Malformations of cortical development can arise when projection neurons generated in the germinal zones fail to migrate properly into the cortical plate. This process is critically dependent on the Reelin glycoprotein, which when absent leads to an inversion of cortical layers and blurring of borders. Reelin has other functions including supporting neuron migration and maintaining their trajectories; however, the precise role on glial fiber-dependent or -independent migration of neurons remains controversial. In this study, we wish to test the hypothesis that migrating cortical neurons at different levels of the cortical wall have differential responses to Reelin. We exposed neurons migrating across the cortical wall to exogenous Reelin and monitored their migratory behavior using time-lapse imaging. Our results show that, in the germinal zones, exogenous Reelin retarded neuron migration and altered their trajectories. This behavior is in contrast to the response of neurons located in the intermediate zone (IZ), possibly because Reelin receptors are not expressed in this zone. In the reeler cortex, Reelin receptors are expressed in the IZ and exposure to exogenous Reelin was able to rescue the migratory defect. These studies demonstrate that migrating neurons have nonequivalent responses to Reelin depending on their location within the cortical wall.
Abnormal neuron migration in the developing cortex can cause disabilities, such as mental retardation and epilepsy (Ross and Walsh 2001; Francis et al. 2006; Guerrini and Marini 2006). Studying and understanding neuronal migration are vital to the prevention of these disorders. Normal neuron migration may be divided into 3 consecutive phases: The generation of neuroblasts within the germinal zones, migration of immature neurons toward the pial surface and final positioning in the cortical plate (CP; Rakic 1988, 2009). These processes work in sequence and unison to place neurons where they belong in an inside-out manner (Takahashi et al. 1999). Disorders of migration are caused by neurons failing to reach their final positions either by remaining at the ventricular surface (periventricular heterotopia), stalling in the white matter (subcortical band heterotopia), or by just forming a disordered CP (Manzini and Walsh 2011).
Endeavours to understand this process through direct imaging of cortical neuron migration has identified key transition points along the migratory route (Rakic et al. 1974; Noctor et al. 2001; Tabata and Nakajima 2003; Noctor et al. 2004; LoTurco and Bai 2006). First, neurons from the last cell division assume a bipolar morphology and migrate along the radial glial fiber. Secondly, the neuron converts to a transitory multipolar morphology within the subventricular zone (SVZ) before reverting to a bipolar shape again in the intermediate zone (IZ) prior to advancement into the CP. Thirdly, the neuron anchors its leading process to the marginal zone (MZ) before “boot-strapping” its cell body up to the top of the CP in a glial-independent manner (somal translocation; Miyata et al. 2001; Nadarajah et al. 2001; Borrell et al. 2006). These coordinated events are controlled by a variety of molecular signals, one of these being the signaling protein Reelin.
Reelin loss leads to disordered neuron positioning in laminated structures of the brain, including the cerebral cortex, hippocampus, and cerebellum (Caviness and Sidman 1973; Caviness and Rakic 1978). In the cortex, the early born neurons fail to split the preplate and new neurons fail to migrate past their predecessors resulting in a disorganized and inverted cortex (Caviness 1982; Goffinet 1984; Sheppard and Pearlman 1997). How Reelin affects neuron positioning is still a subject of debate. Reelin is highly concentrated in the MZ, but also present in the IZ and SVZ, and in the postnatal cortex, in layer V neurons (Alcantara et al. 1998; Yoshida et al. 2006). Therefore, Reelin is well positioned to direct cortical neuron migration and has been implicated in all stages of the migratory process; originating with preplate splitting (Pinto-Lord and Caviness 1979; Sheppard and Pearlman 1997; Magdaleno et al. 2002; Jossin et al. 2004; Nichols and Olson 2010), altering the rate of migration (Hammond et al. 2001; Sanada et al. 2004; Feng et al. 2007; Britto et al. 2011), polarization of multipolar migration (Jossin and Cooper 2011), adhesion and detachment from the radial glial fiber (Pinto-Lord et al. 1982; Hoffarth et al. 1995; Sanada et al. 2004; Voss et al. 2008), and termination of migration (Super et al. 2000; Herrick and Cooper 2004; Feng et al. 2007; Franco et al. 2011; Sekine et al. 2011). However, when Reelin is absent, it has multiple consequences and it is unclear how each of the above processes is directly affected.
A further aspect of the Reelin phenotype that has been less well canvassed is the crossing of aberrant neurons into adjacent layers, resulting in blurred cortical layer boundaries (Hevner et al. 2003; Dekimoto et al. 2010). This suggests that the role of Reelin in organizing neurons into layers may be separate, and independent to affecting neuron migration per se. Supporting evidence is seen in the analysis of single mutants of the Reelin receptors “very low-density lipoprotein receptor” (VLDLR) and “Apolipoprotein E receptor 2” (ApoER2; D'Arcangelo et al. 1995; Howell et al. 1997; Sheldon et al. 1997; Trommsdorff et al. 1999; Hack et al. 2007). Parallel to this, mutants of the downstream activators of Reelin signaling, the Fyn and Src kinases, display varying cortical layering defects. (Soriano et al. 1991; Stein et al. 1994; Yuasa et al. 2004; Kuo et al. 2005). Considered together, these observations indicate that the reeler phenotype may be the sum of multiple effects and, depending on the developmental context, may be manifested as glial-dependent or glial-independent migration (Cooper 2008; Zhao and Frotscher 2010; Franco et al. 2011).
Within the cortical wall itself, the high concentration of Reelin in the MZ has biased the debate toward CP-centric theories of Reelin function. However, outside of the CP, Reelin may have additional roles, evidenced by the IZ and SVZ expression of Reelin receptors VLDLR and ApoER2, and Dab1 (Forster et al. 2002; Hartfuss et al. 2003; Luque et al. 2003; Meyer et al. 2003; Uchida et al. 2009). What may be the function of Reelin outside the MZ? Exogenous expression of Reelin in the ventricular zone (VZ) of Nestin-reelin transgenic mice is capable of rescuing the preplate splitting defect in reeler (Magdaleno et al. 2002), but not layer inversion. Interestingly, Reelin and Notch signaling have been shown to regulate the pace of neurogenesis within the VZ (Lakoma et al. 2011), and in utero electroporation of exogenous Reelin into the VZ at midcorticogenesis can induce discrete spherical cell aggregates with cellular orientation reminiscent of the MZ (Kubo et al. 2010). Recently, we showed that, in reeler, young neurons in the VZ/SVZ behave aberrantly with altered speeds of migration and abnormal radial trajectories, suggesting that Reelin may be important for promoting directionality (Britto et al. 2011). This view is corroborated by the discovery that Reelin can influence the intracellular location of Rap1 in the lower IZ, leading to polarized neuron migration toward the CP (Jossin and Cooper 2011).
In the present study, we further interrogate the functions of Reelin in the VZ/SVZ by exposing slice cultures to exogenous Reelin and address whether exogenous Reelin is capable of influencing migration near the germinal zones. We discovered that exogenous Reelin caused many neurons to assume a multipolar morphology and, despite highly motile activity revealed by time-lapsed imaging, these multipolar neurons did not exhibit an outward trajectory. These effects appeared to be mediated by functional Reelin signaling, as neurons in Dab-null cortices were refractory to exogenous Reelin under similar conditions. In addition, we specifically asked whether neurons entering the CP could respond to exogenous Reelin, as neurons are known to “pile up” underneath the CP in the reeler cortex. Our results demonstrate that exogenous Reelin is capable of rescuing the reeler phenotype by promoting migration into the CP. In conclusion, exogenous Reelin provides a useful tool for studying the differential behaviors of migrating neurons within cortical regions to shed further insights on how Reelin controls neuronal positioning during development.
Materials and Methods
Experiments were undertaken with the approval of The Florey Institute of Neuroscience and Mental Health Animal Ethics Committee and conform to the Australian Code of Practice for the Care and Use of Animals for Scientific Purposes (7th Ed, 2004). Relnrl/J (reeler) mice were bred in standard conditions, and genotyping was carried out using polymerase chain reaction (PCR) primers designed by D'Arcangelo et al. (1996) and conditions as previously described (Britto et al. 2011).
In Utero Labeling of Neurons
Replication-incompetent enhanced green fluorescent protein (GFP)-expressing retrovirus was produced from a stably transfected packaging cell line 293gp NIT-GFP (kind gift from F. Gage). Cells were transiently transfected with glycoprotein of the vesicular stomatitis virus and virus harvested as previously described (Palmer et al. 1999). Morning of vaginal plug detection was defined as embryonic day 0.5 (E0.5), and E14.5 pregnant dams were anesthetized with pentobarbitone (Ilium) and prepared for surgery. Injections of retrovirus GFP were made into the lateral ventricle using a glass micropipette. For in utero electroporations, the chicken beta actin promoter (CAG)-GFP plasmid (Matsuda and Cepko 2004; Addgene plasmid 11150) was injected in the ventricle of E14.5 embryos using a glass micropipette. DNA was visualized using 0.01% Fast Green and electroporated into the cortex with five 50-ms square pulses at 40 V administered at 950-ms intervals using the CUY21EDIT square-wave pulse generator. Embryos were allowed to develop for 24–48 h after surgery before the collection of brain slice cultures.
Electroporation of Brain Slices
Brains from E15.5 embryos were dissected, embedded in 3% low melting point (LMP) agarose, sectioned into 400-μm coronal slices, and collected into artificial cerebral spinal fluid (ACSF). Slices were transferred to an electroporation chamber (NEPA GENE, Japan), where the CAG-GFP plasmid (Matsuda and Cepko 2004) visualized using 0.01% Fast Green (Sigma) was placed in the ventricle using a glass micropipette. DNA was electroporated into the brain slice with five 50-ms square pulses at 30 V administered at 950-ms intervals using the CUY21EDIT square-wave pulse generator (NEPA GENE). Slices were transferred onto millicell inserts and placed in culture media.
Production and Placement of Exogenous Reelin
For the expression of recombinant Reelin, Hek293T cells were transfected with 0.5–1 μg of pCrl (kind gift from T. Curran) or 0.5–1 μg of vector pcDNA3.1 using Lipofectamine 2000 (Invitrogen). One day after transfection, cells were trypsinized, resuspended, and cell pellets created using the hanging drop cell aggregation method. Alternatively, cells were resuspended in a solution containing 20 μl of culture medium and 80 μl of 2% LMP agarose in culture medium. The cell–agarose suspension was set into molds, and the resulting cell block was sectioned at 200 μm and placed as an overlay onto the brain slice cultures. For each transfection, duplicates were produced to confirm transfection efficiency with immunohistochemistry on the HEK293T cells and western blots for secreted Reelin protein in the media.
Time-Lapse Imaging of Coronal Brain Slices and Analysis of Migratory Kinetics
Embryos were collected in ACSF for organotypic slice culture as previously described (Britto et al. 2011). Images were captured using a ×10 objective at 15-min intervals over 12 h periods, and each neuron was manually tracked using in-house software developed in MATLAB® (Britto et al. 2009, 2011). The tracking of somal coordinates through the imaging sequence was used to derive a number of parameters: (1) The net speed of the neuron from its start and end position, (2) the accumulated speed or overall speed of the neuron regardless of the direction of migration, and (3) the trajectory for each neuron in reference to the ventricular surface.
Immunohistochemistry, Western Blotting and Phalloidin Staining
HEK293T cells were fixed in and washed in 0.1 M phosphate buffer (PB) with 0.1% Triton X-100. Cells were exposed to the mouse monoclonal anti-Reelin G10 (1:1000, Chemicon), overnight at 4°C followed by 2–3 h incubation in AlexaFluor™594 (1:500, Molecular Probes). Slices were fixed in 4% paraformaldehyde/0.1 M PB, permeabilized with 0.3% Triton X-100, and incubated overnight with the Nestin (1:200, Millipore) and phospho-Cofilin antibody (Ser3, 1:1000, Santa Cruz Biotechnology). Western blots were performed with the Reelin antibody (1:1000, Millipore) to detect the secretion of Reelin from transfected cells into the medium. For intracellular signaling analysis, protein lysates of cortical slices were prepared using radio immunoprecipitation assay buffer and western blots conducted with β-actin (1:1000, Sigma) and phospho-Src Tyr 416 (1:1000, Cell Signaling). Signal was detected using horseradish peroxidase-coupled goat antimouse antibodies and ECL detection reagent (Amersham). To visualize cellular architecture, tissue was incubated with tetramethylrhodamine isothiocyanate-Phalloidin (Sigma) to label filamentous actin.
Alkaline Phosphatase Fusion Probe Production and Detection
Alkaline phosphatase (AP) fusion proteins were created and used as previously described (Uchida et al. 2009). In brief, HEK293T cells were transfected with fusion constructs using Lipofectamine 2000. Supernatants were collected at 48 and 72 h, filtered, and buffered with 10 mM 4-(2-hydroxyethyl)-1-piperazineethanesulfonic acid. Embryonic brains were cut on the vibratome into 300-μm thick sections of unfixed tissue, incubated with AP fusion proteins for 1 h, and processed for AP detection.
Data and Statistical Analysis
For the long-term migratory analysis, a 2-way analysis of variance (ANOVA) with Bonferroni post hoc tests was performed between overlay (vector and Reelin) and cortical domains (VZ, SVZ, IZ, and CP). For the Dab-1 bin data, a 2-way ANOVA consisting of one between genotypes and another between bins was performed. All other statistical analyses were performed using unpaired, 2-tailed Student's t-test.
Exogenous Reelin Induces a Multipolar Morphology and Alters the Speed of Migration
To manipulate the level of exogenous Reelin and aid the discrete placement of the cell overlay across the cortical wall, HEK293T cells transfected with Reelin or vector (pcDNA3.1) constructs were embedded into low-melting point agarose blocks (Fig. 1A). Reelin expression and release into the media was confirmed by immunohistochemistry on cells in a monolayer (Fig. 1B), cells embedded in the agarose (Fig. 1C), and by western blot analysis of the culture media (Fig. 1D). Cell viability and secretion of Reelin were detected in the medium of agarose-embedded cells cultured after 24 h (data not shown) and in all cases, only the full-length Reelin (>225-kDa marker) was detected in the media. In biochemical terms, there is evidence that the Reelin overlay is capable of activating the Reelin signaling pathway in recipient cells. Src family kinases are activated by phosphorylation of tyr416 in neurons exposed to Reelin (Arnaud et al. 2003; Bock and Herz 2003). This was confirmed in our system by western blot analysis of cortical lysates prepared from slices incubated with either a Reelin or vector overlay. The results showed an increase in the levels of phospho-Src (Tyr 416) from Reelin overlay, but not from vector overlay, slices (Fig. 1D). To eliminate any extraneous factors produced by HEK293T, all data comparisons were conducted between slice cultures with Reelin-transfected cell overlay and vector-transfected cell overlay (referred to as Reelin and vector overlays, respectively). This method of Reelin production has been used effectively in previous studies to evaluate Reelin-induced cellular responses (Howell et al. 1999; Zhao et al. 2004; Kubo et al. 2010).
To monitor the response of cells in the VZ/SVZ to an exogenous source of Reelin, GFP was introduced by in utero electroporations of CAG-GFP (E14.5) prior to harvesting cortical slices at E15.5 and placement of the Reelin or vector overlay for the entire duration of the culture period (schematic Fig. 1E). In the presence of a vector overlay, electroporated GFP+ cells were located in the VZ while outwardly migrating cells transitioned into multipolar cells (arrows) to form the distinct region of the SVZ (Fig. 1F). In contrast, the presence of a Reelin overlay altered the morphology of GFP+ cells to become predominantly multipolar and remain localized in the VZ close to the ventricular surface. Few GFP+ cells separated into the SVZ (Fig. 1G). Labeling of neurons by electroporation leads to an abundance of GFP+ cells that cannot be easily distinguished for the examination of migratory dynamics. To circumvent this problem, in utero labeling using retroviral GFP was performed, and cortical slices containing GFP+ cells were overlayed with either a Reelin or vector overlay (Fig. 1H,I). Radial fibers were labeled throughout the length of the cortical wall, with no discernible difference in morphology between Reelin and vector overlays after 18–24 h of culture. In confirmation, cells exposed to exogenous Reelin exhibited retarded migration, residing in the VZ with distorted morphology and directionality (Fig. 1I).
The loss of radial orientation and change in morphology warranted further investigation of individual migratory behaviors. Time-lapse analysis of retroviral-labeled slices confirmed a loss of radial trajectory of GFP+ cells in the presence of a Reelin overlay, compared with neurons in vector-transfected slices (Fig. 2A,B; Supplementary Movies 1 and 2). These neurons were overwhelmingly multipolar in morphology with a tendency to cluster together (Fig. 2B). Imaging and analysis were conducted as previously described (Britto et al. 2011). Two distinct forms are discernible by morphology during the imaging period; the bipolar cell, with 1 or 2 processes (Fig. 2A), and the multipolar cell that has multiple processes in all orientations (Fig. 2B). A third category, transition cells were those utilizing both modes of migration during the tracking period. The 1.4-fold significant increase in the proportion of cells with a multipolar morphology was more pronounced at higher concentrations of transfected Reelin (0.5 or 1.0 μg cDNA; Fig. 2C), with these cells being located abnormally in the VZ/SVZ (vector n = 3 slices, 112 cells; 0.5 μg Reelin n = 2 slices, 38 cells; and 1.0 μg Reelin n = 3 slices, 107 cells). The increase in multipolar cells was at the expense of bipolar cells that showed a 4-fold reduction in number (Fig. 2C, *P < 0.05, Student's t-test). To determine whether the induced change in morphology was causing a conversion to a basal progenitor fate, immunohistochemistry with Tbr2 was performed (kind gift from R. Hevner) and showed no alteration in Tbr2+ cell distribution (data not shown).
This type of short-term imaging analysis (using the GFP marker) does not readily distinguish between cells that migrate from the VZ to become basal progenitors in the SVZ. However, a recent study demonstrated that a distinction is possible based on the mode of migration and migratory speed (Tabata et al. 2009). This study showed that basal progenitors exit the VZ at greater speeds with the morphology of cells undergoing somal translocation, whereas multipolar cells have slower migration speed and accumulated in the lower part of the SVZ prior to migration toward the CP. In our analysis, a bipolar cell exiting the VZ is classified as having 2 process extensions parallel to the radial fiber, reflecting active locomotion. These cells possess a leading and trailing process and, therefore, unlikely to be cells undergoing interkinetic nuclear migration (IKNM) or division. The ability to dynamically follow cell morphology and behavior in real time permits a differentiation between progenitors undergoing IKNM and postmitotic neurons migrating toward the CP.
Another prominent feature of GFP+ cells within a Reelin-induced cluster is that migrating cells possessed highly motile processes despite an overall significant 1.9-fold decrease in net speed (vector overlay compared with 1.0 μg Reelin overlay). Interestingly, only multipolar, not bipolar, cells showed this reduction in migratory speed across both Reelin doses (Fig. 2D, vector n = 3 slices, 112 cells; 0.5 Reelin n = 2, 38 cells; and 1.0 Reelin n = 3, 107 cells; speed data presented in the figure legend). Net speed is the net vector distance over imaging time; while accumulated speed reflects total distance covered. The reduction in accumulated speed is not as pronounced (1.2-fold) highlighting the loss of directionality of multipolar migration (Fig. 2E).
Although the number of bipolar cells decreased in the presence of exogenous Reelin, no significant change was recorded in the rate of migration. To correlate migratory speeds with their trajectories, bipolar and multipolar cells were classified as exhibiting a positive or negative trajectory at the end of the time-lapse period, with a positive trajectory meaning migration toward the CP and a negative trajectory indicating movement toward the ventricle. This analysis revealed that, in the presence of exogenous Reelin, a greater proportion of multipolar neurons (compared with bipolar neurons) assumed negative trajectories (Fig. 2F,G). This trend was particularly robust at the higher dose of exogenous Reelin.
Given that exogenous Reelin caused multipolar cells to assume lower speeds of migration, we tested to see if this lead to an overall reduced migratory distance. Using retroviral-labeled cortices, the distribution of GFP+ cells across the cortical wall (in the presence of both Reelin and vector overlays) was recorded every 12 h over a 72-h culture period (vector n = 5 slices, 1432 cells and Reelin n = 5 slices, 791 cells; Fig. 3). In the presence of a vector overlay, the proportion of GFP+ cells in the CP increased over time, resulting in an average proportion of 35% of all labeled cells (Fig. 3C,E). The proportion of cells remaining in the IZ was consistent over time and represented approximately 40% of total GFP+ cells. In contrast, cells exposed to a Reelin overlay displayed a significantly reduced number of cells in the IZ at the 24-h time-point; however, this value increased to 48% (1.8-fold) after 72 h in culture (Fig. 3D,F). In addition, significantly reduced proportions of GFP+ cells migrate further into the CP after 48 h in culture (Fig. 3F and Table 1). Therefore, exogenous Reelin results in retarded migratory distances as well as reduced migratory speeds.
|24 h||48 h||72 h|
|VZ||0.13 ± 0.01||0.24 ± 0.04||0.09 ± 0.01||0.18 ± 0.05||0.11 ± 0.01||0.16 ± 0.05|
|SVZ||0.27 ± 0.05||0.46 ± 0.05+||0.21 ± 0.02||0.27 ± 0.06||0.13 ± 0.02||0.18 ± 0.05|
|IZ||0.46 ± 0.05||0.22 ± 0.02++||0.40 ± 0.07||0.45 ± 0.06||0.42 ± 0.04||0.49 ± 0.04|
|CP||0.15 ± 0.06||0.05 ± 0.03||0.30 ± 0.06||0.10 ± 0.04+||0.35 ± 0.06||0.17 ± 0.05+|
|24 h||48 h||72 h|
|VZ||0.13 ± 0.01||0.24 ± 0.04||0.09 ± 0.01||0.18 ± 0.05||0.11 ± 0.01||0.16 ± 0.05|
|SVZ||0.27 ± 0.05||0.46 ± 0.05+||0.21 ± 0.02||0.27 ± 0.06||0.13 ± 0.02||0.18 ± 0.05|
|IZ||0.46 ± 0.05||0.22 ± 0.02++||0.40 ± 0.07||0.45 ± 0.06||0.42 ± 0.04||0.49 ± 0.04|
|CP||0.15 ± 0.06||0.05 ± 0.03||0.30 ± 0.06||0.10 ± 0.04+||0.35 ± 0.06||0.17 ± 0.05+|
VZ: ventricular zone; SVZ: subventricular zone; IZ: intermediate zone; CP: cortical plate.
+P < 0.05; ++P < 0.01.
The influence of Reelin on radial glial fiber formation and morphology (Pinto-Lord et al. 1982; Hartfuss et al. 2003; Hunter-Schaedle 1997) has been reported previously. To assess the morphology of glial fibers in our long-term culture system, immunohistochemistry for Nestin was conducted (Fig. 3G–L). Comparable radial alignment was observed in the IZ with GFP+ cells aligning individual fibers in either the presence of a vector or Reelin overlay. Our results imply that, during midcorticogenesis (E15.5), exogenous Reelin has no effect on preexisting glial fiber morphology, and thus not a cause of disrupted migration.
Reelin-Induced Migratory Defects are Dependent on the Presence of Dab-1
Thus far we have shown that exogenous Reelin caused multipolar cells to be retarded in the distance of migration and induced altered trajectories. To test that this assay is a reflection of altered Reelin signaling, we performed identical experiments on Dab-1 knock-out (Dab-1) cortical brain slices. Wild-type (Fig. 4A–C) and Dab-1 slices (Fig. 4D–F) were collected at E14.5, electroporated with CAG-GFP, and cultured for 40 h in the presence of either a Reelin or vector overlay. The results showed that, after 24 h, wild-type cells were retarded in their migratory distance by the Reelin-containing overlay, while Dab-1 cells appear unaffected (Fig. 4F). Quantitative assessment for each condition was obtained by scoring the proportion of GFP+ cells in 100 μm bins between the SVZ and pial surface (Fig. 4G,H). The migratory distances of wild-type cells with vector overlay (n = 6 slices, 1306 cells) were not significantly different to Dab-1 cells with Reelin overlay (n = 3, 453 cells; non significant, ANOVA). In contrast, wild-type cells with Reelin overlay accumulated in the lower bins and exhibited a decrease in the upper bins 5–7 (n = 5, 405 cells; P < 0.01, ANOVA), when compared with Dab-1 cells with Reelin overlay (n = 5, 833 cells). This analysis shows that the migratory behavior of Dab-1 cells are unaffected by exogenous Reelin and suggests that our previous findings are attributed to Reelin signaling.
Exogenous Reelin Rescues the Migratory Defect in the Reeler Cortex
Our analysis has demonstrated that exogenous Reelin has a negative effect on cells migrating in the germinal zones of the VZ/SVZ. Cells in these regions are known to undergo a bipolar/multipolar transition, and our results have revealed a preponderance of multipolar cells in the presence of a Reelin overlay. However, beneath the CP, Reelin has been suggested to be vital for orientating neurons from multipolar to bipolar morphology prior to entry into the CP (Jossin and Cooper 2011). If this is correct, then exogenous Reelin would be expected to rescue these neurons in the reeler, where there is an overwhelming failure of migration, resulting in a “traffic jam” beneath the CP. To test this hypothesis, we subjected both heterozygous and reeler cortical slices to exogenous Reelin overlays and compared the migration of cortical cells using time-lapse imaging. To capture cells migrating from the upper IZ into the CP, CAG-GFP was introduced by in utero electroporation at E14.5, and embryos were left for 48 h before slices were prepared at E16.5 and cultured for 16 h in the presence of a Reelin or vector overlay. Time-lapse imaging was conducted at 15-min intervals over a 24-h culture period, and an end-point image is shown in Figure 5A–D. In heterozygous slices (Fig. 5A,B; Supplementary Movie 3), although the morphology and directionality of cells were distorted, no significant difference was observed in the migratory speed of GFP+ cells between vector and Reelin overlays (Fig. 5H). In contrast, reeler slices with vector overlay showed slower neuron migration (up to 3-fold speed reduction compared with heterozygous) with accumulation of GFP+ cells beneath the CP, indicative of Reelin deficiency (Fig. 5C and Supplementary Movie 4). However, these neurons appeared to be robustly rescued in speed in the presence of a Reelin overlay (Fig. 5D and Supplementary Movie 5) with GFP neurons migrating at speeds comparable with heterozygous cortical slices (Fig. 5H; speed data presented in the figure legend). Considered together, these experiments suggest that exogenous Reelin is capable of rescuing the “CP phase” of migration that is defective in reeler, resulting in accumulation of neurons underneath the CP.
Superficially, the above results appear at odds with the migratory behavior of cortical neurons located in the germinal zone, where wild-type cells respond to exogenous Reelin by reducing their migratory speeds and distances (Fig. 2). What might be the underlying bases of this discrepancy? We have shown that Reelin signaling is active in the cortex (pSrc levels, Fig. 1); however, our western result provides no indication of intracellular responses from cells located in either the VZ or CP. We know that Reelin signaling leads to serine3 phosphorylation of n-Cofilin, an actin-depolymerizing protein, that promotes disassembly of F-actin (Chai et al. 2009). Utilizing this discovery, we conducted phospho-Cofilin (ser3) immunohistochemistry on heterozygous and reeler cortical slices exposed to conditioned media containing Reelin (CM-Reelin) or vector alone (CM-vector, Supplementary Fig. 1). Within the germinal zone, both heterozygous and reeler slices show minimal staining for phospho-Cofilin (CM-vector) with a marginal increase observed after exposure to Reelin (Supplementary Fig. 1B,F). More notable is the altered phospho-Cofilin staining in the reeler cortex of CP cells exposed to Reelin. Under vector conditions, diffuse staining was observed near the cortical surface, a result consistent with Chai et al. (2009) (Supplementary Fig. 1G); however, this is markedly reduced when exposed to CM-Reelin (Supplementary Fig. 1H). A decrease in phospho-Cofilin levels is associated with increased migration (Chow et al. 2011; Simard et al. 2011), thus corresponding with the increased migratory speed we have shown when reeler cells entering the CP are exposed to exogenous Reelin.
This change in phospho-Cofilin response may also reflect differential distribution of functional Reelin receptors across the cortical wall. We performed AP reporter assays using an AP fusion with the central fragment of Reelin R3-R6 and corresponding AP control construct (Fig. 5E–G). These constructs have previously been used to demonstrate the presence of functional Reelin receptors in the SVZ/IZ domains (Uchida et al. 2009). While a strong AP signal was detected within the SVZ/IZ in the heterozygous cortex (Fig. 5E), a more widespread AP signal was detected across the reeler cortex (Fig. 5F), suggesting that, in reeler, functional Reelin receptors are redistributed across the cortical depth. Control staining using AP alone is provided in Figure 5G. The redistribution could be a primary genetic response due to defective Reelin signaling or, alternatively, be a reflection of malpositioned neuron positions due to the reeler defect. In either event, the shift of Reelin receptors toward the superficial aspects of the cortex can account for the observed rescue of reeler cells entering into the CP.
The present study describes the use of time-lapse imaging to study the effects of exogenous Reelin (delivered by cell overlays) on neurons migrating in different cortical locations at different time-points. Although the results presented here are from a single developmental stage (E15.5), this assay can be readily adapted to test any migratory population during corticogenesis. This is an advantage over transgenic over-expression of Reelin, where Reelin availability is subject to constrains of promoter choice and cell-type specificity (Magdaleno et al. 2002; Kubo et al. 2010). Although Reelin availability may be considered nonphysiological, it nevertheless permits gain-of-function tests to be conducted on all neurons migrating in the same cortical slice. This study extends previous studies using transfected HEK293T cells to deliver Reelin in CM, or as cell cocultures followed by end-point analysis (Howell et al. 1999; Brandes et al. 2001; Kubo et al. 2010). In contrast, we have gone further by using time-lapse imaging to record the migratory responses of neurons exposed to exogenous Reelin, revealing dynamic aspects of neuron migration not detectable in fixed preparations. We also demonstrate that these responses are dependent on activation of the Reelin signaling pathway by using Dab1 mutant controls.
The above caveats notwithstanding our results demonstrate that migrating neurons situated at different levels of the cortical wall have nonequivalent responses to exogenous Reelin. Neurons in the germinal zone responded by reducing their migratory speeds and tended to adopt multipolar morphologies. Given that neurons in these locations normally undergo bipolar to multipolar transitions, excess Reelin appears to promote bipolar to multipolar conversion, or alternatively favor neurons remaining in the multipolar state. These observations are consonant with the opposite phenomenon observed in reeler slices of the same age, where reeler neurons in the VZ/SVZ migrated at excessive speeds, leading to greater distance traveled (Britto et al. 2011). Thus, Reelin hyperfunction appears to stimulate migration in certain neurons but not others, a recurring theme reported by different laboratories working on a range of models (Super et al. 2000; Hammond et al. 2001; Bock and Herz 2003; Jossin 2004; Feng et al. 2007; Hashimoto-Torii et al. 2008).
The observed alterations in morphology and migratory speed in VZ cells are indicative of perturbed migratory guidance cues (Valiente and Marin 2010). Accumulation of neurons with multipolar morphology in the SVZ and IZ has previously been observed following disruption of chondroitin sulfate signaling in the VZ (Ishii and Maeda 2008). In addition, transplantation of Reelin-expressing cells led to the clustering of neurons (Kubo et al. 2010). Recent evidence has disclosed multifaceted roles for Reelin in providing directional cues to polarize neuron orientation, radial migration, and dendrite sprouting (Matsuki et al. 2010; Nichols and Olson 2010; Jossin and Cooper 2011). Although our application of excess Reelin is nondirectional, the induced elaborate extension and retraction of cellular processes and reduction in net movement are indicative of a loss of directionality.
A major surprise is that neurons situated beneath the CP (compared with those in the VZ) did not respond in a similar manner to exogenous Reelin. Here, neurons are known to undergo a second transition multipolar to bipolar conversion stage before migrating into the CP (Tabata and Nakajima 2003; Noctor et al. 2004). In these neurons, we observed that exogenous Reelin did not reduce their migratory speeds; conversely in the reeler cortex, retarded neurons beneath the CP were rescued by increasing migratory speed and distance into the CP. How is it that exogenous Reelin can cause retarded migration into the VZ/SVZ neurons but have no effect on neurons situated beneath the CP? There are 2 issues to consider here. First, we found the distribution of Reelin receptors (detected by Reelin R3-6-AP binding) is dissimilar in the 2 genotypes. In the heterozygous cortex, Reelin-AP binding was detected in the MZ and SVZ/IZ [this study and Uchida et al. 2009]. In contrast, the reeler cortex showed diffused and uniform binding of Reelin-AP across the entire cortical wall. Thus, the differential pattern of Reelin receptors across the cortical depth of the heterozygote might be responsible for the observed sensitivities of migrating neurons in different locations to exogenous Reelin. It would also explain why in the reeler, transitional neurons lying beneath the CP are capable of being rescued. Further understanding of the mechanisms would require information regarding the precise binding locations of exogenous Reelin to their receptors. Secondly, transitional neurons lying beneath the CP are uniquely regulated by various genes (Bielas et al. 2004; LoTurco and Bai 2006) and are known to revert into bipolar shapes with their leading processes anchored to the MZ. This physical arrangement prepares them for the final stage of migration by somal translocation, a Reelin-dependent process (Olson et al. 2006; Cooper 2008; Franco et al. 2011; Sekine et al. 2011). Therefore, it is not surprising that exogenous Reelin was capable of effecting full rescue of neurons bereft of Reelin signaling and arrested in this zone (Morimura and Ogawa 2009; Simo et al. 2010).
Our results provide further impetus to revise the traditional role of Reelin as being solely involved in organizing cortical lamination and neuron positioning (Rice and Curran 2001). It is possible that Reelin has additional roles during cortical development given that Reelin, and its receptors, is transiently expressed in multiple locations in the developing cortex (Alcantara et al. 1998; Bar et al. 2003; Luque et al. 2003; Yoshida et al. 2006; Uchida et al. 2009). The presence of Notch and Reelin-dependent signaling in the germinal zones (Hashimoto-Torii et al. 2008; Keilani and Sugaya 2008) and the recent evidence of Reelin regulating proliferation (Lakoma et al. 2011) add support to this notion. Additionally, the expression of Reelin receptors and Dab-1 in the VZ of the human fetal cortex during early gestation suggests that this function is ancient and conserved (Meyer et al. 2003; Perez-Garcia et al. 2004; Cheng et al. 2011).
In conclusion, the Reelin gene has been identified for nearly 2 decades now (D'Arcangelo et al. 1995; Hirotsune et al. 1995), and yet we are no nearer to understanding its precise function in cortical layering. The initial concept of Reelin, an organizer of cortical layering, was based on the observation of layering defects in the cortex and cerebellum of the null mutant; however, layer inversion may be derivative and secondary to a range of failures in neuron migration (Caviness and Rakic 1978; Caviness et al. 2008; Marin et al. 2010). In this context, reductionist studies such as the present one are useful in piecing together different aspects of Reelin function on different neuronal populations.
The NHMRC and the Victorian Government's Operational Infrastructure Support Program supported this work.
We wish to thank Brian Howell for the gift of Dab-1 mutant mice, Tom Curran for the pCrl plasmid, André Goffinet for the E4 antibody, Robert Hevner for the Tbr2 antibody, and Fred Gage for the retroviral-GFP construct and packaging line. Conflict of Interest: None declared.