Cannabinoid WIN55,212-2 reprograms monocytes and macrophages to inhibit LPS-induced inflammation

Introduction Chronic or uncontrolled activation of myeloid cells including monocytes, macrophages and dendritic cells (DCs) is a hallmark of immune-mediated inflammatory disorders. There is an urgent need for the development of novel drugs with the capacity to impair innate immune cell overactivation under inflammatory conditions. Compelling evidence pointed out cannabinoids as potential therapeutic tools with anti-inflammatory and immunomodulatory capacity. WIN55,212-2, a non-selective synthetic cannabinoid agonist, displays protective effects in several inflammatory conditions by mechanisms partially depending on the generation of tolerogenic DCs able to induce functional regulatory T cells (Tregs). However, its immunomodulatory capacity on other myeloid cells such as monocytes and macrophages remains incompletely understood. Methods Human monocyte-derived DCs (hmoDCs) were differentiated in the absence (conventional hmoDCs) or presence of WIN55,212-2 (WIN-hmoDCs). Cells were stimulated with LPS, cocultured with naive T lymphocytes and their cytokine production and ability to induce T cell responses were analysed by ELISA or flow cytometry. To evaluate the effect of WIN55,212-2 in macrophage polarization, human and murine macrophages were activated with LPS or LPS/IFNγ, in the presence or absence of the cannabinoid. Cytokine, costimulatory molecules and inflammasome markers were assayed. Metabolic and chromatin immunoprecipitation assays were also performed. Finally, the protective capacity of WIN55,212-2 was studied in vivo in BALB/c mice after intraperitoneal injection with LPS. Results We show for the first time that the differentiation of hmoDCs in the presence of WIN55,212-2 generates tolerogenic WIN-hmoDCs that are less responsive to LPS stimulation and able to prime Tregs. WIN55,212-2 also impairs the pro-inflammatory polarization of human macrophages by inhibiting cytokine production, inflammasome activation and rescuing macrophages from pyroptotic cell death. Mechanistically, WIN55,212-2 induced a metabolic and epigenetic shift in macrophages by decreasing LPS-induced mTORC1 signaling, commitment to glycolysis and active histone marks in pro-inflammatory cytokine promoters. We confirmed these data in ex vivo LPS-stimulated peritoneal macrophages (PMΦs), which were also supported by the in vivo anti-inflammatory capacity of WIN55,212-2 in a LPS-induced sepsis mouse model. Conclusion Overall, we shed light into the molecular mechanisms by which cannabinoids exert anti-inflammatory properties in myeloid cells, which might well contribute to the future rational design of novel therapeutic strategies for inflammatory disorders.


Introduction
Inflammation is a vital defense mechanism that can be initiated by the immune system after detecting invading pathogens and/or cellular damage. Inflammatory responses are orchestrated to eliminate the cause of injury and to initiate healing upon tissue damage. Alterations of these tightly regulated processes are associated with different immune-mediated diseases and inflammatory disorders such as autoimmune and allergic diseases, autoinflammatory syndromes or sepsis (1,2). Myeloid cells such as conventional dendritic cells (DCs), monocytes and macrophages are specialized in the initiation and amplification of inflammatory responses. These cells are equipped with a large battery of patternrecognition receptors (PRRs) to sense a plethora of harmful pathogens and damage-associated stimuli, which allows their activation, initiation of inflammatory responses and spreading of the alarm throughout the body (1,3,4). Compelling experimental evidence shows that inflammatory responses orchestrated by myeloid cells are closely linked to their metabolic profile (5,6). For example, the activation of mammalian target of rapamycin complex 1 (mTORC1) on myeloid cells promotes anabolic metabolism and shift from glycolysis towards lactic fermentation as a faster energetic source to induce and sustain the functional adaptations required under a pro-inflammatory scenario (6). Under certain pathological conditions, the chronic or uncontrolled activation of myeloid cells significantly contributes to perpetuate inflammation and tissue damage, which leads to the main clinical manifestations associated to such diseases (2,7). Although current immunosuppressive therapies represent the mainstay of treatment for many inflammatory conditions, their clinical efficacy is still limited for many patients and their overuse is associated with significant adverse effects (8). Therefore, the discovery of novel molecules able to target myeloid cells and dampen inflammation as well as the better understanding of their mode of action at the molecular level might well contribute to the development of alternative therapeutic interventions for many immune-mediated and inflammatory diseases.
The endocannabinoid system (ECS) is a complex signaling network encompassing the cannabinoid receptors (CBRs), the endogenous cannabinoids (anandamide, AEA) and 2-arachidonoylglycerol, 2-AG) and the enzymes involved in their synthesis and degradation (9, 10). In addition, phytocannabinoids from Cannabis sativa L., marijuana, (D 9tetrahydrocannabinol, THC) and synthetic cannabinoids (WIN55,212-2, HU210, or HU308) also activate CBRs regulating proliferation, differentiation, cell survival, metabolism or immunity (9, 10). The therapeutic exploitation of the ECS is an emerging field of research, and different pre-clinical and clinical studies point out cannabinoids as potential therapeutic tools for inflammatory diseases (9-12). At this regard, we have previously demonstrated that human DCs express functional CBRs and that the synthetic cannabinoid WIN55,212-2 induces anti-inflammatory immune responses by promoting metabolic reprogramming in fully differentiated DCs, which eventually led to the generation of functional regulatory T cells (Tregs) with suppressive capacity both in vitro and in vivo (13). We have also shown that WIN55,212-2 is able to restore rhinovirus-induced bronchial epithelial barrier disruption (14). However, the potential capacity of WIN55,212-2 to regulate the generation of human monocyte-derived DCs (hmoDCs) and its anti-inflammatory properties on other myeloid cells such as macrophages remains elusive. Herein, we show for the first time that the differentiation of hmoDCs in the presence of WIN55,212-2 generates tolerogenic DCs (WIN-hmoDCs) that are less responsive to lipopolysaccharide (LPS) stimulation and able to prime functional Tregs. We also uncover that WIN55,212-2 inhibits the pro-inflammatory M1 activation in THP-1 Macrophages (THP-1 MFs) and primary cultures of human monocyte-derived macrophages by mechanisms depending on metabolic and epigenetic rewiring. WIN55,212-2 protects from LPSinduced sepsis in vivo and limits LPS-induced metabolic reprogramming of ex vivo stimulated peritoneal macrophages (PMFs). Our data enhance the knowledge about the molecular mechanisms by which cannabinoids exert anti-inflammatory properties in myeloid cells, which might help to the development of novel therapeutic interventions for inflammatory diseases.
In vitro cell activation was performed with LPS from Escherichia coli (O127:B8, Sigma-Aldrich) and/or IFNg (PeproTech). In ex vivo and in vivo experiments, LPS from E. coli (O155:B5, Sigma-Aldrich) was employed. The cannabinoid agonist WIN55,212-2 (Sigma-Aldrich) was used. The specific doses and time points are specifically detailed for each corresponding experiment.

Generation of human monocyte-derived dendritic cells and WIN-hmoDCs
Peripheral blood mononuclear cells (PBMCs) were isolated from buffy coats of healthy anonymous donors (Transfusion Centre of Madrid, under the approved code "PODIS09") by Ficoll Paque density-gradient. Then, monocytes were purified by positive magnetic isolation using anti-human CD14 microbeads and AutoMACS technology (Miltenyi Biotec). To generate hmoDCs, monocytes were seeded at 1x10 6 cells/mL concentration and differentiated in the presence of 100 ng/mL of granulocytemacrophage colony-stimulating factor (GM-CSF) and 100 ng/mL IL-4 (Preprotech) for 6 days. To generate WIN-hmoDCs, WIN55,212-2 (Sigma-Aldrich) at a final concentration of 50 nM was added at days 0 and 4 of the hmoDCs differentiation.

Coculture experiments
Peripheral blood naïve CD4 + T cells were purified from PBMCs obtained from buffy coats of healthy donors using the "Naïve CD4 + T Cell Isolation Kit" (Miltenyi Biotec). LPS-stimulated hmoDCs and WIN-hmoDCs were cocultured with allogeneic naïve CD4 + T cells (hmoDC:T cell ratio of 1:5) during 5 days. Then, cell-free supernatants were collected for cytokine quantification and cells were stained for flow cytometry analysis.

NF-kB/AP-1 activation assay
The evaluation of NF-kB/AP-1 activation in human macrophages was performed using the THP-1 XBlue ™ monocytic cell line (InvivoGen), which is stably integrated with a NF-kB/AP-1 inducible secreted embryonic alkaline phosphatase (SEAP) reporter construct. Briefly, THP-1 XBlue monocytes were differentiated into human THP-1 MFs as described above. Then, after incubation with the indicated stimuli, the activation of NF-kB/ AP-1 was determined in cell-free supernatants using a SEAP colorimetric detection reagent called QUANTIBlue ™ .

Cytokine quantification
Concentrations of human TNFa, IL-1b, IL-6, IL-10, IL-5, IFNg and murine IL-6 in cell-free supernatants were quantified using sandwich ELISA cytokine kits from BD Biosciences. Murine TNFa and IL-1b levels were assessed by sandwich ELISA using specific kits from Invitrogen. In all cases, the manufacturer protocols were followed with minor modifications.

Flow cytometry
For flow cytometry determinations, monoclonal antibodies were purchased from BioLegend: human anti-FOXP3-Alexa Fluor 488 (A488), anti-CD127-phycoerythrin (PE), anti-CD25allophycocyanin (APC), anti-CD4-peridinin chlorophyll protein complex (PerCP), anti-HLA-DR-Fluorescein-5-isothiocyanate (FITC) anti-CD86-PE and anti-CD83-APC. Staining for CB1 was performed using the HU210-A488 fluorescent probe as previously described (15), A488 alone was use as control. For the detection of CB2 receptor, a first incubation with the primary CB2 polyclonal antibody (PA1744, Invitrogen) was followed by the incubation with an A488-conjugated goat anti-rabbit secondary antibody (Invitrogen). Cell viability was analyzed by propidium iodide (PI, Life technologies) or eFluor660 (Invitrogen) staining. For surface staining, cells were harvested, washed in 2 mM PBS/EDTA with 0.5% bovine serum albumin (BSA) and stained in the darkness for 15 min at room temperature with the corresponding fluorescencelabelled or isotype control antibodies. For the analysis of FOXP3 expression in human T cells cocultured with hmoDCs, a first surface staining with anti-human CD4-PerCP, CD127-PE, and CD25-APC antibodies was followed by subsequent fixation and permeabilization. Then, cells were stained with anti-human FOXP3-A488 according to the manufacturer's recommendations. The corresponding isotype controls were included in every staining step. Flow cytometry analysis was performed with a FACSCalibur cytometer (Becton Dickinson) in the cytometry and fluorescence microscopy unit at the Complutense University of Madrid.

Western Blot
After the corresponding treatments, cells were washed and lysed with RIPA buffer (Thermo Fisher Scientific) supplemented with Protease/Phosphatase Inhibitor Cocktail (Cell Signaling) and 1 mM PMSF (Sigma-Aldrich) for 30 min at 4°C. Protein was quantified with Micro BCA Kit (Pierce) following the manufacturer's protocols. 10 mg of total protein from cell lysates were separated in 10% SDS-polyacrylamide gel electrophoresis (SDS-PAGE) and transferred onto a nitrocellulose membrane (Bio-Rad Laboratories). The membrane was blocked with 5% BSA, 0.1% Tween-20 in Tris buffered saline for 1 h and incubated with the indicated primary antibodies: phospho-Akt (Ser473) (1:1000, Cell Signaling), phospho-p70 S6 Kinase (Thr389) (1:750, Cell Signaling), and bactin (1:15000, Sigma-Aldrich). Then, the membrane was washed and incubated with goat anti-rabbit (1:4000; Bio-Rad Laboratories) or goat anti-mouse (1:2500, Pierce) conjugated with horseradish peroxidase (HRP) as a secondary antibody. The signal was developed with Clarity Western ECL Substrate, detected in a Chemidoc XRS System and analyzed with Image Lab Software (all from Bio-Rad). Briefly, for the quantification of the images obtained, same-size regions of interest (ROIs) were displayed over the different bands and the intensity/mm 2 of each ROI was quantified. After removal of the background value, the quantification of each band was then relativized to the one of its corresponding loading control (b-actin).

Metabolic studies
The determination of glucose concentration in cell-free culture supernatants was assessed by using the Glucose (GO) Assay Kit (Sigma-Aldrich). The metabolic rate was derived mathematically as the percentage of glucose consumed relative to its concentration in medium without cells (2 mg/mL). Lactate concentration was measured in perchloric acid deproteinized culture supernatants by means of a two-step coupled lactate oxidase (Sigma-Aldrich) and HRP (Fisher Scientific) enzymatic reactions. Briefly, lactate was oxidized producing H 2 O 2 which was coupled to the conversion of Amplex Red reagent (Fisher Scientific) to fluorescent resorufin by HRP. The fluorescence intensity of resorufin was analyzed using FLUOstar OPTIMA fluorescence reader (BMG LabTech).
Real-time metabolic profiling of macrophages was performed using a Seahorse XF HS Mini Analyzer (Agilent) following conventional protocols for adherent cells. Briefly, human macrophages were harvested at day 5 of differentiation and seeded (50x10 3 cells/well) in XFp Cell Culture Miniplates with cRPMI supplemented with 100 ng/mL GM-CSF. Cells were incubated 24 h and then activated for 18 h with LPS or LPS+WIN55,212-2 (LPS +WIN), 10 ng/mL and 10 mM respectively. For peritoneal macrophage experiments, mice peritoneal cells were seeded (400x10 3 cells/well) in XFp Cell Culture Miniplates with cRPMI supplemented with 20 ng/mL M-CSF and incubated for 3h. Peritoneal macrophages were purified by adherence through three consecutive washing steps and stimulated for 18 h with LPS or LPS+WIN55,212-2 (LPS+WIN), 10 ng/mL and 10 mM respectively. After stimulation, cRPMI medium was removed, cells were washed and cultured with XF DMEM medium supplemented with 10 mM glucose, 1 mM pyruvate and 2 mM glutamine in a CO 2free incubator for at least 1 h at 37°C before the assay. To accurately monitor glycolysis, we analysed the glycolytic proton efflux rate (glycoPER) at three consecutive stages: basal conditions (no drugs administrated), inhibition of the electron transport chain (0,5 µM Rotenone, ROT) and (0,5 µM Antimycin A, AA) and inhibition of glycolysis (50 mM 2-deoxy-D-glucose, 2-DG) (Sigma-Aldrich).

Quantitative PCR
The RNA from collected samples was isolated with RNeasy Mini Kit (Qiagen) according to the manufacturer's protocol. Next, complementary DNA was generated with a PrimeScript RT Reagent Kit (Takara Bio). Finally, quantitative PCR (qPCR) was performed by using FastStart Universal SYBR Green Master (Roche). The sequences of the primer pairs used are listed in Supplementary Table 1.

Chromatin Immunoprecipitation
For chromatin immunoprecipitation (ChIP), cells were harvested and fixed after stimulation. Samples were lysed and sonicated using a Branson 1200 Ultrasonic Cleaner to obtain chromatin fragments of optimal size (300-800 bp). Sheared chromatin was incubated using 2 mL of anti-H3K27ac antibody (Abcam, ab4729) or 2 mL of anti-IgG antibody (Millipore, 12-370), and immunoprecipitated with Dynabeads A (Invitrogen, 10001D) and Dynabeads G (Invitrogen, 10003D) magnetic beads. ChIP samples were de-cross-linked with proteinase K (Thermo Fisher Scientific) at 65°C for 4 h. Then, DNA was purified with the DNA Clean Kit & Concentrator TM-5 (Zymo Research). Quantitative analysis of the purified ChIP and Input DNAs was performed by using FastStart Universal SYBR Green Master (Roche) by real time qPCR.

Ex vivo culture of PMFs
Female BALB/c mice were anesthetized using isoflurane and subjected to intraperitoneal lavage with 3 mL of 20 mM PBS/EDTA. The peritoneal cells were then washed, seeded in medium at 0,5x10 6 cells/mL and allowed to rest for 3h. Then, PMFs were isolated by adherence after three consecutive washing steps and treated with the indicated stimuli.

In vivo LPS-induced sepsis model
All mice procedures included in this study were reviewed and ethically approved by Universidad Complutense de Madrid (UCM) and Comunidad Autońoma de Madrid (CAM) within the context of project SAF-2017-84978-R and PID2020-114396RB-I00 (CAM:ref.10/ 250312.9/18 and CAM:ref.10/465020.9/21). 6-week-old female BALB/c mice (Charles River) were randomly assigned into 4 groups. Then, mice were intraperitoneally injected with vehicle (PBS+DMSO), WIN55,212-2 (5 mg/kg), LPS (20 mg/kg), or LPS plus WIN55,212-2, respectively. Blood samples were collected 12 h after LPS administration via retroorbital bleeding using lime grass Pasteur pipettes and microtainer tubes (Beckton Dickinson). To collect serum, samples were centrifuged at 10,000 rpm for 10 min at room temperature. Then, the presence of bilirubin in mice serum was determined by its absorbance at 450 nm, lactate dehydrogenase (LDH) activity was analyzed using CyQUANT LDH Cytotoxicity Assay Kit (Fisher) and the concentration of TNFa, IL-1b and IL-6 was measured by ELISA.

Monocyte-derived dendritic cells differentiated in the presence of WIN55,212-2 display anti-inflammatory and tolerogenic features
To determine the capacity of WIN55,212-2 to modulate the phenotype and function of human monocyte-derived DCs (hmoDCs), monocytes were differentiated into DCs under conventional protocols in the absence or presence of this synthetic cannabinoid ( Figure 1A). First, the expression of CBRs in monocytes was confirmed at both mRNA and protein level (Supplementary Figure 1). Then, we used 50 nM of WIN55,212-2 during the differentiation process to generate WIN-hmoDCs, as this concentration showed the most effective results without affecting cell viability during dose-response optimization experiments (Supplementary Figure 2). After 18 h of LPS stimulation, conventional hmoDCs produced higher amounts of the pro-inflammatory cytokines TNFa, IL-1b and IL-6 than WIN-hmoDCs ( Figure 1B). No significant differences were observed in the production of the antiinflammatory cytokine IL-10. The IL-10/TNFa, IL-10/IL-1b and IL-10/IL-6 cytokine ratios were significantly higher in LPS-stimulated WIN-hmoDCs than conventional hmoDCs ( Figure 1C suggesting that WIN-hmoDCs might well preserve their capacity to antigen presentation and T cell activation. To further assay the capacity of the generated WIN-hmoDCs to polarize T cell responses upon LPS stimulation, we performed allogeneic coculture experiments with naïve CD4 + T cells. LPS-stimulated WIN-hmoDCs generated T cells that produced significantly lower amounts of IFNg and IL-5 than LPS-activated hmoDCs, without significant differences observed in the production of IL-10 ( Figure 1D). The IL-10/IL-5 and IL-10/IFNg ratios were significantly higher in T cells generated by WIN-hmoDCs than conventional hmoDCs ( Figure 1E), suggesting the generation of Tregs by WIN-hmoDCs. Supporting these data, the frequency of CD4 + CD127 -CD25 high FOXP3 high Tregs generated by LPSactivated WIN-hmoDCs was significantly higher than by LPSstimulated conventional hmoDCs ( Figure 1F). Collectively, these data indicate that the presence of WIN55,212-2 during monocyte differentiation promotes the generation of DCs with anti-inflammatory properties upon LPS stimulation and with increased capacity to polarize Tregs.
To further verify these results, we generated primary human monocyte-derived macrophages as shown in Figure 2E Figure 2F). Likewise, LPS/ IFNg-activated GM-MFs treated with WIN55,212-2 reduced the production of the pro-inflammatory cytokines TNFa, IL-1b and IL-6 both at the mRNA ( Figure 2G) and protein level ( Figure 2H), thus confirming the potent anti-inflammatory profile. In all cases, LPS/ IFNg-induced IL-10 levels were also inhibited by WIN55,212-2 in a dose-dependent manner (Supplementary Figure 5), suggesting that IL-10-producing M2 phenotypes in human macrophages are not generated under the assayed conditions. Remarkably, we uncovered a significant inverse correlation between IL-1b production and cell viability in GM-MFs stimulated under the different assayed conditions ( Figure 2I), which prompted us to investigate whether WIN55,212-2 could interfere with inflammasome activation and pyroptotic-mediated cell death. For that, we quantified the mRNA expression levels of the inflammasome components NLRP3 and CASP1, and the pore-forming protein GSDMD after macrophage activation under different conditions. LPS/IFNg stimulation significantly enhanced the mRNA expression levels of all the assayed molecules, which were in all the cases significantly impaired by WIN55,212-2 ( Figure 2J), confirming its potential capacity to rescue human macrophages from this inflammatory cell death pathway.

WIN55,212-2 impairs the LPS-induced metabolic and epigenetic reprogramming in human macrophages
To shed light into the molecular mechanisms underlying the anti-inflammatory capacity of WIN55,212-2 in human macrophages, we sought to investigate the potential effects of this cannabinoid on the metabolic pathways linked to the acquisition of inflammatory phenotypes in innate immune cells. For this set of experiments, we chose a lower dose of LPS for macrophage stimulation to avoid bias due to overactivation-induced cell death. Stimulation of GM-MFs with LPS (10 ng/mL) mimics the previous LPS/IFNg activation by significantly inducing the production of pro-inflammatory cytokines without affecting cell viability (Supplementary Figure 6). We initially analysed the mTORC1 signalling pathway as the key regulator of anabolic metabolism linked to inflammatory features in macrophages. WIN55,212-2 inhibited the LPS-induced phosphorylation of Akt and p70S6K ( Figure 3A), indicating a reduced activation of the mTORC1 pathway and suggesting the potential regulation of LPS-induced metabolic reprogramming. The administration of WIN55,212-2 alone showed minimal changes in glucose uptake and lactate production (Supplementary Figure 7). Nonetheless, WIN55212,2 significantly reduced the high glucose consumption and the increased lactate production observed in LPS-stimulated GM-MFs ( Figure 3B), demonstrating the capacity of this cannabinoid to interfere with the glycolytic metabolism. Supporting these data, WIN55,212-2 significantly reduced the mRNA expression levels of the glycolysis-related genes HK2, PFKB3, LDHA1 and HIF1A in LPS-stimulated macrophages, which was accompanied by significantly lower mRNA levels of the pro-inflammatory cytokines TNF, IL1B, IL8 and IL6 ( Figure 3C). To further confirm these results, we performed functional metabolic experiments using a Seahorse bioanalyzer to monitor real-time glycolytic proton efflux rate (glycoPER) in LPS-or LPS + WIN55,212-2-stimulated GM-MFs. Our data showed that commitment to glycolysis was significantly higher in LPS-stimulated than LPS+WIN55,212-2stimulated GM-MFs both at basal conditions (basal glycolysis) and after the inhibition of the electron transport chain (compensatory glycolysis) ( Figure 3D). Remarkably, the production of TNFa, IL-1b and IL-6 by GM-MFs stimulated with LPS or LPS+WIN55,212-2, directly correlated with the amount of glucose consumed by those cells (Figure 3E), indicating that the inhibition of the glucose uptake by human macrophages might well be connected to WIN55,212-2 antiinflammatory effects. As recent evidence links LPS-induced glycolysis with epigenetic changes that govern pro-inflammatory cytokine production in macrophages (18,19), we studied chromatin status near the TNF, IL1B and IL6 gene promoters by quantifying the activation histone mark H3K27ac by ChIP analysis. LPS significantly enhanced the acetylation of H3K27 within the proinflammatory IL1B and IL6 genes, which was significantly reduced by WIN55,212-2 ( Figure 3F). A similar non-significant tendency was also observed within the TNF gene promoter. Collectively, these findings indicate that WIN55,212-2 inhibits the metabolic and epigenetic inflammatory reprogramming induced by LPS in human macrophages.

WIN55,212-2 protects from LPS-induced sepsis in vivo and limits LPS-induced metabolic reprogramming of ex vivo stimulated peritoneal macrophages
To assess the potential in vivo anti-inflammatory properties of WIN55,212-2, we employed a mice model of LPS-induced sepsis. Mice were intraperitoneally injected with PBS, WIN55,212-2 or a lethal dose of LPS alone or in the presence of WIN55,212-2 ( Figure 4A). After 12h, we collected mice serum and quantified the levels of IL-1b, lactate dehydrogenase (LDH) activity and bilirubin levels as hallmarks of the inflammatory response/ inflammasome activation, cell death or organ injury, respectively. Intraperitoneal administration of LPS significantly enhanced the levels of serum IL-1b, which was significantly impaired by WIN55,212-2 ( Figure 4B). Remarkably, not only inflammasomedependent IL-1b release but the levels of other inflammatory mediators such as TNFa and IL-6 were also reduced by the treatment with the cannabinoid (Supplementary Figure 8). Similarly, increased levels of serum LDH and bilirubin were found in LPS-treated mice while mice receiving LPS+WIN55,212-2 showed significantly lower levels of both parameters ( Figure 4C). These data confirmed the in vivo anti-inflammatory properties of WIN55,212-2 in the context of LPS-induce sepsis. As macrophages depict one of the main immune cell populations of the peritoneal cavity (20), we next sought to investigate their potential contribution into such protective anti-inflammatory effects. For that, we isolated peritoneal macrophages (PMFs) and stimulated them ex vivo with medium (control), LPS or LPS+WIN55,212-2 ( Figure 4D). WIN55,212-2 inhibited the production of TNFa and IL-6 by LPS-stimulated PMFs in a dose-dependent manner ( Figure 4E). Likewise, LPS-induced glucose consumption and lactate production was significantly impaired by WIN55,212-2, suggesting that this synthetic cannabinoid also inhibits LPSinduced metabolic rewiring in ex vivo stimulated PMFs ( Figure 4F). Supporting these data, real-time metabolic experiments showed a significantly reduced basal and compensatory glycolysis in LPS+WIN-stimulated PMFs compared LPS-stimulated PMFs, thus confirming our findings ( Figure 4G). Collectively, these data show the in vivo antiinflammatory capacity of WIN55,212-2 and its potential to immunomodulate tissue-specific macrophage populations during LPS-induced inflammatory conditions.

Discussion
In this study, we uncover unpreceded mechanisms about how the synthetic cannabinoid WIN55,212-2 displays immunomodulatory and anti-inflammatory properties by regulating the differentiation of human monocytes into DCs and the activation of pro-inflammatory macrophages. We showed that tolerogenic WIN-hmoDCs are less responsive to LPS and prime functional Tregs. In addition, WIN55,212-2 impaired the pro-inflammatory polarization of human macrophages by inhibiting cytokine production, inflammasome activation and rescuing macrophages from pyroptotic cell death. Mechanistically, WIN55,212-2 induced a metabolic and epigenetic shift in macrophages by decreasing LPS-induced mTORC1 signaling, commitment to glycolysis and active histone marks in proinflammatory cytokine promoters. These data were confirmed in ex vivo stimulated PMFs and supported by the anti-inflammatory capacity of WIN55,212-2 in a LPS-induced sepsis mouse model. Overall, we report novel insights into the anti-inflammatory capacity of WIN55,212-2 in myeloid cells, enhancing the knowledge on cell targets and mechanisms driving its potential beneficial effects in inflammatory conditions. Our findings might well contribute to open new avenues for the rational design of alternative cannabinoid-based therapeutic agents that are not able to cross the blood-brain barrier but retain anti-inflammatory/immunomodulatory properties on peripheral immune cells.
Monocytes are myeloid cells whose immune versatility relies on their capacity to differentiate into DCs or macrophages (4,21). The environmental milieu to what monocytes are exposed is linked to the immunogenicity of the resulting differentiated cells. For example, the presence of stimuli such as dexamethasone (Dex), vitamin D3 (VitD3), rapamycin, or allergoid-mannan conjugates promotes the differentiation of monocytes into tolerogenic DCs or immunosuppressive macrophages (22-26). Conversely, monocytes exposed with whole heat-inactivated Candida albicans or b-glucans from its cell wall give raise to trained monocyte-derived macrophages (27-29). Here, we show that LPS-stimulated WIN-hmoDCs displayed reduced production of the pro-inflammatory cytokines TNFa, IL-1b and IL-6 without changes in IL-10. Supporting these data, previous studies demonstrated that hmoDCs generated in the presence of THC produced less IL-12 after activation. THC also reduced the phagocytic activity, maturation state and the capacity of the differentiated hmoDCs to activate T cells (30). Likewise, WIN-hmoDCs primed FOXP3 + Tregs and displayed reduced capacity to polarize both Th1 and Th2 responses. The induction and/or expansion of Tregs is essential to keep homeostasis in the context of immune-mediated and inflammatory diseases (31)(32)(33), which might be regulated by exogenous signals including cannabinoids like cannabidiol (CBD), JTE907 or O-1966 (34-36). We previously showed that fully differentiated conventional hmoDCs and blood DCs subsets stimulated in vitro with toll-like receptor ligands in the presence of WIN55,212-2 promote functional Tregs (13,37,38), which was further corroborated in vivo in LPS-induced sepsis, peanut-allergic sensitization and anaphylaxis models (13,37). Our current data enhance the knowledge on how WIN55,212-2 could promote Tregs not only by directly acting on fully differentiated DCs, but also by imprinting tolerogenicity during monocyte differentiation into DC.
Macrophages are major producers of cytokines and chemokines that are essential for innate immune responses (3,4). In this work, we focused on the capacity of WIN55,212-2 to impair proinflammatory macrophage polarization. Previous studies reported that AEA, 2-AG, THC or CBD inhibit the production of inflammatory mediators in macrophages (39-41). WIN55,212-2 impaired NF-kB activation and cytokine release in LPS-stimulated TLR3/TLR4-transfected HEK293 cells, hmoDCs and murine macrophages activated with oxidized-LDL (13,42,43). Consistent with these reports, we showed that WIN55,212-2 impaired NF-kB/ AP-1-activation, decreased the expression of activation markers and the production of inflammatory cytokines in classically M1 activated THP-1 MFs, which was comprehensively validated in primary human macrophages (GM-MFs) and murine tissueresident macrophages (PMFs). WIN55,212-2 also restored GM-MFs viability after LPS/IFNg-induced inflammatory cell death, which correlated with a reduced release of IL-1b and impaired expression of NLRP3, CASP1 and GSDMD. These findings suggest that WIN55,212-2 interferes with inflammasome activation and protects from pyroptotic macrophage death, which is congruent with the previous reported capacity of CBD or THC to prevent inflammasome-mediated signalling (41, 44). Nonetheless, we cannot rule out the potential contribution of WIN55,212-2 to interfere with other cell death pathways, since the recently coined term PANoptosis recognizes the complexity of programmed cell death signalling and the potential interplay between its molecular executors (45, 46). In fact, it was recently described that caspase 8 and other apoptotic proteins were involved in the IFNg priming of macrophage TLR-induced death (47). The combination of IFNg with pro-inflammatory cytokines, mainly TNFa, also induces macrophage death and perpetuates inflammation during cytokine shock syndrome (48). Herein, the cannabinoid WIN55,212-2 dosedependently reduced the production of TNFa in LPS/IFNg-treated GM-MFs, suggesting its potential to prevent both cytokine storm and inflammatory cell death.
The anti-inflammatory capacity of cannabinoids has been linked with their ability to interfere with TLR-signalling and to induce metabolic reprogramming in immune cells (9, 49, 50). Our previous findings in human DCs indicate that the WIN55,212-2mediated activation of CB1 as well as of the non-canonical cannabinoid receptor, peroxisome proliferator-activated receptor a (PPARa), leads to a tolerogenic profile characterized by AMPKmediated induction of autophagy and promotion of oxidative phosphorylation over glycolysis (13). Herein, we show that WIN55,212-2 impaired LPS-induced phosphorylation of mTORc1's upstream and downstream signalling molecules, Akt and p70S6K, in human macrophages. Aerobic glycolysis is the main supply for the energetic demands of pro-inflammatory activated macrophages after TLR-engagement (51,52). Our real time metabolic experiments in LPS or LPS+WIN55,212-2-stimulated GM-MFs showed reduced basal and compensatory glycolysis in the latter ones, indicating inhibition of TLR-induced metabolic rewiring. In this context, Lauterbach and others demonstrated that glucose uptake and glycolysis provide metabolic substrates that support LPS-induced inflammatory gene expression through epigenetic histone modifications (18,19). LPS augments H3K27ac at IL1B and IL6 gene promoters in GM-MFs, which was impaired by WIN55,212-2, supporting that WIN55,212-2 may alter the LPSinduced metabolic-epigenetic axis in human macrophages. Beyond the potentiation of inflammatory cytokine expression, the induction of aerobic glycolysis also sustains inflammasome induction and its inhibition compromises IL-1b production (53-55). In addition, activation of mTORc1 pathway is linked to inflammasome activation, gasdermin D oligomerization and pyroptosis (56,57). Altogether these data might well support our findings on the capacity of WIN55,212-2 to prevent inflammatory cell death.
Systemic inflammatory diseases with uncontrolled release of pro-inflammatory mediators (cytokine storm), are lifethreatening conditions that remain poorly managed by current treatments (58)(59)(60). For instance, during endotoxin shock, the cooperative activation of inflammatory cytokine and cell death signalling drives severe tissue injury and multi-organ failure that may be eventually lethal (61,62). Here, WIN55,212-2 abolished the LPS-induced increase of serum IL-1b in mice, confirming its previously reported in vivo anti-inflammatory capacity (13,37,42,(63)(64)(65)(66). Inflammasome activation during sepsis aggravates organ injury and worsens prognosis. Suppression of inflammasome and pyroptosis protects mice from septic myocardial dysfunction, and clinical trial data shows reduced mortality after blockage of IL-1R in septic patients developing macrophage activation syndrome (MAS) (67,68). Patients that develop MAS were characterized by hepatobiliary dysfunction and had significantly higher risk of death (69). We showed that WIN55,212-2-mediated decrease of serum IL-1b was associated with reduced levels of the myocardial and hepatic injury markers LDH and bilirubin. Remarkably, we showed that cytokine production of ex vivo stimulated PMFs was dose-dependently modulated by WIN55,212-2 and associated with a markedly reduced glucose consumption and commitment to glycolysis. Our findings are congruent with the fact that inhibition of glycolysis with 2-DG mimicked WIN55,212-2 immunomodulatory effects in M1-activated murine alveolar macrophages and the capacity of WIN55,212-2 to protect mice from acute lung injury (63, 64).
Collectively, we uncover novel findings on the immunosuppressive capacity of WIN55,212-2 by acting on monocytes and macrophages and provide novel insights into the molecular mechanisms underlying such effects. Our in vivo data confirms WIN55,212-2 antiinflammatory properties and highlights the important involvement of myeloid lineage cells in its potential therapeutic benefits. Our data might well contribute to pave the way for the development of alternative cannabinoid-based therapies for immune-mediated and inflammatory diseases.

Data availability statement
The raw data supporting the conclusions of this article will be made available by the authors, without undue reservation.