Acquired resistance to oestrogen deprivation: role for growth factor signalling kinases/oestrogen receptor cross-talk revealed in new MCF-7X
http://www.100md.com
《肿瘤与内分泌》
Tenovus Centre for Cancer Research, Welsh School of Pharmacy, Cardiff University, Redwood Building, King Edward VII Avenue, Cardiff, UK
This paper was presented at the 1st Tenovus/AstraZeneca Workshop, Cardiff (2005). AstraZeneca has supported the publication of these proceedings.
Abstract
In vitro models of long-term oestrogen deprivation utilise increased oestrogen receptor (ER) and are oestrogen hypersensitive, with emerging evidence that growth factor signalling contributes and interacts with ER. However, such models are commonly derived in the presence of serum growth factors that may force the resistance mechanism. Our new in vitro model, MCF-7X, has thus been developed under conditions of both oestrogen and growth factor depletion. ER expression, serine 118 phosphorylation on this receptor and its transcriptional activity were modestly increased compared to the parental MCF-7 cells, although MCF-7X cells were not oestrogen hypersensitive. Faslodex (0.1 μM) partially decreased ER and its transcriptional activity, with associated decreases in serine 118 phosphorylation. Faslodex inhibited MCF-7X growth by 50% for 10 weeks. Classical growth factor receptors did not impact on MCF-7X growth and only a modest contribution for MAP kinase was revealed using PD98059 (25 μM; 35% inhibition for 3 weeks). However, the phosphatidylinositol-3-OH (PI3)-kinase inhibitor LY294002 (5 μM) inhibited MCF-7X growth by 65% for 10 weeks. In contrast to PD98059, LY294002 also partially-inhibited ER transcriptional activity and decreased serine 167 ER phosphorylation. Co-treatment with faslodex plus LY294002 to decrease activity of both serine 118 and 167 proved superior vs the single agents in decreasing ER transcriptional activity and MCF-7X growth (90% inhibition for 25 weeks). However, triple treatment including PD98059 was required to prevent resistance in MCF-7X, an event dependent on maximal depletion of serine 118 phosphorylation and ER transcriptional activity. Kinases clearly contribute in resistance to oestrogen deprivation, cross-talking with ER signalling via AF-1 phosphorylation. While inhibiting each pathway has potential to treat this state, combined therapy targeting all regulators of ER phosphorylation may be required to block subsequent emergence of resistance.
Introduction
Oestrogen deprivation, achieved clinically in post-menopausal women using aromatase inhibitors or in premenopausal women using ovarian ablation or suppression with luteinizing hormone-releasing hormone superagonists, is of considerable value in breast cancer management. Indeed, emerging data indicate that ‘third generation’ aromatase inhibitors may be superior to tamoxifen in postmenopausal women (Santen 2003). Thus, such inhibitors provide an effective second-line strategy in advanced postmenopausal oestrogen receptor positive (ER+) breast cancer. They can also improve tumour response compared to tamoxifen when used first-line in advanced ER+ disease. Moreover, the updated analysis of the ATAC trial in early breast cancer has shown that anastrazole as a primary adjuvant therapy significantly improved disease-free survival, increased time to recurrence and reduced contralateral invasive breast cancer incidence versus tamoxifen in ER+ postmenopausal disease. Use of aromatase inhibitors following completion of tamoxifen treatment also appears promising (Winer et al. 2005).
Importantly, however, while these data are encouraging, it is evident that oestrogen deprivation is unlikely to provide the whole answer to effective treatment of breast cancer. As with other endocrine strategies, acquisition of resistance remains a significant problem with aromatase inhibitors, where improvement in relapse-free survival, compared with tamoxifen, in the adjuvant setting remains modest and relapse rates in advanced disease are still often measured in months rather than years (Nicholson et al. 2004a, b). It is clearly essential that we elucidate the mechanisms underlying acquired resistance to oestrogen deprivation so that we might intelligently approach treatment or prevention of this state in the future.
Experimental studies focussing on acquisition of resistance to oestrogen deprivation are now emerging, although models remain relatively sparse. Various long-term oestrogen deprived ER+ human breast cancer xenograft models in oophorectomised nude mice have been described previously (Shim et al. 2000). Furthermore, response and resistance specifically to aromatase inhibitors is being examined using mouse xenografts of aromatase-transfected MCF-7Ca breast cancer cells (Brodie et al. 2003). Long-term oestrogen deprived ER+ in vitro models (Martin et al. 2003, Santen et al. 2004) have also been developed from endocrine responsive cell lines, including MCF-7. Such cells are initially growth inhibited in vitro by oestrogen deprivation achieved under phenol red-free, charcoal-stripped serum conditions that deplete the exogenous oestradiol concentration to 10–13 M. However, resistant growth emerges following long-term oestrogen deprivation. A unifying feature of these models of acquired resistance to oestrogen deprivation appears to be a retained mitogenic role for ER (Martin et al. 2003, Santen et al. 2004). Such models commonly exhibit increased ER expression and can be growth inhibited by the pure antioestrogen faslodex. Interestingly, long-term oestrogen deprivation has also invariably been associated with development of hypersensitivity to oestrogens, where exquisitely low residual steroid hormone levels now appear sufficient to support tumour cell growth. Indeed, Santen et al. (2004) have shown oestradiol concentrations as low as 10–14 M maximally stimulate growth of their ER+ long-term oestrogen deprived sub-line, in contrast to higher concentrations required in MCF-7 cells.
Various groups are examining the underlying biology of resistance to oestrogen deprivation and its apparent oestrogen hypersensitivity. Some are examining the role of altered ER structure/function while others are monitoring the impact of aromatase content, although the relevance of these parameters to clinical resistance remains obscure (Fuqua et al. 2000, de Jong et al. 2003). However, a potentially important mitogenic contribution in resistance to oestrogen deprivation is growth factor signalling (Nicholson et al. 2004b). There is supportive evidence from gene transfer studies for a causative association between altered growth factor signalling and oestrogen independent growth (Daly et al. 1991). Further evidence has arisen from other forms of antihormone resistance, notably antioestrogen resistant clinical disease and oestrogen receptor negative states, as well as acquired antioestrogen resistant models. These data cumulatively suggest that increased activation of growth factor signalling pathways, either through enhanced availability of growth factor ligands or via up-regulation/increased activation of their receptors or recruited kinases, can promote antihormone resistance (Nicholson et al. 2004a, b). While such aberrations may be present de novo, adaptive events occurring in such pathways during therapy may also promote the acquired resistant state. As proof of principle, our acquired tamoxifen resistant cell line recruits increased epidermal growth factor receptor (EGFR)/HER2 signalling that emerges during the antihormone responsive phase for its growth (Knowlden et al. 2003). Such signalling appears to promote ER activity in these antioestrogen resistant cells, an event that serves to enhance tamoxifen agonism (Nicholson et al. 2004a). Clearly, if emergent during long-term oestrogen deprivation, adaptive increases in growth factor signalling elements might also allow growth in the presence of very low oestrogen levels and promote oestrogen hypersensitivity. Interestingly, therefore, various models of acquired resistance to oestrogen deprivation have implicated insulin-like growth factor receptor (IGF-1R), HER2 and downstream activation of MAP kinase (MAPK) and phosphatidylinositol-3-OH kinase (PI3K)/AKT, with emerging evidence of such growth factor elements interacting with ER (Stephen et al. 2001, Martin et al. 2003, Santen et al. 2004).
However, most in vitro models employed to date have been developed in the presence of serum growth factors, with in some cases further insulin supplementation. This availability of exogenous growth factors could feasibly force the acquired resistance mechanism and promote oestrogen hypersensitivity. Indeed, in the model from Martin et al. (2003) addition of insulin appeared to promote IGF-1R-mediated super-sensitisation to residual oestrogens. To further decipher resistance to oestrogen deprivation, we have therefore developed a new ER+ acquired resistant in vitro model, MCF-7X (Nicholson et al. 2004b), under conditions of parallel depletion of oestrogen and exogenous growth factors. Using this model we have questioned (i) whether resistance to oestrogen deprivation can still arise in the absence of high levels of exogenous growth factors; (ii) whether ER remains critical to growth and if the resistant phenotype is oestrogen hypersensitive; (iii) whether there is evidence at the receptor/ligand or (iv) intracellular kinase level for growth factor signalling pathway elements contributing to growth; (v) whether such signalling cross-talks with ER and (vi) whether the growth mechanism can be effectively targeted. It is hoped that by further understanding of the inherent resistance mechanism we will be able to determine effective treatment strategies for acquired resistance to oestrogen deprivation.
Acquired resistance to oestrogen deprivation can arise under conditions of exogenous growth factor depletion (MCF-7X cells)
MCF-7 cells were grown long-term in white RPMI medium containing 5% charcoal-stripped, heat-inactivated (65 °C, 40 min) foetal calf serum (termed ‘X’ medium), an approach which depletes oestrogens and exogenous growth factors yet still ensures cell viability and attachment (van der Burg et al. 1988). After an initial phase of >80% growth inhibition for 4 months, resistance was acquired with proliferative activity of the resultant MCF-7X sub-line approaching that of parental MCF-7 cells prior to treatment. Clearly, high levels of exogenous growth factors are not essential for development of resistance to oestrogen deprivation, with input from any residual steroid hormone and presumably predominantly autocrine signalling adequate. This new cell model provides a sobering example of the adaptability of ER+ human breast cancer cells under conditions of substantially depleted exogenous mitogens.
ER signalling is prominent in MCF-7X cells and promotes growth, but oestrogen and exogenous growth factor-depleted conditions fail to support oestrogen hypersensitivity
We have obtained considerable evidence that ER is retained and functional in MCF-7X cells. Triplicate preparations for RT-PCR, Western blotting and immunocytochemistry revealed that these resistant cells express ER that is increased in comparison to the parental cells. However, this ER increase in MCF-7X was modest (elevated by 30%) and immunocytochemistry revealed nuclear intensity rather than percentage positivity changes (Fig. 1A). Adaptive increases in ER expression have also been commonly described for models resistant to oestrogen deprivation developed in the presence of serum growth factors, although these increases are generally more substantial (3–10-fold increases: Martin et al. 2003, Santen et al. 2004). We observed a 5-fold increase in basal ER transcriptional activity, as measured using dual-luciferase oestrogen response element (ERE) reporter assays, in MCF-7X compared to that in MCF-7 cells (Fig. 1B) and, while only low levels of progesterone receptor were apparent, there was a 60% increase in expression of the endogenous oestrogen-regulated gene pS2. Studies from Martin et al. (2003) and Santen et al. (2004) similarly report substantially increased basal ER transcriptional activity in their long-term oestrogen deprived, hypersensitive model. While tamoxifen was without inhibitory effect in MCF-7X, there was a 50% growth inhibition by day 15 with faslodex (Fig. 2A; n=5). In parallel, faslodex partially down-regulated ER, diminished ERE reporter activity by 30% (assessed after 18 h treatment, n=3; Fig. 2B), and reduced endogenous pS2 expression by 40%. Maximal faslodex inhibitory responses were achieved at an equivalent dosage (10–7 M) to MCF-7 cells, and lasted for ~10 weeks in culture before resistance emerged. Faslodex sensitivity is a feature retained by many models of acquired resistance to oestrogen deprivation (Martin et al. 2003, Santen et al. 2004), including an MCF-7/S9 cell line recently developed by Jensen et al. (2003) under serum-free culture conditions. We have demonstrated that the increased ER/ERE signalling within MCF-7X is unlikely to be the consequence of any changes in endogenous aromatase activity. Thus, challenge with the non-steroidal aromatase inhibitor anastrazole (1–300 ng/ml) in the presence or absence of the aromatase substrate androstenedione (10–8 M) was without impact on MCF-7X ER levels, ER-regulated gene expression or growth. In this respect, MCF-7X cells are comparable to MCF-7 cells that lack significant aromatase activity (Brodie & Njar 2000).
Clearly, there is some overlap between MCF-7X and in vitro models generated in the presence of higher levels of exogenous growth factors, with ER signalling being increased and an important growth contributor. Indeed, data from MCF-7X and previous models are cumulatively supportive of the use of faslodex in treating resistance to oestrogen deprivation. In contrast, however, oestrogen and growth factor depleted conditions appear unable to support oestrogen hypersensitivity in MCF-7X. While MCF-7X cells were substantially stimulated by exogenous oestradiol with a maximum 100% increase in growth when compared to control achieved at 10–9 M, they were not growth hypersensitive to sub-physiological oestradiol levels, demonstrating an approximately equivalent concentration response profile to MCF-7 cells over an extensive 10–8–10–19 M oestradiol range (n=3; Fig. 1C). Similarly, promotion of ER transcriptional activity by oestradiol assessed using ERE reporter gene constructs in MCF-7X cells followed an equivalent concentration profile to MCF-7. Our data infer that significant levels of exogenous growth factors are important in the development of oestrogen hypersensitivity in resistant models generated in the presence of stripped serum, while any autocrine signalling in long-term oestrogen deprived cells appears insufficient in this regard.
Classical growth factor receptors and their ligands do not contribute to growth of MCF-7X but may be important in oestrogen hypersensitivity
We have investigated whether growth factor signalling contributes to growth of MCF-7X by extensively profiling key growth factor pathways, first at the receptor/ligand level compared to parental MCF-7 cells. Interestingly, we have been unable to date to demonstrate a dominant role for any classical plasma membrane growth factor receptors, notably EGFR, HER2 or IGF-1R, in promoting MCF-7X growth.
In detail, challenge with an extensive range of peptide growth factors (10 ng/ml; n=5), including various erbB receptor ligands (e.g. epidermal growth factor (EGF), transforming growth factor-, amphiregulin, heregulins and ), failed to obviously promote growth of MCF-7X. In parallel, there was a lack of obvious effect of inhibitors specifically targeting such signalling in replicate experiments. Thus, the selective EGF receptor (EGFR) tyrosine kinase inhibitor gefitinib inhibited MCF-7X growth by only ~20%, using 1 μM previously demonstrated to be effective across our various antioestrogen resistant cell lines (Knowlden et al. 2003, Nicholson et al. 2004a). Indeed, immunocytochemistry revealed a lower basal EGFR expression in MCF-7X versus already low levels in the parental MCF-7 cells, also with a lower basal EGFR activity using a pan-phospho-specific EGFR antibody. While slightly increased HER2 expression was noted in MCF-7X, there was a fall in its activity compared to that in MCF-7 cells and challenge with the humanised HER2-directed monoclonal antibody herceptin (100 nM) failed to exert any growth inhibitory effects in MCF-7X in replicate experiments. These data appear to be in contrast to the oestrogen hypersensitive model derived in the presence of serum growth factors by Martin et al. (2003), where HER2 activity is significantly increased and also growth-contributory, as evidenced by gefitinib inhibitory effects at a higher drug dosage known to abrogate HER2. Further increased HER2 signalling has also been noted on acquisition of resistance to oestrogen deprivation by HER-2 transfected MCF-7 cells in vivo, where challenge using gefitinib was able to delay resistance confirming a growth contribution for such signalling (Massarweh et al. 2003).
Similarly, IGF-1R ligands failed to obviously stimulate MCF-7X growth. Furthermore, only ~15% growth inhibition of MCF-7X was achieved with various selective IGF-1R inhibitors, including AG1024, in several experiments. This was inferior to the growth inhibitory effect of such agents in the parental MCF-7 cell line, where a dominant role for IGF-1R signalling prior to endocrine therapy has previously been established (Nicholson et al. 2004a). In accordance with this poor response, IGF-1R expression was decreased basally in MCF-7X relative to MCF-7 cells. This was apparent at the mRNA level by RT-PCR, while immunocytochemistry demonstrated a substantial decline in plasma membrane-localised receptor. The latter technique was also able to demonstrate decreased IGF-1R activation in MCF-7X versus MCF-7 cells, using antibodies specific for key receptor phosphorylation sites. These data are in contrast to models generated in the presence of serum growth factors that cumulatively suggest some importance for IGF-1R-mediated signalling in adaptation of cells to oestrogen deprivation (Stephen et al. 2001, Martin et al. 2003, Santen et al. 2004), while gene transfer studies overexpressing IGF2, IGF-1R or insulin receptor substrate-1 also suggest a causative relationship with oestrogen independence in vitro (Daly et al. 1991, Guvakova & Surmacz 1997). In our quest to determine if there is any classical receptor input in MCF-7X, we have now profiled the growth effects of many additional growth factors, including basic and acidic fibroblast growth factor (FGFs), FGF7, platelet-derived growth factor (PDGF) and vascular endothelial growth factor (VEGF), also monitoring for respective receptor/ligand expression and where available, employing inhibitory agents. However, none of these studies have suggested a role for classical growth factor/receptor pathways to date. In total, our data in MCF-7X indicate that neither exogenous growth factor ligands nor their classical receptors are essential to maintain growth in an oestrogen deprived environment. However, they clearly appear to be of some importance in facilitating the growth of oestrogen hypersensitive in vitro models derived in the presence of serum growth factors, as evidenced by the emerging relevance of insulin, IGF-1R and HER2 described for such cells.
Our search for additional receptors that may contribute to MCF-7X growth is continuing using Affymetrix Human Genome U133A GeneChip arrays. To date, a receptor that appears more prominent in MCF-7X versus MCF-7 cells is the transferrin receptor (TR). Transferrin is a glycoprotein found in the bloodstream which transports iron. This is delivered to cells via transferrin binding to the target TR that is then internalized (Elliott et al. 1993). MCF-7X cells express markedly increased levels of transferrin receptor both at the mRNA and protein level, with substantial TR immunostaining apparent at their plasma membranes. RT-PCR revealed these cells also produce detectable levels of transferrin. Moreover, the only obvious mitogenic exogenous factor for MCF-7X, other than oestrogen, is transferrin (4 μg/ml) that consistently gives a superior 100% increase in growth compared with parental MCF-7 cells. Interestingly, breast cancer cells have previously been shown to produce transferrin and express TR, an event that may improve iron delivery to facilitate proliferative events (Vandewalle et al. 1989). It is thus feasible that increased TR may facilitate MCF-7X growth under oestrogen and growth factor depleted conditions by improving iron delivery and thereby aiding mitogenic signalling pathways. Such findings may have clinical relevance since TR has been shown to alter during disease progression (Elliott et al. 1993, Inoue et al. 1993), and we have obtained preliminary data showing increased TR mRNA associates with elevated proliferative activity and poorer patient survival in clinical disease.
Role for PI3K/AKT (and MAPK) in promoting growth of MCF-7X
Previous studies from several experimental models have revealed an importance for growth factor pathway kinases in adaptation of ER+ cells to long-term oestrogen deprivation. Of particular note is increased MAPK activity, with an additional role for increased PI3K/AKT signalling (Jeng et al. 2000, Shim et al. 2000, Martin et al. 2003, Santen et al. 2004). Our studies in antioestrogen resistance re-enforce the concept that activation of MAPK and PI3K/AKT can provide proliferative and survival signals in endocrine resistant cells (Knowlden et al. 2003, Nicholson et al. 2004a). In order to assess the relevance of these kinases to the oestrogen-deprived resistant state under conditions of depleted exogenous growth factor inputs, we have employed immunocytochemistry and Western blotting with phospho-specific antibodies in triplicate preparations to monitor activity of the intracellular kinases MAPK and AKT, as well as protein kinase C (PKC), in MCF-7X cells, while also examining the impact of inhibitors of such signalling.
MAPK phosphorylation was detected in MCF-7X but was not obviously increased with respect to that in parental MCF-7 cells. Similarly, there were no MAPK expression changes at the mRNA/protein level. Challenge with a MAP kinase kinase 1 (MEK1) inhibitor PD98059 (25 μM) blocked MAPK activity in MCF-7X and inhibited cell growth by day 15 (n=5). However, growth inhibition was relatively modest (35%) with resistance emerging quickly after only 3 weeks, suggesting the growth-promoting role of MAPK is minor in MCF-7X. Interestingly, oestrogen-deprived models generated in the presence of serum growth factors that are oestrogen hypersensitive commonly show substantially increased activity of this kinase (Martin et al. 2003, Santen et al. 2004), and so the absence of increased MAPK activity in MCF-7X may contribute to its lack of oestrogen hypersensitivity. Increased MAPK activity in some of the oestrogen hypersensitive models may be driven by growth factor ligands and classical receptors, since Martin et al. (2003) showed their 8-fold increased MAPK activity was regulated by insulin/IGF-1R and probably HER2. A role for increased MAPK signalling in facilitating basal growth and oestrogen hypersensitivity in these models has been revealed using MAPK inhibitors, although such signalling did not appear to be the only growth contribution (Jeng et al. 2000, Martin et al. 2003, Yue et al. 2003, Santen et al. 2004). Coupled with our observations in MCF-7X cells, these data suggest that while MAPK inhibitors may be of some value in resistance to oestrogen deprivation, this state is unlikely to be fully-treated using such an approach.
Our studies revealed a more prominent contribution for PI3K/AKT signalling in MCF-7X. These cells express AKT protein, with RT-PCR demonstrating a predominance of the AKT1 and two isoforms rather than AKT3 (although interestingly the latter has been linked previously with oestrogen independence; Faridi et al. 2003). MCF-7X cells have detectable activity of AKT in their cytoplasm, at the plasma membranes and within their nuclei, although neither AKT activity nor expression is elevated with respect to MCF-7 cells. Nevertheless, MCF-7X growth was inhibited by 65% using the PI3K/AKT signalling inhibitor LY294002 (5 μM) by day 15 in parallel with a substantial blockade of AKT activity with this agent (n=5; Fig. 2A). This growth inhibitory response lasted for ~10 weeks before resistance began to emerge. These data implicate PI3K/AKT signalling as an important pathway in promoting MCF-7X growth in the absence of substantial exogenous serum growth factor input. While any upstream regulator of PI3K/AKT signalling in MCF-7X remains as yet unidentified, interestingly there may be some link with transferrin receptor in such cells. Provocatively, PI3K signalling has previously been shown to regulate trafficking of transferrin/transferrin receptor, where PI3K inhibitors can deplete cell surface TR (Jess et al. 1996). We have obtained preliminary evidence supportive of positive TR/PI3K interplay in MCF-7X, since LY294002 abrogates transferrin-induced growth in MCF-7X cells. Interestingly, Santen et al. (2004) and Martin et al. (2003) have demonstrated obviously increased PI3K signalling in their hypersensitive models generated in the presence of serum growth factors. The model by Santen et al. (2004) showed significantly enhanced AKT activation, p70 S6 kinase and eukaryotic initiation factor-4E binding protein (Yue et al. 2003), while Martin et al. (2003) observed altered p90 ribosomal S6 kinase. Challenge with PI3K inhibitors in these cells revealed PI3K signalling was (in addition to MAPK) contributory to their oestrogen hypersensitivity, and again the absence of obvious increases in such signalling in MCF-7X may contribute to its lack of hypersensitivity. In total, the data from MCF-7X and the various oestrogen hypersensitive models indicate inhibitors of PI3K/AKT signalling may be of value in the future treatment of acquired resistance to oestrogen deprivation.
Finally, despite small increases in PKC-delta activity in MCF-7X versus MCF-7 cells, the PKC inhibitor bisindolylmaleimide (0.5 μM) was without impact on growth suggesting that there is no obvious growth regulatory role for PKCs in MCF-7X.
PI3K/AKT (but not MAPK) dependent regulation of Ser167ER phosphorylation in MCF-7X revealed using LY294002
In addition to their direct promotion of growth, the intracellular kinases PI3K/AKT and MAPK are able to interact with ER. Models from Santen et al. (2004) and Martin et al. (2003) generated in the presence of serum growth factors both indicate that such cross-talk may contribute to oestrogen hypersensitive breast cancer cell growth, although the nature of this crosstalk appears to differ between models, occurring in a non-genomic or genomic manner (see below). Given our observations that both PI3K/AKT and ER comprise dominant signalling routes in MCF-7X (with MAPK apparently subservient under basal growth conditions), it is important to investigate whether there is any evidence for kinase/ER cross-talk in MCF-7X and whether it is growth contributory.
A rapid ‘non-genomic’ cross-talk mechanism has been proposed by Santen et al. (2004) associated with adaptation to long-term oestrogen deprivation and gain of oestrogen hypersensitivity in their model. It has been proposed that residual oestrogens bind cytoplasmic ER and this then interacts with and phosphorylates the adapter molecule Shc via Src. This triggers rapid and marked MAPK (and potentially PI3K) activation. Preliminary evidence suggests that the Shc/ER complex is recruited to activated IGF-1R at the plasma membrane upstream of MAPK activation in breast cancer cells (Santen et al. 2004, Song et al. 2004). Such kinase signalling ultimately converges on cell cycle elements to promote proliferation. This appears to occur independently of any direct influence on ER transcription since while clearly oestrogen growth hypersensitive, there is an absence of parallel global hypersensitivity at the level of ERE-regulated transcription (Santen et al. 2004).
An alternative genomic model has been suggested for kinase/ER cross-talk, occurring at the level of nuclear ER and directly influencing ER transcriptional activity. The human ER is a phosphoprotein that is hyperphosphorylated in response to steroid binding and associated receptor conformational changes. Phosphorylation residues including the AF-1 residues serine 167 (Ser167ER), 118 (Ser118ER) and 104/106 have been implicated, but these appear to vary according to cell context (Lannigan 2003). ER phosphorylation acts to regulate aspects of steroid receptor function, notably transcriptional activation of ER-regulated genes. Under conditions of near-complete oestrogen deprivation, it is predicted that ER activity and hence ER-mediated transcription would be largely abrogated. However, it is now evident that various growth factor signalling kinases, in addition to their direct stimulation of proliferation and survival signals, can cross-talk with nuclear ER to phosphorylate key regulatory sites particularly within the AF-1 domain (Lannigan 2003). Of note in this regard is MAPK, implicated in Ser118ER phosphorylation, and PI3K/AKT that targets Ser167ER (Campbell et al. 2001, Chen et al. 2002). Interestingly, we (and others including Shou et al. 2004) have shown that increased nuclear ER phosphorylation on Ser118ER and Ser167ER occurs as a consequence of elevated EGFR/HER2/IGF-1R-regulated MAPK and AKT activity in acquired tamoxifen resistant cells (Nicholson et al. 2004a). This acts to enhance the transcriptional activity of the tamoxifen-ER complex to promote growth. It is conceivable, therefore, that such kinases might also induce nuclear ER phosphorylation and thereby transcriptional activity in a ligand-independent manner, facilitating the action of ER as a transcription factor in the steroid-depleted environment. Such a genomic ER mechanism has recently been implicated in the promotion of oestrogen hypersensitivity by Martin et al. (2003).
Immunofluorescence studies performed in MCF-7X failed to demonstrate obvious cytoplasmic or plasma membrane recruitment of ER either before or after short-term oestradiol challenge, suggestive that any kinase/ER cross-talk is unlikely to occur in a predominantly non-genomic manner in such cells. We have therefore to date focussed our attentions around cross-talk at the nuclear ER level, monitoring impact of kinase inhibitors on ER AF-1 phosphorylation (detected by immunocytochemistry in triplicate), ER-mediated transcription assessed using ERE reporter gene constructs (18 h treatment; n=3), and endogenous pS2 immunostaining. MCF-7X cells exhibited detectable levels of ER phosphorylation in their nuclei on Ser167ER and Ser118ER residues (Fig. 3A,B). In parallel with the adaptive increase in ER expression, there was a 30% increase in Ser118ER phosphorylation in MCF-7X versus the parental MCF-7 cells, although Ser167ER activity was unchanged. The level of Ser118ER phosphorylation achieved was lower than we have noted previously in our various tamoxifen resistant models with their markedly increased EGFR/HER2/MAPK/AKT signalling (Nicholson et al. 2004a), and furthermore a more obvious 3-fold increase in Ser118ER activity has been reported for an oestrogen hypersensitive model generated in the presence of serum growth factors by Martin et al. (2003). The lack of substantial increase in ER AF-1 phosphorylation in MCF-7X may be a consequence of relatively modest ER expression and growth factor kinase activity, as well as the apparent absence of any classical growth factor ligand/receptor input under conditions of oestrogen and serum growth factor depletion.
Challenge with the MEK1 inhibitor PD98059 blocked MAPK activation but was without inhibitory effect on phosphorylation of Ser167ER or Ser118ER. It furthermore failed to inhibit ERE transcription or endogenous pS2 expression. These data indicate that the modest MAPK-mediated basal growth input in MCF-7X is direct, rather than via any impact on nuclear ER signalling. Jeng et al. (2000) also describe that basal growth regulation of their long-term oestrogen deprived cells by MAPK occurs independently of any impact on ER transcriptional activity, although growth stimulation of these cells by exogenous oestrogen appears in part to involve such interplay. Martin et al. (2003) have also failed to demonstrate any impact of MAPK activity on Ser118ER phosphorylation in their oestrogen hypersensitive model. However, they have demonstrated that this kinase does partially contribute to ER transcriptional activity and growth, an event they suggest may occur via MAPK regulating ER coactivator activity or additional AF-1 phosphorylation sites.
In contrast, the PI3K inhibitor LY294002 decreased AKT activity and substantially decreased Ser167ER phosphorylation in MCF-7X (Fig. 3D), although Ser118ER activity was unaffected (Fig. 3E). It also partially reduced ER transcriptional activity both with regards to ERE reporter activity and endogenous pS2 expression (approximately 30% and 20% inhibition respectively; Figs 2B and 3F) versus control (Figs 2B and 3C) and, as previously stated, inhibited MCF-7X growth by 65% (Fig. 2A). These data indicate that there is prominent cross-talk between PI3K/AKT signalling and ER at the genomic level in MCF-7X, where PI3K/AKT regulation of Ser167ER phosphorylation appears to play a role in undermining the inhibitory effects of oestrogen deprivation on ER transcriptional activity and growth. Martin et al. (2003) similarly report that LY294002 impacts adversely on ER transcriptional activity in their oestrogen hypersensitive model, and an AKT/Ser167ER mechanism has been reported to contribute to tamoxifen resistance (Campbell et al. 2001, Nicholson et al. 2004b). It is important to note that ER transfection studies have previously suggested LY294002 (1–25 μM) may, in equivalence to faslodex, be able to act as a competitive inhibitor of oestrogen binding to the ER and thereby subvert ER transcriptional activity. This is in addition to its ability to decrease PI3K/AKT signalling that can influence ER phosphorylation status and transcriptional activity (Pasapera Limon et al. 2003). Our ongoing studies are thus using further PI3K pharmacological inhibitors (notably including wortmannin) unlikely to directly bind ER and displace oestrogen to confirm PI3K/AKT cross-talk with ER in MCF-7X. Encouragingly, studies to date indicate wortmannin also inhibits Ser167ER phosphorylation.
PI3K/AKT and MAPK independent regulation of Ser118ER phosphorylation in MCF-7X revealed using faslodex
In contrast to LY294002, faslodex challenge was associated with reduced ER protein level and decreased Ser118ER activity in MCF-7X (Fig. 3H). This was paralleled by partial inhibitory effects on ER transcriptional activity assessed using ERE reporters and pS2 expression (30% and 40% decreases respectively; Figs 2B and 3I), with subsequent obvious (albeit incomplete) growth inhibition (n=5; Fig. 2A). However, challenge with faslodex did not influence Ser167 activity in MC7-7X cells (Fig. 3G); paradoxical observations given the partial decrease in ER protein level achieved with this agent. Thus, faslodex may even be enabling Ser167 activity on residual ER in MCF-7X cells. Faslodex and a further pure antioestrogen, ICI 164,384, have been reported previously to be able to promote ER serine phosphorylation, although it is uncertain which AF-1 sites are targets for this action and this may also be cell context-specific (Le Goff et al. 1994, Joel et al. 1998, Chen et al. 2002). Potential kinase mediators of Ser167ER phosphorylation in MCF-7X cells are AKT, MAPK-promoted p90 ribosomal S6 kinase, and casein kinase II (Lannigan 2003). As stated above, the lack of effect of PD98059 and inhibitory impact of LY294002 on Ser167ER phosphorylation (paralleled by inhibitory effects recently noted with wortmannin) indicate AKT is the kinase driving Ser167ER activity in MCF-7X cells in the presence of faslodex. How this is achieved remains unknown, since faslodex treatment does not obviously increase AKT activity: perhaps there is in some way improved accessibility of this kinase to the Ser167 residue in the presence of faslodex.
Thus, while PI3K/AKT appears likely to maintain Ser167ER phosphorylation in MCF-7X cells, faslodex challenge suggests an additional growth importance for phosphorylation of Ser118ER. Since faslodex treatment did not alter MAPK or AKT activity, Ser118ER phosphorylation appears independent of these kinases in MCF-7X. It is possible that activity of this site may be promoted by residual steroid hormone in this model. Ser118ER is reported to be a dominant ligand phosphorylated site in MCF-7 and we have recently observed substantial further Ser118ER (but not Ser167ER) activation in MCF-7X cells following oestrogen challenge. Moreover, ligand stimulated Ser118ER phosphorylation has been shown to occur independently of MAPK and PI3K/AKT (Martin et al. 2003) and may be mediated by CDK7/TFIIH (Joel et al. 1998, Chen et al. 2002, Lannigan 2003). Decreases in ER protein may also be contributory to the reduced Ser118ER phosphorylation observed with faslodex in MCF-7X cells.
Residual ER phosphorylation after single agent treatment provides a compensatory survival mechanism for MCF-7X
Our data indicate that both Ser167ER and Ser118ER activation contribute to ER transcriptional activity and growth of MCF-7X, with LY294002 or faslodex revealing these sites have non-overlapping regulatory pathways. We have thus hypothesised that any phosphorylated ER remaining when LY294002 or faslodex are applied singly might provide an important compensatory cell survival mechanism, underlying incomplete inhibitory effects on ER transcriptional activity and growth in MCF-7X and ultimately enabling emergence of resistance after 10 weeks. If this hypothesis were valid, co-targeting the regulatory pathways to decrease activation of both Ser118ER and Ser167ER would be expected to improve inhibition of ER transcriptional activity and growth of such cells. Co-treatment with faslodex plus LY294002 was able to decrease ER phosphorylation on both ER sites (Figs 3J and K) and further decreased ER transcriptional activity using ERE reporter constructs (45% decrease, n=3; Fig. 2B) and pS2 expression (60% depletion; Fig. 3L). This was associated with a superior anti-tumour response in MCF-7X, with 90% growth inhibition by day 15 (n=5; Fig. 2A). Furthermore, resistance was substantially delayed until 25 weeks, where resistant cells were subsequently slower growing than those emerging with the single agents. However, while clearly a superior approach, co-treatment with faslodex plus LY294002 was unable to prevent development of therapeutic resistance in MCF-7X. Profiling of co-treated cells revealed that MAPK activity was not depleted and we therefore surmised that maintenance of such signalling might underpin incomplete tumour response. Triple challenge with the MEK1 inhibitor PD98059 in combination with faslodex plus LY294002 resulted in a further superior anti-tumour effect in MCF-7X compared with faslodex plus LY294002 co-treatment alone. By 15 days cell numbers fell below the initial seeding density (indicative of cell loss) with triple treatment (Fig. 2A), so that cultures could not be maintained past 16 weeks, an effect that was highly reproducible between experiments (n=5). Triple treatment efficiently abrogated therapeutic resistance. Preliminary experiments using wortmannin together with PD98059 and faslodex again demonstrate this is a superior treatment strategy in MCF-7X. The profound inhibitory effect in MCF-7X was associated with the predicted MAPK blockade. Surprisingly, however, there was also a superior depletion of residual Ser118ER activity in MCF-7X versus faslodex plus LY294002 co-treatment (as well as triple treatment maintaining the substantially decreased Ser167ER phosphorylation; Fig. 3M and N). Triple treatment further depleted ERE activity (60% decrease, n=3; Fig. 2B) and maximally-reduced pS2 expression (Fig. 3O). Thus, while MAPK activity does not contribute to basal ER phosphorylation in MCF-7X, it does appear that this kinase interplays with ER via Ser118ER phosphorylation under conditions of faslodex plus LY294002 co-treatment, an event that contributes to cell survival. The mechanism underlying this new coupling remains unknown. However, it is feasible that the low levels of MAPK activity are now sufficient to trigger Ser118ER phosphorylation because there is effective blockade of the basal ER phosphorylation regulators by faslodex plus LY294002 co-treatment, or alternatively that MAPK access to this site is now enabled because of ER conformational changes occurring with co-treatment (Lannigan 2003). As further evidence of this new coupling, we noted that MCF-7X cells ultimately acquiring resistance to faslodex plus LY294002 co-treatment showed parallel obvious increases in MAPK and Ser118ER activity.
We subsequently were able to confirm that inhibition of Ser118ER phosphorylation was required for the catastrophic effects of triple treatment on MCF-7X. LY294002 plus PD98059 co-treatment in the absence of faslodex again maximally decreased kinase activity, but was without inhibitory effect on Ser118ER phosphorylation. This combination treatment was certainly superior in inhibiting growth compared with LY294002 or PD98059 alone (75% inhibition, n=5). This has also been observed by Martin et al. (2003), Yue et al. (2003) and Santen et al. (2004) in their various models. However, in parallel with the lack of effect on Ser118ER phosphorylation, MCF-7X cells always survived LY294002 plus PD98059 co-treatment, with resistance emerging at 12 weeks.
Therapeutic implications
The new MCF-7X model has consolidated the importance of increased ER signalling and the key contribution for intracellular kinases, notably PI3K/AKT, to growth of cells that have acquired resistance to oestrogen deprivation. While there does not seem to be a significant contribution of classical growth factor receptors such as EGFR, HER2 or IGF-1R in MCF-7X cells, PI3K interplay with transferrin receptor may be contributory and requires future detailed investigation. MCF-7X cell studies demonstrate the potential of inhibiting PI3K/AKT signalling to treat this disease and inhibitors of such pathways are emerging, while faslodex responses following clinical acquisition of resistance to aromatase inhibitors have recently been described (Johnston 2004). Importantly, we have also been able to demonstrate that PI3K/AKT cross-talks with ER via Ser167 activation (where any contribution for transferrin receptor in this cross-talk is currently being evaluated through transferrin challenge studies). There is an additional important MAPK/PI3K/AKT-independent regulation of Ser118ER phosphorylation in MCF-7X cells. The impact on ER phosphorylation of agents targeting these dominant regulatory routes (i.e. LY294002 and faslodex) appears central to subsequent inhibitory effects on ER transcriptional activity and growth. Importantly, however, MCF-7X cells reveal that the oestrogen-deprived phenotype is extremely flexible in its use of regulatory pathways to maintain ER phosphorylation during treatment. This readily provides a compensatory mechanism that limits initial therapeutic efficacy of agents applied singly and allows evolution of resistance in MCF-7X. Our data demonstrate that intelligent combination therapy to target alll regulators of ER AF-1 phosphorylation (i.e. PI3K/AKT inhibitor plus MAPK inhibitor plus faslodex in MCF-7X) may be required to abrogate this cell survival mechanism and thereby effectively inhibit acquired resistance to oestrogen deprivation and prevent subsequent therapeutic resistance.
Acknowledgements
Thanks to the Tenovus Cancer Research Centre Tissue Culture and Immunocytochemistry Units; to Dr Kathryn Taylor and Mr Richard McClelland for immunofluorescence and reporter gene assay guidance; to Mr David Burston and Miss Helena Dunn (Welsh School of Pharmacy) for technical assistance.
Funding
This research was generously supported by the Tenovus Organisation. There are no conflicts of interest to declare regarding this work.
References
Brodie AM & Njar VC 2000 Aromatase inhibitors and their application in breast cancer treatment. Steroids 65 171–179.
Brodie AH, Jelovac D & Long B 2003 The intratumoral aromatase model: studies with aromatase inhibitors and antiestrogens. Journal of Steroid Biochemistry and Molecular Biology 86 283–288.
van der Burg B, Rutteman GR, Blankenstein MA, de Laat SW & van Zoelen EJ 1988 Mitogenic stimulation of human breast cancer cells in a growth factor-defined medium: synergistic action of insulin and estrogen. Journal of Cell Physiology 134 101–118.
Campbell RA, Bhat-Nakshatri P, Patel NM, Constantinidou D, Ali S & Nakshatri H 2001 Phosphatidylinositol 3-kinase/AKT-mediated activation of estrogen receptor alpha: a new model for anti-estrogen resistance. Journal of Biological Chemistry 276 9817–9824.
Chen D, Washbrook E, Sarwar N, Bates GJ, Pace PE, Thirunuvakkarasu V, Taylor J, Epstein RJ, Fuller-Pace FV, Egly JM, Coombes RC & Ali S 2002 Phosphorylation of human estrogen receptor alpha at serine 118 by two distinct signal transduction pathways revealed by phosphorylation-specific antisera. Oncogene 21 4921–4931.
Daly RJ, Harris WH, Wang DY & Darbre PD 1991 Autocrine production of insulin-like growth factor II using an inducible expression system results in reduced estrogen sensitivity of MCF-7 human breast cancer cells. Cell Growth and Differentiation 2 457–464.
Elliott RL, Elliott MC, Wang F & Head JF 1993 Breast carcinoma and the role of iron metabolism. A cytochemical, tissue culture, and ultrastructural study. Annals of the New York Academy of Sciences 698 159–166.
Faridi J, Wang L, Endemann G & Roth RA 2003 Expression of constitutively active Akt-3 in MCF-7 breast cancer cells reverses the estrogen and tamoxifen responsivity of these cells in vivo. Clinical Cancer Research 9 2933–2939.
Fuqua SA, Wiltschke C, Zhang QX, Borg A, Castles CG, Friedrichs WE, Hopp T, Hilsenbeck S, Mohsin S, O’Connell P & Allred DC 2000 A hypersensitive estrogen receptor-alpha mutation in premalignant breast lesions. Cancer Research 60 4026–4029.
Guvakova MA & Surmacz E 1997 Overexpressed IGF-I receptors reduce estrogen growth requirements, enhance survival, and promote E-cadherin-mediated cell–cell adhesion in human breast cancer cells. Experimental Cell Research 231 149–162.
Inoue T, Cavanaugh PG, Steck PA, Brunner N & Nicolson GL 1993 Differences in transferrin response and numbers of transferrin receptors in rat and human mammary carcinoma lines of different metastatic potentials. Journal of Cell Physiology 156 212–217.
Jeng MH, Yue W, Eischeid A, Wang JP & Santen RJ 2000 Role of MAP kinase in the enhanced cell proliferation of long term estrogen deprived human breast cancer cells. Breast Cancer Research and Treatment 62 167–175.
Jensen J, Kitlen JW, Briand P, Labrie F & Lykkesfeldt AE 2003 Effect of antiestrogens and aromatase inhibitor on basal growth of the human breast cancer cell line MCF-7 in serum-free medium. Journal of Steroid Biochemistry and Molecular Biology 84 469–478.
Jess TJ, Belham CM, Thomson FJ, Scott PH, Plevin RJ & Gould GW 1996 Phosphatidylinositol 3'-kinase, but not p70 ribosomal S6 kinase, is involved in membrane protein recycling: wortmannin inhibits glucose transport and downregulates cell-surface transferrin receptor numbers independently of any effect on fluid-phase endocytosis in fibroblasts. Cellular Signalling 8 297–304.
Joel PB, Traish AM & Lannigan DA 1998 Estradiol-induced phosphorylation of serine 118 in the estrogen receptor is independent of p42/p44 mitogen-activated protein kinase. Journal of Biological Chemistry 273 13317–13323.
Johnston S 2004 Fulvestrant and the sequential endocrine cascade for advanced breast cancer. British Journal of Cancer 90 S15–S18.
de Jong PC, Blankenstein MA, Nortier JW, Slee PH, van de Ven J, van Gorp JM, Elbers JR, Schipper ME, Blijham GH, Thijssen JH, Lu Q, Jelovac D & Brodie AM 2003 The relationship between aromatase in primary breast tumors and response to treatment with aromatase inhibitors in advanced disease. Journal of Steroid Biochemistry and Molecular Biology 87 149–155.
Knowlden JM, Hutcheson IR, Jones HE, Madden T, Gee JMW, Harper ME, Barrow D, Wakeling AE & Nicholson RI 2003 Elevated levels of epidermal growth factor receptor/c-erbB2 heterodimers mediate an autocrine growth regulatory pathway in tamoxifen-resistant MCF-7 cells. Endocrinology 144 1032–1044.
Lannigan DA 2003 Estrogen receptor phosphorylation. Steroids 68 1–9.
Le Goff P, Montano MM, Schodin DJ & Katzenellenbogen BS 1994 Phosphorylation of the human estrogen receptor. Identification of hormone-regulated sites and examination of their influence on transcriptional activity. Journal of Biological Chemistry 269 4458–4466.
Martin LA, Farmer I, Johnston SR, Ali S, Marshall C & Dowsett M 2003 Enhanced estrogen receptor (ER) alpha, ERBB2, and MAPK signal transduction pathways operate during the adaptation of MCF-7 cells to long term estrogen deprivation. Journal of Biological Chemistry 278 30458–30468.
Massarweh SA, Jiang S, Mohsin SK, Di Pietro M, Wakeling AE, Osborne CK & Schiff R 2003 Resistance to endocrine therapy in a xenograst model of HER-2 overexpressing breast cancer is accompanied by increased HER-2 but loss of IGF-1 receptor expression. Breast Cancer Research and Treatment 82 S1007.
Nicholson RI, Hutcheson IR, Knowlden JM, Jones HE, Harper ME, Jordan N, Hiscox SE, Barrow D & Gee JM 2004a Nonendocrine pathways and endocrine resistance: observations with antiestrogens and signal transduction inhibitors in combination. Clinical Cancer Research 10 346S–354S.
Nicholson RI, Staka C, Boyns F, Hutcheson IR & Gee JM 2004b Growth factor-driven mechanisms associated with resistance to estrogen deprivation in breast cancer: new opportunities for therapy. Endocrine Related Cancer 11 623–641.
Pasapera Limon AM, Herrera-Munoz J, Gutierrez-Sagal R & Ulloa-Aguirre A 2003 The phosphatidylinositol 3-kinase inhibitor LY294002 binds the estrogen receptor and inhibits 17beta-estradiol-induced transcriptional activity of an estrogen sensitive reporter gene. Molecular and Cellular Endocrinology 200 199–202.
Santen RJ 2003 Inhibition of aromatase: insights from recent studies. Steroids 68 559–567.
Santen RJ, Song RX, Zhang Z, Yue W & Kumar R 2004 Adaptive hypersensitivity to estrogen: mechanism for sequential responses to hormonal therapy in breast cancer. Clinical Cancer Research 10 337S–345S.
Shim WS, Conaway M, Masamura S, Yue W, Wang JP, Kmar R & Santen RJ 2000 Estradiol hypersensitivity and mitogen-activated protein kinase expression in long-term estrogen deprived human breast cancer cells in vivo. Endocrinology 141 396–405.
Shou J, Massarweh S, Osborne CK, Wakeling AE, Ali S, Weiss M & Schiff R 2004 Mechanisms of tamoxigen resistance: increased estrogen receptor-HER2/neu cross-talk in ER/HER2-positive breast cancer. Journal of the National Cancer Institute 96 926–935.
Song RX, Barnes CJ, Zhang Z, Bao Y, Kumar R & Santen RJ 2004 The role of Shc and insulin-like growth factor 1 receptor in mediating the translocation of estrogen receptor alpha to the plasma membrane. PNAS 101 2076–2081.
Stephen RL, Shaw LE, Larsen C, Corcoran D & Darbre PD 2001 Insulin-like growth factor receptor levels are regulated by cell density and by long term estrogen deprivation in MCF7 human breast cancer cells. Journal of Biological Chemistry 276 40080–40086.
Vandewalle B, Hornez L, Revillion F & Lefebvre J 1989 Secretion of transferrin by human breast cancer cells. Biochemical and Biophysical Research Communications 163 149–154.
Winer EP, Hudis C, Burstein HJ, Wolff AC, Pritchard KI, Ingle JN, Chlebowski RT, Gelber R, Edge SB, Gralow J, Cobleigh MA, Mamounas EP, Goldstein LJ, Whelan TJ, Powles TJ, Bryant J, Perkins C, Perotti J, Braun S, Langer AS, Browman GP & Somerfield MR 2005 American Society of Clinical Oncology technology assessment on the use of aromatase inhibitors as adjuvant therapy for postmenopausal women with hormone receptor-positive breast cancer: status report. Journal of Clinical Oncology 23 619–629.
Yue W, Wang JP, Conaway MR, Li Y & Santen RJ 2003 Adaptive hypersensitivity following long-term estrogen deprivation: involvement of multiple signal pathways. Journal of Steroid Biochemistry and Molecular Biology 86 265–274.(Cindy M Staka, Robert I N)
This paper was presented at the 1st Tenovus/AstraZeneca Workshop, Cardiff (2005). AstraZeneca has supported the publication of these proceedings.
Abstract
In vitro models of long-term oestrogen deprivation utilise increased oestrogen receptor (ER) and are oestrogen hypersensitive, with emerging evidence that growth factor signalling contributes and interacts with ER. However, such models are commonly derived in the presence of serum growth factors that may force the resistance mechanism. Our new in vitro model, MCF-7X, has thus been developed under conditions of both oestrogen and growth factor depletion. ER expression, serine 118 phosphorylation on this receptor and its transcriptional activity were modestly increased compared to the parental MCF-7 cells, although MCF-7X cells were not oestrogen hypersensitive. Faslodex (0.1 μM) partially decreased ER and its transcriptional activity, with associated decreases in serine 118 phosphorylation. Faslodex inhibited MCF-7X growth by 50% for 10 weeks. Classical growth factor receptors did not impact on MCF-7X growth and only a modest contribution for MAP kinase was revealed using PD98059 (25 μM; 35% inhibition for 3 weeks). However, the phosphatidylinositol-3-OH (PI3)-kinase inhibitor LY294002 (5 μM) inhibited MCF-7X growth by 65% for 10 weeks. In contrast to PD98059, LY294002 also partially-inhibited ER transcriptional activity and decreased serine 167 ER phosphorylation. Co-treatment with faslodex plus LY294002 to decrease activity of both serine 118 and 167 proved superior vs the single agents in decreasing ER transcriptional activity and MCF-7X growth (90% inhibition for 25 weeks). However, triple treatment including PD98059 was required to prevent resistance in MCF-7X, an event dependent on maximal depletion of serine 118 phosphorylation and ER transcriptional activity. Kinases clearly contribute in resistance to oestrogen deprivation, cross-talking with ER signalling via AF-1 phosphorylation. While inhibiting each pathway has potential to treat this state, combined therapy targeting all regulators of ER phosphorylation may be required to block subsequent emergence of resistance.
Introduction
Oestrogen deprivation, achieved clinically in post-menopausal women using aromatase inhibitors or in premenopausal women using ovarian ablation or suppression with luteinizing hormone-releasing hormone superagonists, is of considerable value in breast cancer management. Indeed, emerging data indicate that ‘third generation’ aromatase inhibitors may be superior to tamoxifen in postmenopausal women (Santen 2003). Thus, such inhibitors provide an effective second-line strategy in advanced postmenopausal oestrogen receptor positive (ER+) breast cancer. They can also improve tumour response compared to tamoxifen when used first-line in advanced ER+ disease. Moreover, the updated analysis of the ATAC trial in early breast cancer has shown that anastrazole as a primary adjuvant therapy significantly improved disease-free survival, increased time to recurrence and reduced contralateral invasive breast cancer incidence versus tamoxifen in ER+ postmenopausal disease. Use of aromatase inhibitors following completion of tamoxifen treatment also appears promising (Winer et al. 2005).
Importantly, however, while these data are encouraging, it is evident that oestrogen deprivation is unlikely to provide the whole answer to effective treatment of breast cancer. As with other endocrine strategies, acquisition of resistance remains a significant problem with aromatase inhibitors, where improvement in relapse-free survival, compared with tamoxifen, in the adjuvant setting remains modest and relapse rates in advanced disease are still often measured in months rather than years (Nicholson et al. 2004a, b). It is clearly essential that we elucidate the mechanisms underlying acquired resistance to oestrogen deprivation so that we might intelligently approach treatment or prevention of this state in the future.
Experimental studies focussing on acquisition of resistance to oestrogen deprivation are now emerging, although models remain relatively sparse. Various long-term oestrogen deprived ER+ human breast cancer xenograft models in oophorectomised nude mice have been described previously (Shim et al. 2000). Furthermore, response and resistance specifically to aromatase inhibitors is being examined using mouse xenografts of aromatase-transfected MCF-7Ca breast cancer cells (Brodie et al. 2003). Long-term oestrogen deprived ER+ in vitro models (Martin et al. 2003, Santen et al. 2004) have also been developed from endocrine responsive cell lines, including MCF-7. Such cells are initially growth inhibited in vitro by oestrogen deprivation achieved under phenol red-free, charcoal-stripped serum conditions that deplete the exogenous oestradiol concentration to 10–13 M. However, resistant growth emerges following long-term oestrogen deprivation. A unifying feature of these models of acquired resistance to oestrogen deprivation appears to be a retained mitogenic role for ER (Martin et al. 2003, Santen et al. 2004). Such models commonly exhibit increased ER expression and can be growth inhibited by the pure antioestrogen faslodex. Interestingly, long-term oestrogen deprivation has also invariably been associated with development of hypersensitivity to oestrogens, where exquisitely low residual steroid hormone levels now appear sufficient to support tumour cell growth. Indeed, Santen et al. (2004) have shown oestradiol concentrations as low as 10–14 M maximally stimulate growth of their ER+ long-term oestrogen deprived sub-line, in contrast to higher concentrations required in MCF-7 cells.
Various groups are examining the underlying biology of resistance to oestrogen deprivation and its apparent oestrogen hypersensitivity. Some are examining the role of altered ER structure/function while others are monitoring the impact of aromatase content, although the relevance of these parameters to clinical resistance remains obscure (Fuqua et al. 2000, de Jong et al. 2003). However, a potentially important mitogenic contribution in resistance to oestrogen deprivation is growth factor signalling (Nicholson et al. 2004b). There is supportive evidence from gene transfer studies for a causative association between altered growth factor signalling and oestrogen independent growth (Daly et al. 1991). Further evidence has arisen from other forms of antihormone resistance, notably antioestrogen resistant clinical disease and oestrogen receptor negative states, as well as acquired antioestrogen resistant models. These data cumulatively suggest that increased activation of growth factor signalling pathways, either through enhanced availability of growth factor ligands or via up-regulation/increased activation of their receptors or recruited kinases, can promote antihormone resistance (Nicholson et al. 2004a, b). While such aberrations may be present de novo, adaptive events occurring in such pathways during therapy may also promote the acquired resistant state. As proof of principle, our acquired tamoxifen resistant cell line recruits increased epidermal growth factor receptor (EGFR)/HER2 signalling that emerges during the antihormone responsive phase for its growth (Knowlden et al. 2003). Such signalling appears to promote ER activity in these antioestrogen resistant cells, an event that serves to enhance tamoxifen agonism (Nicholson et al. 2004a). Clearly, if emergent during long-term oestrogen deprivation, adaptive increases in growth factor signalling elements might also allow growth in the presence of very low oestrogen levels and promote oestrogen hypersensitivity. Interestingly, therefore, various models of acquired resistance to oestrogen deprivation have implicated insulin-like growth factor receptor (IGF-1R), HER2 and downstream activation of MAP kinase (MAPK) and phosphatidylinositol-3-OH kinase (PI3K)/AKT, with emerging evidence of such growth factor elements interacting with ER (Stephen et al. 2001, Martin et al. 2003, Santen et al. 2004).
However, most in vitro models employed to date have been developed in the presence of serum growth factors, with in some cases further insulin supplementation. This availability of exogenous growth factors could feasibly force the acquired resistance mechanism and promote oestrogen hypersensitivity. Indeed, in the model from Martin et al. (2003) addition of insulin appeared to promote IGF-1R-mediated super-sensitisation to residual oestrogens. To further decipher resistance to oestrogen deprivation, we have therefore developed a new ER+ acquired resistant in vitro model, MCF-7X (Nicholson et al. 2004b), under conditions of parallel depletion of oestrogen and exogenous growth factors. Using this model we have questioned (i) whether resistance to oestrogen deprivation can still arise in the absence of high levels of exogenous growth factors; (ii) whether ER remains critical to growth and if the resistant phenotype is oestrogen hypersensitive; (iii) whether there is evidence at the receptor/ligand or (iv) intracellular kinase level for growth factor signalling pathway elements contributing to growth; (v) whether such signalling cross-talks with ER and (vi) whether the growth mechanism can be effectively targeted. It is hoped that by further understanding of the inherent resistance mechanism we will be able to determine effective treatment strategies for acquired resistance to oestrogen deprivation.
Acquired resistance to oestrogen deprivation can arise under conditions of exogenous growth factor depletion (MCF-7X cells)
MCF-7 cells were grown long-term in white RPMI medium containing 5% charcoal-stripped, heat-inactivated (65 °C, 40 min) foetal calf serum (termed ‘X’ medium), an approach which depletes oestrogens and exogenous growth factors yet still ensures cell viability and attachment (van der Burg et al. 1988). After an initial phase of >80% growth inhibition for 4 months, resistance was acquired with proliferative activity of the resultant MCF-7X sub-line approaching that of parental MCF-7 cells prior to treatment. Clearly, high levels of exogenous growth factors are not essential for development of resistance to oestrogen deprivation, with input from any residual steroid hormone and presumably predominantly autocrine signalling adequate. This new cell model provides a sobering example of the adaptability of ER+ human breast cancer cells under conditions of substantially depleted exogenous mitogens.
ER signalling is prominent in MCF-7X cells and promotes growth, but oestrogen and exogenous growth factor-depleted conditions fail to support oestrogen hypersensitivity
We have obtained considerable evidence that ER is retained and functional in MCF-7X cells. Triplicate preparations for RT-PCR, Western blotting and immunocytochemistry revealed that these resistant cells express ER that is increased in comparison to the parental cells. However, this ER increase in MCF-7X was modest (elevated by 30%) and immunocytochemistry revealed nuclear intensity rather than percentage positivity changes (Fig. 1A). Adaptive increases in ER expression have also been commonly described for models resistant to oestrogen deprivation developed in the presence of serum growth factors, although these increases are generally more substantial (3–10-fold increases: Martin et al. 2003, Santen et al. 2004). We observed a 5-fold increase in basal ER transcriptional activity, as measured using dual-luciferase oestrogen response element (ERE) reporter assays, in MCF-7X compared to that in MCF-7 cells (Fig. 1B) and, while only low levels of progesterone receptor were apparent, there was a 60% increase in expression of the endogenous oestrogen-regulated gene pS2. Studies from Martin et al. (2003) and Santen et al. (2004) similarly report substantially increased basal ER transcriptional activity in their long-term oestrogen deprived, hypersensitive model. While tamoxifen was without inhibitory effect in MCF-7X, there was a 50% growth inhibition by day 15 with faslodex (Fig. 2A; n=5). In parallel, faslodex partially down-regulated ER, diminished ERE reporter activity by 30% (assessed after 18 h treatment, n=3; Fig. 2B), and reduced endogenous pS2 expression by 40%. Maximal faslodex inhibitory responses were achieved at an equivalent dosage (10–7 M) to MCF-7 cells, and lasted for ~10 weeks in culture before resistance emerged. Faslodex sensitivity is a feature retained by many models of acquired resistance to oestrogen deprivation (Martin et al. 2003, Santen et al. 2004), including an MCF-7/S9 cell line recently developed by Jensen et al. (2003) under serum-free culture conditions. We have demonstrated that the increased ER/ERE signalling within MCF-7X is unlikely to be the consequence of any changes in endogenous aromatase activity. Thus, challenge with the non-steroidal aromatase inhibitor anastrazole (1–300 ng/ml) in the presence or absence of the aromatase substrate androstenedione (10–8 M) was without impact on MCF-7X ER levels, ER-regulated gene expression or growth. In this respect, MCF-7X cells are comparable to MCF-7 cells that lack significant aromatase activity (Brodie & Njar 2000).
Clearly, there is some overlap between MCF-7X and in vitro models generated in the presence of higher levels of exogenous growth factors, with ER signalling being increased and an important growth contributor. Indeed, data from MCF-7X and previous models are cumulatively supportive of the use of faslodex in treating resistance to oestrogen deprivation. In contrast, however, oestrogen and growth factor depleted conditions appear unable to support oestrogen hypersensitivity in MCF-7X. While MCF-7X cells were substantially stimulated by exogenous oestradiol with a maximum 100% increase in growth when compared to control achieved at 10–9 M, they were not growth hypersensitive to sub-physiological oestradiol levels, demonstrating an approximately equivalent concentration response profile to MCF-7 cells over an extensive 10–8–10–19 M oestradiol range (n=3; Fig. 1C). Similarly, promotion of ER transcriptional activity by oestradiol assessed using ERE reporter gene constructs in MCF-7X cells followed an equivalent concentration profile to MCF-7. Our data infer that significant levels of exogenous growth factors are important in the development of oestrogen hypersensitivity in resistant models generated in the presence of stripped serum, while any autocrine signalling in long-term oestrogen deprived cells appears insufficient in this regard.
Classical growth factor receptors and their ligands do not contribute to growth of MCF-7X but may be important in oestrogen hypersensitivity
We have investigated whether growth factor signalling contributes to growth of MCF-7X by extensively profiling key growth factor pathways, first at the receptor/ligand level compared to parental MCF-7 cells. Interestingly, we have been unable to date to demonstrate a dominant role for any classical plasma membrane growth factor receptors, notably EGFR, HER2 or IGF-1R, in promoting MCF-7X growth.
In detail, challenge with an extensive range of peptide growth factors (10 ng/ml; n=5), including various erbB receptor ligands (e.g. epidermal growth factor (EGF), transforming growth factor-, amphiregulin, heregulins and ), failed to obviously promote growth of MCF-7X. In parallel, there was a lack of obvious effect of inhibitors specifically targeting such signalling in replicate experiments. Thus, the selective EGF receptor (EGFR) tyrosine kinase inhibitor gefitinib inhibited MCF-7X growth by only ~20%, using 1 μM previously demonstrated to be effective across our various antioestrogen resistant cell lines (Knowlden et al. 2003, Nicholson et al. 2004a). Indeed, immunocytochemistry revealed a lower basal EGFR expression in MCF-7X versus already low levels in the parental MCF-7 cells, also with a lower basal EGFR activity using a pan-phospho-specific EGFR antibody. While slightly increased HER2 expression was noted in MCF-7X, there was a fall in its activity compared to that in MCF-7 cells and challenge with the humanised HER2-directed monoclonal antibody herceptin (100 nM) failed to exert any growth inhibitory effects in MCF-7X in replicate experiments. These data appear to be in contrast to the oestrogen hypersensitive model derived in the presence of serum growth factors by Martin et al. (2003), where HER2 activity is significantly increased and also growth-contributory, as evidenced by gefitinib inhibitory effects at a higher drug dosage known to abrogate HER2. Further increased HER2 signalling has also been noted on acquisition of resistance to oestrogen deprivation by HER-2 transfected MCF-7 cells in vivo, where challenge using gefitinib was able to delay resistance confirming a growth contribution for such signalling (Massarweh et al. 2003).
Similarly, IGF-1R ligands failed to obviously stimulate MCF-7X growth. Furthermore, only ~15% growth inhibition of MCF-7X was achieved with various selective IGF-1R inhibitors, including AG1024, in several experiments. This was inferior to the growth inhibitory effect of such agents in the parental MCF-7 cell line, where a dominant role for IGF-1R signalling prior to endocrine therapy has previously been established (Nicholson et al. 2004a). In accordance with this poor response, IGF-1R expression was decreased basally in MCF-7X relative to MCF-7 cells. This was apparent at the mRNA level by RT-PCR, while immunocytochemistry demonstrated a substantial decline in plasma membrane-localised receptor. The latter technique was also able to demonstrate decreased IGF-1R activation in MCF-7X versus MCF-7 cells, using antibodies specific for key receptor phosphorylation sites. These data are in contrast to models generated in the presence of serum growth factors that cumulatively suggest some importance for IGF-1R-mediated signalling in adaptation of cells to oestrogen deprivation (Stephen et al. 2001, Martin et al. 2003, Santen et al. 2004), while gene transfer studies overexpressing IGF2, IGF-1R or insulin receptor substrate-1 also suggest a causative relationship with oestrogen independence in vitro (Daly et al. 1991, Guvakova & Surmacz 1997). In our quest to determine if there is any classical receptor input in MCF-7X, we have now profiled the growth effects of many additional growth factors, including basic and acidic fibroblast growth factor (FGFs), FGF7, platelet-derived growth factor (PDGF) and vascular endothelial growth factor (VEGF), also monitoring for respective receptor/ligand expression and where available, employing inhibitory agents. However, none of these studies have suggested a role for classical growth factor/receptor pathways to date. In total, our data in MCF-7X indicate that neither exogenous growth factor ligands nor their classical receptors are essential to maintain growth in an oestrogen deprived environment. However, they clearly appear to be of some importance in facilitating the growth of oestrogen hypersensitive in vitro models derived in the presence of serum growth factors, as evidenced by the emerging relevance of insulin, IGF-1R and HER2 described for such cells.
Our search for additional receptors that may contribute to MCF-7X growth is continuing using Affymetrix Human Genome U133A GeneChip arrays. To date, a receptor that appears more prominent in MCF-7X versus MCF-7 cells is the transferrin receptor (TR). Transferrin is a glycoprotein found in the bloodstream which transports iron. This is delivered to cells via transferrin binding to the target TR that is then internalized (Elliott et al. 1993). MCF-7X cells express markedly increased levels of transferrin receptor both at the mRNA and protein level, with substantial TR immunostaining apparent at their plasma membranes. RT-PCR revealed these cells also produce detectable levels of transferrin. Moreover, the only obvious mitogenic exogenous factor for MCF-7X, other than oestrogen, is transferrin (4 μg/ml) that consistently gives a superior 100% increase in growth compared with parental MCF-7 cells. Interestingly, breast cancer cells have previously been shown to produce transferrin and express TR, an event that may improve iron delivery to facilitate proliferative events (Vandewalle et al. 1989). It is thus feasible that increased TR may facilitate MCF-7X growth under oestrogen and growth factor depleted conditions by improving iron delivery and thereby aiding mitogenic signalling pathways. Such findings may have clinical relevance since TR has been shown to alter during disease progression (Elliott et al. 1993, Inoue et al. 1993), and we have obtained preliminary data showing increased TR mRNA associates with elevated proliferative activity and poorer patient survival in clinical disease.
Role for PI3K/AKT (and MAPK) in promoting growth of MCF-7X
Previous studies from several experimental models have revealed an importance for growth factor pathway kinases in adaptation of ER+ cells to long-term oestrogen deprivation. Of particular note is increased MAPK activity, with an additional role for increased PI3K/AKT signalling (Jeng et al. 2000, Shim et al. 2000, Martin et al. 2003, Santen et al. 2004). Our studies in antioestrogen resistance re-enforce the concept that activation of MAPK and PI3K/AKT can provide proliferative and survival signals in endocrine resistant cells (Knowlden et al. 2003, Nicholson et al. 2004a). In order to assess the relevance of these kinases to the oestrogen-deprived resistant state under conditions of depleted exogenous growth factor inputs, we have employed immunocytochemistry and Western blotting with phospho-specific antibodies in triplicate preparations to monitor activity of the intracellular kinases MAPK and AKT, as well as protein kinase C (PKC), in MCF-7X cells, while also examining the impact of inhibitors of such signalling.
MAPK phosphorylation was detected in MCF-7X but was not obviously increased with respect to that in parental MCF-7 cells. Similarly, there were no MAPK expression changes at the mRNA/protein level. Challenge with a MAP kinase kinase 1 (MEK1) inhibitor PD98059 (25 μM) blocked MAPK activity in MCF-7X and inhibited cell growth by day 15 (n=5). However, growth inhibition was relatively modest (35%) with resistance emerging quickly after only 3 weeks, suggesting the growth-promoting role of MAPK is minor in MCF-7X. Interestingly, oestrogen-deprived models generated in the presence of serum growth factors that are oestrogen hypersensitive commonly show substantially increased activity of this kinase (Martin et al. 2003, Santen et al. 2004), and so the absence of increased MAPK activity in MCF-7X may contribute to its lack of oestrogen hypersensitivity. Increased MAPK activity in some of the oestrogen hypersensitive models may be driven by growth factor ligands and classical receptors, since Martin et al. (2003) showed their 8-fold increased MAPK activity was regulated by insulin/IGF-1R and probably HER2. A role for increased MAPK signalling in facilitating basal growth and oestrogen hypersensitivity in these models has been revealed using MAPK inhibitors, although such signalling did not appear to be the only growth contribution (Jeng et al. 2000, Martin et al. 2003, Yue et al. 2003, Santen et al. 2004). Coupled with our observations in MCF-7X cells, these data suggest that while MAPK inhibitors may be of some value in resistance to oestrogen deprivation, this state is unlikely to be fully-treated using such an approach.
Our studies revealed a more prominent contribution for PI3K/AKT signalling in MCF-7X. These cells express AKT protein, with RT-PCR demonstrating a predominance of the AKT1 and two isoforms rather than AKT3 (although interestingly the latter has been linked previously with oestrogen independence; Faridi et al. 2003). MCF-7X cells have detectable activity of AKT in their cytoplasm, at the plasma membranes and within their nuclei, although neither AKT activity nor expression is elevated with respect to MCF-7 cells. Nevertheless, MCF-7X growth was inhibited by 65% using the PI3K/AKT signalling inhibitor LY294002 (5 μM) by day 15 in parallel with a substantial blockade of AKT activity with this agent (n=5; Fig. 2A). This growth inhibitory response lasted for ~10 weeks before resistance began to emerge. These data implicate PI3K/AKT signalling as an important pathway in promoting MCF-7X growth in the absence of substantial exogenous serum growth factor input. While any upstream regulator of PI3K/AKT signalling in MCF-7X remains as yet unidentified, interestingly there may be some link with transferrin receptor in such cells. Provocatively, PI3K signalling has previously been shown to regulate trafficking of transferrin/transferrin receptor, where PI3K inhibitors can deplete cell surface TR (Jess et al. 1996). We have obtained preliminary evidence supportive of positive TR/PI3K interplay in MCF-7X, since LY294002 abrogates transferrin-induced growth in MCF-7X cells. Interestingly, Santen et al. (2004) and Martin et al. (2003) have demonstrated obviously increased PI3K signalling in their hypersensitive models generated in the presence of serum growth factors. The model by Santen et al. (2004) showed significantly enhanced AKT activation, p70 S6 kinase and eukaryotic initiation factor-4E binding protein (Yue et al. 2003), while Martin et al. (2003) observed altered p90 ribosomal S6 kinase. Challenge with PI3K inhibitors in these cells revealed PI3K signalling was (in addition to MAPK) contributory to their oestrogen hypersensitivity, and again the absence of obvious increases in such signalling in MCF-7X may contribute to its lack of hypersensitivity. In total, the data from MCF-7X and the various oestrogen hypersensitive models indicate inhibitors of PI3K/AKT signalling may be of value in the future treatment of acquired resistance to oestrogen deprivation.
Finally, despite small increases in PKC-delta activity in MCF-7X versus MCF-7 cells, the PKC inhibitor bisindolylmaleimide (0.5 μM) was without impact on growth suggesting that there is no obvious growth regulatory role for PKCs in MCF-7X.
PI3K/AKT (but not MAPK) dependent regulation of Ser167ER phosphorylation in MCF-7X revealed using LY294002
In addition to their direct promotion of growth, the intracellular kinases PI3K/AKT and MAPK are able to interact with ER. Models from Santen et al. (2004) and Martin et al. (2003) generated in the presence of serum growth factors both indicate that such cross-talk may contribute to oestrogen hypersensitive breast cancer cell growth, although the nature of this crosstalk appears to differ between models, occurring in a non-genomic or genomic manner (see below). Given our observations that both PI3K/AKT and ER comprise dominant signalling routes in MCF-7X (with MAPK apparently subservient under basal growth conditions), it is important to investigate whether there is any evidence for kinase/ER cross-talk in MCF-7X and whether it is growth contributory.
A rapid ‘non-genomic’ cross-talk mechanism has been proposed by Santen et al. (2004) associated with adaptation to long-term oestrogen deprivation and gain of oestrogen hypersensitivity in their model. It has been proposed that residual oestrogens bind cytoplasmic ER and this then interacts with and phosphorylates the adapter molecule Shc via Src. This triggers rapid and marked MAPK (and potentially PI3K) activation. Preliminary evidence suggests that the Shc/ER complex is recruited to activated IGF-1R at the plasma membrane upstream of MAPK activation in breast cancer cells (Santen et al. 2004, Song et al. 2004). Such kinase signalling ultimately converges on cell cycle elements to promote proliferation. This appears to occur independently of any direct influence on ER transcription since while clearly oestrogen growth hypersensitive, there is an absence of parallel global hypersensitivity at the level of ERE-regulated transcription (Santen et al. 2004).
An alternative genomic model has been suggested for kinase/ER cross-talk, occurring at the level of nuclear ER and directly influencing ER transcriptional activity. The human ER is a phosphoprotein that is hyperphosphorylated in response to steroid binding and associated receptor conformational changes. Phosphorylation residues including the AF-1 residues serine 167 (Ser167ER), 118 (Ser118ER) and 104/106 have been implicated, but these appear to vary according to cell context (Lannigan 2003). ER phosphorylation acts to regulate aspects of steroid receptor function, notably transcriptional activation of ER-regulated genes. Under conditions of near-complete oestrogen deprivation, it is predicted that ER activity and hence ER-mediated transcription would be largely abrogated. However, it is now evident that various growth factor signalling kinases, in addition to their direct stimulation of proliferation and survival signals, can cross-talk with nuclear ER to phosphorylate key regulatory sites particularly within the AF-1 domain (Lannigan 2003). Of note in this regard is MAPK, implicated in Ser118ER phosphorylation, and PI3K/AKT that targets Ser167ER (Campbell et al. 2001, Chen et al. 2002). Interestingly, we (and others including Shou et al. 2004) have shown that increased nuclear ER phosphorylation on Ser118ER and Ser167ER occurs as a consequence of elevated EGFR/HER2/IGF-1R-regulated MAPK and AKT activity in acquired tamoxifen resistant cells (Nicholson et al. 2004a). This acts to enhance the transcriptional activity of the tamoxifen-ER complex to promote growth. It is conceivable, therefore, that such kinases might also induce nuclear ER phosphorylation and thereby transcriptional activity in a ligand-independent manner, facilitating the action of ER as a transcription factor in the steroid-depleted environment. Such a genomic ER mechanism has recently been implicated in the promotion of oestrogen hypersensitivity by Martin et al. (2003).
Immunofluorescence studies performed in MCF-7X failed to demonstrate obvious cytoplasmic or plasma membrane recruitment of ER either before or after short-term oestradiol challenge, suggestive that any kinase/ER cross-talk is unlikely to occur in a predominantly non-genomic manner in such cells. We have therefore to date focussed our attentions around cross-talk at the nuclear ER level, monitoring impact of kinase inhibitors on ER AF-1 phosphorylation (detected by immunocytochemistry in triplicate), ER-mediated transcription assessed using ERE reporter gene constructs (18 h treatment; n=3), and endogenous pS2 immunostaining. MCF-7X cells exhibited detectable levels of ER phosphorylation in their nuclei on Ser167ER and Ser118ER residues (Fig. 3A,B). In parallel with the adaptive increase in ER expression, there was a 30% increase in Ser118ER phosphorylation in MCF-7X versus the parental MCF-7 cells, although Ser167ER activity was unchanged. The level of Ser118ER phosphorylation achieved was lower than we have noted previously in our various tamoxifen resistant models with their markedly increased EGFR/HER2/MAPK/AKT signalling (Nicholson et al. 2004a), and furthermore a more obvious 3-fold increase in Ser118ER activity has been reported for an oestrogen hypersensitive model generated in the presence of serum growth factors by Martin et al. (2003). The lack of substantial increase in ER AF-1 phosphorylation in MCF-7X may be a consequence of relatively modest ER expression and growth factor kinase activity, as well as the apparent absence of any classical growth factor ligand/receptor input under conditions of oestrogen and serum growth factor depletion.
Challenge with the MEK1 inhibitor PD98059 blocked MAPK activation but was without inhibitory effect on phosphorylation of Ser167ER or Ser118ER. It furthermore failed to inhibit ERE transcription or endogenous pS2 expression. These data indicate that the modest MAPK-mediated basal growth input in MCF-7X is direct, rather than via any impact on nuclear ER signalling. Jeng et al. (2000) also describe that basal growth regulation of their long-term oestrogen deprived cells by MAPK occurs independently of any impact on ER transcriptional activity, although growth stimulation of these cells by exogenous oestrogen appears in part to involve such interplay. Martin et al. (2003) have also failed to demonstrate any impact of MAPK activity on Ser118ER phosphorylation in their oestrogen hypersensitive model. However, they have demonstrated that this kinase does partially contribute to ER transcriptional activity and growth, an event they suggest may occur via MAPK regulating ER coactivator activity or additional AF-1 phosphorylation sites.
In contrast, the PI3K inhibitor LY294002 decreased AKT activity and substantially decreased Ser167ER phosphorylation in MCF-7X (Fig. 3D), although Ser118ER activity was unaffected (Fig. 3E). It also partially reduced ER transcriptional activity both with regards to ERE reporter activity and endogenous pS2 expression (approximately 30% and 20% inhibition respectively; Figs 2B and 3F) versus control (Figs 2B and 3C) and, as previously stated, inhibited MCF-7X growth by 65% (Fig. 2A). These data indicate that there is prominent cross-talk between PI3K/AKT signalling and ER at the genomic level in MCF-7X, where PI3K/AKT regulation of Ser167ER phosphorylation appears to play a role in undermining the inhibitory effects of oestrogen deprivation on ER transcriptional activity and growth. Martin et al. (2003) similarly report that LY294002 impacts adversely on ER transcriptional activity in their oestrogen hypersensitive model, and an AKT/Ser167ER mechanism has been reported to contribute to tamoxifen resistance (Campbell et al. 2001, Nicholson et al. 2004b). It is important to note that ER transfection studies have previously suggested LY294002 (1–25 μM) may, in equivalence to faslodex, be able to act as a competitive inhibitor of oestrogen binding to the ER and thereby subvert ER transcriptional activity. This is in addition to its ability to decrease PI3K/AKT signalling that can influence ER phosphorylation status and transcriptional activity (Pasapera Limon et al. 2003). Our ongoing studies are thus using further PI3K pharmacological inhibitors (notably including wortmannin) unlikely to directly bind ER and displace oestrogen to confirm PI3K/AKT cross-talk with ER in MCF-7X. Encouragingly, studies to date indicate wortmannin also inhibits Ser167ER phosphorylation.
PI3K/AKT and MAPK independent regulation of Ser118ER phosphorylation in MCF-7X revealed using faslodex
In contrast to LY294002, faslodex challenge was associated with reduced ER protein level and decreased Ser118ER activity in MCF-7X (Fig. 3H). This was paralleled by partial inhibitory effects on ER transcriptional activity assessed using ERE reporters and pS2 expression (30% and 40% decreases respectively; Figs 2B and 3I), with subsequent obvious (albeit incomplete) growth inhibition (n=5; Fig. 2A). However, challenge with faslodex did not influence Ser167 activity in MC7-7X cells (Fig. 3G); paradoxical observations given the partial decrease in ER protein level achieved with this agent. Thus, faslodex may even be enabling Ser167 activity on residual ER in MCF-7X cells. Faslodex and a further pure antioestrogen, ICI 164,384, have been reported previously to be able to promote ER serine phosphorylation, although it is uncertain which AF-1 sites are targets for this action and this may also be cell context-specific (Le Goff et al. 1994, Joel et al. 1998, Chen et al. 2002). Potential kinase mediators of Ser167ER phosphorylation in MCF-7X cells are AKT, MAPK-promoted p90 ribosomal S6 kinase, and casein kinase II (Lannigan 2003). As stated above, the lack of effect of PD98059 and inhibitory impact of LY294002 on Ser167ER phosphorylation (paralleled by inhibitory effects recently noted with wortmannin) indicate AKT is the kinase driving Ser167ER activity in MCF-7X cells in the presence of faslodex. How this is achieved remains unknown, since faslodex treatment does not obviously increase AKT activity: perhaps there is in some way improved accessibility of this kinase to the Ser167 residue in the presence of faslodex.
Thus, while PI3K/AKT appears likely to maintain Ser167ER phosphorylation in MCF-7X cells, faslodex challenge suggests an additional growth importance for phosphorylation of Ser118ER. Since faslodex treatment did not alter MAPK or AKT activity, Ser118ER phosphorylation appears independent of these kinases in MCF-7X. It is possible that activity of this site may be promoted by residual steroid hormone in this model. Ser118ER is reported to be a dominant ligand phosphorylated site in MCF-7 and we have recently observed substantial further Ser118ER (but not Ser167ER) activation in MCF-7X cells following oestrogen challenge. Moreover, ligand stimulated Ser118ER phosphorylation has been shown to occur independently of MAPK and PI3K/AKT (Martin et al. 2003) and may be mediated by CDK7/TFIIH (Joel et al. 1998, Chen et al. 2002, Lannigan 2003). Decreases in ER protein may also be contributory to the reduced Ser118ER phosphorylation observed with faslodex in MCF-7X cells.
Residual ER phosphorylation after single agent treatment provides a compensatory survival mechanism for MCF-7X
Our data indicate that both Ser167ER and Ser118ER activation contribute to ER transcriptional activity and growth of MCF-7X, with LY294002 or faslodex revealing these sites have non-overlapping regulatory pathways. We have thus hypothesised that any phosphorylated ER remaining when LY294002 or faslodex are applied singly might provide an important compensatory cell survival mechanism, underlying incomplete inhibitory effects on ER transcriptional activity and growth in MCF-7X and ultimately enabling emergence of resistance after 10 weeks. If this hypothesis were valid, co-targeting the regulatory pathways to decrease activation of both Ser118ER and Ser167ER would be expected to improve inhibition of ER transcriptional activity and growth of such cells. Co-treatment with faslodex plus LY294002 was able to decrease ER phosphorylation on both ER sites (Figs 3J and K) and further decreased ER transcriptional activity using ERE reporter constructs (45% decrease, n=3; Fig. 2B) and pS2 expression (60% depletion; Fig. 3L). This was associated with a superior anti-tumour response in MCF-7X, with 90% growth inhibition by day 15 (n=5; Fig. 2A). Furthermore, resistance was substantially delayed until 25 weeks, where resistant cells were subsequently slower growing than those emerging with the single agents. However, while clearly a superior approach, co-treatment with faslodex plus LY294002 was unable to prevent development of therapeutic resistance in MCF-7X. Profiling of co-treated cells revealed that MAPK activity was not depleted and we therefore surmised that maintenance of such signalling might underpin incomplete tumour response. Triple challenge with the MEK1 inhibitor PD98059 in combination with faslodex plus LY294002 resulted in a further superior anti-tumour effect in MCF-7X compared with faslodex plus LY294002 co-treatment alone. By 15 days cell numbers fell below the initial seeding density (indicative of cell loss) with triple treatment (Fig. 2A), so that cultures could not be maintained past 16 weeks, an effect that was highly reproducible between experiments (n=5). Triple treatment efficiently abrogated therapeutic resistance. Preliminary experiments using wortmannin together with PD98059 and faslodex again demonstrate this is a superior treatment strategy in MCF-7X. The profound inhibitory effect in MCF-7X was associated with the predicted MAPK blockade. Surprisingly, however, there was also a superior depletion of residual Ser118ER activity in MCF-7X versus faslodex plus LY294002 co-treatment (as well as triple treatment maintaining the substantially decreased Ser167ER phosphorylation; Fig. 3M and N). Triple treatment further depleted ERE activity (60% decrease, n=3; Fig. 2B) and maximally-reduced pS2 expression (Fig. 3O). Thus, while MAPK activity does not contribute to basal ER phosphorylation in MCF-7X, it does appear that this kinase interplays with ER via Ser118ER phosphorylation under conditions of faslodex plus LY294002 co-treatment, an event that contributes to cell survival. The mechanism underlying this new coupling remains unknown. However, it is feasible that the low levels of MAPK activity are now sufficient to trigger Ser118ER phosphorylation because there is effective blockade of the basal ER phosphorylation regulators by faslodex plus LY294002 co-treatment, or alternatively that MAPK access to this site is now enabled because of ER conformational changes occurring with co-treatment (Lannigan 2003). As further evidence of this new coupling, we noted that MCF-7X cells ultimately acquiring resistance to faslodex plus LY294002 co-treatment showed parallel obvious increases in MAPK and Ser118ER activity.
We subsequently were able to confirm that inhibition of Ser118ER phosphorylation was required for the catastrophic effects of triple treatment on MCF-7X. LY294002 plus PD98059 co-treatment in the absence of faslodex again maximally decreased kinase activity, but was without inhibitory effect on Ser118ER phosphorylation. This combination treatment was certainly superior in inhibiting growth compared with LY294002 or PD98059 alone (75% inhibition, n=5). This has also been observed by Martin et al. (2003), Yue et al. (2003) and Santen et al. (2004) in their various models. However, in parallel with the lack of effect on Ser118ER phosphorylation, MCF-7X cells always survived LY294002 plus PD98059 co-treatment, with resistance emerging at 12 weeks.
Therapeutic implications
The new MCF-7X model has consolidated the importance of increased ER signalling and the key contribution for intracellular kinases, notably PI3K/AKT, to growth of cells that have acquired resistance to oestrogen deprivation. While there does not seem to be a significant contribution of classical growth factor receptors such as EGFR, HER2 or IGF-1R in MCF-7X cells, PI3K interplay with transferrin receptor may be contributory and requires future detailed investigation. MCF-7X cell studies demonstrate the potential of inhibiting PI3K/AKT signalling to treat this disease and inhibitors of such pathways are emerging, while faslodex responses following clinical acquisition of resistance to aromatase inhibitors have recently been described (Johnston 2004). Importantly, we have also been able to demonstrate that PI3K/AKT cross-talks with ER via Ser167 activation (where any contribution for transferrin receptor in this cross-talk is currently being evaluated through transferrin challenge studies). There is an additional important MAPK/PI3K/AKT-independent regulation of Ser118ER phosphorylation in MCF-7X cells. The impact on ER phosphorylation of agents targeting these dominant regulatory routes (i.e. LY294002 and faslodex) appears central to subsequent inhibitory effects on ER transcriptional activity and growth. Importantly, however, MCF-7X cells reveal that the oestrogen-deprived phenotype is extremely flexible in its use of regulatory pathways to maintain ER phosphorylation during treatment. This readily provides a compensatory mechanism that limits initial therapeutic efficacy of agents applied singly and allows evolution of resistance in MCF-7X. Our data demonstrate that intelligent combination therapy to target alll regulators of ER AF-1 phosphorylation (i.e. PI3K/AKT inhibitor plus MAPK inhibitor plus faslodex in MCF-7X) may be required to abrogate this cell survival mechanism and thereby effectively inhibit acquired resistance to oestrogen deprivation and prevent subsequent therapeutic resistance.
Acknowledgements
Thanks to the Tenovus Cancer Research Centre Tissue Culture and Immunocytochemistry Units; to Dr Kathryn Taylor and Mr Richard McClelland for immunofluorescence and reporter gene assay guidance; to Mr David Burston and Miss Helena Dunn (Welsh School of Pharmacy) for technical assistance.
Funding
This research was generously supported by the Tenovus Organisation. There are no conflicts of interest to declare regarding this work.
References
Brodie AM & Njar VC 2000 Aromatase inhibitors and their application in breast cancer treatment. Steroids 65 171–179.
Brodie AH, Jelovac D & Long B 2003 The intratumoral aromatase model: studies with aromatase inhibitors and antiestrogens. Journal of Steroid Biochemistry and Molecular Biology 86 283–288.
van der Burg B, Rutteman GR, Blankenstein MA, de Laat SW & van Zoelen EJ 1988 Mitogenic stimulation of human breast cancer cells in a growth factor-defined medium: synergistic action of insulin and estrogen. Journal of Cell Physiology 134 101–118.
Campbell RA, Bhat-Nakshatri P, Patel NM, Constantinidou D, Ali S & Nakshatri H 2001 Phosphatidylinositol 3-kinase/AKT-mediated activation of estrogen receptor alpha: a new model for anti-estrogen resistance. Journal of Biological Chemistry 276 9817–9824.
Chen D, Washbrook E, Sarwar N, Bates GJ, Pace PE, Thirunuvakkarasu V, Taylor J, Epstein RJ, Fuller-Pace FV, Egly JM, Coombes RC & Ali S 2002 Phosphorylation of human estrogen receptor alpha at serine 118 by two distinct signal transduction pathways revealed by phosphorylation-specific antisera. Oncogene 21 4921–4931.
Daly RJ, Harris WH, Wang DY & Darbre PD 1991 Autocrine production of insulin-like growth factor II using an inducible expression system results in reduced estrogen sensitivity of MCF-7 human breast cancer cells. Cell Growth and Differentiation 2 457–464.
Elliott RL, Elliott MC, Wang F & Head JF 1993 Breast carcinoma and the role of iron metabolism. A cytochemical, tissue culture, and ultrastructural study. Annals of the New York Academy of Sciences 698 159–166.
Faridi J, Wang L, Endemann G & Roth RA 2003 Expression of constitutively active Akt-3 in MCF-7 breast cancer cells reverses the estrogen and tamoxifen responsivity of these cells in vivo. Clinical Cancer Research 9 2933–2939.
Fuqua SA, Wiltschke C, Zhang QX, Borg A, Castles CG, Friedrichs WE, Hopp T, Hilsenbeck S, Mohsin S, O’Connell P & Allred DC 2000 A hypersensitive estrogen receptor-alpha mutation in premalignant breast lesions. Cancer Research 60 4026–4029.
Guvakova MA & Surmacz E 1997 Overexpressed IGF-I receptors reduce estrogen growth requirements, enhance survival, and promote E-cadherin-mediated cell–cell adhesion in human breast cancer cells. Experimental Cell Research 231 149–162.
Inoue T, Cavanaugh PG, Steck PA, Brunner N & Nicolson GL 1993 Differences in transferrin response and numbers of transferrin receptors in rat and human mammary carcinoma lines of different metastatic potentials. Journal of Cell Physiology 156 212–217.
Jeng MH, Yue W, Eischeid A, Wang JP & Santen RJ 2000 Role of MAP kinase in the enhanced cell proliferation of long term estrogen deprived human breast cancer cells. Breast Cancer Research and Treatment 62 167–175.
Jensen J, Kitlen JW, Briand P, Labrie F & Lykkesfeldt AE 2003 Effect of antiestrogens and aromatase inhibitor on basal growth of the human breast cancer cell line MCF-7 in serum-free medium. Journal of Steroid Biochemistry and Molecular Biology 84 469–478.
Jess TJ, Belham CM, Thomson FJ, Scott PH, Plevin RJ & Gould GW 1996 Phosphatidylinositol 3'-kinase, but not p70 ribosomal S6 kinase, is involved in membrane protein recycling: wortmannin inhibits glucose transport and downregulates cell-surface transferrin receptor numbers independently of any effect on fluid-phase endocytosis in fibroblasts. Cellular Signalling 8 297–304.
Joel PB, Traish AM & Lannigan DA 1998 Estradiol-induced phosphorylation of serine 118 in the estrogen receptor is independent of p42/p44 mitogen-activated protein kinase. Journal of Biological Chemistry 273 13317–13323.
Johnston S 2004 Fulvestrant and the sequential endocrine cascade for advanced breast cancer. British Journal of Cancer 90 S15–S18.
de Jong PC, Blankenstein MA, Nortier JW, Slee PH, van de Ven J, van Gorp JM, Elbers JR, Schipper ME, Blijham GH, Thijssen JH, Lu Q, Jelovac D & Brodie AM 2003 The relationship between aromatase in primary breast tumors and response to treatment with aromatase inhibitors in advanced disease. Journal of Steroid Biochemistry and Molecular Biology 87 149–155.
Knowlden JM, Hutcheson IR, Jones HE, Madden T, Gee JMW, Harper ME, Barrow D, Wakeling AE & Nicholson RI 2003 Elevated levels of epidermal growth factor receptor/c-erbB2 heterodimers mediate an autocrine growth regulatory pathway in tamoxifen-resistant MCF-7 cells. Endocrinology 144 1032–1044.
Lannigan DA 2003 Estrogen receptor phosphorylation. Steroids 68 1–9.
Le Goff P, Montano MM, Schodin DJ & Katzenellenbogen BS 1994 Phosphorylation of the human estrogen receptor. Identification of hormone-regulated sites and examination of their influence on transcriptional activity. Journal of Biological Chemistry 269 4458–4466.
Martin LA, Farmer I, Johnston SR, Ali S, Marshall C & Dowsett M 2003 Enhanced estrogen receptor (ER) alpha, ERBB2, and MAPK signal transduction pathways operate during the adaptation of MCF-7 cells to long term estrogen deprivation. Journal of Biological Chemistry 278 30458–30468.
Massarweh SA, Jiang S, Mohsin SK, Di Pietro M, Wakeling AE, Osborne CK & Schiff R 2003 Resistance to endocrine therapy in a xenograst model of HER-2 overexpressing breast cancer is accompanied by increased HER-2 but loss of IGF-1 receptor expression. Breast Cancer Research and Treatment 82 S1007.
Nicholson RI, Hutcheson IR, Knowlden JM, Jones HE, Harper ME, Jordan N, Hiscox SE, Barrow D & Gee JM 2004a Nonendocrine pathways and endocrine resistance: observations with antiestrogens and signal transduction inhibitors in combination. Clinical Cancer Research 10 346S–354S.
Nicholson RI, Staka C, Boyns F, Hutcheson IR & Gee JM 2004b Growth factor-driven mechanisms associated with resistance to estrogen deprivation in breast cancer: new opportunities for therapy. Endocrine Related Cancer 11 623–641.
Pasapera Limon AM, Herrera-Munoz J, Gutierrez-Sagal R & Ulloa-Aguirre A 2003 The phosphatidylinositol 3-kinase inhibitor LY294002 binds the estrogen receptor and inhibits 17beta-estradiol-induced transcriptional activity of an estrogen sensitive reporter gene. Molecular and Cellular Endocrinology 200 199–202.
Santen RJ 2003 Inhibition of aromatase: insights from recent studies. Steroids 68 559–567.
Santen RJ, Song RX, Zhang Z, Yue W & Kumar R 2004 Adaptive hypersensitivity to estrogen: mechanism for sequential responses to hormonal therapy in breast cancer. Clinical Cancer Research 10 337S–345S.
Shim WS, Conaway M, Masamura S, Yue W, Wang JP, Kmar R & Santen RJ 2000 Estradiol hypersensitivity and mitogen-activated protein kinase expression in long-term estrogen deprived human breast cancer cells in vivo. Endocrinology 141 396–405.
Shou J, Massarweh S, Osborne CK, Wakeling AE, Ali S, Weiss M & Schiff R 2004 Mechanisms of tamoxigen resistance: increased estrogen receptor-HER2/neu cross-talk in ER/HER2-positive breast cancer. Journal of the National Cancer Institute 96 926–935.
Song RX, Barnes CJ, Zhang Z, Bao Y, Kumar R & Santen RJ 2004 The role of Shc and insulin-like growth factor 1 receptor in mediating the translocation of estrogen receptor alpha to the plasma membrane. PNAS 101 2076–2081.
Stephen RL, Shaw LE, Larsen C, Corcoran D & Darbre PD 2001 Insulin-like growth factor receptor levels are regulated by cell density and by long term estrogen deprivation in MCF7 human breast cancer cells. Journal of Biological Chemistry 276 40080–40086.
Vandewalle B, Hornez L, Revillion F & Lefebvre J 1989 Secretion of transferrin by human breast cancer cells. Biochemical and Biophysical Research Communications 163 149–154.
Winer EP, Hudis C, Burstein HJ, Wolff AC, Pritchard KI, Ingle JN, Chlebowski RT, Gelber R, Edge SB, Gralow J, Cobleigh MA, Mamounas EP, Goldstein LJ, Whelan TJ, Powles TJ, Bryant J, Perkins C, Perotti J, Braun S, Langer AS, Browman GP & Somerfield MR 2005 American Society of Clinical Oncology technology assessment on the use of aromatase inhibitors as adjuvant therapy for postmenopausal women with hormone receptor-positive breast cancer: status report. Journal of Clinical Oncology 23 619–629.
Yue W, Wang JP, Conaway MR, Li Y & Santen RJ 2003 Adaptive hypersensitivity following long-term estrogen deprivation: involvement of multiple signal pathways. Journal of Steroid Biochemistry and Molecular Biology 86 265–274.(Cindy M Staka, Robert I N)