Neuronal migration contributes to the establishment of mammalian brain. The extracellular protein Reelin sends signals to various downstream molecules by binding to its receptors, the apolipoprotein E receptor 2 (ApoER2) and very low-density lipoprotein receptor and exerts essential roles in the neuronal migration and formation of the layered neocortex. However, the cellular and molecular functions of Reelin signaling in the cortical development are not yet fully understood. Here, to gain insight into the role of Reelin signaling during cortical development, we examined the migratory behavior of Apoer2-deficient neurons in the developing brain. Stage-specific labeling of newborn neurons revealed that the neurons ectopically invaded the marginal zone (MZ) and that neuronal migration of both early- and late-born neurons was disrupted in the intermediate zone (IZ) in the Apoer2 KO mice. Rescue experiments showed that ApoER2 functions both in cell-autonomous and noncell-autonomous manners, that Rap1, integrin, and Akt are involved in the termination of migration beneath the MZ, and that Akt also controls neuronal migration in the IZ downstream of ApoER2. These data indicate that ApoER2 controls multiple processes in neuronal migration, including the early stage of radial migration and termination of migration beneath the MZ in the developing neocortex.
Neuronal migration plays essential roles in the establishment of the mammalian cerebral cortex. In embryonic brains, excitatory neurons generated in the ventricular zone (VZ) or the subventricular zone (SVZ) migrate radially toward the brain surface in a birth-date-dependent “inside–out” manner, to form the cortical plate (CP) (Ayala et al. 2007; Cooper 2008; Rakic 2009). In this process, neurons born in the early and late developmental stages, the future deep- and superficial-layer neurons, respectively, exhibit distinct migration patterns to reach the top of the CP. Early-born neurons mainly use the somal translocation mode, which is thought to be independent of the radial glial fibers (Nadarajah et al. 2001). Late-born neurons undergo complex changes in their migratory modes. After their final cell division, they transform into multipolar cells, exhibiting multipolar migration just above the VZ or in the multipolar cell accumulation zone (MAZ) for about 24 h (Tabata and Nakajima 2003; Tabata et al. 2009). They then transform to a bipolar morphology suitable for directional migration along the radial glial fibers (locomotion mode) (Rakic 1972; Nadarajah et al. 2001). Finally, when their leading processes reach the most superficial layer of the cortex, the marginal zone (MZ), the locomoting neurons switch to the terminal translocation mode, in which their somas move quickly toward the surface of the CP in a radial glia–independent manner (Nadarajah et al. 2001) to complete their migration just beneath the MZ.
Reelin is an extracellular protein secreted mainly from the Cajal–Retzius (C–R) cells in the MZ during cortical development (Bar et al. 1995; D'Arcangelo et al. 1995; Ogawa et al. 1995; Nakajima et al. 1997). Reelin is one of the major signals required for the formation of the neuronal layers in the mammalian central nervous system (CNS). Depletion of Reelin results in neuronal migration defects in several brain structures, including the cerebral cortex, cerebellum, hippocampus, midbrain, and olfactory bulb (Curran and D'Arcangelo 1998; Frotscher 1998; Tissir and Goffinet 2003; Katsuyama and Terashima 2009), and the spinal cord (Yip et al. 2000). Reelin exerts its functions via binding to 2 main lipoprotein receptors, apolipoprotein E receptor 2 (ApoER2) and very low-density lipoprotein receptor (VLDLR) (D'Arcangelo et al. 1999; Hiesberger et al. 1999; Trommsdorff et al. 1999), and then phosphorylates intracellular adaptor protein disabled homolog 1 (Dab1), which then interacts with various downstream molecules (Honda et al. 2011; Sekine et al. 2014). Although numerous components of the Reelin signaling pathway have been discovered, the roles of Reelin signaling in cortical development at the cellular and molecular levels are not yet fully understood.
Previous studies investigating the functions of the Reelin receptors during embryonic development using mutant mouse lines demonstrated that ApoER2 is required for the migration of late-generated neurons and for layer formation by these neurons (Trommsdorff et al. 1999; Benhayon et al. 2003; Hack et al. 2007). However, how ApoER2 is involved in the cellular behavior remains unknown. In the present study, we visualized the behavior of each migrating neuron in the Apoer2 KO cortex using a stage-specific in utero electroporation (IUE) technique (Tabata and Nakajima 2001) and found that ApoER2 was essential for both neuronal migration and the appropriate termination of the migrating neurons in the developing cortex.
Materials and Methods
The method used to maintain the colony of Apoer2 KO mice has been previously described (Hirota et al. 2015). The day of vaginal plug detection was considered to be embryonic day 0 (E0). All animal experiments were performed under the control of the Keio University Institutional Animal Care and Use Committee in accordance with the Institutional Guidelines on Animal Experimentation at Keio University.
The embryos and neonates were placed on ice for anesthesia and perfusion-fixed with 4% paraformaldehyde (PFA) in a 0.1-M sodium phosphate buffer (pH 7.4). Brains were postfixed in the same fixative overnight at 4°C, replaced in a 30% sucrose solution with phosphate buffered saline (PBS), embedded in OCT compound (Sakura), and frozen in liquid nitrogen. The frozen sections were then cut with a cryostat (CM3050 S; Leica) into 20-μm-thick sections. For immunostaining, after being rinsed in PBS, the sections were incubated for 30 min in blocking solution (10% donkey or goat serum and 0.1% Triton X-100 in PBS), then overnight with the primary antibodies, and then for 2 h at room temperature with species-specific secondary antibodies (Invitrogen or Jackson ImmunoResearch Laboratories). The following primary antibodies were used in this study: Cux1 (rabbit, 1/200, Santa Cruz, cat# sc-13024), Tbr1 (rabbit, 1:1000, gift from R. Hevner), Reelin (goat, 1:200, R&D Systems, cat# AF3820), green fluorescent protein (GFP; chick, 1:2000, abcam, cat# ab13970), MAP2 (mouse, 1:400, Santa cruz, cat# sc32791), chondroitin sulfate proteoglycan (CSPG; mouse, 1:200, Sigma, cat# 8035), NeuN (mouse, 1:300, Millipore, cat# MAB377), Ctip2 (rat, 1:500, abcam, cat# ab18465), RORβ (mouse, 1:400, Perseus Proteomics, cat# PP-N7927-00), nestin (mouse, 1:300, R&D, cat# MAB2736), Foxp2 (rabbit, 1:400, abcam, cat# ab16046), chondroitin sulfate (2B6) (mouse, 1:20, cosmo bio, cat# PRPG-BC-M02). For control staining for Tbr1, normal rabbit IgG (Chrompure Rabbit IgG, whole molecule, Jackson ImmunoResearch Laboratories, cat# 011-000-003) was used. For nuclear staining, DAPI (4,6-di-amidino-2-phenylindole, dihydrochloride; Invitrogen) was used. Confocal images were obtained with the FV1000 (Olympus) microscopes. The pictures were processed using the Adobe Photoshop software (Adobe Systems).
pCAGGS-EGFP has been described previously (Niwa et al. 1991). We purchased a synthesized gene encoding modified chondroitinase ABC (ChABC) Y1330 (Zhao et al. 2011) from FASMAC. Y1330 was then cloned into pCAGGS to generate pCAGGS-ChABC. The Tα1 expression vector has been described previously (Sekine et al. 2012). To generate pTα1-ApoER2-HA, the ApoER2-HA from pCAGGS1-ApoER2-HA (Hirota et al. 2015) was cloned into the Tα1 vector. To generate CAG-ApoER2ΔC, mouse ApoER2ΔC lacking the cytoplasmic domain by deletion of residues 764–870, was amplified by PCR and cloned into pCAGGS. The Tα1-Cre vector was obtained as a kind gift from Drs Sakakibara and Miyata. To generate the pTα1-Rap1-63E, a constitutively active mutant of Rap1, Rap1-63E (Kitayama et al. 1990), was cloned into the Tα1 vector. pCALNL-Akt and pCALNL-integrin α5-GFFKA have been described previously (Sekine et al. 2012). pCAGGS-N-cadherin have also been described previously (Kawauchi et al. 2010), as has pCAGGS-Reelin (Kubo et al. 2010).
In Utero Electroporation
Pregnant mice were deeply anesthetized and their intrauterine embryos were surgically manipulated as described previously (Nakajima et al. 1997). In utero electroporation to transfect vectors into the embryonic neocortex and hippocampus was performed as described previously (Tabata and Nakajima 2001, 2008; Tomita et al. 2011; Kitazawa et al. 2014). Approximately 1–2 μL of the plasmid solution (1.5 mg/mL CAG-driven enhanced green fluorescent protein expression vector, CAG-EGFP) (Niwa et al. 1991) containing 0.01% fast green solution was injected into the lateral ventricle of the intrauterine embryos; electronic pulses (31 V, 50 ms, 950-ms intervals, 4 times) were then applied using an electroporator (CUY-21SC; NEPA GENE). The concentrations of the plasmids were as follows. For the ChABC expression experiment (Fig. 5W,X): (1) 1.5 mg/mL of pCAGGS-EGFP; (2) 1.5 mg/mL of pCAGGS-EGFP and 2.0 mg/mL of pCAGGS-ChABC. For the ApoER2 expression experiment (Fig. 6A–C): (1) 1.5 mg/mL of pCAGGS-EGFP; (2) 1.5 mg/mL of pCAGGS-EGFP and 2.5 mg/mL of pTα1-ApoER2-HA. For the ApoER2ΔC expression experiment (Fig. 6D,E): (1) 1.5 mg/mL of pCAGGS-EGFP and 2.5 mg/m pCAGGS; (2) 1.5 mg/mL of pCAGGS-EGFP and 2.5 mg/m pCAGGS-ApoER2ΔC. For the Apoer2 KO rescue experiments (Fig. 7): (1) 1.5 mg/mL of pCAGGS-EGFP; (2) 1.5 mg/mL of pCAGGS-EGFP and 2.5 mg/mL pTα1-Rap1-63E; (3) 1.0 mg/mL of pCAGGS-EGFP, 2.0 mg/mL Tα1-Rap1 63E, 2.0 mg/mL pCALNL-Akt, 1.5 mg/mL pCALNL-Akt, and 1.0 mg/mL Tα1-Cre; (4) 1.0 mg/mL of pCAGGS-EGFP and 1.0 mg/mL of pCGGS-N-cadherin; (5) 0.25 mg/mL of pCALNL-EGFP, 2.0 mg/mL pCALNL-integrin α5-GFFKA, and 0.75 mg/mL Tα1-Cre; (6) 0.25 mg/mL of pCALNL-EGFP, 2.0 mg/mL pCALNL-integrin α5-GFFKA, 2.0 mg/mL pCALNL-Akt, and 0.75 mg/mL Tα1-Cre; (7) 0.25 mg/mL of pCALNL-EGFP, 2.0 mg/mL pCALNL-Akt, and 0.75 mg/mL Tα1-Cre. For Reelin overexpression experiment (Fig. 8): 1.5 mg/mL of pCAGGS-EGFP and 6.0 mg/mL of pCAGGS-Reelin.
The distributions of the GFP-labeled cells were analyzed on coronal sections at the level of the dorsal recess of the third ventricle of the embryonic brain (E15.5, E16.5, and E18.5), or at the level of the caudal part of the somatosensory cortex (caudal to the rostral end of hippocampus) at P0. For the neocortex, the distance from the top of the CP to the bottom of the IZ was divided into 10 bins. For the hippocampus, the distance from the top of the stratum pyramidale (SP) to the ventricle was divided into 5 bins. The distances from the top of the CP (neocortex) or SP (hippocampus) to the nuclei of the migrating cells were measured using the ImageJ software (NIH). All data were expressed as the mean ± standard error of mean (SEM). For direct comparisons, the data were analyzed by unpaired two-tailed Student's t-test. For multiple comparisons, the data were analyzed by one-way ANOVA with Tukey's post hoc test. A P-value of <0.05 was considered significant.
Cortical Neurons Overmigrate Into the MZ in the Apoer2 KO Mice
Our recent observation that ApoER2 is expressed in the MZ (Hirota et al. 2015) of the developing cortex led us to examine whether neuronal migration beneath the MZ might be affected by Apoer2 deletion. Both immunostaining for Cux1 (a marker of superficial-layer neurons) (Nieto et al. 2004) and nuclear staining showed an increased number of neurons in the MZ of the neonatal Apoer2 mutants as compared with that in the control mice (Fig. 1A–J). To further examine the cell type of the ectopic cells in the MZ of the Apoer2 KO mice, we performed double staining for Reelin (expressed in the C–R cells) (D'Arcangelo et al. 1995; Ogawa et al. 1995) and Tbr1 (strongly expressed in the C–R cells and deep-layer neurons and weakly expressed in the other cortical layers and in the IZ at P0) (Hevner et al. 2001, 2003) on P0. Examination revealed no differences in the number of strongly double-positive C–R cells for Reelin and Tbr1 in the Apoer2 KO mice (Fig. 1K–Q), suggesting that the C–R cells were not affected. Strongly Tbr1-positive, but Reelin-negative neurons were almost absent from the MZ in the control animals, but were detected in the Apoer2 mutants (Fig. 1K–P,R). These results suggest that both superficial and deep-layer neurons were ectopically localized in the MZ on P0 in the Apoer2 mutants. Since the number of cells in the MZ was not increased in the later postnatal stages in the Apoer2 KO mice (Hack et al. 2007) (data not shown), the ectopic cells might be eliminated from the MZ by cell death.
Next, to examine whether the ectopic cells in the MZ were caused by invasion of the MZ by the CP neurons, the migrating neurons were visualized using IUE. IUE was performed at E12.5 and E14.5 to label the early-born and relatively late-born neurons, respectively, and the mouse brains were fixed 6 or 5 days later. In both cases, very few GFP-labeled neurons were found in the MZ of the control animals, whereas the number of GFP-labeled cells in the MZ was significantly increased in the Apoer2 mutants (Fig. 2A–F). These results suggest that both early-born and late-born neurons overmigrate into the MZ in Apoer2 mutants.
Migration of Early-Born CP Neurons is Affected in the Apoer2 KO Mice
The Apoer2 mutant mice showed an abnormal laminar phenotype of the neocortex, in which early-born superficial-layer neurons were divided into 2 bands, and the late-born deep-layer neurons were shifted to a superficial position (Trommsdorff et al. 1999; Hack et al. 2007). However, it remains unclear whether the abnormal positioning of early-born neurons in the Apoer2 mutants is secondary to the disturbed migration of late-born neurons or whether ApoER2 is required for radial migration of early-born neurons in the IZ in a cell-autonomous manner. To examine the effect of Apoer2 deletion in the neurons generated in the early developmental stages, we labeled early-born neurons by IUE at E12.5 and analyzed the distribution of the labeled cells at different stages. Observation conducted 2 days later showed that the distribution of the labeled cells was largely unchanged in the Apoer2 mutants as compared with that in the control mice (Fig. 3A,B). Examination conducted at E15.5 revealed that the majority of the labeled control cells had entered the CP. In contrast, in the Apoer2 mutants, a large number of the labeled cells remained in the IZ (Fig. 3C,D). MAP2 staining confirmed that in the Apoer2 mutants, neuronal migration was affected before the cells entered the CP (Fig. 3G,H). At E18.5, almost all of the control labeled cells were located in the superficial half of the cortical wall, whereas in the Apoer2 mutants, the majority of the cells had shifted to the deeper part of the cortical wall (Fig. 3E,F). Quantification confirmed that neuronal migration was significantly affected at E15.5 and E18.5 by Apoer2 deletion (Fig. 3I,J). Furthermore, the CP thickness was reduced in the Apoer2 mutants at E15.5, but not at E14.5, suggesting that the number of neurons entering the CP between these stages was reduced in the Apoer2 mutants (Fig. 3A–D,K). These results suggest that ApoER2 is required for the migration of early-born neurons.
Defective Migration of Late-Born CP Neurons in the Apoer2 KO Mice
A previous study using BrdU labeling showed that ApoER2 was required for the migration of late-born neurons to the superficial layers of the cortex (Hack et al. 2007). We therefore further examined the migration profile of the late-born neurons using IUE at E14.5. Migration was largely normal when the examination was conducted 2 days after the IUE (Fig. 4A,B,I). However, examination on day 3 after the IUE showed migration defect in the Apoer2 KO mice. In the control mice, some GFP+ cells had already entered the CP, whereas in the Apoer2 KO mice, almost all the labeled cells remained in the deeper part of the cortical wall (Fig. 4C,D). On day 4 or 5 after the IUE, the number of labeled cells reaching the superficial part of the CP in the Apoer2 KO mice was less than that in the control mice, and a significant number of cells remained in the deep part of the cortical wall in Apoer2 KO mice (Fig. 4E–H,J), consistent with the previous observation that injection of BrdU into the Apoer2 mutant at E14.5 labeled 2 separate bands of superficially and deeply located neurons on P0 (Hack et al. 2007). Both the GFP+ neurons that had invaded the MZ and those that were stacked in the IZ expressed Cux1, suggesting that these neurons include a common population (Hack et al. 2007). These results indicate that ApoER2 controls the migration of both early-born and late-born neurons in the IZ.
Radially Migrating Neurons Tend to Stop Along the MAP2+/CSPG+ Structures in the Apoer2 KO Mice
We previously showed that in the reeler cortex, neuronal migration tended to be prevented by internal plexiform zones (IPZs) (Tabata and Nakajima 2002), which refer to irregular structures consisting mainly of dendrites (Pinto Lord and Caviness 1979; Tabata and Nakajima 2002). To examine whether neuronal migration might be inhibited by similar structures in the Apoer2 mutants, we immunostained the Apoer2 mutant cortex on P0 with anti-MAP2 and anti-CSPG antibodies, in order to visualize the IPZ; we found that there were cell-sparse regions showing positive staining for both CSPG and MAP2 at the border between the CP and IZ in the Apoer2 mutants (Fig. 5A,B). Then, we examined whether the migrating neurons might be affected by these MAP2+/CSPG+ structures observed in the sections from the P0 mice that had already been electroporated with GFP at E14.5. In the control sections, the majority of the labeled neurons had reached just beneath the MZ by this stage, whereas in the Apoer2 mutant cortex, many labeled neurons stopped migration and tended to align beneath the MAP2+/CSPG+ structures (Fig. 5C–N). These observations suggest that neuronal migration was disturbed, at least in part, by the MAP2+/CSPG+ structures in the deep part of the cortical wall in the Apoer2 mutant mice. Indeed, when we examined the time-point at which the CSPG+ structures emerge in the Apoer2 KO mice, we found that while the CSPG+ staining pattern was indistinguishable between the control and Apoer2 KO mice at E15.5, at E16.5, the CSPG+ region was clearly larger in the Apoer2 KO cortex as compared with the control cortex (Fig. 5O–V). Thus, the MAP2+/CSPG+ structures emerge in relation to the time-point at which migration defect occurs in the Apoer2 KO mice.
Since CSPG is known to function as a repulsive guidance cue for axons and neurons during normal development and after trauma in the adult (Miller and Hsieh-Wilson 2015), it is possible that CSPG itself could inhibit neuronal migration in the Apoer2 KO mice. To address this possibility, we examined whether expression of an optimized chondroitinase ABC (ChABC), a bacterial enzyme with modifications for efficient translation and folding in mouse cells which degrades CSPGs by cleaving glycosaminoglycan chains (Zhao et al. 2011), in migrating neurons would enable the cells to pass through the CSPG-positive region in the Apoer2 KO cortex. Expression of ChABC successively eliminated the CSPG-positive signals with emergence of the digested CSPGs (5W,X), suggesting that CSPG itself does not have a major role in arresting neuronal migration.
Cell-Autonomous Function of ApoER2 in Migratory Neurons
Next, to examine whether ApoER2 controls the migratory behavior in a cell-autonomous manner, we performed a rescue experiment using an ApoER2-expressing vector. Introduction of ApoER2 into migrating neurons in the Apoer2 KO cortex at E14.5 resulted in a significant decrease in the number of neurons overmigrating into the MZ at E18.5 (Fig. 6A). ApoER2 expression also partially rescued the migration defect in the IZ (Fig. 6B,C), while transfection of the ApoER2-expressing vector at the same concentrations as those used in these experiments had no effect in the control mice (Fig. 6A–C).
It was previously reported that a VLDLR deletion mutant lacking the cytoplasmic domain exerted a dominant-negative effect (Jossin and Cooper 2011). Similarly, we prepared a mutant form of ApoER2 (ApoER2ΔC) by deleting its cytoplasmic domain, and found that expression of this construct impaired neuronal radial migration in the IZ (Fig. 6D,E). These results suggest that ApoER2 exerts its functions, at least to some extent, in a cell-autonomous manner.
Rap1, Integrin, and Akt are Involved in the Termination of Migration Beneath the MZ, and Akt Also Controls Neuronal Migration in the IZ Downstream of ApoER2
Previous reports have shown that Reelin controls neuronal migration in the most superficial part of the CP via the Rap1/N-Cadherin and Crk/CrkL-C3G-Rap1-integrin α5β1 pathways (Franco et al. 2011; Sekine et al. 2012). Thus, we tested whether activation of these pathways could rescue the defect of termination of neuronal migration beneath the MZ in the Apoer2 KO cortex. Although the expression of constitutively active forms of Rap1 (CA-Rap1; Rap1-63E), integrin α5 (CA-integrin α5; integrin α5-GFFKA), or wild-type N-cadherin alone did not clearly rescue the overmigration of neurons into the MZ, expression of CA-Rap1 or CA-integrin α5 together with wild-type Akt, which also functions downstream of the Reelin signal (Chai et al. 2009; Feng and Cooper 2009; Jossin and Cooper 2011), partially rescued the phenotype (Fig. 7A). These results suggest that Rap1 and integrin α5 control termination of neuronal migration beneath the MZ cooperatively with Akt downstream of ApoER2.
A recent article showed that Akt regulated radial neuronal migration during neocortical development in a microtubule-mediated manner (Itoh et al. 2016). We found that expression of Akt significantly restored the migration defect in the Apoer2 KO cortex (Fig. 7B,C). At the concentration used in this study, expression of the Akt-expressing vector had no effect on the radial migration in the control mice (Fig. 7B,C). These results suggest that Akt regulates radial migration downstream of the Reelin signal.
ApoER2 is Required for the Neuronal Aggregation Caused by Ectopically Expressed Reelin In Vivo
We previously showed that ectopically expressed Reelin in vivo causes the migrating neurons to form cellular aggregates, wherein neurons stop migrating just along the ectopically expressed Reelin, recapitulating the birthdate-dependent inside–out pattern of neuronal alignment (Kubo et al. 2010). These ectopic aggregates resembled the structures around the MZ, including the central MZ-like cell body-sparse region surrounded by densely packed neuronal cell bodies. To examine whether deletion of Apoer2 would affect the aggregate formation, plasmids expressing Reelin together with GFP were electroporated into the cortex of the Apoer2 KO mice at E14.5 and the brains were analyzed at P1.5. The findings revealed that the ectopically expressed Reelin caused cellular aggregation in the IZ of control littermates, but not in the Apoer2 KO mice (Fig. 8), consistent with the previous observation that suppression of ApoER2 resulted in an altered distribution of neurons within the ectopic aggregates (Kubo et al. 2010). These results imply that ApoER2 is required for the Reelin-dependent neuronal aggregation, which recapitulates the formation of a cell-dense “primitive cortical zone” at the top of the CP (Sekine et al. 2011), suggesting the importance of ApoER2 for the events that normally occur beneath the MZ during cortical development.
Disrupted Neuronal Migration in the Hippocampal CA1 in the Apoer2 KO Mice
A previous article showed that hippocampal development was also disrupted in the Apoer2 KO mice (Trommsdorff et al. 1999), and we recently showed that ApoER2 is strongly expressed in the developing hippocampus (Hirota et al. 2015). Thus, we next examined whether Apoer2 deletion might also affect neuronal migration in the hippocampal CA1 region, the neuronal migration profile in which was recently well-described (Kitazawa et al. 2014). NeuN staining showed that the alignment of mature neurons in the SP at P0 was disorganized in the Apoer2 mutants. In the control mice, strongly NeuN-positive mature neurons were localized only in the deep part of the SP (Fig. 9A,B), whereas in the Apoer2 KO mice, they were distributed in both the superficial and deep parts of the SP (Fig. 9C,D). Tbr1 staining also showed disturbed alignment of neurons in the CA1 of the Apoer2 KO mutant mice, while no obvious differences were noted for the Ctip2-positive neurons (Fig. 9E–J).
We then investigated the migration of neurons from the VZ using IUE at E14.5. Distribution of the GFP-labeled cells was slightly shifted to a deeper position in the Apoer2 KO mice at 2 days after the IUE (Fig. 9K,L,Q). At 5 days after the IUE, the neuronal migration defect in the Apoer2 KO mice became clear. In the control mouse cortex, most GFP+ cells entered the SP, whereas in the Apoer2 KO mice, many labeled cells remained in the stratum oriens (SO) (Fig. 9M,N,R). Ctip2 staining revealed that neuronal differentiation was at least not severely affected in the Apoer2 KO (Fig. 9O,P). These results suggest that ApoER2 is required for neuronal migration in the developing CA1 of the hippocampus.
To uncover the roles of ApoER2 in cerebral cortical development, we examined Apoer2 KO phenotypes using stage-specific labeling of newborn neurons by IUE. In the Apoer2 KO mice, a considerable number of both early- and late-born neurons were found to abnormally invade the MZ. In addition, the Apoer2 mutants showed a defect in neuronal migration during both the early and late stages of cortical development and in the CA1 of the developing hippocampus, suggesting that ApoER2 is required for both proper migration and proper positioning of the neurons.
Previous studies showed that deep-layer marker-positive neurons were ectopically located in the more superficial part of the Apoer2 KO cortex (Benhayon et al. 2003; Hack et al. 2007). Birthdate labeling with BrdU or 3H-thymidine at E12.5 followed by analysis at P0 also showed that early-born neurons became scattered throughout the cortex in the Apoer2 KO mice, whereas the labeled cells were mainly localized in the deep part of the CP in the control mice (Trommsdorff et al. 1999; Hack et al. 2007). These observations suggest ApoER2 is required for normal positioning of early-born neurons, however, it was not clear whether the abnormal positioning of the early-born neurons was caused by a failure of radial neuronal migration in the IZ or was secondary to the disturbed migration of the late-born neurons. In this study, our direct examination using stage-specific labeling by IUE clearly demonstrated abnormal migration of both early and late-born neurons in the IZ of the Apoer2 KO mice. Early-born neurons migrate to the pial surface mainly by somal translocation (Nadarajah et al. 2001), which is a radial fiber-independent migration mode. It has been reported that the Reelin signal promotes somal translocation through the CrkL/C3G/Rap1/N-Cadherin pathway (Ballif et al. 2004; Olson et al. 2006; Voss et al. 2008; Franco et al. 2011). Migration failure of the early-born neurons in the Apoer2 KO mice suggests that ApoER2 is also involved in the somal translocation of neurons. Consistent with this idea, ApoER2 begins to be expressed in the deeper half of the IZ from E14 (Hirota et al. 2015), when early-born neurons migrate through the IZ. Reelin signal has also been shown to be required for the terminal translocation (Olson et al. 2006; Franco et al. 2011; Sekine et al. 2012), which resembles but is different from somal translocation and enables inside–out positioning of the neurons in the cortex (Olson et al. 2006; Franco et al. 2011; Sekine et al. 2012). We previously showed that Reelin stimulates the Dab1-Crk/CrkL-C3G-Rap1 pathway to activate integrin α5β1, which promotes neuronal adhesion to fibronectin during terminal translocation (Sekine et al. 2012). In addition, co-transfection of ApoER2 and Dab1 clearly promoted cell adhesion via integrin/fibronectin in a reconstruction experiment conducted using a non-neuronal cell line (Sekine et al. 2012). These results lend support to the idea that ApoER2 is also required for neuronal migration at the top of the CP.
In this study, we found abnormal invasion of the MZ by neurons in the Apoer2 KO mice. A previous study showed, using Reelin receptor-mutant mice, that migrating neurons invaded the MZ in the Vldlr KO mice (Hack et al. 2007). Together with the fact that VLDLR is expressed in the MZ (Hirota et al. 2015), these findings support the notion that Reelin binding to VLDLR is required for migrating neurons to finish their migration precisely just beneath the MZ. Our new finding of the abnormal invasion of neurons into the MZ in Apoer2 KO mice suggests that ApoER2 is also involved in this process. This is consistent with our previous observations that single knockdown of ApoER2 affected neuronal migration at the top of the CP (Kubo et al. 2010) and that ApoER2 is also expressed in the MZ in the developing cortex (Hirota et al. 2015). Indeed, Reelin induces degradation of Dab1 via a proteasome-dependent system mediated by SOCS7-Cul5-Rbx2 to allow proper termination of neuronal migration after they reach the top of the CP (Arnaud et al. 2003; Feng et al. 2007; Simo et al. 2010; Simo and Cooper 2013). These results suggest that both ApoER2 and VLDLR are involved in the termination of migration at the top of CP via Dab1 degradation. Furthermore, the present results show that expression of CA-Rap1 or CA-integrin α5 together with wild-type Akt partially rescued the overmigration of neurons into the MZ in the Apoer2 KO cortex. Considering that Reelin activates integrin α5β1 through the Crk/CrkL-C3G-Rap1 pathway during terminal translocation (Sekine et al. 2012), it is possible that the same pathway may also control proper termination of neurons beneath the MZ downstream of Reelin. Of note, we recently found that in a knock-in mouse with deletion of the C-terminal region of Reelin, the superficial-layer neurons abnormally invaded the MZ (Layer I) postnatally (Kohno et al. 2015), suggesting that proper Reelin signaling is also required for maintenance of the cell-sparse MZ (Layer I).
Several previous studies have reported on the roles of Reelin signaling in neuronal migration in the IZ. The existence of a small amount of Reelin in the deep IZ, in addition to its strong expression in the MZ, has been observed repeatedly (Yoshida et al. 2006; Uchida et al. 2009; Hirota et al. 2015). ApoER2 is also strongly expressed in the deep IZ from E14.0 to the neonatal stages (Uchida et al. 2009; Hirota et al. 2015), suggesting that Reelin and ApoER2 exert functions in this region. Consistent with this idea, a previous paper showed migration failure of the superficial-layer neurons, which are supposed to move through this region during the late stages of neuronal migration, in the Apoer2 KO mice (Hack et al. 2007). Another paper also showed that the migration pattern of neurons was altered in the VZ/SVZ of reeler mice (Britto et al. 2011). Our present observation that ectopically expressed Reelin failed to cause cellular aggregation in the Apoer2 KO cortex suggests the requirement of ApoER2 for migrating neurons to respond to Reelin in the IZ. We also found that expression of wild-type ApoER2 in migrating neurons resulted in only partial rescue of the neuronal migration in the Apoer2 KO cortex (Fig. 6B,C), which may suggest that ApoER2 controls neuronal migration in both cell-autonomous and noncell-autonomous manners. In the present study, we found the presence of MAP2- and CSPG-double-positive structures in the deep part of the Apoer2 KO cortex in the late migration stage and that the superficial-layer neurons tended to stop and accumulate beneath these structures, suggesting that neuronal migration is likely to be inhibited, at least in part, by these structures in the Apoer2 KO mice, similar to the neurons in the reeler mouse which tend to stop along the IPZ (Tabata and Nakajima 2002). These MAP2- and CSPG-double-positive abnormal structures in the Apoer2 KO mice emerge in the later stages, but not in the early stages of cortical development (Fig. 5O–V), while the IPZs in the reeler mice are already formed in the rather early stages of development, suggesting that relatively late-generated neurons might mainly contribute to the formation of the MAP2- and CSPG-double-positive abnormal structures in the Apoer2 KO mice. Digestion of endogenous CSPG by expression of ChABC failed to rescue the migration defect in the Apoer2 KO mice (Fig. 5W,X), suggesting that CSPG itself is not the main cause of the arrest of neuronal migration. The CSPG/MAP2-positive structures may inhibit the neuronal migration as a physical obstacle, at least in part, and some other component(s) than CSPG contained in these structures might inhibit the neuronal migration in the Apoer2 KO cortex. No increase in the area of CSPG/MAP2-positive regions was observed in the hippocampal CA1 of the Apoer2 KO mice (data not shown). Considering that hippocampal neurons exhibit a unique migration mode (“climbing mode”), different from that in the neocortex (Kitazawa et al. 2014), it is likely that the mechanisms underlying the control of neuronal migration by ApoER2 differ between the neocortex and hippocampus. Another possible mechanism underlying the migration defect in the IZ of the Apoer2 KO mice is that ApoER2 controls neuronal migration in a cell-autonomous manner, since it is strongly expressed in the multipolar cells in the deep part of the IZ or MAZ (Hirota et al. 2015) and exogenous Reelin influences the migratory behavior of the neurons in this region in cultured brain slices (Britto et al. 2014). This is also consistent with previous and present observations that overexpression of a mutant VLDLR or ApoER2 lacking the cytoplasmic domain, which would exert a dominant-negative on the Reelin signal, in the migrating neurons inhibited their migration (Jossin and Cooper 2011) (Fig. 6D,E). Considering that Akt expression partially rescued the migration of Apoer2-deficient neurons in the IZ, it is thought that ApoER2 controls neuronal migration via an Akt-mediated pathway.
In conclusion, this study suggests that ApoER2 controls both neuronal migration in the IZ and termination of migration just beneath the MZ and provides new insights into the mechanisms underlying correct formation of the layered cortex.
MEXT/JSPS KAKENHI JP 25113525, JP 25440114, JP 25123719, and JP 15H01219 to Y.H., JP 15H01293, JP 15K09723, JP 16K09997, and JP 26430075 to K.-i.K., JP 16H06482, JP 15H02355, and JP 15H01586 to K.N., Keio Gijuku Academic Development Funds to Y. H, K.-i.K. and K.N., Keio Gijuku Fukuzawa Memorial Fund for the Advancement of Education and Research to Y. H. and K.N., Mochida Memorial Foundation to Y.H., Narishige Neuroscience Research Foundation to Y.H., Brain Science Foundation to Y.H., Program for the Advancement of Keio Next Generation Research Projects to K.-i.K., Takeda Science Foundation to K.N., and the Naito Foundation to K.N.
We thank Dr Jun-ichi Miyazaki for the providing pCAGGS, Dr Hitoshi Kitayama for Rap1-63E, Dr Robert F. Hevner for the anti-Tbr1 antibody. Conflict of Interest: None declared.