Increased expression of MUSASHI1 in epithelial breast cancer cells is due to down regulation of miR-125b

Musashi1 (MSI1) is an oncogenic protein with a crucial role in the proliferation and characteristics of the epithelial cells in breast cancer. The change in expression of MSI1 has a role in solid tumor progression. There are different factors that regulate MSI1 expression in various cancer tissues including microRNAs which are considered as one of the most important of these factors. The aim of our study is identification of the molecular cause of maximal expression of MSI1 in epithelial breast cancer cell lines. Among predicted microRNAs, miR-125b, miR-637 and miR-802 were able to significantly reduce the luciferase activity. In addition, the relative expression of these three miRNAs were measured in the cancerous cell lines that results showed a significant reduction in expression of all microRNAs. On the other hand, only the overexpression of miR-125b caused a change in the expression pattern of MSI1 in breast epithelial cancer cell lines. Accordingly, our results demonstrated that the exogenous expression of miR-125b decreased not only the MSI1 protein but also expression of epithelial markers in breast cancer cells. The results of luciferase reporter assay showed that MSI1 is a direct target for miR-125b in epithelial breast cancer cells. Moreover, higher amount of MSI1 in those cell lines seems due to the reduced amount of miR-125b, which is responsible for epithelial features of those kinds of cancer cells. Therefore, the modulation of miR-125b may be a potential approach to help to combat against epithelial breast tumors.


Background
Breast cancer has the highest incidence of female cancers worldwide. Analysis of global gene expression and DNA microarrays from cancer tissues asserted that a variety of genes are causative for this type of cancer. Accordingly, breast cancer is divided into five groups; a) HER2-positive, b) luminal A, c) luminal B, d) basal-like and e) normal-like subtypes [1].
In mammalian cells, the family of musashi (MSI) as RNA-binding proteins are highly conserved and have two members, musashi1 (MSI1) and musashi2 (MSI2). MSI1 has a main function in self renewal and maintenance of stem cells [2,3]. MSI1 increases proliferation of mammary stem cells through triggering of two signaling pathways; Notch and Wnt [4]. Moreover, MSI1 is strongly expressed in multiple solid tumors such as brain [5][6][7], lung [8], ovary, squamous oral, colorectal and renal cancers [9][10][11][12]. On the other hand, MSI2 is reported to be upregulated in hematopoietic cancer cells [13]. The expression of MSI1 in HER2-positive breast cancer cell lines is correlated with HER2 and knock down of MSI1 reduces colony expansion of spheroid cultures in tumor cells [14]. Furthermore, MSI protein is essential for the preservation of epithelial property of breast cancer cells whereas overexpression of MSI2 regulates in vivo processing of epithelial-to-mesenchymal transition (EMT). Conversely, knocking down of MSI1 or MSI2 in the BT474 cell line enhances mesenchymal markers and decreases epithelial markers expression [15].
MicroRNAs belong to noncoding RNAs that control gene expression mostly through binding to 3'UTR of target genes. Generally, miRNAs regulate target gene expression via reduction of mRNA stability or translation [16][17][18]. MicroRNAs have diverse roles in various steps of breast cancer progression, which were indicated in recent studies [19][20][21]. 3`-untranslated region (3'UTR) of MSI1 has a length of 1800 nucleotides and is predicted to contain different sites for binding of endogenous miR-NAs. The documentation shows that MSI1 has an oncogenic function in breast cancer progression, therefore any modulation in MSI1 expression could be considered as a possible therapeutic approach against breast cancer progression. Thus, the aim of our study was identification of putative miRNAs, which modulate MSI1 transcript in breast cancer cells. For this purpose, we first predicted the miRNAs that bind to 3'UTR of MSI1 using bioinformatic databases and then confirmed the direct binding of selected miRNAs to MSI1 3'UTR by luciferase assay. Finally, overexpression of miR-125b in human breast epithelial cancer cells modulated the expression of MSI1 as well as epithelial markers.

Results
Up-regulation of MSI1 was only detected in epithelial breast cancer cell Analysis of transcript and protein of MSI1 were carried out in several breast cancer cell lines with different origins. The results demonstrated that MSI1 was highly expressed in T-47D and MCF-7 cells (epithelial cell line) but not in SKBR3 (HER2-positive cell line) and MBA-MD-231 cells (mesenchymal cell line) (Fig. 1).

Bioinformatics prediction of microRNAs targeting MSI1
For prediction of those microRNAs which target MSI1, the following conditions were considered: the minimum folding free energy, context++ score, conservation, number and site seed type of miRNA binding to MSI1 3′-UTR [22]. Due to maximal expression of MSI1 in MCF-7 and T-47D cancer cell lines (epithelial cell lines) compared to MCF-10A (normal breast cell line), the levels of predicted miRNAs were assessed using a GEO dataset GSE70480, which was related to the expression profile of miRNAs in aforementioned cell lines. Based on extracted data, miR-125b-5p, miR-802, miR-490-3P, miR-338-3p and miR-637 were selected due to their effect on MSI1 downregulation (  from the list as it had previously reported that its expression was high in MCF-7 similar to what reported for MSI1. Therefore, miR-342-3p could not be supposed to target MSI1 [23]. Also, the selection of miR-802 for further analysis was due to reduced expression in MCF-7 compared to normal breast cell line [24].
Down-regulation of predicted microRNAs were verified in epithelial breast cancer cells Since expression of MSI1 was high in both epithelial cancer cells (T47-D and MCF-7), the levels of predicted mature miRNAs were assessed in those cell lines (Fig. 3). Mature miRNAs expression were measured with MiRCURY LNA primer mix (Exicon, USA) and normalized with MiRCURY LNA U6 (Exicon, USA). However among them, only miR-125b-5p, miR-637 and miR-802 were selected for luciferase reporter assay because these microRNAs had lower expression values than others in both epithelial cancer cells compared to MCF-10A.

Confirmation of direct interaction of selected microRNAs with MSI1 transcript
To confirm direct binding of miR-125b-5p, miR-637 and miR-802 to 3′-UTR of MSI1 transcript (Supplementary Table 1), HEK293T cells were transfected with plasmids expressing precursors of both microRNAs as well as wild type or mutated MSI1 3'UTR. Transfected cells were harvested, lysed and the cell lysates were used for dual-luciferase reporter assay. Consistent with bioinformatics data, selected microRNAs significantly repressed luciferase activity. Elimination of microRNAs binding sites from 3'UTR of MSI1 resulted in significant increase in luminescent signals (Fig. 4). Therefore, direct targeting of MSI1 transcript by those microRNAs was verified.
Elevated amount of miR-125b was able to inhibit MSI1 expression in cancer cells To evaluate which one of microRNAs has a potential to negatively modulate the expression of MSI1 in epithelial cancer cells, MCF-7 and T-47D cells were transfected with plasmids expressing miRNA precursors. Expression of mature miRNAs in breast cancer cell lines after transfection were analyzed (Fig. 5). According to our data, ectopic expression of miR-125b reduced MSI1 transcript ( Fig. 6 A, B) and its protein compared to negative control (Universal scramble) in both cell lines ( Fig. 6 C, D). Surprisingly, in contrast to luciferase data, overexpression of miR-637 and miR-802 were not able to decrease the MSI1 expression in breast cancer cells. Consequently, we decided to use miR-125b for evaluating the effect of the microRNA on expression of epithelial markers in breast cancer cells. Ectopic expression of miR-125b reduced epithelial markers In previous studies, it was shown that knockdown of MSI1 in epithelial cells such as BT474 was responsible for reduction in epithelial markers including E-cadherin and epithelial cellular adhesion molecule (EpCAM) and CD24 [15]. Inversely, there was an increase in mesenchymal markerssuch as Fibronectin and jagged1 in similar circumstance. Our data were evident that ectopic expression of miR-125b in MCF-7 and T-47D cell lines gave rise to a reduction in E-cadherin, EpCAM and CD24 (Fig. 7a, c). However, increasing expression rate of miR-125 could not able to change the expression of mesenchymal markers (Fig. 7b, d). Moreover, Immunoblotting of E-cadherin in MCF-7 cells confirmed the reduction of epithelial markers followed by ectopic expression of miR-125b (Fig. 7e). Change of CD24 expression in EGFP positive population cells, which were transfected with miR-125b.pBUD and Usc. Pbud were evaluated by flow cytometry. A transfection efficiency of 30% was obtained in MCF7 cells (supplementary figure  2). Analysis of fluorescence intensity revealed that expression of CD24 was reduced after overexpression of miR-125b (Fig. 8). Reduction of MSI1 intensity in EGFP positive cell population in MCF-7 cells transfected with miR-125b.pBUD were shown in supplementary figure 3. Also, MCF-7 cells were transfected with EGFP.pBUD plasmid (USc.pBUD) as negative control. EGFP positive and negative cells have the same intensity for MSI1 (supplementary figure 4). Flow cytometry results were shown in supplementary figure 5. Eventually, according to the results, we concluded that down-regulation of the miR-125b is responsible for uprising of MSI1, which is crucial for maintenance of the epithelial phenotype in breast cancer cells (Fig. 9).

Discussion
RNA binding protein, MSI1, is highly expressed in different solid tumors [25,26]. MSI1 has a crucial role in stem cell and cancer stem cell proliferation [27,28]. MSI1 protein and notch signaling pathway are two key regulators for asymmetric division of the stem cell population in mammary epithelial cells [29]. In breast cancer, a positive correlation between expression of ErbB2 and MSI1 was observed. It has demonstrated that MSI1 expression is increased in spheroid culture derived from breast cancer lines such as T-47D and MCF-7 [14]. In breast cancer, MSI1 has also a critical role in EMT.
Hence expression of MSI1 in T-47D and MCF-7 Similarly, miR-637 has also antitumor activity, such that its expression was reduced in hepatocellular carcinoma cell line (HCC) and clinical specimen of cancer patients (reference). The experiments revealed that overexpression of miR-637 in some hepatocellular carcinoma cell lines gave rise to a reduction in proliferation of cancer cells by regulation of signal transducer and activator of transcription 3 (STAT3) [30]. Furthermore, it was indicated that miR-637 has a tumor suppressor role, which prevents migration and invasion of cancer cells by suppression of AKT1 in human glioma [31]. Recently, it was reported that overexpression of miR-125b decreased the proliferation of breast cancer cells and promoted their apoptosis by regulation of G2/M phase. miR-125b controls cell cycle through regulation of a number of target genes such as multiple EGF-like-domains 9 (MEGF9), cyclin J (CCNJ), aminopeptidase A (ENPEP) and casein kinase 2-alpha (CK2-α) [32]. Likewise, it was reported that ectopic expression of miR-125b could prevent migration and proliferation of breast cancer cells through regulation of KIAA1522 [33].
The long 3'UTR fragment of MSI1 coding sequence has several potential binding sites for microRNA. In the present research, utilizing in silico analysis of a GEO dataset (GSE70480), we predicted several miRNAs that were able to bind 3'UTR of MSI1. Expression of the predicted miRNAs (miR-125b-5p, − 802, − 637, − 490-3p and − 338-3p) was assessed in some epithelial cancer cell lines in which MSI1 expression was high. Amongthem, three miRNAs (miR-125b-5p, − 802, and − 637) with low expression rate in tested breast cancer cells were chosen for further investigation. All of three selected miRNAs were shown to bind directly to MSI1 3'UTR as verified by luciferase reporter assay. Among the aforementioned miRs, miR-125b was able to regulate MSI1 expression whereas overexpression of miR-637 and -802 could not able to modulate transcript of MSI1 in MCF-7 and T-47D cells. Our further in silico analysis using RNAbinding protein-specific database (http://rbpdb.ccbr. utoronto.ca/) indicated that RNA binding motif protein X-Linked (RBMX) could attach to the same sequence of MSI1 3'UTR, which covers the miR-637 binding site. Interestingly, RBMX are actively expressed in breast cancer cells according to the human protein atlas data (https://www.proteinatlas.org) [34,35]. In this respect, several independent studies have indicated that there is a counteraction between RNA binding protein and miR-NAs for accessibility to 3'UTR of target genes [36,37]. Therefore, it is more likely that RBMX compete with miR-637 for binding to its target sequence in MSI1 3'UTR. However to validate such hypothesis, further analyses are needed to be performed. On the other hand, following down regulation of MSI1 by miR-125b MCF-7 and T-47D cell lines, the expression of epithelial markers including E-cadherin, CD24 and EpCAM was also diminished. But conversely, transcript level of mesenchymal marker (Vimentin) was not affected in these cell lines, presumably due to the lack of enough expression of Vimentin as reported previously [38].
However, our data were in good agreement with the recent study that down regulation of MSI1 in epithelial cell line, BT474, caused a reduction in epithelial markers [15]. It seems that this effect is cell line dependent, as transfection of miR-125b inhibitor in triple negative breast cancer lines (TNBC) was responsible for reduction in cellular proliferation and process of EMT [39]. In these lines such as MDA-MB-468 and MDA-MB-231, the cells are negative for estrogen receptors, progesterone receptors, and excess HER2 protein, [39]. Similarly, in an agreement to our data, overexpression of miR-125 in Hs578T cells inhibited the EMT through a reduction in MAP 2 K7 [40].

Conclusions
Collectively, we indicated that MSI1 is a potential target of miR-125b. Hence, miR-125b could be considered as a potential factor for regulation of the progress of a number of breast cancer types and is potentially applicable for therapeutic purposes. However the exact mechanism underlying this phenomenon should be investigated in further studies as it would need to be critically looked at while comparing the same phenomenon with primary cells.

Construction of expression plasmids
For construction of the plasmids expressing human miRs-802, − 637, −125b, miRNA precursors were PCR amplified by specific primers (Table 1, A) and using human genome as a template. Then, the desired sequences were cloned into SalI and XbaI sites of pBUD4.1 expression plasmid under CMV promoter. All restriction enzymes were obtained from Thermo Scientific, USA. To verify the expression of the constructs, the recombinant vectors were double digested with SalI and XbaI restriction enzymes. The recombinant plasmids were confirmed by restriction digestion (Supplementary figure 6). For negative control plasmid, universal scramble sequence was cloned into EGFP.pBUD plasmid (USc.p-BUD) as reported previously [41]. Schematic representation of recombinant plasmids are displayed in supplementary figure 7. For Luciferase reporter assay, the multiple cloning site in pSICHECK2 plasmid (Promega, USA) was modified by cloning a synthetic DNA fragment into SalI and NotI sites in multiple cloning site and resulted vector was named as pSICHECK2.MCS. The synthetic DNA, which was constructed by annealing two oligonucleotides (  (Supplementary figure 8). Mutated type of MSI1 3'UTR for each miRNA binding site was constructed with site directed mutagenesis protocol [42] (Table 1, C) and these mutated fragments were cloned into pSICHECK2.MCS plasmid. Schematic representation of recombinant plasmids for luciferase assay were shown in Supplementary figure 9. The constructed

Transcript analysis
The extraction of total RNA was achieved using TRI reagent (Thermo Scientific, 15,596-018) and cDNAs were synthesized using random hexamer primers and implementing MMLV reverse transcriptase according to the manufacturer's protocol (TaKaRa, RR037A, Japan). To get rid of DNA contamination, DNase treatment was performed during cDNA synthesis. At the next step, 50 ng of cDNA was used for real-time PCR using SYBR green (TaKaRa, RR820Q, Japan) and specific primers on step one plus Real-Time PCR System (Applied Biosystems, USA). Primers were designed by the Beacon designer (Version 7.2, USA) and purchased from Metabion Company (

Luciferase reporter assay
For Luciferase assay, HEK293T were seeded in 24-well plate to reach enough confluence for 24 h. Then, cells were co-transfected with pSICHECK2.MCS.MSI1 and each of precursor expressing plasmid. Besides, plasmid contained mutated type of MSI1 3'UTR for each miRNA binding site along with precursor expressing plasmid were co-transfected into HEK293T cells. Cotransfections were accomplished using Lipofectamine 2000 (Invitrogen, USA) according to its protocol. Approximately 2 days post-transfection, the cells were harvested and analyzed with dual-Luciferase reporter assay system (Promega, USA).

Immunofluorescence analysis
Approximately 48 h post-transfection, miR-125b.pBUD transfected MCF-7 cells were treated with 4% paraformaldehyde for 30 min at 25°C. Permeabilization of the cells were achieved through incubating with 0.4% Triton X-100 for 20 min. After cell washing, primary antibody was diluted by blocking solution (BSA, 5 mg/mL) and used for cell staining. The cells were then incubated at 4°C overnight. In the next step, the cells were stained with secondary antibody for 1 h at 37°C and finally, they were stained with DAPI (D9542, Sigma) at final concentration of 1 μg/mL at room temperature for 5 min. Both first and secondary antibodies were Musashi-1 antibody (1:100, R&D: AF2628) and goat anti-rabbit IgG-FITC (1: 100, Santa Cruz: Sc-2012). Images were captured under a fluorescent microscope (Olympus, Center Valley, PA, USA) using Olympus DP70 camera. CD24 expression was assayed using flow cytometry. Co-transfected MCF-7 cells with miR-125b.pBUD and pBUD.USc were detached using triple (Gibco, 12,604-012) after 48 h. Treated cells were stained with PE labeled mouse Anti-Human CD24 (BD Pharmingen™, 555,428, USA) for 30 min. After washing, the cells were evaluated by FACS Caliber flow cytometer (Becton Dickinson, USA and obtained results were analyzed with the ModFit LT™ v 4.0 program.

Statistical analysis
SPSS Statistical software (v. 20, IBM Corp., USA) was used to perform the data analysis and all data were reported as mean ± standard error of mean (SEM). Independent student t-test and one-way analysis of variance (ANOVA) were utilized to determine the statistical differences between groups, where p < 0.01 and p < 0.05 have been applied for indication of statistically significant differences.