Brca1 is required for embryonic development of the mouse cerebral cortex to normal size by preventing apoptosis of early neural progenitors

The cerebral cortex, which underlies higher brain functions, has undergone a large expansion in size during mammalian evolution, most notably in the primate lineage (Rakic, 1988; Caviness and Takahashi, 1995; Northcutt and Kaas, 1995; Rakic, 1995). Although many intrinsic and extrinsic factors may influence cortical size and cytoarchitecture, such as patterns of neuronal migration(Letinic et al., 2002; Kriegstein and Noctor, 2004; Bystron et al., 2006), thalamic afferents (Windrem and Finlay,1991; Dehay et al.,2001) and the diversification of subventricular zone neural progenitors (Smart et al.,2002; Haubensak et al.,2004; Miyata et al.,2004; Noctor et al.,2004; Fish et al.,2008), an increase in neuron number during brain development and evolution is ultimately controlled by the number and modes of division of neural progenitors in the embryonic ventricular and subventricular zones(Götz and Huttner, 2005; Kriegstein et al., 2006; Fish et al., 2008).

According to the radial unit hypothesis, the expansion of the neocortex can be explained by increased proliferation of ventricular zone neural progenitors early in development, which dramatically affects the resulting surface area of the neocortex and the number of radially organized columns it contains(Rakic, 1988; Rakic, 1995). On a progenitor cell-intrinsic level, features such as apical-basal polarity, mitotic spindle orientation and cell cycle length are involved in shifting progenitors from proliferation to differentiation, and thus influence cortical size(Bond and Woods, 2006; Buchman and Tsai, 2007; Dehay and Kennedy, 2007; Farkas and Huttner, 2008; Fish et al., 2008).

Programmed cell death, or apoptosis, is prominent in neural progenitors(Thomaidou et al., 1997) and appears to play an important role in the development of the cerebral cortex as manipulation of this process can greatly alter cortical size(Haydar et al., 1999; Rakic, 2005). Loss of progenitor apoptosis by knockout of caspase 3(Kuida et al., 1996) or caspase 9 (Kuida et al., 1998)leads to massive overgrowth and folding of the neocortex. Constitutive activation of the Notch signaling pathway in ventricular zone progenitors induces apoptosis, resulting in a reduction in cortical size(Yang et al., 2004). Ephrin A/EphA signaling has also been shown to influence apoptosis and cortical size,whereby gain of EphA receptor function leads to increased apoptosis and a reduction in cortical size, and loss of EphA function leads to a reduction in apoptosis and overgrowth of the neocortex(Depaepe et al., 2005). The cortical surface area reduction seen upon increased Notch and Ephrin A/EphA signaling, and the dramatic cortical surface area increase and convolutions seen in the caspase-deficient mice, are both in accordance with the radial unit hypothesis, which predicts that the loss or increase of early progenitors owing to an increase or decrease, respectively, in apoptosis strongly influences cortical size (Haydar et al.,1999; Rakic,2005). These observations implicate apoptosis as a developmental,and possibly evolutionary, regulator of cortical size(Haydar et al., 1999).

BRCA1, mutations in which predispose women to breast cancer(Miki et al., 1994; Chen et al., 1996; Ruffner and Verma, 1997; Turner et al., 2004), has been implicated in numerous cellular processes, including DNA damage repair(Scully et al., 1997a; Venkitaraman, 2001; Deng and Wang, 2003), cell cycle checkpoints (Wang et al.,2004), transcriptional control(Scully et al., 1997b; Starita et al., 2005),centrosome duplication (Hsu and White,1998; Xu et al.,1999b; Starita et al.,2004) and mitotic spindle assembly(Joukov et al., 2006). Furthermore, Brca1 removal causes increased apoptosis in mammary glands (Xu et al., 1999a),epidermis (Berton et al., 2003)and neuroepithelial cells (Gowen et al.,1996; Xu et al.,2001). Brca1 is widely expressed in proliferating tissues and is also expressed in the ventricular neuroepithelium of the neural tube(Korhonen et al., 2003).

The relevance of apoptosis in cortical development(Haydar et al., 1999), the increased apoptosis of neuroepithelial cells observed in systemic Brca1 knockouts (Gowen et al.,1996; Xu et al.,2001), and the expression pattern of Brca1 in neural progenitors, when considered together, raise the intriguing possibility that Brca1 might have a role in the development of the cerebral cortex. Moreover, the apparent link between BRCA1 and the primary microcephaly genes ASPM (Bond et al., 2002; Zhong et al.,2005) and MCPH1 (Xu et al., 2004; Lin et al.,2005), whereby knockdown of either microcephaly gene leads to a decrease in BRCA1 levels, also leads to the question of whether BRCA1has a role in brain size regulation. In this context, it is interesting to note that BRCA1 has undergone positive selection in the primate lineage (Huttley et al.,2000), raising the possibility that it might have been selected owing to a function in brain development. Here, we have investigated this issue by conditional ablation of Brca1 in cortical progenitors of the dorsal telencephalon during early brain development in the mouse embryo.

Mouse lines were maintained in strict pathogen-free conditions in the animal facility of the Max Planck Institute for Molecular Cell Biology and Genetics, Dresden, Germany. All experiments were performed in accordance with German animal welfare legislation.

A conditional Brca1 knockout mouse line(Xu et al., 1999a) obtained from the United States National Cancer Institute, Mouse Models of Human Cancers Consortium, Mouse Repository (strain number 01XC8; strain nomenclature: STOCK Brca1^tm2Cxd) was used for this study. The Emx1-Cre mouse line was used as a deleter(Iwasato et al., 2000). In all experiments, mice with the genotype Brca1^loxP/loxPEmx1-Cre^-/- were used as control, and mice with the genotype Brca1^loxP/loxP Emx1-Cre^+/-were used as the conditional knockout. The Z/EG reporter line(Novak et al., 2000) was used to test the specificity of Emx1-Cre. Mouse lines and embryos were genotyped for the various alleles by PCR following previously published protocols (Xu et al., 1999a; Iwasato et al., 2000; Novak et al., 2000). The day of the vaginal plug was defined as embryonic day (E) 0.5, and the day of birth was defined as postnatal day (P) 0.5.

In situ hybridization for Brca1 was performed using a digoxigenin-labeled cRNA antisense probe against exon 11 on 10 μm cryosections by standard methods. Three antisense probes, corresponding to nt 77-948, nt 1008-1912 and nt 2010-2981 of Brca1 exon 11, were mixed in equal amounts to a final concentration of ∼400 ng/ml for the hybridization.

Embryos (E11.5-16.5) and postnatal mice (P4.5, P8.5) were dissected in ice-cold PBS and fixed overnight at 4°C with 4% paraformaldehyde in 120 mM phosphate buffer pH 7.4. For cryosectioning, the samples were further equilibrated at 4°C in 30% sucrose in PBS and then embedded in Tissue-Tek O.C.T. compound (Sakura Finetek) and stored at -20°C. All cryostat sections were cut at 10 μm. For Vibratome sectioning, the fixed samples were embedded in 4% agarose in PBS and cut at 70 μm for E13.5 heads and at 50 μm for E16.5 and postnatal brains. Vibratome sections were collected and stored in PBS until further processing.

Nissl staining was performed using cresyl violet acetate (Sigma) in 40 mM sodium acetate buffer on P4.5 brain Vibratome sections attached to Superfrost Plus microscope slides (Thermo Scientific). Immunofluorescence staining was performed by permeabilizing with 0.3% Triton X-100 in PBS for 30 minutes,quenching with 10 mM NH₄Cl in PBS for 30 minutes, followed by blocking, primary and secondary antibody incubations, and washing in a solution of 0.2% gelatin, 300 mM NaCl and 0.3% Triton X-100 in PBS. For Vibratome sections, the primary antibody was incubated for 2 days at 4°C and the secondary antibody for 2 hours at room temperature. For cryosections,the primary antibody was incubated for 2 hours at room temperature and the secondary antibody for 1 hour at room temperature. All cryosection immunofluorescence staining was performed after antigen retrieval by boiling in a microwave oven in Target Retrieval Solution Citrate pH6 (Dako). All immunofluorescence stainings were co-stained with DAPI (Sigma).

The following primary antibodies were used: Brn1 (sc-6028, Santa Cruz),FoxP2 (ab16046, Abcam), Brn2 (sc-6029, Santa Cruz), Tbr1 (AB9616, Millipore),calretinin (7699/4, Swant), Tuj1 (unconjugated MMS-435P, Alexa Fluor 488-conjugated A488-435L, Covance), active caspase 3 (C8487, Sigma), Pax6(PRB-278P, Covance), nestin (ab6142, Abcam), BrdU (ab6326, Abcam),phosphohistone H3 (06-570, Millipore), Tbr2 (ab23345, Abcam), p53 (sc-6243,Santa Cruz) and PCNA (PC10, Dako).

Terminal deoxynucleotidyl transferase dUTP nick-end labeling (TUNEL) was performed on cryosections after antigen retrieval and immunofluorescence staining, using the In Situ Cell Death Detection Kit, TMR red (Roche). EGFP in Emx1-Cre^+/- Z/EG E16.5 brains was detected by intrinsic fluorescence in 50 μm Vibratome sections incubated overnight at 4°C with DAPI in 0.3% Triton X-100 in PBS.

All fluorescence images were acquired by confocal microscopy (Zeiss Axiovert 200M LSM 510). Bright-field images of Nissl staining and in situ hybridization were acquired with an Olympus BX61 microscope fitted with a CCD camera.

BrdU labeling was carried out by intraperitoneal injection of 1.2 mg BrdU(Sigma) in 120 μl PBS into pregnant females 11.5 days post-coitum. Mice were sacrificed 1, 2, 4 or 6 hours after injection, and embryos were processed for cryosectioning as described above. For BrdU immunofluorescence in combination with TUNEL staining, cryosections were subjected to antigen retrieval as described above, then permeabilized with 0.3% Triton X-100 in PBS for 30 minutes, quenched with 10 mM NH₄Cl in PBS for 30 minutes,and then the TUNEL reaction was performed. Cryosections were then subjected to a 30-minute 2 M HCl treatment at 37°C, followed by blocking, primary and secondary antibody incubations, and washing in 10% horse serum in PBS.

Quantification of DAPI, TUNEL, BrdU, phosphohistone H3 and Tbr2 stainings was performed on 200-μm-wide fields with the ventricular surface horizontal. Quantification of Brn1, FoxP2, Brn2 and Tbr1 stainings was performed on 300-μm-wide fields with the pial surface horizontal. Brn1/FoxP2 double-immunofluorescence images were used to define single-positive (Brn1⁺ or FoxP2⁺) and double-positive(Brn1⁺ FoxP2⁺) cells and to determine their distribution across the cortical wall. The ventricular surface length was measured by tracing the apical surface from the most dorsal point down to the attachment site of the ganglionic eminence using ImageJ (NIH). For each anatomical location, the mean was taken from the left and right ventricle, and then the mean was taken from a total of three embryos each for the control and conditional knockout. For all quantifications, the embryos and brains were taken from at least two litters, except for those shown in Fig. 3, which were from one litter. All embryonic cell counting was performed on images from the dorsal telencephalon in regions represented by the white box in Fig. 3A′. P-values were calculated by Student's t-test.

To study the role of Brca1 in the development of the cerebral cortex, Brca1 was selectively ablated in the mouse embryonic dorsal telencephalon (the future neocortex, hippocampus and olfactory bulb) by crossing a mouse line carrying a Brca1 allele in which exon 11 is flanked by loxP sites (Xu et al.,1999a) with an Emx1-Cre knock-in line(Iwasato et al., 2000). In the resulting Brca1 conditional exon 11-deleted mice, the Brca1 protein lacks ∼60% of its sequence, including various domains crucial for its function (Xu et al., 1999a). Recombination by Emx1-Cre in the dorsal telencephalon is known to start at E9.5 (Iwasato et al.,2000) and hence occurs in neuroepithelial cells, the primary neural progenitors of the neocortex. This enabled us to specifically study the function of Brca1 in these progenitors and their progeny during cortical development.

In the experiments described below, mice homozygous for the floxed Brca1 allele but lacking an Emx1-Cre allele(Brca1^loxP/loxP Emx1-Cre^-/-) were used as the control, and mice homozygous for the floxed Brca1 allele but carrying an Emx1-Cre allele (Brca1^loxP/loxPEmx1-Cre^+/-) were used as the conditional knockout(cKO).

In situ hybridization for Brca1 mRNA on E11.5 cryosections using an antisense probe against exon 11 showed that in the E11.5 control telencephalon, Brca1 is expressed in the ventricular zone (VZ) but not in the neuronal layers (Fig. 1A, left), as previously reported(Korhonen et al., 2003). Brca1 mRNA was predominantly located in the basal (abventricular)half of the VZ, which is likely to reflect the specific expression of Brca1 in late G1, S and G2 phase(Gudas et al., 1996; Vaughn et al., 1996). Upon conditional ablation by Emx1-Cre, Brca1 mRNA was not detected in the E11.5 dorsal telencephalon but was still expressed in the ganglionic eminences of the ventral telencephalon(Fig. 1A, right). This demonstrated the regional specificity of Brca1 ablation, consistent with specific expression of Emx1-Cre in the dorsal telencephalon(Iwasato et al., 2000), which we confirmed for E16.5 embryos by crossing to the Z/EG reporter line(Novak et al., 2000), in which EGFP is overexpressed upon loxP recombination(Fig. 1B).

To examine the effects of conditional Brca1 ablation on the neocortex, we first analyzed the size and gross morphology of the P4.5 brain. Compared with the control, Brca1-ablated brains showed a reduction in size specific to the cerebral cortex (Fig. 1C). The whole-brain weight of Brca1-ablated mice was also significantly reduced (Fig. 1D). Analysis of anatomically matched, Nissl-stained coronal sections of P4.5 brains revealed that the reduction in size upon Brca1 ablation affected the neocortex and hippocampus(Fig. 1E-F′). Importantly, the decrease in neocortex size concerned both the lateral(Fig. 1E,F, dotted lines) and radial (Fig. 1E,F, solid lines)dimension. Remarkably, conditionally Brca1-ablated mice developed to adulthood without obvious effects on health and fertility. A reduced cerebral cortex was not observed in Brca1 heterozygous, Emx1-Cre-positive (Brca1^loxP/wtEmx1-Cre^+/-) P4.5 mice (data not shown), indicating that the conditional removal of one Brca1 allele, or the expression of Cre recombinase as such, had no effect on cerebral cortex size.

To investigate the nature of the reduction in the radial thickness of the Brca1-ablated neocortex, we analyzed P4.5 and P8.5 coronal brain sections for the expression of markers of certain cortical layers. We performed double immunofluorescence for FoxP2, which is predominantly expressed by early-born, deep-layer neurons(Ferland et al., 2003),together with Brn1 (Pou3f3 - Mouse Genome Informatics), which is predominantly expressed by later-born, upper-layer neurons(He et al., 1989; McEvilly et al., 2002), and also for Tbr1, which is strongly expressed by layer VI and layer III neurons(Bulfone et al., 1995; Hevner et al., 2001), together with Brn2 (Pou3f2), which is predominantly expressed by later-born,upper-layer neurons (He et al.,1989; McEvilly et al.,2002). This revealed the preservation of cortical layering in the conditional Brca1 knockout, with Brn1⁺ and Brn2⁺ neurons still preferentially found in the upper layers and FoxP2⁺ and Tbr1⁺ neurons preferentially in the deeper layers (Fig. 2A,B; see Fig. S1A in the supplementary material).

Cortical organization was further examined by immunofluorescence staining on P4.5 coronal brain sections for calretinin (calbindin 2), which, at early postnatal stages, is expressed by neurons in layer I (specifically Cajal-Retzius cells) (Del Río et al., 1995; Fonseca et al.,1995), pyramidal-like neurons in layer V(Schierle et al., 1997) and interneurons (Rogers, 1992; Wonders and Anderson, 2006). The abundance and organization of calretinin⁺ neurons appeared unaffected in the Brca1-ablated neocortex as compared with the control (see Fig. S2 in the supplementary material).

Given the reduction in the overall thickness of the Brca1-ablated neocortex, we next quantified the number of neurons positive for the aforementioned markers to see whether this reduction occurred equally in all cortical layers or in only a particular layer or neuron subpopulation. We defined three neuronal subpopulations, i.e. the Brn1⁺ (green) and FoxP2⁺ (red) single-positive and the FoxP2⁺Brn1⁺ double-positive (yellow) neurons, the distribution of which across the cortical wall is illustrated by the colored dots in Fig. S1B in the supplementary material. Interestingly, at P4.5(Fig. 2C) and P8.5 (see Fig. S1C in the supplementary material), whereas the number of FoxP2⁺and of FoxP2⁺ Brn1⁺ neurons in a 300-μm-wide segment of the cortical wall was reduced to almost half upon Brca1 ablation,the number of Brn1⁺ but FoxP2^- neurons, which were predominantly found in the upper-most layer, was unaffected. Similarly, at P8.5, whereas the number of Tbr1⁺ and of Tbr1⁺Brn2⁺ neurons in a 300-μm-wide segment of the cortical wall was reduced to almost half upon Brca1 ablation, the number of Brn2⁺ but Tbr1^- neurons, which were predominantly found in the upper-most layer, showed only a slight (statistically insignificant)reduction (Fig. 2D). Taken together, these data indicated that the reduction in overall cortical thickness (Fig. 1E-F′; Fig. 2A,B; see Fig. S1A in the supplementary material) was not due to a reduction in late-born neurons, but primarily reflected a decrease in earlier-born neurons.

To elucidate the developmental basis of the reduction in radial thickness of the postnatal neocortex (Fig. 2), and also to investigate the reduction in the lateral expansion of the neocortex (Fig. 1E-F′), we analyzed the effects of Brca1 ablation during embryonic cortical development. We first quantified the lateral extension of the VZ of the E13.5 dorsal telencephalon, which reflects neural progenitor number. For this purpose, E13.5 brains were subjected to serial Vibratome coronal sectioning followed by DAPI staining, and the length of the ventricular surface was compared between the control and Brca1-ablated dorsal telencephalon at three defined, anatomically corresponding regions along the rostrocaudal axis(Fig. 3A-A″). At each of these three regions, the length of the ventricular surface was significantly reduced, by one-third to half, upon Brca1 ablation(Fig. 3B-B″).

We complemented this analysis of VZ lateral extension by determining the abundance of DAPI-stained nuclei in the radial dimension of the progenitor layers. Using Tuj1 (Tubb3) immunostaining to distinguish the preplate from the subventricular zone (SVZ) and VZ, this revealed a decrease in the radial thickness of the progenitor layers in the E13.5 Brca1-ablated dorsal telencephalon (Fig. 3C,C′), which is likely to reflect the significant reduction in progenitor nuclei number found upon quantification(Fig. 3D, gray columns). By contrast, there was no detectable difference in the number of neuronal nuclei in the preplate at this developmental stage(Fig. 3D, green columns). Taken together, the reduction in both lateral ventricular length(Fig. 3A-B″) and progenitor layer radial thickness (Fig. 3C,D) at E13.5 indicates a decrease in the number of neocortical progenitors at this relatively early stage of neurogenesis.

The decrease in neocortical progenitor number upon Brca1 ablation could be due to reduced proliferation and/or cell death. In fact, DAPI staining did show numerous pyknotic nuclei in the progenitor layers of the Brca1-ablated E13.5 dorsal telencephalon(Fig. 3C′, arrowheads). Given also that Brca1 removal in other systems has been reported to result in increased apoptosis (Gowen et al., 1996; Xu et al.,1999a; Xu et al.,2001; Berton et al.,2003), we analyzed Brca1 control and cKO embryos at three different stages of cortical neurogenesis (E11.5, E13.5, E16.5) for apoptosis by TUNEL staining. Numerous TUNEL-positive pyknotic nuclei were detected in the Brca1-ablated cortical wall at E11.5 and E13.5, being present in both the progenitor layers (VZ plus SVZ) and the preplate, as identified by Tuj1 immunostaining (Fig. 4A). The majority of the cells with TUNEL-positive nuclei also showed immunoreactivity for active caspase 3 (data not shown). By contrast, very few apoptotic nuclei were detected in the E16.5 Brca1-ablated cortical wall (Fig. 4A); at this stage,a decrease in the radial thickness of the cortical plate was evident.

Quantification of TUNEL-positive nuclei in the progenitor layers, i.e. the VZ at E11.5 and the VZ and SVZ at E13.5 and E16.5, in the Brca1-ablated dorsal telencephalon revealed a dramatically increased absolute number (Fig. 4B, left)and proportion (Fig. 4B, right)at the early stages of neurogenesis (E11.5 and E13.5) as compared with the control. Both the VZ (Fig. 4B,gray column segments) and SVZ (Fig. 4B, white column segments) were found to contain apoptotic cells. There was also a substantial number of apoptotic nuclei in the preplate(51±4 and 45±11 nuclei per field at E11.5 and E13.5,respectively; data not illustrated). By contrast, at E16.5, the absolute number and proportion of apoptotic nuclei in the progenitor layers were much lower, albeit still significantly higher than in the control(Fig. 4B).

Whereas the TUNEL-positive cells in the preplate were most likely to be postmitotic neurons (Fig. 4A),the apoptotic cells in the progenitor layers could be progenitors, newborn neurons, or both. To directly investigate apoptosis of VZ progenitors, we performed immunofluorescence for the VZ progenitor marker Pax6(Götz et al., 1998) in combination with TUNEL staining. This revealed a small number of TUNEL-positive cells with Pax6 immunoreactivity (see Fig. S3 in the supplementary material), suggesting that VZ progenitors underwent apoptosis upon conditional Brca1 ablation (although we cannot exclude the possibility that the apoptotic cells also included newborn neurons that had inherited the Pax6 protein). However, most apoptotic nuclei in the VZ,although apparently surrounded by nestin⁺ cytoplasm (see Fig. S3 in the supplementary material), did not show Pax6 immunoreactivity, which most likely reflected proteolysis during the process of apoptosis(Taylor et al., 2008).

To obtain more compelling evidence for the apoptosis of VZ progenitors, we performed BrdU pulse-chase at E11.5 and determined the time course of the appearance of TUNEL-positive BrdU-labeled cells. Embryos were analyzed 1, 2, 4 and 6 hours after a single BrdU administration for BrdU-labeled nuclei and TUNEL staining. BrdU immunofluorescence showed that 1 and 2 hours after BrdU administration, most labeled nuclei were found in the basal half of the VZ(Fig. 5A,B), whereas after 4 and 6 hours, most labeled nuclei were located in its apical half(Fig. 5C,D), documenting the basal-to-apical interkinetic nuclear migration of VZ progenitors during their G2 phase (Götz and Huttner,2005). The proportion of BrdU-labeled nuclei was essentially the same for control and Brca1-ablated dorsal telencephalon(Fig. 5F). Importantly,analysis of the Brca1-ablated dorsal telencephalon for BrdU-labeled nuclei in the VZ that became TUNEL-positive(Fig. 5E) showed that these appeared rapidly after BrdU administration, being observed at the earliest time point analyzed (1 hour after BrdU) and increasing thereafter(Fig. 5G). By contrast, a much lower proportion of TUNEL-positive BrdU-labeled nuclei and no increase over time were observed for the control (Fig. 5G). Given the average length of the S and G2 phases of VZ progenitors in the E11.5 dorsal telencephalon of ∼3 hours and ∼2 hours, respectively (Takahashi et al.,1995; Calegari et al.,2005), these data imply that most of the TUNEL-positive BrdU-labeled cells observed during the BrdU pulse-chase were VZ progenitors in the S or G2 phase (rather than neurons derived therefrom), and hence provide strong evidence for substantial apoptosis of VZ progenitors upon Brca1 ablation.

To address whether the reduction in neocortical progenitors in the Brca1-ablated dorsal telencephalon was caused by reduced proliferation, we performed immunofluorescence for the M-phase marker phosphohistone H3 in combination with DAPI staining(Fig. 6A), and quantified the mitotic index for apical mitoses and basal mitoses in the VZ (E11.5) or VZ plus SVZ (E16.5) (Fig. 6B). At both E11.5 and E16.5 there was no significant difference in the mitotic index of either class of mitoses between the control and Brca1-ablated dorsal telencephalon. However, in line with the reduction in the number of nuclei per field in the progenitor layers that we observed at E11.5(Fig. 6A,C; see also Fig. 3D and Fig. 7B for E13.5) but no longer at E16.5, the abundance (as opposed to the mitotic index) of apical and basal mitoses at E11.5 was decreased accordingly (data not shown). We conclude that, in line with the observation that in the E11.5 Brca1-ablated dorsal telencephalon the labeling index 1 hour after BrdU administration is the same as in the control (Fig. 5F), cell cycle progression was unaffected upon Brca1ablation.

To analyze whether the reduction in neocortical progenitor number and radial thickness of the neocortex upon Brca1 ablation was attributable to a specific lineage change at the progenitor level, we analyzed the basal (or intermediate) progenitor lineage, which is one of the major lineages of neocortical progenitors(Haubensak et al., 2004; Miyata et al., 2004; Noctor et al., 2004). Basal progenitors are delaminated non-epithelial progenitors that (1) are downstream to apical progenitors (neuroepithelial cells, radial glial cells), (2) are thought to generate the majority of neocortical neurons, and (3) are characterized by the expression of Tbr2 (Eomes)(Haubensak et al., 2004; Englund et al., 2005; Attardo et al., 2008). We performed immunofluorescence for Tbr2 on control and Brca1-ablated E13.5 dorsal telencephalon (Fig. 7A) and quantified the percentage of Tbr2⁺ nuclei over total nuclei as revealed by DAPI staining(Fig. 7B, top). No significant change was found in the percentage of Tbr2⁺ cells in the cortical wall, indicating that the Brca1 ablation that occurred in neuroepithelial cells did not affect the subsequent generation of Tbr2⁺ basal progenitors from apical progenitors. As shown in Fig. 3D, the number of nuclei per field in the Brca1-ablated E13.5 dorsal telencephalon was reduced, which mostly reflected the decrease in the radial thickness of the progenitor layers (Fig. 7B,bottom).

Increased apoptosis caused by Brca1 removal is known to involve the p53 (Trp53) pathway (Lakin and Jackson, 1999; Xu et al.,2001; Fridman and Lowe,2003). We therefore performed immunofluorescence for p53 in the E11.5 control and Brca1-ablated dorsal telencephalon. In the control VZ, p53 immunoreactivity was low and, when detectable, was concentrated in the perinuclear cytoplasm (Fig. 8A,B). By contrast, Brca1-ablated VZ progenitors(identified by PCNA immunostaining) often showed nuclear p53 immunoreactivity(Fig. 8C,D). Thus, Brca1 ablation in the dorsal telencephalon results in the nuclear translocation of p53 and hence in the activation of the p53 pathway(Haupt et al., 1997; Fuchs et al., 1998).

Our study shows that Brca1, the gene of which shows evidence of positive selection during primate evolution(Huttley et al., 2000), has a role in controlling the size of the cerebral cortex during mouse embryonic development by preventing the apoptosis of early cortical progenitors and their immediate progeny. The reduction in size of the Brca1-ablated neocortex is remarkable in that it appears to be proportional, affecting both radial thickness and lateral extension, and in that cortical layering, despite this reduction, remains largely intact. The nature of this size reduction is consistent with the apoptosis that occurs in early, but not late, progenitors and their immediate progeny, and is in accordance with the radial unit hypothesis (Rakic, 1988; Rakic, 1995), which predicts that a depletion of early progenitors results in a decrease in the lateral expansion of the neocortex. The observation that the upper-most cortical layers are not reduced upon Brca1 ablation is consistent with the low levels of apoptosis we found in late progenitors and the neurons derived therefrom.

The apoptosis observed at early developmental stages upon Brca1ablation occurred not only in neurons, as indicated by the abundant TUNEL staining in the E11.5-13.5 preplate, but also in apical progenitors, as directly shown by the appearance of TUNEL staining in BrdU pulse-labeled VZ progenitors, and presumably also in basal progenitors as suggested by the substantial TUNEL staining in the SVZ. Taken together, the high levels of apoptosis observed in BrdU pulse-labeled S- and G2-phase cells, and the specific expression of Brca1 in late G1, S and G2 phase(Gudas et al., 1996; Vaughn et al., 1996), raise the possibility that the main function of Brca1 in preventing apoptosis of neural progenitors is in the S or G2 phase. We did not observe any effect on the mitotic index of apical and basal progenitors or on the Tbr2 lineage,indicating that Brca1 ablation did not affect cell cycle progression as such. Importantly, not only the nature of the reduction in size of the postnatal neocortex upon Brca1 ablation, but also its magnitude, were fully consistent with the number of apoptotic nuclei observed for both embryonic neural progenitors and neurons.

The question arises as to what underlies the high levels of apoptosis in early but not late progenitors upon Brca1 ablation. One possibility is that Brca1 ablation acutely causes apoptosis in the immediate downstream progenitors and neurons, and later there is an adaptation or compensation mechanism that renders progenitors and the neurons derived therefrom less susceptible to apoptosis. Another, perhaps more intriguing possibility is that there is a difference in the properties of the early versus late cortical progenitors in either their sensitivity to Brca1ablation or in an intrinsic readiness to undergo apoptosis. At early stages of cortical development (up to E13.5), most apical progenitor cells divide symmetrically to proliferate (Haubensak et al., 2004), and it might be that this expanding progenitor population is more sensitive to Brca1 ablation than the more committed, differentiating progenitors that are prevalent at later stages.

Another major difference between early and late apical progenitors is the transition from neuroepithelial to radial glial cells(Kriegstein and Götz,2003; Götz and Huttner,2005), and one may hypothesize that the latter are less sensitive to Brca1 ablation. However, we find this unlikely as high levels of progenitor apoptosis were observed at E13.5, when a large proportion of apical progenitors are radial glial cells(Götz and Huttner,2005).

A further question that arises is the molecular cause of the increased apoptosis observed upon Brca1 ablation. In light of the observation that many VZ progenitors in the Brca1-ablated E11.5 dorsal telencephalon exhibited nuclear localization of p53, there are two major possibilities. First, reflecting the function of Brca1 in DNA damage repair(Scully et al., 1997a; Venkitaraman, 2001; Deng and Wang, 2003), and given that Brca1 removal causes increased genetic instability(Shen et al., 1998; Xu et al., 1999b), there might be an increase in cells in the dorsal telencephalon that are accumulating DNA damage and undergoing apoptosis by the p53 pathway. Second, there is evidence for p53 function in mouse embryonic development, as p53-deficient mice may exhibit defects in neural tube closure(Armstrong et al., 1995; Sah et al., 1995). Increased apoptosis caused by constitutively active Notch signaling also results in nuclear localization of p53, and this increase in apoptosis is diminished in a p53-null background (Yang et al.,2004). Interestingly, BRCA1 has been reported to play a significant role in the differentiation of mammary stem and progenitor cells(Liu et al., 2008). These lines of evidence raise the possibility that Brca1 might have a specific role in the development of the cerebral cortex, distinct from its function in DNA damage repair.

Alterations in the extent of apoptosis of cortical progenitors during embryonic development have been shown to greatly affect cortical size(Kuida et al., 1996; Kuida et al., 1998; Yang et al., 2004; Depaepe et al., 2005), and it has been hypothesized that regulation of neural progenitor apoptosis,especially at early developmental stages, might have contributed to the changes in neocortex size that occurred during mammalian evolution(Haydar et al., 1999; Rakic, 2005). The present study adds Brca1 to the list of key molecules required for the cerebral cortex to develop to normal size by preventing early progenitor apoptosis. Importantly, the fact that not all early progenitor cells underwent apoptosis upon Brca1 ablation, which has also been reported with regard to the increased apoptosis upon gain of EphA receptor function(Depaepe et al., 2005) and activation of Notch signaling (Yang et al., 2004), raises the possibility that there is a subpopulation of progenitors that are particularly apoptosis prone, or that there is a stochastic mechanism that renders a certain proportion of progenitors to undergo apoptosis readily. Be this as it may, our data underscore the notion that the genetic changes that underlie cortical expansion might in particular affect the early progenitor pool (Rakic,1988; Rakic, 1995)and implicate Brca1 as an essential component to ensure the full complement of early cortical progenitors during embryonic development.