Oncogenomics

Oncogene (2010) 29, 6071–6083; doi:10.1038/onc.2010.333; published online 16 August 2010

Biological reprogramming in acquired resistance to endocrine therapy of breast cancer

H Aguilar1, X Solé2,3, N Bonifaci2,3, J Serra-Musach2,3, A Islam4, N López-Bigas4, M Méndez-Pertuz5, R L Beijersbergen6, C Lázaro7, A Urruticoechea1 and M A Pujana1,2,3

  1. 1Translational Research Laboratory, Catalan Institute of Oncology, Bellvitge Institute for Biomedical Research (IDIBELL), L’Hospitalet, Barcelona, Spain
  2. 2Biomarkers and Susceptibility Unit, Catalan Institute of Oncology, IDIBELL, L’Hospitalet, Barcelona, Spain
  3. 3Biomedical Research Centre Network for Epidemiology and Public Health, Catalan Institute of Oncology, IDIBELL, L'Hospitalet, Barcelona, Spain
  4. 4Research Unit on Biomedical Informatics, Department of Experimental and Health Science, Pompeu Fabra University, Barcelona Biomedical Research Park, Barcelona, Spain
  5. 5Epithelial Carcinogenesis Group, Molecular Pathology Programme, Spanish National Cancer Research Centre, Madrid, Spain
  6. 6Division of Molecular Carcinogenesis, Center for Biomedical Genetics and Cancer Genomics Center, Netherlands Cancer Institute, Amsterdam, The Netherlands
  7. 7Molecular Diagnostics Unit, Hereditary Cancer Program, Catalan Institute of Oncology, IDIBELL, L’Hospitalet, Barcelona, Spain

Correspondence: Dr Urruticoechea or Dr MA Pujana, Translational Research Laboratory, Catalan Institute of Oncology, Bellvitge Institute for Biomedical Research, Gran Vía 199-203, L’Hospitalet, Barcelona 08907, Spain. E-mails: anderu@iconcologia.net or mapujana@iconcologia.net

Received 16 February 2010; Revised 22 June 2010; Accepted 28 June 2010; Published online 16 August 2010.

Top

Abstract

Endocrine therapies targeting the proliferative effect of 17β-estradiol through estrogen receptor α (ERα) are the most effective systemic treatment of ERα-positive breast cancer. However, most breast tumors initially responsive to these therapies develop resistance through molecular mechanisms that are not yet fully understood. The long-term estrogen-deprived (LTED) MCF7 cell model has been proposed to recapitulate acquired resistance to aromatase inhibitors in postmenopausal women. To elucidate this resistance, genomic, transcriptomic and molecular data were integrated into the time course of MCF7–LTED adaptation. Dynamic and widespread genomic changes were observed, including amplification of the ESR1 locus consequently linked to an increase in ERα. Dynamic transcriptomic profiles were also observed that correlated significantly with genomic changes and were predicted to be influenced by transcription factors known to be involved in acquired resistance or cell proliferation (for example, interferon regulatory transcription factor 1 and E2F1, respectively) but, notably, not by canonical ERα transcriptional function. Consistently, at the molecular level, activation of growth factor signaling pathways by EGFR/ERBB/AKT and a switch from phospho-Ser118 (pS118)- to pS167-ERα were observed during MCF7–LTED adaptation. Evaluation of relevant clinical settings identified significant associations between MCF7–LTED and breast tumor transcriptome profiles that characterize ERα-negative status, early response to letrozole and tamoxifen, and recurrence after tamoxifen treatment. In accordance with these profiles, MCF7–LTED cells showed increased sensitivity to inhibition of FGFR-mediated signaling with PD173074. This study provides mechanistic insight into acquired resistance to endocrine therapies of breast cancer and highlights a potential therapeutic strategy.

Keywords:

aromatase inhibition; breast cancer; estrogen receptor; fibroblast growth factor receptor; long-term estrogen-deprived; MCF7

Top

Introduction

Endocrine therapies are the most effective systemic treatment of estrogen receptor α (ERα)-positive breast cancer, and over two-thirds of patients are considered to present with this kind of disease (EBCTCG, 1998; Chlebowski et al., 2002; Winer et al., 2002). Two major strategies mediate the efficacy of these therapies. Drugs directed at ERα, mainly tamoxifen and fulvestrant, impede its binding to 17β-estradiol (17βE2) and, as a result, canonical ERα-dependent transcriptional regulation (Dowsett et al., 2005b). In contrast, the activity of aromatase inhibitor (AIs), which are the most effective treatment of breast cancer in postmenopausal women (the largest group of patients), is based on almost complete deprivation of estrogen production (Geisler et al., 2002). However, although endocrine therapies are initially effective, resistance occurs both in the form of tumor relapse after excision during adjuvant treatment and as a near-universal event when tumors cannot be excised. Importantly, acquired resistance is not commonly associated with the conversion to ERα-negative of previous ERα-positive breast tumors. Nevertheless, changes in ERα expression have been found in some series (Johnston et al., 1995).

Current literature supports the hypothesis that acquired resistance is mainly mediated by molecular events that—particularly in the case of resistance to AIs—lead to constitutive activation of ERα and growth factor signaling pathway cross-talk (Clarke et al., 2003; Martin et al., 2003; Yue et al., 2003; Jelovac et al., 2005; Normanno et al., 2005; Sabnis et al., 2005; Santen et al., 2005; Masri et al., 2008). On the basis of these studies, several clinical trials have attempted to overcome resistance through combination with growth factor signaling inhibitors (Johnston et al., 2005; Massarweh and Schiff, 2006). However, results of these trials are relatively insufficient and, therefore, deeper knowledge of the mechanism of acquired resistance is needed. In this study, we use a well-known model of acquired resistance to AIs (that is, MCF7–LTED) to comprehensively analyze the mechanism underpinning adaptation to estrogen deprivation in ERα-positive breast cancer. The novelty of our approach derives from two key characteristics: the description of dynamic genomic, transcriptomic and molecular changes; and the integrative analysis of these biological data levels to delineate the adaptive phenomena. Together, our data suggest a mechanism for acquired resistance that is mostly independent of canonical ERα transcriptional function and coordinated across biological levels.

Top

Results

Molecular changes during MCF7–LTED adaptation are consistent with growth signaling pathway cross-talk

To generate a breast cancer cell model of acquired resistance to AIs, MCF7 cells were cultured in a medium depleted of 17βE2, as described previously (Katzenellenbogen et al., 1987; Welshons and Jordan, 1987; Masamura et al., 1995; Jeng et al., 1998; Stephen et al., 2001; Chan et al., 2002). After initial quiescence for approximately 60 days, cell cultures adapted at 90 days, defined by a diminished proliferation-dependent response from 17βE2 (Figure 1a). The proliferation profile and time lapse for resistance are in agreement with previous studies (Yue et al., 2003; Martin et al., 2005; Shaw et al., 2006; Sadler et al., 2009). At the latter time point evaluated in this study, MCF7–LTED cells are thought to be in 17βE2-hypersensitive status, which should be acquired at between 3–6 months in 17βE2-deprived culture and precedes acquisition of full 17βE2 independence (Masamura et al., 1995; Santen et al., 2001; Chan et al., 2002; Martin et al., 2003).

Figure 1.
Figure 1 - Unfortunately we are unable to provide accessible alternative text for this. If you require assistance to access this image, please contact help@nature.com or the author

Cellular and molecular phenotypic characterization of the MCF7–LTED model. (a) 17βE2 dose–response growth curves of MCF7 and MCF7–LTED (at 90 days of 17βE2 deprivation) cell cultures, measured with MTT-based assays. (b) Western blot results of growth factor signaling pathway components involved in acquired resistance, compared to the original MCF7 cell line, and TUBA levels as control. Results are shown for whole-cell extracts and the antibodies (ab) used for each experiment are indicated in parentheses (see supplementary Materials and methods for details). (c) Western blot results of total progesterone receptor (PGR) and ERα, and ERα phosphoisoforms in whole-cell extracts of MCF7 and MCF7–LTED cell extracts. (d) Western blot results for total and phosphoisoforms of ERα, and glyceraldehyde 3-phosphate dehydrogenase (GAPDH) as control, in subcellular fractionations of MCF7 and MCF7–LTED extracts. EGFR, epidermal growth factor receptor.

Full figure and legend (144K)

Molecular characterization of the MCF7–LTED model showed activation of AKT in the form of phospho-Serine473 (pS473) and overexpression of total epidermal growth factor receptor, ERBB2 and ERBB4 but not ERBB3 (Figure 1b). Activation of AKT and overexpression of EGF and ERBB receptors are consistent with previous observations in MCF7–LTED cells and related breast cancer conditions (Masamura et al., 1995; Jeng et al., 1998; Shim et al., 2000; McClelland et al., 2001; Knowlden et al., 2003; Osborne et al., 2005; Yue et al., 2005; Song et al., 2006; Beeram et al., 2007; Lewis-Wambi et al., 2008; Santen et al., 2008; Ghayad et al., 2009). Notably, ERBB2 amplification was associated with resistance to endocrine therapies (Ellis et al., 2006), and treatment with the mTOR inhibitor RAD001 increased letrozole efficacy in a neoadjuvant setting of ERα-positive breast cancer (Baselga et al., 2009).

Despite the observed increase in growth factor receptor levels, the largest molecular differences relative to the parental MCF7 cells were in progesterone receptor and ERα expressions. Consistent with previous studies (Jeng et al., 1998; Shim et al., 2000; Stephen and Darbre, 2000; Martin et al., 2003; Shaw et al., 2006; Santen et al., 2008; Sadler et al., 2009), progesterone receptor was found to be strongly underexpressed, whereas ERα was strongly overexpressed in MCF7–LTED extracts (Figure 1c). Next, evaluation of activated forms of ERα revealed increased pS167 and pS305 but not pS118 (Figure 1c). Importantly, the major phosphorylation site in response to 17βE2 was previously demonstrated to be S118, whereas S167 is only modified in response to MAPK/AKT pathway signaling (Lannigan, 2003). In addition, pS305 was associated with resistance to tamoxifen treatment (Michalides et al., 2004). According to these observations, subcellular fractionation assays showed relatively higher levels of ERα in the cytoplasm of MCF7–LTED cells (Figure 1d). Higher levels of pS305 were also observed in the nuclear fraction of MCF7–LTED cells, whereas the level of pS118 was higher in the cytoplasm of the parental MCF7 cells (Figure 1d). Collectively, these data support the relevance of ERα and growth factor signaling pathway cross-talk in acquired resistance (Dowsett et al., 2005a; Massarweh and Schiff, 2007) and, accordingly, the validity of the MCF7–LTED model for delineating the underlying mechanism.

Dynamic and extensive genomic changes, including ESR1, during adaptation to 17βE2 deprivation

Although molecular alterations are seen to occur during the establishment of the MCF7–LTED model, the extent, time course and biological impact of genomic changes contributing to adaptation are unknown. To evaluate these changes, DNA samples from MCF7 cultures were taken at five time points between 0–150 days after 17βE2 deprivation and subsequently hybridized against a high-density single nucleotide polymorphism platform (see Materials and methods). Copy number alterations were defined with the Bayesian approach in the Genome Alteration Detection Analysis algorithm (Pique-Regi et al., 2008) and change trends computed using the Goodman–Kruskal gamma (Γ) coefficient. With |Γ|=1 (P<0.001), a dynamic shift through adaptation to 17βE2 deprivation was observed at many loci (Figure 2a). A marked transition was observed at the 90-day time point, akin with the resistance phenotype defined above. Notably, between days 0 and 150, a total of ~360 and ~550 Mega bases (Mb) were lost and gained, respectively. Changes involved most chromosomes, except 9, 10, 16, 17 and 19 (data not shown).

Figure 2.
Figure 2 - Unfortunately we are unable to provide accessible alternative text for this. If you require assistance to access this image, please contact help@nature.com or the author

Dynamic genomic copy number changes during MCF7–LTED adaptation. (a) Heatmap of copy number changes between 0–150-days time points of 17βE2 deprivation. Categorical changes are shown as indicated in the inset. (b) Left panel, dynamic copy number changes at chromosome 6q including the ESR1 locus, as marked by a vertical dashed orange line. Relative genomic gain is indicated by horizontal dashed red lines. Right panel, dynamic allelic imbalances at chromosome 6q including the ESR1 locus. (c) Profiles of ESR1 expression probes as indicated in the inset.

Full figure and legend (269K)

Although a large proportion of these genomic changes might be randomly selected (that is, passenger) relative to adaptation to 17βE2 deprivation—due to inherent genomic instability and clonal selection—the transcriptomic and biological annotation data analyses presented below point to causally implicated changes (that is, drivers). Of the putative driver loci, the region containing ESR1 showed a significant gain at the days 120–150 time point relative to 0–90 (Figure 2b, left panel). Thus, the maximum gain in chromosome 6q was detected at rs6939257 (log R ratio=0.38, corresponding to a 2.4-fold gain at 150 days), which is located ~40kb proximal to the 5′-exon of ESR1. Importantly, consistent results were observed across time points and concurrently highlighted by allelic imbalances at the same region (Figure 2b, right panel). Intriguingly, ESR1 locus gain appeared to be accompanied by accentuated losses of distal parts of the same chromosome arm (Figure 2b). The relatively late time points at which this gain was manifested relative to ESR1 overexpression (Figure 2c) suggest that transcriptional regulation has a major role at the transition, whereas genomic gain might participate in the settlement of resistance. Notably, all ESR1 probes in the microarray platform showed a trend for overexpression but none passed significance correction using the false discovery rate approach (nominal P-values <0.05 for seven out of nine probes); however, quantitative analysis revealed significant overexpression (subsequent sections). Together, these observations suggest that dynamic genomic and transcriptional changes has an important role in acquired resistance.

Correlation between genomic and transcriptomic changes, and the link to non-genomic function of ERα

To corroborate genomic changes and assess their consequences at the transcriptomic level, we examined genome-wide expression profiles at similar time points to those indicated above. Differential expression was defined using linear regression modeling with estimates of significance corrected by the false discovery rate approach (Benjamini and Hochberg, 1995). Consistent with the changes described above, a marked transition of transcriptomic profiles was observed at day 90 (Figure 3a, top panel). Examination of MKI67 expression as a mitotic marker in the same dataset revealed profiles consistent with an initial quiescent stage and subsequent higher proliferative potential (Figure 3a, bottom panel; cubic function P-values <0.05). Finally, in further support of the profiles observed, a significant agreement was observed between genomic copy number and transcriptomic changes: whereas 140 and 191 loci/genes show concordant genomic loss/underexpression and genomic gain/overexpression, respectively, 58 and 23 show discordant results, respectively (P=10−37) (Supplementary Table 1).

Figure 3.
Figure 3 - Unfortunately we are unable to provide accessible alternative text for this. If you require assistance to access this image, please contact help@nature.com or the author

Dynamic transcriptomic changes during MCF7–LTED adaptation. (a) Top panel, heatmap of transcriptomic changes between 0–180 days time points of 17βE2 deprivation. The color code of log2 expression ratios is shown in the bottom right panel. Bottom panel, profiles of MKI67 expression probes as indicated in the inset. (b) Expression profiles of genes relevant to ERα genomic function (GATA3) versus genes relevant to growth factor signaling pathways (EGFR and RAF1). Corresponding microarray probes are indicated in the inset. (c) Expression profiles of genes previously associated with ‘proliferation-on’ transcriptional programs. (d) Results of qRT–PCR assays for key genes in MCF7–LTED adaptation (left panel, transcription factors, additional validations are shown in subsequent figure; right panel, proliferation-associated genes). Results are representative of two independent experiments. Asterisks indicate significant differences (* P<0.05; **P<0.01). (e) Western blot results showing RAF1 overexpression in MCF7–LTED extracts relative to parental MCF7.

Full figure and legend (191K)

As supported by the molecular data presented above, acquired resistance may be mediated by constitutive activation of ERα and growth factor signaling pathway cross-talk (Martin et al., 2003; Jelovac et al., 2005; Sabnis et al., 2005; Santen et al., 2005; Masri et al., 2008). Thus, the identities of genes differentially expressed during MCF7–LTED adaptation may support the important role of the non-genomic function of ERα. GATA3, which encodes for a key transcription factor for luminal differentiation tightly coregulated with and by ERα in breast cancer cells (Eeckhoute et al., 2007; Kouros-Mehr et al., 2008), showed significant underexpression during MCF7–LTED adaptation (Figure 3b). In contrast, RAF1, which encodes for a key factor in mitogen-activated protein kinase signaling and the associated resistance to tamoxifen treatment (Iorns et al., 2008), showed significant overexpression (Figure 3b). Evaluation of these observations—and of ESR1—by quantitative reverse transcription polymerase chain reaction (quantitative reverse transcriptase (qRT)–PCR) and/or by western blotting assays corroborated the overexpression of ESR1 and RAF1 in MCF7–LTED cells (Figure 3d, left panel, and e, and Supplementary Table 2). Although a trend was observed, GATA3 expression change was not confirmed by qRT-PCR assays (Figure 3d, left panel), which might be because of alternative splicing patterns for this gene (Stamm et al., 2006) detected in MCF7 cells (Hoch et al., 1999). Finally, in accordance with the molecular characterization above (Figure 1b), EGFR also showed significant overexpression in the microarray dataset (Figure 3b). Taken together, these data further support pathway cross-talk, whereas suggest a transition toward a transcriptomic program independent of canonical ERα transcriptional function.

Transcriptional reprogramming independent of canonical ERα function linked to breast cancer prognostic and predictive signatures

To delineate the regulation of transcriptomic changes during MCF7–LTED adaptation, the representation of predicted transcription factor binding sites (TFBSs) at promoters was evaluated using predictions from the STORM algorithm (Schones et al., 2007) and position frequency matrices from TRANSFAC (Matys et al., 2006) and JASPAR (Bryne et al., 2008). Thus, both under- and overexpressed gene sets during adaptation were found to be enriched in a TFBS of ZFX (Figure 4a, bottom panels), a transcription factor involved in stem cell self-renewal (Galan-Caridad et al., 2007). The underexpressed set was found to be enriched in TFBSs principally of the interferon regulatory transcription factor (IRF) family (Figure 4a). Notably, IRF1 activity was previously demonstrated to be critical in response to endocrine therapies through promoting apoptosis (Gu et al., 2002; Bouker et al., 2004; Bowie et al., 2004). In contrast, the overexpressed set principally showed enrichment in E2Fs and SP1 TFBSs (Figure 4a). IRF1 probes in the microarray dataset indicated underexpression but none reached statistical significance: however, qRT–PCR and/or western assays confirmed the predicted expression changes for IRF1 as well as for E2F1/E2F1 and MYC (Figure 4b). According to these observations, the overexpressed set showed enrichment in the Gene Ontology biological process ‘DNA metabolic process’ (false discovery rate-adjusted P=0.01) and its corresponding parent terms (data not shown). Moreover, gene markers of cell cycle or proliferation were found in this set: for example, CCNA1, PCNT, POLD1 and TERF1 (Figure 3d), three of which were corroborated by qRT–PCR assays (Figure 3d, right panel). Together, these data suggest a link with a cell ‘proliferation-on’ transcriptional program with inhibition of apoptosis, as commonly observed in poor prognosis and non-response treatment expression profiles in breast cancer (Solé et al., 2009).

Figure 4.
Figure 4 - Unfortunately we are unable to provide accessible alternative text for this. If you require assistance to access this image, please contact help@nature.com or the author

Transcriptional reprogramming during MCF7–LTED adaptation. (a) Predicted TFBSs of TRANSFAC and JASPAR at two promoter cut-offs (between −1000 and 100bp, and −600 and 100bp relative to transcription start sites) for under- and over-expressed gene sets. False discovery rate (FDR)-adjusted P values are indicated as shown in the colored scale bar. (b) Left panel, results of qRT–PCR assays for E2F1 and IRF1. Results are representative of two independent experiments. Asterisks indicate significant differences (*P<0.05; **P<0.01). Right panel, Western blot results showing E2F1 and MYC overexpression in MCF7–LTED extracts relative to parental MCF7. (c) Results of expression correlation analyses across breast tumors (van de Vijver et al., 2002) between predicted TFs above and under- (green curves) or overexpressed (red curves) gene sets relative to 5000 randomly chosen equivalent sets (black curves). The graphs show real Pearson's correlation coefficient (PCC) values and, for random sets, the average distribution. Specific microarray probes are detailed in parentheses.

Full figure and legend (285K)

To further evaluate the depicted transcriptional program, we examined expression correlations between gene sets and the corresponding TF genes across a large series of breast tumors (van de Vijver et al., 2002). Significant correlations between the expression profiles of under- and/or overexpressed sets and E2F1, IRF1, IRF7 and SP1 were identified relative to randomly selected equivalent gene sets (P<10−3) (Figure 4c, top left panels show results for E2F1). In addition, the overexpressed set showed coexpression with MYC (Figure 4c, top right panels), which is consistent with TFBSs predictions (Figure 4a) and the role of MYC regulation of poor outcome signatures (Adler et al., 2006; Wolfer et al., 2010). Significant coexpression was also observed for ZFX (Figure 4c, bottom left panels), although with an opposite pattern that suggests a differential effect on transcriptional regulation during acquired resistance. A similar relationship was observed for coexpression with ESR1 (Figure 4c, right bottom panels), in agreement with the acquisition of a transcriptional program that may be mostly independent of canonical ERα transcriptional function, which therefore resembles ERα negativity (subsequent sections).

Transcriptional reprogramming was further evaluated using data on chromatin immunoprecipitation assays (that is, genomic response elements) and expression profiles linked to cell proliferation and canonical ERα transcriptional function (Balciunaite et al., 2005; Carroll et al., 2006; Xu et al., 2007). In agreement with TFBSs predictions and breast cancer profiles, the underexpressed set showed infra-representation of E2F1-AP2 response elements (P=0.041), whereas the overexpressed set showed enrichment of the same elements (P=0.022). Analysis of an independent dataset (Balciunaite et al., 2005) showed similar results for E2F4 response elements (underexpressed set, infra-representation P=0.072; overexpressed set, supra-representation P=0.003). However, neither under- nor overexpressed sets showed variation relative to random in the content of transcripts regulated through 17βE2 (Pgreater than or equal to0.30). Moreover, both sets showed infra-representation of ERα response elements detected in MCF7 cells (P<0.010). Collectively, these observations further support the acquisition of a cell proliferation-on transcriptional program that is mostly independent of canonical ERα transcriptional function.

Mechanistic insight into response to endocrine therapies

The genomic, transcriptomic and molecular changes observed during MCF7–LTED adaptation are compatible with common features of breast cancer prognostic and predictive signatures (Solé et al., 2009). Further insight into breast cancer biology and treatment was obtained through the examination of expression profiles that characterize ERα tumor status and letrozole and tamoxifen responses (van de Vijver et al., 2002; Ma et al., 2004; Miller et al., 2007). This analysis was performed using the non-parametric algorithm in the gene set enrichment analysis tool (Subramanian et al., 2005). When comparing ERα-positive and -negative tumors, significant asymmetrical distributions of expression values were detected: genes overexpressed during MCF7–LTED adaptation had higher expression values in ERα-negative tumors (P=0.024), whereas underexpressed genes had higher expression values in ERα-positive tumors (P=0.023) (Figure 5a). These results support the data obtained from the transition of ERα-positive MCF7 to MCF7–LTED cells which, although maintaining ERα positivity, as commonly occurs in the clinical setting, suggests a different role of ERα. In further support of this transition and its potential clinical impact, significant overexpression of key genes recently described in triple-negative breast tumors was also detected, including the therapeutic target fibroblast growth factor receptor 2 (FGFR2) (Figure 5b) (Turner et al., 2010a). In this set, the expression changes of CSNK1D and FGFR2 were assessed and subsequently corroborated by qRT–PCR and western blot assays (Figure 5c).

Figure 5.
Figure 5 - Unfortunately we are unable to provide accessible alternative text for this. If you require assistance to access this image, please contact help@nature.com or the author

Association with breast cancer pathological and clinical features. (a) In GSEA, the results of the statistical analysis of gene expression data are rank-ordered from the largest positive change to the largest negative change (the red-blue profile at the bottom of each graph). Next, a gene set is mapped to this distribution (black bar code) to assess potential bias using an enrichment score that reflects the degree to which this set is overrepresented at the extremes (top or bottom) of the entire ranked list. Here, GSEA outputs including the enrichment scores and the corresponding P values of the non-parametric analysis are shown. Association with ERα tumor status: left panel, overexpressed genes during MCF7–LTED adaptation had higher expression values in ERα-negative than in ERα-positive tumors; right panel, conversely, underexpressed genes during MCF7–LTED adaptation had higher expression values in ERα-positive than in ERα-negative tumors. (b) Expression profiles during MCF7–LTED adaptation of key genes overexpressed in triple-negative breast tumors (Turner et al., 2010a). (c) Top panel, results of qRT–PCR assays for CSNK1D, FGFR2 and IGFBP5. Results are representative of two independent experiments. Asterisks indicate significant differences (**P<0.01). Bottom panel, Western blot results showing CSNK1D and FGFR2 overexpression in MCF7–LTED extracts relative to parental MCF7 cells. (d) Association with recurrence after tamoxifen treatment: left panel, overexpressed genes during MCF7–LTED adaptation had higher expression values in recurrent tumors. A trend was observed for the underexpressed set (right panel). (e) Expression profiles during MCF7–LTED adaptation of immune-response genes overexpressed early and strongly in tamoxifen-treated normal mammary epithelial cells (Schild-Hay et al., 2009). (f) Association with early transcriptional responses to letrozole treatment: left panel, over-expressed genes during MCF7–LTED adaptation had higher expression values in pre-treatment than post-treatment samples (0–14 days). No significant differences were observed for the underexpressed set (right panel).

Full figure and legend (247K)

Examination of tamoxifen treatment response (Ma et al., 2004) revealed association of overexpressed genes with recurrence (P=0.008; underexpressed genes showed a trend with non-recurrence, P=0.13): specifically, overexpressed genes during MCF7–LTED adaptation had higher expression levels in tumors that were recurrent (Figure 5d), in accordance with their link to cell proliferation and ERα-negative tumor status. In addition, endorsing the importance of long-term transcriptional reprogramming in the clinical scenario, genes overexpressed early and strongly with tamoxifen treatment in normal mammary epithelial cells (Schild-Hay et al., 2009) showed significant underexpression in MCF7–LTED adaptation: 6 out of 20 genes showed this transition (concordance between studies P=10−5), namely BST2, IFI27, IFIT1, OAS1, RSAD2 and STAT1 (Figure 5e). These genes are known to participate in the immune response (Schild-Hay et al., 2009), which is in agreement with regulation through IRFs as detected above. At early time points of MCF7–LTED adaptation, these genes showed relative overexpression (Figure 5e) compatible with observations in normal mammary epithelial cells and with their role in apoptosis (Schild-Hay et al., 2009). Similar behavior, although not genome-wide statistically significant, was observed for a recently identified determinant of tamoxifen sensitivity, IGFBP5 (Ahn et al., 2010); nonetheless, qRT–PCR assays for this gene were consistent with underexpression in MCF7–LTED cells (Figure 5c).

Examination of early transcriptional response to letrozole treatment (before and after 14 days of treatment (Miller et al., 2007)) revealed significant differences in the expression of the overexpressed set between pre- and post-treatment samples (P=0.022) (Figure 5f). Conversely to the immune response set above, genes downregulated immediately after letrozole treatment exhibited overexpression in MCF7–LTED adaptation, again akin to the role of this set in cell proliferation and consistent with the profiles at early MCF7–LTED time points (Figure 3). Although underexpressed genes during MCF7–LTED adaptation did not show association with letrozole treatment (P=0.30; Figure 5f), expression data for wider time points after treatment will be necessary to comprehensively evaluate this subject, as this study shows dynamic profiles in acquired resistance. Therefore, although endocrine therapies may temporarily reduce tumor growth or size, this study proposes that the expression of key gene sets will be recovered and eventually increase—or, conversely, be reduced for the immune response set—to produce resistant tumors.

Potential second-line therapeutic strategies

Previous MCF7–LTED models have shown sensitivity to further antiestrogenic treatment with fulvestrant (ICI 182780), which acts as a pure antagonist leading to ERα degradation (Santen et al., 2001; Chan et al., 2002; Martin et al., 2003; Dowsett et al., 2005b). Following these observations, and as a potential second-line therapeutic strategy, we assessed the effect of ICI 182780 using late-stage (that is, 18 months) MCF7–LTED cells and their parental MCF7 cells. Notably, MCF7–LTED cells were relatively less sensitive to ICI 182780 (Figure 6a, including controls of ERα and CDK4 levels, right panels), which further supports the hypothesis of full 17βE2 independence at a late time point, but it is in apparent contradiction with previous evidence (Santen et al., 2001; Chan et al., 2002; Martin et al., 2003). An explanation for this discrepancy may be found in the examination of different time points, as there are possibly dynamic profiles after six months in 17βE2-deprived culture. Next, having identified the activation of growth factor signaling pathways and, in particular, of FGFR2/FGFR2 overexpression (Figure 5c), we sought to evaluate the effects of targeting FGFR-mediated signaling using the inhibitor PD173074 (Korc and Friesel, 2009). In this study, MCF7–LTED cells showed increased sensitivity to PD173074 relative to parental cells in a dose–response manner (Figure 6b). Importantly, similar effects have recently been described in models of triple-negative and endocrine-resistant (to tamoxifen) breast cancer (Turner et al., 2010a, 2010b), which further endorses the existence of reprogramming across different biological levels but partially independent of canonical ERα transcriptional function.

Figure 6.
Figure 6 - Unfortunately we are unable to provide accessible alternative text for this. If you require assistance to access this image, please contact help@nature.com or the author

Second-line therapeutic strategies according to the MCF7–LTED model. (a) Typical proliferation profiles of MCF7–LTED cells (17βE2 deprived for 18 months) and their parental MCF7 cells in response to different concentrations of ICI 182780, measured with MTT-based assays. Curves, which are significantly different between cell lines (ANOVA P=0.001 but not different between replicates, P>0.10), include the mean and standard deviation calculated from three replicates at each point. Controls of ICI 182780 treatment in both cell lines are shown in the right panel. (b) Typical proliferation profiles of MCF7–LTED cells (17βE2 deprived for 18 months) and their parental MCF7 in response to different concentrations of PD173074, measured with MTT-based assays. Curves, which are significantly different between cell lines (ANOVA P=3.3 × 10−6 but not different between replicates, P>0.30), include the mean and standard deviation calculated from three replicates at each point.ICI

Full figure and legend (82K)

Top

Discussion

When not fully excised, almost all breast tumors showing ERα positivity develop resistance to endocrine therapies. Metastatic breast cancer, in which the spread of cancer cells impedes surgical removal, is the paradigmatic clinical scenario that recapitulates this fact. In the majority of these cases, most effective first-line therapies achieve patient response rates no higher than 50% (Jonat et al., 1995). Even after a period of disease control, resistance appears in those cases previously considered sensitive, and second-line therapies achieve disease control rates of little higher than 20% for shorter periods of time (Rose et al., 2003). At the molecular level, however, ERα usually remains expressed in tumors that were initially sensitive but subsequently develop resistance. Our data are consistent with this observation, providing an explanation based on biological reprogramming independent of canonical ERα transcriptional function.

One of the most common hypotheses for the mechanism of resistance to AIs is that, following 17βE2 deprivation, growth factor signaling cross-talk modulates ERα, turning it into a supersensitive receptor able to regulate transcription in the presence of traces of estrogens or even independently of them. This theory implies that, although resistant to estrogen deprivation, these tumors are still dependent on ERα as a major driver of cell biology. The apparent sensitivity of MCF7–LTED cells to further antiestrogenic treatment such as fulvestrant (Santen et al., 2001; Chan et al., 2002; Martin et al., 2003) substantiates this theory, leading to the proposal of clinical trials including these treatments and/or growth factor signaling inhibitors. However, these approaches have shown limited efficacy; in fact, fulvestrant only elicits a response in approximately 7% of ERα-positive tumors that regrow under AI therapy, although its clinical use is justified based on a clinical benefit rate of 32% (Chia et al., 2008). On the other hand, trials with combinations of AIs and growth factor signal inhibitors have not improved outcomes in most cases (Johnston et al., 2005), except for recent results on novel combinations that are still under research (Baselga et al., 2009; Johnston et al., 2009). In this scenario, the dynamic profiles of our model coincide with previous findings concerning the relevance of growth factor signaling but, importantly, our data also support a different explanation of the role of ERα: firstly, the switch from pS118- to pS167-ERα, an isoform known to be related to MAPK/AKT signaling (Campbell et al., 2001), may indicate that sustained ERα expression is more a consequence of the activation of other pro-growth signals than a reflection of the importance of canonical ERα function; secondly, the fact that both over- and underexpressed gene sets during MCF7–LTED adaptation showed infra representation of ERα-responsive elements and were not enriched in 17βE2-mediated regulation from original MCF7 data (Carroll et al., 2006), suggests a molecular landscape in which ERα, despite sustained expression, is no longer a highly active transcription modulator. This is in further agreement with the distinct increasing relevance of other transcription factors involved in cell proliferation, such as E2Fs, which has a key role in poor prognosis and non-response treatment signatures in breast cancer (Chang and Nevins, 2006; Vuaroqueaux et al., 2007; Yu et al., 2007; Zhang et al., 2007; Shen et al., 2008; Solé et al., 2009). In addition, when compared with tumor transcriptome profiles, the MCF7–LTED model bears resemblance to ERα-negative profiles.

The biological reprogramming toward a phenotype similar to ERα-negative conditions is akin to the relatively lower sensitivity of late-stage MCF7–LTED cells to fulvestrant but their increased sensitivity to FGFR-mediated signaling inhibition. Although these observations further underline the importance of considering dynamic profiles beyond short-term adaptation to first-line endocrine therapies, they also highlight a common scenario with related clinical settings. Recently, FGFR1 amplification and overexpression has been revealed as key in resistance to adjuvant tamoxifen (Turner et al., 2010b), and integrative profiling has shown FGFR2 overexpression to be a promising therapeutic target in triple-negative breast cancer (Turner et al., 2010a). Accordingly, the expression of FGFR1 and FGFR2 is partially regulated by E2F1 (Tashiro et al., 2003; Kanai et al., 2009). Furthermore, targeting FGFR-mediated signaling that acts in an autocrine manner in a subset of breast cancers of poor prognosis has been shown to impair tumor outgrowth and metastasis (Dey et al., 2010). These data, together with observations linking FGFR2 and risk of breast cancer (Easton et al., 2007; Hunter et al., 2007; Meyer et al., 2008), further support that FGFR-mediated signaling is central in breast carcinogenesis.

Data obtained at different levels suggest a role of ERα expression increase linked to its non-genomic function. Indeed, genomic analysis supports the potential importance of ESR1 genomic amplification in acquisition of resistance. ESR1 locus gain, and its relevance in prognosis or prediction of response to endocrine therapies, is an area of open scientific debate. Two studies described a 11–20% frequency of amplification and improved prognosis for tumors showing this alteration, substantiating the importance of this biologic event on resistance to tamoxifen (Holst et al., 2007; Tomita et al., 2009). Nevertheless, other reports appear to contradict these findings (Adelaide et al., 2008; Brown et al., 2008; Horlings et al., 2008; Reis-Filho et al., 2008; Vincent-Salomon et al., 2008). Our work indicates that ESR1 gain may be a dynamic process; as such, ERα expression may respond to different events during resistance acquisition, with amplification being a possible landmark of estrogen independence. Owing to the potential clinical impact of this finding, further exploration using clinical samples is warranted.

Combined examination of those genes significantly deregulated during MCF7–LTED adaptation and those for which expression change is reported shortly after initiation of endocrine therapy is important, given the relevance of these early changes in the prediction of subsequent responses (Miller et al., 2007; Schild-Hay et al., 2009). Our data recapitulate the early transcriptional changes but, given the dynamic behavior of these alterations, also suggest that caution should be taken when assessing their relevance to the overall benefit of the treatment in the clinical setting. This study expands on previous molecular and clinical observations proposing a mechanism of acquired resistance coordinated across biological levels and mostly independent of canonical ERα transcriptional function.

Top

Materials and methods

Genomic and transcriptomic analyses

Genomic DNAs extracted with Qiagen DNeasy were hybridized onto CNV370-Quad 3 arrays (Illumina). Raw copy number values were obtained using the BeadStudio software with standard parameters. Copy number segmentation was computed using the Genome Alteration Detection Analysis algorithm (Pique-Regi et al., 2008). Regions with consistent alteration over time points were computed using the Γ coefficient with categorized changes −1, 0 or 1. Expression microarray data have been deposited at Gene Expression Omnibus GSE20361. The RMA method implemented in BioConductor (Gentleman et al., 2004) was applied for background adjusting, normalization and summarization of microarray probes (Gautier et al., 2004). For each probe—as appropriate, given potential transcriptional complexity and technical caveats (Alvarez et al., 2009)—a linear model was fitted with expression as the response variable and time as the independent variable. To evaluate correlation between copy number and expression changes, probes were mapped to their genomic position in the May 2006 (hg18) version of the human genome and P-values computed using a χ2 goodness-of-fit test. For TF correlations, an expression dataset was used (van de Vijver et al., 2002) without filtering by ERα status and individually considering probes matching each Entrez gene identity. Differences in PCC distributions were assessed using the Mann–Whitney non-parametric test.

Top

Conflict of interest

The authors declare no conflict of interest.

Top

References

  1. Adelaide J, Finetti P, Charafe-Jauffret E, Wicinski J, Jacquemier J, Sotiriou C et al. (2008). Absence of ESR1 amplification in a series of breast cancers. Int J Cancer 123: 2970–2972. | Article | PubMed | ChemPort |
  2. Adler AS, Lin M, Horlings H, Nuyten DS, van de Vijver MJ, Chang HY. (2006). Genetic regulators of large-scale transcriptional signatures in cancer. Nat Genet 38: 421–430. | Article | PubMed | ISI | ChemPort |
  3. Ahn BY, Elwi AN, Lee B, Trinh DL, Klimowicz AC, Yau A et al. (2010). Genetic screen identifies insulin-like growth factor binding protein 5 as a modulator of tamoxifen resistance in breast cancer. Cancer Res 70: 3013–3019. | Article | PubMed | ChemPort |
  4. Alvarez MJ, Sumazin P, Rajbhandari P, Califano A. (2009). Correlating measurements across samples improves accuracy of large-scale expression profile experiments. Genome Biol 10: R143. | Article | PubMed | ChemPort |
  5. Balciunaite E, Spektor A, Lents NH, Cam H, Te Riele H, Scime A et al. (2005). Pocket protein complexes are recruited to distinct targets in quiescent and proliferating cells. Mol Cell Biol 25: 8166–8178. | Article | PubMed | ChemPort |
  6. Baselga J, Semiglazov V, van Dam P, Manikhas A, Bellet M, Mayordomo J et al. (2009). Phase II randomized study of neoadjuvant everolimus plus letrozole compared with placebo plus letrozole in patients with estrogen receptor-positive breast cancer. J Clin Oncol 27: 2630–2637. | Article | PubMed | ChemPort |
  7. Beeram M, Tan QT, Tekmal RR, Russell D, Middleton A, DeGraffenried LA. (2007). Akt-induced endocrine therapy resistance is reversed by inhibition of mTOR signaling. Ann Oncol 18: 1323–1328. | Article | PubMed | ChemPort |
  8. Benjamini Y, Hochberg Y. (1995). Controlling the false discovery rate: a practical and powerful approach to multiple testing. J Roy Statist Soc Ser B 57: 289–300. | ISI |
  9. Bouker KB, Skaar TC, Fernandez DR, O'Brien KA, Riggins RB, Cao D et al. (2004). Interferon regulatory factor-1 mediates the proapoptotic but not cell cycle arrest effects of the steroidal antiestrogen ICI 182780 (faslodex, fulvestrant). Cancer Res 64: 4030–4039. | Article | PubMed | ISI | ChemPort |
  10. Bowie ML, Dietze EC, Delrow J, Bean GR, Troch MM, Marjoram RJ et al. (2004). Interferon-regulatory factor-1 is critical for tamoxifen-mediated apoptosis in human mammary epithelial cells. Oncogene 23: 8743–8755. | Article | PubMed | ISI | ChemPort |
  11. Brown LA, Hoog J, Chin SF, Tao Y, Zayed AA, Chin K et al. (2008). ESR1 gene amplification in breast cancer: a common phenomenon? Nat Genet 40: 806–807; author reply 810–2. | Article | PubMed | ChemPort |
  12. Bryne JC, Valen E, Tang MH, Marstrand T, Winther O, da Piedade I et al. (2008). JASPAR, the open access database of transcription factor-binding profiles: new content and tools in the 2008 update. Nucleic Acids Res 36: D102–D106. | Article | PubMed | ChemPort |
  13. Campbell RA, Bhat-Nakshatri P, Patel NM, Constantinidou D, Ali S, Nakshatri H. (2001). Phosphatidylinositol 3-kinase/AKT-mediated activation of estrogen receptor α: a new model for anti-estrogen resistance. J Biol Chem 276: 9817–9824. | Article | PubMed | ISI | ChemPort |
  14. Carroll JS, Meyer CA, Song J, Li W, Geistlinger TR, Eeckhoute J et al. (2006). Genome-wide analysis of estrogen receptor binding sites. Nat Genet 38: 1289–1297. | Article | PubMed | ISI | ChemPort |
  15. Chan CM, Martin LA, Johnston SR, Ali S, Dowsett M. (2002). Molecular changes associated with the acquisition of oestrogen hypersensitivity in MCF-7 breast cancer cells on long-term oestrogen deprivation. J Steroid Biochem Mol Biol 81: 333–341. | Article | PubMed | ISI | ChemPort |
  16. Chang JT, Nevins JR. (2006). GATHER: a systems approach to interpreting genomic signatures. Bioinformatics 22: 2926–2933. | Article | PubMed | ChemPort |
  17. Chia S, Gradishar W, Mauriac L, Bines J, Amant F, Federico M et al. (2008). Double-blind, randomized placebo controlled trial of fulvestrant compared with exemestane after prior nonsteroidal aromatase inhibitor therapy in postmenopausal women with hormone receptor-positive, advanced breast cancer: results from EFECT. J Clin Oncol 26: 1664–1670. | Article | PubMed | ChemPort |
  18. Chlebowski RT, Col N, Winer EP, Collyar DE, Cummings SR, Vogel III VG et al. (2002). American Society of Clinical Oncology technology assessment of pharmacologic interventions for breast cancer risk reduction including tamoxifen, raloxifene, and aromatase inhibition. J Clin Oncol 20: 3328–3343. | Article | PubMed | ChemPort |
  19. Clarke R, Liu MC, Bouker KB, Gu Z, Lee RY, Zhu Y et al. (2003). Antiestrogen resistance in breast cancer and the role of estrogen receptor signaling. Oncogene 22: 7316–7339. | Article | PubMed | ISI | ChemPort |
  20. Dey JH, Bianchi F, Voshol J, Bonenfant D, Oakeley EJ, Hynes NE. (2010). Targeting fibroblast growth factor receptors blocks PI3K/AKT signaling, induces apoptosis, and impairs mammary tumor outgrowth and metastasis. Cancer Res 70: 4151–4162. | Article | PubMed | ChemPort |
  21. Dowsett M, Martin LA, Smith I, Johnston S. (2005a). Mechanisms of resistance to aromatase inhibitors. J Steroid Biochem Mol Biol 95: 167–172. | Article | PubMed | ChemPort |
  22. Dowsett M, Nicholson RI, Pietras RJ. (2005b). Biological characteristics of the pure antiestrogen fulvestrant: overcoming endocrine resistance. Breast Cancer Res Treat 93 (Suppl 1): S11–S18. | Article | PubMed | ChemPort |
  23. Easton DF, Pooley KA, Dunning AM, Pharoah PD, Thompson D, Ballinger DG et al. (2007). Genome-wide association study identifies novel breast cancer susceptibility loci. Nature 447: 1087–1093. | Article | PubMed | ISI | ChemPort |
  24. EBCTCG (1998). Tamoxifen for early breast cancer: an overview of the randomised trials. Early Breast Cancer Trialists’ Collaborative Group. Lancet 351: 1451–1467. | Article | PubMed | ISI
  25. Eeckhoute J, Keeton EK, Lupien M, Krum SA, Carroll JS, Brown M. (2007). Positive cross-regulatory loop ties GATA-3 to estrogen receptor α expression in breast cancer. Cancer Res 67: 6477–6483. | Article | PubMed | ChemPort |
  26. Ellis MJ, Tao Y, Young O, White S, Proia AD, Murray J et al. (2006). Estrogen-independent proliferation is present in estrogen-receptor HER2-positive primary breast cancer after neoadjuvant letrozole. J Clin Oncol 24: 3019–3025. | Article | PubMed | ChemPort |
  27. Galan-Caridad JM, Harel S, Arenzana TL, Hou ZE, Doetsch FK, Mirny LA et al. (2007). Zfx controls the self-renewal of embryonic and hematopoietic stem cells. Cell 129: 345–357. | Article | PubMed | ISI | ChemPort |
  28. Gautier L, Cope L, Bolstad BM, Irizarry RA. (2004). affy--analysis of Affymetrix GeneChip data at the probe level. Bioinformatics 20: 307–315. | Article | PubMed | ISI | ChemPort |
  29. Geisler J, Haynes B, Anker G, Dowsett M, Lonning PE. (2002). Influence of letrozole and anastrozole on total body aromatization and plasma estrogen levels in postmenopausal breast cancer patients evaluated in a randomized, cross-over study. J Clin Oncol 20: 751–757. | Article | PubMed | ISI | ChemPort |
  30. Gentleman RC, Carey VJ, Bates DM, Bolstad B, Dettling M, Dudoit S et al. (2004). Bioconductor: open software development for computational biology and bioinformatics. Genome Biol 5: R80. | Article | PubMed
  31. Ghayad SE, Vendrell JA, Larbi SB, Dumontet C, Bieche I, Cohen PA. (2009). Endocrine resistance associated with activated ErbB system in breast cancer cells is reversed by inhibiting MAPK or PI3K/Akt signaling pathways. Int J Cancer 126: 545–562. | Article | ChemPort |
  32. Gu Z, Lee RY, Skaar TC, Bouker KB, Welch JN, Lu J et al. (2002). Association of interferon regulatory factor-1, nucleophosmin, nuclear factor-kappaB, and cyclic AMP response element binding with acquired resistance to Faslodex (ICI 182780). Cancer Res 62: 3428–3437. | PubMed | ChemPort |
  33. Hoch RV, Thompson DA, Baker RJ, Weigel RJ. (1999). GATA-3 is expressed in association with estrogen receptor in breast cancer. Int J Cancer 84: 122–128. | Article | PubMed | ISI | ChemPort |
  34. Holst F, Stahl PR, Ruiz C, Hellwinkel O, Jehan Z, Wendland M et al. (2007). Estrogen receptor α (ESR1) gene amplification is frequent in breast cancer. Nat Genet 39: 655–660. | Article | PubMed | ISI | ChemPort |
  35. Horlings HM, Bergamaschi A, Nordgard SH, Kim YH, Han W, Noh DY et al. (2008). ESR1 gene amplification in breast cancer: a common phenomenon? Nat Genet 40: 807–808; author reply 810–2. | Article | PubMed | ChemPort |
  36. Hunter DJ, Kraft P, Jacobs KB, Cox DG, Yeager M, Hankinson SE et al. (2007). A genome-wide association study identifies alleles in FGFR2 associated with risk of sporadic postmenopausal breast cancer. Nat Genet 39: 870–874. | Article | PubMed | ChemPort |
  37. Iorns E, Turner NC, Elliott R, Syed N, Garrone O, Gasco M et al. (2008). Identification of CDK10 as an important determinant of resistance to endocrine therapy for breast cancer. Cancer Cell 13: 91–104. | Article | PubMed | ChemPort |
  38. Jelovac D, Sabnis G, Long BJ, Macedo L, Goloubeva OG, Brodie AM. (2005). Activation of mitogen-activated protein kinase in xenografts and cells during prolonged treatment with aromatase inhibitor letrozole. Cancer Res 65: 5380–5389. | Article | PubMed | ChemPort |
  39. Jeng MH, Shupnik MA, Bender TP, Westin EH, Bandyopadhyay D, Kumar R et al. (1998). Estrogen receptor expression and function in long-term estrogen-deprived human breast cancer cells. Endocrinology 139: 4164–4174. | Article | PubMed | ISI | ChemPort |
  40. Johnston S, Pippen Jr J, Pivot X, Lichinitser M, Sadeghi S, Dieras V et al. (2009). Lapatinib combined with letrozole versus letrozole and placebo as first-line therapy for postmenopausal hormone receptor-positive metastatic breast cancer. J Clin Oncol 27: 5538–5546. | Article | PubMed | ChemPort |
  41. Johnston SR, Martin LA, Head J, Smith I, Dowsett M. (2005). Aromatase inhibitors: combinations with fulvestrant or signal transduction inhibitors as a strategy to overcome endocrine resistance. J Steroid Biochem Mol Biol 95: 173–181. | Article | PubMed | ChemPort |
  42. Johnston SR, Saccani-Jotti G, Smith IE, Salter J, Newby J, Coppen M et al. (1995). Changes in estrogen receptor, progesterone receptor, and pS2 expression in tamoxifen-resistant human breast cancer. Cancer Res 55: 3331–3338. | PubMed | ISI | ChemPort |
  43. Jonat W, Kaufmann M, Blamey RW, Howell A, Collins JP, Coates A et al. (1995). A randomised study to compare the effect of the luteinising hormone releasing hormone (LHRH) analogue goserelin with or without tamoxifen in pre- and perimenopausal patients with advanced breast cancer. Eur J Cancer 31A: 137–142.
  44. Kanai M, Tashiro E, Maruki H, Minato Y, Imoto M. (2009). Transcriptional regulation of human fibroblast growth factor receptor 1 by E2F-1. Gene 438: 49–56. | Article | PubMed | ChemPort |
  45. Katzenellenbogen BS, Kendra KL, Norman MJ, Berthois Y. (1987). Proliferation, hormonal responsiveness, and estrogen receptor content of MCF-7 human breast cancer cells grown in the short-term and long-term absence of estrogens. Cancer Res 47: 4355–4360. | PubMed | ISI | ChemPort |
  46. Knowlden JM, Hutcheson IR, Jones HE, Madden T, Gee JM, Harper ME et al. (2003). Elevated levels of epidermal growth factor receptor/c-erbB2 heterodimers mediate an autocrine growth regulatory pathway in tamoxifen-resistant MCF-7 cells. Endocrinology 144: 1032–1044. | Article | PubMed | ISI | ChemPort |
  47. Korc M, Friesel RE. (2009). The role of fibroblast growth factors in tumor growth. Curr Cancer Drug Targets 9: 639–651. | Article | PubMed | ChemPort |
  48. Kouros-Mehr H, Kim JW, Bechis SK, Werb Z. (2008). GATA-3 and the regulation of the mammary luminal cell fate. Curr Opin Cell Biol 20: 164–170. | Article | PubMed | ChemPort |
  49. Lannigan DA. (2003). Estrogen receptor phosphorylation. Steroids 68: 1–9. | Article | PubMed | ISI | ChemPort |
  50. Lewis-Wambi JS, Cunliffe HE, Kim HR, Willis AL, Jordan VC. (2008). Overexpression of CEACAM6 promotes migration and invasion of oestrogen-deprived breast cancer cells. Eur J Cancer 44: 1770–1779. | Article | PubMed | ChemPort |
  51. Ma XJ, Wang Z, Ryan PD, Isakoff SJ, Barmettler A, Fuller A et al. (2004). A two-gene expression ratio predicts clinical outcome in breast cancer patients treated with tamoxifen. Cancer Cell 5: 607–616. | Article | PubMed | ISI | ChemPort |
  52. Martin LA, Farmer I, Johnston SR, Ali S, Dowsett M. (2005). Elevated ERK1/ERK2/estrogen receptor cross-talk enhances estrogen-mediated signaling during long-term estrogen deprivation. Endocr Relat Cancer 12 (Suppl 1): S75–S84. | Article | PubMed | ChemPort |
  53. Martin LA, Farmer I, Johnston SR, Ali S, Marshall C, Dowsett M. (2003). Enhanced estrogen receptor (ER) α, ERBB2, and MAPK signal transduction pathways operate during the adaptation of MCF-7 cells to long term estrogen deprivation. J Biol Chem 278: 30458–30468. | Article | PubMed | ISI | ChemPort |
  54. Masamura S, Santner SJ, Heitjan DF, Santen RJ. (1995). Estrogen deprivation causes estradiol hypersensitivity in human breast cancer cells. J Clin Endocrinol Metab 80: 2918–2925. | Article | PubMed | ISI | ChemPort |
  55. Masri S, Phung S, Wang X, Wu X, Yuan YC, Wagman L et al. (2008). Genome-wide analysis of aromatase inhibitor-resistant, tamoxifen-resistant, and long-term estrogen-deprived cells reveals a role for estrogen receptor. Cancer Res 68: 4910–4918. | Article | PubMed | ChemPort |
  56. Massarweh S, Schiff R. (2006). Resistance to endocrine therapy in breast cancer: exploiting estrogen receptor/growth factor signaling crosstalk. Endocr Relat Cancer 13 (Suppl 1): S15–S24. | Article | PubMed | ChemPort |
  57. Massarweh S, Schiff R. (2007). Unraveling the mechanisms of endocrine resistance in breast cancer: new therapeutic opportunities. Clin Cancer Res 13: 1950–1954. | Article | PubMed | ChemPort |
  58. Matys V, Kel-Margoulis OV, Fricke E, Liebich I, Land S, Barre-Dirrie A et al. (2006). TRANSFAC and its module TRANSCompel: transcriptional gene regulation in eukaryotes. Nucleic Acids Res 34: D108–D110. | Article | PubMed | ISI | ChemPort |
  59. McClelland RA, Barrow D, Madden TA, Dutkowski CM, Pamment J, Knowlden JM et al. (2001). Enhanced epidermal growth factor receptor signaling in MCF7 breast cancer cells after long-term culture in the presence of the pure antiestrogen ICI 182780 (Faslodex). Endocrinology 142: 2776–2788. | Article | PubMed | ISI | ChemPort |
  60. Meyer KB, Maia AT, O'Reilly M, Teschendorff AE, Chin SF, Caldas C et al. (2008). Allele-specific up-regulation of FGFR2 increases susceptibility to breast cancer. PLoS Biol 6: e108. | Article | PubMed | ChemPort |
  61. Michalides R, Griekspoor A, Balkenende A, Verwoerd D, Janssen L, Jalink K et al. (2004). Tamoxifen resistance by a conformational arrest of the estrogen receptor α after PKA activation in breast cancer. Cancer Cell 5: 597–605. | Article | PubMed | ISI | ChemPort |
  62. Miller WR, Larionov AA, Renshaw L, Anderson TJ, White S, Murray J et al. (2007). Changes in breast cancer transcriptional profiles after treatment with the aromatase inhibitor, letrozole. Pharmacogenet Genomics 17: 813–826. | Article | PubMed | ChemPort |
  63. Normanno N, Di Maio M, De Maio E, De Luca A, de Matteis A, Giordano A et al. (2005). Mechanisms of endocrine resistance and novel therapeutic strategies in breast cancer. Endocr Relat Cancer 12: 721–747. | Article | PubMed | ISI | ChemPort |
  64. Osborne CK, Shou J, Massarweh S, Schiff R. (2005). Crosstalk between estrogen receptor and growth factor receptor pathways as a cause for endocrine therapy resistance in breast cancer. Clin Cancer Res 11: 865s–8670s. | PubMed | ChemPort |
  65. Pique-Regi R, Monso-Varona J, Ortega A, Seeger RC, Triche TJ, Asgharzadeh S. (2008). Sparse representation and Bayesian detection of genome copy number alterations from microarray data. Bioinformatics 24: 309–318. | Article | PubMed | ChemPort |
  66. Reis-Filho JS, Drury S, Lambros MB, Marchio C, Johnson N, Natrajan R et al. (2008). ESR1 gene amplification in breast cancer: a common phenomenon? Nat Genet 40: 809–810; author reply 810–2. | Article | PubMed | ChemPort |
  67. Rose C, Vtoraya O, Pluzanska A, Davidson N, Gershanovich M, Thomas R et al. (2003). An open randomised trial of second-line endocrine therapy in advanced breast cancer: comparison of the aromatase inhibitors letrozole and anastrozole. Eur J Cancer 39: 2318–2327. | Article | PubMed | ISI | ChemPort |
  68. Sabnis GJ, Jelovac D, Long B, Brodie A. (2005). The role of growth factor receptor pathways in human breast cancer cells adapted to long-term estrogen deprivation. Cancer Res 65: 3903–3910. | Article | PubMed | ISI | ChemPort |
  69. Sadler AJ, Pugazhendhi D, Darbre PD. (2009). Use of global gene expression patterns in mechanistic studies of oestrogen action in MCF7 human breast cancer cells. J Steroid Biochem Mol Biol 114: 21–32. | Article | PubMed | ChemPort |
  70. Santen R, Jeng MH, Wang JP, Song R, Masamura S, McPherson R et al. (2001). Adaptive hypersensitivity to estradiol: potential mechanism for secondary hormonal responses in breast cancer patients. J Steroid Biochem Mol Biol 79: 115–125. | Article | PubMed | ISI | ChemPort |
  71. Santen RJ, Song RX, Masamura S, Yue W, Fan P, Sogon T et al. (2008). Adaptation to estradiol deprivation causes up-regulation of growth factor pathways and hypersensitivity to estradiol in breast cancer cells. Adv Exp Med Biol 630: 19–34. | PubMed |
  72. Santen RJ, Song RX, Zhang Z, Kumar R, Jeng MH, Masamura A et al. (2005). Long-term estradiol deprivation in breast cancer cells up-regulates growth factor signaling and enhances estrogen sensitivity. Endocr Relat Cancer 12 (Suppl 1): S61–S73. | Article | PubMed | ChemPort |
  73. Schild-Hay LJ, Leil TA, Divi RL, Olivero OA, Weston A, Poirier MC. (2009). Tamoxifen induces expression of immune response-related genes in cultured normal human mammary epithelial cells. Cancer Res 69: 1150–1155. | Article | PubMed | ChemPort |
  74. Schones DE, Smith AD, Zhang MQ. (2007). Statistical significance of cis-regulatory modules. BMC Bioinformatics 8: 19. | Article | PubMed | ChemPort |
  75. Shaw LE, Sadler AJ, Pugazhendhi D, Darbre PD. (2006). Changes in oestrogen receptor-α and -β during progression to acquired resistance to tamoxifen and fulvestrant (Faslodex, ICI 182780) in MCF7 human breast cancer cells. J Steroid Biochem Mol Biol 99: 19–32. | Article | PubMed | ISI | ChemPort |
  76. Shen R, Chinnaiyan AM, Ghosh D. (2008). Pathway analysis reveals functional convergence of gene expression profiles in breast cancer. BMC Med Genomics 1: 28. | Article | PubMed | ChemPort |
  77. Shim WS, Conaway M, Masamura S, Yue W, Wang JP, Kmar R et al. (2000). Estradiol hypersensitivity and mitogen-activated protein kinase expression in long-term estrogen deprived human breast cancer cells in vivo. Endocrinology 141: 396–405. | Article | PubMed | ISI | ChemPort |
  78. Solé X, Bonifaci N, López-Bigas N, Berenguer A, Hernández P, Reina O et al. (2009). Biological convergence of cancer signatures. PLoS One 4: e4544. | Article | PubMed | ChemPort |
  79. Song RX, Fan P, Yue W, Chen Y, Santen RJ. (2006). Role of receptor complexes in the extranuclear actions of estrogen receptor α in breast cancer. Endocr Relat Cancer 13 (Suppl 1): S3–13. | Article | PubMed | ChemPort |
  80. Stamm S, Riethoven JJ, Le Texier V, Gopalakrishnan C, Kumanduri V, Tang Y et al. (2006). ASD: a bioinformatics resource on alternative splicing. Nucleic Acids Res 34: D46–D55. | Article | PubMed | ChemPort |
  81. Stephen R, Darbre PD. (2000). Loss of growth inhibitory effects of retinoic acid in human breast cancer cells following long-term exposure to retinoic acid. Br J Cancer 83: 1183–1191. | Article | PubMed | ChemPort |
  82. Stephen RL, Shaw LE, Larsen C, Corcoran D, Darbre PD. (2001). Insulin-like growth factor receptor levels are regulated by cell density and by long term estrogen deprivation in MCF7 human breast cancer cells. J Biol Chem 276: 40080–40086. | Article | PubMed | ISI | ChemPort |
  83. Subramanian A, Tamayo P, Mootha VK, Mukherjee S, Ebert BL, Gillette MA et al. (2005). Gene set enrichment analysis: a knowledge-based approach for interpreting genome-wide expression profiles. Proc Natl Acad Sci USA 102: 15545–15550. | Article | PubMed | ChemPort |
  84. Tashiro E, Minato Y, Maruki H, Asagiri M, Imoto M. (2003). Regulation of FGF receptor-2 expression by transcription factor E2F-1. Oncogene 22: 5630–5635. | Article | PubMed | ChemPort |
  85. Tomita S, Zhang Z, Nakano M, Ibusuki M, Kawazoe T, Yamamoto Y et al. (2009). Estrogen receptor α gene ESR1 amplification may predict endocrine therapy responsiveness in breast cancer patients. Cancer Sci 100: 1012–1017. | Article | PubMed | ChemPort |
  86. Turner N, Lambros MB, Horlings HM, Pearson A, Sharpe R, Natrajan R et al. (2010a). Integrative molecular profiling of triple negative breast cancers identifies amplicon drivers and potential therapeutic targets. Oncogene 29: 2013–2023. | Article | ChemPort |
  87. Turner N, Pearson A, Sharpe R, Lambros M, Geyer F, Lopez-Garcia MA et al. (2010b). FGFR1 amplification drives endocrine therapy resistance and is a therapeutic target in breast cancer. Cancer Res 70: 2085–2094. | Article | ChemPort |
  88. van de Vijver MJ, He YD, van't Veer LJ, Dai H, Hart AA, Voskuil DW et al. (2002). A gene-expression signature as a predictor of survival in breast cancer. N Engl J Med 347: 1999–2009. | Article | PubMed | ISI | ChemPort |
  89. Vincent-Salomon A, Raynal V, Lucchesi C, Gruel N, Delattre O. (2008). ESR1 gene amplification in breast cancer: a common phenomenon? Nat Genet 40: 809; author reply 810–2. | Article | PubMed | ChemPort |
  90. Vuaroqueaux V, Urban P, Labuhn M, Delorenzi M, Wirapati P, Benz CC et al. (2007). Low E2F1 transcript levels are a strong determinant of favorable breast cancer outcome. Breast Cancer Res 9: R33. | Article | PubMed | ChemPort |
  91. Welshons WV, Jordan VC. (1987). Adaptation of estrogen-dependent MCF-7 cells to low estrogen (phenol red-free) culture. Eur J Cancer Clin Oncol 23: 1935–1939. | Article | PubMed | ChemPort |
  92. Winer EP, Hudis C, Burstein HJ, Chlebowski RT, Ingle JN, Edge SB et al. (2002). American Society of Clinical Oncology technology assessment on the use of aromatase inhibitors as adjuvant therapy for women with hormone receptor-positive breast cancer: status report 2002. J Clin Oncol 20: 3317–3327. | Article | PubMed | ISI | ChemPort |
  93. Wolfer A, Wittner BS, Irimia D, Flavin RJ, Lupien M, Gunawardane RN et al. (2010). MYC regulation of a ‘poor-prognosis’ metastatic cancer cell state. Proc Natl Acad Sci USA 107: 3698–3703. | Article | PubMed
  94. Xu X, Bieda M, Jin VX, Rabinovich A, Oberley MJ, Green R et al. (2007). A comprehensive ChIP-chip analysis of E2F1, E2F4, and E2F6 in normal and tumor cells reveals interchangeable roles of E2F family members. Genome Res 17: 1550–1561. | Article | PubMed | ChemPort |
  95. Yu JX, Sieuwerts AM, Zhang Y, Martens JW, Smid M, Klijn JG et al. (2007). Pathway analysis of gene signatures predicting metastasis of node-negative primary breast cancer. BMC Cancer 7: 182. | Article | PubMed | ChemPort |
  96. Yue W, Wang J, Li Y, Fan P, Santen RJ. (2005). Farnesylthiosalicylic acid blocks mammalian target of rapamycin signaling in breast cancer cells. Int J Cancer 117: 746–754. | Article | PubMed | ChemPort |
  97. Yue W, Wang JP, Conaway MR, Li Y, Santen RJ. (2003). Adaptive hypersensitivity following long-term estrogen deprivation: involvement of multiple signal pathways. J Steroid Biochem Mol Biol 86: 265–274. | Article | PubMed | ChemPort |
  98. Zhang Z, Chen D, Fenstermacher DA. (2007). Integrated analysis of independent gene expression microarray datasets improves the predictability of breast cancer outcome. BMC Genomics 8: 331. | Article | PubMed | ChemPort |
Top

Acknowledgements

This study was funded by the Spanish Ministry of Health CIBERESP and grant 06/0545 and the Ministry of Science and Innovation (MICINN) grant SAF06/05399. HA was supported by a MICINN postdoctoral fellowship and MAP was a ‘Ramón y Cajal’ Researcher with the MICINN.

Supplementary Information accompanies the paper on the Oncogene website