m6A Readers, Writers, Erasers, and the m6A Epitranscriptome in Breast Cancer

Abstract

Epitranscriptomic modification of RNA regulates human development, health, and disease. The true diversity of the transcriptome in breast cancer including chemical modification of transcribed RNA (epitranscriptomics) is not well understood due to limitations of technology and bioinformatic analysis. N-6-methyladenosine (m6A) is the most abundant epitranscriptomic modification of mRNA and regulates splicing, stability, translation, and intracellular localization of transcripts depending on m6A association with reader RNA binding proteins. m6A methylation is catalyzed by the METTL3 complex and removed by the specific m6A demethylase ALKBH5, with FTO’s role as an ‘eraser’ uncertain. In this review, we provide and overview of epitranscriptomics related to mRNA and focus on m6A in mRNA and its detection. We summarize current knowledge on altered levels of writers, readers, and erasers of m6A and their roles in breast cancer and association with prognosis. We summarize studies identifying m6A peaks and sites in genes in breast cancer cells.

Introduction

Breast cancer is the most commonly diagnosed cancer worldwide, although countries with the highest human development index, related to life expectancy, education and wealth, have the highest incidence of breast cancer (BC) (Wilkinson and Gathani 2021). Breast tumors are characterized by cellular heterogeneity, abnormal gene expression, and chaotic signaling associated with genetic, epigenetic, and epitranscriptomic changes (Lüönd, et al. 2021; Yang, et al. 2021). Breast cancer is divided into four major clinical subtypes based on varying expression of hormone receptors (HR) estrogen receptor α (ERα, referred to as ER+) and progesterone receptor(PR) and human epidermal growth factor receptor (HER2 ERBB2-)) (Dai, et al. 2016) (Dunnwald, et al. 2007), or the lack of these receptors in triple negative breast cancer (TNBC). These molecular markers inform treatment options. Due to early diagnosis and improved treatment options, the death rate from BC has decreased significantly during the last 20 years (Siegel, et al. 2021; Siegel, et al. 2022; Sung, et al. 2021). Based on individual risk assessment, patients with HR+/HER2- primary tumors receive 5–10 years of single agent endocrine therapies (ET): an aromatase inhibitor (AI, e.g., letrozole) for postmenopausal women or tamoxifen (TAM) and ovarian suppression for premenopausal women (Burstein, et al. 2016; Burstein, et al. 2010; Burstein, et al. 2014). For metastatic ER+ BC, patients are treated with combined ET and CDK4/CDK6 inhibitors, PI3K inhibitors, PARP inhibitors, and or anti-PD-L1 immunotherapy, depending on molecular profiling (reviewed in (Loibl, et al. 2021)). Although the PAM50 risk of recurrence score (PAM50-ROS) is an example of ‘precision medicine’ and patient-customized-therapy’ with a predictive recurrence value in ET-treated ERα+ BC patients (Giorgi Rossi, et al. 2021), > 43,700 women in the U.S. will die of metastatic BC in 2022 (Siegel et al. 2022).

If a breast tumor expresses little (< 1%) or no ER/PR, but has a high expression of HER2, the patient is treated with the humanized monoclonal antibodies trastuzumab and pertuzumab directed against the extracellular domain of HER2 (ERB2) (von Minckwitz, et al. 2017) and Trastuzumab deruxtecan (T-DXd), an antibody-drug conjugate with a topoisomerase I inhibitor payload for metastatic of unresectable disease (Bardia, et al. 2022). Patients with TNBC receive chemotherapy +/− radiation therapy as first line therapy (Bianchini, et al. 2016). New FDA-approved therapeutic options used in combination with chemotherapy for patients with local recurrent unresectable or metastatic TNBC include PD-L1 inhibitors atezolizumab and pembrolizumab (Torres and Emens 2022), PARP inhibitors (olaparib or talazoparib, depending on germline BRCA gene mutations) (Cardoso, et al. 2020), and Sacituzumab govitecan (SG) that is antibody conjugate with anti-trophoblast cell surface antigen-2 (Trop-2), targeting Trop-2-expressing cancer cells, linked with cytotoxic moiety SN-38 (govitecan) with topoisomerase I inhibitor activity (Michaleas, et al. 2022). A recent pilot clinical trial using personalized immunotherapy, infiltrating lymphocytes(TILs), showed that six of 42 patients with metastatic BC developed objective cancer regression (Zacharakis, et al. 2022).

For ~ 25–40% of BC patients with ER+ breast tumors, ET-resistance develops and overall survival (OS) for patients with metastatic disease is only ~ 44.6 mos. (Polley, et al. 2021). Significant progress has been made in understanding the molecular mechanisms for ET resistance (reviewed in (Clarke, et al. 2015)) and includes mutations in ESR1 and ESR1 fusion transcripts in ~30% of metastatic tumors from patients on aromatase inhibitor ET therapy (reviewed in (Piscuoglio, et al. 2018; Toy, et al. 2017)). Increased progression-free survival (PFS) has been realized with the addition of cyclin-dependent kinase 4 & 6 (CDK4 & 6) inhibitors to ET for women with advanced BC (Cardoso et al. 2020; Sledge, et al. 2020). However, metastatic spread occurs early in tumor progression and likely precedes detection of the primary tumor (Celià-Terrassa and Kang 2016; Sanger, et al. 2011). Thus, identifying features in the primary tumor that reflect its metastatic potential and blocking progression is key to preventing recurrent disease. Whether epitranscriptomic RNA modifications in breast tumors may hold clues to metastatic potential and the mechanism(s) and targets that these modifications regulate are currently unknown. Analysis of BC data in ONCOMINE revealed changes in the transcript levels of writers, readers, and erasers of the most common RNA modification, N-6-methyladenosine (m6A) with two readers, RBM15B and ZC3H13, positively associated with ER and PR expression and with longer relapse-free survival (RFS) (Zhang, et al. 2020). However, there are conflicting data with respect to associations of these and other m6A writers, readers, and erasers in BC (Liu, et al. 2022a). There are few studies of how the position of m6A affects transcript splicing, stability, or translation in BC cells or tumors, indicting the need for closer examination of the importance of the regulation of the epitranscriptome in subtype specific BC and metastatic disease.

An Overview of Epitranscriptomics

Over 160 chemical modifications of transcribed RNA have been identified (Mathlin, et al. 2020). The proteins that add the chemical modification, recognize the specific RNA base modification, and remove that modification are termed ‘writers, readers, and erasers’. The term epitranscriptomics refers to these chemical modifications of transcribed RNA and includes modifications of adenosine (A), guanosine (G), cytidine (C), and uridine (U) (reviewed in (McCown, et al. 2020)). The Modomics database of RNA modification pathways summarizes these modifications (Boccaletto, et al. 2022). These modifications affect the structure of RNA and its interactions with proteins, including readers (Jones, et al. 2022; Kierzek, et al. 2022; Szabat, et al. 2022). Transfer RNAs (tRNAs) are the most extensively modified cellular RNAs with over 100 chemical modifications (Boccaletto and Bagiński 2021) and each tRNA containing 11–13 modifications (Delaunay and Frye 2019). The tRNA chemical modifications within the D-, T-, and anticodon- loops regulate tRNA stability, folding, and interaction with the Elongator complex for translational fidelity (Pereira, et al. 2018). 130 modifications have been identified in rRNA, including isomerization of uridine to pseudouridine (Ψ) and 2′-O-methylation of the ribose (Am) (Delaunay and Frye 2019). Mitochondrial-encoded rRNAs are also highly modified (Laptev, et al. 2020). There are thirteen known internal mRNA modifications located in the 5’ and 3’ UTRs as well as exons and introns (Anreiter, et al. 2021). m6A is the most common dynamic modification of the transcriptome and regulates the function and processing of mRNAs, long non-coding RNAs (lncRNAs), and primary microRNAs (pri-miRNAs) to precursor (pre)-miRNAs (Zaccara, et al. 2019). In mammalian mRNA, ~ 0.1%–0.4% of RNA are exposed to m6A modification, with an average of 3–5 m6A sites per transcript (Roundtree, et al. 2017a). m6A is also enriched in circRNAs, lncRNAs, and U6 snRNA (Zaccara et al. 2019). m6A is enriched in the 3’UTR and within exons, highlighting its role in regulating mRNA processing, alternative splicing, stability, and translation (Kretschmer, et al. 2018; Roundtree et al. 2017a). The ‘landscape’ of m6A modification is species- and cell-type specific (Liu, et al. 2019a; Wang and Wang 2020) and plays an important role in various cancers (reviewed in (Niu, et al. 2022))

m6A is dynamic and is added by a METTL3 ‘writer’ complex at DRACH (RRACH) (D or R = A/G/U, R = A/G, H = A/C/U) consensus motifs and removed by ‘erasers’ FTO and ALKBH5 (Knuckles and Bühler 2018). FTO and ALKBH5 are α-ketogluterate-dependent dioxygenases sensitive to D-2-hydroxygluterate-mediated inhibition connecting glucose metabolism to m6A (Su, et al. 2018). Demethylases FTO and ALKBH5 are the ‘erasers’ of m6A methylation (Zhao, et al. 2020), although since FTO is nuclear, it may only demethylate m6Am in the 5’ N-terminal cap of mRNA (Mauer, et al. 2019). The role of m6A depends on the proteins that recognize and ‘read’ this marker, including YTHDC1–3, YTHDF1–3, and HNRNPA2B1 (Lewis, et al. 2017; Licht and Jantsch 2016). An additional number of RNA binding proteins and transcription factors, including glucocorticoid receptor (GR), also bind m6A (An, et al. 2020).

The mechanisms regulating and roles of epitranscriptomic mRNA modifications other than m6A are largely unknown (Zhao et al. 2020). Common epitranscriptomic marks in mRNA are involved in translation efficiency, i.e., m6A, m6Am, m1A, m5C, Ψ, and hm5C, RNA structure and stability, export, and degradation (Table 1) (Kumar and Mohapatra 2021). The biological roles of the epitranscriptome in human biology and disease and the writers, readers, and erasers of these modifications is of great research interest. New strategies are being developed to address the technical challenges in identifying and understanding the biological importance of the epitranscriptome

Table 1:

Selected mRNA epitranscriptomics (Wiener and Schwartz 2021)

modification	Role in mRNA	Writer	Readers	Erasers	Role in breast cancer
Ψ (Pseudouridine)	enriched in alternatively spliced regions (Martinez, et al. 2022); enhanced mRNA translation (Hoernes, et al. 2019), role in mRNA not well-understood (Borchardt, et al. 2020)	PUS1, PUS3, PUS7, PUS7L, PUS10, DKC1, TRUB1,(Anreiter et al. 2021; Chen et al. 2021a; Chen, et al. 2019b; Pereira et al. 2018) Ψ is also converted from uridine through a post-transcriptional isomerization reaction (Adachi, et al. 2019).	Unknown for mRNA (Levi and Arava 2021)	Unknown for mRNA (Cerneckis, et al. 2022)	Role in mRNA modification is unknown (Xue, et al. 2022)
m5C (5-methylcytosine)	mRNA stability, translation, and export (Chen et al. 2021a; Chen, et al. 2021b; Yang, et al. 2017); enriched in CG-rich regions in the 3’UTRs of genes and immediately downstream of translation initiation sites and has conserved, tissue-specific and dynamic features across mammalian transcriptomes (Yang et al. 2017).	NSUN2 (Yang et al. 2017)	ALYREF/THOC4, the mammalian mRNA export adaptor (Yang et al. 2017)	TET2 (Chen et al. 2021b)	m5C writer NSUN2 was upregulated in breast cancer and promoted proliferation, migration, invasion and tumorigenicity (reviewed in (Li, et al. 2022c))
m1A (1-methyladenosine)	affects mRNA structure and interaction with RNA binding proteins (Dominissini, et al. 2016).	TRMT6 TRMT61A TRMT61B (mt) (Li, et al. 2016b; Pereira et al. 2018)	YTHDF1–3(Anreiter et al. 2021)	ALKBH1 (Liu, et al. 2016), ALKBH3 (Chen, et al. 2019c)	Demethylation of m1A promotes breast cancer invasion (reviewed in (Kumari, et al. 2021))
m6A (N6-Methyladenosine)	splicing, stability, translation, degradation (Dinescu, et al. 2019)	METTL3 METTL14 WTAP VIRMA RBM15 RBM15B ZC3H13 WTAP CBLL1 METTL16 (Balacco and Soller 2019; Licht and Jantsch 2016; Wei, et al. 2018; Wen, et al. 2018)	YTHDF1–3, YTHDC1–2, HNRNPA2B1, IGFBP1–3, PRRC2A, FMR1, ELAVL1, HNRNPC, HNRNPG (Alarcón et al. 2015a; Anreiter et al. 2021; Bi, et al. 2019; Chen, et al. 2019a; Yang, et al. 2018; Zhou, et al. 2019)	FTO, ALKBH5 (Licht and Jantsch 2016)	Writers, readers, and erasers have roles in breast cancer (Reviewed in (Deng, et al. 2018; Fang, et al. 2021))

METTL3 methyltransferase complex (MTC)

METTL3 is the ‘writer’ of the m6A modification in mRNA. The METTL3 complex (875 kDa) includes the METTL3-METTL14 heterodimeric complex and auxiliary proteins that enhance the m6A methyltransferase activity: WTAP (Wilms tumor suppressor-1–associated protein), CBLL1 (Cbl Proto-Oncogene Like 1, HAKAI), VIRMA (Vir Like m6A Methyltransferase Associated), RBM15 and RBM15B (RNA Binding Motif Protein 15), and ZC3H13 (Zinc Finger CCCH-Type Containing 13) (reviewed in (Covelo-Molares, et al. 2021; Zaccara et al. 2019)) (Table 2). Crystal structure studies revealed that METTL3 and METTL14 interact over a large surface area via their methyltransferase domains (MTD) to bind RNA (Zaccara et al. 2019). METTL3 catalyzes the m6A methylation by taking the methyl from the cofactor S-adenosylmethionine (SAM) and releasing S-adenosylhomocysteine (SAH) while METTL14 is catalytically inactive (Wang, et al. 2016; Zhou and Pan 2016). METTL14 stabilizes METTL3 and contributes to RRACH (R = A/G, H = A/U/C) site binding. METTL3/METTL14 was also shown to preferentially bind RNA G-quadruplex motifs (G4) that are enriched in human gene promoters, suggesting that G4s are involved in transcription regulation by the MTC (Yoshida, et al. 2022). The MTC is localized to gene promoters to deposit m6A on nascent transcripts, including enhancer RNAs (eRNAs) and promoter upstream transcripts (PROMPTs) (Xu, et al. 2022), and plays a role in RNA polymerase II pause release (Akhtar, et al. 2021).

Table 2:

METTL3 methyltransferase complex proteins and their role in breast cancer.

Protein (other names)	Function	Breast cancer activity
METTL3	m6A methyltransferase	Oncogene (Cai et al. 2018; Pan et al. 2021; Wang et al. 2020a) Tumor suppressor in TNBC (Shi et al. 2020)
METTL14	Heterodimer partner for METTL3 m6A methyltransferase	Oncogenic (Sun et al. 2020).
WTAP	Stabilizes the METTL3-METTL14 heterodimer to nuclear speckles and interacts with splicing factors (Horiuchi, et al. 2021)	Oncogene, but inhibits metastasis (Wang et al. 2022)
CBLL1 (Hakai)	E3 ubiquitin ligase, but its enzymatic activity is dispensable for its function in stabilizing the METTL3 complex by interacting with WTAP and VIRMA (Zhang, et al. 2022)	Tumor suppressor and low expression associated with TAM resistance (Zheng, et al. 2021a); higher expression associated with higher RFS (Zheng, et al. 2021b)
VIRMA (KIAA1429)	Cofactor that guides the METTL3 complex to bind RNA	Oncogene (Qian et al. 2019); upregulated in breast tumors (Fang et al. 2021)
ZC3H13 (KIAA0853)	Component of the WTAP complex	Oncogene, positively associated with ER/PR expression (Zhang et al. 2020); high expression correlated with reduced OS (Mu, et al. 2022; Tai, et al. 2022)
RBM15/15B	RNA binding proteins that interact with METTL14	RBM15 is high in basal-like breast tumors whereas RBM15B was higher in Luminal A and normal breast-like subtypes of breast cancer (Zhang et al. 2020).

The N-terminal coiled-coil domain of WTAP interacts with the N-terminal leader helix of METTL3 and is required for methyltransferase activity (Zaccara et al. 2019). WTAP regulates METTL3/METTL14 localization to nuclear speckles for pre-mRNA interaction, methylation, and splicing (Ping, et al. 2014). Nuclear speckles are dynamic interchromatin granules that function as sites for RNA 5’capping, 3’end processing, splicing and m6A methylation (reviewed in (Galganski, et al. 2017; Ilık and Aktaş 2021)). Recent studies show that METTL3 undergoes phase separation in HEK-293T nuclei supporting the scaffolding of the METTL3 complex to promote m6A mRNA modification (Han, et al. 2022). CBLL1 (HAKAI) is an E3 ubiquitin ligase that interacts with WTAP and is required to writer-complex interactions (Zaccara et al. 2019). ZC3H13 also interacts with WTAP and with RBM15/15B and is required for m6A methylation (Zaccara et al. 2019).

BioID, a proximity-dependent labelling approach in which a bait protein is fused to an E. coli biotin ligase (BirA*) is used to label proteins within a 10 nm radius (Liu, et al. 2018), and coupled to liquid chromatography-tandem mass spectrometry (LC-MS/MS) was used to identify METTL3 interacting proteins in HEK-293T cells (Covelo-Molares et al. 2021). In addition to the proteins listed in Table 2, a number of other proteins involved in transcription, e.g., STAT5B, ZEB2; chromatin remodeling, i.e., BCL7C; DNA replication, i.e., CHFR and MCMBP; and protein modification, e.g., SUGT1, DNAJA2, FKBP4/5, were identified (Covelo-Molares et al. 2021). These results have not been validated or extended in other cell lines, nor in BC cells.

A novel circRNA from the METTL3 gene, circMETTL3 (623 nt originating from exon-2 and exon-3) was identified as highly expressed in in breast tumors (Li, et al. 2021b). Experiments in ZR-75–1 and SUM1315 BC cells demonstrated that siRNA targeting METTL3 inhibited cell proliferation, migration, invasion, and xenograft tumor growth in vivo (Li et al. 2021b). circMETTL3 was reported to sponge miR-31–5p, thus increasing its target CDK1 which drives BC progression (Li et al. 2021b).

MTC transcripts and proteins in breast cancer

Examination of BC data in The Cancer Genome Atlas (TCGA) showed decreased transcript levels of METTL3, METTL14, and WTAP (1.575-, 1.200- and 1.883-fold) in breast tumors compared to normal controls (Wu, et al. 2019). Similar findings for METTL3, METTL14, and WTAP using the ONCOMINE BC data were reported; additionally, ZC3H13 transcript was also decreased, but RBM15B and VIRMA transcripts were increased (Zhang et al. 2020). qRT-PCR examination of 36 breast tumor samples (no histopathological data were included) identified significant reductions in the expression of METTL3, METTL14, and WTAP (Wu et al. 2019). These breast tumors had reduced global m6A levels in total RNA (primarily rRNA and tRNA with ~ 5% mRNA), although the inter-sample variability was high, with some tumor samples showing elevations in gene transcript and in m6A (Wu et al. 2019).

Other studies reported increased METTL3 in BC cells and tumors. Increased METTL3 mRNA expression in breast tumors correlated with increased expression of Late Endosomal/Lysosomal Adaptor, MAPK and MTOR Activator 5 (LAMTOR5, HBXIP) in BC cell lines and patient breast tumor arrays (TMA) (Cai, et al. 2018). This study demonstrated that HBXIP increased let-7g which directly targeted and reduced METTL3 in MCF-7 luminal A (ER+/PR+, HER2-) BC cells and that METTL3-induced m6A methylation of HBXIP increased HBXIP protein expression (Cai et al. 2018). METTL3 transcript levels were higher in GEPIA database, in breast tumors, in BC cell lines (including MCF-7 and MDA-MB-231 TNBC cells) compared to MCF-10A (immortalized ‘normal’ breast epithelial) cells (Wang, et al. 2020a). Knockdown of METTL3 by shRNA (~ 60% decrease in protein) in MCF-7 and MDA-MB-231 cells inhibited cell proliferation and reduced BCL2,an apoptosis regulator (Wang et al. 2020a). High METTL3 transcript was associated with lower RFS in the Kaplan-Meier plotter (Gyorffy, et al. 2010) used in the study (Wang et al. 2020a). Higher METTL3 protein correlated with higher levels of m6A (dot blot) in BC cell lines (MCF-7, T47D (luminal A), MDA-MB-231, SKBR3 (ER-/PR-/HER2(ERBB2)+, and BT474 (ER+/PR+/HER2+)) and knockdown of METTL3 (shRNA) increased sensitivity to growth inhibition by chemotherapeutics in vitro: ADR, 5-FU, cisplatin, paclitaxel, and carboplatin (Li, et al. 2022b). METTL3 promoted double-stranded DNA (dsDNA) break repair and homologous recombination repair (HHR) after DNA damage induced by etoposide (ETO) and adriamycin in MCF-7 and MDA-MB-231 cell lines (Li et al. 2022b). The mechanism involved METTL3 m6A methylation of epidermal growth factor (EGF) which increased EGF translation via interaction with m6A reader YTHDC1. In turn, EGF increased the expression of RAD51, a DNA repair enzyme, RAD51 recombinase (Li et al. 2022b). Inhibiting DNA repair is an important target in BC to enhance the efficacy of chemotherapy, e.g., the use of PARP1 inhibitors (McCann, et al. 2019). METTL3 expression was upregulated in BC stem cells (BCSC) and increased expression of AF4/FMR2 family member 4 (AFF4) which binds the MYC and SOX2 gene promoters to regulate ‘stemness’ (Gao, et al. 2020)

METTL14 mRNA and protein expression was increased in breast tumors and some BC cell lines (Sun, et al. 2020). LINC00942 (LNC942) is upregulated in BC was shown to promote METTL14-mediated m6A methylation by interacting with METTL14 and upregulating the transcription of specific target genes, CXCR4 and CYP1B1, that play roles in EMT and BC tumorigenesis (Sun et al. 2020). Upregulation of oncogenic lncRNA UCA1 in T47D BC cells inhibited METTL14 transcription by recruiting DNMT1, DNMT3A/B to increase METTL14 promoter DNA methylation. Consequently the decrease in METLL14 reduced m6A in pri-miR-375 which increased pre-miR-375 and mature miR-375 that reduced SOX12 translation, resulting in increased proliferation and invasion, and reducing apoptosis (Zhao, et al. 2022).

Aurora kinase A (AURKA) positively regulated METTL14 protein expression in BC cells in a protein kinase-independent manner by inhibiting its ubiquitination-mediated degradation to increase m6A content in the DROSHA mRNA transcript resulting in increased DROSHA mRNA stability and binding of IGF2BP2 in MDA-MB-231 BCSC (Peng, et al. 2021).

Nuclear WTAP protein expression was higher in breast tumor specimen (n = 347) compared to noncancerous breast tissue (n = 23) with higher staining associated with tumor size > 2cm, higher tumor grade (II or III), and in ER- tumors compared to TNBC which had the highest WTAP staining (Wang, et al. 2022). We reported that WTAP transcript abundance was higher in LCC9 endocrine-resistant BC cells compared to MCF-7 parental cells (Klinge, et al. 2019b). Whether ER inhibits WTAP expression in a ligand-dependent manner is unknown. HER2 status had no correlation with WTAP staining (Wang et al. 2022). Interestingly, higher WTAP staining was observed in LN metastatic negative compared to LN metastatic positive breast tumors. The authors suggested this result to indicate a role for WTAP as a suppressor of metastasis; however, no metastatic tumors from distant site were examined (Wang et al. 2022). There was no association WTAP protein expression and breast cancer patient survival (Wang et al. 2022).

Hypoxia increases the expression of METTL14 in MDA-MB-231 cells (Panneerdoss, et al. 2018). Knockdown of METTL14 inhibited the growth, migration, and invasion of MDA-MB-231, MDA-MB-468, and BT-549 TNBC cells in vitro and MDA-MB-231 tumor xenografts in vitro by ~ 30% (Panneerdoss et al. 2018). The authors found that METTL14 and the m6A eraser ALKBH5 constitute a positive feedback loop with the RNA binding protein HuR (ELAV1) to regulate the stability of target genes, i.e., cyclin D1, SMAD3, cyclin E1, TGFβ1, VEGFA, HMGA2, and PDGF, in MDA-MB-231 cells (Panneerdoss et al. 2018).

METTL3 methylation of m6A on pri-miRNAs (Alarcón, et al. 2015b) and RNA-dependent interaction of the m6A reader HNRNPA2B1 with DGCR8, a component of the DROSHA complex, stimulate processing of selected pri-miRNA-m6A to precursor miRNA (pre-miRNA) (Alarcón, et al. 2015a). METTL3 protein expression was increased in adriamycin-resistant MCF-7 breast cancer cells (MCF-7/ADR) and enhanced pri-miR-221–3p m6A methylation and increased mature miR-221–3p expression which decreased Homeodomain Interacting Protein Kinase 2 (HIPK2) protein expression in MCF-7/ADR cells (Pan, et al. 2021).

VIRMA (KIAA1429) is the largest (202 kDa) component of the METTL3 complex and acts as a scaffold and bridges the METTL3/METTL14/WTAP complex with RNA and is upregulated in BC at the transcript and protein levels (Qian, et al. 2019). Knockdown of VIRMA (KIAA1429) in MCF-7 and MDA-MB-231 BC cells inhibited proliferation, migration, invasion, and colony formation (Qian et al. 2019). VIRMA was shown to have the most copy number variations (CNV) among m6A readers, writers, and erasers (Zhao, et al. 2021). Further studies are needed to validate this report.

RNA binding proteins as m6A readers

While studies have identified altered m6A on transcripts in human disease, their potential mechanism of action and involvement in human pathologies is not well understood. A focus of m6A research has been characterizing the protein complexes responsible for identifying the reversible and dynamic m6A mark in mRNA and investigating the downstream effects resulting from m6A modification (reviewed in (Zaccara and Jaffrey 2020; Zaccara et al. 2019)). At the time of this review, 20 RNA-binding proteins (RBPs) have been identified to bind to m6A-methylated regions of the mRNA transcript and regulate transcript outcome for protein expression (Du, et al. 2016; Wang, et al. 2014; Wang, et al. 2015; Xiao, et al. 2016). m6A readers with known relevance to breast cancer, their cellular location, and molecular functions are summarized in Table 3. These m6A readers have been shown to play a role in embryonic stem cell fate (Batista, et al. 2014; Lasman, et al. 2020) energy homeostasis (Edupuganti, et al. 2017), human circadian rhythm (Robinson, et al. 2019), and various cancers (Jing, et al. 2021; von Hagen, et al. 2021). Evidence suggesting that the dysregulation of m6A readers plays a pivotal role breast cancer progression will be described below.

Table 3. RNA-binding proteins identified as m6A readers.

Each m6A reader, location, function, and role in breast cancer are listed.

Enzymes	Location	Cellular function	Additional roles in breast cancer
YTHDF1	Cytoplasm	RNA translation efficiency (Wang et al. 2015)	YTHDF1 mediated translation of Spred2, a gene that encodes for SPRED2, a negative regulator of ERK signaling, in murine bone marrow-derived macrophages (BMDMs) (Yin, et al. 2021). Elongation and translation of KRT7, which is higher in BT-549 cells, is mediated by the YTHDF1/eEF-1 complex (Chen et al. 2021a)
YTHDF2	Cytoplasm	RNA degradation (Du et al. 2016; Wang et al. 2014)	YTHDF2 is upregulated in TCGA-BRCA analysis of TNBC tumor tissue (Wang, et al. 2020b).
YTHDF3	Cytoplasm	RNA degradation or translation efficiency (Li, et al. 2017; Shi et al. 2017)	Knockdown of YTHDF3 in MDA-MB-231 and BT-549 cells decreased the expression of EMT network protein ZEB1 (Lin et al. 2022).
YTHDC1	Nucleus	Alternative splicing (Patil et al. 2016; Roundtree, et al. 2017b)	YTHDC1 is downregulated in TCGA-BRCA analysis of TNBC tumor tissue (Wang et al. 2020b).
YTHDC2	Cytoplasm	RNA translation efficiency (Hsu et al. 2017; Mao, et al. 2019)	YTHDC2 is downregulated in TCGA-BRCA analysis of TNBC tumor tissue (Wang et al. 2020b)
IGF2BP1 (IMP1, CRD-BP)	Cytoplasm	RNA stability (Fakhraldeen, et al. 2015; Stohr, et al. 2012; Zhang et al. 2021a)	IGF2BP1 interacted with lncRNA KB-1980E6.3 to stabilize MYC in breast cancer stem cells (Zhu, et al. 2021). Decreased expression of IGF2BP1 by miR-873–5p in BT-549 cells reduces cMYC expression (Tang, et al. 2020). Phosphorylation of tyrosine 260 on a truncated form of IGF1BP2 (ΔN-IMP1), which is expressed in ER+ cell lines MCF-7 and ZR-75–1, is required for MYC mRNA stabilization (Abdullah, et al. 2018)
IGF2BP2	Cytoplasm	RNA stability (Dai, et al. 2015; Pu, et al. 2020; Simon, et al. 2014)	Overexpression of IGF2BP2 (p62/IMP2) in MDA-MB-231 cells increased MYC expression and cell migration, while reducing cell adhesion (Li, et al. 2015)
IGF2BP3	Cytoplasm	RNA stability (Yang et al. 2020)	IGF2BP3 interacts with circFOXK2 and miR-370 to promote cell migration and invasion in BT-549 cells (Zhang, et al. 2021b). IGF2BP3 depletion repressed TRIM25 expression and reduced MCF-7 and MDA-MB-231 proliferation (Wang, et al. 2019).
HNRNPA1	Nucleus	Alternative splicing (Loh, et al. 2015)	HNRNPA1 is higher in breast, colorectal, and prostate cancer cells and associated with increased expression of type I and type II methyltransferases (PRMTs) (Li et al. 2021a). Alternative splicing of CD44 was mediated by HNRNPA1 in MDA-MB-231 cells and was associated with metastatic progression (Loh et al. 2015).
HNRNPA2B1	Nucleus	Alternative splicing and miRNA processing (Alarcón et al. 2015a)	Based in TCGA breast cancer datasets, HNRNPAB is increased with elevated histological grade and higher in MDA-MB-468 TNBC cells compared to MCF-7 cells (An et al. 2021). Plays a role in endocrine-therapy resistance in an MCF-7 cell model (Klinge et al. 2019a; Petri et al. 2021).
HNRNPC	Nucleus	m6A switch recognition, Alternative splicing (Liu et al. 2015)	HNRNPC is upregulated in TCGA-BRCA analysis of TNBC tumor tissue (Wang et al. 2020b).
FMR1	Nucleus/Cytoplasm	RNA trafficking, translation regulation (Darnell et al. 2011)	Genotype analysis of BRCA patients identified an association between FMR1 expression and the BRCA1/2 mutation (Tea, et al. 2013).
LRPPRC	Nucleus/Mitochondria	mRNA maturation (Cui, et al. 2019)	LRPPRC suppression by miR-1 in stem cells isolated from MCF-7 cells induced mitochondrial damage (Zhang, et al. 2019).
PRRC2A	Nucleus/Cytoplasm	RNA stabilization (Chen et al. 2019a)	PRRC2A is upregulated in breast cancer compared to normal samples in TCGA breast cancer data (Yang et al. 2021)
SND1	Nucleus/Cytoplasm	Alternative splicing, RNA editing, and miRNA-mediated mRNA degradation (Paukku, et al. 2008; Scadden 2005)	SND1 interacted with Linc00668 and induced SMAD2, SMAD3, and SMAD4 transcription in MDA-MB-231 cells. Depletion of SND1 in MDA-MB-231 and MCF-7 cells reduced migratory ability and invasiveness (Qian, et al. 2020; Yu et al. 2015). Higher SND1 (IHC tumor vs. normal) was associated with advanced stages and larger breast tumor size (Gu, et al. 2018).
ELAV1 (HuR)	Nucleus/Cytoplasm	RNA stability (Ling, et al. 2020)	ELAV1 stabilized and promoted transcription of ZEB1 in MDA-MB-436 and MDA-MB-453 cells, promoting migratory ability and invasiveness (Luo, et al. 2020).
EIF3	Cytoplasm	RNA translation initiation	EIF3m knockdown in MDA-MB-231 and MDA-MB-468 cells decreased cell proliferation, migration, and invasion (Han et al. 2020). EIF3D knockdown inhibits activation of Wnt/β-catenin signaling in MDA-MB-231 and MCF-7 cells (Fan and Guo 2015). EIF3e knockdown reduces levels of PARP1 in MDA-MB-231 cells inducing senescence (Morris et al. 2021).

YTH Domain m6A-binding proteins

The earliest identified m6A readers contain a broadly conserved YT521-B homology (YTH) domain with tryptophan residues forming a hydrophobic pocket that specifically recognizes and encapsulates m6A-containing RNA (Luo and Tong 2014). Five YTH domain-containing proteins have been identified in the mammalian genome; YTHDC1, YTHDC2, and three YTHDF paralogs - YTHDF1, YTHDF2, and YHDF3 (reviewed in (Zaccara and Jaffrey 2020)). In 2012, the binding affinity of YTHDC1, YTHDF2, and YTHDF3 to m6A was validated using affinity chromatography followed by mass spectrometry (MS) (Dominissini, et al. 2012). The direct binding of these reader proteins as well as YTHDC2 have been validated using individual-nucleotide-resolution UV crosslinking and immunoprecipitation (iCLIP) (Patil, et al. 2016). YTHDF family members are primarily cytoplasmic and contain the YTH domain located at the C-terminus as well as a larger domain enriched with Q, N, and P residues (Meyer and Jaffrey 2017). Studies suggest that m6A-containing mRNAs have a reduced half-life compared to mRNAs without m6A (Batista et al. 2014; Geula, et al. 2015); however, each YTHDF paralog has different functions related to both mRNA stability and translation efficiency. Binding of YTHDF1 to m6A in the 3’ untranslated region (UTR) recruits eukaryotic translation initiation factor 3 (eIF3) to facilitate 5’ cap and polyA tail addition, as well as the forming of the cyclic structure to promote translation (Wei, et al. 2022). In contrast, binding of YTHDF2 to m6A sites in the 3’UTR leads to shortening of the polyA tail length and localizes m6A-modified mRNA transcripts to processing bodies in the cytoplasm, promoting mRNA degradation (Wang et al. 2014; Wei et al. 2022). A study in human astrocytes exploring the role of YTHDF2 in mRNA degradation showed that YTHDF3 influences proliferation and migration of glioma cells by suppressing UBX domain protein 1 (UBXN1), thus inducing NF-kB activation (Chai, et al. 2021). The role of YTHDF3 is less clear, although it appears to share many targets with YTHDF1 and YTHDF2, suggesting dynamic and cooperative activity between the YTHDF proteins regulating translation and decay of common mRNAs (Shi, et al. 2017).

YTHDC1 is predominantly nuclear (Patil et al. 2016). An investigation of YTHDC1 nuclear interacting proteins in HeLa cells revealed the enrichment of splicing enhancer binding SR proteins, SRSF1 (ASF/SF2), SRSF3 (SRp20), SRSF7 (9G8), SRSF9 (SRp30c), and SRSF10 (SRp38), identifying a role for YTHDC1 in alternative splicing (Xiao et al. 2016). The function of cytoplasmic YTHDC2 is less defined, but it includes a DEAD-box RNA helicase domain that may be targeted to specific mRNAs by m6A (Patil et al. 2016). A direct role for YTHDC2 for translation efficiency was discovered using luciferase-based tethered reporter assays in HeLa cells (Hsu, et al. 2017).

IGF2BP family m6A-binding proteins

Insulin-like growth factor-2 (IGF2) mRNA–binding proteins (IGF2BPs) constitute a conserved family of single-stranded RBPs that specifically recognize m6A-RNA and influence transcript outcomes (Bell, et al. 2013; Du, et al. 2021; Korn, et al. 2021). The role of IGF2BPs in regulating transcript outcomes has been studied since 2002, well before IGF2PBs were discovered to be m6A readers, and several synonyms for IGF2BP exist in the literature, including, but not limited to, IMP, CRD-BP, VICKZ and ZBP (Bell et al. 2013). IGF2BP1/2/3 are predominantly cytoplasmic and modulate cell function during development and cancer (Samuels, et al. 2020; Stöhr, et al. 2012; Zhou, et al. 2022). These three IGF2BP proteins have a 56% aa sequence identity with two RNA-recognition motifs (RRMs) at their N-termini and hnRNP-K homology (KH) domains at their C-termini (Bell et al. 2013). This sequence homology suggests that the IGF2BP proteins share biochemical functions. IGF2BPs form ribonucleoprotein (RNP) granules in the cytoplasm and encapsulate their target mRNAs in protein-RNA complexes (mRNPs), preventing premature decay (Kobel, et al. 2007; Lemm and Ross 2002; Vikesaa, et al. 2006).

RNA-pulldown experiments using methylated single-stranded RNA bait followed by mass spectrometry, confirmed that IGF2BPs selectively bind m6A-RNA independent of RNA secondary structure and depletion of IGF2BPs globally downregulated target gene expression (Huang, et al. 2018). In human endometrial cancer cells, IGF2BP1 binds to the m6A mark on PEG10 mRNA and recruits polyadenylate-binding protein 1 (PABC1) to stabilize the transcript, promoting PEG10 protein expression (Zhang, et al. 2021a). IGF2BP2 stabilized tuberous sclerosis complex 1 (TSC1) and peroxisome proliferator activated receptor-γ (PPARγ) contributing to mouse M2 macrophage formation (Wang, et al. 2021). In human colon cancer cells, IGF2BP3 stabilized and prevented degradation of CCND1 and VEGFA, increasing proliferation and angiogenesis (Yang, et al. 2020). Collectively, these studies established a role for IGF2BPs in mRNA stability and show that IGFBP1/2/3 bind to transcripts harboring the m6A mark and prevent mRNA degradation in human disease, including breast cancer as will be described below.

HNRNP m6A-binding proteins

The role of heterogeneous ribonucleoproteins (HNRNPs) in alternative splicing, mRNA packaging, and transport contributing to proteomic diversity in eukaryotes is well studied (Geuens, et al. 2016; Kedzierska and Piekielko-Witkowska 2017). HNRNPs were first identified and characterized in HeLa cells as A, B, and C groups comprising HNRNP particles that form on all nascent transcripts (Beyer, et al. 1977) and act as key proteins in mRNA biogenesis (Choi, et al. 1986; Osheim, et al. 1985). The major classes of HNRNPs each contain conserved domains that facilitate RNA-binding: the RRM, the quasi-RRM, an arginine-glycine-glycine (RGG) box, and a KH domain (Gorlach, et al. 1992; Siomi and Dreyfuss 1995; Swanson and Dreyfuss 1988). HNRNPs A/B are divided into four subgroups: HNRNP A1, A2/B1, A3, and A0, though the more highly expressed A1 and A2/B1 are the most studied for the role they play in mRNA translation (Park, et al. 2015), splicing (Mayeda and Krainer 1992), and trafficking (Shan, et al. 2000). Only in the past 10 years, have HNRNP proteins been associated with m6A-RNA mechanisms. In 2012, Tavazoie and others investigated how HNRNP proteins stabilize mammalian mRNA (Goodarzi, et al. 2012), and in 2015 they established a clear connection between HNRNPA2/B1 and m6A-RNA by showing that HNRNPA2/B1 recognizes m6A in mRNA in MDA-MB-231 cells to mediate alternative splicing (Alarcón et al. 2015a).

HNRNPA2/B1 was reported to bind to m6A on primary microRNA (pri-miRNAs) transcripts in MDA-MB-231 cells and interacts with the DROSHA complex to induce pri-miR to premature (pre)-miR processing (Alarcón et al. 2015b). We reported that HNRNPA2B1 was increased in endocrine-resistant ER+ breast cancer cells, played a role in sensitivity to tamoxifen and fulvestrant, and regulated miRNA expression (Klinge et al. 2019b; Petri, et al. 2021). This will be described in more detail below, HNRNPC contains only one RBD and forms self-oligomers for RNA binding (Swanson, et al. 1987). In 2015, HNRNPC was reported to bind to a secondary RNA structure that occurs when m6A remodels RNA to form truncated hairpins. This m6A-remodelled structure, also known as an m6A-switch, creates accessible U-tracts which significantly improves HNRNPC binding affinity and facilitates RNA splicing (Liu, et al. 2015). HNRNPG is another HNRNP that takes advantage of the m6A-switch mechanism by preferentially binding to target transcripts remodeled by m6A. Depletion of HNRNPG in HEK-293T cells altered the expression and splicing of hundreds of transcripts altered by METTL3 or METTL14 knockdown (Liu, et al. 2017a). Together, these reports validate the connection between m6A-remodelling and HNRNP m6A-readers in regulating mRNA metabolism.

Other m6A-binding proteins

Additional RBPs have been identified and characterized as m6A readers. High-efficiency diazirine-based RNA photo-crosslinking of m6A-mRNA in HeLa cells combined with LC-MS/MS proteomics identified novel m6A-RNA-protein interactions, including the leucine rich pentatricopeptide repeat containing (LRPPRC) and fragile X-messenger ribonucleoprotein 1 (FMR1) proteins as m6A-binders (Arguello, et al. 2017). Gene set enrichment analysis from GEO datasets of lung adenocarcinoma patients found that downregulation of LRPPRC was correlated with immune response activation, suggesting that LRPPRC may regulate the tumor environment (Li, et al. 2021b). FMR1 encodes FMRP, which contains an RGG domain that recognizes the m6A-modified RNA (Myrick, et al. 2015). FMRP associates with CDS of transcripts, stalling ribosomes and regulating translation at the elongation stage (Darnell, et al. 2011). Another RBP that recognizes m6A-RNA is staphylococcal nuclease-like (SN-like) domain-containing protein 1 (SND1), a member of the Tudor family of proteins that binds to methylated DNA. SND1 was first identified in Schizosaccharomyces pombe, but is conserved from yeast to humans (Chen, et al. 2011). Although, the exact mechanism of the m6A/SND1 axis is not understood, a pan-cancer analysis of TGCA datasets shows a correlation between increased SND1 expression and poor OS and RFS in many cancers, including breast cancer (Cui, et al. 2020). Other RBPs with roles in translation have been identified as possible m6A readers. For example, eukaryotic initiation factor 3 (eIF3), a component of the 43S preinitiation complex, interacts with YTHDF1 or YTHDF3 at m6A-modified RNA in the 5’-end of transcripts (Shi et al. 2017; Tarun, et al. 1997). However, binding of eIF3 to m6A in the 5’UTR of transcripts does not require YTHDF1, suggesting that eIF3 directly binds m6A (Meyer, et al. 2015). The identification of RBPs that recognize m6A and their roles in transcriptional-translational regulation require further investigation.

m6A readers in breast cancer

Dysregulation of m6A modification and the expression of m6A readers is associated with cancer progression (Bai, et al. 2019; Liu, et al. 2020b). Analysis of data from The Cancer Genome Atlas (TCGA) for YTHDF1 and YTHDF3 showed increased transcript expression of these readers in breast tumors compared to normal tissue that was correlated with reduced OS of breast cancer patients (Anita, et al. 2020). In a separate study, gene set enrichment analysis (GSEA) of microarray expression datasets from the TCGA database showed distinct enrichment of the WNT signaling pathway in patient samples with high YTHDF1 expression and knockdown of YTHDF1 in MCF-7 cells reduced the expression of WNT pathway proteins (Hu, et al. 2021). YTHDF1 and YTHDF3 were associated with lower OS and higher expression rates of genes related to c-Myc signaling and T cell differentiation in TCGA breast cancer data (Li, et al. 2022a). Despite evidence that YTH domain proteins play an important role in mRNA translation and stability in cancer, few studies have evaluated their specific role(s) in breast cancer (Table 3). The REPIC (RNA EPItransriptome Collection) database only lists one m6A-Seq/Me-RIP-seq data set from breast cancer cells (Liu, et al. 2020a). YTHDF1, with METTL14, promoted double stranded break (DSB) DNA damage repair and YTHDF1 knockdown increased sensitivity to adriamycin, cisplatin, and olaparib in MDA-MB-231 cells (Sun, et al. 2022). FOXM1, which promotes breast cancer metastasis, is an YTHDF1 target and binding of YTHDF1 to FOXM1 mRNA accelerated translation of FOXM1 in MCF-7 cells (Chen, et al. 2022). Binding of YTHDF1 to PKM2 mRNA stabilized the transcript to increase translation efficiency, which increased glycolytic activity in MBA-MD-231 and MCF-7 cells (Yao, et al. 2022).

Depletion of YTHDF2 increased apoptosis in MDA-MB-231 by increasing mitochondrial stress, proteotoxicity, endoplasmic reticulum stress (ERS), and the unfolded protein response (UPR) (Einstein, et al. 2021b). Interestingly, eCLIP for YTHDF identified unique YTHDF target genes in MDA-MB-231, MCF-7, and SKBR3 cells that overlapped with m6A sites identified by m6A-RIP-seq (Einstein et al. 2021b). shYTHDF2 upregulated transcript levels of genes promoting cell migration, wound healing, EMT, and metastasis in MDA-MB-231 cells (Einstein et al. 2021b). In summary, YTHDF2 was demonstrated to bind m6A peak-containing transcripts for degradation in MDA-MB-231 cells, thus ‘safeguarding’ the cells from ERS and proteotoxicity (Einstein et al. 2021b). By binding m6A, YTHDF2 stabilized ATF3 transcripts, thus increasing ATF3 protein, a transcription factor that was identified as a potential driver of TAM-resistance in TAM-resistant MCF-7 cells (Liu, et al. 2021). A study evaluating TNBC tissue samples identified an association between YTHDF3 expression, lymph node metastasis, and higher histological grade (Lin, et al. 2022). YTHDF3 was reported to increase the binding of eIF3 to ST6GALNAC5, GJA1, EGFR, and VEGFA mRNAs in MDA-MB-231 cells, suggesting YTHDF3-stimulated translation of key genes that drive breast cancer brain metastasis (Chang, et al. 2020). YTHDC1 overexpression in MDA-MB-231 cells promoted DNA replication and DNA damage repair and induced resistance to the adriamycin (Zhou, et al. 2021). Taken together, these studies suggest that the YTH domain family of m6A readers plays a dynamic role in altering the expression of genes in pathways associated with breast cancer progression and metastasis.

MYC is an important oncogenic transcription factor in TNBC (Fallah, et al. 2017) and MYC mRNA has been reported as a downstream target of IGF2BP proteins (Huang et al. 2018) as well as YTHDF1 (Zheng, et al. 2022). Binding of IGF2BPs by FGF13-ASI, a long non-coding RNA (lncRNA) that decreases glycolysis and stemness in MDA-MB-468 cells and prevents MYC mRNA stabilization by IGF2BPs (Ma, et al. 2019). IGF2BPs also interact with circCD44 to stabilize MYC mRNA in BT549 cells (Li, et al. 2021a). Further evidence for IGF2BP regulation of MYC mRNA in breast cancer is listed in Table 3.

IGF2BP1 bound directly to m6A sites on MIR210HG, an oncogenic lncRNA that promotes proliferation and EMT in MCF-7 cells (Shi, et al. 2022b). Similar associations between IGF2BP1 and other lncRNAs with roles in breast tumor progression have been established, e.g., upregulation of LINC00483, an oncogenic lncRNA associated with advanced tumor stage (Qiao, et al. 2021). IGF2BP2 stabilized DROSHA mRNA for translation in MDA-MB-231 cells and DROSHA induced the expression of genes promoting sphere formation and cell stemness, i.e. STC1, GRB10, and SLCo4A1 (Peng et al. 2021). METTL3 and IGF2BP3-mediated m6A modification that increased PD-L1, a cell surface immune checkpoint inhibitor, in MDA-MB-231 and HCC38 cells (Wan, et al. 2022). These studies characterizing IGF2BPs and their interaction with m6A in target transcripts suggest lncRNA-epitranscriptomic regulation of breast cancer progression and drug-resistance.

The role of HNRNPs in breast cancer progression and drug-resistance is being actively investigated (An, et al. 2021; Klinge, et al. 2019a; Li, et al. 2021a; Petri et al. 2021). There are 16 HNRNP family members named alphabetically A-U, classified by their RNA binding domains and have diverse structural features and functions (reviewed in (Geuens et al. 2016)). HNRNPA2B1 expression is higher in breast tumors compared to normal breast tissue (Liu, et al. 2019b; Ma, et al. 2020) and regulates stability, alternative splicing, DNA repair, and telomere maintenance (Liu and Shi 2021). HNRNPA2B1 plays a role in the processing of selected pri-miRNAs to pre-miRNAs in HEK-293T cells (Alarcón et al. 2015b). HNRNPA2B1 stimulated miRNA processing in MDA-MB-231 cells by recognizing m6A on pri-miRNAs and interacting with the DROSHA complex protein DGCR8, to increase pri-miRNA processing to pre-miRNAs (Alarcón et al. 2015b). However, some studies suggest that HRNPA2B1 may also interact with negative regulators of miRNA biogenesis and suppress miRNA processing (Suster and Feng 2021). Analysis of the miRome in MCF-7 cells with overexpressed HNRNPA2B1 found up- and down-regulation of hundreds of miRNAs and identified dysregulated pathways associated with breast cancer progression, metastatic spread, and TAM-resistance (Klinge et al. 2019a). Knockdown of HNRNPA2B1 inhibited proliferation in MCF-7 and MDA-MB-231 breast cancer cells by downregulating signal transducer and activator of transcription 3 (STAT3) and extra-cellular-signal-regulated kinase 1/2 (ERK1/2) (Hu, et al. 2017). In addition, overexpression of HNRNPA2B1 increased endocrine resistance, cell motility, and stem cell properties in MCF-7 cells whereas knockdown of HNRNPA2B1 restored TAM and fulvestrant sensitivity to endocrine resistant LCC9 and LY2 breast cancer cells (Petri et al. 2021). Investigating other HNRNPs has revealed novel m6A regulatory mechanisms in breast cancer. For example, HNRNPC expression promoted cell proliferation by inducing MAPK signaling in MDA-MB-231 and MCF-7 cells (Lv, et al. 2021). HNRNPK promotes invasion and metastasis in MDA-MB-231 and MDA-MB-468 cells by interacting with PROX1 and stimulating the WNT pathway (Zhu, et al. 2022).

Emerging roles have been identified for other m6A-associated RBPs in breast cancer progression. FMR1 and LRPPC are over-expressed in breast cancer tissue data from TCGA compared to normal tissue (Zhang et al. 2020). Additionally, GTEx data revealed a higher expression of LRPPC in breast tissue (Zhang et al. 2020). SND1 is upregulated in breast primary invasive ductal carcinomas and induced RhoA degradation, thus disrupting F-actin cytoskeletal organization, by promoting expression of Smurf1, an E3 ubiquitin ligase (Yu, et al. 2015). The role of eIF3 subunits in breast cancer tumorigenesis has been well-documented (Fan and Guo 2015; Han, et al. 2020; Morris, et al. 2021; Zheng, et al. 2016) and there is clear evidence for an m6A-eIF3 interaction (Meyer et al. 2015). Further studies investigating RBPs involved in breast cancer progression, drug-resistance, and metastasis are necessary to define their role in m6A-mediated RNA metabolism in breast cancer.

m6A Erasers in Breast Cancer

ALKBH5 and FTO are members of a conserved family of AlkB non-hem Fe(II)/2-oxoglutarate (2OG) dependent dioxygenases with nine ALKBH homologues with diverse substrates and specificities (reviewed in (Rajecka, et al. 2019)). Knockout phenotypes for ALKBH5 and FTO are divergent: infertility and impaired spermatogenesis versus growth retardation, lean phenotype and impaired adipogenesis, respectively (reviewed in (Rajecka et al. 2019)) ALKBH5 and FTO were described as the two mammalian non-heme Fe²⁺ and α-ketoglutarate (αKG)–dependent dioxygenases that act as m6A mRNA demethylases that catalyze the oxidation of m6A in a two-step reaction that require O₂ and generates CO₂, succinate, and Fe(III) and ultimately yields adenosine and formaldehyde (reviewed in (Meyer and Jaffrey 2014; Purslow, et al. 2021)).

ALKBH5

There is evidence that ALKBH5 is the specific m6A-demethylase whereas FTO is the m6Am demethylase (Mauer and Jaffrey 2018; Mauer, et al. 2017; Mauer et al. 2019). Hypoxia (1% O₂) increased ALKBH expression in MCF-7 luminal A and in MDA-MB-231, SUM159, and MDA-MB-435 TNBC cell lines in a HIF-1α- and HIF-2α- dependent manner and reduced m6A methylation in total RNA (Zhang, et al. 2016). Further experiments in MCF-7 and MDA-MB-231 cells revealed that ALKBH5 demethylated m6A in NANOG mRNA which increased total NANOG mRNA transcript levels and enriched BCSC populations in these cell lines (Zhang et al. 2016). NANOG is a pluripotency factor that promotes breast cancer tumorigenesis and metastasis (Lu, et al. 2014). Knockdown of ALKBH5 in MDA-MB-231 cells reduced mammary tumor growth in vivo (Zhang et al. 2016). Immortalization of primary human mammary epithelial cells (HMECs) by stable expression of hTERT, P53^DD, Cyclin D1, CDK4^R24C, or C-MYC^T58A increased FTO and ALKBH5 while decreasing METTL3 protein expression (Fry, et al. 2018). siALKBH5 increased overall m6A levels and led to reduced long-term viability, migration, and invasion of MDA-MB-231 cancer cells (Panneerdoss et al. 2018). Doxorubicin treatment increased m6A levels in breast tumor patient samples but overexpression of PRMT5 (Protein arginine methyltransferase 5) expression in MDA-MB-231 breast cancer cells decreased m6A levels and decreased m6A methylation of BRCA1 to increase BRCA1 stability by increasing ALKBH5 nuclear translocation through an interaction with ALKBH7 (Wu, et al. 2022). The drug tadalafil was recently demonstrated to inhibit PRMT5 methyltransferase activity and reduce ALKBH5 nuclear localization (Wu et al. 2022). Tadalafil increased total RNA m6A methylation and specific m6A methylation on the 3’-UTR of BRCA1 and decreased BRCA1 mRNA stability and protein expression, enhancing the doxorubicin-induced MDA-MB-231 cell apoptosis (Wu et al. 2022). These findings suggest that blocking nuclear localization of ALKBH5 may enhance chemotherapeutic responses in breast cancer.

FTO

FTO was originally identified in genome-wide association studies (GWAS) as having a modest, but significant, contribution to childhood and adult obesity (Blakemore and Froguel 2008). FTO was reported to share motifs with Fe(II)- and 2-oxoglutarate-dependent oxygenases and showed high mRNA expression in the brain, notably in the hypothalamic nuclei and in the arcuate nucleus that regulates satiety and feeding behaviors (Church, et al. 2009). FTO was identified as an m6A demethylase in vitro and in HeLa and HEK-293T cell studies (Jia, et al. 2011). Case-control analysis of 354 breast tumors and 364 controls revealed that three single nucleotide polymorphisms (SNPs) in intron 1 of FTO were associated with breast cancer risk (Kaklamani, et al. 2011). Another study found an association with an intron 1 SNP in FTO (rs9939609) with breast cancer risk in a large study (62,328 breast cancer cases and 83,817 controls) of women of European ancestry (Zhao, et al. 2016). In women of African ancestry, FTO SNP rs17817449 was associated with both overall breast cancer and ER+ disease (Feng, et al. 2017). Reduced FTO transcript expression was associated with TAM-resistance and early distant relapse time in an analysis of two datasets (GSE26971, GSE58644) (Zhou, et al. 2018).

FTO localizes to nuclear speckles in HeLa cells (Jia et al. 2011) and MCF-7 cells and was shown to affect RNA modifications including the 3-methyluridine/uridine and pseudouridine/uridine ratios in total brain RNA (Berulava, et al. 2013). Transgenic mice with a CRISPR/Cas9 engineered mutation modeling the human FTO-R316Q mutation (mouse FTO-R313A) resulting in a catalytically inactive FTO protein showed reduced bone mineral density and content, reduced body length and mass and reduced viability (Sachse, et al. 2018). Bovine milk consumption was suggested to be an epigenetic regulator of FTO transcription via milk exosomes containing bovine miR-29s that downregulate human cell DNA methyltransferases (DNMTs), thus reducing CpG island DNA methylation in the FTO gene and increasing FTO mRNA and protein (Melnik 2015). Persistent consumption of cow’s milk is associated with hepatocellular cancer, luminal A breast cancer, prostate cancer, and Parkinson’s disease by milk exosomal regulation of mTORC1 signaling (Melnik 2021).

Conclusions as to whether FTO is increased or decreased in breast tumors compared to normal breast depend on the type of study (transcripts versus protein) and publication (Table 4 and below). FTO transcript expression was higher in MCF-7 compared to MDA-MB-231 and HCC1937 TNBC cell lines (Liu, et al. 2017b). In silico examination of FTO transcript expression (GSE9014) found FTO upregulated in breast tumors and protein was higher in IHC staining of three breast tumors vs. normal breast, correspondingly, an m6A dot blot showed lower m6A levels in breast tumors compared to normal breast (Niu, et al. 2019). Likewise examination of Oncomine data identified increased FTO and ALKBH5 in breast tumors (Wu et al. 2019). However, neither FTO nor ALKBH5 transcript expression was associated with OS in the TCGA breast cancer dataset (Chang et al. 2020). Analysis of breast cancer and normal breast tissues in the RNA sequencing expression data in GEPIA found lower FTO but no difference in ALKBH5 in breast tumors (Shi, et al. 2020). However, the distance-metastasis-free survival (DMFS) of the FTO high-expression TNBC patients was shorter than that of the FTO low-expression group, indicating that FTO is a risk factor for DMFS in TNBC.

Table 4:

m6A Erasers ALKBH5 and FTO in breast cancer

Protein (other names)	Function	Breast cancer
ALKBH5 (AlkB Homolog 5, RNA Demethylase)	Dioxygenase that specifically demethylates m6A (Zheng, et al. 2013). ALKBH5 is mainly nuclear and is specifically the m6A demethylase (Mauer et al. 2017).	• IHC staining showed increased ALKBH5 in breast cancer tissues compared to normal breast tissue (Liu et al. 2019b). • ER+ or PR+ status was associated with high mRNA levels of ALKBH5 (Wu et al. 2019). • Transcript expression was down-regulated in breast cancer compared to adjacent normal breast (Zhang et al. 2020). • No difference in transcript expression between breast tumors and normal breast tissue (Shi et al. 2020; Zhao et al. 2021). • Copy number loss in breast tumors (Yang et al. 2021). No difference in copy number variations or somatic mutations of ALKBH5 in breast cancer (Zhao et al. 2021).
FTO (fat mass and obesity associated gene (FTO) Alpha-Ketoglutarate Dependent Dioxygenase)	RNA demethylase that mediates oxidative demethylation of different RNA species, such as mRNAs, tRNAs and snRNAs, and acts as a regulator of fat mass, adipogenesis and energy homeostasis (Jia et al. 2011). FTO targets m6Am (Mauer and Jaffrey 2018; Mauer et al. 2017; Mauer et al. 2019). FTO also demethylates 3-meT in ssDNA and 3-meU in ssRNA (reviewed in (Chen and Du 2019)).FTO demethylates m6A, m6Am, and m1A in the nucleus and cytoplasm (Wei et al. 2018).	• IHC showed FTO staining was both nuclear and cytoplasmic and higher in breast cancer than adjacent normal tissue and the percent of FTO –positive expression was higher in TNBC and HER2+ than ER+/PR+ breast cancer (Tan, et al. 2015). • Another study found reduced FTO IHC staining in breast carcinoma tissues compared to normal breast tissue (Liu et al. 2019b). • Lower transcript expression in breast cancer tissues compared to normal tissue (Zhang et al. 2020; Zhao et al. 2021).

High FTO transcript abundance was associated with ER-/PR-/HER2+ and higher grade (II and III) breast tumors and with lower OS (Niu et al. 2019; Xu, et al. 2020). SKBR3 and MDA-MB-453 ER-/PR-/HER2+ breast cancer cells had higher FTO than MCF-10A, MCF-7, T47D, BT474, MDA-MB-231, or BT549 breast cancer cells (Xu et al. 2020). Lentiviral-mediated stable knockdown of FTO in MCF-7 and MDA-MB-231 cells inhibited cell proliferation, colony formation, and reduced cell apoptosis (Niu et al. 2019). Likewise, shFTO inhibited 4T1 mouse mammary tumor growth in vivo (Niu et al. 2019). RNA-seq in shFTO-MDA-MB-231 cells identified BNIP3, an apoptosis inhibitor, as upregulated by FTO knockdown in an m6A-modification-dependent manner (Niu et al. 2019). Thus, the authors concluded that increased FTO reduced m6A in the BNIP3 mRNA resulting in lower BNIP3 protein and increased breast cancer cell proliferation (Niu et al. 2019). Similarly, knockdown of FTO by siRNA in SKBR3 and MDA-MB-453 cells inhibited cell migration and invasion, although not cell proliferation. Overexpression of FTO in SKBR3 and MDA-MB-453 cells stimulated cell migration and invasion and decreased total m6A abundance (Xu et al. 2020). RNA-seq in FTO-knockdown SKBR3 cells identified miR-181b-3p as upregulated and ARL5B (ADP Ribosylation Factor Like GTPase 5B) as a direct target of miR-181b-3p, although a direct role for FTO removal of m6A in downregulating miR-181b-3p, thus increasing ARL5B in breast cancer, was not proven (Xu et al. 2020).

FTO was reduced while METTL3 and m6A levels were increased in “Lung metastasis breast cancer cells” MDA-MB-231^LMF3 and BT-549^LMF3 which are TNBC cell lines established by three rounds of tail vein injection and “lung metastasized cell” re-injection (Chen, et al. 2021a). However, chemical inhibitors of FTO (CS1 and CS2 at 0.1 and 1 µM) inhibited leukemia stem/initiating cell self-renewal and MDA-MB-231 cell proliferation in vitro and as tumor xenografts in vivo (Su, et al. 2020). While this finding suggests that inhibition of FTO may inhibit breast cancer metastasis, the study of the lung metastatic MDA-MB-231 cells suggest caution. More research is needed on FTO’s mechanism and role in metastatic disease.

m6A in mRNA and its detection

m6A modification of mRNA regulates RNA splicing, stability, translation, and intracellular localization and plays a role in many human diseases including cancers (breast, lung, hepatocellular, renal, prostate, gastric, bladder, and colorectal) and other diseases, e.g., osteoporosis and type II diabetes (reviewed in (Destefanis, et al. 2021; Lan, et al. 2021)). The number of m6A in transcripts is variable, some mRNAs can contain 20 m6A sites, generally found in genes that regulate development and cell fate (Zaccara et al. 2019).

The most common methods to detect m6A in mRNA were recently reviewed (Moshitch-Moshkovitz, et al. 2022; Zaccara et al. 2019). These include m6A-RNA immunoprecipitation (RIP)-seq (called meRIP-seq, m6A-RNA-seq, or m6A-seq), in which an m6A antibody is used to identify bound transcripts by Illumina Seq; m6A- cross-linking and immunoprecipitation (CLIP)-seq which employs UV cross-linking to identify m6A sites at higher specificity compared to m6A-seq; miCLIP (modified CLIP) that relies on inducing specific mutations (C to T conversions) during reverse transcription (Roberts, et al. 2021), LC-MS (Thüring, et al. 2016) and deamination adjacent to RNA modification targets (DART-seq) that uses a fusion protein consisting of the m6A-binding YTH domain tethered to the cytidine deaminase APOBEC1 to direct C-to-U editing at cytidine residues that invariably follow m6A sites (Meyer 2019; Tegowski, et al. 2022). Direct RNA sequencing by Oxford nanopore-based single molecule sequencing, which maintains and directly detects nucleic acid modifications as a single strand of RNA passes through a pore (Kono and Arakawa 2019; Parker, et al. 2020), without the need for added reagents or processing, i.e. antibodies specific for m6A for RNA immunoprecipitation (RIP)-seq or cDNA generation which introduce biases and only localizes the m6A mark within ~ 100 nt, rather than the precise position of the m6A (Pratanwanich, et al. 2021). A computational method (xPore) was demonstrated to identify m6A positions at single base resolution, within the DRACH/RRACH motif (Jenjaroenpun, et al. 2020), from nanopore reads (Pratanwanich et al. 2021).The nascent RNA m6A methylome was also identified by methylation inscribed Nascent Transcripts Sequencing (MINT-Seq) to capture nascent RNAs RNA metabolic labeling used in TT-Seq combined with m6A-RIP-seq (Lee, et al. 2021). An m6A-Atlas of the m6A epitranscriptome includes m6A sites in human, mouse, rat and other species, as well as 46 human cell lines (Tang, et al. 2021). Using the m6A-Atlas, m6A, m1A, m5C, m6Am, m7G, and Ψ modifications were identified in 2,294 gene transcripts involved in pharmacokinetic and pharmacodynamics, e.g., drug metabolizing enzymes and transporters, receptors, and targets (Liu, et al. 2022b). Based on their analysis, the authors concluded that m6A is a biomarker for anti-tumor drugs in TNBC (Liu et al. 2022b).

Roles for m6A mRNA modification in cellular functions

As indicated above, the presence and position of m6A in mRNA transcripts regulates stability, splicing, and intracellular location by the association of reader proteins with the m6A mark. Recently, YTHDC1 and unknown m6A readers recognized m6A and recruit KDM3B and other demethylases of H3K9me2, to the chromatin regions around the transcription start site of CYP2B6 in Huh-7 and HepaRG hepatoma cells, resulting in increased transcription (Isono, et al. 2022). This suggests that protein interactions between chromatin modifiers and m6A readers are a potential mechanism for transcript in initiation as well as transcript stability and splicing. m6A is involved in the maintenance of the “stemness properties” of cancer stem cells and the interaction between CSC and the tumor immune microenvironment (TIME) (reviewed in (Zhang, et al. 2021d)). m6A modification of MYC mRNA regulates transcript stability by binding YTHDF1 to increase translation (Zheng et al. 2022). Interestingly, m6A modification on mRNA accumulates within 2 min. at sites of DNA damage in U2OS cells generated by global UVC irradiation, allowing recruitment of DNA polymerase kappa (Pol κ) (Xiang, et al. 2017). Tonicity-responsive enhancer binding protein (TonEBP) was demonstrated to recruit METTL3 to R-loops, three-stranded DNA-RNA hybrids with ssDNA, by to methylate m6A and stimulates homologous recombination (HR) repair in YTHDC1 dependent manner at DNA double-stranded breaks (Zhang et al. 2020). There are few studies of how transcript-specific m6A affects gene expression or pathways in breast cancer. A recent study reported increased expression of ALKBH5 in HER2+ SKBR3 and BT474 breast cancer cell lines that are trastuzumab- and lapatinib-resistant and in breast tumors from patients who responded poorly to these treatments (Liu, et al. 2022a). m6A-RIP seq identified m6A peaks in the SKBR3 and BT474 and their HER2-therapy resistant derivatives and correlated increased ALKBH5 with increased m6A demethylation and expression of EGFR, FOXO1, and GLUT4 in the resistant cells. RIP-qPCR identified YTHDF2 –m6A interaction stabilized the GLUT4 mRNA transcript. m6A demethylation of GLUT4 by the increase in ALKBH5 in the resistant cell lines resulted in increased glycolysis and Ritonavir, a FDA-approved HIV protease inhibitor that also inhibits GLUT4, restored sensitivity to trastuzumab in the resistant cells (Liu et al. 2022a).

Identification of m6A sites in genes in breast cancer cells

There are relatively few studies identifying m6A sites in genes in breast cancer cell lines and few cell lines have been studied (Table 5). Most studies have used some form of m6A-RIP-seq to identify m6A peaks and a few have knocked down METTL14 or METTL3 as a control to verify m6A identification. In MDA-MB-231 TNBC cells, 15,981 and 17,312 m6A peaks from 6,796 and 7,194 m6A-containing transcripts in control and METTL14-silenced MDA-MB-231 cells, respectively (Panneerdoss et al. 2018). METTL14 target genes including TGFβ1, SMAD3, cyclin E1, cyclin D1, MMP9, VEGFA, and HMGA2 were found to be hyper m6A methylated and showed decreased expression in METTL14 KD cells, suggesting that increased m6A methylation may increase the stability of these transcripts in MDA-MB-231 cells (Panneerdoss et al. 2018). Another study in MDA-MB-231 cells knocked down either METTL3 or the reader IGF2BP2 to identify specific m6A peaks by m6A-RIP-seq and reported that most m6A peaks were identified in the CDS (Wan et al. 2022). Epitranscriptomic microarray and meRIP-seq identified 30 sequences from 24 genes that harbored both hypo-methylation and down-regulated m6A peaks after shMETTL3 in MDA-MB-231 cells including PD-L1 (CD274). IGF2BP2 reduced the CD274 transcript and PD-L1 protein, suggesting that m6A deposition and binding by IGFBP2 enhance transcript stability, and m6A-RIP-qPCR identified binding of IGF2BP3 to CD274 mRNA in an m6A-dependent manner (Wan et al. 2022). In MCF-7 cells, knockdown of METTL3 and METTL14 increased ERα and YTHDC1 protein levels (Lee et al. 2021), which can have secondary effects on measured outcomes. m6A-enchanced cross-linking and immunoprecipitation (m6A-eCLIP) was developed to specifically identify m6A sites in MCF-7 and MDA-MB-231 cells, although the authors did not describe the m6A-containing genes nor compare the two cell lines (Roberts et al. 2021). Secondary lung metastatic MDA-MB-231-LM2 tumor cells showed more m6A peaks (14,873 peaks) and m6A-peak modified genes (4,512 genes) that were unique from the parental MDA-MB-231 cell line; however, MYC-induction of HMECs did not alter m6A peaks (Einstein, et al. 2021a). Unique and common m6A peaks were identified when comparing MDA-MB-231 vs. MCF-7 or SKBR3 cells and between MCF-7 and SKBR3 cells; however, the genes containing different numbers of m6A peaks in the different cell lines were not discussed (Einstein et al. 2021a). Direct RNA seq using nanopore and xPore identified m6A sites in transcripts from HEK-293T cells and compared those in transcripts from cancer cell lines: MCF-7, K562 (leukemia), HCT116 (colon cancer), A549 (non-small cell lung cancer), and HepG2 (hepatocellular) (Pratanwanich et al. 2021). The number of m⁶A sites was lower in MCF-7 cells compared to the other cell lines; however, because the paper was focused on the computational analysis, the authors did not discuss which transcripts showed reduced m6A, nor interpret their data relative to breast tumor samples.

Table 5:

Detection of m6A peaks in genes in breast cancer cells.

Method of m6A-RIP-seq	Breast cancer dell line(s)	Other treatments	Comments, reference and data availability
Magna MeRIP™ m6A kit (Millipore) Illumina	MDA-MB-231	siMETTL14	GEO database GSE81164 (Panneerdoss et al. 2018)
ONT Nanopore direct RNA-seq with xPore computational analysis	MCF-7 with two replicates		The data are available through the ENA (PRJEB40872) (Pratanwanich et al. 2021).
m6A-enchanced cross-linking and immunoprecipitation (m6A-eCLIP)	MCF-7 and MDA-MB-231 Performed 3 biological replicates/cell line		Observed divergent numbers of m6A peaks in the three replicates: GEO database GSE147440 (Roberts et al. 2021).
Magna MeRIP™ m6A kit (Millipore) Illumina HiSeq 2000 system	BT-549 and BT-549LMF3 cells		m6A-seq analysis identified 47, 539 and 38, 383 m6A peaks from 6,284 to 5,373 m6A-modified transcripts in BT-549 and BT-549LMF3 cells, respectively (Chen et al. 2021a). KRT7 and KRT7-AS showed increased expression, m6A deposition, and increased transcript stability by binding IGF2BP1/HuR and increased interaction of YTHDF1/eEF-1 with KRT7 mRNA to increase elongation in BT-549LMF3 cells (Chen et al. 2021a)
M6A-RIP-seq (Mouse mAb anti-m6A (Synaptic Systems; 202 011)-Illumina HiSeq4000	MYC-induced human mammary epithelial cells (HMECs); MDA-MB-231, MDA-MB-231-LM2, MCF-7, SKBR3 cells in biological duplicates		Unique and common m6A peaks were identified when comparing MDA-MB-231 vs. MCF-7 or SKBR3 cells and between MCF-7 and SKBR3 cells (Einstein et al. 2021a). GEO: GSE137258
Magna MeRIP™ m6A kit (Millipore) Illumina HiSeq X Ten platform	MDA-MB-231	+/− shMETTL3 or IGF2BP2	No information about data availability (Wan et al. 2022). .

Using MINT-seq, ~ 3–5 times more m6A peaks were identified in the nascent transcripts of MCF-7 cells compared to MeRIP-seq (Lee et al. 2021). The authors reported pervasive m6A signals on nascent RNAs in MCF-7 cells, including 2,207 enhancer RNAs (eRNAs) and 1,201 upstream antisense RNA (uaRNAs) (eRNAs and uaRNAs are non-coding RNAs whose expression is correlated with the activity of functional enhancers by modifying chromatin accessibility at promoters, stabilizing enhancer-promoter interactions, transcription factor-DNA interactions, and RNA Pol II stability (reviewed in (Cardiello, et al. 2020; Kim, et al. 2015; Li, et al. 2016a)). Approximately 18.8% of putative active eRNAs contained one or more m6A peaks (Lee et al. 2021). The m6A-marked transcripts had higher abundance, suggesting the m6A increased eRNA stability. Estradiol (E2, 100 nM for 24 h) did not significantly alter m6A peak levels in the 226 eRNAs induced by E2 in MCF-7 cells (Lee et al. 2021). The authors concluded that m6A modification is “largely hard wired to nascent sequences based on sequences, e.g., the RRACH motif” (Lee et al. 2021), as originally proposed (Darnell, et al. 2018). The RRACH motif f (R = A, G; H = A, C, U) (Jenjaroenpun et al. 2020) is recognized by the METTL3 complex (Knuckles and Bühler 2018; Licht and Jantsch 2016).

Conclusions and future directions

Abundant data supports the role of epitranscriptomic modification of mRNA and the readers, writers, and erasers of m6A in various diseases and cancers, including breast cancer. The number of publications on m6A in breast cancer is doubling each year (Figure 2) with 64 papers in 2022 at the time of submission of this review. Whether m6A is “hard wired” by the RRACH motif (Darnell et al. 2018) or modified depending on regulation of the METTL3 complex, other RNA binding proteins, RNA structure, intranuclear localization, e.g., condensates, and the epigenome remain to be examined in a cell- and tissue-type manner in breast cancer. How the “expososome”, including the microbiome, environmental chemical exposures, and endocrine- and metabolism- disrupting chemicals (Koual, et al. 2020; Koval, et al. 2022), affect these events to regulate the epitranscriptome in breast cancer is unknown. The advent of direct RNA sequencing and single cell RNA sequencing promises to yield new understanding of epitranscriptomics not only of m6A but less abundant modifications in breast cancer progression, endocrine therapy resistance, and metastasis. For example, identifying specific sites of m6A that impact on transcript stability and reader binding will be important in determining the role of m6A on endocrine-resistance in breast cancer as seen for HER2 resistance and GLUT4 expression (Liu et al. 2022a). Future studies to determine the mechanism by which gene- and position- specific m6A sites affects transcript stability and protein abundance, may employ inducible Crispr/dCAs13b-METTL3 and Crispr/dCas13b-ALKBH5 programmed to respectively methylate and demethylate specific m6A sites on a transcript (Li, et al. 2020; Shi, et al. 2022a; Zhang, et al. 2021c). The regulation and selectivity, role of specific cell types within a breast tumor and the tumor microenvironment, nuclear-mitochondrial or retrograde signaling of epitranscriptomic regulation of mRNA transcript processing and pathway regulation remains to be elucidated.

Figure 2: Number of publications on m6A AND breast cancer in PubMed by year.

Figure 1: Readers, writers, and erasers of m6A in mRNA regulate transcript fate.

Shown is the METTL3 m6A methyltransferase (writer) complex based on (Bawankar, et al. 2021). Readers are shown by their predominant subcellular location and roles in mRNA events, e.g., splicing, pri-miR-to-pre-miR-processing, nuclear export, mRNA stabilization or decay, translation initiation, elongation, or silencing. ALKBH5 is a specific m6A demethylase whereas there is controversy as to FTO’s specificity for m6A demethylation as described in the text. Created with BioRender.