The role of RNA demethylase ALKBH5 in cancer and its potential for treatment

Breast cancer represents the major female malignancy, affecting one in 20 women worldwide, with an even higher incidence in high-income countries [70]. ALKBH5 amounts are increased in immortalized and oncogenically-transformed human mammary epithelial cells in comparison with their primary cell precursors [71]. Besides, ALKBH5 is upregulated in breast cancer samples compared to noncancerous tissue [72]. Knockdown of ALKBH5 inhibits breast cancer cell viability, colony formation and migration, confirming the oncogenic properties of ALKBH5 in this malignancy [72]. In breast cancer, there is a small subset of cells, called breast cancer stem cells (BCSCs), which could both self-renew and differentiate into cancer cells [73]. BCSCs are resistant to chemotherapy and survive the treatment, leading to tumor recurrence and metastasis [73]. Zhang et al. found that NANOG, a core pluripotency factor, is regulated by ALKBH5 in breast cancer cells [74]. Indeed, ALKBH5 overexpression decreases m⁶A methylation of NANOG mRNA at 3′ untranslated region (3′UTR), increasing NANOG mRNA stability and levels, and induces BCSC enrichment in hypoxic breast cancer cells [74]. Later, these authors found another pluripotency factor, KLF4, is also positively regulated by ALKBH5-mediated m⁶A methylation in breast cancer, contributing to the overall regulation of the BCSC phenotype by ALKBH5 [75].

Glioma is a central nervous system tumor that seriously endangers the physical and mental health of patients [76]. ALKBH5 is upregulated in glioma and of significance in regulating the metabolism and development of glioma [77]. Specifically, ALKBH5 demethylates the transcript of glucose-6-phosphate dehydrogenase (G6PD), the rate-limiting enzyme of the pentose phosphate pathway (PPP), and enhances its mRNA stability, thus promoting oxidative PPP flux and stimulating the aggravation of glioma [77]. In addition, AKT2 mRNA was identified as another critical target of ALKBH5 that contributes to the tumor-promoting effects of ALKBH5 in glioma cells [78].

Glioblastoma (GBM) represents the commonest primary brain cancer with high heterogeneity and poor prognosis, making up 54% of all gliomas [79, 80]. GBM stem-cell like cells (GSCs) constitute an important factor in maintaining tumor growth and inducing tumor recurrence [81]. ALKBH5 is overexpressed in GSCs and correlated with reduced patient survival in GBM [82]. It demethylates the nascent transcript of FOXM1 and promotes the interaction of FOXM1 pre-mRNA with the nuclear RNA binding protein HuR, resulting in increased FOXM1 expression, which is critical for GSC growth and self-renewal [82]. ALKBH5 also plays an important role in radioresistance and invasiveness of GSCs [83]. ALKBH5 enhances radioresistance by modulating genes involved in homologous recombination (HR) and contributes to the aggressiveness of GBM by upregulating YAP1 expression [83]. Temozolomide, an alkylating agent, is a first-line chemotherapeutic orally administered following surgical excision in GBM [83]. However, temozolomide resistance is frequently found in the middle and late stages of chemotherapy [84]. ALKBH5 could demethylate SOX2 mRNA, enhancing SOX2 expression and promoting temozolomide resistance in GBM [85].

Lung cancer ranks first among cancers in terms of morbidity and mortality globally, with non-small-cell lung cancer (NSCLC) representing the main histological subtype [86]. ALKBH5 undergoes ectopic upregulation in NSCLC, and is tightly associated with reduced patient survival [87]. ALKBH5 promotes proliferation and reduces apoptosis in NSCLC cells by repressing TIMP3 mRNA stability and translation [87]. UBE2C, an oncogene that selectively represses autophagy and promotes cell proliferation in NSCLC, is epitranscriptionally stabilized with reduced m⁶A amounts within its mRNA because of ALKBH5 upregulation in NSCLC [88]. Meanwhile, m⁶A demethylation caused by ALKBH5 stabilizes oncogenic drivers, including SOX2, SMAD7 and MYC, via the m⁶A reader protein YTHDF2 in KRAS mutant NSCLC [89].

In addition to NSCLC, ALKBH5 is elevated in lung adenocarcinoma cells submitted to intermittent hypoxia [90]. ALKBH5 knockout in lung adenocarcinoma cells increases m⁶A abundance in FOXM1 transcript and decreases its translation, leading to reduced cell proliferation and invasion [90]. ALKBH5-dependent m⁶A demethylation of the lncRNA RMRP also has an oncogenic effect in lung adenocarcinoma [91]. RMRP deficiency in lung adenocarcinoma cell lines suppresses cell proliferation, migration and invasion, and promotes cell apoptosis, mimicking the effects of ALKBH5 inhibition [91].

As a malignant neoplasm affecting women with a high death rate [92], the development ovarian cancer was reported to be tightly associated with m⁶A modification. A study examined the expression profiles of AlkB family members in ovarian serous carcinoma, and revealed that cases highly expressing ALKBH5 have reduced overall survival (OS) and progression-free survival (PFS) [93]. ALKBH5 expression is increased in epithelial ovarian cancer in comparison with noncancerous ovarian tissue [94]. ALKBH5 inhibits autophagy and promotes malignancy in ovarian cancer cells by stabilizing BCL2 mRNA and promoting BCL2 binding to BECN1 [94]. Besides, ALKBH5-mediated mRNA demethylation increases NANOG expression and enhances aggressiveness of ovarian cancer cells [95].

Acute myeloid leukemia (AML) represents a fatal hematologic cancer featuring uncontrolled expansion of poorly differentiated myeloid cells, with hematopoietic stem and progenitor cells (HSPCs) retaining their self-renewal ability and impaired myeloid differentiation [96]. Kwok and collaborators demonstrated ALKBH5 is commonly deleted in AML, particularly in individuals with TP53 mutations, based on the cancer genome atlas (TCGA) data [97], implying ALKBH5’s tumor-suppressive function in AML. However, recent studies showed that ALKBH5 is actually upregulated in human AML, with its elevated amounts associated with reduced patient survival [98, 99]. These studies all demonstrated that ALKBH5 is selectively required for self-renewal of leukemia stem cells (LSCs) but not for normal hematopoiesis [98, 99]. Mechanistically, ALKBH5 enhances malignancy in AML by post-transcriptionally regulating TACC3 and AXL [98, 99].

Gastric cancer (GC) represents a very common malignancy of the digestive tract, ranking fifth and fourth in terms of incidence and mortality among all cancers, respectively [100]. ALKBH5 is overexpressed in GC and downregulates m⁶A levels of the lncRNA NEAT1, leading to the upregulation of NEAT1 expression [101]. Increased NEAT1 then positively affects the expression of EZH2, a subunit of the polycomb repressive complex, by acting as a scaffold and thus promoting GC invasion and metastasis [101]. Colorectal cancer is the third incident and second deadliest malignancy worldwide, representing about one in 10 cancer cases and deaths [100]. ALKBH5 knockdown suppresses the proliferative and migratory abilities of colorectal cancer cells partially through ALKBH5-related NEAT1 downregulation [102], which is consistent with the finding in GC cells [101].

Besides the above cancers, ALKBH5 is also correlated with poor prognosis and/or plays an oncogenic role in other malignancies, including endometrial cancer [103, 104], renal cell carcinoma [105], hepatitis B virus (HBV) related hepatocellular carcinoma (HCC) [106], osteosarcoma [107], uveal melanoma [108], pituitary adenomas [109], esophageal squamous cell carcinoma [110], cervical cancer [111] and oral squamous cell carcinoma [112]. During the investigation of underlying mechanisms, a set of functional targets of ALKBH5 were identified in these cancers (Table 2; Fig. 3).

While ALKBH5 has overt oncogenic roles in most cancers, it is a tumor suppressor in others (Fig. 2; Table 2). Reduced ALKBH5 amounts has been described in the following malignancies.

Bladder cancer (BC) is among the most prevalent cancers globally, with approximately 573,000 new cases reported in 2020 [100]. Chen et al. demonstrated that ALKBH5 expression in BC shows positive correlations with grade and M1 stage, indicating ALKBH5 might be an oncogene in BC [113]. However, other studies demonstrated that ALKBH5 is actually a tumor suppressor in BC according to cell and animal assays. Yu et al. found that reduced ALKBH5 amounts in BC are associated with decreased patient survival [114]. Functional assays showed ALKBH5 silencing increases the proliferative, migratory and invasive abilities of BC cells and reduces their cisplatin chemosensitivity via a CK2α-related glycolytic pathway [114]. Jin and colleagues examined the role of m⁶A modification in BC, and revealed ALKBH5, but not METTL3, inhibits cell adhesion of BC by repressing ITGA6 expression [115]. Specifically, ALKBH5 decreases m⁶A methylations in the 3′UTR of the ITGA6 transcript, suppressing ITGA6 protein synthesis mediated by YTHDF1 and YTHDF3 [115].

Pancreatic cancer (PC) represents one of the rare cancers whose mortality and incidence rates are almost identical, ranking seventh among deadliest cancers globally [116]. Studies revealed enhanced survival in PC cases highly expressing ALKBH5 [117, 118]. ALKBH5 exerts tumor-suppressive effects via the demethylation of the lncRNA KCNK15-AS1, controlling KCNK15-AS1-dependent cell migration and invasion in PC [119]. In addition, ALKBH5 post-transcriptionally enhances PER1 expression in an m⁶A-YTHDF2-dependent fashion, activating ATM-CHK2-P53/CDC25C signaling and suppressing PC cell proliferation [120]. Moreover, ALKBH5 acts as a tumor suppressor in pancreatic ductal adenocarcinoma (PDAC) and sensitizes PDAC cells to gemcitabine by reducing WIF1 mRNA methylation and modulating the Wnt pathway, which leads to the downregulation of C-MYC, Cyclin D1, MMP-2 and MMP-9 [121].

Although most reports showed that ALKBH5 is an oncogene in NSCLC [87–89], as stated above, another work reported its tumor suppressive role in NSCLC. Jin et al. showed ALKBH5 inhibits tumor cell proliferation and metastasis by decreasing YTHDFs-dependent YAP1 expression in NSCLC [122]. Meanwhile, ALKBH5 decreases YAP1 activity by HuR-dependently controlling miR-107/LATS2 expression [122]. In osteosarcoma, Yuan and co-workers discovered downregulated levels of ALKBH5 in osteosarcoma cell/tissue specimens in comparison with noncancerous osteoblast cell/tissue samples [123], which was contrary to the observation of a previous study [107]. According to Yuan’s study, ALKBH5 overexpression markedly reduces the proliferative, migratory and invasive features of osteosarcoma cells, while triggering apoptosis [123]. Similar to the finding in NSCLC [122], ALKBH5 suppresses osteosarcoma progression via m⁶A-related direct/indirect YAP1 regulation [123]. Specifically, YAP1 expression is directly suppressed through mRNA m⁶A methylation, alongside indirect suppression by ALKBH5-regulated miR-181b-1 [123].

ALKBH5 downregulation was demonstrated in clinical colon cancer tissue samples, which was tightly associated with distant metastasis and disease stage [124]. Further functional assays confirmed ALKBH5 overexpression inhibits colon cancer cell invasion in vitro and metastasis in experimental animals [124]. However, the underlying mechanism remains to be elucidated. In addition, ALKBH5 also functions as a tumor suppressor in clear cell renal carcinoma (ccRCC) and HCC. In ccRCC, reduced ALKBH5 gene expression is associated with shorter OS, indicating that ALKBH5 could be used as a prognostic biomarker [125]. In HCC, ALKBH5 suppresses the proliferative and invasive properties of cancer cells by m⁶A-mediated inhibition of LYPD1 [126].

Hypoxia, a feature of most tumors, results from fast cell division and abnormal angiogenesis [137]. Hypoxia-inducible factors (HIFs) control transcriptional responses to local hypoxia in cancer and induce tumor progression by affecting cell metabolism, promoting angiogenesis, controlling stem cell state and other cellular processes [137]. HIFs consist of α (HIF-α) and β (HIF-β) subunits and interact with canonical DNA sequences termed hypoxia-response elements (HREs) in the promoters or enhancers of target genes, upregulating 150 genes or more [137]. An early study found that ALKBH5 is a direct target of HIF-1α, indicating that ALKBH5 may be involved in the regulation of cellular responses to hypoxia [138]. Later, it was demonstrated in human breast cancer cells that ALKBH5 is significantly upregulated under hypoxic conditions, while knockdown of HIF-1α and/or HIF-2α abrogates this effect [74, 75]. Similarly, hypoxia and HIFs elevate ALKBH5 expression in lung adenocarcinoma cells [90], renal cell carcinoma cells [105], endometrial cancer stem cells [104] and pituitary adenoma cells [109], leading to invasive tumor phenotype, enhanced tumor initiation capacity, sustained stem-like state and poor clinical outcomes.

Epigenetic modifications, including DNA methylation, histone modifications and nucleosome positioning, affect gene expression with no alteration of the underlying DNA sequence [139]. Dysregulation of these processes is frequently observed in human cancers, some of which are responsible for ALKBH5 dysregulation in cancers.

Histone modifications and associated enzymes are involved in chromatin compaction, nucleosome dynamics and transcription activation or suppression [140]. In leukemia-initiating cells, the promoter regions of ALKBH5 have markedly elevated amounts of active histone markers, including H3K9ac, H3K4me3, H3K4me2 and H3K79me2, and reduced levels of the repressive histone marker H3K27me3 and H3K9me3, leading to a higher degree of chromatin openness [99]. Consistently, a set of H3K4 histone methyltransferases (MLL1 and MLL2) and H3K9 histone demethylases (JMJD2B and JMJD1C) were identified as ALKBH5 promoter-binding proteins in AML cells [99]. Further investigation confirmed knockdown of MLL1 and MLL3, but not MLL2 and MLL4, markedly reduces ALKBH5 amounts in leukemia cells [99]. Moreover, histone demethylase KDM4C controls ALKBH5 expression in leukemia by reducing H3K9me3 levels and enhancing chromatin accessibility of the ALKBH5 locus, which promotes the binding of the transcriptional factor MYB and the C-terminal domain (CTD) of the RNA polymerase II to the ALKBH5 promoter [99]. These results showed ALKBH5 overexpression is controlled by chromatin state in the pathogenesis of AML.

Epigenetic dynamics of ALKBH5 has also been researched in HBV-HCC, uveal melanoma (UM) and NSCLC. The HBV X protein (HBx) was found to induce an elevated degree of epigenetic H3K4me3 modification of chromatin by upregulating WD-40 Repeat Protein 5 (WDR5), contributing to HBV-induced hepatocellular carcinogenesis [141]. It was demonstrated that highly expressed ALKBH5 in HBV-HCC occurs via HBx- and WDR5-dependent H3K4me3 modification of the ALKBH5 promoter following HBV infection [106]. Elevated ALKBH5 amounts then enhance HBx mRNA expression, forming a positive-feedback loop that contributes to HBV-associated liver carcinogenesis [106]. In UM, EP300-induced H3K27 acetylation activation upregulates ALKBH5 expression [108]. ALKBH5 level is starkly decreased in UM cells after treatment with C646, a histone acetyltransferase suppressor specific to EP300 [108]. Many NSCLC cases contain concurrent KRAS mutation, alongside loss of LKB1, which negatively regulates the total 5mC DNA methylation [142–144]. LKB1 deficiency upregulates ALKBH5 by inducing DNA hypermethylation of the transcriptional repressor CTCF binding region on ALKBH5 gene’s promoter, preventing CTCF binding while enhancing active histone modifications such as H3K4me3, H3K9ac and H3K27ac [89].

Transcription factors bind DNA in a sequence-specific manner to control chromatin and transcription, forming a complicated system that impact gene expression [145]. During the investigation of ALKBH5 promoter-binding proteins in AML cells, a set of transcription factors was identified, including RUNX2, MYB and PBX3 [99]. Further investigation confirmed MYB silencing markedly decreased ALKBH5 mRNA and protein amounts in leukemia cells [99]. The well-known transcription activator P53 is mutated or suppressed in about half of cancers [120]. A TCGA dataset assessment demonstrated elevated and reduced ALKBH5 amounts in the P53 wild-type and mutant groups of PC cells, respectively [120]. P53 interacts with the ALKBH5 promoter, transcriptionally activating ALKBH5 and indirectly reducing m⁶A amounts in PC cells [120].

While ALKBH5 acts as an important factor controlling the expression and functions of a number of ncRNAs, some ncRNAs also influence its expression and function. The crosstalk between ALKBH5 and ncRNAs appears to form a feedback loop with extensive impacts in cancers.

LncRNAs, i.e., noncoding transcripts of more than 200 nucleotides, might play the roles of enhancers, scaffolds or decoys by physically binding other RNAs or proteins [146]. In GBM cells, the lncRNA FOXM1-AS (antisense to FOXM1) enhances ALKBH5 binding to FOXM1 nascent mRNA [82]. FOXM1-AS suppression mimics ALKBH5 depletion in modulating FOXM1 methylation and biosynthesis [82]. Similarly, the lncRNA GAS5-AS (antisense to GAS5) enhances GAS5 stability by binding to ALKBH5 and controlling m⁶A modifications of GAS5 in cervical cancer [147]. In GBM cells with temozolomide-resistance, the lncRNA SOX2OT functions as a scaffold that recruits ALKBH5, positively regulating its function and expression [85].

MiRNAs are 18–25 nucleotide-long ncRNAs that inhibit gene expression by directly interacting with complementary target mRNAs, which are then degraded or translationally inhibited [148]. ALKBH5 mRNA was recognized as a miR-193a-3 target in both glioma and esophageal squamous cell carcinoma (ESCC), with its 3′UTR being a miR-193a-3p binding site [78, 149]. miR-193a-3p transfection results in substantial ALKBH5 downregulation, while miR-193a-3p inhibition upregulates ALKBH5 [78, 149]. Meanwhile, in ESCC, ALKBH5 suppresses the maturation of primary miR-193a-3p by decreasing its m⁶A modification in turn, suggesting a feedback regulation between miR-193a-3p and ALKBH5 [149].

CircRNAs represent single-stranded, covalently closed RNAs [150]. They interact with and sequester target proteins to adequate subcellular locations, and modulate some protein–protein and protein-RNA interactions [150]. CircRNA cIARS (hsa_circ_0008367) controls sorafenib-induced autophagy and ferroptosis in HCC cells through interaction with ALKBH5 [151], which inhibits autophagy by demethylating BCL2 mRNA and promoting BCL2-BECN1 binding [94]. cIARS interacts with ALKBH5 directly and represses its regulatory effect on BCL2-BECN1 autophagy regulatory complex, leading to increased autophagic flux and ferritinophagy [151].

DDX3, a DEAD-box RNA helicase, interacts with ALKBH5 in an RNA-independent manner and modulates the demethylation activities of ALKBH5 [152]. The ATP and DSBH domains of DDX3 and ALKBH5, respectively, were required for this interaction [152]. Later, another study confirmed that ALKBH5 is directly regulated by DDX3 in oral squamous cell carcinoma cells [112]. However, the study showed that ATP domain of DDX3 was not essential for DDX3 mediated up-regulation of ALKBH5 [112]. In human ovarian cancer, ALKBH5 and TLR4 amounts are elevated in cancer cells after co-culture with alternatively activated (M2) macrophages [95]. Specifically, TLR4 positively regulates ALKBH5 through NF-κB signaling [95].

To identify specific ALKBH5 suppressors, a high-throughput in silico screening of 144,000 preselected molecules was performed, with further analysis by the enzyme inhibition assay [162]. The authors ultimately identified two ALKBH5 inhibitors, including 2-[(1-hydroxy-2-oxo-2-phenylethyl)sulfanyl]acetic acid (Fig. 5a, Inhibitor A; IC50 = 0.84 μM) and 4-[(furan-2-yl)methyl]amino-1,2-diazinane-3,6-dione (Fig. 5a, Inhibitor B; IC50 = 1.79 μM), which were tested for activity in leukemia and GBM cell lines [162]. In leukemia HL-60, CCRF-CEM and K562 cells, theses inhibitors blunted cell proliferation in the low micromolar range (IC50 = 1.38–16.5 μM) [162]. However, another leukemia cell line, Jurkat, and glioblastoma A-172 cells had low to no susceptibility, suggesting ALKBH5 suppression by small-molecule inhibitors as a feasible cancer-cell-type-selective anti-proliferative approach [162]. A novel sodium channel blocker, named MV1035 (Fig. 5a, Inhibitor C), potently inhibits ALKBH5 activity by an alternative off-target interaction with ALKBH5, with the binding site of MV1035 in ALKBH5 partially overlapping with that of 2OG [163]. It was demonstrated that the inhibitory effects of MV1035 on glioma cell migration and invasion rely on ALKBH5 inhibition but not sodium channel blocking [163]. Besides, a specific inhibitor of ALKBH5, named ALK-04, could potentiate the efficacy of anti–PD-1 immunotherapy in melanoma treatment, providing a rational for future combination drug therapy [133].

In addition to naturally-derived small-molecule inhibitors of ALKBH5, developing synthetic small-molecule inhibitors represent another approach to inhibit ALKBH5. ALKBH5 requires an interaction with 2OG and ferrous iron to function as an m⁶A demethylase [68]. Therefore, it would be reasonable to design suppressors by taking into consideration the reported binding pocket of this protein. Particularly, a 2OG-tethering approach of concurrently occupying both the 2OG- and substrate-binding sites is of interest. Linking 2OG derivatives with substrate analogs was successfully utilized to develop selective histone demethylase suppressors [164, 165] and AlkB enzyme inhibitors [166], which can be used to design specific ALKBH5 inhibitors in the future.

Given the tumor suppressive role of ALKBH5 in a small number of cancers, activators of ALKBH5 are needed in such cases. However, pharmacologic activators specifically targeting ALKBH5 have not been reported to date, raising the need for developing ALKBH5 activators. The ability to bind metal cofactor (Fe²⁺) and the secondary substrate (2OG) precedes binding of the m⁶A-containing oligonucleotide to ALKBH5 and is crucial for demethylation activity of ALKBH5 [167]. Therefore, the binding sites of Fe²⁺ and 2OG could be used as target areas for identifying potential compounds as activators of ALKBH5 by high-throughput screens. Flavin mononucleotide, a metabolite produced by riboflavin kinase, was found to mediate substantial photochemical demethylation of m⁶A residues of RNAs in live cells, indicating that regulation of m⁶A modification by external small organic molecules is possible [168]. It also provided insights into the development of powerful small molecules as RNA demethylases and specific functional probes targeting m⁶A modification for biomolecular and treatment purposes [168].

Besides small molecule-based occupancy-driven pharmacology as an approach to inhibit ALKBH5 functions, PROTAC represents a new technological method for degrading proteins of interest (Fig. 4b). A PROTAC is a heterobifunctional molecule that consists of a protein comprising a ligand for interacting with a target protein and another interacting with an E3 ubiquitin ligase, with a linker utilized for connection (Fig. 5b, right panel) [169, 170]. Indeed, a PROTAC recruits the E3 ligase to the protein of interest and induces its ubiquitination and degradation [169, 170]. The main advantages of PROTACs include their high selectivity, the potential to target non-druggable proteins, and the capacity of overcoming resistance to small-molecule suppressors via targeting of mutant proteins [170]. Using PROTACs for degrading proteins critical for tumorigenesis represents a potential therapeutic approach in cancer. PROTACs have be successfully used in solid (targeting AR, ER, FAK and P38) and hematological (targeting BRD4, BTK, BCR-ABL and CDK-6) tumors [170]. Importantly, the first PROTACs against AR and ER have been applied clinically [170]. Whether such approach would degrade ALKBH5 in cancer deserves further attention.

A potential concern is that targeting ALKBH5 utilizing the abovementioned techniques or reagents would affect the global demethylation of mRNAs, including both oncogenes and tumor suppressor genes, or genes with other functions, which may lead to potential deleterious effects [171]. According to previous studies (Table 2; Fig. 3), the tumor-promoting or suppressing role of ALKBH5 is largely mediated by a few critical targets, suggesting that modulating the m⁶A on specific RNAs could alleviate tumor progression with reduced side effects. Therefore, developing a site-specific m⁶A modulation system may provide a more potent tool for correcting dysregulated m⁶A modifications in cancers.

The development of Cas13, a novel RNA-targeting CRISPR-associated nuclease, enables the targeting of dynamics of endogenous RNAs [172, 173]. Li et al. have recently developed a CRISPR–Cas13b-based tool, named dm⁶ACRISPR, to specifically demethylate target mRNAs [171]. The dm⁶ACRISPR fusion protein was constructed by linking catalytically inactive Cas13 enzyme (dCas13b) to ALKBH5 (Fig. 5c, left panel) [171]. dm⁶ACRISPR effectively demethylated target transcripts without significant off-target effects, and was successfully applied to target oncogenic mRNAs, such as EGFR and MYC mRNAs, to decrease HeLa cell proliferation [171]. In addition, another team designed programmable m⁶A modification utilizing the CRISPR-based METTL3 methyltransferase domain and METTL3:METTL14 methyltransferase complex, named dCas13-M3nls and dCas13-M3M14nes respectively, which can install m⁶A on targeted mRNAs (Fig. 5c, right panel) [174]. Together, these technologies could work together to add or delete functional essential m⁶A sites on specific RNAs. Considering that CRISPR-based gene therapy has achieved great success, with many ongoing phase I/II clinical studies [175], CRISPR-based m⁶A targeting may provide a powerful strategy for m⁶A-based precision medicine.

The finding of upstream regulators of ALKBH5 in cancers (Table 2; Fig. 4) provides an alternative approach to inhibit ALKBH5 by targeting its regulatory machinery. These ALKBH5 regulators, including multiple epigenetic modulators, transcription factors and non-coding RNAs, are also dysregulated in specific cancers and exert similar effects on cellular phenotypes as ALKBH5. Inhibitors of these regulators, like DNA methylation inhibitor 5-azacytidine (5-aza) [89] and histone acetyltransferase EP300 inhibitor C646 [108], as well as miRNA miR-193a-3p mimic [78, 149] are able to indirectly or directly suppress ALKBH5 expression in cancer cells (Fig. 5d). Numerous epigenetic, HIF-targeted and miRNA-directed therapies have been examined in clinical trials [176–178], showing potential of targeting these kinds of molecules for cancer treatment, which might exert anti-tumor effects partly through altering ALKBH5-mediated m⁶A demethylation.

In the setting that ALKBH5 functions as a tumor suppressor, inhibiters targeting negative regulators of ALKBH5 demonstrate great pharmacological potential as cancer remedy. This approach has proven to be beneficial in the case of activating tumor suppressor p53 for cancer treatment. Several compounds have been developed to specifically target MDM2, a negative regulator of p53, leading to p53 stabilization and activation [179]. For ALKBH5, epigenetic modulators, e.g., protein arginine methyltransferase PRMT7, that function to repress ALKBH5 transcription have been investigated with mechanistic insights gained. PRMT7 binds to the BRG1-related hSWI/SNF chromatin remodeling complex and carries out the methylation of histones H2A Arg-3 (H2AR3) and H4 Arg-3 (H4R3), leading to negative regulation of ALKBH5 expression [180]. Thus, inhibitors targeting PRMT7 may effectively reverse the inhibition of ALKBH5 transcription. In addition, SUMO E2 conjugation enzyme UBC9 and E3 ligase PIAS4 mediates ALKBH5 SUMOylation at lysine residues K86 and K321, which markedly inhibits demethylase activity of ALKBH5 by blocking its binding to m⁶A RNA species [57], indicating the design of small-molecule inhibitors targeting the modulators of ALKBH5 post-translational modifications might be another intriguing strategy to reactivate ALKBH5. Further understanding of regulatory mechanisms on the expression, activation and stabilization of ALKBH5 may provide more ideal targets to up-regulate its level and enhance its tumor-suppressive activities in some cancers.

Gene therapy is the treatment of a genetic disease by repairing or reconstructing defective genetic material [181]. For cancers caused by a deficiency in ALKBH5, such as bladder cancer [114, 115] and pancreatic cancer [119–121], the most direct way to restore the expression and function of ALKBH5 is to introduce wild-type ALKBH5 into cancer cells. This can be achieved by in vivo delivery of ALKBH5 gene into the target tumor cells by vectors based on retroviruses, adenoviruses or adeno-associated viruses (AAVs) [181]. Due to their gene delivery efficacy, lack of pathogenicity, and strong safety profile, AAV vectors may be an excellent platform for delivering therapeutic genes [182]. Recombinant AAV vectors are generated by replacing the endogenous rep and cap genes with an expression cassette consisting of a promoter driving a transgene of interest and a poly(A) tail (Fig. 5e) [183]. The viral vectors can be administered intratumorally and are able to infect cancer cells to achieve ectopic expression of the therapeutic transgene [183]. AAV vector-mediated gene delivery was recently approved for the treatment of inherited blindness and spinal muscular atrophy [183, 184]. Additionally, AAV vectors have been increasingly employed in a variety of preclinical tumor models to deliver a wide array of therapeutic genes and tumor suppressor genes, such as TP53 [182]. The application of gene therapy to restore ALKBH5 expression in cancers as well as its clinical efficacy and biosafety await further validation.