Estrogen Receptor a Rapidly Activates the IGF-1 Receptor Pathway*

Estrogen and insulin-like-growth factor 1 (IGF-1) are potent mitogenic stimuli that share important properties in the control of cellular proliferation. However, the coupling between the signaling cascades of estrogen receptors a and b and the IGF-1 receptor (IGF-1R) is poorly understood. Therefore, we selectively transfected estrogen receptor a or b in COS7 and HEK293 cells, which contain IGF-1R. In presence of estrogen receptor a but not b , 17 b -estradiol (E2) rapidly induces phosphorylation of the IGF-1R and the extracellular sig-nal-regulated kinases 1/2. Furthermore, upon stimulation with E2, estrogen receptor a but not b bound rapidly to the IGF-1R in COS7 as well as L6 cells, which express all investigated receptors endogenously. Control experiments in the IGF-1R-deficient fibroblast cell line R 2 showed that after stimulation with E2 only estrogen receptor a bound to the transfected IGF-1R. Overexpression of dominant negative mitogen-activated protein kinases kinase inhibited this effect. Finally, estrogen receptor a but not b is required to induce the activation of the estrogen receptor-responsive reporter ERE-LUC in IGF-1-stimulated cells. Taken together, these data demonstrate that ligand bound estrogen receptor a is required for rapid activation of the IGF-1R Immunoprecipitation— with expression plasmids for either ER a or ER b and stimulated with 10 2 9 M E2 for 0–60 min. Phosphorylation of IGF-1R was detected by immunoprecipitation of the IGF-1R b -subunit and subsequent Western blot analysis with an antibody against phosphotyrosine residues. As a control, cells were cotreated with antagonists of the IGF-1R ( H , H 1356, 10 2 7 M ) and estrogen receptors ( ICI , ICI 182,780, 10 2 8 M ). These results are representative of at least three experiments. caused by nonspecific cross-reactions of the antibodies. The results from both experimental approaches were in correspondence (Fig. 2, lower panels ). The observed coimmunoprecipitation was sensitive to coincubation with the estrogen receptor antagonist ICI 182,780, the IGF-1R antagonists H1356, and the MAPKK inhibitor PD 98059 (40). Phosphorylation of ERK1/2 Is Required for Binding of ER a to the IGF-1R— To assess the physiological relevance of E2-induced IGF-1R phosphorylation, we analyzed signaling events downstream of the IGF- 1R. Phosphorylation of ERK1/2 is part of the signaling cascade downstream of the IGF-1R. Phosphorylation of ERK1/2 by E2 was detected by immunoblotting with an phospho-specific antibody against tyrosine/ threonine phosphorylated ERK1/2. COS7 cells were transfected with expression plasmids for either ER a or ER b . ERK1/2 phosphorylation was determined after 0–60 min stimulation with 10 2 9 M E2. In the presence of ER a E2 rapidly induced ERK1/2 phosphorylation (Fig. 3), which reached its maximum after 15 min. In cells expressing only ER b E2 failed to induce this reaction. ERK1/2 phosphorylation in the presence of ER a was blocked by cotreatment with the IGF-1R antagonist H 1356. Control experiments using E2 and PD 98059 revealed that MAPKK is involved in the process. Identical results were obtained when the same experimental procedure was performed in HEK293 cells (data not shown). 2 or constitutively active MAPKK ( MAPKK , Cells were incubated with 10 9 vehicle alone for Protein were subjected to immunoprecipitation with antibody with ysis with an These representative of at least three experiments. activates the IGF-1R signaling cascade. The Ras/Raf cascade is a subsequent target of IGF-1R signaling and activates MAPK that in turn stimulates the activity of ER a , thus establishing a positive feedback loop between estrogen receptor and receptor tyrosine kinase signaling.

Estrogen as well as insulin-like growth factor 1 (IGF-1) 1 are potent mitogens that are involved in a large array of processes that control proliferation and differentiation in mammalian cells (1,2). Both mitogens act through receptor-mediated signaling pathways. The cross-talk between these two signaling pathways is currently under investigation (3)(4)(5)(6). Estrogen is a steroid hormone that binds to members of the nuclear receptor superfamily (7), whereas IGF-1 as a peptide-growth factor binds to a transmembrane tyrosine kinase receptor, which signals via a series of phosphorylation events (2).
Two different estrogen receptors, ER␣ and ER␤, which are encoded by genes located on different chromosomes, have been identified so far (8,9). Sequence analysis demonstrates a high degree of homology between ER␣ and ER␤ in the DNA-binding domain and the ligand-binding domain. However, there are significant differences in regions that would be expected to influence transcriptional activity. The ability of estrogen receptors to activate target gene transcription has been attributed to two regions: the N-terminal activation function 1 (AF-1) and the ligand-dependent AF-2, which is localized in the C-terminal hormone-binding domain (10,11). AF-1 and AF-2 can activate transcription independently and synergistically, and they act in a promoter-and cell-specific manner (12,13). Phosphorylation of a serine residue at position 118 is required for full action of the AF-1 (14). Both AF-1 and AF-2 are required to enhance transcription of target genes through AP-1 sites (15). Interestingly, ER␣ and ER␤ act differently at AP-1 sites (16), which may be due to differences in their AF domains (17). ER␣ and ER␤ can form homo-and heterodimers (18), and thus the tissue-specific distribution of the estrogen receptor subtypes plays an important role in the different responses to 17␤estradiol. ER␣ appears to play a major role in the regulation of reproductive events. ER␣ knockout female mice are completely infertile, whereas ER␤ knockout mice females display severe but incomplete infertility (19,20). Under pathological conditions such as endothelial denudation of the rat carotid artery ER␣ mRNA is constitutively expressed, whereas the expression of ER␤ mRNA is increased (21). The cross-talk between IGF-1R and estrogen receptor-signaling pathways results in synergistic growth stimulation (22)(23)(24). The mitogenic response of IGF-1 appears to be initiated by ligand-induced tyrosine autophosphorylation of the IGF-1R, which is important for the selective recruitment of downstream signaling molecules and results in activation of the Ras/Raf/MAPK signaling cascade. Because activation of the estrogen receptors by growth factors like IGF-1 involves the MAPK pathway by direct phosphorylation of AF-1 (5,25), this is one of the potential interaction sites of both signaling pathways. On the other hand, the modulation of IGF-1R and membrane-associated IGF-binding proteins as a response to stimulation by estrogen have been described. Estrogen induces tyrosine phosphorylation of IGF-1 and insulin receptor substrate-1, which is followed by enhanced MAPK activation (4, 26 -28).
The aim of this study was to gain further insight into the interactions of the IGF-1 and the estrogen signaling pathways. For that, the mechanisms involved in the cross-talk between the estrogen receptor and the IGF-1R were investigated. In addition, we addressed the question of whether ER␣ and ER␤ differ in their potential to interact with the IGF-1R signaling pathway.
were obtained from the American Type Culture Collection (Manassas, VA). COS7 cells were maintained in Dulbecco's modified Eagle's medium (Life Technologies, Inc.) supplemented with 10% charcoalstripped, estrogen-free fetal calf serum (c.cpro, Hamburg, Germany). HEK293 were maintained in modified Eagle's medium (Life Technologies, Inc.) supplemented with 1% nonessential amino acids and 10% charcoal-stripped fetal calf serum. R Ϫ is a fibroblast cell line derived from a transgenic mouse that does not express the IGF-1R (29) but contains both ER␣ and ER␤ as we confirmed by Western blotting. R Ϫ cells were a kind gift from Dr. R. Baserga and were maintained in Dulbecco's modified Eagle's medium supplemented with 10% charcoalstripped, estrogen-free fetal calf serum. The myogenic cell line L6 (30), which expresses ER␣, ER␤, and the IGF-1R without transfection, was obtained from the American Type Culture Collection. L6 skeletal myoblasts were cultured as described before (31). All media contained 25 g/ml Gentamycin (Life Technologies, Inc.). Phenol red-free medium was used throughout all experiments because phenol red acts as a weak estrogen (32).
Immunoprecipitation-Cellular lysates (400 g of protein) were incubated with the respective antibody (described above) against ER␣, ER␤, and the IGF-1R. Cellular extracts were incubated with the primary antibody for 24 h at 4°C. Depending on the antibody used, 40 l of protein A or protein G-Sepharose were added to the lysate followed by incubation for 3 h at 4°C. Samples were centrifuged in a cooled microcentrifuge (4°C) for 2 min at 12,000 ϫ g. Pellets were washed twice with RIPA buffer (50 mM NaCl, 20 mM Tris, pH 7.4, 50 mM NaF, 50 mM EDTA, 20 mM Na 4 P 2 O 7 , 1 mM Na 3 VO 4 , 1% Triton X-100, 1 mM phenylmethylsulfonyl fluoride, 0.6 mg/ml leupeptin, 10 g/ml aprotinin) and resuspended in 50 l of total volume. SDS loading buffer was added, and samples were heated for 5 min at 95°C. Sepharose beads were pelleted by centrifugation in a microcentrifuge at 12,000 ϫ g for 5 min at 4°C. The supernatants were then analyzed by SDS-polyacrylamide gel electrophoresis and subsequent immunoblotting as described above.
Transfections-Cells were grown to an approximate confluence of 70% and transfected using a liposome-conjugated transfection technique according to the manufacturer's instructions (DOTAP; Roche Molecular Biochemicals). Both COS7 and HEK293 cells were transfected with expression plasmids for human ER␣. The data presented in this study were obtained with the expression plasmid HE0 (34). This plasmid contains the human wild-type ER␣ cDNA-clone OR8 as described in Ref. 8, which is known to contain a point mutation in its hormone-binding domain (Gly 400 3 Val 400 ), which alters its ligand binding properties (16,35). Human ER␣ expressed by HE0 has a slightly reduced transactivation response compared with the human wild-type ER␣. All experimental data shown are based on experiments performed with HE0 and were reproduced at least once with the expression plasmid for the human wild-type ER␣ (HEG0). The results of transfection and immunoprecipitation studies were found to be the same with both constructs with the only difference that transactivation responses of HEG0 were slightly higher than the data obtained with HE0 (HE0 and HEG0 were kind gifts from Dr. Pierre Chambon). In experiments that included ER␤, pCMV29 (9) (kind gift from Dr. G. Kuiper) was transfected that expresses the rat ER␤ cDNA. To confirm the dependence of the observed effect on the presence of IGF-1R protein, R Ϫ cells were transfected with a plasmid containing the IGF-1R precursor cDNA under the control of the SV40 early promoter (29) (kind gift from Dr. R. Baserga). Plasmids expressing dominant negative (K97S-MAPKK) and constitutively active mutants (S222A-MAPKK) of Xenopus MAPKK (36) were utilized to determine the influence of the activity of MAPKK on interactions between estrogen receptor and IGF-1 signaling pathways.
Transfection Assay-Cells were transfected with ERE-LUC (containing three copies of an estrogen-responsive element from the Xenopus vitellogenin promoter, driving expression of the luciferase gene, kindly provided by Dr. C. Glass) and the respective estrogen receptor expression vectors. Cells were harvested 24 h after transfection, and luciferase activity was determined in a luminometer (Optocom 1, MGM Instruments, Hamden, CT) as described in Ref. 37. Cells were cotransfected with pL7RH-Gal (SV40 promoter including a nuclear localization signal driving the ␤-galactosidase cDNA). Transfection efficiency was determined by staining of a subset of the transfected cells. Cells were washed with phosphate-buffered saline and then fixed in 0.5% glutaraldehyde for 10 min followed by three more washes. Then they were incubated overnight at 37°C in a staining solution containing 15 mM K 3 Fe(CN) 6 , 15 mM K 2 Fe(CN)6 3 H 2 O, 0.15 mM MgCl 2 , 1% Me 2 SO, and 1 mg/ml 5-bromo-4-chloro-3-indolyl ␤-D-galactopyranoside (X-gal). Nuclei of cells stained for ␤-galactosidase activity were counted, and the results of each luciferase determination were normalized for the transfection efficiency.
Inhibitory Substances-The IGF-1R antagonist H 1356 is an IGF-1 peptide analogue with the amino acid sequence CYAAPLKPAKSC (38) (Bachem, Heidelberg, Germany) and was added in a subset of transfection studies to investigate the effect of IGF-1R inhibition on the experimental results. The estrogen receptor antagonist ICI 182,780 (39) (kind gift from Dr. A. Wakeling), which completely antagonizes ER␣ as well as ER␤, was used to study inhibition of the estrogen receptors. An equal volume of vehicle alone (0.1% ethanol) was added to control cells. The MAPKK inhibitor PD 98059 (40) (Calbiochem, Bad Soden, Germany) was utilized to investigate the influence of MAPKK inhibition.
Statistical Analysis-All reported values are the means Ϯ S.D. Statistical comparisons were made by Student's t test, two-sided with adjustment for multiple comparisons. Statistical significance was assumed if a null hypothesis could be rejected at the p Ͻ 0.01 level. All materials were obtained from Merck if not otherwise specified RESULTS IGF-1R Autophosphorylation by E2 Requires ER␣ but Not ER␤-In a first step we elucidate the role of ER␣ and ER␤ in E2-mediated activation of IGF-1R signaling. We investigated whether the phosphorylation of IGF-1R by E2 (4) depends on the presence of estrogen receptors. To distinguish between the role of ER␣ and ER␤ in the E2-dependent activation of IGF-1R phosphorylation we used the cell line COS7, which expresses the IGF-1R but is devoid of estrogen receptors. These cells were transfected with expression plasmids for either ER␣ or ER␤ and were subsequently stimulated with 10 Ϫ9 M E2 for 0 -60 min. In a series of immunoprecipitation experiments with an IGF-1R antibody and subsequent immunoblotting against phosphotyrosine residues, phosphorylation of IGF-1R was detected (Fig. 1). In the presence of ER␣, stimulation with E2 led to a time-dependent induction of IGF-1R phosphorylation in COS7. In contrast, E2 did not stimulate receptor phosphorylation in the presence of ER␤. No induction of tyrosine phosphorylation after stimulation with E2 was observed in cells that had not been transfected. The ER␣-dependent induction of tyrosine phosphorylation after E2 treatment was inhibited by cotreatment with the IGF-1 peptide analogue H 1356, an antagonist of the IGF-1R (38), and by ICI 182,780, an antagonist of both estrogen receptor subtypes (39). Identical results were obtained when the same experimental procedure was performed in the cell line HEK293 (data not shown).
E2 Stimulates Binding of ER␣ but Not ER␤ to the IGF-1R-Because the E2-dependent induction of IGF-1R phosphorylation was sensitive to inhibition of both the IGF-1R and the estrogen receptor, we performed a series of coimmunoprecipitation experiments of these receptors to elucidate the underlying mechanisms. Three groups of cells were employed in these experiments: 1) COS7 cells served as models for controlled expression of either ER␣ or ER␤. 2) As a model for controlled IGF-1R expression we used R Ϫ cells, an embryonic fibroblast cell line derived from a transgenic animal that does not express IGF-1R but does express ER␣ and ER␤ (29). One subset of these cells was transfected with an expression plasmid for the IGF-1R.
3) The myogenic cell line L6 (30) was utilized as an example for a cell line expressing all receptors endogenously (31,41,42). Stimulations were performed with 10 Ϫ9 M E2 for 0 -60 min. Total cellular protein was harvested and subjected to immunoprecipitation with antibodies against ER␣ or ER␤. The precipitated fraction was used for subsequent immunoblotting with an IGF-1R antibody as shown in Fig. 2. Coimmunoprecipitation of IGF-1R with the ER␣ was detected 5 min after stimulation with 10 Ϫ9 M E2 in COS7 ( Fig. 2A) cells transfected with ER␣ and in R Ϫ cells (Fig. 2B) transfected with the IGF-1R. In L6 cells (Fig. 2C) this signal was detectable after 30 min. L6 cells express all investigated receptors under the control of endogenous promoters. We therefore assume that the onset of this interaction may take considerably longer. No interaction between ER␤ and the IGF-1R was observed in any experiments (Fig. 2, upper panels). In R Ϫ cells, in the absence of the IGF-1R, no interaction of the IGF-1R with ER␣ was detected (Fig. 2B, upper panel). The sequence of antibodies used for precipitation and immunoblotting was reversed to exclude the possibility of false positive results caused by nonspecific cross-reactions of the antibodies. The results from both experimental approaches were in correspondence (Fig. 2, lower panels). The observed coimmunoprecipitation was sensitive to coincubation with the estrogen receptor antagonist ICI 182,780, the IGF-1R antagonists H1356, and the MAPKK inhibitor PD 98059 (40).
Phosphorylation of ERK1/2 Is Required for Binding of ER␣ to the IGF-1R-To assess the physiological relevance of E2-induced IGF-1R phosphorylation, we analyzed signaling events downstream of the IGF-1R. Phosphorylation of ERK1/2 is part of the signaling cascade downstream of the IGF-1R. Phosphorylation of ERK1/2 by E2 was detected by immunoblotting with an phospho-specific antibody against tyrosine/ threonine phosphorylated ERK1/2. COS7 cells were transfected with expression plasmids for either ER␣ or ER␤. ERK1/2 phosphorylation was determined after 0 -60 min stimulation with 10 Ϫ9 M E2. In the presence of ER␣ E2 rapidly induced ERK1/2 phosphorylation (Fig. 3), which reached its maximum after 15 min. In cells expressing only ER␤ E2 failed to induce this reaction. ERK1/2 phosphorylation in the presence of ER␣ was blocked by cotreatment with the IGF-1R antagonist H 1356. Control experiments using E2 and PD 98059 revealed that MAPKK is involved in the process. Identical results were obtained when the same experimental procedure was performed in HEK293 cells (data not shown).
The estrogen receptor signaling pathway has been shown to be activated by binding of E2 to its receptor as well as by phosphorylation of ER␣ by MAPK (5). Because our results point toward an activation of the IGF-1R and and its downstream signaling cascade by an interaction of the IGF-1R with the estrogen receptor-ligand complex, it was important to analyze whether ER␣ phosphorylation by MAPK also influences this receptor interaction. We therefore transfected HEK293 cells with plasmids expressing ER␣ together with plasmids expressing either dominant negative MAPKK or constitutively active MAPKK (29). Coimmunoprecipitation experiments with the IGF-1R and either ER␣ or ER␤ were performed. As shown in Fig. 4, expression of dominant negative MAPKK (Fig. 4A) prevents an interaction between ER␣ and the IGF-1R, whereas in the presence of constitutively active MAPKK (Fig. 4B) binding of ER␣ to the IGF-1R was observed even in the absence of ligand.
Because the expression of constitutively active MAPKK induced an interaction of ER␣ with the IGF-1R even in the absence of ligand, we investigated whether ER␤ may also interact with the IGF-1R under the conditions of constitutively active MAPKK in HEK293 cells. In the absence and presence of constitutively active MAPKK, ER␤ did not interact with the IGF-1R regardless of stimulation with E2 (Fig. 4C). However, ER␤ was detectable in samples from protein lysates that were not subjected to immunoprecipitation.
IGF-1 Leads to Ligand-independent Activation of ER␣ but Not ER␤-In our previous experiments we observed an activation of the IGF-1 signaling pathway by E2 via ER␣. In contrast, ER␤ was not involved in the interaction with the IGF-1 signaling pathway. As described previously, transcriptional activation of estrogen-responsive genes is induced by growth factors like IGF-1 in vivo (36,43). Further-more, transactivation of estrogen-responsive reporter constructs has been demonstrated to be inducible by growth factors like IGF-1 (6) or epidermal growth factor even in the absence of ligand (25). As a consequence of our previous observations, we investigated whether ER␣ and ER␤ differ in their potential to activate an estrogen-responsive gene after stimulation with IGF-1. To test this hypothesis a reporter construct containing an estrogen-responsive enhancer element driving a luciferase cDNA (ERE-LUC) (44) was employed.
In a first step we investigated whether the expression of an estrogeninducible reporter after incubation with E2 or IGF-1 depends on the integrity of the IGF-1 signaling pathway. R Ϫ cells were transfected with an expression plasmid for the IGF-1R and cotransfected with the reporter plasmid ERE-LUC (Fig. 5, R Ϫ ). Stimulation with E2 (10 Ϫ9 M) resulted in a 6.3 Ϯ 0.06-fold induction of reporter expression in the presence of IGF-1 compared with a 5.6 Ϯ 0.03-fold induction in the absence of the IGF-1R. Furthermore, we observed that in R Ϫ cells transfected with the IGF-1R plasmid ERE-LUC expression was induced 5.0 Ϯ 0.06-fold after incubation with IGF-1 (10 Ϫ8 M), whereas in cells lacking the IGF-1R IGF-1 did not influence reporter expression. Expression of the same reporter plasmid lacking estrogen-responsive elements (TK-LUC) was neither influenced by incubation with E2 nor by incubation with IGF-1.

DISCUSSION
The interaction of ER␣ and ER␤ with membrane-associated signaling pathways is currently under investigation but is still poorly understood (3,4,6,26,27). In our understanding of the role and function of estrogen receptors on the cell membrane level two major points deserve further attention: the interaction between estrogen receptors and tyrosine kinase receptor signaling pathways as well as the differences of ER␣ and ER␤ in coupling to the cell membrane.
The data presented here contribute new observations to these points by demonstrating that: 1) E2 activates the IGF-1R signaling cascade exclusively via ER␣, 2) binding of ER␣ to the IGF-1R is a potential mechanism for estrogen to activate the IGF-1R, 3) phosphorylation of ERK1/2 is required for binding of ER␣ to the IGF-1R, and 4) ER␣ but not ER␤ is the target of ligand-independent activation by IGF-1 signal transduction via the MAPK. Rapid, so called nongenomic effects of estrogen, which occur within seconds to minutes, are part of a variety of cellular responses to estrogen. In particular, rapid effects of estrogen on NO release (45,46), on regulatory processes involved in NO formation (47), on calcium homeostasis (48 -50), and on cAMP accumulation (51) have been described up to now, but the exact mechanisms of these processes are incompletely understood. Therefore, the observation that 1) ligand-induced ER␣ binds to a transmembrane receptor and that 2) E2 rapidly activates phosphorylation of the IGF-1R in the presence of ER␣ may offer a potential mechanism to gain further insight into the previous findings. Our data provide a model for how ligandbound ER␣ activates growth factor signaling cascades in a rapid nongenomic fashion: by binding to the IGF-1R followed by its phosphorylation. Recently, it was reported that estrogen stimulates the phosphorylation of the membrane bound IGF-1R in the uterus after 6 h and involves the phosphatidylinositol 3-kinase signaling pathway (4,26). Our findings extend these observations by demonstrating a rapid effect of estrogen on IGF-1R phosphorylation within minutes. The presence of a putative membrane-bound ER (51)(52)(53)(54)(55) has to be discussed in the light of these data.
Previously, it was demonstrated that the estrogen-independent phosphorylation and activation of estrogen receptors can be mediated via the MAPK pathway (5,25). MAPKs such as ERKs are important regulators in signaling pathways in response to a wide array of extracellular stimuli. This observa- tion raises the possibility that estrogen receptors can be activated by a number of growth factors using the MAPK pathway. Interestingly, activation of MAPK pathway is also induced by estrogen itself (33,56,57) and requires mobilization of intracellular calcium (58). Thus it appears that a feed-forward system exists where E2 activates MAPK, an event that in turn enhances transcriptional activity of the estrogen receptor. In the present study we show that E2-induced MAPK activation is mediated by ER␣ and does not involve ER␤. We were able to demonstrate that binding of ER␣ to the IGF-1R depends on MAPKK activity and is induced via the MAPK pathway. As a consequence, the binding of ER␣ to the transmembrane IGF-1R is regulated by the MAPK pathway under these conditions. In our study we found important differences between the two estrogen receptor subtypes in the ability to interact with the IGF-1 pathway. ER␤ was not able to bind to the IGF-1R, and in consequence no phosphorylation and activation of IGF-1R was observed. ER␣ and ER␤ display significant differences in the transcriptional activation of target genes upon stimulation with IGF-1. Treatment with IGF-1 only stimulated transcriptional activity of ER␣ but not of ER␤ in different cell lines such as COS7, HEK293, L6, or R Ϫ . These results suggest different regulatory functions of the two estrogen receptor subtypes. The differences may be in part due to different AF-1 domains because ER␣ reveals constitutive AF-1 activity, whereas ER␤ lacks constitutive AF-1 activity (15,59). These data demonstrate that the tissue-specific distribution pattern of ER␣ and ER␤ may be crucial to understand the proliferative response to IGF-1 and E2 in a variety of hormone-dependent malignant diseases.
In summary, we were able to demonstrate that after ligand binding and phosphorylation by the MAPK cascade ER␣ is able to bind to the IGF-1R. This binding induces autophosphorylation of the IGF-1R and thus its activation. IGF-1R activation stimulates MAPKK and consequently phosphorylation of ERK1/2. Activation of ERK1/2 may, in turn, lead to phosphorylation of ER␣ and provides a possible mechanism for ligandindependent activation of ER␣ but not of ER␤ (Fig. 6). Further investigations, however, are required to understand how estrogen and the expression patterns of the two estrogen receptor subtypes as well as IGF-1R signal transduction cooperate in the regulation of growth and differentiation. FIG. 6. Suggested model of a mechanism for the interaction between estrogen receptor and IGF-1R signaling. ER␣ binds to the IGF-1R and activates the IGF-1R signaling cascade. The Ras/Raf cascade is a subsequent target of IGF-1R signaling and activates MAPK that in turn stimulates the activity of ER␣, thus establishing a positive feedback loop between estrogen receptor and receptor tyrosine kinase signaling.