Newborn neurons migrate along the processes of radial glial cells (RGCs) to reach their final positions in the cortex. Here, we visualized individual migrating neurons and RGCs using in utero electroporation. We show that branching of migrating neurons and RGCs is closely correlated spatiotemporally with the distribution of Reelin. Time-lapse imaging revealed that the leading processes of migrating neurons gave rise to increasingly more branches once their growth cones contacted the Reelin-containing marginal zone. This was accompanied by translocation of the nucleus and gradual shortening of the leading process. Absence of Reelin in reeler mice altered these processes resulting in misorientation, loss of bipolarity, and aberrant migration of cortical neurons. Moreover, in reeler, the branching of the basal processes of RGCs in the marginal zone was severely disrupted. Consistent with previous reports, we show that in dissociated reeler cortical cultures, exposure to recombinant Reelin enhanced dendritic complexity and glial branching. Our results suggest that Reelin induces branching of the leading processes of migrating neurons and that of basal processes of RGCs when they arrive at the Reelin-containing marginal zone. Branching of these processes may be crucial for the termination of nuclear translocation during the migratory process and for correct neuronal positioning.
The mammalian cerebral cortex consists of 6 well-organized layers extending from the pial surface to the white matter. The formation of this highly ordered structure requires the proper migration of postmitotic neurons from their sites of origin to their final destinations (Rakic 1988; Walsh and Goffinet 2000; Nadarajah and Parnavelas 2002). Disruption of neuronal migration leads to malformation of the cerebral architecture and to functional deficits (Gleeson and Walsh 2000; Ayala et al. 2007). Previous studies demonstrated that neuronal migration is precisely orchestrated by various intrinsic and extrinsic factors. Reelin is one of the crucial molecules involved in neuronal migration and cortical lamination.
Reelin is an extracellular matrix protein secreted by Cajal–Retzius (CR) cells located in the marginal zone (MZ) of the developing forebrain (D'Arcangelo et al. 1995, 1997). Absence of Reelin results in migration defects of many, but not all, neurons and in altered layer formation in laminated brain structures, such as the cerebral cortex, hippocampus, and cerebellum (Caviness and Rakic 1978; Del Rio et al. 1997; Curran and D'Arcangelo 1998; Frotscher 1998). In the Reelin-deficient mouse mutant reeler, many neurons show abnormal polarity orientation and dendritic abnormalities as well as malformation of the radial glial scaffold (Caviness and Sidman 1973; Goffinet and Lyon 1979; Frotscher et al. 2003; Hartfuss et al. 2003; Tissir and Goffinet 2003; Jossin 2004; Herz and Chen 2006; Frotscher 2010). Reelin exerts its function by binding to two lipoprotein receptors, very low-density lipoprotein receptor (VLDLR) and apolipoprotein-E receptor type 2 (ApoER2) (D'Arcangelo et al. 1999; Hiesberger et al. 1999; Trommsdorff et al. 1999). Binding of Reelin to these two receptors induces phosphorylation of the intracellular adaptor protein Disabled1 (Dab1) (Howell et al. 1997, 1999; Sheldon et al. 1997; Ware et al. 1997; Trommsdorff et al. 1999) by Src-family kinases (SFKs), particularly Src and Fyn (Arnaud et al. 2003; Bock and Herz 2003). Various signaling pathways downstream of Dab1 connect Reelin to the dynamics of the actin- and microtubule cytoskeleton (D'Arcangelo 2006; Chai et al. 2009; Förster et al. 2010; Frotscher 2010; Zhao and Frotscher 2010). Mutant mice deficient in Dab1 or in both VLDLR and ApoER2 show a reeler-like phenotype (Pinto Lord and Caviness 1979; Howell et al. 1997; Sheldon et al. 1997; Trommsdorff et al. 1999). Recent in vitro studies reported Reelin-induced branching of radial glial cell (RGC) processes and that of neuronal dendrites (Niu et al. 2004, 2008; Jossin and Goffinet 2007; Matsuki et al. 2008). Although these studies have demonstrated that Reelin is closely associated with neuronal migration and dendritic development, little is known about the underlying molecular and cellular mechanisms in space and time.
Using timed in utero electroporation in combination with immunocytochemistry for Reelin and time-lapse imaging, we demonstrate that intensive branching of migrating neurons and RGCs was closely correlated spatiotemporally with the distribution of endogenous Reelin. Our data also show that dendritic arbor complexity is inversely correlated with the length of the leading process of migrating neurons in their “terminal translocation” stage. In reeler mice, cortical neurons were unable to reach their defined destination and ectopically stopped the migratory process in the deep layers of the cortical plate (CP). Many of them lost their characteristic bipolarity and were misoriented. Moreover, branching of RGCs was strikingly impaired in reeler mutant mice leading to a dramatic reduction in the number of basal branches and end-feet. We hypothesize that Reelin promotes the branching of basal RGC processes, resulting in the stabilization of the RGC cytoskeleton, thus providing an appropriate template for migrating neurons. By inducing the branching of the leading processes of neurons, Reelin may contribute to anchoring these processes to the MZ matrix, which maintains neuronal bipolarity and orientation. Branching of the leading processes, in turn, may terminate the final phase of neuronal migration by nuclear translocation.
Materials and Methods
Reeler mice were purchased from the Jackson Laboratory (Bar Harbour). POMC-EGFP transgenic mice were generously provided by Dr. G. L. Westbrook (Vollum Institute). The animals were bred in the Experimental Animal Center of Lanzhou University and in the Experimental Animal Center of ZMNH. The animals were reared on a normal 12 h light/dark schedule. Noon of the date on which the vaginal plug was detected in the morning was designated as embryonic day 0.5 (E0.5). The first neonatal day was considered to be postnatal day 0 (P0). All experiments were carried out in accordance with the principles and guidelines of the Ethics Committees of Lanzhou University and Hamburg University. The genotype of the mutants was confirmed by PCR analysis of genomic DNA, as described (Deller et al. 1999; Overstreet et al. 2004).
The plasmid used in the present study, pCAG-GFP (Matsuda and Cepko 2004), was purchased from Addgene (Plasmid 11150, Addgene), which contains CAG (cytomegalovirus immediate-early enhancer/chicken β-actin hybrid) promoter (Niwa et al. 1991). The plasmid was purified using a Qiagen plasmid maxi kit.
In Utero Electroporation
In utero electroporation was performed according to the procedure described elsewhere previously (Tabata and Nakajima 2001), with slight modifications. Timed-pregnant females were deeply anesthetized by intraperitoneal injection of chloral hydrate (4.3 mg per 10 g of body weight). Then, the uterine horns of the pregnant mice were carefully exposed via a midline abdominal incision by pinching gaps between embryos with ring-forceps. 1–1.5 μL of plasmid DNA solution in normal saline was injected into the lateral ventricle of mouse embryos using a mouth-controlled pipette system and a pulled-glass micropipette. Fast Green solution (0.1%) was added to the plasmid solution in a ratio of 1 : 10 to monitor the injection. The heads of injected embryos in the uterus were placed between the tweezers-type electrodes. Electronic pulses (30–50 V; 50 ms) were charged 5 times at intervals of 900 ms with an electroporator (Gene Pulser Xcell Electroporation System). The uterine horns were placed back into the abdominal cavity to allow the embryos to continue normal development until the desired time of observation.
Slice Co-Culture and Live Imaging
Slice co-cultures were prepared as described previously (Zhao et al. 2004). To visualize individual granule cells in the reeler dentate gyrus (DG) for time-lapse imaging, POMC-EGFP transgenic mice, in which newborn granule cells in DG express GFP under the POMC promoter (Overstreet et al. 2004), were crossbred with reeler mice. For the preparation of co-cultures, newborn POMC-EGFP transgenic reeler mouse pups (P0) and young postnatal wild-type mice (P7) were used. Brains were removed following decapitation under hypothermic anesthesia. The hippocampi were dissected and sliced (300 μm) perpendicular to their longitudinal axis with a McIlwain tissue chopper. Hippocampal slices of newborn POMC-EGFP transgenic reeler mice were positioned in close vicinity to the outer molecular layer of the DG of slices from P7 wild-type mice so that the DG of the reeler slice came in direct contact with the Reelin-containing MZ of the wild-type slice. Slices were placed onto Millipore membranes and transferred to a 6-well plate with 1 mL/well nutrition medium (25% heat-inactivated horse serum, 25% Hank's balanced salt solution, 50% minimal essential medium, 2 mm glutamine, pH 7.2). Slices were first put into an incubator containing 5% CO2 at 37°C for 6 to 8 h to allow for recovery of the slices from injury during preparation. Then, the slices were transferred to a small incubator for live imaging using a confocal microscope (Improvision LiveCell Spinning Disk, PerkinElmer). The incubator was supplied with 5% CO2 and heated to 37°C. Time-lapse images were collected every 7 min for more than 12 h. Following imaging, the cultures were immediately fixed with 4% paraformaldehyde (PFA) at 4°C overnight. After rinsing in 0.1 m PB, immunostaining for Reelin and GFP was performed in the fixed cultures. Then, the cultures were counterstained with DAPI and mounted in Moviol. Images were acquired using a Zeiss LSM 510 confocal microscope.
Production of Recombinant Reelin
Preparation of Reelin-containing supernatants and control supernatants was performed as described previously (Förster et al. 2002; Zhao et al. 2004). In brief, stably transfected HEK 293-cells expressing either full-length Reelin or control cells transfected with a control vector (MOCK) were grown in Dulbecco's modified eagle medium (DMEM, low glucose, Invitrogen) with 10% fetal bovine serum, low glucose (Invitrogen), 100 U/mL penicillin, and 100 μg/mL streptomycin (Invitrogen), and 0.9 g/L G418 (Invitrogen) for 2 days to reach full confluence. The medium was then replaced by serum-free DMEM, and the cells were incubated for 2 days at 37°C in 5% CO2. Conditioned medium was collected and centrifuged. The supernatant was concentrated 10-fold using 100-kDa cut-off centrifugal filters (Millipore), sterile filtered, and stored at −80°C until used.
Neuronal Culture and Treatment with Recombinant Reelin
Dissociated cultures of neocortex were prepared as described previously (Chai et al. 2009). For primary cell cultures, 6 brains of E17.5 reeler mutants were used. The neocortex was dissected and put into ice-cold Hanks' Balanced Salt Solution (HBSS, Invitrogen). After removal of the meninges, the neocortex were collected and maintained in 15 mL Falcon tubes (BD Biosciences) containing 3 mL of 0.5% Trypsin and 0.53 mM EDTAx4Na (Invitrogen) and incubated at 37°C for 10 min. The neocortex was then washed twice in ice-cold HBSS and manually triturated with a polished glass Pasteur pipet for several times. Then, the cells were centrifuged at low speed (800g) for 3 min to discard dead cells. The sediment was collected in fresh Falcon tubes and centrifuged again. The supernatant was discarded, and cell pellets were resuspended in 1 mL of Neurobasal-A medium (Invitrogen), supplemented with 2% B27 (Invitrogen), 1 mM Glutamax (Invitrogen), 100 U/mL penicillin, and 100 μg/mL streptomycin (Invitrogen). Cortical cells were counted using a hemocytometer and suspended at a density of 5 × 105 cells in 0.5 mL aliquots on glass coverslips inserted in 24-well plates. The coverslips had been coated overnight with a solution of 20 μg/mL poly-l-ornithine (Sigma) prepared in borate buffer, pH 8.1. After 2 h of incubation in a humidified incubator at 37°C and 5% CO2 to allow for cell attachment, the medium was changed to remove dead cells. Then, 500 μL of Reelin-containing supernatant or control supernatant was added to medium. After another 24 h of incubation, cultures were fixed with 4% PFA in 0.1 m PBS, pH 7.4, for 15 min at room temperature (RT) and rinsed in 0.1 m PBS for 1 h.
After permeabilization and preincubation with blocking solution (5% normal goat serum and 1% bovine serum albumin in 0.1 m PB containing 0.2% Triton X-100) at RT for 1 h, the cultures were immunostained with primary antibodies against Microtubule-associated protein 2 (MAP2) and brain lipid-binding protein (BLBP) at 4°C overnight and subsequently with Alexa Fluor-488-conjugated secondary antibodies at RT for 2 h. After rinsing in 0.1 m PB for 2 h, cells were mounted in Moviol on glass slides.
The transfected brains were harvested and fixed by immersion in 4% PFA at different time points after electroporation. The fixed brains were coronally cut for 50 μm with a vibratome (Leica, VT1000S). Free-floating sections were blocked with 5% normal goat serum in 0.1 m PB for 30 min. The sections were then incubated with the solutions containing primary antibodies: rabbit anti-green fluorescent protein (GFP; 1 : 1000; Invitrogen), mouse anti-Reelin G10 (1 : 1000; Chemicon), rabbit anti-BLBP (1 : 500; Chemicon), mouse anti-Nestin (1 : 1000; Chemicon), and goat anti-doublecortin (DCX; 1 : 500; Santa Cruz Biotechnology) in 0.1 m PB overnight at 4°C. After washing in 0.1 m PB, the sections were incubated with the secondary antibodies: goat anti-rabbit Alexa Fluor 488 or 568, goat anti-mouse Alexa Fluor 568, and donkey anti-goat Alexa Fluor 568 (1 : 300, Molecular Probes) overnight at 4°C. After rinsing in 0.1 m PB for 2 h, the nuclei in some sections were labeled with propidium iodide (PI; 2.5 mg/mL; Invitrogen) or 4′, 6-diamidino-2-phenylindole (DAPI; 1 : 10 000; Roche Molecular Biochemicals). The sections were mounted in Moviol. Images were acquired using a Zeiss LSM 510 confocal microscope.
To analyze the morphological features of GFP-positive cells, optical z-stacks were acquired at 1-μm intervals in sections using a Zeiss LSM 510 confocal microscope with a 63×, 1.4 numerical aperture objective. These z-series were projected into a single plane (flattened) using maximum pixel intensity algorithm by the Zeiss ZEN 2010 software (Carl Zeiss). GFP-positive neurons and RGCs were traced using Neuromantic V1.6.3 software (http://www.rdg.ac.uk/neuromantic). The length of the leading processes and branches of neurons and RGCs were measured using ImageJ software (National Institutes of Health). For dendritic morphology analysis, 2 parameters were analyzed: total length of branches and number of bifurcations. A primary branch was defined as the length of any GFP-labeled structure originating from the leading process that is longer than 2.5 μm. Thus, small unstable filopodia shorter than 2.5 μm protruding from the leading processes of migrating neurons were excluded. The length of the residual leading process is measured from its origin at the soma to the first branching point.
For analysis of the cell culture experiments, the processes of MAP2-positive neurons and BLBP-positive glial cells were quantified with SIS software using images acquired on an epifluorescence microscope with a 60× objective lens (BZ-9000). A total of 200 neurons and 100 glial cells of each condition were measured. Student's t-test was used for statistical analysis. Results were presented as mean ± SD; P < 0.05 was considered to be statistically significant.
Defined Cohorts of Neurons by GFP Labeling using Timed In Utero Electroporation
In the brains of wild-type mice transfected at E14.5 and fixed at E17.5, numerous GFP-positive cells were observed throughout the different layers of the developing cortex (Fig. 1A). The GFP-positive cells in the CP showed the typical bipolar morphology of migrating neurons with a thick leading process pointing toward the pial surface and a thin trailing process directed to the ventricle (Fig. 1A inset). Some GFP-positive cells already reached the upper portion of the CP, and their leading processes approached or invaded the MZ (Fig. 1A). At 5 days after electroporation, the somata of the majority of GFP-positive cells accumulated immediately beneath the MZ by forming a dense layer. They gave rise to primitive stem dendrites invading the MZ (Fig. 1B). At P5, 10 days after transfection, most GFP-positive cells showed the typical morphology of differentiated neurons with an apical dendritic tuft in the MZ (Fig. 1C).
The morphology of these GFP-labeled cells indicated that they were neurons. To prove this possibility, immunostaining for doublecortin (DCX), a microtubule-associated protein being generally recognized as a marker for postmitotic migrating neurons (Francis et al. 1999; Gleeson et al. 1999), was performed in the transfected cortex. Our results showed that the majority of GFP-labeled cells in the CP, fixed at 3 days after electroporation, were DCX-positive (Fig. 1D), indicating that cells in brain sections transfected with pCAG-GFP at E14.5 and fixed 3 days or later were indeed neurons.
RGCs Were GFP-Labeled using In Utero Electroporation
In the brains of wild-type mice transfected at E17.5 and fixed at E18.5, numerous GFP-positive cells with long processes were observed in the cerebral cortex (Fig. 2A). They showed a typical bipolar radial glial morphology, characterized by a long ascending process spanning the whole developing cortical wall up to the pial surface and a short apical process contacting the ventricle and terminating by an end-foot. The oval somata of these GFP-positive cells were located in the VZ (Fig. 2B). Within the IZ and CP, the thin ascending processes of the GFP-labeled cells showed a straight course toward the MZ without branching. In contrast, upon reaching the MZ, the ascending processes of the GFP-labeled cells branched extensively and formed a distinct end-foot at the tip of each branch (Fig. 2B–D).
To confirm the radial glial nature of these GFP-labeled cells, we performed immunocytochemistry to detect the expression of 2 markers for RGCs: BLBP (Feng et al. 1994) and Nestin (Lendahl et al. 1990). Our results showed that the majority of GFP-labeled cells in brains transfected at E17.5 were BLBP-positive (Fig. 2C) and Nestin-positive (Fig. 2D), indicating that they were indeed RGCs.
Reelin Promotes Branching of Neurons and Dendritic Development In Vivo
It has been demonstrated previously that migrating neurons change their migratory behavior from glia-guided locomotion to somal translocation upon contact of the leading process with the CP–MZ boundary. In this way, cortical neurons complete their migration (Nadarajah et al. 2001). To examine Reelin's function in this process, we classified the transfected neurons in wild-type cortex by their positions with respect to the CP and MZ: 1) migrating neurons with both somata and leading processes within the CP (Fig. 3A), 2) migrating neurons with their somata in the upper CP but their leading processes in contact with the Reelin-enriched MZ (Fig. 3B), and 3) neurons with their somata near the CP–MZ boundary (Fig. 3C). Drawings of representative neurons from each group are shown in Figure 3D.
We found significant differences in the branching pattern between neurons whose leading processes were in contact with the Reelin-rich MZ and those with their leading processes restricted to the CP. The leading processes of most neurons crawling through the CP displayed a straight orientation toward the MZ without bifurcation (Fig. 3A), whereas the leading processes of migrating neurons developed V-shaped bifurcations as soon as they contacted the Reelin-rich MZ (Fig. 3B). Dendritic arbors that maintained contact with Reelin became increasingly branched as their parent soma approached the MZ and the residual leading process became shorter. The somata of migrating neurons stopped at the top of the CP without entering the Reelin-rich MZ (Fig. 3C).
To quantify the morphological differences between neurons within the CP and those contacting the MZ, we measured the lengths of leading processes and their apical branches and counted the number of bifurcation points. The length of leading processes restricted to the CP was 65.73 ± 10.88 μm. The length of the residual leading process correlated negatively with the total length of all branches (n = 42 cells for this group, r = −0.700, P < 0.001; Fig. 3E) and the number of bifurcations (n = 39 cells for this group, r = −0.701, P < 0.001; Fig. 3F). In summary, the close spatiotemporal relationship of Reelin in the MZ and the initiation of branching suggest that it is Reelin that induces branching of the leading processes, thus contributing to the early development of the apical dendritic arbor.
Reelin Induces Branching of Leading Processes of Migrating Neurons in Slice Culture
As a next step, we wanted to visualize the Reelin-induced branching dynamics using real-time microscopy. For this purpose, slice co-cultures were performed, an experimental model in which the migration defect of granule cells in the reeler DG was rescued by Reelin from a co-cultured wild-type hippocampus (Zhao et al. 2004). To visualize the granule cells in the reeler DG in the living slice, reeler mice were crossbred with POMC-EGFP transgenic mice, in which EGFP is specifically expressed in dentate granule cells under the POMC promoter. The EGFP-positive reeler hippocampus (P0) was then co-cultured with P7 wild-type hippocampus, which provided a Reelin-rich MZ. Time-lapse imaging was performed for more than 12 h. With this approach, we were able to monitor the migration process and branching of the migrating neurons. Neurons that did not yet contact the Reelin-containing zone showed their typical bipolar morphology, and their leading processes did not give rise to branches (Fig. 4A). However, the leading process started to branch as soon as its growth cone came in contact with the Reelin-containing zone. The more the soma moved toward the Reelin-containing zone, the more branches were formed in that zone, accompanied by shortening of the leading process (Supplementary Video 1, Fig. 4A). The migration was terminated as the soma reached the border of the Reelin-containing zone and the leading process disappeared. As a result, only its branches remained in the Reelin-containing zone (Fig. 4A–C). These results indicate that Reelin induces the dynamic branching of the leading process during somal translocation, the last step of the migratory process. We have shown previously that the effects were mediated via canonical Reelin signaling involving ApoER2, VLDLR, and the adapter molecule Dab1 (Zhao et al. 2006).
Absence of Reelin Results in Severe Migration Defects and Morphological Abnormalities of Neurons
It is well known that in reeler mice deficient in Reelin neuronal migration is severely altered and cortical layers are nearly inverted. In addition, dendrites and axons of the neurons are misoriented. However, the time point at which these abnormalities occur has remained unclear. To address this issue, we compared the positioning and morphology of GFP-labeled neurons in reeler mice with those in wild-type littermates at 2 critical time points, i.e., 3 and 5 days after electroporation at E14.5, respectively. In the wild-type cortex fixed at E17.5, GFP-labeled neurons were found in all layers of the cerebral cortex (Fig. 1A), most of them migrating through the different layers of CP. Some neurons had already reached their final destination, future layer 2/3 of the cerebral cortex directly underneath the MZ. In the reeler cortex fixed at E17.5, we similarly observed numerous GFP-labeled neurons traversing the CP (Fig. 5A), displaying the typical bipolar morphology as known from migrating neurons in the wild-type cortex (Fig. 5A,B). However, the vast majority of GFP-labeled neurons were found in the deep portion of the CP. No GFP-labeled neurons were observed that reached the border between CP and MZ (Fig. 5A). These observations suggest that the transformation of migrating neurons from the multipolar to the bipolar stage in the intermediate zone (IZ) and the migration from IZ to CP were not affected in reeler. However, the migration within CP appeared disrupted. One possibility might be that neurons in the reeler CP migrated slower than their counterparts in the wild-type cortex. Alternatively, neurons in reeler cortex were unable to migrate further toward superficial layers. To examine these possibilities, we fixed brains at P0, i.e., 5 days after electroporation. At this time point, most of the GFP-labeled neurons in the wild-type cortex had reached their final destination, forming a densely packed cell layer underneath the MZ (Fig. 1B). In contrast, in the reeler cortex, most GFP-labeled neurons were loosely scattered throughout the deep portion of the CP, far from their definitive destination, the upper cortical layers. At this time point, we also observed severe morphological abnormalities of the neurons in the reeler cortex transfected at E14.5 when compared with wild-type mice. GFP-labeled neurons in the wild-type cortex were characterized by their bipolarity and preferential polarity orientation. Thus, they give rise to a short, thick apical dendrite, the former leading process that forms an apical tuft in layer 1, the former MZ containing Reelin (cf. Figs 1B,C and 5E). The thin axon (the former trailing process) emerges from the opposite side of the cell body (Figs 1C and 5E). No basal dendrites were observed at this stage. In the reeler cortex, many neurons have lost their bipolarity and preferential orientation, and many apical dendrites are oriented toward the white matter (Fig. 5C,D,F). Moreover, the axons of many neurons in the reeler cortex do not originate from the soma, but from one of the dendrites. The above-mentioned results suggest that Reelin does not only play a crucial role in the migratory process but is involved in neuronal polarization, polarity orientation, and axon formation.
Reelin Induces Branching of RGC Processes In Vivo
To examine whether Reelin also induces branching of RGCs, we performed in utero electroporation with pCAG-GFP in embryonic wild-type mice and reeler mutants at E17.5 and fixed the tissue at E18.5. The sections were double stained with antibodies against Reelin and GFP. We found that in wild-type mice, RGCs projected long and thin processes to the pial surface, spanning the entire developing CP. The segments of radial glial fibers that were located in the IZ and CP were straight with characteristic varicosities, whereas those contacting the Reelin-rich MZ gave rise to extensive branches (Figs 2 and 6A,C). Each branch formed an end-foot that attached these processes to the pial surface (Figs 2B–D and 6A). In reeler mice, radial glial fibers also extended into the MZ but rarely exhibited branches (Fig. 6B). Quantitative analysis showed that each wild-type radial glial fiber gave rise to various branches in the MZ ranging from 6 to 12. In contrast, the number decreased significantly to 1.85 ± 0.96 in reeler mice (Fig. 6E). Total branch length, total number of branches and average branch length of glial fibers were dramatically reduced in reeler mutants compared with wild-type mice (Fig. 6D–F). Furthermore, branches of radial glial fibers in the MZ mainly ran perpendicular to the pial surface, whereas their course was more irregular in reeler. Notably, the first branch points in wild-type mice were close to the CP–MZ boundary, whereas the few branch points in reeler were often found within the MZ (Fig. 6G).
The observed disruption of the radial glial scaffold in Reelin-deficient reeler mice indicates that Reelin is indispensable for the branching of basal processes of RGCs. The total length of branches of a wild-type RGC was 402.69 ± 131.07 μm, i.e., approximately 6 times the width of the MZ.
Reelin Promotes Dendritic Arborization and Glial Process Outgrowth In Vitro
To determine whether Reelin directly promotes dendritic development independent of the neuron's integration into its normal environment and developing neural circuitry, we established primary neural cultures of neocortex from E17.5 reeler mice and treated them with recombinant Reelin for 24 h. To analyze dendritic development and glial process outgrowth separately, antibodies against MAP2, a marker for neurons, and BLBP, a specific marker for glial cells in the embryonic neocortex, were used in this study.
Treatment of recombinant Reelin significantly induced process formation, elongation, and branching of both neurons and glial cells (Fig. 7A–D). After treatment, the number of terminal tips of neuronal and glial processes, an index of process formation and branching, increased from 2.1 ± 0.6 to 3.4 ± 1.0 (P < 0.001) and from 1.8 ± 0.4 to 2.1 ± 0.6 (P < 0.05), respectively (Fig. 7E). Average length of processes increased from 7.4 ± 2.5 to 13.7 ± 2.8 μm in neurons (P < 0.001) and from 14.8 ± 4.2 to 18.7 ± 6.0 μm in glial cells (P < 0.001; Fig. 7F). The total process length per neuron increased from 20.3 ± 6.5 to 45.0 ± 14.1 μm (P < 0.001), whereas the total process length per glial cell increased from 27.5 ± 8.2 to 38.2 ± 16.1 μm (P < 0.05; Fig. 7G). These results indicate that Reelin promotes both dendritic complexity and glial process outgrowth.
During development, postmitotic neurons migrate along radial glial fibers from the VZ toward the MZ, forming the well-organized 6-layered structure of the mammalian cerebral cortex in an “inside-out” manner (Rakic 1972; Nadarajah and Parnavelas 2002). Reelin, synthesized and secreted by CR cells, is located in the MZ of the developing cerebral cortex as revealed by in situ hybridization and immunocytochemistry for Reelin (D'Arcangelo et al. 1995, 1997; Ogawa et al. 1995; de Bergeyck et al. 1997; Schiffmann et al. 1997; Drakew et al. 1998). Previous studies demonstrated that Reelin is required for correct positioning of neurons in the developing neocortex (Rakic and Caviness 1995). However, the precise mechanism of Reelin's action on neuronal migration and layer formation has remained unclear. By combining in utero electroporation with immunocytochemistry against Reelin and by studying migrating neurons using real-time microscopy, we provide evidence that Reelin promotes the branching of neurons and RGCs in vivo, which might contribute to the completion of the migratory process and correct positioning of radially migrating cortical neurons.
In Utero Electroporation Labels Individual Neural Cells at Defined Points in Time
In utero electroporation is a simple and powerful method, which has proven to efficiently introduce genes into cells of the embryonic cortex (Saito and Nakatsuji 2001; Tabata and Nakajima 2001). During the past decade, it has been successfully applied to study the proliferation, migration, and differentiation of neurons (Tabata and Nakajima 2002; Olson et al. 2006; Chen et al. 2008; Kubo et al. 2010; Jossin and Cooper 2011). Thus, this technique enables the investigator to visualize the morphology of individual neurons and RGCs in different developmental stages. In previous studies, DiI labeling was used to trace RGCs (Miyata et al. 2001; Noctor et al. 2002; Hartfuss et al. 2003). However, passive lateral diffusion of the tracer did not result in the complete staining of cell types with branches, and staining with this lipophilic dye is incompatible with immunocytochemical techniques employing Triton X-100 treatment (Matsubayashi et al. 2008; Staffend and Meisel 2011). Various antibodies against specific markers expressed in different developmental phases were also used to label diverse cohorts of neurons or to identify distinct neural cell types. However, nearly all the cells of the selected population are immunolabeled, which makes it difficult to trace individual cells with all their processes. Therefore, in utero electroporation appears to be an appropriate method to trace individual neurons and RGCs.
Different Neural Cells Were Labeled at Different Embryonic Stages
In the present study, we transfected embryonic brains with pCAG-GFP at distinct embryonic days using in utero electroporation. Different experimental results were obtained when the transfected brains were harvested after different survival times. When the electroporation was performed at E14.5 and the transfected brains were fixed at E17.5, numerous migrating bipolar neurons with a long, thick leading process and a short, thin trailing process were labeled by GFP throughout the entire CP, whereas in brains fixed 5 days after transfection, the majority of the somata of GFP-labeled cells accumulated immediately beneath the MZ. This indicates that newly born neurons require 4–5 days to complete their migration. When we fixed E14.5-transfected brains at P5, the majority of the neurons were differentiated with abundant branches. When electroporation was performed at embryonic stage E17.5 and the transfected brains were fixed 24 h later, the majority of GFP-positive cells were identified as RGCs, many of them exhibiting the typical spindle-shaped soma, a short apical process contacting the ventricular surface, and a long ascending basal process spanning the whole developing cortical wall. Thus, in utero electroporation allowed us to study defined types of cells as well as different stages of their migration and differentiation during the development of the cerebral cortex.
Reelin Induces Branching of Migrating Neurons and RGCs
Previous studies on the role of Reelin during embryonic and early postnatal development have focused on radial neuronal migration and the formation of the laminated cortical structure (Zhao et al. 2004; Hack et al. 2007; Frotscher 2010; Zhao and Frotscher 2010; Jossin and Cooper 2011). Here, we focused on Reelin's role in process differentiation of late-generated cortical neurons and RGCs. In agreement with previous reports, we found that addition of recombinant Reelin to the culture medium promoted dendritic extension and branching of reeler cortical neurons and glial cells in vitro (see Förster et al. 2002; Niu et al. 2004; Niu et al. 2008).
The thin radial glial fibers follow a straight course traversing the IZ and CP and then branch extensively as soon as they reach the Reelin-rich MZ. They terminate with bulbous swellings, the characteristic radial glial end-feet. A similar branching pattern was found for the leading processes of neurons. Our results raise the possibility that the Reelin-induced bifurcation of the leading process, which results in two branches thinner than the main leading process, forms a mechanical obstacle, preventing further migration by nuclear translocation. Along this line, branching of the radial fiber may similarly be involved in ceasing the migration process, probably facilitating detachment of the neuron from the radial glial fiber. At least, this hypothesis would be compatible with the results in reeler mice where the branching of leading processes and radial glial fibers is severely altered and many neurons “overmigrate” and populate the MZ, a cell-poor layer in wild-type animals. This interpretation is also concordant with previous studies in which Reelin was found to induce the detachment of migrating neurons from their radial glial substrate (Dulabon et al. 2000; Hack et al. 2002; Sanada et al. 2004).
Our real-time microscopy studies have shown that the assumed sequence of events, i.e., branching of the leading process upon contact with the MZ followed by nuclear translocation associated with the shortening of the leading process, does take place in one and the same migrating neuron as it approaches its definitive destination. In future real-time microscopy studies, migrating neurons and radial glial fibers might be labeled by different fluorescent proteins in order to get more insight into the interactions of both cell types in the terminal phase of the migratory process. The paucity of branches of leading processes and radial glial fibers and their misorientation in reeler mutants strongly suggests that Reelin not only induces branching—as was shown before (Niu et al. 2004; Olson et al. 2006) and confirmed by the present culture experiments—but is also important for anchoring the leading processes to the MZ by these branches. Attachment to the MZ, in turn, might create the tension of the leading process that is required for nuclear translocation to occur. Finally, it may not only be ramification of the leading process in the MZ that attaches it to the cortical surface. Another likely presumption is the direct involvement of Reelin in morphological alterations of the migrating neurons themselves. Chai et al. (2009) have previously shown that Reelin phosphorylates the actin-depolymerizing protein cofilin. Phosphorylation of cofilin renders it unable to depolymerize actin, thereby stabilizing the actin cytoskeleton (Arber et al. 1998; Yang et al. 1998). Consisting with this function of Reelin, p-cofilin immunostaining was strongest in the leading processes approaching the Reelin-rich MZ (Chai et al. 2009). Immunostaining for p-cofilin was dramatically reduced in reeler tissue, likely resulting in an altered cytoskeletal stability of the leading process. This in turn would be compatible with the altered orientation of many apical dendrites, the former leading processes, in reeler. In the wild-type cortex, apical dendrites are regularly directed toward the pial surface, resulting from the Reelin-induced branching of the leading process and its stabilization via cofilin phosphorylation.
In conclusion then, our data provide evidence for a role of Reelin in inducing the branching of migrating neurons and RGCs. Thus, Reelin may act as a “stop signal” in two ways, on the one hand by inducing branching, on the other by phosphorylating cofilin, which stabilizes the actin cytoskeleton thus interfering with the flow of actin filaments required for the translocation of the nucleus during migration (Chai et al. 2009, 2014; He et al. 2010). Finally, we would like to point out that Reelin's effect on the branching of leading processes of neurons and the processes of RGCs should not be generalized. Other guiding molecules and mechanisms are involved in the formation of axonal projections, axonal collateralization, and targeting, as for instance in the development of thalamocortical projections (Molnár et al. 1998; Higashi et al. 2005).
This work was supported by grants from the Deutsche Forschungsgemeinschaft (FR 620/12-1 to M.F.) and the National Natural Science Foundation of China (No. 31071873 to S.Z.).
Michael Frotscher is Senior Research Professor of the Hertie Foundation. The authors are grateful to Ms. Ludmila Butenko for her excellent technical assistance and Ms. Pingjing Guo for genotyping the mutants. The authors thank the members of the Core Facility Imaging (UMIF) for their help with real-time microscopy. Conflict of Interest: None declared.