Cannabinoid receptor type 2 activation induces a microglial anti-inflammatory phenotype and reduces migration via MKP induction and ERK dephosphorylation

First, we determined the effects of JWH015 on MKP-1 and MKP-3 expression in LPS-stimulated microglia. LPS did not significantly change the levels of MKP-1 expression compared to the medium control group at the tested time points (15–60 min, Figures 1A). However, MKP-1 expression was significantly increased in LPS + JWH015 only at 15 min incubation time point compared to the 0 time point (the medium control group, 1.22 ± 0.04 of medium control group, p < 0.05; Figures 1A). This increased MKP-1 expression in LPS + JWH015 group was also significantly different from the LPS alone group at the same time point (15 min, 1.22 ± 0.04 vs. 1.04 ± 0.02 of medium control group respectively, p < 0.05, Figures 1A). LPS did not significantly change the levels of MKP-3 expression compared to the medium control group at the tested time points (Figures 1B). MKP-3 expression was significantly increased in LPS + JWH015 at 15 and 60 min incubation time points (1.45 ± 0.14 and 1.42 ± 18 of medium control group respectively, p < 0.05; Figures 1B). This increased MKP-3 expression in LPS + JWH015 group was also significantly different from the LPS alone group at the 15 min incubation time point (15 min, 1.45 ± 0.14 vs. 1 ± 0.07 of medium control group respectively, p < 0.05, Figures 1B).

MKP-1 has been shown to be inducible (for example by LPS) but not constitutively expressed in macrophages [30]. To test whether MKP-1 and MKP-3 are constitutively expressed in microglial cells we incubated primary microglia in serum free medium (DMEM) or in DMEM plus 10% fetal bovine serum (S+DMEM) for 1.5 or 20 h. We observed that primary microglial cells constitutively expressed MKP-1 and we found no differences in the DMEM and S+DMEM groups or 1.5 and 20 h incubation periods (1.5 hr DMEM vs. S+DMEM = 0.97 ± 0.1 vs. 0.98 ± 0.2; 20 hr DMEM vs. S+DMEM = 0.75 ± 0.02 vs. 0.72 ± 0.02). We also confirmed a constitutive expression of MKP-3 in microglia with no differences between DMEM and S+DMEM group or 1.5 and 20 h incubation periods (1.5 hr DMEM vs. S+DMEM = 1.3 ± 0.04 vs. 1.4 ± 0.3; 20 hr DMEM vs. S+DMEM = 1.1 ± 0.03 vs. 1.1 ± 0.02). These results also suggest that our microglial cells are in a quiescent rather than a primed state prior to our treatments.

MKP-1 decreases MAPKs, such as p38, c-Jun N-terminal kinase (JNK) or ERK1/2 [28, 29]. MKP-3 is a specific and major negative regulators of the ERK pathway [25, 26]. We investigated the functional implications of MKP-1 and MKP-3 modulation by studying ERK1/2 phosphorylation. LPS did not induce any significant change in t-ERK1/2 at any time point studied and JWH015 did not modify this (Figure 2A). The expression of t-ERK1/2 was not significantly different at any incubation time point in the LPS + JWH015 group compared to the medium control group, except at the 120 min time point in LPS + JWH015 compared to medium control group (Figure 2A). No differences were found in t-ERK1/2 expression between LPS alone and LPS + JWH015 groups in any of the incubation time points (Figure 2A).

LPS induced a significant increase in p-ERK1 expression at 30 min (1.7 ± 0.11 of medium control group, p < 0.05; Figures 2A and 2B), and in p-ERK2 at the 30 min incubation time point compared to the medium control group (1.85 ± 0.23 of medium control group, p < 0.05; Figures 2A and 2C). The expression of p-ERK1/2 was not significantly different in the LPS + JWH015 group compared to the medium control group at any of the incubation time points. Moreover, p-ERK1/2 expression in LPS+JWH015 was significantly reduced at the 30 min incubation time point compared to the LPS alone group (p-ERK1 = 1.7 ± 0.11 vs. 1.17 ± 0.08 of medium control group respectively, p-ERK2 = 1.85 ± 0.23 vs. 1.11 ± 0.1 of medium control group respectively, p < 0.05; Figures 2A, B and 2C).

ERK mediates the LPS-induced microglial pro-inflammatory factor, TNF [21, 31, 32]. We further investigated the functional implications of p-ERK modulation by studying TNF expression. LPS induced an increase in TNF (17 kD [monomer] and 34 kD [dimer]) expression at the 60 min incubation time point compared to the medium control (34 kD = 1.15 ± 0.03, and 17 kD = 1.43 ± 0.03 of medium control group, p < 0.05, Figures 2A, D and 2E). TNF expression was not significantly different between the LPS + JWH015 group compared to the medium control group. Furthermore, 34 kD TNF expression was significantly reduced in LPS + JWH015 group compared to the LPS alone group at the 60 and 120 min incubation time points (0.93 ± 0.04 vs. 1.14 ± 0.3 and 0.91 ± 0.05 vs. 1.05 ± 0.1 of medium control group 60 and 120 min respectively, p < 0.05; Figures 2A and 2D) and 17 kD TNF expression was significantly reduced in LPS + JWH015 group compared to the LPS alone group at the 60 min incubation time point (0.90 ± 0.14 vs. 1.43 ± 0.03 of medium control group 60 min, p < 0.05; Figure 2A and 2E). We confirmed the dependence of TNF expression on ERK phosphorylation in LPS-stimulated microglia by inhibiting MAP/ERK kinase (MEK) with UO126 (1 μM). TNF expression was reduced in the LPS + UO126 group compared to the LPS alone group at the 60 min incubation time point (40.6 ± 12.1% inhibition, p < 0.05, data not shown). UO126 completely inhibited p-ERK in the LPS + UO126 group compared to the LPS alone group at the 30 and 60 min incubation time points (data not shown).

In these studies we challenged JWH015's effects on MKP-1 and MKP-3 using three different compounds claimed to be MKP-1 inhibitors. Cells were pre-incubated with or without PSI2106, Ro-31-8220 or triptolide (1 and 10 μM) for 30 min, then LPS with or without JWH015 (1 μM) was added for 15 min. The expression of both MKP-1 and MKP-3 were significantly enhanced in LPS+JWH015 group (p < 0.05) as shown previously. MKP-1 expression induced by JWH015 in LPS-stimulated microglia was not modified by 1 μM PSI2106 or Ro-31-8220 (2 ± 0.13 vs. 1.95 ± 0.02 and 2.02 ± 0.3 of medium control, LPS+JWH015 vs. LPS+JWH015+PSI2106 or Ro-31-8220 respectively), but it was inhibited following treatment with the dose of 10 μM (2 ± 0.13 vs. 1.6 ± 0.02 and 1.5 ± 0.05 of medium control, LPS+JWH015 vs. LPS+JWH015+PSI2106 or Ro-31-8220 respectively, p < 0.05, Figures 3A and 3C). PSI2106 or Ro-31-8220 did not modify JWH015's effects on MKP-3 at 1 μM (1.43 ± 0.13 vs 1.58 ± 0.09 and 1.64 ± 0.16 of medium control, LPS+JWH015 vs. LPS+JWH015+PSI2106 or Ro-31-8220 respectively) or 10 μM concentration (Figures 3D and 3F). However, triptolide (only at 10 μM) completely inhibited the JWH015-induced MKP-1 (Figures 3A and 3B) and MKP-3 (Figures 3D and 3E) expression in LPS-stimulated microglia. Triptolide at 1 μM did not inhibit JWH015-induced MKP-1 (2 ± 0.13 vs. 1.93 ± 0.04 of medium control, LPS+JWH015 vs. LPS+JWH015+triptolide respectively) or MKP-3 (1.43 ± 0.13 vs 1.55 ± 0.17 of medium control, LPS+JWH015 vs. LPS+JWH015+triptolide respectively) in LPS-treated microglia.

We evaluated the expression of p-ERK1/2 in the above mentioned groups to confirm the functional implications of the inhibition of MKP-1 and/or MKP-3. Triptolide (10 μM), which completely inhibited the induction of both MKP-1 and MKP-3 by JWH015, induced a significant increase in p-ERK1 (1.14 ± 0.02 of LPS + JWH015 control group, p < 0.05; Figures 4A and 4B) and p-ERK2 (1.12 ± 0.02 of LPS + JWH015 control group, p < 0.05; Figures 4A and 4B). Triptolide at 1 μM concentration did not significantly modify the expression of p-ERK1 (1.11 ± 0.06 of LPS + JWH015 control group) or p-ERK2 (1.09 ± 0.08 of LPS + JWH015 control group) in LPS plus JWH015-treated cells. Unexpectedly, neither PSI2106 nor Ro-31-8220 (which effectively blocked only the JWH015-induced MKP-1 but not MKP-3 expression) modified p-ERK1 (1 μM: 1.04 ± 0.05 and 0.99 ± 0.06 of LPS + JWH015 control group, LPS+JWH015+PSI2106 or Ro-31-8220 respectively; Figure 4A and 4C) or p-ERK2 (1.03 ± 0.06 and 0.99 ± 0.02 of LPS + JWH015 control group, LPS+JWH015+PSI2106 or Ro-31-8220 respectively; Figure 4A and 4C). These data together suggest that the induction of MKP-3 rather than MKP-1 is the mechanism by which JWH015 reduces p-ERK1/2 expression. LPS plus triptolide (10 μM) in the absence of JWH015, did not modify the expression of MKP-1, MKP-3, t-ERK and p-ERK when compared to the LPS alone group (data not shown). Since our results suggest that MKP-3 rather than MKP-1 is responsible for the p-ERK1/2 reduction induced by JWH015, we challenged the effects of JWH015 on MKP-3 expression with specific cannabinoid antagonists. The selective CBR2 antagonist, AM630, but not the CBR1 antagonist, AM281, significantly blocked the JWH015-induced MKP-3 in LPS stimulated microglia (1.12 ± 0.02 vs. 1.05 ± 0.02 of LPS alone group, LPS + JWH015 1 μM in absence or presence of AM630 1 μM respectively, p < 0.05; Figure 5).

We observed that JWH015 (1 μM) does not induce microglial chemotaxis (compared to medium as chemoattractant) in LPS-stimulated microglia (data not shown). To determine the effects of JWH015 on microglial migration, we performed a JWH015 dose-response pre-treatment experiments using LPS-stimulated microglia and ADP 10 μM as chemoattractant. JWH015 reduced the ADP-induced microglial migration in a significant and dose related fashion. The numbers of migrated LPS-stimulated cells in JWH015 0.1 and 1 μM groups were significantly lower compared to the LPS alone group (79 ± 10 cells in LPS alone group vs. 42 ± 7 and 40 ± 6 cells in LPS+JWH015 0.1 and 1 μM respectively, p < 0.05; Figure 6A). The effect observed in LPS+JWH015 1 μM group was blocked by the CBR2 selective antagonist AM630, but not by the CBR1 selective antagonist AM281 (45 ± 3 vs. 73 ± 4 or 54 ± 3 cells in LPS+JWH015+AM630 or AM281 respectively; Figure 6B). We have previously shown that minocycline (60 μM) reduced cell migration towards ADP (10 μM) using non-stimulated primary microglia [33]. Herein, we used minocycline as a positive control for our cell migration experiments using LPS-stimulated primary microglia. We observed that minocycline (60 μM) significantly reduced the migration of LPS-stimulated microglia towards ADP (85 ± 4 vs. 24 ± 4 cells LPS alone and LPS plus minocycline groups respectively, p < 0.05, data not shown).

Cell migration is a p-ERK dependent phenomenon [23, 24] that to our knowledge, has not yet been studied in LPS-activated primary microglia. We studied the effects of a selective MEK inhibitor, UO126, on LPS-stimulated microglial migration. The inhibition of MEK and subsequently inhibition of ERK phosphorylation with UO126 reduced in a significant manner the number of LPS-stimulated microglial cells that migrated towards ADP (77 ± 11 vs. 23 ± 7 cells, LPS alone and LPS + UO126 groups respectively, p < 0.05; Figure 6C). Together these data confirm that p-ERK1/2 regulates LPS-stimulated microglial migration and suggest that JWH015's effects on LPS-stimulated microglial cells are due to MKP-1/3 induction, and subsequent ERK dephosphorylation.

The following experiments were designed to test the functional implications of JWH015 effects on MKP-1/3 expression (which results in ERK dephosphorylation) and LPS-stimulated microglial migration. JWH015-reduced microglial migration was challenged with triptolide 10 μM, which completely blocked JWH015's effects on microglial MKP-1 and MKP-3 expression. The number of cells that migrated towards ADP was significantly lower in the LPS+JWH015 group compared to the LPS alone group (47 ± 3 vs. 67 ± 3 cells respectively, p < 0.05, Figures 7A and 7B). This effect of JWH015 was significantly inhibited by triptolide (67 ± 4 cells compared to LPS + JWH015, p < 0.05; Figures 7A and 7B). Triptolide in the presence of LPS did not affect microglial migration compared to the LPS alone control group (60 ± 2 cells, Figures 7A and 7B).

The CBR1 antagonist AM281 (1-(2,4-Dichlorophenyl)-5-(4-iodophenyl)-4-methyl-N-4-morpholinyl-1H-pyrazole-3-carboxamide), the CBR2 antagonist AM630 (6-Iodo-2-methyl-1- [2-(4-morpholinyl)ethyl]-1H-indol-3-yl](4-methoxyphenyl)methanone) and the CBR2 agonist JWH015 ((2-Methyl-1-propyl-1H-indol-3-yl)-1-naphthalenylmethanone) were obtained from Tocris, Ellisville, MI, USA. Drugs were diluted in dimethylsulfoxide (DMSO) and then in saline to a final concentration of 100 μM (5% DMSO). The final concentration of DMSO in cell culture with cannabinoid treatments was never higher than 0.05%, and this was used as the control group (medium control). JWH015 was chosen because we have shown in vivo that it selectively acts on spinal microglial CBR2 in L5 nerve transection rats and reduces spinal Iba-1 expression in association with its anti-allodynic effects [13]. We chose the specific antagonists because we also showed in the same study that AM630 blocked JWH015's effects, while AM281 did not. JWH015 at 5 μM concentration has been shown to effectively reduce pro-inflammatory factors in microglia with CBR2 selectivity [16]. We chose a maximum dose of 1 μM for JWH015 (and equimolar doses for the antagonists) to assure its specificity.

Since ERKs are the only known substrates of MEK, the MEK inhibitor, UO126 (Cell Signaling, Danvers, Massachusetts, USA) was used to produce ERK inactivation. PSI2106 [52] (kindly provided by Dr. Key Brummon, University of Pittsburgh, Pennsylvania, USA), Ro-31-8220 [29] and triptolide [28, 53, 54] were used as MKP-1 inhibitors. UO126, PSI2106 and triptolide (Sigma, St. Louis, Missouri, USA) were diluted in DMSO and then in saline to a final concentration of 100 μM (1% DMSO). The final concentration of DMSO in cell culture with these drug treatments was never higher than 0.01%, and this was used as control group (medium control). Ro-31-8220, minocycline, adenosine diphosphate (ADP) and lipopolysaccharide (LPS, 0111:B4 serotype) were diluted in saline (Sigma).

After approval by the Institutional Animal Care and Use Committee at Dartmouth College (Hanover, New Hampshire, USA), highly purified primary microglial cultures were prepared using postnatal day (P) 2–3 Harlan Sprague-Dawley pups (Indianapolis, Indiana, USA) as described previously [33, 47]. Briefly, pups were decapitated and the cerebral cortices were removed; meninges were dissected away; cortical tissue was minced with a sterile scalpel blade and digested with trypsin/EDTA 1× (Mediatech, Herndon, Virginia, USA) for 15 min at 37°C. The supernatant was discarded and 5 mL Dulbecco's modified Eagle's medium (DMEM; Mediatech) supplemented with 10% charcoal-stripped fetal bovine serum (FBS; Hyclone, Logan, Utah, USA), 1.1% GlutaMax (Gibco-Invitrogen, Carlsbad, California, USA), and 1% penicillin/streptomycin (100 U/mL penicillin and 100 μg/mL streptomycin; Mediatech) containing 2000 U DNase (Sigma) were added to the tissue on ice. The tissue was triturated with a 5 mL pipette. The tissue clumps were allowed to settle and the supernatant was removed to a sterile 50 mL conical tube on ice between triturations. Triturations were repeated until no tissue clumps were observed. The final volume was diluted to 25 mL with media and centrifuged at 310 g for 15 min. The supernatant was discarded and the cells resuspended in media. A small aliquot of cells was stained for trypan blue (Sigma) exclusion and cells were plated at 1 × 10⁶ cells per 75-cm² flask. Cultures were maintained at 37°C and 5% CO₂. Media was changed every 3–4 days. After 8 days in vitro (DIV 8), the flasks were confluent with astrocytes and microglia. Flasks were lightly shaken by hand for 1 min and the media containing microglia was removed and centrifuged at 310 g for 15 min. Cultures were found to be 99% microglia by staining with OX-42 antibody (generous gift from Dr William Hickey) a marker for the CR3/CD11b receptor.

Confluent DIV 8 primary microglia were shaken, counted and plated for at least 1.5 hr in serum free medium (SFM) at a density of 400 × 10³ cells/mL. The cells were treated with different drugs (see below) using a time course incubation (15, 30, 60 and 120 min). Following treatments, the culture plates were briefly centrifuged; supernatants were removed and 60 μL of 1× Laemmli buffer (Bio-Rad, Hercules, California, USA) containing 2-mercaptoethanol (ME) (Sigma) was added to each well. Protein expression was assessed using western blot analysis. Briefly, samples and standard protein markers were subjected to sodium dodecyl sulfate-polyacrylamide gel electrophoresis (10% gels; Bio-Rad) and transferred to polyvinylidene difluoride (Bio-Rad) membranes. Non-specific binding was blocked by incubation with 5% bovine serum albumin in Tris-buffered saline-Tween 20 (0.05%; Sigma) at 22°C, then membranes were incubated overnight or for 36–40 hr at 4°C with rabbit anti-phospho-ERK 44/42 (phospho-mitogen-activated protein kinase, p-ERK, 1:500; Cell Signaling), mouse anti-total-ERK 44/42 (total mitogen-activated protein kinase, t-ERK, 1:1000; Cell Signaling), rabbit anti-MKP-1 (mitogen-activated protein kinase phosphatase-1, MKP-1, 1:400; Santa Cruz, California, USA), goat anti-MKP-3 (mitogen-activated protein kinase phosphatase-3, MKP-3, 1:400; Santa Cruz, USA) or rabbit anti-TNF (1:1000; Peprotech, Rocky Hill, New Jersey, USA). The next day (or 36–40 hr), blots were incubated for 1 h at 22°C with goat anti-rabbit, mouse or donkey anti-goat horseradish peroxidase-conjugated secondary antibodies (1:3000; Pierce, Rockford, Illinois, USA), visualized with SuperSignal West Femto Maximum Sensitivity Substrate (Pierce) for 5 min and imaged using the Syngene G-Box (Synoptics, Frederick, Maryland, USA). After incubation with the first primary and secondary antibodies, three blots were incubated for 25 min at 37°C in stripping buffer. Blots were visualized using SuperSignal West Femto Maximum Sensitivity Substrate (Pierce) for 5 min and we observed that the antibodies were completely removed. Therefore, the same stripping procedure was used to re-probe with another primary and secondary antibodies. This methodology allowed us to observe the effects of the drug treatments in different proteins from the same treated cells. Finally, blots were subsequently stripped and re-probed with mouse anti-beta-actin antibody (1:3000; Abcam, Cambridge, Massachusetts, USA), and this was used as the protein loading control. Band intensity was assessed using the analysis software package provided with the Syngene G-Box and data were quantified as relative intensity of band of interest divided by intensity of beta-actin. Normalization of p-ERK and t-ERK were also conducted against beta-actin, then p-ERK data were quantified as relative intensity of band divided by intensity of t-ERK. Data were expressed as relative intensity normalized to beta actin control and to the control group for each experiment ± SEM. Incubation of blots with only secondary antibodies did not show any protein band.

The following treatment groups were performed. Medium control (time 0) or LPS (5 ng/ml) alone groups were used as the control groups to compare the effects of selective microglial CBR2 activation on MKP-1/3 (n = 5–7), t-ERK (n = 5–7), p-ERK (n = 5–7) and TNF (n = 4) using a group with LPS + JWH015 (1 μM). To test the specificity of the CBR2 agonist (JWH015), LPS + JWH015 (1 μM) group was compared to LPS + JWH015 (1 μM) + AM281 or AM630 (1 μM, n = 9) groups. The dependence of TNF production on p-ERK has been previously shown [21, 32]. To confirm this in our preparation, we used the MEK inhibitor, UO126 (1 μM, n = 6) in LPS-stimulated cells. Additionally, to probe that CBR2 activation reduces p-ERK by inducing MKP-1 and/or MKP-3, we challenged the effects of JWH015 (1 μM) in the presence of LPS with different drugs claimed to be MKP-1 inhibitors. Thus, the LPS + JWH015 (1 μM) group was compared with LPS + JWH015 (1 μM) + Ro-31-8220, triptolide or PSI2106 (1–10 μM, n = 3). All data were normalized to control groups which were given a value of 1.

Our laboratory has previously characterized and optimized the incubation times and chemoattractants that allow optimum cell migration susceptible to pharmacological modulation [33, 47]. We used Costar Transwell^® plates (6.5 mm diameter insert, 8.0 μm pore size, polycarbonate membrane; Corning Inc., Corning, New York, USA). For this group of experiments, the bottom chamber of these plates contained 10 μM ADP, a potent microglial chemoattractant released from injured neurons [33, 41, 47, 55]. Confluent DIV 8 primary microglia were shaken, washed with PBS, counted using trypan blue, then placed in SFM (DMEM). Cells were resuspended at 100 × 10³ cells in 200 μL SFM and treated with different drugs. To assess the effects of cannabinoid or other pharmacological treatments on primary microglia, cells were pre-treated with drugs for 2 h in the presence of LPS (5 ng/ml). Then, cells were centrifuged, medium containing the drugs was discarded and fresh SFM was added. Cells were counted using trypan blue to insure survival post-treatment (> 95% viability).

Cells were added to the top chamber (100 × 10³ cells in 200 μL SFM) of a transwell plate with fibronectin-coated membranes with ADP (10 μM, 500 μL in SFM) in the bottom chamber. Cells were then allowed to migrate towards ADP for 1 or 2 h at 37°C and 5% CO₂. Following migration, the medium in the top chamber was aspirated and the membrane gently wiped with a cotton swab to remove the cells that did not migrate. The membranes were first rinsed with PBS, the cells were then fixed with 2% formaldehyde in PBS, permeabilized with 0.01% Triton X-100 (Sigma) in PBS, and finally stained with crystal violet (Sigma). The membranes were then dried and mounted on microscope slides. Images of nine random fields (20× objective) for each membrane were captured via a Q-Fired cooled CCD camera attached to an Olympus microscope and counted by hand with aid of SigmaScan Pro imaging analysis software. Counts for all nine fields were averaged to give a mean cell count for each membrane. All experiments were completed at least three times (n = 3).

To study the effects of CBR2 activation on cell migration in LPS-stimulated microglia we used the following groups: LPS + JWH015 (0.01–1 μM, n = 9). To test the specificity of the CBR2 agonist, we challenged its effects with specific CBR1 (AM281) and CBR2 (AM630) antagonists by using the following groups: LPS + JWH015 (1 μM) + AM281 (1 μM, n = 4) or AM630 (1 μM, n = 6). To test whether p-ERK is involved in microglial migration we used a specific MEK inhibitor, UO126 (1 μM, n = 6) in LPS-stimulated cells. We also used a positive control group using LPS-stimulated microglia + minocycline (60 μM, n = 6), a dose that our laboratory has shown to be effective in reducing microglial migration in non-stimulated microglia [33]. To further test whether JWH015's effects on MKP-1/3 were causatively linked with JWH015's effects on microglial migration, we used Triptolide (10 μM, 15 min pre-treatment, n = 6) in LPS-stimulated cells + JWH015 (1 μM) incubated for 2 hr in the migration well. It has been shown that JWH015 act as chemoattractant in human monocytes when used at 20 μM concentration in parallel to an induction in ERK phosphorylation. However, JWH015 does not induce chemotaxis at 5–10 μM concentrations, [56]. Therefore, we chose 1 μM as our highest JWH015 dose tested.