Synthesis of Sphingosine Is Essential for Oxidative Stress-Induced Apoptosis of Photoreceptors

Abstract

Purpose.: Oxidative stress is involved in inducing apoptosis of photoreceptors in many retinal neurodegenerative diseases. It has been shown that oxidative stress increases in photoreceptors the synthesis of ceramide, a sphingolipid precursor that then activates apoptosis. In several cell types, ceramide is converted by ceramidases to sphingosine (Sph), another apoptosis mediator; hence, this study was undertaken to determine whether Sph participates in triggering photoreceptor apoptosis.

Methods.: Rat retina neurons were incubated with [³H]palmitic acid and treated with the oxidant paraquat (PQ) to evaluate Sph synthesis. Sph was added to cultures with or without docosahexaenoic acid (DHA), the major retina polyunsaturated fatty acid and a photoreceptor survival factor, to evaluate apoptosis. Synthesis of Sph and sphingosine-1-phosphate (S1P), a prosurvival signal, were inhibited with alkaline ceramidase or sphingosine kinase inhibitors, respectively, before adding PQ, C₂-ceramide, or Sph. Apoptosis, mitochondrial membrane polarization, cytochrome c localization, and reactive oxygen species (ROS) production were determined.

Results.: PQ increased [³H]Sph synthesis in photoreceptors and blocking this synthesis by inhibiting alkaline ceramidase decreased PQ-induced apoptosis. Addition of Sph induced photoreceptor apoptosis, increased ROS production, and promoted cytochrome c release from mitochondria. Although DHA prevented this apoptosis, inhibiting Sph conversion to S1P blocked DHA protection.

Conclusions.: These results suggest that oxidative stress enhances formation of ceramide and its subsequent breakdown to Sph; ceramide and/or Sph would then trigger photoreceptor apoptosis. Preventing Sph synthesis or promoting its phosphorylation to S1P rescued photoreceptors, suggesting that Sph is a mediator of their apoptosis and modulation of Sph metabolism may be crucial for promoting photoreceptor survival.

Apoptosis of photoreceptors leads to retinal dysfunction in neurodegenerative disorders, such as retinitis pigmentosa and macular degeneration. ^1,2 Oxidative stress has a decisive role in activating this death; expression of oxidative stress-related genes increases during the progression of retinal neurodegeneration ³ ; production of reactive oxygen species (ROS) is a common feature during photoreceptor cell death in several animal models of retina degeneration ^4,5 ; and antioxidants ameliorate the progression of this neurodegeneration. ^6–8 Oxidative stress also triggers the apoptosis of photoreceptors in vitro. ⁹  

Many of the steps leading to photoreceptor death have been elucidated; however, the mediators involved in activating this death remain uncertain. A novel group of signaling molecules are the sphingolipids ceramide, sphingosine (Sph), and sphingosine-1-phosphate (S1P), bioactive lipids that participate in regulating cell death, survival, and proliferation in several cell types. ^10–13 Ceramide is an endogenous mediator of apoptosis. ^14–16 Trophic factor removal, oxidative stress, and other stress signals increase intracellular ceramide, ^11,17–19 which inhibits proliferation and induces death in many cell systems, including neurons. ^20,21 Sphingosine (Sph), synthesized through the breakdown of ceramide catalyzed by ceramidases, also increases in the early steps of the apoptotic pathway, and its addition activates apoptosis and inhibits proliferation in many cell types. ²² Different apoptotic stimuli induce a rapid increase in ceramide and Sph—the elevation of the former preceding that of the later, suggesting that Sph may result from ceramide degradation. ^23,24 In turn, Sph phosphorylation generates S1P, with roles completely opposite to those of ceramide and Sph, promoting survival and proliferation. 

Little is known concerning the involvement of sphingolipids in the activation of photoreceptor death. Decreasing ceramide levels prevents photoreceptor death in a Drosophila model of photoreceptor degeneration. ²⁵ The finding that a form of retinitis pigmentosa originates in a mutation of a gene encoding an enzyme of sphingolipid metabolism, the overexpression of which protects cells from oxidative stress-induced apoptosis, ^26,27 points to a role for these lipids in retinal degeneration. We have shown that oxidative stress increases ceramide synthesis in rat retina photoreceptors before activating apoptosis, whereas inhibiting this increase prevents their death. ²⁸ Oxidative stress also augments ceramide levels and induces apoptosis in the 661W cone-like cell line. ²⁹ These findings support ceramide as a key mediator in apoptosis induction in photoreceptors. 

Sph functions in the retina are virtually unknown. In Drosophila phototransduction mutants, an Sph-enriched diet enhances photoreceptor degeneration ²⁵ whereas overexpression of neutral ceramidase suppresses retinal degeneration. ^25,30 We investigated whether Sph participates in inducing photoreceptor death in the vertebrate retina. Our results show that oxidative stress increased the synthesis of [³H]Sph in rat retina photoreceptors in culture, before the onset of apoptosis; conversely, blocking Sph synthesis by inhibiting alkaline ceramidase protected photoreceptors from oxidative stress-induced apoptosis. Exogenous addition of Sph triggered photoreceptor apoptosis. Docosahexaenoic acid (DHA), the major polyunsaturated fatty acid in the retina and a survival factor for photoreceptors, prevented this death, but its protection was blocked by inhibiting Sph phosphorylation, and hence S1P synthesis. These results suggest for the first time that Sph may be a crucial mediator of photoreceptor death on oxidative damage. 

Materials and Methods

Albino Wistar rats were bred in our own colony. All proceedings concerning animal use were performed in accordance with the guidelines of the ARVO Statement for the Use of Animals in Ophthalmic and Vision Research. Plastic culture 35-mm diameter dishes (CellStar) were from Greiner Bio-One (Frickenhausen, Germany). Dulbecco's modified Eagle's medium (DME) was from Life Technologies (Grand Island, NY). Bovine serum albumin (fraction V; fatty acid-free; low endotoxin, tissue culture tested), paraquat dichloride (PQ; methyl viologen, 1,1′-dimethyl4,4′-bipyridinium dichloride), poly-l-ornithine, trypsin, trypsin inhibitor, transferrin, hydrocortisone, putrescine, insulin, polyornithine, selenium, gentamicin, 4,6-diamidino-2-phenylindole (DAPI), monoclonal anti-syntaxin clone HPC-1, docosahexaenoic acid (DHA), and paraformaldehyde were from Sigma Chemical Co. (St. Louis, MO). C₂-acetylsphingosine (C₂-ceramide), Sph, (1S,2R)-d-erythro-2-(N-myristoyl-amino)1-phenyl-1-propanol (MAPP), dl-threo-dihydro-sphingosine (DHS), fumonisin B1, and kainic acid were from Biomol (Plymouth Meeting, PA). A tyramide signal amplification kit and [9,10-³H]palmitic acid (45 mCi/mmol) were from NEN (Boston, MA). Monoclonal antibody against rhodopsin, Rho4D2 and polyclonal antibody anti-SphK1 were generously supplied by Robert Molday (University of British Columbia, Vancouver, BC, Canada) and Lina Obeid (Medical University of South Carolina, Charleston, SC), respectively. Rabbit polyclonal antibody against cytochrome c was from Santa Cruz Biotechnology (Santa Cruz, CA). Secondary antibody Cy2-conjugated goat anti-rabbit was from Jackson ImmunoResearch (West Grove, PA). A red mitochondria stain (MitoTracKer Red CMXRos), Alexa Fluor-488–conjugated annexin V, propidium iodide (PI), terminal deoxynucleotidyl transferase, recombinant 5-bromo-2-deoxyuridine-5-triphosphate (BrdUTP), terminal deoxy-nucleotidyl transferase (TdT) buffer, and the fluorescent dye 2,7-dichloro-dihydrofluorescein diacetate (H₂DCFDA) were from Invitrogen-Molecular Probes (Eugene, OR). Solvents were HPLC grade and all other reagents were analytical grade. 

Neuronal Cultures

Purified cultures of rat retinal neurons were prepared as previously described. ^31,32 Approximately 0.5 × 10⁵ cells/cm² were seeded on 35-mm diameter dishes, which had previously been sequentially treated with polyornithine- and Schwannoma-conditioned medium. ³³ The cells were then incubated in a chemically defined medium. ³¹  

Addition of PQ

PQ (48 μM, final concentration in the incubation medium, in calcium-magnesium–free Hanks' solution) was added to 3-day cultures. Neurons were incubated for different times and then scraped for lipid extraction or fixed. 

Addition of C₂-ceramide

Because of its extremely hydrophobic characteristics, natural, long-chain ceramides are frequently replaced in in vitro experiments by short-chain acyl ceramides like acetylsphingosine (C₂-ceramide, C₂-Cer), which is more water soluble and cell-permeable and has been shown to produce apoptosis in several cell systems, ³⁴ including photoreceptors. ²⁸ A 10 μM C₂-Cer solution was added to 3-day cultures, in an ethanol:calcium-magnesium–free Hanks' solution (1:8). ²⁸ The same volume of the vehicle was added to the control cultures. The cultures were then incubated for 24 hours before fixation. 

Inhibition of Sph Synthesis

To test whether Sph was required for ceramide or PQ-induced apoptosis, we inhibited ceramide hydrolysis to Sph using MAPP, a cell-permeable, specific inhibitor of alkaline ceramidase. ³⁵ Three-day cultures were treated with 10 μM MAPP and 30 to 60 minutes later either with C₂-Cer or with PQ for 24 hours. 

Effect of PQ on Sph Synthesis

To evaluate Sph synthesis exclusively in photoreceptors, 3-day neuronal cultures were treated with kainic acid (0.25 mM) for 3 hours, to selectively induce amacrine cell death. ³⁶ [³H]palmitic acid (2.5 μCi/dish), resuspended in the neuronal incubation medium and complexed with BSA (2:1 molar ratio), was added to the cultures, which were returned to the incubator for 30 minutes to allow fatty acid uptake. Cultures were then treated with or without 48 μM PQ, for 45 minutes 2, 4, and 6 hours. The incubation medium was removed and neuronal lipids were extracted. 

Addition of Sph

Sph was solubilized in an ethanol:calcium-magnesium–free Hanks' solution (1:100), and the effect on photoreceptor apoptosis of concentrations from 1 to 10 μM was evaluated. For subsequent experiments, 3-day neuronal cultures were treated with 5 μM Sph (final concentration in the culture medium) for 6, 12, or 24 hours; a similar volume of the vehicle was added to control cultures. To evaluate whether Sph-induced apoptosis is due to its conversion to ceramide, ceramide synthase and alkaline ceramidase, which can catalyze this conversion, were inhibited adding 10 μM fumonisin B1 or 10 μM MAPP, respectively, before treatment with Sph. 

Effect of DHA

DHA, at a 6.7 μM final concentration, was added at day 1 in culture, complexed with BSA, in a 2:1 (fatty acid to BSA) molar ratio. ³² The same volume of a BSA solution was added to control cultures. To evaluate whether DHA stimulated the synthesis of S1P, catalyzed by sphingosine kinase 1 (SphK1) to protect photoreceptors from Sph-induced death, 1 μM DHS, a SphK1 inhibitor, was added to cultures, with or without DHA, before Sph treatment. 

Effect of DHA on Sphingosine Kinase Expression

To evaluate DHA effect on SphK1 expression in photoreceptors, amacrine neurons were eliminated by kainic acid treatment. ³⁶ Fresh media with or without DHA was added at day 3 and 0.1 μM MG-132, ³⁷ a proteosome inhibitor, was added 6 hours later to inhibit SphK1 degradation and improve its detection. Cells were fixed 18 hours later. SphK1 expression was evaluated by immunocytochemistry with a specific polyclonal antibody anti-SphK1. 

Lipid Extraction and Analysis

Neurons were washed with ice-cold phosphate-buffered saline (PBS; 0.9% NaCl in 0.01 M NaH₂PO₄ [pH 7.4]), scraped and transferred to glass tubes. After a 10-minute centrifugation at 1000 rpm, the supernatant was removed, cells were resuspended in 2% acetic acid in methanol and lipids were extracted. ³⁸ Mild alkaline treatment was then performed, incubating lipid extracts with 0.1 M NaOH in a methanol solution for 10 minutes at 37°C. Incubation was stopped by adding an equal volume of 0.1 N HCl and chloroform. Lipids were then separated by TLC using butanol:acetic acid:water (3:1:1). ³⁹ Unlabeled lipids obtained from bovine retina were used as carriers. Lipid spots, visualized with iodine vapors, were identified with standards and scraped to vials to determine the incorporated radioactivity by liquid scintillation counting. 

Immunocytochemical Methods

Cultures were fixed for at least 1 hour with 2% paraformaldehyde in PBS, at room temperature and permeated with Triton X-100 (0.1% in PBS) for 15 minutes. The photoreceptors were identified by immunocytochemistry with the monoclonal antibody Rho4D2, by their morphology and other criteria as described elsewhere. ^40,41 Amacrine cells were identified with the anti-syntaxin monoclonal antibody, HPC-1. ^32,42 A Cy-2-conjugated goat anti-mouse was used as a secondary antibody. Tyramide signal amplification was occasionally used to improve visualization, according to the procedure described by the manufacturer. Control immunocytochemistry was performed by omitting either the primary or the secondary antibody from the cells. Cultures were analyzed by phase and fluorescence microscopy (Eclipse E600 microscope with a C-C Phase Contrast Turrent Condenser; and a Y-FL Epi-Fluorescence Attachment; Nikon, Tokyo, Japan; and a Laser Scanning Confocal Microscope [LSCM]; DMIRE2/TSCSP2; Leica, Deerfield, IL) with a 63× water objective; x–y sections were collected and processed with LCS software (Leica). 

Evaluation of Apoptosis

Apoptosis was determined by terminal deoxynucleotide transferase dUTP nick end labeling method (TUNEL) to evaluate DNA integrity; DAPI staining to assess the amount of fragmented or pyknotic nuclei and annexin/PI labeling, to evaluate phosphatidylserine translocation to the plasma membrane outer leaflet, a hallmark of apoptosis. 

For TUNEL staining, the cells were fixed at day 4 with 2% PF for 15 minutes and then stored in 70% ethanol for 48 hours at −20°C. The cells were then preincubated with 1× TdT buffer for 15 minutes and incubated with the TdT reaction mixture (0.05 mM BrdUTP, 0.3 U/μL TdT in TdT buffer) at 37°C in a humidified atmosphere for 1 hour. The reaction was stopped by a 15-minute incubation with stop buffer (300 mM NaCl, 30 mM sodium citrate [pH 7.4]) at room temperature. The negative control was prepared by omitting TdT. BrdU uptake was determined with an anti-bromodeoxyuridine (anti-BrdU) monoclonal antibody. 

Nuclear integrity was established by staining the cell nuclei with DAPI, a fluorescent dye that binds to DNA. Briefly, the cells were permeated with 0.1% Triton X-100, washed with PBS, and incubated with DAPI for 20 minutes. The amount of photoreceptors or amacrine cells with fragmented or condensed (pyknotic) nuclei was counted in cultures double-labeled with DAPI and with either Rho4D2 or HPC-1, to unambiguously identify cells as either photoreceptors or amacrine neurons, respectively, and thus establish the total amount of each cell type. The percentage of photoreceptors or amacrine neurons with fragmented or pyknotic nuclei was then calculated, taking into account the amount of Rho4D2- or HPC-1-labeled cells, respectively. 

For annexin V staining, cultures were incubated with Sph for 12 hours; the incubation medium was removed and the dishes were washed twice with ice-cold PBS. The cells were then incubated with a 1:4 dilution of annexin V conjugate in annexin-binding buffer (10 mM HEPES, 140 mM NaCl, and 2.5 mM CaCl₂ [pH 7.4]) at room temperature in the dark for 15 minutes. PI was added immediately after, and the cells were incubated for another 15 minutes in the same conditions. They were then washed in cold PBS and fixed in PF in annexin-binding buffer for 1 hour. After they were washed, labeling with annexin V, PI, or both was analyzed. 

Evaluation of Mitochondrial Membrane Potential and Cytochrome c Translocation

To assess the amount of cells preserving their mitochondrial membrane potential, the cultures were incubated for 30 minutes before fixation with the fluorescent probe (0.1 μg/mL; MitoTracker Red; Invitrogen-Molecular Probes). The amount of photoreceptors displaying fluorescent mitochondria with respect to the total number of photoreceptors was determined. 

To evaluate whether Sph induces cytochrome c release from mitochondria, the cells were incubated with the mitochondrial stain 12 hours after Sph addition, fixed with methanol for 15 minutes and immunocytochemical labeling of cytochrome c was performed with a polyclonal antibody. The localization of mitochondria and cytochrome c was determined by confocal microscopy. 

Determination of ROS

The effect of Sph on ROS production was evaluated by treating 3-day cultures with or without Sph for 5.5 hours and then incubating them with 10 μM 2,7- H₂DCFDA for 30 minutes at 37°C in the dark. The cells were washed twice with cold PBS, scraped with 100 μL DMSO, and resuspended in PBS. The protein content was determined, ⁴³ and the same amount of protein per sample was then used for measuring fluorescence. Fluorometric measurements were performed (SLM model 4800 fluorometer; SLM Instruments, Urbana, IL) at 480 to 520 nm. The temperature was set at 25°C in a thermostat-controlled circulating water bath (Haake, Darmstadt, Germany). 

Statistical Analysis

For cytochemical studies, 10 fields per sample, randomly chosen, were analyzed in each case. Each value represents the average of at least three experiments, with three to four dishes for each condition ± SD. Statistical significance was determined by Student's two-tailed t-test. 

Results

Effect of PQ on Sph Synthesis

PQ induces the apoptosis of photoreceptor and amacrine cells, the two major cell types in our pure neuronal cultures. ⁹ To evaluate whether oxidative stress prompted an increase in Sph levels in photoreceptors, we eliminated amacrine neurons before adding [³H]palmitic acid and PQ to the photoreceptor-enriched cultures. PQ rapidly increased Sph levels; 45 minutes after its addition, [³H]Sph was almost two times higher in the PQ-treated photoreceptors than in the control cells (Fig. 1A). At this time point, neurons still showed no indication of apoptosis, which takes longer to be evident. ²⁸ The elevation in [³H]Sph levels was transient, since no significant differences with control cells were observed 2, 4, or 6 hours after PQ addition (not shown). This suggests that PQ rapidly promoted Sph synthesis, which preceded the onset of apoptosis. 

Figure 1.

 

Effect of Sph synthesis in the induction of apoptosis in retina photoreceptors. (A) Retina neuronal cultures were treated at day 3 with kainic acid, to eliminate amacrine neurons, supplemented 3 hours later with [³H]palmitic acid (2.5 μCi/dish), complexed with BSA, and treated after 30 minutes without (−) or with 48 μM PQ. Lipids were then extracted, subjected to alkaline hydrolysis and separated by thin layer chromatography. *Statistically significant difference compared with control (P < 0.05). Phase (B, D) and fluorescence (C, E) micrographs of cultures treated without (B, C) and with (D, E) 10 μM MAPP and 30 minutes later with 10 μM C₂-Cer (Cer) for 24 hours. Nuclei were labeled with DAPI (C, E). After Cer treatment, MAPP-lacking cultures had many pyknotic nuclei (arrowheads), which were markedly reduced in MAPP-treated cultures. (F) The percentage of apoptotic photoreceptors in control cultures (−) and in cultures treated with Cer and with MAPP or with MAPP+Cer (MAPP Cer) was determined by quantifying the number of pyknotic or fragmented nuclei with the DNA probe DAPI. (G) The percentage of photoreceptors preserving mitochondrial membrane potential was quantified by analyzing the presence of fluorescent mitochondria, labeled with red fluorescent stain. Scale bar, 15 μm. *Statistically significant differences compared with control cells (P < 0.01).

Inhibition of Sph Synthesis on Ceramide- and PQ-Induced Apoptosis

To establish whether ceramide conversion to Sph was essential to induce apoptosis, we inhibited this hydrolysis with MAPP, a well-established inhibitor of alkaline ceramidase, ⁴⁴ and then treated neurons with C₂-Cer. C₂-Cer induced photoreceptor apoptosis (Figs. 1B, 1C, arrowheads) whereas addition of MAPP before C₂-Cer treatment markedly diminished it (Figs. 1D, 1E). Approximately 20% of the photoreceptors were apoptotic in the control cells and MAPP-supplemented cultures and C₂-Cer addition almost doubled this amount (Fig. 1F). MAPP addition before C₂-Cer treatment reduced significantly, although not completely, ceramide-induced apoptosis, with only 25% of photoreceptors being apoptotic (Fig. 1F). Since C₂-Cer-induced photoreceptor apoptosis is associated with the loss of mitochondrial membrane potential, ²⁸ we explored whether blocking Sph synthesis would prevent this loss. As observed in Figure 1G, whereas C₂-Cer decreased the percentage of photoreceptors preserving mitochondrial membrane potential, MAPP prevented the decrease. 

We then investigated whether ceramide breakdown to Sph is necessary for oxidative stress-induced apoptosis. Three-day neuronal cultures were treated with or without MAPP, before addition of PQ. After 24 hours, many TUNEL-positive cells were observed in PQ-treated cultures, and almost 50% of the total photoreceptors were apoptotic (Figs. 2A, 2B, 2E). MAPP addition before PQ treatment markedly diminished the amount of TUNEL-labeled cells (Figs. 2C, 2D), and photoreceptor apoptosis remained almost at control values (∼20%; Fig. 2E). This finding suggests that PQ induction of photoreceptor apoptosis necessitates, at least to some extent, the conversion of ceramide to Sph. 

Figure 2.

 

Effect of inhibiting Sph synthesis on PQ-induced apoptosis of photoreceptors. Phase (A, C) and fluorescence (B, D) micrographs of cultures treated without (A, B) and with (C, D) MAPP and 1 hour later with 48 μM PQ for 24 hours. Apoptosis was then analyzed by TUNEL assay (B, D), and (E) the percentage of apoptotic photoreceptors was determined with DAPI. (F) Percentage of photoreceptors preserving mitochondrial membrane potential. Cultures lacking MAPP showed many TUNEL-positive cells after PQ treatment, which decreased when MAPP was added before PQ. *Statistically significant differences compared with controls (P < 0.01). Scale bar, 15 μm.

PQ-induced photoreceptor apoptosis is associated with mitochondrial depolarization. ^9,45 In control and MAPP-treated cultures approximately 85% of photoreceptors preserved their mitochondrial potential, and PQ decreased this percentage to 55% (Fig. 2F). MAPP addition prevented mitochondrial depolarization; 82% of photoreceptors retained their mitochondrial membrane potential in these cultures, despite PQ treatment (Fig. 2F). 

Effect of Sph Exogenous Addition on Apoptosis

We then evaluated the effect of exogenous addition of Sph on photoreceptor apoptosis. Photoreceptors develop normally for 3 to 4 days in the chemically defined medium used, which lacks photoreceptor trophic factors. ³² Few TUNEL-positive photoreceptors and annexin-labeled cells were observed in day 4 control cultures (Figs. 3A, 3B, 3E, 3F). Treatment with 5 μM Sph during 24 hours markedly increased the number of TUNEL-labeled cells (Figs. 3C, 3D). Many annexin-labeled cells were already present, though very few showed PI-labeled nuclei (Figs. 3G, 3H) only 12 hours after addition of Sph. Sph-induced apoptosis increased in a concentration-dependent manner (Fig. 3O); at concentrations higher than 5 μM almost all cells evidenced necrotic features, losing their integrity and normal morphology (not shown). 

Figure 3.

 

Effect of exogenous Sph addition on neuronal apoptosis. Neuronal cultures were treated at day 3 without (A, B, E, F) or with (C, D, G, H) 5 μM Sph. Phase (A, C, E, G) and fluorescence (B, D, F, H) micrographs show TUNEL (A–D) and annexin/PI (E–H) labeling, 24 and 12 hours after Sph addition, respectively. Sph increased the amount of TUNEL-positive cells (compare B, D). Annexin (green fluorescence and PI; red fluorescence)-labeled control cells were very few (E, F) but increased after addition of Sph (H, arrowheads). Confocal images (I–N) show cytochrome c (green, I, L), mitochondrial stain labeling (red, J, M), and merge (K, N) in photoreceptors in control cells (I–K) and after 12 hours of Sph treatment (L–N). Cytochrome c and mitochondrial labeling colocalized in control cells (I–K, thin arrows), but this colocalization was no longer observed after Sph treatment (L–N, arrowheads). (O) Cultures were treated at day 3 with increasing concentrations of Sph for 24 hours, and the percentage of apoptotic photoreceptors was determined with DAPI. (P) Cultures were treated at day 3 with Sph for 5.5 hours and then incubated with 10 μM H₂DCFDA for 30 minutes at 37°C. Relative fluorescence in control cells (−) and Sph-treated (Sph) cultures was then determined. *Statistically significant differences compared with control cells (P < 0.01). Scale bar: (A–H) 15 μm; (L–N) 10 μm.

We also investigated whether Sph induces cytochrome c release from mitochondria. In control cells, cytochrome c colocalized with fluorescence-labeled mitochondria (MitoTracker, Invitrogen-Molecular Probes; Figs. 3I, 3J, 3K, thin arrows). A 12-hour treatment with Sph led to the loss of this colocalization (Figs. 3L, 3M, 3N, arrowheads), suggesting that Sph promotes cytochrome c translocation from mitochondria to cytosol. Approximately 6 hours after Sph treatment, we evaluated the effect on ROS production, which occurs as an early step in the apoptotic process. Sph markedly increased ROS levels, which were approximately four times higher than in control cultures (Fig. 3P). 

Few photoreceptors had fragmented or pyknotic nuclei in day 4 control cultures (Figs. 4A, 4C, arrowheads) and their number increased approximately 2.5 times after addition of Sph (Fig. 4B, 4D, 4G, arrowheads). This increase was parallel to an increased mitochondrial depolarization; whereas most cells in controls retained their mitochondrial membrane potential, showing brilliant fluorescent mitochondria when labeled with the mitochondrial stain, Sph decreased the amount of photoreceptors preserving mitochondrial potential (Figs. 4E, 4F, 4H). 

Figure 4.

 

Effect of Sph on photoreceptor apoptosis and mitochondrial membrane potential. Phase (A, B) and fluorescence (C–F) micrographs showing nuclei labeled with DAPI (C, D) and labeled mitochondria (E, F) in day 4 neuronal cultures without (A, C, E) or with (B, D, F) 5 μM Sph for 24 hours. In control cells, most photoreceptors were viable and preserved their mitochondrial membrane potential (A, C, E, arrows). Sph induced photoreceptor apoptosis and mitochondrial membrane depolarization (B, D, F, arrowheads). (G) The percentage of apoptotic photoreceptors in (−) control and (Sph) Sph-treated cultures was determined with DAPI. (H) The percentage of photoreceptors preserving mitochondrial membrane potential was quantified. *Statistically significant difference compared with control (P < 0.01). Scale bar, 15 μm.

Sph may induce apoptosis, not by itself but through its conversion to ceramide, catalyzed by ceramide synthase. To exclude this possibility, we added fumonisin B1 before Sph addition, to inhibit this conversion. Apoptotic photoreceptors increased from 23.2% ± 3.1% in controls to 51.7% ± 0.7% after Sph treatment and remained at 49.3% ± 1.6%, regardless of fumonisin B1 addition (not shown). Some isoforms of alkaline ceramidase also exhibit ceramide synthase activity, so we inhibited this activity with MAPP ⁴⁶ ; as observed with fumonisin, MAPP addition had no protective effect on Sph-induced apoptosis, since 49.3% ± 3.5% of photoreceptors were apoptotic in MAPP, Sph-treated cultures (not shown). Hence, exogenously added Sph, not its conversion to ceramide, induces photoreceptor apoptosis. 

Sph also promoted apoptosis in amacrine neurons; the percentage of these neurons with fragmented or pyknotic nuclei increased from 9.7% ± 5.2% to 38% ± 4.8% after Sph addition (n = 3, not shown). This increase was parallel to a decrease in the amount of amacrine neurons preserving mitochondrial membrane potential, though less pronounced than in photoreceptors (not shown). 

DHA Prevention of Photoreceptor Apoptosis

We have shown that DHA promotes photoreceptor survival during early development in vitro and on C₂-Cer- and PQ treatments. ^9,28,32,46 We now investigated its effect on Sph-induced apoptosis. DHA sharply reduced the number of TUNEL-positive cells (compare Figs. 5D, 5E) and the amount of annexin-labeled cells present after Sph treatment (Figs. 5J, 5K, arrowheads). Apoptotic photoreceptors increased from approximately 20% in control cells to more than 50% after Sph addition (Fig. 5Q). DHA prevented this increase, with only approximately 30% of total photoreceptors being apoptotic despite Sph treatment (Fig. 5Q). DHA also prevented Sph-induced mitochondrial membrane depolarization in photoreceptors. Sph reduced the percentage of photoreceptors preserving their mitochondrial potential compared with controls, but DHA maintained this percentage at control levels (Fig. 5R). 

Figure 5.

 

Effect of DHA and of the inhibition of sphingosine kinase on Sph-induced apoptosis of photoreceptors. Phase (A–C, G–I) and fluorescence (D–F, J–L) micrographs of cultures treated at day 3 with (C, F, I, L) or without (A, B, D, E, G, H, J, K) the sphingosine kinase inhibitor DHS, then supplemented (B, C, E, F, H, K, I, L) or not (A, D, G, J) with 6.7 μM DHA and finally treated with Sph (A–L). Apoptosis was determined 24 hours later by TUNEL (D–F) and annexin/PI (J–L) labeling. DHA-supplemented, Sph-treated cultures (B, E) showed very few TUNEL-positive cells, compared with Sph-treated cultures (A, D); pretreatment with DHS increased the amount of TUNEL-labeled cells after Sph addition, despite DHA supplementation (C, F). The large number of PI (red), annexin (green)-labeled apoptotic photoreceptors in Sph-treated cultures (arrowheads in G, J) decreased noticeably after DHA supplementation (arrowhead in H, K); however, DHS increased photoreceptor apoptosis in DHA-supplemented, Sph-treated cultures (arrowheads in I, L). (Q) Percentage of apoptotic photoreceptors determined with DAPI. (R) Percentage of photoreceptors preserving mitochondrial membrane potential. *Statistically significant differences compared with control cells (P < 0.01). Scale bar, 15 μm.

Inhibition of S1P Synthesis Blocked DHA Protection from Sph-Induced Apoptosis

Our previous finding that DHA enhances ceramide glucosylation to decrease ceramide levels, and thus protects photoreceptors from ceramide-induced apoptosis, ²⁸ suggests that DHA may modulate the activity of other enzymes of sphingolipid metabolism to prevent Sph-induced apoptosis. A promising pathway to explore was Sph phosphorylation to synthesize S1P. This reaction is catalyzed by SphK, which has two isoforms, SphK1 and SphK2. Synthesis of S1P by SphK1 has been associated with antiapoptotic properties. ^47,48 We evaluated whether inhibiting this enzyme with DHS would block DHA protection. As mentioned, few TUNEL or annexin-positive photoreceptors were present in DHA-supplemented cultures, regardless of Sph treatment (Figs. 5E, 5K, respectively). Addition of DHS blocked DHA protection, markedly increasing the amount of TUNEL- and annexin-labeled cells (Figs. 5F, 5L, respectively). Only 30% of photoreceptors were apoptotic in DHA-supplemented, Sph-treated cultures. DHS almost doubled this percentage, increasing photoreceptor apoptosis to the values found in Sph-treated cultures lacking DHA (Fig. 5Q). DHS also blocked DHA preservation of mitochondrial membrane potential (Fig. 5R); the percentage of photoreceptors maintaining their mitochondrial potential decreased from almost 80% in DHA-supplemented, Sph-treated cultures, to 50% after DHS addition. Hence, S1P synthesis was required for DHA protection. 

We then evaluated whether DHA enhances SphK1 expression to increase S1P levels. After eliminating amacrine neurons, the cultures were supplemented with DHA and then with the proteasome inhibitor MG132, to increase SphK1 lifetime. Control cultures showed a very low expression of SphK1 (Figs. 5M, 5O). DHA not only increased this expression but also promoted SphK1 translocation to the plasma membrane (Figs. 5N, 5P, arrows). 

Discussion

Our previous finding that ceramide acts as a signal in the induction of photoreceptor death ²⁸ suggested that other sphingolipids may participate in regulating their death and/or survival. We show here for the first time that Sph plays a relevant role in activating photoreceptor apoptosis on oxidative stress. PQ rapidly increased Sph endogenous levels before inducing photoreceptor apoptosis. Conversely, inhibiting Sph synthesis by blocking ceramide breakdown noticeably reduced oxidative stress-induced apoptosis of photoreceptors. These results suggest that in addition to enhancing the synthesis of ceramide, oxidative stress promoted its subsequent breakdown to Sph, and either the latter or both ceramide and Sph activated photoreceptor death. Exogenous addition of Sph also triggered photoreceptor apoptosis; DHA prevented it, but inhibiting the synthesis of S1P blocked this protection. These results imply that keeping Sph levels low, either by preventing its synthesis or by promoting its conversion to S1P contributes to protect photoreceptors from oxidative stress. 

Intensive research has established the roles of several sphingolipids as signals, at times antagonistic, in the regulation of cell death. Ceramide was the first shown to be an endogenous mediator of apoptosis on a myriad of cell stressors. ^12,49,50 Its role as a key mediator in the apoptosis of photoreceptors has recently been uncovered; oxidative stress rapidly increases de novo synthesis of ceramide in these cells and lowering ceramide levels either by blocking its synthesis or by enhancing its glucosylation prevents their apoptosis. ²⁸ Sph has also been shown as a mediator in the induction of apoptosis. Since it is not confined to membrane fractions as ceramide, Sph makes an ideal second messenger ⁵¹ ; rapidly produced and accumulated during the early stages of apoptosis, ^52,53 both its accumulation and exogenous addition induce apoptosis in many cell types. ^22,54 Blockade of its synthesis attenuates apoptosis on different death stimuli. ^55,56 The role of Sph in photoreceptor death was unknown. To be a mediator of oxidative stress-induced apoptosis, induction of oxidative stress should increase Sph levels; our data demonstrate that PQ treatment rapidly increased the amount of [³H]Sph in photoreceptors, with levels higher than in controls already detectable less than an hour after PQ addition. The rapid increase in Sph levels in other cell systems after apoptosis induction, which precedes cell morphologic changes ^55,56 led to the proposal that Sph regulates early events in the apoptotic program, acting either independently or together with ceramide. ²² No evidence of apoptosis was found in photoreceptors at the time of Sph increase, the onset of morphologic changes that accompany their apoptosis occurring at least 2 hours after addition of PQ. ²⁸ This result suggests that Sph is an early signal in the activation of photoreceptor death after oxidative stress. 

Sph is exclusively synthesized from ceramide through the action of ceramidases, which are classified according to their pH optimum in alkaline, acidic, and neutral. ^57,58 Mounting evidence points to their relevant role in controlling cell death and survival. ⁵⁹ An increase in their activity has been described in phorbol ester-induced apoptosis of HL-60 leukemia cells. ²³ Transgenic overexpression of neutral ceramidase, even in cells of tissues distant from photoreceptors, prevents their degeneration in a Drosophila phototransduction mutant. ^60,61 Our results indicate a key role for alkaline ceramidase in the regulation of photoreceptor death; its inhibition significantly decreased ceramide-induced apoptosis of photoreceptors, though did not completely prevent it. This finding implies that, although ceramide induced photoreceptor apoptosis, it had to be at least partially hydrolyzed to Sph to achieve the full apoptotic effect. Inhibiting alkaline ceramidase also protected photoreceptors from oxidative stress-induced apoptosis, supporting the proposal that Sph, either by itself or, more probably, together with ceramide, is an essential mediator in apoptosis induction by this damage. Ceramide intracellular concentrations are frequently more than an order of magnitude higher than those of Sph ¹³ ; hence, the breakdown of only a fraction of the ceramide generated by oxidative stress would significantly increase Sph levels, turning on the apoptotic pathways it signals. PQ-induced apoptosis involves mitochondrial membrane permeabilization and cytochrome c translocation. ^9,44 Inhibition of Sph synthesis in PQ-treated cultures preserved mitochondrial membrane potential, suggesting that Sph may act upstream of mitochondria, inducing its depolarization and subsequently promoting photoreceptor death. 

Our results also demonstrate that exogenous addition of Sph, at concentrations reported to cause cell death in many cell types, induced apoptosis of photoreceptors and amacrine neurons in culture. Sph is rapidly taken up by cultured cells, diffuses through different cellular compartments and can be efficiently metabolized to ceramide. ⁶² Whether ceramide and Sph act together or separately to activate apoptosis seems to depend on the cell type or the apoptotic insult. An increase in ceramide precedes that of Sph with different apoptotic stimuli, ^44,52,63 suggesting the latter may result from ceramide hydrolysis, and both precede the onset of apoptosis. ²⁴ However, Sph has also been reported to act independently from and earlier than ceramide. ⁵¹ Sph, and not its conversion to ceramide, is responsible of the apoptotic signal in several tissues. ⁴⁴ Exogenously added Sph was able by itself to induce photoreceptor apoptosis, as evidenced by the absence of a protective effect when its conversion to ceramide was inhibited. Sph addition alters the integrity of several intracellular membrane systems, increasing lysosomal and mitochondrial permeability and leading to Golgi fragmentation. ⁶⁴ In photoreceptors, Sph addition induced mitochondrial membrane depolarization and cytochrome c release, suggesting the mitochondrial pathway was involved in the induction of apoptosis, as occurs in several cell types. ^24,44 The increased ROS production induced by Sph further supports mitochondrial pathway involvement in Sph-induced apoptosis. Mitochondria are among the major cellular sources of ROS and both Sph and ceramide have been shown to increase ROS production. ⁶⁵ Exogenous addition and endogenous production of Sph may not necessarily activate the same intracellular pathways, as different subcellular compartments may be involved. However, oxidative stress-induced apoptosis also involved mitochondrial depolarization whereas inhibition of Sph synthesis prevented it, suggesting at least several steps of the apoptotic pathway may be shared. 

Our present data evidenced DHA prevented Sph-induced apoptosis of photoreceptors. We explored whether DHA lowers Sph levels by promoting its phosphorylation to increase intracellular S1P, which has an antiapoptotic effect in many cell systems. ^66,67 Inhibiting SphK1 completely blocked DHA's protective effect. The increased apoptosis was accompanied by mitochondrial membrane depolarization, suggesting that the accumulation of either Sph, ceramide, or both leads to mitochondrial permeabilization and then triggers apoptosis. These results underscore the relevance of lowering Sph levels, either by preventing its synthesis or by enhancing its conversion to S1P, to promote photoreceptor survival. They also support the involvement of S1P and SphK1 activity in DHA protection. We recently showed that exogenous addition of S1P prevents photoreceptor apoptosis during development in vitro and established that DHA upregulates the expression of SphK1 to promote their differentiation. ⁶⁸ We have now demonstrated that DHA increased the expression of SphK1 in photoreceptors at the earlier culture times used in these studies and promoted its translocation to the plasma membrane, known to be involved in SphK1 activation. ⁶⁷ The enhanced expression and activity of this enzyme may increase S1P levels and thus prevent photoreceptor apoptosis in response to oxidative stress. The dynamic balance between intracellular levels of S1P and those of ceramide and Sph have been proposed to constitute the so-called sphingolipid rheostat. Since these sphingolipids regulate opposing signaling pathways, their relative intracellular levels are crucial in determining cell fate. ⁶⁹  

In conclusion, our present and previous results emphasize the key role of sphingolipid molecules and the relative balance between them in regulating the final fate of photoreceptors. They also point to a crucial role for Sph, either by itself or combined with ceramide, as an intracellular signal involved in determining photoreceptor death after oxidative damage. 

Footnotes

 Supported by CONICET fellowships (CEA, GEM, DLA); and grants from the Secretaria de Ciencia y Tecnología, Universidad Nacional del Sur; FONCyT; and CONICET, Argentina. NPR and LEP are CONICET Independent Researchers.

Footnotes

 Disclosure: C.E. Abrahan, None; G.E. Miranda, None; D.L. Agnolazza, None; L.E. Politi, None; N.P. Rotstein, None

The authors thank Silvia Antollini for help with fluorescence measurements and biochemist Beatriz de los Santos for expert technical assistance.