c-Src Modulates Estrogen-Induced Stress and Apoptosis in Estrogen-Deprived Breast Cancer Cells

Developing drugs that target the estrogen receptor (ER) either directly (tamoxifen) or indirectly (aromatase inhibitors) has improved the prognosis of breast cancer (1, 2). Although aromatase inhibitors show considerable advantages over tamoxifen with respect to patient disease-free survival and tolerability, acquisition of resistance to all forms of endocrine treatments is inevitable (3, 4). Multiple mechanistic changes are involved in antihormone resistance, which provides the scientific rationale for the clinical development of additional targeted therapies (5, 6). It is well-known that the biologic actions of E₂ are mediated through the ER, which functions in the nucleus as ligand-dependent transcription factors to promote gene transcription and stimulation of cell growth (7). Paradoxically, laboratory evidence shows that E₂ can induce apoptosis in sensitive antihormone-resistant cells in vivo (8–10). This new targeted strategy provides novel therapeutic approaches to endocrine-resistant breast cancer. A recent phase II clinical trial reports that E₂ provides a clinical benefit for patients with aromatase inhibitor–resistant advanced breast cancer (11). In addition, the laboratory results on E₂-induced apoptosis using antihormone-treated MCF-7 cells have been used to explain the reduction of breast cancer and the reduction in mortality observed in postmenopausal hysterectomized women in their 60s treated with conjugated equine estrogen (CEE) when compared with a placebo-treated control (12). The antitumor action of CEE is observed, not only during CEE treatment but also for 6 years after treatment. These data suggest a cidal effect for CEE and has been noted recently (13). These encouraging clinical results prompted us to investigate the mechanisms underlying E₂-induced apoptosis to increase the therapeutic benefits of E₂ in aromatase inhibitor–resistant breast cancer.

Experimental evidence has established the oncogene, c-Src, as a critical component of multiple signaling pathways that regulate proliferation, survival, angiogenesis, and metastasis (14, 15). Increased c-Src activity is believed to play an important role in the development and progression of breast cancer (16), and c-Src has been considered as a survival signal for endocrine resistant breast cancer cells (17). Therefore, a c-Src inhibitor administered as a single-agent or in combination with other antihormone therapy has the potential to enhance the inhibitory effects of antihormones and delay antihormone resistance (18). These observations highlight c-Src as an important therapeutic target for the treatment of human breast cancer.

Mitochondria are important intracellular organelles involved in apoptosis via an intrinsic pathway (19). Although the molecular mechanisms of E₂-induced apoptosis are not fully understood, evidence indicates that mitochondria-related caspase pathways are involved (20, 21). Similarly, a variety of events in apoptosis focus on mitochondria, including the loss of mitochondrial transmembrane potential, release of cytochrome c, and participation of pro- and antiapoptotic Bcl-2 family proteins (22, 23). However, accumulating evidence suggests that the endoplasmic reticulum where members of the Bcl-2 family of proteins localize is also a major point of integration of pro-apoptotic signaling or damage sensing (24, 25). The endoplasmic reticulum senses local stress such as unfolded protein (UPR) through a set of pathways known as the UPR (26), which activates 3 transmembrane sensors PRK-like endoplasmic reticulum kinase (PERK), inositol-requiring 1 alpha (IRE-1α), and activating transcription factor 6 (ATF-6) in endoplasmic reticulum (26). Depending on the duration and degree of stress, the UPR can provide either survival signals by activating adaptive and antiapoptotic signals or death signals by inducing cell death programs (27, 28).

We have found that E₂ changes the cell number according to the treatment period in long-term E₂-deprived breast cancer cell lines MCF-7:5C and MCF-7:2A (25). E₂ has the capacity to decrease around 80% of cell number in MCF-7:5C cells after 7 days treatment, whereas in MCF-7:2A cells after 2-week treatment (29). Unexpectedly, the c-Src inhibitor effectively rescues the decreasing of cell number by E₂ in 2 long-term E₂-deprived cell lines (29). The goal of this study is to identify the mechanisms underlying the early stage of E₂-induced apoptosis and the function of c-Src in the process of E₂-initiated apoptosis. To that end, we show that E₂ triggers endoplasmic reticulum stress and oxidative stress, which activate 2 main apoptotic pathways, the mitochondrial (intrinsic) and death receptor (extrinsic) pathways, whereas c-Src plays an essential role in mediating stress responses induced by E₂ in MCF-7:5C cells. These findings have important clinical implications for the appropriate application of combination therapies in advanced aromatase inhibitor–resistant breast cancer.

Estradiol was purchased from Sigma-Aldrich. c-Src inhibitor PP2 was purchased from CalBiochem. ERα antibody was from Santa Cruz Biotechnology. Total mitogen-activated protein kinase (MAPK), phosphorylated MAPK, phosphorylated c-Src, phosphorylated eIF2α, total eIF2α, and IRE1α antibodies were from Cell Signaling Technology. Total c-Src mouse antibody was from Millipore. Estrogen dendrimer conjugate (EDC) was a kind gift by Dr. J.A. Katzenellenbogen (University of Illinois at Urbana-Champaign, Urbana, IL).

Estrogen-deprived MCF-7:5C cells were maintained in estrogen-free RPMI-1640 medium supplemented with 10% dextran-coated charcoal-stripped FBS as previously described (20). The DNA fingerprinting pattern of cell line is consistent with the report by the American Type Culture Collection (29). The DNA content of the cells, a measure of proliferation, was determined by using a DNA fluorescence quantitation kit (29).

Briefly, MCF-7:5C cells were treated with vehicle (0.1% EtOH) and E₂ (10⁻⁹ mol/L), respectively. Cells were harvested and gradually fixed with 75% EtOH on ice. After staining with propidium iodide (PI), cells were analyzed using a FACSort flow cytometer (Becton Dickinson), and the data were analyzed with ModFit software.

The FITC Annexin V Detection Kit I (BD Pharmingen) was used to quantify apoptosis by flow cytometry according to the manufacturer's instructions. In brief, MCF-7:5C cells were treated with different compounds, respectively. Cells were suspended in 1× binding buffer and 1 × 10⁵ cells were stained simultaneously with fluorescein isothiocyanate (FITC)-labeled Annexin V (FL1-H) and PI (FL2-H). Cells were analyzed using FACSort flow cytometer (Becton Dickinson).

Mitochondrial membrane potential was measured by flow cytometry using the cationic lipophilic green fluorochrome rhodamine-123 (Rh123; Molecular Probes) as previously described (20). Disruption of ΔΨ_m is associated with a lack of Rh123 retention and a decrease in fluorescence.

Intracellular reactive oxygen species (ROS) were detected by fluorescent dye 2′,7′-dichlorofluorescein diacetate (H₂DCFDA, Invitrogen; ref. 30). Briefly, MCF-7:5C cells were treated with E₂ for different time points using vehicle (0.1% EtOH) cells as control. Cells were loaded with 1 μmol/L CM-H₂DCFDA for 10 minutes and washed with PBS twice. Then, cells were monitored at fluorescence 530 nm and an excitation wavelength of 488 nm through flow cytometry.

Proteins were extracted in cell lysis buffer (Cell Signaling Technology) supplemented with Protease Inhibitor Cocktail (Roche) and Phosphatase Inhibitor Cocktail Set I and Set II (Calbiochem). The immunoblotting was conducted as previously described (29).

Transient transfection assay was conducted using a dual-luciferase system (Promega). To determine ER transcriptional activity, cells were transfected with an estrogen response element (ERE)-regulated (pERE (5×) TA-ffLuc plus pTA-srLuc) dual-luciferase reporter gene sets. The cells were treated with E₂ for 24 hours following the transfection. Then, the cells were harvested and processed for dual-luciferase reporter activity, in which the firefly luciferase activity was normalized by Renilla luciferase activity.

Total RNA, isolated with an RNeasy Micro Kit (Qiagen), was converted to first-strand cDNA using a kit from Applied Biosystems. Quantitative real-time PCR assays were done with the SYBR Green PCR Master Mixes (Applied Biosystems) and a 7900HT Fast Real-time PCR System (Applied Biosystems). All primers were synthesized in Integrated DNA Technologies. The sequence of primers is shown in the Supplementary Table S1. All the data were normalized by 36B4.

MCF-7:5C cells were treated with different compounds for 72 hours. Cells were harvested in TRIzol. Total RNA was isolated with an RNeasy Micro Kit. These long RNA samples were first converted into a library of cDNA fragments. Sequencing adaptors were subsequently added to each cDNA fragment and a 2 × 100 bp paired-end sequence was obtained from each cDNA using high-throughput sequencing technology (Illumina GAII). An average of 73.8 million such reads was produced for each sample. The resulting sequence reads were aligned to reference genome build hg19 using TopHat 1.3.0 (31), a splice junction aligner. Transcript abundance was estimated as Fragments Per Kilobase of exon per Million fragments mapped (FPKM), using Cufflinks 1.0.3 (32). Additional analysis was conducted with the alternative expression analysis by sequencing (Alexa-seq) software package as previously described (33). Gene expression measures were compared between Cufflinks and Alexa-seq for the set of 17,993 overlapping genes. Correlations were excellent with Spearman correlations of 0.955 to 0.971 for the 6 samples. Pathway analysis was conducted with DAVID (34) on lists of differentially expressed gene lists.

All reported values are the means ± SE. Statistical comparisons were determined with 2-tailed Student t tests. Results were considered statistically significant if the P < 0.05.

It is well-documented that E₂ stimulates growth and prevents apoptosis in wild-type breast cancer cells and estrogen-responsive osteoblast cells (35, 36). In contrast, physiologic concentrations of E₂ induce apoptosis in long-term E₂-deprived breast cancer cells (20, 21). c-Src plays a critical role in relaying ER signaling pathways in breast cancer cells (37). To investigate the function of E₂ and c-Src in long-term E₂-deprived breast cancer cells MCF-7:5C, a specific c-Src tyrosine kinase inhibitor, PP2, was used to block phosphorylation of c-Src (Fig. 1A). It also effectively abolished the growth pathways including the MAPK and phosphoinositide 3-kinase (PI3K)/AKT pathways in MCF-7:5C cells (Fig. 1A). E₂ activated c-Src through ER as 4-hydroxytamoxifen (4-OHT) completely suppressed phosphorylation of c-Src (Fig. 1B). Although our previous finding showed that E₂ initiates apoptosis in MCF-7:5C cells (20), E₂ was able to activate nongenomic (Supplementary Fig. S1A) and genomic pathways in MCF-7:5C cells (Fig. 1C). These actions were blocked by the c-Src inhibitor, PP2 (Fig. 1C and Supplementary Fig. S1A). Even though the characteristic E₂-induced apoptosis occurs after 72-hour treatment (20), cell numbers were initially increased by E₂ with a high percentage in S-phase (Fig. 1D). All of these results suggested that E₂ caused an imbalance between growth and apoptosis in MCF-7:5C cells.

We have shown that long-term E₂ deprivation increases c-Src activity (29). Therefore, we addressed the question of whether the c-Src inhibitor, PP2, in combination with E₂ would enhance apoptosis in MCF-7:5C cells. Unexpectedly, the c-Src inhibitor blocked apoptosis initiated by E₂ (Fig. 2A and Supplementary Fig. S1D). To confirm that inhibition of c-Src could block E₂-induced apoptosis, a specific siRNA was used to knock down c-Src in MCF-7:5C cells (Fig. 2B), which reduced the percentage of Annexin V binding induced by E₂ (Fig. 2C). Further experiments showed that E₂ disrupted mitochondrial membrane potential (ΔΨ_m) after 48-hour treatment, which was measured by flow cytometry using Rh123 (Fig. 2D). The c-Src inhibitor PP2 and 4-OHT both prevented reduction of Rh123 retention induced by E₂ (Fig. 2D). These data showed that E₂-triggered apoptosis use the c-Src tyrosine kinase pathway. To evaluate the role of the nongenomic pathway in E₂-induced apoptosis, studies were completed with a synthetic ligand, EDC, that only activates the nongenomic pathway at certain concentration (38). The results showed that EDC (10⁻⁸ mol/L) activated the nongenomic pathway incorporating c-Src (Supplementary Fig. S2). Importantly, EDC had no capacity to activate endogenous E₂ target gene pS2 and did not induce apoptosis in MCF-7:5C cells (Supplementary Fig. S2). All of these findings suggested that the nongenomic pathway does not play a critical role in triggering E₂-induced apoptosis.

The ER is the initial site for E₂ to induce apoptosis as anti-estrogens ICI 182,780 and 4-OHT completely block apoptosis triggered by E₂ (ref. 20 and Supplementary Fig. S3A). In addition to the mediation of ER growth pathways, c-Src is involved in the process of ligand-activated ER ubiquitylation (39). Therefore, blockade of c-Src tyrosine kinase with PP2 further increased ERα protein and mRNA expression levels in MCF-7:5C cells (Fig. 3A). E₂ activated ERE activity, which could be blocked by 4-OHT but not by PP2 (Fig. 3B). It was interesting to find that the c-Src inhibitor alone could upregulate E₂-inducible gene pS2 and was additive with E₂ to elevate pS2 mRNA level (Fig. 3C). Another important ER target gene progesterone receptor (PR) has been regarded as an indicator of a functional ER pathway, as expression of PR is regulated by E₂. Although the c-Src inhibitor alone did not elevate PR expression, it dramatically synergized with E₂ to upregulate PR mRNA (Fig. 3D). All of these results showed that blockade of c-Src increased expression of classical ER target genes. It also implied that classical ER pathway might not directly involve in the E₂-induced apoptosis.

To further investigate the mechanisms of the suppression of E₂-induced apoptosis by PP2, RNA-seq analysis was conducted to examine the genes regulated by E₂ to trigger apoptosis in MCF-7:5C cells. A wide range of apoptosis-related genes was activated by E₂ (Fig. 4A), which were functionally classified into 3 groups: TP53-related genes (such as TP63, PMAIP1, and CYFIP2), stress-related genes (such as HMOX1, PPP1R15A, ZAK, NUAK2, etc.), and inflammatory response–related genes (such as LTB, FAS, TNFRSF21, CXCR4, etc.). Most were stress-related genes (Supplementary Fig. S3B). Consistent with the biologic experiments, 4-OHT and PP2 both blocked apoptosis-related genes induced by E₂ but to a different extent in MCF-7:5C cells (Fig. 4A). The majority of these apoptosis-related genes were confirmed by real-time PCR with similar changes noted as in RNA-seq analysis (Fig. 4B–D and Supplementary Fig. S4). E₂ dramatically increased p63 mRNA levels (Fig. 4B) but did not arrest cells in the G₁ phase. In fact, S-phase was markedly elevated in MCF-7:5C cells (Fig. 1D). Heme oxygenase 1 (HMOX1), which is active at high concentrations of heme, catalyzes the degradation of heme and is thought to function as an oxidative stress indicator (40). In breast cancer cells, cytochrome c is a major source of heme protein found in the inner membrane of the mitochondrion. E₂ markedly increased HMOX1 in MCF-7:5C cells (Fig. 4C), thereby confirming that E₂ may damage the mitochondria and caused cytochrome c release. In contrast to MCF-7:5C cells, E₂ decreased HMOX1 levels in wild-type MCF-7 cells (Supplementary Fig. S5A) and clearly did not change HMOX1 expression in another long-term E₂-deprived cell line MCF-7:2A (Supplementary Fig. S5B), both of MCF-7 and MCF-7:2A do not undergo apoptosis after exposure to E₂ in the first 3 days. In addition, E₂ upregulated TNF family members (such as TNFα, LTA, and LTB), which were abolished by 4-OHT and PP2 (Fig. 4D and Supplementary Fig. S6A and S6B). Low dose of TNFα activated pro-apoptotic pathways in MCF-7:5C cells and inhibited cell growth (Supplementary Fig. S6C and S6D). All of these data suggested that E₂ widely activated intrinsic and extrinsic apoptosis pathways and c-Src was directly involved in mediating apoptosis.

ROSs are the product of oxidative stress by mitochondria, whereas an increase in ROS contributes to degenerative changes in mitochondrial function (41). Under physiologic conditions, cellular ROS levels are tightly controlled by low-molecular-weight radical scavengers and by a complex intracellular network of enzymes such as catalases (CAT) and superoxide dismutases (SOD). Under conditions of lethal stress, ROSs are considered as key effectors of cell death (42). Intracellular ROSs were detected by CM-H₂DCFDA through flow cytometry (Fig. 5A). Detectable ROS appeared after 48 hours of treatment with E₂. The production of ROS reached a peak after 72-hour treatment (Fig. 5A and B). Blocking ER (by 4-OHT) and c-Src (by PP2) abolished ROS generation induced by E₂ (Fig. 5C), indicating that both ER and c-Src were upstream signals of ROS. Free-radical scavengers Mn-TBAP, catalase, and sodium formate (SF) that respectively act on superoxide radical (O₂⁻), H₂O₂, and hydroxyl radical (OH⁻) were used to suppress the production of ROS. Our results suggested that H₂O₂ and OH⁻ were the major sources of ROS induced by E₂. This conclusion was based on the observation that catalase and sodium formate inhibited E₂-induced apoptosis, whereas Mn-TBAP was less effective (Fig. 5D). The RNA-seq analysis showed that E₂ did not significantly regulate antioxidant enzymes such as catalases (CAT) and superoxide dismutases (SOD) in MCF-7:5C cells (data not shown). Our results suggest that E₂ has the potential to damage mitochondria to cause oxidative stress.

Our previous global gene array data show that E₂ activates genes related to endoplasmic reticulum stress in MCF-7:5C cells (25). To relieve stress, sensors of UPRs are activated as initial responses (43). In this study, a significant induction of UPR sensors, inositol-requiring protein 1 alpha (IRE1α) and PERK/eukaryotic translation initiation factor-2α (eIF2α), by E₂ occurred after 24 hours of treatment and was further increased by prolonging treatment times in MCF-7:5C cells (Fig. 6A). The anti-estrogen 4-OHT completely abolished the response (Fig. 6A). The PERK inhibitor blocked phosphorylation of eIF2α and prevented E₂-induced apoptosis (Fig. 6B and C), confirming that endoplasmic reticulum stress was important in the apoptosis initiated by E₂. Phosphorylated eIF2α closely associates with an important cellular energy sensor, adenosine monophosphate (AMP)-activated protein kinase (AMPK) to regulate protein translation and apoptosis (44). AMPK, which phosphorylates many metabolic enzymes to stimulate catabolic pathways and increases the capacity of cells to produce ATP (45), was significantly activated after 48-hour treatment with E₂ (Fig. 6D). The c-Src inhibitor, PP2, blocked the phosphorylation of eIF2α but not IRE1α induced by E₂ (Fig. 6E). PP2 also prevented the activation of AMPK after E₂ treatment (Fig. 6F). All of these data indicate that c-Src acts as an important transducer in the protein kinase pathways (eIF2α and AMPK) of stress response (Fig. 6E and F) that result in apoptosis.

We have previously investigated the inhibitory effects of E₂ on long-term endocrine-resistant breast cancer tumor growth in vivo (8–10). And we have confirmed that this therapeutic effect is related with the apoptosis induced by E₂ (20). This scientific discovery has been used in the clinical trials to treat aromatase inhibitor–resistant patients with breast cancer and 30% of patients receive benefit (11). The potential limitation on translational research in the treatment of hormone-responsive breast cancer is that only 4 ER-positive breast cancer cell lines are available to use routinely (46). Only MCF-7 of the 4 produces the phenotype of E₂-induced apoptosis observed clinically (20, 21). The purpose of establishing long-term E₂ deprivation in vitro models is to mimic administration of an aromatase inhibitor that reduces levels of circulating estrogen in clinical studies (47). After a period of proliferative quiescence lasting a few months, the return of proliferation is similar to the relapses observed 12 to 18 months after primary hormonal therapy in patients. Multiple pathways are involved in the adaptive response to the pressure of E₂ deprivation (48). Although MCF-7 cells grown long-term have been shown to differ substantially in various properties depending upon the number of passages and geographic source of the cell lines, induction of apoptosis by physiologic concentrations of E₂ is the common characteristic of these in vitro model systems (20, 21). Nevertheless, how E₂ induces apoptosis is at present unclear. Our new observation (29) that a c-Src inhibitor paradoxically can block E₂-induced apoptosis naturally demands further study. We examined this aspect of c-Src pharmacology to describe fully this phenomenon and gain an insight into the convergence of ER and c-Src pathways for the modulations of an apoptotic trigger in breast cancer. Here, for the first time, we document that c-Src participates in the mediation of stress responses induced by E₂ to widely activate apoptosis-related genes involved in the intrinsic and the extrinsic apoptosis pathways.

The ER is the initial point for E₂ to induce apoptosis, as anti-estrogens ICI 182,780 and 4-OHT completely block apoptosis triggered by E₂ (ref. 20 and Supplementary Fig. S3A). Contradictory to the traditional apoptosis mechanism caused by cytotoxic chemotherapy with cell-cycle arrest, E₂-induced apoptotic cells simultaneously undergo proliferation with an increased S-phase of cell cycle resulting in increased cell number despite p53 family members being upregulated (Figs. 1D and 4B). E₂ exerts a dual function on MCF-7:5C cells, with both initial proliferation and the apoptosis. In other words, the initial response of E₂ to stimulate growth is the upregulating of classical transcriptional activity by ER (Fig. 3B) without any detected apoptotic changes in the first 24 hours. Activation of apoptotic genes appeared after 48-hour treatment with E₂ (data not shown) and reached a peak by 72 hours (Fig. 4B–D). Consistently, characteristic apoptosis occurred at 72 hours (Fig. 2A). These data suggest that the higher rate of proliferation by E₂ might activate other pathways to trigger apoptosis. Our data show that E₂ caused endoplasmic reticulum stress, which activated the UPR within 24 hours (Fig. 6A). The initial aim of UPR is to restore normal function of the cell; however, if the damage is too severe to repair, the UPR ultimately initiates cell death through activation of the apoptotic pathway (49).

c-Src functioned as an important downstream signal of ER in MCF-7:5C cells, which was activated by E₂ (Fig. 1B, Supplementary Fig. S1A–S1C) and showed multiple levels of association with ER (Figs. 1B and C, 2A, 3A, C, and D). An important finding in this study is that c-Src tyrosine kinase is critical for E₂-induced apoptosis (Fig. 2A, C, and D). This, therefore, raised the question of the actual role played by c-Src in the process of apoptosis induced by E₂. c-Src mediated PI3K/AKT and MAPK growth pathways by E₂ (Fig. 1C). However, specific inhibitors of PI3K/Akt (LY294002) and MAPK (U0126) could inhibit cell growth but did not prevent E₂-induced apoptosis in MCF-7:5C cells (Supplementary Fig. S7), which imply that MAPK/Akt growth pathways are not directly involved in the apoptosis-induced by E₂. In MCF-7:5C cells, E₂ activated the nongenomic pathway after 10-minute treatment and the c-Src inhibitor blocked the nongenomic pathway (Supplementary Fig. S1A and S1B). Detectable elevation of c-Src phosphorylation appeared after 30-minute treatment with E₂ (Supplementary Fig. S1B). Consistent stimulation of c-Src appeared after 24-hour treatment and gradually increased when extending to 48 hours (Fig. 1B and Supplementary Fig. S1C). All of these data suggest that c-Src activation is a direct effect resulting from E₂. To further explore the function of the nongenomic pathway in the process of E₂-induced apoptosis, EDC was used to treat MCF-7:5C cell, which is very ineffective in stimulating transcription of endogenous E₂ target genes (38). The EDC (10⁻⁸ mol/L) activated the nongenomic pathway but without capacity to activate genomic pathway and did not induce apoptosis in MCF-7:5C cells (Supplementary Fig. S2). All of these results suggest that the nongenomic pathway does not play a critical role in the E₂-induced apoptosis. Interestingly, the EDC could continuously activate c-Src and Akt but without any effect on MAPK after 24-hour treatment (Supplementary Fig. S2E), which may be resulted from enhanced association between ERα and membrane growth factor receptor (48).

In addition, E₂ activated classical ERE activity but the c-Src inhibitor could not block the response (Fig. 3B). Furthermore, the c-Src inhibitor collaborated with E₂ to upregulate endogenous ER target genes pS2 and PR (Fig. 3C and D). All of these results imply that classical ER transcriptional pathways are not directly involved in E₂-induced apoptosis. Similarly, Zhang and colleagues reported that the inhibitory effects of E₂ on cell growth are independent of the classical ERE-regulated transcriptional genes (50). Our global gene array data suggest that E₂ signaling can occur through a nonclassical transcriptional pathway involving the interaction of ER with other transcription factors such as activator protein-1 (AP-1) and Sp1, which may regulate stress responses (25). In the present study, E₂ initiated UPR (Fig. 6A), increased ROS production (Fig. 5A), and widely activated apoptosis-related genes (Fig. 4A). The c-Src was involved in the stress responses and inhibition of c-Src decreased the expression of apoptosis-related genes induced by E₂, which are critical mechanisms for the blockade of c-Src to prevent E₂-induced apoptosis.

Overall, E₂ induces endoplasmic reticulum and mitochondrial stresses in MCF-7:5C cells, which subsequently upregulates apoptosis-related genes to activate intrinsic and extrinsic apoptotic pathways. Unexpectedly, c-Src tyrosine kinase plays a critical role in the stress response induced by E₂. These data clearly raise a concern about the ubiquitous use of c-Src inhibitors to treat patients with advanced aromatase inhibitor–resistant breast cancer, thereby undermining the beneficial effects of E₂-induced apoptosis.

No potential conflicts of interest were disclosed.

The views and opinions of the author(s) do not reflect those of the U.S. Army or the Department of Defense.

Conception and design: P. Fan, J. Katzenellenbogen, J.W. Gray, V.C. Jordan

Development of methodology: P. Fan, O.L. Griffith, F.A. Agboke, X. Zou, J. Katzenellenbogen, J.W. Gray

Acquisition of data (provided animals, acquired and managed patients, provided facilities, etc.): P. Fan, F.A. Agboke, X. Zou, R.E. McDaniel, K. Creswell, J.W. Gray, V.C. Jordan

Analysis and interpretation of data (e.g., statistical analysis, biostatistics, computational analysis): P. Fan, O.L. Griffith, P. Anur, K. Creswell, J. Katzenellenbogen, J.W. Gray, V.C. Jordan

Writing, review, and/or revision of the manuscript: P. Fan, O.L. Griffith, J. Katzenellenbogen, V.C. Jordan

Administrative, technical, or material support (i.e., reporting or organizing data, constructing databases): O.L. Griffith, R.E. McDaniel, V.C. Jordan

Study supervision: P. Fan, V.C. Jordan

Synthesis of EDC: S.H. Kim

V.C. Jordan is supported by the Department of Defense Breast Program under Award number W81XWH-06-1-0590 Center of Excellence; subcontract under the SU2C (AACR) Grant number SU2C-AACR-DT0409; the Susan G Komen For The Cure Foundation under Award number SAC100009; GHUCCTS CTSA (Grant # UL1RR031975); and the Lombardi Comprehensive Cancer Center Support Grant (CCSG) Core Grant NIH P30 CA051008. J.W. Gray is supported by a Stand Up to Cancer Dream Team Translational Cancer Research Grant, a Program of the Entertainment Industry Foundation (SU2C-AACR-DT0409). J. Katzenellenbogen is supported by NIH grant (PHS 5R37DK015556). O.L. Griffith was supported by a fellowship from the Canadian Institutes of Health Research (CIHR).

The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.