Cerebral Cortex Advance Access originally published online on February 8, 2006
Cerebral Cortex 2007 17(1):175-191; doi:10.1093/cercor/bhj136
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Immunocytochemical Distribution of the Cannabinoid CB1 Receptor in the Primate Neocortex: A Regional and Laminar Analysis
1 Department of Neuroscience, University of Pittsburgh, Pittsburgh, PA 15213, USA, 2 Center for the Neural Basis of Cognition, University of Pittsburgh, Pittsburgh, PA 15213, USA, 3 Department of Psychiatry, University of Pittsburgh, Pittsburgh, PA 15213, USA
Correspondence: David A. Lewis, Department of Psychiatry, University of Pittsburgh, 3811 O'Hara Street, W1651 BST, Pittsburgh, PA 15213, USA. Email: lewisda{at}upmc.edu.
 |
Abstract |
|---|
 Top Abstract IntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Delta-9-tetrahydrocannabinol (
9-THC) has profound effects on higher cognitive functions, and exposure to 9-THC has been associated with the appearance or exacerbation of the clinical features of schizophrenia. These actions appear to be mediated via the CB1 receptor, the principal cannabinoid receptor expressed in the brain. However, the distribution of the CB1 receptor in neocortical regions of the primate brain that mediate cognitive functions is not known. We therefore investigated the immunocytochemical localization of the CB1 receptor in the brains of macaque monkeys and humans using antibodies that specifically recognize the N- or C-terminus of the CB1 receptor. In monkeys, intense CB1 immunoreactivity was observed primarily in axons and boutons. Across neocortical regions of the monkey brain, CB1-immunoreactive (IR) axons exhibited considerable heterogeneity in density and laminar distribution. Neocortical association regions, such as the prefrontal and cingulate cortices, demonstrated a higher density, and exhibited a unique laminar pattern of CB1-IR axons, compared with primary sensory and motor cortices. Similar regional and laminar distributions of CB1-IR axons were also present in the human neocortex. CB1-IR axons had more prominent varicosities in human tissue, but this difference appeared to represent a postmortem effect as similar morphological features increased in unperfused monkey tissue as a function of postmortem interval. In electron microscopy studies of perfused monkey prefrontal cortex, CB1 immunoreactivity was predominantly found in axon terminals that exclusively formed symmetric synapses. The high density, distinctive laminar distribution, and localization to inhibitory terminals of CB1 receptors in primate higher-order association regions suggests that the CB1 receptor may play a critical role in the circuitry that subserves cognitive functions such as those that are disturbed in schizophrenia.
Key Words: cholecystokinin • GABA • human • interneurons • monkey • parvalbumin
|
Introduction |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Marijuana, a preparation of the hemp plant Cannabis sativa, is one of the oldest known recreational drugs, and today it is one of the most widely abused illicit drugs (Watson and others 2000
). Delta-9-tetrahydrocannabinol (9-THC), the chief psychoactive cannabinoid in cannabis, has profound effects on mood and a number of cognitive functions (reviewed in Childers and Breivogel 1998; Ameri 1999). In addition, cannabis use has been associated with both an increased risk of, and greater symptom severity in, psychiatric disorders such as schizophrenia (reviewed in Smit and others 2004). Thus, understanding the molecular mechanisms by which cannabinoids elicit their effects is of substantial interest and clinical importance.
The demonstration of selective and specific binding of a radiolabeled synthetic derivative of 9-THC to brain tissue revealed the presence of a central brain cannabinoid receptor (designated CB1) (Devane and others 1988). This finding led to the cloning of the brain CB1 receptor (Matsuda and others 1990) and the identification of a peripheral CB2 receptor (Munro and others 1993), both of which are G protein–coupled receptors. The discovery of the endogenous cannabinoids anandamide (Devane and others 1992) and 2-arachidonoyl-glycerol (Mechoulam and others 1995) and the development of selective synthetic ligands that bind to the 2 receptor types (Rinaldi-Carmona and others 1994, 1998) soon followed.
Most of the physiological and behavioral effects of cannabinoids appear to be mediated by the CB1 receptor (Zimmer and others 1999), which is highly expressed and widely distributed in the brain (Herkenham and others 1991; Matsuda and others 1993; Glass and others 1997). In particular, high levels of the CB1 receptor are expressed in neocortical association areas such as the prefrontal cortex and the cingulate cortex (Herkenham and others 1991; Matsuda and others 1993; Glass and others 1997), which are known to mediate executive functions. Other regions involved in cognitive functioning, such as the hippocampus, basal ganglia, and cerebellum, also express high levels of the CB1 receptor (Herkenham and others 1991; Matsuda and others 1993; Glass and others 1997; Biegon and Kerman 2001). Therefore, CB1 receptors in these regions may mediate certain deficits in cognitive functions observed following cannabinoid administration in humans and animals (Winsauer and others 1999; Schneider and Koch 2003; D'Souza and others 2004).
In rodents, the CB1 receptor is almost exclusively expressed by gamma-aminobutyric acid (GABA) interneurons in the neocortex, hippocampus, and basal nuclei of the amygdala. Indeed, in situ hybridization experiments in the mouse neocortex and hippocampus have demonstrated that 100% of neurons that express high levels of CB1 mRNA also express mRNA for the 65-kDa isoform of glutamic acid decarboxylase 65, a synthesizing enzyme of GABA (Marsicano and Lutz 1999). Furthermore, dual-label in situ hybridization and dual-label electron microscopy experiments in the rodent neocortex, hippocampus, and amygdala revealed that the CB1 receptor is preferentially expressed by, and predominantly localized in, the terminals of the subtype of GABA basket interneurons that contain the neuropeptide cholecystokinin (CCK). In contrast, CB1 is not found in GABA neurons containing the calcium-binding protein parvalbumin (PV) (Katona and others 1999, 2001; Marsicano and Lutz 1999; Hajos and others 2000; Bodor and others 2005). Consistent with this anatomical localization of the CB1 receptor, electrophysiological studies have demonstrated that CB1 agonists affect GABA neurotransmission. For instance, in in vitro slices of rodent neocortex, hippocampus, and amygdala, CB1 agonists inhibit the release of GABA from neurons and reduce the amplitude of inhibitory postsynaptic currents (Katona and others 1999, 2001; Hajos and others 2000; Trettel and others 2004; Bodor and others 2005). Furthermore, systemic administration of CB1 agonists decreases GABA levels in the rat neocortex in vivo as measured by microdialysis (Pistis and others 2002).
Together, these observations suggest that the endogenous cannabinoid system plays an integral role in modulating GABA synaptic neurotransmission. However, understanding the involvement of CB1 receptor–mediated signaling in cognitive functions and in the impairment of these abilities in schizophrenia requires knowledge of the anatomical localization of this receptor in the primate neocortex, especially in areas such as the prefrontal cortex where GABA neurotransmission is critical for cognitive functions (Sawaguchi and others 1988; Rao and others 2000). However, in the 2 studies that previously investigated the immunocytochemical distribution of the CB1 receptor in the monkey neocortex, the antibody that was used produced a pattern of labeling not entirely consistent with electrophysiological and pharmacological affects of CB1 agonists or the anatomical localization of the CB1 receptor in rodents (Lu and others 1999; Ong and Mackie 1999). Furthermore, although the immunocytochemical localization of CB1 in CCK interneurons in the human hippocampus (Katona and others 2000) and the autoradiographic distribution of CB1 receptor–binding sites across certain human neocortical regions (Glass and others 1997; Biegon and Kerman 2001) appear to be conserved from the rodent (Herkenham and others 1991; Katona and others 1999), the immunocytochemical distribution of the CB1 receptor in the human brain has not been examined outside of the hippocampus. Consequently, we used immunocytochemical techniques and antibodies that specifically recognize the N- or C-terminus of the CB1 receptor to examine the regional and laminar distribution of the CB1 receptor in the neocortex of macaque monkeys and humans. In addition, we used immunoelectron microscopy in order to determine the cellular localization and cell types that express the CB1 receptor in the monkey dorsolateral prefrontal cortex.
|
Materials and Methods |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Light Microscopy
Perfused Monkey Specimens
Eight adult male long-tailed macaque monkeys (Macaca fascicularis) were utilized for light microscopy. Housing and experimental procedures were conducted in accordance with United States Department of Agriculture and National Institutes of Health guidelines and with approval of the University of Pittsburgh's Institutional Animal Care and Use Committee. Monkeys were deeply anesthetized with ketamine hydrochloride (25 mg/kg) and sodium pentobarbital (30 mg/kg), intubated, mechanically ventilated with 28% O2/air, and perfused transcardially with ice-cold 1% paraformaldehyde in 0.1 M phosphate buffer (pH 7.4) followed by 4% paraformaldehyde in phosphate buffer, as previously described (Oeth and Lewis 1993). Brains were immediately removed, blocked into 5- to 6-mm-thick coronal or sagittal blocks, and postfixed for 6 h in phosphate-buffered 4% paraformaldehyde at 4 °C. Tissue blocks were subsequently immersed in cold sucrose solutions of increasing concentrations (12%, 16%, and 18%) and then stored at –30 °C in a cryoprotectant solution containing glycerine and ethylene glycol in dilute phosphate buffer. We have previously shown that this storage procedure does not affect immunoreactivity for a number of antigens (see Cruz and others 2003). Tissue blocks from either the left or the right hemisphere were sectioned coronally or sagittally at 40 µm on a cryostat, and every 10th section was stained for Nissl substance with thionin.
Human Autopsy Subjects
Brain specimens from 6 (4 male and 2 female) human subjects (20–38 years of age; postmortem interval [PMI] 4.5–8.5 h) were obtained from autopsies conducted at the Allegheny County Coroner's Office, Pittsburgh, PA, following informed consent for brain donation from the next of kin. None of the subjects had a history of psychiatric or neurologic disorders as determined by information obtained from clinical records and a structured interview conducted with a surviving relative. Neuropathological examinations revealed no abnormalities in any subject. All procedures were performed with the approval of the University of Pittsburgh's Institutional Review Board for Biomedical Research.
Following retrieval of brain specimens, the left hemisphere was cut into 1.0-cm coronal blocks and immersed in phosphate-buffered 4% paraformaldehyde for 48 h at 4 °C. Tissue blocks were subsequently immersed in graded sucrose solutions and stored as described earlier. Tissue blocks containing the superior frontal gyrus (dorsolateral prefrontal cortex) or the calcarine sulcus (primary visual cortex) were sectioned coronally as described earlier.
Unperfused Monkey Specimens
One adult male long-tailed macaque monkey (M. fascicularis) was used to investigate the effect of PMI on CB1 immunoreactivity. The monkey was deeply anesthetized as described earlier and the brain was immediately removed, cut into 3- to 5-mm-blocks, and immersed in ice-cold 0.01 M phosphate-buffered saline (PBS; pH 7.4). Following PMIs of 2, 12, 24, or 48 h, adjacent tissue blocks were transferred to cold phosphate-buffered 4% paraformaldehyde for 48 h. Tissue blocks were then immersed in sucrose solutions and stored as described earlier. Blocks containing the dorsolateral prefrontal cortex for each PMI were sectioned on a cryostat as described earlier.
Immunocytochemistry
For perfused monkeys, free-floating tissue sections were thoroughly washed in 0.01 M PBS and then treated for 30 min with a blocking solution containing 0.3% Triton X-100 and 4.5% normal donkey sera (NDS) and normal human sera (NHuS) at room temperature in PBS. Tissue sections were subsequently incubated for 48 h at 4 °C in PBS containing 0.3% Triton X-100, 3% NDS and NHuS, 0.05% bovine serum albumin (BSA), and an affinity-purified polyclonal rabbit anti-CB1 antibody raised against either the N-terminus of the human CB1 receptor (residues 1–99; Sigma, St. Louis, MO; diluted 1:1000), the entire C-terminus of the rat CB1 receptor (anti-CB1–CT) (residues 401–473; diluted 1:5000), or the last 15 amino acid residues of the rat CB1 receptor (anti-CB1–L15) (diluted 1:5000). Both the C-terminus antibodies were kindly provided by Dr Ken Mackie (University of Washington, Seattle, WA). Sections were then incubated for 1 h in a biotinylated anti-rabbit IgG secondary antibody made in donkey (diluted 1:200; Jackson ImmunoResearch, West Grove, PA) followed by processing with the avidin–biotin–peroxidase method (Hsu and others 1981) using the Vectastain Avidin-Biotin Elite Kit (Vector Laboratories, Burlingame, CA). The immunoperoxidase reaction was visualized using 3,3'-diaminobenzidine (DAB; Sigma), mounted on gel-coated slides, and air-dried. The DAB reaction product was stabilized by serial immersions in osmium tetroxide (0.005%) and thiocarbohydrazide (0.5%) as previously described (Lewis and others 1986).
Human and unperfused monkey tissue sections were processed with the anti-CB1–L15 in the same manner except that they were initially pretreated with 1% hydrogen peroxide for 15 min to remove endogenous peroxidase activity.
Analysis of Neocortical Laminar Patterns
The laminar patterns of CB1 immunoreactivity observed across different neocortical regions were assessed using a Microcomputer Imaging Device (MCID) system (Imaging Research Inc, London, ON, Canada). Slide-mounted sections containing selected neocortical regions were transilluminated on a light box, and images were captured by a video camera under precisely controlled conditions and digitized. Within each cytoarchitectonic region, a rectangle 
1-mm wide extending from the pial surface to white matter was drawn and the optical density within each rectangle was measured. Optical density values were divided into 100 bins using Matlab software (The MathWorks, Natick, MA) and normalized within each cytoarchitectonic region by dividing the value of each bin by the value of the maximum bin. Laminar boundaries were determined by calculating the percent depth of each layer in adjacent Nissl-stained sections using the Neurolucida program (MicroBrightfield, Inc., Colchester, VT). Values from layer 1 were not reported due to edge effects that produced high optical density values even though layer 1 did not contain CB1-immunoreactive (IR) axons in most regions. Optical density values for each selected region were obtained from a single traverse from one animal each. A systematic analysis across multiple animals was not performed due to a lack of available tissue containing each region studied from every animal. However, the overall density and laminar patterns were confirmed in at least 2 animals for each region qualitatively.
Electron Microscopy
Animals and Tissue Preparation
For electron microscopy studies, 2 additional adult male long-tailed macaque monkeys (M. fascicularis) were deeply anesthetized as described earlier and transcardially perfused with room temperature 1% paraformaldehyde and 0.05% glutaraldehyde in 0.1 M phosphate buffer followed by 4% paraformaldehyde and 0.2% glutaraldehyde in the same buffer as previously reported (Melchitzky and others 2005). Brains were immediately removed, blocked coronally into 5-mm-thick blocks, and postfixed for 2 h in phosphate-buffered 4% paraformaldehyde at 4 °C. Tissue blocks were subsequently washed in 0.1 M phosphate buffer, and blocks containing prefrontal cortex area 46 were sectioned coronally at 50 µm on a vibrating microtome.
Immunocytochemistry
Free-floating tissue sections were initially treated with 1% sodium borohydride in 0.1 M phosphate buffer for 30 min followed by several washes in 0.1 M phosphate buffer to improve antigenicity and reduce nonspecific immunoreactivity (Sesack and others 1998). Sections were incubated for 30 min in a blocking solution containing 0.2% BSA, 0.04% Triton X-100, 3% NDS, and 3% NHuS in 0.01 M PBS (pH 7.4) to further reduce nonspecific labeling. Sections were subsequently incubated overnight in blocking solution containing the anti-CB1–CT primary antibody (diluted 1:5000). On the following day, sections were rinsed in PBS and incubated for 1 h in blocking solution containing a biotinylated anti-rabbit IgG secondary antibody made in donkey (diluted 1:200; Jackson ImmunoResearch). Following rinses in PBS, sections were processed with the avidin–biotin–peroxidase method and visualized with DAB as described earlier. Tissue sections were then postfixed in 2% osmium tetroxide for 1 h in phosphate buffer, dehydrated in ascending alcohol solutions, and embedded in Epon 812 (EM bed 812; Electron Microscopy Sciences, Fort Washington, PA) as previously described (Sesack and others 1995).
Sampling Regions and Procedures
For each animal, trapezoid blocks were cut from layer 4 of area 46 (Fig. 1A) and sectioned on a Reichert ultramicrotome (Nussloch, Germany) at 80 nm. Two to four ultrathin sections were serially collected on 200-mesh copper grids and counterstained with uranyl acetate and lead citrate. For each trapezoid block 2–3 grids, separated by at least 10 grids, were examined. Grids were examined on an FEI Morgagni transmission microscope (Hillsboro, OR) and micrographs containing CB1-labeled structures were captured as digital images using an AMT XP-60 digital camera (Danvers, MA) and stored for later analysis.
|
Identification of Neuronal and Synaptic Elements
Neuronal elements encountered in electron micrographs were identified according to previous descriptions (Peters and others 1991). Briefly, perikarya were identified by the presence of a nucleus. Dendritic shafts were identified by the presence of synaptic inputs, mitochondria, microtubules, and neurofilaments, whereas dendritic spines were characterized by the absence of both organelles and microtubules. Asymmetric synapses (Gray's type I) were identified by the widening and parallel spacing of apposed plasmalemmal surfaces, the presence of a prominent postsynaptic density, and round small synaptic vesicles. Symmetric synapses (Gray's type II) were identified by the presence of intercleft filaments, a thin postsynaptic density, and pleomorphic small synaptic vesicles.
Antibody Specificity
The specificity of the antibodies raised against the entire C-terminus of the rat CB1 receptor has been verified by several lines of evidence. First, when tested in tissue from CB1 knockout mice, no immunolabeling was observed (Hajos and others 2000; Katona and others 2001). Second, western blotting of rat brain homogenates produced bands at the predicted molecular weights based on the amino acid sequence of the CB1 receptor. Third, the CB1 antibody labeled ATt20 cells transfected with the CB1 receptor but did not label untransfected cells (Hajos and others 2000; Katona and others 2001). In addition, we preadsorbed the anti-CB1–CT and anti-CB1–L15 antibodies with 1 µg/mL of their respective cognate peptides. We also preadsorbed the anti-CB1–L15 antibody with the anti-CB1–CT fusion peptide and vice versa (see Results).
Nomenclature
Neocortical regions were identified on adjacent Nissl-stained sections based on previously published cytoarchitectonic and connectional analyses in the macaque monkey and the atlas of Paxinos and others (2000) (for regional abbreviations see Table 1). Frontal lobe cytoarchitectonic delineations used the criteria and terminology of Barbas and Pandya (1989) and Carmichael and Price (1994). Regions of the cingulate cortex were identified following the divisions of Vogt and others (1987) and Morecraft and others (2004). In the temporal lobe, the terminology of Amaral and others (1987) was used for the subregions of the entorhinal cortex, and regions of the superior temporal sulcus (TE and TEO) were defined according to Seltzer and Pandya (1989). The auditory regions in the superior temporal gyrus were identified according to the cytoarchitectonic criteria of Galaburda and Pandya (1983), using the updated parcellation and nomenclature of Hackett and others (1998). In this nomenclature, AI and R compose the core auditory cortex; association belt regions are designated RM, AL, and ML; and parabelt regions are designated RP and CP. Divisions of the insular cortex were classified according to Mesulam and Mufson (1982). In the parietal postcentral gyrus, primary somatosensory areas 3, 1, and 2 were identified based on the studies of Burton and others (1995) and Cipolloni and Pandya (1999). Regions of the posterior parietal cortex were defined in accordance with Cavada and Goldman-Rakic (1989) in which area 5 corresponds to area PE and areas 7m, 7a, and 7b correspond to areas PGm, PG, and PF of Pandya and Seltzer (1982), respectively. Regions of the intraparietal sulcus were designated as physiologically defined areas MIP, VIP, and LIP (reviewed in Colby and Goldberg 1999) but are architectonically analogous to area PEa and POa of Pandya and Seltzer (1982) and area 7ip of Cavada and Goldman-Rakic (1989). Visual regions of the parietooccipital cortex (MT, MST, FST) were defined based on the studies of Boussaoud and others (1990) and Lewis and Van Essen (2000). The laminar boundaries of area 17 were delineated based on the divisions of Fitzpatrick and others (1985). The nuclear divisions of the amygdala were defined according to Amaral and Bassett (1989) and Pitkanen and Amaral (1998). Regions of the hippocampal formation and layers of the dentate gyrus (DG) and CA regions followed the terminology of Alonso and Amaral (1995) and Jongen-Relo and others (1999). Cerebellar lobules and folia were defined according to the atlas of Madigan and Carpenter (1971). Finally, laminar boundaries in human neocortical regions were identified using published cytoarchitectonic criteria (Table 1) (Braak 1976; Rajkowska and Goldman-Rakic 1995).
|
Photography
Brightfield photomicrographs were obtained with a Zeiss Axiocam camera. Digital electron micrographs and brightfield photomontages were assembled, and the brightness and contrast were adjusted in Adobe Photoshop.
|
Results |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Specificity of the CB1 Antibodies
The specificity of the C-terminus CB1 antibodies was tested by omission of the primary antibodies and by preadsorption of the anti-CB1–CT and anti-CB1–L15 antibodies with their respective fusion proteins. All specific immunoreactivity was eliminated in monkey tissue under these conditions. When the anti-CB1–L15 antibody was preadsorbed with the anti-CB1–CT fusion peptide, all specific immunoreactivity was eliminated. However, when the anti-CB1–CT antibody was preadsorbed with the anti-CB1–L15 fusion peptide, some specific labeling of axons remained, indicating that the overlapping but shorter anti-CB1–L15 fusion peptide contained some, but not all, epitopes recognized by the anti-CB1–CT antibody. Furthermore, although raised against different amino acid sequences of the CB1 receptor, all 3 antibodies produced identical patterns of immunoreactivity in the neocortex (Fig. 1). Specifically, each of the CB1 antibodies labeled numerous thin, highly varicose axons with a distinctive laminar pattern of distribution (Fig. 1A–C). In addition, dense perisomatic arrays of labeled processes around unlabeled cell bodies (Fig. 1D–F) and a small number of intensely immunoreactive cell bodies restricted to neocortical layers 2 and superficial 3 were observed with all 3 antibodies (Fig. 1A–C). Consistent with other reports (Hajos and others 2000; Katona and others 2001), the C-terminus antibodies (Fig. 1A,B,D,E) produced more intense immunoreactivity with a better signal-to-noise ratio than the N-terminus antibody (Fig. 1C,F). We therefore used the C-terminus antibodies in our analyses.
Distribution of CB1 Immunoreactivity in Monkey Neocortex
Regional Densities
The overall distribution of CB1-IR axons varied substantially across different neocortical regions in each of the animals examined (Table 2). Specifically, the highest densities of CB1-IR axons in the neocortex were present in higher-order association regions, such as the prefrontal cortex, whereas the primary visual cortex had the lowest density of CB1-IR axons. Within the frontal lobe, dorsolateral prefrontal area 46, followed by dorsal area 9 and rostral area 10, contained the highest densities of CB1-IR axons. From these areas the density of CB1-IR axons progressively declined both caudally (across areas 8, 6, and 4) and ventrally (across areas 11 and 13) (Fig. 2). The primary motor cortex (area 4) contained the lowest overall density of CB1-IR axons in the frontal lobe (Table 2; Fig. 2). Regions of the cingulate cortex also expressed a high density of CB1-IR axons (Table 2; Fig. 3). For instance, the densities CB1-IR axons in areas 24b–d, 23b–d, and 31 were similar to that observed in prefrontal areas 9 and 10. However, areas 24a, 23a, 29, and 30 exhibited a lower density of CB1 immunoreactivity (Table 2; Fig. 3).
|
|
|
In the medial temporal cortex, all fields of the entorhinal cortex (EI and EC illustrated) and adjacent area 36 exhibited a high density of CB1-IR axons (Table 2; Figs. 3 and 8A). Visual area TEO in the inferior temporal cortex contained an intermediate density of CB1-IR axons, whereas rostral region TE (Fig. 3) contained a moderate to high CB1-IR axon density. In the superior temporal gyrus, auditory association areas contained a high CB1-IR axon density, with parabelt regions RP and CP exhibiting slightly higher CB1-IR axon densities than the adjacent belt regions AL and ML (Table 2; Figs. 3 and 7A). Primary auditory regions R and AI exhibited only an intermediate density of CB1-IR axons (Table 2; Figs. 3 and 7A). The dysgrangular region of the insula (Idg) contained a lower density of CB1-IR axons than the prefrontal cortex, and labeled axon density was further reduced in the granular region of the insula (Ig) (Table 2; Fig. 3).
|
|
In the postcentral gyrus of the parietal cortex, primary somatosensory areas 3, 1, and 2 had a relatively low density of CB1-IR axons, similar to that observed in primary motor area 4 (Table 2; Fig. 3). Area SII on the dorsal bank of the lateral fissure showed a slightly higher density of CB1-IR axons. Areas 7 and 5 of the posterior parietal cortex contained a high density of CB1-IR axons similar to that observed in the prefrontal cortex, whereas areas MIP, VIP, and LIP in the intraparietal sulcus exhibited only an intermediate expression of CB1-IR axons (Table 2).
The lowest overall density of CB1-IR axons in the neocortex was present in the primary visual cortex (area 17), with area 18 demonstrating a slightly higher density of CB1-IR axons (Table 2). Higher-order visual areas MT, MST, and FST demonstrated intermediate densities of CB1-IR axons (Table 2).
Laminar Patterns
The laminar distribution of CB1-IR axons also varied substantially across different neocortical regions in each of the animals examined. Most association areas, such as the prefrontal cortex (Figs. 1 and 2), parietal area 7m (Fig. 4A), and cingulate areas 23b–d, 24b–d, and 31 (Fig. 3), had a similar laminar distribution of CB1-IR axons. For example, in prefrontal area 46 (Fig. 1), the density of CB1-IR axons was lowest in layer 1 and progressively increased from superficial to deep across layers 2 and 3. Layer 4 contained a very dense band of immunoreactive axons and varicosities and layer 6 contained a band of lower density. The low level of immunoreactivity in layer 5 sharply demarcated the borders with layers 4 and 6. The presence of distinct radial fibers also distinguished layers 3 and 5. The somewhat lower overall density of CB1-IR axons in some other association regions was due to a lower density of positive axons in layer 4. For instance, areas 24a (Fig. 4B) and 23a (Fig. 3) of the cingulate cortex showed a fairly homogenous distribution of CB1-positive axons across layers 2–6.
|
In contrast, in the EI and EC subdivisions of the entorhinal cortex (Fig. 4C), layers 2 and 5 contained the highest density of CB1-IR axons and varicosities, and the density was lower in layers 3 and 6. The lower density of radially oriented CB1-IR axons in layer 3 in EC and layers 3 and 4 in EI highlighted distinct borders with layers 2 and 5. Layer 2 of EI contained patches of high and low densities of CB1-IR axons that correspond to the islands of multipolar cells separated by cell-sparse zones that are characteristic of this region (Fig. 3). Layer 1 of the entorhinal cortex contained diffuse horizontally oriented CB1-IR axons unlike other neocortical association regions (Fig. 4).
The laminar pattern of CB1-IR axons was quite different in primary motor (Fig. 2) and somatosensory cortices (Fig. 5A), where the density of CB1-IR axons was greatest in layers 2–3 and 5–6 and lowest in layer 4. The primary visual cortex was unique in that the highest density of CB1-IR axons was observed in layers 5–6, layers 1–3 exhibited diffuse horizontally oriented axons, and layer 4 demonstrated a complete absence of CB1-IR axons (Fig. 5B). The laminar distribution of CB1-IR axons in higher-order visual areas such as area FST was homogeneous across layers (Fig. 5C).
|
These qualitative regional differences in laminar distribution of CB1-IR axons were confirmed by optical density measures (Fig. 6). For example, optical density across layers in area 46 (Fig. 6A), area 24a (Fig. 6B), area 2 (Fig. 6C), and area 17 (Fig. 6D) precisely matched the laminar distribution of CB1-IR axons observed in these areas qualitatively.
|
The differences in density and laminar distribution of CB1-IR axons were so striking that many boundaries between cytoarchitectonic regions were clearly delineated by CB1 immunoreactivity (Fig. 7). For instance, the boundary between primary auditory cortex AI and the adjacent auditory association belt cortex area ML in the superior temporal gyrus was clearly distinguished by the appearance of a dense band of CB1-IR axons in layer 4 of ML (Fig. 7A). In the posterior parietal cortex, the border between area 5 and MIP could be delineated by the disappearance of a dense band of CB1-IR axons in layer 4 of MIP (Fig. 7B). In the visual cortex, the area 17/18 border was clearly identified by an increase in axon density in layers 2–4 of area 18 (data not shown).
Other Brain Regions
In order to place the relative density of CB1-labeled structures in the neocortex in a broader context, we also evaluated a sampling of other brain regions.
Hippocampal Formation
The hippocampus contained overall a high density of CB1-IR axons that was similar to the density of labeled axons observed in the prefrontal cortex (Table 2). A dense meshwork of immunolabeled beaded axons was present throughout all regions of the hippocampal formation (Table 2; Fig. 8). In the CA1–CA3 fields, the pyramidal cell layer (PCL) demonstrated a high density of CB1-IR axons that surrounded unlabeled pyramidal cells (Fig. 8A,B). The stratum lacunosum-moleculare (SLM) also exhibited a high density of CB1-IR axons (Table 2; Fig. 8A,B), whereas the stratum oriens (SO) and stratum radiatum (SR) in CA1–CA3 of the hippocampus showed a slightly lower density of CB1-IR axons (Table 2; Fig. 8A,B). The stratum lucidum (SL) of the CA3 region demonstrated the lowest density of CB1-IR axons (Table 2; Fig. 8A,B). CB1-IR axons appeared to radiate through the SL in parallel with pyramidal cell dendrites, whereas in the SO, SR, and SLM regions, CB1-IR axons appeared to travel perpendicular to dendrites originating from pyramidal cells in the PCL (Fig. 8B). The subicular regions demonstrated a high density of CB1-IR axons, although the density was somewhat lower than that present in the CA regions (Fig. 8A).
The DG was distinguished by an extremely low density of CB1 labeling in the granule cell layer (GL), which appeared as a prominent immunonegative strip at low magnification (Fig. 8A). Similar to the PCL in the CA regions, granule cells were always immunonegative for CB1; however, in contrast to the PCL, CB1-IR axons appeared to radiate through the GL without surrounding unlabeled granule cells except at the border of the infragranular plexus (Fig. 8C). The outer molecular layer (OML) contained a very high density of CB1-IR axons, whereas the molecular layer (MLdg) contained a slightly lower density of CB1-IR axons (Fig. 8A,C). The polymorphic layer (PLdg) had a somewhat lower density of CB1-IR axons compared with MLdg (Fig. 8C).
Amygdala
The pattern of CB1 immunolabeling clearly followed the boundaries of several nuclei in the monkey amygdala (Figs. 3 and 9C). The most intense CB1 immunoreactivity was observed in the cortical-like basolateral complex (Table 2; Figs. 3 and 9C), where a dense meshwork of varicose axons surrounded CB1-immunonegative cell bodies (Fig. 9C). In the basolateral complex, the density of CB1-IR axons was greatest in the Bmc, Bi, ABmc, and Lvi nuclei (Fig. 3), whereas the more ventral and lateral ABpc, Bpc, Ldi, and Lv nuclei contained slightly fewer CB1-IR axons. The overall density of CB1-IR axons in the basolateral nuclei was similar to that observed in the prefrontal cortex (Table 2). In the striatal-like central (Ce) and medial (Me) nuclei of the monkey amygdala, no CB1-IR axons were evident (Table 2; Figs. 3 and 9C). In these nuclei, light labeling above that in the white matter was observed; however, this immunoreactivity was morphologically indistinct.
|
Basal Ganglia, Cerebellum, Thalamus, and Claustrum
As summarized in Table 2, CB1 immunoreactivity was present throughout the basal ganglia, but each nucleus exhibited a pattern of immunoreactivity that differed from other components of the basal ganglia and from regions of the neocortex. The caudate nucleus (Cd) and putamen (Pu) contained sparsely distributed, thin, varicose, labeled processes, as well as a diffuse immunoreactivity that was morphologically indistinct (Table 2; Figs. 3 and 9A,B). The globus pallidus (GP), which exhibited the most intense CB1 immunoreactivity of the brain regions examined (Table 2; Figs. 3 and 9B), contained a very dense meshwork of thin, smooth, CB1-IR processes that encircled large unstained fascicles, cell bodies, and wooly fibers (Fig. 9B). The substantia nigra pars reticulata (SNr) demonstrated CB1 immunoreactivity identical to that observed in the GP (data not shown). The thalamus (Th) appeared to be completely devoid of CB1-IR axons (Fig. 3). The claustrum (Cl) exhibited a high density of CB1-IR highly varicose, axonal processes (Table 2), similar to those present in the adjacent insular cortex (Fig. 9A). As in the neocortex, axons in the Cl were found to surround immunonegative cell bodies.
In the cerebellum, most lobules contained a high intensity of CB1 immunoreactivity with a distinctive laminar pattern (Fig. 10). The intensity of CB1 immunoreactivity was greatest in the molecular layer and lowest in the granular layer (Table 2; Fig. 10). Intense CB1 immunoreactivity coated the basilar portion of Purkinje cell bodies in a triangular cap-like fashion reminiscent of the pinceau synapses furnished by basket interneurons (Fig. 10B). However, Purkinje cell bodies and dendrites were always CB1 immunonegative (Fig. 10B). Although this pattern of CB1 immunoreactivity was present in most lobules, lobule X was notable for containing a very low level of CB1 immunoreactivity in the molecular layer (Fig. 11).
|
|
Comparison with Human Neocortex
The distribution of CB1-IR axons in the human prefrontal and primary visual cortices was very similar to that observed in homologous regions of the monkey neocortex with a few exceptions. As in the monkey, human prefrontal area 46 exhibited a much higher density of CB1-IR axons than primary visual cortex area 17 (Fig. 12). In the human prefrontal cortex (Fig. 12A), layer 5 contained the lowest density of CB1-IR axons, layers 2–3 contained a moderate density of CB1-IR axons, and layer 4 contained the highest density of CB1-IR axons, similar to that observed in the monkey. However, in contrast to the monkey, the density of CB1-IR axons in layer 6 did not appear to be greater than in layer 5. In human area 17 (Fig. 12B), the laminar distribution of CB1-IR axons was identical to the monkey. However, layers 1–3 contained a greater density CB1-IR axons and layer 4 appeared to contain slightly more CB1-IR radial axons, whereas layers 5–6 exhibited a lower density of CB1-IR axons than in the monkey (Fig. 12B). Finally, the morphology of CB1-IR axons in the human neocortex differed slightly from that observed in the monkey. In the human neocortex (Fig. 13C), the intervaricose segments of CB1-IR axons were less distinct and boutons appeared to be larger and swollen compared with those in the perfused monkey neocortex (Fig. 13A). These morphological differences appeared to reflect a postmortem effect. In unperfused monkey tissue with different postmortem delays before fixation, the intervaricose segments of CB1-IR axons became less distinct and boutons became more swollen with longer delays (Fig. 13B).
|
|
Electron Microscopy
Electron microscopy studies were performed in layer 4 of monkey area 46 in order to assess the cellular distribution of the CB1 receptor (Fig. 1A). CB1 immunoreactivity was found in axon terminals forming symmetric synapses (Gray's type II) (Fig. 14A). Asymmetric synapses (Gray's type I), dendrites, and dendritic spines did not appear to be immunoreactive for the CB1 receptor (Fig. 14A). When CB1 immunoreactivity was observed in cell bodies, they had the characteristic features of GABA neurons such as invaginated nuclei. The immunoperoxidase reaction product in cell somas was associated with the Golgi apparatus and endoplasmic reticulum but was not found near the plasma membrane (Fig. 14B).
|
|
Discussion |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
The results of this study demonstrate that 1) the distribution of CB1-IR axons is heterogeneous across neocortical regions of the macaque monkey with regions associated with higher cognitive functions, such as the prefrontal cortex, containing higher densities of CB1-IR axons than primary motor and sensory cortices; 2) different neocortical regions exhibit distinctive laminar distributions of CB1-IR axons, which precisely mark the cytoarchitectonic boundaries between many regions; 3) the density and distribution of CB1-IR axons also differ substantially across other regions of the primate brain; 4) the regional and laminar distributions of CB1-IR axons in the human neocortex are quite similar to those in monkey, although the morphology of labeled axons is altered by postmortem delay and; 5) in the monkey neocortex, CB1 immunoreactivity is primarily contained in cells and axon terminals that have the morphologic features characteristic of GABA neurons.
Methodological Considerations
Several lines of evidence indicate that the antibodies used in this study are selective for the central CB1 receptor. First, 3 antibodies raised against different portions of the CB1 receptor produced identical patterns of immunoreactivity. Second, preadsorption with the cognate peptides eliminated specific immunoreactivity. Third, previous studies confirmed the specificity of the anti-CB1–CT antibody by demonstrating the elimination of immunoreactivity in CB1 knockout mice, the identification of bands of appropriate molecular weight in western blot analysis, and the labeling of cultured cells transfected with the CB1 receptor (Hajos and others 2000; Katona and others 2001).
Previous studies employing immunocytochemistry in both rat (Tsou and others 1998) and monkey (Lu and others 1999; Ong and Mackie 1999) using antibodies raised against the first 77 N-terminus amino acids of the CB1 receptor reported a significantly lower density and less distinct laminar pattern of axon labeling and a much greater density of labeled cells including interneurons and pyramidal neurons. These differences in patterns of immunoreactivity could reflect the phosphorylation state of the C-terminus of the CB1 receptor, which mediates CB1 receptor internalization (Hsieh and others 1999). As suggested by Egertová and Elphick (2000), the last 13 amino acids of the CB1 receptor contain 6 sites that when phosphorylated may render the CB1 receptor nonimmunoreactive to antibodies directed against the C-terminus. Therefore, internalized receptors in cell bodies may not be recognized by C-terminus antibodies. However, we observed immunoreactivity within cell bodies, including labeling associated with cytoplasmic organelles, as has been reported in rodents using the same C-terminus antibodies (Hajos and others 2000; Katona and others 2001; Bodor and others 2005). Furthermore, in the present study we observed identical patterns of immunoreactivity with both C-terminus and N-terminus antibodies. Together, these data indicate that the C-terminus antibodies used in this study recognize CB1 receptors whether they are inserted in the plasma membrane or internalized.
Some studies suggest that CB1 receptors might be present presynaptically in pyramidal cell axon terminals where they could affect glutamate release. For instance, low levels of CB1 mRNA have been observed in pyramidal neurons (Marsicano and Lutz 1999), and the application of CB1 agonists has been reported to reduce the amplitude of excitatory postsynaptic currents (Auclair and others 2000; Hajos and others 2001). These findings could reflect alternative splicing of the N-terminus CB1 receptor to form a variant of the CB1 receptor (CB1A) (Shire and others 1995; Egertová and Elphick 2000). In the proposed CB1A splice variant, the first 61 amino acids of the CB1 receptor are replaced by 28 amino acids that are unrelated to the full CB1 receptor. It could be argued that it is the CB1A splice variant that is expressed by pyramidal neurons; however, the C-terminus antibodies used in this study would recognize both isoforms of the CB1 receptor, yet only inhibitory neurons, axons, and terminals were found to contain CB1 immunoreactivity. Alternatively, a yet unidentified cannabinoid receptor (or receptors) might be present in glutamatergic neurons (Begg and others 2005). This hypothesis is supported by findings that the ability of WIN 55212-2 (a CB1 receptor agonist) to block inhibitory postsynaptic currents is abolished in CB1 knockout mice, whereas its ability to prevent excitatory postsynaptic currents persists (Hajos and others 2001). In addition, WIN 55212-2 binds to tissue of CB1 knockout mice, but with lower affinity than in wild-type mice (Breivogel and others 2001).
Regional and Laminar Distributions of CB1-Containing Axons
CB1-IR axons were present throughout the monkey and human neocortex but showed marked regional variations in overall densities and laminar distributions. For example, the overall density of CB1-IR axons was much higher in association areas, such as the prefrontal cortex, than in primary sensory and primary motor regions. The distribution of CB1 immunoreactivity across monkey neocortical regions was broadly similar to that observed in the rat with several C-terminus antibodies (Egertová and Elphick 2000; Hajos and others 2000; Katona and others 2001; Bodor and others 2005) and to the distribution of CB1 receptor–binding sites (Herkenham and others 1991). However, some species differences in the laminar distribution of CB1-IR axons are worth noting. For instance, Egertová and Elphick (2000) reported that CB1-IR axons were most densely localized in layers 2–3 and 6 and least dense in layer 4 of the rat frontal and cingulate cortices, whereas the monkey layer 4 contained the highest density of CB1-IR axons in these regions. Furthermore, the dense band of CB1-IR axons in layer 5A, bordered by sparse axon labeling in layers 4 and 5B, in the rat somatosensory cortex (Bodor and others 2005) differs from the relatively similar density of axons across layers 2–3 and 5–6 and sparse axonal labeling in layer 4 in the monkey primary somatosensory cortex. In contrast to the differences observed in the laminar distribution of CB-IR axons in the monkey and rat, the laminar distribution of CB1-IR axons across homologous regions of the monkey and human is nearly identical. Furthermore, the laminar distribution of CB1 receptor–binding sites in the human revealed by autoradiography (Glass and others 1997) is quite similar to that of CB1-IR axons observed here.
The various laminar distributions of CB1-IR axons across the monkey neocortex raise the question of whether some CB1-IR axons could arise from extrinsic sources. Although the band of CB1-IR axons in layer 4 of some regions suggests a thalamic source for these axons, the absence of labeling in terminals forming asymmetric synapses (this study; Katona and others 1999, 2001; Hajos and others 2000; Bodor and others 2005) and the complete absence of CB1 immunoreactivity in the thalamus excludes this source. In contrast, the presence of CB1 immunoreactivity in cell bodies and terminals with the morphological features of GABA neurons suggests that the most likely source of neocortical CB1-IR axons is intrinsic inhibitory local circuit interneurons. This idea is supported by a recent study demonstrating that the axons of intracellularly labeled CB1-IR cells in layers 2–3 of the mouse neocortex arborized extensively within layers 2–3 but also projected into layers 4–6 (Galarreta and others 2004). Furthermore, recent evidence suggests that neocortical neurons containing CB1 receptors belong to the subpopulation of large basket GABA neurons that contain the neuropeptide CCK (Bodor and others 2005). Indeed, experiments in the rodent neocortex revealed that nearly all neurons expressing high levels of CB1 mRNA also express CCK mRNA and that inhibitory terminals containing CB1 immunoreactivity also contain CCK immunoreactivity (Marsicano and Lutz 1999; Bodor and others 2005). Furthermore, both the laminar pattern of CB1-IR neurons and axons and the perisomatic arrays formed by CB1-IR processes observed in this study directly match the findings of CCK-IR structures in monkey neocortex (Oeth and Lewis 1990, 1993).
Functional Implications
In the rodent, activation of CB1 receptors inhibits the release of GABA from presynaptic terminals and reduces GABAA receptor–mediated inhibitory postsynaptic currents in pyramidal neurons in the neocortex (Galarreta and others 2004; Trettel and others 2004; Bodor and others 2005