(CANCER RESEARCH (SUPPL.) 46, 4278s-4281s, August 1986) Comparison of Immunochemical and Radioligand Binding Assays for Estrogen Receptors in Human Breast Tumors1

Giovanni Di Fronzo,2 Patrizia Miodini, Moreno Brivio, Vera Cappelletti, Danila Corallini, Giovanna Granata, and Enrico Ronchi Oncologia Sperimentale C, Istituto Nazionale per lo Studio e la Cura dei Tumori, Via Venzian 1, 20133 Milano, Italy

Abstract patients, for whom most investigators have estimated mainly the unoccupied fraction of ER binding sites, whereas the anti We have compared a new enzyme immunoassay (EIA) for estrogen body recognizes indistinctly apo- and holoreceptors (11); more receptors (ER) with our conventional radioligand binding assays (multi point dextran-coated charcoal assay for cytoplasmic ER and hydroxyl- over, irrespective of the occupancy of endogenous ligands, from a theoretical point of view monoclonal antibodies could also apatite exchange assay for nuclear ER). Cytoplasmic ERs were measured in 76 human breast cancer specimens by EIA and by five-point Scatchard detect those receptors which for some obscure reason lack analysis. The correlation between the two assays yielded a straight line binding activity; therefore, for clinical purposes traditional with a slope of 0.92 (r = 0.95; P < 0.001); conversely, in 31 nuclear salt binding methods should not be neglected until immunochemical extracts, linear regression analysis of hydroxylapatite exchange assay techniques have been extensively evaluated in comparative stud data with EIA showed a clear correlation (r = 0.93; P < 0.001) but a ies performed on human tissue specimens. slope of 1.7, demonstrating that EIA detects more ER sites. The binding In the present study we compared the new EIA with conven of the antibody to the cytoplasmic ER molecules was investigated by tional assays in a series of 76 breast cancer patients using a sucrose density gradient analysis, which showed that EIA recognizes five-point DCC assay and a HAP exchange assay for the cyto- both cytoplasmic forms (9 and 3S), but does not distinguish between solic and nuclear fractions, respectively. We also checked the them. Advantages and drawbacks of this method are discussed with ability of the antibody to recognize either occupied or unoccu respect to its application for routine receptor determination for clinical management of breast cancer patients. pied receptors and differential molecular forms of ER recovered from the cytoplasmic fraction and identified by sucrose density gradient analysis. Introduction Our present knowledge concerning the biochemical proper ties and the mechanism of action of ER3 has been obtained by Materials and Methods use of methods based on the interaction of a radioactive specific Materials. 16a-['25I]Iodoestradiol (200 Ci/mmol in ethanol solution) ligand with the receptor protein. Despite the large number of and 17/3-[2,4,6,7-3H]estradiol (85-110 Ci/mmol in toluene ethanol results obtained by this method, the need has emerged for an solution) were obtained from the Radiochemical Centre (Amersham, alternative assay that will recognize receptors irrespective of England). Radioinert 17/3-estradiol, diel liyIstübestrol,Norit-A char their hormonal binding activity. In fact, the binding activity can coal, gelatin, dextran T70, human -,-globulin, and bovine serum albu be affected by many factors, such as the concentration of min were obtained from Sigma Chemical Co. (St. Louis, MO). Abbott ER-EIA monoclonal kits were purchased from Abbott Laboratories exogenous reduced sulfhydrylic groups, temperature, pH and Diagnostic Division (North Chicago, IL). All other chemicals were of ionic strength of the medium, and even by endogenous hor analytical grade. mones that occupy the binding site and do not allow the Tumor Cytosol Preparation. Tissue samples obtained at surgery were detection of the receptor under normal assay conditions (1). immediately frozen and stored in liquid nitrogen for not more than 4 The limitations inherent in traditional binding assays have been weeks. Homogenization was performed by mechanical disruption with recently overcome by the preparation of a series of monoclonal a microdismembrator (Braun, Melsungen, West Germany), and the antibodies to ER, which provide an alternative approach for obtained powder was resuspended in 8 volumes of ice-cold 20 imi K2HPO„,pH7.4 at 4°C,1 mM EDTA, 3 mM NaN3, 10% glycerol (v/ the detection, purification, and characterization of ER (2, 3). A solid-phase, sandwich type EIA for human ER has been devel v), and 1 mM thioglycerol buffer. Homogenates were spun at 1,000 x g for 10 min to separate the nuclear fraction; thereafter, the superna- oped by Abbott Laboratories using two antibodies that recog tants were centrifuged at 100,000 x g for l h at 2°Cusinga TFT 50.13 nize different antigenic sites on the ER molecule. This assay fixed angle rotor in a Centrikon T-2000 ultracentrifuge (Kontron AG, allows the processing of human tumors for ER detection in an Zurich, Switzerland). The clear supernatant, freed of fat and referred easier and quicker way. to here as the cytosolic fraction, was used for DCC and EIA assays. The clinical usefulness of biochemical ER assays in specimens DCC Assay. For saturation analysis, cytosols were incubated with 5 of human breast cancer has been clearly established by corre concentrations of iodinated ligands ranging from 2.5-0.156 nM. Non lations of patient responses to endocrine therapies (4-6). There specific binding was assayed by running a duplicate set of tubes con is also evidence that patients with certain ER concentrations in taining a 100-fold M excess of radioinert 17/3-estradioI. Incubations their tumors have a better prognosis than do ER-negative cases were performed overnight at 0-4°Cand were stopped by adding to the (7-10). Since these results have been obtained using radioligand reaction mixtures a suitable volume of DCC suspension (0.5% w/v Norit-A, 0.05% w/v dextran T70, and 0.1% w/v gelatin in 20 mM binding assays, some caution should be used in the clinical K2HPO4, pH 7.4 at 4'C, 1 mM EDTA, and 3 mM NaN3 buffer). After interpretation of results obtained by immunochemical tech 15 min the charcoal was pelleted by centrifugation at 2000 x g for 10 niques. This could be the case for premenopausal breast cancer min. The clear supernatant was used for assay of bound radioactivity ' Presented at the Symposium on "Estrogen Receptor Determination with in a Multiprias I (Packard United Technologies, Downers Grove, IL) Monoclonal Antibodies," December 14, 1984, Monte Carlo. Supported in part gamma counter. Counting efficiency (82%) was checked periodically by grant PFO 84.00808.44 from the Consiglio Nazionale delle Ricerche, Rome, using the formula of Eldridge and Growther (12). Where specified, an Italy. exchange assay was performed essentially as described above but incu 2To whom requests for reprints should be addressed. bating for 2 h at 25°C.Total and nonspecific binding data were analyzed 3The abbreviations used are: ER, estrogen receptor(s); EIA, enzyme immu noassay; DCC, dextran-coated charcoal; HAP, hydroxylapatite. according to the method of Scatchard (13) and expressed as fmol of 4278s

Downloaded from cancerres.aacrjournals.org on October 1, 2021. © 1986 American Association for Cancer Research. IMMUNOCHEMICAL AND RADIOLIGAND BINDING ASSAYS FOR ER specifically bound iodinated steroid per mg of cytosolic protein. Total protein content was measured using the Coomassie blue G250 dye (Bio- 400 Rad, Munich, West Germany) essentially according to the method of Bradford (14) using a mixture of human albumin and human -y-globulin (5:3) for the calibration protein curve. Nuclear HAP Exchange Assay. The nuclear myofibrillar pellet was quickly washed twice with 10 mM K2HPO4, pH 7.4 at 4°C,5 mM MgCl2, 3 mM NaNj, 250 mM sucrose, and 0.1% Triton X-100 buffer and twice with the same buffer but without Triton X-100. The washed nuclear myofibrillar pellet was extracted with a suitable volume of homogenization buffer containing 0.5 M KC1 with vortexing at 20-min intervals, and after l h extracts were spun at 100,000 x g for 20 min to obtain a soluble nuclear extract. HAP (DNA grade Bio-Gel HTP, Bio-Rad, Richmond, CA) was washed several times according to the method of Carola and McGuire (15). Nuclear extracts were incubated with a HAP slurry in 50 mM Tris-HCl, pH 7.2 at 4'C-10 mM K2HPO4- 100 200 30O 400 3 mM NaNj buffer for 45 min with gentle vortexing and finally centri- DCC (fmol/mg protein) fuged at 800 x g for 5 min to discard the supernatant. Nuclear pellets Fig. 1. Correlation of cytopiasmic ER-EIA values with a multipoint DCC were incubated with 5 nM 17/3-[3H]estradiol in the presence or absence assay on 76 human breast cancer specimens. The plot was drawn by linear of a 200-fold M excess of diethylstilbestrol for 2 h at 30°C.At the end regression analysis of the data. Data on abscissa and ordinate are expressed in of the incubation, pellets were washed twice with the HAP suspension fmol of specifically bound steroid per mg of total protein. buffer supplemented with 1% Tween 80 to remove unbound ligand and finally extracted overnight with ethanol at room temperatures. The Table 1 Influence of cytosolic total protein content on ER-EIA values alcohol extracts were counted in 5 ml of a commercial scintillation cocktail (Pico-Fluor 15; Packard, Downers Grove, IL) using a Packard Dilution ER-EIA value ER-EIA value 300D spectrometer with automatic quench correction on each sample. factor1 (fmol/ml ofcytosol)466' (fmol/ml ofcytosol)356 ER values were expressed in fmol/mg DNA. DNA concentrations were determined on nuclear pellets by the diphenylamine method essentially 1.3 309 267 178* 178* according to the method of Burton (16). 24 EIA. The EIA was performed on the cytosolic and nuclear fractions 90 89 6.7 48 53 by use of an Abbott ER-EIA monoclonal kit. A Quantum I spectropho- 10Experimental 36Theoretical 36 tometer (Abbott) was used to measure the absorbance at 492 nm and " Mean of triplicate determinations. for data reduction. ER concentrations could be quantified only in the * Reference value (corresponding to cytosol sample with a protein concentra range of 0-500 fmol/ml of cytosol. tion of 1.97 mg/ml) used to calculate theoretical ER values. Sucrose Density Gradient Analysis. Cytosols undergoing sucrose density gradient analysis were incubated with 17/3-pH]estradiol in the absence or presence of a 200-fold M excess of diethylstilbestrol at 0- obtained with the first 2 assays and those obtained with the 4"C for 4 h. Unbound and weakly bound ligands were removed by latter (r = 0.999 and 0.998, respectively), and in both cases the treating the reaction mixtures (300 ^1)with a DCC pellet obtained from slope was about 1. a 500-fil suspension essentially as described for the saturation analysis. The influence of total protein content in cytosols processed Clear supernatants (200 ^1)were layered on 5-20% (w/w) linear sucrose by the EIA was studied by serial dilutions with homogenization gradients (4 ml) prepared in the homogenization buffer without thio- buffer. ER levels of the undiluted cytosol and five serially glycerol. Gradients were spun at 370,000 x g for 16 h at 3°Cusing an diluted samples were separately determined by EIA. The third SW60TÃŒrotorin an L8-70 ultracentrifuge (Beckman, Palo Alto, CA). sample was chosen as a reference, since its protein concentra A mixture of 14C-labeled human 7-globulin and MC-labeIed bovine tion fell within the range recommended by the instructions for serum albumin, prepared by reductive alkylation according to the the kit. Table 1 reports the results of the EIA expressed as fmol method of Rice and Means (17), was run as an internal marker. per ml of cytosol together with the dilution factors and theo Fractions of seven drops each were collected from the bottoms of the tubes. Radioactivity counting was carried out as described for nuclear retical ER values calculated starting from the reference value. ER, except that appropriate 3H/I4C standard efficiency curves were Differences between theoretical and experimental data were used for counting of dual-labeled samples. more evident for the less diluted samples, corresponding to a total protein concentration higher than 2 mg/ml. The comparative study between the EIA and the radioligand Results binding assay was carried out also on 31 nuclear extracts. The Cytoplasmic ER determinations of 76 consecutive breast linear regression of values expressed as fmol per mg of DNA cancer specimens were carried out by EIA and multipoint yielded a correlation coefficient of r = 0.93 and P < 0.001. Scatchard analysis, and the results were compared. Values Although we used the HAP exchange assay for detection of obtained by these methods were expressed in fmol ER per mg total nuclear sites, ER concentrations obtained by the EIA were cytosol protein and plotted (Fig. 1). A very clear correlation always higher than the corresponding ones determined by our was noted between the 2 sets of values as demonstrated by nuclear assay; however, there was a significant correlation be regression curve analysis (r = 0.95; P < 0.001). A straight line tween the data obtained by these two methods (r = 0.93; P < with a slope of 0.92 was obtained, meaning that on the average 0.001). Linear regression analysis of cases yielded a straight the EIA and our radioligand binding assay gave the same line but with a slope of 1.74, indicating the ability of the EIA estimation of receptor concentrations. In a small subgroup of to detect a greater number of ER sites. The results are reported patients (three premenopausal, four postmenopausal), ER lev in Fig. 2. els were determined by normal and exchange assays and the Some of the tumors were also characterized by sucrose den results were compared to values obtained with the EIA in order sity gradient analysis. For most of these cases the sedimentation to assess the contribution of occupied sites. No differences were profile showed the presence of both main cytopiasmic molecular found since there was a good correlation between the results forms (9 and 3S). A typical sedimentation pattern is shown in 4279s

Downloaded from cancerres.aacrjournals.org on October 1, 2021. © 1986 American Association for Cancer Research. IMMUNOCHEMICAL AND RADIOLIGAND BINDING ASSAYS FOR ER Conventional low-temperature binding assays do not permit the detection of all occupied ER sites, which may be the case 2000 in premenopausal women and necessitates the use of high- temperature exchange assays. Similar problems occur when o measuring ER in tumors from postmenopausal patients treated o> with the antiestrogen, tamoxifen. Difficulties become even ^ n«31 greater for the estimation of nuclear receptors, most of which I 1000 r =0.93 are occupied by endogenous ligands. Despite the development slope= 1.74 of many exchange methods for the detection of occupied nuclear ER sites without significant loss of binding activity, their rou tine analysis is still a problem for many laboratories. Recently, the assay of ER other than by the interaction with its ligand has been made possible by the production of mono 1000 2000 clonal antibodies against purified human ER. These monoclo nal antibodies are now available as a solid-phase immunochem- HAP (fmol/mg DNA) ical assay. In the present study we have compared the EIA with Fig. 2. Correlation of nuclear ER-EIA values with results obtained by the our conventional 5-point Scatchard analysis by measuring cy HAP exchange assay on 31 human breast cancer specimens. The plot was drawn by linear regression analysis of data. Data on abscissa and ordinate are expressed toplasmic ER in 76 breast cancer patients. Linear regression in fmol of specifically bound steroid per mg of DNA. analysis demonstrated a clear and significant correlation be tween the two methods, which detected similar levels of ER as

1000800eoo4002007.1 shown by the slope value of about 1; moreover, this method is S 4.6S2.8 not influenced by the molecular form of the ER, as demonstra ted by sucrose density analysis performed in parallel on some S"ii of our samples, which yielded ER values very close to those found with the EIA despite the prevalence of one form over the other. \ iiM; We did not find a better correlation with the EIA using the exchange assay than with low-temperature methods, which <8. generally allow the detection only of unfilled sites. A possible explanation is that in our breast cancer patients the fraction of 8 S,'« occupied sites was very low and did not contribute significantly I f <«/ to the total ER levels. In the case of nuclear ER, the overesti- \\"l't Õ\t mation of binding sites obtained with the EIA with respect to 'l 1/'' 1 • the HAP assay was due to the fact that in the nucleus a larger \ /i/ fraction of receptors are expected to be saturated by endogenous steroids. Discrepancies between the two methods could there i\ \ x fore depend on an incomplete exchange with endogenous hor t ^\1 mones or on a temperature-dependent inactivation of the recep 1 \ \ 1\v^ tor during the radioligand binding assay. The EIA is therefore \\VTOP5 a reliable technique for measuring nuclear ER, and it allows a simplification of the experimental procedure with respect to the rather cumbersome HAP exchange assay; however, it ap peared that the total cytosol protein content influenced ER- 1030fraction 15 2O 25 number EIA results outside a certain concentration range. In fact, Fig. 3. Typical sedimentation profile of cytoplasmic ER from human breast cytosol protein values higher than 2 mg/ml caused the EIA to cancer. Cytosols were incubated with 5 nM 17/3-[3H]-estradiol for 4 h at 0-4*C in overestimate ER levels, whereas at low protein concentrations the presence or absence of a 200-fold M excess of diethylstilbestrol. At the end of the protein assay was not always reliable. the incubation, reaction mixtures were transferred onto preformed DCC pellets for removal of unbound and weakly bound ligands. Arrows, sedimentation posi Conversely, in the case of cytoplasmic ER, we would not tions of 14C-labeled 7-globulin and 14C-labeled bovine serum albumin run as recommend the introduction of the EIA for routine use despite internal markers. the good correlation found by Scatchard analysis. In our opin ion, before switching to immunochemical methods, ER data Fig. 3. ER concentrations obtained from sedimentation profiles obtained by the EIA should prove to be at least as reliable as did not differ from EIA counterparts, suggesting indirectly that those obtained with the radioligand binding assay for predicting the monoclonal antibodies recognize but do not distinguish response to hormone therapy and prognosis. between the 9 and 3S receptor. The routine use of the EIA does not actually represent a simpler way to obtain receptor data for management of breast cancer patients, since recent data support the need to measure Discussion also the progesterone receptor (10, 18, 19); unfortunately, the In the last few years, the need for an alternative approach to only way to measure the progesterone receptor is with radioli conventional radioligand binding assays has become particu gand binding assays (20); however, we suggest the use of the larly evident. In fact, results obtained by these methods are EIA for a quick and reliable estimation of nuclear ER, whereas influenced by slight variations in the experimental procedure in the case of cytoplasmic ER we think that until there are data and should not be used by laboratories lacking experience. showing a correlation between EIA results and clinical follow- 4280s

Downloaded from cancerres.aacrjournals.org on October 1, 2021. © 1986 American Association for Cancer Research. IMMUNOCHEMICAL AND RADIOLIGAND BINDING ASSAYS FOR ER up, immunochemical assays should not replace conventional receptor status and prognosis in early breast cancer. Lancet, /: 995-997, 1979. radioligand binding techniques. 9. Crowe, J. P., Hubay, C. A., Pearson, O. H., Marshall, J. S., Rosenblatt. J., Mansour, E., Hermann, R. E., Jones, J. S., Flynn, W. J., McGuire, W. L., and participating investigators. Estrogen receptor status as a prognostic Acknowledgments indicator for stage I breast cancer patients. Breast Cancer Res. Treat., 2: 171-176, 1982. The authors thank B. Johnston for editing and preparing the manu 10. Di Fronzo, G., Cappelletti, V., Coradini, D., Ronchi, E., and Scavone, G. Prognostic significance of progesterone receptors alone or in association with script. estrogen receptors in human breast cancer. Tumori, 70: 159-164, 1984. 11. Tate, A. C.. Greene, G. L., De Sombre, E. R., Jensen, E. V., and Jordan, V. C. Differences between estrogen and antiestrogen-estrogen receptor com References plexes from human breast tumors identified with an antibody raised against the estrogen receptor. Cancer Res., 44: 1012-1018, 1984. 1. Sakai, F., and Saez, S. Existence of receptors bound to endogenous estradici 12. Eldridge, J. S., and Growther, P. Absolute determination of '"I. Nucleonics, in breast cancers of premenopausal and postmenopausal women. Steroids, 22: 56-59, 1984. 27:99-110, 1979. 13. Sfaldarti, G. The attraction of proteins for small molecules and ions. Ann. 2. Greene, G. L., Nolan, C. Engler, P., and Jensen, E. V. Monoclonal antibodies NY Acad. Sci., 51:660-672, 1979. to human estrogen receptor. Proc. Nati. Acad. Sci. USA, 77: 5115-5119, 14. Bradford, M. M. A rapid and sensitive method for the quantification of 1980. microgram quantities of protein utilizing the principle of protein dye binding. 3. Greene, G. L., and Jensen, E. V. Monoclonal antibodies as probes for Anal. Biochem., 72: 248-254, 1976. estrogen receptor detection and characterization. J. Steroid Biochem., 16: 15. Carola, R. E., and McGuire, W. L. An improved assay for nuclear estrogen 353-359, 1982. receptor in experimental and human breast cancer. Cancer Res., 37: 3333- 4. De Sombre, E. R., Greene, G. L., and Jensen, E. V. Estrogen receptors and 3337, 1977. the hormone dependence of breast cancer. In: M. J. Brennan, C. M. McGrath, 16. Burton, K. H. A study of the conditions and mechanism of diphenylamine and M. A. Rich (eds.). Breast Cancer: New Concepts in Etiology and Control, reaction for the colorimetrie estimation of deoxyribonucleic acid. Biochem. pp. 69-87. New York: Academic Press. Inc., 1980. J., 62: 315-323, 1956. 5. Heuson, J. C., Longeva!, E., Mattheiem, W. H., Deboel, M. C., Sylvester, 17. Rice, R. H., and Means, G. E. Radioactive labeling of proteins in vitro. J. R. J., and Leclercq, G. Significance of quantitative assessment of estrogen Biol. Chem., 246: 831-832, 1971. receptors for endocrine therapy in advanced breast cancer. Cancer (Phila.), 18. Horwitz, K. B., and McGuire, W. L. Estrogen control of progesterone 39: 1971-1977, 1977. receptor in human breast cancer: correlation with nuclear processing of 6. Horwitz, K. B., McGuire, W. L., Pearson, O. H., and Segaloff, A. Predicting estrogen receptor. J. Biol. Chem., 253: 2223-2228, 1978. response to endocrine therapy in human breast cancer. A hypothesis. Science 19. Magdelenat, H.. Pouillart, P., Jouve, M., Palangie, T., Garcia Girali, E., (Wash. DC), 189: 726-727, 1975. Bretaudeau, B., and Asselain. B. Progesterone receptor as a more reliable 7. Bishop, H. M., Estone, C. A., Blarney, R. W.. and Haybittle, J. L. Relation prognostic parameter than estrogen receptor in patients with advanced breast ship of estrogen receptor status to survival in breast cancer. Lancet, 2: 283- cancer (letter). Breast Cancer Res. Treat., 2: 195-196, 1982. 284, 1979. 20. Pichón,M. F., and Milgrom, E. Characterization and assay of progesterone 8. Cooke, T., George, D., Shields, R., Maynard, P., and Griffiths, K. Estrogen receptor in human mammary carcinoma. Cancer Res., 37: 464-471, 1977.

4281s

Downloaded from cancerres.aacrjournals.org on October 1, 2021. © 1986 American Association for Cancer Research. Comparison of Immunochemical and Radioligand Binding Assays for Estrogen Receptors in Human Breast Tumors

Giovanni Di Fronzo, Patrizia Miodini, Moreno Brivio, et al.

Cancer Res 1986;46:4278s-4281s.

Updated version Access the most recent version of this article at: http://cancerres.aacrjournals.org/content/46/8_Supplement/4278s

E-mail alerts Sign up to receive free email-alerts related to this article or journal.

Reprints and To order reprints of this article or to subscribe to the journal, contact the AACR Publications Subscriptions Department at [email protected].

Permissions To request permission to re-use all or part of this article, use this link http://cancerres.aacrjournals.org/content/46/8_Supplement/4278s. Click on "Request Permissions" which will take you to the Copyright Clearance Center's (CCC) Rightslink site.

Downloaded from cancerres.aacrjournals.org on October 1, 2021. © 1986 American Association for Cancer Research.