Anti-microRNA-222 (Anti-miR-222) and -181B Suppress Growth of Tamoxifen-resistant Xenografts in Mouse by Targeting TIMP3 Protein and Modulating Mitogenic Signal*

Background: MicroRNAs-221/222 and 181b are up-regulated in tamoxifen-resistant breast cancer. Results: Anti-miRs-222/181b regressed tamoxifen-resistant xenografts. Down-regulation of TIMP3, a common target of these miRs, facilitated growth factor signaling by regulating metalloproteases. Conclusion: Survival of tamoxifen-resistant breast cancer is dependent on miR-mediated suppression of TIMP3. Significance: Anti-miRs-222/221 and -181b can be used to render tamoxifen-resistant tumors responsive to the drug. We have shown earlier that miR-221 and -222 are up-regulated in tamoxifen-resistant MCF-7 (OHTR) cells and Her2-positive human breast tumors when compared with Her2 negative tumors. In this study, we report markedly enhanced expression of miR-181b in OHTR cells and endocrine-resistant tumors. Further, anti-miR-222 or -181b in combination with tamoxifen suppressed growth of tamoxifen-resistant xenografts in mice. Luciferase reporter assay and expression analysis showed that TIMP3, a tissue metalloproteinase inhibitor, is a common target of miR-221/222 and -181b. In situ hybridization and immunohistochemical analysis demonstrated reciprocal relationships between TIMP3 and miR-221/222/181b expression in primary human breast carcinomas. Ectopic expression of TIMP3 inhibited growth of the OHTR cells, and its depletion in MCF-7 cells reduced sensitivity to tamoxifen in vitro and in vivo. EGF-induced MAPK and AKT phosphorylation were significantly higher in OHTR cells and miR-221/222-overexpressing MCF-7 cells than in control cells, which suggests modulation of mitogenic signaling by TIMP3 and the miRs. On the contrary, phosphoMAPK and phosphoAKT levels were diminished in TIMP3-overexpressing OHTR cells and increased in TIMP3-depleted MCF-7 cells. Low levels of estrogen or tamoxifen elicited similar differences in phosphoMAPK levels in these cells. Reduced levels of TIMP3 facilitated growth of tamoxifen-resistant cells by alleviating its inhibitory effect on ADAM10 and ADAM17, which are critical for OHTR cell growth. In conclusion, miR-221/222 and -181b facilitate growth factor signaling in tamoxifen-resistant breast cancer by down-regulating TIMP3, and corresponding anti-miRs can be used to render these tumors responsive to tamoxifen.

The estrogen dependence of breast cancer was first recognized more than a century ago when oophorectomy was shown to cause regression of mammary tumors in a subset of premenopausal women (1). Subsequent studies have consistently demonstrated an increased risk of breast cancer associated with elevated levels of estrogen (2,3). Selective estrogen receptor (ER) 4 modulators such as tamoxifen, the first targeted therapy developed against ER-positive breast cancers, are currently the only adjuvant endocrine therapy approved for premenopausal women with this cancer subtype (4). Unfortunately, about 30% of women either fail to respond to tamoxifen or become resistant over time. Multiple mechanisms responsible for endocrine resistance have been proposed that include deregulation of various components of the ER pathway, alterations in cell cycle and cell survival signaling molecules, and the activation of alternative pathways that can facilitate tumor survival in the presence of the drug (for review, see Ref. 5).
Recently, we have explored the potential role of specific microRNAs (miRs) in inducing tamoxifen resistance in breast cancer (9). miRs are small, non-coding RNA molecules that generally block translation by imperfect base pairing to the 3Ј-untranslated regions (3Ј-UTR) of specific mRNAs or inducing mRNA degradation (6). Each miR is thought to regulate multiple genes (7), and each mRNA could harbor seed sequence for multiple miRs. Usually, mRNAs harboring multiple, nonoverlapping binding sites for one or more miRs are more likely to be regulated by miRs than those containing a single miR site (8). Our study showed that miR-221 and -222 are up-regulated in tamoxifen-resistant breast cancer cell lines and Her2-positive primary human breast tumors (9). These two miRs can also confer resistance to this drug when ectopically expressed in tamoxifen-sensitive cells (9,10). This is achieved, at least in part, by targeting cell cycle inhibitory protein p27/Kip1, which is significantly reduced in tamoxifen-resistant cells. Besides miR-221/222, the expression of miR-181b was also significantly augmented in the resistant cells. Here, we demonstrate that treatment with anti-miR-222 or anti-miR-181b can render tamoxifen-resistant xenografts in mice sensitive to the drug and cause suppression of tumor growth and that the tissue metalloprotease inhibitor TIMP3, a direct target of miR-221/222 and miR-181b, plays a key role in this process. Our study also demonstrates that these miRs confer tamoxifen resistance in breast cancer cells by facilitating growth factor signaling, which is ameliorated by high levels of TIMP3 or depletion of the metalloprotease ADAM17, a target of TIMP3, in tamoxifen-sensitive cells.

EXPERIMENTAL PROCEDURES
Cell Culture and Tissue Procurement-Tamoxifen-sensitive MCF-7 cells and resistant OHT R cells were obtained from Dr. Kenneth P. Nephew (Indiana University) and maintained as described (11). T47D cells were obtained from the ATCC (American Type Culture Collection) and maintained as instructed. Primary human breast samples were obtained from the Stephanie Spielman Tissue Bank (Protocol 2003C0036).
Plasmids and Transfections-TIMP3-3ϫFLAG plasmid and primers were described previously (12). For pre-miR transfection, cells were transfected with 60 nM pre-miR or negative control RNA (Applied Biosystems). To knock down TIMP3, stable MCF-7 cells were made using pLKO.1-TIMP3shRNA vector (Thermo Fisher Scientific), and the corresponding empty vector pLKO.1. miR 221/222 binding site on the TIMP3 3Ј-UTR was mutated by site-directed mutagenesis. Primers are provided in the supplemental material. MISSION esiRNA (Sigma) was used for transient knockdown of ADAM17 and ADAM10 from OHT R cells.
TaqMan MicroRNA Assays-Total RNA was reverse-transcribed using the TaqMan microRNA reverse transcription kit, subjected to real-time PCR using the TaqMan microRNA assay kit (Applied Biosystems), and normalized to snRNA RNU6B (13). Reactions were performed using the Stratagene Mx3000 in quadruplicates.
RT-PCR Analysis-TIMP3 and ␤-actin mRNA was measured in cDNA synthesized from DNase-treated total RNA using SYBR Green chemistry.
Western Blot Analysis-Whole cell or tissue extracts were prepared in the cell lysis buffer followed by immunoblotting as described (14). Protein expression levels were quantified by the ImageJ software (rsbweb.nih.gov/ij).
In Situ Hybridization and Immunohistochemical Analysis-These analyses were done as described (15).
Cell Proliferation Assay-Cell proliferation was monitored using the cell proliferation reagent kit I (3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide (MTT); Roche Applied Science). Cells (4000/well) seeded in 96-well plates were serum-starved overnight and treated with 5 M tamoxifen. To measure cell proliferation, 10 l of MTT labeling reagent I was added to each well at the indicated time and incu-bated at 37°C for 4 h followed by the addition of 100 l of solubilization reagent in each well. Absorbance was measured at 570 nm in the ELISA reader (TriStar; Berthold Technologies) after overnight incubation.
Cell Migration Assay-For cell migration, OHT R cells stably expressing pCMV3ϫFLAG or TIMP3 (1 ϫ 10 5 ) were placed in a serum-free medium in 24-well Transwell inserts of 8-m pore size (Corning Costar Corp.) for 24 h. The bottom chambers contained serum-supplemented medium that acted as a chemoattractant. The migration of cells was allowed to proceed for 48 h at 37°C. Cells that migrated to the bottom of the insert were fixed, stained, and counted, and the percentage of migration was determined. Each experiment was performed at least three times.
Wound Healing Assay-Cells were grown in 6-well chambers to 80% confluency. A scratch wound was inflicted diagonally in the monolayer with a pipette tip. The image of the wound was captured at the beginning and every 24 h under the microscope (15).
Growth Factor Treatment-Cells at 60% confluency were starved in phenol red-free medium containing 1% serum for 1 day for EGF treatment (0.1 ng and 1 ng/ml) and 4 days for 17-␤-estradiol (50 pM) or 4-OH-tamoxifen (50 nM) (Sigma) treatment. Whole cell extracts were analyzed by Western blotting.
Mouse Mammary Tumor Induction and Tamoxifen Treatment-Derivatives of MCF-7 or OHT R cells (5 ϫ 10 6 ) were grafted in mammary fat pads of ovariectomized BALB/c nude mice (Charles River Laboratories) with subcutaneous supplementation of 0.72 mg of 17-␤-estradiol pellet (Innovative Research of America). When the tumor reached ϳ100 mm 3 in volume, a 5-mg, 60-day sustained release pellet of tamoxifen citrate (Innovative Research of America) was subcutaneously implanted in each mouse. Length (L) and width (W) of the tumors were measured twice a week, and tumor volume was calculated using the formula LW 2 /6, as described previously (17).
Anti-miR Treatment in Vivo-Xenografts were induced in 5-6-week-old female athymic nude mice with OHT R cells. All the mice received tamoxifen citrate (as described) when the tumor reached ϳ100 mm 3 in volume, and the mice were randomized into three groups. The anti-miR-222 or -181b or negative control oligonucleotide (50 l of 50 nM solution containing 12 l of Optifect) (from Ambion) was injected directly into the tumors (18), once a week for 4 weeks, and tumors were measured weekly. Tumors were harvested and snap-frozen for further studies upon harvest.
Statistical Analysis-Statistical significance of the data was analyzed by unpaired Student's t test, and p Յ 0.05 was considered to be statistically significant and marked as * (Figs. 1 and 3-5). All real-time RT-PCR and Western blot analyses were repeated at least twice. Representative data from 2-3 reproducible experiments are presented, where error bars represent S.D. or S.E. in animal experiments. resistant breast cancer cell lines and tumors, we hypothesized that the corresponding anti-miRs could sensitize the resistant tumors to tamoxifen. To test this hypothesis, OHT R cell xenografts induced in mouse mammary fat pads (ϳ100 mm 3 ) were treated with tamoxifen citrate along with negative control anti-miR or anti-miR-222 or -miR-181b once a week for four consecutive weeks. A significant decrease (40 -45%, p Ͻ 0.001) in the tumor size was observed in mice treated with either anti-miR-181b or anti-miR-222 when compared with the control group (Fig. 1A). It is noteworthy that in both anti-miR-222-treated and anti-miR-181-treated groups, tumor growth was significantly reduced beginning with the 2nd week of treatment when compared with the control group and that some tumor growth regressed completely after four treatments. A marked decrease in the levels of miR-222 and -181b was observed in the recovered tumors treated with the corresponding anti-miR (Fig. 1B).

Anti-miR-222/181 Inhibits Tamoxifen-resistant Tumor Growth
TIMP3 Is a Common Target of miR-221, -222, and -181-To determine whether these three miRs target any common protein and therefore a common pathway, we performed a bioinformatics search (TargetScan, Pictar, and RNA22). Among the several candidate targets, we focused on TIMP3 because the 3Ј-UTR of human TIMP3 (NM_000362) mRNA contained one miR-221/miR-222 (nucleotides 2443-2449) and two miR-181b (nucleotides 3499 -3505, 3573-3579) seed sequences (Fig. 1C). To verify that TIMP3 is a direct target of these miRs in breast cancer cell lines, MCF-7 cells were transfected with pISO-TIMP3-Luc plasmid harboring 3Ј-UTR of TIMP3 (12). Luciferase expression was reduced significantly (ϳ50%, Fig.  1D) when transfected with miR-221, -222, or -181b when compared with the negative control miRs (Fig. 1E). Conversely, deletion or mutation of the miR seed sequences blocked the inhibitory effects of all three miRs, and luciferase activity was completely recovered. Further analysis demonstrated ϳ50% reduction of endogenous TIMP3 mRNA level in OHT R cells as well as miR-221/222-overexpressing cells when compared with MCF-7 cells and when miR-181b was transiently overexpressed (Fig. 1E). In addition, a marked reduction (75%) in TIMP3 protein was observed in OHT R cells when compared with an ϳ50% reduction when individual miRs were overexpressed (Fig. 1F). These data suggest a cumulative inhibitory effect of miR-221/222 and -181 on TIMP3 expression in the OHT R cells. Analysis of the tumor tissues recovered from the mice also revealed an increase in the TIMP3 level in the anti-miR-222 (n ϭ 2) and -181b (n ϭ 2)-treated group when compared with the control group (n ϭ 4) (Fig. 1G), confirming that TIMP3 is a direct target of these miRs in breast tumor.

miR-181b Is Up-regulated in HER2/neu-positive Primary
Breast Tumors When Compared with the Her2/neu-negative Tumors-Increased expression or signaling via EGF receptor/ HER2 pathway has been associated with both experimental and clinical endocrine resistance (19). Therefore, it was of interest to determine whether the increase in miR-181b level is a signature of tamoxifen-resistant mammary tumors using HER2/ neu-positive and -negative primary human breast cancer samples. A significant increase (3.5-fold, p ϭ 0.008) in miR-181b expression ( Fig. 2A) in the HER2/neuϩ tumor samples (n ϭ 24) was observed when compared with the HER2/neu Ϫ samples (n ϭ 25), which is consistent with the increase in miR-221/222 levels in HER2/neuϩ primary breast cancer (9). An inverse correlation between TIMP3 and miRs levels was also demonstrated in primary breast cancer tissues. Tumors with markedly high levels of these miRs exhibited negligible TIMP3 expression (Fig. 2B). In contrast, relatively high levels of TIMP3 expression correlated with low miR expression in some tumors.
Ectopic Expression of TIMP3 Sensitizes Breast Cancer Cells to Tamoxifen-Next, we determined the role of TIMP3 in altering sensitivity of breast cancer cells to tamoxifen in vitro and in vivo. The proliferation of stable FLAG-tagged TIMP3-overexpressing OHT R cells (ϳ3-fold overexpression, Fig. 3A) was significantly reduced in response to 5 M tamoxifen when compared with the control cells (Fig. 3B). In Transwell migration assays where cells were allowed to migrate in response to serum, TIMP3 markedly impeded cell migration in the presence or absence of tamoxifen (Fig. 3C). In wound healing assays, where a scratch-inflicted wound was allowed to heal over time (Fig. 3D), the rate of wound healing after 72 h was comparable between the vector-transfected and OHT R /TIMP3 cells, but the healing process was delayed markedly in the latter cells upon 1.0 M tamoxifen treatment.
To confirm the role of TIMP3 in altering tamoxifen resistance in vivo, the control and OHT R /TIMP3 cells were used to induce tumors in mammary fat pads of mice. Because TIMP3 overexpression impeded tumor growth (supplemental Fig. S1), we generated TIMP3-depleted stable MCF-7 cells (MCF-7-shTIMP3) and the respective control (MCF-7-Vec) cells (Fig.  3E). Although tamoxifen inhibited growth of both cell lines, MCF-7-shTIMP3 cells were relatively less sensitive to the drug when compared with the controls (Fig. 3F). In addition, tamox-ifen treatment resulted in significant reduction of MCF-7-Vec tumors in mice. On the contrary, tumors induced with MCF-7-shTIMP3 cells continued to grow with time, and significant increase in tumor growth was observed when compared with the controls in the presence of the drug at all time points beginning day 60 after injection (Fig. 3G). TIMP3 level at harvest was higher in the control group when compared with the tumors induced with TIMP3-depleted cells (Fig. 3H).
To test further the hypothesis that TIMP3 overexpression can alter response to tamoxifen, TIMP3 was overexpressed in T47D cells (T47D/TIMP3) that are relatively more resist- ant to endocrine therapy and exhibit significantly reduced basal TIMP3 expression when compared with MCF-7 cells (Fig. 4, A and B). In vitro proliferation assay revealed increased (30%) sensitivity of the T47D/TIMP3 cells to tamoxifen-induced growth inhibition when compared with the control cells (Fig. 4C). In vivo treatment with tamoxifen caused significant regression of the T47D/TIMP3-induced tumors, but not the control tumors (Fig. 4D). These data support the hypothesis that TIMP3 could play a key role in mediating response of breast cancer cells to tamoxifen.
TIMP3 Overexpression Curtails Tamoxifen-induced Growth Factor Signaling in OHT R Cells-Endocrine therapy renders the resistant cells dependent on alternate growth factor signaling, such as the EGF pathway, due to accompanying inactivation of ER and therefore reduced availability of estrogen for growth. A marked increase in mitogenic signaling coupled with EGF-mediated MAPK phosphorylation has been demonstrated earlier in tamoxifen-resistant cells (19). Metalloproteases such as ADAM17 and ADM10 are known to facilitate EGF signaling by shedding of EGF receptor ligands, thereby activating the downstream signaling events (20 -22). These metalloproteases are inhibited in cells expressing endogenous tissue metalloprotease inhibitors, such as TIMP3, resulting in altered regulation of several biological functions (23). We hypothesized that reduction in TIMP3 levels and the resultant increase in the metalloprotease activity could contribute to enhanced growth factor signaling in the OHT R cells. Indeed, we observed significantly high levels of ADAM17 and ADAM10 proteins in OHT R cells (p Ͻ 0.05), whereas expression of these metalloproteases is very low in the MCF-7 cells (Fig. 5A). In addition, the activity of MMP9, a matrix metalloprotease, is also significantly elevated in OHT R and miR-221/222-overexpressing MCF-7 cells when compared with the respective control cells (supplemental Fig. S3). Treatment with 0.1 ng/ml EGF induced a significant increase (2.5-fold) in phosphoMAPK level (normalized to total MAPK) in the OHT R but not in the MCF-7 cells (Fig.  5B, compare lanes 2 and 5). This effect was further amplified at 1 ng/ml EGF, and phosphoMAPK levels were consistently higher in OHT R cells when compared with the control cells under all treatment conditions. PhosphoMAPK level was reduced (ϳ65%) in OHT R /TIMP3 cells when compared with the OHT R /Vec cells at both 0.1 ng/ml and 1.0 ng/ml EGF concentrations (Fig. 5C, compare lanes 2 and 3 with lanes 5  and 6). In contrast, depletion of TIMP3 in the MCF-7 cells resulted in increased EGF-induced MAPK phosphorylation. At 0.1 and 1.0 ng/ml EGF, the corresponding level of phos-phoMAPK was 1.5-and 1.8-fold higher in the TIMP3-depleted cells when compared with the control cells (Fig. 5D, panels i and ii). Because TIMP3 is a target of miR-221/222, we explored the possibility that miR-221/222 overexpression can mimic the effect of TIMP3 knockdown in MCF-7 cells in response to EGF. Indeed, phosphoMAPK level was markedly elevated in MCF-7 cells overexpressing miR-221/222 when treated with 0.1 and 1.0 ng/ml EGF (1.7-and 2.5-fold, respectively) but not in control cells (Fig. 5E). Our effort to study the effect of miR-181b on growth factor stimulation failed due to lethality of transient or stable miR-181b overexpression in MCF-7 cells. A sharp decline in ER␣ level within 36 h of miR-181b transfection (supplemental Fig. S2) could be attributed to miR-181 seed sequence on 3Ј-UTR of ER␣ mRNA (TargetScan4.0). Indeed, miR-181b has been shown to suppress estrogen-dependent proliferation of MCF-7 cells earlier (24). We, therefore, knocked down miR-181b in OHT R cells using anti-miR-181b. The MAPK phosphorylation at 0.1 and 1.0 ng/ml EGF was reduced by ϳ50 and 25%, respectively, in anti-miR-181b-transfected cells when compared with the respective controls (Fig. 5F). miR-181b expression was reduced by 60% in cells transfected with the corresponding anti-miR (Fig. 5F, bottom panel).
Both estrogen and tamoxifen elicit rapid increase in MAPK phosphorylation in the tamoxifen-resistant cells when compared with the sensitive cells (19). We explored the possibility that overexpression of miR-221/222 in MCF-7 cells mimics these characteristics of the resistant cells, which could be blocked by overexpressing TIMP3 in OHT R cells.
Indeed, phosphoMAPK was markedly augmented (6 -7-fold, when compared with untreated cells) in MCF-7 cells/miR-221/222-overexpressing cells in response to 17-␤-estradiol within 3-5 min as opposed to 3-4-fold increase in control cells (Fig. 6A). Similarly, treatment with 50 nM tamoxifen elicited robust MAPK phosphorylation (5-7-fold, when compared with untreated cells) in MCF-7 cells/miR-221/ 222-overexpressing cells (Fig. 6B), supporting the role of these miRs in agonistic activity of tamoxifen in the resistant cells. Treatment of OHT R /TIMP3 cells with 17-␤-estradiol or 50 nM tamoxifen resulted in ϳ50% reduction in MAPK activation when compared with OHT R /vector cells (Fig. 6, C  and D, compare lane 2 with lane 5 and lane 3 with lane 6). In summary, growth factor signaling is facilitated in OHT R cells due to reduced TIMP3 level that, in turn, is regulated by miR-221/222 and -181b in these cells.

EGF-induced AKT Phosphorylation Is Inversely Modulated by miR-221/222 and TIMP3 in Breast Cancer Cells-Constitu-
tive activation of PKB/AKT is a hallmark of tamoxifen-resistant cells that leads to aggressive cell proliferation. Growth factors such as EGF can significantly increase Ser-473 phosphorylation of AKT in the resistant cells (25). We observed increase in basal (1.9-fold) and EGF-induced (0.1 and 1 ng/ml) AKT phosphorylation (Ser-473) in the OHT R cells (1.9-and 3.5-fold, respectively) when compared with the MCF-7 cells (Fig. 7A). The basal phosphoAKT level was also higher (1.8-fold) in MCF-7

FIGURE 5. EGF-induced signaling is curtailed in TIMP3-overexpressing OHT R cells or enhanced in miR-221/222-overexpressing MCF-7 cells.
A, analysis of ADAM10 and ADAM17 expression in MCF-7 and OHT R cells by quantitative RT-PCR and Western blotting. B, MCF-7 and OHT R cells were treated with the indicated concentrations of EGF for 15 min. PhosphoMAPK (p-p42/44-MAPK) and total MAPK were analyzed in whole cell extracts by Western blotting. C, OHT R -FLAG and OHT R -TIMP3 cells were treated with EGF as indicated and analyzed for phosphoMAPK. D, panel I, MCF-7-Vec and MCF-7-shTIMP3 were treated with EGF and analyzed for MAPK phosphorylation. Panel ii, TIMP3 expression was determined by quantitative RT-PCR. E, MCF-7-miR-221/222 and MCF-7 vector cells were treated with the indicated EGF concentrations and analyzed for MAPK phosphorylation. miR-221/222 expression is shown in the bar diagram. F, OHT R cells transiently transfected with anti-miR-181b or control anti-miR were treated with EGF and analyzed for MAPK phosphorylation. miR-181b expression is shown in the bar diagram. In panels B-F, band intensity was quantified using ImageJ software, and phosphoMAPK was normalized to total MAPK. Numbers indicate -fold change in phosphoMAPK when compared with the untreated cells.
ADAM17 and ADAM10 Are Essential for Growth and Migration of Tamoxifen-resistant Cells-To test our hypothesis that ADAM17 and ADAM10 are required for growth factor signaling and proliferation of tamoxifen-resistant cells, ADAM17 and ADAM10 were depleted from OHT R cells using siRNA. A 35 and 50% reduction in ADAM10 and ADAM17, respectively, at the protein level was achieved upon siRNA transfection in OHT R cells (Fig. 8A). A 30% decrease in cell proliferation was observed in both cells when compared with the scramble siRNA-transfected controls (Fig. 8B). Further growth inhibition was not observed in the presence of externally added tamoxifen, demonstrating that these metalloproteases are important for maintenance of the resistant cells. The Transwell migration of ADAM17-depleted OHT R cells was inhibited by 64% (p Ͻ 0.0001), and migration of ADAM10depleted cells was inhibited by 38% (p Ͻ 0.001) (Fig. 8C). Migration of ADAM17-depleted cells was further reduced by 21 and 38% in the presence of 1.0 and 2.0 M tamoxifen, respectively, when compared with untreated control. Similar inhibition was also observed with ADAM10-depleted cells (22% at 1.0 M tamoxifen and 46% at 2.0 M tamoxifen). The control siRNA-treated cells demonstrated a 15 and 31% decrease in migration when treated with 1.0 and 2.0 M tamoxifen, respectively. Wound healing assay demonstrated ϳ20% inhibition in healing of ADAM17-and ADAM10-depleted cells in the absence of tamoxifen, whereas an ϳ50% inhibition was observed in the presence of 2.0 M tamoxifen when compared with control siRNA-transfected cells (Fig. 8D).
We further tested growth factor signaling in these metalloprotease-depleted cells by assessing AKT and MAPK phosphorylation levels. A marked decrease in AKT phosphorylation  (60%, p Ͻ 0.01) was observed in ADAM17-depleted OHT R cells when compared with the scramble siRNA-transfected cells when treated with 1.0 ng/ml EGF (Fig. 8E), whereas MAPK phosphorylation was reduced by ϳ30% (p Ͻ 0.05) (Fig. 8E). On the contrary, both AKT and MAPK phosphorylation decreased slightly (ϳ12%) in ADAM10-depleted cells upon EGF treatment when compared with control siRNA-treated cells. These data suggest that the metalloproteases play a key role in growth factor signaling and survival of the tamoxifen-resistant cells, where ADAM17 have a predominant effect on growth factor signaling.

DISCUSSION
Although endocrine therapy that blocks the ER pathway is a very effective treatment for ER-positive breast cancers, the response has not been uniform due to de novo or acquired resistance to this therapy. Extensive studies led to identification of several alternative growth factor signaling pathways that are aberrantly activated in the resistant tumors (for review, see Ref. 5) as well as the role of macroautophagy in protection against tamoxifen-induced cell death and developing anti-estrogen resistance (26). Recently, we observed deregulation of several miRs in tamoxifen-resistant breast cancer, specifically significant elevation in the expression of miR-221, -222, and -181b (9). Here, we have demonstrated increased sensitivity of the tamoxifen-resistant xenografts in mice to the drug upon combined treatment with the corresponding anti-miRs. In breast cancer cell lines and in primary human breast tumors, these three miRs were found to target and regulate TIMP3. In addition, alteration in TIMP3 level in breast cancer cells could modulate metalloprotease activity and mitogenic signaling, thereby contributing to tamoxifen sensitivity both in vitro and in vivo. The increased expression of ADAM17 and ADAM10 in OHT R cells coupled with the marked inhibition of cell growth and migration of cells depleted of these proteins demonstrate the dependence of these cells on the metalloproteases for survival. These data also suggest that these proteins, particularly ADAM17, play a predominant role in growth factor signaling in OHT R cells, and increased expression combined with miR-mediated TIMP3 suppression facilitates growth of the resistant cells. Based on these observations, we conclude that the suppression of miR expression by anti-miRs could sensitize resistant tumors to tamoxifen. A recent study demonstrated that unconjugated tiny locked nucleic acid oligonucleotides could be taken up by mouse mammary tumors following systemic delivery that led to long term miR silencing (27). A similar approach could further establish the beneficial role of knocking down miR-221/222/181 in sensitizing resistant tumors to tamoxifen. miR-221 and -222 are closely located on human chromosome X and appear to be transcribed as a single primary transcript. A recent in vitro study demonstrated that ER␣ suppresses miR-221/-222 levels through the recruitment of nuclear receptor corepressor (NCoR) and silencing mediator of retinoic acid and thyroid hormone receptor (SMRT) (28). It can be speculated that inhibition of ER␣ by anti-estrogenic compounds could release the suppression and increase miR-221/222 expression upon prolonged exposure to the drugs. Similarly, the increase in miR-181b expression in MCF10A cells overexpressing Src was attributed to STAT3 activation (29). However, in hepatocellular carcinoma, the TGF␤ pathway was found to up-regulate miR-181b expression (12).
We were the first to demonstrate the role of miR-221/222 in conferring resistance to tamoxifen in breast cancer (9). Subsequently, up-regulation of miR-221 and -222 has been implicated in resistance to drugs such as fulvestrant (30) and cisplatin (31) in breast cancer, castration-resistant prostate cancer (32), TNF-related apoptosis-inducing ligand (TRAIL)-resistant non-small cell lung cancer cells (33), and radiation-resistant gastric carcinoma cells (34). The ability of miR-221/222 to confer resistance has, however, been attributed mostly to targeting p27/Kip1 in breast cancer (9,10) and non-small cell lung cancer, ER␣ in breast cancer (10), and phosphatase and tensin homolog (PTEN) in gastric cancer (34). Here, we have shown for the first time that miR-221 and -222 can confer resistance to tamoxifen in breast cancers by regulating TIMP3 levels. Further, unlike miR-221/222, up-regulation of miR-181 family has not been frequently correlated with drug resistance. Interestingly, miR-181b expression is significantly elevated in hepatocellular carcinoma (12), breast cancer (35), and pancreatic cancer (36). This study has also opened up the possibility that increased levels of miR-181 in different cancer types could contribute to resistance against other potent anticancer drugs by targeting TIMP3.
Although a close correlation exists among high levels of TIMP3 mRNAs, success of adjuvant endocrine therapy (37,38), and distant metastasis-free survival (39), the mechanism underlying this observation has not been studied in detail. In this study, we provided strong evidence for miR-mediated regulation of TIMP3 level and offered mechanism of its action through inhibition of metalloproteases in facilitating growth of the resistant cells. In summary, we have identified miRs and their targets that contribute to tamoxifen resistance in cell culture models, mouse xenograft models, as well as in primary breast tumors. Identification of the molecular signature of a select panel of deregulated microRNAs that can potentially target multiple genes and modulate several biological pathways could lead to a novel regulatory mechanism of drug resistance and open up alternate strategies for treatment of patients resistant to this important class of drugs.