Involvement of Double-Stranded RNA-Dependent Protein Kinase in ER Stress-Induced Retinal Neuron Damage

Abstract

purpose. To clarify whether the activation of double-stranded RNA-dependent protein kinase (PKR) participates in the cell death induced by endoplasmic reticulum (ER) stress, the authors used cultured retinal ganglion cells (RGC-5, a rat ganglion cell line transformed with the E1A virus) in vitro and the effect of a PKR inhibitor (an imidazolo-oxindole derivative) on N-methyl-d-aspartate (NMDA)–induced retinal damage in mice in vivo.

methods. In RGC-5 culture, cell damage was induced by tunicamycin (an ER stress inducer), and cell viability was measured by Hoechst 33342, YO-PRO-1, or propidium iodide (PI) double staining or by the resazurin-reduction test. Levels of glucose-regulated protein (GRP) 78/BiP, activating transcription factor 4 (ATF4), C/EBP-homologous protein (CHOP) and the phosphorylated form of PKR were analyzed by immunoblot. The PKR inhibitor and two siRNAs that recognize nonoverlapping sequences of rat PKR were tested for their effects on tunicamycin-induced cell death. In vivo, retinal cell damage was induced by intravitreal injection of NMDA (20 nmol/eye) in mice. To examine its effect in vivo, the PKR inhibitor (1 nmol/eye) was intravitreally injected with NMDA, and ganglion cell layer cell loss and inner plexiform layer thinning were evaluated 7 days after NMDA injection.

results. Treatment with tunicamycin at 1, 2, and 4 μg/mL for 24 hours increased the number of YO-PRO-1 and PI-positive (apoptosis or necrosis indicator) cells in a concentration-dependent manner. Immunoblotting analysis showed that tunicamycin at 2 μg/mL induced BiP, ATF4, and CHOP protein production and PKR phosphorylation. Both the PKR inhibitor (0.03–1 μM) and the PKR knockdown (using siRNA) inhibited tunicamycin-induced RGC-5 cell death. The same inhibitor also reduced NMDA-induced retinal damage in vivo. The PKR inhibitor reduced the tunicamycin-induced increase in CHOP but not that in BiP protein production.

conclusions. These results indicate that inhibiting PKR activation is neuroprotective against ER stress-induced retinal damage, suggesting that PKR activation may be involved in the mechanisms underlying ER stress-induced cell death.

Endoplasmic reticulum (ER) stress has recently been linked to the pathogenesis of several diseases, including vascular and neurodegenerative diseases such as stroke, Alzheimer disease, and Parkinson disease. ^{1 2 3} Little is known about any role of ER stress in retinal damage. 

Retinal ganglion cell (RGC) death is a common feature of many ophthalmic disorders, such as glaucoma, optic neuropathies, and of various retinovascular diseases (diabetic retinopathy, retinal vein occlusions). A variety of factors, including oxidative stress, ⁴ excitatory amino acids, ⁵ and nitric oxide, ⁶ have been reported to induce retinal cell death. These reports emphasize the importance of a better understanding of the precise mechanisms underlying retinal diseases. 

Recently, it has been reported that one of the proapoptotic proteins involved in ER stress-mediated apoptosis (tunicamycin-induced apoptosis) is a double-stranded RNA-dependent protein kinase (PKR), as identified using a randomized ribozyme library. ⁷ PKR, an interferon-induced protein kinase identified initially in a study of responses to viral infection, is activated by the extensive secondary structure of viral RNA. ⁸ On binding to double-stranded RNA, PKR is autophosphorylated, and it then increases the cellular sensitivity to apoptotic stimuli through a number of putative pathways, including the phosphorylation of eukaryotic initiation factor 2α (p-eIF2α). ^{9 10} Thus, PKR is involved in the apoptosis induced not only by viral infection but also by ER stress. Hence, the purpose of the present study was to examine whether a PKR inhibitor (an imidazolo-oxindole derivative that acts as an ATP-binding site-directed inhibitor of PKR) or PKR silencing (by means of siRNA) might inhibit the retinal neuronal death induced by either ER stress (tunicamycin) or N-methyl-d-aspartate (NMDA). 

Materials and Methods

All experiments were performed in accordance with the ARVO Statement for the Use of Animals in Ophthalmic and Vision Research and were approved and monitored by the Institutional Animal Care and Use Committee of Gifu Pharmaceutical University. 

Dulbecco modified Eagle medium (DMEM) was purchased from Sigma-Aldrich (St. Louis, MO). The drugs used and their sources were as follows: double-stranded RNA-dependent protein kinase (PKR) inhibitor {8-[imidazol-4-ylmethylene]-6H-azolidino[5,4-g]benzothiazol-7–1} ¹¹ and tunicamycin were obtained from Calbiochem (San Diego, CA) and Wako (Osaka, Japan), respectively; isoflurane was acquired from Nissan Kagaku (Tokyo, Japan); and fetal bovine serum (FBS) was obtained from Valeant (Costa Mesa, CA). 

Retinal Ganglion Cell Line Culture

Cultures of RGC-5 were maintained in DMEM supplemented with 10% FBS, 100 U/mL penicillin (Meiji Seika Kaisha, Ltd., Tokyo, Japan), and 100 μg/mL streptomycin (Meiji Seika Kaisha, Ltd.) in a humidified atmosphere of 95% air and 5% CO₂ at 37°C. RGC-5 cells were passaged by trypsinization every 3 days, as described in a previous report. ¹²  

Cell Viability Assay after Tunicamycin

RGC-5 cells were plated at a density of 1000 cells/well in 96-well culture plates (Falcon 3072; Becton Dickinson Labware, Oxnard, CA). Twenty-four hours later, cells were washed twice with DMEM and then immersed in DMEM supplemented with 1% FBS. One hour later, tunicamycin at 1 to 4 μg/mL was added to the media. Vehicle or PKR inhibitor at 0.03 to 1 μM was added to the media 1 hour before, concomitantly with, or 6 hours after tunicamycin treatment. Twenty-four hours after the addition of tunicamycin, cell viability was measured according to one of two methods. The first was a single-cell digital imaging-based method using fluorescent staining of nuclei. Briefly, cell death was assessed on the basis of combination staining with fluorescent dyes, namely, Hoechst 33342 (Molecular Probes, Eugene, OR) and YO-PRO-1 (YO; Molecular Probes) or propidium iodide (PI; Molecular Probes), observations made using an inverted epifluorescence microscope (IX70; Olympus, Tokyo, Japan). At the end of the culture period, Hoechst 33342 and YO-PRO-1 or PI dyes were added to the culture medium (8 μM, 0.1 μM, and 1.5 μM, respectively) for 30 minutes. Images were collected with a digital camera (Coolpix 4500; Nikon, Tokyo, Japan). At least 400 cells per condition were counted by a masked observer who used image-processing software (Image-J, version 1.33f; National Institutes of Health, Bethesda, MD). Cell mortality was quantified by expressing the number of YO-PRO-1- or PI-positive cells as a percentage of the number of Hoechst 33342-positive cells. 

As the second method for measuring cell viability, the effect on cell viability of tunicamycin treatment was quantitatively assessed by examining the fluorescence intensity changes induced by the cellular reduction of resazurin to resorufin. These experiments were performed in DMEM at 37°C. Cell viability was assessed by the use of 10% resazurin solution for 3 hours at 37°C, and then cells were examined for fluorescence at 560/590 nm. Fluorescence was expressed as a percentage of that in control cells in DMEM containing 1% FBS (after subtraction of background fluorescence). 

RNA Interference

For rat PKR and negative control siRNAs, we used a duplex (Stealth PKR RNAi; Invitrogen, Carlsbad, CA) and a control duplex (Stealth RNAi Negative Control Duplex; no. 12935 to 200; Low GC Duplex; Invitrogen), respectively. Sense and antisense strands of rat PKR siRNA were: sequence 1, 5′-UACUUUGUGUAUCUGGGAGUAUUUG-3′ (sense) and 5′-CAAAUACUCCCAGAUACACAAAGUA-3′ (antisense); sequence 2, 5′-AAUUCCAUUUGGAUAAAGAGGCACC-3′ (sense) and 5′-GGUGCCUCUUUAUCCAAAUGGAAUU-3′ (antisense). 

Transfection with siRNA In Vitro

RGC-5 cells were seeded at a density of 1000 cells/well into 96-well plates (for cell viability assay) or at a density of 5000 cells/well into 24-well plates (for assessment of the effects of PKR silencing) using the standard medium. Twenty-four hours later, cells were washed twice with medium (Opti-Mem I; Invitrogen) supplemented with 1% FBS without antibiotics and then were immersed in the same medium. Reagent (Lipofectamine 2000; Invitrogen) was used as the transfection agent. PKR and control siRNAs, each at a concentration of 10 nM, were transfected by incubation for 6 hours, according to the manufacturer’s instructions (Invitrogen). Subsequently, the medium containing siRNA and complex (Lipofectamine 2000; Invitrogen) was replaced by DMEM supplemented with 1% FBS. Forty-eight hours after the infection, tunicamycin was added to each well, and incubation continued for another 24 hours. 

In Vivo NMDA–Induced Retinal Damage

Male adult ddY mice weighing 36 to 43 g each (Japan SLC, Hamamatsu, Japan) were used for these experiments and were kept under controlled lighting conditions (12 hours light/12 hours dark). Anesthesia was induced with 3.0% isoflurane and maintained with1.5% isoflurane in 70% N₂O and 30% O₂, delivered through an animal general anesthesia machine (Soft Lander; Sin-ei Industry Co. Ltd., Saitama, Japan). Body temperatures were maintained between 37.0°C and 37.5°C with the aid of a heating pad and a heating lamp. Retinal damage was induced by injection (2 μL/eye) of NMDA (Sigma-Aldrich) at 10 mM dissolved in 0.01 M phosphate-buffered saline (PBS) injected into the vitreous body of the left eye under the anesthesia described. One drop of levofloxacin ophthalmic solution (Santen Pharmaceuticals Co. Ltd., Osaka, Japan) was applied topically to the treated eye after the intravitreous injection. The PKR inhibitor (1 nmol/eye) or an identical volume of vehicle (0.5% DMSO in PBS) was coadministered with NMDA injection at 20 nmol/eye. 

Immunoblotting

RGC-5 cells were lysed using a cell-lysis buffer (RIPA buffer [R0278; Sigma] with protease [P8340; Sigma] and phosphatase inhibitor cocktails [P2850 and P5726; Sigma]; and 1 mM EDTA). Cell lysates were solubilized in SDS-sample buffer, separated on 10% SDS-polyacrylamide gels, and transferred to PVDF membrane (Immobilon-P; Millipore, Bedford, MA). Transfers were blocked for 1 hour at room temperature (5% Blocking One-P; Nakarai Tesque, Inc., Kyoto, Japan) in 10 mM Tris-buffered saline with 0.05% Tween 20 (TBS-T) and were incubated overnight at 4°C with the primary antibody. Transfers were then rinsed with TBS-T and incubated for 1 hour at room temperature in horseradish peroxidase goat anti–rabbit or goat anti–mouse (Pierce, Rockford, IL) diluted 1:2000. Immunoblots were developed with chemiluminescence (Super Signal West Femto Maximum Sensitivity Substrate; Pierce) and visualized with the aid of a digital imaging system (FAS-1000; Toyobo Co., Ltd., Osaka, Japan). Primary antibodies used were as follows: mouse anti–BiP (BD Bioscience, San Jose, CA), rabbit anti–ATF4 (Santa Cruz Biotechnology, Santa Cruz, CA), rabbit anti–phospho-PKR (Abcam, Cambridge, MA), rabbit anti–PKR (Cell Signaling, Beverly, MA), mouse anti–CHOP (Santa Cruz Biotechnology), and rabbit anti–actin (Santa Cruz Biotechnology). 

Immunostaining

To clarify whether NMDA or tunicamycin induced PKR phosphorylation in the mouse retina in vivo and whether a PKR inhibitor would prevent PKR phosphorylation, immunocytochemistry was performed. At 12 or 24 hours after intravitreal injection of NMDA (20 mmol/eye) or tunicamycin (1 μg/eye), with or without PKR inhibitor (1 nmol/eye), eyes were enucleated, fixed in 4% paraformaldehyde overnight at 4°C, immersed in 20% sucrose for 48 hours at 4°C, and embedded in optimum cutting temperature (OCT) compound (Sakura Finetechnical Co., Ltd., Tokyo, Japan). Transverse, 10-μm–thick cryostat sections were cut and placed onto slides (Mas Coat). Sections were subsequently processed for immunocytochemistry using antibodies against phospho-PKR (Abcam) diluted 1:500 in immunoreaction enhancer solution A (CanGet signal immunostain; Toyobo Co., Ltd.). Sections were incubated with biotin-conjugated secondary antibody for 1 hour at room temperature and were visualized with an immunodetection kit (Vector M.O.M.; Vector, Burlingame, CA). 

Histologic Analysis of Mouse Retina

Seven days after the NMDA or tunicamycin injection, eyeballs were enucleated for histologic analysis. In mice under anesthesia produced by intraperitoneal injection of sodium pentobarbital (80 mg/kg), each eye was enucleated and kept immersed for at least 24 hours at 4°C in a fixative solution containing 4% paraformaldehyde. Six paraffin-embedded sections (3-μm thick) cut through the optic disc of each eye were prepared in a standard manner and were stained with hematoxylin and eosin. Retinal damage was evaluated as described previously, with three sections from each eye used for the morphometric analysis. Light microscope images were photographed, and the cell counts in the ganglion cell layer (GCL), at a distance between 350 and 650 μm from the optic disc, were measured on the photographs in a masked fashion by a single observer (Y.I.). Data from three sections (selected randomly from the six sections) were averaged for each eye, and the values obtained were used to evaluate the cell count in the GCL. 

Statistical Analysis

Data are presented as the mean ± SEM. Statistical comparisons were made, using a Student’s t-test or a Dunnett test (by means of Stat View version 5.0; SAS Institute Inc., Cary, NC). P < 0.05 was considered statistically significant. 

Results

Retinal Cell Death and Changes in Endoplasmic Reticulum Stress-Related Protein Induced by Tunicamycin

The time-course of the changes in cell morphology occurring after tunicamycin treatment at 2 μg/mL are shown in Figure 1A . An increase in nonadherent cells was observed 12 hours after tunicamycin treatment compared with that in nontreated control cells. Representative fluorescence stainings of nuclei—using Hoechst 33342, YO-PRO-1, and propidium iodide (PI) dyes—are shown in Figure 1B . Nontreated control cells displayed normal nuclear morphology and negative staining with YO-PRO-1 dye (which stains early apoptotic and later-stage cells) and PI dye (which stains late-stage apoptotic cells) (Fig. 1B , upper panels). Treatment with tunicamycin led to shrinkage and condensation of nuclei and to positive staining with each of these dyes (Fig. 1B , lower panels). 

Changes in the protein levels of BiP, pATF-4, and CHOP 24 hours after tunicamycin treatment at 1–4 μg/mL are shown in Figure 1C . BiP, ATF-4, and CHOP were all concentration dependently increased, whereas actin levels remained unchanged. 

Phosphorylation of a PKR in RGC-5 Cells Is Induced by Tunicamycin

Although no phosphorylated PKR was observed in vehicle-treated normal cells, marked and slight increases in phosphorylated PKR were noted at 6 and 24 hours, respectively, after tunicamycin treatment (Fig. 2A) . 

Effects of a PKR Inhibitor and PKR Silencing on the RGC-5 Cell Death Induced by Tunicamycin

Representative fluorescence stainings of nuclei with Hoechst 33342 and YO-PRO-1 dyes are shown in Figure 2B . Nontreated cells showed normal nuclear morphology and negative staining with YO-PRO-1 dye (which stains early-stage apoptotic and necrotic cells). Treatment with tunicamycin at 2 μg/mL for 24 hours led to shrinkage and condensation of nuclei and to positive staining with YO-PRO-1 dye. Treatment with a PKR inhibitor at 0.3 and 1 μM reduced the tunicamycin-induced morphologic changes in nuclei and the number of cells stained with YO-PRO-1. The number of cells exhibiting YO-PRO-1 fluorescence was counted, and the positive cells were expressed as the percentage of YO-PRO-1– to Hoechst 33342–positive cells (Fig. 2C) . After treatment with tunicamycin for 24 hours, the percentage of YO-PRO-1–positive cells was 8.2% ± 0.5% (n = 8), whereas in the control group (supplemented with 1% FBS) it was 0.5% ± 0.1% (n = 8). Treatment with the PKR inhibitor at 0.03 to 1 μM significantly reduced the increase in YO-PRO-1–positive cells induced by tunicamycin (in a concentration-dependent manner). Treatment with the PKR inhibitor at 1 μM, either concomitantly with or 6 hours after tunicamycin, protected against RGC-5 cell death at 24 hours after tunicamycin (Fig. 2D) . As shown in Figures 3A and 3B , each of the PKR siRNAs (#1 and #2) that recognize nonoverlapping sequences of rat PKR decreased the PKR protein expression and led to a significant level of resistance to the tunicamycin-induced decrease in cell viability. Some standard siRNAs have been reported to induce the activation of PKR/interferon pathways as nonspecific stress responses resulting in growth inhibition and cytotoxicity. ¹³ To avoid such undesirable responses, we used duplex and control duplex siRNA (Stealth; Invitrogen) with chemical modifications to enable the elimination of nonspecific stress response affecting the PKR/interferon pathways. ¹⁴ In the present study, negative control siRNA (Stealth; Invitrogen) at 10 nM did not increase PKR protein at 48 hours after treatment compared with the nontreated control (Shimazawa et al., unpublished data, 2006). This result indicates that under our experimental conditions, Stealth siRNA does not induce the activation of PKR/interferon pathways. 

Effects of a PKR Inhibitor on the Tunicamycin-Induced Increases in BiP and CHOP Proteins in RGC-5 Cells

To determine the mechanism underlying the action of tunicamycin, we examined the effect of the PKR inhibitor on the increases in BiP and CHOP proteins induced in RGC-5 cells by treatment with tunicamycin at 2 μg/mL for 24 hours. Tunicamycin induced BiP and CHOP proteins in such cells (Fig. 4) . Pretreatment with the PKR inhibitor at 1 μM significantly reduced the increase in CHOP protein but not that in BiP protein. On the other hand, treatment with the same PKR inhibitor alone (at 0.3 or 1 μM) had little effect on the production of these proteins (vs. vehicle-treated controls). Staurosporine (200 nM), an apoptosis inducer, did not increase BiP or CHOP protein, but it did induce apoptotic cell death in this condition. 

Phosphorylation of a PKR in Mouse Retina after NMDA or Tunicamycin

No phosphorylated PKR was observed in the vehicle-treated normal retina, but phosphorylated PKR was observed in both the GCL and the inner plexiform layer (IPL) at 12 and 24 hours after intravitreal injection of NMDA (Fig. 5) . Phosphorylated PKR was also observed in GCL and IPL at 12 and 24 hours after tunicamycin injection (Fig. 5) . The PKR inhibitor (1 nmol), when coadministered with NMDA or tunicamycin, inhibited the increases in phosphorylated PKR in mouse retina induced by NMDA and tunicamycin (Fig. 5) . 

Effect of a PKR Inhibitor on the Retinal Damage Induced by Intravitreal Injection of NMDA

Intravitreal injection of NMDA at 20 nmol/eye decreased the cell count in the GCL (vs. nontreated normal retinas; Figs. 6A 6B ). Coadministration of the PKR inhibitor at 1 nmol/eye significantly reduced the NMDA-induced loss of GCL cells (Figs. 6A 6B) . 

Discussion

In the present study, we demonstrated a close association between PKR and ER stress-induced retinal damage and showed that the inhibition of PKR reduced such damage. 

Nutrient deprivation and agents that cause unfolded or misfolded proteins in the ER can activate the ER stress response, though the cell normally survives the insult. ¹ However, excessive or prolonged ER stress can induce cell death, usually in the form of apoptosis. In the present study, we found that tunicamycin induced the ER stress-associated proteins BiP, ATF-4, and CHOP in cultured RGC-5 cells. In our previous study, the phosphorylation of eukaryotic translation initiation factor 2 kinase (eIF2α) was increased concomitantly with the increases in the expressions of BiP, ATF-4, and CHOP proteins from 6 hours onward. ¹⁵ BiP, ATF-4, and the phosphorylation of eIF2α increased time-dependently throughout the 24-hour tunicamycin treatment period, whereas actin levels remained unchanged. CHOP was first detected 6 hours after the addition of tunicamycin, and it persisted thereafter. On the other hand, a slight increase in cell death was observed 12 hours after tunicamycin treatment, and it was time-dependently increased thereafter. ¹⁵ BiP acts as an ER-resident molecular chaperone induced by ER stress, and this protein refolds the unfolded proteins, thereby tending to maintain homeostasis in the ER. ^{16 17} CHOP is a member of the CCAAT/enhancer-binding protein family that is induced by ER stress and participates in ER-mediated apoptosis; hence, CHOP may be a key molecule in retinal cell death. ¹⁸  

PKR is involved in the phosphorylation of eIF2α. ^{19 20} Overexpression of PKR leads to apoptosis. ²¹ In recent years, PKR has been reported to be involved in Alzheimer disease, ^{7 22 23 24} Huntington disease, ²⁵ Parkinson disease, ²⁶ and amyotrophic lateral sclerosis, ²⁷ indicating that PKR may be implicated in the neuronal damage induced by ER stress in these neurologic diseases. In the present study, PKR phosphorylation was observed in RGC-5 cells 6 and 24 hours after tunicamycin treatment. We therefore tested, both in vitro and in vivo, whether the activation of PKR might participate in the retinal neuron death induced by ER stress. In the in vitro study, both the PKR inhibitor and the knockdown of PKR (using siRNA) inhibited tunicamycin-induced RGC-5 cell death, and the inhibitor attenuated the tunicamycin-induced increase in CHOP protein but not the tunicamycin-induced BiP protein production. In our previous study, we reported that the same PKR inhibitor protects against the cell death induced by tunicamycin in a different cell line, SH-SY5Y cells. ²⁸ Furthermore, treatment with the PKR inhibitor at 1 μM, either concomitantly with or 6 hours after tunicamycin, protected against RGC-5 cell death 24 hours after tunicamycin. This result suggests that PKR may operate in the late phase. In this study, pretreatment with the PKR inhibitor at 1 μM reduced the expression of CHOP protein 24 hours after tunicamycin. However, treatment with the PKR inhibitor (1 μM) 6 hours after tunicamycin, when CHOP protein was being induced, also protected against RGC-5 cell death. This result suggests that CHOP production does not necessarily have to be decreased for cell survival to be enhanced. Indeed, pretreatment with the PKR inhibitor at 0.3 μM also inhibited tunicamycin-induced RGC-5 cell death, but it did not reduce the increase in CHOP protein. Although CHOP-deficient mice have been reported to show resistance to ER stress-induced cell death, a dimerization partner such as C/EBPβ is needed for the induction of such cell death. ²⁹ In addition, CHOP protein undergoes stress-inducible phosphorylation by stress-inducible members of the p38 mitogen-activated protein kinase (MAPK) family, and such phosphorylation is associated with an enhancement of transcriptional activation by CHOP. ³⁰ Takizawa et al. ³¹ reported that a dominant-negative mutant of PKR inhibited the apoptosis and the p38 MAPK activation induced by apoptosis signal-regulating kinase 1 (ASK1), a member of the MAPK kinase kinase (MAPKKK) family, which is activated by a variety of apoptosis-inducers. Both ASK1 and PKR are known to bind proteins associated with death receptors, such as tumor necrosis factor (TNF) receptor–associated protein 2 (TRAF2). ^{32 33} During ER-stress, ASK1 is recruited to oligomerized inositol-requiring enzyme-1 (IRE1) complexes containing TRAF2, thereby activating this kinase and causing downstream activation of c-Jun N-terminal kinase (JNK) and p38 MAPK. ³⁴ Thus, PKR may activate the ASK1-p38 MAPK/-JNK signaling pathways to execute apoptosis. Furthermore, aggregated β-amyloid peptide has been reported to activate PKR through its phosphorylation or cleavage through calcium release from the ER, with activation of caspase-8 and caspase-3 as upstream signals. ²³ In another possible mechanism, the eIF2α phosphorylation induced by activated PKR might result in an upregulation of CHOP, a proapoptotic transcription factor. In addition, excessive ER stress leads to activation of PKR-like ER kinase (PERK), just as it does to activation of PKR. ^{35 36} The luminal ER stress-sensing domains of PERK regulate its dimerization, and this leads to activation of its protein kinase activity under ER stress. Activation of PERK induces phosphorylation of eIF2α, contributing to a suppression of protein translation after the initiation of ER stress. Therefore, even when PKR activation is inhibited by a PKR inhibitor or by PKR downregulation (using siRNA), the phosphorylation of eIF2α may not be reduced. Accordingly, eIF2α may not be a target molecule through which activated PKR executes cell death, at least in ER stress. However, further studies will be needed to clarify the precise mechanisms. Regarding the specificity of the PKR inhibitor against PKR, we must consider the possibility of effects on other targets as a limitation. 

The present results were obtained in cultured RGC-5 cells, and it could be agreed that the protective effect observed in vitro cannot be extrapolated directly to animal models in vivo. However, we have already confirmed that intravitreal injection of NMDA leads to increases in X box binding protein (XBP-1) mRNA splicing and in BiP and CHOP proteins in the mouse retina, representing activation of the unfolded protein response (UPR) signaling pathway. ¹⁵ Moreover, expression of the CHOP gene is reportedly increased in the rat retina after intravitreal injection of NMDA. ³⁷ Furthermore, Awai et al. ³⁸ found that treatment with MK-801, an NMDA-receptor antagonist, inhibited the increases in CHOP mRNA and protein in the mouse retina observed after intravitreal injection of NMDA and that CHOP-deficient mice were resistant to NMDA-induced retinal damage. However, CHOP-deficient mice partially suppressed NMDA-induced retinal cell death; therefore, other pathways (e.g., leading to mitochondrial dysfunction) may be engaged in the induction of this cell death. These findings indicate that NMDA can cause ER stress in the retina and that the neurotoxicity induced by NMDA is caused in part by a mechanism dependent on the induction of CHOP protein through excessive ER stress. On the other hand, the neuroprotective effect of the PKR inhibitor on tunicamycin-induced RGC-5 cell death did not appear to depend on CHOP suppression. Although little is known about the precise mechanisms responsible for the NMDA-induced activation of ER stress, NMDA is known to cause intracellular Ca²⁺ overload and increased NO production, resulting in apoptotic cell death. ^{39 40} Several lines of study suggest that intracellular Ca²⁺ overload and excessive production of NO deplete Ca²⁺ in the ER, thereby resulting in ER stress. ^{41 42} Recently, Uehara et al. ⁴³ reported that in primary cortical culture, even mild exposure to NMDA induces apoptotic cell death through an accumulation of polyubiquitinated proteins and increases in XBP-1 mRNA splicing and CHOP mRNA, representing activation of the UPR signaling pathway. Furthermore, they found that NO induces S-nitrosylation of protein-disulfide isomerase (PDI), an enzyme that assists in the maturation and transport of unfolded secretory proteins and thereby helps to prevent the neurotoxicity associated with ER stress. S-nitrosylated PDI exhibits reduced enzymatic activity and induces cell death through the ER stress pathway. These mechanisms may contribute to the activation of ER stress in the retina after NMDA stimulation. Collectively, the results indicate that NMDA can cause ER stress in the retina and that the neurotoxicity induced by NMDA is caused in part by a mechanism dependent on ER stress. 

In the present study, the phosphorylation of PKR was increased in the inner retina in mice in vivo after intravitreal injection of either NMDA or tunicamycin. The PKR inhibitor, when coadministered with NMDA or tunicamycin, inhibited the increase in phosphorylated PKR. This result indicates that PKR may be activated in the mouse retina in vivo during ER stress. The PKR inhibitor also reduced NMDA-induced retinal damage in mice in vivo, but this was a partial effect. Possibly, the observed difference in potency between the in vitro and in vivo situations may be attributed to the poor distribution or rapid excretion in ocular tissues of the PKR inhibitor after its intravitreal injection. Taken together, our results indicate that ER stress plays a role in retinal ganglion cell death and that PKR activation forms part of the underlying mechanisms. 

In conclusion, we have identified a close association between PKR and ER stress–induced retinal damage, and our results suggest that PKR might be a good target in the search for better treatments for retinal diseases. 

 Supported by a donation from the late Emeritus Professor Akihide Koda and by a research grant from the Ministry of Education, Science, Sports, Culture of Japanese Government (Grant 17591848).

 Submitted for publication September 19, 2006; revised February 14 and March 19, 2007; accepted June 12, 2007.

 Disclosure: M. Shimazawa, None; Y. Ito, None; Y. Inokuchi, None; H. Hara, None

 The publication costs of this article were defrayed in part by page charge payment. This article must therefore be marked “advertisement” in accordance with 18 U.S.C. §1734 solely to indicate this fact.

 

Figure 1.

 

RGC-5 cell death and changes in ER-stress related proteins induced by tunicamycin. (A) Representative phase-contrast microscopy showing time-course of changes in RGC-5 cells after addition of tunicamycin at 2 μg/mL (upper panels, low power; lower panels, high power). Bar = 250 μm. (B) Representative fluorescence microscopy showing nuclear stainings for Hoechst 33342, YO-PRO-1, and PI 24 hours after tunicamycin treatment at 2 μg/mL. (C) Representative immunoblots showing protein levels (BiP, ATF4, CHOP, and actin) 24 hours after tunicamycin treatment at 1–4 μg/mL.

Figure 2.

 

Phosphorylation of a PKR and its inhibitory effect on tunicamycin-induced cell damage in RGC-5 cells. (A) Phosphorylation of PKR in RGC-5 cells is induced by tunicamycin. Representative band images show immunoreactivity toward phosphorylated PKR and total PKR. (B–D) Effects of a PKR inhibitor on the RGC-5 cell death induced by tunicamycin. (B) Representative fluorescence microscopy showing nuclear stainings for Hoechst 33342 and YO-PRO-1 after 24-hour tunicamycin treatment. The PKR inhibitor was added 1 hour before the tunicamycin. (C) Number of cells exhibiting YO-PRO-1 fluorescence was counted, and positive cells were expressed as the percentage of YO-PRO-1 to Hoechst 33342. Each column represents the mean ± SEM (n = 8). **P < 0.01 versus tunicamycin alone (Dunnett test). (D) The PKR inhibitor at 1 μM was added 1 hour before, concomitantly with, or 6 hours after tunicamycin at 2 μg/mL. Number of cells exhibiting YO-PRO-1 fluorescence was counted, and the positive cells were expressed as the percentage of YO-PRO-1 to Hoechst 33342. Each column represents the mean ± SEM (n = 8). **P < 0.01 versus tunicamycin alone (Dunnett test).

Figure 3.

 

Effect of PKR siRNAs on tunicamycin-induced RGC-5 cell death. (A) Representative band images show immunoreactivities toward PKR and actin. PKR protein was decreased by 48-hour treatment with one of two PKR siRNAs. (B) RGC-5 cells were transiently transfected with control siRNA or one of two PKR siRNAs. After 48 hours, cells were treated with 2 μg/mL tunicamycin and maintained in this condition for another 24 hours. Each column represents the mean ± SEM (n = 8). *P < 0.05 and **P < 0.01 versus control siRNA-treated group (Dunnett test).

Figure 4.

 

Effect of a PKR inhibitor on the tunicamycin-induced increases in BiP and CHOP proteins in RGC-5 cells. (A) Representative band images show immunoreactivities against BiP and CHOP. Quantitative analysis of band densities for BiP (B) and CHOP (C). Data are expressed as mean ± SEM (n = 4) of values (in arbitrary units) obtained from each band. **P < 0.01 versus tunicamycin alone (Dunnett test).

Figure 5.

 

Phosphorylation of a PKR in mouse retina is induced by NMDA and by tunicamycin. Retinal cross-sections were labeled with antibody against phosphorylated PKR. Phospho-PKR–like immunoreactivity was increased in the inner retina 12 and 24 hours after intravitreal injection of NMDA (20 nmol/eye) or of tunicamycin (1 μg/eye). Coadministration of PKR inhibitor (1 nmol/eye) with the NMDA or tunicamycin inhibited the increase in phosphorylated PKR. Bar = 25 μm.

Figure 6.

 

Effect of a PKR inhibitor on the mouse retinal damage induced by intravitreal injection of NMDA. (A) Representative photographs showing nontreated normal retina, NMDA-injected vehicle-treated retina, and NMDA-injected PKR inhibitor-treated retina. PKR inhibitor at 1 nmol/eye or vehicle was coadministered with the NMDA (20 nmol/eye). Quantitative analysis of cell number in the GCL (B). Each column represents the mean ± SEM (n = 12). **P < 0.01 versus NMDA plus vehicle-treated group (Student’s t-test). Bar = 25 μm.

The authors thank Neeraj Agarwal (Department of Pathology and Anatomy, UNT Health Science Center, Forth Worth, TX) for the kind gift of RGC-5.