Oncogenic p95HER2/611CTF primes human breast epithelial cells for metabolic stress-induced down-regulation of FLIP and activation of TRAIL-R/Caspase-8-dependent apoptosis

Oncotarget93688 www.impactjournals.com/oncotarget www.impactjournals.com/oncotarget/ Oncotarget, 2017, Vol. 8, (No. 55), pp: 93688-93703 Oncogenic p95HER2/611CTF primes human breast epithelial cells for metabolic stress-induced down-regulation of FLIP and activation of TRAIL-R/Caspase-8-dependent apoptosis Rosa Martín-Pérez1,2,3, Rosario Yerbes1, Rocío Mora-Molina1, Ana Cano-González1, Joaquín Arribas4,5,6,7, Massimiliano Mazzone2,3, Abelardo López-Rivas1,7 and Carmen Palacios1 1Centro Andaluz de Biología Molecular y Medicina Regenerativa-CABIMER, CSIC-Universidad de Sevilla-Universidad Pablo de Olavide, Sevilla, Spain 2Lab of Tumor Inflammation and Angiogenesis, VIB, Leuven, Belgium 3Lab of Tumor Inflammation and Angiogenesis, Department of Oncology, KU Leuven, Leuven, Belgium 4Preclinical Research Program, Vall d’Hebron Institute of Oncology (VHIO), Barcelona, Spain 5Department of Biochemistry and Molecular Biology, Universitat Autónoma de Barcelona, Campus de la UAB, Bellaterra, Spain 6Institució Catalana de Recerca i Estudis Avançats (ICREA), Barcelona, Spain 7Centro de Investigación Biomédica en Red-Oncología (CIBERONC), Carlos III Health Institute, Madrid, Spain Correspondence to: Carmen Palacios, email: [email protected] Abelardo López-Rivas, email: [email protected] Keywords: p95HER2/611CTF, metabolic stress, TRAIL-R, FLIP, mTOR Received: January 10, 2017 Accepted: September 16, 2017 Published: October 03, 2017 Copyright: Martín-Pérez et al. This is an open-access article distributed under the terms of the Creative Commons Attribution License 3.0 (CC BY 3.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited. ABSTRACT Oncogenic transformation triggers reprogramming of cell metabolism, as part of the tumorigenic process. However, metabolic reprogramming may also increase the sensitivity of transformed cells to microenvironmental stress, at the early stages of tumor development. Herein, we show that transformation of human breast epithelial cells by the p95HER2/611CTF oncogene markedly sensitizes these cells to metabolic stress induced by the simultaneous inhibition of glucose and glutamine metabolism. In p95HER2/611CTF-transformed cells, metabolic stress activates a TNF related apoptosis-inducing ligand (TRAIL)-R and caspase-8-dependent apoptotic process that requires prior down-regulation of cellular FLICE-like inhibitor protein (c-FLIP) levels. Importantly, sustained mTOR activation is involved in FLIP down-regulation and apoptosis induced by metabolic stress. In vivo experiments in immunodeficient mice demonstrate a requirement for caspase-8 in restraining primary tumor growth of xenografts with p95HER2/611CTF-transformed cells. Collectively, these data define a critical role of the extrinsic pathway of apoptosis in the control of tumor initiation by microenvironmental cues. INTRODUCTION Amplification of a genomic region on chromosome 17q12 containing the HER2/ERBB2 gene, a member of the ERBB receptor family, has been observed in about 25% of breast tumors and this is associated with aggressive disease and lower survival of patients [1]. Despite the remarkable success of anti-HER2/ERBB2 therapies, patients with advanced HER2-positive breast cancer frequently display primary resistance [2]. Moreover, in patients initially sensitive to these agents, acquired resistance may emerge over time [2]. A potential mechanism of resistance to HER2/ERBB2-directed therapies is the expression of an amino terminally truncated form of HER2/ERBB2 known as p95HER2/611CTF generated by alternative initiation of translation [3]. This truncation leads to Research Paper Oncotarget93689 www.impactjournals.com/oncotarget spontaneous homodimerization which results in a gain- of-function compared with wild-type HER2/ERBB2 [3, 4]. p95HER2/611CTF expression leads to deregulation of intracellular signalling pathways, such as the extracellular signal-regulated kinase (ERK) and the phosphoinositide-3-kinase (PI3K)/protein kinase B (AKT)/ mechanistic target of rapamycin (mTOR) pathways [5] that control cell metabolism, growth and proliferation. Clinically, patients with high-p95HER2/611CTF tumors had significantly worst outcome compared with those with low-p95HER2/611CTF tumors [5]. To survive the hostile environment of tumors a number of intracellular signalling pathways are activated in tumor cells, facilitating tumor growth [6]. Although an increased consumption of glucose through aerobic glycolysis is important to meet the increased energetic and biosynthetic demands of tumor cells [7], glutamine is also an essential nutrient to support cell survival and growth of many tumor cells [8]. Glutamine contributes substantially to maintain macromolecular synthesis, glutathione levels and bioenergetics of proliferating tumor cells [8]. Importantly, prior to metabolic adaptation, oncogenic transformation renders tumor cells initially sensitive to nutrient shortage, in particular to glucose and glutamine deprivation [9, 10]. However, the molecular determinants controlling the cell death response of oncogene-transformed cells to nutrient deprivation remain to be identified. The so-called extrinsic pathway of apoptosis utilises membrane-localized death receptors of the tumor necrosis factor (TNF) receptor superfamily to activate the caspases cascade and apoptosis upon ligand binding [11]. Activation of TRAIL receptors leads to the formation of a death-inducing signaling complex (DISC), which includes the receptor itself, the adapter molecule FADD and procaspase-8 [12]. Processing and activation of caspase-8 at the DISC leads to a cascade of apoptotic events which results in the death of the cell. At the DISC level, the apoptotic signal may be inhibited by cFLIP, the homologue of vFLIP in vertebrate cells [13]. Interestingly, in recent years there have been some reports on the involvement of TRAIL receptors in cell fate decisions after endoplasmic reticulum stress [14, 15]. In addition, the induction of an integrated stress response by certain antitumor drugs triggers cell death in different tumor cells through the activation of TRAIL-R2/DR5-dependent apoptosis [16]. Here, we examined the impact of inhibiting both glycolysis and glutamine metabolism on the viability of human breast epithelial cells transformed by the p95HER2/611CTF oncogene. We found that metabolic stress induces TRAIL-R2/DR5 clustering at the plasma membrane and down-regulates FLIP expression, which leads to activation of a TRAIL-R/caspase-8-dependent apoptotic program in p95HER2/611CTF-transformed cells. Remarkably, when glycolysis is inhibited addiction of p95HER2/611CTF-transformed cells to glutamine relies on the sustained activation of mTOR complexes. In addition, in vivo experiments demonstrate a key role of caspase-8 in the prevention of p95HER2/611CTF- mediated tumor formation. Collectively, these results point to a role of the extrinsic apoptotic pathway in the control of tumor growth by microenvironmental factors. RESULTS Expression of p95HER2/611CTF oncogene sensitizes human breast epithelial cells to metabolic stress-induced cell death through a TRAIL-R and caspase-8-dependent apoptotic pathway MCF10A cells, a non-transformed human mammary epithelial cell line, have been used as a model system to investigate the biological properties of oncogenes in tumor initiation and progression [17–19]. We have examined the impact of p95HER2/611CTF expression on the cellular response of MCF10A cells to metabolic stress. Inhibition of glucose metabolism with 2-deoxyglucose (2DG) concurrently with deprivation of glutamine markedly induced apoptosis in p95HER2/611CTF cells as compared to control cells or cells expressing wildtype HER2/ERBB2 (wt-HER2) (Figure 1A–1B). These observations were extended to the 184A1 cell line, another immortalized, non-tumorigenic human breast epithelial cell line. We generated a bulk population of 184A1 cells transformed by retroviral infection with an empty vector (mock) or the p95HER2/611CTF-encoding vector. Initial studies with these cell lines demonstrated a lack of sensitivity of p95HER2-184A1 cells to starvation (Supplementary Figure 1A). Western blot analysis in the bulk population of p95HER2-184A1 cells revealed a markedly reduced HER2 expression as compared to p95HER2-MCF10A cells that may be relevant in the lack of sensitivity to starvation-induced apoptosis. To circumvent this problem we selected different clones from the bulk populations of mock and p95HER2 cells which were subsequently used to determine HER2 expression and their apoptotic response to starvation. Results shown in Supplementary Figure 1B demonstrate that whereas mock clones were resistant to starvation, metabolic stress activated an apoptotic process in the p95HER2 clones. In both cellular models, apoptosis was inhibited in the presence of the HER2/ERBB2 tyrosine kinase inhibitor lapatinib (Figure 1C, Supplementary Figure 1C), strongly supporting the hypothesis that sensitivity to metabolic stress results from the constitutive activation of p95HER2/611CTF tyrosine kinase. Induction of the extrinsic apoptotic pathway by TRAIL receptor activation has been observed in cells undergoing endoplasmic reticulum (ER) stress [14, 15]. We next studied the role of the TRAIL system in the differential activation of apoptosis by the combination Oncotarget93690 www.impactjournals.com/oncotarget of 2DG and glutamine deprivation in p95HER2/611CTF cells. Compared to mock cells, p95HER2/611CTF- transformed cells expressed higher TRAIL-R2/DR5 levels at the cell surface (Supplementary Figure 2A, left panel) and consequently showed an increased sensitivity to exogenous TRAIL (Supplementary Figure 2A. right panel). Remarkably, metabolic stress induced caspase-8 activation specifically in p95HER2/611CTF cells (Figure 2A), suggesting that differential activation of the extrinsic apoptotic pathway may underlie the increased sensitivity of these cells to metabolic stress. Up-regulation of TRAIL-R2/DR5 expression following ER stress has been associated to activation of apoptosis in various cell types [14, 15]. However, we did not observe an increase in either total (Supplementary Figure 2B) or surface (Supplementary Figure 2C) TRAIL-R2/DR5 levels following metabolic stress in p95HER2/611CTF-transformed cells. In contrast, treatment with thapsigargin, a potent inducer of ER stress, significantly enhanced plasma membrane levels of TRAIL-R2/DR5 (Supplementary Figure 2C). Likewise, expression levels of FADD and caspase-8, two key components of the death-inducing signaling complex (DISC) in the extrinsic apoptotic pathway, remained unaffected after metabolic stress (Supplementary Figure 2B). However, silencing caspase-8 expression by RNA interference markedly inhibited metabolic stress-induced apoptosis in p95HER2/611CTF cells (Figure 2B, Supplementary Figure 3A). Furthermore, simultaneous knockdown of both TRAIL-R1/DR4 and TRAIL-R2/DR5 significantly reduced apoptosis upon metabolic stress (Figure 2B, Supplementary Figure 3B), suggesting a role of the TRAIL system in this cell death process. Importantly, silencing TRAIL expression by siRNA did not prevent apoptosis induced by 2DG in glutamine deprived-cells (Figure 2C) indicating that a Figure 1: Increased sensitivity of p95HER2/611CTF-overexpressing cells to metabolic stress. (A) Cells were cultured for 30 h in medium containing the indicated additions and apoptosis was determined as described in Materials and Methods. Western blotting shows the expression of p-HER2 (Tyr1248) and total HER2. (B) Apoptosis in mock and p95HER2/611CTF cells incubated for 30 h either in complete or Stv medium (glutamine-free medium + 10 mM 2DG) in the presence or absence of Q-VD (20 µM). (C) p95HER2/611CTF cells were incubated in control or Stv medium either in the presence or absence of Lapatinib (5 µM), for 30 h to measure apoptosis (left panel) or for 16 h to analyze p-HER2 expression (right panel). Error bars, standard deviation (SD) from three independent experiments. ***P < 0.001, **P < 0.01, *P < 0.05. Oncotarget93691 www.impactjournals.com/oncotarget ligand-independent mechanism of apoptosis through the extrinsic pathway was activated during metabolic stress, specifically in p95HER2/611CTF cells. Caspase-8 knockdown promotes tumor growth in a xenograft model of p95HER2/611CTF- expressing cells Nutrient deprivation is a common feature of rapidly growing tumors [20]. To translate our in vitro findings into an in vivo model of primary tumor growth we first knockeddown stably caspase-8 expression in p95HER2/611CTF- overexpressing breast epithelial cells with shRNA by lentiviral infection and determined the apoptotic response of these cells to metabolic stress. As shown in Figure 3A, stable silencing of caspase-8 expression markedly reduced the sensitivity of p95HER2/611CTF cells to metabolic stress, further confirming the results obtained with siRNA oligonucleotides (Figure 2B and Supplementary Figure 3A). It has been shown that HER2/ERBB2 activation in MCF10A cells initiates a sequence of events that is characteristic of neoplastic progression in early-stage epithelial tumours [18]. In addition, expression of the p95HER2/611CTF oncogene promotes the development of aggressive and invasive mammary tumors in transgenic animals [5]. In this regard, p95HER2/611CTF-expressing MCF10A cells shows anchorage-independent growth in soft agar assays (Supplementary Figure 4A) and invasive growth in Matrigel (Supplementary Figure 4B), two hallmarks of cell transformation. To examine the role of the extrinsic apoptotic pathway in the tumoral behavior of p95HER2/611CTF-transformed cells in an in vivo setting we then injected control (Scr) or caspase-8- silenced p95HER2/611CTF cells subcutaneously into SCID-beige mice. In agreement with the data about the anchorage-independent growth of p95HER2/611CTF cells, control cells gave rise to detectable tumors within 6 weeks of injection (Figure 3B). Strikingly, we observed that caspase-8 knockdown clearly increased p95HER2/611CTF-mediated tumorigenicity. Thus, mice injected with caspase-8-silenced p95HER2/611CTF cells developed tumors with an increased volume (Figure 3B) Figure 2: Role of the extrinsic pathway of apoptosis in metabolic stress-induced cell death. (A) Mock or p95HER2/611CTF cells were cultured in starvation medium for the indicated times. Following these treatments, caspase-8 activation was assessed by western blotting. Results are representative of three independent experiments. (B) p95HER2/611CTF cells were transfected for 48 h either with a scrambled oligonucleotide (Scr) or with the indicated siRNAs targeting TRAIL-R1, TRAIL-R2, or Caspase-8. Cells were then cultured either in complete or Stv medium, or treated with soluble TRAIL (50 ng/mL) for 30 h and apoptosis was determined (left panel). Error bars, SD from three independent experiments. **P < 0.01, *P < 0.05. Protein knockdown was assesed by western blotting (right panel). (C) p95HER2/611CTF cells transfected with either a siRNA targeting TRAIL or a scrambled oligonucleotide were incubated for 30 h in complete or starvation medium and apoptosis was determined (left panel). Results show the average and range of two independent experiments. TRAIL knockdown was assessed by RT-qPCR (right panel) as described in Materials and Methods. Oncotarget93692 www.impactjournals.com/oncotarget and weight (Figure 3C) compared with the control group. These results were in line with our in vitro data on the key role of the extrinsic apoptotic pathway in the hypersensitivity of p95HER2/611CTF-transformed cells to metabolic stress. Our data suggest that prior to metabolic adaptation and tumor vessels formation, primary tumor development will be markedly dependent on nutrient availability in the tumor microenvironment which in turn controls the activity of the TRAIL-R/ caspase-8 system. Indeed, nuclei staining with Hoechst demonstrated the presence of fragmented nuclei in sections from control p95HER2/611CTF- shScr tumors that were markedly reduced in p95HER2/611CTF-shCasp8 tumors (Figure 3D). These results suggested that caspase-8 knockdown is instrumental to bypass cell death in line with the in vitro experiments. In addition, analysis of CD31 staining in tumor sections showed a significant increase in vessels formation in shCasp8 tumors as compared to shScr tumors (Figure 3D). Therefore, increased tumor growth in mice injected with p95HER2/611CTF-shCasp8 cells is likely inducing angiogenesis, thus overcoming their demand for oxygen and nutrient supply. p95HER2/611CTF expression promotes TRAIL-R2/DR5 clustering at the plasma membrane upon metabolic stress To cope with environmental or intrinsic stresses, mammalian cells induce an adaptive response through the activation of a family of eIF2α protein kinases [21]. Upon eIF2α phosphorylation global protein translation is inhibited to prevent aggravation of stress. To get further insight into the mechanism of metabolic stress-induced Figure 3: Caspase-8-dependent inhibition of tumor growth. (A) p95HER2/611CTF cells stably expressing a Scrambled or a Caspase 8-targeting shRNA were cultured either in complete or starvation medium and apoptosis was measured (left panel). Caspase-8 knockdown was determined by western blotting (right panel). Error bars, SD from three independent experiments. ***P < 0.001. (B) Subcutaneous growth of p95HER2/611CTF cells stably expressing a Scrambled or a Caspase 8-targeting shRNA injected in SCID-bBeige mice. Representative pictures were taken at day 30 after injection. (C) End-stage tumor weight in mice injected with p95HER2/611CTF cells stably expressing either a Scrambled oligonucleotide or a Caspase-8-targeting shRNA (n = 8). Graphs shown mean ± SEM. *P < 0.05. (D) Representative images and quantification of CD31+ tumor vessel area in mice injected either with p95HER2shScr or p95HER2shCasp8 cells (n = 8). The vessel area was calculated by the percentage of CD31 area per field. Oncotarget93693 www.impactjournals.com/oncotarget apoptosis in p95HER2/611CTF cells we investigated the role of the eIF2α kinases activated following ER stress or aminoacid deprivation, PERK and GCN2, respectively. Remarkably, we did not observe differences between mock and p95HER2/611CTF cells in the activation of eIF2α kinases upon metabolic stress, as determined by the phosphorylation of the eukaryotic initiation factor 2 α (eIF2α) (Supplementary Figure 5A). Furthermore, PERK silencing did not result in the inhibition of apoptosis following metabolic stress in p95HER2/611CTF cells (Supplementary Figure 5B). In contrast, PERK knockdown markedly reduced CAAT/enhancer binding protein homologous protein (CHOP) up-regulation and apoptosis induced by thapsigargin (Supplementary Figure 5B), a widely used activator of ER stress [22]. To further analyze the role of ER stress in the apoptotic response of p95HER2/611CTF cells to metabolic stress we silenced the expression of the other two UPR sensors of ER stress, Ire1α and activating transcription factor-6 (ATF6), before incubating the cells in glutamine-free medium with 2DG. As shown in Supplementary Figure 5C, silencing of either Ire1α or ATF6 did not inhibit metabolic stress-induced apoptosis in p95HER2/611CTF cells, further excluding a role of the UPR in this cell death process. The general control nonderepressible 2 (GCN2) signalling pathway is a major regulatory mechanisms for amino acid sensing that can play an adaptive role to sustain cell viability in response to glutamine deprivation [23]. However, GCN2 has been also involved in the apoptotic response of tumor cells to glutamine deprivation [24]. To investigate the role of the GCN2 pathway in metabolic stress-induced apoptosis in p95HER2/611CTF cells, GCN2 expression was silenced by siRNA interference prior to incubating the cells in starvation medium. Despite inhibiting CHOP induction upon glutamine deprivation, GCN2 knockdown failed to provide protection against starvation (Supplementary Figure 5D) thus excluding a proapoptotic role of the GCN2 pathway in p95HER2/611CTF cells subject to metabolic stress. Ligand-independent assembly of the DISC has been demonstrated in the TNF family of death receptors, most likely due to the homotypic association of receptors mediated by the pre-ligand-binding assembly domain (PLAD) [25]. Moreover, the DISC components may colocalize in an intracellular membrane fraction in breast epithelial cells in the absence of TRAIL [26]. Furthermore, it was recently shown that ER stress activates a ligandindependent TRAIL-R2/DR5 multimerization in the Golgi apparatus leading to activation of the extrinsic apoptotic pathway [15]. Therefore, we investigated if metabolic stress-induced apoptosis in p95HER2/611CTF cells involved intracellular TRAIL-R2/DR5 clustering. We first examined co-localization of TRAIL-R2/DR5 with Golgi marker GM130 in starved cells or in cells treated with ER stress inducer thapsigargin. As shown in Figure 4A, TRAIL-R2/DR5 co-localized with GM130 in p95HER2/611CTF cells treated with thapsigargin, as previously demonstrated in other cell lines [15]. However, in contrast to the ER stress inducer, metabolic stress did not induce co-localization of TRAIL-R2/DR5 with the Golgi marker in p95HER2/611CTF cells (Figure 4A). Generation of SDS-stable aggregates of TRAIL-R2/ DR5 is a requirement in the activation of caspase-8 and apoptosis following TRAIL binding to this receptor [27]. To further investigate the role of TRAIL-R2/DR5 in the mechanism underlying ligand-independent apoptosis upon metabolic stress in p95HER2/611CTF cells we first determined by gel electrophoresis formation of SDS- insoluble aggregates at different times following metabolic stress. As shown in Figure 4B, SDS-stable aggregates were readily detected upon treatment of p95HER2/611CTF cells with 2DG in glutamine-free medium. Clustering of TRAIL receptors by TRAIL into high-molecular-weight complexes leads to DISC formation where caspase-8 processing and activation takes place [27]. To determine whether metabolic stress induced ligand-independent DISC assembly at the cell surface, control or starved cells were incubated with biotin-labelled TRAIL (bio-TRAIL) at 4°C to facilitate loading of TRAIL receptors with bio- TRAIL. At low temperature bio-TRAIL does not induce DISC assembly but it should allow the isolation of preformed DISC at the cell surface. Remarkably, results shown in Figure 4C demonstrate that metabolic stress induced DISC formation in p95HER2/611CTF cells but not in mock cells. If the different DISC formation between p95HER2/611CTF and mock cells is the result of the higher expression levels of TRAIL-R2/DR5 at the cell surface in p95HER2/611CTF cells is an issue that requires further investigation. FLIP is a master regulator of metabolic stressinduced apoptosis in p95HER2/611CTF- transformed cells When examining DISC assembly we observed an important loss of FLIP proteins specifically in cell lysates from p95HER2/611CTF cells upon metabolic stress (Figure 4C, lower panel). These results prompted us to investigate the role of FLIP in the enhanced sensitivity of p95HER2/611CTF cells to metabolic stress. Time course analysis of FLIP protein expression in both mock and p95HER2/611CTF cells incubated in glutamine-free medium with 2DG showed a marked down-regulation of both FLIP(L) and FLIP(S) levels in p95HER2/611CTF cells when compared with mock cells (Figure 5A). Proteasome-mediated degradation was involved in the loss of FLIP proteins in starved p95HER2/611CTF cells as it was prevented in the presence of the proteasome inhibitor MG132 (Figure 5B, left panel) but not by a lysosomal inhibitor (Figure 5B, right panel). Importantly, over-expression of FLIP(L) protein in p95HER2/611CTF cells significantly Oncotarget93694 www.impactjournals.com/oncotarget reduced apoptosis when these cells where incubated in glutamine-free medium with 2DG (Figure 5C), which supported a central role of FLIP in modulating the response to metabolic stress. Conversely, silencing FLIP(L) expression in mock MCF10A cells markedly sensitized these cells to metabolic stress (Figure 5D), further indicating that maintenance of cellular FLIP(L) levels plays a protective role in the apoptotic response to metabolic stress in human breast epithelial cells. mTOR activity governs metabolic stress-induced down-regulation of FLIP levels and apoptosis in p95HER2/611CTF-expressing cells Because of its ability to form homodimers maintained by disulphide bonds, p95HER2/611CTF is a constitutively active form of HER2 that activates multiple signalling pathways [5]. Amongst these, stimulation of the MAPK/Erk and PI3K/Akt pathways converges in the activation of the mTORC1 complex [28]. Deregulated activation of the mTORC1 complex leads to unrestrained protein synthesis and consequently to increased sensitivity to ER stress [29]. Likewise, in cells with mTOR over-activation, the maintenance of a high rate of protein synthesis under glucose deprivation leads to ATP depletion and cell death [9]. To determine the role of mTORC1 in the differential activation of apoptosis between mock and p95HER2/611CTF cells following metabolic stress, we first examined mTORC1 activity by measuring the phosphorylation state of the mTORC1 substrates p70S6K and 4E-BP1. Interestingly, whereas levels of phosphorylated p70(S6K) and 4E-BP1 were Figure 4: Metabolic stress induces TRAIL DISC formation at the plasma membrane in p95HER2/611CTF cells. (A) Confocal microscopy analysis of TRAIL-R2 distribution in p95HER2/611CTF cells either incubated in complete or starvation medium or treated with 100 nM Thapsigargin for 21 h, in the presence of Z-VAD (20 µM). Merged images show labelling for DAPI, TRAIL-R2 and the Golgi Marker GM130. (B) p95HER2/611CTF cells were incubated either in complete or starvation medium for the indicated times, or treated with 100 nM Thapsigargin for 16 h before collecting samples to determine TRAIL-R2 levels. (C) Mock or p95HER2/611CTF cells were cultured either in complete or starvation medium, in the presence of Q-VD (20 µM) for 24 h prior to incubation on ice with bio-TRAIL (1 µg/ml) for 60 min. Unstimulated receptor controls (US) represent the addition of bio-TRAIL to an equivalent volume of lysate isolated from cells kept on ice without bio-TRAIL. DISC isolation and analysis was performed as described in Materials and Methods (upper panel). Lower panel shows western blot analysis of proteins in cell lysates. Data shown are representative of 3 independent experiments. Oncotarget93695 www.impactjournals.com/oncotarget markedly reduced in mock cells upon metabolic stress, mTORC1 activity remained elevated for up to 16 h in p95HER2/611CTF cells (Figure 6A). Similar results were obtained when AktSer473 phosphorylation was determined to assess mTORC2 activity (Figure 6A). Altogether, these results suggest that deregulated activation of both mTOR complexes as a result of p95HER2/611CTF signalling makes these activities unresponsive to metabolic stress. We next examined the role of mTOR activation in metabolic stress-induced apoptosis by using torin1, a highly potent and selective ATP-competitive mTOR inhibitor which fully inhibits mTORC1 and mTORC2 complexes [30]. As shown in Figure 6B and Supplementary Figure 1D, right panel, torin1 completely inhibited mTORC1 activities in p95HER2/611CTF cells. Furthermore, torin1 also efficiently inhibited mTORC2 activity as indicated by the complete inhibition of AktSer473 phosphorylation (Figure 6B, right panel and Supplementary Figure 1D). Strikingly, we observed a marked inhibition of starvation-induced apoptosis by torin1 (Figure 6B and Supplementary Figure 1D, left panels), suggesting an important role of mTOR in this cell death process. Interestingly, results shown in Figure 6B demonstrate that FLIP(L) and FLIP(S) down-regulation occurs in a mTOR-dependent manner as inhibition of mTOR by torin1 significantly restored FLIP levels in p95HER2/611CTF cells. The role of mTORC1 and mTORC2 activities in the sensitivity of p95HER2/611CTF-transformed cells to metabolic stress was further studied in experiments silencing Raptor or Rictor expression with a shRNA lentiviral vector. As shown in Figure 6C, Raptor Figure 5: FLIP levels control cell death induced by metabolic stress. (A) FLIP levels in mock and p95HER2/611CTF cells cultured under starvation conditions. (B) FLIP levels in cells were cultured either in complete (NT) or Starvation medium for 16 h, in the presence or absence of 10 µM MG132 (left panel) or 400 nM Bafilomycin A (right panel). Results are representative of three independent experiments. (C) Apoptosis was assessed in FLIPL–overexpressing p95HER2/611CTF cells cultured in Stv or treated with TRAIL (50 ng/mL) for 30 h. Insert shows FLIPL overexpression as measured by western-blotting. (D) Apoptosis in mock cells transfected for 16 h either with a Scrambled oligonucleotide or siRNAs targeting FLIPL, Caspase-8, or both and then cultured in complete or Stv medium for 6 h (upper panel). Protein knockdown was determined 16 h after transfection by western blotting (lower panel). Error bars, SD from three independent experiments. ***P < 0.001, **P < 0.01, *P < 0.05. Oncotarget93696 www.impactjournals.com/oncotarget knockdown partially reduced the sensitivity of cells to glycolysis inhibition in glutamine-deprived medium, similarly to what is observed in cultures treated with rapamycin. Interestingly, silencing Rictor expression markedly inhibited FLIP down-regulation and apoptosis upon metabolic stress (Figure 6C), almost like the inhibition observed with Torin-1 (Figure 6B and 6C). Although both mTOR complexes seem to counteract the activation of apoptosis after metabolic stress, our results underscore the prominent role of mTORC2 activation as an effector of the signalling pathway leading to FLIP downregulation and apoptosis activation in p95HER2/611CTF- transformed cells undergoing metabolic stress. DISCUSSION Uncontrolled cell proliferation in tumors results in the development of aberrant vasculature in which the delivery of nutrients is drastically reduced leading to microenvironmental stress. To cope with this extrinsic form of stress, tumor cells must activate different adaptive responses [6]. However, different observations indicate that prior to adaptation oncogene-transformed cells may become very sensitive to nutrient deprivation and this represent a potential Achilles’heel in tumor progression [31–33]. Although MCF10A cells express markers not observed in mammary gland tissue [34] and show genetic abnormalities [35], they recapitulate many features of normal breast epithelium when grown in three-dimensional (3D) cultures, including the formation of growth-arrested polarized acini with a hollow lumen and basal deposition of basement membrane components [17]. Moreover, MCF10A cells and isogenic progression series have recently been used to identify driver genes associated with tumor progression from preinvasive to invasive cellular phenotypes as seen in primary breast cancer [19]. Interestingly, it has been demonstrated that Her2 activation in MCF10A cells initiates a sequence of events that is characteristic of neoplastic progression Figure 6: Role of mTOR complexes in apoptosis induced by metabolic stress. (A) Cells were cultured in Stv medium for the indicated times. Following these treatments, mTORC1 and mTORC2 activation (p-p70/S6K, p-4EBP1 and p-AktSer473), was assessed by western blotting. Results are representative of 4 independent experiments. (B) Left panel shows apoptosis in p95HER2/611CTF cells cultured in complete or stv medium in the presence or absence of Torin-1 (250 nM). FLIP levels and mTORC1/mTORC2 activities were determined by western blot analysis (right panels). (C) Apoptosis in p95HER2/611CTF cells stably expressing a Scrambled, Raptor or Rictor shRNA cultured either in complete or stv medium (left panel). Where indicated, Rapamycin (5 µM) or Torin (250 nM) were added to the medium. Levels of FLIP, Raptor and Rictor proteins upon 16 h of treatment are shown in the right panel. Error bars, SD from three independent experiments. ***P < 0.001, **P < 0.01, *P < 0.05. (D) Schematic overview of the differential response to metabolic stress in control and p95HER2/611CTF cells. Oncotarget93703 www.impactjournals.com/oncotarget 48. Bar-Peled L, Sabatini DM. Regulation of mTORC1 by amino acids. Trends Cell Biol. 2014; 24:400–406. 49. Moloughney JG, Kim PK, Vega-Cotto NM, Wu CC, Zhang S, Adlam M, Lynch T, Chou PC, Rabinowitz JD, Werlen G, Jacinto E. mTORC2 Responds to Glutamine Catabolite Levels to Modulate the Hexosamine Biosynthesis Enzyme GFAT1. Mol Cell. 2016; 63:811–826. 50. Facchinetti V, Ouyang W, Wei H, Soto N, Lazorchak A, Gould C, Lowry C, Newton AC, Mao Y, Miao RQ, Sessa WC, Qin J, Zhang P, et al. The mammalian target of rapamycin complex 2 controls folding and stability of Akt and protein kinase C. EMBO J. 2008; 27:1932–1943. 51. Kim SJ, DeStefano MA, Oh WJ, Wu CC, Vega-Cotto NM, Finlan M, Liu D, Su B, Jacinto E. mTOR complex 2 regulates proper turnover of insulin receptor substrate-1 via the ubiquitin ligase subunit Fbw8. Mol Cell. 2012; 48:875–887. 52. Harper N, Farrow SN, Kaptein A, Cohen GM, MacFarlane M. Modulation of tumor necrosis factor apoptosis-inducing ligandinduced NF-kappa B activation by inhibition of apical caspases. J Biol Chem. 2001; 276:34743–34752. 53. Yerbes R, Lopez-Rivas A, Reginato MJ, Palacios C. Control of FLIP(L) expression and TRAIL resistance by the extracellular signal-regulated kinase1/2 pathway in breast epithelial cells. Cell Death Differ. 2012; 19:1908–1916. 54. Hughes MA, Langlais C, Cain K, MacFarlane M. Isolation, characterisation and reconstitution of cell death signalling complexes. Methods. 2013; 61:98–104. 55. Wiznerowicz M, Trono D. Conditional suppression of cellular genes: lentivirus vector-mediated drug-inducible RNA interference. J Virol. 2003; 77:8957–8961. 56. Muthuswamy SK, Li D, Lelievre S, Bissell MJ, Brugge JS. ErbB2, but not ErbB1, reinitiates proliferation and induces luminal repopulation in epithelial acini. Nat Cell Biol. 2001; 3:785–792. www.impactjournals.com/oncotarget/ Oncotarget Supplementary Materials Oncogenic p95HER2/611CTF primes human breast epithelial cells for metabolic stress-induced down-regulation of FLIP and activation of TRAIL-R/Caspase-8-dependent apoptosis SUPPLEMENTARY MATERIALS Supplementary Figure 1: Apoptotic response to starvation in 184A1/p95HER2 cells. (A) MCF10A/p95HER2 (left panel) or 184A1/p95HER2 cells (middle panel), were cultured in starvation medium for 48 or 72 hours, respectively, and apoptosis was determined as described in Materials and Methods. Error bars, SD from four independent experiments. ***P < 0.001 (MCF10A/p95HER2). Results also show the average and range of two independent experiments (184A1/p95HER2). Western blotting shows p95HER2 expression in both cell lines (right panel). (B) Mock/184A1 and p95HER2/184A1 clones were cultured in starvation medium for 48 hours and apoptosis was determined. Error bars, SD from three independent experiments. **P < 0.01; *P < 0.05. Western blotting shows the expression of total HER2 and p-HER2 (Tyr1248) in 184A1 clones. (C) 184A1/p95HER2 (Clone 5) cells were treated with or without Lapatinib (5 µmol/L) for 16 hours before culturing them in control or starvation (Stv) medium for 48 hours to determine apoptosis. Error bars, SD from four independent experiments. **P < 0.01. Western blots show the expression of total HER2 and p-HER2 (Tyr1248) in 184A1/p95HER2 (clone 5) cells treated with or without Lapatinib (5 µmol/L) for 16 hours. (D) 184A1/p95HER2 (clone 5) cells were cultured under control or starvation (Stv) conditions in the presence or absence of Torin (250 nmol/L) for 48 hours before measuring apoptosis. Error bars, SD from four independent experiments. *P < 0.05. Western blots show the inhibition of mTORC1 and mTORC2 activities in 184A1/p95HER2 (clone 5) cells treated with or without Torin (250 nmol/L) for 16 hours. Results are representative of three independent experiments. Supplementary Figure 2: TRAIL-R2/DR5 levels in response to starvation. (A) (Left panel) Cell surface expression of TRAIL-R2/DR5 was measured by flow cytometry with TRAIL-R2-PE antibody as described in the Materials and Methods section. Cells incubated with control IgG-PE antibody alone, were used as a control for background fluorescence of cells (dashed line). Data shown are representative of three independent experiments. (Right panel) MCF10A mock and p95Her2 cells were treated with the indicated doses of TRAIL for 24 h. Apoptosis was determined by subG1 analysis as previously described. Error bars represent SD from three independent experiments. **P < 0,01. (B) MCF10A mock or p95Her2 cells were cultured in starvation medium for the indicated times. Following this incubation, p95-Her2, TRAIL-R2, pro-Caspase 8 or FADD levels were assessed by Western blotting analysis. Results are representative of three independent experiments. (C) p95Her2 cells were incubated either in complete or starvation medium, or treated with 100 nM thapsigargin, for 20 hours. TRAIL-R2 surface expression was assessed by flow cytometry as in (A). Data shown are representative of 2 independent experiments. Supplementary Figure 3: Role of the TRAIL-R/Caspase-8 apoptotic pathway in metabolic stress-induced cell death. p95HER2/611CTF cells were transfected for 48 h either with a scrambled oligonucleotide (Scr) or with the indicated siRNAs targeting Caspase-8 (A), TRAIL-R1 or TRAIL-R2 (B). Cells were then cultured either in complete or Stv medium for 30 h and apoptosis was determined. Error bars, SD from three independent experiments. ***P < 0.001; **P < 0.01; *P < 0.05. Protein knockdown was assessed by western blotting. Supplementary Figure 4: Tumorigenic potential of p95HER2/611CTF-transformed cells. (A) Tumorigenic potential of mock and p95HER2/611CTF cells was determined with a soft agar colony formation assay as described in the Material and Methods section. Left panel show a picture of visible colonies after staining with crystal violet. Middle and right panels show representative images and quantification of mock and p95HER2/611CTF colonies obtained by phase contrast microscopy (10×). Data shown are representative of three independent experiments. (B) Phase contrast microscopy images of 3D cultures of mock, p95HER2/611CTF and MDA-MB231 cells growing in Matrigel for 18 days. Images shown are representative of two independent experiments. Supplementary Figure 5: Role of stress pathways in starvation-induced apoptosis. (A) Mock and p95HER2/611CTF cells were cultured in starvation medium for the indicated times. Following these incubations, eIF2α phosphorylation was assessed by Western blotting. Results are representative of two independent experiments. p95HER2/611CTF cells were transfected either with a scrambled (Scr) oligonucleotide or siRNAs targeting PERK (B), Ire1α or ATF6 (C) or GCN2 (D) for 24 h. Cells were then incubated either in complete or starvation medium, or treated with 100 nM thapsigargin, for 30 h. Apoptosis was measured as previously described. Error bars, SD from three independent experiments (B) or average and range of 2 independent experiments (C and D). ***P < 0.001; *P < 0.05. Protein knockdown was determined after 16 h of treatment by Western blotting. ATF6 knockdown was assessed by RT-qPCR as described in Materials and Methods.