Scepticism regarding the E-screen has also been expressed because different clones of MCF-7 cultured in identical conditions showed distinct differences in the proliferative response to estradiol and to the xenoestrogens, p-nonyl-phenol and bisphenol-A [49], as well as to commercial resin-based dental restorative materials [41]. However, this is a common problem when working with cancer cells, and one can overcome it through meticulously uniform cell stocks. Furthermore, apart from MCF-7 cells, other oestrogen-responsive breast cancer cell lines have been used in the E-screen assay, such as T-47D [43] or ZR-75-1 [50].

The proliferation of the Ishikawa human endometrial cancer cell line has also been proposed to be used for the evaluation of oestrogenic activity [51], but there were indications that the response of this cell line is not specific for oestrogenic molecules [52] in contrast to breast cancer ones. However, Ishikawa cells have been shown to respond to oestrogen and phytoestrogens with a potent induction of alkaline phosphatase (ALP) activity, which is oestrogen specific [53, 54]; hence, it can be used for the screening of potentially oestrogenic compounds [19, 55].

Probably, the most important drawback to the use of both the E-screen and the ALP-induction assays as rapid screening tools is that they are time consuming (the assessment can take from 3 to 6 days, depending on the protocol variation). Accordingly, analysis of oestrogen-regulated gene or protein expression in various cell types can be used as an alternative. For example, expression in MCF-7 cells of the genes coding for the progesterone receptor [56] and for the trefoil peptide pS2 [57, 58] or prolactin production by rat pituitary cells [59, 60] has been proposed as tools to study an oestrogen-specific response. However, these assays are not always as sensitive as the E-screen [32], and they require the use of laborious and/or expensive techniques such as northern blotting or real-time PCR; hence, they are not appropriate for high-throughput screening.

The use of genetically engineered mammalian cell systems was intended to solve some of the above problems. In most of the cases, cells are transfected with an oestrogen-inducible reporter gene, or they are co-transfected with an ER-construct and an ERE-containing reporter gene, similarly to the approaches described above in yeast. The reporter genes usually are designed for measuring CAT or luciferase activity, which due to their high sensitivity offer the possibility to identify even weak oestrogens [15]. The transfection can be transient [61] or stable [19], the latter being more advantageous in terms of reproducibility, as well as rapidity, once the stable line is ready for use [32]. The parental cells used for transfection can be either ER negative, such as HeLa [61] or HEK-293 [19], or ER responsive, like MCF-7 or MG-63 [62]. Consequently, it is clear that apart from the high-throughput capability and the rapidity of these assays (typically gene expression can be assessed within 24 h), their main advantage is their versatility, allowing for separate tests for the various ER subtypes and EREs, recognising both oestrogens and anti-oestrogens and giving the choice of selecting an ER-naive cell context, such as in the case of HeLa cells, or a more physiological context, such as that of MCF-7 cells [15]. Still, these assay systems are artificial, and there are reports regarding the irreversible silencing of the reporter gene after treatment with anti-oestrogens, such as 4-hydroxy-tamoxifen [63, 64].