Introduction MicroRNAs (miRNAs) are a class of evolutionarily conserved regulatory RNAs that pleiotropically suppress gene expression at post-transcriptional level [1]. MiRNAs control the expression of 10–30% of the human transcriptome and are crucial regulators of both physiologic and pathologic processes [2]–[4]. In cancer, the spectrum of miRNAs expressed in neoplastic cells differs dramatically from that found in normal cells and it is now well established that miRNAs play fundamental roles in essentially all aspects of tumor biology [5], [6]. In breast cancer, divergent miRNA expression between normal and neoplastic tissues has been demonstrated, as well as differential miRNA expression among the molecular subtypes of breast cancer, including luminal A, luminal B, Her2+ and basal-like [7], [8]. MiRNAs have been shown to play an important role in breast cancer initiation and progression. For example, overexpression of miR-21 in breast carcinomas has been shown to target important tumor-suppressor genes such as PTEN, PDCD4, and TPM1, and was associated with advanced clinical stage, lymph node metastasis, and poor patient prognosis [9], [10]. MiR-10a was reported to be overexpressed in about 50% of metastatic breast cancer and transcriptionally activated by the pro-metastatic transcription factor TWIST1 [11]. Reduced expression of miR-126 and miR-335 in the majority of primary breast tumors from relapsed patients was reported, and simultaneous loss of miR-126 and miR-335 expression was associated with poor distal metastasis-free survival [12]. Oncogene regulation by miRNAs has also been reported, including tyrosine kinase receptors HER-2 and HER-3 by miR-125b and miR-205, respectively [13], [14], and the miR-200 family, known to reduce cell migration and invasiveness by targeting ZEB transcription factor members, was suppressed in metastatic breast cancer [15], [16]. miRNA regulation by estrogen receptor-alpha (ERα), the most important prognostic and therapeutic indicator in breast cancer, has recently been described by us and others [17]–[20]. Specifically, the majority of miRNAs upregulated by ERα are key components of a negative feedback loop that restrict E2 action and thus play a tumor suppressive role. In this regard, ERα-activation of let-7 family members limits the expression of oncogenes, such as Ras and c-Myc, and promotes differentiation of cancer cells [18]; ERα-mediated activation of the miR-17/92 cluster functions as a tumor suppressing mechanism in breast cancer through the downregulation of cyclin D1 and AIB1 by the miR-17/20/106 family and the direct suppression of ERα mediated by miR-18 and miR-19 [19]. We and others have described a double-negative feedback loop involving E2-suppressed microRNAs that target ERα, specifically miR-206 and miR-221&222, resulting in upregulation of ERα expression and low miRNA level in luminal A-type breast cancers [17], [21]. Recent works from our group have shown that miR-191 is highly induced in several human solid tumors including colon, lung, pancreas, prostate, and stomach cancer [22], as well as acute lymphocytic leukemia (ALL)-associated hematopoietic malignancies [23]. We have also reported a strong positive correlation between miR-191 expression and ERα levels in breast tumors [7], suggesting an oncogenic function for this miR. A role for miR-191 in tumorigenesis is further strengthened by several findings, including that miR-191 is induced by a dioxin family carcinogen, the miR is hypomethylated and overexpressed in liver cancer [24], [25], and miR-191 inhibition decreases cell proliferation and tumor growth of hepatocellular carcinoma cells [24]. Furthermore, miR-191 overexpression promotes cell growth and suppresses apoptosis of gastric cancer cells [26]. However, in ovarian and thyroid follicular cancer, miR-191 represses MDM4 or CDK6 expression, respectively, thereby delaying cancer progression and tumor-related death [27], [28]. These contradictory findings indicate that the precise role for miR-191 in human neoplasia may be tumor-type specific and not well understood. In this current study, we report a positive association between ERα expression and miR-191 and miR-425, two intronic miRNAs hosted by the putative protein coding gene DALR anticodon binding domain containing 3 (DALRD3), and further show direct control of the miR-191/425/DALRD3 transcriptional unit by the E2/ERα axis. We evaluated that the estrogen dependent activation of miR-191/425 induces proliferation in part by targeting the estrogen modulated tumor-suppressor gene, EGR1. We also demonstrated that, when constitutively expressed in highly aggressive ERα negative breast cancer cells, the miR-191/425 cluster reprograms gene expression to impair tumorigenicity and metastatic potential through the suppression of several different oncogenic proteins. Results miR191/425 cluster is positively correlated with ERα levels MiR-191 and miR-425 are highly conserved miRNAs found on human chromosome 3 within the first intron of DALRD3 (Figure S1). Given their genomic organization and proximity, we hypothesized that miR-191 and miR-425 are co-transcribed and transcriptionally dependent on the host gene DALRD3. We examined expression of mature miR-191, miR-425, and DALRD3 mRNA in 20 different normal human tissues using qRT-PCR (Figure S2A). Both miRNAs were detected in all tissues and, their levels of expression were highly correlated, as shown by scatter plot analyses, (R2 = 0.7351; p 1.2) after 6 days of hormone starvation and 6 h of E2 treatment. Blue (down) and yellow (up) columns represent all miRNA genes regulated after hormone starvation. (B) Mature miR-191, miR-425 and DALRD3 mRNA levels were determined by qRT-PCR (* p-value 1.5 fold; p-value 0.001) (Table S2). Functional profiling of these genes defined that the greatest proportion of them is associated with cell adhesion, adherens junction followed by phosphatidylinositol signaling (Figure 4B). We used qRT-PCR to validate the modulation of over 20 genes identified in the microarray analyses or to their related molecular pathways in two different breast cancer cell lines (Figure 4C). Expression of many genes involved in promoting growth and metastasis of breast cancer cells was found to be downmodulated by miR-191/425 cluster: CCND1, CCND2, E2F1, CSDA and API5, regulatory proteins of the cell cycle progression and apoptosis [40]–[42]; FSCN1, TNC, VEGFA, CDC42 and SOX4, which have roles in angiogenesis and migration, and are involved in filopodia/invadopodia formation [43]–[48]; the protooncogene MYC, which initiates the transcription of a large set of genes involved in cell growth by stimulating metabolism and protein synthesis [49]; and SATB1, which reprograms gene expression to enhance aggressive histomorphological features and invasive capabilities [50]. We also found that miR-191/425 cluster represses cell-structure and adhesion genes typical of invasive breast cancer cells such as fibronectin, an ECM adhesive glycoprotein, and vimentin, the intermediate filament protein of mesenchymal cells, which together provide cellular integrity and resistance against stress [51]. Finally, miR-191/425 cluster upregulates zonula occludens-1 (ZO-1), a component of the tight junction barrier in epithelial and endothelial cells [52]; E-cadherin (CDH1), an important marker of epithelial tumor progression; and β-catenin (CTNN1) a component of wnt pathway that drives progression in various cancers [53]. 10.1371/journal.pgen.1003311.g004 Figure 4 miR-191/425 cluster alters gene expression profile of highly aggressive MDA-MB-231. (A, B) Unsupervised clustering of genes differentially expressed between miR-191/425 and scrambled oligonucleotide control transfected MDA-MB-231 cells. (C) qRT-PCR for the miR-191 and miR-425 modulated genes in two different aggressive breast cancer cells after miR-191/425 over-expression. Gene expression levels are reported as relative expression to GAPDH levels. All graphs showed p-values 90% was verified by fluorescent microscopy and confirmed by real-time PCR for miRs expression. Proliferation assays MDA-MB-231 cells, previously transfected with miR-191 or miR-425 precursors for 72 h, were plated (3000 per well) in 96-well plates and grown for 96 hours after transfection (final miRNA concentration of 100 nM) in normal culture conditions. MCF7 in normal culture conditions (+E2) transfected with anti-miR-191/425 and CTR oligonucleotide or in hormon deprivation conditions (−E2) transfected with miR-191/425 and CTR oligonucleotide were plated in 96-well plates and grown for 96 hours after transfection. Cell proliferation was documented every 24 hours for 4 days using a 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide assay kit (Promega, Madison, WI), and absorbance at 490 nm was evaluated by a SpectraMax 190 microplate reader (Molecular Devices, Sunnyvale, CA). Supporting Information Figure S1 miR-191/425 genomic locus. Schematic representation of the human (A) and murine (B) genomic locus of miR-191/425 cluster. miRNAs are represented with red lines. Green boxes represent the CpG islands. Arrowheads indicate the direction of the transcription. (TIF) Click here for additional data file. Figure S2 Co-expression of miR-191 and miR-425 with their host gene, DALRD3, in normal tissues and breast cancer cells. (A) Quantitative RT-PCR on mature miR-191, miR-425 and DALRD3 mRNA levels in 20 normal human tissues. (B) XY scatter plots to define the correlation between miR-191/425/DALRD3 expression in human normal tissues. (C) Expression levels of miR-191 and miR-425 in human breast cancer cells by qRT-PCR. All error bars indicate s.d. (TIF) Click here for additional data file. Figure S3 Expression of DALRD3 mRNA in breast cancer specimens and cancer cells. (A) DALRD3 transcript expression with different probes in breast tumor subtypes from Oncomine analysis. The first author and statistical significance are indicated. (B) SYBR qRT-PCR to discriminate the expression levels of the two main splicing variants of DALRD3 in 15 breast cancer cells. Isoform1 represents the splicing variants that may be responsible for the transcription of miR-191/425 cluster. (TIF) Click here for additional data file. Figure S4 miR-191 and miR-425 in situ hybridization (ISH) in human breast cancer. (A) In situ hybridization analysis of miR-191 and miR-425 expression in breast cancer tissues with different ERα expression status. Bars represent 200 µm. Two different cores for each microRNA and scrambled control oligonucleotide are represented for each category. Results are reported in the table as a percentage of the total number of ERα positive and ERα negative cores. (B) Co-labeling for miR-191 and miR-425 in human ERα positive breast tissue. Large and small arrows indicate tumor and stroma cells, respectively. (TIF) Click here for additional data file. Figure S5 miR-191/425 and estrogen regulation. (A) qRT-PCR on TFF1/pS2 and mature miR-17 upon E2 (10 nM) stimulation. MCF7 cells were hormone starved for 6 days and treated daily with estrogen for 72 h. (B) qRT-PCR on the primary precursor of mir-191 and miR-425 after E2 (10 nM) stimulation. (C) qRT-PCR for both splicing variant1 ad 2 of DALRD3 after hormone stimulation of MCF7 cells. (D) qRT-PCR for total DALRD3, splicing variants1 and 2, and TFF1/pS2 after hormone starvation of MCF7 cells (NT: untreated; HS: hormone starved). Error bars indicate s.d. and * represent p-value 1.2 and p-value 1.2. (XLS) Click here for additional data file. Table S2 Gene expression signature of miR-191 and miR-425 in MDA-MB-231 cells. All modulates genes with a fold change and p-value higher of 1.5 and 0.05 were reported. Modulated targets were obtained by comparison between miR-down-modulated gene list and Tagetscan predicted targets. (XLS) Click here for additional data file.
