Increased Expression of Estrogen Receptor β mRNA in Tamoxifen-resistant Breast Cancer Patients

Introduction

The nonsteroidal antiestrogen tamoxifen is currently the treatment of choice for hormone-dependent breast cancer; clinically, around two-thirds of all breast cancer patients with ER 3 -α⁺ tumors will initially benefit from tamoxifen treatment. However, most eventually develop tamoxifen resistance (1) . Acquired resistance is likely to be multifactorial, but at present, this is incompletely understood. Nevertheless, a number of mechanisms have been proposed. These include local metabolism of tamoxifen to less potent or unstable metabolites, loss/mutation of target receptor, or altered signal transduction (reviewed in Ref. 2 ). More specifically, alterations in protein kinase A signaling pathways (3) and overexpression of TGF-β1 (4) have also been implicated. Mutations in ER-α have also been considered, although the significance of such mutations has largely been discounted after the observation that these occur at low frequency, with only about 10% of tamoxifen-resistant breast tumors harboring ER-α mutations (5) . However, all of these proposals were suggested before the identification of a second ER, ER-β (6) , which may represent a novel prognostic factor in patients with breast cancer (7 , 8) . An inverse relationship has been identified between ER-β and expression of PR, which is generally a good prognostic marker in breast cancer (7) . Furthermore, coexpression of ER-α and ER-β is associated with tumors with a poor prognosis, i.e., those that are lymph node positive and of higher grade (8) . Further still, when ER signaling mechanisms are considered, in particular, signaling from an AP-1 response element, this may have important implications for antiestrogen resistance because ER subtypes signal in opposite ways from AP-1: when bound to ER-α, E2 activates transcription; whereas with ER-β, transcription is inhibited (9) . However, using transient transfection assays, all classes of antiestrogens bound to ER-β are potent transcriptional activators at an AP-1 site, acting as estrogen agonists rather than antagonists (9) . These observations raised the question of whether increased expression of ER-β might be associated with the development of tamoxifen resistance observed in many breast cancer patients. To address this issue, we have quantitatively determined ER-β mRNA expression in breast tumors from tamoxifen-sensitive and -resistant patients, as well as tamoxifen-sensitive breast cancer cell line MCF-7 and its tamoxifen resistant clone, clone 9 (10) .

Materials and Methods

Tumor Samples.

We identified snap-frozen primary breast tumor biopsies from two age-matched groups of postmenopausal breast cancer patients operated on between 1997 and 1998. The first group comprised those patients who showed a clinical response to tamoxifen, i.e., no tumor recurrence (n = 8); the second group consisted of biopsies from age-matched patients who subsequently developed tamoxifen resistance (defined as disease recurrence while receiving tamoxifen; n = 9). All patients were alive except for two patients from the tamoxifen-resistant group. Clinical details are presented in Table 1 ⇓ . Ethical approval was obtained, and all patients gave informed consent.

Cell Culture.

The tamoxifen-sensitive human breast cancer cell line MCF-7 was obtained from frozen stocks within our laboratory. Clone 9, a tamoxifen-resistant cell line isolated by cDNA transfection of tamoxifen-sensitive MCF-7 cells (10) , was a gift from Prof. A. L. Harris (Imperial Cancer Research Fund Laboratories, University of Oxford, Oxford, United Kingdom). This cell line is tamoxifen-resistant both in vitro and in vivo when grown as xenografts in nude mice (10) . Both cell lines were maintained in 25-cm² tissues culture flasks in DMEM supplemented with 2 mm glutamine, 100 μg/ml streptomycin, 100 units/ml penicillin, and 10% heat-inactivated fetal bovine serum (all Life Technologies, Inc., Paisley, United Kingdom). Cultures were passaged at least once per week by routine trypsinization.

RNA Extraction and cDNA Synthesis.

Frozen breast tissue was pulverized using a mortar and pestle, and total RNA was extracted with Trizol (Life Technologies, Inc.) according to the manufacturer’s instructions. Cell lines were lysed in situ by adding Trizol directly to the culture vessel. One μg of RNA was used as a template for first strand synthesis as described previously (11) .

RT-PCR Amplification.

Primers to detect fragments of the ER-β gene were designed from a published human sequence (6) and spanned exons 4–6. The sequences were 5′-GGCCGACAAGGAGTTGGTA-3′ (nucleotides 762–780) and 5′-AAACCTTGAAGTAGTTGCCAGGAGC-3′ (nucleotides 995-1020), giving an amplified product of 257 bp. The PCR reaction contained 2 units of BIOTAQ; 10× PCR buffer (containing 1.5 mm MgCl₂; both from Bioline, London, United Kingdom); 0.5 μg of each oligonucleotide primer; 200 μm each of dATP, dCTP, dGTP and dTTP; and 2 μl of nascent cDNA and sterile distilled water to bring the volume to 50 μl. ER-β transcripts were analyzed in parallel on at least two separate occasions in a thermal cycler (Hybaid OmniGene, Teddington, United Kingdom) with the following cycle: a denaturation step of 94°C for 2 min; followed by 30 cycles of 94°C for 30 s, 65°C for 30 s, and 72°C for 30 s and concluded with a final primer extension step of 72°C for 5 min. Primers and reaction conditions for TGF-β1 have been described previously (11) , except that the reactions described in this study were performed over 30 amplification cycles. Positive controls comprised human testis cDNA (ER-β) and human breast fibroblast cDNA (TGF-β1), respectively. Negative controls included substitution of RNA or cDNA with distilled water. These were consistently negative. PCR products (20 μl) were resolved on a 1.2% agarose gel in Tris-borate-EDTA buffer and visualized by ethidium bromide staining under UV illumination. To confirm cDNA integrity, fragments of the housekeeping gene GAPDH were amplified concurrently.

Confirmation of Product Identity.

Restriction digests were performed on representative ER-β PCR products to confirm their identity. Five μl of amplified product were digested overnight at 37°C with either SacI or EcoRI restriction endonucleases (Life Technologies, Inc.), yielding products of 195 and 62 bp or 212 and 45 bp, respectively. For TGF-β1, digestion with BamHI yielded fragments of 241 and 95 bp (Ref. 9 ; data not shown). Digested products were electrophoresed through a 2% agarose gel and visualized as described above.

PCR Quantification and Statistical Analysis.

Reference standards of GAPDH PCR products were prepared as described previously (12) and electrophoresed alongside ER-β and TGF-β1 PCR products. Using UVIband software (UVItec, Ltd., Cambridge, United Kingdom), an arbitrary value of 100 was ascribed to the GAPDH reference standard, and the fluorescence intensity of ER-β and TGF-β1 was expressed as a percentage of this standard. Each analysis was performed independently by two observers (V. S. and D. S. W.) and on two separate occasions after which a mean value was obtained. The two-tailed Mann-Whitney test was used to test the difference between groups. Results were considered to be significant at a probability level (P) of ≤ 0.05. Statistical analysis was performed using the Arcus software package for Windows (Research Solutions, Cambridge, United Kingdom).

Results

Quantitative Analysis of ER-β mRNA in Tamoxifen-sensitive and -resistant Breast Tumors.

A representative agarose gel showing PCR products for ER-β is presented in Fig. 1a ⇓ , and its identity was confirmed by restriction mapping (Fig. 1b) ⇓ . All tamoxifen-resistant breast tumors expressed transcripts for ER-β, as compared with the tamoxifen-sensitive group, in which seven of eight samples expressed this gene. Furthermore, the levels of ER-β mRNA appeared to be increased in the tamoxifen-resistant samples, despite apparently equivalent levels of expression of the constitutively expressed housekeeping gene GAPDH across both cohorts (Fig. 1a) ⇓ . To determine the relative expression of ER-β mRNA in individual samples, ER-β PCR products were quantified against a GAPDH reference standard (Fig. 2a) ⇓ . In the tamoxifen-resistant group, the median relative expression of ER-β mRNA was 33.64 (range, 21.97–53.87). This was significantly greater (P = 0.001) than that of the tamoxifen-sensitive breast tumors (median, 17.28; range, 0–31.91).

TGF-β1 Expression in Tamoxifen-sensitive and -resistant Breast Tumors.

To verify that our RT-PCR method was a reliable indicator of gene expression in the two patient groups, cDNA was amplified with oligonucleotide primers designed to detect TGF-β1. This factor was selected because previous studies have shown its overexpression in tamoxifen-resistant tumors (4) . In accord with these results, TGF-β1 mRNA was significantly overexpressed in the tamoxifen-resistant group (P = 0.02). This is illustrated in Fig. 2b ⇓ .

ER-β mRNA Expression in Tamoxifen-sensitive and -resistant MCF-7 Variants.

To test our hypothesis that increased expression of ER-β may be responsible for tamoxifen resistance, ER-β mRNA was quantified in the tamoxifen-sensitive cell line MCF-7 and its tamoxifen-resistant clone, clone 9. As shown in Fig. 3 ⇓ , ER-β mRNA was detected in both MCF-7 and clone 9. Both cell lines showed equivalent levels of expression of the constitutive GAPDH gene. However, the median relative expression of ER-β mRNA in clone 9 was 56.21 (range, 36.1–74.32). This was over 2-fold greater than that observed in MCF-7 (median, 24; range, 15.3–39.43; P = 0.05). However, as predicted, the highest expression of ER-β was observed in the testes control (P = 0.03 versus MCF-7).

Discussion

Despite the clear benefits of tamoxifen in extending the disease-free and overall survival of breast cancer patients, the development of antiestrogen resistance remains a considerable dilemma. Clinically, over 60% of all breast cancer patients with ER-α⁺ lesions will initially benefit from tamoxifen. However, most of these patients eventually relapse (1) , with acquisition of tamoxifen resistance typically observed after about 15 months (13) . Despite intense research, the factor(s) responsible for acquired resistance remains poorly understood. In the present study, we have evaluated and compared the expression of ER-β mRNA in primary breast tumors from tamoxifen-sensitive and -resistant breast cancer patients using RT-PCR.

When primary tumors from the tamoxifen-sensitive and -resistant groups were compared, ER-β mRNA was significantly overexpressed in the latter group. This was confirmed with in vitro studies in which ER-β mRNA expression in tamoxifen-sensitive and -resistant cell lines was measured. In wild-type MCF-7 cells, ER-β mRNA was detected; however, this was significantly up-regulated by approximately 2-fold in the tamoxifen-resistant clone, clone 9. This observation is in accord with a suggested role for ER-β as a poor prognosticator in breast cancer (7 , 8) and supports our hypothesis that ER-β may be involved in tamoxifen resistance.

Although we observed a clear up-regulation of ER-β mRNA in tamoxifen-resistant tumors and in the tamoxifen-resistant cell line clone 9, it is unlikely that ER-β is the universal resistance mediator. Indeed, as a control group to validate the mRNA quantification, we studied the relative expression of TGF-β1 and confirmed previous findings that this is also up-regulated in tamoxifen-resistant breast tumors (4) . Other members of the TGF-β family have been shown to have similar effects, with a reversal of tamoxifen resistance observed in vivo after treatment with antibodies against TGF-β2 (14) . Also, our previous work has shown that ER-β predominates in normal breast tissue, where its exclusive expression is common (8) . This is in contrast to breast tumors, where exclusive expression of ER-β occurs only sporadically, 4 with coexpression of both ER subtypes more frequently observed (8) . In the present study, only a single tumor from the tamoxifen-resistant group failed to coexpress ER-α and ER-β. This sample also expressed the lowest levels of ER-β. It is known that ratios of ER-α:ER-β alter with tumor progression, with the relative expression of wild-type ER-β inversely related to the tumor grade (15 , 16) . Therefore, depending on the relative expression of ER-α and ER-β during the evolution of a tumor, a differential response to antiestrogen therapy may thus be observed. This has particular relevance when the ER signaling mechanisms are considered. Both ER subtypes can form DNA-binding homo- and heterodimers (17, 18, 19) , resulting in signaling from either ER-α, ER-β, or ER-α/ER-β. Any alterations in the expression of ER subtypes within the same breast tumor may allow the interaction of ER-α and ER-β proteins, leading to differential responses to antiestrogen. This could be addressed in vitro by measuring the ER-α/ER-β profiles over time in breast cancer cell lines cultured continuously in the presence of tamoxifen. Additionally, the availability of specific receptors for ligand binding may influence the action of antiestrogens as antagonists or agonists, particularly in view of opposing downstream effects mediated by ligand bound receptor signaling via the classic estrogen response element, and from AP-1 (9) . This has been substantiated by transient transfection assays that reveal that antiestrogen-ER-β ligand receptor complexes are agonists of AP-1-driven reporter genes, whereas no transcriptional activity is observed when the same ligand-receptor complex signals from a conventional estrogen response element (20) . Also, the possibility of novel response elements that interact preferentially with antiestrogen-bound ER-β cannot be excluded. Further work is required to resolve this question.

The intracellular environment also deserves consideration. Perhaps those tumors that overexpress ER-β may also possess a different complement of transcription factors and coactivator/corepressor proteins, both necessary components in the transcriptional cascade. The ratios of these proteins may vary, depending on the composition of the particular ligand-receptor complex (i.e., agonist/antagonist, ER-α or ER-β; Ref. 21 ), with a shift in the balance of corepressor:coactivator proteins associated with antiestrogen resistance. Levels of the ER-regulated PR should also be considered, because it is generally accepted that expression of PR in human breast tumors is a good prognostic marker. Although PR data were not available for our entire study population, an inverse relationship was apparent between ER-β and PR in over 50% of the tamoxifen-resistant cohort, which was not evident in tumors from the tamoxifen-sensitive group (data not shown). This is consistent with observations that levels of ER-β mRNA were significantly reduced in those tumors that are ER-α⁺ and PR⁺ (7) .

One of the drawbacks of all types of PCR-based work is that there is no guarantee that levels of mRNA and protein are directly correlated. Clearly it would have been helpful to measure both protein and RNA in fragments of the same sample, but the amounts of tissue were limiting. Additionally, the current lack of a reliable commercial antibody specific for human ER-β precludes this type of study at the present time.

In summary, although our conclusions are based on a small number of tumors, we propose a novel mechanism of tamoxifen resistance, which may be related to overexpression of ER-β. Breast tumor biopsies are now routinely screened for ER-α. The additional measurement of ER-β in those biopsies at the time of presentation may be invaluable in predicting patient response to endocrine therapy.