GCN2iB

BCR-ABL tyrosine kinase inhibition induces metabolic vulnerability by preventing the integrated stress response in K562 cells

A B S T R A C T
The integrated stress response (ISR) is a cellular process that is characterized by activation of eukaryotic initiation factor (eIF)2a kinases and subsequent induction of activating transcription factor (ATF)4. The ISR plays an important role in protecting cells from tumor-related metabolic stresses, such as nutrient deprivation and perturbed proteostasis. Here, we showed that disruption of the ISR, together with increased cellular stress vulnerability, was produced by pharmacological inhibition of BCR-ABL, the oncogenic driver in chronic myeloid leukemia (CML). Treatment of CML-derived K562 cells with BCR-ABL tyrosine kinase inhibitors, including imatinib, dasatinib, nilotinib and ponatinib, prevented activation of eIF2a kinases, protein kinase-like endoplasmic reticulum kinase (PERK) and general control non- derepressible 2, and downstream ATF4 induction during metabolic stress. Prevention of ATF4 induction likely occurred as a result of the combinatorial suppression of the eIF2a kinase and phosphoinositide 3- kinase/mammalian target of rapamycin signaling pathways. In addition, we found that pharmacological inhibition of PERK mimicked BCR-ABL inhibition to enhance apoptosis induction under stress conditions. These findings indicate that the ISR is under the control of BCR-ABL and may foster adaptation to tumorigenic stresses in CML cells.

1.Introduction
The integrated stress response (ISR) is a cellular adaptation mechanism to various tumor-related metabolic stresses, such as endoplasmic reticulum (ER) stress or amino acid deprivation, and plays important roles in cell proliferation and survival [1e3]. Under distinct stress conditions, ISR is activated by four different eukaryotic initiation factor (eIF)2a kinases, including general con- trol nonderepressible (GCN)2 [2], protein kinase-like endoplasmic reticulum kinase (PERK) [3], double-stranded RNA-dependent protein kinase [4] and heme-regulated inhibitor [5]. These eIF2a kinases commonly phosphorylate eIF2a at Ser51, reducing general protein synthesis and subsequently promoting translation of acti- vating transcription factor (ATF)4 mRNA [6,7]. ATF4 increases transcription of genes for stress adaptation such as protein folding, amino acid metabolism and redox metabolism [6,7].The ISR can be triggered by oncogene activation that alters nutrient demands and translation control in cancer cells [8,9]. For example, PIK3CA-mutated colon cancer cells show high glutamine dependency, and mutated PIK3CA contributes to the glutamine metabolism though ISR activation [10]. Transformation and tumor growth induced by c-Myc overexpression can be promoted by PERK-initiated ISR activation to cope with perturbed proteostasis [11]. These previous findings have revealed a link between ISR activation and adaptation to oncogenic signal-associated stress. However, it is not known whether activated oncogenes are always involved in the control of ISR activation.In this study, we focused on the BCR-ABL fusion gene, the product of the Philadelphia chromosome, which is expressed in most chronic myeloid leukemia (CML) patients [12]. The BCR-ABL protein possesses constitutively activated tyrosine kinase activity[13] that transmits oncogenic signals to various downstream ef- fectors, such as Ras [14,15], phosphoinositide 3-kinase (PI3K) [16,17] and signal transducer and activator of transcription (STAT)5 [18,19]. Here, we showed that BCR-ABL was involved in regulation of the ISR through its tyrosine kinase activity. Indeed, BCR-ABL tyrosine kinase inhibitors suppressed activation of ISR, resulting in enhanced cell death under metabolic stress conditions.

2.Materials and methods
2.1.Cell culture and reagents
Human CML cell line K562 was obtained from the American Type Culture Collection (Manassas, VA, USA). K562 cells were maintained in RPMI 1640 (Wako Pure Chemical Industries, Osaka, Japan) with 5% fetal bovine serum (Merck, Burlington, MA, USA), as described previously [20]. Imatinib, dasatinib, nilotinib, ponatinib, BKM120, MK2206, AZD6244, Rapamycin, TG101348, PP242 and BEZ235 (Selleck Chemicals, Houston, TX, USA), GSK2656157 (Merck, Burlington, MA, USA), tunicamycin (Nacalai Tesque, Kyoto, Japan) and thapsigargin (Wako) were dissolved in DMSO (<0.5% final concentration) as a stock solution and were added to culture medium. Histidinol (Merck) was dissolved in sterilized distilled water. 2.2.Immunoblot analysis Immunoblot analysis was performed as described previously [20]. Cell lysates were prepared using SDS lysis buffer (62.5 mM TriseHCl [pH 6.8], 2% SDS, 50 mM dithiothreitol, 10% glycerol). Protein concentrations were determined using the Bio-Rad Pro- tein Assay (Hercules, CA, USA). Protein samples were subjected to SDS-PAGE and subsequently transferred to nitrocellulose mem- branes. Membranes were incubated with the following primary antibodies: phospho-GCN2 (Thr899) and PERK (Abcam, Cam- bridge, UK), ATF6 (Proteintech, Rosemont, IL, USA), GCN2, ATF4, c- ABL, phospho-c-ABL (Y412), X-box binding protein (XBP)1s, inositol-requiring enzyme (IRE)1, phospho-Akt (Ser473), Akt, phospho-p44/42 MAPK (ERK1/2) (Thr202/Tyr204), p44/42 MAPK (ERK1/2), phospho-STAT5 (Tyr694), STAT5, phospho-4EBP1 (Ser65), 4EBP1, phospho-S6 (Ser235/236), S6, cleaved-poly (ADP-ribose) polymerase (PARP) and ribosomal protein S3 (Cell Signaling Technology, Danvers, MA, USA). The specific signals were detected using Western Lighting plus ECL (Perkin Elmer, Waltham, MA, USA). 2.3.Detection of apoptotic cells Cells were seeded at 2 × 104/well in 96-well plates and treated with the indicated concentration of reagents. After 18 h, the nuclei were stained with 10 mg/mL Hoechst 33342 (Thermo Fisher Scien- tific) for 10 min. Fluorescent images were acquired by IN Cell Analyzer 6000 (GE Healthcare, Little Chalfont, UK). The percentage of apoptotic cells was calculated by dividing the number of apoptotic cells by the total number of cell. Results are shown by boxplot diagram (12 fields of view in total under identical condi- tions, 3 fields of view/well). The boxplot shows the median and dispersion of each group. The box represent distributions of data from 25% to 75% and the whiskers show the range of 1.5-fold of the box. 2.4.Cell proliferation assay Cells were seeded at 1.5 × 104/well in 96-well plates and treated with the indicated concentrations of reagents. After 48 or 72 h, cell viability was measured by the CellTiter-Glo Luminescent Cell Viability Assay (Promega, Madison, WI, USA). The cell viability is shown as the percentage of the controls. 3.Results 3.1.BCR-ABL inhibition prevents activation of PERK and GCN2 To investigate the effects of BCR-ABL inhibition on the PERK signaling pathway, we treated human CML K562 cells with BCR-ABL inhibitors imatinib (1 or 10 mM), dasatinib (10 or 100 nM) or ponatinib (10 or 100 nM) for 18 h under ER stress conditions of glucose withdrawal (Fig. 1A) as well as tunicamycin (10 mg/mL), an N-linked glycosylation inhibitor [21] (Fig. 1B). Activation of the PERK signaling pathway by glucose withdrawal and tunicamycin treatment was demonstrated by phosphorylation of PERK, which was seen as a bandshift from lower (underphosphorylated) to higher (phosphorylated), and by induction of ATF4. PERK Fig. 1. BCR-ABL inhibitors prevent eukaryotic initiation factor 2a kinases activation and activating transcription factor 4 (ATF4) induction. K562 cells were treated with imatinib (1 or 10 mM), dasatinib (10 or 100 nM) or ponatinib (10 or 100 nM) for 18 h under normal (Nor) or the following stress conditions: (A) glucose (Glc) withdrawal; (B) 10 mg/ mL tunicamycin (TM) addition; (C) glutamine (Gln) withdrawal; and (D) 2 mM histidinol (His) addition. The cell lysates were subjected to immunoblot analysis with specific antibodies, as indicated. GCN2, general control nonderepressible 2; PERK, protein kinase-like endoplasmic reticulum kinase; RPS3, ribosomal protein S3. phosphorylation and ATF4 induction were prevented by any of the BCR-ABL inhibitors. BCR-ABL inhibition was confirmed by the decreased level of tyrosine phosphorylation of BCR-ABL proteins (Fig. 1A and B).We also examined the effects of BCR-ABL inhibition on activation of the PERK signaling pathway at an earlier time point of 6 h, using BCR-ABL inhibitor dasatinib and chemical stressor tunicamycin (Fig. S1). Dasatinib prevented tunicamycin-induced ATF4 expression, with marginal inhibition of PERK phosphorylation, suggesting that dasatinib also inhibited PERK-independent mechanisms of ATF4 in- duction. Dasatinib also inhibited emergence of active forms of transcription factors XBP1s and p50-ATF6, which are the respective downstream effectors of IRE1-and ATF6-originated UPR signaling pathways under ER stress conditions [22] (Fig. S1). Thus, BCR-ABL inhibition can prevent the production of effector transcription fac- tors via any of the three branches of the UPR.Next, we examined whether BCR-ABL inhibition affected activa- tion of the GCN2 signaling pathway during amino acid deprivation. We cultured K562 cells for 18 h under GCN2-activating stress condi- tions of glutamine withdrawal (Fig. 1C), as well as the addition of chemical stressor histidinol, mimicking histidine limitation [23] (Fig.1D). Both glutamine withdrawal and histidinol addition activated the GCN2 signaling pathway, as monitored by GCN2 phosphorylation and ATF4 induction. In the presence of BCR-ABL inhibitors imatinib, dasatinib or ponatinib, stress-induced GCN2 phosphorylation and ATF4 induction were suppressed, with decreased tyrosine phos- phorylation of BCR-ABL proteins. These results indicate that BCR-ABL inhibition in K562 cells prevents both PERK- and GCN2-initiated ISR during ER perturbation and amino acid limitation. 3.2. ATF4 induction is sensitive to PI3K/mammalian target of rapamycin (mTOR) inhibition BCR-ABL activates multiple signaling pathways, including JAK/STAT, Raf/MEK/ERK and PI3K/Akt/mTOR [14e19]. By screening of several inhibitors against BCR-ABL downstream kinases, we found that PI3K inhibitor BKM120 and mTOR inhibitor rapamycin pre- vented ATF4 induction in K562 cells treated with tunicamycin for 18 h (Fig. S2). In agreement, ATF4 induction in tunicamycin-treated K562 cells was suppressed by BEZ235 and PP242, which are se- lective inhibitors of PI3K and mTOR, respectively (Fig. 2A). Similar suppression of ATF4 induction was observed in K562 cells treated with histidinol for 18 h (Fig. 2B). These kinase inhibitors, however, exhibited weak effects on PERK and GCN2 phosphorylation in tunicamycin- and histidinol-treated K562 cells, respectively (Fig. 2A and B), as compared with BCR-ABL tyrosine kinase in- hibitors (Fig. 1).Previous studies have shown that prevention of ATF4 induction is strongly associated with emergence of the hypophosphorylated, activated form of 4E-BP1, a repressor of translation initiation, when mTOR is inhibited [24,25]. Treatment of K562 cells with BEZ235 or PP242 induced hypophosphorylation of 4EBP1 regardless of the presence of chemical stressors tunicamycin and histidinol (Fig. 2C). BEZ235 and PP242 also led to hypophosphorylation of ribosomal protein S6, another downstream molecule of mTOR signaling, but hardly affected phosphorylation status of BCR-ABL (Fig. 2C). Similar to BEZ235 and PP242, the BCR-ABL tyrosine kinase inhibitor dasa- tinib induced hypophosphorylation of 4EBP1 as well as S6 (Fig. 2D). These results suggest that inhibition of PI3K and mTOR was involved in prevention of stress-induced ATF4 in BCR-ABL-inhibited K562 cells, possibly through 4EBP1 activation. 3.3.Apoptosis induction by BCR-ABL inhibition under stress conditions We examined the viability of K562 cells after treatment with BCR-ABL tyrosine kinase inhibitor dasatinib (10 or 100 nM) under stress conditions induced by tunicamycin or histidinol for 18 h.Fig. 2. Phosphoinositide 3-kinase/mammalian target of rapamycin pathway and integrated stress response activation. K562 cells were treated with 500 nM BEZ235 or 1 mM PP242 (AeC), or dasatinib (Dasa: 10 or 100 nM) (D) alone (Nor), or together with 10 mg/mL tunicamycin (TM) or 2 mM histidinol (His) for 18 h. The cell lysates were subjected to immunoblot analysis with specific antibodies, as indicated. ATF4, activating transcription factor 4; GCN2, general control nonderepressible 2; PERK, protein kinase-like endoplasmic reticulum kinase; RPS3, ribosomal protein S3.Dasatinib induced apoptosis more strongly under stress than normal conditions, as determined by nuclear condensation and fragmentation assay, as well as immunoblot analysis of cleaved PARP (Fig. 3AeC). Such increased apoptosis also occurred in K562 cells treated with dasatinib under stress conditions of glucose deprivation as well as another chemical stressor thapsigargin (Figs. S3A and B). Similar results were obtained with BCR-ABL in- hibitors imatinib, ponatinib or nilotinib under stress conditions induced by tunicamycin or histidinol (Fig. 3D). Furthermore, cell survival assay after 72-h exposure revealed that K562 cells were sensitized to BCR-ABL inhibitors dasatinib and ponatinib under stress conditions (Fig. 3E). 3.4.Apoptosis induction by ISR inhibition To examine the impact of ISR on cell survival under stress con- ditions, we used GSK2656157, a highly selective inhibitor of PERK [26]. Treatment of K562 cells with GSK2656157 prevented PERK phosphorylation and ATF4 induction under PERK-activating stress conditions (using tunicamycin, Fig. 4A). GSK2656157 did not pre- vent GCN2 phosphorylation or ATF4 induction under GCN2- activating stress conditions (using histidinol, Fig. S4A). GSK2656157 hardly affected phosphorylation status of BCR-ABL under any conditions examined (Figs. 4A and S4A). Thus, GSK2656157 selectively inhibited the PERK/ATF4 pathway in K562 cells. Importantly, GSK2656157 treatment for ≥48 h induced apoptosis under stress conditions induced by tunicamycin (Fig. 4Band C) and reduced cell survival dose dependently (Fig. 4D). GSK2656157 had no effect on stress conditions induced by histi- dinol (Figs. S4B and C). These results indicate that ISR inhibition under stress conditions can lead to the death of K562 cells. 4.Discussion We have shown that BCR-ABL tyrosine kinase inhibition pre- vents ISR activation in K562 cells. Indeed, activation of major eIF2a kinases PERK and GCN2 was suppressed by BCR-ABL inhibition under metabolic stress conditions that induced ER perturbation and amino acid deprivation, respectively. Importantly, preventing the ISR by BCR-ABL inhibition was associated with enhanced cell death under the stress conditions. Increased cell death was also observed in K562 cells treated with PERK inhibitor GSK2656157 during ER stress, supporting the idea that stress vulnerability induced by BCR-ABL inhibition occurs through prevention of the ISR signaling pathway. During ER stress, BCR-ABL inhibition also prevented activation of the IRE1 and ATF6 signaling pathways, which, along with the PERK signaling pathway, constitute the three branches of UPR. These findings suggest that BCR-ABL counteracts metabolic vulnerability through regulating multiple stress signaling pathways.In addition to the ISR signaling, BCR-ABL likely uses mTOR signaling to regulate ATF4 expression. Indeed, we found that inhi- bition of the PI3K/mTOR signaling pathway, especially inhibition of Fig. 3. Apoptosis induction by BCR-ABL inhibitors under stress conditions. (AeC) K562 cells were treated with dasatinib (Dasa: 10 or 100 nM) alone (Nor), or together with 10 mg/mL tunicamycin (TM) or 2 mM histidinol (His) for 18 h. After staining with Hoechst 33324, cell nuclei were observed using an IN Cell Analyzer 6000 (A). The apoptotic populations were calculated and shown in boxplot diagrams (B). The cell lysates were subjected to immunoblot analysis with anti-cleaved PARP antibody (C). (D) K562 cells were treated with 1 mM imatinib (Ima), 100 nM ponatinib (Pona) or 100 nM nilotinib (Nilo) alone (Nor), or together with 10 mg/mL TM or 2 mM His for 18 h and the apoptotic populations were determined as in (B). (E) K562 cells were treated with Dasa or Pona at indicated dose alone (Nor), or together with 10 mg/mL TM or 2 mM His for 72 h. Cell viability was determined by the Cell Titer-Glo luminescent cell viability assay. PARP, poly (ADP-ribose) polymerase.Fig. 4. Apoptosis induction by PERK inhibition. (AeC) K562 cells were treated with GSK2656157 (GSK; +, 300 nM) alone (Nor) or together with tunicamycin (TM; 10 mg/mL). After 6 h treatment, cell lysates were subjected to immunoblot analysis with specific antibodies, as indicated (A). After 48 h treatment, apoptosis induction was evaluated with nuclear staining with Hoechst 33324 (upper boxplot diagram) and immunoblot analysis of cleaved PARP (below), as in Fig. 3B and C, respectively (B, C). (D) After 48 or 72 h treatment of K562 cells with GSK2656157 (0, 100, 300 nM) alone (Nor) or together with 10 mg/mL TM, cell viability was determined by the Cell Titer-Glo luminescent cell viability assay. ATF4, activating transcription factor 4; PARP, poly (ADP-ribose) polymerase; PERK, protein kinase-like endoplasmic reticulum kinase; RPS3, ribosomal protein S3 the downstream mTOR kinase, prevented stress-induced ATF4 expression. This inhibitory effect occurred without affecting acti- vation of eIF2a kinases PERK and GCN2 during metabolic stress. Instead, mTOR inhibition, like BCR-ABL inhibition, led to activation of hypophosphorylation of translation repressor 4EBP1. These ob- servations are consistent with the previous finding that 4EBP1 negatively regulates ATF4 expression under the control of mTORC1, a complex containing mTOR [25]. Notably, mTORC1 activity is regulated downstream of various oncogenic signals [27], raising the possibility that activated oncogenes, including BCR-ABL, may commonly regulate ATF4 expression via mTOR signaling. In contrast, at present, little is known about the mechanisms of how activation of the ISR signaling pathway, especially activation of eIF2a kinases PERK and GCN2, is regulated by BCR-ABL as well as other activated oncogenes. Further studies will be needed to un- derstand the precise molecular link between ISR and oncogenic signals. With the emergence of imatinib and subsequent generation of BCR-ABL inhibitors, such as dasatinib and ponatinib, CML has become well-controlled and overall survival has improved signifi- cantly [28]. However, it remains difficult to cure CML even with long-term therapy and the appearance of acquired drug resistance is a serious problem [29]. Under GCN2iB these circumstances, it is desirable to develop new curative therapeutic approaches that eradicate re- sidual CML cells. In this study, we found that apoptosis induction of CML cells can be enhanced by BCR-ABL inhibition under ISR- activating conditions. Interestingly, ISR-activating drugs such as mipsagargin, a thapsigargin derivative, are currently under devel- opment as anticancer drugs [30]. Thus, combination of BCR-ABL inhibitors and chemical stressors may provide a novel, feasible approach to combat CML.