Abstract

Neocortical projection neurons arise from a pseudostratified ventricular epithelium (PVE) from embryonic day 11 (E11) to E17 in mice. The sequence of neuron origin is systematically related to mechanisms that specify neuronal class properties including laminar fate destination. Thus, the neurons to be assembled into the deeper layers are the earliest generated, while those to be assembled into superficial layers are the later generated neurons. The sequence of neuron origin also correlates with the probability of cell cycle exit (Q) and the duration of G1-phase of the cell cycle (TG1) in the PVE. Both Q and TG1 increase as neuronogenesis proceeds. We test the hypothesis that mechanisms regulating specification of neuronal laminar destination, Q and TG1 are coordinately regulated. We find that overexpression of p27Kip1 in the PVE from E12 to E14 increases Q but not TG1 and that the increased Q is associated with a commensurate increase in the proportion of exiting cells that is directed to superficial layers. We conclude that mechanisms that govern specification of neocortical neuronal laminar destination are coordinately regulated with mechanisms that regulate Q and are independent of mechanisms regulatory to cell cycle duration. Moreover, they operate prior to postproliferative mechanisms necessary to neocortical laminar assembly.

Introduction

In the mouse, neocortical projection neurons are produced in the pseudostratified ventricular epithelium (PVE) over a period of 7 days, from embryonic day (E) 11 to E17, spanning 11 integer cell cycles (His, 1904; Sauer, 1935, 1936; Boulder Committee, 1970; Takahashi et al., 1994, 1995). On each successive day of neuronogenesis there is a systematic progression in the laminar destination of neocortical neurons produced (Fig. 1A). Specifically, the earliest born neurons are destined for the subplate and layer VI, while the later formed neurons are destined for progressively more superficial layers (Sidman and Rakic, 1973; Rakic, 1976; Caviness, 1982; Takahashi et al., 1999). It is widely accepted that the time of origin of a neuron (i.e. a neuron's ‘birthday’) is intimately linked to mechanisms of specification of its laminar fate; however, the length of the cell cycle (TC), G1-phase (TG1) and the proportion of daughter cells exiting the cell cycle (Q) also increase with each day of the 7 day neuronogenetic period (Caviness et al., 1995; Takahashi et al., 1996a,b). Thus, the laminar fate of a newly born neuron is not only associated with its birthday but also with specific parameters of the proliferative population. These proliferative factors, including TC, TG1 and Q, are not normally dissociable from one another during normal development. Therefore, enquiry into the mechanisms linking one or more of these factors to mechanisms of laminar fate specification has not been feasible in vivo.

Figure 1.

Experimental design in relation to p27Kip1 overexpression. (A) A diagrammatic illustration of the period of p27Kip1 overexpression and the overall schedule of neocortical neuronogenesis (modified from Sidman and Rakic, 1973; Rakic, 1976; Caviness, 1982; Takahashi et al., 1999). Neocortical neurons arise from a pseudostratified ventricular epithelium (PVE) in the course of 11 integer cell cycles during the interval embryonic day 11 (E11) to E17 (Sidman and Rakic, 1973; Rakic, 1976; Caviness, 1982; Takahashi et al., 1999). There is a systematic progression in the laminar fate of neurons produced during each successive cell cycle and on each successive day of neuronogenesis. Thus, the earliest born neurons are destined for the deeper layers while the later formed neurons are destined for progressively more superficial layers (Sidman and Rakic, 1973; Rakic, 1976; Caviness, 1982; Takahashi et al., 1999). The diagram applies specifically to the medial cortical zone (MCZ) and neocortical field 1. In this region, infragranular neurons (IG neurons; layers V and VI) are produced mainly during cell cycles 1–6 and granular and supragranular neurons (SG neurons; layers IV, III and II) mainly during cycles 9–11. We induced p27Kip1 overexpression in the PVE of the embryonic brain by oral administration of doxycycline to pregnant dams, twice daily. The p27Kip1 overexpression was induced for a 3 day period from E12 to E14 when cycles 3–8 are in progress. Our earlier work showed that in this paradigm, p27Kip1 protein levels increased by 3-fold from E13 onwards and p27Kip1 mRNA returned to basal levels 12 h after the last doxycycline dosage (Mitsuhashi et al., 2001). The cell cycle parameters and the probability of cell cycle exit were estimated for cycles 7/8 on E14. During these cell cycles, the PVE of the MCZ produces neurons destined for both SG and IG layers of field 1, with a significant bias in favor of the IG layers (Takahashi et al., 1999). We reasoned that if p27Kip1 overexpression produced a shift in layer destination of the neurons generated in the MCZ on E14, analysis of the proportion of neurons that take up residence in SG versus IG layers in field 1 at maturity would facilitate a quantitative estimation of the direction and magnitude of the shift. (B–E) Schematic illustrations of the double S-phase labeling paradigms for the estimation of cell cycle parameters, neuronogenetic interval, probability of cell cycle exit and laminar fate of cells. We used four protocols in which S-phase markers iododeoxyuridine (IdU, solid arrowhead) and bromodeoxyuridine (BrdU, open arrowhead) were administered to pregnant mothers for estimation of cell cycle parameters on E14 (B), the neuronogenetic gradient on E16 (C), the number of cells in the Q-fraction on E14 (D) and the number and distribution of cells exiting the cell cycle on E14 in the neocortex of mice on P21 (E).

Figure 1.

Experimental design in relation to p27Kip1 overexpression. (A) A diagrammatic illustration of the period of p27Kip1 overexpression and the overall schedule of neocortical neuronogenesis (modified from Sidman and Rakic, 1973; Rakic, 1976; Caviness, 1982; Takahashi et al., 1999). Neocortical neurons arise from a pseudostratified ventricular epithelium (PVE) in the course of 11 integer cell cycles during the interval embryonic day 11 (E11) to E17 (Sidman and Rakic, 1973; Rakic, 1976; Caviness, 1982; Takahashi et al., 1999). There is a systematic progression in the laminar fate of neurons produced during each successive cell cycle and on each successive day of neuronogenesis. Thus, the earliest born neurons are destined for the deeper layers while the later formed neurons are destined for progressively more superficial layers (Sidman and Rakic, 1973; Rakic, 1976; Caviness, 1982; Takahashi et al., 1999). The diagram applies specifically to the medial cortical zone (MCZ) and neocortical field 1. In this region, infragranular neurons (IG neurons; layers V and VI) are produced mainly during cell cycles 1–6 and granular and supragranular neurons (SG neurons; layers IV, III and II) mainly during cycles 9–11. We induced p27Kip1 overexpression in the PVE of the embryonic brain by oral administration of doxycycline to pregnant dams, twice daily. The p27Kip1 overexpression was induced for a 3 day period from E12 to E14 when cycles 3–8 are in progress. Our earlier work showed that in this paradigm, p27Kip1 protein levels increased by 3-fold from E13 onwards and p27Kip1 mRNA returned to basal levels 12 h after the last doxycycline dosage (Mitsuhashi et al., 2001). The cell cycle parameters and the probability of cell cycle exit were estimated for cycles 7/8 on E14. During these cell cycles, the PVE of the MCZ produces neurons destined for both SG and IG layers of field 1, with a significant bias in favor of the IG layers (Takahashi et al., 1999). We reasoned that if p27Kip1 overexpression produced a shift in layer destination of the neurons generated in the MCZ on E14, analysis of the proportion of neurons that take up residence in SG versus IG layers in field 1 at maturity would facilitate a quantitative estimation of the direction and magnitude of the shift. (B–E) Schematic illustrations of the double S-phase labeling paradigms for the estimation of cell cycle parameters, neuronogenetic interval, probability of cell cycle exit and laminar fate of cells. We used four protocols in which S-phase markers iododeoxyuridine (IdU, solid arrowhead) and bromodeoxyuridine (BrdU, open arrowhead) were administered to pregnant mothers for estimation of cell cycle parameters on E14 (B), the neuronogenetic gradient on E16 (C), the number of cells in the Q-fraction on E14 (D) and the number and distribution of cells exiting the cell cycle on E14 in the neocortex of mice on P21 (E).

Proliferative parameters TC, TG1 and Q are regulated by the concerted actions of cyclins, cyclin-dependent kinases (cdk) and their inhibitors. The cdk inhibitor p27Kip1 is a critical modulator of the G1- to S-phase transition (Sherr and Roberts, 1999). Interestingly, it is also implicated in mechanisms of neuronal specification (Durand et al., 1997, 1998; Ohnuma et al., 1999, 2001, 2002; Livesey and Cepko, 2001; Vernon et al., 2003). We have developed a transgenic mouse model in which p27Kip1 is overexpressed selectively in response to doxycycline in the neuroepithelium of the embryonic brain during specific periods of neocortical neuronogenesis. In a previous report (Mitsuhashi et al., 2001) using this mouse model we characterized the expression of the p27Kip1 transcript and protein in the E12–E13 ventricular zone (VZ) in response to different doxycycline dosing schedules. Following only 26 h of doxycycline exposure (an interval corresponding to ∼2 cell cycle durations) beginning on E12, the 2 h bromodeoxyuridine (BrdU) labeling index in the VZ was reduced, although the width of the VZ was unchanged (Mitsuhashi et al., 2001). We interpreted the observation that VZ width did not change to indicate that Q was not altered by this brief interval of p27Kip1 overexpression. This interpretation relating to Q required that the decrease in the BrdU labeling index would be explained by a prolongation of TG1. In the present study, we re-examined the consequences of p27Kip1 overexpression for a full 2 day interval with actual direct measurements of both Q and TG1. At variance with our earlier interpretation, the present experiments show that p27Kip1overexpression increases Q but does not modify TG1. Moreover, the increase in Q is associated with a commensurate increase in the relative proportion of postmitotic neurons directed to the supragranular layers of the neocortex.

Materials and Methods

Animals

Double transgenic (DT) mice, in which p27Kip1 can be overexpressed selectively and electively in the neuroepithelium of the embryonic brain, were produced by mating the hemizygous tetOp27Kip1/– and PnestinrtTA/– FVB lines described previously (Mitsuhashi et al., 2001). The plug date was defined as embryonic day 0 (E0). Wild-type (WT) littermates were used as controls. Each experiment was initiated by induction of p27Kip1 overexpression in response to orally administered doxycycline hydrochloride (dox; Sigma) to pregnant dams at a dose of 25 μg/g body wt, twice daily from E12 to E14 (Fig. 1). All of the experimental procedures were in full compliance with institutional guidelines and the NIH Guide for the Care and Use of Laboratory Animals.

Induction of p27Kip1 Overexpression and S-phase Labeling

Each experiment was initiated by induction of p27Kip1 expression by oral dox administration to pregnant dams (Fig. 1A). The following four iododeoxyuridine–bromodeoxyuridine (IdU and BrdU, respectively) double S-phase labeling paradigms were employed.

Cell Cycle Kinetics (Fig. 1B)

A single i.p. injection of IdU (Sigma; 50 μg/g) was administered to pregnant dams carrying E14 mice at 07:00 h. It was followed at 09:00 h by a single injection of BrdU (Sigma; 50 μg/g). Mice were killed at 09:30 h (Hayes and Nowakowski, 2000).

Neuronogenetic Interval (Fig. 1C)

BrdU was administered to pregnant dams carrying E16 mice as a single i.p. injection at either 09:00 or 21:00 h on E16. Offspring were killed on postnatal day 4 (P4) and the distribution of BrdU-labeled cells was mapped in the cerebral cortex with respect to cortical cytoarchitectonic fields (Fig. 2).

Figure 2.

Effect of p27Kip1 overexpression on total neurogenetic interval. Schematic illustration of the labeling patterns in the cerebral hemisphere of postnatal day 4 (P4) mice produced by a single administration of bromodeoxyuridine (BrdU) either at 9:00 h (A, C) or 21:00 h (B, D) on embryonic day 16 (E16) (Fig. 1C). The position of BrdU-labeled cells in the cortical gray matter is illustrated in drawings of sections taken from the level of somatosensory cortex following the 09:00 or 21:00 injections (A and B, respectively) in the DT mice. Solid arrows point to the lateral extent of the BrdU labeling in each section, which corresponds to the position of the ‘wave front’ of labeled cells in the medial-to-lateral axis. With four coronal sections as A and B taken at approximately equal intervals (broken lines i–iv in C, D; section iii is presented in A, B) in double transgenic and wild-type littermates, the distribution of BrdU-positive cells on a surface view of the cortical area map was reconstructed (C, D). The neocortical area map (C, D) is adapted from the architectonic map of the adult mouse brain (Caviness, 1975) as seen from superior–lateral aspect. The solid lines indicate the approximate position of the ‘wave front’ of BrdU-labeled cells in the two BrdU injection paradigms. The BrdU-labeling displays an ascending rostrolateral to caudomedial gradient, approximated by the gray shading, which corresponds to the transverse neurogenetic gradient. In both the genotypes, following the 9:00 h BrdU injection, the front of BrdU-labeled cells is located a short distance from the rhinal fissure along an axis that passes through insular field 14 laterally and the lateral temporal and perirhinal field 36 posteriorly (solid line in C). Following the 21:00 h BrdU injection, the front of BrdU-labeled cells is located in the lateral parietal fields 3a and 40 medially and the lateral occipital field 18a posteriorly in both the genotypes (solid line in D). Thus, the position of the proliferative ‘wave fronts’ produced by the 9:00 h to 21:00 h BrdU injections is an identical in both the genotypes. CC, corpus callosum; CTX, cortex; HI, hippocampus. Scale bars in A, B = 100 μm.

Figure 2.

Effect of p27Kip1 overexpression on total neurogenetic interval. Schematic illustration of the labeling patterns in the cerebral hemisphere of postnatal day 4 (P4) mice produced by a single administration of bromodeoxyuridine (BrdU) either at 9:00 h (A, C) or 21:00 h (B, D) on embryonic day 16 (E16) (Fig. 1C). The position of BrdU-labeled cells in the cortical gray matter is illustrated in drawings of sections taken from the level of somatosensory cortex following the 09:00 or 21:00 injections (A and B, respectively) in the DT mice. Solid arrows point to the lateral extent of the BrdU labeling in each section, which corresponds to the position of the ‘wave front’ of labeled cells in the medial-to-lateral axis. With four coronal sections as A and B taken at approximately equal intervals (broken lines i–iv in C, D; section iii is presented in A, B) in double transgenic and wild-type littermates, the distribution of BrdU-positive cells on a surface view of the cortical area map was reconstructed (C, D). The neocortical area map (C, D) is adapted from the architectonic map of the adult mouse brain (Caviness, 1975) as seen from superior–lateral aspect. The solid lines indicate the approximate position of the ‘wave front’ of BrdU-labeled cells in the two BrdU injection paradigms. The BrdU-labeling displays an ascending rostrolateral to caudomedial gradient, approximated by the gray shading, which corresponds to the transverse neurogenetic gradient. In both the genotypes, following the 9:00 h BrdU injection, the front of BrdU-labeled cells is located a short distance from the rhinal fissure along an axis that passes through insular field 14 laterally and the lateral temporal and perirhinal field 36 posteriorly (solid line in C). Following the 21:00 h BrdU injection, the front of BrdU-labeled cells is located in the lateral parietal fields 3a and 40 medially and the lateral occipital field 18a posteriorly in both the genotypes (solid line in D). Thus, the position of the proliferative ‘wave fronts’ produced by the 9:00 h to 21:00 h BrdU injections is an identical in both the genotypes. CC, corpus callosum; CTX, cortex; HI, hippocampus. Scale bars in A, B = 100 μm.

The Number of Cells in the Q Fraction (NQ) (Fig. 1D)

IdU was administered as a single i.p. injection to pregnant dams carrying E14 mice at 07:00 h. This was followed by sequential i.p. administrations of BrdU every 3 h from 09:00 to 24:00 h corresponding to an interval longer than TC – TS (‘birth hour method’: Takahashi et al., 1996a,b). This design identifies a cohort of cells that were in S-phase between 07:00 and 09:00 h and that exited the cell cycle following the S-phase. The cohort is labeled only with IdU (Fig. 3C). Since we count only IdU-only labeled cells, the cells that reenter S-phase (rather than exiting the cell cycle) ‘disappear’, as they will be double-labeled with BrdU and IdU (Fig. 3C).

Figure 3.

Effect of p27Kip1 overexpression on cell output. Micrographs of 4 μm thick coronal sections through the heads of embryonic day 14 wild-type (WT, A) and double transgenic (DT, B) littermates taken from approximately the mid-hemisphere level and processed for iododeoxyuridine (IdU) and bromodeoxyuridine (BrdU) immunohistochemistry. A higher magnification view (C) illustrates the IdU-only labeled cells (solid arrowheads, blue cells) and BrdU-labeled cells (open arrowheads, blue/brown cells), which can be reliably distinguished and counted in these preparations. The labeled cells are distributed throughout the telencephalic neuroepithelium and demarcate it from the marginal zones, which contain only unlabeled cells. Histological appearance of the cerebral wall was similar in the WT and DT mice (compare A to B). The IdU-only labeled cells were counted in the ventricular zone (VZ) of the cerebral wall at two regions along the medial-lateral axis, the medial and lateral cortical zones (MCZ and LCZ, respectively). These cells are referred to as Q-fraction or Q cells. The distribution of the Q cells in the VZ and the intermediate zone (IZ) of the cerebral wall was similar in the WT and DT littermates both in the MCZ (D) and the LCZ (E) suggesting that the rates of cell exit from the VZ and migration through the IZ were not influenced by the p27Kip1 overexpression. Scale bars in A and B = 100 μm, in C = 10 μm.

Figure 3.

Effect of p27Kip1 overexpression on cell output. Micrographs of 4 μm thick coronal sections through the heads of embryonic day 14 wild-type (WT, A) and double transgenic (DT, B) littermates taken from approximately the mid-hemisphere level and processed for iododeoxyuridine (IdU) and bromodeoxyuridine (BrdU) immunohistochemistry. A higher magnification view (C) illustrates the IdU-only labeled cells (solid arrowheads, blue cells) and BrdU-labeled cells (open arrowheads, blue/brown cells), which can be reliably distinguished and counted in these preparations. The labeled cells are distributed throughout the telencephalic neuroepithelium and demarcate it from the marginal zones, which contain only unlabeled cells. Histological appearance of the cerebral wall was similar in the WT and DT mice (compare A to B). The IdU-only labeled cells were counted in the ventricular zone (VZ) of the cerebral wall at two regions along the medial-lateral axis, the medial and lateral cortical zones (MCZ and LCZ, respectively). These cells are referred to as Q-fraction or Q cells. The distribution of the Q cells in the VZ and the intermediate zone (IZ) of the cerebral wall was similar in the WT and DT littermates both in the MCZ (D) and the LCZ (E) suggesting that the rates of cell exit from the VZ and migration through the IZ were not influenced by the p27Kip1 overexpression. Scale bars in A and B = 100 μm, in C = 10 μm.

Identification of a Cohort of cells that Exited the Cell Cycle on E14 in the P21 Cortex (Fig. 1E)

The design is identical to that for NQ estimation (Fig. 1D) except that the 2 h cohort of IdU-only labeled cells is examined in the cerebral cortex on P21 (Fig. 4D,E) (Takahashi et al., 1999).

Figure 4.

Effect of p27Kip1 overexpression on laminar fate of neocortical neurons. Effects of p27Kip1 overexpression during the embryonic period on the gross appearance of the brain, cytoarchitecture of the neocortex and laminar fates of cells generated on embryonic day 14 (E14) examined at postnatal day 21 (P21). The brains of wild-type (WT) and double transgenic (DT) mice appear similar in size and shape at P21 (A). B and C are micrographs of 4 μm thick coronal sections through field 1 (B) and field 40 (C) of WT and DT mice stained with basic fuchsin to reveal cortical lamination. The cytoarchitecture of fields 1 and 40 is preserved in the DT mice. However, the thickness of the cortical gray matter is reduced dramatically in the DT mouse in field 1 compared to the WT littermates (B). The thickness of the gray matter in field 40 is similar in the DT and WT mice (C). The reduction in the thickness of field 1 is caused mainly by a marked reduction in the thickness of SG layers (layers II/III and IV; B). The thickness of IG layers (layers V and VI) is similar in the DT and WT littermates in fields 1 and 40 (B, C). Micrographs of 4 μm thick coronal sections through field 1 of P21 WT and DT littermates processed for iododeoxyuridine (IdU) and bromodeoxyuridine (BrdU) immunohistochemistry. At higher magnification (box insert), IdU-only (blue labeled) cells are readily distinguished from cells labeled with BrdU (brown containing). (D). Quantitative analysis of the distribution of the IdU-only labeled cells (i.e. cells that exited the cell cycle on E14) in field 1 at P21 (E) revealed that the majority of the cells was distributed in the IG layers in the WT cortex (blue line) and in the SG layers in the DT cortex (red line) at P21. Thus, layer destination of cells generated on E14 was shifted toward the SG layers in the DT mice. ML = molecular layer (layer I). Scale bar in A = 5 mm; BD = 100 μm.

Figure 4.

Effect of p27Kip1 overexpression on laminar fate of neocortical neurons. Effects of p27Kip1 overexpression during the embryonic period on the gross appearance of the brain, cytoarchitecture of the neocortex and laminar fates of cells generated on embryonic day 14 (E14) examined at postnatal day 21 (P21). The brains of wild-type (WT) and double transgenic (DT) mice appear similar in size and shape at P21 (A). B and C are micrographs of 4 μm thick coronal sections through field 1 (B) and field 40 (C) of WT and DT mice stained with basic fuchsin to reveal cortical lamination. The cytoarchitecture of fields 1 and 40 is preserved in the DT mice. However, the thickness of the cortical gray matter is reduced dramatically in the DT mouse in field 1 compared to the WT littermates (B). The thickness of the gray matter in field 40 is similar in the DT and WT mice (C). The reduction in the thickness of field 1 is caused mainly by a marked reduction in the thickness of SG layers (layers II/III and IV; B). The thickness of IG layers (layers V and VI) is similar in the DT and WT littermates in fields 1 and 40 (B, C). Micrographs of 4 μm thick coronal sections through field 1 of P21 WT and DT littermates processed for iododeoxyuridine (IdU) and bromodeoxyuridine (BrdU) immunohistochemistry. At higher magnification (box insert), IdU-only (blue labeled) cells are readily distinguished from cells labeled with BrdU (brown containing). (D). Quantitative analysis of the distribution of the IdU-only labeled cells (i.e. cells that exited the cell cycle on E14) in field 1 at P21 (E) revealed that the majority of the cells was distributed in the IG layers in the WT cortex (blue line) and in the SG layers in the DT cortex (red line) at P21. Thus, layer destination of cells generated on E14 was shifted toward the SG layers in the DT mice. ML = molecular layer (layer I). Scale bar in A = 5 mm; BD = 100 μm.

Tissue Processing and Histology

P4 and P21 mice were anesthetized (Ketamine, 50 μg/g; Ketalar, Abbott; Xylazine 10 μg/g; Rompun, Bayer; i.p.) and perfused through the heart with 4% paraformaldehyde in phosphate buffer, pH 7.2. Brains were removed, embedded in paraffin wax. E14 mice were removed by hysterotomy from anesthetized dams and decapitated. The entire heads were embedded in paraffin wax. The postnatal brains and embryonic heads were sectioned at a thickness of 4 μm in the coronal plane. The sections were processed for IdU–BrdU double immunohistochemistry as described below. Tail (postnatal) or trunk (embryo) samples were collected for genotyping from anesthetized mice prior to tissue fixation.

IdU–BrdU Immunohistochemistry

Paraffin-embedded sections were cleared in Histoclear (National Diagnostic) and xylene (Fisher Scientific), rehydrated in graded ethanol and PBS. The sections were immersed in 5% acetic acid overnight, then washed with distilled water, and treated with 0.2% trypsin (37°C, 20 min) and 2 N HCl (30 min). The sections from E14 mice were microwaved in a solution of 0.01 M sodium citrate, pH 6.2, for 15 min and not trypsinized, as trypsinization adversely affected the integrity of the sections. Non-specific antibody reaction was blocked with 1.5% normal horse serum in PBS (30 min). Sections were incubated with mouse monoclonal antibody Br3 (Caltag Lab., 0.025% in PBS, 1.5% normal horse serum and 0.5% Tween-20) for 30 min, biotinylated anti-mouse IgG (0.5% in PBS) for 45 min and then with Vector ABC-peroxidase solution (ABC Peroxidase Elite kit, Vector) for 60 min. The sections were reacted with diaminobenzidine (DAB, 0.05%, Sigma) and H2O2 (0.01%) for 8– 15 min, rinsed with PBS, immersed in 5% acetic acid for 30 min and washed with distilled water. The sections were then incubated with mouse monoclonal antibody IU4 (Caltag Lab., 0.025% in PBS and 0.5% Tween-20) for 30 min, biotinylated anti-mouse IgG (0.5% in PBS) for 45 min and then with vector ABC alkaline phosphatase solution (ABC-Alkaline phosphatase kit, Vector) for 60 min. The sections were reacted with Alkaline Phosphatase Substrate (Vector Blue-Alkaline Phosphatase Substrate Kit III, Vector) for 15 min. The sections were rinsed with distilled water and coverslipped with Crystal Mount (Biomedia).

TUNEL Histochemistry

Paraformaldehyde-fixed, paraffin-embedded, 4 μm thick, coronal sections were processed for terminal deoxynucleotidyl transferase-mediated dUTP-biotin nick-end labeling (TUNEL) according to the manufacturer's instructions (ApopTag kit, Intergen Purchase). The sections were counterstained with 0.1% aqueous basic fuchsin. TUNEL-positive cells were counted in the cortical gray matter (layer VI to the pial surface) of fields 1 and 40 (Verney et al., 2000).

Analysis of IdU–BrdU-labeled Cells

The monoclonal antibody IU4 detects both IdU and BrdU. IU4-labeled cells stain blue due to the alkaline phosphatase reaction product. Br3 detects only BrdU. Br3-labeled cells stain brown due to the DAB reaction product (Hayes and Nowakowski, 2000). Thus, although the IdU–BrdU labeling paradigms (Fig. 2B,D,E) yield three types of labeled cells — IdU-only, BrdU-only, and BrdU and IdU double-labeled cells — only the IdU-only labeled cells (blue, solid arrowhead in Fig. 3C), which were in S-phase at the time of the IdU injection and exited the S-phase by the time of the BrdU injection, can be considered to be specifically identifiable. Both the BrdU-only labeled and IdU–BrdU-double labeled cells (blue/brown, open arrowhead in Fig. 3C) would have been in S-phase at the time of the BrdU injection (Nowakowski et al., 1989; Takahashi et al., 1999; Hayes and Nowakowski, 2000). Due to the antibody specificity, we attach no significance to the double labeled versus BrdU-only labeled population. Note, specifically, that our assay depends only on identifying the IdU-only labeled cells. There are also cells not labeled with either IdU or BrdU. Those cells were not in S-phase when effective labeling concentrations of either tracer were available (Fig. 3AC).

The analyses were performed at two locations along the medial–lateral axis of the VZ, namely in the medial cortical zone (MCZ) and the lateral cortical zone (LCZ; Fig. 3A,B). We counted all three types of labeled cells in the MCZ and the LCZ of E14 mice (Fig. 3A,B) within a sector that was 100 μm in its medial–lateral dimension and 4 μm (corresponding to section thickness) in its rostral–caudal dimension (Fig. 3C). The radial dimension of the sector was divided into bins. Each bin is 10 μm high. The bins were numbered 1, 2, 3, etc., from the ventricular margin outward. Following the IdU-only labeled cell counts, the coverslips were removed and the sections were stained with 0.1% aqueous basic fuchsin so that all cells in the VZ could be counted. The number of unlabeled cells was obtained by subtracting IdU and/or BrdU labeled cells from all cells.

Labeled cells were counted also in the medially located field 1 and laterally located field 40 (Caviness, 1975) in P21 cerebral cortex (Fig. 4D,E). For convenience and consistency with our previous work (Takahashi et al., 1999; Caviness et al., 2003), we divided the cortical gray matter into granular–supragranular (SG: layers IV–II/III) and infragranular (IG: layers VI–V) layers. In order to correlate judgments of the architectonic landmarks of the two cortical fields with the various labeling patterns in P21 mice, we stained the IdU–BrdU-labeled sections with basic fuchsin as described above (Fig. 4B,C). In these sections, we identified the border between layers V (internal pyramidal cell layer) and IV (internal granular cell layer) based on cell morphology (large pyramidal cells in layer V and smaller granular cells in IV) and cell packing density (relatively low packing density in layer V compared to that in layer IV). The entire gray matter between the pial surface and white matter/layer VI border was divided into bins as in the E14 studies. We divided the cortical gray matter into SG and IG layers (Fig. 4B,C). We confirmed the registration of the IdU-only labeled cells to IG and SG layers, retrospectively by re-registering the bins with cortical layers (Fig. 4D, Table 4). We also counted the total number of nuclei (i.e. all nuclei) in IG and SG (Table 3).

Cell counts were performed on non-adjacent sections to avoid counting the same cell more than once. The analysis was performed on data collected from the brains of five embryos/mice in each genotype, one or two mice from each of four or five different litters. In each brain, four or five non-adjacent coronal sections were analyzed. Statistical analyses were performed using t-test and ANOVA (Microsoft Excel). In one case a statistical outlier (>3 SD from the mean) was removed from the analysis. A P < 0.05 was used as a significance level in all cases.

Analysis of Cell Cycle Kinetics

To measure the length of the cell cycle (TC) and its constituent phases [TG1, TS (the length of S-phase) and TG2+M (combined lengths of G2- and M-phases)], three types of labeled cells in the labeling paradigm of Figure 2B were counted in the VZ: IdU-only labeled cells (NI), BrdU-only and BrdU and IdU double-labeled cells (NB) and all cells (NT). The number of unlabeled cells (N0) was calculated by N0 = NT – NI – NB (Table 1). Kinetic parameters are derived according to a modification of the double-labeling algorithm formulated by Hayes and Nowakowski (2000). We determined that TG2+M was ∼2 h, as the majority of M-phase cells (99%) were labeled with IdU-only during the 2 h labeling period (Takahashi et al., 1995) in both WT and DT mice. The growth fraction (GF; Takahashi et al., 1995) was ∼1.0, as cumulative BrdU administration over 15 h labeled all cells in the VZ of WT and DT mice (data not shown). Since the number of cells after M-phase is doubled by mitosis, the number of unlabeled cells that contributes to the cell cycle is N0/2. Similarly the total number of cells that contributes to the cell cycle is NT – N0/2. With these considerations, TC and TS are estimated from the number of cells of each labeling group according to the following equations. 

(1)
\[NI/(NT{-}N0/2){=}2/T_{C}T_{C}{=}2(NT{-}N0/2)/NI\]
 
(2)
\[NI/NB{=}2/T_{S}T_{S}{=}2NB/NI\]
Note that both (1) and (2) are approximations because we ignore the 30 min survival time. This was done because there is an 10–15 min interval required for effective labeling concentrations for S-phase markers in the VZ (Hayes and Nowakowski, 2000).

Table 1

Analysis of cell cycle kinetics and cell output


 
Genotype
 
NI
 
NB
 
NT
 
N0
 
MCZ WT 14.8 ± 0.7 19.5 ± 2.8 137.3 ± 10.0 103.0 ± 9.2 
 DT 15.6 ± 0.9 20.3 ± 0.9 125.1 ± 8.1 89.2 ± 8.5 
LCZ
 
WT 11.9 ± 1.4 24.8 ± 3.4 125.7 ± 5.2 89.0 ± 4.1 
 DT
 
11.9 ± 1.3
 
22.7 ± 3.5
 
118.3 ± 6.4
 
83.8 ± 5.2
 

 
Genotype
 
NI
 
NB
 
NT
 
N0
 
MCZ WT 14.8 ± 0.7 19.5 ± 2.8 137.3 ± 10.0 103.0 ± 9.2 
 DT 15.6 ± 0.9 20.3 ± 0.9 125.1 ± 8.1 89.2 ± 8.5 
LCZ
 
WT 11.9 ± 1.4 24.8 ± 3.4 125.7 ± 5.2 89.0 ± 4.1 
 DT
 
11.9 ± 1.3
 
22.7 ± 3.5
 
118.3 ± 6.4
 
83.8 ± 5.2
 

Mean ± SEM values of IdU-labeled (NI), BrdU- or IdU+BrdU-labeled (NB) and all cells (NT) per unit area of the VZ in embryonic day 14 wild-type (WT) and double transgenic (DT) mice produced by the S-phase labeling protocol shown in Figure 1B. The number of unlabeled cells (N0) was calculated using the formula N0 = (NT −NI – NB). The values are shown separately for the medial cortical zone (MCZ) and lateral cortical zone (LCZ) of the cerebral wall.

Analysis of Neurogenetic Interval

The BrdU-labeled sections of P4 brains were photographed using a Nikon Eclipse E400 microscope and a digital camera (SPOT RT Slider Camera, RT230-2, Diagnostic Instruments). The images were processed by SPOT Advanced v.3.5 software (Diagnostic Instruments) and Adobe Photoshop 7.0 (Adobe). The position of the BrdU-labeled cells in the cortical gray matter was superimposed on the outlines of the sections (Fig. 2A,B). The BrdU labeling pattern reconstructed from the serial sections was superimposed on a surface view of the cortical area map (Fig. 2C,D).

Results

The Double Transgenic (DT) Mouse Line and Induction of p27Kip1

The mating between the hemizygous tetOp27Kip1/– and PnestinrtTA/– mice was as frequently associated with vaginal plugs as mating among WT mice, although the incidence of successful impregnation was lower in the former. Once pregnancy was established, however, the course of gestation was not altered by the hemizygous condition and parturition occurred on E19, just as for the WT mice. In addition, the number of live born offspring per litter was not different in the hemizygous matings compared to the WT matings. Finally, the DT, hemizygote and WT genotype ratios from tetOp27Kip1/– × PnestinrtTA/– mating approximated the expected Mendelian ratio of 1:2:1.

Cell Cycle Kinetics

We used a double S-phase labeling method to estimate cell cycle parameters in the VZ of the cerebral wall in E14 DT and WT littermates exposed to dox from E12 onwards (Fig. 1B). We chose E14 for analysis because it is the time when the laminar fate of neurons shifts from layer V to IV in the somastosensory cortex of normal mouse (Takahashi et al., 1999) (Fig. 1A). The analyses were performed at MCZ and LCZ (Fig. 3A,B). The progenitors in the LCZ are developmentally ‘in advance’ of those in the MCZ by at least 24 h, with respect to neuronogenetic schedule (Takahashi et al., 1995, 1996a,b, 1999). Thus, this single experiment is the equivalent of examining the same area of cortex at both E14 and E15, and the analyses in the two zones provide a method to determine if the effects of p27Kip1 overexpression are dependent upon the stage of maturation of the progenitors.

The number of IdU- and BrdU-labeled (i.e. BrdU+ or IdU+/BrdU+) cells was recorded in the MCZ and LCZ of DT and WT littermates (Table 1) to estimate the length of cell cycle (TC) and its phases [TG1, TS (the duration of S-phase) and TG2+M (combined lengths of G2- and M-phases)] according to a method described previously (Hayes and Nowakowski, 2000). The cell cycle parameters did not differ between DT and WT mice either in the MCZ or the LCZ (Table 2). Thus, overexpression of p27Kip1 from E12 to E14 does not alter either the total length of cell cycle or the lengths of its constituent phases on E14.

Table 2

Cell cycle parameters


 
Genotype
 
TC (h)
 
TG2+M (h)
 
TS (h)
 
TG1 (h)
 
NQ (cells)
 
Q
 
MCZ WT 11.7 ± 1.0 2.0 2.7 ± 0.4 7.1 ± 0.7 10.5 ± 0.4 0.35 
 DT 10.5 ± 1.0 2.0 2.6 ± 0.2 5.8 ± 0.9 12.5 ± 0.8* 0.40* 
LCZ
 
WT 14.6 ± 2.0 2.0 4.5 ± 0.8 8.1 ± 1.4 9.6 ± 0.3 0.40 
 DT
 
13.6 ± 2.2
 
2.0
 
4.1 ± 1.0
 
7.2 ± 1.5
 
10.5 ± 1.0
 
0.41
 

 
Genotype
 
TC (h)
 
TG2+M (h)
 
TS (h)
 
TG1 (h)
 
NQ (cells)
 
Q
 
MCZ WT 11.7 ± 1.0 2.0 2.7 ± 0.4 7.1 ± 0.7 10.5 ± 0.4 0.35 
 DT 10.5 ± 1.0 2.0 2.6 ± 0.2 5.8 ± 0.9 12.5 ± 0.8* 0.40* 
LCZ
 
WT 14.6 ± 2.0 2.0 4.5 ± 0.8 8.1 ± 1.4 9.6 ± 0.3 0.40 
 DT
 
13.6 ± 2.2
 
2.0
 
4.1 ± 1.0
 
7.2 ± 1.5
 
10.5 ± 1.0
 
0.41
 

Mean ± SEM values of cell cycle parameters (the length of cell cycle TC; S phase TS; combined length of G2 and M phase TG2+M; G1 phase TG1) in the medial cortical zone (MCZ) and lateral cortical zone (LCZ) of embryonic day 14 wild-type (WT) and double transgenic (DT) littermates. The number of Q cells (NQ) was calculated using the S-phase labeling protocol shown in Fig. 1D. The Q fraction was calculated from NQ and NP+Q (= 2 × NI) and is significantly larger in the MCZ of the DT mice. No other comparisons are statistically significant.

*

P <0.05; t-test.

Duration of the Neuronogenetic Interval

Since the cell cycle parameters did not change in the DT mice, the total number of integer cell cycles executed over the duration of the neuronogenetic interval itself would not be expected to change. We tested that prediction next. Neuronogenesis is completed throughout the neocortical PVE of the mouse early on E17 (Caviness, 1982; Takahashi et al., 1995; Miyama et al., 1997). The process of termination of neuronogenesis follows a transverse neurogenetic gradient (Bayer and Altman, 1991) such that ‘wave fronts’ of initiation and termination of neuronal production advance along the rostrolateral to caudomedial axis of the hemisphere (Miyama et al., 1997). The position of the termination wave front is particularly clear at the end of the neuronogenetic period because it marks the border between an area of the PVE that continues to produce neurons and an area in which the PVE has involuted and is no longer producing cells. If p27Kip1 overexpression altered the neuronogenetic interval, the position of the termination wave front at the end of the neuronogenetic period would be displaced in the DT mice compared to the WT littermates, and the degree of displacement would be commensurate with the extent of alteration in the neuronogenetic interval.

We visualized the position of the termination wave front in DT and WT mice at the end of the neuronogenetic period by using a BrdU labeling method (Fig. 1C). We administered BrdU at 09:00 or 21:00 h on E16 and observed the distribution of BrdU-positive cells in the neocortical gray matter in serial coronal sections of the brain on P4 (see examples in Fig. 2A,B). The position of BrdU-positive cells in the cortical gray matter was recorded with respect to the section outlines (Fig. 2A,B). The lateral extent of the spread of the BrdU label (the termination wave front) for representative cases from the 09:00 and 21:00 injections is indicated by arrows at the level of the somatosensory (barrel field) cortex in Figure 2A,B. The termination wave front is located more medially following the 9:00 injection than the 21:00 injection at this coronal level (compare Fig. 2A to 2B), as predicted by the neurogenetic gradient. We reconstructed the termination wave front from the serial sections and superimposed it on a surface view of the cortical area map (Fig. 2C,D). In both WT and DT mice, following the 09:00 h BrdU injection, the front of labeled cells was located a short distance from the rhinal fissure along a plane that passed through insular field 14 laterally and the lateral temporal and perirhinal field 36 posteriorly (Fig. 2C), in accordance with the predicted labeling pattern (Caviness, 1975; Miyama et al., 1997). Following the 21:00 h BrdU injection, the termination wave front was located further into the lateral parietal fields 3a and 40 medially and the lateral occipital field 18a posteriorly in both the genotypes (Fig. 2D). Thus, we found that the tangential and radial distributions of the E16 BrdU labeled cells in the P4 neocortex were indistinguishable between DT (Fig. 2A,B) and WT littermates (data not shown). Therefore, within the limits of resolution of the methods used here, neither the duration of the neuronogenetic interval nor the slope of the transverse neurogenetic gradient were altered by the overexpression of p27Kip1.

Probability of Cell Cycle Exit

Each of the two daughter cells resulting from a round of cell division has a quantifiable probability of re-entry into or exit from the cell cycle. These probability values are designated as P (for reentry) and Q (for exit) (Takahashi et al., 1996a,b). P + Q is always 1.0, as these two fates are complementary and cell death in the PVE is small, i.e. <1% per cell cycle (Haydar et al., 2000b; Cai et al., 2002). P and Q are calculated based on an estimation of the number of cells in the P and Q fractions (NP and NQ, respectively). This is accomplished by labeling a cohort of cells undergoing S-phase over an experimentally defined interval and following the cohort as it executes G2-, M- and G1-phases so as to distinguish cells that exit the cell cycle from those that re-enter it (Takahashi et al., 1996a,b). We calculated NQ and NP in E14 DT and WT embryos exposed to dox from E12 to E14 using the IdU–BrdU labeling protocol. We calculated NQ using the labeling protocol shown in Figure 1D. In this protocol cells stained blue (labeled only with IdU) are the NQ cells (Fig. 3AC) (Takahashi et al., 1996a,b). We calculated NP+Q by doubling the value of NI (Table 1) obtained using the labeling protocol shown in Figure 1B. Q was calculated using the formula Q = NQ/(NP + NQ). We found that there was a statistically significant increase (P-value = 0.03, t-test) in Q in MCZ of DT mice compared to WT littermates (Table 2), whereas the measures of Q in the LCZ of DT and WT littermates were not different.

We also recorded the pattern of distribution of the NQ cells in the VZ and the intermediate zone (IZ) as they exited the VZ and migrated toward the marginal zones. The distribution of these cells represented the distance of migration ∼6 h after exiting cell cycle. The cells which exited cell cycle have left the VZ and migrated across the IZ to traverse ∼90% (MCZ) and ∼70% (LCZ) of the height of the cerebral wall at the time of measurement (Takahashi et al., 1996a,b). Therefore the pattern of distribution represented the migratory rate of these cells, which was indistinguishable in DT and WT littermates (Fig. 3D,E). Moreover, the architectonic appearance of the developing cortex in DT animals was normal, including cortical plate and subplate and the course of the sagittal stratum subjacent to the cortical strata. There was no suggestion of architectonic anomalies such as are seen in reeler (Caviness, 1982) and other ‘cortical mutants’ (Walsh, 2000; Ohshima et al., 2001; Hammond et al., 2004) where there is disorder of migration and postmigration mechanisms of laminar assembly. Therefore, patterns of cell exit from the VZ and migration across the IZ as well as postmigration mechanisms of laminar assembly are unperturbed by p27Kip1 overexpression. The analysis of postnatal migration and laminar assembly will be considered later in the Discussion.

Cytoarchitecture of the Cerebral Cortex at Postnatal Day 21

We examined the histology of the cerebral cortex in the DT and WT littermates on P21. There were no differences between WT and DT mice in the gross appearance of the brain as seen in a dorsal view (Fig. 4A). In histological sections, however, there are clear differences in the cortical width. Thus, the neocortex is distinctly thinner in the DT, especially in field 1 (Fig. 4B) albeit less in field 40 (Fig. 4C). This is a quantitative difference only. In particular, the stratification and laminar cytoarchitectonic patterns are similar in the DT and WT mice in neocortical fields 1 and 40 (Fig. 4B,C). Moreover, regionally distinguishing architectonic features — e.g. the barrel patterns of area 3, and the full array of sublaminar patterns through the parietal, frontal temporal, occipital and medial hemispheric fields — were also preserved in the DT mice (data not shown). Thus, the p27Kip1 overexpression from E12 to E14 did not produce gross malformations or architectonic pattern abnormalities of the cerebral cortex or other brain regions at P21. We address the quantitative issues related to cortical thickness in the next section.

Cortical Thickness and Neuronal Number at P21

Overexpression of p27Kip1 was associated with a modest reduction in the radial thickness and the total number of neurons in the cortex in field 1 in the DT mice (∼ 8% and ∼10%, respectively) (Table 3). This was due entirely to a more substantial reduction in the thickness and number of cells of the SG layers (∼24% and ∼ 18%, respectively). There was no detectable difference in the thickness of the IG layers of field 1. There was no detectable difference in either the full cortical thickness or cell numbers of field 40 overall or that of either the SG or IG layers in that field.

Table 3

The thickness and cell number of P21 cortex

  Field 1
 
   Field 40
 
   

 

 
WT
 
DT
 
P
 
% change
 
WT
 
DT
 
P
 
% change
 
Thickness of ctx (μm)a Total 830 ± 14 759 ± 31 0.04 −8* 977 ± 28 939 ± 632 0.55  
 SG 475 ± 24 408 ± 16 0.03 −24* 521 ± 8 523 ± 65 0.97  
 IG 355 ± 23 475 ± 20 0.08  457 ± 22 417 ± 38 0.16  
Cell number (cells/area)b
 
Total 179 ± 6 161 ± 6 0.001 −10** 175.6 ± 7 173.4 ± 7 0.76  
 SG 82.2 ± 2.8 67.7 ± 2.8 0.00009 −18** 83.1 ± 3.3 86.8 ± 6.8 0.46  
 IG
 
97.0 ± 3.6
 
93.2 ± 2.3
 
0.23
 

 
92.5 ± 4.8
 
86.6 ± 2.4
 
0.16
 

 
  Field 1
 
   Field 40
 
   

 

 
WT
 
DT
 
P
 
% change
 
WT
 
DT
 
P
 
% change
 
Thickness of ctx (μm)a Total 830 ± 14 759 ± 31 0.04 −8* 977 ± 28 939 ± 632 0.55  
 SG 475 ± 24 408 ± 16 0.03 −24* 521 ± 8 523 ± 65 0.97  
 IG 355 ± 23 475 ± 20 0.08  457 ± 22 417 ± 38 0.16  
Cell number (cells/area)b
 
Total 179 ± 6 161 ± 6 0.001 −10** 175.6 ± 7 173.4 ± 7 0.76  
 SG 82.2 ± 2.8 67.7 ± 2.8 0.00009 −18** 83.1 ± 3.3 86.8 ± 6.8 0.46  
 IG
 
97.0 ± 3.6
 
93.2 ± 2.3
 
0.23
 

 
92.5 ± 4.8
 
86.6 ± 2.4
 
0.16
 

 

Mean ± SEM values of the cortical thicknesses and mean of cell numbers in field 1 and field 40 of postnatal day 21 wild type (WT) and double transgenic (DT) littermates.

a

Statistical analyses with respect to thickness were performed by t-test.

b

Statistical analyses with respect to cell number by ANOVA.

‘% change’ refers to the WT – DT difference as a percent of WT.

*

P < 0.05;

**

P < 0.01.

This contrast in the response to p27Kip1 overexpression between medial and lateral localized cortical fields and of SG with respect to IG layers within field 1 is predicted from a model formalized earlier (Caviness et al., 2000, 2003). It will be considered further in the Discussion.

Laminar Fates of Cells ‘Born’ on E14

We next examined if the laminar fates of cells that exited the cell cycle (NQ cells) on E14 were different between the DT and WT littermates on P21, using a modification of the double-S-phase labeling method (Figs 1E, 4D,E) that was used to estimate NQ on E14 (Fig. 1D). In this method, the E14 NQ cells can be recognized in the P21 cortex as blue cells (Fig. 4D,E). These cells appear to be neocortical projection neurons based on the morphology of the somata (data not shown) and also based on previous reports that the output of the neocortical VZ on E14 is virtually exclusively projection neurons (Takahashi et al., 1995; Qian et al., 2000; Malatesta et al., 2003). It is possible that some of the NQ cells are interneurons. However, the interneurons constitute only ∼15–25% of the total number of neocortical neurons in rodents (Ren et al., 1992; Beaulieu, 1993) and are unlikely to introduce significant bias in our data. Therefore, we consider the E14 NQ cells to be projection neurons.

We recorded the number and radial distribution of the NQ cells in neocortical fields 1 and 40 in WT and DT mice (field 1; Fig. 4E). First, we measured the total number of NQ cells in the gray matter in fields 1 and 40. We found that the total number increased by ∼52% in field 1 (destination of cells originating in the MCZ; Takahashi et al., 1999) and ∼43% in field 40 (destination of cells originating in the LCZ; Takahashi et al., 1999) in the DT mice relative to the WT littermates (Table 4). The differences were statistically significant in both the fields. The increase in NQ cells in field 1 of the DT mice was consistent with the increases in NQ and Q in the MCZ of DT mice at E14 (Table 2). However, the increase in NQ cells in field 40 was not paired with an increase in NQ or Q in the LCZ of DT mice at E14. Plausibly the discrepancy here is attributable to the variability of measurement in the relatively small numbers of DT and WT littermates recovered for the E14 studies.

Next, we examined the distribution of the E14 NQ cells in the SG and IG layers. We found that in the WT mice ∼22% of the NQ cells were distributed to SG layers in field 1 and ∼56% in field 40 (Table 4), reflecting the maturity differences produced by the down gradient position of field 1 relative to field 40 along the transverse neurogenetic gradient. In the DT mice, by contrast, nearly 42% of the NQ cells were found in SG layers in field 1 and 65% in field 40. Thus, there was an increase in the proportion of the NQ cells of the cohort that had been directed to SG layers in the DT mice in both fields 1 and 40 compared to the WT mice (Fig. 4E). This corresponded to a 94% increase in the proportion of NQ cells directed to SG layers in field 1 of the DT mice compared to the WT mice. In field 40, by contrast, although 65% of the NQ cells were directed to the SG layers in the DT mice, this represented only a 20% increase over the WT mice (Table 4).

Table 4

Laminar fate of neurons born at E14


 
Genotype
 
Cell number
 
% change
 
IG cells
 
SG cells
 
% SG cells
 
% change
 
Field 1 WT 6.8 ± 0.4 51.9 ± 16.9 5.3 ± 0.6 1.4 ± 0.6 21.5 ± 8.1 94.2 ± 150.7 
 DT 10.3 ± 0.1**  6.1 ± 0.8 4.2 ± 0.9* 41.7 ± 7.0**  
Field 40
 
WT 5.9 ± 0.7 43.4 ± 19.9 2.9 ± 0.7 3.0 ± 0.2 56.2 ± 9.5 19.9 ± 20.1 
 DT
 
8.5 ± 0.7*
 

 
3.1 ± 0.6
 
5.4 ± 0.6**
 
64.7 ± 5.8
 

 

 
Genotype
 
Cell number
 
% change
 
IG cells
 
SG cells
 
% SG cells
 
% change
 
Field 1 WT 6.8 ± 0.4 51.9 ± 16.9 5.3 ± 0.6 1.4 ± 0.6 21.5 ± 8.1 94.2 ± 150.7 
 DT 10.3 ± 0.1**  6.1 ± 0.8 4.2 ± 0.9* 41.7 ± 7.0**  
Field 40
 
WT 5.9 ± 0.7 43.4 ± 19.9 2.9 ± 0.7 3.0 ± 0.2 56.2 ± 9.5 19.9 ± 20.1 
 DT
 
8.5 ± 0.7*
 

 
3.1 ± 0.6
 
5.4 ± 0.6**
 
64.7 ± 5.8
 

 

The number and laminar fate of a cohort of cells that exited the cell cycle on embryonic day 14 and came to reside in the infragranular (IG) or granular–supragranular (SG) layers of cortical fields 1 and 40 at postnatal day P21. The proportion of cells destined to the SG layers (% SG cells) was calculated as a percentage of the total cell number.

*

P < 0.05;

**

P < 0.01; t-test.

These findings indicate that on average in the DT mice an NQ cell produced on E14 in the MCZ has a 94% greater chance of residing in SG layers and a NQ cell from the LCZ a 20% greater chance of residing in the SG layers compared to corresponding cells in the WT (Table 4). We encountered substantial variability among the litters in the relative distribution of the NQ cells, in keeping with the expected variation in relative maturity among the litters (Theiler, 1972; Takahashi et al., 1999). We exploited this by correlating the percentage of NQ cells directed to SG layers in DT versus WT littermate pairs in fields 1 and 40 (Fig. 5A). The broken line (with 45° slope; Fig. 5A) denotes the ‘no effect’ line, representing a hypothetical situation, in which DT and WT littermates have identical values (i.e. p27Kip1 overexpression has no effect). However, the slope of the regression line obtained with the actual data is significantly (t-test, P < 0.01) lower than that of the ‘no effect’ line, indicating that the percentage of cells directed to the SG layers was greater in DT mice compared to the WT littermates. Thus, it is evident that the field 1 values diverge from the ‘no effect’ line to a greater extent than the field 40 values, reflecting differences in the maturational state of these two areas (Miyama et al., 1997). This is consistent with the findings from E14 where NQ and Q were significantly higher in the DT mice in the MCZ (precursor of field 1) and not in the LCZ (precursor of field 40) (Table 2). Thus, overexpression of p27Kip1 from E12 to E14 not only shifted the destination of cells produced on E14 towards more superficial layers of the neocortex but also did so in a developmentally regulated manner. That is, there is progressively smaller effect in more mature regions of the cortex (at the time of p27Kip1 induction) as the transverse neurogenetic gradient shifts to formation of principally SG neurons.

Figure 5.

The regulatory linkage among p27Kip1 overexpression, cell output and cell fate. (A) Regression analysis of the percentage of Q cells directed to granular and supragranular (SG) layers in fields 1 and 40 in wild type (WT; x-axis) and double transgenic (DT; y-axis) littermate pairs. The broken line (with 45° slope) denotes the ‘no effect’ line, representing a hypothetical function, in which DT and WT littermates have identical values (i.e. no effect of p27Kip1 overexpression). The slope of the regression line obtained with the actual data is significantly (t-test, P = 0.01) lower than that of the ‘no effect’ line indicating that the percentage of cells directed to the SG layers was greater in DT mice compared to the WT littermates. Moreover, the field 1 values diverge from the ‘no effect’ line to a greater extent than the field 40 values, which tend to approximate the line closely. (B) The percentage of cells that exited the cell cycle on embryonic day 14 (E14) and that occupied SG layers at postnatal day 21 (percentage SG cells) in the WT and DT mice is plotted as function of the probability of cell cycle exit (Q). The solid line represents data from CD1 strain of mouse obtained in our earlier work (Takahashi et al., 1996a, 1999). The data from the current study are plotted as broken line. Non-linear regression curves are obtained for both the sets of data. Q values were measured in the ventricular zone of the cerebral wall in the medial and lateral cortical zones (MCZ and LCZ, respectively) on E14. The MCZ and LCZ are precursors of fields 1 and 40, respectively. MCZ and LCZ are separated from each other along the transverse neurogenetic gradient such that the LCZ progenitors are ‘in advance’ of the MCZ progenitors with regard to cell cycle kinetics. The correlation between Q and percentage SG cells observed in the WT mice also applies to the DT mice despite the p27Kip1 overexpression-induced increase in Q. Thus, when Q increases, percent SG increases proportionately, both in the MCZ and the LCZ and both in the WT and DT mice. This indicates that the probability of origin of a given laminar neuron class is a function of the probability of cell cycle exit, regardless of position along the transverse neurogenetic gradient. The dotted lines in BE demarcate the boundary between IG and SG layers.

Figure 5.

The regulatory linkage among p27Kip1 overexpression, cell output and cell fate. (A) Regression analysis of the percentage of Q cells directed to granular and supragranular (SG) layers in fields 1 and 40 in wild type (WT; x-axis) and double transgenic (DT; y-axis) littermate pairs. The broken line (with 45° slope) denotes the ‘no effect’ line, representing a hypothetical function, in which DT and WT littermates have identical values (i.e. no effect of p27Kip1 overexpression). The slope of the regression line obtained with the actual data is significantly (t-test, P = 0.01) lower than that of the ‘no effect’ line indicating that the percentage of cells directed to the SG layers was greater in DT mice compared to the WT littermates. Moreover, the field 1 values diverge from the ‘no effect’ line to a greater extent than the field 40 values, which tend to approximate the line closely. (B) The percentage of cells that exited the cell cycle on embryonic day 14 (E14) and that occupied SG layers at postnatal day 21 (percentage SG cells) in the WT and DT mice is plotted as function of the probability of cell cycle exit (Q). The solid line represents data from CD1 strain of mouse obtained in our earlier work (Takahashi et al., 1996a, 1999). The data from the current study are plotted as broken line. Non-linear regression curves are obtained for both the sets of data. Q values were measured in the ventricular zone of the cerebral wall in the medial and lateral cortical zones (MCZ and LCZ, respectively) on E14. The MCZ and LCZ are precursors of fields 1 and 40, respectively. MCZ and LCZ are separated from each other along the transverse neurogenetic gradient such that the LCZ progenitors are ‘in advance’ of the MCZ progenitors with regard to cell cycle kinetics. The correlation between Q and percentage SG cells observed in the WT mice also applies to the DT mice despite the p27Kip1 overexpression-induced increase in Q. Thus, when Q increases, percent SG increases proportionately, both in the MCZ and the LCZ and both in the WT and DT mice. This indicates that the probability of origin of a given laminar neuron class is a function of the probability of cell cycle exit, regardless of position along the transverse neurogenetic gradient. The dotted lines in BE demarcate the boundary between IG and SG layers.

Apoptosis

P4 is the time of maximum apoptotic cell death in the developing mouse neocortex, especially in fields 1 and 40 (Verney et al., 2000). No significant difference was detected between DT and WT littermates in the numerical density of TUNEL+ profiles either in field 1 (mean ± SEM, DT 4.36 × 10−3 ± 1.46 × 10−3, WT 3.66 × 10−3 ± 0.51 × 10−3 cells/unit area, t-test, P = 0.661) or field 40 (mean ± SEM, DT 2.46 × 10−3 ± 0.92 × 10−3, WT 3.38 × 10−3 ± 0.55 × 10−3 cells/unit area, t-test, P = 0.418). There was also no difference in the laminar distribution of TUNEL+ profiles between WT and DT littermates in either cortical field (data not shown).

Discussion

The direct experimental measures of TG1 and Q undertaken here are based on a recently developed method (Hayes and Nowakowski, 2000), which is both more efficient and sensitive than the methods we have used previously. Importantly in the present study we overexpressed p27Kip1 for a 48 h period from E12 to E14, corresponding to the interval through which p27Kip1 expression ascends from its low to high asymptote in the normal mouse (Delalle et al., 1999). As with other developmentally regulated processes, p27Kip1 expression in WT was found in the prior study (Delalle et al., 1999) to reach asymptote in the upgradient LCZ well in advance of its asymptote in the downgradient MCZ. We find that TG1 is not altered by the p27Kip1 overexpression but that Q is increased. These findings are supported by a recent report showing that both TC and TG1 are unchanged in the neocortical VZ of the p27Kip1 null mouse (Goto et al., 2005). Thus, cell cycle kinetic parameters are not altered even by extreme disregulation of p27Kip1 expression, whether the disregulation is either up or down. In a previous report (Mitsuhashi et al., 2001) we interpreted the reduction in BrdU LI following 26 h of p27Kip1 overexpression as indication of prolongation of TG1 without alteration of Q. Our present findings, based on direct measurements of Q and TG1, show that Q and not TG1 is modulated by p27Kip1 overexpression; this corrects our prior misinterpretation of the limited data we had previously.

The capacity of p27Kip1 to increase Q is developmentally regulated, i.e. strongly apparent in the developmentally early MCZ but not detectable in the developmentally late LCZ. The increase in Q is dissociated from the kinetic operation of the cell cycle in that TG1 (and therefore TC) and the overall neuronogenetic interval remain unchanged. This means that the number of cell cycles in the neuronogenetic interval remains fixed at 11 just as in the normal animal. This also means that the path of ascent of Q as a function of cell cycle sequence through the neuronogenetic interval is altered. This path increases non-linearly from 0.0 (before the first neurons are born) to 1.0 with the final cycle when the last cell division occurs (Takahashi et al., 1996a,b; Nowakowski et al., 2002). In the DT animal the rate of ascent with cycle is increased over the interval of p27Kip1 overexpression and must therefore be somewhat slowed subsequently. Correspondingly the ascent of Q in the p27Kip1/– mouse must rise abnormally slowly initially but then accelerate with respect to WT in the terminal cycles (Goto et al., 2005). It is to be noted in this regard that variations in cell cycle parameters, in contrast to variations in Q, will have little effect upon cell production in the neocortical PVE. Thus, output per cycle is an exponential function of Q and total output is simply the cumulative output of all cycles (Takahashi et al., 2000). Variations in cycle duration may affect linearly the output in time but will have no effect upon output per cycle or total output for a series of cycles. As an example in point, in the Emx2–/– mutant there is a profound reduction in neuronal production but cycle durations are unaltered (Mallamaci et al., 2000). Q, by default the affected parameter, has not been measured. Plausibly the same will hold for Pax6–/– and other mutants which are similarly associated with massive limitations of neuronal production (O'Leary and Nakagawa, 2002; Grove and Fukuchi-Shimogori, 2003; Shin et al., 2004).

Finally, we have observed that the effect of p27Kip1 overexpression upon Q is dissociated from post-proliferative histogenetic mechanisms involved in migration and postmigration histogenetic mechanisms of cortical pattern formation and from postmigratory histogenetic cell death. These processes appear to proceed normally in all respects in the E14 DT mouse embryo. Moreover, in confirmation, the stratification and laminar cytoarchitectonic patterns (such as barrel patterns of area 3) are preserved completely in the P21 animal. Recently p27Kip1 has been found to play a role as modulator of cell motility in cultured fibroblasts, a role mediated by RhoA activation (Besson et al., 2004). However, in the present study, p27Kip1 overexpression in the DT mice would have returned to near normal levels by E16, prior to the onset of robust neocortical neuronal migration. Our previous study (Mitsuhashi et al., 2001) showed that p27Kip1 mRNA levels began to rise 6 h after a single dose of dox, reached 300-fold of baseline expression at 12 h and returned to essentially baseline expression levels 48 h after the dox dose. Therefore, it is unlikely that p27Kip1 overexpression affects migration in this experiment model.

p27Kip1, Q and Neuronal Laminar Destination

The size of the 2 h (cells stained blue) cohort arising on E14 and distributed within fields 1 and 40 at P21 is larger in DT than WT animals. We attribute this to an increase in Q, occurring in response to p27 Kip1 overexpression. We note, moreover, that there is also a substantial increase in the proportion of the cohort that is distributed to SG layers in DT P21 animals. Because TG1 and therefore TC are unaltered at E14 in the DT embryos, the cohort in the P21 cortex must have arisen from the same cell cycle in the 11 cycle sequence in WT and DT animals. However, the composition of this cohort has been altered such that an increased proportion of its cells is destined for SG layers in both fields 1 and 40 though to a larger degree in field 1. For reasons cited above, this phenomenon is not an artifact or a disturbance of migration, postmigration mechanisms of laminar assembly or abnormal patterns of cell death. It cannot be simply incidental to the E14 cohort representing an increased fraction of the total numbers of cells formed on E14 and subsequently in that the architectonic features of these layers are normal. The patterning behavior of the SG cells in DT is indistinguishable from those in WT. Thus, it is the IG–SG fate specification profile of the E14 cohort that is changed. In the DT animals, where cells are specified with the characteristics of IG or SG cells, their histogenetic behavior is appropriate to that specification.

Thus, p27Kip1 overexpression is found here to be associated with an alteration in the proportionate laminar destination of cells with respect to the cell cycle number of origin. However, the cardinal insight here is that the proportionate laminar destination of cells is not altered with respect to the value of Q of the cell cycle of origin. In other words, of four possible ‘counting’ methods — embryonic age, cell cycle number, the length of G1, and Q (the probability of exiting the cell cycle) — it is the last that is apparently the determinant for laminar position. The evidence for this assertion is based upon the interpretation of the data presented here in the context of previously published data from the WT CD1 strain mouse (Takahashi et al., 1996a,b, 1999), in which the relationship between Q and the percentage of the cohort that is destined to be SG cells was first determined (percentage SG cells; Fig. 5B). The percentage of SG cells from the E14 cohort and measured at P21 approximate closely a sigmoidal curve (Fig. 5B; solid line; R2 = 0.99). This suggests a tight regulatory linkage between Q and laminar destination. The data from the present study illustrating the effect of p27Kip1 overexpression also approximates closely a sigmoidal curve (R2 = 0.96). This is represented as a broken line since the y-axis values of its left and right tails would be 0% and 100%, respectively (Takahashi et al., 1996a,b, 1999). The curve for the present transgenic mouse line is displaced to the left, presumably reflecting both differences in the strains and in the methods used, which are recognized to be differentially sensitive to leading and trailing edges of the S phase labeled cohort (Hayes and Nowakowski, 2000). Thus, the proportions of neurons fated by specification for SG layer positions under conditions of p27Kip1 overexpression are also correlated to Q. The most striking feature of Figure 5A is the effect that the maturity of the neocortex has on the shift produced by the upregulation of p27Kip1. The less mature cortex is more dramatically affected. This is reasonable because the less mature cortex in WT would be producing fewer cells ‘fated’ to become SG, and, thus, a shift towards SG production in DT would be more dramatic.

These findings imply that mechanisms that regulate Q are coordinate with those that regulate specification of laminar destination and operate before the histogenetic events of laminar assembly and cell differentiation (Fig. 6). The expression of Tis21, an antimitogen, appears to be selectively sensitive to the transition between specification and readiness to exit the cycle. Its profile of expression approximates the measured profile of advance of Q with cell cycle (Iacopetti et al., 1999; Calegari and Huttner, 2003; Haubensak et al., 2004; Kosodo et al., 2004). These investigators propose that over the course of the G1 phase of the cell cycle there is cumulative synthesis of a substance, as yet unidentified, that coordinately increases Q and the expression of Tis21 (Calegari and Huttner, 2003). Our data are in accord with this model in that they identify p27Kip1 as a substance that drives Q upward as its levels increase. Moreover, we also showed that the linkage between Q and laminar destination is driven by p27Kip1 overexpression. For the present it is premature to suggest that p27Kip1 or substances regulating its synthesis are the exclusive determinants of the advance of Q. The linkage of the p27Kip1 drive of Q may be downstream of the proliferative drive originating with the FGFR1 receptor (Shin et al., 2004). It must be upstream of the graded concentrations of transcription factors known to be involved in neuron specification (Grove and Fukuchi-Shimogori, 2003). The actual linkages between Q and regulation of these concentration gradients are as yet unknown.

Figure 6.

Schematic diagram of proliferative model. The diagram schematizes the comparative interrelationship in DT relative to WT mice in expression level of p27Kip1 mRNA (A) to the progression of Q with cell cycle (B) and finally to neuron output from the PVE (C). All three levels are aligned with respect to the 11 cycles of the neuronogenetic sequence as the ordinate, which is the same in WT and DT PVE. Our previous investigations have determined for the WT that the peak level of p27Kip1 expression (Delalle et al., 1999), the ascent of Q through the steady-state value of 0.05 (Takahashi et al., 1996a) and the period of transition from predominantly IG to SG cell output (Takahashi et al., 1999) occurs in the cycle 6–8 interval. For the purposes of the present findings overexpression of p27Kip1 is induced in the interval leading up to cycles 6–7 (gray area enclosed by dashed curve in A). Transcription factors including Emx2 and Pax6, possibly driven by the Notch signaling system, strongly modulate the proliferative process (Grove and Fukuchi-Shimogori, 2003) and may do so via regulating expression of p27Kip1 or other regulators of Q. Induction of p27Kip1 overexpression in turn induces an increase Q (gray area enclosed below dashed curve in B). The increased Q in turn leads to an overall decreased cumulative output that was limited to the output of SG cells (dashed line in C) in these experiments. However, the ratio of SG with respect to IG cells arising in corresponding cell cycles is increased by p27Kip1 overexpression induced increased rate of advance of Q [compare gray (DT) to white (WT) diamond in C].

Figure 6.

Schematic diagram of proliferative model. The diagram schematizes the comparative interrelationship in DT relative to WT mice in expression level of p27Kip1 mRNA (A) to the progression of Q with cell cycle (B) and finally to neuron output from the PVE (C). All three levels are aligned with respect to the 11 cycles of the neuronogenetic sequence as the ordinate, which is the same in WT and DT PVE. Our previous investigations have determined for the WT that the peak level of p27Kip1 expression (Delalle et al., 1999), the ascent of Q through the steady-state value of 0.05 (Takahashi et al., 1996a) and the period of transition from predominantly IG to SG cell output (Takahashi et al., 1999) occurs in the cycle 6–8 interval. For the purposes of the present findings overexpression of p27Kip1 is induced in the interval leading up to cycles 6–7 (gray area enclosed by dashed curve in A). Transcription factors including Emx2 and Pax6, possibly driven by the Notch signaling system, strongly modulate the proliferative process (Grove and Fukuchi-Shimogori, 2003) and may do so via regulating expression of p27Kip1 or other regulators of Q. Induction of p27Kip1 overexpression in turn induces an increase Q (gray area enclosed below dashed curve in B). The increased Q in turn leads to an overall decreased cumulative output that was limited to the output of SG cells (dashed line in C) in these experiments. However, the ratio of SG with respect to IG cells arising in corresponding cell cycles is increased by p27Kip1 overexpression induced increased rate of advance of Q [compare gray (DT) to white (WT) diamond in C].

The capacity of p27Kip1 overexpression to drive this linkage appears to be developmentally regulated. Our earlier studies (Delalle et al., 1999) determined that p27Kip1 mRNA expression in the PVE maximizes in the LCZ earlier than in the MCZ on E14 as Q approaches 0.5 with cell cycle number 8. Expression then declines as Q continues its ascent to 1.0 over the final three cell cycles (Delalle et al., 1999) (Fig. 6A). Here at E14 we find a prominent effect of p27Kip1 upon Q in the MCZ but not in the LCZ. We hypothesize that this could be because other cdk inhibitors take over the control of the restriction point and therefore the control of Q late in the course of neuronogenesis. Candidates include, inter alia, both the inhibitors of CDKs4/6 (Ink4 proteins: p16Ink4a, p15Ink4b, p18Ink4c) and the inhibitors of cyclin E-Cdk2 complex (p21Cip1 and p57Kip1 in addition to p27Kip1), all of which are variably expressed in regulated fashion in proliferative populations of the CNS (Watanabe et al., 1998; Sherr and Roberts, 1999; Zindy et al., 1999). We suggest this occurs as Q increases above the value of 0.5 and only the last formed cells destined for supragranular layers are being formed. Thus, beyond this stage, p27Kip1 may be replaced by other cdk inhibitors as regulators of Q in the final phase of ascent of this parameter.

The present observation linking neocortical laminar fate specification to p27Kip1 has precedent in other systems. For example, under optimum growth conditions in vitro, O-2A oligodendrocyte progenitor cells withdraw from cell cycle and differentiate into oligodendroglial cells after a set number of cycles. As cycles proceed there are rising levels of p27Kip1 and both cycle withdrawal and differentiation are triggered after p27Kip1 level reaches a plateau. It has been postulated that this reflects a cycle counting mechanism mediated at a threshold level of p27Kip1 (Durand et al., 1997; Durand and Raff, 2000). It was not resolved in these studies whether the role of p27Kip1 is only to trigger cycle arrest or is also directly involved in mechanisms leading to differentiation (Galderisi et al., 2003). This distinction has been approached by experiments with CG-4 cells obtained from p27Kip1 null animals. Although CG-4 cells lack p27Kip1, a small fraction arrest and differentiate into normal astrocytes under deprived conditions of in vitro culture. These observations suggested that p27Kip1 primarily induces cell cycle arrest and only indirectly through cycle arrest does differentiation proceed (Tikoo et al., 1997; Casaccia-Bonnefil et al., 1999). In Xenopus, however, the homologous protein, p27Xic1, may be more directly required for specification of neural fate (Ohnuma et al., 1999; Vernon and Philpott, 2003). Whereas the protein operates as both an inhibitor of cycle progression and in cell fate specification, the active domains of the protein serving the two functions are only partially overlapping (Ohnuma et al., 1999). These observations place p27Kip1, and perhaps related cdks, within a linkage that coordinates the mechanisms regulatory to specification and others regulatory to the proliferative process. The linkage of p27Kip1 overexpression, the associated advance in Q and modulation of the laminar fate are consistent with these observations from other experimental systems.

p27Kip1 and the Proliferative Model

Our proliferative model for neocortical neuronogenesis was formalized earlier (Caviness et al., 2000, 2003; Nowakowski et al., 2002) from observations based upon kinetic parameters and Q in normal embryos and subsequently confirmed and further validated in Ts16 trisomy (Haydar et al., 2000a) and extended in the p27Kip1 overexpression model. l. This model (Figs 1A, 6AC) integrates the behavior of these parameters across the entire neuronogenetic interval in mouse. In brief, the founder population and progeny on average execute 11 integer cell cycles before exhaustion of proliferative activity in the PVE. The fractional advance of Q with each cell cycle determines the total number of cycles (11 in the normal mouse neocortex).

Pertinent to the present findings, the model predicts that should Q be increased at a rate ahead of its normal progression, there would be an early increase in cell cycle output (Caviness et al., 2003) (Fig. 6A,B). However, this would be at the expense of the size of the overall proliferative pool so that the total number of later formed neurons would be reduced (Fig. 6C). The effect should be more salient in down-gradient regions, i.e. more salient in less mature regions of the cerebral wall, than in up-gradient, more mature regions. These predictions are largely confirmed here where the rate of progression of Q with cell cycle has been increased by overexpression of p27Kip1 in the early part of the neuronogenetic interval. Thus, from the perspective of the size of the E14 labeled cohort counted in the P21 cortex, its size is increased in DT and to a greater degree in down-gradient field 1 than in upgradient field 40. However, in contrast to the size of this cohort directed to the cortex at E14, the total number of cells that will be directed to the cortex over the full remaining neuronogenetic interval is reduced (Table 3). The effect upon total cell numbers is detected only with respect to SG layers in field 1, which are largely being formed on and after E14. An effect is not detectable in field 40, which has largely completed its cycle of neuronogenesis by E14 (by which time a large proportion of cells has already been designated to SG layers).

Overall these studies illustrate the predominant importance of the parameter Q as determinant of the rate and number of neurons arising in the course of cell proliferation in the PVE. In principle, an accelerated progression of Q with successive cycles predicts an initial increased output per cycle, which then causes a reduction in the size of the proliferative population. This in turn leads to an ultimate reduction in total output (Caviness et al., 2003). In contrast, a retardation in the progression of Q with successive cycles predicts the opposite (Caviness et al., 2003). Here Q was increased by only ∼10% at what was probably the point of maximum effect on E14 yet there was substantial reduction in the cells delivered to the cortex. In the p27Kip1–/– (Fero et al., 1996; Kiyokawa et al., 1996; Nakayama et al., 1996; Goto et al., 2004) and the Ts16 mice (Haydar et al., 1996, 2000a) predictions of the opposite condition of retarded progression of Q with cycle are confirmed. In the p27Kip1–/–all other proliferative parameters, including founder population, cycle duration and growth fraction, are normal and the cortex is increased in cell number (Goto et al., 2005). This consequence of the knockout indicates that p27Kip1 exerts a regulatory effect upon Q that is independent of that of other CDK inhibitors, at least in its progression through the earlier cycles of the neuronogenetic interval. Whereas our model scales Q to neuronal production, available observations relating to the coordinate interactions of CDK inhibitors do not yet allow a scaling of the independent contribution of p27Kip1 to the progression of Q.

In Ts16 the effect of changes in Q are offset by a smaller founder population and reduced growth fraction. The number of cells delivered through the early cycles is substantially reduced but the number largely normalizes with the later cycles (Haydar et al., 1996, 2000a). Precocious acceleration rather than retardation, in the progression of Q with cell cycle in particular may underlie a wide range of genetically determined human disorders classified as microcephaly vera, as well as a host of other microcephalies that bear no indication of actively destructive pathological process (Mochida and Walsh, 2001). In particular, relative attenuation of SG layers, as would be predicted from this model, is characteristic of certain forms of human microcephaly vera (Urich, 1976).

We thank Genevieve Stein-O'Brien for the graphics work for Figure 6.This work was supported by US Public Health Service Grants NS12005 (V.S.C.), NS43246 (P.G.B.). T. Tarui was supported by a fellowship from the William Randolph Hearst Foundation.

References

Bayer SA, Altman J (
1991
) Neocortical development. New York: Raven Press.
Beaulieu C (
1993
) Numerical data on neocortical neurons in adult rat, with special reference to the GABA population.
Brain Res
 
609
:
284
–292.
Besson A, Gurian-West M, Schmidt A, Hall A, Roberts JM (
2004
) p27Kip1 modulates cell migration through the regulation of RhoA activation.
Genes Dev
 
18
:
862
–876.
Boulder Committee (
1970
) Embryonic vertebrate nervous system: revised terminology.
Anat Rec
 
166
:
257
–262.
Cai L, Hayes NL, Takahashi T, Caviness VS Jr, Nowakowski RS (
2002
) Size distribution of retrovirally marked lineages matches prediction from population measurements of cell cycle behavior.
J Neurosci Res
 
69
:
731
–744.
Calegari F, Huttner WB (
2003
) An inhibition of cyclin-dependent kinases that lengthens, but does not arrest, neuroepithelial cell cycle induces premature neurogenesis.
J Cell Sci
 
116
:
4947
–4955.
Casaccia-Bonnefil P, Hardy RJ, Teng KK, Levine JM, Koff A, Chao MV (
1999
) Loss of p27Kip1 function results in increased proliferative capacity of oligodendrocyte progenitors but unaltered timing of differentiation.
Development
 
126
:
4027
–4037.
Caviness VS (
1975
) Architectonic map of neocortex of the normal mouse.
J Comp Neurol
 
164
:
247
–263.
Caviness VS (
1982
) Neocortical histogenesis in normal and reeler mice: a developmental study based upon [3H]thymidine autoradiography.
Dev Brain Res
 
4
:
293
–302.
Caviness V, Takahashi T, Nowakowski R (
1995
) Numbers, time and neocortical neuronogenesis: a general developmental and evolutionary model.
Trends Neurosci
 
18
:
379
–383.
Caviness V, Takahashi T, Nowakowski R (
2000
) Neuronogenesis and the early events of neocortical histogenesis. In: Development of the neocortex (Goffinet A, Rakic P, eds), pp. 107–143. Berlin: Springer Verlag.
Caviness VS, Goto T, Tarui T, Takahashi T, Bhide PG, Nowakowski RS (
2003
) Cell output, cell cycle duration and neuronal specification: a model of integrated mechanisms of the neocortical proliferative process.
Cereb Cortex
 
13
:
592
–598.
Delalle I, Takahashi T, Nowakowski RS, Tsai LH, Caviness VS Jr (
1999
) Cyclin E-p27 opposition and regulation of the G1 phase of the cell cycle in the murine neocortical PVE: a quantitative analysis of mRNA in situ hybridization.
Cereb Cortex
 
9
:
824
–832.
Durand B, Raff M (
2000
) A cell-intrinsic timer that operates during oligodendrocyte development.
Bioessays
 
22
:
64
–71.
Durand B, Gao FB, Raff M (
1997
) Accumulation of the cyclin-dependent kinase inhibitor p27/Kip1 and the timing of oligodendrocyte differentiation.
EMBO J
 
16
:
306
–317.
Durand B, Fero ML, Roberts JM, Raff MC (
1998
) p27Kip1 alters the response of cells to mitogen and is part of a cell-intrinsic timer that arrests the cell cycle and initiates differentiation.
Curr Biol
 
8
:
431
–440.
Fero M, Rivkin M, Tasch M, Porter P, Carow C, Firpo E, Polyak K, Tsai L-H, Broudy V, Perlmutter R, Kaushansky K, Roberts J (
1996
) A syndrome of multiorgan hyperplasia with features of gigantism, tumorigenesis, and female sterility in p27Kip1-deficient mice.
Cell
 
85
:
733
–744.
Galderisi U, Jori FP, Giordano A (
2003
) Cell cycle regulation and neural differentiation.
Oncogene
 
22
:
5208
–5219.
Goto T, Mitsuhashi T, Takahashi T (
2005
) Altered patterns of neuron production in the p27Kip1 knockout mouse. Dev Neurosci (in press).
Grove EA, Fukuchi-Shimogori T (
2003
) Generating the cerebral cortical area map.
Annu Rev Neurosci
 
26
:
355
–380.
Hammond V, Tsai LH, Tan SS (
2004
) Control of cortical neuron migration and layering: cell and non cell-autonomous effects of p35.
J Neurosci
 
24
:
576
–587.
Haubensak W, Attardo A, Denk W, Huttner WB (
2004
) Neurons arise in the basal neuroepithelium of the early mammalian telencephalon: a major site of neurogenesis.
Proc Natl Acad Sci USA
 
101
:
3196
–3201.
Haydar TF, Blue ME, Molliver ME, Krueger BK, Yarowsky PJ (
1996
) Consequences of trisomy 16 for mouse brain development: corticogenesis in a model of Down syndrome.
J Neurosci
 
16
:
6175
–6182.
Haydar TF, Nowakowski RS, Yarowsky PJ, Krueger BK (
2000
a) Role of founder cell deficit and delayed neuronogenesis in microencephaly of the trisomy 16 mouse.
J Neurosci
 
20
:
4156
–4164.
Haydar TF, Wang F, Schwartz ML, Rakic P (
2000
b) Differential modulation of proliferation in the neocortical ventricular and subventricular zones.
J Neurosci
 
20
:
5764
–5774.
Hayes N, Nowakowski R (
2000
) Exploiting the dynamics of S-phase tracers in developing brain: interkinetic nuclear migration for cells entering vs leaving the S-phase.
Dev Neurosci
 
22
:
44
–55.
His W (
1904
) Die Entwicklung des Menschlichen Gehirns wahrend der ersten Monate. Leipzig: von S. Hirzel.
Iacopetti P, Michelini M, Stuckmann I, Oback B, Aaku-Saraste E, Huttner WB (
1999
) Expression of the antiproliferative gene TIS21 at the onset of neurogenesis identifies single neuroepithelial cells that switch from proliferative to neuron-generating division.
Proc Natl Acad Sci USA
 
96
:
4639
–4644.
Kiyokawa H, Kineman R, Manova-Todorava K, Soares V, Hoffman E, Ono M, Khanam D, Hayday A, Frohman L, Koff A (
1996
) Enhanced growth of mice lacking the cyclin-dependent kinase inhibitor function of p27Kip1.
Cell
 
85
:
721
–732.
Kosodo Y, Roper K, Haubensak W, Marzesco AM, Corbeil D, Huttner WB (
2004
) Asymmetric distribution of the apical plasma membrane during neurogenic divisions of mammalian neuroepithelial cells.
EMBO J
 
23
:
2314
–2324.
Livesey FJ, Cepko CL (
2001
) Vertebrate neural cell-fate determination: lessons from the retina.
Nat Rev Neurosci
 
2
:
109
–118.
Mallamaci A, Muzio L, Chan CH, Parnavelas J, Boncinelli E (
2000
) Area identity shifts in the early cerebral cortex of Emx2–/– mutant mice.
Nat Neurosci
 
3
:
679
–686.
McConnell SK (
1989
) The determination of neuronal fate in the cerebral cortex.
Trends Neurosci
 
12
:
342
–349.
Mitsuhashi T, Aoki Y, Eksioglu YZ, Takahashi T, Bhide PG, Reeves SA, Caviness VS, Jr. (
2001
) Overexpression of p27Kip1 lengthens the G1 phase in a mouse model that targets inducible gene expression to central nervous system progenitor cells.
Proc Natl Acad Sci USA
 
98
:
6435
–6440.
Miyama S, Takahashi T, Nowakowski RS, Caviness V (
1997
) A gradient in the duration of the G1 phase in the murine neocortical proliferative epithelium.
Cereb Cortex
 
7
:
678
–689.
Mochida GH, Walsh CA (
2001
) Molecular genetics of human microcephaly.
Curr Opin Neurol
 
14
:
151
–156.
Nakayama K, Ishida N, Shirane M, Inomata A, Inoue T, Shishido N, Horii I, Loh D, Nakayama K-I (
1996
) Mice lacking p27Kip1 display increased body size, multiple organ hyperplasia, retinal dysplasia, and pituitary tumors.
Cell
 
85
:
707
–720.
Nowakowski R, Lewin S, Miller M (
1989
) Bromodeoxyuridine immunohistochemical determination of the lengths of the cell cycle and the DNA-synthetic phase for an anatomically defined population.
J Neurocytol
 
18
:
311
–318.
Nowakowski R, Caviness V Jr, Takahash IT, Hayes N (
2002
) Population dynamics during cell proliferation and neuronogenesis in the developing murine neocortex. In: Cortical development: from specification to differentiation (results and problems in cell differentiation) (Hohmann C, ed.), pp. 1–22. New York: Springer-Verlag.
O'Leary DD, Nakagawa Y (
2002
) Patterning centers, regulatory genes and extrinsic mechanisms controlling arealization of the neocortex.
Curr Opin Neurobiol
 
12
:
14
–25.
Ohnuma S, Philpott A, Wang K, Holt CE, Harris WA (
1999
) p27Xic1, a Cdk inhibitor, promotes the determination of glial cells in Xenopus retina.
Cell
 
99
:
499
–510.
Ohnuma S, Philpott A, Harris WA (
2001
) Cell cycle and cell fate in the nervous system.
Curr Opin Neurobiol
 
11
:
66
–73.
Ohnuma S, Hopper S, Wang KC, Philpott A, Harris WA (
2002
) Co-ordinating retinal histogenesis: early cell cycle exit enhances early cell fate determination in the Xenopus retina.
Development
 
129
:
2435
–2446.
Ohshima T, Ogawa M, Veeranna, Hirasawa M, Longenecker G, Ishiguro K, Pant HC, Brady RO, Kulkarni AB, Mikoshiba K (
2001
) Synergistic contributions of cyclin-dependant kinase 5/p35 and Reelin/Dab1 to the positioning of cortical neurons in the developing mouse brain.
Proc Natl Acad Sci USA
 
98
:
2764
–2769.
Rakic P (
1976
) Differences in the time of origin and in eventual distribution of neurons in areas 17 and 18 of visual cortex in Rhesus monkey.
Exp Brain Res Suppl
 
1
:
244
–248.
Ren JQ, Aika Y, Heizmann CW, Kosaka T (
1992
) Quantitative analysis of neurons and glial cells in the rat somatosensory cortex, with special reference to GABAergic neurons and parvalbumin-containing neurons.
Exp Brain Res
 
92
:
1
–14.
Sauer FC (
1935
) Mitosis in the neural tube.
J Comp Neurol
 
62
:
377
–405.
Sauer FC (
1936
) The interkinetic migration of embryonic epithelial nuclei.
J Morphol
 
60
:
1
–11.
Sherr CJ, Roberts JM (
1999
) CDK inhibitors: positive and negative regulators of G1-phase progression.
Genes Dev
 
13
:
1501
–1512.
Shin DM, Korada S, Raballo R, Shashikant CS, Simeone A, Taylor JR, Vaccarino F (
2004
) Loss of glutamatergic pyramidal neurons in frontal and temporal cortex resulting from attenuation of FGFR1 signaling is associated with spontaneous hyperactivity in mice.
J Neurosci
 
24
:
2247
–2258.
Sidman RL, Rakic P (
1973
) Neuronal migration, with special reference to developing human brain: a review.
Brain res
 
62
:
1
–35.
Takahashi T, Nowakowski RS, Caviness VS Jr (
1994
) Mode of cell proliferation in the developing mouse neocortex.
Proc Natl Acad Sci USA
 
91
:
375
–379.
Takahashi T, Nowakowski R, Caviness V (
1995
) The cell cycle of the pseudostratified ventricular epithelium of the murine cerebral wall.
J Neurosci
 
15
:
6046
–6057.
Takahashi T, Nowakowski R, Caviness V (
1996
a) The leaving or Q fraction of the murine cerebral proliferative epithelium: a general computational model of neocortical neuronogenesis.
J Neurosci
 
16
:
6183
–6196.
Takahashi T, Nowakowski RS, Caviness VS Jr (
1996
b) Interkinetic and migratory behavior of a cohort of neocortical neurons arising in the early embryonic murine cerebral wall.
J Neurosci
 
16
:
5762
–5776.
Takahashi T, Goto T, Miyama S, Nowakowski RS, Caviness VS Jr (
1999
) Sequence of neuron origin and neocortical laminar fate: relation to cell cycle of origin in the developing murine cerebral wall.
J Neurosci
 
19
:
10357
–10371.
Takahashi T, Nowakowski R, Caviness V (
2000
) Neocortical neuronogenesis: regulation, control points and a strategy of structural variation. In: The handbook of developmental cognitive neuroscience (Nelson C, Luciana M, eds), pp. 3–23. Cambridge, MA: MIT Press.
Theiler K (
1972
) The house mouse. Development and normal stages from fertilization to 4 weeks of age. Berlin: Springer Verlag.
Tikoo R, Casaccia-Bonnefil P, Chao M, Koff A (
1997
) Changes in cyclin-dependent kinase 2 and p27kip1 accompany glial cell differentiation of central glia-4 cells.
J Biol Chem
 
272
:
442
–447.
Urich H (
1976
) Malformations of the nervous system, perinatal damage and related conditions in early life. In: Greenfield's neuropathology (Blackwood W, Corsellis J, eds), pp. 361–469. London: Edward Arnold.
Verney C, Takahashi T, Bhide P, Nowakowski RS, Caviness V (
2000
) Independent controls for neocortical neuron production and histogenetic cell death.
Dev Neurosci
 
22
:
125
–138.
Vernon AE, Philpott A (
2003
) A single cdk inhibitor, p27Xic1, functions beyond cell cycle regulation to promote muscle differentiation in Xenopus.
Development
 
130
:
71
–83.
Vernon AE, Devine C, Philpott A (
2003
) The cdk inhibitor p27Xic1 is required for differentiation of primary neurones in Xenopus.
Development
 
130
:
85
–92.
Walsh CA (
2000
) Genetics of neuronal migration in the cerebral cortex.
Ment Retard Dev Disabil Res Rev
 
6
:
34
–40.
Watanabe G, Pena P, Shambaugh GE 3rd, Haines GK 3rd, Pestell RG (
1998
) Regulation of cyclin dependent kinase inhibitor proteins during neonatal cerebella development.
Brain Res Dev Brain Res
 
108
:
77
–87.
Zindy F, Cunningham JJ, Sherr CJ, Jogal S, Smeyne RJ, Roussel MF (
1999
) Postnatal neuronal proliferation in mice lacking Ink4d and Kip1 inhibitors of cyclin-dependent kinases.
Proc Natl Acad Sci USA
 
96
:
13462
–13467.

Author notes

1Neurology, Massachusetts General Hospital, Boston, MA, USA, 2Pediatrics, Keio University School of Medicine, Tokyo, Japan and 3Neuroscience & Cell Biology, UMDNJ-RWJ Medical School, Piscataway, NJ, USA