Issues arising when interpreting results from an in vitro assay for estrogenic activity

Concern about possible adverse effects caused by the inadvertent exposure of humans and wildlife to endocrine-active chemicals, has led some countries to develop an in vivo-in vitro screening program for endocrine effects. In this paper, a previously described estrogen-inducible recombinant yeast strain (Saccharomyces cerevisiae) is used to investigate a number of issues that could potentially lead to the mislabeling of chemicals as endocrine disruptors. The chemicals studied were: 17beta-estradiol, dihydrotestosterone, testosterone, estradiol-3-sulfate, 4-nonylphenol, 4-tert-octylphenol, 4-tert-butylphenol, bisphenol-A, methoxychlor, 2,2-bis(p-hydroxyphenyl)-1,1,1-trichloroethane, butyl benzyl phthalate, 4-hydroxytamoxifen, and ICI 182,780. Alterations in assay methodology (for example, incubation time, initial yeast cell number, and the use of different solvents) did not affect the potency of bisphenol-A and 4-nonylphenol relative to 17beta-estradiol, but did alter the apparent potency of butyl benzyl phthalate. Other issues (including the metabolic activation of methoxychlor, the chemical purity of a steroid metabolite and unusual chemical artifacts observed with alkylphenolic chemicals) which affect data interpretation are described. Many of the issues raised will also affect other in vitro assays for endocrine activity, and some will be relevant to the interpretation of data from in vivo assays. These examples illustrate that considerable care and thought must be applied when interpreting results derived from any single assay. Only by using a suite of assays will we minimize the chances of wrongly labeling chemicals as endocrine disruptors.