Activation of the UPR Protects against Cigarette Smoke-induced RPE Apoptosis through Up-Regulation of Nrf2*

Background: The unfolded protein response (UPR) has been implicated in retinal cell death but the mechanism is unknown. Results: Cigarette smoke extract induces RPE cell apoptosis, which is alleviated by enhancing the UPR function. Conclusion: UPR activation is required for RPE survival through up-regulation of Nrf2. Significance: Enhancing Nrf2 and adaptive UPR protects the RPE against oxidative injury and apoptosis. Recent studies have revealed a role of endoplasmic reticulum (ER) stress-induced unfolded protein response (UPR) in the regulation of RPE cell activity and survival. Herein, we examined the mechanisms by which the UPR modulates apoptotic signaling in human RPE cells challenged with cigarette smoking extract (CSE). Our results show that CSE exposure induced a dose- and time-dependent increase in ER stress markers, enhanced reactive oxygen species (ROS), mitochondrial fragmentation, and apoptosis of RPE cells. These changes were prevented by the anti-oxidant NAC or chemical chaperone TMAO, suggesting a close interaction between oxidative and ER stress in CSE-induced apoptosis. To decipher the role of the UPR, overexpression or down-regulation of XBP1 and CHOP genes was manipulated by adenovirus or siRNA. Overexpressing XBP1 protected against CSE-induced apoptosis by reducing CHOP, p-p38, and caspase-3 activation. In contrast, XBP1 knockdown sensitized the cells to CSE-induced apoptosis, which is likely through a CHOP-independent pathway. Surprisingly, knockdown of CHOP reduced p-eIF2α and Nrf2 resulting in a marked increase in caspase-3 activation and apoptosis. Furthermore, Nrf2 inhibition increased ER stress and exacerbated cell apoptosis, while Nrf2 overexpression reduced CHOP and protected RPE cells. Our data suggest that although CHOP may function as a pro-apoptotic gene during ER stress, it is also required for Nrf2 up-regulation and RPE cell survival. In addition, enhancing Nrf2 and XBP1 activity may help reduce oxidative and ER stress and protect RPE cells from cigarette smoke-induced damage.

and 3) ATF6-chaporene pathway (14,15). The latter two pathways primarily regulate ER chaperones, such as GRP78 and p58IPK, thereby increasing the capacity of protein folding (15)(16)(17)(18)(19). The early activation of PERK increasing the eIF2␣ phosphorylation reduces protein translation rate, which enables the ER to recover from its stressed condition. However, persistent ER stress activates the pro-apoptotic gene C/EBP homologous protein (CHOP), a downstream factor of the PERK pathway, resulting in apoptosis and cell death (14,20,21).
Aqueous whole-phase cigarette smoke extract (CSE) has been widely used to study the effects of cigarette smoke in cultured cells and disease models, such as chronic obstructive pulmonary disease, cancer, and cardiovascular diseases (22)(23)(24)(25)(26)(27)(28). The extract comprises water-soluble components of both the particulate and gas phases of cigarette smoke, which were believed to be present in the bloodstream of smokers (29). Importantly, it contains high levels of pro-oxidant compounds and free radicals that induce oxidative injury of various cells and tissues including RPE cells (30). In human bronchial epithelial and lung cells, CSE elicits ER stress likely through ROS resulting in CHOP-dependent apoptosis (31)(32)(33). In agreement with this finding, recent studies show that cigarette smoke induces ER stress in the RPE (34 -36). In cultured RPE cells, hydroquinone, a potent oxidant in cigarette smoke, differentially regulates the UPR pathways by up-regulating CHOP expression but suppressing XBP1-mediated adaptive UPR signaling (36). Overexpressing XBP1 or decreasing CHOP level by reducing ER stress or inhibition of its upstream regulator, eIF-2a phosphorylation, protects cells from hydroquinone-induced apoptosis (36), suggesting a role of the UPR in regulating the RPE cell survival during oxidative injury.
In the present study, we examined the role of the UPR factors in RPE cells challenged with CSE and intended to understand the interactions between oxidative stress and ER stress and signaling pathways of Nrf2, XBP1s, and CHOP in CSE-induced apoptosis of RPE cells.

EXPERIMENTAL PROCEDURES
Preparation of CSE-Cigarette smoke extract (dissolved in DMSO, 40 mg/ml total particular matter, nicotine content ϭ 6%) was purchased from Murty Pharmaceuticals (Lexington, KY) and was kept at Ϫ20°C. Before each treatment, CSE was freshly prepared into working solutions and diluted with HEPES buffer as previously described (37).
Cell Culture-Human RPE (ARPE-19) cells (American Type Culture Collection, ATCC, Manassas, VA) were cultured with DMEM/F12 medium containing 10% fetal bovine serum (FBS) and 1% antibiotic/antimycotic. While growing to 70% confluence, cells were starved overnight with low-serum (1% FBS) DMEM/F12 medium to reduce antioxidant pool in the medium and make the cells quiescent and then subjected to desired treatment.
Construction and Transduction of Adenoviruses-Recombinant adenovirus expressing human Nrf2 was constructed using the AdEasy system ((Agilent Technologies) according to the manufacturer's instructions. Briefly, full-length human Nrf2 gene was cloned using primers: forward primer ACCGCCAC-CATGGATTTGATTGACATAC; and reverse primer CTC-GAGCTAGTTTTTCTTAACATCTGGCT, and inserted into the KpnI-XhoI sites of the vector pShuttle-CMV. The resultant plasmid was co-transformed with pAdEasy-1 adenoviral vector into BJ5183 Escherichia coli-competent cells by electroporation. The recombinant adenoviral plasmids were then transfected into the packing cell line 293AD to generate recombinant adenoviruses. Construction of adenoviruses expressing human spliced XBP1 was described elsewhere (38). Large-scale preparation of adenoviruses was completed using 293AD cells. Harvested adenoviruses were tittered and purified using Adeno-X Maxi Purification kit (Clontech Laboratories, Mountain View, CA). ARPE-19 cells were transduced with adenoviruses as described previously (38). Adenovirus expressing LacZ was used as the control. After 24 h of transduction, cells were starved with 1% FBS DMEM/F12 medium followed by CSE treatment.
Small-interfering RNAs-Cells were transfected with siRNAs against human Nrf2, XBP1 or CHOP (Santa Cruz Biotechnology, Santa Cruz, CA) using Lipofectamine 2000 (Invitrogen, Carlsbad, CA) following the manufacturer's instruction and previously described (39). Lipofectamine 2000 or a nonsilencing scrambled siRNA with oligonucleotide sequence that does not recognize any known homology to mammalian genes was used as a negative control. The knockdown efficiency was determined by target protein levels using Western blotting.
MTT (3-(4,5-Dimethylthiazolyl-2)-2,5-diphenyltetrazolium Bromide) Assay-MTT assay was performed to assess the viability of ARPE-19 cells as described previously (38). Cells were seeded into a 96-well plate for each assay. Culture medium was refreshed, and 10 l of MTT reagent was added to 100 l of medium in each well. Cells were incubated at 37°C for 2 h. The precipitation was then dissolved by adding 100 l of detergent reagent into each well and incubated at room temperature for 4 h. The absorbance was measured at 570 nm with a microplate reader (Perkin Elmer, Waltham, MA), and data were analyzed.
TUNEL (Terminal Deoxynucleotidyl Transferase dUTP Nick End Labeling) Assay-TUNEL assay was performed to detect apoptosis using the In Situ Cell Death Detection Kit, TMR red (Roche Diagnostics Corp., Indianapolis, IN) following the manufacturer's protocol (40). Briefly, cells on coverslips were fixed with 4% paraformaldehyde (PFA) for 1 h, permeabilized in 0.1% citrate buffer containing 0.1% Triton X-100 for 2 min on ice, then incubated in TUNEL reaction mix containing nucleotides and terminal deoxynucleotidyl transferase (TdT) at 37°C for 1 h. Incubation without the TdT enzyme was conducted as negative control. After incubation, the coverslip was mounted onto a slice using mounting medium containing 4Ј-6-diamidino-2-phenylindole (DAPI; Vector Laboratories, Burlin-game, CA) and observed under an Olympus AX70 microscope (Olympus, Japan).
In Situ Trypan Blue Staining-After treatment, ARPE-19 cells were stained in situ with 0.04% Trypan Blue in DMEM/F12 medium for 15 min (41). Trypan Blue-stained cells and total cells were counted per 10ϫ field under an invert microscope (Zeiss, Germany). At least 5 fields were counted and averaged for each replicate, and results were obtained from three independent experiments.
Reverse Transcription Polymerase Chain Reaction (RT-PCR)-Total RNA from ARPE-19 cells was extracted using the E.Z.N.A. Total RNA Kit I (Omega Bio-Tek, Norcross, GA) according to the manufacturer's protocol. cDNA synthesis was performed using the Maxima First Strand cDNA Synthesis Kit (Fermentas, Glen Burnie, MD). PCR was performed using PCR FIGURE 1. Cigarette smoking extract induced ER stress and apoptosis in ARPE-19 cells. ARPE-19 cells were exposed to CSE (0.004 g/ml-320 g/ml) for up to 24 h. A, protein levels of ER markers or pro-apoptosis markers were determined by Western blotting in cells treated with CSE for 6 h ( p-eIF2␣, eIF2␣ and GRP78) or 12 h (ATF4 and CHOP). B, expression of p-eIF2␣ (normalized with eIF2␣), GRP78, ATF4 and CHOP were quantified by densitometry of Western blots. C, protein levels of ER stress markers, Nrf2, p-p38, and cleaved caspase-3 were determined by Western blotting in cells treated with 320 g/ml of CSE for 6 -24 h. D, expression of p-eIF2␣ (normalized with eIF2␣), GRP78, ATF4, CHOP, ATF6, Nrf2, p-p38, and cleaved caspase-3 were quantified by densitometry. E, apoptosis of ARPE-19 cells was examined by TUNEL assay after CSE treatment for 24 h. Scale Bar: 50 m. F, quantification of TUNEL-positive cells. G, cell viability of ARPE-19 cells with or without 320 g/ml of CSE treatment for 24 h. All data were expressed as mean Ϯ S.D., from three independent experiments. *, p Ͻ 0.05; **, p Ͻ 0.01 versus control.
Intracellular ROS and Mitochondrial Morphology Analysis-Levels of intracellular reactive oxygen species (ROS) were assessed using CellROX (Fluorescence Probes, Invitrogen). Briefly, cells were incubated with CellROX Deep Red Reagent (5 M) for 30 min (43) and then incubated with MitoTracker Green FM (Invitrogen) at 500 nM for another 30 min to determine morphologic changes of the mitochondria and the distribution of ROS (44). After three washes with PBS, cells were observed and imaged under a Zeiss LSM confocal microscope. ROS levels were measured fluorescence density and quantified using Image-J software.
Statistical Analysis-All quantitative data are presented as mean Ϯ S.D. Statistical analyses were performed using unpaired Student's t test for two group data and one-way analysis of variance (ANOVA) with Bonferroni's multiple compar-ison test for three groups or more. Differences were considered statistically significant at p Ͻ 0.05.

CSE Induces ER Stress and Apoptosis in ARPE-19 Cells-To
determine if CSE is sufficient to induce ER stress, ARPE-19 cells were exposed to a broad range of doses (0.004 -320 g/ml) of CSE for 24 h. This dose range overlaps with the plasma levels of water-soluble components of cigarette smoke in smokers (37), and moreover, the concentrations of nicotine in the CSE solutions (0.24 ng/ml-19.2 g/ml) overlap with plasma levels of nicotine found in smokers (45). Results showed that 80 g/ml-320 g/ml of CSE significantly increased expression of GRP78 and phosphorylation of eIF2␣, while CSE increased ATF4 and CHOP expression only at 320 g/ml (Fig. 1, A and B). To determine the time course of CSE's effects on ER stress, cells were exposed to 320 g/ml of CSE for 6 -48 h. Results showed that level of GRP78 and phosphorylation of eIF2␣ increased at 6 h and continued to increase until 24 h. The levels of ATF4 and CHOP increased from 6 h, peaked at 12 h, and slightly declined at 24 h. In addition, the level of activated ATF6, another major ER stress marker, was increased from 6 to 24 h. Similar changes were observed in Nrf2, a major transcription factor that up-regulates antioxidant genes, which may reflect increased oxidative stress upon CSE exposure (Fig. 1, C and D).
To determine whether CSE exposure induces apoptosis in RPE cells, activation of caspase-3, a key mediator of apoptosis, was examined by Western blot analysis of cleaved caspase-3. Results show that the level of cleaved caspase-3 significantly increased only after CSE (320 g/ml) treatment for 24 h (Fig. 1, C and D). Activation of p38 MAPK (P38), another potential signaling molecule involved in apoptosis, was observed CSEchallenged cells. TUNEL assay showed that the number of apoptotic cells was significantly increased after CSE treatment ( Fig.  1, E and F). Cell viability measured by MTT assay was markedly decreased by CSE exposure (Fig. 1G).

Antioxidant or Chemical Chaperone Ameliorates CSE-induced ER Stress and Apoptosis-N-Acetyl-cysteine (NAC) is an antioxidant that reduces oxidative stress induced by CSE (30).
To determine whether ER stress induced by CSE is mediated by oxidative stress, we pre-treated ARPE-19 cells with 1 mM NAC and then exposed the cells with 320 g/ml of CSE for 6 -24 h. Activation of the UPR and apoptosis was determined by Western blot analysis. Results show that NAC reduced the expression of p-eIF2␣, ATF4, CHOP, and alleviated activation of caspase-3 compared with CSE treatment only (Fig. 2, A and B), suggesting a potential role of oxidative stress in CSE-induction of ER stress. As expected, NAC also reduced the induction of Nrf2 by CSE (Fig. 2C). TUNEL assay and MTT demonstrated that pre-treatment with NAC protected ARPE-19 cells from apoptosis (Fig. 2, D and E) and alleviated the reduction of cell viability and mitigated cell death caused by CSE (Fig. 2, F  and G).
To decipher the importance of ER stress in CSE-induced apoptosis, TMAO, a chemical chaperone that facilitates protein folding thereby reducing ER stress, was applied in our study. ARPE-19 cells were pre-treated with TMAO and then challenged with CSE. In TMAO-pretreated cells, CSE induced much lower expression of ATF4, CHOP, and cleaved caspase-3 (Fig. 3, A and B). TUNEL assay confirmed the anti-apoptotic effect of TMAO (Fig. 3, C and D). These results indicate that ER stress plays a role in CSE-induced apoptosis in RPE cells.
Chemical Chaperone Decreases ROS Levels and Attenuates Mitochondrial Changes Caused by CSE-Evidence shows that CSE, which contains potent oxidants, leads to depletion of GSH and subsequent mitochondrial damage in ARPE-19 cells (30).
To determine reducing ER stress by chemical chaperone affects ROS generation and mitochondria, ROS levels and mitochondrial morphology were measured in ARPE-19 cells after CSE treatment for 6 h, with or without pretreatment of NAC or TMAO. After 6 h of CSE treatment, APRE-19 cells showed significantly increased ROS, most of which co-localized with mitochondria, suggesting that mitochondria is a major source of ROS generation in CSE-treated cells (Fig. 4A). Significant lower levels of ROS were observed in cells pre-treated with TMAO or NAC, which indicates an inhibitory effect of chemical chaperone on oxidative stress (Fig. 4B). Additionally, using MitoTracker staining of live cells, we observed changes in mitochondrial morphology after CSE treatment. As shown in Fig. 4A, mitochondria became fragmented and more diffusely distributed in CSE-treated cells. These changes were reduced in cells with NAC or TMAO treatment, indicating  FEBRUARY 27, 2015 • VOLUME 290 • NUMBER 9 that inhibiting ER stress may protect mitochondria and alleviate ROS generation.

The UPR Regulates Nrf2 and Protects RPE Cells
Activation of XBP1 Is Important for Cell Survival after CSE Treatment-As a master coordinator of the adaptive UPR, XBP1 plays a vital role in maintaining the ER function by inducing ER chaperone and ERAD protein expression. RT-PCR and Western blotting results showed that 320 g/ml or 400 g/ml of CSE treatment for 24 h stimulated XBP1 mRNA splicing (Fig.  5A) and increased its protein expression (Fig. 5B). To determine the role of XBP1 in CSE-induced cell apoptosis, ARPE-19 cells were transduced with adenovirus expressing spliced (active form) XBP1 (Ad-XBP1) or LacZ (Ad-LacZ) as control. Adenoviral transduced cells were then treated with 320 g/ml of CSE for 24 h. Western blotting showed that overexpressing XBP1 reduced the expression of CHOP, p-p38, and cleaved caspase-3 (Fig. 5, C and D). TUNEL assay (Fig. 5, E and F) and in situ Trypan Blue staining (Fig. 5, G and H) showed that CSE induced less apoptosis and cell death in Ad-XBP1-treated cells than in control cells. These results indicate that XBP1 overexpression attenuates ER stress and protects cells from apoptosis.
To elucidate the role of endogenous XBP1, siRNA was used to knock down XBP1 gene in ARPE-19 cells and the knockdown efficiency was examined by Western blotting. Due to the short half-life of XBP1 proteins (22 min for XBP1s and 11 min for XBP1u), MG132, a proteasome inhibitor, was used to inhibit protein degradation. Specifically, cells were treated with XBP1 siRNA or control, then exposed to CSE for 24 h. During the last 4 h of the treatment, 10 M of MG132 was added to culture medium. Our results show that protein level of XBP1s was reduced by 60% (Fig. 6, A and B). In XBP1 siRNA-treated cells, CSE induced less eIF2␣ phosphorylation, more p-p38 expression, and strongly increased caspase-3 activation (Fig. 6, C and  D). However, down-regulation of XBP1 did not affect CSE-induced CHOP expression (Fig. 6, C and D). CSE also induced more morphologic changes and shrunken cells (Fig. 6E), and more reduction in cell viability in the group of XBP1-knock- down (Fig. 6F). TUNEL assay (Fig. 6, G and H) and Trypan Blue staining (Fig. 6, I and J) showed that CSE induced more apoptosis and cell death in the cells treated with XBP1 siRNA, suggesting endogenous XBP1 has a protective role in RPE cells challenged with CSE.
CHOP Down-regulation Reduced Nrf2 Level and Exacerbated CSE-induced Cell Death-Compelling evidence shows that activation of CHOP is a key step in the execution of ER stressassociated apoptosis (14). Our study found that CSE induced CHOP expression in APRE-19 cells. To determine the role of CHOP in CSE-induced apoptosis, siRNA was applied to knock down CHOP expression. In CHOP siRNA-treated cells, CSEinduced CHOP expression was dramatically decreased compared with control cells (Fig. 7, A and B), and the expressions of CSE-induced Nrf2, p-eIF2␣, and p-p38 were also reduced (Fig.  7, A and B). Although we had expected that CHOP inhibition protects ARPE-19 cells form CSE induced cell apoptosis and cell death, in contrast to our expectation, CSE induced more  FEBRUARY 27, 2015 • VOLUME 290 • NUMBER 9 caspase-3 activation (Fig. 7, A and B), more morphologic changes and shrunken cells (Fig. 7C), more reduction of cell viability (Fig. 8D), more apoptosis (Fig. 7, E and F), and more cell death (Fig. 7, G and H) in CHOP-deficient cells compared with controls. These data indicate that although CHOP was generally considered to be a transcription factor involved in apoptosis, knockdown of CHOP may have an aggravating effect rather than a protective effect on CSE-induced apoptosis in RPE cells.

The UPR Regulates Nrf2 and Protects RPE Cells
Nrf2 Regulates CHOP Expression and Apoptosis in RPE Cells-Nrf2 is a central regulator of the anti-oxidant system, inducing expression of genes encoding phase II detoxification enzymes and antioxidant proteins (46). To determine whether down-regulation of Nrf2 expression by CHOP deletion is

The UPR Regulates Nrf2 and Protects RPE Cells
responsible for enhanced caspase-3 activation and apoptosis, we used siRNA to knock down the Nrf2 gene in ARPE-19 cells. We found that in the cells pre-treated with Nrf2 siRNA, CSE caused more robust CHOP and p-p38 expression and caspase-3 activation compared with mock controls (Fig. 8, A and B). In line with these changes, TUNEL and Trypan Blue staining assays showed more apoptotic cells (Fig. 8, C and D) and more dead cells (Fig. 8, E and F) in Nrf2 siRNA group after CSE treatment. These results suggest that Nrf2 down-regulation may contribute to aggravated apoptosis caused by CHOP inhibition.
To further validate the role of Nrf2 on CHOP expression and apoptosis induced by CSE, we overexpressed Nrf2 using adenovirus in ARPE-19 cells. Western blotting showed that Nrf2 overexpression remarkably reduced expression CHOP, p-p38, cleaved caspase-3 induced by CSE when compared with control adenovirus (Fig. 9, A and B). TUNEL assay showed that overexpression of Nrf2-protected cells from CSE-induced apoptosis (Fig. 9, C and D). In situ Trypan Blue staining showed decreased number of dead cells in Nrf2 overexpression group after CSE treatment (Fig. 9, E and F). Together, these data revealed a potent effect of Nrf2 on reducing CHOP expression and apoptosis in CSE-challenged RPE cells.

DISCUSSION
Cigarette smoking is the most important environmental risk factor for AMD, and has been confirmed to induce damage in  FEBRUARY 27, 2015 • VOLUME 290 • NUMBER 9 RPE cells (7)(8)(9)(10). Studies have shown that CSE induces ER stress markers such as phosphorylation of eIF2␣ and the nuclear form of ATF6 in lung lysates from mice (47), GRP78 and CHOP in human bronchial epithelial cells (hBEC) and in the lungs of mice exposed to cigarette smoke (32). Recent work by our group and others also suggests induction of ER stress by cigarette smoke in the RPE (34 -36), however, the mechanisms by which ER stress regulates signaling pathways implicated in stress response and apoptosis are not well understood. In the present study, we found that CSE increased the expression of GRP78 and p-eIF2␣, and induced ATF4 and CHOP expression in a time-and dose-dependent manner in cultured human RPE cells. The increased p-eIF2␣ reflects the activation of PERK in UPR, which indirectly inactivates eIF2␣ and inhibits mRNA translation (14). At the same time, p-eIF2␣ activates transcription factor ATF4. ATF4 induces its downstream transcription factor CHOP, which is considered to control genes involved in apoptosis (14). Our study shows that CSE induced apoptosis and reduced the cell viability in ARPE-19 cells, indicating that CSE induces cell apoptosis when persistent ER stress exits. We also found that CSE induced ATF6 and Nrf2, two protective factors expressed during ER stress, in ARPE-19 cells at 6 h of CSE exposure. This result may indicate an auto-protective mechanism of ARPE-19 cells in early stages of ER stress.

The UPR Regulates Nrf2 and Protects RPE Cells
CSE contains a large amount of free radicals that can induce oxidative stress (29, 48 -51). As described above, oxidative stress can induce ER stress. However, how oxidative stress affects CSE-induced ER stress in ARPE-19 cells is not clear. In our study, antioxidant NAC alleviated CSE-induced ER stress, protected RPE cells from apoptosis, and increased cell viability in ARPE-19 cells, suggesting that oxidative stress is involved in the induction of CSE-induced ER stress. We also found that chemical chaperone TMAO significantly attenuated CSE-induced ER stress, and protected ARPE-19 cells from CSE-induced apoptosis, indicating that ER stress is important to CSEinduced cell apoptosis. Previous studies have shown that CSE disturbs oxidative balance, including increased lipid peroxidation (4-HNE) and mitochondrial superoxide production, as well as decreased intracellular glutathione (GSH) in RPE cells (30). In the present study, we found that CSE increased ROS accumulation in ARPE-19 cells, confirming the CSE induced oxidative stress. The accumulated ROS partly co-located with mitochondria, causing mitochondria shrank and fragmented. Pre-treatment of cells with the antioxidant NAC or the chemical chaperone TMAO alleviated CSE-induced ER stress and ROS accumulation, and protected cells from apoptosis. Interestingly, our preliminary study indicates that inducing mitochondrial ROS generation was sufficient to activate the UPR signaling (data not shown). In addition, CSE induced a reduction in mitochondrial potential, which was partially prevented by NAC but not TMAO (data not shown). These results indicate that in RPE cells ER stress and mitochondria-associated oxidative stress are closely associated and interdependently regulate each other, both contributing to apoptosis induced by CSE.
The early stage of UPR is the predominant adaptive cytoprotective response against ER stress. Studies have shown that the transcription factor XBP1s activates UPR genes, such as chaperon p58IPK, to relieve ER stress (18,19). Although hydroquinone, a potent oxidant in cigarette smoke, suppressed XBP1mediated adaptive UPR signaling (36), we found that CSE increased the expression of UPR factor XBP1s. In our study we found that overexpression of XBP1s reduced the CSE-induced CHOP and p-p38, protecting ARPE-19 cell from CSE-induced apoptosis. In contrast, knockdown of XBP1 decreased the baseline level of p-eIF2␣ and CSE induced eIF2␣ phosphorylation, up-regulated P38 activation, and increased the sensitivity of ARPE-19 cell to CSE-induced cell apoptosis. However, knockdown of XBP1 did not induce more CHOP expression. Our results suggest that knockdown of XBP1 enhances the sensitiv- FIGURE 9. Overexpression of Nrf2 reduces CSE induced ER stress and protects RPE cells from apoptosis. ARPE-19 cells were transduced with Ad-Nrf2 or Ad-LacZ for 24 h, followed by CSE treatment for additional 24 h. A, protein level of Nrf2, CHOP, p-p38, and cleaved-caspase-3 were determined by Western blotting. B, semiquantification of Nrf2, CHOP, p-p38, and cleaved-caspase-3 (normalized with ␤-actin) by densitometry. C, apoptosis was examined by TUNEL assay after CSE treatment for 24 h. Scale Bar: 100 m. D, quantification of TUNEL-positive cells from C). E, cell death was examined using in situ Trypan Blue staining after CSE treatment for 24 h. Arrow shows Trypan Blue-stained cells. Scale Bar: 100 m. F, quantification of dead cells (Trypan Blue-stained cells). All data were expressed as mean Ϯ S.D., from three independent experiments. *, p Ͻ 0.05; **, p Ͻ 0.01 versus Ad-lacZ; #, p Ͻ 0.05; ##, p Ͻ 0.01, versus Ad-LacZ ϩ CSE. FEBRUARY 27, 2015 • VOLUME 290 • NUMBER 9 ity of ARPE-19 cells to CSE-induced apoptosis, which may act through an alternative pathway to CHOP. These results confirmed the protective effect of UPR factor XBP1s in cultured human RPE cells during CSE exposure.

The UPR Regulates Nrf2 and Protects RPE Cells
CHOP is considered a major pro-apoptotic gene induced by ER stress. Evidence demonstrates that activation of CHOP induces a variety of downstream genes involved in apoptosis, while deficiency of CHOP protects cells from ER stress-induced apoptosis (52)(53)(54)(55)(56). However, some recent studies have shown discrepant results. For instance, Esposito et al. showed that knock-out of CHOP exaggerated LPS-induced inflammation and kidney injury in mice, suggesting that CHOP inhibits the inflammatory response in renal cells and at the same time provides a protective effect against kidney injury (57). Chen et al. found that CHOP deficiency enhanced apoptosis in hippocampal cells and impaired memory-related behavioral performances in mice with tunicamycin treatment (58). Recently, Nashine et al. observed that CHOP deficiency did not protect but exacerbate rod photoreceptor death in retina degeneration mice, with up-regulation of p-eIF2a and decreased spliced XBP1, which indicated that CHOP may be a survival factor for rod photoreceptors in severe retinitis pigmentosa mice (59). Furthermore, Cano et al. showed that CHOP knockdown decreased cell viability in ARPE-19 cells when exposed to 500 g/ml CSE (35). Likewise, we found that CSE induced more apoptosis and cell injury in CHOP-deficient cells. These observations suggest that despite the pro-apoptotic effect of CHOP, a baseline level of CHOP seems to be required for photoreceptor and RPE cell survival. However, it remains unclear why loss of CHOP results in more cell death in stressed retina and RPE.
To address this question, we evaluated the expression of Nrf2, p-eIF2␣, and p-p38 in cells treated with siRNA against CHOP after exposure to CSE. Nrf2 is a central regulator of the anti-oxidant system, inducing expression of genes encoding phase II detoxification enzymes and antioxidant proteins to attenuate oxidative stress and protect cell survival (46). Recently, Lei Wang et al. found that Nrf2 was altered in human AMD specimens, and Nrf2 deficiency promoted cellular oxidative damage and a pro-inflammatory environment in CSE exposed RPE cells (60). In our study, we found that in CHOP knockdown cells, p-p38, a MAPK factor involved in cell death, was decreased, which may indicate that knockdown of CHOP to some extend alleviates cell damage through MAPK pathway; meanwhile, Nrf2 and p-eIF2␣ were significantly decreased. Overexpression of Nrf2 significantly reduced CHOP-dependent apoptosis in RPE cells; in contrast, knockdown of Nrf2 exacerbated CSE induced cell apoptosis, suggesting that CHOP knockdown exacerbated cell death at least partially through the down-regulation of Nrf2. In addition, the decreased p-eIF2␣ level may lead to a decreased capacity to slow down protein formation, which would increase the unfolded protein accumulation, thereby exacerbating ER stress and cell damage.
In summary, our study showed that CSE induces ER stress associated with oxidative stress and mitochondrial damage, playing an important role in apoptosis of RPE cells death. The CSE induced UPR and CHOP activation is essential for induction of Nrf2, which suppresses ER stress and protects against apoptosis. Activation of XBP1, the central regulator of the adaptive UPR, is also important for RPE survival during ER and oxidative stress. Thus, harnessing the protective factors such as XBP1 and Nrf2 may provide novel therapeutic targets to protect RPE cells from cigarette smoke-associated damage of RPE cells.