Abstract
Cross-talk between insulin-like growth factor (IGF)- and estrogen receptor (ER)-signaling pathways results in synergistic growth. We show here that estrogen enhances IGF signaling by inducing expression of three key IGF- regulatory molecules, the type 1 IGF receptor (IGFR1) and its downstream signaling molecules, insulin receptor substrate (IRS)-1 and IRS-2. Estrogen induction of IGFR1 and IRS expression resulted in enhanced tyrosine phosphorylation of IRS-1 after IGF-I stimulation, followed by enhanced mitogen-activated protein kinase activation. To examine whether these pathways were similarly activated in vivo, we examined MCF-7 cells grown as xenografts in athymic mice. IRS-1 was expressed at high levels in estrogen-dependent growth of MCF-7 xenografts, but withdrawal of estrogen, which decreased tumor growth, resulted in a dramatic decrease in IRS-1 expression. Finally, we have shown that high IRS-1 expression is an indicator of early disease recurrence in ER-positive human primary breast tumors. Taken together, these data not only reinforce the concept of cross-talk between IGF- and ER-sig-naling pathways, but indicate that IGF molecules may be critical regulators of estrogen-mediated growth and breast cancer pathogenesis.
INTRODUCTION
One of the most important predictive and prognostic markers in human breast cancer is the estrogen receptor (ER). Long before the receptor was cloned, or its function understood, antihormonal therapies were used to treat breast cancer, and long-term studies have since shown the importance of antiestrogens such as tamoxifen (Tam) in reducing both disease progression and contralateral breast cancer (1). Due to the proliferative action of estrogen in vitro, much research has examined estrogen’s ability to control proliferation via regulation of autocrine and paracrine growth factors (2). From these studies data have emerged supporting a role for insulin-like growth factors (IGFs) in estrogen action from in vitro, in vivo, and clinical studies (3).
The IGF family plays an important role in the growth of both normal and neoplastic cells (4). Gene-knockout studies have revealed the complexity of IGF signaling and have shown that IGFs can have mitogenic, transforming (probably via a permissive action), and antiapoptotic functions (5). Examination of IGF expression in primary human breast tumors has shown that at all stages of IGF action, expression of the components of the effector pathways are correlated with ER expression and have prognostic significance. For instance, IGF-binding protein (IGFBP) expression in primary human breast tumors is correlated with ER status, and high IGFBP-3 and IGFBP-4 levels are associated with poor prognostic markers (6, 7). The type 1 IGF receptor (IGFR1) is expressed in a high percentage of primary breast tumors, and expression is positively correlated with ER status (8). One of the downstream signaling molecules of IGFR1, insulin receptor substrate 1 (IRS-1), is expressed in human breast cancer, and a high level of expression is an indicator of early disease recurrence in small tumors (7).
Animal studies support a role for IGFs in the pathogenesis of breast cancer. MDA-231 xenograft tumor growth can be inhibited by blockade of IGFR1 with the monoclonal antibody αIR3 (9). We followed this observation by showing that neutralization of IGF action with IGFBP-1 inhibited growth of MDA-231 and also ascites growth of MDA-435A (10). More recent evidence that endocrine IGFs are important in breast tumor growth is shown by the reduced growth of MCF-7 xenografts in mice that lack circulating IGF-I (11). Taken together, these data show that circulating IGFs are important in these xenograft models of breast cancer. Clinically, this is important, considering the ability of Tam to consistently reduce serum IGF-I levels in breast cancer patients (12).
Strengthening the clinical and animal studies that support a role for IGFs in breast cancer pathogenesis are a number of in vitro studies that not only confirm the potent mitogenic effects of IGFs on breast cancer cells, but also highlight considerable synergism between IGF and ER signaling. Estrogen can affect IGF action and growth by altering expression of IGFR1 (13, 13A ), IGFR2 (14), IGF-II (15), and IGFBPs (16, 17). It is not surprising, therefore, that overexpression of certain IGF family members (IGF-II and IRS-1) results in enhanced growth and reduced estrogen requirements (18–22). Conversely, decreased expression of IRS-1 inhibits breast cancer cell proliferation and causes cell death in serum-free conditions (23).
In an attempt to understand cross-talk and synergism between IGF and ER signaling, we examined regulation of IGF family members by estrogen and antiestrogens. We present data here confirming that IGFR1 is an estrogen-regulated protein, but we also show that expression of the major downstream targets of IGFR1, IRS-1, and IRS-2 are regulated by estrogen. Increased expression of IGFR1, IRS-1, and IRS-2 by estrogen results in enhanced IRS phosphorylation, which leads to greater mitogen-activated protein kinase (MAPK) activity. Furthermore, we have shown that IRS-1 is regulated by estrogen in a xenograft model of human breast cancer, and that in ER-positive human breast tumors high IRS-1 expression is associated with poor disease-free survival (DFS). Taken together, these data support a role for IGFs in ER-mediated growth, with components of IGF signaling pathways being key targets for ER action and thus, in part, responsible for estrogen’s growth promoting effects.
RESULTS
Enhancement of MCF-7 Cell Growth by Coincubation with Estrogen and IGF-I
MCF-7 cells were incubated with estrogen (1 nm), IGF-I (5 nm), or a combination for 5 days, and cell number was counted. Estrogen and IGF-I treatment resulted in a 2.2-fold increase in cell number compared with cells incubated in serum-free medium (SFM) (Fig. 1). Coincubation with both mitogens resulted in greater proliferation and a 5-fold increase in cell number. This result was seen consistently, with the effect of estrogen and IGF-I being greater than estrogen or IGF-I alone.
Enhancement of MCF-7 Cell Growth by Coincubation with Estrogen and IGF-I MCF-7 cells were incubated in SFM, E2 (1 nm), IGF-I (5 nm), or a combination for 5 days, and then cell number was counted by hemocytometer. Bars represent the average of triplicate cells ± sem. This figure is representative of three independent experiments.
Estrogen Increases IRS-1 Expression in ER-Positive Breast Cancer Cell Lines
Three ER-positive human breast cancer cell lines (MCF-7, T47D, and ZR-75) and an ER-negative breast cancer cell line (MDA-435A) were incubated with or without estrogen for increasing lengths of time (2, 4, and 6 days), and extracts were immunoblotted for IRS-1 expression (Fig. 2). Estrogen increased IRS-1 expression in all three ER-positive cell lines, but had no effect in MDA-435A cells. A consistent decrease in IRS-1 expression was seen in all ER-positive cells grown in SFM, probably representing depletion of residual estrogen. Further experiments did indeed show that antiestrogens reduced this basal IRS-1 expression (data not shown and next figure).
Estrogen Increases IRS-1 Expression in ER-Positive Breast Cancer Cell Lines Three ER-positive (MCF-7, T47D, and ZR-75) and one ER-negative (MDA-435A) breast cancer cell line were stimulated with or without E2 (1 nm) for increasing periods of time (2, 4, and 6 days). Cells were lysed and immunoblotted for IRS-1. This result is representative of two independent experiments.
Expression of IRS-1, IRS-2, and IGFR1 Is Regulated by Estrogen
MCF-7 cells were incubated with estrogen and the pure steroidal antiestrogen ICI 182780 (ICI) for increasing lengths of time (8, 24, and 48 h) and analyzed for IGFR1, IRS-1, and IRS-2 expression. IGFR1 mRNA- and IGF-I-binding sites have previously been shown to be up-regulated after estrogen treatment of breast and endometrial cancer cells (13, 13A, 24). Confirming this observation, we saw that after 48 h, estrogen increased IGFR1 protein expression (Fig. 3), and that this stimulation was specifically competed by ICI. No difference in IGFR1 expression was seen at 8 and 24 h of estrogen treatment. Expression of IRS-1 and IRS-2 was also increased by estrogen and specifically inhibited by ICI. In some instances, IRS expression in the presence of antiestrogen was actually reduced below that of cells in SFM [compare IRS-1 expression at 48 h SFM vs. estradiol (E2)+ICI], again suggesting that there is residual estrogen present. However, in contrast to the relatively late induction of IGFR1, IRS-1 and IRS-2 levels were also increased at earlier time points (8 and 24 h). Up-regulation of IGFR1 and IRS expression was not simply due to the fact that cells were proliferating faster in the presence of estrogen, since growth stimulation by IGF-I treatment resulted in no change in IGFR1 expression and an actual decrease in IRS-1 expression (data not shown). Additionally we have shown that IRS1 expression is decreased during S-phase (data not shown), ruling out the possibility that estrogen induction of IRS-1 is simply due to the fact that cells are proliferating faster and that more cells are in S-phase.
Estrogen Increases Expression of IGFR1, IRS-1, and IRS-2 in MCF-7 Cells MCF-7 cells were stimulated with E2 (1 nm) and ICI (1 μm) alone or in combination for increasing periods of time (8, 24, and 48 h). Cells were lysed and immunoblotted for IGFR1, IRS-1, and IRS-2. Results shown are representative of four independent experiments.
Estrogen Increases Expression of IGFR1 and IRS-1 mRNA
MCF-7 cells were treated with estrogen and antiestrogen for 24 and 48 h, total RNA was isolated, and mRNA expression was examined by RNAse protection assay. Estrogen resulted in increased mRNA expression of both IRS-1 and IGFR1 at both time points (Fig. 4A). Additionally, both antiestrogens, Tam and ICI, effectively reversed the induction. The expression is represented graphically after correcting for the loading control (Fig. 4B). The lower induction of IRS-1 expression at 48 h represents experimental variation and was not seen in all experiments. As a control, we analyzed IGFR1 and IRS-1 mRNA expression after exposure of cells to IGF-I. In this instance there was no change in either mRNA (data not shown), indicating again that estrogen-induction is not simply a result of estrogen-stimulated proliferation.
Estrogen Increases IGFR1 and IRS-1 mRNA Expression in MCF-7 Cells A, MCF-7 cells were stimulated with E2 (1 nm), Tam (1 μm), or ICI (1 μm), alone or in combination for 24 or 48 h. RNA was prepared and 20 μg analyzed by RNAse protection assay. 36B4 was used as a loading control. B, Graphical representation of data in panel A after densitometry and correction for 36B4 expression. This figure is representative of three independent RNAse protection assays. C, MCF-7 cells were pretreated with (Act D) or without (SFM) actinomycin D (2 μg/ml) for 30 min and then stimulated with E2 (1 nm) in the absence or presence of actinomycin D for 24 h. Cells were lysed and immunoblotted for IRS-1 and IGFR1.
Having shown that IRS-1 mRNA and protein expression was induced relatively quickly after estrogen stimulation, we examined whether this was a transcriptional effect. MCF-7 cells were stimulated with estrogen in the presence or absence of actinomycin D (2 μg/ml). The induction of IGFR1 and IRS-1 was completely blocked by actinomycin D (Fig. 4C), suggesting that estrogen induces IGFR1 and IRS-1 mRNA by a transcriptional mechanism. However, there was a slight discordance between the level of estrogen induction of IGFR1 protein (Fig. 3) and IGFR1 mRNA (Fig. 4A), suggesting that there may be nontranscriptional mechanisms of regulation.
Estrogen Increases IGF Signal Transduction Events
We sought to examine the effect of altering IRS-1 and IGFR1 levels on IGF signal transduction by first preincubating cells with estrogen, or with estrogen and ICI for 48 h, and then stimulating with IGF-I for 10 min. As expected, estrogen increased expression of IRS-1 and IGFR1, which was blocked by ICI (Fig. 5, panels 2 and 3). Treatment of cells with IGF-I for 10 min resulted in tyrosine phosphorylation (PY) of a single protein of approximately 180 kDa (panel 1). We have recently shown by immunoprecipitation that this tyrosine-phosphorylated protein is IRS-1, and that while MCF-7 and other ER-positive breast cancer cells express IRS-2, that IRS-2 is not activated in these cells (25). After exposure to estrogen, IGF-I stimulation resulted in enhanced phosphorylation of IRS-1 compared with cells stimulated with IGF-I but without estrogen pretreatment. The increased phosphorylation of IRS-1 generally mirrored the increase in IRS-1 protein seen by immunoblotting. Preincubation of cells with estrogen and ICI resulted in reduced IRS phosphorylation compared with cells in estrogen.
Estrogen Increases IGF Action in MCF-7 Cells MCF-7 cells stimulated with E2 (1 nm) or E2 and ICI (1 μm) for 48 h (48 h). At the end of the 48 h, cells were stimulated with IGF-I (5 nm) for 10 min (10 m). Cells were lysed and immunoblotted with antibodies to phosphotyrosine (PY, panel 1), IRS-1 (panel 2), IGFR1 (panel 3), phospho-specific MAPK (panel 4), or total MAPK protein (panel 5). The figure is representative of three independent experiments.
To examine whether estrogen enhancement of IRS expression and phosphorylation resulted in increased downstream signaling, we measured activation of MAPK using a phospho-specific MAPK antibody. We have previously determined in our laboratory that there is good concordance between the phosphospecific MAPK antibodies and actual measurement of MAPK activity (measured by the ability to phosphorylate myelin basic protein). Estrogen enhancement of IRS-1 phosphorylation resulted in an increase in IGF-I-induced phospho-MAPK (panel 4). Densitometric analysis revealed that IGF induction of MAPK activity in the presence of estrogen was 5-fold higher than IGF activation of MAPK in cells grown in SFM. Again, this increase in MAPK activity was inhibited by ICI. A similar pattern of estrogen induction of IRS phosphorylation and MAPK activity was seen at shorter times of estrogen treatment (8 and 24 h; data not shown) although the absolute level of induction was smaller than that seen at 48 h. Total levels of MAPK protein were not affected under any of the conditions tested (panel 5).
IRS-1 Expression and Phosphorylation Are Estrogen Regulated in an in Vivo Model of Human Breast Cancer
As we have previously shown that IRS-1, and not IRS-2, is the major downstream signaling molecule in MCF-7 cells (25), we further investigated regulation of IRS-1 expression in vivo in a xenograft model of human breast cancer growth. We examined tumors grown in the presence or absence of estrogen for both total IRS-1 protein expression and tyrosine phosphorylation of IRS-1. Similar to cells grown in vitro, tumors growing in the presence of estrogen (+E2) had high levels of IRS-1 expression (Fig. 6A, second panel). When the estrogen pellet was removed (−E2), there was a dramatic reduction in IRS-1 expression. MCF-7 cells grown in vitro were used as a positive control. In two independent experiments eight of eight tumors grown in the presence of estrogen expressed high levels of IRS-1. When estrogen was removed, there was a greater than 95% reduction of IRS-1 expression in 22 of 23 tumors. We then performed antiphosphotyrosine analysis on tumors growing in the presence or absence of estrogen, which revealed an immunoreactive species of the same mol wt as IRS-1 (Fig. 6A, top panel). Immunoprecipitation with IRS-1 antibodies followed by antiphosphotyrosine immunoblotting revealed a band that comigrated with IRS-1 immunoprecipitated from MCF-7 cells grown in vitro, indicating that this was indeed tyrosine-phosphorylated IRS-1 (Fig. 6B). Immunoprecipitation from two individual xenografts grown in the presence of estrogen showed tyrosine-phosphorylated IRS-1 whereas we detected absolutely no tyrosine-phosphorylated IRS-1 in xenografts grown in the absence of estrogen.
IRS-1 Expression and Activation Are Regulated by Estrogen in MCF-7 Xenograft Tumors MCF-7 cells were grown as flank tumors in nude mice. A, Four individual MCF-7 xenografts were harvested either when growing in the presence of estrogen (+E2), or when estrogen was removed (−E2). Tumors were crushed under liquid nitrogen, lysed, and immunoblotted for phosphotyrosine (PY), IRS-1, active MAPK, and total MAPK. MCF-7 cells grown in vitro stimulated with IGF-I for 10 min were used as a positive control (MCF7L). This figure shows representative samples of eight tumors +E2 and 23 tumors −E2. B, Individual xenograft tumors from Fig. 6A that were grown in the presence (+E2, tumors 1 and 2) or absence of estrogen (−E2, tumors 5 and 6) were immunoprecipitated with antibodies to IRS-1 and immunoblotted with an antibody to PY.
In vitro data from Fig. 5, panel 4, and the presence of tyrosine- phosphorylated IRS-1 in estrogen-treated tumors suggested that IRS-1 was in an active signaling cascade; therefore, we analyzed activation of downstream MAPK in the same tumors. The tumors that were growing in the presence of estrogen (+E2) and had high levels of tyrosine-phosphorylated IRS-1 (second panel) also had high levels of phosphorylated MAPK (Fig. 6A, third panel). In absolute contrast, removal of estrogen (−E2), which resulted in loss of IRS-1 tyrosine phosphorylation and expression, resulted in complete loss of MAPK activity. Immunoblotting for total MAPK levels indicated that there was no change in total MAPK expression among the tumors and served as loading control. While there are many upstream activators of MAPK, in all samples we saw an absolute correlation between IRS-1 expression, tyrosine phosphorylation, and MAPK activity (data not shown).
IRS-1 Expression Affects Breast Cancer Recurrence in ER-Positive Breast Cancer Patients
Since our in vitro and xenograft studies show that ER and IGF may act synergistically to enhance growth, we reasoned that tumors expressing both ER and IRS-1 would have a growth advantage reflected by early recurrence after surgery. To test this hypothesis, we examined the influence of IRS-1 expression on DFS in ER-positive breast cancer patients. We have previously reported that IRS-1 expression positively correlates with ER levels (7). Indeed, after cutpoint analysis, breast tumor samples were separated into high or low IRS-1 expression by median IRS-1 levels, and nearly all of the ER-negative tumors had low IRS-1 expression. Furthermore, in the subset of ER-positive patients, patients with high IRS-1 expression had significantly shortened DFS (P = 0.035) than patients with low IRS-1 levels (Fig. 7). In the ER-negative subset of patients, IRS-1 levels had no effect upon DFS; however, the low number of samples did not provide enough data to reach statistical significance. The fact that ER-positive tumors with high IRS-1 expression have poor DFS suggests that the relationship between ER and IRS-1 that we have observed in vitro and in xenograft models is also relevant in human primary breast cancers.
High IRS-1 Expression Is Associated with Reduced DFS in ER-Positive Breast Cancer Patients Patients with lower IRS-1 levels (solid line, n = 48) had significantly better DFS (P = 0.035 by log rank test) than patients with higher IRS-1 levels (dotted line, n = 99). The cutoff used to separate tumors expressing high levels of IRS-1 was the median of all samples analyzed. For graphing, follow up is truncated at 10 yr.
DISCUSSION
While the potent mitogenic effect of estrogen has been known for a long time, the mechanism of estrogen-mediated proliferation remains unclear. Estrogen acts through a nuclear hormone receptor, which upon activation can induce transcription of hormone-responsive genes. Many candidates were initially found to be involved in mitogenesis, including DNA-synthesizing enzymes, other hormone receptors, and autocrine and paracrine growth factors and their receptors (26). Further studies showing that overexpression of some of these growth factors and receptors caused estrogen-independent growth (19, 20) led to the “autocrine hypothesis”: estrogen stimulation of cells results in enhanced growth factor secretion, which then acts upon the same cells to stimulate proliferation (27). However, it is clear from recent studies that estrogen can directly affect the cell-cycle machinery before increased growth factor secretion is observed (28–30), suggesting that estrogen-mediated growth probably represents a combination of an early direct effect upon the cell cycle machinery and a late effect upon growth factor signaling.
Previous reports have shown that estrogen may regulate IGF activity by altering expression of IGFBPs and IGFR1 mRNA and IGF-binding sites (13, 13A, 24). We confirm reports of estrogen induction of IGFR1 expression but also show for the first time that estrogen can induce expression of the downstream signaling molecules, IRS-1 and IRS-2. Estrogen induction of IRS-1 expression was associated with increased tyrosine phosphorylation of IRS-1 after IGF stimulation and correlated with enhanced downstream MAPK activation. Thus, efficacy of IGF signaling, i.e. the maximum achievable activity, was increased. We would predict that other IGF-signaling pathways will be similarly affected and these are under investigation. While the increase in IRS-1 expression generally mirrored the increase in tyrosine phosphorylation, we cannot rule out the possibility that the increase in tyrosine phosphorylation of IRS-1 results from a change in stoichiometry or sites of phosphorylation. Additionally, there may be other factors controlling phosphorylation of IRS-1. For instance, decreased phosphorylation of IRS may be mediated by antiestrogen-induction of a specific tyrosine phosphatase activity, which has been proposed previously by other groups (22, 31, 32).
Synergism between estrogen and IGF has been shown in a number of model systems including normal breast (33), normal uterus (34), endometrial cancer cells (24), and breast cancer cells (13). In many of these systems, and in the data we present here, cotreatment with estrogen and IGF-I causes growth and signaling that is greater than IGF-I alone. While we hypothesize that estrogen actually sensitized cells to IGFs by up-regulating expression of IGF-signaling components, testing this hypothesis in growth assays is problematic due to the potential of IGFBPs to influence low levels of IGF interaction with the receptor, and the fact that expression of the IGFBPs are regulated by estrogen. However, our data showing enhanced growth with E2 and IGF-I compared with either ligand alone is completely consistent with the observation that increased levels of IRS-1 achieved by transfection result in enhanced growth of these cells and decreased requirements for estrogen (35).
It has recently been suggested that the main mechanism of estrogen-mediated growth is through early activation of cyclin-dependent kinases (Cdk-2 and Cdk-4), phosphorylation of pRB, and increased expression of cyclins within 2–8 h of hormone stimulation (28–30). As changes in growth factors tend to occur at later time points, their importance in estrogen-mediated growth has been challenged (29). However, the short time for induction of IRS expression and activation (8 h), similar to the recently reported early estrogen induction of tyrosine phosphorylation of IGFR1 and IRS-1 in the rat uterus (34), suggests that early-growth factor signaling activation may be an important component in estrogen-mediated growth. While regulation of cell-cycle regulatory components is probably sufficient to send cells through one round of the cell cycle, it is probably not sufficient for maximum estrogen-mediated growth. There must be, in addition, some level of growth factor signaling to allow cells to pass the restriction point of the cell cycle (36). We believe that the early induction of IRS-1, which has been shown to be a crucial rate-limiting step in IGF signal transduction in MCF-7 cells (18, 23), may “sensitize” cells to autocrine, paracrine, or endocrine sources of IGFs, which results in early activation of IGF signaling that allows cells to pass the restriction point. In contrast to the autocrine hypothesis (27, 37), estrogen treatment of cells increases expression of IGF signaling components within the cell and thus enhances sensitivity to any source of IGFs. After the initial entry into and movement through the G1 phase of the cell cycle, estrogen then causes a relatively late (24–48 h) induction of other growth-signaling proteins, e.g. IGFR1 and IGF-II (38), which then act in an autocrine or paracrine manner and cause synergistic and maximal proliferation.
While our studies indicate that estrogen induces IGFR1 and IRS-1 expression by a transcriptional mechanism, we do not know whether this is a direct effect of ER upon either promoter. For instance, estrogen induction of cyclin D1 expression probably does not result from a direct effect of ER upon the promoter, but rather from estrogen induction of other cis-acting factors such as c-myc or AP-1, which are both increased by estrogen (39, 40) and have been shown to increase cyclin D1 mRNA expression (30, 41). The IRS-1 promoter does have four consensus half-estrogen response elements, supporting the possibility of direct ER regulation, probably through synergism between multiple ERs (42). Furthermore, the IRS-1 promoter contains several AP-1 and SP-1 sites (43), which in other promoters have been shown to interact with the ER and activate transcription in a synergistic manner (44, 45).
Consistent with estrogen regulation of IRS-1 expression in vitro, IRS-1 expression is regulated in the MCF-7 xenograft model of breast cancer. In the presence of estrogen, and when the tumor is growing exponentially, IRS-1 is both expressed at high levels and tyrosine phosphorylated and is thus presumably involved in an active mitogenic signal transduction cascade. Indeed, active IRS-1 is associated with downstream MAPK activity. These data support previous work showing that estrogen increases tyrosine phosphorylation of IGFR1 and IRS-1 in the epithelial layer of the rat uterus (34). We do not know at present the factors responsible for the activation of IRS-1 in the MCF-7 xenograft. It may be a result of 1) autocrine or paracrine IGF expression from within the xenograft, 2) endocrine IGF-I from the host mouse, or 3) an unrelated IGF event, since IRS molecules are involved in signaling by several cytokines and other ligands (46). Interestingly, growth of MCF-7 xenograft tumors is retarded in mice lacking circulating IGF-I (11), suggesting that endocrine IGF-I may affect the proliferation of these cells. Removal of estrogen resulted in no detectable IRS-1, as was observed in vitro.
Considering the evidence indicating that ER regulates IRS-1 expression in vitro and in vivo, and after our initial observation that high IRS-1 expression is a poor prognostic indicator in small tumors (7), we reanalyzed the same data set of node-negative patients to determine whether IRS-1 levels have prognostic significance in ER-positive tumors. Analysis revealed that indeed when IRS-1 levels were examined in this set of patients, high IRS-1 expression was associated with a shortened DFS. These data support the concept that IRS-1 is an important molecule in ER-mediated growth, and that when IRS-1 is coexpressed with ER, tumor recurrence is more frequent.
In summary, we provide evidence that IGFR1, IRS-1, and IRS-2 are estrogen-regulated proteins. Increased expression of all of these components leads to enhanced IGF signaling, resulting in synergistic growth. Combined with the ability of IGF-I to transcriptionally activate the ER (47–50), this reveals complex cross-talk and synergism between these important signal transduction pathways that results in both pathways reinforcing each other. Further evidence for the importance of IGF components in ER action and breast cancer growth is provided by the prognostic significance of IRS-1 expression in ER-positive breast cancer patients. Thus, this and other data provide strong motivation for development of strategies to inhibit IGF action in human breast cancer.
MATERIALS AND METHODS
Materials
All materials and chemicals were purchased from Sigma Chemical Co. (St. Louis, MO) unless otherwise noted. IGF-I was purchased from GroPep Pty. Ltd. (Adelaide, Australia). ICI 182780 was a kind gift from Zeneca Pharmaceuticals (Macclesfield, U.K.). All tissue culture reagents were purchased from Life Technologies (Gaithersburg, MD) unless otherwise stated.
Cell Lines
MCF-7 cells have been maintained in our laboratory for many years (51). ZR-75 and T47D human breast cancer cells were purchased from the American Type Culture Collection (Manassas, VA). MDA-MB-435A cells were kindly provided by Nils Brunner (Finsen Laboratories, Copenhagen, Denmark). Cells were routinely maintained in improved MEM (IMEM) + 10% fetal bovine serum (Summit Biotechnology, Ft. Collins, CO) + 2 mm glutamine + 50 I.U./ml penicillin, 50 μg/ml streptomycin. SFM consisted of IMEM + 10 mm HEPES pH 7.4, 1 μg/ml transferrin, 1 μg/ml fibronectin, 2 mm glutamine, 50 I.U./ml penicillin, 50 μg/ml streptomycin, and trace elements (Biofluids, Rockville, MD).
Cell Stimulation and Lysis
Cells were plated at 5 × 105 cells in 6-cm dishes (Becton Dickinson and Co., Lincoln Park, NJ) and allowed to adhere overnight. The next day the medium was changed to SFM, and 24 h later cells were stimulated with various ligands for the indicated times. The concentrations were: E2, 10−9m; Tam, 10−6m; ICI, 10−6m; and IGF-I, 5× 10−9m. After stimulation cells were washed twice with cold PBS and then lysed in 150 μl of TNESV buffer with fresh protease inhibitors (50 mm Tris-HCl, pH 7.4, 1% NP-40, 3 mm EDTA, 100 mm NaCl, 10 mm sodium orthovanadate, 1 mm phenylmethylsulfonylfluoride, 20 μg/ml leupeptin, and 20μ g/ml aprotinin). Lysates were clarified by centrifugation at 14,000 × g for 15 min at 4 C, and lysates were stored at −20 C. Protein concentrations were determined by the bichionic acid method according to the manufacturer’s instructions (Pierce Chemical Co., Rockford, IL).
Immunoblotting and Immunoprecipitation
Total protein (50 μg) was resuspended in denaturing sample loading buffer (3% dithiothreitol, 0.1 m Tris-HCl, pH 6.8, 4% SDS, 0.2% bromophenol blue, 20% glycerol), separated by 8% SDS-PAGE, and electrophoretically transferred to nitrocellulose overnight at 4 C. The membrane was blocked with 5% milk-TBST (0.15 m NaCl, 0.01 m Tris-HCl, pH 7.4, 0.05% Tween 20). For anti-PY immunoblot, the membrane was incubated with a 1000:1 dilution of hrp-linked primary antibody (RC20, Transduction Laboratories, Inc., Lexington, KY) in TBST. Bands were visualized by ECL according to the manufacturer’s instructions (Pierce Chemical Co.). For activated MAPK (New England Biolabs, Inc., Beverly, MA) immunoblot, the membrane was incubated with a 1000:1 dilution of primary antibody in TBST. All other antibodies were diluted in TBST + 5% milk and used at a concentration of 1000:1 for IRS-1, IRS-2, and total MAPK (Upstate Biotechnology, Inc., Lake Placid, NY) and 200:1 for IGFR1 (Santa Cruz Biotechnology, Inc., Santa Cruz, CA).
RNAse Protection Assay
MCF-7 cells were plated at 3 × 106 cells in 15-cm dishes (Becton Dickinson and Co.) and allowed to adhere overnight. Cells were harvested by trypsin/EDTA and pelleted in 15-cm tubes. Total RNA was prepared by Qiagen RNeasy Midi Kit (Qiagen, Valencia, CA) according to the manufacturer’s instructions and checked for integrity by separation on a 1% agarose gel. Ribonuclease (RNAse) protection was performed according to our previously published method (15), and RNA loading was normalized to mRNA of the ribosomal protein 36B4 (52), which has previously been shown to be not regulated by estrogen. The IRS-1 cDNA was generated by PCR from MCF-7 genomic DNA using a 5′-primer containing an XbaI restriction site (5′-AGTTTCTAGACTCCAGCCCTGTTTGCATGT-3′) and a 3′-primer with an EcoRI restriction site (5′-CGAAGAATTCGTCAGCCCGCTTGTTGATGT-5′). The probe for IGFR1 (53) has been detailed previously.
Analysis of Human Tumors
Measurement of IRS-1 expression in 200 node-negative breast cancer patients and association with other clinical and laboratory factors have been described previously (7). DFS was defined as the time from date of diagnosis to the date of first recognition of relapse or last contact (censored). ER-positive (≥3 fmol/mg) tumors were dichotomized into those with IRS-1 levels above or below median IRS-1 value (0.61 arbitrary units) of the entire sample. DFS curves were estimated by the Kaplan-Meier method (54) and compared using the log-rank test.
Nude Mouse Model System
MCF-7 cells were grown in nude mice as xenografts as previously described in detail (55). Estrogen supplementation was provided in the form of a 3-week release 0.25-mg E2 pellet (Innovative Research of America, Rockville, MD). When the tumors had reached 8–10 mm in size, the mice were randomized into continued estrogen treatment or removal of the E2 pellet. After a further month, the mice were killed and the tumors were removed and treated as above (see Immunoblotting and Immunoprecipitation).
Acknowledgments
We thank Drs. S. Oesterreich and G. Chamness for helpful comments and suggestions.
This work was supported by Public Health Service Grants P01CA-30195 and P50CA-58183–06 and Cancer Center Support Grant P30CA-54174 from the National Cancer Institute (NIH).
These authors contributed equally to the manuscript.
Early Breast Cancer Trialists’ Collaborative Group
Van Den
McGuire
