The core function of writers is to transfer the methyl group (-CH₃) from S-adenosylmethionine (SAM, the methyl donor) to the N⁶ position of adenine (A) in RNA molecules, a process executed by a multi-subunit complex [9], where the methyltransferase-like 3/14 (METTL3/14) heterodimer serves as the core catalyst. Both METTL3 and METTL14 contain SAM-binding domains, though only METTL3 possesses direct methyltransferase activity. Although METTL14 is lack of intrinsic catalytic activity, it can facilitate substrate recognition by assisting METTL3 to identify specific RNA sequence motifs [10]. Wilms’ tumor 1 associated protein (WTAP) does not directly participate in catalytic reactions; instead, it primarily functions to recruit the METTL3/14 heterodimer to the transcriptional sites in cell nucleus (such as the spliceosome-associated regions) and regulate their subcellular localization [11]. Moreover, VIRMA recruits the METTL3/METTL14/WTAP complex to preferentially modify 3^′UTRs and regions near the stop codons. Some other regulatory proteins including vir-like m6A methyltransferase associate (KIAA1429) and RNA-binding motif protein15 (RBM15) contribute to complex assembly and functional regulation, thereby ensuring precise m6A deposition. Notably, METTL5 and ZC3H13 govern the m6A modification of 18S and 28S ribosomal RNAs, while METTL16 regulates m6A on non-coding RNAs [12].

The dynamic reversibility of m6A is mediated by fat mass and obesity-associated protein (FTO) and alkB homolog H5 (ALKBH5), two $\alpha$-ketoglutarate ($\alpha$-KG) and Fe (II)-dependent dioxygenases. FTO catalyzes stepwise oxidative demethylation of m6A via N6-hydroxymethyladenosine (hm6A) and N6-formyladenosine (f6A) intermediates, and finally converts m6A to adenosine [13]. It exhibits broad substrate specificity, targeting m6A-modified mRNAs, tRNAs, snRNAs, as well as other RNA modifications such as m6Am and m1A. ALKBH5 directly converts m6A to adenosine. It primarily targets spliceosome-associated regions in the cell nucleus, and preferentially acts on nuclear RNAs (such as pre-mRNA), thereby regulating RNA alternative splicing and nuclear export [14].

m6A-modified RNAs are decoded by specialized “reader” proteins that trigger downstream biological responses, and the key families include YT521-B homology (YTH) domain proteins, insulin-like growth factor 2 mRNA-binding proteins (IGF2BPs), and heterogeneous nuclear ribonucleoproteins (HNRNPs). YTH m6A RNA-binding protein (YTHDF)1/2/3 are mainly located in the cytoplasm, targeting the 3^′UTR region of mRNA at the m6A site. YTHDF1 recruits ribosomes or translation initiation factors (such as eukaryotic initiation factors 3 (eIF3)) to enhance the mRNA translation efficiency. In contrast, YTHDF2 facilitates mRNA decay by recruiting the CCR4-NOT deadenylase complex and other regulatory proteins. YT521-B homology domain-containing family 3 (YTHDF3) exerts dual functions: it synergizes with YTHDF1 to promote mRNA translation, and also interacts with YTHDF2 to enhance YTHDF2-mediated RNA degradation. YTHDC1/2 are mainly located in the cytoplasm, and YTHDC1 regulates pre-mRNA splicing and nuclear export [15]. The IGF2BP family proteins (IGF2BP1–3) recognize m6A modification sites through their unique domains. After binding to mRNA, they can enhance its stability and promote its translation process [16]. Heterogeneous nuclear ribonucleoprotein A2B1 (HNRNPA2B1) regulates pre-miRNA processing, whereas HNRNPC and HNRNPG mediate alternative splicing of m6A-containing transcripts [17]. In addition, HUR, eIF3, and Prrc2a further expand the functional repertoire of m6A recognition, linking RNA methylation to diverse cellular processes from translation control to epigenetic memory. The roles of different regulatory factors in m6A modification are summarized in Table 1.

Liquid chromatography-tandem mass spectrometry (LC-MS/MS) and colorimetric assays are commonly employed for global m6A quantification. LC-MS/MS involves the enzymatic digestion of RNA, followed by chromatographic separation and mass spectrometric analysis to accurately determine the ratio of m6A to total adenosine [18]. Colorimetric assays, which are based on ELISA-like principles, utilize m6A antibodies and color development for semi-quantitative detection [19]. Both approaches are straightforward and appropriate for assessing the overall m6A level. However, they cannot provide transcript- or site-specific information, which limits their utility in mechanistic studies.

MeRIP-seq (m6A-seq) remains the most widely used method for m6A detection, which combines m6A-specific antibody-based immunoprecipitation of methylated RNA fragments with high-throughput sequencing for transcriptome-wide mapping of m6A modifications [20]. This approach benefits from a well-established workflow and commercially available reagents, but suffers from several limitations, including low resolutions (typically spanning 100–200 nucleotides), inability to pinpoint the modified adenosines at single-base resolution, low reproducibility, and requirement for a substantial amount of input RNA. To overcome the resolution constraint, a m6A individual-nucleotide-resolution cross-linking and immunoprecipitation (miCLIP) method was developed [21], which employs UV cross-linking to covalently link antibodies to m6A sites, introducing mutations or truncations during reverse transcription to allow base-resolution mapping. However, the miCLIP method is still highly dependent on antibody quality and cross-linking efficiency, and struggles in resolving closely adjacent m6A sites.

Advancements in epitranscriptomics have led to the emergence of several novel m6A detection technologies. In 2019, Meyer [22] developed DART-seq, which employs a fusion protein of the YTH domain and APOBEC1 enzyme to identify m6A sites in an antibody-free manner. In the same year, researchers introduced MAZTER-seq and m6A-REF-seq, both leveraging the m6A-sensitive MazF ribonuclease for site-specific cleavage within defined motifs [23]. Subsequent innovations include m6A-SEAL-seq, which employs the demethylase FTO for selective chemical labeling [24], and m6A-SAC-seq based on selective allyl chemical labeling [25]. Beyond these next-generation sequencing-based methods, nanopore direct RNA sequencing represents a third-generation strategy that enables the detection of modifications in native RNA, and does not require reverse transcription, antibodies, or enzymatic steps [26]. This method provides single-molecule resolution and can simultaneously identify a broad spectrum of RNA modifications beyond m6A.

Despite these advancements, m6A detection is still faced with multiple challenges, including limited resolution and specificity, antibody dependency, sample throughput restrictions, lack of data standardization, difficulties in detecting isoform-specific modifications, and high sequencing costs [27]. Looking forward, the field is moving toward the integration of multiple technologies, single-cell multi-omics analyses, and clinical translation. In parallel, the expansion of public datasets and development in deep learning are expected to yield more accurate and efficient algorithms for m6A identification and quantification.

The m6A modification has been reported to participate in nearly all aspects of RNA metabolic regulation, including mRNA polyadenylation, pre-mRNA splicing, nuclear export of mRNA, as well as mRNA stability and translation under normal physiological conditions [28]. m6A modification plays a pivotal role in regulating critical biological processes such as cell proliferation, differentiation, development, and senescence [29]. Dysregulation of m6A modification is closely linked to the pathogenesis and progression of multiple diseases, particularly in cancers. During tumorigenesis, aberrant expression or functional perturbation of m6A regulatory enzymes, including “writers” (such as METTL3 and METTL14), “erasers” (such as FTO and ALKBH5), and their associated regulators, disrupts global and transcript-specific m6A methylation landscapes. This dysregulation subsequently modulates a broad spectrum of malignant phenotypes, encompassing cancer initiation, progression, metastasis, metabolic reprogramming (such as glycolysis), therapy resistance, immune evasion, cancer stem cell self-renewal, and remodeling of the tumor microenvironment. These findings collectively underscore the significant therapeutic potential of targeting the dysregulated m6A machinery. The role of m6A modification in other types of cancer has been reviewed in detail by other studies [30, 31].

Aberrant m6A modification frequently occurs under CRC, and is closely associated with tumor initiation, progression, metastasis, and drug resistance. Therefore, comprehensive insights into the mechanistic roles of m6A modification in CRC hold significant theoretical and clinical importance in elucidating the disease pathogenesis and identifying novel therapeutic targets.

Numerous studies have demonstrated aberrant expression of m6A modification proteins in both CRC tissues and cell lines, and their dysregulation is mechanistically linked to CRC and disease progression. It has been shown that most m6A-related genes such as METTL3, WATP, and YTHDF1 are significantly up-regulated in tumor tissues, while METTL14, YTHDF3, and ALKBH5 were downregulated in CRC as determined with multiple databases (The Cancer Genome Atlas (TCGA), Gene Expression Omnibus (GEO), Human Protein Atlas (HPA) and TMA) [32, 33]. An analysis of MeRIP-seq data revealed the presence of m6A peaks in most mRNAs in CRC patients, and CRC samples had 1343 out-of-balance m6A peaks (625 significantly up-regulated and 718 significantly down-regulated) [34]. Wang et al. [35] further demonstrated that elevated m6A modification levels and expression of METTL3, METTL16, and WTAP are significantly associated with poor clinical outcomes in CRC patients, suggesting their potential as prognostic biomarkers. The dysregulation of m6A modification proteins confer their oncogenic or anti-oncogenic functions in CRC pathogenesis and progression.

Consistently, it has been demonstrated that METTL3 and WTAP are overexpressed in CRC tissues and strongly correlated with poor prognosis in CRC patients. From a functional perspective, these m6A writers act as oncogene by promoting CRC cell invasion, migration, tumor progression, cancer stem cell maintenance, and drug resistance [36, 37, 38]. In contrast to METTL3, METTL14 shows opposite expression patterns and functions in CRC. METTL14 is downregulated and regarded as an independent risk factor in CRC. It has been shown that KDM5C-mediated demethylation of H3K4me3 in the METTL14 promoter reduces METTL14 transcription. Moreover, METTL14 increases the m6A methylation of SOX4 mRNA in a YTHDF2-dependent manner, thereby inhibiting SOX4-mediated EMT process and the PI3K/AKT pathway, ultimately suppressing CRC tumorigenesis [39].

FTO is significantly upregulated in CRC tissues and cells, where it promotes tumor cell proliferation while suppressing apoptosis [40, 41]. In contrast, the role of another demethylase ALKBH5 in CRC remains controversial: one study proposed that ALKBH5 downregulation is significantly associated with poor prognosis in CRC patients [42], while another study demonstrated that elevated ALKBH5 expression represents unfavorable prognosis in CRC [43].

m6A modification plays an important role in the occurrence and development of CRC by modifying the mRNA of some key CRC genes, affecting their stability, translation efficiency, splicing, and other processes, thereby regulating their expression levels (Table 2, Ref. [41, 42, 43, 44, 45, 46, 47, 48, 49, 50, 51, 52, 53, 54, 55, 56, 57]). METTL3 affects the m6A modification level of Myc proto-oncogene protein (MYC), influences the stability of MYC mRNA in an IGF2BP1-dependent manner, and promotes CRC cell proliferation, migration, and invasion [44]. Moreover, it can increase the transcripts of Sry-related HMG box (SOX)2 and maintain the mRNA stability of SOX2 together with IGF2BP2, thereby affecting the expression of downstream target genes MYC, CCND1, and pOU5F1, and further influencing tumorigenesis and tumor metastasis [45]. METTL3 further drives CRC progression by regulating proliferation-related genes. Overexpression of METTL3 was found to enhance m6A modification in the 5^′UTR region of cyclin E1 (CCNE1) mRNA, thereby increasing its stability and ultimately promoting the proliferation and colony formation capabilities of CRC cells [46]. METTL3 also promotes m6A modification on lncRNA hepatocyte nuclear factor 1-$\alpha$ (HNF1A)-AS1 and YPEL5 mRNA with IGF2BP2 and YTHDF2 to regulate HNF1A-AS1 stability and YPEL5 translation. Subsequently, HNF1A-AS1 and YPEL5 regulate CCNB1 and PCNA expression, thereby stimulating the expansion of CRC cells [47, 58]. Stromal antigen 3 (STAG3) is also regulated by m6A modification. METTL3 knockdown reduces m6A methylation and regulates STAG3 protein expression together with the reader protein IGF2BP2, thereby inhibiting CRC cell proliferation and migration while promoting apoptosis [48]. However, it has been identified that METTL14 regulates the expression of the tumor suppressor Krüppel-like factor 4 (KLF4) by altering its m6A methylation [49].

Zhou et al. [50] showed that silencing of WTAP potently restrained the CRC tumorigenesis in virto and in vivo. Mechanically, WTAP regulates SOD2 m6A modification in an IGF2BP3-dependent manner to maintain its mRNA stability. Notably, SOD2 overexpression could rescue the tumor-suppressive effects induced by WTAP knockdown. The m6A methyltransferase KIAA1429 acts as an oncogenic factor in CRC. It has been suggested that KIAA1429 promotes the proliferation, colony formation, and growth of CRC cells by regulating the level of m6A and expression of Wee1-like protein kinase 1 (WEE1), SIRT1, and NF$\kappa$B1 [51, 52]. Another methyltransferase RBM15 can bind to E2F transcription factor 2 (E2F2) to increase its m6A binding capability and stabilize the corresponding E2F2 mRNA formation to result in malignant cellular processes in CRC cells [53].

Numerous studies have demonstrated that FTO plays pivotal roles in regulating cell cycle progression and apoptosis in CRC. Mechanistically, FTO reduces the m6A modification level of the key oncogenes MYC and myeloid zinc finger 1 (MZF1), thereby increasing the proportion of S-phase CRC cells [41]. Downregulation of ALKBH5 is significantly associated with poor prognosis in CRC patients, and ALKBH5 has been revealed to suppress the proliferation, migration, and invasion capacities of LOVO and RKO cells. In terms of mechanism, ALKBH5 may destabilize PHF20 mRNA by removing m6A modification in the 3^′UTR region [42]. In contrast, another study demonstrated that elevated ALKBH5 expression generally represents unfavorable prognosis in CRC [43]. Shen et al. [54] found that ALKBH5 promotes the proliferation, migration, and invasion abilities of CRC cells in vitro and enhances subcutaneous tumor growth in vivo. Mechanistically, ALKBH5 can perform post-transcriptional activation of RAB5A by m6A demethylation through a YTHDF2-mediated pathway.

Reader proteins regulate RNA metabolic processes through m6A and are involved in the regulation of CRC. YTHDF1 directly binds to m6A-modified SH3TC2 mRNA and enhances its expression, thereby accelerating cell cycle progression and promoting tumor growth in CRC [55]. Yang et al. [56] demonstrated that IGF2BP3 promotes the cell cycle progression and proliferation of CRC cells by recognizing m6A modification in CCND1 and enhancing its mRNA expression.

Extensive research has demonstrated that m6A modification regulates CRC progression through multiple signaling pathways (Table 3, Ref. [59, 60, 61, 62, 63, 64, 65, 66, 67, 68]). METTL3 modulates the m6A modification level of crumbs protein homolog 3 (CRB3) mRNA in CRC cells. The m6A-YTHDF2 axis suppresses CRB3 protein translation efficiency, ultimately reducing the phosphorylation levels of key Hippo pathway components (MST1, LATS1, MOB1, and YAP) [59]. METTL3 upregulates plasminogen activator (PLAU) mRNA in an m6A-dependent manner, and then participates in the MAPK/ Extracellular signal-regulated kinase (ERK) pathway to promote angiogenesis and metastasis in CRC [60]. METTL3 also regulates the m6A modification of Membrane-bound erythropoietin-producing hepatocellular receptor tyrosine kinase class A2 (EphA2) and vascular endothelial growth factor A (VEGFA), and stabilizes their transcripts through IGF2BP2/3 to prevent their degradation, thereby activating both the PI3K/AKT and ERK1/2 signaling pathways and promoting vasculogenic mimicry formation in CRC [61]. Notably, METTL3 upregulates JAK1 and Signal transducer and activator of transcription 3 (STAT3) expression through both m6A-dependent and -independent mechanisms, and cooperates with NF-$\kappa$B to enhance STAT3 transcription, thereby activating the p-STAT3 signaling pathway and driving CRC progression [62]. These findings underscore the mechanistic complexity of m6A-mediated regulation in CRC. However, some conflicting results demonstrate that METTL3 downregulation activates the p-p38 and p-ERK signaling pathways, and inhibition of p38 or ERK significantly reverses the enhanced migration and invasion capacities of CRC cells induced by METTL3 knockdown [63].

YTHDF1 is widely involved in the regulation of signaling pathways. It can promote the expression of ARHGEF2, TCF7L2, T-cell factor 4 (TCF4), and Glucocorticoid modulatory element-binding proteins 2 (GMEB2) in an m6A-dependent manner, thereby activating multiple oncogenic signaling pathways including RhoA, Wnt/$\beta$-catenin, and NF-$\kappa$B, and ultimately promoting tumor cell growth, migration, invasion in CRC [64, 65, 66]. YTHDF2 is also involved in regulating the Wnt/$\beta$-catenin signaling pathway. It can regulate the expression of Glycogen synthase kinase 3$\beta$ (GSK3$\beta$) and six-transmembrane epithelial antigen of the prostate family member 3 (STEAP3), which influence Wnt/$\beta$-catenin to regulate CRC progression. Intriguingly, YTHDF2 is regulated by miR-6125 and interacts with lncRNA STEAP3-AS1 to modulate Wnt/$\beta$-catenin signaling [67, 68].

There have been emerging studies revealing the dual substrate specificity of m6A modification in CRC, as it mediates both protein-coding transcripts and non-coding RNAs to orchestrate oncogenic programs (Table 4, Ref. [58, 69, 70, 71, 72, 73, 74, 75, 76, 77, 78, 79, 80, 81, 82]). It has been revealed that porcine retina-derived POU domain factor 1 (POU6F2)-AS1 may function as a novel oncogenic lncRNA in CRC with potential diagnostic and therapeutic value. In vitro studies have demonstrated that POU6F2-AS1 promotes the growth of both CRC cells and patient-derived organoids (PDOs). METTL3 induces the modification of m6A and stability of POU6F2-AS1 through IGF2BP2, and upregulation of POU6F2-AS1 could tether YBX1 to the FASN promoter to induce its transcriptional activation, thus facilitating the growth and lipogenesis of CRC cells [69]. Similarly, the methyltransferase KIAA1429 regulates POU6F2-AS1 to modulate CRC cell survival, migration, and metastasis, underscoring the diversity of m6A-mediated lncRNA regulation in cancer [70]. METTL3-mediated m6A modification can induce abnormal LINC02418 expression in CRC, where LINC02418 regulates the proliferation and migration of CRC cells by interacting with YBX1 to activate CTNNB1 transcription, and promoting the interaction between IGF2BP1 and CTNNB1 mRNA to enhance CTNNB1 stability [71]. It has also been confirmed that METTL3 is critical for maintaining the RNA stability of alpha/beta hydrolase domain-containing protein 11-antisense RNA 1 (ABHD11-AS1) and Family with sequence similarity 83 member H (FAM83H)-AS1 through IGF2BP2/3, where ABHD11-AS1 enhances FOXM1 stability and FAM83H-AS1 specifically binds to PTBP1 to regulate its phosphorylation and splicing function in CRC cell proliferation and migration and suppress ferroptosis [72, 73].

It has been demonstrated that METTL14 significantly suppresses CRC cell proliferation and migration potential through YTHDF2-mediated inhibition of the oncogenic lncRNA XIST [74]. ZCCHC4, a newly identified m6A methyltransferase, has been functionally validated in CRC, where it downregulates the lncRNA GhrelinOS (GHRLOS) to modulate KDM5D expression, thereby promoting CRC cell survival and migration [75].

Guo et al. [76] has confirmed that ALKBH5 increases the expression level of lncRNA nuclear paraspeckle assembly transcript 1 (NEAT1) by lowering its m6A enrichment to enhance cell proliferation and tumor growth while inhibit apoptosis in CRC cells. Inversely, another study reported that ALKBH5-mediated demethylation of m6A on lncRNA Cardiac mesoderm enhancer-associated non-coding RNA (CARMN) can enhance its stability through YTHDF2 recognition, thereby sustaining the expression level of CARMN. Functionally, the CARMN-miR-5683-FGF2 axis suppresses CRC cell proliferation, invasion, and migration potential [77].

Circular RNAs (circRNAs) have emerged as pivotal players in deciphering cancer pathogenesis due to their exceptional stability and resistance to RNase R digestion. Circ ubiquitin-like with PHD and ring finger domains 2 (UHRF2), a novel circRNA, plays a critical oncogenic role in CRC progression by facilitating the binding of IGF2BP1 to DDX27 mRNA, thereby enhancing its stability and expression.

Circ_0003215 is significantly downregulated in CRC and demonstrates strong negative correlations with tumor size, TNM stage, and lymph node metastasis. Mechanistically, YTHDF2 binds to circ_0003215 and promotes its RNA degradation. Functionally, circ_0003215 acts as a tumor suppressor by sponging miR-663b to upregulate DLG4 expression and G6PD ubiquitination [78]. Zeng et al. [79] demonstrated that YTHDC1 facilitates cytoplasmic translocation of m6A-modified circ fibronectin type III domain containing 3B (FNDC3B), thereby enhancing the stability of RNF41 mRNA to promote ASB6 degradation, and ultimately inhibiting CRC stemness and metastasis. CircEZH2 interacts with IGF2BP2, impeding its ubiquitination-dependent degradation, and also acts as a miR-133b sponge to upregulate IGF2BP2 expression. Taken together, the circEZH2-IGF2BP2 axis stabilizes cyclic-AMP response element-binding protein 1 (CREB1) mRNA and exacerbates CRC cell proliferation and migration [80].

Dysregulation of YAP has significant implications for the pathological biology and progression of CRC. The lncRNA pituitary tumor-transforming 3, pseudogene (PTTG3P) undergoes m6A methylation mediated by METTL3 and IGF2BP2 to enhance its stability, which promotes glycolytic CRC reprogramming and proliferation in cooperation with YAP1 [38]. YTHDF3 has been identified as a novel target of YAP, and their mRNA and protein levels have significant positive correlations. The circ-YAP transcript encodes a novel 220-amino acid truncated protein (YAP-220aa) that functions as a competitive inhibitor of LATS1, which promotes YAP dephosphorylation and nuclear translocation, thereby activating a pro-metastatic transcriptional program. Notably, YTHDF3 can cooperate with eIF4G to enhance the translation of circ-YAP, while the YAP/TEAD transcription factor complex simultaneously upregulates circ-YAP expression, establishing a self-reinforcing positive feedback loop to promote liver metastasis of CRC [81]. It has also been found that the m6A reader protein PRRC2A directly targets CSNK1E (encoding CK1$ϵ$) to maintain its RNA stability, and high levels of CK1$ϵ$ can simultaneously activate both the Wnt and YAP signaling pathways, thereby promoting the progression of CRC [82].

The regulation of lncRNAs and circRNAs by m6A modification offers novel perspectives for CRC diagnosis. However, its clinical application is faced with two major challenges: the stability of non-coding RNAs and insufficient biomarker specificity due to tumor heterogeneity. Therefore, integration of m6A sequencing with liquid biopsy techniques (such as exosomal RNA analysis) may enhance the sensitivity of early detection in future studies. Notably, although many lncRNAs and circRNAs demonstrate promising potential in mechanistic research, their validity as independent prognostic indicators still requires large-scale validation with collaboration to mitigate sampling bias.

Cancer cells mainly rely on glycolysis for energy supply, which enables cancer cells to tolerate hypoxic environments, while also increases the lactate level in the microenvironment of cancer cells, thereby stimulating tumor growth and development (Table 5, Ref. [38, 83, 84, 85, 86, 87, 88, 89, 90, 91, 92, 93]) [94]. There has been accumulating evidence demonstrating that m6A modification regulates tumor glycolysis through multiple pathways, thereby exerting critical impacts on cancer proliferation, metastasis, and therapeutic response. In CRC cells, METTL3 directly targets lncRNA PTTG3P, HK2, glucose transport protein type 1 (GLUT1) and PNN genes, enhancing their mRNA stability by m6A modification through IGF2BP2/3 or YTHDF1, ultimately promoting glycolysis and proliferation in CRC cells [38, 83, 84]. The methyltransferase KIAA1429 also enhances HK2 mRNA methylation in an m6A-dependent manner, leading to increased transcription level and stability of HK2. This metabolic reprogramming accelerates aerobic glycolysis in CRC cells, thereby promoting tumor aggressiveness [85]. METTL14 inhibits glycolysis, and wild-type p53 activates its transcription, which promotes the biogenesis of pri-miR-6769b and pri-miR-499a. Ultimately, the miR-6769b-3p-GLUT3 and miR-499a-3p-PGAM1 axes suppress aerobic glycolysis and malignant phenotypes in p53 WT CRC cells [86]. Wu et al. [87] proposed that ALKBH5/IGF2BPs bind to mjC domain-containing protein 8 (JMJD8) and modulate its stability in an m6A-dependent manner, increasing glycolysis and accelerating the development of CRC by enhancing the enzymatic activity of PKM2. Interestingly, the synthesized ALKBH5 mRNA-loaded folic acid-modified exosome–liposome hybrid nanoparticles were found to significantly inhibit the progression of CRC.

In CRC patients, cancer metastasis is the primary cause of deaths, with the liver and lungs being the most common target organs for metastasis. A previous study has shown that METTL3 facilitates epithelial-mesenchymal transition of pulmonary metastasis by targeting the m6A-Snail- CXC motif chemokine ligand (CXCL)2 axis to recruit M2-type immunosuppressive macrophages [88]. LINC00460 promotes EMT, CRC cell proliferation, migration, invasion in vitro, and tumor growth and metastasis in vivo. Mechanistically, METTL3 elevates high mobility group A1 (HMGA1) m6A methylation, while LINC00460 directly interacts with IGF2BP2 and DHX9 to bind HMGA1 mRNA, thereby enhancing its stability [89]. High mobility group A2 (HMGA2) (a member of the HMGA1 protein family) has been confirmed to play an oncogenic role in CRC liver metastasis. Mechanistic studies have revealed that circNSUN2 forms a complex with IGF2BP2 to directly bind and stabilize HMGA2 mRNA, thereby driving liver metastasis in CRC [90]. IGF2BP2 also binds to two m6A motifs on LncRNA $\beta$-secretase (BACE1)-AS for its stabilization, promoting TUFT1 expression via a ceRNA network involving miR-214-3p, which in turn facilitates CRC liver metastasis [91]. Overexpression of lncRNA SLERT was found to promote liver metastasis in CRC. Mechanistically, SLERT binds to the methyltransferase RBM15, indicating its ability to stabilize hormonally up-regulated neu-associated kinase (HUNK) mRNA, which will lead to reduced HUNK expression and enhanced metastatic potential in CRC [95].

Tumor-associated macrophages (TAMs) can reduce the antitumor activity of T cells.

Angiogenesis plays an important role in the occurrence and development of CRC. Ye et al. [92] found that WTAP overexpression exacerbates CRC cell proliferation, migration, invasion, and angiogenesis. WTAP upregulates VEGF expression in a YTHDC1-dependent manner, which subsequently activates the MAPK signaling pathway to drive angiogenesis.

Upregulation of glutaminolysis facilitates energy generation and biomass accumulation, thereby providing energy for CRC cell proliferation. Han et al. [93] demonstrated that FTO promotes ATF4 expression through m6A-dependent stabilization of its mRNA, which can counteract YTHDF2-mediated decay under glutaminolysis inhibition. Functionally, ATF4 activates DDIT4 transcription to inhibit mTOR signaling, thereby triggering pro-survival autophagy in CRC cells. Fig. 2 shows the schematic diagram illustrating the role of m6A modification in CRC.

Fig. 2.

Roles of m6A modification in CRC progression. (A) m6A modification proteins function as oncogene or tumor suppressors. (B) m6A modification orchestrates CRC progression through master signaling pathways. (C) m6A modification regulates CRC progression through non-coding RNA. (D) m6A modification governs CRC progression by pivotal biological events.

METTL3 and METTL14 are core components of methyltransferases, and can form a heterodimer to exert consistent regulatory functions to control other biological events. However, they play opposite roles in CRC: METTL3 generally acts as a promoter, while METTL14 mostly functions as a suppressor, which may be attributed to the differences in their specific upstream and downstream mechanisms within cells. Methyl-CpG binding protein 2 (MeCP2) is a methylated DNA-binding protein, and has been reported to have oncogenic functions in gastric cancer, CRC, hepatocellular carcinoma, and osteosarcoma. It was identified that MeCP2 can bind to METTL14 to occupy the interaction interface between METTL3 and METTL14, indicating a competitive relationship between MeCP2 and METTL3 for METTL14. Furthermore, MeCP2 acts as an oncogene in CRC progression by interacting with METTL14 and reducing m6A modification [49]. Some studies have shown that METTL3 can promote glycolysis by affecting the downstream target genes such as lncRNA PTTG3P, HK2, GLUT1, and pNN, thereby facilitating the proliferation of CRC cells [38, 83, 84]. In contrast, METTL14 inhibits glycolysis via pri-miR-6769b and pri-miR-499a, thus suppressing the progression of CRC [86]. Therefore, the contrasting roles of METTL3 and METTL14 in CRC may be ascribed to their distinct upstream binding factors and downstream regulatory pathways.

In addition, the demethylases FTO and ALKBH5 exhibit paradoxical roles in CRC. Many studies have demonstrated that FTO plays a promoting role in CRC progression. However, another study showed that FTO may act as anti-tumor factor in CRC progression. FTO expression was found to prevent YTHDF2-mediated degradation of the anti-proliferative and pro-apoptotic factor potassium channel tetramerization domain containing 15 (KCTD15) mRNA. Furthermore, KCTD15 overexpression downregulates HDAC1 expression and enhances p53 acetylation, thereby establishing a tumor-suppressive signaling cascade [57]. We hypothesize that these paradoxical roles may be attributed to two factors. On the one hand, it may be caused by differences in experimental samples; on the other hand, it may be associated with the microenvironment for the CRC cells. Specifically, in tumors with abundant immune cell infiltration, FTO tends to regulate apoptosis-associated genes (such as KCTD15).

ALKBH5 plays even more complex bidirectional roles in CRC. First, different studies have reported inconsistent expression patterns of ALKBH5 in CRC, with some showing its upregulation while others displaying its downregulation. Second, numerous studies have documented both pro-tumorigenic and anti-tumorigenic effects of ALKBH5 in CRC. It remains challenging to elucidate the mechanism underlying such paradoxical roles. We suspect that one possible reason may be the variations in clinical stages across different studies. For instance, in early-stage CRC, ALKBH5 appears to suppress cell proliferation by inhibiting PHF20 [42]; in contrast, in advanced metastatic lesions, it seems to promote tumor invasion by activating RAB5A [54]. This stage-dependent effect suggests that the clinical utility of ALKBH5 as a therapeutic target must be evaluated in the context of tumor progression.

The main treatment for CRC is usually surgery combined with radiotherapy, chemotherapy, targeted therapy, and immunotherapy [96]. Chemotherapy is key for advanced/metastatic CRC, and the first-line drugs such as 5-fluorouracil (5-FU) and oxaliplatin (OX) form the FOLFOX regimen (OX inhibits DNA replication) [97]. Immunotherapy, especially the use of immune checkpoint inhibitors (ICIs) such as PD-1/PD-L1 inhibitors (approved by the FDA in 2014), shows promise for metastatic CRC by activating T cells in the tumor microenvironment [98]. Although immunotherapy has advanced significantly in CRC treatment, its clinical use is limited by low response rate, unique toxicity, and acquired resistance. Thus, targeted therapy has become the optimal option for personalized and comprehensive CRC treatment [99]. It targets dysregulated pathways (such as EGFR and Wnt/$\beta$-catenin) with drugs such as cetuximab (for mCRC with wild-type RAS/BRAF, blocking EGFR signaling) and bevacizumab (inhibiting angiogenesis). The treatment of CRC has been reviewed in detail by several papers [100, 101, 102]. While these drugs have demonstrated clinical efficacy and can improve patient survival, their therapeutic potential is significantly compromised by the frequent emergence of drug resistance in practice.

Chemoresistance to OX and 5-FU agents in CRC cells represents a major therapeutic challenge in advanced CRC, and is the primary cause of treatment failure and poor patient prognosis. Lan et al. [103] demonstrated that OX-resistant CRC tissues exhibit elevated total m6A RNA levels and increased expression of METTL3. Mechanistically, M2-polarized TAMs enhance the OX resistance via elevating METTL3-mediated m6A modification of tumor necrosis factor receptor associated factor 5 (TRAF5) in cells [103]. Zhang et al. [104] and Li et al. [105] observed upregulation of both METTL3 expression and global m6A level in 5-FU-resistant CRC cell lines. METTL3 increases lactate dehydrogenase A (LDHA) transcription by stabilizing HIF-1$\alpha$ and triggers the translation of LDHA mRNA via YTHDF1. LDHA catalyzes the conversion of pyruvate to lactate, thereby promoting glycolysis and inducing resistance to 5-FU. This pathway concurrently involves metabolic reprogramming (lactate accumulation) and DNA damage repair. Targeting METTL3 can simultaneously reverse both the metabolic phenotype and drug resistance, which has been supported by early-phase clinical data on METTL3 inhibitors in solid tumors. METTL3 also regulates the expression of RAD51-associated protein 1 (RAD51AP1), promoting 5-FU resistance in CRC cells by attenuating DNA damage accumulation and cell apoptosis. FTO also enhances chemoresistance in CRC. It has been shown that FTO promotes 5-FU resistance in CRC cells by participating in DNA damage, apoptosis, and iron homeostasis through the demethylation of G6PD/ Poly (ADP-ribose) polymerase 1 (PARP1), SIVA1, and nuclear protein 1 (NUPR1) [40, 106, 107].

Wnt/$\beta$-catenin signaling plays an important role in maintaining CRC, and Sec62 upregulation is associated with the chemoresistance of CRC and poor outcomes of CRC patients. Furthermore, upregulation of Sec62 by METTL3-mediated m6A modification promotes the stemness and chemoresistance of CRC by binding to $\beta$-catenin and enhancing Wnt signaling. Thus, the m6A modification-Sec62-$\beta$-catenin molecular axis may act as a therapeutic target in improving the treatment of CRC [108]. YTHDF1 is positively associated with cisplatin resistance of CRC cells, and binds to the 3^′UTR of GLS1 to promote the synthesis of GLS1 protein, which finally mediates cisplatin resistance through the GLS1-glutamine metabolism axis [109]. Another study found that YTHDF3 is highly expressed in OX-resistant (OXAR) CRC tissues and cells. YTHDF3 is enriched in the 5^′UTR of RNA from m6A-methylated tumor resistance genes (ATPase 1 (ATP7A), dual-specificity tyrosine-regulated kinase 1B (DYRK1B), Somatic excision repair cross-complementing rodent repair deficiency (ERCC1)), and promotes the translation of target genes by recruiting eIF3A. EIF2 $\alpha$ kinase (eIF2AK2) enhances the stability of the YTHDF3/eIF3A complex in OXAR CRC cells, which collaboratively regulate the translation process of target genes in oxaliplatin-resistant CRC [110]. YTHDF3 is specifically overexpressed in resistant cells, and its interaction with eIF3A can be disrupted by small molecules, providing a well-defined molecular interface for precision drug design. IGF2BP3 is upregulated in HCT8/T cells and affects doxorubicin (Dox) sensitivity. It can recognize m6A-modified ATP-binding cassette sub-family B member 1 (ABCB1) mRNA, enhance its stability, and promote its expression. Upregulation of ABCB1 leads to DOX resistance in CRC cells [111].

The TAMs, which consist of extracellular matrix, myofibroblasts, cytokines, fibroblasts, neuroendocrine cells, adipocytes, immune-related cells, and blood vessels, induce phenotypic changes in cancer cells and immune cells through complex molecular mechanisms to promote immune escape [112]. In CRC cells, deletion of METTL3/METTL14 causes dysfunction of CD8+ T cells, thereby restricting the anti-tumor T cell response in the tumor immune microenvironment. Wang et al. [113] found that deficiency of METTL3/METTL14 in CRC significantly increases the infiltration of CD8+ T cells and promotes the secretion of cytokines such as IFN-$\gamma$, CXCL9, and CXCL10, thereby enhancing the response of patients to anti-PD-1 therapy. T regulatory (Treg) cells express the immune checkpoint receptor of CTLA-4 and promote tumor immunological tolerance. In CRC, Circ Quiescin Q6 sulfhydryl oxidase 1 (QSOX1) is methylated under the action of METTL3, and its stability is improved after binding to IGF2BP2. The m6A-modified CircQSOX1 can recruit miR-326 and miR-330-5p to increase the expression of PGAM1 and enhance glycolysis to promote lactate production, thereby facilitating immune escape of CRC, which weakens the therapeutic effect of anti-CTLA-4 [114]. In a CRC cell/T cell co-culture system, METTL16 can regulate immune evasion by controlling PD-L1 RNA stability. Furthermore, METTL16 overexpression in CRC cells induces a decrease in the proportion of PD-1 positive T cells [115].

Zhai et al. [43] reported that elevated ALKBH5 expression represents unfavorable prognosis in CRC. ALKBH5 demethylates the m6A of axis inhibition protein 2 (AXIN2) mRNA, leading to its dissociation from IGF2BP1 and subsequent degradation, which upregulates the downstream Wnt/$\beta$-catenin target gene dickkopf-1 (DKK1) and recruits myeloid-derived suppressor cells (MDSCs), thereby fostering an immunosuppressive tumor microenvironment. Finally, vesicle-like nanoparticles encapsulating ALKBH5-siRNA or anti-DKK1 antibodies were found to significantly enhance the efficacy of anti-PD-1 therapy in suppressing CRC growth by potentiating anti-tumor immunity [43].

In terms of targeted therapy drugs and resistance to them, lncRNA MIR100HG sustains cetuximab resistance and facilitates invasion and migration of CRC cells both in vitro and in vivo. HNRNPA2B1 interacts with MIR100HG to maintain the mRNA stability of TCF7L2, which activates Wnt/$\beta$-catenin signaling to confer resistance to cetuximab [116]. IGF2BP3 acts as an m6A reader, and collaborates with METTL14 to enhance the mRNA stability and translation of EGFR, thereby activating the EGFR signaling pathway. Moreover, IGF2BP3 contributes to the resistance of CRC cells to the EGFR-targeted antibody cetuximab [117].

Notably, drug-resistant cancer cells, particularly those in a mesenchymal state with high metastatic potential, generally exhibit marked vulnerability to ferroptosis. This unique vulnerability makes pharmacological induction of iron oxidation a promising therapeutic strategy against refractory malignancies [118]. Curdione can induce ferroptosis to trigger necrotic death of CRC cells. Mechanistically, it promotes the expression of METTL14 and YTHDF2, which in turn elevates m6A methylation and upregulates the expression of light chain subunit solute carrier family 7 member 11 (SLC7A11) and Human class I homeobox A13 (HOXA13) genes [119].

Zhang et al. [120] found that targeting AKT could significantly induce GPX4-dependent ferroptosis and suppress CRC growth both in vitro and in vivo. AKT inhibition downregulates FTO expression, increasing the m6A modification of GPX4 mRNA. The m6A-marked GPX4 transcripts are subsequently recognized and degraded by YTHDF2, which subsequently suppresses CRC progression. The demethylase ALKBH5 also exhibits pro-ferroptotic effects on CRC cells. ALKBH5 is downregulated in CRC and its overexpression reduces SLC7A11 transcription by erasing m6A modification, thus promoting ROS release and ferroptosis [121]. The schematic diagram for the role of m6A modification in CRC drug resistance is shown in Table 6 (Ref. [103, 104, 105, 106, 107, 108, 109, 110, 111, 113, 114, 115, 116, 117, 119, 120, 121]) and Fig. 3.

Fig. 3.

Role of m6A modification in CRC drug resistance. (A) Chemoresistance to chemotherapeutic agents 5-FU and OX. (B) Resistance to targeted therapy drugs. (C) Immune escape (tumor microenvironment). (D) Vulnerability to ferroptosis.

While the role of the m6A regulatory axis in drug resistance has been well-established, its clinical translation potential remains not fully explored. To date, inhibitors targeting m6A modification have been extensively reported [122]. It has been shown that STM2457 has great impacts on cell growth suppression and apoptosis of CRC cells in vitro and subcutaneous xenograft growth in vivo mediated by the METTL3/ASNS axis [123]. In 2019, a new, specific, and cell active FTO inhibitor compound FB23/FB23-2 was reported. Compared with FB23, FB23-2 has a slightly lower inhibitory activity but a significantly higher cell uptake rate [124]. Lin et al. [40] demonstrated that FB23-2 could increase the sensitivity of CRC cells to 5-FU both in vitro and in vivo by inhibiting FTO. Therefore, targeting m6A regulators with small molecules or modulating their expression represents a promising therapeutic strategy for cancer treatment and is worth of significant scientific interests.

In the treatment of certain cancers (such as oral squamous cell carcinoma), STM2457 has shown promising outcomes. The combination of STM2457 and anlotinib (targeting EGFR) can exert a multifaceted antitumor effect [125]. Despite the promising prospects of m6A-targeted drugs, their clinical translation is still confronted with significant challenges in terms of toxicity, bioavailability, and patient stratification.

The primary concern is “on-target/off-tumor” toxicity. Small-molecule inhibitors targeting m6A modification proteins may exert certain off-target effects on the enzymes with similar structures. Moreover, since m6A modification plays a fundamental role in normal physiology, such inhibitors may trigger systemic toxicity. For example, inhibition of METTL3 can impair postnatal liver regeneration in mice [126], while knockout of YTHDF2 causes hematopoietic system failure [127], indicating the risk of non-target tissue damage. Future studies should prioritize the development of tissue-specific delivery systems to mitigate such risks. In addition, utilization of well-established drugs with existing safety certifications can bypass the “de novo” development of small molecules, which is expected to accelerate clinical application.

In terms of bioavailability, although small-molecule inhibitors (such as FB23-2) exhibit robust activity in vitro, their in vivo efficacy is often compromised by poor solubility, low metabolic stability, and insufficient tumor penetration. Encouragingly, Huang et al. [128] recently proposed a novel strategy to circumvent this bottleneck. They found that homoharringtonine, a clinically approved drug, does not directly inhibit FTO enzymatic activity; instead, it promotes FTO proteasomal degradation, thereby upregulating the global m6A level in acute monocytic leukemia.

Precise patient stratification remains a critical bottleneck. The functional roles of m6A regulators are highly context-dependent, showing significant variations across different cancer types and even subtypes [129]. Currently, there is still a lack of validated biomarkers for prediction of the treatment response or drug resistance, making it difficult to identify patients who would truly benefit from m6A-targeted therapies. The treatment of CRC is developing toward greater precision and individualization, and biomarker-driven trial design has become a focus of current research. Molecular biomarkers such as MSI/dMMR, BRAF, KRAS, and HER-2 have been widely used to guide treatment decisions, while the development of liquid biopsy technologies such as ctDNA has provided new tools for CRC management [130]. Therefore, combining m6A expression profiles with these biomarkers may facilitate the clinical application of m6A modification.

Finally, in current CRC treatment strategies, single therapeutic approaches usually have certain limitations, and the optimization of combined therapies will become a future hotspot. Multiple strategies, including immunotherapy combinations, chemotherapy combined with targeted therapy, and novel ADC drugs, have shown significant efficacies [131]. For example, regarding microsatellite instability-high (MSI-H) and microsatellite stable (MSS) subtypes, a key research direction is to determine whether further combining PD-1 antibodies with chemotherapy plus bevacizumab can improve the efficacy. The complexity of the m6A regulatory network, its functional redundancy, and the dynamic evolution of the epitranscriptome under therapeutic pressure all make it necessary to develop combination therapies. For instance, YTHDF1-mediated chemoresistance is closely associated with glutamine metabolism, suggesting the potential of combining its inhibitors with GLS1 blockers. Similarly, the immunosuppressive microenvironment induced by METTL14 deficiency may be reversed through combination with PD-1 inhibitors. Nevertheless, combination regimens may also introduce new complexities regarding optimal dosage determination and additive toxicity [132].

In our view, the field of m6A research in CRC is at a critical juncture. While preclinical data are robust, clinical translation largely lags behind due to insufficient understanding of the dose-dependent effects and long-term safety. To successfully integrate m6A-targeted strategies into the therapeutic arsenal for CRC, these challenges must be addressed by developing more specific inhibitors, advanced drug delivery systems, and reliable predictive biomarkers, as well as by better understanding the biological mechanisms through which m6A regulates CRC.