Myeloid But Not Endothelial Expression of the CB2 Receptor Promotes Atherogenesis in the Context of Elevated Levels of the Endocannabinoid 2-Arachidonoylglycerol

The endocannabinoid 2-arachidonoylglycerol (2-AG) is an inflammatory mediator and ligand for the cannabinoid receptors CB1 and CB2. We investigated the atherogenic mechanisms set in motion by 2-AG. Therefore, we created two atherosclerotic mouse models with distinct cell-specific knockouts of the CB2 receptor on either myeloid or endothelial cells. These mice were treated with JZL184, resulting in elevated plasma levels of 2-AG. After a high-fat high-cholesterol diet, atherosclerotic plaques were analyzed. The atherogenic effect of 2-AG was abrogated in mice lacking myeloid expression of the CB2 receptor but not in mice lacking endothelial expression of the CB2 receptor. In vitro, treatment of human monocytes with 2-AG led to the increased production of reactive oxygen species (ROS) and IL-1β. In conclusion, 2-AG shows an atherogenic effect in vivo, dependent on the presence of the CB2 receptor on myeloid cells. In addition, our in vitro data revealed 2-AG to promote inflammatory signalling in monocytes. Graphical abstract 2-Arachidonoylglycerol shows an atherogenic effect that is abrogated in mice lacking myeloid expression of the CB2 receptor. Supplementary Information The online version contains supplementary material available at 10.1007/s12265-022-10323-z.


Introduction
The endocannabinoid 2-arachidonoylglycerol (2-AG) is a mediator of inflammation and ligand to the cannabinoid receptors CB1 and CB2, which are expressed on peripheral myeloid and human vascular endothelial cells [3,31]. The CB2 receptor has traditionally been considered to exert antiinflammatory and atheroprotective effects, such as decreasing macrophage accumulation in lesions, decreasing smooth muscle cell content in lesions, and decreasing apoptosis [28,31,32]. However, some conflicting data suggest that 2-AG can actually promote atherosclerosis and that the CB2 receptor may initiate inflammation as well as monocyte chemotaxis, infiltration, and activation under certain experimental conditions [5,7,9,12,22,24,30,40]. Multiple molecular pathways that promote atherogenesis have been discovered to date, yet the molecular effects of 2-AG in atherogenesis still need further exploration. Such molecular mechanisms include the generation of reactive oxygen species, the secretion of myeloperoxidase, and the activation of the IL-6/ IL-1β/CRP axis [2,6,14,16,26,27,33,35,36,39].
In the present study, we investigated the role of the CB2 receptor on myeloid and vascular endothelial cells in promoting atherosclerosis under elevated 2-AG levels. This was investigated by using two cell-specific CB2 knockout mouse models. Both mouse models had an atherogenic background (ApoE −/− ) and, additionally, distinct cell-specific knockouts of the CB2 receptor on either myeloid (ApoE −/− LysM cre CB2 fl/fl ) or endothelial (ApoE −/− Tie2 cre CB2 fl/fl ) cells. Mice were treated with intraperitoneal injections of JZL184, an inhibitor of the 2-AGdegrading enzyme monoacylglycerol lipase (MAGL), which led to elevated 2-AG levels. Meanwhile, they were fed a high-fat and high-cholesterol diet for 4 weeks. The volume and composition of the plaques formed were analyzed. In vitro, human monocytes were stimulated with 2-AG and a CB2 receptor antagonist. We investigated the production of reactive oxygen and nitrogen species (ROS, RNS), the transcription of NADPH oxidases (NOX), the release of myeloperoxidase (MPO), and the production of IL-1β under stimulatory conditions.

Materials and Methods
A detailed materials and methods section including a graphical summary of the experimental plan is available as an online supplement. In brief, two mouse models with an atherogenic background (ApoE −/− ; C57BL/6 J genetic background; Charles River, Wilmington, USA) and distinct cell-specific knockouts of the CB2 receptor (CB2 gene) on either myeloid (ApoE −/− LysM cre CB2 fl/fl ) or endothelial (ApoE −/− Tie2 cre CB2 fl/fl ) cells were created using the cre-lox system. Littermates without cell-specific cre-recombinase activity, and thus without any CB2 knockout (ApoE −/− CB2 fl/fl ), served as the control groups for both knockout strains. For each strain, the mice were divided into two groups, one of which was treated with 5 mg/kg body weight i.p. of the MAGL inhibitor JZL184 (Selleckchem, Munich, Germany), while the other group received the same volume of the vehicle PBS/Kolliphor/DMSO (AppliChem, Darmstadt, Germany) for 4 weeks. We chose a short treatment period of 4 weeks because we expected an increased plaque burden after treatment with JZL184. This protocol was established and validated in an earlier trial [12]. Approval was granted by the responsible German authority, the North Rhine Westphalian State Agency for Nature, Environment and Consumer Protection (reference number 84-02.04.2014.A419).
Data are presented as the mean ± SEM. Data were analyzed using Microsoft Excel (Microsoft, Redmond, USA) and GraphPad Prism software (GraphPad Software, San Diego, USA). For the comparison of continuous and normally distributed variables between two groups, an unpaired Student's two-sided t test was applied. For the comparison of three or more groups and normally distributed variables, a one-way ANOVA and subsequent Bonferroni correction were performed. In case of lacking Gaussian distribution, Kolmogorov-Smirnov test was applied for the comparison of two groups, and Kruskal-Wallis test with subsequent Dunn's multiple comparisons for nonparametric testing was performed for three or more groups. p values < 0.05 were considered statistically significant.

Results
In vivo, the impact of elevated 2-AG levels on atherogenesis was assessed in ApoE-deficient-mice (ApoE −/− ) combined with a distinct cell-specific knockout of the CB2 receptor on either myeloid (ApoE −/− LysM cre CB2 fl/fl ) or endothelial cells (ApoE −/− Tie2 cre CB2 fl/fl ). The effects of 2-AG in vitro were examined in human monocytes.

Pharmacological Inhibition of MAGL Increases 2-AG Levels in Both Myeloid and Endothelial CB2 Receptor Knockout Models
Two different mouse models with cell-specific knockouts of the CB2 receptor in either myeloid cells (ApoE −/− LysM cre CB2 fl/fl ) or endothelial cells (ApoE −/− Tie2 cre CB2 fl/fl ) were treated with the MAGL inhibitor JZL184 or vehicle for 4 weeks. Cre-negative littermates (ApoE −/− LysM wt CB2 fl/fl and ApoE −/− Tie2 wt CB2 fl/fl ) served as controls. All animals were fed a high-fat and high-cholesterol diet for the duration of the treatment.
Meanwhile, the plasma concentrations of arachidonic acid, as well as palmitoylethanolamide, were unaffected by JZL184 treatment (Supplementary Table S3).

Blood Pressure, Heart Rate, and Body Weight
Several clinical parameters, blood pressure, heart rate, and body weight, were acquired for mice of all genotypes and treatments. None of them, neither for the myeloid CB2 knockouts, the endothelial CB2 knockouts, nor the cre-negative animals, was affected by JZL184 treatment, as detailed in Supplementary Table S4.

Ex Vivo Assessment of Reactive Oxygen Species
In order to gain further mechanistic insights into the effects of 2-AG on atherogenesis, the formation of reactive oxygen species (ROS), as a marker of local inflammation, was measured in mice with/without elevated 2-AG levels in the presence and absence of the CB2 receptor on myeloid cells.

2-AG Increases ROS and RNS Production In Vitro
These differences in ROS production in vivo prompted us to examine the influence of 2-AG on ROS and RNS production in vitro. Therefore, we stimulated THP-1 cells with increasing concentrations of 2-AG and measured the ROS/RNS as well as superoxide production. Further, we aimed to determine the effect of CB2 on ROS/RNS production by the costimulation of THP-1 cells with 2-AG in combination with AM630, which is an inverse agonist for the CB2 receptor.

Neither the Release of MPO Nor the Transcription of NADPH Oxidase Seems To Be Responsible for the Increase of ROS and RNS
To further investigate the pathological mechanism behind the 2-AG dependent elevation of ROS/RNS and superoxide, we took a closer look at the generation of these species. Molecular oxygen is converted to superoxide by the NADPH oxidase 2 (NOX-2) or through its subunits cytochrome b-245 light chain (CYPA) and cytochrome b-245 heavy chain (CYBB). After the conversion of superoxide to hydrogen peroxide, it is converted to ROS by myeloperoxidase (MPO). Therefore, we measured CYBA, CYBB, and MPO concentration in the supernatants of THP-1 cells after stimulation with 2-AG.

Stimulation with 2-AG Increases the Concentration of IL-1β In Vitro
The IL-1β/IL-6 axis is an important signaling pathway for human and murine atherosclerosis. Earlier studies have shown a link between IL-1β and the ECS for inflammatory diseases other than atherosclerosis as well as neurological diseases [18,34]. This prompted us to examine whether the IL-1β/IL-6 pathway is involved in mediating endocannabinoid-associated effects in the presented cellular and murine models: IL-1β was measured in supernatants from PBMC after co-stimulation with 2-AG and LPS in various concentrations. We observed a dose-dependent increase in IL-1β release upon stimulation with 2-AG and LPS (2-AG [10 µM], 1864 ± 168.2 pg/ml, n = 12, p < 0.01; 2-AG [1 µM], 1383 ± 58.9 pg/ml, n = 12, p < 0.05; DMSO, 1329 ± 101.5 pg/ml) (Fig. 4a).

Discussion
The present study aimed to elucidate the influence of 2-AG and its receptor CB2 on atherogenesis. In the early stages of atherosclerosis, an atherogenic environment leads to activation of endothelial cells and increases leukocyte adhesion, while later on, monocytes are guided into the intima by chemokines and sustain continuous inflammation [20]. In the present study, we demonstrated that elevated concentrations of the endocannabinoid 2-AG promote atherogenesis in ApoE-deficient mice, accompanied by a significant increase in macrophages and neutrophils infiltrating the vessel wall. In the applied in vivo model, the effect of 2-AG was mitigated in animals lacking the myeloid CB2 receptor. In those mice, the plaque burden decreased, as did the number of infiltrating immune cells. These findings corroborate previous data by our own group and by others, where 2-AG was also reported to expedite atherogenesis, whereas inhibition of the CB2 receptor abolished this effect [12].
In our present study, pharmacological inhibition of MAGL also resulted in a higher collagen content in murine atherosclerotic plaques. Myeloid-specific depletion of the CB2 receptor abolished this effect. These data agree with previous findings by Vujic et al., where a constitutional knockout of MAGL in ApoE −/− mice (with subsequently elevated 2-AG plasma levels) led to an increased collagen content in atherosclerotic plaques [40].
However, an increased collagen content in atherosclerotic plaques must not be mistaken as a surrogate for an increased plaque stability in rodent models. There are counteracting roles of MMPs, either beneficial by mediating intimal thickening or harmful by thinning the extracellular matrix. Besides, there are several differences in the expression patterns and the pathophysiological role between humans and mice. For example, while MMP-1 is abundantly expressed in human endothelial cells and monocytes and contributes to plaque instability, MMP-1 expression in mice is sparse. However, in one study, artificial expression of human MMP-1 in ApoE −/− mice resulted in smaller atherosclerotic plaques [19,25]. These data are supported by the present study, where MMP-1 release from human monocytes increased upon stimulation with 2-AG, whereas 2-AG enhanced collagen deposition in WT mice. Therefore, the question whether a higher collagen content translates into increased plaque stability will have to be assessed in future experiments using the appropriate animal models or virtual histology studies in humans.
Interestingly, CB2 receptor deficiency in endothelial cells did not provide any benefit towards the plaque burden during the pharmacological elevation of 2-AG, although levels of infiltrating monocytes and neutrophils were found to be decreased. Thus, myeloid, and not endothelial, expression of the CB2 receptor appears to promote atherogenesis in the context of elevated 2-AG levels.
In an earlier study, we demonstrated that 2-AG promotes endothelial dysfunction and leukocyte recruitment [8]. 2-AG might contribute to atherosclerosis via endothelial cell activation, but in our present study, endothelial CB2 receptor activation does not appear to modulate the atherosclerotic plaque burden. One possible explanation might be that the endothelial damage in the present ApoE knockout mouse model is already so pronounced and that endothelial CB2 activation has little or no further impact on endothelial dysfunction. Meanwhile, CB2 receptor activation appears to be a fundamental mechanism in the activation of myeloid cells in the context of atherogenesis. This might explain why the knockout of endothelial CB2 receptor expression does not change the overall atherosclerotic plaque burden, while the knockout of myeloid CB2 receptor expression does so.
Treatment with JZL184 increased the amount of neutrophils within atherosclerotic plaques compared to DMSO control, even though this effect was not statistically significant in mice lacking the endothelial or myeloid CB2 receptor. While past studies on CB receptor expression on neutrophils were inconclusive, there is growing consensus that neutrophils do not express CB1 and CB2 receptors [4,37]. However, murine CB2 receptor depletion increased neutrophil recruitment in a murine model of endotoxemia, probably mediated by expanded MMP-9 secretion [15]. Another study demonstrated that 2-AG activates human neutrophils, which promotes the transmigration of additional neutrophils independent of the CB receptors [4]. Altogether, the observed accumulation of neutrophils might be a CB receptor-independent effect.
The CB1 and the CB2 receptors are both subject to autoinduction by their respective ligands [1,11]. In line with these data, we demonstrated in an earlier study that a decrease in 2-AG results in reduced CB2 receptor expression in vascular tissue of ApoE −/− mice. However, CB1 receptor expression in vascular tissue was marginal in this trial and not dependent on 2-AG levels [11]. On the contrary, another study demonstrated that the CB2 activity upregulates CB1 receptor expression in Jurkat cells [1]. Altogether, evidence about the interplay between CB1 and CB2 receptors in the context of elevated endogenous ligands is scarce [1]. Further research is needed, taking into account the cell-specific expression patterns of the respective receptors.
We and others have reported 2-AG to induce macrophage migration in vitro in a CB2-dependent fashion, which might explain the in vivo effects from our current study [12,17]. Moreover, the present study introduces two possible triggers for the atherogenic effect of 2-AG that are independent of the CB2 receptor: the production of ROS and the secretion of IL-1β.
It is well-known that ROS production is associated with endothelial dysfunction and inflammation and plays a crucial part in atherogenesis [16,36]. In addition, several studies provide evidence that the ECS is involved in the modulation of ROS/RNS production [21]. Therefore, we decided to quantify ROS production in murine aortic tissue and in cultured human monocytes. While 2-AG promoted ROS production in vitro, this effect was less pronounced in vivo. In vitro, the inverse CB2 receptor agonist AM630 attenuates basal MPO production, while 2-AG stimulation had no effect.
Several studies have underlined the importance of the IL-1β/IL-6 axis in atherosclerotic disease. For example, the CANTOS trial showed beneficial effects of targeting IL-1β [33]. For IL-6, a distinct homeostasis appears to be necessary to prevent atherogenesis: high levels of IL-6 are clearly connected to atherosclerosis, but on the other hand, atherosclerosis is enhanced in IL-6-deficient mice [6,36]. The endocannabinoid system influences the expression of IL-6 in several ways. While Chang et al. postulated that 2-AG hampers the expression of IL-6, Saroz et al. showed that activation of the CB2 receptor increases the induction of IL-6 [2,35]. We measured the levels of IL-6 in mice treated with JZL184. Interestingly, we could not detect any effect on IL-6 levels in JZL184-treated animals.
Meanwhile, our in vitro data suggest that 2-AG activates the NLRP3 inflammasome in a dose-dependent and CB receptor-independent fashion, leading to the liberation of proinflammatory IL-1β. While this is, to our knowledge, the first report to link 2-AG to NLRP3 activation, one previous study has demonstrated that the endocannabinoid N-arachidonoylethanolamine (AEA) promotes the excretion of IL-1β in a diabetic mouse model [13].
We subsequently measured IL-1β in plasma samples of atherosclerotic mice, but IL-1β levels remained consistently below the assay's lower limit of detection. Therefore, the activation of the IL-1β/IL-6 axis by 2-AG is an interesting novel finding in vitro, but it remains unlikely to be a driver of atherosclerosis in our mouse model. The current study underlines the importance of the endocannabinoid system to the pathophysiology of atherosclerosis in a murine model. Several studies by our lab and by others have also highlighted the importance of the endocannabinoid system to human cardiovascular disease [10,23,29,30,38]. However, the endocannabinoid system is difficult to use as a drug target, given its ubiquitous expression and its cell-and tissue-specific effects in various pathologies. Thus, cell-specific analyses of ECS effects are warranted to pave the way for the development of targeted therapies. Our present study may represent the first step in this quest.

Study Limitations
In the present study, we examined cellular and molecular effects that are crucial for the development of atherosclerosis. All data were generated in vivo in mice and in vitro in human models. Although those models are established and well-studied, the mechanisms may still vary in humans.

Conclusions
Taken together, data from the present study demonstrate that 2-AG promotes atherogenesis, whereas the absence of the myeloid CB2 receptor attenuates this effect. On the other hand, the endothelial CB2 receptor appears not to be involved in this mechanism. Three potential atherogenic effects of 2-AG were substantiated by our in vitro experiments: the increase in macrophage migration, the production of ROS, and the activation of the NLRP3 inflammasome. Meanwhile, the pathophysiological relevance of these effects in human atherosclerosis remains to be elucidated.