Cannabinoids ablate release of TNFα in rat microglial cells stimulated with lypopolysaccharide

Abstract

Upon activation, brain microglial cells release proinflammatory mediators, such as TNFα, which may play an important role in eliciting neuroinflammatory processes causing brain damage. As cannabinoids have been reported to exert anti-inflammatory and neuroprotective actions in the brain, we here examined the effect of both synthetic and endogenous cannabinoids on TNFα release elicited by bacterial endotoxin lypopolysaccharide (LPS) in cultured microglia. Exposure of primary cultures of rat cortical microglial cells to LPS significantly stimulated TNFα mRNA expression and release. The endogenous cannabinoids anandamide and 2-arachidonylglycerol (2-AG), as well as the synthetic cannabinoids (+)WIN 55,212-2, CP 55,940, and HU210, inhibited in a concentration-dependent manner (1–10 μM) the LPS-induced TNFα release. Unlike the high-affinity cannabinoid receptor agonist (+)WIN 55,212-2, the low-affinity stereoisomer (−)WIN 55,212-2 did not exert any significant inhibition on TNFα release. Given this stereoselectivity, the ability of (+)WIN 55,212-2 to inhibit LPS-induced TNFα release from microglia is most likely receptor-mediated. By RT-PCR we found that the two G_i/o protein-coupled cannabinoid receptors (type 1 and 2) are both expressed in microglial cultures. However, selective antagonists of type 1 (SR141716A and AM251) and type 2 (SR144528) cannabinoid receptors did not affect the effect of (+)WIN 55,212-2. Consistent with this finding is the observation that the ablative effect of (+)WIN 55,212-2 on LPS-evoked release of TNFα was not sensitive to the G_i/o protein inactivator pertussis toxin. In addition, the cAMP elevating agents dibutyryl cAMP and forskolin both abolished LPS-induced TNFα release, thus rendering unlikely the possibility that (+)WIN 55,212-2 could ablate TNFα release through the inhibition of adenylate cyclase via the G_i-coupled cannabinoid receptors type 1 and 2. In summary, our data indicate that both synthetic and endogenous cannabinoids inhibit LPS-induced release of TNFα from microglial cells. By showing that such effect does not appear to be mediated by either CB receptor type 1 or 2, we provide evidence suggestive of the existence of yet unidentified cannabinoid receptor(s) in brain microglia. GLIA 41:161–168, 2003. © 2003 Wiley-Liss, Inc.

INTRODUCTION

Parenchymal microglial cells are ubiquitously distributed in the central nervous system (CNS), where they comprise up to 20% of the total non-neuronal cell population. In response to CNS insults, such as ischemia and trauma or bacterial endotoxins, microglia become activated and produce various cytokines and neurotoxic substances implicated today in the exacerbation of the neuropathological damage accompanying neurological disorders, including AIDS dementia, stroke, and Alzheimer's disease (Gehrmann et al., 1995). For example, microglia release substantial amounts of proinflammatory cytokines, including TNFα, in response to exposure to the Gram-negative bacterial endotoxin lypopolysaccharide (LPS) (Zielasek and Hartung, 1996). TNFα is a pivotal mediator of inflammatory processes in CNS. It induces the expression of cell adhesion molecules in endothelial cells and astrocytes, thereby favoring leukocyte infiltration and inflammatory brain damage (Merrill and Benveniste, 1996). TNFα is also capable of inducing apoptosis of myelin-producing oligodendrocytes (Kassiotis et al., 1999) and of silencing survival signaling in neurons (Venters et al., 2000). In addition, TNFα has the ability to stimulate the expression of inducible nitric oxide synthase (NOS), which generates the cytotoxic oxygen radical nitric oxide (Heneka et al., 1998). Therefore, pharmacological modulation of microglial TNFα release may be of therapeutic relevance in the clinical management of diverse neurological diseases such as multiple sclerosis (MS), stroke, and Alzheimer's disease.

Cannabinoids, such as delta-9-tetrahydrocannabinol, the principal psychoactive component of Cannabis sativa (marijuana), and its synthetic (e.g., HU-210 and WIN 55212-2) or their endogenous counterparts (anandamide and 2-arachidonylglycerol), have been reported to exert neuroprotective actions, in part linked with their capability to modulate cytokine negatively, including TNFα production (Molina-Holgado et al., 1997; Shohami et al., 1997; Leker et al., 1999; Achiron et al., 2000). In addition, both synthetic and endogenously occurring cannabinoids exert most of their central and peripheral effects by binding to specific G-protein–coupled receptors (Axelrod and Felder, 1998; Di Marzo et al., 1998). To date, two distinct G-protein receptors with seven transmembrane domains have been identified as cannabinoid receptors, namely cannabinoid receptor type 1 (CB1) (Matsuda et al., 1990) and type 2 (CB2) (Munro et al., 1993), both of which bind cannabimimetic agents. CB1, also known as the central cannabinoid receptor, is primarily expressed within the CNS, including microglia (Galiegue et al., 1995; Waksman et al., 1999). Conversely, the peripheral receptor CB2 is the predominant form of the receptor expressed within the immune system (Galiegue et al., 1995). Activation of either CB1 and CB2 receptors has been shown to trigger several signaling pathways, including inhibition of adenylate cyclase activity through a pertussis toxin-sensitive GTP-binding protein (G_i).

In vitro delta-9-tetrahydrocannabinol (THC), as well as anandamide, a putative endogenous cannabinoid receptor ligand, inhibits the production of TNFα in murine macrophage cultures (Gallily et al., 2000). In cultured astroglia, synthetic cannabinoids exert inhibitory effects on isoproterenol-induced cAMP accumulation (Sagan et al., 1999) as well as on LPS-stimulated NO production (Waksman et al., 1999). It has also been reported that synthetic cannabinoids, such as levonantradol, CP55,940, and, to a much lesser extent, THC and the endogenous cannabinoid anandamide, partially counteract LPS-induced upregulation of TNFα mRNA (Puffenbarger et al., 1999). To extend these findings, we explored the effects of cannabinoids on the release of the mature TNFα protein. Here we show that synthetic cannabinoids, in particular (+)WIN 55.212-2, can modulate in a stereoselective manner TNFα release elicited by LPS in primary cultures of rat cortical microglia cells. Further, we report that the inhibitory effect of (+)WIN 55,212-2 on LPS-induced TNFα release, although stereoselective, does not likely involve either cannabinoid receptor type 1 or 2. Altogether our study shows that cannabinoids negatively modulate release of TNFα through a mechanism likely involving yet unidentified cannabinoid receptors.

MATERIALS AND METHODS

Reagents and Media

Lipopolysaccharide from E. coli stereotype O26:B6, anandamide, and WIN 55212-2 (R)-(+)- and R-(−)-[2,3-dihydro-5-methyl-3-[(4-morpholinyl)methyl]pyrrolo[1,2,3-de]1,4-benzoxazinyl]-(1-naphthalenyl)methanone-mesylate] and 2-arachidonylglycerol were purchased from Sigma Chemical (St. Louis, MO). CP55,940 [(−)-cis-3-[2-Hydroxy4-(1,1-dimethylheptyl) phenyl]-trans-4-(3-hydroxypropyl) cyclohexanol)], HU 210 [(−)-11-hydroxy-delta(8)-tetrahydrocannabinol-dimethylheptyl], and AM251 [N-(piperidin-1-yl)-5-(4-iodophenyl)-1-(2,4-dichlorophenyl)-4-methyl-1H-pyrazole-3-carboxamide] were purchased from Tocris Cookson (St. Louis, MO). SR141716A [N-(piperidin-1-yl)-5-4-(chlorophenyl)-1-(2-4-dichlorophenyl)-4-methyl-1H-pyrazole-3-carboxamidehydrochloride] and SR144528 [N-[(1S)-endo-1,3,3-trimethylbicyclo[2.2.1]heptan-2-yl]5-(4-choro-3-methylphenyl)-1-(4-methylbenzyl)pyrazole-3–carboxamide] were a kind gift from Sanofi (Montpellier, France). Pertussis toxin was obtained from List (Campbell, CA). SQ 22536 [9-(tetrahydro-2-furanyl)-9H-purin-6-amine] and 2′-5′-dideoxyadenosine were purchased from Calbiochem (La Jolla, CA). Dulbecco's modified Eagle's medium (DMEM), fetal calf serum, L-glutamine, and penicillin/streptomycin were purchased from Biochrom (Berlin, Germany). Culture dishes and plates were purchased from Iwaki (Japan). ELISA plates (Maxisorp) were purchased from Nunc (Roskilde, Denmark). RT-PCR reagents were purchased from Promega (Madison, WI). Oligonucleotides were purchased from MWG (Ebersberg, Germany).

Microglial Cell Cultures

Rat primary cortical microglial cells were obtained from cortices of neonatal SD rats (Charles River) as previously described (Levi et al., 1993). Briefly, cells were dissociated from cerebral hemispheres by trypsination (0.025% trypsin at 37°C for 15 min), trituration, and filtration through 100 μm cell strainers. Cells were plated at the density of 10⁵ cells/cm² in culture medium [Dulbecco's modified Eagle medium high-glucose formula (4.5 g/l) supplemented with heat-inactivated 10% fetal calf serum, 2 mM glutamine, 100 U/ml penicillin, and 0.1 mg/ml streptomycin]. Culture medium was changed after 3 days and then twice a week. After 2 weeks, cultures contained glial cells including ameboid microglia mostly localized on the top of the cellular layer. The loosely adherent microglial cells were recovered by mild shaking. After centrifugation (100 g for 5 min), cell viability was determined by trypan blue exclusion and viable cells were resuspended in culture medium and plated at a final density of 2 × 10⁵ cells/cm² onto 48-well plates or onto 35 mm dishes. Non-adherent cells were removed 30 min after plating by changing the medium and adherent cells were incubated for 24 h in culture medium before being used for the treatments. More than 95% of the adherent cells were positive for Bandeiraea Simplicifolia isolectin B4, a specific microglial marker (Fig. 1A). Minimal contamination by astrocytes and oligodendrocytes was detected by using indirect immunofluorescence with antibodies against glial fibrillar acidic protein (polyclonal; Dako, Denmark) and CNPase (monoclonal; clone 11-5b, Sigma), respectively (data not shown).

Figure 1. A: Microglial cell cultures stained with the peroxidase-conjugated isolectin B4. Scale bar, 30 μm. B: Detection of type 1 and type 2 cannabinoid receptors mRNAs by RT-PCR in microglial cultures (+, cDNA; −, no cDNA). C: TNFα released from microglial cultures upon 2-h challenge with increasing concentrations of LPS. Cultured rat microglial cells were treated with HU210, CP55,940, (+)WIN 55212-2 (D), anandamide (Ana), or 2-arachidonylglycerol (2-AG; E) at the indicated concentration 15 min before LPS exposure (1 μg/ml). TNFα release, expressed as percent of LPS-treated cells, was assayed by ELISA in culture media 2 h following LPS stimulation. Values ± SE for 100% release of TNFα induced by LPS were 1,366 ± 204 pg/10⁵ cells. Data are the mean of at least five independent observations ± SE. Asterisk, P < 0.01 with respect to LPS-treated (Dunnett's test after ANOVA). None of the cannabinoids tested had any significant effect on the basal levels of TNFα release (74 ± 9.7 pg/10⁵ cells in untreated cultures). F: Effect of (+)WIN55212-2 (W) on expression of TNFα mRNA in microglia stimulated with LPS (L). Rat microglial cells were exposed to WIN55212-2 (10 μM) 15 min before stimulation with LPS (1 μg/ml for 1 h). Expression of TNFα and actin mRNA was determined by quantitative RT-PCR. The results show a representative experiment.

Isolectin B4 Staining

Twenty-four hours after plating, microglial cells were fixed with 4% paraformaldehyde (15 min at room temperature) before incubation in PBS containing 1% bovine serum albumin (BSA) for 1 h at room temperature. Subsequently, cells were incubated for 1 h at room temperature with biotin-conjugated isolectin B4 (Sigma, 5 μg/ml) from Bandeiraea Simplicifolia. After washes in PBS, cells were incubated with streptoavidin-peroxidase for 45 min and peroxidase activity was revealed by incubation with 3,3-diamnobenzidine tetrahydrochloride (0.7 mg/ml) and urea hydrogen peroxide (1.6 mg/ml) in 60 mM Tris buffer at pH 7.4. The preparations were mounted in Eukit (Kindler, Freiburg, Germany) and digital images were obtained using a digital camera (Olympus DP10) with an Olympus microscope (I × 50).

Experimental Conditions for Treatments

All cannabinoids, 2′,5′-dideoxyadenosine, SQ 22536, and forskolin were dissolved in dimethyl sulfoxide (DMSO). The final concentration of DMSO in the culture media never exceeded 0.2%, a concentration that did not affect LPS-evoked or basal TNFα release (data not shown). Cannabinoids were added to the culture medium 15 min prior to LPS challenge (2 h at 37°C). The antagonists of cannabinoid receptors (AM 251, SR141716A, and SR144528) and the demylate cyclase (AC) inhibitors (SQ 22536 or 2′,5′-dideoxyadenosine) were added to the culture medium 15 min prior to (+)WIN 55,212-2. Pertussis toxin was reconstituted in sterile PBS (0.1 mg/ml) and stored at 4°C for no longer than 2 weeks. Exposure to pertussis toxin was conducted overnight (16–18 h). Subsequently, the growth medium containing the toxin was removed and replaced with fresh medium before challenge with LPS and (+)WIN 55,212-2.

TNFα ELISA

After 2 h following LPS stimulation, supernatants were collected and stored at −20°C until assay. Medium samples were assayed for TNFα content by using Duo set enzyme-linked immunoassay (ELISA) kits specific for rat TNFα according to the manufacture's instructions (Bender, Vienna, Austria).

Reverse Transcription and PCR Analyses

Total RNA was isolated from microglial cell cultures by using Trizol (Gibco-BRL) according to the manufacturer's instruction. The intactness of total RNA used for the reverse transcription (RT) reaction was tested by electrophoresis on a 1.1% agarose gel. First strand cDNA was prepared by RT of 1 μg of total RNA by using oligodT to prime Moloney murine leukemia virus reverse transcriptase. The 25 μl reaction contained the following reagents obtained from Promega: 50 mM Tris (pH 8.3), 75 mM KCl, 3 mM MgCl, 1 mM dithiotreitol, 0.5 mM deoxynucleoside triphosphates, 5 μM random primers, and 25 U RNasin. The samples were incubated for 60 min at 42°C, and the reaction was then terminated by heating to 94°C for 5 min. The RT-PCR was performed on identical amounts of RNA for each sample. Primers used were designed from sequences in the Genebank database using Primer Express software (Perkin-Elmer). Primer sequences were the following: rat actin sense primer 5′-TCATGAAGTGTGACGTTGACATCCGT-3′ and antisense 5′-CCTAGAAGCATTTGCGGTGCACGATG-3′ (285 bp product); rat TNFα sense primer 5′-CAGACCCTCACACTCAGATCATCTT-3′ and antisense primer 5′-CAGAGCAATGACTCCAAAGTAGACCT-3′ (479 bp product); CB1 receptor sense primer 5′-CGTAAAGACAGCCCCAAT and antisense primer 5′-CTGGGTCCCAGCCTGAAT-3′ (398 bp product); CB2 receptor sense primer 5′-TTTCACGGTGTGGACTCC and antisense primer 5′-TAGGTAGGAGATCAACGC (212 bp product). For PCR amplification of specific cDNAs, the 50 μl reactions were prepared on ice and contained the following reagents: 0.1 mM each of dCTP, dGTP, dATP, and dTTP, 2.5 mM MgCl₂, 50 mM KCl, 10 mM Tris (pH 9.0), 0.1% Triton-X, 500 nM of each primer, 1 U of Taq polymerase, and 40 ng of the cDNA synthesized in the RT reaction. The number of cycles and reaction temperature conditions were optimized to provide a linear relationship between the amount of input template and the amount of PCR product. Amplification programs were the following: 22 cycles of 94°C for 45 s, 65°C for 1 min, and 68°C for 1 min for actin; 26 cycles of 94°C 1 min, 60°C for 1 min, and 72°C for 1 min for TNFα; 32 cycles of 94°C for 1 min, 64°C for 45 s, and 72°C for 1 min for CB1 receptor; 32 cycles of 94°C for 1 min, 58°C for 45 s, and 72°C for 1 min for CB2 receptor. PCR products were separated by electrophoresis through a 1.4% agarose gel and stained with ethidium bromide. An image of the gel was digitally captured by using PhotoCapt 99.01 software (Bioprofil). The identities of the amplicones were confirmed with restriction digests.

Statistical Analysis

Results are presented as the mean ± standard error mean (SEM) of at least five independent observations unless otherwise noted. By using Instat software (Graphpad), experimental groups with multiple treatments were analyzed by analysis of variance (ANOVA) followed by Dunnet's test or Bonferroni's test.

RESULTS

Cannabinoids Ablate LPS-Induced TNFα Release in Cultured Rat Microglial Cells

In cultures highly enriched in microglia, as assessed by isolectin B4 staining (Fig. 1A), we found that release of TNFα in response to 2-h exposure to LPS occurs in a concentration-dependent (0.1–10 μg) fashion (Fig. 1C). In addition, as shown in Figure 1B, cultured microglia express both CB1 and CB2 mRNA. We thus elected to use 2-h exposure of LPS at the dose of 1 μg/ml throughout the study and we used WIN 55,212-2, CP 55,940, and HU210 as prototypes to determine effects of cannabinoid receptor agonists on cytokine production in endotoxin-activated microglia. As shown in Figure 1D, both (+)WIN 55,212-2 and CP55,940, as well as HU210, exerted a concentration-dependent (1–10 μM) inhibitory effect on LPS-induced TNFα release by microglial cells. At 10 μM, all three synthetic cannabinoids ablated LPS-induced TNFα release. A similar effect was also observed with the putative endogenous cannabinoids, anandamide and 2-arachidonylglycerol (Fig. 1E). Furthermore, LPS induced significant upregulation of TNFα mRNA expression in microglial cells (Fig. 1F). While 10 μM (+)WIN 55,212-2 almost completely inhibited LPS-induced TNFα release, this same concentration only minimally affected LPS-induced TNFα mRNA upregulation (Fig. 1F).

WIN 55,212-2 Effects on LPS-Evoked Release of TNFα Are Stereoselective and Not Modified by CB Receptor Antagonists

We compared the effects of the two stereoisomers of WIN 55,212-2 on LPS-evoked release of TNFα. We observed that unlike (+)WIN 55,212-2, (−)WIN 55,212-2 did not display a significant inhibitory effect on LPS-evoked release of TNFα (Fig. 2). Given this stereoselectivity, the capability of (+)WIN 55,212-2 to inhibit LPS-induced TNFα release from microglia is most likely receptor-mediated.

Figure 2. Stereoselectivity of the effect of WIN 55212-2 on LPS-induced TNFα release. Microglial cells were exposed to the two stereoisomers (−)WIN 55212-2 and (+)WIN 55212-2 at the indicated concentration 15 min before LPS treatment (1 μg/ml for 2 h). Release of TNFα is expressed as percent of LPS-treated cells. Values ± SE for 100% release of TNFα induced by LPS were 1,143 ± 276 pg/10⁵ cells. Data are the mean ± SE of at least four independent observations. Asterisk, P < 0.01 with respect to LPS-treated (Dunnett's test after ANOVA).

To determine whether CB1 and/or CB2 receptors mediate the inhibitory effects of (+)WIN 55,212-2 on LPS-evoked TNFα release, we treated microglial cells challenged with LPS with (+)WIN 55,212-2 in the presence of the receptor-selective antagonists SR141716A and AM251 for CB1 receptors and SR144528 for CB2 receptors. Both SR141716A and AM251, as well as SR144528, even if used at an excess concentration of 10 μM, did not modify the inhibitory effects of (+)WIN 55,212-2 (Fig. 3) and HU210 (not shown) on TNFα release.

Figure 3. Cannabinoid receptor antagonists do not reverse the inhibitory effects of (+)WIN 55212-2 on TNFα release. Microglial cells were exposed to LPS (1 μg/ml for 2 h) and (+)WIN55212-2 in the presence or absence of 10 μM receptor-selective antagonist SR 1417164A (SR1; A), AM251 (B), or SR 144528 (SR2; C). The effect of (+)WIN55212-2 on LPS-induced TNFα release was not significantly affected by the presence of the CB receptor antagonists (Bonferroni's test after ANOVA). Release of TNFα was expressed as percent of LPS-treated cells. Values ± SE for 100% release of TNFα induced by LPS are 1,796 ± 208 pg/10⁵ cells (A and C) and 1,033 ± 112 pg/10⁵ cells (B). Bars represent the mean ± SE of at least five independent observations.

(+)WIN 55,212-2 Inhibitory Effects Are Not Sensitive to Pertussis Toxin

Previous studies have demonstrated the functional coupling of CB1 and CB2 receptor to the inhibition of adenylate cyclase via pertussis toxin-sensitive G_i-proteins (Bayewitch et al., 1995; Kaminski, 1998). Accordingly, overnight pretreatment with petussis toxin has been shown to inhibit both G_i/o protein function coupled with CB receptors and CB1-mediated effects on LPS-induced nitric oxide production in cultured microglial cells (Sagan et al., 1999; Waksam et al., 1999). We thus evaluated whether pertussis toxin could prevent the inhibitory effects of (+)WIN 55,212-2 on LPS-evoked TNFα release. As shown in Figure 4, 18-h pretreatment with pertussis toxin (100 ng/ml) did not prevent the ablative effects on LPS-evoked TNFα release of (+)WIN 55,212-2.

Figure 4. Pertussis toxin does not affect the inhibitory effects of (+)WIN 55212-2. Microglial cells were incubated with pertussis toxin (ptx; 100 ng/ml) for 18 h before being challenged with LPS in the presence or absence of (+)WIN 55212-2. Release of TNFα was expressed as percent of LPS-treated cells. Values ± SE for 100% release of TNFα induced by LPS were 1,138 ± 170 pg/10⁵ cells. Bars represent the mean ± SE of at least five independent observations.

WIN 55,212-2 Inhibition of TNFα Release Induced by LPS Does Not Occur Through Adenylate Cyclase

Both forskolin and dibutyryl-cAMP, which elevate intracellular cAMP by independent mechanisms, ablated LPS-evoked TNFα release (Fig. 5A). Thus, the inhibitory effect of (+)WIN 55,212-2 on TNFα release is unlikely the result of the negative modulation of cAMP levels as it would occur through the stimulation of the G_i/o-coupled receptors CB1 and CB2. However, as CB1 receptor stimulation can cause intracellular cAMP accumulation in certain experimental conditions (Glass and Felder, 1997; Calandra et al., 1999), we questioned whether adenylate cyclase (AC) function is required to mediate (+)WIN 55,212-2 inhibitory effects on TNFα release. To this end, we employed the specific cell-permeable AC inhibitors SQ 22536 (300 μM) and 2′,5′-dideoxyadenosine (10 μM) (Haslam et al., 1978). As shown in Figure 5B, neither molecules modified the effect of (+)WIN 55,212-2 on TNFα release in microglia cultures activated by LPS.

Figure 5. A: Effects of cAMP-elevating agents on LPS-induced TNFα release. Microglial cells were exposed to cAMP-elevating agents forskolin (fsk; 20 μM) or dibutyryl cAMP (BtAMP; 2 mM) 15 min before LPS. Both forskolin and dibutyryl cAMP ablated LPS-evoked TNFα release. Values ± SE for 100% release of TNFα induced by LPS are 1,144 ± 276 pg/10⁵ cells. The data are the mean ± SE of at least three independent observations. Asterisk, P < 0.01 with respect to LPS treated. B: Effects of inhibitors of adenylate cyclase on LPS-evoked release of TNFα. Rat brain cultured microglia were exposed with SQ 22536 (150 μM) or 2′-5′-dideoxyadenosine (DDA, 20μM) 15 min before stimulation with LPS (1 μg/ml for 2 h). The effect of (+)WIN 55212-2 on TNFα release was not significantly affected by the presence of either SQ 22536 or 2′-5′-dideoxyadenosine (Bonferroni's test after ANOVA). Values ± SE for 100% release of TNFα induced by LPS are 1,102 ± 166 pg/10⁵ cells. Bars represent the mean ± SE of at least five independent observations.

DISCUSSION

Several evidences indicate that exogenous and endogenous cannabinoids modulate the function of immune cells in experimental in vivo and in vitro models (Kaminski, 1998; De Petrocellis et al., 2000). In particular, either synthetic and endogenous cannabinoids have been shown to modulate TNFα mRNA expression in cultured microglial cells stimulated with LPS (Puffenbarger et al., 2000). In this study, we report that both endogenous (anadamide and 2-AG) as well as synthetic cannabinoids (CP55,940, HU-210, and WIN 55,212-2) have the ability to ablate the release of TNFα elicited by LPS in cultured rat microglial cells. The inhibitory effect of WIN 55,212-2 on TNFα release can only in part be explained as an effect on TNFα mRNA expression as we show that the upregulation of TNFα mRNA levels induced by LPS in microglia is only minimally counteracted by (+)WIN 55,212-2, a finding nevertheless in line with previous observations (Puffenbarger et al., 2000).

We addressed the issue of the stereoselectivity of the cannabinoid effects by comparing the enantiomeric pairs (−)WIN 55,212-2 and (+)WIN 55,212-2. Unlike the high-affinity cannabinoid receptor agonist (+)WIN 55,212-2, the low-affinity stereoisomer (−)WIN 55,212-2 did not exert any significant inhibitory effects on TNFα release. These data, by showing an enantiomeric stereoselectivity, are suggestive of a receptor-mediated effect of (+)WIN 55,212-2 on TNFα release. Cannabinoids exert many of their effects through activation of two distinct G_i/o protein-coupled receptors, namely CB1 and CB2 (Matsuda et al., 1990; Munro et al., 1993). In highly purified neonatal rat microglial cells, CB1 receptor mRNA and protein are expressed (Sinha et al., 1998) and functional, as its stimulation exerts inhibitory effects on endotoxin-stimulated NO production (Waksman et al., 1999). By RT-PCR we confirmed that CB1 mRNA is expressed in our microglia preparations; in addition, we also detected CB2 mRNA. To assess whether the inhibitory effect of (+)WIN 55,212-2 is exerted through CB1 and/or CB2 receptors, we challenged microglia cultures with (+)WIN 55,212-2 in the presence of SR141716A or its analog AM251, both selective CB1 antagonist, or SR144528, a selective CB2 antagonist (Rinaldi-Carmona et al., 1995; Gatley et al., 1996). Neither SR141716A and AM251 nor SR144528 modified the inhibitory effect of (+)WIN 55,212-2 on TNFα release evoked by LPS. In addition, we observed that (+)WIN 55,212-2 ablative effects on TNFα release were unaffected by pretreatment with the G_i/o proteins inhibitor pertussis toxin. Taken together, these observations further indicate that the inhibitory effect of (+)WIN 55,212-2 unlikely involves CB1 and/or CB2 receptors, which are known to be functionally coupled with G_i/o proteins.

Since we observed that cAMP-elevating agents have an ablative effect on LPS-induced TNFα release, the inhibitory effect of (+)WIN 55,212-2 on TNFα release is unlikely mediated through a negative modulation of cAMP levels as it would occur through a CB1 receptor- and/or CB2 receptor-mediated mechanism. However, as CB1 receptor stimulation can augment cAMP accumulation in certain experimental conditions (Glass and Felder, 1997), we tested the possibility that (+)WIN 55,212-2 could inhibit TNFα release by increasing intracellular cAMP via adenylate cyclase stimulation. Pretreatment of microglial cultures with the adenylate cyclase-specific inhibitors SQ 22536 and 2′,5′-dideoxyadenosine did not modify the inhibitory effects of (+)WIN 55,212-2, thus excluding a stimulatory effect of (+)WIN 55,212-2 on adenylate cyclase activity. Also noteworthy is that although cannabinoids have been reported to increase CB1 receptor-mediated sphingomyelin turnover, this effect is antagonized by SR141716A, thus potentially excluding influences of this pathway on the effects here observed (Sanchez et al., 1999).

In summary, our data show that both synthetic and endogenous cannabinoids have the ability to ablate LPS-induced release of TNFα in microglial cells. While the stereoselectivity of the inhibitory effect of (+)WIN 55,212-2 on TNFα release is suggestive of a receptor-mediated effect, the fact that such an effect is not pertussis toxin-sensitive or counteracted by specific antagonists of the CB1 and CB2 receptors implies the involvement of receptor(s) yet to be identified. This is not unprecedented as, for instance, cannabinoids have been shown to stimulate signal transduction pathways and modulate cytokine mRNA expression through mechanisms not involving CB1 or CB2 receptors (Felder et al., 1992; Puffenbarger et al., 2000; Breivogel et al., 2001).

Definition of the mechanisms underlying the ablative effect of cannabinoids on the production of proinflammatory mediators, such as TNFα, in CNS cells may well provide insights relevant to the application of selected cannabinoid-like molecules as therapeutic agents in neurological conditions associated with inflammatory processes such as multiple sclerosis, stroke, and Alzheimer's disease.

Acknowledgements

The authors thank M. Fabris for technical support and E. Polazzi for advice on culturing techniques.