Mechanisms underlying differential response to estrogen-induced apoptosis in long-term estrogen-deprived breast cancer cells

Models of long-term estrogen-deprived breast cancer cells are utilized in the laboratory to mimic clinical aromatase inhibitor-resistant breast cancer and serve as a tool to discover new therapeutic strategies. The MCF-7:5C and MCF-7:2A subclones were generated through long-term estrogen deprivation of estrogen receptor (ER)-positive MCF-7 cells, and represent anti-hormone-resistant breast cancer. MCF-7:5C cells paradoxically undergo estrogen-induced apoptosis within seven days of estrogen (estradiol, E2) treatment; MCF-7:2A cells also experience E2-induced apoptosis but evade dramatic cell death until approximately 14 days of treatment. To discover and define the mechanisms by which MCF-7:2A cells survive two weeks of E2 treatment, systematic experiments were performed in this study. The data suggest that MCF-7:2A cells employ stronger antioxidant defense mechanisms than do MCF-7:5C cells, and that oxidative stress is ultimately required for MCF-7:2A cells to die in response to E2 treatment. Tumor necrosis factor (TNF) family member activation is also essential for E2-induced apoptosis to occur in MCF-7:2A cells; upregulation of TNFα occurs simultaneously with oxidative stress activation. Although the unfolded protein response (UPR) signaling pattern is similar to that in MCF-7:5C cells, it is not sufficient to cause cell death in MCF-7:2A cells. Additionally, increased insulin-like growth factor receptor β (IGF-1Rβ) confers a mechanism of growth and anti-apoptotic advantage in MCF-7:2A cells.


Introduction
Aromatase inhibitor-resistant breast cancer cells are modeled in vitro by long-term E 2 -deprived breast cancer cell lines. The MCF-7:WS8 cell line represents a clone of the estrogen receptor (ER)-positive cell line MCF-7 that is highly sensitive to E 2 -stimulated growth (1). The MCF-7:5C and MCF-7:2A subclones are derived from the parental MCF-7 cell line through long-term E 2 deprivation (1)(2)(3)(4). MCF-7:5C cells express wild-type ER at a higher level than the parental line, and are progesterone receptor (PR)-negative (3). These cells grow in the absence of E 2 , and do not respond to 4-hydroxytamoxifen (4-OHT) (2,3). MCF-7:2A cells can induce expression of PR and express both wild-type (66 kDa) and mutant (77 kDa) ER (4,5). The mutant ER contains a repeat of exons 6 and 7 and cannot bind E 2 nor anti-estrogens; it is expressed 4-to 10-fold lower than the wild-type ER (6). The total ER level of MCF-7:2A cells is higher than in parental MCF-7 cells, and they also grow in E 2 -free media. 4-OHT and pure anti-E 2 are able to block their growth (4,5).
In addition to the different responses to anti-E 2 observed in MCF-7:5C versus MCF-7:2A cells, they also have different apoptotic responses to E 2 . The MCF-7:5C cells undergo apoptosis and die during the first week of E 2 treatment, whereas the MCF-7:2A cells die later, after two weeks of E 2 treatment (7). MCF-7:5C cell response to estrogens and anti-estrogens has been extensively studied in our lab; the data show that these cells undergo E 2 -induced apoptosis through mechanisms associated with endoplasmic reticulum stress (ERS) and oxidative stress (8,9). Thus far, there has been less focus on the classification and mechanisms of the MCF-7:2A response.
Network enrichment analyses done using gene arrays in timecourse experiments show overexpression of apoptotic-and stress-related pathways in the MCF-7:5C cells after 24-96 h of E 2 treatment; however, these analyses show the MCF-7:2A cells expressing more genes associated with glutathione metabolism during this time period of E 2 exposure (Fig. 1). This suggests that the two cell lines respond to E 2 treatment using different signaling pathways. The MCF-7:5C cells respond by quickly inducing apoptosis, while the antioxidant pathway may be more relevant to the MCF-7:2A cells. Experiments were designed to interrogate the apoptotic, stress and antioxidant pathways in both cell lines to distinguish signaling mechanisms in response to E 2 .
The concept of E 2 -induced death is important because of its clinical relevance. A clinical study published in 2009 (10) compared two doses of E 2 for second-line treatment after breast cancer patients had failed aromatase inhibitor therapy. The authors showed that after long-term anti-hormone therapy, no response is lost with the lower dose of E 2 ; overall Mechanisms underlying differential response to estrogen-induced apoptosis in long-term estrogen-deprived breast cancer cells about 30% of women responded to E 2 treatment. The goal of this study is to uncover the mechanisms preventing the other 70% of patients from responding, and perhaps find ways to circumvent their resistance. To this end, MCF-7:2A cells were used as a model for E 2 -deprived breast tumors with the ability to evade E 2 -induced apoptosis in the clinic.

Materials and methods
Cell culture. All cell lines were cultured in phenol red-free RPMI-1640 media supplemented with 10% charcoal-stripped fetal bovine serum (SFS). Media and treatments were replaced every three days. Estradiol (E 2 ) (Sigma-Aldrich, St. Louis, MO, USA), buthionine sulfoximine (BSO) (Sigma-Aldrich), and combinations were dissolved in ethanol and then in media. AG1024 (Calbiochem, San Diego, CA, USA) was dissolved in DMSO and then in media.
Cell cycle analysis. MCF-7:2A cells were cultured in dishes and treated with vehicle (0.1% ethanol) or E 2 (10 -9 mol/l, 1 nM). Cells were harvested after 24 h, fixed in 75% ethanol on ice, stained with propidium iodide and sorted using FACS flow cytometry (Becton Dickinson, San Jose, CA, USA). Results were analyzed using CellQuest software.
RT-PCR. Cells were harvested using TRIzol, and RNA was isolated using RNeasy mini kit (Qiagen, Valencia, CA, USA). RNA was reverse transcribed to cDNA using a kit (Applied Biosystems, Foster City, CA). SYBR-Green (Applied Biosystems) was used for quantitative real-time polymerase chain reaction (RT-PCR) in a 7900HT Fast Real-Time PCR system (Applied Biosystems).
Statistical analysis. Values reported are means ± standard deviation (SD). Significant differences were found by Student's t-test. P-values <0.05 were considered to indicate a statistically significant difference.   . 5B), suggesting an earlier protective mechanism inherent in these cells to prevent oxidative stress longer than MCF-7:5C cells.

Discussion
This study investigated the mechanisms through which MCF-7:2A cells evade E 2 -induced apoptosis in vitro as a means to understand resistant breast cancer cells after   long-term anti-hormone therapy in the clinic. After failure on an aromatase inhibitor, approximately 30% of breast cancer patients will respond to treatment with E 2 (10); their nascent or remaining breast tumors will become cytostatic or disappear with physiological levels of E 2 . Further, E 2 replacement therapy (ERT) has been shown to reduce the risk of breast cancer in hysterectomized post-menopausal women (12), perhaps due to E 2 -deprived breast cancer cells undergoing E 2 -induced apoptosis before resulting in clinically apparent disease. This study sought to discriminate between E 2 -deprived breast tumors that will quickly respond to treatment with E 2 versus those that will respond more slowly and less dramatically. We modeled these different scenarios with MCF-7:5C and MCF-7:2A cell lines, respectively.
We have found that the UPR, associated with endoplasmic reticulum stress (ERS), is a fundamental element in E 2 -induced MCF-7:5C cell apoptosis (8). In this setting, E 2 triggers UPR and rapidly causes apoptosis within one week of treatment. Two main sensors of the UPR, IRE1α and PERK are activated in both cell lines similarly. PERK activation is confirmed by elevated p-eIF2α, since eIF2α is phosphorylated by activated PERK. In MCF-7:2A cells, the same sensors are activated as in MCF-7:5C cells (Fig. 3), but significant cell death is not apparent at the same timepoint ( Fig. 2A). Despite similar signaling patterns, the biological responses between the two cell lines differ. Our data suggested that another mechanism was preventing cell death after E 2 -induced UPR in MCF-7:2A cells.
Oxidative stress is a critical pathway for MCF-7:2A cells to undergo E 2 -induced apoptosis. MCF-7:2A cells inherently exhibit stronger survival and antioxidant mechanisms than MCF-7:5C cells (Figs. 4-6). This relationship is consistent with previously published data showing that MCF-7 cells with higher levels of glutathione peroxidase-1 (GSHPx-1) can survive better under oxidative stress conditions, such as hydrogen peroxide treatment (13), and that MCF-7 cells can increase antioxidant enzymes (i.e. manganese superoxide dismutase, MnSOD) to prevent TNF-mediated apoptosis (14). Activation of E 2 -induced apoptosis in MCF-7:2A cells also seems to require TNF family member upregulation (Fig. 4A  and B). Oxidative stress occurs concurrently with upregulation of apoptosis-related genes in the TNF family. Whether increased TNFα causes oxidative stress or oxidative stress causes increased TNFα is not yet documented in this setting.
Additionally, B cell lymphoma 2 (BCL2) plays a role in preventing cell death caused by oxidative stress (15). In fact, MCF-7:2A cells exhibit 3.76-fold and 3.02-fold higher basal BCL2 and B cell lymphoma extra large (BCL-xL, BCL2L1) mRNA levels than MCF-7:5C cells, respectively (Table I), providing support for the idea of a stronger survival signal.
Other data from our lab shows that MCF-7:2A cells exhibit 6.19-fold higher glutathione peroxidase 2 gene (GPX2) over MCF-7:5C cells (Table II), illustrating more evidence in favor of increased protection from E 2 -induced oxidative stress and apoptosis in this context.
Increased IGFR promotes anti-hormone resistance in breast cancer, likely through growth factor receptor crosstalk and aberrant ER, MAPK, and AKT signal transduction pathway activation (16)(17)(18). Our data correlate with these findings in that higher IGF-1Rβ mRNA and protein expression confer a growth advantage and apoptotic resistance in MCF-7:2A cells despite treatment with E 2 (Fig. 7). This suggests an IGF-1Rβ signaling pathway that can circumvent normal ER signaling in long-term estrogen-deprived breast cancer cells. Studies using hepatocellular carcinoma cells (HCC) have demonstrated that IGF-1R overexpression can potentially cause increased glutathione transferase (GST) and protection from oxidative stress (19). Although  Chemokine (C-C motif) ligand 5

CCL5
Oxidative stress responsive genes -50.23 Global gene expression analyses were performed, and oxidative stress-related genes were ranked by fold change of MCF-7:2A expression over MCF-7:5C expression. Notably, GPX2 shows the highest fold change.
this mechanism is shown in liver cancer cells, it may apply to our models of breast cancer as well. Perhaps the higher level of IGF-1Rβ in MCF-7:2A cells generates the increased glutathione levels necessary to escape cell death in the presence of E 2 . The evidence thus far shows that TNF family member gene expression, protection against oxidative stress, and growth factor signaling are major mechanisms underlying the different biological responses to E 2 seen in MCF-7:2A cells versus MCF-7:5C cells. Despite similar UPR signaling patterns, MCF-7:2A cells resist ERS-induced death longer and stronger than MCF-7:5C cells. Additional studies may provide further insight into the connection between IGF-1Rβ and glutathione in MCF-7:2A cells, and how this relationship functions in the presence and absence of a stressor such as E 2 . In order to effectively treat breast cancer patients who have undergone exhaustive anti-hormone treatment, and to explain why ERT can prevent breast cancer in some post-menopausal women, the examination of breast cancer cell models of E 2 deprivation is proving invaluable. By understanding mechanisms that prevent apoptosis in these breast cancer cells, we can translate key findings into clinical practice.