Proliferation assays for estrogenicity testing with high predictive value for the in vivo uterotrophic effect

Proliferation assays based on human cell lines are the most used in vitro tests to determine estrogenic properties of compounds. Our objective was to characterise to what extent these in vitro tests provide alternatives for the in vivo Allen and Doisy test, a uterotrophic assay in immature or ovariectomised rodents with uterus weight as a crucial read-out parameter. In the present study four different human cell lines derived from three different female estrogen-sensitive tissues, i.e. breast (MCF-7/BOS and T47D), endometrial (ECC-1) and ovarian (BG-1) cells, were characterised by investigating their relative ERα and ERβ amounts, as the ERα/ERβ ratio is a dominant factor determining their estrogen-dependent proliferative responses. All four cell lines clearly expressed the ERα type and a very low but detectable amount of ERβ on both the mRNA and protein level, with the T47D cell line expressing the highest level of the ERβ type. Subsequently, a set of reference compounds representing different modes of estrogen action and estrogenic potency were used to investigate the proliferative response in the four cell lines, to determine which cell line most accurately predicts the effect observed in vivo. All four cell lines revealed a reasonable to good correlation with the in vivo uterotrophic effect, with the correlation being highest for the MCF-7/BOS cell line (R²=0.85). The main differences between the in vivo uterotrophic assay and the in vitro proliferation assays were observed for tamoxifen and testosterone. The proliferative response of the MCF-7/BOS cells to testosterone was partially caused by its conversion to estradiol by aromatase or via androstenedione to estrone. It is concluded that of the four cell lines tested, the best assay to include in an integrated testing strategy for replacement of the in vivo uterotrophic assay is the human MCF-7/BOS breast cancer cell line.