Complex RNA Changes in Female Reproduction and Women's Pregnancy and Gynecological Disorders

Apart from m⁶A modification, adenosine-to-inosine (A-to-I) RNA editing is one of the most common post-transcriptional RNA modifications in mammals [98]. This process is performed by the adenosine deaminase acting on the RNA family that acts on double-stranded RNA, including the catalytically active adenosine deaminase acting on RNA1 (ADAR1), ADAR2 and ADAR3, which usually lacks catalytic activity [99]. Biochemically, ADAR enzymes convert adenosine in RNA into inosine through hydrolytic deamination [100]. Because inosine resembles guanosine in its structure, the translation machinery often interprets it as guanosine, which can introduce nonsynonymous changes in protein coding sequences [101]. Besides changing the amino acid’s identity and expanding proteomic diversity, A-to-I editing can affect RNA splicing, secondary structures, transcript stability and translational efficiency [102]. Editing from A to I is another common occurrence in ncRNAs and may interfere with either Drosha or Dicer processing. This positively affects miRNA maturation and target recognition while negatively influencing circRNA biogenesis [103].

Beyond m⁶A, another prevalent and functionally significant RNA modification is A-to-I editing. A-to-I editing plays a central role in the female reproductive system’s functioning and maladies. In oocytes, inosine marks are enriched in coding regions and are preferentially found at wobble sites, which suggests that A-to-I editing may alter codon usage and affect the clearance and stability of maternal RNA in meiosis [104]. Ovarian granulosa cell functioning also requires ADAR1-mediated editing, as the granulosa cell-specific deletion of ADAR1 impairs ovulation, delays oocyte maturation and disrupts the expression of inflammation-related genes [105]. When A-to-I editing is unregulated, it can promote the development of POCS [106]. Patients with polycystic ovary syndrome experience increasing RNA editing sites in adipose tissue and the abnormally reduced expression of ADAR2 in granulosa cells [107]. ADAR1 influences the RNA editing of substrates, such as Eukaryotic Translation Initiation Factor 2-Alpha Kinase 2 (EIF2AK2), which increases expression and activates Mitogen-Activated Protein Kinase (MAPK) signalling. Hence, the disease-associated pathology is worsened. Thus, this site may have good potential for clinical diagnoses [108].

A-to-I editing also shapes the immune microenvironment outside its direct functioning in the regulation of reproductive endocrines. ADAR1 can edit endogenous double-stranded RNA, including Alu-derived transcripts, converting them to less immunogenic versions. This prevents them from being recognised by innate immune sensors, such as melanoma differentiation-associated protein 5, thus diminishing self-directed inflammation and immune activation [32]. A-to-I editing also interacts with other RNA modifications such as m⁶A. Together, these form an integrated post-transcriptional regulatory network [109]. Selectively targeting ADAR enzymes is becoming a viable research approach because of its importance to oocyte development and immune homeostasis [110]. Inhibitors of ADAR1 have been created, and these epitranscriptomic therapies may provide novel biomarkers and precision treatments for polycystic ovary syndrome and numerous female reproduction-related diseases [111].

One of the most prevalent RNA modifications found in eukaryotes is pseudouridine (Ψ), termed RNA’s fifth nucleotide when first identified in the 1950s. The Ψ isomer of uridine is a C-glycoside. The conversion of uridine to Ψ affects ribose conformation and increases the backbone’s rigidity, thus stabilising RNA duplex structures [112]. There are two major routes to pseudouridylation. One modification route uses standalone Ψ syntheses that directly encounter the substrate and catalyse the chemical transformation. The other pathway relies on the box H ACA ribonucleoprotein complex containing the Dyskerin (DKC1) protein, which facilitates site-specific catalysis [113]. The modification appears irreversible because the carbon–carbon bond formed between the base and ribose by Ψ is more chemically inert and more difficult to break than the carbon–nitrogen bond in uridine. In fact, there is no identified eraser, and only a few reader proteins have been reported, such as Profilin 1 (PFN1), polyadenylate binding protein 1 (PABPC1) and the yeast factor Prp5 [38].

Additionally, Ψ is in different areas of mRNA and regulates many post-transcriptional processes, namely splicing, translation and transcript stability. While being translated, codons with Ψ can encourage aberrant base pairing. This can cause amino acid misincorporation or stop codon readthrough. Moreover, Ψ protects RNA from degradation by ribonucleases. This increases the half-life of transcripts [112,113]. Changes in environmental signals and cellular stress can alter Ψ landscapes. Disruptions to these programmes occur via mutations in Ψ synthases or through their aberrant expression, leading to a range of human diseases [113].

An alteration of Ψ has significant clinical applications in female reproductive disorders and gynaecologic cancers. Because human cells lack the requisite enzymes to metabolise C-glycoside compounds efficiently, excess Ψ is excreted and may accumulate in many human body fluids [113]. Elevated Ψ has been detected in plasma samples collected before a diagnosis of ovarian cancer, indicating a relationship between dysregulated Ψ and preclinical disease progression that may present an early warning biomarker [114]. A candidate prognostic marker and therapeutic target in ovarian cancer, Ψ synthase 7 (PUS7) promotes tumour cell proliferation [115]. The enrichment of DKC1 protein levels is also related to breast cancer progression via effects on RNA biogenesis and telomerase activity regulation [116]. The findings enhance our understanding of how Ψ modification functions at the cellular level, provide an alternative view of disease evolution within the female reproductive system and bolster prospects for personalised clinical applications.

At the fifth carbon of cytosine, 5-Methylcytidine (m⁵C) is a post-transcriptionally regulated RNA chemical modification. This modification was again found in tRNA, rRNA and mRNA and plays an important role in epitranscriptomic regulation [117]. The installation of m⁵C requires coordination between various protein complexes. Major methyltransferases include members of the NSUN family and DNA methyltransferase 2 (DNMT2), otherwise called tRNA aspartic acid methyltransferase 1 (TRDMT1). The NSUN family has seven members, NSUN1 to NSUN7, and is named after the highly conserved NOL1/NOP2/SUN domain. The core enzymes involved in the mRNA m⁵C modification include NSUN2 and NSUN6 [118]. This modification is reversible because ten-eleven translocation (TET) and AlkB homologue 1 (ALKBH1) can oxidise m⁵C back to 5-hydroxymethylcytidine, reversing the methylation event [119]. Functionally, m⁵C mainly works through recognition by reader proteins that determine RNAs’ fate. As an example, the nuclear reader Aly/REF export factor (ALYREF) acts as an mRNA nuclear export factor by recognising m⁵C sites, whereas the cytoplasmic reader Y-box-binding protein 1 (YBX1) recruits cofactors to enhance mRNAs’ stability and translation. These mechanisms display how essential m⁵C is to maintaining the balance of the RNA metabolism [45,120].

The m⁵C modification in female reproductive physiology and pathology is spatio–temporally regulated. The levels of m⁵C vary in oocytes and early embryos. NSUN5-mediated rRNA methylation plays a crucial role in the stability of maternal mRNA and when the loss function NSUN5 can cause ovarian insufficiency and developmental arrest [121]. During the endometrial cycle, epithelial cytoskeleton remodelling and adhesion molecule expression regulated by NSUN2-mediated m⁵C modification are needed to establish normal endometrial receptivity and embryo implantation [122]. Abnormal m⁵C modification is associated with different gynaecologic diseases. Altered TRDMT1 activity in granulosa cells can disrupt DNA damage repair mechanisms and lead to premature ovarian insufficiency [123]. In cancers such as cervical cancer, the high expression of NSUN2 causes aberrant methylation of the oncogenic mRNAs Keratin 13 (KRT13) and Leucine-rich repeat-containing eight volume-regulated anion channel A (LRRC8A). These methylated transcripts are recognised and stabilised by the reader protein YBX1, enhancing tumour cell proliferation, invasion and chemotherapy resistance. These findings suggest that the m⁵C pathway may serve as a novel biomarker and therapeutic target for gynaecologic diseases [124,125].

N7-Methylguanosine (m7G) is a common epitranscriptomic modification in RNA. It involves methylation at the N7 position of guanosine catalysed by specific methyltransferases [126]. In addition, m7G appears mainly at the 5 prime cap and internal sites of eukaryotic mRNA and is also found in the internal positions of rRNA, tRNA and miRNA [127]. Key m7G regulatory proteins include complexes formed by methyltransferase-like 1 (METTL1) with WD repeat domain 4 (WDR4), Williams–Beuren syndrome chromosome region 22 (WBSCR22) with tRNA methyltransferase activator subunit 11-2 (TRMT112) and RNA guanine-7 methyltransferase (RNMT) with its activating mini-protein (RAM) [128]. The METTL1 WDR4 heterodimer installs m7G at the tRNA variable loop G46, G quadruplex structures of primary microRNAs (pri-miRNAs) and internal mRNA sites, regulating RNA stability, nuclear export and translation efficiency [129]. WBSCR22 and TRMT112 work together to catalyse m7G at G1639 of 18S rRNA, supporting ribosome biogenesis and maturation [130]. RAM and RNMT function to create an m7G cap at the 5′ end of the mRNA. Furthermore, its interaction with eukaryotic translation initiation factor 4E (eIF4E) directly influences nuclear export, and translation initiation states this assertion [131]. Outside these roles, m7G and N6,2 prime-O-dimethyladenosine (m⁶Am) stave off the degradation mediated by mRNA-decapping enzyme 2 (DCP2), thus maintaining the stability of RNA [132].

The m7G alteration and dysregulation of regulatory protein are partially implicated in the development of tumours and remodelling of the immune microenvironment in female reproductive malignancies [133]. M7G-associated lncRNAs and miRNAs in endometrial cancer may be used to build prognostic models, with expression levels significantly correlated with immune infiltration profiles and drug sensitivity [134]. Evidence from breast cancer research has suggested that m7G modification facilitates the infiltration of multiple immune cells. Examples include naïve B cells, CD4 plus memory T cells, CD8 plus T cells, NK cells and M1 macrophages. Most importantly, nuclear cap-binding protein 1 (NCBP1) mRNA is identified as a key m7G target [135]. The m7G tRNA modification mediated by METTL1 can affect the translation of chemokines such as CXCL8, which will recruit polymorphonuclear myeloid-derived suppressor cells (PMN-MDSCs) to create an immunosuppressive tumour microenvironment [136]. In ovarian cancer, prognostic models based on m7G regulatory genes can effectively predict patients’ survival outcomes [137]. The m7G modification, along with m⁶A and m⁵C modifications on RNA forming multi-layered networks, underlies the escape ability of the immune system and therapy resistance in gynaecologic malignancies [138]. The precise function of m7G in shaping a tumour’s immune microenvironment indicates its potential as a biomarker and a therapeutic target. Monitoring its modification status may help guide precision immunotherapy [139].

N1 methyladenosine (m¹A) refers to the methylation modification that occurs at the first nitrogen of adenosine. This modification takes place in both coding and non-coding RNAs [140]. M¹A carries explain a positive charge under physiological conditions, which disrupts standard Watson–Crick base pairing, alters the secondary structure of RNA and affects RNA–protein interactions. M¹A is involved in controlling RNA stability and translation efficiency, as well as other internal metabolism processes [141]. Methyltransferases, demethylases and reader proteins stringently control the balance of m¹A. Key methyltransferases are tRNA methyltransferase 6 (TRMT6), TRMT61A, TRMT10C, TRMT61B and nucleomethylin (NML). Demethylases that erase m¹A include AlkB homologue 1 (ALKBH1), ALKBH3, ALKBH7 and fat-mass and obesity-associated protein (FTO). Recognising proteins include YTH domain-containing proteins YTHDF1, YTHDF2, YTHDF3 and YTHDC1 [142].

In the female reproductive system and gynaecologic diseases, m¹A modification and its regulatory factors reveal clear clinical and pathological importance [143]. In ovarian and breast cancers, the demethylase ALKBH3 removes m¹A modifications from GC-rich regions of colony stimulating factor 1 (CSF1) mRNA, markedly prolonging its half-life and promoting macrophage recruitment and cancer cell invasion within a tumour’s microenvironment [144]. High expression of the methyltransferase TRMT6 in ovarian cancer is tied to poor patient survival, while microRNA 1915p can target TRMT6 to reduce its oncogenic effects [145]. In cervical cancer, the methyltransferase TRMT10C works as an oncogene with elevated expression tied to worse survival, whereas TRMT6, TRMT61A, ALKBH3, YTHDC1 and YTHDF2 display lower expression and act as tumour suppressors [146]. In addition, in endometrial cancer, m¹A-related lncRNAs can be used to build prognostic models tied closely to immune infiltration patterns and drug sensitivity [147]. In summary, m¹A modification adjusts metabolic networks and immune infiltration in gynaecologic malignancies and is becoming a target for developing new therapeutics and combinatory immunotherapies [148].

N4-acetylcytidine (ac4C) is an RNA epitranscriptomic modification. The modification was identified in the tRNA of yeast in 1966 and was confirmed in the tRNA of bacteria in 1972 and in 18S rRNA of eukaryotes 1978 [149,150,151]. As the sole known RNA modification in eukaryotes, ac4C is produced by N-acetyltransferase 10 (NAT10), which has acetyltransferase, tRNA binding and RNA helicase domains. Studies have demonstrated that ac4C exists in different types of RNAs, including viral RNAs [152,153]. It assists with the assembly of RNA duplex and modulates translation. The ac4C modification in the anticodon loop of tRNA enhances codon recognition, while ac4C modification in the decoding centre of 18S rRNA helps the ribosome to select the correct codons. Transcript stability is enhanced due to ac4C modifications of mRNA coding regions. These mechanisms maintain and enhance, respectively, the fidelity and efficiency of target protein translation [154,155].

Regarding female reproductive health, NAT10-mediated ac4C modification is associated with oogenesis, embryonic development and various gynaecological diseases [153]. NAT10 expression slowly drops between the germinal vesicle stage and metaphase II (MII) during oocyte maturation, and this event influences meiotic progression by stabilising O-GlcNAcase (OGA) mRNA [156]. The loss of NAT10 arrests follicular development at the primary stage and results in premature ovarian insufficiency [157]. In early embryonic development, ac4C stabilises OCT4 mRNA, allowing embryonic stem cell self-renewal and pluripotency [158], and stabilises NOP2 mRNA to regulate the morula-to-blastocyst transition [159]. NAT10 is upregulated in gynaecologic malignancies, such as cervical and ovarian cancer, tying to poor prognoses, and promotes tumour cell proliferation and invasion through ac4C modification [160,161]. However, in animal models of premature ovarian insufficiency, elevated ac4C levels are tightly tied to ovarian microenvironmental damage [162]. Taken together, ongoing research on ac4C and related modifications points to the potential for RNA epitranscriptomic-based precision diagnostics and therapeutics in future gynaecologic clinical practice.

Moreover, 5-methoxycarbonylmethyl-2-thiouridine (mcm⁵s²U) is a tRNA modification found at the wobble uridine (U34) position in the anticodon loop. Its biosynthesis requires tRNA methyltransferase 9-like protein (TRM9L) and AlkB homologue 8 (ALKBH8) to work together [163,164,165]. In eukaryotes, mcm⁵s²U stabilises the translation reading frame, and its absence reduces A-site selection efficiency by the ribosome, causing +1 frameshift errors [166]. Studies illustrate that mcm⁵s²U cooperates with m⁶A, promoting ribosome A-site binding efficiency to inhibit m⁶A-mediated mRNA decay to control oncogenic mRNA stability [167,168]. Models of non-associative learning have proposed that elevated mcm⁵s²U levels correlate with the increased synthesis of polyglutamine proteins and more neuronal excitability, suggesting a possible role in neural plasticity [169].

Although the involvement of mcm⁵s²U in essential biological processes has been elucidated, the possibilities for mischief in female reproductive physiology and gynaecologic diseases have not been studied directly. Due to the crucial role it plays in the maintenance of translation fidelity and the regulation of cellular differentiation [167,168], further studies could explore the control mechanisms during embryonic development, oogenesis and cyclic endometrial changes. As associations between mcm⁵s²U, m⁶A and poor prognoses in breast cancer have been reported, a systematic study of this modification in the initiation and progression of gynaecologic tumours could provide traction for the development of new biomarkers and therapeutic targets [167,170].

RNA modifications form another layer of gene expression regulation. Studies show that various RNA modifications do not exist in isolation, but rather in complex interactive networks in cells. These interactions are primarily mediated through the coordinated activities of writers, erasers and readers, which collectively determine the fate of RNA at multiple levels, including splicing, stability, translation and localisation [171].

To establish a unified framework for understanding these diverse processes, we propose that RNA modifications in the female reproductive system operate as an integrated ‘Epitranscriptomic Code’. This framework consists of three hierarchical layers: the ‘Molecular Input Layer’, where diverse modifications (m⁶A, m⁵C, Ψ, etc.) are deposited by specific writers in response to hormonal and environmental signals; the ‘Regulatory Processing Layer’, where reader proteins interpret these marks to modulate the post-transcriptional life cycle of key reproductive transcripts; and the ‘Functional Output Layer’, where these coordinated changes culminate in physiological milestones such as oocyte maturation or pathological states similar to tumorigenesis. This multi-layered network ensures that female reproductive cells can achieve rapid, reversible and precise gene expression control, moving beyond isolated mechanisms to a systems-level regulatory logic [172].

In this context, the cooperative and antagonistic interactions between modification enzymes are particularly significant, as this dynamic is essential for fine tuning RNA activity. Mechanistically, such crosstalk can occur through several modes, including competition for overlapping or structurally constrained RNA regions, RNA structural remodelling induced by specific modifications and feedback regulation among modification enzymes [31,173].

For instance, ALKBH5 (m⁶A demethylase) and NSUN4 (m⁵C methyltransferase) are regulated mainly at the protein level. They mainly affect mRNA splicing, protein acetylation and ubiquitination. These two modification systems create post-transcriptional feedback loops through the mutual modification of each other’s effector transcripts, the ALKBH5 transcripts carrying m⁵C modifications and the NSUN4 transcripts carrying m⁶A sites. This type of reciprocal regulation represents a typical feedback mechanism that contributes to the maintenance of epitranscriptomic homeostasis.

The methyltransferases METTL14 and METTL16 and the m⁵C methyltransferases NSUN4 and NSUN5 co-regulate each other closely, related to mitochondrial functioning and post-translational processes such as phosphorylation and SUMOylation [174]. In addition, m6A writers, such as METTL3/METTL14, can regulate the expression of other RNA modification enzymes via the m6A-dependent control of RNA stability and translation, further strengthening the interconnected regulatory network [172].

Analysing transcriptomes with nanopore direct RNA sequencing, we found a negative correlation between m⁶A and pseudouridine (Ψ) within the same transcript, suggesting Ψ-enriched transcripts have fewer m⁶A sites. This observation may be explained by competition for structurally constrained regions or steric hindrances between modification machineries [175]. Inside polysomes, however, the two modifications display a strong synergism, jointly enhancing translation. Functionally, Ψ improves codon–anticodon stability and translation fidelity, whereas m6A promotes translation initiation, indicating a division of labour between these modifications [176]. In this hierarchy, Ψ contributes the most to translation efficiency over m⁶A [177].

The quantitative evaluation of RNA modifications has been enabled through mass spectrometry [178]. Knocking out METTL3 causes compensatory increases in Ψ levels in high m⁶A messages, which suggests m⁶A may inhibit Ψ installation. Conversely, the loss of the Ψ synthase TRUB1 unexpectedly decreases m⁶A levels, suggesting a reciprocal regulatory network between the two systems [177].

A-to-I editing refers to the process of converting adenosine (A) into inosine (I) catalysis by a family of enzymes known as the ADARs. Both methylation on adenosine and m⁶A modification occur, meaning that the two modifications are negatively correlated and often mutually exclusive in transcripts. Mechanistically, m6A disrupts the local RNA duplex structures required for ADAR binding, thereby inhibiting A-to-I editing [173]. The knockdown of METTL3 markedly increases global A-to-I editing, though FTO knockdown reduces A-to-I levels. These modifications show widespread antagonistic interactions [173,179].

In a study reporting on early-stage lung adenocarcinoma, an expression signature based on m⁶A- and m⁵C-related regulatory genes has strong prognostic capabilities [180]. Hepatocellular carcinoma (HCC) oncogenes, such as epidermal growth factor receptor (EGFR), are regulated in part by m⁶A and m⁵C modification to drive tumour progression. ALYREF-mediated m⁵C modification exerts a stabilising effect on EGFR mRNA, leading to the upregulation of EGFR expression [181]. In contrast, m⁶A seems to exert incremental effects at experimental concentrations. METTL3 enhances the translation efficiency of EGFR, while YTHDF2, through its binding to m⁶A sites, recruits degradation complexes that speed up the degradation of the EGFR transcript [182,183]. This illustrates a coordinated regulatory model in which m⁵C primarily stabilises transcripts, whereas m6A dynamically balances translation and decay [120].

In contrast, in glioblastoma (GBM), the m⁶A modification system and A-to-I editing have cell type-specific functions. In GBM stem cells (GSCs), METTL3 and METTL14 preserve m⁶A levels to inhibit self-renewal and tumorigenesis, while their downregulation facilitates GSC proliferation and stemness [184]. In differentiated GBM cells, METTL3 can enhance the presence of the ADAR1 protein via the m⁶A-YTHDF1 axis; in turn, ADAR1 binds to and stabilises cyclin-dependent kinase 2 (CDK2) mRNA, facilitating cell cycle progression and tumour growth [185]. These findings further support a context-dependent interaction between RNA modifications, where the same modification can exert distinct biological effects depending on the cellular state [172].

It has been demonstrated that genetic alterations, such as the PIK3CA mutation and PTEN loss, drive the activation of the phosphatidylinositol 3-kinase/protein kinase B/mammalian target of rapamycin (PI3K/Akt/mTOR) pathway in ovarian cancer [186]. Epitranscriptomic studies indicate that m1A modification can promote tumour progression via this pathway in gastric and bladder cancers. In addition, the m⁶A demethylase ALKBH5 modulates autophagy in ovarian cancer through the miR-7/EGFR axis [187]. These findings suggest that m⁶A, m1A, and other RNA modifications may provide an additional layer of regulation over the PI3K signalling pathway.

Outside cancer, modification crosstalk is also important in non-malignant diseases. In cellular senescence models, NSUN2-mediated m⁵C and METTL3/METTL14-mediated m⁶A cooperatively modify the 3′UTR of p21 mRNA, mutually promoting each other and enhancing p21 translation, a mechanism important for oxidative stress-induced ageing [188].

In patients with premature ovarian insufficiency (POI) and chemotherapy-induced POI models, ovarian granulosa cells show elevated m⁶A levels and reduced FTO expression, accompanied by impaired proliferation and increased apoptosis [189]. Since oxidative stress is a key pathological factor in POI, multiple insults can induce excessive ROS accumulation in granulosa cells, leading to mitochondrial dysfunction and follicle depletion [190,191]. Because of the close relationship between oxidative stress and cellular senescence, Li et al. showed that m⁶A/m⁵C cooperative modification regulates p21 to influence cell ageing, a mechanism likely relevant in granulosa cells from POI patients. Therefore, the dysregulation of m⁶A/m⁵C crosstalk may contribute to the ovaries’ functional decline. Therefore, the dysregulation of RNA modification crosstalk may represent an important epitranscriptomic mechanism underlying ovarian ageing and functional decline.

Taken together, these findings show that complex networks of RNA modifications play important roles in disease development and progression. However, several key questions remain unresolved, including the temporal hierarchy of modification deposition, the structural determinants governing modification interplay and the quantitative contribution of each modification to gene regulation. Addressing these issues will require integrative approaches that combine high-resolution sequencing, mass spectrometry and functional epitranscriptomic editing technologies.

Oogenesis is a controlled biological process that starts with primordial germ cells during embryogenesis and progresses through mitosis to meiosis and growth to produce mature oocytes or eggs that can be fertilised. In addition, the development of multi-celled embryos can also support [192]. In mammals, the growing oocyte undergoes extensive transcription and accumulation of RNA molecules known as maternal factors that continue until oocyte growth, meiotic maturation and the early period after maturation. During classification methylation, when fully meiotically matured oocytes are transcriptionally silent, these maternal mRNAs are used for protein synthesis as the primary source of genetic information [157,193,194]. Upon entering meiotic maturation, fully grown oocytes exhibit very little transcriptional activity [195]. Hence, oocyte quality and the ability of early embryos to develop depend less on the new transcription of genes than the spatio–temporal control of pre-stored maternal RNAs in the oocyte, particularly their stability, localisation, translated activation and final degradation [196,197]. The development of oocytes does not happen alone; instead, it runs with the development of granulosa cells (GCs) in the microenvironment prevalent in the ovarian follicle. In addition, RNA modifications have significant control functions [198,199].

Early embryonic development after fertilisation involves a shift from the reliance on maternal RNA to the activation of transcription from the embryonic genome. This transition is accompanied by the extensive reprogramming of gene expression. RNA modifications are an important layer of epitranscriptomic regulation and have important functions in controlling early post-fertilisation transcriptional activation and gene expression.

Research shows that m⁶A changes dynamically during the maternal-to-zygotic transition (MZT). It can be inherited from maternal mRNA and can be added to transcripts synthesised after fertilisation, supporting regulation in two directions. That is, m⁶A promotes the decay of a subset of maternal mRNAs, allowing the removal of transcripts that are no longer required. Conversely, m⁶A stabilises a smaller group of mRNAs and enhances their translation, supporting the processes needed for embryonic development. In mice, m⁶A-associated factors such as Ythdc1 and Ythdf2 serve as reader proteins during preimplantation development and assist MZT [233]. In early mammalian embryos, m⁶A also influences RNA chromatin interactions and reshapes transcriptional control networks, modulating embryonic genome activation and cell fate decisions [234].

During the oocyte-to-embryo transition in mice and humans, m⁶A and N6,2′-O-dimethyladenosine (m⁶Am) show species-specific regulatory patterns. Zygotic genome activation (ZGA) marks a major control point during which the m⁶A and m⁶Am methylation landscape changes greatly. Transcripts with these marks are typically more abundant and more translated. Moreover, m⁶A appears to play a dual role in transposable element RNA regulation by promoting its activity before ZGA and curbing runaway activity during ZGA [235]. M⁶A is present in pig embryos from the one-cell stage to the blastocyst stage with an upward shift from the morula to blastocyst stage. Inhibiting m⁶A methylation causes developmental arrest, abnormal lineage allocation and increased apoptosis and autophagy at the blastocyst stage, displaying that m⁶A is needed for normal development [236]. Y-box binding protein 1 (YBX1) influences the ZGA and histone modification states in pig embryos by regulating the expression of METTL3 and IGF2BP1 [237]. Single-cell transcriptomic analyses have also characterised the gene expression dynamics of MZT in early pig embryos [238].

In addition, m⁵C helps maintain maternal RNA stability and supports the transition to ZGA. The m⁵C methylation produced by NSUN6 may add a distinct layer of cell division control during human embryonic development [209]. The m⁵C binding protein ALYREF promotes mRNA export from the nucleus to the cytoplasm. YBX1 helps secure orderly MZT by stabilising maternal RNAs marked by m⁵C. YBX1 is linked to translational repression in zebrafish embryos, whereas in mammals, it may also take part in ZGA through an m⁶A-related route [45,207].

Additionally, ac4C plays a part in the key stages of embryonic development. The ac4C writer NAT10 is needed for ZGA and blastocyst formation in mouse embryos. The loss of NAT10 leads to arrest at the two-cell or morula stage and disrupts both maternal mRNA decay and ZGA, showing that ac4C assists with development by regulating mRNA stability and transcriptional activity [239]. Mechanistic work further suggests that NAT10 modulates ac4C on specific targets, such as Nop2, thus altering mRNA stability and protein production. By these means, NAT10 influences cell fate decisions and promotes the transition from the morula to the blastocyst [159].

A-to-I RNA editing is another RNA modification that assists with early embryonic development. Single-cell transcriptomic analyses indicate that editing events in early human embryos are strongly stage specific. At the eight-cell stage, editing is extensively reprogrammed, and editing levels drop at most sites. Editing also tends to occur at splice sites of lncRNAs, and editing levels tie closely to lncRNA splicing indices, pointing to its function in transcriptional control through the modulation of lncRNA splicing [240]. Outside transcription-related effects, A-to-I editing may also influence maternal RNA stability after fertilisation by shifting codon usage patterns and, thus, promoting maternal RNA clearance. For example, inosine modifications in oocyte transcripts are enriched at wobble positions within coding regions and are markedly reduced when CCR4-NOT transcription complex subunit 6 like (Cnot6l) is absent. These results give credence to the idea that A-to-I editing alters codon usage in a manner favourable for maternal RNA stability and MZT [104,241,242].

In embryos, epigenetic mechanisms, such as histone modifications and DNA methylation, aid in genome reprogramming and transcriptional activation [243,244]. The embryonic genome becomes activated efficiently through RNA modifications and related mechanisms, allowing cells to develop proper fates. Regulatory mechanisms for RNA modifications differ in their species and cell type specificity, demonstrating that such mechanisms are subject to the species context and developmental stage.

When taken as a whole, RNA modifications have a powerful effect on fertilisation and embryonic genome activation at multiple levels, including RNA stability, splicing, translation and transcriptional control. The interplay between gene and chromatin dynamics not only reflects developmental competence but also enables the coordination between histone modifications] and DNA methylation to achieve the timing and efficiency of ZGA. By shaping genome activation, RNA stability and cell cycle regulation, RNA modifications are needed for essential early developmental processes. The disruption of these pathways can lead to developmental arrest and implantation failure and can impair lineage differentiation and overall embryo quality. Further work could define the stage-specific functions of different RNA modifications and their regulators across embryonic development to identify new targets and strategies for diagnosing and treating developmental disorders.

Endometrial receptivity is a prerequisite for successful embryo implantation and the maintenance of pregnancy. Its establishment depends on tightly controlled ovarian steroid hormone signalling and the dynamic remodelling of the epitranscriptome [245]. During normal pregnancy, m⁶A levels in the endometrium increase progressively, and the spatial distribution of related regulators changes in a gestational stage-dependent manner. This balance is disrupted in infertility, with reduced transcript levels of METTL16 and WTAP, abnormally elevated ALKBH5 and decreased IGF2BP2, changes that are tied to the dysregulation of immune-related signalling pathways [246].

Between days 16 and 25 of pregnancy, METTL3 is highly expressed in luminal and glandular epithelial cells of the endometrium, and its expression can be induced by oestrogen and progesterone. METTL3 supports CTGF expression by keeping m⁶A on CTGF mRNA, which promotes endometrial epithelial proliferation and helps create an endometrial environment that allows implantation [247]. Further studies report that METTL3 is markedly reduced in the endometria of patients with recurrent implantation failure (RIF) and stage-IV endometriosis (EMs)-associated infertility. METTL3 levels correlate positively with progesterone receptor (PGR) and MYC levels and correlate negatively with ELF3. A uterus-specific conditional knockout of Mettl3 driven by Pgr-Cre shows that the loss of METTL3 leads to complete implantation failure and female infertility. Mechanistically, METTL3-dependent m⁶A promotes the degradation of oestrogen-responsive transcripts such as Elf3 and Celsr2 by targeting their three prime untranslated regions, thus limiting excessive oestrogen signalling. Simultaneously, METTL3 maintains the expression of PGR and its downstream target Myc, supporting normal progesterone responsiveness. This coordinated balance between oestrogen and progesterone receptor pathways is needed for uterine receptivity and for decidualisation, thus supporting female fertility at the organism level [248]. Recent work further indicates that the loss of METTL3 induces SUV39H1-mediated histone H3 lysine 9 trimethylation (H3K9me3) across the gene body of Sequestosome 1 (SQSTM1, also known as P62). Meanwhile, METTL3 deficiency stabilises P62 mRNA in an m⁶A-dependent manner. These changes lead to P62 accumulation, the activation of autophagy and inflammatory programmes and a marked reduction in embryo adhesion. Mice that are conditional knockouts for Mettl3 in their uteruses display increased P62, tumour necrosis factor-α (TNF-α) and the abnormal development of endometrial glands, which leads to implantation failure [249].

At the mechanistic level, METTL3 directly installs m⁶A on progesterone receptor (PGR) mRNA and enhances the translation of PGR protein in a YTHDF1-dependent manner. This process is crucial for the in vitro decidualisation of endometrial stromal cells [250]. Nonetheless, METTL3’s impact is context dependent. In recurrent implantation failure (RIF) patients, pathological increases in METTL3 expression and global m⁶A levels in the endometrium are associated with decreased receptivity. In this context, increased METTL3 enhances m⁶A on homeobox A10 (HOXA10) mRNA and accelerates its decay, thus reducing β3 integrin expression. This change is accompanied by the abnormal activation of empty spiracles’ homeobox 2 (EMX2) signalling and finally impairs implantation [251]. Unlike the complex role of METTL3, the loss of METTL14 drives implantation failure through an alternative route. METTL14 deficiency leads to the abnormal activation of oestrogen receptor alpha (ERα) signalling and ERα phosphorylation and is accompanied by the pathological activation of innate immune pathways. These changes disrupt immune tolerance at the maternal foetal interface and lead to implantation failure [252].

Outside m⁶A, m⁵C also contributes to the regulation of endometrial receptivity. The m⁵C methyltransferase NSUN2 is highly expressed in epithelial cells during the proliferative phase. Its overexpression markedly reduces embryo attachment and is linked to recurrent implantation failure. Mechanistically, NSUN2 installs m⁵C on claudin 4 (CLDN4) mRNA and suppresses Sirtuin 4 (SIRT4), which increases histone H3 lysine 9 acetylation (H3K9ac) and jointly elevates CLDN4 expression. NSUN2 is also reported to promote exon-skipping events in the signal transducer and the activator of transcription 1 (STAT1) and matrix metalloproteinase 14 (MMP14). These alterations impair endometrial receptivity [122].

In conjunction, m⁶A and m⁵C RNA modifications work through major regulators such as METTL3, METTL14 and NSUN2 to coordinate the establishment and maintenance of endometrial receptivity at the post-transcriptional level. These enzymes influence hormone receptor signalling, cell adhesion, splicing and the immune microenvironment to support successful implantation. Imbalance in the RNA modification network, whether excessive activity or loss of function, can lead to impaired receptivity and infertility-associated pathology. These findings show the importance of maintaining a dynamic epitranscriptomic balance for normal reproductive outcomes.

In early mammalian embryogenesis, m⁶A supports successful MZT by regulating the timing of maternal messenger RNA clearance and the activation of the zygotic genome. The loss of METTL3-mediated m⁶A in mouse oocytes impairs maternal mRNA decay, delays ZGA and leads to developmental arrest or embryonic lethality [253]. Simultaneously, YTHDF2 promotes the removal of maternal transcripts by recognising m⁶A, a process needed for early zebrafish development and for maternal mRNA clearance in goat embryos [254,255]. In addition, IGF2BP2 serves as a key maternal activator at the two cell stages by regulating specific targets that support blastocyst formation, whereas IGF2BP3 keeps maternal RNA stable to sustain early zebrafish development [256,257]. Outside m⁶A, m⁵C assists with preimplantation development. The m⁵C methyltransferase NSUN5 (NOL1/NOP2/Sun domain family member 5) is greatly upregulated at the two-cell stage, and its loss reduces blastocyst formation, lowers cell numbers and increases apoptosis. At the mechanistic level, NSUN5 regulates the expression of Hippo pathway components LATS1 (large tumour suppressor 1) and LATS2, affecting YAP1 (yes-associated protein 1) nuclear localisation and the expression of the trophoblast marker CDX2. Accordingly, NSUN5 supports the proper segregation of the inner cell mass and trophoblast lineages [258].

In trophoblast stem cells, m⁶A has dual effects on pluripotency and differentiation. METTL3-mediated m⁶A can reduce the stability of pluripotency factor mRNAs and promote their decay, allowing the transition of mouse embryonic stem cells (mESCs) from a naive state towards differentiation [259]. The loss of METTL3 impairs the self-renewal of mESCs and leads to embryonic lethality [260]. Likewise, METTL14 deficiency markedly delays post-implantation development and impairs the maturation of ectoderm-derived lineages [261]. At the rRNA level, METTL5 installs m⁶A on 18S rRNA and influences the differentiation potential of mESCs. METTL5 loss leads to defects in germ layer differentiation and the abnormal accumulation of c-MYC protein [262]. METTL16 could also influence early development by regulating methionine adenosyltransferase 2A (MAT2A) mRNA abundance [263]. YTHDC1 helps mature oocytes through the regulation of alternative polyadenylation and splicing, while YTHDC2 supports meiosis and embryonic development by modulating mRNA translation efficiency and stability [27,212].

As placental development proceeds, m⁶A plays a key role in trophoblast invasion, migration and the maintenance of functions at the maternal–foetal interface. In placental tissue, m⁶A sites are enriched in the three prime untranslated regions and in coding sequences near the stop codon. Abnormal m⁶A levels are tightly linked to foetal growth restriction and preeclampsia, and placental FTO expression is also tied to foetal growth [264], while placental FTO expression levels also correlate with foetal growth [265]. METTL3 enhances trophoblast invasiveness by methylating myosin light chain kinase (MYLK) mRNA, an effect needed for normal placental functioning [266]. YTHDC1 assists with placental homeostasis by promoting the degradation of circMPP1 through pathways involving nuclear factor kappa B (NF-κB) and mitogen-activated protein kinase 3 (MAPK3) signalling [267]. Differences in m⁶A levels within the five prime untranslated regions are tied to foetal weight, suggesting that translation-related mechanisms assist with placental functioning [264].

Demethylases exhibit different functional patterns during placental development. Placental FTO expression correlates positively with newborn birth weight and could help regulate foetal growth [265]. ALKBH5 regulates trophoblast invasiveness by demethylating cysteine-rich angiogenic inducer 61 (CYR61) mRNA. Elevated ALKBH5 expression has been tied to recurrent miscarriage [268]. ALKBH5 is also involved in placental regulation by modulating the m⁶A modification of the mRNA of the peroxisome proliferator-activated receptor gamma (PPARG) in preeclampsia [269]. FTO can additionally impact the growth and differentiation of embryonic stem cells through the demethylation of long interspersed nuclear element 1 (LINE1) RNA [270]. Aside from methylation and demethylation, A-to-I RNA editing also contributes to placental biology. In the placenta, endogenous double-stranded RNA (dsRNA) is modified by the editor ADAR1 to maintain innate immune homeostasis. The loss of ADAR1 leads to heightened interferon signalling and activates an integrated stress response (ISR) in junction zone trophoblasts, affecting placental function and leading to lethality. The phenotype in question is associated with the accumulation of double-stranded RNA derived from the three-prime untranslated region(s) of interferon-stimulated genes (ISGs), suggesting the important role of RNA editing in regulating immunity at the maternal–foetal interface [271].

Taken together, RNA modifications form a multi-level regulatory network that spans the reproductive cycle, including oogenesis, fertilisation, embryonic development, the establishment of uterine receptivity and placental formation (Figure 2). From primordial germ cell differentiation to MZT and from shaping endometrial receptivity to maintaining immune homeostasis at the maternal foetal interface, m⁶A, m⁵C, ac4C and A-to-I RNA editing work together. By controlling transcript stability, translation efficiency and the timing of RNA decay, these mechanisms secure the accurate control of timing and location for germ cell fate decisions and embryonic development. The disruption of these epitranscriptomic pathways is tightly linked to obstetric and gynaecologic disorders, including polycystic ovary syndrome, premature ovarian insufficiency, recurrent implantation failure and preeclampsia. It offers a basis for understanding the molecular causes of reproductive dysfunction.

Given the critical role of these RNA modifications in maintaining normal reproductive physiology, it is not surprising that their dysregulation is increasingly implicated in a spectrum of obstetric and gynaecologic disorders. The following sections will explore how these molecular aberrations contribute to the pathogenesis of such diseases, from gynaecologic cancers to pregnancy-related complications.

Cervical cancer (CC) remains a major cause of cancer-related death among women worldwide, and its development is tightly linked to persistent infections with high-risk human papillomavirus (HPV) types [274,275]. M⁶A is a key regulator of CC progression because it shapes the expression of cancer-promoting and tumour-suppressing genes. Among writers, METTL3 is greatly upregulated in CC tissue and is tied to a poor prognosis. METTL3 enhances the stability of target transcripts such as apoptotic chromatin condensation inducer 1 (ACIN1) and cathepsin L (CTSL) in an m⁶A-dependent manner, thus promoting the proliferation, invasion and migration of tumour cells [276,277]. Cell division cycle 25B (CDC25B), a key regulator of the G2/M phase transition, is recognised by YTHDF1 after its mRNA is modified by METTL3, which promotes translation and accelerates cell cycle progression [278]. In addition, METTL3 can regulate the m⁶A modification levels of targets, such as nuclear receptor subfamily 4 group A member 1 (NR4A1) and thioredoxin domain-containing protein 5 (TXNDC5), thus promoting malignant tumour behaviour by facilitating NR4A1 degradation and enhancing TXNDC5 stability, respectively [279,280]. METTL14 is also highly expressed in CC and increases the stability of tripartite motif-containing 11 (TRIM11) mRNA through m⁶A, which promotes tumour progression by activating AKT signalling via the ubiquitination of the phosphatase PHLPP1 [281]. ZC3H13 regulates centromere protein K (CENPK) and cytoskeleton-associated protein 2 (CKAP2) through m⁶A. The ZC3H13 CENPK axis supports cancer stem-like traits and chemotherapy resistance, while the ZC3H13 CKAP2 axis mainly enhances proliferation, invasion and migration [273,282]. RNA-binding motif protein 15 (RBM15) is elevated in HPV-positive cervical squamous cell carcinoma and adenocarcinoma and promotes proliferation, migration and metastasis by modulating m⁶A on targets including Otubain 2 (OTUB2) and decorin (DCN) [283,284].

Among m⁶A readers, YTHDF1 is highly expressed in CC and promotes the translation of RANBP2 mRNA by recognising its m⁶A sites. This effect enhances tumour cell proliferation, migration and invasion through the regulation of RAN GTPase activity [285]. YTHDF2 reduces the stability of axis-inhibition protein 1 (AXIN1) mRNA, thus promoting the epithelial–mesenchymal transition (EMT) and chemotherapy resistance [286]. YTHDF3 is elevated in radioresistant CC cells and contributes to homologous recombination repair by enhancing the translation of RAD51 homologue D (RAD51D), which supports the resistance to radiotherapy [287]. Members of the IGF2BP family are often upregulated across cancers. In CC, IGF2BP3 serves as an oncogene and stabilises cancer-promoting transcripts, such as MYC, by binding m⁶A. IGF2BP2 can bind m⁶A-marked circARHGAP12 and increase the stability of forkhead box protein M1 (FOXM1) mRNA, promoting CC progression [28,87,288]. HPV E6 and E7 oncoproteins can increase IGF2BP2 expression to stabilise host MYC mRNA and support aerobic glycolysis. In addition, the viral E7 transcript, itself, depends on METTL14-mediated m⁶A and recognition by IGF2BP1 to maintain stability [289,290].

Among m⁶A demethylases, FTO is overexpressed in CC. By reducing m⁶A on cancer-promoting transcripts, such as E2F transcription factor 1 (E2F1) and MYC, FTO enhances their translation and promotes tumour cell proliferation and migration [291]. FTO can also reduce m⁶A on β-catenin mRNA to increase its expression, which elevates the activity of excision repair cross-complementation group 1 (ERCC1) and contributes to the resistance to radiotherapy and chemotherapy [292]. Conversely, ALKBH5 is often described as tumour suppressive in CC, and low ALKBH5 levels are tied to a poor prognosis. ALKBH5 can suppress glycolysis by reducing m⁶A on pyruvate dehydrogenase kinase 4 (PDK4) mRNA and can affect the lipid metabolism by promoting the decay of sirtuin 3 (SIRT3) mRNA [293,294]. However, in HPV E7-driven CC, ALKBH5 has also been reported to stabilise p21-activated kinase 5 (PAK5) mRNA through m⁶A-dependent demethylation and, thus, promote tumour progression. ALKBH5 expression also differs substantially between HPV-positive and HPV-negative cells [295].

Outside m⁶A, the role of m⁵C in CC is receiving increasing attention. NSUN2, a major m⁵C methyltransferase, is greatly upregulated in CC tissue. NSUN2 installs m⁵C on keratin 13 (KRT13) and leucine-rich repeat-containing 8A (LRRC8A) mRNAs, enhances their binding to the reader YBX1 and increases transcript stability, promoting tumour cell proliferation, migration and invasion [124,125]. Evidence for m1A regulators in CC is more limited. However, tRNA methyltransferase 10C (TRMT10C) is reported to be highly expressed and associated with a poor prognosis. Silencing TRMT10C suppresses tumour cell proliferation and migration [146].

M⁶A on ncRNAs also contributes to CC progression. After METTL3-mediated m⁶A, lncRNA FOXD2-AS1 becomes more stable and promotes tumour progression by recruiting lysine-specific demethylase 1 (LSD1) and suppressing p21 expression [84]. ZFAS1 works in an m⁶A-dependent manner to sponge miR-647 and relieve the repression of its downstream targets [296]. KCNMB2-AS1 and IGF2BP3 form a positive feedback loop. KCNMB2-AS1 increases IGF2BP3 expression through a ceRNA mechanism, while IGF2BP3 stabilises KCNMB2-AS1 in an m⁶A-dependent manner, together driving CC cell proliferation [288]. Among circRNAs, circARHGAP12 interacts with IGF2BP2 through m⁶A to enhance FOXM1 mRNA stability. Moreover, circCCDC134 maintains its own stability through ALKBH5-regulated m⁶A and promotes metastasis by recruiting p65 to increase the transcription of hypoxia-inducible factor 1-alpha (HIF1α) [87,89]. In addition, Piwi-interacting RNA (piRNA)-14633 can increase METTL14 expression, elevating cytochrome P450 family 1 subfamily B member 1 (CYP1B1) mRNA and protein levels and promoting malignant phenotypes in CC cells [94].

From a translational perspective, m⁶A regulators show promise for a diagnosis, prognosis and treatment response assessment in CC. Prognostic models based on m⁶A-related genes such as ZC3H13 and YTHDC1 can help predict patient survival. The evaluation of combinations of METTL3 MDSCs and CD33-positive myeloid-derived suppressor cells (MDSCs) will be more helpful [297,298]. STM2457, an METTL3 inhibitor, and CWI1-2, an IGF2BP2 inhibitor, show antitumor effects in acute myeloid leukaemia. Rhein, an anthraquinone compound, has the potential to treat breast cancer. Further studies should investigate whether these agents can act on CC and other tumour types [299,300,301]. In addition, the findings in gastric cancer and leukaemia imply that the m⁶A-targeted therapies mediated by nanocarriers may be highly specific and less toxic, given the pro-tumour function of YTHDF1 in CC. This strategy could indicate a possible route for precision therapy in CC [302,303].

Recent investigations have initiated the mapping of the RNA modification landscape in CC. Writers such as METTL3 and METTL14 for m⁶A have been implicated in tumorigenesis via the stabilisation of the HPV E6 and E7 mRNAs, while ALKBH5 and FTO have been reported to be tumour suppressive in some settings.

Other than m⁶A, the m⁵C NSUN2 culprit and the m1A TRMT6 and TRMT61A culprits have also been implicated in cervical cancer cell models. However, there are many limitations. Numerous studies have focused on a single enzyme and a linear downstream target, while the potential competition or cooperation of marks, as shown by m⁶A and m⁵C, is often overlooked. Whether HPV16-positive and HPV18-positive tumours display distinct epitranscriptomic features also remains insufficiently compared. A significant portion of existing research relies heavily on in vitro models (e.g., SiHa, HeLa or SKOV3 cell lines) and high-dose knockdown/overexpression experiments, which may not faithfully recapitulate the physiological complexity of the human reproductive microenvironment. The therapeutic potential of combining small-molecule inhibitors targeting FTO or METTL3 with PD-1 and PD-L1 antibodies or platinum-based chemotherapy also requires validation. Future studies should integrate CLIP-seq and RNA-seq to map interactions among RNA modifications and use patient-derived organoids and HPV molecular subtyping cohorts to clarify tumours’ heterogeneity. Developing cervical cancer-specific inhibitors of RNA modification enzymes and testing strategies that combine these agents with tumour microenvironment remodelling may offer new approaches to overcome the resistance to chemoradiotherapy.

Within the m⁶A methyltransferase complex, METTL3 mainly works as a cancer-promoting factor in ovarian cancer. METTL3 is highly expressed in ovarian cancer tissue and is tightly linked to a higher tumour grade. Hua et al. demonstrated that METTL3 promotes ovarian cancer growth and invasion by increasing the translation of AXL receptor tyrosine kinase and facilitating EMT [304]. METTL3 can also enhance tumour cell survival and proliferation by activating the phosphoinositide 3-kinase/protein kinase B/mechanistic target of rapamycin (PI3K/AKT/mTOR) signalling. In ovarian cancer, METTL3 can promote the maturation of miR-126-5p in an m⁶A-dependent manner, which suppresses PTEN and activates this pathway. METTL3 may also promote progression by regulating AKT phosphorylation and downstream effectors [68,305]. In endometrioid epithelial ovarian cancer, METTL3 can work without METTL14 and WTAP and can regulate m⁶A on genes such as EIF3C and AXL [306]. Conversely, METTL14 is usually considered tumour suppressive in ovarian cancer. METTL14 expression is reduced in ovarian cancer tissue, and METTL14 can inhibit G1-phase proliferation by decreasing the stability of TROAP mRNA [307]. WTAP is the main regulatory subunit of the m⁶A writer complex. Although WTAP lacks catalytic activity, it helps maintain complex functioning by directing METTL3 and METTL14 to nuclear speckles. Yu et al. reported that WTAP is markedly elevated in high-grade serous ovarian carcinoma (HGSOC) and that higher WTAP levels are tied to lymph node metastasis and shorter overall survival. In vitro, WTAP knockdown suppresses proliferation and migration and induces apoptosis, indicating the cancer-promoting function of WTAP in ovarian cancer progression [308].

Moreover, m⁶A readers have key functions in ovarian cancer progression. In cisplatin-resistant ovarian cancer cells, increased m⁶A on TRIM29 mRNA recruits YTHDF1 and enhances translation, maintaining stem-like features [308]. YTHDF1 can also bind EIF3C mRNA and increase its translation, supporting the progression of ovarian cancer [309]. YTHDF2 forms a double-negative feedback loop with miR-145 and promotes proliferation and migration by reducing global m⁶A levels [310]. The E3 ubiquitin ligase FBW7 (F-box and WD repeat domain-containing 7) can counter this effect by inducing the proteasomal degradation of YTHDF2 and restricting its cancer-promoting effects [311]. In addition, IGF2BP1 works as an m⁶A reader that can stabilise SRF mRNA to promote ovarian cancer progression or stabilise MDM2 mRNA to facilitate immune evasion and tumour growth [312,313].

The m⁶A demethylase FTO appears to be tumour suppressive in high-grade serous ovarian carcinoma. Huang et al. found that FTO reduces m⁶A and the stability of PDE1C and PDE4B mRNAs, strengthens cyclic adenosine monophosphate (cAMP) signalling and suppresses the self-renewal of ovarian cancer stem cells. In contrast to its pro-tumour roles reported in leukaemia and glioblastoma, these findings show that FTO functions vary by tissue [314]. Conversely, ALKBH5 mainly shows pro-tumour activity in ovarian cancer. Jiang et al. reported that ALKBH5 is highly expressed and is regulated by the toll-like receptor 4 (TLR4) and NF-κB signalling axis in the tumour’s microenvironment. Through m⁶A demethylation, ALKBH5 increases NANOG expression and promotes tumorigenesis [315]. ALKBH5 can also form a positive feedback loop with HOXA10 and drive cisplatin resistance by demethylating Janus kinase 2 (JAK2) mRNA [316].

M⁶A is also involved in shaping the immune microenvironment and chemotherapy-related phenotypes in ovarian cancer. Such m⁶A regulators as ZC3H13, YTHDF1 and IGF2BP1 are tied to tumour immune infiltration [317]. In addition, elevated m⁶A on FZD10 mRNA has been linked to resistance to poly (ADP-ribose) polymerase (PARP) inhibitors by increasing transcript stability and activating Wnt and β-catenin signalling [318].

Bioinformatic analyses using public resources, such as TCGA and GEO, indicate that the dysregulated expression of m⁶A regulators is tightly linked to ovarian cancer prognoses. Fan et al. used LASSO Cox regression to identify IGF2BP1, VIRMA and ZC3H13 for a risk score model. Higher expression was tied to worse outcomes, and these genes were enriched in WNT signalling and cancer-related pathways [319]. Zhu et al. broadened prognostic models by including KIAA1429, YTHDC2, HNRNPC and WTAP and noted that the resulting risk score predicted survival regardless of age and tumour stage. In this model, a higher expression of KIAA [320].

Other RNA marks outside m⁶A also promote ovarian cancer’s progression. In platinum resistance, m⁵C regulates the NSUN2 ALYREF LGR4 axis that enhances the tolerance to cisplatin. NSUN2 adds m⁵C to the coding region of leucine-rich repeat-containing G-protein coupled receptor 4 (LGR4) mRNA, a leucine-rich repeat-containing G-protein coupled receptor that is recognised by ALYREF. This interaction increases transcript stability and supports nuclear export, leading to the activation of Wnt and β-catenin signalling and, finally, the development of a platinum-resistant phenotype [321].

Additionally, m7G plays a role in ovarian cancer. METTL1 has been reported to increase BRCA1 expression by installing m7G on breast cancer susceptibility gene 1 (BRCA1) mRNA and promoting proliferation and tumour growth in high-grade serous ovarian cancer. A vaginal microbiota-derived metabolite, 5-formamidoimidazole-4-carboxamide ribotide (FAICAR), can bind METTL1 and inhibit its methyltransferase activity, reduce m7G on BRCA1 and suppress tumour progression. These findings suggest that host microbe interactions can influence ovarian cancer through epitranscriptomic regulation [322]. In the regulation of stem-like traits, the ac4C acetyltransferase NAT10 increases the stability of cytoplasmic activation/proliferation-associated protein 1 (CAPRIN1) mRNA through ac4C and thus promotes ovarian cancer cell migration, invasion and maintenance of stem-like properties [323]. Currently, the evidence linking METTL3 to ovarian cancer progression is robust across multiple independent cohorts, whereas the roles of rarer modifications such as ac4C and m7G in reproductive health remain in the exploratory phase, primarily supported by descriptive sequencing data rather than mechanistic validation [134,304].

The aberrant regulation of RNA methylation is an important molecular mechanism in endometrial cancer (EC). That is, m⁶A and its regulators show changed expression patterns in EC tissue and drive tumour proliferation, invasion and metastasis by affecting mRNA stability, translation and signalling pathway activity. Low METTL3 expression or functional defects in METTL14 are common in EC. Reduced m⁶A can enhance AKT signalling by decreasing the AKT-negative regulator PHLPP2 and increasing the positive regulator mTORC2, promoting EC cell proliferation and tumorigenicity [324]. Conversely, WTAP is highly expressed in EC and promotes progression by installing m⁶A within the three prime untranslated regions of Caveolin-1 (CAV-1) mRNA and reducing its expression, which activates NF-κB signalling [325].

M⁶A readers have dual effects on EC. YTHDF2 works as a tumour suppressor and can bind m⁶A-marked insulin receptor substrate 1 (IRS1) mRNA and promote its decay, limiting IRS1 and AKT signalling [326]. Conversely, IGF2BP1 works as an oncogene and is highly expressed in EC. IGF2BP1 recognises m⁶A sites in the three prime untranslated regions of Paternally Expressed Gene 10 (PEG10) mRNA and recruits Polyadenylate-binding protein 1 (PABPC1) to increase PEG10 mRNA stability, which promotes EC cell proliferation. Higher IGF2BP1 expression is also tied to worse outcomes [327]. Reduced m⁶A levels may further promote EC proliferation and tumour formation through a pregnancy-associated plasma protein-A (PAPPA) and insulin-like growth factor binding protein 4 (IGFBP4) axis that modulates extracellular signal-regulated kinase (ERK) along with NF-κB and AKT signalling [328].

Among erasers and related enzymes, both FTO and ALKBH5 drive EC progression. FTO limits the decay of Homeobox B13 (HOXB13) mRNA through demethylation, increases HOXB13 protein levels and activates Wnt signalling to enhance invasion and metastasis [329]. ALKBH5 increases the stability and translation of insulin-like growth factor 1 receptor (IGF1R) mRNA through demethylation and promotes EC cell proliferation [329]. Under hypoxia, ALKBH5 can also maintain the expression of SRY-box transcription factor 2 (SOX2) mRNA and strengthen stem-like phenotypes in EC [330]. In addition, Peptidylarginine Deiminase 2 (PADI2) can activate ERK1 and ERK2 signalling through MEK1 citrullination and enhance the IGF2BP1-mediated stabilisation of SOX2 mRNA. This PADI2 MEK1 ERK IGF2BP1 SOX2 axis may drive malignant progression in EC [331].

Prognostic models based on m⁶A regulator expression could aid in EC risk assessment. Low FTO expression and high RBM15 expression have been tied to poor prognoses and a higher FIGO grade in endometrial adenocarcinoma. A risk score derived from FTO, RBM15 and YTHDF1 can predict survival outcomes. Various target genes, including IGF1 and IRS1, can influence energy metabolism, RNA stability and connective tissue formation. The progression of this malignancy may also be influenced by FTO through its effects on EMT and chemotherapy sensitivity, highlighting the possibility that targeting m⁶A enzymes and associated pathways may be useful for the design of precision treatment strategies in endometrial adenocarcinoma [332].

RNA modifications beyond m⁶A also promote EC progression. Post-transcriptional modifications such as m⁵C, m¹A and m⁷G can modulate the stability and function of lncRNAs or messenger RNAs. The accumulation of such modifications can assist with the remodelling of the immune microenvironment, control of the cell cycle and response to treatment in EC. A risk model based on m⁵C-related lncRNAs correlates with PD-L1 expression and immune cell infiltration and can predict patient outcomes [333]. Patterns associated with m1A have been linked to immune pathway enrichment and tumour mutational burden and can distinguish between immune-inflamed and immune-cold phenotypes [147]. M7G is a type of cap found on the messenger RNA molecule of eukaryotic cells. Such m7G-related mRNAs have been connected to tumours’ progression through effects on stemlike properties, immune infiltration and drug sensitivity [334]. Collectively, our findings reveal that multiple RNA modifications collaborate to regulate EC development and progression and that these findings may identify new targets for patient stratification and precision therapy.

Choriocarcinoma is a highly invasive tumour of trophoblastic origin that invades blood vessels early and metastasises widely. The tumour that arises due to the uncontrolled multiplication of cytotrophoblasts and syncytiotrophoblasts can be gestational or non-gestational. The gestational type tends to be relatively chemosensitive; however, nearly 30% show systemic metastasis on the first presentation and 5% display multidrug resistance at presentation. The form that is not gestational has a more aggressive course: 84% of patients eventually die from the disease despite multiagent chemotherapy, indicating broad resistance to therapy. Distinguishing these two origins remains a clinical challenge that can affect the choice of therapy and prognosis assessment [335,336]. Recent findings suggest that m⁶A likely plays a role in the progression of choriocarcinoma. This modulation may occur through an ncRNA network, though the associated mechanistic details remain elusive.

METTL3, the m⁶A writer, was identified as a critical regulator of choriocarcinoma progression. METTL3 plays a role in processing pri-miR-935, which reduces the expression of gap junction protein GJA1 (Cx43). This suppression weakens intercellular communication and enhances the growth and invasion of choriocarcinoma cells [93]. METTL3 enhances the maturation of pri-miR-21 and boosts miR-21-5p levels. The latter targets the 3′-untranslated region of hypoxia-inducible factor asparagine hydroxylase (HIF1AN) mRNA, enhancing its decay. The reduced expression of HIF1AN relieves the inhibition of HIF1A, enhancing both HIF1A activity and vascular endothelial growth factor (VEGF) expression. Optimised angiogenesis through this pathway supports the growth and spread of tumours [337].

Some ncRNAs play tumour-suppressive roles in choriocarcinoma. The upregulation of miR-373-3p in cancer inhibits invasion and metastasis by targeting transforming growth factor β receptor II (TGFβR2), reducing TGFβ signalling and the epithelial–mesenchymal transition [338]. According to these observations, m⁶A regulatory networks exert context-dependent bidirectional effects in choriocarcinoma on distinct RNA substrates and cell contexts.

Besides ncRNA modulation, m⁶A directly modulates the stability and expression of protein-coding transcripts. In an m⁶A-dependent manner, phospholipase Cε1 (PLCE1) mRNA is stabilised by METTL14. Subsequently, PLCE1 induces VEGF secretion with Ras-associated protein 1 (Rap1) signalling, which promotes proliferation, the epithelial–mesenchymal transition and angiogenesis [339]. Previously, METTL3 was shown to interact with YTHDF3 to deposit m⁶A onto phosphoglycerate kinase 1 (PGK1) mRNA, thereby enhancing both transcript stability and translational efficiency to drive glycolytic reprogramming and malignant progression [340].

There is still insufficient knowledge about how epitranscriptomics influence the disease. Thus far, research has focused on m⁶A, especially writers (METTL3 and METTL14) and readers (YTHDF3). The systematic characterisation of demethylases, including FTO and ALKBH5, along with other alternative YTH family members, has yet to be achieved. CircRNA and lncRNA are excellent candidates for other marks, such as m⁵C and A-to-I RNA editing, and other modifications that remain to be deciphered. Broadening the research scope beyond m⁶A-centric paradigms for the construction of the interactomes of multiple RNA modifications will enable a better understanding of epitranscriptomic governance in choriocarcinoma and facilitate the development of precision therapeutics.

While the oncogenic roles of RNA modifications in gynaecologic malignancies are well established, their influence extends to a broader range of reproductive health issues. The precise spatio–temporal regulation of the epitranscriptome is equally critical for the maintenance of pregnancy. Disruption of these finely tuned mechanisms can lead to a variety of pregnancy-related disorders, including preeclampsia, gestational diabetes mellitus and miscarriage, as discussed below.

Preeclampsia (PE) is a pregnancy disorder characterised by fresh hypertension, along with proteinuria or end-organ dysfunction. Approximately 2% to 5% of pregnancies worldwide are affected by this condition. It is a leading cause of maternal and perinatal mortality [341]. The inadequate trophoblast invasion at an early stage and the failure to remodel the uterine spiral arteries are believed to be the underlying causes of pathogenesis, leading to ischemia and hypoxia of the placenta. These insults lead to the haemodynamic release of antiangiogenic factors and proinflammatory mediators, ultimately resulting in maternal endothelial injury, systemic inflammation and multi-organ dysfunction [342]. At present, screenings and prevention strategies are limited. Low-dose aspirin confers advantages only to certain high-risk groups, and delivery is the sole curative intervention. In addition, affected pregnancies bear a notably higher long-term cardiovascular risk [343]. Deciphering epigenetic regulatory mechanisms, such as RNA modifications, represents a major goal for understanding disease pathophysiology and forging new therapeutic interventions.

Analyses of epitranscriptomic patterns reveals heterogenous m⁶A in PE-induced placentas. Taniguchi et al. linked m⁶A deposition within placental mRNAs’ 5′-untranslated region (5′UTR) to foetal growth restriction and PE [264]. Zhu et al. reported increased levels of m⁶A and METTL3 expression in trophoblast cells from the placenta of PE [344]. Conversely, Bian et al. demonstrated that via an IGF2BP1-HMGN3 axis, reduced m⁶A could block trophoblast invasion [345]. This variation likely mirrors the heterogeneity underlying PE pathogenesis. However, some studies focusing on biomarkers are limited by small cohort sizes and a lack of multicentre validation, which weakens the statistical power and clinical generalisability of the findings [346].

The activity of m⁶A writers and readers modifies trophoblast behaviour via specific target transcripts. METTL3 promotes trophoblast invasion by stabilising myosin light chain kinase (MYLK) mRNA; disruption of this mechanism is associated with PE [266]. The activity of m⁶A writers and readers modifies trophoblast behaviour via specific target transcripts. METTL3 promotes trophoblast invasion by stabilising myosin light chain kinase (MYLK) mRNA; disruption of this mechanism is associated with PE [347]. Further studies depict that METTL3 stabilises acyl-CoA synthetase long-chain family member 4 (ACSL4) mRNA and promotes trophoblast ferroptosis, thus accelerating PE development [348]. The uplifted expression of METTL14 is also associated with the dysfunction of trophoblasts through m⁶A deposition on forkhead box O3a (FOXO3a) mRNA and the increased stability of transcripts [349]. The overexpression of RBM15 enhances the binding of YTHDF2 to CD82 mRNA, thereby speeding up its degradation and reducing the expression of CD82 mRNA [350].

According to the impact of the writer and the reader, an m⁶A eraser possesses dual regulatory abilities in PE. PE placentas demonstrate increased levels of ALKBH5, which may worsen the disease by reducing the stability of the peroxisome proliferator-activated receptor γ (PPARγ) mRNA [269]. Nevertheless, ALKBH5 enhances the expression of placenta-specific 8 (PLAC8) by eliminating m⁶A from PLAC8 mRNA, thus facilitating trophoblast invasion and migration, which may be advantageous in PE [351]. ALKBH5 confers a context-dependent function, so the directionality of activity is determined by the specific cell type and cell-type-specific downstream targets.

Beyond mRNAs, m⁶A regulates the function of non-coding RNAs. Wang et al. uncovered that the maintenance of the stability of the HOXD cofactor cluster antisense RNA 1 (HOXD-AS1) was influenced by METTL3-mediated m⁶A modification on long non-coding RNAs. HOXD-AS1 sequesters cellular microRNA miR-135a and promotes the expression of β-transducin repeat-containing protein (β-TRCP) and NF-κB signalling to modulate PE progression [352]. Zhang et al. discovered that the m⁶A manipulation of linc01116 takes place with the help of IGF2BP2. Linc01116, which is lessened in PE placentas, sponges miR-210-3p to promote angio-associated migratory cell protein (AAMP) expression, which promotes trophoblast angiogenesis and invasion [353]. A circSETD2-miR-181a-5p-myeloid cell leukaemia 1 (MCL1) axis represents another regulatory node regarding circRNA regulation. The METTL3 m⁶A-mediated stabilisation of circSETD2 functions as a competing endogenous RNA that binds to miR-181a-5p, relieving MCL1 suppression and alleviating trophoblast apoptosis [354]. Likewise, in placentas with PE, m⁶A mediated by METTL14 on circPAPPA2 increases, with the stability of circPAPPA2 being sustained by IGF2BP3 in an m⁶A-dependent manner. PE loss hampers circPAPPA2 and alters trophoblast invasion via IGF2BP3 expression [355].

Family members NOP2/Sun RNA methyltransferase and m⁵C participate in PE pathology. Enoxaparin enhances trophoblast functioning by activating NSUN2-mediated m⁵C and stabilising the mRNA of paired box 3 (PAX3) [356]. NSUN5 is associated with decidualisation by regulating the cellular energy metabolism through its interaction with ATP synthase F1 subunit beta (ATP5B) [357]. An R295C variant at rs77133388 in NSUN5 dampens IL-11Rα-JAK2-STAT3-Cyclin D3 signalling, impairs decidual cell polyploidisation and thereby increases the risk for PE [358]. The m⁵C regulatory network contributes to the pathogenesis of PE. Similarly to PE, other common pregnancy complications, such as gestational diabetes mellitus, exhibit profound alterations in the epitranscriptomic landscape, highlighting the broad relevance of RNA modifications in gestational health.

Gestational diabetes mellitus (GDM) is the most common metabolic complication in pregnancy. This condition occurs when there is worsening physiological insulin resistance during pregnancy, together with insufficient compensatory ability of pancreatic β-cells [359]. Despite the widespread acceptance of the diagnostic criteria laid down by IADPSG and WHO 2013, considerable heterogeneity in prevalence estimates is produced by regional differences in screening strategies. GDM increases the risk of hypertensive disorders during pregnancy and foetal overgrowth in the mother. It also increases both the mother’s and child’s risk for metabolic disease through in utero metabolic programming. Understanding the molecular mechanisms of GDM, especially epigenetic RNA modification leading to glucose homeostasis, will help ameliorate diagnosis and treatment [360].

The expression of METTL14, a component of the m6A methyltransferase complex, is altered in GDM. Placental tissue from patients with GDM has lower METTL14 expression and m6A deposition on BMP- and activin-membrane-bound inhibitor (BAMBI) mRNA and BAMBI expression. Alterations in these pathways may modulate transforming growth factor-β (TGF-β) and Wnt signalling, impair β-cell function and promote insulin resistance, which raises the risk of GDM [361]. METTL14 in GDM pathogenesis through the XIST-miR-497-5p-FOXO1 axis. In high-glucose-cultured HTR-8/SVneo cells, METTL14 mediates the m⁶A-dependent silencing of X-inactive specific transcript (XIST) in high-glucose cultured HTR-8/SV neo cells. The silencing of this gene decreases the sequestration of microRNA-497-5p (miR-497-5p) by XIST and increases miR-497-5p levels, downregulating forkhead box O1 (FOXO1). These modifications increase the proliferation and migration of trophoblasts while reducing apoptosis, which may inhibit GDM development [362]. M⁶A on chemokine ligand 5 (CCL5) by METTL14 controls CCL5 mRNA stability and influences the proliferation, migration and apoptosis of trophoblasts, thereby altering GDM progression [363].

Outside METTL14, METTL3 is active in the regulation of GDM. The M⁶A methylation of hsa_circ_0072380 by METTL3 reduces circular RNA stability and may affect the GDM mechanism by altering trophoblast proliferation, migration and invasion [364]. Regarding m⁶A demethylases, FTO demethylates the mRNA of salt-inducible kinase 1 (SIK1), ultimately reducing its stability and expression. Thus, cytotrophoblast fusion is inhibited, leading to defective placental syncytialisation [365]. Concerning m⁶A demethylases, FTO demethylates the mRNA of salt-inducible kinase 1 (SIK1), ultimately reducing its stability and expression. Thus, cytotrophoblast fusion is inhibited, leading to defective placental syncytialisation [366]. RBM15 regulates the m⁶A level of claudin 4 (CLDN4) gene expression and plays a role in the glucose and lipid metabolism in the offspring of GDM mice. It might cause insulin resistance [367]. Hesperidin is a therapeutic, natural compound of potential. Trophoblast cells from obese GDM patients have increased m⁶A levels; hesperidin exerts protective effects through autophagy and m⁶A methylation [368].

The complicated and rich world of idolatries advances as we enter the world of idolatries. Under high-glucose conditions, NSUN2 modifies PTEN-induced putative kinase 1 (PINK1) mRNA and alters mitophagy, events that impair trophoblast functioning. The inhibition of NSUN2 enhances the glucose metabolism and decreases inflammation in pregnant GDM mouse models [369]. Furthermore, GDM placenta tissue accumulates n4-acetylcytidine. N-acetyltransferase 10 (NAT10) alters mesenchyme homeobox 2 (MEOX2) mRNA while regulating human umbilical vein endothelial cell (HUVEC) angiogenesis. NAT10 inhibition alleviates the endothelial dysfunction induced by high glucose concentrations [370].

RNA modifications operate as fundamental modulators of GDM, functioning across trophoblast functions, placental angiogenesis and insulin sensitivity through marks including m⁶A, m⁵C and ac4C. Nevertheless, much evidence is derived from in vitro cell systems and animal models, while crosstalk among modification enzymes and pregnancy-stage-specific modification dynamics awaits elucidation. Prospective investigations warrant the integration of multi-omics approaches to clarify GDM-specific interactions among RNA modifications and the development of targeted interventions against key enzymes capable of acting across the placental barrier. Such advances may help curtail the pathological transmission of in utero metabolic programming and improve long-term outcomes for both mother and child.

Miscarriage denotes involuntary pregnancy loss before 24 weeks of gestation, affecting approximately 23 million cases globally each year (44 cases per minute), representing over 15% of all clinically confirmed pregnancies. Recurrent spontaneous abortion (RSA) traditionally refers to the loss of three or more consecutive pregnancies; however, the European Society of Human Reproduction and Embryology revised this definition in 2018 to encompass two or more non-consecutive pregnancy losses, a condition affecting roughly 1–2% of women of reproductive age [371,372,373]. Although chromosomal abnormalities, endocrine disorders and anatomical factors constitute established aetiologies, approximately half of recurrent miscarriages are unexplained. RNA modifications in embryo development and pregnancy maintenance have drawn growing research interest, offering new insights into molecular bases of idiopathic pregnancy loss.

The expression of METTL3 is reduced in placental villi from patients with RSA. The m⁶A modifications induced by METTL3 generally destabilise Zinc Finger and BTB domain containing 4 (ZBTB4) mRNA and inhibit its expression. Decreased METTL3 activity allows the accumulation of ZBTB4, which decreases the proliferation and migration of trophoblasts [374]. Conversely, environmental pollutants such as benzo(a)pyrene (BaP) and its metabolite benzo(a)pyrene-7,8-dihydrodiol-9,10-epoxide (BPDE) upregulate METTL3 expression, increasing the m⁶A levels of lncRNA lnc-HZ09 and enhancing stability. This change inhibits the phospholipase d1 (PLD1)/RAC family small GTPase 1 (RAC1)/cell division cycle 42 (CDC42) signalling axis, reducing trophoblast migration and invasion and leading to RSA [375]. Exposure to BPDE damages pyroptosis via the lnc-HZ14/Z-DNA binding protein 1 (ZBP1)/NLRP3 axis. ZBP1 transcription is promoted by lnc-HZ14 through interferon regulatory factor 1 (IRF1). It is also involved in NLRP3 mRNA m⁶A modification and stabilisation through METTL3, which leads to pregnancy loss [376].

At the same time, METTL14 expression increased in human trophoblasts and villous tissue from recurrent miscarriage patients exposed to BPDE. The METTL14 gene leads to increased lnc-HZ01’s m⁶A modification and enhanced RNA stability. This process creates a feedback loop involving MXD1, METTL14 and lnc-HZ01. The abnormal triggering of this loop significantly inhibits the growth of trophoblasts. Mechanistically, lnc-HZ01 upregulates transcription factor c-JUN and deubiquitinating enzyme ubiquitin-specific protease 36 (USP36), which increases the mRNA transcription and protein stability of MAX dimerisation protein 1 (MXD1). Later, MXD1 activates lnc-HZ01 expression by inducing METTL14 transcription to catalyse m⁶A methylation [377].

The literature presents paradoxical findings regarding the role of the m⁶A demethylase ALKBH5 in RSA, highlighting a critical need for the rigorous appraisal of the experimental models. ALKBH5 is notably upregulated in placental villous tissue from RSA patients, leading to a global reduction in m⁶A levels. This downregulation shortens the half-life of cysteine-rich protein 61 (CYR61) mRNA, disrupting trophoblast invasion while reducing VEGF secretion, which impairs macrophage recruitment and M2 differentiation at the maternal–foetal interface [377]. Conversely, Zheng et al. demonstrated that ALKBH5 expression was markedly decreased in the trophoblasts of RSA patients. In their model, hypoxia induces the nucleocytoplasmic translocation and enrichment of ALKBH5, which removes m⁶A marks to enhance the expression of matrix metalloproteinase 9 (MMP9) and integrin α1 (ITGA1) via the SMAD1/5 signalling axis, promoting trophoblast viability and invasion [378]. The varying findings stem from cell-type specificity and oxygen–environment dependence. ALKBH5 acts through the VEGF and SMAD1/5 pathways in uterine stromal cells and extravillous trophoblasts. Moreover, hypoxia-inducible factor 1-alpha (HIF-1α) signalling is activated by different oxygen tensions, modulating early pregnancy trophoblast activity. The specificity and environmental dependence of target gene selection suggest that ALKBH5 is context dependent and may modulate pregnancy maintenance in opposite ways. First, the spatio–temporal heterogeneity of placental development means that ALKBH5 activity may vary significantly between the first and second trimesters. Second, the methodological differences in m6A quantification (e.g., MeRIP-seq vs. LC-MS/MS) and the diverse cell types (primary villi vs. immortalised cell lines such as HTR-8/SVneo) utilised across studies likely contribute to these inconsistent findings. Therefore, future researchers must employ single-cell resolutions and standardised functional assays to resolve these contradictions and clarify whether ALKBH5 serves as a protective or detrimental factor in the maternal–foetal interface [162].

The downregulation of another m⁶A demethylase, FTO, occurs in the maternal–foetal interface of spontaneous abortion patients. FTO deficiency leads to abnormal m⁶A build-up in villous tissue and oxidative stress, which downregulates human leukocyte antigen-G (HLA-G) and VEGFR expression while upregulating matrix metalloproteinases (MMP) 7 and MMP9. These changes affect maternal–foetal immune tolerance and angiogenesis, thus causing a miscarriage [379].

The m7G modification in unexplained recurrent spontaneous abortions or unexplained recurrent spontaneous abortions (URSA) is being researched. Transcriptomic analyses of decidua depicted the upregulation of Sm-like protein 1 (LSM1) with the downregulation of La-Related Protein 1 (LARP1) and Nuclear Cap-Binding Protein Subunit 2 (NCBP2) in URSA patient decidua. Possible diagnostic biomarkers could be these three proteins. Furthermore, URSA patients displayed a low density of decidual regulatory T cells (Tregs) and a high density of helper T cells (Th). Notably, LSM1 was negatively correlated with the infiltration of Tregs, while LARP1 and NCBP2 exhibited a positive correlation with Tregs [380]. According to these observations, m7G modification helps regulate Treg homeostasis and the maternal–foetal immune tolerance imbalance. This offers a new epitranscriptomic perspective in understanding the mechanisms of immune dysregulation in URSA, which indicates that m7G regulatory factors may be targeted for precise diagnosis and therapy in URSA.

RNA modifications also play a pivotal role in the pathogenesis of chronic reproductive endocrine disorders, such as PCOS and premature ovarian insufficiency. While pregnancy disorders often involve acute placental or immune dysregulation, endocrine disorders typically reflect long-term disturbances in ovarian and metabolic homeostasis, both underpinned by complex epitranscriptomic networks (Table 2).

PCOS is the most common endocrinological and metabolic disorder among women of reproductive age and affects nearly 10% of the world’s women. This condition is associated with hyperandrogenism, ovulatory dysfunction and polycystic ovarian morphology. PCOS is a genetic predisposition and metabolic disorder affected by the environment. The core pathophysiological mechanisms, that is, insulin resistance (IR), compensatory hyperinsulinemia and adipose tissue dysfunctions, set up a vicious cycle between the metabolic and reproductive axes [394,395]. With RNA modifications being important molecular mechanisms that affect epigenetic regulation in metabolic homeostasis and germ cell development, understanding their functional roles in PCOS pathophysiology could unravel molecular aetiology and suggest targeted interventions.

In a meta-analysis of 2548 PCOS patients, Wojciechowski et al. found that the fat mass and obesity associated gene (FTO) rs9939609 variant has almost double the genetic effect on body mass index (BMI) and body weight in PCOS compared to the general female population. This observation indicates that the different metabolic environments or genetic backgrounds of PCOS influence FTO variations in obesity-related traits. This study mainly included European white populations and may not be applicable to Asian groups [396]. A separate study of Chinese women with a PCOS diagnosis found no significant independent association of FTO variation with the obesity phenotype, suggesting the effect is ethnic specific [397]. The granulosa cell (GC) expression of FTO in patients with PCOS is increased, showing a negative correlation with intracellular m⁶A levels and a positive correlation with follicular fluid androgen concentrations. Through its mechanism of action, androgen receptor (AR) activation triggers increased FTO levels. Subsequently, FTO enhances steroid hormone biosynthesis, which worsens hyperandrogenism while creating a positive feedback loop [226]. In another context, FTO reduces the m⁶A modification of the flotillin 2 (FLOT2) mRNA and enhances its stability, promoting GC proliferation, etc. This FTO/FLOT2 axis is considered crucial in the pathological development of PCOS [227].

Apart from FTO, the altered expression of IGF2BP2 leads to GC dysfunctioning in PCOS. The expression of IGF2BP2 is upregulated in the GCs of PCOS patients and promotes cell proliferation via the mRNA binding and stabilisation of cyclin D2 (CCND2) and SERPINE1 mRNA binding protein 1 (SERBP1) mRNAs [398]. In PCOS patients, GCs exhibit increased expression of YTHDF2. The translation-activating protein for mitochondrial mss51 mRNA of this reader is recognised, bound and downregulated, impairing mitochondrial function and glycolysis and causing GC dysfunction [384].

Intestinal microbe dysbiosis is closely associated with the m⁶A epitranscriptomic abnormalities of PCOS. Investigations have demonstrated that obese PCOS patients exhibit aberrant gut microbiota diversity accompanied by reduced serum butyrate levels. Butyrate supplementation inhibits the expression of METTL3, reduces m⁶A modification on FOS-like antigen 2 (FOSL2) mRNA and inhibits NLRP3 inflammasome activation and inflammatory cytokines, such as IL-6 and TNF-α, which improves ovarian functioning and inflammatory status [227]. In luteinised GCs from non-obese PCOS patients, for example, FOXO3 mRNA, m⁶A modification levels are significantly downregulated, which boosts FOXO3 mRNAs’ stability and YTHDF2-mediated RNA degradation and expression and influences apoptosis metabolism and proliferation [228].

A PCOS patient may be phenotypically obese or lean, insulin resistant or noninsulin resistant. The precise regulatory schemes of m⁶A modifications among subtypes have yet to be defined [67]. The m⁶A-modifying enzymes and reader proteins have the potential of being useful diagnostic biomarkers and therapeutic targets for PCOS. However, large-cohort studies are needed to confirm the molecular relationships of genetic variations in FTO with phenotypes such as ovulatory dysfunction and the regulatory roles of dietary habits and metabolic states on m⁶A modifications in women with PCOS [61].

NAT10-mediated ac4C modification also regulates PCOS. PCOS patient ovarian tissues exhibit significantly reduced NAT10 expression and ac4C levels. NAT10 restricts abnormal GC proliferation by enhancing the mRNA stability of C-X-C motif chemokine ligand 14 (CXCL14). Animal experiments have established that localised ovarian NAT10 overexpression ameliorates cystic follicle formation and hyperandrogenism via the NAT10/CXCL14 axis [399]. Simultaneously, GC pyroptosis in PCOS is controlled by NSUN7. Its expression is elevated in PCOS patients, and this enzyme stabilises the mRNA of NLRP3 through m⁵C catalysis and promotes the assembly of the NLRP3 inflammasome and cellular pyroptosis. The inhibition of NSUN7 ameliorates hormonal imbalances and the pathology of the ovary, suggesting that this axis is a potential target for PCOS [231]. Furthermore, significant A-to-I RNA editing alterations occur in the GCs of PCOS patients. Transcriptome analysis identified 545 differential editing sites that are significantly enriched in genes nucleoporin 43 (NUP43) and retinoblastoma binding protein 4 (RBBP4), which are involved in apoptosis and necroptosis. The downregulation of ADARB1, an RNA editing enzyme, in the PCOS condition might indicate that A-to-I editing might influence PCOS development by altering the expression of GC survival [106]. This study offers significant insights into the regulatory mechanism of RNA modifications in PCOS. It will help develop disease-specific therapeutics and add to the understanding of the molecular heterogeneity of the condition.

Premature ovarian insufficiency (POI), defined as ovarian failure before the age of 40, presents as amenorrhea and elevated FSH. Its global prevalence approximates 1–3.5%. However, this figure shows significant ethnic and geographic variations. The depletion of the primordial follicle pool is accelerated, the premature atresia of follicles occurs, and meiotic regulation is disturbed. Infertility and long-term complications such as osteoporosis, cardiovascular disease and neurocognitive impairment are associated with these processes [400,401]. While investigations of FMR1 premutation, genetic defects in meiosis and the like targeted variations at the DNA sequence level, more recent evidence points to epitranscriptomic regulation, particularly RNA modifications, to maintain follicular homeostasis by regulating mRNA stability and translational efficiency, as well as the metabolic reprogramming of oogenesis-associated genes. Understanding the RNA modification regulatory networks involved in the pathogenesis of POI is important for revealing the molecular mechanism of ovarian ageing, further screening of early diagnostic biomarkers and developing precise intervention strategies for POI.

The genetic mechanisms linked to POI are strongly associated with YTHDC2 loss-of-function mutations. The exome sequencing of early-onset POI patients revealed the presence of the missense mutation c.2567C > G (p.P856R) and nonsense mutation c.1129G > T (p.E377) in YTHDC2. The p.P856R mutation is within the helicase-associated domain (HA2 domain), which impairs RNA processing due to the misfolding of RNA helicase conformational flexibility. The nonsense mutation p.E377 truncates the helicase core region [385]. YTHDC2 is essential for the completion of meiosis in mammalian germ cells. Animal studies reveal that the ovaries of female mice lacking Ythdc2 are significantly smaller, with a histology showing the arrest of follicular development at the primordial and primary follicle stages. These follicles cannot continue the secondary or antral stages, exhausting the follicle pool and ovarian reserve. In male germ cells, Ythdc2 absence does not prevent spermatogonial entry into meiosis but leads to arrest at the pachytene-to-diplotene transition, preventing normal diplotene spermatocyte formation and producing abnormal metaphase figures [402]. The use of single-cell transcriptomic analysis reveals that germ cells missing YTHDC2 display mixed mitotic and meiotic transcriptomic profiles after meiotic initiation, indicating the failure to accomplish the mitosis-to-meiosis transition and thus arrests in pre-meiotic [203].

The use of single-cell transcriptomic analysis reveals that germ cells missing YTHDC2 display mixed mitotic and meiotic transcriptomic profiles after meiotic initiation, indicating failure to accomplish the mitosis-to-meiosis transition and thus arrest in pre-meiotic [189]. FTO protects GCs from oxidative stress damage via the miR-642a-5p/FOXO1 signalling axis by stabilising exosomal circular RNA BRCA1 (circBRCA1) [225].

FTO also affects ovarian ageing via m⁶A demethylation. In aged GCs and FTO, downregulation results in abnormal m⁶A-fos mRNA accumulation, increased stability and high expression, causing GC senescence. ROS also activate this pathway through FTO downregulation [403]. In contrast, FTO promotes MMP2 mRNA degradation in a YTHDF2-dependent manner and activates the ERK signalling pathway to delay GC ageing [404]. Essential to primordial follicle pool formation, nuclear-localised FTO is involved in early folliculogenesis. It regulates the alternative splicing of CDK5 via m⁶A demethylation, promoting GC proliferation and cell cycle progression [204]. Additionally, the m⁶A demethylation of long interspersed nuclear element 1 (LINE1) RNA mediated by FTO in embryonic stem cells modifies chromatin’s accessibility, which is involved in regulating oocyte development and early embryogenesis. It awaits to be elucidated at the earliest opportunity that its specific regulatory networks pervade ovarian ageing or the follicular microenvironment [270]. The m⁶A demethylase ALKBH5 is involved in the development of germ cells. The deletion of ALKBH5 significantly enhances m⁶A modification levels in mouse testicular cells and causes the aberrant expression of genes associated with spermatogenesis and apoptosis, impeding male germ cells’ functioning [22]. The mechanisms of ALKBH5 in POI have yet to be defined. Nonetheless, its evolutionarily conserved function to maintain m⁶A modification homeostasis in germ cells suggests that it might regulate ovarian functions.

Iatrogenic POI is a common complication among young cancer survivors. Alkylating agents such as cyclophosphamide (CTX) and selective oestrogen receptor modulators such as tamoxifen (TAM) damage ovarian functioning by disrupting m⁶A homeostasis. Specifically, CTX upregulates GC m⁶A levels in a time- and dose-dependent manner, markedly suppressing demethylase FTO expression while elevating the methyltransferases METTL3 and METTL14 [230]. Maternal tamoxifen exposure during pregnancy downregulates FTO and ALKBH5 while upregulating METTL3 and METTL14, precipitating aberrant primordial follicle assembly and diminishing ovarian reserves in offspring [405]. Environmental endocrine disruptors also interfere with ovarian m⁶A modifications: 4-vinylcyclohexene diepoxide (VCD) reduces global ovarian m⁶A levels in rats [406], whereas the pyrethroid pesticide fenvalerate (FEN) elevates METTL3 and m⁶A reader protein YTHDF2 expression, promoting P-body assembly, disrupting primordial follicle formation and increasing offspring’s POI risk [206].

Beyond m⁶A, m⁵C modifications and their methyltransferases participate in POI pathology. The m⁵C methyltransferase NSUN5 maintains maternal mRNA stability by catalysing m⁵C modifications, ensuring a successful maternal-to-zygotic transition (MZT). Its deletion reduces m⁵C levels, accelerates maternal mRNA degradation and inhibits folliculogenesis, oocyte maturation and embryonic development [121]. TRDMT1 mediates m⁵C modifications in GCs, participating in the DNA damage repair induced by ROS and suppressing oxidative stress-related apoptosis. Functional defects in this enzyme are strongly associated with POI onset and progression [123]. Ac4C modifications also participate in POI pathological processes. In a cyclophosphamide-induced POF mouse model, total mRNA ac4C modification levels in granulosa cells rose markedly, with enhanced ac4C modification at specific sites of senescence inhibitor P16 mRNA. Electroacupuncture treatment reduces ac4C modifications by suppressing NAT10 expression, diminishing P16 mRNA stability and expression, alleviating the inhibition of cyclin-dependent kinase 6 (CDK6) and cyclin D1 (CCND1), restoring granulosa cell proliferation and ameliorating ovarian function. This observation illuminates the reversibility of epitranscriptomic regulation in iatrogenic POI treatment [162].

Endometriosis (EM) refers to the aberrant implantation of endometrial-like tissue outside the uterine cavity as a chronic oestrogen-dependent condition. EM affects approximately 10% of the world’s reproductive-age women. It presents as pelvic pain that progresses and includes infertility. Pathogenesis involves retrograde menstruation, immune evasion, inflammatory microenvironment remodelling and progesterone resistance, culminating in cyclical bleeding, fibrosis and neovascularisation within ectopic lesions [407,408]. Genetic variation alone does not completely determine the disease risk of susceptibility loci identified by genome-wide association studies, implicating the epigenetic regulation of the pathogenesis of the disease. Investigations have confirmed that epigenetic mechanisms modulate the hormonal responsiveness and invasive phenotypes of endometrial cells. Epitranscriptomic regulation is influenced by RNA modifications that illustrate endometriosis’s molecular pathophysiology.

The pathological progression of EMs involves aberrant m⁶A modification profiles. Patients with endometriosis have lower levels of m⁶A in ectopic endometrial tissue due to lower METTL3 expression. This attenuation weakens DiGeorge syndrome critical region gene 8 (DGCR8)-mediated pri-miR-126 maturation, promoting the invasion and migration of endometrial stromal cells (ESCs) [402]. Conversely, METTL3 has a dual regulatory effect on EMs. This enzyme works with YTHDF2 to alter the m⁶A on sirtuin 1 (SIRT1) mRNA, modulating the SIRT1/FoxO3a signalling axis to trigger cellular senescence, inhibiting the migration, invasion and proliferation of ESCs [388]. METTL3 also provides a protective effect by regulating the alternative splicing of lncRNA MIR17HG [409]. Aside from altering the biological behaviour of ESCs, m⁶A participates in immune microenvironment remodelling. Lactic acid derived from ectopic ESCs upregulates METTL3 expression and promotes the m⁶A modification of Tribbles pseudokinase 1 (Trib1), activating the ERK/STAT3 signalling pathway, inducing M2 macrophage polarisation and fostering ectopic lesions’ progression [391]. Other regulatory factors of m⁶A are involved in EM pathology. The expression of METTL14 is significantly decreased in ectopic endometrium from EM patients. The individual or combined loss of METTL14 and METTL3 promotes the proliferation and invasion of ESC via m⁶A-dependent mechanisms [410]. Other regulatory factors of m⁶A are involved in EM pathology. The expression of METTL14 is significantly decreased in an EM patient’s ectopic endometrium. The individual or combined loss of METTL14 and METTL3 promotes the proliferation and invasion of ESC via m⁶A-dependent mechanisms [392].

Patients with EM exhibit the downregulation of the m⁶A demethylase FTO. The overexpression of FTO decreases the m⁶A levels in the mRNA of autophagy-related gene 5 (ATG5), which further increases autophagy, reduces the expression of pyruvate kinase M2 (PKM2) and ultimately reduces the glycolysis, proliferation and migration of ectopic endometrial stromal cells (EESCs) [389]. In terms of regulating the energy metabolism, METTL3 enhances the recognition of m⁶A by the E3 ubiquitin ligase RNF43 and promotes the ubiquitination and degradation of NDUFS1. This inhibits oxidative phosphorylation in ectopic ESCs that modulate energy metabolism and growth in cells [411]. Aberrant m⁶A changes not only promote local EM lesion progression but also greatly contribute to EM-related infertility. Such patients exhibit elevated METTL3 expression and m⁶A levels while the expression of FOXO1 is reduced [11]. Mechanistic investigations demonstrate that METTL3-mediated m⁶A impairs ESC decidualisation capacity through YTHDF2-dependent FOXO1 mRNA degradation pathways, consequently affecting embryo implantation [387]. Hypoxia-induced ALKBH5 upregulation stabilises histone methyltransferase EZH2 mRNA, elevating H3K27me3 levels at decidualisation markers, including IGFBP1 and prolactin (PRL), producing decidualisation defects and embryo implantation failure [390].

Diagnostic applications indicate that the aberrant expression of m⁶A regulatory factors, such as HNRNPA2B1 and HNRNPC, may associate with immune responses in EM and serve as potential biomarkers for diagnoses [412]. M⁶A modification operates through regulating ESC proliferation, invasion, immune microenvironment remodelling, energy metabolism and decidualisation throughout the onset, progression and the infertility associated with EM. Interventions targeting m⁶A-modifying enzymes and downstream effectors may offer therapeutic avenues for EM.

Ac4C modification also participates in the progression of EM. NAT10-mediated ac4C levels are markedly elevated in ectopic endometrial tissue from EM patients. Ac4C advances the proliferation, epithelial–mesenchymal transition (EMT) and cell cycle progression of endometrial epithelial cells by stabilising transforming growth factor-β1 (TGFβ1) mRNA and enhancing target gene expression [413].

Adenomyosis refers to the benign invasion of the endometrial glands and stroma into the myometrium [414]. Adenomyosis affects parous perimenopausal women who present abnormal uterine bleeding and severe dysmenorrhea. Imaging techniques widely detect adenomyosis in younger reproductive-age women. This condition is closely associated with miscarriages and infertility. Molecular pathogenesis encompasses dysregulated sex hormone receptor signalling, chronic inflammation, aberrant extracellular matrix deposition and angiogenesis. However, current theories fail to adequately account for the correlations between diverse clinical presentations and imaging phenotypes [415]. Adenomyosis commonly presents with endometriosis and uterine fibroids, which complicates the clinical diagnosis and treatment of pelvic disorders [416]. The current understanding of how adenomyosis develops is limited and utilises classical molecular pathways. A probe into new biological levels, such as epitranscriptomic regulation, is likely to explain adenomyosis by describing the nature of the disease.

The abnormal expression of m⁶A modification and regulators are involved in adenomyosis pathology. Bioinformatics analysis demonstrates that adenomyosis patients specifically display the downregulation of METTL3, ZC3H13, FTO and YTHDC1, resulting in reduced overall m⁶A modification levels in their endometrial tissue [393]. There is a significant negative correlation between the expression levels of IGF1 and D-dopachrome tautomerase (DDT) with METTL3. Such factors may enhance endometrial dysfunctioning by inducing epithelial cell proliferation and migration. Total m⁶A levels are similarly reduced within the myometrium, and 11 m⁶A regulatory factors (including METTL3 and FTO) demonstrate differential expression. Researchers also have identified possible targets, such as cadherin 3 (CDH3), sodium channel β-subunit 4 (SCN4B) and placenta-specific protein 8 (PLAC8), which may be involved in regulating cell adhesion, muscle contraction and the immune response [393].

As discussed in previous sections, in reproductive disorders, such as preeclampsia, gestational diabetes, recurrent miscarriage, PCOS and EM, different regulatory patterns of RNA modifications are observed, such as m⁶A, m⁵C, ac4C and m7G. These discussions highlighted the molecular mechanisms of four key modification enzymes, METTL3, FTO, NSUN2 and NAT10, in governing key biological events including trophoblast invasion, maternal–foetal immune tolerance, follicular microenvironment homeostasis and endometrial receptivity that may epitranscriptomically underlie these diseases (Figure 4).

No new data were created or analysed in this study. Data sharing is not applicable to this article.