Radioimmunohistochemistry of Epidermal Growth Factor Receptor in Breast Cancer

Abstract

Context.—Conflicting reports of epidermal growth factor receptor (EGFR) expression in breast cancer and inconstant relationships with established prognostic indicators and outcomes suggest difficulties with EGFR measurement. Objective.—To compare EGFR measurement in a panel of cell lines and in breast carcinomas by radioimmunohistochemistry (R-IHC), conventional immunohistochemistry (IHC), and a ligand binding (LB) assay. Design.—Eight EGFR-expressing cell lines and 50 primary breast carcinoma specimens were analyzed for EGFR by IHC, R-IHC, and LB assays. A further 153 primary breast cancer specimens were analyzed by R-IHC alone. Results.—All 3 assays were in good agreement for the cell lines. In the subset of the 50 carcinoma specimens, EGFR was detected by LB assays in 19 (38%) and by IHC in 24 (48%). However, R-IHC detected EGFR in 46 (92%) of 50 and in 186 (92%) of all 203 carcinoma specimens. The LB assay agreed poorly with R-IHC of carcinomas, possibly because the LB assay is sensitive to EGFR-expressing nontumor breast parenchyma in the tissue analyzed. Both IHC and R-IHC on carcinoma specimens agreed better, but 26 carcinoma specimens (52%) in which EGFR was not detectable by IHC had a 10-fold range in receptor level detectable by R-IHC. Conclusion.—To elucidate the role of EGFR or other growth factor receptors in breast cancer requires accurate, sensitive receptor assays. With its dynamic range, R-IHC returned meaningful results over the entire range of expression actually present in breast cancer, which LB assays and IHC failed to do.

Supplementary Methods, Tables 1-6, Figures 1-5 from Signal Transduction Networks in Cancer: Quantitative Parameters Influence Network Topology

“…The relationships between immunoblot intensity was assumed to be linear as the range in observed intensities in ErbB1 (i.e., O(10 3 )) corresponds to the range in observed ErbB1 expression using a ligand binding assay. 39 The same conversion factor was used to estimate the relative measurement of Grb2, Shc (p45 and p52 isoforms), and Irs-1 expression among cell lines. The predicted dynamic ranges in protein expression were adjusted such that the lowest measured value for a protein corresponds to the estimated detection limit.…”