A key discovery that sparked bench-to-bedside breakthroughs in the field of breast cancer was the recognition of the hormone-dependence of many breast cancers. Observations made by Cooper and Beatson correlated the size of breast tumors with the phases of the menstrual cycle and showed that ovariectomy caused tumor regression and improved prognosis 12. We now know that estrogen and its receptor, estrogen receptor-α (ERα), underlie these effects through the transcriptional regulation of genes involved in cell proliferation and differentiation. Understanding the mechanisms of estrogen and ERα action created the foundation for the design of therapies that interfere with estrogen signaling and block tumor growth. These include: reduction of endogenous estrogens via aromatase inhibitors (exemestane, anastrazole, letrozole) and/or ovariectomy; interference of ER-mediated transcriptional control via selective ER modulators (tamoxifen); and degradation of the receptor via selective ER downregulator compounds (fulvestrant). These approaches are generally successful at prolonging patient survival for those tumors expressing ERα and have less toxic side effects than chemotherapy. For instance, it was estimated that tamoxifen has saved the lives of 400,000 women since its introduction in the 1970s 3.

Has this achievement tempted us into complacency with regard to the extent of our understanding of ERα's role in the pathogenesis of breast cancer? Given the large, on-going research effort and number of publications devoted to estrogen in breast cancer (26,303 articles entered into PubMed as of June 2007), some individuals would say no. Yet, a recent report by Holst et al. 4 in Nature Genetics was the first to investigate whether a common mechanism of oncogene activation, gene amplification, occurred at the ERα gene locus during tumor progression. Their work is an important scientific contribution that expands upon prior studies demonstrating ESR1 gene amplification in breast cancer cell lines and in some advanced tumors 567.

The pattern of ERα expression in normal breast tissue compared with precancerous and cancerous lesions is strikingly different. In normal breast tissue, ERα expression is restricted to a small proportion of non-proliferating luminal epithelial cells, typically at low to intermediate levels 89. However, in more than half of premalignant lesions and carcinomas, this dissociation breaks down and the receptor is detected in proliferating cells, generally at higher levels 8. Additionally, there is a striking increase in the intracellular amount of ERα protein 10. A significant unknown in the field of breast cancer is what drives the change in ERα expression and distribution in breast lesions. Holst et al. 4 used fluorescence in situ hybridization (FISH) to probe a tissue microarray containing 2,222 invasive breast cancers and 295 normal, pre-malignant, and pre-invasive samples and found ESR1 gene amplification in 358 samples (21%) of the 1,739 invasive breast carcinomas with analyzable FISH data. Virtually all (99%) cases with amplification exhibited correspondingly high ERα protein levels as measured by immunohistochemistry. Characterization of the ESR1 amplicon at 6q25.1 by PCR-based methods found that it was relatively small and did not extend into any other genes. Furthermore, ESR2, which encodes a second ER, ERβ, was not amplified. Amplification of other known oncogenes (HER2/neu, MDM2, MYC, EGFR) was detected in invasive cancer samples, although these were found to be independent of ESR1 amplification. Interestingly, ESR1 amplification was observed in proliferative benign breast lesions (36.4% of papillomas and 8.3% of usual ductal hyperplasia) and carcinomas in situ (35% ductal and 33% lobular) in addition to more advanced tumors 4. While these studies require independent validation, the data provide evidence that amplification of ERα appears in early lesions and may contribute, in part, to the appearance of high levels of ERα in breast tumorigenesis.

Gene amplification alone, however, cannot explain all cases involving high ERα protein levels. Only 54% of cancers with high ERα expression also had gene amplification 4. The remaining 46% showed high ERα expression without gene amplification 4. This suggests that other mechanisms contribute to high ERα protein levels, such as altered regulation of ESR1 transcription, mRNA stability, or ERα protein turnover. For example, recent studies have demonstrated that disruption of caveolin-1 and micro-ribonucleic acid 206 can increase ERα levels 1112. How such upstream factors regulate the ERα gene and protein is not well understood and needs further attention.

A significant point raised by the finding of ESR1 amplification in early lesions is whether high levels of ERα expression are a cause or consequence of malignant transformation. Studies of HER2/neu provide a clear example where overexpression of an amplified gene product is oncogenic 13. Could this also be the case for ERα? ESR1 gene amplification was identified in several benign proliferative breast lesions, which increase a patient's risk of cancer 14. Studies have shown that high levels of ERα are present in benign epithelium of women with breast cancer compared to controls and there is an inability to downregulate the receptor in response to estrogen in these cases, supporting a potential role for ERα overexpression in breast cancer risk 1516. Transgenic mouse models also indicate that overexpression of ERα is sufficient for the development of ductal hyperplasia, lobular hyperplasia, and ductal carcinoma in situ 17.

How high ERα levels might contribute to tumorigenesis is less understood. The simplest explanation is that the presence of additional receptors supports a more robust response to estrogen. An alternative and intriguing possibility comes from analogy of studies conducted on ErbB2/Her2-neu. Proteomic analysis of ErbB2 protein interactions showed that elevated concentrations of ErbB2 lead to promiscuous interactions and promote activation of distinct signaling pathways 18. In this model, overexpression of the oncoprotein, resulting from amplification or other processes, could lead to an expansion of its regulatory role by permitting protein interactions that activate non-canonical signaling pathways. Similar findings have been reported for ERα in a breast cancer cell model system of ERα overexpression in which the mechanism of transactivation and target gene regulation differ when ERα protein levels are elevated 1920. These studies of ErbB2 and ERα overexpression raise the interesting scenario that perhaps ERα in normal breast epithelium is maintained at restrictive levels that are necessary to promote differentiation. When the ERα protein concentration increases during tumorigenesis, promiscuous interactions with coregulatory proteins or DNA could lead to the activation of proliferative signaling pathways, which, at normal levels of expression, would be too weak to occur. This scenario would predict that amplification or overexpression of ERα would be causally related to the high proliferative capacity of ER+ cells. This possibility remains to be tested.

Classification of tumors into subtypes helps predict therapeutic responses and patient survival. Categorizing breast tumors as either ERα positive or negative by immunohistochemistry has proved clinically useful in determining which patients would benefit from endocrine therapy. More recently, microarray analysis has further refined the groupings of breast tumors on the basis of distinct gene expression profiles: basal-like, HER2+/ER-, normal breast-like, luminal A, and luminal B 21. The latter has clearly shown that the ER positive cohort is not a single group of patients. Both luminal A and B subtypes are ER+; however, patients with luminal B tumors have poorer outcomes. The ER+ cohort can also be subdivided into IE and IIE subtypes. The group IIE tumors are similar to subtype B and express more proliferative genes 22. The same proliferative gene signature was shown by Dai et al. 23 to be a marker of poor outcome in patients with tumors expressing high levels of ERα for their age. Although it is currently standard practice to offer hormonal therapy to all patients categorized as ER+, these and other studies demonstrate marked heterogeneity within this group in terms of gene expression profiles and patient survival.

Holst et al. 4 analyzed the clinical utility of classifying tumors based on ESR1 amplification. Phenotypes associated with ESR1 amplification included low tumor grade and lack of lymph node metastases, both positive prognostic indicators. Furthermore, tumors with ESR1 amplification were associated with longer survival in patients treated with adjuvant tamoxifen compared with non-ESR1 amplified and ER-negative tumors. However, there was no statistically significant difference in survival for patients with cancers having ESR1 amplification compared to patients with non-ESR1 amplified cancers containing the highest level of ERα protein (P = 0.09). Thus, the classification of tumors based on ESR1 amplification does not yield more clinical information than does the current method of tumor characterization based on ERα protein levels.

While all breast cancers are analyzed for the expression of ERα, steroid receptor status is not routinely measured for benign breast lesions. Depending on the level of suspicion, biopsy-proven benign lesions can either be surgically excised or followed with imaging. One histological group whose management is currently under debate comprises benign papillary lesions, which includes papilloma 24. Holst et al. 4 showed that ESR1 amplification occurs in 8 of 22 (36%) benign papilloma samples. Furthermore, elevated ERα protein levels have been demonstrated for papillomas and are associated with increased proliferation 25. Measurement of ESR1 gene amplification or ERα protein levels for papillary lesions may be potentially useful since the presence of amplification or overexpression would argue in favor of surgical excision instead of follow-up imaging.

Over 100 years have passed since the discovery of the importance of estrogen and, later, ERα to the growth of breast tumors. Since that time, tremendous advances have been made in our understanding of the molecular mechanisms of ERα activity and in the application of this knowledge to the development of therapies for the prevention and treatment of breast cancer. The recent discovery of ERα amplification in early breast lesions by Holst et al. is an important reminder that, despite our perception that we understand how ERα contributes to pathogenesis, there are still major questions that remain unanswered and breakthroughs to be made. Major clinical dilemmas still revolve around how better to predict response to hormonal therapy and how to fight endocrine resistance. Thus, in 2007, the question, "How does ER contribute to breast cancer?" remains one worth asking.

ER = estrogen receptor; FISH = fluorescence in situ hybridization.

The authors declare that they have no competing interests.