Estrogen receptors alfa (ERα) and beta (ERβ) differentially regulate proliferation and apoptosis of the normal murine mammary epithelial cell line HC11

Estrogen receptors alfa (ERa) and beta (ERb) differentially regulate

proliferation and apoptosis of the normal murine mammary epithelial cell

line HC11

Luisa A Helguero1, Malin Hedengran Faulds1, Jan-Ake Gustafsson1 and Lars-Arne Haldosen*,1

1Department of Medical Nutrition, Karolinska Institutet, NOVUM, Huddinge SE-141 86, Sweden

The mitogenic effect of 17b-estradiol (E2) on the breast ismediated by estrogen receptor alfa (ERa), hence ERaantagonists are effective in the treatment of breast cancer.The possible use of estrogen receptor beta (ERb) as atarget in treatment of breast cancer is under investigation.The mouse mammary cell line HC11 expresses both ERsand was used to study the role of the two receptors inproliferation. E2 had no effect on proliferation. The ERa-selective agonist 4,40,400-(4-propyl-(1H)-pyrazole-1,3,5-triyl)trisphenol (PPT) stimulated proliferation. TheERb-selective agonist 2,3-bis(4-hydroxy-phenyl)-propio-nitrile (DPN) inhibited cell growth and induced apoptosis.PPT upregulated while DPN downregulated cyclin D1and proliferating cell nuclear antigen (PCNA). Uponinhibition of ERa expression with RNA interference, E2caused a decrease in cyclin D1 and PCNA, and increasedapoptosis. When ERb expression was blocked, E2 inducedproliferation and cells gained the capacity to grow in softagar. In summary, in HC11 mammary epithelial cells,ERa drives proliferation in response to E2 while ERb isgrowth inhibitory. The lack of effect of E2 on HC11 cellgrowth is the result of the combined actions of ERa(proliferation) and ERb (apoptosis). We suggest that useof ERb agonists will be a useful addition in treatment ofbreast cancer, which, at present, is only aimed atinhibition of ERa.Oncogene (2005) 24, 6605–6616. doi:10.1038/sj.onc.1208807;published online 20 June 2005

Keywords: estrogen receptors; mammary epithelial cells;estradiol; estrogen receptor-selective ligands; prolifera-tion; apoptosis

Introduction

Estrogens, acting through estrogen receptor alfa (ERa)(Greene et al., 1986) and beta (ERb) (Kuiper et al.,1996), play a key role in mammary gland developmentand morphogenesis. Studies of ERa and ERb knockoutmice have confirmed the importance of ERa in ductal

elongation during puberty (Bocchinfuso et al., 2000) andrevealed a role for ERb in differentiation of theepithelium (Forster et al., 2002). Both receptors bind17b-estradiol (E2) with the same affinity, but they canactivate certain promoters differentially (Saville et al.,2000) and on some promoters ERb can behave as anegative modulator of ERa activity (Weihua et al., 2003;Strom et al., 2004). About 2/3 of breast tumors expressERa and 70% of these respond to the antiestrogentamoxifen (Clarke et al., 1996). Prolonged treatmentwith tamoxifen leads to resistance to the drug despite thecontinued presence of estrogen and progesterone recep-tors (Clarke et al., 1996). At some promoters, tamoxifenis an agonist in the presence of ERb and it is possiblethat ERb plays an antiproliferative, proapoptotic roleduring tamoxifen therapy (Iwao et al., 2000a, b; Speirset al., 2004).

In recent years, much effort has been invested inthe development of ERa- and ERb-specific agonistsand antagonists. Several such compounds have beendeveloped, and this has permitted dissection of thespecific functions of each receptor. One ERa-specificagonist, 4,40,400-(4-propyl-(1H)-pyrazole-1,3,5-triyl)tri-sphenol (PPT), is 410-fold more potent in binding toERa than ERb (Stauffer et al., 2000), whereas 2,3-bis(4-hydroxy-phenyl)-propionitrile (DPN) binds to ERbwith an affinity 72-fold higher than to ERa (Meyerset al., 2001). Both ligands induce expression of anestrogen response element (ERE)-luciferase reportergene with potencies similar to that of E2; hence, theyare considered to be ERa- and ERb-selective agonists,respectively.

In the present study, the HC11 mouse cell line wasused to evaluate the effect of estrogen on mammaryepithelial cells. This cell line was cloned from theCOMMA-D1 cell line obtained from the mammaryepithelium from a mid-pregnant BALB/c mouse (Ballet al., 1988). These cells have stem cell-like propertiessuch as growth factor- and hormone-dependent expres-sion of markers for different epithelial lineages (Deug-nier et al., 1999) and the ability to reconstitute partiallythe ductal epithelium in a cleared mammary fat pad(Humphreys and Rosen, 1997).

The present report is a study of the roles of ERaand ERb in E2-induced proliferation. The action ofeach receptor was delineated with the use of ERa- and

Received 31 January 2005; revised 28 April 2005; accepted 6 May 2005;published online 20 June 2005

*Correspondence: L-A Haldosen;E-mail: [email protected]

Oncogene (2005) 24, 6605–6616& 2005 Nature Publishing Group All rights reserved 0950-9232/05 $30.00

www.nature.com/onc

ERb-specific agonists and with ERa- and ERb-specificsmall inhibitory RNA (siRNA). We show that ERamediates proliferation while ERb is proapoptotic. Inaddition, we found that inhibition of ERb expressioncan lead to cellular transformation.

Results

ERa and ERb colocalize in HC11 mammary epithelialcells

We have previously demonstrated by Western blottingthat HC11 mammary epithelial cells express bothERa and ERb (Faulds et al., 2004). In this report, weused double immunofluorescence staining to studycoexpression of these receptors (Figure 1a–e). Highlevels of ERa surrounding the nucleus were detected inmost of the cells, (panel a). ERb was also detected inmost cells with higher intensity around the nucleus(panel b). Merged images are shown in panels c and e,where it is evident that although both receptorscolocalize throughout the cell, ERa has a higher nucleardistribution than ERb. Both receptors were found indefined areas of the nucleus in a speckle-like pattern.The extent of staining differed between cells. At present,the reason for this is unclear but could be due to aninfluence from cell–cell contact and cell cycle stage onER expression.

The staining surrounding the nucleus was studiedwith lamin B (red) as a marker for nuclear membranelocalization (Figure 1f ). In some cells, ERb (green)colocalized with this component of the nuclear lamina(yellow), indicating that at least ERb can be found onthe inner side of the nuclear envelope. These resultsshow that ERa and ERb colocalize in the nucleus(speckles), with a strong signal around the nuclearmembrane, which in some cases may be on the inner sideof the membrane and also in the cytoplasm.

Specificity of anti-ERa (MC-20) and anti-ERb (503)antibodies was evaluated on mammary glands fromERaþ /þ , ERa�/� and ERbþ /þ , ERb�/� mice,respectively (Figure 1g and h) and on ER-negative HeLacells (not shown). A specific nuclear staining andoccasional cytosolic signal along ducts was observedwith MC-20 only in ERaþ /þ (panel g) but not inERa�/� (panel g inset) or HeLa cells. 503 antibodygave both nuclear and cytosolic staining in ERbþ /þ(panel h) but not in ERb�/� mammary glands (panel h,inset) or HeLa cells.

Differential activation of ERa or ERb with selectiveligands results in opposing responses in terms ofproliferation and apoptosis

The effects of E2, the ERa-selective ligand PPT or theERb-selective ligand DPN on cell proliferation wereevaluated by cell counting following 48 h incubationwith 1 or 10 nM of each compound (Figure 2a). E2 didnot increase proliferation, with 10 nM PPT there was a50% increase in cell number (Po0.001) and, strikingly,

with DPN, cell number decreased 25–30% (Po0.001).A combination of PPT and DPN had similar effect asE2, resulting in no change in cell number.

Expression of the proliferation associated proteinscyclin D1 and PCNA was evaluated by Western blot24 h after exposure to 10 nM E2, PPT or DPN. CyclinD1 expression was up-regulated by both E2 and PPT,but in DPN-treated cells its levels were lower than thecontrol (Figure 2b). PCNA expression also decreasedfollowing DPN treatment. No change or a slightdecrease was observed with E2 and a slight increasewith PPT (Figure 2c). To further study if DPNtreatment was affecting exit from G1, we evaluated thenumber of cells in mitosis following 6 and 24 hincubation. Mitotic index was measured as cells inmitosis/total number of cells in cultures stained withhematoxylin (10 random fields were evaluated). It wassignificantly higher following PPT treatment (6 h: 270.6and 24h: 1.370.2; Po0.05); DPN decreased the numberof mitoses at an early time point but no significant effectwas observed after 24 h (6 h: 0.470.3; Po0.05 and 24h:1.070.4); no significant differences were observed withE2, although there was a tendency to decreased numberof mitoses at earlier time points (6 h: 0.670.3 and 24 h:0.970.4). Ki-67 was also evaluated; only DPN caused asignificant decrease after 6 h (62716 vs 91715%;Po0.01). Taken together, these results indicate thatE2 and PPT have a similar effect on cyclin D1expression while DPN plays an opposing role bytransiently preventing the cells from entering the cellcycle. The lack of effect of E2 on PCNA despiteinduction of cyclin D1 indicates that E2 does signal cellsto proliferate but that exit from G1 is inhibited. Theseresults imply that estrogen signaling via ERa and ERbacts at different steps of the cell cycle where ERa favorsentry into the cell cycle while ERb transiently inhibits it.

To further substantiate the effects of ER-selectiveligands on cell proliferation, we used the MTS assay,which determines metabolic activity by reduction oftetrazolium to a colored formazan product by dehy-drogenases. This is supposed to give an indirectestimation of cell proliferation. The effect of the threeligands was studied in a range of concentrations from0.1 to 100 nM. As shown in Figure 3, no effect wasobserved with E2, while a significant stimulatoryresponse of 40–50% was obtained with both PPTand DPN (100 nM, Po0.001). With a combination ofPPT plus DPN (PPTþDPN), instead of a synergisticstimulatory effect, there was no stimulation of cellmetabolism. We can speculate that ER homodimerformation is favored by binding of selective ligands andthis might induce expression of different sets of genesthan those induced by the ERa–ERb heterodimers.Thus, DPN increases cell metabolism but inhibits cellproliferation, whereas PPT increases both parameters.However, combined treatment with PPT and DPN, orexposure to E2, that is, simultaneous activation of ERaand ERb, results in lack of effect on either proliferationor metabolism. We can also conclude that, under ourexperimental conditions, the MTS assay does notreliably measure E2-induced proliferation.

ERa and ERb in cell proliferation and apoptosisLA Helguero et al

6606

Oncogene

Figure 1 ERa and ERb localization in HC11 cells. (a) ERa antibody (MC20) plus anti-rabbit IgG-TRITC. (b) ERb antibody (503)plus anti-chicken IgG-FITC. (c) Merged image from panels a and b. Magnification: � 20. (d) Staining of nuclei with DAPI (e). Mergedimage of ERa (red) and ERb (green) staining. Magnification: � 60. (f) Merged image of ERb (green) and lamin b (red). Magnification:� 60. (g, h) Wild-type mammary glands were used as positive controls for anti- ERa (MC-20) or anti- ERb (503) antibodies. Mammaryglands from ERa�/� or ERb�/� mice (inset) were used as negative controls for MC-20 and 503 antibodies, respectively.Magnification: � 40


6607

Oncogene

The negative effect on cell number observed withDPN could also be the result of an induction ofapoptosis. To test this hypothesis, the number ofapoptotic cells after 24 and 48 h incubation with 10 nM

E2, PPT or DPN was measured with the TUNELmethod (Figure 4). After 24 h, both E2 and DPNsignificantly increased apoptosis (Po0.05 and Po0.001,respectively) with highest values (four-fold increase overcontrol) observed following DPN treatment. With E2,the increase was 2.5-fold. There was no significantincrease in apoptosis after PPT treatment. Only DPNcaused a significant increase in apoptosis after 48 h(Po0.001).

During involution, mammary epithelial mass isreduced by apoptosis. Bcl-2 family of proteins play animportant role in this process and evidence pointstoward the regulation of some members of this family bysex steroid hormones (Schorr et al., 1999b). Thus, E2-and DPN-mediated apoptosis was also studied byanalysing the expression of the proapoptotic proteinBax and the E2-regulated antiapoptotic protein Bcl-x(Leung and Wang, 1999) between 2 and 24 h oftreatment (Figure 5). Bax expression was slightly higherthan control throughout the whole time period studied,independently of the ligand. However, a significantincrease was observed in cells treated with E2 or DPNfor 24 h. The expression pattern of Bcl-x, one of theinhibitory partners to Bax, was more complex. Activa-tion of ERb by DPN significantly decreased expressionof Bcl-x at 6 h and values remained low up to 18 h, whileduring the same time interval, incubation with PPTtended to keep Bcl-x levels high. E2 had the same effectas PPT at 6 h while it decreased Bcl-x levels at 18 h. At24 h after exposure to any of the ligands, Bcl-x washighly upregulated. The study of Bax and Bcl-x proteinlevels in HC11 cells indicates that the critical regulatoryperiod for Bax-mediated apoptosis is between 6 and

Figure 2 Effect of different ER ligands on cell proliferation. (a)HC11 cells were treated with 1 or 10 nM E2 (�), PPT (’), DPN(J) or a combination of equal amounts of PPTþDPN (&) inexperimental medium. After 2 days, cells were detached andcounted. Proliferation index was calculated comparing to theuntreated control set to 1. (b, c) Western blot analysis of whole-cellextracts from HC11 cells treated for 24 h with 10 nM E2, PPT orDPN. Membranes were blotted with anti-cyclin D1 antibody (b) oranti-PCNA antibody (c). The intensity of the bands was normal-ized to that of b-actin. Normalized values were compared to theuntreated control, arbitrarily set to 1. *Po0.01; **Po0.05;***Po0.001

Figure 3 Effect of different ER ligands on cell metabolism. HC11cells were plated on 96-well plates and treated with increasingconcentrations of E2 (�), PPT (’), DPN (J) or a combination ofequal amounts of PPTþDPN (&). The ligands were added to theexperimental medium (2% HS) and the incubation proceeded for48 h. Metabolism index, measured by MTS assay, was calculatedcomparing to the untreated control set to 1. *Po0.01; **Po0.05;***Po0.001

Figure 4 Effect of E2, PPT and DPN on apoptosis of HC11 cells.A concentration of 10 nM E2, PPT or DPN was added to theexperimental medium (2% HS) and 24 or 48 h later, cells werefixed. TUNEL-fluorescein assay was performed and nuclei werecounterstained with DAPI. A total of 10 random fields at � 40magnification were counted for TUNEL-positive cells/number oftotal nuclei. A representative experiment of three is shown.**Po0.05; ***Po0.001


6608

Oncogene

18 h. During this time period, Bax levels increased withall of the treatments. However, the apoptosis inhibitorBcl-x also increased in the case of PPT treatment butdecreased with E2 (18 h) or DPN (6 and 18 h). Theseresults are in agreement with the finding that gain offunction of Bcl-2, another antiapoptotic protein partnerto Bax, is more potent than loss of Bax function inregulating mammary epithelial cell survival in vivo(Schorr et al., 1999a).

ERa and ERb play different roles in mammary epithelialcellular proliferation and apoptosis

The role of ERa and ERb in E2-induced cell prolifera-tion was studied with RNA interference (RNAi) toblock alternatively ERa (siERa) or ERb (siERb). Thespecificity and efficiency of each siRNA were tested intransient transfection experiments in HC11 cells to-gether with YFP-mERa or GFP-mERb expressionvectors. Figure 6a shows that siERa and siERbdecreased ERa and ERb expression, respectively, whilethe control siRNAs with a point mutation (mutsiERaand mutsiERb) or the mock vector had no effect. Later,stably transfected HC11 cell lines (HC11-siERa, HC11-siERb or HC11-mock) were obtained and the efficiencyof each siRNA was evaluated by Western blot of whole-cell extracts (Figure 6b). The left panel shows thespecificity of anti-ERa (MC-20) or anti-ERb (14C8)antibodies. Neither ERa nor ERb was detected in NIH-3T3 or HeLa cell extracts. A 67 kDa band correspond-ing to ERa was observed in T47-D and HC11 cells. Aband of 58–60 kDa corresponding to ERb was observedin T47-D cells. Results from stably transfected cell lines

are shown in the right panel. ERa expression was lowerin HC11-siERa compared to HC11 or HC11-mock cells,while ERb expression was lower in HC11-siERbcompared to HC11 or HC11-mock cells.

Cell proliferation was evaluated on HC11 control andsiRNA stably transfected cells treated with 10 nM E2(Figure 7a, left panel). Inhibition of ERa expression(HC11-siERa) had a negative effect on cell number.That is, when E2 was bound to ERb, cell numberdecreased by about 30% (Figure 7a, left panel; Po0.05).On the contrary, E2 increased cell number by 50%(Po0.05) in HC11-siERb. No differences between thecontrol HC11 and HC11-mock cells were observed. Torule out the possibility of artifacts due to siRNAexpression per se, mutsiERa and mutsiERb weretransiently transfected into HC11 cells. Following this,cells were treated with 10 nM E2 and counted 48 h later(Figure 7a, right panel). Similar results as those withstably transfected cells were obtained in HC11 cellstransiently transfected with siERa or siERb but noeffect was evident with the mutated siRNAs. Theseobservations support the idea that stimulation of theERb pathway decreases cell number, while the oppositeis observed when ERa is activated.

Cyclin D1 and PCNA expression was evaluated byWestern blot of whole-cell extracts from each stable cellline, following 24 h treatment with 10 nM E2 (Figure 7b).As previously shown, an increase of cyclin D1 levels wasobserved after E2 incubation in HC11 or HC11-mockcells (in accordance with Figure 2b). In HC11-siERacells, E2 downregulated cyclin D1 expression(Figure 7b). Surprisingly, in E2-treated HC11-siERbcells, cyclin D1 was not upregulated. Expression ofPCNA was slightly downregulated in HC11 or mock-transfected cells, and HC11-siERa cells. PCNA was notdownregulated in HC11-siERb (Figure 7b). Theseobservations are consistent with the fact that E2 actingthrough ERb (HC11-siERa cells) inhibits proliferation.

The effect of siERa and siERb on apoptosis of E2-treated HC11 cells was evaluated with TUNEL assay(Figure 7c). Apoptosis index did not vary betweenuntreated wild-type HC11 cells and the stable cell lines(not shown). Conversely, following 48 h incubation withE2, a significant (Po0.001) increase was observed inHC11-siERa cells and lower levels in HC11-siERb cellscompared to the untreated control.

Decreased ERb expression favors epithelialtransformation

Impaired regulation of proliferation/apoptosis pathwaysis one factor influencing cell transformation. Conse-quently, effect of loss of ERb expression on HC11 celltransformation was investigated. Figure 8a shows arepresentative picture of growth in soft agar, whereanchorage-independent growth was evident in HC11-siERb cells (arrows) but not in HC11, HC11-siERa ormock-transfected cells (only HC11 is shown for com-parison). In the case of HC11-siERb cells, a significantincrease in the percentage of colonies was observedfollowing 15 days treatment with 10 nM E2 (Figure 8b,

Figure 5 Expression of Bcl-2 family members. After 2, 6, 18 and24 h treatment with 10 nM E2, PPT or DPN, western blots ofwhole-cell extracts were developed with anti-Bax or Bcl-xantibodies. The intensity of the bands was normalized to that ofb-actin in the same sample. Normalized values were compared tothe untreated control, arbitrarily set to 1. *Po0.01


6609

Oncogene

Po0.05). Treatment with 10 nM E2 together with 10 nM

4OH-tamoxifen (Tam) or ICI 182 780 (ICI) partiallyinhibited colony formation by HC11-siERb cells(Figure 8b). No effect was observed after antiestrogentreatment of wild-type HC11 or the other stable celllines (not shown). The effect of 10 nM E2 on the numberof cells per colony was also studied in HC11-siERb cells.The most frequent value observed in 75% of thecolonies in the untreated group was six cells per colony(maximum¼ 14) while E2 increased the cell number to11.5 (maximum¼ 20). When cells were treated with acombination of 10 nM E2 and 10 nM Tam or ICI, thenumber of cells decreased to 5.5 (maximum¼ 8) and 7.5(maximum¼ 11), respectively. Thus, E2 exerted asignificant tendency to increase the number of cells ineach colony, which was blocked by Tam or ICI(ao0.05), suggesting that these effects are ER mediated.The results shown here indicate that an imbalance

between ERa and ERb expression due to reduced ERblevels can lead to cell transformation.

Discussion

The mouse mammary cell line, HC11, is known toexpress both ERa and ERb (Faulds et al., 2004) andreflects the situation in the late pregnant and lactatingstates, where the mammary glands are resistant to theproliferative effects of estrogen (Shyamala and Ferenc-zy, 1982), but express both receptors. The mechanismsresponsible for resistance to E2-induced proliferationare not known but there have been suggestions ofreduction in coactivators and increase in corepressors.In this study, we show that both ERa and ERb arefunctional and can be activated in HC11 cells. Thereason for resistance to E2-induced proliferation is the

Figure 6 Inhibition of ERa and ERb expression by si RNAs. (a) HC11 cells were transiently transfected with YFP-ERa alone or incombination with MOCK, siERa or mutsiERa vectors, as described in Materials and methods. Another group of cells was transfectedwith GFP-ERb alone or together with MOCK, siERb or mutsiERb vectors. After 18 h, cells were fixed in 10% formalin and imagescaptured with a confocal microscope (Leica). (b) ER expression in stably transfected HC11 cells. Whole-cell extracts were separated ina 10% SDS–PAGE, blotted to PVDF membranes and ERa or ERb expression detected with MC-20 or 14C8 antibodies, respectively.The left panel shows the specificity of each antibody. NIH-3T3 and HeLa cells were used as negative and T47-D cells as positivecontrols for ERa (MC-20 antibody) and ERb (14C8 antibody). The right panel shows the results obtained with stably transfected celllines. Blots are representative of two independent experiments


6610

Oncogene

concomitant activation of ERa and ERb, which haveopposing cellular responses with regard to proliferationand apoptosis.

We found that ERa and ERb are expressed by allHC11 cells and colocalize in the nucleus and perinuclearregion. Despite this, HC11 cell growth was neitherstimulated nor inhibited by E2 at concentrationsranging from 0.1 to 100 nM. The ERa-selective ligandPPT (Stauffer et al., 2000) and the ERb-selective ligandDPN (Meyers et al., 2001), which, based on reporterassays, were both considered to be agonists with asimilar potency to that of E2, were used to study theeffects exerted by one or the other receptor. Uponexposure of the cells to PPT, there was an increase in cellnumber of nearly 50% while DPN caused a decrease of30% compared to the untreated cells. These resultsindicate that activation of ERb leads to inhibition ofproliferation and increase in apoptosis. Both selectiveligands had positive effects on cell metabolism measuredas dehydrogenase activity. Activation of ERa indepen-dently of ERb, and vice versa, increased cell metabolism,but activation of both receptors simultaneously (E2 or

PPTþDPN) had no effect. One hypothesis in line withthe idea of selective estrogen receptor modulators(SERMs), is that – depending on the type of ligand –different sets of coactivator/corepressor proteins will berecruited to the transcription complexes, resulting intranscription of different genes depending on the type ofER dimer-coregulator complex formed. In fact, ligandssuch as estrone and genistein induce SRC-1 recruitmentto ERb with much higher efficiency compared to ERa(Margeat et al., 2003) and selective ligands such as PPTshow some quantitative differences in their ability torecruit members of the SRC/p160 family (Kraichelyet al., 2000). In addition, our results show that the MTSassay cannot be used to measure accurately estrogenicregulation of cell proliferation.

ERa and ERb activate transcription from consensusand nonconsensus ERE reporter genes and it haspreviously been shown that the transactivation potencyof the two receptors differs at different EREs (Hall andKorach, 2002) and that this potency is greatly influencedby the structure of the ligand. The proliferative effects ofE2 are mediated not only by transcriptional activation

Figure 7 Inhibition of ERa or ERb results in a differential response to E2. (a) Wild-type HC11 cells and the stably siERa- or siERb-transfected cells (left panel) were seeded in 24-well plates and treated with or without 10 nM E2 supplied in experimental medium. After48 h cells were counted. Proliferation index was calculated comparing the effect exerted by E2 in each cell line to its untreated control.Mean7s.e. of three independent experiments is shown. The right panel shows, HC11 cells transiently transfected with siERa, siERband the control mutated sequences (mutsiERa and mutsiERb). Following 48 h in the presence of E2, cells were counted, normalized tofirefly luciferase activity and, for each transfection, values were related to the untreated group. Mean7s.e. of two independentexperiments carried out in sextuplicate is shown. Differences between treatments and control were analysed using Student’s t-test.**Po0.05; ***Po0.001. (b) Cyclin D1 and PCNA expression in siERa or siERb stably transfected HC11 cell lines after 24 hincubation with 10 nM E2. b-Actin was used as loading control. Blots are representative of three experiments. (c) TUNEL-assayfollowing 48 h incubation with 10 nM E2. TUNEL positive cells/total number of cells in 10 random fields at � 40 magnification werecounted; apoptosis index was calculated comparing to each untreated cell line. **Po0.05; ***Po0.001. The experiment is arepresentative of two


6611

Oncogene

from ERE promoters but also by the interaction of ERswith AP-1 sites via protein–protein interactions of ERwith AP-1 binding proteins like jun and fos (Kushneret al., 2000). The two receptors can have opposingactivities on AP-1-driven promoters, such as the cyclinD1 promoter (Liu et al., 2002) where ERa is an activatorwhile ERb is repressive. Our results with HC11 cellsindicate that the opposing activities of ERa and ERb onAP-1 sites are responsible for the resistance to E2-induced proliferation. As reported in human breastcancer cell lines (Frasor et al., 2003), E2 and PPTincreased cyclin D1 expression in HC11 cells while DPNreduced the level of cyclin D1. Interestingly, PCNAexpression as well as mitotic index was increased onlywith PPT treatment but not with E2, and a decrease ofthese parameters was observed with DPN. These resultssuggest a stimulation of cyclin D1 promoter by ERa anda block in G1 downstream of cyclin D1 by ERb.

Treatment of HC11 cells with 10 nM E2 during 24–48 h influenced both proliferation and apoptosis. Therewas increased cyclin D1 expression, as well as increasedapoptosis. With TUNEL assay, it was clear that DPN

induced apoptosis while PPT did not. We conclude fromthese data that there are two E2-activated pathways,that is, proliferation (ERa) and apoptosis (ERb), andactivation of both pathways by E2 is responsible for thelack of effect on cell number.

Throughout postlactation involution of the mammarygland, apoptosis is strongly regulated by expression ofBcl-2 family proteins (Schorr et al., 1999a, b). In ourstudy, the role of two members of this family wasinvestigated. At 18 and 24 h, the proapoptotic proteinBax was upregulated by E2 and DPN. However, theantiapoptotic protein Bcl-x was downregulated between6 and 18 h by DPN and not by PPT. Thus, in the first24 h after administration of ligands, apoptosis prevailsin DPN-treated cells. Apoptosis remained high in DPN-treated cells 48 h after treatment although Bax and Bcl-xremained high. The reason for sustained apoptosisdespite increase in Bcl-x has not been investigated, butit is possible that at these later time points othermembers of the Bcl-2 family may have becomeimportant players, or that the initial downregulationof Bcl-x is enough to start the apoptosis process. In the

Figure 8 Inhibition of ERb expression leads to anchorage-independent growth. (a) Representative pictures of growth in 0.7% nobleagar with 10 nM E2. Arrows show colonies formed after 15 days. Magnification: � 10, inset � 40. (b) Number of colonies formedfollowing 15 days of exposure to 10 nM E2, E2 plus same concentration of Tam or ICI. Percent (%) colonies formed/total cells: 10random fields at � 10 magnification were counted for each treatment and mean7s.d. of duplicates is shown. The figure isrepresentative of two independent experiments. þ þ þPo0.001 vs HC11, HC11-mock or HC11-siERa; ***Po0.001 and **Po0.05 vsHC11-siERb plus E2; #Po0.01 vs ICI


6612

Oncogene

normal mouse mammary gland, 48 h after involution,both Bax and Bcl-x mRNA were higher than duringlactation (Schorr et al., 1999b).

The antiapoptotic protein Bcl-2 is an E2-induced gene(Zhang et al., 1999). Its presence in breast cancer ishighly correlated with ER expression and favorableprognosis (Gasparini et al., 1995; Elledge et al., 1997).This is in agreement with the inhibitory effect of the ERantagonist tamoxifen on Bcl-2 expression (Zhang et al.,1999) as shown in the MCF-7 cell line. Bcl-2 has beenshown to be an ERb-repressed gene in the ventralprostate (Imamov et al., 2004); so it is possible that theeffect of tamoxifen could be due to its action as an ERbagonist at AP-1-regulated promoters (Kushner et al.,2000).

Blocking ERa or ERb expression with siRNAsconfirmed what was found with ERa- or ERb-specificligands. Inhibition of ERa expression in HC11-siERacells decreased proliferation in an E2-dependent way,while blocking ERb expression in HC11-siERb cellscaused an increase in cell number. These results implythat, in a nonmalignant cell line, the ratio of ERa toERb in cells determines whether or not E2 will induceproliferation. On the other hand, the inhibitory responseachieved by E2 in HC11-siERa cells was not only due todecreased proliferation but also due to, as a consequenceof reduced ERa levels, E2 binding to ERb andactivation of apoptosis, as was as seen with the selectiveERb ligand DPN.

To date, there is no consensus on the role of ERb inbreast cancer (Fuqua et al., 2003). One study shows thatERa mRNA is upregulated in ER positive breastcancers while ERb is downregulated (Iwao et al.,2000a, b). Other studies, using RT–PCR, have shownthat ERb was expressed at much lower rates than ERa(Dotzlaw et al., 1997; Speirs et al., 1999). Shaaban et al.analysed ERa and ERb protein expression in 53 normalbreasts and compared the results with those from acohort of histologically distinct breast lesions ofdifferent prognostic risk (Shaaban et al., 2003). Theauthors suggested that the tissue concentration, relativeoccurrence and/or interaction of these two types of ERsmay play an important role in modulating mammarytumorigenesis.

The results reported in this study, in which E2 actingthrough ERb has a negative effect on epithelial cellgrowth, are in accordance with findings by Omoto et al.(2003) in MCF-7 cells and Strom et al. (2004), whorecently showed that overexpression of ERb in T47-Dcells reduces cell proliferation with a concomitantdecrease of many regulatory components of the cellcycle. The inhibitory effects of E2 have also beenreported by Lanari’s group in a well-established mousemammary cancer model. These tumors of ductalhistological type express high levels of estrogen andprogesterone receptors (Molinolo et al., 1987) andregress with E2 treatment administered subcutaneouslydue to an increase in apoptosis and the cell cycleinhibitory proteins p21waf, p27 and p53 (Vanzulli et al.,2002). Growth inhibition in response to E2 was alsoobserved in primary cultures established from these

tumors (Lamb et al., 2003). The inhibitory effects of E2are in line with the observation that DPN inducesapoptosis in HC11 cells.

HC11 cells do not form tumors when transplantedsyngeneically, in fact they have the ability to repopulatepartially a cleared mammary fat pad giving origin tonormal ductal structures. However, when ERb expres-sion was blocked, HC11-siERb cells were able to formcolonies in soft agar, indicating that HC11 cells weretransformed. This result could be explained by thefindings of Forster et al. on mammary glands frompregnant ERb knockout mice where decreased levels ofadhesion molecules such as E-cadherin, connexin 32,occludin and integrin a2 as well as altered tight junctions(Forster et al., 2002), indicative of a less differentiatedphenotype, were observed. Studies in MCF-7 breastcancer cell lines transfected with ERb showed a decreasein the number of colonies growing in soft agar (Omotoet al., 2003). Interestingly, during carcinogenesis, therelative ratio of ERa/ERb expression was also reportedto increase (Leygue et al., 1998).

There are few nonmalignant cell lines reported thatexpress both ERs. This makes the HC11 cell line a goodmodel for studies on the interplay between ERa andERb in epithelial growth. In this study, we focused onthe proliferation response to E2 and found that the ‘yin-yang’ relationship between both ERs not only prevailsat the promoter level as reported earlier for cyclin D1.Estrogen also simultaneously activates different signal-ing pathways that, probably depending on ERa andERb levels, modulate the normal epithelial response toE2, which may be altered at some stage of breast cancerdevelopment.

Materials and methods

Hormones, reagents and antibodies

RPMI 1640 medium without phenol red, L-glutamine, trypsin-EDTA and gentamycin were from Gibco, BRL. Epidermalgrowth factor (EGF), insulin, E2, DAPI, diaminobenzidine(DAB) and gelded horse serum (HS) were from Sigma. Fetalbovine serum (FBS) was from Integro (Dieren, Holland). PPTwas kindly provided by KaroBio, Sweden and DPN was fromTocris. The primary antibodies used were rabbit anti-mouseERa (MC-20, Santa Cruz Biotechnology, Santa Cruz, CA,USA), the noncommercial chicken anti-human ERb (503)antibody developed in our own laboratory, anti-human laminB (Santa Cruz), rabbit anti-human cyclin D1 (H-295, SantaCruz), mouse anti-human PCNA (PC-10, Santa Cruz), rabbitanti-human Bax (Upstate Biotechnology), rabbit anti-humanBcl-x (B22630, Transduction Labs), mouse anti-human ERb(14C8, GeneTex) and anti-mouse b-actin antibody (ab8227,GeneTex). The secondary antibodies were goat anti-rabbit-TRITC (Sigma) or anti-chicken-FITC (Zymed).

Cell culture

HC11 cells were grown and maintained in complete medium(RPMI 1640, L-glutamine, 50 mg/ml gentamycin, 10% FBS,10 ng/ml EGF and 5mg/ml insulin). After 24 h attachment, themedium was changed for experimental medium (2% HS, 5 mg/


6613

Oncogene

ml insulin and L-glutamine) with or without the addition ofE2, PPT or DPN.

Note on the use of HS: Preliminary experiments indicatedthat 2% steroid stripped FBS (ssFBS) was not a good choice ofmedium to study proliferation in HC11 cells. The cells did notgrow well under these conditions and the results were notreproducible. Cell number was 3875% (n¼ 3) lower with 2%ssFBS compared to 2% HS; cell metabolism was 6474%(n¼ 4) lower. On the other hand, metabolic rate was 2872%(n¼ 4) lower with 2% HS compared to the same concentrationof normal FBS. Based on the fact that results with HS werereproducible and that E2 levels in HS are lower than in FBS(8 pM E2 in HS vs 134 pM in FBS) (Visser-Wisselaar et al.,1997), we decided to use 2% HS in order to have a goodwindow to see stimulatory effects.

Immunofluorescence

Cells were grown on glass microscope coverslides. At 70%confluence experimental medium was added. 24 h later, thecells were fixed in 10% buffered formalin and antigen wasretrieved in 10mM citrate buffer, pH 6.5, for 10min. Cells werepermeabilized with 0.5% Triton X-100 and unspecific bindingwas blocked with 10% HS in 0.1% Tween 20–PBS. Afterincubation with the antibodies, the slides were mounted inanti-fading medium (Vectashield Hardset medium, VectorLabs) and visualized in a Leica confocal microscope.

Immunohistochemistry

Paraffin-embedded tissues were hydrated through decreasingethanol washes, stained with the indicated primary antibodiesand the signal visualized with biotin–streptavidin–peroxidasemethod (ABC, Vector) and DAB substrate. In selectedexperiments, cells were grown on coverslides to 70% con-fluence, treated with 10 nM E2, PPT or DPN for 6 and 24 h andstained with Ki-67 antibody (Dako). Number of Ki-67 positivenuclei/total nuclei was counted in 10 random fields at� 20magnification. Sections were counterstained with hematoxylin.

Cell counting

A total of 10 000 cells were seeded in 24-well plates in completemedium. 24 h later, the medium was replaced by experimentalmedium with or without hormone addition and, after 48 h,cells were detached and counted in a Burker chamber. Theexperiments were conducted in sextuplicate. Proliferationindex was calculated comparing to the untreated controlwithin the same experiment, which was arbitrarily set to 1.Except where indicated, the mean7s.e. of at least fourexperiments is shown.

MTS assay

A total of 1000 cells were seeded in complete medium on 96-well plates. After 24 h, the hormones were added in experi-mental medium. The cells were treated for 48 h with onemedium change every 24 h and cell metabolism was measuredwith the MTS reagent (Promega, Madison, WI, USA)according to the manufacturer’s instructions. Reduction ofthe MTS tetrazolium compound was measured at 490 nmusing a Spectramax 250 spectrophotometer (Molecular De-vices Corporation). The experiments were conducted inoctuplicate. The proliferation index was calculated for eachtreatment comparing to the untreated control within the sameexperiment, which was arbitrarily set to 1. The mean7s.e. of atleast six experiments is shown.

TUNEL assay

Cells were grown on glass microscopy coverslides. At 70%confluence, they were treated with 10 nM E2, PPT or DPN inexperimental medium. After 24 or 48 h, the cells were fixed in10% buffered formalin. TUNEL assay was performedaccording to the manufacturer’s protocol (Roche Diagnostics,Sweden). Apoptosis index was quantified by counting thenumber of apoptotic nuclei/number of total nuclei stained withDAPI; 10 random fields at � 40 magnification, with approxi-mately 30–40 cells, were counted for each treatment. Apoptosisindex was calculated comparing to the control arbitrarily setto 1. Unless stated otherwise, the mean7s.d. of an experimentrepresentative of three is shown.

Whole-cell extracts

Cells were washed with ice-cold PBS, spun in a microcentrifugeand resuspended in lysis buffer (1% NP-40, 50 mM Tris-HCl,pH 7.5, 140mM NaCl, 2 mM EDTA, 10 mg/ml leupeptin, 10 mg/ml pepstatin, 0.2U/ml aprotinin, 1mM PMSF and 1mM

Na3VO4). The lysate was incubated for 20 min on ice andcentrifuged at 20 000 g for 20min at 41C. Protein quantifica-tion was performed using Bradford reagent (BioRad).

Western blotting

Whole-cell extracts (75 mg of protein) were resolved on 10 or12% SDS–PAGE and transferred onto a PVDF membrane bysemidry blotting. The membranes were blocked with 5% (w/v)milk protein dissolved in PBS and incubated overnight at 41Cwith the indicated primary antibodies. The secondary anti-bodies were coupled to horseradish peroxidase (Sigma). Theluminescent signal was detected with the enhanced chemilu-minescence kit (ECL, Amersham). Band intensity wasquantified with Density One software (BioRad). Differenceswere calculated first by normalizing the intensity of each bandto that of b-actin in the same sample; thus, intensity variationswill be independent of loading efficiency. Second, folddifference was calculated setting the untreated (normalized)control to 1. Experiments were repeated at least times timesand mean7s.d. is shown in bar graphs.

Construction of siRNAs

pSUPER (Oligoengine, Seattle, WA, USA) was used toexpress the siRNAs. The sequences were chosen with ‘siRNASelection Program’ (Whitehead Institute for BiomedicalResearch, 2003). The custom-designed oligos were fromDNA Technology A/S, Denmark. The sequence to blockERa (siERa) was 50-GCTCCTGTTTGCTCCTAAC-30

(homologous to bp 1395–1417 of the mouse ERa cDNA)and to block ERb (siERb) was 50-GTGCCAGCGAGCAGGTGCA-30 (homologous to bp 472–494 of the mouseERb cDNA). The specificity of these oligos was controlled byinserting a point mutation in base 9 (T for G in mutsiERa andG for T in mutsiERb). Generation of the plasmids was carriedout following the manufacturer’s suggestions (oligoengine.com).

Transient transfections

HC11 cells were grown in complete medium until 50%confluence, and then transfection was performed with Fugene(Roche). A double transient transfection was carried out tocontrol the blockade efficiency of each siRNA: 1 mg pCMX-YFP-mERa (YFP-mERa) with 0.5 mg siERa or mutsiERa;alternatively the same amounts of pCMX-GFP-mERb


6614

Oncogene

(GFP-mERb) with siERb or mutsiERb. The cells were grownfor 24 h, fixed in 10% buffered formalin and differencesbetween YFP or GFP intensity were evaluated by observationof random fields under the same microscope settings. Tocontrol the specificity of siRNAs in the response to E2 in cellcounting experiments, cotransfection of 0.5 mg of eachpSUPER-siRNA plasmid together with 100 ng of pGL3-control vector coding the firefly luciferase gene (Promega)was performed. Cells were simultaneously transfected andtreated with or without 10 nM E2. After 2 days, cells werecounted and number of cells/ml was normalized to luciferaseactivity in each well. Two independent experiments werecarried out, each comprising sextuplicates for each treatment.Following normalization, proliferation index was calculatedcomparing to the untreated control within the same experi-ment, which was arbitrarily set to 1. The mean7s.e. of twoexperiments is shown.

Stable transfections

HC11 cells were cotransfected with 10 mg of the describedpSUPER-siRNA vectors and the same amount of pEF6/V5-HisC (Invitrogen, with blasticidin resistance). The transfectionmethod was calcium phosphate precipitation as describedpreviously (Petersen and Haldosen, 1998). Cells were routinelymaintained in complete medium plus blasticidin S (15 mg/ml;Invitrogen). The experiments were performed in the absence ofselection antibiotic.

Growth in soft agar

Six-well plates were prepared by adding a 2 ml bottom layer of1.3% noble agar dissolved in RPMI 1640 medium. Aftersolidification, 0.7% noble agar was diluted in experimentalmedium and 1� 105 cells were added when the temperaturereached 401C. The mix was poured on top of the bottom layer.Next, 200ml of experimental medium with or without 10 nM E2was added and replaced every 2 days. The cultures were grownin a humidified chamber at 371C with 5% CO2. Anchorage-

independent growth was quantified after 15 days by countingnumber of colonies with more than four cells and number ofcells/colony in 10 random fields at � 10 magnification.Experiments were performed in duplicate and two independentexperiments were carried out.

Statistical analysis

Results from treatment with E2, PPT and DPN (comparing tountreated controls) were analysed with one-way ANOVA andTukey’s multiple post test. The experiments with siRNA-transfected cells, where the effects of E2 on each cell line werecompared to untreated controls, were analysed with Student’st-test. Statistical significance of differences between number ofcolonies grown in soft agar was evaluated by one-wayANOVA and Tukey’s multiple post test; differences betweennumber of cells/colony were analyzed by w2 test for trend witha 95% confidence interval.

AbbreviationsERa, estrogen receptor alfa; ERb, estrogen receptor beta; E2,17b-estradiol; PPT, 4,40,400-(4-propyl-(1H)-pyrazole-1,3,5-triyl)trisphenol; DPN, 2,3-bis(4-hydroxy-phenyl)-propionitrile;ERE, estrogen response element; HS, horse serum; PCNA,proliferating cell nuclear antigen; RNAi, RNA interference;siRNAs, small inhibitory RNAs; SERMs, selective estrogenreceptor modulators.

Acknowledgements

We are grateful to Dr Margaret Warner for encouragingdiscussions throughout the development of the manuscript, DrMichel Tujague for his assistance with confocal Imaging, CissiGardmo for help with transient transfections and Dr KatarinaPettersson, who kindly provided pCMX-YFP-mERa andpCMX-GFP-mERb vectors. This work was supported byfunds from Swedish Cancer Fund, Magnus Bergvalls Founda-tion and Karolinska Institutet.

References

Ball R, Friis RR, Schoenenberger CA, Doppler W and GronerB. (1988). EMBO J., 7, 2089–2095.

Bocchinfuso WP, Lindzey JK, Hewitt SC, Clark JA, Myer PH,Cooper R and Korach KS. (2000). Endocrinology, 141,

2982–2994.Clarke R, Skaar T, Leonessa F, Brankin B, James M, Brunner

N and Lippman ME. (1996). Cancer Treat. Res., 87,

263–283.Deugnier MA, Faraldo MM, Rousselle P, Thiery JP and

Glukhova MA. (1999). J. Cell Sci., 112, 1035–1044.Dotzlaw H, Leygue E, Watson PH and Murphy LC. (1997).

J. Clin. Endocrinol. Metab., 82, 2371–2374.Elledge RM, Green S, Howes L, Clark GM, Berardo M,

Allred DC, Pugh R, Ciocca D, Ravdin P, O’Sullivan J,Rivkin S, Martino S and Osborne CK. (1997). J. Clin.Oncol., 15, 1916–1922.

Faulds MH, Olsen H, Helguero LA, Gustafsson JA andHaldosen LA. (2004). Mol. Endocrinol., 18, 412–421.

Forster C, Makela S, Warri A, Kietz S, Becker D, Hultenby K,Warner M and Gustafsson JA. (2002). Proc. Natl. Acad. Sci.USA, 99, 15578–15583.

Frasor J, Danes JM, Komm B, Chan KC, Lyttle CRand Katzenellenbogen BS. (2003). Endocrinology, 144,

4562–4574.

Fuqua SA, Schiff R, Parra I, Moore JT, Mohsin SK, OsborneCK, Clark MG and Allred DC. (2003). Cancer Res., 63,2434–2439.

Gasparini G, Barbareschi M, Doglioni C, Palma PD, MauriFA, Boracchi P, Bevilacqua P, Caffo O, Morelli L andVerderio P. (1995). Clin. Cancer Res., 1, 189–198.

Greene GL, Gilna P, Waterfield M, Baker A, Hort Y andShine J. (1986). Science, 231, 1150–1154.

Hall JM and Korach KS. (2002). J. Biol. Chem., 277,44455–44461.

Humphreys RC and Rosen JM. (1997). Cell Growth Differ., 8,839–849.

Imamov O, Morani A, Shim GJ, Omoto Y, Thulin-AnderssonC, Warner M and Gustafsson JA. (2004). Proc. Natl. Acad.Sci. USA, 101, 9375–9380.

Iwao K, Miyoshi Y, Egawa C, Ikeda N and Noguchi S.(2000a). Int. J. Cancer, 88, 733–736.

Iwao K, Miyosh Y, Egawa C, Ikeda N, Tsukamoto F andNoguchi S. (2000b). Cancer, 89, 1732–1738.

Kraichely DM, Sun J, Katzenellenbogen JA and Katzenellen-bogen BS. (2000). Endocrinology, 141, 3534–3545.

Kuiper GG, Enmark E, Pelto-Huikko M, Nilsson S andGustafsson JA. (1996). Proc. Natl. Acad. Sci. USA, 93,

5925–5930.


6615

Oncogene

Kushner PJ, Agard DA, Green GL, Scanlan TS, Shiau AK,Uht RM and Webb P. (2000). J. Steroid Biochem. Mol. Biol.,74, 311–317.

Lamb CA, Helguero LA, Fabris V, Colombo L, MolinoloAA and Lanari C. (2003). Breast Cancer Res. Treat., 79,25–35.

Leung LK and Wang TT. (1999). Br. J. Cancer, 81, 387–392.Leygue E, Dotzlaw H, Watson PH and Murphy LC. (1998).

Cancer Res., 58, 3197–3201.Liu MM, Albanese C, Anderson CM, Hilty K, Webb P, Uht

RM, Price RH, Pestell RG and Kushner PJ. (2002). J. Biol.Chem., 277, 24353–24360.

Margeat E, Bourdoncle A, Margueron R, Poujol N, CavaillesV and Royer C. (2003). J. Mol. Biol., 326, 77–92.

Meyers MJ, Sun J, Carlson KE, Marriner GA, Katzenellenbo-gen BS and Katzenellenbogen JA. (2001). J. Med. Chem., 44,4230–4251.

Molinolo AA, Lanari C, Charreau EH, Sanjuan N andPasqualini CD. (1987). J. Natl. Cancer Inst., 79, 1341–1350.

Omoto Y, Eguchi H, Yamaguchi Y and Hayashi S. (2003).Oncogene, 22, 5011–5020.

Petersen H and Haldosen LA. (1998). Exp. Cell Res., 243,

347–358.Saville B, Wormke M, Wang F, Nguyen T, Enmark E, Kuiper

G, Gustafsson JA and Safe S. (2000). J. Biol. Chem., 275,

5379–5387.Schorr K, Li M, Bar-Peled U, Lewis A, Heredia A, Lewis B,

Knudson CM, Korsmeyer SJ, Jager R, Weiher H and FurthPA. (1999a). Cancer Res., 59, 2541–2545.

Schorr K, Li M, Krajewski S, Reed JC and Furth PA. (1999b).J. Mammary Gland Biol. Neoplasia, 4, 153–164.

Shaaban AM, O’Neill PA, Davies MP, Sibson R, West CR,Smith PH and Foster CS. (2003). Am. J. Surg. Pathol., 27,

1502–1512.Shyamala G and Ferenczy A. (1982). Endocrinology, 110,

1299–1256.Speirs V, Carder PJ, Lane S, Dodwell D, Lansdown MR and

Hanby AM. (2004). Lancet Oncol., 5, 174–181.Speirs V, Parkes AT, Kerin MJ, Walton DS, Carleton PJ, Fox

JN and Atkin SL. (1999). Cancer Res., 59, 525–528.Stauffer SR, Coletta CJ, Tedesco R, Nishiguchi G, Carlson K,

Sun J, Katzenellenbogen BS and Katzenellenbogen JA.(2000). J. Med. Chem., 43, 4934–4947.

Strom A, Hartman J, Foster JS, Kietz S, Wimalasena J andGustafsson JA. (2004). Proc. Natl. Acad. Sci. USA, 101,1566–1571.

Vanzulli S, Efeyan A, Benavides F, Helguero LA, Peters G,Shen J, Conti CJ, Lanari C and Molinolo A. (2002).Carcinogenesis, 23, 749–758.

Visser-Wisselaar HA, Van Uffelen CJ, Van Koetsveld PM,Lichtenauer-Kaligis EG, Waaijers AM, Uitterlinden P,Mooy DM, Lamberts SW and Hofland LJ. (1997).Endocrinology, 138, 1180–1189.

Weihua Z, Andersson S, Cheng G, Simpson ER, Warner Mand Gustafsson JA. (2003). FEBS Lett., 546, 17–24.

Zhang GJ, Kimijima I, Onda M, Kanno M, Sato H, WatanabeT, Tsuchiya A, Abe R and Takenoshita S. (1999). Clin.Cancer Res., 5, 2971–2977.


6616

Oncogene

Estrogen receptors alfa (ERα) and beta (ERβ) differentially regulate proliferation and apoptosis of the normal murine mammary epithelial cell line HC11

Documents