Breast Cancer Research and Treatment 68: 127–138, 2001.© 2001 Kluwer Academic Publishers. Printed in the Netherlands.

Report

Induction of apoptosis in mammary gland by a pure anti-estrogen ICI182780

K. B. Lim, C. Y. Ng, C. K. Ong, C. S. Ong, E. Tran, T. T. T. Nguyen, G. M. T. W. Chan, andH. HuynhLaboratory of Molecular Endocrinology, Division of Cellular and Molecular Research, National Cancer Centre ofSingapore, Singapore

Key words: apoptosis, anti-estrogens, mammary gland

Summary

The antiestrogen, ICI 182780 (ICI) proves to be clinically useful for the treatment of estrogen receptor positivebreast tumours. We report the assessment of the in vivo and in vitro effects of ICI on apoptosis of breast epithelialcells. In vivo, administration of rats with ICI for 3 weeks resulted in a reduction in the size of the lobular structureswith the rate of mammary epithelial apoptosis equivalent to 10, 35 and 45% on treatment with 1, 1.5 and 2 mgICI per kg body weight, respectively. Concomitantly, these treatment led to a 2.0-, 2.2- and 2.5-fold increase inBax. Similar elevations were also observed in Bad levels which increased 1.7-, 2.6- and 2.7-fold respectively inthe ICI treatment as compared to controls. This also resulted in a dose dependent decrease in Bcl-2 and Bcl-xLprotein expressions. Growth inhibition and induction of apoptosis were also observed in the MCF-7 cells followingin vitro treatment with ICI. This is closely associated with [1] the down-regulation of Bcl-2 and Bcl-xL proteinsand [2] upregulation of Bax and Bad, whose gene products are known to be involved the regulation of apoptosis inmammalian cells. Stable over-expression of Bcl-2 resulted in protection of MCF-7 cells from apoptosis and growthinhibitory effects of ICI. Conversely, reduction of Bcl-2 by antisense transfection make MCF-7 cells more sensitiveto ICI-induced growth inhibition and apoptosis. These findings suggest that modulation of Bax, Bcl-xL, Bcl-2 andBad proteins by ICI may be, in part, responsible for the anti-proliferative and apoptotic effect of ICI seen clinicallyand in animal models.

Introduction

Apoptosis plays an important role during develop-ment, organ involution, and in many diseases, includ-ing cancer [1, 2]. Regulation of apoptosis is a complexprocess which involves a number of cellular genes,including Bcl-2 [3, 4], and Bcl-2 related family mem-bers such as Bcl-xL, Bcl-xs, Bad, and Bax [5]. TheBcl-2 gene product protects cells against apoptosis ina variety of experimental systems. Over-expression ofBcl-2 has been shown to suppress the initiation of ap-optosis in response to a number of stimuli, includinganticancer drugs [6, 7, 8, 9]. Furthermore, inhibitionof Bcl-2 expression by antisense oligonucleotide [10,11] and dexametasone [12] has been shown to promoteapoptosis and increase sensitivity to chemotherapy-induced apoptosis. Thus, cancer cells may primarily

depend on Bcl-2 or related family members to pre-vent cell death. It has been shown that cells from avariety of human cancers including breast may havedecreased ability to undergo apoptosis in responseto some physiological stimuli [2, 13] and hence adefect in apoptosis may lead to the development ofcancer. Therefore, identification of agents that negat-ively regulate the Bcl-2 pathway in breast cancer, thustriggering apoptosis, serves to provide a therapeuticapproach leading to growth arrest of breast cancercells.

Steroid hormones are potent regulators of growth,differentiation and apoptosis in steroid-dependent celltypes and tissues such as mammary gland. In women,a high proportion of primary breast cancers has de-tectable levels of estrogen receptor (ER) and requiresestrogen for tumour proliferation. Current treatments

128 KB Lim et al.

have been directed toward interruption of estrogen byoophorectomy or the use of antiestrogens [14, 15]. ICI182780 (FaslodexTM, AstraZeneca, Cheshire UnitedKingdom) is a novel, steroidal estrogen antagonistthat is devoid of estrogen activity in preclinical mod-els [16]. Multiple changes in ER function after ICI182780 treatment appear to contribute to the blockageof estrogen action. These include impaired dimeriza-tion, increased turnover, and disrupted nuclear local-ization [17, 18, 19]. The transcription of ER-regulatedgenes is inhibited as a consequence of the downregula-tion of ER by ICI. Tamoxifen-resistance cell lines andtumours selected in vivo remain sensitive to growthinhibition by ICI [20, 21, 22]. Breast cancer patientstreated with ICI have a significant decline in ER andprogesterone receptor [23]. Treatment of ICI also res-ulted in declines in the levels of estrogen-stimulatedproteins, pS2, and Ki-67, a cell cycle marker of pro-liferation [23]. In vitro, ICI has been shown to actas a growth inhibitor even in the complete absenceof estrogen stimuli [24, 25], and to actively regulategene expression in a direction opposite to that of es-trogens [25]. We and others have shown that ICI pos-sesses a stronger antineoplastic activity than tamox-ifen and also absence of tamoxifen-related uterine sideeffects [26].

Although ICI has been shown to be a potent growthinhibitor and inducer of apoptosis in ER-positivebreast cancer, little is known about the biochemicalmechanisms of in vivo apoptosis induced by ICI. Ourpresent study demonstrates that in vivo treatment ofrats with ICI resulted in the apoptosis of breast epi-thelial cells. In vitro, treatment of MCF-7 cells withICI also resulted in apoptosis and growth inhibition.Both in vivo and in vitro effects of ICI on apoptosiswere closely linked with the down-regulation of Bcl-2 and Bcl-xL expressions and increase in Bax andBad. Over-expression of Bcl-2 in MCF-7 cells resultedin significant inhibition of ICI-induced apoptosis andalso protected the MCF-7 cells against the inhibitoryeffects of ICI. Thus, the imbalance between Bax, Bad,Bcl-xL and Bcl-2 expressions during ICI treatmentmay play an important role in ICI-induced growth in-hibition and apoptosis as observed in our in vitro andin vivo studies.

Materials and methods

Animal experiments were approved by Local AnimalCare Committee. Ovary intact Sprague-Dawley rats,60 days old at the beginning of the experiments were

obtained from Charles River, Quebec. To investig-ate the effects of ICI 182780 on breast epithelial cellapoptosis, rats (n= 12) were weekly injected with1 mg (n= 12), 1.5 mg (n= 12), and 2 mg (n= 12) ICI182780 per kg body weight (AstraZeneca, Pharma-ceuticals) dissolved in castor oil for 3 weeks. Controlrats (n= 12) received only castor oil. Animals weresacrificed by carbon dioxide at the end of the experi-ment. The mammary tissue was excised, trimmed andfrozen in liquid nitrogen and stored at −70◦C for RNAextraction. Part of the mammary tissue was fixed in10% buffered formalin for histochemical studies.

Immunohistochemistry and histology

Fixed mammary tissue was routinely processed in atissue processor and embedded in paraffin. Sections of5 µm were cut and stained with haematoxylin-eosin.Examination of the slides was performed by light mi-croscopy. The ImmunoCruz Staining System was usedfor immunohistochemical study. To evaluate the Ki-67labelling index, 500 epithelial cells were counted foreach group in randomly chosen fields at a × 400 mag-nification. The Ki-67 labelling index was expressedas the number of clearly labelled Ki-67 reactive nuc-lei in 500 cells counted. Significance difference wasdetermined by Mann–Whitney U -test.

Quantitation of apoptosis

For mammary tissues 5 µM sections were used. Frag-mented DNA were labelled using the ApoAlert DNAfragmentation assay (Clontech Laboratories, PaloAlto, CA) which is based on the terminal deoxynuc-loetidyl transferase-mediated dUTP nick end labelling(TUNEL) as described [27]. Labelling indices wereobtained by counting the number of labelled cellsamong at least 100 epithelial cells per region andexpressed as percentage values.

Western blotting

To determine the changes in the expression of Bcl-2,Bad, Bax and Bcl-xL in the mammary gland, mam-mary tissue was homogenized in lysis buffer (1 mMCaCl2, 1 mM MgCl2, 1% NP-40, 1 µg/ml leupeptin,1 µg/ml aprotinin, 1 µM PMSF, and 100 µM NaVO4).Proteins were subjected to western blot analysis asdescribed [28]. Blots were incubated with eitherrabbit anti-Bcl-2 (1 µg/ml), anti-Bad (1 µg/ml) andanti-Bcl-xL (0.75µg/ml), mouse anti-Bax (1 µg/ml)and mouse anti-α tubulin (0.5 µg/ml) antibodies and

Induction of apoptosis by a pure anti-estrogen ICI 182780 129

horseradish peroxidase-conjugated donkey anti-mouseor anti-rabbit secondary antibody (1:7500). All anti-bodies were obtained from Santa Cruz. Blots werevisualized with a chemiluminescent detection systemas described by the manufacturer (ECL, Amersham).

Cell culture, quantitation of apoptosis and thymidineincorporation assay

Human breast cancer MCF-7 cells were maintained asmonolayer cultures in Alpha Modified Eagle’s Media(α-MEM) (Gibco, Grand Island, NY) supplementedwith 10% fetal bovine serum (FCS) (Gibco, GrandIsland, NY). Confluent stock cultures were trypsin-ized and plated at 2.5 × 104 cells per well in 24-welldishes (Becton Dickinson, Lincoln Park, New Jersey)in α-MEM supplemented with 2.5% FCS. After 48 hcell monolayers were rinsed twice with serum-free α-MEM and incubated for a further 24 h in α-MEM sup-plemented with 2.5% double charcoal stripped serum.Cells were then washed twice with serum-, estrogen-,and phenol red-free (SEPF) α-MEM, and then incub-ated for 48 h in the presence or absence of variousconcentrations of ICI 182780 under study, as indicatedin figure legends. Cells were harvested for determina-tion of cell number. For apoptosis determination, cellswere plated and treated as described above. Treatedcells were harvested by scraping in cold phosphatebuffered saline and cytoplasmic cells extracts wereprepared according to the manufacturer’s protocol andwere equalized on the basis of cell number. Apop-tosis was measured using a cell death ELISA (RocheMolecular Biochemicals) which measures cytoplas-mic histone-bound DNA generated during apoptoticDNA fragmentation and not free histone or DNA thatcould be released during non-apoptotic cell death.Samples from triplicate well were run in duplicate onthe ELISA.

To determine the changes in Bcl-2, Bad, Bax andBcl-xL proteins, MCF-7 cells were plated at a dens-ity of 5 × 106 cells per 75 mm2 flask and treated withindicated concentrations of ICI 182780 for 48 h as de-scribed above. Treated cells were harvested and lysedin above lysis buffer and cell lysates were analyzed bywestern analysis as described above.

MCF-7 cell stable transfectant cell lines

The entire coding region of Bcl-2 cDNA [29] wascloned into the mammalian expression vector pcDNA-3.1 (Invitrogen, Carlsbad, CA) in a sense (pcDNA3.1/

Bcl-2S) and antisense (pcDNA3.1/Bcl-2AS) orient-ation. The recombinant plasmid pcDNA3.1/Bcl-2Sand pcDNA3.1/Bcl-2AS sequences were confirmed bysequencing. MCF-7 cells were seeded at 2 × 105 in100 mm culture dishes in 90% α-MEM (Life Techno-logies, Inc.) containing 10% FCS with Garamycine24 h prior to transfection. Cells were transfected with5 µg of either pcDNA3.1/Bcl-2S or pcDNA3.1/Bcl-2AS or pDNA3.1 control plasmid DNA and 28 µl ofLipofectamine reagent (Life Technologies) followingmanufacturer’s recommendations. Forty-eight hourspost-transfection, cells were subcultured at a ratio of1:10 and replaced with growth medium containing800 µg/ml G418 (Calbiochem, La Jolla, CA). After4 weeks, clones were isolated, expanded and assayedfor Bcl-2 expression by western blot analysis.

Statistical analysis

Differences in Ki-67 labelling index, apoptotic la-belling index, thymidine incorporation into the cells,expression of gene in the Bcl-2 family and Bcl-2 pro-tein among parental MCF-7 cells, sense Bcl-2 andantisense Bcl-2 transfectants were analysed by theMann–Whitney U -test.

Results

Control mammary gland had a sparse cluster of epi-thelial tubules surrounded by a small amount of con-nective tissue which was in turn embedded in a largefat pad. The epithelial ducts possesses small lumens,lined by cuboidal cells with dark stained nuclei (Fig-ure 1A). Treatment of ovary intact animals with ICI182780 resulted in a marked atrophy of the mam-mary gland (Figure 1B, 1C and 1D). High magnific-ation view revealed that the ICI-induced pattern wascharacterized by a decreased size of the lobular struc-tures, which consisted of small atrophic alveoli, linedby atrophic and low cuboidal cells. The acinar epi-thelial cells were apparently inactive, with diminishedquantity of cytoplasm (data not shown).

We determined the effect of ICI 182780 on Ki-67expression in mammary tissue. Figure 1E shows theresults of an experiment where mammary tissues werecollected from rats treated with vehicle and variousdoses of ICI for immunohistochemical analysis of Ki-67 expression. ICI significantly decreased (p < 0.01)the number of epithelial cells expressing Ki-67 ascompared to controls.

130 KB Lim et al.

Figure 1. Effects of ICI 182780 treatment on mammary gland histology and proliferation of mammary epithelial cells. Rats were treated witheither vehicle (A), 1 mg ICI (B), 1.5 mg ICI (C) and 2 mg ICI (D) for 3 weeks as described under Materials and methods. Mammary sectionswere stained with haematoxylin-eosine. Note ICI caused marked atrophy of the mammary gland, small atrophic alveoli, lined by atrophic andlow cuboidal cells. Magnification × 200. Effect of ICI on mammary epithelial cell proliferation is shown in (E). Ki-67 labelling indices weredetermined as described under Materials and methods. ICI was very effective in reducing proliferation of epithelial cells. Bars with differentletters are significantly different from one another at (p < 0.01) as determined by the Mann–Whitney U -test. Data are expressed as the mean ofeight samples ± SEM. Results shown are representative of three independent experiments.

The TUNEL assay was performed on mammarysections to determine whether blocking of ER activ-ity with ICI caused breast epithelial cells to undergoapoptosis. In normal mammary tissue, the rate of ap-optosis was extremely low (Figure 2A). In rats treatedwith 1, 1.5 and 2 mg ICI for 3 weeks, approximately

10, 35 and 45% of apoptotic epithelial cells wereobserved respectively (Figure 2C).

Cellular proteins of critical importance in regula-tion of apoptosis include those of Bcl-2 [3, 4], Bcl-xL,Bad, and Bax [5]. Western blot analysis was con-ducted to determine if ICI-induced apoptosis in the

Induction of apoptosis by a pure anti-estrogen ICI 182780 131

Figure 2. Effects of ICI treatment on apoptosis of mammary epithelial cells. Sixty days old rats were treated with either vehicle (A) or 2 mgICI for 3 weeks (B) as described under Materials and methods. Formalin-fixed, paraffin-embedded mammary tissue sections were used todetect mammary apoptosis in situ by the TUNEL method. The sections were analysed using fluorescence microscopy with an FITC filter at themagnification of 320X. Representative sections are shown. The nuclei of apoptotic cells contain fluorescein-labeled DNA and appear bright.Quantitation of apoptotic is shown in (C). Quantitation of apoptosis was performed as described under Materials and methods. ICI-inducedapoptosis in rat mammary gland in dose dependent manner. Bars with different letters are significantly different from one another at (p < 0.01)as determined by the Mann–Whitney U -test. Data are expressed as the mean of eight samples ± SEM. Results shown are representative of threeindependent experiments.

132 KB Lim et al.

mammary gland was associated with the changes ofthese proteins. Administration of ICI resulted in adose-dependent increase in expressions of Bax andBad while Bcl-xL and Bcl-2 expressions were inhib-ited (Figure 3). Bax expression levels increased 2.0-,2.2 and 2.5-fold in mammary gland of rats receiv-ing 1, 1.5, and 2 mg ICI respectively compared withcontrol animals. Similarly, Bad levels increased 1.7-,2.6 and 2.7-fold respectively in the ICI-treated groups(Figure 3C). Treatment with 1.0 mg ICI caused a 45%decrease in Bcl-2 and 2 mg ICI resulted in a 70%decrease compared with controls (Figure 3D). A sig-nificant decreased in Bcl-xL protein was observed ata dose of 2 mg ICI (p < 0.05). All changes in Bad,Bax and Bcl-2 expression in treatment groups com-pared with control groups were statistically significant(P < 0.01) as determined by Mann–Whitney U -test.

We have previously shown that in vitro ICI treat-ment inhibited human MCF-7 breast cancer cell pro-liferation and subsequent induction of apoptosis [30].Extending from our previous observations, we ex-amined the changes in the expression of genes in-volved in apoptosis. A significant number of MCF-7cells revealed apoptosis associated with morpholo-gic changes with characteristic rounding up after ICItreatment. A quantitative ELISA assay was used tomeasure cytoplasmic histone bound DNA complexesgenerated during apoptotic DNA fragmentation [31,32]. As shown in Figure 4A, a dose-dependent in-crease in apoptosis was observed following exposureof MCF-7 cells to ICI for 48 h. A 3-fold increasein apoptosis was observed in cells treated with 5 nMICI compared to vehicle-treated controls. This correl-ated well with the inhibitory effects of ICI on MCF-7proliferation (Figure 4B).

The apoptotic events in the mammary gland arewell known to be associated with changes in Bax,Bcl-xL, Bad and Bcl-2 expressions. This have led usto postulate whether ICI-induced apoptosis in MCF-7 cells was also related to the possible modulationin the levels of these proteins. This possibility wastested with the expression of these proteins in MCF-7 cells treated with or without ICI and examined bywestern blot analysis. As shown in Figure 5, treatmentof MCF-7 cells with ICI resulted in down-regulationof Bcl-2 and Bcl-xL in a dose dependent manner butthis contrasted with the significantly elevated levels ofBax and Bad proteins. The results suggest that ICI ex-hibits inhibitory effects on anti-apoptotic proteins andinduces the accumulation of well known apoptotic andproapoptotic proteins.

Since ICI potently inhibited Bcl-2 expression andover-expression of Bcl-2 suppressed apoptosis [6, 7,8, 9], we examined the possible involvement of Bcl-2 in the induction of apoptosis and growth inhibition

Figure 3. Effects of ICI on Bax, Bad, Bcl-xL and Bcl-2 levels inthe rat mammary gland. Rats were treated with either vehicle andindicated doses of ICI for 3 weeks as described under Materialsand methods. Tissue lysates from mammary tissue were analyzedby western blotting. Blots were incubated with anti-α-tubulin (A),anti-Bax (B), anti-Bad (C), anti-Bcl-2 (D) and anti-Bcl-xL (E) an-tibodies. ICI induced Bax and Bad accumulation while Bcl-2 andBcl-xL levels were decreased by ICI. Quantitative analysis of theabove proteins was determined by scanning of the blots densito-metrically (F). Bars with different letters are significantly differentfrom one another at (p < 0.01) as determined by the Mann–WhitneyU -test. Data are expressed as the mean of eight samples ± SEM.Results shown are representative of three independent experiments.

Induction of apoptosis by a pure anti-estrogen ICI 182780 133

of MCF-7 cells by ICI. MCF-7 cells were transfec-ted with sense and antisense Bcl-2 expressing vectorswith subsequent isolation of clones stably expressingsense and antisense Bcl-2. Figure 6 shows the levelsof Bcl-2 protein in 2 sense-transfected clones, whichover-express 4-fold and 2.5-fold higher levels of Bcl-2, respectively, compared to the levels of parentalMCF-7 cells. The two antisense-transfected clones,however, expressed only 20% and 35% Bcl-2 respect-ively, compared to the Bcl-2 levels of parental MCF-7cells and mock-tranfection cells. Figure 7A showssense Bcl-2 clones were protected from (8% and 10%versus 38% in parental cells) while antisense Bcl-2clones were more sensitive to (60% and 62% versus38% in parental cells) the growth inhibitory actionof 0.5 nM ICI to an extent compared to growth in-hibition in MCF-7 cells. Bcl-2 antisense transfection

Figure 4. Induction of apoptosis in MCF-7 cells in response toantiestrogen ICI 182780. Cells were grown and treated with in-dicated doses of ICI as described under Materials and methods.Apoptosis induced by various doses ICI (A) was quantitated by celldeath ELISA where absorbance at 405 nm is a direct measure ofDNA fragmentation. Plates treated in parallel were used to meas-ure [3H]-thymidine incorporation (B). Bars with different letters aresignificantly different from one another at (p < 0.01) as determinedby the Mann–Whitney U -test. All experiments were repeated atleast three times. The results from one representative experimentare shown. Data are expressed as the mean of eight samples ± SEM.

also slightly affected the basal proliferation of MCF-7 cells. MCF-7 cells transfected with control vectorexhibited similar sensitivity to growth inhibition andinduction of apoptosis by treatment with ICI.

The effects of Bcl-2 over-expression and under-expression in MCF-7 cells on the ability of ICI in-duced apoptosis were also examined. Results in figure7B showed that Bcl-2 over-expression in MCF-7 cellsled to a significant decrease in the ICI-induced stim-ulation of apoptosis (2.2-fold increase in apoptosiscompared to 3.2-fold apoptosis in the parental MCF-7cells by 48 h) while reduction in Bcl-2 expression byantisense transfection resulted in a significant increasein the ICI-induced apoptosis (4.1-fold increase in ap-optosis compared to 3.2-fold apoptosis in the parentalMCF-7 cells by 48 h ICI treatment).

Discussion

Many therapeutic agents induce tumour regressionthrough their ability to activate the apoptotic path-way. Apoptosis is regulated by specific cellular path-ways, including Bcl-2, Bax, Bad, Bcl-xL and Bcl-xs.The over-expression of Bcl-xs causes breast cancercells to be more sensitive to apoptosis [33] and over-expression of Bcl-2 expression prevents apoptosis inresponse to a variety of stimuli [10, 34, 12]. In thepresent study, we report that treatment with ICI in-duced atrophy of the mammary gland, which wascharacterized by a decrease in the size, and numberof the lobular structure and diminution of secretoryactivity. Histological examinations indicate a reduc-tion in epithelial component of the mammary gland,and TUNEL analysis confirmed the apoptotic death ofmammary epithelial cells. The apoptotic effects of ICIare associated with the decrease in expression of Bcl-2 and Bcl-xL, increase in Bax and Bad proteins andthese correlate well with the anti-proliferative efficacyof ICI. This is consistent with previous studies show-ing that apoptosis is induced in estrogen dependenthuman breast tumours during regression in response totamoxifen treatment [35, 36], an agent routinely usedin breast cancer treatment. The up-regulation of Baxand Bad during ICI-induced apoptosis in mammarygland is similar with previous report showing Bcl-xL, Bax and Bad being upregulated during mammaryinvolution [37].

Despite the high rate of apoptosis as observedin the mammary gland by ICI treatment, not allthe epithelial cells underwent apoptosis. ER staining

134 KB Lim et al.

Figure 5. Effects of ICI on Bax, Bad, Bcl-xL and Bcl-2 levels in MCF-7 cells. Cells were grown and treated with indicated doses of ICIas described under Materials and methods. Cell lysates were analyzed by western blotting. Blots were incubated with anti-α-tubulin (A),anti-Bcl-xL (B), anti-Bax (C), anti-Bcl-2 (D) and anti-Bad (E) antibodies. ICI 182780 increased Bax and Bad accumulation while Bcl-2 andBcl-xL levels were decreased by ICI. Quantitative analysis of the above proteins was determined by scanning of the blots densitometrically(F). Bars with different letters are significantly different from one another at (p < 0.01) as determined by the Mann–Whitney U -test. Data areexpressed as the mean of six samples ± SEM. Results shown are representative of three independent experiments.

on ICI-treated mammary glands reveal that someductal epithelial cells weakly expressed ER-α (datanot shown). It is also possible that there may be sub-populations of cells which are sensitive to ICI andundergo apoptosis rapidly whereas a separate popula-tion of cells are less sensitive to ICI. The latter may beless differentiated than the former one and the degreeof differentiation determines whether they are more orless susceptible to anti-estrogen induced apoptosis. Itis also possible that the remaining cells require higherdose of anti-estrogen exposure or longer treatment forapoptosis.

Our in vitro study shows that ICI causes a greaterthan 3-fold increase in apoptosis along with signific-

antly decrease in MCF-7 cell proliferation as meas-ured by cell death enzyme-linked immunosorbent as-say. This is closely linked with the down-regulation ofBcl-2 and Bcl-xL expression and increase in Bax andBad. Over-expression of Bcl-2 in MCF-7 cells resultsin a significant inhibition of ICI-induced apoptosis andalso protects the MCF-7 cells against the inhibitoryeffects of ICI. Down-regulation of Bcl-2 by antisensetransfection increases the sensitivity of MCF-7 cells toICI-induced apoptosis.

The exact molecular mechanisms underlying ICI-induced apoptosis in normal breast epithelium andbreast cancer cells are not fully understood. However,estrogens seems to play a central role in the coordin-

Induction of apoptosis by a pure anti-estrogen ICI 182780 135

Figure 6. Expression of Bcl-2 in MCF-7 cells transfected with sense and antisense Bcl-2 cDNAs. MCF-7 cells were transfected with sense andantisense Bcl-2 vectors as described under Materials and methods. Western blot analysis with Bcl-2 antibody of total cell lysate (50 µg/lane)extracted from parental MCF-7 cells (P), pcDNA3.1 (V), Bcl-2 antisense (Bcl-2 AS) and Bcl-2 sense (Bcl-2 S) transfected cell lines is shownin (A). Protein loading amounts were compared by blotting with anti-α-tubulin antibody. Quantitative analysis of Bcl-2 was determined byscanning of the blots densitometrically (B). Bars with different letters are significantly different from one another at (p < 0.01) as determinedby the Mann–Whitney U -test. Data are expressed as the mean of eight samples ± SEM. Results shown are representative of three independentexperiments.

ated regulation of cell proliferation and apoptosis ofbreast epithelial cells. The inactivation of estrogenreceptor activity by ICI (reviewed in [38]) may leadto an alteration in the expression of pro-apototic andanti-apoptotic proteins involved in apoptosis.

Although an increase in Bcl-2 observed in MCF-7 cells attenuated ICI-induced stimulation of apop-tosis, ICI-induced apoptosis and growth inhibition arenot completely prevented. The observation can beexplained in part that ICI also down-regulated Bcl-xL protein while the levels of apoptotic protein Baxand proapoptotic protein Bad were significantly el-evated (Figure 5). Since over-expression of Bcl-2 orBcl-xL, in relation to Bax promotes survival, butover-expression of Bax accelerates cell death [39], up-regulation of Bax and Bad and down-regulation ofBcl-xL and Bcl-2 following ICI treatment would pro-mote apoptosis. Thus, the imbalance between Bax,Bad, Bcl-xL and Bcl-2 expressions during ICI treat-ment is believed to play a significant role in ICI-

induced apoptosis as observed in other experimentalsystems [40].

We have previously reported that the inhibitory ef-fect of the anti-estrogen ICI 182780 on MCF-7 cells ismediated in part by the induction of IGFBP-3 gene ex-pression [25] and the apoptotic effect of the antiestro-gen on breast cancer cell proliferation is also mediatedby IGFBP-3 [30]. IGFBP-3 may induce apoptosis in-directly by sequestering anti-apoptotic IGFs from theIGF-IR [30]. Recent report has shown that IGFBP-3can also induce apoptosis and modulates expressionof Bcl-2-related proteins in a p53-independent man-ner [41]. It is possible that ICI up-regulates IGFBP-3which, in turn, modulates the ratio of the pro-apoptoticproteins, Bax and Bad, and anti-apoptotic Bcl-2 andBcl-xL proteins both in vivo and in MCF-7 cells. Sincethe active TGF-β signaling pathway exists in MCF-7 cells and in normal mammary cells, upregulationof IGFBP-3 by ICI may also facilitate the synergismbetween IGFBP-3 and TGF-β in their growth inhib-

136 KB Lim et al.

Figure 7. Proliferative behaviour of the clones expressing sense and antisense Bcl-2 in the presence and absence of ICI 182780. ParentalMCF-7 (P), pcDNA3.1 vector (V), Bcl-2 sense (Bcl-2 S) and antisense (Bcl-2 AS) expressing cells were grown in the absence (Ctrl) or thepresence of 0.5 nM ICI (0.5 nM [ICI]) for 48 h as described under Materials and methods. Cell proliferation was determined by measuring[3H]-thymidine incorporation as described under Materials and methods (A). (B) Apoptosis induced by 0.5 nM ICI was quantitated by celldeath ELISA where absorbance at 405 nm is a direct measure of DNA fragmentation. All experiments were repeated at least three times. Theresults from one representative experiment are shown. Data are expressed as the mean of eight samples ± SEM. Bars with different letters aresignificantly different from one another at (p < 0.01) as determined by the Mann–Whitney U -test. Bcl-2 over-expressing cells were partiallyprotected from ICI-induced growth inhibition and ICI-induced apoptosis while cells expressed low Bcl-2 as a result of antisense exhibited moresensitive to growth inhibition and apoptosis induced by ICI.

ition and apoptosis as described by Fanayan et al.[42].

Target disruption of IGF-IR resulted in inductionof apoptosis [43, 44] and increased sensitivity to ap-optotic stimuli, including chemotherapy drugs. Down-regulation of the IGF-IR function causes a massiveapoptosis of tumour cells both in vivo and in vitro(reviewed in [45]). Our previous study demonstratedthat ICI was able to decrease IGF-IR expression [46].It is possible that the reduction in IGF-IR followingICI treatment may prevent the breast epithelial cellsfrom responding to anti-apoptotic factors, IGFs. The

ability of ICI to inactivate PI-3 kinase by reducingIGF-IR and IGF availability through up-regulation ofIGFBP-3 may be important for ICI-induced apoptosis.This hypothesis is based on the observation that thereis a direct link between PI-3 kinase and apoptosis-regulating protein Bcl family of proteins through Aktphosphorylation of Bad [47, 48]. There is evidencethat anti-apoptotic signaling through the IGF-IR is as-sociated with changes in the expression of Bcl-2 andBcl-xL [49, 50]. It is possible that the changes in Bcl-2and Bcl-xL following ICI treatment may in part be dueto the changes in IGF-IR expression.

Induction of apoptosis by a pure anti-estrogen ICI 182780 137

The observation that ICI-induced growth inhibitionand apoptosis are associated with down-regulation ofBcl-2 and Bcl-xL and up-regulation of Bax and Bad isimportant because the ratio of pro-apoptotic Bax-likeproteins to anti-apoptotic Bcl-2 is a crucial determin-ant of both cellular susceptibility to apoptosis [40] andradiosensitivity of breast tumours [51, 52]. The datasuggest that ICI can be used in conjunction with radio-therapy in the treatment of breast cancer. Furthermore,Bcl-2 expression in human breast cancers has been as-sociated with a good prognosis, while decreased Baxexpression has been linked to poor clinical outcome(Reviewed in [37]). The data suggest that studying thechanges of these proteins in breast cancer specimensduring the course of anti-estrogen treatment may helpin the prediction of the clinical outcome.

In summary, treatment of rats with a pure anti-estrogen ICI 182780 caused programmed cell deathof individual alveolar cells which was correlated withincreased expression levels of the death inducers, Baxand Bad and decreased expression of survival factors,Bcl-xL and Bcl-2. The ICI induced apoptosis in MCF-7 cells was closely linked with the down-regulation ofBcl-2 and Bcl-xL proteins and upregulation of Bax andBad. Stable overexpression of Bcl-2 resulted in protec-tion of MCF-7 cells from ICI-induced apoptosis andICI-induced growth inhibition. These findings sug-gest that modulation of Bax, Bcl-xL, Bcl-2 and Badproteins by ICI may be, in part, responsible for theapoptotic and anti-proliferative effects of ICI seen inanimal model, clinically and in vitro.

Acknowledgements

We would like to thank Drs A. Wakeling and B. Vose(Astra-Zeneca Pharmaceuticals) for preformulated ICI182,780. This work was supported by grants fromthe BMRC-NSTB and the National Medical ResearchCouncil of Singapore to Hung Huynh.

References

1. Steller H: Mechanisms and genes of cellular suicide. Science267: 1445–1449, 1995

2. Thompson CB: Apoptosis in the pathogenesis and treatmentof disease. Science 2675: 1456–1462, 1995

3. Korsmeyer SJ: Bcl-2 initiates a new category of oncogenes:regulators of cell death. Blood 804: 879–886, 1992

4. Hockenbery DM, Zutter M, Hickey W, Nahm M, KorsmeyerSJ: Bcl-2 protein is topographically restricted in tissues char-acterized by apoptotic cell death. Proc Natl Acad Sci USA8816: 6961–6965, 1991

5. Boise LH, Gonzalez-Garcia M, Postema CE, Ding L, LindstenT, Turka LA, Mao X, Nunez G, Thompson CB: Bcl-x, a Bcl-2-related gene that functions as a dominant regulator of apoptoticcell death. Cell 744: 597–608, 1993

6. Hockenbery D, Nunez G, Milliman C, Schreiber RD,Korsmeyer SJ: Bcl-2 is an inner mitochondrial membrane pro-tein that blocks programmed cell death. Nature 348: 334–336,1990

7. Cox AL, Skipper J, Chen Y, Henderson RA, Darrow TL,Shabanowitz J, Engelhard VH, Hunt DF, Slingluff CL Jr:Identification of a peptide recognized by five melanoma-specific human cyotoxic T cell lines. Science 264: 716–719,1994

8. Hickman JA: Apoptosis induced by anticancer drugs. CancerMetastasis Rev 112: 121–139, 1992

9. Fisher TC, Milner AE, Gregory CD, Jackman AL, AherneGW, Hartley JA, Dive C, Hickman JA: Bcl-2 modulation of ap-optosis induced by anticancer drugs: resistance to thymidylatestress is independent of classical resistance pathways. CancerRes 53: 3321–3326, 1993

10. Berchem GJ, Bosseler M, Sugars LY, Voeller HJ, Zeitlin S,Gelmann EP: Androgens induce resistance to Bcl-2-mediatedapoptosis in LNCaP prostate cancer cells. Cancer Res 554:735–738, 1995

11. Carson WE, Haldar S, Baiocchi RA, Croce CM, Caligiuri MA:The c-kit ligand suppresses apoptosis of human natural killercells through the upregulation of Bcl-2. Proc Natl Acad SciUSA 91: 7553–7557, 1994

12. Lotem J, Sachs L: Regulation of Bcl-2, Bcl-xL and Bax inthe control of apoptosis by hematopoietic cytokines and dexa-methasone. Cell Growth Differ. 66: 647–653, 1995

13. Carson DA, Ribeiro JM: Apoptosis and disease. The Lancet341: 1251–1254, 1993

14. Early breast cancer trialists’ collaborative group: Systemictreatment of early breast cancer by hormonal, cytotoxic, orimmune therapy. The Lancet 339: 1–15, 71–85, 1992

15. Howell A, DeFriend D, Robertson J, Blamey R, Walton P:Response to a specific antiestrogen (ICI 182780) in tamoxifenresistant breast cancer. The Lancet 345: 29–30, 1995

16. Wakeling AE, Dukes M, Bowler J: A potent specific pure anti-estrogen with clinical potential. Cancer Research 51: 3867–3873, 1991

17. Parker MG: Action of ‘pure’ antiestrogens in inhibiting estro-gen receptor action. Breast Cancer Res Treat 262: 131–137,1993

18. Pink JJ, Jordan VC: Models of estrogen receptor regulation byestrogens and antiestrogens in breast cancer cell lines. CancerRes 56: 2321–2330, 1996

19. Dauvois S, White R, Parker MG: The antiestrogen ICI 182780disrupts estrogen receptor nucleocytoplasmic shuttling. J CellSci 106 (Pt 4): 1377–1388, 1993

20. Hu XF, Veroni M, De Luise M, Wakeling A, Sutherland R,Watts CK, Zalcberg JR: Circumvention of tamoxifen resist-ance by the pure anti-estrogen ICI 182,780. Intern J Cancer555: 873–876, 1993

21. Osborne CK, Jarman M, McCague R, Coronado EB, Hilsen-beck SG, Wakeling AE: The importance of tamoxifen meta-bolism in tamoxifen-stimulated breast tumor growth. CancerChemother Pharmacol 342: 89–95, 1994

22. Osborne CK, Coronado–Heinsohn EB, Hilsenbeck SG, Mc-Cue BL, Wakeling AE, McClelland RA, Manning DL, Nich-olson RI: Comparison of the effects of a pure steroidal anti-estrogen with those of tamoxifen in a model of human breastcancer. J Natl Cancer Inst 87: 746–750, 1995

138 KB Lim et al.

23. DeFriend DJ, Howell A, Nicholson RI, Anderson E, Dow-sett M, Mansel RE, Blamey RW, Bundred NJ, Robertson JF,Saunders C: Investigation of a new pure antiestrogen (ICI182780) in women with primary breast cancer. Cancer Res542: 408–414, 1994

24. Vignon F, Bouton MM, Rochefort H: Antiestrogens inhibit themitogenic effect of growth factors on breast cancer cells in thetotal absence of estrogens. Biochem Biophys Res Commun146: 1502–1508, 1987

25. Huynh H, Yang X, Pollak M: Estradiol and antiestrogens reg-ulate a growth inhibitory insulin- like growth factor bindingprotein 3 autocrine loop in human breast cancer cells. J BiolChem 271, 1016–1021, 1996

26. Huynh H, Pollak M: Uterotrophic actions of estradiol andtamoxifen are associated with inhibition of uterine IGF bind-ing protein 3 gene expression. Cancer Res 54: 3115–3119,1994

27. Nickerson T, Pollak M, Huynh H: Castration-induced apop-tosis in the rat ventral prostate is associated with increased ex-pression of genes encoding insulin-like growth factor bindingproteins 2,3,4 and 5. Endocrinology 1392: 807–810, 1998

28. Huynh H, Larsson C, Narod S, Pollak M: Tumour suppressoractivity of the gene encoding mammary-derived growth inhib-itor. Cancer Res 55: 2225–2231, 1995

29. Seto M, Jaeger U, Hockett RD, Graninger W, Bennett S,Goldman P, Korsmeyer SJ: Alternative promoters and exons,somatic mutation and deregulation of the Bcl-2-Ig fusion genein lymphoma. EMBO J 71: 123–131, 1988

30. Nickerson T, Huynh H, Pollak M: Insulin-like growth factorbinding protein-3 induces apoptosis in MCF7 breast cancercells. Biochem Biophys Res Commun 237: 690–693, 1997

31. Leist M, Gartner F, Bohlinger I, Tiegs G, Wendel A: Applic-ation of the cell death ELISA for the detection of tumournecrosis factor induced DNA fragmentation in murine modelsof inflammatory organ failure. Biochemica 11: 20–22, 1994

32. Wang TT, Phang JM: Effects of estrogen on apoptotic path-ways in human breast cancer cell line MCF-7. Cancer Res 55:2487–2489, 1995

33. Sumantran VN, Ealovega MW, Nunez G, Clarke MF, WichaMS: Overexpression of Bcl-xs sensitizes MCF-7 cells tochemotherapy-induced apoptosis. Cancer Res 55: 2507–2510,1995

34. Mandal M, Kumar R: Bcl-2 expression regulates sodiumbutyrate-induced apoptosis in human MCF-7 breast cancercells. Cell Growth Differ 73: 311–318, 1996

35. Bardon S, Vignon F, Montcourrier P, Rochefort H: Steroidreceptor-mediated cytotoxicity of an antiestrogen and an anti-progestin in breast cancer cells. Cancer Res 475: 1441–1448,1987

36. Warri AM, Huovinen RL, Laine AM, Martikainen PM,Harkonen PL: Apoptosis in toremifene-induced growth inhib-ition of human breast cancer cells in vivo and in vitro. J NatlCancer Inst 85: 1412–1418, 1993

37. Schorr K, Li M, Krajewski S, Reed JC, Furth PA: Bcl-2 genefamily and related proteins in mammary gland involution andbreast cancer. J Mammary Gland Biol Neoplasia 42: 153–164,1999

38. Howell A, Osborne CK, Morris C, Wakeling AE: ICI 182,780 (Faslodex): development of a novel, ‘pure’ antiestrogen.Cancer 894: 817–825, 2000

39. Sedlak TW, Oltvai ZN, Yang E, Wang K, Boise LH, ThompsonCB, Korsmeyer SJ: Multiple Bcl-2 family members demon-strate selective dimerizations with Bax. Proc Natl Acad SciUSA 92: 7834–7838, 1995

40. Oltvai ZN, Milliman CL, Korsmeyer SJ: Bcl-2 heterodimer-izes in vivo with a conserved homolog, Bax, that acceleratesprogrammed cell death. Cell 744: 609–619, 1993

41. Butt AJ, Firth SM, King MA, Baxter RC: Insulin-like growthfactor-binding protein-3 modulates expression of Bax andBcl-2 and potentiates p53-independent radiation-induced ap-optosis in human breast cancer cells. J Biol Chem 275:39174–39181, 2000

42. Fanayan S, Firth SM, Butt AJ, Baxter RC: Growth inhibitionby insulin-like growth factor-binding protein-3 in T47D breastcancer cells requires transforming growth factor-beta (TGF-beta) and the type II TGF-beta receptor. J Biol Chem 275:39146–39151, 2000

43. Lamm GM, Christofori G: Impairment of survival factor func-tion potentiates chemotherapy-induced apoptosis in tumourcells. Cancer Res 584: 801–807, 1998

44. Liu Y, Lehar S, Corvi C, Payne G, O’Connor R: Expressionof the insulin-like growth factor I receptor C terminus as amyristylated protein leads to induction of apoptosis in tumorcells. Cancer Res 583: 570–576, 1998

45. Baserga R: The contradictions of the insulin-like growth factor1 receptor. Oncogene 19: 5574–5581, 2000

46. Huynh H, Nickerson T, Yang X, Pollak M: Regulation ofinsulin-like growth factor I receptor by the pure antiestrogenICI 182,780. Clin Cancer Res. 2: 2037–2042, 1996

47. Datta SR, Dudek H, Tao X, Masters S, Fu H, Gotoh Y,Greenberg ME: Akt phosphorylation of BAD couples sur-vival signals to the cell- intrinsic death machinery. Cell 912:231–241, 1997

48. Zha J, Harada H, Yang E, Jockel J, Korsmeyer SJ: Serinephosphorylation of death agonist BAD in response to sur-vival factor results in binding to 14–3-3 not Bcl-xL. Cell 874:619–628, 1996

49. Singleton JR, Dixit VM, Feldman EL: Type I insulin-likegrowth factor receptor activation regulates apoptotic proteins.J Biol Chem 271: 31791–31794, 1996

50. Parrizas M, LeRoith D: Insulin-like growth factor-1 inhib-ition of apoptosis is associated with increased expressionof the Bcl-xL gene product. Endocrinology 138: 1355–1358,1997

51. Sakakura C, Sweeney EA, Shirahama T, Igarashi Y, HakomoriS, Nakatani H, Tsujimoto H, Imanishi T, Ohgaki M, OhyamaT, Yamazaki J, Hagiwara A, Yamaguchi T, Sawai K, Taka-hashi T: Overexpression of Bax sensitizes human breast cancerMCF-7 cells to radiation-induced apoptosis. Intern J Cancer671: 101–105, 1996

52. Harima Y, Harima K, Shikata N, Oka A, Ohnishi T, Tanaka Y:Bax and Bcl-2 expressions predict response to radiotherapyin human cervical cancer. J Cancer Res Clin Oncol 1249:503–510, 1998

Address for offprints and correspondence: Hung Huynh, Labor-atory of Molecular Endocrinology, Division of Cellular and Mo-lecular Research, National Cancer Centre of Singapore, Singa-pore 169610; Tel: +65 436 8347; Fax.: +65 226 5694; E-mail:cmrhth@nccs.com.sg