Cannabidiol protects retinal neurons by preserving glutamine synthetase activity in diabetes.

PURPOSE: We have previously shown that non-psychotropic cannabidiol (CBD) protects retinal neurons in diabetic rats by inhibiting reactive oxygen species and blocking tyrosine nitration. Tyrosine nitration may inhibit glutamine synthetase (GS), causing glutamate accumulation and leading to further neuronal cell death. We propose to test the hypothesis that diabetes-induced glutamate accumulation in the retina is associated with tyrosine nitration of GS and that CBD treatment inhibits this process.
METHODS: Sprague Dawley rats were made diabetic by streptozotocin injection and received either vehicle or CBD (10 mg/kg/2 days). After eight weeks, retinal cell death, Müller cell activation, GS tyrosine nitration, and GS activity were determined.
RESULTS: Diabetes causes significant increases in retinal oxidative and nitrative stress compared with controls. These effects were associated with Müller cell activation and dysfunction as well as with impaired GS activity and tyrosine nitration of GS. Cannabidiol treatment reversed these effects. Retinal neuronal death was indicated by numerous terminal deoxynucleotidyl transferase dUTP nick end-labeling (TUNEL)-labeled cells in diabetic rats compared with untreated controls or CBD-treated rats.
CONCLUSIONS: These results suggest that diabetes-induced tyrosine nitration impairs GS activity and that CBD preserves GS activity and retinal neurons by blocking tyrosine nitration.

Introduction

Diabetic retinopathy (DR) is the leading cause of blindness in working-age adults, affecting nearly 21 million people in the United States alone (American Diabetes Association). The early clinical features of DR in patients with diabetes as well as observations in experimental diabetes include vascular permeability and vitreoretinal neovascularization secondary to vascular dysfunction as well as retinal glial dysfunction and neuronal cell loss (reviewed in [1,2]). The biochemical mechanisms involved in diabetic retinopathy are complex and include the activation of several cellular pathways (reviewed in [3]). Previously, we and others have shown that an increase in peroxynitrite, as indicated by tyrosine nitration, correlates with accelerated retinal endothelial cell death, breakdown of the blood-retinal barrier (BRB), and accelerated neuronal cell death in experimental models of diabetes, inflammation, and neurotoxicity [4-15]. These studies suggest peroxynitrite plays a key role in mediating different aspects of DR. However, the causal role of diabetes-induced tyrosine nitration in mediating Müller glial cell injury and dysfunction has not been elucidated.

In response to hyperglycemia-induced oxidative stress, both microglial and macroglial cells are activated, and the function of macroglia in transporting glutamate by glutamate transporters and in metabolizing glutamate by glutamine synthetase (GS) may be impaired [16-18]. This may lead to glutamate accumulation, such as that reported in the vitreous humor of diabetic patients [19] and in the retina of diabetic animals [16,20]. Recent studies demonstrated that GS is susceptible to tyrosine nitration, which subsequently can impair the enzyme activity [21,22]. Together, these observations prompted us to study the role of diabetes-induced tyrosine nitration in mediating glial injury and GS dysfunction.

Cannabinoids are known to possess therapeutic properties, including anti-oxidant, anti-inflammatory, and N-methyl-D-aspartic acid (NMDA) receptor-activation blocking activity [23-25]. Non-psychotropic cannabidiol (CBD) has been shown to prevent neuronal damage to the central nervous system in gerbils caused by cerebral ischemia [26]. We recently demonstrated the neuroprotective effect of CBD via antioxidant and anti-inflammatory action in rat models of NMDA-induced retinal neurotoxicity and lipopolysaccharide (LPS)-induced neurotoxicity [9,15,27] as well as the anti-inflammatory and BRB-preserving effects in diabetic rats [12]. However, the mechanism of the neuroprotective effect of CBD via preserving glial function in diabetic retina has not been studied. The present study evaluates the ability of CBD to reduce oxidative and nitrative stress, preserve GS function, and prevent neuronal cell death in experimental diabetes.

Methods

Experimental animals and retina isolation

Eight-week-old male Sprague Dawley rats (≥200 g) were obtained from Charles River (Wilmington, MA) and made diabetic by tail-vein injection of streptozotocin (STZ; Sigma, St. Louis, MO) 65 mg per kg of bodyweight in 0.1 M citrate-buffered saline, pH 4.5. All procedures involving animals were performed in accordance with the ARVO Statement for the Use of Animals in Ophthalmic and Vision Research and with Medical College of Georgia (Augusta, GA) guidelines. Diabetes was confirmed by detection of glucose in the urine and blood of injected animals (>250 mg/dl). Three sets of animals were prepared for a total of 48 rats to study the effects of eight weeks of induced diabetes. The following groups were prepared: untreated controls, controls treated with CBD, untreated diabetics, and diabetics treated with CBD. The CBD-treated groups received intraperitoneal injections of CBD (10 mg/kg) every other day. Cannabidiol was obtained from the National Institute of Drug Abuse (Research Triangle Park, NC), and a fresh solution in 0.25 ml of 1:1:18 alcohol:cremorphol:Ringer solution was prepared. Control groups received vehicle injections at the same time points. Streptozotocin-injected animals had significant increases in blood glucose level (495±35 mg/dl) compared with untreated controls (135±7 mg/dl). Treatment with CBD did not alter blood glucose levels in diabetic animals (455±37 mg/dl) or in treated controls (125±5 mg/dl). After eight weeks of diabetes, eyes were enucleated and retinas were dissected for further analyses.

Glutamine synthetase activity

Frozen retinas were thawed and pulse-sonicated in ice-cold sonication buffer (PBS with 1 mM 2- mercaptoethanol). The ability of an aliquot to convert 14C-glutamate to 14C-glutamine was measured by a modification of a published method [28]. Briefly, 30 µl of sample were reacted with an equivalent volume of reaction buffer (100 mM imidazole HCl, pH 7.2, 30 mM MgCl2, 20 mM ATP, 8 mM NH4Cl, 1 mM 2-mercaptoethanol, and 14 mM 14C-glutamate; Specific Activity, 5×105 CPM/ml) for 20 min at 37 °C. The reaction was stopped by adding 600 µl of 2% perchloric acid (PCA). In the control reactions, PCA was added at the beginning of the incubation period. Glutamine was then separated from glutamate by anion exchange chromatography (AG 1-X8 Resin Acetate form; Bio-Rad, Hercules, CA) and quantified by liquid scintillation counting [29]. Numbers were normalized to proteins measured by DC Protein Assay (Bio-Rad).

Immunoprecipitation and western blot analysis of glutamine synthetase

Individual rat retinas were dissected and homogenized in a Mini-Bead beater with treated Ottawa sand in 250 μl of modified RIPA buffer supplemented with inhibitors for proteases and phosphatases as described previously [12]. Retinal protein extract was determined by DC Protein Assay (Bio-Rad). The supernatants containing 500 μg of protein were combined with 5 µl of polyclonal anti-glutamine synthetase Santa Cruz Biotechnology (Santa Cruz, CA) and 50 µl of protein A/G agarose (Santa Cruz) and mixed overnight at 4 °C. The immunoadsorbents were recovered by centrifugation for 5 min at 700× g and washed three times in modified RIPA buffer and twice in 50 mM Tris (pH 7.5) containing 0.1% (w/v) sodium dodecyl sulfate and 150 mM NaCl. The samples were eluted into 60 µl of sodium dodecyl sulfate loading buffer and subjected to sodium dodecyl sulfate PAGE. The membranes were incubated with polyclonal anti-nitrotyrosine (1:1,000; Upstate Biotechnology, Millipore, Billerica, MA) and then with peroxidase-conjugated goat antirabbit (1:5,000) for detection of immunoreactive bands by ECL advance chemiluminescence (GE Healthcare, Piscataway, NJ).

Terminal dUTP nick end-labeling analysis

Terminal dUTP nick end-labeling (TUNEL) analysis was performed using the ApopTag Fluorescein In Situ Apoptosis Detection Kit (Millipore) following the manufacturer’s directions as described previously [9]. Briefly, OCT-frozen eye sections (10 µm) from each group were fixed using paraformaldehyde (PFA) and ethanol:acetic acid (2:1). Then, the samples were incubated with Terminal Deoxynucleotidyl Transferase followed by incubation with anti-digoxigenin conjugate. Propidium iodide 1 µg/ml was added as a nuclear counter stain. On completion of the TUNEL assay, coverslips were applied using VECTASHIELD Mounting Medium for fluorescence (Vector Laboratories, Burlingame, CA). Each section was systematically scanned for positive green fluorescent cells in retinal layers indicating apoptosis. Images were obtained using an AxioObserver.Z1 Microscope (Carl Zeiss, Oberkochen, Germany) with 200× magnification. Four to five animals were used for each group, and the number of TUNEL positive cells were counted in four fields of the mid-peripheral retina and calculated as the number per mm2 of retinal area.

Immunolocalization studies

The distribution of nitrotyrosine, glial fibrillary acidic protein (GFAP), and caspase-3 in frozen eye sections was analyzed using immunolocalization techniques as described previously [12]. Retinal sections were fixed with 4% paraformaldehyde and then reacted with a polyclonal rabbit anti-nitrotyrosine (Millipore), mouse anti-GFAP (Cell Signaling Technology, Danvers, MA) antibody, or rabbit caspase-3 antibody (Cell Signaling Technology) followed by Oregon Green-conjugated goat antirabbit or antimouse antibody (Molecular Probes, Carlsbad, CA). Data (10 fields/retina, n=6 in each group) were analyzed using fluorescence microscopy and UltraVIEW morphometric software to quantify the intensity of immunostaining. For colocalization studies of caspase-3 within retinal ganglion cells, retina sections were stained with mouse Brn-3 antibody (Santa Cruz Biotechnology) followed by Texas Red-conjugated antibody (Molecular Probes).

Dichlorofluorescein assay

Dichlorofluorescein (DCF) is the oxidation product of the reagent 2’,7’-dichlorofluorescin diacetate (H2DCFDA; Molecular Probes, Eugene, OR), a marker of cellular oxidation by hydrogen peroxide and peroxynitrite [12]. Earle’s balanced salt solution containing H2DCFDA was incubated with retina sections, and the fluorescence of DCF was measured and analyzed. The average retinal fluorescence intensity (10 fields/retina, n=6 in each group) was analyzed using AxioObserver.Z1 Microscope and Axio-software (Carl Zeiss).

Data analysis

The results are expressed as mean±SEM. Differences among experimental groups were evaluated by performing an ANOVA (ANOVA), and the significance of differences between groups was assessed by a post-hoc test (Fisher’s PLSD) when indicated. Significance was defined as p<0.05.

Results

Cannabidiol reduces diabetes-induced oxidative and nitrative stress

We have previously shown the neuroprotective effects of CBD in short-term diabetes via inhibiting oxidative and nitrative stress [12]. Here, we tested the antioxidant effect of CBD after eight weeks of diabetes. As shown in Figure 1A, diabetes-induced neurotoxicity involved significant tyrosine nitration within retinal layers with the strongest immunoreactivity in the ganglion cell layer. Quantitative analysis showed that levels of tyrosine nitration increased ~1.6-fold in the diabetic retinas in comparison with the controls (Figure 1B). This tyrosine nitration was almost completely eliminated by CBD (10 mg/kg/2 days). The antioxidant effect of CBD was further confirmed by blocking the fluorescence of DCF, a general marker for both oxidative and nitrative stress in diabetic retinas (Figure 1C,D). The treated control rat retinas were not affected by CBD treatment.

Figure 1

Cannabidiol (CBD) reduces oxidative and nitrative stress in diabetic retinas. A: Representative images show the distribution of nitrotyrosine immunolocalization in different retinal layers, the ganglion cell layer (GCL), the inner plexiform layer (IPL), the inner nuclear layer (INL), and the outer nuclear layer (ONL), and the retinal pigment epithelium (RPE) (magnification, 200×). B: Morphometric analysis of fluorescence intensity in serial sections of rat eyes shows that diabetic rats had a significant increase in fluorescence compared with controls. Treatment with CBD (10 mg/kg/2 days) inhibited nitrotyrosine formation in the diabetic rats but not in the normal controls. Data shown are the mean±SEM of six or seven animals in each group (*p<0.05). C: Representative images show the distribution of dichlorofluorescein (DCF) fluorescence in different retinal layers, the GCL, the IPL, the INL, and the ONL, and the RPE (magnification, 200×). D: CBD reduces peroxides in the retinas of diabetic rats as represented by morphometric analysis of DCF fluorescence showing that diabetic rats had a significant increase in fluorescence compared with controls. Treatment with CBD (10 g/kg/2 days) inhibited reactive oxygen species formation in diabetic rats but not normal controls. Data shown is the mean±SEM of five or six animals in each group (*p<0.05).

Cannabidiol prevents diabetes-induced Müller glial cell injury

Glial activation, as indicated by GFAP, is a common response to stress conditions. There are two types of glial cells in the retina: astrocytes and Müller cells. Therefore, we assessed glial injury in response to the diabetic insult by immunolocalization of GFAP. Astrocytes were notably positively and equally labeled with GFAP in all groups. As shown in Figure 2, only retinas from the diabetic group demonstrated an increase in the intensity of GFAP immunoreactivity in the filaments of Müller cells that extended from the nerve fiber layer and inner plexiform layer into the outer nuclear layer of retina as compared with controls or the CBD-treated group.

Figure 2

Cannabidiol (CBD) prevents Müller cell activation in diabetic animals. Representative images of glial fibrillary acidic protein (GFAP) showing abundant immunofluorescence at the end-feet of the Müller cells and the radial processes stained intensely throughout both the inner and outer retina in the diabetic retinas compared with normal controls. This effect was blocked by treatment with CBD (10 mg/kg/2days, i.p.). Similar results were obtained from five additional animals per group.

Cannabidiol prevents diabetes-induced glutamine synthetase nitration and restores its activity

Diabetes-induced peroxynitrite formation and its subsequent alteration of protein function via tyrosine nitration are well documented [30]. Recent studies have demonstrated that GS is a susceptible target for tyrosine nitration [21]. Therefore, we evaluated the specific tyrosine nitration levels of GS and the extent to which its activity can be altered in diabetic rat retinas. As shown in Figure 3, diabetes caused significant tyrosine nitration (2.3-fold) of GS that was significantly reduced by treatment with CBD. We next evaluated the effects of tyrosine nitration on GS activity. Indeed, diabetes-induced GS tyrosine nitration was positively correlated with a significant inhibition (40%) of GS activity (Figure 4), and treatment with CBD restored this activity in the diabetic animals. These results suggest a causal role of tyrosine nitration in impairing the function of GS, which can lead to the accumulation of glutamate and possibly cause neurotoxicity.

Figure 3

Cannabidiol (CBD) reduces GS nitration in diabetic (D) animals. Immunoprecipitation with anti-glutamine synthetase (GS) and western blot analysis using anti-nitrotyrosine antibody show that diabetes significantly increased the tyrosine nitration of GS compared with normal retinas. This effect was blocked by treatment with CBD (10 mg/kg/2days, i.p.; n=4–6 retinas/group, *p<0.05, versus control [standard error of mean]).

Figure 4

Cannabidiol (CBD) restores diabetes-impaired glutamine synthetase (GS) activity. Glutamine synthetase activity measured by the ability of the sample to convert 14C-glutamate to 14C-glutamine demonstrated significant inhibition of GS activity in diabetic rat retinas compared with controls. The GS activity was restored by treating the diabetic animals with CBD (10 mg/kg/2days, i.p.; n=4–5 retinas/group, *p<0.05, versus control [standard error of mean]).

Cannabidiol prevents diabetes-induced neuronal cell death and activation of caspase-3

We next evaluated neuronal death after eight weeks of diabetes. Our results demonstrated that diabetic rat retina showed significant increases in TUNEL positive cells (~8-fold) mainly in retinal ganglion cells and inner retinal layers compared with controls. Treatment with CBD blocked neuronal cell death in diabetic animals but did not affect treated controls (Figure 5A,B). Neuronal cell death in diabetic animals was further confirmed by prominent immunostaining of caspase-3, a known marker for apoptosis, within the ganglion cell layer (GCL) as indicated by the specific retinal ganglion cell marker Brn-3. The ganglion cell layer notably includes ~35%–40% displaced amacrine cells in addition to retinal ganglion cells. Treatment with CBD blocked neuronal cell death in diabetic animals but did not affect treated controls (Figure 5C,D).

Figure 5

Retinal neuroprotective effect of cannabidiol (CBD) in experimental diabetes. A: Representative images show the terminal deoxynucleotidyl transferase dUTP nick end-labeling (TUNEL) labeling of frozen eye sections from the diabetic rats (eight weeks) in different retinal layers. TUNEL-positive cells (arrows) were distributed mainly in the inner retinal layers. B: Statistical analysis of TUNEL-positive nuclei in various groups. At least four fields per mid-peripheral retina were counted for each retina from one animal. (n=4–5 retinas/group, *p<0.05, versus control [standard error of mean]). Treating the diabetic animals with CBD (10 mg/kg/2 days, i.p.) prevented neuronal death. C: Representative images show the localization of the apoptotic marker caspase-3 in the ganglion cell layer and the inner retinal layer in diabetic retina sections but not in other groups. D: Enlarged window of retina sections showing colocalization (yellow) of the apoptotic marker caspase-3 (green) within the retinal ganglion cell layer labeled with Brn-3 (red). The layers shown are the ganglion cell layer (GCL), the inner plexiform layer (IPL), the inner nuclear layer (INL), and the outer nuclear layer (ONL).

Discussion

Diabetes-induced retinal oxidative and nitrative stress have been well documented in patients and animals and have been positively correlated with neuronal cell death [12,13,31-34]. In response to neuronal injury, glial cells including microglial and macroglial cells are activated. This might be followed by neuroinflammation, during which activated microglial cells release TNF-alpha and migrate toward dying neurons to further exacerbate the damage [35]. However, the effects of diabetes-induced oxidative and nitrative stress on macroglial activation and how this can affect neuronal function have not been fully elucidated. Indeed, our results showed a significant increase in oxidative and nitrative stress as indicated by significant increases in DCF fluorescence and nitrotyrosine as well as prominent Müller glial cell activation compared with controls. Exposure of retinal Müller glial cells to high glucose levels stimulates oxidative stress and peroxynitrite formation ( [36], unpublished data). However, peroxynitrite produced by glial cells is not toxic by itself but causes activation and expression of proinflammatory cytokines [37]. Our previous studies have shown that Müller cells are not among the retinal cell population undergoing apoptosis early in diabetes [13]. Our current study demonstrated that Müller cells are activated as evidenced by an enhanced intensity of GFAP immunoreactivity in the filaments of Müller cells in diabetic retinas that was blocked by CBD treatment.

Previous studies have documented the adverse effects of diabetes on the function of Müller cells in transporting glutamate by glutamate transporter or in metabolizing glutamate by GS [16-18,20]. Although alterations in glutamate transporter activity during diabetes remain controversial, impairment of GS activity has been previously reported [38-40]. Interestingly, recombinant GS enzyme from E. coli, rat liver, or mammalian GS has been reported to be a susceptible target for tyrosine nitration that might reduce its activity [21,22,41]. Therefore, we investigated GS nitration and its impact on GS activity in diabetic rat retinas. Our results showed a 2.3 fold increase in GS tyrosine nitration that was associated with a significant reduction (40%) in GS activity in diabetic retinas compared with controls. Our results lend further support to previous reports showing that diabetes can alter glial function and impair GS activity [38,42,43]. Although the concept of GS nitration and the subsequent impairment of its activity has been demonstrated at the recombinant protein level, we believe our study provides the first experimental evidence in a diabetic model. Further studies of human samples should provide clinical evidence and implications for GS nitration.

Tyrosine nitration and the subsequent loss of protein function have been documented in response to peroxynitrite [11,30,44-46]. Furthermore, the impact of GS nitration and its impairing activity is evidenced by glutamate accumulation, as reported in the vitreous humor of diabetic patients [19] and in the retinas of diabetic animals [16,17,20]. Glutamate excitotoxicity occurs via the activation of NMDA receptors to induce calcium influx and the release of superoxide and nitric oxide, leading to the formation of peroxynitrite and neuronal death [47,48]. Diabetes-impaired GS activity should lead to the accumulation of glutamate and the formation of peroxynitrite, which in turn can sustain tyrosine nitration and the inhibition of GS activity. This vicious cycle of glial dysfunction will result in cell death and the injury of adjacent retinal neurons. Therefore, we next evaluated neuronal cell death in the diabetic animals. Indeed, our results showed significant increases in TUNEL-positive cells that were mainly localized in retinal ganglion cells and inner retinal layers in the retinas of diabetic animals compared with controls. Additional immunolocalization studies using caspase-3, a known apoptotic marker, and Brn-3, the specific retinal ganglion cell marker, confirmed apoptosis of ganglion cells in the diabetic animals. As further support, previous studies have demonstrated retinal ganglion cell loss in response to STZ diabetes within the same time frame, eight weeks [49], that continues to happen later during the progression of the disease [32]. Retinal ganglion cells represent about 60%–65% of neurons in addition to displaced amacrine cells in the ganglion cell layer. These findings suggest a loop where diabetes-induced oxidative and nitrative stress alter the function of Müller cells by impairing GS activity, leading to glutamate neurotoxicity and sustaining retinal neuronal death.

Treating diabetic animals with CBD blocked the increases in oxidative and nitrative stress and significantly reduced the number of apoptotic cells. Neurons are highly susceptible to oxidative stress, which can induce apoptosis; therefore, it is likely that diabetes-induced oxidative stress leads to neuronal injury. Several reports have described the neuroprotective effects of CBD via blocking reactive oxygen species or nitrotyrosine formation in glutamate-induced cell death in neuron cultures and in an NMDA-induced neurotoxicity [9,12,15,25]. Here, we demonstrate a novel role of CBD in restoring GS activity by reducing its tyrosine nitration in diabetic animals. This effect was associated with a significant reduction in Müller glial cell activation, which confirms the preservation of its morphology and function in the diabetic animals. Together, our present findings suggest that CBD represents novel therapeutics in the treatment of diabetes and stress-mediated retinal damage. Furthermore, CBD is an attractive medical alternative to smoked marijuana or plant extract because of its lack of psychoactive effect and because it is well tolerated in humans when administered chronically [50,51]. In addition, CBD has been approved for the treatment of inflammation, pain, and spasticity associated with multiple sclerosis in humans (reviewed in [52]). In conclusion, the data presented here provide experimental evidence that diabetes-activated retinal glial cells represent a central player in retinal neurodegeneration.