YTHDF1 promotes breast cancer cell growth, DNA damage repair and chemoresistance

Chemoresistance represents a major obstacle to the treatment of human cancers. Increased DNA repair capacity is one of the important mechanisms underlying chemoresistance. In silico analysis indicated that YTHDF1, an m6A binding protein, is a putative tumor promoter in breast cancer. Loss of function studies further showed that YTHDF1 promotes breast cancer cell growth in vitro and in vivo. YTHDF1 facilitates S-phase entry, DNA replication and DNA damage repair, and accordingly YTHDF1 knockdown sensitizes breast cancer cells to Adriamycin and Cisplatin as well as Olaparib, a PARP inhibitor. E2F8 is a target molecule by YTHDF1 which modulates E2F8 mRNA stability and DNA damage repair in a METTL14-dependent manner. These data demonstrate that YTHDF1 has a tumor-promoting role in breast cancer, and is a novel target to overcome chemoresistance.


INTRODUCTION
Chemoresistance remains one of the major hurdles in the treatment of human malignant tumors including breast cancer. Enhanced DNA damage repair capacity endues tumor cells with resistance to chemotherapy and radiotherapy. Actually, DNA damage response (DDR) is employed by normal cells to maintain genomic integrity, and defect in DNA damage repair may lead to genomic instability and malignant transformation. In tumor cells, however, DDR deficiency or repression can be utilized to increase the sensitivity to DNA-damaging drugs. Therefore, targeting factors involved in DNA damage repair is an important strategy to overcome chemoresistance and radioresistance [1].
Here a novel function of YTHDF1 in human tumors was identified. YTHDF1 promotes DNA damage repair, especially DSBs HR repair, in a METTL14-dependent manner. Accordingly, YTHDF1 induces the resistance to DNA-damaging drugs Adriamycin and Cisplatin, and PARP inhibitor Olaparib.

RESULTS
In silico analysis indicates YTHDF1 as a potential tumor promoter in breast cancer Databases were used to analyze the potential role of YTHDF1 in breast cancer. In TCGA database, cancer tissues expressed higher level of YTHDF1 compared to normal tissues (Fig. 1A). In Curtis database, YTHDF1 was overexpressed in invasive ductal carcinoma, invasive lobular carcinoma, mucinous carcinoma, and medullary carcinoma, but not in benign breast neoplasm ( Supplementary Fig. 1). Higher YTHDF1 level was correlated with shorter overall survival (OS), recurrence-free survival (RFS), and distal metastasis-free survival (DMFS) in the TCGA database and Kaplan Meier plotter website (Fig. 1B, C). GSEA in TCGA database indicated a relation of YTHDF1 to cell cycle, DNA replication and DNA damage repair (Fig. 1D). Furthermore, GSEA revealed a relation of YTHDF1 to mRNA export and processing (Supplementary Fig. 2). Finally, YTHDF1 was positively correlated with factors related to cell cycle, DNA replication and DNA damage repair in CCLE database (Fig. 1E). These findings indicated that YTHDF1 functioned as a putative tumor promoter in breast cancer.

YTHDF1 induces breast cancer cell growth in vitro and in vivo
To verify the tumor-promoting role of YTHDF1, breast cancer cell lines MDA-MB-231, MCF7 and HS578T were selected to stably knockdown YTHDF1 expression ( Supplementary Fig. 3, Supplementary file 1). As shown by CCK8 and Colony formation tests, YTHDF1 knockdown remarkably suppressed cell growth in vitro ( Fig. 2A, B). To the contrary, YTHDF1 overexpression promoted cell growth ( Supplementary Fig. 4). YTHDF1 knockdown interfered with mammosphere formation, indicating that YTHDF1 contributed to the tumorigenic growth (Fig. 2C). MDA-MB-231 cells were inoculated into immune-deficient mice to evaluate whether YTHDF1 affects cell growth in vivo. Consistently, YTHDF1 knockdown impeded tumor growth significantly (Fig. 2D). Tumors with YTHDF1 knockdown had less cells with positive staining of phosphorylated Histone 3 (p-H3), a factor reflecting cell division, compared to those with control knockdown (Fig. 2E).
YTHDF1 promotes S-phase entry and DNA replication Based on the in silico analysis, the effect of YTHDF1 on the cell cycle and DNA replication was investigated. YTHDF1 knockdown in MDA-MB-231 cells reduced S-phase fraction while increased G1-phase fraction (Fig. 3A). This finding was supported by EDU staining. YTHDF1 knockdown diminished the fraction of EDU-positive cells (Fig. 3B). Mechanistically, YTHDF1 knockdown downregulated CDK2 and Cyclin E2 whereas upregulated P21, a CDK inhibitor involved in S-phase (Fig. 3C, Supplementary file 1). YTHDF1 knockdown downregulated a series of factors related to DNA replication including PCNA, RFC4, MCM4, GINS1, and POLE2 (Fig. 3C, D). Similar findings were observed in YTHDF1-knockdowned MCF7 cells ( Supplementary Fig. 5, Supplementary file 1). These data indicated that YTHDF1 promoted S-phase entry and DNA replication. In silico analysis indicates YTHDF1 as a potential tumor promoter in breast cancer. A YTHDF1 mRNA expression was investigated in TCGA database and compared between breast cancer samples (n = 1101) and normal mammary tissues (n = 113). M: median value. B The association of YTHDF1 with OS was analyzed in TCGA database. C The association of YTHDF1 with OS, RFS and DMFS was analyzed in Kaplan Meier plotter website. D GSEA in TCGA database was performed. YTHDF1-related enrichment plots were shown. E A heat map was generated from CCLE database indicating the correlation of YTHDF1 with factors related to cell cycle, DNA replication and DNA damage repair.
YTHDF1 enhances E2F8 mRNA stability dependent on METTL14 As an m6A "reader", the effect of YTHDF1 on E2F8 mRNA stability was evaluated. As expected, YTHDF1 knockdown significantly shortened E2F8 mRNA half-lives (Fig. 6A). RIP test further confirmed the binding of YTHDF1 to E2F8 mRNA in the predicted m6A sites (Fig. 6B). To investigate the involvement of METTL14 in this process, METTL14 was stably knockdowned in breast cancer cells (Fig. 6C, Supplementary file 1). As shown in Fig. 6D, METTL14 knockdown similarly reduced E2F8 mRNA stability. METTL14 knockdown interfered with the binding of YTHDF1 to E2F8 mRNA (Fig. 6E), and suppressed E2F8 expression in both mRNA and protein levels (Fig. 6F, Supplementary file 1). These data indicated that YTHDF1 impeded E2F8 mRNA decay dependent on METTL14. METTL14 is involved in DNA replication, DNA damage repair and chemoresistance As METTL14 coordinated with YTHDF1 to modulate E2F8, the involvement of METTL14 in YTHDF1-associated phenotypes was further clarified. METTL14 was positively correlated with factors related to cell cycle, DNA replication and DNA damage repair in the CCLE database (Fig. 7A). Consistently, METTL14 knockdown in MDA-MB-231 cells reduced cell growth (Fig. 7B, Supplementary Fig. 9B) and repressed DNA replication (Fig. 7C, D). METTL14-knockdowned cells exhibited more γ-H2AX staining after Adriamycin treatment, indicating that METTL14 knockdown interfered with DNA damage repair (Fig. 7E). As a consequence, METTL14 knockdown enhanced the sensitivity to Adriamycin, Cisplatin, and Olaparib (Fig. 7F). These data demonstrated that METTL14 promoted DNA replication, DNA damage repair, and chemoresistance.

DISCUSSION
Several lines of evidence indicate that YTHDF1 functions as a putative tumor-promoter in breast cancer. First, YTHDF1 is overexpressed in breast cancer tissues. Second, higher YTHDF1 expression is correlated with shorter survival. Third, YTHDF1 promotes breast cancer cell growth in vitro and in vivo. YTHDF1 facilitate S-phase entry and DNA replication through modulating a series of factors related to S-phase, including Cyclin E2, CDK2, P21 and PCNA. Notably, DNA replication is closely associated with DNA damage repair, and HR is active in S-phase of the cell cycle [17,18].
The link between DNA damage repair and m6A modification has been revealed by several recent studies. METTL3 is involved in nucleotide excision repair of UV-induced DNA damage [19]. Moreover, METTL3 and m6A binding protein YTHDC1 are responsible for DSBs HR repair in human sarcoma U2OS cells; METTL3 depletion enhances the sensitivity to DNA-damaging therapy [20]. The finding that YTHDF1 contributes to HR repair and chemoresistance in breast cancer further confirms the association of m6A modification with DNA damage repair. Mechanistically, YTHDF1 upregulates HR-related factors RAD51 and BRCA1, as well as BRCA2, BARD1, and PALB2 in a METTL14-dependent manner.
PARP inhibition is synthetic lethal with DSBs HR deficiency. Currently, PARP inhibitors are prescribed to patients with germline BRCA1/2 mutations [21]. In tumors with other HR defects such as germline PALB2 or somatic BRCA1/2 mutations, Olaparib is also effective [22]. However, PARP inhibitors are less effective on HRproficient tumors. YTHDF1 knockdown sensitizes HR-proficient cells to Olaparib, mainly due to the suppression of HR repair. Clearly, targeting HR repair is a potential strategy to extend the use of PARP inhibitors to HR-proficient tumors [23][24][25].
E2F8 functions as a tumor promoter in breast cancer by promoting cell proliferation [26,27]. As a downstream target of YTHDF1-METTL14, E2F8 was first identified to be involved in DNA damage repair and chemoresistance in breast cancer. Actually, the role of E2F8 in DNA damage repair remains largely unknown in human tumors. In a study using U2OS cells, E2F8 is shown to be required for the cell-cycle response to DNA damage [28]. Unlike E2F8, E2F1, another member in the E2F family, has been demonstrated to play an important role in DNA damage repair [29].
In summary, YTHDF1 for the first time was uncovered to promote S-phase entry, DNA replication and DNA damage repair. YTHDF1 is not only a tumor promoter but also a target to overcome chemoresistance in breast cancer. Furthermore, the role of YTHDF1 in DNA replication and repair is dependent on METTL14 and mediated by E2F8.

In silico analysis
The Cancer Genome Atlas (TCGA) database was interrogated for YTHDF1 mRNA expression in breast cancer samples (n = 1101) and normal mammary tissues (n = 113). YTHDF1 expression in Curtis database was investigated in various breast tumors and normal mammary tissues in Oncomine (https://www.oncomine.org/). The correlation of YTHDF1 with survival was analyzed in the TCGA database and Kaplan-Meier plotter website (http://kmplot.com/analysis/). GSEA was conducted in TCGA database. The samples were divided into high and low groups according to gene expression. Gene pathways differentially expressed between high and low groups were analyzed. The Cancer Cell Line Encyclopedia (CCLE) database was interrogated for gene expression in human breast cancer cell lines. Correlation between two genes was analyzed by Pearson statistics.

In vivo animal study
Female BALB/c nude mice (5~6 weeks of age, 18-20 g) were purchased from Weitong Lihua (Beijing, China). All animals were dealt with according to the Animal Ethics Committee of China Medical University. Before tumor cell inoculation, mice were randomized into different groups (five in each group). 1×10 6 MDA-MB-231 cells were resuspended in 100 μl PBS with 50% Matrigel (Corning Costar, USA), and injected into the mammary fat pad of the mice. The length and width of tumors were measured with a vernier caliper every week. Tumor volume was calculated by the formula: V = 1/2×length×width 2 . The investigator was blinded to the group allocation of the animals during the experiment. No statistical method was used to predetermine the sample size for the xenograft mice experiment, which was based on previous experimental observations. The sample size of each experiment is shown in the legend. No data were excluded from the analysis.

Immunohistochemistry
The tissues were first fixed in 4% paraformaldehyde for 72 h, and then dehydrated. Next the tissues were embedded in paraffin and then sliced into 4 μm sections. The sections were deparaffinized and hydrated with xylene and gradient alcohol, respectively. Three percentage H 2 O 2 was used to eliminate endogenous peroxidase activity. The sections were further repaired with citrate buffer, and then blocked by BSA. Subsequently, the sections were incubated with primary antibody (anti-phosphorylated Histone 3, ThermoFisher, PA5-17869, USA, 1:200) overnight at 4°C, incubated with goat anti-rabbit IgG and streptavidin peroxidase (SP) complex at 37°C for 30 min, and stained with DAB reagent. At last, the sections were re-stained with hematoxylin, dehydrated with gradient alcohol, mounted, and photographed under a microscope (LEICA DM2500 LED). Phosphorylated Histone 3-positive cells in five randomly selected fields were counted.
Cell cycle assay EDU staining 5×10 5 cells were trypsinized, counted, and seeded into 24-well culture plates for 24 h. After addition of 50 μM EDU (Ribobio, China), cells were incubated for another 2 h. Then cells were washed with PBS, fixed with 4% paraformaldehyde for 30 min and permeabilized with 0.5% Triton X-100 for 10 min. Subsequently, cells were treated with 200 μl 1×Apollo reaction cocktail for 30 min at room temperature in the dark. The DNA contents were stained by 200 μl 1×Hoechst 33342 for 30 min at room temperature in the dark. Cells were observed under a laser scanning confocal focus microscope (FV-1000; Olympus). Positive cells in random five fields were counted.

Western blot
Cells were lysed by RIPA lysis with 1% PMSF on ice for 1 h, and centrifuged with 12,000× g at 4°C for 40 min. The protein concentration was determined by a BCA protein assay kit (Wanleibio, China). 30 μg protein was separated on 10% SDS-PAGE gel and transferred into a polyvinylidene difluoride (PVDF) membrane in a wet electron transfer device. 5% skimmed Apoptosis assay 5×10 5 cells were harvested and washed with PBS. Following the standard protocol, an Annexin V-FITC/PI Apoptosis Kit (KeyGEN BioTECH, China) was used to analyze cell apoptosis. Cells were incubated in 500 μl binding buffer containing 5 μl Annexin V-FITC and 5 μl PI at room temperature in the dark for 15 min. The number of apoptotic cells was measured by an Accuri C6 Plus Flow Cytometer (BD).

mRNA stability assay
Cells were harvested and seeded into 6-well culture plates for 24 h. After addition of 5 μg/ml actinomycin D (Sigma, A9415), cells were cultured for 3 h or 6 h, and collected. Total RNA was extracted for reverse transcription and Real-time PCR.
RNA immunoprecipitation (RIP) assay 5×10 6 cells were harvested and washed with cold PBS. According to the standard instructions of the Imprint RNA RIP kit (Sigma), cells were lysed with the lysis buffer in −80°C freezer overnight. 5 μg antibody (anti-YTHDF1: Cell Signaling Technology, #86463, USA, 1:50) was prebound to 20 μl magnetic beads in RIP wash buffer with rotation for 30 min at room temperature, and then the cell lysates were added at 4°C with rotation overnight. Then RNA was extracted and dissolved with RNase-free water. The enrichment of certain fragments was determined by real-time PCR. Primer used for E2F8 quantification was designed as follows: forward: 5'-AGTGCTTTGTATCTTTAAGGAAGCCC-3', reverse: 5'-AGCAGTAAAGTCGTGG-GAGGT-3'. IgG was used as the negative control.

Statistical analysis
All cell experiments were performed triplicate. The data are expressed as mean ± SD. GraphPad prism 8 was used to analyze the data. Differences were analyzed by two-sided Student's t test or two-way ANOVA when the variance is similar between the groups. P value < 0.05 was considered statistically significant.

DATA AVAILABILITY
The published article includes all data sets generated/analyzed for this study.