MicroRNA-221/222 Negatively Regulates Estrogen Receptorα and Is Associated with Tamoxifen Resistance in Breast Cancer*

A search for regulators of estrogen receptor α (ERα) expression has yielded a set of microRNAs (miRNAs) for which expression is specifically elevated in ERα-negative breast cancer. Here we show distinct expression of a panel of miRNAs between ERα-positive and ERα-negative breast cancer cell lines and primary tumors. Of the elevated miRNAs in ERα-negative cells, miR-221 and miR-222 directly interact with the 3′-untranslated region of ERα. Ectopic expression of miR-221 and miR-222 in MCF-7 and T47D cells resulted in a decrease in expression of ERα protein but not mRNA, whereas knockdown of miR-221 and miR-222 partially restored ERα in ERα protein-negative/mRNA-positive cells. Notably, miR-221- and/or miR-222-transfected MCF-7 and T47D cells became resistant to tamoxifen compared with vector-treated cells. Furthermore, knockdown of miR-221 and/or miR-222 sensitized MDA-MB-468 cells to tamoxifen-induced cell growth arrest and apoptosis. These findings indicate that miR-221 and miR-222 play a significant role in the regulation of ERα expression at the protein level and could be potential targets for restoring ERα expression and responding to antiestrogen therapy in a subset of breast cancers.

Estrogen receptor ␣ (ER␣) 3 is an important marker for prognosis and is predictive of response to endocrine therapy in patients with breast cancer. Although the majority of primary breast cancers are ER␣-positive and respond to antiestrogen therapy, up to one-third of patients with breast cancer lack ER␣ at the time of diagnosis, and a fraction of breast cancers that are initially ER␣-positive lose ER␣ expression during tumor progression (1). These patients fail to respond to antiestrogen therapy and have a poor prognosis. Previous studies have shown that ER␣ absence is a result of hypermethylation of CpG islands in the 5Ј-regulatory regions of ER␣ in a fraction of breast cancers (1). However, the molecular mechanism of the rest of the ER␣-negative cases and the molecule(s) involving ER␣ hypermethylation remain largely unknown (1).
MicroRNAs (miRNAs) are a new class of small (ϳ22 nucleotide) noncoding RNAs and negatively regulate protein-coding gene expression by targeting mRNA degradation or translation inhibition (2)(3)(4)(5). Frequent deregulation of miRNAs has been detected in breast cancer, and some are associated with breast cancer metastasis and poor prognosis, suggesting an important role of miRNAs in breast oncogenesis and cancer progression (6 -9). In this study, we performed miRNA profiling in ER␣negative versus ER␣-positive human breast cancer cell lines and primary tumors and identified the deregulation of a panel of miRNAs in ER␣-negative breast cancer. Of the elevated miRNAs, miR-221 and miR-222 were found to directly regulate ER␣ expression by interaction with the 3Ј-untranslated region (3Ј-UTR) of ER␣. Ectopic expression of miR-221 and/or miR-222 reduced ER␣ levels in MCF-7 and T47D cells, whereas knockdown of miR-221 and/or miR-222 restored ER␣ expression and tamoxifen sensitivity in MDA-MB-468 cells. These results indicate that miR-221 and miR-222 could play a pivotal role in the regulation of ER␣ expression in a subset of breast cancers.
RNA Isolation and miRNA Microarray and Northern Blot Analyses-Total RNA was isolated from cell lines and tissue samples using TRIzol reagents (Invitrogen). miRNA microarray and Northern blot analyses were performed as described previously (10 -12). Briefly, a custom miRNA microarray platform containing 515 miRNAs was hybridized with [␥-32 P]ATPlabeled low molecular weight RNAs. To ensure accuracy of the hybridizations, each labeled RNA sample was hybridized with three separate membranes. Northern blot analysis was performed by separation of total RNA on 15% denaturing polyacrylamide gel and hybridized with the probes indicated in the figure legends. The probe sequences were as follows: miR-221, 5Ј-GAAACCCAGCAGACAATGTAGCT-3Ј; miR-222, 5Ј-ACCCAGTAGCCAGATGTAGCT-3Ј; and U6, 5Ј-CGTTC-CAATTTTAGTATATGTGCTGCCGAAGCGA-3Ј. The blot was quantified using ImageQuant software (GE Healthcare).
Cell Viability and Apoptosis Assays-Cell viability was examined by 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide assay as described previously (10). Cells were seeded in a 96-well plate. After a 24-h incubation, the cells were treated with tamoxifen (5, 10, and 20 M) or a dimethyl sulfoxide control for 48 h and then subjected to 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide and apoptotic detection assay using a Cell Death Detection ELISAPLUS kit (Roche Applied Science) according the manufacturer's protocol. Each experiment was repeated three times in triplicate. The results are expressed as the enrichment factor relative to the untreated controls.
Statistical Analysis-Statistical significance was analyzed by unpaired Student's t test, and p Յ 0.05 was considered to be statistically significant.

miR-221 and miR-222 Are Highly Expressed in ER␣-negative Breast Cancer Cell Lines and Primary Tumors-In an attempt
to identify the miRNAs that contribute to regulation of ER␣ expression in breast cancer, we performed miRNA profiling in ER␣-positive versus ER␣-negative breast cancer cell lines as well as primary tumors. RNAs isolated from a total of five cell lines and 10 primary tumors were hybridized to a custom miRNA microarray platform containing 515 miRNAs. After three times of hybridization, quantification, and normalization, a dozen miRNAs, especially miR-221 and miR-222, were elevated in the ER␣-negative cell lines and primary tumors compared with ER␣-positive breast cancers (Fig. 1A). Consistent with the miRNA microarray data, Northern blot analysis revealed the expression of miR-221 and miR-222 in five of eight ER␣-negative cell lines examined, with higher levels in MDA-miR-221/222 Targets ER␣ FIGURE 1. Frequently increased expression of miR-221 and miR-222 in ER␣-negative breast cancer. A, partial heat map of miRNA microarray analysis of ER␣-positive versus ER␣-negative breast cancer cell lines and primary tumors. Several miRNAs were significantly elevated in ER␣-negative cells. B and C, elevated levels of miR-221 and miR-222 in ER␣-negative breast cancer cell lines and primary tumors. Total RNAs from the cell lines and primary tumors were subjected to Northern blot (B) and quantitative RT-PCR (C) analyses. U6 small nuclear RNA (snRNA) and glyceraldehyde-3-phosphate dehydrogenase (GAPDH) were used as loading controls. The blots were quantified by dividing miR-221 and miR-222 signals by U6 and by dividing ER␣ by glyceraldehyde-3-phosphate dehydrogenase. ER␣-negative tumors overexpressing both miR-221 and miR-222 are labeled by asterisks, and the tumors also expressing ER␣ mRNA are indicated by triangles (C). M stands for marker. D, representation of the inverse correlation of expression of ER␣ and miR-221/miR-222. Breast tumor specimens were immunohistochemically stained with anti-ER␣ antibody (first and third panels). The second and fourth panels are the same specimens that were hybridized with the LNA-miR-221 and LNA-miR-222 probes using miRNA locked nucleic acid in situ hybridization as described under "Experimental Procedures." MB-468, Hs578T, and MDA-MB-231 cells (Fig. 1B). Notably, all four ER␣-positive breast cancer lines had very low levels of miR-221 and miR-222 (Fig. 1B). Furthermore, RT-PCR, immunostaining, and miRNA in situ hybridization analyses revealed overexpression of miR-221 and miR-222 in 13 of 25 (52%) ER␣negative primary tumors, 11 of which had ER␣ mRNA expression (Fig. 1C). In contrast, of 16 ER␣-positive tumors examined, only four expressed moderate levels of miR-221 and miR-222  (ER␣1; upper), and the other is not (ER␣ 2; lower). Mutants of pmiR-ER␣1-5 (seed sequence mutation) and pmiR-ER␣1-3 are shown (middle), and the mutant nucleotides are labeled in red. B, miR-221 and miR-222 inhibit ER␣ expression in a dose-dependent manner. MCF-7 cells were transiently transfected with indicated amounts of BLOCK-iT-miR-221 and BLOCK-iT-miR-222 expression plasmids. Following a 72-h incubation, cells were subjected to immunoblotting with anti-ER␣ (first panel) and ␤-actin (second panel) antibodies and quantitative RT-PCR analysis (third through fifth panels). Expression of transfected miR-221 and miR-222 is shown in the third and fourth panels. U6 small nuclear RNA (snRNA) is a loading control. Quantification was done by dividing ER␣ signals by actin. M stands for marker. C, decrease in ER␣ protein but not mRNA levels by stable expression of miR-221 or miR-222 in MCF-7 and T47D cells. Following transfection of BLOCK-iT-miR-221 or BLOCK-iT-miR-222, cells were selected with blasticidin. Stably transfected cells were subjected to quantitative RT-PCR (upper panels), Western blot (middle panels), and RT-PCR (lower panels) analyses. Dividing ER␣ signals by actin (Western) and glyceraldehyde-3-phosphate dehydrogenase (GAPDH; RT-PCR) was used to quantify the protein and mRNA levels of ER␣, respectively. D, immunofluorescence staining of parental MCF-7 cells (panels A-D) and cells transiently transfected with the GFP vector (panels E-H), GFP-miR-221 (panels I-L), or GFP-miR-222 (panels M-P) with anti-ER␣ antibody (panels C, G, K, and O). Cells transfected with vector (e.g. expressing only GFP) exhibited the same levels of ER␣ as did parental cells (panels A-D). ER␣ signals in miR-221-and miR-222-transfected cells (arrowheads) were significantly lower than in untransfected surrounding cells (arrows). Quantitation of ER␣-positive cells is shown in the bar graph. DAPI, 4Ј,6-diamidino-2-phenylindole. E, ectopic expression of miR-221 and miR-222 reduces tamoxifen-induced cell death. MCF-7 and T47D cells were stably transfected with miR-221, miR-222, and the vector alone as described for C. Following treatment with or without tamoxifen, cell viability was examined by 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide assay. The experiment was repeated three times in triplicate. Asterisks indicate p Ͻ 0.05. The gels show expression of transfected miR-221 and miR-222. (Fig. 1, C and D), suggesting that miR-221 and miR-222 could regulate ER␣ expression possibly through inhibition of ER␣ translation.

miR-221/222 Targets ER␣
ER␣ Protein but Not mRNA Is Suppressed by miR-221 and miR-222-Because miRNAs negatively regulate their target genes through base-pairing interaction between their seed sequence and the 3Ј-UTR of target genes, we searched the miRNA TargetScan Database and found that two sequence motifs of the 3Ј-UTR of ER␣ match miR-221 and miR-222 seed sequences, one of which is conserved between human, mouse, and rat ( Fig. 2A). To examine if ER␣ is indeed regulated by miR-221 and miR-222, we ectopically expressed miR-221 and miR-222 in ER␣-positive MCF-7 and T47D cells (Fig. 1B), in which endogenous miR-221 and miR-222 are undetectable by Northern blotting. Transient transfection of increasing amounts of miR-221 and/or miR-222 in MCF-7 cells reduced ER␣ expression in a dose-dependent manner (Fig. 2B). Furthermore, stably miR-221-and miR-222-transfected MCF-7 and T47D cells decreased the protein but not mRNA levels of ER␣ (Fig. 2C). Because the BLOCK-iT plasmids express green fluorescent protein (GFP), we transiently transfected MCF-7 cells with BLOCK-iT-miR-221 and BLOCK-iT-miR-222 as well as the GFP vector alone. Immunofluorescence staining with anti-ER␣ antibody revealed that ER␣ levels were considerably reduced in the cells expressing miR-221 or miR-222 compared with the cells transfected with or without the GFP vector (Fig. 2D). As a result, miR-221-and miR-222-transfected MCF-7 and T47D cells became resistant to tamoxifen-induced cell death (Fig. 2E).

Knockdown of miR-221 and miR-222 in MDA-MB-468 Cells Partially Restores ER␣ Expression and Tamoxifen
Sensitivity-Having demonstrated the miR-221 and miR-222 repression of ER␣ expression at the protein but not mRNA level, we reasoned that up-regulation of miR-221 and miR-222 is responsible for a subset of ER␣ protein-negative/mRNA-positive breast cancers. To test this hypothesis, we examined expression of ER␣ protein and mRNA in 12 breast cancer cell lines. Western blot and RT-PCR analyses revealed that six ER␣ proteinnegative cell lines expressed ER␣ mRNA with more abundance in MDA-MB-468 and MCF-10A cells (Fig. 4A). Both cell lines also had high levels of miR-221 and miR-222 (Fig. 1B). Thus, we  4B). Immunoblot analysis showed that ER␣ protein was partially restored in miR-221 and/or miR-222 knockdown cells but not in control 2Ј-O-Me-treated cells (Fig. 4C). However, there was no significant difference between individual knockdown of miR-221 and miR-222 and their combination (Fig. 4C). Similar results were obtained in the MCF-10A cell line (Fig. 4C). Res-  toration of ER␣ could not be achieved by knockdown of miR-221 and/or miR-222 in ER␣ mRNA-negative cell lines such as MDA-MB-231 and SKBr3 (data not shown). These findings further support the notion that ER␣ is a direct target of miR-221 and miR-222 at the translation level.
We next examined whether the miR-221-and/or miR-222knocked down MDA-MB-468 cells became sensitive to tamoxifen. As shown in Fig. 4 (D and E), knockdown of miR-221 or miR-222 reduced MDA-MB-468 cells resistant to tamoxifeninduced cell growth arrest and apoptosis. Cells with knock-  down of both miR-221 and miR-222 became more vulnerable to tamoxifen-inhibited cell growth (Fig. 4D) and tamoxifen-induced apoptosis (Fig. 4E) compared with cells with knockdown of either one alone.

DISCUSSION
Because expression of ER␣ is a main predictor of response to endocrine therapy, lack of expression of ER␣ is a major mechanism of tamoxifen resistance in breast cancer. In this respect, the loss of ER␣ gene expression has been associated with the aberrant methylation of its CpG islands and histone deacetylation in a fraction of breast cancers (17)(18)(19). A recent report showed that miR-206 represses ER␣ mRNA and protein expression (20). In this study, we have demonstrated frequent up-regulation of miR-221 and miR-222 in ER␣-negative breast cancer cell lines and primary tumors. miR-221 and miR-222 inhibit ER␣ expression at the protein but not mRNA level, indicating the suppression of ER␣ by these two miRNAs at the translational level. Taken collectively, these studies indicate that miRNAs are important regulators of ER␣ and could be major determinants of ER␣ status in human breast cancer.
Previous studies have focused primarily on ER␣ protein expression in breast cancer. Several reports have shown that a subset of ER␣ protein-negative breast cancer cell lines and primary tumors express ER␣ mRNA (21)(22)(23)(24)(25). However, the mechanisms by which the mRNA of ER␣ does not translate to protein are unclear. It was speculated that lack of ER␣ protein is not due to lack of ER␣ gene expression or methylation of its promoter, but might be due to post-transcriptional or post-translational mechanisms (23)(24)(25). Our study has shown that miR-221 and miR-222 inhibit ER␣ translation by direct interaction with the 3Ј-UTR of ER␣ and thus provide a molecular mechanism of ER␣ regulation at the post-transcriptional level in breast cancer.
It has been well documented that each miRNA negatively regulates hundreds of protein-coding genes. Recently, miR-221 and miR-222 have been shown to repress CDK inhibitory proteins p27 Kip1 and p57 as well as the c-Kit receptor, leading to cell proliferation and survival and inhibition of differentiation (26 -31). In this study, we identified ER␣ as a direct target of miR-221 and miR-222. Knockdown of miR-221 and miR-222 restores ER␣ protein expression and sensitizes MDA-MB-468 cells to tamoxifen-induced cell growth arrest and apoptosis (Fig. 4), whereas ectopic expression of miR-221 and miR-222 in MCF-7 and T47D cells reduces the ER␣ protein level and renders the cells resistant to tamoxifen (Fig. 2E). Although miR-221 and miR-222 have an identical eight-nucleotide seed sequence and redundantly regulate p27, p57, and c-Kit (26 -31) as well as ER␣, the effect of the combined knockdown of miR-221 and miR-222 on tamoxifen-induced cell death is more significant than that of knockdown of either one alone (Fig. 4, D  and E). This suggests that miR-221 and miR-222 might target different genes because the rest of nucleotide sequences of miR-221 and miR-222 are quite different ( Fig. 2A).
In summary, we have demonstrated that miR-221 and miR-222 are frequently up-regulated in ER␣-negative breast cancer cell lines and primary tumors. The elevated level of miR-221 and miR-222 is responsible for a subset of ER␣-negative breast tumors that express ER␣ mRNA. Furthermore, overexpression of miR-221 and miR-222 contributes to tamoxifen resistance through negative regulation of ER␣, whereas knockdown of miR-221 and/or miR-222 restores ER␣ expression and tamoxifen sensitivity. Therefore, miR-221 and miR-222 could serve as potential therapeutic targets for a subset of ER␣-negative breast cancers.