Abstract

Subplate cells in the mouse are generally defined as cells located in the subplate layer between the white matter and layer 6a. They are some of the earliest born and maturing cells of the cerebral cortex. The postnatal subplate layer in mouse contains neurons with expression of the presynaptic protein complexin 3 (Cplx3), connective tissue growth factor (CTGF), the orphan nuclear receptor Nr4a2 (Nurr1), and the G-protein-coupled lysophosphatidic acid receptor 1 (Lpar1/Edg2). All 4 of these molecular markers show layer 6b-restricted expression at young postnatal ages, with CTGF expression being the most widespread in the young postnatal subplate. However, all 4 markers overlap in their expression pattern to varying degrees. Here we demonstrate with bromodeoxyuridine birthdating that cells labeled with any 1 of these molecular subplate markers are indeed generated at E11.5 or E12.5 in the mouse. Furthermore, we demonstrate a correlation between gene expression and cell birthdates. Lpar1-GFP cells are preferentially generated on E11.5, whereas Cplx3 or Nurr1-positive cells are equally generated during the 2-day peak of subplate neurogenesis (E11.5–E12.5). Our study also demonstrates that early-born subplate neurons labeled by Cplx3, Nurr1, and Lpar1-GFP survive preferentially after the first postnatal week compared with other subplate neurons.

Introduction

Mouse subplate cells are generally defined as cells located in the subplate layer between the white matter (WM) and layer 6a and are among the cortical cells that are born earliest and mature first (Angevine and Sidman 1961; Price et al. 1997; Smart et al. 2002; Bystron et al. 2008). Subplate neurons are recognized to be important players in cortical development and maturation, with distinct roles at different developmental ages (Kanold and Luhmann 2010). At embryonic ages, subplate cells are involved in thalamocortical and corticofugal axon guidance, including establishment of topographical projections (McConnell et al. 1989; Ghosh et al. 1990; Grant et al. 2012). Perinatally, subplate neurons play a role in generating cortical oscillations (Dupont et al. 2006; Yang et al. 2009), and postnatally, they are essential for the maturation of plasticity and the inhibitory circuitry in cortical layer 4 (Kanold et al. 2003; Kanold and Shatz 2006). In the adult, they may support cortico-cortical connectivity (Kostovic and Rakic 1990; Friedlander and Torres-Reveron 2009; Suarez-Sola et al. 2009; Kostovic et al. 2011).

In contrast to their many and well-characterized functions, there are conflicting reports of significant cell death during the early postnatal period (Price et al. 1997) or very little cell death (Valverde et al. 1995). More in agreement with the latter, there are also reports of persistent immunohistochemical or in situ hybridization labeling of the postnatal and adult subplate layer (Arimatsu et al. 2003; Heuer et al. 2003; Watakabe et al. 2007; Hoerder-Suabedissen et al. 2009).

All of the early birthdating studies on rodent preplate and subplate cells were of rats. Using tritiated thymidine, Rickmann et al. (1977) and Bayer and Altman (1990) determined that rat preplate neurons are generated from E14 (plug date = E1). Cells born on E14 are destined for the future marginal zone, whereas cells born on E15 eventually contribute to both marginal zone and subplate zone (Rickmann et al. 1977; Bayer and Altman 1990). Some cells located in the late embryonic or postnatal rat subplate zone are also generated after E15 (Rickmann et al. 1977; Raedler and Raedler 1978). Furthermore, Bayer and Altman (1990) followed the distribution of E14-born neurons in the lateral embryonic cortex and concluded that they are dispersed in the cortical plate at E16 but aggregate at the lower edge of the cortical plate by E17 and form a separate band below the cortical plate at E18, thereby highlighting that subplate cells may be born early but that the subplate layer is not formed as a cytoarchitectonically distinct structure until after the cortical plate has emerged.

In the first autoradiographic study in mouse, Smart and Smart (1982) did not report a continued contribution to the marginal zone and subplate layer in their analysis of juvenile (P22) mouse brains. Depending on the mediolateral position, cells within the subplate layer were primarily generated at E11 and E12 (plug date = E1), with a few more contributed by E13 divisions.

In this study, we determined the distribution of cells in the P1 mouse cortex that were born between E10.5 and E13.5 and, in particular, show that there is continuous addition of cells to the subplate zone and underlying WM during this 72 h period for the NIHS mice used in this study. We also examined possible relationships between birthdates and molecular marker expression in subplate.

Robertson (2000) described somatostatin, neuropeptide Y, and acetylcholine esterase as functional markers of the embryonic rat subplate; yet, none of these distinguishes the subplate layer from the overlying cortical plate in postnatal (P21) brains. This prompted our search for molecular markers of the postnatal subplate (Hoerder-Suabedissen et al. 2009), in which we reported the presynaptic protein complexin 3 (Cplx3), connective tissue growth factor (CTGF), and the orphan nuclear receptor Nr4a2 (Nurr1), among others, as molecules that are consistently expressed in the mouse subplate layer during the postnatal period. Additionally, our screen identified the G-protein-coupled lysophosphatidic acid receptor 1 (Lpar1/Edg2) as another putative marker of the postnatal subplate (Hoerder 2007). None of these, however, is expressed from the time of subplate cell generation. Thus, we set out to determine the birthdates of the marker-labeled subplate neurons. We also examined whether subplate neurons with different gene expressions can be distinguished based on their birthdates.

Mouse subplate neurons do not just lose the expression of some of their putative functional markers during the postnatal period, but, according to some studies, the majority of the early-born cells disappear outright from the postnatal cortex, similar to cat (Luskin and Shatz 1985a, 1985b). Price et al. (1997) analyzed the distribution of bromodeoxyuridine (BrdU)-labeled nuclei in the postnatal mouse (Balb/c) cortex and concluded that cells born on E11 (E1 = day following overnight mating, referred to as E10.5 in this study) had completely disappeared from the cortex by P8. Similarly, the majority of E12- and E13-born cells contributed to the subplate layer at birth, but not at all by P21. At P8, some E12- and very few E13-born cells were still present in the subplate. We replicated some of these findings, extending them to a general loss of early-born neurons during the first postnatal week, and determined whether there is a differential survival among marker-labeled neurons of the mouse subplate.

Materials and Methods

Definition of Subplate

For the purposes of this study, the subplate zone or layer was defined as a 50 µm thick band directly above the WM (Hoerder-Suabedissen and Molnár 2012) and is equivalent to layer 6b. Anything above that was considered cortical plate. All 4 molecular markers used here label a thin band of cells in the subplate location with very few cells above (in layer 6a) or below (in WM). Neurogenesis was also monitored in a 50 µm thick band below the subplate (i.e. in WM) as surviving subplate neurons in primates and rodents have been reported in the WM (Judas et al. 2010; Viswanathan et al. 2012). Finally, as all 4 subplate markers used here are known to label cells in the adult subplate layer, we also determined whether marker-labeled early-born (E11.5 and E12.5) subplate neurons preferentially survive after the first postnatal week.

Animals

All animal experiments were approved by a local Ethics Review Committee and conducted in accordance with personal and project licenses under the UK Animals (Scientific Procedures) Act (1986). Wild-type (WT) NIHS females (supplied by Harland, UK) were mated for 12 h overnight with Lpar1-GFP males [Tg(Lpar1-EGFP)GX193Gsat, on an NIHS background] and checked for plugs. Midday after the mating was considered E0.5, and pregnant dams were injected i.p. with 100 mg/kg BrdU in sterile saline (BD Biosciences, Oxford, UK) on E10.5, E11.5, E12.5, or E13.5. For the following experiments, only pups born on E20 were used (i.e. 10 of 11 litters, with the remaining litter being born a day early). WT (female) and Lpar1-GFP (male) pups were killed by cervical dislocation at midday on P1 or P8, the brains were dissected out and washed in ice-cold 0.1 M phosphate-buffered saline (PBS), and immersion-fixed in 4% paraformaldehyde (PFA) in 0.1 M PBS, pH 7.4 (4% PFA; TAAB, Aldermaston, UK) for 24 h. Furthermore, Lpar1-GFP and WT pups from similar overnight matings but without BrdU injections were transcardially perfused with 4% PFA and postfixed in 4% PFA overnight for marker colocalization analysis. For temporal GFP expression analysis, additionally, adult Lpar1-GFP males were obtained. Timed-pregnant females were killed by cervical dislocation and embryos harvested at E13.5 and E15.5. Embryonic heads were postfixed in 4% PFA overnight.

GFP-Expression Analysis

Brains of Lpar1-GFP animals aged E13.5, E15.5, P0, P2, P4, P6, P8, P10, P14, P21, and adult (n= 2 brains or more for each age) were cut to 50 µm coronally on a vibrating microtome. Following counterstaining with DAPI (4′,6-diamidino-2-phenylindole, 5 µg/mL, Invitrogen), GFP+ cell distribution was documented on an epifluorescence microscope. The distribution of GFP+ cells and GFP+NeuN+ neurons (discussed subsequently) was quantified from confocal images at P8 (n= 3 brains). For this, 50 µm wide bins centered on SP, lower L6a, and upper L6a and L5 were used. Furthermore, at P8, anti-GABA immunohistochemistry was performed on glutaraldehyde-fixed (4% PFA + 0.25% glut) brains (n= 3), following the protocol discussed subsequently, and quantified from confocal images.

Immunohistochemistry

Immunohistochemistry was performed on sections from 4% PFA-fixed brains cut coronally at 50 μm on a vibrating microtome (VT1000S, Leica), for brains aged either P1 or P8. Fluorescent double immunohistochemistry was performed on 3 to 4 free-floating sections per brain and antibody combination. The sections included the anterior commissure (AC) and ranged from approximately 250 μm anterior to AC to approximately 500 μm posterior to AC. Immunohistochemistry was performed against Cplx3, CTGF, or Nurr1 (WT brains) or GFP (GFP brains)—subsequently referred to as “subplate marker”—and BrdU. Alternatively, immunohistochemistry was performed against the pan-neuronal marker NeuN and BrdU. DNA was denatured in 1 M HCl (Sigma, Gillingham, UK) at 38.5°C for 1 h. Following washes with 0.1 M PBS, sections were blocked in blocking solution [2% donkey serum (Sigma) and 0.1% Triton X-100 (BDH, Poole, UK)] for 2 h at room temperature (RT) before being incubated overnight at 4°C with the primary antibodies in the blocking solution. The following antibodies were used: anti-BrdU (11200, Promega, 1 : 500), anti-Cplx3 (122301, Synaptic Systems, 1:1000), anti-CTGF (sc-14939, Santa Cruz, 1:500), anti-Nurr1 (AF2156, R&D Systems, 1:100), anti-GFP (A11122, Molecular Probes, 1:500), and anti-NeuN (ab104225, Abcam, 1:500). Note, whenever anti-CTGF was used, that the Triton X-100 concentration in the blocking solution was increased to 1%. Following further washes, the sections were incubated with secondary antibody in the blocking solution [biotinylated donkey anti-mouse (Abcam, ab7060, 1:100) and Alexa488-conjugated donkey anti-rabbit (Molecular Probes, A21206, 1:500) or donkey anti-goat (Molecular Probes, A11055, 1:500) for 2 h at RT]. Biotinylated secondary antibody was further labeled with streptavidin-cy3 (1:500; Jackson ImmunoResearch, Newmarket, UK) in blocking solution for 2 h at RT.

Alternatively, the proportion of subplate neurons expressing of each of the various subplate markers or pairs of markers was determined, using P1 and P8 Lpar1-GFP or WT brains. Sections were immunoreacted for NeuN (1:1000 Chemicon, MAB377 or 1:500 Abcam, ab104225) and the subplate markers (either singly or in pairs. To determine the expression overlap of the various subplate markers, sections from P8 Lpar1-GFP brains were immunoreacted for combinations of 2 subplate markers (i.e. CTGF and Cplx3 or Nurr1 and Cplx3). The same antibodies and concentrations as mentioned earlier were used in addition to biotinylated donkey anti-rabbit (1:100, Abcam, ab6801) and streptavidin-cy5 (1:200, Jackson ImmunoResearch). Anti-GABA (Sigma, A2052, 1:2500) immunohistochemistry was performed on P8 Lpar1-GFP brains that were fixed with 4% PFA and 0.25% glutaraldehyde. The same protocol as described earlier was followed with the exception of the initial denaturation step. Sections were counterstained with DAPI (5 µg/mL, Invitrogen).

For antibody combinations see Table 1, and for details about the brains used for BrdU immunohistochemistry see Table 2.

Table 1

Overview of antibodies used for this study, including manufacturer, product code, and concentration it was used at, as well as which secondary antibody it was combined with for which experimental paradigm

Primary antibody Product code Manufacturer Concentration Secondary antibody Experiment 
BrdU 11200 Promega 1:500 D anti-m biotin Birthdating 
Cplx3 122301 Synaptic Systems 1:500 (BrdU) or 1:1000 D anti-rb Al488 or biotin Birthdating and marker colocalization 
CTGF Sc-14939 Santa Cruz 1:500 D anti-gt Al488, Al568, biotin Birthdating and marker colocalization 
GFP A11122 Molecular Probes 1:500 D anti-rb Al488 Birthdating 
NeuN MAB377 Chemicon 1:100–1:1000 G anti-m Al568 Marker colocalization 
NeuN/Fox3 Ab104225 Abcam 1:500 D anti-rb Al488, Al568 Birthdating 
Nurr1 AF2156 R&D Systems 1:100 D anti-gt Al488, Al568, biotin Birthdating and marker colocalization 
GABA A2052 Sigma 1:2500 G anti-rb Al546 Colocalization with Lpar1-GFP 
Primary antibody Product code Manufacturer Concentration Secondary antibody Experiment 
BrdU 11200 Promega 1:500 D anti-m biotin Birthdating 
Cplx3 122301 Synaptic Systems 1:500 (BrdU) or 1:1000 D anti-rb Al488 or biotin Birthdating and marker colocalization 
CTGF Sc-14939 Santa Cruz 1:500 D anti-gt Al488, Al568, biotin Birthdating and marker colocalization 
GFP A11122 Molecular Probes 1:500 D anti-rb Al488 Birthdating 
NeuN MAB377 Chemicon 1:100–1:1000 G anti-m Al568 Marker colocalization 
NeuN/Fox3 Ab104225 Abcam 1:500 D anti-rb Al488, Al568 Birthdating 
Nurr1 AF2156 R&D Systems 1:100 D anti-gt Al488, Al568, biotin Birthdating and marker colocalization 
GABA A2052 Sigma 1:2500 G anti-rb Al546 Colocalization with Lpar1-GFP 

Note: D, donkey; rb, rabbit; m, mouse; gt, goat; Al, Alexa.

Table 2

Summary of the number (n) of brains used for each antibody and injection–survival age combination

Antibody Strain Injection age Brain stage Litter 
BrdU Edg2-WT E10.5 P1 (n= 2) 1 litter (J) 
E13.5 P1 (n= 2) 1 litter (O) 
Edg2-GFP E10.5 P1 (n= 3) 1 litter (Q) 
E13.5 P1 (n= 3) 1 litter (O) 
Cplx3 + BrdU Edg2-WT E11.5 P1 (n= 6) 3 litters (H, M, N) 
P8 (n= 6) 3 litters (M, N, R) 
E12.5 P1 (n= 6) 3 litters (G, L, K) 
P8 (n= 6) 3 litters (K, L, T) 
CTGF + BrdU Edg2-WT E11.5 P1 (n= 4) 2 litters (H, M) 
P8 (n= 4) 3 litters (M, N, R) 
E12.5 P1 (n= 4) 2 litters (G, L) 
P8 (n= 4) 2 litters (K, L) 
GFP + BrdU Edg2-GFP E11.5 P1 (n= 6) 3 litters (H, M, N) 
P8 (n= 6) 3 litters (M, N, R) 
E12.5 P1 (n= 6) 3 litters (G, K, L) 
P8 (n= 6) 3 litters (K, L, T) 
NeuN + BrdU Edg2-WT E11.5 P1 (n= 4) 2 litters (H, M) 
P8 (n= 4) 2 litters (R, N) 
E12.5 P1 (n= 4) 2 litters (G, L) 
P8 (n= 4) 2 litters (L, K) 
Nurr1 + BrdU Edg2-WT E11.5 P1 (n= 6) 3 litters (H, M, N) 
P8 (n= 6) 3 litters (M, N, R) 
E12.5 P1 (n= 6) 3 litters (G, L, K) 
P8 (n= 5) 3 litters (K, L, T) 
Antibody Strain Injection age Brain stage Litter 
BrdU Edg2-WT E10.5 P1 (n= 2) 1 litter (J) 
E13.5 P1 (n= 2) 1 litter (O) 
Edg2-GFP E10.5 P1 (n= 3) 1 litter (Q) 
E13.5 P1 (n= 3) 1 litter (O) 
Cplx3 + BrdU Edg2-WT E11.5 P1 (n= 6) 3 litters (H, M, N) 
P8 (n= 6) 3 litters (M, N, R) 
E12.5 P1 (n= 6) 3 litters (G, L, K) 
P8 (n= 6) 3 litters (K, L, T) 
CTGF + BrdU Edg2-WT E11.5 P1 (n= 4) 2 litters (H, M) 
P8 (n= 4) 3 litters (M, N, R) 
E12.5 P1 (n= 4) 2 litters (G, L) 
P8 (n= 4) 2 litters (K, L) 
GFP + BrdU Edg2-GFP E11.5 P1 (n= 6) 3 litters (H, M, N) 
P8 (n= 6) 3 litters (M, N, R) 
E12.5 P1 (n= 6) 3 litters (G, K, L) 
P8 (n= 6) 3 litters (K, L, T) 
NeuN + BrdU Edg2-WT E11.5 P1 (n= 4) 2 litters (H, M) 
P8 (n= 4) 2 litters (R, N) 
E12.5 P1 (n= 4) 2 litters (G, L) 
P8 (n= 4) 2 litters (L, K) 
Nurr1 + BrdU Edg2-WT E11.5 P1 (n= 6) 3 litters (H, M, N) 
P8 (n= 6) 3 litters (M, N, R) 
E12.5 P1 (n= 6) 3 litters (G, L, K) 
P8 (n= 5) 3 litters (K, L, T) 

Note: The “litter” column indicates how many litters the brains were derived from, and which litters these were, to help identify when littermates were used for different survival (P1 and P8) ages. For each brain used, 4 sections were stained and analyzed for each antibody combination quantified.

Imaging and Analysis

Fluorescent-labeled sections were imaged on a confocal laser scanning microscope (LSM710, Zeiss) in the motor cortex (in future referred to as medial cortex) and at the lateral edge of the neocortex (referred to as lateral cortex; Supplementary Fig. S1) in areas corresponding to putative future somatosensory and auditory cortex. Imaging intensity and filter cut-offs were selected to minimize bleedthrough. All images were subsequently imported in Photoshop CS3 and intensity- and contrast-adjusted.

Determining Marker Colocalization

To determine the proportion of subplate neurons that are marker-positive, NeuN+ cells that were also CTGF+ or Nurr1+ or Cplx3+ (in WT) or GFP+ (in Lpar1-GFP) were determined within a 50 µm wide band in the primary somatosensory cortex of P1 or P8 brains. For analysis of marker colocalization on P8 Lpar1-GFP sections, GFP+ cells and CTGF+ and NeuN+ or CTGF+ and Cplx3+ or Cplx3+ and Nurr1+ cells were quantified.

Determining Localization of Marker Expression

For Cplx3, Nurr1, and Lpar1-GFP, the maximal vertical thickness of the strongly labeled band of cells above the WM was measured in medial and lateral (not Nurr1) cortices at P1 and P8. Measurements were taken from the middle of the cell body/nucleus to avoid artifacts based on the larger volume of labeling for the Cplx3+ or GFP+ cells and their asymmetric cell shape.

Determining BrdU-Labeled Cell Distribution Across Time and Space

For each brain, 3–8 single plane images (square image of 375 µm side length and 5 µm confocal imaging depth) were analyzed for the distribution of BrdU-labeled cells using Cell Profiler (Lamprecht et al. 2007). Cell locations were binned into 50 μm bins across the cortical depth present in the image (bin 1 was fully in WM, bin 2 covered SP, bin 3 was located just above SP, etc.). Additionally, for E13.5 BrdU injections, the mean area and form factor [4π area/(perimeter)2 = 1 for circle] of the labeled nuclei were determined for all fully labeled nuclei located in each of the 50 μm wide bins.

Determining Birthdate of Marker-Labeled Cells

BrdU-labeled cells within a 50 µm wide band above the WM were assessed for colocalization with the subplate marker label (or general marker for neurons, NeuN) and vice versa. Alternatively, for sections stained with CTGF, only BrdU+ nuclei were identified and assessed for colabeling with CTGF.

For all of the above, “fully labeled BrdU nuclei” refer to approximately circular objects (diameter 6.25–12.5 µm) with at least 75% of the structure filled and a relatively smooth outer edge (see arrowheads in Fig. 1).

Figure 1.

Distribution of BrdU+ nuclei in the lateral cortex of the mouse for various injection ages of BrdU. (A–A′′′) Confocal plates of lateral mouse cortex at P1, showing the distribution of BrdU+ nuclei for injections at (A) E10.5, (A′) E11.5, (A′′) E12.5, and (A′′′) E13.5. The location of the subplate layer is demarcated by thin white lines. (B) A graphic representation of the distribution of BrdU+ nuclei in the lateral cortex at P1 (given as mean ± SD for each bin across 4 images for each brain and n= 10 brains each for E11.5 and E12.5 and n= 5 brains each for E10.5 and E13.5). The confocal images were divided into 50 μm horizontal bins such that the border between the lowest and the second-lowest bin was centered on the border of the WM with the subplate. Following an E10.5 injection, very few BrdU+ nuclei could be found anywhere in the cortex, whereas after an E11.5 injection, the majority of BrdU+ nuclei were located within the subplate bin. Following an E12.5 injection, the majority of BrdU+ nuclei were within layer 6, although a considerable number of nuclei were also found in the subplate and some in the WM. By E13.5, the majority of labeled cells were in layers 6 and 5, with some more labeled nuclei also found in the WM. (C) A graphic representation of the distribution of BrdU+ nuclei in the medial cortex for E11.5 or E12.5 injection and survival to either P1 or P8 (mean ± SD for each bin across 4 images for each brain and n= 10 brains for each injection and survival). Gray labeling indicates the approximate anatomical layer equivalent of the bins. (D) A graphic representation of the distribution of BrdU+ nuclei in the lateral cortex for E11.5 or E12.5 injection and survival to either P1 or P8 (mean ± SD for each bin across 4 images for each brain and n= 10 brains for each injection and survival). Gray labeling indicates the approximate anatomical layer equivalent of the bins. Scale bar = 50 μm applies to all images. WM, white matter; SP, subplate; L6, layer 6a; L5, layer 5; CP, cortical plate.

Figure 1.

Distribution of BrdU+ nuclei in the lateral cortex of the mouse for various injection ages of BrdU. (A–A′′′) Confocal plates of lateral mouse cortex at P1, showing the distribution of BrdU+ nuclei for injections at (A) E10.5, (A′) E11.5, (A′′) E12.5, and (A′′′) E13.5. The location of the subplate layer is demarcated by thin white lines. (B) A graphic representation of the distribution of BrdU+ nuclei in the lateral cortex at P1 (given as mean ± SD for each bin across 4 images for each brain and n= 10 brains each for E11.5 and E12.5 and n= 5 brains each for E10.5 and E13.5). The confocal images were divided into 50 μm horizontal bins such that the border between the lowest and the second-lowest bin was centered on the border of the WM with the subplate. Following an E10.5 injection, very few BrdU+ nuclei could be found anywhere in the cortex, whereas after an E11.5 injection, the majority of BrdU+ nuclei were located within the subplate bin. Following an E12.5 injection, the majority of BrdU+ nuclei were within layer 6, although a considerable number of nuclei were also found in the subplate and some in the WM. By E13.5, the majority of labeled cells were in layers 6 and 5, with some more labeled nuclei also found in the WM. (C) A graphic representation of the distribution of BrdU+ nuclei in the medial cortex for E11.5 or E12.5 injection and survival to either P1 or P8 (mean ± SD for each bin across 4 images for each brain and n= 10 brains for each injection and survival). Gray labeling indicates the approximate anatomical layer equivalent of the bins. (D) A graphic representation of the distribution of BrdU+ nuclei in the lateral cortex for E11.5 or E12.5 injection and survival to either P1 or P8 (mean ± SD for each bin across 4 images for each brain and n= 10 brains for each injection and survival). Gray labeling indicates the approximate anatomical layer equivalent of the bins. Scale bar = 50 μm applies to all images. WM, white matter; SP, subplate; L6, layer 6a; L5, layer 5; CP, cortical plate.

Figures for publication were assembled and contrast- and intensity-adjusted using Photoshop CS3. Statistical analysis was performed using the open source software package “R” (http://www.r-project.org/).

Results

For the following, the definition of layer 6b and subplate is used interchangeably and refers to a thin band of cells above the WM, which is separated from layer 6 proper by a narrow, cell sparse region during the first 2 postnatal weeks (Hoerder-Suabedissen et al. 2009). The term “WM cells” refers to cells whose soma is located entirely within the WM. For this study, this is further restricted to the upper 50 μm of the WM, owing to imaging restrictions.

GFP+ Cell Distribution and Characterization in the Lpar1-GFP Mouse Line

GFP expression is driven from a genomically integrated copy of bacterial arteficial chromosome (BAC) RP23-149O20, in which GPF is inserted into the first codon of the lpar1/edg2 gene. Only male animals of the Lpar1-GFP mouse line are GFP+. Based on both GFP expression and GFP-genotyping results, female littermates are always GFP-negative (i.e. WT; n> 250), suggesting that the GFP-containing BAC is inserted into the y-chromosome. GFP is evident in E13.5 brains and labels cells in the ventricular zones and preplate. By E15.5, the GFP+ cells are located both in the marginal zone and in the subplate, but are absent from the cortical plate. Furthermore, GFP+ cells are no longer evident in the ventricular zones. Up until P2, GFP+ cells in the cortical wall remain evident both in the subplate and in the marginal zone, but not in the cortical plate or developing cortical layers 2–6a. By P6, GFP+ cells are no longer visible in the marginal zone, but faint GFP+ cells are additionally visible in cortical layers 2–6a. These non-SP GFP+ cells are much less abundant than those in the subplate (Supplementary Fig. S2). Immunohistochemistry against GABA, performed at P8, revealed that approximately half of the faint non-SP GFP+ cells are GABAergic (49 ± 13%, n= 295 GFP+ cells), but none of the GFP+ cells (n= 373) in the subplate expresses GABA (n= 3 brains, Supplementary Fig. S2).

GFP intensity decreases with age, but native GFP fluorescence remains evident in the subplate and very faintly in some layer 5 cells into adulthood. Outside of the cerebral cortex, GFP+ cells are abundant in the dentate gyrus and hippocampus, as well as in the amygdala, anterior olfactory area, habenula, preoptic nucleus, and bed of stria terminalis but not in any of the primary sensory nuclei of the thalamus. Furthermore, the skull, pia, and choroid plexus are strongly GFP-labeled in embryonic and young postnatal heads.

Timing of Subplate Cell Generation

Subplate cells are classically thought of as “the earliest generated cells of the cerebral cortex,” with neurogenesis starting at E10.5 in mice (Angevine and Sidman 1961; Smart and Smart 1982; Price et al. 1997). Here, we analyzed the distribution of fully BrdU-labeled cells deriving from BrdU injections at E10.5, E11.5, E12.5, or E13.5 in both medial (motor) and lateral (somatosensory or auditory) cortices.

Following an E10.5 injection, very few fully labeled BrdU+ cells could be detected in the cerebral cortex at P1, whereas numerous subcortical structures contained many fully labeled nuclei (see arrowheads in Fig. 1AA’’’ for examples of fully labeled nuclei). In the lateral cortex, only 1 ± 1 cells (uncorrected profile count, mean ± SD on average/image, where each image is a 375 µm by 375 µm wide field of view) were found to be fully labeled following an E10.5 BrdU injection, whereas 11 ± 6 cortical cells were labeled by E11.5 BrdU injections and 50 ± 30 and 45 ± 16 cortical cells were fully labeled by E12.5 and E13.5 BrdU injections, respectively (n= 5 brains for E10.5 or E13.5 and n= 12 brains for E11.5 or E12.5, see Fig. 1AA’’’, numbers given apply to images taken in the lateral cortex).

The distribution of BrdU+ cells labeled at P1 changed for injections made over this embryonic time period (Fig. 1). The peak of the distribution was within the subplate for E10.5- and E11.5-born cells, irrespective of whether the analysis was performed in the medial or lateral cortex. However, only in the medial cortex did the peak of the E12.5 BrdU+ cell distribution also include the subplate. The peak of the BrdU+ cell distribution at E12.5 for the lateral cortex and at E13.5 for lateral and medial cortices was in the overlying cortical layers. Furthermore, at E12.5 and E13.5, a substantial number of BrdU+ nuclei were located within the WM (Fig. 1B). BrdU+ nuclei contributing to this distribution calculation were not assessed for whether they were in neurons, but there was no significant difference in the area occupied or the roundness (form factor) of BrdU+ nuclei when comparing between the WM and subplate. Based on the distribution of BrdU+ nuclei, it was decided to analyze double labeling of BrdU and SP markers in the primary somatosensory cortex at P1 and P8, following the injection of BrdU at E11.5 or E12.5.

Are Previously Published Markers for Subplate Cells an “Anatomical Compartment” Marker or Do They Label Early-Born Neurons?

Cplx3 labels a thin band of cells above the WM in the postnatal rodent cortex (Hoerder-Suabedissen et al. 2009; Wang et al. 2011), with onset of expression between P0 and P2 in a latero-medial gradient. By P8, Cplx3 also faintly labels pyramidal cells in layer 5. The band of labeled subplate cells is approximately 50 µm thick (49 µm medially at P8 and 43 and 59 µm laterally at P1 and P8; n= 2 brains). Cplx3 only very rarely labels cells within the WM. Nurr1, in contrast, labels cells in a thin band of cells above the WM in the medial cortex, whereas it labels cells in all infragranular layers in the lateral cortex (Arimatsu et al. 2003; Hoerder-Suabedissen et al. 2009). Expression levels are higher in the nonsubplate cells. The medial band of Nurr1 expression is narrower than that of Cplx3 (38 and 37 µm at P1 and P8). Nurr1 and Cplx3 expression partially colocalizes in cells located within the thin band above the WM (Fig. 2 and Table 3) (Hoerder-Suabedissen et al. 2009). CTGF also labels a thin band of cells above the WM (Fig. 2) and has been reported to colocalize with Nurr1 in mouse (Watakabe et al. 2007). Very rarely does it label cells outside of the 50 µm thick subplate band. Lpar1-GFP also labels a band of cells above the WM (Supplementary Fig. S2), although this band is slightly wider than that of Cplx3 and Nurr1 (51 and 58 µm medially at P1 and P8 and 56 and 53 µm laterally at P1 and P8; n= 2 brains). Additionally, weaker GFP expression is seen in some cells scattered throughout all cortical layers (discussed earlier), about half of which are GABAergic (Supplementary Fig. S2). However, this nonsubplate expression is not evident at P1. For Cplx3, Nurr1, and Lpar1-GFP, the definition of subplate as a 50 µm thick band of cells excludes <7% of strongly marker-labeled cells in the medial cortex at either P1 or P8.

Table 3

Summary of the percentage of neurons within the subplate that are colabeled with either a single subplate marker or any (technically feasible) combination of 2 subplate markers (i.e. excluding CTGF+Nurr1+ as both antibodies are raised in goat)

Age Colabel single % of neurons Colabel double % of neurons 
P1 Cplx3+ 16 ± 9 (n= 6) Cplx3+CTGF+ Failed 
CTGF+ 38 ± 8 (n= 3) Cplx3+GFP+ 4 ± 3 (n= 3) 
GFP+ 33 ± 9 (n= 8) Cplx3+Nurr1+ 9 ± 9 (n= 3) 
Nurr1+ 34 ± 13 (n= 7) CTGF+GFP+ 23 ± 6 (n= 3) 
  CTGF+Nurr1+ Not possible 
  GFP+Nurr1+ 6 ± 4 (n= 4) 
P8 Cplx3+ 43 ± 9 (n= 6) Cplx3+CTGF+ Failed 
CTGF+ 66 ± 13 (n= 3) Cplx3+Nurr1+ 15 ± 9 (n= 3) 
GFP+ 29 ± 13 (n= 7) Cplx3+GFP+ 22 ± 5 (n= 3) 
Nurr1+ 32 ± 9 (n= 6) CTGF+GFP+ 30 ± 10 (n= 3) 
  CTGF+Nurr1+ Not possible 
  GFP+Nurr1+ 8 ± 5 (n= 3) 
Age Colabel single % of neurons Colabel double % of neurons 
P1 Cplx3+ 16 ± 9 (n= 6) Cplx3+CTGF+ Failed 
CTGF+ 38 ± 8 (n= 3) Cplx3+GFP+ 4 ± 3 (n= 3) 
GFP+ 33 ± 9 (n= 8) Cplx3+Nurr1+ 9 ± 9 (n= 3) 
Nurr1+ 34 ± 13 (n= 7) CTGF+GFP+ 23 ± 6 (n= 3) 
  CTGF+Nurr1+ Not possible 
  GFP+Nurr1+ 6 ± 4 (n= 4) 
P8 Cplx3+ 43 ± 9 (n= 6) Cplx3+CTGF+ Failed 
CTGF+ 66 ± 13 (n= 3) Cplx3+Nurr1+ 15 ± 9 (n= 3) 
GFP+ 29 ± 13 (n= 7) Cplx3+GFP+ 22 ± 5 (n= 3) 
Nurr1+ 32 ± 9 (n= 6) CTGF+GFP+ 30 ± 10 (n= 3) 
  CTGF+Nurr1+ Not possible 
  GFP+Nurr1+ 8 ± 5 (n= 3) 

Note: The numbers in parenthesis indicate the number of brains used.

Figure 2.

Some overlap in the expression of Cplx3, CTGF, Nurr1, and Lpar1-GFP in the P8 mouse somatosensory cortex. (A) Single confocal image showing the overlap among Nurr1 (red), Cplx3 (white), and Lpar1-GFP (green). (B) Single confocal image showing the overlap among Cplx3 (red), CTGF (white), and Lpar1-GFP. (C) Single confocal image demonstrating that most CTGF (red) and Lpar1-GFP (green) cells are NeuN-positive (white) neurons. The location of the subplate layer is demarcated by thin white lines. Insets in (AC) are higher magnification of nearby triple- (arrowhead) and double (arrow)-labeled cells. (D) A schematic representation of the overlapping expression of the 4 molecular markers. Note that the overlap between CTGF and Nurr1 was not directly assessed, and quadruple or quintuple overlap is inferred from the total proportion of triple-labeled cells observed. Scale bar = 50 μm applies to all images.

Figure 2.

Some overlap in the expression of Cplx3, CTGF, Nurr1, and Lpar1-GFP in the P8 mouse somatosensory cortex. (A) Single confocal image showing the overlap among Nurr1 (red), Cplx3 (white), and Lpar1-GFP (green). (B) Single confocal image showing the overlap among Cplx3 (red), CTGF (white), and Lpar1-GFP. (C) Single confocal image demonstrating that most CTGF (red) and Lpar1-GFP (green) cells are NeuN-positive (white) neurons. The location of the subplate layer is demarcated by thin white lines. Insets in (AC) are higher magnification of nearby triple- (arrowhead) and double (arrow)-labeled cells. (D) A schematic representation of the overlapping expression of the 4 molecular markers. Note that the overlap between CTGF and Nurr1 was not directly assessed, and quadruple or quintuple overlap is inferred from the total proportion of triple-labeled cells observed. Scale bar = 50 μm applies to all images.

All 4 subplate markers overlap to varying degrees in their expression, in a pattern that is summarized in Figure 2 and Table 3. At P1, GFP+ or Nurr1+ cells make up roughly one-third each of NeuN+ neurons (33% and 34%, n= 8 or 7 brains, respectively), with relatively little overlap (6% of neurons are double-labeled for GFP and Nurr1 at P1, n= 3 brains). Neither proportion of neurons labeled (n= 7 or 6 brains) nor relative overlap change much by P8 (n= 3 brains). In contrast, Cplx3+ labels approximately 15% of neurons at P1 (n= 6 brains), but nearly half of neurons (43%, n= 6 brains) by P8. An increase in the number of labeled neurons is also seen for CTGF, which labels 38% of neurons at P1 (n= 3 brains) but nearly two-thirds (66%, n= 3 brains) of neurons at P8. CTGF, therefore, labels the largest group of subplate neurons at either age, but this is more pronounced at P8.

The proportion of neurons double-labeled with either Cplx3 or CTGF and Nurr1 or Lpar1-GFP also changes in this time frame. At either P1 or P8, almost all Lpar1-GFP+ cells are also CTGF+ (23% and 30% of neurons at P1 and P8, respectively), but only rarely are Lpar1-GFP+ cells Nurr1+ (6% and 8% of neurons at P1 and P8, respectively). The percentage of double labeling changes for Cplx3+GFP+ neurons between P1 and P8. At P1, 4% of neurons are GFP+Cplx3+, but this increases to 22% by P8. In terms of the GFP+ population of neurons, this means that Cplx3 coexpression increases from about 16% of GFP+ neurons to 87% of GFP+ neurons between P1 and P8. A similar, although less dramatic, increase in colabeling with Cplx3 also occurs in the population of Nurr1+ neurons. Nearly twice as many Nurr1+ neurons are Cplx3+ at P8 than at P1 (49% vs. 27% at P8 and P1, respectively), which corresponds to an increase in the overall population of neurons that are double-labeled with Cplx3+Nurr1+ (15% and 9% at P8 and P1, respectively). This suggests that Cplx3 expression specifically increases in those cells already expressing Lpar1-GFP or Nurr1 and is consistent with the overall time course of expression onset for Cplx3.

Furthermore, at P8, an attempt was made to quantify the relative overlap between any 2 subplate markers and Lpar1-GFP. Almost all GFP+ cells that are Cplx3+ are also CTGF+, but only rarely Nurr1+ (CTGF+GFP+Cplx3+/GFP+Cplx3+ = 88 ± 10% vs. Nurr1+GFP+Cplx3+/GFP+Cplx3+ = 15 ± 15%, mean ± SD, n= 3 brains). See Figure 2D for an accurate schematic representation of each of the various single-, double-, or triple-labeled cell groups at P8.

Thus, given the diverse distribution and onset of expression of these 4 subplate markers, we asked whether they are all expressed in “early-born” neurons of the cerebral cortex. The peak of neurogenesis for cells located within the subplate is on E11.5 and E12.5 (plug = E0.5) in this study, and therefore colocalization analysis was restricted to these ages. Cplx3, CTGF, Nurr1, and Lpar1-GFP are all expressed in some cells labeled with BrdU injections at either E11.5 or E12.5 (Fig. 3).

Figure 3.

Each of the 4 subplate markers, Cplx3, CTGF, Nurr1, and Lpar1-GFP, label some early-born cells in the P1 mouse cortex. (A–A′′′) Single confocal planes showing immunohistochemical labeling for BrdU (injected at E11.5), their matched planes showing immunohistochemical labeling of Cplx3, CTGF, Nurr1, or Lpar1-GFP (B–B′′′), as well as the merger of the 2 images (C–C′′′). Scale bar = 25 μm applies to all images. Numerical summary of the extent of colocalization following E11.5 or E12.5 injection is given in Figure 4.

Figure 3.

Each of the 4 subplate markers, Cplx3, CTGF, Nurr1, and Lpar1-GFP, label some early-born cells in the P1 mouse cortex. (A–A′′′) Single confocal planes showing immunohistochemical labeling for BrdU (injected at E11.5), their matched planes showing immunohistochemical labeling of Cplx3, CTGF, Nurr1, or Lpar1-GFP (B–B′′′), as well as the merger of the 2 images (C–C′′′). Scale bar = 25 μm applies to all images. Numerical summary of the extent of colocalization following E11.5 or E12.5 injection is given in Figure 4.

Differences in the Timing of Birth of Various Subplate Cell Populations

The percentage of BrdU+ cells within the subplate that were also positive for Lpar1-GFP or Nurr1 or Cplx3 was determined for BrdU injections at E11.5 and E12.5 (Fig. 4). Overall, Lpar1-GFP cells appear to be preferentially generated on E11.5 (GFP+BrdU+/BrdU+ is 27 ± 20% for E11.5 and 4 ± 6% for E12.5, lateral and medial data are pooled; P< 0.001, pairwise t-test, Holm's multiple testing correction). No statistically significant timing effect could be detected for Cplx3 and Nurr1 (Cplx3+BrdU+/BrdU+ is 11 ± 11% for E11.5 and 5 ± 9% for E12.5 and Nurr1+BrdU+/BrdU+ is 12 ± 16% for E11.5 and 7 ± 8% for E12.5), although a trend toward preferential generation on E11.5 is also evident for Cplx3+ cells, consistent with its frequent coexpression in Lpar1-GFP cells. This conclusion is based on data from littermates at P1 [of the same litter, female pups (WT) were used for Cplx3 and Nurr1 and male pups (Lpar1-GFP) were used for GFP]. A similar calculation was not possible for CTGF, as it was not feasible to determine exactly how many cells were labeled with CTGF at P1 because of the very restricted area of CTGF labeling in each cell.

Figure 4.

Quantification of colocalization of NeuN+ neurons or subplate marker-labeled cells and BrdU. (A) A bar graph (mean ± SD) giving the percentage of NeuN+ neurons that colocalize with BrdU+ nuclei for injections at E11.5 and E12.5 and survival until either P1 or P8. Note that the percentage of BrdU-labeled neurons decreases significantly between P1 and P8, following an E11.5 injection of BrdU. (BD) Quantification of colocalization of subplate marker-labeled cells with BrdU as a bar graph (mean ± SD). (C) Note that significantly more Lpar1-GFP cells are born on E11.5 than on E12.5. (E) The percentage of BrdU-labeled nuclei that are also NeuN+ neurons as a bar graph (mean ± SD). Virtually all BrdU+ nuclei at P8 are neuronal, whereas a small proportion of BrdU+ nuclei at P1 were not colabeled by NeuN. (FH) Quantification of colocalization of the BrdU-labeled cells with any of the 3 subplate markers as a bar graph (mean ± SD). For Nurr1 (G) and Cplx3 (H), significantly more BrdU+ cells are colabeled by either of the subplate markers at P8 than at P1, irrespective of injection age. A similar result is seen for Lpar1-GFP, following E11.5 BrdU injections. This may suggest a preferential survival of marker expressing, early-born cells compared with nonmarker expressing early-born cells. *P< 0.05 and ***P< 0.001 (pairwise t-test using Holm's multiple testing correction).

Figure 4.

Quantification of colocalization of NeuN+ neurons or subplate marker-labeled cells and BrdU. (A) A bar graph (mean ± SD) giving the percentage of NeuN+ neurons that colocalize with BrdU+ nuclei for injections at E11.5 and E12.5 and survival until either P1 or P8. Note that the percentage of BrdU-labeled neurons decreases significantly between P1 and P8, following an E11.5 injection of BrdU. (BD) Quantification of colocalization of subplate marker-labeled cells with BrdU as a bar graph (mean ± SD). (C) Note that significantly more Lpar1-GFP cells are born on E11.5 than on E12.5. (E) The percentage of BrdU-labeled nuclei that are also NeuN+ neurons as a bar graph (mean ± SD). Virtually all BrdU+ nuclei at P8 are neuronal, whereas a small proportion of BrdU+ nuclei at P1 were not colabeled by NeuN. (FH) Quantification of colocalization of the BrdU-labeled cells with any of the 3 subplate markers as a bar graph (mean ± SD). For Nurr1 (G) and Cplx3 (H), significantly more BrdU+ cells are colabeled by either of the subplate markers at P8 than at P1, irrespective of injection age. A similar result is seen for Lpar1-GFP, following E11.5 BrdU injections. This may suggest a preferential survival of marker expressing, early-born cells compared with nonmarker expressing early-born cells. *P< 0.05 and ***P< 0.001 (pairwise t-test using Holm's multiple testing correction).

Postnatal Survival of Early-Born Cells of the Cortex

As previously reported by Price et al. (1997), the majority of cells labeled with BrdU between E10.5 and E12.5 can no longer be detected in the cerebral cortex of mice by P8. Here, we can broadly confirm that the peak of labeled cell distribution remains within the subplate in the medial cortex, but the overall number of fully labeled BrdU nuclei is much smaller at P8 than at P1 (Fig. 1). Moreover, significantly fewer NeuN+ neurons in the subplate layer are BrdU+ at P8 compared with P1 (3% vs. 8% for E11.5 injections and 8% vs. 11% for E12.5 injections, P< 0.005 for both, pairwise t-test with Holm's multiple testing correction; Fig. 4A), ruling out a purely “dilution” effect because of brain expansion during this time frame.

In contrast, there is no significant difference in the proportion of marker-labeled cells that are also BrdU+ between P1 and P8 (assessed only for Cplx3, Nurr1, or Lpar1-GFP; Fig. 4BD).

During this time frame, the average area occupied by each fully labeled BrdU+ nucleus increases by a factor of 1.4 (BrdU injection at E12.5, averaged across all images for litter L, n= 1193 and 364 cells for P1 and P8, respectively). The area occupied by Nurr1, a nuclear marker, also increases between P1 and P8 (not quantified), so this probably corresponds to an overall increase in cell volume during this 1-week period.

Enhanced Postnatal Survival of the Selected Subplate Marker Expressing Cells

In contrast to the above finding of a decrease in the number of early-born cortical neurons during the first postnatal week, the subplate layer remains visible well past the first postnatal week, and all 4 subplate markers, CTGF, Cplx3, Nurr1, and Lpar1-GFP (identified in a postnatal screen of subplate markers), continue to label neurons within the narrow band of cells between the WM and layer 6a at P8 (Fig. 2) and even in the adult (Hoerder-Suabedissen et al. 2009). Overall, the percentage of subplate marker-labeled BrdU+ nuclei within the subplate is higher at P8 than at P1 (Fig. 4DF), although this enhanced survival of subplate marker expressing BrdU+ cells differs between the 3 subplate markers studied here. For Cplx3 and Nurr1, significantly more BrdU+ cells are marker-positive at P8 than at P1 for either injection time point (P< 0.05), but the difference is only significant for E11.5 injections when comparing double labeling with Lpar1-GFP (P< 0.001 vs. P= 0.27, pairwise t-test using Holm's multiple correction testing), possibly a consequence of the preferential E11.5 birth of Lpar1-GFP neurons. The effect was most pronounced for Cplx3+ cells. On average, more than 4 times as many BrdU+ cells expressed Cplx3 at P8 than at P1 (for an E11.5 injection, Cplx3+BrdU+/BrdU+ is 11 ± 11% at P1 and 46 ± 42% at P8), possibly partly a consequence of the temporally coincident increase in neurons labeled with Cplx3 between P1 and P8. Similarly, on average, 2–3 times as many BrdU+ cells expressed Lpar1-GFP or Nurr1 at P8 compared with P1 (for an E11.5 injection, GFP+BrdU+/BrdU+ is 27 ± 20% at P1 and 64 ± 43% at P8 and similarly Nurr1+BrdU+/BrdU+ is 12 ± 16% at P1 and 37 ± 38% at P8), despite the average number of neurons labeled with either marker staying steady in the same time frame (GFP+NeuN+/NeuN+ is 33 ± 9% and 29 ± 8% at P1 and P8, respectively, and similarly Nurr1+NeuN+/NeuN+ is 34 ± 13% and 32 ± 9%; Table 3). All quantification is based on double immunohistochemistry so it is not possible to state whether the marker-positive surviving subplate cells are single-, double-, or triple-labeled for the various subplate markers, nor is it possible to state that all marker-labeled BrdU+ cells are neurons, although this is very likely as no BrdU+NeuN− cells were found at P8 (n= 188 BrdU+ cells within the subplate in 8 brains).

Discussion

Timing of Subplate Cell Generation

In this mouse strain (NIHS), the peak of neurogenesis for cells located in the subplate compartment is at E11.5 and E12.5, with few cells generated on E10.5 or E13.5 when analysis is performed at P1. However, more cells within the subplate compartment were generated at E13.5 than at E10.5. Additionally, cells born at E13.5 contributed to both the overlying cortex and presumed neurons within the WM in the P1 cortex. Similar numbers of WM neurons were generated at E12.5 and E13.5, but only very few were generated earlier than that. We did not test whether all WM BrdU+ cells are neurons, but their size and nuclear shape are not different from those in subplate or lower layer 6, suggesting a neuronal identity. Furthermore, no BrdU+ nuclei that were NeuN− were observed in the WM following E11.5 and E12.5 injections on those sections stained for NeuN.

Continuous addition of neurons to the subplate layer has been observed in primates (Smart et al. 2002; Bystron et al. 2008) and rats (Rickmann et al. 1977; Raedler and Raedler 1978), but not in mice (Smart et al. 2002). However, a further broadening of the subplate layer after the formation of the cortical plate has been noted in mice (Smart et al. 2002) as well as in primates and humans (Kostovic and Rakic 1990; Smart et al. 2002; Bystron et al. 2008). Our study was centered on the thin cortical layer 6b as subplate. This is indeed the site of the expression of the selected genes of this study. However, the term subplate is often used more liberally. In particular, it may be important to reach a consensus whether the later born, presumed neuronal cells of the WM should be classed as subplate cells (or deep subplate cells) or continue to be regarded as a separate group of cells. Notably, the molecular markers used in this study do not label many cells in the WM, although occasionally labeled neurons both above and below the dense band of labeling are observed. But, in contrast, the human subplate is subcompartmentalized with regard to molecular markers as well (Wang et al. 2010).

Subplate Markers Label Early-Born Neurons

Here we demonstrate that all 4 subplate markers (Cplx3, CTGF, Nurr1, and Lpar1-GFP) do not just label cells within the anatomical compartment of the subplate, but genuinely label early-born cells commonly considered to be subplate cells. The 4 subplate markers also occasionally label neurons above or below the subplate layer, and for such cells, we did observe some colabeling with BrdU from either E11.5 or E12.5 injections as well, although this was not quantified, suggesting that these are still “subplate cells” despite being located outside of the subplate layer as defined here.

Postnatal Survival of Early-Born Cells of the Cortex

Despite the confirmed decline of birthdated subplate neurons (Price et al. 1997), a considerable number of early-born cells survive past the first postnatal week. In this context, it is relevant to note that the volume of the nucleus increases from P1 to P8, which may lead to a decrease in the density of BrdU+ labeling, which could confound the classification as a fully versus partially labeled nucleus, as done in this study. The images here were acquired as 5 μm thick optical sections as it was noted in a preliminary study that thinner optical sections at higher magnification never resulted in “fully” labeled nuclei, suggesting a nonuniform distribution of BrdU in the nucleus and hence possibly the appearance of being “not fully labeled” if the BrdU is distributed in a larger volume at P8, giving rise to an underestimate of the number of BrdU-labeled cells at P8. In contrast, the increase in volume would correspond to an increase in cell counts per image, as the optical dissector method was not used. Additionally, the extra survival by 1 week may further reduce the amount of BrdU incorporated into the DNA of each cell as DNA repair may remove some BrdU nucleotides (Cooper-Kuhn and Kuhn 2002; Sauerzweig et al. 2009). Lastly, the volume of the brain expands during the first postnatal week, but this can be neglected, as the percentage of neurons that were BrdU-labeled also decreased in this time window, and one would expect an equal decrease in cell density for BrdU+ and NeuN+ cells within the subplate unless there is an as-yet unidentified group of later born neurons that only settle in the subplate region after P1.

Enhanced “Survival” of Subplate Neurons with Selected Marker Expression

Overall, BrdU+ cells expressing any 1 of the 3 subplate markers assessed here (Lpar1-GFP, Cplx3, and Nurr1) preferentially survive during the first postnatal week. In this context, it is highly relevant that all 3 genes (as well as CTGF) were identified in a screen of subplate gene expression using P8 tissue (Hoerder-Suabedissen et al. 2009). Our comparisons between the embryonic and adult gene expression pattern (Belgard et al. 2011; Oeschger et al. 2011) demonstrated that there are relatively few genes that are continuously expressed in subplate until adulthood (Hoerder-Suabedissen, Oeschger et al. in preparation). It would be interesting to compare whether postnatally expressed genes identified in an equivalent screen using E15 tissue (Oeschger et al. 2011) preferentially label those subplate neurons that do not survive the first postnatal week.

The enhanced survival of Cplx3+ cells compared with either Nurr1+ or Lpar1-GFP+ cells may reflect the later onset of Cplx3 expression (more neurons are Cplx3+ at P8 than at P1), rather than a genuine survival effect. Without a causal link between gene expression and survival, this cannot be untangled. On the contrary, Nurr1 and Lpar1-GFP start to be expressed in embryonic cortical cells, and the proportion of neurons labeled with either marker does not change during the first postnatal week. The proportion of BrdU+-labeled cells out of total neurons does decrease in the same time frame, strongly suggesting that neurons expressing either Nurr1 or Lpar1-GFP (or both) preferentially survive the first postnatal week, whereas those neurons not expressing at least 1 of the subplate markers are more likely to disappear in the same time frame.

Technical Considerations

BrdU is a commonly used thymidine analog that integrates into DNA during any process of DNA synthesis (S-phase as well as DNA repair). As this study was designed to identify neurons undergoing their final round of DNA synthesis and cell division at a particular embryonic time point but analyzed much later, cell cycle markers could not be utilized. BrdU is the only commonly used DNA analog that can be combined with immunohistochemistry (Taupin 2007), an essential criterion given that the aim of the study was colocalization with molecular markers of a subgroup of mature neurons. For the same reason, it was essential to obtain sufficient BrdU labeling to be able to distinguish fully labeled nuclei from those that were partially labeled and presumed to have undergone one or more further cell divisions. As BrdU is mutagenic, teratogenic, and may influence neuronal cell division as well as cell migration (Taupin 2007; Duque and Rakic 2011), it is advisable to use as low a dose as possible. Here we trialled i.p. injections of 50 and 100 mg/kg maternal body weight, but the lower dose resulted in insufficient labeling of BrdU nuclei. With this systemically administered dosage, the BrdU concentration in the embryonic brain is likely to be high enough for good DNA labeling for not more than 2 h (Taupin 2007); therefore, the cells labeled here following a single dose of BrdU will probably be <10% of the cells undergoing cell division in any of the 24 h periods investigated. This is consistent with the low percentage of neurons labeled with a single BrdU injection observed here.

The BrdU antigen retrieval protocol used here is slightly different compared with the more commonly used ones (Wojtowicz and Kee 2006). Wojtowicz and Kee (2006) suggest the use of 1 M HCl for 30 min at 45°C; while this works well for colabeling with either Nurr1 or GFP, it is unsuitable for colabeling with Cplx3 (data not shown). Thus, the protocol was modified to 1 M HCl for 1 h at 38.5°C, which is compatible with all 5 colabeling antibodies used in the present study (Cplx3, CTGF, GFP, NeuN, and Nurr1).

BrdU incorporated into double-stranded DNA may be removed as part of normally occurring DNA repair processes (Taupin 2007); therefore over time, the density of BrdU in an initially fully labeled nucleus is expected to decrease. This may have contributed to the less-dense BrdU label in the P8 nuclei because of the extra week of survival and DNA repair, although BrdU label has been detected in neurons in the mouse cerebral cortex more than 60 days after an embryonic BrdU injection (Miller and Nowakowski 1988).

The gender split between WT (female) and GFP (male) is unavoidable when using littermates, as only males are GFP+ in the Lpar1-GFP line. The alternative would be to use WT males from separate litters, but that increases the risk of slightly different times of BrdU injection as the mating window was 12 h long and would rely on the assumption that the offspring of an Lpar1-GFP×NIHS cross are otherwise identical to NIHS×NIHS.

The images for analysis were acquired using consecutive scans on a laser scanning confocal microscope. The settings were carefully selected to avoid bleedthrough of the very bright GFP-Alexa488 label onto the cy3 channel used for BrdU. Additionally, the filtering settings within Cell Profiler were chosen such that what appeared to be very faintly labeled nuclei were rejected as they could either represent scattered light from a BrdU nucleus outside the focal plane or be derived from bleedthrough of the Alexa488 channel. Thus, every attempt has been made to minimize the risk of falsely identifying BrdU+ cells.

Scanning across a square field of view of 425 μm side length and 5 μm optical section thickness was selected to optimize for fully labeled BrdU nuclei. Scanning at higher magnification and thinner optical sections resulted almost exclusively in partially labeled nuclei that made it impossible to reliably distinguish between fully labeled and <50% labeled nuclei. This, however, may result in an underestimation of BrdU+ nuclei if they are densely packed. Cell Profiler settings were optimized for identifying circular nuclei above a certain size (6.25 µm diameter) and rejecting elongated shapes deriving from either blood vessels or possibly BrdU+-labeled glial nuclei. But to avoid falsely accepting blood vessels as BrdU+ nuclei, some genuine BrdU+ nuclei in extremely close apposition were falsely rejected (see an example at the left edge of bin 6 in Supplementary Fig. S1B). This false rejection rate should be consistent and may lead to an underestimate of the number of BrdU+ cells in the densely packed cortices of E12.5- and E13.5-injected animals (and in particular possibly underestimate the total number of BrdU+ nuclei within the peak of each BrdU+ nuclei distribution). However, this does not impact on the main conclusion as cell profiler was only used to determine the distribution of BrdU+ nuclei following different injection time points.

It was noted that the area occupied by BrdU- or Nurr1-labeled nuclei increased between P1 and P8. This is likely to lead to a less-dense BrdU label and possibly more nuclei that are classified as “partially labeled” because the BrdU is nonuniformly distributed. One could test whether more nuclei are classified as fully BrdU-labeled if 150–200 mg/kg BrdU was given; however, this will also increase the likelihood of brain malformations and neurotoxic effects, and we therefore decided to consistently use the 100 mg/kg dose of BrdU irrespective of postnatal survival.

Conclusion

The peak of neurogenesis within the subplate layer is on E11.5 and continues into E12.5 in medial locations. Very few cortical cells are born on E10.5, but cell birth on E13.5 still contributes some cells to the subplate layer. Moreover, cells generated on E12.5 and E13.5 contribute more to the underlying WM population than to the subplate.

All 4 molecular markers of the postnatal subplate used here (Cplx3, CTGF, Nurr1, and Lpar1-GFP) are present in some early-born subplate cells. In particular, Lpar1-GFP-positive cells are preferentially born on E11.5. The proportion of marker-labeled BrdU+ nuclei increases between P1 and P8, suggesting an enhanced survival of Nurr1 and Lpar1-GFP marker-labeled cells during the first postnatal week and possibly also for Cplx3 marker-labeled cells.

Lastly, Cplx3, CTGF, Nurr1, and Lpar1-GFP each label partially overlapping groups of neurons within the subplate layer. While both Nurr1 and Lpar1-GFP are expressed from embryonic stages and label a fixed proportion of neurons at the beginning and end of the first postnatal week, CTGF and Cplx3 protein levels only become detectable around the time of birth and both markers label a larger proportion of neurons at P8 compared with P1. Consequently, there is a temporal dynamic to the degree of coexpression between these 4 subplate neuron markers.

Supplementary Material

Supplementary material can be found at:http://www.cercor.oxfordjournals.org/.

Funding

This study was funded by the Medical Research Council (G00900901 and G0700377 to Z.M.).

Notes

The authors wish to thank Dr Helen Stolp and Professor Ray Guillery for their comments on the manuscript. Conflict of Interest: None declared.

References

Angevine
JB
Jr
Sidman
RL
Autoradiographic study of cell migration during histogenesis of cerebral cortex in the mouse
Nature
 , 
1961
, vol. 
192
 (pg. 
766
-
768
)
Arimatsu
Y
Ishida
M
Kaneko
T
Ichinose
S
Omori
A
Organization and development of corticocortical associative neurons expressing the orphan nuclear receptor Nurr1
J Comp Neurol
 , 
2003
, vol. 
466
 
2
(pg. 
180
-
196
)
Bayer
SA
Altman
J
Development of layer I and the subplate in the rat neocortex
Exp Neurol
 , 
1990
, vol. 
107
 
1
(pg. 
48
-
62
)
Belgard
TG
Marques
AC
Oliver
PL
Abaan
HO
Sirey
TM
Hoerder-Suabedissen
A
Garcia-Moreno
F
Molnar
Z
Margulies
EH
Ponting
CP
A transcriptomic atlas of mouse neocortical layers
Neuron
 , 
2011
, vol. 
71
 
4
(pg. 
605
-
616
)
Bystron
I
Blakemore
C
Rakic
P
Development of the human cerebral cortex: Boulder Committee revisited
Nat Rev Neurosci
 , 
2008
, vol. 
9
 
2
(pg. 
110
-
122
)
Cooper-Kuhn
CM
Kuhn
HG
Is it all DNA repair? Methodological considerations for detecting neurogenesis in the adult brain
Brain Res Dev Brain Res
 , 
2002
, vol. 
134
 
1–2
(pg. 
13
-
21
)
Dupont
E
Hanganu
IL
Kilb
W
Hirsch
S
Luhmann
HJ
Rapid developmental switch in the mechanisms driving early cortical columnar networks
Nature
 , 
2006
, vol. 
439
 
7072
(pg. 
79
-
83
)
Duque
A
Rakic
P
Different effects of bromodeoxyuridine and [3H]thymidine incorporation into DNA on cell proliferation, position, and fate
J Neurosci
 , 
2011
, vol. 
31
 
42
(pg. 
15205
-
15217
)
Friedlander
MJ
Torres-Reveron
J
The changing roles of neurons in the cortical subplate
Front Neuroanat
 , 
2009
, vol. 
3
 pg. 
15
 
Ghosh
A
Antonini
A
McConnell
SK
Shatz
CJ
Requirement for subplate neurons in the formation of thalamocortical connections
Nature
 , 
1990
, vol. 
347
 
6289
(pg. 
179
-
181
)
Grant
E
Hoerder-Suabedissen
A
Molnár
Z
Front Neurosci
 , 
2012
, vol. 
6
 pg. 
53
  
Development of the corticothalamic projections
Heuer
H
Christ
S
Friedrichsen
S
Brauer
D
Winckler
M
Bauer
K
Raivich
G
Connective tissue growth factor: a novel marker of layer VII neurons in the rat cerebral cortex
Neuroscience
 , 
2003
, vol. 
119
 
1
(pg. 
43
-
52
)
Hoerder
A
Mouse cortical subplate neurones: molecular markers, connectivity and development
 , 
2007
Oxford
University of Oxford
Hoerder-Suabedissen
A
Molnár
Z
The morphology of mouse subplate cells with identified projection targets changes with age
J Comp Neurol
 , 
2012
, vol. 
520
 
1
(pg. 
174
-
185
)
Hoerder-Suabedissen
A
Wang
WZ
Lee
S
Davies
KE
Goffinet
AM
Rakic
S
Parnavelas
J
Reim
K
Nicolic
M
Paulsen
O
, et al.  . 
Novel markers reveal subpopulations of subplate neurons in the murine cerebral cortex
Cereb Cortex
 , 
2009
, vol. 
19
 
8
(pg. 
1738
-
1750
)
Judas
M
Sedmak
G
Pletikos
M
Early history of subplate and interstitial neurons: from Theodor Meynert (1867) to the discovery of the subplate zone (1974)
J Anat
 , 
2010
, vol. 
217
 
4
(pg. 
344
-
367
)
Kanold
PO
Kara
P
Reid
RC
Shatz
CJ
Role of subplate neurons in functional maturation of visual cortical columns
Science
 , 
2003
, vol. 
301
 
5632
(pg. 
521
-
525
)
Kanold
PO
Luhmann
HJ
The subplate and early cortical circuits
Annu Rev Neurosci
 , 
2010
, vol. 
33
 (pg. 
23
-
48
)
Kanold
PO
Shatz
CJ
Subplate neurons regulate maturation of cortical inhibition and outcome of ocular dominance plasticity
Neuron
 , 
2006
, vol. 
51
 (pg. 
627
-
638
)
Kostovic
I
Judas
M
Sedmak
G
Developmental history of the subplate zone, subplate neurons and interstitial white matter neurons: relevance for schizophrenia
Int J Dev Neurosci
 , 
2011
, vol. 
29
 
3
(pg. 
193
-
205
)
Kostovic
I
Rakic
P
Developmental history of the transient subplate zone in the visual and somatosensory cortex of the macaque monkey and human brain
J Comp Neurol
 , 
1990
, vol. 
297
 
3
(pg. 
441
-
470
)
Lamprecht
MR
Sabatini
DM
Carpenter
AE
CellProfiler: free, versatile software for automated biological image analysis
Biotechniques
 , 
2007
, vol. 
42
 
1
(pg. 
71
-
75
)
Luskin
M
Shatz
CJ
Neurogenesis of the cat's primary visual cortex
J Comp Neurol
 , 
1985a
, vol. 
242
 
4
(pg. 
611
-
631
)
Luskin
M
Shatz
CJ
Studies of the earliest-generated cells of the cat's visual cortex: cogeneration of subplate and marginal zones
J Neurosci
 , 
1985b
, vol. 
5
 
4
(pg. 
1062
-
1075
)
McConnell
SK
Ghosh
A
Shatz
CJ
Subplate neurons pioneer the first axon pathway from the cerebral cortex
Science
 , 
1989
, vol. 
245
 
4921
(pg. 
978
-
982
)
Miller
MW
Nowakowski
RS
Use of bromodeoxyuridine-immunohistochemistry to examine the proliferation, migration and time of origin of cells in the central nervous system
Brain Res
 , 
1988
, vol. 
457
 
1
(pg. 
44
-
52
)
Oeschger
FM
Wang
WZ
Lee
S
Garcia-Moreno
F
Goffinet
AM
Arbones
ML
Rakic
S
Molnar
Z
Gene expression analysis of the embryonic subplate
Cereb Cortex.
 , 
2011
 
.
Price
DJ
Aslam
S
Tasker
L
Gillies
K
Fates of the earliest generated cells in the developing murine neocortex
J Comp Neurol
 , 
1997
, vol. 
377
 
3
(pg. 
414
-
422
)
Raedler
E
Raedler
A
Autoradiographic study of early neurogenesis in rat neocortex
Anat Embryol
 , 
1978
, vol. 
154
 
3
(pg. 
267
-
284
)
Rickmann
M
Chronwall
BM
Wolff
JR
On the development of non-pyramidal neurons and axons outside the cortical plate: the early marginal zone as a pallial anlage
Anat Embryol
 , 
1977
, vol. 
151
 
3
(pg. 
285
-
307
)
Robertson
RT
Do subplate neurons comprise a transient population of cells in developing neocortex of rats?
J Comp Neurol
 , 
2000
, vol. 
426
 
4
(pg. 
632
-
650
)
Sauerzweig
S
Baldauf
K
Braun
H
Reymann
KG
Time-dependent segmentation of BrdU-signal leads to late detection problems in studies using BrdU as cell label or proliferation marker
J Neurosci Methods
 , 
2009
, vol. 
177
 
1
(pg. 
149
-
159
)
Smart
IH
Dehay
C
Giroud
P
Berland
M
Kennedy
H
Unique morphological features of the proliferative zones and postmitotic compartments of the neural epithelium giving rise to striate and extrastriate cortex in the monkey
Cereb Cortex
 , 
2002
, vol. 
12
 
1
(pg. 
37
-
53
)
Smart
IH
Smart
M
Growth patterns in the lateral wall of the mouse telencephalon: I. Autoradiographic studies of the histogenesis of the isocortex and adjacent areas
J Anat
 , 
1982
, vol. 
134
 (pg. 
273
-
298
)
Suarez-Sola
ML
Gonzalez-Delgado
FJ
Pueyo-Morlans
M
Medina-Bolivar
OC
Hernandez-Acosta
NC
Gonzalez-Gomez
M
Meyer
G
Neurons in the white matter of the adult human neocortex
Front Neuroanat
 , 
2009
, vol. 
3
 pg. 
7
 
Taupin
P
BrdU immunohistochemistry for studying adult neurogenesis: paradigms, pitfalls, limitations, and validation
Brain Res Rev
 , 
2007
, vol. 
53
 
1
(pg. 
198
-
214
)
Valverde
F
López-Mascaraque
L
Santacana
M
DeCarlos
J
Persistence of early-generated neurons in the rodent subplate: assessment of cell death in neocortex during the early postnatal period
J Neurosci
 , 
1995
, vol. 
15
 (pg. 
5014
-
5024
)
Viswanathan
S
Bandyopadhyay
S
Kao
JP
Kanold
PO
Changing microcircuits in the subplate of the developing cortex
J Neurosci
 , 
2012
, vol. 
32
 
5
(pg. 
1589
-
1601
)
Wang
WZ
Hoerder-Suabedissen
A
Oeschger
FM
Bayatti
N
Ip
BK
Lindsay
S
Supramaniam
V
Srinivasan
L
Rutherford
M
Mollgard
K
, et al.  . 
Subplate in the developing cortex of mouse and human
J Anat
 , 
2010
, vol. 
217
 
4
(pg. 
368
-
380
)
Wang
WZ
Oeschger
FM
Montiel
JF
Garcia-Moreno
F
Hoerder-Suabedissen
A
Krubitzer
L
Ek
CJ
Saunders
NR
Reim
K
Villalon
A
, et al.  . 
Comparative aspects of subplate zone studied with gene expression in sauropsids and mammals
Cereb Cortex
 , 
2011
, vol. 
21
 
10
(pg. 
2187
-
2203
)
Watakabe
A
Ichinohe
N
Ohsawa
S
Hashikawa
T
Komatsu
Y
Rockland
KS
Yamamori
T
Comparative analysis of layer-specific genes in mammalian neocortex
Cereb Cortex
 , 
2007
, vol. 
17
 
8
(pg. 
1918
-
1930
)
Wojtowicz
MJ
Kee
N
BrdU assay for neurogenesis in rodents
Nat Protoc
 , 
2006
, vol. 
1
 (pg. 
1399
-
1405
)
Yang
JW
Hanganu-Opatz
IL
Sun
JJ
Luhmann
HJ
Three patterns of oscillatory activity differentially synchronize developing neocortical networks in vivo
J Neurosci
 , 
2009
, vol. 
29
 
28
(pg. 
9011
-
9025
)