Biological and pharmacological roles of m⁶A modifications in cancer drug resistance

Abstract

Cancer drug resistance represents the main obstacle in cancer treatment. Drug-resistant cancers exhibit complex molecular mechanisms to hit back therapy under pharmacological pressure. As a reversible epigenetic modification, N⁶-methyladenosine (m⁶A) RNA modification was regarded to be the most common epigenetic RNA modification. RNA methyltransferases (writers), demethylases (erasers), and m⁶A-binding proteins (readers) are frequently disordered in several tumors, thus regulating the expression of oncoproteins, enhancing tumorigenesis, cancer proliferation, development, and metastasis. The review elucidated the underlying role of m⁶A in therapy resistance. Alteration of the m⁶A modification affected drug efficacy by restructuring multidrug efflux transporters, drug-metabolizing enzymes, and anticancer drug targets. Furthermore, the variation resulted in resistance by regulating DNA damage repair, downstream adaptive response (apoptosis, autophagy, and oncogenic bypass signaling), cell stemness, tumor immune microenvironment, and exosomal non-coding RNA. It is highlighted that several small molecules targeting m⁶A regulators have shown significant potential for overcoming drug resistance in different cancer categories. Further inhibitors and activators of RNA m⁶A-modified proteins are expected to provide novel anticancer drugs, delivering the therapeutic potential for addressing the challenge of resistance in clinical resistance.

Introduction

Estimated 600,000 people die from cancer each year, which is still a challenging problem that scientists are desperate to resolve [1, 2]. Oncotherapy is currently divided into five mainstream approaches: surgical resection, chemotherapy, radiotherapy, biological immunotherapy, and targeted therapy [3, 4]. Although there have been numerous breakthroughs for specific cancer categories, most strategies still are not as effective as expected. The major reason for treating cancer failure is the lacked understanding of the molecular mechanisms of therapeutic resistance. Resistance to chemotherapy drugs is usually divided into two main categories: acquired and intrinsic [5]. Intrinsic resistance, also called primary resistance, is a consequence of genetic alterations before treatment. Acquired drug resistance is caused by drug treatment and is also known as secondary resistance. Both are due to mutations and/or epigenetic changes in the genome of cancer cells. In the process of drugs binding to target and function, multiple mechanisms must be involved, including altered metabolism, transport, and varied target proteins [6]. Additionally, impaired apoptosis, augmented populations of cancer stem cells (CSCs), altered expression of oncogene/tumor suppressors, and manipulated tumor immune microenvironment (TIME) are also the dominant causes in charge of diminishing antitumor drug efficacy [7, 8]. Nevertheless, these are only influencing factors of therapy-resistant cancers, and the specific mechanism for therapy-resistant are unknown.

Researchers have identified more than 160 different chemically RNA modifications, creating a novel frontier called epitranscriptomics [9]. N⁶-methyladenosine (m⁶A) RNA modification has been identified as one of the most pervasive and abundant RNA modifications in eukaryotic messenger RNA (mRNA) [10, 11] and viral nuclear RNA [12, 13] since discovered in the 1970s. The process of m⁶A modification is dynamic and reversible, which is regulated by methylases (“writers”) and demethylases (“erasers”) (Table 1). m⁶A is installed by writers including methyltransferase-like (METTL) 3 [14], METTL14 [15], Wilms tumor 1-associated protein (WTAP) [17], KIAA1429 [18], METTL16 [16], RBM15 [20], and ZC3H13 [21]. m⁶A is removed by erasers such as fat mass and obesity-associated protein (FTO) [22] and alkB homolog 5 (ALKBH5) [23]. Different families of m⁶A reader proteins are capable of recognizing RNAs modified with m⁶A. One type of natural m⁶A reader protein contains the YT521-B homology (YTH) domain [33], and heterogeneous nuclear ribonucleoproteins (HNRNPs) belong to the other type, which mainly regulated alternative splicing or processing of target transcripts [29]. Other subfamily members are insulin-like growth factor 2 (IGF2) mRNA binding proteins (IGF2BP1/2/3) [31], and eIF3 [32].

Table 1 The role of m⁶A modification in the cancer biological functions

Emerging evidence indicated that m⁶A modifications were strongly associated with therapy resistance. In several neoplasms, m⁶A regulators (writers, erasers, and readers) are frequently overexpressed, regulating oncoprotein expression, enhancing cancer inception, and cell multiplication [34]. m⁶A modulates multiple anticancer resistance, including drug transport and metabolism, target receptors, cancer stemness, DNA damage repair, and cell death [35,36,37,38]. In addition, m⁶A is closely related to the immune response in the tumor microenvironment, providing new prospects for tumor immunotherapy [39]. Importantly, small-molecule activators and inhibitors of m⁶A regulators have recently been revealed to possess considerable anticancer effects when applied alone or in combination with other anticancer agents, suggesting the novel function of m⁶A in anticancer drug resistance [40]. This review primarily introduced the significant role of m⁶A modification in tumor drug resistance, reviewed the mechanisms of RNA m⁶A modification associated with drug resistance, and further discussed the strategies targeting the m⁶A change in predicting and treating cancer resistance (Fig. 1).

Fig. 1

m⁶A-mediated biological processes of drug resistance. m⁶A was involved in several aspects of drug pharmacokinetics. m⁶A modifications upregulated drug transporters (e.g., ABCB1, ABCC1, ABCC10), facilitating ATP-driven drug efflux. m⁶A was also engaged in regulating several drug-metabolizing enzymes (e.g., CYP2C8 and UGT2B7) that affected the efficacy of chemotherapeutic drugs. Some drug targets (e.g., EGFR) were regulated by m⁶A and affected cancer development. Additionally, m⁶A also participated in activating downstream effects, which were embodied in the following three aspects. Firstly, m⁶A could selectively upregulate the p53 (R273 H) protein, releasing prohibited anti-apoptotic proteins (e.g., BCL-2, IAPs). Secondly, m⁶A altered the expression of various key signaling molecules (e.g., ULK1, FOXO3) in autophagy and ultimately regulated autophagy through light chain 3-II (LC3-II). Thirdly, m⁶A modification activated oncogenic bypass signaling through key molecules (e.g., IGF1R, DUXAP8) and promoted cell stemness, which became an important barrier to drug resistance. Immune cell infiltration and cytokine secretion in the tumor microenvironment were also regulated by m⁶A, which was relevant for cancer immunotherapy. The m⁶A modification of exosomal non-coding RNA was implicated in multiple biological processes in tumors and was associated with resistance to multiple anticancer drugs

Mechanisms of m⁶A-mediated drug resistance

Cancer resistance is caused by a variety of factors, such as individual differences in drug sensitivity, tumor location, tissue spectrum, tumor aggressiveness, and alterations in intracellular molecules [3, 41]. The mechanism of m⁶A-mediated drug resistance was embodied in drug pharmacokinetics, tumor cells, and tumor microenvironment. Deciphering the impact of m⁶A modifications on the mechanisms of resistance to anticancer therapy could offer more prospects for individualized tumor treatment.

m⁶A modulation in drug pharmacokinetics

m⁶A modulated aberrant drug transport and metabolism

Several membrane transporter proteins work together to promote drug efflux and resistance to chemotherapeutics. Most drug efflux experiments have focused on the role of the ATP-binding cassette (ABC) proteins [42]. Multidrug resistance (MDR) is mediated by a wide range of ABC transporters, such as ABCB1 (MDR1), ABCC1 (MRP1), ABCC10 (MRP7), and others [43, 44]. Recently, researchers have demonstrated that RNA m⁶A modifications regulated the expression of ABC family proteins through either direct impact on tumor transcripts or indirect effects on upstream signaling pathways. For instance, m⁶A upregulated estrogen-related receptor gamma (ERRγ) in chemo-resistant cancer cells. ERRγ not only directly enhanced ABCB1 transcription but also indirectly by further strengthening the interaction with p65 [45]. Besides, METTL3 m⁶A-dependently enhanced translation of ABCD1, leading to migration and spheroid formation in clear cell renal cell carcinoma (ccRCC) [46]. Notably, exosomal-FTO facilitated ABCC10 of recipient cells via FTO/YTHDF2/ABCC10 axis, eventually leading to gefitinib resistance in non–small cell lung cancer (NSCLC) [47]. Excluding drug transport, the efficacy of chemotherapeutic drugs is determined by the effects of drug metabolism, such as bioactivation, catabolism, conjugation, and elimination [48]. Recent studies have revealed that the m⁶A modification had a negative regulatory effect on regulating drug metabolism. For example, METTL3/14 depletion upregulated cytochrome P450 family member cytochrome P450 2C8 (CYP2C8), whereas FTO depletion suppressed it. Mechanically, YTHDC2 promoted CYP2C8 mRNA degradation by recognizing the m⁶A in CYP2C8 mRNA [49]. Another drug metabolism enzyme, carboxylesterase 2 (CES2), exhibits the exact mechanism of negative regulation by m6A as CYP2C8 [50]. UDP-glucuronosyltransferases (UGTs) are enzymes that catalyze the glucuronidation of various endogenous and exogenous compounds. In Huh-7 cells, the m⁶A regulator-mediated methylation modification also showed a negative correlation with UGT2B7 [51]. In summary, m⁶A modifications are novel regulators of drug transport and metabolism, contributing to the practice of personalized medicine.

m⁶A drove drug target alterations

Alterations to drug targets, such as mutations or changes in expression levels, impact drug response and resistance [52]. For example, the TP53 gene coding for the p53 protein and mutant p53 proteins augmented cancer progression and generated drug resistance. METTL3-mediated m⁶A produced the p53 R273H mutant protein, causing MDR in colon cancer cells (Fig. 1) [53]. Epidermal growth factor receptor (EGFR) is another potential therapeutic target whose activation led to tumor cell proliferation, evasion of apoptosis, angiogenesis, and metastasis [54]. METTL3 augmented the translation efficiency of EGFR, followed by rebound activation of RAF/MEK/ERK, resulting in acquired PLX4032 resistance in melanoma (Fig. 1) [55]. Furthermore, YTHDF1 and YTHDF2 impacted cancer via binding m⁶A sites in the 3′-UTR of EGFR transcription and contributed to aberrant activities of downstream signal pathways [56, 57]. m⁶A-induced alterations in p53 protein and EGFR drug targets affect the efficacy of anticancer drugs, which may enable us to develop effective strategies to reverse the alterations in drug targets.

m⁶A modulation in tumor cells

m⁶A regulated DNA damage repair

An ocean of chemotherapeutic agents primarily targeting genomic DNA can result in DNA lesions and inhibit transcription and replication [58]. m⁶A methyltransferase METTL3 facilitated oxaliplatin resistance in gastric cancer (GC) stem cells by substantial DNA damage repair [59]. Furthermore, METTL3 enhanced the expression of UBE2B, a crucial enzyme involved in DNA damage repair, thereby triggering multifarious drug resistance [60,61,62]. Additionally, other m⁶A regulators, YTHDF1 and ALKBH5, were also engaged in chemoresistance (including adriamycin, cisplatin, and olaparib) by enhancing DNA damage repair in breast cancer (BC) [63, 64].

m⁶A activated downstream effects

Anticancer drugs result in tumor cells’ death upon binding to their cellular targets. The m⁶A modification affected a diverse array of downstream impacts, including demolition of apoptosis, activation of autophagy, and energizing of oncogenic bypass signaling, which was a crucial part of current cancer therapy [65, 66].

m⁶A mediated cell apoptosis

Cell sensitivity to anticancer drugs was primarily determined by the upregulation of anti-apoptotic proteins, including B-cell lymphoma 2 (BCL-2), IAPs, and FLIP [67, 68]. Remarkably, m⁶A modification had a differential effect on BCL-2 expression according to the type of cancer. Recent research revealed that overexpression of FTO was accompanied by BCL-2 upregulation [69], which was consistent with the trend of regulation of BCL-2 by ALKBH5 found in epithelial ovarian cancer (EOC) [70]. Consequently, RNA m⁶A modification was inversely correlated with BCL-2 expression and anti-apoptosis. Nonetheless, varied results were found that m⁶A also positively influenced the expression of anti-apoptotic proteins. Wang et al. found METTL3 knockdown dramatically augmented apoptosis capabilities in BC by decreasing BCL-2 expression [71]. In esophageal cancer, NSCLC, and GC, reduced expression of m6A positively correlated with the decrease of the anti-apoptotic protein BCL-2, contributing to the activation of apoptosis [72,73,74]. Overall, the m⁶A modification modulated apoptosis based on the cancer context, uncovering the dual role of m⁶A in tumor cells.

m⁶A mediated cell autophagy

Autophagy is a lysogenic process that permits cells to own stress-coping strategies by degrading damaged organelles and accumulated proteins, which could result in cancer resistance treated with anticancer drugs [75,76,77,78]. m⁶A modification acted as a double-edged sword in autophagy regulation. In some cases, the RNA m⁶A modification inhibited autophagy (Fig. 2A). Light chain 3B (LC3B) was a well-known autophagy biomarker in the cytoplasmic matrix [79]. In hepatocellular carcinoma (HCC), METTL3 depletion promoted the LC3-II accumulation by reducing the stability of FOXO3 mRNA through a YTHDF1-dependent mechanism [80]. Jin et al. [81] validated that FTO enhanced LC3B II accumulation by slowing the decay rate of unc-51-like kinase 1 (ULK1) transcripts in a YTHDF2-dependent manner. By the same mechanism, FTO enhanced the translation of autophagy-associated gene-5 (ATG5) and ATG7 mRNAs and promoted an increase of LC3-II [82]. Conversely, m⁶A modification promoted autophagy in some cases (Fig. 2B). ALKBH5 activated the EGFR-PIK3CA-AKT-mTOR pathway and specifically cemented the BCL-2 mRNA stability to slow the autophagy in EOC [70]. The latest study found that m⁶A reader YTHDF3 promotes autophagy by recognizing the METTL3-mediated m⁶A modification site around the FOXO3 mRNA stop codon, providing new evidence for a dual role in m⁶A autophagy [83].

Fig. 2

Dual effects of m⁶A in autophagy. On the one hand, the m⁶A modification inhibits autophagy. In hepatocellular carcinoma (HCC), METTL3 enhanced forkhead box O3 (FOXO3) mRNA stability and inhibited light chain 3-II (LC3-II) accumulation through a YTHDF1-dependent mechanism. The overexpression of FTO induced YTHDF2-dependent inhibition of unc-51-like kinase 1 (ULK1) mRNA decay and promoted LC3-II accumulation and autophagy. With the help of YTHDF2, FTO also increased the translation of autophagy-associated gene-5 (ATG5) and ATG7 mRNAs and promoted autophagosome assembly. On the other hand, m⁶A modification also promotes autophagy. In epithelial ovarian cancer (EOC), ALKBH5 slowed autophagy by cementing B-cell lymphoma 2 (BCL-2) mRNA stability and activating the EGFR-PIK3CA-AKT-mTOR pathway. Additionally, the m⁶A reader YTHDF3 promoted autophagy through the upregulation of FOXO3 mRNA translation

m⁶A regulated oncogenic bypass signaling

Even though targeted therapies enabled tumor cells to be sensitive to chemotherapy, drug resistance remained a significant obstacle owing to the activation of oncogenic bypass pathways (including Wnt/β-catenin, PI3K/AKT, MAPK, or c-MET signaling) [84,85,86]. ALKBH5 suppressed m⁶A modification of the WIF-1 mRNA to promote its transcription, which probably interfered with the Wnt signaling and led to chemosensitivity [87]. Besides, Xu et al. [88] revealed that the elevated level of m⁶A in circular RNA (circRNA)-SORE enhanced its stability, allowing it to induce sorafenib resistance by acting as a microRNA (miRNA) sponge to isolate miR-103a-2-5p and miR-660-3p, thereby competitively activating the Wnt/β-catenin pathway. YTHDC2, the m⁶A reader protein, regulated irradiation efficacy via IGF1R-AKT/S6 pathway, leading to radiotherapy resistance of nasopharyngeal carcinoma (Fig. 1) [89]. Alternatively, m⁶A modification-mediated DUXAP8 regulated malignant phenotype and chemoresistance of HCC through miR-584-5p/MAPK1/ERK pathway (Fig. 1) [90]. Beyond that, chidamide reduced c-MET expression by lowering m⁶A methylation, which increased crizotinib sensitivity in NSCLC cells in a c-MET/HGF-dependent manner [91]. NF-κB activating protein (NKAP), as a reader of m⁶A, promoted SLC7A11 mRNA splicing and maturation, thereby enhancing cell resistance to ferroptosis inducers [92]. Overall, the m⁶A mutation activated the oncogenic bypass pathway, circumventing the classical drug targets, which could be considered in targeted therapy to avoid or overcome drug resistance (Fig. 3).

Fig. 3

m⁶A-regulated oncogenic bypass signaling. Downregulation of ALKBH5 led to the downregulation of WIF-1 mRNA expression, thus activating the Wnt pathway. The elevated levels of m⁶A in circRNA-SORE enhanced its stability and allowed it to competitively activate the Wnt/β-certain pathway by acting as a miRNA sponge. YTHDC2 promoted radiotherapy resistance by activating the IGF1R-AKT/S6 signaling axis. m⁶A modification-mediated DUXAP8 contributed to chemoresistance via miR-584-5p/MAPK1/ERK. Chidamide decreased c-MET expression and increased crizotinib sensitivity by reducing m⁶A methylation. NKAP promoted SLC7A11 mRNA splicing and maturation, thereby inhibiting ferroptosis

m⁶A affected the sustainment of cell stemness

CSCs represent a small population of tumor cells sustaining versatility and promoting tumor progression and drug resistance [93, 94]. METTL3 was involved in regulating the stemness and chemosensitivity of colon cancer through the upregulation of LGR5 [95]. Aside from that, METTL3 facilitated oxaliplatin resistance in CD133+ stem cells by promoting PARP1 mRNA stability and increased base resection repair pathway activity [59]. Liu and his team [96] identified a crucial regulatory METTL14-miR99a-5p-TRIB2 feedback circuit that promoted cancer stemness and radioresistance in esophageal squamous cell carcinoma (ESCC). m⁶A modification of circHPS5 expedited cytoplasmic output and facilitated (epithelial-to-mesenchymal transition) EMT and CSC phenotypes, further accelerating HCC cell tumorigenesis [97]. HNRNPA2B1 promoted CD44⁺/CD24⁻/^low CSC and altered the EMT markers to initiate acquired endocrine resistance by activating ser/thr kinase growth factor signaling pathways [98]. The researches about m⁶A and stemness are still quite insufficient; thus, linking m⁶A modifications to CSCs in tumor drug resistance may be a new direction for future studies.

m⁶A modulation in the tumor microenvironment

m⁶A altered the TIME

An increasing number of studies demonstrated that the alteration of m⁶A regulated the TIME features [99], making the m⁶A regulator a promising immunotherapy target. Abnormal expression of METTL3 in various cancers played a dual part in the infiltration of immune cells. On the one hand, METTL3 was significantly downregulated in testicular germ cell tumor tissues, which positively correlated with the tumor-infiltrating levels of CD8+ T cells, CD4+ T cells, and NK cells [100]. On the other hand, the depletion of METTL3 or METTL14 tumors increased the infiltration of cytotoxic CD8+ T cells and elevated secretion of interferon-gamma (IFN-γ), CXCL9, and CXCL10 in the TIME, thus enhancing the reaction to anti-programmed cell death protein 1 (PD-1) treatment in pMMR-MSI-L colorectal cancer (CRC) [101]. WTAP was overexpressed in GC and negatively associated with T cell infiltration and T cell-induced immunity, indicating an unfavorable prognosis [102]. The depletion of FTO reprogrammed the immune response and enhanced T-cell toxicity by suppressing the expression of immune checkpoint genes, especially LILRB4 [103]. In melanoma, combining FTO inhibition with blocking the PD-1/PD-L1 checkpoint may relieve the resistance to immunotherapy [104]. In addition to regulating immune checkpoint blockade, FTO functioned as an essential epitranscriptomic regulator by regulating glycolytic metabolism and suppressing the function of CD8+ T cells [105]. ALKBH5, another m⁶A eraser, correlated positively with Treg cell infiltration. Melanoma patients treated with anti-PD-1 therapy benefited from ALKBH5 deletion [106]. Furthermore, the latest research found that a large number of immune checkpoint receptors (including PD-1, TIM-3, and CTLA-4) as well as lymphocytes infiltrating (such as B cells, T cells, macrophages, and dendritic cells) positively correlated with the level of m⁶A readers YTHDF1, and YTHDF2 in respective cancer type, including glioma, NSCLC, kidney renal clean cell carcinoma and BC [107,108,109,110]. Despite the different TIME among tumor types and individual responses, correcting m⁶A regulator disorder was a feasible strategy for cancer immunotherapy (Fig. 4).

Fig. 4

m⁶A-mediated alterations in the tumor immune microenvironment. METTL3 was significantly downregulated in testicular germ cell tumor tissues, which positively correlated with the level of tumor infiltration by CD8+ T cells, CD4+ T cells, and natural killer (NK) cells. WTAP was overexpressed in granulosa cells (GCs) and negatively correlated with T cell infiltration and T cell-induced immunity. In skin cutaneous melanoma (SKCM) patients, the number of infiltrating regulatory T cells (Tregs) and polymorphonuclear myeloid-derived suppressor cells (PMN-MDSCs) was significantly decreased in ALKBH5 knockout (KO) tumors, while dendritic cells (DCs) were significantly elevated. In kidney renal clear cell carcinoma (KIRC), downregulation of YTHDF2 positively correlated with lymphocyte infiltration (e.g., B cells, T cells, macrophages, neutrophils, and dendritic cells). In breast cancer (BC), high expression of YTHDF1 distinctly exhibited higher infiltration scores of activated memory CD4 + T cells and M1 macrophages but low infiltration levels of activated NK cells. METTL3 was highly expressed in mismatch-repair-proficient or microsatellite instability-low colorectal cancer (CRC) patients, and decreased interferon-γ (IFN-γ) Chemokine (C-X-C motif) ligand 9 (CXCL9) and CXCL10 secretion in TIME

m⁶A modified exosomal non-coding RNA

Exosomes are nano-sized extracellular vesicles that contain constituents of origin cells, which are essential for tumor-stroma cellular communication for mediating pigmentation-induced tumor resistance [111, 112]. However, the role of exosomal non-coding RNAs on tumor drug resistance has not been investigated until recently. Liu and colleagues [113] identified METTL3 positively modulated pri-miR-320b maturation process, which was associated with peritumoral lymphangiogenic activity and lymph node metastasis. Besides, METTL3 promoted the exosomal miR-181b-5p in cancer-associated fibroblasts (CAFs) and suppressed CRC cell sensitivity to 5-fluorouracil (5-FU) via the METTL3/miR-181d-5p axis [114]. In NSCLC, the miR-4443 level was significantly upregulated in cisplatin-resistant tumor-released exosomes. Mechanistically, overexpression of miR-4443 inhibited FSP1-mediated ferroptosis induced by cisplatin treatment in vitro and promoted tumor growth via METLL3-mediated m⁶A manner in vivo [115]. Exosome-transmitted circVMP1 was also involved in cisplatin resistance by targeting the miR-524-5p-METTL3/SOX2 axis [116]. Another research showed that exosomal long-noncoding RNAs (lncRNAs) might be a controller in regulating drug resistance. They discovered adipocyte exosomes contained the LncRNA package released by multiple myeloma (MM) cells through METTL7A-mediated methylation resulting in therapeutic resistance [117].

m⁶A induced specific drug resistance

Emerging researches show that m⁶A RNA methylation is involved in drug resistance of multiple cancer chemotherapeutic agents by regulating the expression of different targets or pathways. Elevated levels of m⁶A due to METTL7B overexpression in lung adenocarcinoma (LUAD) induced gefitinib and osimertinib resistance in a ROS-scavenging-dependent manner [118]. YTHDF2-mediated endoribonucleolytic cleavage of m⁶A-modified circASK1 also contributed to LUAD gefitinib resistance [119]. ALKBH5-mediated m⁶A demethylation stabilizes CASC8 transcription, ultimately leading to cisplatin resistance in ESCC [120]. Furthermore, YTHDF2 increased CDKN1B mRNA degradation in an m⁶A-dependent manner, which promoted intrahepatic cholangiocarcinoma (ICC) progression and reduced sensitivity to cisplatin treatment [121]. m⁶A modifications also play an integral part in tamoxifen resistance, a classical chemotherapeutic agent in breast cancer treatment [122]. METTL3 promoted the translation of AK4 mRNA by increasing m⁶A levels and facilitated ROS production and activation of p38, ultimately resulting in tamoxifen resistance [123]. Tamoxifen resistance was also caused by the m⁶A reader HNRNPA2B1 regulating downstream targets through activation of the ser/thr kinase growth factor signaling pathway [98]. In treating glioblastoma multiforme (GBM) with temozolomide, METTL3 increased the m6A modification of histone modify-related gene transcripts leading to the development of chemoresistance [124]. In ccRCC, YTHDC1 acted as an m⁶A reader and regulated the sensitivity of tyrosine kinase inhibitors (TKI) such as sunitinib through the YTHDC1/ANXA1 axis [125]. In conclusion, research on the molecular mechanisms of m⁶A in different chemotherapeutic agents has attracted increasing attention, offering new prospects and potential therapeutic targets for reversing therapeutic resistance (Fig. 5).

Fig. 5

m⁶A-induced specific drug resistance. Specific chemotherapy drug resistance associated with m⁶A and related regulators in esophageal squamous cell carcinoma (ESCC), lung adenocarcinoma (LUAD), intrahepatic cholangiocarcinoma (ICC), glioblastoma multiforme (GBM), breast cancer (BC), clear cell renal cell carcinoma (ccRCC)

Targeting the m⁶A modification to surmount anticancer resistance

As discussed above, m⁶A modifications had a dual function in driving drug resistance, yet obscure behind the molecular mechanisms. In addition to mutations in m⁶A, each tumor’s m⁶A regulators had a different function [126], drawing researchers’ attention to the regulating agency of m⁶A regulators in targeted therapy (Table 2).

Table 2 The role and regulatory mechanism of m⁶A regulator in cancer drug resistance

Targeting methyltransferase

METTL3

As an m⁶A writer, METTL3 regulated cancer initiation and progression, including glioblastoma, BC, HCC, leukemia, and other cancer cells [142,143,144,145]. Silencing METTL3 could reverse cancer cells’ resistance to radiotherapy/chemotherapy even though its biological effects were likely organ/lineage-specific. A recent study proposed that the elevated expression of METTL3 enhanced SOX2 mRNA stability. Specifically, silencing METTL3 enhanced the sensitivity of (glioblastoma stem cells) GSCs to γ-H2AX and efficient DNA repair, resulting in rescuing glioblastomas’ radiosensitivity [146]. Furthermore, silencing METTL3 promoted temozolomide’s sensitivity, inhibited proliferation, and facilitated apoptosis. Taketo’s study [62] showed that cancer cells were more sensitive to chemotherapy and radiotherapy when METTL3 was suppressed. Their study affirmed that METTL3 was linked to the alternative expression of MAPK cascades, especially in patients treated with gemcitabine, 5-FU, and cisplatin. Meanwhile, Uddin and colleagues [53] demonstrated that METTL3 catalyzed a preferential pre-mRNA splicing in the point-mutated codon 273 (G > A) of TP53. Whereafter, the enlarged translation of mutant p53 protein-induced MDR as a result. m⁶A was recruited to the translation initiation complex in a METTL3-mediated manner and directly promoted yes-associated protein (YAP) translation. Additionally, the stability of MALAT1 was increased by METTL3/YTHDF3 complex, which also promoted YAP expression via the MALAT1-miR-1914-3p-YAP axis. The amplified YAP expression induced DDP resistance and metastasis [128]. Meanwhile, m⁶A also developed resistance to other chemotherapeutic drugs in NSCLC. Chidamide downregulated c-MET expression by decreasing its mRNA m⁶A methylation, thereby increasing the sensitivity of NSCLC cells to crizotinib in a c-MET−/HGF-dependent manner [91]. By eliminating METTL3-mediated FOXO3 mRNA stabilization in the hypoxic tumor microenvironment, METTL3 depletion significantly enhanced the drug resistance of HCC to sorafenib, which confirmed FOXO3 as a crucial m⁶A modification downstream molecule in the sorafenib resistance of HCC [80]. The latest study revealed the potential function of METTL3 in adriamycin resistance (ADR) in BC. METTL3-mediated m⁶A regulated MALAT1 expression, thereby recruiting E2F1 and promoting AGR2 expression, which resulted in ADR in BC [127]. A recent study in GC showed that the reader IGF3BP1 recognized METTL3-mediated m6A modification on apoptotic protease-activating factor 1-binding lncRNA to maintain its stability, which inhibited GC cell apoptosis and led to multidrug resistance [147]. Notably, m⁶A-targeted transcription factors differed across cancer phenotypes, and further studies on the regulatory mechanism of action are necessary to develop more treatments targeting METTL3.

WTAP

WTAP is another essential m⁶A methyltransferase complex interacting with METTL3 and METTL14 to pre-RNAs/hnRNAs for catalytic activity. The targeting WTAP knockdown significantly reduced m⁶A modification and increased apoptosis [17]. Bansal et al. [148] hypothesized that excessive expression of the WTAP was associated with an oncogenic role in leukemogenesis. Its abnormal elevated expression correlated with a poor prognosis of acute myeloid leukemia (AML). They also predicted that WTAP was an HSP90 client protein, which maintained the stability of many oncoproteins and inhibited the anticancer efficiency of etoposide. After silencing WTAP, K562 cells showed significant apoptosis activity after etoposide treatment. A combined application of etoposide and WTAP inhibitors would escalate AML cell apoptosis. Circ0008399 (a novel circular RNA) promoted the expression of the target gene TNFAIP3 by increasing its mRNA stability in an m⁶A-dependent manner. As a result, WTAP diminished bladder cancer (BLCA) chemosensitivity to CDDP via the circ0008399/WTAP/TNFAIP3 pathway [129]. Ma et al. [130] suggested that WTAP-mediated DUSP6 upregulation contributed to carcinogenesis and drug resistance of nasal-type natural killer/T-cell lymphoma, providing a rationale for developing innovative avenues of antitumor therapeutics for natural killer/T-cell lymphoma (NKTCL). Likewise, WTAP bound to the m⁶A modified site of DLGAP1-AS1 contributed to stability, promoting BC-ADR through WTAP/DLGAP1-AS1/miR-299-3p feedback loop [131].

Targeting demethylase

FTO

Demethylase FTO played an oncogenic role in BC, AML, and other malignant tumors [149,150,151]. FTO-mediated m⁶A modification was also associated with drug resistance in various cancers, such as MM, glioblastoma, and melanoma. YAN et al. [69] confirmed that the TKI-tolerance phenotype emerged in leukemia patients because the overexpression of FTO caused m⁶A reduction. Signal transducers and activators of transcription 3 (STAT3) were constitutively active in several cancer types, and such hyperactivity was associated with an adverse clinical outcome [152]. Wang et al. [134] found increased expression of FTO and STAT3 in doxorubicin-resistant BC cells, and STAT3 bound to the FTO promoter to positively accommodate FTO expression. Moreover, FTO was involved in STAT3-mediated doxorubicin resistance and impaired doxorubicin sensitivity in BC cells. The overexpressing of FTO in cervical squamous cell carcinoma (CSCC) was resistant to radiotherapy and chemotherapy by the FTO-mediated mRNA demethylation and ERCC1 activity [135]. Interestingly, FTO was set up at high concentrations in patients’ MM cells and bone marrow tissues. Further analysis showed that FTO promoted bortezomib resistance by destabilizing SOD2 expression through an m⁶A-dependent manner, which might open up innovative therapeutic options [133]. JPX, a non-coding RNA adjacent to the X-inactive specific transcript, was entangled in tumor progression. It appeared that JPX interacted with the mRNA of phosphoinositide-dependent kinase-1 (PDK1) and promoted its stability and expression. Furthermore, JPX demethylated PDK1 mRNA, through its interaction with FTO alpha-ketoglutarate-dependent dioxygenase, contributed to the enhanced demethylation. Consequently, JPX exerted its GBM positive effects via the FTO/PDK1 axis and directly stabilized the PDK1 mRNA in temozolomide drug resistance [132]. Besides, the knockdown of FTO decreased the stability of PD-1, CXCR4, and SOX10, increasing RNA attenuation via m⁶A reader YTHDF2. It also sensitized melanoma cells to IFN-γ and anti-PD-1 therapy.

ALKBH5

ALKBH5, another m⁶A modification demethylase, was related to the onset, development, and prognosis of colon cancer, BLCA, EOC, and oral squamous cell carcinoma (OSCC) [153,154,155]. The downregulation of FTO and ALKBH5 in ovarian cancers with breast-cancer susceptibility gene 2 (BRCA2) mutations enhanced FZD10 mRNA m⁶A modifications, which ultimately reduced the sensitivity of PARPi via the Wnt/β-catenin pathway [138]. Moreover, ALKBH5 promoted cisplatin resistance in cancer cells [136]. HOXA10, the upstream transcription factor of ALKBH5, could form a loop with ALKBH5. In this way, ALKBH5 and HOXA10 together activated the JAK2/STAT3 signaling pathway, mediating JAK2 m⁶A demethylation and promoting EOC resistance to cisplatin. A recent study found that ubiquitin-specific proteases (USPs) were associated with T-cell acute lymphoblastic leukemia (T-ALL) occurrence and chemoresistance. ALKBH5 exhibited a carcinogenic effect on cancers and improved USP mRNA’s stability, resulting in GC resistance [137]. Multiple neoplasms expressed the human RNA helicase DDX3, essential for cell proliferation, invasion, and metastasis. By directly regulating ALKBH5, DDX3 could decrease m⁶A methylation of FOXM1 and NANOG transcripts, giving rise to cisplatin resistance in OSCC cells [139]. Likewise, the deletion of the m⁶A demethylase ALKBH5 sensitized tumors to cancer immunotherapy, suggesting that ALKBH5 may be a potential target to improve the outcome of immunotherapy for melanomas, CRC, and other underlying cancers [106]. In pancreatic cancer (PC), ALKBH5-mediated m6A modification caused DDIT4-AS1 overexpression, and DDIT-AS1 increased cancer stemness and led to gemcitabine resistance by destabilizing DDIT4 and activating the mTOR pathway [156].

Targeting other m⁶A regulators

So far, strategies targeting m⁶A mainly relied on the regulation of methyltransferase (such as METTL3 and WTAP) and demethylase. However, multiple sources of evidence suggested that other m⁶A modulators also had great potential as drug-therapeutic targets. For instance, the depletion of METTL14, core subunits of RNA methyltransferase, dramatically slowed tumor growth and prolonged the survival in mice bearing CT26 CRC and B16 melanoma [101]. m⁶A reader protein also played a pivotal role in drug resistance. In NSCLC, Keap1 was degraded following YTHDF1 depletion, facilitating Keap1-Nrf2-AKR1C1 axis cells and resulting in cisplatin resistance [140]. MicroRNA-145 could abrogate YTHDF2’s role as an oncogene in HepG2 cells associated with HCC [157]. In CRC, hypoxia-induced antisense lncRNA STEAP3-AS1 competed with YTHDF2 to STEAP3 mRNA binding site, protecting STEAP3 mRNA from m⁶A-mediated degradation and leading to high STEAP3 protein expression. Followed by this, activation of the Wnt/β-catenin pathway contributed to CRC progression [158]. Moreover, paclitaxel, 5-FU, and cisplatin were more effective in cell lines that lacked the m⁶A reader protein HNRNPC [30]. IGF2BP3, another m⁶A reader, was bound to the m⁶A modification region of ABCB1 mRNA and increased chemoresistance in CRC cells [141]. These studies illustrated that HNRNPC and IGF2BP3 could be latent biomarkers for chemoresistance.

m⁶A-targeted compounds

FTO inhibitors

Rhein was the first identified inhibitor for FTO in vitro and in vivo, which was neither a structural mimic of 2OG nor a chelator of the metal ion. Rhein blocked FTO demethylase by competitively binding its catalytic domain instead [159]. In therapy, the rhein-TKI combination synthetically eradicated relapsed/refractory leukemia [69], while rhein exposure increased the level of m⁶A in leukemia. In contrast, no growth arrest was observed after 24 hours of 20 μM rhein, proposing the anticancer therapy of rhein. Ascorbic acid also enhanced the activity of 2OG-dependent dioxygenases. In BC, ascorbic acid analog MO-I-500 exhibited antiproliferative activity in an FTO-dependent manner [160, 161]. However, rhein, as well as MO-I-500, was a broad-spectrum 2-OG inhibitor, which tremendously reduced their applications. In a high-throughput fluorescence polarization assay, meclofenamic acid (MA), a non-steroidal anti-inflammatory drug, was selected as the inhibitor of FTO. Moreover, the ethyl ester form of MA (MA2) upgraded levels of m⁶A modification in mRNA [162]. Additionally, MA2 inhibited self-renewal and tumorigenesis of GSCs in a GSC-xenograft mouse model and prolonged survival [163]. Of note, MA2 enhanced the antitumor effect of chemotherapy in glioma [164]. As a result of the specific inhibitory property of MA, higher potency derivatives were designed and synthesized. A new MA-derived inhibitor, FB23, directly bound to FTO and selectively inhibited its activity, which possessed 140-fold over that of MA. The benzohy-droxamic acid, termed FB23–2, was a further practical analog of FB23 [165]. FB23–2 exhibited FTO-dependent anti-leukemia effects broadly and targeted the same signaling pathways as FB23. Dac51, another small-molecule analog of FB23, could modulate the tumor microenvironment via inhibiting FTO and mounting CD8+ T cell infiltration, contributing to a remarkable antitumor efficac y[105]. FTO-04 demonstrated robust inhibition of neurosphere formation in patient-derived GSCs but did not inhibit the growth of healthy human neural stem cells. On the side, FTO-04-mediated inhibition of FTO increased m⁶A modification and demethylated N6,2′-O-dimethyladenosine (m⁶A_m) levels of GSCs [166]. Nafamostat mesylate often was applied in treating pancreatitis and cancers. The combination of thermodynamic and enzymatic activity provided insight into the FTO inhibition of nafamostat mesylate [167]. R-2-hydroxyglutarate (R-2HG) was architecturally and chemically similar to another inhibitor, 2OG. R-2HG inhibited FTO’s enzymatic activity by competitive inhibition and proved the overall antitumor effect. As a result of the R-2HG therapeutic regimen, m⁶A modification levels increased. Meanwhile, aerobic glycolysis was suppressed by inhibiting FTO activity and downstream signaling molecules, consisting of MYC, CEBPA, PFKP, and LDHB [168, 169]. CS1 and CS2 displayed a much higher efficacy. Consequently, two highly efficacious FTO inhibitors were named CS1 and CS2. They displayed a much higher efficacy in inhibiting AML cells’ viability than two previously reported FTO inhibitors (FB23–2 and MO-I-500) [103]. Therefore, FTO represented a modern therapeutic potential to target cancer therapy, and more clinical studies were required to confirm the long-term side effects of these inhibitors.

METTL3 inhibitors

Bedi et al. [170] reported a virtual screening method for almost 4000 adenosine derivatives to identify potential METTL3 inhibitors. Their best compound, S-adenosyl-L-methionine (SAM) mimic, was the first small molecule to inhibit METTL3. METTL3 inhibitors possessed excellent ligand efficiency, and their binding patterns were validated by protein crystallography. Respective RNA m⁶A methyltransferase inhibitors displayed anticancer abilities. Accompanied by the selective reduction of m⁶A levels on known leukemogenic mRNAs, STM2457 treatment reduced AML growth and increased differentiation and apoptosis [171]. Another METTL3 chemical inhibition, UZH1a, reduced the m⁶A/A ratio in mRNAs of different cell lines, revealing the potential implications of METTL3 inhibition in tremendous disease models [172].

Other m⁶A regulator activators and inhibitors

Using silico-based discovery could identify small-molecule ligands binding to the METTL3–14-WTAP complex. Primarily, SAM bonded with Asp377 and acted as a hydrogen bond donor to the Asp395 of METTL3 protein. Similarly, four compounds bound to the extent of the METTL3 enzyme relating to Asp295, Phe534, Arg536, and Asn539. METTL3-METTL14 RNA m⁶A methyltransferase complex activators provoked cells to modify mRNA m⁶A [173]. Their potential anticancer effects needed more experiments to prove. Li and his team [106] identified a small molecule inhibitor of ALKBH5 by using the X-ray crystal structure in silico screening of compounds and named ALK-04. Compound libraries verified this specific inhibitor. Subsequent proof found that melanoma tumor growth was significantly reduced in mice applying the ALK-04 compared to the control group. This study also provided evidence for ALKBH5 inhibitors combined with immunotherapy against melanoma. BTYNB has been identified by compound library screening with its ability to inhibit c-Myc and IGF2BP1 protein selectively [174]. The small molecule BTYNB also destabilized E2F1 mRNAs by impairing the IGF2BP1-RNA association, which interfered with cellular protein synthesis and tumor growth [174]. Table 3 collates the identified m⁶A-targeted compounds.

Table 3 Identified m⁶A-targeted compounds

Conclusion and perspective

Despite considerable research underway to understand the function of m⁶A modifications in cancer proliferation and drug resistance, many questions remain unanswered. For example, as a broad RNA modification in eukaryotic messenger RNA, will the m⁶A regulator targeted compounds be a good candidate in tumor therapy? How to focus and target key molecules? How to specifically target the regulatory axis involved in m⁶A to reverse drug resistance in tumor tissue?

The practical significance of m⁶A modifications and regulators heralded a new dawn for targeting m⁶A regulators in therapy. However, few m⁶A-phenotype associated inhibitors and activators are clinically applicable. Followings might be responsible for this plight. Firstly, due to lacking study on cellular activity, how these compounds actually affect methylation levels is elusive. Secondly, adenosine analogs have poor cell permeability and pharmacokinetics, complicating their potential use. Thirdly, tumor heterogeneity and rare predictors mound a barrier between the targeted compounds and distinct cancers, contributing to poor clinical applicability. Therefore, further screening of potential agents is needed. For the precise regulation of m⁶A modifications (global and/or targeted), protein-protein interactions (PPI) or protein-nucleotide interactions would be promising strategies. Further studies on tumor biology, the development of high-quality chemical probes, and preclinical studies will help to identify precise biomarkers, which are crucial for individualized treatment, improved outcomes, and potential toxicity prediction. In addition, most of the reported targeted compounds are cytotoxic, whereas non-cytotoxic inhibitors that modulate the immune system also represent a promising combination. For example, the ALKBH5 inhibitor ALK-04 showed significant synergy with anti-PD-1 therapy while without cytotoxicity in vivo. Overall, the clinical application of compounds targeting m⁶A is still in its infancy. As the understanding of epigenomics in cancer grows, there is great promise for those therapy-resistant patients accompanied with abnormal m⁶A manners.