Received 17 June 2008 | Accepted 24 November 2008 | Published 3 December 2008
1Department of Ophthalmology, Medical College of Georgia, Augusta, GA; 2School of Medicine, Medical College of Georgia, Augusta, GA; 3Department of Neurology, The Second Affiliated Hospital, Sun Yat-Sen University, Guangzhou, P. R. China; 4Program in Clinical and Experimental Therapeutics, University of Georgia, Athens, GA; 5V.A. Medical Center, Augusta, GA
Correspondence to: Gregory I. Liou, Department of Ophthalmology, Medical College of Georgia, 1120 15th Street, Augusta, GA, 30912; Phone: (706) 721-4599; FAX: (706) 721-7913; email: firstname.lastname@example.org
Purpose: Degenerative retinal diseases are characterized by inflammation and microglial activation. The nonpsychoactive cannabinoid, cannabidiol (CBD), is an anti-inflammatory in models of diabetes and glaucoma. However, the cellular and molecular mechanisms are largely unknown. We tested the hypothesis that retinal inflammation and microglia activation are initiated and sustained by oxidative stress and p38 mitogen-activated protein kinase (MAPK) activation, and that CBD reduces inflammation by blocking these processes.
Methods: Microglial cells were isolated from retinas of newborn rats. Tumor necrosis factor (TNF)-α levels were estimated with ELISA. Nitric oxide (NO) was determined with a NO analyzer. Superoxide anion levels were determined by the chemiluminescence of luminol derivative. Reactive oxygen species (ROS) was estimated by measuring the cellular oxidation products of 2’, 7’-dichlorofluorescin diacetate.
Results: In retinal microglial cells, treatment with lipopolysaccharide (LPS) induced immediate NADPH oxidase-generated ROS. This was followed by p38 MAPK activation and resulted in a time-dependent increase in TNF-α production. At a later phase, LPS induced NO, ROS, and p38 MAPK activation that peaked at 2-6 h and was accompanied by morphological change of microglia. Treatment with 1 μM CBD inhibited ROS formation and p38 MAPK activation, NO and TNF-α formation, and maintained cell morphology. In addition, LPS-treated rat retinas showed an accumulation of macrophages and activated microglia, significant levels of ROS and nitrotyrosine, activation of p38 MAPK, and neuronal apoptosis. These effects were blocked by treatment with 5 mg/kg CBD.
Conclusions: Retinal inflammation and degeneration in uveitis are caused by oxidative stress. CBD exerts anti-inflammatory and neuroprotective effects by a mechanism that involves blocking oxidative stress and activation of p38 MAPK and microglia.
Degenerative retinal diseases such as uveitis, glaucoma, macular degeneration, and diabetic retinopathy all involve inflammation with activated microglia . Inflammation is an active defense reaction against diverse insults, designed to remove or inactivate noxious agents and to inhibit their detrimental effects. Although inflammation serves as a protective function in controlling infections and promoting tissue repair, it can also cause tissue damage and disease. Following brain injury, inflammation occurs in response to glutamate, reactive oxygen species (ROS), nitric oxide (NO), and cytokines including tissue necrosis factor (TNF)-α, released from activated microglia or macrophage, leading to neurodegeneration .
To understand how inflammation affects retinal function in degenerative retinal diseases, it is necessary to examine the processes and signaling pathways during inflammation with in vivo and in vitro models. Endotoxin-induced uveitis (EIU) in rodents is an in vivo model for acute ocular inflammation induced by systemic or local injection of lipopolysaccharide (LPS) [3,4]. EIU is characterized by a breakdown of the blood–ocular barrier  with inflammatory cell infiltration involving the anterior and posterior segments of the eye  and accelerated death of retinal ganglion cells . To further elucidate the molecular events of retinal inflammation, LPS-activated cultured retinal microglial cells have been used as a model to simulate neuroinflammation . The p38 mitogen-activated protein kinase (p38 MAPK), a stress-activated serine/threonine protein kinase, is a downstream target of proinflammatory cytokines and oxidative stress. In addition, activation of p38 MAPK has been also implicated in both induction of inflammatory mediators and transcription-independent effects such as induction of actin reorganization and cellular motility [7-9].
The neuroprotective effects of a nonpsychoactive cannabinoid, cannabidiol (CBD), are largely mediated by its ability to scavenge ROS . We have shown that CBD reduces diabetes- and glutamate-induced ROS formation, p38 MAPK activation, blood–retina barrier breakdown, and retinal degeneration [11,12]. Cannabinoids are known to serve as an anti-inflammatory by modulating the activity of cerebral microglia during inflammation . To date, however, the cellular and molecular mechanism by which CBD reduces inflammation in degenerative retinal diseases is still unclear. In the present study, we test the hypothesis that retinal inflammation and degeneration are initiated by oxidative stress, which activates p38 MAPK, and causes cytokine release that eventually leads to the activation of microglial cells and neurodegeneration. We also show that the neuroprotective and anti-inflammatory effects of CBD involve reducing oxidative stress and modulating p38 MAPK activation in EIU model and LPS-treated retinal microglial cells.
This study used inbred male, 8-10-week-old Sprague-Dawley (SD) rats, each weighing approximately 250 g (Charles River, Durham, NC). The animals were treated in accordance with the ARVO Statement for the Use of Animals in Ophthalmic and Vision Research. Three sets of animals were prepared for a total of 72 rats to study the effect of CBD on EIU. The CBD-treated control or uveitis group received one intraperitoneal injection of CBD (National Institute of Drug Abuse, Research Triangle Park, NC) at 5 mg/kg bodyweight in a 0.25 ml solution that contained 1 part alcohol to 1 part Cremophor EL to 18 parts Ringer solution. This dose was selected based on previous studies showing maximum protection of CBD [11,12]. Control groups received vehicle injections. One hour after CBD or vehicle injection, rats were anesthetized with an intraperitoneal injection of 40 mg/kg Nembutal (Abbott Laboratories, Abbott Park, IL). LPS (Sigma, St. Louis, MO), from Salmonella typhimurium prepared in sterile saline at 1 mg/ml, was given at 0.35 mg/kg bodyweight [13,14]. Experiments were performed 24 h after LPS injection, using 3–6 rats for each group in each experiment.
Immunohistochemistry was performed to identify macrophages or activated microglia on the flat-mounted retinas  using monoclonal antibody CD11b. This antibody reacts with the CR3 complement receptor expressed on monocytes, granulocytes, macrophages, dendritic cells, natural killer (NK) cells, and a subset of lymphocytes. Incubation with the primary and Texas-red fluorescence-labeled secondary antibodies at 4 °C was performed overnight. Immunostained, flat-mounted retinas were placed on slides with the inner side of the retina facing up, covered, and examined with confocal microscopy.
Microglial cells were isolated from retinas of newborn SD rats according to a previous procedure , with minor modifications. Briefly, retinas were dissected within 48 h from newborn SD rat pups. Tissues were collected into 0.01 M PBS (0.01 M NaH2PO4/Na2HPO4, 0.15 M NaCl, pH 7.4) and washed with ice-cold 0.01 M PBS, then digested with 0.125% trypsin for 3–5 min before mixing with a 1:1 blend of DMEM and F12 medium (Invitrogen, Carlsbad, CA) containing 10% fetal bovine serum (FBS; Atlanta Biologicals, Atlanta, GA) and 1% penicillin/streptomycin (Mediatech, Herndon, VA). Retina pieces were triturated by passing through a disposable pipette several times until cells were dispersed. Cells were then filtered through a 100 μm mesh, collected by centrifugation, resuspended in culture medium, and plated onto T150 cell culture flasks (Corning, Corning, NY) at a density of 2×105 cells/cm2. All cultures were maintained in a humidified CO2 incubator at 37 °C and 5% CO2 and fed on the third day, then once every 4 days. After 2 weeks, microglial cells were harvested by shaking the flasks at 100 rpm for 1 h. The cell suspension was centrifuged and the detached cells were replated in a 1:1 mixture of DMEM and F12 medium plus 10% FBS overnight and then in serum-free, low-protein medium (Cellgro Complete containing 0.1% BSA; Mediatech) at designated densities for various experiments. Immunocytochemical studies showed that more than 95% of the cultured cells stained positively for CD11b (Figure 1), with staining localized to the cell membrane. Almost none of these cells showed positive staining for glial fibrillary acidic protein (GFAP), indicating that the majority of the isolated cells were microglia and were not contaminated with astrocytes or Müller cells (data not shown).
Microglia collected from culture flasks were seeded at a density of 5×105 cells/well in 24 well tissue culture plates, or 1×105 cells/well in 96 well plates. One day after seeding, the culture wells were washed with Cellgro Complete (Mediatech) and incubated in the same medium with various treatments. For microglia treatment, LPS at the final concentration of 30 ng/ml (Escherichia coli 0111:B4; Sigma) was added to each well for 24 h. CBD (Cayman Chemical, Ann Arbor, MI) was dissolved in dimethyl sulfoxide (DMSO) as 30 mM stock solution and used at a final concentration of 1 μM. Apocynin, an NADPH oxidase assembly inhibitor, was dissolved in DMSO as 3M stock solution and a range of concentrations of 15, 100, and 200 µM was used. Thenoyltrifluoroacetone (TTFA), an inhibitor of mitochondrial oxidase, was used at 15 μM. Finally, SB 203580, a specific inhibitor of p38 MAPK, was supplied as a 1 mg/ml solution in DMSO and used at a final concentration of 10 μM. At indicated time points, 50 μl aliquots of incubation medium were taken and analyzed for ROS, TNF-α, and NO. Cells were used for superoxide anion assay by chemiluminescence. Also at each time point, cells were washed, homogenized, and subjected to western analysis as will be described.
Microglia collected from culture flasks were seeded at a density of 2×105 cells/chamber in 4 chambered slides. The purity of the microglial cultures and their morphological responses to LPS and CBD treatments were ascertained using immunocytochemical staining analysis for CD11b, a microglial marker, or for F-actin distribution using fluorescent phallotoxins (Molecular Probes, Eugene, OR). The purity and morphology of microglia in culture was examined with confocal microscopy.
Dissected individual rat retinas or washed cultured cells were homogenized in a Mini-Bead beater with treated Ottawa sand in modified RIPA buffer (Upstate, Lake Placid, NY), containing 50 mM Tris, 150 mM NaCl, 1 mM EDTA, 1% Nonidet P-40, 0.25% deoxycholate, supplemented with 40 mM NaF, 2 mM Na3VO4, 0.5 mM phenylmethylsulfonyl fluoride and 1:100 (v/v) of proteinase inhibitor cocktail (Sigma). Insoluble material was removed by centrifugation at 12,000x g at 4 °C for 30 min. Protein was determined by DC Protein Assay (Bio-Rad, Hercules, CA). Antibodies for β-actin (Sigma), phospho-p38 MAPK, p38 MAPK, and caspase-3 were purchased from Cell Signaling Technology (Beverly, MA). Protein of 50−100 μg was boiled in Laemmli sample buffer, separated by SDS–PAGE on a 10% or gradient gel (4 to 20%; Bio-Rad), transferred to nitrocellulose membrane and incubated with specific antibodies. The primary antibody was detected using a horseradish peroxidase-conjugated goat antirabbit antibody and ECL advanced chemiluminescence (Amersham BioSciences, Buckinghamshire, UK). Intensity of immunoreactivity was measured by densitometry (n=3–5 in each group).
TNF-α levels in rat serum or supernatants of culture medium were estimated with ELISA kits (R&D, Minneapolis, MN) per the manufacturer’s instructions. Standards and samples were added and bound by the immobilized antibody. Samples were washed by the washing solution provided in the ELISA kit, and an enzyme-linked polyclonal antibody specific for the cytokine was added to the wells followed by a substrate solution, yielding a colored product. The intensity of the color was measured at 450 nm. The sample levels were calculated from a standard curve and were corrected for protein concentration.
Nitrite (NO2-), the stable breakdown product of NO, was quantitatively reduced to NO and could be quantified by a chemiluminescence detector after reaction with ozone in an NO analyzer (Sievers, GE Analytical Instruments, Boulder, CO). Briefly, medium of cultured microglial cells were deproteinized, and samples containing NO2- were refluxed in glacial acetic acid containing sodium iodide. The amount of NO generated was calculated as the difference in basal and LPS-stimulated NO levels.
The superoxide anion levels were determined in the LPS-treated cells in the presence or absence of 1 μM CBD, 15 μM apocynin (an NADPH oxidase assembly inhibitor), 15 μM TTFA (a mitochondrial oxidase inhibitor), or 100 U/ml superoxide anion dismutase. Cell medium in 96 well plates was replaced with 50 μl of Earle’s balanced salt solution. An equal volume of the same solution containing 800 μM of highly sensitive luminol derivative L-012 (Pure Chemical Industries, Osaka, Japan)  was added to each well. Chemiluminescence was measured with a Synergy-2 plate reader (Bio-Tek Instruments, Winooski, VT).
Dichlorofluorescein (DCF) is the oxidation product of the reagent 2’,7’-dichlorofluorescin diacetate (H2DCFDA; Molecular Probes), a marker of cellular oxidation by hydrogen peroxides and peroxynitrite . Briefly, a 5 mM solution of H2DCFDA was prepared in absolute ethanol and stored under N2 at −20 °C, in the dark. Earle’s balanced salt solution containing 5 μM H2DCFDA was directly applied to and incubated with frozen eye sections for 30 minutes in the dark at room temperature. The sections were then washed with Earle’s balanced salt solution . The fluorescence of DCF was measured and analyzed in eye sections or in cell lysates. The average retinal fluorescence intensity (6 fields/retina) was analyzed using fluorescence microscopy and Ultra-View morphometric software (Molecular Devices, Sunnyvale, CA). The fluorescence of the cell lysates was determined using a spectrofluorometer (Synergy-2 plate reader; BioTek Instruments Inc, Winooski, VT) with 488 nm excitation and 530 nm emission.
Nitrotyrosine immunoreactivity was measured as an indicator for peroxynitrite formation. The distribution of nitrotyrosine in frozen eye sections was analyzed using immunolocalization techniques with confocal microscopy using antibodies specific for nitrotyrosine . Frozen eye sections were fixed with 4% paraformaldehyde then reacted with a polyclonal rabbit antinitrotyrosine antibody (Upstate Biotechnology) followed by Oregon green-conjugated goat antirabbit antibody (Molecular Probes). Data (6 fields/retina; n=6 in each group) were analyzed using fluorescence microscopy and Ultra View Morphometric software to quantify intensity of immunostaining.
Eyes were mounted in optimal cutting temperature (OCT) and 10 μm sections were collected and stored at −80 °C. TUNEL was performed in these frozen sections using the Apop TAG in situ cell death detection kit (TUNEL-FITC; Chemicon International) .
The results are expressed as mean ±SEM. Differences among experimental groups were evaluated by ANOVA, and the significance of differences between groups was assessed by the posthoc test (Fisher’s PLSD) when indicated. Significance was defined as p<0.05.
We examined the retinas from LPS-treated rats to test the hypothesis that activated retinal microglia or infiltrated macrophages are present in the retina during inflammation. Serum and retinas were collected for TNF-α measurement 24 h after footpad injection of LPS. The retinas were also examined by immunofluorescence using an antibody for CD11b, a marker for microglia or macrophages. As shown in Figure 2A, in the normal, untreated retina, the microglia, mainly localized in the ganglion and inner plexiform layers, appear to have long, thin processes. There were few CD11b-positive round cells attached to the retinal blood vessels. Upon LPS treatment, there was a dramatic increase in the number and intensity of the CD11b-positive cells in the flat-mounted retina. CD11b+ cells that attached to the retinal blood vessels began to migrate into the retina. At 24 h, we frequently observed attachment of an increasing number of round cells to the retinal blood vessels, accompanied by infiltration of some cells into the retina and occasionally by focal accumulation of these cells around the blood vessels. We found few CD11b+ round cells in 5 mg/kg CBD-pretreated uveitic rats.
In response to LPS challenge, the immune cells, including macrophages, released proinflammatory cytokines to amplify the inflammation reactions, both locally and systematically. As shown in Figure 2B,C, the serum or retinal levels of TNF-α were 30–40% higher than control 24 h after LPS injection. In LPS-treated rats that were pretreated with CBD, the rise of TNF-α was prevented.
Activated microglia has been reported to play a key role in causing inflammation and neurodegeneration in uveitis and other models . Therefore, we examined the behavior of cultured rat retinal microglial cells in response to LPS treatment to determine how inflammation occurs, and how CBD blocks this event. Rat retinal microglial cells were prepared according to an established protocol . Immunocytochemical studies showed that the prepared cells express CD11b (Figure 2A). Treatment with 30 ng/ml LPS induced microglia activation as indicated by enhanced expression of CD11b and morphological changes (Figure 2B). These effects were prevented by pretreatment of cells with 1 μM CBD (Figure 2C). The activation effect of LPS and the prevention effect of CBD in microglial cells were further confirmed by the increased expression and re-arrangement of F-actin, as compared with the control (Figure 2D-F).
To test the hypothesis that LPS treatment induces TNF-α production, we treated retinal microglial cells with LPS. We then collected culture medium at 5, 15, 20, and 30 min and 1, 2, 4, 6, 12, and 24 h after treatment and assayed for TNF-α. In the medium of LPS-treated cultures, levels of TNF-α began to increase significantly at or after 4 h of LPS treatment, peaking at 6 h, then leveled off within the 24 h period (Figure 3A). Early treatment (0–30 min) with LPS did not induce TNF-a production in microglial cells (data not shown). The levels of TNF-α in control culture medium without LPS treatment did not change throughout the experimental period (Figure 3A and data not shown). We next evaluated the effects of CBD in reducing TNF-α production. As shown in Figure 3B, pretreatment with 1 µM CBD blocked TNF-α production after 6 h of LPS treatment in retinal microglial cells.
It is known that many stress factors, including oxidative stress, lead to inflammation. Therefore, we determined ROS formation after LPS treatment in the presence or absence of CBD. Chemiluminescence assay showed that stimulation of the cells with LPS caused immediate and significant increases in superoxide anion formation (0–30 min), gradually decreasing at 60 min (Figure 4A). Pretreatment of microglial cells with CBD significantly reduced ROS formation during this period. To determine the source of superoxide anion formation, we treated microglia cells were pretreated with 15 μM apocynin (a specific NADPH oxidase assembly inhibitor), 15 μM TTFA (a mitochondrial oxidase inhibitor), or 100 U/ml superoxide anion dismutase (SOD). Chemiluminescence assay showed that stimulation of the cells with LPS caused about 50% increase in the ROS formation at 30 min (Figure 4B). Apocynin at 15 μM significantly reduced LPS-induced superoxide anion formation similar to CBD effects. In contrast, inhibiting mitochondrial oxidase with 15 μM TTFA did not alter LPS-induced superoxide anion formation. While CBD or apocynin did not alter superoxide levels in the controls, treatment with TTFA increased superoxide levels in the control. This may be due to a nonspecific effect of TTFA. The signal was blocked completely by SOD, indicating specificity of the detection of superoxide anion.
Cytokine release has been reported to be mediated by p38 MAPK or oxidative stress in macrophages . Immunoblot assays showed that stimulation of the cells with LPS caused minimal activation of p38 MAPK at 15 min, slowly increased at 30 min, and further increased at 60 min (Figure 5A). Pretreatment of CBD at 1 μM or apocynin at 200 μM blocked p38 MAPK activation in response to LPS. These results suggest that in response to LPS treatment, oxidative stress causes p38 MAPK activation in retinal microglial cells. To understand the regulation of cytokine release in LPS-treated retinal microglial cells, we sought to determine whether the release of TNF-α was mediated by oxidative stress as well as p38 MAPK. The roles of NADPH oxidase and p38 MAPK were examined by using apocynin and the p38 MAPK inhibitor SB203580, respectively . Pretreatment of rat retinal microglial cells with SB203580 at 10 μM significantly decreased the LPS-induced release of TNF-α (Figure 5B). Pretreatment with apocynin at 200 μM, although less effective as compared with SB203580, also significantly decreased the LPS-induced release of TNF-α. Pretreatment with both apocynin and SB203580 did not further decrease TNF-α beyond the effect of SB203580 alone, suggesting that the effects of oxidative stress and p38 MAPK activation are related, and not independent, events.
Treatment with both apocynin and SB203580 did not alter TNF-α levels in the controls. In the same study, pretreatment with CBD at 1 μM almost completely blocked the release of TNF-α (Figure 3 B), suggesting that CBD has a potent effect in blocking oxidative stress and p38 MAPK activation.
To study the effects of LPS and CBD on the release of NO, we collected culture medium at 5, 15, 20, and 30 min and 1, 2, 4, 6, 12, and 24 h after LPS treatment. NO levels in the culture medium were determined using an NO analyzer. The levels of NO in control culture medium without LPS treatment did not change throughout the experimental period. In the LPS-treated cultures, LPS-induced activation of microglia involved a significant late increase in the NO level in culture medium (Figure 6A). The increment of NO levels became significant only after 6 h of LPS treatment and continued for 12 h. LPS-treated rat retinal microglial cells pretreated with CBD for 60 min at 1 µM reduced NO at 6 or 12 h. Peroxynitrite, the combination product from superoxide and NO, increased oxidative and nitrative stresses in the cells. To study the effects of LPS and CBD on the oxidative and nitrative stresses, we collected culture medium at various time points after LPS treatment for fluorescence measurement of DCF (Figure 6B). LPS induced DCF fluorescence as early as 15 min, followed by a further increase at a later phase of 6-12 h. CBD at 1 µM reduced DCF fluorescence during this period. We next evaluated the effects of LPS on activation of p38 MAPK in the late phase (at 2, 6, or 12 h). Immunoblotting of phospho-p38 MAPK showed a time-dependent activation that peaked at 2-6 h in rat retinal microglial cells. Pretreatment with 1 μM CBD significantly reduced it (Figure 6C). LPS treatment did not affect the levels of p38 MAPK or actin proteins (data not shown). Together, these results suggest that LPS induces a second phase of increased oxidative stress and p38 MAPK activation, which coincides with cytokine release and morphological changes with F-actin rearrangement in the microglial cells (Figure 2E).
To elucidate the mechanism and the consequences of inflammation in the uveitic retina and the neuroprotective effects of CBD, we determined the oxidative and nitrative stresses in the rat retina in the presence or absence of CBD 24 h after the injection of LPS. As shown in Figure 7A, images from uveitic retinas of 6 rats showed increased DCF fluorescence in the inner and outer plexiform and the outer segment layers. Quantitative analysis showed a 1.5 fold increase (p<0.05) in the fluorescence intensity in the uveitic retinas compared to normal controls (Figure 7B). Treatment with CBD blocked uveitis-induced oxidative stress as indicated by significant inhibition of fluorescence accumulation in these retinas. We further confirmed the antioxidant effects of CBD by measuring tyrosine nitration. As shown in Figure 7C, images from uveitic retinas of 6 rats showed significant tyrosine nitration within retinal layers with strongest immunoreactivity within the plexiform layers. Quantitative analysis showed that levels of tyrosine nitration were increased 2.8 fold in the uveitic retinas in comparison with the controls. This tyrosine nitration was almost completely eliminated by CBD (Figure 7D).
Our results in cultured retinal microglial cells suggest that endotoxin-induced oxidative stress induces retinal inflammation via activation of p38 MAPK. Activated p38 MAPK is not only an important factor for microglial activation, it also plays an important role in cell survival and death. We therefore examined the role of p38 MAPK in uveitic retinas and determined whether CBD treatment blocks this pathway. Western blot analysis of the phosphorylation of p38 MAPK in uveitic retinas showed an estimated 1.8 fold increase at 24 h after LPS injection (Figure 8A). Treatment with CBD significantly blocked the increases in phosphorylation of p38 MAPK. Accelerated death of retinal ganglion cells and neurons has been reported to be significant in uveitis in vitro . Therefore, we tested the hypothesis that CBD prevents death of neurons in LPS-induced uveitis in rat retinas. Retinal neuronal cell death was determined using TUNEL-FITC analysis on frozen eye sections and western analysis of retinal caspase-3. As shown in Figure 8B, LPS administration in rats resulted in the induction of neuronal death as indicated by the significant number of TUNEL positive cells (green, see white arrows) in the retinal ganglion cell layer. In these LPS-treated rats, the activated p38 MAPK was colocalized with TUNEL positive neurons (Figure 8C). CBD blocked the activation of p38 MAPK and neuronal cell death in LPS-injected animals. Caspase-3 is an intracellular cysteine protease that exists as a pro-enzyme and becomes activated during the cascade of apoptosis. We confirmed LPS-induced apoptosis by significant increases of cleaved caspase-3 expression in LPS-treated animals (Figure 8D). Treatment with 5 mg/kg CBD significantly reduced the number of TUNEL-positive cells and expression of caspase-3.
In the present study, we have demonstrated that stress-activated retinal microglial cells play a key role in mediating retinal inflammation. Using primary culture of rat retinal microglia, we have shown that treatment with LPS first induces oxidative stress, then p38 MAPK activation, leading to accumulation of TNF-α. Retinal inflammation was maintained with nitrative stress, another phase of p38 MAPK activation, and resulted in neurodegeneration. These effects were blocked by treatment with the nonpsychoactive CBD in vitro and in vivo. To our knowledge, this is the first work to study the molecular and cellular processes involved in the stress-activated retinal inflammation and the antiinflammatory and neuroprotective effects of CBD in the retina.
Inflammatory retinal diseases including uveitis, glaucoma, macular degeneration, and diabetic retinopathy are characterized by microglial activation. Microglial cells are commonly described as the central nervous system equivalent of tissue macrophages, which, upon tissue injury, become activated to release glutamate, ROS, and cytokines, leading to inflammation, vascular dysfunction, and neurodegeneration . To test the hypothesis that activated microglia, similar to those in the injured brain, are present in the retina during inflammation, we examined the effects of LPS on rat retinas. Indeed, there was a dramatic increase in infiltrated macrophages in the retina and serum levels of TNF-α in response to LPS injection. In agreement, immune cells including macrophages are known to release proinflammatory cytokines to amplify the inflammatory reactions both locally and systematically. Moreover, our results showed that LPS-treated microglial cells exhibited significant increases in TNF-α and NO levels, indicating that retinal microglial cells provide a feasible model for studying the molecular mechanism and the role of microglia in mediating retinal inflammation. In addition, the use of primary culture of retina microglia offers the advantage of studying the response of a pure population of microglia to controlled and designated treatments.
Using primary culture of rat retinal microglial cells, we showed that treatment with low levels of LPS induced an immediate and significant increase in superoxide anion formation which was blocked by treatment with 1 μM CBD. To identify the source of superoxide anion formation, we compared the antioxidant effects of CBD with TTFA, an inhibitor of mitochondrial oxidase or apocynin, an inhibitor of NADPH oxidase. The results showed comparable effects of CBD to apocynin in inhibiting superoxide anion formation. However, TTFA did not prevent superoxide anion formation in LPS-treated cells. Together, these results suggest that in response to LPS, activation of NADPH oxidase is a major source of superoxide anion. In agreement, other studies have shown that microglial NADPH oxidase is a major source of ROS formation in the LPS model of inflammation [6,19,20]. The antioxidant effects of CBD have been attributed to its ability to directly scavenge ROS . However, this scavenging activity requires higher levels of CBD (2.5–5 μM) . Thus, the antioxidant effects of CBD at 1 μM could be at least partially attributed to its effects in inhibiting NADPH oxidase activity rather than entirely ascribed to ROS scavenging activity.
The receptor mainly responsible for LPS recognition is the toll-like receptor 4 (TLR4), which triggers a variety of intracellular signaling cascades, including MAPK cascade, leading to the induction of transcription of target genes involved in the innate immune response . Therefore, we examined the effects of LPS on p38 MAPK activation in cultured microglial cells. Our results showed an early (30 min) and time-dependent effect of LPS in activating p38 MAPK, which was blocked by treatment with CBD or apocynin (Figure 5A). These experiments have clearly demonstrated the causal relationship of oxidative stress and p38 MAPK activation.
Activation of p38 MAPK has been recognized as a major event involving MAPK in the signaling cascade of TNF-α induction in stimulated macrophages and cerebral microglia [7,9]. In agreement, our results showed that LPS treatment induced a time-dependent induction of TNF-α release that peaked after 6 h and was reduced by specific inhibitors of p38 MAPK (10 μM SB203580) and NADPH oxidase (200 µM apocynin). Furthermore, we demonstrate that LPS-induced release of TNF-α release is blocked almost totally by treatment with 1 µM CBD, clearly showing that CBD exerts more potent effects than apocynin or SB203580.
Induction of inducible nitric oxide synthase and production of NO have been implicated as a mechanism by which activated microglia kill neurons . Therefore, we investigated the effects of LPS on NO production in cultured microglial cells. Our results showed significant and dramatic increases in NO after 6–12 h but not at the early phase of LPS treatment (0–30 min). This effect was in parallel with a second wave of increased formation of ROS and peroxynitrite as indicated by increased DCF fluorescence and nitrotyrosine levels in vitro and in vivo. The late phase effects of LPS were prevented by CBD treatment in cultured microglial cells and in vivo. p38 MAPK, a stress-activated protein kinase, is a downstream target of proinflammatory cytokines including TNF-α  and oxidative stress . The activation of p38 MAPK has been described to induce transcription-independent effects such as induction of actin reorganization and cellular motility . In agreement, our results showed increased expression and redistribution of F-actin in LPS-activated microglia (Figure 2E) within the same time frame as late phase activation of p38 MAPK after 6 h (Figure 6C). Our findings indicate a biphasic pattern of p38 MAPK activation, including an early phase (30–60 min) and a late phase (2–6 h). This suggests that the first phase occurs as a result of oxidative stress to induce cytokine release and the second phase is induced by TNF-α stimulation and maintained by ROS formation, forming the autoregulatory loop of TNF-α, sustaining its own biosynthesis. Similar observations for p38 MAPK have been reported after stimulation with cytokines or vasoactive agents [24-26]. Figure 9 illustrates the pathways of LPS-induced oxidative stress leading to TNF-α release, biphasic p38 MAPK, and microglial activation as well as the mechanism by which CBD blocks these processes.
The anti-inflammatory effect of CBD is not mediated by the known cannabinoid receptors because CBD does not bind well to these receptors. Here, we show that CBD blocks retinal inflammation and microglia activation via inhibition of NADPH oxidase activity and early p38 MAPK phosphorylation, resulting in blocking TNF-α release. The superior effect of CBD over inhibitors of NADPH oxidase and p38 MAPK could be due to modulation of microglial activity by inhibiting an unidentified cannabinoid receptor , or by enhancing endogenous adenosine signaling . Ongoing studies by our group are in progress to further elucidate the role of adenosine signaling in the anti-inflammatory effect of CBD in models of retinal inflammation.
Our results have demonstrated a critical role of oxidative stress in activating microglial cells and causing inflammation in vitro. To determine the consequences of such cellular events in vivo, we determined steady-state levels of inflammatory and oxidative stress markers in relation to retinal neuronal death. The neural retina has a high content of polyunsaturated fatty acids and hence is extremely susceptible to oxidative and nitrative insults . NO and peroxynitrite formation have been reported to play a critical role in the pathogenesis of LPS-induced uveitis . In agreement, our results showed a 1.5 fold increase in oxidative stress and a 2.8 fold increase in peroxynitrite formation as indicated by DCF fluorescence and nitrotyrosine, respectively (Figure 7). These effects were associated with increased inflammation (Figure 1), p38 MAPK activation, and retinal neurodegeneration (Figure 8). Similarly, accelerated death of retinal cells has been reported in rats after LPS treatment . These results confirm our findings that LPS-induced oxidative and nitrative stress activate microglial cells, which causes neuronal degeneration. Our present finding that CBD blocked oxidative and nitrative stress, macrophage infiltration, TNF-α production, and prevented retinal neurodegeneration suggest that CBD represents novel therapeutics in the treatment of inflammation-mediated retinal damage. Furthermore, CBD is an attractive medical alternative to smoked marijuana or plant extract because of its lack of psychoactive effect  and is well tolerated in humans . In conclusion, the data presented here provide evidence that stress-activated retinal microglial cells and their inactivation by CBD represent a central player in retinal inflammation and neuroprotection, respectively.
This work was supported by Knights Templar Educational Foundation of Georgia (G.I.L.), American Diabetic Association (G.I.L.), a Scientist Development Grant from the American Heart Association (A.B.E.), Juvenile Diabetic Research Foundation (A.B.E.), and University of Georgia Alliance Research Foundation (A.B.E.).