Photo-oxidative Stress Down-modulates the Activity of Nuclear Factor-κB via Involvement of Caspase-1, Leading to Apoptosis of Photoreceptor Cells*

The mechanisms of photoreceptor cell death via apoptosis, in retinal dystrophies, are largely not understood. In the present report we show that visible light exposure of mouse cultured 661W photoreceptor cells at 4.5 milliwatt/cm2 caused a significant increase in oxidative damage of 661W cells, leading to apoptosis of these cells. These cells show constitutive expression of nuclear factor-κB (NF-κB), and light exposure of photoreceptor cells results in lowering of NF-κB levels in both the nuclear and cytosolic fractions in a time-dependent manner. Immunoblot analysis of IκBα and p50, and p65 (RelA) subunits of NF-κB, suggested that photo-oxidative stress results in their depletion. Immunocytochemical studies using antibody to RelA subunit of NF-κB further revealed the presence of this subunit constitutively both in the nucleus and cytoplasm of the 661W cells. Upon exposure to photo-oxidative stress, a depletion of the cytoplasmic and nuclear RelA subunit was observed. The depletion of NF-κB appears to be mediated through involvement of caspase-1. Furthermore, transfection of these cells with a dominant negative mutant IκBα greatly enhanced the kinetics of down modulation of NF-κB, resulting in a faster photo-oxidative stress-induced apoptosis. Taken together, these studies show that the presence of NF-κB RelA subunit in the nucleus is essential for protection of photoreceptor cells against apoptosis mediated by an oxidative pathway.

The mechanisms of photoreceptor cell death via apoptosis, in retinal dystrophies, are largely not understood. In the present report we show that visible light exposure of mouse cultured 661W photoreceptor cells at 4.5 milliwatt/cm 2 caused a significant increase in oxidative damage of 661W cells, leading to apoptosis of these cells.

These cells show constitutive expression of nuclear factor-B (NF-B), and light exposure of photoreceptor cells results in lowering of NF-B levels in both the nuclear and cytosolic fractions in a time-dependent manner. Immunoblot analysis of IB␣ and p50, and p65
(RelA) subunits of NF-B, suggested that photo-oxidative stress results in their depletion. Immunocytochemical studies using antibody to RelA subunit of NF-B further revealed the presence of this subunit constitutively both in the nucleus and cytoplasm of the 661W cells. Upon exposure to photo-oxidative stress, a depletion of the cytoplasmic and nuclear RelA subunit was observed. The depletion of NF-B appears to be mediated through involvement of caspase-1. Furthermore, transfection of these cells with a dominant negative mutant IB␣ greatly enhanced the kinetics of down modulation of NF-B, resulting in a faster photo-oxidative stress-induced apoptosis. Taken together, these studies show that the presence of NF-B RelA subunit in the nucleus is essential for protection of photoreceptor cells against apoptosis mediated by an oxidative pathway.
Nuclear factor-B (NF-B) 1 is a widely distributed transcription factor that plays a role in the regulation of a number of cellular and viral genes involved in early defense reactions in higher organisms (1). NF-B exists in an inactive form bound to the inhibitory protein IB␣ or IB␤ (2)(3)(4). Treatment of cells with inducers such as lipopolysaccharide, interleukin-1, and tumor necrosis factor-␣ (TNF-␣), generally result in degradation of IB proteins. This releases NF-B of its inhibitory constraint, facilitating its translocation to the nucleus, resulting in regulation of expression of genes encoding cytokines, hematopoietic growth factors, and cellular adhesion molecules. NF-B exhibits its DNA binding activity in its dimeric form, and the most commonly occurring dimer is that of the p50 and the p65 (RelA) subunits. NF-B has been shown to be constitutively active in several cell types, including B cells (5), thymocytes (6), and neurons (7).
Most of the earlier studies on NF-B focused on its role in immunological and inflammation responses (1,8,9). Recent reports suggest that NF-B is also activated by oxidative signaling (10 -13). It has been suggested in many of these studies that reactive oxygen intermediates (ROI) may be involved in the activation of NF-B. Another area of research, where NF-B involvement is gaining momentum, is the regulation of apoptosis. One of the earliest significant observations in this direction was made by Beg et al. (14), who demonstrated extensive apoptosis of liver cells leading to embryonic death of mice lacking the RelA subunit. Subsequent work by Beg and Baltimore (15) demonstrated that treatment of RelA-deficient (RelA Ϫ/Ϫ ) mouse fibroblasts and macrophages, with TNF-␣, resulted in a significant reduction in cell viability. Along similar lines, Wang et al. (16), Van Antwerp et al. (17), and Liu et al. (18) showed a role of NF-B in suppression of TNF-␣induced apoptosis. There is also evidence of pro-apoptotic aspects of RelA activity. For instance, it was shown that serum starvation of 293 cells causes cell death accompanied by the activation of RelA containing NF-B (19).
In several experimental models of retinal dystrophies, including certain forms of retinitis pigmentosa, photoreceptor cells of the retina have been shown to undergo apoptosis (20 -23). A number of studies have identified the primary mutations involved in the etiology of these diseases (24). These studies, however, do not completely explain the ultimate phenotypic manifestation of the disease namely apoptosis of photoreceptor cells. The difficulty in studying the disease process of retinal dystrophies is further compounded by non availability of a homogenous permanent photoreceptor cell line, since retina is a complex tissue of multiple cell types. In the current studies we have used a transformed mouse photoreceptor cell line 661W. 2 In vivo studies have shown that exposure of rats to constant light results in apoptosis of photoreceptor cells (25)(26)(27)(28)(29). Even moderate intensities of light exposure have been shown to damage the retinas of rats (30). Since then, light has been extensively studied as an initiator of retinal cell death in a number of in vivo (31)(32)(33) and in vitro experimental conditions.
In the current study, we assessed the contribution of an oxidative stress paradigm to the propensity of photoreceptor cells to proceed to cell death via apoptosis, using cultured photoreceptor cells. We provide evidence in this paper that visible light exposure to photoreceptor cells results in oxidative damage leading to apoptosis via down-modulation of NF-B. NF-B, which was constitutively expressed in the 661W cells, was found to be progressively down-regulated upon exposure of the cells to light. By immunocytochemistry using NF-B RelA antibody, the NF-B activity appeared to be localized both in the nucleus and cytoplasm of dark-exposed 661W cells. Upon exposure to light the nuclear and cytoplasmic NF-B RelA immunolabeling was largely diminished in these cells. Pretreatment of the cells with various antioxidants prevented to a great extent the down-modulation of NF-B and also protected the cells from apoptosis. Furthermore, transient transfection of the 661W cells with a dominant negative IB␣⌬N (super-repressor) caused a rapid decline in NF-B binding activity in the cells, leading to a faster kinetics of photo-oxidative stressinduced apoptosis. Down-modulation of NF-B in these cells appears to be mediated by caspase-1. Our results suggest that NF-B, which is constitutively expressed in 661W photoreceptor cells, undergoes degradation when subjected to oxidative stress leading to apoptosis of the photoreceptor cells. Thus, the presence of NF-B in the nucleus is essential for photoreceptor cell survival and protection against oxidative stress induced apoptosis.

EXPERIMENTAL PROCEDURES
Materials-The following materials were purchased from the indicated sources: fetal bovine serum from JRH Biosciences, Lenexa, KS; paraformaldehyde and H 2 O 2 from EM Sciences, Gibbstown, NJ; HEPES, phenylmethylsulfonyl fluoride, ALLN, and dithiothreitol from Sigma; poly(dI-dC)⅐poly(dI-dC) from Amersham Pharmacia Biotech; and polynucleotide kinase from New England Biolabs, Beverly, MA.
Antibodies-p50 subunit of NF-B, a goat polyclonal IgG; p65, of NF-B, a rabbit polyclonal IgG; and IB␣ rabbit polyclonal IgG were from Santa Cruz Biotechnology, Santa Cruz, CA. GAPDH (chicken anti-rabbit GAPDH immunoaffinity-purified monospecific antibody) was kindly supplied by Drs. Glaser and Cross (34). ␤-Tubulin, a mouse monoclonal antibody, was from Sigma. Peroxidase-labeled secondary antibodies either anti-rabbit IgG or anti-mouse IgG were from Kirkegaard and Perry Laboratories Inc., Gaithersburg, MD. Anti-cyclin D1 antibody was against amino acids 1-295, which represents full-length cyclin D1 of human origin, and was obtained from Santa Cruz Biotechnology. Fluorescein isothiocyanate-labeled anti-rabbit IgG from Vector Laboratories, Burlingame, CA.
Cell Culture-The 661W cells were originally isolated from a transgenic mouse line expressing the construct HIT1 comprising of SV40 T-antigen driven by the human interphotoreceptor retinol binding protein promoter (35). The construct HIT1 resulted in SV40 T-antigen expression and retinal and brain tumors. 661W cells are routinely grown in complete medium consisting of Dulbecco's modified Eagle's medium containing 10% fetal bovine serum and 1% penicillin/streptomycin, at 37°C in a humidified atmosphere of 5% CO 2 and 95% air.
Light Exposure of the Cells-The 661W cells were seeded either in 35/60/100-mm tissue culture dishes or on round coverslips kept in 35-mm dishes and exposed to fluorescent visible light at 4.5 milliwatt/ cm 2 for varying durations up to 4 h at 37°C in tissue culture. The accompanying control cells were shielded from light for similar intervals and left in similar conditions as the cells in light-exposed paradigm.
3Ј End Labeling of Fragmented DNA by Terminal Deoxynucleotidyl Transferase-mediated Fluoresceinated dUTP Nick End Labeling (TUNEL)-The TUNEL procedure as described by Gavrieli et al. (36) was employed to study apoptosis, using a commercially available apo-ptosis kit (in situ cell death detection kit, Boehringer Mannheim) as per the supplier's instructions.
Measurements of Membrane Lipid Peroxidation and Reduced Glutathione (GSH) Levels-The membrane lipid peroxidation of light-exposed cultured cells was studied by measuring the malonyldialdehyde levels by a colorimetric method involving thiobarbituric acid adduct formation (37). The GSH levels in light-exposed cells was studied by using the 5,5Ј-dithiobis(2-nitrobenzoic acid) reagent (38).
Immunoblot Analysis-Protein extracts from 661W cultured cells exposed to light were subjected to immunoblot analysis (39,40) using specific antibodies for IB␣, p50, and RelA subunit of NF-B at 1:500 dilution. Cytoplasmic extracts were used for IB␣ analysis, whereas nuclear extracts were used to study RelA and p50 subunits of NF-B. Control blots were run using total cellular extracts and an antibody to GAPDH at 1:1000 dilution. The binding of primary antibodies was detected by using peroxidase labeled appropriate secondary antibodies, which were detected by using diaminobenzidine as substrate.
Immunolocalization Studies-The 661W cells were exposed to light and fixed in 4% paraformaldehyde. The immunofluorescence for p65 subunit of NF-B was done by using a specific antibody against p65 and a fluorescein isothiocyanate-labeled goat anti-rabbit secondary antibody (20). The immunofluorescent cells were photographed using a Nikon Microphot-FXA photomicroscope.
Preparation of Cytoplasmic and Nuclear Extracts-The 661W cells were exposed to light for the desired amount of time, and the nuclear and cytoplasmic extracts were prepared (41). Briefly, the cells were suspended in 100 l of buffer C (10 mM HEPES, pH 7.9, 1.5 mM MgCl 2 , 10 mM KCl, 10% glycerol, 1 mM dithiothreitol, 0.5 mM phenylmethylsulfonyl fluoride) and incubated on ice for 15 min. 3 l of 10% Nonidet P-40 was added to the suspension and briefly vortexed. Following this, the nuclei were pelleted by centrifugation at low speed. The supernatant (cytoplasmic extract) was collected and stored at Ϫ80°C. The nuclear pellet was resuspended in 70 l of buffer D (20 mM HEPES, pH 7.9, 400 mM NaCl, 1.5 mM MgCl 2 , 0.2 mM EDTA, 20% glycerol, 1 mM dithiothreitol, 0.5 mM phenylmethylsulfonyl fluoride). The suspension was incubated for 20 min at 4°C followed by a centrifugation at 8000 g for 5 min. The supernatant containing the nuclear protein extract was transferred to a fresh microcentrifuge tube and stored at Ϫ80°C. Protein concentrations of the cytoplasmic and the nuclear extracts were measured with a detergent-compatible Protein Assay Kit (Bio-Rad), using bovine serum albumin as a standard.
Electrophoretic Mobility Shift Assays (EMSA)-A double-stranded oligonucleotide containing the NF-B DNA-binding consensus sequence, 5Ј-AGT TGA GGG GAC TTT CCC AGG C-3Ј, and a doublestranded mutant oligonucleotide, 5Ј-AGT TGA GGC GAC TTT CCC AGG C-3Ј (Santa Cruz Biotechnology) were used to study the DNA binding activity of NF-B by EMSA as described (42). For supershift assay, 4 g of nuclear extract was incubated with 1 g of antibodies for 30 min at room temperature and analyzed by EMSA.
Transfection of 661W with the Dominant Negative IB␣-The dominant negative IB␣ (super-repressor) construct IB␣⌬N was obtained from Dr. Dean Ballard, Vanderbilt University, Nashville, TN (43). IB␣⌬N is a deletion mutant in which the N-terminal 36 amino acids are deleted from the IB␣ protein. The 661W cells were transiently transfected with the construct using the LipofectAMINE reagent (Life Technologies, Inc.), as per the manufacturer's instructions using 8 l of LipofectAMINE and 5 g of the IB␣ super-repressor plasmid DNA for 1 ml of transfection mix. The untransfected control cells were treated in a similar manner except for the exclusion of the plasmid DNA. The transfected cells and their controls were used 48 h post-transfection for making either total cellular extract for immunoblot analysis or cytoplasmic and nuclear extracts for EMSAs and for TUNEL assay as described above.

Membrane Lipid Peroxidation and Depletion of Glutathione
Are Observed in Cultured Photoreceptor Cells Exposed to Light-The 661W cells were exposed to light for up to 4 h, and membrane lipid peroxide formation and GSH levels were measured. There was almost a 2-fold increase in malonyldialdehyde formation following light exposure, as compared with controls ( Table I) and inclusion of NAC in the medium before light exposure of cultured cells, prevented the increase in MDA levels (Table I). These results indicate that photic injury to photoreceptor cells occurs due to a possible involvement of an oxidative pathway. To explore this possibility further, we used other anti-oxidants such as thiourea and mannitol in our studies. As shown, photo-oxidative stress resulted in significant lowering of GSH levels as compared with control cells maintained in dark. The presence of thiourea (7 mM) in the medium of 661W cells was protective against photo-oxidative damage, as seen by the maintenance of GSH levels close to control values (Table I).
NF-B Activity Is Down-modulated upon Exposure of Photoreceptor Cells to Light-The photoreceptor cells were exposed to light for various time intervals for 15, 30, and 60 min. There was no change in NF-B activity up to 30 min of light exposure in both the nucleus (Fig. 1a, lanes 2 and 3 for nucleus) and cytoplasm ( Fig. 1a, lanes 6 and 7) compared with dark-exposed control cells (Fig. 1a, lanes 1 and 5, for nucleus and cytoplasm, respectively). Upon 60 min of light exposure, there was approximately 80% loss of NF-B binding activity in both the nucleus and cytoplasm (lanes 4 and 8, respectively). These results indicate that the cultured photoreceptor cells express NF-B constitutively and that the activity of NF-B decreases on exposure to photo-oxidative stress.
The specificity of the binding of NF-B was shown by competition with cold NF-B consensus oligonucleotide (Fig. 1b, lanes 2-4). As expected, a lack of competition was observed with cold NF-B mutant oligonucleotide (Fig. 1b, lanes 6 -8).
The DNA protein complex seen in the EMSA appears to be a heterodimer of p50 and p65 subunits of NF-B, as revealed by a decrease in the binding upon additional incubation with the antibodies to the p65 and p50 ( Fig. 1c) subunits, in a supershift assay. An unrelated antibody, anti-cyclin D1 used as a negative control, did not inhibit the DNA protein complex formation (Fig. 1c). These results confirm the identity of the NF-B DNAprotein complex seen in the EMSAs.

Pretreatment of Photoreceptor Cells with Antioxidants Protects against Down-modulation of NF-B upon Exposure to
Light-In order to establish the involvement of oxidative damage in the lowering of NF-B activity during conditions of photo-oxidative stress, we studied the effects of antioxidants, namely NAC, mannitol, and thiourea under these conditions (Fig. 2a). Lanes 1, 5, and 9 represent NF-B binding activity in dark-exposed cells. Lanes 2, 6, and 10 represent NF-B binding activity in dark-exposed cells in the presence of NAC, mannitol, and thiourea, respectively. The presence of these anti-oxidants did not appreciably alter the NF-B binding activity in darkexposed control cells. As expected, the activity of NF-B decreased on exposure to photo-oxidative stress (lanes 3, 7, and 11). The inclusion of NAC, mannitol, and thiourea in the growth medium prior to light exposure partially protected against the down-modulation of NF-B (Fig. 2a, lanes 4, 8, and  12, respectively) in light-exposed 661W cells. The differences in the extent of protection of NF-B levels in these groups may be attributed to differences in efficacy of these anti-oxidants in affording protection against photo-oxidative damage. These results indicate that oxidative damage plays a major role in decreasing the NF-B activity in cultured photoreceptor cells Effect of Photo-oxidative Stress on NF-B Levels in Madin-Darby Canine Kidney (MDCK) Cells-To assess the specificity of response of 661W cells to photo-oxidative stress, we studied the effect of light exposure on MDCK cells, using them as an unrelated control. The light exposure of MDCK cells did not cause a decrease in NF-B binding activity, in both the nucleus and cytoplasm (data not shown). These results indicate that the cell-specific response of 661W cells to light is different from that of MDCK cells.
Immunoblot Analysis of IB␣, p50, and RelA Subunit of NF-B in 661W Cells Exposed to Light-To further confirm the down-modulation of NF-B, the protein levels of IB␣, p50, and RelA subunits of NF-B were studied in 661W cells exposed to light, with or without pretreatment with NAC and thiourea, by immunoblot analysis using specific antibodies. The light-exposed cells showed lowering of IB␣, p50, and RelA subunit of NF-B, as compared with dark controls. Pretreatment of the cells with both anti-oxidants protected the levels of NF-B p50, RelA, and IB␣ subunits, albeit partially, upon exposure to light (Fig. 3). To ensure that light exposure does not result in a generalized protein degradation, a control protein GAPDH was included, which was not greatly altered in all samples under these experimental conditions. On quantitation, there was a Х50% decrease of IB␣ protein levels on 2-h light exposure. On the other hand, there was Х90% decrease in p50 and RelA subunit with no change in GAPDH levels under similar conditions. Inclusion of anti-oxidants protected to a large extent against degradation of these proteins. IB␣ was protected 100%, whereas p50 and p65 were protected to 40 -45% of control values. Based on the results of GAPDH protein levels, these data suggest that down-modulation of IB␣ and p50 and p65 subunit of NF-B is not due to random protein 0.0258 Ϯ 0.005 (n ϭ 3) Light-exposed 0.598 Ϯ 0.040 (n ϭ 4) a 0.010 Ϯ 0.003 (n ϭ 3) a N-Acetylcysteine (2 mM) pretreated and light-exposed 0.197 Ϯ 0.067 (n ϭ 4) b ND c Thiourea (7 mM) pretreated and light-exposed ND 0.037 Ϯ ND a Significant at p Ͻ 0.01. b Significant at p Ͻ 0.05. c ND ϭ not determined.

degradation.
Cultured Photoreceptor Cells Undergo Cell Death via Apoptosis upon Light Exposure-To establish that oxidative damage along with a down-modulation of NF-B results in apoptosis of these cells, we studied the effect of photo-oxidative stress on apoptosis of these cells in presence or absence of the antioxidant, NAC. We found that exposure of 661W cells to light up to 1 h did not result in any significant apoptosis of cultured photoreceptor cells (Fig. 4A). However, after 2 and 4 h of light exposure, many cells underwent apoptosis (Fig. 4, C and E, respectively), compared with dark-exposed control cells as seen by incorporation of fluoresceinated dUTP in the nuclei of apoptotic cells containing fragmented DNA. Approximately, 80% of the cells were seen to undergo apoptosis in 661W cells exposed to light for 4 h (Fig. 4E). Inclusion of NAC protected the cells from apoptosis at the 1-, 2-, and 4-h intervals of light For quantitation of the bands of the autoradiogram, a value of 100% was taken for dark-exposed controls and the values for other samples were calculated as percent of control and are shown below each lane. "D" and "L" represent dark-exposed controls and light-exposed cells respectively. b, the identity of the shifted band seen in the EMSA was confirmed to be that of NF-B by competition EMSA using molar excesses of cold consensus and mutant NF-B oligonucleotides. Lanes 1 and 5 show NF-B levels in dark-exposed 661W cells in the absence of any competitor oligonucleotide. Competition EMSA using 50-, 100-, and 200-fold molar excess of cold consensus NF-B oligonucleotide resulted in reduction of the NF-B binding to the consensus sequence (Fig. 1b,  lanes 2-4). On the other hand competition with cold mutant NF-B oligonucleotide using 50-, 100-, and 200-fold molar excess did not result in a reduction of the NF-B binding to consensus sequence (Fig. 1b,  lanes 6 -8). c, furthermore, the specificity of NF-B binding was established by super shift assays using p50 and p65 antibodies. Anticyclin D1 was used as an unrelated antibody serving as a negative control. The results show that both p65 and p50 antibodies resulted in decrease in NF-B band intensity, whereas anticyclin D1 did not have any effect on the binding reaction (c).

FIG. 2. Pretreatment of 661W cells with antioxidants protects
against the down-modulation of nuclear NF-B levels upon light exposure. a, pretreatment of 661W cells with NAC, mannitol, and thiourea protects against the down-modulation of NF-B levels (lanes 4, 8, and 12, respectively) compared with light-exposed 661W cells (lanes 3, 7, and 11, respectively). NF-B binding activity in dark-exposed controls are shown in lanes 1, 5, and 9 and dark-exposed control cells treated with NAC (2 mM), mannitol (50 mM), and thiourea (7 mM) are shown in lanes 2, 6, and 10, respectively. "D" and "L" represent darkexposed controls and light-exposed cells respectively. "NAC," "Man," and "Thio" indicate N-acetylcysteine, mannitol, and thiourea, respectively. b, 661W cells were also treated with 300 M H 2 O 2 for 30, 60, and 120 min, and NF-B binding activity was studied by EMSA. There was little change in NF-B binding activity upon treatment with H 2 O 2 for 30 and 60 min in both the nucleus (Fig. 2b, lanes 2 and 3) and cytoplasm (Fig. 2b, lanes 6 and 7), compared with the activity in the untreated cells (Fig. 2b, lanes 1 and 5 corresponding to nucleus and cytoplasm, respectively). A modest increase in the NF-B binding activity was seen after 120 min of treatment with H 2 O 2 in both nucleus (lane 4) and cytoplasm (lane 8).
FIG. 3. Immunoblot analysis of NF-B p50, RelA, IB ␣ subunit in 661W cells exposed to photo-oxidative stress. The cells were exposed to light for 2 h in presence or absence of NAC (2 mM) and thiourea (7 mM), and the levels of NF-B subunits p50, RelA, and IB␣ subunit were studied by immunoblot analysis. There was a decrease in the levels of IB␣, p50, and RelA subunit of NF-B upon light exposure (lane 2), and these were protected to varying extents by pretreatment of the cells with NAC and thiourea (lanes 3 and 4), respectively. The GAPDH levels were not altered between the experimental groups studied. "D" and "L" represent dark-and light-exposed 661W cells, respectively. "Nac" and "Thio" indicate N-acetylcysteine and thiourea, respectively. exposure, respectively (Fig. 4, B, D, and F). As expected, darkexposed control cells did not show any TUNEL-positive apoptotic cells (data not shown).
Photo-oxidative Stress and Immunocytochemical Localization of NF-B RelA Subunit in 661W Cells-To study the effect of photo-oxidative stress on the localization of RelA subunit of NF-B, immunocytochemistry was performed in these cells in presence and absence of NAC, using RelA-specific antibody. It was seen that RelA was present in the nucleus and also in the cytoplasm of dark-exposed control cells (Fig. 5A). Upon exposure to light, the nuclear and cytoplasmic labeling of NF-B diminished to a large extent (Fig. 5B). But, in the presence of NAC, a number of cells exposed to light retained positive immunoreactivity of RelA, in the nucleus as well as cytoplasm (Fig. 5C). There were still some cells, which had diminished RelA immunoreactivity in their nuclei and cytoplasm (Fig. 5C). These data indicate that light exposure caused the degradation of RelA subunit from the nucleus as well as cytoplasm by an oxidative pathway.
Pretreatment with the Proteasome Inhibitor ALLN Does Not Protect 661W Cells against Photo-oxidative Stress-induced Apoptosis-Since our results suggested that preservation of NF-B levels in the cells protects against apoptosis, we investigated the down-regulation of constitutive NF-B activity and its effect on apoptosis, by a proteasome inhibitor, ALLN. ALLN pretreatment of the 661W cells did not protect the NF-B binding activity, upon exposure to light (Fig. 6a, lanes 4 and 8 for nucleus and cytoplasm, respectively) compared with NF-B binding activity in both the nucleus and the cytoplasm of 661W cells exposed to light without any pretreatment (Fig. 6a, lanes   3 and 7 for nucleus and cytoplasm, respectively). ALLN treatment to cells maintained in the dark also did not alter their NF-B activity (Fig. 6a, lanes 2 and 6 for nucleus and cytoplasm, respectively).
The TUNEL assay revealed that ALLN pretreatment also caused some of the cells maintained in the dark, to undergo apoptosis (Fig. 6b, panel B). Furthermore, ALLN treatment did not protect against photo-oxidative stress induced apoptosis (Fig. 6b, panel D) compared with cells exposed to light without any pretreatment of ALLN (Fig. 6b, panel C). ALLN, by virtue FIG. 4. Fluorescent TUNEL labeling of 661W cells after exposure to light for various time durations in presence or absence of NAC. 661W cells with or without pretreatment with NAC were exposed to light for various durations, fixed with 4% paraformaldehyde, and processed for TUNEL labeling. There was a time-dependent increase in the number of the cells labeled with the fluoresceinated dUTP suggestive of apoptosis (white arrows) in 661W cells, with increasing duration of light exposure for 1 h (A), 2 h (C), and 4 h (E). Not all the cells were undergoing apoptosis in 2-h light-exposed group (C, arrowheads). Inclusion of NAC in the culture medium before light exposure, protected the cells from apoptosis (B, D, and F) for 1, 2, and 4 h of light exposure, respectively. There were also a few apoptotic cells in the NAC pretreated cells exposed to light for 4 h (F, arrow). Non-apoptotic cells were devoid of fluorescence.

FIG. 5. Photo-oxidative stress and immunocytochemical localization of NF-B RelA subunit in 661W cells.
The RelA subunit of NF-B was predominantly present in the nucleus and in the cytoplasm of dark-exposed control cells (A, arrows). Upon exposure to light, the nuclear and cytoplasmic labeling diminished considerably (B, arrows).
In the presence of NAC, a number of cells showed positive immunoreactivity in the nucleus as well as cytoplasm (C, arrows). Some nuclei devoid of RelA were also seen in NAC pretreated cells exposed to light (C, arrowheads). These data indicate that light exposure brought about lowering of NF-B levels in the nucleus as well as cytoplasm by an oxidative pathway. of being a proteasome inhibitor, could block IB␣ degradation, thereby, NF-B activation could be inhibited thus leading to apoptosis. This provides further confirmation of our hypothesis that NF-B is involved in blocking apoptosis.

Transient Transfection of 661W Cells with a Dominant Negative IB␣ Accelerates Photo-oxidative Stress-induced Apoptosis by Down-modulating NF-B Binding
Activity-To further delineate the role of NF-B in apoptosis we also transfected the cells with IB␣ super-repressor and examined its effects on NF-B binding activity and apoptosis upon exposure to photooxidative stress. To determine this, 661W cells were transiently transfected with IB␣ super-repressor and evaluated for their IB␣ and IB␣⌬N levels, NF-B binding activity, and DNA fragmentation by TUNEL. Immunoblot analysis of the transfected cells revealed higher protein levels of IB␣ in these cells, compared with the mock-transfected controls (Fig. 7a,  compare lane 2 with lane 1). A faster moving band corresponding to the super-repressor was also detected in the transfected cells representing a truncated form of IB␣, the IB␣⌬N, but not in the mock-transfected cells (Fig. 7a, compare lane 2 with  lane 1). A control protein, ␤-tubulin, did not appear to change between the two samples (Fig. 7a, bottom panel).
The NF-B binding activity in the untransfected cells was higher than that of the super-repressor transfected cells (Fig.  7b, lanes 1 and 2). Upon exposure to light, there was a rapid decline in NF-B activity in the transfected cells compared with that in the untransfected cells, both in the nucleus and cytoplasm (Fig. 7b, lanes 4, 6, and 8 compared with lanes 3, 5,  and 7 for nucleus; lanes 12, 14, and 16 compared with lanes 11,  13, and 15 for cytoplasm, for 15, 30, and 60 min of light exposure, respectively). Quantitation of a typical EMSA autoradiogram showed practically no change in NF-B binding activity of control cells exposed to light up to 30 min, with a 75-85% decrease in binding activity upon 60 min of light exposure in nucleus as well as cytoplasm. In contrast, in IB␣ super-repressor transfected cells, there was a rapid loss of NF-B binding activity, which decreased by 40 and 95% at 30 and 60 min in the nucleus and by 50,80, and 90% at 15, 30, and 60 min of light exposure in the cytoplasm. There was no decrease in the NF-B binding activity of nucleus at 15 min of light exposure in the transfected cells.
The TUNEL assay revealed that mere transfection of the IB␣⌬N super-repressor caused a number of cells to undergo apoptosis, even without being exposed to light (Fig. 7c, panel  E). There was a rapid increase in the number of TUNEL positive cells in the transfected cells by as early as 15 min of exposure to light (Fig. 7c, panel F), when compared with untransfected 661W cells (Fig. 7c, panel B). It reached a maximum by 30 min of light exposure in the IB␣ super-repressor transfected group (Fig. 7c, panel G), in contrast with just a few apoptotic cells in the untransfected 661W cells exposed to light for 60 min (Fig. 7c, panel D).
Thus, transfection with the super-repressor hastens the kinetics of down-modulation of NF-B binding activity as well as apoptotic cell death of 661W cells upon exposure to light.

Pretreatment with the Caspase-1 Inhibitor (YVAD-CMK) Protects against Down-modulation of NF-B and Apoptosis of 661W
Cells upon Exposure to Light-In order to identify proteases involved in the down-regulation of NF-B in the 661W cells, we studied the effect of caspase inhibitors on NF-B binding activity and apoptosis of 661W cells exposed to photooxidative stress. Pretreatment of 661W cells with the caspase-1 inhibitor, YVAD-CMK (100 M) could protect against downregulation of NF-B upon exposure of the cells to light in both nucleus (Fig. 8a, compare lane 4 to lane 3) and cytoplasm (Fig.  8a, compare lane 8 to lane 7). On the other hand, inclusion of caspase-3 inhibitor, DEVD-CHO did not protect the levels of NF-B in the light-exposed cells (data not shown).
To further study if the caspase-1 inhibitor could protect against light-induced apoptosis, we performed TUNEL assay of these cells exposed to light in presence of the inhibitor. As seen in Fig. 8b, pretreatment with the caspase-1 inhibitor could protect against photo-oxidative stress induced apoptosis (panel C) as compared with 661W cells exposed to light in absence of caspase-1 inhibitor (panel B), and dark-exposed controls (panel A). Similar studies with caspase-3 inhibitor showed that it did not render protection of 661W cells against photo-oxidative stress-induced apoptosis (data not shown). These studies indicate that down-modulation of NF-B may be due to an activation of a specific caspase, namely, caspase-1.

DISCUSSION
Apoptosis of photoreceptor cells is a common phenotype in retinal dystrophies, shared by patients with age-related macular degeneration and autosomal dominant retinitis pigmentosa, a family of disorders characterized by photoreceptor cell degeneration and a corresponding loss of vision. Point mutations of several genes involved in visual transduction have been identified in retinitis pigmentosa and other forms of retinal dystrophies (24). These studies, while pointing to the primary cause of the disease, do not provide clues to the downstream effectors that play a crucial role in the progression of the disease, ultimately leading to photoreceptor cell death by apoptosis.
FIG. 6. NF-B binding activity and apoptosis of ALLN-pretreated, dark-and light-exposed 661W cells. a, 661W cells were pretreated with ALLN (100 M) and exposed to light for 1 h. Pretreatment with ALLN did not protect against the down-regulation of NF-B binding activity upon light exposure (lanes 4 and 8 for nucleus and cytoplasm, respectively). Lanes 1 and 5 represent NF-B binding activity in nucleus and cytoplasm of dark-exposed control cells. Lanes 2 and 6 show the NF-B binding activity in ALLN-pretreated, dark-exposed control cells. The NF-B binding activity was decreased upon exposure to light (lanes 3 and 7 for nucleus and cytoplasm, respectively). b, TUNEL assay of ALLN-pretreated, dark-and light-exposed 661W cells. Pretreatment with ALLN caused a few cells to undergo apoptosis even in cells maintained in dark (B, arrows), compared with the untreated controls (A). Upon light exposure, a number of cells were TUNELpositive (C, arrows). Pretreatment with ALLN did not protect against photo-oxidative stress-induced apoptosis (D, arrows).
Light has been extensively used as a initiator of photoreceptor cell death in a number of in vivo and in vitro experimental conditions (31)(32)(33). In vivo studies have also shown that exposure of rats to constant light results in apoptosis of photoreceptor cells (25)(26)(27)(28)(29)(30). Production of lipid hydroperoxides has been observed in light-exposed retinas (44). Retina has been shown to be susceptible to lipid peroxidation (45,46) despite having high levels of antioxidants (47)(48)(49).
In the current study, we have assessed the involvement of oxidative damage in the apoptotic cell death of 661W photoreceptor cells in an in vitro model. The cell line expresses several markers of differentiated photoreceptors, including opsin, arrestin, rds/peripherin, phosducin, and interphotoreceptor retinol binding protein. 2 The 661W cell line, therefore, is a valua-  11 and 13). A significant decrease in the NF-B binding activity in the untransfected cells was seen only upon 60 min of light exposure in both nucleus (lane 7) and cytoplasm (lane 15). For quantitation of the bands of the autoradiogram, a value of 100% was taken for dark treated controls for both the transfected and untransfected cells and the values for other samples were calculated as percent of control for each group and are shown below each lane. "D" and "L" represents dark and light, respectively. c, TUNEL assay of super-repressor IB␣⌬N transfected 661W cells exposed to light for various time durations. There were no TUNEL-positive cells in the untransfected cells maintained in the dark (A) and in those exposed to light for 15 (B) and 30 (C) min. There was induction of apoptosis only after 60 min of light exposure (D, arrows), in the untransfected cells. In contrast, transfection with the IB␣ super-repressor caused a few cells to undergo apoptosis even in the cells maintained in the dark (E, arrows). There was a rapid increase in the number of TUNEL positive cells, in the transfected cells exposed to light for 15 (F), 30 (G), and 60 min (H) of light exposure (arrows). ble cell line to study photoreceptor cell death. When the cells exposed to light for 4 h were subjected to an apoptotic TUNEL assay, it was seen that 80% of the cells were labeled with fluoresceinated dUTP, suggestive of apoptosis. In a time course study, the extent of apoptosis of 661W cells had a direct correlation to duration of light exposure, and inclusion of NAC was protective of this effect. This experimental system thus, afforded a convenient model to analyze, the molecular events leading to apoptosis of photoreceptors by an oxidative pathway.
The important role of oxidative events has been further emphasized for a variety of biological processes, such as signal transduction and gene expression (19). A few studies have demonstrated that in particular, activation of NF-B requires oxidative signaling, i.e. its expression is dependent on the redox state of the cell (1,10). Treatment of several cell lines with H 2 O 2 resulted in activation of NF-B (50,51), and this can be blocked by inclusion of antioxidants (52)(53)(54).
By electrophoretic mobility shift assay, we found constitutive NF-B activity, both in the nucleus and the cytoplasm of 661W photoreceptor cells. Visible light exposure to photoreceptor cells creates conditions of photoxidative stress leading to oxidative damage. This causes the cells to proceed to cell death via apoptosis. The NF-B activity in 661W cells was found be progressively down-regulated upon exposure of the cells to light. Pretreatment of the cells with antioxidants namely NAC, thiourea, and mannitol partially prevented the down-modulation of NF-B and NAC also protected the cells from apoptosis, indicating involvement of hydroxyl radicals and superoxide anions in the pathway leading to cell death via apoptosis. The NF-B activity profile in the photoreceptor cells exposed to light appears to be radically different from that seen in the unrelated MDCK cells. While the light exposure of 661W cells leads to decrease in NF-B binding activity and apoptosis, the same stimulus does not greatly alter the nuclear NF-B activity and does not lead to cell death in the MDCK cells. This points to a cell type specificity of 661W cells' response to light.
In the cell, NF-B is stored in the cytoplasm in its inactive state by interaction with IB␣. On activation, IB␣ undergoes degradation through an ubiquitin-dependent pathway (55,56), allowing translocation of NF-B to nucleus (57,58) and subsequently binding to DNA regulatory elements within NF-B target genes. Under our experimental conditions we find degradation not only of IB␣, but also the NF-B p50 and RelA subunits. How the p50 and p65 subunits of NF-B are degraded in this oxidative stress paradigm is not fully understood. It is plausible that exposure of photoreceptor cells to light causes activation of one or more proteases which leads to degradation of not only IB␣, but also the p50 and p65 subunits. One of the protease responsible for degradation of NF-B proteins in the 661W cells appears to be the interleukin 1␤-converting enzyme also called caspase-1 according to the new nomenclature. Recently, it was shown that caspase-3, which activates apoptosis, can also induce proteolytic cleavage of IB␣ (59). In addition, the ubiquitin-conjugating enzymes that control IB␣ degradation are known to be induced by ROI (60). Thus, in our system, it appears that exposure of photoreceptor cells to light generates ROI, which activates caspase-1, resulting in proteolytic cleavage of NF-B proteins leading to apoptosis. This is consistent with studies which show that NF-B activation is essential for cell survival (61).
The role of NF-B in apoptosis is not very clear with reports of both pro-and anti-apoptotic aspects appearing in literature. It was demonstrated recently that NF-B is needed for TNF-␣ mediated induction of IAP-2, a protein belonging to the inhibitors of apoptosis (IAP) protein family. When overexpressed in mammalian cells, cIAP-2 activates NF-B and suppresses TNF cytotoxicity (62). Exposure of human alveolar epithelial (A 549) cells to hyperoxia resulted in activation of NF-B leading to necrotic cell death (63). These authors speculated that apoptosis occurs in the absence of NF-B activation but protection from cell death by NF-B is limited to apoptosis. It was further shown that TNF-␣ induces cell death in RelA Ϫ/Ϫ mouse fibroblast cells, whereas RelA ϩ/ϩ cells were unaffected, demonstrating the role of RelA in protection of the cells from TNF-␣ induced apoptosis (15). A similar antiapoptotic role of the RelA subunit was also demonstrated by Wang et al. (16) and Van Antwerp et al. (17). These studies indicate that either the inhibition of NF-B RelA subunit or the prevention of its translocation to nucleus is essential to induce apoptosis (15). However, several groups have suggested that NF-B may function pro-apoptotically under some conditions and in certain cell lines (19,64,65).
In the 661W cells the NF-B protein does not seem to follow the prototype pathway of its activation, upon exposure to an oxidative stimuli. For instance we do not observe an increase in NF-B binding activity in these cells treated with an oxidizing agent like H 2 O 2 . Similarly, exposure of 661W cells to photooxidative stress, which causes a degradation of IB␣, does not activate NF-B, but in fact degrades the p50 and p65 subunits. It therefore appears that in these cells, oxidative damage activates other cellular mechanisms that trigger selective protein degradation, culminating in apoptosis. It needs to be emphasized that NF-B signaling is indeed essential for survival of these cells, as seen by the induction of apoptosis in the 661W cells upon transfection with the IB super-repressor. But the overwhelming stimulus for cell death appears to result from the oxidative degradation of NF-B, as seen by the greatly accelerated kinetics of apoptosis in the super-repressor-trans- . Inclusion of the caspase-1 inhibitor in dark-exposed control cells did not alter NF-B levels in both nuclear (lane 2 compared with lane 1) and cytoplasmic fractions (lane 6 compared with lane 5), respectively. "D" and "L" represent dark-and light-exposed 661W cells. b, fluorescent TUNEL labeling of caspase-1 inhibitor pretreated 661W cells exposed to light. The cells maintained in the dark were devoid of apoptotic cells (A, arrowheads). A number of TUNEL-positive cells were seen in the lightexposed group (B, arrows). Pretreatment with the caspase-1 inhibitor was protective against apoptosis as seen by decreased TUNEL labeling in these cells (C, arrowheads). A few cells undergoing apoptosis were also seen in this group (C, arrows). fected 661W cells. A similar response was observed when the cells were pretreated with the proteasome inhibitor, ALLN, which also inhibits calpain I and II, cathepsin B, cathepsin L, and neutral cysteine proteases. Treatment with ALLN would slow down the degradation of IB␣, thereby perturbing the NF-B signaling in the cell and exacerbate cell death via apoptosis. The results of ALLN treatment further suggest that the down-modulation of RelA subunit is not due to generalized protein degradation, but may indeed require an activation of a specific caspase involved in apoptosis (66). Evidence is accumulating to attest to the fact that a constitutive expression/ activation of NF-B is essential for a cell to survive an apoptotic insult (61,67,68). All these observations underscore the importance of the protective role of NF-B against photo-oxidative stress induced apoptosis in the 661W cells. A schematic representation of the molecular events occurring in the course of photo-oxidative stress-induced apoptosis of photoreceptor cells are shown in Fig. 9.
Taken together, these results suggest that the RelA subunit of NF-B constitutively expressed in 661W photoreceptor cells may be important for photoreceptor cell survival. Exposure of the cells to visible light creates conditions of photo-oxidative stress, causing the production of reactive oxygen intermediates leading to oxidative damage as evidenced by depletion of glutathione and increase in malonyldialdehyde formation. These oxidative events further result in the down-modulation of NF-B (predominantly the RelA subunit), thereby exacerbating the oxidative damage and channeling the cells along a pathway of cell death, via apoptosis.