1.  INTRODUCTION

       We provide an initial characterisation of two new mammalian ETS transcription factors, ELF3 and ELF5. Expression of ELF3 and ELF5 appears to be restricted to the epithelial cells of multiple organs, and we are examining their role in normal mammary differentiation and function, and in mammary neoplasia. We show evidence that both ELF3 and ELF5 are able to positively regulate transcription of the whey acidic protein (WAP) promoter in mammary epithelial cells, independently of hormone treatment.

1.1  Background

       One of the major milk proteins produced by mammary epithelial cells during pregnancy and lactation is whey acidic protein (WAP). WAP expression is very low until mid-gestation, after which time the combined actions of the glucocorticoid and prolactin pathways produce a several thousand-fold increase in transcription. Transcription is also partly dependent upon the binding of factors to the mammary-cell activating factor (MAF) site 1,2, a conserved ETS-like element present in the WAP proximal promoter and in the regulatory regions of other mammary specific genes.

     The ETS family of transcription factors regulate gene expression during normal biological processes such as haemopoiesis, angiogenesis and cartilage/skeletal development, and aberrant forms have also been implicated with various types of neoplasia. Despite the large variety of ETS factors, functions in epithelial cells, such as those in the mammary gland, have not been examined extensively. Overexpression of PEA3 is, however, associated with aggressive subsets of breast carcinoma that also overexpress erbB2 . PEA3 does not appear to be involved in normal mammary function, but may be required for early mammary development.We have recently cloned two new ETS transcription factors, ELF3  and ELF5 . The HUGO Nomenclature Committee has upheld the ELF naming convention despite ELF3 being cloned and differently named by other groups (ESX  , ESE-1, jen  , ERT  ).

2. EXPRESSION OF ELF3 AND ELF5
      
      We examined the expression of ELF3 and ELF5 in mouse organs by Northern blot analysis of poly A⁺ RNA. The expression patterns of these two genes were similar, but both differ significantly to the expression of other ETS family members. Most other ETS factors are expressed strongly in haemopoietic compartments, but ELF3 and ELF5 levels were undetectable in adult thymus or spleen (Figs la and 2a), nor in peripheral lymphocytes of leukaemic cell lines (data not shown). Strong expression, however, was observed in a subset of organs that contain secretory epithelial cells, such as the lung, stomach, prostate and mammary gland.
       The mammary gland undergoes distinct phases of growth, differentiation and regression; from virginal mice to pregnancy, parturition and weaning. Examination of expression levels through these stages revealed that ELF3 is present in virginal mammary gland and during early pregnancy, when cells are proliferating, but drops in late pregnancy when epithelial cells undergo terminal differentiation and begin to produce milk components (Fig 1b). Expression remains low during lactation, but reappears upon weaning, when most of the epithelial cells in the mammary gland are undergoing apoptosis. Another group’s study made similar findings 10 . In situ expression analysis in human mammary gland confirms that ELF3 is expressed specifically in the epithelial cells of the ductules and lobular structures (Fig 1C).
          One possible explanation for the pattern of ELF3 expression is that it is associated with a small proportion of epithelial stem cells, rather than the differentiated functional cells, and that these cells survive apoptosis to re-colonise the mammary gland during the next pregnancy.

     ELF5 is also expressed specifically in human mammary gland epithelial cells (Fig 2b), and, in two time points, mammary ELF5 expression is increased during pregnancy in the mouse (Fig 2a), suggesting that the temporal pattern of ELF5 expression may be very different to that of ELF3.

3.  FUNCTION OF ELF3 AND ELF5 IN MAMMARY GLAND

     Expression of the WAP gene has been studied extensively as it is highly specific to mammary epithelial cells and is induced several thousand fold during late pregnancy and lactation. The WAP promoter responds to lactogenic hormones, but also contains other important transcription factor elements that are essential for expression. In vitro studies indicate that the MAF site is critical for maximal WAP expression, independently of hormone treatment 1. Mouse transgenic studies indicate that the site is important for expression of WAP during pregnancy, but not critical during lactation, when the promoter is most active and under the control of hormones2.The MAF site is similar to others described as binding sites for the ETS family of transcription factors. We have investigated the ability of ELF3 and Recombinant ELF3 and ELF5 proteins are able to strongly interact with the ETS/1 site, and also more weakly with the ETS/2 site, in electrophoretic mobility shift assay (data not shown). Consistent with the binding, we found that both ELF3 (fig 3a) and ELF5 (fig 3b) were able to transactivate a WAP promoter construct in a mammary epithelial cell line, although ELF3 was much more potent than ELFS. Both factors were able to function independently of lactogenic hormones. A longer promoter construct, containing more distal hormone response elements, behaved similarly (data not shown).Point mutations (fig 3c) of either ETS/1 or ETS/2 alone were enough to disable most transactivation, indicating the ELF3 and ELF5 operate through these sites. Both mutations together had only a slightly greater effect, suggesting that the ETS/1 and ETS/2 sites work cooperatively, both being required for maximal promoter activity.

4. CONCLUSION

    Expression of mammary-specific genes is under the control of both hormone dependent and independent mechanisms, neither of which may be totally mammary-specific. It seems that a complex interplay of transcription factors may be responsible for mammary expression. Our challenge is to determine whether ELF3 and/or ELF5 are important for the expression of mammary-specific genes in vivo, and whether they cooperate with other stimuli, such as prolactin. Because the mammary gland is not a static tissue, but can continually undergo cycles of growth, differentiation and function, it will be interesting to examine whether ELF3 and ELF5 operate cooperatively or differentially during these different phases.