Multilevel regulation of N⁶-methyladenosine RNA modifications: Implications in tumorigenesis and therapeutic opportunities

Abstract

N⁶-methyladenosine (m⁶A) RNA modification is widely perceived as the most abundant and common modification in transcripts. This modification is dynamically regulated by specific m⁶A “writers”, “erasers” and “readers” and is reportedly involved in the occurrence and development of many diseases. Since m⁶A RNA modification was discovered in the 1970s, with the progress of relevant research technologies, an increasing number of functions of m⁶A have been reported, and a preliminary understanding of m⁶A has been obtained. In this review, we summarize the mechanisms through which m⁶A RNA modification is regulated from the perspectives of expression, posttranslational modification and protein interaction. In addition, we also summarize how external and internal environmental factors affect m⁶A RNA modification and its functions in tumors. The mechanisms through which m⁶A methylases, m⁶A demethylases and m⁶A-binding proteins are regulated are complicated and have not been fully elucidated. Therefore, we hope to promote further research in this field by summarizing these mechanisms and look forward to the future application of m⁶A in tumors.

Introduction

Epigenetics is the study of reversible, heritable changes in genes in the absence of nuclear DNA sequence changes. This field mainly includes DNA modifications (such as methylation) and histone and chromatin modifications. In the 1970s, scientists identified complex base-methyl nucleoside patterns on RNA that were identified as N⁶-methyladenosine (m⁶A).¹ Although m⁶A has been known for decades, the significance of its biological properties has become better understood in recent years due to technical breakthroughs. More than 160 different RNA modifications have been identified to date,^2,3 and of these, m⁶A is the most abundant modification among messenger RNAs (mRNAs), transfer RNAs (tRNAs), ribosomal RNAs (rRNAs) and small nuclear RNAs (snRNAs).4, 5, 6 This modification is widely distributed in more than 7000 mRNAs and 300 noncoding RNA (ncRNAs) transcripts in human cells and is enriched in the 3′-untranslated regions (3′ UTRs) of linear RNAs and near the last exon in ncRNAs.7, 8, 9, 10 Studies have shown that m⁶A plays a critical role in all aspects of RNA metabolism, including stability, splicing, nuclear export and translation.11, 12, 13, 14, 15 In this review, we focus on the mechanisms through which m⁶A methylation is regulated and the role of m⁶A in tumor progression.

m⁶A writers, erasers, and readers

Many studies have confirmed that m⁶A is a reversible modification. Methylases, demethylases and m⁶A-binding proteins are the main proteins involved in the regulation of this dynamic process and are also known as the “writers”, “erasers” and “readers” of m⁶A, respectively (Fig. 1). m⁶A modification is involved in a series of important biological processes.¹⁶

Figure 1.

Mechanisms and molecular functions of m⁶A modification. METTL3/METTL14/WTAP and other methylases form a “writer” complex, which is responsible for the adding of m⁶A to mRNA. This modification is removed by the demethylation enzymes FTO and ALKBH5. In the nucleus, YTHDC1 binds to target mRNA with m⁶A modifications to promote its splicing and also interacts with SRSF3 to promote the binding of target mRNAs to SRSF3 and NXF1, which facilitates the transport of mRNA from the nucleus to the cytoplasm. In the cytoplasm, YTHDF1 promotes mRNA translation with the help of YTHDF3. YTHDF2 recognizes m⁶A and promotes mRNA degradation.

Methylases, also known as the “writer” complex, have a core of methyltransferase-like protein 3/14 (METTL3/14) and WT1-associated protein (WTAP),¹⁷ and RNA-binding motif protein 15/15 B (RBM15/15 B), zinc finger CCCH-type containing 13 (ZC3H13) and vir-like m⁶A methyltransferase associated (VIRMA/KIAA1429) participate in the formation of subunits of this complex.18, 19, 20 Recently, human CCHC zinc finger-containing protein (ZCCHC4) was identified as a novel m⁶A writer that mediates the methylation of 28 S rRNAs.²¹ METTL16 (METTL3 homolog) is considered the “writer” of precursor mRNA and several ncRNAs, including U6 snRNA and long noncoding RNAs (lncRNAs).^22,23 More “writers” in this “writer” complex may be discovered in the future.

Human obesity-associated protein (FTO) and ALKB homolog 5 (ALKBH5), known as the “erasers” of m⁶A, are selective demethylation enzymes that regulate gene expression and cell fate through the oxidative removal of methyl groups from m⁶A.^24,25 In addition, ALKB homolog 1 (ALKBH1) can demethylate the methylcytidine (mC) of mRNA in mammalian cells.²⁶ ALKB homolog 3 (ALKBH3) regulates 1-meA and 3-meC demethylation in endogenous methylated RNA,²⁷ and ALKB homolog 7 (ALKBH7) acts as an RNA demethylase and thereby controls newborn mitochondrial RNA processing and mitochondrial activity.²⁸ These findings add to our understanding of the dynamic modifications of m⁶A.

m⁶A “readers”, which constitute a type of m⁶A-binding protein mainly in the YT521-B homology (YTH) and heterogeneous nuclear ribonucleoprotein (hnRNP) domain families, selectively recognize RNA with m⁶A modification to mediate its degradation, splicing and translation.²⁹ Several m⁶A-binding proteins have been identified, and these include YTH family proteins (YTHDC1/2 and YTHDF1/2/3).^13,30,31 YTHDC1 promotes the splicing of target mRNA by recruiting the pre-mRNA splicing factor SRSF3 while preventing SRSF10 mRNA binding.³² Moreover, the binding of YTHDC1 to SRSF3 promotes the interaction between SRSF3 and nuclear RNA export factor 1 (NXF1) on target mRNAs and transfers mRNA from the nucleus to the cytoplasm.¹³ YTHDF1 enhances the translation efficiency of target mRNA with the help of YTHDF3.¹⁴ Moreover, YTHDF2 can selectively identify and bind mRNA with m⁶A modification to promote degradation.¹¹ Additionally, insulin-like growth factor 2 mRNA-binding proteins (IGF2BP1/2/3),^33,34 hnRNPA2B1 andeukaryotic translation initiation factor 3 (eIF3) have been identified as m⁶A “readers”.^35,36

m⁶A-mediated molecular mechanisms

Early studies indicated that m⁶A modification is involved in pre-mRNA splicing.³⁷ METTL3 significantly affects the p53 mRNA splicing pattern and apoptosis.⁹ FTO regulates the exon splicing of the lipid-forming regulator RUNX1T1 by regulating the m⁶A levels near the splicing site.³⁸ Subsequent studies further confirmed the regulatory effect of FTO on mRNA splicing.³⁹ ALKBH5, another demethylase, is also reportedly involved in the regulation of target mRNA splicing.²⁵ The m⁶A “reader” protein hnRNP regulates the splicing of target mRNAs,⁴⁰ and hnRNP G interacts synergistically with RNA polymerase II (RNAPII) to regulate alternative splicing within the transcriptome.⁴¹ In addition to mammals, m⁶A modification also affects the splicing of mRNA in Drosophila melanogaster and thus affects the expression of sex-determining core genes.¹² Further research on the association between m⁶A and mRNA splicing is needed to reveal its potential role.

“Writers”, “readers”, and “erasers” reportedly regulate mRNA export. The main complex involved in mRNA export is TREX, and the m⁶A methylase complex can regulate mRNA export by recruiting TREX.⁴² METTL3 reportedly regulates the nuclear export of mature mRNA of clock gene Per2.⁴³ ALKBH5 knockout results in accelerated mRNA nuclear export, which suggests that m⁶A plays an important role in regulating mRNA export.²⁵ YTHDC1 interacts with SRSF3 to mediate the export of methylated mRNA from the nucleus to the cytoplasm.⁴⁴

The most identified function of m⁶A on RNA metabolism is its effect on mRNA stability, as demonstrated by metabolic radioisotope labeling studies conducted in the 1970s.⁴⁵ m⁶A modification is negatively correlated with mRNA stability and gene expression in mouse embryonic stem cells (mESCs).⁴⁶ Many subsequent studies have confirmed the effect of m⁶A modification on mRNA stability.47, 48, 49 The demethylation enzymes FTO and ALKBH5 also play a role by regulating the stability of mRNA.^50,51 m⁶A modifications on target mRNAs are recognized by different RNA-binding proteins and mediate different fates, and of these proteins, proteins belonging to the YTHDF family selectively recognize and accelerate target mRNA degradation.^11,52, 53, 54 Unlike the YTHDF family, the IGF2BP family primarily promotes the stability and storage of target mRNAs.³⁴ Therefore, scientists are also considering the possibility of using m⁶A as a marker for the half-life of mRNA.

A large number of studies have linked m⁶A to an increased translation rate. METTL3 enhances mRNA translation by recruiting the translation initiation factor eIF3, whereas deletion of METTL3 inhibits translation.^55,56 Both FTO and ALKBH5 reportedly regulate downstream mRNA translation.^57,58 YTHDF1 actively promotesprotein translation by recognizing m⁶A modifications on target mRNAs.¹⁴ Under heat shock conditions, YTHDF2 recognizes m⁶A modifications located at the 5′UTR of the transcript and facilitates translation.⁵⁹ YTHDF3 regulates mRNA translation due to its interaction with translation-related proteins.⁶⁰ The influence of m⁶A on mRNA metabolism may affect more than one aspect, and different RNA-binding proteins that simultaneously recognize an m⁶A modification might mediate different metabolic processes.

Dysfunction of m⁶A in tumors

Dysregulation of the m⁶A pathway plays an important role in the occurrence and development of diseases, and many studies have shown that m⁶A can promote the progression of tumors, such as breast cancer,⁶¹ lung cancer,⁶² glioblastoma,^63,64 acute myeloid leukemia,^65,66 and liver cancer.⁶⁷ Many reviews have addressed these diseases in detail, and they will thus not be included in this paper. In this review, we focus on the role of upstream factors in tumor regulation through m⁶A.

During the period of tumor occurrence and development, many environmental factors, such as cigarette smoke, pollution and heavy metals, play an important or even decisive role. One study found that exposure to particulate matter (PM), sodium arsenite, bisphenol A (BPA), vinclozolin and other environmental toxicants leads to a significant decrease in the global m⁶A levels.⁶⁸ In long-term diethylnitrosamine (DEN) exposure-induced hepatocellular carcinoma (HCC), FTO is decreased, which promotes HCC development by regulating energy homeostasis and glucose metabolism.⁶⁹ Fusaric acid (FA), a food-borne mycotoxin, downregulates p53 by promoting hypermethylation and reducing the m⁶A level of the p53 promoter, which results in promotion of the occurrence and development of hepatocellular carcinoma.⁷⁰ FA can also regulate the expression of m⁶A-related enzymes to affect the overall m⁶A level in hepatic cells.⁷¹ This finding suggests that FA may regulate the expression of p53 in the same way, but the specific mechanism still needs to be further explored. Fumonisin B is a common contaminant of cereal grains, and exposure to FB increases the overall levels of m⁶ A modification and oxidative stress in human hepatoma (HepG2) cells.⁷² FTO is inhibited by the ethyl ester form of meclofenamic acid (MA2), which enhances the function of the chemotherapy drug temozolomide in restraining the proliferation of glioma cells.⁷³

In the tumor microenvironment, many factors play a role by regulating m⁶A modification. m⁶A is also changed as a result of endogenous stimuli such as heat shock, glucose starvation and oxidative stress.⁷⁴ Glucose starvation triggers autophagy and activates the nuclear factor kappa B (NF-κB) pathway and thereby FTO. FTO has been instrumental in the development of melanoma and anti-PD-1 resistance therapies.⁷⁵ The hypoxia-inducible factors HIF-1α and HIF-2α promote the mRNA and protein expression of NANOG, which encodes pluripotency factors, by enhancing ALKBH5 in breast cancer cells, and NANOG increases the percentage of breast cancer stem cells (BCSCs) and promotes their growth and metastases.⁷⁶ Under the condition of the accumulation of reactive oxygen species (ROS) induced by hypoxia or chemotherapy, the expression of YTHDF1 is downregulated, which in turn promotes the expression of NF-E2 p45-related factor 2 (Nrf2) and its antioxidant genes downstream of aldo-keto reductase 1C1 (AKR1C1) and ultimately induces resistance to cisplatin in non-small cell lung cancer (NSCLC).⁷⁷ Stress granules (SGs) are dynamic structures that include stalled translation mRNA and can affect the translation and stability of mRNA.⁷⁸ In human osteosarcoma cells, oxidative stress promotes formation of the METTL3/METTL14/WTAP complex, which mediates m⁶A modification at the 5′ UTR of mRNA. This modification pattern allows the transport of mRNA from the translatable pool to SGs. Thereafter, the modified mRNA is identified by YTHDF3, and the translation process is blocked.⁷⁹ Under hypoxic conditions, HIF-1α induces YTHDF1 expression, and subsequently, YTHDF1 promotes the translation of ATG2A and ATG14 (autophagy-related genes) in an m⁶A-dependent manner, which contributes to the progression of HCC.⁸⁰ Free radicals, as specific regulatory factors, can directly or indirectly regulate the posttranslational modification of histones⁸¹; therefore, these factors may regulate the expression of m⁶A-related enzymes similarly in response to oxidative stress lesions in the body.

The above discussion notes that many factors, such as environmental factors and oxidative stress, can regulate tumor growth by regulating m⁶A. Although it is well known how they regulate m⁶A, the role that m⁶A plays in this process is less clear. More studies are needed to address this complex question and to uncover the complex functions of m⁶A in tumorigenesis and development.

Multilevel regulation of m⁶A in tumors

Most studies on m⁶A have focused on the regulation of its downstream transcripts, but the potential regulatory mechanism of m⁶A itself has not been fully revealed. To date, some studies have revealed these mechanisms from different aspects.

Transcription-level regulation of m⁶A molecules by transcription factors or histone modifications

The expression of m⁶A can be regulated by changes in transcription factors and alterations in the chromatin state. The initiation of eukaryotic transcription is very complex, and transcription factors play an important role in this process. A previous study found that ALKBH5 could be regulated as a direct transcriptional target of hypoxia-inducible factor-1 (HIF-1) (a core transcription factor that regulates oxygen homeostasis) under hypoxia-induced conditions.⁸² In a hypoxic microenvironment, HIF-1α and HIF-2α promotes the expression of NANOG by upregulating ALKBH5 in breast cancer cells, which leads to the promotion of tumor growth and metastasis.⁷⁶ Subsequently, TFEB bind to the conserved element E-box of the ALKBH5 promoter to activate its transcription constitutionally. In contrast, TFEB inhibits METTL3 expression by decreasing the mRNA stability of METTL3.⁸³ In epithelial ovarian cancer (EOC), ALKBH5 mediates the upregulation of HOXA10 by maintaining the stability of HOXA10 mRNA, and upregulated HOXA10 interacts with the TAAA region of the ALKBH5 promoter as a transcription factor. The positive feedback loop increases the proliferation and cisplatin resistance of EOC cells in vivo and in vitro.⁸⁴ Zfp217 acts as a transcription factor and binds to the promoter of the FTO gene to activate the transcription of FTO and thus modulate m⁶A mRNA modification.⁸⁵

In addition, previous studies have shown that histone H3 trimethylation at lysine 36 (H3K36me3) can be a determinant of m⁶A and found that most m⁶A peaks overlap with H3K36me3 modifications. METTL14 writes m⁶A modifications on new transcripts by directly recognizing and binding H3K36me3.⁸⁶ Another study showed that lysine-specific histone demethylase 5C (KDM5C) inhibits METTL14 transcription and promotes the metastasis of colorectal cancer (CRC) by demethylating H3K36me3 in the METTL14 promoter.⁸⁷ The acetylation of histone H3K27 in the METTL3 promoter is mediated byp300 and promotes the transcription of METTL3. Subsequently, increased METTL3 expression promotes tumor angiogenesis and glycolysis in GC by enhancing the stability of HDGF mRNA in an IGF2BP3-dependent manner.⁸⁸ KDM4C also reportedly regulates the expression of ALKBH5 by increasing the alteration in the chromatin state of ALKBH5. This process is achieved by reducing the H3K9me3 levels and promoting MYB and Pol II recruitment. ALKBH5 subsequently affects the mRNA stability of the receptor tyrosine kinase AXL inan m⁶A-dependent manner and maintains leukemia stem cell (LSC) function.⁸⁹ A recent study demonstrated that HBX mediates H3K4me3 modification of the ALKBH5 gene promoter in HBV infection and subsequently promotes the expression of ALKBH5 in a WDR5-dependent manner, and the resulting increased ALKBH5 promotes high expression of HBX in an m6A-dependent manner; this positive feedback pathway ultimately promotes HBV-induced hepatocellular carcinoma.⁹⁰ Cigarette smoke condensate (CSC) increases the expression of METTL3 by reducing the methylation of its promoter, and subsequently, METTL3 promotes the development and progression of pancreatic cancer by inducing the excessive maturation of miR-25-3p.⁹¹ Further studies have shown that CSC also induces ALKBH5 CpG island hypomethylation, which results in reductions in the LINC00278-SORF1 micropeptide levels, downregulation of LINC00278 mRNA, and thereby promotion of the progression of esophageal squamous cell carcinoma (ESCC)⁹² (Fig. 2).

Figure 2.

Transcription-level regulation of m⁶A molecules. The transcription level of m⁶A molecules is mainly regulated by two aspects: 1. Activation of transcription factors. The transcription factor TFEB can bind to the conserved E-box in the promoter to promote ALKBH5 expression. Hypoxia-inducible factor-1 (HIF-1) directly binds to the ALKBH5 promoter and promotes its transcription, HOXA10 promotes ALKBH5 transcription by binding to the ALKBH5 promoter, and Zfp217 binds to the FTO promoter to activate its transcription. 2. Histone modification. The histone demethylase KDM4C regulates ALKBH5 transcription by modifying histone H3K36me3 in the ALKBH5 promoter. HBX mediates the modification of K3K4me3 in the ALKBH5 promoter to promote ALKBH5 transcription, and KDM5C promotes its transcription through demethylated METTL14 histone H3K36me3 modification. P300 regulates METTL3 transcription by mediating the acetylation of METTL3 histone K27. Cigarette smoke condensate (CSC) causes hypomethylation of METTL3 and ALKBH5 CpG islands and regulates their transcription level.

The above discussion notes that the transcriptional activity and chromatin accessibility of m⁶A can affect its expression. In addition, m⁶A modification is inversely correlated with chromatin alterations. Specifically, m⁶A modification sites are located on chromosome-associated regulatory RNAs (carRNAs), and decreased m⁶A modification levels in carRNAs will increase their expression and promote an open chromatin state and downstream transcription.⁹³ METTL3 also regulates m⁶A modifications on the histone methyltransferase EZH2, which in turn regulates the H3K27me3 levels.⁹⁴ These findings link chromatin state dynamics to the regulation of m⁶A enzyme expression.

Posttranscriptional regulation of m⁶A molecules by ncRNAs

Previous reports have suggested that m⁶A regulates ncRNAs, including their splicing, transportation, degradation and expression.⁹⁵ Interestingly, the level of m⁶A is also affected by ncRNAs. To date, most studies on the regulatory effect of ncRNAs on m⁶A have focused on microRNAs (miRNAs), which can regulate m⁶A expression through sequence pairing, promote the binding of METTL3 to miRNA target genes, and thus change the global m⁶A modification level.⁹⁶ In NSCLC, miR-33a directly targets 3′ UTR of METTL3 mRNA and reduces its expression, which eventually leads to attenuation of the proliferation of NSCLC cells.⁹⁷ In breast cancer, metformin reduces the expression of METTL3 by targeting miR-483-3p, which reduces the m⁶A levels and ultimately inhibits the proliferation of breast cancer cells. In this process, miR-483-3p also functions by binding to the 3′ UTR of METTL3.⁹⁸ MiR-320d inhibits the expression of KIF3C by targeting the 3′ UTR of METTL3 and decreasing its expression, which inhibits the progression of prostate cancer.⁹⁹ Other studies found that HBXIP upregulates the expression of METTL3 in breast cancer cells by inhibiting miRNA let-7g, which downregulates METTL3 expression by binding to its 3′ UTR. In turn, METTL3 promotes HBXIP expression in an m⁶A-dependent manner and eventually forms the positive feedback loop HBXIP/let-7g/METTL3/HBXIP, which accelerates the proliferation of breast cancer cells.¹⁰⁰ Moreover, another research group found that HBXIP silencing in gastric cancer (GC) reduces METTL3 expression, inhibits GC cell proliferation, migration, and invasion, and promotes apoptosis.¹⁰¹ Additionally, in HCC cells, HBXIP promotes the expression of METTL3 at the mRNA and protein levels and then promotes the metabolic reprogramming of HCC cells.¹⁰² However, whether HBXIP also regulates METTL3 in a miRNA-dependent manner remains unclear. A study found that miR-4429 inhibits the progression of GC by targeting METTL3 and inhibiting its mRNA and protein expression.¹⁰³ METTL3 is significantly elevated in hepatoblastoma (HB), suggesting a poor prognosis. Further investigation revealed that METTL3 may serve as a downstream target of miR-186, which inhibits tumor invasion by directly targeting METTL3.¹⁰⁴ miR-600 also reportedly inhibits lung cancer progression by decreasing METTL3 expression.¹⁰⁵

In addition to METTL3, METTL14 is directly regulated by miR-103-3P, which also functions by targeting its 3′ UTR.¹⁰⁶ Similarly, miR-29a decreases the expression of WTAP and ERK by binding to their 3′ UTRs, inhibits the proliferation and metastasis of glioblastoma stem cells (GSCs), and promotes apoptosis.¹⁰⁷ However, whether WTAP plays a role in an m⁶A-dependent manner still needs to be further studied. The one report showed that miR-145 reduces the expression of YTHDF2 by targeting its 3′ UTR, and YTHDF2 increases the total m⁶A level in HCC and inhibits the occurrence, proliferation, invasion and metastasis of HCC.¹⁰⁸ Lysine-specific demethylase 5 (KDM5) is highly expressed in prostate cancer (PCa). This demethylase inhibits miR-495 transcription and expression by binding to its promoter. Subsequently, miR-495 inhibits MOB3B expression by targeting YTHDF2. This miR-495/YTHDF2/m6A-MOB3B axis augments PCa tumor growth.¹⁰⁹ In thyroid cancer (TC), IGF2BP2 acts as a direct target of miR-204. MALAT1 competitively binds to IGF2BP2 with miR-204 to promote MYC expression, which promotes TC proliferation, migration, and invasion.¹¹⁰ The miR-96 antagomir promotes cancer progression in CRC by regulating m⁶A modification in the MYC transcript, and this result is achieved by the regulation of FTO in an AMPKα2-dependent manner.¹¹¹

It is clear that the above mentioned ncRNAs, particularly miRNAs, appear to preferentially target the 3′ UTR of m⁶A enzymes to regulate their expression. However, whether this “preference” applies to other ncRNAs is unknown. In addition, as discussed above, miRNAs always appear to negatively regulate m⁶A modification, but the conclusion drawn is probably one-sided, and it is that additional positive regulators have not yet been detected. In addition, whether ncRNAs bind to m⁶A enzymes to regulate their expression by promoting their mRNA degradation, inhibiting their translation, or in other manners remains unclear. These questions are equally worth exploring.

In summary, the above-described results indicate that the proper biological function of m⁶A may require the appropriate regulation of transcription factors and histone modifications. Dysregulation of these processes may lead to changes in m⁶A and result in tumors or autoimmune diseases. Therefore, additional studies are needed to uncover the mechanisms of m⁶A-modified transcription, and it is believed that more scholars will reveal its regulatory mechanism in the future, which will improve our understanding and utilization of m⁶A modification (Fig. 3).

Figure 3.

m⁶A is regulated by ncRNAs. Mature miRNA binds to the 3′ untranslated region of m⁶A mRNA in the cytoplasm and inhibits its expression. Specifically, miR-33a can directly target the 3′ UTR of METTL3 mRNA and inhibit its expression. The miRNA let-7g downregulates its expression by targeting the 3′ UTR of METTL3, whereas mir-145 downregulates its expression by targeting the 3′ UTR of YTHDF2. In addition, the expression of METTL14, FTO, YTHDF2 and IGF2BP is also negatively regulated by miRNAs.

Posttranslational modifications of m⁶A molecules

In terms of posttranslational modifications (PTMs), SUMOylation modifications reportedly regulate m⁶A-modified molecules. SUMOylation is the process of attaching small ubiquitin-related modifier (SUMO) to protein substrates at specific lysine residues.^112,113 This reversible PTM can change the stability, localization, protein–protein interactions and activity of the target protein.114, 115, 116, 117

Du Y et al found that METTL3 could be modified by the SUMO1-specific protease SENP1 on lysine residues K177, K211, K212 and K215. The methyltransferase activity of METTL3 is reduced after SUMOylation. However, the mechanism through which SUMOylation affects the activity of METTL3 remains unclear,¹¹⁸ and SUMOylation of METTL3 does not change its stability, localization, or interaction with METTL14 and WTAP. This type of modification was verified in the NSCLC cell line H1299.¹¹⁸ Mitogen stimulation upregulates UBC9 in liver cancer cells, which modifies METTL3 through the small ubiquitin modifier SUMO1; then, SUMOylated METTL3 regulates EMT progression by controlling Snail mRNA homeostasis.¹¹⁹ ROS can promote the SUMOylation of ALKBH5 by activating ERK/JNK signaling. SUMOylated ALKBH5 cannot bind to its substrate, which significantly increases the m⁶A level on the entire mRNA and eventually leads to the rapid and effective induction of genes involved in various biological processes, including DNA damage repair.¹²⁰ An analysis of disease-free survival in patients with lung adenocarcinoma showed that high expression of YTHDF2 and SUMO1 predict poor prognosis. Subsequently, the researchers found that the SUMOylation site of YTHDF2 is on K571 in vivo and in vitro. The SUMOylation modification of YTHDF2 has little effect on its ubiquitination and localization but significantly increases its binding affinity with m6A-modified mRNA, leading to gene expression disorder and cancer progression.⁶² However, the mechanisms through which SUMOylation affects the activity of m⁶A molecules remain unclear. Methyl transferase domain (MTD, residues 369–590) and two Cys-Cys-Cys-His (CCCH)-type zinc finger motifs (ZNF, 259–340) are needed for the correct functions of METTL3 and METTL14.¹²¹ The researchers found that the SUMOylation modification sites are close to the ZNF motif of METTL3 but far from the MTD domain; therefore, they speculate that SUMOylation may spatially affect the binding of ZNF and MTD to substrate mRNA and ultimately inhibits m⁶A methyltransferase activity,¹¹⁸ but additional research is needed to confirm this hypothesis.

In addition to SUMOylation, the ubiquitination/autophagy-lysosome pathway of the m⁶A reader IGF2BP2 could be blocked by LINRIS(upregulated in CRC tissues from patients with poor overall survival (OS), which facilitates Myc-mediated glycolysis and promotes the progression of CR.¹²² Phosphorylation also reportedly regulates m⁶A modification. S43, S50 and S525 of METTL3 and S306 and S341 of WTAP could be phosphorylated by ERK, and USP5 subsequently stabilizes METTL3 by deubiquitination,¹²³ which results in maintenance of the stability of the m⁶A methyltransferase complex to ensure the level of m⁶A on intracellular mRNA.¹²³ EGFR/SRC/ERK signaling phosphorylates YTHDF2 at serine 39 and threonine 381, which stabilizes YTHDF2. YTHDF2 is needed for GBM cell proliferation, invasion, and tumorigenesis.¹²⁴

PTMs are essential for protein activity and stability. However, few posttranscriptional modifications have been found on m⁶A enzymes; thus, future research on PTMs and m⁶A should first pay attention to whether other PTM modifications exist on m⁶A enzymes. Another difficulty is the relationship between these posttranslational modifications because whether they function through synergy or antagonism is unclear. This important problem also needs to be resolved in the future (Table 1).

Table 1.

Summary of m⁶A posttranslational modifications.

Molecule	Modification	Modification site	Mechanism of modification	Functions of modification
METTL3	SUMOylation	Lysine residues K177, K211, K212 and K215	m6A methyl transferase activity is significantly inhibited	Regulation of EMT progression118,119
METTLE	Phosphorylation	S43/S50/S525	Control of m6A methyl transferase activity	Loss of METTTL3/WTAP phosphorylation leaves mouse embryonic stem cells in a pluripotent state123
WTAP	Phosphorylation	S306/S341	Maintenance of the stability of the m6A methyltransferase complex	Stabilization of the m6A mRNA level in mouse embryonic stem cells123
ALKBH5	SUMOylation	/	SUMOylated ALKBH5 cannot bind to its substrate	Rapid and efficient induction ofDNA damage repair inside genes120
YTHDF2	SUMOylation	K571	Increase the binding affinity to m6A-modified mRNA	Promotion of gene expression disorders and cancer progression62

Regulation of the function of m⁶A molecules by ncRNAs

As detailed in the previous discussion, ncRNAs can regulate the expression of m⁶A molecules by binding to their 3′ UTR; in addition, ncRNAs can regulate the levels of m⁶A by recruiting m⁶A molecules. The lncRNA GAS5-AS1 is the antisense RNA of GAS5. A recent study found that GAS5-AS1 combines with ALKBH5 and that ALKBH5 regulates the m⁶A modification of GAS5 mRNA to enhance its stability in a YTHDF2-dependent manner. The increased GAS5 ultimately plays an antitumor role by inhibiting the proliferation, invasion, migration and metastasis of cervical cancer cells.¹²⁵ A novel circRNA, HSA_circ_0008399 (circ 0008399), binds to WTAP and promotes formation of the WTAP/METTL3/METTL14 m⁶A methyltransferase complex. Subsequently, the mRNA stability of TNF alpha-induced protein 3 (TNFAIP3) is increased in an m⁶A-dependent manner, and its expression is increased to reduce the chemotherapy sensitivity of metastatic bladder cancer (BC) to cisplatin (CDDP).¹²⁶ The lncRNA ARHGAP5-AS1 is highly expressed and related to autophagy impairment in gastric cancer. In the nucleus, ARHGAP5-AS1 recruits METTL3 to increase the m⁶A levels of the target mRNA ARHGAP5 and thus increase its stability. Upregulated ARHGAP5 promotes chemoresistance and suggests poor prognosis in gastric cancer.¹²⁷ The lncRNA FOXM1 (FOXM1-AS) enhances FOXM1 expression by promoting the binding of ALKBH5 to the nascent transcripts of FOXM1 and reducing its m⁶A levels and then promotes glioblastoma stem-like cell (GSC) tumorigenesis.¹²⁸ Similarly, the lncRNA GATA3-As recruits KIAA1429, a “reader” of m⁶A, to GATA3 pre-mRNA, which results in GATA3 pre-mRNA degradation and RNA-binding protein HuR separation. Promotion of the growth and metastasis of HCC indicates an adverse prognosis in liver cancer.¹²⁹

A study conducted by Zhu et al revealed another interesting consequence: the lncRNA LINC00266-1 encodes a 71-amino-acid peptide named RNA-binding regulatory peptide (RBRP), which interacts with IGF2BP1 and promotes binding to target transcripts, such as c-Myc mRNA, by increasing the mRNA stability and expression of c-Myc to promote colorectal cancer (CRC) tumorigenesis.¹³⁰ The m⁶A reader IGF2BP1 is recruited by the lncRNA KB-1980E6.3 to retain the stability of c-Myc mRNA. These factors form the lncRNA KB-1980E6.3/IGF2BP1/c-Myc signaling axis that maintains the stemness of BCSCs.¹³¹

These results show another mechanism through which ncRNAs regulate m⁶A. Specifically, ncRNAs regulate the expression of m⁶A modification molecules, and the functions of m⁶A enzymes can also be regulated through recruitment or interaction. However, whether this regulation is accidental or normal in physiological processes is unclear, and whether the same ncRNAs can not only regulate the production of m⁶A molecules but also participate in the regulation of their function remains to be determined. Do ncRNAs regulate m⁶A in other ways? Additional research might be needed to answer these questions.

Regulation of the interaction of m⁶A molecules with target mRNAs

METTL3 is related to the maintenance of leukemia stages, is located at transcriptional start sites (TSSs) with clear transcription factor and histone modification characteristics, and regulates the expression of mRNAs encoded by its target genes. However, METTL3 mapping to the transcription start site must depend on CEBPZ, a CCAAT box-binding factor that recruits METTL3 to chromatin sites.¹³² Recent studies showed that the target gene 5′ untranslated nuclear cap-binding subunit 3 (NCBP3) recruits METTL3 in response to hypoxia stress/stress and increases the mRNA m⁶A levels in cardiomyocytes. The NCBP3/METTL3/eIF4A2 regulatory axis plays an important role in the hypoxia stress of cardiomyocytes.¹³³ In intestinal epithelial cells infected by E. coli K88, the transcription factor FOXO6 recruits METTL3 to form the METTL3/FOXO6 complex. FOXO6 is responsible for the transcriptional activation of GPR161, and METTL3 is responsible for the m⁶A modification of new GPR161 transcripts, promoting GPR161 transcription and subsequently regulating the expression of β-defensin.¹³⁴ DDX46 is a member of the RNA helicase dead-box family. DDX46 binds to the CCGGUU conserved motif of mRNA encoding antiviral molecules. When viral infection occurs, DDX46 recruits the m⁶A demethylase ALKBH5 and increases its binding. The DDX46-binding antiviral molecule mRNA is demethylated to induce nuclear retention, which blocks the protein expression of these antiviral molecules, reduces the production of interferon, and ultimately inhibits the natural antiviral immune response.¹³⁵ DDX5, another member of the DDX family, interacts with METTL3 after viral infection and regulates the methylation of antiviral mRNA molecules by affecting the formation of the METTL3-METTL14 heterodimer complex, which negatively regulates natural immunity and promotes viral infection.¹³⁶ ALKBH5 is directly regulated by DDX3, which leads to decreased m⁶A methylation in FOXM1 and then ascent NANOG transcript and thus contributes to tumor chemoresistance.¹³⁷

METTL3 interacts with viral RNA-dependent RNA polymerase 3D during EV71 C4 subtype infection, and this interaction induces SUMOylation and ubiquitination of the polymerase and ultimately promotes polymerase stability and viral replication.^138,139 During hepatitis B virus (HBV) infection, the viral HBV X (HBx) protein interacts with METTL3 and METTL14, which stimulates their nuclear import and achieves cotranscriptional m⁶A modification of viral RNAs.¹⁴⁰

These reports show another mechanism in which m⁶A enzymes work by binding to each other and suggest that m⁶A is not an independent modification pattern. It is more likely one part of a biological process. The mechanism through which m⁶A collaboratively functions with other molecules may be of interest (Fig. 4).

Figure 4.

Multilevel regulation of m⁶A. In the nucleus, transcription factors bind to the promoter of m⁶A molecules and activate their transcription. Histone modifications such as methylation or acetylation also affect the transcription level of m⁶A molecules. After transcription, ncRNAs, particularly miRNAs, regulate their expression by binding to mRNAs of m⁶A molecules in the cytoplasm. After translation, the polypeptide chain of m⁶A molecules is regulated by posttranslational modifications, such as phosphorylation, SUMOylation, and acetylation, and then folds into the correct functional structure. When it plays a regulatory role, the interaction between m⁶A molecules and target mRNA is regulated by ncRNAs and other molecules, which ultimately affects the m⁶A modification level on target mRNA, mediatesits splicing, nuclear output, stability, translation and other metabolic processes and affects tumor progression.

Future prospects

In this review, we summarize relevant reports on m⁶A regulation from the perspective of m⁶A molecule transcription and expression levels, posttranscriptional modification levels, and other mechanisms. At the transcriptional and expression levels, many ncRNAs and transcription factors reportedly regulate the expression of m⁶A molecules. Are other molecules involved in this regulation? How are the upstream signaling pathways associated with these molecules regulated? In addition, many studies have shown that a specific miRNA can target multiple genes, and a gene can also be targeted by multiple miRNAs. Is this complex phenomenon involved in the m⁶A modification process? How can we explain this phenomenon in the future? These questions may complicate the mechanisms involved in m⁶A modification, and this difficult problem will be faced in future research. Regarding modifications after translation, SUMOylation and phosphorylation reportedly regulate the activity and stability of m⁶A molecules. More than 300 different posttranscriptional modifications have been identified, and these include phosphorylation, acetylation, ubiquitination, carboxylation, ribosylation and disulfide bond formation. In addition to the several modification methods of m⁶A molecules found thus far, other posttranscriptional modification methods may be involved in the regulation of m⁶A molecules, and this possibility is worth further exploration. m⁶A molecules show regulatory functions by interacting with nucleic acids, such as noncoding RNAs, or ribonucleic acid-binding proteins, such as transcription factors, but the mechanism through which m⁶A binds to these substances and the regulation of this process remain unanswered questions that can be addressed in future research. In addition to specific regulatory mechanisms, internal and external factors in tumor development can participate in the regulation of m⁶A modifications and thus play a role in the regulation of tumor progression. It is worth noting that the specific molecular mechanisms of these regulatory processes have not been fully elucidated. Furthermore, whether m⁶A molecules are regulated by only a single factor and whether m⁶A molecules are feedback regulated by these influencing factors in tumorigenesis are worthy of future consideration and research efforts.

A growing body of evidence shows that m⁶A modification plays a dual role in cancer. On the one hand, m⁶A modification regulates the expression of oncogenes or tumor suppressor genes and thus affects tumor progression, as has been elucidated in many excellent reviews. On the other hand, the m⁶A levels and the expression and activity of m⁶A molecules can be regulated to affect the role of m⁶A in cancer, as summarized in this paper.

An increasing number of m⁶A inhibitors have been reported, and some inhibitors have been applied to tumor therapy and have achieved good results. However, many molecules are involved in the dynamic regulation of m⁶A modification, and many of these may change during tumor occurrence and development. Moreover, most “writers” and “erasers” need “readers” to play a role, which makes the targeting of future therapies a major problem. Which m⁶A molecules play a major role in tumor development? If multiple m⁶A molecules change, is it necessary to use multiple m⁶A inhibitors in combination? These problems make it difficult to target m⁶A molecules in tumor therapy. However, studying more upstream influencing factors of m⁶A, obtaining a better understanding of m⁶A and its role in tumors, and making improvements upstream may lead to the avoidance of side effects and treatment difficulties caused by the use of multiple inhibitors. In addition, the simultaneous regulation of m⁶A upstream factors and application of m⁶A inhibitors may achieve better results in the treatment of tumors.

Author contributions

LF drafted the manuscript. RD, BC, ML and JT discussed and revised the manuscript. SW designed the study and revised the manuscript. All the authors read and approved the final manuscript.

Conflict of interests

The authors declare that there are no conflicts of interests.

Funding

This work was supported by the Research Project of Jiangsu Commission of Health (China) (No. K2019019).