Background: HER-2/neu, which encodes a receptor tyrosine kinase, is amplified and overexpressed in 20%-25% of human breast cancers. Such tumors are often resistant to hormone therapy. Despite a general inverse association between HER-2/neu amplification/overexpression and estrogen receptor (ER) and/or progesterone receptor (PR) expression, a fraction of patients are both HER-2/neu- and hormone receptor (HR)-positive. The efficacy of hormone therapy in this group is currently a matter of debate. To better understand the relationship between HER-2/neu positivity and HR expression, we analyzed HER-2/neu, ER, and PR as continuous variables in breast cancer cell lines and two cohorts of primary breast cancer patients.
Methods: HER-2/neu and ER/PR expression was analyzed by enzyme-linked immunosorbent assay (ELISA) and enzyme immunoassay (EIA), respectively, in 14 human breast cancer cell lines, some of which had been transfected with the HER-2/neu gene. For the clinical study population, HER-2/neu protein levels were assessed by ELISA (cohort A, n = 665), and HER-2/neu gene copy number was determined using fluorescence in situ hybridization (cohort B, n = 894). ER/PR expression was analyzed by EIA (cohort A) or radioligand binding (cohort B). Associations between HER-2/neu and ER/PR expression were analyzed using Spearman's rho correlation and the chi-square test, and absolute levels were compared using the Mann-Whitney U test. All statistical tests were two-sided.
Results: HR-positive human breast cancer cell lines transfected with the HER-2/neu gene expressed statistically significantly lower levels of ER and PR than parental lines. In the clinical cohorts, levels of HER-2/neu overexpression and gene amplification were inversely correlated with ER/PR levels (Cohort A [n = 112]: for ER, r = -0.34, P<.001; for PR, r = -0.24, P =.010. Cohort B [n = 188]: for ER, r = -0.39, P<.001; for PR, r = -0.26, P<.001). Among patients with HR-positive tumors, HER-2/neu-positive tumors had statistically significantly lower ER/PR levels than HER-2/neu-negative ones (Cohort A: for ER, median = 25 fmol/mg [interquartile range [IQR] = 13-78] versus median = 38.5 fmol/mg [IQR = 17-99] and P =.031; for PR, median = 35 fmol/mg [IQR = 12-119] versus median = 88.5 fmol/mg [IQR = 22-236] and P<.001. Cohort B: for ER, median = 44 fmol/mg [IQR = 13-156] versus median = 92 fmol/mg [IQR = 35-235] and P<.001; for PR, median = 36 fmol/mg [IQR = 13-108] versus median = 84 fmol/mg [IQR = 24-250] and P<.001). Patients with higher levels of HER-2/neu overexpression or amplification had statistically significantly lower levels of ER/PR than patients with lower levels of HER-2/neu overexpression or amplification.
Conclusion: Because absolute HR levels are strongly related to response to hormone therapy in primary and advanced breast cancer, reduced ER/PR expression may be one mechanism to explain the relative resistance of HER-2/neu-positive:HR-positive tumors to hormone therapy.