RNA N6-methyladenosine methylation and skin diseases

Abstract

Skin diseases are global health issues caused by multiple pathogenic factors, in which epigenetics plays an invaluable role. Post-transcriptional RNA modifications are important epigenetic mechanism that regulate gene expression at the genome-wide level. N6-methyladenosine (m6A) is the most prevalent modification that occurs in the messenger RNAs (mRNA) of most eukaryotes, which is installed by methyltransferases called “writers”, removed by demethylases called “erasers”, and recognised by RNA-binding proteins called “readers”. To date, m6A is emerging to play essential part in both physiological processes and pathological progression, including skin diseases. However, a systematic summary of m6A in skin disease has not yet been reported. This review starts by illustrating each m6A-related modifier specifically and their roles in RNA processing, and then focus on the existing research advances of m6A in immune homeostasis and skin diseases.

Keywords:

• N6-methyladenosine (m6A)
• RNA methylation
• skin diseases
• skin cancers

1. Background

Skin is the largest organ of the body that directly contacts with the external environment and functions as the first defensing line against injury and infection from the outside world. The pathogenesis of most skin diseases is primarily an inflammatory response caused by internal and external factors. Heritable changes to gene expression resulted from environmental factors like radiation (e.g. UVB [1], X-rays, gamma rays [2] etc.), air pollution [3], and individual lifestyle [4,5], either at the mRNA level or protein level without altering the DNA sequences, are known as epigenetics. Epigenetic mechanisms that regulate expression of key genes from complex signalling networks during different skin conditions are emerging to play critical roles in the initiation and progression of skin diseases.

Epigenetic events affect the expression of related genes by controlling the access of transcription regulators to chromatin and involving complex interactions among transcription factors, chromatin remodellers, histone modifications, DNA methylation and non-coding RNA molecules [6]. RNA modifications have also been identified recently to be closely related to cell homeostasis and the development of diseases. According to the Modomics database, more than 170 RNA modifications have been identified (https://iimcb.genesilico.pl/modomics/). In eukaryotes, most RNA modifications occur in transfer RNA (tRNA) and ribosomal RNA (rRNA) [7], and N6-methyladenosine (m6A) modification is the predominant and well-studied decoration in mRNA and long non-coding RNA (lncRNA) [8,9]. Given the universality, reversibility, and multifunctionality of RNA m6A modifications, investigations on RNA m6A modifications provide new perspectives on the aetiology of skin diseases and associated therapies. This review systematically summarises the evolution, composition, and regulatory mechanisms of RNA m6A, as well as latest m6A studies in dermatosis, enabling a comprehensive understanding of m6A function in dermatosis and providing insights into more precise diagnosis and therapies.

2. Dynamic regulation of N6-methyladenosine modification

m6A modification of RNA is dynamically reversible that correlates with the expression of intracellular m6A demethylases and methyltransferases. Methyltransferases (writers) methylate the sixth N atom of adenine (A) on RNA, and demethylases (erasers) demethylate the m6A-modified RNA. Methylation recognition proteins (readers) recognise m6A modifications of RNA and modulate a range of downstream biological processes: RNA degradation, nucleation, translation, and cleavage (Figure 1).

Figure 1. The regulatory process and function of m6A modification. m6A is installed in the RRACH motif of the RNA (R represents the A and G bases, A represents the A base to be modified by methylation, C represents the C base and H represents the A, C, or U base) by m6A methyltransferases (Writers), which consists of METTL3/14, WTAP, METTL16, ZCCHC4, and VIRMA. m6A is removed by RNA demethylases (Erasers) FTO and ALKBH5. The m6A reader proteins, which specifically recognise and bind RNAs with m6A modifications, mediate a variety of processes related to RNA metabolic homeostasis, such as RNA alternative splicing, export, translation, and degradation.

Figure 2. m6A modifications regulate immune homeostasis. A. m6A increases the translation of co-stimulatory molecules such as CD40 and CD80 to promote the maturation of DC; On the other hand, m6A promotes the translation of lysosomal cathepsins, enhances the degradation of foreign antigens, and weakens the antigen presentation function of DCs. B. The m6A modification of IRAK3 promotes its self-degradation, abolishes the inhibitory effect of IRAK3 on the TLR signalling pathway, and facilitates the differentiation of Monocyte-macrophages into M1. C. ① m6A promotes the degradation of SOCS family genes, prevents its inhibitory effect on IL-7-regulated STAT5 activation, and eventually induces T cell differentiation into CD4 + T helper cell; ② The functional maintenance of CD4+ regulatory T cells requires the m6A-mediated decay of SOCS family genes to activate the IL-2-STAT5 signalling pathway. ③ The VHL-HIF-1α axis inhibits the promoting effect of GAPDH on METTL3 which can facilitate follicular helper T cell differentiation by increasing the degradation of ICOS (a key molecule in Tfh cell development). D. m6A increases the degradation of genes that negatively regulate the cell cycle, thus promoting the pro-B cell transition.

2.1. Evolution and distribution of RNA m6A

The m6A modification of mRNA was initially discovered in mouse L cells and Novikoff hepatoma cells in 1974 [9,10]. FTO, an obesity-related protein, was the firstly identified m6A-linked protein that demethylases RNA [11]. High-throughput sequencing technology subsequently facilitated the investigation of m6A, the modification of which was shown to be involved in the regulation of gene expression. With the first m6A sequencing paper published by Rechavi and colleagues in 2012, a series of enzymes and proteins involved in the modification of m6A were identified and functionally defined [12].

mRNA m6A is sequence-specific distributed with a conserved RRACH motif (R represents A or G, and H represents A, C or U). Results comparing different human tissues via MeRIP-seq and m6A-seq techniques revealed that the proportion of RRACH motifs and distribution of m6A varies greatly between tissues, suggesting the potential involvement of m6A in tissue development [13]. m6A modifications are preferentially located near the 3′ untranslated regions (3’UTR), stop codons, protein-coding regions (CDS), and long exons of RNA, indicating their diverse biological functions [8,12]. Evidence suggested that the m6A sites near terminator stop codons mainly regulate genes that maintain essential cellular functions, such as chromatin organisation, and cellular catabolic processes, whereas m6A enriched in other regions of genes are more likely to be associated with tissue-specific functions [13,14].

2.2. m6A writer\eraser\reader

2.2.1. m6A writer

Multiple m6A writers, mainly including methyltransferase-like 3 (METTL3), methyltransferase-like 14 (METTL14), and Wilms tumour 1-associating protein (WTAP), form a methyltransferase complex that recognises and binds to the RRACH motif of the RNA, resulting in an m6A modification. METTL3 contains the highly conserved binding site of S-adenosylmethionine (SAM) and DPPW domain with catalytic function to specifically transfer -CH3 provided by S-adenosylmethionine to adenine of mRNA sequence - a consensus sequence of either AAC or GAC (middle A is methylated) [15]. METTL14 is a homolog of METTL3 that has no methyltransferase activity but shares 43% identity with METTL3 [15,16]. However, when METTL3 and METTL14 form complex with a stoichiometry of 1:1, the activity of methyltransferase is significantly enhanced, with METTL14 structurally supporting the catalysis of METTL3 and assisting with the substrate recognition [15–17]. WTAP, initially found as the pre-mRNA splicing regulator, was proved to affect the process of RNA methylation in 2014 [16,18]. WTAP participates by specifically binding to RNA and recruiting METTL3 and METTL14 to achieve catalytic effects [19,20].

In addition to the aforementioned conventional m6A methyltransferase, some other subordinate methyltransferases were recently discovered to also play an exceptional role in m6A modification. Vir-like m6A Methyltransferase Associated (VIRMA, also known as KIAA1429), another subunit component of the m6A methyltransferase complex, was found to recruit the core component METTL3/METTL14/WTAP for regioselective methylation near the 3’UTR and the stop codon [21]. Methyltransferase-like 16 (METTL16) methylates U6 snRNA and methionine adenosyltransferase 2 A (MAT2A) by targeting UACAGAGAA sequences [22]. The C-terminal region of zinc finger CCCH domain-containing protein 13 (ZC3H13) interacts with WTAP-Virilizer-Hakai and mediates its nuclear localisation to facilitate the onset of m6A modifications [23]. Another zinc finger protein, zinc finger CCHC-type containing 4 (ZCCHC4), methylates the AAC motif of human 28SrRNA [24]. RNA binding motif protein 15 (RBM15) and its paralogue RBM15B recruits the methyltransferase complex to the long non-coding RNA X-inactive specific transcript (XIST), leading to the formation of its m6A modification [25]. There are subunits like Cbl proto-oncogene like 1 (CBLL1, also known as Hakai) that also interact tightly with the core components [26,27], the specific mechanisms of which remain to be further investigated.

2.2.2. m6A eraser

m6A erasers remove m6A modifications from RNA, also known as demethylases. The only two erasers identified to date are fat mass and obesity-associated protein (FTO) and alkB homolog 5 (ALKBH5), which both belong to the Fe(ii) and α-ketoglutarate-dependent AlkB family. Early studies identified an important role for FTO in the regulation of energy homeostasis and obesity [28,29]. The demethylation function of FTO controls exon splicing of the adipogenic regulator RUNX1 partner transcriptional co-repressor 1 (RUNX1T1), thereby regulating adipogenesis [30]. FTO demethylates N1-methyladenosine (m1A) in tRNA as well as internal m6A, and cap N6, 2-O-dimethyladenosine (m6Am) in mRNA and snRNA [31]. The mechanism of FTO demethylation is to catalyse the oxidative demethylation of m6A to N6-hydroxymethyladenosine (hm6A), N6-formyladenosine (f6A) and A stepwise [32]. In contrast, ALKBH5 interacts with intranuclear RNA directly to remove the m6A modification without intermediate steps [33]. Further discussion about m6A erasers can be found in the subsequent sections.

2.2.3. m6A reader

The m6A readers are RNA-binding proteins that specifically recognise and bind RNAs with m6A modifications, thereby mediating a variety of processes related to RNA metabolic homeostasis. The binding of readers to targeted RNA can be subsequently facilitated by m6A modifications which alter the structure of local RNAs and raise the chromatin accessibility to reveal RNA binding motifs [34,35]. A family of proteins whose structures contain the YT521-B YTH domain, the intracytoplasmic proteins YTH N6-methyladenosine RNA binding protein 1 (YTHDF1), YTHDF2, YTHDF3, and the intranuclear proteins YTH domain containing 1 (YTHDC1) and YTHDC2, form a hydrophobic pocket to recognise the m6A of GA-containing RNA [36]. Heterogeneous nuclear ribonucleoprotein (HNRNP, including HNRNPC/HNRNPG/HNRNPA2B1), the insulin-like growth factor 2 mRNA-binding proteins (IGF2BPs, including IGF2BP1/2/3), and proline rich coiled-coil 2 A (Prrc2a) were also identified lately to be involved in biological regulation as readers. Although intensive studies of RNA-binding proteins have led to deep understanding of m6A biological functions, a series of proteins that recognise and bind mRNA methylation sites with unknown functions still remain to be sequentially explored.

YTHDF1-3 share the same YTH domain and coordinate with each other during the regulation of gene expression. YTHDF3 promotes protein synthesis in synergy with YTHDF1, and affects methylated mRNA decay mediated through YHTDF2 [37]. The liquid-liquid phase separation of YTHDF proteins can be facilitated by m6A-modified mRNAs [38]. The translational efficiency of m6A-modified mRNA depends on the sequence context and the binding site of other YTHDF proteins [39]. YTHDC1 mainly works in the nucleus, while YTHDC2 predominantly works in the cytoplasm. Whether they have interaction with each other remains to be determined.

3. Biological function (Table 1)

3.1. Effects of m6A on mRNA

Writers install m6A on mRNA, erasers remove m6A, and readers recognise mRNA with m6A modifications, all at the post-transcriptional level. m6A modification regulates mRNA position, stability, translocation, shearing, and translation, which allows for a more rapid response to stimuli to regulate protein production and execute localised control than at the transcriptional level.

Table 1. Biological functions of m6A-related proteins.

Categories	Gene name	Location	Function	Preferred binding site	Refs.
Writer	METTL3	Nucleus; polysomes	Modulates m6A modification of RNA; Promotes RNA translation	AAC or GAC; GGAC motif	[15, 40]
	METTL14	Nucleus	Provides a binding scaffold for RNA and enhances the catalytic activity of METTL3		[15]
	METTL16	Nucleus	Installs m6A to U6 snRNA and MAT2A; Promotes or inhibits mRNA shearing	UAC(m6A)GAGAA	[22, 41]
	WTAP	Nucleus	Recruits METTL3 and METTL14 to the RNA locus		[18]
	VIRMA	Nucleus	Recruits the methyltransferase complex to the RNA 3′UTR and the stop codon		[21]
	RBM15	Nucleus	Recruits methyltransferase complexes to XIST		[25]
	ZC3H13	Nucleus	Mediates the nuclear localisation of the WTAP-Virilizer-Hakai complex		[23]
	ZCCHC4	Nucleus	Modulates m6A modification of 28S rRNA	AAC motif	[24]
Eraser	FTO	Nucleus	Demethylates m6A of DNA and RNA		[32]
	ALKBH5	Nucleus	Demethylates m6A of RNA; Exports nuclear RNA		[33]
Reader	YTHDF1	Cytoplasm	Promotes mRNA translation	GGAC motif	[42]
	YTHDF2	Cytoplasm	Promotes mRNA degradation	GAC motif	[43]
	YTHDF3	Cytoplasm	Promotes mRNA translation	GGAC motif	[44]
	YTHDC1	Nucleus	Help with RNA splicing and nuclear export; Regulates X chromosome silencing	GGAC motif	[25, 45, 46]
	YTHDC2	Cytoplasm	Promotes mRNA translation	GGACU motif	[47]
	HNRNPA2B1	Nucleus	Regulates selective splicing of mRNA and processing of precursor micro RNA	The RGAC Motif (R represents a purine)	[48]
	HNRNPC	Nucleus	Promotes mRNA stabilisation and translation; nuclear RNA export	Poly(U) RNA	[49, 50]
	HNRNPG	Nucleus	Facilitates RNA selective shearing	AGGAC motif	[34]
	IGF2BPs [IGF2BP1/2/3]	Nucleus; Cytoplasm	Promotes mRNA stabilisation and translation	GGAC motif	[51]
	Prrc2a	Cytoplasm	Regulates mRNA stability	GGACU motif	[52]
	FMR1	Cytoplasm	Negatively regulates mRNA translation	GGACU motif	[53, 54]
	eIF3	Cytoplasm	Promotes mRNA translation	GAC motif in 5′UTR	[55]

3.1.1. mRNA stabilisation and shearing

The density of m6A on mRNA is related to the half-life of RNA. The amino-terminal domain of YTHDF2 binds to methylated mRNA to reduce its stability and promote its degradation [43,56]. In contrast, PRRC2A enhances RNA stabilisation by competitively binding to the YTHDF2-associated RNA binding site [52]. IGF2BPs act as m6A binding proteins to protect mRNA from stress-mediated degradation and reinforce its translation [51].

Intracellular METTL16 is essential in maintaining SAM homeostasis through a negative feedback regulatory mechanism involving RNA shearing [41]. In the absence of SAM, METTL16 targets the MAT2A 3’UTR and mediates the shearing of MAT2A to regulate SAM synthetase gene expression. However, when intracellular SAM is plentiful, the m6A occupancy of SAM synthetase by the worm m6A writer METT-10, a homolog of METTL16, results in the inhibition of its normal RNA shearing and subsequently reduction of SAM abundance [22]. YTHDC1 collaborates with the pre-mRNA splicing factor serine and arginine-rich splicing factor 3 (SRSF3, also known as SRp20) to regulate the shearing process of mRNAs that contain exon inclusion [45]. HNRNPG, which selectively shears RNA to regulate gene expression, has also been reported recently [34,57].

3.1.2. Translational control of mRNA

In addition to acting as a Writer, METTL3 promotes epidermal growth factor (EGFR) and the Hippo pathway effector TAZ translation by recruiting eukaryotic initiation factor 3 (eIF3) to the transcription initiation complex in human lung cancer cells [40]. In contrast to the RNA degradation-promotive function of YTHDF2, YTHDF1 recruits ribosomes to accelerate RNA translation and protein production [42,58]. Activities of YTHDF1 and YTHDF3 have been reported to synergistically improve translation efficiency [44]. YTHDC2 specifically recognises and binds to the GGACU motif of the target RNA, which improves translation efficiency and reduces its RNA abundance [47]. Fragile X mental retardation protein (FMR1), a sequence-context-dependent reader, binds to ribosomes and inhibits mRNA translation [53,54]. hnRNPC specifically binds to the 29 nt sequence in the 3’UTR region of amyloid beta precursor protein (APP) mRNA, thereby stabilising APP mRNA and facilitating its translation [49]. Multiple intracellular stresses can increase m6A modification of the 5’UTR, the binding of which to eIF3 can induce translation of intracellular mRNA in a cap-independent manner [55].

3.1.3. Transport of mRNA

Other than splicing, YTHDC1 also acts as an mRNA nuclear transporter to facilitate the binding of methylated mRNA to dephosphorylated SRSF3, which then exits from the nucleus to the cytoplasm with the help of nuclear mRNA export receptor nuclear RNA export factor 1 (NXF1) [46]. The deficiency of ALKBH5 causes a dramatic accumulation of cytoplasmic RNA due to faster nuclear RNA export [33]. The heterotetramer of hnRNPC1/C2 selectively binds over 200 to 300 nucleotide regions of mRNA and mediates their nuclear export [50].

3.2. Effects of m6A on non-coding RNAs (lncRNA/microRNA/circRNA)

YTHDC1 recognises m6A modifications of XIST and promotes XIST-mediated X chromosome silencing [25]. Knockdown of the FTO gene in HEK293 cells results in increased expression of intracellular miRNAs with elevated m6A modification, as revealed by RNA-seq, suggesting that m6A may potentially regulate the metabolic processes of miRNAs [59]. The binding of HNRNPA2B1 to the m6A tag of the nascent microRNA promotes miRNA maturation by facilitating the selectable shearing event of pri-miRNA via the microprocessor protein DiGeorge syndrome critical region gene 8 (DGCR8) [48,60]. YTHDF3, together with the translation initiation factor eukaryotic translation initiation factor 4 gamma 2 (Eif4g2), recognises circRNAs that are rich in m6A modifications to facilitate their translation in response to external environment stresses [61].

4. m6A functions in growth, development, and immune homeostasis

4.1. Growth and development

Wang et al. found that knocking down of Mettl3 and Mettl14 in mouse embryonic stem cells (mESC) significantly reduces the level of m6A modification and impairs the self-renewing ability of cells [62]. Batista et al. reported that knockdown of METTL3 in mESC and hESC blocks cell differentiation [63]. A similar study by Geula et al. demonstrated that ablation of METTL3 in naive embryonic stem cells dramatically diminishes their pluripotency, leading to early embryonic death [64]. Deficiency of METTL3 in mouse epidermal progenitor cells results in a marked defect in hair follicle morphogenesis during skin development [65]. m6A maintains the balance between haematopoietic and endothelial cells by regulating the expression of notch receptor 1 (NOTCH1) gene [66]. Together, m6A methylation modifications participates indispensably in the regulation of cell growth, differentiation, and pluripotency.

m6A also plays a role in fertility. METTL3 and METTL14 work together to regulate sperm formation [67]. YTHDC2 affects early spermatogenesis and promotes sperm cell development, high expression of which promotes maturation of oocytes in meiosis of mice [47,68]. Deficiency of ALKBH5 leads to apoptosis of spermatocytes and abnormal spermatogenesis in mice, thereby disrupting reproductive function [33].

m6A modifications rise significantly during postnatal brain development, suggesting an important impact of m6A modifications on neurodevelopment [8,12]. Prrc2a is a novel m6A reader in neuronal cells that controls oligodendrocyte specification and myelination by stabilising Olig2 mRNA [52]. FTO is highly expressed in mouse neuronal cells and adult neural stem cells (aNSC), where it regulates postnatal growth, promotes aNSC proliferation and differentiation, and affects learning and memory capacity [69–71]. Deletion of either METTL3 or METTL14 causes severe defects in brain cortical development, and METTL3 deficiency alone also causes apoptosis of cerebellar granulosa cells (CGCs), leading to cerebellar hypoplasia [72,73]. Besides, ALKBH5 regulates postnatal maturation of the mouse cerebellum by maintaining the expression of differential cell fate determination genes [74].

Recent research demonstrated that the m6A modification of RNAs is crucial for haematopoietic development. Haematopoietic stem/progenitor cells (HSPCs) are derived from haemogenic endothelium during embryogenesis via the endothelial-to-haematopoietic transition (EHT) [75]. The continuous activation of Notch signalling could repress the emergence of HSPCs by blocking EHT in METTL3-deficient embryos [75,76]. METTL3 inhibits the endothelial Notch signalling pathway via YTHDF2-mediated mRNA decay [76]. Consistently, YTHDF2 was identified to repress inflammatory pathways in HSCs, which act as a protective factor in long-term HSC maintenance [77]. Gao et al. also found that conditional deletion of METTL3 in murine foetal liver activates the aberrant innate immune response, leading to haematopoietic failure and perinatal lethality [78]. These findings suggest that m6A maintains the balance of gene expression in endothelial and haematopoietic cells during EHT. In addition, shRNA-mediated depletion of METTL3 promotes cell differentiation and reduces cell proliferation in human haematopoietic stem/progenitor cells [79]. However, another study showed that deletion of YTHDF2 prevents the mRNA degradation of Wnt-target genes and survival-related genes, thus promoting the expansion and regenerative capacity of HSCs under stress conditions [80]. Taken together, the discoveries mentioned above suggest the complexity of m6A RNA methylation in the different developmental stages of haematopoiesis.

4.2. Immune homeostasis (Figure 2)

m6A modification and related proteins are involved in the activation, differentiation, and function of immune cells, which are vital for the pathogen recognition and cytokine secretion of immune system. Dendritic cells (DCs) initiate the immune response by processing antigens and activating T cells via pathogen-specific MHC-peptide complexes and co-stimulatory molecules (CD40, OX40L, 4-IBBL, etc.). Recent studies have reported that m6A modifications are required for the maturation and antigen presentation of DCs. Wang et al. found that knockout of METTL3 in DCs not only reduces the expression of m6A modified MHC class II (I-Ab), TLR signalling adaptor Tirap, and co-stimulatory molecules, but also blocks the Toll-like receptor 4/Nuclear factor kappa B subunit 1 (TLR4/NF-κB) signalling pathway via the restriction of inflammatory factors interleukin 6 (IL-6) and IL-12 expression, which ultimately leads to aberrant antigen presentation of DCs and activation of T cells [81]. On the other hand, Han et al. reported that deletion of YTHDF1 in DCs markedly increases the cross-presentation of tumour antigens and anti-tumour responses of CD8 T-cell in vivo by inhibiting the translation of lysosomal cathepsins, suggesting that the variation in the functional regulation of DCs by m6A is closely tied to contextual differences [82].

Monocyte-macrophages are characterised by diversity and plasticity, which can differentiate towards two polarised activation states, M1 or M2, when exposed to preferable stimulations [83]. The TLR4 signalling pathway is known to play a key role in the polarisation of macrophages towards proinflammatory M1 [84]. A recent study illustrated that METTL3 mediates RNA degradation of IL-1 receptor-associated kinase 3 (IRAK3), an inhibitor of TLR signalling pathway, by m6A modification, to regulate the process of macrophage polarisation [85]. FTO silencing inhibits the NF-kappa B signalling pathway and decreases the mRNA stability of STAT1 and PPAR-gamma via YTHDF2, preventing macrophage activation [86]. A following study by Dong et al. found that knockdown of METTL14 in tumour-associated macrophages results in the inhibition of CD8 + T-cell activation and impairment of CD8 + T-cell killing performance [87]. Taking together, these studies revealed that m6A modification regulates the polarisation and immunomodulatory function of macrophages.

m6A-induced mRNA degradation of suppressor of cytokine signalling (SOCS) family genes suppresses T cell proliferation and differentiation by preventing the counteract of IL-7-regulated signal transducer and activator of transcription 5 (STAT5) activation [88]. Similarly, CD4+ regulatory T cells (Tregs) rely on m6A to maintain their inhibition function via SOCS gene targeted IL-2-STAT5 signalling pathway [89]. Additionally, the m6A-dependent mRNA degradation inhibits the differentiation of follicular helper T (Tfh) cells through von Hippel-Lindau (VHL)-HIP-1α axis [90]. By increasing the abundance of m6A RNA modifications in thymocytes, the deletion of ALKBH5 increased the expression of the Notch signalling molecules Jagged1 and Notch2. In contrast, the impairment of Notch signalling promoted the expansion of γδ T cells, which are abundant in the mucosa and improve defence against gastrointestinal infections [91]. Moreover, in neuroinflammation, ALKBH5 deficiency reduces the expression of interferon-γ (IFN-γ) and C-X-C motif chemokine ligand 2 (CXCL2) by increasing their mRNA m6A modifications, resulting in diminished CD4+ T cell responses and reduced recruitment of neutrophils [92].

m6A is also involved in regulating two major transitions during B cell development. mRNA decay mediated by YTHDF2 promotes the transition of B cell from pro-B to large-pre-B. The large-pre-B-to-small-pre-B transition, instead, relies on the activation of key transcription factors mediated by METTL14 [93].

The majority of innate lymphoid cells that mediate antiviral and anti-tumour immunity are natural killer (NK) cells. IL-15 is a critical cytokine that regulates NK cell survival, proliferation, and effector function [94]. METTL3-mediated m6A methylation controls NK cells’ receptivity to IL-15 via safeguarding signalling pathways downstream of IL-15R in the tumour microenvironment [95]. STAT5-YTHDF2 forms a positive feedback loop downstream of IL-15 in NK cells to regulate NK cell survival and effector functions [96]. These observations provide insights into the positive regulatory effects of m6A methylation on homeostasis, anti-tumour, and anti-viral activities of NK cells.

5. RNA m6A in skin diseases (Table 2)

5.1. Arsenite-induced skin diseases

Arsenite, a toxic substance that is widespread in nature, was used to treat syphilis and psoriasis in the late 1940s, before safer and more effective treatments were available. Emerging work showed that low doses of arsenite have a protective effect in promoting cell growth and proliferation, while higher concentration causes significantly adverse effects to the skin, including hyperpigmentation, keratosis pilaris, Bowen disease, squamous cell carcinoma, and basal cell carcinoma [109]. To probe the role of m6A in the biphasic dose-dependent effects of arsenite, researchers treated human Hacat cells and A549 cell with different concentrations of arsenite. In cells treated with low dose of arsenite, the expression of methyltransferase complex was increased and the espression of demethylases was decreased, resulting in mild oxidative stress and markedly accelerated cell proliferation due to the increased m6A modification of target genes. When exposed to high concentrations, however, cells have reduced m6A modification accompanied by reactive oxygen species (ROS) overload to induce apoptosis [110–112]. Previous studies revealed that arsenite is capable of inducing malignant transformation of susceptible cells, in which the involved mechanism is most likely the regulation of the p53 signalling pathway by m6A under ROS stress [12,100,113]. Knockdown of METTL3 in arsenite-induced transformed cells reverses the malignant phenotype attributed to the regulation of cell proliferation and apoptosis pathways by m6A-mediated miRNAs, suggesting that miRNAs may also be important mediators linking m6A to arsenite toxicity [114]. Moreover, the demethylase FTO is effective against the neurological disorders caused by arsenite-induced dopamine dysregulation, further indicating the critical role of m6A regulation in arsenite-related disease [115].

Table 2. m6A in Skin diseases.

Skin disease types	Tissues or cells	regulators	expression	Role	Underlying mechanisms	Refs.
Systemic Lupus Erythematosus	Peripheral blood cells	YTHDF2 and ALKBH5	down	Risk factors	Enhances the inflammatory response	[97,98]
Psoriasis	Psoriasis vulgaris skin	m6A	up	Risk factors	Promotes inflammatory response by regulating IL-17A and TNF-α expression	[99]
Arsenite-induced skin diseases	Hacat	METTL3, METTL14, WTAP; FTO	up; down	Oncogene	Regulate p53 pathway mediated cell proliferation	[100]
Melanoma	Melanocytes	METTL3	up	Oncogene	Promotes tumour colony formation and invasiveness by promoting metastasis-related MMP2 and N-cadherin accumulation	[101]
Primary Acral Melanoma	Melanocytes	METTL3	up	Oncogene	Promotes the progression of melanoma by mediating the methylation of TXNDC5	[102]
Melanoma	Melanocytes	FTO	up	Oncogene	Accelerates tumour growth by promoting the expression of pro-tumorigenic genes PD-1, CXCR4 and SOX10	[103]
Ocular melanoma	Melanocytes	YTHDF1	down	Suppressor	Inhibits tumour growth by promoting translation of the oncogenic factor HINT2	[104]
Ocular melanoma	Ocular melanoma stem-like cells and ocular melanoma tissues	YTHDF3	up	Oncogene	Enhances tumorigenicity of cancer stem cells in ocular melanoma by promoting CTNNB1 translation	[105]
Cutaneous squamous cell carcinoma	cSCC cells	METTL3	up	Oncogene	Promotes tumour proliferation and self-renewal by regulating the expression of DNp63	[106]
Merkel cell carcinoma	MCC cells	YTHDF1	up	Oncogene	Enhances tumour proliferation by promoting the expression of translation initiation factor eIF3	[107]
Kaposi’s sarcoma	Virus-infected cells	YTHDC1	up	Oncogene	Promotes KSHV transcript shearing	[108]

5.2. Systemic lupus erythematosus

Systemic lupus erythematosus (SLE) is a multi-system autoimmune disease characterised by abnormal activation of polyclonal B cells, production of autoantibodies, and deposition of immune complexes. Genetic variation, epigenetic pathways, gender, and environmental factors, corporately contribute to the pathogenesis of SLE [116]. Although epigenetics-based diagnosis and treatment have achieved considerable progress in SLE, uncovering the underlying mechanisms and addressing effective therapeutic interventions remain a huge challenge [117–120]. Li et al. primarily illustrated the potential link between m6A modifications and SLE, pointing out the way for more in-depth studies [121]. mRNA expression of METTL3, MTEEL14, WTAP, ALKBH5, FTO, and YTHDF2 in peripheral blood was subsequently reported to be significantly lower in SLE patients compared to normal controls, with ALKBH5 being negatively correlated with anti-dsDNA levels and closely associated with clinical symptoms. Given that anti-dsDNA is one of the most important markers to evaluate the SLE activity, ALKBH5 may be a potential marker to predict prognosis of SLE [97,98]. The methyltransferase complex is stabilised via the S43/S50/S525 phosphorylation of METTL3 and S306/S341 of WTAP by ERK, followed by the deubiquitination by USP5 [122]. Lactate accumulation induces METTL3 upregulation via H3K18 lactylation [123]. The binding of WTAPP1 to WTAP with the help of CCHC-type zinc finger nucleic-acid binding protein (CNBP) regulates the translation of WTAP [124]. WTAP can also be controlled by the mTORC1/S6K axis [125] and PIWI-interacting RNAs (piRNAs) [126]. The expression of ALKBH5 can be regulated by DNA methylation at the promoter [127] and chromatin accessibility by histone modification [127]. While the expression of FTO is regulated by miR-96/AMPKα2 in colorectal cancer [128] and by WNT/β-catenin in lung adenocarcinoma [129]. The expression of YTHDF2 is subjected to histone lactylation in ocular melanoma [130] and EGFR/SRC/ERK pathway in Glioblastoma [131]. In summary, these mechanisms could potentially contribute to the reduced expression of m6A proteins in SLE, which need to be further elucidated.

Zhao et al. explored potential integrative models of m6A immunity in SLE through bioinformatics analysis. They found that insulin-like growth factor binding protein 3 (IGFBP3) may contribute to the diagnosis and treatment of SLE [132]. Studies referred above collectively illustrate the distinct correlation between RNA m6A and SLE, but the exact mechanism of m6A in the pathogenesis of SLE remains elusive that warrants future in-depth investigation.

5.3. Psoriasis

Psoriasis, a chronic and recurrent inflammatory skin disease that has a prevalence of 2.8% and predominately occurred in adolescents, is characterised by excessive proliferation and abnormal activation of keratin-forming cells, as well as overwhelming infiltration of immune cells [133,134]. Current research on the epigenetic mechanisms of psoriasis vulgaris-the most prevalent type of psoriasis-is mainly focussed on DNA methylation, histone modifications, and non-coding RNA, with RNA modifications being rarely reported [135]. Combined results from MeRIP-seq and RNA-seq of skin Psoriasis Vulgaris revealed that genes with higher m6A modifications tend to have more mRNA expression. Among these genes, IL-17A and tumour necrosis factor-alpha (TNF-α) from TNF-α/IL-23/Th17 axis, which have significantly higher m6A modifications in psoriasis vulgaris relative to normal skin tissue, have been shown to promote the proliferation of keratinocytes to form inflammatory lesions. This study demonstrated that regulation of IL-17A and TNF-α expression is the potential mechanism of m6A to promote the formation of inflammatory lesions in Psoriasis Vulgaris [99]. Furthermore, Arf GAP with GTP-binding protein-like domain -AS1 (AGAP2-AS1) in skin tissues of Psoriasis patients was suggested to promote keratinocyte proliferation via the miR-424-5p/AKT/mTOR axis, while Xian et al. found that METTL3-mediated m6A modification repressed AGAP2-AS1 expression through YTHDF2-dependent regulation of AGAP2-AS1 RNA stability [136].

5.4. Skin cancers

Accumulating evidence suggested the intense association between post-transcriptional m6A modification of RNA and skin cancer aetiology. Melanoma, a highly malignant tumour of melanocytic origin that often occurs in the skin, is predicted to become the second leading cancer with high mortality rate by 2040 [137]. Early diagnosis and intervention of melanoma are extremely important due to the unavailable of a specific cure other than complete surgical resection. Consistent with what was elucidated in acute myeloid leukaemia (AML) and lung cancer, METTL3 was found to promote colony formation of melanoma cells in 2019 [40,79,101]. In addition, METTL3 also enhances the invasion and metastasis of melanoma by regulating the accumulation of matrix metallopeptidase 2 (MMP2) and N-cadherin independent of its m6A catalytic activity [101], or by regulating specific m6A-methylated transcripts like thioredoxin domain-containing protein 5 (TXNDC5) [102]. Another study showed that METTL3 and eukaryotic translation initiation factor 3 subunit h (eIF3h) interact to form a closed-loop complex that contributes to the translation enhancement of a large set of oncogenes [138]. Moreover, eIF3h has recently been reported to be highly expressed in melanoma and positively correlated with the proliferation and migration of melanoma [139]. FTO, which facilitates the expression of tumorigenic genes programmed cell death protein 1 (PD-1), C-X-C motif chemokine receptor 4 (CXCR4), and SRY-box transcription factor 10 (SOX10) by reducing the mRNA m6A level, contributes to melanoma tumorigenesis and anti-PD-1 treatment resistance [103]. FTO inhibitors are thereby suggested to potentially improve the anti-tumour effect of anti-PD-1 therapy. YTHDF3 facilitates the protein expression of catenin beta 1 (CTNNB1) to promote the propagation and migration of ocular melanoma via m6A methylation [105]. Conversely, YTHDF1 inhibits the growth of ocular melanoma as a result of its ability to promote translation of the tumour suppressor histidine triad nucleotide-binding protein 2 (HINT2) [104]. Recently, Lin et al. performed copy number variation (CNV) and single nucleotide polymorphism (SNP) genomic variant analysis of m6A-related genes in combination with the data of their expression in cutaneous melanoma (CM), revealing that m6A-regulated genes have high genomic mutagenicity and that patients with higher frequency of mutations in m6A-regulated genes have a worse prognosis [140]. Meanwhile, a subgroup of m6A-correlated transcripts that contributes to better prognosis were also identified, which motivates the efforts to seek novel approaches for CM management [140].

Cutaneous squamous cell carcinoma (cSCC) is the second leading pervasive non-melanoma skin cancer worldwide characterised by the malignant proliferation of keratin-forming cells [141]. METTL3 is highly expressed in human cSCC cells, the depletion of which significantly inhibits the proliferation and self-renewal of tumour cells. The phenotype of cSCC cells was effectively restored when exogenous Delta Np63 (DNp63) cDNA was added after METTL3 knockdown, indicating that METTL3 exerts pro-tumorigenic effects via the regulation of DNp63 [106].

Merkel cell carcinoma (MCC) is a malignant Merkel cell polyomavirus (MCPyV)- associated tumour that is usually found on the exposed skin areas of elderly female patients [142]. A recent study identified a possible connection between m6A and MCC disease. m6A was detected in the transcripts of MCPyV, and YTHDF1 expression was elevated in MCC cells. The silence of YTHDF1 in MCC cells attenuates tumour cell proliferation by suppressing the expression of eIF3 [107].

Kaposi’s sarcoma (KS) is a multiple idiopathic haemorrhagic sarcoma of the skin, the key aetiological factor of which is generally considered to be the Kaposi’s sarcoma- associated herpesvirus (KSHV) [143]. The genome of KSHV in latency is silenced as a result of repressive epi-modifications. A large amount of m6A present in the KSHV genome on the pre-mRNA of viral replication transcription activator (RTA) has been reported to direct the transcriptomic control of herpesvirus replication. Two of the m6A sites that are located in exon 2 and intron near the shear site are recognised by YTHDC1, which subsequently recruits the shear factors SRSF3 and SRSF10 that are essentially required for the expression of RTA. The amount of pre-RTA mRNA is drastically reduced, and RTA protein expression is suppressed under circumstances where either the two m6A sites are mutated, or METTL3 is deleted. In contrast, the FTO ablation has the opposite effect. These observations suggest that m6A promotes KSHV replication and tumour development by facilitating the shearing and stabilisation of RTA pre-mRNA [108,144].

6. Small-molecule inhibitors of m6A regulators (Table 3)

Several small molecules have recently been identified to have biological functions during immune responses and tumorigenesis by targeting m6A modification proteins. Selberg and colleagues screened and identified compounds that activate m6A methylation by binding to the active site of the METTL3-METTL14 complex [145]. STM2457, N-substituted amide of ribofuranuronic acid analogues of adenosine, Cpd-564, and SAH display efficient inhibition effect against METTL3 in multiple diseases [146–151]. Small molecules discovered to suppress m6A modification predominantly disrupted FTO function by binding to the different structures of FTO protein [152–159]. ALK-04 inhibits the ALKBH5 function, the combination of which with GVAX/anti-PD-1 immunotherapy has a synergistic effect against melanoma [160]. BTYNB was found to selectively inhibit IGF2BP1 and disrupt its interaction with oncogenic c-Myc mRNA [161,162]. Although mainly being discussed in tumorous disorders, these small molecules and their targets could be further tested in skin diseases for therapeutic strategies.

Table 3. Small-molecule inhibitors of m6A regulators.

Compounds	Target	Functions	Refs.
Piperidine derivative and piperazine derivative compounds	METTL3-METTL14 complex	Promote mRNA m6A modification	[145]
STM2457	METTL3	Reduces the expression of SP1, BRD4, HOXA10, MYC, IFIT2, PTCH1, and GLI2	[146–148]
N-substituted amide of ribofuranuronic acid analogues of adenosine	METTL3	Reduces mRNA m6A modification	[149]
SAH	METTL3	Reduces the stability of Bcl2 mRNA	[150]
Cpd-564	METTL3	Reduces the m6A modification and stability of TAB3	[151]
FB23-2	FTO	Reduces the expression of c-Myc and CEBPA and activates the expression of RARA and ASB2	[152]
Dac51	FTO	Reduces the expression of transcription factors c-Jun, JunB, and C/EBPβ	[153]
Bisantrene	FTO	Reduces LILRB4 expression	[154]
R-2HG	FTO	Induces mRNA degradation of c-Myc and CEBPA	[155]
N-CDPCB and CHTB	FTO	Binds to FTO and increases RNA m6A modification	[156,157]
Meclofenamic acid	FTO	Reduces MYC, miR-155, and miR-23a expression, and increases p53 expression	[158,159]
ALK-04	ALKBH5	Suppresses Mct4 expression	[160]
BTYNB	IGF2BP1	Destabilizes E2F1, c-Myc, and eEF2 mRNA	[161,162]

7. Conclusion

In Summary, we systematically discussed the classification of m6A-regulating proteins and their biological function in physiological processes and skin pathology. m6A erasers, writers, and readers context-dependently and synergistically regulate the expression of critical transcription factors and cellular signalling pathways in dermatosis pathogenesis via an m6A-mediated RNA stabilising, shearing, transporting, and translation. Given the extensive implementation of epitranscriptomics studies, RNA m6A modification has been emerging as crucial post-transcriptional regulator of gene expression. Aberrant expression of m6A related proteins-writers, erasers, and readers-mediates the imbalance of multiple disease-associated targets and downstream signalling pathway by affecting mRNA stability and translation efficiency. The generalisability and content dependency of m6A determine that attentions should be paid to the clarification of quantitative m6A level at different sites and distinctive function of a given m6A modification in variable cells and disorders, including skin disease. The work of m6A in dermatology has been largely focussed on autoimmune disease and cancer. Whether m6A also participates in the pathogenesis of other skin diseases like infectious dermatosis, genodermatosis, and metabolic dermatosis, especially the respective underlying mechanisms, requires to be further elucidated. Besides the well-studied m6A modifiers addressed in this review, some newly identified regulators that have been barely discussed before, may have underestimated effects in cell homeostasis or pathological progress of dermatosis. m6A-based target identification and clinical interventions are now showing great potential in the management of dermatological diseases for both diagnostic and therapeutic purposes.

Abbreviations
m6A	=	N6-methyladenosine
mRNA	=	messenger RNA
tRNA	=	transfer RNA
rRNA	=	ribosomal RNA
lncRNA	=	long non-coding RNA
3’UTR	=	3’ untranslated regions
CDS	=	coding regions
METTL3	=	methyltransferase-like 3
METTL14	=	methyltransferase-like 14
WTAP	=	Wilms tumour 1-associating protein
SAM	=	S-adenosylmethionine
VIRMA	=	Vir-like m6A Methyltransferase Associated
METTL16	=	methyltransferase-like 16
MAT2A	=	methionine adenosyltransferase 2A
ZC3H13	=	zinc finger CCCH domain-containing protein 13
ZCCHC4	=	zinc finger CCHC-type containing 4
RBM15	=	RNA binding motif protein 15
XIST	=	X-inactive specific transcript
CBLL1	=	Cbl proto-oncogene like 1
FTO	=	fat mass and obesity-associated protein
ALKBH5	=	alkB homolog 5
RUNX1T1	=	RUNX1 partner transcriptional co-repressor 1
hm(6)A	=	N(6)-hydroxymethyladenosine
f(6)A	=	N(6)-formyladenosine
m1A	=	N1-methyladenosine
m6Am	=	N6, 2-O-dimethyladenosine
YTHDF1	=	YTH N6-methyladenosine RNA binding protein 1
YTHDC1	=	YTH domain containing 1
HNRNP	=	heterogeneous nuclear ribonucleoprotein
IGF2BPs	=	the insulin-like growth factor 2 mRNA-binding proteins
Prrc2a	=	proline rich coiled-coil 2A
SRSF3	=	serine and arginine rich splicing factor 3
EGFR	=	epidermal growth factor
eIF3	=	eukaryotic initiation factor 3
FMR1	=	fragile X mental retardation protein
APP	=	amyloid beta precursor protein
NXF1	=	nuclear RNA export factor 1
DGCR8	=	DiGeorge syndrome critical region gene 8
Eif4g2	=	eukaryotic translation initiation factor 4 gamma 2
mESC	=	mouse embryonic stem cells
NOTCH1	=	notch receptor 1
aNSC	=	adult neural stem cells
CGCs	=	cerebellar granulosa cells
HSPCs	=	Haematopoietic stem/progenitor cells
EHT	=	endothelial-to-haematopoietic transition
DCs	=	Dendritic cells
TLR4	=	The Toll-like Receptor 4
NF-κB	=	Nuclear factor kappa B subunit 1
IL-6	=	interleukin 6
IRAK3	=	IL-1 receptor-associated kinase 3
SOCS	=	suppressor of cytokine signalling
STAT5	=	signal transducer and activator of transcription 5
Tregs	=	CD4+ regulatory T cells
Tfh	=	follicular helper T
VHL	=	von Hippel-Lindau
IFN-γ	=	interferon-γ
CXCL2	=	C-X-C motif chemokine ligand 2
NK cells	=	natural killer cells
ROS	=	reactive oxygen species
SLE	=	Systemic lupus erythematosus
IGFBP3	=	insulin like growth factor binding protein 3
TNF-α	=	tumour necrosis factor alpha
AGAP2	=	Arf GAP with GTP-binding protein-like domain
AML	=	acute myeloid leukaemia
MMP2	=	matrix metallopeptidase 2
TXNDC5	=	thioredoxin domain-containing protein 5
eIF3h	=	eukaryotic translation initiation factor 3 subunit h
PD-1	=	programmed cell death protein 1
CXCR4	=	C-X-C motif chemokine receptor 4
SOX10	=	SRY-box transcription factor 10
CTNNB1	=	catenin beta 1
HINT2	=	histidine triad nucleotide-binding protein 2
CNV	=	copy number variation
SNP	=	single nucleotide polymorphism
CM	=	cutaneous melanoma
cSCC	=	Cutaneous squamous cell carcinoma
DNp63	=	Delta Np63
MCC	=	Merkel cell carcinoma
MCPyV	=	Merkel cell polyomavirus
KS	=	Kaposi’s sarcoma
KSHV	=	Kaposi’s sarcoma associated herpesvirus
RTA	=	replication transcription activato

Disclosure statement

No potential conflict of interest was reported by the authors.

Additional information

Funding

This work was supported by the National Natural Science Foundation of China (No. 82030097 and 81874243), CAMS Innovation Fund for Medical Sciences (CIFMS) (2019-I2M-5-033), the Key project for international and regional cooperation in science and technology innovation of Hunan province (2019WK2081), the Project for leading talents in science and technology in Hunan province (2019RS3003), the National Natural Science Foundation of China (No. 81874253), Excellent postdoctoral innovative talents of Hunan province in 2020 (No. 2020RC2014), Natural Science Foundation of Hunan Province China (No. 2021JJ40837).