The mammalian cerebral cortex can be tangentially subdivided into tens of functional areas with distinct cyto-architectures and neural circuitries; however, it remains elusive how these areal borders are genetically elaborated during development. Here we establish original bacterial artificial chromosome transgenic mouse lines that specifically recapitulate cadherin-6 (Cdh6) mRNA expression profiles in the layer IV of the somatosensory cortex and by detailing their cortical development, we show that a sharp Cdh6 gene expression boundary is formed at a mediolateral coordinate along the cortical layer IV as early as the postnatal day 5 (P5). By further applying mouse genetics that allows rigid cell fate tracing with CreERT2 expression, it is demonstrated that the Cdh6 gene expression boundary set at around P4 eventually demarcates the areal border between the somatosensory barrel and limb field at P20. In the P6 cortical cell pellet culture system, neurons with Cdh6 expression preferentially form aggregates in a manner dependent on Ca2+ and electroporation-based Cdh6 overexpression limited to the postnatal stages perturbs area-specific cell organization in the barrel field. These results suggest that Cdh6 expression in the nascent cortical plate may serve solidification of the protomap for cortical functional areas.
The cerebral cortex (isocortex or neocortex) has 6 major layers in its radial dimension and scores of functional areas in its tangential dimension with each area harboring specialized roles along unique cyto-architecture and neuronal connectivity. To elaborate the mammal-specific characteristics during early development, the cortical ventricular zone (VZ) generates pyramidal neurons in a birthday-dependent manner and neurons that belong to each layer move along the radial glia fibers in an inside-out manner to form the cortical plate (CP) with obscure areal boundaries at the perinatal stages (Rakic 1988; Rakic et al. 2009). At the later developmental stages, individual cortical area demarcated by abrupt border is arranged with many of differentiated interneurons and/or axonal terminals, most of which are originated from extra cortical regions such as basal ganglia and thalamus (O'Leary and Nakagawa 2002; López-Bendito and Molnár 2003; Markram et al. 2004).
Many studies have revealed that transcription factors expressed in the cortical VZ and their upstream/downstream genes (i.e. “protomap” mechanism) likely exert dominant roles in CP arealization (Rakic 1988; Bishop et al. 2000; Fukuchi-Shimogori and Grove 2001; Monuki and Walsh 2001; Grove and Fukuchi-Shimogori 2003; Funatsu et al. 2004; Hamasaki et al. 2004; Mallamaci and Stoykova 2006; Armentano et al. 2007; O'Leary et al. 2007). Supporting the scenario, it is recently suggested that a transcription factor Bhlhb5 whose expression appeared to demarcate putative primary sensory areas in the CP at the postnatal day 4 (P4) plays a role in specifying some sets of areal identities (Joshi et al. 2008). Additionally, a lifespan analysis of intraneocortical connections and several gene expression profiles such as RORβ and Id2 from the early embryonic stages to adulthoods revealed significant correlations between gene expression boundaries and developing cortical areal borders (Dye et al. 2011a, 2011b). These clearly indicated that functional areal borders could be determined by combinatorial patterns of gene expression boundaries in the nascent CP. However, it still remains elusive whether those gene expression boundaries at the perinatal stages immediately elaborate the functional areal borders, since individual cortical cell might dynamically change their gene expression during arealization at around the expression border and/or the positions of gene expression boundaries could be balanced due to the activity-dependent machineries (i.e. “protocortex” mechanism).
Vertebrate classic cadherins comprise as many as 20 subclasses, each of which can confer unique adhesiveness to cells, whereby cell population that expresses the same sets of cadherin subclasses actively segregates each other (Nose et al. 1988; Takeichi 1995). Quantitative differences of a given cadherin subclass also contribute to segregation of cell populations (Steinberg and Takeichi 1994). Each classic cadherin subclass shows spatio-temporally restricted expression profiles, thereby plays diverse roles in neural development including regional boundary formation and/or maintenance by regulating cell–cell adhesiveness (Redies 2000; Takeichi 2007). For instance, during mouse embryonic development, cadherin-6 (Cdh6) expression boundary appeared to coincide with a cell lineage restricted boundary between forebrain and midbrain at the 5-somite stage (Inoue et al. 2000). R-cadherin (Cdh4) and Cdh6 delineated another cell lineage restricted boundary between the putative cerebral cortex and lateral ganglionic eminence and such differential expression profiles were confirmed to be essential to maintenance of the neuromere boundary at the embryonic day 10.5 (E10.5; Inoue et al. 2001). Interestingly, in the cerebral cortex, Cdh6, Cdh8, and Cdh11 mRNA expression differentially demarcated sub-regions at the first postnatal week (Suzuki et al. 1997; Inoue et al. 1998). Especially, Cdh6 and Cdh8 expression showed mutually exclusive pattern at the somatosensory area: the former seemed to mark the putative barrel field whereas the latter appeared to localize into the limb field (Suzuki et al. 1997; Inoue et al. 2008a, 2008b; Dye et al. 2011a, 2011b). This complementary expression profile might define the future functional areal border between the barrel and limb field. However, the difficulty in long-term cell fate tracing has prevented detailed analysis of cellular and/or expression dynamics in cortical arealization.
A rodent's cortical area, barrel field is a somatosensory-receptive field of whiskers and forms the topographic map in which each whisker is represented in a discrete anatomical module “barrel” consisting of axons from the ventral posterior medial (VPM) thalamic nucleus and their surrounding neurons in the layer IV (Woolsey and Van der Loos 1970; Petersen 2007). Due to its clear structure that dynamically develops at the first postnatal week, when several gene expression boundaries are established in the cerebral cortex, the barrel field provides a perfect model in investigating the relationship between gene expression profiles and functional area formation.
In the present study, we found that in the layer IV of the mouse somatosensory cortex, a Cdh6 gene expression boundary was formed at the first postnatal week on the basis of Cdh6 mRNA in situ hybridization (ISH) as well as reporter expression analyses in bacterial artificial chromosome-transgenic (BAC-Tg) mouse lines that selectively recapitulate Cdh6 mRNA expression profiles in layer IV neurons by reporter integrations. Additionally, by generating a Cdh6-BAC-CreERT2 Tg mouse line, which allows us to label Cdh6-expressing layer IV neurons in a temporally specific manner by tamoxifen (TM) administration and to trace their areal distributions during long-lasting cortical development, we indicated that the Cdh6 expression boundary cells once positioned at P4 precisely contributes to an areal border cells between the somatosensory barrel and limb field at P20. To our knowledge, this is the first demonstration of the experimental evidence that CP gene expression boundary established at the perinatal stage immediately pre-patterns the future functional areal border. Furthermore, we found that cells with Cdh6 expression preferentially form aggregates in dissociated cell pellets from the P6 CP in a manner dependent on Ca2+ and that ectopic Cdh6 expression limited to the postnatal stages disrupts area-specific organization in the somatosensory barrel field. These results highlight the possible roles of Cdh6 in establishing and/or maintaining the areal pattern in the postnatal CP.
Materials and Methods
In situ Hybridization
Twenty-five–micrometer cryosections from postnatal brains were put on slide glasses coated with VECTABOND reagent (VECTOR LABORATORIES). Cdh6 RNA probe preparations, pretreatment of sections, hybridization, and the probe detections were performed as described previously (Inoue et al. 1998; Ishii et al. 2000).
Generation of Cdh6-Bacterial Artificial Chromosome-Transgenic (BAC-Tg) Mouse Lines
All animal experiments in this study conform to Japanese governmental guidelines and have been approved by the Animal Care and Use Committee of the National Institute of Neuroscience, Japan (Projects 2007022 and 2011007).
Establishment of the Cdh6-BAC-LacZ Tg line, in which Escherichia coli beta galactosidase (β-gal; LacZ) is expressed on the basis of a BAC clone RP23-78N21 with the 3′ coverage of the Cdh6 locus that contains cis-regulatory modules for the somatosensory cortex, was reported in the earlier studies (Inoue et al. 2008a, 2008b).
For generation of the Cdh6-BAC-nLacZ Tg line, the same BAC clone RP23-78N21 was modified by means of homologous recombination so that LacZ reporter with the nuclear localization signal (NLS) is expressed in vivo as the Cdh6-BAC-LacZ Tg line. Briefly, two 5′ phosphorylated oligonucleotides with NarI and BamHI linkers encoding MAPKKKRKVKILD, the underlined part of which corresponds to the NLS (Nar-NLS-Bam-F: 5′-GGCGCCATGGCCCCAAAAAAGAAGAGAAAGGTAAAGATCTTGGATCC-3′; Nar-NLS-Bam-R: 5′-GGATCCAAGATCTTTACCTTTCTCTTCTTTTTTGGGGCCATGGCGCC-3′), were purchased (Invitrogen), annealed, and cloned into the blunted BamHI site of the pBluescript II-SK vector (Stratagene). Those clones with an insertion of the NLS containing fragment in a T7 to T3 direction were then selected by sequencing and the ScaI–BamHI fragment was excised. In parallel, a SmaI fragment from the pBGZ40 (Inoue et al. 2008a) was cloned into the blunted BamHI site of the pBluescript II-SK vector in the same coding direction to the NLS-containing fragment and the ScaI–BamHI fragment was excised to replace it with the aforementioned NLS-containing ScaI–BamHI fragment, yielding pBSII-nLacZpA, in which MAPKKKRKVKILD is conjugated in frame to the fourth amino acid “P” from the LacZpA cassette of pBGZ40. Next, the portion of nLacZpA was cloned into a shuttle vector in which the LacZ cassette was excised from the Rec:LacZ→Cdh6-ATG Neo vector (Inoue et al. 2008a). The fragment containing the homology arms, nLacZpA and Neo cassettes was then isolated by means of electrophoresis and was electroporated into the recombinogenic EL250 strain that harbors the BAC clone RP23-78N21 for homologous recombination (Lee et al. 2001; Asami et al. 2011). Those BAC clones with correct modification were selected by polymerase chain reaction (PCR) and were confirmed by pulse field gel electrophoresis. For PCR, the same combinations of primers were used as previously listed (Inoue et al. 2008a). To obtain Tg founders, modified BAC was purified, linearized, and microinjected into mouse-fertilized eggs as detailed previously (Inoue et al. 2008a).
To express CreERT2 under the control of Cdh6 cis-regulatory regions for the somatosensory cortex, the BAC clone RP23-78N21 was again modified by using the EcoRI fragment from pCre-ERT2 (Feil et al. 1997). Processes for the BAC modification were totally the same except for a step by which a loxP site in the BAC vector pBACe3.6 was replaced to a Neo cassette via homologous recombination to prevent eventual deletions of chromosomally integrated BACs by CreERT2. Genotypes of Tg founder generated were determined by PCR primers for the CreERT2 cassette (Cre-for: 5′-ACCTGAAGATGTTCGCGATTATCT-3′; Cre-rev: 5′-ACCGTCAGTACGTGAGATATCTT-3′). For western blot analysis of Tg founders, postnatal cerebral cortices were sampled for sodium dodecyl sulfate polyacrylamide gel electrophoresis and the blots were detected with anti-Cre rabbit polyclonal antibody (69050-3; Novagen) in a 1:1000 dilution.
Labeling Cdh6-Expressing Cells with Temporal Specificity by Using Cdh6-BAC-CreERT2 Tg Mouse Line
In order to label Cdh6-expressing cells with temporal specificity, Cdh6-BAC-CreERT2; ROSA26R males were bred with B6C3F1 wild-type females to get progenies carrying Cdh6-BAC-CreERT2; ROSA26R. Vaginal plugs were checked in the morning and the noon of the day was defined as E0.5. The birthday was designated as the postnatal day 0 (P0). For those embryos carrying Cdh6-BAC-CreERT2; ROSA26R, TM was administered to timed pregnant females by intraperitoneal injection at 9:00 am with the dose being 1 mg per 10 g body weight. For postnatal pups carrying Cdh6-BAC-CreERT2; ROSA26R, TM was intraperitoneally injected at 9:00 am with the dose being 0.375 mg per 10 g body weight.
Staining Brain Slices for β-Gal Reporter Expressions
In order to prepare brain slice samples, whole brains were fixed for 50min on ice with phosphate-buffered saline (PBS) containing 1% paraformaldehyde (PFA), 0.2% glutaraldehyde, 2 mM MgCl2, 5 mM EGTA, and 0.02% NP40. After washing 3 times using PBS containing 0.02% NP40 (=WB), brains were embedded in 2% agarose/PBS. Slices of 350–550 µm thickness were prepared by using a microslicer DTK-3000 (D.S.K). These slices were then incubated at 37°C with the staining buffer containing 5 mM K3Fe(CN)6, 5 mM K4Fe(CN)6·3H2O, 2 mM MgCl2, 0.01% sodium deoxycholate, 0.02% NP40 and 1 mg/mL X-gal (Calbiochem) for a defined time: those from Cdh6-BAC-CreERT2; ROSA26R mice were incubated for 3h at P7 and for 1.5h at P20, while those from Cdh6-BAC-LacZ mice were stained for 1.5h at any postnatal stages. Stained slices were washed twice by WB and post-fixed overnight at 4°C with PBS containing 10 mM EDTA, 1% PFA, 0.2% glutaraldehyde, 2 mM MgCl2, 5 mM EGTA, and 0.02% NP40. After washing 3 times by WB, they were stored at 4°C with WB containing 1 mM EDTA. In order to analyze them, they were put on 2% agarose/PBS plate and pictures were taken under the stereoscopic microscopes (MZ FLIII, Leica) equipped with a CCD camera (DFC300FX, Leica).
Quantification of β-Gal-Positive Cell Number and Intensity in Cdh6-BAC-NLacZ Mice
Whole brains from Cdh6-BAC-nLacZ mice were fixed in 2% PFA/PBS for 12h at 4°C. After washing in PBS solution twice, fixed brains were immersed in 30% sucrose/PBS solution overnight at 4°C. Then, they were frozen in Tissue-Tek OCT compound (Sakura) and sectioned at 20 μm by a cryostat (Leica) to collect the brain cryosections on slide glasses coated with VECTABOND reagent (VECTOR LABORATORIES). After washing by PBS solution twice, sections were fixed for 20 min with PBS containing 1% PFA, 0.2% glutaraldehyde, 2 mM MgCl2, 5 mM EGTA, and 0.02% NP40 at room temperature. After these samples were washed twice, they were incubated with staining buffer containing 5 mM K3Fe(CN)6, 5 mM K4Fe(CN)6·3H2O, 2 mM MgCl2, 0.01% sodium deoxycholate, 0.02% NP40, and 1 mg/mL X-gal (Calbiochem) for 60min at room temperature. They were washed by PBS solution twice to stop the reaction and post-fixed with PBS containing 1% PFA, 0.2% glutaraldehyde, 2 mM MgCl2, 5 mM EGTA, 0.02% NP40, and 10 mM EDTA at 4°C. Stained sections were then immersed to graded series of ethanol, followed by xylen, and mounted with Enterannew (Merck).
In order to count β-gal-positive nuclei, images in the somatosensory cortex were captured by a microscope (Leica DM5000B) equipped with a CCD camera (DFC300FX). Brains from 3 different pups were collected and 3 sections per animal were analyzed. We defined rectangular regions for the measurement in the layer IV of the putative barrel field, the layer IV of the putative limb field, and the thalamic nuclei. Briefly, a 100 × 150 µm window for the layer IV of barrel and limb field and a 100 × 100 µm window for the thalamus were selected and the number of β-gal-positive nucleus from each section was counted to obtain the average number from each window.
To compare expression levels of each β-gal-positive nucleus in the putative barrel field with those in the putative limb field, color images for analysis were transformed into gray scale mode, in which β-gal signal intensity was in proportion to the depth of black color. The intensity of each β-gal-positive nucleus was categorized by means of the K value (100%–0%), which shows the depth of black color in Photoshop CS4 software (Adobe): Level 5, Level 4, Level 3, Level 2, and Level 1 corresponds to the K value 100–90, 90–80, 80–70, 70–60, and 60–0%, respectively.
Quantification and Statistical Processing in Slice Samples
For definition of the measuring area, we chose a β-gal-stained P20 slice from Cdh6-BAC-LacZ Tg mice within which barrel structures and their boundaries in the layer IV are clearly identifiable due to the strong β-gal signals from the thalamocortical axon terminals (Fig. 2F asterisks, Supplementary Fig. 2). The rostrocaudal position of this slice was at Bregma −0.70 mm along The Mouse Atlas 2nd Edition (Paxinos and Franklin 2001) and we could easily recognize this position by its clear β-gal staining of the anterior dorsal thalamic nuclei and the anterior ventral thalamic nuclei (Supplemental Fig. 2B). Next, we captured the whole image of this slice with its resolution being 1392 × 1040 pixels and defined a folding-fan–shaped area in the image with its radial length (y-axis) being 50 pixels and its tangential length (x-axis) being 700 pixels along the cortical layer IV (Supplementary Fig. 2C). This defined area was enough to reproducibly trace the transition of β-gal signal intensities along the layer IV of the somatosensory area in the Cdh6-BAC-LacZ Tg mice (Supplementary Fig. 2C). Then the same area at the same slice level from Cdh6-BAC-CreERT2; ROSA26R-brain was analyzed to measure the staining intensities along the radial direction (y-axis) at each tangential level (x-axis) by using Plot Profile tool of Image J (National Institutes of Health) (Supplemental Fig. 2D). For convenience sake, x-axis was subdivided into 10 bins, each of which has equal width and intensities of each bin were quantified. After the minimum value of 10 bins was subtracted from each value, the values of each sample were averaged and the calculated values were plotted for each bin. For statistical processing, paired T-test was used.
Preparation of Cortical Cell Pellet and Primary Culture
A cortical hemisphere from P6 Cdh6-BAC-nLacZ Tg mouse was divided into 3 fragments along the anterior–posterior axis and the middle fragment containing the S1 barrel-limb field was dissected out for cell dissociation. Briefly, ∼12 fragments from ∼6 pups were collected in ice-cold Hepes buffered Calcium Magnesium Free-solution (HCMF) containing 1 mM EDTA to remove the pia matter. These fragments were then transferred to 3.5 cm dish with 2 mL of HCMF containing 1 mM EDTA and were minced into pieces by micro-scissors. After adding 0.2 mL of 0.25% trypsin stock solution to the dish, these tissue fragments were incubated at 37°C for 30 min with gentle agitation to completely digest the cell–cell contact proteins. Followed by further treatment with DNase I under the presence of 5 mM MgCl2 at 37°C for 10 min, these tissue fragments were thoroughly dissociated into single cells by mechanical pipetting in a BAS-coated test tube at 4°C. After straining the cell suspension by the cell strainer (Falcon 2450), cells were centrifuged at 4°C and ∼3 µL of the cell pellet was put on the membrane filter (JHWP01300, Milliopore) floating on the culture medium DH10 that contains DMEM (Nissui) and HAMF12 (Nissui) in 1:1 ratio with 10% fetal calf serum in each well of a 24-well dish. In 1 experiment, 12 pellets were incubated on normal medium, while another 12 pellets were cultured on the medium containing 2 mM EGTA in a CO2 incubator at 37°C for ∼72h. The leftover was again suspended into DH10 and was plated on 3.5 cm dish or chamber slide glass for ∼72h.
Double Staining of Primary Culture Cells with X-Gal and Anti-Neurofilament Antibody
Primary culture cells plated on 3.5 cm dish or chamber slide glass were washed twice with HCMF and histochemically stained for LacZ expression as described above. After the post-fixation process, cells on 3.5 cm dish or chamber slide glass were washed 3 times with Tris-buffered saline containing 0.1% Tween 20 (TBST) and were blocked with TBST containing 2% skim milk and 0.5% dimethyl sulfoxide (TBSTMD) for 1h at room temperature. Cells were then incubated with 1:100 diluted hybridoma supernatant for anti-neurofilament monoclonal antibody (2H3; DSHB, University of Iowa) in TBSTMD at 4°C overnight. After washing twice with TBST, cells were incubated with 1:300 diluted Cy3-conjugated anti-mouse IgG secondary antibody (Chemicon) in TBSTMD for 3h at room temperature. Cells were then washed with TBST for 3 times and were treated with distilled water for 2 times. Finally, the aqueous mounting medium, PermaFluor (Thermo), was dropped to place a cover slip on the dish or chamber slide glass. Photo images were captured by a microscope (DM5000B, Leica) equipped with a CCD camera (DFC300FX, Leica).
In utero Electroporation-Based Cdh6 Overexpression in the Postnatal CP
ICR or B6C3F1 female mice were mated and a tet-ON inducible plasmid vector for Cdh6 that harbors the Tol2 transposable elements (details were described in Kawakami and Noda 2004; Sato et al. 2007) was co-injected with the Tol2-based rtTA expression vector (Kawakami and Noda 2004; Sato et al. 2007) and a transposase expression vector (Kawakami and Noda 2004; Sato et al. 2007) into the E13.5–14.5 telencephalic cavity to deliver these vectors to the ventricular cells by means of in utero electroporation. After birth, doxycycline (Sigma) was intraperitoneally administrated everyday from P2 to P6 with the dose being 25 µg per 10 g body weight to bi-directionally induce EGFP and Cdh6 expressions via the tet responsible element in those cells mainly settled along the CP layer II/III–IV (see also Supplementary Fig. 4 for the experimental design).
Cytochrome Oxidase Staining
P7 electroporated brain samples were cryosectioned into 16 µm thickness. To grasp EGFP/Cdh6-positive cell distribution patterns in individual section, fluorescent images were captured by a microscope (DM5000B, Leica) equipped with a CCD camera (DFC300FX, Leica) before the CO staining process.
For the CO staining, each section was soaked into the CO mixture solution that contains 0.5 mg of cytochrome C (Sigma-Aldrich), 0.5 mg of diaminobenzidine (Sigma-Aldrich), and 40 mg of sucrose per 1 mL of phosphate buffer (pH 7.4). The staining solution was replaced every 12h for ∼2 days at room temperature. Those sections developed sufficient staining signals were then washed 3 times with phosphate buffer (pH 7.4), rinsed once by distilled water, and mounted by the PermaFluor (Thermo). Photoimages were collected by the microscope (DM5000B, Leica) equipped with a CCD camera (DFC300FX, Leica).
The Infraorbital Nerve Transection
P0 Cdh6-BAC-nLacZ Tg mouse pups were anesthetized by hypothermia, and the left infraorbital nerve (ION) was exposed by making a skin incision behind the whisker pad and cut with microscissors. After recovery from anesthesia, the pups were returned to their mothers. P7 brains were fixed with 4% paraformaldehyde in PBS by intracardial perfusion and 12 µm coronal sections were stained using standard methods.
Immunohistochemistry for the Brain Sections
Primary antibodies and dilutions used were guinea pig anti-vesicular glutamate transporter 2 (VGLUT2, 1:6000, Millipore), rabbit anti-RORβ (1:6000, diagenode), and chicken anti-beta galactosidase (βGal, 1:3000, abcam). Appropriate secondary antibodies were from the Molecular Probes Alexa series and used at dilutions of 1:600.
Cdh6 mRNA Expression Profile Changes Dynamically in the Mouse Somatosensory Cortex at the Postnatal Stages
Early studies have demonstrated that mouse Cdh6 mRNA is expressed in a high lateral to low medial gradient in the CP at the embryonic stages (Inoue et al. 1998). It was also shown that Cdh6 mRNA expression demarcates putative somatosensory and auditory regions and neurons strongly express Cdh6 are confined to their layers II–IV at the postnatal stages (Suzuki et al. 1997; Inoue et al. 1998). However, it remained elusive how Cdh6 expression boundaries among sensory areas are established and maintained in the mouse cerebral cortex.
To analyze details of Cdh6 mRNA expression patterns in the mouse cerebral cortex at the postnatal stages, we collected brain samples at various developmental stages and performed ISH in frontal sections at a comparable level. At the postnatal day 0 (P0), when each layer is not distinguishable in the CP, we found that Cdh6 was broadly expressed in a high lateral to low medial gradient in the CP (Fig. 1A). At P3, when the deep-layer patterning becomes apparent, Cdh6 mRNA expression was found in a similar gradient pattern in the upper tier of the CP (Fig. 1B). Notably, a sharp Cdh6 expression boundary first became evident in the layer IV at P5, when all layer structures segregated each other in the CP (Fig. 1C). At P7, this expression boundary stood more prominent in the layer IV and the somatosensory barrel development appeared to occur within the region that strongly expresses Cdh6 mRNA (Fig. 1D). However, such developmental processes were hard to monitor at the cellular resolutions in ISH analyses as cell population that expresses Cdh6 mRNA in layers II/III was very close to that in layer IV and most of neurons in the cortical layers II–IV were Cdh6 mRNA positive at the postnatal stages.
A Cdh6-BAC::LacZ Transgenic Mouse Line Specifically Recapitulated Cdh6 mRNA Expression in the Layer IV of Somatosensory Cortex at the Postnatal Stages
To better discriminate Cdh6 expression profiles in the cortical layer IV from that in layers II/III, we next examined the beta-galactosidase (β-gal) reporter expression patterns in a Cdh6-BAC-LacZ transgenic (Tg) mouse line that selectively recapitulates Cdh6 mRNA expression in the somatosensory cortex under the control of Cdh6 cis regulatory modules (e.g. enhancers and promoters) included in a BAC clone RP23-78N21 (Fig. 2A) (Inoue et al. 2008b). As the results, we confirmed that 6 out of 7 founder BAC-Tg lines showed a similar β-gal expression pattern in the somatosensory area (Inoue et al. 2008b) and we detailed expression profiles from Line 7 in this study.
By analyzing reporter expression patterns in the somatosensory cortex during the postnatal stages, we found that β-gal was expressed in a high lateral to low medial gradient at P3 in the Tg mice, as was the case with Cdh6 mRNA expression (Fig. 2B). The same pattern was observed at P0, P1, and P2 (data not shown). At P4, in the layer IV of the putative barrel field, conspicuous β-gal staining was observed, yet we could not tell from these observations whether the number of cells that began to express β-gal increased or expression level of each β-gal-positive cell rose during development (Fig. 2C). Notably, in the layers II/III and V/VI, the expression level of β-gal was much weaker than that of Cdh6 mRNA, setting off the layer IV β-gal expression in the BAC-Tg line from this stage. By P5, a sharp β-gal expression boundary was formed in the layer IV and the boundary became more evident at P5 than P4 (Fig. 2C and D). Coincidently, along the sharp Cdh6 expression boundary, the vibrissal barrels development occurred in those cell populations that expressed strong β-gal at P7 (Fig. 2E). At P20, when the somatosensory barrel and limb field become morphologically apparent with functional circuitries, considerable amount of β-gal signals were detected although Cdh6 mRNA expression was relatively weak at the adult stages (GENSAT Images 26035, 26036, 26037, 26982, and 26983 for adult brains: http://www.ncbi.nlm.nih.gov/projects/gensat/; ALLEN INSTITUTE for BRAIN SCIENCE database: http://www.brain-map.org/). Each barrel structure became more prominent at P20 than at P7 probably because β-gal staining was accumulated at the tips of thalamocortical axons from the VPM nucleus (Fig. 2F). Collectively, it was found that, in the layer IV, a β-gal expression boundary was formed at P4 and this boundary was maintained until P20 to correlate with the functional areal border between the barrel and limb field.
A Cdh6-BAC-NLacZ Tg Mouse Line Revealed Qualitative and Quantitative Expression Profile of Cdh6 in the Layer IV of Somatosensory Cortex
While Cdh6-BAC-LacZ reporter expression specifically recapitulated Cdh6 mRNA expression profiles in the layer IV of somatosensory cortex, Cdh6-BAC-LacZ Tg mouse also showed strong β-gal reporter expression in the VP nucleus (Supplementary Fig. 1D). Thus the β-gal expression profiled from the layer IV of the somatosensory barrel field might reflect only the β-gal signals from VPM axons in the Cdh6-BAC-LacZ Tg mouse (Supplementary Fig. 1C). It was also difficult to precisely measure the expression level from individual β-gal-positive neuron in the developing somatosensory cortex. We therefore newly generated Cdh6-BAC-nLacZ mice, in which nuclear localizing signal allowed active accumulation of β-gal protein into cell nuclei (Fig. 3A).
First, we sought to confirm whether the β-gal expression in Cdh6-BAC-nLacZ mice was controlled under cis-regulatory modules included in the Cdh6-BAC clone RP23-78N21 by analyzing Tg brain slice/sections at various developmental stages. Consequently, it was found that β-gal-positive cells in Cdh6-BAC-nLacZ mice were located within the cortical layers II–IV, antero-ventral/antero-dorsal thalamic nuclei (AV/AD), and VPM and that these patterns were quite similar to that of Cdh6-BAC-LacZ mice (Figs 2E and 3D1, Supplementary Fig. 1C and D). In the magnified view, the β-gal signal was localized to the AV/AD and cortical layer IV cell nuclei at high intensity and layer II/III cell nuclei at low intensity (Fig. 3). Importantly, in Cdh6-BAC-nLacZ mice, the anterior commissure was β-gal negative while Cdh6-BAC-LacZ mice showed strong expression there, indicating that β-gal proteins were not localized in axons but in nucleus in the Cdh6-BAC-nLacZ Tg mouse line (Supplementary Fig. 1A and B). Taken together, we clarified that β-gal expression in Cdh6-BAC-nLacZ mice was precisely controlled by Cdh6 cis-regulatory modules in the BAC clone RP23-78N21 and the staining signal was dominantly localized to each cell nucleus. Among 4 Tg founders generated, Lines 1 and 4 recapitulated Cdh6 mRNA expression in the somatosensory cortex and Line 1 showed strong and accurate β-gal expression. We hence hereafter used Line 1 for further analysis.
We next examined how the sharp β-gal expression boundary in Cdh6-BAC-nLacZ mice was established during postnatal development. To this end, we rigorously compared transition of β-gal-positive cell number and compared intensity of each nucleus in 3 defined regions (Regions 1, 2, and 3 in Fig. 3B1, C1, and D1) from the somatosensory cortex and 1 region (Region 4 in Fig. 3B1, C1, and D1) from AV during development. As the consequence, at P3, we realized that the number of β-gal-positive nucleus was smaller than the negative one and the intensity of β-gal-positive nucleus was relatively weak in Region 1 (Fig. 3B2). Also, in Region 3, we found that β-gal-positive nuclei were limited in number and the expression level of individual nucleus was very low (Fig. 3B4). At this stage, boundary made by β-gal-positive nuclei was unclear in Region 2 (Fig. 3B3). At P5, the number of β-gal-positive nucleus in Regions 1 and 3 appeared to increase and quantitative difference in β-gal-positive nuclei between Regions 1 and 3 was more prominent at P5 than at P3 with the intensity of individual β-gal-positive nucleus in Region 1 being stronger than that in Region 3 (Fig. 3C2, C3, and C4). At P7, both the number of β-gal-positive nuclei and intensity of each nucleus seemed to increase exclusively in Region 1 and the difference in their parameter was more prominent at the boundary region 2 (Fig. 3D2, D3, and D4).
For further quantification, we averaged the number of β-gal-positive nucleus in the 3 regions (Regions 1, 3, and 4) at P3, P5, and P7 (Fig. 4A). In Region 1, it was revealed that the number of β-gal-positive nucleus progressively increased during cortical development. As the results, the number of β-gal-positive nucleus at P5 reached twice as large as that in the Region 1 at P3 (Fig. 4A1), while in Region 3, the number of β-gal-positive nucleus just enlarged gradually (Fig. 4A2). Increasing the rate of the number of β-gal-positive nucleus in Region 1 was hence larger than that in Region 3 during these developmental stages (Fig. 4A1 and A2). On the other hand, the number of β-gal-positive nucleus remained unchanged in the thalamic region 4 during development (Fig. 4A3). We also classified intensity of each β-gal-positive nucleus in both Regions 1 and 3 at these stages with the highest expression being defined as level 5 (Fig. 4B). At P3, 27.9% (n = 3), 43.8% (n = 3), and 28.3% (n = 3) of all β-gal-positive nuclei showed Level 3, 2, and 1 intensity, respectively, in Region 1 (Fig. 4B1). Meantime, nearly 80% (n = 3) of β-gal-positive nuclei showed the lowest level 1 expression and the rest of nuclei expressed level 2 intensity in Region 3 (Fig. 4B4). In the cortical region 1, almost 40% (n = 3) of all β-gal-positive nuclei expressed Level 4 intensity and another 40% (n = 3) nuclei showed Level 3 expression at P5 (Fig. 4B2). On the other hand, in the cortical region 3, at P5, 6.3% (n = 3), 35.6% (n = 3), 39.5% (n = 3), and 18.6% (n = 3) of β-gal-positive nuclei expressed Level 4, 3, 2, and 1 intensity, respectively (Fig. 4B5). At this stage, it was confirmed that intensity of each β-gal-positive nucleus in Region 1 was higher than that in Region 3 (Fig. 4B2, B5, and C2). At P7, in the cortical region 1, β-gal-positive nuclei with Level 5, 4, and 3 intensity were 44.7% (n = 3), 40.3% (n = 3), and 14.5% (n = 3) of all β-gal-positive nuclei, while in Region 3, the percentage of β-gal-positive nuclei was 30.1% (n = 3) for Level 4, 50.9% (n = 3) for Level 3, and 16.1% (n = 3) for Level 2 (Fig. 4B3 and B6). This demonstrated that the difference of β-gal signal intensity between the cortical regions 1 and 3 was maintained at P7 (Fig. 4B3, B6, and C3).
From these results, we concluded that the increasing rate of β-gal-positive nuclei in the cortical region 1 was larger than that in the cortical region 3 and that the expression level of each nucleus in Region 1 was higher than that in Region 3, although β-gal-positive number and intensity of each nucleus increased both in Regions 1 and 3 during cortical development (Fig. 4A and B). Thus, the sharp expression boundary observed in ISH analysis for Cdh6 mRNA or Cdh6-BAC-LacZ mice could be formed by differences from the increasing rate of β-gal-positive nuclei as well as the expression level of each nucleus between Regions 1 and 3 during cortical development (Fig. 4C).
Generation of Cdh6-BAC-CreERT2 Tg Mouse Lines
The ISH analysis for Cdh6 mRNA or the reporter expression analysis of Cdh6-BAC-LacZ Tg and Cdh6-BAC-nLacZ Tg mice indicated that the sharp Cdh6::LacZ gene expression boundary formed as early as P5 in the layer IV (i.e. Region 2 in Fig. 3C1) might coincide with the functional areal border between the barrel and limb field at the adult stage. However, it is still unclear if the Cdh6::LacZ expression boundary at P5 constituted a compartment boundary for neocortical areas since cells might change their gene expression during cortical development at around the expression border and the position where Cdh6 expression boundary formed at around P5 could be balanced over development due to transcriptional regulatory mechanisms.
To further examine the relationship between Cdh6::LacZ expression boundary at P5 and functional cortical area border at the adult stages, we newly generated 8 Tg founders of Cdh6-BAC-CreERT2 Tg, which can label Cdh6-expressing cells in a temporally specific manner by means of TM administrations (Feil et al. 1997; Fig. 5A). To detect CreERT2 proteins in the cerebral cortex of Cdh6-BAC-CreERT2 Tg mice, we first carried out western blot analysis on brain lysate with an antibody for Cre (Novagen). Consequently, among 8 Tg lines generated, Line 8 showed a highest level of CreERT2 protein expression (data not shown). Next, to investigate whether cells were stably labeled on the basis of cis-regulatory modules of Cdh6-BAC clone RP23-78N21, we administered TM to the pregnant Cdh6-BAC-CreERT2; ROSA26R females at E10.5 and detected the initial population of marked cells in the E12.5 Cdh6-BAC-CreERT2; ROSA26R embryos. By comparing the β-gal expression pattern to Cdh6-BAC-LacZ embryos at E12.5, we found that labeled cells were located only within Cdh6-expressing domain such as the first branchial arch and dorsal midline cells in Tg Lines 4, 5, 6, and 8 (an arrow and arrowheads in Fig. 5B and C). Notably, compared with the Cdh6-BAC-LacZ Tg line, relatively small number of cells was marked in the first branchial arch at E12.5 by TM administration at E10.5, highlighting the temporal specificity of cell labeling in using Cdh6-BAC-CreERT2 Tg lines. To further examine whether it is possible to mark cells in the cerebral cortex at the postnatal stages under the control of cis-regulatory modules in the Cdh6-BAC clone RP23-78N21, we administered TM to Cdh6-BAC-CreERT2; ROSA26R pups at P5 and analyzed the distribution of the initially labeled cell population at P7. As the result, Cdh6-BAC-CreERT2-marked cells in Tg mouse Line 8 were positioned within the Cdh6-expressing region including the barrel field at P7 (Fig. 5D and E). We finally checked the distribution of β-gal-positive cells without TM administration and confirmed that there were no ectopically marked cells in the cerebral cortex. This demonstrated that the labeling system functioned precisely in our Cdh6-BAC-CreERT2 Tg mouse line 8 if mated with the Rosa26R mouse line (Fig. 5F and G). Hereafter, we used Tg mouse line 8 out of 8 founders for further analysis because of its high spatio-temporal efficiency and accuracy in labeling the Cdh6-expressing cells during cortical development.
Cdh6::LacZ Gene Expression Boundary Formed at Around P4 Corresponded to a Functional Areal Boundary Between the Barrel and Limb Field at P20
To clarify whether the Cdh6::LacZ gene expression boundary found at around P5 corresponded to a functional areal border between the barrel and limb field at P20, we labeled Cdh6::LacZ-expressing cells at different postnatal stages by administering TM to Cdh6-BAC-CreERT2; ROSA26R mice and examined the distribution pattern of labeled cells at P20. We first injected TM to Cdh6-BAC-CreERT2; ROSA26R pups at P4 and P5, when the sharp Cdh6::LacZ expression boundary becomes clear, and checked the β-gal-positive cell distribution pattern at P20. As the result, we found that the density of labeled cells was high in the barrel field while it was low in the limb field and a remarkable change of marked cell density was observed at P20 (Fig. 6C and D). This sharp staining boundary appeared to fall at the area where prominent change in the β-gal signal from tips of thalamocortical axons was observed at the border between the barrel and limb field in Cdh6-BAC-LacZ mice (Supplementary Fig. 2). To further quantify the density of Cdh6-BAC-CreERT2-marked cells in the layer IV of somatosensory cortex, we plotted intensities of β-gal-positive cells along the tangential coordinate (bins 1–10) within a rectangular area. Consequently, we found that a sharp staining boundary marked by TM administration in Cdh6-BAC-CreERT2; ROSA26R at P4 and P5 located between bins 7 and 8. We also confirmed that a remarkable change in β-gal signals in P20 Cdh6-BAC-LacZ mice occurred between bins 7 and 8 (Supplementary Fig. 2). These results supported that the location where the remarkable change in marked cell density was observed coincides with the border between the somatosensory barrel and limb field. Notably, the β-gal staining pattern at the boundary area was unchanged after the short-term labeling. Collectively, it was concluded that the sharp Cdh6::LacZ expression formed in the layer IV by P5 during cortical development corresponded to the somatosensory areal border between the barrel and limb field at P20.
On the other hand, there was no significant difference in marked cell density at the P20 areal border when labeling by TM injection was performed at P0, P1, and P2, which apparently contrasted to the marking results from P4 and P5 (Fig. 6A). Also, intensity of β-gal signals at P20 did not show the remarkable change between bins 7 and 8 when we administered TM to Cdh6-BAC-CreERT2; ROSA26R pups at P0, P1, and P2 (Fig. 6A). When we injected TM at P3, the β-gal staining pattern at P20 showed intermediate results: 8 out of 15 cases showed the P0–P2-type non-boundary phenotype and 7 out of 15 cases indicated the P4–5-type pattern with the apparent expression boundary at P20 (Supplementary Fig. 3). This could be explained by our finding that the Cdh6::LacZ expression profile changes dynamically from P3 to P5 (Figs 2 and 3) or the property that the CreERT2-mediated recombination occurs in a wide time window (e.g. 12–24 h) after TM administration (Zervas et al. 2004). Whatever the case might be, these results suggest the trend that Cdh6::LacZ expression border, which corresponds to a cell fate restricted boundary between the functional somatosensory barrel and limb field, is established immediately at around P4.
Cdh6::LacZ-Positive Cortical Neurons Formed Aggregates in the Pellet Culture, but not Under the Presence of EGTA
Further to examine whether the Cdh6::LacZ expression border established at around P4, which corresponds to a cell fate restricted boundary between the functional somatosensory barrel and limb field at P20 (Fig. 6), has some physiological significance in the process of cortical arealization, we dissociated the cells from the P6 Cdh6-BAC-nLacZ Tg mouse brain and evaluated Cdh6::LacZ-positive cell distributions in the pellet culture system. By using the system, it was already shown that dissociated mouse limb mesenchymal cells expressing cadherin-11 were completely sorted from those without cadherin-11 expression to form cell aggregates, which highlights the role of cadherins in limb tissue formation and/or maintenance (Kimura et al. 1995). In the present study, we dissociated cells from the barrel cortex at P6 and their pellets were incubated on the filter membrane floating over the medium with or without EGTA. Simultaneously, dissociated cells were plated on culture dishes to check their viability in vitro. After 72-h incubation in vitro, we confirmed that ∼40% of plated cells on culture dishes strongly express LacZ and that all LacZ-positive cells harbor immunoreactivity with the anti-neurofilament monoclonal antibody 2H3 (arrows in Fig. 7A–C), indicating that these cells stably maintain neuronal character after dissociation. We then processed the cortical cell pellets for LacZ staining after 72-h incubation in vitro and found out that as many as 30 LacZ-positive cell bodies aggregate to generate sphere-like cyto-architectures in the pellet without EGTA (Fig. 7D and E), whereas LacZ-positive cell bodies distributed randomly under the presence of EGTA (Fig. 7F and G) with 100% reproducibility. Notably, the number of LacZ-positive nuclei was at a similar level in a pellet regardless of EGTA. Considering the selective chelating action of EGTA on Ca2+ under physiological conditions, cell aggregation activity observed in the pellet culture might dominantly be dependent on Ca2+. This implicates a role of cadherin cell adhesion system in establishing and/or maintaining the areal border and/or barrel-specific cyto-architecture demarcated by their expressions during the cortical development.
Electroporation-Based Cdh6 Overexpression in the Postnatal CP Revealed a Role of Cdh6 in Areal Protomap Solidification
In order to more directly test the role of cadherins in areal map elaboration, we first analyzed postnatal brains from Cdh6 gene-targeting mice, yet we could detect few apparent phenotypes in their neocortical arealization probably due to redundant cadherin subclass expressions in the developing CP (data not shown, Inoue et al. 2001; Osterhout et al. 2011). We thus next sought to perturb cadherin gene expression patterns in the postnatal CP by means of Cdh6 gene overexpression experiments. Since we had already noticed that simple in utero electroporation of Cdh6 gene expression vectors into the telencephalic VZ at E13.5–E14.5 caused defects in neuronal emigration from VZ and/or radial migration modes before cells reach the CP (data not shown), here we applied a tet-ON inducible expression system with the Tol2 transposable elements that allows efficient integration of expression vector plasmids into chromosomes as well as strict control of exogenous gene expression on-set during the development (Supplementary Fig. 4; Kawakami and Noda 2004; Sato et al. 2007). By electroporating such a mixture of expression vector plasmids into the E13.5–14.5 telencephalic VZ in utero and administrating doxycycline (Dox) at the postnatal stages, we could minimize defects in neuronal migration and confirm that most of electroporated cells normally settle in the CP layer II/III–IV at P7 (Fig. 8B; Supplementary Fig. 5B). Then we closely looked at how ectopic expression of Cdh6 limited to the postnatal stages affects CP cyto-architectures by the cytochrome oxidase (CO) staining method. As the results, we found that the CO staining pattern along the cortical layer IV was totally disrupted only along the area where intense Cdh6 expression was induced by Dox in the postnatal CP layer II/III–IV (Fig. 8D1, D2; Supplementary Fig. 5D1, D2). Importantly, the CO staining pattern at the control side on the same section was completely normal (Fig. 8C1, C2; Supplementary Fig. 5C1, C2). Moreover, in the control experiments where the Tol2 empty expression vector was electroporated, no perturbation was observed in the CO staining pattern (Supplementary Fig. 6). Together with the aforementioned results from CP cell pellet culture, it is suggested that strong Cdh6 expression on the CP cell surface is sufficient to elaborate the area-specific cyto-architecture in the CP and that Cdh6 expression dynamics in the postnatal CP harbors a role in areal map formation and/or maintenance.
The ION Transection Resulted in Reduction of Cdh6 Expression along the CP Layer IV Neurons in the Barrel Field
Finally, we examined the effects of neuronal activity on Cdh6 expression in the postnatal CP by means of unilateral ION section at P0. As was reported in the previous study by Jabaudon et al. (2012), we confirmed that ION transection at P0 resulted in reduction of the cortical RORβ as well as thalamic VGLUT2 staining signal only at the contralateral side of the operation (Fig. 9A and B). Then we evaluated Cdh6 expression profiles recapitulated in the Cdh6-BAC-nLacZ Tg mouse line along the cortical layer IV neurons within the barrel field after ION transection and found that the Cdh6 expression level at P7 was deteriorated specifically and significantly at the side where the thalamocortical inputs were deprived (Fig. 9C). These results highlight that activity-dependent machinery at least in part plays a role in maintaining the Cdh6 expression dynamics in the postnatal CP.
In the present study, ISH analysis for Cdh6 mRNA as well as the reporter expression profiling in the Cdh6-BAC-LacZ or Cdh6-BAC-nLacZ Tg mice revealed that a sharp Cdh6 gene expression boundary is formed in the cortical layer IV by P5. By means of the cell fate tracing strategy with Cdh6-BAC-CreERT2 Tg mice, we further indicated that the sharp Cdh6 expression boundary formed at around P4 corresponds to the future functional areal border between the barrel and limb field at P20 which can morphologically be identifiable by the maturation of thalamocortical axon terminals from VP nucleus. Cells dissociated from the P6 CP showed Ca2+-dependent segregation and Cdh6 overexpression limited to the postnatal stages perturbs CP cell organizations in the somatosensory barrel field. These results provide evidence that a functional areal border is ascribed to a cadherin gene expression boundary demarcated at the early postnatal stages in CP.
Gene Expression Code of Classic Cadherins in the CP at the Perinatal Stages Might Outline the Future Functional Area Map
The protomap hypothesis originally suggested that differences among discrete functional areas are genetically set in the cortical primordium, although the initial map might be modified in the CP by cellular dynamics and/or neuronal connectivity (Rakic 1988; Rakic et al. 2009). In the embryonic cortical VZ, none of genetically distinct areas defined by region-specific gene expressions were observed while gradients of transcriptional factors such as Pax6, Emx2, Sp8, and Coup-tf1 along the rostrocaudal and mediolateral axes in the VZ were shown to convey positional identities in the cortical progenitors (Bishop et al. 2000; Fukuchi-Shimogori and Grove 2001; Monuki and Walsh 2001; Grove and Shimogori 2003; Funatsu et al. 2004; Hamasaki et al. 2004; Mallamaci and Stoykova 2006; Armentano et al. 2007; O'Leary et al. 2007). In the postnatal CP, transcription factors and/or cell adhesion/repulsion molecules initially showed graded distributions in the VZ/CP dynamically change their expression pattern to harbor sharp expression boundaries (Bulchand et al. 2003; Nakagawa and O'Leary 2003; Cang et al. 2005; Takeuchi et al. 2007; Joshi et al. 2008; Huang et al. 2009; Bedogni et al. 2010) and the lifespan analysis of several gene expression profiles and intraneocortical connections revealed that several gene expression domains actually seemed to correlate with the future areas from the perinatal stages to adult stages (Dye et al. 2011a, 2011b). For instance, RORβ expression boundary seemed to co-register the medial boundary of the barrel field from the perinatal stages to adulthood (Nakagawa and O'Leary 2003; Hirokawa et al. 2008; Dye et al. 2011a, 2011b). However, the direct linkage between these gene expression profiles and functional area map largely remained elusive.
In the present study, we showed in mouse that a Cdh6 sharp expression boundary formed at around P4, when the functional barrel circuitries with integrated interneurons are still under development (Rice et al. 1985; Agmon et al. 1993; Itami et al. 2007; Kichula and Huntley 2008; Espinosa et al. 2009; Sehara et al. 2010), precisely matches to the areal border between the somatosensory barrel and limb field in the layer IV at P20 (Figs 1–3 and 6). Notably, classic cadherins including Cdh6 appear to demarcate other functional areal borders in the developing mouse cortex. For example, Cdh6 expression domain correlates to the developing auditory region and Cdh8 and Cdh11 expression differentially delineate the frontal cortex at the postnatal stages (Suzuki et al. 1997). Curiously, ferret classic cadherins showed a conserved expression pattern in the developing CP compared with mouse ones and some of these apparently delineate functional areas. For instance, Cdh8 and Cdh20 expression complementarily demarcate the V1/V2 areal border in the developing cortex (Suzuki et al. 1997; Krishna et al. 2009). Combinatorial expression code of classic cadherins might thus emerge as the first recognizable protomap in the nascent CP beyond species.
Genetic Machineries that Immediately Restrict Cell Lineages within Functional Areas at the Early Postnatal Stages
Previous series of studies demonstrated that quantitative and qualitative difference of classic cadherins could mediate the selective adhesion of cells in vitro (Nose et al. 1988; Steinberg and Takeichi 1994). It was also known that individual classic cadherin showed a distinct expression pattern in restricted subdivisions of central nervous system (CNS) and differential and/or complementary expression patterns among classic cadherins have important roles in CNS regionalization and/or compartmentalization by controlling cell adhesiveness (Inoue et al. 1997, 2000; Redies 2000; Inoue et al. 2009). For example, the complementary expression pattern of Cdh4 and Cdh6 was responsible for maintaining the compartment boundary between the cortex and the striatum during embryonic stages (Inoue et al. 2001). Combinatorial expression code of classic cadherins could therefore endow specific adhesive properties to each putative functional area in the nascent CP. As for the developing border between the somatosensory barrel and limb field, zone with the strong Cdh6-expressing cells at P6 indeed harbored differential adhesiveness to form aggregates in the pellet culture system (Fig. 7). In addition, the postnatal stage-specific induction of ectopic Cdh6 expression perturbed the somatosensory barrel structure (Fig. 8; Supplementary Fig. 5), implicating the functional relevance of differential cadherin expression profiles in the nascent CP. Could this be the sole genetic machinery for cortical protomap formation?
In the nascent CP, EphA receptors and ephrinA ligands also showed complementary expression patterns at the postnatal stages (Cang et al. 2005; Depaepe et al. 2005). Eph/ephrin signaling generally functions as repellant which prevented cells or axons from entering inappropriate territories and are involved in establishment and/or maintenance of CNS compartment by actively controlling cell sorting and cell–cell communications (Mellitzer et al. 1999; Xu et al. 1999; Wilkinson 2001; Solanas et al. 2011). Functional imaging analysis further revealed that the primary visual field (V1) was rotated and shifted medially in the ephrin-A2/A3/A5 triple knockout mice, while ectopically overexpressed ephrinA ligands in the lateral cortex disrupted the retinotopic map of V1 (Cang et al. 2005). Hence, with classic cadherins, the Eph/ephrin system might have significant roles in the formation and/or maintenance of the functional areal borders in the nascent CP by cooperatively controlling cell adhesive properties for dynamic sorting (Solanas et al. 2011). Curiously, a previous report indicated that programmed cell death might be activated at the complementary expression boundary between ephrin-A5 and EphA7 in the nascent CP to prevent cells intermingling between the future functional areas (Depaepe et al. 2005). Such machinery could synergistically facilitate patterned cell segregation and/or protomap formations in the CP. In this context, noticeable is the recent report that a transcription factor RORβ is sufficient to induce the periodic barrel-like clustering of neurons in the mouse CP (Jabaudon et al. 2012). As was also discussed by Jabaudon et al., identification of RORβ downstream targets would further reveal the genetic machinery that is required to elaborate the cortical area-specific cyto-architecture during development and our ION transection results (Fig. 9) as well as information from Cdh6 enhancer analysis for the somatosensory barrel-layer IV-specific expression (Terakawa et al. 2011) now raise a tantalizing possibility that Cdh6 expression in the developing barrel field is a direct downstream target of RORβ whose expression is sufficient to emerge barrel-like cell clusters where precise gene expression level of RORβ is modulated by neuronal activity from the thalamocortical axons at the postnatal stages (Jabaudon et al. 2012).
In summary, we demonstrated that the future areal border between the somatosensory barrel and limb field in the developing CP can be specified by the sharp Cdh6 gene expression boundary formed as early as P4. We also found that Cdh6 strongly positive neurons dissociated from the P6 CP were Ca2+ dependently sorted to form clusters in cell pellet culture system and that the ectopic induction of Cdh6 expression in the postnatal CP disrupted the barrel cluster organizations in layer IV. These results suggest that cadherin expression profiles in the nascent CP are pertinent to define a rigid protomap of the functional areas by conferring specific adhesive properties to each CP neuron. Identification of upstream regulators for cortical area-specific expression of cadherins would be an important next step in fully understanding genetic cascades for the cortical area map elaboration and its evolutionary traits.
This work was supported by grants from Takeda Science Foundation, The Nakatomi Foundation and Research Foundation ITSUU Laboratory, Program for Promotion of Fundamental Studies in Health Sciences of the National Institute of Biomedical Innovation (05-32), Intramural Research Grant (24-12) for Neurological and Psychiatric Disorders of NCNP and Grant-in-Aid for Scientific Research from JSPS (#21500333; #24300130) to T.I.
We thank Dr. Robb Krumlauf for pBGZ40 vector, Dr. Pierre Chambon for pCre-ERT2, Dr. Neal Copeland for the recombineering-related materials including EL250 bacterial stain, and Drs. Koichi Kawakami and Yoshiko Takahashi for Tol2 related materials. We also acknowledge Drs. Takayuki Sota, Shinichi Kohsaka, Keiji Wada, and Dr. Hoshino's lab members for discussions and encouragements. Conflict of Interest: None declared.