The Variability of Estrogen Receptors in Metastatic Breast Cancer Michael J. Brennan, Milwaukee, Wisconsin William L. Donegan, MD,* Milwaukee, Wisconsin Dougias E. Appieby, MS,+ Milwaukee, Wisconsin
The most important recent biochemical contribution to the management of the patient with metastatic breast cancer has been the determination of estrogen receptor protein (ERP) in breast cancer tissue. It is a considerable aid in the selection of patients for hormonal manipulation. Response rates to hormonal therapy in patients with significant ERP levels are approximately 55 per cent, whereas response rates in patients with insignificant ERP levels are only about 5 per cent [I]. Since metastatic disease may occur in sites not easily accessible to biopsy, it is important to measure the ERP of the primary tumors for this purpose; it is also important to appreciate that dissimilarities may exist in this regard between primary and metastatic tumors. The degree and importance of this variability has been the object of recent study and is the topic of this report. Singhakowinta et al [2] in 1976 reported that in eleven of their twelve patients with multiple metastases there was no individual discrepancy of ERP status at the various biopsy sites. Leclerq et al [3] used twenty-three mastectomy specimens to compare the ERP levels in the primary tumors and the invaded axillary nodes. He reported that twenty-two of the twenty-three had similar levels and that there was a statistically significant correlation for receptor concentrations. The present report provides additional information on the ERP content of primary breast cancers and their metastases. Material and Methods A laboratory for assay of ERP in tumor tissue using the sucrose diffusion method of Jenson and associates was established at Columbia Hospital in Milwaukee in early 1975. This laboratory has maintained quality control by exchange of specimens with other laboratories and has
From the Department of Surgery, The Medical College of Wisconsin, and the Department of Laboratory Medicinet. Columbia Hospital, Milwaukee, Wisconsin. Reprint requests should be adbessed to William L. Donegan, MD, Division of Surgery, BOX 145, Medical College of Wisconsin, 8700 West Wisconsin Avenue, Milwaukee, Wisconsin 53226. * Professor of Clinical Oncology, American Cancer Society.
served as both a research and service facility. From February lo,1975 to December 30,1976,439 determinations were made of ERP levels in tumors from patients with breast cancer. Eighty-four patients had metastatic tumors tested. From this group, patients were excluded if an assay of the primary tumor was not performed (51 patients), if the patient was receiving hormone therapy when the tumor was biopsied (3), or if inappropriate transport of the tumor to the laboratory cast doubt upon interpretation of the results (1). The present report pertains to the twenty-nine remaining patients in whom ERP levels were determined in both the primary tumor and at least one metastasis. (Table I.) In twenty-eight cases invaded axillary nodes were sources of metastatic tumor; in twenty-seven of these cases nodes were from the mastectomy specimen that contained the primary tumor. Eleven of the patients had multiple metastases tested, including one with a chest wall metastasis and another with cells from a pleural effusion. Seven patients were premenopausai, twenty-two postmenopausal. Significant ERP levels, that is, those predictive of clinically useful frequencies of response to endocrine or hormonal therapy, were defined pragmatically by Block, Jensen, and Poiley [4] as more than 250 femtomoies/gm tissue for premenopausai patients and more than 750 femtomoles/gm tissue for postmenopausal and previously castrated individuals. Rarely do clinical responses result at lower concentrations. The processing of tumor specimens was conducted as follows: After the presence of malignancy had been verified histologically with frozen sections, the first step of the assay was homogenization of the frozen tumor specimens. The homogenates were then centrifuged at 210,000 X g for 30 minutes to obtain the cytosol fractions. Two fractions were taken. Into one fraction CI-628 (Parke-Davis) was added to block the specific estrogen binding sites. Both fractions of cytosol were next incubated with 3H-estradiol-17P. After incubation the two fractions were placed on a sucrose density gradient and centrifuged at 250,000 X g for 16 hours. The sucrose columns were fractionated into thirty-six 0.1 ml portions, and the radioactivity of each fraction was determined in a scintillation counter. By this means it was possible to differentiate 4s from 8s complexes. The amount of specific binding (ERP) was then determined indirectly by subtracting the amount of labeled estradiol bound to nonspecific sites from the labeled estradiol bound to all sites [I].
The American Journal of Surgery
Estrogen Receptors
TABLE I
Patient 1 2 3 4 5 6 7 6 9 10 11 12 13 14 15 16
Absence (-) and Presence (+) of ERP in 29 Primary Breast Cancers and Their Metastases ERP of Primary -
17 18
ERP and Sites of Metastases Axillary Nodes (l-6) Other --------(-)‘ --++
19 20
-
--++ +
22 21 23 24 25 26
+ +
c++,t + ++++++ -
27 28 29
+ + +
-+ + ++++
(-)+
Note: Results were dissimilar in 10 of 29 patients. Lack of concurrence between metastases was found in 3 of 10 patients with multiple metastases. Chest wall. t Pleural effusion. t Contralateral axillary nodes. l
TABLE II
ERP Status according to Patient Subgroups
ERP Status of Primary All Patients ERP(+) 5 ERP (-) 24
ERP Status of Metastases ERP (-) ERP (+ and -) ERP (+) 2 5
2 17
1 2
Patients with Multiple Metastases ERP(+) 2 1 ERP(-) 8 1
0 5
1 2
Premenopausal Patients ERP(+) 2 ERP(-) 5
0 2
1 3
1
Postmenopausal Patients ERP(+) 3 ERP i-i 19
2 3
1 14
2
per cent) was similar to the consistency overall between primary tumors and metastases (65.5 per cent). Consistently similar ERP levels were found in metastases of 71 per cent of ERP-negative primaries and 40 per cent of ERP-positive ones. (Table I.) At least some metastases reflected the parent tumor’s ERP status in 79 per cent of positive primaries and 60 per cent of negative primaries. Four of seven premenopausal women and six of twenty-two postmenopausal patients failed to have consistent ERP levels in all tumor specimens. To demonstrate that inconsistently reacting cases were not those with marginally positive or negative ERP levels, the values for the ten dissimilar cases are graphically presented in Figure 1.
Results
Comments and Conclusions
The primary tumors of five patients contained significant concentrations of ERP, the metastases of ten did so. The result of the assays of the primary and metastatic tumors were similar in nineteen of the twenty-nine cases (65.5 per cent). (Table I.) In seven cases the metastatic tumors did not have the same significance level. In three cases tumor from some invaded nodes reacted similarly to the primary tumor, but tumor from other nodes did not; these three cases were classified as dissimilar reactions for purposes of statistical analysis. A similar result, that is, 60 per cent correspondence, was observed in ten patients with determinations from multiple metastases. (Table II.) In cases with multiple metastases, if any metastasis reflected the ERP content of the primary, 75.9 per cent of all cases (22 of 29) could have been considered to have primaries and metastases that reacted similarly. However, among the ten patients with determinations from multiple metastases, the consistency of ERP among metastases (70
In this study ten of twenty-nine patients (34.5 per cent) with breast cancer had metastases with an ERP level that was not reflective of the ERP level of the primary tumor. This lack of correlation held true whether primary tumors were ERP-positive or ERP-negative. The results are at odds with the reports of two studies previously cited [.2,3]. Rosen et al [5], however, observed a proportion of inconsistencies between primaries and metastases similar to ours in an identical number of patients using the dextran-charcoal method and a different level of significance for ERP assay. A factor in this discrepancy may be the problem of accurately assessing what portion of the specimen is nonneoplastic tissue. Uniformity in this respect might improve the accuracy of ERP concentration comparisons. It is relevant that the frequency of ERP positivity depends upon the criteria adopted; some investigators reporting high frequencies consider any detectable level of ERP significant [3].
Volume 137, February 1979
261
Brennan, Donegan, Appleby
l
=
Remenopwsol
x = poslmenopatJsol
PRIMARY ERP LEVEL (femfomoles~m tissue) Figure 1. Ten cases of breast cancer are shown with wide dMerences in ERP concenfratlons between p&nary tumors and metastases. More than one metastasis was evaluated in four cases. Levels of significance were those accepted by Block, Jensen, and Po//ey [4] at 750 femtomoles/gm tissue for postmenopausal women and 250 femtomohdgm tissue for premenopausal women.
Of the twenty-nine cases, only five primary tumors (17.2 per cent) were ERP-positive according to customary criteria. This ratio is lower than in most other studies, but it should be recognized that the study group was selected as patients with proven metastases who had the appropriate sampling. It may not be representative of the total spectrum of persons with breast cancer. Secondly, the concentration of ERP had to be greater than a threshold limit to be considered positive. The present study clearly does not support the assertion of Leclerq et al [3] that the receptor content of “. . . a primary tumor and its axillary nodal metastases could be regarded as one.” Instead, it is consistent with the concept that a breast cancer is multiclonal, composed of both ERP-negative and ERP-positive cells in varying proportions. These cells may metastasize. or grow at widely varying rates and differ in response to hormonal manipulation. Although the ERP status of the primary and at least some of its metastases are in accord more often than not, lack of concurrence is frequent. This underscores the importance of not relying solely on past evaluation of a primary tumor removed with mastectomy when metastases become evident and are accessible. In this setting, evaluation of the metastasis provides
262
a more accurate reflection of its potential hormonal responsiveness. Unfortunately, it may not offer a more reliable guide to the responsiveness of other metastases than is provided by the primary. In our series the ERP status of one metastasis reflected that of all others tested in little more than two thirds of the cases; therefore, the accuracy of predicting overall patient benefit may not be greatly improved. The data presented help to explain a number of clinical observations. Among these are the response to hormones of metastases from an occasional primary tumor devoid of ERP, the failure of metastases from more than one third of ERP-containing tumor to respond, and the occurrence of mixed responses to hormonal therapy. Summary
Estrogen receptor protein (ERP) concentrations were determined by the sucrose diffusion method in primary tumors and one or more metastases in twenty-nine patients with breast cancer. Concurrence of ERP concentrations between primaries and at least some metastases was found in 76 per cent of cases. Multiple metastases were assayed in ten cases, three of which demonstrated highly variable concentrations. It was concluded that clinically significant differences in ERP concentrations often exist between primary breast cancers and their metastases as well as between different metastases from the same tumor, accounting for the lack of responsiveness of some ERP-“positive” tumors and for mixed responses to hormonal or endocrine therapy. Assay of an isolated metastasis may be no more reliable in predicting overall patient benefit from therapy than assay of the primary itself. References 1. McGuire WL, Carbone PP. Vollmer EP: Estrogen Receptors in Human Breast Cancer. New York, Raven Press, 1975. 2. Singhakowinta A, Potter HG. Buroker JR, et al: Estrogen receptors and natural course of breast cancer. Ann Surg 183(l): 84, 1978. 3. Leclerq G, Heuson J, Deboel M, Matteiem W: Oestrogen receptors in breast cancer: a changing concept. Br A&d J 1: 185, 1975. 4. Block GE, Jensen E, Polley T: The prediction of hormonal dependency of mammary cancer. Ann Surg 182(3): 342, 1975. 5. Rosen PP, Henendez-Botet CJ, Urban JA, et al: Estrogen receptor protein (ERP) in multiple tumor specimens from individual patients with breast cancer. Cancer 39: 2194, 1977.
The American Journal of Surgery
The most important recent biochemical contribution to the management of the patient with metastatic breast cancer has been the determination of estrogen receptor protein (ERP) in breast cancer tissue. It is a considerable aid in the selection of patients for hormonal manipulation. Response rates to hormonal therapy in patients with significant ERP levels are approximately 55 per cent, whereas response rates in patients with insignificant ERP levels are only about 5 per cent [I]. Since metastatic disease may occur in sites not easily accessible to biopsy, it is important to measure the ERP of the primary tumors for this purpose; it is also important to appreciate that dissimilarities may exist in this regard between primary and metastatic tumors. The degree and importance of this variability has been the object of recent study and is the topic of this report. Singhakowinta et al [2] in 1976 reported that in eleven of their twelve patients with multiple metastases there was no individual discrepancy of ERP status at the various biopsy sites. Leclerq et al [3] used twenty-three mastectomy specimens to compare the ERP levels in the primary tumors and the invaded axillary nodes. He reported that twenty-two of the twenty-three had similar levels and that there was a statistically significant correlation for receptor concentrations. The present report provides additional information on the ERP content of primary breast cancers and their metastases. Material and Methods A laboratory for assay of ERP in tumor tissue using the sucrose diffusion method of Jenson and associates was established at Columbia Hospital in Milwaukee in early 1975. This laboratory has maintained quality control by exchange of specimens with other laboratories and has
From the Department of Surgery, The Medical College of Wisconsin, and the Department of Laboratory Medicinet. Columbia Hospital, Milwaukee, Wisconsin. Reprint requests should be adbessed to William L. Donegan, MD, Division of Surgery, BOX 145, Medical College of Wisconsin, 8700 West Wisconsin Avenue, Milwaukee, Wisconsin 53226. * Professor of Clinical Oncology, American Cancer Society.
served as both a research and service facility. From February lo,1975 to December 30,1976,439 determinations were made of ERP levels in tumors from patients with breast cancer. Eighty-four patients had metastatic tumors tested. From this group, patients were excluded if an assay of the primary tumor was not performed (51 patients), if the patient was receiving hormone therapy when the tumor was biopsied (3), or if inappropriate transport of the tumor to the laboratory cast doubt upon interpretation of the results (1). The present report pertains to the twenty-nine remaining patients in whom ERP levels were determined in both the primary tumor and at least one metastasis. (Table I.) In twenty-eight cases invaded axillary nodes were sources of metastatic tumor; in twenty-seven of these cases nodes were from the mastectomy specimen that contained the primary tumor. Eleven of the patients had multiple metastases tested, including one with a chest wall metastasis and another with cells from a pleural effusion. Seven patients were premenopausai, twenty-two postmenopausal. Significant ERP levels, that is, those predictive of clinically useful frequencies of response to endocrine or hormonal therapy, were defined pragmatically by Block, Jensen, and Poiley [4] as more than 250 femtomoies/gm tissue for premenopausai patients and more than 750 femtomoles/gm tissue for postmenopausal and previously castrated individuals. Rarely do clinical responses result at lower concentrations. The processing of tumor specimens was conducted as follows: After the presence of malignancy had been verified histologically with frozen sections, the first step of the assay was homogenization of the frozen tumor specimens. The homogenates were then centrifuged at 210,000 X g for 30 minutes to obtain the cytosol fractions. Two fractions were taken. Into one fraction CI-628 (Parke-Davis) was added to block the specific estrogen binding sites. Both fractions of cytosol were next incubated with 3H-estradiol-17P. After incubation the two fractions were placed on a sucrose density gradient and centrifuged at 250,000 X g for 16 hours. The sucrose columns were fractionated into thirty-six 0.1 ml portions, and the radioactivity of each fraction was determined in a scintillation counter. By this means it was possible to differentiate 4s from 8s complexes. The amount of specific binding (ERP) was then determined indirectly by subtracting the amount of labeled estradiol bound to nonspecific sites from the labeled estradiol bound to all sites [I].
The American Journal of Surgery
Estrogen Receptors
TABLE I
Patient 1 2 3 4 5 6 7 6 9 10 11 12 13 14 15 16
Absence (-) and Presence (+) of ERP in 29 Primary Breast Cancers and Their Metastases ERP of Primary -
17 18
ERP and Sites of Metastases Axillary Nodes (l-6) Other --------(-)‘ --++
19 20
-
--++ +
22 21 23 24 25 26
+ +
c++,t + ++++++ -
27 28 29
+ + +
-+ + ++++
(-)+
Note: Results were dissimilar in 10 of 29 patients. Lack of concurrence between metastases was found in 3 of 10 patients with multiple metastases. Chest wall. t Pleural effusion. t Contralateral axillary nodes. l
TABLE II
ERP Status according to Patient Subgroups
ERP Status of Primary All Patients ERP(+) 5 ERP (-) 24
ERP Status of Metastases ERP (-) ERP (+ and -) ERP (+) 2 5
2 17
1 2
Patients with Multiple Metastases ERP(+) 2 1 ERP(-) 8 1
0 5
1 2
Premenopausal Patients ERP(+) 2 ERP(-) 5
0 2
1 3
1
Postmenopausal Patients ERP(+) 3 ERP i-i 19
2 3
1 14
2
per cent) was similar to the consistency overall between primary tumors and metastases (65.5 per cent). Consistently similar ERP levels were found in metastases of 71 per cent of ERP-negative primaries and 40 per cent of ERP-positive ones. (Table I.) At least some metastases reflected the parent tumor’s ERP status in 79 per cent of positive primaries and 60 per cent of negative primaries. Four of seven premenopausal women and six of twenty-two postmenopausal patients failed to have consistent ERP levels in all tumor specimens. To demonstrate that inconsistently reacting cases were not those with marginally positive or negative ERP levels, the values for the ten dissimilar cases are graphically presented in Figure 1.
Results
Comments and Conclusions
The primary tumors of five patients contained significant concentrations of ERP, the metastases of ten did so. The result of the assays of the primary and metastatic tumors were similar in nineteen of the twenty-nine cases (65.5 per cent). (Table I.) In seven cases the metastatic tumors did not have the same significance level. In three cases tumor from some invaded nodes reacted similarly to the primary tumor, but tumor from other nodes did not; these three cases were classified as dissimilar reactions for purposes of statistical analysis. A similar result, that is, 60 per cent correspondence, was observed in ten patients with determinations from multiple metastases. (Table II.) In cases with multiple metastases, if any metastasis reflected the ERP content of the primary, 75.9 per cent of all cases (22 of 29) could have been considered to have primaries and metastases that reacted similarly. However, among the ten patients with determinations from multiple metastases, the consistency of ERP among metastases (70
In this study ten of twenty-nine patients (34.5 per cent) with breast cancer had metastases with an ERP level that was not reflective of the ERP level of the primary tumor. This lack of correlation held true whether primary tumors were ERP-positive or ERP-negative. The results are at odds with the reports of two studies previously cited [.2,3]. Rosen et al [5], however, observed a proportion of inconsistencies between primaries and metastases similar to ours in an identical number of patients using the dextran-charcoal method and a different level of significance for ERP assay. A factor in this discrepancy may be the problem of accurately assessing what portion of the specimen is nonneoplastic tissue. Uniformity in this respect might improve the accuracy of ERP concentration comparisons. It is relevant that the frequency of ERP positivity depends upon the criteria adopted; some investigators reporting high frequencies consider any detectable level of ERP significant [3].
Volume 137, February 1979
261
Brennan, Donegan, Appleby
l
=
Remenopwsol
x = poslmenopatJsol
PRIMARY ERP LEVEL (femfomoles~m tissue) Figure 1. Ten cases of breast cancer are shown with wide dMerences in ERP concenfratlons between p&nary tumors and metastases. More than one metastasis was evaluated in four cases. Levels of significance were those accepted by Block, Jensen, and Po//ey [4] at 750 femtomoles/gm tissue for postmenopausal women and 250 femtomohdgm tissue for premenopausal women.
Of the twenty-nine cases, only five primary tumors (17.2 per cent) were ERP-positive according to customary criteria. This ratio is lower than in most other studies, but it should be recognized that the study group was selected as patients with proven metastases who had the appropriate sampling. It may not be representative of the total spectrum of persons with breast cancer. Secondly, the concentration of ERP had to be greater than a threshold limit to be considered positive. The present study clearly does not support the assertion of Leclerq et al [3] that the receptor content of “. . . a primary tumor and its axillary nodal metastases could be regarded as one.” Instead, it is consistent with the concept that a breast cancer is multiclonal, composed of both ERP-negative and ERP-positive cells in varying proportions. These cells may metastasize. or grow at widely varying rates and differ in response to hormonal manipulation. Although the ERP status of the primary and at least some of its metastases are in accord more often than not, lack of concurrence is frequent. This underscores the importance of not relying solely on past evaluation of a primary tumor removed with mastectomy when metastases become evident and are accessible. In this setting, evaluation of the metastasis provides
262
a more accurate reflection of its potential hormonal responsiveness. Unfortunately, it may not offer a more reliable guide to the responsiveness of other metastases than is provided by the primary. In our series the ERP status of one metastasis reflected that of all others tested in little more than two thirds of the cases; therefore, the accuracy of predicting overall patient benefit may not be greatly improved. The data presented help to explain a number of clinical observations. Among these are the response to hormones of metastases from an occasional primary tumor devoid of ERP, the failure of metastases from more than one third of ERP-containing tumor to respond, and the occurrence of mixed responses to hormonal therapy. Summary
Estrogen receptor protein (ERP) concentrations were determined by the sucrose diffusion method in primary tumors and one or more metastases in twenty-nine patients with breast cancer. Concurrence of ERP concentrations between primaries and at least some metastases was found in 76 per cent of cases. Multiple metastases were assayed in ten cases, three of which demonstrated highly variable concentrations. It was concluded that clinically significant differences in ERP concentrations often exist between primary breast cancers and their metastases as well as between different metastases from the same tumor, accounting for the lack of responsiveness of some ERP-“positive” tumors and for mixed responses to hormonal or endocrine therapy. Assay of an isolated metastasis may be no more reliable in predicting overall patient benefit from therapy than assay of the primary itself. References 1. McGuire WL, Carbone PP. Vollmer EP: Estrogen Receptors in Human Breast Cancer. New York, Raven Press, 1975. 2. Singhakowinta A, Potter HG. Buroker JR, et al: Estrogen receptors and natural course of breast cancer. Ann Surg 183(l): 84, 1978. 3. Leclerq G, Heuson J, Deboel M, Matteiem W: Oestrogen receptors in breast cancer: a changing concept. Br A&d J 1: 185, 1975. 4. Block GE, Jensen E, Polley T: The prediction of hormonal dependency of mammary cancer. Ann Surg 182(3): 342, 1975. 5. Rosen PP, Henendez-Botet CJ, Urban JA, et al: Estrogen receptor protein (ERP) in multiple tumor specimens from individual patients with breast cancer. Cancer 39: 2194, 1977.
The American Journal of Surgery
