Epitranscriptomic modifications in programmed cell death: mechanistic insights and implications for liver diseases.

Introduction
Epitranscriptomic modifications, an essential regulatory layer in gene expression, encompass diverse chemical alterations introduced into RNA molecules during posttranscriptional processing. These modifications exert precise spatiotemporal control over gene expression by modulating RNA secondary structures, metabolic stability, splicing fidelity, translational efficiency, and ribonucleoprotein interactions [1]. The advent of high-throughput sequencing technologies has revolutionized the epitranscriptomics field, establishing its central role in shaping global gene regulatory networks [2]. Among these modifications, N6-methyladenosine (m6A) represents the most extensively characterized mark, demonstrating functional significance in diverse physiological and pathological contexts, including adipogenesis [3], immune dysregulation [4], neuropsychiatric disorders [5], and oncogenic transformation [6]. Beyond m6A, emerging modifications such as 5-methylcytosine (m5C), N4-acetylcytidine (ac4C), pseudouridylation, and adenosine-to-inosine (A to I) editing are gaining momentum. These modifications collectively orchestrate RNA fate determination through structural remodeling and functional reprogramming, thereby regulating critical cellular processes [7, 8]. Notably, recent breakthroughs reveal that epitranscriptomic modifications can regulate the expression levels of programmed cell death (PCD)-related genes, the translational efficiency of proteins, and the activation status of associated signaling pathways through multiple mechanisms, thereby precisely modulating the process of PCD [9–13]. The intricate crosstalk between these modifications and PCD constitutes a core molecular mechanism underlying disease pathogenesis and progression, providing new potential therapeutic targets for various diseases.
PCD represents an evolutionarily conserved, genetically encoded process of active cellular demise that maintains tissue homeostasis through precise responses to molecular signals and microenvironmental cues [14]. On the basis of distinct morphological hallmarks and molecular execution mechanisms, PCD is categorized into classical apoptosis, pyroptosis, ferroptosis, cuproptosis, and disulfidptosis. While these modalities differ in triggering stimuli and effector pathways, they share core events including plasma membrane integrity disruption and release of damage-associated molecular patterns, which collectively initiate localized or systemic inflammatory responses [15, 16]. As a central regulatory axis of biological systems, PCD not only orchestrates tissue morphogenesis during embryogenesis and immune clonal selection, but also safeguards adult tissue integrity through elimination of senescent, damaged, or infected cells. Dysregulation of this machinery is pathologically linked to oncogenesis, neurodegenerative disorders, and cardiovascular dysfunction [17–19]. Emerging evidence highlights the regulatory control of PCD pathways by epitranscriptomic modifications, with their dynamic interplay being intricately linked to liver disease progression.
Liver diseases, encompassing a spectrum of pathologies characterized by hepatocyte injury, inflammatory cell infiltration, and activation of hepatic stellate cells (HSCs), include metabolic dysfunction-associated steatotic liver disease (MASLD), viral hepatitis, alcohol-associated liver disease (ALD), hepatic fibrosis, hepatocellular carcinoma (HCC), and others. These conditions cumulatively impair liver function and disrupt its architecture. Annually, approximately 2 million deaths are attributed to liver diseases, accounting for 4% of global mortality, thus posing a significant challenge to global public health [20]. Current clinical therapeutic strategies encompass direct-acting antivirals (DAAs, such as sofosbuvir/velpatasvir for HCV) [21], metabolic modulators (such as the peroxisome proliferator-activated receptor (PPAR)-α agonist pemafibrate) [22], multikinase inhibitors (such as sorafenib, lenvatinib) [23], and immune checkpoint inhibitors (such as the combination therapy of atezolizumab and bevacizumab) [24]. Although these modalities have improved clinical outcomes, critical challenges persist, including therapeutic resistance, interpatient heterogeneity in treatment response, and extrahepatic toxicity from prolonged medication [25]. Emerging evidence demonstrates that epitranscriptomic modifications dynamically regulate PCD-associated pathways, profoundly influencing liver disease progression. This regulatory crosstalk suggests that targeting the “epitranscriptomic modifications–PCD” regulatory axis may represent a novel strategy to overcome current therapeutic limitations. This review systematically dissects the dynamic regulatory mechanisms of epitranscriptomic modifications, integrating their interactions with the PCD signaling network from multiple dimensions. It focuses on the molecular pathological characteristics of the RNA modification–PCD regulatory axis in liver diseases and provides a comprehensive overview of drugs targeting epitranscriptomic regulatory nodes.

Molecular mechanisms of epitranscriptomic modifications

Molecular mechanisms of m6A modification
m6A is among the most prevalent and extensively studied RNA modifications in eukaryotes, characterized by methylation at the N6 position of adenosine. This modification is broadly distributed across diverse RNA species, including messenger RNA (mRNA), ribosomal RNA (rRNA), long noncoding RNA (lncRNA), circular RNA (circRNA), and microRNA (miRNA) [26]. Evolutionarily conserved from yeast to humans, m6A is consistently detectable in mRNA and preferentially enriched within the RRACH consensus motif (R = G or A; H = A, C, or U), with pronounced accumulation in the coding sequence (CDS) and 3′ untranslated region (UTR) [27]. Functionally, m6A operates throughout the entire RNA life cycle, governing RNA stability, splicing, nuclear export, and translational efficiency, thereby exerting a profound influence on gene expression [28].
m6A methylation is a dynamically reversible process orchestrated by three functional groups of regulators-methyltransferases (writers), demethylases (erasers), and reader proteins (readers) [29]. The “writers” are assembled into the methyltransferase complex (MTC), whose catalytic core is a 1:1 heterodimer of methyltransferase-like 3 (METTL3) and methyltransferase-like 14 (METTL14). METTL3 serves as the catalytic subunit that donates the methyl group from S-adenosylmethionine (SAM), whereas METTL14 functions as a structural scaffold to facilitate substrate recognition. Additional co-factors—including Wilms tumor 1-associated protein (WTAP), VIRMA, RBM15/15B, and ZC3H13—recruit the core complex to specific RNA loci, thereby conferring site selectivity [30]. The “erasers” comprise FTO and ALKBH5, which catalyze α-ketoglutarate- and Fe2⁺-dependent oxidative demethylation. FTO removes methyl groups from m6A, N1-methyladenosine (m1A), and N6,2′-O-dimethyladenosine (m6Am), whereas ALKBH5 specifically demethylates m6A in nascent nuclear RNAs, modulating the assembly of RNA-processing complexes and thus the dynamics of methylation [31]. “Reader” proteins recognize m6A-modified RNAs and transduce downstream signals. Among them, the YTH-domain family (YTHDF1/2/3 and YTHDC1/2) is the most extensively characterized. YTHDF2, the first reader whose function was elucidated, recruits the CCR4–NOT deadenylase complex to accelerate mRNA degradation. YTHDF1 enhances the translational efficiency of its targets, while YTHDF3 cooperates with both YTHDF1 and YTHDF2 to fine-tune translation or decay rates. YTHDC1 participates in mRNA splicing and nuclear export. In addition, the IGF2BP family (IGF2BP1/2/3) stabilizes m6A-marked RNAs and governs their translation and localization. Heterogeneous nuclear ribonucleoproteins (HNRNPs), exemplified by HNRNPA2B1, function as m6A readers that recruit splicing factors to regulate alternative splicing of pre-mRNAs [32]. Through these interactions, reader proteins activate downstream pathways that govern RNA metabolism and biological function (Fig. 1) (Table 1).
Dynamic equilibrium of m6A modification is indispensable for maintaining RNA metabolism and cellular homeostasis; its dysregulation has been closely implicated in the onset and progression of various liver diseases. Recent studies have revealed that METTL3 depletion attenuates m6A modification of Lats2 mRNA, thereby enhancing YAP phosphorylation, blocking its nuclear translocation, and repressing the expression of profibrotic genes, ultimately mitigating hepatic fibrosis [33]. Conversely, FTO-mediated demethylation of m6A promotes hepatic lipogenesis and lipid droplet expansion via the SREBP1c pathway while suppressing CPT-1–dependent fatty-acid oxidation, leading to excessive lipid accumulation in hepatocytes and accelerating the development of MASLD [3]. Additionally, ALKBH5 exerts bidirectional effects on hepatic fibrosis through m6A modification: on one hand, it suppresses Hedgehog signaling by upregulating PTCH1 and reduces Drp1 methylation to inhibit HSCs activation, thereby ameliorating fibrosis; on the other hand, in radiation-induced hepatic fibrosis, ALKBH5 demethylates TIRAP mRNA, activating downstream cascades that exacerbate fibrogenesis [34]. Collectively, m6A modification functions as a reversible epitranscriptomic mechanism that orchestrates gene expression and cellular functions through intricate regulatory networks. Aberrant m6A regulation plays a pivotal role in hepatic pathogenesis, offering promising diagnostic biomarkers and therapeutic targets for liver diseases.

Molecular mechanisms of m5C modification
As one of the core forms of RNA epitranscriptomic modifications, m5C modification refers to the chemical process of transferring a methyl group to the fifth carbon atom of cytosine in RNA molecules under the catalysis of specific enzymes, resulting in the formation of m5C [35]. This modification is widely present in various types of RNA molecules, including mRNA, rRNA, and transfer RNA (tRNA) [35]. In terms of distribution characteristics, the localization of m5C modification in RNA is heterogeneous. Some studies have found that it is mainly enriched in the CDS and around the translation initiation site [36], while others have indicated that it is more concentrated in the UTRs [37]. These distribution differences may be associated with the functional specificity under various physiological and pathological conditions. Regarding biological functions, m5C modification exhibits specificity in different types of RNA. In tRNA, it maintains structural stability and optimizes codon–anticodon pairing to ensure accurate translation; in rRNA, it participates in ribosome assembly and maturation to maintain protein synthesis capacity; in mRNA, it regulates stability, splicing, and translation efficiency, thereby engaging in key biological processes such as cell metabolism, proliferation, and differentiation [38].
The regulatory mechanisms of m5C modification also involve methyltransferases (writers), demethylases (erasers), and reader proteins (readers). Methyltransferases use SAM as the methyl donor to catalyze the transfer of a methyl group to cytosine, forming m5C. Among the known methyltransferases, proteins from the NOL1/NOP2/SUN domain (NSUN) family (NSUN1-7) and members of the DNA methyltransferase (DNMT) family are involved, with NSUN2 and NSUN6 being the primary methyltransferases for mRNA m5C modification [39]. The erasure of m5C modification mainly involves the action of TET family proteins and demethylases such as ALKBH1. TET proteins play a crucial role in DNA demethylation by oxidizing 5-methylcytosine (5mC) stepwise to 5-hydroxymethylcytosine (5hmC), 5-formylcytosine (5fC), and 5-carboxylcytosine (5caC), ultimately achieving demethylation [40]. Similarly, in RNA, TET proteins may also remove m5C modification via a similar oxidation pathway, thereby regulating the methylation level of RNA [40]. In addition, the α-ketoglutarate-dependent dioxygenase ALKBH1 has been found to participate in the demethylation of RNA and may play a role in the erasure of m5C modification [41]. Reader proteins can recognize and bind to m5C-modified RNA, thereby affecting processes such as RNA stability and nuclear export. Known binding proteins for m5C-modified RNA include the Aly/REF export factor (ALYREF) and Y-box binding protein 1 (YBX1). Moreover, according to a recent study, the m6A-binding protein YTHDF2 has a conserved residue in its hydrophobic pocket that can also bind to m5C-modified RNA [42] (Fig. 1) (Table 2).
As one of the important epitranscriptomic modifications, m5C has increasingly emerged as a key player in liver diseases in recent years. NSUN2 stabilizes PKM2 mRNA by increasing the m5C modification at the C773 site in the 3′-UTR of PKM2 mRNA, thereby upregulating PKM2 expression, enhancing glycolysis, and ultimately promoting the growth and metastasis of HCC [43]. ALYREF stabilizes EGFR mRNA by directly binding to the m5C modification site in the 3′-UTR region of EGFR mRNA, thereby activating the STAT3 signaling pathway and ultimately promoting the progression of HCC [44]. Additionally, in MASLD, m5C methylation exhibits a positive correlation with the mRNA expression of lipid metabolism-related genes, including Acaa1b and Cyp4a10, and is significantly enriched in lipid metabolism pathways. These findings suggest that m5C-mediated epitranscriptomic modifications may promote the progression of MASLD by modulating the expression of critical genes involved in lipid metabolic processes. Overall, these findings reveal the central role of m5C modification in liver diseases and provide new insights for targeting epitranscriptomic modifications in treating liver diseases.

Molecular mechanisms of ac4C modification
ac4C is a highly evolutionarily conserved RNA chemical modification widely present in various RNA molecules from prokaryotes to eukaryotes, including major RNA types such as tRNA, rRNA, mRNA, and lncRNA [45]. Its core chemical characteristic is the acetylation modification of the N4 nitrogen atom of cytidine, and it is currently the only confirmed form of acetylation modification in eukaryotic RNA [45]. From a functional perspective, ac4C exerts important effects in physiological and pathological processes such as embryonic development, cell differentiation, stress response, and tumorigenesis by dynamically regulating RNA stability, translation efficiency, and signal pathway participation through multiple mechanisms. Specifically, in tRNA, ac4C is mainly located in the anticodon loop, which optimizes codon–anticodon pairing to maintain translation fidelity. In rRNA, it is enriched in the helix 45 region of 18S rRNA, which participates in ribosome assembly and structural stability. In mRNA, ac4C is predominantly found in the CDS and the 3′-UTR, regulating gene expression by enhancing RNA stability and translation efficiency [46].
Current studies indicate that NAT10, as the only identified ac4C writer protein, catalyzes the acetylation process by utilizing acetyl-CoA as the acetyl group donor and providing energy through the hydrolysis of adenosine triphosphate (ATP) or guanosine triphosphate (GTP) to covalently attach the acetyl group to the N4 position of RNA cytidine [47]. During the ac4C modification of tRNA, NAT10 forms a functional complex with THUMP domain-containing protein 1 (THUMPD1), which acts as a molecular chaperone by mediating the precise localization and conformational stabilization of tRNA through its THUMP domain [48]. In contrast, the ac4C modification of rRNA relies on the antisense sequence of small nucleolar RNA (snoRNA) to complementarily pair with the target rRNA, thereby guiding NAT10 to specific modification sites [49]. Although the ac4C modification mechanisms of tRNA and rRNA have been shown to involve cofactor regulation, whether similar cofactors are involved in the ac4C modification of mRNA remains to be elucidated. Using biotin-labeled oligonucleotides with or without ac4C modification, researchers performed RNA affinity purification followed by mass spectrometry analysis, which revealed eukaryotic elongation factor 2 (eEF2) as a specific binding partner of ac4C-modified RNA. Furthermore, RNA pull-down assay was conducted to validate the binding affinity of eEF2 to ac4C-modified HMGB2 mRNA. A series of functional experiments were carried out to elucidate the crucial role of eEF2 as an ac4C modification reader in the process of RNA translation [50]. Moreover, research has identified SIRT7 as the only potential ac4C deacetylase reported to date. SIRT7 binds to 18S rRNA through its catalytic domain key residue H187 and RNA-binding motif, and removes the ac4C modification in a NAD⁺-independent manner, thereby accelerating rRNA degradation. Mutations in H187 or the RNA-binding motif significantly weaken its deacetylation activity [51]. However, the function of SIRT7 as a specific ac4C eraser is still under further validation and exploration (Fig. 1) (Table 3).
As a dynamic and reversible epitranscriptomic modification, ac4C modification profoundly influences the progression of liver diseases. NAT10 catalyzes the ac4C modification of Srebp-1c mRNA, thereby enhancing its stability and expression. This leads to the upregulation of lipogenic enzymes and promotes the development of MASLD and metabolic dysfunction-associated steatohepatitis (MASH) [52]. Additionally, NAT10 catalyzes the ac4C modification in the CDS region of HMGB2 mRNA, enhancing the binding of eEF2 to the modified site and thereby promoting the translation of HMGB2. This process drives the proliferation and metastasis of HCC [50]. Moreover, NAT10 catalyzes the ac4C modification of Srebf1 and Scap mRNA, enhancing their mRNA stability and translation efficiency, thereby promoting hepatic lipogenesis. Inhibition of NAT10 reduces this modification and lowers the expression of lipogenesis-related genes, providing a novel therapeutic strategy for MASLD [53]. In summary, NAT10-mediated ac4C modification finely regulates RNA function and offers potential therapeutic targets for diagnosing and treating liver diseases.

Molecular mechanisms of pseudouridylation
Pseudouridylation refers to the RNA modification process in which uridine is isomerized to pseudouridine (Ψ) by the catalysis of pseudouridine synthase (PUSs). As one of the most conserved and abundant RNA modifications, it is widely present in tRNA, rRNA, mRNA, and small nuclear RNA (snRNA), where it regulates RNA structure and function and participates in various physiological and pathological processes [54]. The distribution and function of pseudouridylation are RNA-specific. In tRNA, pseudouridylation is mainly located in the TΨC loop (Ψ 55) and the anticodon loop (Ψ 34–39), where it stabilizes the L-shaped structure, optimizes codon–anticodon pairing, and thereby enhances translation accuracy. In rRNA, pseudouridylation is enriched in the peptidyl transferase and decoding centers, participating in ribosome assembly and enhancing translation efficiency. In mRNA, pseudouridylation is distributed in the CDS and the UTRs, where it functions by improving stability, regulating translation initiation and termination, and modulating splicing. In snRNA, pseudouridylation promotes spliceosome assembly by stabilizing interactions with precursor mRNA [55].
The installation of pseudouridylation primarily relies on the catalytic action of PUS or the H/ACA small nucleolar ribonucleoprotein complex (H/ACA snoRNP, which is composed of snoRNA and four core proteins: DKC1, NOP10, NHP2, and GAR1) [56]. The RNA-independent pathway is accomplished independently by enzymes from the PUS family. PUS enzymes recognize specific sequences or structures in the substrate RNA through their conserved catalytic domains, inducing isomerization of the glycosidic bond of uridine to form Ψ with a C5–C1′ glycosidic bond linkage [57]. In contrast, the RNA-dependent pathway relies on the H/ACA snoRNP, which uses the guide sequence of snoRNA to form complementary pairing with the target RNA, precisely locating the modification site. Subsequently, Cbf5 catalyzes the rotation isomerization of the uridine glycosidic bond to form Ψ [56]. Both mechanisms achieve the conversion from uridine to Ψ by flipping the glycosidic bond within the molecule, significantly enhancing RNA stability and functional conformation. They participate in ribosome biogenesis, RNA translation regulation, and dynamic modulation of RNA–protein interaction networks. To date, no classical reader or eraser proteins for pseudouridylation have been identified (Fig. 1) (Table 4).
Pseudouridylation profoundly influences RNA metabolism and liver diseases through a sophisticated regulatory network. PUS1 is highly expressed in HCC, which promotes the progression of HCC by enhancing the translation efficiency of oncogenes such as IRS1 and c-Myc through pseudouridylation of their mRNAs [58]. Additionally, DKC1, which mediates pseudouridylation of rRNA and spliceosomal mRNAs, is involved in ribosome biogenesis and pre-mRNA maturation. Upregulation of DKC1 expression can promote the malignant proliferation and survival of HCC cells. Moreover, oxidative modification of PDIA3 releases its inhibitory effect on DKC1, further enhancing its oncogenic activity [59]. A thorough investigation into the dynamic changes, regulatory mechanisms, and impacts on key signaling pathways of pseudouridylation in liver diseases is an important direction for uncovering new pathogenic mechanisms and identifying potential diagnostic and therapeutic targets.

Molecular mechanisms of A-to-I editing
A-to-I editing is one of the most prevalent RNA modifications in eukaryotes. It is catalyzed by adenosine deaminases acting on RNA (ADARs), which deaminate adenosine to inosine. This modification can influence RNA function since inosine is recognized as guanosine during translation and RNA processing [30]. A-to-I editing is widely distributed across various types of RNA, including the CDS, introns, and 3′-UTRs of mRNAs, with double-stranded RNA (dsRNA) formed by Alu repeat sequences being a major target. In addition, precursors of miRNAs, lncRNAs, and circRNAs are also frequently modified, forming a complex epitranscriptomic regulatory network [8]. A-to-I editing plays a crucial role in multiple biological processes, significantly impacting gene expression regulation, RNA stability, splicing, translation efficiency, and the dynamic modulation of RNA–protein interaction networks.
The molecular mechanism of A-to-I editing primarily relies on the recognition and catalysis of dsRNA structures by ADAR proteins. The ADAR family comprises three subtypes: ADAR1, ADAR2, and ADAR3, among which ADAR1 and ADAR2 possess catalytic activity. ADAR1 exists in two isoforms, p110 and p150, both containing three double-stranded RNA-binding domains (dsRBDs). The p150 isoform also has a Zα domain for recognizing Z-RNA. ADAR2 contains two dsRBDs and is mainly localized in the nucleolus, where it can regulate its own function through self-controlled pre-mRNA splicing [60]. These isoforms anchor to the target dsRNA regions via dsRBDs and catalyze the deamination of the amino group at the C6 position of adenosine to generate inosine. During translation, ribosomes recognize inosine as guanosine, leading to codon reassignment or alteration in the function of non-coding RNAs [61]. Additionally, ADAR1 can inhibit the abnormal activation of innate immune sensors such as MDA5 and PKR by editing endogenous retrotransposons, thereby preventing excessive production of type I interferons and maintaining immune homeostasis [62]. Although ADAR3 lacks catalytic activity, it competitively binds to dsRNA through its Arg-rich domain, thereby inhibiting the editing functions of ADAR1 and ADAR2 and serving as a major negative regulator of the editing process [60] (Fig. 1) (Table 5).
In various liver diseases, the A-to-I editing landscape undergoes significant alterations, profoundly affecting the pathological processes of these diseases. ADAR1-mediated A-to-I editing regulates the abundance of small RNAs such as snoRNA and Y RNA in hepatocytes, thereby influencing mRNA translation efficiency and the transcriptional function of RNA polymerase III. It also participates in maintaining the normal response of hepatocytes to interferon. Its absence leads to hepatocyte growth arrest, translational inhibition, and disruption of the small RNA landscape [63]. ADAR1 can also edit the RNA of hepatitis B virus (HBV), thereby disrupting host immune recognition and suppressing the interferon response, thus facilitating HBV replication. The HBx protein of HBV can transcriptionally upregulate ADAR1 expression to enhance this immune evasion mechanism, while ADAR1 inhibitors can promote HBV clearance by activating the immune response [64]. In addition to these effects, A-to-I editing mediated by ADARs is dysregulated in hepatocellular carcinoma. Overexpression of ADAR1 and downregulation of ADAR2 induce gene-specific editing changes, such as hyperediting of FLNB and hypoediting of COPA, which exert oncogenic and tumor-suppressive effects, respectively. These differential expression patterns are closely associated with increased incidence of cirrhosis, higher risk of postoperative recurrence, and poor prognosis in patients [65]. In summary, A-to-I editing influences gene expression and cellular functions through a sophisticated regulatory network, providing potential molecular targets for targeted therapy of liver diseases.

Other epitranscriptomic modifications
The term 2′-OH-acylation refers to the selective acylation reaction of the 2′-hydroxyl (2′-OH) group in RNA molecules, serving as a chemical methodology for direct RNA functionalization. Unlike traditional RNA modification approaches that rely on solid-phase synthesis or enzymatic conversion, 2′-OH acylation offers a versatile alternative for efficiently modifying both synthetic RNA and transcribed RNA, positioning it as an emerging research focus in RNA chemistry. The 2′-OH group represents a hydrolytically vulnerable site in RNA. By introducing 2′-OH acylation modifications, this reactive moiety can be sterically blocked, significantly enhancing RNA stability in vitro and in vivo. Reversible 2′-OH acylation further functions as a “molecular switch,” enabling precise spatiotemporal control over RNA functionality [66]. Previous studies demonstrate that acylation shields the nucleophilicity of 2′-OH, thereby inhibiting its participation in spontaneous hydrolysis via intramolecular attack on the adjacent 3′-phosphoester bond. This extends RNA lifespan in aqueous solution at 37 °C by up to sevenfold. Concurrently, a high degree of acylation effectively obstructs cleavage sites for RNases [67]. With continuous advancements in reagent design and application technologies, 2′-OH acylation is poised to play an increasingly critical role in both fundamental RNA biology research and applied RNA technologies.
In addition to the several RNA epitranscriptomic modifications mentioned above, m1A, N7-methylguanosine, ribose methylation, and guanine-to-uracil transversion are also important types of RNA modifications. These modifications play a key role in the metabolism, stability, and functional regulation of RNA. For example, m1A modification may affect the translation efficiency and stability of mRNA; N7-methylguanosine modification plays an important role in the nuclear export and translation initiation of mRNA; ribose methylation modification and guanine-to-uracil transversion are involved in the structural and functional regulation of RNA [68–70]. The diversity and complexity of these modifications further enrich the RNA regulatory network, providing more possibilities for the fine regulation of gene expression.

Epitranscriptomic modifications of PCD

Epitranscriptomic modifications of ferroptosis
Ferroptosis is an iron-dependent form of PCD triggered by the accumulation of iron-catalyzed lipid peroxidation. It is distinct from traditional cell death modes such as apoptosis and pyroptosis [71]. The core characteristics of ferroptosis include iron ion-dependent lipid reactive oxygen species (ROS) accumulation, mitochondrial structural abnormalities, and disruption of cell membrane integrity, accompanied by glutathione (GSH) depletion and increased lipid peroxidation products [72]. The regulatory mechanisms of ferroptosis involve three major core metabolic networks. Firstly, iron metabolism imbalance is one of the key factors. Increased iron uptake mediated by transferrin receptor 1 (TFR1) and iron release promoted by ferritinophagy mediated by nuclear receptor coactivator 4 (NCOA4) expand the labile iron pool within the cell, accelerating ROS generation through the Fenton reaction [73]. Secondly, the blockage of cystine uptake by the cystine/glutamate antiporter (system Xc−, composed of SLC7A11/SLC3A2) leads to reduced GSH synthesis, which in turn inhibits the activity of glutathione peroxidase 4 (GPX4), preventing it from effectively clearing lipid peroxides (l-OOH) [74]. Lastly, lipid metabolism disorder is also an important mechanism of ferroptosis. Acyl-CoA synthetase long-chain family member 4 (ACSL4) and lysophosphatidylcholine acyltransferase 3 (LPCAT3) mediate the incorporation of polyunsaturated fatty acids (PUFAs) into membrane phospholipids, which are then catalyzed by lipoxygenases (LOXs) to generate lipid peroxides, disrupting membrane structure [75]. In addition, the key regulatory pathways of ferroptosis include the GPX4 pathway (core antioxidant defense), the FSP1-CoQ10 pathway (non-GPX4-dependent lipid peroxidation inhibition), and the DHODH pathway (mitochondrial antioxidant) [76]. These pathways collectively maintain lipid redox balance to suppress the occurrence of ferroptosis. Ferroptosis is closely related to a variety of diseases, including liver diseases. Studies have found that APE1 regulates the sensitivity of HCC cells to ferroptosis by modulating the NRF2/SLC7A11/GPX4 axis through its redox activity, and inhibiting APE1 can enhance the antitumor efficacy of ferroptosis inducers [77]. ATF4 inhibits ferroptosis by inducing SLC7A11 expression, thereby suppressing the occurrence and development of HCC [78]. Moreover, ferroptosis plays a key role in the pathogenesis of MASLD through iron-dependent lipid peroxidation. It can exacerbate hepatic oxidative stress, inflammation, and cell damage, promoting the progression of the disease from simple steatosis to MASH. Targeting ferroptosis-related pathways may provide potential therapeutic strategies for MASLD [79]. In summary, in liver diseases such as hepatocellular carcinoma and MASLD, inhibiting ferroptosis helps protect hepatocyte function and reduce liver tissue damage, thereby providing new intervention targets and therapeutic approaches for the treatment of these diseases.
m6A methylation and ferroptosis are current research hotspots, jointly influencing disease progression through a multi-level regulatory network. Notably, m6A methyltransferase METTL3 plays a significant role in regulating ferroptosis in sepsis-related lung injury. On the one hand, METTL3-mediated m6A methylation is enriched in ACSL4 and upregulates its expression by modulating mRNA stability via a YTHDC1-dependent pathway, thereby promoting mitochondria-associated ferroptosis [80]. On the other hand, METTL3 can induce the upregulation of HIF-1α expression through an IGF2BP2-dependent mechanism. As a key regulator of cellular glycolysis, the upregulation of HIF-1α enhances glycolysis and reduces oxidative phosphorylation, and this metabolic reprogramming leads to the accumulation of ROS. Meanwhile, the upregulation of HIF-1α also promotes ferroptosis by downregulating GPX4 [81]. However, METTL3 also has an inhibitory effect on ferroptosis. It can bind to the 3′-UTR region of GPX4 mRNA, enhance its translation efficiency, and maintain the expression level of GPX4, thereby inhibiting ferroptosis [82]. In addition to METTL3, METTL14 plays a significant role in regulating cellular ferroptosis. It can bind to the mRNA of ferroptosis-related genes SAT1 and ACSL4 and increase their m6A methylation. After recognition by IGF2BP2, the stability of the two is enhanced, leading to the upregulation of SAT1 and ACSL4 expression and the promotion of ferroptosis [83]. Conversely, METTL14 can destabilize lncRNA TUG1 via m6A methylation. This reduces TUG1′s ability to promote GDF15 ubiquitination and degradation, leading to higher GDF15 levels and activation of the nuclear factor E2-related factor 2 (NRF2) pathway, which exerts an inhibitory effect on ferroptosis [84]. Additionally, FTO inhibits ferroptosis by activating the P21/NRF2 signaling pathway. On the one hand, FTO can activate P53 by mediating P53 m6A demethylation, thereby upregulating the P21/NRF2 pathway. On the other hand, FTO can also directly mediate the m6A demethylation of P21 and NRF2, activating this P21/NRF2 pathway, promoting the expression of ferroptosis-related proteins such as SLC7A11, GPX4, and FTH1, and inhibiting ferroptosis [85]. In contrast, the m6A demethylase ALKBH5 binds to specific m6A modification sites in the 3′-UTR region of ACSL4 mRNA, promoting the stability of ACSL4 mRNA, thereby upregulating ACSL4 expression and promoting the ferroptosis process [86]. YTHDF1 is negatively correlated with cell ferroptosis in prostate cancer. Specifically, YTHDF1 recognizes and binds to the m6A modification sites of PD-L1 mRNA, enhancing the stability of PD-L1 mRNA, helping cancer cells evade the cytotoxic effects of effector T cells, and indirectly inhibiting CD8+ T cell-mediated ferroptosis [87]. YTHDF2 can recognize and bind to m6A-modified GCH1 mRNA, promoting its translation. The generated GCH1 can participate in the synthesis of BH4 with antioxidant activity, thereby inducing ferroptosis in cardiomyocytes [88]. IGF2BP1 can bind to AIFM2 mRNA and enhance its stability through m6A modification, promoting AIFM2 expression, thereby reducing intracellular iron ion levels and ROS generation, increasing SOD and GSH content, upregulating GPX4 and SLC7A11 expression, and inhibiting the occurrence of ferroptosis [89] (Fig. 2).
m5C modification regulates ferroptosis through multidimensional mechanisms. As a core methyltransferase, NSUN2 regulates the stability of SLC7A11 mRNA through m5C modification, thereby affecting the expression level of the SLC7A11 protein. Knockdown of NSUN2 reduces the expression of SLC7A11, increases the levels of lipid ROS and lipid peroxidation, and significantly enhances the sensitivity of endometrial cancer cells to ferroptosis [90]. Moreover, NSUN2 can catalyze the m5C modification of the 3′-UTR of Gpx4 mRNA, enhancing the binding of SBP2 to Gpx4 mRNA and thereby promoting the expression of the GPX4 protein. When NSUN2 expression is downregulated, the m5C modification level of Gpx4 mRNA decreases, leading to reduced expression of the GPX4 protein and increased cellular sensitivity to ferroptosis [91]. In addition to NSUN2, YBX1 also plays a key role in the regulation of cellular ferroptosis. At the transcriptional level, YBX1 directly binds to the promoter region of RAC3 to promote its expression and enhance cellular antioxidant defense capabilities [92]. At the posttranscriptional level, YBX1 specifically recognizes the m5C modification sites of SLC7A11 and G6PD mRNA, stabilizes target mRNA by recruiting ELAVL1, and thereby inhibits the occurrence of ferroptosis [93]. Furthermore, overexpression of the demethylase ALKBH1 leads to a decrease in overall protein translation rate, rendering cells resistant to Erastin-induced ferroptosis but more sensitive to RSL3-induced ferroptosis, thereby playing an important role in cellular ferroptosis [94].
ac4C modification and A-to-I editing play key roles in the regulatory mechanisms of cellular ferroptosis, and both exert significant effects on the ferroptosis process through NAT10 and ADAR1, respectively. Research has shown that NAT10 enhances the stability of ferroptosis suppressor protein 1 (FSP1) mRNA by mediating ac4C modification, thereby increasing FSP1 expression levels and effectively inhibiting ferroptosis in CRC cells [95]. In studies on nasopharyngeal carcinoma, NAT10 promotes the expression of SLC7A11 through ac4C modification, exerting an anti-ferroptotic effect [96]. In ovarian cancer, NAT10 enhances the stability and translation efficiency of ACOT7 mRNA by increasing its ac4C modification, thereby affecting fatty acid metabolism and inhibiting ferroptosis [97]. In addition, ADAR1 is closely associated with cellular ferroptosis. The absence of ADAR1 leads to inactivation of the FAK/AKT signaling axis, which relieves the inhibition of monounsaturated fatty acids synthesis by AKT, causing lipid metabolic imbalance and ultimately promoting ferroptosis in CRC cells [98]. In breast cancer research, ADAR1 negatively regulates the expression of miR-335-5p, which targets the Sp1 transcription factor and thereby affects GPX4 expression. When ADAR1 expression is reduced, miR-335-5p expression increases, inhibiting the Sp1/GPX4 axis and promoting ferroptosis [99]. Furthermore, snoRNA SNORA56 promotes the translation of glutamate-cysteine ligase (GCLC) by pseudouridylation of the U1664 site in 28S rRNA in CRC. GCLC is the key rate-limiting enzyme for GSH synthesis, and its increased expression can inhibit ferroptosis, thereby promoting CRC cell proliferation and tumorigenesis [100].

Epitranscriptomic modifications of cuproptosis
Cuproptosis is a form of PCD triggered by copper ion overload and has garnered widespread attention in recent years. Its core characteristics include mitochondrial dysfunction, disruption of cell membrane integrity, and a significant increase in oxidative stress levels. Copper ions initiate protein aggregation and proteotoxic stress by binding to acylases in the tricarboxylic acid (TCA) cycle, ultimately leading to cell death [101]. The regulatory mechanisms of cuproptosis involve multiple levels, including copper metabolism imbalance, mitochondrial metabolic disorder, and protein homeostasis disruption. After Cu2⁺ enters the mitochondria, it is reduced to the more toxic Cu⁺ by FDX1. Meanwhile, FDX1 catalyzes the lipoic acid modification of TCA cycle enzymes such as DLAT, promoting the abnormal oligomerization of copper–lipoic acid complexes. This directly disrupts the assembly of respiratory chain complexes, accumulating oxidative and proteotoxic stress within the cell [102]. The key regulatory pathways of cuproptosis include both mitochondrial-dependent and non-mitochondrial-dependent routes. In the mitochondrial-dependent pathway, copper ions inhibit the activity of mitochondrial respiratory chain complexes, disrupting electron transfer and ATP synthesis, thereby causing cellular energy metabolic disorders. In the non-mitochondrial-dependent pathway, copper ions regulate the intracellular redox state, activating various stress pathways such as NRF2/ARE and p53, affecting cell survival and death [103]. Moreover, cuproptosis is closely related to multiple cellular metabolic processes, such as lipid, amino acid, and nucleotide metabolism. The disruption of these metabolic processes further exacerbates the occurrence of cuproptosis. Previous studies have found that cuproptosis is closely related to liver diseases. For example, Wilson’s disease is caused by abnormal copper metabolism leading to hepatic copper accumulation, and the zinc transporter ZnT1, which acts as a copper transporter, may become a potential therapeutic target for Wilson’s disease and other liver diseases by regulating cuproptosis [104]. In MASLD, cuproptosis interferes with lipid metabolism and insulin sensitivity, driving disease progression. In HCC, cuproptosis-related genes affect tumor cell proliferation, metastasis, and sensitivity to chemotherapeutic drugs [105]. Research on the mechanisms and regulatory pathways of cuproptosis is still in its infancy. In-depth studies of the molecular mechanisms and regulatory networks of cuproptosis are of great significance for the development of new therapeutic strategies for liver diseases.
The interplay between m6A modification and cuproptosis influences cell fate through multilevel dynamic regulation. Research has shown that METTL16 is highly expressed in gastric cancer and is closely associated with cuproptosis in gastric cancer cells. Specifically, copper stress induces lactylation of METTL16 at lysine 229 (K229), which significantly enhances the methyltransferase activity of METTL16. This subsequently increases the transcriptional and protein expression levels of FDX1 through m6A modification. FDX1, as a key regulator of cuproptosis, efficiently reduces relatively stable Cu2+ to the more cytotoxic Cu⁺, ultimately inducing cuproptosis [12]. In addition, YTHDF1 plays an indispensable regulatory role in the process of cuproptosis. Studies have shown that YTHDF1 can tightly bind to FDX1 mRNA molecules, maintaining the stability of FDX1 mRNA and preventing its premature degradation by nucleases within the cell. This upregulates its expression and enhances the sensitivity of glioma cells to cuproptosis [106]. The cuproptosis-related gene LIPT1 is downregulated in bladder cancer. However, restoring its expression level through genetic intervention effectively inhibits the proliferation of bladder cancer cells and promotes cuproptosis. Further investigation into the molecular mechanisms revealed that YTHDF2 plays a key regulatory role in this process. YTHDF2, with its highly specific recognition ability, precisely identifies and binds to the m6A modification sites on LIPT1 mRNA molecules, promoting the degradation of LIPT1 mRNA and reducing the expression level of LIPT1, thereby inhibiting the cuproptosis process in bladder cancer cells [107] (Fig. 3).
Existing literature primarily focuses on the regulation of cuproptosis by m6A modification, whereas the role of m5C in cuproptosis has been less studied. Previous studies have shown that, in cholangiocarcinoma, NSUN5 enhances the stability of GLS mRNA through m5C modification, thereby increasing the accumulation of GLS protein. GLS, a key protein inhibiting cuproptosis, enables cholangiocarcinoma cells to resist copper-induced cell death through its upregulated expression, thus promoting tumor progression [108]. It is worth noting that the roles of ac4C modification, pseudouridylation, and A-to-I editing in cellular cuproptosis have not yet been systematically investigated. However, given the crucial roles of these modifications in cellular activities and the widespread regulatory mechanisms of RNA modifications, it is reasonable to speculate that they may indirectly influence the cuproptosis process by regulating genes related to copper metabolism. For example, certain RNA modifications may alter the expression levels of copper ion transporter genes, affecting the uptake, storage, and excretion of copper ions by cells, and thereby influencing the homeostasis of intracellular copper ions and ultimately affecting cuproptosis. Alternatively, they may modulate the sensitivity of cells to cuproptosis by regulating key genes in signaling pathways related to cuproptosis. Although research in this area is still limited, it provides potential directions and ideas for future studies.

Epitranscriptomic modifications of disulfidptosis
Disulfidptosis is a novel form of PCD first reported in 2023. It is driven by the collapse of the cytoskeleton caused by the abnormal accumulation of disulfide bonds within the cell and is characterized by unique morphological features such as cell shrinkage, F-actin contraction, and detachment from the cell membrane. This form of cell death is distinctly different from other types of cell death, which cannot be blocked by ferroptosis inhibitors (Fer-1, DFO), apoptosis inhibitors (Z-VAD-FMK), or autophagy inhibitors (CQ). Moreover, it is independent of ATP levels and represents an entirely new type of cell death [109]. Mechanistically, under glucose deprivation and NADPH depletion conditions, cells with high expression of SLC7A11 take up large amounts of cystine, leading to cystine accumulation and abnormal disulfide cross-linkages. These cross-linkages disrupt the actin cytoskeleton structure, ultimately triggering cell death. Additionally, it has been confirmed that the actin remodeling mediated by the Rac1-WRC-Arp2/3 signaling pathway regulates this process [110]. Currently, research on disulfidptosis is still in its early stages. However, its unique molecular mechanisms and close association with tumor metabolism have made it a hot topic in the field of cell death and anticancer drug development, providing a new direction for targeting tumors with high expression of SLC7A11. Meanwhile, studies have found that disulfidptosis is closely related to the occurrence and development of liver diseases. The underlying mechanisms may influence hepatocyte metabolism and cytoskeleton stability, offering a new perspective for the research and treatment of liver diseases [111].
Study on epitranscriptomic modifications and disulfidptosis is also in the initial stage, but the prospect of its interaction is foreseeable. As an important RNA demethylase, FTO is closely associated with the occurrence of disulfidptosis. Research has found that FTO is highly expressed in uveal melanoma (UM) tissues and is closely correlated with early recurrence, increased invasiveness, and poor prognosis of UM tumors. Further exploration of the mechanism of action of FTO in UM cells revealed that high expression of FTO can reduce the m6A level within cells, thereby affecting the expression and function of a series of downstream genes. To verify the role of FTO in the process of cellular disulfidptosis, researchers treated UM cells with the FTO inhibitor meclofenamic acid (MA) and found that the m6A level within cells was significantly restored, the expression level of SLC7A11 was markedly upregulated, and a series of typical disulfidptosis features, such as cell shrinkage and disruption of the actin cytoskeleton, appeared in the cells, strongly confirming that inhibiting FTO can induce tumor cells to undergo disulfidptosis [112]. In addition, ACTN4, one of the disulfidptosis-related genes, also plays an important role in lung adenocarcinoma. It may affect tumor development by regulating the function and infiltration of immune cells, and its expression level is significantly correlated with 20 m6A-related genes, which further reveals the close connection between disulfidptosis and epigenetic modifications in the process of tumor occurrence and development [113]. During the occurrence and development of breast cancer, disulfidptosis-related lncRNAs interact with epigenetic modifications to jointly affect the biological behavior of tumors. Studies have shown that the expression levels of four key lncRNAs, GATA3-AS1, LINC00511, LINC01488, and LINC02188, vary significantly among different breast cancer subtypes, and they exhibit diverse correlations with m6A, m1A, and m5C modification genes of RNA. This correlation enables these lncRNAs to serve as potential biomarkers for accurately predicting breast cancer subtypes, thereby providing a new perspective and strategy for the diagnosis and personalized treatment of breast cancer [114]. However, research on the mechanisms of epigenetic modifications in regulating disulfidptosis is still limited (Fig. 4).

Epitranscriptomic modifications of apoptosis
Apoptosis is a highly regulated PCD mediated by cysteine aspartate proteases (caspases). It maintains homeostasis by eliminating damaged or abnormal cells and is distinct from passive death modes such as necrosis. Its characteristics include cell shrinkage, chromatin condensation, nuclear fragmentation, and the formation of apoptotic bodies through membrane blebbing. The initiation of apoptosis relies on two major pathways: extrinsic and intrinsic [115]. The intrinsic pathway (mitochondrial pathway) is triggered by DNA damage, oxidative stress, and other factors. Bax/Bak mediate mitochondrial membrane permeabilization, which releases cytochrome c to form the apoptosome and activate caspase-9. The extrinsic pathway (death receptor pathway) is initiated by ligands such as FasL and TNF-α binding to death receptors such as Fas and TNFR1. This binding recruits FADD to form the death-inducing signaling complex (DISC), which activates caspase-8 [116]. Bcl-2 family proteins are key regulators of the intrinsic pathway. They precisely control mitochondrial membrane permeability through the dynamic balance between pro-apoptotic molecules (such as Bax and Bak) and anti-apoptotic molecules (such as Bcl-2 and Bcl-XL), determining the cell’s fate of survival or death [116]. Apoptosis is closely related to the occurrence and development of various liver diseases. In diseases such as alcohol-associated liver disease (ALD), MASLD, and drug-induced liver injury, hepatocyte apoptosis exacerbates liver damage through mechanisms involving mitochondrial dysfunction, oxidative stress, and inflammatory responses. Therefore, regulating apoptosis-related pathways may provide new therapeutic targets for the treatment of liver diseases [117].
m6A modification plays an important role in apoptosis by dynamically regulating the stability, transport, and translation efficiency of RNA. In the context of cancer, METTL3 enhances the translation efficiency of Bcl-2 mRNA by catalyzing its m6A modification, thereby affecting the proliferation and apoptosis of breast cancer cells [118]. By modifying hepatoma-derived growth factor (HDGF) mRNA with m6A, METTL3 enhances the binding of IGF2BP3 to the m6A sites, stabilizing HDGF mRNA. Subsequently, nuclear HDGF activates glycolysis-related proteins to promote glycolysis, thereby inhibiting apoptosis in gastric cancer cells [119]. In noncancerous diseases, METTL3 stabilizes TIMP2 mRNA via m6A modification. This subsequently upregulates the expression of Notch3 and Notch4, promoting podocyte injury. Conversely, knocking down TIMP2 significantly alleviates kidney injury and podocyte apoptosis [120]. In colitis, METTL14 stabilizes Nfkbia mRNA, thereby inhibiting the nuclear factor kappa-light-chain-enhancer of activated B cells (NF-κB) signaling pathway, downregulating the expression of anti-apoptotic genes, and promoting apoptosis in intestinal epithelial cells [121]. In diabetic nephropathy, METTL14 reduces the stability of α-klotho mRNA, leading to its downregulation, increased pro-apoptotic factors, inhibition of cell proliferation, and accelerated apoptosis [122]. Unlike METTL3 and METTL14, FTO in breast cancer induces degradation of BNIP3 mRNA by removing m6A modifications in its 3′-UTR, resulting in downregulation of BNIP3 expression. As a pro-apoptotic gene, the downregulation of BNIP3 prevents the normal activation of apoptosis-related proteins such as caspase-3 and fails to effectively inhibit the expression of the anti-apoptotic protein Bcl-2, thereby inhibiting apoptosis in breast cancer cells [123]. Additionally, the reader protein YTHDF1 regulates apoptosis in liver cells by controlling the translation of MFG-E8. Its absence fails to effectively activate the FAK-STAT3 signaling pathway, leading to impaired mitochondrial function, increased ROS levels, and exacerbated apoptosis in liver cells [124]. YTHDF2 affects apoptosis in tumor cells such as MYC-driven breast cancer through complex mechanisms. Its inhibition or absence activates the epithelial–mesenchymal transition (EMT) signaling pathway, causing cells to adopt a mesenchymal phenotype and increasing overall translation rates. However, excessive translation leads to endoplasmic reticulum stress, accumulation of unfolded proteins, and activation of c-Jun N-terminal kinase (JNK), ultimately triggering apoptosis. Meanwhile, the translation of pro-apoptotic mRNAs such as CST3 increases within the cell, participating in caspase-mediated cell death and further promoting the apoptotic process [125] (Fig. 5).
m5C exerts multilevel regulatory effects in the apoptosis regulatory network. During tumorigenesis and progression, NSUN2 is closely associated with apoptosis. Glucose induces the oligomerization and activation of NSUN2 by binding to its N-terminal amino acid domain, thereby maintaining the m5C modification and stability of TREX2 mRNA. TREX2 degrades cytoplasmic dsDNA, thereby inhibiting the activation of the cGAS/STING pathway, blocking the release of type I interferons and the apoptosis pathway, and thus maintaining the survival of tumor cells [126]. Additionally, NSUN2 activates the MAPK/ERK signaling axis in an m5C-dependent manner. Knockdown of NSUN2 inhibits the phosphorylation of ERK1/2, reverses the Bcl-2/Bax expression ratio, and induces caspase-9-dependent apoptosis [127]. Moreover, NSUN5 is highly expressed in clear cell renal cell carcinoma (ccRCC), where it inhibits cancer cell invasion, proliferation, and migration. It also promotes apoptosis by activating the p53 signaling pathway, thereby suppressing the progression of ccRCC [128]. ALYREF forms an RNA–protein complex with LINC02159, specifically recognizes the m5C modification sites on YAP1 mRNA, enhances its stability, promotes YAP1 protein expression, and activates the Hippo/YAP and Wnt/β-catenin pathways, ultimately endowing tumor cells with apoptosis resistance [129].
The ac4C modification dynamically regulates the stability and translational activity of target gene mRNA, playing an important role in the process of apoptosis. Previous studies have shown that NAT10 can regulate apoptosis through the p53 pathway. On the one hand, NAT10 acts as a co-activator of p53, recruiting the histone acetyltransferase PCAF to catalyze the acetylation of lysine 120 (K120) on p53. On the other hand, NAT10 uses its E3 ligase activity to promote the degradation of Mdm2, inhibiting Mdm2-mediated ubiquitination of p53, thereby maintaining the homeostasis of p53 protein under basal conditions. When DNA is damaged, NAT10 undergoes nucleocytoplasmic translocation and forms a functional complex with p53, co-activating pro-apoptotic effectors such as PUMA and cell cycle arrest factors such as p21, driving the apoptotic cascade [130]. In addition, in CRC, NAT10 specifically binds to the 3′-UTR region of KIF23 mRNA, catalyzes its ac4C modification, and enhances its stability, thereby promoting the expression of the KIF23 protein. KIF23 activates the Wnt/β-catenin signaling axis, inducing the accumulation of β-catenin in the nucleus, which in turn upregulates the expression of proliferation-related genes such as cyclin D1 and c-Myc, as well as anti-apoptotic genes such as Bcl-xL and Survivin, endowing tumor cells with apoptosis resistance [131]. In non-small cell lung cancer, NAT10 maintains the high expression of enolase 1 mRNA by catalyzing its ac4C modification, thereby driving glycolytic metabolic reprogramming. Knockdown of NAT10 reverses this effect, ultimately inducing apoptosis through metabolic stress [132].
Pseudouridylation, as an important form of epitranscriptomic regulation, dynamically reshapes RNA functions and participates in the precise regulation of apoptosis. Studies have found that PUS1 is highly expressed in HCC tissues and cell lines, and its expression level is significantly correlated with tumor stage and patient overall survival. Functional studies have shown that knockdown of PUS1 significantly inhibits HCC cell proliferation and induces apoptosis. Further research indicates that PUS1 can affect HCC progression by regulating the mTOR and MYC signaling pathways. Knockdown of PUS1 downregulates the expression of c-MYC and mTOR, thereby affecting the mRNA expression of downstream genes and inducing apoptosis [133]. In addition, PUS7 is highly expressed in CRC cell lines and tissues and may inhibit the activation of caspase-9 by activating the PI3K/Akt/mTOR signaling pathway. Treatment with the PI3K inhibitor LY294002 can reverse the aforementioned effects and thereby promote apoptosis [134]. Research on lung adenocarcinoma has revealed a unique mechanism of another Ψ-modifying enzyme, dyskerin pseudouridine synthase 1 (DKC1). As a core component of the telomerase ribonucleoprotein, downregulation of DKC1 expression leads to impaired stability of TERC and reduced telomerase activity, ultimately triggering telomere crisis. These changes ultimately induce cellular senescence and apoptosis, characterized by an increase in senescence-associated β-galactosidase-positive cells and elevated expression of senescence and apoptosis markers such as P21, as well as DNA damage markers such as γH2A.X [135].
A-to-I RNA editing, mediated by the ADAR protein family through base conversion mechanisms, plays an important role in apoptosis. In inflammatory diseases, ADAR1 maintains cell survival through a dual mechanism. On the one hand, its Zα domain specifically recognizes and binds to the Z-form nucleic acid conformation, catalyzing A-to-I editing of endogenous Alu repeat element dsRNA, thereby disrupting its recognition interface with ZBP1. On the other hand, RNA secondary structure remodeling blocks the formation of the ZBP1–RIPK1 complex. When ADAR1 undergoes a Zα domain mutation, the editing efficiency of Alu elements is reduced, leading to the accumulation of unmodified Alu dsRNA and triggering apoptosis via the ZBP1–MAVS axis [136]. Additionally, in esophageal squamous cell carcinoma, ADAR2 modifies specific sites of IGFBP7 mRNA through A-to-I RNA editing, protecting it from proteolysis by matriptase, thereby maintaining the integrity of IGFBP7. The intact form of IGFBP7 can bind to and inhibit the activation of IGF1R, blocking the downstream Akt signaling pathway and thereby inducing apoptosis [137].

Epitranscriptomic modifications of pyroptosis
Pyroptosis is a form of pro-inflammatory programmed necrosis mediated by gasdermin family proteins. It is triggered by the recognition of intracellular pathogens or endogenous danger signals and is characterized by cell swelling, membrane rupture, and the release of pro-inflammatory factors, which distinguish it from other cell death modes such as apoptosis [138]. Its typical features include osmotic imbalance and cell lysis caused by the formation of transmembrane pores by gasdermin proteins, as well as the release of pro-inflammatory factors such as IL-1β and IL-18, ultimately leading to local inflammatory responses. The complete rupture of the cell membrane mediated by NINJ1 protein is a key terminal step [139]. At the regulatory level, the activation of gasdermin serves as the core hub. Inflammatory caspases such as caspase-1, 4/5/11, or caspase-3 can all initiate the death program by cleaving substrates such as GSDMD or GSDME to produce pore-forming fragments that target the cell membrane. Relevant pathways include the canonical pathway (inflammasomes such as NLRP3 activate caspase-1), the noncanonical pathway (cytoplasmic LPS directly activates caspase-4/5/11), and other pathways such as caspase-3 cleaving GSDME. These pathways mediate the pyroptosis process through cross-regulation [140]. Pyroptosis plays an important role in the body’s defense against pathogen infection. However, excessive or abnormal pyroptosis can also lead to tissue damage and the development of various diseases. Studies have found that pyroptosis is closely related to the occurrence and development of various liver diseases. In MASH, pyroptosis exacerbates liver injury through interactions with lipotoxicity, oxidative stress, and gut microbiota imbalance via canonical, noncanonical, and caspase-mediated pathways. In viral hepatitis and liver fibrosis, pyroptosis also drives the pathological process through mechanisms such as the release of pro-inflammatory factors [141]. Therefore, regulating pyroptosis-related pathways provides new therapeutic targets for the treatment of liver diseases.
m6A methylation and pyroptosis are closely regulated. m6A methylation can regulate cell sensitivity to pyroptosis by influencing the dynamic changes of NLRP3 inflammasome, thereby affecting the pyroptosis process. For example, METTL3 methylates NLRP3 mRNA in diabetes-related periodontitis through an IGF2BP3-dependent pathway. This modification significantly enhances the stability of NLRP3 mRNA, leading to its upregulation. As a key component of the inflammasome, the upregulation of NLRP3 activates caspase-1, promoting the production of pro-inflammatory cytokines such as IL-1β and IL-18, ultimately triggering pyroptosis in macrophages [142]. In Crohn’s disease research, METTL3 methylates circPRKAR1B, promoting its interaction with SPTBN1, inhibiting autophagy, and facilitating pyroptosis mediated by the NLRP3 inflammasome [143]. In liver fibrosis, METTL3 increases the expression of MALAT1 through m6A methylation. Subsequently, the binding of MALAT1 to PTBP1 promotes the degradation of USP8 mRNA, reducing USP8 expression. This inhibits the ubiquitination and degradation of TAK1, increasing the stability and expression of TAK1 protein, activating the NF-κB signaling pathway and NLRP3 inflammasome, promoting the release of inflammatory factors such as IL-1β and IL-18, and ultimately triggering pyroptosis in macrophages [144]. In diabetic cardiomyopathy research, METTL14 upregulates the m6A modification level of terminal differentiation-induced noncoding RNA (TINCR), enhancing its stability and promoting its expression. TINCR directly interacts with NLRP3 mRNA, effectively increasing the stability of NLRP3 mRNA and preventing its premature degradation, thereby upregulating NLRP3 expression and promoting pyroptosis [145]. WTAP is highly expressed in diabetic nephropathy, promoting the m6A methylation and upregulation of NLRP3 mRNA, thereby activating the NLRP3 inflammasome and promoting cell pyroptosis and inflammatory response. Additionally, the histone acetyltransferase p300 promotes WTAP transcription, further influencing WTAP’s regulation of cell pyroptosis [146]. In sepsis-induced acute kidney injury, IGF2BP1 recognizes the m6A modification site in the 3′-UTR region of E2F1 mRNA, stabilizing E2F1 mRNA and increasing E2F1 protein expression. As a transcription factor, E2F1 binds to the MIF promoter, promoting MIF transcription. MIF, as a subunit of the NLRP3 inflammasome, interacts with NLRP3 to activate the NLRP3 inflammasome and trigger cell pyroptosis [147]. YTHDF1 can recognize the m6A modification site of WWP1 mRNA, promoting its translation and upregulating WWP1 expression. As an E3 ubiquitin ligase, WWP1 promotes the ubiquitination of NLRP3, thereby inhibiting the activation of the NLRP3 inflammasome and caspase-1-mediated GSDMD cleavage, ultimately inhibiting cell pyroptosis [148]. FTO can regulate the expression of MEG3 by affecting the stability of MEG3 RNA, and also promotes neuronal pyroptosis through the NLRP3/caspase-1/GSDMD signaling pathway, thereby affecting ischemic brain injury [149] (Fig. 6).
ac4C modification, as a core form of RNA acetylation, demonstrates regulatory potential in the process of pyroptosis through a dynamic regulatory network mediated by NAT10. Studies have shown that NAT10 is significantly downregulated in neutrophils of sepsis mice, leading to decreased stability and protein expression levels of ULK1 mRNA. This failure to effectively inhibit the STING-IRF3 signaling pathway results in increased activation of the NLRP3 inflammasome, which promotes pyroptosis. In contrast, overexpression of NAT10 can significantly inhibit pyroptosis by restoring ULK1 expression, improving the survival rate of sepsis mice, and reducing lung tissue damage [150]. In severe acute pancreatitis, NAT10 also catalyzes the ac4C modification of NLRP3 mRNA, enhancing its stability and expression levels. This promotes the overactivation of the NLRP3 inflammasome and accelerates pyroptosis in pancreatic cells [151]. In a kidney injury model, NAT10 maintains the ac4C modification of ULK1 mRNA, extending its half-life and allowing ULK1 to continuously inhibit the NLRP3 inflammasome, thereby reducing pyroptosis in renal tubular epithelial cells [152].
Research on A-to-I RNA editing in pyroptosis remains limited. Studies have found that the expression of ADAR1 in peripheral blood mononuclear cells of sepsis patients is downregulated and significantly negatively correlated with the levels of pyroptosis markers. Further investigation into the mechanism revealed that ADAR1 can regulate the biogenesis of miR-21 through its RNA editing function, thereby negatively regulating the expression of A20, a negative regulator of the NLRP3 inflammasome. Therefore, downregulation of ADAR1 leads to upregulation of miR-21 and downregulation of A20, ultimately promoting the activation of the NLRP3 inflammasome and the occurrence of pyroptosis [153]. Additionally, in the pathological process of neuropathic pain (NP), ADAR3 is closely associated with pyroptosis. In a mouse model of NP induced by chronic constriction injury, ADAR3 expression levels are reduced, while overexpression of ADAR3 can alleviate NP symptoms and reduce inflammatory responses. Further research has shown that ADAR3 can directly target NLRP3 and negatively regulate its expression, inhibiting the activation of the NLRP3 inflammasome, thereby reducing the expression of pyroptosis-related proteins such as caspase-1, IL-1β, IL-18, and GSDMD, and alleviating neuronal pyroptosis [154].
Although there is currently no direct evidence that m5C modification and pseudouridylation affect pyroptosis, given the important roles of these two modifications in RNA function regulation, we can reasonably speculate that both m5C modification and pseudouridylation may indirectly regulate the pyroptosis process by acting on RNA related to pyroptosis, thereby exerting influence at the levels of expression and function. Specifically, m5C modification is highly likely to regulate the mRNA encoding pyroptosis-related proteins, affecting their stability or translation efficiency, thereby intervening in the pyroptosis process. Additionally, m5C modification may also indirectly influence the release of inflammatory factors during pyroptosis by regulating RNA related to the inflammatory response, thereby affecting the occurrence and development of pyroptosis. In terms of pseudouridylation, it may regulate the activation of inflammasomes by affecting the stability or translation efficiency of mRNA related to inflammatory factors, thereby playing a key role in the regulation of pyroptosis. Moreover, the level of pseudouridylation in cellular RNA is mainly regulated by enzymes such as PUS. Abnormal expression or activity changes of these regulatory factors are likely to interfere with the normal regulation of pyroptosis. However, the specific mechanisms by which m5C modification and pseudouridylation affect pyroptosis still need to be further revealed through more research.

Epitranscriptomic modifications of PCD in liver diseases

Epitranscriptomic modifications of PCD in MASLD
MASLD, formerly known as nonalcoholic fatty liver disease (NAFLD), is one of the most common chronic liver diseases worldwide. Characterized by hepatic steatosis, defined as the accumulation of triglycerides in more than 5% of hepatocytes, MASLD can progress to MASH and hepatic fibrosis if left untreated. Some patients may eventually develop cirrhosis, and in advanced stages, HCC may arise as a severe complication, significantly impacting patients’ health and quality of life [155]. The pathogenesis of MASLD is complex, involving the interplay of various factors including insulin resistance, lipid metabolism disorders, oxidative stress, mitochondrial dysfunction, and imbalance of the gut–liver axis [156]. Current treatments for MASLD are primarily based on lifestyle interventions, including dietary adjustments and increased physical activity. Additionally, drugs such as GLP-1 receptor agonists, PPAR agonists, and FXR agonists have shown promising therapeutic potential through mechanisms that regulate lipid metabolism, reduce inflammation, and improve insulin sensitivity [157].
Epitranscriptomic modifications, as an essential component of gene expression regulation, are closely and intricately linked to PCD, jointly influencing the development and progression of MASLD. METTL3, through m6A modification, regulates the expression of FAS, which is one of the key mechanisms inducing hepatocyte apoptosis in MASLD. Research has shown that METTL3 is upregulated in a high-fat diet-induced MASLD mouse model, and its silencing significantly improves plasma lipid levels and reduces hepatic lipid accumulation. In vitro experiments have demonstrated that silencing METTL3 can increase hepatocyte viability, inhibit apoptosis, and reduce lipid concentrations, while downregulating the levels of markers associated with lipogenesis. Further mechanistic studies have revealed that METTL3 promotes the m6A methylation of FAS mRNA, enhancing its stability and thereby increasing FAS expression. Given that FAS plays a crucial role in fatty acid synthesis, its upregulated expression may exacerbate lipid deposition and metabolic disorders within hepatocytes, ultimately leading to increased apoptosis [158]. Moreover, FTO, the fat mass and obesity-associated gene, is closely related to the occurrence and development of MASLD. FTO can promote the development of MASLD by increasing IR, oxidative stress, and lipid deposition within hepatocytes. Studies have shown that FTO is upregulated in MASLD animal models, and its overexpression exacerbates oxidative stress and lipid deposition in hepatocytes, leading to hepatocyte dysfunction and increased apoptosis. Additionally, FTO increases lipid accumulation by promoting the maturation of SREBP-1c, which in turn enhances the transcriptional activity of lipid droplet-associated protein CIDEC, further aggravating lipid accumulation within hepatocytes. This lipid accumulation and metabolic disorder ultimately leads to IR and apoptosis in hepatocytes, thereby promoting the progression of MASLD [159]. Furthermore, FTO can reduce the m6A methylation level of the SLC7A11 gene, decreasing its stability and inhibiting intracellular GSH maintenance. This leads to a decline in cellular antioxidant capacity, rendering cells more susceptible to ferroptosis and thus promoting the progression of MASLD. For instance, inhibiting FTO with arbutin can suppress cellular ferroptosis, reduce hepatic lipid deposition, and ameliorate MASLD symptoms [160]. Recent studies have also demonstrated that YTHDF2 can participate in the disease progression of MASLD by regulating ferroptosis. Specifically, YTHDF2 recognizes m6A-modified target mRNAs and promotes the degradation of mRNAs encoding key ferroptosis factors GPX4 and SLC7A11, thereby enhancing ferroptosis and exacerbating liver injury. ZHX2 can reverse this process by inhibiting the transcription of YTHDF2 [161]. This ZHX2–YTHDF2–ferroptosis regulatory axis provides a new perspective for elucidating the pathogenesis of MASLD and developing therapeutic strategies (Fig. 7).

Epitranscriptomic modifications of PCD in viral hepatitis
Viral hepatitis, a global public health challenge, is caused by five types of hepatitis viruses: A, B, C, D, and E. These viruses induce infections and varying degrees of liver diseases through different transmission routes. Among them, hepatitis A and E are predominantly acute infections, whereas hepatitis B, C, and D are more likely to become chronic, potentially progressing to liver cirrhosis or cancer, which severely impacts patients’ quality of life and life expectancy [162]. Viral hepatitis is characterized by its high contagiousness, complex transmission routes, and high incidence rate. Although the widespread use of vaccines has achieved significant success in preventing certain types of viral hepatitis, there are still many susceptible individuals in areas with low vaccination coverage and among high-risk groups. For those types of viral hepatitis that are not preventable by vaccines, the application of direct-acting antiviral drugs has increased the cure rate, bringing new hope to patients. However, the high cost and drug resistance issues of these drugs still need attention [163]. Therefore, continuously strengthening vaccination, optimizing treatment plans, and improving public health strategies are crucial for reducing the burden of viral hepatitis and achieving global health goals.
In the complex pathological process of viral hepatitis, epitranscriptomic modifications profoundly affect PCD and immune responses by regulating the virus–host interaction network. For instance, in HBV, m6A modification regulates the maturation of miR-146a-5p through METTL3, thereby participating in the process of hepatocyte apoptosis. Specifically, m6A modification mediated by METTL3 can promote the maturation of miR-146a-5p, which in turn enhances HBV replication, promotes the release of inflammatory factors, and exacerbates hepatocyte apoptosis. This modification facilitates the production of mature miR-146a-5p, creating a favorable microenvironment for viral replication. It not only intensifies hepatocyte apoptosis and inhibits cell proliferation but also exacerbates liver injury by increasing the levels of pro-inflammatory cytokines such as TNF-α and IL-6 [164]. In addition, regarding m5C modification, the m5C modification of HBV mRNA catalyzed by the host methyltransferase NSUN2 plays a crucial regulatory role. This modification affects disease progression in two dimensions. First, it promotes the nuclear export of viral RNA and the translation of HBx protein, ensuring viral genome replication and protein synthesis. Second, inhibiting RIG-I recognition of viral RNA blocks the transcriptional activation of antiviral factors such as IFN-β, thereby weakening the innate immune response [165]. This dual effect may suppress immune-mediated PCD, such as apoptosis and pyroptosis, and create conditions for immune evasion of persistent HBV infection. However, the specific molecular pathways by which various epitranscriptomic modifications, including m5C, regulate programmed death in viral hepatitis still require further in-depth functional studies to be elucidated (Fig. 7).

Epitranscriptomic modifications of PCD in ALD
ALD is one of the most common chronic liver diseases globally. Chronic excessive alcohol consumption, the primary etiology of ALD, induces hepatic metabolic disturbances and a cascade of pathological alterations. The pathogenesis of ALD is intricate, involving the interplay of multiple factors. The metabolism of ethanol in the liver generates acetaldehyde and ROS, which directly injure hepatocytes, impair mitochondrial function, and trigger oxidative stress. These changes activate hepatic stellate cells, promoting hepatic fibrosis development. Moreover, alcohol disrupts the intestinal barrier function, leading to gut microbiota dysbiosis. As a result, intestinal microbes and their metabolites translocate to the liver, activating the immune system and eliciting inflammatory responses. Inflammation plays a pivotal role in the progression of ALD, with the infiltration of various immune cells exacerbating hepatocyte injury. In recent years, research has also highlighted the significant roles of the gut–liver axis and adipose tissue–liver axis in the occurrence and development of ALD. The clinical management of ALD primarily relies on alcohol abstinence, nutritional support, and symptomatic treatment. Future research needs to explore the pathogenesis of ALD and develop effective treatment strategies to improve patient outcomes.
In ALD, epitranscriptomic modifications can participate in the process of PCD by regulating the expression of target genes, thereby affecting disease progression. For example, in alcohol-associated steatohepatitis (ASH), METTL3 promotes the biogenesis of miR-34a-5p through m6A modification of pri-miR-34a. MiR-34a-5p then inhibits SIRT1 expression, thereby enhancing pyroptosis, releasing inflammatory factors in Kupffer cells, and driving the progression of ASH. This process can be reversed by silencing METTL3 [166]. In addition, alcohol can induce E3 ubiquitin ligase STUB1-mediated ubiquitination and degradation of METTL3, while disrupting the protective binding of HSP70 to METTL3, decreasing hepatic m6A modification levels. The reduction in m6A modifications weakens the degradation of target gene ATF3 mRNA by its reader protein YTHDF2, resulting in upregulated ATF3 expression. This exacerbates hepatic steatosis, neutrophil infiltration, and the release of inflammatory factors such as TNF-α and IL-1β. These pathological processes, accompanied by the activation of mechanisms related to PCD, ultimately drive the progression of ASH [167]. Thus, m6A modifications interact intricately with PCD and inflammatory responses in liver diseases by regulating the stability of downstream target genes, influencing disease occurrence and development. While the roles of other epitranscriptomic modifications in PCD during ALD remain under investigation, their potential regulatory mechanisms offer novel avenues for unraveling disease pathogenesis and developing innovative therapeutic approaches. The complex interplay between ALD, epitranscriptomic modifications, and PCD will be clarified with the continuous advancement of RNA modification detection technologies (Fig. 7).

Epitranscriptomic modifications of PCD in hepatic fibrosis
Hepatic fibrosis is a pathological process characterized by excessive deposition of extracellular matrix (ECM) following chronic liver injury, such as MASLD, ALD, and viral hepatitis, with the activation of HSCs as the core driving mechanism [168]. Quiescent HSCs are transformed into myofibroblasts under the stimulus of injury signals, and by secreting collagen, they disrupt the liver architecture, which may eventually progress to cirrhosis and even hepatocellular carcinoma [169]. Current treatments include anti-inflammatory, antioxidant, and antifibrotic drugs, as well as etiology-targeted approaches (e.g., antiviral therapy, metabolic improvement, etc.). Recently, it has been discovered that targeting the Wnt/β-catenin, YAP signaling pathways, and lactate dehydrogenase A can exert therapeutic effects by modulating HSC glycolysis, sensitivity to apoptosis, and senescence. Strategies based on cellular interventions (e.g., inducing HSC quiescence, apoptosis, or senescence) are also under investigation [170]. However, treatment of hepatic fibrosis still faces challenges such as the complexity of disease mechanisms, limited efficacy of single-target drugs due to the involvement of multiple cells and pathways, drug side effects, and the heterogeneity of fibrosis [169].
In hepatic fibrosis, activated HSCs are the primary source of ECM. Inhibition or clearance of activated HSCs can reverse hepatic fibrosis, and eliminating activated HSCs is considered an effective antifibrotic strategy [171]. METTL3 plays an important role in the progression of hepatic fibrosis by regulating related molecules that affect pyroptosis and inflammatory responses. Studies have found that METTL3 increases the lncRNA MALAT1 level through m6A modification. MALAT1 interacts with PTBP1, promoting the degradation of USP8 mRNA, thereby enhancing the stability of TAK1 protein. As a key regulatory kinase, changes in TAK1 activity significantly influence cellular inflammatory responses and pyroptosis. When TAK1 is activated, it promotes pyroptosis and inflammatory responses in macrophages, activates HSCs, and further exacerbates hepatic fibrosis [144]. After treatment with ferroptosis inducers, the level of m6A modification in HSCs significantly increases, mainly owing to the upregulation of the methyltransferase METTL4 and the downregulation of the demethylase FTO. m6A modification stabilizes BECN1 mRNA to activate autophagy, while YTHDF1, as a key m6A reader protein, recognizes and binds to the m6A binding sites on BECN1 mRNA, thereby enhancing its stability and promoting autophagy activation, ultimately leading to ferroptosis in HSCs. This mechanism has been verified in animal models and clinical samples: inhibiting m6A modification can block autophagy activation and weaken ferroptosis, providing a potential therapeutic target for hepatic fibrosis [172]. Similarly, downregulation of FTO plays a key role in DHA-induced ferroptosis in HSCs. Specifically, DHA increases m6A modification by reducing FTO expression, thereby stabilizing BECN1 mRNA, which is a key regulator of autophagy. YTHDF1 binds to BECN1 mRNA, extending its half-life, thereby enhancing autophagy and ultimately leading to ferroptosis in HSCs [173]. During the progression of hepatic fibrosis, YTHDF2 is significantly upregulated in liver tissue. It binds to and regulates the translation of ACSL4 mRNA in an m6A modification-dependent manner, promoting the activation and ferroptosis of HSCs. Specifically, YTHDF2 recognizes the m6A modification sites on ACSL4 mRNA, enhances its stability, and promotes its translation, thereby increasing the expression of ACSL4 protein. ACSL4, as a key regulator of ferroptosis, can exacerbate the ferroptosis process in HSCs and promote the progression of hepatic fibrosis when upregulated [174]. In addition, the absence of IGF2BP3 in HSCs reduces m6A modification on Jag1, inhibiting the Notch pathway and decreasing the expression of the downstream molecule Hes1. As a transcription factor, the reduced expression of Hes1 suppresses the expression of GPX4. GPX4 is a negative regulator of ferroptosis, and its decreased expression promotes ferroptosis in HSCs, thereby alleviating hepatic fibrosis [175] (Fig. 7).

Epitranscriptomic modifications of PCD in HCC
HCC is one of the most common malignancies with high incidence and mortality rates worldwide, posing a serious threat to human health. As a primary liver cancer originating from the liver, its pathogenesis is extremely complex, involving the interplay of multiple factors that ultimately disrupt the balance between the inactivation of tumor suppressor genes and the activation of oncogenes, leading to the abnormal activation of molecular signaling pathways and the dysregulation of HCC cell differentiation and angiogenesis [176]. Owing to the insidious onset of HCC and the neglect of early screening, patients are often diagnosed at an advanced stage, resulting in poor treatment outcomes and a 5-year survival rate of less than 20% [177]. In recent years, with the continuous progress of medical technology, new systemic treatment methods have emerged, bringing new hope to patients with HCC. However, the survival rate of patients with HCC remains a cause for concern.
m6A modification plays a significant role in HCC apoptosis by regulating key gene expression. Studies have shown that METTL3 is significantly upregulated in HCC and promotes the degradation of tumor suppressor gene SOCS2 mRNA in a YTHDF2-dependent manner, thereby inhibiting its expression and subsequently suppressing apoptosis and promoting tumor progression. Additionally, METTL3 collaborates with YTHDF1 to enhance the translation of Snail protein, driving epithelial–mesenchymal transition and metastasis, indirectly reducing cell apoptosis sensitivity. Conversely, downregulation of METTL14 leads to impaired maturation of miR-126, potentially weakening its pro-apoptotic function and thereby promoting HCC metastasis. Additionally, FTO is overexpressed in HCC, promoting the translation of PKM2 mRNA by removing its m6A modification, enhancing glycolysis and tumor growth, with high PKM2 expression associated with apoptosis resistance. Downregulation of YTHDF2 results in increased mRNA stability of IL11 and SERPINE2, remodeling the tumor microenvironment and inhibiting apoptosis. IGF2BPs, as m6A recognition proteins, stabilize the mRNA of oncogenes such as MYC, further reinforcing apoptosis resistance [178]. m6A modification also plays a crucial role in ferroptosis in HCC. Studies have shown that ferroptosis inducers can significantly increase the overall level of m6A modification in HCC cells, primarily through upregulation of the m6A methyltransferase WTAP. WTAP-mediated m6A modification promotes the translation of autophagy-related gene ATG5 mRNA through the recognition protein YTHDC2, thereby activating autophagy and ultimately leading to ferroptosis. In in vivo models, knockdown of YTHDC2 significantly weakens the inhibitory effect of ferroptosis inducers on HCC tumor growth, indicating that m6A modification has potential therapeutic applications in HCC treatment through regulating autophagy-dependent ferroptosis [179] (Fig. 7).
In addition to m6A, m5C modification also plays a key role in regulating various types of cell death in HCC. Studies have shown that in HCC, m5C modification regulates the ferroptosis process through the NSUN2/YBX1/RNF115 axis. NSUN2 catalyzes the m5C modification of RNF115 mRNA. YBX1 recognizes the modified site in its 3′-UTR and interacts with the translation initiation factor EIF4A1 to promote mRNA circularization and translation, significantly upregulating the expression of RNF115 protein. RNF115, as an E3 ubiquitin ligase, ubiquitinates mitochondrial protein DHODH at K27, inhibiting its autophagic degradation and thereby maintaining mitochondrial function and resisting ferroptosis [180]. In HCC, ac4C modification regulates apoptosis through the NAT10/eEF2/HMGB2 axis. Studies have shown that NAT10 is markedly upregulated in HCC tissues and selectively installs ac4C within the CDS of HMGB2 mRNA. eEF2, as an ac4C reader, promotes the translation of HMGB2 mRNA by recognizing the ac4C site, upregulating the expression of HMGB2 protein. The high expression of HMGB2 inhibits the activation of key proteins in the mitochondrial apoptosis pathway, reducing the release of cytochrome C and the caspase cascade, thereby suppressing apoptosis in HCC cells. Targeting inhibition of NAT10 or using panobinostat can reduce the level of ac4C modification, downregulate HMGB2 expression, restore mitochondrial membrane potential, and activate caspase-3/9, significantly inducing apoptosis in HCC cells. This study reveals a new mechanism by which ac4C modification inhibits apoptosis in HCC cells through epigenetic regulation, providing a theoretical basis for targeting the NAT10-ac4C axis for anti-HCC therapy [50] (Fig. 7).
In summary, epitranscriptomic modifications, as a crucial posttranscriptional regulatory mechanism, profoundly influence the activation and execution of multiple PCD pathways in liver diseases. These modifications finely regulate the fate of hepatocytes by altering the RNA stability, localization, splicing, and translation efficiency of key PCD-related genes. Elucidating the regulatory network of epitranscriptomic modifications in PCD not only provides a novel perspective for understanding the pathogenesis of liver diseases but also paves a highly promising path for developing new diagnostic biomarkers and therapeutic strategies targeting the RNA modification–PCD axis.

Drugs targeting epitranscriptomic modifications
This section thoroughly reviews the latest advancements in developing drugs that target key regulatory proteins associated with epitranscriptomic modifications. These drugs can significantly impact RNA metabolism and gene expression by modulating these modifications. They have shown impressive therapeutic efficacy in models of MASLD, HCC, CRC, and other related diseases, opening up new avenues and strategies for precision medicine in these areas. However, further in-depth and detailed research is urgently needed to fully explore the clinical potential of these drugs, including their safety, efficacy, and pharmacokinetic properties.

Drugs targeting m6A modification

FTO demethylase inhibitors
FTO is the first identified m6A demethylase, and its dysregulation is closely associated with various pathological processes. Drugs targeting FTO can significantly increase the m6A methylation levels of key disease-related RNAs within cells by specifically inhibiting its demethylase activity, thereby significantly suppressing disease progression. Entacapone, an FDA-approved drug initially used for treating Parkinson’s disease, has recently been repurposed as an effective FTO inhibitor in recent studies. In MASLD models, Entacapone reduces the stability of SREBF1 and ChREBP mRNA by inhibiting FTO activity, thereby decreasing the expression of lipogenic genes. This mechanism effectively reduces hepatic triglyceride accumulation and alleviates high-fat diet-induced hepatic steatosis, providing strong support for a potential therapeutic strategy for MASLD [181]. Arbutin, a natural antioxidant extracted from bearberry (Arctostaphylos uva-ursi), also possesses various pharmacological properties, including antioxidant, anti-inflammatory, and antibacterial effects. Studies on MASLD have shown that arbutin can inhibit FTO activity to enhance the m6A methylation level of SLC7A11, stabilize its mRNA, and promote protein expression. This mechanism helps to inhibit ferroptosis, reduce oxidative stress, and decrease lipid accumulation, ultimately improving MASLD symptoms. Arbutin also regulates energy metabolism by increasing mitochondrial content and ATP levels, and improves insulin resistance, further supporting its potential therapeutic value in MASLD [119]. CS2, as an FTO inhibitor, holds potential application value in the treatment of HCC. CS2 increases the m6A methylation level of GPNMB mRNA by inhibiting FTO activity, thereby promoting its degradation and reducing GPNMB protein accumulation. Further studies have shown that CS2 enhances the sensitivity of HCC cells to immune checkpoint inhibitors and sorafenib treatment, indicating that CS2 not only directly inhibits the proliferation and metastasis of HCC cells but also enhances the efficacy of immunotherapy by modulating the tumor immune microenvironment [182]. Rhein, a major component of several traditional Chinese medicines such as Rheum palmatum, aloe, and Sennae folium, possesses a variety of pharmacological properties, including anti-inflammatory, anti-tumor, anti-fibrotic, and antioxidant effects. Rhein has been identified as a selective FTO inhibitor that competitively binds to the catalytic domain of FTO, effectively suppressing its m6A demethylase activity on mRNA substrates. However, it has relatively poor selectivity and may affect other epigenetic proteins [183].
In addition, MA is a nonsteroidal anti-inflammatory drug that has been proven to be a highly selective inhibitor of FTO. MA can dose-dependently inhibit FTO′s demethylation activity on mRNA containing m6A, but it does not inhibit ALKBH5-mediated m6A demethylation reactions, even at high concentrations, demonstrating high selectivity for FTO [184]. On the basis of the structural complex of MA bound to FTO, researchers have designed MA analogs FB23-2 and Dac51. Compared with MA, their activities are significantly enhanced. FB23-2, which introduces an additional five-membered heterocyclic ring on the dichlorobenzene of MA, acts as a highly selective FTO inhibitor targeting m6A modifications on mRNA. This compound minimally impacts other proteins and normal human bone marrow cell proliferation. In cellular models, FB23-2 inhibits proliferation of NB4 and MONOMAC6 cell lines, elevates m6A levels in the mRNA transcriptome, accelerates myeloid differentiation of NB4 cells, and induces apoptosis. In animal experiments, it can inhibit the proliferation of primary acute myeloid leukemia (AML) cells, prolong the survival of patient-derived AML xenografted mice, and reduce the malignancy of leukemia; Dac51 can increase the m6A abundance in B16-OVA cell mRNA, significantly reduce the glycolytic capacity and metabolite levels of tumor cells, and has low toxicity to normal cells. In terms of immune regulation, it can enhance IFN-γ production in patient-derived organoids. When combined with immune checkpoint inhibitors, it suppresses MC38 tumor growth, elicits a robust and durable memory T-cell response, and markedly curtails tumor recurrence [185] (Table 6).

ALKBH5 demethylase inhibitors
ALKBH5 is another m6A demethylase associated with tumor immune evasion and metabolic reprogramming. MV1035 is a small-molecule compound based on an imidazobenzoxazine-5-sulfone scaffold. In the U87 glioblastoma cell line, although MV1035 was initially designed as a sodium channel blocker, its inhibitory effects on cell migration and invasion are independent of sodium channel blockade. Through virtual screening of the entire proteome using the SPILLO-PBSS software, ALKBH5 was identified as a potential target of MV1035. Further studies have shown that MV1035 can compete with the 2-oxoglutarate binding region of ALKBH5, inhibit its catalytic activity, and lead to an increase in methylated RNA levels [186]. In addition, 20m is a novel ALKBH5 inhibitor targeting mRNA demethylation, obtained through fluorescence polarization screening, structural optimization, and structure–activity relationship analysis. It exhibits extremely high inhibitory activity against ALKBH5, with an IC50 value of only 0.021 μM in fluorescence polarization assays and high selectivity, showing weak inhibitory effects on FTO and other members of the AlkB subfamily. Critically, cellular studies demonstrate that 20m stabilizes ALKBH5 substrates in HepG2 cells and significantly elevates m6A levels on cellular mRNA [187] (Table 6).

METTL3 methyltransferase inhibitors
METTL3 is the core catalytic subunit of the m6A methylation complex, and its overexpression is associated with tumorigenesis. STM2457 is a highly selective catalytic inhibitor of METTL3 that effectively suppresses m6A methyltransferase activity on mRNA, thereby reducing cellular m6A methylation levels. In metabolic dysfunction-associated steatotic liver disease-associated hepatocellular carcinoma (MASLD-HCC), STM2457 inhibits METTL3 activity to downregulate the expression of cholesterol biosynthesis-related genes, thereby reducing the accumulation of cholesterol and cholesteryl esters. STM2457 inhibits tumor cell proliferation and enhances the antitumor immune response by restoring the function of cytotoxic CD8+ T cells. Additionally, STM2457 has been shown to significantly improve therapeutic efficacy when combined with anti-PD-1 immune checkpoint blockade therapy, inhibiting the growth of MASLD-HCC and providing a new strategy for its treatment [188]. STC-15 is an orally bioavailable, highly selective METTL3 inhibitor. Preclinical studies have found that multiple AML cell lines and AML patient-derived xenograft models are sensitive to the inhibitory effects of STC-15. It can downregulate the expression of the anti-apoptotic protein BCL-2 and demonstrates a high degree of synergy when used in combination with the BCL-2 inhibitor venetoclax, significantly extending animal survival and showing promising therapeutic efficacy against AML [189] (Table 6).

Reader protein inhibitors
m6A-binding proteins can recognize m6A modifications and regulate RNA metabolism, including proteins from the YTH and IGF2BP families. Salvianolic acid (SAC) is a natural compound extracted from the traditional Chinese medicine Salvia miltiorrhiza, which has pharmacological properties such as anti-inflammatory, antioxidant, and immune regulatory effects. SAC can act as a selective inhibitor of YTHDF1, reducing its binding to and translational regulation of m6A-modified mRNA, thereby exerting anti-inflammatory effects. In a study, SAC was used to test its effects on gluten-induced intestinal inflammation. The results showed that SAC significantly reduced the inflammatory response caused by gluten digestion products, decreasing the expression of inflammation-related proteins and mRNA such as XPO1, NFκB, and IL8. Additionally, SAC-treated mice exhibited protective effects against inflammation in intestinal morphology, characterized by the restoration of the villus height-to-crypt depth ratio and reduced inflammatory cell infiltration. These findings indicate that SAC, as a YTHDF1 inhibitor, showed good anti-inflammatory effects in both in vitro and in vivo models without any observed significant toxicity or side effects [190]. CWI1-2 is a small-molecule IGF2BP2 inhibitor derived from NSC69557, which competitively inhibits the binding of IGF2BP2 to RNA by binding to a hydrophobic pocket near the KH4 domain of IGF2BP2. In vitro experiments showed that CWI1-2 has significant inhibitory effects on AML cells, inducing cell differentiation and apoptosis, reducing glutamine uptake, impairing mitochondrial function, and decreasing ATP production, with greater sensitivity in AML cells with high IGF2BP2 expression. In vivo experiments indicated that CWI1-2 effectively inhibited leukemia cell engraftment, delayed leukemia onset, and prolonged mouse survival without significant side effects. Moreover, CWI1-2 exhibited synergistic inhibitory effects on AML cell growth when used in combination with daunorubicin or homoharringtonine, making it a highly promising lead compound for AML treatment [191] (Table 6).

Drugs targeting m5C modification
SU056 is a small-molecule inhibitor targeting YBX1 that disrupts its binding to the m5C-modified RNF115 mRNA, thereby weakening its interaction with EIF4A1. This interference suppresses RNF115 mRNA circularization and translation, and reduces RNF115 expression. The decrease in RNF115 attenuates its K27 ubiquitination-mediated control of DHODH, promotes autophagic degradation of DHODH, and enhances hepatocellular carcinoma cell sensitivity to ferroptosis, ultimately inhibiting HCC progression [180]. MALAT1-IN1, a preclinical-stage small-molecule inhibitor targeting MALAT1, abrogates its m5C modification, destabilizing the transcript and downregulating SLC7A11 expression to potentiate ferroptosis sensitivity. In HCC models, combined treatment with MALAT1-IN1 and sorafenib markedly suppresses tumor growth and overcomes drug resistance [192]. At present, therapeutic agents that directly target RNA m5C modification in liver diseases remain exceedingly scarce. Nevertheless, small-molecule compounds that modulate m5C writers or readers hold promise for correcting aberrant RNA methylation and ameliorating pathological processes such as hepatic fibrosis and hepatocellular carcinoma, offering a novel conceptual framework for the development of epitranscriptomic therapies (Table 7).

Drugs targeting ac4C modification
The core strategy for targeting ac4C modification is to inhibit NAT10. Direct inhibition of NAT10′s acetyltransferase activity reduces ac4C modification on mRNA. By decreasing ac4C levels, the stability and translation efficiency of target gene mRNA are reduced, thereby inhibiting tumor-related signaling pathways [50, 193]. Remodelin is a small molecule NAT10 inhibitor that can inhibit the acetyltransferase activity of NAT10 by interacting with its acetyl-CoA binding site, thereby reducing the ac4C modification levels on RNA. In vitro experiments have shown that Remodelin can significantly inhibit the proliferation of bladder cancer cells, suppress cell activity in a dose-dependent manner, and induce apoptosis. In vivo experiments further confirmed that Remodelin, in combination with cisplatin, can significantly inhibit tumor growth without causing significant pathological changes [194]. In addition, the histone deacetylase (HDAC) inhibitor panobinostat mainly regulates gene expression and cell function by inhibiting HDAC activity and increasing the acetylation levels of histones and nonhistone proteins. Although panobinostat’s core mechanism of action is HDAC inhibition, studies have found that it can also effectively inhibit NAT10-mediated ac4C modification. Specifically, panobinostat can bind tightly to the catalytic pocket of NAT10, blocking its acetyltransferase activity and significantly reducing the ac4C modification levels of RNA in HCC cells. Moreover, after treatment with panobinostat, protein synthesis and HMGB2 protein expression in HCC cells were significantly inhibited, thereby weakening the proliferation and invasion capabilities of HCC cells. In vivo experiments showed that panobinostat inhibited the growth of HCC xenografts in a dose-dependent manner, significantly reduced the number of intrahepatic tumor nodules and pulmonary metastases, and did not cause significant weight loss or other adverse reactions, demonstrating good anti-HCC efficacy and safety [50] (Table 7).

Drugs targeting pseudouridylation
Pyrazofurin is a drug with a unique mechanism of action that was initially investigated as an antitumor and antiviral agent. Recent studies have found that pyrazofurin can reduce the overall Ψ levels in cellular RNA by inhibiting the PUS activity of DKC1, thereby weakening the stability of ribosomal protein mRNA targeted by DKC1 and decreasing the expression of these proteins, thus inhibiting the proliferation of CRC cells. Moreover, pyrazofurin can be combined with the MEK1/2 inhibitor trametinib to exert a synergistic inhibitory effect on the growth of CRC cells, providing a new strategy for treating CRC [195] (Table 7).

Drugs targeting A-to-I editing
Drugs targeting A-to-I editing mainly focus on the development of ADAR1 inhibitors. These inhibitors specifically block the activity of the ADAR1 protein, thereby inhibiting its A-to-I RNA editing function and effectively suppressing abnormal RNA editing events in tumor cells. A small-molecule compound, 8-azaadenosine, has been shown to have significant inhibitory activity against A-to-I editing. In thyroid cancer research, 8-azaadenosine significantly reduces the invasiveness of cancer cells by inhibiting ADAR1 activity. Specifically, this inhibitor significantly decreases the proliferation, three-dimensional growth, invasion, and migration capabilities of thyroid cancer cells, with inhibitory effects verified in both two-dimensional and three-dimensional in vitro cell culture systems. Moreover, in in vivo xenograft tumor models, 8-azaadenosine treatment significantly delays tumor growth without any observed significant adverse reactions [196] (Table 7).

Summary of clinical trials involving drugs targeting epitranscriptomic modifications
In the preceding sections, we discussed the pivotal role of epitranscriptomic modifications in the pathologies of various diseases and summarized relevant clinical trial data to further evaluate the clinical potential of targeting epitranscriptomic modifications. These studies encompass a range of therapeutic applications, including the treatment of diverse diseases, research methodologies, and stages of clinical development. The compiled data provide valuable insights into the therapeutic efficacy and advancements of epitranscriptomic modification-targeted approaches, highlighting their clinical significance and potential future directions for research and therapeutic applications (Table 8).

Conclusions
Epitranscriptomic modifications, as one of the core mechanisms for gene expression regulation, precisely control the metabolic fate of RNA through dynamic and reversible chemical modifications, ranging from splicing, stability, translational efficiency, to subcellular localization, forming a complex posttranscriptional regulatory network. In recent years, breakthroughs in high-throughput sequencing technologies and single-cell spatial omics have not only revealed the key roles of epitranscriptomic modifications in various diseases such as obesity, immune dysregulation, neurodegenerative diseases, and malignancies but also highlighted their fine-tuning ability in PCD. PCD, as an active and orderly process of cell death, is crucial for maintaining intracellular homeostasis, tissue and organ development, and normal physiological functions of the body. The role of epitranscriptomic modifications in PCD pathways is exceptionally complex and closely intertwined with the occurrence and development of liver diseases. During the progression of liver diseases, epitranscriptomic modifications may become an important driving factor for disease deterioration through certain abnormal modification changes, accelerating hepatocyte damage and lesions. However, at the same time, their unique regulatory mechanisms also provide potential targets for treating liver diseases, bringing new hope for developing innovative therapeutic strategies.
Although current research on epitranscriptomic modifications has yielded fruitful results, most studies have primarily focused on in-depth analysis of the molecular mechanisms of these modifications and exploration of their regulatory pathways and modes of action within cells. However, relevant research is still relatively limited in practical application, especially in developing drugs for various liver diseases. Existing treatments for liver diseases often face numerous challenges, such as limited efficacy, significant side effects, or the development of drug resistance, which make it challenging to meet clinical needs. Therefore, in-depth research on drug treatment strategies targeting epitranscriptomic modifications has significant theoretical importance and clinical application value.
In summary, this review provides a comprehensive and systematic overview of the molecular mechanisms of epitranscriptomic modifications, delves into the complex regulatory network between these modifications and PCD, and elaborates on their roles in liver diseases. Additionally, it summarizes the current drugs targeting epitranscriptomic modifications. This work aims to offer new theoretical insights for further investigation into the pathogenesis of liver diseases, to pave new directions for the development of safer and more effective therapeutic drugs and strategies for liver diseases, to promote innovative development in the field of liver disease treatment, and ultimately to improve the prognosis and quality of life of patients with liver diseases.