RNA m⁶A modification is regulated by a cohort of enzymes that catalyze the installation, removal, recognition and function of these modifications (35,54). Writers, erasers and readers constitute the trio of components involved in m⁶A regulation (Fig. 1). The installation of m⁶A modification is carried out by a group of enzymes known as writers (55). For example, the m⁶A writer complex includes methyltransferase-like 3 (METTL3), METTL14 and Wilms tumor 1-associating protein (WTAP) among others. These enzymes collaboratively catalyze the methylation process, with METTL3 serving as the catalytic subunit (56,57). Accessory proteins such as vir-like m⁶A methyltransferase associated protein (VIRMA), RNA binding motif protein 15 (RBM15) and METTL16 further modulate the specificity and efficiency of m⁶A installation (58). Conversely, eraser enzymes such as fat mass and obesity-associated protein (FTO) and AlkB homolog 5 (ALKBH5) are responsible for demethylating m⁶A, removing the methyl group and thus reversing the modification (56,59). This dynamic installation and removal establish a regulatory mechanism that finely tunes RNA function in response to cellular conditions.

The fate of m⁶A-modified RNA is tightly regulated by various reader proteins. For instance, YTH domain-containing proteins, recognize and bind to m⁶A-modified sites to influence mRNA stability and translation efficiency (56). Specifically, YTH domain family protein 1 (YTHDF1) primarily facilitates translation by interacting with translation initiation factors to enhance the translation efficiency of m⁶A-modified mRNAs; YTHDF2 is well-known for its role in promoting mRNA decay, targeting m⁶A-modified transcripts for degradation; by contrast, YTHDF3 acts as a multifaceted adapter that can coordinate with either YTHDF1 to promote translation or with YTHDF2 to expedite mRNA decay, thereby fine-tuning the balance between mRNA stability and translation (60). YTH domain containing protein 1 (YTHDC1) is involved in the regulation of splicing, while YTHDC2 affects multiple aspects of RNA metabolism, with reported roles in splicing and translation (61). The insulin-like growth factor 2 mRNA-binding proteins, which also recognize m⁶A sites, are known to enhance the stability and storage of their target mRNAs (62).

Additionally, heterogeneous nuclear ribonucleoproteins (hnRNPs), such as heterogeneous nuclear ribonucleoprotein C/G (hnRNPC/G) and heterogeneous nuclear ribonucleoprotein A2/B1 (hnRNPA2B1), recognize m⁶A modifications and influence miRNA processing and splicing events by facilitating the recruitment of splicing factors and modulating alternative splicing patterns (63,64). Eukaryotic initiation factor 3 can bind to m⁶A sites, particularly near the 5' untranslated region (UTR) of mRNAs, to directly promote translation initiation in a cap-independent manner, enhancing the translation of specific transcripts under certain conditions (65). Fragile X mental retardation protein is another recognized m⁶A reader that plays a role in neuronal mRNA localization and stability, thereby influencing synaptic function and plasticity (66). This regulatory capacity of m⁶A modification, which spans mRNA splicing, localization, stability and translation efficiency, allows cells to finely tune gene expression in response to developmental cues and environmental stresses.

The m⁶A modification serves as a fundamental regulator of RNA metabolism, exerting precise control over gene expression and a broad range of cellular processes. It governs nearly every stage of the mRNA life cycle: in the nucleus, it determines exon inclusion during splicing and facilitates mRNA export to the cytoplasm (67-70). Once in the cytoplasm, m⁶A directly influences transcript stability and translation efficiency, thereby dictating mRNA half-life and protein synthesis levels (71,72). This multifaceted regulation is indispensable for directing critical cellular behaviors, including stem cell renewal, neurogenesis and other fate decisions, ensuring proper development and physiological function (73,74).

The dynamic nature of m⁶A is crucial for cellular adaptation to various cues. Under stress conditions such as heat shock or oxidative stress, rapid m⁶A remodeling enables selective stabilization or degradation of key transcripts, facilitating a swift response (75,76). Notably, this regulation can be recursive, as exemplified by an m⁶A-modified transcript regulating the stability of its reader, IGF2BP1, under heat stress (77). m⁶A also contributes to maintaining circadian rhythms by modulating the stability of clock gene transcripts, thereby ensuring cellular homeostasis (78,79). This capability to adapt both to intrinsic cycles and extrinsic stressors underscores m⁶A's essential role in preserving the balance and functionality of cellular environments.

The regulatory scope of m⁶A extends beyond mRNA to include non-coding RNAs (ncRNAs). It influences the biogenesis of microRNAs (miRNAs) and the function of long non-coding RNAs (lncRNAs) and circular RNAs (circRNAs), which themselves act as scaffolds, sponges, or epigenetic regulators (52,80-82). Furthermore, extensive crosstalk exists between the epitranscriptome and other epigenetic layers. m⁶A can recruit chromatin modifiers to influence histone modifications and DNA methylation, thereby impacting transcription and chromatin accessibility (45,83). It also modulates diverse RNA-RNA, RNA-DNA and RNA-protein interactions crucial for various cellular processes and affects the formation of RNA-DNA hybrids (R-loops), which have implications for transcription and genome stability (84-86).

Taken together, the m⁶A modification is a versatile and powerful regulator of gene expression. Its ability to integrate signals and coordinate complex cellular programs makes it indispensable for normal physiology. However, this same versatility means that its dysregulation can have profound consequences, directly contributing to the pathogenesis of human diseases, including cancer, by disrupting the precise control of oncogenic and tumor-suppressive pathways.

The dysregulation of RNA m⁶A modification has emerged as a critical mechanism in tumorigenesis, adding a new dimension to our understanding of gene expression control in cancer beyond traditional genetic and epigenetic alterations (87-89). By disrupting cellular homeostasis, aberrant m⁶A modification contributes markedly to tumor initiation and progression (90-92). Accumulating evidence has demonstrated the implications of RNA m⁶A modification in solid tumors and hematological malignancies. This section briefly summarizes the functional roles and clinical relevance of m⁶A dysregulation in cancer.

The sophisticated regulatory functions of m⁶A in normal cellular processes are frequently subverted in oncogenesis (49,93). By controlling mRNA stability, splicing, localization and translation, m⁶A directly influences the expression of oncogenes and tumor suppressor genes, thereby driving cancer hallmarks such as uncontrolled proliferation, invasion and metastasis (94-96). This is mechanistically demonstrated in diverse types of cancer. For instance, METTL3 promotes the epithelial-mesenchymal transition by adding m⁶A marks to Forkhead box protein O1 (FOXO1) mRNA, leading to its YTHDF2-mediated decay (97). Another study identified Mettl3 as a crucial regulator for transitioning murine stem cells from the naive to the primed pluripotent state by promoting mRNA degradation of key pluripotency factors, linking cancer stem cell features to m⁶A modification (98). In hepatocellular carcinoma, WTAP facilitates disease progression via an m⁶A-dependent mechanism that epigenetically silences the tumor regulator erythroblast transformation specific 1 (ETS1) mRNA (99). Conversely, ALKBH5 maintains glioblastoma stem-like cells by demethylating and stabilizing FOXM1 transcripts (100). Loss of YTHDF2 enhances macrophage recruitment and mitochondrial respiration of CD8⁺ T cells, ultimately inhibiting the growth of melanoma tumor cells and improving survival in immunocompetent models (101). Wang et a (102) demonstrated that m⁶A-mediated control of mRNA translation influences cancer-relevant gene expression, emphasizing its role in oncogenesis. The examples from various other solid tumors that are extensively reviewed elsewhere, underscore that m⁶A modifiers can function as context-dependent oncogenes or tumor suppressors (103,104).

Similarly, m⁶A dysregulation is a recognized hallmark of hematological malignancies, as detailed in other specialized reviews (34,105). It disrupts normal hematopoiesis and immune surveillance to promote leukemogenesis and progression (34,105,106). The roles of m⁶A regulators are often complex and dualistic. For example, FTO has been demonstrated to promote the development of acute myeloid leukemia (AML) by reducing m⁶A levels on transcripts, including ASB2 and RARA, underscoring the oncogenic potential of FTO in AML (107). Conversely, other studies showed that METTL3, a core component of the m⁶A writer complex, controls myeloid differentiation in both normal hematopoietic and leukemia cells (108,109). Depletion of METTL3 promotes cell differentiation and reduces cell proliferation, indicating a tumor-suppressive role of m⁶A modulation in this context (108). IGF2BP3 is overexpressed in AML and drives disease progression by binding to and stabilizing RCC2 transcripts in an m⁶A-dependent manner (110). These studies illustrate the dual and context-dependent roles of m⁶A modifications in hematopoiesis and leukemia.

Given this paradigm of m⁶A's importance in oncology and hematology, its investigation in MM is a logical and critical frontier. The ability of m⁶A modifications to alter the expression of critical oncogenes and tumor suppressors, as demonstrated in other cancers, strongly positions it as a key player in MM pathogenesis, progression and drug resistance. The following sections will focus exclusively on synthesizing the growing evidence for m⁶A's role in this specific hematological cancer.

The writer complex, responsible for installing m⁶A marks, is frequently upregulated in MM. Pre-clinical studies predominantly report oncogenic functions for these writers, though clinical validation is ongoing. For instance, studies indicate METTL3 enhances the maturation of oncogenic miR-182-5p via m⁶A modification, leading to downregulation of the proliferation inhibitor calcium/calmodulin-dependent protein kinase II inhibitor 1 (CAMK2N1) (Fig. 2A) (114). In a separate proposed axis, METTL3 stabilizes Yin Yang 1 (YY1) mRNA and promotes the maturation of primary-miR-27a-3p, suggesting a potential feedback loop that drives proliferation and stemness (Fig. 2B) (115). Furthermore, METTL3 upregulates the translation initiation regulator basic leucine zipper and W2 domains 2 (BZW2) through m⁶A methylation, accelerating MM cell proliferation and inhibiting apoptosis (Fig. 2C) (116). In addition, METTL3 collaborates with the oncogenic lncRNA metastasis associated lung adenocarcinoma transcript 1 (MALAT1) to promote MM cell proliferation and MALAT1 overexpression rescues the effects of METTL3 knockdown (117). The potential oncogenic role of METTL3 is further supported by evidence that the antidiabetic drug metformin exerts anti-myeloma effects partly by impairing METTL3-mediated m⁶A modification of genes such as thyroid hormone receptor associated protein 3, RNA binding motif protein 25 and ubiquitin specific peptidase 4 (118-120). It is important to note that these METTL3-dependent axes are primarily supported by individual in vitro and xenograft studies. Their relative contributions and clinical relevance in heterogeneous MM patient populations require independent validation and further investigation.

Other writers also contribute to MM pathogenesis in pre-clinical contexts. For example, METTL5 has been shown to drive MM progression by enhancing the translation of selenoproteins such as selenophosphate synthetase 2 (SEPHS2), which is critical for mitigating oxidative stress (121). The accessory writer protein VIRMA (KIAA1429) is overexpressed in MM and experimental evidence suggests it promotes tumorigenesis by enhancing m⁶A modification and expression of the oncogene forkhead box M1 (FOXM1) via the reader YTHDF1 (122). Beyond regulating oncogene expression, VIRMA has been linked to suppression of ferroptosis to promote MM cell survival (123). Mechanistically, VIRMA is stabilized by the lncRNA FEZF1-AS1 and, in turn, enhances m⁶A-dependent translation of OTU deubiquitinase, ubiquitin aldehyde binding 1 (OTUB1). OTUB1 then deubiquitinates and stabilizes the ferroptosis defense protein solute carrier family 7 member 11 (SLC7A11), protecting MM cells. The core writer complex component WTAP also appears to function as a critical oncogene. A recent study identified WTAP as overexpressed in MM patients and associated with poor survival (124). Mechanistically, this study reported that WTAP installs m⁶A modifications on microtubule-associated protein 6 domain containing 1 mRNA, thereby regulating the Hippo signaling pathway to drive MM cell proliferation. Notably, the writer complex can be regulated post-translationally; research indicates protein arginine methyltransferase 1 methylates and stabilizes WTAP, which in turn enhances m⁶A modification of NADH:ubiquinone oxidoreductase core subunit S6 mRNA, boosting its expression and activating oxidative phosphorylation to promote tumorigenesis (125).

Collectively, while these studies identify promising oncogenic roles for m⁶A writers in MM, the current mechanistic understanding is predominantly derived from in vitro and mouse xenograft models. Independent replication and validation in primary patient samples are essential to establish their significance and prioritize translational efforts.

The functional output of m⁶A methylation is largely determined by reader proteins, whose dysregulation in MM orchestrates diverse oncogenic programs. A pivotal player is YTHDF2, whose elevated expression correlates with poor patient prognosis (126). Loss- and gain-of-function experiments established a causal role for YTHDF2 in driving MM proliferation both in vitro and in vivo. Mechanistic insights revealed that YTHDF2 recognizes m⁶A motifs on the signal transducer and activator of transcription 5A (STAT5A) transcript, promoting its degradation (126). Subsequent rescue experiments confirmed that STAT5A acts as a tumor suppressor by attenuating ERK phosphorylation via interaction with the mitogen-activated protein kinase kinase 2 (MAP2K2) promoter, thereby delineating the YTHDF2-STAT5A-MAP2K2-p-ERK axis as a key proliferative driver (Fig. 3A) (126). Parallel investigations established that YTHDF2 promotes G₁/S phase transition of MM cells by degrading the tumor suppressor early growth response 1 (EGR1). This model is solidified by the finding that EGR1 knockdown abrogates the cytostatic effect of YTHDF2 depletion, as EGR1 transactivates p21 and represses the CDK2/cyclin E1 complex (Fig. 3B) (127).

The oncogenic role of readers extends to genetically defined MM subgroups. IGF2BP1 is markedly overexpressed in patients with chromosome 1q gain, where it predicts an inferior clinical outcome (128). Mechanistically, IGF2BP1 directly binds to m⁶A sites within the cell division cycle 5-Like (CDC5L) mRNA, enhancing its stability and translation. The functional dependency of this axis was confirmed by the fact that genetic or pharmacological inhibition of IGF2BP1 suppressed the pro-tumor effects, revealing the IGF2BP1-CDC5L axis as a vulnerability in this high-risk population (128).

A paramount clinical challenge in MM is the bone disease and m⁶A readers are critically involved in this destructive process (129,130). The reader hnRNPA2B1 is a key orchestrator of myeloma bone disease, as its expression correlates with the number of osteolytic lesions. hnRNPA2B1 drives bone destruction via a dual exosomal miRNA mechanism. It promotes the maturation and exosomal packaging of specific miRNAs, such as miR-92a-2-5p and miR-373-3p, in an m⁶A-dependent manner. Upon delivery to the bone marrow niche, these miRNAs disrupt bone homeostasis: miR-92a-2-5p suppresses the osteoclast inhibitor IRF8 in monocytes/macrophages, enhancing osteoclastogenesis and bone resorption, while miR-373-3p inhibits the master osteogenic transcription factor RUNX2 in mesenchymal stem cells, impairing bone formation (131). Cell-autonomously, HNRNPA2B1 also enhances MM cell survival by stabilizing interleukin enhancer binding factor 3 mRNA, leading to AKT3 activation, a pathway corroborated by positive immunohistochemical correlations in patient samples (132). Another study revealed that hnRNPA2B1 sustains the activity of the central kinase CK2 and CK2 inhibition induces ER stress and autophagy to suppress proliferation and promote apoptosis in hnRNPA2B1-overexpressing MM cells, identifying CK2 as a critical downstream effector (133).

The clinical relevance of m⁶A readers is further highlighted by the overexpression of leucine rich pentatricopeptide repeat containing, another m⁶A-binding protein, in advanced-stage MM, where its knockdown induces apoptosis and cell cycle arrest, positioning it as a potential therapeutic node (134). The pro-tumorigenic functions of m6A readers such as YTHDF2, IGF2BP1 and HNRNPA2B1 have been confirmed in multiple preclinical studies through genetic manipulations and mechanistic experiments and are associated with patient prognosis. However, current understanding heavily relies on a limited set of cell line models and preclinical animal studies. To establish these readers as reliable clinical therapeutic targets, further validation in more diverse MM models is required.

The erasers ALKBH5 and FTO, which mediate m⁶A demethylation, are frequently overexpressed in MM. Pre-clinical evidence suggests they may function as oncogenes by stabilizing a network of transcripts crucial for survival, stemness and stress adaptation. ALKBH5 promotes tumorigenesis through multiple, distinct mechanisms. It stabilizes the lncRNA small nucleolar RNA host gene 15 (SNHG15) in an m⁶A-dependent manner (135). Studies using cell line models indicate that this ALKBH5-SNHG15 axis subsequently facilitates the expression of the histone methyltransferase SET domain containing 2 (SETD2), altering the H3K36me3 landscape to enhance chromatin accessibility and transcriptional elongation (Fig. 4A). In a separate pathway, integrated MeRIP-seq and functional analyses from one study identified Salvador family WW domain containing protein 1 (SAV1), a core component of the HIPPO tumor suppressor pathway, as a key ALKBH5 target. ALKBH5-mediated demethylation destabilizes SAV1 mRNA, leading to Yes-associated protein (YAP) activation and MM progression (136). Consistent with effects on cellular plasticity, ALKBH5 depletion suppressed the MM stem cell-like phenotype, as evidenced by reduced expression of pluripotency factors (NANOG, SOX2 and OCT4), linking epitranscriptomic regulation to cellular plasticity (Fig. 4B). Mechanistically, suppression of SAV1 perturbs Hippo signaling and enhances YAP activity and ALKBH5 depletion exerted an anti-myeloma effect by promoting expression of p21, p53 upregulated modulator of apoptosis and Bcl2 associated X (BAX) (136). Furthermore, another mechanistic report suggests ALKBH5 drives oncogenic signaling by reducing m⁶A levels on the 3'UTR of TNF receptor associated factor 1 (TRAF1) mRNA, thereby enhancing its stability and activating both NF-κB and MAPK pathways to promote growth and survival (Fig. 4C) (137). These findings from pre-clinical studies suggest that ALKBH5 functions as an oncogene in MM and could serve as a potential biomarker and therapeutic target.

The other major eraser, FTO, exerts its oncogenic role by establishing a hypomethylated transcriptome (138). FTO is upregulated in MM and promotes tumor growth and metastasis. It specifically demethylates heat shock factor 1 (HSF1) mRNA, shielding it from YTHDF2-mediated decay and thus augmenting the heat shock response, a critical survival pathway under proteotoxic stress (139). Importantly, FTO activity is metabolically coupled. Isocitrate dehydrogenase 2 (IDH2), which generates the essential FTO cofactor α-ketoglutarate, is highly expressed in progressive MM. IDH2 depletion increased global m⁶A levels and impaired growth, mechanistically through suppressing FTO-mediated hypomethylation and stabilization of Wnt family member 7B mRNA, which activates pro-tumorigenic Wnt signaling (140,141). The translational promise of targeting m⁶A erasers is underscored by the finding that FTO inhibition synergizes with the frontline therapeutic drug bortezomib to suppress myeloma growth in vivo, presenting a compelling combinatorial strategy (139).

Taken together, the ALKBH5-SNHG15-SETD2, ALKBH5-SAV1-YAP and ALKBH5-TRAF1 axes, along with FTO-mediated regulation of HSF1 and Wnt signaling, represent important emerging mechanistic insights. However, the current evidence for these oncogenic roles of m⁶A erasers is predominantly derived from in vitro and mouse xenograft models. Their overarching significance, interdependence and clinical relevance as therapeutic targets in primary MM require independent replication and further validation in patient cohorts.

The epitranscriptomic landscape of MM extends beyond m⁶A to include other chemically distinct modifications, such as N4-acetylcytidine (ac⁴C), 5-methylcytidine (m⁵C) and N1-methyladenosine (m¹A), which are increasingly recognized as potential contributors to disease pathology. The ac⁴C writer N-acetyltransferase 10 (NAT10) is highly expressed in MM and associated with poor prognosis (142). NAT10 promotes MM cell proliferation by acetylating and stabilizing B-cell lymphoma-extra-large (BCL-XL) mRNA, a key anti-apoptotic regulator, thereby enhancing cell survival through the PI3K-AKT pathway. The dependency on this axis was confirmed by the potent anti-myeloma activity of Remodelin, a specific NAT10 inhibitor, which induced apoptosis both in vitro and in vivo (142). A separate study further showed that NAT10 also acetylates and stabilizes G protein-coupled receptor 37 (GPR37) mRNA (143). This NAT10-GPR37 axis drives MM progression by promoting cell cycle progression, glycolysis and immune evasion, while inhibiting apoptosis, effects that were recapitulated in a xenograft model.

Similarly, the role of m⁵C modification is being unraveled through bioinformatics and immunological profiling. Studies mapping the m⁵C regulome have identified distinct patient clusters with unique immune microenvironment signatures and metabolic features (144). A risk model based on m⁵C-related genes effectively stratified patients and high m⁵C scores were associated with specific immune cell infiltration patterns, suggesting that m⁵C modifications contribute to the immunosuppressive niche in MM (144). Importantly, the primary m⁵C methyltransferase NSUN2 is overexpressed in MM and correlates with poor prognosis (145). Functional studies demonstrate that NSUN2 drives MM progression by installing m⁵C modifications on huntingtin interacting protein 1 mRNA, thereby enhancing its stability and promoting tumor cell proliferation both in vitro and in vivo.

The m¹A modification also holds significant prognostic and functional relevance in MM. Systematic analysis has identified specific m¹A regulators (such as TRMT61A/B and YTHDF1-3) that are dysregulated in MM and are associated with patient outcomes (146). Notably, high expression of the m¹A reader YTHDF2 was strongly associated with poor survival and was shown to promote MM cell proliferation while inhibiting apoptosis. Furthermore, YTHDF2 overexpression increased global m¹A levels and its co-expressed partner SRSF10 was identified as a potential downstream effector (146). This positions the m¹A machinery as a novel layer of epitranscriptomic regulation in MM pathogenesis.

Collectively, these findings from recent pre-clinical and bioinformatic studies illuminate an expanding epitranscriptomic network in MM. While m⁶A remains the most extensively characterized, the reported contributions of ac⁴C and m⁵C underscore the broader regulatory potential of RNA modifications. Targeting these pathways represents a promising but early-stage frontier for developing novel therapeutic strategies aimed at overcoming drug resistance and improving patient outcomes.

Bioinformatics analyses have suggested a link between the m⁶A machinery and the immunosuppressive tumor microenvironment in MM. For example, a study reported that a risk signature based on high expression of HNRNPC, HNRNPA2B1 and YTHDF2, coupled with low ZC3H13 expression, was associated with poor prognosis (154). Patients stratified into this high-risk group showed enrichment for hallmark MM pathways (such as MYC signaling and unfolded protein response) and, importantly, a markedly depleted immune landscape. This 'immune-cold' phenotype was characterized by a broad reduction in infiltrating immune cells, including both effector (such as CD8⁺ T cells and NK cells) and suppressive subsets (such as myeloid-derived suppressor cells and regulatory T cells) and by downregulation of HLA class II molecules (such as HLA-DMA, HLA-DMB, HLA-DPA1 and HLA-DQB1) on tumor cells, potentially impairing antigen presentation and the initiation of adaptive immune responses (154). While this suggests a role for m⁶A in immune evasion, the underlying mechanisms and the clinical utility of this signature require prospective validation.

Functional studies in selected models implicate multiple m⁶A modifiers as direct mediators of chemoresistance. A study by Wang et al (155) provides a mechanistic example of how m⁶A dysregulation directly confers chemoresistance in MM. The researchers established a bortezomib-resistant model in vitro by chronically exposing the human MM cell line RPMI 8226 to progressively increasing concentrations of the drug. Comparative analysis revealed that the resistant cells exhibited significant upregulation of the m⁶A eraser FTO alongside concomitant downregulation of the antioxidant enzyme SOD2 (Fig. 5). The investigators proposed a direct causal relationship: FTO demethylates m⁶A sites on SOD2 mRNA, thereby promoting its decay. In this model system, ectopic overexpression of FTO in sensitive cells induced a resistant phenotype, while genetic knockdown of FTO or exogenous restoration of SOD2 resensitized resistant cells. Pharmacological inhibition of FTO with FB23-2 reversed resistance in vitro and in vivo (155), suggesting the FTO-SOD2 axis as a potential therapeutic target. This finding presents a seemingly paradoxical phenomenon, as diminished antioxidant capacity would typically be expected to exacerbate bortezomib-induced oxidative stress and cell death (156,157). The study attributed this to adaptive cellular rewiring, where chronically low SOD2 drives compensatory upregulation of pro-survival signaling networks (such as NF-κB and MAPK) (155). This model aligns with prior observations that alterations in oxidative stress management correlate with therapeutic response (158-160).

Notably, the regulation of oxidative stress in MM reveals a complex picture. On one hand, METTL5 upregulates SEPHS2 expression, thereby enhances antioxidant defense (121). By contrast, FTO reduces SOD2 expression, diminishing antioxidant defense yet being reported to contribute to bortezomib resistance (155). This paradox may be explained by the dual role of reactive oxygen species (ROS), which at optimal levels can drive pro-tumorigenic signaling (161,162). Thus, cancer cells may manipulate different antioxidant pathways to maintain ROS within a survival-favorable range. Future studies are warranted to validate this effect and to determine how redox homeostasis can be therapeutically targeted (163,164).

ALKBH5 has also been implicated in resistance to targeted therapies. A recent study demonstrated that ALKBH5 overexpression attenuates the efficacy of the histone deacetylase inhibitor romidepsin by reducing m⁶A methylation on FOXM1 mRNA, thereby stabilizing it (165). Conversely, romidepsin treatment or ALKBH5 knockout increases FOXM1 m⁶A modification, accelerating its decay and synergistically enhancing apoptosis (165). This positions ALKBH5 as a key modulator of drug response in pre-clinical models.

Drug resistance is further proposed to be mediated through m⁶A-dependent regulation of key signaling hubs and non-coding RNAs. The reader protein HNRNPA2B1 binds to m⁶A sites on Toll-like receptor 4 (TLR4) mRNA, enhancing its stability and expression (166). Given that TLR4 signaling promotes a pro-survival and resistant state in MM, this HNRNPA2B1-TLR4 axis represents one mechanism that may contribute to microenvironment-associated resistance (167,168). Similarly, circular RNAs are co-opted in model systems; circ_0000337 is upregulated in bortezomib-resistant cells, where it sponges miR-98-5p, leading to increased expression of the DNA repair protein DNA2. The stability of circ_0000337 itself is regulated by m⁶A modification, creating a positive feedback loop that enhances DNA repair capacity (169). In another layer, METTL3 installs m⁶A marks on the lncRNA H19, which sequesters miR-184, relieving the suppression of the oncogenic coactivator-associated arginine methyltransferase 1 (CARM1) and driving a bortezomib-resistant phenotype in experimental models (170).

This concept extends to predicting response to combination therapies. In a cohort of patients relapsing after bortezomib-based regimens, a high m⁶A risk score (with an Area Under the Curve of 0.9) was associated with non-response to salvage therapy with daratumumab, carfilzomib, lenalidomide and dexamethasone (DARA-KRD) (113). Although mechanisms such as FTO-SOD2 have been demonstrated to mediate resistance in preclinical models, these findings are primarily based on single studies or limited cohorts. The prevalence of these pathways in patient populations, particularly among those with different types of resistance and their interactions with classical resistance mechanisms remain insufficiently elucidated and warrant validation in large-scale, prospective clinical cohorts.

The bone marrow adipocyte niche has emerged as a proposed source of exosome-mediated transfer of chemoresistance, a process suggested to be governed by epitranscriptomic regulation. Adipocyte-derived exosomes are reported to enrich MM cells with the lncRNAs LOC606724 and SNHG1, which inhibit apoptosis (171). Furthermore, the methyltransferase METTL7A, whose activity is potentiated by EZH2 in these models, promotes the m⁶A modification of these lncRNAs, thereby enhancing their packaging into exosomes. Upon delivery to MM cells, these m⁶A-modified lncRNAs drive c-Myc expression (171). This paradigm highlights a potential mechanism of niche-induced drug resistance that warrants further investigation.

Overall, preclinical studies suggest an interplay between m⁶A modification, non-coding RNA networks and the tumor microenvironment may create a framework for drug resistance in MM. Targeting key nodes within this epitranscriptomic network may offer future opportunities for resensitizing refractory MM, pending further validation of their clinical relevance.

A systematic approach is required to navigate the landscape of m⁶A regulators and identify the most promising candidates for drug discovery in MM. The present review proposed a multi-dimensional assessment framework based on the following criteria: First, the strength of MM-specific preclinical validation. This encompasses evidence from genetic dependency screens (such as CRISPR/Cas9 knockout) in MM cell lines, demonstration of robust anti-tumor efficacy in MM-specific in vivo models and clear correlation with poor prognosis, drug resistance, or key MM hallmarks in patient cohorts. Second, chemical tractability and drug development stage. This criterion assesses the availability of selective, potent and bioavailable small-molecule inhibitors, PROTAC degraders, or other pharmacological modalities targeting the m⁶A regulator. Third, the potential for integration with existing MM therapies. This evaluates whether preclinical models demonstrate synergistic effects when combining modulation of the m⁶A target with backbone MM therapies, such as proteasome inhibitors (bortezomib), immunomodulatory drugs (lenalidomide), or monoclonal antibodies (daratumumab). Targets that address core resistance mechanisms (such as FTO in bortezomib resistance) or modulate the tumor microenvironment to enhance therapy are considered particularly attractive.

Targeting the catalytic core of m⁶A regulatory enzymes with small-molecule inhibitors represents the most direct path for clinical translation (Table I) (128,152,173-198). Applying the proposed translational framework, several targets emerge with high priority.

The most compelling current evidence points to targeting the m⁶A writer METTL3. Its MM-specific validation is robust, as knockdown consistently inhibits the proliferation and survival of MM cells. Furthermore, metformin's partial anti-myeloma effect, mediated through the impairment of METTL3 function, provides indirect translational support (117-119). Regarding chemical tractability, the METTL3 inhibitor STM2457 is currently in Phase I clinical trials for AML. This agent has demonstrated potent anti-leukemic effects by reducing global m6A levels and downregulating key oncogenic transcripts such as BRD4 and SP1 (184). In terms of potential for therapeutic integration, given METTL3's well-defined oncogenic role in MM, where it drives proliferation by stabilizing transcripts such as YY1 and BZW2 and promoting the maturation of miR-182-5p, its inhibitors are hypothesized to synergize with proteasome inhibitors, which induce proteotoxic and endoplasmic reticulum stress (115,116). Therefore, evaluating METTL3 inhibitors (such as STM2457, UZH1) in MM models, particularly their efficacy in combination with bortezomib, is a high-priority preclinical direction.

FTO is another high-priority target. MM-specific evidence shows that FTO is upregulated in patients and promotes bortezomib resistance through mechanisms including the disruption of the oxidative stress modulator SOD2 (Fig. 5) (155). The chemical tractability of FTO is supported by several inhibitors (such as FB23-2, CS1/CS2), which have demonstrated anti-tumor effects in leukemia models (173,176-178). The potential for integration is highly compelling. Synthetic small-molecule inhibitors like FB23-2 can selectively block FTO activity, leading to enhanced decay of oncogenic transcripts such as MYC. Accordingly, FTO inhibition warrants MM-specific validation as a resensitization strategy, including testing in acquired bortezomib-resistant models and in combination regimens with proteasome inhibitors.

The eraser ALKBH5 also presents a viable target. Its MM-specific validation is underscored by its role in driving tumorigenesis and mediating resistance to agents such as the histone deacetylase inhibitor romidepsin by stabilizing FOXM1 mRNA (165). However, available ALKBH5 inhibitors (such as IOX1) are generally considered tool compounds and improved potency/selectivity and MM-specific pharmacological validation will be required. Modulating ALKBH5 may still be attractive for combination strategies if on-target activity and tolerability can be established (194).

Although the development of inhibitors targeting m⁶A readers is still in its early stages, it offers unique opportunities for precision therapy. For instance, targeting IGF2BP1, which stabilizes CDC5L mRNA in high-risk 1q-amplified MM (128); inhibiting HNRNPA2B1, a driver of bone disease and survival signaling (131-133); or blocking YTHDF2, which promotes proliferation by degrading tumor suppressors such as STAT5A and EGR1 (126,127) could each provide a highly targeted strategy with a potentially wider therapeutic window. Nevertheless, reader-targeting chemical matter remains relatively immature and robust pharmacodynamic assays and MM-specific validation will be essential to assess clinical feasibility.

Collectively, the inhibitors for METTL3 and FTO possess the strongest immediate translational rationale. However, it is critical to note that the current assessment of their efficacy largely draws on research from other hematologic malignancies such as AML. Specific pharmacodynamic data in MM, optimal combination regimens and potential resistance mechanisms await elucidation through future systematic preclinical studies dedicated to MM.

Beyond small molecules, genetic and targeted protein degradation technologies offer alternative paths for precise and potent modulation of the m⁶A machinery (Table II) (199-209). CRISPR-Cas9-mediated knockout of METTL3 or ALKBH5 has proven effective in impairing cancer progression in various models, though in vivo delivery to MM cells in the bone marrow remains a challenge (194,210). More sophisticated epigenetic editing tools, such as dCas13b-ALKBH5 fusion systems, enable site-specific m⁶A demethylation without altering the DNA sequence, offering the potential to reversibly silence specific driver genes such as IRF4 while sparing normal hematopoiesis (208,211).

Proteolysis-targeting chimeras (PROTACs) represent a groundbreaking modality to achieve complete and selective degradation of target proteins. PROTACs such as WD6305, which recruit E3 ubiquitin ligases to degrade METTL3, have demonstrated potent anti-leukemic effects in vivo (199,212,213). This strategy could be particularly effective against MM cells that rely on METTL3 for survival, potentially overcoming resistance mechanisms that arise with catalytic inhibitors. Similarly, antisense oligonucleotides (ASOs) designed to silence METTL3 mRNA expression provide high specificity and could mitigate off-target effects associated with small molecules (214).

These next-generation approaches are poised to address the issues of tumor heterogeneity and acquired resistance that often plague conventional therapies. The future clinical application of these modalities will hinge on overcoming the critical challenge of delivery, efficiently and specifically targeting MM cells within the bone marrow niche. Advances in lipid nanoparticles, viral vectors, or cell-specific targeting moieties will be essential to unlock the full potential of genetic and degradation-based epitranscriptomic therapy for MM.

Notably, for targets such as ALKBH5 and YTHDF2, the rationale for targeted therapies primarily relies on genetic approaches (such as knockdown/knockout experiments) and the efficacy of tool compounds in non-MM models. Conclusive evidence of 'chemical tractability', such as highly selective and potent lead compounds, remains lacking in the context of MM. Furthermore, direct single-cell epitranscriptomic mapping of m⁶A marks (rather than regulator expression) in MM subtypes remains limited, representing a key technical and conceptual gap.

Collectively, the targeting of the m⁶A epitranscriptome opens a new frontier in MM therapeutics. The diversity of druggable targets, from writers and erasers to readers, provides multiple avenues to disrupt the post-transcriptional networks that sustain this disease. The immediate challenge lies in advancing the most promising agents, particularly METTL3 inhibitors, into robust preclinical validation in MM-specific models. Success in this endeavor promises to move beyond sequential therapy toward rational, biomarker-driven combinations that anticipate and delay resistance mechanisms. By precisely manipulating the RNA regulatory code, we may ultimately shift the treatment paradigm from managing relapse to achieving durable remissions in MM.

The translational potential of targeting the m⁶A epitranscriptome must be balanced against the risk of on-target toxicities, as its core regulators (such as METTL3, FTO and ALKBH5) play indispensable roles in normal hematopoiesis and immune function (34,105,106). Mitigating potential adverse effects, such as bone marrow suppression, is therefore paramount. A viable path forward hinges on two complementary strategies: The development of highly selective inhibitors to minimize off-target effects and the pursuit of rational combination therapies. Combining an m⁶A-targeted agent with a backbone MM therapy could allow for lower, more tolerable doses of each drug while achieving synergistic efficacy. Ultimately, the clinical success of this approach will depend on integrating predictive m⁶A-based biomarkers to identify patients whose tumors are most dependent on these pathways, thereby maximizing the therapeutic window.