|Home | About | Journals | Submit | Contact Us | Français|
Interferon-γ (IFN-γ) has potent antiviral activity in neurons which is affected by the production of nitric oxide (NO). This study examines the interactions between cannabinoid receptor-1 (CB1), IFNγ–induced pathways, and inhibition of vesicular stomatitis virus (VSV) replication in neuronal cells. CB1 is abundantly expressed in neurons of the CNS and the NB41A3 neuroblastoma cell line. CB1 activation of NB41A3cells by the synthetic cannabinoid, WIN55,212-2, is associated with an inhibition of Ca2+ mobilization, leading to diminished nitric oxide synthase (NOS)-1 activity and the production of NO, in vitro. This ultimately results in antagonism of IFN-γ–mediated antiviral activity and enhanced viral replication. Therefore, activation of cells expressing CB1 by endogenous (or exogenous) ligands may contribute to decreased inflammation and to increased viral replication in neurons and disease in the CNS.
Intranasal administration of vesicular stomatitis virus (VSV) to immunocompetent mice results in an acute infection that can lead to host death (33,34,63). The expression of the neuronal isoform of nitric oxide synthase (NOS-1) is critical for host survival and elimination of VSV infection from infected neurons (44). NOS-1 activity is regulated by many factors, including substrate, tetrahydrobiopterin, protein inhibitor of NOS-1, and cellular calcium-activated kinases. (9,18,21,32,39,57). NOS-1 accumulates in neurons exposed to inflammatory cytokines such as interleukin (IL)-12 and IFN-γ (13,16,17,43,79). This study was performed to investigate if a lipid mediator, naturally produced in the CNS, which is capable of binding to neurons, was able to alter the cell-autonomous antiviral response of neurons to IFN-γ.
Cannabinoids derive from three sources of lipid molecules: endogenously synthesized endocannabinoids, compounds from some plants (e.g., marijuana), or pharmaceutically manufactured chemicals. Two receptors that are expressed on distinct cell types have been well described: CB1 expressed by neurons, and CB2 expressed by cells of the reticuloendothelial system including microglia (48,60,61,73,74). The functions of these receptors are distinct, although the same signaling pathways are used: the serpentine 7-transmembrane receptor is G-protein coupled, negatively regulates Ca2+ channels resulting in inhibition of Ca2+ release, activates a signal transduction cascade of Raf-1, MEK, and ERK, as well as adenyl cyclase, ultimately activating protein kinase A (5,22,50,51).
Signaling through the CB1 receptor is associated with sensory responses (77) including analgesia (56), hypothermia (12,76), immobility, euphoria (47), and hyperphagia (12,30). In contrast, activation of cells expressing the CB2 receptor has indicated that CB2 is a negative regulator of monocyte and microglial activation (23,48), and hence is immuno-dampening. Many agonists or antagonists are able to bind and activate both, but selective receptor agonists target either immune or neuronal responses.
Endocannabinoids may be released locally and have paracrine effects in the CNS (4) or may have affects on infiltrating inflammatory cells (24,68,73,74); alternatively, ingested or administered drugs may have systemic effects (19). Determining the contribution of the distinct receptors is potentially ambiguous when you consider both the ability of unselective ligands to trigger both receptors and the regulation of cell-autonomous innate immune responses to viral infections of neurons in the CNS. To begin to dissect this, in this study we have examined the in vitro effects of synthetic cannabinoids and an antagonist on the ability of IFN-γ to inhibit VSV replication in neuronal cells. We have found that cannabinoid signaling suppresses the production of NO and antagonizes the beneficial antiviral activity of IFN-γ, resulting in enhanced production of viral progeny.
NB41A3, HN9e, N18, L929, and RAW 264.7cells were purchased from ATCC, and were maintained as previously described (37). All tissue culture reagents were obtained from Mediatech (Manassas, VA). VSV, Indiana serotype, San Juan strain, originally obtained from Alice S. Huang (The Children's Hospital, Boston, MA) was used in all viral studies. Recombinant murine IFN-γ was purchased from R&D Systems (Minneapolis, MN) and used at a concentration of 20ng/mL. Agonists WIN55,212-2 (WIN55) and anandamide (AEA), and the CB1 antagonist AM251 were purchased from Tocris USA (Ellisville, MO). U0126, a MEK1/2 inhibitor, was purchased from Cell Signaling Technology (Danvers, MA). Preliminary experiments were performed to optimize the doses of all drugs; drug ranges were initially based on published studies. Propridium iodide (PI) and 7-nitroindazole (7-NI) were purchased from Sigma. CalciumGreen-AM, was purchased from Invitrogen Molecular Probes (Carlsbad, CA).
Cell viability was determined by the Cell Titer 96 Aqueous Non-Radioactive Cell Proliferation colorimetric assay (Promega, Fitchburg, WI) following cannabinoid agonist and antagonist treatments. Triplicate samples of 1×104 cells were seeded per time point with 100μL media in 96-well flat-bottom plates, then 20μl of freshly mixed MTS reagent was added to each well and incubated for 4h. Absorbance was determined at 490nm using a microplate reader (Model 550; Bio-Rad, Hercules, CA), and the results expressed as the mean absorbance of triplicate experiments±SE. Based on viability assays (Fig. 1) we concluded that treatment of NB41A3cells with up to 10μM WIN55 is not harmful.
mRNA levels were determined by multiplex RT-PCR using the Qiagen One-Step RT-PCR kit (Germantown, MD). Total RNA was used in each reaction. The following program was used to carry out the reaction: lid preheated to 105°C then 50°C for 30min, 95°C for 15min, 39 cycles of 94°C for 1min, 55°C for 1min, 72°C for 1min, then 72°C for 10min, and 4°C hold. Products were resolved by 1.5% agarose gel electrophoresis, stained using ethidium bromide, and visualized using a UV transilluminator. Results were digitally photographed using a Kodak DC290 camera controlled by Kodak 1D analytical software (Eastman Kodak, Rochester, NY). Primers used for RT-PCR are shown in Table 1.
Viral titers were determined in triplicate using serially diluted culture supernatants on monolayers of L929cells as previously described (43).
Cell lysates were prepared in the presence of both protease and phosphatase inhibitors, protein concentrations were determined using the Bio-Rad DC protein assay, and samples were boiled for 5min in SDS sample buffer with 2-mercaptoethanol. Samples were resolved by 10%, 12%, or 14% SDS-PAGE electrophoresis, transferred to Transblot™ membranes (Bio-Rad), and incubated for either 1h at ambient temperature or overnight at 4°C with primary antibodies, and then probed with horseradish peroxidase (HRP)-conjugated secondary antibodies for 1h at ambient temperature. Membranes were visualized with Pierce PicoWest Enhanced™ Chemiluminescent Reagents (Pierce Immnuochemical, Rockford, IL, USA). Film exposure was on Kodak BioMax™ (Eastman Kodak, Rochester, NY, USA) film for 3min. Protein band densities were quantitated using UN-SCAN-IT™ software.
Primary antibody reagents included rabbit anti-N-terminus of CB1 (Invitrogen BioSource; 1:1000). Mouse mAb anti-GAPDH (ABCAM, 1:100,000), sheep anti-VSV (1:100,000; a generous gift from Dr. Alice S. Huang), rabbit anti-ERK (p42/p44 MAP kinase; 1:1000), and mAb anti-phospho-ERK (1:1000 for Western blot and 1:50 for immunohistochemistry) were purchased from Cell Signaling. Mouse mAb 5A7, anti-IL-12Rβ2 was the generous gift of Jerome Ritz, Dana-Farber Cancer Institute (75). HRP-conjugated anti-mouse and anti-rabbit secondary antibodies (1:2000) were purchased from Vector (Vector Laboratories, Burlingame, CA, USA) and Zymax (Zymax: Spectrum Laboratories, Rancho Dominguez, CA, USA), and HRP-conjugated donkey anti-sheep (1:4000) was purchased from Jackson ImmunoResearch (West Grove, PA).
Nitrite levels were measured by the Nitrite/Nitrate Colorimetric Assay Kit from Cayman Chemical (Ann Arbor, MI). NB41A3cells were treated with IFN-γ, and either 10μM WIN55 or 400μM 7-NI 24h later. Supernatants were collected and then incubated for 2h at room temperature with nitrate reductase. Greiss reagent was added to triplicate samples, and the product was read with a Bio-Rad Microplate model 550 Reader at 540nm.
NB41A3cells (50,000cells/well incubated overnight) were treated for 1, 5, and 15min with either 10μM WIN55 or 400μM N-methyl D-aspartate (Sigma), or mock-treated for 1, 5, and 15min. Before fixation, cells were incubated with 10μM CalciumGreen-AM (Invitrogen Molecular Probes) for 1h. CalciumGreen-AM is a cell permeant ester that is cleaved and sequestered in the cytosol. Cell nuclei were labeled with PI. Cells were assessed by confocal microscopy on a Leica instrument (LeicaMicrosystems, Inc., Bannockburn, IL, USA).
In some experiments, NB41A3cells were cultured on glass slides coated with poly-D-lysine. Twenty-four hours after WIN55 treatment, cells were fixed in 4% paraformaldehyde, and then permeabilized with 0.1% Triton X-100. In other experiments, NB41A3cells were plated on fibronectin-coated Aclar film and treated for 1 hour with either WIN55 or AEA. Cells were then fixed and stained with primary antibodies. Primary rabbit anti-CB1 N-terminus, mAb anti-IL-12Rβ2, or mAb anti-phospho-ERK were utilized. Following treatment with secondary goat anti-rabbit AlexaFluor-488 (Invitrogen Molecular Probes) or secondary rabbit anti-mouse AlexaFluor-546, cells were covered with Vectashield Hardset (Vector Labs, Burlingame, CA) with DAPI and/or propidium iodide, and analyzed by confocal microscopy on a Leica confocal microscope.
Wells of NB41A3cells were treated with 10μM WIN55 or 10μM AEA for 1, 5, 5, 30, and 60min. In some samples, 10μM AM251 (a CB1 antagonist) or 10μM U0126 (a MEK1/2 antagonist) were added 1h prior to WIN55 administration to inhibit ERK phosphorylation. Cell lysates were isolated for Western blot in the presence of protease and phosphatase inhibitors (Calbiochem). Techniques described above in the sections on Western blot and immunohistochemistry were utilized to visualize phospho-ERK1/2 and total ERK1/2 in NB41A3cells.
To determine if there was a significant difference among the groups, means±SEM of each group were compared using a parametric ANOVA. The results were considered significant for a p value<0.05.
NB41A3cells, a mouse neuroblastoma line, has been used extensively for studies of VSV infection, of responses to cytokines, and of dynamics of NOS-1 activity (8,13–15,36,37,44,78,79). To determine if the NB41A3cell line was appropriate for our experimental system, the reverse transcription polymerase chain reaction (RT-PCR) was performed to detect CB1 or CB2 mRNA transcripts. NB41A3, RAW264.7, N18, HN9e, and L929cells were used for RNA extraction. As expected, NB41A3, HN9e, and N18cells showed significant expression of CB1 mRNA transcripts, while RAW264.7 and L929cells did not (Fig. 2A). RAW264.7cells served as a positive control for CB2 mRNA expression (not shown). NB41A3cells did not express detectable levels of CB2 mRNA transcripts.
In an effort to determine if the CB1 mRNA transcripts in NB41A3cells were translated into functional receptor proteins we examined the CB1 protein and localization by fluorescence microscopy and Western blot analysis. Confocal microscopy indicated that CB1 protein was localized to the plasma membrane and vesicles within NB41A3cells (Fig. 2B and C). Western blot experiments showed CB1 protein in whole-cell lysates. CB1 was not degraded following treatment of NB41A3cells with the receptor agonist WIN55 (Fig. 2D). These data confirm that CB1 receptor protein is expressed by the NB41A3cell line.
We used immunofluorescence and confocal microscopy to examine the distribution of the CB1 receptor following agonist treatment. NB41A3cells were plated on fibronectin-coated Aclar film and treated for 1h with either a synthetic cannabinoid (WIN55), or the endocannabinoid AEA. Cells were then fixed and stained using primary antibodies against CB1 and IL-12Rβ2, expressed on the plasma membranes of neuronal cells (37), as a positive control. As shown in Fig. 3, there was no evidence of endocytosis of the receptor CB1 receptor following treatment.
Instead, we observed significant morphological changes in WIN55- and AEA-treated cells when compared to untreated cells. This change in morphology was not associated with apoptosis. Furthermore, the CB1 protein was mobilized along the plasma membrane (Fig. 3), into patches and caps. We interpret these data to be consistent with the concept that CB1 ligand-receptor binding leads to cytoskeletal rearrangement, which results in change in the cell's shape and receptor distribution into patches; this is indicative of functional activation of the CB1 receptor by its ligand.
One of the major signal transduction pathways CB1 uses is the mitogen-activated protein kinase (MAP kinase) cascade. We treated NB41A3cells with either WIN55 or AEA for 1, 5, 15, 30, or 60min. Whole-cell lysates were collected and Western blots were performed using anti-p42/44 (ERK) and anti-phospho-p42/22 (pERK). A significant increase in pERK was seen at 1 and 5min after addition of the agonist, and to some extent at the 15-min time point, in both WIN55- (not shown) and AEA-treated samples (Fig. 4A). These data indicate that in neuronal cultures, CB1 signals through the MAP kinase cascade.
In order to further assess CB1 activation of MAP kinases, AM251 (a CB1 receptor antagonist) or U2106 (a MEK-1/2 inhibitor) was added to NB41A3 cultures before the addition of WIN55. MEK is upstream from ERK in the MAP kinase cascade. When these protein samples were analyzed by Western blot, the induction of pERK was inhibited by both AM251 and U2106 to levels below the level of detection both at 1min and at 15min after WIN55 stimulation (Fig. 4B). We interpret these data to indicate that AM251 blocked WIN55 binding to CB1, and that inhibition of the upstream enzyme, MEK, by U2106, prevented ERK phosphorylation. In addition, these data are consistent with the interpretation that the response to CB1 agonist is not delayed in the presence of antagonists, but rather is prevented. Therefore WIN did not successfully signal (induce phosphorylation of ERK1/2) in cells treated with either a receptor or kinase antagonist.
Generally, when ERK is phosphorylated, it undergoes nuclear translocation and acts as a transcription factor (70). To confirm and extend the Western blot analysis we examined the cellular location of pERK by confocal microscopy. NB41A3cells were cultured on fibronectin-coated Aclar film. Cells were stimulated with WIN55 for 1, 15, or 30min and were then fixed, and stained with anti-ERK and anti-pERK antibodies. Our data from these immunofluorescence experiments are consistent with the results from the Western blot assays (Fig. 4). The 3D confocal images indicate that pERK translocated to the nucleus as early as 1min post-WIN55 treatment (Fig. 5). Taken together, these data provide strong evidence for CB1 signaling through the MAP kinase pathway in neuroblastoma cells. The activation of this pathway may be central to CB1's many broad effects including its role in immunomodulation.
Neuronal signaling relies on stringent control of intracellular Ca2+ levels through a variety of mechanisms (e.g., NMDA receptors and GQ proteins). Previous reports have shown that under a variety of physiological conditions CB1 ligands modulate Ca2+ influx into neurons (26,45). CalciumGreen-AM is a cell permeant ester that is cleaved in the cytosol and cannot transverse the membrane once inside the cell. Free Ca2+ binds CalciumGreen-AM, inducing fluorescence at 488nm. In an effort to determine if CB1 activation had an effect on intracellular Ca2+ levels in this neuronal cell line, NB41A3cells were incubated with 10μM CalciumGreen-AM for 1h and then treated with 10μM WIN55 or 400μM N-methyl D-aspartate (NMDA) for 1, 5, or 15 minutes. The cells were then fixed and stained with propidium iodide to indicate nuclei. The data shown in Fig. 6 indicate that NMDA rapidly induced an influx of Ca2+ into the cytosol, shown as bright green staining at 1min after addition of NMDA, fading over the observation period. WIN55 binding of CB1 does not increase intracellular Ca2+ levels, and appear very similar to untreated neuronal cells (top far right panel), which are detected only by virtue of nuclear staining.
In neurons, the antiviral effect of IFN-γ involves an increase in NO via increased neosynthesis and stability of NOS-1 (14,79). NOS-1 is a Ca2+- and calmodulin-dependent enzyme (9,21,57). Since WIN55 may limit intracellular Ca2+ levels (Fig. 6), one potential mechanism through which it might be antagonizing the IFN-γ antiviral response is by inhibiting the synthesis of NO. In order to test this hypothesis, we examined NO production, by measuring nitrite levels in supernatants of NB41A3cells treated with both IFN-γ and WIN55, compared to cells treated solely with IFN-γ. Treatment with IFN-γ and WIN55 results in diminished production of NO (Fig. 7) when compared to neuronal cells that were treated only with IFN-γ.
We have previously shown that IFN-γ induces a potent antiviral response in neurons, and leads to a~100-fold reduction in VSV viral titers over untreated cells. In order to examine if CB1 activation modified the IFN-γ–induced antiviral response, NB41A3cells were pretreated with IFN-γ for 24h, followed by a 24-h treatment with WIN55, then the cells were infected with VSV at a multiplicity of infection of 0.1 for 7h. Supernatants were collected, and viral titers were analyzed by plaque assay. Although WIN55 treatment itself had no discernible effect on the neuronal response to viral infection (p=0.42), we found that when the NB41A3cells were pretreated with IFN-γ and subsequently treated with WIN55, there was a 10-fold increase in infectious virus released, compared to NB41A3cells that were treated solely with IFN-γ (Fig. 8A) (p=0.00053). While CB1 activation antagonized the IFN-γ–induced antiviral response, it does not completely abrogate the antiviral activity of IFN-γ.
IFN-γ treatment of neurons inhibits viral replication by inhibiting VSV protein synthesis in infected NB41A3cells (15,43). In order to elucidate the mechanism by which CB1 partially antagonizes the IFN-γ response, NB41A3cells were pretreated with IFN-γ followed by WIN55. The cells were subsequently infected with VSV at a multiplicity of infection of 1.0. At 5h post-infection, whole-cell lysates were collected and protein samples were analyzed by Western blot (Fig. 8B). These data are consistent with the viral replication assays. WIN55 treatment on its own did not exert any detectable changes in viral protein concentrations (p=0.539), while IFN-γ treatment reduced synthesis of all five viral proteins (p=0.00969). Viral protein production was intermediate in cells that were pretreated with IFN-γ followed by WIN55 treatment (p=0.0039), a result of partial antagonism of the IFN-γ–induced viral protein synthesis inhibition (Fig. 8C). These data indicate that CB1 agonists antagonize the IFN-γ–induced antiviral response in neuronal cells in vitro.
Treatment of neuronal cells expressing the CB1 receptor (Fig. 2) with WIN55 results in patching and capping of the receptor (Fig. 3), activation of the MAP kinase cascade (Fig. 4), phosphorylation of ERK (Fig. 5), and no increase in cellular Ca2+ concentrations (Fig. 6). Neuronal cells treated with WIN55 produce less nitrite, the stable end-product of NOS-1 (Fig. 7) and, most importantly, WIN55 antagonizes the inhibition of VSV replication, which is associated with IFN-γ treatment (Fig. 8), resulting in 10-fold enhanced viral yields. These data are consistent with a model in which cannabinoid regulation of intracellular Ca2+ levels in neurons diminishes NOS-1 activity and NO production, and therefore suppresses the inhibition of VSV replication which would otherwise be effected by IFN-γ.
Endogenous cannabinoids are released following injury to the brain (50,62). Cannabinoids may contribute to neurogenesis by antagonizing NO production (41) and may be protective in ischemia (7,71) and excitotoxicity (28,42).
An indirect anti-inflammatory effect of cannabinoids has been associated with activation of the nuclear transcription factor PPAR family (10,59,72). CB2 receptor activation may lead to release of endogenous opioids, which inhibit inflammatory pain (35,64). Somewhat unexpectedly, the anti-nociceptive and anti-pyretic effects of acetaminophen may be due to binding CB1 receptors (2,66).
Cannabinoids have been shown to be neuroprotective in several inflammatory or neurodegenerative diseases including Alzheimer's disease, Parkinson's, experimental autoimmune allergic encephalomyelitis, stroke, and excitotoxicity (25,27,28,38,49,65). These lipid molecules may target microglial activation (6,60,68,69,73). WIN55 may also induce PGE2 (52). Thus, where inflammation contributes to pathology, cannabinoids may be beneficial.
In an infection model in which inflammation contributes to pathology (Theiler's murine encephalomyelitis virus infection [TMEV]), WIN55 treatment ameliorates clinical disease (20,53). Another potentially beneficial cannabinoid effect may be protection of BBB integrity during infections in which microglial NOS-2 (iNOS) is overactive, such as Borna virus infection (31). In both TMEV and Borna viral disease, the effect is probably attributable to CB2-expressing cells.
In contrast, Δ9-tetrahydrocannabinol treatment decreases host resistance to HSV-2 infection in both mice and guinea pigs (11,54), possibly by inhibiting host inflammatory immune responses against the virally-infected cells; like VSV, herpesviruses are sensitive to the antiviral effects of NO (1,40). However, cannabinoids may contribute to syncytia formation in HIV-E (58), which is associated with disease pathology; the mechanism of this phenomenon may be suppressing inhibition of viral gene expression. In the case of VSV, WIN55 treatment resulted in enhanced viral replication (Fig. 8).
We speculate that in models of infection in which the replication is sensitive to NO-associated inhibition [roughly half of viruses studied (3,29,46,67)], cannabinoids promote viral replication and disease. Replication may be enhanced by the paracrine release of cannabinoids in response to viral infection–associated neuronal injury (50,62). However, in those other viral infections in which inflammation is deleterious, cannabinoids are potentially beneficial.
Since cannabinoids are both widely used recreational drugs (55), and their receptors are therapeutic targets (60), understanding the impact of agonists and antagonists for the CB1 and CB2 receptors on the outcome of viral infections, especially those in the CNS, is important. Work is in progress to examine the in vivo effects and to distinguish between effects on cell autonomous pathways from inflammation-associated contributions of the CB2-bearing cells.