0 20 40 60 80 100
% of proliferating cells which are ER+ve prolactin receptor and PR (Brisken et al. 1998, 1999) and may be a protective mechanism to prevent uncontrolled proliferation. Other protective mechanisms reported in the rodent gland implicate transforming growth factor p (TGFp) as a factor limiting ERa+ve cell proliferation (Ewan et al. 2005) and the need for ERa+ve regulated epithelial proliferation to have permissive signals to the stroma via production of amphiregulin which stimulates fibroblast EGFR and production of stromal growth factors (Wiesen et al. 1999; Luetteke et al. 1999). There is also evidence that ER+ve cells acquire the ability to proliferate in human premalignant and malignant lesions, which may be a mechanism whereby the ERa+ve cell can adapt and grow in a relatively low estrogen environment (Shoker et al. 1999; Clarke et al. 1997b). This may be because of the induction of amphiregulin since this important growth factor is increased in the estrogen stimulated normal human mammary gland and in HELU (Wilson et al. 2006; Lee et al. 2005b) On average, HELU were reported to have 27% of ERa+ve cells which were dividing (<5% in normal TDLU), and we have shown that hyperplasia of usual type ADH, DCIS and invasive cancer have increased numbers of these cells (Shoker et al. 2000, 1999; Lee et al. 2005a). It is of interest that treatment of the rodent mammary gland with the carcinogen MNU increases the numbers of ER+ve dividing cells (Sivaraman et al. 2001).
Another potential mechanism for resistance to ED is expansion of a resistant stem cell population within premalignant breast lesions. We have demonstrated that ERa+ve cells have some features of stem or precursor cells (Clarke et al. 2005) and cells which retain DNA radiolabel for long periods (a feature of stem cells) are ERa+ve (Clarke et al. 2005; Zeps et al. 1996). In normal tissue, homeostasis ERa+ve putative stem cells rarely divide, accounting for less than 4% of proliferating cells. However, a rare ERa+ve stem cell division event may be asymmetric, i.e. one daughter cell replacing the ERa+ve stem cell while the other daughter cell is a ERa-ve transit amplifying cell stimulated to further divide by the para-crine signals (discussed above) from the ERa+ve cells. Asymmetric cell division is often regulated by the Delta-Notch signalling pathway in development (Okano et al. 2005). We found that the ERa+ve putative stem cells expressed Musashi, a protein predicted to switch on the Notch signal only in the ERa+ve putative stem cell and not in the ERa-ve transit amplifying daughter cell (Clarke et al. 2003). In ERa+ve breast cancers, where the ERa+ve are highly proliferative, the Notch signalling pathway is highly activated, suggesting that the ERa+ve cells in cancers are symmetrically dividing and giving rise to two ERa+ve daughter cells (Clarke et al. 2003; Styli-anou et al. 2006). We speculate that this may be an early event that occurs in HELU and causes expansion of the ERa+ve cell population via a signalling pathway that may be ED-resistant.
Thus, cellular mechanisms of resistance to ED may include changes in the proliferation status of the ERa+ve cells and its putative stem cell characteristics however. There is no doubt that most ERa+ve tumours in postmenopausal women are responding to the prevailing estradiol concentration since treatment with AIs preoperatively and preoperative withdrawal of HRT reduces cell proliferation within most ER+ve tumours (Dow-sett et al. 2006; Prasad et al. 2003).
Molecular Changes in Response to Estrogen Deprivation
MCF-7 cells are a well-described breast cancer cell-line derived from a patient with a malignant pleural effusion. The cell line is ERa+ve and the ERa+ve cells are capable of dividing and thus experiments on this line may be seen as potentially modelling what happens in tumours and possibly premalignant lesions. In an important experiment, Masamura et al. (1994) showed that MCF-7 cells were able to adapt to a change in estrogen concentration. Wild type MCF-7 cells proliferate maximally at physiological concentrations of estradiol (about 10-9 M). When estrogen was removed from the medium, the cells ceased to divide for 3 months or so but then began to grow again (Fig. 7). A repeat of the estradiol dose response curve at this time indicated that the cells were growing in response to minute amounts of estrogen in the medium and proliferated maximally at 10-13 M concentrations (Fig. 7). These experiments were repeated by Martin et al. (2003) with similar results. Both groups have found that the MCF-7 cells respond to reduced estrogen by increasing nuclear ERa concentration and activity, activation of membrane ER alpha and increased activity of growth factor receptors and activation of the PI3Kinase and MAPKinase signal transduction pathways (Martin et al. 2003; Santen et al. 2004, 2005). Sabnis et al. (2005) also showed increased growth factor receptor and signal transduction factor activity in their MCF-7 cells transfected with the gene for the aroma-tase enzyme and grown in estrogen-depleted conditions, but these cells were not sensitive to low estrogen. Other potential molecular mechanisms of increased cell sensitivity to estradiol include a reduction in NOD1 (da Silva Correia et al. 2006) and loss of nuclear PELP1 (Vadlamudi et al. 2005; Gururaj et al. 2006). Thus, it is possible that precursor lesions in the breast become sensitive to lower concentrations of estradiol, as demonstrated in tumour lines, although at present there is no experimental evidence for this phenomenon in this situation.
The mechanisms of resistance outlined above involve an increase in the number of ER+ve cells, an increase in ER concentration and activation potentially at relatively low E2 concentrations
through signal transduction pathways within the cell. However, there is evidence that the ER may be activated by paracrine pathways from other cells in the breast including adipocytes, immune cells exemplified by macrophages and fibroblasts (Fig. 8). These cell types secrete adipokines, cytokines and growth factors which bind to cell surface receptors, and most have been reported to activate ER by various signal transduction pathways. Because of the difficulty of experiments involving premalignant lesions, most have been performed on ER+ve mammary tumour cell lines or primary tumours. Adipocytes secrete a large number of signalling molecules in response to the metabolic state of the body (Rajala and Scherer 2003). Su-pernatants from adipocytes or co-cultures with adipocytes stimulates mammary tumour cell line growth in-vitro (Iyengar et al. 2003, 2005; Manabe et al. 2003; Chamras et al. 1998). An important adipokine from the viewpoint of breast cancer is the polypeptide leptin which is secreted in response to increasing weight. Several studies have shown that leptin stimulates the growth of mammary tumours via its cell surface receptor, OB-Rb, and receptor-deficient mice do not develop oncogene-induced mammary tumours (Hu et al. 2002; Dieudonne et al. 2002; Cleary et al. 2004).
There is evidence that the ER is activated via the STAT and MAP kinase pathways by leptin (Yin et al. 2004; Catalano et al. 2004) and leptin stimulation can cause resistance to the pure antiestrogen fulvestrant (Garofalo et al. 2004). Leptin also alters the intracellular concentration of the energy-sensing enzyme complex AMP-activated protein kinase (AMPK). In energy-restricted states, AMPK increases in concentration and switches off cell growth via inhibition of the AKT/MTOR pathway (Hardie 2005). However, the opposite occurs in energy-replete states and is a potential mechanism whereby obesity may activate ER and mammary cell growth. Macrophages are found in tumours, premalignant lesions and the normal breast. They may be recruited, in part, because mammary epi-thelia secrete colony-stimulating factor. It is of interest that the breast does not develop in the absence of CSF and inhibition of CSF can cause tumour regression in experimental models (Lin et al. 2002; Aharinejad et al. 2004). Co-cultures of mammary tumour cells with macrophages stimulate growth via production of a variety of cytokines including tumour necrosis factor alpha (TNFa) (Hagemann et al. 2004). Cytokines via their epithelial cell surface receptors activate the transcription factor NFKp which, in turn, can activate ER
Fig. 8 Simplified view of potential pathways for activation of ERa. A aromatase
Fig. 8 Simplified view of potential pathways for activation of ERa. A aromatase and stimulate growth and antiestrogen resistance (Hagemann et al. 2005; Rubio et al. 2006; Zhou et al. 2005; Biswas et al. 2005). Inhibition of NFKp with the polyphenol parathenolide is reported to reverse resistance to antiestrogens (Riggins et al. 2005; de Graffenried et al. 2004). The NFKp system is a potent tumour survival pathway in response to cytokine stimulation. Since estrogen is immuno-suppressive through this pathway, ED alone may activate it. In chronic immunostimulated states such as rheumatoid arthritis, estrogen analogues with immunosuppressive but no growth factor activity are being developed (Keith et al. 2005).
Fibroblasts produce several growth factors capable of stimulating epithelial cell growth, some of which can be shown to activate the ER, including fibroblast growth factors (FGF) (McLeskey et al. 1998; Thottassery et al. 2004) and hepatocyte growth factor (HGF) (Rayala et al. 2006; Zhang et al. 2002). Numerous experiments concerning fibroblast epithelial interactions are summarised in recent reviews of Barcellos-Hoff and Medina (2005) and Haslam and Woodward (2003). Important experiments from the point of view of human fibroblast-epithelial interactions are those of Kuperwasser et al. (2004). They were able to grow normal and premalignant human mammary cells in immune-deprived mice where the mammary fat pad was humanised by transplantation of normal human mammary fibroblasts. Epithelial growth occurred in this system, particularly if the fibroblasts were irradiated before implantation and/or transfected with genes for HGF and TGFp, indicating the importance of stroma for normal and premalignant epithelial cell growth. Thus, the secretions of adipocytes, macrophages and fibroblasts may induce epithelial proliferation by paracrine mechanisms and cause resistance to ED in postmenopausal women (Fig. 8). Some of this stimulation could be related to obesity, and recent studies indicating that obesity is associated with macrophage infiltration and activation may also be relevant to mammary tumour progression (Weisberg et al. 2003).
Was this article helpful?