Detection of breast tumor progesterone receptor isoforms with the PR-enzyme immunoassay kit (Abbott). Effect of heat shock proteins on epitope recognition.

Eighty-nine human breast tumor progesterone receptors were assayed both by a radioligand assay (RLA; 3H-ORG-2058) and by the enzyme immunoassay from Abbott Laboratories (PR-EIA). The correlation obtained between the two methods was EIA = 0.83 RLA + 4.1 fmol/mg protein (r = 0.83). Great discrepancies were observed with EIA/RLA ratios varying from 0.5 to 2. After KCl 0.4 mol/l dissociation and chromatographic separation of 8 PR isoforms [12], the two PR polymeric forms (isoforms 1 and 2) which remained bound to the hsp90 heat shock protein were not or only partially detected by EIA, whereas PR-hsp70 isoforms were highly detected with EIA/RLA ratios increased up to 3.8. Free PR-A and PR-B proteins and the PR-truncated form (52 kDa) were never detected by EIA. Thus, the final result of PR assay using the Abbott method depends directly on the amount of PR-hsp70 isoforms produced through KCl dissociation during the overnight incubation of PR with the KD68 coated beads.

Chromatographic separation of eight progesterone receptor isoforms in human breast tumors, and detection by radioligand and monoclonal antibodies. Association with hsp90 and hsp70 heat shock proteins.

It has been shown previously that untransformed molybdate-stabilized breast cancer progesterone receptor (PR) complexes can be dissociated by KCl (0.4 to 1M) into eight different intermediate forms, or isoforms (680-600-361-224-193-119-88-52 kDa), separated by high performance size exclusion chromatography and with the use of computer assisted smoothing and deconvolution procedures (H Cren et al [1993] J Chromatogr 615, 23-36). The purpose of this work was to study the constitution of each isoform by using different monoclonal antibodies (mabs) raised against PR-A/B (JZB39 and KD68), against PR-B (PR6 and KC 146), and against hsp90 and hsp70 heat shock proteins (9D2 and Ab72, respectively). The differential recognition of nontransformed molybdate-stabilized PR isoforms by either radioligand (RLA, 3H-Org2058) or by an enzyme immunoassay (PgR-EIA Abbott) showed the presence of two different PR isoforms in the non-dissociated PR heteropolymeric peak eluted from a TSK-3000 SW column. After PR dissociation by 0.4 M KCl and interaction of PR isoforms with the different mabs, the presence of PR-A, PR-B, hsp90 and hsp70 was studied. Results showed that hsp90 was present in isoforms 1 (680 kDa), 2 (600 kDa) and 3 (361 kDa) exclusively, whereas hsp70 remained strongly bound to isoforms 4 (224 kDa) and 5 (193 kDa). Isoforms 6 (119 kDa) reacted with PR6 antibody and represented the PR-B protein, whereas isoform 7 (88 kDa) represented PR-A protein. Isoform 8 (52 kDa) was not detected by mabs and represented a truncated form of PR. Detection of isoform 1 either by RLA or by EIA showed ratios EIA/RLA approximately 1 or 2, and these values suggested that this heteropolymeric form may contain a dimeric structure. From these observations, a model is proposed for the composition of each PR isoform obtained from dissociation of breast tumor PR.

Use of automated smoothing and deconvolution procedures for the determination of human breast cancer estrogen and progesterone receptor isoforms, after high-performance size-exclusion chromatography.

Automated smoothing and deconvolution procedures were used to analyse complex chromatographic patterns of human breast cancer estrogen and progesterone receptors, obtained by size-exclusion chromatography. By injecting, first, different known amounts of a radio-iodinated protein in a TSK G-3000 SW column, constructed complex chromatograms (twelve chromatograms) with known peak positions were obtained, and were further treated by mathematical functions to determine a smooth-deconvolution strategy, which could be used with unknown chromatographic patterns. The determination of peak areas by a "curve-fit" program showed a good correlation with the amounts of protein injected (r = 0.91). Human breast cancer estrogen receptors (56 cases) and progesterone receptors (45 cases) were chromatographed in a TSK G-3000 SW column, and further analysed with the smoothing and deconvolution procedures: eight different estrogen receptor isoforms were detected with molecular masses ranging from 530,000 [Stoke's radius (Rs) = 7.7 nm] to 23,000 (Rs = 2.6 nm), and eight progesterone receptor isoforms were observed with masses ranging from 680,000 (Rs = 8.6 nm) to 50,000 (Rs = 3.1 nm). The dissociation effect of KCl (0 to 1 M) on receptor structure yielded different proportions of receptor isoforms, but did not modify their different peak positions.

Evolution of immunoreactivity of monoclonal antibodies H222 and/or D547 used in the detection of breast cancer estrogen receptors. Varying reactivity of receptor isoforms.

From 1984 to 1990, human breast cancer estrogen receptors have been measured both by a radioligand assay (RLA[3H]estradiol) and by an enzyme immunoassay (Abbott ER-EIA kit). The ratio EIA/RLA results increased continuously from 1.04 (1984) to 1.87 (1990), and this evolution was consistent with the last trial of the E.O.R.T.C. receptor study group (Trial 1989-II, EIA/RLA = 2.5). Dilution studies of cytosols with the current ER-EIA kits showed an important parallelism defect of the standard curve, the final result of cytosols (fmol/mg protein) obtained from the upper part of the curve (between 100 and 500 fmol/ml) being 1.5 to 2 times higher than the results obtained from readings of the lower part of the standard curve (between 0 and 50 fmol/ml). Chromatographic experiments were carried out during 1986 and the measures of binding sites by RLA and of immunoreactive sites by EIA on chromatographic fractions were compared. Identical results were obtained with EIA and RLA, either on polymeric forms of the estrogen receptor, or on monomeric forms obtained after dissociation by 0.4 M KCl. The same experiments performed during 1990 showed that, in the chromatographic fractions, the concentration of immunoreactive sites was twice as large as that of ligand-binding sites, detected by tritiated estradiol. Furthermore, the detection of polymeric and monomeric receptor isoforms by monoclonal antibodies varied, and was increased by the presence of KCl (0.4 M) and/or bovine serum albumin (BSA) (1 mg/ml) in the cytosol. These findings showed that the large differences between enzyme immunoassay and ligand-binding assay results currently observed were due to differential reactivity of monoclonal antibodies for the estrogen receptor standard provided in the ER-EIA kits and for the estrogen receptor present in cytosols from human breast cancers, suggesting modifications of immunoreactivity of the monoclonal antibodies actually provided in the ER-EIA kits.

Immunoradiometric assay of pS2 protein in breast cancer cytosols.

We studied the ELSA-pS2 immunoradiometric kit (CIS Bio International) for pS2 protein assay in breast cancer cytosols according to classic validation methods. In addition, we studied correlations between pS2, steroid receptors, and cathepsin-D assays. Repeatability (CV = 1.5% to 4.8%) and reproducibility (CV = 1.6% to 4.9%) were good. The results were linearly related to pS2 concentrations between 205 and 2200 ng/L; the detection limit was 40 ng/L. The accuracy of the assay was measured by assessing recovery; analytical recoveries were near 100% throughout the standard curve. The use of different compounds for cytosol preparation (Tris 10 mmol/L or phosphate 25 mmol/L, KCl 0.4 mol/L, bovine serum albumin 1 g/L) had no effect on pS2 results. pS2 was assayed in breast tumor cytosols from 197 postmenopausal and 92 premenopausal patients. The mean value was 24 micrograms/g of protein; the median and 25th and 75th percentiles were 6, 1, and 23 micrograms/g protein, respectively. We observed a relation between concentrations of pS2 and those of estrogen and progesterone receptors, but there was no relationship between the concentrations of pS2 and cathepsin-D.

Estradiol and progesterone receptors in cultured normal human breast epithelial cells and fibroblasts: immunocytochemical studies.

The estradiol (E2) and progesterone (P) receptors (ER and PR) were studied in normal human breast epithelial (HBE) cells and fibroblasts cultured separately in our laboratory from surgical reductive mammoplasty samples. Immunocytochemical studies were performed on cytospun cells using the anti-ER antibody H222 Sp gamma and the anti-PR antibodies JZB39 and KD68. A specific immunostaining was observed for ER and PR in HBE cells. This immunostaining was nuclear, varying from cell to cell in positivity and intensity of staining. Moreover, ER and PR immunostaining was hormone-modulated: it increased in E2-treated cells and decreased after addition of the progestin R5020. In fibroblasts, a weak ER immunostaining and a stronger PR immunostaining could be observed; however it was not modified by either E2 or progestogen treatment. Thus, in normal breast epithelial cells, E2 stimulates both its own receptor and PR, whereas the progestin R5020 lowers ER and PR content. In contrast, ER and PR content in normal breast fibroblasts seem to be independent of E2 or P action.