Differential spatial expression and activation pattern of EGFR and HER2 in human breast cancer

Abstract

Overexpression of HER family members is a well established prognostic factor and identifies potential targets for antibody-based receptor blocking strategies. While several studies have analyzed the expression of HER2 and other HER-family members in malignant tumors, considerably less is known about their expression and activation in non-involved breast tissue from breast cancer patients. We have therefore investigated the differential expression of EGFR, HER2, and their tyrosine-kinase activated forms (ptyr-1248 Her-2 and ptyr-845 EGFR) in 63 tumor specimen containing: a) malignant epithelium, b) in non-malignant tissue located at the peritumoral margin, and c) in uninvolved breast tissue obtained from tissue distant from the tumor. Using immunohistochemistry (IHC), we found significantly higher HER2 protein expression levels in malignant epithelium than in marginal and peripheral non-malignant epithelium (p=1.3 x 10(-10) Fisher's exact test). Epithelial EGFR expression did not differ between the three tissue types, but stromal EGFR protein was significantly more common in marginal and peripheral tissues when compared to tumor tissues (p=0.008, Fisher's exact test). When analyzing activated receptor forms, we found epithelial ptyr-1248 HER2 expression in one tumoral, one marginal and one peripheral sample. We did not observe ptyr-845 EGFR in any of the samples analyzed. We found a significant overall correlation between epithelial and stromal EGFR expression (r=0.442; p<0.0001; Spearman's Rho), and between stromal EGFR expression and normal tissue type (r=0.170; p<0.02; Spearman's Rho). Epithelial HER2 expression and normal tissue type (r=0.492; p<0.0001; Spearman's Rho) were inversely correlated. Taken together, we have observed a differential expression pattern of EGFR, HER2, and activated HER2 that is dependent on the spatial relation to a malignant tumor. Our findings of decreased intratumoral EGFR expression and the absence of activated EGFR suggests that, in contrast to HER2, EGFR inhibition might not be an ideal target for antibody therapy.