Cannabinoid Type 1 Receptor is Undetectable in Rodent and Primate Cerebral Neural Stem Cells but Participates in Radial Neuronal Migration

Cannabinoid type 1 receptor (CB1R) is expressed and participates in several aspects of cerebral cortex embryonic development as demonstrated with whole-transcriptome mRNA sequencing and other contemporary methods. However, the cellular location of CB1R, which helps to specify molecular mechanisms, remains to be documented. Using three-dimensional (3D) electron microscopic reconstruction, we examined CB1R immunolabeling in proliferating neural stem cells (NSCs) and migrating neurons in the embryonic mouse (Mus musculus) and rhesus macaque (Macaca mulatta) cerebral cortex. We found that the mitotic and postmitotic ventricular and subventricular zone (VZ and SVZ) cells are immunonegative in both species while radially migrating neurons in the intermediate zone (IZ) and cortical plate (CP) contain CB1R-positive intracellular vesicles. CB1R immunolabeling was more numerous and more extensive in monkeys compared to mice. In CB1R-knock out mice, projection neurons in the IZ show migration abnormalities such as an increased number of lateral processes. Thus, in radially migrating neurons CB1R provides a molecular substrate for the regulation of cell movement. Undetectable level of CB1R in VZ/SVZ cells indicates that previously suggested direct CB1R-transmitted regulation of cellular proliferation and fate determination demands rigorous re-examination. More abundant CB1R expression in monkey compared to mouse suggests that therapeutic or recreational cannabis use may be more distressing for immature primate neurons than inferred from experiments with rodents.


Introduction
The cerebral cortex is composed of a complex network of excitatory projection neurons and inhibitory interneurons that are generated, migrate to predetermined destination and establish specific synaptic connections through precisely orchestrated processes, and dysfunction during assembly may cause neuropsychiatric disorders such as lissencephaly, polymicrogyria and autism spectrum disorder [1][2][3][4][5][6][7]. Neural progenitors undergoing repeated cycles of proliferation in the ventricular zone (VZ) undergo interkinetic nuclear translocation whereby the postmitotic nucleus moves within the vertically elongated cell body to the basal region of the VZ during the S phase of the cell cycle and then it returns to the apical region for mitosis [8]. During interkinetic nuclear translocation, cells stay attached to the ventricular surface through adherence junctions between adjacent cells and do not actually migrate but instead, interphase nuclei may be displaced out from the apical region by other dividing and growing cells [9]. After a postmitotic cell detaches from the ventricular surface, it undergoes active migration towards the developing cortical plate where it settles as excitatory projection neuron. A fundamental detail of the molecular mechanism of neuronal nucleus movement and cell body locomotion is coordination of the cytoskeleton reformation by the centrosome-a highly conservative organelle that consists of two centrioles surrounded with matrix proteins and serves as the primary microtubule-organizing center in animal cells [10][11][12]. For example, cells undergoing mitosis cannot actively migrate because their cytoskeletons and particularly centrosomes are engaged in chromosome separation. Interphase neural stem cells (NSCs) in the VZ also do not actively migrate while they have both centrioles included in the structure of primary cilia as basal bodies and daughter centrioles making them unavailable for the cytoskeleton reformation crucial for cell migration. Release of the centrosome from the eliminated primary cilium and its relocation ahead of the migrating nucleus takes place simultaneously with initiation of migration in postmitotic neurons [10,[13][14][15][16]. When the migration ends, the centrosome finalizes its function for nuclear movement, attaches to the cell membrane and participates in the recovery of the primary cilium [17,18]. In the present electron microscopic study, we identified the centrosome position in the analyzed cells and used them as a fiduciary organelle for precise identification of the cellular developmental stage.
Retrograde synaptic signaling through CB 1 R is well documented, while many potential functions of the endocannabinoid system are less well understood [19][20][21]. CB 1 R is expressed in the mammalian brain early in development and are involved in vital and complex developmental events such as axonal growth and path finding [22][23][24] and in neuronal migration [25,26]. Endocannabinoidand CB 1 R-mediated promotion of neural progenitor differentiation into astroglial cellular line has been suggested [27][28][29], however, these developmental events demand more detailed investigation. For example, labeling by anti-CB 1 R sera was observed with light microscopy in neural progenitors and radial glia in vitro and in vivo [30][31][32]; but, precise CB 1 R cellular location, which determines the receptor function, remains to be demonstrated in NSCs. In the present study we analyzed CB 1 R expression and the subcellular location in the proliferating NSCs and postmitotic neurons migrating through transient embryonic zones of the cerebral cortex during the middle neurogenic stage in mice and rhesus macaques. Combining immunohistochemical (IHC) labeling with electron microscopic 3D reconstruction from uninterrupted serial ultrathin sections allowed us to estimate CB 1 R expression and distribution in distinct cells at specific developmental stages.

Results
In silico analysis of our previous quantitative whole-transcriptome mRNA sequencing data [33] demonstrated varied levels of expression of endocannabinoid system genes in mouse embryo cortical zones. More specifically, we found dramatic upregulation of the CB 1 R gene (Cnr1) in parallel with presumed prospective pyramidal neurons relocation from VZ to cortical plate (CP). Low content of CB 1 R in the VZ/subventricular zone (SVZ) raises the question about when CB 1 R is first expressed, which we addressed below. Expression of other genes of the endocannabinoid system was low or absent in all analyzed embryonic cerebral zones with the exception of an increase of Mgll-an endocannabinoid degrading enzyme-in samples combining CP and the marginal zone (MZ), while Mgll remains very low in VZ/SVZ and intermediate zone (IZ). This may be due either to Mgll expression in the interneurons tangentially migrating through the MZ, or its fast upregulation in projection neurons settling in the CP (Figure 1). Low expression of genes encoding endocannabinoid synthetizing enzymes indicates that the ligands of CB 1 R likely are not locally synthetized in the embryonic neocortex. They might be delivered to developing brain through blood circulation from other organs and/or mother organism through placental penetration or breast feeding during postnatal development [34,35].

Figure 1.
Quantitative whole-transcriptome mRNA sequencing data of the endocannabinoid system genes in the cortical zones from the embryonic day (E) 14.5 mouse embryo. X axis shows endocannabinoid system genes: Cnr1 and Cnr2 cannabinoid type 1 and type 2 receptors, respectively; endocannabinoid synthesizing enzymes Dagla and Daglb, sn-1-specific diacylglycerol lipase alpha and beta, respectively; endocannabinoid metabolizing enzymes Mgll and Faah, monoacylglycerol lipase and fatty acid amide hydrolase, respectively. Y axis shows the abundance of the genes by calculating reads per kilobase of the exon model per million mapped reads; average ± SEM from 3 embryos analyzed [33]. Notice dramatic upregulation of CB1R gene expression in parallel with neuronal maturation and relocation from VZ to CP. Expression of other genes of the endocannabinoid system is low if any in all analyzed embryonic cerebral zones making doubt for the capacity of immature cells to synthesize endocannabinoids. Increase of endocannabinoid degrading enzyme Mgll in CP/MZ may be due to its expression in the interneurons tangentially migrating through MZ or projection neurons settling in CP.
Next, we performed IHC labeling using different anti-CB1R sera followed by several visualization and amplification methods such as immunofluorescence, immunoperoxidase and silver amplification of ultra-small gold particles in the mouse embryo brain. These results indicate that antibodies generated against the last 15 amino acids (L15) of the CB1R C-terminus also stain radial processes in addition to the previously demonstrated selective labeling of CB1R in cortico-fugal axons and interneuron cell bodies [23,24]. Double immunolabeling for fluorescence and electron microscopy demonstrated colocalization of anti-CB1R staining with the radial glia cell marker Glast but not with neuronal markers TuJ1 and doublecortin ( Figure 2). Nevertheless, our control tests show equal staining of radial processes in wild type mouse embryos and CB1R −/− littermates, identifying this as artifactually positive CB1R immunolabeling in radial glia cells ( Figure 3). Selective axonal labeling was observed in wild type mice with different CB1R antibodies generated against C-terminus or amino-terminus (NH) while it was abolished in CB1R −/− embryos or by preincubation of the antibodies with their respective fusion peptides ( Figure 4). Out of several antibodies tested in this and our previous studies [24,36,37], CB1R antibody generated against last 31 amino acids (L31) provides maximal selectivity with minimal background staining that determined our choice of this antibody for electron microscopic 3D analysis (see below). Quantitative whole-transcriptome mRNA sequencing data of the endocannabinoid system genes in the cortical zones from the embryonic day (E) 14.5 mouse embryo. X axis shows endocannabinoid system genes: Cnr1 and Cnr2 cannabinoid type 1 and type 2 receptors, respectively; endocannabinoid synthesizing enzymes Dagla and Daglb, sn-1-specific diacylglycerol lipase alpha and beta, respectively; endocannabinoid metabolizing enzymes Mgll and Faah, monoacylglycerol lipase and fatty acid amide hydrolase, respectively. Y axis shows the abundance of the genes by calculating reads per kilobase of the exon model per million mapped reads; average ± SEM from 3 embryos analyzed [33]. Notice dramatic upregulation of CB 1 R gene expression in parallel with neuronal maturation and relocation from VZ to CP. Expression of other genes of the endocannabinoid system is low if any in all analyzed embryonic cerebral zones making doubt for the capacity of immature cells to synthesize endocannabinoids. Increase of endocannabinoid degrading enzyme Mgll in CP/MZ may be due to its expression in the interneurons tangentially migrating through MZ or projection neurons settling in CP.
Next, we performed IHC labeling using different anti-CB 1 R sera followed by several visualization and amplification methods such as immunofluorescence, immunoperoxidase and silver amplification of ultra-small gold particles in the mouse embryo brain. These results indicate that antibodies generated against the last 15 amino acids (L15) of the CB 1 R C-terminus also stain radial processes in addition to the previously demonstrated selective labeling of CB 1 R in cortico-fugal axons and interneuron cell bodies [23,24]. Double immunolabeling for fluorescence and electron microscopy demonstrated colocalization of anti-CB 1 R staining with the radial glia cell marker Glast but not with neuronal markers TuJ1 and doublecortin ( Figure 2). Nevertheless, our control tests show equal staining of radial processes in wild type mouse embryos and CB 1 R −/− littermates, identifying this as artifactually positive CB 1 R immunolabeling in radial glia cells ( Figure 3). Selective axonal labeling was observed in wild type mice with different CB 1 R antibodies generated against C-terminus or amino-terminus (NH) while it was abolished in CB 1 R −/− embryos or by preincubation of the antibodies with their respective fusion peptides ( Figure 4). Out of several antibodies tested in this and our previous studies [24,36,37], CB 1 R antibody generated against last 31 amino acids (L31) provides maximal selectivity with minimal background staining that determined our choice of this antibody for electron microscopic 3D analysis (see below). double-labeled for CB1R-L15 (immunogold/silver particles, arrows) and Glast (diffuse DAB immunoperoxidase labeling) in the thalamus. (c1,c2,d1,d2) Double immunofluorescence reveals dissimilar patterns for CB1R-L15 labeling versus neuronal markers TuJ1 (c1, c2) and doublecortin (d1,d2) in the ganglionic eminence (ge), thalamus (tha) and hippocampus (hip). Ventricular surfaces are underlined with white dotted lines. Abbreviations: f, fimbria; N, cell nucleus. Scale bars: (a,c,d) 100 μm and (b) 1 μm.   In accord with the transcriptome data ( Figure 1), our IHC labeling identified distinct amounts of CB1R in the embryonic cortical zones ( Figure 5). In both the mouse and macaque monkey, VZ and SVZ are mostly CB1R-immunonegative whereas bundles of horizontal CB1R-positive axons emanating from cortical projection neurons are abundant in the IZ [23]. The MZ contains numerous CB1R-positive cell bodies of tangentially migrating interneurons [24]. Mouse CP demonstrated minor CB1R accumulation, whereas CB1R-positive globules in radially migrating cell bodies (see below for a detailed electron microscopic analysis) are numerous in the macaque monkey ( Figure 5). . CB 1 R immunolabeling in E15 mouse hippocampal fimbria with CB 1 R-L15 (a-d) and CB 1 R-NH (e-h) antibodies and in E14 neocortex with CB 1 R-L31 (i-l) antibodies. Selective axonal labeling (arrows) in the fimbria (f) and neocortical intermediate zone (iz) was observed in wild type embryos with light (a,e,i) and electron microscopy (b,f,j). In accord with the location of CB 1 R C-terminus on the inner surface of the cell membrane and NH-terminus on the outer surface, the DAB-Ni immunoperoxidase reaction end-product (arrowheads) was located inside or outside of the axons, correspondingly. Immunolabeling was abolished by preincubation of the antibodies with the corresponding fusion peptides (c,g,k) and in CB 1 R −/− littermate embryos (d,h,l). Unlabeled fimbria in (c,d,g,h) are indicated with ovoids. Border between subventricular zone (SVZ) and intermediate zone is indicated with a dashed line. Abbreviations: ge, ganglionic eminence; hip, hippocampus; tha, thalamus. Scale bars: (a,c-e,g-i,k,l) 50 µm and (b,f,j) 0.5 µm.
In accord with the transcriptome data ( Figure 1), our IHC labeling identified distinct amounts of CB 1 R in the embryonic cortical zones ( Figure 5). In both the mouse and macaque monkey, VZ and SVZ are mostly CB 1 R-immunonegative whereas bundles of horizontal CB 1 R-positive axons emanating from cortical projection neurons are abundant in the IZ [23]. The MZ contains numerous CB 1 R-positive cell bodies of tangentially migrating interneurons [24]. Mouse CP demonstrated minor CB 1 R accumulation, whereas CB 1 R-positive globules in radially migrating cell bodies (see below for a detailed electron microscopic analysis) are numerous in the macaque monkey ( Figure 5). Our light microscopic observations indicate that VZ/SVZ cells were mostly immunonegative with the exception of sporadic positive vertical processes. The positions of the cell bodies emitting these processes were not visible with light microscopy provoking the hypothesis that they are emitted by VZ radial glia cells. To try to identify these cell bodies, we applied correlative light/electron microscopic analysis and unexpectedly detected CB1R-expressing growth cones directed towards the ventricular surface identifying the vertical CB1R-positive processes to be unusually oriented growing axons ( Figure 6). Most likely, they represent disoriented cortico-fugal axons emitted by developing pyramidal neurons located in the CP [23]. Electron microscopic 3D reconstruction shows that several vertical axons may continuously contact each other revealing reciprocal affinity and contact navigation of growing CB1R-expressing axons (Figure 6). Although a detailed study of this phenomenon is beyond the scope of the present article, atypical CB1R-expressing axons in the VZ/SVZ indicate a technical opportunity to identify rare CB1R-containing structures further Our light microscopic observations indicate that VZ/SVZ cells were mostly immunonegative with the exception of sporadic positive vertical processes. The positions of the cell bodies emitting these processes were not visible with light microscopy provoking the hypothesis that they are emitted by VZ radial glia cells. To try to identify these cell bodies, we applied correlative light/electron microscopic analysis and unexpectedly detected CB 1 R-expressing growth cones directed towards the ventricular surface identifying the vertical CB 1 R-positive processes to be unusually oriented growing axons ( Figure 6). Most likely, they represent disoriented cortico-fugal axons emitted by developing pyramidal neurons located in the CP [23]. Electron microscopic 3D reconstruction shows that several vertical axons may continuously contact each other revealing reciprocal affinity and contact navigation of growing CB 1 R-expressing axons (Figure 6). Although a detailed study of this phenomenon is beyond the scope of the present article, atypical CB 1 R-expressing axons in the VZ/SVZ indicate a technical opportunity to identify rare CB 1 R-containing structures further confirming immunonegativity of majority of VZ/SVZ cells.
As a next step, we applied electron microscopic 3D reconstruction to analyze cells at distinct developmental stages. A total of 67 randomly chosen cells from wild type mice and monkeys were analyzed; see Supplementary Table S1 for detailed morphologic characterization of each cell. Mitotic and postmitotic cells in the VZ were either spherical or emitting radial processes (n = 9 cells analyzed in this article; see also [38]). Out of a total of 52 interphase cells from SVZ, IZ and CP of wild type mice and monkeys, 50 showed pronounced vertical elongation of the cell body and the nucleus. Out of two non-vertical cells located in the IZ, one was multipolar (cell #M14) with unpronounced orientation of processes [39], and the other one was bipolar and horizontally oriented (cell #M15)-presumed to be a tangentially migrating interneuron, similar to horizontal cells in MZ (n = 6 cells). Among the vertical cells from SVZ, IZ and CP, 39 were bipolar; 7 were multipolar with 3 or 4 processes and situated in the basal layer of CP-likely finalizing those vertical migration; 3 were unipolar with basally directed process-likely undergoing vertical somal translocation [40]; and one cell (#R10) in SVZ was unipolar with the apical process. Thus, bipolar cells predominate in developing neocortex, reflecting the massive vertical migration taking place [2].
In accordance with previous publications [10], we observed a characteristic behavior of centrosomes that we used for more substantial identification of the cellular developmental stages. Namely, early prophase cells (n = 2) showed centrioles included in the cilia apparatus and duplication of centrosomes located close to the ventricular surface; prometaphase cells (n = 2) demonstrated two centrosomes separated for prospective daughter cells and absence of primary cilia. In numerous analyzed VZ and SVZ cells (n = 24) in mice and macaque monkeys, we did not find centrosomes in the vicinity of the cell body that indicates they were distantly located in truncated processes; while two cells showed centrosomes in the apical processes indicating that these cells had detached from the ventricular surface and were preparing for basal migration (Figure 7). So, most cells in VZ/SVZ of E16 mice and E45 monkeys are likely in the cycle of interkinetic nuclear translocation or in the stage of initiation of vertical migration. In contrast, in the analyzed vertical cells in IZ and CP (n = 36), we identified 30 centrosomes, 26 of which were located basally relative to the nucleus and 4 showed the position aside of the nucleus (Supplementary Table S1 and Figures 8-10). Although centrosomes were missed in some incompletely reconstructed processes, the predomination of basal centrosome position confirms massive basal migration of projection neurons in IZ and CP. Thus, cellular morphology and intracellular location of centrosomes identify interkinetic nuclear translocation of VZ/SVZ cells, initiation of basal migration in SVZ cells and massive locomotion through the IZ and CP.        Despite an extensive search, we did not find credible CB 1 R immunolabeling in VZ or SVZ mitotic and postmitotic cells (n = 24 cells occasionally chosen for 3D reconstruction). The trace labeling observed (about one locus of the antibodies binding per 100 µm 3 of cytoplasm) was similar in wild type and CB 1 R −/− mice indicating that this is the level of non-selective background (Supplementary Tables S1 and S2, Figures 10 and 11). The earliest observation of CB 1 R-L31 immunoreaction end-product depositions is in the cytoplasm of IZ cells. Neurons migrating radially through IZ and CP contained intracellular vesicles surrounded by the immunoreaction end-product forming typical CB 1 R-positive globules previously detected only in the interneurons [36,41]. CB 1 R-positive globules were numerous in nearly all migrating neurons in monkeys, whereas in mice only 4 out of 21 vertical migrating cells contained similar globules that show less intensive labeling. Thus, CB 1 R was undetectable in VZ and SVZ cells while they were undergoing repetitive circles of proliferation or initiating migration, whereas postmitotic neurons migrating radially through IZ and CP express CB 1 R and this expression was more abundant in macaque monkeys than in mice. Despite an extensive search, we did not find credible CB1R immunolabeling in VZ or SVZ mitotic and postmitotic cells (n = 24 cells occasionally chosen for 3D reconstruction). The trace labeling observed (about one locus of the antibodies binding per 100 μm 3 of cytoplasm) was similar in wild type and CB1R −/− mice indicating that this is the level of non-selective background (Supplementary Tables S1 and S2, Figures 10 and 11). The earliest observation of CB1R-L31 immunoreaction endproduct depositions is in the cytoplasm of IZ cells. Neurons migrating radially through IZ and CP contained intracellular vesicles surrounded by the immunoreaction end-product forming typical CB1R-positive globules previously detected only in the interneurons [36,41]. CB1R-positive globules were numerous in nearly all migrating neurons in monkeys, whereas in mice only 4 out of 21 vertical migrating cells contained similar globules that show less intensive labeling. Thus, CB1R was undetectable in VZ and SVZ cells while they were undergoing repetitive circles of proliferation or initiating migration, whereas postmitotic neurons migrating radially through IZ and CP express CB1R and this expression was more abundant in macaque monkeys than in mice.  Out of 12 migrating neurons randomly chosen for 3D reconstruction from the IZ of CB 1 R −/− mice, eight emitted processes in directions that deviated from basal or apical. One cell had only the apical process, which is also unusual for IZ cells; and only three cells showed a normal bipolar vertical morphology (Supplementary Table S2; Figure 11). This contrasts to neurons in the IZ of wild type mice and monkeys where these neurons were mostly bipolar (9 out of 14 cells analyzed; Figure 10). Neurons migrating radially through the CP of CB 1 R −/− mice (n = 4) apparently recovered the typical bipolar phenotype (Figure 11). Thus, projection neurons in CB 1 R −/− mice show morphological features of temporarily disoriented migration such as a non-vertically shaped cell body and an increased number of processes emanating in random directions. Delay of correct neuronal migration may contribute to developmental disorders of cerebral cortex and later behavioral consequences for the animal.

Discussion
Here we applied several methods, such as in silico analysis of whole-transcriptome mRNA data and light and electron microscopic IHC to estimate expression of the endocannabinoid signaling system and particularly CB 1 R in distinct cells at specific developmental stages in mouse and rhesus macaque. Using morphological and ultrastructural criteria such as the laminar position and shape of the cell body and intracellular position of the centrosome, we identified stages of projection neuron generation and migration. We found that CB 1 R in the cell body accumulates in parallel with cell-autonomous development, namely: mitotic and postmitotic cells in VZ and SVZ were immunonegative, whereas the cells migrating through IZ and CP contain CB 1 R-positive intracellular vesicles. Visible CB 1 R immunoreactivity of vertical processes emanating from radial glia cells may be non-selective due to nonspecific immunolabeling conditions. Simultaneously, relatively rare authentic CB 1 R-positive vertical processes in the VZ/SVZ may represent disoriented axons most likely emanating from cortical projection neurons [23]. We conclude that the low levels or absence of CB 1 R from proliferating VZ/SVZ cells indicates that previously suggested direct influence of the cannabinoid signaling system and particularly CB 1 R to cell proliferation and developmental fate of NSCs [29] should be rigorously examined. Observed actions of cannabis and endocannabinoids on proliferating neurons [27,28] may be less direct and, for example, may be transmitted through other molecules, cells, organs or the maternal organism. Potential actions of CB 1 R in later prenatal and postnatal proliferation zones such as the neocortical outer SVZ or hippocampal dentate gyrus [42] demands further investigation using methods with high cellular and subcellular resolution.
Consistent with previous studies of developing mice [25,26], here we showed that initial expression of the CB 1 R gene and accumulation of the protein in the cell body coincided with the initiation of postmitotic neurons migration towards their final positions in the developing cortical plate. Block of endocannabinoid signaling in CB 1 R −/− mice causes a deviation from the correct vertical migration that is detectable as a non-vertical shape of the cell bodies, which also emit numerous disoriented processes. Such a cell migration disorder may be temporary and its overall impact only moderate for the animal's ontogenesis consistent with the relatively normal cortical development of CB 1 R −/− mice [43]. Nevertheless, as we show here, CB 1 R expression in cerebral projection neurons is more pronounced in monkeys than in the mouse as judged by the number of CB 1 R-positive cells and quantity and intensity of labeling of CB 1 R-containing intracellular vesicles in each cell. Therefore, medicinal or recreational cannabis use may affect immature primate neurons more than is apparent from rodent experiments. CB 1 R-containing vesicles may play a role in intracellular self-regulation sharing pathways similar to those used in slow self-inhibition, which is elicited by cell autonomous Ca 2+ -dependent production of endocannabinoids and K + channel activation, as previously detected in cholecystokinin/CB 1 R expressing interneurons and a subpopulation of cortical pyramidal neurons in adult mice [44,45]. Fluctuations of Ca 2+ concentration in the somata may influence neuronal migration, likely through the reformation of a microtubule and other components of the cytoskeleton [46][47][48]. Thus, CB 1 R-containing intracellular vesicles in migrating neurons represent a potential molecular substrate for endocannabinoid participation and cannabis disruption of cytoskeleton reorganization through regulation of intracellular Ca 2+ oscillations.

In Silico Quantification of Cannabinoid System Genes Expression in the Mouse Embryo Cerebral Cortex
Quantitative whole-transcriptome mRNA sequencing data were obtained from Rakic lab website https://medicine.yale.edu/lab/rakic/transcriptome/ (accessed on 10 July 2013). The data are also publicly available at the online base of National Center for Biotechnology Information at https://www.ncbi.nlm. nih.gov.

Animals
Animal protocols for non-human primates and mice were complied with the NIH guidelines for animal care and use and were approved by the Institutional Animal Care and Use Committee of Yale University (Protocol #2018-10297.A2 approved 22 July 2020; protocol #2018-10750.A3 approved 25 June 2020). Timed-pregnant macaque and mice were housed in the vivarium of Yale University on a recommended diet and socializing. For terminal surgery, the animals were anaesthetized with sodium pentobarbital (3 mL/kg) or euthasol (0.5 mL/kg). Two monkey embryos at E45 were collected through cesarean sections performed by veterinarians. Fetal age was estimated based on weeks after ovulation, crown-rump length and anatomical landmarks. No apparent abnormalities were noted at the time of tissue collection. Monkey embryos were decapitated, brains were removed and one hemisphere was used for ex vivo analyses for other projects while the other was immersed in fixative containing 4% paraformaldehyde, 0.2% picric acid and 0.2% glutaraldehyde in 0.1 M phosphate buffer (pH 7.4) for 3-4 days at 4 • C. Timed-pregnant mice were killed with cervical translocation and the embryos exteriorized from the uterus and decapitated, and their brains removed and fixed as above. Wild type C57BL6 mouse embryos at E14 (n = 4), E15 (n = 8) and E16 (n = 4) and CB 1 R-knock out (CB 1 R −/− ) embryos in C57BL6 background (generated at NIMH, Bethesda, MD [43]) at E14 (n = 4) and E15 (n = 3) were used for morphologic analyses.

IHC Labeling for Light and Electron Microscopy
IHC labeling was performed as previously described [37,49,50]. Briefly, coronal 80-m-thick slices from the embryo cerebral cortex were cut by a vibratome. For double immunofluorescence labeling, slices were blocked in solution of 5% bovine albumin with 0.05% Triton X-100 and immersed in the mixture of rabbit polyclonal antibodies generated against the last 15 amino acids of rat CB 1 R (CB 1 R-L15; dilution 1:1000 [51]), with guinea pig Glast polyclonal antibodies (Synaptic Systems, Goettingen, Germany; catalog number 250114; 1:5000), monoclonal mouse neuronal class III-tubulin antibodies (TuJ1; Covance, Berkeley, CA; catalog number MMS-435P; 1:1000), or goat doublecortin (C-18) polyclonal antibodies (Santa Cruz Biotechnology, Dallas, TX; catalog number sc-8066; 1:1000) overnight at room temperature. Then slices were immersed in a mixture of the corresponding combination of secondary antibodies conjugated with Alexa A-594 and Alexa A-488 (all from Molecular Probes, Eugene, OR; 1:300) and mounted on microscope slides. Slices were evaluated and photographed in microscope Axioplan 2 equipped with the Apotom system (Zeiss, Jena, Germany). For electron microscopy, vibratome slices were cryoprotected in 30% sucrose, freeze-thawed over liquid nitrogen, blocked in 5% bovine albumin and incubated in rabbit CB 1 R-L15 (1:1000), rabbit polyclonal antibodies generated against amino-terminus of rat CB 1 R (CB 1 R-NH; 1:2000 [52]) or guinea pig polyclonal antibodies generated against last 31 amino acids of mouse CB 1 R (CB 1 R-L31; Frontier Science Co, Ltd., Japan; catalog number CB1-GP-Af530-1; 1:2000) overnight at room temperature. Then slices were immersed in respective solutions of biotinylated secondary antibodies (Jackson Immunoresearch Inc., West Grove, PA; 1:300) and developed by the Elite ABC kit (Vector Laboratories, Burlingame, CA) with Ni-intensified 3,3 -diaminobenzidine-4HCl (DAB-Ni) as a chromogen. For double immunolabeling, we first applied CB 1 R-L15 serum as above and goat antirabbit IgGs conjugated with 1 nm gold particles (1:80) with the subsequent silver intensification kit R-Gent SE-LM (all from Aurion, Wageningen, The Netherlands). Then, Glast was visualized with made-in-guinea pig antibodies and the immunoperoxidase reaction as above with 3,3 -diaminobenzidine-4HCl (DAB) as a chromogen. Specificity of CB 1 R antibodies was tested in CB 1 R −/− embryos and by preincubation of the primary antibodies (dilution as above) with corresponding fusion proteins (concentration 1 g/mL) for 2 h before the immunoreaction. See also our previous articles [37,53] for control tests in CB 1 R −/− animals. Slices were post-fixed with 1% OsO 4 , dehydrated and embedded in Durcupan (ACM; Fluka, Buchs, Switzerland) on microscope slides. Selected neocortex segments were photographed with an Axioplan 2 microscope (Zeiss, Jena, Germany).

Serial Ultrathin Sectioning and Electron Microscopic 3D Reconstruction
For electron microscopic investigations, fragments from neocortical zones were re-embedded into Durcupan blocks and cut by a Reichert ultramicrotome into 70-nm-thick sections. Series of 150-200 consecutive sections were collected on one-slot grids covered with Butvar B-98 films (EMS, Hatfield, PA, USA), stained with lead citrate and evaluated in the JEM 1010 electron microscope (JEOL, Tokyo, Japan) equipped with the Multiscan 792 digital camera (Gatan, Pleasanton, CA, USA) or Talos L120C electron microscope (ThermoFisher Scientific, Boston, MA, USA). Serial images of mitotic or interphase cells from identified zones of the embryo neocortex were made with 12,000× or 15,000× magnification of JEM1010 or with 3400× of the Talos L120C electron microscope. 3D reconstruction of cell bodies with proximal segments of the processes and measurement of the volume and volume of the nucleus or chromosomes were performed using the computer program Reconstruct 1.1.0.0. (Boston, MA) [54,55] publicly available at http://www.bu.edu/neural/Reconstruct.html.
Mitotic phases of the 3D reconstructed cells were identified by the intracellular position of duplicated centrosomes and condensed chromosomes. Projecting processes were identified by their continuity with the cell membrane and a diameter of at least 0.5 µm in contrast to filopodia of 0.1-0.2 µm diameter. Growth cones were identified through the termination of the process with numerous filopodia [23]. The migration stage of interphase cells was determined through the laminar position and shape of the cell body with projecting processes and the intracellular location of the centrosome (if identified) relative to the nucleus. Centrosome position relative to the nucleus was defined as basal, apical or aside of the nucleus. The following morpho-functional/developmental stages of centrosomes were identified in the interphase cells: (1) both centrioles are in the cytoplasm; (2) the mother centriole is attached to the cilial vesicle in the cytoplasm; (3) the mother centriole is attached to the cell membrane, no cilia is formed and (4) both centrioles are included in the cilia apparatus as the basal body and daughter centriole [17]. Following migration, stages of interphase cells were identified: (1) interkinetic nuclear translocation-VZ/SVZ cells with adherent junctions and cilium on the ventricular surface; (2) initial vertical migration-bipolar cells in VZ/SVZ detached from the adherent junctions with the centrosome in the apical segment; (3) locomotion-bipolar vertical cells in IZ/CP with the centrosome located basally or aside of the nucleus; (4) terminal vertical migration-multipolar vertical cells in CP; (5) multipolar migration-multipolar cells in IZ [39]; (6) somal translocation-radially migrating cells in IZ/CP with the single basal process [40] and (7) lateral migration-horizontal bipolar cells in IZ or MZ. We considered interkinetic nuclear translocation and initial radial migration stages of VZ/SVZ cells indistinguishable from each other if centrosomes were not detected in the serial sections. Anti-CB 1 R DAB-Ni depositions in each cell were traced for 3D reconstruction and counted as "single" depositions if local staining were identified in at least three serial sections; merging conglomerates of the immunoreaction end-product around intracellular vesicles were counted as "globule" depositions. Numbers of "single" and "globule" depositions from each cell were normalized per 100 µm 3 of the analyzed cytoplasm and average ± SEM was calculated for each embryonic cortical zone in wild type animals or totally for CB 1 R −/− embryos. Although such quantification obviously underestimates immunolabeling in the cells with a high level of CB 1 R expression due to merging of adjacent DAB-Ni depositions, it is conclusive for low labeling.