In the present study, it could be proven that the endocannabinoid system is involved in two different canine inflammatory CNS diseases.

In the last decades, evidence has shown that the functions of the endocannabinoid system are not limited to the CNS but are involved in the whole organism of mammals [42]. Although the endocannabinoid system has been widely studied for its involvement in regulation of neurotransmission [6, 43, 44], increasing evidence supports its involvement in immunomodulation [5, 15, 18, 19, 23, 25] and neuroprotection [45–47]. CB receptors are expressed by leukocytes [14, 20], mice with deficient CB receptors display an altered, usually more severe, inflammatory phenotype [18]. Therefore, the enhancement of the endocannabinoids or the activation of cannabinoid receptors may have valuable therapeutic effects [48].

In accordance with previous studies in other species [39, 49, 50] and in dogs [34], overall CSF levels of AEA were in the picomolar range (pM) and of total AG in nanomolar range (nM). Levels of 2-AG in unstimulated tissues and cells are usually much higher than those of AEA, and are in principle sufficient to permanently activate both cannabinoid receptors [11, 51]. Interestingly, 2-AG spontaneously isomerizes to its biologically inactive form 1-AG, by acyl migration [39]. This phenomenon is of particular importance for quantification of 2-AG. Therefore, both 1-AG and 2-AG were measured and summarized to total AG, since such calculation has been proven to be more accurate [39].

The highest levels of AEA and total AG in CSF were found in dogs with Intraspinal Spirocercosis and differed significantly from all other groups. High values occurred also in SRMA A group that showed significantly higher levels of both AEA and total AG compared to SRMA Tr. Spirocercosis in dogs has been mostly associated with the presence of esophageal granulomas [52], and although the disease is frequently subclinical, esophageal dysphagia is considered the clinical hallmark [52]. Aberrant migration of the worm to unusual anatomical structures results in atypical clinical signs [28]. Neurological deficits presented in the IS group are similar to previous findings of extradural and intraspinal Spirocerca lupi migration [30, 53]. Although the reason for aberrant migration is still unclear [28], it possibly occurs via the intercostal arteries, originating from the thoracic aorta, through their spinal branches and into the extradural space [53]. The worm possibly penetrates the dura and pia mater to enter the spinal cord parenchyma [29]. Eosinophils play a major role in dealing with elimination of parasites [54]. CSF samples from the IS group showed mixed neutrophilic pleocytosis with moderate to severe eosinophilia and foamy macrophages. CSF findings agree with previous CSF analysis in single cases of dogs with intraspinal migration [29, 55]. This disease was chosen for the current study, to include a canine population with inflammatory CNS disease and eospinophilia in CSF samples to compare with SRMA dogs and a clear neutrophilic pleocytosis. All CSF samples in the SRMA A group displayed a neutrophilic pleocytosis, the key feature of the acute stage of SRMA [32, 56, 57].

ECs are lipids that are able to modulate cell migration through specific receptors [1]. Both in vitro and in vivo studies have shown CB2 receptor involvement in eosinophil migration [1], which is supported by the current study. Human peripheral eosinophils express functional CB2 receptors that mediate chemotaxis towards 2-AG, but not AEA [58]. Interestingly, another study showed that 2-AG induces migration of eosinophils in a dose-dependent manner and that the concentration required appears to be pathophysiologically relevant [59]. Since 2-AG is a fully effective CB2 agonist [60], several studies have focused on the immunomodulatory effects of 2-AG, while information of AEA in this aspect is lacking. Moreover, 2-AG has been proven to be produced and released from platelets [61], macrophages [61, 62], endothelial cells [63, 64], glial cells [65], macrophages [62] and adipocytes [66]. The main 2-AG effects on neutrophils seem to be independent of CB activation in humans [67]. Although 2-AG does not seem to have a chemotactive effect on neutrophils, it has been proven to activate them [67].

CSF levels of AEA and 2-AG in healthy dogs were low as already determined in previous measurements [34]. Such low values were also found in dogs under treatment with glucocorticosteroids (SRMA Tr group). Increased levels of AEA have been found in CSF in dogs with epilepsy and correlated with disease severity and duration [34]. AEA seems to regulate seizure threshold in epilepsy [34]. AEA is a high affinity ligand for CB1 receptors [60], mostly expressed in CNS tissue and regulating neurotransmission [68]. AEA CSF levels in SRMA were much higher than in dogs with epilepsy. 2-AG is the fully effective CB2 agonist [60], but AEA has been linked to activate CB2 receptors in pathological conditions [69] and could enhance or regulate the fulminant inflammatory response in SRMA.

AEA concentration in serum was significantly higher in all groups (Intraspinal Spirocercosis, SRMA A, and SRMA Tr) compared to healthy controls. Moreover, dogs suffering from SRMA in the acute phase and those with Intraspinal Spirocercosis showed significantly higher concentrations of total AG than healthy controls in serum. SRMA is a systemic immune-mediated disorder [32]. Even when the main lesions are found in the cervical leptomeninges, SRMA causes systemic inflammatory lesions of the vessels of the heart and intestine [70, 71]. Albeit migration and persistence of larvae or adult worms are responsible for intraspinal lesions in Spirocercosis [72], the tissue damage is not restricted to the spinal cord in these dogs. The systemic nature of both diseases might explain the increased levels of both ECs in serum.

Strong CB2 immunoreaction was found in leukocytes within the marginal zone of splenic follicles and hepatocytes and Kupffer cells as previously described [20, 73, 74]. Previous studies performed in healthy dog tissues have shown strong CB2 immunoreaction in B cell zones of lymphoid follicles, and diffuse immunoreaction in cells in the dermis, including perivascular cells with mast cell morphology, fibroblasts, and endothelial cells [75]. There is however a lack of studies describing CB2 receptor expression in the CNS of dogs. Our study showed moderate to strong CB2 expression on more than 90% ramified glial cells within the grey and white matter of the cervical spinal cord sections of healthy dogs. Moreover, slight to moderate immunoreaction was found in the neurons in the ventral and dorsal horns. CB2 receptor expression, was initially thought to occur only in peripheral tissues (mostly immune organs and cells) [1, 3, 14, 20]. However, several studies have demonstrated CB2 expression in microglia [76–79], oligodendrocytes, astrocytes and endothelial cells in the healthy CNS [80–83]. The expression of CB2 receptors in neurons remains controversial [84]. Nonetheless, functional CB2 receptors have been found in neuronal subpopulations in cerebellum, brainstem, cerebral cortex [85–88], although less concentrated than CB1 receptors [85]. In SRMA and IS CB2 expression seemed to be upregulated both in glial cells and neurons. However, CB2 expression depended on lesion severity and was even downregulated in glial cells in regions with severe necrosis and cell loss. Interestingly, CB2 positive glial cells underwent changes in the morphology in SRMA and IS toward a round morphology showing less and shorter cytoplasmic extensions. Although further studies are needed to exactly confirm the type of glial cells expressing CB2, it might be speculated that the change in morphology is a consequence of astrogliosis and/or correlated with the occurrence of activated microglia. In other studies, upregulation of CB2 receptor expression in microglia, macrophages, neurons and astrocytes upon activation has been demonstrated [5, 86, 89, 90]. An increased production of AEA and 2-AG has also been described in several other pathologies of the CNS [48, 91, 92]. The understanding of the role of AEA in inflammation is particularly important regarding its interaction with microglia [93]. Depending on the relative quantities of 2-AG and AEA that accumulate in certain CNS areas, their competition at CB2 receptors may determine the extent to which CB2 receptors will regulate microglial cell behavior and phenotype, towards a M1 pro-inflammatory phenotype or a M2 anti-inflammatory phenotype [93]. Indeed, CB2 can be induced on demand during early inflammatory events and it has been linked to attenuation of pro-inflammatory cytokine production by microglia, counteracting neuronal damage [1, 42]. On the other hand, upon CB2 activation microglia have also been shown to cause pro-inflammatory actions by eliciting cell migration [65].

CB2 was strongly expressed on infiltrating leukocytes (i.e. neutrophils, lymphocytes, plasma cells, eosinophils and macrophages) in spinal cord lesions of dogs with SRMA and IS. Interestingly, arachnoid mesothelial-like cells and cells in the pia matter near the lesions expressed also CB2. Immune cells express high levels of CB2, and even a hierarchy of CB2 receptor expression has been described within the immune system (B cells > natural killer cells > monocytes > neutrophils > CD8 lymphocytes > CD4 lymphocytes) [14]. The level of expression, however, is dependent on the activation state of the cell and the type of stimuli [20]. CB2 receptor expression in macrophages has been proven to be modulated upon activation. CB2 is detected at low levels in resting cells, while it is present at high levels in activated macrophages [5, 89].

Endocannabinoids have been shown to downregulate inflammation in numerous experimental models [20–26]. Activation of the cannabinoid system has been linked to decreased inflammatory cell recruitment and enhanced anti-inflammatory cytokine production [18]. Evidence shows that exogenous application of AEA and 2-AG exerts anti-inflammatory effects by decreasing the production of inflammatory mediators [19]. Interestingly, increased levels of AEA and 2-AG have been found in the spinal cord of rats after moderate-severe spinal cord injury (SCI) [48, 94]. The rapid activation of cannabinoid receptors by endocannabinoids after SCI is thought to be an endogenous protective response [48]. In another study, Arevalo-Martin and colleagues showed that one single injection of 2-AG 30 minutes after the induction of the lesion, protects the white matter from secondary damage, reducing the lesion expansion and myelin loss [46], while CB1 and CB2 antagonists increased myelin damage [48]. Strikingly, administration of CB1 and CB2 antagonists impaired the spontaneous motor recovery observed in rats after SCI [48].

Although several studies support endocannabinoids as anti-inflammatory mediators, several in vitro and in vivo experiments have reported a pro-inflammatory role of the endocannabinoid system in the development of inflammation [18]. These pro-inflammatory effects have been associated with enhanced leukocyte recruitment and activation, production of reactive oxygen species and release of pro-inflammatory cytokines [1, 58, 59, 65, 95, 96]. Nonetheless, most of these pro-inflammatory effects attributed to endocannabinoids involve 2-AG, but not AEA [18]. Indeed, AEA is a potent anti-inflammatory endocannabinoid and acts practically on all cell subsets except natural killer cells and B cells, while 2-AG exerts both pro- and anti-inflammatory effects which seem to be strictly dependent on cell type [97]. The anti-inflammatory effects of AEA may explain the fact that in SRMA Tr AEA serum levels remained high, while the pro-inflammatory total AG levels decreased notably in serum in the SRMA Tr group.

The endocannabinoid system response in inflammation is indeed dependent on type and state of the disease. While CB2 receptor activation in general mediates immunosuppressive effects limiting inflammation and being associated with tissue damage in several pathological conditions, in some disease states activation of the CB2 receptor may enhance or even trigger tissue damage [84]. In physiological conditions, the endocannabinoid production by neurons is high and in microglia low [98]. In diseased CNS tissue with an activated immune system the cell-specific expression profile of cannabinoid receptors changes, resulting in higher expression of CB2 receptors in activated microglia [16, 98], which can secondarily activate astrocytes leading to further induction of expression of inflammatory factors [5]. Such an upregulation on activated microglia cells was also seen in the evaluated canine inflammatory CNS diseases and could be the reason for the excessive immune response in SRMA. As the disorder progresses, the blood brain barrier (BBB) becomes partially disrupted and blood macrophages, B cells, T cells, natural killer cells and other infiltrating leukocytes start upregulating CB2 [5, 42, 93]. The activation of the receptors further stimulates the migration and/or activation of these cells into the nervous tissue and towards the endocannabinoids produced by microglia, neurons [42] and leukocytes [61, 62, 98], which might explain eosinophilia in CSF samples of IS dogs in the current study after extremely high endocannabinoid levels.

Although the initial enhancement of the endocannabinoid system has been proven to have an endogenous protective response and attempts to control inflammation, as the disease progresses, a later dysregulation of the endocannabinoid system during neuroinflammation leads to increased CB2 receptor expression, excessive endocannabinoid production leading to an exacerbated inflammatory response. Thus, both CB2 agonists and antagonists might be beneficial in counteracting the inflammatory consequences depending on the disease phase [42]. Increased AEA, total AG and upregulation of CB2 receptors might therefore explain the waxing and waning course of SRMA.