Endocrine-disrupting compounds (EDC) are predominantly investigated with respect to their ability to mimic or block estrogenic actions. However, it is well-known that EDC can act as agonists or antagonists of androgen- and estrogen-response systems. For that reason, there is an obvious need for bioassays providing the possibility of detecting (anti-)estrogenic and (anti-)androgenic effects. The retinol-binding protein (RBP) seems to be a useful molecular biomarker for assessing all modes of action of EDC, because it is regulated by sex steroid hormones. This study was conducted to establish RBP as a biomarker for determination of (anti-)estrogenic and (anti-)androgenic effects of EDC using a Xenopus laevis primary hepatocyte culture system. It could be shown that RBP mRNA expression in X. laevis hepatocytes was stimulated by estrogens in a dose-dependant manner whereas a combination of estrogen and androgen or estrogen and anti-estrogen treatment suppressed estrogenic stimulating effects. Androgens testosterone and dihydrotestosterone were able to reduce RBP mRNA expression and the anti-androgen vinclozolin could abolish the mRNA synthesis-suppressing activity of the androgen dihydrotestosterone. These results clearly demonstrated that RBP mRNA expression patterns in Xenopus laevis hepatocytes have different modes of (anti-)estrogenic and (anti-)androgenic action and can be used for examination of suspected EDC. Moreover, water samples from sewage-treatment plant effluents were applied to liver cells and expression levels of RBP and estrogen receptor mRNA (a known estrogenic biomarker) were detected. These samples had high estrogenicity but caused low to moderate induction of RBP mRNA synthesis, leading to the conclusion that RBP levels represent the sum of all possible effects (estrogenic and other effects) of EDC in environmental samples.