Hypermethylation of lncRNA MEG3 impairs chemosensitivity of breast cancer cells

Abstract Background Chemoresistance posed a barrier to successful treatment of breast cancer (BC), and lncRNA MEG3 has been documented to implicate in BC development. However, whether MEG3 methylation, which led to low MEG3 expression, was relevant to BC progression and chemoresistance remained uncertain. Methods In the aggregate, 374 pairs of tumor tissues and adjacent normal tissues were collected from pathologically confirmed BC patients, and four BC cell lines, including MDA‐MB‐231, Bcap‐37, MCF‐7, and SK‐BR‐3, were purchased. Moreover, methylation‐specific polymerase chain reaction (PCR) was adopted to evaluate the methylation status of BC tissues and cell lines, and chemo‐tolerance of BC cell lines was assessed by performing MTT assay. Concurrently, transwell assay and scratch assay were carried out to estimate the migratory and invasive capability of BC cell lines. Results Methylated MEG3, lowly expressed MEG3, large tumor size (≥2 cm), advanced TNM grade and lymphatic metastasis were potentially symbolic of poor prognosis among BC patients (P < .05). Besides, MDA‐MB‐231 cell line exhibited the strongest resistance against paclitaxel, adriamycin, and vinorelbine (P < .05), while MCF‐7 cell line seemed more sensitive against these drugs than any other BC cell line (P < .05). Furthermore, pcDNA3.1‐MEG3 and 5‐Aza‐dC markedly sensitized MDA‐MB‐231 and MCF‐7 cell lines against the drug treatments (P < .05). Simultaneously, proliferation and metastasis of the BC cell lines were slowed down under the force of pcDNA3.1‐MEG3 and 5‐Aza‐dC (P < .05). Conclusion Preventing methylation of MEG3 might matter in lessening BC chemoresistance, owing to its hindering proliferation and metastasis of BC cells.


| INTRODUC TI ON
Breast cancer (BC), one common malignancy among global women, was usually sparked by abnormalities in breast epithelium. According to the statistics publicized by International Agency for Research on Cancer (IARC) in 2008, BC incidence has reached nearly 22.9% among worldwide females. 1 Despite standard treatments established by National Comprehensive Cancer Network (NCCN) guideline, annually there were 20%-40% of BC patients who suffered from BC recurrence and metastasis, 2 and the 5-year overall survival metastatic ones hovered around 25%. 3 Hence, novel therapeutic means for BC were in desperate need, given the unoptimistic outcome of BC patients.
Breast cancer onset, a cumulative process, was correlative with aberrant activation of oncogenes and inactivation of anti-oncogenes, which argued the significance of elucidating biomarkers that mirrored progression of BC. Long non-coding RNAs (lncRNAs), a class of non-coding RNAs longer than 200 nt, were acknowledged to modify the expression of protein-coding genes in considerable aspects, and their abnormal expression was sometimes a reflection of cancer development and chemoresistance. 4 For instance, highly expressed lncRNA HOX transcript antisense RNA (HOTAIR) and lncRNA H19 were found to quicken neoplastic growth and to facilitate tumor transfer. 5,6 Moreover, doxorubicin/etoposide tolerance of squamous cancer cells was intensified when lncRNA cancer up-regulated drug resistant (CUDR) was highly expressed, which blocked apoptosis of tumor cells. 7 Concerning lncRNA matriarchal-expressed gene 3 (MEG3) studied here, its expression was down-regulated and even lost in tumor tissues of brain, bladder, bone marrow, breast, cervix, colon, liver, lung, and prostate, [8][9][10][11][12] as compared with normal tissues of these organs. 13 In addition, over-expressed MEG3, as disclosed by in vitro experiments, could impede BC proliferation and metastasis by enhancing transcription of p53-targeting genes (eg, p21, Maspin, and KA11). 14 The MEG3 was also able to reverse cisplatin-tolerance of lung adenocarcinoma cells by inducing cell apoptosis, 15 and silencing of MEG3, conversely, improved cisplatin-resistance of non-small cell lung cancer cells via activation of WNT/β-catenin signaling. 16 Based on the above, there presented a probability that expressional loss of MEG3 might advance neoplastic (eg, BC) progression and boost chemoresistance of malignancies.
It was noteworthy that hypermethylation of MEG3 promoter was a principal cause of lowly expressed MEG3 in various disorders (eg, BC). 17,18 To be specific, the methylation rate of MEG3 promotor was higher in pituitary tumor than in normal pituitary gland. 19 And hypermethylation of MEG3 promotor was also discoverable among patients who were tortured by neurofibroma, meningioma, myelodysplastic syndrome, acute myelogenous leukemia, and acute myeloid leukemia. 10,[20][21][22] It was, hence, conjectured that MEG3 methylation might also account for the role of lowly expressed MEG3 in regulating tumor activity, such as metastasis and chemoresistance. Nonetheless, up to now a limited number of investigations have been undertaken to verify this point. In response, clinical and in vitro experiments were designed here to overcome this handicap.

| Acquisition of clinical specimens
From March 2008 to January 2014, 374 pairs of tumor tissues and adjacent normal tissues were collected from pathologically confirmed BC patients, who received treatments in the medical oncology of Minhang Hospital. The BC patients were graded according to severity, based on the staging standard revised by World Health Organization (WHO) in 2003. 23 The participants all complied with following items: (a) their diagnostic results agreed with criteria detailed in Chinese Anti-Cancer Association Guidelines and Specifications on Breast Cancer Diagnosis and Treatment 24 ; (b) they hardly underwent any treatments before; (c) they were able to communicate with medical staff smoothly; and (d) they have signed informed consents. Additionally, the subjects were eliminated from this project if they: (a) showed disorders in bilateral breasts, liver, and kidney, or (b) were complicated by other malignancies. Last but not the least, procedures of this project have obtained approvals from Minhang Hospital and its affiliated ethics committee.

| Reverse-transcription polymerase chain reaction
Breast Cancer tissues and cell lines were firstly lysed after supplementation of TRIzol reagent (Invitrogen), so as to extract total RNAs. Quality of the RNAs was guaranteed via UV spectrophotometer, and then, the RNAs were reversely transcribed into cDNAs (Invitrogen). With cDNAs as the template, real-time PCR was launched as per specifications of Power SYBR Green kit (TaKaRa), and the reaction condition was designed as (a) pre-denaturation at 95°C for 30 seconds, and (b) 40 cycles of denaturation at 95°C for 5 seconds, annealing at 60°C for 30 seconds, and extension at 72°C for 30 seconds. Finally, expression of MEG3 (upstream primer: 5′-CTGCCCATCTACACCTCACG-3′, downstream primer: 5′-CTCTCCGCCGTCT-GCGCTAGGGGCT-3′) as relative to GAPDH (upstream primer: 5′-GTCAACGGATTTGGTCTGTATT-3′, downstream primer: 5′-AGTCTTCTGGGTGGCAGTGAT-3′) was calculated, after being normalized based on 2 −ΔΔCT method.

| Cell apoptosis assay
The digested BC cells were centrifuged at 1000 g, and then, binding buffer was added to re-suspend the cells. After that, cell suspension, mixed by Annexin V-FITC (Invitrogen), was placed quiescently in the dark for 10 minutes. Then, the BC cells were stained by propidium iodide (PI) dye solution at room temperature for another 10 minutes.
Apoptotic rate of the BC cells was measured utilizing flow cytometry (Thermo Fisher), with an excitation (Ex) wavelength of 448 nm and an emission wavelength (Em) of 530 nm.

| MTT assay for assessing chemosensitivity of BC cells
Breast Cancer cells seeded at the concentration of 5 × 10 3 /well were

| Transwell invasion assay
On one hand, 100 μL matrigel, diluted by serum-free and high-glucose DMEM at a ratio of 1:8, was added into each Transwell chamber, and the gel was incubated at 37°C until solidification. Then,

| Cell scratch test
Well-grown MDA-MB-231 and MCF-7 cell lines, adjusted to a concentration of 2 × 10 5 /mL, were incubated into 12-well plates. When the cells grew to about 80%-90% confluence, a uniform line was drawn on the back of each well with a 10-μL pipette tip. At the time points of 0 hour and 24 hours, photographs of the scratches were taken under the microscope, and Image J software was employed to measure the width of scratches.

| Statistical analyses
All experimental data were analyzed using SPSS 13.0 statistical software. Comparisons between/among measurement data, expressed as mean ± standard deviation (SD), were accomplished by means of t test or one-way analysis of variance (ANOVA), while the enumeration data were contrasted through usage of chi-square test. The statistical difference was considered as significant in the context of P < .05.

| MEG3 was lowly expressed and methylated in BC tissues and cell lines
It was demonstrated that the expression level of MEG3 was obviously decreased in BC tissues when compared with adjacent normal tissues (P < .05) ( Figure 1A). At the meantime, the methylation rate of MEG3 was significantly raised in BC tissues as compared with adjacent normal tissues (P < .05) ( Figure 1B).   Table 2). The Kaplan-Meier analysis also vividly showed that BC patients who were featured by MEG3 methylation or low MEG3 expression were associated with shorter longevity than those with non-methylated MEG3 or high MEG3 expression (P < .05) ( Figure 1E).

| Inhibitory effect of chemotherapeutic drugs on growth of BC cell lines
As illustrated in Figure

| Influence of MEG3 and 5-Aza-dC on viability, apoptosis, migration, and invasion of BC cell lines
Over-expressed MEG3 was found to impair viability of MDA-MB-231 and MCF-7 cell lines (P < .05), while lowly expressed MEG3 reinforced   Our investigation also revealed that MEG3 served as an effective restraint of BC metastasis and proliferation (Figures 1-5 In addition, EMT was increasingly certified as an accompanying phenomenon when tumor cells exhibited drug-resistant inclinations. 41 For instance, epithelial-derived malignant cells not only expressed biomarkers that were indicative of mesenchymal differentiation, but also became less sensitive to chemo-treatments. [42][43][44] Hence, the incremental chemosensitivity of BC cells might, to some extent, result from the inhibition of MEG3 on EMT process of cancer cells ( Figure 5). Besides EMT, drug resistance of cancer cells could also be facilitated by restraining apoptosis of tumor cells, and lncRNAs appeared as crucial biomarkers in modulating apoptosis of neoplastic cells. For instance, knockout of lnc-AK022798 in cisplatin-tolerant gastric cancer cells was found to heighten Caspase-3/8 expression and to propel apoptosis of the tumor cells. 45 Here, over-expressed MEG3 (ie, pcDNA-MEG3 transfection) and de-methylated MEG3

| D ISCUSS I ON
(ie, 5-Aza-dC treatment) were found to elevate the apoptotic rate of MDA-MB-231 and MCF-7 cell lines ( Figure 3C,D), which suggested that enhancive BC apoptosis might be another mechanism explaining MEG3' reinforcing BC chemosensitivity. However, here we failed to figure out downstream molecules that aided MEG3 to play such roles, which awaited further researches.
Furthermore, epigenetic silencing of tumor-suppressive genes was, to some degree, insinuated to result from super-methylation of CpG island (CGI) in the promoter of these genes. Taking MEG3 for instance, its promoter was abundant with CGI which carried IG-DMR and MEG3-DMR, and methylation of its promoter could trigger in NSCLC tissues (ie, 96%) exceeded that in normal tissues (ie, 68%), and MEG3 expression was lower in NSCLC tissues than in normal tissues. 47 Results of in vitro studies also pointed out that lowly expressed MEG3 in ovarian cancer cells was relevant to hypermethylation in the promoter of MEG. 48 Consistently, the methylation rate of MEG3 in BC tissues was also obviously higher than that in adjacent non-tumor tissues ( Figure 1B). What's more, BC cell lines were methylated as relative to normal cell line ( Figure 1D), and de-methylation of MEG3 was capable of accelerating metastasis and enhancing chemosensitivity of BC cell lines ( Figure 2D). All these indicated that MEG3 methyla- Thirdly, animal models were not established to verify whether