Breast Cancer Research and Treatment 16: 273-278, 1990. © 1990 Kluwer Academic Publishers. Printed in the Netherlands.
Report
Nuclear binding of the estrogen receptor: A potential predictor for hormone response in metastatic breast cancer
Nicholas J. Robert, Leslie Martin 1, C. Douglas Taylor 1, James Popkin ~, David R. Parkinson 2, Charles White ~, Constantine A. Pappas ~, Hisashi Tamura 1, Rebecca Gelman 3, Joseph Cohen ~ and Shanthi Raam 1 Division of Hematology/Oncology, New England Medical Center, Boston, MA, USA; i Tufts University School of Medicine, Boston, MA, USA; 2Biological Therapy Unit, M.D. Anderson Cancer Center, Houston, TX, USA; 3Division of Biostatistics, Dana Farber Cancer Institute, Boston, MA, USA
Key words: estrogen receptor, hormone therapy, immunohistochemistry, nuclear binding Abstract
We have previously described an in vitro immunohistochemical test employing anti-receptor antibodies, for demonstrating the nuclear binding characteristics of estrogen receptors (ER) in breast carcinomas. Based on a retrospective analysis of twenty-five patients with estrogen receptor-positive ( E R + ) breast cancer who were treated with hormone therapy and whose clinical responses were evaluable, we were able to demonstrate that this test may be valuable to predict which, among the E R + tumors (whether or not they are progesterone receptor positive, P R + ) , are likely to respond to hormone therapy and which may fail. While tumors in which ER exhibited abnormalities in nuclear binding behavior (ligand-independent nuclear binding or no nuclear binding) failed hormone therapy (16 out of 19 patients), those in which nuclear binding of ER appeared normal (ligand-dependent) in the in vitro test, responded to hormone therapy (5/6 patients). While our previous report dealt with the procedural details, specificity of the reagents, and the design of the study, this report addresses the clinical aspects of this study and response correlation.
Introduction
Endocrine therapy elicits variable responses in patients with metastatic breast cancer. Although many factors appear to play a role in predicting responses, the most notable parameter has been the E R + status of the tumor [1]; in this group 30-70% of the patients will respond to hormone therapy. Two major functions have been associated with estrogen-receptors: binding of the hormone, and binding to the specific sites in the target cell nucleus as a receptor-hormone complex to ini-
tiate other estrogen-dependent molecular events such as progesterone receptor synthesis. With the availability of anti-receptor antibodies which recognize hormone-free as well as nuclear ER, and an in vitro test utilizing these antibodies [2-5], it has been possible to study the nuclear binding characteristic of ER in cryosections from tumors which were stored frozen for a long period of time (--- 10 years). In these tumors, the ER is still well preserved and is able to bind the hormone with high affinity. With the help of this procedure we demon-
Address for offprints: N.J. Robert, Division of Hematology/Oncology, New England Medical Center, 750 Washington Street, Boston,
MA 02111, USA
274
N J Robert et al.
strated three subclasses based on the in vitro nuclear binding characteristics of ER: A) Tumors in which E R exhibits normal liganddependent nuclear binding requiring both the presence of estrogen and exposure in vitro to temperature of 37°C for thirty minutes; B) Tumors in which E R binds to the nuclei in the absence of estrogen, but requires exposure to 37°C for thirty minutes thus exhibiting abnormal ligand-independent nuclear binding; C) Tumors in which ER is unable to bind to the nucleus. In all these three subclasses the hormone binding characteristics of the cytoplasmic E R was unimpaired and was typical of the pattern described for Type I E R in normal target cells [2-5]. Twenty-five patients out of 88 E R + tumors classified as such were evaluable for clinical response, and the results
of the in vitro nuclear binding test were correlated with clinical response to hormone therapy.
Material and methods
The test procedure, specificity of the results, criteria for classification of the E R + tumors on the basis of nuclear binding parameters, and the reproducibility of the classification procedure, have been described in detail previously [2-5] and therefore will not be included in this report. Briefly, the cryosections of the tumors were exposed either to ligand-free saline or saline with an estrogenic ligand (estrogen or DES) at 37° C for thirty minutes in a moist chamber. A progestin was used in experimental controls. Cryosections were also directly processed without prior incubation. After the in-
Table 1. Clinical data Patient Age (yrs)
ER/PR
Nuclear binding DFI (months)
Metastatic Site(s)
Therapy
Response
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25
1418 +/NK 12/9/NK 52/216 106/119/280/63/491 19/NK 695+/NK 230/89/104 20/99 27/6 37/NK 415/249 231/8/NK 727/37/NK 68/5 93/67 363/81 157/363 38/NK
+ (NL) +(L) + (NL) +(NL) + (NL) +(L) + (L) + (NL) +(L) +(NL) +(NL) + (NL) +(NL) +(L) +(L) + (NL) -
Lymph node Skin Lymph node, Adrenal Breast, Skin Bone Bone Bone Skin, Bone Skin Lung Lymph node, Lung Bone Bone Skin Skin Lymph node Bone, Lung Bone Multiple Bone Bone Bone Bone Ascites Bone, Lung
Tam Tam Tam Tam Tam Tam Aminoglutethemide Megace Oophorectomy Tam Tam Tam Tam Tam Tam Tam Tam Tam Tam Tam Tam Tam DES Tam Adrenalectomy
PD PD PD R PD NR R PD PD PD R PD R PD R NR NR R PD PD PD R R PD PD
65 61 55 55 58 53 58 65 51 75 64 62 71 62 75 47 78 64 66 71 46 60 63 67 73
0 18 18 45 32 4 NK 46 1 0 22 0 0 6 48 90 19 57 0 0 4 42 48 0 7
+ (L) - ligand dependent nuclear binding. + (NL) - ligand independent nuclear binding. - - no nuclear binding. N K - not known. DFI - disease free interval. ER/PR - estrogen receptor/progesterone receptor. NR - no response. PD - progressive disease. R - response. T a m - tamoxifen. Multiple - bone, lymph nodes, pleura, liver.
Nuclear binding of the estrogen receptor
275
Fig. 1A. Immunofluorescenceis seen in the cytoplasmof the tumor cells (No estrogenicligand). Fig. lB. Highlylocalizedintra nuclear fluorescenceis evident when the tumour cellswere exposed to estrogenicligands at 37° C. Method: Indirect immunofluorescenceutilizing anti-receptor antibodies and FITC-labelledanti-rabbit IgG. Experimental controls included incubation with progestogenic ligand, replacing anti-receptor antibody with normal rabbit serum, or excluding primary antibody and using only the FITC-labelledanti-rabbit IgG. Further details are givenin reference [5]. cubation, the sections were fixed and processed for immunohistochemistry. The E R was localized using anti-receptor antibodies and indirect immunofluorescence [2-5]. We examined the clinical history on a total of 262 patients with breast cancer whose tumors were available frozen in our tumor bank, for objective evidence of metastatic disease, positive estrogenreceptor status (>-5 femtomoles/mg cytosol protein), hormonal treatment for a minimum of 4-6 weeks, and whether or not they were evaluable for clinical response to hormonal therapy as defined by the UICC standards [6]. Twenty-five patients satisfied all these criteria for inclusion in this study (Table 1). These twenty-five patients were among the 88 cases for which in vitro nuclear binding data were also available [5]. The patients were divided into 2 subgroups: responders (subjective and objective), and non-responders. Stable disease status was classified as non-response. The in vitro binding test yielded three sub-classes: 1. Ligand-dependent nuclear binding (See Fig. 1A, B). 2. Ligand-independent nuclear binding. 3. No nuclear binding. The clinical response evaluation was done at various medical centers by the medical oncologists without prior knowledge of the in vitro nuclear binding data. In most cases (84%), the primary
biopsy was examined for nuclear binding, whereas in others recurrences were examined because the primaries were unavailable. In a few cases where serial studies were possible, there was good correlation of the nuclear binding features between the primary and recurrent sites. The statistical analysis for the correlative data was by two-sided Fishers' exact test.
Results
Table 1 lists the clinical features of the patients, response data, and the classification based on nuclear binding. One patient was male, and among the women, 92% were post-menopausal. Bone metastasis was the most common site (52%), followed by soft tissue sites (skin, lymph node) in 36%. Eighty percent received tamoxifen as hormone therapy. Eight patients out of twenty-five were responders; of these, five had bone metastasis and three had soft-tissue metastasis (Table 1). Based on the in vitro nuclear binding test (Table 2), six patients had normal E R with ligand-dependent nuclear binding, ten patients had abnormal E R with ligand-independent nuclear binding, and nine patients had non-nuclear-binding E R (Table 2). Five out of six with ligand-dependent nuclear binding E R were responsive to hormone therapy; none
276
N J Robert et al.
of the ten patients with ligand-independent nuclear binding of E R responded to hormone treatment; only three out of nine patients who had non-nuclear-binding E R were responsive. Thus, clinical response rate among patients with abnormalities in the nuclear binding of E R was only 16% (3/19), while among these with ligand-dependent nuclear binding ER, 86% (5/6) responded to hormonal treatment. These data were statistically significant (p = 0.011). This group was also analyzed for prediction of response on the basis of E R level equal to or exceeding 100 femtomoles/mg, and on the basis of presence or absence of PR. As illustrated in Table 3A and 3B, a high level of E R was not predictive of clinical response. As seen from the data in Table 4, 40% of the progesterone receptor positive tumors responded, while 29% lacking PR also responded. Thus, in this small group in which PR level was known in 17 cases, the PR status was not highly predictive of response to hormone therapy. However, among these seventeen patients with known PR status, all cases who had ligand-dependent nuclear binding were responders (100%); among the patients with abnormal nuclear binding of ER, one out of twelve (8%) responded to hormone therapy (Table 4).
Discussion
In this small series, the pattern of nuclear binding of E R appeared predictive for hormone response among the E R + tumors, irrespective of the PR status or the level of ER ( - 100 femtomoles/mg).
Leak et al., who evaluated patients for soluble and nuclear ER status with a biochemical ligand binding test, made a similar observation [7]. In that study a response rate of 71% was reported when the tumors had both soluble and nuclear ER, and 19% when only one of the two varieties was present. About 30% response has been reported among PR-negative tumors, and the same frequency of therapy failures among P R + tumors [8]. Other factors such as histological grade, menopausal status, disease-free interval, sites of metastasis, and type of hormonal treatment may all play a role in predicting response. Clearly, a larger group of patients needs to be studied to evaluate the interaction of all these parameters and to further document the wider application of this in vitro test. The most interesting significance of this in vitro nuclear binding test may be its ability to explain clinical response/non-response behavior in biological terms. In target tissues, the tissue requires estrogen as well as the binding of the receptorhormone complex (transformed ER) to the nuclear acceptor sites for triggering hormone-dependent events in the cell. A tumor which contains E R with an unimpaired hormone binding site but impaired nuclear binding may be hormone-independent, if it can by-pass the requirement for activated estrogen receptor. Similarly, a tumor equipped with E R which can bind to nuclear acceptor sites in the absence of estrogen also bypasses the requirement for estrogen and is estrogen-independent. Antiestrogens such as tamoxifen will not be effective for treating these tumors either by virtue of their inability to reach the nucleus and exhibit their anti-
Table 2. Nuclear binding assay and tumor response Tumor Status
Response
No response
Progressive disease
Normal binding: Ligand dependent nuclear binding
5
0
1
Total (% response)
6 (83%) I p = 0.011
Abnormal binding: Ligand independent nuclear binding No nuclear binding
0
2
8
10 (0%)
3
1
5
9 (33%)
Nuclear binding of the estrogen receptor Table 3A. Estrogen receptor level and nuclear binding assay
polyclonal antibody reagent used in our study identifies the E R in frozen tumors [2] and in tissue culture cells maintained in estrogen-free medium [4] in the cytoplasm and hormone-complexed, in vitro transformed, ER in the nucleus [2-5]. The reason for this apparent difference in E R distribution between these antibody reagents has not yet been resolved. It is to be noted that Wilkstrom et al. have reported results similar to ours using monoclonal antibodies to GR [12]. Among the patients with metastatic disease, in addition to predicting which E R + patients will benefit from hormone therapy, this in vitro test may be useful to define, among the primary tumors, which patients would be better candidates for hormone therapy in an adjuvant setting. The need for better markers for predicting hormoneresponse among the E R + tumors continues to be felt [13], and this test, once its usefulness is documented in a larger patient population, may provide that clue needed for identifying that sub-set of patients who may respond to hormone treatment.
Estrogen receptor level range (mean) fmol/mg Normal binding: Ligand dependent nuclear binding Abnormal binding: Ligand independent nuclear binding No nuclear binding
20q595 (242)
12-727 (178) 8-1418 (230)
Table 3B. Estrogen receptor level and tumor response Estrogen Response receptor level (fmol/mg)
No response Total or progression (% response)
< 100 > 100
9 8
5 3
277
14 (36%) 11 (27%)
estrogenic characteristics (in tumors with non-nuclear-binding ER) or by their nuclear binding being inconsequential (as in tumors with ligand-independent nuclear binding ER). Receptor mutants with nuclear binding features similar to those described in this report have been described for glucocorticoid receptors (GR) in lymphomas resistant to dexamethasone therapy [9]. Dickinson et al. [10] have described estrogen-independent synthesis of growth factors in MCF-7 cells with the insertion of H-ras oncogene. Unlike the commercially available monoclonal anti-ER antibodies [11] which locate hormone-free receptors in the nucleus, the
Acknowledgements Supported by a grant from the National Cancer Institute.
Table 4. Progesterone receptor assay, nuclear binding assay, and tumor response
ER+/PR+ ER+/PRNormal binding: Ligand dependent nuclear binding Abnormal binding: Ligand independent nuclear binding No nuclear binding ER - estrogen receptor. PR - progesterone receptor.
Response
No response
Progression
Total (% response)
4 2
1 2
5 3
10 (40%) 7 (29%)
5
0
0
5 (100%)
0 1
2 1
8 0
10 "1 2 ~(8%)
278
N J R o b e r t et al.
References 1. Bloom ND, Tobin EH, Schreibman B, Degenshein GA: The role of progesterone receptors in the management of advanced breast cancer. Cancer 45: 2992-2997, 1980 2. Tamura H, Raam S, Sneedy A, Pappas C: An update on the immunohistochemical localization of estrogen receptors in mammary carcinomas utilizing polyclonal anti-receptor antibodies. Eur J Cancer Clin Onco120: 1261-1277, 1984 3. Raam S, Peters L, Rafkind I, Putnum E, Longcope C, Cohen JL: Simple methods of production and characterization of rabbit antibodies for human breast estrogen receptors. Mol Immunol 18: 143-156, 1981 4. Raam S, Richardson GS, Bradley F et aL : Translocation of cytoplasmic estrogen receptors to the nucleus: immunohistochemical demonstration utilizing rabbit antibodies to estrogen receptors of mammary carcinomas. Breast Cancer Res Treat 3: 179-199, 1983 5. Raam S, Robert N, Pappas CA et al. : Defective estrogen receptors in human mammary cancers: their significance in defining hormone dependence. J Natl Cancer Inst 80: 756761, 1988 6. Hayward JL, Carbone PP, Heuson J-C, Kumaoka S, Segaloft A, Rubens RD: Assessment of response to therapy in advanced breast cancer. Cancer 39: 1289-1294, 1977 7. Leak RE, Laing L, Caiman KC, MacBeth FR, Crawford
8.
9.
10.
11.
12.
13.
D, Smith DC: Estrogen-receptor status and endocrine therapy of breast cancer: response rates and status stability. Br J Cancer 43: 59-66, 1981 King RJB, Stewart JF, Millis RR, Rubens RD, Hayward JL: Quantitative comparison of estradiol and progesterone receptor contents of primary and metastatic human breast tumors in relation to response to endocrine treatment. Breast Cancer Res Treat 2: 339-346, 1982 Miesfeld R, Okret S, Wikstrom A, Wrange O, Gustafsson J, Yamamoto KR: Characterization of a steroid hormone receptor gene and mRNA in wild type and mutant cells. Nature 32: 779-782, 1984 Dickson RB, Bates SE, McManaway ME, Lippman ME: Characterization of estrogen responsive transforming activity in human breast cancer cell lines. Cancer Res 46: 17071713, 1986 King WL, Greene GL: Monoclonal antibodies localize estrogen receptor in the nuclei of target cells. Nature 307: 745-747, 1984 Wikstrom A-C, Bakke O, Okret S, Bronnegard M, Gustafsson J-A: Intracellular localization of the glucocorticoid receptor: evidence for cytoplasmic and nuclear localization. Endocrinology 120: 1232-1242, 1987 McGuire WL: Adjuvant therapy in node-negative breast cancer (editorial), N Engl J Med 320: 525-527, 1989
Report
Nuclear binding of the estrogen receptor: A potential predictor for hormone response in metastatic breast cancer
Nicholas J. Robert, Leslie Martin 1, C. Douglas Taylor 1, James Popkin ~, David R. Parkinson 2, Charles White ~, Constantine A. Pappas ~, Hisashi Tamura 1, Rebecca Gelman 3, Joseph Cohen ~ and Shanthi Raam 1 Division of Hematology/Oncology, New England Medical Center, Boston, MA, USA; i Tufts University School of Medicine, Boston, MA, USA; 2Biological Therapy Unit, M.D. Anderson Cancer Center, Houston, TX, USA; 3Division of Biostatistics, Dana Farber Cancer Institute, Boston, MA, USA
Key words: estrogen receptor, hormone therapy, immunohistochemistry, nuclear binding Abstract
We have previously described an in vitro immunohistochemical test employing anti-receptor antibodies, for demonstrating the nuclear binding characteristics of estrogen receptors (ER) in breast carcinomas. Based on a retrospective analysis of twenty-five patients with estrogen receptor-positive ( E R + ) breast cancer who were treated with hormone therapy and whose clinical responses were evaluable, we were able to demonstrate that this test may be valuable to predict which, among the E R + tumors (whether or not they are progesterone receptor positive, P R + ) , are likely to respond to hormone therapy and which may fail. While tumors in which ER exhibited abnormalities in nuclear binding behavior (ligand-independent nuclear binding or no nuclear binding) failed hormone therapy (16 out of 19 patients), those in which nuclear binding of ER appeared normal (ligand-dependent) in the in vitro test, responded to hormone therapy (5/6 patients). While our previous report dealt with the procedural details, specificity of the reagents, and the design of the study, this report addresses the clinical aspects of this study and response correlation.
Introduction
Endocrine therapy elicits variable responses in patients with metastatic breast cancer. Although many factors appear to play a role in predicting responses, the most notable parameter has been the E R + status of the tumor [1]; in this group 30-70% of the patients will respond to hormone therapy. Two major functions have been associated with estrogen-receptors: binding of the hormone, and binding to the specific sites in the target cell nucleus as a receptor-hormone complex to ini-
tiate other estrogen-dependent molecular events such as progesterone receptor synthesis. With the availability of anti-receptor antibodies which recognize hormone-free as well as nuclear ER, and an in vitro test utilizing these antibodies [2-5], it has been possible to study the nuclear binding characteristic of ER in cryosections from tumors which were stored frozen for a long period of time (--- 10 years). In these tumors, the ER is still well preserved and is able to bind the hormone with high affinity. With the help of this procedure we demon-
Address for offprints: N.J. Robert, Division of Hematology/Oncology, New England Medical Center, 750 Washington Street, Boston,
MA 02111, USA
274
N J Robert et al.
strated three subclasses based on the in vitro nuclear binding characteristics of ER: A) Tumors in which E R exhibits normal liganddependent nuclear binding requiring both the presence of estrogen and exposure in vitro to temperature of 37°C for thirty minutes; B) Tumors in which E R binds to the nuclei in the absence of estrogen, but requires exposure to 37°C for thirty minutes thus exhibiting abnormal ligand-independent nuclear binding; C) Tumors in which ER is unable to bind to the nucleus. In all these three subclasses the hormone binding characteristics of the cytoplasmic E R was unimpaired and was typical of the pattern described for Type I E R in normal target cells [2-5]. Twenty-five patients out of 88 E R + tumors classified as such were evaluable for clinical response, and the results
of the in vitro nuclear binding test were correlated with clinical response to hormone therapy.
Material and methods
The test procedure, specificity of the results, criteria for classification of the E R + tumors on the basis of nuclear binding parameters, and the reproducibility of the classification procedure, have been described in detail previously [2-5] and therefore will not be included in this report. Briefly, the cryosections of the tumors were exposed either to ligand-free saline or saline with an estrogenic ligand (estrogen or DES) at 37° C for thirty minutes in a moist chamber. A progestin was used in experimental controls. Cryosections were also directly processed without prior incubation. After the in-
Table 1. Clinical data Patient Age (yrs)
ER/PR
Nuclear binding DFI (months)
Metastatic Site(s)
Therapy
Response
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25
1418 +/NK 12/9/NK 52/216 106/119/280/63/491 19/NK 695+/NK 230/89/104 20/99 27/6 37/NK 415/249 231/8/NK 727/37/NK 68/5 93/67 363/81 157/363 38/NK
+ (NL) +(L) + (NL) +(NL) + (NL) +(L) + (L) + (NL) +(L) +(NL) +(NL) + (NL) +(NL) +(L) +(L) + (NL) -
Lymph node Skin Lymph node, Adrenal Breast, Skin Bone Bone Bone Skin, Bone Skin Lung Lymph node, Lung Bone Bone Skin Skin Lymph node Bone, Lung Bone Multiple Bone Bone Bone Bone Ascites Bone, Lung
Tam Tam Tam Tam Tam Tam Aminoglutethemide Megace Oophorectomy Tam Tam Tam Tam Tam Tam Tam Tam Tam Tam Tam Tam Tam DES Tam Adrenalectomy
PD PD PD R PD NR R PD PD PD R PD R PD R NR NR R PD PD PD R R PD PD
65 61 55 55 58 53 58 65 51 75 64 62 71 62 75 47 78 64 66 71 46 60 63 67 73
0 18 18 45 32 4 NK 46 1 0 22 0 0 6 48 90 19 57 0 0 4 42 48 0 7
+ (L) - ligand dependent nuclear binding. + (NL) - ligand independent nuclear binding. - - no nuclear binding. N K - not known. DFI - disease free interval. ER/PR - estrogen receptor/progesterone receptor. NR - no response. PD - progressive disease. R - response. T a m - tamoxifen. Multiple - bone, lymph nodes, pleura, liver.
Nuclear binding of the estrogen receptor
275
Fig. 1A. Immunofluorescenceis seen in the cytoplasmof the tumor cells (No estrogenicligand). Fig. lB. Highlylocalizedintra nuclear fluorescenceis evident when the tumour cellswere exposed to estrogenicligands at 37° C. Method: Indirect immunofluorescenceutilizing anti-receptor antibodies and FITC-labelledanti-rabbit IgG. Experimental controls included incubation with progestogenic ligand, replacing anti-receptor antibody with normal rabbit serum, or excluding primary antibody and using only the FITC-labelledanti-rabbit IgG. Further details are givenin reference [5]. cubation, the sections were fixed and processed for immunohistochemistry. The E R was localized using anti-receptor antibodies and indirect immunofluorescence [2-5]. We examined the clinical history on a total of 262 patients with breast cancer whose tumors were available frozen in our tumor bank, for objective evidence of metastatic disease, positive estrogenreceptor status (>-5 femtomoles/mg cytosol protein), hormonal treatment for a minimum of 4-6 weeks, and whether or not they were evaluable for clinical response to hormonal therapy as defined by the UICC standards [6]. Twenty-five patients satisfied all these criteria for inclusion in this study (Table 1). These twenty-five patients were among the 88 cases for which in vitro nuclear binding data were also available [5]. The patients were divided into 2 subgroups: responders (subjective and objective), and non-responders. Stable disease status was classified as non-response. The in vitro binding test yielded three sub-classes: 1. Ligand-dependent nuclear binding (See Fig. 1A, B). 2. Ligand-independent nuclear binding. 3. No nuclear binding. The clinical response evaluation was done at various medical centers by the medical oncologists without prior knowledge of the in vitro nuclear binding data. In most cases (84%), the primary
biopsy was examined for nuclear binding, whereas in others recurrences were examined because the primaries were unavailable. In a few cases where serial studies were possible, there was good correlation of the nuclear binding features between the primary and recurrent sites. The statistical analysis for the correlative data was by two-sided Fishers' exact test.
Results
Table 1 lists the clinical features of the patients, response data, and the classification based on nuclear binding. One patient was male, and among the women, 92% were post-menopausal. Bone metastasis was the most common site (52%), followed by soft tissue sites (skin, lymph node) in 36%. Eighty percent received tamoxifen as hormone therapy. Eight patients out of twenty-five were responders; of these, five had bone metastasis and three had soft-tissue metastasis (Table 1). Based on the in vitro nuclear binding test (Table 2), six patients had normal E R with ligand-dependent nuclear binding, ten patients had abnormal E R with ligand-independent nuclear binding, and nine patients had non-nuclear-binding E R (Table 2). Five out of six with ligand-dependent nuclear binding E R were responsive to hormone therapy; none
276
N J Robert et al.
of the ten patients with ligand-independent nuclear binding of E R responded to hormone treatment; only three out of nine patients who had non-nuclear-binding E R were responsive. Thus, clinical response rate among patients with abnormalities in the nuclear binding of E R was only 16% (3/19), while among these with ligand-dependent nuclear binding ER, 86% (5/6) responded to hormonal treatment. These data were statistically significant (p = 0.011). This group was also analyzed for prediction of response on the basis of E R level equal to or exceeding 100 femtomoles/mg, and on the basis of presence or absence of PR. As illustrated in Table 3A and 3B, a high level of E R was not predictive of clinical response. As seen from the data in Table 4, 40% of the progesterone receptor positive tumors responded, while 29% lacking PR also responded. Thus, in this small group in which PR level was known in 17 cases, the PR status was not highly predictive of response to hormone therapy. However, among these seventeen patients with known PR status, all cases who had ligand-dependent nuclear binding were responders (100%); among the patients with abnormal nuclear binding of ER, one out of twelve (8%) responded to hormone therapy (Table 4).
Discussion
In this small series, the pattern of nuclear binding of E R appeared predictive for hormone response among the E R + tumors, irrespective of the PR status or the level of ER ( - 100 femtomoles/mg).
Leak et al., who evaluated patients for soluble and nuclear ER status with a biochemical ligand binding test, made a similar observation [7]. In that study a response rate of 71% was reported when the tumors had both soluble and nuclear ER, and 19% when only one of the two varieties was present. About 30% response has been reported among PR-negative tumors, and the same frequency of therapy failures among P R + tumors [8]. Other factors such as histological grade, menopausal status, disease-free interval, sites of metastasis, and type of hormonal treatment may all play a role in predicting response. Clearly, a larger group of patients needs to be studied to evaluate the interaction of all these parameters and to further document the wider application of this in vitro test. The most interesting significance of this in vitro nuclear binding test may be its ability to explain clinical response/non-response behavior in biological terms. In target tissues, the tissue requires estrogen as well as the binding of the receptorhormone complex (transformed ER) to the nuclear acceptor sites for triggering hormone-dependent events in the cell. A tumor which contains E R with an unimpaired hormone binding site but impaired nuclear binding may be hormone-independent, if it can by-pass the requirement for activated estrogen receptor. Similarly, a tumor equipped with E R which can bind to nuclear acceptor sites in the absence of estrogen also bypasses the requirement for estrogen and is estrogen-independent. Antiestrogens such as tamoxifen will not be effective for treating these tumors either by virtue of their inability to reach the nucleus and exhibit their anti-
Table 2. Nuclear binding assay and tumor response Tumor Status
Response
No response
Progressive disease
Normal binding: Ligand dependent nuclear binding
5
0
1
Total (% response)
6 (83%) I p = 0.011
Abnormal binding: Ligand independent nuclear binding No nuclear binding
0
2
8
10 (0%)
3
1
5
9 (33%)
Nuclear binding of the estrogen receptor Table 3A. Estrogen receptor level and nuclear binding assay
polyclonal antibody reagent used in our study identifies the E R in frozen tumors [2] and in tissue culture cells maintained in estrogen-free medium [4] in the cytoplasm and hormone-complexed, in vitro transformed, ER in the nucleus [2-5]. The reason for this apparent difference in E R distribution between these antibody reagents has not yet been resolved. It is to be noted that Wilkstrom et al. have reported results similar to ours using monoclonal antibodies to GR [12]. Among the patients with metastatic disease, in addition to predicting which E R + patients will benefit from hormone therapy, this in vitro test may be useful to define, among the primary tumors, which patients would be better candidates for hormone therapy in an adjuvant setting. The need for better markers for predicting hormoneresponse among the E R + tumors continues to be felt [13], and this test, once its usefulness is documented in a larger patient population, may provide that clue needed for identifying that sub-set of patients who may respond to hormone treatment.
Estrogen receptor level range (mean) fmol/mg Normal binding: Ligand dependent nuclear binding Abnormal binding: Ligand independent nuclear binding No nuclear binding
20q595 (242)
12-727 (178) 8-1418 (230)
Table 3B. Estrogen receptor level and tumor response Estrogen Response receptor level (fmol/mg)
No response Total or progression (% response)
< 100 > 100
9 8
5 3
277
14 (36%) 11 (27%)
estrogenic characteristics (in tumors with non-nuclear-binding ER) or by their nuclear binding being inconsequential (as in tumors with ligand-independent nuclear binding ER). Receptor mutants with nuclear binding features similar to those described in this report have been described for glucocorticoid receptors (GR) in lymphomas resistant to dexamethasone therapy [9]. Dickinson et al. [10] have described estrogen-independent synthesis of growth factors in MCF-7 cells with the insertion of H-ras oncogene. Unlike the commercially available monoclonal anti-ER antibodies [11] which locate hormone-free receptors in the nucleus, the
Acknowledgements Supported by a grant from the National Cancer Institute.
Table 4. Progesterone receptor assay, nuclear binding assay, and tumor response
ER+/PR+ ER+/PRNormal binding: Ligand dependent nuclear binding Abnormal binding: Ligand independent nuclear binding No nuclear binding ER - estrogen receptor. PR - progesterone receptor.
Response
No response
Progression
Total (% response)
4 2
1 2
5 3
10 (40%) 7 (29%)
5
0
0
5 (100%)
0 1
2 1
8 0
10 "1 2 ~(8%)
278
N J R o b e r t et al.
References 1. Bloom ND, Tobin EH, Schreibman B, Degenshein GA: The role of progesterone receptors in the management of advanced breast cancer. Cancer 45: 2992-2997, 1980 2. Tamura H, Raam S, Sneedy A, Pappas C: An update on the immunohistochemical localization of estrogen receptors in mammary carcinomas utilizing polyclonal anti-receptor antibodies. Eur J Cancer Clin Onco120: 1261-1277, 1984 3. Raam S, Peters L, Rafkind I, Putnum E, Longcope C, Cohen JL: Simple methods of production and characterization of rabbit antibodies for human breast estrogen receptors. Mol Immunol 18: 143-156, 1981 4. Raam S, Richardson GS, Bradley F et aL : Translocation of cytoplasmic estrogen receptors to the nucleus: immunohistochemical demonstration utilizing rabbit antibodies to estrogen receptors of mammary carcinomas. Breast Cancer Res Treat 3: 179-199, 1983 5. Raam S, Robert N, Pappas CA et al. : Defective estrogen receptors in human mammary cancers: their significance in defining hormone dependence. J Natl Cancer Inst 80: 756761, 1988 6. Hayward JL, Carbone PP, Heuson J-C, Kumaoka S, Segaloft A, Rubens RD: Assessment of response to therapy in advanced breast cancer. Cancer 39: 1289-1294, 1977 7. Leak RE, Laing L, Caiman KC, MacBeth FR, Crawford
8.
9.
10.
11.
12.
13.
D, Smith DC: Estrogen-receptor status and endocrine therapy of breast cancer: response rates and status stability. Br J Cancer 43: 59-66, 1981 King RJB, Stewart JF, Millis RR, Rubens RD, Hayward JL: Quantitative comparison of estradiol and progesterone receptor contents of primary and metastatic human breast tumors in relation to response to endocrine treatment. Breast Cancer Res Treat 2: 339-346, 1982 Miesfeld R, Okret S, Wikstrom A, Wrange O, Gustafsson J, Yamamoto KR: Characterization of a steroid hormone receptor gene and mRNA in wild type and mutant cells. Nature 32: 779-782, 1984 Dickson RB, Bates SE, McManaway ME, Lippman ME: Characterization of estrogen responsive transforming activity in human breast cancer cell lines. Cancer Res 46: 17071713, 1986 King WL, Greene GL: Monoclonal antibodies localize estrogen receptor in the nuclei of target cells. Nature 307: 745-747, 1984 Wikstrom A-C, Bakke O, Okret S, Bronnegard M, Gustafsson J-A: Intracellular localization of the glucocorticoid receptor: evidence for cytoplasmic and nuclear localization. Endocrinology 120: 1232-1242, 1987 McGuire WL: Adjuvant therapy in node-negative breast cancer (editorial), N Engl J Med 320: 525-527, 1989
