METTL3 Depletion Contributes to HR+/HER2- Breast Cancer Progression and Drug Resistance via m6A Modication of Constituents of the CDKN1A/EMT and BAX/caspase-9/-3/-8 Signalling Pathways

Chemotherapy is an important strategy for the treatment of hormone receptor positive / human epidermal growth factor receptor 2 negative (HR+/HER2-) breast cancer (BC), but this subtype has a low response to chemotherapy. Growing evidence indicates that N6-methyladenosine (m6A) is the most common RNA modication in eukaryotic cells and that methyltransferase-like 3 (METTL3) participates in tumour progression in several cancer types. Therefore, exploring the function of METTL3 in HR+HER2- BC initiation and development is still signicant.

ovarian suppression) are the gold standard of treatment of BC and the backbone of adjuvant therapies for these patients with signi cantly decreased risk of recurrence and death(4), up to 20% of patients will eventually relapse (5,6). Single-agent chemotherapy is an essential treatment option for endocrineresistant or treatment-refractory disease (7). However, response rates to these therapies are low. Reported progression-free survival (PFS) ranges from 4.0 to 6.3 months with second-line chemotherapy and from 2.4 to 5.5 months with third-line chemotherapy(8). Similarly, the association of pathological complete response (pCR) with disease-free survival (DFS) or overall survival (OS) in HR+/HER2− BC following neoadjuvant systemic therapy is relatively low compared to that of the other two subtypes of BCs (9).
Thus, there is an imperative need to further improve our insight into the tumour biology and the tumour cell response to chemotherapy for HR+/HER2− BC.
N 6 -methyladenosine (m6A) is the most prevalent messenger RNA (mRNA) modi cation in eukaryotes and is added to mRNA molecules by the N 6 -adenosine methyltransferase complex, which consists of methyltransferase-like 3 (METTL3), methyltransferase-like 14 (METTL14) and Wilms tumour 1 associated protein (WTAP) (10). METTL3 and METTL14 are two active methyltransferases that form a heterodimer to catalyse m6A RNA methylation, while WTAP interacts with this complex and substantially affects mRNA methylation (11). Two m6A demethylases fat mass and obesity-associated (FTO) and AlkB homolog 5 (ALKBH5) have been discovered since 2011, revealing the dynamic nature of m6A modi cation (12). Some cellular proteins have been found to preferentially bind m6A-modi ed RNA, whereas others been characterized to speci cally recognize m6A-modi ed mRNA and accelerate the decay of the mRNA (13). These results indicated that this chemical modi cation is common and important in a variety of biological processes.
With the elucidation of the mechanisms involved in m6A modi cation, a recent report described the role of the m6A modi cation in multiple tumours (14). Although studies on the function of m6A in BC are still in their early stages, there is growing evidence showing that m6A plays a critical role in many aspects of BC, including tumorigenesis (15), metastasis(16), prognosis (17), and treatment resistance(18). Reduced expression of METTL3 promotes metastasis of BC by increasing COL3A1 expression (19). Hypoxia stimulates ALKBH5, which stabilizes NANOG mRNA and induces a phenotype associated with BC stem cells (BCSCs) and lung metastasis (20,21). YTHDF3 promotes BC metastasis to the brain by inducing m6A-enriched gene translation(16). Moreover, m6A modi cation patterns have therapeutic implications and correlate with drug resistance. HNRNPA2B1 and METTL3 overexpression in MCF-7 cells reduces their sensitivity to 4-hydroxytamoxifen and/or fulvestrant (22,23). In triple-negative BC (TNBC) cells, IGF2BP3 promotes chemoresistance to doxorubicin (DOX) and mitoxantrone by regulating ABCG2 expression (24). Therefore, it is possible to determine why HR+/HER2− BC is insensitive to chemotherapy treatment by further exploring the role of m6A modi cation in this subtype of BC.
In this study, we investigated the potential effect of m6A methylation on the sensitivity of HR+/HER2− BC to chemotherapy. Our data revealed that chemotherapy decreased the levels of the m6A modi cation, which was dependent on METTL3 expression. We demonstrated that METTL3 facilitates HR+/HER2− BC progression via its downstream target cyclin-dependent inhibitor kinase 1A (CDKN1A), which mediates epithelial-mesenchymal transition (EMT). In addition, METTL3 regulated BAX/caspase3/8/9 signalling in an m6A-independent manner. Overall, our results suggested that METTL3 plays multifunctional roles in the progression of HR+/HER2− BC, indicating that METTL3 is a promising biomarker for predicting the e cacy of chemotherapy as well as a potential therapeutic target for reversing chemotherapy resistance in HR+/HER2− BC.

Methods
Human BC tissues and cell lines Thirteen pairs of primary BC tissues collected before and after treatment were obtained from the Second Xiangya Hospital of Central South University (Hunan, China) from August 2020 to March 2021. All individuals with BC were diagnosed for the rst time, only received chemotherapy prior to surgery, and had histologically con rmed BC. All patients provided written informed consent, which was conducted in accordance with the Declaration of Helsinki. BC cell lines (MCF-7 and T47D) were obtained from the Shanghai Type Culture Collection of the Chinese Academy of Sciences and were grown in RPMI 1640 medium (Gibco, Carlsbad, CA, USA) supplemented with 1% penicillin/streptomycin (Shanghai, Beijing, China). All cells in this study were incubated in 37°C incubators with 5% carbon dioxide and routinely tested for mycoplasma.

Cell transduction
Stable knockdown and overexpression of METTL3 were achieved with lentiviral-based delivery of shorthairpin RNA (shRNA) and overexpression vectors, respectively. The shRNA sequences were subcloned into a lentiviral expression vector containing GFP by Shanghai Genechem Co., Ltd. (Shanghai, China).
Lentiviral transduction was performed according to the manufacturer's instructions. All constructed vectors were veri ed by DNA sequencing.

Western blotting
Total proteins from cell lines and tissues were extracted with RIPA buffer and then quanti ed by BCA analysis. Subsequently, 20 µg of total protein per sample (10 µL per lane) was separated using sodium lauryl sulfate-polyacrylamide gel electrophoresis (10% polyacrylamide gel) before the proteins were transferred to a PVDF membrane. After incubation with primary antibodies overnight, the membranes were then incubated with secondary antibody. Finally, target protein bands were detected using a chemiluminescence system. The antibodies used targeted the following proteins: AKT (Cell Signaling Quantitative real-time polymerase chain reaction (qRT-PCR) Total RNA from cell lines and tissues was isolated using TRIzol reagent (Thermo Fisher Scienti c, Beijing, China). A PrimeScript RT kit (Thermo Fisher Scienti c, Beijing, China) was used for cDNA synthesis. Realtime quantitative PCR analysis was performed using a SYBR Premix Kit (Abclonal, Wuhan, China). Each sample was run in triplicate, and the expression levels were normalized to those of GAPDH using relative quantitative methods. All PCR primers (Well Biological Science, China) are listed as follows: Cells were seeded into 96-well plates at a density of 1 × 10 3 cells per well and cultured in an incubator (37°C with 5% CO2) for 24 h, 48 h, and 72 h, after which cell proliferation was examined on a microplate reader by measuring the absorbance at a wavelength of 450 nm. To assess colony formation, cells were seeded into 6-well plates at a density of 1 × 10 3 cells per well and cultured in an incubator (37°C with 5% CO2). After 14 days, the cells were xed and stained with Giemsa stain, modi ed solution (Sigma), and colonies containing 50 or more cells were counted.

Cell cycle analysis
For cell cycle analysis, cells were seeded in 6-well plates at a density of 2-3 ×10 5 per well and grown tõ 70% con uence. Cells were then harvested and suspended in complete medium. The cell suspension was centrifuged at 1300 rpm at 4°C, washed once in D-Hanks buffer, counted and resuspended to a density of 3-6 ×10 6 cells/ml. Ice-cold 70% ethanol was added dropwise to x the cells, which were then centrifuged, washed once with D-Hanks, stained with 20 µg/ml propidium iodide (Sigma, P4170) in D-Hanks containing 50 µg/ml RNase A (Fermentas, EN0531) for 1 h at room temperature and analysed using ow cytometry.

Wound healing assay
For the wound healing assay, cells were seeded and cultured until a 90% con uent monolayer was formed. Cells were then scratched by a sterile pipette tip and treated as indicated in the text in FBS-free medium. Cell migration distances into the scratched area were measured in 10 randomly chosen elds under a microscope.

TUNEL assay
A TUNEL kit (Roche) was used according to the manufacturer's instructions. In brief, cells were xed with 4% paraformaldehyde, permeated with 0.1% Triton X-100 in PBS, incubated with 50 µl of TUNEL reaction mixture for 1 h at 37°C, and then washed with PBS 3 times. A total of 1,000 cells were counted, and the percentage of apoptotic cells was quanti ed.

Immunohistological staining
Fresh tumour tissues were excised from nude mice and xed with 4% paraformaldehyde. Then, the sections were dehydrated with a graded alcohol series, cleared in xylene, embedded in para n, and sliced into sections with a microtome at a thickness of approximately 4 mm. After more treatments with xylene and an alcohol gradient, the sections were dewaxed and hydrated. Ki-67 antibody was added and then incubated overnight, and secondary antibody was incubated for 1 h. Diaminobenzidine (DAB) was used for colour development, and the slides were stained with haematoxylin before they were mounted in neutral resin. With a common microscope, 5 40× high-de nition elds were randomly selected for counting the total number of cells and the number of Ki-67-positive cells (which have brown nuclei).
In vivo tumour xenograft model

Bioinformatics analysis
The expression pro les of m6A-related mRNAs were obtained from the GSE87455 and GSE763 datasets (https://www.ncbi.nlm.nih.gov/geo/), which comprised 69 pairs of pre-and post-chemotherapy BC tissues and MCF-7 cells treated with DOX, respectively. Differential expression of m6A-related genes between the pre-treatment and post-treatment tissues were determined by R software (http://www.rproject.org/). The association of METTL3 expression with patient OS was analysed with Kaplan-Meier Plotter (http://kmplot.com/analysis/index.php?p=service&cancer=pancancer_rnaseq) based on the TCGA data. Gene ontology (GO) analysis and Kyoto Encyclopedia of Genes and Genomes (KEGG) enrichment analysis of differentially expressed genes (DEGs) were performed using the DAVID database (https://david.ncifcrf.gov/).

Statistical analysis
Experimental data are presented as the mean ± standard deviation (SD) and were analysed using GraphPad Prism 8.0 software. All in vitro results are representative of at least three independent trials. Two-group comparisons were assessed by the Mann-Whitney U test or Student's t test, and paired t tests were employed for paired BC and corresponding chemo-only BC samples. A two-tailed p-value of 0.05 was considered statistically signi cant.

Results
Chemotherapy-induced decreases in METTL3 expression correlated with a poor prognosis in HR+HER2-BC To identify the crucial regulator of m6A modi cation in the progression of HR+HER2-BC, we screened the expression pro les of m6A writers (METTL3, METTL14, and WTAP) and m6A erasers (FTO and ALKBH5) in the GSE87455 platform. The results of the analysis showed that METTL3 was reduced and FTO was signi cantly increased in the chemo-only group, whereas METTL14, WTAP and ALKBH5 expression levels showed no signi cant differences between the untreated specimens and chemo-only tumour specimens ( Figure 1A). We also examined these m6A-related genes in the GSE763 dataset (Additional le 1A), and only the pattern of METTL3 expression was consistent with that in the GSE87455 cohort. Next, we detected METTL3 expression in 13 pairs of pre-and post-chemotherapy HR+HER2-BC tumour tissues and observed that METTL3 expression was de cient in the post-chemotherapy tissues compared to the pre-treatment tissues ( Figure 1B, C). In addition, the downregulation of METTL3 in tissues after chemotherapy was consistent with the expression pattern in the MCF-7/ADR cell line ( Figure 1D). MCF-7 and T47D cell lines treated with chemotherapeutic drugs, including DOX ( Figure 1E, F), paclitaxel (PTX) (Additional le 1B) and cisplatin (DDP) (Additional le 1C), corroborated these ndings. RT-PCR and western blot con rmed that METTL3 expression was reduced at the mRNA and protein levels in the DOXtreated group. Furthermore, individuals with HR+HER2-BC and low METTL3 expression had a worse prognosis, and METTL3 expression was correlated with recurrence-free survival (RFS) ( Figure 1G, H).

METTL3 depletion attenuated the drug sensitivity of HR+/HER2-BC cells by promoting tumour proliferation and metastasis and inhibiting apoptosis
We rst detected the global m6A level in BC tissues and cell lines using an m6A RNA methylation quanti cation kit. The levels of m6A-modi ed RNA were remarkably higher in untreated BC tissues than in  Figure  2E, Additional le 2C). The levels of m6A modi cation were decreased upon METTL3 inhibition ( Figure  2F). Notably, METTL3 knockdown obviously enhanced cell proliferation but reduced the sensitivity of MCF-7 cells to DOX as shown in the CCK-8 assays ( Figure 2G, H). As anticipated, METTL3 knockdown enhanced colony formation, especially in the DOX-treated group ( Figure 2I), whereas METTL3 overexpression evoked the opposite effects ( Figure 2J). Furthermore, a cell cycle assay demonstrated that silencing METTL3 prominently increased the proportion of cells in S phase ( Figure 2K). However, METTL3 upregulation did not obviously reduce the proportion of S phase cells ( Figure 2L). Similarly, we also observed a protective effect of METTL3 in T47D cells ( Figure 2SD-I).
Cancer metastasis is a critical characteristic of drug resistance in individuals with BC. A wound healing experiment was utilized to assess the effects of METTL3 on BC cell mobility. The results demonstrated that depletion of METTL3 induced, but overexpression of METTL3 weakened, MCF-7 and T47D cell migration in vitro ( Figure 3A, Additional le 3A). Moreover, an increase in the number of apoptotic cells was observed among METTL3-overexpressing MCF-7 and T47D cells ( Figure 3B, Additional le 3B). These results were further substantiated by the in vivo results. The same numbers of METTL3overexpressing and control MCF-7 cells were injected into immunocompromised nude mice.
Overexpression of METTL3 suppressed tumour cell seeding, which led to a lower average tumour volume and weight ( Figure 3C-E). Immunohistochemistry (IHC) con rmed that the expression of Ki-67, an indicator of tumour growth, was lower in the tumours derived from the METTL3-overexpressing cells ( Figure 3F). Taken together, these ndings suggest the anti-proliferative and migratory roles of METTL3 in BC cells.
Multidimensional sequencing identi es CDKN1A as a downstream target of METTL3 To better understand the possible mechanism by which METTL3 participates in HR+HER2-BC progression, we rst performed DEG analysis of the GSE87455 and GSE763 datasets and identi ed 4049 and 1783 DEGs in these two cohorts, respectively. There were 498 overlapping genes altered by chemotherapeutic drugs in both datasets, among which cyclin-dependent kinase inhibitor 1A (CDKN1A), DNA topoisomerase II alpha (TOP2A) and cell division cycle 20 (CDC20) were ltered by the criteria |log 2 FC|≥1 and p-value < 0.05 ( Figure 4A). GO and KEGG enrichment analysis revealed that the DEGs were mostly linked to EMT, transcriptional activation of cell cycle inhibitor p21 and transcriptional activation of p53 responsive genes, indicating a regulatory role of METTL3 in the chemoresistance of HR+HER2-BC ( Figure 4B, C). Because CDKN1A exhibited the most distinct differential expression, we chose this gene as a candidate target for subsequent experiments. RT-PCR and western blot experiments con rmed the positive regulation of METTL3 on CDKN1A at the mRNA and protein levels ( Figure 4D; E). Additionally, MeRIP-qPCR using a m6A-speci c antibody showed that the levels of m6A-modi ed CDKN1A were decreased in METTL3-silenced MCF-7 cells ( Figure 4F). These data suggest that METTL3 regulates CDKN1A mRNA expression via m6A methyltransferase activity.
EMT plays a critical role in the drug resistance and metastasis of tumours. We also found that DOX induced a decrease in the expression of E-cadherin and an increase in the expression of N-cadherin and vimentin in MCF-7 and T47D cells ( Figure 4G) indicative of EMT. Similarly, inhibition of METTL3 resulted in increased protein levels of N-cadherin and vimentin, whereas overexpression of METTL3 showed the opposite trends ( Figure 4I, J). Furthermore, pseudopodia and cell spacing in MCF-7 cells were increased in response to METTL3 knockdown ( Figure 4H). Collectively, these results suggested the importance of METTL3-mediated EMT signalling in chemoresistance.

METTL3 regulates EMT in a PI3K/AKT signalling-dependent manner
Despite the observed critical role of METTL3 in the EMT process, whether this activity is speci cally attributed to the METTL3/CDKN1A axis needs to be further explored. HR+HER2-BC is insensitive to chemotherapy, which may be largely related to the mechanism of cellular feedback regulation to drug exposure. There is often a mutation in the PIK3CA gene in HR+HER2-BC, which plays a crucial role in regulating the PI3K/AKT pathway and drug resistance. However, the precise role of DOX-induced PI3K/AKT pathway activity has not been documented. Therefore, we rst determined the effect of DOX intervention on the levels of p-AKT and total AKT. We observed increased levels of p-AKT in MCF-7 and T47D cells treated with DOX, as evidenced by the western blot experiments ( Figure 5A, B). We also assessed the potential effects of METTL3 on the PI3K/AKT pathway. The data showed that METTL3 depletion resulted in upregulation of p-AKT levels. Conversely, MCF-7 cells with stable overexpression of METTL3 showed restrained p-AKT levels compared with those of the other groups ( Figure 5C, D).
Subsequently, we tested the in uence of METTL3 and PI3K/AKT on the EMT pathway using the PI3K inhibitor LY294002. The data showed that the increased N-cadherin and p-AKT levels associated with METTL3 de ciency could be partially attenuated by treatment with LY294002 ( Figure 5E), highlighting the essential role of PI3K/AKT in the METTL3-controlled EMT process. Additionally, we conducted bibliometric analysis and found that CDKN1A is one of the key molecules in the PI3K/AKT pathway (Additional le 3C). Genetic upregulation of CDKN1A also impaired the effects of METTL3 inhibition on p-AKT activation ( Figure 5F). Our results clarify the critical role of CDKN1A and AKT in METTL3-mediated EMT progression in HR+HER2-BC.
BAX activates caspase3/8/9 to facilitate METTL3-mediated tumour cell apoptosis We demonstrated that METTL3 promoted apoptosis ( Figure 3B, Additional le 3B). Given the involvement of the apoptosis signalling pathway in tumour proliferation and metastasis, western blotting was implemented to visualize changes in the expression of apoptosis-related proteins following METTL3 knockdown or overexpression in MCF-7 and T47D cells. We found that increasing METTL3 expression elevated the levels of multiple apoptosis-related proteins, including caspase3, caspase8 and caspase9.
By contrast, knockdown of METTL3 appreciably decreased the expression of these proteins ( Figure 6A, B). BAX is an important protein that promotes apoptosis. As an upstream gene of caspase3, BAX regulates the activation of the caspase family. To further verify the relationship between METTL3 and apoptosis, the expression of apoptosis-related genes was detected at the mRNA and protein levels. We observed that increased METTL3 expression could facilitate the expression of BAX in MCF-7 cells and T47D cells ( Figure 6C, D). Notably, MeRIP-qPCR con rmed the interaction between BAX mRNA and m6A in METTL3-depleted MCF-7 cell lines ( Figure 6E). Therefore, we demonstrate that METTL3 increases BAX/caspase3/8/9 expression, promotes apoptosis and restrains tumorigenesis in HR+HER2-BC cells.

Discussion
In recent decades, substantial improvements have been made in therapeutic interventions for BC, which have increased the survival and quality of life of patients (25). Chemotherapy, as the broadest application of tumour treatment, remains the cornerstone of adjuvant therapy for BC and is widely used in BC patients with high metastatic burden and local advanced disease(26). Nevertheless, recent studies have indicated that the response rate of BC patients with the HR+/HER2-subtype to chemotherapy is low (27). Hence, the demand for elucidating the mechanism of HR+/HER2-BC insensitivity to chemotherapy is crucial. In this study, we rst discovered that m6A modi cations and METTL3 expression were inhibited by chemotherapy; thus, we evaluated the function of METTL3 in regulating HR+/HER2-BC progression, metastasis, and drug resistance. Based on our ndings, we de ned the METTL3/CDKN1A/EMT and METTL3/BAX/caspase3/8/9 axes as novel pathways involved in a potential mechanism of HR+HER2-BC chemoresistance ( Figure 6F). m6A modi cation, one type of RNA epigenetic modi cation, has been identi ed on almost all types of RNAs and has been implicated in a variety of cellular processes, including mRNA stability, splicing, location, and translation, RNA-protein interactions, and pri-miRNA processes(28-33). An increasing number of studies have addressed the pathological signi cance of m6A dysregulation in human diseases, especially in cancers(28-30). The results of our current study showed that the overall level of m6A modi cation was signi cantly downregulated after chemotherapy in HR + /HER2 − BC patients, and treatment of MCF-7 and T47D cells with DOX, PTX and DDP also resulted in a decrease in m6A modi cations. However, the levels of m6A modi cation were not affected by drug intervention in MDA-MB-231 cells. Therefore, our results suggested that chemotherapy-induced changes in m6A levels are a biological difference between HR+HER2-BC and TNBC, especially in terms of responsiveness to chemotherapy. Various studies indicate that the m6A modi cation affects drug sensitivity by regulating ABC transporters either directly at the transcript level or via upstream signalling pathways (31). Recent studies also indicated that the m6A modi cation is involved in the maintenance of CSCs in tumours, leading to drug resistance and recurrence (32). It has also been shown that m6A modi cations can affect the response of BC to endocrine therapy (22). However, there are few studies on the relationship between m6A and chemotherapy response. Therefore, considering the potential role of the m6A RNA modi cation in the development of chemoresistance, it is necessary to illustrate the relationship between these two phenomena.
METTL3, a key component of the N6-methyltransferase complex, has been reported to play an important role in many tumour types (33)(34)(35)(36)(37)(38). Previous studies reported that METTL3 plays an oncogenic role in acute myeloid leukaemia through diverse downstream targets(38), whereas other studies suggested that either increased or decreased METTL3 expression could promote the self-renewal and tumorigenicity of glioma stem-like cells, respectively (37,39). Regarding METTL3 in BC, data from the literature have suggested that METTL3 can promote BC progression by targeting Bcl-2, HBXIP or SOX2(36, 40,41) and that METTL3 could promote adriamycin resistance by accelerating pri-miRNA-221-3p maturation (42).
However, our results illustrated that chemotherapy-mediated depletion of METTL3 plays a signi cant protective role in tumour progression and drug tolerance. Reasonable explanations for these contradictory phenomena could be attributed to recognition by different m6A readers (43). We speculated that the m6A modi cation and METTL3 expression protect some critical genes from degradation or restrain the role of oncogenes by enhancing their recognition by "readers". However, this hypothesis needs more study. In summary, the decreased METTL3 expression is secondary to chemotherapy, which is consistent with the clinical medication pattern, and HR+HER2-BC is the only BC subtype to exhibit this expression pattern. Therefore, METTL3 can be used as a biomarker to predict the sensitivity of HR+HER2-BC to chemotherapy and as a novel target for combination therapy to reverse chemotherapy resistance.
Our results further showed that METTL3 regulates the proliferation, apoptosis, migration and drug resistance of HR + /HER2 − BC through multiple signalling pathways. On the one hand, METTL3 can affect the m6A modi cation of BAX mRNA, thereby promoting activation of the pro-apoptotic caspase cascade and (consequently) apoptosis. Apoptosis is an important mechanism to mitigate the uncontrolled growth of tumour cells and is mainly regulated by the Bcl2 protein family (44). The Bcl2 protein family can be divided into two categories according to their functions: one plays a pro-apoptotic role and includes BAX and Bak, whereas the other plays an anti-apoptotic role and includes Bcl2. Both pathways promote caspase cascades that eventually lead to cell death (44). Our experiment found that METTL3 can promote the expression of BAX and the subsequent activation of caspase3, 8, and 9, leading to apoptosis.
On the other hand, downregulation of METTL3 can regulate CDKN1A expression to affect the EMT process and promote cell proliferation. CDKN1A is one of the key molecules involved in cell cycle progression and was rst identi ed as a tumour suppressor (45). Later, it was found to be involved in pathways related to tumorigenesis and development, such as cell death, DNA replication/repair, gene transcription and cell motility(46). It is believed that the dual role of CDKN1A depends on its cellular localization (47). When in the nucleus, CDKN1A functions a tumour suppressor. However, when CDKN1A is concentrated in the cytoplasm, p53-impaired or p53-de cient cells may acquire carcinogenic properties, which may inhibit apoptosis and promote cell migration and proliferation(48). Studies have shown that miR-33b-3p can promote the survival and cisplatin resistance of A549 human lung cancer cells by targeting CDKN1A after DNA damage (49). It was also found that miR-520g mediated the resistance of colorectal cancer cells to  or oxaliplatin by downregulating of CDKN1A expression (50). These studies suggest that the presence of CDKN1A protects cancer cells from apoptosis after anti-cancer therapy. Therefore, in this study, the changes in CDKN1A expression were caused by chemical drugs, which may stimulate the translocation of CDKN1A protein from the nucleus to the cytoplasm, thereby activating downstream related pathways to reduce the sensitivity of cells to chemotherapy drugs; however, the speci c mechanism is still unknown. In short, the mechanisms of interaction between cell signalling pathways and epigenetic elements are diverse and complex and merit further exploration and veri cation.
This study still has some shortcomings, such as the small sample size and lack of follow-up data. The direct intermolecular regulatory mechanism by which METTL3 affects tumour progression and survival was not clari ed in detail and we will be further studied in the future.

Conclusion
Our data suggested the importance of chemotherapy-induced alterations in m6A modi cations and METTL3 expression in HR+HER2-subtype BC, providing a theoretical basis for METTL3 as a new predictor of chemotherapy response and target for drug therapy.