Novel insights on m6A RNA methylation in tumorigenesis: a double-edged sword

N6-methyladenosine (m6A), the most prevalent modification of mammalian RNA, has received increasing attention. Although m6A has been shown to be associated with biological activities, such as spermatogenesis modulation, cell spermatogenesis and pluripotency, Drosophila sex determination, and the control of T cell homeostasis and response to heat shock, little is known about its roles in cancer biology and cancer stem cells. Recent articles have noted that some genes have abnormal m6A expression after tumorigenesis, including genes ABS2, RARA, MYB, MYC, ADAM19 and FOX1. Abnormal changes in the m6A levels of these genes are closely related to tumour occurrence and development. In this review, we summarized the ‘dual edge weapon’ role of RNA methylation in the tumorigenesis. We discussed RNA methylation could lead to not only tumour progression but also tumour suppression. Moreover, we clarified that the abnormal changes in the m6A enrichment of specific loci contribute to tumour occurrence and development, thereby representing a novel anti-cancer strategy by restoration to balanced RNA methylation in tumour cells.


Background
Approximately 100 different post-transcriptional chemical modifications are present in RNA among all living organisms [1]. N6-methyladenosine (m 6 A), one such modification, was identified in the 1970s as the most abundant internal chemical modification in eukaryotic mRNA [2]. Approximately 0.1-0.4% of adenosine nucleotides in isolated mammalian RNA are chemically modified [3]. Extensive m 6 A modifications are present in the RNA of plants and vertebrates, and these modifications also occur in the RNA of single-celled organisms, such as bacteria and yeast [2,[4][5][6][7]. m 6 A-based modifications occur at a consensus motif, RRm 6
N6-methyladenosine represents one of the most common RNA modifications in eukaryotes, regulating RNA behaviours such as splicing or the ability to code diverse proteins [9,[20][21][22][23][24][25]. Recently, aberrant m 6 A modification in the large internal exon of a tumour suppressor was shown to give rise to premature polyadenylation, leading to membrane associated guanylate kinase (MAGI3) inactivation [31]. In addition, m 6 A regulates other forms of RNA modification. There is a significant negative correlation between two distinct and abundant RNA modifications, m 6 A and adenosine-to-inosine (A-to-I), suggesting a previously underappreciated interplay between them [32]. N6-methyladenosine also affects the function of Long non-coding RNA (LincRNA). LincRNA 1281 is required for proper differentiation of mouse embryonic stem cells, and this critical function relies on sufficient m 6 A modification [33]. m 6 A may also play a functional role by regulating histones and DNA. The homeostatic regulation of S-adenosylmethionine (SAM) synthesis in mammalian cells involves dynamic m 6 A modifications on the 3'UTR of methionine adenosyltransferase 2A (MAT2A) [34]. In this review, we will briefly introduce physiological activities related to m 6 A modification. Then, we will describe in detail the ability of m 6 A modifications, functioning as a double-edged sword, to play a regulatory role in tumorigenesis and development.

Physiologic functions of m 6 A modifications
The dynamic reversibility of m 6 A methylation suggests that it plays an important role in physiological processes. Studies have revealed that m 6 A modifications on Fig. 1 The process of m 6 A RNA modification. The installation, removal and identification of m 6 A are performed by writers, readers, and erasers, respectively. Writers interact with a special sequence of RRACH in mRNA that produces catalytic action mediated by METTL3, METTL14, KIAA1429, ZC3H13, METTL16 and WTAP. m 6 A functions are received by some reader proteins: YT521-B homology (YTH) domain-containing protein, IGF2BP, FMR1, LRPPRC and the heterogeneous nuclear ribonucleoprotein (HNRNP) protein families. Two m 6 A eraser proteins are obesity-associated protein (FTO) and alkB homologue 5 (ALKBH5) mRNAs or non-coding RNAs play important roles in spermatogenesis, T cell homeostasis, Drosophila sex determination, heat shock responses, pluripotency and reprogramming, as well as other processes [27,[35][36][37][38][39] (Table 1).

m 6 A modulates spermatogenesis
The process by which diploid spermatogonial stem cells (SSCs) produce haploid spermatozoa is called spermatogenesis [35]. m 6 A is reportedly present on the key regulatory factors of SSCs/progenitor cells, such as Plzf, Id4, Dnmt3b, and Sohlh2, which control the timing of transcript translation to coordinate normal protein generation, and this modification is essential for mammalian spermatogenesis [39]. m 6 A deletion resulted in the dysregulation of proliferation and differentiation factors of SSC/progenitor cells and SSC depletion [39].

m 6 A influences T cell homeostasis
Peripheral T cells are subject to complex and rigorous regulation, and the interleukin 7(IL-7)/signal transducer and activator of transcription 5(STAT5) signal axis is highly significant for maintaining naive T cell homeostasis and survival [36]. Decreased levels of m 6 A modification on the mRNAs of suppressor of cytokine signalling (SOCS) family genes have been reported to slow mRNA decay and increase protein expression levels (SOCS1, SOCS3 and CISH) in naive T cells [40]. Overexpression of SOCS1, SOCS3 and CISH leads to the inhibition of the downstream signal IL-7/STAT5, preventing naive T cell proliferation and differentiation while maintaining T cell survival [40]. Thus, m 6 A modifications are known to play an important role in T cell homeostasis.

High levels of m 6 A during the heat shock response
The heat shock response is a complex cellular reaction that causes significant changes in protein translation, folding and degradation, thereby mitigating toxic reactions caused by protein misfolding [44]. m 6 A and the heat shock response are linked because m 6 A is the most abundant mRNA post-transcriptional modification. A new report has revealed that m 6 A is preferentially deposited on the 5'UTR of new stress-inducible transcripts, such as Hsp105 (HSPH1), under heat shock stress, and that increased levels of m 6 A modification at the 5'UTR can enhance cap-independent translation initiation [28]. Thus, the mechanistic connection between 5'UTR methylation and cap-independent translation reveals links between the heat shock response and m 6 A [28].

m 6 A influences somatic cell reprogramming and maintains the pluripotency of embryonic stem cells (ESCs)
Epigenetic and epitranscriptomic networks play important roles in somatic cell reprogramming and the maintenance of ESC pluripotency [37]. A new study has revealed that zinc finger protein 217 (ZFP217) activates the transcription of key pluripotency genes and modulates m 6 A deposition on their transcripts [37]. ZFP217 depletion globally enhances m 6 A modification on Nanog, Sox2, Klf4, and c-Myc mRNAs to accelerate their degradation, thus damaging ESC self-renewal and somatic cell reprogramming [37]. This finding represents strong evidence of the close relationships between m 6 A and  [37] somatic cell reprogramming and the maintenance of ESC pluripotency.

Aberrant m 6 A modification contributes to diversified tumours
Given the important role of RNA m 6 A modification in regulating gene expression and various biological processes [2], it is reasonable to speculate that aberrant m 6 A modification plays an important role in human carcinogenesis. However, knowledge of the mechanistic link between m 6 A and human carcinogenesis is rather limited. While investigations addressing this issue are still at an early stage, efforts are underway to explore the biological impacts of m 6 A modifications in cancer. We will summarize recent reports describing our understanding of the biological functions and underlying molecular mechanisms of m 6 A regulatory proteins in various types of cancer and explore new options for cancer treatment ( Fig. 2 and Table 2).

Connection between m 6 A and acute myeloid leukaemia (AML)
AML is one of the most common haematopoietic malignancies and is associated with a high mortality and distinct genetic and molecular abnormalities that lead to unsatisfactory therapeutic effects [45,46]. Only a small proportion of patients survive for more than 5 years with standard chemotherapies [45,46]. Therefore, it is urgent and necessary to explore new treatments for AML based on a deep understanding of the mechanisms of AML occurrence and development. FTO, an obesity risk-associated gene and the first m 6 A eraser to be identified, has been reported to play an important oncogenic role in haematopoietic cell transformation and AML [17,47]. FTO levels in certain subtypes of AML (e.g., t(11q23)/MLL-rearranged, t(15;17), FLT3-ITD, and/or NPM1-mutated) are abnormally elevated, which leads to the downregulation of m 6 A levels on the UTRs of ASB2 and RARA [47]. These low levels of m 6 A reduce the mRNA and protein levels of these two genes [47]. FTO is not the only demethylase with a link with AML; METTL14 also plays an important oncogenic role in this disease by regulating its mRNA targets (e.g., MYB, MYC) through m 6 A modifications, which in turn leads to enhanced MYB and MYC expression and ultimately blocks myeloid differentiation [48]. Additionally, a new report has shown that FTO promotes the stability of MYC mRNA by inhibiting YTHDF2-mediated RNA decay, which is attributed to decreased m 6 A abundance on the 5′-terminal and internal exons of MYC mRNA, indicating that m 6 A modifications on different regions of the same mRNA transcript (e.g., MYC) lead to distinct fates [48,49].
Furthermore, it has been reported that the presence of increased METTL3 levels in AML leads to higher m 6 A methylation levels of BCL2 and PTEN and promotes the translation of BCL2 and PTEN mRNA, which ultimately leads to tumour formation [50]. Recently, another study revealed that METTL3 binds to the promoter region of SP1 with the aid of transcription factor CEBPZ, which enhances m 6 A modification of SP1, strengthens SP1 gene expression, and ultimately leads to myeloid leukaemia [51]. In general, changes in m 6 A modification levels on ASB2, RARA, MYC, MYB, BCL2, SP1and PTEN eventually lead to the occurrence of AML [47][48][49][50].

Aberrant m 6 A in hepatocellular carcinoma (HCC)
HCC is a major type of primary liver cancer, accounting for the 5th highest incidence of malignant tumours worldwide and causing more than 700,000 annual deaths [52]. The prevalence of hepatitis B and C viruses in Asia is the main cause of the high incidence of liver cancer in Asia [52]. Effective interventions are lacking, resulting in high HCC mortality due to metastasis and recurrence; thus, a deeper understanding of the molecular mechanism underlying the occurrence and development of HCC is required. Growing evidence suggests that the occurrence of liver cancer is a multistep process involving complex interactions between genetics, epigenetics and transcriptional changes [53]. Previous studies have shown that DNA hypermethylation occurs on CpG islands of the promoter regions of tumour suppressor genes, such as DLC1, TFPI-2, CDKN2A, and PTEN, ultimately affecting tumorigenesis and development [54][55][56]. It was reported that high levels of the histone methyltransferases EZH2, SUV39H1, SETDB1 and G9a promote the development and metastasis of HCC via epigenetic silencing of critical tumour suppressor genes [57][58][59]. In general, abnormal epigenetic modifications may be important factors in the development of liver cancer.
Many articles note that the development of liver cancer is associated with abnormal m 6 A modifications [60,61]. The high expression of METTL3 in human HCC reportedly leads to increased m 6 A modification levels on the tumour suppressor SOCS2 [60]. Excessive m 6 A chemical modification of SOCS2 is read by YTHDF2, which accelerates the degradation of SOCS2 and eventually leads to the occurrence of HCC [60]. Another report has indicated that decreased METTL14 expression reduces m 6 A modification levels and the expression of microRNA126 (miR126) [61]. Low m 6 A modification levels on miR126 are recognized by DGCR8 and may ultimately promote Cervical cancer β-catenin Anti-oncogene Upregulated FTO represses m 6 A modification of β-catenin and induces chemoradiotherapy resistance [79] liver carcinogenesis [61]. These articles strongly suggest that abnormal m 6 A modification plays an important role in the occurrence and development of HCC.

The foundation role of m 6 A in glioblastoma stem cells (GSCs)
Glioblastomas are primary brain tumours with a high degree of malignancy [62]. The median survival time after diagnosis is usually less than 15 months, even if diagnosis is combined with surgical resection, radiotherapy and chemotherapy [62,63]. GSCs are a group of tumour stem cells with the ability to promote tumour growth and invasion, showing strong resistance to radiotherapy and chemotherapy; these characteristics are the main reasons for the poor prognosis of glioblastoma [64][65][66][67]. Therefore, the issue of stem cell resistance should be explored and addressed to improve therapeutic approaches to glioblastoma [66,68]. RNA epigenetics has become a rapidly developing field in biology and may be valuable for informing glioblastoma treatment.
A study showed that low levels of METTL3 or METTL14, key components of the RNA methyltransferase complex, lead to decreased m 6 A modification levels on ADAM19 and the enhanced expression of ADAM19 in GSCs, ultimately causing glioblastoma [69]. Low m 6 A modification levels on ADAM19 and high mRNA expression levels of ADAM19 may represent a promising target for anti-glioblastoma therapy [69]. Another study revealed that aberrant m 6 A modifications caused by high levels of ALKBH5, an m 6 A demethylase, are an obvious physiological abnormality of GSCs [70]. Decreased m 6 A modification levels of FOXM1 result in enhanced FOXM1 expression levels, which ultimately cause glioblastoma [70]. Aberrant m 6 A modifications caused by different mechanisms in GSCs strongly suggest that this modification is related to the occurrence and development of glioblastoma.

Abnormal m 6 A modification in breast cancer
Among all malignant tumours in women, breast cancer has the highest incidence and leads to the highest number of deaths [71,72]. Although the therapeutic outcomes of early-stage breast cancer are relatively good, treatments for metastasis are not effective [73]. Given the high recurrence and mortality rate, the molecular mechanisms that regulate breast cancer phenotypes need to be carefully delineated and studied to design more effective therapies.
In breast cancer stem cells (BCSCs), ZNF217 has been reported to interact with METTL3 and inhibit the m 6 A methylation of KLF4 and NANOG, which ultimately leads to high expression of KLF4 and NANOG, thus promoting tumorigenesis [74]. Another report indicated that high m 6 A modification levels on hepatitis B X-interacting protein (HBXIP) and the overexpression of HBXIP caused by high METTL3 levels accelerate the proliferation of BCSCs [75]. In addition, a study revealed that high levels of m 6 A modification on MAGI3 lead to premature polyadenylation, switching its functional role from a tumour suppressor gene to a dominant-negative oncogene and ultimately promoting tumorigenesis of breast cancer [31]. The aberrant m 6 A modifications caused by different mechanisms in breast cancer prove that this modification is related to the occurrence and development of breast cancer.

Aberrant m 6 A modification in cervical cancer
Cervical cancer is one of the most common and destructive gynaecological malignancies [76]. Chemoradiotherapy is the major therapy used to treat cervical squamous cell carcinoma [77]. However, chemoradiotherapy resistance is the major cause of treatment failure. Therefore, it is necessary to further understand the molecular mechanisms underlying chemoradiotherapy resistance and explore novel therapeutic treatments for cervical squamous cell carcinoma (CSCC) [78].
In CSCC, the expression of FTO is significantly higher than that in normal tissues, resulting in lower levels of m 6 A modification in β-catenin, which causes decreased expression of β-catenin and chemoradiotherapy resistance. The discovery of this mechanism suggests that MA2, a novel small molecular inhibitor of FTO, may increase the chemoradiotherapy sensitivity of CSCC [79].

METTL3 promotes the translation of oncogenes in human lung cancer
Lung cancer is one of the most common malignant tumours in humans, causing many deaths every year [80,81]. The 5-year survival rate of lung cancer patients is still very low despite continuous improvement and progress in the diagnosis and treatment of lung cancer [82,83]. Non-small-cell lung carcinoma (NSCLC) accounts for 85% of all pathological types of lung cancer [82,83]. Our attention should be focused on the abnormal molecular biological characteristics of NSCLC to find an effective treatment.
Many articles have reported that abnormal m 6 A modifications ultimately affect tumour development. However, one report indicated that in lung cancer, METTL3 associates with translation machinery and enhances the translation of target mRNA (RGFR and TAZ) independent of its methyltransferase activity [84]. Another report also indicated that miR-33a prohibits NSCLC cell proliferation by targeting METTL3, which suggests that miR-33a may be a potential molecule for therapy [85]. Moreover, post-translational modification of METTL3 has been revealed. For example, METTL3 is modified by SUMO1, and SUMOylation of METTL3 decreases m 6 A levels on mRNAs, which ultimately promotes the development of NSCLC [86].

m 6 A leads to the acceleration of tumour formation
In AML, upregulated m 6 A modification on MYB, MYC, BCL2, PTEN and SP1 results in enhancement of the binding capability and translational efficiency of onco-RNA and ribosomes, leading to tumorigenesis [47][48][49][50]. Moreover, in hepatocellular carcinoma, the excessive m 6 A modification of the SOCS2 tumour suppressor gene reduces mRNA stability and accelerates its degradation, which causes tumour progression [60]. In addition, in breast cancer, upregulated m 6 A modification of HBXIP and MAGI3 results in tumour formation [31]. It is worth noting that excessive modification of MAGI3 leads to premature polyadenylation, switching its functional role from that of a tumour suppressor gene to a dominant-negative oncogene, ultimately promoting tumorigenesis [31]. In summary, RNA methylation triggers certain alterations to tumour-specific mRNA behaviour and results in changes in onco-protein expression and biologic activity, thereby accelerating the tumour progression.

m 6 A contributes to the inhibition of tumour development
In contrast, the aberrant decreased m 6 A modification levels on target loci can also disrupt normal RNA functions, which in turn restores normal m 6 A levels on these targets and ideally suppresses tumour formation. The FTO-mediated downregulation of m 6 A modification levels on ASB2 and RARA leads to the downregulation of these anti-oncogenes via RNA and protein degradation, leading to the promotion of tumorigenesis [47]. In hepatocellular carcinoma, the decreased m 6 A modification levels on microRNA126 influence its function as a ceRNA and disrupt its regulation of binding capability, thereby triggering the acceleration of tumour development [61].
Similarly, in GSCs, a long non-coding RNA FOX-M1-AS directly binds to FOXM1 mRNA, enhancing the interaction between ALKBH5 and FOXM1 nascent transcripts and giving rise to reduced m 6 A modification levels and the overexpression of this oncogene [70]. Additionally, decreased m 6 A modification levels enhance the RNA stability of KLF4 and NANOG and ultimately contribute to tumour formation [74]. In summary, it has been revealed that decreased RNA methylation may participate in tumorigenesis. Therefore, a novel therapeutic strategy may involve tumour suppression via enhanced m 6 A modification to balance the transcription of these genes.

Conclusion
In summary, an increasing number of studies has shown that aberrant m 6 A modification is closely related to tumorigenesis, including AML, HCC, GSCs, breast cancer, cervical cancer and lung cancer [50,60,70,75,79]. Moreover, numerous genes modified by m 6 A have been revealed to play regulatory roles in tumour formation, such as BCL2, PTEN, SOCS2, FOXM1 and HBXIP [50,60,70,75,79]. In conclusion, m 6 A modification is a double-edged sword, over-modification of a target gene by m 6 A could result in altered RNA splicing and translational capability, leading to the acceleration of cancer formation, whereas the lack of m 6 A modification at other loci may also contribute to tumorigenesis.
Abnormal levels of m 6 A methylation may give rise to tumour progression. However, we should not ignore the notion that RNA methylation enzymes influence tumorigenesis in an m 6 A-independent manner. For example, in lung cancer, METTL3 directly associates with translation machinery and enhances the translation of target mRNA (RGFR and TAZ) independent of its methyltransferase activity [84].
This review updates our knowledge of the aberrant m 6 A methylation of diverse target loci and discusses its impact on tumour formation. Aberrant levels of m 6 A modification, such as increased or decreased levels, may alter RNA splicing, RNA-coding capability or onco-or tumour suppressor genes. To discover novel tumour therapies based on the evaluation of m 6 A modifications, it should be noted that m 6 A functions as a dual-edged weapon; thus, restoring ideal levels of m 6 A (rather than simply over-supplementing or over-silencing) holds great significance.