Cortical development in the presenilin-1 null mutant mouse fails after splitting of the preplate and is not due to a failure of reelin-dependent signaling

Authors

  • Rita De Gasperi,

    1. Research and Development James J. Peters Department of Veterans Affairs Medical Center, Bronx, New York
    2. Department of Psychiatry, Mount Sinai School of Medicine, New York, New York
    Search for more papers by this author
  • Miguel A. Gama Sosa,

    1. Research and Development James J. Peters Department of Veterans Affairs Medical Center, Bronx, New York
    2. Department of Psychiatry, Mount Sinai School of Medicine, New York, New York
    Search for more papers by this author
  • Paul H. Wen,

    1. Research and Development James J. Peters Department of Veterans Affairs Medical Center, Bronx, New York
    2. Department of Psychiatry, Mount Sinai School of Medicine, New York, New York
    Search for more papers by this author
  • Jingjun Li,

    1. Department of Psychiatry, Mount Sinai School of Medicine, New York, New York
    Search for more papers by this author
  • Gissel M. Perez,

    1. Research and Development James J. Peters Department of Veterans Affairs Medical Center, Bronx, New York
    2. Department of Psychiatry, Mount Sinai School of Medicine, New York, New York
    Search for more papers by this author
  • Tom Curran,

    1. Abramson Research Center, Children's Hospital of Philadelphia, Philadelphia, Pennsylvania
    Search for more papers by this author
  • Gregory A. Elder

    Corresponding author
    1. Research and Development James J. Peters Department of Veterans Affairs Medical Center, Bronx, New York
    2. Abramson Research Center, Children's Hospital of Philadelphia, Philadelphia, Pennsylvania
    3. Rehabilitation Medicine Service, James J. Peters Department of Veterans Affairs Medical Center, Bronx, New York
    • James J. Peters VA Medical Center, Research and Development (3F22), 130 West Kingsbridge Road, Bronx, NY 10468
    Search for more papers by this author

Abstract

Cortical development is disrupted in presenilin-1 null mutant (Psen1−/−) mice. Prior studies have commented on similarities between Psen1−/− and reeler mice. Reelin induces phosphorylation of Dab1 and activates the phosphatidylinositol 3-kinase (PI3K)/Akt pathway. Psen1 is known to modulate PI3K/Akt signaling and both known reelin receptors (apoER2 and VLDLR) are substrates for Psen1 associated γ-secretase activity. The purpose of this study was to determine whether reelin signaling is disrupted in Psen1−/− mice. We show that, while Dab1 is hypophosphorylated late in cortical development in Psen1−/− mice, it is normally phosphorylated at earlier ages and reelin signaling is intact in Psen1−/− primary neuronal cultures. γ-secretase activity was also not required for reelin-induced phosphorylation of Dab1. Unlike reeler mice the preplate splits in Psen1−/− brain. Thus cortical development in Psen1−/− mice fails only after splitting of the preplate and is not due to an intrinsic failure of reelin signaling. Developmental Dynamics 237:2405–2414, 2008. © 2008 Wiley-Liss, Inc.

INTRODUCTION

The presenilin-1 (Psen1) gene was first recognized because mutations in it and a related gene, presenilin-2 (Psen2), cause early onset familial Alzheimer's disease (St George-Hyslop,2000). Mice with null mutations in Psen1 (Psen1−/−) die during late intrauterine life or shortly after birth (Shen et al.,1997; Wong et al.,1997). The embryos are small with a generally hypomorphic caudal body as well as skeletal malformations and cardiovascular defects (Shen et al.,1997; Wong et al.,1997; Nakajima et al.,2004).

Central nervous system (CNS) development is disrupted in Psen1−/− embryos as well, in the telencephalon resulting in a cortical dysplasia that is invariably accompanied by some degree of intracerebral hemorrhage (Shen et al.,1997; Wong et al.,1997). The cellular and molecular basis for Psen1's effects particularly on CNS development remain incompletely understood. However, cell migration defects in the developing neocortex of Psen1−/− mice have been reported in several studies (Shen et al.,1997; Hartmann et al.,1999; Handler et al.,2000; Yuasa et al.,2002; Louvi et al.,2004; Wen et al.,2005; Wines-Samuelson et al.,2005).

The reelin pathway is a key developmental pathway that influences cell migration and cortical histogenesis (Rice and Curran,2001; D'Arcangelo,2006). Hartmann et al. (1999) first commented on similarities between the cortical dysplasia seen in Psen1−/− mice and reeler mice as well as reporting that Cajal-Retzius cells were depleted in Psen1−/− brain. Later studies documented Cajal-Retzius cell loss in Psen1−/− brain (Wines-Samuelson et al.,2005) as well as altered physiological characteristics of those that remain (Kilb et al.,2004).

Reelin acts by binding to two lipoprotein related receptors, the very low-density lipoprotein receptor (VLDLR) and apoER2 (Hiesberger et al.,1999; Trommsdorff et al.,1999). Reelin binding results in tyrosine phosphorylation of the adaptor protein Dab1 (reviewed in Herz and Chen,2006), which leads to recruitment and activation of Src family kinase members. Activation of the phosphatidylinositol 3-kinase (PI3K)/Akt pathway is one of the principle mechanisms of reelin's downstream effects leading to phosphorylation and inactivation of glycogen synthase kinase 3β (GSK-3β).

Psen1 could potentially influence signaling through the reelin pathway by two mechanisms. First, Psen1 is known to modulate signaling through the PI3K/Akt pathway. Indeed several studies have shown that as judged by Akt phosphorylation, basal PI3K activation is reduced in cells lacking Psen1 or both presenilins (Baki et al.,2004; Kang et al.,2005; Zhang et al.,2007), effects that occur independent of Psen1's influence on γ-secretase activity. In addition, Psen1 is functionally best known for its role in regulating intramembrane proteolysis by affecting γ-secretase activity (Iwatsubo,2004). Many γ-secretase substrates exist (Vetrivel et al.,2006) including apoER2 (May et al.,2003) and VLDLR (Hoe and Rebeck,2005). In several settings γ-secretase cleavage results in an intracellular domain (ICD) being released that translocates to the nucleus and acts as a transcriptional regulator. The notch ICD represents the best studied of these (Lai,2004), although transcriptionally active ICDs have been associated with the amyloid precursor protein (Raychaudhuri and Mukhopadhyay,2007), ErbB4 (Sardi et al.,2006), LRP1 (Kinoshita et al.,2003), and the reelin receptor apoER2 (May et al.,2003). What role if any ICDs generated from the reelin receptors play in reelin signaling is unknown. However, their existence suggests an additional mechanism whereby Psen1 might influence signaling through the reelin pathway.

This led us to investigate whether reelin signaling is altered in Psen1−/− brain. Here, we show that, while biochemical signaling through the reeler pathway is disturbed late in cortical development, it is intact at earlier stages. We also show that unlike reeler mice the preplate splits in Psen1−/− brain. These studies thus show that cortical development in the Psen1−/− neocortex fails only after splitting of the preplate and is not due to an intrinsic failure of reelin signaling in Psen1−/− cells.

RESULTS

Dab1 Is Hypophosphorylated in the Cerebral Cortex of E18.5 Psen1−/− Mice but Normally Phosphorylated in E14.5 to E16.5 Brain

Dab1 is an adaptor protein that mediates downstream signaling through the reeler pathway. In reeler mice, throughout the period of neocortical development Dab1 levels are increased and the protein is hypophosphorylated (Rice et al.,1998). Therefore, as an indicator of reelin pathway activation in Psen1−/− embryos, we initially examined the phosphorylation status of Dab1 in embryonic day (E) 18.5 Psen1−/− embryos and controls by immunoprecipitating Dab1 from brain followed by Western blotting with an anti-phosphotyrosine antibody. As shown in Figure 1A, Dab1 was hypophosphorylated in Psen1−/− compared with either wild-type or Psen1+/− embryos while total Dab1 levels were unaltered in Psen1−/− brain with the ratio of p-Dab1/Dab1 significantly decreased in the E18.5 Psen1−/− embryos (Fig. 1C; P = 0.0065, unpaired t-test).

Figure 1.

Decreased phosphorylation of Dab1 in embryonic day (E) 18.5 Psen1−/− mice but normal phosphorylation in E14.5 brain. A,B: Dab1 was immunoprecipitated from E18.5 (A) or E14.5 (B) brains of wild-type (+/+) as well as heterozygous (+/−) and homozygous (−/−) Psen1 mutant mice. Representative Western blots from studies that were performed multiple times are presented probed with an anti-phosphotyrosine antibody. The ∼80-kDa Dab1 band is shown. Lower panels show the same blots reprobed for total Dab1. C: Dab1 phosphorylation is expressed as the ratio of p-Dab1 to total Dab 1. Results are presented ± the standard error of the mean and represent averages of n = 4 wild-type (wt) and 2 Psen1−/− embryos at each age.

Thus, reelin signaling appeared to be defective in Psen1−/− brain, although unlike reeler mice (Rice et al.,1998), total Dab1 levels were not increased in Psen1−/− mice. However, to our surprise when we subsequently examined Dab1 phosphorylation at E14.5, Dab1 was phosphorylated in Psen1−/− brain to a similar level as in controls (Fig. 1B) with no alteration in the level of total Dab1 and no change in the ratio of p-Dab1/Dab1 (Fig. 1C; P = 0.65, unpaired t-test). Similar results were obtained with extracts from E16.5 brain (data not shown).

Because altered reelin levels might explain the Dab-1 hypophosphorylation at E18.5 we also examined reelin expression at E14.5 and E18.5 by Western blotting (Fig. 2). There were no differences in reelin levels between wild-type (wt) and Psen1−/− at either age (P = 0.29, E14.5; P = 0.77, E18.5, unpaired t-tests). Thus, Dab1 is hypophosphorylated in E18.5 Psen1−/− brain but normally phosphorylated at earlier embryonic ages and this change is not due to decreased levels of reelin.

Figure 2.

No change in reelin levels in Psen1−/− brain. A: Reelin levels were determined by Western blotting of extracts from E14.5 or E18.5 brain from wild-type (+/+), heterozygous (+/−), or homozygous (−/−) Psen1 mutant mice. Arrows indicate the ∼180, 250, and 400-kDa reelin bands. Lower panel shows the same blot reprobed for β-tubulin. B: The reelin levels were quantitated as a ratio of the sum of the three reelin bands to β-tubulin (n = 4 wild-type and 2 Psen1−/− at each age). There were no differences between wild-type (wt) and Psen1−/− at either age.

Dab1 Is Normally Phosphorylated in Primary Neuronal Cultures From Psen1−/− Embryos After Stimulation With Reelin

As an indication of the intrinsic responsiveness of Psen1−/− neuronal cells to reelin, we treated primary neuronal cultures from Psen1−/− and wild-type littermate control embryos with recombinant reelin or control supernatants for varying time intervals and monitored Dab1 phosphorylation by Western blotting. After a 15 min exposure to reelin, an approximate five- to sevenfold induction of Dab1 phosphorylation was observed in both wild-type and Psen1−/− cultures (Fig. 3A,B). This induction gradually subsided over 90 min without any significant difference in responsiveness between wild-type and Psen1−/− cultures (Fig. 3B).

Figure 3.

Treatment of Psen1−/− primary neuronal cultures with recombinant reelin induces normal Dab 1 phosphorylation. Primary neuronal cultures were treated with supernatants from reelin transfected or mock-transfected HEK cells. A: Results from a representative experiment are shown. The apparent increase in p-Dab1 in the Psen1−/− mock treated at 45 min represents a loading effect. B: The fold increase in Dab1 phosphorylation is expressed as the ratio of p-Dab1 to total Dab1. Results are presented ± the standard error of the mean and represent averages from 3–6 independent experiments for each time point. Responses in wild-type (wt) and Psen1−/− cultures were compared at each time point using unpaired t-tests. There were no significant differences at any of the time points (P values 0.10 to 0.46). C–E: Primary neuronal cultures were stimulated with reelin or control supernatant for 30 min in the presence or absence of the PI3K inhibitor LY294002. Representative Western blots from studies that were performed multiple times are shown probed for phospho-Dab1/total Dab1 (C), phospho-Akt (p-Akt)/total Akt (D), and phospho-GSK-3β (p-GSK-3β) total GSK-3β (E).

Binding of reelin to VLDLR and apoER2 also stimulates PI3K, an activation that is dependent on Dab1 phosphorylation (Beffert et al.,2002). PI3K stimulation leads to phosphorylation of protein kinase B (Akt) on serine 473, which in turn phosphorylates glycogen synthase kinase-3β (GSK 3β) on serine 9 (Beffert et al.,2002). Interestingly, Psen1 has been reported to affect signaling through the PI3K/Akt pathway (Baki et al.,2004; Kang et al.,2005). Therefore, as a measure of the intactness of downstream reelin signaling in Psen1−/− neurons, we monitored Akt and GSK-3β phosphorylation after reelin stimulation. Reelin induced Akt (Fig. 3D) and GSK-3β (Fig. 3E) phosphorylation in both wild-type and Psen1−/− cultures without any significant differences between genotypes. To test the dependence of Akt and GSK 3β phosphorylation on PI3K activation, cultures were stimulated with reelin in the presence or absence of the PI3K inhibitor LY294002. As shown in Figure 3, the inhibitor blocked reelin-induced Akt (Fig. 3D) and GSK-3β (Fig. 3E) phosphorylation in both wild-type and Psen1−/− cultures while having no effect on reelin-induced phosphorylation of Dab1 (Fig. 3C). Quantitation of duplicate experiments found no differences in the fold induction of Akt or GSK-3β phosphorylation between wild-type and Psen1−/− neuronal cells (unpaired t-tests; P = 0.64 for Akt and P = 0.40 for GSK-3β, data not shown).

γ-Secretase Activity Is Not Required for Reelin-Induced Phosphorylation of Dab1

Both apoER2 (May et al.,2003) and VLDR (Hoe and Rebeck,2005) are substrates for γ-secretase cleavage and the apoER2 cleavage has been reported to generate a transcriptionally active ICD (May et al.,2003). To determine whether γ-secretase activity affects reelin signaling, we treated primary neuronal cultures from wild-type embryos with γ-secretase inhibitors and examined whether reelin induced Dab1 phosphorylation in the presence of inhibitors. Primary neuronal cultures were treated overnight with either the γ-secretase inhibitor XXI or L-685,458 and then stimulated with reelin. Dab1 phosphorylation was induced in the presence of both inhibitors (Fig. 4A). The fold increase in Dab1 phosphorylation was determined by the ratio of p-Dab1 to total Dab1 in control and inhibitor treated cultures (Fig. 4B) and the three groups (n = 3, control and XXI treated; n = 2, L-865,458 treated) were compared using a one-way analysis of variance (ANOVA). No significant differences were found among the groups (F2,6 = 0.78; P = 0.49, no significant differences Tukey post hoc tests). As a control for the adequacy of γ-secretase inhibitor treatment, the same samples were probed for N-cadherin to search for the N-cadherin γ-secretase cleaved C-terminal fragment (CTF; Marambaud et al.,2002). As shown in Figure 4A, an N-cadherin CTF accumulated in the presence of both inhibitors indicating that γ-secretase inhibition had been effective.

Figure 4.

γ-secretase activity is not needed for reelin-induced phosphorylation of Dab 1 in primary neuronal cultures. A: Cultures were treated overnight with the γ-secretase inhibitors XXI or L-685,458 and then stimulated with reelin for 15 min. Blotting for phospho-Dab1 and total Dab1 is shown. The same samples were blotted and probed for N-cadherin. A band for the full-length N-cadherin is visible, and in the γ-secretase inhibitor treated lanes, a band that corresponds to the N-cadherin CTF is indicated. The N-cadherin blot was stripped and reprobed for β-tubulin (β-tub) as a loading control. B: The fold increase in Dab1 phosphorylation is expressed as the ratio of p-Dab1 to total Dab1. C: Extracts from wild-type and Psen1−/− neurons with or without reelin treatment were sequentially probed for apoER2, VLDLR, and β-tubulin (β-tub). D: Cultures were treated overnight with γ-secretase inhibitors and then stimulated with reelin followed by blotting for apoER2. Full-length (FL) apoER2 and the apoER2 CTF are shown. Last lane shows blotting for apoER2 in untreated Psen1−/− neurons. Reprobing for β-tubulin (β-tub) is shown in the lower panel.

As an additional indicator of the effectiveness of γ-secretase inhibition we also assessed apoER2 and VLDLR processing. Western blotting for apoER2 and VLDR showed that both receptors were equally expressed in wild-type and Psen1−/− primary neuronal cultures and that their levels were not affected by treatment with reelin (Fig. 4C). As shown in Figure 4D, a fragment accumulated in the γ-secretase inhibitor treated wild-type cultures that migrated at the expected molecular weight for an apoER2 CTF (May et al.,2003) and a similar fragment appeared in Psen1−/− cultures. We were not able to detect a VLDLR CTF in γ-secretase inhibitor treated cultures. We suspect that this reflects a sensitivity problem related to the generally lower levels of VLDLR expressed in neurons (Trommsdorff et al.,1999) rather than a failure of the inhibitors to affect VLDLR cleavage. These studies thus argue that despite apoER2 and VLDLR being γ-secretase substrates, γ-secretase cleavage is not necessary for reelin activation of Dab1.

Preplate Splits in Psen1−/− Telencephalon.

In developing mouse neocortex, an initial wave of postmitotic cells exits the ventricular zone, migrating radially and establishing a layer known as the preplate (Gupta et al.,2002; Ayala et al.,2007). A second wave of migrating cells splits the preplate into a superficial marginal zone and a deeper subplate. In reeler mice, the preplate forms but never splits and an abnormal layer termed the superplate develops (Sheppard and Pearlman,1997).

Splitting of the preplate can be assessed by chondroitin sulfate proteoglycan (CSPG) immunostaining, which marks the preplate and its derivatives (Sheppard et al.,1991; Bicknese et al.,1994). CSPG immunostaining in the telencephalon from E15.5 wild-type and Psen1−/− embryos is shown in Figure 5. In both wild-type and Psen1−/− brain, CSPG staining appeared in distinct layers corresponding to marginal zone and subplate labeling indicating that splitting of the preplate had occurred in the Psen1−/− telencephalon. We also found no difference between wild-type and Psen1−/− embryos in the pattern of calretinin staining (Supplementary Figure S1, which can be found online), which also labels the preplate and its derivatives (Weisenhorn et al.,1994; Fonseca et al.,1995). Thus unlike reeler mice, the preplate splits in Psen1−/− mice (Supplementary Figure S1).

Figure 5.

Chondroitin sulfate proteoglycan (CSPG) immunostaining reveals splitting of the preplate in Psen1−/− telencephalon. Horizontally cut Vibratome sections from embryonic day (E) 15.5 wild-type or Psen1−/− embryos were immunostained with CSPG (green) along with a 4′,6-diamidino-2-phenylindole (DAPI) nuclear stain (blue). A,B: Sections through the frontal poles are shown. Asterisks (*) indicate the unstained cortical plate. C–H: Sections from the lateral telencephalon of wild-type (C,E,G) or Psen1−/− embryos (D,F,H) are shown labeled for DAPI (C,D), immunostained for CSPG (E,F), or as a merged image (G,H). The marginal zone (MZ), cortical plate (CP), and subplate region (SP) are indicated. Scale bar = 200 μm in A,B, 50 μm in C,D.

As an additional indicator of the status of cell migration in Psen1−/− telencephalon around the time of preplate splitting, we injected bromodeoxyuridine (BrdU) into pregnant female mice at E11.5 and examined BrdU immunostaining at E15. As shown in Figure 6, in wild-type and Psen1−/− embryos the most highly labeled cells (i.e., those cells that left the cell cycle first at E11.5) contributed to cells in the marginal zone and to cells in the subplate region. Some lightly BrdU-labeled cells (indicative of cells that left the cell cycle and migrated later) could be seen in the central portion of the cortical plate. Thus in Psen1−/− embryos, latter generated cells continued to migrate into the cortical plate in a pattern that was not distinguishable between wild-type and Psen1−/− brains, further indicating the intactness of neuronal migration at this stage of development in Psen1−/− brain.

Figure 6.

No difference in bromodeoxyuridine (BrdU) labeling between wild-type and Psen1−/− embryos. Pregnant female mice received one injection of BrdU at embryonic day (E) 11.5 and BrdU immunostaining was performed on sections of E15 brain. A–F: Coronal sections from the lateral telencephalon of wild-type (A,C,E) or Psen1−/− embryos (B,D,F) are shown immunostained for BrdU (A,B), labeled with 4′,6-diamidino-2-phenylindole (DAPI; C,D), or as merged images (E,F). The marginal zone (MZ), cortical plate (CP), subplate region (SP), and intermediate zone (IZ) are indicated. Scale bar = 100 μm.

DISCUSSION

Here we investigated whether altered reelin signaling might play a role in the disrupted cortical lamination found in Psen1−/− mice. While our initial studies did indeed find Dab1 to be hypophosphorylated in Psen1−/− cortex at E18.5, subsequent studies found Dab1 to be normally phosphorylated at earlier ages. We also found that when primary neuronal cultures from Psen1−/− embryos were treated with recombinant reelin, Dab1 was phosphorylated normally and that reelin induced phosphorylation of Akt as well as phosphorylation of its downstream target GSK-3β.

Reelin signaling could be intact at the cellular level but nevertheless perturbed at the tissue level if migrating neurons in Psen1−/− brain were deprived of their source of reelin. Cajal-Retzius cells are present in Psen1−/− brain and express reelin (Kilb et al.,2004; Wen et al.,2005; Wines-Samuelson et al.,2005). However, both loss (Wines-Samuelson et al.,2005) as well as physiological dysfunction (Kilb et al.,2004) of Cajal-Retzius cells has been reported in Psen1−/− brain. Our finding that the preplate is split at E15.5 in Psen1−/− neocortex argues that signaling through the reelin pathway at the tissue level is intact up to this stage of cortical development. These findings are also consistent with the anatomical data that suggest that Cajal-Retzius cell loss is a relatively late event occurring well after the time of preplate splitting (Wines-Samuelson et al.,2005). Cajal-Retzius cell loss might explain the hypophosphorylation of Dab1 at E18.5. However, even at this stage, losses of Cajal-Retzius cells are relatively mild (∼20%; Wines-Samuelson et al.,2005), and as shown in this study reelin levels are not affected making it unlikely that altered availability of reelin explains the profound effects on Dab1 phosphorylation. It is also of note that recent studies have found that even massive losses of Cajal-Retzius cells do not lead to a reeler phenotype (Yoshida et al.,2006).

Multiple studies have addressed the issue of cortical development in Psen1−/− mice (Shen et al.,1997; Handler et al.,2000; Yuasa et al.,2002; Louvi et al.,2004; Wen et al.,2005; Wines-Samuelson et al.,2005; Kim and Shen,2008). That cortical development is altered in the absence of Psen1 is clear. However, the timing of the onset of the developmental defect as well as its cellular and molecular nature remains less clear. Some studies have suggested that neural progenitor cells are depleted in Psen1−/− telencephalon as early as E12.5 (Shen et al.,1997) and that the cortical plate is thinned and disorganized at E14.5 (Handler et al.,2000). In the latter study, an early migration defect was also supported by studies showing that progenitor cells labeled with BrdU at E10.5 were abnormally dispersed across the cortical layers in Psen1−/− telencephalon at E14.5 (Handler et al.,2000). Our own studies described here, however, found no systematic alteration in the distribution of BrdU-labeled cells labeled at E11.5 and examined at E15.

Defects in cortical lamination can be generally divided into the preplate splitters and nonsplitters (Gupta et al.,2002; Ayala et al.,2007), with reeler being the classic nonsplitter (Rice and Curran,2001) along with mice having null mutations in Dab1 (Howell et al.,1997) or combined mutations in the reelin receptors, VLDLR and apoER2 (Trommsdorff et al.,1999). Cdk5 null mutants were the first recognized preplate splitters (Gilmore et al.,1998; Kwon and Tsai,1998) and combined mutations in Cdk5's regulatory subunits p35 and p39 produce a similar phenotype (Ko et al.,2001).

Our results clearly place the Psen1 null mutant among the preplate splitters, thus arguing that the major defects in cortical development are relatively late events. Other studies have documented defects in both radial and tangential migration in Psen1−/− brain (Louvi et al.,2004). However, these effects were only apparent at E16.5 or later also arguing for a relatively late effect of Psen1 on cell migration. Loss of radial glia occurs in Psen1−/− telencephalon (Louvi et al.,2004; Wen et al.,2005). However, this loss is also only clearly evident at E16.5 or later. A relatively late migration defect is also supported by our prior BrdU-labeling studies showing that cells labeled at E12.5 (1 day later than in the present study) are abnormally dispersed across the cortex in the Psen1−/− telencephalon at E18.5 rather than being concentrated in the inner layers as in wild-type mice (Wen et al.,2005). This pattern is most suggestive of a late migration defect in Psen1−/− brain in which later generated unlabeled cells fail to migrate past earlier generated labeled ones. The hypophosphorylation of Dab1 at E18.5 documented in this study can also be seen as further evidence of a late effect. Thus, collectively most evidence supports a predominately late migration defect in the Psen1−/− telencephalon.

At the cellular level, it has been suggested that neurons differentiate prematurely in Psen1−/− brain leading to a premature depletion of neural progenitor cells (Handler et al.,2000), although this effect has been described as being a region-specific effect that becomes normalized later. A similar effect has also been suggested to occur in a neural specific conditional knockout of Psen1 (Wines-Samuelson et al.,2005). The only quantitative analysis of this phenomenon (Wen et al.,2004), however, found no evidence for premature neuronal differentiation in single cell suspensions of Psen1−/− telencephalon suggesting that premature neuronal differentiation is not a general event in Psen1−/− brain and unlikely to explain the altered cortical development.

The causes of the late migration defect in Psen1−/− telencephalon remain unexplained. Similarities between Cdk5 and Psen1 null mutants have been commented on previously (Louvi et al.,2004). Of interest, Cdk5 phosphorylates β-catenin and this phosphorylation has been reported to affect the amount of Psen1 that becomes bound to β-catenin (Kesavapany et al.,2001). In a latter study, these same investigators showed that Cdk5 can phosphorylate a threonine in the Psen1 CTF and that this phosphorylation stabilizes the CTF (Lau et al.,2002). Both observations could place Psen1 in a linear pathway downstream of Cdk5 and notably Cdk5 null mutants show normal Dab1 phosphorylation early in cortical development, but hypophosphorylation late in development (Keshvara et al.,2002).

However, arguing against a Cdk5/ Psen1 connection is the observation that Cdk5 and its activators are expressed primarily in postmitotic neurons (Tsai et al.,1993; Zheng et al.,1998) while Psen1 is expressed in neural progenitor cells and indeed selective expression of Psen1 in neural progenitors cells is sufficient to rescue cortical development in Psen1−/− brain (Wen et al.,2005). By contrast Psen1 expression in postmitotic neurons rescues none of the abnormalities in Psen1−/− brain (Wen et al.,2005), suggesting that the major effects of Cdk5 and Psen1 on cortical development occur in different cell types. In Cdk5 null mutants it has been proposed that the late hypophosphorylation of Dab1 may be caused by late migrating neurons failing to reach the source of reelin (Keshvara et al.,2002) and this remains a possible explanation for the hypophosphorylation of Dab1 late in development in Psen1−/− cortex even if Psen1 and Cdk5 do not interact in a linear pathway.

Functionally Psen1 is best known for its role in γ-secretase activity (Vetrivel et al.,2006). Among the known γ-secretase substrates are several members of the low-density lipoprotein receptor (LDLR) family, including LRP1 (May et al.,2002), and the reelin receptors apoER2 (May et al.,2003) and VLDLR (Hoe and Rebeck,2005). In several settings, γ-secretase cleavage results in an ICD being released that translocates to the nucleus and acts as a transcriptional regulator. The notch ICD represents the best studied of these (Lai,2004), and altered notch signaling has been described in Psen1−/− embryonic brain (Handler et al.,2000; Yuasa et al.,2002). However, transcriptionally active ICDs have been associated with the amyloid precursor protein (Raychaudhuri and Mukhopadhyay,2007), ErbB4 (Sardi et al.,2006), LRP1 (Kinoshita et al.,2003), and apoER2 (May et al.,2003). However, the fact that reelin signaling is intact in Psen1−/− cortex suggests that in the case of apoER2, a γ-secretase generated ICD is unlikely to play a significant role in reelin-induced cortical lamination.

It is perhaps more surprising that reelin induced activation of the PI3K/Akt pathway is intact in Psen1−/− neuronal cells given the reported effects of Psen1 and Psen1 FAD mutants on PI3K/Akt activation (Weihl et al.,1999; De Sarno et al.,2001; Vestling et al.,2001; Baki et al.,2004; Kang et al.,2005; Repetto et al.,2007; Zhang et al.,2007). Indeed several studies have shown that, as judged by Akt phosphorylation, basal PI3K activation is reduced in cells lacking Psen1 or both presenilins (Baki et al.,2004; Kang et al.,2005; Zhang et al.,2007). However, ligand-induced activation of Akt is more variably affected by the absence of presenilins and when affected has been attributed to altered receptor levels rather than interference with signal transduction (Kang et al.,2005; Repetto et al.,2007). While the studies reported here do not negate a role of Psen1 in PI3K/Akt signaling, they show that in the context of reelin signaling, Psen1 is not essential for PI3K/Akt activation in neuronal cells and that the failure of cortical development in Psen1−/− mice is a relatively late event that is not due to an intrinsic failure of reelin signaling.

EXPERIMENTAL PROCEDURES

Animals

Heterozygous Psen1+/− mice (Shen et al.,1997) were mated to produce Psen1−/− embryos with the day a vaginal plug was detected designated as E0.5. Psen+/+ and Psen1+/− littermates were used as controls. Because Psen1+/− mice exhibit no known developmental abnormalities, they were treated as wild-type controls. All protocols were approved by the Institutional Animal Care and Use Committees of the James J. Peters Department of Veterans Affairs Medical Center and the Mount Sinai School of Medicine and were conducted in conformance with the National Institutes of Health “Guide for the Care and Use of Laboratory Animals” (NIH publication 80-23).

Dab1 Immunoprecipitations

For immunoprecipitations (IPs), tissue was homogenized by sonication in 350 μl of 25 mM Tris-HCl, 150 mM NaCl, 1 mM ethylenediaminetetraacetic acid (EDTA), 1% Triton X-100 plus HALT protease inhibitor cocktail (Pierce Biotechnologies, Rockford, IL) and phosphatase inhibitor cocktail I and II (Sigma-Aldrich, St. Louis, MO; IP buffer). The suspension was gently mixed by rotation for 2 hr at 4°C and then centrifuged at 14,000 × g for 10 min at 4°C. The supernatant was saved and protein concentrations were determined using the BCA reagent (Pierce Biotechnologies) according to the manufacturer's instructions. Three hundred micrograms of total protein from each sample was precleared with 20 μl of Protein G beads (Santa Cruz Biotechnology, Santa Cruz, CA) and immunoprecipitated with goat anti-Dab1 polyclonal antibody (1:10 final dilution; Santa Cruz Biotechnology) overnight at 4°C. Immune complexes were precipitated by addition of 20 μl of Protein G beads and the suspension rotated for 2 hr at 4°C. The pellet was collected by centrifugation, washed four times with IP buffer, and resuspended in 20 μl of IP buffer mixed with 20 μl of 4× Laemmli gel loading buffer. The final sample was boiled for 10 min and then centrifuged at 14,000 × g for 10 min at room temperature. Proteins were separated by sodium dodecyl sulfate (SDS) polyacrylamide gel electrophoresis and transferred to polyvinylidene difluoride (PVDF) membranes (Millipore Corporation, Billerica, MA). After blocking for 1 hr with 5% bovine serum albumin in 50 mM Tris HCl pH 7.6, 0.15M NaCl, 0.1% Tween-20 (TBST), the membranes were probed with a mouse monoclonal anti-phosphotyrosine antibody (1:5,000; PY20, Zymed/Invitrogen, Carlsbad, CA) diluted in blocking buffer. After extensive washes, the membranes were incubated for 1 hr with horseradish peroxidase-conjugated goat anti-mouse IgG (1:10,000; GE Amersham, Piscataway, NJ). The bands were visualized with ECL+ reagent (GE Amersham) and exposed to CLXposure film (Pierce). Total Dab1 was determined using a polyclonal rabbit anti-Dab1 antibody (1:3,000; Rockland Immunochemical Inc., Gilbertsville, PA).

Reelin Preparation

To obtain recombinant reelin containing or control supernatants, HEK 293 cells were transfected with the plasmid pCRL (D'Arcangelo et al.,1997) or with pcDNA3 (Invitrogen, Carlsbad, CA) using the Fugene 6 reagent (Roche Applied Science, Indianapolis, IN). Twenty-four hours posttransfection the cells were switched to Neurobasal medium (Invitrogen) for 48 hr. Control and reelin containing supernatants were harvested, centrifuged at 14,000 × g for 15 min, and the supernatants frozen at −80°C. The presence of full-length reelin was verified by Western blotting using the mouse monoclonal anti-reelin antibody G10 (Chemicon, Temecula, CA).

Treatment of Primary Neuronal Cultures With Reelin

Primary neuronal cultures were prepared from the cerebral cortex isolated from E15.5 to E16 embryos as previously described (Gama Sosa et al.,2007) and maintained in Neurobasal medium/B27 supplement (Invitrogen). Cultures were treated after 5 days in vitro (DIV) with reelin or control supernatant and lysed in a cold cell lysis buffer consisting of 10 mM Na phosphate buffer, pH 7.4, 150 mM NaCl, 2 mM EDTA, 1% Triton X-100, 0.25% Na deoxycholate, and 0.5% SDS, supplemented with HALT protease and phosphatase inhibitor cocktails (Pierce; Beffert et al.,2002). The lysates were centrifuged for 15 min at 14,000 rpm and the supernatant saved. When γ-secretase inhibitors were used, cultures were treated overnight with 1 μM of inhibitor XXI (also known as compound E) or L-685,458 (Calbiochem, San Diego, CA). To inhibit PI3K activity cultures were pretreated with 50 μM LY294002 (Cell Signaling Technology, Danvers, MA) for 1 hr.

Dab-1 phosphorylation was assessed by Western blotting using the anti-phosphotyrosine monoclonal antibody 4G10 with the phospho-Dab1 (p-Dab1) band identified by size as described in Bock et al. (Bock et al.,2003). A total of 10–15 μg of protein was transferred onto Nitrocellulose filters (Protran, BA85, 0.45 mm, Schleicher & Schuell/Perkin Elmer Life Sciences, Waltham, MA) and the filters were blocked for 1 hr in phosphate buffer saline (PBS)/5% nonfat dry milk (Upstate). The filters were incubated overnight with anti-phosphotyrosine monoclonal antibody 4G10 (Upstate Biotechnologies/Millipore; 1:1,500 diluted in blocking solution) followed by horseradish peroxidase–conjugated goat anti mouse IgG (GE Amersham; 1:5,000 in blocking solution). Signals were detected using the ECL+ reagent (GE Amersham) and exposure to CLXposure film (Pierce). Total Dab1 was determined independently using rabbit anti-Dab1 antibodies (Rockland; 1:3,000). Levels of Akt and GSK3β and their phosphorylation status were analyzed by Western blot using rabbit monoclonal or rabbit polyclonal anti p-Akt (Ser473), anti-total Akt, anti–p-GSK3β (Ser9) and anti-total GSK-3β (Cell Signaling Technology) as suggested by the manufacturer. Western blotting for apoER2 and VLDLR were performed with a rabbit polyclonal anti-apoER2 antiserum (1:1,500, Novus Biologicals, Littleton, CO) and a mouse monoclonal anti-VLDLR antibody (Santa Cruz Biotechnologies). A rabbit polyclonal anti–β-tubulin antibody (1:1500; Abcam Inc., Cambridge, MA) was used as loading control. For densitometric quantification of bands, nonsaturated chemiluminiscence films were scanned and the images analyzed with ImageQuant TL software (GE Amersham).

Immunohistochemistry

Embryos were fixed in 4% paraformaldehyde in PBS overnight and stored in PBS until processing. Immunohistochemical staining was performed as previously described (Wen et al.,2005). The primary antibodies used were a monoclonal IgM anti-chondroitin sulfate proteoglycan (CSPG) antibody (1:300; CS65, Sigma-Aldrich) and a rabbit polyclonal anti-calretinin antibody (1:500; Chemicon). Immunofluorescence staining was detected with species specific Alexa Fluor secondary antibody conjugates (1:400, Invitrogen) and nuclei were counterstained with 1 μg/ml 4′,6-diamidino-2-phenylindole (DAPI). BrdU immunohistochemistry was performed as previously described (Wen et al.,2002).

Acknowledgements

R.D.G. received a Young Investigator Award from the National Alliance for Research in Schizophrenia and Affective Disorders (NARSAD).

Ancillary