IKKϵ Phosphorylation of Estrogen Receptor α Ser-167 and Contribution to Tamoxifen Resistance in Breast Cancer*

IKKϵ has recently been identified as a breast cancer oncogene. Elevated levels of IKKϵ are associated with cell survival and growth. Here, we show that IKKϵ interacts with and phosphorylates estrogen receptor α (ERα) on serine 167 in vitro and in vivo. As a result, IKKϵ induces ERα transactivation activity and enhances ERα binding to DNA. Cyclin D1, a major target of ERα, is transcriptionally up-regulated by IKKϵ in a phospho-ERα-Ser-167-dependent manner. Further, overexpression of IKKϵ induces tamoxifen resistance, whereas knockdown of IKKϵ sensitizes cells to tamoxifen-induced cell death. These data suggest that ERα is a bona fide substrate of IKKϵ and IKKϵ plays an important role in tamoxifen resistance. Thus, IKKϵ represents a critical therapeutic target in breast cancer.

Breast carcinoma is the most common cancer among women in developed countries, and about 70% of these tumors express estrogen receptor (ER) 3 ␣. ER␣-positive tumors are associated with a well differentiated phenotype and have a better prognosis than ER␣-negative tumors (1). The major reason is that ER␣-positive tumors initially respond well to anti-estrogen agents such as tamoxifen. However, a significant portion of ER␣-positive tumors eventually become resistant to anti-estrogen therapy (2)(3)(4). The underlying molecular mechanisms have been linked to loss of ER␣ expression caused by DNA hypermethylation and deregulation of certain microRNAs (2,5). It is noted that a number of patients with tamoxifen-resistant breast cancer remain ER␣-positive (2,4). Growing evidence shows that ER␣ membrane-initiated steroid signaling activities and cross-talk with growth factor signal transduction pathways may contribute to tamoxifen resistance. Activation of ER␣ outside the nucleus leads to the activation of surface tyrosine kinase receptors (e.g. IGF-IR, EGFR, and HER2) as well as interaction with cellular kinases and adaptor molecules (e.g. c-Src or the p85␣ regulatory subunit of phosphatidylinositol-3-OH kinase), which in turn lead to the activation of mitogen-activated protein kinase (MAPK) and AKT pathways known to enhance cell proliferation and survival (6 -8). The activation of these signaling pathways causes phosphorylation of ER␣ and/or its co-activators and co-repressors, thereby increasing nuclear ER␣ activity (4,9).
IKK⑀ is a member of IB kinase (IKK) family and activates NFB through phosphorylation and degradation of IB (10 -12). It is primarily activated by interferon (IFN), PMA, and activation of the T-cell receptor (13,14). Using complementary genetic approaches, Boehm et al. (15,16) identified IKK⑀ as a breast cancer oncogene, which is frequently amplified/overexpressed in human breast cancer. Recently, we have shown frequent alterations of IKK⑀ in ovarian cancer (17). In this report, we demonstrate that IKK⑀ phosphorylates Ser-167 of ER␣. IKK⑀ induction of ER␣ transactivation and cyclin D1 expression is dependent on phosphorylation of Ser-167. Knockdown of IKK⑀ sensitizes breast cancer cells to tamoxifen-induced cell death and growth arrest, whereas ectopic expression of IKK⑀ exhibits the opposite effect. These data indicate that IKK⑀ plays an important role in regulation of ER␣ activity and tamoxifen resistance, and thus could be a therapeutic target in breast cancer.

EXPERIMENTAL PROCEDURES
Cells and Reagents-All breast cancer cell lines used in this study were purchased from ATCC, maintained in DMEM with 10% fetal bovine serum. MCF10A was maintained in MACo5 with 10% FBS. Cells, when used for ER␣ phosphorylation, tamoxifen treatment, and ER␣ transcriptional activity experiments, were cultured in phenol red-free DMEM with charcoalstripped serum. The IKK⑀ antibody (I-4907) was obtained from Sigma. CCND1 (sc-8396), ER␣ (sc-543), and p-ER␣-Ser-167 (sc-101676) antibodies were purchased from Santa Cruz Biotechnology. Recombinant ER␣ and IKK⑀ were obtained from Thermo Scientific and Cell Signaling, respectively.
In Vitro Kinase Assay, in Vivo [ 32 P]P i Cell Labeling, and Mass Spectrometry-In vitro IKK⑀ kinase assay was performed as described previously (11,17). Briefly, the reaction was carried out in the presence of 10 Ci of [␥-32 P]ATP (NEN) and 3 M cold ATP in 30 l of buffer containing 20 mM Hepes (pH 7.6), 10 mM MgCl 2 , 50 mM NaCl, 0.1 mM sodium vanadate, 20 mM ␤-glycerolphosphate, and 1 mM DTT using GST-ER␣ as substrate. After incubation at 30°C for 30 min, the reaction was stopped by adding protein-loading buffer; the proteins were then separated in SDS-PAGE gels. Each experiment was repeated three times. The relative amounts of incorporated radioactivity were determined by autoradiography and quantitated with a PhosphorImager (Molecular Dynamics).
For in vivo labeling, MCF10A cells were transfected with ER␣/Myc-IKK⑀. After serum starvation overnight, cells were labeled with [ 32 P]P i (0.5 mCi/ml) in phenol red-free MEM without phosphate for 4 h. ER␣ was immunoprecipitated and separated by SDS-PAGE prior to transferring to membrane for detection and quantification of phosphorylated ER␣.
Mass spectrometry was used to map IKK⑀ phosphorylation site of ER␣. After separation of in vitro IKK⑀/ER␣ kinase reactions in SDS-PAGE, ER␣ bands were excised and washed. Proteins were reduced with Tris(carboxyethyl)phosphine and alkylated with iodoacetamide. Samples were digested overnight with modified sequencing grade trypsin (Promega, Madison, WI). Peptides were extracted and concentrated under vacuum centrifugation. A nanoflow liquid chromatograph (U3000, LC Packings/Dionex, Sunnyvale, CA) coupled to an electrospray hybrid ion trap mass spectrometer (LTQ Orbitrap, Thermo, San Jose, CA) was used for tandem mass spectrometry peptide sequencing. Peptides were separated on a C18 reverse phase column (LC Packings C18Pepmap, 75 m ID ϫ 15 cm) using a 40-min gradient from 5% B to 50% B (A: 2% acetonitrile/0.1% formic acid; B: 90% acetonitrile/0.1% formic acid). The flow rate on the analytical column was 300 nl/min. Five tandem mass spectra were acquired for each MS scan using 60 s exclusion for previously sampled peptide peaks. Sequences were assigned using Mascot data base searches. Oxidized methionine, deamidation, carbamidomethyl cysteine, and phosphorylated serine, threonine, and tyrosine were selected as variable modifications, and as many as 2 missed cleavages were allowed. Assignments were manually verified by inspection of the tandem mass spectra and coalesced into Scaffold.
Western Blot, Co-immunoprecipitation (co-IP), Immunofluorescence, Luciferase Reporter Assay, Chromatin Immunoprecipitation (ChIP), and RT-PCR-Western blots, co-IP, immunofluorescence, and luciferase reporter assays were performed as previously described (18). ChIP assay was performed using a kit (Upstate) following the manufacturer's instruction. The MCF10A cells were transfected with indicated plasmids and labeled with [ 32 P]orthophosphate. Following immunoprecipitation with anti-ER␣ antibody, the immunoprecipitates were separated by SDS-PAGE, transferred, and then exposed (top). Expression of transfected plasmids is shown in panels 2 and 3. D, IKK⑀ interacts with ER␣. MCF10A cells were transfected with Myc-IKK⑀ and GFP-ER␣. After a 48-h incubation, cells were lysed and immunoprecipitated with anti-GFP antibody. The immunoprecipitates were immunoblotted with anti-Myc antibody (top panel) and vice versa (panel 2). Panels 3 and 4 show expression of transfected plasmids. Actin was used as a loading control (bottom). E and F, endogenous ER␣ binds to IKK⑀. T47D cells, expressing ER␣ and IKK⑀ (panels 3 and 4), were immunoprecipitated with anti-ER␣ and detected with anti-IKK⑀ antibody (E) and vice versa (F; panels 1 and 2). IgG was used as a control for coimmunoprecipitation, and actin is a loading control (bottom).
Cell Survival, Focus Formation, and Apoptosis Assays-Cell survival was examined by MTT assay as previously described (17). Briefly, cells were diverted into 96-well plates with 1 ϫ 10 4 cells/well. Following a 24-h incubation, cells were treated with different concentrations of tamoxifen for 72 h and then assayed for total cell viability (MTT assay). For focus formation, cells were released from culture flasks with Trypsin-EDTA (Invitrogen) and resuspended at a concentration of 100 cells/ml and placed in a 6-well plate. The cells were allowed to grow for 7 days in the presence or the absence of tamoxifen at which point they were washed with phosphate-buffered saline (pH 7.4) and stained with 1 mg/ml p-iodonitrotetrazolium violet in DMEM (21). Apoptosis was analyzed with Annexin V-fluorescein isothiocyanate apoptosis detection kit following the manufacturer's instruction (BD Pharmingen).
Statistical Analysis-For luciferase activity, cell survival, and apoptosis, the experiments were repeated at least three times in triplicate. The data are represented by means Ϯ S.D. Differences between control and testing cells were evaluated by Student's t test.

IKK⑀ Phosphorylates and Interacts with ER␣-Previous studies
have shown that phosphorylation of ER␣ by serine/threonine protein kinases (e.g. Erk, casein kinase II, pp90 rsk1 , Akt, and IKK␣) induces tamoxifen resistance (9,(22)(23)(24)(25)(26). Frequent alterations of IKK⑀ kinase in breast cancer (15,16) prompted us to examine whether IKK⑀ regulates ER␣. As an initial step, we examined IKK⑀ expression in a panel of breast cancer cell lines. Consistent with a previous report (15), expression of IKK⑀ has no correlation with ER␣ status in breast cancer cell lines (Fig. 1A). Further, in vitro kinase assay was performed by incubation of recombinant IKK⑀ and full-length human recombinant ER␣. Fig. 1B

p-ERα
separated on SDS-PAGE and exposed on the film. Fig. 1C displays that ectopic expression of IKK⑀ induces ER␣ phosphorylation in vivo.
We next examined whether IKK⑀ interacts with ER␣. The coimmunoprecipitation was performed in MCF10A cells, which had been transfected with GFP-ER␣ and Myc-IKK⑀. ER␣ was readily detected in Myc-IKK⑀ immunoprecipitates and vice versa (Fig. 1D). Furthermore, endogenous ER␣ and IKK⑀ were able to form a complex in T47D cells (Fig. 1, E and  F), which express ER␣ as well as high levels of IKK⑀ (Fig. 1A). These findings indicate that ER␣ is a substrate of IKK⑀.
IKK⑀ Phosphorylates ER␣ at Serine 167-To define the site in ER␣ that is phosphorylated by IKK⑀, in vitro IKK⑀ kinase assay was carried out using GST fusion proteins containing different portions of ER␣ as substrates. A potential phosphorylation site(s) was mapped to the amino acid 90 -324 region of ER␣ ( Fig. 2A). Mass spectrometry analysis revealed serine 167 (Ser-167) as a possible phosphorylation site (Fig. 2B). To verify if ER␣-Ser-167 is phosphorylated by IKK⑀, we created non-phosphorylatable ER␣-S167A by converting Ser-167 into alanine and performed in vitro IKK⑀ kinase assay using the wild-type and S167A mutant GST-ER␣ as substrates. Fig. 2C shows that wild-type GST-ER␣ but not GST-ER␣-S167A is phosphorylated by IKK⑀. Furthermore, in vivo [ 32 P]orthophosphate labeling was performed in MCF10A cells, which were transfected with either ER␣ or ER␣-S167A together with and with-out Myc-IKK⑀. We observed IKK⑀ phosphorylation of wild-type but not S167A mutant ER␣ (Fig. 2D).
In addition, immunoblotting was performed with anti-pER␣-Ser-167 antibody in MCF10A cells transfected with ER␣ together with wildtype and dominant-negative (DN) IKK⑀. As shown in Fig. 2E, IKK⑀, but not DN-IKK⑀, phosphorylates ER␣-Ser-167. To further demonstrate that ER␣-Ser-167 is phosphorylated by IKK⑀, we ectopically expressed IKK⑀ in MCF7 cells and then immunoblotted and immunostained with anti-pER␣-Ser-167 antibody. Fig. 2, F and G show elevated pER␣-Ser-167 levels in IKK⑀transfected cells when compared with pCMV vector-treated cells. Moreover, knockdown of IKK⑀ decreases pER␣-Ser-167 in T47D cells (Fig. 6A), which express both ER␣ and high level of IKK⑀ (Fig. 1A). Based on these results, we conclude that IKK⑀ phosphorylates Ser-167 of ER␣ in vitro and in vivo.
IKK⑀ Induces ER␣ Transactivation Activity through Phosphorylation of Ser-167-Previous studies have demonstrated that the ER␣-Ser-167 is one of major phosphorylation sites to activate ER␣ (6,26). Therefore, we further assessed whether IKK⑀ induces ER␣ transactivation activity and, if present, whether the activation depends on phosphorylation of Ser-167. Reporter assay was performed in MCF7 cells transfected with estrogen response element promoter-luciferase (ERE-Luc) reporter together with and without IKK⑀. Fig. 3A shows that ERE-Luc activity was induced by ectopic expression of wild type but not DN-IKK⑀, implying that IKK⑀ activation of ER␣ is kinasedependent. Further, we transfected MCF10A cells with wildtype ER␣ and non-phosphorylatable ER␣-S167A together with and without IKK⑀. Expression of IKK⑀ alone did not induce ERE-Luc activity in MCF10A cells. However, co-expression of ER␣/IKK⑀ but not ER␣-S167A/IKK⑀ significantly stimulated the reporter activity (Fig. 3B), indicating that IKK⑀ activates ER␣ through phosphorylation of Ser-167. (27), we next investigated whether cyclin D1 is FIGURE 2. IKK⑀ phosphorylates ER␣-Ser-167. A, IKK⑀ phosphorylates N-terminal region (amino acids 90 -324) of ER␣ In vitro IKK⑀ kinase assay was performed using different truncated GST-ER␣ fusion proteins as substrates (top). Coomassie Blue staining shows GST-ER␣ fusion proteins used in the kinase assay (bottom). B, tandem mass spectrum analysis. The inset shows the peptide measurement in the survey scan. Arrows indicate the fragment ions that confirm the location of the phosphorylation site as serine 167. The m/z value of the y13 ion reflects that the phosphorylation is not present in that fragment; the m/z values b3 and y14 (2ϩ) indicate the presence of the phosphorylation. The Mascot score for the phosphopeptide was 59 (cutoff score is usually 20 -25 for reliable data); in addition, peptides with one or two methionine oxidations were also observed with Mascot scores of 37 and 51, respectively. The mass measurements of the oxidized forms were accurate to 4.2 ppm and 2.6 ppm. C, IKK⑀ phosphorylates ER␣-Ser-167 in vitro. In vitro IKK⑀ kinase assay was performed using GST-WT-and -S167A-ER␣ as substrates (top). The bottom panel is Coomassie Blue staining. D, IKK⑀ phosphorylates ER␣-Ser-167 in vivo. MCF10A cells were transfected with wild-type and S167A mutant ER␣ together with and without IKK⑀. In vivo labeling was performed as described in Fig. 1C (top).  regulated by IKK⑀, and, if this is the case, whether IKK⑀-regulated cyclin D1 depends on pER␣-Ser-167. MCF7 (ER␣-positive) and MCF10A (ER␣-negative) cells were co-transfected with IKK⑀/ER␣ or IKK⑀/ER␣-S167A. As shown in Fig. 4, A and  B, ectopic expression of IKK⑀ up-regulates cyclin D1 more significantly in MCF7 than MCF10A cells. Further, co-expression  1). GAPDH was used as control (panel 2). Middle panels are immunoblotting analysis with indicated antibodies, and the CCND1 protein levels were quantified. Bottom panels show the quantification of CCND1 mRNA. C, cyclin D1 promoter is activated by IKK⑀ through pER␣-Ser-167. MCF710A cells were transfected with cyclin D1-Luc, ␤-galactosidase, and other indicated plasmids. After 48 h of incubation, the promoter activity was determined as described in Fig. 3. D, IKK⑀ enhances ER␣ binding to cyclin D1 promoter. MCF10A cells were transfected with the indicated plasmids. The ChIP assay was performed as described under "Experimental Procedures." Anti-ER␣ antibody was used for chromatin immunoprecipitation. IgG served as a negative control. The DNA prior to the immunoprecipitation was used as positive controls (input). of IKK⑀ with ER␣, but not ER␣-S167A, induces CCND1 in MCF10A cells (Fig. 4B). However, ectopic expression of IKK⑀ alone also induced cyclin D1 mRNA and protein levels (lane 2 of Fig. 4B), suggesting that IKK⑀ regulation of cyclin D1 could also be mediated by other pathway(s).
IKK⑀ Plays an Important Role in Tamoxifen Sensitivity-Having demonstrated IKK⑀ phosphorylation of ER␣ on Ser-167, a site which has been shown to be involved in tamoxifen resistance (6,18,24), we subsequently examined the role of IKK⑀ in tamoxifen sensitivity. Of ER␣positive cell lines, MCF7 expressed relatively lower levels of IKK⑀ than T47D and MDA-MB-361 (Fig. 1A). Thus, we ectopically expressed IKK⑀ in MCF7 (Fig. 5A). Following tamoxifen treatment, MTT and Annexin V/FACS assays revealed that total cell death and apoptosis were significantly reduced in IKK⑀transfected MCF7 cells when compared with the cells transfected with vector alone (Fig. 5, B and C). Moreover, tamoxifen-induced inhibition of focus formation was reduced in MCF7 cells overexpressing IKK⑀ (Fig. 5D).
To further demonstrate the role of IKK⑀ in tamoxifen sensitivity, we knocked down IKK⑀ in T47D cells by either infection of lentiviruses/ shRNA-IKK⑀ or transfection of siRNA-IKK⑀. The cells infected/ transfected with shRNA-GFP and scramble/mismatched siRNA-IKK⑀ were used as controls. Western blot analysis shows that IKK⑀ was efficiently knocked down by both shRNA-IKK⑀ and siRNA-IKK⑀ as compared with controls (Fig. 6A). Accordingly, pER␣-Ser-167 levels were reduced by depletion of IKK⑀ (Fig. 6A). Further, the knockdown of IKK⑀ by shRNA or siRNA increased tamoxifen-induced total cell death and apoptosis at 10 M (Fig. 6, B and C; p Ͻ 0.05). Tamoxifen-inhibited focus formation was also enhanced by knockdown of IKK⑀ (Fig.  6D). In addition, we noticed that parental T47D cell is less sensitive to tamoxifen than MCF7 (Figs. 5, B-D versus 6. B-D) that could be attributed to elevated level of IKK⑀ in T47D (Fig. 1A).
Furthermore, we have selected 18 ER␣-positive primary breast cancers, 8 of which were tamoxifen-resistant, and examined IKK⑀ and pER␣-Ser-167 levels. Immunoblotting analysis revealed high levels of IKK⑀ and pER␣-Ser-167 in 5 tamoxifenresistant tumors but not the rest of the specimens (Fig. 6E and data not shown), further suggesting that IKK⑀ plays an important role in phosphorylation of ER␣ and tamoxifen resistance.

DISCUSSION
In this study, we demonstrate for the first time the role of IKK⑀ in ER␣ activation and tamoxifen sensitivity in breast cancer. Whereas IKK⑀ is a member of IKK family, it only shares 30% amino acid identity to IKK␣ and IKK␤ in their kinase domains. IKK⑀ has been shown to activate NFB; however, it differs from the IKK␣-IKK␤-IKK␥ complex. In response to proinflammatory stimuli, IKK␣ and IKK␤ phosphorylate Ser-32/36 of IB leading to NFB activation (28). Whereas, IKK⑀ only phosphorylates IB Ser-36 to mediate NFB activation induced by PMA and the activated T-cell receptor (10,12). In addition, IKK⑀ phosphorylates interferon response factors 3/7 (IRF3 and IRF7), STAT1, and tumor suppressor CYLD to regulate the type I interferon response and cell transformation, respectively (13,28,29). We have shown in this report that IKK⑀ phosphorylates ER␣-Ser-167 in vitro and in vivo, leading to activation of ER␣ and up-regulation of cyclin D1. In consid-eration of frequent overexpression/activation of IKK⑀ in breast and ovarian cancer as well as its ability to transform HMEC-MEK1 cells (15)(16)(17), these findings suggest that IKK⑀ exerts its cellular function through regulation of not only canonical NFB pathway but also other important cascades.
Tamoxifen has been the mainstay of hormonal therapy in the treatment of breast cancer. Despite its long-term benefit, some tumors eventually become resistant to tamoxifen and exhibit an estrogen-independent phenotype even though ER␣ is maintained in a subset of cases. One of the underlying molecular FIGURE 6. Knockdown of IKK⑀ sensitizes cells to tamoxifen-induced cell death. A, knockdown of IKK⑀. T47D cells were infected and transfected with lentiviruse/shRNA-IKK⑀ and siRNA-IKK⑀, respectively. As controls, the cells were also treated with shRNA-GFP and scramble/mismatched siRNA-IKK⑀. The methods and nucleotide sequence were described under "Experimental Procedures." The knockdown of IKK⑀ and its effect on pER-Ser-167 were assessed by immunoblotting with the indicated antibodies. B-D, depletion of IKK⑀ enhances the tamoxifen effect on total cell death, apoptosis, and focus formation. IKK⑀ knockdown and control cells were treated with the indicated concentration of tamoxifen and subsequently assayed for total cell death, apoptosis, and focus formation as described in the legend to Fig. 5. Asterisks indicate p Ͻ 0.05. E, IKK⑀ expression levels correlate with pER␣-Ser-167 and tamoxifen sensitivity in ER␣-positive breast cancers. Representative ER␣-positive breast cancer specimens were immunoblotted with indicated antibodies. Tumors expressing high levels of IKK⑀ increase pER␣-Ser-167 and are insensitive to tamoxifen indicated by arrows. mechanisms is the phosphorylation of ER␣ by protein kinases, which include serine/threonine kinases Erk1/2 (Ser-118, Ref. 22 . These phosphorylation events induce ER␣ activity and result in tamoxifen resistance. The ER␣ has an N-terminal domain with a hormone-independent transcriptional activation function (AF-1, amino acids 1-180), a central DNAbinding domain (amino acids 181-263) and a C-terminal ligand-binding domain with a hormone-dependent transcriptional activation function (AF-2; amino acids 302-552, Ref. 32,33). Because Ser-167 is located in the AF-1 region, activation of ER␣ by phosphorylation of Ser-167 is ligandindependent. We demonstrated that expression of IKK⑀ is sufficient to induce ERE-Luc activity and cyclin D1 expression in ER␣-positive cells.
Our data also show that expression of IKK⑀ alone in ER␣negative cells induces cyclin D1 expression and its promoter activity (Fig. 4, B and C), suggesting that IKK⑀ regulates cyclin D1 not only through phosphorylation of ER␣ but also through other molecules. It has been well documented that the NFB pathway transcriptionally regulates cyclin D1 (34 -36). Thus, IKK⑀-induced cyclin D1, in addition to ER␣, is also mediated through activation of the NFB pathway.
It has been shown that the PI3K inhibitor LY29002 and mTOR inhibitor rapamycin efficiently inhibit Akt-and S6K1induced ER␣ activation and cell proliferation, respectively (18,24). Therefore, development of small molecule inhibitors of IKK⑀ could have great potential to surmount IKK⑀-associated tamoxifen resistance. Previous studies have demonstrated that inhibition of IKK⑀ by either shRNA or dominant-negative IKK⑀-K38A reduces cell growth, survival, and invasion and that ectopic expression of IKK⑀ induces cell survival and proliferation (15,17). These findings indicate that IKK⑀ could be a critical therapeutic target. Further investigations are required to identify small molecule inhibitor(s) of IKK⑀ for anticancer drug discovery and elucidate IKK⑀ oncogenic activity in transgenic mouse models.