STEROID RECEPTOR MEASUREMENT IN BREAST CANCERS: COMPARISON BETWEEN LIGAND BINDING AND ENZYME-IMMUNOASSAY IN CYTOSOLIC AND NUCLEAR EXTRACTS Y.A. LUQMANI 1,2, * L. T EMMIM 1 , A. MEMON 1,3 , M.A.A. ALI 1 and A.H. PARKAR 1 1 Kuwait Cancer Control Center, Kuwait 2 Faculty of Allied Health Sciences, Kuwait University, Kuwait 3 Faculty of Medicine, Kuwait University, Kuwait W e have analysed cytoplasmic and nuclear extracts of breast-cancer tissue from a total of 799 patients, measuring both oestrogen and progesterone receptors (ER, PR) using either the ligand binding assay (LBA) or the enzyme immuno- assay technique (EIA). Mean and median receptor levelswere much lower than those widely reported by others. For ER, this may in part be a consequence of the younger median age of the patient group. The frequency of positivity, using consensus cut-off values for clinical evaluation, was also lower than that reported by the EORTC Receptor Study Group. Although the measurements comparing the 2 methods were statistically correlated in terms of positivity, based on the above criteria for clinical assessment, concordance was con- sidered to be relatively poor, particularly for ER when assayed in the same samples by the 2 methods. In cytosolic but not nuclear extracts, the LBA method gave a higher median value for ER than the EIA (except in the group that had EIA values greater than 15 fmol/mg protein); for PR, median valueswere higher with EIA in both cell fractions. There was an excellent correlation between receptor amounts in cytosolic and nuclear extracts for both ER and PR using the EIA; this was significantly better than with LBA. W e also observed a correlation between ER and PR in both cytosolic and nuclear fractions which was most pronounced when the analysis was done by EIA. T he amountsof ER in the cytosolic fraction were also correlated with the those of PR in the nuclear fraction and ER in the nuclear fraction with PR in the cytosolic fraction, but only when the EIA method was used. W e conclude that the EIA method appears to be more sensitive and gives biologically more reliable results. However, the disagreement between the methods may be due to legiti- mate recognition of altered formsof the receptor and may be of biological significance. Although the presence of receptor in the cytosolic fraction isartifactual, itsmeasurement by EIA does parallel the amounts of nuclear receptor, which may be a more relevant biological parameter. Int. J. Cancer 71:526– 538, 1997. r 1997 Wiley-Liss, Inc. The evaluation of steroid receptors in extracts of excised cancer tissue has become a routine procedure for the selection of breast-cancer patients most likely to benefit from endocrine therapy (McGuire and Clark, 1992). The most commonly used quantitative procedure has been the radioligand binding assay (LBA) utilizing 3 H-estradiol (Korenman and Dukes, 1970). This method is labori- ous and requires relatively large amounts of tissue. The result is also dependent on the degree of occupancy of the receptor by endogenous oestrogen; attempts to minimize this by extraction in low-salt buffers are liable to produce variable results and the reported frequency of receptor positivity varies considerably. The isolation of monoclonal antibodies (MAbs) to the oestrogen receptor (ER) (Greene et al., 1980) and to the progesterone receptor (PR) (Greene and Press, 1987) proteins has led to widespread immunohistochemical evaluation. This method has proved quite useful, particularly in demonstrating the nuclear localization of both ER and PR, and several studies have compared this method with the biochemical LBA method (e.g., King et al., 1985; Goussard et al., 1988; Allred et al., 1990). However, the advantages of speed and convenience and the ability to utilize very small biopsies (as only single sections are needed) are outweighed by the subjective nature of assessing immunostaining on a quantitative basis and there is still debate on the use of this technique for routine reporting (Allred, 1993). The enzyme immunoassay technique (EIA) (Leclercq et al., 1986) combines the advantages of using antibody detection with a quantitative assay. It overcomes some of the objections of the 2 preceding methods and it has thus been increasingly adopted for routine analyses. In several studies it has been compared with the immunocytochemical assay (Thorpe, 1987b; Negri et al., 1990; Lerma et al., 1994). It is easier to perform than the LBA and more sensitive in the lower ranges. Epitope recognition should, theoreti- cally, not be affected by the presence of ligand-bound receptor and probably not by minor receptor degradation, either (Greene and Press, 1987; Thorpe, 1988). Several groups have compared the LBA with the EIA method and have reported a reasonable degree of correlation between them (Thorpe, 1988; Leclercq et al., 1986; Jordan et al., 1986; Goussard et al., 1986; Anderson et al., 1988; Foekens et al., 1988; Smyth et al., 1988; Hanna and Mobbs, 1989; Holmes et al., 1990; Jarque et al., 1994). Only cytosolic fractions have been analysed in the vast majority of these studies, with less attention being paid to the nuclear fraction (Romic-Stojkovic and Gamulin, 1980; Thorpe et al., 1986). From immunohistochemical evidence, most, if not all, of the receptor is located in the nucleus of intact cells. Positive correlations between ER amounts in the cytoplasmic fraction (ER c ) and ER concentration in the nuclear fraction (ER n ) have been reported by some (O’Connell et al., 1982; Vandewalle et al., 1983; Thorpe, 1987a) but not all (Romic-Stojkovic and Gamulin, 1980) groups, although ER c measurements are routinely used to assess the amount of cellular ER in breast-tumour biopsies. There are methodological difficulties associated with measurement of ER n , as extensively discussed by Thorpe et al. (1986) and Thorpe (1987a). The high salt extractions that are necessary may affect measure- ment by the LBA technique. The increased time of preparation of nuclear fractions and the use of methods involving hydroxylapatite or nuclear exchange make these unsuitable for routine use. However, the EIA technique, particularly with the use of a Tris buffer system, has been successfully employed for assessment of ER n (Thorpe et al., 1986). In this report we present our experience of routine steroid receptor determinations over a period of 8 years in a hospital laboratory; we compare the LBA and EIA methods for both ER and Abbreviations: ER c and ER n , oestrogen receptor (cytosolic and nuclear); PR c and PR n , progesterone receptor (cytosolic and nuclear); EIA, enzyme immunoassay; LBA, ligand binding assay. *Correspondence to: Kuwait University, Faculty of Allied Health Sci- ences, P.O. Box 31470, Sulaibikhat 90805, Kuwait. Fax: 965 4830937. E-mail: yunus@hsccmail.kuniv.edu.kw Received 15 October 1996; revised 30 December 1996 Int. J. Cancer: 71, 526–538 (1997) r 1997 Wiley-Liss, Inc. Publication of the International Union Against Cancer Publication de l’Union Internationale Contre le Cancer