Abstract

Hidden in plain sight, epitranscriptomics, the study of chemical marks that sculpt RNA function, has emerged as a powerful driver of cancer biology and a rapidly expanding druggable landscape. Among the more than 170 RNA modifications identified to date, N6-methyladenosine (m⁶A) is the most abundant internal modification in eukaryotic mRNA, dynamically influencing transcript stability, translation, and subcellular localization. The machinery responsible for m⁶A includes methyltransferase complexes (“writers”), demethylases (“erasers”), and RNA-binding proteins (“readers”), which collectively determine RNA fate. Recent findings indicate that alterations in m⁶A levels or dysregulation of these regulatory proteins are linked to cancer pathogenesis and impact key cancer hallmarks, including proliferation, metastasis, metabolic reprogramming, and stemness. Importantly, the effect of m⁶A in cancer is strongly context-dependent: the same regulators, such as METTL3, METTL14, ALKBH5, and YTHDF2, may act as oncogenes or tumor suppressors depending on tumor type and cellular environment. This dual behavior underscores both the complexity of m⁶A biology and its potential as a therapeutic target. Early translational efforts, including the development of the METTL3 inhibitor STC-15, are promising but emphasize the urgent need for deeper mechanistic understanding.

Content

Why look beyond DNA?
Gene expression is not a simple, unidirectional process from DNA to protein. While the central dogma provides a foundational framework, DNA is transcribed into RNA, which is then translated into protein, cellular regulation is far more flexible. Each step represents a potential control point, determining when, how, and how much of a gene product is produced in response to cellular needs.
Post-transcriptional modifications occur in a defined temporal sequence that shapes gene expression outcomes, often independently of transcriptional control. Among these, chemical modifications of RNA molecules, collectively studied under epitranscriptomics, play a central role. Epitranscriptomics refers to reversible and dynamic biochemical modifications that influence RNA structure, processing, and function without altering the underlying nucleotide sequence.
Of the >170 chemical modifications described on RNA, N6-methyladenosine (m⁶A) is the most abundant internal mark on eukaryotic mRNA ¹. It consists of a methyl group added to the nitrogen at position six of adenosine residues, preferentially occurring within the RRACH motif (R = A/G; H = A/C/U). m⁶A marks are not randomly distributed but are enriched in defined transcript regions, such as near stop codons and within long exons. Transcriptome-wide studies reveal that m⁶A sites are evolutionarily conserved yet dynamically regulated. Their distribution is highly plastic, varying with cell type, developmental stage, and environmental stimuli. This adaptability allows m⁶A to fine-tune gene expression by modulating nuclear export, RNA stability, splicing, and translation. For a detailed historical and biochemical account of m⁶A, including discovery, consensus motifs, catalytic mechanisms, and detection methods, see reference ².
Disruptions in m⁶A modification patterns or in the activity of associated proteins, including writers (methyltransferases), erasers (demethylases), and readers (binding proteins), have been linked to cancer. Depending on the biological context, the same m⁶A machinery may function as an oncogene in one setting and as a tumor suppressor in another, highlighting its complexity and relevance in cancer biology.
This mini-review provides an overview of m⁶A regulatory machinery and its association with diverse cancer types, emphasizing its context-dependent roles. Finally, current strategies to target m⁶A modifications and recent therapeutic findings are summarized.

Who adds, removes, and reads the mark?
The m⁶A epitranscriptomic machinery can be categorized into three main functional groups: writers, erasers, and readers. Writers deposit m⁶A marks, erasers remove them, and readers interpret them by interacting with downstream effectors to influence RNA fate.
In mammalian cells, most m⁶A marks on mRNA are installed co-transcriptionally by a multi-subunit methyltransferase complex. The central components of this writer complex are the methyltransferase-like 3 and methyltransferase-like 14 heterodimer METTL3–METTL14 ³. METTL3 is the catalytic subunit that binds the methyl donor S-adenosyl methionine (SAM), while METTL14 stabilizes the complex and enhances substrate recognition.
The complex also includes Wilms tumor 1-associated protein WTAP ⁴, which is required for proper nuclear localization and recruitment of target RNAs. Additional associated proteins, such as METTL16 ⁵ and METTL5 ⁶, guide the complex to specific mRNA and rRNA sequences, particularly near 3′ untranslated regions and stop codons, modulating transcript stability, translation, and localization.
Currently, only two demethylases are known to remove m⁶A from RNA: fat mass and obesity-associated protein FTO and AlkB homolog 5 ALKBH5, both Fe²⁺/α-ketoglutarate-dependent dioxygenases. These enzymes catalyze oxidative demethylation, rendering m⁶A modification reversible and tightly regulated ^7,8.
After methylation, m⁶A-marked RNAs are recognized by specific reader proteins, which can be classified into three main groups. The first group comprises YTH domain-containing proteins, including YTHDF1, YTHDF2, YTHDF3, and YTHDC1/YTHDC2, which bind m⁶A sites directly ^9,10. The second group includes the insulin-like growth factor 2 mRNA-binding proteins IGF2BP1, IGF2BP2, IGF2BP3, which use KH domains to recognize m⁶A and generally stabilize RNA transcripts ¹¹. The third group consists of heterogeneous nuclear ribonucleoproteins (hnRNPs), such as HNRNPA2B1, HNRNPC, and RBMX, which recognize structural changes in RNA induced by m⁶A (the “m⁶A switch”) rather than the methyl group itself, regulating splicing, nuclear export, and translation ¹².
Overall, m⁶A acts as a central regulator of post-transcriptional gene expression, dynamically modulating RNA processing, export, translation, and degradation. This enables cells to fine-tune gene output in response to changing physiological and pathological contexts. For additional mechanistic insights, see reference ¹³.

Figure 1. Epitranscriptomic regulation by m⁶A writers, erasers, and readers.
Left panel. Chemical structure of N⁶-methyladenosine (m⁶A), highlighting the methyl group at the N⁶ position. During transcription, writer complexes deposit this mark: the catalytic core consists of the METTL3–METTL14 heterodimer, supported by WTAP, while specialized methyltransferases such as METTL16 and METTL5 target specific RNA families. Removal of the modification is driven by two Fe2+/α-ketoglutarate-dependent dioxygenases, FTO and ALKBH5, commonly termed erasers. A third group, the readers, recognizes the modified base: these include YTH-domain proteins (YTHDF1, YTHDF2, YTHDF3), the IGF2BP family (IGF2BP1, IGF2BP2, IGF2BP3), and several heterogeneous nuclear ribonucleoproteins (hnRNPs). Right panel. Once a reader docks on m⁶A, the transcript's functional fate is determined. Upward and downward arrows indicate stimulatory and inhibitory effects, respectively. Collectively, the dynamic interplay among writers, erasers, and readers fine-tunes RNA stability, localization, and translation efficiency, providing an agile epitranscriptomic layer for gene-expression control. Mechanistic assignments are adapted from reference ¹³.

Can context-specific m⁶A regulation explain conflicting phenotypes in cancer types?
The dysregulation of the machinery responsible for m⁶A modification is strongly associated with tumor development and affects multiple hallmarks of cancer, including uncontrolled proliferation, invasiveness, metabolic reprogramming, stemness, and interactions with the microenvironment. However, the functional role of m⁶A modifications cannot be generalized: the mechanisms of action of writers, erasers, and readers are highly context-dependent. Therefore, it is not possible to draw a clear boundary between the oncogenic and tumor-suppressive effects of a specific alteration. Evidence from the literature shows that both increases and reductions in global m⁶A levels can lead to either protumorigenic or antitumor effects, depending on the biological context.
For instance, acute myeloid leukemia (AML) has been extensively studied. Overexpression of METTL3 promotes leukemogenesis by increasing m⁶A methylation along the coding regions of mRNAs and enhancing translation via the removal of ribosomal stalling ¹⁴. This leads to the upregulation of key oncogenes, including SP1, a critical regulator of MYC expression ¹⁴. The oncogenic activity of METTL3 integrates with other driver mutations, such as MYC, BCL2, and PTEN, whose translation is sustained in an m⁶A-dependent manner ¹⁵. Conversely, depletion of METTL3 induces cell cycle arrest, differentiation, apoptosis, and heightened PI3K/AKT pathway activation, reflecting a reversal of the leukemic phenotype ¹⁵. Functional experiments demonstrate that overexpression of catalytically active METTL3, which increases m⁶A levels, promotes AML cell proliferation, whereas a catalytically inactive form is ineffective ¹⁵.
In contrast, other tumor models illustrate that loss-of-function mutations in writer genes, leading to reduced m⁶A levels, can also promote tumorigenesis. In endometrial cancer, for example, mutations in METTL14 or low expression of METTL3 are associated with increased proliferation and tumorigenicity. Mechanistically, this is mediated by PI3K/AKT pathway activation, through upregulation of the positive regulator mTORC2 and downregulation of the negative regulator PHLPP2 ¹⁶.
Studies in epithelial tumors have revealed that increased m⁶A levels can correlate with worse clinical outcomes. Specifically, depletion of FTO, the demethylase responsible for removing m⁶A, leads to global m⁶A accumulation and alters 3′-end processing of key transcripts in the Wnt signaling pathway, ultimately activating epithelial–mesenchymal transition (EMT) programs ¹⁷.
Similarly, ALKBH5 exemplifies the dual behavior of epitranscriptomic regulators. In AML ¹⁸, breast cancer ¹⁹, glioblastoma ²⁰, and gynecological tumors ²¹, ALKBH5 acts as an oncoprotein, promoting cancer stem cell self-renewal through m⁶A-dependent stabilization of oncogenic transcripts. Conversely, in pancreatic ²², lung ²³, liver ²⁴, bladder ²⁵, and osteosarcoma ²⁶ cancers, ALKBH5 functions as a tumor suppressor. Its downregulation and the consequent rise in m⁶A levels correlate with poor clinical outcomes.
Beyond METTL3, METTL14 also demonstrates context-dependent effects. In colorectal cancer (CRC), METTL14 expression is significantly reduced in tumor samples compared to normal tissues, correlating with poor prognosis. Functionally, METTL14 inhibits migration, invasion, and metastasis. Mechanistically, it promotes degradation of SOX4 mRNA, a known EMT and PI3K/AKT pathway promoter, through m⁶A modification followed by YTHDF2-dependent recognition and degradation ²⁷.
Readers, such as YTHDF2, also display context-specific behavior. In glioblastoma models, YTHDF2 can exert opposing effects depending on cellular subpopulations: in differentiated cells, it promotes degradation of tumor-suppressive mRNAs ²⁸, while in stem-like cells, it stabilizes oncogenic transcripts ²⁹. This duality underscores how readers contribute to tumor plasticity by differentially regulating epitranscriptomic programs within the same tumor.
As summarized in Table 1, m⁶A regulators exhibit remarkable functional heterogeneity in tumors, acting as oncogenes in some contexts and tumor suppressors in others. Their duality derives from selective modulation of transcript stability and translation in an m⁶A-dependent manner. The same protein, such as METTL14 or ALKBH5, may promote tumorigenesis in one cancer type and inhibit it in another, reflecting a strict dependence on molecular and cellular context. Consequently, the pathological role of m⁶A cannot be defined in absolute terms but requires interpretation within the tumor’s genetic and transcriptional landscape. This complexity, while challenging for precision oncology, opens opportunities for context-specific therapeutic strategies targeting the tumor epitranscriptomic profile.


Table 1. Context-dependent effects of m⁶A regulators across different tumor types.
This table summarizes the dual and context-specific roles of m⁶A machinery in various cancers. Columns indicate the tumor type, the overall change in global m⁶A levels (↑ increase, ↓ decrease), the altered regulator (writer, eraser, or reader), the downstream molecular mechanisms affected, and the resulting phenotypic outcome. The examples highlight that the same m⁶A regulator can exert either pro-tumor or anti-tumor effects depending on cellular context, tissue type, and specific transcript targets. References correspond to studies providing the experimental evidence for each observation.

Can m⁶A be safely targeted?
Growing interest in epitranscriptomics as a therapeutic frontier has driven the development of selective inhibitors targeting RNA methylation proteins, particularly writers. A major milestone was the discovery of STM2457, providing experimental evidence that RNA-modifying enzymes can be effectively targeted for cancer therapy. STM2457 selectively binds the METTL3–METTL14 complex, impairing its methyltransferase activity and reducing m⁶A levels on oncogenic transcripts. In preclinical AML models, STM2457 treatment impaired proliferation, promoted differentiation, and induced apoptosis, particularly in stem-like leukemic subpopulations. These effects were linked to defective translation and downregulation of key leukemogenic mRNAs ³⁰. Despite its efficacy, STM2457 has not progressed to clinical development.
This prompted the development of refined molecules, such as STC-15, the only METTL3 inhibitor currently in clinical trials. STC-15 is highly selective and potent against the catalytic METTL3–METTL14 complex ³¹. Phase 1 monotherapy studies established its safety and tolerability in patients with advanced malignancies refractory to standard therapy (NCT05584111, https://www.clinicaltrials.gov/study/NCT05584111). Clinical development continues in an ongoing Phase 1b/2 trial evaluating STC-15 in combination with the anti-PD-1 immune checkpoint inhibitor toripalimab (NCT06975293, https://www.clinicaltrials.gov/study/NCT06975293). This combination specifically targets patients with advanced solid tumors, including NSCLC, melanoma, endometrial cancer, and HNSCC, who progressed despite prior anti-PD-1 therapy, positioning METTL3 inhibition as a potential strategy to overcome immune resistance.
Experimental compounds against demethylases FTO ³² and ALKBH5 ³³ are also being explored, but remain in preclinical development. Effective epitranscriptomic therapy requires careful evaluation of compensatory mechanisms, such as potential activation of alternative writers or erasers, and their impact on tumor heterogeneity to avoid paradoxical effects or resistance.

Where will m⁶A research go next?
The m⁶A RNA modification machinery plays crucial roles in cancer, but its functions are far from straightforward. Rather than acting simply as promoters or blockers of tumor progression, m⁶A regulators function as context-dependent modulators of cancer cell identity. The same protein, whether a writer like METTL14, an eraser like ALKBH5, or a reader like YTHDF2, can produce vastly different outcomes depending on when, where, and on which transcripts it acts.
This perspective fundamentally changes how we interpret m⁶A biology: it is no longer meaningful to categorize these proteins as “good” or “bad” actors. Their functional impact depends heavily on the transcriptional profile of the tumor and its specific mutational landscape. Consequently, m⁶A is not merely a passive chemical mark; it is an active regulator of cancer plasticity, influencing critical hallmarks such as metastasis, stemness, metabolic reprogramming, and immune evasion.
As highlighted in the introduction, gene expression is not a simple linear path from DNA to protein. Post-transcriptional modifications, exemplified by m⁶A, represent a clear demonstration of this regulatory complexity. While the detailed discussion of detection methods is beyond the scope of this mini-review, it is worth noting that transcriptome-wide techniques, such as MeRIP-seq ³⁴ and GLORI ³⁵, have been instrumental in uncovering this additional layer of regulation.
Nevertheless, many questions remain unanswered. Does m⁶A dysregulation precede driver mutations during cancer initiation? How does its pattern evolve throughout tumor progression? The challenge is not to label m⁶A regulators simply as oncogenes or tumor suppressors, but to define the specific contexts in which they act as drivers and therefore represent potential therapeutic targets. Only by embracing this complexity can future research enable the development of therapies capable of targeting not just the genome, but also the dynamic regulatory logic that shapes cancer biology.