© The Author 2013. Published by Oxford University Press. All rights reserved. For Permissions, please email: journals.permissions@oup.com

Carcinogenesis vol.00 no.00 p.1 of 11, 2013doi:10.1093/carcin/bgt067Advance Access publication February 21, 2013

Estrogen receptor β expression induces changes in the microRNA pool in human colon cancer cells

Karin Edvardsson1,2,†, Trang Nguyen-Vu1,†, Sharanya M.Kalasekar1, Fredrik Pontén3, Jan-Åke Gustafsson1,2 and Cecilia Williams1,* 1Department of Biology and Biochemistry, Center for Nuclear Receptors and Cell Signaling, University of Houston, Houston, TX 77204, USA, 2Department of Biosciences and Nutrition, Novum, Karolinska Institutet, Stockholm 141 83, Sweden and 3Department of Pathology and Genetics, Rudbeck Laboratory, Uppsala University, Uppsala 751 85, Sweden

*To whom correspondence should be addressed. Tel: +1(832)842 8807; Fax: +1(713)743 0634; Email: CeciliaWilliams@uh.edu

There is epidemiological, animal and in vitro evidence that estro-gen receptor β (ERβ) can mediate protective effects against colon cancer, but the mechanism is not completely understood. Previous research has indicated critical pathways whereby ERβ acts in an antitumorigenic fashion. In this study, we investigate ERβ’s impact on the microRNA (miRNA) pool in colon cancer cells using large-scale genomic approaches, bioinformatics and focused functional studies. We detect and confirm 27 miRNAs to be significantly changed following ERβ expression in SW480 colon cancer cells. Among these, the oncogenic miR-17–92 clus-ter and miR-200a/b are strongly downregulated. Using target prediction and anticorrelation to gene expression data followed by focused mechanistic studies, we demonstrate that repression of miR-17 is a secondary event following ERβ’s downregulatory effect on MYC. We show that re-introduction of miR-17 can reverse the antiproliferative effects of ERβ. The repression of miR-17 also influences cell death upon DNA damage and medi-ates regulation of NCOA3 (SRC-3) and CLU in colon cancer cells. We further determine that the downregulation of miR-200a/b mediates increased ZEB1 while decreasing E-cadherin levels in ERβ-expressing colon cancer cells. Changes in these genes correspond to significant alterations in morphology and migration. Our work contributes novel data of ERβ and miRNA in the colon. Elucidating the mechanism of ERβ and biomark-ers of its activity has significant potential to impact colon cancer prevention and treatment.

Introduction

Colorectal cancer is the third most common form of cancer among both men and women and corresponds to almost 10% of all reported cancer cases in the world (1,2). Women have a later onset than men (2,3) and younger women (18–44  years old) have a better over-all survival compared with men of the same age. In older women (>50 years old) this, however, is reversed (4,5). Hormone replace-ment therapy and oral contraceptives have been shown to reduce the incidence of colon cancer in women (6–11). Taken together, a role for the hormone estrogen is indicated in the disease develop-ment. Estrogen treatment also reduces the growth of colonocytes in non-malignant mouse cell line and reduces aberrant crypt foci in mice (12).

Estrogen mediates its effects through two nuclear receptors: estrogen receptor α (ERα) and β (ERβ). These receptors bind estro-gen with equal affinity and act as transcription factors by inducing

changes in target gene expression. ERα is highly important in breast cancer where it is upregulated and increases cell proliferation. ERα is widely utilized as a therapeutic target and its activity blocked by the drug (e.g. tamoxifen). The transcriptional effect of ERα is well studied in breast cancer cells (13,14). In colon cancer, however, ERα is not widely expressed. The predominant ER in normal colonic epi-thelium is ERβ, as determined both at the protein (15) and messenger RNA (mRNA) levels (16). Immunohistochemistry has further shown decreased expression of ERβ during colon cancer progression (17). Polymorphism in the ERβ gene affects colon cancer risk and survival (18–20). In vivo animal studies have demonstrated that ERβ mediates the colon cancer protective effect of estrogen (12,21–24). Thus, there is a significant interest to determine the molecular mechanism of this receptor activity in colon cancer. ERβ has been shown to be antipro-liferative in both colon cancer cells (25–28) and in xenografts (26). Recently, genome-wide analysis of protein-coding genes allowed us to identify multiple ERβ-mediated antitumorigenic effects in human colon cancer cell lines and initiate work to dissect its mechanism (29). Exposing the wide and cell-specific impact ERβ has on gene expres-sion (29), we hypothesize that ERβ in part mediates these multiple effects through regulations of microRNA (miRNA).

miRNAs are short (~22 nt), single-stranded non-coding RNA regulators. They regulate gene expression through imperfect complementary binding between the miRNA 5ʹ-sequence and the 3ʹ-untranslated region of the target mRNA. Such binding medi-ates degradation of the mRNA transcript and/or suppression of its translation. miRNA genes are transcribed by RNA polymerase II into primary miRNAs. The nuclear protein DGCR8 and the enzyme Drosha introduce a cleavage to liberate a hairpin loop from the primary miRNAs. The resulting hairpin, precursor miRNA (pre-miRNA), is exported from the nucleus to the cytoplasm. The pro-tein Dicer then cleaves the loop, leaving a short, double-stranded duplex. One strand of the duplex is released and often degraded, whereas the other strand, the mature miRNA, is incorporated into the Argonaute containing miRNA-induced silencing complex, which facilitates the interaction between the mature miRNA and its mRNA target. A miRNA can target several different genes and many of these targets are often in the same pathway (30). As one mRNA can also be targeted by several miRNAs, the total impact of miRNAs on the transcriptome and proteome can be significant. There are about 1500 miRNA genes in the human genome (mir-base.org, January 2012) and together they may target and affect the expression of up to 60% of all protein-coding genes (31). miRNAs have been linked to tumorigenesis due to their frequent localiza-tion at fragile sites and common breakpoint regions (32) and both oncomiRs and tumor suppressor miRNAs have been identified (33). Altered levels of miRNAs are common in tumors, including those of the colon (34–36).

An emerging topic in cell control and other biological processes is the regulatory connection between miRNAs and nuclear receptors (37–42). Estrogen has been implicated to have an effect on miRNAs in breast cancer cells (43,44) where ERα is the predominant ER. Studies have identified ERα-regulated miRNAs (45–50), although it is not clear whether the regulation is directly transcriptional or indirect (51,52). Although recently, ERβ was indicated to affect miRNAs in breast cancer (53), no studies so far have described miRNAs regulated by ERβ in the colon. In this study, we examined changes in the miRNA pool in cells where we previously have characterized the function of ERβ (26,29). In order to estimate the impact of the changed miRNA expression, we correlated changes in the miRNome with the generated transcriptome data and with the ER chromatin-binding data. Finally, we have performed focused functional studies of selected miRNAs repressed by ERβ.

Abbreviations: ER, estrogen receptor; mRNA, messenger RNA; PBS, phos-phate-buffered saline; pre-miRNA, precursor miRNA.

†These authors contributed equally to this work.

Page 1 of 11

Carcinogenesis Advance Access published March 14, 2013 by guest on M

arch 20, 2013http://carcin.oxfordjournals.org/

Dow

nloaded from

K.Edvardsson et al.

Materials and methods

ERβ immunohistochemistryParaffin-embedded tissue arrays were first de-paraffinated in xylene and rehy-drated. Antigen retrieval was achieved by immersing the slides in 10 mM citrate buffer (pH 6.0) for 15 min at 97°C. The tissues were then immersed in cold phosphate-buffered saline (PBS) and treated with 0.5% Triton-X in PBS for 15 min. Endogenous peroxidase was blocked with 0.3% H2O2 in methanol for 30 min. The tissues were incubated for 10 min in 3% bovine serum albumin to block unspecific binding and treated over night with a 1:50 dilution of in-house anti-ERβ antibody ERβ503 (54) in a humid chamber. Secondary biotinylated rabbit antichicken/turkey antibody (Invitrogen, New York, NY) was applied to the slide with the tissues and incubated for 1 h at room temperature. The slide was then treated with the Vectastain ABC solution (Vector Laboratories, Burlingame, CA), and the color was developed with 3,3ʹ-diaminobenzidine (Dako, Carpinteria, CA). Antibody washes were done with a 0.1% Nonidet P 40 substituent solution (Sigma, St Louis, MO) in PBS. The stained tissues were counterstained with hematoxylin, dehydrated, treated with xylene and mounted with Pertex mounting medium. The Olympus BX41 microscope and the Microsuite Basic Edition software from Olympus were used for visualiza-tion and imaging of the tissues.

Cell cultureCell lines (SW480, HT29 and SW620) were obtained from American Type Culture Collection (Manassas, VA). Full-length ERβ was stably expressed via lentivirus transduction in three mixed cell populations of SW480 and HT29, with empty vectors as control, as described previously (26). For SW620 cells, transient expression of ERβ was performed using PC3 construct with or without full-length ERβ as described previously (29). The cells were cultured in RPMI-1640 (Invitrogen, Carlsbad, CA) with 5% fetal bovine serum and 1% penicil-lin-streptomycin (PEST) (Invitrogen) at 37°C with 5% CO2. To investigate ligand-dependent regulations, the medium was first changed to phenol-red free RPMI-1640 with 5% dextran-coated-charcoal-treated fetal bovine serum and 1% PEST 24 h before the experiments. Treatment was done for 24 h in stripped medium with 10 nM E2, 1 μM tamoxifen, 10 nM ICI 182780 or vehicle (EtOH).

Receptor quantificationReceptor expression was quantified using radiolabeled ligand-binding assay. Cells were plated at 500 000 cells per well in six-well plates, incubated at 37°C for 24 h and washed with PBS three times. Cells were then incubated with 0.5 nM [3H]estradiol in the absence or presence of 1 μM unlabeled estradiol (to measure total and unspecific binding, respectively) in 10% DCC-fetal bovine serum for 3 h at 37°C. Cells were then washed with PBS three times, lysed with lysis buffer and lysates were mixed with LSC-cocktail (Emulsifier-Safe; Packard BioScience, Waltham, MA). Radioactivity was measured using liquid scintillation counter (LS-6000-SC; Beckman-Coulter, Brea, CA). Experiments were performed in six replicates, and measurements of total protein and total cell number were performed in parallel.

RNA extractionTotal RNA was extracted with TRIzol (Invitrogen) and purified with miRNe-asy spin columns (Qiagen, Chatsworth, CA), according to standard protocol. On-column DNase I digestion was used to remove remaining genomic DNA. Quantitative and qualitative RNA analyses were performed using NanoDrop 1000 spectrophotometer and the Agilent 2100 BioAnalyzer (Agilent technolo-gies, Palo Alto, CA), respectively. Samples with RNA integrity >9.5 were used.

MiRNA expression analysisLarge-scale profiling of the miRNA pool in E2-treated ERβ-expressing cells and E2-treated control cells was made using TaqMan Low-Density Array miRNA cards, card A  v2.0 (Applied Biosystems, Foster City, CA) without pre-amplification step. Total RNA, 1000 ng each, from two different mixed populations of ERβ-expressing cells and two different mixed populations of control cells was analyzed. Analysis was performed on 7900HT Fast Real-Time PCR System (Applied Biosystems). MiRNAs were considered differen-tially expressed when the following criteria were fulfilled: Ct ≤ 32, 0.7≥FC≥1.3 and no significant difference between biological duplicates.

Complementary DNA synthesis and quantitative real-time–PCRFor miRNA analysis, poly(A) tails were added to 1 µg of total RNA with NCode miRNA First-Strand cDNA Synthesis Kit (Invitrogen) followed by complementary DNA synthesis using corresponding universal primer, according to manufacturer’s protocol. For the real-time PCR reaction, a corresponding universal real-time reverse PCR primer from the kit and a forward primer based on the mature miRNA sequence (www.mirbase.org) were used. U6 was used as reference gene. For mRNA analysis, complementary DNA was synthesized from 1 µg of total RNA using Superscript III (Invitrogen) and random hexamer primers, as described previously (29). 18S and

ARHGDIA were used as reference genes. Real-time PCR was performed in 7500 Fast Real-Time PCR System (Applied Biosystems) with Fast SYBR-Green Master mix (Applied Biosystems) according to conditions specified by the manufacturer. All primer pairs were checked with melting curve analysis. Primer sequences for genes will be provided upon request. Relative expression was calculated as fold change in relation to control using the ΔΔCt-method. Unpaired two-tailed t-test was used to test significance between two parallel treatment groups of identical origin and considered significant if P < 0.05.

Bioinformatics, network and pathway analysisRegulatory networks and biological themes affected by differentially expressed miRNAs were analyzed with Ariadne Pathway Studio v8. Predicted and con-firmed target genes were correlated to previous gene expression studies (29). Chromosomal locations of pre-miRNAs of interest were compared with ERα-, ERβ- and MYC-binding studies (55–57).

Transfection with miRNA mimics and inhibitorsTransfection with Dharmacon miRIDIAN miRNA mimics, hairpin inhibitors and mimic or inhibitor controls were performed using DharmaFECT reagents (Thermo Scientific, Waltham, MA), essentially as described previously (58). In brief, the cells were seeded in six-well plates in stripped serum for 24 h, then trans-fected in antibiotic-free medium with DharmaFECT I (0.1% of total volume) and a final concentration of 25 nM mimic, inhibitor or respective control. Medium was changed to complete medium 24 h after transfection, and the cells were cultured for an additional 24 h before harvest. All experiments were performed in triplicates.

Western blot analysisCells were lysed with RIPA lysis buffer. Protein concentrations were deter-mined using Qubit® Protein Assay Kit (Invitrogen) along with the Qubit® 2.0 Fluorometer (Invitrogen). Approximately, 50 or 100  μg of total protein were loaded onto a 10% polyacrylamide gel. Proteins separated by sodium dodecyl sulfate–polyacrylamide gel electrophoresis were transferred to a polyvinylidene difluoride membrane (Millipore, Billerican, MA). The membrane was blocked with 5% non-fat milk in Tris-buffered saline Tween-20 and then incubated with antibodies against NCOA3, E-cadherin (Santa Cruz Biotechnology, Santa Cruz, CA) or β-actin (catalog number: A2228; Sigma) in Tris-buffered saline Tween-20 overnight. The membrane was then reprobed with appropriate secondary antibod-ies conjugated with horseradish peroxidase for 2 h. Blots were processed using ECL Kit (Thermo Fisher Scientific, Rockford, IL) and exposed to film.

Proliferation studyCells were plated in six-well plates for 24 h, transfected with miRNA mimic or control and cultured for additional 96 h. Following trypan blue staining, the cells were counted with the Countess automated cell counter (Invitrogen). All experiments were performed in triplicates.

Migration assayCell migration was measured using in vitro scratch assays. Cells were seeded in 12-well plates and were allowed to attach for 24 h after which an artificial gap was created in the confluent cell monolayer with a pipette tip. If treated with miRNA mimics, the cells were then transfected 24 h before the gap was introduced. The medium was replaced with fresh medium after the cells were washed twice with pre-warmed PBS to remove the debris and to smooth the edges of the scratch. Initial images of the scratch (0 h) were taken with Olympus 1X51 inverted micro-scope, after which the cells were incubated in a 5% CO2 cell incubator at 37°C. Plates were taken out of the incubator intermittently to measure cell migration; images of the scratch were taken at 4 and 8 h following the initial scratch and image acquisition. Distance of migration and area covered by migrating cells were analyzed quantitatively from the acquired images by using Olympus cellS-ens® digital imaging software and ImageJ software, respectively.

Apoptosis assaySW480 control cells were transfected with miR-17 inhibitor and control for 48 h, then treated with vehicle (dimethyl sulfoxide) or the DNA-damaging agent cisplatin (cis-diammineplatinum(II) dichloride) (Sigma–Aldrich, St Louis, MO) with a final concentration of 10 μg/ml for 24 h. Cells were stained with trypan blue and counted as described previously. Statistical analysis was made with Chi-square test between amount of live and dead control cells with endogenous miR-17 expression, and cells were transfected with miR-17 inhibitor, treated with either dimethyl sulfoxide or cisplatin.

Results

Expression of ERβ in normal colonic epitheliumAs previous studies investigating ERβ expression in colon have not thoroughly differentiated between ERβ wild-type and splice variants, we investigated the expression pattern using our in-house

Page 2 of 11

by guest on March 20, 2013

http://carcin.oxfordjournals.org/D

ownloaded from

ERβ affects miRNAs in colon cancer

wild-type-specific ERβ503 antibody (54,59). This polyclonal anti-ERβ was used on tissue microarrays containing 46 1 mm2 cores of different types of normal tissue and 216 cores representing the 20 most common forms of human cancer. We focused on the 15 sections from normal intestinal tract and 27 sections of colon carcinoma from different individuals. A widespread expression of ERβ was observed in both males and females throughout the normal intestinal tract, exemplified in Figure 1A. Nuclear ERβ expression was present in the epithelial cells of the duodenum, small and large intestine and the rec-tum (Figure 1A), whereas other tissue samples (e.g. liver) on the same array showed no nuclear ERβ staining (Supplementary Figure 1, avail-able at Carcinogenesis Online). Although ERβ was still expressed in the colorectal tumors, the extent of nuclear staining was noticeably lower and weaker than in normal intestine. The cytoplasmic staining, on the other hand, did not seem to decrease in the tumors. Consistent with previous reports (17), the lymphocytes in the normal colon were positive for ERβ, whereas negative for the colorectal cancer tissues. The Human Protein Atlas (http://www.proteinatlas.org) offers a large database with protein profiling of human normal and cancer tissues as well as cell lines (60). Here, three different and characterized anti-bodies against ERα revealed no expression of ERα in tissues from the intestinal tract or in colon cancer (Figure  1A). These findings, together with previous reports, solidly demonstrate that ERβ is the predominant ER in colonic epithelia. This supports the notion that the protective effects of estrogen in colon carcinogenesis are mediated by ERβ in colon epithelial cells and emphasizes the need to elucidate their mechanism(s). As miRNAs affect cancer development and, also, are promising as biomarkers and as therapeutic agents, we investi-gated whether ERβ affects miRNAs in colorectal cancer cells.

Exogenously expressed ERβ are within physiological rangesTo investigate the mechanism of ERβ action, one needs cell lines expressing this receptor. As ERβ is not expressed at sufficient lev-els in colorectal cancer cell lines (29,61) it is necessary to express it exogenously. We have previously established and characterized cell lines with stable ERβ expression (26,29). We and others have noted that ERβ when re-introduced in colon cancer cell lines is ligand independent (25,26,29). To determine whether the level of expression in our cell lines is physiological, we performed radiolabeled ligand-binding assays. Cell lines with or without ERβ were incubated with 1 nM [3H]estradiol in the absence or presence of 1000 nM unlabeled estradiol and total and unspecific binding were recorded, as shown in Figure 1B. The resulting expression levels were calculated and cor-respond to 3 770 ± 940 receptors/cell (6.3 fmol/106 cells) for stably transfected SW480-ERβ, 2 470 ± 597 receptors/cell (4.1 fmol/106 cells) for stably transfected HT29-ERβ cells and 3 325 ± 307 recep-tors/cell (5.5 fmol/106 cells) for transiently transfected SW620, which falls within physiological ranges (62,63). None of the parental or control cell lines showed significant levels of binding, supporting the lack of endogenous ER expression. Thus, the exogenously introduced ERβ is expressed at levels that are physiologically relevant and binds ligand.

Introduction of ERβ in SW480 colon cancer cells changes the miRNA poolTo investigate whether ERβ has an effect on miRNA expression in colon cancer cells, we compared estrogen-treated SW480 control cells to estrogen-treated ERβ-expressing SW480 cells using large-scale miRNA techniques. The results indicate that ERβ expression induced a significant change in the mature miRNA pool; a total of 51 miRNAs were scored as downregulated and 43 as upregulated (Supplementary Table  1, available at Carcinogenesis Online). Confirmations using real-time PCR were performed for a total of 60 miRNAs, out of which 27 confirmed the large-scale analysis results. The regulations of the remaining 33 miRNAs are inconclusive due to issues with primer specificity, detection levels or inconsistent regulations between the three mixed cell populations. Table I summarizes the confirmed miRNAs and Figures 2 and 3A show real-time PCR results. Of the

confirmed miRNAs, 24 were downregulated and miR-205, miR-10a/b were the only three miRNAs confirmed as upregulated. We found the entire miR-17–92 cluster, also known as oncomir-1, to be downregu-lated in ERβ-expressing cells (Figure 3A). This study shows for the first time that introduction of ERβ induces a change of the miRNA pool in colon cancer cells.

The changed miRNA pool contributes to the ERβ-induced transcriptomeMiRNAs can induce degradation of their target mRNAs and we have observed large changes in the mRNA expression as a result of ERβ in these cells (29). To understand the impact of the changed miRNA pool on the SW480 transcriptome, we identified the predicted and/or confirmed target genes of the regulated miRNAs. An anticorre-lation between the ERβ-regulated miRNome and the transcriptome of protein-coding genes, using bioinformatics, indicates that the impact of the changed miRNA pool could be substantial (illustrated in Supplementary Figure  2, available at Carcinogenesis Online). We note several clusters that appear particularly implicated, such as the one of miR-9, which is predicted to target and regulate mul-tiple genes that were regulated by ERβ expression. Other miRNA targets that are implicated in colon carcinogenesis include ANAX1 that was upregulated upon ERβ expression. ANAX1 is a predicted target of miR-196a-1, which is correspondingly downregulated upon ERβ expression. ANAX1 inhibits nuclear factor-κB and inflamma-tory response in the colon and mediates antitumorigenic effects (64). We have previously reported that ERβ mediates anti-inflammatory effects, including the repression of IL-6, and hypothesized that nuclear factor-κB contributes to its tumor-protective properties (29). This was recently confirmed in vivo where knock-out of ERβ in mice increased nuclear factor-κB and IL-6 expression and corresponding inflammatory response in the colon, and significantly enhanced intes-tinal tumor development (24). The impact of miR-196a-1 in this con-text requires further investigation. Further, many tumor suppressors that were upregulated by ERβ, may be targeted by the miRNAs, found to be downregulated in this study. DLC1, which can inhibit prolifera-tion in colon cancer cell lines (65), and cyclin D1 are both predicted targets of miR-19a and tumor suppressor LATS2 (66,67) by miR-25. All these genes are not necessarily regulated only as a consequence of the ERβ-affected miRNAs; however, exploring these networks is likely to contribute to elucidating the mechanism of transcriptional regulation by ERβ. These results may thus serve as a guide for future studies to investigate ERβ transcriptional regulation.

ERβ regulates miRNAs with known ERβ DNA-binding sites nearbyA relevant question is whether the observed change in the miRNA pool is a direct or indirect result of ERβ transcriptional activity. None of the proteins and enzymes in the miRNA processing apparatus, such as DGCR8, Drosha, Dicer or Argonaut, were significantly changed at their transcriptional levels following ERβ expression (Supplementary Table  2, available at Carcinogenesis Online). This suggested that changes in the miRNA pool were a consequence of transcriptional regulation. We investigated whether the regulated miRNAs had an ER chromatin-binding sites close to their chromosomal locations. The high homology of the DNA-binding domain between ERα and ERβ allows these receptors to frequently bind to the same estrogen responsive elements (EREs). Published studies have identified endogenous ERα binding (55,56) and/or exogenous ERβ binding (55,57) to the genome in the human breast cancer cell line MCF-7. No such studies have been performed in colon cancer cells. We may, however, assume that ERβ is capable of binding to the identified chromatin positions also in colon cancer cells. By correlating the chromosomal location of regulated miRNAs with published ERα- and ERβ-binding sites, we could identify miRNAs with one or several proximal and distal ERβ- or ERα-binding sites, as summarized in Table II. The 27 regulated mature miRNA sequences can be coded for by 35 different pre-miRNAs. Out of these 35 pre-miRNAs, 23 had an ERβ-binding site within 100 kb.

Page 3 of 11

by guest on March 20, 2013

http://carcin.oxfordjournals.org/D

ownloaded from

K.Edvardsson et al.

Fig. 1. Characterization of ERβ expression in tissues and cell lines. (A) Endogenous expression in human normal and cancer tissue samples from the intestinal tract samples as measured by immunohistochemical staining. Top panel shows ERβ expression (brown) in duodenum, jejunum-ileum (small intestine), colon

Page 4 of 11

by guest on March 20, 2013

http://carcin.oxfordjournals.org/D

ownloaded from

ERβ affects miRNAs in colon cancer

Table I. ERβ expression in SW480 cells resulted in 27 differentially expressed miRNAs 

Downregulated (ascending order) Upregulated (descending order)

miR-142-3p miR-205miR-19a miR-10bmiR-200a miR-10amiR-19bmiR-18amiR-18bmiR-20bmiR-20amiR-135amiR-200bmiR-192miR-194miR-590-5pmiR-17miR-92amiR-9miR-301amiR-106amiR-196amiR-183miR-30bmiR-140-5pmiR-221miR-9-3miR-25

miRNAs regulated as a consequence of ERβ introduction in SW480 cells and compared with SW480 control cells, both treated with E2. Differentially expressed miRNAs were detected with TaqMan Low-Density Arrays and confirmed with real-time PCR in three separate mixed-clone populations of control and ERβ-transduced cells.

Fig. 2. Differential expression of miRNAs in SW480 cells upon ERβ expression. Fold-change indicates the relative change in miRNA expression in ERβ-expressing SW480 cells compared with control SW480 cells, in presence of E2. Real-time PCR analysis performed in three mixed cell populations confirmed the large-scale screening results. Representative data from one cell population are shown in the figure. (A) miRNAs upregulated in ERβ-expressing cells; (B) miRNAs downregulated in ERβ-expressing cells. Unpaired two-tailed t-test was used to compare differences between the two parallel groups. Significance is presented as *P ≤ 0.05, **P ≤ 0.01, ***P ≤ 0.001.

(large intestine) and rectum. Immunoreactivity for the ERβ antibody is found in a majority of glandular and inflammatory cells of the mucosa. The expression is mainly nuclear with only weak general cytoplasmic positivity. Bottom left panel shows three different examples of ERβ expression in colorectal carcinoma. The nuclear ERβ immunoreactivity is reduced in cancer cells with only a smaller fraction of tumor cells being positive. The cytoplasmic ERβ staining does not appear to be decreased in the tumor cells compared with normal glandular cells. Bottom right panel shows negative staining for ERα antibody (Novacastra #NCL-L-ER-6F11) in the intestinal tract as well as in colon cancer. (B) Endogenous and exogenous ligand-binding activity in human colon cancer cell lines. ER concentrations are calculated in cell lines SW480, HT29 and SW620 before and after exogenous expression of ERβ, using the radiolabeled ligand-binding assay described in Materials and methods. Cell lines show no significant endogenous estrogen-binding capacity, whereas exogenous expression of ERβ results in ligand-binding activity within physiological ranges.

←

Thus, up to 66% of pre-miRNAs are potentially transcriptionally regulated by ERβ. We note that the most upregulated miRNA, miR-10a, has the closest reported ERβ-binding site 23 kb upstream of the miRNA gene (55). However, as previously found for ERβ’s antiproliferative effect and its regulation of protein-coding genes by exogenously expressed ERβ in colon cancer cells, the regulations of miRNAs were not dependent on ligands (E2 [10 nM], tamoxifen [1  μM] or ICI [10 nM]; Supplementary Figure  3A, available at Carcinogenesis Online). Future studies are needed to explore the details of ERβ transcriptional regulation, including defining its chromatin-binding sites in colon cells.

ERβ affects miRNAs through repression of MYCSeveral of the regulated miRNAs do not have a known ERβ-binding site in their proximity. Although the ERs can regulate genes dis-tant from their chromatin-binding site (68), we assumed that miR-NAs were also regulated as a result of secondary events. We have previously observed that the proto-oncogene MYC is strongly downregulated following ERβ expression in these and other cells (13,26,29), and ERβ binds the MYC promoter within 70 kb indicat-ing that MYC is directly regulated by ERβ (55,57). MYC is a key regulator of miRNAs (69–71), which contributes to its oncogenic effect (72). We note that ERβ mediates a 50% decrease in MYC levels in SW480 cells (Figure  3B, left panel). We correlated our regulated miRNAs with MYC chromatin-binding sites (56) and previously published data on MYC and miRNA regulation in other cell lines (69,71,73,74), summarized in Table III. Several MYC-regulated miRNAs (miR-200a, miR-200b, miR-9-3, miR-106a and miR-135a-1) also have ERβ-binding sites in their vicinity and may be regulated by both ERβ and MYC, whereas others, including the miR-17–92 cluster, have only known MYC-binding sites. The over-expression of the latter cluster has, in colon cancer, been associated

Page 5 of 11

by guest on March 20, 2013

http://carcin.oxfordjournals.org/D

ownloaded from

K.Edvardsson et al.

with MYC gene amplification and overexpression (75) and associ-ated with poor survival (76). To explore the influence of MYC on the miRNAs detected in this study in colon, we used small interfer-ing RNA to silence MYC. The silencing of MYC was efficient and reduced its mRNA by 50% within 48 h of small interfering RNA transfection (Figure 3B, right panel). This resulted in a downregu-lation of miR-17, miR-19a and miR-106a reaching a maximum at 72 h after transfection (Figure 3B, right panel and Supplementary

Figure  3B, available at Carcinogenesis Online). MiR-200a/b, on the other hand, was upregulated upon MYC silencing, whereas miR-221 was unaffected (Supplementary Figure  3B, available at Carcinogenesis Online). We conclude that MYC does not contrib-ute to the ERβ-mediated repression of miR-200a/b and miR-221, but that the miR-17–92 cluster and miR-106 are transcriptionally downregulated as a consequence of the ERβ-mediated downregula-tion of MYC.

Fig. 3. ERβ regulates miR-17–92 cascade effects through MYC. (A) The oncogenic miR-17–92 cluster was confirmed to be downregulated in ERβ-expressing SW480 cells (+E2) compared with control SW480 cells (+E2), using real-time PCR. (B) Left panel, introduction of ERβ in SW480 cells (+E2) leads to decreased mRNA levels of MYC; Right panel, small interfering RNA transfection decreased MYC levels by half within 48 h, resulting in reduced levels of miR-17 and 19a within 72 h. (C) Introduction of miR-17 mimic into SW480-ERβ cells reversed the ERβ-antiproliferative effects. (D) Inhibition of miR-17 in SW480 cells (miR-17 inhibitor, no E2) increased cisplatin-induced cell death compared with control cells (inhibitor control, no E2): 41% of the control cells died after cisplatin treatment compared with 56% in miR-17 inhibited cells. Y-axis: Dead cells (%) due to cisplatin treatment compared with vehicle-treated cells. (E) NCOA3 was upregulated upon ERβ expression at both mRNA (left) and protein (right) levels. (F) Regulatory network, based on published literature and our transcriptome and miRNome data, proposing a mechanism for this ERβ-induced cascade. Red fill indicates a gene upregulated by ERβ (our data). Blue fill, a gene downregulated by ERβ (our data). Light blue lines illustrate published effects on gene expression (induction or repression), green lines illustrate known binding of one protein to the chromatin regulatory regions of the other gene (proposing a direct regulation) and dark purple lines symbolize known protein–protein interaction. (G) Inhibition of miR-17 in SW480 cells (no E2) significantly upregulated CLU and NCOA3 mRNA levels (top) and NCOA protein levels (bottom). (H) Overexpression of miR-17 (mimic, no E2) decreased mRNA levels of CLU and NCOA3. Fold-change indicates the relative changes in miRNA or mRNA expression compared with controls. Unpaired two-tailed t-test was used to test the significance of differences between two parallel treatment groups of identical origin. Significance is presented as *P ≤ 0.05, **P ≤ 0.01, ***P ≤ 0.001.

Page 6 of 11

by guest on March 20, 2013

http://carcin.oxfordjournals.org/D

ownloaded from

ERβ affects miRNAs in colon cancer

MiR-17 re-expression can override the ERβ-mediated reduction of proliferationExpression of ERβ resulted in an antiproliferative effect in the colon cancer cells. Our TaqMan Low-Density Array analysis showed that the entire miR-17–92 cluster was downregulated following ERβ expression (Table I). This cluster consists of the following six mem-bers: miR-17, miR-18a, miR-19a, miR-20a, miR-19b-1 and miR-92a-1, and all, except miR-92a, were significantly downregulated when confirmed using real-time PCR (Figure  3A). The members of the miR-17–92 cluster (also called Oncomir-1) are implicated in colon cancer (76) and have been related to enhanced cell prolifera-tion (77,78). We selected to perform experiments using one of its key

members, miR-17, to test whether the downregulation of this clus-ter had a significant role in the ERβ-mediated effect of proliferation. We re-introduced miR-17, using mimic, into ERβ-expressing SW480 cells and investigated whether this could oppose the ERβ-mediated repression of growth. We show that transfection with miR-17 mimic indeed can reverse the antiproliferative effects of ERβ and increase cell proliferation (Figure 3C).

ERβ-mediated repression of miR-17 can enhance apoptosisThe miR-17–92 cluster has also been reported to suppress apopto-sis (77). As one described function of ERβ is increased apoptosis in normal colonocytes (79) and intestines (80), we hypothesized that ERβ may mediate this apoptotic effect through this miRNA clus-ter. We mimicked the effect of ERβ by using a miR-17 inhibitor in SW480 cells. Although we did not observe a significant effect on apoptosis by inhibiting miR-17 in SW480 cells grown under regular conditions, we noted a clear effect on cell death when the cells were challenged with DNA-damaging cisplatin treatment. While 41% of the control cells died following cisplatin treatment, 56% of the miR-17-silenced cells died (Figure 3D). A Chi-square test between treated control cells (inhibitor control) and cells where miR-17 was silenced (miR-17 inhibitor) gave a P -value < 0.001. Our results show that inhibition of miR-17 caused the cells to become more sensitive to cis-platin-induced treatment. This leads us to conclude that ERβ through decreased miR-17 expression can enhance apoptosis after DNA dam-age in colon cells.

MiR-17 mediates the upregulation of NCOA3 and CLU noted after ERβ expressionAlthough the targets of the miR-17–92 cluster in colon are not well characterized, we used anticorrelation between the ERβ-mediated transcriptome in SW480 cells and bioinformatically predicted tar-gets of miR-17 to infer potential effects. Two predicted targets were

Table II. Differentially expressed miRNAs that harbor an ERα or ERβ chromatin-binding site before or after the miRNA locus 

MiRNA ERβ (57) ERβ (55) ERα (55) ERα (56)

US DS US DS US DS US DS

miR-142-3p 186 246 308miR-200a 87 70 88 265 215miR-19b-2 85miR-18b 85miR-20b 85miR-135a-1 177 56 139 56miR-135a-2 4 4.9 4.1 4miR-200b 87 70 87 266 215miR-192 14 42 14miR-194-1 220miR-194-2 205 14 233 14miR-590-5p 57 62 18miR-9-1 196 36 137 52 207 196miR-9-3 228 252 17 17miR-301a 216 181 44 69 146 216miR-106a 84miR-196a-1 132 30 132 96miR-196a-2 146 53 124 166miR-183 174 5.4 139 5.4 174miR-30b 66 65miR-140 110 181 203 281miR-221 78miR-9-3 176 71 89 176 90miR-25 176 71 89 176 90miR-92a-2 85miR-10a 80 23 76 79 43miR-10b 52miR-205 104

US, upstream; DS, downstream.miRNAs with proximal and distal ERα- and ERβ-binding sites according to four different ER-binding studies (55–57). Numbers are given as distance between closest binding site and miRNA locus (kb). miRNAs in bold have a binding site within 30 kb.

Table III. Known MYC-regulated miRNAs that were affected by ERβ expression 

MicroRNA Reference

miR-200a Known MYC-binding site 59 kb US of miRNA locus (56)

miR-200b Known MYC-binding site 58 kb US of miRNA locus (56)

miR-9-3 Expression of miR-9 is activated by MYC, which directly binds to the miR-9-3 locus (73,88)

miR-106a miR-106a was upregulated in the high c-MYC state in a human B-cell line (69)

miR-17–92 cluster MYC activates expression of the miR-17–92 cluster and binds directly to the miR-17–92 locus (69)

miR-221 Human mammary epithelial cells expressing MYC have increased expression of miR-221 (71)

miR-135a-1 miR-135a is upregulated by c-MYC in embryonic stem (ES) cells (74)

US, upstream; DS, downstream.

Page 7 of 11

by guest on March 20, 2013

http://carcin.oxfordjournals.org/D

ownloaded from

K.Edvardsson et al.

implicated in colorectal cancer, the nuclear receptor coactivator NCOA3 (SRC-3) and CLU. NCOA3 may be a direct target of miR-17, based on their complementing nucleotide sequences with one conserved 8mer site at position 714–720 of NCOA3 3ʹ-untranslated region and one conserved 7mer-m8 site at position 1296–1302. In human colon cancer HCT116 cells, miR-17 has been reported to repress CLU levels, which contributed to stimulated angiogenesis and cell growth (81). Both NCOA3 and CLU were upregulated in ERβ-expressing cells compared with controls (29). We confirmed that NCOA3 was upregulated both at the mRNA and protein level (Figure 3E). This led to the hypothesis that ERβ through MYC and miR-17 regulates a cascade of CLU and NCOA3 activity (Figure 3F). As NCOA3 regulates the transcriptional activity of ERβ (82,83), this may also constitute a feed-forward loop. To investigate whether the regulations were a direct consequence of the ERβ-mediated down-regulation of the miR-17–92 cluster, SW480 cells were transfected with a miR-17 inhibitor to emulate the downregulation of miR-17. Indeed, inhibition of miR-17 led to the upregulation of CLU and NCOA3 (Figure 3G) and conversely, when we overexpressed miR-17 in SW480-ERβ cells, using a miR-17 mimic, CLU and NCOA3 levels were correspondingly decreased (Figure 3H). This confirms that their upregulation in ERβ-expressing SW480 cells is a consequence of the repressed miR-17 levels. Thus, we have dissected a pathway where ERβ regulates its own coactivator through MYC and miR-17.

ERβ modulates miR-200a/b and epithelial–mesenchymal transition markersWe noted that when ERβ is expressed, the morphology of the cells is significantly changed (Figure  4A). The cells adopt a more stretched out and adhesive phenotype that also exhibits reduced migration (Figure 4B). This coincides with a strong downregulation of miR-200a and miR-200b in ERβ-expressing cells (Figure 2B). Studies in breast

cancer cells have established that the miR-200-family repress epithe-lial–mesenchymal transition through the downregulation of repressors of E-cadherin expression, ZEB1 and ZEB2 (84). The downregulation of miR-200a/b in SW480 cells coincided with an upregulation of ZEB1 and a downregulation of E-cadherin (Figure 4C), whereas ZEB2 was not expressed at detectable levels. As this pathway has not been stud-ied in colon cancer cells, we investigated whether the changed levels of ZEB1 and E-cadherin were a result of the ERβ-mediated repression of miR200a/b. To test this, we re-introduced miR-200a/b into SW480-ERβ cells using mimics. We show that both miRNAs repressed ZEB1 while increasing E-cadherin mRNA and protein levels (Figure 4D). Thus, the lower level of E-cadherin in ERβ-expressing colon cancer cells is a con-sequence of reduced miR-200a/b levels. Because the miR-200 cluster has de facto ERβ-binding sites 87 kb upstream and 70 kb downstream of the primary miRNA chromosomal region (55), it is possible that miR-200a/b is directly repressed by ERβ. This is further supported by our data that miR-200a was downregulated also in a second colon cancer cell line HT29 (characterized in ref. 29) upon ERβ expression (Figure 4E).

Cell-specific effectsTo explore whether ERβ might have this effect also in advanced colon cancer, we selected the cell line SW620 for further experiments. This cell line originates from the same patient as SW480, but was isolated from a metastatic lymph node 1 year later. Transient transfection of ERβ resulted in higher levels of its mRNA as compared with the stably transfected SW480 cells; functional ERβ receptors, however, showed similar number of expression (Figure 1B). We investigated the regulations of MYC and a number of miRNAs 48 h after ERβ transfection using real-time PCR, however, we did not identify the same changes in SW620. It is possible that ERβ does not mediate the same protective mechanism in more advanced and invasive stages, or that stable expression is required.

Fig. 4. ERβ expression affects miR-200a/b, morphology, migration and epithelial–mesenchymal transition markers. (A) ERβ expression in SW480 cells dramatically changes the cells’ morphology into a more stretched out and differentiated phenotype (in absence of E2). (B) Migration assay of SW480 control and ERβ-expressing cells. A scratch assay (in absence of E2 and measured at 8 h) revealed that the ERβ-expressing cells are less migratory. (C) ZEB1 was upregulated and E-cadherin was downregulated following ERβ expression. (D) Re-introduction of miR-200a/b in SW480-ERβ cells returned ZEB1 and E-cadherin mRNA levels (top) resembling those in control cells (Figure 4C), along with protein levels (E-cadherin, bottom). (E) Expression of miR-200a is downregulated also in a second colon cancer cell line, HT29, following introduction of ERβ. Fold-change (in panels C and D) indicates the relative change in miRNA or mRNA expression compared with controls and normalized to U6 or 18S, respectively. Unpaired two-tailed t-test was used to test the significance, presented as *P ≤ 0.05, **P ≤ 0.01, ***P ≤ 0.001.

Page 8 of 11

by guest on March 20, 2013

http://carcin.oxfordjournals.org/D

ownloaded from

ERβ affects miRNAs in colon cancer

In conclusion, we show that the previously established effect of ERβ on colon cancer proliferation, involves regulation of the miR-17–92 cluster, and that morphology and migration, and regulation of miR-200a/b-ZEB1-E-cadherin levels, are affected by ERβ in SW480 cells.

Discussion

The aim of this study was to investigate whether ERβ expression impacts miRNA regulation in colon cells and to begin dissecting its potential influence. Here, we present evidence that introduction of ERβ in SW480 colon cancer cells induces significant changes in the miRNA pool. To our knowledge, this is the first study to show that expression of ERβ plays an important role in regulating miRNAs in colon cells. We have identified and confirmed 27 miRNAs to be changed following ERβ expression (Figure 2) and most of these miR-NAs are implicated as deregulated in different cancers (85–87). We further show that many of the regulated miRNAs have predicted or confirmed mRNA targets that were concordantly changed upon ERβ expression (Supplementary Figure  2, available at Carcinogenesis Online). Using functional studies, we show that the miRNA regula-tions play a critical role affecting proliferation, apoptosis and migra-tion (Figures 3 and 4). We conclude that parts of the ERβ-induced transcriptome changes and functions are due to changes in miRNA expression, highlighting the complexity of gene regulation.

Twenty-three of the changed miRNAs have an ERβ chromatin-binding site in proximity (<100 kb) to the chromosomal location of the stem-loop of the miRNA (Table II). This observation suggests that several miRNAs are directly regulated by ERβ, offering insight into ERβ function in addition to mRNA transcription. ERβ chromatin-binding sites have only been established for breast cancer cells, and further binding studies in colon cancer cells are needed to prove a direct mechanistic role.

ERβ also regulates several other transcription factors in these cells (29), which may control their own set of miRNAs. We show that sev-eral miRNA regulations, including the repression of the oncogenic miR-17–92 cluster, are a consequence of the ERβ-driven downregula-tion of MYC (Figure 3B and Supplementary Figure 3B, available at Carcinogenesis Online). The miR-17–92 cluster was also reported as regulated by ERβ in breast cancer MCF-7 cells (53), whereas overlap of other miRNAs was low between these two studies. Some miRNAs have both an ERβ and a MYC-binding site in their proximity (e.g. miR-9-3, mirR-106a, miR-135a-1) and it is possible that ERβ fine-tunes their regulation both directly and via MYC.

The proposed target gene of miR-17, NCOA3 and its downstream targets were changed in ERβ-expressing cells (Figure 3E), suggest-ing that ERβ influences these genes through its regulation of MYC/miR-17. We confirmed that NCOA3 was upregulated as a conse-quence of miR-17 silencing and repressed as a consequence of miR-17 induction (Figure 3G and H), as well as downstream target MYB (Supplementary Figure 3D, available at Carcinogenesis Online). As NCOA3 also activates ERβ and functions as a histone acetyltrans-ferase, making target genes available for transcription, this pathway is important for ERβ’s transcriptional action. Further studies are needed to dissect its exact contribution. The upregulation of NCOA3 is likely to be beneficial in a clinical perspective because overexpression of NCOA3 is associated with prolonged overall survival and decreased probability of death in colon cancer (88). We further demonstrate that the ERβ-mediated decrease in cell proliferation can be reversed by re-introduction of miR-17 (Figure 3C). Silencing of miR-17 also increased cell death during DNA-damaging conditions (Figure 3D). Although we have previously suggested that ERβ increases DNA repair in colon cells, we propose that the axis miR-17/CLU is respon-sible for increased apoptotic capacity and that other ERβ-mediated mechanisms, including those described in refs 29 and 89, are respon-sible for increased DNA repair. These separate pathways are of poten-tial importance for future treatments of colon cancer patients who have developed resistance to chemotherapy. As the entire miR-17–92

cluster was downregulated in the ERβ-expressing cells, the com-bined effect of all members may play an even more significant role. We explored whether ERβ would mediate the same mechanism in more advanced cancers. Transient expression of ERβ in the metastatic SW620 cells did not repress MYC nor downstream miRNAs (data not shown). As we and others have found that ERβ frequently represses MYC in less metastatic cell lines, ERβ may not mediate the same protective mechanism in more advanced and invasive stages, which have major abnormalities. This area requires further investigation in the future. This indicates that miR-17 inhibitors in combination with chemotherapy may render colon tumors more sensitive to treatment, whereas ERβ activation, increasing DNA repair and attenuating pro-liferation, may be most beneficial from a prevention or early-stage treatment perspective.

MiR-200a/b was strongly downregulated as a consequence of ERβ expression (Figure  2B) along with increased ZEB1 and decreased E-cadherin (Figure 4C). After introduction of the miR-200a/b mimics, the levels of ZEB1 and E-cadherin returned to the levels resembling those of control cells where ERβ is not expressed (Figure  4D). Although restoration of miR-200a has been shown to repress migration in mammary epithelial cells (90), its downregulation was accompanied by lower migration in SW480 cells. Scratch assay after transient miR-200b mimic transfection indicated no significant change in migration (data not shown). Further studies are needed to explore the functional consequences of this regulation in colon cancer cells. Of note is that the effects of miR-200a/b are multiple and cell specific. We observed that miR200a/b mimics also increased MYC expression (Supplementary Figure  3C, available at Carcinogenesis Online) and that related miR-200c has been reported to increase prolif-eration in colorectal cancer (91). Thus, the ERβ-mediated direct downregulation of MYC is enhanced by the concurrent down-regulation of miR-200a/b. We see a similar pattern indicated in the regulation of another oncogene, MYB: a potential regulation both through a direct mechanism by ERβ binding to the MYB locus (29) as well as through the miR-17 cascade (Figure 3F and Supplementary Figure  3D, available at Carcinogenesis Online). These constitute examples of how ERβ represses oncogenes (MYC and MYB) through multiple pathways.

We did not observe any ligand-dependent regulations by ERβ. This correlates to previous observations that ERβ when expressed exoge-nously in colorectal cancer cell lines often exhibit ligand-independent behavior. This does not imply that ERβ does not rely on conforma-tional changes or is ligand independent when endogenously expressed in vivo. In cell lines, high levels of growth factors and mitogen-acti-vated protein kinase/extracellular signal-regulated kinase pathways along with endogenous estrogen synthesis (92) may activate ERβ. Thereby, high enough levels of the active conformation may be pre-sent to fully activate the limited endogenous templates. By perform-ing radiolabeled ligand-binding assays, we ascertained that the levels of exogenously expressed ERβ are within ‘physiological’ ranges and that it binds ligand (Figure  1B). We cannot, however, exclude that non-genomic effects contribute.

In conclusion, we show novel data demonstrating that introduction of ERβ in colon cancer cells induces changes in the miRNA pool, resulting in subsequent mRNA regulations and functional effects. We illustrate corresponding regulation cascades and functional impacts. We propose that the capacity of miRNAs to target multiple pathways help elucidate the genome-wide antitumorigenic regulations mediated by ERβ in colon cancer cells. As ERβ appears to have a significant potential in preventing or attenuating colon cancer development, our results provide insights into the mechanism whereby ERβ acts in the colon and can aid in the development of better diagnostic tools, pre-vention and treatment of colon cancer.

Supplementary material

Supplementary Figures 1–3 and Tables 1–2 can be found at http://carcin.oxfordjournals.org/

Page 9 of 11

by guest on March 20, 2013

http://carcin.oxfordjournals.org/D

ownloaded from

K.Edvardsson et al.

Funding

Faculty start-up funding (C.W.) from the University of Houston; Texas Emerging Technology Fund (300-9-1958); the Cancer Prevention Research Institute of Texas (RP110444); the Swedish Cancer Society; the Robert A.  Welch Foundation (E-0004 to J.Å.G.); Knut and Alice Wallenberg Foundation (F.P., The Human Protein Atlas project).

Acknowledgements

We thank Dr Anders Ström, Dr Margaret Warner and Ryan Butler (University of Houston, TX) for generous advice; Dr Kenneth Wester (Human Protein Atlas, Uppsala, Sweden) for collaboration on tissue microarray staining and Dr Anna Ehrlund (Karolinska Institutet, Sweden) for critical reading of the manuscript.

Conflict of Interest Statement: None declared (K.E., T.N.V., S.M.K. and F.P.). J.Å.G is stockholder of Karo Bio AB. C.W.  and J.Å.G are consultants to Prometheus Laboratories.

References

1. Ferlay,J.S.H et al. (2008) GLOBOCAN 2008 v1.2, Cancer incidence and mortality worldwide. IARC Scientific Publications No. 10. International Agency for Research on Cancer, 2010, Lyon. Available from: http://globo-can.iarc.fr (22 October 2011, date last accessed).

2. Siegel,R. et al. (2011) Cancer statistics, 2011: the impact of eliminating socioeconomic and racial disparities on premature cancer deaths. CA. Cancer J. Clin., 61, 212–236.

3. Brozek,W. et  al. (2009) Mutual associations between malignancy, age, gender, and subsite incidence of colorectal cancer. Anticancer Res., 29, 3721–3726.

4. Hendifar,A. et al. (2009) Gender disparities in metastatic colorectal cancer survival. Clin. Cancer Res., 15, 6391–6397.

5. Koo,J.H. et al. (2008) Improved survival in young women with colorectal cancer. Am. J. Gastroenterol., 103, 1488–1495.

6. Calle,E.E. et  al. (1995) Estrogen replacement therapy and risk of fatal colon cancer in a prospective cohort of postmenopausal women. J. Natl. Cancer Inst., 87, 517–523.

7. Franceschi,S. et al. (2000) Menopause and colorectal cancer. Br. J. Cancer, 82, 1860–1862.

8. Rossouw,J.E. et  al.; Writing Group for the Women’s Health Initiative Investigators. (2002) Risks and benefits of estrogen plus progestin in healthy postmenopausal women: principal results From the Women’s Health Initiative randomized controlled trial. JAMA, 288, 321–333.

9. Newcomb,P.A. et  al. (2007) Estrogen plus progestin use, microsatellite instability, and the risk of colorectal cancer in women. Cancer Res., 67, 7534–7539.

10. Lin,K.J. et al. (2012) The effect of estrogen vs. combined estrogen-pro-gestogen therapy on the risk of colorectal cancer. Int. J.  Cancer, 130, 419–430.

11. Fernandez,E. et al. (2001) Oral contraceptives and colorectal cancer risk: a meta-analysis. Br. J. Cancer, 84, 722–727.

12. Weige,C.C. et al. (2009) Estradiol alters cell growth in nonmalignant colo-nocytes and reduces the formation of preneoplastic lesions in the colon. Cancer Res., 69, 9118–9124.

13. Williams,C. et al. (2008) A genome-wide study of the repressive effects of estrogen receptor beta on estrogen receptor alpha signaling in breast cancer cells. Oncogene, 27, 1019–1032.

14. Chang,E.C. et al. (2006) Impact of estrogen receptor beta on gene networks regulated by estrogen receptor alpha in breast cancer cells. Endocrinology, 147, 4831–4842.

15. Foley,E.F. et al. (2000) Selective loss of estrogen receptor beta in malignant human colon. Cancer Res., 60, 245–248.

16. Campbell-Thompson,M. et  al. (2001) Expression of estrogen receptor (ER) subtypes and ERbeta isoforms in colon cancer. Cancer Res., 61, 632–640.

17. Konstantinopoulos,P.A. et  al. (2003) Oestrogen receptor beta (ERbeta) is abundantly expressed in normal colonic mucosa, but declines in colon adenocarcinoma paralleling the tumour’s dedifferentiation. Eur. J. Cancer, 39, 1251–1258.

18. Honma,N. et  al. (2013) Estrogen receptor-β gene polymorphism and colorectal cancer risk: effect modified by body mass index and isoflavone intake. Int. J. Cancer, 132, 951–958.

19. Wu,H. et  al. (2012) Association of estrogen receptor beta variants and serum levels of estradiol with risk of colorectal cancer: a case control study. BMC Cancer, 12, 276.

20. Passarelli,M.N. et  al. (2013) Common single-nucleotide polymorphisms in the estrogen receptor β promoter are associated with colorectal cancer survival in postmenopausal women. Cancer Res., 73, 767–775.

21. Giroux,V. et al. (2008) Estrogen receptor beta deficiency enhances small intestinal tumorigenesis in ApcMin/+ mice. Int. J. Cancer, 123, 303–311.

22. Giroux,V. et  al. (2011) Chemopreventive effect of ERβ-Selective ago-nist on intestinal tumorigenesis in Apc(Min/+) mice. Mol. Carcinog., 50, 359–369.

23. Saleiro,D. et  al. (2010) Enhanced induction of mucin-depleted foci in estrogen receptor {beta} knockout mice. Cancer Prev. Res. (Phila)., 3, 1198–1204.

24. Saleiro,D. et al. (2012) Estrogen receptor-β protects against colitis-associ-ated neoplasia in mice. Int. J. Cancer, 131, 2553–2561.

25. Martineti,V. et al. (2005) ERbeta is a potent inhibitor of cell proliferation in the HCT8 human colon cancer cell line through regulation of cell cycle components. Endocr. Relat. Cancer, 12, 455–469.

26. Hartman,J. et al. (2009) Tumor repressive functions of estrogen receptor beta in SW480 colon cancer cells. Cancer Res., 69, 6100–6106.

27. Bielecki,A. et  al. (2011) Estrogen receptor-β mediates the inhibition of DLD-1 human colon adenocarcinoma cells by soy isoflavones. Nutr. Cancer, 63, 139–150.

28. Tu,Z. et  al. (2012) Estrogen receptor β potentiates the antiproliferative effect of raloxifene and affects the cell migration and invasion in HCT-116 colon cancer cells. J. Cancer Res. Clin. Oncol., 138, 1091–1103.

29. Edvardsson,K. et al. (2011) Estrogen receptor β induces antiinflammatory and antitumorigenic networks in colon cancer cells. Mol. Endocrinol., 25, 969–979.

30. Johnson,C.D. et al. (2007) The let-7 microRNA represses cell proliferation pathways in human cells. Cancer Res., 67, 7713–7722.

31. Friedman,R.C. et al. (2009) Most mammalian mRNAs are conserved tar-gets of microRNAs. Genome Res., 19, 92–105.

32. Calin,G.A. et al. (2004) Human microRNA genes are frequently located at fragile sites and genomic regions involved in cancers. Proc. Natl. Acad. Sci. U.S.A., 101, 2999–3004.

33. Farazi,T.A. et  al. (2011) miRNAs in human cancer. J. Pathol., 223, 102–115.

34. Michael,M.Z. et al. (2003) Reduced accumulation of specific microRNAs in colorectal neoplasia. Mol. Cancer Res., 1, 882–891.

35. Nugent,M. et al. (2011) MicroRNAs in colorectal cancer: function, dys-regulation and potential as novel biomarkers. Eur. J.  Surg. Oncol., 37, 649–654.

36. Wu,W.K. et al. (2011) MicroRNA in colorectal cancer: from benchtop to bedside. Carcinogenesis, 32, 247–253.

37. Zhao,C. et al. (2009) A feedback regulatory loop involving microRNA-9 and nuclear receptor TLX in neural stem cell fate determination. Nat. Struct. Mol. Biol., 16, 365–371.

38. Bethke,A. et al. (2009) Nuclear hormone receptor regulation of microR-NAs controls developmental progression. Science, 324, 95–98.

39. Shah,Y.M. et al. (2007) Peroxisome proliferator-activated receptor alpha regulates a microRNA-mediated signaling cascade responsible for hepato-cellular proliferation. Mol. Cell. Biol., 27, 4238–4247.

40. Smith,L.K. et  al. (2010) Glucocorticoids modulate microRNA expres-sion and processing during lymphocyte apoptosis. J. Biol. Chem., 285, 36698–36708.

41. Cochrane,D.R. et al. (2011) Steroid receptors and microRNAs: relation-ships revealed. Steroids, 76, 1–10.

42. Eendebak,R.J. et al. (2011) Nuclear receptors and microRNAs: who regu-lates the regulators in neural stem cells? FEBS Lett., 585, 717–722.

43. Bhat-Nakshatri,P. et  al. (2009) Estradiol-regulated microRNAs control estradiol response in breast cancer cells. Nucleic Acids Res., 37, 4850–4861.

44. Maillot,G. et  al. (2009) Widespread estrogen-dependent repression of microRNAs involved in breast tumor cell growth. Cancer Res., 69, 8332–8340.

45. Castellano,L. et al. (2009) The estrogen receptor-alpha-induced microRNA signature regulates itself and its transcriptional response. Proc. Natl. Acad. Sci. U.S.A., 106, 15732–15737.

46. Kondo,N. et al. (2008) miR-206 expression is down-regulated in estrogen receptor alpha-positive human breast cancer. Cancer Res., 68, 5004–5008.

47. Wickramasinghe,N.S. et  al. (2009) Estradiol downregulates miR-21 expression and increases miR-21 target gene expression in MCF-7 breast cancer cells. Nucleic Acids Res., 37, 2584–2595.

48. de Souza Rocha Simonini,P. et  al. (2010) Epigenetically deregulated microRNA-375 is involved in a positive feedback loop with estrogen recep-tor alpha in breast cancer cells. Cancer Res., 70, 9175–9184.

Page 10 of 11

by guest on March 20, 2013

http://carcin.oxfordjournals.org/D

ownloaded from

ERβ affects miRNAs in colon cancer

49. Di Leva,G. et  al. (2010) MicroRNA cluster 221-222 and estrogen receptor alpha interactions in breast cancer. J. Natl. Cancer Inst., 102, 706–721.

50. Xiong,J. et al. (2010) An estrogen receptor alpha suppressor, micro-RNA-22, is downregulated in estrogen receptor alpha-positive human breast cancer cell lines and clinical samples. FEBS J., 277, 1684–1694.

51. Katchy,A. et  al. (2012) Estradiol-activated estrogen receptor α does not regulate mature microRNAs in T47D breast cancer cells. J. Steroid Biochem. Mol. Biol., 128, 145–153.

52. Klinge,C.M. (2012) miRNAs and estrogen action. Trends Endocrinol. Metab., 23, 223–233.

53. Paris,O. et  al. (2012) Direct regulation of microRNA biogenesis and expression by estrogen receptor beta in hormone-responsive breast cancer. Oncogene, 31, 4196–4206.

54. Saji,S. et al. (2000) Estrogen receptors alpha and beta in the rodent mam-mary gland. Proc. Natl. Acad. Sci. U.S.A., 97, 337–342.

55. Grober,O.M. et al. (2011) Global analysis of estrogen receptor beta binding to breast cancer cell genome reveals an extensive interplay with estrogen receptor alpha for target gene regulation. BMC Genomics, 12, 36.

56. Hua,S. et al. (2008) Genomic analysis of estrogen cascade reveals histone variant H2A.Z associated with breast cancer progression. Mol. Syst. Biol., 4, 188.

57. Zhao,C. et  al. (2010) Genome-wide mapping of estrogen receptor-beta-binding regions reveals extensive cross-talk with transcription factor acti-vator protein-1. Cancer Res., 70, 5174–5183.

58. Aydoğdu,E. et al. (2012) MicroRNA-regulated gene networks during mam-mary cell differentiation are associated with breast cancer. Carcinogenesis, 33, 1502–1511.

59. Esslimani-Sahla,M. et al. (2004) Estrogen receptor beta (ER beta) level but not its ER beta cx variant helps to predict tamoxifen resistance in breast cancer. Clin. Cancer Res., 10, 5769–5776.

60. Uhlen,M. et al. (2010) Towards a knowledge-based Human Protein Atlas. Nat. Biotechnol., 28, 1248–1250.

61. Holbeck,S. et al. (2010) Expression profiling of nuclear receptors in the NCI60 cancer cell panel reveals receptor-drug and receptor-gene interac-tions. Mol. Endocrinol., 24, 1287–1296.

62. Katzenellenbogen,B.S. et  al. (1987) Proliferation, hormonal responsive-ness, and estrogen receptor content of MCF-7 human breast cancer cells grown in the short-term and long-term absence of estrogens. Cancer Res., 47, 4355–4360.

63. Webb,P. et al. (1992) The limits of the cellular capacity to mediate an estro-gen response. Mol. Endocrinol., 6, 157–167.

64. Zhang,Z. et al. (2010) Annexin 1 induced by anti-inflammatory drugs binds to NF-kappaB and inhibits its activation: anticancer effects in vitro and in vivo. Cancer Res., 70, 2379–2388.

65. Wu,P.P. et al. (2009) Restoration of DLC1 gene inhibits proliferation and migration of human colon cancer HT29 cells. Ann. Clin. Lab. Sci., 39, 263–269.

66. Yabuta,N. et al. (2011) The tumor suppressor Lats2 is pivotal in Aurora A and Aurora B signaling during mitosis. Cell Cycle, 10, 2724–2736.

67. Sivarajasingham,N.S. et al. (2003) Identifying a region of interest in site- and stage-specific colon cancer on chromosome 13. Ann. Surg. Oncol., 10, 1095–1099.

68. Carroll,J.S. et al. (2005) Chromosome-wide mapping of estrogen recep-tor binding reveals long-range regulation requiring the forkhead protein FoxA1. Cell, 122, 33–43.

69. O’Donnell,K.A. et al. (2005) c-Myc-regulated microRNAs modulate E2F1 expression. Nature, 435, 839–843.

70. Bui,T.V. et  al. (2010) Myc: maestro of microRNAs. Genes Cancer, 1, 568–575.

71. Kim,J.W. et al. (2010) Myc-induced microRNAs integrate Myc-mediated cell proliferation and cell fate. Cancer Res., 70, 4820–4828.

72. Chang,T.C. et al. (2008) Widespread microRNA repression by Myc con-tributes to tumorigenesis. Nat. Genet., 40, 43–50.

73. Ma,L. et al. (2010) miR-9, a MYC/MYCN-activated microRNA, regulates E-cadherin and cancer metastasis. Nat. Cell Biol., 12, 247–256.

74. Lin,C.H. et  al. (2009) Myc-regulated microRNAs attenuate embryonic stem cell differentiation. EMBO J., 28, 3157–3170.

75. Diosdado,B. et al. (2009) MiR-17-92 cluster is associated with 13q gain and c-myc expression during colorectal adenoma to adenocarcinoma pro-gression. Br. J. Cancer, 101, 707–714.

76. Yu,G. et al. (2012) Prognostic values of the miR-17-92 cluster and its para-logs in colon cancer. J. Surg. Oncol., 106, 232–237.

77. He,L. et al. (2005) A microRNA polycistron as a potential human onco-gene. Nature, 435, 828–833.

78. Zhang,J. et  al. (2012) miR-21, miR-17 and miR-19a induced by phos-phatase of regenerating liver-3 promote the proliferation and metastasis of colon cancer. Br. J. Cancer, 107, 352–359.

79. Yang,L. et al. (2012) Sorghum phenolics demonstrate estrogenic action and induce apoptosis in nonmalignant colonocytes. Nutr. Cancer, 64, 419–427.

80. Schleipen,B. et  al. (2011) ERβ-specific agonists and genistein inhibit proliferation and induce apoptosis in the large and small intestine. Carcinogenesis, 32, 1675–1683.

81. Dews,M. et al. (2010) The myc-miR-17~92 axis blunts TGF{beta} signal-ing and production of multiple TGF{beta}-dependent antiangiogenic fac-tors. Cancer Res., 70, 8233–8246.

82. Wong,C.W. et  al. (2001) Structure-function evaluation of ER alpha and beta interplay with SRC family coactivators. ER selective ligands. Biochemistry, 40, 6756–6765.

83. Leo,C. et al. (2000) Differential mechanisms of nuclear receptor regulation by receptor-associated coactivator 3. J. Biol. Chem., 275, 5976–5982.

84. Gregory,P.A. et al. (2008) The miR-200 family and miR-205 regulate epi-thelial to mesenchymal transition by targeting ZEB1 and SIP1. Nat. Cell Biol., 10, 593–601.

85. Slaby,O. et  al. (2009) MicroRNAs in colorectal cancer: translation of molecular biology into clinical application. Mol. Cancer, 8, 102.

86. Panarelli,N.C. et al. (2011) MicroRNA expression in selected carcinomas of the gastrointestinal tract. Patholog. Res. Int., 2011, 124608.

87. Luo,X. et al. (2011) MicroRNA signatures: novel biomarker for colorectal cancer? Cancer Epidemiol. Biomarkers Prev., 20, 1272–1286.

88. Grivas,P.D. et al. (2009) Estrogen receptor alpha/beta, AIB1, and TIF2 in colorectal carcinogenesis: do coregulators have prognostic significance? Int. J. Colorectal Dis., 24, 613–622.

89. Weige,C.C. et al. (2012) P53 mediates estradiol induced activation of apop-tosis and DNA repair in non-malignant colonocytes. J. Steroid Biochem. Mol. Biol., 128, 113–120.

90. Eades,G. et al. (2011) miR-200a regulates SIRT1 expression and epithelial to mesenchymal transition (EMT)-like transformation in mammary epithe-lial cells. J. Biol. Chem., 286, 25992–26002.

91. Hur,K. et al. (2012) MicroRNA-200c modulates epithelial-to-mesenchy-mal transition (EMT) in human colorectal cancer metastasis. Gut. [Epub ahead of print]

92. Xu,X. et al. (2012) Concentration of endogenous estrogens and estrogen metabolites in the NCI-60 human tumor cell lines. Genome Med., 4, 31.

Received September 14, 2012; revised January 28, 2013; accepted February 10, 2013

Page 11 of 11

by guest on March 20, 2013

http://carcin.oxfordjournals.org/D

ownloaded from