Antagonism of miR-21 Reverses Epithelial-Mesenchymal Transition and Cancer Stem Cell Phenotype through AKT/ERK1/2 Inactivation by Targeting PTEN

Background Accumulating evidence suggested that epithelial-mesenchymal transition (EMT) and cancer stem cell (CSC) characteristics, both of which contribute to tumor invasion and metastasis, are interrelated with miR-21. MiR-21 is one of the important microRNAs associated with tumor progression and metastasis, but the molecular mechanisms underlying EMT and CSC phenotype during miR-21 contributes to migration and invasion of breast cancer cells remain to be elucidated. Methodology/Principal Findings In this study, MDA-MB-231/anti-miR-21 cells were established by transfected hsa-miR-21 antagomir into breast cancer MDA-MB-231 cells. EMT was evaluated by the changes of mesenchymal cell markers (N-cadherin, Vimentin, and alpha-SMA), epithelial cell marker (E-cadherin), as well as capacities of cell migration and invasion; CSC phenotype was measured using the changes of CSC surface markers (ALDH1 and CD44), and the capacity of sphereforming (mammospheres). We found that antagonism of miR-21 reversed EMT and CSC phenotype, accompanied with PTEN up-regulation and AKT/ERK1/2 inactivation. Interestingly, down-regulation of PTEN by siPTEN suppressed the effects of miR-21 antagomir on EMT and CSC phenotype, confirming that PTEN is a target of miR-21 in reversing EMT and CSC phenotype. The inhibitors of PI3K-AKT and ERK1/2 pathways, LY294002 and U0126, both significantly suppressed EMT and CSC phenotype, indicating that AKT and ERK1/2 pathways are required for miR-21 mediating EMT and CSC phenotype. Conclusions/Significance In conclusion, our results demonstrated that antagonism of miR-21 reverses EMT and CSC phenotype through targeting PTEN, via inactivation of AKT and ERK1/2 pathways, and showed a novel mechanism of which might relieve the malignant biological behaviors of breast cancer.


Introduction
Epithelial-mesenchymal transition (EMT) is associated with increased aggressiveness and metastasis in carcinomas, including breast cancer, as it allows cells to migrate and invade surrounding issues and escape into the bloodstream, en route to establishing metastasis. Once these metastatic cells reach their destination, they can undergo mesenchymal-epithelial transition (MET) to establish secondary tumors, resulting in cancer spreading and treatment failure [1][2][3]. On the molecular level, cells undergoing EMT towards a more mesenchymal phenotype involves loss or lowered the expression of epithelial markers such as E-cadherin, and increased the expression of mesenchymal markers such as Ncadherin, Vimentin, and alpha-SMA [2,4,5]. There are studies suggested that EMT in breast cancer is tightly linked to the basallike phenotype breast cancer subgroup and cancer stem cells (CSCs) [6][7][8].
CSCs are predicted to be critical drivers of tumor progression due to CSC characteristics including self-renewal capacity, limitless proliferative potential and metastasis potential, suggesting that targeting CSC characteristics would likely eliminate CSCs which are the ''seeds'' of tumor recurrence and metastasis. Although the CSC hypothesis suggests that tumors can arise from stem/progenitor cells, studies from many laboratories demonstrated that EMT can endow cells which possess CSC characteristics as well as a more motile invasive phenotype [7][8][9][10]. Previously studies have been confirmed that breast cancer contains a CSC-compartment [11][12][13], which can be enriched by purifying Aldehyde dehydrogenase 1 (ALDH1)positive cells [13] or CD44 + /CD24 2/low cells [11] by sorting, and also by purifying sphereforming cells (mammospheres) from parental cells [12]. There are studies demonstrated that EMT phenotype is highest in the CD44 + /CD24 2/low breast cancer CSCs (7), and CSCs enriched from breast tumors and metastatic breast pleural effusions express markers similar to cells that have undergone an EMT [7,14]. Similarly, EMT and CSC markers are also frequently associated with breast cancers that have a propensity to metastasis, such as basal-like phenotype breast cancer subgroup [15] and metaplastic [16] breast cancers.
MicroRNAs (miRNAs) are a family of small non-coding RNA molecules which regulate gene expression by base pairing to the 39-UTR of the target mRNA. Recently, a series of miRNAs have been shown to play critical roles in the progression and metastasis of human malignancy [17,18], including breast cancer. MiR-21 is one of the first miRNAs detected in the human genome, which also is one of miRNAs known to be upregulated in all types of human malignancies [19]. Recent studies indicated that several tumor suppressors including phosphatase and tensin homolog deleted on chromosome ten (PTEN) [20], tumor suppressor gene tropomyosin 1 (TPM1) [21], programmed cell death 4 (PDCD4) [22], maspin [23], and matrix metalloproteinases inhibitors RECK and TIMP3 [24] were targets of miR-21, suggesting that miR-21 is an important oncogenic miRNA which is closely related to tumor growth and metastasis.
There are studies demonstrated that miR-21 is an important component of the cellular signaling circuitry that regulates the EMT program [25][26][27], as well as associates with CSC signatures [26,28], suggesting that miR-21 plays a key role in the process of EMT and acquisition of CSC characteristics, which consistent with our previously study [29], but the underlying mechanisms remain unclear.
These results strongly suggest that miR-21 might via AKT and ERK1/2 pathways by targeting PTEN regulate EMT and CSC phenotype. In the present study, we have revealed that antagonism of miR-21reversed EMT consistent with CSC phenotype via AKT and ERK1/2 pathways by targeting PTEN, indicates a molecular pathway which might relieve the malignant biological behaviors of breast cancers.

Antagonism of miR-21 Reversed CSC Phenotype in MDA-MB-231 Cells
To examine the effects of antagonism of miR-21 on CSC phenotype in MDA-MB-231 cells, the proportion of cells with ALDH1 + (ALDH bright ) and CD44 + /CD24 2/low in MDA-MB-231/anti-miR-21 cells and negative control cells were measured. The percentage of cells with ALDH1 + in MDA-MB-231/anti-miR-21 cells was 0.8560.092, significantly less than 2.32360.315 in negative control groups (p = 0.0051; Fig. 2A, B), by ALDE-FLUOR assay; the percentage of cells with CD44 + /CD24 2/low in MDA-MB-231/anti-miR-21cells was 0.54360.129, also significantly lower than 4.80761.062 in control groups (p = 0.0094; Fig. 2C, D), by FACS analysis. These results suggested that antagonism of miR-21 could decrease the breast cancer CSC proportion, which expressing CSC surface biomarkers ALDH1 + and CD44 + /CD24 2/low . Moreover, the relative mRNA and protein expression of ALDH1 and CD44 in MDA-MB-231/anti-miR-21 cells and control cells were also measured. The relative mRNA levels of ALDH1 and CD44 in MDA-MB-231/anti-miR-21 cells were significantly decreased as compared to control groups (p = 0.0039, p = 0.0034, respectively; Fig. 2E, F), as assessed by real time RT-PCR analysis, which suggested that antagonism of miR-21 could decrease the mRNA expression of ALDH1 and CD44. The protein expression of ALDH1 and CD44 were also decreased accordingly (p = 0.0002, p = 0.0005, respectively; Fig. 2G, H), by Western blot analysis, suggested that antagonism of miR-21 could decrease the protein expression of ALDH1 and CD44. These results indicated that antagonism of miR-21 could reverse CSC phenotype in MDA-MB-231 cells, which further suggested that miR-21 could involve in the regulation of CSC characteristics. More interestingly, the mammosphere number in MDA-MB-231/anti-miR-21 cells was significantly inferior to control groups (p = 0.0017; Fig. 2I, J), indicated that antagonism of miR-21 could decrease formation of mammospheres, which further suggested that miR-21 could regulate CSC characteristics, including the capacities of self-renewal and clonogenicity.

Antagonism of miR-21 Induced Over-expression of PTEN
Recent studies have demonstrated that miR-21 increased cell proliferation, migration and invasion through modulating tumor suppressor gene PTEN [20], but the role of PTEN in miR-21 regulating tumor EMT and CSC phenotype remains to be elucidated. To explore whether antagonism of miR-21 regulate PTEN expression, the mRNA and protein levels of PTEN were measured, by real time RT-PCR analysis and Western blot assay, respectively. As expected, both the mRNA and protein expression of PTEN in MDA-MB-231/anti-miR-21 cells were strongly up-regulated, as compared to control groups (p = 0.003, p = 0.0437, respectively; Fig. 3A, B, D). These results demonstrated that antagonism of miR-21 could up-regulate the expression of PTEN.
Antagonism of miR-21 Inactivated AKT and ERK1/2 AKT and ERK1/2 are two major signaling pathways in regulating cell proliferation, migration and survival, and both also were regulated by PTEN, but the roles of AKT and ERK1/2 pathways in miR-21 regulating tumor EMT and CSC phenotype remains to be elucidated. To determine the signaling molecules that are involving in antagonism of miR-21 reversing EMT and CSC phenotype, the protein levels of phosphorylated AKT (p-AKT) and AKT, phosphorylated ERK1/2 (p-ERK1/2) and ERK1/2 were measured by Western blot analysis. The protein levels of p-AKT and p-ERK1/2 in MDA-MB-231/anti-miR-21 cells were strongly decreased as compared to control groups (p = 0.0173, p = 0.0126, respectively; Fig. 3C, D), confirmed that antagonism of miR-21 could suppress AKT and ERK1/2 activation in process of reversing EMT and CSC phenotype.
Re-expression of miR-21 Induced EMT and CSC Phenotype, Accompanied with Down-expression of PTEN, as well as Activation of AKT and ERK1/2 To further confirm the role of miR-21 in regulating tumor EMT and CSC phenotype, hsa-miR-21 mimics or mimics negative control was transfected into established MDA-MB-231/ anti-miR-21 cells. The expression of miR-21 increased to more than 22-fold after hsa-miR-21 mimics treated, as compared to the negative control (p = 0.0037; Fig. 4A), indicated that hsa-miR-21 mimics transfection could increase the relative expression of miR-21.
Besides, to verify the roles of AKT and ERK1/2 pathways in miR-21 regulating cell proliferation in the cells that mentioned above, the cell growth numbers of the cells were calculated at 0, 12, 24, 36, 48, 60 and 72 h after treatments by cell count. We found that forced antagonism of miR-21 decreased cell proliferation in MDA-MB-231 cells (p = 0.0077; Fig. 6F), while reexpression of miR-21 elevated cell proliferation in MDA-MB-231/ anti-miR-21 cells, during 0-72 h (p = 0.0091; Fig. 6F). The results suggested that miR-21 could regulate the cell propagation to a certain degree in MDA-MB-231 cells. Meanwhile, both AKT and ERK1/2 inhibitions were decreased cell proliferation (p = 0.0109, p = 0.0031, respectively; Fig. 6F), which suggested that the regulation of AKT or ERK1/2 linked to changes in cell proliferation.

Discussion
MiRNAs are noncoding small RNAs that may act as oncogenes or tumor suppressor genes [37,38]. Growing evidence showed that miR-21 over-expression is detected in various kinds of human cancers including breast cancer, and is associated with EMT [25][26][27] and CSC characteristics [26,28]. But the direct roles and central molecular mechanisms of miR-21 in regulating breast cancer EMT and CSC phenotype remains to be elucidated. In this study, we found that antagonism of miR-21 in MDA-MB-231 cells reversed EMT and CSC phenotype, up-regulated the expression of PTEN, as well as inactivated AKT and ERK1/2 pathways. The results demonstrated that antagonism of miR-21 is sufficient to reverse EMT and CSC phenotype through modulating the expression of PTEN and AKT/ERK1/2 pathways, providing some direct evidence and molecular mechanisms that miR-21 is able to regulate EMT and CSC phenotype.
To investigate the role of miR-21 in regulating EMT and CSC phenotype of breast cancer cells, antagomir of miR-21, which can specifically bind to and inhibit endogenous miR-21 molecules, or its negative control oligomer, was transfected into MDA-MB-231 cells. The antagomir decreased 90% of endogenous miR-21 expression in the cells (Fig. 1A). Interestingly, the relative EMT and CSC phenotype were reversed by antagonism of miR-21( Fig. 1B-G; Fig. 2A-J), which consistent with miR-21 involves in promoting EMT and (or) CSC phenotype in pancreatic cancer cells [26,28], thyroid carcinoma cells [27], and our previously study in MCF-7 cells [29]. Accompanied with the results, antagonism of miR-21 significantly increased the expression of PTEN (Fig. 3A, B, D), as well as decreased AKT and ERK1/2 activation (Fig. 3C, D). These results demonstrated that miR-21 plays an important role in mediating EMT and CSC phenotype, accompany by regulating the expression of PTEN, as well as AKT and ERK1/2 activation.
Next, we explored the underlying mechanisms and relative signaling molecules in antagonism of miR-21 reversing EMT and CSC phenotype. In consistent with that miR-21 suppresses the function of tumor suppressor PTEN expression by binding its 39-UTR [20], we proved that antagonism of miR-21 up-regulated PTEN expression (Fig. 3A, B, D). As expected, PTEN downexpression notably blocked miR-21-reversing EMT and CSC phenotype (Fig. 5B, C, E), confirmed that antagonism of miR-21 exhibits its role partly by promoting PTEN expression, and the recovery of PTEN expression would reduce its function of tumor suppressor.
In summary, our results demonstrated that antagonism of miR-21 reverses EMT and CSC phenotype through AKT and ERK1/ 2 pathways inactivation by inducing PTEN expression in MDA-MB-231 cells (Fig. 7). This study showed a direct role and novel mechanism of miR-21 in inversing EMT and CSC phenotype, and the information may be useful to develop a new therapy for treatment of breast cancers in the future.

Reagents and Antibodies
Human has-miR-21 (MIMAT0000076) antagomir and negative control, mimics and negative control were purchased from Ribobio (Guangzhou, China). The PCR primers were synthesized by Sangon Biotech (Shanghai, China). Real time RT-PCR assay kits were purchased from Takara (Dalian, China). Specific siPTEN duplexes and scramble control siRNA sequence (SsiPTEN) were purchased from Ribobio. LY294002 and U0126 were purchased   For mRNAs, total RNA from cells was isolated using TRIzol reagent (Invitrogen). Real time RT-PCR reactions were carried out using SYBRH Premix Ex Taq TM II (Takara). Beta-actin was used as endogenous control.

Western Blot Analysis
Western blot analysis was conducted as our previously study [29,39]. Briefly, the protein was separated by SDS-PAGE (8%, 10%, or 12%) and transferred to PVDF membranes. Nonspecific binding sites were blocked by incubating with TBST containing 5% (w/v) non-fat dried milk. Then incubated with primary antibodies (as described above) and HRP-conjugated anti-rabbit IgG secondary antibody in order, and visualized by ECL chemiluminescence. Beta-actin or GAPDH was used as loading control. The bands were semi-quantified using ImageJ software.

Migration and Invasion Assays
Cell migration and invasion assays were performed as our previously described [29,40]. For invasion assay, 2.5610 4 cells were seeded on an 8-mm pore size Transwell insert (Corning Inc. Corning, NY, USA) coated with ECM (1:7.5) (Sigma), while cell migration assay did not coat with ECM. After 24 h of incubation at 37uC in an atmosphere containing 5% CO 2 , the cells adherent to the upper surface of the filter were removed using a cotton applicator. Then cells were stained with 0.4% crystal violet dissolved in methanol, and the numbers of cells on the bottom were counted.

ALDEFLUOR Assay
The ALDEFLUOR kit was used to isolate the population with high ALDH1enzymatic activity (ALDH bright ) as previously described [13,29]. Briefly, cells were incubated in ALDEFLUOR assay buffer containing ALDH substrate bodipy-aminoacetaldehyde (BAAA, 1 mmol/l). In each experiment, a sample of cells was stained under identical conditions with diethylaminobenzaldehyde (DEAB), a specific ALDH inhibitor, as negative control. Flow cytometry (FACS) analysis was used to measure ALDH bright cell subpopulation.

Construction of siPTEN
Based on human PTEN mRNA sequence (NM_000314), specific siRNA duplexes were designed, synthesized and annealed by Ribobio. The selected RNA duplex (siPTEN) corresponding to nucleotides 1567-1585 of PTEN mRNA is defined as sense: 59-GCUACCUGUUAAAGAAUCAdTdT-39 and antisense: 59-UGAUUCUUUAACAGGUAGCdTdT-39. The scramble control siRNA sequence (SsiPTEN) was also designed by Ribobio, and has no significant homology to any known human gene sequence.
Transfection of siPTEN 5610 5 cells were seeded in six-well plates and grown to 70% confluence. Human siPTEN or SsiPTEN was allowed to form transfection complexes with Lipofectamine TM 2000 (Invitrogen) in free of serum Opti-MEMH I (Invitrogen) at a final concentration of 50 nmol/l, according to the manufacturer's protocol.

Cell Proliferation Assay
Various cells were cultured in 10% FBS culture medium as mentioned in Cell Culture part. The cell growth numbers of the cells were calculated and analyzed by cell count at 0, 12, 24, 36, 48, 60 and 72 h after treatments.

Statistical Analysis
All data were expressed as mean 6 SD and were calculated by using statistics analysis software SPSS 13.0. Statistical difference of each treatment was compared by Student's t test. The P value equal to or less than 0.05 was considered as statistical significance.