Spatial distribution of cannabinoid receptor type 1 (CB1) in normal canine central and peripheral nervous system

The endocannabinoid system is a regulatory pathway consisting of two main types of cannabinoid receptors (CB1 and CB2) and their endogenous ligands, the endocannabinoids. The CB1 receptor is highly expressed in the central and peripheral nervous systems (PNS) in mammalians and is involved in neuromodulatory functions. Since endocannabinoids were shown to be elevated in cerebrospinal fluid of epileptic dogs, knowledge about the species specific CB receptor expression in the nervous system is required. Therefore, we assessed the spatial distribution of CB1 receptors in the normal canine CNS and PNS. Immunohistochemistry of several regions of the brain, spinal cord and peripheral nerves from a healthy four-week-old puppy, three six-month-old dogs, and one ten-year-old dog revealed strong dot-like immunoreactivity in the neuropil of the cerebral cortex, Cornu Ammonis (CA) and dentate gyrus of the hippocampus, midbrain, cerebellum, medulla oblongata and grey matter of the spinal cord. Dense CB1 expression was found in fibres of the globus pallidus and substantia nigra surrounding immunonegative neurons. Astrocytes were constantly positive in all examined regions. CB1 labelled neurons and satellite cells of the dorsal root ganglia, and myelinating Schwann cells in the PNS. These results demonstrate for the first time the spatial distribution of CB1 receptors in the healthy canine CNS and PNS. These results can be used as a basis for further studies aiming to elucidate the physiological consequences of this particular anatomical and cellular distribution.


Introduction
The properties for medical intervention of the plant Marijuana (Cannabis sativa) have been known for centuries [1,2]. Behavioural and pharmacological effects of its most psychoactive component, Δ 9 -tetrahydrocannabinol (THC), can be explained by the activation of receptors localized in the nervous system [3,4] and peripheral tissues [5]. These receptors are known as PLOS  cannabinoid receptors (CBs), and along with their endogenous ligands, the endocannabinoids (ECs), and the enzymes responsible for their synthesis and degradation, constitute the endocannabinoid system [6,7]. In mammalian tissues, two main subtypes of cannabinoid receptors, the cannabinoid receptor 1 (CB 1 ) and cannabinoid receptor 2 (CB 2 ), which are G protein-coupled receptors, have been recognized [5,8] and are responsible for the transduction of different effects of ECs [9]. Furthermore, CB 1 receptors have been shown to be primarily expressed in the central nervous system (CNS) and peripheral nervous system (PNS) while CB 2 receptors are mostly found in cells of the immune system [10,11].
Besides the therapeutical effects of several cannabinoids as antiemetics, analgesics, antispasmodics, appetite-stimulating, and sleeping inductors [12]; THC, a phytocannabinoid partial CB 1 agonist, as well as WIN55,212-2, a synthetic CB 1 agonist, have both been proved to have an anticonvulsant effect in vitro [13] and in rodent models [14] of epilepsy and status epilepticus, being more effective than conventional antiepileptics like phenytoin and phenobarbital [15]. Furthermore, increased levels of anandamide (AEA), an endocannabinoid, have been found in cerebrospinal fluid of dogs suffering from idiopathic epilepsy compared to healthy dogs [16]. Several companies started to sell medical marihuana to be used in pets to treat chronic pain, seizures, inflammation, cancer, diabetes, nausea, anxiety and obesity. There is, however, a lack of reliable research to back those claims regarding the specific distribution of cannabinoid receptors and their associated function according to their presence in different anatomical localization within the healthy nervous system and under pathological conditions.
The expression of CB 1 has been described in brain sections of humans using autoradiography [3,17] and in rhesus monkeys by positron emission tomography (PET) [18]. Immunohistochemistry allows the identification of particular neuronal cells and fibres that express cannabinoid receptors because of its greater resolution [19]. Consequently, CB 1 distribution has been extensively mapped in the mouse [20], rat [19,21] and macaque monkey [22] CNS. In addition, CB 1 expression has also been described particularly in the dorsal horn in rats [19,23] and in the spinal cord of humans [17]. In the species dog, CB 1 receptors were detected in salivary glands [24], hair follicles [25], skin and hippocampus [26]. However, a detailed analysis of the distribution of CB 1 receptors in the CNS and PNS has not been reported in canines so far. It is well established that many conditions in dogs share striking similarities with their human counterparts thus representing suitable translational models for studying human neurological diseases including epilepsy [27], neuropathic pain [28], spinal cord injury [29] and multiple sclerosis, as described in the canine distemper virus (CDV)-induced demyelination model [30]. Thus, the species dog might help to overcome the gap between highly homogenous and standardized rodent models and clinically relevant conditions in humans. Precise knowledge of the distribution of CB 1 within the canine nervous system are therefore of great relevance to design therapeutic strategies to manipulate the effects of the endocannabinoid system [9].
In the current study, we analyzed the spatial distribution of CB 1 receptors in the healthy CNS and PNS of dogs from different ages. This is the first study which characterizes in detail the presence of those receptors under normal circumstances, therewith providing novel insights into the localization of CB 1 receptors for further characterization under pathophysiological conditions. fixed in non-buffered formalin (10%) for at least 48 hours and embedded in paraffin. Serial sections (3 μm thick) were mounted on SuperFrost-Plus slides (Menzel Gläser, Braunschweig, Germany), and stained with hematoxylin and eosin (HE), a complete histological examination was performed in order to confirm the absence of histopathological lesions. Afterwards, the slides where further processed for immunohistochemistry and double immunofluorescence. A total of five dogs of different ages were included, one female and two male six-month-old Beagle dogs, one ten-year-old female Cocker Spaniel and one four-week-old female Leonberger. Tissue samples of the dogs used in this study were included in a previous study [31]. German Animal Welfare Act with the law of animal welfare, Germany (permission number: 33.9-42502-05-13A346), and the ethical guidelines of the University of Veterinary Medicine Hannover were followed for the euthanasia of the dogs. No animals were euthanized for this particular study; samples obtained and previously used in other studies were taken. The study was approved and followed the guidelines of the PhD commission of the University of Veterinary Medicine Hannover, the institutional ethics committee.
Transversal sections were cut through the brain at the level of olfactory bulb, frontal lobes, basal forebrain, thalamus, lateral and medial corpus geniculatum, hippocampus, cerebellum and brainstem. Transversal sections of the cervical, thoracic and lumbar spinal cord with their corresponding dorsal root ganglia were included, as well as a representative section of the sciatic nerve.

Double immunofluorescence staining
Double immunofluorescence staining was performed on representative tissue sections as previously described [32] on 3 μm thick paraffin-embedded sections to demonstrate a possible co-localization of CB 1 with GFAP and CNPase in the CNS, and P0 and p75 NTR in the PNS. Briefly, sections were simultaneously incubated with the respective primary antibodies for 90 min. Cy3-labeled goat anti-mouse (red, 1:200, Alexa Fluor 555 dye, Life Technologies) and Cy2-labeled goat anti-rabbit (green, 1:200, Alexa Fluor 488 dye, Life Technologies) secondary antibodies were used to visualize the respective antigens. Nuclear counterstaining was performed with 0.01% bisbenzimide (H33258, Sigma Aldrich, Taufkirchen, Germany) and sections were mounted with Dako Fluorescent Mounting medium (DakoCytomation, Hamburg, Germany). Antigenic expression was visualized using an inverted fluorescence microscope (BZ-9000E, Keyence GmbH, Neu-Isenburg, Germany) and examined through the BZ-II Analyzer software. All images were acquired with the same microscope settings under which control sections showed no signal. Images were transferred to Adobe Photoshop (San Jose, CA) for cropping, and they were adjusted to optimize contrast and brightness.

Results
The distribution of CB 1 immunoreactivity in anatomically related regions is described below in detail. Importantly, there were few differences in the expression of CB 1 regarding the analysed anatomical localisations in different aged dogs. Generally, strong cytoplasmic CB 1 immunoreactivity was observed in astrocytes both in the white and in the grey matter along the cerebrum (Fig 1A), cerebellum and spinal cord in all dogs, except in the four-week-old dog, in which only scattered astrocytes were slightly positive ( Fig 1B). In addition, the cytoplasm of ependymal cells lining the lateral (Fig 1C and 1D), third, fourth ( Fig 1E) ventricles and the central canal of the spinal cord; as well the choroid plexus ependymal cells (Fig 1F) strongly expressed CB 1 . Strikingly, the cytoplasm of small numbers of neuroglial cells surrounding the fourth ventricle ( Fig 1E) and the central canal of the spinal cord were intensely CB 1 positive. Within the meninges, flattened fibroblast-like cells mostly in the dura mater showed slight cytoplasmic CB 1 immunoreactivity.

Olfactory bulb
In the main olfactory bulb, network of fibres were intensely stained with CB 1 in the glomerular layer (GL) (Fig 2A). CB 1 immunoreactivity occurred in a network of fibres that surrounded unstained neuronal soma ( Fig 2B). Immunoreactivity was also found in the fibres of the internal plexiform layer (IPL) (Fig 2A). In addition, a population of cells within the internal granule cell layer were strongly CB 1 positive (Fig 2C). No immunoreactivity was observed in the external plexiform layer (EPL) or in the mitral cell layer (ML). However, mitral cells axons were moderately CB 1 positive. In the four-week-old dog, only the glomerular layer expressed moderate CB 1 immunoreactivity and all other layers were devoid of immunostaining ( Fig 2D).

Cerebral cortex (neocortex-frontal lobe)
The grey matter of the neocortex expressed strong CB 1 immunoreactivity in the external granular layer (II), external pyramidal layer (III), inner granular layer (IV), inner pyramidal layer (V) and multiform layer (VI) (Fig 3A and 3B). This intense immunoreactivity of the fibres was presented in a dot-like pattern surrounding the unstained neuronal bodies (Fig 3A). The densest expression was found in the II, III, V and VI layers of the frontal lobe while the molecular layer (I) appeared almost devoid of CB 1 immunostaining ( Fig 3B). In the four-week-old dog and the ten-year-old dog, the intensity of the immunoreactivity was lower comparing to the other dogs.

Hippocampus
Within the hippocampus, strong dot-like CB 1 immunostaining was associated with a dense network of fibres in the stratum pyramidale surrounding the unstained pyramidal neuronal bodies (Fig 3C and 3E). A progressive decrease in the immunoreactivity from CA1 to CA4 was seen ( Fig 3C). In the hippocampal polymorphic layer and the molecular layer, the fibres were less intensely stained. In the dentate gyrus, CB 1 immunoreactivity was associated with fibres in the molecular layer with the most intense staining occurring adjacent to the granule cell layer ( Fig 3D). The granule cell layer lacked CB 1 expression ( Fig 3D).   figure B, layer I appears almost devoid of CB 1 immunoreactivity and layer II and III express strong immunoreactivity. Notice that the meninges show positive flattened fibroblast-like cells in the dura matter (B, arrow). The hippocampus shows progressive decrease in the immunoreactivity from C1 to C4 (C). In figure D, the dentate gyrus of the hippocampus depicting strong dot-like CB 1 immunoreactivity in the molecular layer (ML). Interestingly, the granule cell layer (GCL) appears devoid of CB 1 immunoreactivity. The stratum pyramidale shows strong CB 1 immunoreactive fibres surrounding unstained pyramidal neuronal bodies in the C1 (E; arrow). In figure F, strong CB 1 immunoreactivity is observed in fibres of the substantia nigra pars reticulata. IHC was performed using the avidin-biotin-peroxidase complex (ABC) method. WM: white matter; GM: grey matter.

Basal ganglia and lateral and medial geniculate nucleus
Intense CB 1 immunoreactive fibres were observed in the globus pallidus of the basal nuclei. The expression of CB 1 was observed only in scattered fibres in the lateral and medial geniculate nucleus.

Midbrain
Strong CB 1 immunoreactivity was observed in the fibres surrounding unstained neuronal bodies in the substantia nigra, denser towards pars reticulata ( Fig 3F). There were intensely stained fibres from all directions at the level of the oculomotor nucleus and red nucleus. In addition, moderate CB 1 immunoreactivity was observed in the fibres of the periaqueductal gray (PAG) and in the soma of neurons.

Cerebellum
Strong CB 1 immunoreactivity was observed homogenously within the molecular layer of the cerebellar cortex (Fig 4A and 4B). Small numbers of Purkinje cells showed slight cytoplasmic immunoreactivity. Interestingly, strong immunoreactivity was present surrounding the Purkinje cells bodies, particularly in the basal portion of the cells (Fig 4B). The underlying granule cell layer remained negative, with just few scattered positive fibres surrounding unstained cellular bodies ( Fig 4B). In the ten-year-old dog the staining pattern remained alike, however, the molecular layer showed moderate to slight CB 1 immunoreactivity, while other layers remained negative (Fig 4C and 4D), Purkinje cells were surrounded by dot immunoreactivity (Fig 4D).

Medulla oblongata
Surrounding the neuronal bodies of the cochlear nucleus ( Fig 4E) and the nucleus of the spinal tract of the trigeminus, a strong dot-like CB 1 immunoreactivity was observed while the neuronal cytoplasm were completely negative ( Fig 4F).

Spinal cord
Within the grey matter of the cervical, thoracic and lumbar spinal cord, strong CB 1 immunoreactive fibres were observed in the dorsal horn, intermediate region and ventral horn ( Fig 5A). CB 1 dot-like immunostaining was present surrounding the neuronal bodies ( Fig 5B). In addition very few neurons showed slight cytoplasmic immunoreactivity within the ventral and dorsal horns (Fig 5B). In the four-week-old and the ten-year-old dogs, the intensity of the immunoreactivity in the grey matter was lower comparing to the six-month-old dogs (Fig 5C).

Dorsal root ganglia
Within the dorsal root ganglia (DRG) of the cervical, thoracic and lumbar spinal cord segments large neurons showed slight cytoplasmic CB 1 immunoreactivity, while small dark neurons strongly expressed CB 1 (Fig 5D). In addition, satellite cells were strongly immunopositive ( Fig 5D). In the ten-year-old dog the immunostaining pattern remained, nevertheless, the overall DRG immunostaining was weaker (Fig 5E). In the four-week-old, however, only scattered large and small dark neurons and satellite cells were slightly positive ( Fig 5F).  Figure B depicting in detail immunonegative Purkinje cells surrounded by strong immunorreactive fibers particularly in the basal portion (arrow). In the ten-year-old dog, there is a slight immunoreactivity in the molecular layer of the cerebellar cortex (C). Purkinje cells surrounded by a dot-like immunoreactivity appear devoid of immunoreactivity in the ten-yearold dog (D; arrow). The cochlear nucleus in a six-month-old dog showing strong CB 1 immunoreactivity (E). In figure F detail of the cochlear nucleus with strong CB 1 immunoreactivity surrounding the unstained neuronal bodies (arrow). IHC was performed using the avidin-biotinperoxidase complex (ABC) method.
Peripheral nerve CB 1 immunostaining within the thoracic spinal nerve revealed strong expression in randomly distributed Schwann cells ensheathing axons (Fig 5G). In the four-week-old and the ten-yearold dogs, the intensity of the CB 1 immunoreactivity was lower in positive Schwann cells ( Fig  5H and 5I). Moreover, in the ten-year-old dog only few Schwann cells showed a moderate positive immunoreactivity (Fig 5H).

Double immunofluorescence tracking of glial cells expressing CB 1 receptors in the CNS and PNS
In order to specifically identify the glial cell types expressing CB 1 receptors in the CNS and PNS, double immunofluorescence was performed in a representative case of a six-month-old beagle dog. Co-localization of CB 1 with the astrocytic marker GFAP was observed in about 20% of the GFAP-positive astrocytes, indicating that only a subpopulation of astrocytes does express CB 1 receptors (Fig 6A and 6C). On the other hand, no co-expression was present among CB 1 and the mature oligodendrocytic marker CNPase (Fig 6D and 6F).
Interestingly, double immunolabelling of the sciatic nerve showed co-localization of CB 1 and P0, a marker for myelinating Schwan cells, in about 100% of the Schwann cells (Fig 7A  and 7C). On the contrary, no co-expression was found among CB 1 and the non-myelinating Schwann cells marker p75 NTR (Fig 7D and 7F).

Discussion
This study describes the first detailed spatial distribution of CB 1 receptors in the healthy canine CNS and PNS. A commercially available antibody against human CB 1 was used, corresponding to C terminal amino acids 461-472 of human cannabinoid receptor type 1. The CB 1 protein sequence is highly conserved across mammalian species [33], moreover, crossreactivity of this specific CB 1  The presence of cannabinoid receptors in the canine CNS was first reported by the autoradiographic studies of Herkenham in 1990, who found the densest binding of a radiolabeled synthetic cannabinoid in the cerebellar molecular layer, followed by the globus pallidus, substantia nigra pars reticulate, hippocampal dentate gyrus and the neocortex [3]. This pattern, with few variations remained for humans, rhesus monkeys, rats and guinea pigs [3]. Our study using immunohistochemistry shows similar results regarding the overall distribution of CB 1 in healthy CNS of dogs.
In the current study, the distribution of CB 1 immunostaining consisted mostly on strongly positive network of fibres in specific regions such as the olfactory bulb, cerebral cortex, cerebellar cortex, hippocampus, basal ganglia, cochlear nucleus, nucleus of the spinal tract of the trigeminus and grey matter of the spinal cord. This particular distribution might be due to the fact that the CB 1  Strikingly, ependymal cells lining the ventricular system and the central canal of the spinal cord, and a small numbers of neuroglial cells surrounding the fourth ventricle and the central canal expressed the strongest cytoplasmic CB 1 immunoreactivity. Cells surrounding the central canal of the spinal cord are a source of stem/precursor cells that may give rise to neurons, astrocytes, or oligodendrocytes [49]. The ependymal region in the spinal cord has been shown to express CB 1 in rodents [50] and humans [51]. Even a subpopulation of ependymal cells named "CB 1 high cell" has been described in both species [49,51], co-expressing stem/precursor cell markers in rats [49]. Control of proliferation of brain progenitors/stem cells through CB receptor activation has been shown in vitro [52,53]. Furthermore, "CB 1 high cells" proliferate during early postnatal development and after spinal cord injury (SCI) in adult rats, but not in the unlesioned spinal cord [49]. Aguado and others showed that endocannabinoid signalling controls neural progenitor differentiation into astroglial cells in postnatal and adult mice [52]. Nevertheless, further studies are needed to fully understand the potential of cannabinoids on neurogenesis in the dog and other species [49].
In the olfactory bulb, we demonstrated a network of fibres intensely stained with CB 1 in the glomerular layer and moderate staining of the internal plexiform layer surrounding unstained neuronal soma. Despite lack of immunostaining of the mitral cell layer, mitral cell axons were moderately CB 1 positive. Furthermore, a population of cells located in the internal granule cell layer expressed strong cytoplasmic CB 1 immunoreactivity. In mice [54] and rats [19,21], the strongest CB 1 immunoreactivity has been detected in the fibres of the inner granule cell layer, followed by the inner plexiform layer, however, only surrounding unstained cell bodies. Soria-Gómez and others demonstrated that the endocannabinoid system controls food intake via olfactory processes in mice [54]. CB 1 receptors are described to be densely expressed in all regions of the cortex in mice, rats, monkeys and humans [17,19,22,55]. While the general laminar pattern of CB 1 immunoreactivity between species seems preserved [9], the densest expression appears to be in the III and V in primates [9] and within layers II, III and VI in mice [21] and rats [19], while layer I appears almost devoid of CB 1 receptors in these species [9,21]. We found intense CB 1 immunoreactivity presented in a dot-like pattern of the fibres surrounding unstained neuronal bodies, as previously described [21], within layers II, III, IV, V, and VI of the neocortex, with the densest expression found in the II, III, IV and VI layers of the frontal lobe. In humans and monkeys, the laminar pattern has been widely studied and shows distinctive laminar density across the different regions, showing higher CB 1 expression in the prefrontal cortex [22]. In the monkey, neocortex immunoreactivity is primarily contained in cells and axon terminals that show morphological features of GABAergic neurons [22]. Moreover, the majority of high CB 1 expressing cells in the rat forebrain are GABAergic neurons [56].
The strongest immunoreactivity we found in the hippocampus was associated with fibres within the molecular layer of the dentate gyrus with the most intense staining occurring adjacent to the granule cell layer, which lacked CB 1 expression. A similar pattern has been found in macaques [22], mice [20] and rats [19], where the highest CB 1 density was found within the molecular layer of the dentate gyrus, while the granule cell layer appeared to be completely devoid of CB 1 immunoreactivity. In immunohistochemical studies made in rats [19], mice [20, 57] and macaques [22] strong immunoreactivity occurred in the Cornu Ammonis (CA) regions of the hippocampus within the pyramidal layer. Interestingly, the cell bodies of pyramidal neurons in CA1-CA3 fields appeared to be unstained but surrounded by a dense plexus of highly immunoreactive fibres [19][20][21][22]57]. Indeed, we found strong CB 1 immunostaining associated with a dense network of fibres in the stratum pyramidale also surrounding immunonegative pyramidal neuronal bodies. Specific expression of CB 1 receptor has been already reported in the hippocampus of healthy dogs. Dot-like structures with CB 1 immunoreactivity were found lining the external surface of neuronal cell bodies in the 4 regions of the CA where the cytoplasm of neurons did not have CB 1 immunoreactivity [26]. Our findings agree with this previous study. Interestingly, we found a progressive decrease in the immunoreactivity from CA1 to CA4. Campora and others described a similar pattern in the canine CA [26]. In rats [58] and humans [37] most of CB 1 immunoreactive neurons in the hippocampus are GABAergic, and are involved in mechanism by which cannabinoids impair memory and associational processes [37]. Expression of CB 1 is markedly increased specially in the stratum pyramidale (CA1-CA3) and molecular layer of the dentate gyrus in different mouse models of epilepsy [57,59,60] and human patients with epilepsy [60]. This CB 1 upregulation may be a compensatory mechanism of excitatory neurons to strengthen the negative feedback loop of the endocannabinoid system and to down-regulate neurotransmitter release [57].
The subcortical nuclei with the highest level of CB 1 receptor expression are the basal ganglia, including the globus pallidus and substantia nigra pars reticulate in rats [19,21], rhesus monkeys [3] and humans [3,17]; and account for the complex effects of cannabinoids on motor behavior [61][62][63]. In dogs, however, the expression is lower compared to these species [3]. We found intense CB 1 immunoreactive fibres in the globus pallidus and in the fibres surrounding unstained neuronal bodies in the substantia nigra; this immunoreactivity was stronger towards the pars reticulata.
Our results show moderate CB 1 immunoreactivity in fibres and soma of neurons of the periaqueductal grey (PAG). CB 1 immunoreactivity has been found on cell bodies [64] as well as axons and dendrites of the PAG in healthy rats [19,64]. Furthermore, expression of CB 1 immunoreactive neurons is increased after immobilization stress [64]. Stress activates neural systems that inhibit pain sensation depending on neural pathways projecting from cortical neurons to the PAG and descending to the brainstem and spinal cord suppressing nociception [64]. Moreover, antinociceptive effects of cannabinoids in the PAG have been proven in rats [65].
Strong homogeneous staining has been described in the molecular layer of the cerebellar cortex and surrounding the immunonegative Purkinje cells bodies in rats [19], mice [20] and macaques [22], particularly in their basal areas, corresponding to initial axonal segments [20] or basket cell processes [21,22]. We found identical patterns of immunoreactivity in this particular region. Higher receptor-binding levels have been found in the canine cerebellum compared to humans [3], which might induce less motor depression in humans under effects of THC [66,67]. Interestingly, the use of THC and cannabinoid analogs in experimental studies showed ataxia and even prostration at higher dosages in dogs [67,68]. High concentrations of cannabinoid expression in the basal ganglia and cerebellum are consistent with their involvement in the initiation and coordination of movement [3,69] and explain this behavioural changes in dogs at high doses of THC and cannabinoid analogs.
At the level of the medulla oblongata a strong dot-like CB 1 immunoreactivity was observed only surrounding the neuronal bodies of the cochlear nucleus and the nucleus of the spinal tract of the trigeminus. Such beaded fibers were described in the spinal trigeminal tract and spinal trigeminal nucleus in rats [19].
Within the grey matter of all spinal cord sections, strong CB 1 immunoreactive fibres were observed in the dorsal horn, intermediate region and ventral horn. CB 1 dot-like immunostaining was present surrounding the body of groups of neurons. Few neurons showed slight cytoplasmic immunoreactivity within the dorsal and ventral horns. In humans, strong CB 1 immunoreactivity has been found in dorsal horn, lamina X and ventral horn [51]. Immunoreactive cell bodies have been found in the lamina X in rats [23,70] and according to one study through all grey matter [71].
DRG larger cells seamed devoid of immunoreactivity in rats [72]. Interestingly, cultured rat DRG [73] and in-situ hybridization studies [74] showed that most CB 1 immunoreactive neurons are small cells. According to our results, large neurons showed slight cytoplasmic expression and small dark neurons expressed high CB 1 immunoreactivity. Indeed, satellite cells strongly expressed CB 1 . The presence of CB 1 receptors in the DRG and the dorsal horn may explain some analgesic effects of cannabinoids [71]. Cannabinoids have been widely reported to produce antinociception in several animal models [75][76][77][78] and the effects are mediated through CB 1 receptors through peripheral [78], spinal [79] and supraspinal [76] mechanisms.
Moderate to strong cytoplasmic CB 1 immunoreactivity was observed in astrocytes both in the white and grey matter along the cerebrum, cerebellum and spinal cord in all dogs. However, co-expression of CB 1 with the astrocytic marker GFAP was observed only in about~20% of astrocytes in the cerebral white matter. CB 1 expression has been evidenced in-situ in the cytoplasm and processes of astrocytes in rats [43,72,80]. In vitro studies suggest that cannabinoids may influence astrocyte function [81]. Bidirectional neuron-astrocyte communication has been demonstrated [82][83][84]. Furthermore, astrocytes have been shown to be activated by endocannabinoids released by neurons [42]. Forming a "tripartite synapse" where an exchange of information with the synaptic neuronal elements occurs, responding to synaptic activity and thus regulating synaptic transmission [84]. Interestingly, increased CB 1 expression has been shown in astrocytes of the hippocampus of epileptic rats [45]. Therefore, astrocytes should be taken into account when assessing the overall effects of cannabinoids [44] particularly in epilepsy [45].
No co-expression was found among CB 1 and the mature oligodendrocytic marker (CNPase) in the cerebral white matter. Interestingly, previous research has shown CB 1 expression in vitro and in vivo in oligodendrocytes of healthy rat brain and spinal cord [50,[85][86][87]. CB 1 expression was evidenced in oligodendrocyte progenitor cells (OPCs) in cultures [50,85] and in oligodendrocytes of postnatal and adult cerebral [86] and spinal cord white matter in rats [87]. In humans however, CB 1 receptor expression has been found in OPCs and adult oligodendrocytes within multiple sclerosis (MS) plaques, but not in healthy brain tissue [88].
Strikingly, CB 1 immunostaining within the sciatic nerve revealed strong expression in randomly distributed Schwann cells ensheathing axons. Moreover, co-expression of CB 1 and P0, a marker for myelinating Schwan cells was found in 100% of the Schwan cells stained. CB 1 expression in peripheral nerve fibres have been described in rats [71]. However, CB 1 expression in Schwann cells has not been previously reported to our knowledge. Nevertheless, the presence of CB 1 in myelinating Schwann cells might have a role in myelination processes.
Regarding the age of the dogs analysed, we found a lower general CB 1 expression in the fourth-week-old dog. CB 1 expression have been described in the fetal and neonatal human brain showing that the density of receptor expression was generally similar in both [89] or even higher in neonatal human brains [17]. A lower CB 1 receptor expression has been found in aged rats in specific regions, being most prominent in the cerebellum, cerebral cortex [90], basal ganglia [91], and less prominent in the hippocampus [90]. These findings agree with our results and might be related to the decline of motor coordination and cognitive performance observed in normal ageing [92].
The small number of cases of animals with different ages does not allow us to draw definite conclusions regarding particularities among younger and older dogs. However, it is clear that the overall distribution of CB 1 receptors is preserved in dogs at examined ages. The intensity of the expression, however, is known to change during development and aging.

Conclusions
These results represent the first detailed spatial distribution of CB 1 receptors in the healthy canine CNS and PNS. Our results agree with the overall distribution of CB 1 receptors reported in other species. The high CB 1 expression found in the cerebral and cerebellar cortex, Cornu Ammonis (CA) and dentate gyrus of the hippocampus, globus pallidus and substantia nigra spinal cord and DRG might relate to the effects of cannabinoids on cognition, memory, motor functions and pain sensitivity. Moreover, expression on ependymal cells and neuroglial cells relate to the effects on neurogenesis and gliogenesis modulation. Finally, CB 1 expression on myelinating Schwann cells points out potential roles of the encocannabinoid system in myelination. Our results provide a solid basis for further studies to elucidate the physiological consequences and the implication of CB 1 receptors in pathological conditions with the future aim to manipulate them in pharmacotherapy.