Tamoxifen Stimulates the Growth of Cyclin D1–Overexpressing Breast Cancer Cells by Promoting the Activation of Signal Transducer and Activator of Transcription 3

Introduction

Estrogen is required for the normal proliferation and differentiation of breast epithelial cells and is also implicated in the development and progression of breast cancer. The estrogen receptor (ER) is primarily a nuclear protein and the classic view is that its activation involves the binding to estrogen and receptor dimerization. The ER dimer then associates with coregulators, binds to DNA, and activates the transcription of target genes, which results in cell cycle progression ( 1). Ligand-independent activity of the ER is also observed and results from phosphorylation of the ER ( 1). The transcription activity of the ER is mediated by two domains, activation function 1 (AF1) and 2 (AF2). The activities of AF1 and AF2 differ depending on the cell types and can either be dominant or synergize with each other ( 1). In addition, nonclassic pathways of ER-mediated transcription as well as nongenomic roles of the ER have also been reported ( 1). Therefore, although much remain to be understood about the regulation of the ER, the original observation of its activation by binding to estrogen has led to the development of antihormonal therapy for the treatment of ER-positive breast cancers.

Tamoxifen was the first antiestrogen to be developed and has been used for the management of ER-positive breast cancers over the past three decades. Binding of both estrogen and tamoxifen leads to conformational changes of the ER that result in the recruitment of coactivators and corepressors; however, whereas estrogen acts as an agonist whether AF1 or AF2 is dominant, tamoxifen acts as a antagonist on AF2 but as an agonist on AF1. As a result, tamoxifen can lead to repression or activation of estrogen-dependent genes depending on the cellular context and whether AF1 or AF2 is dominant in this tissue. Because the presence of the ER is essential for tamoxifen response, the ER status has been used to identify breast cancer patients who are likely to benefit from tamoxifen treatment. However, despite ER expression, some tumors do not respond or develop resistance to tamoxifen ( 2, 3). Multiple mechanisms have been proposed to contribute to the resistance to tamoxifen, including change in uptake or metabolism of tamoxifen, loss of expression of ER, expression of mutant or variant forms of ER, loss of cofactors, modification of the estrogen response element, and ligand-independent ER activation. In addition, because cyclin D1 binds and activates the ER in a ligand-independent manner and prevents its inhibition by tamoxifen ( 4, 5), the overexpression of cyclin D1 may also contribute to tamoxifen resistance.

Cyclins act as the regulatory subunits of the cyclin-dependent kinases (cdk). Binding of D-type cyclins to cdk4 and cdk6 leads to the phosphorylation of the retinoblastoma protein (Rb), release of the E2F family of transcription factors from Rb, which in turn activates downstream targets of E2F required for the G₁-S phase transition and results in cellular proliferation ( 6, 7). In addition to its original role as a cdk-dependent regulator of the cell cycle, cyclin D1 also affects the activity of various transcription factors in a cdk-independent manner including the ER ( 4, 5, 8– 10). Importantly, cyclin D1 not only interacts with the ER but also with steroid receptor coactivators and P/CAF ( 11, 12). Therefore, by binding to both the ER and its coactivators, cyclin D1 acts as a bridging factor that recruits coactivators to the ER in the absence of estrogen. This effect of cyclin D1 on the ER does not require cdk binding by cyclin D1, because it can be recapitulated with a cyclin D1 mutant that is unable to bind cdk4 ( 5). As cyclin D1 is overexpressed in 35% of breast cancers and that cyclin D1 positivity is tightly linked to ER positivity, this observation has led to the hypothesis that cyclin D1 overexpression may help identify the subset of patients that do not benefit from tamoxifen treatment.

In agreement with this possibility, two recent clinical studies have found a strong link between cyclin D1 overexpression and resistance to tamoxifen ( 13, 14). First, in postmenopausal women, high levels of cyclin D1 protein were found to promote resistance to tamoxifen ( 13). Second, in premenopausal women, elevated levels of cyclin D1 protein also abolished the beneficial effect of tamoxifen on recurrence-free survival and further elevated levels of cyclin D1 gene correlated with a 6.38-fold increase in relative risk of breast cancer recurrence and a 5.34-fold increased in relative risk of death following tamoxifen treatment ( 14). In contrast to the adverse effect of tamoxifen treatment of cyclin D1–overexpressing cancers, in the control group who did not receive tamoxifen, elevated levels of cyclin D1 were found to be associated with a better outcome in both studies ( 13, 14).

The beneficial effect of cyclin D1 overexpression on breast cancer survival was also reported by several others groups ( 15– 20). Notably, a study using microarray analysis of cyclin D1 expression reported that high cyclin D1 expression was associated with low risk of local recurrence of breast cancer, whereas low expression was associated with high risk ( 16). The other studies showed that strong staining of cyclin D1 is associated with inverse tumor grade, smaller tumor size, and improved relapse-free and overall survival of breast cancer patients ( 18, 20). We recently reported that the beneficial effect of cyclin D1 correlates with its ability to repress the expression of the antiapoptotic transcription factor signal transducer and activator of transcription 3 (STAT3; ref. 21). These results suggest that cyclin D1 overexpression induces opposite effects on breast cancer cell survival depending on whether tamoxifen is given or not. We hypothesized that one possible explanation for this opposite effect of cyclin D1 may be that tamoxifen alters the interaction between cyclin D1/STAT3 and cyclin D1/ER. We initiated this study to address this possibility and found that upon tamoxifen treatment, cyclin D1 can no longer inhibit STAT3 and that the resulting activation of STAT3 contributes to tamoxifen resistance.

Materials and Methods

Cell culture, reagent, and transfection. Cells were grown in RPMI medium supplemented with 10% fetal bovine serum (FBS), insulin (5 μg/mL), and antibiotics (Life Technologies, Inc.). For depletion of estrogen, cells were cultured in phenol red–free RPMI 1640 supplemented with 10% charcoal-stripped FBS, insulin (5 μg/mL), and antibiotics. Charcoal stripping of serum was carried out as follows: 1.25 g of activated charcoal and 0.125 g of dextran (both from Sigma) were added to 500 mL FBS, incubated at 55°C for 30 min, and centrifuged at 3,000 rpm for 20 min; the resulting single-stripped serum was collected and the procedure was repeated but with incubation at 37°C for 30 min, followed by centrifugation and filter sterilization of the double-stripped serum.

Tamoxifen was purchased from Sigma. AG490 [a-Cyano-(3,4-dihydroxy)-N-benzylcinnamide] was purchased from Calbiochem.

Transient transfections were performed by lipofection using the FuGENE 6 system as described by the manufacturer (Boehringer Mannheim).

Cell culture on Matrigel. Cells (5 × 10³/mL to 1 × 10⁴/mL) were plated in 500 μL DMEM with 5% FBS and insulin (5 μg/mL) in four-well plates that had been precoated with 250 μL Matrigel (BD Biosciences). The medium was replaced every 3 to 4 days.

3-(4,5-Dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide assay. To determine the percentage of cell survival, T47D and T47D-D1 cells were seeded at 3 × 10⁴/mL in 24-well plates and then treated for indicated time with or without drugs. 3-(4,5-Dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide (MTT; Sigma) solution (50 μL of a 5 mg/mL solution in PBS) was added to each well, and the cells were incubated for 4 h at 37°C. The medium was then aspirated, the cells were lysed in 400 μL DMSO per well and the absorption at 570 nm was determined by Universal Micro plate Reader ELX800 (BIOTEX Instrument, Inc.). The percentage of cell survival was evaluated relative to that of untreated cells.

Quantitation of apoptotic cells. Apoptotic cells were examined morphologically after staining with acridine orange and ethidium bromide. Cells were trypsinized from Matrigel, and 4 μL stock solution containing 100 mg/mL acridine orange and 100 mg/mL ethidium bromide were added to 100 μL of cell suspension. The numbers of total cells and apoptotic cells that showed nuclear shrinkage and apoptotic bodies were counted. The percentage of apoptotic cells was calculated after counting a total of 1,500 cells.

Immunoprecipitation and Western blot analysis. Protein extracts from cell lines were prepared by washing thrice in ice-cold PBS, followed by lysis in 200 μL ice-cold NP40 lysis buffer [50 mmol/L Tris (pH 7.5), 250 mmol/L NaCl, 5 mmol/L EDTA, 0.5% NP40, 50 mmol/L NaF, 0.2 mmol/L Na₃VO₄, 1 g/mL leupeptin, 1 g/mL pepstatin, 100 g/mL phenylmethylsulfonyl fluoride, and 1 mmol/L DTT]. Lysates were centrifuged at 10,000 × g for 20 min at 4°C, and the protein concentrations of the supernatants were determined using the Bio-Rad protein assay. Proteins (15 μg) were separated by SDS-PAGE on 10% acrylamide gels and transferred to nitrocellulose membranes (Perkin-Elmer Life Sciences). Membranes were incubated with rabbit polyclonal anti–cyclin D1 antibody (1:500; Santa Cruz Biotechnology), mouse monoclonal anti-STAT3 antibody (1:1,000; Zymed), rabbit polyclonal anti–phospho-STAT3 antibody (1:500, Santa Cruz Biotechnology), rabbit polyclonal anti-ERα (G-20) antibody (1:500, Santa Cruz Biotechnology), or mouse monoclonal anti-tubulin antibody (1:2,000; Hybridoma Facility, University of Iowa, Iowa City, IA), and antigen-antibody complexes were visualized using the ECL kit (Amersham Pharmacia Biotech). All experiments were performed at least twice. The intensity of the bands was quantified using a Bio-Rad GS-800 densitometer equipped with the Quantity One program (Bio-Rad).

To immunoprecipitate HA–cyclin D1, anti–HA-probe antibody (Y-11, Santa Cruz Biotechnology) was added to the lysates and incubated at 4°C for 90 min. Protein A–Sepharose beads were added to the lysates and incubation was continued overnight at 4°C. Protein A–Sepharose beads were added to the lysates and incubation was continued overnight at 4°C.

Xenograft implantation and measurement of tumor size. Eight-week-old BALB/c nude mice were purchased from the Animal Research Center. A pellet of 17-β-estradiol (0.72 mg/mL) and subsequently a tamoxifen pellet (Innovative Research of America; 2.5 mg/pellet, 60 days release) were inserted s.c. in the upper back of each mouse. To insert the pellets, mice were anesthetized using ketamine and xylazine at a dose of 0.1 mg/kg of body weight. A small incision was made and the pellet was inserted using a precision trochar. The incision was sealed using a clip. T47D or T47D-D1 cell pellets (1 × 10⁷ cells) were mixed with an equal volume of Matrigel (Basement Membrane Matrix, BD Biosciences). The mixture was injected s.c. in the lower back of each animal using a 26-gauge needle 1 week after the insertion of the estrogen pellet. The estrogen pellets were kept throughout the experiment because the growth of T47D cells in mice is dependent on estrogen and the removal of the pellet would lead to tumor regression even in the absence of tamoxifen. Our protocol was approved by the animal ethic committee at Mount Sinai School of Medicine.

Tumor size was measured using a digital caliper. Two independent measurements (length and width) were taken for each tumor weekly and their resulting average was used to obtain tumor volume.

Luciferase assay. T47D-D1 cells were plated in a 24-well plate 24 h before transfection at a density of 3 × 10⁴ cells/mL. Cells were transfected with a vector expressing a firefly luciferase reporter plasmid containing two copies of STAT3 consensus binding sites linked to a thymidine kinase minimal promoter, together with control pTK-Renilla luciferase reporter plasmids using the FuGENE 6 system. STAT3 plasmid was kindly provided by Dr. Olivier Coqueret (Centre Hospitalier Universitaire, France). After 24 h of incubation, tamoxifen was added and incubated for an additional 24 h. Luciferase activity was determined using a dual-luciferase reporter assay system (Promega) and luminometer. Firefly luciferase activity was normalized to the activity of Renilla luciferase and to obtain the relative luciferase activity.

Selection of tamoxifen-resistant clones. Tamoxifen-resistant T47D clones were selected after ∼1 year of culture in phenol red–free RPMI 1640 supplemented with 10% charcoal-stripped FBS, insulin (5 μg/mL), antibiotics, and 1 μmol/L of tamoxifen (Sigma).

Results

Cyclin D1–mediated repression of STAT3 is required for its ability to repress growth of ER-negative cells on Matrigel. To investigate the effect of cyclin D1 overexpression on cell growth, we took advantage of our stable clone of the human ER-negative breast cancer cell line HBL100 overexpressing HA-tagged cyclin D1 (HBL100-D1; ref. 21). Both HBL100 and HBL100-D1 were tested for their ability to form colonies on Matrigel. HBL100 cells formed large and elongated colonies, whereas HBL100-D1 cells formed smaller, rounder colonies characterized by the presence of increased number of floating cells ( Fig. 1A ). This effect was due to increased apoptosis as the percentage of apoptosis in HBL100 cells was 5.3%, whereas it increased to 23.7% in HBL100-D1 cells ( Fig. 1A). We previously reported that the beneficial effect of cyclin D1–overexpressing cells was correlated with the repression of the transcription factor STAT3 ( 21). Consistent with this observation, in HBL100-D1 cells, STAT3 and phospho-STAT3 levels were low compared with HBL-100 parental cell line ( Fig. 1B). To determine whether the growth suppression in cyclin D1–overexpressing cells on Matrigel is due to reduced levels of STAT3, HBL100-D1 cells were transiently transfected with a plasmid where STAT3 is under the control of a constitutive promoter and the morphology of the colonies was analyzed ( Fig. 1C). Forced expression of STAT3 in HBL100-D1 cells resulted in the formation of larger colonies compared with HBL100-D1 cells and, further, the cell survival was increased ( Fig. 1C). To quantify this effect, the number of surviving cells in HBL100, HBL100-D1, and HBL100-D1 transfected with STAT3 was determined by MTT assay. When the percentage of surviving HBL100 cells was set at 100%, the survival of HBL100-D1 cells was reduced to 55% and forced expression of STAT3 in HBL100-D1 rescued survival up to 90% ( Fig. 1D). This result supports the hypothesis that overexpression of cyclin D1 inhibits cell growth by repressing STAT3 expression and further supports our previous findings ( 21).

• Download figure
• Open in new tab
• Download powerpoint

Figure 1.

Repression of STAT3 is required for the inhibition of colony formation induced by cyclin D1 overexpression in ER-negative cells. A, HBL100 and HBL100 cells stably transfected with the plasmid expressing cyclin D1^HA (HBL100-D1) were cultured on Matrigel for 2 wk, and representative photomicrographs were taken. The percentage of apoptotic cells was determined using a fluorescence microscope after staining with acridine orange and ethidium bromide. B, HBL100 and HBL100-D1 cells were harvested and proteins were extracted. Levels of cyclin D1, STAT3, and phospho-STAT3 were determined by Western blot analysis using anti-cyclin D1, anti-STAT3, and anti–phospho-STAT3 antibodies, respectively. Anti–α-tubulin antibody was used as a loading control. C, HBL100-D1 cells were transiently transfected with or without a plasmid expressing STAT3. Cells were then plated on Matrigel and representative photomicrographs were taken at 10 d. D, HBL100, HBL100-D1, and HBL100-D1 transfected with STAT3 were plated on Matrigel and allowed to grow for 10 d. Cells were harvested and cell number was determined by MTT assay. Transfection of STAT3 was confirmed by Western blot analysis using anti-STAT3 antibody.

Overexpression of cyclin D1 in ER-positive breast cancer cells also promotes repression of STAT3 and inhibition of cell growth. To further examine the beneficial effect of cyclin D1 overexpression, we created stable clones of the ER-positive breast cancer cell line T47D overexpressing HA-tagged cyclin D1 ( Fig. 2A ). In these clones, the levels of STAT3 and phospho-STAT3 were reduced ( Fig. 2A) and consistent with a decrease in its transcriptional activity, endogenous cyclin D1 were also reduced because the promoter of cyclin D1 contains a STAT3-binding site ( 22, 23). The growth rate of the four individual clones was then compared with that of the parental cell line in vitro. We found that when cells were grown in vitro, no significant difference in the growth rate was observed between individual cyclin D1–overexpressing clones or between the clones and the parental cell lines ( Fig. 2B). Because the highest expression of cyclin D1 was observed in T47D-D1 clone no. 1, we selected this clone for further study. T47D and T47D-D1 cells were injected in mice and their ability to form tumor in vivo was compared. Tumor formation was monitored by weekly size measurements over a period of 9 weeks. Consistent with our previous observation ( 24), the tumor growth rate in T47D-D1 showed a tendency to be slower compared with parental T47D ( Fig. 2C). To further study the mechanism of growth suppression in cyclin D1–overexpressing tumors, STAT3 and phospho-STAT3 levels were determined in three individual tumors at week 9. We found that the levels of STAT3 and phospho-STAT3 were lower in T47D-D1 tumors compared with that in T47D tumors ( Fig. 2D). This result further supports the hypothesis that overexpression of cyclin D1 inhibits cell growth by repressing STAT3 expression and therefore contributes to the beneficial effect of cyclin D1 on breast cancer survival.

• Download figure
• Open in new tab
• Download powerpoint

Figure 2.

Cyclin D1 overexpression represses cell growth in vivo but not in vitro in ER-positive cells. A, T47D cells and four independent clones of T47D cells stably expressing cyclin D1^HA (T47D-D1) were harvested and proteins were extracted. Levels of cyclin D1, STAT3, and phospho-STAT3 were determined by Western blot analysis using anti-cyclin D1, anti-STAT3, and anti-phospho-STAT3 antibodies, respectively. B, T47D and four clones of T47D-D1 cells were cultured for 1, 3, 6, and 9 d, and cell numbers were determined at each time point. Points, mean of three determinations. C, nude mice (n = 15) were injected with 1 × 10⁷ T47D or T47D-D1 cells. Tumor growth was determined over a period of 9 wk as described in Materials and Methods. Points, means of 15 mice. D, proteins were extracted from three individual xenografts from either the T47D or T47D-D1 groups at week 9. Levels of STAT3 and phospho-STAT3 were determined by Western blot analysis using anti-STAT3 antibody and anti–phospho-STAT3 antibody.

Tamoxifen stimulates the growth of cyclin D1–overexpressing tumors in vivo. In premenopausal women, cyclin D1 overexpression was found to not only lead to tamoxifen resistance but to actually increase the risk for disease recurrence ( 14). This adverse effect of tamoxifen on cyclin D1–overexpressing cancers suggests that tamoxifen may in fact promote the growth of these tumors. We first tested whether this clinical observation can be reproduced using our cell line model. T47D and T47D-D1 cells were injected in mice and tumors were allowed to grow for 10 weeks. At week 11, slow release tamoxifen pellets were inserted s.c. and tumor formation was monitored for an additional 6 weeks. As expected, tamoxifen treatment inhibited further growth of T47D tumors; however, tamoxifen failed to inhibit the growth of T47D-D1 xenografts ( Fig. 3A ). At week 17, the distribution of individual tumor volume was plotted and revealed that the average tumor volume of T47D xenograft was 260 mm³, whereas at week 17, the average tumor volume of T47D-D1 tumors was 550 mm³ ( Fig. 3B). As we observed that the growth of cyclin D1–overexpressing cells on Matrigel inversely correlates with STAT3 levels ( Fig. 1), at week 17, tumors were extracted and the levels of STAT3 were determined in individual tumors. We found that STAT3 levels were consistently elevated in T47D-D1 tumors treated with tamoxifen compared with the level of STAT3 in T47D tumors (the average ratio of STAT3/α-tubulin in T47D/T47D-D1 is 1:3.0; Fig. 3C). In the majority of mice, the levels of phospho-STAT3 were also elevated (data not shown). This observation suggests that tamoxifen treatment not only abolishes the inhibitory effect of cyclin D1 on STAT3 but also stimulates STAT3 expression. As these results indicate that tamoxifen may stimulate the growth of cyclin D1 tumors, we next compared the tumor growth rate of T47D-D1 cells with or without tamoxifen treatment. Consistent with the clinical observation ( 14), when the growth rate of T47D-D1 tumors between week 4 and 9 in absence of tamoxifen was set at 100%, in comparison the relative growth rate of tumor during the same period in presence of tamoxifen treatment was increased to 170% ( Fig. 3D). In addition, the levels of STAT3 in the xenografts from tamoxifen-treated T47D-D1 were elevated compared with that from untreated T47D-D1 (average ratio of STAT3/α-tubulin in untreated T47D-D1 versus tamoxifen-treated T47D-D1 is 1:1.9; data not shown). These results suggest that tamoxifen promotes tumor growth when cyclin D1 is overexpressed possibly through STAT3 up-regulation.

• Download figure
• Open in new tab
• Download powerpoint

Figure 3.

Tamoxifen promotes the growth of cyclin D1–overexpressing tumors in vivo. A, nude mice (n = 9) were injected with 1 × 10⁷ T47D and T47D-D1 cells and tumors allowed to grow for 10 wk. The average tumor volume in each group was adjusted to 100 mm³ to allow for a direct comparison of the two groups, and at week 10, tamoxifen treatment was initiated. Tumor volumes were determined over a total period of 17 wk. Statistical difference in tumor volumes between T47D and T47D-D1 during tamoxifen treatment (from week 11 to 17) was determined using two-sided paired t test. *, P < 0.05. B, the tumor volumes of T47D and T47D-D1 on each mouse at week 17 are shown as a scatter plot. C, proteins were extracted from six individual xenografts from either the T47D (n = 6) or T47D-D1 (n = 6) groups at week 17. Levels of STAT3 were determined by Western analysis using anti-STAT3 antibody. D, nude mice (n = 4) were injected with T47D-D1. At week 4, mice were untreated or treated with tamoxifen by inserting tamoxifen pellet. Tumor growth was determined over a period of 9 wk.

Tamoxifen promotes the growth of cyclin D1–overexpressing cells by up-regulating STAT3. To further examine the growth-promoting effect of tamoxifen on cyclin D1–overexpressing cells, T47D and T47D-D1 cells were treated with various concentrations of tamoxifen in vitro for 7 days and the number of viable cells was determined. In agreement with the results obtained in vivo, tamoxifen at concentrations between 0.01 and 1 μmol/L stimulated cell growth of T47D-D1 cells ( Fig. 4A ). In addition, tamoxifen also stimulated the growth T47D-D1-KE cells that overexpressed a mutant form of cyclin D1 that cannot activate cdk4 (data not shown), indicating that the effect is independent of cdk activation. Further, when T47D-D1 cells were allowed to grow for 3 weeks on Matrigel, the colonies formed by T47D-D1 treated with tamoxifen were clearly larger than the colonies of the untreated control (average ratio of the colony volume; control/tamoxifen 1:6.69; Fig. 4B). Because STAT3 was found to be elevated in the T47D-D1 xenograft treated with tamoxifen ( Fig. 3C), we also determined the levels of STAT3 and phospho-STAT3 in T47D-D1 cells treated with tamoxifen and in agreement with the in vivo observation, tamoxifen treatment led to an increase in STAT3 and also phospho-STAT3 levels ( Fig. 4C). To determine whether the growth-promoting effect of tamoxifen is dependent on STAT3 activity, T47D-D1 cells were treated with or without tamoxifen and STAT3 inhibitor, AG490. We found that upon treatment with AG490 at a concentration of 20 μmol/L, as expected AG490 led to a reduction in cell growth ( 25) and further that the growth-promoting effect of tamoxifen was also significantly abolished ( Fig. 4D). These results indicate that tamoxifen stimulates the growth of cyclin D1–overexpressing cells and that this effect correlates with an increase in STAT3 levels.

• Download figure
• Open in new tab
• Download powerpoint

Figure 4.

Tamoxifen up-regulates STAT3 and leads to proliferation of cyclin D1–overexpressing cells. A, T47D and T47D-D1 cells were plated in a 24-well plate (1.5 × 10⁴ per well) and treated with increasing concentrations of tamoxifen for 7 d. The percentage of growth was determined by MTT assay. Results are presented as the mean from one independent experiment performed in triplicate. B, T47D-D1 cells were plated on Matrigel (5 × 10³ per well, four-well plate) and cultured with or without 0.1 μmol/L tamoxifen. The photomicrographs were taken after 3 wk. C, T47D-D1 cells were plated in 10-cm plate (1 × 10⁵ per well) and treated with increasing concentrations of tamoxifen for 7 d. Proteins were harvested and extracted to determine the levels of STAT3 and phospho-STAT3 by Western analysis using anti-STAT3 and anti–phospho-STAT3 antibody. D, T47D-D1 cells were plated on a 24-well plate (1.5 × 10⁴ per well) and treated with or without tamoxifen in the presence of 0 and 20 μmol/L AG490 for 5 d. At day 5, the percentage of cellular survival was determined by MTT assay. Results are presented as the mean of four determinations.

Tamoxifen promotes the redistribution of cyclin D1 from STAT3 to the ER. Cyclin D1 binds to both the ER and STAT3; however, cyclin D1 binding activates the ER whereas it represses STAT3 ( 4, 5, 26). As tamoxifen treatment rescues the levels of STAT3 in cyclin D1–overexpressing cells, one potential explanation is that the conformational change of the ER upon tamoxifen treatment may enhance its binding to cyclin D1 and as a result sequester cyclin D1 away from STAT3 allowing its reactivation. To test this possibility, T47D-D1 cells were treated with tamoxifen and the association between cyclin D1 and the ER was tested by immunoprecipitation following tamoxifen treatment. We found that tamoxifen increased the interaction between cyclin D1 and the ER ( Fig. 5A, top ). The levels of cyclin D1 following immunoprecipitation were not changed during treatment with tamoxifen ( Fig. 5A, middle). The increase in ER levels in the crude lysate following tamoxifen was observed in both T47D and T47D-D1 cells ( Fig. 5A, bottom, and data not shown). However, when the activation of the ER transcriptional activity was measured using an ERE-luciferase reporter, the activity of the ER was only stimulated in the cyclin D1–overexpressing cells and not in the T47D ( Fig. 5B). Furthermore, no association between cyclin D1 and the ER was detected in the T47D cells (data not shown). These results indicate that despite the ability of tamoxifen to increase the level of the ER in both cell lines, the elevation in ER level alone does not lead to an increase in ER activity in cells that have low level of cyclin D1. In cyclin D1–overexpressing cells, however, tamoxifen stimulates the binding of cyclin D1 to the ER and this results in an increase in the ER activity.

• Download figure
• Open in new tab
• Download powerpoint

Figure 5.

Tamoxifen stimulates the sequestration of cyclin D1 by the ER. A, T47D-D1 cells were transfected with HA-tagged cyclin D1 plasmid and treated with increasing concentrations of tamoxifen for 4 d. To determine the interaction between cyclin D1 and ER, protein extracts were immunoprecipitated (IP) using anti-HA antibody, and the levels of ER bound to cyclin D1 were detected using anti-ER antibody (top). The levels of cyclin D1 on the immunoprecipitate are shown as a control (middle). Crude protein extracts were analyzed to determine ER expression by Western blot (WB) using anti-cyclin ER antibody (bottom). B, T47D and T47D-D1 cells were transfected with a vector expressing ERE-luciferase reporter plasmid. After 24 h, cells were treated with or without tamoxifen and luciferase activity was measured after an additional 24 h. Relative activities of untreated controls in T47D and T47D-D1 cells were set at 1. C, the same protein precipitates as in A were used to detect the levels of STAT3 bound to cyclin D1 using anti-STAT3 antibody (top). Crude protein extracts were analyzed to determine the levels of STAT3 and phospho-STAT3 by Western blot using anti-STAT3 and anti–phospho-STAT3 antibody (middle and bottom). D, T47D and T47D-D1 cells were transfected with a vector expressing a luciferase reporter plasmid containing two copies of STAT3 consensus binding sites linked to a thymidine kinase minimal promoter. After 24 h, cells were treated with or without tamoxifen and luciferase activity was measured after an additional 24 h. Relative activities of untreated controls in T47D and T47D-D1 cells were set at 1.

To further test our hypothesis, the association between cyclin D1 and STAT3 was also tested by immunoprecipitation. We found that using the same extract as in Fig. 5A, tamoxifen reduced the interaction between cyclin D1 and STAT3 ( Fig. 5C, top). We next tested more directly the effect of tamoxifen on STAT3 transcriptional activity using a STAT3 luciferase reporter. T47D and T47D-D1 cells were transfected with a plasmid where the luciferase reporter is under the control of the thymidine kinase minimal promoter containing two STAT3 consensus binding sites ( 26). After 24 h, cells were treated with or without tamoxifen for an additional 24 h and STAT3 activity was measured. We found that at a dose of 3 μmol/L, tamoxifen increased STAT3 activity in T47D-D1 cells by 2-fold, whereas no effect was observed in T47D cells ( Fig. 5D). This higher dose was used because unlike other experiments measuring effect on growth rate over several days, the luciferase assay is performed only 24 h following transfection. This result therefore indicates that in the context of cyclin D1 overexpression, STAT3 transcriptional activity is activated following tamoxifen treatment. The activation of STAT3 transcriptional activity following tamoxifen treatment was further tested in two additional ER-positive cell lines that have elevated levels of cyclin D1, namely MCF-7 and ZR-75.1 cells. We found that in both cell lines, STAT3 transcriptional activity was also stimulated by tamoxifen (data not shown); further, this increase in activity correlated with an increase in STAT3 and phospho-STAT3 in these cells lines (data not shown). These results therefore support the hypothesis that following tamoxifen treatment, cyclin D1 is sequestered from STAT3 and redistributed to the ER.

Cyclin D1 is elevated in tamoxifen-resistant T47D clones. To further test the role of cyclin D1 in tamoxifen resistance, we isolated tamoxifen-resistant clones by long-term incubation of T47D cells in the presence of tamoxifen. Although T47D cells rapidly lost their ability to grow upon depletion of estrogen and addition of tamoxifen, after a period of 1 year, resistant colonies (Tam-R-T47D cells) were identified, suggesting that some isolated cells had gained the ability to grow under these conditions although their growth rate is slow. Upon expansion of individual clones in estrogen-depleted/tamoxifen-containing medium, the level of cyclin D1 in each clone was compared with that in T47D cells grown under the same condition for 9 days. We found that in all resistant clones, cyclin D1 levels were elevated compared with the parental T47D cells ( Fig. 6A ). Quantification of cyclin D1 levels revealed that in average, cyclin D1 was increased by 3.6-fold in tamoxifen-resistant clones compared with T47D cells. One possible explanation for this observation is that the difference in cyclin D1 levels may simply reflect differential growth rates between the two cell types because T47D cannot grow in the absence of estrogen, whereas resistant clones do grow. Conversely, in complete medium, T47D cells grow at a much higher rate than the resistant clones (data not shown); therefore, to circumvent this difficulty, we compared the growth of resistant clones, which were incubated either in estrogen-depleted/tamoxifen-containing medium or complete medium over a period of 5 days. As expected, we found that Tam-R-T47D clones grew although slowly in estrogen-depleted/tamoxifen-containing medium, whereas their growth was accelerated in complete medium ( Fig. 6B). Despite the slower growth of Tam-R-T47D clones in estrogen-depleted/tamoxifen-containing medium, the levels of cyclin D1 were much higher compared with the levels of cyclin D1 in the more rapid growth in complete medium ( Fig. 6C). This result indicates that the levels of cyclin D1 did not simply reflect the growth rate but rather that growth in estrogen-depleted/tamoxifen-containing medium requires elevated levels of cyclin D1. This result further supports the observation reported by others that inhibition of cyclin D1 by siRNA abolishes the ability of tamoxifen-resistant clones to grow in the presence of tamoxifen ( 27). In agreement with our finding that STAT3 and phospho-STAT3 levels were elevated in cyclin D1–overexpressing cells in the presence of tamoxifen, this was also observed in tamoxifen-resistant clones ( Fig. 6C).

• Download figure
• Open in new tab
• Download powerpoint

Figure 6.

Tamoxifen-resistant clones display elevated cyclin D1 expression and activated STAT3. A, T47D cells were incubated in the presence of tamoxifen (Tam) over a period of 1 y. Colonies of tamoxifen-resistant cells (Tam-R-T47D cells) were identified and transferred to a 12-well plate for their expansion. Following their expansion, cells from five individual clones were used for Western analysis of cyclin D1 using anti-cyclin D1 antibody. As a control, T47D cells were incubated in absence of estrogen for 9 d. B, Tam-R-T47D cells were either grown in estrogen-depleted/tamoxifen-containing medium or in complete (10% FBS) medium for 10 d and the medium changed at day 5. After 10 d in culture, 1 × 10⁵ Tam-R-T47D cells were plated in estrogen-depleted/tamoxifen-containing medium or in complete (10% FBS) medium and allowed to grow for 5 d. At day 5, the cells were counted. C, cells obtained from growth conditions described in B were harvested after 5 d in culture and used for Western analysis using anti-cyclin D1, anti-STAT3, and anti–phospho-STAT3 antibodies. D, the model of the interaction between cyclin D1, ER, and STAT3 in the presence or the absence of tamoxifen. See Discussion for details.

Discussion

The best known function of cyclin D1 is in the regulation of the cell cycle and therefore the general expectation was that cyclin D1 overexpression would lead to increased proliferation and therefore bad prognosis. However, several studies have reported a beneficial effect of cyclin D1 overexpression in breast cancer. This unexpected observation may relate to the fact that under the abnormal situation where cyclin D1 is overexpressed, excess cyclin D1 becomes available for binding to other partners and therefore the cdk-independent functions of cyclin D1 need to be also considered in its overall effect in breast cancer. An additional complicating factor in defining the role of cyclin D1 in breast cancer is that its overexpression is linked to tamoxifen resistance ( 28– 33). Therefore, the value of cyclin D1 as a prognostic factor is highly dependent on the type of treatment.

The results presented here offer a potential explanation for the apparent beneficial effect of cyclin D1 overexpression in patients that do not receive tamoxifen as well as an explanation for the adverse effect of tamoxifen on cyclin D1–overexpressing breast cancer in premenopausal women. Figure 6D summarizes a model to explain the dual effect of cyclin D1 overexpression in breast cancer. In the absence of tamoxifen, cyclin D1 preferentially binds to STAT3 and only weakly to the ER. This interaction with STAT3 leads to the repression of STAT3 activity and therefore of STAT3 own transcription. Whereas cyclin D1 can activate the ER and therefore cellular proliferation, in the absence of tamoxifen the proapoptotic activity that results from the repression of STAT3 is able to counteract the proliferation and contribute to the slower tumor growth rate observed in cyclin D1–overexpressing cancers. The growth-inhibitory effect of cyclin D1 was, however, only observed in vivo and not in vitro ( Fig. 2B and C). This finding may be explained by the fact that several transcriptional targets of STAT3 are genes involved in angiogenesis and extracellular matrix degradation. Therefore, because the growth in vitro does not require such events, the lack of growth inhibition despite a reduction in STAT3 levels is not surprising.

In the presence of tamoxifen, our data suggest that the conformational changes of the ER promote a stronger interaction of cyclin D1 with the ER. Alternatively, the transcription of genes specifically in response to tamoxifen may encode proteins that facilitate the binding of cyclin D1 to the ER or prevent its binding to STAT3. Our data cannot distinguish between these two possibilities, which will need to be addressed in the future. Nevertheless, the precise mechanism, the end result of tamoxifen treatment, is the preferential binding of cyclin D1 to the ER rather than to STAT3. The release of the repression of STAT3 that results from this sequestration of cyclin D1 to the ER would allow the transcription of STAT3 and its antiapoptotic effect. In this situation, both the activation of the ER by cyclin D1 and the antiapoptotic activity of STAT3 would contribute to the increased proliferation of cyclin D1–overexpressing tumors following tamoxifen treatment. Therefore, our data fully support the clinical observation of an adverse effect of cyclin D1 gene amplification in premenopausal women treated with tamoxifen ( 14). Further, our data support the use of STAT3 inhibitors to overcome acquired tamoxifen resistance.

The difference between the observed resistance associated with elevated cyclin D1 protein versus the adverse effect associated with elevated cyclin D1 gene raises the possibility that some other genes in the cyclin D1 locus may affect the response. However, because the growth-stimulatory effect of tamoxifen is observed in our cell lines where only the cyclin D1 gene is overexpressed, our data indicate that the effect is due to cyclin D1. Therefore, the difference between the predictive value of cyclin D1 protein and cyclin D1 gene suggests that the method of scoring cyclin D1 intensity by immunohistochemistry need to be refined or combined with fluorescence in situ hybridization to identified the patients at most risk of an adverse response.

In postmenopausal women, cyclin D1 overexpression led to resistance to tamoxifen ( 13). Our data further support the use of aromatase inhibitors as the preferred antihormonal treatment for postmenopausal women.

Although our results support the adverse effect of tamoxifen on cyclin D1–overexpressing tumors ( 17, 29– 33), we also reported previously that cyclin D1–overexpressing cells show an increased sensitivity to the proteasome inhibitor bortezomib ( 21). It is therefore tempting to speculate that the clinical use of bortezomib may also be useful for the treatment of breast cancers resistant to tamoxifen. This possibility will be addressed in the future.

In conclusion, because cyclin D1 overexpression is closely linked to ER positivity in breast cancer ( 34, 35), it is highly likely that premenopausal women with cyclin D1–positive tumors would receive tamoxifen therapy. Strong clinical data has indicated the adverse effect of tamoxifen in this group of patients ( 14). Our data now offer a mechanistic rationale to explain these clinical observations and further support the idea that determining cyclin D1 levels in addition to ER may be critical in the better clinical management of breast cancer using tamoxifen.