VLDLR is not essential for reelin-induced neuronal aggregation but suppresses neuronal invasion into the marginal zone

In the developing neocortex, neurons radially migrate toward the pial surface. During the early stages of neocortical development, neurons migrate in a somal translocation mode, in which neurons detach from the ventricular surface and then shorten their radial process to move their cell bodies to just beneath the marginal zone (MZ) (Nadarajah et al., 2001). During the later stages, neurons sequentially adopt different migration modes, including multipolar migration (Tabata et al., 2009; Tabata and Nakajima, 2003) and radial glia-dependent locomotion (Nadarajah et al., 2001; Rakic, 1972). Upon reaching the top of the cortical plate (CP), neurons are thought to detach from the radial fibers and migrate in a terminal translocation mode, in which the cell bodies move to the most superficial region of the CP by shortening their radial process while retaining their contact with the MZ (Nadarajah et al., 2001; Sekine et al., 2011). Cell bodies of the immature neurons that have stopped migrating beneath the MZ are then densely packed in the most superficial region of the CP, which is termed the primitive cortical zone (PCZ) (Sekine et al., 2011). Subsequently, neurons form a CP in an ‘inside-out’ pattern, that is, the newly arriving neurons pass by their predecessor neurons to settle in the most superficial position in the CP. These birthdate-dependent layer formation and neuronal migration processes are regulated by several signaling cascades, including reelin signaling.

Reelin is a large extracellular protein, which has diverse and essential roles in the mammalian central nervous system, including in neuronal migration, layer formation, axonal and dendritic development and modulation of synapses underlying learning and memory in the mature brain (Hirota and Nakajima, 2017; Lee and D'Arcangelo, 2016; Sekine et al., 2014). During neocortical development, reelin is mainly secreted by the Cajal-Retzius neurons in the MZ, and controls the migratory behavior of excitatory neurons. Reelin signaling is mediated via its two major receptors, apolipoprotein E receptor 2 (ApoER2; also known as Lrp8) and very low density lipoprotein receptor (VLDLR) (D'Arcangelo et al., 1999; Hiesberger et al., 1999; Trommsdorff et al., 1999), and triggers phosphorylation and activation of the cytoplasmic adaptor protein disabled 1 (Dab1), which interacts with various downstream molecules (Honda et al., 2011). Mutant mice for reelin signaling components, including reeler mice lacking in the reelin protein, Apoer2/Vldlr double-knockout mice and Dab1 mutants show severe migration defects with roughly inverted formation of the cortical layers (Caviness, 1976; Hirota and Nakajima, 2017; Howell et al., 1997; Kojima et al., 2000; Sheldon et al., 1997; Trommsdorff et al., 1999; Yoneshima et al., 1997), reflecting the roles of these proteins in the cortical development. Although previous research has identified a large number of downstream effectors of the reelin pathway, the mechanisms at the cellular and molecular levels underlying the regulation of neuronal migration by reelin are not yet fully understood.

Our recent studies have shown that ectopic expression of reelin by in utero electroporation (IUE) (Tabata and Nakajima, 2001, 2008) caused neuronal aggregation in the developing neocortex (Ishii et al., 2015; Kubo et al., 2010). In response to ectopically expressed reelin, neurons oriented their leading processes to assemble in the reelin-rich region, resulting in the formation of a cell-body-sparse and dendrite-rich central region, resembling the structure of the MZ, whereas the cell bodies were densely packed in the peripheral region of the aggregate, resembling the structure of the PCZ. We have recently shown that Apoer2-deficient neurons failed to form neuronal aggregates in response to ectopically expressed reelin (Hirota et al., 2018), suggesting that ApoER2 is required for neuronal aggregation. On the other hand, although VLDLR mRNA and protein are shown to be expressed in the most superficial region of the CP and in the MZ, respectively (Hirota et al., 2015; Trommsdorff et al., 1999; Uchida et al., 2009), whether VLDLR serves functions distinct from those of ApoER2 in this region during the embryonic stages remains unknown. In this study, we investigated the functions of VLDLR and report that, unlike ApoER2, VLDLR is not a prerequisite for neuronal aggregation, but is required for suppressing neuronal invasion into the cell-sparse MZ during neocortical development.

Ectopic expression of reelin by IUE of migrating neurons results in the formation of neuronal aggregates with a cell-body-sparse and dendrite-rich central region that resembles the MZ (Hirota and Nakajima, 2017; Kubo et al., 2010; Matsunaga et al., 2017). This reelin-induced neuronal aggregation does not occur in Apoer2-KO mice, clearly indicating that ApoER2 is essential for this aggregate formation (Hirota et al., 2018). In contrast, whether VLDLR is also involved in this reelin-induced neuronal aggregation has not yet been demonstrated, even though VLDLR is found in abundance in the central region of the reelin-induced neuronal aggregates (Fig. 1A; the specificity of the immunofluorescence signal was confirmed using Vldlr-KO mice in Fig. 1B) (Kubo et al., 2010). To examine whether Vldlr deficiency might affect the pattern of aggregate formation, plasmids expressing reelin together with green fluorescent protein (GFP) were electroporated into the cortex of the Vldlr-KO mice on embryonic day (E) 14.0, and the brains were analyzed on postnatal day (P) 1.5. Ectopic reelin overexpression led to neuronal aggregation with a cell-body-sparse central region in the intermediate zone (IZ) of the control mice (Fig. 1C). In contrast, in the Vldlr-KO mice, which also showed neuronal aggregation in the IZ, the central cell-sparse MZ-like structure was absent (Fig. 1D,E). These results imply that VLDLR is required for the reelin-dependent segregation of the dendrite-rich cell-body-sparse region and cell soma-rich region during the formation of the neuronal aggregates. This segregation recapitulates the segregation between the dendrite-rich MZ and cell-body-dense PCZ in the most superficial region of the CP (Sekine et al., 2011; Shin et al., 2019), suggesting the possibility that VLDLR is involved in the formation of the cell-body-sparse MZ during cortical development.

A previous paper showed that, in Vldlr-KO mice, pyramidal neurons were ectopically localized in layer I in the P7 and adult stages (Hack et al., 2007). Based on these findings, it was previously suggested that Vldlr-deficient neurons may fail to stop migrating in the superficial part of the CP and invade the MZ. However, our recent study showed that impaired reelin signaling caused invasion of the superficial-layer neurons into layer I in the postnatal stages, after normal formation of the MZ, without affecting neuronal migration in the embryonic stages (Kohno et al., 2015). Thus, it is not yet entirely clear whether Vldlr-deficient neurons indeed invade the MZ in the embryonic stages. To address this issue, migrating newborn neurons were labeled with GFP by IUE on E14.5 and the brains were fixed 3 or 4 days later. Three days after the IUE, no alterations in the distribution of the GFP-labeled neurons were observed in the Vldlr-KO mice compared with the control mice (Fig. 2A-C), suggesting that radial migration in the IZ is largely normal in the Vldlr-KO mice. However, examination on day 4 after the IUE revealed altered distribution of the labeled cells in the Vldlr-KO mice. Although most of the GFP-labeled cells reached the most superficial region of the CP in both the control and the Vldlr-KO mice, some Vldlr-deficient neurons migrated to a more superficial position than that observed in the control mice (Fig. 2D-F) and localized in the MZ, whereas in the control mice, very few GFP-labeled neurons were found in the MZ (Fig. 2G-I). Similar results were also obtained when early-born neurons were labeled on E12.5 and examined 6 days later (Fig. 2J-L).

Next, we tried to determine which cell types were located in the MZ of the Vldlr-KO mice on P0 using layer marker staining. Double-immunostaining for reelin (expressed in the Cajal-Retzius neurons) (D'Arcangelo et al., 1995; Ogawa et al., 1995) and Tbr1 revealed strongly Tbr1-positive, but reelin-negative, neurons (deep layer neurons) (Hevner et al., 2001) in the MZ of the Vldlr-KO mice, but not in the MZ of the control mice (Fig. 3A,B,I). The number of Cajal-Retzius neurons (strongly double-positive for reelin and Tbr1) was not altered in the Vldlr-KO mice (Fig. 3I), suggesting that the Cajal-Retzius neurons were not affected. Similarly, Rorb-positive neurons, which are thought to be layer IV neurons or could contain layer V neurons (Nakagawa and O'Leary, 2003; Oishi et al., 2016a,b; Schaeren-Wiemers et al., 1997), and strongly Ctip2-positive (also known as Bcl11b) neurons with large nuclei, likely to be layer V neurons (Arlotta et al., 2005), were also abnormally observed in the MZ of the Vldlr-KO mice (Fig. 3C-F,I). Cux1 staining showed a markedly increased number of Cux1-positive superficial layer neurons (Nieto et al., 2004) in the MZ of the Vldlr-KO mice (Fig. 3G-I), consistent with previous reports of observations in the postnatal brain (Hack et al., 2007). Together, these results indicate that Vldlr deficiency causes a positional shift of both early-born and late-born cortical neurons into the MZ.

A previous paper showed that reelin signaling controls oriented migration and morphology of the neurons during neocortical development (Britto et al., 2011; Jossin and Cooper, 2011). To investigate whether Vldlr deficiency might affect the orientation of neurons that have reached the most superficial region of the CP, the centrosome position in the cells was examined using pericentrin (PCNT) immunostaining. In the control neurons and Vldlr-deficient neurons located within the CP, centrosomes were observed on the pial side of the nuclei (Fig. 4A-E,G,H). In contrast, in the Vldlr-deficient neurons that had entered the MZ, some centrosomes were positioned on the ventricular side of the nuclei, or tangentially (Fig. 4F,I). Associated with this, the primary dendrites of the Vldlr-deficient neurons extended toward the ventricular side, or tangentially (Fig. 4F,I), whereas those of the control neurons were oriented towards the pial side (Fig. 4B,C). These observations suggest that the Vldlr-deficient neurons that had invaded the MZ showed loss of orientation.

Next, to examine whether VLDLR suppresses neuronal invasion into the MZ in a cell-autonomous manner, we performed a rescue experiment. Introduction of a VLDLR-expressing vector into the migrating neurons in the Vldlr-KO cortex on E14.5 resulted in a significant decrease in the number of neurons entering the MZ on E18.5 compared with the observations in the controls (Fig. 5A-C). We confirmed that transfection of the VLDLR-expressing vector at the same concentration as that used in the rescue experiments had no effect in the control mice (Fig. S1). These results suggest that VLDLR controls the suppression of neuronal invasion into the MZ via, at least in part, a cell-autonomous mechanism.

Previous reports have shown that reelin signaling promotes somal and terminal translocation, both of which are thought to be a radial-glia-independent mode of neuronal migration into the superficial part of the CP, via the Rap1 (Rap1a)/N-Cadherin (Cdh2) and Crk/CrkL-C3G-Rap1-integrin α5β1 pathways, respectively (Franco et al., 2011; Sekine et al., 2012). In addition, we recently reported that Rap1, integrin α5 and Akt (Akt1), but not N-cadherin, are involved in the termination of migration beneath the MZ, mediated by ApoER2 (Hirota et al., 2018). Thus, we next examined whether the same molecules might also be involved downstream of VLDLR. Expression of constitutively active forms of integrin α5 (CA-integrin α5; integrin α5-GFFKA) or Rap1 (CA-Rap1; Rap1-63E), or wild-type Akt alone resulted in a significant decrease in the number of neurons invading the MZ in the Vldlr-KO mice (Fig. 6A-E). These results suggest the possibility that VLDLR controls the suppression of neuronal invasion into the MZ via both Rap1/integrin α5- and Akt-mediated pathways, although it cannot be completely ruled out that these pathways might suppress neuronal invasion into the MZ independently from VLDLR.

Reelin is known to promote the development and branching of the apical dendrites of the pyramidal neurons (Chai et al., 2014; Sekine et al., 2011, 2012), which may be involved in the correct positioning of neurons in the most superficial region of the CP. Thus, we next examined whether Vldlr deficiency might affect apical dendrite formation, using IUE on E14.5. Observations conducted on E18.5 showed that the Vldlr-deficient GFP-labeled neurons located in the most superficial region of the CP showed poorly branched apical dendrites in the MZ (Fig. 7B,D) compared with those in the control cases (Fig. 7A,C). The number of apical dendrite branches and the total apical dendrite length were significantly reduced in the Vldlr-KO mice (Fig. 7E,F). We also performed MAP2 staining to confirm defective maturation of dendrites in the MZ of the Vldlr-KO mice in three different regions in the medial-lateral axis (Fig. 8A-G). The ratio of the MAP2 staining intensity in the superficial half of the MZ to that in the deep half of the MZ was significantly decreased in the Vldlr-KO mice in all the three regions examined (Fig. 8H). As Vldlr-deficient neurons invade the MZ (Fig. 2), ectopically localized nuclei in the MZ may cause the alteration in the MAP2 staining intensity in the MZ of the Vldlr-KO mice. Thus, the same analysis using images in which the regions occupied by the nuclei were excluded was performed, which led to the same conclusion (Fig. 8I). These results suggest that VLDLR is required for proper maturation of apical dendrites.

Suppression of invasion of pyramidal neurons into the MZ is a crucial prerequisite for the formation of a fine laminar structure of the neocortex. Several lines of evidence have shown that reelin signaling plays important roles in the suppression of neuronal invasion into the MZ. In reeler mice, Dab1-deficient mice and Apoer2/Vldlr double-KO mice, migrating neurons fail to develop the cell-sparse MZ or show the subsequent layer I formation (Caviness, 1982; Goffinet, 1979; Howell et al., 1997; Kojima et al., 2000; Pinto Lord and Caviness, 1979; Sheldon et al., 1997; Trommsdorff et al., 1999; Yoneshima et al., 1997), suggesting that reelin binding to its receptors and subsequent activation of downstream signaling pathways is required for proper suppression of neuronal invasion into the MZ. Both ApoER2 and VLDLR proteins are expressed in abundance in the dendrites of the neurons in the most superficial region of the CP (Hirota et al., 2015), suggesting that they contribute to this process. However, whether these two receptors have distinct functions at this location during the embryonic stages remained to be elucidated. The present study showed that when reelin was ectopically expressed in the Vldlr-KO cortex, neurons formed aggregates, but these lacked the central cell-sparse MZ-like structure, whereas ApoER2 was required for the neuronal aggregation itself (Hirota et al., 2018). This observation indicates that VLDLR is not essential for the neuronal aggregation, but rather is specifically involved in the segregation of the cell-body-sparse MZ-like region and the cell-body-rich region in the neuronal aggregates, which may correspond to preventing neuronal cell bodies from invading the MZ. These results provide evidence for distinct functions of these two receptors during the last phase of neuronal migration (Fig. 9). Although both receptors are expressed in the reelin-enriched MZ, our previous study has shown that some VLDLR and ApoER2 signals in the MZ are localized in a mutually exclusive manner, suggesting that these two receptors may function on distinct fibers/processes (Hirota et al., 2015). Also, VLDLR exhibits a sixfold lower affinity for reelin than ApoER2 (Andersen et al., 2003), suggesting that these receptors begin to signal downstream events in a spatiotemporally distinct manner. These different properties may contribute to the distinct functions of VLDLR and ApoER2 in the suppression of neuronal invasion into the MZ. Consistent with the function of VLDLR in the aggregation caused by ectopically-expressed reelin, stage-specific labeling of neurons clearly showed that, in the absence of VLDLR, cortical neurons migrated to a more superficial position than that in the normal cortex without affecting radial migration toward the MZ. Our previous study showed that neuronal invasion into the MZ is also found in Apoer2-KO mice (Hirota et al., 2018). However, as Apoer2 deficiency also causes a defect of radial migration in the IZ, which may possibly affect later processes in the development of the CP, it is difficult to conclude that the ApoER2 receptor contributes in a fully cell-autonomous manner to suppression of neuronal invasion into the MZ. The findings of the present study clearly show that reelin signaling indeed has an important role in suppressing neuronal invasion into the MZ via the VLDLR.

How are the neuronal cell bodies suppressed from invading the MZ? Recent studies suggest a link between the positioning of neurons after neuronal migration and neurite development. Deficiency of Gα 13, a heterotrimeric G protein subunit, causes both overmigration and aberrant neurite morphology of cortical neurons (Moers et al., 2008; Scherer et al., 2017). Inhibition of Wwp1/2, E3 ubiquitin ligases, and the intragenic miRNA miR-140 also cause overshooting of migration accompanied by deformed neurites (Ambrozkiewicz et al., 2018). Deficiency of POMT1, the gene responsible for Walker-Warburg syndrome, which is characterized by disturbed dendritic development, caused overmigration of cortical neurons in a mouse model (Beltran-Valero de Bernabe et al., 2002; Judas et al., 2009). Furthermore, a recent report showed that generation of lateral protrusion in the leading processes suppressed the radial migration of newborn interneurons in a postnatal olfactory bulb (Sawada et al., 2018). Similarly, cortical neurons rapidly develop dendritic arbors and arrest the cell body below the first stable branch point in the most superficial region of the CP (O'Dell et al., 2015), suggesting contribution of the branching of leading processes/immature apical dendrites to suppression of neuronal cell body invasion into the MZ. Many reports have demonstrated the requirement for reelin signaling in dendrite formation in various contexts, including in the neocortex and hippocampus in both the embryonic and postnatal stages (Frotscher et al., 2017; Lee and D'Arcangelo, 2016). Live imaging analysis revealed that the leading processes of migrating cortical neurons progressively developed more branches upon contacting the reelin-rich MZ (Chai et al., 2014), suggesting that reelin promotes dendritic growth of neurons that have reached beneath the MZ. Our present results show that Vldlr deficiency causes a defect in the maturation of apical dendrites, associated with invasion of cortical neurons into the MZ. Furthermore, invasion of Vldlr-deficient neurons into the MZ was partially rescued by overexpression of Akt, which has been shown to promote dendrite development upon reelin stimulation through mTor activation in the hippocampal neurons (Jossin and Goffinet, 2007). Thus, the effect of reelin on dendrite development might contribute, at least in part, to the proper positioning of neurons beneath the MZ. Given that reelin increases microtubule assembly in developing dendrites (Meseke et al., 2013), and that Akt controls the microtubule dynamics in migrating cortical neurons in a dynein-dependent manner (Itoh et al., 2016), and that overexpression of Lis1 (Pafah1b1), a regulator of dynein, causes invasion of cortical neurons into the MZ (Katayama et al., 2017), a dynein-microtubule-dependent mechanism might also be involved in the appropriate suppression of neuronal invasion into the MZ.

Cell adhesion might also be involved in the suppression of neuronal invasion into the MZ. Locomoting neurons pause transiently just beneath the PCZ and then switch their migration mode to terminal translocation (Sekine et al., 2011). Reelin signaling triggers this mode switch through activation of integrin α5β1 through the Crk/CrkL-C3G-Rap1 pathway (Sekine et al., 2012). Thereafter, the neurons complete terminal translocation through several mechanisms including N-cadherin-mediated regulation of cell adhesion and Akt-mediated reorganization of the actin cytoskeleton, both of which are also controlled by reelin signaling (Chai et al., 2009; Franco et al., 2011). Our results in this study show that activation of either integrin α5 or Rap1 partially rescued the invasion of neurons into the MZ in the Vldlr-KO cortex (Fig. 6), suggesting the possibility that the Rap1/integrin α5β1 pathway might also contribute, downstream in reelin signaling, to suppression of neuronal invasion. Given that the inhibition of N-cadherin in the migrating neurons resulted in the loss of the cell-sparse central region in the reelin-induced neuronal aggregates (Matsunaga et al., 2017), N-cadherin-mediated cell adhesion is also likely to be involved in preventing neurons from invading the MZ-like region. The relationship among these molecules and how reelin suppresses neuronal invasion into the MZ downstream of VLDLR should be investigated in the future.

In conclusion, this study highlights the distinct functions of the two reelin receptors: VLDLR controls suppression of neuronal invasion into the MZ, whereas ApoER2 promotes neuronal aggregation. Reelin receptor-mediated functions at the final step of neuronal migration revealed in this study provide new insights into the mechanisms underlying correct formation of the layered neocortex.

The method used to maintain the colony of Vldlr KO mice has been previously described (Frykman et al., 1995). The day of vaginal plug detection was considered to be E0. All the 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 Japanese Government Law Concerning the Protection and Control of Animals, and the Japanese Government Notification of Feeding and Safekeeping of Animals. The genotype of the mutants was confirmed by PCR analysis of genomic DNA using the following primers: for wild-type (WT) allele, mVR/E5F2 (TGCAATGGCCAGGATGACTGTG) and mVR/E6R1 (AAACTGGTCAGGTCGGCAGG); for mutant allele, mVR/neo2 (AGCAGCCGATTGTCTGTTGTGC) and mVR/E6R1. The sizes of the PCR products were 418 bp (WT allele) and 600 bp (mutant allele), respectively.

The embryos or 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 for 45 min on ice, replaced in a 25% 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 (5% donkey or goat serum and 0.1% Triton X-100 in PBS), then overnight with the primary antibodies at 4°C, and then for 2 h at room temperature with species-specific secondary antibodies (A21432, A21206, A11005, A11007 and A11008, 1:1000, Thermo Fisher Scientific or 703-485-155, 1:1000, Jackson ImmunoResearch Laboratories). The following primary antibodies were used in this study: VLDLR (goat, 1:600, R&D Systems, AF2258), Tbr1 (rabbit, 1:1000, gift from R. Hevner, University of California, USA), reelin (goat, 1:200, R&D Systems, AF3820), Rorb (mouse, 1:400, Perseus Proteomics, PPN7927-00), Ctip2 (rat, 1:500, Abcam, ab18465), Cux1 (rabbit, 1:200, Santa Cruz Biotechnology, sc-13024), MAP2 (mouse, 1:400, Santa Cruz Biotechnology, sc-32791), pericentrin (PCNT; rabbit, 1:500, Abcam, ab4448), green fluorescent protein (GFP; chick, 1:2000, Abcam, ab13970). 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 and National Institutes of Health Fiji/ImageJ version 2.0. For quantitative analyses of the neuronal morphology, reconstructions were performed from confocal z-series stacks (12-15 μm). The number of dendritic branches and the total apical dendrite length were analyzed using the ‘simple neurite tracer’ plugin of the Fiji/ImageJ software. For quantitative analyses of the MAP2 staining in the MZ, immunofluorescence signal intensities were measured with the Fiji/ImageJ software.

pCAGGS-EGFP has been described previously (Niwa et al., 1991). The Tα1 expression vector has been described previously (Sekine et al., 2012). To generate pTα1-VLDLR, the cDNA of VLDLR (Oka et al., 1994) was cloned into the Tα1 vector. The Tα1-Cre vector was obtained as a kind gift from Drs Sakakibara and Miyata (Nagoya University, Japan). pCALNL-Akt and pCALNL-integrin α5-GFFKA (Sekine et al., 2012), pTα1-Rap1-63E (Hirota et al., 2018), and pCAGGS-Reelin (Kubo et al., 2010) have been described previously.

Pregnant mice were deeply anesthetized and their intrauterine embryos were surgically manipulated as described previously (Nakajima et al., 1997). IUE to transfect vectors into the embryonic neocortex was performed as described previously (Tabata and Nakajima, 2001, 2008). Approximately 1-2 μl of the plasmid solution 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) with a forceps-type electrode (CUY650P3 for E12.5, CUY650P5 for E14.5). The concentrations of the plasmids were as follows. Fig. 1: 1.0 mg/ml of pCAGGS-EGFP and 7.0 mg/ml of pCAGGS-Reelin; Figs 2, 4 and 7: 1.0 mg/ml of pCAGGS-EGFP; Fig. 5: (A, empty vector) 1.0 mg/ml of pCAGGS-EGFP and 1.0 mg/ml of pTα1; (B, VLDLR) 1.0 mg/ml of pCAGGS-EGFP and 1.0 mg/ml of pTα1-VLDLR; Fig. 6: (A, empty vector) 0.25 mg/ml of pCALNL-EGFP, 2.0 mg/ml pCALNL-integrin α5-GFFKA and 0.75 mg/ml Tα1-Cre; (B, CA-integrin α5) 0.25 mg/ml of pCALNL-EGFP, 2.0 mg/ml pCALNL-integrin α5-GFFKA and 0.75 mg/ml Tα1-Cre; (C, Akt) 0.25 mg/ml of pCALNL-EGFP, 2.0 mg/ml pCALNL-Akt and 0.75 mg/ml Tα1-Cre; (D, CA-Rap1) 1.0 mg/ml of pCAGGS-EGFP and 2.5 mg/ml pTα1-Rap1-63E.

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 at E17.5 and E18.5. The whole cortex was divided into 10 equally spaced bins. The distances from the top of the cortex to the nuclei of the migrating cells were measured using the Fiji/ImageJ software. All data were expressed as mean±s.e.m. For direct comparisons, the data were analyzed using an unpaired two-tailed Student's t-test as neither normality nor homogeneity of variances of the dataset was rejected by the Chi-square test and F test, respectively. For multiple comparisons, the data were analyzed by one-way ANOVA with Tukey's post hoc test as neither normality nor homogeneity of variances of the dataset was rejected by the Shapiro-Wilk test and Levene's test, respectively. A P-value less than 0.05 was considered significant.

We thank Dr Jun-ichi Miyazaki for providing pCAGGS, Drs Sakakibara and Miyata for Tα1-Cre vector, Dr. Hitoshi Kitayama for Rap1-63E and Dr Robert F. Hevner for the anti-Tbr1 antibody.

The peer review history is available online at https://dev.biologists.org/lookup/doi/10.1242/dev.189936.reviewer-comments.pdf