Light-induced photoreceptor apoptosis in vivo requires neuronal nitric-oxide synthase and guanylate cyclase activity and is caspase-3-independent.

Apoptosis is the mode of photoreceptor cell death in inherited and induced retinal degeneration. However, the molecular mechanisms of photoreceptor cell death in human cases and animal models of retinal dystrophies remain undefined. Exposure of Balb/c mice to excessive levels of white light results in photoreceptor apoptosis. This study delineates the molecular events occurring during and subsequent to the induction of retinal degeneration by exposure to white light in Balb/c mice. We demonstrate an early increase in intracellular calcium levels during photoreceptor apoptosis, an event that is accompanied by significant superoxide generation and mitochondrial membrane depolarization. Furthermore, we show that inhibition of neuronal nitric-oxide synthase (nNOS) by 7-nitroindazole is sufficient to prevent retinal degeneration implicating a key role for neuronal nitric oxide (NO) in this model. We demonstrate that inhibition of guanylate cyclase, a downstream effector of NO, also prevents photoreceptor apoptosis demonstrating that guanylate cyclase too plays an essential role in this model. Finally, our results demonstrate that caspase-3, frequently considered to be one of the key executioners of apoptosis, is not activated during retinal degeneration. In summary, the data presented here demonstrate that light-induced photoreceptor apoptosis in vivo is mediated by the activation of nNOS and guanylate cyclase and is caspase-3-independent.

and cone photoreceptors. Nyctalopia (night-blindness), an initial symptom of the disease is associated with rod photoreceptor degeneration. As the disease progresses, severe loss of peripheral and central fields occur, due to the subsequent loss of cone photoreceptors. The genetics of RP is complex and to date at least 30 genes have been implicated in the etiology of RP, the majority of which encode photoreceptor-specific proteins that are either components of the phototransduction cascade or structural components of these cells (see the Retinal Information Network available on the University of Texas School of Public Health Web server). Despite the diverse genetics underlying the pathology of RP, one feature common to both human cases and animal models is the apoptotic cell death of rod and cone photoreceptors (1,2). Apoptosis, therefore, represents a common final pathway in the pathology of RP.
Exposure to excessive levels of white light induces photoreceptor apoptosis and has previously been used as a model for the study of retinal degeneration (3). Furthermore, several animal studies have demonstrated that photoreceptors from retinal degeneration mutants are more susceptible to the damaging effects of excessive light than normal photoreceptors (4 -7). The evidence from these animal models suggests that excessive light may enhance the progression and severity of some forms of human RP. The mechanism by which light induces retinal degeneration is at present unclear. A recent study demonstrated that rhodopsin is essential for light-induced retinal degeneration, indicating that signal flow through the phototransduction cascade is necessary to mediate the damaging effects of light (8). Indeed, constitutive signal flow in phototransduction is thought to underlie some forms of inherited retinal disorders (9). This is supported by the significant number of mutated genes that encode components of the phototransduction cascade and are implicated in the etiology of RP or of related conditions. These include rhodopsin, cyclic GMP phosphodiesterase, the rod cGMP-gated channel, and rhodopsin kinase (see the Retinal Information Network available on the Web).
Although the exact mechanism of photoreceptor apoptosis resulting from exposure to excessive levels of light remains undefined, a diverse range of agents have been used to prevent photoreceptor apoptosis in this model (10 -12). These include calcium channel blockers (13) as well as antioxidants (14,15) implicating a role for both intracellular calcium and reactive oxygen species in retinal apoptosis. However, the molecular basis for the rescue of photoreceptors from apoptosis in lightinduced retinal degeneration by these agents has yet to be elucidated.
In this study, the molecular events that occur during lightinduced photoreceptor apoptosis are delineated. We demonstrate that inhibition of NOS not only prevents photoreceptor cell death, but also additional features of apoptosis, including superoxide generation, mitochondrial membrane depolarization, and elevated intracellular calcium levels, implicating a key role for NO in this model. In addition, the potential involvement of cyclic guanosine monophosphate (cGMP), a central component of the phototransduction cascade, is explored by examining the effects of ODQ (1H- (1,2,4)oxadiazolo [4,3-a]quinoxalin-1-one), a potent and selective inhibitor of guanylate cyclase. Inhibition of guanylate cyclase in this study prevents light-induced retinal degeneration. These results suggest that NO mediates cell death in this model through activation of guanylate cyclase, resulting in increased levels of intracellular calcium through cGMP-gated calcium channels. Finally, our results demonstrate that caspase-3 is not activated during photoreceptor apoptosis in this model.

EXPERIMENTAL PROCEDURES
Retinal Light Damage-Adult male Balb-c mice were maintained in the dark for 18 h before being exposed to constant light. Immediately prior to light exposure their pupils were dilated with 5% cyclopentolate. The mice were then exposed to 2 h of cool white fluorescent light at a luminescence level of 5000 lux. The mice were sacrificed after treatment by cervical dislocation at the following time points: 30 min and 1 h after light onset, immediately after light exposure (0 h) and after 6, 14, and 24 h of darkness that followed the 2-h light exposure.
Intraperitoneal Injections-Mice were injected intraperitoneally with the following: N G -nitro-L-arginine methyl ester (L-NAME), 100 mg kg Ϫ1 (Sigma Chemical Co., UK) in PBS or the inactive isomer N G -nitro-Darginine methyl ester (D-NAME), 100 mg kg Ϫ1 (Sigma, UK) in PBS as a control, 7-nitroindazole (7NI), 100 mg kg Ϫ1 (Sigma, UK) in peanut oil or peanut alone as a control, and ODQ, 50 mg kg Ϫ1 (Calbiochem) in Me 2 SO or Me 2 SO alone as a control. All intraperitoneal injections were administered 1 h prior to light exposure.
Terminal Deoxynucleotidyl Transferase-mediated dUTP Nick End Labeling of Fragmented DNA-DNA strand breaks in photoreceptor nuclei were detected by terminal deoxynucleotidyl transferase-mediated dUTP nick end labeling (TUNEL). Briefly, enucleated eyes were fixed in 10% buffered formalin for 24 h, dehydrated, processed, and embedded in paraffin. Sections (5 m) were incubated in 50 l of reaction buffer containing 2.5 mM CoCl 2 , 0.1 unit/ml terminal deoxynucleotidyl transferase (TdT) in a 0.1 M sodium cacodylate (pH 7.0) buffer and 0.75 nM fluorescein-12-dUTP (Roche Molecular Biochemicals, Germany). These sections were incubated at 37°C for 1 h in a humidified chamber. Following several washes in PBS, the sections were mounted in mowiol (Calbiochem) and viewed under a fluorescence microscope (Nikon Eclipse E600) using a fluorescein isothiocyanate filter. Three animals were used for each of the time points; 0, 6, 14, and 24 h after light exposure.
Agarose Gel Electrophoresis-Enucleated eyes were placed in PBS, and retinal dissection was carried out using a watchmaker's forceps. The choroid, sclera, and pigmented epithelium were removed, and the retina was then separated from the vitreous and lens. Retinal DNA was isolated following phenol-chloroform extraction. Briefly, retinas were placed in 150 l of lysis buffer (20 mM EDTA, 100 mM Tris, and 0.8% sodium lauryl sarcosinate) containing proteinase K (20 g/ml) and incubated at 50°C for 18 h, vortexing occasionally. The DNA was then extracted with phenol chloroform and chloroform isoamyl alcohol (v/v, 24:1). DNA was precipitated with ethanol, and the pellet was dissolved in Tris-EDTA. Total RNase-treated DNA was visualized by including ethidium bromide (0.5 g/ml) in the agarose and observed by illuminating on a 302-nm UV transilluminator.
Assessment of Active Caspase-3 Content in Light-induced and Uninduced Balb/c Retina-Retinal dissection was carried out as above. Tissue dissociation was achieved in a 0.25% trypsin solution (Life Technologies, Inc., Paisley, UK). The cells were washed in PBS and fixed in 1% paraformaldehyde at 4°C for 30 min. The cells were then washed again with PBS and then with permeabilization buffer (PB: 10 mM HEPES, 150 mM NaCl, 4% FCS, 0.1% sodium azide, and 0.1% Triton X-100). Cells were resuspended in PB containing 0.125 g of anti-active caspase-3 antibody (PharMingen International, San Diego, CA), or the same concentration of an isotype control (rabbit IgG, Sigma, UK) and incubated for 1 h at 4°C. Following two washes in PB, cells were resuspended in 20 g/ml fluorescein isothiocyanate-conjugated secondary antibody (goat anti-rabbit, Sigma, UK) and incubated for 1 h at 4°C. After a further two washes in PB, cells were resuspended in 0.5 ml of PBS for FACScan analysis on a Becton-Dickinson FACScan flow cytometer.
Determination of Ac-DEVD-pNA Cleavage-Retinas were dissected and washed with cold PBS. Total protein was obtained by homogenizing retinas in 50 l of chilled lysis buffer containing 10 mM HEPES, pH 7.4, 2 mM MgCl 2 , 5 mM EGTA, 50 mM NaCl, 1 mM phenylmethylsulfonyl fluoride, 2 g/ml aprotinin, and 2 g/ml leupeptin. The cells were incubated on ice for 20 min and then lysed by 3-4 cycles of freezing and thawing. Insoluble material was pelleted by centrifuging at 20,000 ϫ g for 15 min at 4°C. The protein content of each sample was determined by the Bio-Rad protein assay (Bio-Rad, Hemel Hempstead, UK) using bovine serum albumin as a standard, and 80 g of protein in 50 l of lysis buffer was dispensed into each well of a microtiter plate. An equal volume of 2ϫ reaction buffer (50 mM HEPES, pH 7.4, 0.2% CHAPS, 20% glycerol, 2 mM EDTA, and 10 mM dithiothreitol) was added to each sample with 50 M caspase-3 substrate-DEVD-pNA (Bahcem, Saffron Walden, UK; 1 mM stock in Me 2 SO). Reactions were incubated at 37°C for 1 h and then cleavage of the peptide substrate DEVD-pNA was monitored by liberation of the chromogenic pNA in a SpectraMax-340 plate reader (Molecular Devices, CA) by measuring absorption at 405 nm.
Analysis of Intracellular ROS Generation-Superoxide anion levels were measured using a modified version of the assay as previously described (16). Briefly, cells (5 ϫ 10 5 ) were loaded with 10 M hydroethidine (DHE) (Molecular Probes), prepared from a 10 mM stock in Me 2 SO for 15 min at 37°C. Superoxide anion oxidizes DHE intracellularly to produce ethidium bromide, which fluoresces upon interaction with DNA. Superoxide anion levels were assessed by monitoring the fluorescence due to ethidium bromide on a Becton-Dickinson FACScan flow cytometer with excitation and emission settings of 488 and 590 nm, respectively.
Intracellular Free Calcium Measurement-Intracellular calcium levels were determined using the intracellular calcium probe, Fluo-3 AM (acetoxymethyl ester) (Molecular Probes, Leiden, The Netherlands). Cells were incubated in darkness with Fluo-3 (250 nM), prepared from a 500 M stock for 15 min at 37°C, and fluorescence was measured in FL-1 (530 nm) on a Becton-Dickinson FACScan flow cytometer with excitation at 488 nm.

Detection of Photoreceptor Apoptosis in Light-induced Balb/c
Mice-Exposure to excessive levels of light results in photoreceptor apoptosis (3). In the present study, Balb/c mice were exposed to 2 h of cool white fluorescent light at a luminescence level of 5000 lux. DNA strand nicking in the outer nuclear layer was assessed by Terminal dUTP nick end labeling (TUNEL) over 24 h in control animals and in animals exposed to 2 h of NO in Caspase-3-independent Retinal Apoptosis in Vivo white fluorescent light. Retinas of control animals showed no TUNEL-positive labeling in photoreceptor nuclei immediately after light exposure (Fig. 1A). At 6 h, positive labeling increased, and at 14 and 24 h, abundant TUNEL-positive cells were apparent in the ONL ( Fig. 1, b-e). Apoptotic cell death was confirmed by DNA agarose gel electrophoresis (Fig. 1B). Biochemically, apoptosis is characterized by internucleosomal DNA cleavage-producing DNA fragments that are multiples of 180 -200 bp, which appear in agarose gel electrophoresis as a ladder pattern. The presence of this typical ladder pattern at 14 and 24 h (lane 3 and 4) confirms apoptotic cell death in this model and verifies the efficacy of the TUNEL assay to detect apoptosis following light exposure.
Inhibition of NOS by L-NAME Prevents Light-induced Photoreceptor Apoptosis-To elucidate the molecular mechanism of light-induced photoreceptor apoptosis we investigated the role of NOS, previously shown to be involved in another model of retinal degeneration where rats were exposed to light for 7 days at an illuminescence of 90 footcandles (17). To investigate the role of NOS in this model of light-induced photoreceptor apoptosis, animals were treated with the NOS inhibitor L-NAME (100 mg kg Ϫ1 ), 1 h prior to light exposure. In treated animals no TUNEL-positive photoreceptor cells were detected immediately following light exposure or following a further incubation of 6, 14, and 24 h in darkness (Fig. 2, e-h). Animals that were injected with D-NAME, the inactive isomer of L-NAME, showed similar TUNEL labeling to untreated light-induced animals ( Fig. 2, a-d). These results suggest a key role for NO in this model of light-induced retinal degeneration.
Elevation of Intracellular Calcium Is an Early Event in Retinal Apoptosis and Is Inhibited by L-NAME-A role for calcium has previously been implicated in retinal cell death in vivo. The ability of calcium channel blockers to prevent rod photoreceptor apoptosis in two different models of retinal degeneration has been demonstrated (13,18). However, measurement of calcium levels during photoreceptor apoptosis has not previously been conducted. In this study, intracellular calcium levels were analyzed using the fluorescent probe Fluo-3 AM (Fig. 3). In-creased levels of calcium were detected after 30 min of light exposure, and the number of cells with elevated calcium continued to increase up to 3 h. This was followed by a significant decrease in intracellular calcium concentration. These data demonstrate that calcium elevation is an early and rapid event in retinal degeneration. This transient elevation of intracellular calcium is blocked by the NOS inhibitor L-NAME, suggesting that elevated intracellular calcium requires NOS activity.
Superoxide Anion Formation and the Collapse of Mitochondrial Membrane Potential (⌬ m )-A key role for ROS has been demonstrated in an in vitro model of retinal degeneration, where antioxidants prevent photoreceptor cell death (19). In vivo, a number of compounds possessing antioxidant properties have prevented photoreceptor apoptosis (see "Discussion"). However, to date, little is known about the oxidative pathways involved in such retinal death, nor have ROS measurements been conducted in an in vivo model of retinal degeneration. In this study, superoxide anion formation was monitored using the probe DHE in control light-induced Balb/c mice and in those treated with L-NAME (Fig. 4a). Increased levels of superoxide were detected following 30 min of light exposure, and the number of cells with elevated superoxide levels increased up to 3 h after the light insult. This was followed by a decrease in intracellular levels. This data demonstrates, therefore, that superoxide production is an early event in light-induced retinal degeneration occurring alongside increased calcium levels. Furthermore, our results show that superoxide generation is inhibited by L-NAME, suggesting that NOS activity is also required for the generation of superoxide.
Mitochondria are the major site of cellular ROS production. The increase in superoxide production observed here prompted the investigation of mitochondrial dysfunction as alterations in mitochondrial membrane potential (⌬ m ) can result in increased ROS generation. The lipophilic probe JC-1 was used to analyze ⌬ m . In the presence of an intact ⌬ m , JC-1 forms J-aggregates, which are associated with a shift in fluorescence emission (590 nm). Thus, a reduction in fluorescence emission at 590 nm can be interpreted as a reduction in ⌬ m . As illustrated in Fig. 4b, a reduction in ⌬ m is evident in a significant number of cells after 30 min of light exposure and 3 h after light exposure 75% of cells have reduced ⌬ m . This reduction in ⌬ m is blocked by L-NAME.
Light-induced Retinal Degeneration Is Accompanied by Increased Expression of nNOS-Although the NOS inhibitor L-NAME is an established inhibitor of in vivo NO production (20), L-NAME shows little or no selectivity for individual NOS isoforms. As both constitutive and inducible isoforms of NOS have been reported in the retina (see "Discussion"), it is unclear which of the three isoforms; inducible nitric-oxide synthase (iNOS), endothelial nitric synthase (eNOS), or neuronal nitric oxide (nNOS) is responsible for the production of NO following light exposure. Indeed, elevated levels of NO may be due to increased NOS protein expression or activation by calcium entry into the cells. Immunoblot analysis revealed no detectable increase in either iNOS or eNOS expression (Fig. 5A, part a) and Fig. 5A (part b). However, a time-dependent increase in nNOS expression was evident (Fig. 5A, part c). Densitometric analysis revealed a 2.5-fold increase in nNOS expression at 3 h, a 3-fold increase at 6 h, a 3.5-fold increase at 14 h, and a 5-fold increase at 24 h. The retinas taken from mice treated with 100 mg kg Ϫ1 of L-NAME 1 h prior to light exposure showed no increase in nNOS expression (Fig. 5A, part d). These results suggest that initial NOS activity is required for up-regulation of nNOS.
7-Nitroindazole Inhibits Light-induced Retinal Degeneration-To clarify the role of nNOS in light-induced retinal degeneration, we investigated the effects of treatment with 7NI, a selective inhibitor of neuronal nitric-oxide synthase that has been employed in several in vivo studies to inhibit the activity of neuronal NOS. 7 NI has high selectivity for nNOS (IC 50 710 nm) over iNOS (IC 50 20 m). In addition, several studies have demonstrated that 7NI had no effect on eNOS in the brain in a number of animal species, including rodents (21)(22)(23). Animals were treated with 100 mg kg Ϫ1 of 7NI or peanut oil 1 h prior to light exposure. Retinas of control animals (injected with peanut oil) had scattered TUNEL labeling at 6 h and abundant TUNEL labeling at 14 and 24 h (Fig. 5B, a-d) similar to untreated light-induced animals (Fig. 1A). No TUNEL-positive photoreceptor cells were detected at 6, 14, and 24 h in the retina of animals treated with 7NI (Fig. 5B, e-h). 7 NI, therefore, prevents light-induced photoreceptor apoptosis thus establishing the role of neuronal NOS in promoting light-induced retinal degeneration.
Inhibition of Guanylate Cyclase by ODQ Blocks Photoreceptor Apoptosis-A major action of NO is to activate the soluble form of the enzyme guanylate cyclase (24,25). Binding of NO to guanylate cyclase increases the latter's activity leading to the formation of cGMP, which may result in excess calcium influx through cGMP-gated channels. To explore the potential involvement of cGMP in photoreceptor apoptosis, we examined the effect of ODQ, a potent and selective inhibitor of NOsensitive guanylate cyclase. Animals were treated with 50 mg kg Ϫ1 of ODQ or Me 2 SO 1 h prior to light exposure. In the Me 2 SO control animals, positive TUNEL labeling was apparent at 6, 14, and 24 h (Fig. 6, a-d), similar to untreated animals (Fig. 1A). In animals treated with ODQ, no TUNEL-positive photoreceptor cells were detected at 6, 14, and 24 h (Fig. 6, e-h). These results demonstrate that guanylate cyclase activity is required for light-induced retinal degeneration.
Light-induced Photoreceptor Degeneration Occurs Independently of Caspase-3 Activation-Caspase-3 is recognized as one of the key executioners of apoptosis. Recently, activation of caspase-3 was reported during photoreceptor apoptosis in transgenic rats with the rhodopsin mutation S334ter (26). In these mutants, pretreatment with the irreversible caspase-3 inhibitor, z-DEVD-fmk prevents photoreceptor apoptosis. However, emerging evidence now suggests that not all cell types require caspase-3 to undergo apoptosis, and it has been reported that oxidative stress can inhibit caspase activation during photoreceptor apoptosis in vitro (27). Because the free radical NO appears to play a key role in light-induced retinal degeneration, we determined the activation status of caspase-3 during photoreceptor apoptosis in this model, using three different analytical techniques (Fig. 7). Caspase-3 is synthesized as a 32-kDa inactive proenzyme and is cleaved at Asp 28 -Ser 29 and Asp 175 -Ser 176 to generate active subunits of 17 and 12 kDa. Analysis of the levels of the 32-and 17-kDa caspase-3 species using Western blot demonstrates the absence of active, 17-kDa caspase-3 in cell lysates taken from the retinas of light-induced Balb/c mice 3, 6, and 14 h after light exposure (Fig. 7A). The murine hematopoietic 32D cell line is included as a control to demonstrate the processing of pro-caspase-3 (32 kDa) to the active 17-kDa fragment as these cells undergo apoptosis following ultraviolet light irradiation.
The activity of caspase-3-like proteases was further assessed by measuring the cleavage of the colorimetric substrate Ac-DEVD-pNA (Fig. 7B). Again no evidence of AcDEVD-pNA cleavage was obtained in light-induced Balb/c mice 3, 6, or 14 h after light exposure. Untreated and anti-Fas IgM-treated Jurkat cells served as negative and positive controls, respectively.
To further confirm the absence of active caspase-3 in the retina of light-induced Balb/c mice, assessment of active caspase-3 content by flow cytometry was carried out using an anti-active caspase-3 antibody (Fig. 7C). This antibody is raised against the active fragment and preferentially recognizes active caspase-3. Untreated and anti-Fas IgM-treated Jurkat cells served as negative and positive controls, respectively. There was no detectable active caspase-3 at 3, 6, or 14 h after light exposure.
Western blot analysis of two caspase-3 substrates, ICAD (inhibitor of caspase-activated Dnase) and PARP (poly (ADPribose)polymerase) was carried out to verify the lack of caspase-3 activity during light-induced retinal degeneration. ICAD is a caspase-3 substrate that exists as a complex with CAD and is cleaved at two sites by caspase-3 during apoptosis resulting in the release of CAD. PARP is cleaved and inactivated by caspase-3 into 25-and 85-kDa fragments. Analysis of the levels of the cleaved products of ICAD (12 kDa) and PARP (85 kDa) by Western blot demonstrates the absence of these fragments in cell lysates taken from the retinas of light-induced BALB/C mice 3, 6, and 14 h after light exposure (Fig. 7, D and E). The murine hematopoietic 32D cell line was employed to demonstrate the processing of both ICAD and PARP to 12-and 85-kDa fragments, respectively, as these cells undergo apoptosis following etoposide treatment. These results taken together with the lack of detection of caspase-3 activity suggest that light-induced retinal degeneration is caspase-3-independent. DISCUSSION In this study, we delineate the molecular events occurring during and subsequent to the induction of retinal degeneration by constant exposure to white light in mice. It has been known for some time that constant exposure to white light results in retinal degeneration (28) and accumulated data to date has shown this to be a useful model for the study of photoreceptor apoptosis in vivo. It has been suggested that some of the molecular defects observed in inherited retinal degenerations may produce signals equivalent to real light, leading to retinal degeneration by a similar mechanism (9).
Elevated intracellular calcium has been suggested to be a key event in apoptosis (29,30). In the central nervous system, it has been demonstrated to play an important role in degeneration during stroke or trauma (31). Elevated calcium is thought to disrupt the membrane potential and outer membrane of the mitochondria and induce the release of several potent apoptosis-inducing factors (32). In this study, intracellular calcium levels were monitored during and following light exposure, and it was determined that elevated calcium is an early and rapid event in photoreceptor apoptosis (Fig. 3). This is indirectly supported by the potent anti-apoptotic effects of calcium channel blockers in both the light-induced model (13) and in inherited retinal degeneration in the rd mouse (18). The results presented here provide a molecular basis for the inhibition of photoreceptor apoptosis by calcium channel blockers.
Intraperitoneal administration of the NOS inhibitor L-NAME inhibited the elevation of intracellular calcium observed, suggesting that NOS activity is required for increased intracellular calcium. To determine the mechanism by which NO regulates intracellular calcium levels in photoreceptors, we examined the potential role of guanylate cyclase, a downstream effector of NO that converts GTP into cGMP. In photoreceptors, cGMP binds cation specific channels on the photoreceptor membrane, maintaining them in an open conformation. Excessive calcium influx could potentially occur, therefore, through cGMP-gated channels following activation of guanylate cyclase by NO. In this study, intraperitoneal administration of ODQ a potent and selective inhibitor of nitric oxide-sensitive guanylate cyclase prevented photoreceptor apoptosis (Fig. 6), demonstrating that NO-sensitive guanylate cyclase plays a key role in light-induced photoreceptor apoptosis. NO possibly acts in this model by activating guanylate cyclase and eliciting a cGMP increase. Indeed, there is considerable evidence from other animal studies supporting a role for increased intracellular cGMP in rod photoreceptors in retinal degeneration. Increased levels of cGMP have been demonstrated to precede photoreceptor degeneration in the rd mouse (33), and rod-cone dysplasia in Irish Setter dogs (34). Accumulation of cGMP in the outer segment would result in an increase in the opening of the cGMP-gated channels and a subsequent increase in intracellular calcium. Such a model would explain the ability of calcium channel blockers, L-NAME and ODQ, to protect against light-induced photoreceptor apoptosis.
Several studies have implicated the importance of oxidative stress in animal models of retinal degeneration in which antioxidants retard or inhibit the degenerative pathology. In the light-induced model, a number of compounds possessing antioxidant properties have successfully prevented photoreceptor degeneration, these include: methylprednisolone (15), dithiomethyleurea (14), and ascorbic acid (10). However, to date, little is known about the oxidative pathways involved in such retinal death, nor have ROS measurements been conducted in vivo. This laboratory recently reported ROS as mediators of photoreceptor apoptosis in vitro and has demonstrated the early and rapid generation of ROS while also demonstrating inhibition of photoreceptor apoptosis with antioxidants (19). In this study, for the first time we undertake the analysis of the superoxide anion in an in vivo model of retinal degeneration. Our data demonstrate that superoxide production is an early event in light-induced retinal degeneration (Fig. 4), occurring alongside increased intracellular calcium levels. These results, taken together with the ability of antioxidants to prevent apoptosis in vivo, establish a critical role for ROS in photoreceptor apoptosis.
Emerging evidence now suggests that mitochondria play a central role in the execution of the apoptotic program. Alterations in ⌬ m have been reported as early events during apoptosis induced by a diverse range of stimuli (35). Consistent with these reports, an early reduction in ⌬ m was observed during photoreceptor apoptosis in this model (Fig. 5). Loss of ⌬ m can result in the release of several potent apoptosis-inducing factors whose presence have been reported in the intermembrane space, these include cytochrome c (36), apoptosis inducing factor (37), procaspase-3 (38), pro-caspase-2 (39), and pro-caspase-9 (39,40). The loss of ⌬ m observed here may occur as a result of elevations in intracellular calcium, because alterations in intracellular calcium can disrupt mitochondrial potential (32,41). Indeed, calcium-induced mitochondrial depolarization has previously been described in rod photoreceptor apoptosis in vitro (42). This study revealed that mitochondria are the initial target site in Ca 2ϩ -induced apoptotic rod cell death and demonstrated that Ca 2ϩ enters the rod photoreceptor outer segment via cGMP-gated channels and subsequently enters the mitochondria, ultimately resulting in mitochondrial depolarization.
Inhibition of NOS by L-NAME in this study was sufficient to prevent all features of apoptosis monitored, implicating a key role for NOS in light-induced photoreceptor apoptosis. However, because all three isoforms of this enzyme have been reported in the retina (43)(44)(45), it was unclear which of the three isoforms was responsible for mediating light-induced photoreceptor apoptosis. Western blot analysis demonstrates the expression of all three isoforms (Fig. 2) in the retina consistent with previous reports. A time-dependent increase in nNOS expression was demonstrated 3 h after the light insult, but there was no detectable increase in expression of either eNOS or iNOS. Pre-treatment of the mice with 7NI, a selective inhibitor of nNOS, was found to prevent photoreceptor apoptosis, demonstrating a prominent role for neuronal derived NO in this model. Activation of caspases, a family of cysteine proteases, has been demonstrated to be a central event in apoptosis in several studies (46). Caspase-3 is considered to be one of the major downstream caspases of the apoptotic machinery and has been shown to play a critical role in apoptosis induced by a range of stimuli in a wide variety of cell types. However, emerging evidence now suggests that not all cell types require caspase activity to undergo apoptosis (47)(48)(49)(50). Indeed, caspase-3-independent apoptosis is well documented in neurons (51)(52)(53)(54). Analysis of caspase-3 activity in this study of light-induced retinal degeneration using a range of different analytical techniques, demonstrated the absence of caspase-3 activation during apoptosis. Furthermore, Western blot analysis of ICAD and PARP, both caspase-3 substrates, revealed that these are not cleaved during photoreceptor apoptosis (Fig. 7). These results suggest that light-induced retinal degeneration is caspase-3independent. This also strongly implies that this cell death is caspase-independent, because caspase-3 activation has been shown to be dependent on prior activation of upstream caspases (caspase-2, 8,9,10) (55)(56)(57). This laboratory has previously provided evidence for a caspase-independent apoptotic pathway in photoreceptors in vitro (27). This study demonstrated oxidative inactivation of caspases. The presence of an essential cysteine residue in the active site of the caspases renders the activity of this family of proteases susceptible to redox modification (58). The key role of NO in photoreceptor apoptosis in this study may provide an explanation for the failure of caspase activation by a similar mechanism. Inhibition of caspase activity by S-nitrosylation, whereby nitric oxide is transferred to the cysteine sulfydryl group, has been reported in other apoptotic models, including neuronal apoptosis (59).
Despite the lack of evidence for caspase-3 involvement in this study, the photoreceptor apoptosis described retains one of the key characteristics of apoptosis attributed to caspase-3 activation; DNA fragmentation. CAD (a caspase-activated DNase that requires cleavage of its endogenous inhibitor (ICAD) in a caspase-3-dependent manner) is at present the only apoptosisspecific endonuclease identified (60). However, there is much evidence to support a role for Ca 2ϩ -dependent endonucleases during DNA fragmentation and death in apoptotic cells (61). Given the evidence for a role of calcium in this model, it is possible that this apoptotic pathway culminates in the activation of calcium-dependent endonucleases.
The results of this and previous in vitro studies (27) are in contrast to the recent reports of activation of caspases during photoreceptor apoptosis in rhodopsin S334ter rats (26) and RCS rats (62) and during photo-oxidative stress in vitro (63). The experiments described by Krishnamoorthy and colleagues (63) were carried out on cells derived from SV40-transformed mouse retina. This cell line is continuously dividing, non-differentiated, and, therefore, may not be directly comparable to the study presented in this paper, which is conducted on fully differentiated, mature, post mitotic photoreceptor cells in vivo. Furthermore, it is likely that the type and the severity of the initiating insult may determine whether the apoptosis is caspase-dependent or caspase-independent. It was recently reported that neurons exposed to a relatively mild nitrosative insult died by caspase-dependent apoptosis while those that were exposed FIG. 7. Light-induced retinal degeneration is caspase-3 independent. A, analysis of caspase-3 activity by immunoblotting analysis. Equivalent quantities of protein from cell lysates taken prior to light exposure (0 h), and 3, 6, and14 h of darkness following light exposure were resolved using SDS-PAGE and transferred to a nitrocellulose membrane. The presence of pro-caspase-3 (32 kDa) and the proteolytically active 17-kDa fragment were determined using an anti-caspase-3 antibody. A second Western blot was carried out with untreated and UV-treated 32D cells, which served as negative and positive controls, respectively, to determine the ability of this antibody to detect active murine caspase-3. The 32D protein lysates were extracted 16 h after a 10-min exposure to ultraviolet irradiation. These cells demonstrate the processing of pro-caspase-3 (32 kDa) to the active 17-kDa fragment as these cells undergo apoptosis. Two separate Western blots were necessary, because the Balb/c samples required longer exposure time to enable detection of pro-caspase-3. The 32-kDa pro-caspase species is present at all time points analyzed; however, the active 17-kDa fragment is absent up to 14 h. A representative result of three independent experiments is shown. B, analysis of caspase-3-like activity by detection of DEVD-pNA cleavage. The measurement of DEVD-pNA cleavage was performed in a spectrophotometric assay by monitoring the liberation of pNA due to caspase activity prior to light exposure (0 h), and 3, 6, and 14 h of darkness following light exposure. Untreated and anti-Fas IgM-treated Jurkat cells served as negative and positive controls, respectively. Jurkat cells were suspended at 5 ϫ 10 5 cells/ml and treated with apoptosis-inducing anti-Fas IgM antibody (300 ng ml Ϫ1 ) for 5 h. Data are expressed as the mean Ϯ S.E. of three independent experiments. The slight increase in activity at 3 h of light exposure is not statistically significant as determined by the student t test (p Ͻ 0.5). C, assessment of active caspase-3 content in light-induced and uninduced Balb/c retina. Active caspase-3 was assessed by FACS scan analysis using an anti-active caspase-3 antibody. Increased fluorescence indicates increased levels of active caspase-3. Untreated and Fas-treated Jurkat cells served as negative and positive controls, respectively. A representative result of three independent experiments is shown. D and E, Western blot analysis was carried out to detect the cleaved products of ICAD (D) and PARP (E). Equivalent quantities of protein from cell lysates taken prior to light exposure (0 h), and 3, 6, and14 h of darkness following light exposure were resolved using SDS-PAGE and transferred to a nitrocellulose membrane. A second Western blot was carried out with untreated and VP-16-treated 32D cells, which served as negative and positive controls, respectively, to determine the ability of the antibodies employed to detect the native and cleaved products of ICAD and PARP. The 32D protein lysates were extracted 16 h after treatment with VP-16 (5 g/ml). These cells demonstrate the processing of PARP (116 kDa) and ICAD (45 kDa) to the 85-and 12-kDa fragments, respectively, as these cells undergo apoptosis. to a high nitrosative insult still died by apoptosis, but caspase inhibitors were no longer protective (64). This is further evidence that the severity of the insult confers the shape of neuronal demise. Indeed, a caspase-independent pathway for apoptosis may be essential for neuronal programmed cell death, because the inherent biochemical and physiological characteristics of neuronal cells, including high lipid concentrations and energy requirements, make them particularly susceptible to free radical mediated insult.
It has been demonstrated in some models that inhibition of apoptosis is sufficient for the prevention of retinal degeneration (65). This is supported by recent studies on congenital stationary night blindness, a disease that involves loss of night vision due to blockage of the phototransduction cascade (66,67). However, daylight vision is maintained as the rod and cone photoreceptors do not undergo apoptosis. Therefore, by blocking the death of rod photoreceptors and the subsequent death of the cones, visual function may be retained. The data presented here not only provides a possible mechanism by which nitric oxide mediates light-induced retinal degeneration but also demonstrates that oxidative stress and elevated calcium contribute significantly to an apoptotic process, which does not involve caspase-3. These findings suggest that therapeutic strategies, which target initiating signals of apoptosis such as calcium channel blockers and antioxidants may be of more therapeutic benefit, than caspase inhibitors.