|Home | About | Journals | Submit | Contact Us | Français|
During corticogenesis, the earliest generated neurons form the preplate, which evolves into the marginal zone and subplate. Lrp12/Mig13a, a mammalian gene related to the Caenorhabditis elegans neuroblast migration gene mig-13, is expressed in a subpopulation of preplate neurons that undergo ventrally directed tangential migrations in the preplate layer and pioneer axon projections to the anterior commissure. As the preplate separates, Lrp12/Mig13a-positive neurons polarize in the radial plane and form a pseudocolumnar pattern, prior to moving to a deeper position within the emerging subplate layer. These changes in neuronal polarity do not occur in reeler mutant mice, revealing the earliest known defect in reeler cortical patterning and suggesting that the alignment of preplate neurons into a pseudolayer facilitates the movement of later-born radially migrating neurons into the emerging cortical plate.
The emergence of the laminar architecture of the mammalian forebrain depends on the spatiotemporal pattern of neurogenesis, migration of postmitotic neurons (Caviness and Sidman 1973; Zecevic et al. 1999; Nadarajah et al. 2001; Molnar et al. 2006), formation of transverse zones of transient neurons (preplate, cortical plate, marginal zone, and subplate) (Caviness 1982; Allendoerfer and Shatz 1994; McConnell et al. 1994; Del Rio et al. 2000; Zecevic and Rakic 2001), and positioning of the principal classes of cortical neurons into neuronal layers. Although these basic steps of corticogenesis are conserved among mammalian species, studies on human and nonhuman primate corticogenesis indicate differences in the period of neurogenesis, the temporal gradient of maturation across the hemisphere, and cellular complexity of the transient cell layers (Bystron et al. 2008). In the mouse, the earliest neurons generated in the ventricular zone (VZ) of the dorsal telencephalon become postmitotic on, or before, the eleventh day of gestation (E11). These early-born polymorphic neurons complete their migrations by E13 and settle in a narrow laminar zone, the preplate (Caviness 1982; Kostovic and Rakic 1990; Wood et al. 1992; McConnell et al. 1994; Del Rio et al. 2000; Zecevic and Rakic 2001). The preplate, now recognized as the initial zone of postmigratory neurons and neuropil, forms between the VZ and the pial surface of the dorsal telencephalon (Bystron et al. 2008). Between E13–E15 in the mouse, radially migrating neurons pattern a second transient layer in the middle of the preplate, the cortical plate. As the cortical plate forms, preplate neurons simultaneously segregate into a superficial marginal zone containing Cajal–Retzius cells and a deeper layer, the subplate (Caviness 1982; Wood et al. 1992; McConnell et al. 1994; Del Rio et al. 2000; Molnar et al. 2006).
Studies on the genetics of cortical development underscore the importance of preplate separation to the development of laminar cytoarchitectonics. During corticogenesis in reeler mutant mice, the segregation of preplate neurons into 2 zones and formation of the cortical plate does not occur and subsequent cortical lamination fails (Caviness and Rakic 1978; Goffinet 1984; Tissir and Goffinet 2003). The Reelin (Reln) protein (RELN) binds to 2 members of the lipoprotein receptor family, the very low-density lipoprotein receptor (VLDLR) and the apolipoprotein E receptor type 2 (APOER2) (D'Arcangelo et al. 1995; Trommsdorff et al. 1999; Herz and Bock 2002; Tissir and Goffinet 2003). Mice lacking Vldlr and ApoER2 have cortical malformations similar to a reeler phenotype. RELN signaling activates nonreceptor tyrosine kinases of the Src and Fyn families, leading to tyrosine phosphorylation of the intracellular adapter DAB1 (Howell et al. 1997; Jossin et al. 2003) and phosphatidylinositol 3-kinase (Jossin and Goffinet 2007). Genetic ablation of Src and Fyn kinases or inhibitors of Src family kinases (SFKs) also lead to defects in the orientation and lamination of cortical neurons (Jossin et al. 2003; Kuo et al. 2005).
The overall plan of cortical connectivity includes the formation of reciprocal connections between the cortex and the thalamus and corticocortical connections between the cerebral hemispheres (McConnell et al. 1994; Xie et al. 2002; Richards et al. 2004; Price et al. 2006). These projections are pioneered by axons of transient populations of neurons in the preplate, and later, by neurons in the subplate (De Carlos and O'Leary 1992; McConnell et al. 1994). Since many local cues and signaling pathways that affect axon navigation and cell migration are evolutionarily conserved, one approach to discovering novel regulators of cortical development is to identify vertebrate genes whose invertebrate homologs affect normal development (Kee et al. 2007). In Caenorhabditis elegans, 3 classes of neurons (anterior lateral microtubule cell, canal-associated neurons, and hermaphrodite specific neuron) and one line of neuroblasts, Q cells and their descendants, undergo extensive migrations during central nervous system (CNS) development (Blelloch et al. 1999; Shakir and Lundquist 2005). QR neuroblasts migrate toward the anterior by a mechanism that involves instructive changes in the level of expression of the transmembrane protein MIG-13 (Sym et al. 1999). In this study, we identify Lrp12/Mig13a, a mammalian gene related to mig-13, as a novel marker for a distinct subpopulation of preplate neurons. To map the pattern of gene expression, we generated lines of bacterial artificial chromosome transgenic mice that express the enhanced green fluorescent protein (Egfp) reporter gene under the control of the native Lrp12/Mig13a locus [Tg(Lrp12/Mig13a-Egfp)] (Gong et al. 2003). In these mice, Lrp12/Mig13a expression is restricted to a transient population of cells in the preplate, which segregate into the subplate during formation of the cortical plate. The present studies reveal a novel series of cell movements that orient the polarity of Lrp12/Mig13a-expressing neurons and their axons just prior to cortical plate assembly. These changes do not occur in reeler mutant mice or after perturbation of the RELN signaling pathway in wild-type mice. Our results suggest an earlier defect than previously recognized in reeler corticogenesis, which involves defects in local cell movements needed to align into a pseudolayer rather than an arrest of neuronal migration along glial fibers.
Tg(Lrp12/Mig13a-Egfp) transgenic mice [CD-1-Tg(RP23-181D20)BG31Gen] were generated as described previously (Gong et al. 2003). The reeler mutation, gene symbol rl or Reln, is maintained on B6C3Fea/a hybrid mouse strain (Jackson Laboratory). To express Lrp12/Mig13a-Egfp in rl/rl mutant mice, we crossed a male Tg(Lrp12/Mig13a-Egfp) with a female B6C3Fea/a-Relnrl/J mouse (Jackson Laboratory [stock number 000235]) and intercrossed F1 offspring to generate B6C3FeTg(Lrp12/Mig13a-Egfp)a-Relnrl/a-Relnrl mice. The genotype of B6C3FeTg(Lrp12/Mig13a-Egfp)a-Relnrl/a-Relnrl and B6C3Fe-Tg(Lrp12/Mig13a-Egfp)a/J embryos was confirmed by polymerase chain reaction (PCR) analysis for Egfp and Relnrn as described (D'Arcangelo et al. 1996; Gong et al. 2003). All animal procedures were performed in accordance with institutional guidelines.
The pregnant female was injected intraperitoneally with thymidine analog 5′-bromo-2′-deoxyuridine (BrdU) (5 mg/g body weight) at 24-h intervals between the 10th (El0.5) and 12th (E12.5) days of gestation. Animals were killed 1–3 days later by an overdose of Pentobarbital (Nembutal; Abbott), and embryos were removed by laparotomy. The embryos/brains were fixed and immunostained with antibodies against BrdU and enhanced green fluorescent protein (EGFP) as described below to correlate Lrp12/Mig13a expression with neurogenesis.
Tg(Lrp12/Mig13a-Egfp) embryos or embryonic brains were dissected in phosphate-buffered saline (PBS) (4 °C), fixed in paraformaldehyde (4%, 4 °C, 1–3 h), immersed in sucrose (30%, 4 °C, overnight), embedded in Neg-50 (Richard-Allan Scientific), and sectioned (20 μm) with a Microm Model HM 500 M Cryostat (GMI, Inc.). Nonspecific immunostaining was blocked by pretreating with normal donkey serum (5% in PBS containing 0.1% Triton-X-100; Jackson ImmunoResearch Laboratories, Inc.). Primary and secondary antibody staining was carried out at 4 oC overnight. The primary antibodies used in this study were anti-GFP rabbit polyclonal antibody (1:2000, Molecular Probes), anti-GFP antibody (sheep polyclonal, 1:250; Biogenesis), anti-LRP12/MIG13A antibody (rabbit polyclonal, 1:50) (Schneider S, Gulacsi A, Gong S, Ayad N, Hatten ME, in preparation), anti-TAG-1 antibodies (mouse monoclonal, Dr Jane Dodd, Columbia University, NY), anti-L1 (mouse monoclonal IgG, 324, 1:20; Dr J. Trotter, University of Mainz, Germany), anti-CALB1 antibody (rabbit polyclonal, 1:1000; Swant), anti-RELN antibody (mouse monoclonal IgG, clone G10, 1:1000; Chemicon/Millipore Biosciences Division, Danvers, MA), anti-CALB2 antibodies (rabbit polyclonal, 1:1000) (Swant), anti-BrdU antibodies (mouse monoclonal IgG, 1:100, Becton Dickson Biosciences), anti-MAP2 antibodies (mouse monoclonal IgG, clone SMI 52, Covance), and anti-GM130 antibodies (mouse monoclonal IgG, BD Biosciences). Secondary antibodies were purchased from Jackson ImmunoResearch and Molecular Probes (Invitrogen Corp.). Nuclei were visualized using 4′,6-diamidino-2-phenylindole (DAPI) (Sigma) or DRAQ5 (Axxora LLC.). Double staining for EGFP and BrdU was performed by immunostaining for EGFP, as described above, postfixing, treating with 4 N HCl/0.5% Tween20 (20 min), and then with 0.1 M Borax buffer pH 8.5 (5 min), and staining with anti-BrdU primary antibody (mouse monoclonal IgG, 1:100, Becton Dickson Biosciences) and anti-mouse secondary antibody as above. Images were acquired using an Axioscope 200 inverted microscope (Zeiss) with a LSM510 confocal laser scanhead (Carl Zeiss Microimaging, Inc.).
Embryos were fixed with paraformaldehyde and immersed in sucrose (30%, 4 °C, overnight), after which 60-μm sections were generated with a Leica RM2265 microtome (Leica Microsystems, Inc.). Hybridization and detection were performed as described (Schaeren-Wiemers and Gerfin-Moser 1993). The Lrp12/Mig13a in situ probe was amplified with the following primers: sense primer 5′-GTCAGTTGATCGCTCAGGG-3′ and antisense primer 5′-GGACAGATACACAAGACCTC-3′.
To assay the effects of SFK inhibitor PP2 (Calbiochem) on Lrp12/Mig13a-positive preplate cells, coronal slices (250 μm) of E12.5 Tg(Lrp12/Mig13a-Egfp) mouse neocortex, generated with a Leica VT100S vibrating blade microtome (Leica Microsystems, Inc.), were cultured in 35% Eagle's Basal Medium, 35% L15, 25% complete Hank's balanced salt solution, 20 mM D-glucose, 1 mM L-Glutamine, 10 mM Hepes, 1× Pen/Strep, 5% heat-inactivated horse serum. For SFK inhibition, E12.5 Tg(Lrp12/Mig13a-Egfp) cortical slices were cultured in medium supplemented with dimethyl sulfoxide with or without 10 μM PP2.
Coronal slices (250 μm) of E12.5 or E13.5 Tg(Lrp12/Mig13a-Egfp) neocortex were maintained at 37 °C in a Warner environmental imaging chamber (Warner Instruments) for 7–15 h. Time-lapse images were acquired with a NLO LSM 510 multiphoton microscope (Zeiss) equipped with a Coherent Chameleon multiphoton 870 nm laser, Zeiss LSM software, Imaris (Bitplane Scientific Solutions) and ImageJ. For Supplementary Movie 1, we used Photoshop CS3 to isolate a series of time-lapse images of a single migrating cell by hand, after which we used Final Cut Pro to sequence the images chronologically and add text, and QuickTime conversion software to export the resulting movie into MPEG-4 format. ImageJ software (http://rsb.info.nih.gov/ij/) was used to determine the trajectory and speed of migration of Lrp12/Mig13a-positive neurons in Tg(Lrp12/Mig13a-Egfp) tissue slices.
To visualize the nucleus and Golgi apparatus of Lrp12/Mig13a-postive neurons in coronal cryosections (20 μm) of Tg(Lrp12/Mig13a-Egfp), B6C3Fe-Tg(Lrp12/Mig13a-Egfp)a/J, or B6C3FeTg(Lrp12/Mig13a-Egfp)a-Relnrl/a-Relnrn embryos, embryonic brains were immunostained with anti-EGFP and anti-GM130 antibodies and counterstained with DAPI. To measure the polarity of preplate neurons relative to the neuroaxis, we drew a line from the center of the DAPI-stained nucleus through the center of the GM130-positive Golgi apparatus. ImageJ software was used to score and bin the orientation of Lrp12/Mig13a-positive preplate cells as illustrated in Figures 7e and 8e and Supplementary Figure 8e. To measure the polarity of Lrp12/Mig13a-postive neurons in slice cultures of E12.5 neocortex, cultured for 0div or 2div, we fixed the slices and resectioned them (20 μm) using a cryostat, immunostained the slices with anti-EGFP and anti-GM130 antibodies and counterstained them with DAPI. The orientation (polarity) of Lrp12/Mig13a-postive cells relative to the neuroaxis was measured as described above.
HEK 293T cells were cotransfected with Flag-tagged full-length Lrp12/Mig13a cDNA and control nontargeting small interfering (siRNA) pool (Thermo Scientific Dharmacon siGENOME Control Pool, Non-Targeting no. 2, Cat no. D-001206-14-05) or Lrp12/Mig13a siRNA pool (Thermo Scientific Dharmacon siGENOME SMARTpool, Mouse C820005L12RIK, Cat no. M-055299-00), using Lipofectamine 2000. Five microgam of DNA/7.5 μL of 20 μM (2 μg) siRNA pool/250 μL of Dulbecco's Modified Eagle's Medium (DMEM) were mixed with 7.5 μL of Lipofectamine 2000/250 μL of DMEM and added to one well of a 6-well plate. Lysates were harvested 24 and 48 h after transfection and analyzed by western blotting. Membranes were immunoblotted with a rabbit antibody to the C-terminus of LRP12/MIG13a and with mouse anti-GAPDH as a loading control.
Two Tg(Lrp12/Mig13a-Egfp) dams were sacrificed at E11.5, embryos were removed by laparotomy and brains were removed by dissection. After embedding in 4% low-melting point agarose, 250-μm coronal slices were prepared by vibratome sectioning (Leica VT1000S) and transferred to Nuclepore Track-Etch Membranes (Whatman) floating on culture medium as described previously (Polleux et al. 2002). Control or Lrp12/Mig13a SMARTpool siRNAs were injected at 10 μM along with the pCIG-tdTomato vector (200 ng/mL), containing the tdTomato reporter gene, kindly provided by Dr Roger Tsien, into the lateral ventricles of the telencephalon. One side of the lateral/dorsolateral telencephalic wall was electroporated with a BTX electroporation system (Electro Square Porator T830) with platinum electrodes (Protech International) (130 V, 2 pulses of 5 ms duration at an interval of 1 s). Slices were cultured for 2 days in vitro (DIV) prior to fixing with 4% paraformaldehyde and embedding in OCT (Optimal Cutting Temperature, Tissue-Tek OCT Compound Histomount), after which the embedded slices were resectioned at 16 μm using a cryostat (Microm Model HM 500 M Cryostat, Leica CM 3050S). Sections were immunolabeled for EGFP and counterstained with DRAQ5 to visualize cell nuclei. tdTomato was visible without antibody amplification.
We analyzed the distribution of Lrp12/Mig13a-positive (EGFP labeled) and siRNA transfected (tdTomato positive) cells in slices electroporated with control siRNA (from n = 5 embryos) or Lrp12/Mig13 siRNA (from n = 5 embryos). We counted the number of transfected, tdTomato-positive cells in 2–3 fields per section (at least 100 cells per section) and calculated the ratio of transfected, tdTomato-positive cells in the lower half of the cortical wall, containing the VZ, to tdTomato-positive cells in the upper half, containing the preplate. The Student's t-test was performed on the 2 sets of data with Microsoft Excel.
We identified murine and human homologs of mig-13 by tBLAST search of human and mouse genome sequences using the protein sequence of C. elegans mig-13 as a query. Mus musculus LRP12/MIG13A is a novel transmembrane domain containing glycoprotein with maximal expression levels during embryonic CNS development (data not shown). The translated sequence for Mus musculus LRP12/MIG13A (Genbank accession number NM 172814) has 92% overall identity to Homo sapiens LRP12/MIG13A. To map the expression of Lrp12/Mig13a on a cellular level, we generated lines of bacterial artificial chromosome transgenic mice that express the Egfp reporter gene under the control of the native Lrp12/Mig13a locus [Tg(Lrp12/Mig13a-Egfp)] (Gong et al. 2003). At the preplate stage, immunostaining coronal sections of E12.5 Tg(Lrp12/Mig13a-Egfp) mouse telencephalon with anti-LRP12/MIG13A and anti-EGFP antibodies revealed a zone of double-labeled cells in the superficial layer of the dorsal telencephalon (Fig. 1a,b). In the E12.5 telencephalic wall, Lrp12/Mig13a messenger RNA also localizes to a superficial cortical layer close to the pial surface (Supplementary Fig. 1).
In the mouse, cumulative labeling with the thymidine analog BrdU has shown that the first neuronal progenitors exit the cell cycle in the rostrolateral aspect of the dorsal telencephalic VZ. Neurons produced at the onset of neurogenesis, E10.5–E11.5 in the mouse, form the preplate layer. To correlate the timing of the penultimate cell division and expression of Lrp12/Mig13a, we injected pregnant Tg(Lrp12/Mig13a-Egfp) dams with BrdU between E10.5 and E12.5 and assayed BrdU accumulation in Lrp12/Mig13a-positive cells during early phases of cortical development. At E13.5, the vast majority of EGFP/BrdU double-labeled cells were observed in embryos exposed to BrdU on E10.5 or E11.5, suggesting that Lrp12/Mig13a-positive cells undergo terminal divisions between E10.5 and E11.5 (Fig. 2a,b). BrdU labeling a day later (E12.5) allowed us to trace the position of preplate-derived Lrp12/Mig13a-positive cells relative to later-born neurons, which form the cortical plate. At E14.5, the preplate has split into the superficial marginal zone and the deeper subplate. Very few double-labeled (BrdU/Lrp12/Mig13a) cells were observed in lateral neocortical areas of E14.5 embryos exposed to BrdU on E12.5, when cells destined for the cortical plate undergo terminal cell division (Fig. 2a,b). While the majority of cells born on E12.5 were located in the cortical plate, BrdU-positive cells were seen positioned beneath the cortical plate, among migrating neurons and subplate neurons. In E11.5 embryos, Lrp12/Mig13a-positive cells coexpressed TUJ1 by immunocytochemistry, which identified them as postmitotic neurons. Lrp12/Mig13a/TUJ1-positive neurons were located above the cortical VZ, in the superficial layer of the E11.5 telencephalic wall (Fig. 2c,d). These experiments show that Lrp12/Mig13a expression is restricted to preplate neurons.
We examined whether Lrp12/Mig13a-positive preplate neurons included precursors of Cajal–Retzius neurons within the marginal zone that express the markers RELN and Calretinin (CALB2) or interneurons that express the marker Calbindin (CALB1) (Meyer and Goffinet 1998; Hevner et al. 2003). At E13.5, Lrp12/Mig13a-expressing cells did not coexpress CALB1, RELN, or CALB2 (Supplementary Fig. 2). RELN-positive Cajal–Retzius cells were superficial to Lrp12/Mig13a-EGFP-positive neurons (Supplementary Fig. 2). Lrp12/Mig13a-positive cells were also distinct from the TAG-1-positive plexiform layer in the intermediate zone (Fig. 3a,b). Lrp12/Mig13a-EGFP-positive preplate neurons coexpressed the microtubule-associated protein MAP2, a marker for mouse preplate neurons that segregate both to the marginal zone and to the subplate (Crandall et al. 1986; Chun et al. 1987). The coexpression of MAP2 by Lrp12/Mig13a-positive cells in the newly formed subplate at E14.5 (Fig. 3c) showed that Lrp12/Mig13a marks a subpopulation of preplate neurons that later reside in the subplate. Lrp12/Mig13a expression in the subplate layer persisted through E18 (Fig. 3d,e) and declined in the early postnatal period (data not shown). Thus, Lrp12/Mig13a-positive neurons constitute a subpopulation of preplate neurons that segregate into the subplate during early cortical development.
During corticogenesis, transient neuronal populations of the preplate and subplate extend descending axons to subcortical targets and corticocortical axons to targets in the ipsilateral or contralateral hemisphere. The latter pioneer axon tracts that project through the anterior commissure. At E13.5, Lrp12/Mig13a-positive preplate neurons extended axons that descended into the intermediate zone before navigating through the subpallium and exiting the cortical hemisphere via the anterior commissure (Fig. 4a,b and Supplementary Fig. 3a). In contrast, thalamocortical fibers immunopositive for anti-CALB2 (Del Rio et al. 2000) projected to the internal capsule (Fig. 4a,b,d); Lrp12/Mig13a-positive fibers were not detected in this pathway. Lrp12/Mig13a-positive fibers did not coexpress the transient axonal glycoprotein (TAG-1) (Denaxa et al. 2001), a marker for corticothalamic projections at the preplate stage (Fig. 4a,c). TAG-1-positive fibers extended through the upper aspect of the intermediate zone, above Lrp12/Mig13a-positive fibers (Fig. 4c). Lrp12/Mig13a-positive axons were immunopositive for antibodies against the cell adhesion molecule L1, a marker for commissural projections (Fukuda et al. 1997; Jones et al. 2002; Morante-Oria et al. 2003) (Supplementary Fig. 3b). These results show that Lrp12/Mig13a is expressed in a subpopulation of preplate neurons, whose axons pioneer projections to the contralateral hemisphere via the anterior commissure. At later stages, this projection forms corticocortical connections with the contralateral temporal lobe.
Although studies on the development of cortical malformations, including those in reeler mice, have defined the principal steps of corticogenesis, the dynamics of cell movements in the preplate zone have not been examined. To examine tangential movements of preplate neurons (Tomioka et al. 2000), we prepared coronal slices of E12.5 Tg(Lrp12/Mig13a-Egfp) mouse forebrain and used multiphoton microscopy to acquire time-lapse images of living Lrp12/Mig13a-positive preplate cells. Approximately 5% of Lrp12/Mig13a-expressing cells migrated along a ventral tangential trajectory in the middle of the preplate layer and at the boundary of the preplate and VZ during the period we imaged (Fig. 5a,b, Supplementary Movie 1). Lrp12/Mig13a-expressing cells were not observed in the superficial subpial aspect of the preplate zone, where Cajal–Retzius cell migration has been reported. During periods of movement, Lrp12/Mig13a-positive, EGFP-labeled neurons polarized and extended and retracted a broad lamellipodium (Fig. 5a,b, Supplementary Movie 1) while migrating at speeds ranging from 0 to 140 μm/h (average velocity, 18 μm/h) (Supplementary Fig. 4a,b). These findings reveal the dynamics of tangential migrations within the preplate layer prior to cortical plate formation.
To determine whether Lrp12/Mig13a functions in cell motility during the preplate to cortical plate transition, we used SMARTpool siRNAs to knockdown Lrp12/Mig13a levels. Cotransfection of a full-length Lrp12/Mig13a cDNA and SMARTpool Lrp12/Mig13a siRNAs into HEK 293T cells resulted in complete knock down of LRP12/MIG13A protein levels within 24 h, suggesting efficient posttranscriptional silencing of Lrp12/Mig13a (Supplementary Fig. 5). To examine the effect of silencing Lrp12/Mig13a expression in developing murine neocortex, coronal slices of E11.5 Tg(Lrp12/Mig13a-Egfp) mouse forebrain were coelectroporated with either control or Lrp12/Mig13a SMARTpool siRNAs and tdTomato as an indicator for electroporated cells. After 2 DIV, the position of cells expressing EGFP as an indicator of endogenous Lrp12/Mig13a expression and the tdTomato reporter was examined by confocal microscopy. In sections of control siRNA electroporated slices, preplate formation occurred normally in both lateral and dorsolateral areas of the telencephalic wall (Supplementary Fig. 6a,b). After electroporation of Lrp12/Mig13a SMARTpool siRNAs, 2 classes of defects in cell migration were observed. First, Lrp12/Mig13a-positive, EGFP labeled cells that expressed tdTomato were located in the deep layers of the neocortex, suggesting a failure to migrate into the preplate zone. Second, cells that expressed only tdTomato also accumulated in the deep layers of the telencephalic wall, suggesting nonautonomous defects in radial migrations (Supplementary Fig. 6c,d). Thus, silencing Lrp12/Mig13a affects cell migrations of Lrp12/Mig13a-expressing cells, as well as the migrations of other cells in the developing cortex. This prompted further analysis of Lrp12/Mig13a-positive cells at the preplate stage of cortical development.
To examine the behavior of preplate neurons during separation of the preplate into a superficial marginal zone and deeper subplate zone, we imaged Lrp12/Mig13a-positive, EGFP-labeled cells in coronal forebrain slices of E13.5 Tg(Lrp12/Mig13a-Egfp) mice. Real-time imaging by multiphoton microscopy revealed dynamic changes in the polarity and motility of Lrp12/Mig13a-positive cells (Fig. 6a,b). Lrp12/Mig13a-positive cells assumed an elongated morphology and aligned along the radius of the cortical wall. In live images, during preplate separation, Lrp12/Mig13a-positive cells drifted into the deeper subplate position (Fig. 6a,b). Thus, as the subplate formed, Mig13a/Lrp12-positive cells moved into the emerging subplate layer.
The changing orientation of preplate neurons at late stages of preplate development followed a spatiotemporal gradient. In dorsal aspects of E12.5–E13.5 telencephalon, Lrp12/Mig13a-positive preplate neurons assumed an apparently random orientation (Fig. 6c). At the onset of cortical plate formation in the ventrolateral telencephalic wall (E13.5), Lrp12/Mig13a-positive preplate neurons polarized toward the pial surface in a pseudocolumnar pattern (Fig. 6d). These radially oriented cells were intercalated with Lrp12/Mig13a-negative cells and some extended neurites into the intermediate zone (stars in Fig. 6d).
To determine the polarity of Lrp12/Mig13a-positive neurons during preplate separation, we immunolabeled cells for the cis-Golgi matrix protein GM130 (Nakamura et al. 1995) in conjunction with the nuclear marker DAPI. To measure the orientation of the Golgi apparatus relative to the neuronal soma, we drew a line from the center of the nucleus to the GM130-immunopositive Golgi apparatus and mapped the orientation of the line relative to the radius of the cortical wall (Fig. 7e). At E12.5, the position of the Golgi apparatus of Lrp12/Mig13a-positive cells was oriented largely parallel to the pial surface (Fig. 7a,d). During the initial stage of preplate separation at E13.5, the majority of Lrp12/Mig13a-positive cells were radially aligned, with their Golgi apparatus positioned between the nucleus and the pial membrane (Fig. 7b,d) and a process extended toward the pia. These cells appeared to intercalate with the arriving cortical plate neurons, which were also radially aligned (Pinto-Lord et al. 1982). These data reveal for the first time a dynamic regulation of the polarity of future subplate neurons during preplate separation (Fig. 7f).
Subsequent to the appearance of the cortical plate, after E13 in mouse, cohorts of cells migrate through the intermediate zone and emerging subplate to the boundary of the cortical plate and marginal zone, where they form a series of neuronal layers. In the reeler mouse, however, later-generated neurons fail to migrate through the emerging subplate, leaving the preplate population as an intact “superplate.” Although cohorts of cells formed at the same time in the 2 genotypes give rise to the same classes of neurons, neurons are improperly oriented and the neuronal layers are inverted in the reeler mutant mouse (Caviness 1982; Goffinet 1984; Tissir and Goffinet 2003). These studies suggested a model in which the cortical malformation seen in reeler mice results from a failure of cortical plate formation. This failure has been attributed to defects in the cessation of the migration of later-generated neurons destined for the cortical plate (Caviness 1982; Goffinet 1984). To examine whether cell movements at an earlier phase, prior to cortical plate formation, were defective in reeler mice, we crossed transgenic Tg(Lrp12/Mig13a-Egfp) mice with mice lacking Reelin. Tg(Lrp12/Mig13a-Egfp) embryos with the reeler (rl/rl) mutation (B6C3Fea-Tg(Lrp12/Mig13a-Egfp)a-Relnrl/a-Relnrl) were identified by PCR (D'Arcangelo et al. 1996; Gong et al. 2003; Jossin and Goffinet 2007). We examined embryos between the ages of E12.5–15.5 and measured the distance of Lrp12/Mig13a-positive preplate neurons from the pial surface, comparing their distribution in rl/rl and wild-type littermates (Supplementary Fig. 7a–d). In wild-type cortex, Lrp12/Mig13a-positive cells progressively moved from subpial locations at E12.5 to deeper regions of cortex at E15.5, while in rl/rl cortex Lrp12/Mig13a-positive cells remained in subpial positions throughout the time periods examined, confirming that they did not form a subplate layer.
We next examined the dynamics of polarity of Lrp12/Mig13a-positive preplate neurons in rl/rl embryonic cortex. Temporal changes in the morphology and orientation of Lrp12/Mig13a-positive neurons were measured by determining the orientation of the Golgi apparatus relative to the radius of the cortical wall (Fig. 8e). At E12.5, a stage prior to preplate splitting, the shape and orientation of Lrp12/Mig13a-positive preplate neurons in rl/rl mutant mouse neocortex (Fig. 8a,d) were indistinguishable from their counterparts in wild-type neocortex (Figs 7a,d and 8d). However, at E13.5, just before the preplate splits, Lrp12/Mig13a-positive neurons in the rl/rl neocortex failed to adopt a bipolar shape and align with the radial axis (Fig. 8b,d) like their counterparts in wild-type neocortex. At E14.5, after the segregation of preplate neurons in the wild-type neocortex, small but significant differences were seen in the morphology and orientation of Lrp12/Mig13a-positive cells in the mutant superplate (Fig. 8c,d) and the wild-type subplate (Figs 7c,d and 8d). After the segregation of preplate neurons in the wild-type neocortex at E15.5, Lrp12/Mig13a-positive cells were randomly oriented both in the wild-type subplate and in the rl/rl superplate. These studies identify the earliest known defect in reeler cortical patterning and demonstrate that Lrp12/Mig13a-positive reeler preplate cells do not undergo temporal changes in neuronal morphology and orientation during early stages of corticogenesis.
Since genetic and pharmacological experiments have shown that inhibitors of SFKs generate a reeler-like cortical malformation (Hanke et al. 1996; Arnaud et al. 2003; Bock and Herz 2003; Jossin et al. 2003), we examined the effect of the SFK inhibitor PP2 on temporal changes in polarity of Lrp12/Mig13a-positive preplate neurons in developing Tg(Lrp12/Mig13a-Egfp) neocortex. When coronal slices of E12.5 Tg(Lrp12/Mig13a-Egfp) mouse forebrain were treated with PP2 for 2 DIV, the temporal changes in Lrp12/Mig13a-positive cell shape and polarity seen in control slice cultures did not occur (Supplementary Fig. 8). In addition, in PP2-treated slices, the preplate did not separate (Supplementary Figs 7e–g and 8c) and Lrp12/Mig13a-positive cells were randomly oriented (Supplementary Fig. 8c,d). These findings suggest that the neuropathological changes that occur in reeler or in mice lacking SFKs might result from a failure of preplate neurons to polarize toward the pial surface during the stage when preplate neurons segregate into the marginal zone and subplate layer (Fig. 8f).
These experiments demonstrate that preplate neurons that express Lrp12/Mig13a form specific axon projections and undergo previously unrecognized Reln-dependent changes in cell polarity just prior to cortical plate formation in developing neocortex. During early and mid-preplate stages, Lrp12/Mig13a-positive neurons extended pioneering axons that project to the contralateral hemisphere. Live imaging of Lrp12/Mig13a-positive, EGFP-labeled neurons in the early preplate revealed that preplate neurons migrated tangentially in the preplate layer. The mode of neuronal movement during these tangential migrations involved the rapid extension and retraction of a broad lamellipodial process. After E13.5, as the cortical plate emerged in the lateral telencephalic wall, these tangential movements paused as Lrp12/Mig13a-positive neurons polarized toward the pial surface. Subsequently, preplate neurons moved into positions above (marginal zone) or below (subplate) the cortical plate, with Lrp12/Mig13a-positive neurons drifting into the lower subplate region. In reeler mutant mice, this transient change in neuronal polarity did not occur and the preplate failed to separate. A similar defect in neuronal polarity occurred in Tg(Lrp12/Mig13a-Egfp) neocortex treated with PP2, an inhibitor of SFKs (Hanke et al. 1996). Taken together, these experiments reveal an earlier defect in reeler corticogenesis than previously realized and suggest that Reln-signaling-dependent changes in cell polarity are involved in the formation of a pseudolayer just prior to assembly of the cortical plate.
Preplate neurons that express the evolutionarily conserved gene Lrp12/Mig13a represent a subpopulation of future subplate neurons that establish connections with the contralateral hemisphere. These data highlight the little-appreciated heterogeneous nature of preplate and subplate neuron populations and provide a novel genetic marker to positively distinguish neurons pioneering projections to the anterior commissure from those pioneering projections toward the thalamus.
The alignment of Lrp12/Mig13a-positive preplate neurons with the radial axis resulted in the formation of a transient columnar arrangement during the segregation of preplate neurons into a superficial marginal zone and an underlying subplate. To our knowledge, this is the first description of the morphology of neurons destined for the subplate during cortical plate formation, and our observations suggest that these active changes in preplate/subplate neuron morphology and orientation might be required for the cortical plate neurons to efficiently migrate past subplate neurons to their final destination beneath the marginal zone. These results also suggest a potential role for neuron–glia interaction and signaling pathways in preplate cell polarity as radial alignment of neurons is largely dependent on their interactions with radial glia in the developing cortex (Rakic 1972).
Transient polarization of Lrp12/Mig13a-positive preplate neurons did not occur in reeler cortex or in cortical slices treated with the SFK inhibitor PP2. This lack of preplate neuron polarization was associated with a failure of preplate splitting. The failure of preplate neurons to polarize in the absence of RELN or SFK signaling could be a consequence of the defects in radial glial morphology previously described in reeler48 or a secondary consequence of the migratory defects observed in neurons of the reeler ectopic cortical plate. However, our data raise the intriguing possibility that RELN signaling between populations of preplate neurons is necessary for the observed changes in polarity and is an early patterning event required for normal cortical development.
The failure of Lrp12/Mig13a-positive preplate neurons to transiently polarize in the absence of RELN or SFK signaling further suggests that these pathways might converge on conserved polarity signaling pathways to control neuronal alignment and polarity during preplate splitting. Support for this idea is provided by recent studies on the role of the conserved mPar6α polarity signaling complex in the polarization and migration of neurons in the developing CNS (Arnaud et al. 2003; Shi et al. 2004; Solecki et al. 2004, 2006) and tyrosine phosphorylation of Par3 by an SFK signaling pathway during the formation of epithelia (Wang et al. 2006). Thus, SFK signaling pathways that involve conserved polarity proteins may control preplate neuron alignment into a pseudoepithelium, which we suggest may be required for proper formation of the cortical plate.
The development of the neocortex involves a series of discrete steps that include the migration of postmitotic neurons from generative zones, and the assembly of transient cell layers, which transform into a series of 6 neuronal layers. The present study reveals dynamic changes in preplate neuron movement and organization. At the preplate stage, neurons within the preplate undergo tangential migration. These migrations halt just prior to preplate separation, when preplate neurons form a columnar arrangement and Lrp12/Mig13a-expressing preplate cells move into the emerging subplate zone. These experiments show that Lrp12/Mig13a-expressing preplate cells constitute a subset of the preplate cell population that segregate into the subplate layer by an active process, which appears to involve polarity signaling pathways. We postulate that the transient polarization of Lrp12/Mig13a-positive preplate neurons allows radially migrating cortical plate neurons to move through the emerging subplate zone. In this model, defects in RELN signaling disrupt a sequence of preplate cell movements and changes in polarity required to coordinate the separation of the preplate and formation of the cortical plate during normal corticogenesis.
National Institutes of Health grant (RO1-NS-15429-27 to M.E.H.); Deutsche Forschungsgemeinschaft to S.S.
We are grateful to Drs Cori Bargmann, Nathaniel Heintz, Hilleary Osheroff, Shiaoching Gong, and Carla Shatz for critical comments on the experiments reported and to Drs Eve-Ellen Govek and Hilleary Osheroff for reading the manuscript. We thank John Baker and Nick Didkovsky for assembling time-lapse images into a MPEG-4 movie, Drs Tom Curran, Tom Jessell, and Jacqueline Trotter for providing anti-RELN, anti-TAG-1, and anti-L1 antibodies, and Dr Roger Tsien for providing tdTomato. We also thank Dr Alison North and Shivaprasad Bhuvanendran in the Bio-Imaging Resource Center at the Rockefeller University for expert assistance collecting multiphoton images and Nawshin Hoque for expert technical assistance. Conflict of Interest: None declared.