Astrocytic expression of cannabinoid type 1 receptor in rat and human sclerotic hippocampi
Abstract
Cannabinoid type 1 receptor (CB1R), which is traditionally located on axon terminals, plays an important role in the pathology of epilepsy and neurodegenerative diseases by modulating synaptic transmission. Using the pilocarpine model of chronic spontaneous recurrent seizures, which mimics the main features of mesial temporal lobe epilepsy (TLE) with hippocampal sclerosis (HS) in humans, we examined the expression of CB1R in hippocampal astrocytes of epileptic rats. Furthermore, we also examined the expression of astrocytic CB1R in the resected hippocampi from patients with medically refractory mesial TLE. Using immunofluorescent double labeling, we found increased expression of astrocytic CB1R in hippocampi of epileptic rats, whereas expression of astrocytic CB1R was not detectable in hippocampi of saline treated animals. Furthermore, CB1R was also found in some astrocytes in sclerotic hippocampi in a subset of patients with intractable mesial TLE. Detection with immune electron microscopy showed that the expression of CB1R was increased in astrocytes of epileptic rats and modest levels of CB1R were also found on the astrocytic membrane of sclerotic hippocampi. These results suggest that increased expression of astrocytic CB1R in sclerotic hippocampi might be involved in the cellular basis of the effects of cannabinoids on epilepsy.
Introduction
Epilepsy is one of the most common neurological diseases affecting 50 million people worldwide. Although seizures are controlled successfully with currently available antiepileptic drugs (AEDs) in most patients, current treatments are ineffective in at least 30% of cases, demonstrating a need for more research to be done to identify better treatments [1]. Normal neuronal activity is a prerequisite for proper brain function. However, excessive neuronal activity may endanger individual neurons, and even entire neuronal networks. Epilepsy manifests as states of pathological hyperexcitability and hypersynchronous activity at the neuronal network level. The endocannabinoid system is an important neuromodulatory system involved in a plethora of physiological functions such as modulation of neurotransmitter release, regulation of pain perception, and cardiovascular, gastrointestinal and liver function [2,3]. It consists of cannabinoid receptors, their endogenous lipid ligands, and the enzymatic machinery for their synthesis, release, and degradation. Cannabinoid Type 1 Receptor (CB1R), which is distributed on both excitatory and inhibitory axon terminals [4], is one of the most highly expressed G-protein coupled receptors in the central nervous system, rivaling the abundance of benzodiazepine, dopaminergic and ionotropic glutamate receptors [5]. Endocannabinoids are released from the postsynaptic neurons in an activity-dependent manner, and bind to presynaptic CB1R. Through activation of CB1R, the endocannabinoid system exerts retrograde inhibition of neurotransmitter release in the central nervous system [6], thereby modulating neuronal excitability [7].
Published work regarding the effects of cannabinoids in epilepsy has been controversial. CB1R agonists display anti-epileptic effects in animal models of TLE [8] as well as in patients suffering from tonic-clonic seizures [9]. Furthermore, CB1Rs present on glutamatergic axon terminals were shown to exert an anticonvulsant effect in the kainic acid model of epilepsy in mice [7,10]. In line with this, CB1R antagonists are proconvulsant in already hyperexcitable networks, such as in pilocarpine treated animals [8,11]. Conversely, proconvulsive effects of CB1R agonists were described as well [12–14] and CB1R antagonists were shown to prevent the long-term increase in seizure susceptibility when applied during a particular time-window [11,15].
Some studies show that recurrent seizures may lead to an adverse reorganization of the endocannabinoid system and to the impairment of its naturally/endogenously protective effect [16,17]. CB1R expression in astrocytes is much lower than in neurons, requiring the use of immune electron microscopy for detection [18]. CB1R activation in astrocytes promotes glutamate release [19], which can enhance neuronal excitability and play an important role in sustaining epileptiform activity [20]. Although CB1R expression has been previously described in hippocampi during epileptogenesis [21–23], to date expression of astrocytic CB1R in the sclerotic hippocampus has not been reported.
The aim of the present study was to explore the expression of astrocytic CB1R in sclerotic hippocampi. Using immunofluorescent double labeling and immunohistochemical electron microscopy, we were able to detect changes in CB1R expression in astrocytes, as well as at symmetric synapses, in the sclerotic hippocampi of rats treated with pilocarpine. The expression of astrocytic CB1R in the hippocampus of TLE patients was also examined.
Materials and methods
Animals
Male Sprague-Dawley rats weighing 210-230 g were used. Animals were housed in individual cages in a controlled environment (constant temperature, 22-25°C; humidity, 50-60%; 12/12 h light/dark cycle) with access to food and water ad libitum. All animal care procedures were in accordance with the guidelines of the National Institute for Health (NIH) Guide and were approved by the Fourth Military Medical University Animal Care Committee. In addition, all efforts were made to minimize animal suffering. Animals were assigned to experimental and control groups. Experimental rats were injected intraperitoneally with lithium chloride (127 mg/kg, Sigma, St. Louis, MO, USA) 18-20 hr prior to intraperitoneal injection of pilocarpine (30 mg/kg, Sigma, St. Louis, MO, USA). Scopolamine methyl bromide (1 mg/kg, Sigma, St. Louis, MO, USA) was administered 30 min prior to pilocarpine to reduce its peripheral muscarinic effects. Rats were closely observed for the occurrence of limbic seizures and status epilepticus (SE) following pilocarpine administration. Body weight-matched control rats were treated in an identical manner, except that they were injected with saline instead of pilocarpine. The convulsive behaviors of seizures were classified as follows: class I: hypoactivity, mouth and facial automatism; class II: head nodding and mastication; class III: forelimb clonus without rearing; class IV: bilateral forelimb clonus and rearing; class V: rearing and loss of posture. Seizures were terminated with diazepam (10 mg/kg, i.p.) when rats experienced class IV seizures for 1 h. Animals that did not reach the 4-5 stage seizure scores were removed from further experiments to diminish any bias.
Electrographic seizures were detected via skull surface electrodes implanted 2 weeks after the initial episode of SE. Briefly, under 10% Chloral Hydrate anesthesia (4 ml/kg, i.p.), the rat was placed on a stereotaxic frame, and electroencephalography (EEG) electrodes were implanted according to the brain atlas (Paxinos and Watson, 1986). Screw electrodes were placed on the skull over the left and right frontal cortex (3.5 mm A, 2.5 mm L) and occipital cortex (-5.0 mm P, 2.5 mm L). Each electrode was then soldered to a connector (Sun company, Beijing, China), which was fixed on the skull with self-curing acrylic resin (ADFA, Shofu Inc., Kyoto, Japan). The animals (n = 3 in each group) were allowed 7 days to recover from the surgery before the start of EEG recording. EEG was recorded eight hours a day for 7 days.
Tissue preparation
Animals were sacrificed 3 days (n = 6) and 4 weeks (n = 9) after pilocarpine or saline administration. Rats were perfused under anesthesia, first with physiological saline (1 min) and then with a fixative containing 0.05% glutaraldehyde (TAAB, UK; only used for electron microscope) and 4% paraformaldehyde for 30 min. Six brains from each group were removed respectively 3 days and 4 weeks after treatment and cryoprotected in 30% sucrose in 0.1 M phosphate buffer at 4°C overnight. Coronal sections of each brain were cut at 12 μm thickness (CM1900, Leica, Heidelberger, Germany) and mounted on gelatin-coated slides for NeuN immunoperoxidase staining and GFAP/CB1R immunofluorescent double labeling. Three brains from 4-week pilocarpine or saline treated groups were used for electron microscopy.
Case and tissue selection
Fifteen patients, 7 men (46.7%), 8 women (53.3%), mean age 25.3 ± 6.28 years, with medically refractory TLE with unilateral hippocampal sclerosis on MRI, 8 (53.3%) with right- and 7 (46.7%) with left-hippocampal resections, were included in this study. Diagnosis was established according to previously reported clinical and electrographic characteristics (Commission on Classification and Terminology of the International League Against Epilepsy, 1989). Patients, who did not achieve seizure control despite adequate use of at least two first-line AEDs in mono- or poly-therapy, and up to toxic levels, were qualified as medically refractory. All patients had extensive presurgical evaluation including high-resolution 3.0T magnetic resonance imaging (MRI), prolonged non-invasive video-EEG recording and neuropsychological testing consistent with unilateral hippocampal sclerosis. Patients with unilateral HS associated with additional abnormalities besides brain atrophy detected by visual inspection were excluded. All patients had undergone anterior temporal lobectomy and hippocampectomy during adulthood for the treatment of refractory TLE. After surgery, all patients were diagnosed with hippocampal sclerosis by two independent pathologists. Hippocampal tissue was dissected and fixed overnight in 10% formalin, and then was cryoprotected in 30% sucrose in 0.1 M phosphate buffer overnight at 4°C. Specimens were cut at 12 μm thickness on a cryostat (CM1900, Leica, Heidelberger, Germany) and mounted on slides coated with polylysine.
Five control patients, 4 men (80.0%), 1 woman (20.0%), mean age 32.2 ± 8.61 years, with traumatic brain injury caused by car accidents, were included in this study. All patients had undergone decompression surgery within six hours post-traumatic brain injury, and then parts of hippocampus were isolated from tissues resected. All specimens were fixed, dehydrated, sliced, and stained like that of TLE patients.
This study has been approved by the ethics committee of the Fourth Military Medical University, and informed consent was given to all patients.
Immunoperoxidase histochemistry and immunofluorescent double labeling
For NeuN immunoperoxidase staining (n = 6; 4 weeks post-SE), sections were blocked for 2 h and then were incubated with mouse anti-NeuN (1:500; Millipore, Billerica, MA, USA, MAB377) 4°C overnight. Slides were then incubated with biotinylated goat anti-mouse IgG (1:300, Jackson, West Grove, PA, USA) and then incubated with avidin-biotin-peroxidase complex (ABC, 1:300, Sigma, St. Louis, MO, USA). The reaction was detected with the glucose oxidase-3, 3-diaminobenzidine method. For negative control slides, the primary antibody was replaced with PBS.
All the prepared slices including epileptic and control rat hippocampus (n = 6; 3 days and 4 weeks post-SE) and TLE patient sclerotic hippocampi (n = 15) and controls (n = 5) were used for GFAP/CB1R immunofluorescent double labeling. Slices were blocked and then incubated overnight in mixed primary antibodies including mouse anti-GFAP (1:1000, Sigma, St Louis, USA, G3893) and rabbit anti-CB1R (1:300, Abcam, Hong Kong, ab23703). Slides were then incubated with biotinylated anti-rabbit IgG (1:300, Vector, Burlingame, CA) for 4 h and finally incubated with donkey anti-mouse IgG conjugated with Alexa Fluor 594 (1:400, Invitrogen, Carlsbad, CA) and streptavidin-cyanine 2 conjugate (1:500, SouthernBiotech, Birmingham, USA) for 4 h. The nuclei were labeled with DAPI (Sigma, St Louis, USA). Slides were examined with a confocal laser scanning microscope (Fluoview 1000; Olympus, Tokyo, Japan). Digital images were captured by Fluoview application software (Olympus, Tokyo, Japan).
Immunohistochemistry for electron microscopy
The brains of epileptic rats (n = 3) and controls (n = 3) were removed and postfixed for 3 h at 4°C. Serial coronal sections of 50 μM thickness were prepared with a vibratome (VT 1000S; Leica, Heidelberger, Nussloch, Germany) for CB1R immunohistochemical labeling. The sections were placed in PBS containing 25% sucrose and 10% glycerol for 1 hr for cryoprotection. After a freeze-thaw treatment, sections were immersed in PBS containing 5% BSA and 5% NGS for 4 h to block nonspecific immunoreactivity. CB1R was detected with the immunogold-silver staining method. Sections were incubated overnight with rabbit polyclonal anti-CB1R (1:300, Abcam, Hong Kong, ab23703), and then were incubated with goat anti-rabbit IgG conjugated to 1.4 nm gold grains (1:100, Nanoprobes, Stony Brook, NY) overnight. Silver enhancement was carried out in the dark with the HQ Silver Kit (Nanoprobes) for visualization of CB1R immunoreactivity. Before and after the silver enhancement step, sections were rinsed several times with deionized water. Immunolabeled sections were fixed with 0.5% osmium tetroxide in 0.1 M phosphate buffer for 1 h, dehydrated in graded ethanol series, then in propylene oxide, and finally flat-embedded in Epon 812. After polymerization, sections were examined under a light microscope. The stratum radiatum of CA1 subregion was selected and trimmed under a stereomicroscope, and mounted onto blank resin stubs. Ultrathin sections were cut with an ultratome (Nova, LKB, Bromma, Sweden) and mounted on mesh grids (six to eight sections/grid). The ultrathin sections were counterstained with uranylacetate and lead citrate and observed under a JEM-1230 electron microscope (JEOL Ltd., Tokyo, Japan).
Immunohistochemistry and electron microscopy quantification
In order to calculate the degree of cell loss a quantitative analysis was conducted. Six corresponding coronal sections through the hippocampus of individual rats were selected for NeuN immunohistochemical and CB1R/GFAP immunofluorescence analysis. Each section was 180 μm apart from the next, ensuring that the same cell would not be counted twice. NeuN positive cells were counted along the cell layer (length = 500 μm) in a predefined field of the hippocampal CA1, CA3, DG and CA4 regions respectively. Both total number of astrocytes and number of CB1R-positive astrocytes were counted respectively in the fixed area (area = 5 × 104 μm2) in a predefined field of hippocampus. All measurements were performed by using Photoshop CS3.0 software (Adobe systems Ltd, USA).
In order to analyze changes in synaptic CB1R expression, electron microscopy quantification was carried out in CA1. Serial ultrathin sections were made from the blocks from stratum radiatum and examined under the electron microscope. Both symmetric and asymmetric synapses established by CB1R positive axon terminals were examined in the stratum radiatum of experimental and control group hippocampi. CB1R stained terminals were analyzed in every 10th section in order, following the rules of systematic random sampling, to avoid sampling of the same axon terminals. Photographs were taken of all terminals in a given area. To assess any changes occurring in the density of CB1R per terminal, the number of gold particles located at the membrane of axon terminals was counted and normalized to 1 μm of the terminal’s perimeter.
Statistical analysis
Data were expressed as mean ± standard deviation (SD) and analyzed using two sample Student’s t-test, unless other specific statistical methods were mentioned. SPSS 16.0 (SPSS Inc, IL, USA) was used for the statistical analysis and differences were considered statistically significant when the P-value was <0.05.
Results
Behavior during SE and spontaneous convulsions (SCs)
About 5 min after pilocarpine injection, all rats developed diarrhea, piloerection, and other signs of cholinergic stimulation. In the following 10 to 20 min, the rats exhibited head bobbing, scratching, chewing, and exploratory behavior. Convulsive and recurrent seizures of stage 4 or higher were observed at 30-50 min after pilocarpine injection. Four weeks after induction of SE, SCs occurred in over 90% of the rats in the pilocarpine-treated group. Animals without SCs were removed from the study.
Characteristics of sclerotic hippocampi and EEG recordings
Both neuronal loss (Figure 1B) and astrocyte proliferation (Figure 2E1, 2E2), which are the main pathological features of hippocampal sclerosis, were found in rat hippocampi 4 weeks after SE induction. Analysis of NeuN immunostaining of low-magnification images from control (Figure 1A) and epileptic (Figure 1B) rats revealed cell loss in the epileptic hippocampus. Higher magnification images show that the cells in the control group were arranged in neat rows in the stratum pyramidale of CA1 (Figure 1A1) and CA3 (Figure 1A2) and the granule cell layer of dentate gyrus (Figure 1A3) and scattered in the CA4 (Figure 1A4). However, four weeks after SE obvious neuronal loss was found in CA1 (Figure 1B1, ,1C)1C) and CA4 (Figure 1B4, ,1F),1F), while neurons in CA3 were only partially affected (Figure 1B2, ,1D)1D) and preserved in the DG (Figure 1B3, ,1E).1E). Such changes are consistent with the human classical hippocampal sclerosis [24,25]. The EEG activity of pilocarpine treated rats differed from controls. In the pilocarpine-pretreated group, obvious spike waves were observed within 22-28 days after SE, on the contrary the control group showed normal EEG activity (Figure 1G).
CB1R expression by astrocytes in rat sclerotic hippocampi
To determine whether CB1R is expressed by astrocytes in sclerotic hippocampi, confocal laser scanning microscopy was used. In 4-week control rats, astrocytes with slender projections (Figure 2A1) and CB1R positive axonal plexus (Figure 2B1) were distributed in the CA1 subregion, but CB1R was not found in astrocytes of the hippocampus (Figure 2D2). Four weeks after SE induction, accrementition of astrocytes (Figure 2E1), disordered nucleus (Figure 2G1), and increased CB1R fluorescence intensity (Figure 2F1) were observed in sclerotic hippocampi. Our results also indicate the presence of CB1R in many GFAP-positive cells in sclerotic hippocampi (Figure 2H2). To determine the time point when astrocytic CB1R expression becomes upregulated, we observed hippocampal astrocytes 3 days after SE induction. We found that astrocytes began to proliferate at this time point (Figure 3A1) and CB1R expression was shown in a few astrocytes (Figure 3D3), while CB1R was not found in astrocytes of the control group. Both the density of astrocytes (Figure 4A) and the percentage of CB1R-positive astrocytes (Figure 4B) 3 days after SE induction were less than in hippocampi examined 4 weeks after SE. No significant differences either in behavioral or morphometric parameters were found between 3-day and 4-week control rats, so to simplify images, only the 4-week control images are shown.
CB1R expression by astrocytes in human sclerotic hippocampi
CB1R was found in some hippocampal astrocytes of 9 cases of TLE patients (Figure 5L1) and the percentage of CB1R positive astrocytes was 35.50% ± 12.36% for the 9 patients, while CB1R was not found in astrocytes of the other 6 cases of TLE patients (Figure 5H1) and 5 cases of controls (Figure 5D1). CB1R found at synapses around astrocytes, in both control and TLE patient hippocampus (Figure 5B, ,5F,5F, ,5J),5J), was distributed in a reticular pattern that was similar to that in rats, and their fluorescence intensity on synapses was much higher than that in astrocytes (Figure 5D1, 5H1, 5L1). Because the resected hippocampal tissue from patients is incomplete, hippocampal subfields could not be distinguished.
Ultrastructural analysis of CB1R distribution in rat sclerotic hippocampi
We then examined the precise subcellular localization of CB1R in the stratum radiatum of the CA1 subregion by using silver enhanced immunogold electron microscopy. CB1R was not found in astrocytes in control tissue (Figure 6A), but sclerotic hippocampi exhibited numerous hypertrophic astrocytes that were positive for CB1R staining (Figure 6B), and some silver grains were also found in astrocytic membranes (Figure 6C). In both control and epileptic tissue, most of the CB1R labeling was seen along the cell membrane; in addition to this, more labeling was detected in inhibitory axon terminals that formed symmetrical synapses (Figure 6D, ,6F)6F) than that detected around excitatory axon terminals forming asymmetrical synapses (Figure 6E, ,6G).6G). Sclerotic hippocampi were marked by a substantial increase in the silver grains in the axon terminals of symmetric synapses (Figure 6B, ,6F,6F, ,6H).6H). On the contrary, no difference was found in asymmetric synapses between control and sclerotic hippocampi (Figure 6G, ,6I6I).
Discussion
In the present study, we report the expression profile of astrocytic CB1R in sclerotic hippocampi. The expression of astrocytic CB1R was increased in hippocampi of TLE rat models 3 days and 4 weeks after SE. Furthermore, CB1R was expressed by astrocytes in the hippocampus of patients with refractory temporal lobe epilepsy, associated with hippocampal sclerosis. CB1R has previously been found mainly in GABAergic interneurons [26], especially in cholecystokinin-containing GABAergic interneurons of the hippocampus [27], and some CB1R expression has been observed in glutamatergic neurons [28,29]. It was reported that no CB1R was expressed in hippocampal astrocytes [30]. Recently, however, some studies in the hippocampus showed that CB1 receptors also mediate the activation of astrocytes [18,19] and play an important role in sustaining epileptiform activity [20]. Therefore identifying CB1R expression characteristics in astrocytes from sclerotic hippocampi is crucial for understanding the role of the endocannabinoid system in the pathophysiology of epilepsy.
In the present study, the hippocampal pathology generated by the administration of pilocarpine is reminiscent to that seen in human hippocampal sclerosis. Significant neuronal loss is observed in the CA1 and CA4 regions of the hippocampus 4 weeks after SE induction, consistent with the classical hippocampal sclerosis of TLE patients [25]. CB1R immunoreactivity was significant in the stratum pyramidale in normal hippocampus, which was similar to what had been previously reported [22,23,31]. We did not find CB1R expression in astrocytes in control brain, presumably because CB1R expression in astrocytes was below detection threshold. However, four weeks after SE induction, the expression of astrocytic CB1R was significantly increased compared with control; moreover, CB1R expression was found in a few reactive astrocytes as early as the third day after SE. To identify the specific localization of CB1R, we carried out an ultrastructural examination of astrocytes in hippocampal CA1 region, a region largely affected by seizures [32], and found that some CB1Rs were located on the astrocytic membrane. We were intrigued to find astrocytic CB1R expression in epileptic rat hippocampi, and were curious to see whether this is a reflection of the human pathology. Upon examining patient hippocampal tissue, expression of astrocytic CB1R was found in 9 cases of TLE patients, but it could not be detectable in another 6 cases of TLE patients and 5 cases of control. The fact that CB1R expression could not be detected in the astrocytes of all TLE patients suggested that astrocytic CB1R expression might be fluctuating in the course of TLE. It is possible that the results observed by Magloczky were due to the detection threshold of the antibody used [30]. Recently, some studies reported that the activation of astrocytic CB1R induces the release of calcium from intracellular stores, which then triggers glutamate release and results in the potentiation of synaptic transmission in neighboring neurons [19,33]. These effects could be enhanced with the increasing expression of CB1R in astrocytes, which may be helpful to further understand the role of CB1R in epilepsy.
Because astrocytes and synapses are closely interconnected in the structure of tripartite synapse, astrocytes are able to exert an essential function in controlling synaptic transmission in the hippocampus [34,35]; similarly, neuronal discharge may also induce a comparable activation of astrocytic networks [36]. In the pilocarpine model of TLE, in addition to the increased expression of astrocytic CB1R in hippocampal sclerotic tissue, presence of CB1R-positive axons, which covered the entire hippocampus, was also significantly increased. Moreover this increase was mainly on the inhibitory presynaptic membrane. GABAergic transmission, while physiologically inhibitory in the adult brain, can become depolarizing and even excitatory in the human epileptic temporal lobe [37]. Thus increasing expression of CB1R on the inhibitory presynaptic membrane may be an endogenous mechanism to inhibit GABA release and form a negative feedback loop to terminate seizure discharge. Actually, when GABA currents exhibit a depolarizing shift (under pathophysiological conditions), benzodiazepines become ineffective and instead contribute to the overall excitation of the neuron-this is thought to be one of the causes of benzodiazepine resistance in a subset of patients [38]. However, in fact benzodiazepines, which strengthen the GABAergic system, are still effective in many epileptic patients with hippocampal sclerosis, and thus the increase of CB1R on the inhibitory presynaptic membrane, which inhibits the GABAergic system, might promote epileptiform discharge in these patients. So whether the increased expression of CB1R on the inhibitory presynaptic membrane is proconvulsant or anticonvulsant becomes uncertain under conditions of hippocampal sclerosis.
There is still no unified consensus as to whether cannabinoids suppress or facilitate epilepsy, which might be related to the differences of CB1R distribution under physiological and pathological conditions. On the one hand, CB1R on the inhibitory presynaptic membrane can inhibit GABA release, thus increasing the excitability of the neuronal network; on the other hand, CB1R on the excitatory presynaptic membrane can inhibit glutamate release, thus reducing the excitability of the neuronal network [39]. In the present study, CB1R expression was increased in astrocytes, which are very abundant in the neuronal network, in hippocampal sclerotic tissue. Since CB1R activation in astrocytes promotes glutamate release [19], its activation could enhance synaptic transmission [33] and play an important role in sustaining the epileptiform activity [20]. Moreover, a recent report showed the existence of functional CB1R signaling in human astrocytes [40], suggesting that an increased expression of astrocytic CB1R in hippocampal sclerotic tissue might be involved in the mechanism underlying the effects of cannabinoids on epilepsy.
Acknowledgements
We are very grateful to Shuli Liang, Professor of Department of Neurosurgery of First Affiliated Hospital of Chinese People’s Liberation Army (PLA) General Hospital for providing the human hippocampal sclerotic tissue. This work was supported by grants from the Natural Science Foundation of China (Nos. 81271432).
References
