Nuclear Translocation of Glyceraldehyde-3-Phosphate Dehydrogenase: A Role in High Glucose-Induced Apoptosis in Retinal Müller Cells

Abstract

purpose. A recent study demonstrated that retinal Müller cells undergo hyperglycemia-induced apoptosis in vitro. Translocation of glyceraldehyde-3-phosphate dehydrogenase (GAPDH) from the cytosol to the nucleus is a critical step in the induction of apoptosis in neuronal cells. R-(−)-deprenyl prevents nuclear translocation of GAPDH and subsequent apoptosis in neuronal cells. In this study, the role of nuclear translocation of GAPDH in hyperglycemia-induced apoptosis in retinal Müller cells and the ability of R-(−)-deprenyl to inhibit the translocation of GAPDH and apoptosis were investigated.

methods. Transformed rat Müller cells (rMC-1) and isolated human Müller cells were cultured in normal glucose, high glucose, and high glucose plus R-(−)-deprenyl for up to 5 days. Subcellular distribution of GAPDH was determined in vitro and in vivo by immunocytochemistry. Apoptosis in tissue cultures was determined by annexin-V staining and caspase-3 activity.

results. Hyperglycemia significantly increased the amount of GAPDH protein in the nucleus above normal within the first 48 hours in rMC-1 and human Müller cells. The addition of R-(−)-deprenyl to these cells incubated in high glucose reduced the amount of GAPDH protein in the nucleus and decreased hyperglycemia-induced apoptosis in both cell types. In vivo studies confirmed the accumulation of GAPDH in nuclei of Müller cells in diabetes.

conclusions. The nuclear translocation of GAPDH in rMC-1 and human Müller cells is closely associated with the induction of apoptosis. R-(−)-deprenyl inhibits nuclear accumulation of GAPDH and subsequent apoptosis in these cells. Therefore, R-(−)-deprenyl offers a strategy to explore the role of GAPDH translocation into the nucleus in the development of diabetic retinopathy.

Müller cells have numerous activities in the retina and are responsible for the modulation of neural activity by regulation of calcium, potassium, and neurotransmitters, and control of the local environment by release of lactate and regulation of pH. ^{1 2} Energy production in Müller cells relies entirely on glycolysis, with low dependency on oxygen. ² Diabetes induces abnormalities in retinal Müller cells, including increased expression of glial fibrillary acidic protein, reduction of glutamine synthetase, and decreased function of glutamate transporter. ^{3 4 5 6} Nevertheless, the contribution of Müller cells to diabetic retinopathy remains unclear. We have shown that Müller cells undergo hyperglycemia-induced apoptosis in vitro, although the mechanism of initiation of the apoptotic cascade is unknown. ⁷  

Glyceraldehyde-3-phosphate dehydrogenase (GAPDH) is a key enzyme in glycolysis, catalyzing the conversion of d-glyceraldehyde-3-phosphate to 1,3-bisphosphoglycerate in a reaction that is reversible. ⁸ The structure of GAPDH, containing a highly reactive cysteine in the active site, allows the enzyme to be redox–sensitive to oxidation and nitrosylation resulting in loss of activity. ⁹ Several other nonglycolytic functions of the enzyme have been identified recently, such as its role in endocytosis, microtubule bundling, phosphotransferase, nuclear RNA export, DNA replication, DNA repair, viral pathogenesis, oncogenesis, and apoptosis (for review see Sirover ¹⁰ ). Alterations of oligomeric structure and subcellular localization may be important in mediating some of these processes. 

The role of GAPDH in apoptosis has been described in neurodegenerative diseases, such as Parkinson’s disease, Huntington’s disease, and Alzheimer’s disease, ^{11 12 13} and appears to rely on the relocalization of the protein within subcellular compartments. In nonapoptotic cells, GAPDH is primarily found in the extranuclear cytoplasm, with only sparse localization to small punctuate areas in the nucleus, whereas in apoptotic cells, the enzyme accumulates in the nucleus. ^{14 15 16 17} The mechanism that triggers GAPDH translocation and its function in the nucleus is not known. 

R-(−)-deprenyl has been found to inhibit nuclear translocation of GAPDH and apoptosis. ¹⁸ The drug was developed as a monoamine oxidase B (MAO-B) inhibitor, and has been used in the treatment of Parkinson’s disease. ¹⁹ Nuclear translocation of GAPDH was observed in neuronal cells undergoing apoptosis in Parkinson’s disease. ¹¹ The effect of R-(−)-deprenyl to increase neuronal cell survival has been found to be independent of MAO-B inhibition. ²⁰ It has been suggested that R-(−)-deprenyl binds to GAPDH, causing the dissociation of the tetramer to a dimer, which inhibits nuclear localization of GAPDH. ^{18 20}  

In this study demonstrates, hyperglycemia stimulated accumulation of GAPDH in the nucleus in retinal Müller cells. The nuclear increase of GAPDH protein was closely associated with apoptosis induction in these cells. Furthermore, R-(−)-deprenyl inhibited both the hyperglycemia-induced nuclear translocation of GAPDH and apoptosis in these cells. 

Materials and Methods

Materials

R-(−)-deprenyl, purified GAPDH from rabbit muscle, 7-amino-4-trifluoro-methylcoumarin (AFC), and a glucose reagent (Infinity) were purchased from Sigma-Aldrich (St. Louis, MO). Caspase-3 substrate was purchased from Calbiochem (San Diego, CA). Annexin-V-phycoerythrin reagents were purchased from R&D Systems (Minneapolis, MN). The following antibodies were used in the study: mouse monoclonal antibody to GAPDH (Chemicon International, Inc., Temecula, CA), rabbit polyclonal antibody to GAPDH (Novus Biologicals, Littleton, CO), mouse monoclonal antibody to glutamate synthetase (GS; BD Transduction Laboratories, Palo Alto, CA), mouse monoclonal antibody to vimentin (Oncogene Research Products, Boston, MA), goat anti-mouse IgG conjugated to Texas red, and goat anti-rabbit conjugated to Alexa Fluor 488 were purchased from Molecular Probes, Inc. (Eugene, OR), and horseradish peroxidase (HRP)–conjugated anti-mouse IgG from Santa Cruz Biotechnology, Inc. (Santa Cruz, CA). 

Culture of Rat Retinal Müller Cells

Rat retinal Müller cells (rMC-1) ²¹ were plated on tissue culture plastic (Falcon Labware; BD Biosciences; Franklin Lakes, NJ) and grown in Dulbecco’s modified Eagle’s medium (DMEM) containing 5 mM glucose, 10% FBS, and 1% penicillin/streptomycin. For treatment conditions, cells were plated at an 80% cell confluence and switched to DMEM containing 5 mM glucose (normal), 25 mM glucose (high glucose), or 25 mM glucose plus 1 nM R-(−)-deprenyl (drug treatment), all supplemented with 2% FBS and 1% penicillin-streptomycin at 37°C in a humidified chamber with a 5% CO₂-95% air mixture. Medium was changed every day. Because of the rate of growth in the cell line, experiments were terminated at 96 hours to ensure that the effects were not due to overgrowth or starvation that might occur at later time points. 

Culture of Freshly Isolated Human Retinal Müller Cells

Donor eyes from a 25-year-old male were obtained through the National Disease Research Interchange (Philadelphia, PA). The donor was a healthy male with no history of diabetes. The eyes were enucleated, packed on wet ice, and received in the laboratory within 12 hours after death. Müller cells were isolated as described with minimal modification. ²² Briefly, retinas were removed and incubated in calcium- and magnesium-free phosphate buffer supplemented with 0.1% trypsin (Difco-BD Biosciences) for 1 hour at 37°C. The retinas were mechanically disrupted and washed in DMEM/Ham’s F12 medium (1:1 ratio, 17.3 mM glucose) supplemented with 20% fetal bovine serum and 1% penicillin-streptomycin. Cells were plated on tissue culture plastic (Falcon Labware; BD Biosciences) in the medium and allowed to attach and grow at 37°C in a humidified chamber with a 5% CO₂-95% air mixture. After attachment, cells were switched to DMEM (5 mM glucose)/Ham’s F12 (1:1 ratio, 7.8 mM glucose) supplemented as stated earlier. After the second passage of cells, cultures were 95% pure for Müller cells as assessed by immunohistochemistry using antibodies to GS, glial fibrillary acidic protein (GFAP), and vimentin, followed by appropriate secondary antibodies for detection. The cultured human Müller cells were not immunoreactive for GS or GFAP (data not shown) but stained positively for vimentin (see 1 1 1 ), which has been observed and described by Winkler et al. ² Experiments were performed on passages 2 to 6. For experiments, cells were switched to DMEM (5 mM glucose)/Ham’s F12 medium (1:1 ratio, 7.8 mM glucose, normal), DMEM/Ham’s F12 medium (1:1 ratio, supplemented to 25 mM glucose, high glucose), or the high-glucose medium (25 mM glucose) with 1 nM R-(−)-deprenyl. All experimental media were supplemented with 2% FBS and 1% penicillin-streptomycin. Cells were grown for up to 5 days at 37°C in a humidified chamber with a 5% CO₂-95% air mixture. Medium was changed every day. 

Confocal Laser Scanning Microscopy of rMC-1 Cells and Isolated Human Müller Cells

Conditioned rMC-1 cells or isolated human Müller cells were grown up to 96 hours in 5 mM glucose, 25 mM glucose, or 25 mM glucose plus 1 nM R-(−)-deprenyl. Cells were fixed with a solution of freshly prepared 4% paraformaldehyde (Electron Microscopy Sciences, Fort Washington, PA) for 10 minutes at room temperature and washed with PBS. Samples were permeabilized with 0.2% (wt/vol) Triton X-100 for 10 minutes and then incubated with 3% BSA overnight. Samples were incubated with primary antibody (rabbit anti-GAPDH for human Müller cells, or mouse anti-GAPDH for rMC-1 cells) for 1 hour at room temperature, washed with PBS, and incubated with 5% normal goat serum for 15 minutes. Samples were incubated with the appropriate secondary antibody for 1 hour at room temperature, followed by extensive rinsing with PBS. Human Müller cells were further stained with an antibody to vimentin to identify Müller cells and labeled with a secondary anti-mouse IgG conjugated to Texas red. Coverslips containing cells were mounted on glass slides that contained a drop of antifade mounting medium (Vectashield; Vector Laboratories, Inc., Burlingame, CA). Cells were viewed by scanning laser confocal microscope (LSM 410; Carl Zeiss Meditec, Göttingen, Germany) using the 568-nm wavelength lines of an argon-krypton laser and an oil objective (100× Plan-Neofluor; Carl Ziess Meditec). 

GAPDH-stained rMC-1 cells were analyzed by fluorescence microscope (Axiphot; Carl Zeiss Meditec, Thornwood, NY) at 40× with excitation of 540 nm and emission of 600 nm and saved digitally by the attached system (SPOT; Diagnostic Instruments, Inc., Sterling Heights, MI). Three random fields of 100 cells were used to assess the number of cells expressing accumulation of GAPDH in the nucleus, as shown in 1 . The percentages were averaged to four independent experiments for each cell condition. 

Cell Fractionation

rMC-1 cells were treated as described earlier. ²¹ At times indicated, cells were washed with Hanks’ buffered saline and scraped in ice-cold homogenization buffer (HB) consisting of 50 mM HEPES (pH 7.5), 0.3 M sucrose, 1 mM EDTA, and protease inhibitors (0.2 mM phenylmethylsulfonyl fluoride [PMSF] and 1 μM leupeptin). Cells were collected by centrifugation, resuspended in HB, and homogenized with 22 strokes of a homogenizer (Dounce; Bellco Glass Co., Vineland, NJ). Nuclear fractions were collected by low-speed spin (3000g), washed in HB once, and resuspended in lysis buffer (LB) containing 50 mM HEPES (pH 7.5), 1% Triton X-100, 150 mM NaCl, 1 mM EDTA, and protease inhibitors. The concentration of protein was determined with a protein dye (Bio-Rad Laboratories, Hercules, CA). 

Western Blot and Immunoblot Analyses

Immunoblot analysis was performed with cytosolic and nuclear fractions, prepared as described earlier. Samples were resolved by polyacrylamide gel and transferred to nitrocellulose by using standard techniques. ^{23 24} Nitrocellulose blots that had been blocked with 5% dry milk dissolved in TBS-T (10 mM Tris [pH 7.4], 150 mM NaCl, 1% Tween 20) were incubated overnight at 4°C with mouse monoclonal antibody to GAPDH diluted in blocking solution. After extensive washing with TBS-T, blots were incubated with anti-mouse IgG HRP-conjugated antibody diluted in blocking solution for 1 hour at room temperature, followed by detection with Western blot reagent (Luminol; Santa Cruz Biotechnology). Densitometric analysis was performed with a scanner (ScanJet 6300C; Hewlett-Packard, Inc., Palo Alto, CA, and Scion Image 1.62C software; Scion Corp, Frederick, MD). Cytosolic and nuclear GAPDH protein concentrations in cells exposed to 5 mM glucose, 25 mM glucose, and 25 mM glucose plus R-(−)-deprenyl were calculated from a standard curve by using defined concentrations of purified GAPDH protein. 

Confocal Laser Scanning Microscopy of Retinal Sections

Male Lewis rats (225–250 g) were randomly assigned to the control group or the diabetic group. Diabetes was induced by single injection of streptozotocin (55 mg/kg body weight, intraperitoneally). Insulin was given as needed to achieve slow weight gain without preventing hyperglycemia and glycosuria (0–2 U of neutral protamine Hagedorn [NPH] insulin subcutaneously, two to three times a week). Treatment of animals conformed to the ARVO Statement for the Use of Animals in Ophthalmic and Vision Research. Before the animals were killed, blood glucose levels were measured, and the percentage of glycated hemoglobin (GHb) was determined by chromatography (Glyc-Affin; Pierce, Rockford, IL). Experimental animals were killed after 12 weeks of diabetes along with age-matched control subjects. Eyes were fixed in methacarn followed by embedding in paraffin. Sections (5 μm thick) of diabetic eyes and control eyes were rehydrated and stained with polyclonal rabbit antibody to GAPDH and mouse monoclonal antibody to GS, as described earlier. 

GAPDH Activity Assay

GAPDH activity was measured as previously described. ⁹ Purified GAPDH, the cytosolic and nuclear fractions (25 μg), was incubated in 50 mM triethlyammonium buffer (pH 7.5) containing 50 mM arsenate, 2.4 mM glutathione, and 10 μg/mL glyceraldehyde-3-phosphate, in a total volume of 100 μL at 37°C. The enzymatic reduction of NAD⁺ to NADH was started by addition of 250 μM NAD⁺. GAPDH activity was monitored by recording the fluorescence emission at 435 nm after excitation at 360 nm. 

Glucose Consumption Measurements

Briefly, medium was collected from treated cells every 24 hours and samples stored at −20°C until assayed. Glucose reagent (Infinity; Sigma-Aldrich) was used to assess glucose concentration in tissue culture medium according to the manufacturer’s instructions. Glucose concentration was monitored by recording the conversion of glucose to glucose-6-phosphate in a NADH-coupled reaction, as measured at 340 nm. A standard plot of glucose concentration and absorbance was produced to determine sample concentration. Measurements are given as the amount of glucose absent from the original medium concentration as consumed in picomoles glucose per cell × 10⁴ per 24 hours. 

Annexin-V Staining

Annexin-V staining was performed as previously described. ⁷ rMC-1 cells (1 × 10⁵/24 mM) or human Müller cells (1 × 10⁴/24 mM) were grown on coverslips and incubated in treatment medium, as described earlier. Briefly, cells grown for 96 hours were washed in annexin-V binding buffer and incubated in 500 μL annexin-V staining solution (1:50 dilution of annexin-V-phycoerythrin in annexin-V binding buffer) for 10 minutes in the dark at 37°C in 5% CO₂. The annexin-V staining solution was removed, and coverslips were immediately mounted on slides with antifade medium (Vectashield; Vector Laboratories) and analyzed by fluorescence microscopy (Axiphot; Carl Zeiss Meditec) at excitation of 540 nm and emission of 600 nm. Human Müller cells were mounted with mounting medium containing 4,6-diamidino-2-phenylindole (DAPI; Vector Laboratories), a nuclear stain. 

Measurement of Caspase-3 Activity

Caspase activity was measured as described previously. ²⁵ Briefly, equal amounts of protein (15 μg) isolated from cells exposed to treatment conditions as described earlier were incubated in lysate buffer (100 mM HEPES [pH 7.5], 10% sucrose, 0.1% 3-([3-cholamidopropyl]dimethylammonio-2-hydroxy-1-propanesulfonate [CHAPS], and 1 mM dithiothreitol [DTT]) containing the fluorogenic caspase substrate, DEVD-AFC (2.5 μM), in a total volume of 100 μL at 32°C for 1 hour. Cleavage of the substrate emitted a fluorescence signal that was quantified by a fluorescence plate reader (excitation: 400 nm, emission: 505 nm; Spectra FluorPlus; Tecan USA, Durham, NC). Activity units were calculated from a standard curve based on defined concentrations of AFC. 

Trypan Blue Assay

Trypan blue was used to assess the percentage of cell death caused by late apoptosis and necrosis. Cells exposed to treatment for 96 hours were collected by a brief trypsin wash. Equal volume of trypan blue dye (Sigma-Aldrich) was added to collected cells. Cells were counted by hemocytometer and assessed for blue inclusion, which is suggestive of a compromised membrane and cell death. Cell death was determined by the percentage of blue cells in total cells. 

Statistical Analysis

Data processing was performed on computer (Excel; Microsoft, Redmond, WA). All quantitative data are expressed as the mean ± SD. Comparison between two groups was analyzed with the Mann-Whitney test (two-tailed). 

Results

Effects of High Glucose on GAPDH Cellular Localization in rMC-1 Cells and of R-(−)-deprenyl on Nuclear Translocation of GAPDH

Previous studies in neuronal cells have suggested that nuclear translocation of GAPDH is essential in apoptosis. ^{15 16 26} We examined whether GAPDH accumulates in the nucleus in rMC-1 cells under high-glucose conditions, and whether R-(−)-deprenyl inhibits this process. GAPDH immunostaining (in red) appears to be primarily cytosolic as seen in control cells grown in 5 mM glucose 1 . At 48 hours in high glucose, the number of rMC-1 cells that stained positive for GAPDH in the nucleus increased by 41% compared with cells incubated in 5 mM glucose 1 1 . Forty-eight hours was the earliest time at which GAPDH nuclear accumulation could be detected in 25 mM glucose (data not shown). R-(−)-deprenyl (1 nM) decreased the high glucose-induced nuclear localization of GAPDH by 63%, compared with cells incubated in 25 mM glucose without drug 1 1 . The concentration of drug was chosen because it exerted the greatest effect on cell survival over 96 hours of exposure (data not shown). 

To confirm further the high glucose-induced increase in the amount of GAPDH protein in the nucleus, nuclear fractions were purified from cells after exposure to treatment conditions for 48 hours, and Western blot analysis was performed 1 . Cells exposed to 5 mM glucose showed a small amount of GAPDH protein in the nuclear fraction (1 , lane 1), whereas cells exposed to 25 mM glucose had an increase of 47% in the amount of GAPDH protein in the nuclear fraction (1 , lane 2). Addition of 1 nM R-(−)-deprenyl resulted in a 53% decrease in GAPDH protein in the nuclear fraction (1 , lane 3) compared with the nuclear fractions from cells exposed to 25 mM glucose and a 30% decrease compared with those exposed to 5 mM glucose. 

In addition, we wanted to determine whether high glucose causes GAPDH nuclear accumulation in primary human Müller cells and whether the translocation can be inhibited by R-(−)-deprenyl. Isolated human Müller cells were cultured in normal glucose, high glucose, or high glucose plus 1 nM R-(−)-deprenyl. We identified Müller cells by using an antibody specific for vimentin (red stain), a marker for Müller cells 1 1 1 . GAPDH localization within the cells was determined by immunostaining with GAPDH specific antibody (green stain, 1 1 1 ). Primary human Müller cells exposed to normal glucose levels demonstrated GAPDH immunostaining primarily in the cytosol, with a relatively small amount entering the nucleus 1 . When glucose concentration in the culture medium was increased to 25 mM, the nuclear localization of GAPDH increased 1 . Addition of R-(−)-deprenyl resulted in the inhibition of the high-glucose–induced nuclear accumulation of GAPDH 1 . 1 1 1 show the overlay of the vimentin and GAPDH staining of human Müller cells in normal glucose, high glucose, or high glucose plus 1 nM R-(−)-deprenyl. 

Diabetes-Induced Increase in GAPDH Protein in Nuclei of Müller Cells in Retina of Diabetic Rats

Retinal sections from three diabetic rats and three age-matched normal rats were analyzed by immunohistochemistry to verify the occurrence of nuclear accumulation of GAPDH due to a high-glucose environment in vivo. Diabetic rats averaged higher blood glucose and glycated hemoglobin levels (372.3 ± 34.3 mg/dL and 11.1% ± 1.0%, respectively) compared with age-matched controls (111.3 ± 8.3 mg/dL and 3.8% ± 0.01%, respectively). Müller cells from both diabetic and control retinal sections were immunostained with an antibody to GS (red stain), a Müller-cell–specific marker ²⁷ 2 . GS staining was detected around and in punctate form within the nuclei of Müller cells in the inner nuclear layer (INL) of the retina as well as in the processes that extended out to the outer nuclear layer (ONL) and the inner plexiform layer (IPL). In the normal retina 2 , the nuclei of Müller cells are arranged in a row in the INL, located within the central portion of this layer. The staining for GAPDH using a GAPDH-specific antibody (green stain) showed a low level of expression. Within the INL, nuclei of Müller cells remained absent of GAPDH. However, other nuclei, such as those of the amacrine, bipolar, and horizontal cells in the INL, contained significant amounts of GAPDH protein compared with nuclei of Müller cells. In the retinal sections from diabetic rats 2 , the nuclei of Müller cells (as identified by GS staining, in red) appeared to be less uniform in shape and more disordered in the localization within the INL 2 . The staining for GAPDH (green) demonstrated an increase in protein expression in the INL, the ONL, and the cytosol of the photoreceptors. Increased magnification of a single Müller cell nucleus from normal retina showed the localization of GS to the nucleus (2 , red) and the lack of GAPDH staining within the nucleus (2 , green). 2 2 2 showed the localization of GS (2 , red) and positive GAPDH staining (2 , green) in a single nucleus of a representative Müller cell in the diabetic retinal section. The nucleus of the Müller cell from a diabetic retina shows an accumulation of GAPDH protein. We analyzed the occurrence of GAPDH accumulation in the nuclei of Müller cells of retinal sections from both normal and diabetic rats. Müller cells that stained positive for nuclear GAPDH, as demonstrated in 2 , were counted. The control retinal sections demonstrated 6.5% ± 0.8% of Müller cells positive for increased nuclear GAPDH of the 1652 nuclei of Müller cells counted, whereas diabetic retinal sections demonstrated an increase in occurrence of GAPDH accumulation by 32% ± 6% of the 1545 nuclei of Müller cells counted. Results of counts of Müller cell nuclei showing an increased amount of GAPDH protein are presented in 2 . 

Expression of Cytosolic GAPDH Protein Remains Stable in Hyperglycemic Conditions at 48 Hours

Increase in GAPDH protein levels has been found to be associated with an increased probability of cell death. ¹² We have shown that GAPDH accumulates in the nucleus under high glucose 1 2 . Because GAPDH is necessary for the glycolytic capacity of the cell, the levels of GAPDH available in the cytosol under these conditions were determined. GAPDH protein concentration in the cytoplasm did not change in rMC-1 cells exposed to 5 mM glucose over the 96 hours of study 3 . Likewise, the amount of cytosolic GAPDH protein in cells exposed to 25 mM (8.95 ± 1.51 ng GAPDH/μg cytosolic protein) was similar to that in cells exposed to 5 mM glucose (8.75 ± 1.26 ng GAPDH/μg cytosolic protein) at 48 hours. In every experiment, the amount of GAPDH protein in the cytoplasm declined in cells exposed to 25 mM glucose over the next 48 hours, and by 96 hours the amount of GAPDH protein (7.37 ± 1.34 ng GAPDH/μg cytosolic protein) was significantly lower than the concentration in cells exposed to 5 mM glucose (8.84 ± 1.45 ng GAPDH/μg cytosolic protein) for the same duration. With the addition of R-(−)-deprenyl, the amount of GAPDH protein (14.94 ± 1.40 ng GAPDH/μg cytosolic protein) increased significantly by 48 hours of exposure and declined over time (9.18 ± 0.95 ng GAPDH/μg cytosolic protein). 

Effect of Elevated Glucose and R-(−)-deprenyl on Specific GAPDH Activity in Cytoplasm and Nucleus

The function of GAPDH in the glycolytic pathway is directly tied to the energy production of the cell. The specific activity of cytosolic GAPDH in rMC-1 cells exposed to 5 mM glucose remained consistent during the time course of the 96 hours (75.8 ± 11.7 and 83.8 ± 24.75 nmol NADH/min per milligram GAPDH at 48 hours and 96 hours, respectively; 4 ), whereas the specific activity of cytosolic GAPDH from cells exposed to 25 mM glucose increased over the initial 48 hours to 96.8 ± 15.5 nmol NADH/min per milligram GAPDH and then declined over the next 48 hours to 84.8 ± 14.9 nmol NADH/min per milligram GAPDH. Specific activity of cytosolic GAPDH in cells exposed to 25 mM plus R-(−)-deprenyl (61.8 ± 10.3 nmol NADH/min per milligram GAPDH) decreased by 36% compared with the activity in cells incubated in 25 mM glucose, and showed an 18% decrease compared with the activity in cells incubated in 5 mM glucose at 48 hours. At 96 hours, cytosol-specific activity of GAPDH in the cells incubated in 25 mM glucose plus R-(−)-deprenyl (86.2 ± 25.8 nmol NADH/min per milligram GAPDH) was not significantly different from the level in cells incubated in 5 mM glucose. R-(−)-deprenyl did not alter glucose consumption (7.4 ± 0.9 pmol glucose/cell × 10⁴ per 24 hours compared with 7.3 ± 1.3 pmol glucose/cell per 24 hours in cells exposed to 25 mM alone) and remained steady over 96 hours. However, cells exposed to 25 mM glucose had an overall increase in glucose consumption compared with cells exposed to 5 mM glucose (2.3 ± 0.5 pmol glucose/cell × 10⁴ per 24 hours). 

GAPDH activity was assayed in the nuclear fraction to determine whether the glycolytic function of GAPDH persists in the translocated protein. Western blot analysis showed that the enzyme was present in the nuclei under all treatments, but GAPDH translocated to the nucleus demonstrated no glycolytic activity from cells in any of the treatment conditions or time points tested 4 . 

Effect of R-(−)-deprenyl on Hyperglycemia-Induced Apoptosis in Retinal Müller Cells

Cells exposed to 25 mM glucose for 96 hours showed an increase in the number of cells that stained positively for annexin-V, an early marker of apoptosis, compared with cells exposed to 5 mM glucose 5 5 . At 48 hours, apoptosis was not detected in cells exposed to 25 mM glucose (data not shown). Cells incubated in 1 nM R-(−)-deprenyl demonstrated a lower number of cells stained for annexin-V 5 . To confirm that high glucose can induce apoptosis in isolated human Müller cells and that R-(−)-deprenyl can inhibit this action, isolated human Müller cells were treated in low glucose, high glucose, and high glucose plus R-(−)-deprenyl medium for 72 hours. Analysis of annexin-V staining shows an increase in apoptosis in human Müller cells exposed to high glucose compared with cells exposed to normal glucose 5 5 , which was prevented by the addition of R-(−)-deprenyl 5 . 

Activity of the executioner caspase-3, a downstream marker of apoptosis, was measured on lysates from rMC-1 and isolated human Müller cells. rMC-1 cells exposed to 25 mM glucose for 96 hours showed a sixfold increase in caspase-3 activity compared with cells exposed to 5 mM glucose for the same time point 5 . Moreover, the use of 1 nM R-(−)-deprenyl resulted in a 37% inhibition of caspase-3 activity compared with samples in 25 mM glucose. Human Müller cells also demonstrated an increase in caspase-3 activity due to exposure to high glucose 5 . Caspase-3 activity increased 46% in high glucose compared with normal glucose. The addition of R-(−)-deprenyl to the high-glucose exposure decreased the activity of caspase-3 by 60%. The amount of caspase-3 activity was higher in the normal level of glucose compared with the activity measured in the rMC-1 cells. 

Total cell death was determined by trypan blue assay 6 . The cells exposed to 5 mM glucose for 96 hours showed a baseline of 6.8% ± 1.4% cell death. Cell death increased to 16.8% ± 4.3% in cells exposed to 25 mM glucose for the same duration. The addition of R-(−)-deprenyl inhibited cell death (7.6% ± 1.3%) to a level that was similar to the level in cells exposed to 5 mM glucose. The resultant death of the cells incubated in 25 mM glucose was not due to an osmolarity phenomenon, because the presence of 20 mM mannitol in the 5 mM glucose medium solution did not induce increased cell death (8.9% ± 2.2%), and the percentage of cell death was comparable to that in cells exposed to 5 mM glucose alone 7 . 

Discussion

Cells in the retina (Müller cells and ganglion cells) and its microvasculature (endothelial cells and pericytes) die at an accelerated rate in diabetes in a process that is consistent with apoptosis. ^{7 28 29} Although the number of retinal cells undergoing diabetes-induced apoptosis is small at any given duration, it is, nevertheless, significantly higher than normal. The mechanism by which hyperglycemia induces cell death in retinal cells is unknown. In this study, Müller cells incubated in elevated glucose underwent apoptosis, and the evidence showed that apoptosis in these cells was initiated through a mechanism that involved nuclear accumulation of GAPDH. Moreover, prevention of GAPDH nuclear accumulation resulted in the survival of these cells. 

The increase of GAPDH protein in the nucleus appeared to be an early event in the hyperglycemia-induced apoptotic process in rMC-1 cells, as well as in freshly isolated human Müller cells. Nuclear accumulation of GAPDH preceded the activation of known apoptosis markers, such as phosphatidylserine residue flip from the inside to the outside of the cell membrane (as seen by annexin-V staining) and the activation of caspase-3. Other groups have stated the importance of nuclear translocation of GAPDH as an early event in apoptosis in neuronal cells and is induced by a variety of death stimuli, such as serum withdrawal and ischemia-reperfusion. ^{14 15 16 18 30 31 32} The effect of elevated glucose on freshly isolated human Müller cells was comparable to that in the rMC-1 cell line. Nuclei of freshly isolated human Müller cells stained strongly positive for nuclear GAPDH when incubated in high glucose. The subsequent apoptosis in human Müller cells, as shown by the increase in caspase-3 activity and annexin-V staining was similar to amounts in rMC-1 cells. These results indicate that rMC-1 cells represent a good model for the study of Müller cell behavior. 

We also have shown that GAPDH nuclear accumulation occurred in vivo in Müller cells of diabetic rats. One third of Müller cells stained positive for nuclear GAPDH in the retina of diabetic rats at 12 weeks of diabetes, compared with a small number in the normal retina. In the normal retina, Müller cells were regularly aligned in the middle of the INL, but in the diabetic retina, this regular order was disrupted, indicating a possible loss of Müller cells. The accumulation of GAPDH in the nucleus of Müller cells may serve as an indicator that these cells will undergo eventual apoptosis, based on our in vitro studies and studies of neurodegenerative diseases. ^{8 14 15 16 17 33 34 35 36} However, GAPDH nuclear translocation is an upstream event, and survival signals may be able to reverse GAPDH nuclear translocation, therefore allowing cells to recover. One in vitro study has demonstrated that serum withdrawal leads to GAPDH nuclear accumulation in neurons that is reversed by the addition of serum, preventing the execution of apoptosis. ³⁷ More studies should be performed to determine the amount of Müller cell loss and its effect on the development of diabetic retinopathy. 

GAPDH nuclear translocation has been associated with apoptosis. All studies so far have demonstrated an increase in GAPDH protein in the nucleus; however, changes in cytosolic GAPDH protein levels during apoptosis varies, depending on the study and cell type. ^{12 16 17 30 31} In our in vitro system using rMC-1 cells, we did not see a significant change in cytosolic GAPDH protein concentration in cells exposed to 25 mM glucose compared with the control. However, GAPDH nuclear accumulation was observed at 48 hours, suggesting an overall increase in GAPDH protein. In addition, staining of rMC-1 cells and freshly isolated Müller cells against GAPDH did not reveal a dramatic loss in cytosolic GAPDH but increase in nuclear GAPDH. However, real-time RT-PCR data comparing GAPDH mRNA levels in rMC-1 cells treated with 5 and 25 mM glucose at different time points showed a tendency for an increased mRNA level in cells exposed to high glucose at an early time points (data not shown). At 96 hours of high glucose incubation, GAPDH protein levels in the cytosol declined significantly below control levels, as would be expected. The term “GAPDH nuclear translocation” has been associated in the literature with induction of apoptosis, but this term may be misleading. It is very well possible that GAPDH accumulates in the nucleus during initiation of apoptosis, either because free movement of GAPDH among cellular compartments is suppressed or because of de novo synthesis of protein. A recent study has shown a specific isoform of GAPDH in the nucleus that may be directly involved in the apoptotic process, suggesting a role for de novo synthesis. ²⁶ More studies should be undertaken to determine the nature of GAPDH in nucleus-related apoptosis. We and others have shown that GAPDH in the nucleus has no glycolytic activity, ^{26 35} but several other functions of GAPDH may occur in the cell (for a review see Ishitani et al. ³² ). Functional determination for GAPDH may depend on the isoform of GAPDH, cellular distribution, oligo formation, ¹⁰ and posttranslational modifications such as APD ribosylation. ³⁸  

R-(−)-deprenyl inhibits the movement of GAPDH into the nucleus and results in cell survival. ^{18 20} Likewise, in our hands, R-(−)-deprenyl prevented high glucose-induced nuclear translocation of GAPDH and apoptosis in rMC-1 cells. Moreover, R-(−)-deprenyl significantly inhibited the amount of GAPDH in the nucleus to levels below that observed in nuclei of cells incubated in a normal glucose concentration. The inhibition of the nuclear accumulation of GAPDH by R-(−)-deprenyl may have resulted in the increased amount of cytosolic GAPDH observed at 48 hours of exposure. The mechanism of R-(−)-deprenyl is not completely understood. It has been reported that R-(−)-deprenyl targets GAPDH and results in the inactivation of the glycolytic function by destroying the tetrameric structure of the enzyme. ^{18 20} R-(−)-deprenyl treatment decreased the cytosol-specific activity of GAPDH that was induced by high glucose, raising the possibility that the enzyme was inactivated by addition of the drug. However, specific activity of GAPDH did not decrease in cells exposed to R-(−)-deprenyl for 96 hours, suggesting that the drug may have other actions, such as promoting synthesis of antiapoptotic proteins. ³⁹ Use of R-(−)-deprenyl in any treatment should be carefully evaluated and limited in concentration to avoid a potential of total inhibition of glycolysis. Impaired cell viability occurred with increased concentrations of R-(−)-deprenyl. 

Overall, GAPDH translocation to the nucleus is an important step in the early hyperglycemia-induced apoptotic process in retinal Müller cells. R-(−)-deprenyl plays a pivotal role in the prevention of apoptosis in cells incubated in diabetic-like concentrations of glucose, seemingly by preventing accumulation of GAPDH in the nucleus. These studies suggest that GAPDH and its nuclear accumulation may play a role previously unanticipated in hyperglycemia-induced death of glial and perhaps other cell types in the retina, and thus offer a new therapeutic target to inhibit the development of diabetic retinopathy. R-(−)-deprenyl or similar agents that limit the mass movement of GAPDH into the nucleus may be useful therapies for diabetic retinopathy. 

 Supported by Juvenile Diabetes Research Foundation International Grant 2-2000-390 (SM), National Institute of Diabetes and Kidney Disease Grant DK57733-01 (SM), and Grants T32 EY07157 (LLK) and EY03523 (VPS) from the Visual Science Training Program, National Eye Institute.

 Submitted for publication November 17, 2003; accepted November 28, 2003.

 Disclosure: L.L. Kusner, None; V.P. Sarthy, None; S. Mohr, 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.

 

Immunostaining of rMC-1 cells and isolated human Müller cells showed nuclear accumulation of GAPDH after 48 hours of high-glucose treatment. rMC-1 cells were exposed to 5 mM glucose (A), 25 mM glucose (B), or 25 mM glucose plus 1 nM R-(−)-deprenyl (C) for 48 hours. Cells were processed for GAPDH staining and then counted for increased GAPDH localization in a field of 100 cells (D) (n = 4 independent experiments). The increase of GAPDH protein in nuclear fractions is illustrated by Western blot analysis of nuclear extracts of rMC-1 cells after 48 hours of treatment (E). rMC-1 cells were exposed to 5 mM glucose (lane 1), 25 mM glucose (lane 2), or 25 mM glucose and 1 nM R-(−)-deprenyl (lane 3). The Western blot is representative of four independent experiments. Isolated human Müller cells were exposed to 7.8 mM glucose (F–H), 25 mM glucose (I–K), or 25 mM glucose plus 1 nM R-(−)-deprenyl (L–N) for 48 hours. Cells were processed for vimentin (red) staining (F, I, L) to identify Müller cells and GAPDH (green) staining (G, J, M) and images were merged (H, K, N).

Figure 2.

 

Presence of GAPDH in nuclei of Müller cells in retinas of normal and diabetic rats. Retinal sections of normal (A) and diabetic (B) rats are represented by immunohistochemistry with antibodies to GS (red) and GAPDH (green). Increased magnification of one nucleus of a Müller cell shows GS (red) alone (C, F), GAPDH (green) alone (D, G), and merged image of both GS and GAPDH (E, H, red and green) (E, H) with normal (C–E) and diabetic (F–H) retinas represented. Müller cells were counted for increased GAPDH nuclear accumulation and graphed in (I) as the number of GAPDH nuclear-positive cells per 100 cells counted.

Figure 3.

 

Expression of cytosolic GAPDH protein at 48 and 96 hours of treatment. rMC-1 cells were exposed to control glucose (5 mM), high glucose (25 mM), or high glucose and 1 nM R-(−)-deprenyl for the indicated times. Data represent the result of 12 independent experiments.

Figure 4.

 

Cytosol- and nuclear-specific GAPDH activity at 48 and 96 hours of treatment. rMC-1 cells were exposed to control glucose (5 mM), high glucose (25 mM), or high glucose, and 1 nM R-(−)-deprenyl for the indicated times. Cytosolic and nuclear fractions of GAPDH are marked accordingly. Data represent the result of 12 independent experiments.

Figure 5.

 

Annexin-V-phycoerythrin staining to assess apoptosis in rMC-1 cells and isolated human Müller cells after 96 hours of treatment. Confluent cultured rMC-1 cells exposed to 5 mM glucose (A), 25 mM glucose (B), or 25 mM glucose and 1 nM R-(−)-deprenyl (C). Human Müller cells exposed to 7.8 mM glucose (D), 25 mM glucose (E), or 25 mM glucose and 1 nM R-(−)-deprenyl (F). Annexin-V–positive cells are stained red and nuclei are stained blue by DAPI.

Figure 6.

 

Caspase-3 activity in Müller cells. rMC-1 cells were exposed to control glucose (5 mM), high glucose (25 mM), or high glucose plus 1 nM R-(−)-deprenyl for 96 hours (A). Data represent the result of 12 independent experiments. Human Müller cells were exposed to control glucose (7.8 mM), high glucose (25 mM), or high glucose plus 1 nM R-(−)-deprenyl for 72 hours (B). Data represent the result of three independent experiments. Activity is presented as picomoles of AFC production per minute per milligram protein.

Figure 7.

 

Percentage of cell death after 96 hours of treatment, as determined by trypan blue. rMC-1 cells were exposed to control glucose (5 mM), high glucose (25 mM), high glucose and 1 nM R-(−)-deprenyl, or control glucose and 20 mM mannitol for 96 hours. Data represent the result of four independent experiments.

 

The authors thank Xia Xi and Denise Ann Hatala for technical assistance; Timothy Kern for discussions; and the National Disease Research Interchange (NDRI, Philadelphia, PA) for human tissue.