Effect of estradiol on estrogen receptor-a gene expression ... - Nature

Oncogene (2003) 22, 7998–8011

& 2003 Nature Publishing Group All rights reserved 0950-9232/03 $25.00 www.nature.com/onc

Effect of estradiol on estrogen receptor-a gene expression and activity can be modulated by the ErbB2/PI 3-K/Akt pathway Gerald E Stoica1, Thomas F Franke2, Maria Moroni1, Susette Mueller1, Elisha Morgan3, Mary C Iann3, Abigail D Winder3, Ronald Reiter1, Anton Wellstein1, Mary Beth Martin1 and Adriana Stoica*,1,3 1

Department of Oncology, Lombardi Cancer Center, Georgetown University, Washington DC 20057, USA; 2Department of Pharmacology, Columbia University, New York 10032, USA; 3School of Nursing and Health Studies, Georgetown University, Washington DC 20057, USA

Epidermal growth factor (EGF), insulin-like growth factor-I (IGF-I), and heregulin-b1 (HRG-b1), can modulate the expression and activity of the estrogen receptor-a (ER-a) via the phosphatidylinositol 3-kinase (PI 3-K)/Akt pathway in the ER-a-positive breast cancer cell line, MCF-7. Estradiol can also rapidly activate PI 3-K/Akt in these cells (nongenomic effect). The recent study examines whether Akt is involved in the ER-a regulation by estradiol (genomic effect). Stable transfection of parental MCF-7 cells with a dominant-negative Akt mutant, as well as the PI 3-K inhibitors wortmannin and LY 294,002, blocked the effect of estradiol on ER-a expression and activity by 70–80 and 55–63%, respectively. Stable transfection of MCF-7 cells with a constitutively active Akt mimicked the effect of estradiol. The changes in ER-a expression and activity were abrogated in response to estradiol by an arginine to cysteine mutation in the pleckstrin homology (PH) domain of Akt (R25C), suggesting the involvement of this amino acid in the interaction between Akt and ER-a. Experiments employing selective ErbB inhibitors demonstrate that the effect of estradiol on ER-a expression and activity is mediated by ErbB2 and not by EGFR. Moreover, anchorage-dependent and -independent growth assays, cell cycle and membrane ruffling analyses showed that Akt exerts estrogen-like activity on cell growth and membrane ruffling and that a selective ErbB2 inhibitor, but not anti-ErbB2 antibodies directed to the extracellular domain, can block these effects. In the presence of constitutively active Akt, tamoxifen only partially inhibits cell growth. In contrast, in cells stably transfected with either a dominant-negative Akt or with R25C-Akt, as well as in parental cells in the presence of a selective ErbB2 inhibitor, the effect of estradiol on anchorage-dependent and -independent cell growth was inhibited by 50–75 and 100%, respectively. Dominant-negative Akt inhibited membrane ruffling by 54%; however, R25C-Akt did not have any effect, suggesting that kinase activity plays an important role in this process. Scatchard analysis demon*Correspondence: A Stoica, E411 Research Building, 3970 Reservoir Rd. NW, Washington, DC 20057, USA; E-mail: [email protected] Received 16 December 2002; accepted 6 May 2003

strated a 67% reduction in estrogen-binding capacity in cells transfected with constitutively active Akt. No change in binding affinity of estradiol to the receptor was observed upon transfection with either Akt mutant. Taken together, our results suggest that estradiol treatment results in binding to membrane ER-a and interaction with a heterodimer containing ErbB2, leading to tyrosine phosphorylation. This results in the activation of PI 3-K and Akt. Akt, in turn, may interact with nuclear ER-a, altering its expression and activity. Oncogene (2003) 22, 7998–8011. doi:10.1038/sj.onc.1206769 Keywords: estradiol; estrogen receptor-a; ErbB2; Akt

Introduction The ER plays a central role in the control of cell growth and its expression is used to predict those patients who will benefit from hormonal therapy. The cell-typespecific antiestrogen tamoxifen competitively binds to estrogen receptor-a (ER-a) and inhibits estrogen-stimulated growth of mammary epithelial cells. Tamoxifen inhibits transcriptional activation of activation function AF-2, but not activation function AF-1 (Berry et al., 1990). ICI 182,780, on the other hand, is a complete antagonist that blocks transcriptional activation of both AF-1 and AF-2 (McDonnell et al., 1995). Although most ER-positive breast cancers initially respond to tamoxifen therapy, tamoxifen-resistant tumors eventually develop (Johnston, 1997). This is thought to result from growth factor-induced ER-a activity through activation of protein kinases resulting in phosphorylation of ER-a (Katzenellenbogen et al., 1997). Therefore, a clearer understanding of the mechanism of action of steroid hormones, as well as their cross talk with other signal transduction pathways is important. The effects of estrogens are mediated primarily by direct binding to the ER, which homodimerizes and interacts with estrogen response elements to stimulate the transcription of target genes. The ER is a phosphoprotein and belongs to a nuclear receptor superfamily of ligand-regulatable transcription factors

Akt modulates ER-a GE Stoica et al

7999

(Tsai and O’Malley, 1994). The ER has an aminoterminal domain with a hormone-independent transcriptional activation function AF-1 (Kato, 2001), a central DNA-binding domain, and a carboxy-terminal ligandbinding domain with a hormone-dependent transcriptional AF-2 (Mangelsdorf and Evans, 1995; Weigel, 1996). In the absence of hormone, the ER is associated with a host of proteins that prevent it from interacting with the cellular transcription apparatus. Upon binding estradiol, the receptor undergoes an activating conformational change, facilitating its association with target genes and permitting it to regulate gene transcription (genomic effects) (Kato et al., 1998). In addition to these genomic estrogenic effects, we and others have demonstrated some effects of estrogen that occur within minutes after estradiol administration (nongenomic effects) (Nemere and Farach-Carson, 1998; Watson and Gametchu, 1999; Molloy et al., 2000; Simoncini et al., 2000; Stoica et al., 2003a). Estradiol can rapidly activate MAPK (Kelly and Levin, 2001; Song et al., 2002) or Akt in several cell lines, including breast cancer (Castoria et al., 2001; Marquez and Pietras, 2001; Stoica et al., 2003a). The nongenomic effects of estrogens have been widely viewed as alternatives to nuclear hormone action. However, more recently, using a two-pulse hormonal schedule and the membrane-impermeable E2: BSA in a transfection system in a nerve cell, it was demonstrated that membrane-limited estradiol was necessary but not sufficient for full transcriptional potency of the second estrogen pulse. Moreover, the first pulse depended on intact signal transduction pathways (Vasudevan et al., 2001). In addition to estrogens, several polypeptide growth factors may also play an important role in the growth regulation of breast cancer by autocrine, paracrine, or endocrine mechanism(s) (Sporn and Roberts, 1985; Dickson and Lippman, 1987; Arteaga et al., 1988; Reddy et al., 1994; Dickson and Lippman, 1995). ER-positive breast cancer cells produce growth factors that may influence the proliferation and responsiveness of the disease (Knabbe et al., 1987). Cell responses to growth factors are mediated by cell-surface receptor tyrosine kinases (RTKs) that possess an intrinsic protein kinase activity. Ligand binding induces activation of tyrosine kinase and subsequent recruitment of target proteins, which initiate a complex signaling cascade of phosphorylation–dephosphorylation reactions (Davis, 1993; Carraway and Cantley, 1994; van der Geer et al., 1994). These signaling cascades propagate signals to the nucleus to elicit changes in gene expression (Sweeney et al., 2001). The ErbB family of RTKs includes four receptors (EGFR/ErbB1, ErbB2/HER-2/Neu, ErbB3/HER3, and ErbB4/HER4) (Muthuswamy et al., 1999) and several ligands (six EGFR ligands and two families of heregulins). Each of these ligands has a different preference to stabilize distinct receptor dimers and each ligandinduced receptor dimer signals through a unique set of pathways by recruiting a different set of effector proteins (Muthuswamy et al., 1999). Unlike homodimers, whose activities are relatively weak, heterodimers

are more potent. Heterodimers between ErbB3 and ErbB2 are the most mitogenic (Alimandi et al., 1995; Pinkas-Kramarski et al., 1996; Zhang et al., 1996; Waterman et al., 1999), leading to proliferation, growth, and transformation. The activation of ErbB RTKs eventually results in the activation of signaling molecules, the most important of which are PI 3-K and Ras (Alroy and Yarden, 1997). Overexpression of ErbBs leads to transformation (Salomon et al., 1995). EGFR and ErbB2 are overexpressed in a variety of human tumors, including breast cancers and are associated with a poor prognosis (Salomon et al., 1987; Salomon and Kidwell, 1988; Hynes and Stern, 1994) and resistance to chemo- and endocrine therapy (Nicholson et al., 1990; Wright et al., 1992; Borg et al., 1994; Leitzel et al., 1995). In addition, ErbB2 expression is inversely correlated with ER and/or progesterone receptor (PR) expression, and is generally more frequent in tumors from node-positive patients than from node-negative patients (Tandon et al., 1989, Zeillinger et al., 1989; Bacus et al., 1996). Cross coupling between the ErbB RTKs and ER signaling pathways occurs in rodent uterine tissues (Zeillinger et al., 1989) and in breast cancer cells (Pietras et al., 1995). ER-a can also modulate ErbB2 gene expression. On the other hand, activation of ErbB2 by heregulin in breast cancer cells leads to cell growth, ER-a regulation, as well as direct and rapid phosphorylation of ER-a on tyrosine residues (Pietras et al., 1995). The serine/threonine protein kinase Akt is downstream of many growth factor signaling cascades (Franke et al., 1995; Andjelkovic et al., 1996), including EGF (Burgering and Coffer, 1995; Martin et al., 2000), insulin-like growth factor-I (IGF-I) (Alessi et al., 1996; Martin et al., 2000), and heregulin (Carraway et al., 1995; Sepp-Lorenzino et al., 1996; Liu et al., 1999). These growth factors may utilize the phosphatidylinositol 3-kinase (PI 3-K)/Akt pathway to activate ER-a and confer hormone-independent growth. Apart from growth factors, estrogen also activates Akt in vascular endothelial cells (Simoncini et al., 2000) and in breast cancer cells (Stoica et al., 2003a). ER-a, but not ER-b can bind, in a ligand-dependent manner, to the regulatory subunit of PI 3-K in human vascular endothelial cells and in fibroblasts transfected with ER-a. Stimulation with physiological concentrations of estradiol increased ER-a-associated PI 3-K activity, leading to activation of Akt and endothelial nitric oxide synthase (Honda et al., 2000). These data suggest that PI 3-K is recruited and activated by a small subset of ligand-bound membrane-associated ERs. The PI 3-K signaling cascade is also involved in the neuroprotective mechanism stimulated by estradiol in rat primary cortical neurons (Lobenhofer et al., 2000). Inhibition of PI 3-K activity in MCF-7 cells prevented estrogeninduced mitogenesis and increased the activation of both activation functions, AF-1 and AF-2 of ER-a (Campbell et al., 2001). However, Akt increased the activity of only AF-1 (Martin et al., 2000; Campbell et al., 2001). Previously, we have demonstrated that the mitogenic growth factors EGF and IGF-I, as well as heregulin-b1 Oncogene


8000

(HRG-b1), can regulate ER-a gene expression and activity via Akt in the hormone-dependent breast cancer cell line, MCF-7 (Martin et al., 2000; Stoica et al., 2003b). Additionally, estradiol can rapidly activate PI 3-K/Akt through the ErbB2 pathway (Stoica et al., 2003a). In this paper, we report that the effect of estradiol on ER-a expression and activity (genomic effect) can also be modulated by ErbB2 and by Akt. Inhibitors of PI 3-K and a dominant-negative Akt or the R25C-Akt mutant inhibit the effect of estradiol on ER-a expression and activity, whereas a constitutively active Akt mutant mimics the effect of estrogen in the absence of the ligand. Selective ErbB inhibitor experiments demonstrate that the effect of estradiol on ER-a expression, activity, and on anchorage-dependent and -independent cell growth is mediated by the ErbB2 RTK. Together, our data suggest that the nongenomic effects of estradiol may complement its genomic effects, as well as cell proliferation. Results Akt can modulate the effect of estradiol on ER-a gene expression To determine whether Akt regulates ER-a gene expression, MCF-7 cells were stably transfected with a dominant-negative (K179M-Akt), a constitutively active

a

MCF-7

MCF-7 / K179M-Akt

(myr-Akt), or a mutant form of Akt with an arginine to cysteine mutation at amino acid 25 in the pleckstrin homology (PH) domain of Akt (R25C-Akt). The R25C mutation was tested because a comparable mutation (R28C) in the PH domain of Bruton’s tyrosine kinase (Btk) results in X-linked immunodeficiency in CBA/N mice (Rawlings et al., 1993). Franke et al. have shown that this PH mutation does not affect its kinase function, but NIH 3T3 cells co-transfected with this Akt mutant failed to respond to platelet-derived growth factor. The expression vector alone was employed as a negative control. Four to six clones, as well as a pool of stably transfected clones, were selected and Akt activity was characterized (Figure 1a) (Martin et al., 2000). Clones containing the constitutively active Akt were identified as m1–m4 and the pooled clone was designated as mp. Clones containing the kinase-defective mutant, K179M, were designated as K1–K4 and the pooled clone was identified as Kp. For R25C, the clones were designated R1–R6 and the pooled clone was identified as Rp (Figure 1b). In cells stably transfected with the dominant-negative mutant, Akt was inactive, and the ability of estradiol to activate Akt was inhibited (Figure 1a). Although R25C exhibits kinase activity (Franke et al., 1995), it prevented estradiol from activating Akt (Figure 1a). In cells transfected with the constitutively active mutant, the kinase was active in

MCF- 7 / R25C-A kt

MCF-7 / myr-Akt P-Akt

C

b

E2

C

E2

C

E2

C

E2

200 ER protein ER mRNA

150

% Control

*** *

100

** *

**

*

50

0 C

E2

K1 K2 Kp K1 K2 Kp R1 R5 Rp R1 R5 Rp m2 m4 m5 mp m2 m4 mp +E2

MCF-7

MCF-7 / K179M-Akt

+E2 MCF-7 / R25C-Akt

+E2 MCF-7 / myr-Akt

Figure 1 Akt modulates ER-a gene expression. (a) MCF-7 cells were stably transfected with the Akt mutants as described in Materials and methods. Pooled clones (mp, Kp, and Rp) were selected, serum starved, and treated with 109 m estradiol (E2) for 10 min. Western blot analysis was performed using an antiphospho-Akt (S473) antibody. Representative immunoblots of three independent experiments. (b) Parental and stably transfected MCF-7 cells with myr-Akt (clones m2, m4, m5, and mp), K179M-Akt (clones K1, K2, and Kp), or with R25C-Akt (clones R1, R5, and Rp) were grown in IMEM supplemented with 5% CCS. At 80% confluence, the medium was changed to phenol-red-free IMEM with 5% CCS. After 2 days, the medium was replaced with a serum-free medium and the cells were treated for 6 h with 109 m estradiol. ER-a protein was measured using an enzyme immunoassay and ER-a mRNA was determined by an RNase protection assay. Results are expressed as percent of control cells and represent the mean value of three independent experiments7s.d. Statistical differences between treatment with estradiol in parental MCF-7 and in MCF-7 cells stably transfected with K179M-Akt or with R25C-Akt were determined using the Student’s t-test, *Po0.05; **Po0.02; ***Po0.0005 Oncogene


8001 PR protein

1000

PR mRNA

% Control

800

600

400

* *

200

* *

***

***

0 C

E2

K1 K2 Kp K1 K2

Kp

R1 R5

+E2 MCF-7

MCF-7 / K179M-Akt

Rp R1 R5 Rp m2 m4 m5 mp m2 m4 m5 mp +E2

MCF-7 / R25C-Akt

+E2 MCF-7 / myr-Akt

Figure 2 Effect of Akt on regulation of the ER-a Activity. Parental and stably transfected MCF-7 cells with myr-Akt (clones m2, m4, m5, and mp), K179M-Akt (clones K1, K2, Kp), or with R25C-Akt (clones R1, R5, and Rp) were grown and treated as described in the legend of Figure 1. PR protein was measured using the enzyme immunoassay and mRNA was measured by an RNase protection assay. Results are expressed as percent of control and represent the mean value of three experiments7s.d. Statistical differences between cells in the presence of K179M-Akt or R25C-Akt (treated or nontreated with estradiol) versus in the absence of K179M-Akt or R25C-Akt were determined using the Student’s t-test as described in Figure 1

nontreated cells as measured by the amount of autophosphorylation of Akt on S473 (Figure 1a). MCF-7 cells transfected with the expression vector alone behaved similar to the parental MCF-7 cell line (Martin et al., 2000). To demonstrate that the effect of estradiol on ER-a expression was mediated by Akt, cells that were stably transfected with the Akt mutants were treated with estradiol, and the effect on ER-a protein and mRNA was measured in serum-starved cells. In the parental MCF-7 cells and in MCF-7 cells stably transfected with the empty vector (data not shown), estradiol treatment resulted in a 60 and 70% (Po0.05) decrease in total receptor protein and mRNA, respectively (Figure 1b). The expression of the kinase inactive Akt mutant in clones K1, K2, and Kp, as well as expression of R25CAkt in R1, R5, and Rp, did not alter ER-a expression, but inhibited the effect of estradiol on receptor protein and mRNA by 70–80 and 60–80% (Po0.05), respectively. The expression of constitutively active Akt mutant in clones m2, m4, m5, and mp decreased ER-a protein and mRNA by 70–80%. Addition of estradiol to these clones did not further decrease ER-a protein and mRNA. These results suggest that the regulation of ERa gene expression by estradiol is modulated by Akt and that the amino acid R25 in the PH domain of Akt plays a significant role in the interaction with ER-a. Estradiol activation of ER-a can be modulated by Akt To demonstrate that the effect of estradiol on ER-a activity can be modulated by Akt, the ability of the two

Akt mutants to inhibit the effect of estradiol was investigated in the K1, K2, and Kp and R1, R5, and Rp clones. The amounts of PR protein and mRNA were measured by an enzyme immunoassay and an RNase protection assay, respectively, and the results were compared to the parental MCF-7 cells (Figure 2). As expected, in the parental MCF-7 cells, estradiol induced PR protein and mRNA by 4- to 5.5-fold, respectively. In MCF-7 cells stably transfected with the dominantnegative or the R25C-Akt mutant, the effect of estradiol was significantly inhibited (50–60%; Po0.05 and 98– 100%; Po0.05, respectively). In MCF-7 cells stably transfected with the constitutively active mutant, a 4and a 3.5-fold increase in PR protein and mRNA, respectively, was observed in the absence of treatment with estradiol (Martin et al., 2000). Addition of estradiol to these clones did not further increase PR protein and mRNA (Figure 2). The expression vector alone did not alter the effect of estradiol on ER-a activity (data not shown), suggesting that estradiol activation of ER-a is modulated by Akt. PI 3-K and ErbB2 Inhibitors but not Anti-ErbB2 antibodies can inhibit ligand-induced ER-a gene expression and activity To determine whether PI 3-K inhibitors can block ligand-induced ER-a expression and/or activity, MCF-7 parental cells were treated with 109 m E2 for 6 h in the presence or absence of 107 m wortmannin and 105 m LY 294,002. ER-a and PR protein were measured using the enzyme immunoassay and RNase protection was Oncogene


8002 ER protein

150

ER mRNA

% Control

*** ***

100

******

*** *** **

**

****

50

0 C

E2

Tam AG30 AG825 EGF

W

LY

Tam AG30 AG825 W

+E2

LY

AG30 AbN12 AbG9 AbN12 AbG9

+EGF

+E2

Figure 3 Effect of PI 3-K, ErbB2 inhibitors, and anti-ErbB2 antibodies on ER-a gene expression. MCF-7 cells were treated with 109 m E2 or 100 ng/ml EGF in the presence or absence of the PI 3-K inhibitors wortmannin (W) (107 m) and LY 294,002 (LY) (105 m), the specific EGFR and ErbB2 inhibitors AG 30 and AG 825 (107 m), respectively, or two human monoclonal anti-ErbB2 antibodies (N12 and G9, 10 mg/ml). ER-a protein was measured using an enzyme immunoassay and ER mRNA was determined by an RNase protection assay. Results are expressed as percent of control cells and represent the mean value of three experiments7s.d. Statistical differences between estradiol- or EGF-treated cells in the absence versus in the presence of the PI 3-K inhibitors, antiestrogen, selective ErbB inhibitors, and anti-ErbB2 antibodies were determined as described in Figure 1

performed to determine ER mRNA, PR mRNA, and pS2 mRNA (Figures 3 and 4). Estradiol decreased ER-a protein and mRNA by 70% and its effect was inhibited by wortmannin and LY 294,002 by 75 and 70% (Po0.05), respectively (Figure 3). Estradiol induced PR protein, mRNA, and pS2 mRNA by 4.5-, 5-, and 3-fold, respectively, and this effect was inhibited by wortmannin and LY 294,002 by about 50% (Po0.05) (Figure 4a, b), suggesting a role for PI 3-K. Several earlier reports demonstrate that estradiol can induce growth factors in ER-positive breast cancer cells that could play a major role in signal transduction (Sporn and Roberts, 1985; Dickson and Lippman, 1987, 1995; Knabbe et al., 1987; Arteaga et al., 1988; Reddy et al., 1992, 1994; Normanno and Ciardiello, 1997). ErbB ligands could represent such candidates. To test this hypothesis, the selective inhibitors for ErbB2 and EGFR, AG825 and AG30, respectively, as well as two anti-ErbB2 human monoclonal antibodies directed against the extracellular domain of ErbB2 (clones N12 and G9-Ab N12 and AbG9), were also used to determine their effects on ER-a regulation by estradiol (Figures 3 and 4a, b). The effect of estradiol on ER-a and PR protein and on ER mRNA, PR mRNA, and pS2 mRNA was blocked by AG825, but not by AG30, suggesting a role for ErbB2, but not EGFR. Moreover, none of the two anti-ErbB2 monoclonal antibodies could block the effect of estradiol on ER and PR protein and mRNA. To determine whether N12 or G9 can block ErbB2 activation in MCF-7 cells by HRG-b1, the cells were transiently transfected with an estrogenresponsive luciferase construct. The transfected cells were treated with estradiol or HRG-b1 in the presence or absence of the two monoclonal anti-ErbB2 antibodies (N12 and G9) (Figure 4c). Following estradiol and HRG-b1 treatment, a 7- and a 2.5-fold induction in luciferase activity were observed (Po0.05). N12 and G9 did not have any effect on luciferase activity, nor on Oncogene

estradiol-treated cells. However, both antibodies were able to block the effect of HRG-b1. To determine whether N12 or G9 can block ErbB2 transactivation by ligand-bound ErbB2 in the ErbB2/ErbB3 heterodimer, we determined the effect of estradiol and HRG-b1 on Akt activity in the presence or absence of the two anti-ErbB2 antibodies (Figure 5). Following estradiol or HRG-b1 treatment (for 10 min), a 5.5- and a 7-fold Akt induction were observed. N12 or G9 did not have any effect on Akt activity, nor on estradiol-treated cells. However, both anti-ErbB2 antibodies were able to block the effect of HRG-b1 on Akt induction. To further investigate the possibility that the secreted growth factors produced by estradiol-treated cells can bind to their cell-surface receptors, regulating growth by autocrine/paracrine stimulation in our system, we determined the effect of estradiol on Akt activity in the presence of cycloheximide (CHX), a potent inhibitor of protein synthesis. When cells were treated with CHX 1 h prior to the addition of estradiol (4 h treatment), no induction of estradiol-induced Akt activity was observed (Figure 5). Taken together, these results suggest that secretion of ErbB ligands and activation of ErbB2 by an autocrine/paracrine effect upon estradiol treatment was not the mechanism of ER-a regulation. Akt and ErbB2 can influence anchorage-dependent and anchorage-independent growth and tamoxifen resistance of MCF-7 cells To determine whether Akt can also regulate anchoragedependent cell growth, parental and MCF-7 cells stably transfected with the Akt mutants were treated with 109 m estradiol in the presence or absence of the two antiestrogens tamoxifen and ICI 182,780 and cell number was determined on days 2–8 (Figure 6). In parental MCF-7 cells, as expected, estradiol stimulated the growth of cells when compared with cells grown in


8003

a

800 PR protein PR mRNA

% Control

600 400

*** * ****

200

*** * **

*** ***

******

0

C

E2

Tam AG30 AG825 EGF

W

LY

ICI AG30 AG825 W +E2

LY AG30 AbN12 AbG9 AbN12 AbG9 +EGF +E2

b 300

200

***

***

pS2mRNA, % Control

400

**

** ***

100

0

C

E2

Tam AG30 AG825 EGF W

LY

ICI AG30 AG825 W +E2

LY AG30 +EGF

Relative Luciferase Activity, % Control

c 1200 *** ***

1000

*** 800 600 * 400 200 0 C

E2 HRG-β1

N12

G9

N12

G9 +E2

N12 G9 +HRG-β1

Figure 4 Effect of PI 3-K, ErbB2 inhibitors, and anti-ErbB2 antibodies on ER-a activity. (a, b) MCF-7 cells were treated as described in Figure 3. PR protein, PR mRNA (a), and pS2 mRNA (b) were measured by an enzyme immunoassay and RNase protection, respectively. (c) Transient transfection of an estrogen-responsive luciferase reporter gene into MCF-7 cells. Transfected cells were treated for 6 h with 109 m estradiol or HRG-b1 in the presence or absence of the two monoclonal anti-ErbB2 antibodies (10 mg/ml). Luciferase activity was measured and normalized to the amount of Renilla. Results are expressed as percent control and represent the mean value of three experiments7s.d. Statistical differences between estradiol-, EGF-, or HRG-b1-treated cells in the absence versus in the presence of the PI 3-K inhibitors, antiestrogen, selective ErbB inhibitors and anti-ErbB2 antibodies were determined as described in Figure 1

an estrogen-depleted medium. Both antiestrogens could block the effect of estradiol. In MCF-7 cells stably transfected with constitutively active Akt (MCF-7/myrAkt), even without hormone treatment, cells grew to a level that was comparable with the degree of growth stimulation induced by estradiol in both parental and MCF-7/myr-Akt cell lines. 4-Hydroxytamoxifen only partially inhibited the growth of MCF-7/myr-Akt cells (50% decrease in cell number on day 8 after treatment). In contrast to 4-hydroxytamoxifen, cells were significantly growth inhibited by the pure antiestrogen ICI 182,780, which inhibits ER-a activity by preventing DNA binding and enhancing degradation (Dauvois et al., 1993). These results suggest that Akt can present estrogen-like activity on cell growth and, in the presence of constitutively active Akt, this effect can only partially be blocked by tamoxifen. In stably transfected MCF-7

cells with dominant-negative Akt (MCF-7/K179M-Akt) or with R25C-Akt, however, the effect of estradiol on cell growth were inhibited by 70–75% (Po0.05), and 40–50% (Po0.05), respectively, and both antiestrogens could block its effect. To further confirm the mitogenic effect of Akt on anchorage-dependent growth as well as the inhibitory effect of a dominant-negative Akt or of the R25C-Akt mutant, cell cycle analysis was performed (Figure 6). After 3 days of treatment with 109 m estradiol, the population of cells in the S þ G2M phases of the cell cycle increased 2.7-fold in parental MCF-7 cells. The effect of estradiol was blocked by both antiestrogens in parental cells and was inhibited significantly (to 1.3- and 1.2-fold, respectively) (Po0.05) in MCF-7 cells stably transfected with either K179M-Akt or R25C-Akt. In contrast, in MCF-7 cells stably transfected with myrOncogene


Relative Akt Activity, % Control

8004 800

600

400

200

0 C

10’ 4h HRG-β1 N12 +E2

G9

CHX AG825 AG825 N12 G9 +E2

CHX N12 G9 +HRG-β1

Figure 5 Effect of anti-ErbB2 antibodies on Akt activity. MCF-7 cells were serum starved for 24 h and treated with 109 m estradiol (for 15 min or 4 h) or HRG-b1 (for 15 min) in the presence or absence of two human monoclonal anti-ErbB2 antibodies (N12 and G9, 10 mg/ml) (20 min–1 h preincubation before treatment) or of the potent inhibitor of protein synthesis cycloheximide (CHX) (25 ng/ml) (1 h prior to the addition of 4 h estradiol treatment). Western blot of cell extracts were probed using antiphospho-Akt (P-Akt)(Ser 473) or control anti-Akt antibodies. The ratio between phosphorylated and total Akt was determined by densitometry and results are presented as percent of control cells. The experiments were repeated three times7s.d

Akt, even in the absence of estradiol, the population of cells in the S þ G2M phases of the cell cycle increased to about 3.5-fold, an effect that could be significantly inhibited by the antiestrogen ICI 182, 780. 4-Hydroxytamoxifen, however, did not block the proliferation in response to myr-Akt as well as ICI 182, 780. Taken together, these results suggest that Akt can influence cell growth and tamoxifen is not capable of fully eliminating this effect. To determine whether the selective ErbB2 inhibitor can also block the effect of estradiol and Akt on anchorage-dependent cell growth, AG825 and the two anti-ErbB2 monoclonal antibodies were added to estradiol in parental MCF-7 cells. Neither AG825 nor the anti-ErbB2 monoclonal antibodies have any effect on cell proliferation (data not shown). While AG825 could block the effect of estradiol on anchoragedependent cell growth, none of the two anti-ErbB2 antibodies (1, 5, and 10 mg/ml) (Figure 6 and data not shown) were able to influence the effect of estradiol. Moreover, AG825 could also block the effect of estradiol in stably transfected cells with the constitutively active Akt, dominant-negative Akt, or R25C-Akt (Figure 6a). In addition to anchorage-dependent growth, we also examined if Akt regulation can affect the anchorageindependent growth potential of MCF-7 cells (Figure 6c). We determined that MCF-7 cells stably transfected with constitutively active Akt showed highly increased (by 20-fold) and estradiol-treated MCF-7 cells stably transfected with either dominant-negative Akt or R25C-Akt showed greatly reduced colony-forming efficiency (60–100% reduction, respectively) in soft agar when compared with parental MCF-7 cells. Similar to the results observed before, estradiol induced a 4-8- and 2-fold increase in colony formation in parental MCF-7 and in K179M-Akt cells, respectively. No effect of estradiol was observed in MCF-7 cells stably transfected Oncogene

with R25C-Akt and addition of estradiol to cells stably transfected with constitutively active Akt did not further increase colony stimulation. The effect of estradiol was also blocked by either both antiestrogens or by the selective ErbB2 inhibitor in parental cells as well as in cells stably transfected with K179M-Akt (Figure 6c). In contrast, in MCF-7 cells stably transfected with constitutively active Akt, only ICI 182,780, and AG825 were able to block colony formation (by 92 and 100%, respectively). In this clone, tamoxifen only partially decreased colony formation in soft agar (by 66%) and inhibited the effect of estradiol by only 42%. The PI 3-K inhibitor, wortmannin also inhibited colony formation upon estradiol treatment by 54, 56, and 51%, respectively, in MCF-7 cells (parental, myr-Akt, and K179M-Akt). These results suggest that Akt can significantly increase anchorage-independent growth of MCF-7 cells and that this effect is mediated by ErbB2 and ER-a. Activation of Akt elicits downregulation of ER-a without changing the binding affinity The effect of estradiol binding to ER-a was tested in the stably transfected MCF-7 cells with constitutively active, dominant-negative, or R25C-Akt (Figure 7). Parental cells bound estradiol with both high affinity and binding capacity (Kd ¼ (2.3871.96) 1010 m). However, MCF-7/myr-Akt cells demonstrate a 67% reduction in estrogen-binding capacity (Bmax) with no change in the affinity of hormone binding (Ki ¼ (2.157 1.63) 1010 m). MCF-7/K179M and MCF-7/R25C elicit a similar Bmax and Kd with the parental cells (see table from Figure 7), suggesting that stable transfection with Akt does not modify the binding affinity of estradiol to ER-a. Akt can influence membrane ruffling in MCF-7 cells Membrane ruffles are specialized plasma membrane ultrastructures of cells, which are thought to have roles in growth, development, and locomotion (Heath and Holifield, 1991). Being associated with cell motility, membrane ruffling may be important in determining the metastatic potential of cells. Ruffles contain fine actin filaments and may be visualized using fluorescently labeled phalloidin. MCF-7 cells, cultured in an estrogendeprived medium, show occasional actin-rich membrane ruffles. Estradiol was reported to cause changes in the state of actin polymerization in these cells (Sapino et al., 1986). To determine whether Akt can modulate membrane ruffling, parental and Akt-transfected MCF-7 cells with or without estradiol treatment were grown on glass cover slips. Cells were fixed, permeabilized, and actin filaments were visualized (Figure 8a). Arrows indicate ruffles, sites of intense F-actin staining, where membranes were rapidly moving prior to fixation. Ruffling activity was confirmed by live cell imaging (data not shown). The area occupied by ruffles was selectively measured by image analysis as described in ‘Materials


8005

MCF-7 parental

MCF-7/myr-Akt

ICI E2+Tam 1000 E2+ICI E2+AG825 E2+AbN12 E2+AbG9 0

AbN12 0

1

2

3

4 5 Days

6

7

Cell Number, % control

E2

ICI

1500

E2+Tam E2+ICI

1000

E2+AG825 500

0

8

1

2

3

4 5 Days

b

E2 1500

Tam ICI

1000

E2+Tam E2+ICI

500

E2+AG825

6

7

8

C E2

1500

Tam ICI

1000

E2+Tam E2+ICI

500

E2+AG825

0 0

1

2

3

4 5 Days

6

7

0

8

1

2

3 4 Days

5

6

c MCF-7

500 400

K179M-Akt 300

MCF-7

4000

myr-Akt

R25C-Akt

200 100

myr-Akt 3000

K179M-Akt R25C-Akt

2000

500

0 C

E2

Tam

ICI

E2+Tam E2+ICI

C

E2

ICI

Tam AG825 W

ICI Tam AG825 +E2

Figure 6 Effect of Akt on anchorage-dependent and anchorage-independent growth of MCF-7 cells. (a) MCF-7, MCF-7/myr-Akt, MCF-7/K179M-Akt, and MCF-7/R27C-Akt cells were plated in six-well plates at a density of 4 105 cells/well. Following treatment with 109 m E2 in the presence or absence of tamoxifen or ICI 182,780 (5 107 m), cells were counted at different times. Results are expressed as percentage of control in the parental cell line and represent the mean value of three independent experiments performed in triplicate7s.d. (b) MCF-7, MCF-7/myr-Akt, MCF-7/K179M-Akt, and MCF-7/R27C-Akt cells were plated and grown as described under Materials and methods. After treatment for 3 days with 109 m estradiol in the presence or absence of tamoxifen or ICI 182,780 (5 107 m), cells were washed twice with PBS, resuspended in 100 ml citrate/DMSO buffer, and DNA analysis was performed after Vindelov staining. The values represent the mean of three independent assays7s.d. (c) Serum-stripped MCF-7 cells (parental and stably transfected with myr-Akt, K179M-Akt, or R25C-Akt) were seeded into 0.35% soft agar with either 109 m estradiol, in the presence or absence of 107 m tamoxifen or ICI 182,780, 5 107 m wortmannin, or 107 m AG825 over a 0.6% agar bottom layer and maintained at 371C. Colonies were quantified after 7–10 days. Each experiment was performed in triplicate. Data shown are expressed as a percent of colonies formed by parental MCF-7 cells and are an average of three independent experiments7s.d 0.3 Bs/FC 0.25 Specific Binding/Free

G2M+S, % Control

C

0 0

AbG9

2000

2000

Tam

Colony Formation, % Control


Tam

500

MCF-7/R25C-Akt Cell Number, % control

2000

E2 1500

MCF-7/K179M-Akt

C

C

2000


a

Bs/Fmyr Bs/FK179M

0.2

Bs/F R25C 0.15 0.1

Cell Line

Kd X 10-10 M

Bmax (% Control)

MCF-7

2.38 +/-1.96

100

MCF-7/myr-Akt

2.15 +/-1.63

33 +/-9

MCF-7/K179M-Akt

2.57 +/-2.12

83 +/-31

MCF-7/R25C

0.98 +/-0.02

96+/-8

0.05 0 0

2

4 6 8 10 Specific Binding (pM)

12

Figure 7 Effect of Akt on estradiol binding to the ER-a. Scatchard analyses of control MCF-7, MCF-7/myr-Akt, MCF-7/K179MAkt, and MCF-7/R25C-Akt cells were performed to evaluate estradiol binding capacity (Bmax) and binding activity (Kd). Cells were grown in six-well plates and incubated with increasing concentrations of 3[H]E2 (1012–107 m) in the absence (BT) or presence (BN) of a 200-fold molar excess of diethylstilbestrol. BS ¼ BTBN, where BT is total binding, BN is nonspecific binding, and BS is specific binding. A representative assay is shown. The experiment was repeated three times Oncogene


8006

a MCF-7/parental

MCF-7/myr-Akt

MCF-7/K179M-Akt

MCF-7/R25C-Akt

C

E2

b

***

Ruffling Activity

30

** 20

*

10

0 C E2 MCF-7

C E2 myr-Akt

C E2 K179M-Akt

C E2 R25C-Akt

Figure 8 Effect of Akt on membrane ruffling activity. Cells (parental (first row) or MCF-7 cells stably transfected with myrAkt (second row), K179M-Akt (third row), and R25C-Akt (fouth row)) were grown on glass cover slips in an IMEM medium supplemented with 5% CCS until semiconfluent. The medium was changed to phenol-red-free for 24 h and then cells were serum starved by further incubation for 24 h in a serum-free medium. They were stimulated for another 24 h with 109 m estradiol (second panel in each row), fixed, and stained as described in Materials and methods. (a) Representative image of three independent experiments. Three-dimensional renderings of confocal z-stacks showing F-actin labeling. Arrows indicate ruffles. Scale bars ¼ 100 mm. (b) Quantification of membrane ruffling activity. The area occupied by ruffles was measured, nuclei counted, and the ruffling activity was expressed as area of ruffles per nuclei7s.d. Asterisks indicate significant differences (Po0.05) compared with MCF-7 control (MCF-7 C)

and methods’ and ruffling activity was expressed as the area of ruffles per nuclei. As expected, in parental MCF7 cells and in MCF-7 cells stably transfected with the empty vector (data not shown), few ruffles could be observed (Figure 8, first row, first panel). Estradiol treatment significantly enhanced the number of ruffles (1.7-fold, P ¼ 0.0449) (Figure 8a, first row, second panel and Figure 8b). MCF-7 cells stably transfected with constitutively active Akt, even in the absence of Oncogene

treatment, showed significantly more membrane ruffles than the parental cells (2.25-fold increase, P ¼ 0.0179) (Figure 8a, second row, first panel and Figure 8b). Estradiol treatment in these clones further increased the ruffles (threefold increase, P ¼ 0.0155) (Figure 8a second row, second panel and Figure 8b). In cells transfected with dominant-negative Akt, the number of ruffles decreased upon estradiol treatment by 54%, P ¼ 0.0124 (Figure 8a third row, second panel and Figure 8b). However, in R25C-Akt transfected cells, no difference in membrane ruffling was observed as compared to parental cells both in the absence as well as in the presence of estradiol (Figure 8a, fourth row, both panels and Figure 8b). Taken together, these results suggest that Akt can modulate membrane ruffling.

Discussion The actions of estrogens are required for the development and maintenance of reproductive tissue, as well as for growth and survival of breast cancer. Although a large body of evidence exists showing that the ER plays a central role in cell proliferation (Vignon et al., 1987; Auricchio et al., 1995; Pietras et al., 1995), the signaling mechanisms responsible for MAPK and Akt activation by estradiol are not fully characterized. The current dogma of steroid hormone action is that estradiol diffuses through the cell membrane into the cytoplasm and enters the nucleus by passive diffusion (GuiochonMantel et al., 1996). It binds directly to the ER, which homodimerizes and interacts with DNA to stimulate transcription and production of growth factors and other proteins (Tsai and O’Malley, 1944; Reddy et al., 1994; Beato et al., 1995). Consistent with this model of estrogen action in MCF-7 cells exposed to estradiol, a slow (more than 2 h of estradiol exposure) but sustained activation of Erk up to 24 h was reported (Keshamouni et al., 2002). In addition to these long-term (genomic) estrogenic effects, we and others have demonstrated some effects of estrogen that occur within minutes after estradiol administration (nongenomic effects) (Nemere and Farach-Carson, 1998; Watson and Gametchu, 1999; Molloy et al., 2000; Simoncini et al., 2000; Stoica et al., 2003a). Estradiol can rapidly activate MAPK (Kelly and Levin, 2001; Song et al., 2002) or Akt in several cell lines, including breast cancer (Castoria et al., 2001; Marquez and Pietras, 2001; Stoica et al., 2003a). In this study, we report that the effect of estradiol on ER-a expression and activity (genomic effect) can be modulated by the ErbB2/PI 3-K/Akt pathway. Inhibitors of PI 3-K and a dominant-negative Akt mutant significantly inhibited the effect of estradiol, whereas a constitutively active Akt mutant mimicked the effect of estradiol in the absence of ligand. The effect of estradiol on ER-a expression and activity was abrogated by an arginine to cysteine mutation in the PH domain of Akt, suggesting the involvement of this amino acid in the interaction between ER-a and Akt. Evidence for


8007 E2 ErbB2-ErbB3 (nonheterodimer ? genomic effects) membrane ER-α-E2 complex P P

P p85 P

PIP3 PLPase

p110

PIP2

PI 3-K

Akt-K (PDK1,2)

PH

Akt

(genomic effects)

(Ser 473, Thr 308)

KD

membrane

RD cytoplasm P P P

? ER-α α

ER− α − E 2

Ser 104 Ser 106 Ser 118 Ser 167 ( RERLAS )

? P P P

nucleus

P P P

transcription GENE

transcription GENE

Figure 9 Molecular mechanism of modulation of ER-a by Akt. Estradiol, bound to membrane ER-a may interact with a heterodimer containing ErbB2, inducing its activation (nongenomic effects). The heterodimer may activate PI 3-K, which in turn can activate Akt. Akt may phosphorylate nuclear ER-a, leading to the modulation of its expression and activity (genomic effects). In this model, the nongenomic effects may complement and/or synergize with the genomic effects

distinct signaling pathways that are coupled is accumulating. We have previously shown that the cross talk between the EGF and IGF-I, as well as the HRG-b1 signal transduction pathways is also mediated by Akt in MCF-7 cells (Martin et al., 2000; Stoica et al., 2003b). Treatment of these cells with either EGF, IGF-I, or HRG-b1, just like estradiol, decreased ER-a protein and mRNA (Stoica et al., 2000a, b) and increased ER-a activity (Stoica et al., 2000a, b) and these effects were blocked by a dominant-negative Akt mutant (Martin et al., 2000; Stoica et al., 2003b). Similar to estradiol, EGF, IGF-I, and HRG-b1 treatment, the expression of a constitutively active Akt decreased the expression of ER-a protein and increased the expression of PR and pS2, providing additional evidence that Akt is a mediator of the estradiol, EGF, IGF-I, and HRG-b1 pathways. Using ER-a mutants, we have demonstrated that Akt activates ER-a through particular serines in the AF-1 domain of the receptor (Martin et al., 2000). Akt activation in response to growth factors such as EGF, IGF-I, HRG-b1, platelet-derived growth factor results via growth factor receptor/PI 3-K (Burgering and Coffer, 1995; Carraway et al., 1995; Franke et al., 1995; Alessi et al., 1996; Andjelkovic et al., 1996; SeppLorenzino et al., 1996; Liu et al., 1999; Martin et al., 2000; Hellyer et al., 2001; Stoica et al., 2003b). In this study, we show that the blockade of the ErbB2, but not the EGFR signaling, inhibits the effect of estradiol on ER-a expression and activity. Moreover, we demonstrate that the genomic effect of estradiol was not inhibited by human monoclonal anti-ErbB2 antibodies directed to the extracellular domain nor by the potent inhibitor of protein synthesis cycloheximide, suggesting

that estradiol-induced growth factors (such as HRG-b1) are not needed to bind to cell surface ErbB2 to activate signaling by autocrine and/or paracrine mechanism. In contrast, the effect of estradiol on Erk activation was shown to be a slow, long-lasting effect, dependent on new protein synthesis in the same MCF-7 cells (Keshamouni et al., 2002). Our results suggest that the nongenomic effects of estradiol are coupled and/or synergize with its genomic effects. Estradiol may mimic the actions of HRG-b1 (Stoica et al., 2003a, b). It binds to membrane ER-a and interacts with a heterodimer with ErbB2. Since MCF-7 cells overexpress ErbB3, but not ErbB4 (Wosikowski et al., 1997), and heterodimers between ErbB3 and ErbB2 are the most mitogenic (Sporn and Roberts, 1985; Knabbe et al., 1987; Davis, 1993; Dickson and Lippman, 1995), we expect that estradiol may activate this heterodimer. Activated ErbB3 may recruit and activate PI 3-K via phosphorylation of YXXM motifs in the ErbB3 carboxy-terminal domain (Kim et al., 1994; Prigent and Gullick, 1994; Soltoff et al., 1994; Migliaccio et al., 1996). Activated PI 3-K, in turn, leads to Akt activation. Akt may phosphorylate ER-a on serines S104, S106, S118, and S167 (Martin et al., 2000), resulting in alteration of gene expression and activity (Figure 9). A cross talk between ErbB2 and the androgen receptor pathway has also been reported (Craft et al., 1999). ErbB2, in the absence of androgen, promotes androgen-independent survival and growth of prostate cancer cells through the Akt pathway and Akt specifically binds to the androgen receptor, phosphorylating serines S213 and S791 (Wen et al., 2000). Therefore, the ErbB2/Akt/steroid receptor cross talk appears to be a more general mechanism in hormone-dependent cancer settings. Oncogene


8008

Our results indicate that tamoxifen can block the effect of estrogen, but not the effect of Akt. Constitutively active Akt leads to an increase in cell number, in the population of cells in the S þ G2M phases of the cell cycle, in colony formation in soft agar, and in membrane ruffling activity. The effect of constitutively active Akt on anchorage-dependent and -independent cell growth can only partially be inhibited by 4hydroxytamoxifen but can totally be blocked by ICI 182, 780. The effect of estradiol on anchorage-dependent and -independent cell growth can be significantly inhibited by a dominant-negative Akt mutant, by R25CAkt (only anchorage-dependent growth), or by a selective ErbB2 inhibitor. However, only the dominant-negative Akt mutant and not R25C-Akt could inhibit ruffling activity, suggesting that the kinase activity is necessary for membrane ruffling. Consistent with our results, activation of the PI 3-K/Akt pathway in the same breast cancer cell line, MCF-7, leads to Bcl-2 expression, which correlates with 4-hydroxytamoxifen resistance (Lobenhofer et al., 2000) and ErbB2-mediated 4-hydroxytamoxifen resistance correlates with Bcl-2 over expression (Wen et al., 2000). The estradiol stimulation of Akt also leads to cyclin D1 transcription and the progression through the G1-S phase of the cell cycle. A hypothetical model of association of ER-a with Src and p85 triggered by estradiol was proposed. The ternary complex between ER-a, p85, and Src may favor hormone activation of Src- and PI 3-K-dependent pathways, which converge on cell cycle progression (Marquez and Pietras, 2001). The MAPK pathway is also thought to play an important role in mediating the effect of estradiol on cell proliferation and activation of ER-a. Although the signaling mechanism(s) responsible for MAPK activation are not fully characterized, estradiol induces a slow but persistent activation of Erk in MCF-7 breast cancer cells, similar to Akt activation, but predominantly mediated through the secretion of heregulin and activation of ErbB2 by an autocrine/paracrine mechanism (Beato et al., 1995). The ability of both pathways to affect ER-a function suggests that ER is a point of convergence of the Akt and MAPK pathways. It is not clear yet as to whether the Akt and MAPK pathways are alternate pathways or whether they cross talk with each other. The interaction between growth factors and estrogen signaling is complex and occurs at multiple levels. Our results show for the first time that ErbB2/ PI-3K/Akt mediated signaling plays a major role in estrogen-mediated signaling in hormone-dependent breast cancer and that activation of this pathway can set the stage for later genomic actions of estradiol. The nongenomic effect of estradiol may complement and/or synergize with its genomic actions. Besides the insights into novel interactions between ER-a and growth factor signaling cascades, interruption of the ErbB2 and Akt signaling pathways may enhance the inhibitory effects of antiestrogens on both ER-a mediated transcription and on tumor cell proliferation. Oncogene

Materials and methods Cell culture Monolayer cultures of MCF-7 cells were grown in improved minimal essential medium (IMEM), supplemented with 5% fetal calf serum (FCS). At 80% confluence, the medium was replaced with phenol-red-free IMEM containing 5% charcoaltreated calf serum (CCS) (Berthois et al., 1986). The calf serum was pretreated with sulfatase and dextran-coated charcoal to remove endogenous steroids (Saceda et al., 1988). After 2 days, the medium was changed into serum-free, phenol-red-free IMEM supplemented with fibronectin, glutamine, HEPES, trace elements, and transferrin, after which 109 m estradiol or 100 ng/ml epidermal growth factor (EGF) was added. Cells were harvested at the times indicated. EGF was purchased from Upstate Biotechnology (Lake Placid, NY, USA); estradiol and wortmannin were purchased from Sigma (St Louis, MO, USA). The selective ErbB receptor inhibitors AG 30 (for EGFR) and AG 825 (for ErbB2) and LY 294,002 were purchased from Calbiochem (San Diego, CA, USA). Monoclonal human anti-ErbB2 antibodies (Ab-4, clones N12 (Stancovski et al., 1991) and G9) are directed to the external domain of ErbB2 and were purchased from Neomarkers (Fremont, CA, USA). Plasmids The probe for the ER, pOR-300, was constructed by subcloning a 300-bp restriction fragment of pOR3 into the pGem4 polylinker regions using the restriction enzymes PstI and Eco RI (Saceda et al., 1988). The clone 36B4 was constructed by subcloning a 220-bp fragment of 36B4 into the PstI restriction site of the pGem polylinker (Saceda et al., 1988). In addition, the clones for pS2 (Garcia-Morales et al., 1994), glyceraldehyde-3-phosphate dehydrogenase (GAPDH) (McLeskey et al., 1993), and PR (Stoica et al., 2000c) were used as previously described. The expression vector for Akt (HA-Akt) (Ahmed et al., 1993), the kinase defective Akt (HA-Akt (K179M)) (Franke et al., 1995), the constitutively active Akt (myrAkt-HA) (Franke et al., 1995), and the Akt mutant carrying an arginine to cysteine mutation at amino acid 25 Akt (R25C) (Franke et al., 1995) mutants were generated as HindIII-BamHI inserts in pCMV-6. ER-a and PR protein assays For analysis of ER-a and PR protein concentration, MCF-7 cells were cultured and treated as described above. The concentration of receptor protein was determined using an enzyme immunoassay kit from Abbott Laboratories (North Chicago, IL, USA) (Saceda et al., 1988). To obtain total receptor protein, the cells were homogenized by sonication in a high salt buffer (10 mm Tris, 1.5 mm EDTA, 5 mm Na2MoO4, 0.4 m KCl, and 1 mm monothioglycerol with 2 mm leupeptin). The homogenate was incubated on ice for 30 min and centrifuged at 100 000g for 1 h at 41C. Aliquots of the total extracts were then analysed according to the manufacturer’s instructions. ER-a binding assays To measure the number of estrogen-binding sites (Bmax) and the dissociation constant (Kd) of the estradiol – ER-a complex in the parental or stably transfected MCF-7 cells with either a constitutively active, a dominant-negative, or an Akt mutant carrying an arginine to cysteine mutant at amino acid 25; a


8009 whole cell, multiple-dose ligand binding assay was used (Stoica et al., 2000c). Cells were plated in six-well plates. At approximately 70% confluence, the cells were incubated for 2 h with various concentrations of [3H] estradiol. A 200-fold excess of diethylstilbestrol was added to distinguish between specific and nonspecific binding. The cells were washed and lysed. The protein content and radioactivity present in each sample was quantified and analysed using the Scatchard method (Scatchard, 1949). Measurement of ER-a mRNA and PR mRNA Total cellular RNA was extracted from MCF-7 cells by the RNazol method. The amounts of ER-a, 36B4, PR, and GAPDH were determined by an RNase protection assay (Saceda et al., 1988). Briefly, homogeneously 32P-labeled antisense cRNA were synthesized in vitro from pOR-300, 36B4, and pGAPDH using T7 polymerase and from PR using SP6 polymerase. Total RNA of 60 mm were hybridized for 12– 16 h to the radiolabeled cRNA. After a 30 min digestion at 251C with RNase A, 32P-labeled cRNA probes protected by total RNA were separated by electrophoresis on 6% polyacrylamide gels. The bands were visualized by autoradiography and quantified using the phospho imager. The amounts of ER-a mRNA and PR mRNA were normalized to the internal control 36B4 and GAPDH, respectively. Anchorage-dependent growth assays MCF-7 cells (parental or stably transfected with myr-Akt, K179M-Akt, or R25C-Akt) were plated at 105 cells/well into six-well plates in IMEM supplemented with 5% FCS. Cells were grown to 40% confluence and the medium was changed to phenol-red-free IMEM supplemented with 5% CCS. After 2 days in this medium, cells were treated with 109 m estradiol in the presence or absence of the antiestrogens ICI 182, 780 or tamoxifen (5 107 m). Medium with the appropriate treatments was replaced every 3 days. Cells were trypsinized at the specific time points and counted with a coulter counter (Coulter Electronics, Inc., Hialeah, FL, USA). Anchorage-independent growth assays The ability of MCF-7 cells (parental or stably transfected with myr-Akt, K179M-Akt, or R25C-Akt) to stimulate anchorageindependent growth was tested by a soft agar assay as described earlier (Wellstein et al., 1990, 1992). Briefly, 20 000 cells (stripped of endogenous steroids, suspended in IMEM phenol-red-free medium supplemented with 5% CCS) in 0.35% bactoagar with different treatments were grown on a bottom layer of solidified 0.6% bactoagar in 35 mm dishes. After 7–10 days, the colonies in the top layer were counted using an inverted microscope equipped with a measuring grid. The size exclusion limit for positive colony counting was 460 mm diameter. The colony formation induced by control cells was used as background and was in the range of 5–10 colonies, depending on the experiment. Cell cycle phase analysis For cell cycle analysis, MCF-7 cells (parental, or stably transfected with the Akt mutants: myr-Akt, K179M-Akt, or R25C-Akt) were grown as described above and treated for 3 days with 109 m estradiol in the presence or absence of the antiestrogens ICI 182,780 or hydroxytamoxifen (5 107 m). The cells were trypsinized to obtain a single cell suspension and verified microscopically after neutralizing with a medium containing serum. The cell number was adjusted to 2 106

cells for each treatment. The cells were washed twice with PBS, centrifuged at 1000 rpm for 5 min and suspended in 0.1 ml citrate/dimethylsulfoxide buffer. The DNA content was measured by Vindelov staining (Vindelov et al., 1983). Western blot analysis MCF-7 cells (parental, stably transfected with either dominant-negative Akt mutant or R25C-Akt) were treated with estradiol (109 m) for 10 min. Cells were lysed in NP-40 lysis buffer, the lysates were heated to 95–1001C for 5 min and equal amounts of protein (100 mg) were loaded onto SDS–polyacrylamide gels. Gels were electrotransferred to nitrocellulose membranes and washed in phosphate-buffered saline five times at room temperature. Membranes were kept in a blocking buffer overnight at 41C and incubated with the primary antibody (antiphospho-Akt antibody (Ser 473)) for 1 h at room temperature. After three additional washes in phosphate-buffered saline, membranes were incubated with the horseradish peroxidase-conjugated secondary antibody (1 : 10 000) in blocking buffer for 1 h at room temperature. Detection was performed by chemiluminescence, using Super Signal chemiluminescent substrate (Pierce, Rockford, IL, USA). Transfections Stable and transient transfection of MCF-7 cells (passage 47) were performed with lipofectamine Plus (Gibco/BRL, Rockville, MD, USA) according to the manufacturer’s instructions and is described elsewhere (Martin et al., 2000). Stably transfected cells were selected in IMEM supplemented with 10% FCS and 500 mg/ml G418 for about 1 month. Colonies (4–6) (clones K1 and K2 for K179M; clones m2, m4, m5, and m6 for myr-Akt; and R1, R3 for R25C), as well as pools of transfected colonies (Kp for K179M, mp for myr-Akt, and Rp for R25C) were picked up, serum starved, and treated with 109 m estradiol and enzymatic activity of Akt was also tested. For transient transfections, MCF-7 cells were plated at a density of 1 106 cells/100-mm dish in phenol-red IMEM containing 10% CCS for 24 h. The medium was changed to phenol-red-free medium containing 10% CCS for 48 h and the cells were transfected with 5 mg of an estrogen-responsiveluciferase reporter and 0.2 mg of Renilla. At 16–18 h after transfection, cells were replenished with phenol-red-free IMEM containing 10% CCS and treated for 6 h with 109 m estradiol or HRG-b1 in the presence or absence of two monoclonal anti-ErbB2 antibodies (N12 and G9). The cells were harvested 24 h later, and luciferase activity was measured. Luciferase activity was normalized to the activity of Renilla. The increase in luciferase activity in response to treatment is expressed relative to untreated controls. Membrane ruffling assay Cells (parental MCF-7 or stably transfected MCF-7 with the empty vector, myr-Akt, K179M-Akt, or R25C-Akt) were grown on glass cover slips in IMEM supplemented with 5% CCS until semiconfluent. The medium was changed to phenolred-free for 24 h and then cells were serum starved by further incubation overnight in a serum-free medium. Cells were stimulated for 24 h with 109 m estradiol and then fixed in 3.7% formaldehyde in PBS for 10 min. Cells were postpermeabilized in 0.5% Triton X-100 in PBS for 5 min and then washed with PBS. Polymerized actin (F-actin) was visualized by incubating with phalloidin-tetramethyl rhodamine isothiocyanate (TRITC) for 20 min. Nuclei were simultaneously labeled with Oncogene


8010 ToPro3 (Molecular Probes, Eugene, OR, USA). Cells were washed three times in PBS, mounted using Anti-Fade (Molecular Probes) and viewed using an Olympus Fluoview laser scanning confocal microscope equipped. Images were collected using identical settings on the confocal microscope, and then were subjected to image analysis using Metamorph Offline (Universal Imaging). Images were thresholded for actin ruffles on the surfaces and margins of cells, excluding background and other F-actin positive structures with less intense staining. The total area occupied by the ruffles and the number of nuclei in each image was determined from three

randomly collected images per treatment. The ruffling activity was expressed as the area occupied by ruffles per nuclei. Acknowledgements We thank Drs S Byers and D Gamett for critically reading the manuscript. This work was supported by grants from Milheim Foundation, and Georgetown University (Dean of Research) (to AS) and, in part, by CDA-DAMD17-00-1-0214 (to TFF). Support for tissue culture, cell cycle analysis, microscopy, and imaging core facilities was provided by P50-CA-58185 and P30-CA-51008.

References Ahmed NN, Franke TF, Bellacosa A, Datta K, GonzalezPortal ME, Taguchi T, Testa JR and Tsichlis PN. (1993). Oncogene, 8, 1957–1963. Alessi DR, Andjelkovic M, Caudwell B, Cron P, Morrice N, Cohen P and Hemmings BA. (1996). EMBO J., 15, 6541–6551. Alimandi M, Romano A, Curia MC, Muraro R, Fedi P, Aaronson SA, Di Fiore PP and Kraus MH. (1995). Oncogene, 10, 1813–1821. Alroy I and Yarden Y. (1997). FEBS Lett., 410, 83–86 (review). Andjelkovic M, Jakubowicz T, Cron P, Ming XF, Han JW and Hemmings BA. (1996). Proc. Natl. Acad. Sci. USA, 93, 5699–5704. Arteaga CL, Coronado E and Osborne CK. (1988). Mol. Endocrinol., 2, 1064–1069. Auricchio F, Di Domenico M, Migliaccio A, Castoria G and Bilancio A. (1995). Cell Growth Differ., 6, 105–113. Bacus SS, Chin D, Yarden Y, Zelnick CR and Stern DF. (1996). Am. J. Pathol., 148, 549–558. Beato M, Herrlich P and Schutz G. (1995). Cell, 83, 851–857 (review). Berry M, Metzger D and Chambon P. (1990). EMBO J., 9, 2811–2818. Berthois Y, Katzenellenbogen JA and Katzenellenbogen BS. (1986). Proc. Natl. Acad. Sci. USA, 83, 2496–2500. Borg A, Baldetorp B, Ferno M, Killander D, Olsson H, Ryden S and Sigurdsson H. (1994). Cancer Lett., 81, 137–144. Burgering BM and Coffer PJ. (1995). Nature, 376, 599–602. Campbell RA, Bhat-Nakshatri P, Patel NM, Constantinidou D, Ali S and Nakshatri H. (2001). J. Biol. Chem., 276, 9817–9824. Carraway III KL and Cantley LC. (1994). Cell, 78, 5–8 (review). Carraway III KL, Soltoff SP, Diamonti AJ and Cantley LC. (1995). J. Biol. Chem., 270, 7111–7116. Castoria G, Migliaccio A, Bilancio A, Di Domenico M, de Falco A, Lombardi M, Fiorentino R, Varricchio L, Barone MV and Auricchio F. (2001). EMBO J., 20, 6050–6059. Craft N, Shostak Y, Carey M and Sawyers CL. (1999). Nat. Med., 5, 280–285. Dauvois S, White R and Parker MG. (1993). J. Cell Sci., 106 (Part 4), 1377–1388. Davis RJ. (1993). J. Biol. Chem., 268, 14553–14556 (review). Dickson RB and Lippman ME. (1987). Endocr. Rev., 8, 29–43 (review). Dickson RB and Lippman ME. (1995). Endocr. Rev., 16, 559– 589 (review). Franke TF, Yang SI, Chan TO, Datta K, Kazlauskas A, Morrison DK, Kaplan DR and Tsichlis PN. (1995). Cell, 81, 727–736. Oncogene

Garcia-Morales P, Saceda M, Kenney N, Kim N, Salomon DS, Gottardis MM, Solomon HB, Sholler PF, Jordan VC and Martin MB. (1994). J. Biol. Chem., 269, 16896–16901. Guiochon-Mantel A, Delabre K, Lescop P and Milgrom E. (1996). J. Steroid Biochem. Mol. Biol., 56 (1–6 Spec No), 3–9 (review). Heath JP and Holifield BF. (1991). Cell Motil. Cytoskel., 18, 245–257 (review). Hellyer NJ, Kim MS and Koland JG. (2001). J. Biol. Chem., 276, 42153–42161. Honda K, Sawada H, Kihara T, Urushitani M, Nakamizo T, Akaike A and Shimohama S. (2000). J. Neurosci. Res., 60, 321–327. Hynes NE and Stern DF. (1994). Biochim. Biophys. Acta., 1198, 165–184 (review). Johnston SR. (1997). Anticancer Drugs, 8, 911–930 (review). Kato S. (2001). Breast Cancer, 8, 3–9 (review). Kato S, Kitamoto T, Masuhiro Y and Yanagisawa J. (1998). Oncology, 55 (Suppl. 1), 5–10 (review). Katzenellenbogen BS, Montano MM, Ekena K, Herman ME and McInerney EM. (1997). Breast Cancer Res. Treat., 44, 23–38 (review). Kelly MJ and Levin ER. (2001). Trends Endocrinol. Metab., 12, 152–156 (review). Keshamouni VG, Mattingly RR and Reddy KB. (2002). J. Biol. Chem., 277, 22558–22565. Kim HH, Sierke SL and Koland JG. (1994). J. Biol. Chem., 269, 24747–24755. Knabbe C, Lippman ME, Wakefield LM, Flanders KC, Kasid A, Derynck R and Dickson RB. (1987). Cell, 48, 417–428. Leitzel K, Teramoto Y, Konrad K, Chinchilli VM, Volas G, Grossberg H, Harvey H, Demers L and Lipton A. (1995). J. Clin. Oncol., 13, 1129–1135. Liu W, Li J and Roth RA. (1999). Biochem. Biophys. Res. Commun., 261, 897–903. Lobenhofer EK, Huper G, Iglehart JD and Marks JR. (2000). Cell Growth Differ., 11, 99–110. Mangelsdorf DJ and Evans RM. (1995). Cell, 83, 841–850 (review). Marquez DC and Pietras RJ. (2001). Oncogene, 20, 5420–5430. Martin MB, Franke TF, Stoica GE, Chambon P, Katzenellenbogen BS, Stoica BA, McLemore MS, Olivo SE and Stoica A. (2000). Endocrinology, 141, 4503–4511. McDonnell DP, Clemm DL, Hermann T, Goldman ME and Pike JW. (1995). Mol. Endocrinol., 9, 659–669. McLeskey SW, Kurebayashi J, Honig SF, Zwiebel J, Lippman ME, Dickson RB and Kern FG. (1993). Cancer Res., 53, 2168–2177. Migliaccio A, Di Domenico M, Castoria G, de Falco A, Bontempo P, Nola E and Auricchio F. (1996). EMBO J., 15, 1292–1300.


8011 Molloy CA, May FE and Westley BR. (2000). J. Biol. Chem., 275, 12565–12571. Muthuswamy SK, Gilman M and Brugge JS. (1999). Mol. Cell. Biol., 19, 6845–6857. Nemere I and Farach-Carson MC. (1998). Biochem. Biophys. Res. Commun., 248, 443–449 (review). Nicholson S, Wright C, Sainsbury JR, Halcrow P, Kelly P, Angus B, Farndon JR and Harris AL. (1990). J. Steroid Biochem. Mol. Biol., 37, 811–814. Normanno N and Ciardiello F. (1997. J. Mammary Gland Biol. Neoplasia, 2, 143–151. Pietras RJ, Arboleda J, Reese DM, Wongvipat N, Pegram MD, Ramos L, Gorman CM, Parker MG, Sliwkowski MX and Slamon DJ. (1995). Oncogene, 10, 2435–2446. Pinkas-Kramarski R, Soussan L, Waterman H, Levkowitz G, Alroy I, Klapper L, Lavi S, Seger R, Ratzkin BJ, Sela M and Yarden Y. (1996). EMBO J., 15, 2452–2467. Prigent SA and Gullick WJ. (1994). EMBO J., 13, 2831–2841. Rawlings DJ, Saffran DC, Tsukada S, Largaespada DA, Grimaldi JC, Cohen L, Mohr RN, Bazan JF, Howard M and Copeland NG, Jenkins NA and Witte ON. (1993). Science, 261, 358–361. Reddy KB, Mangold GL, Tandon AK, Yoneda T, Mundy GR, Zilberstein A and Osborne CK. (1992). Cancer Res., 52, 3636–3641. Reddy KB, Yee D, Hilsenbeck SG, Coffey RJ and Osborne CK. (1994). Cell Growth Differ., 5, 1275–1282. Saceda M, Lippman ME, Chambon P, Lindsey RL, Ponglikitmongkol M, Puente M and Martin MB. (1988). Mol. Endocrinol., 2, 1157–1162. Salomon DS and Kidwell WR. (1988). Cancer Treat Res., 40, 363–389 (review). Salomon DS, Normanno N, Ciardiello F, Brandt R, Shoyab M and Todaro GJ. (1995). Breast Cancer Res. Treat., 33, 103–114 (review). Salomon DS, Perroteau I, Kidwell WR, Tam J and Derynck R. (1987). J. Cell Physiol., 130, 397–409. Sapino A, Pietribiasi F, Bussolati G and Marchisio PC. (1986). Cancer Res., 46, 2526–2531. Scatchard G. (1949). Ann. NY Acad. Sci. USA, 51, 209–216. Sepp-Lorenzino L, Eberhard I, Ma Z, Cho C, Serve H, Liu F, Rosen N and Lupu R. (1996). Oncogene, 12, 1679–1687. Simoncini T, Hafezi-Moghadam A, Brazil DP, Ley K, Chin WW and Liao JK. (2000). Nature, 407, 538–541. Soltoff SP, Carraway III KL, Prigent SA, Gullick WG and Cantley LC. (1994). Mol. Cell. Biol., 14, 3550–3558. Song RX, McPherson RA, Adam L, Bao Y, Shupnik M, Kumar R and Santen RJ. (2002). Mol. Endocrinol., 16, 116– 127. Sporn MB and Roberts AB. (1985). Nature, 313, 745–747 (review).

Stancovski I, Hurwitz E, Leitner O, Ullrich A, Yarden Y and Sela M. (1991). Proc. Natl. Acad. Sci. USA, 88, 8691–8695. Stoica A, Saceda M, Doraiswamy VL, Coleman C and Martin MB. (2000a). J. Endocrinol., 165, 371–378. Stoica A, Saceda M, Fakhro A, Joyner M and Martin MB. (2000b). J. Cell Biochem., 76, 605–614. Stoica A, Katzenellenbogen BS and Martin MB. (2000c). Mol. Endocrinol., 14, 545–553. Stoica EG, Franke TF, Wellstein A, Czubayko F, List HJ, Reiter R, Morgan E, Martin MB and Stoica A. (2003a). Mol. Endocrinol., 17, 818–830. Stoica EG, Franke TF, Wellstein A, Morgan E, Czubayko F, List HJ, Reiter R, Martin MB and Stoica A. (2003b). Oncogene, 22, 2073–2087. Sweeney CJ, Zhu J, Sandler AB, Schiller J, Belani CP, Langer C, Krook J, Harrington D and Johnson DH. (2001). Cancer, 92, 2639–2647. Tandon AK, Clark GM, Chamness GC, Ullrich A and McGuire WL. (1989). J. Clin. Oncol., 7, 1120–1128. Tsai MJ and O’Malley BW. (1994). Annu. Rev. Biochem., 63, 451–486 (review). van der Geer P, Hunter T and Lindberg RA. (1994). Annu. Rev. Cell Biol., 10, 251–337 (review). Vasudevan N, Kow LM and Pfaff DW. (2001). Proc. Natl. Acad. Sci. USA, 98, 12267–12271. Vignon F, Bouton MM and Rochefort H. (1987). Biochem. Biophys. Res. Commun., 146, 1502–1508. Vindelov LL, Christensen IJ and Nissen NI. (1983). Cytometry, 3, 323–327. Waterman H, Alroy I, Strano S, Seger R and Yarden Y. (1999). EMBO J., 18, 3348–3358. Watson CS and Gametchu B. (1999). Proc. Soc. Exp. Biol. Med., 220, 9–19 (review). Weigel NL. (1996). Biochem. J., 319 (Part 3), 657–667 (review). Wellstein A, Fang WJ, Khatri A, Lu Y, Swain SS, Dickson RB, Sasse J, Riegel AT and Lippman ME. (1992). J. Biol. Chem., 267, 2582–2587. Wellstein A, Lupu R, Zugmaier G, Flamm SL, Cheville AL, Delli Bovi P, Basilico C, Lippman ME and Kern FG. (1990). Cell Growth Differ., 1, 63–71. Wen Y, Hu MC, Makino K, Spohn B, Bartholomeusz G, Yan DH and Hung MC. (2000). Cancer Res., 60, 6841–6845. Wosikowski K, Schuurhuis D, Kops GJ, Saceda M and Bates SE. (1997). Clin. Cancer Res., 3 (12 Partt 1), 2405–2414. Wright C, Prasad K and Lennard TJ. (1992). J. Clin. Pathol., 45, 459–460. Zeillinger R, Kury F, Czerwenka K, Kubista E, Sliutz G, Knogler W, Huber J, Zielinski C, Reiner G, Jakesz R, Staffen A, Reiner A, Wrba F and Spona Y. (1989). Oncogene, 4, 109–114. Zhang K, Sun J, Liu N, Wen D, Chang D, Thomason A and Yoshinaga SK. (1996). J. Biol. Chem., 271, 3884–3890.

Oncogene