We read with great interest the Review (The different roles of ER subtypes in cancer biology and therapy. Nature Rev. Cancer 11, 597–608 (2011))1 by Thomas and Gustafsson, in which the authors addressed the mechanisms of action and regulation of oestrogen receptors (ERs), the distinct biological effect of ER isoforms, and the roles of ERs in cancer prognosis and targeted therapies. Regarding the regulation of cellular levels of ERs, the authors systemically stated the potential mechanisms, including promoter methylation, post-transcriptional regulation by specific microRNAs and proteasome-mediated degradation of ER proteins. Although the aforementioned mechanisms are the main pathways of ER regulation, the complete mechanisms of ER upregulation are still not fully understood. One mechanism that may lead to the overexpression of a given gene in neoplastic cells is gene amplification, and ESR1 gene amplification has been observed in some cancer cells2. We believe that this issue, which is not mentioned in the Review, deserves to be discussed when illustrating the mechanisms of upregulation of ERα.

It has recently been reported that expression of ERα, but not expression of ERβ, is driven in some cases by ESR1 amplification2. Holst et al. used tissue microarray analysis of more than 2,000 samples to demonstrate that 20.6% of breast cancers contained ESR1 amplification2; Burkhardt et al. showed that amplification rates of ESR1 in ductal carcinoma in situ (DCIS), DCIS with invasive cancer and the invasive component did not differ significantly from one another, with rates of 19.0%, 24.1% and 21.3%, respectively3. More interestingly, almost all tumours with ESR1 amplification showed ERα protein overexpression2. It is therefore not surprising that ESR1 amplification in breast cancers was found to be associated with improved survival in women who had received adjuvant tamoxifen. In patients with ER-positive disease who received tamoxifen monotherapy, survival was indeed shown to be longer in patients who had ESR1 amplification than in patients who did not have ESR1 amplification2. Another Japanese report also showed that ESR1 amplification, which was found in 22.6% of samples by three-dimensional fluorescence in situ hybridization (FISH) assay, strongly correlates with higher expression levels of ERα, and that patients with ESR1 amplification in tumours apparently experience longer disease-free survival than those without ESR1 amplification4. These observations consistently suggest that ESR1 amplification is helpful in selecting patients who may potentially benefit from endocrine therapy. Besides breast cancer, ESR1 amplification occurs in more than 20% of endometrial carcinomas (established using FISH5) but it rarely occurs (2%) in ovarian cancer6.

However, the relatively high prevalence of ESR1 amplification in breast cancer (about 20%) has been challenged by several groups, who found ESR1 amplification in only approximately 1–5% of breast cancers7,8,9,10 (Table 1). One potential explanation for this is that independent investigators used a variety of different techniques11. The high prevalence of ESR1 amplification is mainly observed by FISH rather than by comparative genomic hybridization (CGH) or by quantitative-PCR. It is likely that contamination of tumour DNA with normal DNA (from the stroma, for example) is not only a challenge for detecting low-level amplicons in array CGH study, but is also a major drawback in quantitative-PCR. It is also difficult to accurately distinguish multiple small signals from a large confluent signal, given the small size of the ESR1 amplicon in breast cancer. The distance between the signals is often smaller than the diameter of one FISH signal. Such clusters are difficult to count, although the tumour appears to be amplified at first sight during a visual inspection. As a result, most ESR1-amplified tumours are considered as unamplified if ERBB2 criteria are applied. New criteria for estimating the ESR1 gene copy number need to enable a more reliable identification of amplified cancers than identification by classical counting.

Table 1 Available data for ESR1 (at locus 6q25.1) amplification in cancer

Because alteration of ER expression is an important step in the development and progression of hormone-related cancers, and because it influences cancer response to endocrine therapy1, ESR1 amplification leading to upregulation of ERα expression might have clinical importance, and further well-designed investigations are needed to resolve this issue.