Activation of type-2 cannabinoid receptor inhibits neuroprotective and antiinflammatory actions of glucocorticoid receptor α: when one is better than two

Abstract

Endocannabinoids (eCBs) and glucocorticoids (GCs) are two distinct classes of signaling lipids that exert both neuroprotective and immunosuppressive effects; however, the possibility of an actual interaction of their receptors [i.e., type-2 cannabinoid (CB₂) and glucocorticoid receptor α (GRα), respectively] remains unexplored. Here, we demonstrate that the concomitant activation of CB₂ and GRα abolishes the neuroprotective effects induced by each receptor on central neurons and on glial cells in animal models of remote cell death. We also show that the ability of eCBs and GCs, used individually, to inhibit tumour necrosis factor-α (TNF-α) and interferon-γ (IFN-γ) production from activated human T lymphocytes is lost when CB₂ and GRα are activated simultaneously. In addition, signal transduction pathways triggered by concomitant activation of both receptors led to increased levels of GRβ, heat-shock proteins-70 and -90, and p-JNK, as well as to reduced levels of p-STAT6. These effects were reversed only by selectively antagonizing CB₂, but not GRα. Overall, our study demonstrates for the first time the existence of a CB₂-driven negative cross-talk between eCB and GC signaling in both rats and humans, thus paving the way to the possible therapeutic exploitation of CB₂ as a new target for chronic inflammatory and neurodegenerative diseases.

Electronic supplementary material

The online version of this article (doi:10.1007/s00018-012-1253-5) contains supplementary material, which is available to authorized users.

Introduction

Endocannabinoids (eCBs) and glucocorticoids (GCs) are two distinct classes of signaling lipids involved in the adaptive responses of the organism to stressful internal and/or environmental challenges. eCBs are local mediators produced on demand, which have been implicated in a wide variety of physiological processes, including nociception, memory and cognition, appetite control, and immunoregulation [1, 2]. The effects of eCBs are mainly mediated through the activation of two G-protein coupled cannabinoid receptors (CB₁ and CB₂), of which CB₁ has a widespread distribution and is mainly active in the whole brain, whereas CB₂ is mainly expressed in the immune and hematopoietic systems [3–5]. The role of CB₂ at peripheral sites is generally antiinflammatory in several chronic inflammatory and neurodegenerative diseases [6–8]. However, recent studies have also reported CB₂ expression in the brain, in both glial and neuronal cells [9, 10], where our group was the first to report a CB₂-dependent protective role in central neurons undergoing remote cell death [11]. GCs are a class of steroid hormones derived from the hypothalamic–pituitary–adrenal axis (HPA) and crucial for the regulation of basal and stress-related homeostasis [12, 13]. Their antiinflammatory effects are mediated by a ubiquitous intracellular glucocorticoid receptor (GR), which functions as a hormone-activated transcription factor of target genes [14]. GR presents two splicing variants (GRα and GRβ), of which GRα is the classic receptor located in the cytoplasm as part of hetero-oligomeric complexes that mainly contain heat shock proteins (hsp); it is translocated to the nucleus upon binding to GCs [14, 15]. Instead, GRβ does not bind to GCs but functions as a modulator of their responses either by acting as a dominant negative inhibitor of GRα-induced transcriptional activity or through a novel, intrinsic and GRα-independent transcriptional activity [16]. GCs and GRα are at the apex of a regulatory network that blocks several inflammatory pathways through genomic or non-genomic mechanisms [13]. Due to their distinct actions in many biological processes, GCs are currently the most widespread antiinflammatory drugs used for the treatment of several chronic immune-mediated and neuroinflammatory diseases [17, 18]. Although the benefits of GC therapy are derived from short-term vascular changes and limited immunosuppression, prolonged or high-dose GC therapy has multiple side effects, including hypertension, dyslipidemia and “GC-induced psychoses” [19]. Additionally, it leads to GC insensitivity or resistance [20]. A link between GCs and eCBs has been disclosed at stress-relevant synapses, where eCBs have been reported as critical regulators of the stress response through their ability to modulate the sensitivity and activation of the HPA axis, as well as by contributing to the process of stress habituation in a CB₁-dependent manner [21, 22]. In the present study, we thus investigated the novel possibility of a GRα/CB₂ interaction by co-administration of the selective GRα agonist methylprednisolone (MPSS) and the selective CB₂ agonist JWH-015 to an animal model of remote cell death as well as to activated human peripheral T-lymphocytes.

Materials and methods

Animals

Experiments were performed using 100 adult male Wistar rats (body weight 200–250 g; Harlan San Pietro al Natisolone, Udine, Italy): 60 were lesioned, and 40 were used as controls (Table 1). Animals were group-housed in standard cages and maintained under a 12-h light–dark cycle in an air-conditioned facility. The Italian Ministry of Health approved the experimental protocol in agreement with the guidelines of the European Communities Council Directive of 24 November 1986 (86/609/EEC) for the care and use of laboratory animals. All efforts were made to minimize the number of animals used and their suffering. For surgical procedures, the rats were deeply anesthetized by intraperitoneal injections of xylazine (Rompun, 10 mg/ml; Bayer) and tiletamine and zolepam (Zoletil 100, 50 mg/ml; Virbac, Italy), and then they were positioned in a stereotaxic apparatus. The skin of the skull was incised, and the occipital bone was drilled and removed. Subsequently, the dura mater was incised in order to expose the cerebellum, and hemicerebellectomy (HCb) was induced by removal of the right cerebellar hemisphere, as described previously [23]. Then, the wound was sutured, and the animals were returned to their cages. For the unlesioned group, surgery was interrupted after the dura lesion was made, and after suturing, the animals were returned to their cages.

Table 1.

Lesion and treatments in the different experimental groups

Group codes	n	Drug treatments for 7 days (i.p./once a day)
CTRL-saline	10	DMSO 5 % in saline
CTRL-MPSS	10	50 mg/kg
CTRL-JWH-015	10	3 mg/kg
CTRL-MPSS + JWH-015	10	MPSS: as above + JWH-015: as above
HCb-saline	15	DMSO 5 % in saline
HCb-MPSS	15	50 mg/kg
HCb-JWH-015	15	3 mg/kg
HCb-MPSS + JWH-015	15	MPSS: as above + JWH-015: as above

CTRL, Control; HCb, hemicerebellectomy; MPSS, methylprednisolone; i.p., intraperitoneal

Peripheral blood cell preparation and stimulation

Peripheral blood lymphocytes were isolated, after venous puncture, from healthy donors and were separated by density gradient over Ficoll-Hypaque (Pharmacia, Uppsala, Sweden), as previously reported [5]. Informed consent was obtained from all the subjects, and the experiments conformed to the principles set out in the WMA Declaration of Helsinki. Briefly, heparinized blood was diluted with one volume of Dulbecco’s phosphate-buffered saline (PBS) and gently layered over the Ficoll. Following centrifugation at 660g for 30 min, cells at the interface of the gradient were collected and washed twice with PBS, and then they were re-suspended in complete RPMI 1640 medium supplemented with 5 % heat-inactivated human serum. For the identification of CD4 and CD8 T lymphocytes, cells were stained with anti-CD3, anti-CD4 and anti-CD8. PBMCs were pre-treated with MPSS and JWH-015, alone or in combination, for 30 min and then were challenged with phorbol-myristate-13-acetate (PMA) and ionomycin (Iono) for 5 h in the presence of Brefeldin A in order to prevent exocytosis of cytokine-containing vesicles.

Drugs and treatments

For the in vivo studies the following GR agonist was used: MPSS (Solumedrol^®, Pharmacia, Nerviano, Italy; 50 mg/kg, i.p./daily for 7 days). Further, the following CB₂ receptor agonist and antagonist, respectively, were used: JWH-015 (3 mg/kg, i.p./daily for 7 days; Cayman Chemicals, Ann Arbor, MI, USA) and SR144528 (SR2; 3 mg/kg, i.p./daily for 7 days; Cayman Chemicals). All animals, either single or double treated, received only one intraperitoneal injection per day. In fact, to avoid stressful conditions due to multiple injections, in double-treated animals, each time, the needle was maintained in the peritoneum and the two solutions were injected separately by changing the syringe. Treatments in the different experimental groups are listed in Table 1.

For the in vitro studies the following GR agonist and antagonist were used: MPSS and RU-486 (Sigma Aldrich). Further, the following CB₂ receptor agonists were used: JWH-015 (Cayman Chemicals) and GP 1a (Tocris Bioscience, Bristol, UK). SR144528 (SR2) was used as CB₂ receptor antagonist. PBMC detailed treatments are reported in Table 2.

Table 2.

Treatments in the different experimental groups

Group codes	n	Drugs treatment for 7 days
CTRL	6	1 μg/ml Brefeldin A for 4 h
PMA/Iono	6	1 μg/ml Brefeldin A for 4 h + 1 μg/ml PMA and Ionomicin for 5 h
PMA/Iono-MPSS	6	1 μg/ml Brefeldin A for 4 h + 1 μg/ml PMA and Ionomicin + MPSS 10 μM for 5 h
PMA/Iono-JWH-015	6	1 μg/ml Brefeldin A for 4 h + 1 μg/ml PMA and Ionomicin + JWH-015 20 nM for 5 h
PMA/Iono-MPSS + JWH-015	6	1 μg/ml Brefeldin A for 4 h + 1 μg/ml PMA and Ionomicin + MPSS 10 μM + JWH-015 20 nM for 5 h
PMA/Iono-MPSS + JWH-015 + RU486	6	1 μg/ml Brefeldin A for 4 h + 1 μg/ml PMA and Ionomicin + MPSS 10 μM + JWH-015 20 nM + RU486 10 μM for 5 h
PMA/Iono-MPSS + JWH015 + SR144528	6	1 μg/ml Brefeldin A for 4 h + 1 μg/ml PMA and Ionomicin + MPSS 10 μM + JWH-015 20 nM + SR144528 1 μM for 5 h
PMA/Iono-GP 1a		1 μg/ml Brefeldin A for 4 h + 1 μg/ml PMA and Ionomicin + GP 1a 1 nM for 5 h
PMA/Iono-MPSS + GP 1a		1 μg/ml Brefeldin A for 4 h + 1 μg/ml PMA and Ionomicin + MPSS 10 μM + GP 1a 1 nM for 5 h

CTRL, control; PMA, phorbol-myristate-13-acetate; Iono, ionomycin; MPSS, methylprednisolone

Antibodies

The following primary antibodies were used: rabbit anti-GRα (1:500; Abcam, UK), rabbit anti-GRß (1:500; Abcam UK), rabbit anti-pSTAT6 (1:500; Abcam, UK), mouse anti-GFAP (MAB360; 1:500; Millipore, USA), mouse anti-CD11b (OX-42, 1:500; Abcam, UK), mouse anti-actin (1:5,000; Sigma Aldrich, Italy), mouse anti-pJNK (1:1,000; Cayman Chemical, USA), rabbit anti-hsp70 (1:200; Santa Cruz Biotechnology, USA), mouse anti-hsp90 (1:1,000; Santa Cruz Biotechnology, USA) and goat anti-cytochrome-c (cyt-c) (1:400; Santa Cruz Biotechnology, USA). Secondary Cy3-, Cy2- and Cy5-conjugated antibodies (1:200) were purchased from Jackson Immunoresearch (West Grove, PA, USA). Horseradish peroxidase-conjugated secondary antibodies (anti-rabbit, 1:5,000, or anti-mouse, 1:5,000) were purchased from Cell Signaling Technology (Boston, MA, USA).

Neurological evaluation

The Neurologic Severity Score (NSS) was used to evaluate neurological conditions in rats [11]. NSS is a composite of motor, sensory, reflex and balance tests in which, for each test, one point is awarded for the inability to perform or for the lack of a tested reflex, and zero points are awarded for success. An NSS of 18 indicates severe injury, whereas a score of zero signifies healthy, uninjured rats. We also calculated the ΔNSS as the difference between the NSS at 24 h after damage and the NSS at successive time points. ΔNSS reflects the progression of recovery. The NSS was evaluated at 24, 72 h, and 5 and 7 days after damage by an investigator who was blind to the experimental groups.

Histology and immunohistochemistry

Animals were perfused transcardially, 7 days after surgery, with 250 ml of saline, followed by 250 ml of 4 % paraformaldehyde under anesthesia that was induced by intraperitoneal injections of Rompun (xylazine, 20 mg/ml, 0.5 ml/kg body weight) and Zoletil (tiletamine and zolazepam, 100 mg/ml, 0.5 ml/kg body weight). Each brain was removed immediately, post fixed in the same paraformaldehyde solution for 2 h, and after three washes in PBS was transferred to 30 % sucrose solution at 4 °C until it sank. Each brainstem was cut into four series of 40-μm-thick transverse sections using a cryostat, and slices were collected in PBS. For each animal, a series of sections involving pontine nuclei (Pn) were processed for immunohistochemical studies. Following incubation with a solution of primary antibodies, the sections were incubated for 2 h at room temperature with specific secondary antibodies. Sections were examined under a confocal laser scanning microscope (Leica SP5, Leica Microsystems, Wetzlar, Germany) equipped with four laser lines: violet diode emitting at 405 nm (for DAPI), argon emitting at 488 nm, and helium/neon emitting at 543 and 633 nm.

Qualitative and quantitative analyses

Qualitative and quantitative observations were limited to the Pn of the experimental side that was projecting to the lesioned hemicerebellum. Using the Stereo Investigator System (MicroBrightField Europe e.K., Magdeburg, Germany), an optical fractionator, stereological design, was applied to obtain unbiased estimates of total Nissl-stained cells. A stack of MAC 5,000 controller modules (Ludl Electronic Products Ltd., Hawthorne, NY, USA) was configured to interface an Olympus BX 50 microscope with a motorized stage and a HV-C20 Hitachi color digital camera with a Pentium II PC workstation. A three-dimensional optical dissector counting probe (x, y, z dimension of 30 × 30 × 10 μm, respectively) was applied. Five sections for each specimen were analyzed, and Pn was outlined using the 4× objective, while the 100× oil immersion objective was used for marking the neuronal cells. The total Pn cell number was estimated according to the formula:

where ∑Q represents the total number of neurons counted in all optically sampled fields of the Pn, ssf is the section sampling fraction, asf is the area sampling fraction, and tsf is the thickness sampling fraction.

To assess cyt-c expression within the Pn, quantitative analyses were performed off-line on confocal images acquired through the 20× objective at the 0.07 zoom factor. Quantitative data were obtained by adopting a different sampling strategy. Three digital square frames (200 × 200 μm) were placed at a regular distance to sample the entire medio-lateral extent of the Pn. All quantitative analyses were conducted blinded to the animal’s experimental group identity.

Cell lysate preparation and Western blot analyses

Rats (n = 5/group) were deeply anesthetized with intraperitoneal injections of Rompun (xylazine, 20 mg/ml, 0.5 ml/kg body weight) and Zoletil (tiletamine and zolazepam, 100 mg/ml, 0.5 ml/kg body weight), and they were killed by decapitation. The brains were dissected, and the Pn were isolated. Both Pn and PBMC were homogenized, and proteins were extracted in RIPA buffer (PBS supplemented with saline, 1 % Nonidet P-40, 0.5 % sodium deoxycholate, 0.1 % SDS and 0.5 M phenylmethylsulfonyl fluoride, 10 lg/ml leupeptin) for 30 min on ice, and then they were centrifuged for 10 min at 4 °C (14,000 rpm). Supernatants were collected, and the protein content was quantified by Bradford’s colorimetric assay (Bio-Rad, Milan, Italy). Each protein sample was separated by SDS–polyacrylamide gel electrophoresis and transferred to a nitrocellulose membrane. Membranes were saturated with 5 % dried nonfat milk and incubated overnight with specific primary antibodies. Membranes were then incubated with the appropriate horseradish peroxidase-conjugated secondary antibodies. Immunoreactive bands were detected using an enhanced chemiluminescence kit (ECL; Amersham Biosciences).

Mitochondrial and cytosolic fractions

Pn were isolated as reported previously [24] and homogenized in buffer A (50 mM Tris-HCl, pH 7.4, 320 mM sucrose, 1 mM EDTA, 1 mM dithiothreitol, 1 mM PMSF) plus protease inhibitor cocktail by 30 strokes with a glass Pyrex microhomogenizer. The homogenate was centrifuged at 1,000×g for 10 min, and the resulting supernatant was centrifuged at 10,000×g for 20 min to obtain the mitochondria-containing pellet and the supernatant. Mitochondria were washed three times with buffer B (10 mM Tris-HCl pH 7.4, 250 mM sucrose, 1 mM EGTA) by centrifugation for 10 min at 10,000×g. The supernatant was centrifuged at 100,000×g for 1 h to generate the cytosolic fraction.

Flow cytometric analysis

PBMCs were washed twice and stained at the cell surface with anti-CD3-PC7, anti-CD4-ECD and anti-CD8-v450. Cells were washed and then fixed with 4 % formaldehyde for 10 min on ice and then were stained intracellularly with anti-TNF-α-PE and anti-IFNγ-APC in 0.5 % saponin at room temperature. Intracellular cytokine production by T lymphocytes was analyzed by flow cytometry (FACSCyan ADP, Beckman Coulter) by gating at both CD4 and CD8 subsets and reporting the percentage of the T cell populations positive for the indicated cytokines. Cells were analyzed using the Flowjo software (TreeStar, Ashland, OR, USA).

Co-immunoprecipitation

PBMCs were lysed in lysis buffer [10 mM Tris-HCl, pH 7.4, 0.1 mM EDTA, 0,1 mM EGTA, 0.1 % SDS, 50 mM NaCl, 5 mM MgCl₂, 1 mM dithiothreitol, 1 mM Na₃VO₄, 5 mM NaF, 1 % protease inhibitor cocktails (Sigma) and 20 mM Na-molybdate]. The lysate was centrifuged for 5 min at 4 °C (800×g), the supernatant was recovered and centrifuged for 30 min at 4 °C (10,000×g), and the protein content was quantified by Bradford’s colorimetric assay (Bio-Rad).

The protein A-agarose (Roche) was washed three times in lysis buffer and incubated for 45 min at 4 °C with 300 μl of cell lysate (2 mg protein) and 300 μl of 10 mM Tris-HCl, pH 7.4, 5 mM EDTA, 0.1 % SDS, 50 mM NaCl, 10 % glycerol, 1 % protease inhibitor cocktails (Sigma) and 20 mM Na-molybdate. After centrifugation for 3 min at 4 °C (10,000×g), 300 μl of a buffer containing 6 μg of anti-GRα was added to the supernatant and incubated overnight at 4 °C. Beads were finally washed three times with lysis buffer, and co-immunoprecipitated proteins were released by boiling in SB (2 % SDS, 20 % glycerol, 100  mM Tris–HCl, pH 6.8, 100  mM dithiothreitol).

Human phospho-kinase array kit

PBMCs were challenged with PMA/iono and/or single- or double-treated with MPSS or JWH-015 and subsequently screened for 46 different kinases by means of a phosphor-kinase array kit (R&D Systems, Oxford, UK), according to manufacturer's instructions. Briefly, capture and control antibodies were spotted in duplicate on nitrocellulose membranes, and cellular extracts were diluted and incubated overnight with the human phospho-kinase array. The array was washed to remove unbound proteins followed by incubation with a cocktail of biotinylated detection antibodies. Streptavidin-HRP and chemioluminescent detection reagents were applied, and a signal was produced at each capture spot corresponding to the amount of each phosphorylated protein bound.

Statistical analysis

Statistical analysis was performed through GraphPad Prism 5.0 (GraphPAD Software for Science, San Diego, CA). Parametric statistical analysis (mean ± SD) was performed using standard methods. Significant differences were calculated using one- or two-way ANOVA, followed by Bonferroni post hoc multiple comparisons. Differences were considered significant at p < 0.05.

Results

Effect of MPSS and JWH-015, alone or in combination, on neuroprotection after focal brain damage

Previous studies from our group clearly demonstrated that HCb induces degeneration in precerebellar nuclei, namely inferior olive (IO) and Pn, contralateral to the cerebellar lesion, that can be reduced by selective stimulation of either GRα or CB₂ receptors [11, 23]. These data suggested a possible synergic neuroprotective effect of a combined GRα and CB₂ stimulation. To test this hypothesis, we analyzed neuronal degeneration, cyt-c release, functional recovery and glial reaction in Pn after HCb upon different treatments (Table 1). In keeping with previous results, a single treatment with MPSS or JWH-015 protected Pn neurons from HCb-induced degeneration by increasing neuronal survival (Fig. 1a), and cyt-c release was reduced in neurons when lesioned animals were treated with either MPSS or JWH-015 (Fig. 1b–c). In addition, functional recovery, measured as NSS, was significantly diminished in MPSS and JWH-015-treated groups at 7 days after lesion, with JWH-015 being more effective in inducing a functional recovery already after 72 h following HCb lesion (Fig. 1d). Concerning HCb-induced glial response (Fig. 2a), we observed a significant decrease in the number of both GFAP astrocytes (Fig. 2b) and OX-42 microglial cells (Fig. 2c) in both the HCb-JWH-015 and HCb-MPSS groups. Then we combined MPSS and JWH-015 in order to ascertain a possible synergic effect in neuroprotection. To this aim, HCb groups were treated singularly or concomitantly with MPSS and/or JWH-015 (Table 1), and neuronal and glial responses were assessed.

Fig. 1.

Combined stimulation of GRα and CB₂ impairs neuron survival and functional recovery after focal brain injury. a Histograms of stereological neuronal count in pontine nuclei of control animals (saline) or HCb animals treated with saline or with MPSS, JWH-015, or a combination of MPSS and JWH-015. One-way ANOVA (F = 239.40; p < 0.0001) followed by Bonferroni multiple comparison test was performed. Data are reported as mean ± SD (n = 10 animals per group). *p < 0.05 versus HCb-saline; ***p < 0.001 versus HCb-saline; ^### p < 0.001 versus HCb-MSS and HCb-JWH-015. b Histograms of densitometric values of cyt-c release, expressed as mean fluorescence of individual cells, normalized to total cellular surface (F/A; n = 250 cells/group) in pontine nuclei of control animals (saline) or HCb animals treated with saline, MPSS, JWH-015 or a combination of MPSS and JWH-015. One-way ANOVA (F = 62.57; p < 0.0001) followed by Bonferroni multiple comparison test was performed. Data are reported as mean ± SD (n = 10 animals per group). ***p < 0.001 versus HCb-saline; ^### p < 0.001 versus HCb-MSS and HCb-JWH-015. c Representative immunoblotting and densitometry of the cytosolic levels of cyt-c in pontine nuclei of control animals (saline) or HCb animals treated with saline, MPSS, JWH-015 or a combination of MPSS and JWH-015, normalized to the loading control (Actin). One-way ANOVA (F = 94.66; p < 0.0001) followed by Bonferroni multiple comparison test was performed. Data are reported as mean ± SD (n = 5 animals per group). *p < 0.05 versus HCb-saline; ^### p < 0.001 versus HCb-MSS and HCb-JWH-015. d Time course of neurological recovery (NSS) in HCb-saline and HCb animals treated with saline, MPSS, JWH-015 or a combination of MPSS and JWH-015. Two-way ANOVA (time × treatment) followed by Bonferroni multiple comparison tests showed an overall significant effect for time (F = 222.2; p < 0.0001) and for treatment (F = 58.51; p < 0.0001), and also interaction (time × treatment) was significant (F = 16.77; p < 0.0001). Data are expressed as mean ± SD (n = 10 per group). *p < 0.05 versus HCb-saline at 7 day; ***p < 0.001 versus HCb-saline at 72 h, 5 and 7 day; ^# p < 0.05 versus HCb-MPSS; ^§§ p < 0.001 versus HCb-JWH-015 at 72 h; ^§§§ p < 0.001 versus HCb-JWH-015 at 5 and 7 day

Fig. 2.

Combined pharmacological activation of GRα and CB₂ impairs astrocytic and microglial responses. a Double-labeled and merged confocal images of Glial fibrillary acidic protein (GFAP; green) and OX-42 (red) plus DAPI-counterstaining (gray) in pontine nuclei of control animals (saline), HCb animals treated with saline or with MPSS, JWH-015 or a combination of MPSS and JWH-015, scale bar = 20 μm. Histograms of the number of GFAP (b) and OX-42-positive cells (c) of control animals (saline), HCb animals treated with saline or with MPSS, JWH-015 or a combination of MPSS and JWH-015. One-way ANOVA (GFAP: F = 34.19, p < 0.0001; OX-42: F = 48.44, p < 0.0001) followed by Bonferroni multiple comparison test was performed. Data are reported as mean ± SD (n = 10 animals per group). **p < 0.01 versus HCb-saline; ***p < 0.001 versus HCb-saline; ^## p < 0.01 versus HCb-MPSS; ^§§§ p < 0.001 versus HCb-JWH-015; ^### p < 0.001 versus HCb-MPSS and HCb-JWH-015

Unexpectedly, simultaneous activation of both GRα and CB₂ by MPSS and JWH-015 failed to provide any neuroprotection after HCb (Fig. 1). Instead, the double treatment (HCb-MPSS + JWH-015) increased neuronal cell death and cyt-c release with respect to each single treatment of HCb groups (Fig. 1a–c). The lack of neuroprotection in double-treated animals was paralleled by the lack of any inhibitory effect on glial activation, inasmuch as both astrocytes and microglial cells remained highly activated in the HCb-MPSS + JWH-015 group (Fig. 2b–c). Furthermore, HCb double-treated groups showed a recovery curve almost identical to that of the HCb-saline group, thus significantly counteracting the positive effects observed in each single-treated group (Fig. 1d).

No significant differences were observed in all the above mentioned parameters among unlesioned animals treated with saline (CTRL-saline), with MPSS (CTRL-MPSS), with JWH-015 (CTRL-JWH-015) or double treated (CTRL-MPSS + JWH-015) (Supplementary Table 1).

Overall, neuroprotection was entirely lost when MPSS and JWH-015 were administered together after HCb, suggesting a functional interaction of a negative nature between GRα and CB₂ receptors.

Effect of MPSS and JWH-015, alone or in combination, on cytokine release from human T cells

To provide further evidence of the interaction between GRα and CB₂, we extended the investigation to an ex vivo human model of inflammation, i.e. activated human peripheral lymphocytes (Table 2). When challenged with 1 μg/ml phorbol-12-myristate-13-acetate (PMA) and 1 μg/ml Iono for 5 h, CD4+ and CD8+ T cells produce high levels of inflammatory cytokines [5], such as TNF-α and IFN-γ (Fig. 3a). As shown in Fig. 3b, 56 and 32 % of CD4+ T cells were positive for TNF-α and IFN-γ, respectively, whereas 44 and 38 % of CD8+ T cells were respectively positive for these cytokines. Of note, administration of MPSS and JWH-015, alone or in combination, in cells not activated by PMA/Iono, did not result in any cytokine production by both lymphocytes populations (Supplementary Fig. 1). We found that MPSS and JWH-015, when administered alone, induced a ~2-fold TNF-α suppression and a ~3-fold IFN-γ inhibition in both CD4+ and CD8+ activated subpopulations. However, treatment with both receptor agonists (PMA/Iono-MPSS + JWH-015) significantly dampened the antiinflammatory activity of GRα and CB₂ receptors alone. In fact, the combination of MPSS and JWH-015 showed a level of TNF-α and IFN-γ production comparable to that of PMA/Iono-stimulated T cells, corroborating the hypothesis of a functional interaction between eCBs and GCs receptors. These data were also confirmed by using another potent and selective CB₂ agonist, GP 1a. In fact, single treatment with GP 1a displayed a strong and significant inhibition of both TNF-α and IFN-γ production, which was comparable to that exerted by JWH-015. More importantly, combination of both MPSS and GP 1a had the same effects of the MPSS and JWH-015 one. Combination of MPSS and GP 1a treatments significantly blunted the immunosuppressive effect of single stimulation of GRα or CB₂, restoring the production of both cytokines (Supplementary Fig. 2).

Fig. 3.

Combined pharmacological activation of GRα and CB₂ impairs cytokine production from peripheral lymphocytes. a Flow cytometric progressive gating and representative dot plots showing TNF-α and IFN-γ production from unstimulated (CTRL) or PMA/Iono-activated CD4+ and CD8+ T lymphocytes treated or not with MPSS, JWH-015 or a combination of them, or GRα antagonist RU486 or with CB₂ antagonist SR144528. b Histograms of TNF-α and IFN-γ production from CD4+ (left) and CD8+ T lymphocytes (right), according to the different experimental conditions. One-way ANOVA (F = 10.06; p < 0.0001) followed by Bonferroni multiple comparison test was performed. Data are reported as mean ± SD (n = 6). **p < 0.01 versus PMA/Iono; ^## p < 0.01 versus PMA/Iono-MPSS and PMA/Iono-JWH-015; ^§§ p < 0.01 versus PMA/Iono-MPSS + JWH-015

In order to gain more insights into the contribution of each receptor to such an interaction, activated CD4+ and CD8+ T cells were pre-treated with the selective GRα antagonist RU-486 or the selective CB₂ antagonist SR144528 before treating them with a combination of MPSS and JWH-015. Interestingly, in double-treated CD4+ and CD8+ T cells pre-treatment with RU-486 did not affect TNF-α and IFN-γ production, whereas pre-treatment with SR144528 significantly reduced the levels of both cytokines. Taken together, these data point to a predominant role of CB₂ in controlling the cross-talk between GRα and CB₂.

Mechanistic insights into the GRα/CB₂ interaction

Several clinically oriented investigations report a correlation between increased levels of GRβ, the dominant negative inhibitor of GRα, with the development of tissue-specific insensitivity to GCs in various disorders, most of them associated with dysregulation of immune function [20, 25]. Thus, we investigated the effects of single or double treatment with JWH-015 or MPSS on the expression of GRα and GRβ in activated human PBMCs (Fig. 4). As could be expected, GRα levels were reduced, although not significantly, following MPSS or JWH-015 treatment, indicating an activation of this protein, which is in keeping with an antiinflammatory effect. Instead, MPSS + JWH-015 treatment did not affect GRα expression levels in activated cells (Fig. 4a). Interestingly, neither MPSS nor JWH-015 alone induced significant changes of GRβ protein levels in activated PBMCs compared to control cells (Fig. 4b); however, GRβ expression was significantly increased when activated PBMCs were treated with a combination of the two drugs, suggesting that GRβ mediates the GRα/CB₂ interaction.

Fig. 4.

Combined stimulation of GRα and CB₂ affects the expression of GRα and GRβ in peripheral lymphocytes. a Representative immunoblotting and densitometry of GRα expression normalized to the loading control (Actin) in PMA/Iono-activated peripheral blood lymphocytes untreated (lane 1) or treated with MPSS (lane 2), JWH-015 (lane 3) or a combination of them (lane 4). One-way ANOVA (F = 19.10; p < 0.01) followed by Bonferroni multiple comparison test was performed. Data are reported as mean ± SD (n = 4). *p < 0.05 versus PMA/Iono-MPSS and PMA/Iono-JWH-015. b Representative immunoblotting and densitometry of GRβ expression normalized to the loading control (Actin) in PMA/Iono-activated peripheral blood lymphocytes untreated (lane 1) or treated with MPSS (lane 2), JWH-015 (lane 3) or a combination of them (lane 4). One-way ANOVA (F = 39.25; p < 0.0001) followed by Bonferroni multiple comparison test was performed. Data are reported as mean ± SD (n = 4). *p < 0.05 versus PMA/Iono-MPSS and PMA/Iono-JWH-015

GRα, besides being regulated by GRβ, is also controlled by hetero-oligomeric complexes containing heat-shock protein 90 (hsp90) and heat-shock protein 70 (hsp70), which prevent its translocation to the cell nucleus and its antiinflammatory action in the absence of GCs [14, 15]. Thus, we investigated the formation of hsp90/hsp70/GRα complexes upon single or double activation of GRα and CB₂ by immunoprecipitation (Fig. 5). Activated PBMCs treated with MPSS or JWH-015 alone displayed low expression of hsp90 (Fig. 5a). Conversely, a ~3-fold increase in hsp90 expression was observed following double treatment with NPSS and JWH-015, and was completely abolished by SR144528 (Fig. 5a), but not by RU-486 (data not shown). Furthermore, while activated PBMCs treated with MPSS showed low expression of hsp70, JWH-015-treated cells showed a significant ~2-fold increase in its expression (Fig. 5b). Such an increase in hsp70 was also observed in double-treated cells and was completely abolished by SR144528 (Fig. 5b). These data suggest a key role for CB₂ in the GC/eCB interactions. To provide further mechanistic insights, we performed a screening of protein kinases that might be engaged by GRα and CB₂ using a phospho-kinase array kit. Interestingly, out of the 44 different phospho-kinases analyzed, only p-JNK and p-STAT6 showed significant differences in double-treated PBMCs compared to single-treated cells (Fig. 6). The proinflammatory p-JNK showed a ~1.5-fold increase when activated PBMCs were treated with MPSS and JWH-015 together, whereas the antiinflammatory Th2-driving p-STAT6 was significantly reduced (approximately by half) by the combination of the two drugs.

Fig. 5.

Combined stimulation of GRα and CB₂ enhances hsp70 and hsp90 expression in peripheral lymphocytes. a Representative immunoblotting and densitometry of GRα-precipitated hsp90 expression in PMA/Iono-activated peripheral blood lymphocytes treated with MPSS (lane 1), JWH-015 (lane 2), a combination of them (lane 3) or CB₂ antagonist SR144528 (lane 4). One-way ANOVA (F = 33.82; p < 0.01) followed by Bonferroni multiple comparison test was performed. Data are reported as mean ± SD (n = 4). **p < 0.01 versus PMA/Iono-MPSS; ^# p < 0.05 versus PMA/Iono-MPSS + JWH-015. b Representative immunoblotting and densitometry of GRα-precipitated hsp70 expression in PMA/Iono-activated peripheral blood lymphocytes treated with MPSS (lane 1), JWH-015 (lane 2), a combination of them (lane 3) or CB₂ antagonist SR144528 (lane 4). One-way ANOVA (F = 19.75; p < 0.01) followed by Bonferroni multiple comparison test was performed. Data are reported as mean ± SD (n = 4 per group). *p < 0.05 versus PMA/Iono-MPSS and PMA/Iono-JWH-015; ^# p < 0.05 versus PMA/Iono-MPSS + JWH-015

Fig. 6.

Combined stimulation of GRα and CB₂ modulates JNK and STAT6 expression in peripheral lymphocytes. a Human phospho-kinase array kit and densitometry graph of p-JNK in PMA/Iono-activated peripheral blood lymphocytes treated or not with MPSS, JWH-015 or a combination of them. One-way ANOVA (F = 30.70; p < 0.01) followed by Bonferroni multiple comparison test was performed. Data are reported as mean ± SD (n = 4 per group). ^# p < 0.05 versus PMA/Iono-MPSS and PMA/Iono-JWH-015. b Human phospho-kinase array kit and densitometry graph of p-STAT6 in PMA/Iono-activated peripheral blood lymphocytes treated or not with MPSS, JWH-015 or a combination of them. One-way ANOVA (F = 34.58; p < 0.01) followed by Bonferroni multiple comparison test was performed. Data are reported as mean ± SD (n = 4 per group). ^## p < 0.01 versus PMA/Iono-MPSS and PMA/Iono-JWH-015

Discussion

This study demonstrates for the first time the existence of a negative interaction between CB₂ and GRα on neuroprotective and antiinflammatory effects, both in rats and humans. Furthermore, our study indicates that such an interaction is mediated by CB₂ and might engage JNK and STAT6 pathways. Different neurodegenerative diseases, including multiple sclerosis, stroke, traumatic brain and spinal cord injury, as well as Alzheimer’s disease, have been recently considered suitable targets for both eCB- and glucortidoid-based drugs [26–33]. In this context, the present findings of that a negative interaction between eCBs and GCs receptors seem particularly relevant. HCb-based remote cell death is a well-established model to address neurodegeneration [34] as it has been associated with necrosis and apoptosis of neurons [35], glial activation and inflammation [23], as well as mitochondrial damage and autophagy [24, 36, 37]. In this model, neuroprotective effects of both GRα and CB₂ have been recently reported in terms of cell survival, glial activation and functional recovery [23, 34, 38], as well as oxidative and nitrative stress [24, 36]. Against this background, we wondered whether GRα and CB₂ ligands and their subsequent signaling could interact with each other. Double treatment resulted in striking differences compared to each treatment alone. In particular cell survival, which was highly enhanced after single treatment with either CB₂ or GRα agonists, was almost abolished after double treatment. Analysis of mitochondrial damage and glial activation yielded similar results, and also functional recovery scores corroborated the morphological and biochemical data. Therefore, our findings, though unexpected, clearly indicate a negative functional interaction exists between CB₂ and GRα in the central nervous system (CNS).

It is now well recognized that most neuroinflammatory diseases of the CNS are characterized by massive infiltrations of discrete populations of peripheral blood immune cells that are recruited within the CNS at the level of the blood brain barrier [39]. This recruitment is often driven by the endogenous immune cells of the CNS, i.e., microglia and astrocytes, which, following activation, locally release chemokines and cytokines that diffuse into the bloodstream, thereby attracting leukocytes at sites of inflammation [40, 41]. Among the recruited leukocytes, CD4 and CD8 T lymphocytes are the main players. Thus, our data on human activated T cells, by extending the evidence of a negative interaction between CB₂ and GRα receptors to humans, widen the clinical implications of this cross-talk. Indeed, cytokine release from these T cell subsets is highly indicative of neurodegenerative-associated inflammatory processes [42, 43]. In the present study, we mainly focused on TNF-α and IFN-γ production by both CD4 and CD8 T lymphocytes since they represent the main cytokines that characterize the Th-1 proinflammatory response and have been reported to be inhibited by either CB₂ or GRα activation [5, 44–46]. In keeping with our observations on the animal model, the concomitant activation of CB₂ and GRα prevented their antiinflammatory effects by blocking the production of both TNF-α and IFN-γ. Furthermore, the isolation of peripheral blood lymphocytes allowed us to unravel the role of the two receptors in mediating such a negative interaction. By using selective antagonists, we found that the negative interaction between CB₂ and GRα was erased only by SR144528, suggesting that the receptor cross-talk is driven by CB₂. The selectivity of drug manipulation of receptor is often a matter of debate. Indeed, JWH-015 could also activate CB₁, but only when used at high nanomolar concentrations (>432 nM) [47], concentrations quite far from those used here. The high affinities toward CB₂ and the very low affinities toward CB₁ of the selective agonists used (JWH-015 and GP 1A) clearly rule out any doubts on the specificity of CB₂ in sustaining the observed effects.

GC resistance or insensitivity has been recently reported to be the main cause of the lack of response to the manifold therapeutic benefits of GR agonists in various disorders of proliferation and chronic inflammatory diseases [20, 25, 48–50]. This can occur through multiple mechanisms, including differences in cellular localization, transcriptional activity and sensitivity of GR isoforms, as well as GR-independent mechanisms such as differences in those chaperones and co-chaperones that could alter GR heterocomplex signaling [51, 52]. Hence, our immunoblotting and immunoprecipitation data of the two main GR subunits (α and β) seem of major relevance, because they demonstrate a significant increase in GRβ after concomitant CB₂ and GRα activation. Recent evidence suggests that the ratio of GRα:GRβ expression is indeed critical to the GCs responsiveness of various cells, where higher ratios correlate with GC sensitivity, while lower ratios correlate with GC resistance [53]. In particular, the increased expression of GRβ in the development of GC-resistant forms of immune-related diseases is increasingly well documented [54–57]. In this context, our findings not only contribute to validate the theory of GRβ as a dominant negative inhibitor of GRα-induced transcriptional activity, but also offer a putative mechanism for the observed negative interaction between CB₂ and GRα.

Furthermore, we also analyzed other mechanisms underlying GC resistance through which CB₂ signaling might affect GRα-induced beneficial effects. Since the integrity of the mature GR heterocomplex is required for optimal binding of GRα agonists and subsequent activation of the transcriptional response, abnormalities in the chaperones and co-chaperones that make up the heterocomplex may contribute to such GC resistance. Indeed, several pieces of evidence showed that alterations in hsp90 and hsp70 were associated with decreased cellular sensitivity to GCs in several disorders, including malignancies, asthma, and multiple sclerosis [48, 58, 59]. In this scenario, our observed increase in both hsp90 and hsp70 following concomitant activation of CB₂ and GRα is of particular relevance in elucidating the molecular mechanism underlying their negative interaction. However, the increase in hsp70 was also observed following single stimulation of CB₂. This finding is in keeping with the previous observation that CB₂ agonism significantly increased the expression of hsp70 in neurons of the same rat model of remote cell death [24]. Therefore, it can be suggested that CB₂ activation per se induces an early increase in hsp70 expression, followed by a marked increase in hsp90 that was indeed observed when GRα was concomitantly activated. The role of CB₂ in mediating such hsp70/90-dependent CB₂/GRα interaction was further confirmed by the complete reversion of receptor cross-talk upon SR144528 pre-treatment.

Taken together, our findings suggest that the lack of neuroprotection and antiinflammation upon simultaneous activation of CB₂ and GRα can occur through at least two mechanisms: one mediated by GRβ and the subsequent alteration of the GRα:GRβ ratio, and the other one involving hsp90- and hsp70-mediated alteration of the mature GR heterocomplex.

In order to further investigate the molecular basis of the CB₂/GRα interaction, we performed a preliminary proteomic multiarray screening of the various phospho-proteins known to be downstream of CB₂ and/or GRα. Interestingly, only JNK and STAT6 were significantly modulated upon concomitant activation of the two receptors compared to their individual stimulation. However, this preliminary screening is far from comprehensive, and further studies are required to clarify the real mechanism at the basis of the observed CB₂/GRα negative cross-talk. The unprecedented evidence of a negative interaction between these two receptors also extends the concept of eCBs as part of a wider regulatory system in which different signaling pathways interact to control manifold biological functions of the body, from mood control to inflammation and neurodegeneration [60–64]. On a final note, it should be mentioned that nowadays there is quite a demand to extend the clinical use of (endo) cannabinoid-oriented drugs to treat neurological diseases where steroid treatment is already a routine. In this light, our present findings call for caution concerning eCB/GC combinations as novel therapeutic approaches; however, they point to a possible exploitation of CB₂-targeted therapeutics to cope with the ever-increasing problem of GC resistance.

Electronic supplementary material

Below is the link to the electronic supplementary material.