Primary carcinoma of the liver has multifactorial pathogenesis and is a malignant tumour occurring in hepatocytes or intrahepatic bile duct epithelial cells, including hepatocellular carcinoma (HCC), hepatobiliary carcinoma and hepatic sarcoma (Gao et al., 2020). HCC occurs in liver cells and is the main histological subtype of the hepatic malignancy (accounting for more than 90% cases of primary carcinoma of the liver). Overwhelming evidence has proved the regulatory roles of ncRNAs related to liver carcinogenesis, and their relationship can be summed up as “frenemy” which is means a kind of love-hate relationship (Wong et al., 2018). The miR-186, which plays the role of anti-hepatoblastoma, is weakly expressed in HCC tissues, and its overexpression inhibits cell aggressive characteristics; however, the increase of its direct target METTL3 reverses the inhibitory effect (Cui et al., 2020). In addition, another study has shown that METTL3 can improve the expression of c-Myc by increasing m6A modification, whereas miR-338-5p inhibit the expression of METTL3 to interfere with the lung cancer progression (Wu et al., 2021). The m6A “writer” METTL14 positively modulate pri-miRNA-126 maturation by interacting with DGCR8 to suppress the metastatic potential of HCC cells. Similar to the interaction of miR-186 and METTL3, the overexpression of pri-miRNA-126 inversely inhibits the repressing effect of METTL14 (Ma et al., 2017).

Thus, hostility alters their normal biological functions, while the friendship represents a positive correlation between ncRNAs and m6A regulators. With the help of m6A modification mediated by METTL3/METTL14, lncRNA LINC00958 (Zuo et al., 2020)and circ-SORE (Xu et al., 2020a)are upregulated because of enhancement of transcript stability. The m6A “reader” IGF2BP1 plays an oncogenic role in HCC progression, but the decrease of lncRNA LIN28B-AS can deplete the expression of IGF2BP1-dependent mRNAs, such as IGF2 and Myc, thereby inhibiting HCC cell proliferation, migration, and invasion (Zhang et al., 2020c). In addition to influencing the expression of molecules, m6A modification and ncRNA also affect their binding to downstream molecules. KIAA1429 is a member of m6A methyltransferases, and GATA3 mRNA is a direct downstream target of KIAA1429-methylation, facilitating the degradation of GATA3. Strikingly, KIAA1429 preferentially induces GATA3 pre-mRNA in a targeted manner under the guidance of the antisense gene lncRNA GATA3-AS to maintain their roles in cancer progression (Lan et al., 2019). In addition, m6A modification targets the 3ʹ-UTR of YAP, which induces the repression effect of miR-582-3p, although whether the m6A regulator is primarily responsible for this essential part of the interaction is unclear until now (Zhang et al., 2018). The complexity of m6A modification and ncRNAs suggest that it is necessary to further study the mechanisms and effects of their crosstalk in liver cancer, thus using the stable and significant indicators to assist early diagnosis and treatment.

Gastric cancer (GC) is a malignant tumour of digestive system occurring from gastric mucosal epithelium and glandular epithelium (Karimi et al., 2014), accounting for the second largest malignant tumour in China. In GC tissues, it is found that MEETL3 promotes the maturation of pri-miR-17-92 (Sun et al., 2020b), and m6A-modified motif-assisted miR-660 to reduce oncogene E2F3 activity (He and Shu 2019). Besides, METTL3 can upregulate the expression of target mRNA SEC62 by facilitating the stabilizing effect of IGF2BP1 on SEC62, but this positive regulation could be inhibited by miR-4429 (He et al., 2019). In contrast, lncRNA ARHGAP5-AS1 is responsible for recruiting METTLE3 to stabilise ARHGAP5 mRNA in the cytoplasm (Zhu et al., 2019). Unlike enhancing downstream target mRNA stability, METTL3 has been reported to combine with the m6A “reader” YTHDF2 to promote the degradation of PTEN mRNA and increase tumour malignancy, and subtly, the oncogenic lncRNA LINC00470 serves as an accelerator in this process (Yan et al., 2020). Many studies have reported that aberrant expressions of lncRNA (Sun et al., 2016), miRNA (Shin and Chu 2014), and circRNA (Shan et al., 2019) play important roles in GC progression and could be regarded as diagnostic/prognostic markers and chemotherapeutic tools. Compared with the complex change of m6A modification in liver cancer, most ncRNAs and protein genes related to m6A are upregulated in GC tissues, suggesting that the m6A-related risk score might be informative for risk assessment and prognostic stratification (Guan et al., 2020).

Bladder cancer is a morbid malignancy of the urinary tract, and 95% of bladder tumors originate from epithelial tissue. Notably, the symptoms of bladder cancer mimic those of a urinary tract infection, increasing the diagnosis difficulty and delaying timely therapy (Li et al., 2019c). Accumulating evidence implies that the abnormal m6A modification of ncRNA is related to tumorigenicity and poor prognosis in bladder cancer. Oncogenic METTL3 facilitates DGCR8 to recognise pri-miRNA and accelerate the maturation of pri-miR221/222, which is targeted to PTEN (Han et al., 2019a). Differently from the direct regulation, the m6A “eraser” FTO regulates the expression of MALAT/miR-384/MAL2 axis by catalysing MALAT1 demethylation in an m6A-dependent manner, indicating the potential of FTO as a diagnostic and prognostic biomarker (Tao et al., 2021). Low expression of METTL14 and decreased global m6A abundance in bladder cancer tissue are also associated with the clinical severity and outcomes of bladder cancer. METTL14 knockout promotes the proliferation, self-renewal, metastasis, and initiation of tumour cells, while overexpression plays an adverse role (Gu et al., 2019). These m6A-related molecules may play important roles in the clinical diagnosis and treatment of bladder cancer in the future.

Breast cancer is an epithelial malignant tumour originating from the terminal ductal lobule unit of the breast. The breast epithelial cells have unlimited replicative potential and other malignant characteristics because of various carcinogenic factors (Nagini 2017). The hostile hypoxic microenvironment is a major reason for the rapid expansion of cancer cells. It has been reported that the lncRNA KB-1980E6.3 that is associated with hypoxia can recruit m6A “reader” IGF2BP1 to stabilize c-Myc mRNA and enhance the malignant characteristics (Zhu et al., 2021). In endocrine-resistant LCC9 breast cancer cells, HNRNPA2B1 expression is higher than that in parental and tamoxifen-sensitive cells, and this increase alters the transcriptome and expression of miRNAs, because HNRNPA2B1 is a “reader” of the m6A mark in pri-miRNAs and is responsible for promoting DROSHA processing of pre-miRNAs (Klinge et al., 2019). Similarly, m6A methyltransferase METTL14 has been demonstrated to be significantly increased in breast cancer tissues with the function of reshaping miRNA profile of cancer cell lines. It has been verified that hsa-miR-146a-5p is a differentially expressed miRNA, modulated by METTL14-induced m6A modification, and thus affects the migration and invasion of breast cancer cells (Yi et al., 2020). In addition, METTL14 is regulated by the oncogenic lncRNA LNC942 to facilitate cell growth and cancer progression. Mechanistically, LNC942 recruits the METTL14 protein and elevated post-transcriptional METTL14-mediated m6A modification levels, thereby stabilising the expression of downstream targets (Sun et al., 2020a). In human epidermal growth factor receptor 2 (HER2)-positive breast cancer, it is reported that m6A “eraser” FTO inhibits miR-181b-3p and upregulates the expression of oncogenic ARL5B (Xu et al., 2020b). Overall, more extensive and in-depth researches on the aberrant m6A modification of ncRNA for the carcinogenesis and development of breast cancer are required.

Carcinoma of large intestine is a malignant gastrointestinal tract tumour that originates from large intestine mucous membrane epithelium and gland with a high metastasis and reoccurrence rate, including colon cancer and carcinoma of the rectum (Muller et al., 2016). Most researchers focus on the relationship between ncRNA m6A modification and metastasis of colorectal cancer (CRC). A novel lncRNA RP11-138 J23.1 (RP11) is demonstrated to positively regulate epithelial-mesenchymal transition (EMT) in CRC cells. Mechanistically, the expression of RP11 is upregulated in an m6A modification-dependent manner; then, the complex composed by RP11 and HNRNPA2B1 prevents Zeb1 degradation, triggering the dissemination of cancer cells (Wu et al., 2019). LncRNA GAS5 promotes YAP ubiquitin-mediated degradation, but YTHDF3 alleviates this effect via lncRNA GAS5 decay, confirming that YTHDF3 is not only a novel target of YAP but also a significant player in YAP signalling by facilitating lncRNA GAS5 degradation (Ni et al., 2019). Overexpression of METTL14 is also reported to correlate with the YAP pathway and can suppress CRC cell growth and metastasis via the miR-375/YAP1 pathway (Chen et al., 2020b). Besides, the expression of oncogene lncRNA XIST negatively correlates with YTHDF2 and METTL14 in CRC tissues because of the m6A modification (Yang et al., 2020c). In contrast, YTHDC1 recognises circNSUN2 and export it to the cytoplasm, eventually participating in cancer progression by forming the cir-cNSun2/IGF2BP2/HMGA2 complex (Chen et al., 2019a). In addition, the oncogene lncRNA LINC00460 combines with IGF2BP2 and upregulates the expression of HMGA1 mRNA and increase its stability, participating in cancer development (Hou et al., 2021). The m6A “reader” YTH protein family and IGF2BP2 have been widely reported in colorectal cancer, and therefore it is reasonable to speculate that the m6A “reader” may play an important role in the progression of colorectal cancer, which deserves deeply study.

Abnormal m6A modification of ncRNAs is also associated with the progression of other cancers. In clear cell renal cell carcinoma, Gu et al. find a novel DNA methylation-deregulated and RNA m6A reader-cooperating lncRNA (DMDRMR), which interacts with IGF2BP3 to stabilize target genes and plays a carcinogenic role (Gu et al., 2021). In thyroid cancer, lncRNA MALAT1 competitively binds to miR-204, upregulates IGF2BP2 and enhances MYC expression in m6A-dependent manner, conferring a stimulatory effect on cancer progression (Ye et al., 2021). In osteosarcoma, ALKBH5 inhibits the degradation of lncRNA PVT1 and leads to the overexpression of lncRNA PVT1. Overexpressed PVT1 can suppress the binding of YTHDF2 with PVT1 to promote cancer cell proliferation (Chen et al., 2020a). This combination of the m6A “reader” and lncRNA, which affects cancer cell proliferation, also occurs in pancreatic cancer. IGF2BP2 and lncRNA DANCR have been reported to interact and promote the cancer stemness-like properties of pancreatic cancer. The adenosine at 664 of DANCR is modified by N6-methyladenosine, and IGF2BP2 serves as a reader for this structural change to stabilise DANCR RNA and promote cell proliferation (Hu et al., 2020). In pancreatic duct epithelial cells, oncogenic miR-25-3p is correlated with poor prognosis in patients with pancreatic cancer. One study has found that NF-κB-associated protein induced by m6A modification facilitates excessive pri-miR-25 maturation under the environment of cigarette smoke condensate; then it activates the oncogenic AKT-p70S6K signalling and provokes malignancy of pancreatic cancer (Zhang et al., 2019a). In head and neck squamous cell carcinoma, oncogenic LNCAROD affects the degree of tumour malignancy. METTL3-and METTL14-induced dysregulated m6A modification could account for the aberrant expression of LNCAROD and lead to the malignant behaviour of tumour cells (Ban et al., 2020). Moreover, Zheng et al. propose that m6A is highly enriched in oncogenic lncRNA FAM225A and enhances RNA stability, contributing to cell proliferation, migration, invasion, and metastasis (Zheng et al., 2019). Similarly, m6A modification improves the transcript stability of lncRNA RHPN1-AS1 and upregulates its expression in epithelial ovarian cancer tissues, facilitating RHPN1-AS1 to act as an oncogene (Wang et al., 2020a). In glioblastoma, it has been found that 13 central m6A methylation regulators are associated with the clinical and molecular phenotype, suggesting that m6A regulators are important participants in malignant progression (Du et al., 2020b). A more specific example is HNRNPC, which directly binds to pri-miR-21 and promotes its expression, playing a role in the metastatic potential of the glioblastoma cell lines. The experimental results show that the expression levels of HNRNPC are higher in highly aggressive cancer cells and are elevated along with the brain tumour grade (Park et al., 2012). In the cancers listed above, the abundance of m6A modification and the enzymes involved in it can affect ncRNA maturation and activation and change their biological functions, which in turn can impact the progression of various cancers (see Figure 2).

Many studies have reported that m6A modification of ncRNAs influences tumorigenesis and cell proliferation; for example, ectopic expression of miR-660 directly binds to oncogene E2F3 to affect cell proliferation owing to the m6A motif in the region of E2F3 3ʹ-UTR (He and Shu 2019). The stabilisation of oncogene SEC62 mRNA could be enhanced by the m6A modification of METTL3 and IGF2BP1; however, it was demonstrated that miR-4429 targets and suppresses METTL3 to repress SEC62 and realise an anti-proliferative effect (He et al., 2019). In addition, the overexpression of miR-145 inhibits cancer cell proliferation, while this suppression is impaired by the overexpression of YTHDF2, resulting in a crucial crosstalk between miR-145 and YTHDF2 by forming a double-negative feedback loop (Li et al., 2020a). Carcinogenic lncRNA THOR has been reported to be related to m6A modification for the first time. YTHDF1 and YTHDF2 can regulate the stability of THOR by reading the m6A motifs, thereby strengthening cell proliferation and helping THOR to realise its oncogenic function (Liu et al., 2020a). In terms of the m6A “reader,” METTL3-induced m6A methylation reduces lncRNA RHPN1-AS1 degradation to help it serve as a ceRNA and activate the PI3K/AKT pathway (Wang et al., 2020a). Similarly, LNC942, lncRNA XIST and lncRNA FAM225A can affect cell proliferation and cancer progression under the m6A “writer”-mediated m6A modification (Sun et al., 2020a; Yang et al., 2020c; Zheng et al., 2019).

With the participation of m6A modification, both lncRNA THOR and lncRNA RHPN1-AS1, regarded as oncogenes, can promote tumorigenesis and cell proliferation, while it is worth noting that LINC00266-1 itself cannot promote tumorigenesis but RBRP encoded by it is a potential oncogene in colorectal cancer. LncRNA LINC00266-1 encodes a 71-amino acid peptide named “RNA-binding regulatory peptide” (RBRP), which binds to the RNA-binding proteins, including the m6A “reader” IGF2BP1. The existence and expression of natural endogenous RBRP has been verified in colorectal, breast, and ovarian cancers and nasopharyngeal carcinoma. RBRP, considered as a regulatory subunit of a m6A “reader,” strengthens m6A recognition by IGF2BP1 on targeted RNAs through the essential G19 residue and promotes tumorigenesis (Zhu et al., 2020a).

To investigate the function of miRNA m6A modification in cell apoptosis, regions upstream and downstream of miRNA 675 m6A modification sites in the H19 locus are mutated with a system of adenine base editors, showing that mutation facilitates cell apoptosis by downregulating the expression of H19 (Hao et al., 2020). Gu et al. (2018) also believed that m6A-modified miRNAs regulated the pathway closely related to cell apoptosis, as verified in arsenite-transformed cells. In prostate cancer, YTHDF2 is regarded as a target of miR-495 to recognise MOB3B m6A modification and inhibit its expression. At the same time, the miR-495 promoter interacts with overexpressed oncogenic KDM5A and restrains the expression of miR-495. It has been reported that the progression of prostate cancer can be motivated by the activation of the KDM5A/miRNA-495/YTHDF2/m6A-MOB3B axis by influencing the apoptosis of cancer cells (Du et al., 2020a). YTHDF2 is also considered as a direct target gene of miR-145 in epithelial ovarian cancer, and its overexpression enhances cell apoptosis (Li et al., 2020a). Besides, cigarette smoke decreases the m6A modification of LINC00278 and the translation of YY1BM encoded by LINC00278, and these dysregulations inhibit the apoptosis of cancer cells (Wu et al., 2020). In addition, overexpressed lncRNA KCNMB2-AS1 plays the role of ceRNA and upregulates the oncogene IGF2BP3 by sponging miR-130b-5p and miR-4294. IGF2BP3 binds to KCNMB2-AS1 m6A sites, forming a positive regulatory loop composed of KCNMB2-AS1 and IGF2BP3 to inhibit cell apoptosis and induce proliferation (Zhang et al., 2020e).

Some researchers have proposed a correlation between m6A and invasion, metastasis, such as the overexpression of FTO, which promotes the migration and invasion of GC cell lines (Xu et al., 2017). ALKBH5 is involved in mediating methylation reversal, and it is reported that ALKBH5-demethylated lncRNA NEAT1 is overexpressed in GC and colon cancer cells, promoting invasion and metastasis (Guo et al., 2020; Zhang et al., 2019b). Yang et al. also determined that the downregulation of ALKBH5 in colon cancer is associated with tumour inhibition. Overexpression of ALKBH5 restrains the invasion and metastasis of colon cancer cells (Yang et al., 2020a). In addition, ALKBH5 demethylates lncRNA KCNK15-AS1 and downregulates its expression in pancreatic cancer tissues, inhibiting cell motility and invasion (He et al., 2018). These interesting roles of ALKBH5 in different cancers warrant further study. The m6A modification level is significantly reduced in HCC tissue, especially in metastatic tissues. It was confirmed that the main factor is the downregulation of METTL14, which mediates miRNA maturation and alters the metastatic capacity of HCC (Ma et al., 2017). LncRNA miR503HG is also reduced in HCC tissues. Further investigation suggested that HNRNPA2B1 degradation is promoted by miR503HG, suppressing the NF-κB pathway and facilitating the invasion and metastasis of HCC cells (Wang et al., 2018). In other tumours, METTL3-methylated regulates the MALAT1-miR-1914-3p-YAP axis and increases YAP activity to enhance drug resistance and metastasis (Jin et al., 2019). METTL3-mediated methylation can also enhance the stability of lncRNA FAM225A, which regulates the expression of ITGB3 by binding to miR-590-3p and miR-1275 as ceRNA, and thus activates the FAK/PI3K/AKT signalling pathway and promotes the invasion and migration of cancer cells (Zheng et al., 2019). Moreover, the carcinogenic lncRNA RHPN1-AS1 and lncRNA THOR have also been found to be associated with cell viability and mobility by ectopic m6A modification, thereby affecting cancer cell invasion and metastasis (Liu et al., 2020a; Wang et al., 2020a).

During epithelial-mesenchymal transition (EMT) progression, the loss of E-cadherin causes cells to acquire mobility and the capacity (Castosa et al., 2018; Ribatti et al., 2020). It has previously been reported that the decrease of METTL3 inhibited the lung cancer cell morphological conversion induced by TGF-β, and augmented the expression changes of EMT-related marker genes, indicating the importance of m6A modification in EMT (Wanna-Udom et al., 2020). As for ncRNA m6A modification, it has been proposed that m6A-methylated lncRNA HOTAIR could upregulate EMT genes and increase the aggressiveness of tumour cells (Peng et al., 2016). In addition, lncRNA RP11 was reported to positively regulate the migration, invasion, and EMT of colorectal cancer cells. It is believed that m6A methylation increases nuclear accumulation and participates in the upregulation of RP11, triggering the spread of colorectal cancer cells (Wu et al., 2019). By analysing the cancer cells and clinical samples, Kandimalla et al. established the RNAMethyPro-a gene expression signature, which is correlated with EMT-related prognosis genes (Kandimalla et al., 2019). The specific mechanisms between ncRNA m6A modification and EMT deserve further study.

Angiogenesis is regarded as a significant step in tumour development, and some of the genes associated with EMT and tumour angiogenesis are m6A target genes involved in tumorigenesis; they are very sensitive to m6A modification. Previously, RNA functional analysis confirm that METTL14 and ALKBH5 can regulate the expression of each other, inhibit the demethylation activity of YTHDF3, and induce aberrant m6A modification to play roles in tumour angiogenesis and metastasis (Panneerdoss et al., 2018). In lung cancer, the anti-cancer gene miR-320b downregulates the expression of IGF2BP2 and thymidine kinase 1 (TK1), thus suppressing angiogenesis and lung cancer growth (Ma et al., 2021). During the process of lung cancer brain metastasis, upregulated miR-143-3p promoted by METTL3-methylation facilitates miRNA splicing and biogenesis, interacts with VASH1 and increases invasion capability and angiogenesis (Wang et al., 2019a). In brain metastasis of breast cancer, transcripts associated with brain metastasis are enriched by YTHDF3 and promote cancer cells to interact with cells in the tumour microenvironment, benefiting angiogenesis and metastasis (Chang et al., 2020).

Cancer stem cells maintain the vitality of the cellular population through self-renewal and infinite proliferation, which are essential for cancer initiation and metastasis. Experimental analysis has shown that ALKBH5 facilitates the proliferation and tumorigenicity of cancer stem cells by interacting with lncRNA FOXM1-AS (Zhang et al., 2017). Moreover, the identified m6A “reader” YTHDF2 can reduce the half-life of diverse m6A transcripts, including the tumour necrosis factor receptor (Tnfrsf2), leading to negative effects on the overall integrity of leukemic stem cell function. Under YTHDF2 depletion, Tnfrsf2 is upregulated to facilitate cell apoptosis in leukemic stem cells (Paris et al., 2019). In addition to leukemic stem cells, YTHDF2 is associated with the liver cancer stem cell phenotype (Zhang et al., 2020b). In breast cancer, a recent study showed that lncRNA KB-1980E6.3 can augment cancer cell self-renewal and maintain the stemness of cancer cells under a hypoxic microenvironment, indicating that targeting the lncRNA KB-1980E6.3/IGF2BP1/c-Myc axis may be a promising therapy for refractory hypoxic tumours (Zhu et al., 2021).

Drug resistance that develops during conventional drug therapy is one of the significant reasons for chemotherapy failure in cancer. As the main component of methylase involved in m6A modification, METTL3 has been widely reported to be associated with drug resistance. In NSCLC, the METTL3/YTHDF3 complex enhances the stability of lncRNA MALAT1, regulating the MALAT1-miR-1914-3p-YAP axis to increase YAP expression and induce drug resistance (Jin et al., 2019); METTL3, as a direct target gene of miR-4443 in tumour exosomes, is involved in regulating FSP1 expression and interferes with the ferroptosis induced by cisplatin, resulting in drug resistance (Song et al., 2021). Furthermore, in GC cells, METTL3 is recruited by lncRNA ARHGAP5-AS1 to stabilize ARHGAP5 mRNA and promotes chemoresistance (Zhu et al., 2019), and consistent with this study, Sun et al. find that elevated METTL3 and levels of m6A modification facilitate the maturation of pri-miR-17-92 to the miR-17-92 cluster, which activates the AKT/mTOR pathway. High levels of METTL3 and the miR-17-92 cluster then enhance the sensitivity of cancer cells to the mTOR inhibitor, everolimus (Sun et al., 2020b).

In tamoxifen-resistant LCC9 breast cancer cells, studies have determined that overexpressed m6A “reader” HNRNPA2B1 can upregulate and downregulate different miRNAs simultaneously and affect the downstream signalling pathways of these miRNAs. Transient overexpression of HNRNPA2B1 reduces the sensitivity of cancer cells to 4-hydroxytamoxifen and fulvestrant, suggesting the potential role of HNRNPA2B1 in endocrine-resistance (Klinge et al., 2019). Oncogene circRNA-SORE is upregulated in sorafenib-resistant HCC cells, acting as ceRNA to sponge miR-103a-2-5p and miR-660-3p, competitively activate the Wnt/β-catenin pathway, and induce sorafenib resistance (Xu et al., 2020a). However, further study shows that interfering the expression of circRNA-SORE has effects on improving sorafenib efficacy, which represents a promising pharmaceutical intervention targeted circRNA-SORE in sorafenib-treated HCC patients. Besides, recent studies have shown that depletion or inhibition of FTO suppresses the expression of immune checkpoint genes and attenuates the self-renewal ability of cancer stem cells, thus overcoming their immune escape (Su et al., 2020). Another study indicated that the m6A modification induced by ‘eraser’ FTO augments melanoma growth, whereas FTO inhibition increases the sensitivity of cancer cells to anti-PD-1 immunotherapy, suggesting that the immune checkpoint molecular inhibitors represented by PD-1/PD-L1 can play an important role with the help of FTO inhibitors (Yang et al., 2019). Although these two studies were based on the demethylation effect of FTO on mRNAs, it is reasonable to believe that FTO inhibitors can also inhibit ncRNA-related targets or signal pathways with the deepening of studies on the crosstalk of FTO and ncRNAs. These findings provide in-depth insights into chemoresistance and support the therapeutic potential of ncRNA m6A modification, especially in combination with existing drugs to target refractory malignant tumours.

The interaction between ncRNAs and m6A modification suggests that they may provide new ideas for the clinical treatment of cancer and even become prognostic biomarkers or therapeutic targets for cancer therapy. Zhang et al. constructed an m6A-score model to quantify m6A modification patterns in cancers. By analysing tumour microenvironment phenotypes and 5-year survival rates in patients with low and high m6A-score subgroups, it was found that patients with low m6A-scores show significant therapeutic advantages, suggesting a crosstalk between m6A modification and tumour microenvironment diversity and complexity (Zhang et al., 2020a). In HCC tissues, m6A methyltransferases WATP and KIAA1429 are related to prognosis (Chen et al., 2019b; Liu et al., 2020b), and METTL3, YTHDF2, ZC3H13 and ALKBH5 are considered independent prognostic factors for overall survival; of these, only METTL3 is an independent prognostic factor for recurrence-free survival (Li et al., 2019b; Liu et al., 2021). In CRC, depletion of METTL14 is associated with poor prognosis (Chen et al., 2020b; Yang et al., 2020c), which is different from the methylated enzymes METTTL3 and KIAA1429 in liver cancer. Notably, m6A “reader” ALKBH5 is associated with TNM stage, tumour size, lymph node metastasis (Tang et al., 2020), and can be regarded as a prognostic indicator in colon cancer (Guo et al., 2020). ALKBH5 is also believed to be a novel biomarker and independent prognostic factor in pancreatic cancer and NSCLC (Tang et al., 2020; Zhu et al., 2020b). Further, studies have revealed that FTO expression is closely correlated with low differentiation, peritumoral lymphovascular invasion, lymph node metastasis and is positively correlated with TNM stage in HCC, GC, HER2-positive breast cancer and bladder cancer, which makes FTO a possible biomarker for diagnostic and prognostic purposes (Cui et al., 2020; Tao et al., 2021; Xu et al., 2017; Xu et al., 2020b). Notably, high expression of HNRNPC in glioblastoma enhances cell invasiveness by regulating the miR-21/PDCD4 axis, and the level of HNRNPC is reported to increase with the increasing grade of brain tumour (Park et al., 2012). Bioinformatics analysis and validation experiments further confirmed the value of high expression HNRNPC in the prognosis of GBM patients at the expression level (Wang et al., 2020b). Interestingly, Ying et al. have identified m6A “reader” HNRNPC genes in the Chinese population and discussed the relationship between single nucleotide polymorphisms and susceptibility to pancreatic cancer, providing a completely novel idea on studying the role of m6A modification in tumorigenesis (Ying et al., 2021). Besides, several studies have analyzed the m6A regulators expression differences and their correlations in normal tissue and tumour tissue based on the clinical data of TCGA and GTEx databases, and it is found that the differential expression of some m6A regulators are closely associated with the risk factors of prognosis in hematologic system tumours and endocrine-system-related tumours (Li et al., 2020b; Zhang et al., 2021). Based on TCGA, GTEx and GEO, a novel risk signature has been constructed to evaluate the relationship between m6A regulators and survival rate of patients in uterine corpus endometrial carcinoma and uterine carcinosarcoma, further inspiring us to combine the bioinformatics with the study of m6A modification (Zou et al., 2021). Taken together, focusing on the expression of m6A regulators and their prognostic value provides us with a potential experimental avenue. It will have good clinical significance to directly or indirectly measure the expression level of m6A, the degree of tumours malignancy and the prognosis of patients by using tumour samples of patients.

ncRNAs modified by abnormal m6A modification could also be regarded as biomarkers, but these ncRNAs show a high degree of specificity and difference in various cancers. A m6A-lncRNA co-expression network has been constructed in primary glioblastoma to identify four m6A-related prognostic lncRNAs: MIR9-3HG, LINC00900, MIR155HG, and LINC00515 (Wang et al., 2021). Using a similar approach, LINC00152, LINC00265, and nine other lncRNAs are documented as biomarkers for predicting the overall survival of lower-grade glioma patients (Tu et al., 2020). In HCC, high expressed circRNA_104075 acts as a ceRNA to sponge miR-582-3p and upregulate YAP expression. The significant changes in downstream pathways caused by circRNA_104075 render it possible to be a new diagnostic marker (Zhang et al., 2018). LncRNA00266-1-encoded peptides RBRP bind to m6A “reader” IGF2BP1 to facilitate tumorigenesis, and the patients with high level of RBRP display an unfavourable prognosis (Zhu et al., 2020a); high expression of LINC00958 also can independently predict poor overall survival in patients with HCC. The good news is to take advantage of the role of LINC00958 as a biomarker, Zuo et al. constructed a novel drug delivery system, a PLGA-based nanoplatform filled with si-LINC00958, with the advantages of controlled release, tumour targeting, and better safety and effectiveness (Zuo et al., 2020). Differently, as a prognostic indicator, lncRNA miR503HG is significantly decreased in HCC tissues, and is positively associated with overall survival and time to recurrence (Wang et al., 2018).

The same phenomenon occurs in GC and cervical cancer. High expression of lncRNA ARHGAP5-AS1 (Zhu et al., 2019)and lncRNA LINC00470 (Yan et al., 2020) is related to the poor prognosis of patients with GC, whereas low expression of miR-660 is strongly linked with a large tumour size, severe lymph node metastasis, advanced TNM stage, and poor outcome (He and Shu 2019). In cervical cancer, lncRNA HOTAIR and lncRNA GAS5 show opposite effects on forecasting the clinical states of patients, high expression of HOTAIR indicates a high degree of malignancy, whereas lncRNA GAS5 is downregulated in cancer tissues, indicating that low GAS5 expression suggests poor prognosis (Peng et al., 2016). The above examples show that under m6A modification, multiple ncRNAs with significant changes may appear in the same cancer, which makes them have great potential for future prognostic diagnosis. However, accuracy cannot be determined solely using existing studies, and is worth further experimental exploration.

Surprisingly, lncRNA LINRIS (Wang et al., 2019b), lncRNA RP11 (Wu et al., 2019), lncRNA LINC00460 (Hou et al., 2021), lncRNA NEAT1 (Guo et al., 2020), and circRNA NSUN2 (Chen et al., 2019a) are all upregulated in patients with CRC and poor overall survival, indicating their promising role as critical prognostic markers and therapeutic targets. In prostate cancer, lncRNA CCAT1 and lncRNA CCAT2 can be regarded as a group variable to independently predict the prognosis of patients (Barros-Silva et al., 2020). Further, evidence confirms that potential prognostic markers also exist in other cancers, such as LINC01234 in NSCLC (Chen et al., 2020d), miR-744-5p in ovarian cancer (Kleemann et al., 2018), lncRNA LNCAROD in head and neck squamous cell carcinoma (Ban et al., 2020), lncRNA PVT1 in osteosarcoma (Chen et al., 2020a), miR-25-3p (Zhang et al., 2019a)and lncRNA DANCR (Hu et al., 2020)in pancreatic cancer (see Figure 2). Notably, several recent studies have used data models to analyse differentially expressed ncRNAs in some cancers to identify potential prognostic or diagnostic markers, such as in ovarian cancer (Li et al., 2021), adrenocortical carcinoma (Jin et al., 2021), kidney renal clear cell carcinoma (Yu et al., 2021), and colorectal cancer (Zuo et al., 2021). This new approach can explore the relationship between ncRNAs and different cancers on a large scale, as well as provide ideas and directions for further molecular mechanism research. To effectuate the role of ncRNAs and m6A regulators as biomarkers, the remarkable relationship between the expression level of m6A and prognosis needs to be explored more specifically. Demonstrating or testing the changes in m6A levels based on existing data in various cancers to evaluate clinical significance maybe more convincing.

Increasing attention has been focused on m6A regulators in hope of finding new treatments. For instance, a study shows that the new sodium channel blocker MV1035 can disturb migration and invasion characteristics by targeting ALKBH5 in glioblastoma (Malacrida et al., 2020). The potential of ALKBH as a drug target has been demonstrated by screening ALKBH blockers from newly synthesised anthraquinone derivatives (Huang et al., 2015). As an important demethylase, FTO has been identified to play an oncogenic role in NSCLC, ovarian cancer, and acute myeloid leukemia, associating with cancer cell proliferation, cancer stem cell maintenance and other malignant characteristics (Huang et al., 2020a; Huang et al., 2019; Li et al., 2019a). In recent years, researchers have used FTO as a target to develop FTO inhibitors, hoping to make new breakthroughs in cancer treatment. It is well known that meclofenamic acid, R-2-hydroxyglutarate, and MO-I-500 inhibit the activity of FTO and display anti-cancer activity (Huang et al., 2015; Singh et al., 2016; Su et al., 2018). Huang et al. find that artificially developed FTO inhibitors FB23 and FB23-2 can directly bind to FTO, selectively inhibit the m6A demethylase activity of FTO, and significantly inhibit cell proliferation in human acute myeloid leukaemia (Huang et al., 2019). Su et al. screen out two small-molecule FTO inhibitors, CS1 and CS2, and showed strong anti-cancer effects in multiple types of cancers. According to a previous study, the self-renewal of leukaemia stem cells can be attenuated and the immune response can be reprogrammed by CS1/2 (Su et al., 2020). A recent trial of FTO inhibitors shows that the ethylester form of meclofenamic acid (MA2) can negatively regulate Myc-miR-155/23a cluster-MXI1 feedback circuit and augment the efficacy of the chemotherapy drug temozolomide (TMZ) on inhibiting proliferation of malignant glioma cells (Xiao et al., 2020). In addition, the protease inhibitor nafamostat mesilate (NM) has been shown to inhibit the growth and metastasis of colorectal cancer and has anti-cancer effects in pancreatic cancer and lung cancer (Han et al., 2019b). Han et al. find that the activity of FTO can be inhibited by NM, indicating that NM might play a role in interfering with cancer progression as an FTO inhibitor. In addition, some biotech or pharmaceutic companies get down to developing high-efficiency small-molecule inhibitors, targeting the m6A regulators, in particular molecules like METTL3 and FTO that vary significantly in many cancers (Huang et al., 2020b). Undoubtedly, it provides a new choice for the effective management of a variety of cancers whether used alone or in combination with anti-tumour drugs, especially under the circumstance of chemotherapy drug resistance.