Glutamate-Induced NFκB Activation in the Retina

Abstract

purpose. To determine the distribution and glutamate-mediated activation of nuclear factor (NF) κB members in the retina and pan-purified retinal ganglion cells (RGCs) and to characterize steps in the signal transduction events that lead to NFκB activation.

methods. Retinal expression patterns and RGCs were evaluated for five NFκB proteins with the aid of immunohistochemistry. Retinal explants or RGCs were treated with glutamate with or without the presence of the NDMA receptor antagonist memantine, the calcium chelator EGTA, or a specific inhibitor for calcium/calmodulin-dependent protein kinase-II (CaMKII). Characterizations of NFκB activation were performed with the aid of electrophoretic mobility shift assays and supershift assays.

results. All five NFκB proteins were present in the retina and in the pan-purified RGCs. In response to a glutamate stimulus, all NFκB proteins except c-Rel were activated. P65 was unique in that it was not constitutively active but showed a glutamate-inducible activation in the retina and in the cultured RGCs. Memantine, EGTA, or autocamtide-2-related inhibitory peptide (AIP) inhibited NFκB activation in the retina. Furthermore, AIP significantly reduced the level of glutamate-induced degradation of IκBs.

conclusions. These data indicate that glutamate activates distinct NFκB proteins in the retina. P65 activation may be especially important with regard to RGC responses to glutamate given that its activity is induced by conditions known to lead to the death of these cells. The NMDA receptor-Ca²⁺-CaMKII signaling pathway is involved in glutamate-induced NFκB activation. Because AIP blocks the degradation of IκB, its regulation is clearly downstream of CaMKII.

The nuclear factor-κB (NFκB), a ubiquitously expressed transcription factor, is a critical regulator of many genes involved in inflammatory processes, cell differentiation, and apoptosis. The factor has been implicated in mechanisms that mediate cell survival and cell death. ¹ In mammals, the NFκB family comprises five members, p65 (RelA), RelB, c-Rel, p50/p105 (NFκB1), and p52/p100 (NFκB 2), which share an N-terminal Rel homology domain allowing dimerization, nuclear localization, and DNA binding. These proteins form homodimers and heterodimers and are retained in an inactive state in the cytoplasm through interaction with inhibitory molecules, called IκBs, which mask NFκB nuclear localization and DNA-binding domains. ² Activation of NFκB can be induced by multiple stimuli, including inflammation, infection, injury, and stress. On stimulation, IκB protein subunits are phosphorylated by IκB kinases (IKKs), followed by polyubiquitination and subsequent rapid degradation through the proteasome. This phosphorylation leads to the release of NFκB, which is then translocated to the nucleus, where it binds to DNA and activates the transcription of target genes. ³ Proapoptotic and antiapoptotic properties have been attributed to NFκB in neurons, ^{3 4 5} and the balance between cell death and survival in response to external stimuli may rely on the activation of distinct NFκB proteins ⁵ ; complete characterization of this has not yet been demonstrated for any cells in the retina. 

Retinal ischemia is a common clinical entity and has been widely studied because of its proposed relationship to, for example, anterior ischemic optic neuropathy, retinal and choroidal vessel occlusion, glaucoma, diabetic retinopathy, retinopathy of prematurity, and traumatic optic neuropathy. ⁶ All these diseases and disorders have been shown to lead to injury or loss of the retinal ganglion cells (RGCs), resulting in blindness. The mechanisms mediating RGC death are still not well understood, and multiple pathogenic mechanisms have been proposed. Glutamate excitotoxicity is one of the most studied models for inducing RGC death. This model is supported by a large body of literature showing that the level of glutamate is elevated in retinal ischemia and that excess glutamate plays a role in the pathogenesis of ischemic retinopathy. ^{6 7 8 9 10 11 12 13 14 15 16 17 18 19 20}  

Ischemic and excitotoxic stressors are some of the known initiators that activate NFκB in neurons. ^{21 22 23 24 25 26 27} For example, NFκB is activated in the RGCs in several model paradigms, including NMDA-induced retinal neurotoxicity (p65), ^{28 29} retinal ischemia and reperfusion injury (p65), ³⁰ diabetic retinopathy (p50 and p65), ³¹ and optic nerve transaction (p50 and p65). ^{32 33} However, the mechanisms underlying NFκB protein activation and cell death/survival signal transduction pathways after these types of injuries remain unclear or controversial. 

Studies have shown that glutamate stimulation can activate NFκB in a Ca²⁺-dependent manner. ^{34 35} Calcium/calmodulin-dependent protein kinase-II (CaMKII), an essential kinase mediating the Ca²⁺ message, has also been implicated in regulating NFκB activation. ^{35 36 37} This enzyme is downstream of glutamate receptors and responds to increases in intracellular Ca²⁺ resulting from the stimulation of NMDA receptors. Several studies in the past decade have implicated CaMKII in regulating cell death/survival responses in a variety of cell systems. ^{38 39 40 41} Inhibition of CaMKII activity with a specific inhibitor, autocamtide-2-related inhibitory peptide (AIP), protects retinal neurons from NMDA-induced retinal neurotoxicity. ⁴² Taken together, we postulate that the NFκB machinery is a prospective target for CaMKII. 

Because the proapoptotic and antiapoptotic properties of NFκB may rely on the activation of distinct NFκB proteins, the focus of the present study was to investigate which NFκB members are present and which are activated in response to excitotoxic stress in the retina, specifically in the RGCs. Subsequently, we investigated whether the NMDA receptor-CA²⁺- CaMKII pathway is indeed involved in regulating the activation of NFκB. 

Materials and Methods

All animals were handled in accordance with policies and procedures recommended by the Institutional Animal Care and Use Committee at the University of Louisville, and all procedures adhered to the ARVO Statement for the Use of Animals in Ophthalmic and Vision Research. 

Retinal Explant Culture

Retinal organ cultures were performed according to previously described protocols with some modification. ^{43 44} Briefly, Sprague-Dawley (SD) rats were killed at postnatal day (P) 14, and their eyes were enucleated. The anterior segment, vitreous body, and sclera were removed, and the retina was mounted immediately on 0.4-μM inserts (Millicell-CM; Millipore, Billerica, MA) with the photoreceptor side down. Retinal explants were cultured in 1.1 mL medium (Neuroabasal-A; Invitrogen, Carlsbad, CA) supplemented with 2% B27, 2% fetal bovine serum (FBS), 1 mM glutamine, and antibiotics. Considering the possible affects of ex vivo culture conditions on NFκB activation that may interfere with the glutamate-induced response, pilot experiments with the aid of electrophoretic mobility shift assay (EMSA) were performed to compare NFκB binding activation in retinas without glutamate treatment at 0, 2, 4, 6, and 20 hours in culture. Because dissection took only 1 to 2 minutes, retinal explants used immediately after dissection (0 hour in culture) for protein extraction should represent the basal level of NFκB in the in vivo condition. No significant change in NFκB activity was observed until 4 hours later in culture (data not shown). Therefore, retinal explants were treated immediately after dissection, with or without glutamate (2 or 5 mM) for 2 to 4 hours, in the presence or absence of CaMKII inhibitor AIP (20 μM; Calbiochem, La Jolla, CA), Ca²⁺ chelator EGTA (2 mM), NMDA-receptor antagonist memantine (20–100 μM; Tocris Cookson Inc., Ellisville, MO), or APMA-KA receptor antagonist DNQX (50 μM; Tocris Cookson Inc.). During treatment, retinal explants were maintained at 37°C in a humidified environment of 5% CO₂ and 95% air. The concentrations of glutamate were selected based on a review of the literature ^{45 46} and our pilot data (not shown) to overstimulate glutamate receptors. Six retinas were used at each time point for each condition. At the indicated times, retinal explants were fixed for sectioning and immunohistochemistry or processed on ice for nuclear and cytoplasmic protein extraction. 

RGC Culture

RGCs isolated from postnatal SD rat retinas were pan purified, as previously described by Barres et al. ^{47 48} Briefly, eyes of Sprague-Dawley rats (P6-P8) were enucleated and rinsed with Dulbecco phosphate-buffered saline (Invitrogen). Retinas were dissected under a microscope and dissociated with the aid of a dissociation kit (Papain Dissociation System; Worthington Biochemicals, Lakewood, NJ) at 37°C for 40 minutes to create a single-cell suspension. RGCs were isolated from this suspension with a sequential immunopanning protocol. ⁴⁷ Purified RGCs were seeded on poly-d-lysine/laminin-coated 12-mm glass coverslips at a density of 2 × 10⁴ RGCs per coverslip. Cells were maintained in B27-supplemented medium (Neurobasal; Invitrogen), containing bovine serum albumin (100 μg/mL), progesterone (60 ng/mL), insulin (5 μg/mL), pyruvate (1 mM), glutamine (1 mM), putrescine (16 μg/mL), sodium selenite (40 ng/mL), transferrin (100 μg/mL), triiodo-thyronine (30 ng/mL), brain-derived neurotrophic factor (BDNF; 50 ng/mL), ciliary neurotrophic factor (CNTF; 20 ng/mL), bFGF (10 ng/mL), forskolin (5 μM), inosine (100 μM), and antibiotics (Sigma-Aldrich, St. Louis, MO). RGCs were identified by cell marker expression, including Thy-1, and by their characteristic cell morphology. The purity of RGCs isolated by this sequential immunopanning was usually greater than 99%. Cultures were maintained at 37°C in a humidified environment of 10% CO₂ and 90% air. Cells in culture for 1 week were treated with 100 μM glutamate ⁴⁹ for 1 to 2 hours and then processed for immunocytochemistry. 

Immunohistochemistry

Expression patterns of the NFκB proteins were assessed in retina and purified RGCs with the aid of double-immunofluorescence labeling using specific antibodies against distinct NFκB members and Thy-1, a marker for RGCs. Whole eyes (P60) and retinal explants (P14) from SD rats were fixed with 4% paraformaldehyde for 2 hours at room temperature, followed by cryoprotection in 30% sucrose at 4°C overnight and sectioning (10 μM). Frozen sections were permeabilized using 0.2% Triton-X-100 (Sigma). Purified RGCs plated on poly-l-lysine/laminin–coated coverslips were fixed with cold acetone/methanol (1:1) at −20°C for 10 minutes. After blocking of nonspecific binding sites, tissue sections or cultured RGCs were incubated with primary antibodies overnight at 4°C. NFκB antibodies used were anti-p50 (H-119), anti-p52 (447), anti-p65(C-20), RelB(C-19), and c-Rel (N-466) polyclonal antibodies (Santa Cruz Biotechnology, Inc., Santa Cruz, CA). Anti-thy-1 was a monoclonal antibody (Chemicon International, Temecula, CA). Primary antibodies were visualized with Cy3-conjugated goat anti-mouse secondary antibody (Chemicon International) or with Alexa 488-conjugated goat anti-rabbit secondary antibody (Molecular Probes, Eugene, OR). Slides were mounted with anti-fade mounting medium (Vector Laboratories, Burlingame, CA) and viewed with the aid of a fluorescence microscope. Images were recorded with equal exposure conditions for each specific antibody. 

Electrophoretic Mobility Shift Assays

Nuclear proteins were extracted from retinal explants using a reagent kit (NE-PER Nuclear and Cytoplasmic Extraction Reagent kit; Pierce Biotechnology, Rockford, IL), according to the manufacturer’s protocol. Concentrations of all protein samples were determined by protein assay (Coomassie Plus Protein Assay; Pierce Biotechnology). Equal amounts of nuclear protein extracts were analyzed for NFκB binding activity with the aid of an EMSA kit (LightShift Chemiluminescent EMSA; Pierce Biotechnology) and a biotin-labeled κB oligonucleotide probe (5′-AGTTGAGGGGACTTTCCCAGGC-3′ [NFκB target underlined]; Panomics, Redwood City, CA). Briefly, 5 μg nuclear protein was combined with 20 fmol biotin-labeled κB probe in reaction buffer (1× binding buffer, 2.5% glycerol, 50 ng/μL poly (dI · dC), 1% NP-40, 2.5 mM dithiothreitol, and 0.5 mM EDTA) in a total volume of 20 μL for 20 minutes at room temperature. Competition with a 200-fold excess of unlabeled NFκB DNA probe was used to demonstrate the specificity of protein-DNA interactions. DNA-protein complexes were resolved on a 6% DNA retardation gel (Invitrogen), transferred to nylon membrane (Pierce Biotechnology), and cross-linked using 254 nm UV light. Biotin-labeled DNA was detected (Chemiluminescent Nucleic Acid Detection Module; Pierce Biotechnology). Relative intensities of the DNA-protein complex bands were estimated quantitatively with the aid of a computerized image analysis system (Alpha Innotech, San Leandro, CA) as integrated density values. 

In supershift experiments, antibodies specific for different members of the NFκB family were selected for their ability to interfere with DNA binding activity. Nuclear proteins were incubated with antibodies (3 μg) against different NFκB subunits overnight at 4°C before the addition of the other components of the reaction mixture. Incubation proceeded for another 20 minutes. Polyclonal anti-p50, anti-p52, anti-p65, anti-RelB, and anti-c-Rel antibodies were the same as used for immunohistochemistry. 

Western Blots

Samples containing equal amounts of cytoplasmic protein were obtained from retinal explants and separated on 10% SDS-PAGE gels and then transferred to polyvinylidene difluoride membranes (Millipore). The blots were blocked overnight at 4°C in 0.1% Tween-20 Tris-buffered saline solution containing 5% nonfat dry milk and then were incubated with anti-IκB-α or anti-IκB-β (Cell Signaling Technology, Inc., Danvers, MA). Antibody binding was detected with horseradish peroxidase-conjugated anti-rabbit (Chemicon International) secondary antibodies and enhanced chemiluminescence Western blotting detection reagents (Amersham Life Science, Buckinghamshire, UK). For quantitative assays, the density of the immunolabeled bands from three independent experiments was calculated with a computerized image analysis system (Alpha Innotech) as integrated density values, normalized to that of β-actin, and compared with that of controls, whose expression level was taken as 1. 

Statistical Analysis

All quantitative data from blots were expressed as mean ± SEM. At least three independent experiments with three to six determinates for each condition were performed. Student’s t-test was used for two-group comparisons. ANOVA was used for multiple comparisons, followed by Newman-Keuls paired comparison. P < 0.05 significance cutoff was used. 

Results

Expression of NFκB Members in Retina and RGCs

Expression patterns of NFκB proteins were investigated in retina and pan-purified RGCs with the aid of immunofluorescence labeling using antibodies specific for distinct members of the NFκB family. As shown (Fig. 1A) , all five members of the mammalian NFκB family were detected in the retina from sections taken from whole eyes. Expression patterns of individual members varied in the retina. Although p65 and c-Rel had the most restricted expressions, largely confined to the ganglion cell layer (GCL), the p50, p52, and RelB members were expressed more widely in retinal layers in addition to the GCL. 

Double-immunofluorescence labeling for NFκB and Thy-1 revealed some colocalization with various members of NFκB in the retinal GCL. These results were clarified through an examination of NFκB members in pan-purified RGCs. For example, the NFκB member p50 exhibited an apparent constitutive or nuclear localization in the GCL and inner nuclear layer (INL) of the retina (Fig. 1A) , and its nuclear localization in the GCL was confirmed with the aid of purified RGCs (Fig. 1B) . Thus, the RGCs contained constitutively active NFκB-p50; p52 expression appeared to be present in the nuclei or cytoplasm of some cells in the GCL of the retina, whereas its presence in the purified RGCs appeared more perinuclear. Perinuclear labeling appeared to colabel with Thy-1 in merged double-labeled cells in the retinal sections, suggesting an RGC cytoplasmic presence. Rel-B labeling was evident in a diffuse pattern throughout the GCL, inner plexiform layer (IPL), INL, and outer plexiform layer (OPL). The most intense labeling was observed in the inner part of the INL. It appeared that some labeling was nuclear, but there was also much cytoplasmic labeling. Perinuclear labeling was evident in the purified RGCs. P65 was detected mainly in GCL and IPL, showing cytoplasmic labeling that colocalized with Thy-1 staining. c-Rel showed a predominant expression pattern in the GCL, with some faint labeling in OPL. Expression patterns in the retina for NFκB members are presented in Table 1 . 

There were no differences in NFκB expression patterns between sections taken from the whole eye and sections taken from retinal explants (<4 hours in culture) or between retinas from P14 and adult animals (data not shown). In summary, these immunofluorescence data demonstrated the presence of all five NFκB proteins in the retina and in the pan-purified RGCs. Moreover, a significant constitutive presence of p50 was demonstrated in nuclei within the retina GCL and INL and in the purified RGCs. Although nuclear and perinuclear distributions of p52 and Rel-B were observed throughout the retina, p65 and c-Rel primarily had a cytoplasmic presence in GCL. 

Activation of NFκB in Retina in Response to Glutamate Treatment

To determine the steps in the signal transduction pathway for the activation of NFκB, retinal explants were treated with or without glutamate (2 and 5 mM) in the presence or absence of AIP (20 μM) for 4 hours, when ex vivo culture conditions caused no significant change in the level of NFκB activity compared with retinal explants at 0 hour (data not shown). Nuclear protein extracts were obtained, and their NFκB-binding activities were assayed by EMSA. As shown (Fig. 2A) , specific protein-DNA interactions (labeled lanes) were demonstrated by competition with a 200-fold excess of the unlabeled NFκB probe (unlabeled lanes). The two bands (upper and lower) that changed in response to glutamate treatment reportedly represent different NFκB dimers. ³⁵ A basic level of constitutive NFκB-binding activity was detected with a pan-NFκB probe in control retinal explants (lane 1; 4 hours without glutamate), which confirmed the immunolabeling data (nuclear labeling in fixed tissue). Glutamate at concentrations of 2 or 5 mM induced significant increases in the level of NFκB-probe binding activity (Figs. 2A , lanes 2, 3; 2B). Application of the CaMKII inhibitor AIP significantly reduced this glutamate-elicited NFκB activation (Figs. 2A , lane 4; 2B), which indicated an involvement of CaMKII in the activation of NFκB in some part of the retina. 

Degradation of IκB in Response to Glutamate Treatment

To confirm the involvement of CaMKII in the activation of NFκB, cytoplasmic extracts were prepared from retinal explants 1 to 2 hours after glutamate exposure with or without AIP (20 μM). Western blot analysis was performed with the aid of specific antibodies for IκB-α or IκB-β. Blots were analyzed with a densitometer. Glutamate-mediated activation of NFκB was associated with reduced levels of IκBα and IκBβ after 1 to 2 hours of exposure (Fig. 3) . In contrast, in the presence of AIP, reductions in the levels of IκBα and IκBβ were not evident. These results indicated that the glutamate-induced degradation of IκB was downstream of CaMKII; in the CaMKII-containing cells of the INL and GCL, this enzyme has an important role in the regulation of NFκB activity. 

Characterization of NFκB Activation Elicited by Glutamate in Retina

To investigate which distinct NFκB proteins are important components of the signaling machinery subsequent to glutamate stimulation, the molecular composition of NFκB complexes activated by glutamate was assessed with the aid of supershift assays. Antibodies specific for different members of the NFκB family were selected for their ability to interfere with NFκB probe-binding activity. Nuclear extracts obtained from control and glutamate-treated retinal explants were first incubated with specific antibodies against p50, p52, p65, RelB, or c-Rel and observed by EMSA. The p50, p52, and RelB antisera inhibited the formation of NFκB complex from the control retinal explants, but p65 and c-Rel antibodies did not modify the binding activity of NFκB (Figs. 4A , − lanes; 4B, Control panel). These data indicated that p50, p52, and RelB were involved in NFκB constitutive activity in the resting state. In glutamate-stimulated retinal explants, antibodies to p65, p50, p52, or RelB reduced NFκB binding activity, whereas antibodies to c-Rel did not show any modification to NFκB binding (Figs. 4A , + lanes; 4B, Glutamate panel). The result indicated that c-Rel was not involved in glutamate-elicited NFκB activation. In contrast, p50, p52, and RelB revealed constitutive and inducible activation. Of particular interest, p65 showed evidence of inducible activity only, implicating an important role for this member of the NFκB family in the signaling response to glutamate. In summary, c-Rel was not implicated in constitutive or induced activation, though it was present in the retina. The p65 member was implicated in inducible activity only, and the p50, p52, and RelB members were constitutively active but showed additional inducible activation in response to a glutamate stimulus. 

Activation of NFκB, Especially p65, in RGCs in Response to Glutamate Treatment

To investigate whether NFκB is activated in response to glutamate stimulation specifically in RGCs, we used immunolabeling with retinal explants and pan-purified RGCs to determine whether p65 was inducible in these cells. Retinal explants and purified RGCs were treated with or without glutamate for 2 to 4 hours, and the activation of NFκB was assayed by double immunolabeling with antibodies to p65 and Thy-1. Glutamate treatment resulted in a translocation of the p65 from the cytoplasm to the nucleus of cells in the GCL of the retinal explants (Fig. 5A) . This glutamate-induced translocation also occurred in pan-purified RGCs (Fig. 5B) . 

NMDA-Receptor and Ca²⁺ Signaling Involvement in the Activation of NFκB

The finding that the specific CaMKII inhibitor AIP reduced NFκB activation in response to glutamate stimulation (Fig. 2)led us to further investigate whether this NFκB activation was NMDA receptor and Ca²⁺ mediated. Retinal explants were treated with glutamate in the presence or absence of the noncompetitive NMDA receptor antagonist memantine, the AMPA-KA receptor antagonist DNQX, or the Ca²⁺ chelator EGTA. NFκB binding activity was assessed by EMSA with the use of retinal nuclear extracts. Blocking the NMDA receptor with memantine significantly inhibited NFκB activation (Fig. 6) . DNQX did not change NFκB binding activity (data not shown), suggesting that glutamate-induced NFκB activation was effected through stimulation of the NMDA receptor. Chelation of extracellular Ca²⁺ also inhibited NFκB activation (Fig. 6) . Together, these data indicated that the NMDA receptor-Ca²⁺-CaMKII signaling pathway was involved in glutamate-induced activation of NFκB. 

Discussion

The present study demonstrated the presence of all five NFκB proteins in the retina, though different patterns of expression for each protein were observed. Although NFκB p65 and c-Rel are primarily restricted to the GCL, p50, p52, and RelB are present in additional layers. The different expression patterns may reflect distinct cellular phenotypes in the retina. The presence of all five NFκB proteins, specifically in the RGCs, is also demonstrated with the aid of pan-purified RGCs. 

We used more than one technique to confirm the observation of constitutive activity of NFκB. Thus p50, p52, and RelB show basal activity with EMSA and can be seen in the nuclei of cells within the retina. P50 is the best example of this phenomenon because it is readily identified in the nuclei of cells in the GCL of the retina and in the nuclei of pan-purified RGCs. In contrast, c-Rel and p65 did not show evidence of constitutive activity. The finding of constitutive activity of NFκB in the retina is consistent with earlier indications in neurons of the hippocampus and cerebral cortex. ⁵⁰ It has been suggested that constitutive NFκB activity is the result of ongoing synaptic activity. ^{22 50 51} However, the demonstration that p65 did not show constitutive activity in the retina is not consistent with other studies that indicate p50/p65 is the major NFκB dimer functioning in synaptic transmission. ^{21 22 35 50} Whether the discrepancy was the result of cell- or tissue-type specificity is unknown. It has also been shown that constitutive activity of NFκB is required for neuronal survival in other central nervous system (CNS) locations, ⁵² but further studies are required to demonstrate such a role in retinal neurons. 

Results demonstrating that Rel-B and p52, in particular, are constitutively active in the retina are novel. Rel-B is unique in that it does not homodimerize; in addition, it is unable to heterodimerize with c-Rel or p65. ² Rel-B forms heterodimers with p100, p52, and p50, and Rel-B/p52 or Rel-B/p50 heterodimers have been implicated in constitutive activity in multiple tissues. ^{2 53 54 55} Therefore, these results in the retina are consistent. 

Earlier studies have shown that the loss of Rel-B results in increased inflammatory infiltration in multiple organs; this phenotype is exaggerated in the p50 knockout mouse, indicating that Rel-B and p50 cooperate in the regulation of genes that limit inflammation. ^{56 57} This could be one of the mechanisms underlying immunoprivilege in the CNS, including the retina, because the high level of RelB/p50 constitutive activation shown here might have endowed the retina with an inflammation-suppressive or an immunosuppressive microenvironment. ^{58 59} This is an area that can be explored further in the retina. Homodimers of p52 or p50, which lack a transactivation domain, have no intrinsic ability to drive transcription. In fact, binding of p52 or p50 homodimers to κB sites of resting cells leads to the repression of gene expression. ² Whether and under what conditions this occurs in the retina and its RGCs must be further studied. 

Retinal ischemia, in particular, has been associated with increased levels of retinal glutamate and, ultimately, cell death. In the models used here, glutamate treatment eventually leads to the death of the RGCs. ^{45 46 49} In response to glutamate treatment, p65, p50, RelB, and p52 are activated. It is to be especially noted that, among glutamate-activated NFκB proteins, p65 shows only inducible activity. This is further confirmed in purified RGC cultures. Indeed, previous studies have shown that the expression and activity of NFκB p65 increase in RGCs and INLs in retinal ischemia-reperfusion ³⁰ and NMDA-induced retinal neurotoxicity models. ^{28 29} Furthermore, studies on other CNS neurons also reveal that ischemic and glutamate stimuli primarily activate p65 and p50. ^{5 25 60 61} Together, these results may imply a specific and important but prospective role for p65 with respect to the death of RGCs. Based on the data presented here that p50 and p52 exhibit inducible activity, it is possible that p65/p50, p65/p52, or both are relevant complexes for further investigation in RGCs. In addition, our data indicate that the glutamate-activated dimers of RelB and p52, or RelB and p50, may also exist, though their roles remain to be further identified. Given that NFκB protein dimers are retained inactive in the cytoplasm by interaction with the inhibitory molecules IκBs, it is not surprising to demonstrate that the activation of NFκB mediated by glutamate correlates with degradation of IκBα and IκBβ in the retina. 

It has been reported that glutamate-induced NFκB is activated in a Ca²⁺-dependent manner ^{34 35} and that glutamate receptors (NMDA, AMPA, and kainate subtypes) ^{4 22 34 62} may be involved. As an essential kinase mediating the Ca²⁺ message, CaMKII has also recently been shown to play an important role in mediating NFκB activation in other neurons. ^{35 63} In the present study, we have shown an involvement of the NMDA receptor-Ca²⁺- CaMKII signaling pathway in NFκB activation in the retina. In addition, we have demonstrated that the inhibition of NFκB activity through treatment with AIP significantly reduces the level of glutamate-induced IκBα and IκBβ degradation. These indicate that IκB could be a direct substrate for CaMKII, or that some other substrate, such as IKKα or IKKβ, is downstream of CaMKII. This idea is supported by other studies showing that IKKα and IKKβ are phosphorylated by CaMKII. ^{64 65} To our knowledge, the results reported here provide the first evidence for an involvement of CaMKII in promoting IκB degradation and, therefore, regulation of NFκB activation in the retina in response to an excitotoxic stimulus. This part of the signaling pathway is present within the cell cytoplasm. Therefore, cytoplasmic CaMKII is seen as a key control point in the glutamate-induced activation of NFκB in retina, including RGCs, given that CaMKII is known to be present in cells of the INL and GCL. 

The regulation of neuronal survival or death by NFκB may depend on the activation of a distinct combination of subunits, resulting in the differential regulation of target genes and the induction of diverse genetic programs that dictate the fate of cells within the retina. For example, excitotoxic stimulation-induced activation of NFκB p65/p50 may switch on the expression of those κB-responsive genes involved in the control of neuronal cell death, including various proapoptotic genes such as p53 and c-Myc, or the Fas ligand and its receptor (FAS/CD95), which could mediate a cell death response, as reported elsewhere. ^{25 66 67} However, the inclusion of c-Rel as part of NFκB dimers can reportedly provide a neuroprotective effect. In this case, antiapoptotic genes such as manganese superoxide dismutase, Bcl-XL, and Bfl-1 are direct transcriptional targets of c-Rel protein. ^{68 69 70 71 72}  

Although we did not investigate the role of a specific NFκB protein and its target gene(s) that control the cell death/survival pathways, our study may provide some insight into the mechanisms underlying NFκB activation and neuronal death/survival responses. It has been shown that the activation of distinct NFκB subunits and proapoptosis/survival properties may be stimuli specific. ^{5 73 74} Although some studies suggest that NFκB activation is prosurvival for RGCs, these studies were conducted using nonexcitotoxic stimuli, such as optic nerve transaction ³² and serum deprivation. ⁷⁵ On the other hand, it is well documented that excitotoxic simulation induces NFκB (p65 and p50) activation and neuronal death, ^{25 66 67} including RGCs in retina. ^{28 29 30} Our findings reveal that p65 and p50 are activated but that c-Rel is not involved in the response to glutamate stimulation, suggesting that glutamate induces proapoptotic NFκB subunit activation. Taken together, our findings that AIP, an inhibitor of CaMKII, protects retinal neurons from NMDA-induced excitotoxicity ⁴² and inhibits glutamate-induced NFκB activation, as shown in the present study, could indicate that neurotoxic-glutamate-induced NFκB activation plays a role in mediating neuronal cell death in the retina. However, this must be confirmed through further assays. As for Rel-B and p52, which are also shown to be activated in the retina when subjected to a glutamate stimulus, their prospective roles in regulating retinal neuronal death/survival pathways are unknown. Further studies, with the aid of conditional knockouts or siRNA-knockdowns of specific NFκB proteins, are now needed to identify the particular roles of the distinctive NFκB protein in the regulation of death and survival pathways. Thus, future studies should seek to show how these distinct NFκB proteins, and combinations thereof, can affect proapoptotic, antiapoptotic, and prosurvival gene cascades. 

 Supported in part by National Institutes of Health Grants R01EY017594, P20RR016481, and P30ESO14443.

 Submitted for publication July 11, 2008; revised September 12, 2008; accepted December 16, 2008.

 Disclosure: W. Fan, None; N.G.F. Cooper, 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.

 

(A) Double-immunofluorescence labeling for NFκB (green) and Thy-1 (red) in fixed tissue sections of retina. NFκB presents a labeling pattern of cytoplasm, nuclei, or both. Colocalization (yellow) of NFκB and Thy-1 is present in the RGC layer. Although all five NFκB proteins were present in the retina and the RGC layer, the expression pattern varied, with p65 and c-Rel mostly restricted to the GCL and the other three members in additional layers. NFκB p50 exhibited significant constitutive nuclear localization. (B) RGCs were purified from postnatal rat eyes (P6-P8) using the two-step immunopanning method. Cells were cultured for 1 week before immunostaining for NFκB proteins. RGCs were identified by positive Thy-1 staining. All five NFκB members were present in RGCs. Labeling patterns for each NFκB protein in RGCs were similar in vitro and in vivo. Scale bars: (A) 50 μm, (B) 25 μm.

Table 1.

 

Expression of NFκB Proteins in Retina

	p50	p52	p65	Rel-B	c-Rel
GCL	+	+	+	+	+
IPL	−	+	+	+	−
INL	+	+	−	+	−
OPL	−	+	+/−	+	−
ONL	+	−	−	−	−
IS/OS	−	+	−	+	−

 Immunofluorescence labeling indicating presence (+) or absence (−) in retinal layers. IS/OS, inner segment and/or outer segment.

Figure 2.

 

Retinal explants were treated with or without glutamate (2 and 5mM) for 4 hours in the presence or absence of the CaMKII inhibitor AIP (20 μM). (A) Nuclear extracts from the retinas were analyzed by EMSA with the aid of biotin-labeled κB oligonucleotide probe (5′-AGTTGAGGGGACTTTTCCCAGGC-3′ [NFκB target underlined]). Competition with a 200-fold excess of unlabeled NFκB DNA probe demonstrated the specific protein-DNA interaction. The two bands (upper and lower) may represent different NFκB dimers. (B) Densitometric analyses of NFκB activation from EMSA results showing a significant increase in NFκB binding activity in retina in response to glutamate stimulation. AIP inhibited glutamate-induced NFκB activation. Values of binding activity were expressed as a fold change of control values (without glutamate treatment), which were taken as 1. Data are mean ± SEM from at least three independent experiments conducted in different retinal explants. *P < 0.05 compared with controls; **P < 0.05 compared with glutamate-treated groups (ANOVA).

Figure 3.

 

Effects of glutamate and the CaMKII inhibitor AIP on IκB degradation in retinas. Cytoplasmic extracts were prepared from retinal explants 2 hours after glutamate exposure, with or without AIP (20 μM), and were immunoblotted with specific antibody against IκB-α (A) or IκB-β (B). For quantitative assays, densities of the immunolabeled bands from three independent experiments were calculated with a computerized image analysis system as the integrated density values, normalized to those of β-actin and compared with those of controls, whose expression level was taken as 100%. *P < 0.01 compared with control or AIP-treated retinas (ANOVA).

Figure 4.

 

(A) EMSA and supershift analyses were performed in retinal explants with or without glutamate treatment (2 mM, 4 hours). The molecular composition of the NFκB complexes was investigated by incubating nuclear extracts in the presence of antibodies against p50, p65, p52, RelB, and c-Rel. (B) Densitometric analyses of NFκB activation from EMSA results. Binding activity values were expressed as fold change of control values (without glutamate treatment), taken as 1. Data are mean ± SEM of three independent experiments conducted in different retinal explants. *P < 0.05 compared with corresponding binding values obtained in the absence of an antibody (ANOVA). Although p50, p52, and RelB were implicated in constitutive NFκB activity in control retinal explants, they also showed inducible activation in response to glutamate treatment. In contrast to p50, p52, and RelB, the p65 protein showed only inducible activity. c-Rel was not involved in glutamate stimulation in retinal explants.

Figure 5.

 

(A) Retinal explants were treated with or without glutamate (2 mM, 4 hours). Double-immunofluorescence labeling for NFκB and Thy-1 plus DAPI staining for nuclei in fixed-retina sections showed that glutamate treatment caused increased immunolabeling and nuclear localization of p65. Colocalization of p65 and DAPI was evident in cells within the retinal ganglion cell layer in glutamate-treated retina (arrows). (B) Purified RGCs were treated with glutamate (100 μM, 2 hours). Glutamate treatment resulted in the appearance of p65 in the nuclei. Scale bars: (A) 50 μm, (B) 25 μm.

Figure 6.

 

Retinal explants were treated with or without glutamate (2 mM) for 4 hours, in the presence or absence of the Ca²⁺ chelator EGTA (2 mM) or the NMDA antagonist memantine (50 μM). (A) Nuclear extracts from the retinal explants were analyzed by EMSA. (B) Densitometric analyses of NFκB activation from EMSA results showed a significant reduction in glutamate-induced NFκB-binding activity by EGTA or memantine. Values of binding activity were expressed as a fold change of control values (without glutamate treatment), taken as 1. Data are mean ± SEM of three independent experiments conducted in different retinal explants. *P < 0.05 compared with controls. **P < 0.05 compared with glutamate-treated retinas (ANOVA).