- Email: [email protected]
Adaptive hypersensitivity to estradiol: potential mechanism for secondary hormonal responses in breast cancer patients
Journal of Steroid Biochemistry & Molecular Biology 79 (2001) 115–125
Adaptive hypersensitivity to estradiol: potential mechanism for secondary hormonal responses in breast cancer patients夽 Richard Santen a,∗ , Meei-Huey Jeng b , Ji-Ping Wang c , Robert Song c , Shigeru Masamura d , Robert McPherson c , Steven Santner e , Wei Yue c , Woo-Shin Shim f a
Division of Endocrinology, University of Virginia Health System, P.O. Box 800379, Jefferson Park Avenue, Charlottesville, VA 22908, USA b Department of Medicine, Section of Hematology/Oncology, Indiana University, Indianapolis, IN 46202-5121, USA c Division of Hematology/Oncology, University of Virginia Health System, P.O. Box 800747, Charlottesville, VA 22908, USA d Department of General Surgery, School of Medicine, Keio University, Shinanomachi 35, Shinjuku, Tokyo 160-8582, Japan e Karmanos Cancer Institute, 110 East Warren, Detroit, MI 48201, USA f Department of General Surgery, Gacheon Medical College Ghil Hospital, Incheon 400-170, South Korea
Abstract Women with hormone dependent breast cancer initially respond to hormone deprivation therapy with tamoxifen or oophorectomy for 12–18 months but later relapse. Upon secondary therapy with aromatase inhibitors, patients often experience further tumor regression. The mechanisms responsible for secondary responses are unknown. We postulated that hormone deprivation induces hypersensitivity to estradiol. Evidence of this phenomenon was provided in a model system involving MCF-7 cells grown in vitro and in xenografts. To determine if the ER transcriptional process is involved in hypersensitivity, we examined the effect of estradiol on ER reporter activity, PgR, PS2, and c-myc as markers and found no alterations in hypersensitive cells. Next, we examined whether MAP kinase may be upregulated in the hypersensitive cells as a reflection of increased growth factor secretion or action. Basal MAP kinase activity was increased both in vitro and in vivo in hypersensitive cells. Proof of principle studies indicated that an increase in MAP kinase activity induced by TGF␣ administration caused a two- to three-fold shift to the left in estradiol dose response curves in wild type cells. Blockade of MAP kinase with PD98059 returned the shifted curve back to baseline. These data suggested that MAP kinase overexpression could induce hypersensitivity. To determine why MAP kinase was increased, we excluded constitutive receptor activity and growth factor secretion by the demonstration that the pure anti-estrogen, ICI 182780, could inhibit MAP kinase activation. We also excluded hypersensitivity to estradiol induced growth factor secretion, and thus MAP kinase activation, since estradiol stimulated MAP kinase at 24, 48, and 72 h at the same concentrations in hypersensitive as in wild type cells. Surprisingly, a series of experiments suggested that MAP kinase increased in hypersensitive cells as a result of estrogen activation via a non-genomic pathway. We examined the classical signal pathway in which SHC is phosphorylated and binds to SOS and GRB-2 to activate Ras, Raf, and MAP kinase. With 5–20 min of exposure, estradiol caused binding of SHC to the estrogen receptor, phosphorylation of SHC, binding of GRB-2 to SOS, and activation of MAP kinase. All of these affects could be blocked by ICI 182780. Taken together, these observations suggest that the cell membrane ER pathway may be responsible for upregulation of MAP kinase and hypersensitivity in cells adapted to estradiol deprivation. © 2002 Elsevier Science Ltd. All rights reserved. Keywords: Aromatase; Breast cancer; Estrogen; Hypersensitivity; MAP kinase; Non-genomic estrogen effects; Proliferation
1. Introduction Breast cancer will be diagnosed in 180,000 patients in the United States in 2000 and nearly 44,000 will die of this disease [1]. One-third of women with locally advanced or metastatic disease experience tumor regression in response to hormonal therapy [2]. However, adaptive mechanisms intervene and these tumors begin to re-grow within a period 夽 Proceedings of the Symposium: ‘Aromatase 2000 and the Third Generation’ (Port Douglas, Australia, 3–7 November 2000). ∗ Corresponding author. Tel.: +1-804-924-2961; fax: +1-804-924-9616. E-mail address: [email protected] (R. Santen).
of 12–18 months on average. Secondary hormonal therapies frequently induce additional responses of similar duration [2]. The mechanisms whereby secondary responses occur have never been satisfactorily explained. Clinical observations with inhibitors of aromatase, the enzyme catalyzing the rate limiting step in estradiol biosynthesis, suggested to us that hypersensitivity to estradiol might explain these secondary responses. In post-menopausal women with breast cancer, inhibition of aromatase suppresses estrogen production [2]. In initial clinical studies with the first generation inhibitor, aminoglutethimide, we were surprised to observe secondary tumor regressions in women who had previously responded to surgical oophorectomy but later relapsed [3].
0960-0760/02/$ – see front matter © 2002 Elsevier Science Ltd. All rights reserved. PII: S 0 9 6 0 - 0 7 6 0 ( 0 1 ) 0 0 1 5 1 - 0
116
R. Santen et al. / Journal of Steroid Biochemistry & Molecular Biology 79 (2001) 115–125
Following initial treatment with oophorectomy, estrogen levels fell from premenopausal levels of 50–600 pg/ml to approximately 10 pg/ml and in response, tumors regressed for a period of 12–18 months [4]. Upon relapse after oophorectomy, estradiol levels remained at the 10 pg/ml level. Subsequent treatment with aromatase inhibitors lowered estradiol further to 1–5 pg/ml and caused secondary tumor regressions. These observations first led us to consider the possibility that tumors might adapt by becoming hypersensitive to estradiol upon hormonal deprivation. Adaptation to hormonal therapy may also occur in women receiving adjuvant hormonal therapy prior to the appearance of metastatic or locally advanced disease [5]. In such women, tamoxifen reduces the rate of recurrence and enhances survival if administered for 1 to 5 years. When given for 10 years as opposed to 5, no additional benefit is observed and an increased number of recurrences and deaths have been reported [6,7]. Observations in xenograft models of breast cancer suggest that these neoplasms may adapt to tamoxifen such that this anti-estrogen becomes an estrogen agonist after a period of 6 months to 1 year [8]. We and others have suggested that hypersensitivity to the estrogenic effects of tamoxifen may occur in women receiving adjuvant tamoxifen. This mechanism might explain the adverse effects of 10 versus 5 years of adjuvant tamoxifen therapy [9,10]. Taken together, these observations emphasize the fact that tumors can adapt to their environment and develop means to re-grow under conditions that previously limited growth. An understanding of these adaptive processes would provide targets for development of specific therapies to inhibit re-growth and to enhance the efficacy of hormonal therapies.
2. Estradiol hypersensitivity hypothesis Our hypothesis suggests that breast tumors adapt to estrogen deprivation by developing increased sensitivity to estradiol [11]. Re-growth then occurs because of the ability of tumors to utilize very small amounts of estrogen to induce a proliferative response. Differing levels of sensitivity to hormones is a widely recognized phenomenon. The range of sensitivity to hormones varies substantially among different tissues and under a wide spectrum of physiologic conditions [12,13]. Several examples of hypersensitivity to estradiol have been described. A genetically engineered yeast system described by McDonnell et al. responds to levels as low as 10−14 M estradiol [13]. A prolactin producing pituitary tumor described by Chun et al. exhibits enhanced proliferation when exposed to 10−14 M levels of both DES and estradiol [14]. With respect to androgens, clones of Shionogi breast cancer cells exhibit marked ranges in sensitivity to the proliferative responses produced by testosterone [15,16]. The range in sensitivity of various clones varies by a factor of 1000 and some clones respond to levels of testosterone of 10−12 M.
The set point for sensitivity to various hormones can be dynamically determined and altered by the level of ambient hormonal exposure. Prepubertal girls are highly sensitive to the suppressive effects of estradiol on gonadotropin secretion [17]. Upon exposure to increasing amounts of estrogen, their level of sensitivity of the feedback mechanism diminishes [18]. Seasonally breeding ewes exhibit exquisite sensitivity to estradiol during the non-breeding season with much reduced sensitivity during the breeding season [19]. Secondary amenorrhea in women is associated with changes in the set point of sensitivity to estradiol [20]. Several experiments demonstrate shifts in dose response curves to the right or left as a reflection of change in sensitivity [12]. Sensitivities to both glucocorticoid and progesterone are enhanced when specific DNA binding proteins interact with elements as far as 1–3 kb upstream of the promoter regions for their respective receptors [21,22]. Multiple other mechanisms are possible, including events occurring well downstream of the estrogen receptor. The ability of breast cancer cells to adapt to lower levels of estradiol has clinical relevance regarding design of hormonal therapies for patients. The adaptive hypersensitivity concept suggests that the greater the suppression of estradiol, the more effective would be the anti-tumor efficacy of aromatase inhibitors. This raised the possibility that the highly potent third generation aromatase inhibitors would be more effective than the first generation inhibitor, aminoglutethimide or tamoxifen. Large clinical trials confirmed this supposition [23–26]. An additional implication of the hypersensitivity hypothesis relates to choice of inhibitor dosage. It was reasoned that breast cancer cells might respond to amounts of estradiol well below the level of detection with sensitive plasma radioimmunoassays. Notably, in the cell culture model, 10−14 M estradiol, a concentration well below assay sensitivity, stimulates cell proliferation [11]. Based upon this reasoning, it would be logical to compare a dose of aromatase inhibitor which suppresses estradiol to undetectable levels with a much higher dose of the same inhibitor. Clinical trials to test this concept involved 945 breast cancer patients and compared responses in women receiving 0.5 mg of letrozole, a dose which suppresses plasma estradiol to undetectable levels, with a five-fold higher dose, 2.5 mg daily. In both studies, the 2.5 mg dose resulted either in prolongation of patient survival (P < 0.01) or in a higher percent of objective response rate (12.5% versus 23.5%, P < 0.01). These observations supported the clinical relevance of the concept of adaptive hypersensitivity [24–26].
3. Model systems of adaptive hypersensitivity Observations in models of breast and prostate carcinoma support the concept of adaptive hypersensitivity [27–36]. We utilized MCF-7 breast cancer cells grown in culture under conditions of estradiol deprivation as a model [11]. We call these LTED cells to reflect their Long-Term
R. Santen et al. / Journal of Steroid Biochemistry & Molecular Biology 79 (2001) 115–125
117
Fig. 1. Increased basal growth rate and inhibition by ICI 182780 in LTED cells as compared with wild type MCF-7 cells. (A) Wild type cells were deprived of estrogen for 3 days and plated onto six-well plates at the same time as LTED cells, at a density of 6 × 10−5 cells per well. The next day (day 1), cells were lysed and cell nuclei were then counted every day for the next 10 days using a Coulter counter. (B) LTED cells were plated onto six-well plates at a density of 1 × 10−5 cells per well. ICI 182780 at various concentrations was added the next day for 6 days. Media were changed every other day and cells were lysed and cell nuclei were then counted using the Coulter counter. Medium was changed every other day. Data are presented as mean ± S.E. from duplicate wells and S.E. when error bars were sufficiently large to visualize. Reprinted with the permission of publisher and authors from [54].
Estradiol Deprivation (we will use this term throughout this manuscript). These cells initially stop growing under conditions of reduced estrogen achieved by charcoal stripping of serum used in the culture media. Later, they grow maximally under these conditions and cannot be further stimulated by exogenous estrogen but are inhibited with “pure” anti-estrogens (Fig. 1A and B). To explain these observations, we hypothesized that LTED cells are hypersensitive to the residual estradiol present in culture media and consequently, can grow in charcoal stripped serum without addition of exogenous estradiol [11]. Three experimental approaches supported this possibility. Firstly, when special methods were used to remove all residual estrogen from the media, LTED cells grew minimally without exogenous estradiol. Under these conditions, LTED cells were stimulated to grow in response to levels of estradiol of 10−14 M whereas wild type cells required 10−11 M (Fig. 2). Secondly, when the residual estrogen was antagonized by adding a pure anti-estrogen, LTED cells grew only minimally. Addition of estradiol under these conditions caused LTED cells to re-grow at two log lower concentrations of estradiol than similarly treated wild type cells (Fig. 3) [37]. Thirdly, LTED cell xenografts in vivo respond to lower doses of estradiol for growth than wild type cells [38]. The in vivo studies utilized xenografts of wild type and LTED cells implanted into both sides of the backs of castrated nude mice. A detailed description of the methods and results has been previously reported [38]. Briefly, initial studies validated the ability to “clamp” estradiol levels at 1.25, 2.5, 5, 10, and 20 pg/ml using silastic implants filled with mixtures of cholesterol and estradiol. Since such low levels cannot be reliably measured in castrate nude mouse plasma, we utilized uterine weight as a bioassay to demonstrate linearity of plasma estradiol levels. Prior published studies which induced much higher levels of estradiol
(i.e. 20–800 pg/ml) had demonstrated linearity between the cholesterol/estradiol ratio in the silastic implants and plasma estradiol as measured by RIA [11]. We next implanted two LTED and two wild type xenografts onto the backs of each castrate female mouse using our newly developed
Fig. 2. E2 dose–response curves in LTE deprived cells using serum-free medium and wild type MCF-7 cells grown in dextran coated charcoal (DCC) striped serum containing medium. LTE deprived cells (50,000 cells per flask) were seeded in medium containing DCC stripped serum into five replicate T-25 flasks for each group. After 1 day of acclimatization, they were transferred to serum-free medium for 1 week. The media were then changed to experimental media (serum-free medium containing ethanol vehicle 0.1%); E2 at 10−15 to 10−6 M, or ICI 164384 at 10−6 M. Cells were re-fed with experimental serum-free media 3 days later and counted after another 3 days. Wild type cells were grown in media containing DCC treated serum. Data are illustrated as a percentage of the maximum response (±SEM) to normalize data between these two experiments. (䊉) Wild type MCF-7 cells; (䊊) LTE deprived cells. Reprinted with the permission of publisher and authors from [11].
118
R. Santen et al. / Journal of Steroid Biochemistry & Molecular Biology 79 (2001) 115–125
Fig. 3. Dose response effects of estradiol on growth of LTED and wild type MCF-7 cells grown in the presence of 10−9 M ICI 182780.
heteroimplant technique. This method eliminates the confounding effect of the individual animal on tumor growth since both wild type and LTED xenografts are present in the same animal. We then allowed tumors to become established over a 1 month period. This is a key step which eliminates the confounding effect of estradiol on establishment of tumor rather than on proliferation. After 1 month, animals received silastic implants to clamp estradiol at levels of 1.25, 2.5, 5, 10, and 20 pg/ml and tumor volumes were measured by calipers weekly for 8 weeks.
As shown in Fig. 4, these in vivo findings paralleled those routinely observed in vitro. Specifically, LTED tumors grew in response to much lower concentrations of estradiol than wild type as a reflection of hypersensitivity. The LTED cell xenografts grew to a greater extent than wild type xenografts in response to estradiol levels of 1.25 and 2.5 pg/ml (P < 0.0001 for both). When E2 levels increased further, a peak growth rate was reached and then growth diminished, producing a typical bell shaped dose response curve (see Fig. 4). Specifically, both LTED and wild type xenografts grew equally at the 5 pg/ml dosage but at the 10 and 20 pg/ml doses, the wild type tumors grew to a greater extent than LTED. These results indicate that LTED cells exhibit increased sensitivity to both the initial stimulatory and later inhibitory levels of estradiol. The process of development of hypersensitivity could represent either cell selection or adaptation. Supporting the concept of adaptation was the plasticity in reversing and re-inducing the hypersensitive phenotype. When re-exposed to estradiol containing media in vitro for periods of 1 month or longer, the LTED cells reverted back to the wild type phenotype [11]. Deprivation of estradiol again induced the hypersensitive phenotype. The process of reversion and induction could be reproducibly repeated over a period of 2 years, demonstrating the plasticity of this system and supporting the concept of adaptation rather than cell selection [11].
Fig. 4. (A) Growth curves in wild type and LTED tumors in oophorectomized nude mice receiving only cholesterol containing SILASTIC brand implants (vehicle control) and implants maintaining plasma E2 levels at 1.25, and 2.5 pg/ml. The statistical significance of the differences between wild type and LTED tumors is indicated on each panel. The accompanying bars ± SEM represent mean area under the curve for each group and are shown to illustrate variance among groups. The black bars are representative of volumes of LTED tumors and the cross-hatched bars of wild type tumors. The statistical significance indicated represents paired comparisons between integrated tumor volumes and not between mean areas under the various curves. Integrated tumor volumes were significantly higher in the LTED than wild type tumors in response to 1.25 and 2.5 pg/ml but not in oophorectomized animals. (B) Growth curves in wild type and LTED tumors with plasma E2 clamped at 5, 10 and 20 pg/ml. Reprinted with the permission of publisher and authors from [38].
R. Santen et al. / Journal of Steroid Biochemistry & Molecular Biology 79 (2001) 115–125
The process of dynamic adaptation to hormones is not limited to our breast tumor model. Martin et al. also demonstrated development of hypersensitivity in MCF-7 cells upon deprivation of estradiol [39]. Hormone dependent prostate cancer cells also adapt to conditions of low androgen in culture by increasing their sensitivity to androgens [40–42]. Kirschenbaum et al. [40] demonstrated growth stimulation with 10−14 M DHT in LnCAP cells which were conditioned by long-term exposure to low androgen concentrations. Wild type cells required 10−10 M. Kokontis et al. [41,42] also demonstrated adaptive hypersensitivity in cloned LnCAP prostate cancer cells. Upon androgen deprivation, these cells developed one log enhanced sensitivity to the synthetic androgen R-1881 and then reverted to a two to three log less sensitive state upon androgen re-exposure. The observations in prostate cancer cell lines suggest that the process of adaptive hypersensitivity may be common to hormone dependent cancer and not limited to one breast cancer cell line. This phenomenon could explain why men with prostate cancer respond initially to castration, later relapse, and then experience secondary tumor regression upon blockade of androgen action with potent anti-androgens [43].
4. Mechanisms of hypersensitivity Our laboratory has been focusing upon potential mechanisms to explain hypersensitivity to estradiol in the LTED model system. To facilitate these experiments, we have recently modified experimental conditions in our model to alleviate problems from residual estrogen contamination of culture media. LTED cells grow in standard charcoal stripped serum as a result of residual estrogen and growth is inhibited with anti-estrogens (Figs. 1 and 3) [37]. Sources of the residual estrogenic activity in culture media are unknown but recent reports suggest that components of plastic and many other industrial substances act as estrogen agonists. Stimulatory effects from these contaminants are not surprising since the LTED cells are sensitive to concentrations of estradiol as low as 10−15 M (Fig. 2) [11,44–48]. Our method of enhanced stripping of sera with charcoal, as originally reported [11], or use of serum-free media results in a variable rate of growth of LTED cells in the absence of exogenous estrogen. Accordingly, we have developed a practical strategy to block residual estrogen activity with pure anti-estrogens. This method involves antagonism of residual estrogen by addition of identical amounts of anti-estrogen to wild type and LTED cells. We can then precisely determine dose responses to exogenous estradiol in wild type and LTED cells in a reproducible fashion (Fig. 3). Under these conditions, the LTED cells respond to two to three log lower concentrations of estrogen than do wild type cells and are not effected by variable levels of residual estrogen in media. In other experiments, we do not antagonize residual estrogen with anti-estrogens so that we can probe the factors stimulating growth of LTED cells under these conditions.
119
As a conceptual means to facilitate analysis, we arbitrarily divide estradiol mediated events into genomic and non-genomic events. Both pathways can interact with classical growth factor pathways with “cross talk” at an ER transcriptional level or at the level of the cell cycle. A variety of studies indicate that growth factor secretion or action can be stimulated by estradiol or can be regulated independently of estradiol [49–53]. The mechanism for mediation of hypersensitivity could involve either genomic or non-genomic events acting at upstream or downstream sites of action.
5. Genomic effects of estradiol We initially focused upon events involving the estrogen receptor and its regulation of transcription as a possible inducing hypersensitivity. We demonstrated that long-term deprivation of estradiol causes an upregulation of estrogen receptor content from 3- to 20-fold (Fig. 5). These results included assessment of the number of binding sites by radioligand analysis [11] and level of ER alpha message as well as quantitation of the amount of ER by western analysis using antibodies recognizing various epitopes of the ER (Fig. 5) [54]. In addition, we found that basal rates of transcription of the ER were increased and could be inhibited by the pure anti-estrogen, ICI 182780 [54]. Even though upstream processes were altered by estradiol deprivation, we considered it important to determine if these receptor changes were mechanistically liked to the process of hypersensitivity. To do this, we measured several parameters reflecting estrogen receptor mediated transcription. Specifically, we administered estradiol in doses from 10−14 to 10−6 M in LTED and wild type cells and measured ER reporter gene activity as well as c-myc, progesterone receptor, PS2, and MAP kinase. We observed similar dose response stimulation of several ER reporter constructs, of progesterone receptor content, of c-myc message levels, of PS2 and of activated MAP kinase content (data not shown). Taken together, these results suggested that hypersensitivity is not mediated primarily at the level of ER transcription through classical genomic effects of estradiol.
6. Growth factor pathways We then questioned whether growth factor pathways might be turned on in LTED cells and whether this might be the mechanism for enhanced sensitivity to estradiol. Initial studies examined the effect of inhibitors of several steps in the growth factor pathway on cell proliferation in LTED cells. Growth factors bind to cell membrane receptors, activate receptor tyrosine kinases, cause isoprenylation of Ras, and activate MAP kinase by phosphorylation. We found that inhibitors of each of these steps produced a dose dependent inhibition of tritated thymidine uptake in LTED cells as evidence of a reduction in cell proliferation [55] (Fig. 6).
120
R. Santen et al. / Journal of Steroid Biochemistry & Molecular Biology 79 (2001) 115–125
transcription. Under basal conditions, ER reporter activity was higher in LTED than in wild type cells (Fig. 8). However, this enhanced basal activity could not be inhibited by the MAP kinase inhibitor, PD98059 (Fig. 8). On the other hand, estradiol stimulated reporter activity could be partially but not completely inhibited by PD98059 (Fig. 9). These observations suggested that elevated MAP kinase can partially influence estradiol stimulated but not basal ER transcription. Effects of MAP kinase at this step do not explain hypersensitivity since the dose response effects of estradiol on ERE reporter activity are not right shifted by MAP kinase inhibition.
7. Proof of principle experiment
Fig. 5. ER protein expression in wild type and LTED cells. Wild type MCF-7 cells were deprived of estrogen and treated with or without 10−10 M E2 for 2 days. Cell cytosols were collected and 50 g protein (B and C) or 100 g protein (D) were loaded per lane and separated by electrophoresis on 10% polyacrylamide gels containing 1% SDS and transferred onto nitrocellulose membranes. Western blot analysis was performed as described in the original publication. ER antibody C314 was used in panel B and ER antibodies H226, D547, H222 and D75 were used in panels C and D. An epitope map of these antibodies is shown in panel A. Regions A–F and exons 1–8 are marked. Numbers marked above each antibody were the estimated amino acid ranges for the epitope locations. Ponseau S staining of the nitrocellulose membranes after transfer of proteins showed equal loading of proteins for all lanes. Reprinted with the permission of publisher and authors from [54].
These observations provided evidence that growth factor pathways played a major role in mediating cell proliferation in LTED cells. As further evidence of this possibility, we measured the levels of MAP kinase in LTED cells in vitro in comparison with wild type cells and found a marked increase [55] (Fig. 7). In vivo studies using immunohistochemistry to detect activated MAP kinase also demonstrated increased activation of MAP kinase in LTED xenografts. Independent of estradiol dosage, the percentage of MAP kinase positive cells in LTED cell xenografts ranged from 18 to 22% [38]. In contrast, wild type xenografts from castrated animals contained only 3% of MAP kinase activated cells when given vehicle as a control. The wild type xenografts exhibited a maximum dose dependent increase to 6% when exposed to increasing concentrations of estradiol [38]. We next questioned whether elevated MAP kinase might act to phosphorylate the ER and enhance its level of
Our observations linked hypersensitivity to MAP kinase overexpression but did not provide evidence of causality. Accordingly, we designed a proof of principle experiment to increase the level of MAP kinase and examine the effect of this maneuver on sensitivity to estradiol. To do this, we administered TGF␣ and determined that the optimal dose necessary to stimulate MAP kinase was 10 ng/ml and that the MAP kinase inhibitor PD98059 blocked this response (data not shown). We then demonstrated that wild type MCF-7 cells exposed to 10 ng/ml TGF␣ grew in response to two log lower concentrations of estradiol than did the same cells not given TGF␣. The induction of hypersensitivity reflected an increase in MAP kinase rather than another effect of TGF␣. Co-administration of the MAP kinase inhibitor, PD98059, completely abrogated the effects on TGF␣ on hypersensitivity [37]. These data suggested that MAP kinase might be involved in estradiol hypersensitivity observed in the LTED cells. To provide direct evidence of this, we administered a more potent MAP kinase inhibitor, U 0126, to LTED cells and demonstrated a one log shift toward the right in the estradiol dose response curve in them. This provided direct biologic evidence that elevated MAP kinase increased the level of sensitivity of estradiol in LTED breast cancer cells.
8. Downstream mechanisms The data presented indicate that growth factor pathways, when upregulated, can enhance the level of sensitivity to estradiol. The mechanism to explain this effect was suggested by a series of observations by Leone et al. [56,57]. These investigators provided evidence that Ras mediated events act synergistically with c-myc to stimulate production of the important cell cycle protein, E2F1. If this hypothesis is correct, less c-myc would be necessary to stimulate cell proliferation if Ras pathways were activated. Our studies and those of Watson et al. [58] demonstrated that estradiol stimulates c-myc and that this proto-oncogene is involved in estradiol stimulated proliferation [54]. Both c-myc and the Ras/MAP kinase pathways are activated in LTED cells [54,55]. It would be expected then that E2F1 would be
R. Santen et al. / Journal of Steroid Biochemistry & Molecular Biology 79 (2001) 115–125
121
Fig. 6. (A) Diagrammatic representation of the pathway involving growth factors binding to the plasma membrane, autophosphorylation of the receptor via tyrosine kinase, isoprenylation of Ras, activation of Raf, and activation of MAP kinase which then mediates enhancement of cell proliferation. (B) Effect of various inhibitors on 3 h thymidine uptake in LTED cells. LTED cells were plated onto six-well plates. The next day, cells were washed and fresh phenol red-free IMEM containing various inhibitors at indicated concentrations were added to the monolayers. After 18 h, cells were labeled with 3H-thymidine in the presence or absence of various inhibitors for 4 h. Cell monolayers were then processed for the uptake level of 3H-thymidine uptake. Data are presented as a mean of duplicate samples. The experiments were repeated at least two times. Reprinted from [55].
122
R. Santen et al. / Journal of Steroid Biochemistry & Molecular Biology 79 (2001) 115–125
Fig. 7. (A) Activated and total MAP kinase activity in wild type (WT) and LTED cells. (B) Results from the scanning densitometry of the activated MAP kinase as shown in the blots in A. Reprinted from [55].
increased in LTED as opposed to wild type cells and indeed this proved to be the case [59]. Based upon these data, our current working hypothesis to explain adaptive hypersensitivity is that activation of growth factor/MAP kinase pathways acts synergistically with c-myc at the level of the cell cycle to increase E2F1 levels. This synergy sensitizes cells to the proliferative effects of estradiol and allows growth of LTED cells in the presence of very low levels of estradiol.
9. Non-genomic effects of estradiol Our studies reviewed to date indicated that LTED cells are characterized by upregulation of MAP kinase activity. We initially postulated that this might have resulted from constitutive activation of growth factor receptors or growth factor secretion. We reasoned that either of these two possibilities
Fig. 8. Effect of PD98059 on ERE-CAT activity in both wild type and LTED cells under basal (i.e. estrogen depleted) culture conditions. Cells were plated and transfected with pERE-tk-CAT and pCMVbeta gal and treated with PD98059 at various concentrations for 2 days. Cells were collected and assayed for CAT activity using equal amounts of beta galactosidase. Reprinted from [55].
could be excluded if a pure anti-estrogen suppressed MAP kinase activity. Notably, ICI 182780 reduced the basal levels of MAP kinase in LTED cells by approximately two-third. Consequently, we reasoned that the estrogen receptor must somehow be involved in this process but that genomic effects on transcription were not involved. Another possibility was that estradiol exerted non-genomic effects on MAP
Fig. 9. Effect of PD98059 on E2 induced ERE-CAT activity and activated MAP kinase level in both wild type and LTED cells. (A) Wild type MCF-7 cells were deprived of estrogen and transfected with pERE-tk-CAT and pCMV beta gal. Cells were then treated with various concentrations of E2 in the absence or presence of PD98059 (10 g/ml) for 2 days. Cells were then collected and assayed for CAT activity using equal amounts of beta galactosidase. (B) Wild type MCF-7 cells were deprived of estrogen for 2 days and treated with E2 in the absence or presence of PD98059 (10 g/ml) for 12 h. Total MAP kinase and activated MAP kinase were measured using specific antibodies. Reprinted from [55].
R. Santen et al. / Journal of Steroid Biochemistry & Molecular Biology 79 (2001) 115–125
kinase in LTED cells. To test this, we examined the ability of estradiol to stimulate MAP kinase over a 120 min period. We observed an increase in MAP kinase at 15 min and further increases at 60–120 min. Dose response studies demonstrated peak effects at 10−10 M estradiol [60]. To further dissect the mechanism of non-genomic estrogen effects, we examined the possibility that estradiol could activate a classical pathway involving SHC, phosphorylated SHC, GRB-2 and SOS in LTED cells. Within 10 min of exposure to estradiol, a complex involving phosphorylated SHC and the ER alpha formed and SOS bound to GRB-2. These effects could be abrogated by co-administration of the pure anti-estrogen ICI 182780 [60]. Taken together, our data support the possibility that the LTED cells co-opt a classical growth factor pathway to allow estradiol to exert non-genomic effects on MAP kinase.
123
cells predominantly utilize the genomic pathway. Through downstream interactions on the cell cycle, MAP kinase mediated events synergize with the genomic effects of estradiol on transcription. Under these circumstances, smaller amounts of estradiol are required to induce cell proliferation than in wild type cells. The series of studies presented focus on the ability of breast cancer cells to adapt to various therapies. The possibility that tumors become hypersensitive to estradiol provides a plausible explanation why third generation aromatase inhibitors are more effective than the first and second generation inhibitors. The magnitude of estradiol inhibition is greater with the third generation inhibitors. In the long run, means to counteract the adaptive process will be necessary to prevent progression of breast cancer cells to a hormone independent state. Further work to explain this adaptive process should result in development of better therapies for patients and perhaps even cure if applied at early stages of disease.
10. Summary Our data indicate that breast cancer cells can adapt to long-term estradiol deprivation and develop means to re-grow in the presence of very small concentrations of estradiol. Adaptation involves upstream mechanisms with an increase in estrogen receptor level and in basal transcription rates. However, this does not appear to be the primary event mediating hypersensitivity since LTED cells do not exhibit enhanced sensitivity when levels of ER reporter activity, progesterone receptor levels, c-myc levels, or MAP kinase activation serve as endpoints of ER transcriptional rates. Overexpression of growth factor pathways are involved in the hypersensitivity process. The LTED cells depend upon activation of growth factor pathways for proliferation when exogenous estradiol is not added to culture media. Proof of principle experiments demonstrate that administration of TGF␣ can shift estradiol dose response curves two logs to the left and that this effect can be abrogated by co-administration of a MAP kinase inhibitor. Finally, hypersensitivity to estradiol can be reduced in LTED cells with administration of a MAP kinase inhibitor. Taken together, these results suggest participation of growth factor pathways in the process of adaptive hypersensitivity. Measurements of the transcription factor E2F1 suggest that the actual process of hypersensitivity occurs at the level of the cell cycle and involves synergistic interactions between Ras mediated events and c-myc. A key question is why MAP kinase is elevated in LTED cells. This does not appear to be due to constitutive activation of growth factor receptors or growth factor secretion since MAP kinase activity can be blocked by a pure anti-estrogen. Non-genomic effects of estradiol on the other hand appear to be responsible. We have presented evidence for this which includes activation of the SHC, phosphorylated SHC, SOS, GRB-2, MAP kinase pathway. Based upon these data, we propose a working model for hypersensitivity in which LTED cells utilize both non-genomic and genomic pathways of estrogen action whereas the wild type MCF-7
References [1] S.H. Landis, T. Murray, S. Bolden, P.A. Wingo, Cancer statistics, CA Cancer J. Clin. 48 (1998) 6–29. [2] R.J. Santen, A. Manni, H. Harvey, C. Redmond, Endocrine treatment of breast cancer in women, Endocrinol. Rev. 11 (1990) 221–265. [3] R.J. Santen, T.J. Worgul, A. Lipton, H. Harvey, A. Boucher, E. Samojlik, S.A. Wells, Aminoglutethimide as treatment of postmenopausal women with advanced breast carcinoma, Ann. Int. Med. 96 (1982) 94–101. [4] R.J. Santen, T.J. Worgul, E. Samojlik, A. Interrante, A.E. Boucher, A. Lipton, H.A. Harvey, D.S. White, E. Smart, C. Cox, S. Wells, A randomized trial comparing surgical adrenalectomy with aminoglutethimide plus hydrocortisone in women with advanced breast cancer, N. Engl. J. Med. 305 (1981) 545–551. [5] Early Breast Cancer Collaborative Group, Tamoxifen for early breast cancer: an overview of the randomized trials, Lancet 351 (1998) 1451–1467. [6] B. Fisher, J. Dignam, J. Bryant, et al., Five versus more than five years of tamoxifen therapy for breast cancer patients with negative lymph modes and estrogen receptor-positive tumors, J. Natl. Cancer Inst. 88 (1996) 1529–1542. [7] H.J. Stewart, A.P. Forrest, D. Everington, C.C. McDonald, J.A. Dewar, R.A. Hawkins, R.J. Prescott, W.D. George, Randomised comparison of 5 years of adjuvant tamoxifen with continuous therapy for operable breast cancer: the Scottish Cancer Trials Breast Group, Br. J. Cancer 74 (1996) 297–299. [8] M.M. Gottardis, S.Y. Jiang, M.-H. Jeng, V.C. Jordan, Inhibition of tamoxifen-stimulated growth of an MCF-7 tumor variant in athymic mice by novel steroidal antiestrogens, Cancer Res. 49 (1989) 4090– 4093. [9] R.J. Santen, Long term tamoxifen therapy: can an antagonist become an agonist? (Editorial), J. Clin. Endocrinol. Metab. 81 (1997) 2027– 2029. [10] K.B. Horwitz, When tamoxifen turns bad (Editorial), Endocrinology 136 (1995) 821–823. [11] S. Masamura, S.J. Santner, D.F. Heitjan, R.J. Santen, Estrogen deprivation causes estradiol hypersensitivity in human cancer cells, J. Clin. Endocrinol. Metab. 80 (1995) 2918–2925. [12] S.S. Simons, H. Oshima, D. Szapary, Higher levels of control: modulation of steroid hormone-regulated gene transcription, Mol. Endocrinol. 6 (1992) 995–1002.
124
R. Santen et al. / Journal of Steroid Biochemistry & Molecular Biology 79 (2001) 115–125
[13] D.P. McDonnell, E. Vegeto, B.W. O’Malley, Identification of a negative regulatory function for steroid receptors, Proc. Natl. Acad. Sci. U.S.A. 89 (1992) 10563–10567. [14] T. Chun, D. Gregg, D.K. Sarkar, J. Gorski, Differential regulation by estrogens of growth and prolactin synthesis in pituitary cells suggests that only a small pool of estrogen receptors is required for growth, Proc. Natl. Acad. Sci. U.S.A. 95 (1998) 2325–2330. [15] F. Labrie, R. Veilleux, A wide range of sensitivities to androgens develops in cloned Shionogi mouse mammary tumor cells, Prostate 8 (1986) 293–300. [16] F. Labrie, R. Veilleux, A. Fournier, Low androgen levels induce the development of androgen-hypersensitive cell clones in Shionogi mouse mammary carcinoma cells in culture, J. Natl. Cancer Inst. 80 (1988) 1138–1147. [17] H.E. Kulin, M.M. Grumbach, S.L. Kaplan, Changing sensitivity of the pubertal hypothalamic feedback mechanism in man, Science 166 (1969) 1012–1013. [18] J. Maruca, H.E. Kulin, S.J. Santner, Perturbations of negative feedback sensitivity in agonadal patients undergoing estrogen replacement therapy, J. Clin. Endocrinol. Metab. 56 (1983) 3–59. [19] S.J. Legan, F.J. Karsch, D.L. Foster, The endocrine basis of seasonal reproductive function in the ewe: a marked change in response to the negative feedback action of estradiol on luteinizing hormone secretion, Endocrinology 101 (1977) 818–824. [20] R.J. Santen, J.N. Friend, D. Trojanowski, B. Davis, E. Samojlik, C.W. Bardin, Prolonged negative feedback suppression after estradiol administration: proposed mechanism of eugonadal secondary amenorrhea, J. Clin. Endocrinol. Metab. 47 (1978) 1220–1229. [21] M.M. Montano, W.L. Kraus, B.S. Katzenellenbogen, Identification of a novel transferable cis element in the promoter of an estrogen-responsive gene that modulates sensitivity to hormone and antihormone, Mol. Endocrinol. 11 (1997) 330–341. [22] D.A. Jackson, C.D. Collier, H. Oshima, S.S. Simons, Modulation of TAT gene induction by glucocorticoids involves a neutralizing sequence, J. Steroid Biochem. Mol. Biol. 66 (1998) 79–91. [23] M. Gershanovich, H.A. Chaudri, D. Campos, H. Lurie, A. Bonaventura, M. Jeffrey, F. Buzzi, I. Bodrogi, H. Ludwig, P. Reichardt, N. O’Higgins, G. Romieu, P. Friederich, M. Lassus, Letrozole, a new oral aromatase inhibitor: randomised trial comparing 2.5 mg daily, 0.5 mg daily and aminoglutethimide in postmenopausal women with advanced breast cancer, Ann. Oncol. 9 (1998) 639–645. [24] J. Bergh, J. Bonneterre, H.J. Illiger, et al., Vorozole (RivizorTM ) versus aminoglutethimide (AG) in the treatment of postmenopausal breast cancer relapsing after tamoxifen, Proc. Am. Soc. Clin. Oncol. 33 (1997) 155a (Abstract No. 543). [25] P. Dombernowsky, I. Smith, G. Falkson, R. Leonard, L. Panasci, J. Bellmunt, W. Bezwoda, G. Gardin, A. Gudgeon, M. Morgan, A. Fornasiero, W. Hoffman, J. Michel, T. Hatschek, T. Tjabbes, H.A. Chaudri, U. Hornberger, P.F. Trunet, Letrozole, a new oral aromatase inhibitor for advanced breast cancer: double-blind randomized trial showing a dose effect and improved efficacy and tolerability compared with megestrol acetate, J. Clin. Oncol. 16 (1998) 453–461. [26] R.J. Santen, H.A. Harvey, Use of aromatase inhibitors in breast carcinoma, Endocr. Relat. Cancer 6 (1999) 75–92. [27] W.V. Welshons, A.C. Jordan, Adaptation of estrogen-dependent MCF-7 cells to low estrogen (phenol red-free) culture, Eur. J. Cancer Clin. Oncol. 23 (1987) 1935–1939. [28] N. Brunner, V. Boulay, A. Fojo, C.E. Freter, M.E. Lippman, R. Clarke, Acquisition of hormone-independent growth in MCF-7 cells is accompanied by increased expression of estrogen-regulated genes but without detectable DNA amplifications, Cancer Res. 53 (1993) 283–290. [29] R. Clarke, N. Brunner, B.S. Katzenellenbogen, E.W. Thompson, J.F. Norman, C. Koppi, S. Paik, M.E. Lippman, R.B. Dickson, Progression of human breast cancer cells from hormone-dependent to hormone-independent growth both in vitro and in vivo, Proc. Natl. Acad. Sci. U.S.A. 96 (1989) 3649–3653.
[30] B.S. Katzenellenbogen, K.L. Kendra, M.J. Norman, Y. Berthois, Proliferation, hormonal responsiveness, and estrogen receptor content of MCF-7 human breast cancer cell lines grown in the short-term and long-term absence of estrogens, Cancer Res. 47 (1987) 4355–4360. [31] M.E. Herman, B.S. Katzenellenbogen, Alterations in transforming growth factor -␣ and - production and cell responsiveness during the progression of MCF-7 human breast cancer cells to estrogen-autonomous growth, Cancer Res. 54 (1994) 5867–5874. [32] H. Cho, P.A. Ng, B.S. Katzenellenbogen, Differential regulation of gene expression by estrogen in estrogen-growth independent and -dependent MCF-7 human breast cancer cell sublines, Mol. Endocrinol. 5 (1991) 1323–1330. [33] C.S. Murphy, J.J. Pink, V.C. Jordan, Characterization of a receptor-negative, hormone-unresponsive clone derived from a T-47-D human breast cancer cell line kept under estrogen-free conditions, Cancer Res. 50 (1990) 7285–7292. [34] R.J. Daly, P.D. Darbre, Cellular and molecular events in loss of estrogen sensitivity in ZR-75-1 and T-47-D human breast cancer cells, Cancer Res. 50 (1990) 5868–5875. [35] V.C. Jordan, The control of hormone dependent breast cancer—are we talking about estrogen alone, Eur. J. Cancer Clin. Oncol. 24 (1988) 1245–1248. [36] R.I. Nicolson, J.M.W. Gee, D.L. Manning, A.E. Wakeling, M.M. Montano, B.S. Katzenellenbogen, Responses to pure antiestrogens (ICI 164384, ICI 182780) in estrogen-sensitive and -resistant experimental and clinical breast cancer, Ann. New York Acad. Sci. 761 (1995) 148–163. [37] W. Yue, J.-P. Wang, L. Freeman, R.J. Santen, Activation of MAP kinase in estrogen dependent human breast cancer cells during regrowth in the absence of exogenous estrogen, Endo. Abstract P1 (1998) 126. [38] W.-S. Shim, M. Conaway, S. Masamura, W. Yue, J.-P. Wang, R.J. Santen, Estradiol hypersensitivity and MAP kinase expression in long term estrogen deprived human breast cancer cells in vivo, Endocrinology 141 (2000) 396–405. [39] L.A. Martin, C.M.W. Chan, I. Farmer, C. Marshall, M. Dowsett, Molecular changes occurring during lone-term oestrogen deprivation, Aromatase 2000 and the 3rd generation (Abstract), 3–7 November 2000, Port Doughlas, Australia. [40] A. Kirschenbaum, M. Ren, A.C. Levine, Enhanced androgen sensitivity in serum-free medium of a subline of the LnCAP human prostate cancer cell line, Steroids 58 (1993) 439–444. [41] J. Kokontis, K. Takakura, N. Hay, S. Liao, Increased androgen receptor activity and altered c-myc expression in prostate cancer cells after long-term androgen deprivation, Cancer Res. 54 (1994) 1566–1573. [42] J. Kokontis, N. Hay, S. Liao, Progression of LnCAP prostate tumor cells during androgen deprivation: hormone-independent growth, repression of proliferation by androgen, and role for p27kip1 in androgen-induced cell cycle arrest, Mol. Endocrinol. 12 (1998) 941– 953. [43] R.J. Santen, Endocrine treatment of prostate cancer, J. Clin. Endocrinol. Metab. 75 (1992) 685–689. [44] A.M. Soto, H. Justicia, J.W. Wray, C. Sonnenschein, P-nonyl-phenol: an estrogenic xenobiotic released from modified polystyrene, Environ. Health Perspect. 92 (1991) 167–173. [45] A.M. Soto, C. Sonnenschein, K.L. Chung, M.F. Fernandez, N. Olea, F.O. Serrano, The E-SCREEN assay as a tool to identify estrogens: an update on estrogenic environmental pollutants, Environ. Health Perspect. 103 (Suppl. 7) (1995) 113–122. [46] A.V. Krishnan, P. Stathis, S.F. Permuth, L. Tokes, D. Feldman, bis-Phenol-A: an estrogenic substance is released from polycarbonate flasks during autoclaving, Endocrinology 132 (1993) 2279–2286. [47] L.F. Ren, M.A. Marquardt, J.J. Lech, Estrogenic effects of nonylphenol on PS2, ER and MUC1 gene expression in human breast cancer cells—MCF-7, Chem. Biol. Interact. 104 (1997) 55–64.
R. Santen et al. / Journal of Steroid Biochemistry & Molecular Biology 79 (2001) 115–125 [48] K. Connor, K. Ramamoorthy, M. Moore, M. Mustain, I. Chen, S. Safe, T. Zacharewski, B. Gillesby, A. Joyeux, P. Balaguer, Hydroxylated polychlorinated biphenyls (PCBS) as estrogens and antiestrogens-structure–activity relationships, Toxicol. Appl. Pharmacol. 145 (1997) 111–123. [49] R.B. Dickson, M.E. Lippman, Growth factors in breast cancer, Endocr. Rev. 16 (1995) 559–589. [50] C.A. Lange, J.K. Richer, T.J. Shen, K.B. Horwitz, Convergence of progesterone and epidermal growth factor signaling in breast cancer-potentiation of mitogen-activated protein kinase pathways, JBC 273 (1998) 31308–31316. [51] R.M. Elledge, W.L. McGuire, C.K. Osborne, Prognostic factors in breast cancer, Semin. Oncol. 19 (1992) 244–253. [52] S. Kato, H. Endoh, Y. Masuhiro, T. Kitamoto, S. Uchiyama, H. Sasaki, S. Masushige, Y. Gotoh, E. Nishida, H. Kawashima, Activation of the estrogen receptor through phosphorylation by mitogen-activated protein kinase, Science 270 (1995) 1491–1494. [53] G. Bunone, P.A. Briand, R.J. Miksicek, D. Picard, Activation of the unliganded estrogen receptor by EGF involves the MAP kinase pathway and direct phosphorylation, EMBO J. 15 (1996) 2174–2183. [54] M.-H. Jeng, M.A. Shupnik, T.P. Bender, E.H. Westin, D. Bandyopadhyay, R. Kumar, S. Masamura, R.J. Santen, Estrogen receptor expression and function in long-term estrogen-deprived human breast cancer cells, Endocrinology 139 (1998) 4164–4174.
125
[55] M.-H. Jeng, W. Yue, A. Eischeid, J.-P. Wang, R.J. Santen, Role of MAP kinase in the enhanced cell proliferation of long term estrogen deprived human breast cancer cells, Breast Cancer Res. Treatment 62 (2000) 167–175. [56] G. Leone, J. DeGregori, R. Sears, L. Jakoi, J.R. Nevins, Myc and Ras collaborate in inducing accumulation of active cycling E/Cdk2 and E2F, Nature 387 (1997) 422–425. [57] G. Leone, J. DeGregori, R. Sears, L. Jakoi, J.R. Nevins, Myc and Ras collaborate in inducing accumulation of active cyclin E/Cdk2 and E2F, Nature 387 (1997) 932. [58] P.H. Watson, R.T. Pon, R.P.C. Shiu, Inhibition of c-myc expression by phosphorothioate antisense olingonucletide identifies a critical role for c-myc in the growth of human breast cancer, Cancer Res. 51 (1991) 3996–4000. [59] W. Yue, R.J. Santen, S. Manna, R. Kumar, Cell cycle events associated with the ability of hormone dependent breast cancer cells to grow in the absence of exogenous estrogen, in: Proceedings of the 81st Annual Meeting of the Endocrine Society, San Diego, CA, 1999. [60] R.X. Song, R.A. McPherson, W. Yue, J.-P. Wang, R.J. Santen, SHC phosphorylation and interaction with ER alpha are required for estrogen-induced MAP kinase phosphorylation in MCF-7 human breast cancer cells, in: Proceedings of the 82nd Annual Meeting of the Endocrine Society, Abstract 1401, Toronto, Canada, 2000, p. 338.
Adaptive hypersensitivity to estradiol: potential mechanism for secondary hormonal responses in breast cancer patients夽 Richard Santen a,∗ , Meei-Huey Jeng b , Ji-Ping Wang c , Robert Song c , Shigeru Masamura d , Robert McPherson c , Steven Santner e , Wei Yue c , Woo-Shin Shim f a
Division of Endocrinology, University of Virginia Health System, P.O. Box 800379, Jefferson Park Avenue, Charlottesville, VA 22908, USA b Department of Medicine, Section of Hematology/Oncology, Indiana University, Indianapolis, IN 46202-5121, USA c Division of Hematology/Oncology, University of Virginia Health System, P.O. Box 800747, Charlottesville, VA 22908, USA d Department of General Surgery, School of Medicine, Keio University, Shinanomachi 35, Shinjuku, Tokyo 160-8582, Japan e Karmanos Cancer Institute, 110 East Warren, Detroit, MI 48201, USA f Department of General Surgery, Gacheon Medical College Ghil Hospital, Incheon 400-170, South Korea
Abstract Women with hormone dependent breast cancer initially respond to hormone deprivation therapy with tamoxifen or oophorectomy for 12–18 months but later relapse. Upon secondary therapy with aromatase inhibitors, patients often experience further tumor regression. The mechanisms responsible for secondary responses are unknown. We postulated that hormone deprivation induces hypersensitivity to estradiol. Evidence of this phenomenon was provided in a model system involving MCF-7 cells grown in vitro and in xenografts. To determine if the ER transcriptional process is involved in hypersensitivity, we examined the effect of estradiol on ER reporter activity, PgR, PS2, and c-myc as markers and found no alterations in hypersensitive cells. Next, we examined whether MAP kinase may be upregulated in the hypersensitive cells as a reflection of increased growth factor secretion or action. Basal MAP kinase activity was increased both in vitro and in vivo in hypersensitive cells. Proof of principle studies indicated that an increase in MAP kinase activity induced by TGF␣ administration caused a two- to three-fold shift to the left in estradiol dose response curves in wild type cells. Blockade of MAP kinase with PD98059 returned the shifted curve back to baseline. These data suggested that MAP kinase overexpression could induce hypersensitivity. To determine why MAP kinase was increased, we excluded constitutive receptor activity and growth factor secretion by the demonstration that the pure anti-estrogen, ICI 182780, could inhibit MAP kinase activation. We also excluded hypersensitivity to estradiol induced growth factor secretion, and thus MAP kinase activation, since estradiol stimulated MAP kinase at 24, 48, and 72 h at the same concentrations in hypersensitive as in wild type cells. Surprisingly, a series of experiments suggested that MAP kinase increased in hypersensitive cells as a result of estrogen activation via a non-genomic pathway. We examined the classical signal pathway in which SHC is phosphorylated and binds to SOS and GRB-2 to activate Ras, Raf, and MAP kinase. With 5–20 min of exposure, estradiol caused binding of SHC to the estrogen receptor, phosphorylation of SHC, binding of GRB-2 to SOS, and activation of MAP kinase. All of these affects could be blocked by ICI 182780. Taken together, these observations suggest that the cell membrane ER pathway may be responsible for upregulation of MAP kinase and hypersensitivity in cells adapted to estradiol deprivation. © 2002 Elsevier Science Ltd. All rights reserved. Keywords: Aromatase; Breast cancer; Estrogen; Hypersensitivity; MAP kinase; Non-genomic estrogen effects; Proliferation
1. Introduction Breast cancer will be diagnosed in 180,000 patients in the United States in 2000 and nearly 44,000 will die of this disease [1]. One-third of women with locally advanced or metastatic disease experience tumor regression in response to hormonal therapy [2]. However, adaptive mechanisms intervene and these tumors begin to re-grow within a period 夽 Proceedings of the Symposium: ‘Aromatase 2000 and the Third Generation’ (Port Douglas, Australia, 3–7 November 2000). ∗ Corresponding author. Tel.: +1-804-924-2961; fax: +1-804-924-9616. E-mail address: [email protected] (R. Santen).
of 12–18 months on average. Secondary hormonal therapies frequently induce additional responses of similar duration [2]. The mechanisms whereby secondary responses occur have never been satisfactorily explained. Clinical observations with inhibitors of aromatase, the enzyme catalyzing the rate limiting step in estradiol biosynthesis, suggested to us that hypersensitivity to estradiol might explain these secondary responses. In post-menopausal women with breast cancer, inhibition of aromatase suppresses estrogen production [2]. In initial clinical studies with the first generation inhibitor, aminoglutethimide, we were surprised to observe secondary tumor regressions in women who had previously responded to surgical oophorectomy but later relapsed [3].
0960-0760/02/$ – see front matter © 2002 Elsevier Science Ltd. All rights reserved. PII: S 0 9 6 0 - 0 7 6 0 ( 0 1 ) 0 0 1 5 1 - 0
116
R. Santen et al. / Journal of Steroid Biochemistry & Molecular Biology 79 (2001) 115–125
Following initial treatment with oophorectomy, estrogen levels fell from premenopausal levels of 50–600 pg/ml to approximately 10 pg/ml and in response, tumors regressed for a period of 12–18 months [4]. Upon relapse after oophorectomy, estradiol levels remained at the 10 pg/ml level. Subsequent treatment with aromatase inhibitors lowered estradiol further to 1–5 pg/ml and caused secondary tumor regressions. These observations first led us to consider the possibility that tumors might adapt by becoming hypersensitive to estradiol upon hormonal deprivation. Adaptation to hormonal therapy may also occur in women receiving adjuvant hormonal therapy prior to the appearance of metastatic or locally advanced disease [5]. In such women, tamoxifen reduces the rate of recurrence and enhances survival if administered for 1 to 5 years. When given for 10 years as opposed to 5, no additional benefit is observed and an increased number of recurrences and deaths have been reported [6,7]. Observations in xenograft models of breast cancer suggest that these neoplasms may adapt to tamoxifen such that this anti-estrogen becomes an estrogen agonist after a period of 6 months to 1 year [8]. We and others have suggested that hypersensitivity to the estrogenic effects of tamoxifen may occur in women receiving adjuvant tamoxifen. This mechanism might explain the adverse effects of 10 versus 5 years of adjuvant tamoxifen therapy [9,10]. Taken together, these observations emphasize the fact that tumors can adapt to their environment and develop means to re-grow under conditions that previously limited growth. An understanding of these adaptive processes would provide targets for development of specific therapies to inhibit re-growth and to enhance the efficacy of hormonal therapies.
2. Estradiol hypersensitivity hypothesis Our hypothesis suggests that breast tumors adapt to estrogen deprivation by developing increased sensitivity to estradiol [11]. Re-growth then occurs because of the ability of tumors to utilize very small amounts of estrogen to induce a proliferative response. Differing levels of sensitivity to hormones is a widely recognized phenomenon. The range of sensitivity to hormones varies substantially among different tissues and under a wide spectrum of physiologic conditions [12,13]. Several examples of hypersensitivity to estradiol have been described. A genetically engineered yeast system described by McDonnell et al. responds to levels as low as 10−14 M estradiol [13]. A prolactin producing pituitary tumor described by Chun et al. exhibits enhanced proliferation when exposed to 10−14 M levels of both DES and estradiol [14]. With respect to androgens, clones of Shionogi breast cancer cells exhibit marked ranges in sensitivity to the proliferative responses produced by testosterone [15,16]. The range in sensitivity of various clones varies by a factor of 1000 and some clones respond to levels of testosterone of 10−12 M.
The set point for sensitivity to various hormones can be dynamically determined and altered by the level of ambient hormonal exposure. Prepubertal girls are highly sensitive to the suppressive effects of estradiol on gonadotropin secretion [17]. Upon exposure to increasing amounts of estrogen, their level of sensitivity of the feedback mechanism diminishes [18]. Seasonally breeding ewes exhibit exquisite sensitivity to estradiol during the non-breeding season with much reduced sensitivity during the breeding season [19]. Secondary amenorrhea in women is associated with changes in the set point of sensitivity to estradiol [20]. Several experiments demonstrate shifts in dose response curves to the right or left as a reflection of change in sensitivity [12]. Sensitivities to both glucocorticoid and progesterone are enhanced when specific DNA binding proteins interact with elements as far as 1–3 kb upstream of the promoter regions for their respective receptors [21,22]. Multiple other mechanisms are possible, including events occurring well downstream of the estrogen receptor. The ability of breast cancer cells to adapt to lower levels of estradiol has clinical relevance regarding design of hormonal therapies for patients. The adaptive hypersensitivity concept suggests that the greater the suppression of estradiol, the more effective would be the anti-tumor efficacy of aromatase inhibitors. This raised the possibility that the highly potent third generation aromatase inhibitors would be more effective than the first generation inhibitor, aminoglutethimide or tamoxifen. Large clinical trials confirmed this supposition [23–26]. An additional implication of the hypersensitivity hypothesis relates to choice of inhibitor dosage. It was reasoned that breast cancer cells might respond to amounts of estradiol well below the level of detection with sensitive plasma radioimmunoassays. Notably, in the cell culture model, 10−14 M estradiol, a concentration well below assay sensitivity, stimulates cell proliferation [11]. Based upon this reasoning, it would be logical to compare a dose of aromatase inhibitor which suppresses estradiol to undetectable levels with a much higher dose of the same inhibitor. Clinical trials to test this concept involved 945 breast cancer patients and compared responses in women receiving 0.5 mg of letrozole, a dose which suppresses plasma estradiol to undetectable levels, with a five-fold higher dose, 2.5 mg daily. In both studies, the 2.5 mg dose resulted either in prolongation of patient survival (P < 0.01) or in a higher percent of objective response rate (12.5% versus 23.5%, P < 0.01). These observations supported the clinical relevance of the concept of adaptive hypersensitivity [24–26].
3. Model systems of adaptive hypersensitivity Observations in models of breast and prostate carcinoma support the concept of adaptive hypersensitivity [27–36]. We utilized MCF-7 breast cancer cells grown in culture under conditions of estradiol deprivation as a model [11]. We call these LTED cells to reflect their Long-Term
R. Santen et al. / Journal of Steroid Biochemistry & Molecular Biology 79 (2001) 115–125
117
Fig. 1. Increased basal growth rate and inhibition by ICI 182780 in LTED cells as compared with wild type MCF-7 cells. (A) Wild type cells were deprived of estrogen for 3 days and plated onto six-well plates at the same time as LTED cells, at a density of 6 × 10−5 cells per well. The next day (day 1), cells were lysed and cell nuclei were then counted every day for the next 10 days using a Coulter counter. (B) LTED cells were plated onto six-well plates at a density of 1 × 10−5 cells per well. ICI 182780 at various concentrations was added the next day for 6 days. Media were changed every other day and cells were lysed and cell nuclei were then counted using the Coulter counter. Medium was changed every other day. Data are presented as mean ± S.E. from duplicate wells and S.E. when error bars were sufficiently large to visualize. Reprinted with the permission of publisher and authors from [54].
Estradiol Deprivation (we will use this term throughout this manuscript). These cells initially stop growing under conditions of reduced estrogen achieved by charcoal stripping of serum used in the culture media. Later, they grow maximally under these conditions and cannot be further stimulated by exogenous estrogen but are inhibited with “pure” anti-estrogens (Fig. 1A and B). To explain these observations, we hypothesized that LTED cells are hypersensitive to the residual estradiol present in culture media and consequently, can grow in charcoal stripped serum without addition of exogenous estradiol [11]. Three experimental approaches supported this possibility. Firstly, when special methods were used to remove all residual estrogen from the media, LTED cells grew minimally without exogenous estradiol. Under these conditions, LTED cells were stimulated to grow in response to levels of estradiol of 10−14 M whereas wild type cells required 10−11 M (Fig. 2). Secondly, when the residual estrogen was antagonized by adding a pure anti-estrogen, LTED cells grew only minimally. Addition of estradiol under these conditions caused LTED cells to re-grow at two log lower concentrations of estradiol than similarly treated wild type cells (Fig. 3) [37]. Thirdly, LTED cell xenografts in vivo respond to lower doses of estradiol for growth than wild type cells [38]. The in vivo studies utilized xenografts of wild type and LTED cells implanted into both sides of the backs of castrated nude mice. A detailed description of the methods and results has been previously reported [38]. Briefly, initial studies validated the ability to “clamp” estradiol levels at 1.25, 2.5, 5, 10, and 20 pg/ml using silastic implants filled with mixtures of cholesterol and estradiol. Since such low levels cannot be reliably measured in castrate nude mouse plasma, we utilized uterine weight as a bioassay to demonstrate linearity of plasma estradiol levels. Prior published studies which induced much higher levels of estradiol
(i.e. 20–800 pg/ml) had demonstrated linearity between the cholesterol/estradiol ratio in the silastic implants and plasma estradiol as measured by RIA [11]. We next implanted two LTED and two wild type xenografts onto the backs of each castrate female mouse using our newly developed
Fig. 2. E2 dose–response curves in LTE deprived cells using serum-free medium and wild type MCF-7 cells grown in dextran coated charcoal (DCC) striped serum containing medium. LTE deprived cells (50,000 cells per flask) were seeded in medium containing DCC stripped serum into five replicate T-25 flasks for each group. After 1 day of acclimatization, they were transferred to serum-free medium for 1 week. The media were then changed to experimental media (serum-free medium containing ethanol vehicle 0.1%); E2 at 10−15 to 10−6 M, or ICI 164384 at 10−6 M. Cells were re-fed with experimental serum-free media 3 days later and counted after another 3 days. Wild type cells were grown in media containing DCC treated serum. Data are illustrated as a percentage of the maximum response (±SEM) to normalize data between these two experiments. (䊉) Wild type MCF-7 cells; (䊊) LTE deprived cells. Reprinted with the permission of publisher and authors from [11].
118
R. Santen et al. / Journal of Steroid Biochemistry & Molecular Biology 79 (2001) 115–125
Fig. 3. Dose response effects of estradiol on growth of LTED and wild type MCF-7 cells grown in the presence of 10−9 M ICI 182780.
heteroimplant technique. This method eliminates the confounding effect of the individual animal on tumor growth since both wild type and LTED xenografts are present in the same animal. We then allowed tumors to become established over a 1 month period. This is a key step which eliminates the confounding effect of estradiol on establishment of tumor rather than on proliferation. After 1 month, animals received silastic implants to clamp estradiol at levels of 1.25, 2.5, 5, 10, and 20 pg/ml and tumor volumes were measured by calipers weekly for 8 weeks.
As shown in Fig. 4, these in vivo findings paralleled those routinely observed in vitro. Specifically, LTED tumors grew in response to much lower concentrations of estradiol than wild type as a reflection of hypersensitivity. The LTED cell xenografts grew to a greater extent than wild type xenografts in response to estradiol levels of 1.25 and 2.5 pg/ml (P < 0.0001 for both). When E2 levels increased further, a peak growth rate was reached and then growth diminished, producing a typical bell shaped dose response curve (see Fig. 4). Specifically, both LTED and wild type xenografts grew equally at the 5 pg/ml dosage but at the 10 and 20 pg/ml doses, the wild type tumors grew to a greater extent than LTED. These results indicate that LTED cells exhibit increased sensitivity to both the initial stimulatory and later inhibitory levels of estradiol. The process of development of hypersensitivity could represent either cell selection or adaptation. Supporting the concept of adaptation was the plasticity in reversing and re-inducing the hypersensitive phenotype. When re-exposed to estradiol containing media in vitro for periods of 1 month or longer, the LTED cells reverted back to the wild type phenotype [11]. Deprivation of estradiol again induced the hypersensitive phenotype. The process of reversion and induction could be reproducibly repeated over a period of 2 years, demonstrating the plasticity of this system and supporting the concept of adaptation rather than cell selection [11].
Fig. 4. (A) Growth curves in wild type and LTED tumors in oophorectomized nude mice receiving only cholesterol containing SILASTIC brand implants (vehicle control) and implants maintaining plasma E2 levels at 1.25, and 2.5 pg/ml. The statistical significance of the differences between wild type and LTED tumors is indicated on each panel. The accompanying bars ± SEM represent mean area under the curve for each group and are shown to illustrate variance among groups. The black bars are representative of volumes of LTED tumors and the cross-hatched bars of wild type tumors. The statistical significance indicated represents paired comparisons between integrated tumor volumes and not between mean areas under the various curves. Integrated tumor volumes were significantly higher in the LTED than wild type tumors in response to 1.25 and 2.5 pg/ml but not in oophorectomized animals. (B) Growth curves in wild type and LTED tumors with plasma E2 clamped at 5, 10 and 20 pg/ml. Reprinted with the permission of publisher and authors from [38].
R. Santen et al. / Journal of Steroid Biochemistry & Molecular Biology 79 (2001) 115–125
The process of dynamic adaptation to hormones is not limited to our breast tumor model. Martin et al. also demonstrated development of hypersensitivity in MCF-7 cells upon deprivation of estradiol [39]. Hormone dependent prostate cancer cells also adapt to conditions of low androgen in culture by increasing their sensitivity to androgens [40–42]. Kirschenbaum et al. [40] demonstrated growth stimulation with 10−14 M DHT in LnCAP cells which were conditioned by long-term exposure to low androgen concentrations. Wild type cells required 10−10 M. Kokontis et al. [41,42] also demonstrated adaptive hypersensitivity in cloned LnCAP prostate cancer cells. Upon androgen deprivation, these cells developed one log enhanced sensitivity to the synthetic androgen R-1881 and then reverted to a two to three log less sensitive state upon androgen re-exposure. The observations in prostate cancer cell lines suggest that the process of adaptive hypersensitivity may be common to hormone dependent cancer and not limited to one breast cancer cell line. This phenomenon could explain why men with prostate cancer respond initially to castration, later relapse, and then experience secondary tumor regression upon blockade of androgen action with potent anti-androgens [43].
4. Mechanisms of hypersensitivity Our laboratory has been focusing upon potential mechanisms to explain hypersensitivity to estradiol in the LTED model system. To facilitate these experiments, we have recently modified experimental conditions in our model to alleviate problems from residual estrogen contamination of culture media. LTED cells grow in standard charcoal stripped serum as a result of residual estrogen and growth is inhibited with anti-estrogens (Figs. 1 and 3) [37]. Sources of the residual estrogenic activity in culture media are unknown but recent reports suggest that components of plastic and many other industrial substances act as estrogen agonists. Stimulatory effects from these contaminants are not surprising since the LTED cells are sensitive to concentrations of estradiol as low as 10−15 M (Fig. 2) [11,44–48]. Our method of enhanced stripping of sera with charcoal, as originally reported [11], or use of serum-free media results in a variable rate of growth of LTED cells in the absence of exogenous estrogen. Accordingly, we have developed a practical strategy to block residual estrogen activity with pure anti-estrogens. This method involves antagonism of residual estrogen by addition of identical amounts of anti-estrogen to wild type and LTED cells. We can then precisely determine dose responses to exogenous estradiol in wild type and LTED cells in a reproducible fashion (Fig. 3). Under these conditions, the LTED cells respond to two to three log lower concentrations of estrogen than do wild type cells and are not effected by variable levels of residual estrogen in media. In other experiments, we do not antagonize residual estrogen with anti-estrogens so that we can probe the factors stimulating growth of LTED cells under these conditions.
119
As a conceptual means to facilitate analysis, we arbitrarily divide estradiol mediated events into genomic and non-genomic events. Both pathways can interact with classical growth factor pathways with “cross talk” at an ER transcriptional level or at the level of the cell cycle. A variety of studies indicate that growth factor secretion or action can be stimulated by estradiol or can be regulated independently of estradiol [49–53]. The mechanism for mediation of hypersensitivity could involve either genomic or non-genomic events acting at upstream or downstream sites of action.
5. Genomic effects of estradiol We initially focused upon events involving the estrogen receptor and its regulation of transcription as a possible inducing hypersensitivity. We demonstrated that long-term deprivation of estradiol causes an upregulation of estrogen receptor content from 3- to 20-fold (Fig. 5). These results included assessment of the number of binding sites by radioligand analysis [11] and level of ER alpha message as well as quantitation of the amount of ER by western analysis using antibodies recognizing various epitopes of the ER (Fig. 5) [54]. In addition, we found that basal rates of transcription of the ER were increased and could be inhibited by the pure anti-estrogen, ICI 182780 [54]. Even though upstream processes were altered by estradiol deprivation, we considered it important to determine if these receptor changes were mechanistically liked to the process of hypersensitivity. To do this, we measured several parameters reflecting estrogen receptor mediated transcription. Specifically, we administered estradiol in doses from 10−14 to 10−6 M in LTED and wild type cells and measured ER reporter gene activity as well as c-myc, progesterone receptor, PS2, and MAP kinase. We observed similar dose response stimulation of several ER reporter constructs, of progesterone receptor content, of c-myc message levels, of PS2 and of activated MAP kinase content (data not shown). Taken together, these results suggested that hypersensitivity is not mediated primarily at the level of ER transcription through classical genomic effects of estradiol.
6. Growth factor pathways We then questioned whether growth factor pathways might be turned on in LTED cells and whether this might be the mechanism for enhanced sensitivity to estradiol. Initial studies examined the effect of inhibitors of several steps in the growth factor pathway on cell proliferation in LTED cells. Growth factors bind to cell membrane receptors, activate receptor tyrosine kinases, cause isoprenylation of Ras, and activate MAP kinase by phosphorylation. We found that inhibitors of each of these steps produced a dose dependent inhibition of tritated thymidine uptake in LTED cells as evidence of a reduction in cell proliferation [55] (Fig. 6).
120
R. Santen et al. / Journal of Steroid Biochemistry & Molecular Biology 79 (2001) 115–125
transcription. Under basal conditions, ER reporter activity was higher in LTED than in wild type cells (Fig. 8). However, this enhanced basal activity could not be inhibited by the MAP kinase inhibitor, PD98059 (Fig. 8). On the other hand, estradiol stimulated reporter activity could be partially but not completely inhibited by PD98059 (Fig. 9). These observations suggested that elevated MAP kinase can partially influence estradiol stimulated but not basal ER transcription. Effects of MAP kinase at this step do not explain hypersensitivity since the dose response effects of estradiol on ERE reporter activity are not right shifted by MAP kinase inhibition.
7. Proof of principle experiment
Fig. 5. ER protein expression in wild type and LTED cells. Wild type MCF-7 cells were deprived of estrogen and treated with or without 10−10 M E2 for 2 days. Cell cytosols were collected and 50 g protein (B and C) or 100 g protein (D) were loaded per lane and separated by electrophoresis on 10% polyacrylamide gels containing 1% SDS and transferred onto nitrocellulose membranes. Western blot analysis was performed as described in the original publication. ER antibody C314 was used in panel B and ER antibodies H226, D547, H222 and D75 were used in panels C and D. An epitope map of these antibodies is shown in panel A. Regions A–F and exons 1–8 are marked. Numbers marked above each antibody were the estimated amino acid ranges for the epitope locations. Ponseau S staining of the nitrocellulose membranes after transfer of proteins showed equal loading of proteins for all lanes. Reprinted with the permission of publisher and authors from [54].
These observations provided evidence that growth factor pathways played a major role in mediating cell proliferation in LTED cells. As further evidence of this possibility, we measured the levels of MAP kinase in LTED cells in vitro in comparison with wild type cells and found a marked increase [55] (Fig. 7). In vivo studies using immunohistochemistry to detect activated MAP kinase also demonstrated increased activation of MAP kinase in LTED xenografts. Independent of estradiol dosage, the percentage of MAP kinase positive cells in LTED cell xenografts ranged from 18 to 22% [38]. In contrast, wild type xenografts from castrated animals contained only 3% of MAP kinase activated cells when given vehicle as a control. The wild type xenografts exhibited a maximum dose dependent increase to 6% when exposed to increasing concentrations of estradiol [38]. We next questioned whether elevated MAP kinase might act to phosphorylate the ER and enhance its level of
Our observations linked hypersensitivity to MAP kinase overexpression but did not provide evidence of causality. Accordingly, we designed a proof of principle experiment to increase the level of MAP kinase and examine the effect of this maneuver on sensitivity to estradiol. To do this, we administered TGF␣ and determined that the optimal dose necessary to stimulate MAP kinase was 10 ng/ml and that the MAP kinase inhibitor PD98059 blocked this response (data not shown). We then demonstrated that wild type MCF-7 cells exposed to 10 ng/ml TGF␣ grew in response to two log lower concentrations of estradiol than did the same cells not given TGF␣. The induction of hypersensitivity reflected an increase in MAP kinase rather than another effect of TGF␣. Co-administration of the MAP kinase inhibitor, PD98059, completely abrogated the effects on TGF␣ on hypersensitivity [37]. These data suggested that MAP kinase might be involved in estradiol hypersensitivity observed in the LTED cells. To provide direct evidence of this, we administered a more potent MAP kinase inhibitor, U 0126, to LTED cells and demonstrated a one log shift toward the right in the estradiol dose response curve in them. This provided direct biologic evidence that elevated MAP kinase increased the level of sensitivity of estradiol in LTED breast cancer cells.
8. Downstream mechanisms The data presented indicate that growth factor pathways, when upregulated, can enhance the level of sensitivity to estradiol. The mechanism to explain this effect was suggested by a series of observations by Leone et al. [56,57]. These investigators provided evidence that Ras mediated events act synergistically with c-myc to stimulate production of the important cell cycle protein, E2F1. If this hypothesis is correct, less c-myc would be necessary to stimulate cell proliferation if Ras pathways were activated. Our studies and those of Watson et al. [58] demonstrated that estradiol stimulates c-myc and that this proto-oncogene is involved in estradiol stimulated proliferation [54]. Both c-myc and the Ras/MAP kinase pathways are activated in LTED cells [54,55]. It would be expected then that E2F1 would be
R. Santen et al. / Journal of Steroid Biochemistry & Molecular Biology 79 (2001) 115–125
121
Fig. 6. (A) Diagrammatic representation of the pathway involving growth factors binding to the plasma membrane, autophosphorylation of the receptor via tyrosine kinase, isoprenylation of Ras, activation of Raf, and activation of MAP kinase which then mediates enhancement of cell proliferation. (B) Effect of various inhibitors on 3 h thymidine uptake in LTED cells. LTED cells were plated onto six-well plates. The next day, cells were washed and fresh phenol red-free IMEM containing various inhibitors at indicated concentrations were added to the monolayers. After 18 h, cells were labeled with 3H-thymidine in the presence or absence of various inhibitors for 4 h. Cell monolayers were then processed for the uptake level of 3H-thymidine uptake. Data are presented as a mean of duplicate samples. The experiments were repeated at least two times. Reprinted from [55].
122
R. Santen et al. / Journal of Steroid Biochemistry & Molecular Biology 79 (2001) 115–125
Fig. 7. (A) Activated and total MAP kinase activity in wild type (WT) and LTED cells. (B) Results from the scanning densitometry of the activated MAP kinase as shown in the blots in A. Reprinted from [55].
increased in LTED as opposed to wild type cells and indeed this proved to be the case [59]. Based upon these data, our current working hypothesis to explain adaptive hypersensitivity is that activation of growth factor/MAP kinase pathways acts synergistically with c-myc at the level of the cell cycle to increase E2F1 levels. This synergy sensitizes cells to the proliferative effects of estradiol and allows growth of LTED cells in the presence of very low levels of estradiol.
9. Non-genomic effects of estradiol Our studies reviewed to date indicated that LTED cells are characterized by upregulation of MAP kinase activity. We initially postulated that this might have resulted from constitutive activation of growth factor receptors or growth factor secretion. We reasoned that either of these two possibilities
Fig. 8. Effect of PD98059 on ERE-CAT activity in both wild type and LTED cells under basal (i.e. estrogen depleted) culture conditions. Cells were plated and transfected with pERE-tk-CAT and pCMVbeta gal and treated with PD98059 at various concentrations for 2 days. Cells were collected and assayed for CAT activity using equal amounts of beta galactosidase. Reprinted from [55].
could be excluded if a pure anti-estrogen suppressed MAP kinase activity. Notably, ICI 182780 reduced the basal levels of MAP kinase in LTED cells by approximately two-third. Consequently, we reasoned that the estrogen receptor must somehow be involved in this process but that genomic effects on transcription were not involved. Another possibility was that estradiol exerted non-genomic effects on MAP
Fig. 9. Effect of PD98059 on E2 induced ERE-CAT activity and activated MAP kinase level in both wild type and LTED cells. (A) Wild type MCF-7 cells were deprived of estrogen and transfected with pERE-tk-CAT and pCMV beta gal. Cells were then treated with various concentrations of E2 in the absence or presence of PD98059 (10 g/ml) for 2 days. Cells were then collected and assayed for CAT activity using equal amounts of beta galactosidase. (B) Wild type MCF-7 cells were deprived of estrogen for 2 days and treated with E2 in the absence or presence of PD98059 (10 g/ml) for 12 h. Total MAP kinase and activated MAP kinase were measured using specific antibodies. Reprinted from [55].
R. Santen et al. / Journal of Steroid Biochemistry & Molecular Biology 79 (2001) 115–125
kinase in LTED cells. To test this, we examined the ability of estradiol to stimulate MAP kinase over a 120 min period. We observed an increase in MAP kinase at 15 min and further increases at 60–120 min. Dose response studies demonstrated peak effects at 10−10 M estradiol [60]. To further dissect the mechanism of non-genomic estrogen effects, we examined the possibility that estradiol could activate a classical pathway involving SHC, phosphorylated SHC, GRB-2 and SOS in LTED cells. Within 10 min of exposure to estradiol, a complex involving phosphorylated SHC and the ER alpha formed and SOS bound to GRB-2. These effects could be abrogated by co-administration of the pure anti-estrogen ICI 182780 [60]. Taken together, our data support the possibility that the LTED cells co-opt a classical growth factor pathway to allow estradiol to exert non-genomic effects on MAP kinase.
123
cells predominantly utilize the genomic pathway. Through downstream interactions on the cell cycle, MAP kinase mediated events synergize with the genomic effects of estradiol on transcription. Under these circumstances, smaller amounts of estradiol are required to induce cell proliferation than in wild type cells. The series of studies presented focus on the ability of breast cancer cells to adapt to various therapies. The possibility that tumors become hypersensitive to estradiol provides a plausible explanation why third generation aromatase inhibitors are more effective than the first and second generation inhibitors. The magnitude of estradiol inhibition is greater with the third generation inhibitors. In the long run, means to counteract the adaptive process will be necessary to prevent progression of breast cancer cells to a hormone independent state. Further work to explain this adaptive process should result in development of better therapies for patients and perhaps even cure if applied at early stages of disease.
10. Summary Our data indicate that breast cancer cells can adapt to long-term estradiol deprivation and develop means to re-grow in the presence of very small concentrations of estradiol. Adaptation involves upstream mechanisms with an increase in estrogen receptor level and in basal transcription rates. However, this does not appear to be the primary event mediating hypersensitivity since LTED cells do not exhibit enhanced sensitivity when levels of ER reporter activity, progesterone receptor levels, c-myc levels, or MAP kinase activation serve as endpoints of ER transcriptional rates. Overexpression of growth factor pathways are involved in the hypersensitivity process. The LTED cells depend upon activation of growth factor pathways for proliferation when exogenous estradiol is not added to culture media. Proof of principle experiments demonstrate that administration of TGF␣ can shift estradiol dose response curves two logs to the left and that this effect can be abrogated by co-administration of a MAP kinase inhibitor. Finally, hypersensitivity to estradiol can be reduced in LTED cells with administration of a MAP kinase inhibitor. Taken together, these results suggest participation of growth factor pathways in the process of adaptive hypersensitivity. Measurements of the transcription factor E2F1 suggest that the actual process of hypersensitivity occurs at the level of the cell cycle and involves synergistic interactions between Ras mediated events and c-myc. A key question is why MAP kinase is elevated in LTED cells. This does not appear to be due to constitutive activation of growth factor receptors or growth factor secretion since MAP kinase activity can be blocked by a pure anti-estrogen. Non-genomic effects of estradiol on the other hand appear to be responsible. We have presented evidence for this which includes activation of the SHC, phosphorylated SHC, SOS, GRB-2, MAP kinase pathway. Based upon these data, we propose a working model for hypersensitivity in which LTED cells utilize both non-genomic and genomic pathways of estrogen action whereas the wild type MCF-7
References [1] S.H. Landis, T. Murray, S. Bolden, P.A. Wingo, Cancer statistics, CA Cancer J. Clin. 48 (1998) 6–29. [2] R.J. Santen, A. Manni, H. Harvey, C. Redmond, Endocrine treatment of breast cancer in women, Endocrinol. Rev. 11 (1990) 221–265. [3] R.J. Santen, T.J. Worgul, A. Lipton, H. Harvey, A. Boucher, E. Samojlik, S.A. Wells, Aminoglutethimide as treatment of postmenopausal women with advanced breast carcinoma, Ann. Int. Med. 96 (1982) 94–101. [4] R.J. Santen, T.J. Worgul, E. Samojlik, A. Interrante, A.E. Boucher, A. Lipton, H.A. Harvey, D.S. White, E. Smart, C. Cox, S. Wells, A randomized trial comparing surgical adrenalectomy with aminoglutethimide plus hydrocortisone in women with advanced breast cancer, N. Engl. J. Med. 305 (1981) 545–551. [5] Early Breast Cancer Collaborative Group, Tamoxifen for early breast cancer: an overview of the randomized trials, Lancet 351 (1998) 1451–1467. [6] B. Fisher, J. Dignam, J. Bryant, et al., Five versus more than five years of tamoxifen therapy for breast cancer patients with negative lymph modes and estrogen receptor-positive tumors, J. Natl. Cancer Inst. 88 (1996) 1529–1542. [7] H.J. Stewart, A.P. Forrest, D. Everington, C.C. McDonald, J.A. Dewar, R.A. Hawkins, R.J. Prescott, W.D. George, Randomised comparison of 5 years of adjuvant tamoxifen with continuous therapy for operable breast cancer: the Scottish Cancer Trials Breast Group, Br. J. Cancer 74 (1996) 297–299. [8] M.M. Gottardis, S.Y. Jiang, M.-H. Jeng, V.C. Jordan, Inhibition of tamoxifen-stimulated growth of an MCF-7 tumor variant in athymic mice by novel steroidal antiestrogens, Cancer Res. 49 (1989) 4090– 4093. [9] R.J. Santen, Long term tamoxifen therapy: can an antagonist become an agonist? (Editorial), J. Clin. Endocrinol. Metab. 81 (1997) 2027– 2029. [10] K.B. Horwitz, When tamoxifen turns bad (Editorial), Endocrinology 136 (1995) 821–823. [11] S. Masamura, S.J. Santner, D.F. Heitjan, R.J. Santen, Estrogen deprivation causes estradiol hypersensitivity in human cancer cells, J. Clin. Endocrinol. Metab. 80 (1995) 2918–2925. [12] S.S. Simons, H. Oshima, D. Szapary, Higher levels of control: modulation of steroid hormone-regulated gene transcription, Mol. Endocrinol. 6 (1992) 995–1002.
124
R. Santen et al. / Journal of Steroid Biochemistry & Molecular Biology 79 (2001) 115–125
[13] D.P. McDonnell, E. Vegeto, B.W. O’Malley, Identification of a negative regulatory function for steroid receptors, Proc. Natl. Acad. Sci. U.S.A. 89 (1992) 10563–10567. [14] T. Chun, D. Gregg, D.K. Sarkar, J. Gorski, Differential regulation by estrogens of growth and prolactin synthesis in pituitary cells suggests that only a small pool of estrogen receptors is required for growth, Proc. Natl. Acad. Sci. U.S.A. 95 (1998) 2325–2330. [15] F. Labrie, R. Veilleux, A wide range of sensitivities to androgens develops in cloned Shionogi mouse mammary tumor cells, Prostate 8 (1986) 293–300. [16] F. Labrie, R. Veilleux, A. Fournier, Low androgen levels induce the development of androgen-hypersensitive cell clones in Shionogi mouse mammary carcinoma cells in culture, J. Natl. Cancer Inst. 80 (1988) 1138–1147. [17] H.E. Kulin, M.M. Grumbach, S.L. Kaplan, Changing sensitivity of the pubertal hypothalamic feedback mechanism in man, Science 166 (1969) 1012–1013. [18] J. Maruca, H.E. Kulin, S.J. Santner, Perturbations of negative feedback sensitivity in agonadal patients undergoing estrogen replacement therapy, J. Clin. Endocrinol. Metab. 56 (1983) 3–59. [19] S.J. Legan, F.J. Karsch, D.L. Foster, The endocrine basis of seasonal reproductive function in the ewe: a marked change in response to the negative feedback action of estradiol on luteinizing hormone secretion, Endocrinology 101 (1977) 818–824. [20] R.J. Santen, J.N. Friend, D. Trojanowski, B. Davis, E. Samojlik, C.W. Bardin, Prolonged negative feedback suppression after estradiol administration: proposed mechanism of eugonadal secondary amenorrhea, J. Clin. Endocrinol. Metab. 47 (1978) 1220–1229. [21] M.M. Montano, W.L. Kraus, B.S. Katzenellenbogen, Identification of a novel transferable cis element in the promoter of an estrogen-responsive gene that modulates sensitivity to hormone and antihormone, Mol. Endocrinol. 11 (1997) 330–341. [22] D.A. Jackson, C.D. Collier, H. Oshima, S.S. Simons, Modulation of TAT gene induction by glucocorticoids involves a neutralizing sequence, J. Steroid Biochem. Mol. Biol. 66 (1998) 79–91. [23] M. Gershanovich, H.A. Chaudri, D. Campos, H. Lurie, A. Bonaventura, M. Jeffrey, F. Buzzi, I. Bodrogi, H. Ludwig, P. Reichardt, N. O’Higgins, G. Romieu, P. Friederich, M. Lassus, Letrozole, a new oral aromatase inhibitor: randomised trial comparing 2.5 mg daily, 0.5 mg daily and aminoglutethimide in postmenopausal women with advanced breast cancer, Ann. Oncol. 9 (1998) 639–645. [24] J. Bergh, J. Bonneterre, H.J. Illiger, et al., Vorozole (RivizorTM ) versus aminoglutethimide (AG) in the treatment of postmenopausal breast cancer relapsing after tamoxifen, Proc. Am. Soc. Clin. Oncol. 33 (1997) 155a (Abstract No. 543). [25] P. Dombernowsky, I. Smith, G. Falkson, R. Leonard, L. Panasci, J. Bellmunt, W. Bezwoda, G. Gardin, A. Gudgeon, M. Morgan, A. Fornasiero, W. Hoffman, J. Michel, T. Hatschek, T. Tjabbes, H.A. Chaudri, U. Hornberger, P.F. Trunet, Letrozole, a new oral aromatase inhibitor for advanced breast cancer: double-blind randomized trial showing a dose effect and improved efficacy and tolerability compared with megestrol acetate, J. Clin. Oncol. 16 (1998) 453–461. [26] R.J. Santen, H.A. Harvey, Use of aromatase inhibitors in breast carcinoma, Endocr. Relat. Cancer 6 (1999) 75–92. [27] W.V. Welshons, A.C. Jordan, Adaptation of estrogen-dependent MCF-7 cells to low estrogen (phenol red-free) culture, Eur. J. Cancer Clin. Oncol. 23 (1987) 1935–1939. [28] N. Brunner, V. Boulay, A. Fojo, C.E. Freter, M.E. Lippman, R. Clarke, Acquisition of hormone-independent growth in MCF-7 cells is accompanied by increased expression of estrogen-regulated genes but without detectable DNA amplifications, Cancer Res. 53 (1993) 283–290. [29] R. Clarke, N. Brunner, B.S. Katzenellenbogen, E.W. Thompson, J.F. Norman, C. Koppi, S. Paik, M.E. Lippman, R.B. Dickson, Progression of human breast cancer cells from hormone-dependent to hormone-independent growth both in vitro and in vivo, Proc. Natl. Acad. Sci. U.S.A. 96 (1989) 3649–3653.
[30] B.S. Katzenellenbogen, K.L. Kendra, M.J. Norman, Y. Berthois, Proliferation, hormonal responsiveness, and estrogen receptor content of MCF-7 human breast cancer cell lines grown in the short-term and long-term absence of estrogens, Cancer Res. 47 (1987) 4355–4360. [31] M.E. Herman, B.S. Katzenellenbogen, Alterations in transforming growth factor -␣ and - production and cell responsiveness during the progression of MCF-7 human breast cancer cells to estrogen-autonomous growth, Cancer Res. 54 (1994) 5867–5874. [32] H. Cho, P.A. Ng, B.S. Katzenellenbogen, Differential regulation of gene expression by estrogen in estrogen-growth independent and -dependent MCF-7 human breast cancer cell sublines, Mol. Endocrinol. 5 (1991) 1323–1330. [33] C.S. Murphy, J.J. Pink, V.C. Jordan, Characterization of a receptor-negative, hormone-unresponsive clone derived from a T-47-D human breast cancer cell line kept under estrogen-free conditions, Cancer Res. 50 (1990) 7285–7292. [34] R.J. Daly, P.D. Darbre, Cellular and molecular events in loss of estrogen sensitivity in ZR-75-1 and T-47-D human breast cancer cells, Cancer Res. 50 (1990) 5868–5875. [35] V.C. Jordan, The control of hormone dependent breast cancer—are we talking about estrogen alone, Eur. J. Cancer Clin. Oncol. 24 (1988) 1245–1248. [36] R.I. Nicolson, J.M.W. Gee, D.L. Manning, A.E. Wakeling, M.M. Montano, B.S. Katzenellenbogen, Responses to pure antiestrogens (ICI 164384, ICI 182780) in estrogen-sensitive and -resistant experimental and clinical breast cancer, Ann. New York Acad. Sci. 761 (1995) 148–163. [37] W. Yue, J.-P. Wang, L. Freeman, R.J. Santen, Activation of MAP kinase in estrogen dependent human breast cancer cells during regrowth in the absence of exogenous estrogen, Endo. Abstract P1 (1998) 126. [38] W.-S. Shim, M. Conaway, S. Masamura, W. Yue, J.-P. Wang, R.J. Santen, Estradiol hypersensitivity and MAP kinase expression in long term estrogen deprived human breast cancer cells in vivo, Endocrinology 141 (2000) 396–405. [39] L.A. Martin, C.M.W. Chan, I. Farmer, C. Marshall, M. Dowsett, Molecular changes occurring during lone-term oestrogen deprivation, Aromatase 2000 and the 3rd generation (Abstract), 3–7 November 2000, Port Doughlas, Australia. [40] A. Kirschenbaum, M. Ren, A.C. Levine, Enhanced androgen sensitivity in serum-free medium of a subline of the LnCAP human prostate cancer cell line, Steroids 58 (1993) 439–444. [41] J. Kokontis, K. Takakura, N. Hay, S. Liao, Increased androgen receptor activity and altered c-myc expression in prostate cancer cells after long-term androgen deprivation, Cancer Res. 54 (1994) 1566–1573. [42] J. Kokontis, N. Hay, S. Liao, Progression of LnCAP prostate tumor cells during androgen deprivation: hormone-independent growth, repression of proliferation by androgen, and role for p27kip1 in androgen-induced cell cycle arrest, Mol. Endocrinol. 12 (1998) 941– 953. [43] R.J. Santen, Endocrine treatment of prostate cancer, J. Clin. Endocrinol. Metab. 75 (1992) 685–689. [44] A.M. Soto, H. Justicia, J.W. Wray, C. Sonnenschein, P-nonyl-phenol: an estrogenic xenobiotic released from modified polystyrene, Environ. Health Perspect. 92 (1991) 167–173. [45] A.M. Soto, C. Sonnenschein, K.L. Chung, M.F. Fernandez, N. Olea, F.O. Serrano, The E-SCREEN assay as a tool to identify estrogens: an update on estrogenic environmental pollutants, Environ. Health Perspect. 103 (Suppl. 7) (1995) 113–122. [46] A.V. Krishnan, P. Stathis, S.F. Permuth, L. Tokes, D. Feldman, bis-Phenol-A: an estrogenic substance is released from polycarbonate flasks during autoclaving, Endocrinology 132 (1993) 2279–2286. [47] L.F. Ren, M.A. Marquardt, J.J. Lech, Estrogenic effects of nonylphenol on PS2, ER and MUC1 gene expression in human breast cancer cells—MCF-7, Chem. Biol. Interact. 104 (1997) 55–64.
R. Santen et al. / Journal of Steroid Biochemistry & Molecular Biology 79 (2001) 115–125 [48] K. Connor, K. Ramamoorthy, M. Moore, M. Mustain, I. Chen, S. Safe, T. Zacharewski, B. Gillesby, A. Joyeux, P. Balaguer, Hydroxylated polychlorinated biphenyls (PCBS) as estrogens and antiestrogens-structure–activity relationships, Toxicol. Appl. Pharmacol. 145 (1997) 111–123. [49] R.B. Dickson, M.E. Lippman, Growth factors in breast cancer, Endocr. Rev. 16 (1995) 559–589. [50] C.A. Lange, J.K. Richer, T.J. Shen, K.B. Horwitz, Convergence of progesterone and epidermal growth factor signaling in breast cancer-potentiation of mitogen-activated protein kinase pathways, JBC 273 (1998) 31308–31316. [51] R.M. Elledge, W.L. McGuire, C.K. Osborne, Prognostic factors in breast cancer, Semin. Oncol. 19 (1992) 244–253. [52] S. Kato, H. Endoh, Y. Masuhiro, T. Kitamoto, S. Uchiyama, H. Sasaki, S. Masushige, Y. Gotoh, E. Nishida, H. Kawashima, Activation of the estrogen receptor through phosphorylation by mitogen-activated protein kinase, Science 270 (1995) 1491–1494. [53] G. Bunone, P.A. Briand, R.J. Miksicek, D. Picard, Activation of the unliganded estrogen receptor by EGF involves the MAP kinase pathway and direct phosphorylation, EMBO J. 15 (1996) 2174–2183. [54] M.-H. Jeng, M.A. Shupnik, T.P. Bender, E.H. Westin, D. Bandyopadhyay, R. Kumar, S. Masamura, R.J. Santen, Estrogen receptor expression and function in long-term estrogen-deprived human breast cancer cells, Endocrinology 139 (1998) 4164–4174.
125
[55] M.-H. Jeng, W. Yue, A. Eischeid, J.-P. Wang, R.J. Santen, Role of MAP kinase in the enhanced cell proliferation of long term estrogen deprived human breast cancer cells, Breast Cancer Res. Treatment 62 (2000) 167–175. [56] G. Leone, J. DeGregori, R. Sears, L. Jakoi, J.R. Nevins, Myc and Ras collaborate in inducing accumulation of active cycling E/Cdk2 and E2F, Nature 387 (1997) 422–425. [57] G. Leone, J. DeGregori, R. Sears, L. Jakoi, J.R. Nevins, Myc and Ras collaborate in inducing accumulation of active cyclin E/Cdk2 and E2F, Nature 387 (1997) 932. [58] P.H. Watson, R.T. Pon, R.P.C. Shiu, Inhibition of c-myc expression by phosphorothioate antisense olingonucletide identifies a critical role for c-myc in the growth of human breast cancer, Cancer Res. 51 (1991) 3996–4000. [59] W. Yue, R.J. Santen, S. Manna, R. Kumar, Cell cycle events associated with the ability of hormone dependent breast cancer cells to grow in the absence of exogenous estrogen, in: Proceedings of the 81st Annual Meeting of the Endocrine Society, San Diego, CA, 1999. [60] R.X. Song, R.A. McPherson, W. Yue, J.-P. Wang, R.J. Santen, SHC phosphorylation and interaction with ER alpha are required for estrogen-induced MAP kinase phosphorylation in MCF-7 human breast cancer cells, in: Proceedings of the 82nd Annual Meeting of the Endocrine Society, Abstract 1401, Toronto, Canada, 2000, p. 338.