Pharmacological blockade of cannabinoid receptor 2 signaling does not affect LPS/IFN-γ-induced microglial activation

Cannabinoid receptor 2 (CB2) signaling modulates microglial responses to inflammatory stimuli. Our previous studies demonstrated that genetic deletion of CB2 inhibits microglial activation during inflammatory stimulation of toll-like receptors (TLRs) or in neurodegenerative conditions. However, we cannot exclude developmental effects of the constitutive CB2 knockout (CB2−/−), which could mediate compensatory outcomes in CB2−/− mice. In the present study, we therefore tested whether acute pharmacological inhibition of CB2 receptor has a similar effect on microglial activation as in CB2−/− in response to inflammatory stimulation. Our findings suggest that the CB2-specific antagonist SR144528 has little or no effect on LPS/IFN-γ-induced activation in primary microglia or organotypic hippocampal slice cultures at nanomolar concentrations. We show that SR144528 did not alter LPS/IFN-γ-mediated microglial cytokine secretion, Iba1 and CD68 staining intensity or morphology at 1 and 10 nM. Although SR144528 suppressed LPS/IFN-γ-induced microglial activation at 1 µM, this anti-inflammatory effect was not dependent on CB2 receptors and exceeded the Ki on CB2 receptors by more than a thousand-fold. Thus, SR144528 does not mimic the anti-inflammatory effects observed in the CB2−/− microglia after LPS/IFN-γ stimulation. Therefore, we propose that the deletion of CB2 probably triggered an adaptive mechanism, making microglia less responsive to inflammatory stimulation.

These results were somewhat surprising because CB2 agonists also produce anti-inflammatory effects on microglia and macrophages in response to inflammatory stimuli.For example, JWH-015 and AM1241 suppressed inflammatory responses to IFN-γ or LPS/IFN-γ stimulation 7,15 .Moreover, JWH-133 increased the production of the anti-inflammatory cytokine IL-10 in TLR4-stimulated macrophages, and co-administration of the CB2 antagonist SR144528 blocked this effect 9 .In vivo, CB2 activation with JWH-133 similarly improved cognition in an AβPP/PS1 mouse model and lowered inflammatory cytokine expression and microglia reactivity 8 .Additionally, the CB2 agonist 0-1966 reduced the number of activated microglia and enhanced motor function in another model of traumatic brain injury 16 .
Considering the above-mentioned limitations, it is not surprising that genetic deletion and acute pharmacological inhibition can have different outcomes 17 .Our findings of reduced microglial activation in CB2 −/− mice raised the question of whether this effect was caused by the acute lack of CB2 signaling or was the result of compensatory mechanisms.To address this, we now analyzed the effect of acute treatment of LPS/IFN-γ-stimulated microglia with the potent selective CB2 antagonist/inverse agonist SR144528 18,19 .SR144528 is among the most commonly used antagonists, due to its high affinity and selectivity for CB2 receptors at subnanomolar concentrations 18 , with an inhibitory constant (Ki) of 0.6 nM 18 .

Results
Pharmacological blockade of CB2 signaling has no effect on LPS/IFN-γ-induced microglial cytokine secretion.To assess whether acute pharmacological inhibition of CB2 signaling has similar effects on LPS/IFN-γ-induced microglial activation as the constitutive deletion of CB2, we tested the effects of increasing concentrations of SR144528 18 (1 nM-1 µM) on LPS/IFN-γ-induced cytokine secretion in wildtype (WT) primary microglia (Fig. 1).Stimulation with LPS/IFN-γ significantly increased the secretion of all cytokines tested (p < 0.0001).This response was not affected by pre-treatment with SR144528 at nanomolar concentrations, where the compound is CB2 selective.Only a concentration of 1 µM significantly reduced LPS/IFN-γ-induced secretion of TNF-α (p = 0.0008), IL-6 (p = 0.0089), and CCL2 (p = 0.0140).This concentration exceeds the Ki of SR144528 on the CB2 receptors by more than three orders of magnitude, suggesting that the observed effect could be CB2-independent.

SR144528 (1 µM) does not affect cell viability in LPS/IFN-γ-stimulated microglia.
To test whether the reduced levels of cytokines measured after treatment with 1 µM SR144528 was due to an effect on cell viability, we quantified cell death in primary microglia using flow cytometry (Fig. 3).We found a significant reduction in the percentage of live cells after LPS/IFN-γ stimulation (p = 0.0019), which was not affected by pre-treatment with 1 µM SR144528 (Fig. 3).This data indicates that the effect of 1 µM SR144528 was not due to cytotoxicity, thus supporting the finding of potential immune-modulatory off-target effects at micromolar concentrations.

SR144528 does not affect LPS/IFN-γ-induced microglial Iba1 and CD68 intensities in
OHSCs.We also investigated another endpoint of LPS/IFN-γ stimulation, namely the microglial marker Iba1 (ionized calcium-binding adapter molecule 1) and the lysosomal marker and indicator for phagocytosis CD68, in activated microglia.For this purpose, we used organotypic hippocampal slice cultures (OHSCs), a model that supports the physiological maintenance of all hippocampal cell types in culture.We pre-treated OHSCs with SR144528 (1, 10 nM and 1 µM), followed by LPS/IFN-γ stimulation for 16 hr (Fig. 4) and analyzed Iba1 and CD68 staining intensities within microglia in the CA1 pyramidal layer of the treated OHSCs (Fig. 4B-C).Our result showed that both Iba1 and CD68 intensities were significantly increased after LPS/IFN-γ stimulation.However, pre-treatment with SR144528 did not considerably change Iba1 or CD68 intensities at 1 nM and 10 nM (Fig. 4D-E).Here, 1 µM of SR144528 reduced Iba1 but not CD68 intensity in LPS/IFN-γ stimulated OHSCs (Fig. 4D-E).This data additionally supports that blockade of CB2 signaling by SR144528 at nanomolar concentrations does not influence microglial activity after LPS/IFN-γ stimulation.

SR144528 does not considerably alter LPS/IFN-γ-induced activated microglial morphology in
OHSCs.Finally, we determined microglial morphology using the MotiQ plugin 20 to characterize 3D morphometric parameters of microglia in LPS/IFN-γ-stimulated OHSCs pre-treated with SR144528 (Fig. 5B).We observed that LPS/IFN-γ-stimulated microglia exhibited ameboid-like features such as fewer branches (Fig. 5C), shorter tree lengths (Fig. 5D), and lower cell volume (Fig. 5F).These parameters were not altered by SR144528 (1 and 10 nM).In contrast, the ramification index-a measure of the overall complexity of cell shape was significantly increased with 10 nM of SR144528 pre-treatment (Fig. 5E), whereas all the parameters tested were significantly reduced with 1 µM of SR144528 pre-treatment (Fig. 5C-F).This finding further indicates that SR144528 at nanomolar concentrations does not generally influence the LPS/IFN-γ-induced ameboid-like microglial morphology, which is a common feature of activated microglia.

Discussion
In this study, we asked if acute pharmacological inhibition of the CB2 receptor mimics the anti-inflammatory effect observed in constitutive CB2 deletion (CB2 −/− ) following stimulation with LPS/IFN-γ.Our results show that the blockade of CB2 signaling with the CB2 antagonist, SR144528 did not recapitulate the result of CB2 −/− with respect to modulating LPS/IFN-γ-induced microglial activation.It is likely that the constitutive lack of CB2 receptors affects microglia development or functions, making them less responsive to inflammatory stimuli.In addition, our data emphasize the potential off-target effects of SR144528 at micromolar concentrations.
We analyzed several aspects of microglial activation in primary microglia and OHSCs pre-treated with SR144528 and stimulated with LPS/IFN-γ.SR144528 is the most widely-used CB2 antagonist for pharmacological  www.nature.com/scientificreports/studies due to its high potency and selectivity for both mouse and human CB2 receptors 18,21 .It has been employed in numerous studies, often to block the effects of CB2 agonists 9,11,22 .The SR144528 concentration required to produce its half maximum inhibition (Ki) 18 is 0.6 nM.This compound thus blocks CB2 receptors at low nanomolar concentrations.We showed that pre-treatment with SR144528 at 1 nM and 10 nM concentrations did not affect the secretion of inflammatory cytokines IL-6, TNF-α, and CCL2 in LPS/IFN-γ-stimulated microglia.Furthermore, SR144528 at 1 nM and 10 nM did not alter staining intensities of pro-inflammatory microglial activation markers, Iba1, which is often increased during microglia activation 14,23 and CD68, which is also enhanced during inflammation 10 .Additionally, we assessed the morphological changes of microglia resulting from LPS/IFN-γ stimulation in a 3D reconstruction analysis.It is of common notion that the shape of microglia gives helpful information about their activation state 14,[23][24][25] .After LPS/IFN-γ stimulation of OHSCs, we observed activated morphological changes characterized by fewer microglial branch numbers, shorter tree lengths, reduced ramification, and cell volume in the stimulated OHSCs.Pre-treatment with SR144528 at 1 and 10 nM failed to produce a difference in all but one of these parameters, supporting its inability to generally prevent LPS/IFN-γ-induced microglial activation.Only the ramification index was increased by 10 nM SR144528, indicating that either CB2 signaling has a distinct influence on this morphology parameter or reflects a particularly sensitive off-target effect.Together, these data strongly suggest that acute inhibition of CB2 signaling does not modulate microglial responses to LPS/IFN-γ, which is in contrast with our previous results 14 showing that the genetic deletion of CB2 suppressed LPS/IFN-γ-mediated microglial activation.
It is noteworthy that a high concentration of 1 µM SR144528 suppressed LPS/IFN-γ-induced microglial cytokine secretion, Iba1 intensity and morphological alterations, an effect that was not dependent on the CB2 receptor.This finding is relevant, because many studies in microglia or macrophages used SR144528 at 1 µM or higher concentrations to reverse the effect of CB2 agonists in inflammatory settings 9,11,22,26 .Even though SR14428 is an excellent CB2-specific antagonist, some off-target effects have been described at micromolar concentrations 21 .Considering that 1 µM is more than a thousand-fold greater than its Ki, it is likely that SR144528 had CB2-independent off-target immune-modulatory functions.Indeed, data from CB2 ligand profiling revealed that SR144528 has off-target activities on other proteins, namely adenosine A3 receptor (A 3 AR) and genotype/stimulation (representative data from two independent microglia preparations).Data are presented as mean ± SD.Two-way ANOVA followed by Tukey's multiple comparisons.#### p < 0.0001, ## p < 0.01, # p < 0.05 indicate significance to the untreated control group.Significant differences between LPS/IFN-γ vs. LPS/IFN-γ pre-treated with SR144528 are indicated with *** p < 0.001, ** p < 0.01, * p < 0.05.phosphodiesterase 5 (PDE5) 21 .Another study showed that SR144528 and AM-251 are inhibitors of the enzyme acyl CoA: cholesterol acyltransferase (ACAT) 27 .While it was beyond the scope of our present study to identify the CB2-independent mechanisms of 1 µM SR144528 in microglia, these off-targets might be involved.One study demonstrated a pro-inflammatory effect of the A 3 AR agonist, CI-IB-MECA, on LPS-induced TNF-α secretion in peritoneal macrophages 28 .Another study showed that inhibiting ACAT1 in macrophages overloaded with cholesterol produced a lower inflammatory profile 29 .
Where genetic knockout models are important tools for understanding specific roles of genes and their underlying mechanisms in neuropathological or inflammatory conditions [30][31][32][33] , the permanent genetic alteration can initiate adaptive developmental changes 34 .It can thus be difficult to distinguish between direct effects due to the targeted gene deletion and indirect effects, requiring other means of functional gene or protein invalidation, such as pharmacological agonists and antagonists 35 .Here, we found that acute SR144528 treatment did not affect LPS/IFN-γ-induced microglial activation, whereas constitutive deletion of CB2 reduced TLR-induced microglial activation in our previous study 14 , suggesting that the latter is not caused by an acute lack of CB2 signaling during stimulation.Instead, these findings indicate that CB2 −/− microglia are primed differently, making them somewhat resistant to LPS/IFN-γ stimulation.A few studies have reported discrepancies between genetic deletion and pharmacological inhibition of proteins 17,35 .For example, differential effects of soluble epoxide hydrolase inhibition and genetic deletion was shown on cardiac fibrosis 17 while another study demonstrated that the (A) Experimental setup.Primary microglia were pre-treated with 1 µM SR144528 for 15 min, followed by LPS/IFN-γ stimulation for 16 h.DRAQ7™ was used to stain dead cells (B) Gating strategy used for the stained microglia.(C) Quantification of the percentage of live microglia cells (DRAQ7 negative).N = 6-9 samples/ stimulation (from at least two independent microglia preparations).Data are presented as mean ± SD.One-way ANOVA followed by Tukey's multiple comparisons.## p < 0.01, # p < 0.01 indicate significance to the untreated control group.lysophosphatidate LPA 1 receptor antagonist Ki16425 did not totally mimic the effects of the constitutive lack of LPA1 receptors on depression-like behavioral tests in mice 35 .
In our previous studies, we reported that primary microglia from CB2 −/− mice showed decreased secretion of inflammatory cytokines and decreased surface expression of activation markers following LPS/IFN-γ stimulation 6 .Recently, we revealed that the lack of CB2 attenuated TLR-induced microglial activation 14 .These findings suggested that the CB2 −/− mice are less prone to inflammatory stimulation of microglia.Somewhat contradictory to these observations, CB2 activation with CB2 agonists also exhibited anti-inflammatory effects on microglia and macrophages in response to inflammatory stimuli.For example, the CB2 agonists JWH-015 and AM1241 were shown to dampen inflammatory responses to IFN-γ or LPS/IFN-γ stimulation in primary microglia and N9 microglia respectively 7,15 .Additionally, JWH-133 was reported to increase IL-10 secretion in LPS/IFN-γ-stimulated macrophages, which was reversed by SR144528 9 .Based on these findings, it is evident that pharmacological activation/inactivation of CB2 and genetic manipulations can have different outcomes.Our results are most likely due to compensatory mechanisms, as we have previously reported about 200 genes (including gene ontology terms for response to bacteria and viruses) that are differentially expressed between unstimulated WT and CB2 −/− microglia 14 .Some of these genetic changes probably make CB2 −/− microglia less sensitive to inflammatory stimulation.

Methods
Mice.C57BL/6J (WT) and B6.cg-Cnr2tm1Zim (CB2 −/− ) 30 mice were used in this study.The C57BL/6J mice were originally procured from a Charles River commercial breeder and bred in-house.The B6.cg-Cnr2tm1Zim mice were bred homozygously and backcrossed to C57BL/6J line every six generations to reduce the chance of genetic drift.All animals were housed in the animal facility unit of the University of Bonn under specific pathogen-free conditions.They were subjected to a 12-h dark/light cycle with ad libitum access to food and water.Animal care and organ retrievals were performed according to the guidelines of the European Communities Directive 86/609/E.E.C. and the German Animal Protection Law regulating animal research and were approved by the Landesamt für Natur-, Umwelt-, und Verbraucherschutz (LANUV NRW), Germany (01_Organentnahme).The study was conducted in accordance with ARRIVE guidelines.

Figure 3 .
Figure 3. 1 µM of SR144528 does not affect the percentage of live cells in LPS/IFN-γ stimulated microglia.(A)Experimental setup.Primary microglia were pre-treated with 1 µM SR144528 for 15 min, followed by LPS/IFN-γ stimulation for 16 h.DRAQ7™ was used to stain dead cells (B) Gating strategy used for the stained microglia.(C) Quantification of the percentage of live microglia cells (DRAQ7 negative).N = 6-9 samples/ stimulation (from at least two independent microglia preparations).Data are presented as mean ± SD.One-way ANOVA followed by Tukey's multiple comparisons.## p < 0.01, # p < 0.01 indicate significance to the untreated control group.

Figure 4 .Figure 5 .
Figure 4. SR144528 does not affect microglial CD68 and Iba1 intensities in LPS/IFN-γ-stimulated OHSCs.(A) Experimental setup.OHSCs were pre-treated with SR144528 at the indicated concentrations for 15 min, followed by LPS/IFN-γ stimulation for 16 h.After stimulation, OHSCs were stained for microglial markers and imaged using confocal microscopy.(B) Representative images of stimulated OHSCs showing DAPI (blue), Iba1 (cyan), and CD68 (magenta), scale bar = 100 µm, 10× magnification.The white dotted box indicates the region from which representative microglia shown in panel C were obtained.(C) Representative images from CA1 pyramidal microglia showing Iba1 and CD68 staining at 40× magnification, scale bar = 30 µm.Quantification of (D) Iba1 and (E) CD68 intensities.N ≥ 40 microglial cells/stimulation (representative data from two independent OHSCs preparations).Data are presented as mean ± SD.One-way ANOVA followed by Tukey's multiple comparisons was used for normally distributed data, while Kruskal-Wallis test followed by Dunn's multiple comparisons tests was used for data that were not normally distributed.#### p < 0.0001, ### p < 0.001 indicate significance to the untreated control group.Significant difference between LPS/IFN-γ vs. LPS/IFN-γ pre-treated with SR144528 is indicated with * p < 0.05.