Originally published as JCO Early Release 10.1200/JCO.2006.08.3485 on November 20 2006

This Article Full Text (PDF) Purchase Article View Shopping Cart Alert me when this article is cited Alert me if a correction is posted Services Email this article to a colleague Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Save to my personal folders Download to citation manager Rights & Permissions Citing Articles Citing Articles via Google Scholar Google Scholar Articles by Dowsett, M. Search for Related Content PubMed PubMed Citation Articles by Dowsett, M.

Estrogen Receptor: Methodology Matters

Mitch Dowsett

Royal Marsden Hospital, London, United Kingdom

The estrogen receptor (ER) has been our most important biomarker in breast cancer for more than 30 years, largely because of its role in indicating the potential of patients to benefit from endocrine therapy. The substantial survival benefit from tamoxifen in ER-positive but not ER-negative tumors,¹ and the extension of this benefit with aromatase inhibitors in postmenopausal patients, has made treatment with one of these well-tolerated endocrine agents mandatory in ER-positive disease; therefore, ER measurement has been rendered essential for all invasive primary breast carcinomas. ER-positive tumors have a better prognosis in the early years after diagnosis, which is at least partly due to their slower proliferation, although it has become clear that recurrence-free survival curves come together after about 15 years,¹ which suggests that ER status—surprisingly—may have little significance for the establishment of micrometastases before diagnosis.

The identification of tests that can be implemented widely with reproducible accuracy to allow reliable identification of ER-positive patients is of major importance. Thus, the article by Cheang et al² in this issue, which reports that the use of a new antibody to ER for immunohistochemistry (IHC) has better performance than the antibody in widespread use, merits detailed assessment. The most important question for most readers will be whether the data are sufficient for the laboratories at their disposal to consider changing their established methodology. In thinking about this and related issues, it is instructive to consider the stepwise (and sometimes "jumpwise") changes in methodology that have led to ER status being nearly universally measured by IHC on fixed tissues. Data from the early studies remain some of the most solid in establishing the relationship between treatment benefit and ER status.

By the beginning of the 1970s, the ER had been identified and the theory had been established that breast cancers with a high content of the receptor protein would respond to endocrine treatment; however, there was little evidence to support this hypothesis.³ In 1974, overwhelming supporting evidence was presented at a meeting held by the Treatment Committee of the National Cancer Institute Breast Cancer Task Force, comprising many of the oncologists destined to become major authorities in breast cancer. Consistent data showing clinically important separation of responders from nonresponders according to the detection of ER in their primary tumors was presented by several groups and summarized by McGuire et al.⁴ The following response rates were observed for ER-positive versus ER-negative tumors: surgical therapies (eg, adrenalectomy), 55% v 8%; additive medical therapies (eg, androgens), 60% v 8%; and antiestrogens, 40% v 19%. In most countries, analyses of ER rapidly became an important component of a breast cancer patient's work-up, particularly if she was under consideration for surgical endocrine therapy.

From the 1970s to the early 1990s, assays depended on homogenization of a fresh-frozen portion of the tumor and preparation of cytosol by centrifugation for the performance of a ligand-binding assay involving incubation with radioactively labeled estradiol. Separation of receptor-bound estradiol from the unbound fraction was most frequently achieved with a suspension of dextran-coated charcoal (DCC). The DCC adsorbed the unbound estradiol and gave its name to the assay that became the most widely used service assay until the early 1990s. It is of passing interest to note that because this methodology required fresh tissue, it resulted indirectly in collections of large series of excess tumor tissue that have provided the resource for some of the recent seminal publications on expression microarrays, as well as for the article by Cheang et al.²

Improved sensitivity of the DCC assay created the issue (which remains important at present) of whether the very low levels of ER that were previously undetectable should now be considered positive. The Oxford overview analysis, which has been so important in creating practice-changing analyses of pooled clinical data, avoided describing tumors as ER negative, and rather described tumors with values less than 10 fmol/mg protein (the most frequently used positive/negative cutoff) as ER poor.¹ The presence of progesterone receptor in ER-negative tumors is associated with an enhanced chance of response⁵ but may mostly indicate a false-negative ER assay rather than a biologic quandary.

Interestingly, an artifact resulting from the DCC methodology resulted in a misunderstanding of the distribution of ER in cells. Until 1984, the receptor, when unbound from ligand, was believed to exist in the cytoplasm because any ligand-bound receptor was found in the nuclear fraction. However, the availability of monoclonal antibodies to ER revealed that it was an almost exclusively nuclear protein,⁶ and that the earlier results were due to the dissociation of unbound receptor from the nucleus during homogenization. The nuclear retention of ligand-bound receptor during homogenization seems to relate to the tight association of this complex onto estrogen response elements that are upstream of estrogen-responsive genes.

The same monoclonal antibodies led to the development of an enzyme immunoassay (EIA) as a more precise and less labor-intensive alternative to the DCC. Like the DCC, the EIA required the derivation of a cytosol and was quantitative; excellent linear correlations were reported between the two methodologies.⁷ In the mid-1990s, a second ER was discovered and named ER beta.⁸ Initial beliefs that it would be important to measure both ER alpha and ER beta for greater diagnostic accuracy have not been borne out; the latter shows relatively low expression in primary breast carcinomas.⁹ Reference to the excellent correlation between the DCC (which measured the sum of both receptors) and EIA (which was specific for ER alpha) suggested from the beginning that ER beta was unlikely to have a significant diagnostic role.

IHC application of monoclonal anti-ER antibodies was initially successful only on frozen sections, but the introduction of antigen retrieval methods and the creation of new antibodies allowed application on routinely fixed tissues. Since the early to mid-1990s this IHC approach has become widespread because of the ease of the assays, inexpensive reagents, use of minimal amounts of tissue, and applicability to routine histopathologic samples. Two antibodies have been particularly widely used during the last few years: 1D5 and 6F11. The authors of the report in this issue selected the former for comparison with a new antibody, SP1. The authors assert that 1D5 and 6F11 have been found to have similar sensitivities in clinical studies. One study (which they do not cite in this regard) does indeed indicate this similarity, with 592 samples showing 97.5% concordance between the two antibodies.¹⁰ However, one of the two studies they cite studied just 20 tumors¹¹ and the other did not assess the 6F11 antibody at all.¹² The experience of others is that 6F11 is a more sensitive antibody than 1D5. For example, the United Kingdom National External Quality Assessment Scheme found that 55% to 77% of centers using 6F11 had satisfactory performance compared with just 35% of centers using 1D5.¹³ Those using SP1 had between 70% and 74% satisfactory performance. Thus, any gain shown in the current report of SP1 over 1D5 may not extend to an advantage over 6F11, and the conclusion that SP1 represents a new standard for immunohistochemical ER assessment may be overstated.

The tumor samples selected for the comparison were surgical biopsies from more than 4,000 primary breast cancer patients presenting between 1986 and 1992. A study of this size obviously avoids some of the uncertainties of small studies, but any deficits in study design may also be more strongly embedded in the resultant data. The material was from frozen stores of residual tissue that was excess to the requirements of the DCC assay, results from which were therefore available for comparison. The data are therefore only directly validated for tumors of a size that allowed this storage of excess material, but there is no reason to suppose that the results may not be extrapolated to smaller tumors. The frozen tissue was fixed in formalin and tissue microarrays (TMAs) derived from them, presumably recently and for the purposes of this current study, although neither of these points is stated. The authors provide convincing data that the comparison between 1D5 and SP1 on TMAs does not create any biases compared with an evaluation based on sections—an important issue given that it is the latter that are used in diagnostic analyses. However, these same data remind one that false negatives can easily be obtained on TMAs for markers showing significant within-tumor heterogeneity; therefore, results from TMAs should not be considered as equivalent to diagnostic results unless demonstrated to be so. Similarly, the 96% agreement for ER that they found between the duplicate cores from a duplicate-redundancy TMA confirms general experience that ER is estimated with sufficient reliability for most research purposes on a single TMA. Our own experience is that this does not extend to progesterone receptor because of the different nature of the intratumoral heterogeneity. The significance of heterogeneity should be considered for all markers evaluated by TMA.

The use of old frozen tissues that may create freezing artifacts, for example, raises two potentially important issues. First, any such artifacts might differentially affect the epitope targeted by 1D5 rather than SP1. It should be noted, however, that the data presented in Table A1 (online only) on their earlier findings is consistent with the currently reported relative differences between the antibodies; this earlier work was performed on tissues that were fixed in a conventional fashion. Second, there is a larger than expected proportion of DCC-positive tumors that are IHC negative. For 1D5 this is 22% and for SP1 it is 14%; both antibodies showed lower positivity rates in this study than in the earlier study. The authors note that this might be due to the freezing before fixation. This creates bias in the comparison of the IHCs with the DCC assay, which showed a notably better prediction of survival than either antibody.

This comparison would have been better accomplished by using the antibodies on conventionally fixed pieces of the same tumors. This deficit is unfortunate because the relationship of data from the DCC assay with clinical response was evaluated and confirmed in numerous studies internationally; IHC assays have not been subject to the same degree of direct evaluation against response. Another frequently cited study of validation of IHC related to clinical outcome in the adjuvant setting also used unconventional materials (histologic sections from the frozen pellet resulting from centrifugation of tumor homogenates for the DCC), and like the current study, did not use tissues from a randomized trial of treatment compared with surgery alone to allow benefit from tamoxifen in the adjuvant setting to be distinguished from intrinsic prognosis.¹⁴

So, should laboratories switch to SP1 on the basis of this report? My own laboratory will not switch immediately, but the data are sufficient for us to seriously consider performing our own evaluation of it against the 6F11 antibody. Preferably, we will use material from a randomized trial of tamoxifen versus surgery alone.¹⁵

What of the long-term future of ER measurements? First, assays of the highest quality are likely to remain important for the foreseeable future, not only because of their predictive value but also in recognition of ER as the dominant biologic determinant of breast cancer¹⁶ in a world that will be investigating and applying an increasing number of molecular markers. Second, it is likely that we will begin to recognize the information that quantitative measures of ER hold and use assays that assess this as well. For many years laboratories have strived to achieve some sort of quantitative assessment of the amount of ER in tumors, but there has been very little use of that information in patient management. Although there has been evidence for a greater benefit from tamoxifen in tumors with higher levels of ER this has not helped choose whether to treat with endocrine therapy. There is however increasing interest in the development of prognostic scores for ER-positive patients to consider the added value of adjuvant chemotherapy in women already destined for endocrine therapy.^17,18 This is of increasing importance with the excellent prognosis of endocrine-treated patients diagnosed at an increasingly early stage. The quantitative expression of ER when measured by quantitative reverse transcriptase polymerase chain reaction, including formalin-fixed tissues, can make a major contribution to such evaluations.¹⁸ There is also a small amount of data from a neoadjuvant study that the relationship between ER level and response to tamoxifen may differ from the response to aromatase inhibitors.¹⁹ If confirmed in the adjuvant setting, this would make the demand for quantitative assays even greater. IHC assays are ill-suited for this purpose no matter which antibody they use.

We will need to judge these new molecular assays carefully: their assessment in archival tissue blocks from completed adjuvant trials of endocrine therapy will provide a more rigorous assessment than that applied to our currently used IHCs.

Author's Disclosures of Potential Conflicts of Interest

The author indicated no potential conflicts of interest.

NOTES

published online ahead of print at www.jco.org on November 20, 2006.

REFERENCES

1. Early Breast Cancer Trialists' Collaborative Group (EBCTCG): Effects of chemotherapy and hormonal therapy for early breast cancer on recurrence and 15-year survival: An overview of the randomised trials. Lancet 365:1687-1717, 2005[CrossRef][Medline]

2. Cheang MCU, Treaba DO, Speers CH et al: Immunohistochemical detection using the new rabbit monoclonal antibody SP1 of estrogen receptor in breast cancer is superior to mouse monoclonal antibody 1D5 in predicting survival. J Clin Oncol 24:5637-5644, 2006[Medline]

3. Fotherby K, James F: Biochemistry of steroids in subjects with cancer, in Stoll BA (ed): Endocrine Therapy in Malignant Disease. London, United Kingdom, Saunders, 1972, pp 25-51

4. McGuire WL, Carbone PP, Sears ME, et al: Estrogen receptors in human breast cancer: An overview, in McGuire WL, Carbone PP, Vollmer EP (eds). Estrogen Receptors in Human Breast Cancer. New York, NY, Raven Press, 1974, pp 1-8

5. McGuire WL, Horwitz KB, Pearson OH, et al: Current status of estrogen and progesterone receptors in breast cancer. Cancer 39:2934-2947, 1977[CrossRef][Medline]

6. King WJ, Greene GL: Monoclonal antibodies localize oestrogen receptor in the nuclei of target cells. Nature 307:745-747, 1984[CrossRef][Medline]

7. Thorpe SM: Monoclonal antibody technique for detection of estrogen receptors in human breast cancer: Greater sensitivity and more accurate classification of receptor status than the dextran-coated charcoal method. Cancer Res 47:6572-6575, 1987[Medline]

8. Kuiper GG, Enmark E, Pelto-Huikko M, et al: Cloning of a novel receptor expressed in rat prostate and ovary. Proc Natl Acad Sci U S A 93:5925-5930, 1996[Abstract/Free Full Text]

9. Balfe P, McCann A, McGoldrick A, et al: Estrogen receptor alpha and beta profiling in human breast cancer. Eur J Surg Oncol 30:469-474, 2004[CrossRef][Medline]

10. Kaplan PA, Frazier SR, Loy TS, et al: 1D5 and 6F11: An immunohistochemical comparison of two monoclonal antibodies for the evaluation of estrogen receptor status in primary breast carcinoma. Am J Clin Pathol 123:276-280, 2005[CrossRef][Medline]

11. Vassallo J, Pinto GA, Alvarenga JM, et al: Comparison of immunoexpression of 2 antibodies for estrogen receptors (1D5 and 6F11) in breast carcinomas using different antigen retrieval and detection methods. Appl Immunohistochem Mol Morphol 12:177-182, 2004[Medline]

12. Pertschuk LP, Feldman JG, Kim YD, et al: Estrogen receptor immunocytochemistry in paraffin embedded tissues with ER1D5 predicts breast cancer endocrine response more accurately than H222Sp gamma in frozen sections or cytosol-based ligand-binding assays. Cancer 77:2514-2519, 1996[CrossRef][Medline]

13. Barnett S: The breast hormonal receptor module. Immunocytochemistry 4:140-144, 2006

14. Harvey JM, Clark GM, Osborne CK, et al: Estrogen receptor status by immunohistochemistry is superior to the ligand-binding assay for predicting response to adjuvant endocrine therapy in breast cancer. J Clin Oncol 17:1474-1481, 1999[Abstract/Free Full Text]

15. Dowsett M, Houghton J, Iden C, et al: Benefit from adjuvant tamoxifen therapy in primary breast cancer patients according oestrogen receptor, progesterone receptor, EGF receptor and HER2 status. Ann Oncol 17:818-826, 2006[Abstract/Free Full Text]

16. Perou CM, Sorlie T, Eisen MB, et al: Molecular portraits of human breast tumours. Nature 406:747-752, 2000[CrossRef][Medline]

17. Van de Vijver MJ, He YD, van't Veer LJ, et al: A gene-expression signature as a predictor of survival in breast cancer. N Engl J Med 347:1999-2009, 2002[Abstract/Free Full Text]

18. Paik S, Tang G, Shak S, et al: Gene expression and benefit of chemotherapy in women with node-negative, estrogen receptor-positive breast cancer. J Clin Oncol 24:3726-3734, 2006[Abstract/Free Full Text]

19. Ellis MJ, Coop A, Singh B, et al: Letrozole is more effective neoadjuvant endocrine therapy than tamoxifen for ErbB-1- and/or ErbB-2-positive, estrogen receptor-positive primary breast cancer: Evidence from a phase III randomized trial. J Clin Oncol 19:3808-3816, 2001[Abstract/Free Full Text]

This Article Full Text (PDF) Purchase Article View Shopping Cart Alert me when this article is cited Alert me if a correction is posted Services Email this article to a colleague Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Save to my personal folders Download to citation manager Rights & Permissions Citing Articles Citing Articles via Google Scholar Google Scholar Articles by Dowsett, M. Search for Related Content PubMed PubMed Citation Articles by Dowsett, M.