used in combination with other agents. The United States Food and Drug Administration has approved Glasdegib in combination with low-dose Ara-C for newly diagnosed elderly AML patients (≥75 years old) or those who cannot tolerate intensive induction therapy and have significant comorbidities.²⁹¹ Cyclopamine, a natural plant steroidal alkaloid and SMO inhibitor, has been shown to reduce the resistance of CD34+ leukemia cell lines to cytarabine and DXR.²⁹²

7.1.3 Targeting the NF-κB signaling pathway

Parthenolide (PTL) is an NF-κB kinase activation inhibitor that can inhibit LSC proliferation and induce G2 phase cell cycle arrest. PTL is the first effective natural drug targeting LSCs, meaning it preferentially kills leukemia stem/progenitor cells with minimal impact on noncancer cells.²⁹³ PTL treatment significantly reduces the engraftment potential of AML patient-derived LSCs, but its poor druggability limits clinical application. Ghantous et al.^{42, 294} identified a new compound, dimethylamino parthenolide, which, as an NF-κB inhibitor, exhibits similar activity to PTL and selectively eradicates LSCs. DMAMCL (ACT001) also eliminates LSCs by inhibiting NF-κB activity.²⁹⁴ Regulating aberrant signaling pathways in AML through these inhibitors can effectively target and clear LSCs, thus improving AML prognosis. Therefore, targeted inhibition of signaling pathways remains a promising direction for future research.

7.1.4 Potential challenges and limitations of targeted signaling pathway therapies

Targeted signaling pathway therapies for leukemia have made significant advances in recent years, yet they continue to face several potential challenges and limitations. First, the heterogeneity of leukemia is considerable, with tumor cells from different patients potentially exhibiting distinct signaling pathway abnormalities. This heterogeneity implies that a single targeted therapy may not be uniformly effective across all patients, necessitating personalized treatment approaches to address the specific signaling abnormalities present in individual cases. Second, targeted signaling pathway therapies may induce compensatory mechanisms or activate alternative signaling pathways within tumor cells, potentially reducing therapeutic efficacy or leading to resistance. For instance, drugs targeting a specific signaling pathway might enable cancer cells to continue growth and survival through other pathways. Additionally, targeted therapies may have off-target effects on normal cells, as some signaling pathways are also crucial for normal cellular functions. Therefore, improving the specificity and selectivity of these drugs is essential to minimize adverse effects on healthy tissues. Finally, the long-term efficacy and safety of targeted signaling pathway therapies require further investigation to assess their impact on patients' quality of life and potential long-term toxicities. In conclusion, optimizing the effectiveness of targeted signaling pathway therapies for leukemia involves addressing these challenges and continually refining treatment strategies.

7.2.1 Inhibitors of m6A methylation methyltransferases

Yankova et al.²⁴⁸ identified a METTL3 inhibitor, STM2457, which reduces the methyltransferase activity of METTL3 by binding to the SAM binding site. This leads to decreased m6A methylation of METTL3 target genes such as BRD4, SP1, and c-Myc, thereby inhibiting their translation and inducing apoptosis while suppressing AML cell proliferation.²⁴⁸ UZH1a, a selective, cell-permeable non-nucleoside METTL3 inhibitor, has a stereoisomer UZH1b. Takayuki Murata et al.²⁹⁶ found that increasing doses of UZH1a decreased m6A methylation levels of mRNA transcripts in MOLM-13 cells in a dose-dependent manner and downregulated METTL3 activity, thus promoting leukemia cell apoptosis and inhibiting disease progression. Additionally, Ding et al.⁶⁰ discovered that eltrombopag could block the methyltransferase activity of METTL3 by binding to the METTL3 subunit in the METTL3-14 complex. This results in reduced m6A methylation of target gene mRNAs, promoting AML cell apoptosis and inhibiting cell proliferation. The study also found that eltrombopag has a synergistic effect with other leukemia treatments such as venetoclax and azacitidine, delaying leukemia progression.^{297, 298}

7.2.2 Inhibitors of m6A demethylases

The nonsteroidal anti-inflammatory drug meclofenamic acid (MA) inhibits FTO activity. Huang et al.²⁹⁹ optimized MA to produce two derivatives, FB23 and FB23-2, which specifically inhibit FTO-mediated m6A demethylation. This results in increased m6A levels in MYC, CEBPA, RARA, and ASB2 mRNA transcripts, downregulation of MYC and CEBPA proteins, and upregulation of RARA and ASB2 proteins, thereby inhibiting AML cell proliferation.²⁹⁹ R-2HG, an FTO inhibitor produced by isocitrate dehydrogenase 1/2 (IDH1/2), also inhibits FTO competitively, promotes m6A methylation, decreases the stability of MYC and CEBPA mRNAs, and suppresses AML cell proliferation. Moreover, R-2HG enhances AML cell sensitivity to low-methylation agents such as azacitidine and decitabine, promoting AML cell apoptosis and helping prevent HMA-induced drug resistance.³⁰⁰ Another study developed two FTO inhibitors, CS1 and CS2, which inhibit LILRB4 expression and activate T lymphocyte toxicity. CS1 and CS2 bind directly to FTO residues such as H231, E234, K216, S229, and H231, inhibiting its methyltransferase activity and increasing the m6A abundance of MYC, CEBPA, and snRNA mRNAs, thereby inhibiting leukemia progression.³⁰¹ Subsequently, Prakash et al.³⁰² designed compound 11b, which increases m6A methylation levels in FTO, downregulates MYC, and upregulates RARA protein levels, thereby inhibiting AML cell survival. This suggests the potential for new synthetic inhibitors in clinical applications.³⁰² A team designed the FTO inhibitor GNPIPP12MA, which increases m6A mRNA modification and reduces LSC expression, targeting LSCs and leukemia cells.³⁰³ Lai et al.³⁰⁴ discovered a new selective ALKBH5 inhibitor, TD19, which selectively binds to ALKBH5 residues C100 and C26. By inhibiting ALKBH5's binding to target mRNA, TD19 downregulates ALKBH5's m6A demethylase activity, exerting antileukemia effects.³⁰⁴ Additionally, Wang et al.^{55, 305} developed the ALKBH5 selective inhibitor DDO-2728, which increases m6A abundance on TACC3 mRNA transcripts, decreases mRNA stability, and promotes AML cell apoptosis. DDO-2728 significantly inhibited the growth of primary cells in xenografted mice, showing antileukemia effects. Both TD19 and DDO-2728, as newly developed targeted agents, offer high selectivity and efficacy compared with dual FTO/ALKBH5 inhibitors like IOX3, MV1035, ALK-04, and cmp-3. They selectively target ALKBH5 without affecting FTO, providing more specific regulation of AML cells and promoting tumor cell apoptosis.

7.2.3 Inhibitors of m6A reader proteins

Inhibitors of m6A reader proteins play significant roles in the treatment of various malignancies. YTHDF1 is dispensable for normal hematopoiesis but significantly promotes AML progression. Molecular docking studies have revealed that tegaserod can specifically and concentration-dependently bind directly to the YTH domain of YTHDF1. This binding blocks the interaction between YTHDF1 and m6A-modified target genes, inhibiting the translation of downstream target genes and suppressing AML progression.²⁶⁰ Dixit et al.³⁰⁶ studied Linsitinib in glioblastoma, an IGF1/IGF1R inhibitor, and found that it reduces glioblastoma stem cell activity by downregulating YTHDF2 protein levels, thereby inhibiting tumor cell proliferation and slowing disease progression. Wang et al.³¹⁰ found that the YTHDF2 inhibitor, DC-Y13, directly inhibits the binding of YTHDF2 to m6A-containing RNA, thereby suppressing tumor growth. More interestingly, DC-Y13-27 can further enhance the effects of radiotherapy and anti-PD-L1 therapy. The combination of DC-Y13-27 and anti-PD-L1 significantly slowed the growth of MC38 tumors, while the triple therapy of DC-Y13-27, IR, and anti-PD-L1 produced the strongest antitumor effects.¹¹⁷ Feng et al.³⁰⁷ developed a small-molecule IGF2BP2 inhibitor, JX5, for ALL, which lowers m6A methylation levels by inhibiting IGF2BP2. JX5 weakens the stability of NOTCH1 mRNA, thereby downregulating its expression and inhibiting leukemia cell proliferation.³⁰⁷ In AML research, Weng et al.²⁶³ designed a small-molecule IGF2BP2 inhibitor, CWI1-2, which binds directly to IGF2BP2 and competitively inhibits its m6A-binding protein activity. This suppresses AML cell proliferation and shows significant antileukemia effects. The study also found that CWI1-2 in combination with other leukemia chemotherapy drugs, such as DXR and vincristine, exhibits good therapeutic effects in AML treatment.²⁶³ However, there is limited research on the application of reader protein inhibitors in AML treatment, suggesting that m6A reader proteins may have more complex mechanisms in hematologic malignancies, particularly AML, which warrants further investigation. Overall, the use of m6A-related inhibitors, either alone or in combination with chemotherapy, demonstrates significant anti-AML potential and lays the groundwork for the development of new targeted drugs.

7.2.4 Clinical significance of m6A-targeted therapies

Targeted m6A therapies have demonstrated significant potential in the treatment of leukemia. m6A modification is a prevalent epigenetic alteration found on RNA that plays a crucial role in regulating gene expression, cellular differentiation, and tumorigenesis. In leukemia cells, aberrant m6A modifications are often closely associated with cancer development and progression. Therefore, targeted m6A therapies have the potential to improve leukemia treatment outcomes by modulating these modifications. First, targeting m6A can regulate leukemia cell proliferation and apoptosis by inhibiting or activating specific m6A methyltransferases (e.g., METTL3) or demethylases (e.g., FTO). By altering m6A modification levels, these therapies can correct abnormal transcription and translation processes in leukemia cells, thereby suppressing tumor growth. Second, m6A modifications have profound effects on posttranscriptional regulation; thus, m6A-targeted drugs may enhance therapeutic efficacy by influencing signaling pathways and transcription factors in leukemia cells. For example, modulating m6A modifications could improve cellular sensitivity to chemotherapeutic agents, reduce drug resistance, and consequently enhance treatment effectiveness. Finally, combining m6A-targeted therapies with other treatment strategies (e.g., targeted drug therapy, immunotherapy) may further enhance the overall efficacy of leukemia treatment. By integrating various therapeutic approaches, a more comprehensive attack on leukemia can be achieved, potentially improving patient survival rates and quality of life. In conclusion, m6A-targeted therapies offer new hope for leukemia patients, and future research will focus on optimizing these strategies and evaluating their practical applications in clinical settings.

7.2.5 Potential challenges and limitations of m6A-targeted therapies

m6A-targeted therapies, as an emerging therapeutic strategy, hold significant promise but also encounter several challenges and limitations. First, the complexity and diversity of m6A modifications mean that our understanding of their mechanistic roles remains incomplete. m6A modifications not only affect RNA stability and translation but are also involved in various cellular processes, thus targeting specific m6A sites may inadvertently impact other normal biological functions. Second, the selectivity and specificity of regulatory factors involved in m6A modifications (such as methyltransferases and demethylases) require further investigation to minimize side effects and enhance therapeutic efficacy. Additionally, m6A-targeted therapies may face challenges related to drug delivery and biodistribution, as these therapeutics need to precisely reach and act within target cells. Finally, genetic variability among individuals may influence therapeutic responses, with some patients potentially developing resistance, leading to variability in treatment efficacy. Therefore, a thorough exploration of these potential challenges is essential before advancing m6A-targeted therapies, and appropriate strategies must be developed to improve their clinical safety and effectiveness.