Reelin is an important protein that is indispensable for cortical lamination. In the absence of Reelin, cortical layers fail to form due to inappropriate neuron migration and positioning. The inversion of cortical layers is attributed to failure of neurons to migrate past earlier-generated neurons although how Reelin-insufficiency causes this is unclear. The issue is complicated by recent studies showing that very little Reelin is required for cortical layering. To test how variation in the number of Reelin-producing cells is linked to cortical lamination, we have employed Reelin+/+ ↔ Reelin−/− chimeras in which the number of Reelin-expressing neurons is adjusted. We found that the Reeler phenotype was rescued in chimeras with a large contribution of Reelin+/+ neurons; conversely in chimeras with a weak contribution by Reelin+/+ neurons, the mutant phenotype remained. However, increasing the number of Reelin+/+ neurons beyond an unknown threshold resulted in partial rescue, with the formation of a correctly layered secondary cortex lying on top of an inverted mutant cortex. Therefore, the development of cortical layers in the correct order requires a minimal level of Reelin protein to be present although paradoxically, this is insufficient to prevent the simultaneous formation of inverted cortical layers in the same hemisphere.
The arrangement of cortical neurons into discrete layers is dependent on the presence of the secreted glycoprotein Reelin (Rice and Curran 2001; Tissir and Goffinet 2003). Being produced by Cajal–Retzius (CR) cells, the protein is highly concentrated in the marginal zone (MZ), but despite its remote location, it appears to have a signaling effect on neurons and radial glia situated at some distance away (Hartfuss et al. 2003; Perez-Garcia et al. 2004). Reelin receptors and their cytoplasmic effector Disabled-1 (Dab1) are expressed by neuroblasts originating from the ventricular zone (VZ) and, it is believed, that these neuroblasts respond to Reelin by outward migration toward the cortical plate (CP) situated beneath the MZ. Reelin is indispensable for cortical layering; this is affirmed by layer inversion in mice deficient in genes of the Reelin-signaling pathway, which include Reelin, Dab1, the very low-density lipoprotein receptor (VLDLR), and the apolipoprotein E receptor 2 (ApoER2) (D'Arcangelo et al. 1995; Sheldon et al. 1997; Howell et al. 1997; Trommsdorff et al. 1999).
Despite half a century of research, it remains a mystery how Reelin signaling is translated into the proper assembly of neurons into sequential layers over developmental time. At present, there are a number of working models of Reelin function. According to the “chemoattractant” model (Sheppard and Pearlman 1997), Reelin in the MZ sends a chemoattractive cue that attracts migrating neurons to the MZ, causing early-born neurons to migrate past the subplate and late-born neurons to push past early-born neurons. In the absence of Reelin, the late-born neurons become arrested under previously generated layers, causing layer inversion. On the other hand, the “stop-signal” model posits that neurons migrate into the CP until they detect the presence of Reelin in the MZ, whereupon they cease migration and detach from the supporting radial glia. Without this stop signal, neurons migrate past the CP to infiltrate MZ, fail to split the preplate, and cause later-born neurons to be log-jammed underneath the early-born neurons (Ogawa et al. 1995; Pearlman and Sheppard 1996; Dulabon et al. 2000).
Despite their simplicity, neither of the above models is consistent with a number of observations from recent studies. First, the distribution of Reelin is not confined to a localized source in the MZ and has diffused expression, particularly at later stages of corticogenesis, in deeper layers of the CP (Alcantara et al. 1998; Meyer et al. 2002). This expression pattern provides evidence against a chemoattractive or inhibitory function for Reelin. Second, in transgenic mice, forced expression of Reelin driven by the nestin promoter in the VZ did not attract neurons to this region or stop neurons migrating past it; instead the ectopic Reelin was able to rescue certain aspects of cortical layering such as the formation of an ectopic subplate (Magdaleno et al. 2002). Third, Reelin in the MZ is mostly secreted by CR cells originating from the cortical hem (which express p73 and Wnt3a) (Meyer et al. 2002; Takiguchi-Hayashi et al. 2004; Yoshida et al. 2006) and by calretinin-positive CR cells from the ventral palium (Bielle et al. 2005). Removal of hem-CR cells by genetic ablation of p73 or Wnt3a in this population causes a large decline in the number of Reelin-secreting cells; yet the cortices of these animals remain properly laminated suggesting that extremely low doses of Reelin are necessary and sufficient (Meyer et al. 2004; Yoshida et al. 2006). Together, the above considerations suggest that Reelin function is complex and needs to take into account both the location and the quantity of Reelin expression (Magdaleno et al. 2002; D'Arcangelo 2006; Yoshida et al. 2006).
One way of dissecting complex gene functions is to use genetic mosaics to unveil genotype-specific contributions to animal patterning (Rossant and Spence 1998). In higher vertebrates, mouse chimeras offer a useful tool for studying the developmental behavior of cortical neurons with different genotypes (Hammond et al. 2001, 2004). In the effort to understand Reelin function, we have constructed Reelin+/+ ↔ Reelin−/− chimeras to create mosaicism in the cortex. Since chimeras can be produced with different strengths of contribution by wild-type and mutant neurons, they offer the unique opportunity of testing the hypothesis that critical thresholds of Reelin levels are important for the formation of a laminated cortex.
Materials and Methods
Reelin+/+ ↔ Reelin−/− chimeras (Reelin−/− chimeras) were generated using 2 methods, morula aggregation and microinjection of embryonic stem (ES) cells into blastocysts. To generate aggregation chimeras, 8-cell stage morulae from H253 mice carrying the lacZ reporter gene were fused with Reelin−/− morulae (Mintz 1970; Tan et al. 1993). The latter were produced from intercrossing heterozygous Reelin+/− mice (carrying the Orléans mutation). The H253 morulae carrying the lacZ reporter have normal Reelin alleles (confirmed by polymerase chain reaction [PCR]) and are homozygous for lacZ in females or hemizygous in males (Tan et al. 1993; Sturm et al. 1997). These cells are designated as being wild-type for the Reelin gene, and throughout the study, cells expressing β-galactosidase (βgal) are therefore considered to be Reelin-positive. Chimeric hosts that were Reelin−/− (identified by genotyping) were selected for the study, whereas Reelin+/+ hosts were used as controls. Fused morulae (produced by fusing wild-type morula to Reelin−/− morulae) were cultured overnight to the blastocyst stage before implanting into pseudopregnant foster females. To produce weaker chimeras, ES cells (4–14) carrying the lacZ reporter gene were injected into Reelin−/− blastocysts. Once again, these ES cells have normal Reelin alleles and were derived from the H253 transgenic mouse line with the XlacZ/XlacZ genotype (Tan and Breen 1993; Sturm et al. 1997). Previous experiments using this cell line have shown that they are capable of colonizing the central nervous system, and the genetically marked cells migrated in essentially the same patterns as host cells (Tan et al. 1998; Reese et al. 1999). All chimeric pups were reared until postnatal day (P) 22–24 before genotype analysis, X-gal histochemistry, βgal expression, and immunocytochemical analysis.
To determine whether Reelin+/+ and Reelin−/− neurons migrate to their expected layer positions in the chimeric brains, mice were treated in utero (at gestational ages E13.5 or E16.5) with 5-bromodeoxyuridine (BrdU) at a dose of 100 μg/g body weight (intraperitoneally, dissolved in sterile saline containing 0.007 N NaOH). Neurons containing the BrdU label are born during the labeling period when they undergo their final cell division.
To isolate DNA for PCR genotyping, tissue (toe or tail) was digested in 200 μl PCR lysis buffer (50 mM KCl, 10 mM Tris–HCl, pH 8.3, 2 mM MgCl2, 0.45% Nonidet P-40, and 0.45% Tween-20) with Proteinase K (200 μg/ml) at 56 °C overnight. All DNA primers were used at a final reaction concentration of 10 ng/μl. For genotyping of the Reeler mice, Reelin forward and reverse primers were designed either side of the L1 insertion site. L1 insertion at this site leads to exon skipping and the 220-bp deletion in the Reelin gene that results in the Reeler phenotype (Takahara et al. 1996). An additional reverse primer was designed to the 5′ end of the L1 element. Primers for the Reelin gene were as follows—Reelin: forward, 5′-CGACTGCTCTGTCTTCAGTCACG-3′ and reverse, 5′-GGTGGCAGCTTGCCTTATCTG-3′, and the L1 element: reverse, 5′-GCTGCCTCAGTGCCTCTGTG-3′.
Heat-inactivated DNA (2.5 μl) was added to a reaction mix containing primers (10 ng/μl) 100 mM Tris–HCl, pH 8.3, 500 mM KCl (1× PCR Buffer II, Applied Biosystems), 1.5 mM MgCl2, 100 μM deoxynucleotide triphosphates, and 1.25 U Taq DNA polymerase (Applied Biosystems). Amplification of the samples was carried out after 1 min at 94 °C using 40 cycles consisting of annealing at 61 °C for 30 s, extension at 72 °C for 30 s, and denaturing at 94 °C for 30 s. PCR products were analysed on a 4% agarose gel. Product sizes were 188 bp (wild-type allele) and 160 bp (L1 element in mutant Reelin allele).
To confirm the genotype of chimeric tissues, βgal-positive and βgal-negative areas were laser-captured from X-gal-reacted 60 μm sections mounted on gelatin-coated slides. The slides were dried overnight in a dessicator. A PALM Microbeam Microlaser System and PALM RoboSoftware (Millennium Science) were used to isolate tissue from regions containing 1) blue and white cells and 2) white cells only. The laser-dissected tissue (typically 0.5 mm2) was collected into the cap of a 200-μl tube containing 30 μl of PCR lysis buffer. After collection, the tubes were spun at low speed to collect tissue at the bottom of the tube and the tubes were then stored at 4 °C until DNA extraction. For digestion, Proteinase K (200 μg/ul) was added, the tubes were incubated at 56 °C overnight, and DNA was purified by phenol–chloroform extraction followed by chloroform extraction and ethanol precipitation. Precipitated DNA was resuspended in 4 μl of water, and 2 μl was used for PCR.
Following a lethal injection of Nembutal (Merial Australia, Pty. Ltd.), mice at P22-24 were intracardially perfused for 7 min with 4% paraformaldehyde in 0.1 M Sorensen's phosphate buffer, pH 7.4, and the brains were post-fixed for 7 mins. Following cryoprotection with 20% sucrose in phosphate buffer, brains were embedded in OCT (Tissue-Tek, Torrance, CA) for cryosectioning. Thin coronal sections (14 μm) of the cortex were prepared for immunocytochemistry, and thicker sections (60 μm) were obtained for X-gal histochemistry to visualize the wild-type cells labeled with βgal. Sections for X-gal histochemistry were incubated overnight at 37 °C in a solution of 0.1% 4-chloro-5-bromo-3-indolyl-β-D-galactopyranoside (X-gal) containing 2 mM MgCl2, 5 mM ethyleneglycol-bis(2-aminoethylether)-N,N,N',N'-tetraacetic acid, 0.01% (w/v) sodium deoxycholate, 0.02% (w/v) Nonidet P-40, 5 mM K3Fe(CN)6, and 5 mM K4Fe(CN)6.6H2O in 0.1 M Sorenson's phosphate buffer. The X-gal was prepared as a 4% stock in dimethylformamide and was added to the mixture just before use.
For immunofluorescence studies, 14 μm coronal sections were mounted on 3-aminopropyltriethoxy-silane-coated slides and dried for 2 h before incubation in primary antibodies. To expose the BrdU, brain sections were preincubated with 2 N HCl at 37 °C for 45 mins before addition of the primary antibody. All primary antibodies were diluted in 0.1 M phosphate-buffered saline with 0.3% Triton X-100. Primary antibodies included a rabbit polyclonal antibody to Calbindin 9 Swant, Switzerland; 1:1000), a rabbit polyclonal antibody to βgal (5 Prime → 3 Prime, Boulder, CO; 1:500), a mouse monoclonal antibody to BrdU (Becton Dickinson, San Jose, CA; 1:30), a mouse monoclonal antibody to NeuN (Chemicon, Temecula, CA; 1:200), a mouse monoclonal antibody to 2′3′-cyclic nucleotide 3′-phosphodiesterase (CNPase; Chemicon, Temecula, CA; 1:200), and a rabbit polyclonal to Sez6 (1:1000: Gunnersen et al. 2007).
Secondary antibodies were biotinylated anti-rabbit (Vector Laboratories, Burlingame, CA; 1:200), Alexa Fluor 594-conjugated goat anti-mouse IgG (Molecular Probes, Eugene, OR; 1:300), and Alexa Fluor 594-conjugated goat anti-rabbit IgG (Molecular Probes 1:500). With Calbindin, BrdU, Sez6, NeuN, and CNPase antigens, simultaneous visualization of βgal was achieved with fluorescein–avidin D (Vector Laboratories, Burlingame, CA; 1:200). Incubation with primary antibody overnight was followed by 1 h in secondary antibodies before coverslipping with 10% (w/v) Mowial (Hoechst, Australia) in 25% glycerol, 0.1 M Tris–HCl, pH 8.5, and 2.5% (w/v) 1,4-diazobicyclo-[2.2.2]-octane (Dabco; Sigma).
To quantify the layer positions of BrdU-positive cells, montages were created using Microsoft Powerpoint (Seattle, WA) from images taken within the somatosensory cortex. The cortical thickness was divided into 10 bins (bin 1 is closest to the pial surface), and the number of positive cells in each bin was counted. The data were entered into Microsoft SigmaPlot and plotted as a histogram.
Identification and Genotyping of Chimeras
Chimeras containing lacZ-expressing (and therefore Reelin+/+) cells were identified by their distinctive patchy coat colors. The genotype of the non-lacZ positive contribution to these chimeras may be one of 3 possible genotypes from the Reelin+/− intercross: Reelin−/−, Reelin+/−, or Reelin+/+. Genuine Reelin chimeras (lacZ-expressing Reelin+/+ cells on Reelin−/− background) were distinguished from non-Reelin chimeras (containing Reelin+/− or Reelin+/+ backgrounds) in the following manner. Following X-gal histochemistry, 60 μm coronal sections of brain tissue were examined for blue and white cortical areas. Blue areas (BA) were deemed to be Reelin+/+, whereas the white areas (WA) represent host tissue with 3 possible genotypes: Reelin−/−, Reelin+/−, or Reelin+/+. To distinguish between host genotypes, WA (containing no blue cells whatsoever) were sampled by laser capture, and the genotype was compared with the genotype of tissue containing both non-blue and blue cells (white area + blue area; WA + BA; Fig. 1A). In WA-only tissue, cells were identified to be Reelin−/− (160-bp PCR fragment; lane 1, chimera #81.5), Reelin+/+ (188-bp fragment; lane 7, chimera #81.1), or Reelin+/− (lanes 3 and 5, chimeras #81.4 and #74.1). In each case, the corresponding result from the WA + BA tissue gave the expected wild-type band arising from the contribution of wild-type lacZ-expressing cells (lanes, 2, 4, 6, and 8; upper band, 188 bp). A nonspecific band common to lanes containing WA + BA was also noted. A total of 50 chimeras were created, consisting of 8 Reelin+/+ ↔ Reelin+/+, lacZ, 31 Reelin+/− ↔ Reelin+/+, lacZ, and 11 Reelin−/− ↔ Reelin+/+, lacZ.
Different Cortical Phenotypes Are Associated With Varying Degrees of Cortical Chimerism
Chimeras were generated by aggregating together Reelin−/− and Reelin+/+, lacZ morulae (8 plus 8 blastomeres) or by injection of 4–14 lacZ-expressing ES cells into non-lacZ-positive blastocysts (from Reelin+/− intercrosses). Wild-type chimeras generated previously using either method have been shown to exhibit normal patterns of cortical neuron migration and layering (Tan et al. 1998; Hammond et al. 2001, 2004). Except where indicated (e.g., Fig. 1C), all chimeras under discussion contain a mixture of neurons in their cortices that are either Reelin−/− or Reelin+/+ (see Materials and Methods). Chimeras arising from heterozygous host embyos with Reelin+/− genotype (Reelin+/− ↔ Reelin+/+, lacZ) were sectioned and reacted with X-gal but were not analyzed further.
Reelin chimeras analyzed between P22 and P24 were grouped into 3 different categories depending on the proportion of Reelin+/+ (βgal-positive) cells present (strong: rescued, intermediate: compound, and weak: unrescued). In rescued chimeras with a high contribution from Reelin+/+ cells (Fig. 1B), the cortical organization is similar to that of wild-type chimeras (Reelin+/+ ↔ Reelin+/+), with unaffected cortical lamination revealed by BrdU birthdating (results not shown). Radial columns run continuously from deep to superficial layers (indicated by dotted lines, Fig. 1B), as seen in wild-type chimeras (Fig. 1C). Despite the strong presence of Reelin+/+ cells, a greater proportion of cells in these cortices are of Reelin−/− genotype (βgal-negative cells revealed by bisbenzamide staining; data not shown). Thus, it would appear that in these cortices, the strong contribution by Reelin+/+ cells has resulted in the development of a normal-appearing cortex with properly laminated neurons. One possibility is that the amount of Reelin secreted by the wild-type cells in these chimeras exceeds the threshold level required for proper cortical lamination. In contrast, cortices from intermediate Reelin chimeras (lower contribution from Reelin+/+ cells) appear to have incomplete columns with the great majority of βgal-positive cells clustered near the white matter (WM; Fig. 1D). A minority of βGal-positive cells are distributed in the middle and superficial cortical layers (Fig. 1D) with the majority of them in the deeper layers. These intermediate chimeras have 2 cortices within a single hemisphere, with an additional 6-layered cortex occupying the upper half (previously termed a supercortex; Hammond et al. 2001) and another 6-layered structure in the lower half of the hemisphere. In the third category, weak Reelin chimeras have a sparse number of Reelin+/+ cells per hemisphere (Fig. 2). These chimeric brains exhibit a number of features suggestive of the Reeler phenotype, indicating that the small number of wild-type (green) cells (Fig. 2A, arrows) are insufficient for rescue. In addition, these mice exhibited ataxic behavior that provided additional evidence for lack of rescue. CNPase immunoreactivity for myelin indicates the presence of oblique fiber bundles running throughout all cortical layers similar to that previously reported for the Reeler mouse (Fig. 2A and A’; Mikoshiba, Takamatsu, Tsukada 1985). Birthdating of neurons born on E16.5 with BrdU indicates that neurons of both genotypes (wild-type, yellow; mutant, red) are inappropriately distributed near the ventricle, suggesting inverted positioning of these neurons in the deeper part of the cortex (Fig. 2B and B’). In addition, the normally cell-sparse MZ is infiltrated by a large number of cell bodies, similar to cortices with mutations in Reelin signaling (Fig. 2C, arrowheads). Together, these results indicate that despite the presence of a small number of wild-type neurons containing normal alleles of the Reelin gene in the cortex, the overwhelming Reeler phenotype in their cortices remains unrescued.
Intermediate Chimeras (Reelin+/+ ↔ Reelin−/−) Form Mirrored Cortices
Previously, we demonstrated that when Dab1+/+ and Dab1−/− neurons inhabit the same cortex in Dab1 chimeras, the Dab1+/+ neurons invariably formed a correctly layered cortex that we termed as “supercortex,” sitting on top of another 6-layered mutant cortex that is inverted and made up by Dab1−/− neurons (Hammond et al. 2001). The net result was a 12-layer compound cortex in mirror image. In the present study, a compound cortex was formed in intermediate Reelin+/+ ↔ Reelin−/− chimeras (n = 2), suggesting the importance of a critical threshold in Reelin levels for this phenotype. The overall thickness of the cortical wall was similar in all 3 types of chimeras, but in the compound cortices of intermediate chimeras, the supercortex occupies roughly the upper half of the cortical thickness (Fig. 3). Layering order in the compound cortices was examined using a combination of BrdU birthdating and molecular markers. Calbindin is known to differentially stain the neuropil of the supragranular layers of pyramidal neurons (Fig. 3A, bracket; Demeulemeester et al. 1989; van Brederode et al. 1991), and in Reeler cortex, this band of calbindin immunoreactivity is abnormally displaced to the deeper layers (Fig. 3B, bracket; Magdaleno et al. 2002). In the compound cortex of intermediate chimeras, calbindin staining of the neuropil is observed as 2 horizontal bands, one close to the pial surface and the other close to the WM (Fig. 3C, brackets). This suggests that layers II/III pyramidal neurons are present in both locations, with one set of supragranular neurons in the correct superficial position and a second set of supragranular neurons in the incorrect deeper position. These neuronal positions were confirmed by marking of layer II/III neurons with BrdU birthdating at E16.5 and the brains examined at P24. The results show that BrdU-labeled neurons are biased toward the superficial position in wild-type cortices (Fig. 3D, arrows) and toward deeper layers of the Reeler cortex (Fig. 3E, arrows). In the compound cortex, they are present in both superficial and deep positions in a mirrored arrangement (Fig. 3F, arrows). Interestingly, the majority of the BrdU-labeled cells in the supercortex were of mutant genotype (Fig. 3F, arrows), whereas those in the cortex underneath were a mixture of mutant (Fig. 3F, arrows) and wild-type cells (Fig. 3F, arrowheads). Thus, it would appear that layer II/III neurons in these brains can either migrate to their correct superficial positions or be found abnormally distributed in deeper positions close to the WM. To investigate how infragranular neurons belonging to layers V/VI are spatially related to layer II/III neurons, staining of layer V/VI neurons was performed with Sez-6, a strong marker for layer V/VI pyramidal neurons (Lein et al. 2004; Gunnersen et al. 2007). Although Sez-6 staining in the wild-type cortex is correctly distributed to deep layers (Fig. 3G, bracket) and in the Reeler brain the majority of these neurons are wrongly distributed to the superficial layers (Fig. 3H, bracket), the compound cortex shows a merging of these 2 opposite positions into a single middle band (Fig. 3I, bracket). This would suggest that in relation to the layer II/III neurons, the layer V/VI neurons now occupy 2 adjacent bands, one in correct position relative to the layer II/III neurons of the supercortex and the other in the inverted position relative to layer II/III neurons of the inverted mutant cortex. Alternatively, they may have merged together to form a single central band. These observations suggest that in the compound cortex, there are 2 cortices sharing the same hemisphere, one correctly layered supercortex in the upper half and a mirrored inverted cortex in the lower half.
Compound Cortices Exhibit Other Reeler Phenotypes
So far, the results suggest that in a compound cortex, there are 2 cortical compartments with a correctly layered supercortex sitting on top. Underneath this lies a mutant cortex that comprises neurons in inverted positions. This unique arrangement of cortical neuron positioning suggests that although a low level of Reelin signaling may have successfully ushered in certain aspects of cortical development, this developmental program has been incompletely executed. If this is correct, other elements of the Reeler phenotype should also be present. To examine this, we studied the myelination patterns by staining for myelin-associated enzyme CNPase that is normally found in the plasma membrane of oligodendrocytes (Sprinkle 1989; Thompson 1992). In the wild-type cortex, CNPase immunoreactivity is associated with myelin-producing cells found in deeper layers and also in the subcortical WM, with hardly any positive fibers present in superficial layers (Fig. 4A; Vogel et al. 1988). In the Reeler cortex, thicker and oblique fiber bundles can be observed throughout the cortical depth (Fig. 4B; Mikoshiba, Takamatsu, Tsukada 1985). In Reelin chimeras, however, different patterns of myelination are observed, depending on the degree of chimerism as indicated by the presence of βgal-positive neurons. In one intermediate chimera, the 2 hemispheres showed different chimeric strengths. In the chimera shown (Fig. 4C), the rescued hemisphere is more populated with wild-type cells. In this hemisphere, both the hippocampus and the cortex display normal features implying full rescue of the Reeler phenotype. The cortex on this side shows normal ordering of myelinated fibers, being restricted mainly to the lower layers, whereas the hippocampus displays a normal stratum pyramidale and dentate gyrus (Fig. 4C, βgal-positive cells are green). This rescue is further demonstrated at higher power in Figure 4D where βgal-positive cells (double-labeled: yellow) are evenly arranged in the CA1. In the opposite hemisphere, the number of βgal-positive cells is clearly reduced, and this is associated with defects in both the cortex and the hippocampus (Fig. 4C and E). In this hemisphere, a compound cortex is present with abnormal fiber fasciculation revealed by CNPase staining (Fig. 4C). Similar to the Reeler cortex (Fig. 4B), there are thick fiber bundles within the gray matter and also beneath the pial surface. In addition, there are clustered bundles marking the boundary separating the supercortex from the mutant cortex (Fig. 4C, arrowheads). This notion of a secondary WM underpinning the supercortex is also present in another compound cortex (Fig. 4F). In addition to the cortex, the hippocampus also display features of incomplete rescue, with splitting of the pyramidal cell layer (Fig. 4E, arrows and arrowheads). Furthermore, the MZ for the compound cortex is infiltrated by excess neurons (Fig. 4G and inset), another symptom of failed Reelin signaling.
Reelin+/+ and Reelin−/−Neurons Have Nonequivalent Positional Responses to Reelin Insufficiency
Previous rescue experiments, using forced expression of Reelin in Reeler mice or the addition of recombinant Reelin to a brain slice preparation, have been designed to test the effects of Reelin protein on neurons with Reelin−/− genotypes only (Magdaleno et al. 2002; Jossin et al. 2004). In the present study, both Reelin−/− and Reelin+/+ neurons are exposed to Reelin insufficiency during development. Since Reelin is secreted and believed to act non-cell-autonomously on its target cells (Jossin et al. 2004; Zhao et al. 2006), it would be of interest to ascertain if neurons of both genotypes are equally represented in a cohort of BrdU-labeled neurons (injected at E16.5 to label late-born neurons) with respect to their positions. In these intermediate chimeras, the vast majority of BrdU-labeled neurons belong to Reelin−/− genotype. In the upper half, the correctly layered cortex displays more BrdU-positive Reelin−/− neurons (Fig. 5B, red: arrows) compared with the bottom-half containing the mutant cortex (Fig. 5D, red: arrows). Quantification of all BrdU-positive Reelin−/− neurons (2 sections from 2 hemispheres) showed a greater proportion of these neurons in the normal cortex compared with the mutant cortex, in their expected layer II/III positions, respectively (Fig. 5C). In contrast, Reelin+/+ neurons that were not labeled by BrdU (i.e., born outside the E16 window) displayed unequal distribution patterns. In the bottom-half of the cortex (Fig. 5D), Reelin+/+ neurons (green) outnumber Reelin+/+ neurons in the top-half (Fig. 5B). These results indicate that both Reelin−/− and Reelin+/+ neurons have unequal responses in positional distribution in these intermediate chimeras.
In the present study, we have used chimeras to dissect Reelin function in a mixed cortical environment. Given that recent studies have questioned the requirement for Reelin to be present in the correct location, or even the correct amount of Reelin for cortical lamination (Magdaleno et al. 2002; D'Arcangelo 2006; Yoshida et al. 2006), it is of interest to test the effect of varying Reelin dosage by comparing juvenile brains with different chimeric strengths. Using a chimeric approach, we demonstrate that it is possible to vary the contribution of Reelin+/+ neurons in essentially Reelin−/− brains. This was achieved in Reelin+/+ ↔ Reelin−/− chimeras by titrating the contribution of Reelin+/+ blastomeres or ES cells. Admittedly, the present study only examined juvenile brains, and therefore, no information is available on the percentage and distribution of Reelin-positive CR cells in the embryo (Supplementary Fig. 1). Nonetheless, the resulting chimeras displayed varying cortical phenotypes that are directly related to the relative contribution of Reelin+/+ neurons to Reelin−/− hosts. In strong chimeras, where a larger number of Reelin+/+ neurons are present, the Reeler phenotype was essentially rescued. Cortical organization and layer order were restored in a pattern that was largely indistinguishable from wild type. Even though the proportion of Reelin+/+ neurons was less than half in these cortices, it would appear that the amount of Reelin produced was sufficient for ordered cortical patterning. Conversely, when the contribution of Reelin+/+ neurons was low, as seen in the unrescued chimeras, the Reeler phenotype was seen to persist, behaviorally and morphologically. In these chimeras, the Reelin+/+ neurons comprised a minority, and it may be concluded that the rescue of cortical layering is quantitatively related to the amount of Reelin+/+ neurons present.
A third phenotype was seen in chimeras of intermediate strength with respect to Reelin+/+ neurons. Numerically, Reelin+/+ neurons in these chimeras are clearly in the minority, and it may be presumed that the quantity of available Reelin was below the threshold levels necessary to effect full rescue. However, unlike the unrescued chimeras, there appears to be partial rescue in the intermediate chimeras, with aspects of both correctly layered and inverted cortices present in the same hemisphere. A mixture of normal and Reeler-like phenotype neuroblasts beneath the MZ has previously been reported in E18 chimeras (Mikoshiba, Yokoyama, et al. 1985). By contrast, our juvenile chimeras demonstrate that the correct and incorrect layers are not integrated, rather they are segregated and stacked in the same hemisphere. The correct layers were situated in the superficial aspect, forming a supercortex lying on top of an additional layer of WM. Underneath this additional WM was a mutant cortex containing inverted cortical layers, with the supragranular neurons lying next to the ventricle. This mirror effect of cortical layering has previously been reported in Dab1+/+ ↔ Dab1−/− chimeras (Hammond et al. 2001), suggesting that similar mechanisms are responsible for generating 2 sets of cortical layers within a single hemisphere when wild-type neurons are mixed together with neurons mutant for Reelin signaling. It is interesting that in both instances, the correctly layered cortex is situated in the upper half and the inverted cortex in the lower half. Taken together with Dab1+/+ ↔ Dab1−/− chimeras, this arrangement has implications for recent models of Reelin function. For example, it has been hypothesized that Reelin is required only for the somal translocation phase of layer VI neurons, and subsequently, for detachment and somal translocation of layers II–V neurons (Cooper 2008). This hypothesis has some support from the Dab1 chimeras, where the mutant neurons lacking Dab1 are consistently trapped in the mutant cortex underneath the properly layered supercortex. This hypothesis would also predict that where Reelin is insufficient, then some neurons would be rescued, whereas others would not, with the unrescued one forming an inverted cortex closer to the ventricle. Together, it would appear that the positions of the inverted layers in both types of chimeras are consistent with the “detach and go” model for Reelin function (Cooper 2008).
However, there is a major point of difference between the 2 types of chimeras. In the supercortex of Dab1 chimeras, the correctly layered neurons are invariably wild-type and the inverted cortex contains mainly mutant neurons, suggesting that only neurons capable of activating Reelin signals are properly positioned. In the intermediate Reelin chimeras, both the correctly layered supercortex and the inverted mutant cortex contain both Reelin+/+ and Reelin−/− neurons, suggesting the Reelin genotype per se is of no consequence to partial rescue. Instead, partial rescue may represent an incomplete response by competent neurons to a limited supply of Reelin. Partial rescue due to insufficient Reelin has also been reported when exogenous Reelin is introduced into Reeler mutants by forced expression (Magdaleno et al. 2002). In these animals, only the preplate splitting, but not cortical layering, was successfully rescued. Thus, too little Reelin appears insufficient for directing proper cortical layering, but once a threshold level of Reelin is available, the stimulus appears to be sufficient for forming the correct layers.
It remains curious why in the intermediate chimeras the βgal-positive cells (Reelin+/+) are mostly segregated to the lower half of the compound cortex, next to the ventricle (e.g., Fig. 1D). The layer identities of these neurons are unclear, although BrdU labeling at E16.5 would suggest that some of them are supragranular (Fig. 5D). A current hypothesis postulates that supragranular neurons depend strongly on Reelin for proper positioning in their terminal translocation phase (Cooper 2008). Hence, although many of supragranular neurons successfully reach their destinations in the supercortex, the current results would suggest that a large number also fail due to Reelin insufficiency. Curiously, the above argument does not seem to apply to Reelin−/− supragranular neurons labeled with BrdU at E16 (Fig. 5). This cohort of neurons appears to favor the upper half of the compound cortex, with the opposite picture being true for Reelin+/+ supragranular neurons. One possible explanation for this is selective cell death of Reelin+/+ neurons in the upper half of the cortex and Reelin−/− neurons in the lower half. The mechanisms underlying these dichotomous effects require further study. Finally, differential adhesive properties of neurons with different Reelin genotypes might also account for the asymmetric positioning of βGal-positive (wild-type) neurons. For instance, it has been demonstrated that Reelin−/− neurons have greater adhesive properties (Hoffarth et al. 1995), although at face value, this should produce the opposite effect for Reelin+/+ neurons.
In conclusion, the present study has demonstrated that quantitative levels of Reelin might be an important consideration in the assembly of cortical layers. The comparison of cortical phenotypes between mice of different chimeric strengths would suggest that a certain threshold level of Reelin is required for correct layering. Below this threshold, Reelin function is complex, and within the same cortex, both correct and incorrect formation of cortical layers can coexist. Thus, it would appear that when Reelin levels are insufficient, the stimulus to form a set of cortical layers in correct order is undiminished although incomplete.
Program 301204 and Project Grants 454571 from the National Health and Medical Research Council.
We wish to thank Andre Goffinet for the gift of Reeler mice, Frank Weissenborn for expert technical assistance, and Fiona Christensen for making the chimeras. Conflict of Interest: None declared.