relb nf-b represses estrogen receptor expression via induction

MOLECULAR AND CELLULAR BIOLOGY, July 2009, p. 3832–3844 Vol. 29, No. 140270-7306/09/$08.00�0 doi:10.1128/MCB.00032-09Copyright © 2009, American Society for Microbiology. All Rights Reserved.

RelB NF-�B Represses Estrogen Receptor � Expression via Inductionof the Zinc Finger Protein Blimp1�‡

Xiaobo Wang,1†§ Karine Belguise,1†¶ Christine F. O’Neill,1� Nuria Sanchez-Morgan,1Mathilde Romagnoli,1 Sean F. Eddy,1# Nora D. Mineva,1 Ziyang Yu,1 Chengyin Min,1

Vickery Trinkaus-Randall,2 Dany Chalbos,3 and Gail E. Sonenshein1*Department of Biochemistry and Women’s Health Interdisciplinary Research Center1 and Departments of Ophthalmology and

Biochemistry,2 Boston University School of Medicine, Boston, Massachusetts 02118, and INSERM, U896,Universite Montpellier 1, CRLC Val d’Aurelle Paul Lamarque, Montpellier, France3

Received 8 January 2009/Returned for modification 9 March 2009/Accepted 30 April 2009

Aberrant constitutive expression of NF-�B subunits, reported in more than 90% of breast cancers and multipleother malignancies, plays pivotal roles in tumorigenesis. Higher RelB subunit expression was demonstrated inestrogen receptor alpha (ER�)-negative breast cancers versus ER�-positive ones, due in part to repression of RelBsynthesis by ER� signaling. Notably, RelB promoted a more invasive phenotype in ER�-negative cancers viainduction of the BCL2 gene. We report here that RelB reciprocally inhibits ER� synthesis in breast cancer cells,which contributes to a more migratory phenotype. Specifically, RelB is shown for the first time to induce expressionof the zinc finger repressor protein Blimp1 (B-lymphocyte-induced maturation protein), the critical mediator of B-and T-cell development, which is transcribed from the PRDM1 gene. Blimp1 protein repressed ER� (ESR1) genetranscription. Commensurately higher Blimp1/PRDM1 expression was detected in ER�-negative breast cancer cellsand primary breast tumors. Induction of PRDM1 gene expression was mediated by interaction of Bcl-2, localized inthe mitochondria, with Ras. Thus, the induction of Blimp1 represents a novel mechanism whereby the RelB NF-�Bsubunit mediates repression, specifically of ER�, thereby promoting a more migratory phenotype.

NF-�B is a structurally and evolutionary conserved family ofdimeric transcription factors with subunits having an N-termi-nal region of approximately 300 amino acids that shares ho-mology with the v-Rel oncoprotein (17, 44). The conserved Relhomology domain is responsible for DNA binding, dimeriza-tion, nuclear translocation, and interaction with inhibitory pro-teins of NF-�B (I�Bs). Mammals express five NF-�B members,including c-Rel, RelB, RelA (p65), p50, and p52, which canform either homo- or heterodimers. RelB differs from theother members in that it only binds DNA as a heterodimerwith either p52 or p50 and interacts only poorly with theinhibitory protein I�B�. In most untransformed cells, otherthan B lymphocytes, NF-�B complexes are sequestered in thecytoplasm bound to specific I�B proteins.

Aberrant activation of NF-�B has been implicated in thepathogenesis of many carcinomas (37). Constitutive activation

of c-Rel, RelA, p50, and p52 was first detected in breast cancer(9, 30, 45). RelB appeared to have more limited involvementand functioned predominantly in lymphoid organs and theirmalignancies (25, 58). For example, Stoffel et al. found p50/RelB complexes in mucosa-associated lymphoid tissue lym-phoma (46), and RelB complexes were implicated in Notch1-induced T-cell leukemia (53). More recently, RelB has beenimplicated in carcinomas of the breast and prostate. RelB wasthe most frequently detected NF-�B subunit in nuclear prep-arations from advanced prostate cancer tissue and correlateddirectly with Gleason score (26), suggesting an association withprostate cancer progression. We observed elevated nuclearRelB levels in mouse mammary tumors driven by ectopic c-Relexpression in transgenic mice or after carcinogen treatment(14, 40). More recently, we demonstrated that RelB synthesis,which is mediated via synergistic transactivation of the RELBpromoter by p50/RelA NF-�B and c-Jun/Fra-2 AP-1 com-plexes, was selectively active in estrogen receptor � (ER�)-negative versus -positive breast cancers and led to the induc-tion of transcription of the BCL2 gene (56).

While studying the mechanism of the inverse correlationbetween RelB and ER� levels in breast cancer, more recentlywe observed that RelB complexes robustly inhibited ER�(ESR1) gene expression. Since RelB is not by itself inhibitory,we postulated that it controls the expression of an intermediatenegative regulatory factor that represses ER� gene transcrip-tion. TransFac analysis identified putative binding sites for theB lymphocyte-induced maturation protein (Blimp1), which isexpressed from the PRDM1 gene (24) and has been reported tobe regulated by NF-�B, although it is not known whether thiscontrol is exerted directly or indirectly (22). The Zn fingerprotein Blimp1 is a repressor of transcription and has been

* Corresponding author. Mailing address: Boston University School ofMedicine, Department of Biochemistry K615, 72 E. Concord Street, Bos-ton, MA 02118. Phone: (617) 638-4120. Fax: (617) 638-4252. E-mail:[email protected].

† X.W. and K.B. contributed equally to this study.‡ Supplemental material for this article may be found at http://mcb

.asm.org/.§ Present address: Department of Biological Chemistry, Center for

Cell Dynamics, Johns Hopkins School of Medicine, 755 North WolfeStreet, Baltimore, MD 21205.

¶ Present address: INSERM U896, IRCM, Parc Euromedecine-Vald’Aurelle, 34298 Montpellier, France.

� Present address: Center for Molecular Medicine, Maine MedicalCenter Research Institute, 81 Research Drive, Scarborough, ME04074.

# Present address: Genstruct, Inc., One Alewife Center, Cam-bridge, MA 02140.

� Published ahead of print on 11 May 2009.

3832

on February 17, 2018 by guest

http://mcb.asm

.org/D

ownloaded from

http://mcb.asm.org/

shown to function as a master regulator of development ofantibody-secreting B lymphocytes and more recently of T-cellhomeostasis and function (6, 28), and specification of primor-dial germ cells (33, 54). Blimp1 was originally identified as asilencer of IFN-� gene transcription (24). Although the exactmechanism by which repression occurs is not fully understood,Blimp1 possesses DNA-binding activity and can recruit histonedeacetylases, histone methyltransferases, and the corepressorGroucho (19, 39, 63). We demonstrate here for the first timethat Blimp1 functions as a potent repressor of ER� synthesis inbreast cancer cells and is induced by activation of a Bcl-2/Raspathway by RelB.

MATERIALS AND METHODS

Cell culture and treatment conditions. ER�-positive MCF-7, T47D, ZR-75,and BT474 cells and ER�-negative Hs578T and MDA-MB-231 cells were pur-chased from the American Type Culture Collection (ATCC; Manassas, VA) andmaintained in standard culturing medium as recommended by the ATCC. TheNF639 cell line, kindly provided by Philip Leder (Harvard Medical School,Boston MA) was derived from a mammary gland tumor in an ERBB2 transgenicmouse and cultured as described previously (18). Hs578T cell lines containinghuman pRELB-siRNA or control pRELB sense (si-Control) were established asdescribed previously (56) and grown in the presence of 1 �g/ml of puromycin(Sigma, St. Louis, MO). MCF-7 cell lines containing pBABE and pBABE-Bcl-2were established by retroviral infection, as described previously (56), and grownin the presence of 1 �g/ml of puromycin. Hs578T or NF639 cell lines carryingpSM2c-sh-BCL2 (V2HS_111806; Open Biosystems) or scrambled control wereestablished by retroviral infection, as described above. MCF-7 cells containinginducible C4bsrR(TO) or C4bsrR(TO-RelB) were established as described pre-viously (56). Expression of RelB was induced with 1-�g/ml doxycycline treatmentfor 48 h.

Plasmids and transfection analysis. A long promoter B ER� construct (ER�-proB) was cloned via PCR from a full-length ER� promoter (kindly provided byRonald Weigel, University of Iowa College of Medicine, Iowa City) using PfuHotstart Ultra with the following PCR primers (restriction sites are italicized,and the ER� sequence is underlined): KpnI-proB-3489, 5�-CGCTCGGTACCTCCCATGCTCACCAAGCC-3�; and HindIII-proB-1814, 5�-CGCTCAAGCTTTCAAGCCTACCCTGCTGG-3�.

After amplification, the product was digested with KpnI and HindIII andsubcloned into pGL3Basic; sequencing confirmed that no mutations were intro-duced during amplification. The pC4bsrR(TO-RelB) construct was generated bysubcloning the RelB cDNA into the EcoRI restriction site of the doxycyclineinducible pC4bsrR(TO) retroviral vector. The Ras C186S mutant and the ER�expression vector were kindly provided by Mark R. Philips (NYU School ofMedicine, New York, NY) and Pierre Chambon (Strasbourg, France), respec-tively. The Ras-EGFP-C1 vector was generously provided by Wen-Luan WendyHsiao (Hong Kong Baptist University). The Ras CA (RasV12) vector was aspreviously described (20). The human full-length PRDM1 construct in pcDNA3was as reported (39). The TBlimp vector expressing a dominant-negative (dn)-Blimp1 protein, containing the DNA-binding domain of Blimp1 in the Vxy-purovector, and the 7-kb human PRDM1 promoter construct in pGL3 were kindlyprovided by Kathryn Calame (Columbia University, NY) (1, 49). pRELB senseand pRELB-siRNA containing control or si-RELB oligonucleotides in pSIREN-RetroQ vector were described previously (41). The human full-length BCL2constructs in pcDNA3 or pBabe were as reported (42). The BCL-2 mutant L14G,with the indicated point mutation, was kindly provided by Tristram G. Parslow(Emory University School of Medicine, Atlanta, GA). The pRc/CMV-Bcl-2,pRc/CMV-Bcl-acta, and pRc/CMV-Bcl-nt constructs, which were kindly pro-vided by David W. Andrews (McMaster University, Hamilton, Ontario, Canada),were as previously described (64). Briefly, the Bcl-acta (mitochondrion-localizedBcl-2 mutant) and the Bcl-nt (cytoplasm-localized Bcl-2 mutant) were estab-lished by exchanging the Bcl-2 carboxy-terminal insertion sequence for an equiv-alent sequence from Listeria ActA or by deletion of the Bcl-2 carboxy-terminalinsertion sequence, respectively. The simian virus 40 �-galactosidase (SV40-�-Gal) reporter vector was as previously reported (57). The estrogen responseelement (ERE)-TK luciferase construct was as reported elsewhere (18). Fortransfection into six-well or P100 plates, 3 or 10 �g, respectively, of total DNAwas used. For transient transfection, cultures were incubated for 48 h in thepresence of DNA and Fugene6 (Roche Diagnostics Co., Indianapolis, IN) or

Geneporter2 (Gene Therapy Systems, Inc., San Diego, CA) transfection reagent.All transient-transfection reporter assays were performed, in triplicate, a mini-mum of three times. Luciferase assays were performed as described previously(40, 45). Cotransfection of the SV40-�-Gal expression vector was used to nor-malize for transfection efficiency, as described previously (40). PRDM1 smallinterfering RNA (siRNA) duplex sequences have been previously described (61).Duplexes (0.8 nM final) were introduced in cells using Lipofectamine 2000(Invitrogen) according to the manufacturer’s protocol.

Isolation of membrane and cytosolic fractions. Membrane and cytosolic pro-teins were extracted by using a Mem-PER eukaryotic membrane protein extrac-tion kit (Pierce, catalog no. 89826) according to the manufacturer’s protocol.

Immunoblot analysis, antibodies, and immunoprecipitation. Whole-cell ex-tracts (WCEs) were prepared in radioimmunoprecipitation assay buffer (10 mMTris [pH 7.5], 150 mM NaCl, 10 mM EDTA, 1% NP-40, 0.1% sodium dodecylsulfate [SDS], 1% sodium sarcosyl, 0.2 mM phenylmethylsulfonyl fluoride, 10�g/ml of leupeptin, 1 mM dithiothreitol). Nuclear extracts were prepared aspreviously described (18). Immunoblotting was performed as we have describedin an earlier study (18). Antibodies to RelB (sc-226), Bcl-2 (sc-492), Ras (sc-35),or lamin B (sc-6217) were purchased from Santa Cruz Biotechnology (SantaCruz, CA). Antibody to ER� was purchased from NeoMarker (Fremont, CA).Antibodies to E-cadherin, �-catenin, and Ras (catalog no. 610001) used inimmunoprecipitation analyses were purchased from BD Transduction Labora-tories (Franklin Lakes, NJ), and antibody to �-actin (AC-15) was obtained fromSigma. Antibodies to mouse Blimp1 (NB600-235) and human Blimp1 (ab13700)were purchased from Novus Biologicals and Abcam, respectively. For immuno-precipitation, WCEs were prepared in lysis buffer (50 mM Tris [pH 7.5], 150 mMNaCl, 5 mM EGTA, 1% NP-40, 0.5% deoxycholate, 0.1% SDS, 10 mM NaF, 4mM Na3VO4, 0.5 mM phenylmethylsulfonyl fluoride, 10 �g/ml of leupeptin, 10mM �-glycerophosphate) and precleared WCEs (500 �g) were incubated over-night at 4°C with antibodies against Bcl-2 or Ras or the corresponding normalimmunoglobulin G (IgG) in NETN buffer (20 mM Tris [pH 7.5], 100 mM NaCl,0.5% NP-40, and 1 mM EDTA plus protease and phosphatase inhibitors). Im-munoprecipitates were collected using protein A-Sepharose beads, washed fourtimes in NETN buffer, and then subjected to immunoblotting.

EMSA. Nuclear extracts were prepared, and samples (5 �g) were subjected tobinding and electrophoretic mobility shift assay (EMSA) as described previously(45). The sequences of the oligonucleotides used were as follows: c-MYC Blimp1site, 5�-ACAGAAAGGGAAAGGACTAGCGGATC-3�; Puta-1 Blimp1 site, 5�-GATCCGGAAAGGAAAGGGGATCTG-3�; Puta-2 Blimp1 site, 5�-GATCCAGTTGAGAAAGTGGTCAAG-3�; and Puta-3 Blimp1 site, 5�-GATCCGCCAGAGAAAGCTGTACTG-3�. The Oct-1 sequence was as reported previously (45).For supershift analysis, 200 ng of goat anti-Blimp1 antibody (Abcam) was incu-bated overnight with 5 �g of nuclear extract at 4°C; labeled probe was thenadded, and the mixture was incubated at room temperature for 30 min, asdescribed previously (57).

ChIP assay. Cells were cross-linked with 1% formaldehyde at room temper-ature for 10 min, and the reaction was stopped by the addition of glycine to a finalconcentration of 125 mM. After three washes in cold phosphate-buffered saline(PBS), cell pellets were resuspended in chromatin immunoprecipitation (ChIP)lysis buffer (50 mM HEPES [pH 7.5], 140 mM NaCl, and 1% Triton X-100 plusprotease inhibitor cocktail) and incubated 30 min on ice. After sonication andcentrifugation, the supernatants were incubated with 2 �g of salmon spermDNA, 150 �g of bovine serum albumin, and 50 �l of protein G-Sepharose for 2 hat 4°C and then centrifuged to preclear. The resulting supernatants were immu-noprecipitated overnight at 4°C with 2 �g of Blimp1 antibody (ab13700) plus 50�l of protein G-Sepharose and 2 �g of salmon sperm DNA. Immunoprecipitateswere washed sequentially for 5 min each time at 4°C with ChIP lysis buffer, ChIPlysis high-salt buffer (50 mM HEPES [pH 7.5], 500 mM NaCl, 1% Triton X-100),ChIP wash buffer (10 mM Tris [pH 8], 250 mM LiCl, 0.5% NP-40, 1 mM EDTA)and twice with Tris-EDTA buffer. Immunoprecipitates were then eluted twicewith elution buffer (50 mM Tris [pH 8], 1% SDS, 10 mM EDTA) at 65°C for 10min, and both input and pooled eluates were incubated overnight at 65°C toreverse the cross-linking. DNA was purified with QIAquick PCR purification kitand 1 �l from 50 �l of DNA extraction was used for PCR. (Information on PCRprimers and conditions for the ChIP assay is presented in the supplementalmaterial.)

RT-PCR analysis. RNA, isolated by using TRIzol reagent (Invitrogen), wasquantified by measuring the A260. RNA samples, with A260/A280 ratios between1.8 and 2.0, were treated with RQ1 RNase-free DNase (Promega Corp., MadisonWI). For reverse transcription-PCR (RT-PCR), 1 �g of RNA was reverse tran-scribed with Superscript RNase H-RT in the presence of 100 ng of randomprimers (Invitrogen). PCR for different targets was performed in a thermal

VOL. 29, 2009 RelB REPRESSES ER� TRANSCRIPTION BY INDUCING Blimp1 3833


http://mcb.asm

.org/D

ownloaded from

http://mcb.asm.org/

cycler. (Information on PCR primers and conditions for RT-PCR analysis ispresented in the supplemental material.)

Migration assays. Suspensions of 1 to 2.5 � 105 cells were layered, in tripli-cate, in the upper compartments of a Transwell (Costar, Cambridge, MA) on an8-mm-diameter polycarbonate filter (8-�m pore size) and incubated at 37°C forthe indicated times. All migration assays were performed three independenttimes, each in triplicate. Migration of the cells to the lower side of the filter wasevaluated with the acid phosphatase enzymatic assay using p-nitrophenyl phos-phate and determination of the optical density at 410 nm. Assays were performedthree independent times, each in triplicate, and the mean the standard devi-ation (SD) are presented.

Confocal laser scanning microscopy. ZR-75 cells were transfected with 1 �g ofRas-green fluorescent protein (GFP) expression plasmid or pEGFP-C1 EVDNA by using Fugene. At 24 h, cultures were incubated for 15 min in 100 nMMitoTracker (Molecular Probes), washed with PBS, and incubated in fixativesolution (2% electron microscopy-grade paraformaldehyde, PBS, 5 mM MgCl2).Cells were blocked in PBS–2% bovine serum albumin for 30 min and incubatedin blocking buffer containing Bcl-2 antibody (Santa Cruz catalog no. sc-492) for1 h at 37°C. Cells were then washed in PBS, incubated in PBS-bovine serumalbumin containing an Alexa 647 secondary antibody (Invitrogen catalog no.A2143) for 1 h, washed again in PBS, and mounted in antifade medium (Invitro-gen) containing DAPI (4�,6�-diamidino-2-phenylindole) to stain the nuclei. Flu-orescent signals were visualized by using a Zeiss LSM 510 200M confocal mi-croscope (Thornwood, NY) with a �63 objective lens. Z-sections were taken atan optical slice of 2 �m to determine the localization of MitoTracker (red), Bcl-2(purple), and Ras-GFP (green). Confocal settings were optimized to control forsignal crossover. Detector gain and amplitude offset were set to maximize thelinear range without saturation and were kept consistent throughout experi-ments. Images were stacked, composites were generated, and an orthogonal sliceanalysis was performed on the composite by using LSM 510 image software.

RESULTS

RelB represses synthesis of ER�. The inverse relationshipbetween NF-�B and ER� that typifies many breast cancer cellsand tissues has previously been related to the inhibition ofNF-�B synthesis or activity by ER� signaling (23, 51, 56).Unexpectedly, we observed the reciprocal repression of ER�activity by RelB in cultures of stable clones of ER�-positiveMCF-7 cells that ectopically express elevated RelB levels, asdescribed previously (56). Specifically, MCF-7 clones RELB(1)and RELB(2) showed marked reduction of ERE-driven re-porter activity compared to control MCF-7 EV(1) and EV(2)clones, with parental pcDNA3 empty vector (EV) DNA (Fig.1A). To verify that the observed effects of RelB on EREactivity were not due to clonal selection, transient-transfectionanalysis was performed using ER�-positive ZR-75 cells withvectors expressing either RelB or binding partners p50 or p52alone, or in combination. RelB expression resulted in a sub-stantial decrease in ERE-driven reporter activity, which wasfurther decreased upon coexpression of p50 or p52 (Fig. 1B).(Similar data were obtained by using transient transfection ofMCF-7 cells [data not shown]). To verify the effects of RelB onendogenous expression of ER� target genes, RNA was iso-lated from MCF-7/RELB(1) and MCF-7/EV(1) cells. RT-PCRanalysis confirmed decreased RNA levels of target genesCATHEPSIN D, Retinoic Acid Receptor � (RAR�), pS2, andMTA3 in MCF-7/RELB(1) compared to MCF-7/EV(1) cells(Fig. 1C). Similarly, endogenous mRNA levels of CATHEPSIND and RAR� were decreased in ZR-75 and MCF-7 cells upontransient transfection of RelB/p50 or RelB/p52 (Fig. 1D, upperand lower panels). To determine whether RelB decreases ER�levels, transient transfection analysis of vectors expressingRelB in the presence of either p50 or p52 was performed inZR-75 and MCF-7 cells. Decreased endogenous levels of ER�

protein (Fig. 1E) and ER� mRNA (Fig. 1F) were seen withboth RelB complexes. (All protein and RT-PCR samples inFig. 1E and F, respectively, were from the same gels.) ER� hastwo major promoters, A and B, of which the latter is predom-inantly used in breast cancer cells. Cotransfection of RelB andp52 expression vectors with the ER� promoter B reporterdecreases its activity 5-fold in MCF-7 and ZR-75 cells (Fig.1G). Lastly, we monitored the effects of RelB on endogenousER� levels and ER� promoter B reporter activity in the stableMCF-7/RELB(1) and RELB(2) clones and the control MCF-7/EV(1) and EV(2) clones. RelB expression reduced levels ofER� protein (Fig. 1H) and RNA (Fig. 1I), as well as promoteractivity (Fig. 1J). Thus, RelB complexes robustly inhibit ER�expression and activity.

RelB induces expression of the zinc finger repressor proteinBlimp1. Since RelB has a transactivation domain and has notbeen shown to function itself directly as an inhibitor of tran-scription, we tested the hypothesis that it controls the expres-sion of an intermediate negative regulatory factor that re-presses ER�. TransFac analysis (at 80% certainty) of promoterB sequences identified three putative binding sites for Blimp1,a master regulator of B and T-cell development. The Zn fingerprotein Blimp1, which functions as a repressor of key regula-tory genes in these cells, is expressed from the gene termedPRDM1 (human) or Blimp1 (mouse) (24). Since expression ofBlimp1 has not been reported in epithelial cells, we first testedfor endogenous Blimp1 protein in nuclear extracts of humanand mouse breast cancer cells. Substantial Blimp1 levels weredetectable in ER�-negative Hs578T human breast cancer cells,and more moderate levels were detectable in ER�-negativeMDA-MB-231 cells (Fig. 2A). Very low, but detectable Blimp1levels were seen with ER� positive ZR-75, MCF-7, T47D, andBT474 human breast cancer cell lines (better seen on darkerexposures and see Fig. 2G below). The murine NF639 Her-2/neu-driven ER� low breast cancer cell line displayed a mod-erate level of Blimp1 protein. RT-PCR confirmed the presenceof PRDM1 mRNA in human breast cancer cells and demon-strated that the levels of expression were much higher in ER�-negative versus ER�-positive lines (Fig. 2B). To test whetherthe inverse correlation between Blimp1 and ER� predicted bythese findings exists in primary human breast cancers, microar-ray gene expression datasets publicly available at www.oncomine.org were analyzed. The levels of PRDM1 mRNAwere significantly higher in ER�-negative versus ER�-positivebreast cancers (Fig. 2C, left panel) in the Van de Vijver_Breastcarcinoma microarray data set (reporter number NM_001198)(50) (P � 1.3e�6 [Student t test]), consistent with the analysisof the cell lines. The data from an additional microarray study(Bittner_Breast, reporter number 228964_at, publicly availableat www.oncomine.org) confirmed these findings (P � 7.3e�7

[Student t test]) (Fig. 2C, right panel). Thus, PRDM1 geneexpression occurs in patient samples and breast cancer celllines, with higher levels in ER�-negative cells, a finding con-sistent with the pattern of RelB expression (56).

Previously, we prepared stable Hs578T transfectants ex-pressing either RELB siRNA (Hs578T RELB siRNA) or senseRELB control siRNA (Hs578T control siRNA) (56). Analysisof these cells demonstrated that knockdown of RelB leads todecreased levels of PRDM1 mRNA (Fig. 2D) and Blimp1protein (Fig. 2E). Conversely, RelB/p52 expression in human

3834 WANG ET AL. MOL. CELL. BIOL.


http://mcb.asm

.org/D

ownloaded from

http://mcb.asm.org/

FIG. 1. RelB represses ER� synthesis in breast cancer cells. (A) Stable MCF-7 clones expressing either RELB or EV DNA, isolated asdescribed previously (56) and termed RELB(1), RELB(2), EV(1), and EV(2), were transfected with 0.5 �g of ERE-TK luciferase reporter,0.5 �g of SV40-�-Gal. Normalized ERE activity, set to 100% in the stable EV cells, was decreased in both RelB-expressing lines (mean the SD from three separate experiments). (B) ZR-75 cells were transiently transfected with 0.5 �g of ERE-TK luciferase DNA plus 0.5 �gof the indicated NF-�B subunit expression vectors, 0.5 �g of SV40-�-Gal, and EV pcDNA3 DNA to a 3.0 �g of DNA total. Normalized EREactivity, set to 100% in the EV-transfected cells, was reduced by RelB alone or in combination with p50 or p52 (mean the SD from threeseparate experiments). (C) RNA, isolated from EV(1) and RELB(1) stable MCF-7 clones, was subjected to RT-PCR analysis for expressionof ER� target genes CATHEPSIN D (CATHD), RAR�, pS2, MTA3, and GAPDH, as a loading control. (D) ZR-75 or MCF-7 cells weretransiently transfected with 3 �g of each of the vectors expressing RelB and either p50 or p52 or 6.0 �g of EV DNA. Isolated RNA wassubjected to RT-PCR analysis for RAR�, CATHEPSIN D (CATHD), and GAPDH. (E and F) ZR-75 or MCF-7 cells were transientlytransfected with 3.0 �g of vectors expressing RelB, and p50 or p52 or EV DNA. WCEs and RNA were isolated, which were subjected toimmunoblot analysis for the expression of ER� and �-actin (E) and to RT-PCR analysis for the expression of ER� and GAPDH (F),respectively. (Samples in panels E and F were run on the same gels, and the lanes were brought into contiguous positions.) (G) ZR-75 andMCF-7 cells were transiently transfected with 0.5 �g of ER� proB luciferase reporter plus 0.25 �g of vectors expressing RelB and p52 or0.5 �g of EV pcDNA3 DNA, 0.5 �g of SV40-�-Gal, and EV DNA to a 3.0 �g of DNA total. Normalized proB activity values are presentedas the means the SD from three experiments (control EV DNA set to 1). (H and I). Protein and RNA, isolated from RELB(1), RELB(2),EV(1), and EV(2) stable MCF-7 clones, were subjected to immunoblotting for the expression of ER�, RelB, and �-actin (H), and toRT-PCR analysis for the expression of ER� and GAPDH (I), respectively. (J) RELB(1), RELB(2), EV(1), and EV(2) MCF-7 clones weretransiently transfected with 0.5 �g of ER� proB luciferase reporter, 0.5 �g of SV40-�-Gal, and EV DNA to a 3.0 �g of DNA total.Normalized proB activity values are presented as the means the SD from three experiments (control EV DNA set to 1).



http://mcb.asm

.org/D

ownloaded from

http://mcb.asm.org/

FIG. 2. RelB induces Blimp1 in breast cancer cells. (A) Nuclear extracts were isolated from the indicated human and mouse breast cancercell lines and subjected to immunoblotting for Blimp1 and �-actin, which confirmed equal loading. (B) RNA, isolated from the indicatedhuman breast cancer lines, was subjected to RT-PCR for PRDM1 for either 28 or 30 cycles (as indicated) and for GAPDH levels. (C) In theleft panel, a box plot of data from the Van de Vijver_Breast carcinoma microarray data set (reporter number NM_001198) (50) was accessedby using the Oncomine Cancer Profiling Database (www.oncomine.org) and is plotted on a log scale. The data set includes 69 ER�-negativeand 226 ER�-positive human primary breast carcinoma samples. A Student t test, performed directly through the Oncomine 3.0 software,showed the difference in PRDM1 expression between the two groups was significant (P � 1.3e�6). In the right panel, a box plot of data fromthe Bittner study (see Results) showed that the difference in PRDM1 expression between the two groups was significant (P � 7.3e�7).(D) RNA was isolated from stable Hs578T transfectants expressing either RELB siRNA (Hs578T RELB siRNA) or sense RELB (Hs578Tcontrol siRNA) (Con) and analyzed for PRDM1 levels. (E) Nuclear proteins were isolated from Hs578T RELB siRNA and Hs578T controlsiRNA (Con) cells and analyzed for Blimp1, lamin B, and RelB levels by immunoblotting. (F) ZR-75, MCF-7, and NF639 cells weretransiently transfected with 3 �g each of RelB and p52 expression vectors or 6 �g of EV DNA and RNA subjected to RT-PCR for humanPRDM1 or mouse Blimp1 RNA levels. (G) ZR-75, MCF-7, and NF639 cells were transiently transfected with 3 �g each of RelB and p52expression vectors or 6 �g of EV DNA. Nuclear extracts (ZR-75 and MCF-7 [left panels]) and WCEs (NF639 [right panels]) were analyzedby immunoblotting for Blimp1 and either lamin B or �-actin, respectively. (H and I). Nuclear extracts and RNA, isolated from RELB(1),RELB(2), EV(1), and EV(2) stable MCF-7 clones, were subjected to immunoblotting for the expression of Blimp1 and lamin B (H) and toRT-PCR analysis for the expression of PRDM1 and GAPDH (I), respectively. (J) MCF-7 cells were transiently transfected with 0.5 �g of ER�proB luciferase reporter, 0.5 �g of SV40-�-Gal, and EV DNA to a 3.0 �g of DNA total. Normalized proB activity values are presented asthe means the SD from three experiments (control EV DNA set to 1).



http://mcb.asm

.org/D

ownloaded from

http://mcb.asm.org/

ZR-75 and MCF-7 cells and in mouse NF639 cells effectivelyinduced PRDM1 or Blimp1 RNA, respectively (Fig. 2F), andBlimp1 protein (Fig. 2G). Furthermore, RelB increased thelevels of Blimp1 protein (Fig. 2H) and PRDM1 RNA (Fig. 2I)in the stable MCF-7/RELB(1) and RELB(2) clones, indicatingfunctional RelB complexes can induce Blimp1 expression inER�-positive or -low breast cancer cells.

Next, we sought to test whether the ER� promoter is indeedrepressed by ectopic Blimp1 expression and to localize thefunctional Blimp1 binding element(s). An 4-fold reduction inactivity of an ER� proB construct (Fig. 2J), which contains allthree putative Blimp1 binding sites, was seen with ectopicBlimp1 expression in cotransfection analysis in MCF-7 cellscompared to EV DNA. The three Blimp1 sites all contain thecore GAAA sequence but differ in surrounding sequences (seeMaterials and Methods). To identify the functional sites, com-petition and supershift EMSAs were performed with nuclearextracts of ZR-75 cells ectopically expressing Blimp1 proteinor control EV DNA and an oligonucleotide containing theBlimp1 binding site of the c-MYC gene as a probe. Putative site2, but not sites 1 or 3, competed well for binding to the c-MYCBlimp1 site (Fig. 3A). When used as a probe, the putative site2 oligonucleotide effectively bound Blimp1 protein (Fig. 3B),as judged by a supershift EMSA (Fig. 3C), confirming that thissite is a bona fide Blimp1 element. The position of this con-firmed Blimp1 site is indicated on the map in Fig. 3D. ChIPanalysis confirmed intracellular binding of Blimp1 followingectopic expression to the ER� promoter B region containingthe element in MCF-7 cells (Fig. 3E, upper panels). Further-more, ChIP analysis in Hs578T cells, which expressed the high-est levels of Blimp1, confirmed binding of endogenous Blimp1to this ER� promoter B region (Fig. 3E, lower panels).

To assess the functional role of the low endogenous levels ofBlimp1 in the ER�-positive lines, a dominant-negative form ofBlimp1, termed TBlimp or dnBlimp (1), which contains onlythe DNA-binding domain, was used. Ectopic expression of thednBlimp caused an increase in ER� protein (Fig. 3F), indicat-ing that endogenous Blimp1 represses ER� in these lines.Consistently, as seen in Fig. 4A below, knockdown of Blimp1levels using an siRNA strategy in NF639 cells induced ER�protein levels. Conversely, ectopic Blimp1 expression substan-tially reduced ER� expression in ZR-75 and MCF-7 cells (Fig.3G), indicating that it can repress the endogenous ER� gene inboth lines. Blimp1 expression also led to reduced ER� RNAlevels in ZR-75 cells, while inhibition of Blimp1 activity withthe dnBlimp led to their induction (Fig. 3H). Of note, expres-sion of the dnBlimp ablated the decrease in ER� expressioninduced by ectopic RelB in MCF-7 cells (Fig. 3I), furtherimplicating Blimp1 in RelB-induced downregulation of ER�gene expression. Thus, Blimp1 is expressed in breast cancercells and mediates repression of ER�.

Blimp1 promotes a more migratory phenotype in breastcancer cells via inhibition of ER�. In primordial germ cells,the absence of Blimp1 led to defects in cell migration (33).Given our previous findings showing that ER�-negative breastcancer cells have a more migratory phenotype than ER�-pos-itive ones, we next tested whether Blimp1 regulated migrationand reduced expression of epithelial markers, which are re-quired for maintenance of a nonmigratory phenotype. First, ansiRNA was used to reduce Blimp1 levels in NF639 cells, which

are highly invasive and display robust Blimp1 expression.Knockdown of Blimp1 in NF639 cells decreased migration(Fig. 4A, left panel) and increased expression of E-cadherinand ER� (Fig. 4A, right panel). Conversely, ectopic Blimp1expression in ER�-positive ZR-75 cells, which repressed ER�levels (Fig. 3G), substantially increased migration (Fig. 4B, leftpanel) and commensurately decreased levels of E-cadherinand �-catenin (Fig. 4B, right panel). We next assessed the roleof ER� in the ability of Blimp1 to promote a more migratoryphenotype. Expression of ER� prevented the increased abilityof ZR-75 cells to migrate resulting from Blimp1 expression(3.1-fold versus 1.1-fold) (Fig. 4C, left panel), as well as thedecrease in E-cadherin and �-catenin induced by Blimp1 (Fig.4C, right panel). Lastly, dnBlimp expression induced the levelsof E-cadherin expression in ZR-75 and MCF-7 cells, and theseincreases were prevented upon knockdown of ER� usingsiER� RNA (Fig. 4D). Together, these results demonstrateBlimp1 induces a more migratory phenotype and implicate theinhibition of ER� in this control.

Bcl-2 induces Blimp1 via functional interaction with Ras.Previously, we demonstrated BCL2 is a critical RelB targetthat mediates the migratory phenotype of breast cancer cells(56). For example, knockdown of Bcl-2 in Hs578T cells im-paired the ability of the cells to migrate, whereas stable ectopicexpression of Bcl-2 in MCF-7 cells (MCF-7/Bcl-2 versus MCF-7/pBABE stable cells) was sufficient to promote a more migra-tory phenotype via decreasing E-cadherin levels (56). Thus, wetested the hypothesis that Bcl-2 mediates the induction ofBlimp1 by RelB. Knockdown of Bcl-2 with the Bcl2 shRNA inNF639 cells led to increased ER� levels (Fig. 5A, left panel),whereas Bcl-2 overexpression decreased the ER� level inZR-75 and MCF-7 cells (Fig. 5A, right panel), indicating thatBcl-2 can negatively regulate ER� expression. Consistently,knockdown of Bcl-2 levels led to downregulation of Blimp1protein and Blimp1 RNA levels in NF639 cells (Fig. 5B, upperpanels), whereas ectopic Bcl-2 expression robustly inducedtheir levels in these cells (Fig. 5B, lower panels). Similarly,Bcl-2 led to substantial increases in the levels of PRDM1 RNA(Fig. 5C, upper panels) and Blimp1 protein expression (Fig.5C, lower panels) in both ZR-75 and MCF-7 cells, while ec-topic BCL2 shRNA, which inhibited RelB/p52-mediated in-duction of Bcl-2 (56), prevented the induction of Blimp1 byRelB/p52 in ZR-75 cells, as well as the reduction of ER� levels(data not shown). Lastly, the dnBlimp1 robustly impaired thepreviously observed ability of Bcl-2 to enhance migration ofMCF-7 cells in MCF-7/Bcl-2 cells (Fig. 5D).

Wild-type (WT) Bcl-2, which has been shown to interactwith and activate Ras in the mitochondria of cells, leads to theinduction of NF-�B and AP-1, whereas a Bcl-2 14G mutantunable to associate with Ras is functionally inactive (7, 15, 35).Notably, Ohkubo et al. identified an AP-1 element upstream ofthe Blimp1 promoter (34), leading us to hypothesize a role forBcl-2–Ras interaction in the activation of Blimp1. As a firsttest, we compared the effects of the WT and the Bcl-2 14Gmutant on the PRDM1 promoter reporter in ZR-75 cells.Transfection of a vector expressing WT Bcl-2 resulted in asubstantial induction of the PRDM1 promoter activity inZR-75 cells (mean of 3.5- 1.7-fold, P � 0.05), whereas theBcl-2 14G mutant led to a reduction in activity of 5-fold[mean of 0.2- 0.06-fold, P � 0.002]. To assess for association



http://mcb.asm

.org/D

ownloaded from

http://mcb.asm.org/

FIG. 3. Blimp1 represses ER� gene expression in breast cancer cells. (A) Nuclear extracts, prepared from ZR-75 cells transfected withBlimp1 expression vector, were used in competition EMSAs with the Blimp1 element from the c-MYC gene as a probe and a 50-fold molarexcess of oligonucleotides containing the c-MYC or putative ER� promoter Blimp1 sites 1, 2, or 3 (P1 to P3). (B) Nuclear extracts, preparedfrom ZR-75 cells transfected with EV (�) or Blimp1 expression vector (�), were used in EMSAs with the putative ER� gene P1, P2, andP3 Blimp1 sites as a probe. (C) Nuclear extracts, prepared from ZR-75 cells transfected with Blimp1 expression vector, were used insupershift EMSAs in the absence (�) or presence (�) of 200 ng of Blimp1 antibody and the putative ER� gene Blimp1 P2 site as a probe.(D) ER� promoter A and B regions with confirmed Blimp1 site indicated with the element sequence given below, where the core isunderlined. (E) ChIP analyses were performed on MCF-7 cells transiently transfected with Blimp1 (top panels) or Hs578T cells (bottompanels) using anti-Blimp1 antibody or the corresponding normal IgG as indicated. The DNA, extracted from the immunoprecipitates or theinput, was amplified by PCR using pairs of primers specific to the confirmed Blimp1 site of ER� promoter (Blimp1 site) (left panels) or anupstream region as negative control (right panels). (F) ZR-75 and MCF-7 cells were transiently transfected with 5.0 �g of Vxy-puro TBlimpvector expressing the 35-kDa dnBlimp protein or EV DNA, and WCEs were subjected to immunoblotting for ER�, dnBlimp, and �-actin.(G) ZR-75 and MCF-7 cells were transiently transfected with 5.0 �g of Blimp1 expression vector or EV DNA, and WCEs were subjectedto immunoblotting for ER�, Blimp1, and �-actin. Densitometry indicated the ER� levels in the ZR-75 and MCF-7 cells were reduced to28.0 3.4 and 73.9 5.0, respectively, upon ectopic Blimp1 expression compared to EV DNA, set at 100. (H) ZR-75 cells were transientlytransfected with 5.0 �g of vector expressing either Blimp1 or dnBlimp protein (left or right panels, respectively) or the corresponding EVDNA, and RNA was subjected to RT-PCR for ER� and GAPDH. (I) MCF-7 cells carrying an inducible C4bsrR(TO-RelB) RelB construct,plated in the absence or presence of 1 �g of doxycycline/ml, were transiently transfected with 6.0 �g of dnBlimp expression vector. After48 h, WCEs were analyzed for levels of ER�, RelB, dnBlimp, and �-actin.



http://mcb.asm

.org/D

ownloaded from

http://mcb.asm.org/

of Bcl-2 and Ras in breast cancer cell lines, WCEs were pre-pared from ZR-75 cells transfected with vectors expressingH-Ras and Bcl-2. Immunoprecipitation of Bcl-2 led to copre-cipitation of Ras, whereas control rabbit IgG protein did not(Fig. 6A, left panel). Furthermore, coimmunoprecipitationanalysis of endogenous proteins in NF639 cell extracts similarlydetected association between endogenous Ras and Bcl-2 (Fig.6A, right panels). Confocal microscopy confirmed the partialassociation of transfected Ras-GFP with Bcl-2 in the mito-chondria of ZR-75 cells (Fig. 6B and see the projection imagesin Fig. S1 in the supplemental material).

Interestingly, ectopic Bcl-2 expression led to enhancedmembrane and reduced cytosolic localization of endogenousRas protein in ZR-75 and MCF-7 cells (Fig. 6C). To evaluatethe requirement for Ras membrane localization, we comparedthe effects on ER� levels of expression of the Ras WT versusa Ras C186S mutant, which is unable to localize to the mem-brane (5, 8). The Ras C186S mutant, which interacts with Bcl-2(data not shown), was unable to reduce ER� levels or toinduce Blimp1 expression in contrast to findings with Ras WT(Fig. 6D). Furthermore, Ras C186S prevented the reduction inER� levels mediated by Bcl-2 (Fig. 6E) and activation of the

PRDM1 promoter (Fig. 6F), suggesting that it functions as adominant-negative variant. Lastly, we tested the role of mito-chondrial localization using two Bcl-2 mutants: Bcl-acta, whichlocalizes to the mitochondria, and Bcl-nt, which predominantlylocalizes to the cytoplasm (64). Whereas Bcl-acta and Bcl-2WT cooperated with Ras to induce the PRDM1 promoter,Bcl-nt was unable to induce its activity (Fig. 6G), a findingconsistent with an important role for colocalization in themitochondria.

To test whether Ras signaling is sufficient to induce Blimp1expression, constitutively active Ras was ectopically expressedin ZR-75 and MCF-7 cells; potent upregulation of Blimp1 wasobserved (Fig. 7A). We next assessed data publicly available onmicroarray gene expression profiling of primary human mam-mary epithelial cells following transformation with variousgenes. Consistent with the data presented above, a significantinduction of PRDM1 mRNA expression was detected upon theexpression of activated H-Ras versus GFP (P � 1.0e�7), butnot with E2F3, activated �-catenin, c-Myc, or c-Src (2) (Fig.7B). A significant difference was also observed in PRDM1mRNA levels between the means of the H-Ras group versus allothers combined (P � 1.7e�10). Lastly, we sought to elucidate

FIG. 4. Blimp1 induces a migratory phenotype in breast cancer cells via repression of ER�. (A) NF639 cells were transiently transfected withcontrol siRNA (si-con) or siBlimp1 RNA. (Left panel) Cells were subjected, in triplicate, to a migration assay for 4 h. Cells that migrated to thelower side of the filter were quantified by spectrometric determination of the optical density at 410 nm. The average migration from threeindependent experiments the SD is presented relative to si-con (set at 100%) (*, P � 7.0e�5). (Right panel) WCEs were subjected toimmunoblotting for E-cadherin (E-Cadh), ER�, Blimp1, and �-actin. (B) ZR-75 cells were transiently transfected with 6 �g of EV or Blimp1expression vector. (Left panel) Transfected ZR-75 cells were subjected, in triplicate, to a migration assay for 24 h, as in panel A. *, P � 0.0007.(Right panel) WCEs were analyzed by immunoblotting for Blimp1, E-cadherin, �-catenin (�-Caten), and �-actin. (C) ZR-75 cells were transientlytransfected with 6 �g of Blimp1 expression vector or EV DNA in the absence or presence of ER� expression vector (1.5 �g) or EV DNA, asindicated. (Left panel) After 24 h, ZR-75 cells were subjected to migration assays for 24 h, as in panel A. The enhanced migration induced byBlimp1 is significantly decreased by ER�. *, P � 0.0001. (Right panel) After 48 h, WCEs were prepared and analyzed by immunoblotting forE-cadherin, �-catenin, ER�, Blimp1, and �-actin. (D) ZR-75 and MCF-7 cells were transiently transfected with dnBlimp (�) or EV DNA (�).After 24 h, ER� validated Stealth RNAi (�) or Stealth RNAi negative control (�) were introduced in cells using a Lipofectamine RNAiMAXreagent by reverse transfection as previously described (56). After 24 h, WCEs were prepared and analyzed by immunoblotting for E-cadherin,ER�, and �-actin.



http://mcb.asm

.org/D

ownloaded from

http://mcb.asm.org/

the role of Ras on cell migration induced by Bcl-2 in MCF-7cells using the dominant-negative function of the Ras C186Smutant. Ras C186S mutant, but not Ras WT, inhibited theBcl-2-mediated decrease in E-cadherin and ER� levels (Fig.7C) and the ability of MCF-7 cells to migrate (Fig. 7D), afinding consistent with the effects seen above in ZR-75 cells(Fig. 6D to F). Together, these data strongly argue that Bcl-2association with Ras, likely in the mitochondrial membrane,leads to the induction of the levels of Blimp1 and to a moremigratory phenotype (see the scheme in Fig. 7E).

DISCUSSION

Our findings identify a novel mechanism whereby the RelBNF-�B family member acts to repress ER� gene expression viainduction of the zinc finger protein Blimp1. Previously, wedemonstrated RelB induces transcription of the BCL2 gene viainteraction with C/EBP binding to an upstream CREB site (56)(Fig. 7E). Here, we show that the association of Bcl-2 with Ras,activates this proto-oncogene and leads to expression of the Znfinger repressor Blimp1. In turn, Blimp1 binds to an elementupstream of ER� promoter B, reducing promoter activity andER� levels and thereby causing a reduction in the levels ofE-cadherin and �-catenin and a commensurate increase inmigratory phenotype in breast cancer cells (see the scheme inFig. 7E). To our knowledge, this is the first study identifying acrucial role for Blimp1 in repression of ER� transcription andalso in increased migration of cancer cells. Indeed, Blimp1 is awell-known zinc finger transcriptional repressor that is criticalfor terminal differentiation of B cells into immunoglobulin-secreting plasma cells and functions by attenuating prolifera-tion while promoting the differentiated B-cell gene expressionprogram (10). More recently, Blimp1 has been implicated inT-cell homeostasis (16), where it similarly represses the expres-sion of genes promoting proliferation, e.g., interleukin-2 (IL-2)and Bcl-6 while enhancing IL-10 production. Interestingly,studies have also indicated that Blimp1 is a critical determinantof primordial germ cells differentiation (33). The absence ofBlimp1 led to the formation of primordial germ cells clustersand to defects in primordial germ cell migration, as well as toapoptosis (33). Our findings identify one mechanism for theeffects of Blimp1 on migration of breast cancer cells related, inpart, to the downregulation of ER� gene expression. It re-mains to be determined whether similar regulation of ER� byBlimp1 occurs in B cells and plays a role in autoimmunedisease. Overall, our data identify Blimp1 as a potent negativeregulator of ER� expression and thereby as an activator of themigration of breast cancer cells.

RelB has been reported to activate gene transcription viaseveral different mechanisms. RelB has been shown to inducetranscription of the c-Myb oncogene via releasing pausing thatoccurs during elongation (47). The activation of the Blc, Elc,Sdf-1, and Slc promoters by RelB required IKK� (4). Interest-ingly, activation of the dapk1, dapk3, c-flip, and birc3 promotersby RelB can be prevented by Daxx, which interacts selectivelywith RelB but not with RelA or c-Rel (36). In addition to theBCL2 gene (56), RelB induces transcription of several otherprosurvival genes, including MNSOD in prostate cancer (60)and MNSOD and SURVIVIN in breast cancer (29). However,RelB was unable to bind to the NF-�B elements upstream of

FIG. 5. Induction of Blimp1 is mediated via Bcl-2. (A) For the left pan-els, NF639 cells were transiently transfected with expression vectors for eithercontrol shRNA (contl shRNA) or Bcl2 shRNA, and WCEs were analyzed forER�, Bcl-2 (which confirmed effective knockdown), and �-actin protein lev-els. For the right panels, ZR-75 and MCF-7 cells were transiently transfectedwith EV or full-length Bcl-2 expression vector. WCEs were analyzed for ER�and �-actin protein levels and for Bcl-2 (which confirmed the expectedchanges in levels [data not shown]). (B) NF639 cells were transiently trans-fected with expression vectors for either control shRNA (contl shRNA), Bcl2shRNA (upper panels), or EV or full-length Bcl-2 (bottom panels). For theleft panels, WCEs were analyzed for Blimp1 and �-actin. For the right panels,RNA was subjected to RT-PCR analysis determine the levels of Blimp1 andGapdh. (C) RNA (top panels) and nuclear extracts (bottom panels) wereprepared from ZR-75 and MCF-7 cells transfected with control EV DNA orvector expressing Bcl-2 and subjected to RT-PCR for PRDM1 and GAPDHexpression (upper panels) and Blimp1 and lamin B levels (lower panels),respectively. (D) Stable mixed populations of MCF-7 cells expressing Bcl-2 orpBabe (pB) DNA were transiently transfected with EV DNA (�) or vectorexpressing dnBlimp (�). Cells were subjected, in triplicate, to a migrationassay for 24 h, as in Fig. 4A. The increase in migration induced by Bcl-2 issignificantly reduced by dnBlimp, P � 3.8e�7. Western blot analysis con-firmed the expression of the dnBlimp (not shown).



http://mcb.asm

.org/D

ownloaded from

http://mcb.asm.org/

FIG. 6. Bcl-2 interaction with Ras induces Blimp1. (A) WCEs (500 �g), prepared from ZR-75 cells transiently transfected with Bcl-2 and Rasexpression vectors (left panels) or from untransfected NF639 cells (right panels), were immunoprecipitated with rabbit anti-Bcl-2 or mouse anti-Rasantibodies or the corresponding normal IgG as indicated. Immunoprecipitated proteins were analyzed by immunoblotting for Bcl-2 and Ras. WCEs (5�g) were used as input. (B) ZR-75 cells, grown on coverslips in six-well dishes, were transfected with 1 �g of Ras-GFP expression plasmid. After stainingwith MitoTracker, the cells were fixed and subjected to immunohistochemistry for Bcl-2 using an Alexa 647-labeled secondary antibody (purple).Mitochondria are labeled in red with MitoTracker, Ras-GFP is green, and nuclei are labeled with DAPI stain (blue). A composite image was generatedfrom the z-sections and an Orthoganol slice analysis was performed on the composite where XZ is depicted on the horizontal and YZ is depicted on thevertical axis outside the dimension of the composite. Colocalization of Ras and Bcl-2 to the mitochondria is demonstrated by the overlap of signalsyielding white spots. Three-dimensional projection images were made from the composites. The Inset shows a 45° rotation of the image of the cell, witharrows indicating areas of colocalization of Ras and Bcl-2 in the mitochondria. (C) ZR-75 and MCF-7 cells were transiently transfected with EV or avector expressing full-length Bcl-2. After 48 h, cytosol and membrane fractions were prepared as described in Materials and Methods and subjected toimmunoblotting for Ras and �-actin. Immunoblotting for VDAC, a mitochondrial ion channel membrane protein, confirmed effective separation of themembrane from the cytoplasmic compartment (data not shown). (D) ZR-75 cells were transiently transfected with EV (�) or vectors expressing Ras WTor Ras C186S mutant protein. After 48 h, WCEs and nuclear extracts were analyzed for levels of ER� and �-actin (upper panels) and Blimp1 and laminB (lower panels). (E) ZR-75 cells were transiently transfected with EV or vector expressing full-length Bcl-2 in the absence (�) or presence (�) of avector expressing Ras C186S mutant protein. After 48 h, WCEs were analyzed for ER� protein levels. (F) ZR-75 cells were transiently transfected intriplicate, with 0.5 �g of the 7-kb human PRDM1 promoter-Luc vector, 0.5 �g of Bcl-2 with 0.5 �g of Ras WT (Ras) or 0.5 �g of Ras C186S (C186S)expression vector, 0.5 �g of SV40-�-Gal, and pcDNA3 (EV) to make a total of 3.0 �g of DNA. The luciferase and �-Gal activities were determined, andnormalized PRDM1 promoter activity values are presented as the means the SD (control EV DNA set to 1). (G) ZR-75 cells were transientlytransfected in triplicate with 0.5 �g of the 7-kb human PRDM1 promoter-Luc vector; 0.5 �g of Ras expression vector with 0.5 �g of either EV DNA,WT Bcl-2, Bcl-acta (mitochondrion-localized Bcl-2 mutant), or Bcl-nt (cytoplasm-localized Bcl-2 mutant) expression vectors (64); 0.5 �g of SV40-�-Gal;and pRc/CMV (EV) to make a total of 3.0 �g of DNA. At 48 h after transfection, luciferase and �-Gal activities were determined, and normalizedPRDM1 promoter activity values are presented as the means the SD from three separate experiments (control EV DNA set to 1). Using aMann-Whitney test, we determined a P value of 0.05 between each condition, except for Ras plus WT Bcl-2 and Ras plus Bcl-acta.



http://mcb.asm

.org/D

ownloaded from

http://mcb.asm.org/

the Bcl-xl promoter and led to decreased Bcl-xl gene expres-sion, in contrast to the potent activation seen upon coexpres-sion of either c-Rel or RelA (21). In fibroblasts, RelB has alsobeen reported to suppress expression of several genes, includ-ing IL-1�, IL-1�, and tumor necrosis factor alpha, indirectlyvia modulating I�B� stability (59) and to inactivate p65 via theformation of dimeric complex (27). We report here that RelBcan repress expression of the ER� promoter via activation ofthe zinc finger repressor protein Blimp1. Consistent with thisfinding, ER� and RelB levels display an inverse pattern ofexpression in many breast cancer cells and tissues, includinginflammatory breast cancers (23, 30, 52, 56).

To our knowledge, this is the first report of a transcription

factor that acts as a repressor of ER� promoter activity. Pre-vious work showing repression of ER� promoter transcriptionin breast cancer cells implicated DNA methyltransferases andhistone deacetylases in the regulation (43, 62). To date, theidentified transcription factors controlling ER� promoter ac-tivity have all been shown to act as positive regulators. de-Graffenried et al. (12) and Clark et al. (13) have identified SP1,USF-1, and ER� itself as essential factors required for fullER� gene transcription. The estrogen receptor factor ERF-1and the estrogen receptor promoter B-associated factorERBF-1 have been implicated in the positive control of ER�promoter A and B activity, respectively (11, 48). Moreover, ourprevious study also indicated that the Forkhead box O protein

FIG. 7. Induction of Blimp1 is mediated via Bcl-2 and Ras activation. (A) ZR-75 and MCF-7 cells were transiently transfected with EV orvector expressing constitutively active Ras (Ras CA), and nuclear extracts were analyzed for Blimp1 and lamin B. (B) A box plot of data from theBild_CellLine microarray data set of primary human mammary epithelial cells transfected with the vectors expressing GFP, E2F3, �-catenin(�-cat), c-Myc, c-Src, or H-Ras (reporter number 228964_at) was accessed for PRDM1 mRNA levels by using the Oncomine Cancer ProfilingDatabase and is plotted on a log scale (2). A Student t test, performed directly through the Oncomine 3.0 software, showed that the difference inPRDM1 mRNA levels between the means of the H-Ras versus all others groups combined was significant (*, P � 1.7e�10). (C and D) Stable mixedpopulations of MCF-7 cells expressing Bcl-2 or pBabe DNA were transiently transfected with 6 �g of vectors expressing Ras WT or Ras C186S,as indicated. (C) After 48 h, WCEs were prepared and analyzed by immunoblotting for E-cadherin (E-Cadh), ER�, Bcl-2, and �-actin.(D) Alternatively, cells were subjected to migration assays. Analysis of variance shows that the difference between MCF-7/pBABE � EV andMCF-7/Bcl-2 � Ras C186S was not statistically significant (P � 0.84), whereas the differences between MCF-7/pBABE � EV and eitherMCF-7/Bcl-2 �EV or MCF-7/Bcl-2 � Ras WT were significant, as indicated. (E) Summary scheme. Bcl-2, induced by RelB/p52 as shownpreviously (56), interacts with and activates Ras, apparently in the mitochondrial membrane. In turn, Ras induces expression of Blimp1, whichrepresses ER�.



http://mcb.asm

.org/D

ownloaded from

http://mcb.asm.org/

3a (FOXO3a) activates ER� promoter B transcription inbreast cancer cells (18).

Activation of Ras via interaction with Bcl-2 has been previ-ously reported in the context of Ras-mediated apoptosis of Tcells (15, 38) and of adipocytes (31), although the findings weresomewhat contradictory. Ras–Bcl-2 interaction in T cellsblocked Ras-induced apoptosis, whereas in adipocytes it led tothe induction of apoptosis via inactivation of Bcl-2 prosurvivalfunction. Bcl-2 has also been shown to interact with Raf, al-though the resulting effects differed depending upon the cel-lular conditions. In murine myeloid progenitor cells, Bcl-2 tar-geted Raf to mitochondrial membranes, allowing this kinase toprotect against apoptosis by phosphorylating BAD (55),whereas Taxol-induced activation of Raf was accompanied bythe loss of Bcl-2 antiapoptotic function in MCF-7 cells (3). Ofnote, we observed that a constitutively active Raf could simi-larly induce Blimp1 expression, implicating this Ras-inducedpathway in signaling events (results not shown). Interestingly,El-Ashry and coworkers showed that inhibition of Ras or Rafactivities led to the reexpression of ER� in MCF-7 cells (32).Our findings that Blimp1 mediates repression of ER� canexplain these observations. Overall, our study indicates thatRelB-mediated induction of Bcl-2 leads to the repression ofER� transcription and thus to a more migratory phenotype ofbreast cancer cells by activating Ras.

ACKNOWLEDGMENTS

We thank Mark R. Philips, Pierre Chambon, Kathryn Calame, Wen-Luan Wendy Hsiao, David W. Andrews, and Tristram G. Parslow forproviding cloned DNAs and Philip Leder for providing cell lines.

These studies were supported by National Institutes of Health(NIH) grants P01 ES11624, R01 CA129129, R01 CA36355, and R01EY06000; NIH training grant T32 HL007501; and a fellowship fromthe American Institute for Cancer Research with funding from theDerx Foundation.

REFERENCES

1. Angelin-Duclos, C., K. Johnson, J. Liao, K. I. Lin, and K. Calame. 2002. Aninterfering form of Blimp-1 increases IgM secreting plasma cells and blocksmaturation of peripheral B cells. Eur. J. Immunol. 32:3765–3775.

2. Bild, A. H., G. Yao, J. T. Chang, Q. Wang, A. Potti, D. Chasse, M. B. Joshi,D. Harpole, J. M. Lancaster, A. Berchuck, J. A. Olson, Jr., J. R. Marks, H. K.Dressman, M. West, and J. R. Nevins. 2006. Oncogenic pathway signaturesin human cancers as a guide to targeted therapies. Nature 439:353–357.

3. Blagosklonny, M. V., T. Schulte, P. Nguyen, J. Trepel, and L. M. Neckers.1996. Taxol-induced apoptosis and phosphorylation of Bcl-2 protein involvesc-Raf-1 and represents a novel c-Raf-1 signal transduction pathway. CancerRes. 56:1851–1854.

4. Bonizzi, G., M. Bebien, D. C. Otero, K. E. Johnson-Vroom, Y. Cao, D. Vu,A. G. Jegga, B. J. Aronow, G. Ghosh, R. C. Rickert, and M. Karin. 2004.Activation of IKK� target genes depends on recognition of specific �Bbinding sites by RelB:p52 dimers. EMBO J. 23:4202–4210.

5. Cadwallader, K. A., H. Paterson, S. G. Macdonald, and J. F. Hancock. 1994.N-terminally myristoylated Ras proteins require palmitoylation or a polyba-sic domain for plasma membrane localization. Mol. Cell. Biol. 14:4722–4730.

6. Calame, K. L., K. I. Lin, and C. Tunyaplin. 2003. Regulatory mechanismsthat determine the development and function of plasma cells. Annu. Rev.Immunol. 21:205–230.

7. Choi, J., K. Choi, E. N. Benveniste, S. B. Rho, Y. S. Hong, J. H. Lee, J. Kim,and K. Park. 2005. Bcl-2 promotes invasion and lung metastasis by inducingmatrix metalloproteinase-2. Cancer Res. 65:5554–5560.

8. Choy, E., V. K. Chiu, J. Silletti, M. Feoktistov, T. Morimoto, D. Michaelson,I. E. Ivanov, and M. R. Philips. 1999. Endomembrane trafficking of ras: theCAAX motif targets proteins to the ER and Golgi. Cell 98:69–80.

9. Cogswell, P. C., D. C. Guttridge, W. K. Funkhouser, and A. S. Baldwin, Jr.2000. Selective activation of NF-�B subunits in human breast cancer: poten-tial roles for NF-�B2/p52 and for Bcl-3. Oncogene 19:1123–1131.

10. Davis, M. M. 2007. Blimp-1 over Budapest. Nat. Immunol. 8:445–447.11. deConinck, E. C., L. A. McPherson, and R. J. Weigel. 1995. Transcriptional

regulation of estrogen receptor in breast carcinomas. Mol. Cell. Biol. 15:2191–2196.

12. deGraffenried, L. A., S. G. Hilsenbeck, and S. A. Fuqua. 2002. Sp1 is essentialfor estrogen receptor alpha gene transcription. J. Steroid Biochem. Mol.Biol. 82:7–18.

13. deGraffenried, L. A., T. A. Hopp, A. J. Valente, R. A. Clark, and S. A. Fuqua.2004. Regulation of the estrogen receptor alpha minimal promoter by Sp1,USF-1, and ER�. Breast Cancer Res. Treat. 85:111–120.

14. Demicco, E. G., K. T. Kavanagh, R. Romieu-Mourez, X. Wang, S. R. Shin, E.Landesman-Bollag, D. C. Seldin, and G. E. Sonenshein. 2005. RelB/p52NF-�B complexes rescue an early delay in mammary gland development intransgenic mice with targeted super-repressor I�B-alpha expression andpromote carcinogenesis of the mammary gland. Mol. Cell. Biol. 25:10136–10147.

15. Denis, G. V., Q. Yu, P. Ma, L. Deeds, D. V. Faller, and C. Y. Chen. 2003.Bcl-2, via its BH4 domain, blocks apoptotic signaling mediated by mitochon-drial Ras. J. Biol. Chem. 278:5775–5785.

16. Fink, P. J. 2006. T cells join the Blimp-1 brigade. Nat. Immunol. 7:445–446.17. Grilli, M., J. J. Chiu, and M. J. Lenardo. 1993. NF-�B and Rel: participants

in a multiform transcriptional regulatory system. Int. Rev. Cytol. 143:1–62.18. Guo, S., and G. E. Sonenshein. 2004. Forkhead box transcription factor

FOXO3a regulates estrogen receptor alpha expression and is repressed bythe Her-2/neu/phosphatidylinositol 3-kinase/Akt signaling pathway. Mol.Cell. Biol. 24:8681–8690.

19. Gyory, I., J. Wu, G. Fejer, E. Seto, and K. L. Wright. 2004. PRDI-BF1recruits the histone H3 methyltransferase G9a in transcriptional silencing.Nat. Immunol. 5:299–308.

20. Jeay, S., S. Pianetti, H. M. Kagan, and G. E. Sonenshein. 2003. Lysyl oxidaseinhibits ras-mediated transformation by preventing activation of NF-�B.Mol. Cell. Biol. 23:2251–2263.

21. Jiang, H. Y., C. Petrovas, and G. E. Sonenshein. 2002. RelB-p50 NF-�Bcomplexes are selectively induced by cytomegalovirus immediate-early pro-tein 1: differential regulation of Bcl-x(L) promoter activity by NF-�B familymembers. J. Virol. 76:5737–5747.

22. Johnson, K., M. Shapiro-Shelef, C. Tunyaplin, and K. Calame. 2005. Reg-ulatory events in early and late B-cell differentiation. Mol. Immunol. 42:749–761.

23. Kalaitzidis, D., and T. D. Gilmore. 2005. Transcription factor cross-talk: theestrogen receptor and NF-�B. Trends Endocrinol. Metab. 16:46–52.

24. Keller, A. D., and T. Maniatis. 1991. Identification and characterization of anovel repressor of beta-interferon gene expression. Genes Dev. 5:868–879.

25. Lernbecher, T., U. Muller, and T. Wirth. 1993. Distinct NF-�B/Rel tran-scription factors are responsible for tissue-specific and inducible gene acti-vation. Nature 365:767–770.

26. Lessard, L., L. R. Begin, M. E. Gleave, A. M. Mes-Masson, and F. Saad.2005. Nuclear localization of nuclear factor-�B transcription factors in pros-tate cancer: an immunohistochemical study. Br. J. Cancer 93:1019–1023.

27. Marienfeld, R., M. J. May, I. Berberich, E. Serfling, S. Ghosh, and M.Neumann. 2003. RelB forms transcriptionally inactive complexes with RelA/p65. J. Biol. Chem. 278:19852–19860.

28. Martins, G. A., L. Cimmino, M. Shapiro-Shelef, M. Szabolcs, A. Herron, E.Magnusdottir, and K. Calame. 2006. Transcriptional repressor Blimp-1 reg-ulates T-cell homeostasis and function. Nat. Immunol. 7:457–465.

29. Mineva, N. D., X. Wang, S. Yang, H. Ying, Z.-X. J. Xiao, M. F. Holick, andG. E. Sonenshein. Inhibition of RelB by 1,25-dihydroxyvitamin D3 promotessensitivity of breast cancer cells to radiation. J. Cell Physiol., in press.

30. Nakshatri, H., P. Bhat-Nakshatri, D. A. Martin, R. J. Goulet, Jr., and G. W.Sledge, Jr. 1997. Constitutive activation of NF-�B during progression ofbreast cancer to hormone-independent growth. Mol. Cell. Biol. 17:3629–3639.

31. Navarro, P., A. M. Valverde, M. Benito, and M. Lorenzo. 1999. ActivatedHa-ras induces apoptosis by association with phosphorylated Bcl-2 in amitogen-activated protein kinase-independent manner. J. Biol. Chem. 274:18857–18863.

32. Oh, A. S., L. A. Lorant, J. N. Holloway, D. L. Miller, F. G. Kern, and D.El-Ashry. 2001. Hyperactivation of MAPK induces loss of ER� expression inbreast cancer cells. Mol. Endocrinol. 15:1344–1359.

33. Ohinata, Y., B. Payer, D. O’Carroll, K. Ancelin, Y. Ono, M. Sano, S. C.Barton, T. Obukhanych, M. Nussenzweig, A. Tarakhovsky, M. Saitou, andM. A. Surani. 2005. Blimp1 is a critical determinant of the germ cell lineagein mice. Nature 436:207–213.

34. Ohkubo, Y., M. Arima, E. Arguni, S. Okada, K. Yamashita, S. Asari, S.Obata, A. Sakamoto, M. Hatano, J. O.-Wang, M. Ebara, H. Saisho, and T.Tokuhisa. 2005. A role for c-fos/activator protein 1 in B lymphocyte terminaldifferentiation. J. Immunol. 174:7703–7710.

35. Omidvar, N., S. Kogan, S. Beurlet, C. le Pogam, A. Janin, R. West, M. E.Noguera, M. Reboul, A. Soulie, C. Leboeuf, N. Setterblad, D. Felsher, E.Lagasse, A. Mohamedali, N. S. Thomas, P. Fenaux, M. Fontenay, M. Pla,G. J. Mufti, I. Weissman, C. Chomienne, and R. A. Padua. 2007. BCL-2 andmutant NRAS interact physically and functionally in a mouse model ofprogressive myelodysplasia. Cancer Res. 67:11657–11667.

36. Puto, L. A., and J. C. Reed. 2008. Daxx represses RelB target promoters viaDNA methyltransferase recruitment and DNA hypermethylation. GenesDev. 22:998–1010.



http://mcb.asm

.org/D

ownloaded from

http://mcb.asm.org/

37. Rayet, B., and C. Gelinas. 1999. Aberrant rel/nfkb genes and activity inhuman cancer. Oncogene 18:6938–6947.

38. Rebollo, A., D. Perez-Sala, and A. C. Martinez. 1999. Bcl-2 differentiallytargets K-, N-, and H-Ras to mitochondria in IL-2 supplemented or deprivedcells: implications in prevention of apoptosis. Oncogene 18:4930–4939.

39. Ren, B., K. J. Chee, T. H. Kim, and T. Maniatis. 1999. PRDI-BF1/Blimp-1repression is mediated by corepressors of the Groucho family of proteins.Genes Dev. 13:125–137.

40. Romieu-Mourez, R., D. W. Kim, S. M. Shin, E. G. Demicco, E. Landesman-Bollag, D. C. Seldin, R. D. Cardiff, and G. E. Sonenshein. 2003. Mousemammary tumor virus c-rel transgenic mice develop mammary tumors. Mol.Cell. Biol. 23:5738–5754.

41. Schjerven, H., T. N. Tran, P. Brandtzaeg, and F. E. Johansen. 2004. De novosynthesized RelB mediates TNF-induced up-regulation of the human poly-meric Ig receptor. J. Immunol. 173:1849–1857.

42. Sentman, C. L., J. R. Shutter, D. Hockenbery, O. Kanagawa, and S. J.Korsmeyer. 1991. bcl-2 inhibits multiple forms of apoptosis but not negativeselection in thymocytes. Cell 67:879–888.

43. Sharma, D., J. Blum, X. Yang, N. Beaulieu, A. R. Macleod, and N. E.Davidson. 2005. Release of methyl CpG binding proteins and histonedeacetylase 1 from the estrogen receptor alpha (ER) promoter upon reac-tivation in ER-negative human breast cancer cells. Mol. Endocrinol. 19:1740–1751.

44. Siebenlist, U., G. Franzoso, and K. Brown. 1994. Structure, regulation andfunction of NF-�B. Annu. Rev. Cell Biol. 10:405–455.

45. Sovak, M. A., R. E. Bellas, D. W. Kim, G. J. Zanieski, A. E. Rogers, A. M.Traish, and G. E. Sonenshein. 1997. Aberrant nuclear factor-�B/Rel expres-sion and the pathogenesis of breast cancer. J. Clin. Investig. 100:2952–2960.

46. Stoffel, A., M. Chaurushiya, B. Singh, and A. J. Levine. 2004. Activation ofNF-�B and inhibition of p53-mediated apoptosis by API2/mucosa-associatedlymphoid tissue 1 fusions promote oncogenesis. Proc. Natl. Acad. Sci. USA101:9079–9084.

47. Suhasini, M., C. D. Reddy, E. P. Reddy, J. A. DiDonato, and R. B. Pilz. 1997.cAMP-induced NF-�B (p50/relB) binding to a c-myb intronic enhancer cor-relates with c-myb up-regulation and inhibition of erythroleukemia cell dif-ferentiation. Oncogene 15:1859–1870.

48. Tanimoto, K., H. Eguchi, T. Yoshida, K. Hajiro-Nakanishi, and S. Hayashi.1999. Regulation of estrogen receptor alpha gene mediated by promoter Bresponsible for its enhanced expression in human breast cancer. NucleicAcids Res. 27:903–909.

49. Tunyaplin, C., M. A. Shapiro, and K. L. Calame. 2000. Characterization ofthe B lymphocyte-induced maturation protein-1 (Blimp-1) gene, mRNAisoforms and basal promoter. Nucleic Acids Res. 28:4846–4855.

50. van de Vijver, M. J., Y. D. He, L. J. van’t Veer, H. Dai, A. A. Hart, D. W.Voskuil, G. J. Schreiber, J. L. Peterse, C. Roberts, M. J. Marton, M. Parrish,D. Atsma, A. Witteveen, A. Glas, L. Delahaye, T. van der Velde, H. Bartelink,S. Rodenhuis, E. T. Rutgers, S. H. Friend, and R. Bernards. 2002. A geneexpression signature as a predictor of survival in breast cancer. N. Engl.J. Med. 347:1999–2009.

51. Van den Eynden, G. G., I. Van der Auwera, S. Van Laere, C. G. Colpaert, P.van Dam, S. Merajver, C. G. Kleer, A. L. Harris, E. A. Van Marck, L. Y.Dirix, and P. B. Vermeulen. 2004. Validation of a tissue microarray to study

differential protein expression in inflammatory and non-inflammatory breastcancer. Breast Cancer Res. Treat. 85:13–22.

52. Van Laere, S. J., I. Van der Auwera, G. G. Van den Eynden, H. J. Elst, J.Weyler, A. L. Harris, P. van Dam, E. A. Van Marck, P. B. Vermeulen, andL. Y. Dirix. 2006. Nuclear factor-�B signature of inflammatory breast cancerby cDNA microarray validated by quantitative real-time reverse transcrip-tion-PCR, immunohistochemistry, and nuclear factor-�B DNA-binding.Clin. Cancer Res. 12:3249–3256.

53. Vilimas, T., J. Mascarenhas, T. Palomero, M. Mandal, S. Buonamici, F.Meng, B. Thompson, C. Spaulding, S. Macaroun, M. L. Alegre, B. L. Kee, A.Ferrando, L. Miele, and I. Aifantis. 2007. Targeting the NF-�B signalingpathway in Notch1-induced T-cell leukemia. Nat. Med. 13:70–77.

54. Vincent, S. D., N. R. Dunn, R. Sciammas, M. Shapiro-Shalef, M. M. Davis,K. Calame, E. K. Bikoff, and E. J. Robertson. 2005. The zinc finger tran-scriptional repressor Blimp1/Prdm1 is dispensable for early axis formationbut is required for specification of primordial germ cells in the mouse.Development 132:1315–1325.

55. Wang, H. G., U. R. Rapp, and J. C. Reed. 1996. Bcl-2 targets the proteinkinase Raf-1 to mitochondria. Cell 87:629–638.

56. Wang, X., K. Belguise, N. Kersual, K. H. Kirsch, N. D. Mineva, F. Galtier,D. Chalbos, and G. E. Sonenshein. 2007. Oestrogen signalling inhibits inva-sive phenotype by repressing RelB and its target BCL2. Nat. Cell Biol.9:470–478.

57. Wang, X., and G. E. Sonenshein. 2005. Induction of the RelB NF-�B subunitby the cytomegalovirus IE1 protein is mediated via Jun kinase and c-Jun/Fra-2 AP-1 complexes. J. Virol. 79:95–105.

58. Weih, F., D. Carrasco, S. K. Durham, D. S. Barton, C. A. Rizzo, R. P. Ryseck,S. A. Lira, and R. Bravo. 1995. Multiorgan inflammation and hematopoieticabnormalities in mice with a targeted disruption of RelB, a member of theNF-�B/Rel family. Cell 80:331–340.

59. Xia, Y., S. Chen, Y. Wang, N. Mackman, G. Ku, D. Lo, and L. Feng. 1999.RelB modulation of I�Balpha stability as a mechanism of transcriptionsuppression of interleukin-1� (IL-1�), IL-1�, and tumor necrosis factoralpha in fibroblasts. Mol. Cell. Biol. 19:7688–7696.

60. Xu, Y., F. Fang, D. K. St Clair, S. Josson, P. Sompol, I. Spasojevic, and W. H.St. Clair. 2007. Suppression of RelB-mediated manganese superoxide dis-mutase expression reveals a primary mechanism for radiosensitization effectof 1�,25-dihydroxyvitamin D3 in prostate cancer cells. Mol. Cancer Ther.6:2048–2056.

61. Yan, J., J. Jiang, C. A. Lim, Q. Wu, H. H. Ng, and K. C. Chin. 2007. BLIMP1regulates cell growth through repression of p53 transcription. Proc. Natl.Acad. Sci. USA 104:1841–1846.

62. Yan, L., S. J. Nass, D. Smith, W. G. Nelson, J. G. Herman, and N. E.Davidson. 2003. Specific inhibition of DNMT1 by antisense oligonucleotidesinduces re-expression of estrogen receptor-alpha (ER) in ER-negative hu-man breast cancer cell lines. Cancer Biol. Ther. 2:552–556.

63. Yu, J., C. Angelin-Duclos, J. Greenwood, J. Liao, and K. Calame. 2000.Transcriptional repression by blimp-1 (PRDI-BF1) involves recruitment ofhistone deacetylase. Mol. Cell. Biol. 20:2592–2603.

64. Zhu, W., A. Cowie, G. W. Wasfy, L. Z. Penn, B. Leber, and D. W. Andrews.1996. Bcl-2 mutants with restricted subcellular location reveal spatially dis-tinct pathways for apoptosis in different cell types. EMBO J. 15:4130–4141.



http://mcb.asm

.org/D

ownloaded from

http://mcb.asm.org/

relb nf-b represses estrogen receptor expression via induction

Documents