MYC and DNMT3A‐mediated DNA methylation represses microRNA‐200b in triple negative breast cancer

Abstract Triple‐negative breast cancer (TNBC) is the most aggressive breast cancer subtype with a poor prognosis. The microRNA‐200 (miR‐200) family has been associated with breast cancer metastasis. However, the epigenetic mechanisms underlying miR‐200b repression in TNBC are not fully elucidated. In this study, we found that MYC proto‐oncogene, bHLH transcription factor (MYC) and DNA methyltransferase 3A (DNMT3A) were highly expressed in TNBC tissues compared with other breast cancer subtypes, while miR‐200b expression was inhibited significantly. We demonstrated that MYC physically interacted with DNMT3A in MDA‐MB‐231 cells. Furthermore, we demonstrated that MYC recruited DNMT3A to the miR‐200b promoter, resulting in proximal CpG island hypermethylation and subsequent miR‐200b repression. MiR‐200b directly inhibited DNMT3A expression and formed a feedback loop in TNBC cells. MiR‐200b overexpression synergistically repressed target genes including zinc‐finger E‐box‐binding homeobox factor 1, Sex determining region Y‐box 2 (SOX2), and CD133, and inhibited the migration, invasion and mammosphere formation of TNBC cells. Our findings reveal that MYC can collaborate with DNMT3A on inducing promoter methylation and miR‐200b silencing, and thereby promotes the epithelial to mesenchymal transition and mammosphere formation of TNBC cells.


| INTRODUCTION
Breast cancer is a kind of heterogeneous disease, being classified into distinct subtypes based on the molecular markers variants. Triple-negative breast cancer (TNBC), characterized by the absence of the oestrogen receptor (ER), the progesterone receptor (PR), and the human epidermal growth factor receptor 2 (HER2) expression, was known as the most aggressive breast cancer subtype. 1 TNBC patients have a higher risk of distant metastasis and a poorer overall survival than other breast cancer patients, partly due to lacking effective targeted therapies. 2 Despite the great effort made on researching TNBC, the molecular mechanisms underlying TNBC aggressive behaviour are still obstacles to TNBC treatment.
MicroRNAs (miRNAs) are small non-coding RNAs that function through binding to the 3′-untranslated region (3′-UTR) of target mRNAs, resulting in mRNA degradation or translation inhibition. [3][4][5] Increasing evidence has revealed that deregulated miRNAs are involved in human carcinogenesis and cancer stem-like cells by interacting with the downstream genes. 4 11 MiR-200 family members inhibit epithelial-to-mesenchymal transition (EMT) by directly targeting zinc-finger E-box-binding homeobox factor 1 (ZEB1) and 2 (ZEB2). [12][13][14] Recent reports have demonstrated that miR-200 expression is inhibited by several transcription factors. 11 ZEB1 and ZEB2 repress the miR-200 promoter activity by binding to the E-box elements, suggesting a feedback loop with miR-200. [15][16][17] Apart from transcription factors aforementioned, epigenetic mechanisms, including DNA methylation as well as histone methylation and acetylation are all implicated in miR-200 silencing. [18][19][20] For example, DNA methyltransferase 1 (DNMT1) and EZH2 mediated DNA methylation and H3K27 trimethylation silence miR-200b/200a/429 gene and contribute to the progression of gastric cancer and glioblastoma. 21 However, few studies have focused on the association of miR-200b repression with DNMT3A-mediated DNA methylation, which is crucial for digging out the mechanisms of DNA methylation in regulating miRNAs. 22 Previous studies have demonstrated that the activation of MYC proto-oncogene, bHLH transcription factor (MYC) gene contributes to cancer genesis. 23,24 However, the roles of MYC in epigenetically regulating miRNAs are largely unclear. In this study, we aimed to reveal the mechanism of MYC and DNMT3A inducing miR-200b repression, which was involved in EMT and mammosphere formation of TNBC cells.  instructions. β-actin and U6 were used as internal control for MYC, DNMT3A and miR-200B respectively. The relative expression levels of the target genes were analysed using the comparative threshold cycle (2 −ΔΔCt ) method as previously described. 25 The primers for real-time PCR are provided in Table S2.

| Immunofluorescence and immunohistochemistry staining
The cells were fixed in 4% formaldehyde in PBS for 15 minutes, per-  At least six clones were subjected to sequencing analyses (Invitrogen, Shanghai, China).

| Dual-Luciferase Reporter
The 3′-UTR of DNMT3A mRNA containing the wild-type or mutant miR-200b-binding site were chemically synthesized by Invitrogen

| Co-immunoprecipitation assays
Co-immunoprecipitation was conducted as preciously described. 26 Briefly, cells were rinsed with ice-cold PBS and lysed in RIPA buffer.
The cells lysates were pre-cleared by adding 100 μL of Protein A/G agarose beads, and then subjected to overnight incubation with specific antibodies anti-MYC antibody (N262; Santa Cruz Biotechnology) or anti-DNMT3A antibody (IMG268; Imgenex). Immunocomplexes were precipitated by incubating with Protein A/G agarose beads, followed by washing with ice-cold lysis buffer. Subsequently, the agarose beads with immune complexes were boiled in 2× SDS sample buffer. Then, the immune complexes were analysed by western blotting.

| Chromatin immunoprecipitation assay
Chromatin immunoprecipitation (ChIP) was performed as previously described with minor modification. 18,26 Briefly, cells were cross-linked with 1% formaldehyde at 37°C for 10 minutes. After removing the medium, cells were resuspended in 100 μL lysis buffer per 1 × 10 6 cells and lysed on ice for 15 minutes. Then, the samples were sonicated to shear chromatin to fragments ranging from 200 to 1000 base pairs. Subsequently, the chromatin fragments were used for immunoprecipitation with antibodies as follows: anti-MYC antibody (N262; Santa Cruz Biotechnology) and anti-DNMT3A antibody (IMG268; Imgenex). The primers used for PCR analysis for the proximal promoter region of the miR-200b were as follows: forward primer, 5′-CACCGCCTCC-CATTGTC-3′; reverse primer, 5′-CACAGGAAGTCAGTTCAGACC-3′.

| Sphere-formation assays
Mammospheres were cultured as we previously described. 7

| Migration and invasion assays
The cells were planted in six-well plates and transfected as indicated.

| Statistical analysis
All statistical analyses were performed with SPSS software (version 18.0; SPSS Inc., Chicago, IL, USA). All data were represented as mean ± SD from at least three independent experiments. The statistical significance in mean values was determined with Student's ttest (two-tailed). For skewed distribution data, statistical significance was determined with nonparametric test. P < 0.05 indicated a statistically significant difference.

| High MYC, DNMT3A levels and low miR-200b levels in TNBC
We began the study with evaluating MYC, DNMT3A and miR-200b levels in breast cancer tissues. We found that miR-200b expression in TNBC was reduced compared with in ER+ and ER-PR-HER2+ (simplified as HER2+) breast cancer ( Figure 1A). The RNA-sequencing data of TCGA showed DNMT3A and MYC were highly expressed in TNBC tissues compared with in other breast cancer subtypes ( Figure 1B and C). The IHC analysis showed that MYC and DNMT3A were up-regulated in TNBC tissues than other However, MYC and DNMT3A levels in breast cancer cell lines did not show a significant subtype-specific signature ( Figure S1B and C). In addition, there was no significant difference in miR-200b promoter methylation levels among the breast cancer subtypes according to TCGA methylation database ( Figure S1D).

| MYC recruits DNMT3A to miR-200b promoter region inducing gene silencing
To investigate whether MYC is involved in epigenetic mir-200b repression, we knocked down MYC expression with siRNA in MDA-MB-231 and BT549 cells ( Figure 3A and B). MSP and bisulfite sequencing analysis showed that miR-200b methylation levels were reduced following MYC knockdown ( Figure 3C and D). Subsequently, we found the expression of miR-200b, miR-200a and miR-429 was significantly increased following MYC knockdown ( Figure 3E and Figure S3). In addition, DNMT3A knockdown did not significantly reduce MYC levels in TNBC cells ( Figure S4). It seemed that   and E and Figure S7A). The number of mammospheres represents the self-renewal ability of the cells, and the diameter of mammospheres indicates the self-renewal ability of each sphere-generating cell. 33,34 The cells overexpressing miR-200b formed fewer and smaller mammospheres than the scramble groups. Notably, DNMT3A silencing also resulted in lower mammosphere-forming efficiency ( Figure 5F and G). In addition, mouse xenograft experiments showed that miR-200b overexpression inhibited tumour growth in vivo ( Figure S7B, C and D). Furthermore, western blotting analysis showed that miR-200b overexpression may inhibited EMT and self-renewal ability in vivo ( Figure S7E). Above all, miR-200b overexpression or DNMT3A silencing suppressed the EMT and mammosphere formation of TNBC cells.

| DISCUSSION
The miR-200 family expression levels have been proved to be different among breast cancer subtypes. 27,35 Using the public prediction database, MiR-200b/c/429 is predicted to directly targeting DNMT3A. Furthermore, miR-200b has a stronger inhibitory effect on mammosphere formation than other members of miR-200 family. 36 Although all miR-200 family members were repressed in TNBC cell lines, only miR-200b repression (~100-fold) was consistently observed as previously reported. 30 As a result, miR-200b was selected to be main research molecular. In this study, miR-200b expression in TNBC tissues was inhibited compared with other breast cancer subtypes, while DNMT3A and MYC expression were elevated significantly. Our study showed that the TNBC cells showed a lower expression level of miR-200b than other breast cancer cells and normal breast cells as previously described. 30 Furthermore, we explored the interaction of MYC, DNMT3A and miR-200b in TNBC cells, and revealed that miR-200b was epigenetically suppressed by MYC and DNMT3A-mediated promoter methylation.
As a transcription factor, MYC can promote some target genes expression by dimerizing with Max, and repress others through interacting with Miz-1, Sp1 and Smad2. [37][38][39] Recent studies have shown that MYC is associated with epigenetic regulators in silencing miRNAs.
In hepatocellular carcinoma, MYC repressed miR-101 expression by recruiting EZH2 to the promoter regions of miR-101. 40 During Helicobacter pylori-related carcinogenesis, MYC, DNMT3B and EZH2 interacted with each other and led to let-7c silencing by inducing histone methylation and DNA hypermethylation. 41 However, the role of MYC in epigenetically regulating miR-200 expression is still unclear.
A previous study has demonstrated that MYC represses p21Cip1 transcription through recruitment of DNMT3A. 42 MYC directly binds to and represses the miR-200b promoter in endometrial carcinoma cells. 43 Furthermore, DNMT3A knockdown increases miR-200b expression. 44 Then, we wondered whether MYC is responsible for the epigenetic repression of miR-200b via interacting with DNMT3A.
Here, we reported that MYC was associated with DNMT3A in MDA-MB-231 cells. The complex of MYC and DNMT3A co-occupied the miR-200b promoter and resulted in miR-200b silencing dependent on DNMT3A-mediated promoter methylation.
However, there was a discrepancy between BSP results in breast cancer cell lines and TCGA methylation sequencing data in breast cancer tissues. This may be attributed to the different CpG islands included in BSP (Region A) and TCGA methylation sequencing (Region B) ( Figure S8). A previous study has demonstrated that the miR-200b promoter possesses many CpG islands which were methylated differently. 45 It is unclear whether methylation levels of these CpG sites (Region B) are associated with miR-200b expression. The incorporation of more CpG sites into methylation sequencing will contribute to the validation of the conclusion.
Moreover, we demonstrated miR-200b directly inhibited DNMT3A expression and then formed a feedback loop, which contributed to further miR-200b repression and DNMT3A overexpression by a self-reinforcing system. 46 We also try to validate the conclusion in other breast cancer cell lines such as MCF-7. Because of the low methylation level of miR-200b promoter, the feedback loop did not exist in MCF-7 cells. We demonstrated that miR-200b overexpression or DNMT3A knockdown inhibited EMT and selfrenewal of TNBC cells. 11 Thus, miR-200 repression or DNMT3A overexpression may be responsible for the aggressive behaviours of TNBC, paving the way for developing novel targeted agents and designing therapeutic strategy. Delivering chemically modified miRNA mimics by nanoparticles seems to be a promising therapy to overcome the repression of tumour suppressor miRNAs. Another therapy strategy is the drugs that can affect miRNAs regulation. 47 Interestingly, a latest study has shown that HDAC inhibitors for epigenetic regulation of miR-200 are potential therapy against TNBC. 48 However, the instability of delivery in vivo remains a challenge of the miRNA-based therapy.
To be concluded, we found that the feedback loop between miR-200b and DNMT3A played an important role in EMT and mammosphere formation of TNBC cells. Importantly, we demonstrated that miR-200b was epigenetically silenced by MYC and DNMT3Amediated DNA methylation. We believed the newly identified feedback loop provides a new insight into the pathogenesis of TNBC, and represents potential therapeutic targets for the treatment of TNBC.