Estrogen-Induced Apoptosis in Breast Cancers Is Phenocopied By

Published OnlineFirst January 17, 2019; DOI: 10.1158/1541-7786.MCR-18-0481

Cell Fate Decisions Molecular Cancer Research Estrogen-Induced Apoptosis in Breast Cancers Is Phenocopied by Blocking Dephosphorylation of Eukaryotic Initiation Factor 2 Alpha (eIF2a) Protein Surojeet Sengupta1, Catherine M. Sevigny1, Poulomi Bhattacharya2, V. Craig Jordan2, and Robert Clarke1

Abstract

Approximately 30% of aromatase-inhibitor–resistant, tion factor 4 and C/EBP homologous protein that facilitate estrogen receptor–positive patients with breast cancer ben- apoptosis. Notably, we recapitulated this phenotype of efit from treatment with estrogen. This enigmatic estrogen MCF7:5C in two other endocrine therapy–resistant cell lines action is not well understood and how it occurs remains (MCF7/LCC9 and T47D:A18/4-OHT) by increasing the elusive. Studies indicate that the unfolded protein response levels of phospho-eIF2a using salubrinal to pharmacolog- and apoptosis pathways play important roles in mediating ically inhibit the enzymes responsible for dephosphoryla- estrogen-triggered apoptosis. Using MCF7:5C cells, which tion of eIF2a, GADD34, and CReP. RNAi-mediated ablation mimic aromatase inhibitor resistance, and are hypersensi- of these genes induced apoptosis that used the same sig- tive to estrogen as evident by induction of apoptosis, we naling as salubrinal treatment. Moreover, combining define increased global protein translational load as the 4-hydroxy tamoxifen with salubrinal enhanced apoptotic trigger for estrogen-induced apoptosis. The protein kinase potency. RNA-like endoplasmic reticulum kinase pathway was activated followed by increased phosphorylation of eukaryotic Implications: These results not only elucidate the mechanism initiation factor-2 alpha (eIF2a). These actions block global of estrogen-induced apoptosis but also identify a drugable protein translation but preferentially allow high expression target for potential therapeutic intervention that can mimic the of specific transcription factors, such as activating transcrip- beneficial effect of estrogen in some breast cancers.

Introduction resistant breast cancer cells and xenografts (7–12). Using long- term estrogen-deprived (LTED) cells, such as MCF7:5C (7) and fi The bene cial effects of high-dose estrogen therapy in some MCF7/LTED (8), studies have confirmed that low doses of estro- breast cancers, the standard therapy for postmenopausal patients gen can induce intrinsic and extrinsic apoptotic activities (7, 8). with breast cancer prior to the introduction of tamoxifen, have MCF7:5C cells were derived from MCF7 cells by long-term culture – been widely overlooked until relatively recently (1 3). Recent in estrogen-free conditions (7, 13). Stress responses were activated clinical trials have reported a 30% to 50% clinical response rate during estrogen-induced apoptosis in MCF7:5C cells (14) and after estrogen therapy for heavily pretreated estrogen receptor contrasted with apoptosis induced by cytotoxic drugs such as – (ER) positive breast cancers that become resistant to aromatase paclitaxel (15). Therefore, it is important to determine the precise – fi inhibitors (AI; refs. 4 6). Comparable clinical bene t was also mechanism that induces apoptosis by estrogen in LTED breast – observed with lower doses of estrogen in AI resistant breast cancer. cancers (4). In MCF7:5C cells, studies suggest that ER and its classical in vitro in vivo Laboratory studies and have recapitulated the nuclear function is necessary to induce apoptosis (16). Planar beneficial effects of estrogen treatment in endocrine therapy– class I estrogens such as estradiol (E2) induce apoptosis within 48 hours of treatment (15). Class II estrogens like bisphenol initially act as an antagonist during the first week of treatment (17) 1Department of Oncology, Georgetown-Lombardi Comprehensive Cancer but induce apoptosis later because of a delay in the apoptotic 2 Center, Georgetown University Medical Center, Washington D.C. Breast trigger (18). Medical Oncology, The University of Texas MD Anderson Cancer Center, Unfolded protein response (UPR) and stress signaling play a Houston, Texas. critical role in endocrine-resistant breast cancers (19–22). Global Note: Supplementary data for this article are available at Molecular Cancer analysis of gene expression after E2 treatment in MCF7:5C cells Research Online (http://mcr.aacrjournals.org/). demonstrate that UPR, which follows an endoplasmic reticulum Corresponding Author: Surojeet Sengupta, Georgetown University, 3970 (EnR) stress, is involved in estrogen-induced apoptosis (14). UPR Reservoir Rd NW, Research Bldg. W 405B, Washington, DC 20057. Phone: 202- has three distinct adaptive pathways that are primarily cytopro- 687-7451; E-mail: [email protected] tective. These pathways are highly coordinated and act to mitigate doi: 10.1158/1541-7786.MCR-18-0481 the protein load using three sensors, namely inositol-requiring 2019 American Association for Cancer Research. protein 1 alpha (IRE1-a), activating transcription factor 6 (ATF6),

918 Mol Cancer Res; 17(4) April 2019

UPR and Estrogen-Induced Apoptosis

and protein kinase RNA-like endoplasmic reticulum kinase change in expression of transcripts was determined as described (PERK; ref. 23). However, failure to restore protein synthesis previously and used the ribosomal protein 36B4 mRNA as the homeostasis following prolonged EnR stress can lead to an internal control (33). Sequences of the primers used for ATF4, apoptotic cell death (24). Activation of the PERK pathway XBP1 (spliced), CHOP, PPP1R15A (GADD34), PPP1R15B increases phosphorylation of the serine-51 residue in eukaryotic (CReP), IRE1a and 36B4 genes can be found in Supplementary initiation factor 2 alpha (eIF2a) protein that inhibits global Table S1. protein synthesis and helps to attenuate the protein load within the EnR (23). However, sustained phospho-eIF2a–mediated Puromycin labelling translational repression can also initiate cell death (25). Dephos- One million MCF7:5C cells were plated in 10-cm diameter phorylation of eIF2a is catalyzed by two specific regulatory plates and treated with vehicle (0.1% ethanol) or 10 nmol/L subunits of protein phosphatase 1 (PP1) complex, known as 17b-estradiol for varying time. Puromycin labeling was per- GADD34 (growth arrest and DNA damage 34) and CReP (con- formed by adding 10 mg/mL puromycin for the last 10 minutes stitutive repressor of eIF2a phosphorylation; refs. 26–28). before harvest to assess global protein translation (34). Cells were Although GADD34 expression is induced by stress (29), CReP lysed using 1 cell lysis buffer (Cell Signaling Technology) is constitutively expressed and maintains a balance in phospho- containing protease inhibitors (Roche Diagnostics) and phospha- eIF2a levels (28). We have investigated the role of downstream tase inhibitors I and II (EMD Chemicals Inc.). Whole-cell proteins targets of the PERK pathway in the context of E2-induced apo- were isolated and probed for total puromycinylated protein ptosis and examined whether this apoptosis response is replicated by Western blotting analysis. Western blot images were scanned by activating eIF2a signaling. and quantified using Image J software. The blot was stained with ponceau S prior to probing with antipuromycin to normalize Materials and Methods the puromycinylated proteins for each sample lane. Cell culture and reagents siRNA Cell culture media was purchased from Invitrogen Inc. and FCS Two specific ON-Target plus siRNAs were purchased for each was obtained from HyClone Laboratories. Breast cancer cells gene from Dharmacon Inc. targeting PPP1R15A (catalog no. MCF-7:5C (7, 30) and MCF7/LCC9 (31) were derived from MCF7 J004442-05 and J00442-08) and PPP1R15B (catalog no. cells. T47D:A18/4-OHT cells are derived from T47D:A18 cells, by J015013-05 and J015013-07). These siRNAs were used to long-term culture in presence of 4-hydroxy tamoxifen (32). Cells deplete protein levels of GADD34 and CReP protein were maintained in phenol red–free IMEM media supplemented levels, respectively. A nontargeting siRNA from Dharmacon with 10% charcoal dextran-treated FCS. The identities of all the Inc. (catalog no. D-001810-10) was used as a control. Trans- cell lines were authenticated using short tandem repeat profiling fection of siRNA was performed using Dharmafect I Reagent and cells were regularly tested for Mycoplasma contamination (Dharmacon Inc.), and the cells were harvested after 48 hours before and after completion of experiments. Salubrinal (catalog of transfection for protein extraction. no. 2347) was purchased from Tocris Bioscience. Puromycin was purchased from Thermo Fisher Scientific (A1113803). All the Western blotting analysis experiments were performed at least three times, in triplicate to Total proteins from whole cells were extracted using RIPA confirm the results. buffer containing protease inhibitors (Roche Diagnostics) and phosphatase inhibitors I and II (EMD Chemicals Inc.). Total Crystal violet cell density assay protein (15–25 mg) was run on the gels and transferred onto Cell growth assays were performed using a crystal violet assay. A nitrocellulose membranes. Membranes were subsequently total of 1.5 104 cells were seeded in each well of a 24-well plate blocked with 5% nonfat dry milk in TBS and probed with primary and specific treatments were started after 24 hours; media was and secondary antibodies. Specific bands were visualized using changed every 48 hours. Cell growth was assessed after 5 or 6 days chemiluminescence (Thermo Fisher Scientific). Details of the of treatment by washing the cells once with 1X PBS and subse- antibodies used are given in Supplementary Table S2. quently stained with crystal violet followed by washing with deionized water. Cells stained with crystal violet were permeabi- Statistical analysis lized by citrate buffer. Absorbance was measured at 562 nm using Statistical significance was estimated using two-tailed, un- a spectrophotometer. paired Student t test for pairwise comparison wherever relevant. For multiple comparisons, one-way ANOVA with two-tailed RNA isolation and real time PCR post hoc Dunnett test was performed. A P value of <0.05 was Total RNA was isolated using TRIzol reagent (Invitrogen) and considered as statistically significant. RNAeasy Kit (Qiagen) according to the manufacturer's instruc- tions. RT-PCR was performed as described previously (33). Brief- ly, cDNA was generated from 1 mg of total RNA in a total volume of Results 20 mL using high-capacity cDNA Reverse Transcription Kits Estrogen induces global protein synthesis and activates the (Applied Biosystems). Subsequently, the cDNA was diluted to PERK pathway of UPR in MCF7:5C cells 500 mL and RT-PCR was performed using ABI QuantStudio 12K We investigated the time-dependent effects of E2 on global Flex Real-Time PCR System (Applied Biosystems). Each well protein synthesis in MCF7:5C cells. As evident by the incorpo- contained 10 mL of PowerUp SYBR Green PCR Master Mix ration of puromycin during protein synthesis, a significant (Applied Biosystems), 500 nmol/L each of forward and reverse increase (P < 0.01 at 24 hours vs. vehicle) in the global protein primers, and 5 mL of diluted cDNA in 20 mL final volume. The synthesis rate was observed until 48 hours of E2 treatment.

www.aacrjournals.org Mol Cancer Res; 17(4) April 2019 919

Sengupta et al.

However, global protein synthesis was suppressed by 40% versus of eIF2a (35, 36) and was elevated at the early time points vehicle at 72 hours (P ¼ 0.01) and 96 hours (P < 0.01) post-E2 following initiation of E2 treatment (Fig. 1B). In contrast, a treatment (Fig. 1A). Phosphorylated PERK and phosphorylated signiﬁcant increase (P < 0.05) in ATF4 mRNA was detected only eIF2a expression levels were elevated after E2 treatment. The after 48–96 hours of E2 treatment (Fig. 1C). Conversely, proa- increase in phospho-eIF2a expression at 48 hours preceded the poptotic CHOP mRNA was increased 12-fold (P < 0.01) after attenuated rate of protein synthesis seen at 72 and 96 hours. ATF4 48 hours of E2 treatment; expression remained elevated until protein is preferentially translated following phosphorylation 96 hours (P < 0.01; Fig. 1D). CHOP protein levels were

Figure 1.

E2-induced apoptosis of MCF7:5C cells is triggered by high protein translation and activation of PERK pathway. A, A representative Western blot of puromycin- labeled total proteins after indicated time of estrogen (17b-estradiol; E2, 10 nmol/L) treatment. The graph on the right panel shows the percent of puromycin- labelled proteins (average SD, n ¼ 3) versus vehicle-treated cells. The intensity of each lane was normalized by the ponceau S stained lanes. B, Assessment of PARP cleavage, PERK activation, and its downstream effectors (phospho-eIF2a, ATF4, and CHOP) by Western blotting analysis. C and D, Transcripts levels of ATF4 (C), and CHOP (D) were assessed using RT-PCR after indicated time of estrogen (10 nmol/L) treatment (, P < 0.01 versus vehicle treatment).

920 Mol Cancer Res; 17(4) April 2019 Molecular Cancer Research

UPR and Estrogen-Induced Apoptosis

substantially upregulated after 72 hours, corresponding with the resistant to both 4-hydroxy tamoxifen and fulvestrant (Supple- cleavage of PARP, an indicator of apoptosis (Fig. 1B). We also mentary Fig. S2A and S2B) and are insensitive to E2,wealso measured the mRNA levels of other UPR sensors, and found observed a concentration-dependent reduction in cell number that E2 treatment upregulated IRE1a (P < 0.01 at 48 and 72 hours) with salubrinal treatment (Fig. 3B and C). Phospho-eIF2a and the ratio of spliced XBP1/total XBP1 levels (P < 0.01 at levels were elevated after 48 hours of salubrinal treatment 48 hours) in a time-dependent manner (Supplementary in all cell lines, as were ATF4 and CHOP expression and Fig. S1A and S1B). ATF6 transcripts were significantly elevated PARP cleavage (Fig. 3D–F). Salubrinal treatment increased (P < 0.01) at 48 and 72 hours after E2 treatment but declined at Annexin V-positive and PI–positive cells in MCF7:5C (Supple- 96 hours posttreatment (Supplementary Fig. S1C). mentary Fig. S3A and S3B), LCC9 cells (Supplementary Fig. S3C and S3D), and T47D:A18/4-OHT (Supplementary Fig. S3E 17b-estradiol upregulates GADD34 mRNA but the protein level and S3F). is unaltered PERK-mediated phosphorylation of eIF2a occurs during cell E2 and salubrinal similarly regulate UPR-related genes stress and upregulates GADD34 expression (29, 37, 38). E2 treatment in MCF7:5C cells induces several UPR GADD34 protein functions as a negative feedback loop by genes (14). We studied the induction of some key UPR genes dephosphorylating eIF2a and promoting recovery from UPR after 24 and 48 hours of estrogen and salubrinal treatment. by derepressing protein translation (29, 39). GADD34 expres- A similar pattern of gene regulation was seen with E2 and sion is induced transcriptionally by ATF4 (38) and by enhanced salubrinal treatment. The magnitude of induction by estrogen translation of GADD34 mRNA. The latter happens via an and salubrinal was similar for ATF4 (P < 0.05 at 48 hours of 0 alternate 5 upstream open reading frame (uORF) that is E2 and P < 0.01 at 48 hours of salubrinal), PPP1R15A (P < 0.05 inhibited during unstressed conditions (37, 40). We measured at 48 hours of E2 and P < 0.005at48hoursofsalubrinal), the time-dependent regulation of GADD34 and CReP mRNA IRE1a (P < 0.05 at 48 hours of E2 and P < 0.005 at 48 hours and protein levels following estrogen treatment. GADD34 of salubrinal), and spliced-XBP1 (P < 0.05 at both 48 hours of mRNA expression was significantly upregulated after 48 hours E2 and salubrinal; Fig. 4A–D). Induction of CHOP mRNA was of E2 treatment (P < 0.01) and continued to increase with markedly higher (Fig. 4E) after salubrinal treatment (P < 0.01 at treatment time (72 and 96 hours; P < 0.01). In marked contrast, 24 and 48 hours) compared with E2 (P < 0.05 at 48 hours) GADD34 protein levels remained unaltered at all the time treatment. Levels of PPP1R15B, PERK, and total XBP1 did not points evaluated (Fig. 2A and B). No change in the mRNA or change significantly after E2 or salubrinal treatment (Supple- protein levels of CReP was observed after estrogen treatment mentaryFig.S4A,S4B,andS4C). (Fig. 2A and B).

E2 and salubrinal activate the mitochondrial apoptotic pathway Salubrinal induces apoptosis in MCF7:5C, LCC9, and We further investigated how E2 and salubrinal induced T47D:A18/4-OHT cells apoptosis. Previous studies have reported that E2-induced apo- E2-induced apoptosis in MCF7:5C cells was preceded by an ptosis is mediated by the intrinsic mitochondrial pathway, increased phosphorylation of eIF2a.GADD34andCReP,the since this effect is dependent on increased expression of the regulatory subunits of protein phosphatase complex, are proapoptotic proteins Bim and Bax (7). We conﬁrmed that responsible for eIF2a dephosphorylation. Therefore, we tested both Bim and Bax protein levels were elevated 48–72 hours whether inhibition of GADD34 and CReP can increase phos- after estrogen treatment (Fig. 5A). We next determined whether pho-eIF2a levels and induce apoptosis in the absence of estro- salubrinal-induced apoptosis in MCF7:5C and LCC9 cells uses gen. Salubrinal, a pharmacologic inhibitor of both GADD34 the same mechanism. Indeed, BIM and BAX protein levels were and CReP (41), produced a concentration-dependent decrease elevated in a concentration-dependent manner after 48 hours in cell number in MCF7:5C cells (Fig. 3A) over a 6-day period. of salubrinal treatment (Fig. 5B). We also measured levels of In LCC9 (27) and T47D:A18/4-OHT cells (32), which are the proapoptotic protein BAK and antiapoptotic BCL2 and

MCF7:5C Figure 2. A B Estrogen induces GADD34 mRNA 25 ** but not protein in MCF7:5C cells. A, CReP E2 (hrs) Levels of GADD34 (protein product 20 Veh 4244872 96 ppp1R15a gene) and CReP (protein GADD34 product ppp1R15b gene) mRNA 15 GADD34 were measured by RT-PCR after indicated time of E2 (17b-estradiol, P < 10 nmol/L) treatment ( , 0.01 10 ** CReP vs. vehicle treatment). B, Levels of ** GADD34 and CReP protein were 5 β-Actin assessed by Western blotting Fold change vs. Veh Fold change vs. analysis after indicated time of E2 (17b-estradiol, 10 nmol/L) 0 treatment. b-Actin was measured Veh 4 24 48 72 96 to conﬁrm equal protein loading. Hours after E2 treatment

www.aacrjournals.org Mol Cancer Res; 17(4) April 2019 921

Sengupta et al.

Figure 3. Salubrinal, concentration-dependently, decreases cell number and induces apoptosis in MCF7:5C, LCC9, and T47D:A18/4-OHT cells by elevating phosho-eIF2a. Cell number was measured after a 6-day treatment with indicated concentrations of salubrinal in MCF7:5C (A), LCC9 (B), and T47D: A18/4-OHT cells (C). Crystal violet method was used to assess the cell numbers and the data are represented as percent cell number of vehicle treatment. Level of PARP cleavage, phospho-eIF2a, ATF4, and CHOP was assessed by Western blotting analysis in MCF7:5C (D), LCC9 (E), T47D: A18/4-OHT cells (F), after 48 hours of treatment with vehicle or indicated concentrations of salubrinal. b-Actin was used to conﬁrm equal protein loading.

BCLXL proteins. No change in expression was observed after E2 MCF7:5C and LCC9 cells. Treatment with 1 mmol/L 4OHT, in treatment in MCF7:5C cells or after salubrinal treatment in combination with suboptimal concentrations of salubrinal MCF7:5C and LCC9 cells (Fig. 5A and B). (1.6 mmol/L and 3.1 mmol/L), signiﬁcantly (P < 0.05) potentiated salubrinal-induced cell death in both MCF7:5C and Depletion of individual regulatory subunits, GADD34 and LCC9 cells (Fig. 7A and B). This effect was also evident when CReP, induces apoptosis in MCF7:5C cells 1 mmol/L fulvestrant was used in combination with suboptimal The regulatory subunits GADD34 and CReP are responsible concentrations of salubrinal (Supplementary Fig. S5A and for dephosphorylation of eIF2a. To determine whether inhibi- S5B). These drug combinations did not produce any noticeable tion of both GADD34 and CReP is necessary for increased differences in the levels of phospho-eIF2 alpha, ATF4, or phospho-eIF2a levels and apoptosis, we used siRNAs to deplete cleaved PARP; a moderate increase in CHOP protein expression their expression in MCF7:5C cells. Two different siRNA sequen- was observed in the combination treated cells compared with ces were used for each GADD34 and CReP. Depletion of either salubrinal alone (Fig. 7C). GADD34 or CReP led to PARP cleavage in MCF7:5C cells. Notably, CReP levels increased when GADD34 was depleted. The increase in phospho-eIF2a, ATF4, and CHOP was greater in Discussion CReP-depleted cells than in GADD34-depleted cells (Fig. 6). This study establishes activated PERK-mediated phosphorylation of eIF2a as a key regulator of estrogen-induced apoptosis Effect of salubrinal is potentiated by 4-hydroxy tamoxifen in susceptible breast cancer cells. This apoptotic response is cotreatment phenocopied by inhibiting dephosphorylation of eIF2a in E2-induced apoptosis is completely blocked by 4OHT in breast cancer cells with acquired resistance to estrogen and MCF7:5C cells (7, 18). We tested whether 4OHT has any antiestrogen (Fig. 8). The PERK pathway of UPR activates modulatory effect on salubrinal-induced apoptosis in downstream signaling cascades that can restore proteostasis

922 Mol Cancer Res; 17(4) April 2019 Molecular Cancer Research

UPR and Estrogen-Induced Apoptosis

A B C ATF4 PPP1R15A ** 4.5 IRE1α 2.5 ** 2.5 ** (GADD34) 4.0 24 hrs 2.0 * 2.0 24 hrs 3.5 24 hrs * 48 hrs * * 3.0 * 48 hrs * 48 hrs 1.5 1.5 2.5 * 2.0 1.0 1.0 1.5 1.0 0.5 0.5 Veh vs. Fold change Fold change vs. Veh vs. Fold change 0.5 Fold change vs. Veh vs. Fold change 0.0 0.0 0.0 Veh E2 Salubrinal Veh E2 Salubrinal Veh E2 Salubrinal

D 4.5 E XBP1(s) 25 CHOP 4.0 * ** 3.5 24 hrs 20 24 hrs 3.0 48 hrs * * 48 hrs 15 2.5 * 2.0 ** 10 1.5 1.0 5

Fold change vs. Veh vs. Fold change * 0.5 Veh vs. Fold change 0.0 0 Veh E2 Salubrinal Veh E2 Salubrinal

Figure 4. Similar upregulation of UPR genes by estrogen and salubrinal in MCF7:5C cells. Regulation of UPR genes ATF4 (A), PPP1R15A (B), IRE1a (C), XBP1 (spliced; D), and CHOP (E) was measured using RT-PCR after 24 and 48 hours of treatment with vehicle (Veh), 10 nmol/L estrogen (E2), or 25 mmol/L salubrinal (, P < 0.05; , P < 0.01 versus respective vehicle treatment). and may function as a prosurvival signal (23). However, tion of eIF2a is critical for estrogen-induced apoptosis as prolonged and unmitigated activation of the PERK pathway pharmacologic inhibition of PERK blocked estrogen-induced can induce apoptosis (24, 42–44). A previous study, using eIF2a phosphorylation and apoptosis (14). In addition, we MCF7:5C cells, conﬁrmed that PERK-mediated phosphoryla- reported that estrogen-induced apoptosis is a delayed process

A MCF7:5C B MCF7:5C LCC9 Salubrinal (mmol/L) Salubrinal (mmol/L) E2 (10 nmol/L) Hours Veh 42448 72 96 Veh 6.25 12.5 25 Veh 6.25 12.5 25 Bax Bax

Bim Bim

Bak Bak

Bcl2 Bcl2

BclxL BclxL

β-Actin β-Actin

Figure 5. Similar regulation of intrinsic (mitochondrial) apoptotic markers by estrogen and salubrinal in MCF7:5C cells. Protein levels of proapoptotic markers

(Bim, Bak, and Bax) and antiapoptotic markers (Bcl2 and Bclxl) were measured after indicated time of 10 nmol/L estrogen (E2) treatment in MCF7:5C cells (A), and after indicated concentration of salubrinal treatment for 48 hours in MCF7:5C and LCC9 cells (B).

www.aacrjournals.org Mol Cancer Res; 17(4) April 2019 923

Sengupta et al.

GADD34 protein levels (40, 47) and the recovery of repression of protein synthesis (47). In MCF7:5C cells, it is likely that this mechanism is compromised during E2 treatment resulting in loss of this feedback mechanism (Fig. 8). Further studies are warranted to understand this mechanism precisely. E2-induced apoptosis is blocked by inhibition of the c-src activity in MCF7:5C cells (16). Multiple cellular functions are regulated by c-src (48) including estrogen-induced growth in MCF7 cells (49). Src inhibition leads to suppressed AKT phosphorylation in MCF7:5C cells (50). Because activated AKT influ- ences E2-induced ER binding at the genome level (51), inhibition of c-src could regulate E2-induced transcription, reduce protein load, and fail to reach the threshold required to trigger sustained UPR and apoptosis in E2-treated MCF7:5C cells. Conversely, E2 and triphenylethylenes show a differential delay in triggering apoptosis (18). For example, bisphenol induces apoptosis in MCF7:5C cells on day 5 as compared with day 2 with E2 treatment (17, 18). We also showed that bisphenol is not as effective in recruiting either ER or its coactivators to the promoter of the estrogen-responsive gene TFF1. Moreover, the levels of TFF1 mRNA were markedly lower as compared with E2 treatment (17). Because of its diminished estrogenic property, bisphenol may take longer to accumulate the threshold levels of unfolded proteins required for PERK activation leading to a delayed trigger and apoptosis in MCF7:5C cells. Elevated levels of phosphorylated-eIF2a preceded E2- induced apoptosis in MCF7:5C cells. Hence, we examined whether inhibiting the regulatory subunits of the PP1a enzyme Figure 6. complex (GADD34 and CReP) that facilitate dephosphoryla- Depletion of GADD34 and CReP induces apoptosis in MCF7:5C cells. a a Assessment of PARP cleavage, phospho-eIF2a, ATF4, and CHOP by Western tion of eIF2 (26, 28; Fig. 8) can increase the phospho-eIF2 blotting analysis after 24 hours of siRNA-mediated depletion of either levels. We also determined whether this inhibition is sufficient GADD34 or CReP in MCF7:5C cells. to initiate a signaling cascade similar to E2 in MCF7:5C cells, and so induce apoptosis. We used both salubrinal, a pharma- cological agent known to inhibit GADD34 and CReP (41), and that distinctly contrasts with paclitaxel-induced apoptosis in two different siRNAs each against GADD34 and CReP. Inhibit- the same cells (15). However, the underlying mechanism ing GADD34 and CReP by salubrinal in MCF7:5C cells responsible for this temporal lag was unclear. Here, we show markedly elevated the phospho-eIF2a levels and also induced that the delayed trigger coincides with the time needed to apoptosis as evident by PARP cleavage and decreased cell accumulate proteins after E2 treatment. Thus, a threshold level number after 6 days (Fig. 3A and B). Remarkably, induction mayberequiredtoinitiateEnRstressandactivatetheUPR.In of apoptosis by salubrinal was not restricted to MCF7:5C MCF7:5C cells, E2 stimulates global protein synthesis at early cells. LCC9 cells (31), and T47D:A18/4-OHT (32) that are not time points, but this is followed by a significant suppression of susceptible to E2-induced apoptosis (Supplementary Fig. S6) protein translation after approximately 48 hours. This temporal but are E2-independent and cross-resistant to tamoxifen pattern of global protein synthesis coincided with high PERK and fulvestrant, also undergo apoptosis after salubrinal treat- phosphorylation and a subsequent increase in phosphorylated ment (Fig. 3). Comparing the individual role of GADD34 eIF2a, a direct target of PERK. Phospho-eIF2a canthenatten- and CReP using siRNA-mediated knockdown revealed that uate global protein synthesis (45, 46). The timing of this trigger depletion of either gene product induced apoptosis in was consistent with our previous reports, where E2-induced MCF7:5C cells (Fig. 6). However, CReP depletion induced a apoptosis was rescued by 4OHT at early time points but not more robust activation of phospho-eIF2a, whereas a partial with 48 hours of estrogen treatment (15). compensatory mechanism was evident when GADD34 was Expression of GADD34 increases after eIF2a phosphoryla- depleted, as CReP protein was expressed at a higher level than tion and acts as a feedback inhibitor of eIF2a dephosphoryla- the control siRNA-treated cells, and phospho-eIF2a was only tion (29, 39). In MCF7:5C cells, despite a dramatic E2-induced marginally elevated. increase in GADD34 mRNA, its protein levels remain unaltered The salubrinal-mediated increase in phospho-eIF2a was fol- (Fig. 2A and B). This observation could reflect a critical event lowed by elevated protein levels of ATF4, CHOP, and cleaved that renders the cells unable to use feedback to reduce the PARP in all three cell lines (Fig. 3D, E and F), implicating the phospho-eIF2a levels and restore proteostasis. Consequently, same mechanism of apoptosis as occurs with inhibition of cells would experience a prolonged and unabated UPR, perhaps GADD34 and CReP, and with E2 treatment. We also found that sufficient to induce apoptosis. Translation of GADD34 is pref- salubrinal treatment did not significantly decrease the number erentially enhanced under stress conditions (37), by ribosome of immortalized mammary epithelial cells (MCF10A cells; Sup- reinitiation at the 50 untranslated region that ensures high plementary Fig. S6) suggesting that it may not be toxic to normal

924 Mol Cancer Res; 17(4) April 2019 Molecular Cancer Research

UPR and Estrogen-Induced Apoptosis

Figure 7. 4-Hydroxy tamoxifen potentiates apoptotic effect of salubrinal in MCF7:5C and LCC9 cells. Cell growth was measured using the crystal violet method, after a 6-day period of treatment with 4-hydroxy tamoxifen (1 mmol/L) and salubrinal alone or in combination as indicated in MCF7:5C cells (A)and LCC9 cells (B). 17b-Estradiol (10 nmol/L) was used as a control (, P < 0.05). C, Assessment of PARP cleavage, phospho-eIF2a,totaleIF2a,ATF4, and CHOP by Western blotting analysis, in MCF7:5C and LCC9 cells after 48 hours of treatment with 4-hydroxy tamoxifen (1 mmol/L) and salubrinal (25 mmol/L) alone or in combination as indicated. b-Actin was used to confirm equal protein loading. mammary cells. Therefore, inhibitors of GADD34 and/or CReP The mechanism of estrogen-induced apoptosis in vivo that mimic E2-induced apoptosis may be potentially advanta- was initially reported to be extrinsic (8) but was later shown geous in the clinic. For example, these agents could extend the to occur through the mitochondrion (intrinsic pathway; benefit of tumor regression to a wider patient population, ref. 7) and followed later by the extrinsic mechanism (15). including patients whose tumors are not inherently susceptible The role of mitochondrion-mediated apoptosis in estrogen- to E2. treated MCF7:5C was confirmed (7), as proapoptotic proteins The transcript levels of several signature stress-related genes BIM and BAX were upregulated but antiapoptotic proteins that were previously reported to be regulated by estrogen in BCL2 and BCLXL were not altered. Notably, CHOP is a direct MCF7:5C cells (14, 16, 17) were similarly regulated by salu- transcriptional inducer of BIMandcanmediateapoptosis brinal alone in MCF7:5C cells (Fig. 4A–E). This observation by EnR stress (42). Indeed, BIM protein levels increased after indicates that these genes are regulated by the downstream 48–72 hours of E2 treatment and preceded upregulation of effectors of phospho-eIF2a and may not be primary targets of CHOP (Fig. 5A). We found that in both MCF7:5C and LCC9 estrogen. Indeed, most of the observed effects of E2 in MCF7:5C cells salubrinal treatment upregulated BIM and BAX protein cells are likely driven by elevated levels of phospho-eIF2a. whereas BCL2 and BCLXL remained unaltered (Fig. 5B),

www.aacrjournals.org Mol Cancer Res; 17(4) April 2019 925

Sengupta et al.

Intrinsic CHOP apoptosis

Global translation ATF4

P Estrogen siRNA p-eIF2α

GADD34 Salubrinal p-PERK CReP

P P siRNA eIF2α

Global translation

Figure 8. Mechanism depicting estrogen-induced apoptosis in MCF7:5C cells and the apoptosis, which occurs with pharmacologic (salubrinal) and genetic inhibition (siRNA) of GADD34 and CReP that mimics estrogen action.

suggesting a similar mechanism was engaged as in E2-treated Authors' Contributions MCF7:5C cells. In LCC9 cells, pharmacologic inhibition of Conception and design: S. Sengupta, V.C. Jordan, R. Clarke BCL2 induces apoptosis and alters the autophagy pathway in a Development of methodology: S. Sengupta, V.C. Jordan way that prevents the ability to fuel the cellular metabolism Acquisition of data (provided animals, acquired and managed patients, provided facilities, etc.): S. Sengupta, C.M. Sevigny, P. Bhattacharya using degraded products (52). Analysis and interpretation of data (e.g., statistical analysis, biostatistics, It is postulated that 5 years of E2-deprived environment computational analysis): S. Sengupta, C.M. Sevigny, P. Bhattacharya, R. Clarke is required in the clinic to reprogram breast cancer cells to Writing, review, and/or revision of the manuscript: S. Sengupta, C.M. Sevigny, induce apoptosis in response to subsequent E2 treatment (53). V.C. Jordan, R. Clarke Targeting GADD34/CReP could be beneficial since sensitiza- Administrative, technical, or material support (i.e., reporting or organizing data, constructing databases): S. Sengupta tion to E2-therapy may not be required because GADD34/ Study supervision: S. Sengupta, R. Clarke CReP inhibition also induces apoptosis in E nonsensitive 2 Others (Provided cell lines to Dr. Sengupta to conform with the requirements ER-positive breast cancer cells. GADD34 knockout mice do of the referees): V.C. Jordan not show any discernible differences in embryonic development or early adult life (27, 54); therefore, selective targeting Acknowledgments of GADD34 may be less toxic than inhibiting both the regulatory subunits, GADD34 and CReP. Overall, this study The authors thank Karen Creswell and Dan Xun for their help at the defines the trigger of E -induced apoptosis in susceptible Flow Cytometry Shared Resource at Georgetown-Lombardi Comprehensive 2 Cancer Center. This work was supported by Public Health Service breast cancer cells and provides a drugable target that can Awards (U54-CA149147 and U01-CA184902, to R. Clarke), Department be potentially used for therapeutic intervention mimicking of Defense Breast Program (W81XWH-18-1-0722, to R. Clarke), in part by E2-induced apoptosis. GUMC Dean's Pilot Project Award (to S. Sengupta), and the Lombardi Comprehensive Cancer Center Support Grant (CCSG) NIH (P30 CA051008). Disclosure of Potential Conflicts of Interest R. Clarke has ownership interest (including stocks and patents) in American The costs of publication of this article were defrayed in part by the Gene Technologies. No potential conflicts of interest were disclosed by the other payment of page charges. This article must therefore be hereby marked authors. advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact. Disclaimer The views and opinions of the author(s) do not reflect those of the US Army Received May 8, 2018; revised November 15, 2018; accepted January 10, or the Department of Defense. 2019; published first January 17, 2019.

926 Mol Cancer Res; 17(4) April 2019 Molecular Cancer Research

UPR and Estrogen-Induced Apoptosis

References 1. Haddow A, Watkinson JM, Paterson E, Koller PC. Influence of synthetic 23. Walter P, Ron D. The unfolded protein response: from stress pathway to oestrogens on advanced malignant disease. Br Med J 1944;2:393–8. homeostatic regulation. Science 2011;334:1081–6. 2. Ingle JN, Ahmann DL, Green SJ, Edmonson JH, Bisel HF, Kvols LK, et al. 24. Tabas I, Ron D. Integrating the mechanisms of apoptosis induced by Randomized clinical trial of diethylstilbestrol versus tamoxifen in post- endoplasmic reticulum stress. Nat Cell Biol 2011;13:184–90. menopausal women with advanced breast cancer. N Engl J Med 1981;304: 25. Moreno JA, Radford H, Peretti D, Steinert JR, Verity N, Martin MG, et al. 16–21. Sustained translational repression by eIF2alpha-P mediates prion neuro- 3. Kennedy BJ. Hormone therapy for advanced breast cancer. Cancer 1965;18: degeneration. Nature 2012;485:507–11. 1551–7. 26. Connor JH, Weiser DC, Li S, Hallenbeck JM, Shenolikar S. Growth arrest 4. Ellis MJ, Gao F, Dehdashti F, Jeffe DB, Marcom PK, Carey LA, et al. Lower- and DNA damage-inducible protein GADD34 assembles a novel sig- dose vs high-dose oral estradiol therapy of hormone receptor-positive, naling complex containing protein phosphatase 1 and inhibitor 1. aromatase inhibitor-resistant advanced breast cancer: a phase 2 random- Mol Cell Biol 2001;21:6841–50. ized study. JAMA 2009;302:774–80. 27. Kojima E, Takeuchi A, Haneda M, Yagi A, Hasegawa T, Yamaki K, et al. The 5. Iwase H, Yamamoto Y, Yamamoto-Ibusuki M, Murakami KI, Okumura Y, function of GADD34 is a recovery from a shutoff of protein synthesis Tomita S, et al. Ethinylestradiol is beneficial for postmenopausal patients induced by ER stress: elucidation by GADD34-deficient mice. FASEB J with heavily pre-treated metastatic breast cancer after prior aromatase 2003;17:1573–5. inhibitor treatment: a prospective study. Br J Cancer 2013;109:1537–42. 28. Jousse C, Oyadomari S, Novoa I, Lu P, Zhang Y, Harding HP, et al. 6. Zucchini G, Armstrong AC, Wardley AM, Wilson G, Misra V, Seif M, et al. A Inhibition of a constitutive translation initiation factor 2alpha phospha- phase II trial of low-dose estradiol in postmenopausal women with tase, CReP, promotes survival of stressed cells. J Cell Biol 2003;163: advanced breast cancer and acquired resistance to aromatase inhibition. 767–75. Eur J Cancer 2015;51:2725–31. 29. Novoa I, Zeng H, Harding HP, Ron D. Feedback inhibition of the 7. Lewis JS, Meeke K, Osipo C, Ross EA, Kidawi N, Li T, et al. Intrinsic unfolded protein response by GADD34-mediated dephosphorylation mechanism of estradiol-induced apoptosis in breast cancer cells resistant of eIF2alpha. J Cell Biol 2001;153:1011–22. to estrogen deprivation. J Natl Cancer Inst 2005;97:1746–59. 30. Jiang SY, Wolf DM, Yingling JM, Chang C, Jordan VC. An estrogen receptor 8. Song RX, Mor G, Naftolin F, McPherson RA, Song J, Zhang Z, et al. Effect of positive MCF-7 clone that is resistant to antiestrogens and estradiol. long-term estrogen deprivation on apoptotic responses of breast cancer Mol Cell Endocrinol 1992;90:77–86. cells to 17beta-estradiol. J Natl Cancer Inst 2001;93:1714–23. 31. Brunner N, Boysen B, Jirus S, Skaar TC, Holst-Hansen C, Lippman J, et al. 9. Yao K, Lee ES, Bentrem DJ, England G, Schafer JI, O'Regan RM, et al. MCF7/LCC9: an antiestrogen-resistant MCF-7 variant in which acquired Antitumor action of physiological estradiol on tamoxifen-stimulated resistance to the steroidal antiestrogen ICI 182,780 confers an early cross- breast tumors grown in athymic mice. Clin Cancer Res 2000;6:2028–36. resistance to the nonsteroidal antiestrogen tamoxifen. Cancer Res 1997;57: 10. Balaburski GM, Dardes RC, Johnson M, Haddad B, Zhu F, Ross EA, et al. 3486–93. Raloxifene-stimulated experimental breast cancer with the paradoxical 32. Park WC, Liu H, Macgregor Schafer J, Jordan VC. Deregulation of estrogen actions of estrogen to promote or prevent tumor growth: a unifying concept induced telomerase activity in tamoxifen-resistant breast cancer cells. in anti-hormone resistance. Int J Oncol 2010;37:387–98. Int J Oncol 2005;27:1459–66. 11. Liu H, Lee ES, Gajdos C, Pearce ST, Chen B, Osipo C, et al. Apoptotic action 33. Sengupta S, Biarnes MC, Clarke R, Jordan VC. Inhibition of BET proteins of 17beta-estradiol in raloxifene-resistant MCF-7 cells in vitro and in vivo. impairs estrogen-mediated growth and transcription in breast cancers by J Natl Cancer Inst 2003;95:1586–97. pausing RNA polymerase advancement. Breast Cancer Res Treat 2015;150: 12. Obiorah IE, Fan P, Sengupta S, Jordan VC. Selective estrogen-induced 265–78. apoptosis in breast cancer. Steroids 2014;90:60–70. 34. Schmidt EK, Clavarino G, Ceppi M, Pierre P. SUnSET, a nonradio- 13. Lewis JS, Osipo C, Meeke K, Jordan VC. Estrogen-induced apoptosis in a active method to monitor protein synthesis. Nat Methods 2009;6: breast cancer model resistant to long-term estrogen withdrawal. 275–7. J Steroid Biochem Mol Biol 2005;94:131–41. 35. Lu PD, Harding HP, Ron D. Translation reinitiation at alternative open 14. Ariazi EA, Cunliffe HE, Lewis-Wambi JS, Slifker MJ, Willis AL, Ramos P, reading frames regulates gene expression in an integrated stress response. et al. Estrogen induces apoptosis in estrogen deprivation-resistant breast J Cell Biol 2004;167:27–33. cancer through stress responses as identified by global gene expression 36. Vattem KM, Wek RC. Reinitiation involving upstream ORFs regulates across time. Proc Natl Acad Sci U S A 2011;108:18879–86. ATF4 mRNA translation in mammalian cells. Proc Natl Acad Sci U S A 15. Obiorah I, Sengupta S, Fan P, Jordan VC. Delayed triggering of oestrogen 2004;101:11269–74. induced apoptosis that contrasts with rapid paclitaxel-induced breast 37. Lee YY, Cevallos RC, Jan E. An upstream open reading frame regulates cancer cell death. Br J Cancer 2014;110:1488–96. translation of GADD34 during cellular stresses that induce eIF2alpha 16. Fan P, Griffith OL, Agboke FA, Anur P, Zou X, McDaniel RE, et al. c-Src phosphorylation. J Biol Chem 2009;284:6661–73. modulates estrogen-induced stress and apoptosis in estrogen-deprived 38. Ma Y, Hendershot LM. Delineation of a negative feedback regulatory loop breast cancer cells. Cancer Res 2013;73:4510–20. that controls protein translation during endoplasmic reticulum stress. 17. Sengupta S, Obiorah I, Maximov P, Curpan R, Jordan VC. Molecular J Biol Chem 2003;278:34864–73. mechanism of action of bisphenol and bisphenol A mediated by oestrogen 39. Brush MH, Weiser DC, Shenolikar S. Growth arrest and DNA damage- receptor alpha in growth and apoptosis of breast cancer cells. inducible protein GADD34 targets protein phosphatase 1 alpha to the Br J Pharmacol 2013;169:167–78. endoplasmic reticulum and promotes dephosphorylation of the alpha 18. Obiorah IE, Jordan VC. Differences in the rate of oestrogen-induced subunit of eukaryotic translation initiation factor 2. Mol Cell Biol 2003;23: apoptosis in breast cancer by oestradiol and the triphenylethylene bisphe- 1292–303. nol. Br J Pharmacol 2014;171:4062–72. 40. Young SK, Willy JA, Wu C, Sachs MS, Wek RC. Ribosome reinitiation directs 19. Clarke R, Tyson JJ, Dixon JM. Endocrine resistance in breast cancer–an gene-specific translation and regulates the integrated stress response. overview and update. Mol Cell Endocrinol 2015;418:220–34. J Biol Chem 2015;290:28257–71. 20. Clarke R, Cook KL. Unfolding the role of stress response signaling in 41. Boyce M, Bryant KF, Jousse C, Long K, Harding HP, Scheuner D, et al. A endocrine resistant breast cancers. Front Oncol 2015;5:140. selective inhibitor of eIF2alpha dephosphorylation protects cells from ER 21. Clarke R, Cook KL, Hu R, Facey CO, Tavassoly I, Schwartz JL, et al. stress. Science 2005;307:935–9. Endoplasmic reticulum stress, the unfolded protein response, autophagy, 42. Puthalakath H, O'Reilly LA, Gunn P, Lee L, Kelly PN, Huntington ND, et al. and the integrated regulation of breast cancer cell fate. Cancer Res 2012;72: ER stress triggers apoptosis by activating BH3-only protein Bim. Cell 2007; 1321–31. 129:1337–49. 22. Cook KL, Shajahan AN, Warri A, Jin L, Hilakivi-Clarke LA, Clarke R. 43. Sano R, Reed JC. ER stress-induced cell death mechanisms. Biochim Glucose-regulated protein 78 controls cross-talk between apoptosis and Biophys Acta 2013;1833:3460–70. autophagy to determine antiestrogen responsiveness. Cancer Res 2012;72: 44. Woehlbier U, Hetz C. Modulating stress responses by the UPRosome: a 3337–49. matter of life and death. Trends Biochem Sci 2011;36:329–37.

www.aacrjournals.org Mol Cancer Res; 17(4) April 2019 927

Sengupta et al.

45. Harding HP, Zhang Y, Ron D. Protein translation and folding are coupled epidermal growth factor receptor 2 in breast cancer cell lines. Eur J Cancer by an endoplasmic-reticulum-resident kinase. Nature 1999;397:271–4. 2012;48:3488–98. 46. Harding HP, Zhang Y, Bertolotti A, Zeng H, Ron D. Perk is essential for 51. Bhat-Nakshatri P, Wang G, Appaiah H, Luktuke N, Carroll JS, Geistlinger translational regulation and cell survival during the unfolded protein TR, et al. AKT alters genome-wide estrogen receptor alpha binding and response. Mol Cell 2000;5:897–904. impacts estrogen signaling in breast cancer. Mol Cell Biol 2008;28: 47. Reid DW, Tay AS, Sundaram JR, Lee IC, Chen Q, George SE, et al. 7487–503. Complementary roles of GADD34- and CReP-containing eukaryotic ini- 52. Schwartz-Roberts JL, Shajahan AN, Cook KL, Warri A, Abu-Asab M, Clarke tiation factor 2alpha phosphatases during the unfolded protein response. R. GX15–070 (obatoclax) induces apoptosis and inhibits cathepsin D- and Mol Cell Biol 2016;36:1868–80. L-mediated autophagosomal lysis in antiestrogen-resistant breast cancer 48. Thomas SM, Brugge JS. Cellular functions regulated by Src family kinases. cells. Mol Cancer Ther 2013;12:448–59. Annu Rev Cell Dev Biol 1997;13:513–609. 53. Jordan VC. The new biology of estrogen-induced apoptosis applied to treat 49. Castoria G, Migliaccio A, Bilancio A, Di Domenico M, de Falco A, Lombardi and prevent breast cancer. Endocr Relat Cancer 2015;22:R1–31. M, et al. PI3-kinase in concert with Src promotes the S-phase entry of 54. Harding HP, Zhang Y, Scheuner D, Chen JJ, Kaufman RJ, Ron D. Ppp1r15 oestradiol-stimulated MCF-7 cells. EMBO J 2001;20:6050–9. gene knockout reveals an essential role for translation initiation factor 2 50. Fan P, McDaniel RE, Kim HR, Clagett D, Haddad B, Jordan VC. Modulating alpha (eIF2alpha) dephosphorylation in mammalian development. therapeutic effects of the c-Src inhibitor via oestrogen receptor and human Proc Natl Acad Sci U S A 2009;106:1832–7.

928 Mol Cancer Res; 17(4) April 2019 Molecular Cancer Research

Estrogen-Induced Apoptosis in Breast Cancers Is Phenocopied by Blocking Dephosphorylation of Eukaryotic Initiation Factor 2 Alpha (eIF2α) Protein

Surojeet Sengupta, Catherine M. Sevigny, Poulomi Bhattacharya, et al.

Mol Cancer Res 2019;17:918-928. Published OnlineFirst January 17, 2019.

Updated version Access the most recent version of this article at: doi:10.1158/1541-7786.MCR-18-0481

Supplementary Access the most recent supplemental material at: Material http://mcr.aacrjournals.org/content/suppl/2019/01/17/1541-7786.MCR-18-0481.DC1

Cited articles This article cites 54 articles, 23 of which you can access for free at: http://mcr.aacrjournals.org/content/17/4/918.full#ref-list-1

Citing articles This article has been cited by 1 HighWire-hosted articles. Access the articles at: http://mcr.aacrjournals.org/content/17/4/918.full#related-urls

E-mail alerts Sign up to receive free email-alerts related to this article or journal.

Reprints and To order reprints of this article or to subscribe to the journal, contact the AACR Publications Department at Subscriptions [email protected].

Permissions To request permission to re-use all or part of this article, use this link http://mcr.aacrjournals.org/content/17/4/918. Click on "Request Permissions" which will take you to the Copyright Clearance Center's (CCC) Rightslink site.