Aromatase regulation and breast cancer

The aromatase enzyme complex is responsible for the conversion of androgens to oestrogens within the human body. The regulation and inhibition of its activity have attracted considerable research interest in recent years because of its crucial role in regulating oestrogen synthesis in postmenopausal women. Oestrogens support the development and growth of hormone-dependent breast tumours (James & Reed, 1980) and several potent aromatase inhibitors are now used therapeutically (Bhatnagar & Miller, 1999; Miller, 1996). Since the first demonstration that androgens could be converted to oestrogens in peripheral extraglandular tissues (Siiteri & MacDonald, 1973) research has been carried out to identify sites within the body in which the aromatase enzyme occurs and also to identify the factors that regulate its activity. We review the progress that has been made in recent years in understanding the complex tissue-specific regulation of aromatase expression and activity in relation to breast cancer development. Evidence is also presented showing how information about the regulation of aromatase activity has provided an explanation for the earlier observations that obesity, ageing and stress can increase aromatase activity. In addition, we attempt to demonstrate how an understanding of the regulation of aromatase activity in peripheral tissues may implicate the immune system in breast cancer development. Initial studies to determine whether aromatase and oestrogens might show some abnormality in malignant breast tissues started with a number of investigations in which tissue oestrogen concentrations were measured. While such measurements are more difficult to accomplish than plasma assays, the results from several studies were remarkably consistent (Bonney et al., 1983; Van Landeghem et al., 1985; Vermeulen et al., 1986; Pasqualini et al., 1989). Concentrations of the biologically active oestrogen, oestradiol, were found to be consistently higher in malignant than normal breast tissue. Also of importance, the concentrations of oestrogens in tissues were at least 10-fold higher than levels detected in plasma. Surprisingly, concentrations of oestrogens in breast tissues from menopausal women are similar to those from premenopausal women in spite of the much lower levels of these hormones present in blood after the menopause (Thijssen et al., 1987; Blankenstein et al., 1992). The consistent findings of high concentrations of oestrogens in normal and malignant breast tissues vs. plasma levels prompted a number of studies to investigate the origins of oestrogens in these tissues. There are two possible routes by which oestrogens can originate in tissues, namely, uptake from the circulation or synthesis in situ. Oestrogen receptors (ER) would be involved in the retention of oestrogens in target tissues, subsequent to uptake from the circulation. The finding that tumour oestrogen concentrations were similar in ER+ and ER– tumours (Fishman et al., 1977) suggested that formation in situ would be a more important source of tumour oestrogen than uptake from the circulation. Although experiments to differentiate between uptake vs. synthesis in situ are difficult to execute, a number of techniques have been employed, both in human and in animal models, to quantify the contribution that these different routes make to tumour oestrogen concentrations. For postmenopausal women, a double isotopic infusion technique was developed in an attempt to differentiate between these two sources of oestrogen (Reed et al., 1989). The effects of different aromatase inhibitors on oestrogen synthesis in situ were also examined. This isotopic technique involved the infusion of 4–14C-labelled oestrone over a 10–12-h period, prior to operation for breast tumour removal, to ensure that an isotope steady-state was achieved. By isolating and quantifying the levels of 4–14C-oestrone in tumour and plasma, the extent of uptake of oestrone from the plasma can be determined. In order to quantify the extent of oestrogen synthesis in situ, 3H androstenedione was infused together with the 4–14C-oestrone. Aromatization of 3H-androstenedione in breast tumours and in other body tissues results in a small proportion of this isotope being converted to 3H-oestrone. Again, it is possible to isolate and quantify the level of 3H-oestrone resulting from the aromatization of 3H-androstenedione in tissues and in plasma. From a knowledge of the tissue : plasma ratio for 4–14C-oestrone, it is possible to calculate the level of 3H-oestrone that should be present in tissues due to uptake of this isotope from the circulation. It is assumed that any 3H-oestrone present in tissues, in excess of that calculated to be present due to uptake, would be due to synthesis within the tumour. Using this technique in postmenopausal women with advanced breast cancer, synthesis of oestrone in situ from androstenedione was found to account for the major part (84 ± 3%) of the oestrone detected in breast tumours. However, in a small subset of subjects no synthesis in situ was detected. In an attempt to determine whether oestrogen formed in situ as a result of aromatase activity exerted a biological effect, the activity of this enzyme was related to DNA polymerase α activity, which is a marker of cellular proliferation. Overall, a weak but significant correlation (r = 0·57, P < 0·05) was found suggesting that sufficient oestrogen could be synthesized in situ to exert a biological effect (James et al., 1989). Reinvestigation of subjects after treatment with the aromatase inhibitor, 4-hydroxy androstenedione revealed a marked reduction in 3H-oestrone formation in situ. In most breast tumours, a marked decrease in tissue oestrone concentrations also occurred during aromatase inhibitor therapy (Reed et al., 1991). Using a similar double isotopic infusion technique, Miller and colleagues have recently investigated the effect of one of the third generation of aromatase inhibitors, letrozole, on oestrone synthesis in situ (Miller, 1999). Twenty-four women were investigated before and after 3 months of aromatase inhibitor therapy. Evidence for oestrone synthesis in situ was obtained in 30 tumours with the extent of synthesis ranging from 25 to 75% of the total 3H-oestrone detected. Therapy with letrozole resulted in a decrease in oestrone synthesis in situ in all but one of the subjects investigated. Thus, there is now convincing evidence from these human studies that synthesis of oestrone in situ, as a result of aromatase activity, can make an important contribution to the total level of oestrogen present in breast tumours. A number of animal models have also been developed in an attempt to resolve the questions as to whether oestrone synthesis in situ makes a significant contribution to tumour levels and also whether such synthesis is of any biological relevance (Lu et al., 1996). In one of the most elegant studies to date, MCF-7 breast cancer cells that had been transfected with aromatase cDNA (A+) or sham transfected (A–) were inoculated into nude mice (Yue et al., 1998). Three groups of mice were inoculated with A+ cells on one flank and A– cells on the other flank or homoimplants of A+ or A– cells on both flanks. The aromatase substrate androstenedione was injected into all groups of mice for the duration of the study. Little peripheral conversion of androstenedione occurred because insufficient oestrogen was synthesized to support the growth of MCF-7 (A–) cells. In contrast, the A+ cells in the heteroimplant mice grew rapidly with the average weight of A+ tumours being almost eight-fold greater than that of the A– tumours in the same animals. The results from this animal model convincingly demonstrate that oestrogen synthesis in situ predominates over uptake from plasma to support the growth of A+ inoculated tumours. Within the body, the aromatase enzyme complex is located in several tissues including adipose and muscle tissues. Adipose tissue is a major site of the aromatase enzyme with most of the activity being located in the stromal cells which surround the adipocytes (Ackerman et al., 1981). Aromatase activity is also detectable in approximately two-thirds of resected breast tumours (Miller et al., 1990). Immunocytochemical studies to examine the location of aromatase in breast tumours have provided evidence for a stromal (Santen et al., 1994; Sasano et al., 1994) and epithelial (Esteban et al., 1992; Lu et al., 1996) location of this enzyme. Biochemically measured aromatase activity, however, was only found to correlate with immunocytochemically detected aromatase in stromal cells (Santen et al., 1994). Measurements of aromatase activity in fibroblasts derived from breast tumours (tumour fibroblasts), normal tissue proximal to tumours (proximal fibroblasts) or epithelial MCF-7 breast cancer cells revealed a much higher level of aromatase activity in fibroblasts (Singh et al., 1997). Thus, most results suggest that the stromal compartment, rather than the malignant epithelial cells, is the major site for the conversion of androgens to oestrogens in breast tumours. The identification of factors that can regulate aromatase activity has attracted much interest over the last 5–10 years. Considerable progress has been made in understanding the regulation of aromatase activity in normal and malignant breast tissues. This knowledge has provided insight into how factors that stimulate this activity could be associated with an increased risk of breast cancer. When it first became possible to reliably measure the low levels of aromatase activity that are present in normal and malignant breast tissues, there was considerable debate as to whether such low activity could give rise to sufficient oestrogen to support tumour growth (Bradlow, 1982). However, the evidence from human studies and animal models indicating that synthesis in situ was important stimulated a search to identify factors that would support a crucial role for aromatase activity in breast tumour development. A key development in this area was the important finding made by Miller and colleagues that the site of a tumour within the breast influenced aromatase activity in tissue adjacent to the tumour (Miller & O'Neill, 1987; O'Neill et al., 1988). In their study, a series of breasts were obtained from women undergoing mastectomy. The position of the tumours within the breast was noted and aromatase was then measured in the four breast quadrants. In all cases, highest aromatase activity was detected in the breast quadrant that contained the tumour or on which the tumour impinged. This finding has now been confirmed in two further independent studies at the expression and activity levels (Bulun et al., 1993; Purohit et al., 1995) with only one study producing conflicting results (Thijssen et al., 1991). Similar studies have revealed that the location of breast tumours may also influence the activity of other key enzymes involved in oestrogen synthesis within the breast, i.e. oestradiol dehydrogenase and oestrone sulphatase (James et al., 1986; Chapman et al., 1995). The consistent evidence that tumour location is related to increased aromatase activity in adjacent tissues suggests that tumours may arise in regions of high aromatase activity within the breast. Alternatively, a more likely explanation for the relationship detected between tumour location and aromatase activity is that tumours produce and/or contain factors that are able to stimulate aromatase activity. Using breast cyst fluid or conditioned medium (CM) from cultured fibroblasts derived from breast tissues, which had previously been found to stimulate aromatase activity, a number of cytokines, including interleukin (IL)-1, IL-6 and tumour necrosis factor α (TNFα) were identified as potent inducers of aromatase activity (Reed et al., 1992; Macdiarmid et al., 1994). Furthermore, the soluble IL-6 receptor (IL-6sR) was found to markedly potentiate the ability of IL-6 to stimulate aromatase activity in fibroblasts (Singh et al., 1995). Prostaglandin E2 (PGE2) is also thought to be involved in the regulation of aromatase activity in breast tissues (Zhao et al., 1996). It is now well established that expression of the aromatase gene is regulated in a tissue-specific manner and this topic has been the subject of a recent review (Sasano & Harada, 1998). Aromatase transcripts at sites where the gene is expressed contain different 5′ untranslated first exons due to the use of different promoters to regulate expression. In adipose tissue expression is regulated by promoter (P) I.4 (Mahendroo et al., 1993). The 5′ upstream region of this promoter contains a glucocorticoid response element and a GAS (interferon-γ activating sequence) element which can bind transcription factors of the signal transducer and activator (STAT) family (Zhao et al., 1995). Cytokines, in the presence of glucocorticoids, regulate aromatase gene expression via PI.4 (Fig. 1). Recently, expression of aromatase transcripts has been examined in breast tumours and breast adipose tissue from women with breast cancer. In breast adipose tissue from tumour- and nontumour bearing quadrants, there was a reduction in the level of aromatase PI.4 transcripts but an increase in PII and PI.3 transcripts (Utsumi et al., 1996). Expression of the aromatase gene via PII and PI.3 is regulated by cAMP (Mahendroo et al., 1993). Zhao and colleagues have obtained evidence that PGE2 may be the major factor regulating aromatase expression via PII and PI.3 (Zhao et al., 1996). Oestrogen synthesis in breast tissues. Highest levels of aromatase activity are detected in stromal fibroblasts. In malignant breast tissues, aromatase gene expression is thought to be regulated to a greater extent by promoter (P) PI.3 compared to PI.4. PI.3 expression is stimulated by cAMP and factors, such as prostaglandin E2 (PGE2) which increase cAMP levels. In normal breast tissue, aromatase gene expression is regulated mainly by PI.4, expression of which is stimulated mainly by cytokines such as interleukin 6 (IL-6) and tumour necrosis factor α (TNFα). Androstendione (A) is converted to oestrone (E1) by the aromatase enzyme complex. In epithelial cells, E1 is converted to the biologically active oestrogen, oestradiol (E2) by oestradiol 17β-hydroxysteroid dehydrogenase type 1 (E2DH type 1). Because all three promoters appear to be employed, albeit at different levels, in normal and malignant breast tissues, regulation in vivo may depend on the relative concentration of such factors as cytokines and PGE2. The mechanism responsible for the promoter switching that is thought to occur in malignant breast tissues remains to be elucidated. In an attempt to further insight into this and colleagues cultured stromal fibroblasts from normal and malignant breast tissues et that fibroblasts from normal or malignant tissues might be Alternatively, fibroblasts from malignant breast tissue might to different transcription compared to those from normal tissue. investigate these possible examined the response of fibroblasts derived from breast tumours, tissue proximal to tumours and reduction to factors to stimulate expression from PI.4 or The results from these experiments convincingly that fibroblasts from all three of breast tissue a similar It was that in and of fibroblasts was the most likely explanation for the in promoter in different tissues. The results from their study an important in understanding the regulation of aromatase in different breast tissues. However, further investigations are to these results and also to of concentration of different factors in breast tissues. In a similar series of experiments fibroblasts derived from the different of breast tissues, results from other investigations have revealed that is the most potent of aromatase activity in fibroblasts derived from breast tumours IL-6 are the major of aromatase activity in fibroblasts derived from tissues proximal to tumours and reduction et al., 1999). While there is evidence that PGE2 can increase aromatase activity, the fibroblasts used by Zhao et were derived from adipose tissue from women undergoing reduction or reduction (Zhao et al., 1996). of aromatase activity in such fibroblasts is further by the recent finding that PGE2 may in exert part of its effect on aromatase activity via of IL-6 synthesis (Singh et al., 1999). It is that knowledge of IL-6 and PGE2 concentration in normal and malignant breast tissues is in order to the of the tissue-specific regulation of aromatase activity. There is evidence that breast fibroblasts and breast tissue can produce significant of only IL-6 but also et al., et al., 1995). There is also evidence for high levels of PGE2 being detected in malignant breast tissues but this prostaglandin is measurements of its concentrations in such tissue difficult et al., In an attempt to some of the associated with such concentrations of PGE2 were measured in from cultured breast tissue fibroblasts and breast cancer & 1995). significant of PGE2 was detected in from ER+ breast cancer cells but low levels were in from breast fibroblasts. However, in of breast fibroblasts and breast cancer cells, of PGE2 by fibroblasts was almost experiments revealed that this was due to the of PGE2 to PGE2 by Thus, high concentrations of IL-6 and its soluble receptor are likely to be present within malignant breast tissues, the with to PGE2 concentrations of from breast fibroblasts have revealed that cytokines, such as and such as can be by these cells. However, in to their by there is now evidence that cells of the immune such as and may be the major source of aromatase factors (Reed & 1997). to of the of breast tumours can and et al., 1988). important to the role of such cells in aromatase activity from a study in which its activity was measured in breast tissue from a had previously breast by In normal breast adipose tissue, aromatase activity is low 3 et al., 1995). In contrast, much higher levels of aromatase activity were detected in breast tissue from the subject had breast of the breast tissue revealed the presence of tissue with and evidence to support a role for cells of the immune system as a source of aromatase factors was obtained by from stimulated peripheral blood and In the presence of from stimulated and stimulated aromatase activity to a greater than previously et al., 1995). Evidence in support of IL-6 from being an important of aromatase activity has been obtained from experiments in mice with the et al., 1995). In these which oestradiol levels were higher than found in and were associated with a marked increase in IL-6 levels. or of the the effects of that the immune cells are involved in regulating the increased of oestradiol in The identification of cytokines as important of aromatase activity has provided a possible mechanism to account for a number of observations related to the peripheral activity of this enzyme. It is well that peripheral aromatase activity is related to body weight and ageing et al., et al., concentrations of IL-6 with ageing et al., and there is evidence that of IL-6 and are also increased in subjects et al., et al., 1997). Thus, the increase in the conversion of androstenedione to oestrone that is detected in and subjects may be related to in It is well that of the with ageing et al., derived from has a role in regulating of cytokines such as IL-6 by the response et al., 1994). The decrease in may a type response and may account for the increased plasma IL-6 concentration detected in which glucocorticoid can also the of cytokines such as Although there is evidence that stress may be associated with an increased risk of breast cancer et al., 1995). It is possible that at least in part via of IL-6 and increased aromatase activity, could increase breast tissue to oestrogen by such a In subjects with of oestrogen is markedly et al., and plasma concentrations of both IL-6 and are in this et al., 1990). Thus, a effect of these cytokines on peripheral aromatase activity a likely explanation for the high plasma oestrogen concentration detected in subjects with A number of potent aromatase inhibitors are in use as for the treatment of postmenopausal women with hormone-dependent breast cancer. While the results from this type of therapy are the to tumour remains at months (Santen & 1999). It is that there is an for of therapy for the treatment of women with hormone-dependent and breast tumours. that are considerable the oestrogen and of this It has been that is the & 1998). This the growth of breast cancer cells in et al., and the growth of a number of inoculated tumours in vivo in mice et al., 1994). It is thought to by the and malignant cells to Furthermore, it can and is undergoing a as an inhibitor 1998). in the of has been stimulated by studies carried out by and by this have obtained convincing evidence that a reduction in the formation of and is associated with an increased risk of breast cancer et al., 1996). in use for the treatment of advanced breast cancer that such as also to of receptors or of these receptors et al., 1999). The was examined that and related might the ability of cytokines to stimulate aromatase activity. Using fibroblasts derived from breast tumours, and IL-6 stimulated aromatase activity by and (Fig. and almost the ability of cytokines to stimulate aromatase activity et al., 1999). the concentrations the have any significant effect on or Although the extent of resulting from of these cells to PGE2 was lower than that by cytokines, and also PGE2 activity. In to the effects of its which is thought to had a much lower ability to stimulated aromatase activity. For fibroblasts derived from reduction tissue, in the presence of stimulated aromatase activity by In this both and aromatase activity and also stimulated activity. of aromatase in fibroblasts by or Aromatase activity was stimulated tumour necrosis factor α its soluble receptor or prostaglandin E2 and of aromatase activity by IL-6 and PGE2. with et al., 1999). the recent in understanding the of aromatase activity by such factors as cytokines and oestrogen it has possible to a model to suggest how or body weight influence the ability of the immune system to a the development of breast cancer. in recent evidence has to a model in which stromal fibroblasts have a key role in oestrogen from androstenedione in breast tumours. Because these cells are located in to epithelial cells, oestrogen synthesized in fibroblasts could in a manner to stimulate the of epithelial cells. While fibroblasts can produce IL-6 and which could in an manner to aromatase activity, there is evidence that and may be the major source of the immune system is to have to rather than tumour it is difficult to the effects that cytokines have on oestrogen synthesis with their However, the finding that the oestrogen may and PGE2 receptors a possible explanation for this of regulation of aromatase activity in breast fibroblasts. of androstenedione (A) to oestrone as a result of aromatase activity, occurs mainly in the stromal within the breast, with of oestrone to oestradiol (E2) in epithelial tumour cells. and are all important sources of factors that can stimulate aromatase activity including interleukin 6 tumour necrosis factor α (TNFα) and prostaglandin E2 The oestrogen can the ability of IL-6 and PGE2 to stimulate aromatase activity via and It is possible that body factors or to by of may influence the ability of the immune via cytokines, to regulate aromatase activity. In oestrogens can be almost via the or et al., 1996). such as and its activity are potent oestrogens & The formation of or has a role in regulating the activity of oestrogen and the ability of cytokines and PGE2 to regulate tumour aromatase activity. It is now well that women or women on low have a lower risk of breast cancer than women or women high In women or those on low of oestrogens via the predominates with subsequent formation of (Fishman et al., et al., Thus, in such may be synthesized to and PGE2 receptors such that any over of these factors result in oestrogen In contrast, in women or subjects on high that are associated with synthesis of any cytokines or PGE2 within the breast could result in increased of oestrogens and subsequent risk of tumour development. which have been in breast cancer, can also decrease oestrogen but increase et al., 1995). While studies are to further evidence to support this it a explanation for some of the factors that are to the risk of breast cancer. the last a understanding of the factors involved in regulating peripheral aromatase activity has It is now that cytokines such as IL-6 and and PGE2 have important in regulating aromatase activity. the complex by which aromatase activity is regulated in peripheral tissues may to the development of to activity in a tissue-specific A number of potent aromatase inhibitors are now in use but these are associated with a number of Aromatase inhibitors in use be used for the treatment of such as in premenopausal women, due to their effect on it should be possible to or other small that can or receptors involved in of aromatase activity. may in a tissue-specific manner but should also be of for the complex by which aromatase activity is