Comparative measurement of progesterone receptors in breast cancer by biochemical and immunoenzymatic assays.

The measurement of progesterone receptors (PR) by enzyme immunoassay (Abbott Laboratories, EIA monoclonal) and biochemical assay using a tritiated ligand (promegestone, R5020) was studied and compared by using the statistical method of Passing and Bablok. In order to improve the reliability of the biochemical method, data were analysed using a hyperbolic model which avoids the need to determine non-specific binding experimentally. The comparison of hyperbolic (Y) and Scatchard (X) plots gave a regression curve of Y = 0.93 x + 1.34 fmol/mg of protein. Cytosols from 70 human breast cancers homogenized in the absence of KCl were assayed for PR by both the EIA (Y) and biochemical (X) methods. The linear regression obtained gave Y = 1.21 X + 1.97 fmol/mg of protein. In the presence of 0.4 M KCl-Tris buffer, the corresponding result for 80 human breast cancers was Y = 3.11 X + 1.91 fmol/mg of protein. The slopes of the two regression lines obtained in the presence or absence of KCl were significantly different. Results of the biochemical and EIA methods were similar in the absence of KCl, whereas EIA gave higher values when KCl was used. This discrepancy probably stem from the methodological differences between the two methods: the biochemical assay measures active steroid binding sites while EIA measures antigenic activity. The authors conclude that clinical studies are required before using high-salt extraction buffer in routine PR determination by the EIA method; this will result in improvements in the determination of the hormone dependence and/or the prognosis of human breast cancers.