MicroRNA-191, an estrogen-responsive microRNA, functions as an oncogenic regulator in human breast cancer

Carcinogenesis, Aug 2013

Estrogen- and microRNA-mediated gene regulation play a crucial role in breast cancer biology. However, a functional link between the two major players remains unclear. This study reveals miR-191 as an estrogen-inducible onco-miR in breast cancer, which promotes several hallmarks of cancer including enhanced cell proliferation, migration, chemoresistance and survival in tumor microenvironment. miR-191 is a direct estrogen receptor (ER) target and our results suggest existence of a positive regulatory feedback loop. We show miR-191 as critical mediator of estrogen-mediated cell proliferation. Investigations of mechanistic details of miR-191 functions identify several cancer-related genes like BDNF, CDK6 and SATB1 as miR-191 targets. miR-191 and SATB1 show inverse correlation of expression. miR-191-mediated enhanced cell proliferation and migration are partly dependent on targeted downregulation of SATB1. Further, functional validation of estrogen:miR-191:SATB1 link suggests a cascade initiated by estrogen that induces miR-191 in ER-dependent manner to target SATB1, a global chromatin remodeler, thereby contributing to estrogen-specific gene signature to regulate genes like ANXA1, PIWIL2, CASP4, ESR1/ESR2, PLAC1 and SOCS2 involved in breast cancer progression and migration. Overall, the identification of estrogen/ER/miR-191/SATB1 cascade seems to be a significant pathway in estrogen signaling in breast cancer with miR-191 as oncogenic player.

A PDF file should load here. If you do not see its contents the file may be temporarily unavailable at the journal website or you do not have a PDF plug-in installed and enabled in your browser.

Alternatively, you can download the file locally and open with any standalone PDF reader:

https://carcin.oxfordjournals.org/content/34/8/1889.full.pdf

MicroRNA-191, an estrogen-responsive microRNA, functions as an oncogenic regulator in human breast cancer

NehaNagpal 1 Hafiz M.Ahmad 0 1 BhuvanMolparia 1 RituKulshreshtha 1 0 School of Life Sciences, Jawaharlal Nehru University , New Delhi 110067, India 1 Department of Biochemical Engineering and Biotechnology, Indian Institute of Technology , New Delhi 110016, India The Author 2013. Published by Oxford University Press. All rights reserved. For Permissions, please email: - *To whom correspondence should be addressed. Tel:+91-11-26591017; Fax: +91-11-26582282; Email: Estrogen- and microRNA-mediated gene regulation play a crucial role in breast cancer biology. However, a functional link between the two major players remains unclear. This study reveals miR191 as an estrogen-inducible onco-miR in breast cancer, which promotes several hallmarks of cancer including enhanced cell proliferation, migration, chemoresistance and survival in tumor microenvironment. miR-191 is a direct estrogen receptor (ER) target and our results suggest existence of a positive regulatory feedback loop. We show miR-191 as critical mediator of estrogen-mediated cell proliferation. Investigations of mechanistic details of miR-191 functions identify several cancer-related genes like BDNF, CDK6 and SATB1 as miR-191 targets. miR-191 and SATB1 show inverse correlation of expression. miR-191-mediated enhanced cell proliferation and migration are partly dependent on targeted downregulation of SATB1. Further, functional validation of estrogen:miR-191:SATB1 link suggests a cascade initiated by estrogen that induces miR-191 in ER-dependent manner to target SATB1, a global chromatin remodeler, thereby contributing to estrogen-specific gene signature to regulate genes like ANXA1, PIWIL2, CASP4, ESR1/ESR2, PLAC1 and SOCS2 involved in breast cancer progression and migration. Overall, the identification of estrogen/ER/miR-191/SATB1 cascade seems to be a significant pathway in estrogen signaling in breast cancer with miR-191 as oncogenic player. Introduction Estrogen plays an important role in development and progression of hormone-dependent breast cancer (1). It binds and activates nuclear transcription factors (estrogen receptor [ER] and ) to induce a whole plethora of genes involved in the tumorigenesis process (2). We now know that apart from protein coding genes, estrogen also regulates a class of small non-coding RNAs called microRNAs (miRNAs) (3). Conversely, few miRNAs have been identified that target the transcription factors ER- and and thus impact estrogen action (1). However, estrogen:miRNA correlation has not much been explored beyond this point leaving a lacuna regarding functional relevance of this association. Several miRNAs have been reported before to be critical components of various signaling pathways and thus act as onco-miRs or tumor suppressors (46). Thus, it is conceivable that miRNAs regulated by estrogen might act as mediators of its varied functions in breast cancer biology. Abbreviations: BCL11A, B-cell lymphoma/leukemia 11A; BDNF, brainderived neurotrophic factor; cDNA, complementary DNA; CDK6, cell division protein kinase 6; CHIP, chromatin immunoprecipitation; E2, 17-estradiol; EGR1, early growth response protein 1; ER, estrogen receptor; ERE, estrogen response element; FOXP1, forkhead box protein P1; miRNAs, microRNAs; MM231, MDA-MB-231; MM453, MDA-MB-453; MTT, 3-(4,5-dimethylthiazole-2-yl)-2,5-diphenyl tetrazolium bromide; MN, monoclonal; PC, polyclonal; PLCD1, phosphoinositide phospholipase C-delta-1; qRTPCR, quantitative reverse transcriptionPCR; siRNAs, small interfering RNAs; Tm, tamoxifen; UTR, untranslated region. Our results in this study identify miR-191 as an estrogen-inducible miRNA in ER-dependent manner in breast cancer. In the context of clinical breast cancer biology, patients with ER+ tumors show higher miR-191 levels as compared with ER patients (7,8). Interestingly, miR-191 is reported to be overexpressed in several cancers including breast, colon, lung, liver, prostate, pancreas, stomach, ovarian cancer, pituitary adenoma and acute myeloid leukemia though downregulated in severe medulloblastomas, thyroid follicular tumor and melanoma (820). Considering an extensive miR-191:cancer correlation, with minimal reports on its functions, it certainly becomes particularly attractive molecule to work on. We, thus, sought to decode its function in breast cancer with respect to its regulation, cellular functions and target transcripts. Based on our results, miR-191 emerges as oncogenic miRNA in breast cancer involved in controlling various hallmarks of cancer. Exploration of its targetome reveals several genes involved in cancer as its bonafide targets including a chromatin remodeler and transcription factor, SATB1. Functional relevance of estrogen:miR-191 link shows miR-191 as critical mediator of estrogen-mediated cell proliferation. Importantly, overexpression of SATB1 in ER+ cells partially inhibits miR-191-mediated cell proliferation and migration. Additionally, a partial overlap is seen in estrogen, miR-191 and SATB1 gene signatures. Thus, estrogen:ER:miR-191:SATB1 pathway seems to be a significant pathway for breast cancer biology. Materials and methods The breast cancer cell lines (MCF7 and MDA-MB-231 [MM231]) were a kind gift from Dr Mircea Ivan (Indianapolis University, source ATCC). MDA-MB-453 (MM453) and T47D were obtained from NCCS, Pune Cell Repository. MCF10A were a kind gift from Dr Annapoorni Rangarajan (IISc, Bangalore, India). MCF7, MM453, T47D and MM231 were maintained in RPMI 1640 (GIBCO) medium. For MCF10A, cells were grown in Dulbecco's modified Eagle's medium:F12 (GIBCO) supplemented with growth factors hEGF (Sigma; 20 ng/ml), hydrocortizone (Sigma; 0.5 g/ml) and insulin (Sigma; 10 g/ml). The media was supplemented with 100 U/ml penicillin, 100 g/ml streptomycin and 10% fetal bovine serum and incubated at 37C in 5% CO2 incubator (Shell labs). Transient and stable transfections Cells were seeded (5 105 cells per well) in six-well plates. Different concentrations (1030 nM) of pre-miR-191, anti-miR-191 (3070 nM) or their respective controls (Ambion, Exiqon) were transfected using lipofectamine 2000 in various cell lines used. In case of plasmids, 2.5 g of plasmid was transfected into six-well plates using lipofectamine 2000 (Invitrogen). In case of small interfering RNAs (siRNAs), 100 nM of each of the siRNA (Sigma) or universal ctrl siRNA (Sigma) was transfected using lipofectamine 2000. After 6 h, media was changed and the samples were assayed after 2448 h for transcript levels (through quantitative reverse-transcriptionPCR [qRTPCR]) or other cellular assays. Each experiment was repeated three times. Construction of stable polyclonal celllines The miR-191 precursor along with its flanking sequences (~150200 Nt) was amplified from MCF7 genomic DNA using PCR and then cloned in vector pBABE-puro and further confirmed through restriction digestion and sequencing. The miR-191 overexpressing clone or the empty pBABEpuro parent vector as control was then transfected into MCF7 using lipofectamine 2000 (Invitrogen). The stable cell lines monoclonals (MNs) and polyclonals (PCs) were made by selecting cells in 1 g/ml puromycin for up to 1 month. Cell proliferationassay In vitro proliferation was measured by seeding cells of different cancer cell lines (1 104 cells per well) in 24-well plates in triplicates, and level of miR-191 was modulated by transfecting them with 30 nM of pre-miR-191, anti-miR-191 and their scramble controls using lipofectamine 2000 and cell proliferation was scored 72 h post-transfection using 3-(4,5-dimethylthiazole2-yl)-2,5-diphenyl tetrazolium bromide (MTT), quantified at 595 nm. Stable overexpressing clones of miR-191 were also seeded in the same manner (1 104 cells per well) in 24-well plates in triplicates and scored for cell proliferation for five consecutive days using MTT, quantified at 595 nm. The SATB1 is overexpressed in these stable overexpressing clones by transfecting them with the SATB1 complementary DNA (cDNA) clone (Origene) using lipofectamine 2000 and cell proliferation was scored 72 h post-transfection using MTT, quantified at 595 nm. siRNAs against ER-//ctrl (Sigma) were used to downregulate the endogenous level of ER-/. Cell proliferation was scored 72 h post-transfection, whereas estrogen treatment was given 48 h posttransfection. The experiment was repeated three times in triplicates. Cell viability assay Cell viability was determined by staining a small aliquot of cell suspension with 0.5% trypan blue saline solution and then examining/counting the cells in a hemocytometer. The experiment was repeated three times in triplicates. Colony formationassay MCF7 cells transiently transfected with miR-191 mimics (pre-191) or antisense (anti-miR-191) or respective control oligos (ctrl and Nctrl) and 24 h post-transfection cells were seeded (2500 cells per well) in six-well plates in duplicate and allowed to grow with media changes every 3 days till the colonies become visible by eye. Colonies were then stained with crystal violet postfixation with formaldehyde. Colonies were imaged and quantified. The experiment was repeated three times. Soft agar assay Cells (5 104 cells per well in a six-well plate) were resuspended in the medium containing 0.7% agar so that final agar concentration would be 0.35%. This cell suspension was then placed over a layer of 0.5% agar in the medium and cultured at 37C, 5% CO2 with media changes every 3days till the colonies become visible by eye. After 34 weeks, media was aspirated and cells were stained with crystal violet (1:1000 in Dulbecco's phosphate-buffered saline). The experiment was repeated three times. Cell cycle analysis MCF7 cells were transiently transfected with miR-191 mimics (pre-191) or antisense (anti-miR-191) or respective control oligos (ctrl and Nctrl) and 48 h post-transfection cells were fixed with the 70% ethanol (in Dulbecco's phosphate-buffered saline). Cells were then stained with propidium iodide for 30 min at 37C followed by cell cycle analysis. Analysis of cell cycle was done by fluorescence-activated cell sorting (BD Bioscience). The data obtained were analyzed using Cyflogic software. The experiment was repeated three times. Scratch assay Cells were seeded (4 105 cells per well in a six-well plate) in duplicate. Scratches were created by the use of a 10 l sterile pipette tip, and any cellular debris was removed by washing with phosphate-buffered saline. The scratched layer was then incubated in medium containing 10% fetal bovine serum for 72 h and photographed under a light microscope. The area of migration was then estimated by choosing an arbitrary unit scale. The experiment was repeated three times in triplicates. Boyden chamber assay Trans-filter migration assays were performed on MCF7 cells in serum-free Dulbecco's modified Eagle's medium with 8.0 m pore inserts on a 24-well Transwell (BD Bioscience). The cells were transfected with pre-191, anti-191 alone or along with SATB1 and their respective controls. After 24 h post-transfection, the cells were transferred to the upper chamber of the transwell for migration. Migrating cells were fixed with formalin followed by staining with crystal violet dye (0.1%). The stained membrane was then washed and quantification was done postelution with 10% acetic acid (595 nm; iMark microplate reader [Bio-Rad]). The experiment was repeated three times in triplicates. Hypoxia/stress/drug treatment Cells were seeded in 24-well plate (1 104 cells per well) in triplicates. For hypoxic stress, the cells were placed in hypoxia chamber (Invivo200; Ruskin) maintained at 0.2% oxygen, 5% CO2, for various period of times as indicated. The cells growing under normal conditions (21% oxygen) were taken as control. For serum starvation, the full medium was replaced with that of without serum medium for up to 5days. For drug treatment, different concentration of drugs such as doxorubicin (12 M), bleomycin (35 M) and mitomycin C (1.53 M) was added. Cells treated with equal concentration of phosphatebuffered saline were treated as controls. Cell proliferation was quantified after 72 h post-transfection (in transients) and for 45 days (in stables) by MTT assay. The experiment was repeated three times. Total RNA isolation Total RNA isolation from cell lines was done using Trizol reagent (Invitrogen), according to manufacturers instructions. Microarray expression profiling and analysis Cells were transfected with miR-191 mimics or scramble controls (30 nM each) using lipofectamine 2000 reagent (Invitrogen). After 48 h post-transfection, total RNA was isolated and sent for microarray analysis to Ocimum Biosolutions. The gene expression data were generated using Affymetrix Human Gene 1.0 ST arrays and Affymetrix Expression Console was used for quality control analysis. A1.5-fold change cut off resulted in False Discovery Rate < 0.001 when compared. R-programming language was used for statistical analysis and GenowizTM software was used for biological analysis. A total of 757 genes were upregulated, whereas 463 genes were downregulated in comparison with that of control when miR-191 was overexpressed (Supplementary Data 5, available at Carcinogenesis Online). Stem-loop RTPCR Taqman assay (Applied Biosystems) or stem-loop RTPCR was done to determine the level of miR-191 and RNU6B (used for normalization) in all the samples. cDNA was made with RT primer (miR-191RT and U6RT) specific for miR-191 and U6, respectively. PCR was performed using specific forward and a common stem-loop reverse universal primer for both miR-191 and U6. The sequences of the primers are given in Supplementary Data 8, available at Carcinogenesis Online. Hormonal treatment MCF7 cells were grown in phenol redfree RPMI (GIBCO) media supplemented with 10% dextran-stripped serum (Life Technologies), 100 U/ ml penicillin, 100 g/ml streptomycin and grown at 37C and 5% CO2 in incubator (Shell labs). After 4 days, cells were seeded (5 105 cells per well) in six-well plates. Twenty-four hours postseeding estrogen/tamoxifen (Tm)/PPT (ER- agonist)/ERB041 (ER- agonist) treatment was given (concentration mentioned in the graphs) and the samples were assayed after 048 h of hormonal treatment for miR-191 levels. Target prediction The probable targets for miR-191 were selected on the basis of prediction by established target prediction programs. The targets were selected if they were found to be predicted by at least four target prediction programs out of 11 available target prediction programs TargetScan, PITA, PicTar, miRanda, mirTarget2, NBmirTar, RNAhybrid, MiTarget, MicroInspector, RNA22 and DIANA MicroT (21). The list (Supplementary Data 6, available at Carcinogenesis Online) was further narrowed down by considering targets that have previously known to be associated with cancer. qRTPCR Cells were transfected with miR-191 mimic (30 nM) and antisense (60 nM) (sense and scramble controls) using lipofectamine 2000 reagent (Invitrogen) and after an incubation of 48 h, total RNA was isolated and reverse transcribed using Revertaid first strand cDNA synthesis kit (Fermentas) using oligo-deoxythymidine as primer. cDNA formed was then further amplified for the predicted genes with respective primers using Power Sybr Green master mix kit (Ambion). Glyceraldehyde 3-phosphate dehydrogenase was used as control for the normalization of the data. Alist of primer sequences used for amplification is given in Supplementary Data 8, available at Carcinogenesis Online. Construction of 3 untranslated region- or promoter-luciferase constructs To determine whether miR-191 downregulates its target transcripts through its direct binding to the 3 untranslated region (UTR), the region of 3UTRs of the genes (SATB1, cell division protein kinase 6 [CDK6], B-cell lymphoma/ leukemia 11A [BCL11A], forkhead box protein P1 [FOXP1], phosphoinositide phospholipase C-delta-1 [PLCD1], brain-derived neurotrophic factor [BDNF] and early growth response protein 1 [EGR1]) containing miR-191 binding site was amplified (using Taq polymerase; Fermentas) and cloned in the luciferase reporter vectors (pMir-Report) downstream of a firefly luciferase gene. To further confirm whether the predicted target is the functional target of miR191, the miR-191 binding site in the 3UTR of the gene (SATB1) was mutated through site-directed mutagenesis using inverse PCR and luciferase activity is observed (see Supplementary Data 8, available at Carcinogenesis Online, for details of primers). For promoter analysis, we amplified the miR-191 promoter fragments predicted to encompass estrogen-responsive elements. Two estrogen response elements (EREs) upstream of miR-191, with 100% match, with ER- and consensus were cloned downstream of a luciferase promoter vector PGL3-tk-luciferase (Promega). All the clones were confirmed with PCR, restriction digestion and sequence analysis. Dual luciferaseassay For luciferase assays, MCF7 cells (10000 cells per well) were cotransfected with 3UTR luciferase constructs along with pBABE/ pBABE-191vector (for target confirmation) and cotransfected with predicted ERE luciferase constructs/parent vector (PGL3-tk-Luciferase) along with ER-//parent vector (ctrl) using lipofectamine 2000 (Invitrogen). pRL-TK was transfected in all the wells for normalization of transfection efficiency. The activities of firefly (Photinus pyralis) and renilla (Renilla reniformis) luciferase were quantified with the dual luciferase reporter assay (Promega) after 48 h post-transfection. Western blot MCF7 cells were grown in phenol redfree RPMI (GIBCO) media supplemented with 10% dextran-stripped serum (Life Technologies), 100 U/ml penicillin, 100 g/ml streptomycin and grown at 37C and 5% CO2 in incubator (Shell labs). After 4days, cells were seeded (5 105 cells per well) in six-well plates. Twenty-four hour postseeding estrogen treatment (109 M) was given and the samples were assayed after 048 h of hormonal treatment for ER levels. Further for SATB1 levels, cells were transfected with differential level of miR-191 (pre191/antisense miR-191 and their respective controls) or siRNA for ER-//control and assayed after 48 h of post-transfection. Cells were lysed; protein concentration was determined by using Bradford reagent (SigmaAldrich). Equal amount of protein lysates was separated with 10% sodium dodecyl sulfatepolyacrylamide gel electrophoresis and transferred to nitrocellulose membrane (Amersham Hybond ECL). The membrane was then probed with a specific primary antibody at a dilution of 1:1000 (SATB1; Cell Signaling Technology) or 1:3000 (-actin; Invitrogen) or 1:2000 (ER- or ; Santa Cruz Biotechnology) or 1:2000 (Tubulin; Cell Signaling Technology) followed by washing and incubation with respective secondary antibody (anti-rabbit, horseradish peroxidase-linked; SATB1, ER-, ER- and anti-mouse horseradish peroxidase-linked; -actin and Tubulin). The specific protein band was visualized by autoradiography using an ECL kit (Amersham ECL Prime). Chromatin immunoprecipitation MCF7 cells were (cultured in phenol redfree medium with dextran-stripped serum for 34 days) treated with 108 M estrogen for 13 h, then crosslinked (formaldehyde), washed with phosphate-buffered saline and resuspended in the RIPA buffer. Sonication was performed to obtain DNA fragment of 100500 bp fragments (sonicated lysate was confirmed for fragment size through 2% agarose gel). For chromatin immunoprecipitation (CHIP), specific antibodies against ER-/ (Santa Cruz Biotechnology), salmon sperm DNA (Sigma) and protein A-sepharose beads (30 l) were added to the chromatin extract and incubated overnight, washed and eluted with 0.5% w/v sodium dodecyl sulfate solution. Decrosslinking was done at 65 for 45 h and DNA was then purified with PCR purification kit (HiMedia). Rabbit IgG was used as control antibody, whereas the chromatin extract without any antibody/beads treatment was used as positive control. For DNA sequencespecific quantification, qRTPCR was done with equal amount of chromatin extract using sequence-specific primers (Supplementary Data 8, available at Carcinogenesis Online). The experiment was repeated twice. Levels of miR-191 have been reported to be high in various solid cancers including breast cancer. miR-191 levels were quantitated in a panel of breast cancer cell lines. As previously reported, we noticed higher miR191 levels in all cancer cell lines tested as compared with non-tumorigenic human breast epithelial cell line, MCF10A. Interestingly, levels of miR-191 were much higher in ER+ (MCF7 and T47D) versus ER cell lines (MM231 and MM453) (Figure 1A, upper panel; Supplementary Data 1, available at Carcinogenesis Online). We, thus, sought to find the functional relevance of this observation and scored for impact of miR191 overexpression/inhibition on various cancer properties. Oncogenic miR-191 imparts multiple hallmarks ofcancer Cell proliferation and anchorage independence. The expression of miR-191 was transiently modulated in a panel of breast cancer cell lines using miR-191 mimics (pre-191) or LNA-antisense (anti-191) along with the respective controls (ctrl and Nctrl) and cell proliferation was measured. Remarkably, we found increased proliferation on miR-191 overexpression as opposed to reduced proliferation on its inhibition in ER+ cell lines (MCF7 and T47D), whereas ER cell lines (MM231 and MM453) showed negligible effects (Figure 1A, lower panel; Supplementary Data 1, available at Carcinogenesis Online). As MCF7 shows maximal proliferative effects following modulation of miR-191 levels, we chose to focus on MCF7 for further studies. We generated PC (ctrl and PC191 [miR-191 overexpressing]) and MN (ctrl and MN191 [miR-191 overexpressing]) stable cell lines in MCF7 to perform various cellular assays. In accordance to the transient data, miR-191-overexpressing PC191 and MN191 showed more proliferation (1.7- and 1.68-fold, respectively) than the control (Figure1B and C; Supplementary Data 1, available at Carcinogenesis Online). Interestingly, miR-191 promoted MCF7 cell proliferation in a dose-dependent manner when tested through transient assays (Supplementary Data 2, available at Carcinogenesis Online). Colony formation assay was performed using stable MCF7 cell lines or MCF7 cells transiently transfected with pre- or anti-miR-191. miR-191 overexpressors formed more (~50% more) colonies than controls, whereas the vice versa was seen on miR-191 downregulation through anti-191 (~28% lesser colonies than control) (Figure1D). We next used a soft agarbased clonogenic assay to determine whether miR-191 overexpression promotes anchorage-independent growth in MCF7 cells. This assay revealed that both miR-191 overexpressors significantly induced colony formation in soft agar (Figure1E). These results further validate a pro-proliferative effect of miR-191 in breast cancer. Migration. We next assayed whether miR-191 imparts the capability of migration to the breast cancer cells. The scratch assay was performed and around 1.5-fold wound healing was observed on miR-191 overexpression, whereas healing got reduced to 0.56-fold on miR-191 inhibition, compared with that of their respective controls (Figure 1F). Role of miR-191 in supporting the cell migration was then further evaluated by Boyden chamber assay and it was found that rate of cell migration is increased to ~1.6-fold by miR191 overexpression and got decreased to 0.4-fold on its inhibition (Figure 1G). Thus, both scratch and Boyden chamber assays show that miR-191 overexpression augments cellular migration in breast cancer. Stress/hypoxia resistance and chemoresistance. Several oncogenes have been reported before to impart survival and growth advantage to the cancer cells (22,23). We, thus, tested whether miR191 promotes survival of cells under conditions present in the tumor microenvironment (nutrient deprivation and hypoxia) or promote aggressiveness by imparting drug resistance. MCF7 cells transfected with miRNA oligos (sense, antisense or ctrl) were exposed to various conditions/drugs and 48 h later, dead cells were quantified using trypan blue assay (Figure 1H) or fluorescence-activated cell sorting analysis (Figure1I). Notably, miR-191-overexpressing cells showed better survival (P < 0.05, n = 3), whereas its inhibition led to the reverse (P < 0.05, n = 3), compared with that of control under all the conditions tested signifying role of miR-191 in promoting chemoresistance and cell survival under tumoral stress (Figure 1H and I and Supplementary Data 2, available at Carcinogenesis Online). Notably, the various cellular assays were also performed in T47D, another ER+ cell line, and we found that miR-191 promoted proliferation, migration, stress and drug resistance in this cell line similar to MCF7 (Supplementary Data 2a and b, available at Carcinogenesis Online). Transcriptional regulation of miR-191: estrogen and tumor microenvironment MiRNAs are RNA polymerase II regulated genes and thus are prone to dynamic regulation by various factors (24). Considering that no major amplifications/deletions/translocations have been reported for miR-191 locus so far, we next evaluated the contribution of tumor microenvironment or hormonal factors (estrogen) for high miR-191 levels in breast cancer. Interestingly, we found that conditions usually present in tumor microenvironment (hypoxia and nutrient deprivation) induced miR-191 transcripts up to 34fold (Figure2A). Estrogen has been widely reported to be an important factor involved in breast cancer proliferation by regulating specific gene expression (25). Thus, it was imperative to study if estrogen was involved in regulation of miR-191 levels. Estrogen treatment of MCF7 cells brought about significant increase (~5-fold) in miR-191 levels that could be reversed following treatment with an estrogen antagonist (Tm) (Figure 2B). A time coursedependent miR-191 induction was seen in response to estrogen treatment (Figure 2C; Supplementary Data 1, available at Carcinogenesis Online). Next, to confirm whether the effect of estrogen is through its receptors ER- and/or ER-, we transiently modulated the levels of receptors in MCF7 cells and checked miR-191 levels. Indeed, overexpression of ER-/ or their agonists (PPT for ER- and ERB041 for ER-) induced miR-191 levels, whereas their knockdown using siRNAs against ER- and or ER inhibition with Tm treatment reduced miR-191 levels significantly (Figure 2D; Supplementary Data 1, available at Carcinogenesis Online). Estrogen is known to regulate gene expression by binding to ERs, which subsequently binds to ERE present in the promoter region (2,3,26). In silico analysis of region ~5 kb upstream and 1 kb downstream of miR-191 using Promo 3.0 software showed presence of seven putative ERE sequences showing <1% dissimilarity with the ERE consensus (Figure 2E; Supplementary Data 3, available at Carcinogenesis Online) (27). The regions (A: 25 bp upstream and B: 4.2 kb upstream of 5 end of pre-miR-191) bearing ERE were cloned in luciferase promoter vector and scored for luciferase activity as described in Materials and methods. The region Ashowed up to 5-fold induction with both ER- or , whereas the region B showed up to 5-fold induction with ER- and only 2-fold induction with ER- consistent with 100% match with ER- consensus (Figure2F). Further, to check dynamic recruitment of ER-/ to the ERE elements following estrogen treatment, we resorted to CHIP experiment. CHIP analysis was done in MCF7 cells with or without estrogen treatment with the antibodies specific to ER- and or IgG control. The results show that following estrogen treatment, various putative EREs show dynamic recruitment of ER- or and the impact of ER- is much more pronounced on the induction compared with that of ER- (Figure2G and H). Thus, the induction of miR-191 by 17-estradiol (E2) is mediated by binding of ER-/ to various EREs present upstream of miR-191. MiR-191 exists as a cluster with miR-425. We, thus, measured levels of miR-425 and found that miR-425 is also induced by serum starvation and hypoxia (Supplementary Data 4, available at Carcinogenesis Online). We also show that miR-425 is ER inducible and shows higher levels in ER+ (MCF7) cell line as compared with ER (MM231) cell line (Supplementary Data 4, available at Carcinogenesis Online). This suggests that apparently miR-191 and miR-425 are co-regulated as part of a cluster through a common promoter. MiR-191 targetome MiRNAs are post-transcriptional gene regulators that work either through direct transcript binding or indirectly by affecting the expression levels of transcription factors/regulators that translate to elaborate effects on the transcriptome (28,29). To evaluate effects of miR-191 on transcriptome, we transiently overexpressed mature-miR-191 oligos or scramble in MCF7 cells and performed gene expression profiling. Atotal of 757 genes were found to be induced and 463 genes were found to be downregulated by >1.5-fold on miR-191 overexpression (Supplementary Data 5, available at Carcinogenesis Online). Correlations made to the breast cancer database (G2SBC) identified 78 genes commonly altered in breast cancer to be differentially regulated on miR-191 overexpression (Supplementary Data 5, available at Carcinogenesis Online) (30). Interestingly, analysis of the data using DAVID bioinformatics resources showed olfactory transduction pathway (P = 1.0E-7) followed by the calcium signaling pathway (P = 9.2E-2) as the most highly represented pathways for the genes induced by miR-191 (31). A pie chart showing specific functional categories of genes altered in miR-191-overexpressing cells as compared with control cells (P < 0.05) is presented (Supplementary Data 5, available at Carcinogenesis Online). Then, we wanted to find the direct targets of miR-191. Astringent criteria of prediction by at least four target prediction programs resulted in identification of 40 putative targets of miR-191 (Supplementary Data 6, available at Carcinogenesis Online) (21). Due to large number of these targets that were experimentally difficult to contemplate, we further focused on those that were previously known to be associated with cancer. We finally shortlisted seven targets (BCL11A, BDNF, CDK6, SATB1, EGR1, FOXP1 and PLCD1) for confirmation by experiments. MCF7 cells were transfected with miR-191 (sense, antisense or scramble controls) and after 48 h post-transfection, the transcript levels of the shortlisted targets were determined by qRTPCR (Figure 3A). Interestingly, although some of the targets (SATB1, PLCD1 and CDK6) show downregulation, the others (such as BCL11A and BDNF) show the opposite trend on miR-191 overexpression. EGR1 and FOXP1 remain largely unaffected. To further confirm direct targeting of these transcripts by miR-191, a 3 UTR-luciferase reporter assay was performed for seven putative target transcripts (SATB1, PLCD1, BDNF, BCL11A, FOXP1, EGR1 and CDK6). The 3 UTR-luciferase assay results showed that SATB1 and CDK6 are downregulated by miR-191, whereas BDNF is induced by miR-191 in agreement with the qRTPCR data (Figure3B). BCL11A, PLCD1, FOXP1 and EGR1 remain unaffected or showed minimal effects in luciferase assay. Afigure representing the confirmed miR-191 targets SATB1, CDK6 and BDNF is shown highlighting the miR-191 recognition elements (Figure 3C). Thus, overall, based on above results, we confirm SATB1, CDK6 and BDNF as target transcripts of miR-191 in MCF7 cells with maximal effects been observed for SATB1. SATB1: a confirmed target of miR-191 SATB1 is special AT-rich binding protein that holds a controversial though significant role in breast cancer biology (32). We, thus, dissected miR-191:SATB1 functional correlation. We noticed that sequence of both miR-191 and its binding site (8mer) in SATB1 3 UTR is highly conserved across various species (Supplementary Data 7, available at Carcinogenesis Online). Next, miR-191 (sense, antisense along with the controls) was transiently overexpressed in MCF7 cells and SATB1 transcript levels were measured 48 h posttransfection. miR-191 overexpression decreased SATB1 transcript up to 0.3-fold, whereas miR-191 antisense brought about its induction up to 1.33-fold (Figure4A). Similar effects were also seen at the protein level by western blotting using SATB1 antibody (Figure4B). The miR-191-overexpressing stable cell lines (PCs and MNs) also showed strong downregulation of SATB1 both at RNA and protein levels, thus reassuring our presumption of SATB1 as direct miR-191 target (Figure4A and B). Further validation of the same was done by scoring unmutated or mutated SATB1 3 UTR (achieved through sitedirected mutagenesis at the miR-191 binding site in the 3 UTR) for luciferase activity on cotransfection with pBABE-miR-191 plasmid or pBABE alone. As anticipated, miR-191 overexpression brought about a decrease in the luciferase activity with wild-type miR-191 3 UTR construct, whereas no inhibition was observed when the mutated 3 UTR was used (Figure4C and D). These data collectively confirm that miR-191 directly targets SATB1 by binding to its 3 UTR at position 717724. We anticipate that a miRNA and its target will show inverse correlation of expression. In accordance to this, SATB1 showed much higher levels in ER (MM231) as opposed to ER+ (MCF7) cell line (Figure4E) in contrast with the level of miR-191. Consistent to this result, we found that overexpression/inhibition of both ER- or in MCF7 cells modulated SATB1 levels exactly in contrast with that of miR-191 levels under the same conditions (Figure4F). Similarly, SATB1 levels were also found to be downregulated following hypoxia (48 h) or serum starvation treatment (72 h) in contrast with miR-191 induction under same conditions (Figure4G). These results show that miR-191 and SATB1 show inverse correlation of expression under various conditions tested, thus confirming SATB1 as bonafide miR191 target. Functional repercussions of miR-191-mediated downregulation of SATB1 To check if miR-191-mediated cellular effects are mediated through downregulation of SATB1, miR-191 oligos were transfected in MCF7 cells with or without SATB1 overexpression and subsequently scored for cell proliferation/migration. Interestingly, miR-191 (pre-191; 30 nM) mediated (~1.7-fold; P = 4.7559E-05) increase in cellular proliferation was reduced to 1.17-fold (P = 0.00085624), when SATB1 protein was replenished in miR-191-overexpressing MCF7 cells (Figure 5A). Similar results were obtained with miR-191-overexpressing stable cell lines (Figure 5A). The results were further confirmed with soft agar assay (Figure5B). This suggests that miR-191 mediated increase in cellular proliferation is partially dependent on targeted downregulation of SATB1. The effects on cell migration were checked using wound healing and Boyden chamber assay. MCF7 cells were transfected with miR-191 oligos (sense, antisense or controls) with and without SATB1 and scored for cellular migration (Figure5C and D). Interestingly, miR-191 mimics mediated increase in cellular migration was inhibited when transfected along with SATB1. In contrast, anti-miR-191 mediated decrease in cellular migration was enhanced on treatment with SATB1 (Figure 5C). Further experiments on stable cell lines using wound healing assay are also in accordance showing much faster healing of the scratch in miR-191 overexpressors that was inhibited significantly on SATB1 overexpression (Figure 5D). Thus, miR-191-induced cell proliferation and migration seems to be partially dependent on SATB1 downregulation. SATB1 as a chromatin remodeler is known to affect global gene signatures (32). Comparison of the miR-191 gene expression signature with that of SATB1 showed common 13 genes that were inversely regulated by SATB1 and miR-191 (Figure5E) (32). We confirmed the expression level of three such genes: SOCS2 (Suppressor of cytokine signaling 2a candidate tumor suppressor gene involved in cell proliferation and tumor growth in breast cancera SATB1-inducible gene), PLAC1 (placenta-specific 1 genea candidate oncogene in breast cancer, ER-inducible and SATB1-repressed gene) and CASP4 (Caspase-4, involved in the signal transduction pathways of apoptosis, necrosis and inflammation) (3335). We show that miR-191 inhibition or SATB1 overexpression brings about induction of tumor suppressor SOCS2 and downregulation of oncogene PLAC1, whereas vice versa was seen on miR-191 overexpression in MCF7 cells (Figure 5F). However, expression of CASP4 was found to be downregulated by both miR-191 and SATB1 overexpression. Collectively, these results of shared gene expression signature evoke a possibility that miR-191 overexpression may partly affect SATB1 gene signature in breast cancer to implement its functions. Estrogen-mediated cellular proliferation is miR-191 dependent We next wanted to check whether estrogen-mediated cell proliferation is miR-191 dependent. We compared the rate of E2-mediated cell proliferation in presence of miR-191 mimics, antisense or scramble controls and found that effect of E2 on cell proliferation was more when miR-191 was transiently or stably overexpressed, whereas a reduction was seen in presence of antisense miR-191 (Figure6A and B). A similar trend was evident when colony formation assay was performed for MCF7 cells treated with estrogen in presence of pre191 or anti-191. Pre-191 treatment promoted ~48% increase in colony number, whereas anti-191 treatment showed 33% less colonies as compared with their respective controls in presence of estrogen, thus emphasizing the role of miR-191 as a mediator of estrogen-induced cell proliferation (Figure6C). Furthermore, we compared the estrogen/ER gene signature with miR-191 signature and identified 51 genes that were commonly regulated (Figure6D (25,3638)). We confirmed transcript levels of few genes known to promote E2/ER-mediated cell proliferationESR1, ESR2, piwi-like 2, PLAC1 and AnxA1 (35,36,39,40). Estrogen treatment of MCF7 cells mediated induction of these genes that was reversed by treatment with estrogen antagonist (Tm) (Figure6E). We then checked expression of these genes in response to miR-191 overexpression (Figure6F). Interestingly, the qRTPCR data in agreement with microarray data further confirmed that miR-191 induced the above-said E2-inducible genes suggesting that E2-mediated cellular effects may partially bear miR-191 signature. More importantly, miR191-mediated ESR1 and ESR2 induction suggests a positive feedback loop involved in miR-191 induction. Overall, through our studies, we show miR-191 as an estrogeninducible miRNA involved in regulating several hallmarks of cancer by targeting genes involved in breast cancer pathogenesis. Aproposed model detailing miR-191 regulation, function and targets as part of estrogen signaling in breast cancer is shown (Figure6G). Discussion Estrogen plays an important role in the development and progression of breast cancer; however, the exact mechanism remains unclear. Till date, estrogen and miRNA correlation has been limited to reports of aberrant miRNA expression in estrogen-dependent breast cancer or several miRNAs targeting ERs. The results presented in this study identify functional relevance and targetome of estrogen-regulated signature (E). qRTPCR data showing transcript levels of SOCS2, PLAC1 and CASP4 on miR-191 or SATB1 overexpression/inhibition (F). Glyceraldehyde 3-phosphate dehydrogenase has been used for normalization of qRTPCR data. The graphical data points in A, C and F represent mean SD of at least three independent experiments. (*P > 0.05, **P < 0.05). Error bars denote SD. miRNA, miR-191 as an important oncogenic mediator of estrogen signaling involved in imparting aggressive phenotype to breast cancer. We show that estrogen induces miR-191 through dynamic binding of ERs to specific EREs in the promoter of miR-191. In accordance, higher levels of miR-191 in ER+ versus ER patients have been reported by other groups emphasizing its potential as a diagnostic marker (7,8). Experiments involving ER-/ overexpression, inhibition, use of their specific agonists or CHIP assay convincingly demonstrate that miR-191 expression can be stimulated by both ER- and in a ligand-dependent manner. Recent studies demonstrate that both the ERs can bind to the same ERE and thus share a large number of common target genes, although both ERs also have their distinct downstream targets to regulate unique functions (41,42). This is particularly interesting due to the fact that both the ERs are considered to display opposite functions, with ER- affecting the ER--mediated transcription and eventually cell proliferation (4346). However, the cellular effects of estrogen are largely dependent on relative levels of the two ERs displaying complexities associated with differential functions when ER- is expressed alone versus when it is co-expressed with ER-. MCF7 cells used in this study are ER-+/ (4749). Thus, estrogen-mediated MCF7 cell proliferation may involve ER--mediated induction of miR-191 leading to pro-proliferative effects. We further show that miR-191 is involved in a positive feedback loop with ERs in MCF7. Such interplay may have a significant effect on the pathogenesis of hormonedependent breast cancer. Besides estrogen, we show the contribution of hypoxia and serum starvation, common tumor microenvironment factors for miR-191 induction that might contribute to its high levels in other solid tumors. A recent paper demonstrated hypomethylation of miR-191 locus in hepatocellular carcinoma responsible for its high levels; however, its extension to breast cancer needs confirmation (50). Decoding miR-191 cellular functions in breast cancer revealed it to be a onco-miR involved in imparting enhanced cell proliferation, migration, chemoresistance and improved survival in tumor microenvironment. The dissection of the molecular mechanisms underlying multiple functions of miR-191 was done through identification of its direct and indirect targets. The gene signature showed predominance of various signaling pathways like calcium, G Protein CoupledReceptor and olfactory signal transduction indicating important role of miR-191 in cellular signal transduction. Estrogen too is known to mediate effects either in ER-dependent manner or independently by membrane activation leading to activation of intracellular calcium signaling pathway (51). Interestingly, microarray data show that miR191 overexpression modulates up to 30 members of olfactory receptor family. Although the significance of this is too early to comment, a recent report proposes the role of olfactory receptor family in breast cancer progression (52). We confirmed CDK6, BDNF and SATB1 as direct miR-191 targets. Arecent study also demonstrated CDK6 as miR-191 target in thyroid follicular neoplasia (11). Thus, overexpressed miR-191 can be one of the reasons for surprisingly low levels of CDK6 in breast cancer (53). The miRNA overexpression/inhibition showed huge modulation in transcript levels of PLCD1 and BCL11A, but the 3 UTR-luciferase experiments showed minimal effects. Thus, it is probable that miR191 indirectly affects the expression of these genes. BDNF is an estrogen-inducible gene and adds to the list of gene signature shared by miR-191 and estrogen (54). BDNF is also recently shown to be associated with poor prognosis in breast cancer by promoting oncogenic properties (55). We focused our study on SATB1, a key protein involved in breast cancer pathogenesis. Through our results we show that targeted downregulation of SATB1 by miR-191 in MCF7 (ER+) cells is required for cells to enhance cell proliferation and migration. In agreement to our results, high SATB1 levels have been previously reported to be associated with positive prognosis in ER+ breast tumors (56,57). Other miRNAs, miR-7 and miR-155 also have been recently reported to target SATB1 but in ER breast cancer (58). SATB1 functions as a chromatin remodeler and thus impacts global gene expression. Presumably, estrogen-mediated miR-191 induction may bring about chromatin reorganization by targeted downregulation of SATB1 and thus regulates specific gene expression. Accordingly, we observed that on miR-191 overexpression, up to 13 targets of SATB1 being affected, mainly involved in cell proliferation (PLAC1, SOCS2) and apoptosis (CASP4). We also observed lot of commonality in estrogen and miR-191-regulated genes. Overall, through our work in this study, we highlight miR-191 as an important mediator of estrogen signaling in promoting the aggressive phenotype of ER-responsive breast cancer. The study needs to be extended to in vivo mouse models to score therapeutic potential of miR-191 in ER receptor status-dependent manner. Overall, we have identified miR-191 as an estrogen-regulated oncomiR in breast cancer. Widespread impact of miR-191 on cellular functions and pro-proliferative effects in several cancers (Nagpal et al., unpublished data) calls attention to fully dissect its functions and targets both in vitro and in vivo in rest 12 cancers, where its altered levels have been reported before. Based on our studies, anti-miR-191 therapy may prove useful for treatment of hormone receptorpositive breast cancer. Supplementary material Supplementary Data 18 can be found at http://carcin.oxfordjournals. org/ Funding Department of Science and Technology, Government of India (SR/ FT/LS-059/2009 to R.K.). Acknowledgements N.N.thanks Centre for Scientific and Industrial Research for Senior Research Fellowship. H.A. thanks Department of Biotechnology for Senior Research Fellowship. We would like to acknowledge late Megha Bhasme for her valuable contributions in this study. Author contributions R.K.conceived the project and supervised the whole study. N.N.performed all the experiments except generation of miR-191-overexpressing clones and microarray experiments were performed by H.A. B.M.did microarray analysis. R.K.and N.N.wrote the manuscript. Conflict of Interest Statement: None declared.


This is a preview of a remote PDF: https://carcin.oxfordjournals.org/content/34/8/1889.full.pdf

Neha Nagpal, Hafiz M. Ahmad, Bhuvan Molparia, Ritu Kulshreshtha. MicroRNA-191, an estrogen-responsive microRNA, functions as an oncogenic regulator in human breast cancer, Carcinogenesis, 2013, 1889-1899, DOI: 10.1093/carcin/bgt107