SEARCH

SEARCH BY CITATION

Keywords:

  • astrocytes;
  • cytokines;
  • methanandamide;
  • microglia;
  • neuroinflammation;
  • reactive gliosis

Abstract

  1. Top of page
  2. Abstract
  3. Materials and methods
  4. Results
  5. Discussion
  6. Acknowledgements
  7. References

Brain injuries as well as neurodegenerative diseases, are associated with neuro-inflammation characterized by astroglial and microglial activation and/or proliferation. Recently, we reported that lipopolysaccharide (LPS)-activation of microglia inhibits junctional channels and promotes hemichannels, two connexin43 functions in astrocytes. This opposite regulation is mediated by two pro-inflammatory cytokines, interleukin-1 beta and tumor necrosis factor-alpha, released from activated microglia. Because cannabinoids (CBs) have anti-inflammatory properties and their receptors are expressed by glial cells, we investigated on primary cortical cultures the effects of CB agonists, methanandamide and synthetic CBs on (i) cytokines released from LPS-activated microglia and (ii) connexin43 functions in astrocytes subjected to pro-inflammatory treatments. We observed that CBs inhibited the LPS-induced release of interleukin-1 beta and tumor necrosis factor-alpha from microglia. Moreover, the connexin43 dual regulation evoked by the pro-inflammatory treatments, was prevented by CB treatments. Pharmacological characterizations of CB actions on astrocytic connexin43 channels revealed that these effects were mainly mediated through CB1 receptors activation, although non-CB1/CB2 receptors seemed to mediate the action of the methanandamide. Altogether these data demonstrate that in inflammatory situations CBs exert, through the activation of different sub-types of glial CB receptors, a regulation on two functions of connexin43 channels in astrocytes known to be involved in neuron survival.

Abbreviations used:
CB(s)

cannabinoid(s)

CM

conditioned medium

CP

CP-55,940

Cx43

connexin43

DIV

days in vitro

DMEM

Dulbecco’s Modified Eagle Medium

Etd

ethidium

FCS

fetal calf serum

GFAP

glial fibrillary acid protein

GJC

gap junction channel

HC

hemichannel

IL-1β

interleukin-1 beta

LPS

lipopolysaccharide

Meth

Methanandamide

MG

microglia

Mix

mixture of cytokines

Panx1

Pannexin1

PBS

phosphate buffer saline

SR1

CB1 receptor antagonist SR-141716A

SR2

CB2 receptor antagonist SR-144528

TNF-α

tumor necrosis factor-alpha

WIN

WIN 55,212-2

Most neurodegenerative diseases and brain injuries are associated with neuro-inflammation (Block and Hong 2005). This process also implies a reactive gliosis, supported by the activation and/or the proliferation of both astrocytes and microglia (MG) (Ridet et al. 1997; Streit et al. 1999; Sofroniew 2005). An important characteristic of astrocytes is the high level of connexin expression, in particular connexin43 (Cx43) which is the only connexin present in primary cultures of astrocytes (Giaume and McCarthy 1996; Koulakoff et al. 2008). This plasma membrane protein is involved in two modes of cell-to-cell communication: (i) a direct intercellular communication through gap junction channels (GJCs) (Rouach et al. 2002) and (ii) an autocrine and/or paracrine communication mediated by Cx43 hemichannels (HCs) allowing exchanges between the extracellular and intracellular compartments (Spray et al. 2006).

Recently, using co-cultures of astrocytes and MG we have reported that MG activated by lipopolysaccharide (LPS) down-regulates astrocytic Cx43 expression and induces an opposite regulation of the two communicating functions of Cx43: inhibition of Cx43 GJCs and activation of Cx43 HCs (Retamal et al. 2007). This effect is mediated by at least two pro-inflammatory cytokines released from activated MG, interleukin-1 beta (IL-1β) and tumor necrosis factor-alpha (TNF-α) (Même et al. 2006). Indeed, a dual regulation of Cx43-based channels can be produced by treating pure astrocyte cultures with either the mixture of these two cytokines (Mix) or with the conditioned medium (CM) harvest from LPS-activated MG. Importantly, this effect is blocked by the combination of an antagonist of the IL-1β receptor and the soluble receptor for TNFα (Même et al. 2006).

As both Cx43 GJCs and HCs might play a role in neurodegenerative processes (Nakase and Naus 2004; Orellana et al. 2009), an important issue is to identify compounds that could interact with Cx43 functions by preventing the opposite regulation of GJCs and HCs triggered by pro-inflammatory agents. Appropriate candidates for this purpose are cannabinoids (CBs) as they have a potential therapeutic application through an anti-inflammatory action (Croxford 2003; Cabral and Griffin-Thomas 2008) and because astrocytes and MG are one of their targets in the central nervous system (Stella 2004). Indeed an increasing number of studies revealed the anti-inflammatory properties of CBs are exerted both in periphery and in CNS (for reviews see Klein et al. 2003; Klein 2005; Pacher et al. 2006; Walter and Stella 2004; Zurier 2003). For example, an active plant derived CB compound, Δ9-tetrahydrocannabinol, can significantly reduce the symptoms generated in an animal model of multiple sclerosis (Lyman et al. 1989), through modulation of the neuroinflammation (Baker et al. 2000;Baker and Pryce 2008). Indeed, CBs act by blocking the pro-inflammatory products released from activated MG (Coffey et al. 1996; Waksman et al. 1999) switching their phenotype toward an anti-inflammatory phenotype (Stella 2009). Moreover, in vitro studies have shown that the synthetic CB, WIN 55,212-2 (WIN), reduces (i) the release of TNF-α from LPS-activated MG (Facchinetti et al. 2003) and (ii) the release of IL-1β from the stimulated human astrocytes (Sheng et al. 2005). Glial cells express CBs receptors in both healthy and pathological CNS. Thus, it was shown that MG express both CB1 and CB2 receptors (Waksman et al. 1999; Facchinetti et al. 2003) although CB2 receptors are primarily expressed by activated MG in a variety of inflammatory situations (Stella 2004, 2009), like in Alzheimer’s disease (Benito et al. 2003). In contrast, CB1 receptors are expressed by astrocytes (Molina-Holgado et al. 2002; Salio et al. 2002), whereas CB2 receptors have not been found in these glial cells (Walter and Stella 2003). Accordingly, we have used methanandamide (Meth), a non-hydrolysable analogue of the endogenous CB, anandamide, as well as two synthetic CBs: WIN and CP-55,940 (CP), to assess their effects on the opposite regulation exerted by either activated MG co-cultured with astrocytes or pro-inflammatory cytokines on Cx43 GJC and HC activities in primary cultures of astrocytes. Here, we report that CBs prevent GJC and HC opposite regulation by acting either at the level of MG or astrocytes. In addition, we pharmacologically characterized these CB actions on astrocytes using the available antagonists of CB1 or CB2 receptors. These findings might help identify new therapeutic targets in glial cells to counteract deleterious aspects induced by reactive gliosis in neuroglial interaction.

Materials and methods

  1. Top of page
  2. Abstract
  3. Materials and methods
  4. Results
  5. Discussion
  6. Acknowledgements
  7. References

Animals

All cell cultures were performed from OF1 mice (Charles River, L’Arbresle, France). Experiments have been carried out in accordance with the European Community Council Directives of November 24th 1986 (86/609/EEC) and all efforts have been made to minimize the number of animals used and their suffering.

Cell cultures

Primary astrocyte and mixed cultures

Primary astrocyte cultures and mixed cultures were prepared from cortex of newborn (1–2 days) mice as previously described (Même et al. 2006). Briefly, cells were grown to confluence in Dulbecco’s Modified Eagle Medium (DMEM; Sigma-Aldrich, St Quentin Fallavier, France), supplemented with penicillin (5 U/mL), streptomycin (5 μg/mL) (Invitrogen, Carlsbad, CA, USA) and 10% fetal calf serum (FCS; Hyclone, Logan, UT, USA). Between 6 and 8 days in vitro (DIV), cytosine/arabinoside (5 μM; Sigma-Aldrich) was added to the culture medium for 48 h. The medium was changed twice a week and cells were used between 21 and 28 DIV.

Mixed cultures of astrocytes and MG were obtained using the same protocol, but the cytosine/arabinoside treatment was omitted, which allows for the proliferation of MG on the astrocyte monolayer. The cells were used at the same DIV (21–28 DIV) (see Même et al. 2006).

MG cultures and conditioned medium production

To prepare MG in cultures, cerebral hemispheres were dissected from newborn mice (P1) after removal of meninges. After dissociation, cells were seeded into 100 mm in diameter plastic culture dishes (Nunc, Roskilde, Denmark) at the density of 3 × 106cells/dish in DMEM, containing 10% heat-inactivated FCS (Abcys, Paris, France), as previously described (Calvo et al. 2000). Medium was changed at days 1 and 3, and cells were collected at day 10 by shaking culture dishes to detach cells adhering to the astrocyte monolayer. The collected population resulted in more than 98% of cells bearing the Mac-1 antigen, a specific marker of mononuclear cells (Calvo et al. 2000). Freshly collected MG were used to make CM harvested from activated MG. They were seeded in DMEM containing 5% FCS (1.7 × 106 cells/mL/dish in 35 mm dishes) and treated with LPS (10 ng/mL) for 6 h. The resulting supernatants, which constitute the conditioned medium from activated MG, were collected, filtered (0.22 μm) and stored at −20°C before their use.

Cell treatments

Mixed cultures and astrocyte cultures were subjected to different pro-inflammatory treatments, applied for 24 h into the culture medium. Mixed cultures were treated with LPS (10 ng/mL), whereas astrocyte cultures were treated with either CM (diluted ¼) from LPS-activated MG or the Mix (IL-1β plus TNF-α, 10 ng/mL for each). CB agonists (i.e. methanandamine, WIN 55,212-2, the inactive enantiomer S-(−)-WIN 55,212-2 and CP-55,940), were pre-incubated 1 h prior to each pro-inflammatory treatment (previously described above), and they were kept together in the culture medium for 24 h. Antagonists for the CB1 receptor (SR-141716A) and CB2 receptors (SR-144528) were co-incubated with CB agonists (1 h prior to the different pro-inflammatory treatments) and were also kept into the culture medium for 24 h.

For the treatments of MG cultures, cells were incubated with LPS (10 ng/mL) for 6 h. CB agonists and antagonists were pre-incubated 1 h prior to the LPS and were still present in the medium for 6 h.

Detection of IL-1β and TNF-α production by ELISA

ELISA cytokine titrations were performed with the different conditioned media harvested from MG cultures. Amounts of IL-1β and TNF-α were determined in supernatants by means of specific mouse cytokine ELISA kits (BD Bioscience, Bedford, MA, USA). The colorimetric reaction was obtained after incubation with tetramethylbenzidine substrate reagent (BD Bioscience). Doubling dilutions of recombinant mouse IL-1β or TNF-α ranging from 5 to 2000 pg/mL were used as standards. Optical density was then measured at 405 nm with Dynatech MR5000 Reader (Long Island Scientific, East Setauket, NY, USA).

Scrape loading/dye transfer technique

Experiments were performed on primary astrocytes or mixed cultures, grown into 35 mm in diameter plastic culture dishes (Nunc), as previously described (Même et al. 2006). Briefly, cells were firstly incubated, at 20–22°C for 10 min, in HEPES buffered salt solution containing (in mM): NaCl, 140; KCl, 5.5; CaCl2, 1.8; MgCl2, 1; glucose, 10; HEPES, 10 at pH 7.35. Then cells were washed in Ca2+-free HEPES solution for 1 min and scrape loading/dye transfer (SL/DT) was achieved in the same Ca2+-free solution containing 1 mg/mL Lucifer yellow (LY, Sigma-Aldrich). After 1 min, they were washed with the HEPES solution and LY loaded in the cells was allowed to diffuse through GJCs during 8 min. In all experiments, dye coupling through GJC was assessed 9 min after scraping by taking five successive fluorescent images captured using an inverted microscope equipped for epifluorescence (Diaphot-Nikon, Tokyo, Japan) and a camera (DXM1200) connected to an image analysis system equipped with a software (Lucia-Nikon, Tokyo, Japan). Dye diffusion was then quantified by measuring fluorescence areas, expressed in pixel/cm2. In all experiments, the fluorescence of the first raw of cells initially loaded during the scrape was measured in the presence of the gap junction inhibitor carbenoxolone (30 μM), and was subtracted from the total fluorescence area.

Ethidium bromide uptake

For dye uptake experiments, astrocytes cultured on coverslips (14 mm in diameter; Gassalem, Limeil-Brévannes, France), were exposed to either 0.5 μM Ethidium Bromide (Molecular Probes, Eugene, OR, USA) for 10 min at 37°C. Then, cells were washed with Hank’s balanced salt solution in mM: NaCl: 137; KCl: 5.4; Na2HPO4: 0.34; KH2PO4: 0.44, at pH 7.4 supplemented with 1.2 mM CaCl2. To visualize ethidium (Etd) uptake, coverslips were mounted in Fluoromount (Southern Biotech, Birmingham, AL, USA) and examined by epifluorescence (518 nm excitation and 605 nm emission) using an inverted microscope (Daiphot-Nikon) equipped with a CCD camera (Nikon) associated with image analyzer software (Lucia-Nikon). Captured images of Etd uptake were analyzed by counting the number of Etd positive-cells per field, using Image J program (NIH software; Scion Corp., Frederik, MA, USA). For each experimental condition, ten microscopic fields per coverslip were taken arbitrary and averaged. Data were expressed as the number of positive-Etd cells per field.

Western blotting

Cells grown into a 35 mm in diameter plastic culture dishes, were collected by scraping with a rubber policeman in a small volume of phosphate buffer saline (PBS) containing orthovanadate (1 mM), α-glycerophosphate (10 mM), and complete mini protease inhibitor (Roche Diagnostics, Basel, Switzerland). They were then boiled 5 min with Laemmli medium before sonication. Protein concentration was determined with the Bradford method, using bovine serum albumin as a standard. Proteins were separated by electrophoresis on 10% polyacrylamide gels and transferred onto nitrocellulose. Membranes were saturated with 5% fat-free dried milk in triphosphate buffer solution and exposed overnight to mouse anti-Cx43 antibody (1 : 250, BD Bioscience) at 4°C. They were then washed and exposed to peroxydase-conjugated goat anti-mouse IgG (1 : 2500, Santa Cruz Biotech, Santa Cruz, CA, USA). Immunoreactive bands were visualized with the chemiluminescence detection kit (ECL; Amersham Pharmacia Biotech, Piscataway, NJ, USA). Semi-quantitative densitometry analysis was performed after scanning the bands, with image-analysis software (NIH Image).

Immunofluorescence and confocal microscopy

For all immunostaining experiments, cells grown on coverslips (14 mm in diameter; Gassalem) were rinsed with PBS, fixed at 20–22°C with 2% paraformaldehyde for 30 min, and washed two times with PBS. They were incubated in PBS-glycine (0.1 M), three times for 5 min each and then in PBS-0.1% Triton-X100, containing 10% normal goat serum (Zymed Laboratories Inc., San Francisco, CA, USA), for 30 min.

To perform co-labeling of glial fibrillary acid protein (GFAP) and Cx43, cells were incubated for 2 h at 20–22°C with mouse anti-GFAP (1 : 500; ICN Chemicals, Irvine, CA, USA) and rabbit anti-Cx43 antibodies (1 : 500; Zymed Laboratories Inc.) diluted in PBS-0.1% Triton-X100 with 2% normal goat serum. After three rinses in PBS-0.1% Triton-X100, cells were then incubated with goat anti-mouse IgG conjugated to Alexa Fluor 488 (1 : 1500; Molecular Probes) and goat anti-rabbit IgG conjugated to Alexa Fluor 555 (1 : 1500; Molecular Probes), diluted in the same solution, in order to reveal GFAP (in green) and Cx43 (in red), respectively. After several washes in PBS, coverslips were mounted with Fluoromount (Southern Biotech) and examined by epifluorescence. Finally, cells were examined with a confocal laser-scanning microscope (Leica TBCS SP2, Wetzlar, Germany) with a ×63 objective. Stacks of consecutive confocal images taken at 500-nm intervals were acquired sequentially with two lasers (argon 488 nm and helium/neon 543 nm) and Z projections were reconstructed using Leica Confocal Software.

Quantitative real-time PCR

Total RNA was extracted from astrocyte cultures using the ‘RNeasy lipid tissue kit’ (Qiagen, Courtaboeuf, France). Reverse transcription was performed for 1 μg of RNA. Q-PCR was conducted using the following primers: Panx1f 5′-CCTGCAGAGCGAGTCTGGAA-3′ and Panx1r 5′-TGCGGGCAGGTACAGGAGTA-3′ (working concentration 600 nM); Cx43f 5′-GAGATGCACCTGAAGCAGATTGAA-3′ and Cx43r 5′-GATATTCAGAGCGAGAGACACCAA-3′ (working concentration 300 nM); Hprt.f 5′-GTTGGATACAGGCCAGACTTTGTTG-3′ and Hprt.r 5′-GATTCAACTTGCGCTCATCTTAGGC-3′ (working concentration 300 nM), using SYBR Green PCR master kit (Applied Biosystems, Foster City, CA, USA). PCR cycling conditions were 50°C for 2 min, 95°C 10 min, and 40 cycles of 95°C for 15 s and 60°C for 1 min. Experiments were performed in triplicate on an LC480 Roche Light cycler. The relative abundance of amplified Cx43 or Panx1 cDNA was calculated as 2−ΔCt, where ΔCt (change in cycle threshold) equals Ct in treated cells minus Ct in control cells. Results are expressed as means of 2−ΔCtPx1 or 2−ΔCtCx43/2−ΔCtHprt values. Experiments were performed three times in independent cell cultures.

Statistical analysis

For each group, data are expressed as mean ± SEM and n refers to the number of independent experiments. For all experiments, each treatment was compared with its respective control using a one-way anova followed in case of significance by a Bonferroni post hoc test to compare the mean values. All statistical tests were performed on raw data and using Graphpad Prism software (San Diego, CA, USA). Differences were considered significant at *< 0.05, **< 0.01 and ***< 0.001.

Chemicals

The two synthetic CB agonists, WIN-55,212-2 {R-(+)-(2,3-dihydro-5-methyl-3-[(4-morpholinyl}methyl]pyrol [1,2,3-de]-1,4-benzoxazin-6-yl)(l-naphthalenyl) methanonemesylate} and CP-55,940 {[1α,2-(R)-5-(1,1-dimethylheptyl)-2-[5-hydroxy-2-(3-hydroxypropyl)cyclohexyl]-phenol} were from Tocris (Avonmouth, UK). Meth, the inactive enantioner: S-(−)-WIN 55,212-2, lypopolysaccharide (LPS), and Lucifer yellow (LY) were from Sigma-Aldrich (St Quentin Fallavier, France). The two cytokines IL-1β and TNF-α were from Roche Diagnostics (Meylan, France). The CB1 receptor antagonist: SR-141716A {[N-(piperidin-1-yl)-5-(4-chlorophenyl)-1-(2,4-dichlorophenyl)-4-methyl-1 H-pyrazole-3-carboxamidehydrochloride]}, and the CB2 receptor antagonist: SR-144528 {N-[(1S)-endo-1,3,3-trimethyl bicycle [2.2.1] heptan-2-yl]-5-(4-chloro-3-methylphenyl)-1-(4-methylbenzyl)-pyrazole-3-carboxamide}, were provided by Sanofi-Aventis Recherche (Bagneux, France).

Results

  1. Top of page
  2. Abstract
  3. Materials and methods
  4. Results
  5. Discussion
  6. Acknowledgements
  7. References

Cannabinoids prevent the inhibition of astrocyte gap junction channel activity induced by LPS-activated microglia

The effects of Meth, a non-hydrolysable analogue of anandamide, and two synthetic CBs: WIN and CP were tested on astrocyte GJC activity using astrocyte-MG co-cultures (mixed cultures). Untreated mixed cultures (control) presented a high level of gap junctional communication (see Fig. 1a-i). As previously described (Même et al. 2006), activation of MG by 10 ng/mL of LPS applied on mixed cultures, for 24 h, induced a strong decrease of astrocyte GJC permeability for LY (n = 16 and 17 for control and LPS, respectively, < 0.001; Fig. 1a-i,ii and b). By contrast, 24 h co-incubation with LPS (10 ng/mL) plus (i) Meth (10 μM) (ii) WIN (1 μM) or (iii) CP (1 μM), caused a preventive effect on GJC inhibition. Indeed, instead of the 59% decrease induced by LPS application alone, the inhibition of dye coupling reached only (i) 23% with addition of Meth (n = 6, < 0.05 compared to LPS alone; Fig. 1a-iii and b), (ii) 30% in the presence of WIN (n = 6, < 0.05 compared to LPS alone) and (iii) 31% in presence of CP (n = 5, < 0.01 compared to LPS alone) (Fig. 1b). Interestingly, when Meth and Win were applied together on LPS treated mixed cultures, the preventive effect on the LPS-induced GJC inhibition, previously described for each compound, was significantly increased as GJC activity reached to the same level as observed in untreated astrocytes (n = 5, < 0.01 compared to LPS + WIN; Fig. 1b), indicating an additive effect of the endogenous and synthetic CBs. Co-incubation of Meth plus CP on LPS treated astrocytes induced a similar additive effect (n = 4, < 0.01 compared to LPS + CP; Fig. 1b).

image

Figure 1.  Preventive effect of cannabinoids agonists on the LPS-activated microglia-induced GJC inhibition in astrocytes. (a) Representative images depicting the LY diffusion through astrocytic gap junction channels (GJC) in mixed glial cultures. In untreated condition (Control; a-i), mixed cultures exhibited a high level of GJC, which was strongly inhibited by the addition of bacterial lipopolysaccharide (LPS, 10 ng/mL for 24 h; a-ii). The co-incubation with LPS and either Meth (10 μM; a-iii) or the synthetic cannabinoid, WIN (1 μM; a-iv) partially prevented the LPS-induced decrease in GJC. Scale bar = 100 μm. (b) Graph showing the effect of either Meth, or the two synthetic cannabinoids, WIN and CP, on the LPS-induced inhibition of dye coupling in mixed cultures of MG and astrocytes. GJC in astrocytes was evaluated by quantification of LY diffusion from the scrape line, each bar represents the fluorescence area in untreated mixed cultures (C for control, white bar), or treated (i) with LPS alone (10 ng/mL; black bar), (ii) with LPS plus Meth (10 μM), WIN (1 μM) or CP (1 μM) (grey bars) and finally (iii) with LPS plus both Meth and WIN or plus both Meth and CP (hatched bars). Note that the effects of Meth plus each synthetic cannabinoid are additive. Data, expressed as pixels per cm2, are means ± SEM obtained from at least four independent experiments. ***< 0.001, **< 0.01 and *< 0.05 as compared to the indicated groups (one way anova followed by the Bonferroni post hoc test).

Download figure to PowerPoint

Cannabinoids reduced the production of the two pro-inflammatory cytokines, IL-1β and TNF-α, released by LPS-activated microglia

It was previously shown that the activation of MG by LPS increases the release of two pro-inflammatory cytokines, IL-1β and TNF-α, which are responsible for the inhibition of GJC permeability in astrocytes (Même et al. 2006; Retamal et al. 2007). Moreover, endogenous and synthetic CBs activate cannabinoid receptors in MG (Stella 2004). Accordingly, in pure MG cultures, the effects of Meth and synthetic CBs (WIN or CP) were assessed on the production of these two cytokines induced by LPS in pure MG cultures.

Effect of methanandamide on cytokine production

As previously reported (Même et al. 2006), treatment of pure MG cultures with LPS (10 ng/mL, 6 h) induced a basal level of secreted IL-1β and TNF-α (2.7 ± 0.4 ng/mL and 20.0 ± 3.2 ng/mL, Control; Table 1), detected by ELISA measurement from the culture medium. The incubation of Meth (10 μM) significantly reduced the production of both IL-1β (−75%) and TNF-α (−58%) by the LPS-stimulated MG (< 0.05, n = 10 and 11, respectively), while 1 μM Meth did not significantly modify the level of secretion of these two cytokines (Table 1). The pharmacological action of Meth which leads to inhibition of both IL-1β and TNF-α was tested using CB1 and CB2 receptors antagonists (SR1 and SR2, respectively). The addition of SR1 in LPS-stimulated MG partially prevented, without statistical significance, the inhibition of IL-1β and TNF-α secretion induced by Meth (SR1 + Meth; Table 1). Moreover, addition of SR2 in LPS-stimulated MG did not modify the inhibition of IL-1β and TNF-α secretion induced by Meth (SR2 + Meth; Table 1). These observations are in agreement with a previous report obtained at the level of IL-1β mRNA expression (Puffenbarger et al. 2000) which showed that the inhibition of IL-1β mRNA expression by anandamide was mediated neither through CB1 receptor nor through CB2 receptor.

Table 1.   Inhibitory effect of the endogenous cannabinoid, methanandamide, on IL-1β and TNF-α production by LPS-activated microglia
 Endogenous CB
ControlMeth (1 μM)Meth (10 μM)Meth (10 μM) + SR1 (1 μM)Meth (10 μM) + SR2 (1 μM)
  1. Quantifications of the levels of the two cytokines were performed using the ELISA technique. The IL-1β and TNF-α levels, measured from medium of cultured MG activated by LPS (10 ng/mL for 6 h), were taken as control. Data, expressed as ng/mL, were mean ± SEM from n independent measures. ***< 0.001, **< 0.01 and *< 0.05 as compared to control (one way anova followed by the Bonferroni post hoc test).

  2. Meth, methanandamide; SR1, CB1 receptor antagonist; SR2, CB2 receptor antagonist.

IL-1β (ng/mL)2.7 ± 0.4 (n = 15)3.1 ± 0.7ns (n = 4)0.7 ± 0.2*** (n = 11)1.6 ± 0.5ns (n = 11)1.0 ± 0.2** (n = 6)
TNF-α (ng/mL)20.0 ± 3.2 (n = 12)12.1 ± 3.6ns (= 4)8.4 ± 1.4* (= 10)13.7 ± 3.7ns (= 8)6.2 ± 1.7* (= 5)
Effect of synthetic CBs on cytokine production

The incubation of the synthetic CB agonist WIN (10 μM, 6 h) in LPS stimulated MG also caused a significant inhibition of production by 62% and 66% (< 0.05 and < 0.01 respectively, Table 2). In contrast, WIN used at 1 μM had no significant effect on either IL-1β or TNF-α IL production. Moreover, SR1 and SR2 were tested on the WIN effect on cytokine production and both antagonists were found to be uneffective (Table 2). Finally, addition of CP (10 μM) did not significantly reduce the production of these cytokines. Altogether, these data are in agreement with previous studies performed in rat cortical MG for TNF-α release (Facchinetti et al. 2003) and at mRNA level (Puffenbarger et al. 2000).

Table 2.   Inhibitory effect of the synthetic cannabinoid, WIN 55,212-2, on IL-1β and TNF-α production from LPS-activated microglia
 Synthetic CB
ControlWIN (1 μM)WIN (10 μM)WIN (10 μM) + SR1 (1 μM)WIN (10 μM) + SR2 (1 μM)
  1. Quantifications of the levels of the two cytokines were performed using the ELISA technique. The IL-1β and TNF-α levels, measured from medium of cultured MG activated by LPS (10 ng/mL for 6 h), were taken as control. Data, expressed as ng/mL, were mean ± SEM from n independent measures. ***< 0.001, **< 0.01 and *< 0.05 as compared to the control group (one way anova followed by the Bonferroni post hoc test).

  2. Meth, methanandamide; SR1, CB1 receptor antagonist; SR2, CB2 receptor antagonist.

IL-1β (ng/mL)3.4 ± 0.4 (= 24)2.2 ± 0.7ns (= 11)1.3 ± 0.5* (= 10)1.2 ± 0.4* (= 10)1.1 ± 0.8* (= 6)
TNF-α (ng/mL)25.0 ± 2.9 (= 22)16.5 ± 1.2ns (= 3)8.5 ± 1.1*** (= 17)14.5 ± 3.6ns (= 8)5.6 ± 2.6** (= 5)

Finally, 24-h treatment of MG with either Meth or synthetic cannabinoids used at the above indicated effective concentrations did not affect the cell survival as the MG density measured after 4′,6-Diamidino-2-phenylindole (DAPI) labeling was similar as untreated condition (Meth 105 ± 13%, n = 3; WIN 107 ± 10%, n = 3; CP 95 ± 8%, n = 3).

Cannabinoids prevent the inhibition of astrocyte gap junction channels permeability induced by conditioned medium harvested from LPS-activated microglia

Untreated primary cultures of astrocytes (Control) showed a high value of LY fluorescence area that reflects a strong level of GJC activity (Fig. 1a). As previously observed (Même et al. 2006), diluted CM (1/4), harvested from LPS-activated MG cultures, applied for 24 h on astrocytes strongly reduced the intercellular diffusion of LY by 48% compared to control astrocytes (n = 13 both for control and CM, < 0.001, Fig. 2a). Interestingly, the co-incubation with diluted CM of Meth (10 μM) significantly prevented the inhibition of GJC activity induced by the CM treatment alone. Indeed, the inhibition of astrocyte dye coupling was only 25% when Meth was added with CM (Meth + CM, n = 8, < 0.01 compared to CM alone) (Fig. 2a).

image

Figure 2.  Cannabinoid agonists prevent the inhibition of GJCs induced by conditioned medium harvested from LPS-activated MG applied on cultured astrocytes. (a) Graph showing the effect of methanandamide (Meth) on the inhibition of GJC activity induced by 24 h-treatment of astrocyte cultures with diluted CM (¼). Each bar represents the fluorescence area measured after CM treatment (CM; black bar) as compared to untreated culture (C for control; white bar) and after the incubation with CM of Meth (10 μM; grey bar). Data expressed as pixels per cm2, are the mean ± SEM, were obtained from 7 to 13 independent measures. (b) Graph showing the effect of synthetic CBs on the GJC inhibition induced by the CM treatment. Each bar represents the fluorescence area measured after CM-treatment (CM; black bar), as compared to untreated culture (C for control; white bar) and after the co-treatment with CM of the two synthetic CBs, WIN or CP (1 μM each; grey bars). Data, expressed in pixels per cm2, are mean ± SEM obtained from four to eight independent measures. ***< 0.001, **< 0.01 and *< 0.05 as compared to the indicated groups (one way anova followed by the Bonferroni post hoc test).

Download figure to PowerPoint

Synthetic CBs were then tested on the CM-induced GJC inhibition. The 58% inhibition of GJC compared to control, obtained by CM application alone (n = 8 both for control and CM, < 0.001; Fig. 2b) was significantly prevented by the co-incubation with CM of the synthetic CBs, WIN or CP (1 μM for each). Indeed, this inhibition only reached 18% for WIN + CM (n = 5, < 0.01 compared to CM) and 17% for CP + CM, (n = 4, < 0.05 compared to CM) (Fig. 2b).

Cannabinoids prevent the opposite regulation of Cx43 channels induced by pro-inflammatory cytokines

Recently, it has been reported that treatment of astrocytes with a mixture of IL-1β and TNF-α (Mix; 10 ng/mL for each) induces an opposite regulation of Cx43 GJCs and HCs activities (Retamal et al. 2007). Indeed, this treatment inhibits gap junctional communication as assessed by dye coupling experiments while it increases Etd uptake through open HCs. Accordingly, we tested whether treatments with either Meth or synthetic CBs could modulate the Mix effect on these two functions of Cx43 channels.

Methanandamide GJC permeability

Application of Mix (IL-1β plus TNF-α, 10 ng/mL each) for 24 h in astrocyte cultures induced a 41% inhibition of gap junctional communication as compared to control untreated astrocytes (n = 12 for both Mix and control; < 0.001, Fig. 3a-i,ii and c). The co-application of 10 μM Meth with Mix, prevented partially but significantly, the strong inhibition of GJCs induced by Mix alone (10% inhibition, n = 6, < 0.001 compared to Mix alone; Fig. 3a-iii and c). A lower dose of 1 μM of Meth also significantly reduced the Mix-induced GJC inhibition (25% inhibition n = 12, < 0.05 compared to Mix alone; not shown).

image

Figure 3.  Cannabinoid agonists prevent the dual regulation of Cx43 functions evoked by IL-1β and TNF-α in astrocyte cultures. (a and b) Representative images depicting the LY diffusion through GJCs (a) and the Etd uptake via HCs (b) in pure astrocyte cultures. 24 h-treatment with the mixture of cytokines (Mix, IL-1β plus TNF-α; 10 ng/mL each) reduced the LY diffusion (a-ii) associated with an increase in Etd uptake (b-ii) as compared to untreated astrocytes (a-i, b-i). The co-addition with Mix, of either the endogenous CB: Meth (Mix + Meth) or the synthetic CB, WIN (Mix + WIN) reversed both the inhibition of GJC function (a-iii, a-iv) and the increase in HC activity (b-iii, b-iv). (c and d) Graphs showing the preventive effects of Meth on the Mix-induced Cx43 dual regulation, i.e. the inhibition of GJC function (c) versus the increase in HC activity (d). Each bar represents the fluorescence area (c) or the level of Etd uptake (d), measured in untreated astrocyte cultures (C for control, white bars, c, d), or treated with Mix alone (black bars, c, d), or co-treated with Mix plus10 μM Meth (grey bars, c, d). Data, expressed in pixels per cm2 (for c) and in number of Etd positive nuclei/field (for d), are the mean ± SEM obtained from 7 to 12 and 3 to 6 independent measures, respectively. (e and f) Graphs showing the preventive effect of synthetic cannabinoids on the Mix-induced Cx43 dual regulation, i.e. the inhibition of GJC function (e) versus the increase in HC activity (f). Each bar represents the fluorescence area (for e) or the level of Etd uptake (for f), measured in untreated astrocyte cultures (C for control, white bars, e, f), treated with Mix alone (Mix, black bars; e, f), or co-treated with Mix plus 1 μM WIN (first grey bars; e, f) or Mix plus 1 μM CP (second grey bar, e), respectively. Data, expressed in pixels per cm2 (for e) and in number of Etd positive nuclei/field (for f), are the mean ± SEM obtained from 7 to 11 and 3 independent measures, respectively. ***< 0.001, **< 0.01 and *< 0.05 as compared to the indicated groups (one way anova followed by the Bonferroni post hoc test).

Download figure to PowerPoint

The effects of this non-hydrolysable analogue of anandamide applied alone on GJC activity were also investigated in order to exclude direct effects of this compound. Incubation for 24 h with increased concentrations of Meth failed to induce significant modifications in GJC activity, as the fluorescence area values were 2.51 ± 0.16 × 105 pixels/cm² (n = 9) in control untreated astrocytes and 2.50 ± 0.21 × 105 pixels/cm² (n = 5), 2.7 ± 0.10 × 105 pixels/cm² (n = 5) and 2.95 ± 0.18 × 105 pixels/cm² (n = 3) after the incubation with 1, 10 and 100 μM Meth, respectively (n.s. compared to control, for each dose). These observations, obtained with mouse cortical astrocytes, are in agreement with a previous work in which anandamide, instead of Meth, was reported to be active on GJCs of rat striatal astrocytes but ineffective on rat cortical astrocytes (Venance et al. 1995).

HC activity

As previously mentioned in astrocytes, Cx43 at plasma membrane can form HCs which allow the communication between intracellular and extracellular compartments (Sáez et al. 2005). Accordingly, we measured the uptake of Etd, considered as a functional index of Cx43 HC activity (Contreras et al. 2002), in cultured astrocytes following CB applications, under pro-inflammatory conditions. As described by Retamal et al. (2007), application of Mix for 24 h on astrocyte cultures induced a prominent increase in Etd uptake as compared to control untreated astrocytes (923% increase, n = 4 and 6 for Mix and control respectively, < 0.001; Fig. 3b-i,ii and d). Moreover, co-application of Mix and Meth (10 μM) almost completely abolished the Mix effect on evoked Etd uptake (88% decrease, n = 3, < 0.001 compared to Mix treatment alone; Fig. 3b-iii and d).

Synthetic cannabinoid agonists GJC permeability

When synthetic CBs were applied, the Mix-induced decrease in GJC permeability, compared to control untreated astrocytes (49% inhibition, n = 11 both for Mix and control, < 0.001; Fig. 3e) was also partially prevented. Indeed, the co-application of Mix and either 1 μM WIN or 1 μM CP led to significantly reduced values of GJC inhibition compared to Mix treatment alone (21% and 25% inhibition compared to Mix alone, n = 8 and 7, < 0.01 and < 0.05, for WIN + Mix and CP + Mix, respectively; Fig. 3a-iv and e). A significant impairment of the Mix-induced GJC inhibition was also observed with the addition of a lower concentration of WIN (100 nM, 23% inhibition, n = 8, < 0.05 compared to mix alone; not shown), which excluded any unspecific effect of this compound. In addition, no significant reversion of the Mix-induced GJC inhibition was detected when the inactive enantiomer form of this compound, S-(−)-WIN 55,212-2, was applied with Mix, at the following concentrations: 100 nM, 1 μM and 10 μM. The corresponding values of the GJC inhibition were 52%, 40% and 44%, compared to control (n = 5, 6 and 5 respectively, n.s. for each concentration compared to Mix alone, not shown). These data demonstrated that WIN exerted a specific effect as a CB agonist, to prevent the mix-induced GJC inhibition.

Moreover, to exclude a proper effect of synthetic CBs on GJC permeability, increasing concentrations of either WIN or CP alone were applied for 24 h in astrocyte cultures. In agreement with a previous study performed with rat striatal astrocytes in cultures (Venance et al. 1995), the measured GJC level did not show any significant differences for WIN (2.16 ± 0.25 × 105 pixel/cm², n = 5 and 2.08 ± 0.30 × 105 pixels/cm², n = 6 for 1 and 10 μM of WIN, respectively) and CP treatments (2.24 ± 0.29 × 105 pixel/cm², n = 5 and 2.16 ± 0.24 × 105 pixel/cm², n = 5 for 1 and 10 μM CP, respectively) as compared to untreated astrocytes (2.05 ± 0.15 × 105 pixel/cm²; n = 13).

Finally, the effects of co-applications of Meth with each synthetic cannabinoids (WIN and CP) were tested on Mix-induced inhibition of GJC. When Meth and WIN were co-applied on astrocytes, the Mix-induced inhibition of dye coupling was totally prevented (2.75 ± 0.28 × 105 pixel/cm2, n = 4, < 0.05 compared to WIN alone). Similarly, Meth and CP also exerted an additive preventive effect (2.30 ± 0.13 × 105 pixel/cm2, n = 4, < 0.05 compared to CP alone).

HC activity

The effect of synthetic CBs on the increased activity of Cx43 HCs induced by Mix was also investigated. As already described above, 24 h application of Mix on astrocyte cultures, induced an important increase in Etd uptake as compared to control untreated astrocytes (413% increase, n = 8 and 10 for Mix and control, respectively, < 0.001; Fig. 3f). The co-application of Mix with the synthetic CB receptor agonist, WIN (1 μM) abolished the Mix-induced Etd uptake (n = 6, < 0.001 compared to Mix; Fig. 3b-iv and f). In order to complete this observation a range of increasing doses of WIN was tested from 0.1 nM to 10 μM on Mix-evoked Etd uptake. Low WIN concentrations (0.1 nM, 1 nM, 10 nM) were found to be uneffective on Etd uptake induced by Mix (544%, 370% and 452% increase compared to the basal level, n = 4 for each dose), while application of 100 nM WIN reduced the Mix-induced Etd uptake, but without statistical significance (120% increase compared to the basal level, n = 4). Finally, application of 3 μM and 10 μM WIN reduced significantly the mix-induced Etd uptake, and stronger than 1 μM WIN, as the uptake level went below the basal uptake (−50%, and −90% compared the basal level for 3 and 10 μM, respectively, n = 3 for each).

Cannabinoid treatments do not prevent the inhibition of Cx43 expression in mix-treated astrocytes

It was previously reported that the Mix-induced dual regulation of Cx43 GJCs and HCs is associated with a down-regulation of the total level of Cx43 expression (Même et al. 2006; Retamal et al. 2007). Because both synthetic and endogenous CBs prevent the opposite regulation of Cx43-based channels induced by the mixture of cytokines, we assessed the consequences of CBs treatments on Cx43 levels, using either immunofluorescence or western blotting technical approaches.

As expected, double immunofluorescence labeling with a GFAP monoclonal antibody and total Cx43 polyclonal antibody revealed a decrease of Cx43 level in Mix-treated (24 h) astrocytes compared to untreated ones (Fig. 4a and b). Co-treatment for 24 h of astrocyte cultures with Mix plus WIN (1 μM) did not modify the levels and pattern of Cx43 bands (Fig. 4c).

image

Figure 4.  The recovery action of cannabinoids on the Mix-induced opposite regulation of Cx43 functions is not associated with a recovery of Cx43 levels in astrocyte cultures. (a–c) Confocal images depicting the double immunofluorescence staining using an anti-GFAP polyclonal antibody and an anti-Cx43 monoclonal antibody. GPAP-positive astrocytes (green) showed a high level of Cx43 reactivity (red) in untreated conditions (a). 24 h-Mix treatment induced a drastic decrease in Cx43 reactivity (Mix; b) which was not prevented by co-treatment with 1 μM WIN (Mix + WIN; c). (d) Western-blot analysis of total Cx43 protein from extract of cultured astrocytes showing that Mix treatment induced a strong decrease in Cx43 expression level as compared to untreated astrocytes (C for control). The co-treatment with either the synthetic cannabinoid, WIN (1 μM) or the endogenous cannabinoid, Meth (10 μM) did not prevent the reduction of Cx43 expression levels. NP, P1 and P2 refer to the non-phosphorylated and phosphorylated isoform of Cx43, respectively.

Download figure to PowerPoint

In addition, western-blot analyses were performed to quantify the total level of Cx43 after these latter treatments applied on pure astrocytes. As typically observed, the polyclonal antibody detected three Cx43 isoform bands, one unphosphorylated (NP) and two phosphorylated isoforms (P1 and P2) (see Fig. 4d). In parallel, a western blot analysis of α-tubulin protein was performed on the same samples and was taken as an internal control for the quantification (Fig. 4d). The western blot assay indicated that Mix treatment induced a drastic reduction in Cx43 protein signal (Fig. 4d), as previously described (Même et al. 2006), whereas 24 h co-treatment with Mix plus either WIN (1 μM) or Meth (10 μM) did not modify the total Cx43 level (Fig. 4d; data not shown for 1 μM CP). Finally, both Meth (10 μM) and WIN (1 μM), each applied alone for 24 h on cultured astrocytes did not modify the total Cx43 level (data not shown). The quantification by optic density of (i) the three Cx43 bands together and (ii) the α-tubulin band were expressed as a ratio of the total Cx43 density as compared to α-tubulin density. Accordingly, the Cx43/α-tubulin ratio was significantly reduced by 58% in total homogenates of 24 h Mix-treated astrocyte compared to control untreated astrocytes (0.42 ± 0.08 vs. 1.00 ± 0.12, n = 9 for untreated and Mix-treated astrocytes, respectively, < 0.01). The co-treatment of astrocytes with Mix and Meth (10 μM for 24 h), did not modify significantly the Cx43/α-tubulin ratio (0.54 ± 0.15, n = 4) compared to astrocytes treated with Mix alone, as this ratio was significantly lower as compared to control (46% decrease, < 0.05). Similarly, addition of WIN (1 μM, 24 h) with Mix did not significantly change the Cx43/α-tubulin ratio (0.38 ± 0.08, n = 6) compared to that measured after Mix treatment alone. Indeed, this ratio was at a significant low level compared to control (62% decrease, < 0.01). Also, CP (1 μM) did not affected the Cx43/α-tubulin ratio compared to Mix treatment alone (0.44 ± 0.16, n = 4). Finally, the ratio of the non-phosphorylated Cx43 isoform versus α-tubulin was also quantified after Mix treatment and the different cannabinoids treatments in astrocyte cultures. 24 h Mix treatment induced a strong decrease in this ratio as compared to control (0.23 ± 0.09 and 0.68 ± 0.12, n = 9 for Mix-treated and untreated astrocytes, respectively; < 0.05). Addition of cannabinoid agonists with Mix treatment, i.e. Meth (10 μM), WIN (1 μM) and CP (1 μM), did not significantly modify the non-phosphorylated Cx43/α-tubulin ratios (0.32 ± 0.14, n = 4; 0.17 ± 0.06, n = 6 and 0.13 ± 0.09, n = 4, respectively) compared to mix-treated astrocytes.

In order to address the mechanisms involved in GJC inhibition induced by both pro-inflammatory cytokines and conditioned medium (harvested from LPS-activated microglia), we measured the level of Cx43 transcription in pure astrocytes, by quantitative RT-PCR. As compared to untreated conditions, Cx43 transcription in astrocytes decreased by 4.9 ± 0.1 (mean ± SEM, n = 3) fold in presence of Mix, and by 2.8 ± 1.3 (mean ± SEM, n = 3) fold in presence of conditioned medium. Thus, a down-regulation of Cx43 transcription accounts for the inhibition of GJC permeability by Mix treated astrocytes.

The molecular mechanisms involved in the induction of HC activities in astrocytes following Mix treatments is still unknown, and the molecular composition of such HC is still a matter of debate (Iglesias et al. 2009). They have been described to be composed by Cx43 or Pannexin1 (Panx1), a new protein forming non-selective, high-conductance channels permeable to ATP (Scemes et al. 2007). Presently, Panx1 transcription was assessed by quantitative RT-PCR to determine whether HC activation in astrocytes following Mix treatment could involve an increase in Panx1 mRNA, which could account for the increase in cell permeabilisation. As compared to untreated cultures, Panx1 transcription in astrocytes treated with Mix demonstrated a 1.5 ± 0.1 (mean ± SEM, n = 3) fold decrease, and a 1.8 ± 0.1 (mean ± SEM, n = 3) fold decrease when astrocytes were treated with conditioned medium. These results demonstrate that pro-inflammatory treatments on astrocytes do not increase the level of Panx1 expression. Accordingly, the participation of Panx1 in the Mix-induced increase in HC activity appears poorly prominent.

Pharmacological characterization of the cannabinoid preventive effect exerted on the opposite regulation of Cx43-based channels induced by the mixture of pro-inflammatory cytokines

Two CB receptor antagonists the CB1 antagonist, SR-141716 (SR1; Rinaldi-Carmona et al. 1994) and the CB2 antagonist SR-144528 (SR2; Rinaldi-Carmona et al. 1998), were used to characterize what sub-type(s) of CB receptor(s) is (are) involved in the preventive effects, described above, for endogenous and synthetic CBs on the Mix-induced opposite regulation of GJCs versus HCs. For this purpose, we added into the culture medium either SR1 or SR2 (i) with Meth plus Mix and (ii) with WIN plus Mix, both for 24-h incubation.

Methanandamide

While Mix treatment inhibited gap junctional communication by 41% compared to control untreated astrocytes (n = 11 both for Mix and control, < 0.001; Fig. 5a), the co-application of Meth with Mix, significantly prevented this effect, as the inhibition was only 21% (n = 7, < 0.001 compared to Mix; Fig. 5a). In the presence of the CB1 receptor antagonist, SR1 (1 μM), the preventing effect of methanandamide on Mix-induced GJC inhibition was not affected, as the inhibition of dye coupling remained very low (22% inhibition, n = 5, < 0.05 compared to Mix; Fig. 5a). Similarly, 24 h incubation of the CB2 receptor antagonist SR2 (1 μM) with Mix and Meth did not modify the recovery effect of Meth on GJC activity as the inhibition of dye coupling was 16% (n = 7, < 0.01 compared to Mix Fig. 5a). Moreover, when SR1 and SR2 were added together with Mix and Meth, the recovery effect of methanandamide on dye coupling was also maintained, with a 23% inhibition of dye coupling (< 0.05 compared to Mix-treated group, n = 4; Fig. 5a). This result suggests that the preventing effect of Meth on the Mix-induced inhibition of GJC function does not involve the activation of either CB1 or CB2 receptors.

image

Figure 5.  Pharmacological characterization of the preventive effects of endogenous or synthetic cannabinoids on the Mix-induced inverse regulation of Cx43 functions in astrocyte cultures. (a) Graphs showing the pharmacological characterization of the reverse effect of the endogenous CB, Meth, on the Mix-induced GJC inhibition. Each bar represents the quantification of the fluorescence area from SL/DT assay performed either on untreated astrocytes (C for control, white bar), or on astrocytes treated with Mix alone (black bar) or co-treated with (i) Mix plus Meth (10 μM), (ii) Mix plus Meth plus the CB1 receptor antagonist, SR-141716A (Meth 10 μM + SR1 1 μM), (iii) Mix plus Meth plus the CB2 receptor antagonist, SR-144528 (10 μM Meth + SR2 1 μM), and finally (iv) Mix plus Meth plus the CB1 receptor antagonist, plus the CB2 receptor antagonist (Meth 10 μM + SR1 1 μM + SR2 1 μM) (four successive grey bars, respectively). Data, expressed in pixels per cm2, are the mean ± SEM obtained from 4 to 11 independent experiments. (b) Graphs showing the pharmacological characterization of the reverse action the endogenous CB, Meth, on the mix-induced Etd uptake increase. Each bar represents the level of Etd uptake evaluated either from untreated astrocyte (C for control, with bar), or from astrocytes treated with Mix alone (black bar) or co-treated with (i) Mix plus Meth (10 μM), (ii) Mix plus Meth plus the CB1 receptor antagonist, SR-141716A (Meth 10 μM + SR1 1 μM), (iii) Mix plus Meth plus the CB2 receptor antagonist, SR-144528 (Meth 10 μM + SR2 1 μM), and finally (iv) Mix plus Meth plus the CB1 receptor antagonist, plus, the CB2 receptor antagonist (Meth 10 μM + SR1 1 μM + SR2 1 μM) (four successive grey bars, respectively). Data, expressed in number of Etd positive nuclei/field, are the mean ± SEM obtained from three independent measures. (c) Graphs showing the pharmacological characterization of the reverse effect of the synthetic CB, WIN 55,212-2, on the Mix-induced GJC inhibition. Each bar represents the quantification of the fluorescence area from SL/DT assay performed either on untreated astrocyte (C for control, withe bar) or on astrocytes treated with Mix alone (black bar) or co-treated while (i) Mix plus WIN (1 μM), (ii) Mix plus WIN plus the CB1 receptor antagonist, SR-141716A (WIN 1 μM + SR1 1 μM), and finally (iii) Mix plus WIN plus the CB2 receptor antagonist, SR-144528 (Meth 10 μM + SR2 1 μM), (three successive grey bars, respectively). In addition, the effects on fluorescence area of the CB1 and CB2 antagonists, respectively, applied alone on astrocytes were determined (SR1 1 μM; SR2 1 μM, two hatched bars, respectively) Data, expressed in pixels per cm2, are the mean ± SEM obtained from 3 to 11 independent measures. (d) Graph showing the pharmacological characterization of the reverse effect of the synthetic CB, WIN on the Mix-induced HC increase. Each bar represents the level of Etd uptake evaluated either from untreated astrocytes (C for control, white bar), or from astrocytes treated with Mix alone (black bar) or co-treated with (i) Mix plus WIN (1 μM), (ii) Mix plus WIN plus the CB1 receptor antagonist, SR-141716A (1 μM WIN + 1 μM SR1), and finally (iii) Mix plus WIN plus the CB2 receptor antagonist, SR-144528 (10 μM Meth + 1 μM SR2), (three successive grey bars, respectively). Furthermore, the effects on Etd uptake of the CB1 and CB2 antagonists, respectively applied alone on astrocytes were measured (1 μM SR1; 1 μM SR2, two hatched bars, respectively) Data, expressed in number of Etd positive nuclei/field, are the mean ± SEM obtained from three independent measures. ***< 0.001, **< 0.01 and *< 0.05 as compared to the indicated groups (one way anova followed by the Bonferroni post hoc test).

Download figure to PowerPoint

The pharmacological characterization of the preventing effect of Meth was similarly assessed on the Mix-induced increase in HC activity. Mix treatment alone increased by 1025% the Etd uptake compared to control untreated astrocytes (n = 3, < 0.001). As illustrated in figure 5(b), this effect was reversed by Meth (88% reduction, n = 3, < 0.001 compared to Mix). Interestingly, addition of SR1 (1 μM), completely abolished the recovery effect induced by Meth on the Mix-induced increase in the Etd uptake level, which was at a similar level to that observed in Mix-treated astrocytes (+988% as compared to control, n = 3). In contrast, co-incubation of Mix and Meth with SR2 (1 μM) did not modify the preventive effect exerted by Meth on the Mix-induced Etd increase. Indeed, the HC activity was maintained at a low level (n = 3, < 0.01 compared to Mix; Fig. 5b). Finally, we tested the consequence on Etd uptake of the co-incubation of both SR1 and SR2 with Meth and Mix. In these conditions, the preventive effect of Meth on Mix-induced Etd uptake was partially abolished as the uptake level was increased by 760% compared to control (n = 3, < 0.05 Mix and < 0.05 compared to Mix + meth; Fig. 5b). These observations indicate that the preventive effect of Meth on Etd uptake induced by Mix is mediated through the activation of a CB1 receptor in astrocytes.

Synthetic cannabinoids

The identity of the CB-R sub-type responsible for the preventing effect of the synthetic CB WIN on the Mix-induced opposite regulation of Cx43-based channels was also questioned. As illustrated in figure 5(c), Mix treatment induced a strong inhibition (47% compared to control) of GJC activity (n = 9, < 0.001) compared to control untreated astrocytes (n = 11). This effect was significantly prevented by the addition of 1 μM WIN (21% decrease, n = 9, < 0.01 compared to Mix), as described above (see Fig. 3e). The addition for 24 h of SR1 with WIN (1 μM) and Mix significantly abolished the preventive effect of WIN (< 0.05), as previously shown. Indeed, in the presence of this CB1 receptor antagonist GJC inhibition was maintained to the similar low level as observed with Mix treatment alone (40% inhibition, n = 6, n.s. compared to Mix; Fig. 5c). In contrast, the addition of SR2 (1 μM) with WIN (1 μM) and Mix for 24 h did not modify the preventive effect of this CB agonist on the Mix-induced GJC inhibition (Fig. 5c). In fact, the reduction of dye coupling was only 25% in the presence of SR2 (n = 3, < 0.05 compared to Mix; Fig. 5c). Intrinsic action of these antagonists on astrocyte GJC activity was excluded as neither SR1 nor SR2 applications performed at the concentrations indicated above, induced any significant modification on GJC permeability as compared to untreated condition (8% and 10%, reduction, n = 4 and 3 for SR1 and SR2 alone, respectively, ns for each compared to control; Fig. 5c).

Finally, the identification of the CB receptor sub-type responsible for the preventive effect of WIN on the Mix-induced increase in Etd uptake was also investigated. As previously shown, Mix treatment on astrocytes induced a significant increase in Etd uptake (+1500%) compared to control untreated condition (< 0.001, n = 3 for each; Fig. 5d). This effect was significantly reversed by co-application of Mix and WIN (1 μM) as Etd uptake was reduced by 98% (< 0.001 compared to Mix treated group, n = 3; Fig. 5d). The co-addition of SR1 (1 μM) with WIN and Mix completely abolished the preventive effect of WIN, as in this condition the increase in Etd was 4670% (n = 3, < 0.001 as compared to WIN + Mix), a similar level to that measured in astrocytes treated with Mix alone (Fig. 5d). On the contrary, co-incubation of SR2 (1 μM) with WIN (1 μM) plus Mix did not modify the recovery effect of WIN on Mix-evoked Etd uptake increase (Fig. 5d). Finally, neither SR1 nor SR2 had a proper effect on the basal level of Etd uptake in cultured astrocytes (Fig. 5b). These observations indicate that the preventing effect of endogenous as well as synthetic CBs on GJC and HC activities are mediated by CB1 receptors in astrocytes.

Cannabinoids inhibit the LPS-induced TNF-α and IL-1β secretion from microglia

Lipopolysaccharide treatment (10 ng/mL) in astrocyte/MG induced a strong inhibition of GJC activity while a such concentration of LPS does not produce similar effect in primary cultures of astrocytes (Même et al. 2006). Indeed, a 1000-fold higher LPS concentration was required to obtain a substantial decrease in GJC function. Accordingly, the opposite regulation of astroglial Cx43 channel functions is solely triggered by LPS-activated MG, and mediated by the two released pro-inflammatory cytokines, IL-1β and TNF-α (Retamal et al. 2007).

First, the effect of CBs was assessed on the LPS-induced IL-1β and TNF-α release in cultured MG. Co-incubation with LPS of the CBs agonists on MG cultures reduced the production of both cytokines. These data complete a previous study performed on rats MG in which endogenous and synthetic CBs were shown to ablate the LPS-induced TNF-α release (Facchinetti et al. 2003). Moreover, CB treatments with Δ9-tetrahydrocannabinol, Meth or CP, were also shown to restrain the LPS-induced increase at mRNA levels of these two cytokines (Puffenbarger et al. 2000). Finally, it is noteworthy that the amounts of IL-1β and TNF-α measured in the present study (Table 1 and 2) under LPS treatment are in the same range than the lowest concentrations required to inhibit GJC function (Même et al. 2006). Interestingly, their levels measured when LPS-activated MG are co-treated with endogenous or synthetic CBs (Table 1 and 2) were ineffective on GJC function, as tested by the application of the resulting collected CM on astrocyte cultures. This suggests that in astrocyte/MG co-cultures, the lack of effect of LPS on astroglial Cx43 GJC is primarily because of the low level of IL-1β and TNF-α secretion that is maintained under threshold by cannabinoid stimulation.

Cannabinoids prevent the dual regulation of astrocytic Cx43 channel functions induced by pro-inflammatory treatments

In order to determine the effects of CB-R stimulation on astrocytes themselves, and especially on Cx43 channels, we used primary cultures of astrocytes upon which two pro-inflammatory treatments were applied: (i) conditioned medium harvested from LPS-activated MG, and (ii) the mixture of IL-1β and TNF-α. In these situations, both Meth and synthetic CBs partially prevented the inhibition of GJC and the activation of HCs induced by these treatments. These observations demonstrated that astrocytes are also a target for CBs to control the regulation performed by pro-inflammatory cytokines on the two Cx43 channel functions. However, the pharmacological profile of Meth and synthetic CBs was distinct and their effects were additive indicating that they act on different receptors (see below). In addition, while cytokines treatment alone reduces the level of Cx43 expression studied by western blotting or immunofluorescence experiments, CBs application does not restore Cx43 expression to control level. A similar observation was previously reported, as either CM or Mix treatment reduces Cx43 level whereas pharmacological treatments that prevent the opposite regulation of Cx43 HCs and GJCs do not restore the expression level (Retamal et al. 2007). Indeed, the opposite regulation of Cx43 GJCs and HCs was not observed in the presence of the p38 MAPK inhibitor SB202190. Interestingly, in astrocytes CBs exert a protective role on oxidative stress by inhibiting the p38 MAPK pathway (Carracedo et al. 2004). Thus it is likely that the preventing effect of CBs on Mix or CM treatment could act through the inhibition of this signaling pathway to abrogate the opposite regulation of Cx43 channels, without changing its protein levels.

It has been recently shown that Panx, a new family of protein, are able to exert an hemichannel function in untreated astrocytes (Iglesias et al. 2009). This activity involved the Panx1 channels which are expressed in astrocytes (Scemes et al. 2007). Here, we measured the Panx1 mRNA expression in astrocytes subjected to pro-inflammatory treatments (conditioned medium or Mix). Under these two treatments Panx1 mRNA expression was strongly decreased compared to Cx43 mRNA suggesting a poor involvement of Panx1 in the evoked Etd uptake described under our pro-inflammatory conditions. Moreover, we have previously reported that the Mix-triggered Etd uptake was (i) absent in Cx43 knock-out astrocytes and (ii) specifically inhibited by the Cx43 hemichannel blockers (Retamal et al. 2007) which also argue for the involvement of Cx43 HCs rather than Panx1 HCs.

Pharmacological characterization of the modulations of glia functions exerted by cannabinoids

The pharmacological profile of cannabinoid effects was determined using agonists and antagonist at either CB1 or CB2 receptors. In microglia, the inhibition of LPS-evoked IL-1β and TNF-α release induced by Meth did not involve the activation of CB1 or CB2 receptors. Similar results were obtained for the WIN-mediated inhibition of the LPS-induced release of these cytokines. These observations are in agreement with a previous study (Puffenbarger et al. 2000) and with the general statement about the cannabinoid system expressed in MG (Stella 2004).

In astrocytes, the preventive effect of WIN on the opposite regulation of Cx43 channels was well antagonized by application of SR1, but not by SR2. This indicates a CB1-mediated pathway and suggests the absence of CB2 receptor expression, as already reported (Walter and Stella 2003). In contrast, while the effect of Meth on the Mix-induced HC activation is completely blocked by SR1, its preventive action on the GJC inhibition is not antagonized by SR1 or SR2. This suggests that a non-CB1/CB2 receptor is activated by Meth and that the regulation of the two types of Cx43 channels operate through different pathways. Interestingly, Retamal et al. (2007) reported that the opposite regulation of Cx43 HCs versus GJCs shares the same p38/MAPK pathway, but that, in addition, nitric oxide production was also involved in HC regulation (see also Orellana et al. 2009). Based on these observations, we hypothesize that Meth may activate CB1 receptors linked to p38/MAPK but also may stimulate a non-CB1/CB2 receptor that affects the redox status generated by the Mix treatment.

The pharmacological profile described here for endogenous and synthetic CBs treatments is rather complex but is consistent with the actual state of knowledge about the expression of CB receptors in astrocytes. Indeed, in addition to the well-documented expression CB1 receptors in astrocytes (see Stella 2004) there is now evidence for the expression of other CB receptors distinct from CB1 and CB2 (Brown 2007). Interestingly, the recent report of an inhibition of a potassium conductance in rat cortical astrocytes by anandamide also supports the notion that a non-CB1/CB2 receptor is involved (Vignali et al. 2009) and confirms the previous description of an anandamide receptor-mediated inhibition of GJC in striatal astrocytes that was not reproduced by synthetic cannabinoids (Venance et al. 1995). Finally, the finding that the effects of Meth and synthetic cannabinoids on Cx43 channels are additive strongly argue in favor of the co-expression of two types of CB receptors in astrocytes.

Relevance for brain pathologies

Astrocytes are known to play a role in neuronal survival (Kirchhoff et al. 2001; Giaume et al. 2007). Interestingly, a typical property of astrocytes compared to neurons or other glial cells sub-types is their large amount of Cxs expression (Giaume and McCarthy 1996). In pathological situations, these astroglial Cxs are known to play a role in neuronal fate (Nakase and Naus 2004; Orellana et al. 2009). However, their involvement in neuroprotection and/or neurotoxicity is still debated in the literature (Rouach et al. 2002; Nakase and Naus 2004; Farahani et al. 2005) likely because up-to-now the HC function of Cxs in astrocytes was not fully taken into account in this context (see Orellana et al. 2009). Reactive gliosis and brain inflammation are associated with most, if not all, brain injuries and pathologies. In addition, there is a growing amount of evidence suggesting that CBs may be neuroprotective in CNS inflammatory conditions (Correa et al. 2007; Cabral and Griffin-Thomas 2008). Along the present study, we observed that CBs, by acting on the two main actors of reactive gliosis (i.e. MG and astrocytes), prevent the opposite regulation Cx43 channels. HC activation in astrocytes could play a crucial role in the reinforcement of the neuronal death, because of their capacity to release glutamate (Ye et al. 2003) and ATP (Cotrina et al. 1998). Thus, the prevention of the HC activation by pro-inflammatory treatment may represent a unexplored strategy against neuronal damage. Accordingly, the determination of astroglial CB receptors sub-types that are activated differentially by synthetic and endogenous CBs may contribute to define therapeutic targets to prevent a deleterious function of Cx43, i.e. HC activity, while maintaining the other, i.e. GJC activity.

Discussion

  1. Top of page
  2. Abstract
  3. Materials and methods
  4. Results
  5. Discussion
  6. Acknowledgements
  7. References

Reactive gliosis is a trademark of many, if not all, brain injuries and pathologies. This process is closely associated with brain inflammation and mainly supported by the change in phenotype of two sub-types of glial cells, astrocytes and MG. The understanding of cellular and molecular steps that link their respective activation status is important to dissect the mechanisms leading to neuronal dysfunction and death. Accordingly, the objective of the present study was to determine how anti-inflammatory agents, like cannabinoids, interfere with a change in the functional status of astrocytes triggered by pro-inflammatory treatments. Here, this question was addressed by focusing on the previously reported cellular interaction based upon LPS activation of MG that triggers a dual regulation of two functions of Cx43 channels in astrocytes (Retamal et al. 2007). Using either (i) astrocytes/MG co-cultures treated with LPS or (ii) pure cultures of astrocytes treated with pro-inflammatory cytokines, we demonstrated that the glial actions of endogenous and synthetic cannabinoids are processed at two levels. Firstly, these cannabinoids interfere with the release of two pro-inflammatory cytokines from MG and thus, play the role of an immune modulator (see Klein et al. 2003; Pacher et al. 2006). Second, they counteract the effect of these cytokines by acting directly at the level of astrocytes. Altogether, these observations strength the emerging concepts which display cannabinoids (i) as modulators of the brain immune system (Correa et al. 2005) and (ii) as therapeutic agents for ablating neuroinflammation (Cabral and Griffin-Thomas 2008).

Acknowledgements

  1. Top of page
  2. Abstract
  3. Materials and methods
  4. Results
  5. Discussion
  6. Acknowledgements
  7. References

This work was supported by an INSERM/CONICYT cooperative grant, FONDECYT N°1070591 (Chile) and by the CRPCEN (France). N.F. was a recipient of a fellowship from France Alzheimer and J.A.O. was supported by a grant from INSERM (Département des relations internationales). We gratefully thank Dr Francis Barth (Sanofi Company) for his generous gift of the CB receptor antagonists.

References

  1. Top of page
  2. Abstract
  3. Materials and methods
  4. Results
  5. Discussion
  6. Acknowledgements
  7. References
  • Baker D. and Pryce G. (2008) The endocannabinoid system and multiple sclerosis. Curr. Pharm. Des. 14, 23262336.
  • Baker D., Pryce G., Croxford J. L., Brown P., Pertwee R. G., Huffman J. W. and Layward L. (2000) Cannabinoids control spasticity and tremor in a multiple sclerosis model. Nature 404, 8487.
  • Benito C., Núñez E., Tolón R. M., Carrier E. J., Rábano A., Hillard C. J. and Romero J. (2003) Cannabinoid CB2 receptors and fatty acid amide hydrolase are selectively overexpressed in neuritic plaque-associated glia in Alzheimer’s disease brains. J. Neurosci. 23, 1113611141.
  • Block M. L. and Hong J. S. (2005) Microglia and inflammation-mediated neurodegeneration: multiple triggers with a common mechanism. Prog. Neurobiol. 76, 7798.
  • Brown A. J. (2007) Novel cannabinoid receptors. Br. J. Pharmacol. 152, 567575.
  • Cabral G. A. and Griffin-Thomas L. (2008) Cannabinoids as therapeutic agents for ablating neuroinflammatory disease. Endocr. Metab. Immune Disord. Drug Targets 8, 159172.
  • Calvo C. F., Cesselin F., Gelman M. and Glowinski J. (2000) Identification of an opioid peptide secreted by rat embryonic mixed brain cells as a promoter of macrophage migration. Eur. J. Neurosci. 12, 26762684.
  • Carracedo A., Geelen M. J., Diez M., Hanada K., Guzmán M. and Velasco G. (2004) Ceramide sensitizes astrocytes to oxidative stress: protective role of cannabinoids. Biochem. J. 380, 435440.
  • Coffey R. G., Snella E., Johnson K. and Pross S. (1996) Inhibition of macrophage nitric oxide production by tetrahydrocannabinol in vivo and in vitro. Int. J. Immunopharmacol. 18, 749752.
  • Contreras J. E., Sánchez H. A., Eugenin E. A., Speidel D., Theis M., Willecke K., Bukauskas F. F., Bennett M. V. and Sáez J. C. (2002) Metabolic inhibition induces opening of unapposed connexin 43 gap junction hemichannels and reduced gap junctional communication in cortical astrocytes in culture. Proc. Natl Acad. Sci. USA 99, 495500.
  • Correa F., Mestre L., Molina-Holgado E., Arévalo-Martín A., Docagne F., Romera E., Molina-Holgado F., Borell J. and Guaza C. (2005) The role of cannabinoids system on immune modulation: therapeutic implications on CNS inflammation. Mini Rev. Med. Chem. 5, 671675.
  • Correa F., Docagne F., Mestre L., Loría F., Hernangómez M., Borrell J. and Guaza C. (2007) Cannabinoid system and neuroinflammation: implications for multiple sclerosis. Neuroimmunomodulation 14, 182187.
  • Cotrina M. L., Lin J. H.-C. and Nedergaard M. (1998) Cytoskeletal assembly and ATP release regulate astrocytic calcium signaling. J. Neurosci. 18, 87948804.
  • Croxford J. L. (2003) Therapeutic potential of cannabinoids in CNS disease. CNS Drugs 17, 179202.
  • Facchinetti F., Del Giudice E., Furegato S., Passarotto M. and Leon A. (2003) Cannabinoids ablate release of TNFalpha in rat microglial cells stimulated with lipopolysaccharide. Glia 41, 161168.
  • Farahani R., Pina-Benabou M. H., Kyrozis A., Siddig A., Barradas P. C., Chiu F. C., Calvalcante L. A., Lai J. C., Stanton P. K. and Rozental R. (2005) Alterations in metabolism and gap junction expression may determine the role of astrocytes as “good samaritans” or executioners. Glia 50, 351361.
  • Giaume C. and McCarthy K. D. (1996) Control of gap-junctional communication in astrocytic networks. Trends Neurosci. 19, 319325.
  • Giaume C., Kirchhoff F., Matute C., Reichenbach A. and Verkhratsky A. (2007) Glia: the fulcrum of brain diseases. Cell Death Differ. 14, 13241335.
  • Iglesias R., Dahl G., Qui F., Spray D. C. and Scemes E. (2009) Pannexin 1: the molecular substrate of Astrocyte “hemichannels”. J. Neurosci. 29, 70927097.
  • Kirchhoff F., Dringen R. and Giaume C. (2001) Pathways of neuron-astrocyte interactions and their possible role in neuroprotection. Eur. Arch. Psychiatry Clin. Neurosci. 251, 159169.
  • Klein T. W. (2005) Cannabinoid-based drugs as anti-inflammatory therapeutics. Nat. Rev. Imnunol. 5, 400411.
  • Klein T. W., Newton C., Larsen K., Lu L., Perkins I., Nong L. and Friedman H. (2003) The cannabinoid system and immune modulation. J. Leukoc. Biol. 74, 486496.
  • Koulakoff A., Ezan P. and Giaume C. (2008) Neurons control the expression of connexin 30 and connexin 43 in mouse cortical astrocytes. Glia 56, 12991311.
  • Lyman W. D., Sonett J. R., Brosnan C. F., Elkin R. and Bornstein M. B. (1989) Delta 9-tetrahydrocannabinol: a novel treatment for experimental autoimmune encephalomyelitis. J. Neuroimmunol. 23, 7381.
  • Même W., Calvo C., Froger N., Ezan P., Amigou E., Koulakoff A. and Giaume C. (2006) Proinflammatory cytokines released from microglia inhibit gap junctions in astrocytes: potentiaton by β-amyloid. FASEB J. 20, 494496.
  • Molina-Holgado F., Molina-Holgado E., Guaza C. and Rothwell N. J. (2002) Role of CB1 and CB2 receptors in the inhibitory effects of cannabinoids on lipopolysaccharide-induced nitric oxide release in astrocyte cultures. J. Neurosci. Res. 67, 829836.
  • Nakase T. and Naus C. C. (2004) Gap junctions and neurological disorders of the central nervous system. Biochim. Biophys. Acta 1662, 149158.
  • Orellana J. A., Sáez P. J., Shoji K. F., Schalper K. A., Palacios-Prado N., Velarde V., Giaume C., Bennett M. V. and Sáez J. C. (2009) Modulation of brain hemichannels and gap junction channels by pro-inflammatory agents and their possible role in neurodegeneration. Antioxid. Redox Signal. 11, 369399.
  • Pacher P., Bátkai S. and Kunos G. (2006) The endocannabinoid system as an emerging target of pharmacotherapy. Pharmacol. Rev. 58, 389462.
  • Puffenbarger R. A., Boothe A. C. and Cabral G. A. (2000) Cannabinoids inhibit LPS-inducible cytokine mRNA expression in rat microglial cells. Glia 29, 5869.
  • Retamal M. A., Froger N., Ezan P., Sáez P. J., Sáez J. C. and Giaume C. (2007) Cx43 hemi- and full channels in astrocytes are oppositely regulated by pro-inflammatory cytokines released from activated microglia. J. Neurosci. 27, 13781392.
  • Ridet J. L., Malhotra S. K., Privat A. and Gage F. H. (1997) Reactive astrocytes: cellular and molecular cues to biological function. Trends Neurosci. 20, 570577.
  • Rinaldi-Carmona M., Barth F., Héaulme M. et al. (1994) SR141716A, a potent and selective antagonist of the brain cannabinoids receptor. FEBS Lett. 350, 240244.
  • Rinaldi-Carmona M., Barth F., Millan J. et al. (1998) SR 144528, the first potent and selective antagonist of the CB2 cannabinoid receptor. J. Pharmacol. Exp. Ther. 284, 644650.
  • Rouach N., Calvo C., Glowinski J. and Giaume C. (2002) Brain macrophages inhibit gap junctional communication and down-regulate connexin43 expression in cultured astrocyte. Eur. J. Neurosci. 15, 403407.
  • Sáez J. C., Retamal M. A., Basilio D., Bukauskas F. F. and Bennett M. V. (2005) Connexin-based gap junction hemichannels gating mechanisms. Biochim. Biophys. Acta 1711, 215224.
  • Salio C., Doly S., Fischer J., Franzoni M. F. and Conrath M. (2002) Neuronal and astrocytic localization of the cannabinoid receptor-1 in the dorsal horn of the rat spinal cord. Neurosci. Lett. 329, 1316.
  • Scemes E., Suadicani S. O., Dahl G. and Spray D. C. (2007) Connexin and pannexin mediated cell-cell Communications. Neuron Glia Biol. 3, 199208.
  • Sheng W. S., Hu S., Min X., Cabral G. A., Lokensgard J. R. and Peterson P. K. (2005) Synthetic cannabinoids WIN 55,212-2 inhibits generation of inflammatory mediators by IL-1beta stimulated human astrocytes. Glia 49, 211219.
  • Sofroniew M. V. (2005) Reactive astrocytes in neural repair and protection. The Neuroscientist 11, 400407.
  • Spray D. C., Ye Z. C. and Ransom B. R. (2006) Functional connexin “hemichannels”: a critical appraisal. Glia 54, 758773.
  • Stella N. (2004) Cannabinoid signaling in glial cells. Glia 48, 267277.
  • Stella N. (2009) Endocannabinoid signaling in microglial cells. Neuropharmacology 56, 244253.
  • Streit W. J., Walter S. A. and Pennell N. A. (1999) Reactive microgliosis. Prog. Neurobiol. 57, 563581.
  • Venance L., Piomelli D., Glowinski J. and Giaume C. (1995) Inhibition by anandamide of gap junctions and intercellular calcium signalling in striatal astrocytes. Nature 376, 590594.
  • Vignali M., Benfenati V., Caprini M., Anderova M., Nobile M. and Ferroni S. (2009) The endocannabinoid anandamide inhibits potassium conductance in rat cortical astrocytes. Glia 57, 791806.
  • Waksman Y., Olson J. M., Carliste S. J. and Cabral G. A. (1999) The central cannabinoid receptor (CB1) mediates inhibition of nitric oxide production by rat microglial cells. J. Pharmacol. Exp. Ther. 288, 13571366.
  • Walter L. and Stella N. (2003) Endothelin-1 increases 2-arachidonyol glycerol (2-AG) production in astrocytes. Glia 44, 8590.
  • Walter L. and Stella N. (2004) Cannabinoids and neuroinflammation. Br. J. Pharmacol. 141, 775785.
  • Ye Z. C., Wyeth M. S., Baltan-Tekkok S. and Ransom B. R. (2003) Functional hemichannels in astrocytes: a novel mechanism of glutamate release. J. Neurosci. 23, 35883596.
  • Zurier R. B. (2003) Prospects for cannabinoids as anti-inflammatory agents. J. Cell. Biochem. 88, 462466.