Blue Light–Induced Generation of Reactive Oxygen Species in Photoreceptor Ellipsoids Requires Mitochondrial Electron Transport

Abstract

purpose. To investigate whether photoreceptor ellipsoids generate reactive oxygen species (rOx) after blue light illumination.

methods. Cultured salamander photoreceptors were exposed to blue light (480 ± 10 nm; 10 mW/cm²). The light-induced catalytic redox activity in the culture was monitored with the use of 3,3′-diaminobenzidine (DAB). Tetramethylrhodamine ethyl ester (TMRE) and 2′,7′-dichlorodihydro-fluorescein acetate (DHF-DA) were used as probes to measure the mitochondrial membrane potential and intracellular rOx, respectively.

results. A significant deposit of DAB polymers was found in the culture after exposure to blue light. Basal levels of rOx were observed in photoreceptor ellipsoids when cells were stained with DHF-DA. This staining colocalized with TMRE. After exposure to blue light, a sharp increase of rOx immediately occurred in the ellipsoids of most photoreceptors. When the light intensity was reduced, the response kinetics of rOx generation were slowed down; however, comparable amounts of rOx were generated after a standard time of exposure to light. The production of rOx in photoreceptors was markedly decreased when an antioxidant mixture was included in the medium during exposure to light. Rotenone or antimycin A, the respiratory electron transport blockers at complex I and III, respectively, significantly suppressed the light-evoked generation of rOx.

conclusions. A robust amount of rOx is produced in the ellipsoid when photoreceptors are exposed to blue light. This light-induced effect is antioxidant sensitive and strongly coupled to mitochondrial electron transport. The cumulative effect of light on rOx generation over time may implicate a role for mitochondria in light-induced oxidative damage of photoreceptors.

Oxidative stress is a common contributor to many different neurodegenerative diseases, including retinal disorders such as age-related macular degeneration (AMD). ^{1 2} The retina is particularly susceptible to oxidative stress due to its high levels of polyunsaturated fatty acids, photosensitizers (e.g., retinal photosensitizers), and pigments (e.g., mitochondrial pigments), its high consumption of oxygen, and its exposure to visible light. ^{1 2 3 4} Several epidemiologic studies suggest that long-term history of exposure to light may have some impact on the incidence of AMD. ^{3 5} In addition, there is compelling evidence showing that exposure of the eye to visible light creates conditions of photo-oxidative stress, leading to oxidative damage of photoreceptors. ^{1 2 3 6 7 8}  

Blue light-induced damage of retinal cells has been widely used as a model for both in vivo and in vitro studies of retinal degeneration. Although several possible receptor candidate molecules have been proposed, including rhodopsin, ^{1 2 3 4 6 7 8 9} the molecular step leading to production of rOx remains elusive. However, we know that (1) photoreceptors contain numerous chromophores and photosensitizers, including the mitochondrial pigments in the inner segments, that, once activated, may induce or generate rOx ⁴ ; (2) the ellipsoid area of the photoreceptor inner segments are densely packed with mitochondria, ¹⁰ and this geometrical arrangement, the so-called mitochondrial cluster, impedes the stochastic behavior of individual mitochondria. This reinforces interactions between mitochondria, ¹¹ so the rOx generated locally may induce a widespread release of rOx throughout the mitochondrial network, a mechanism referred to as rOx-induced rOx release, ¹² leading to a robust rOx catastrophe; (3) mitochondria themselves have the capability to generate rOx ¹³ ; and (4) significant levels of flavoproteins are reduced in cells when exposed to blue light, resulting in the production of rOx, as shown in several nonretinal, cultured cells. ¹⁴  

Therefore, we sought to determine whether exposure to blue light would result in a significant increase in intracellular rOx in photoreceptor ellipsoids, and if so, what mechanism(s) were involved. 

Materials and Methods

Cell Dissociation and Culture

All procedures in this study adhered to the ARVO Statement for the Use of Animals in Ophthalmic and Vision Research. The protocol developed by MacLeish and Townes-Anderson ¹⁵ was used with slight modifications. The retina, separated from the eyecup, was incubated in low-calcium (0.5 mM CaCl₂/1.0 mM EDTA) Ringer containing 20 U/mL cysteine-pretreated papain (Worthington, Freehold, NJ) for 10 minutes at room temperature (RT). The retinas were then rinsed in culture medium and triturated. Cell suspensions were plated into a 96-well plate and maintained in a humidified chamber at 21°C in room air. 

Immunocytochemical Detection of Isolated Cones in Culture

Immunolabeling was performed on the isolated cells in culture as follows: (1) fixation with 10% buffered formalin for 10 minutes at RT, (2) three 3-minute rinses with phosphate-buffered saline (PBS), (3) permeabilization with 1% Triton X-100 in PBS for 10 minutes at RT, (4) three 3-minute rinse with PBS, (5) incubation in 10% normal goat serum in PBS (GSP) for 20 minutes at RT, (6) incubation in antibodies against mouse recoverin (a gift of Alexander Dizhoor) in 1.5% GSP overnight at 4°C, (6) three 3-minute rinses with PBS, and (7) incubation in secondary antibodies (Alexa Fluor 488; Molecular Probes, Eugene, OR) in 1.5% GSP for 60 minutes at 37°C. Immunofluorescence was visualized by an inverted microscope (Axiovert S100; Zeiss, Thornwood, NY) equipped with a cooled, digital charge-coupled device (CCD) camera (Roper Scientific, Trenton, NJ). 

Cytochemical Detection of rOx

Before exposure to light, cultured photoreceptors were incubated in the dark with 1 mM 3,3′-diaminobenzidine (DAB) ¹³ for 10 minutes at RT with gentle agitation. The 96-well plate was then placed under an inverted microscope, and cells were exposed to blue light (480 ± 10 nm; 10 mW/cm²) for 30 minutes through an objective lens underlying the selected well. 

Visualization of Mitochondrial Cluster in the Photoreceptors

All fluorescent dyes were purchased from Molecular Probes. Cells were loaded with 50 nM tetramethylrhodamine ethyl ester (TMRE) for 60 minutes ^{16 17} and perfused with culture medium containing TMRE (50 nM), with or without carbonyl cyanide m-chlorophenylhydrazone (CCCP) or antimycin A. To visualize cells, they were illuminated with a mercury light source with the intensity adjusted below 0.5 mW/cm² with neutral-density filters. The image acquisition rate was set at one frame per minute. These conditions reduce to a large extent the formation of rOx during visualization. Fluorescence images were obtained with an inverted microscope (Axiovert S100; Zeiss) using a cooled, digital CCD camera (Roper Scientific). A 520/610-nm band-pass filter was used for excitation and detection, and the images were further processed on computer (MetaView or MetaFluor software; Universal Imaging Corp., West Chester, PA). 

Fluorescence Detection of rOx

Cells were loaded in the dark with 25 μM dihydrofluorescein acetate (DHF-DA) for 30 minutes at RT and then perfused continuously for at least 10 minutes to wash away the dye remaining in the solution. A 480/535-nm band-pass filter was used for excitation/detection. Note that the wavelength used for fluorescence excitation is the same as that used to induce production of rOx. During the experiments, the intensity of fluorescent light was raised to 10 mW/cm² to induce rOx. Otherwise, the intensity was kept below 0.5 mW/cm², and images were acquired and processed as described earlier. 

Results

Salamander Photoreceptors in Culture

The cells used in this study were primary cultures of photoreceptors prepared from salamander retinas. During preparation, most photoreceptors (both rods and cones) lost their outer segments and distal axon terminals (rods). However, these cells survived, grew neurites, and were viable in culture for at least 7 days. The photoreceptors can be recognized by their ellipsoids, a morphologic marker specific to photoreceptors (Fig. 1A , left), and we estimated that approximately 80% of cells in culture are photoreceptors. However, it is difficult to distinguish the different subtypes of photoreceptors (e.g., rods, cones) by visual inspection once they lose the outer segments after dissociation and are grown in culture. To distinguish the cones from rods in our cultured photoreceptor preparation, an antibody against mouse recoverin, shown to selectively stain cones, but not rods, in intact salamander retina (Zhang J-H, Wu SM, unpublished result, 2002), was applied to our preparation. We found the rod-cone ratio in culture was approximately 6.0:4.0, similar to an earlier study conducted by MacLeish and Townes-Anderson, ¹⁵ in which they reported the rod-cone ratio to be 5.5:4.5. The preparation contained a very low percentage (<1%) of retinal pigment epithelium, as indicated by cell size, shape, and presence of melanin pigment. 

By electron microscopy, the ellipsoid has been shown to contain a high density of mitochondria that are packed together in a dense mass at the distal end of the inner segment. ¹⁰ This mitochondrial cluster could be easily visualized by staining the photoreceptors with TMRE, a lipophilic, cationic fluorescent probe. ^{16 17} After staining with TMRE, images made with a cooled, digital CCD camera showed a red mass approximately 1 to 3 μm in diameter localized within the ellipsoids (Fig. 1A , right). CCCP, a proton ionophore, or antimycin A, an inhibitor of respiratory electron transport, reduced the intensity of TMRE staining (Fig. 1B) , suggesting that the aggregated fluorescence in the ellipsoid represents the accumulation of TMRE molecules in the mitochondria. With this approach, photoreceptor mitochondria were reliably localized and photoreceptors could therefore be easily distinguished from other retinal cells. Most of the experiments were conducted within 2 to 4 days after plating. 

Effect of Light on Catalytic Redox Activity in Culture

Using an in situ staining method with DAB as a cytochemical probe to detect the redox level, ¹⁴ we found that catalytic redox activity in cultured photoreceptors was enhanced when cells were exposed to blue light (480 ± 10 nm; 10 mW/cm²). We chose 480 nm for our illumination experiments, simply because the same λ is used for excitation of dichlorofluorescein (DCF; see description later). DAB is a substrate for oxygen transferases (such as peroxidase and catalase) or oxidases, owing to its electron donor properties, and is water soluble. When oxidation occurs, the products are polymerized and become water insoluble, leaving osmiophilic, colored reaction products at or close to the reaction site. Figure 2 shows the cytochemical detection of redox activity induced by blue light in culture that includes DAB. When irradiated with blue light for 30 minutes in the presence of DAB, significant deposits of brown reaction product were found in the culture (Fig. 2A2) when compared with the control (Fig. 2A1) . The light dependency of the reaction product was clearly demonstrated when blue light was directed onto one well of a 96-well plate through an objective lens underlying the well. The cells that were exposed to blue light displayed a significant amount of brown precipitate, whereas the production of precipitate was greatly reduced in the region not directly exposed to light. A sharp contrast was clearly seen at the boundary (Fig. 2A3 , white dotted line) of the irradiated area. Note that no reaction product was generated in culture under blue-light illumination in the absence of DAB, and no light-induced depositions were seen in wells that contain DAB but with no cells present (data not shown). However, precipitate was dramatically reduced when cells were pretreated with 1% buffered formalin for 1 hour before exposure to light and DAB (Fig. 2A4) . This confirms that the light-induced oxidation of DAB is enzyme-mediated, in that it has been shown that catalase and peroxidase are highly sensitive to aldehyde fixation. ¹⁷ At higher magnification (400×), large amount of precipitate induced by light was seen in the ellipsoids of most photoreceptors (Fig. 2B , arrowheads) and also in the cell debris (Fig. 2B) . The mechanism for the light-induced production of DAB polymers by cell debris is not clear, but we noted that it is sensitive to fixative, suggesting that the cell debris may contain residual enzymes and/or organelles that are functionally redox active (compare Figs. 2A2 and 2A4 ). Qualitatively speaking, these data strongly suggest that a significant amount of rOx is generated by photoreceptors when exposed to blue-light illumination. To determine the source of the rOx generated within photoreceptors and, therefore, to investigate more directly the mechanism of rOx generation, a commonly used, membrane-permeable, rOx-sensitive fluorescent dye was used and fluorescence measurements were made with a cooled, digital CCD camera. The results are presented in the following sections. 

Fluorescence Detection of rOx Generated in Photoreceptors

Direct evidence that light stimulated rOx production in photoreceptors was obtained by fluorescence microscopy. To monitor intracellular rOx, cells were loaded in the dark with a nonfluorescent, membrane-permeable derivative of dihydrofluorescein, DHF-DA, ¹⁸ for 30 minutes and then continuously perfused for at least 10 minutes to wash away the dye remaining in the solution. DHF-DA will not fluoresce until first hydrolyzed by an intracellular esterase once inside the cell (DHF-DA→DCFH), and then oxidized by intracellular rOx (DCFH→DCF). Images were acquired with a cooled, digital CCD camera. After staining, a significant amount of DCF fluorescence was seen in the ellipsoids of photoreceptors (Fig. 3A3) . The staining with DCF (Fig. 3A3) was superimposed (Fig. 3A4) on that of TMRE (Fig. 3A2) , a mitochondria-specific dye. Taken together, these data demonstrate a basal level of redox activity within the mitochondria. To see whether exposure to light would cause further enhancement of DCF fluorescence, DCFH-loaded cells were exposed to blue-light illumination (480 ± 10 nm; 10 mW/cm²; 10 ms duration for each exposure to light) at 1 Hz. A significant enhancement of fluorescence intensity was observed in almost every photoreceptor (Fig. 3B1) , indicating that rOx was generated during the exposure to light. After reaching its peak amplitude, the fluorescence intensity gradually returned to its basal level. The exact mechanism of this decaying phase of fluorescence intensity during exposure to light is not clear. However, it may reflect a dynamic balance between the rate of oxidization of DCFH by the light-induced rOx in competition with the reduction of DCF by reducing agents present in photoreceptors, and also the rate at which oxidized dye escapes from the cells and is carried away by the perfusate, as suggested by others. ¹⁸ The amplitude of the peak response and the response kinetics vary between cells (Fig. 3B1) . A scatterplot was created, in which the peak response amplitude versus time to reach peak response after exposure to light was plotted to present the results obtained from five different experiments (n = 87; Fig. 3B2 ). Data obtained from the same experiment are labeled in the same color. A reciprocal relationship clearly exists. Cells with a smaller response to blue-light stimuli also display slower response kinetics. These data suggest that the variations in light-induced fluorescence between photoreceptors are not simply due to the experimental conditions, such as the loading efficiency of DHF-DA, or the age or condition of the cells. Instead, it may suggest that rOx production in photoreceptors involves more than one mechanism or pathway or, perhaps, subtype of photoreceptors, ¹³ and that each mechanism, pathway, or subtype of photoreceptors makes different contributions to the rOx induced by light. The results shown in Figure 3C also demonstrate that the light-induced effect was intensity dependent. Cells exposed to blue light at 1 Hz at a low intensity (0.5 mW/cm²; 100 ms duration) displayed slower response kinetic (red traces) compared with the control (blue traces; 10 mW/cm²; 10 ms duration). However, comparable amounts of rOx were generated after a minimum time (≥3 minutes) of exposure to light, suggesting a cumulative effect of light on rOx generation. 

To confirm that the increase in DCF fluorescence induced by light in photoreceptors actually reflects an increase in intracellular rOx level, the illumination experiments were performed in the presence of a mixture of antioxidants (1 mM ascorbic acid, 250 U/mL catalase, 1 mM Trolox and 500 μM 4-OH-TEMPO), a mixture that has been shown to quench rOx effectively. ³⁵ Cells pretreated for 30 minutes with antioxidant mixture were incubated for another 30 minutes with DHF-DA, followed by continuous perfusion for at least 10 minutes to remove the excess DHF-DA. Throughout the experiment, antioxidant mixture was included in the perfusate. Data were collected from four separate experiments with the antioxidant mixture (n = 122) and were plotted together with data obtained from the control (Fig. 3B2) for comparison. The resultant scatterplot is shown in Figure 3D . In the presence of the antioxidant mixture, the average response was reduced to 22%, whereas the time to reach peak amplitude was increased 27-fold compared with the control. 

To compare the generation of rOx production evoked by green light (535 ± 10 nm; 30 mW/cm²) with that evoked by blue light (480 ± 10 nm; 10 mW/cm²), the following experiment was performed. After 1 second of exposure to bright green light, the intensity of DCF fluorescence increased by 18% on average. By contrast, after exposure to bright blue light for 1 second, a 170% increase in DCF intensity on average was observed in the same cells (data not shown). These data suggest that blue light was much more efficient than green light in evoking rOx in our in vitro preparations of photoreceptors, which are missing a large proportion of the outer segments. 

Influence of Mitochondrial Electron Transport on Light-Induced Production of rOx

During light-induced formation of rOx, we observed that fluorescent DCF increased first in the ellipsoid and then spread throughout the whole cell. A typical example is shown in higher magnification in Figure 4 . In the dark, DCF fluorescence of low intensity was observed in the ellipsoid (Fig. 4A , left). After exposure to light, a significant increase of fluorescence was seen in the whole cell, particularly in the ellipsoid (Fig. 4A , right). 

It has been reported that mitochondria generate rOx due to the leakage of electrons during electron transport. ¹³ Therefore, we sought to determine whether the rOx induced by light was driven by mitochondrial respiration. We examined two electron transport blockers that act at different sites, 10 μM rotenone (r) at complex I and 10 μM antimycin (a) at complex III. ¹³ They were applied individually, along with 10 μM oligomycin (o), ¹³ an adenosine triphosphatase (ATPase) inhibitor, added to prevent the rundown of adenosine triphosphate (ATP), and their effect on light-induced rOx in photoreceptors was examined. Cells pretreated for 30 minutes with r + o or a + o were incubated for another 30 minutes with DHF-DA, followed by continuous perfusion for at least 10 minutes to remove the excess DHF-DA. During these experiments, r + o or a + o was included in the perfusate. Representative experiments are shown in Figures 4B1 (r + o) and 4B2 (a + o). Compared with the control (Fig. 3B1) , the average response (Figs. 3B1 4B1 4B2 , thick traces) from cells treated with r + o or a + o was significantly lower. Because of the inherent variation in fluorescence response in the control (Fig. 3B2) , scatterplots were also prepared (Figs. 4B3 , r + o, and 4B4 , a + o). For comparison, the data collected from five and four experiments of r + o and a + o, respectively, were plotted together with data obtained from the control (Fig. 3B2) . Clearly, the light-induced increase in DCF fluorescence was significantly reduced or suppressed by these respiration blockers. On average, the light-evoked DCF fluorescence was reduced to 35% by r + o (n = 100) and to 9% by a + o (n = 45). It appears that a + o has a greater inhibitory effect than r + o, which may simply reflect the site of action of the inhibitors relative to that of light (see the Discussion section). Taken together, these results suggest that the light-induced generation of rOx in photoreceptors is strongly coupled with mitochondrial electron transport. 

A potential problem with using DHF-DA as the rOx sensor is that high-intensity excitation light may cause the photoconversion of DCFH to DCF. To determine to what extent the DCF detected in our illumination experiments (Figs. 3 4) was caused by photoconversion, we conducted the following experiments. DCFH, obtained from DHF-DA after removing the acetate ester with esterase (33 U/mL), was exposed to the same light stimulus as used in the illumination experiments (Figs. 3 4) , both in the presence and absence of antimycin or antioxidant mixture. We found that in the cell-free system, antimcyin A or antioxidant mixture has no effect on the increase in DCF fluorescence induced by light (data not shown). This is in contrast to the observations shown in Figures 3 and 4 where the light-evoked increase in fluorescence intensity of DCF in photoreceptors is markedly suppressed by an antioxidant mixture or antimycin A. Therefore, we conclude that the increase in DCF fluorescence in photoreceptors during exposure to blue light is mainly due to oxidation caused by the rOx generated in photoreceptors. 

Discussion

There were two major findings in this study: Significant amounts of rOx were generated in photoreceptor ellipsoids (a region that is enriched with mitochondria) after exposure to blue light, as measured with the oxidation-sensitive probe DAB or the fluorescent dye DHF-DA, and this light-induced rOx was blocked by inhibitors of respiratory electron transport, suggesting that light-induced production of rOx is strongly coupled to mitochondrial electron transport. 

How does blue light lead to production of rOx in isolated photoreceptors that retain inner segments but with a minimal or absent outer segment? To address this question, two interrelated molecular pathways must be considered: the excitation of receptor by blue light and the subsequent signaling to the mitochondrial electron transport chain. Given that blue light was more efficient than green light in the generation of rOx in our culture preparation, we considered at least three candidate molecules as possible blue light receptors: the visual pigments (e.g., rhodopsin) in the inner segments, ²¹ flavins concentrated in peroxisomes and/or mitochondria, ¹⁴ and mitochondrial pigments (e.g., cytochrome c). ¹³ A recent study conducted on isolated frog rods showed that the rhodopsin in the outer segment, when activated by blue light, can produce oxidative stress, not involving the biochemical cascade for visual signal transduction. ⁹ Whether a similar mechanism can be applied to the visual pigments present in the inner segments is not known. Alternatively, photoreduction of flavins or inactivation of the mitochondrial pigments (e.g., cytochrome) may disturb the normal redox potential of cells, leading to production of rOx. Although the actual molecular identity of the blue-light receptor is not yet determined, it is clear that signals downstream to receptor activation involve mitochondrial electron transport, as demonstrated by our data showing that the light-induced production of rOx was suppressed when respiratory electron transport blockers were included in the medium during exposure to light. 

A further question is how mitochondrial electron transport is involved in rOx production. The factors that control the rate of rOx production by mitochondria in photoreceptors are poorly understood, and a full discussion of the possible mechanisms involved is beyond the scope of this article. However, two important clues can be inferred from our data if the mitochondria is considered to be the main subcellular site that generates rOx: (1) The number of reduced respiratory chain carriers must be increased during exposure to light; light could have a direct effect on the half-life of the reduction state of the electron carriers or an indirect effect through second messengers; and (2) the site(s) that generate rOx in response to light are downstream of complex III, because light-induced rOx was partially suppressed by rotenone, an inhibitor of respiration at complex I, ¹³ and was completely suppressed by antimycin, an inhibitor of respiration at complex III. ¹³ A series of studies published by Chen ²² showed that exposure of the retina to short-wave optical radiation inhibits cytochrome oxidase (see the 1993 review article ²² and the relevant articles cited therein). An increased production of rOx by mitochondria, due to a decrease of V _max of cytochrome oxides, is expected. ²³ Whether cytochrome oxidase is involved in generation of rOx by light, as shown in the current study, warrants further investigation. 

The functional consequence on light-induced photoreceptor damage of the rOx generated by exposure to blue light may have several aspects. Tissue damage produced by rOx, such as hydrogen peroxide, superoxide, and hydroxyl radicals, has been implicated in the pathogenesis of both acute and chronic insults of the central nervous system(CNS), and multiple studies suggest that the rOx generated by mitochondria is involved. ^{24 25 26 27} Alternatively, rOx may act as a second messenger; a series of reports by Grimm et al., ²⁸ who used a transgenic mouse model, suggest that c-Fos may be involved in light-induced photoreceptor apoptosis. rOx has been implicated as having a role in the regulation of c-Fos expression. ²⁹ Also, Krishnamoorthy et al. ³⁰ suggested that the downregulation of NF-κB is mediated by rOx and that the presence of the NF-κB RelA subunit is essential for protection of the photoreceptor against rOx-induced apoptosis. However, in their study, the subcellular site of rOx generation was not examined. It has been proposed that the redox potential in the nucleus is able to alter transcriptional responses that, in turn, decide the apoptosis resistance. ^{31 32} Given that blue light has been shown to trigger apoptosis in photoreceptors, ^{2 28 33 34} it will be interesting to determine whether the mechanisms proposed in several cellular models of rOx-induced apoptosis ^{35 36 37 38} can be applied to the photoreceptor apoptosis induced by light. Clearly, our findings suggest a role for mitochondria in cases of retinal degeneration caused by light-induced oxidative damage. 

Our cultured cell preparation contains both types of photoreceptors: rods (∼60%) and cones (∼40%). After exposure to blue light, a significant increase in DCF fluorescence, although variable, was seen in virtually every photoreceptor studied, suggesting it was very likely that we were sampling both types of photoreceptors. However, the question of whether the variations depicted in Figure 3B2 arose from different subtypes of photoreceptors, with each subtype of photoreceptors having a different sensitivity to blue light, is not clear. The answer to this question is potentially important because, in vivo, photoreceptors are electrically coupled together through gap junctions, and gap junction intercellular communication (GJIC) has been shown to play an important role in maintaining tissue homeostasis. ^{39 40} The rOx itself, or its metabolites, or the downstream messengers initiated by rOx, may propagate through GJIC to neighboring photoreceptors, resulting in spreading the so-called “kiss-of-life” or “kiss-of-death” signals, a phenomenon called the bystander effect. ^{39 40}  

To investigate the question of photoreceptor subtypes, an ideal experiment would be to record from photoreceptors prelabeled with noninvasive surface markers specific to each subtype of photoreceptors. Our initial trial, using 4,4′-diisothiocyanostilbene-2,2′-disulfonic acid (DIDS), a fluorescent stilbene dye reported to bind selectively to cones in Xenopus retina, ⁴¹ failed to reveal any staining in our culture, and we are not aware of any other surface markers that can be used for this purpose. The antibodies against recoverin used in this study for selective labeling of cone cells would not be suitable for this experiment, because staining with these antibodies requires prefixation and permeabilization, which kills the cells. We also tried to solve this problem by growing the dissociated photoreceptors on a sterile coverslip that has a microgrid design enabling quick location of cells, allowing individual cells to be relocated after fluorescence measurements (Cellocate; Eppendorf, Fremont, CA). The poor adhesion of cultured photoreceptors to this type of coverslip, even after coating with poly-l-lysine or Sal-1, ¹⁵ made it technically impossible to use this procedure. 

Various amphibian retinal preparations have been used for vision research for many years because of their several advantages (e.g., accessibility due to large cell size, availability for long-term cell culture, and the well-known electrical properties of retinal cells) over that of mammalian retina. ¹⁷ In addition, the success of using transgenic Xenopus to study photoreceptor degeneration has made this preparation even more attractive, ⁴² but is the amphibian retina a suitable model for the studies of light-induced damage? Given that very few studies of light-induced damage have been conducted on this preparation, the answer remains elusive. In an early in vivo study conducted on frogs by one of the authors, after 14 months under constant room light, swelling was observed in the RPE, whereas photoreceptors appeared normal. In addition, membrane renewal rates were altered, suggesting an alternation in inner segment metabolism. ⁴³ These studies suggest that in the living animals with RPE intact and all the protective mechanisms in place, metabolic alterations occur but are not extensive enough to initiate cell death. However, several studies conducted on the same species reached different conclusions. ^{44 45 46} In summary, it was found in these studies that the light induced significant accumulation of lipid peroxidation products in photoreceptors, and the lipid peroxidation was a factor in the degeneration of photoreceptors. ^{44 45 46} This notion is reinforced by a recent study conducted on isolated frog rods. A significant increase in rOx was detected in both outer and inner segments of photoreceptors. The generation of rOx in the outer segments required rhodopsin activation, whereas the source of rOx generated in the photoreceptor ellipsoid was not determined, but the results suggested that mitochondrial metabolism may be involved. ⁹ Our work confirms this suggestion. 

In conclusion, after exposure to blue light the production of rOx in different compartments (e.g., outer segment versus inner segment) of photoreceptors involves different mechanisms, and the pathophysiological role that rOx derived from different compartments plays in light damage warrants further investigation. 

 Supported in part by grants from Prevent Blindness America, the Lions Eye Bank Foundation, and International Retina Research Foundation; Core Grant P30-EY02520 and Grant EY04446 from the National Eye Institute; and a grant from the Retina Research Foundation of Houston.

 Submitted for publication July 30, 2002; revised October 10 and October 16, 2002; accepted October 17, 2002.

 Disclosure: J.-H. Yang, None; S.F. Basinger, None; R.L. Gross, None; S.M. Wu, 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.

 

The photoreceptors in culture can be identified by the ellipsoids, which are densely packed with mitochondria. (A) Micrograph of a 24-hour culture of photoreceptors (left) and a fluorescence image (right) of the same cells stained with TMRE, a mitochondria-specific dye, used to monitor the membrane potential (ΔΨ_m) of the mitochondria. The photoreceptor image showed a fluorescent mass approximately 1 to 3 μm in diameter localized to the ellipsoid. Bar, 10 μm. (B) Time course of ΔΨ_m photoreceptors stained with TMRE. ΔΨ_m, measured in arbitrary fluorescence (F) units, and recorded in cultured photoreceptors perfused with culture medium (top) or with the addition of 10 μM CCCP (middle) or antimycin A (bottom) in the presence of TMRE. Arrows: starting point of drug application; each trace represents the measurement from one cell. CCCP or antimycin A caused a significant reduction in TMRE fluorescence, indicating a decrement in ΔΨ_m, owing to the abolishment of the proton gradient.

Figure 2.

 

Light induced the formation of oxidative products of DAB. (A1, A2) Cells exposed to blue light (480 ± 10 nm; 10 mW/cm²) for 30 minutes displayed a significant amount of precipitate, the oxidative product of DAB (A2), compared with the control (A1). (A3) The formation of the DAB oxidation product was light dependent. Blue light was directed onto one well of a 96-well plate through an objective lens. The cells that were directly exposed to light showed a significant amount of precipitate; by contrast, little precipitate was seen in the region that was outside the projection field of the lens. A sharp contrast was seen at the boundary of the irradiated area (dotted line). (A4) The production of DAB oxidation product was significantly reduced when the cells were pretreated with 1% buffered formalin for 1 hour before exposure to light+DAB, suggesting that the light-induced formation of oxidative polymer may be enzyme mediated. (B) Light-induced precipitate was seen in the ellipsoids of the photoreceptors (arrowheads) and in the cell debris. Magnification, ×400; bar, 10 μm.

Figure 3.

 

Fluorescence detection of light-induced rOx in photoreceptors. (A1) A photomicrograph of cultured salamander photoreceptors. Bar, 10 μm. (A2–A4) Colocalization of DCF fluorescence with that of TMRE in cultured photoreceptors. After loading cells with DHF-DA, bright green fluorescence (DCF) appeared in the ellipsoids of photoreceptors (A3). The fluorescence in the photoreceptors was localized to the mitochondria because the fluorescence of DCF (green; A3) was colocalized (orange; A4) with that of TMRE (red; A2). (B1) The DCF fluorescence (F) in photoreceptors was significantly enhanced when cells were exposed to blue light. The intensity of DCF fluorescence on exposure to light was quantified (arbitrary units) and plotted versus time after exposure to light (left). Each trace represents the measurement from a single cell and the thick black trace is the average of the multiple traces. Images of photoreceptors acquired before (inset 1) and after (inset 2) exposure to light. Pseudocolor was used to represent the fluorescence intensity of DCF. (B2) The data collected from five separate experiments are plotted as a scatterplot, in which the peak amplitude of the DCF response is plotted versus time to reach the peak response of DCF after exposure to light. Data obtained from the same experiment are labeled in the same color. (C) The kinetics of rOx production by light was intensity dependent. Cells that were exposed to light of a lower intensity (0.5 mW/cm²; 100-ms duration of exposure to light at 1 Hz) displayed slower response kinetics (red traces) compared with the control (blue traces; 10 mW/cm²; 10 ms duration of exposure to light at 1 Hz). However, comparable amounts of rOx were generated in photoreceptors after a minimum time (≥3 minutes) of exposure to light, suggesting that the effect of light on the generation of rOx is cumulative. (D) Antioxidant mixture suppressed the light-induced increase of DCF in photoreceptors. The data collected from four separate experiments with antioxidant mixture are plotted as a scatterplot in which the peak amplitude of the DCF response is plotted versus the time to reach the peak response of DCF after exposure to light. Data from the same experiment are labeled in the same color. For comparison, the data from the control (B2) are also plotted in the same plot (yellow circles). Compared with the control, the light-induced increase of DCF was greatly reduced and accompanied by a slowdown in response kinetics.

Figure 4.

 

Light-induced produc-tion of rOx is coupled with mitochondrial electron transport. (A) Images of the same photoreceptor acquired before (left) and after (right) exposure to light. A significant enhancement of DCF was noted, particularly in the ellipsoid. (B1–B4) Inhibitors of respiratory electron transport blocked the light-induced generation of rOx in photoreceptors. Compared with the control (see Fig. 3B1 ), the light-induced increase of DCF was largely reduced by 10 μM of rotenone (r), an inhibitor of respiratory electron transport at complex I (B1) or 10 μM of antimycin A (a), an inhibitor at complex III (B2), which were applied individually with 10 μM oligomycin (o), an inhibitor of ATPase. Each trace represents the measurement from a single cell, and the thick black trace is the average of the multiple traces from the same experiment. (B3, B4) The data collected from five and four separate experiments of r + o and a + o, respectively, are plotted as a scatterplot in which the peak amplitude of the DCF response is plotted versus time to reach the peak response of DCF after exposure to light. Data obtained from the same experiment are labeled in the same color. For comparison, the data obtained from the control (Fig. 3B2) are plotted in the same plot (yellow circles). In the presence of r + o (B3) and a + o (B4), the light response was significantly suppressed in most of the photoreceptors (B4). The arrow in B4 indicates that the change of DCF fluorescence in 10 of 45 photoreceptors measured was not detectable.