Article Text


ER-α36, a novel isoform of ER-α66, is commonly over-expressed in apocrine and adenoid cystic carcinomas of the breast
  1. Semir Vranic1,2,
  2. Zoran Gatalica1,
  3. Hao Deng3,
  4. Snjezana Frkovic-Grazio4,
  5. Lisa M J Lee1,5,
  6. Olga Gurjeva6,
  7. Zhao-Yi Wang7
  1. 1Department of Pathology, Creighton University Medical Center, Omaha, Nebraska, USA
  2. 2Department of Pathology, Clinical Center of the University of Sarajevo, Sarajevo, Bosnia and Herzegovina
  3. 3Department of Pathology and Pathophysiology, School of Medicine, Jianghan University, Wuhan, China
  4. 4Department of Pathology, Institute of Oncology, Ljubljana, Slovenia
  5. 5Division of Anatomy, College of Medicine, The Ohio State University, Columbus, Ohio, USA
  6. 6Emergency Cardiology Department, Ukrainian Strazhesko Institute of Cardiology, Kiev, Ukraine
  7. 7Department of Medical Microbiology and Immunology, Creighton University Medical School, Omaha, Nebraska, USA
  1. Correspondence to Semir Vranic, Department of Pathology, Clinical Center of the University of Sarajevo, BA-71000 Sarajevo, Bosnia and Herzegovina; semir.vranic{at}


Background ER-α36 is a novel 36 kDa isoform of the full-length oestrogen receptor alpha (ER-α66). ER-α36 primarily localises to the cytoplasm and the plasma membrane, and responds to membrane-initiated oestrogen and antioestrogen signalling pathways.

Aim To examine the expression of ER-α36 in apocrine and adenoid cystic carcinoma of the breast, both of which are consistently ER-α66 negative and currently lack effective targeted therapeutic options.

Methods 19 pure apocrine carcinomas (17 invasive and two in-situ carcinomas) and 11 adenoid cystic carcinomas of the breast were evaluated for ER-α36 expression, along with expressions of ER-α66, progesterone receptor (PR) and androgen receptor (AR) using immunohistochemical methods.

Results All pure apocrine carcinomas showed a characteristic steroid receptor expression profile (ER-α66 and PR negative, AR strongly positive). ER-α36 expression was detected in 18/19 pure apocrine carcinomas (94.7%, 95% CI 75.1 to 98.7) in predominantly membranous and cytoplasmic distribution. When positive, pure apocrine carcinomas uniformly (100% of cells) expressed ER-α36. All adenoid cystic carcinomas were uniformly negative for all three classic steroid receptors, but ER-α36 was detected in 8/11 cases (72.7%, 95% CI 42.8 to 90) with the similar sub-cellular pattern of expression as in the pure apocrine carcinomas. When positive, adenoid cystic carcinomas expressed ER-α36 in the majority of cells (average 76%).

Conclusion ER-α36, a novel isoform of ER-α66, is frequently over-expressed in apocrine and adenoid cystic carcinomas of the breast. These results indicate a potential for a novel targeted treatment in these cancers.

  • Breast cancer
  • ER-α36
  • adenoid cystic carcinoma
  • apocrine carcinoma
  • breast
  • breast pathology
  • immunohistochemistry
  • steroid receptors

Statistics from


Oestrogen receptor alpha (ER-α) expression in breast cancer is the most important predictor of antioestrogen therapy.1 Approximately 75% of all breast carcinomas are ER-α positive, which is a good predictor of the treatment response.2–4 Current immunohistochemical methods use antibodies against the full-length receptor protein (ER-α66, see figure 1), but there are at least three different isoforms of ER-α66 described in humans. Two isoforms have been detected in breast cancers.5 6 These are ER-α46 and the novel isoform, ER-α36 which we recently identified and cloned.7 8 They are generated through multiple promoter usage or alternative splicing.5 Although most steroid hormone receptors primarily localise in the nuclei, additional oestrogen receptors have been reported to exist in the cytoplasm and on the plasma membrane.3

Figure 1

Domain structures of oestrogen receptor (ER)-α66 and ER-α36. Arrows indicate the specific antibodies and their corresponding epitopes.

The novel 36 kDa isoform ER-α36 lacks the transcriptional activation domains found in ER-α66. The DNA-binding domain, dimerisation, partial ligand-binding domains as well as three myristoylation sites near the N-terminus are retained in ER-α36.7 8 While ER-α66 is predominantly detected in the cell nucleus, ER-α36 mainly localises to the cytoplasm and on the plasma membrane.7 8 ER-α36 has been shown to transduce the membrane-initiated steroid signalling (MISS) cascade, and function as a dominant-negative effector of oestrogen-dependent and independent transactivation mediated by ER-α66.7 We have recently demonstrated the expression of ER-α36 in both ER-α66 positive (MCF-7) and negative breast cancer cell lines (MDA-MB-231), as well as in a proportion of invasive breast carcinomas of no special type.8 9

In this study, we focused on adenoid cystic carcinoma (ACC) and pure apocrine carcinomas (PAC), two special types of breast carcinomas that are characteristically negative for ER-α66,10–12 and showed that the majority of them express ER-α36 protein.

Materials and methods

Breast tissue samples

Formalin-fixed, paraffin-embedded tissues (11 adenoid cystic carcinomas and 19 apocrine carcinomas) were selected from the files of the Departments of Pathology at Creighton University Medical Center (Omaha, Nebraska, USA), the Institute of Oncology Ljubljana (Slovenia), Thomas Jefferson University Hospital (Philadelphia, Pennsylvania, USA) and Kansas University Medical Center (Kansas City, Kansas, USA), after the approval of the Institutional Review Board of Creighton University School of Medicine. None of the 30 patients included in this study had received neoadjuvant therapy.


Immunohistochemical assays for ER-α66 (clone 6F11, Ventana Medical Systems, Tucson, Arizona, USA), progesterone receptor (PR; clone 16, Ventana Medical Systems), androgen receptor (AR; clone AR441, DakoCytomation, Carpinteria, California, USA), epidermal growth factor receptor (EGFR; (DAKO EGFR PharmDX diagnostic kit; DakoCytomation) and Her-2/neu (Clone CB11, Ventana Medical Systems) expression were previously performed as a part of clinical evaluation. The expression of ER-α66, PR and AR (the percentage of cells with the nuclear staining), and EGFR and Her-2/neu (the extent and the intensity of the membranous staining) were measured on an automated cellular imaging system (ACIS, ChromaVision Medical Systems, San Juan Capistrano, California, USA). The results of ER-α66, PR, AR, EGFR and Her-2/neu expression for both cohorts have recently been reported.11 12

Antibody used for the detection of ER-α36 was custom-made by Alpha Diagnostic International (San Antonio, Texas, USA); it was raised against the 20 unique amino acids at the C-terminal of ER-α36.8 9 Immunohistochemical staining with this ER-α36 specific antibody was performed using the rabbit ImmunoCruz Staining System (Santa Cruz, sc-2051, California, USA) according to the manufacturer's instructions. Briefly, slides with 5 μm thick formalin-fixed paraffin-embedded tissue sections were deparaffinised and hydrated in sequential treatment of xylene, ethanol and water. Citrate buffer (0.01 M citric acid, pH 8.0) was used to retrieve epitopes in a pressure cooker. Endogenous peroxidase activity was quenched with 3% H2O2 before anti-ER-α36 (at 1:100 dilution) antibody was applied. Biotinylated secondary antibody and streptavidin-horseradish peroxidase were added subsequently, and 3,3′-diaminobenzidine tetrahydrochloride was used as substrate for chromogenic visualisation before counterstaining with haematoxylin. The extent and cellular distribution of staining was evaluated by two investigators on a double-headed microscope. The subcellular localisation and percentage of positive cells was recorded.

For ERα-36 expression, only the membranous/cytoplasmic pattern of staining was considered specific.8 9 Weak (1+) intensity was defined as faint and incomplete membrane positivity. Moderate (2+) intensity and strong (3+) staining were both varying degrees of circumferential staining of membranes and the cytoplasm. The tumour was considered positive if a proportion of stained cells exceeded 1% at any intensity.9

Statistical analysis

To access the differences between proportions, two-sided Fisher's exact tests were used; α was set at a level of 0.05. The kappa statistic was applied to determine the consistency among the tests and 95% CIs are given around the estimates. Statistical analyses were carried out using SPSS V.17.0.


Patient characteristics

The mean age was 62 years (range 53–71 years) for patients diagnosed with ACC and 74 years for PAC (range 63–90 years).

Special breast cancer subtype characterisation

All PACs (including two in-situ carcinomas) were characteristically ER-α66 negative, PR negative and AR positive.11 Her-2/neu over-expression (3+) was observed in 10/19 cases (52.6%, 95% CI 31.53 to 72.80); EGFR expression was observed in 17/19 (89.5%, 95% CI 68.30 to 96.79). Of the two apocrine carcinomas in-situ, one over-expressed Her-2/neu (score 3+) and the other was positive for EGFR (score 3+).

All ACC samples exhibited a characteristic triple-negative breast cancer profile (ER-α66 negative, PR negative, Her-2/neu negative). EGFR was detected in 7/11 (63.6%, 95% CI 34.89 to 84.83) ACCs, as reported recently.12

Table 1 summarises the immunohistochemical characteristics of these two cancer types.

Table 1

Immunohistochemical characteristics of 30 breast samples

ER-α36 protein expression in PAC and ACC

ER-α36 protein was detected in 18/19 PACs (94.7%, 95% CI 75.13 to 98.77) and in 8/11 cases of ACC (72.7%, 95% CI 42.81 to 90.08). Moderate to strong membranous and cytoplasmic expression (figure 2C,D) was seen in both cancer types. The average percentage of ER-α36 positive cells was 100% for PAC and 76.2% for ACC. Nuclear staining was absent in all but one case of PAC, which exhibited both nuclear and cytoplasmic/membrane distribution of the ER-α36.

Figure 2

Representative H&E slides of: (A) a case of adenoid cystic carcinoma of the breast; (B) a pure apocrine carcinoma of the breast; (C) immunohistochemical staining of the tissue samples for ER-α36, demonstrating strong membrane and cytoplasmic staining in an adenoid cystic carcinoma; (D) a pure apocrine carcinoma of the breast.

We also analysed our cases for the co-expression of ER-α36 with EGFR and Her-2/neu proteins because of reported interactions between membrane steroid receptors and growth factors with tyrosine kinase activity.13 Co-expression of ER-α36 and EGFR was observed in 16 cases of PAC and 7 cases of ACC. Although co-expression of ER-α36 and EGFR in both types of breast carcinomas was statistically significant (p=0.018), the measure of agreement was only marginally convincing (κ=0.52, 95% CI 0.12 to 0.93) (table 2). Co-expression of ER-α36 and Her-2/neu however was not found to have a significant relationship (p=1.00).

Table 2

Relationship between oestrogen receptor (ER)-α36 and epidermal growth factor receptor (EGFR) expression

Adjacent normal breast tissue (ductal epithelium) was negative or showed reduced expression of ER-α36 in comparison to expression in the malignant epithelium.


ER-α66 expression is routinely evaluated by immunohistochemistry in all breast cancers since it has been shown to play a pivotal role2–4 in patient treatment and outcome. Approximately 75% of all breast carcinomas are regarded as ER-α66 positive, and are thus amenable to targeted therapy with antioestrogens.2–4

The ER-α gene (ESR1) is composed of six functional domains encoded by eight exons that commonly produce a 66.2 kDa protein (ER-α66).5 Numerous ER-α mRNA transcript isoforms of the gene have been described,6 and are generated through alternative splicing or promoter usage.5 Several isoforms of the ER-α protein have been isolated in humans, including ER-α36, recently identified and cloned by our group.7–9 This novel ER-α isoform is a product of a transcript initiated from a previously unidentified promoter in the first intron of the ER66 gene and is identical to the ER-α66 encoded by exons 2–6 of the ER66 gene, resulting in a smaller, 36 kDa protein (ER-α36).7 8 In the experimental cell models of breast cancer, it served as a membrane-based oestrogen receptor mediating MISS and as a dominant-negative modulator of ER-α66 mediated transcription activity.7 We have previously reported that ER-α36 is frequently detected in invasive breast carcinomas of no special type,9 and that it may also play a role in prognosis.14

Our current study showed that ER-α36 was also frequently expressed in pure apocrine carcinoma and adenoid cystic carcinomas, which are characteristically negative for the full length ER-α66.10–12

ER-α36 was over-expressed in the neoplastic cells in comparison to adjacent normal ductal epithelia, emphasising its role in tumour progression. It was almost exclusively expressed on the plasma membrane and in the cytoplasm; nuclear staining was observed in only one case of PAC. These results are in accordance with our recently published experimental and clinical investigations in the breast cancer cell lines and human breast cancers, where ER-α36 expression was localised primarily to the plasma membrane (50%) and the cytoplasm (40%), with little or no nuclear staining (up to 10%).7–9 Subcellular localisation of ER-α36 is in sharp contrast to the nuclear localisation of the full-length ER-α66. Nuclear expression of ER-α66 is considered a diagnostic hallmark for ‘ER-positive’ tumours, although functionally active extranuclear pools of ER-α66 exist.4 15 Cytoplasmic ER-α66 expression can be observed in breast cancer cells after long-term treatment with tamoxifen, coinciding with resistance to the drug.16 Shi et al indicated the importance of ER-α36 in development of endocrine resistance in a subgroup of invasive breast carcinomas that exhibit co-expression of ER-α66 and ER-α36.14

Identification of the ER-α36 isoform in PAC explains reported discrepancies between mRNA detection and negative ER protein expression. Bratthauer et al reported that ER mRNA was detectable in all cases of apocrine carcinomas despite a complete absence of ER protein using the commercially available immunohistochemical assay against ER-α66.17 The authors used RT-PCR amplification of ER mRNA using primers that covered the first and second exon of the ER-α66 messenger RNA, and concluded that the immunohistochemical absence of ER was not a consequence of an abnormal transcript but did not elaborate. We hypothesise that detectable mRNA in apocrine carcinomas might well be one of the alternatively spliced isoforms of ER-α including currently studied ER-α36. ER-α36 protein could not have been detected due to the lack of antibody specificity; the ER-α66 antibody used in their study (clone CC4-5, Novocastra Laboratories) recognises the N-terminal of the ER-α66 which is not shared between the full-length and ER-α36 isoform (see figure 1).18 Other commonly used commercial antibodies (eg, 1D5, SP1, RB-9016) recognise the epitopes located on either the N- or C-terminal of the ER-α gene which are not represented in either ER-α36 or ER-α46 isoforms.5

Functional importance of ER-α36 is reportedly related to the non-genomic (non-classic) ER activities, among which activation of the mitogen-activated protein kinase (MAPK/ERK) signalling pathway plays a major role.8 The MAPK/ERK signalling pathway is activated in response to both oestrogens (eg, 17β-oestradiol) and antioestrogens (eg, tamoxifen) which might be of particular importance for ER-α66 negative breast carcinomas since this subgroup might still respond to antioestrogen based therapy.8 Previous experiments demonstrated that the antioestrogens induce a stronger and a more prolonged activation of the MAPK/ERK signalling pathway than the oestrogens.8

Our study also revealed a significant relationship between ER-α36 and EGFR expression, indicating a close interaction of the two membrane-initiated signalling.13 19 Although Her-2/neu can also be actively involved in membrane-initiated steroid signalling, we could not identify a significant correlation between ER-α36 and Her-2/neu in our study.

In summary, membranous ER-α36 is over-expressed in breast tumours showing a complete absence of the full-length ER-α66 protein, indicating a potential for targeting the non-genomic growth signalling.

Take-home messages

  • Adenoid cystic carcinoma and pure apocrine carcinoma are two rare, special types of breast cancer characteristically negative for oestrogen receptor-α66 (ER-α66).

  • The novel 36 kDa isoform of ER-α66 (ER-α36) lacks the transcriptional activation domains of ER-α66 and predominantly localises to the cytoplasm and on the cell membrane. It mediates non-classic (non-genomic) oestrogen signalling.

  • ER-α36 is commonly over-expressed in adenoid cystic carcinoma and pure apocrine carcinoma of the breast, which may indicate a potential for novel treatment strategies in these cancers.


We thank Drs Ossama Tawfik, Juan Palazzo, Janez Lamovec and Eduardo Eyzaguirre for contributing cases for the study and Dr Fadila Serdarević for statistical expertise. We also acknowledge Mr Mirza Bašić for technical assistance.


View Abstract


  • Preliminary results were in part presented at the 97th Annual Meeting of the United States and Canadian Academy of Pathology, Denver, Colorado, USA in March 2008, and the 31st Annual San Antonio Breast Cancer Symposium, in December 2008.

  • Funding SV was a research fellow at Creighton University Medical Center, Omaha, Nebraska, USA, and had been supported by a UICC American Cancer Society Beginning Investigators Fellowship (ACSBI) (ACS/08/004) funded by the American Cancer Society.

  • Competing interests None declared.

  • Ethics approval This study was conducted with the approval of the Institutional Review Board of Creighton University School of Medicine.

  • Provenance and peer review Not commissioned; externally peer reviewed.

Request permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.