Reelin binds to very low–density lipoprotein receptor and apolipoprotein E receptor 2, thereby inducing mDab1 phosphorylation and activation of the phosphatidylinositide 3 kinase (PI3K) pathway. Here we demonstrate that Reelin activates the mitogen-activated protein kinase/extracellular signal-regulated kinase (ERK) pathway, which leads to the phosphorylation of Erk1/2 proteins. The inhibition of Src family kinases (SFK) blocked Reelin-dependent Erk1/2 activation. This was also shown in neuronal cultures from mDab1-deficient mice. Although rat sarcoma viral oncogene was weakly activated upon Reelin treatment, pharmacological inhibition of the PI3K pathway blocked Reelin-dependent ERK activation, which indicates cross talk between the ERK and PI3K pathways. We show that blockade of the ERK pathway does not prevent the chain migration of neurons from the subventricular zone (SVZ) but does inhibit the Reelin-dependent detachment of migrating neurons. We also show that Reelin induces the transcription of the early growth response 1 transcription factor. Our findings demonstrate that Reelin triggers ERK signaling in an SFK/mDab1- and PI3K-dependent manner and that ERK activation is required for Reelin-dependent transcriptional activation and the detachment of neurons migrating from the SVZ.
The migration and positioning of neurons are essential steps during the development of the Central Nervous System (CNS). Several molecules direct and modulate these migratory processes, including Netrins, Slits, Ephrins, and Reelin, through the activation of complex signaling cascades (Hatten 1999; Huber and others 2003; Marin and Rubenstein 2003). Reelin is a large extracellular protein that controls neuronal migration in a variety of laminated brain regions (D'Arcangelo and others 1997; Rice and Curran 2001). reeler mice, defective in Reelin, show devastating layering alterations in many regions of the brain (Rice and Curran 2001). Near its N-terminal region, Reelin contains a segment with no homology to other proteins that include the CR-50 epitope, which is essential for Reelin functions (Nakajima and others 1997; Kubo and others 2002). Although the exact functions of Reelin in vivo remain to be elucidated, it has been proposed that this protein is a stop signal or a detachment factor (Curran and D'Arcangelo 1998; Frotscher 1998; Dulabon and others 2000; Hack and others 2002). For instance, in the olfactory bulb (OB), Reelin produced by mitral cells induces the shift from tangential/chain to radial/individual neuronal migration, and the detachment of precursors in the rostral migratory stream (RMS) originated in the subventricular zone (SVZ) (Alcantara and others 1998; Hack and others 2002).
Reelin binds the apolipoprotein E receptor 2 (ApoER2) and the very low–density lipoprotein receptor (VLDLR), thereby triggering the phosphorylation of the adapter protein mDab1 (D'Arcangelo and others 1999; Hiesberger and others 1999; Howell and others 1999; Benhayon and others 2003). The double mutant for apoER2 and Vldlr (Trommsdorff and others 1999) or the mDab1 knockout mouse have a migratory phenotype that is indistinguishable from that of reeler mice (Howell and others 1997; Sheldon and others 1997; Ware and others 1997). Moreover, it has also been proposed that Reelin binds α3β1 integrins (Dulabon and others 2000). mDab1 phosphorylation is performed by the Src family kinases (SFK), which are activated by Reelin in a mDab1-dependent manner, indicating that mDab1 is both a substrate and an activator of SFK in neurons (Arnaud and others 2003; Bock and Herz 2003). Moreover, tyrosine phosphorylation of mDab1 is essential for Reelin function (Howell and others 2000).
Phosphorylated mDab1 binds to several intracellular signaling molecules, such as non-catalytic region of tyrosine kinase adapter protein 2 (Nckβ), the chicken tumor virus number 10 [CT10] regulator of kinase family of adapter proteins, the guanosine triphosphatease-activating protein disable 2 interaction protein, or the p85 subunit of phosphatidylinositide 3 kinase (PI3K) (Bock and others 2003; Homayouni and others 2003; Pramatarova and others 2003; Huang and others 2004). The binding of mDab1 to p85 may activate the p110 PI3K subunit and the subsequent phosphorylation of Akt1 (Beffert and others 2002; Ballif and others 2003; Bock and others 2003).
To further elucidate the functions of Reelin in neuronal migration and CNS development, we analyzed the downstream signaling events activated by this protein in neurons. We recently showed that Reelin induces the phosphorylation of the microtubule-associated protein 1B (MAP1B) and that MAP1B-deficient mice show some similarities to the migration deficits observed in reeler mice (Gonzalez-Billault and others 2005). In the present study, we show that Reelin activates the mitogen-activated protein kinases (MAPK), extracellular signal-regulated kinase 1 (Erk1), and Erk2 proteins, in a way that depends on both SFK and mDab1 phosphorylation. We also demonstrate a Reelin-dependent transcriptional activation of the early gene early growth response 1 (Egr-1) (also known as Zif268, Krox-24, NGFI-A, or Zenk), a transcriptional target of ERK (Harada and others 1996; Watson and others 1997; Hodge and others 1998), and that the Reelin-induced detachment of cells migrating from the SVZ requires MAPK Erk1/2 activation.
Materials and Methods
Antibodies and Reagents
Antiphospho-p44/42MAP Kinase (Thr202/Tyr204) and anti-phospho-Akt (Ser473) antibodies were purchased from Cell Signaling Technology (Danvers, MA), antibodies against Egr-1 (588) and Akt1 (C-20) were from Santa Cruz Biotechnology (Santa Cruz, CA), mouse monoclonal antibody against Erk (pan Erk) was from Transduction Laboratories, mouse monoclonal antibody against β Tubulin isotype III was from Sigma (St. Louis, MO) (clone SDL.3D10) or from Covance (San Diego, CA) (TUJ1), anti-mDab1 antibody was purchased from ExAlpha Biologicals (Maynard, MA), anti-phosphotyrosine (clone 4G10) was from Upstate Biotechnology, anti-pan-rat sarcoma viral oncogene (Ras) (Ab-3) antibody was from Oncogene Science (Cambridge, MA), affinity purified antibody anti-mDab1 (B3) used in western blots (WBs) was a generous gift from J.A. Cooper (Seattle, WA), mouse monoclonal anti-Reelin antibody (clone G10) was provided by A.M. Goffinet, and mouse monoclonal anti-Reelin blocking antibody (clone CR-50) was provided by K. Nakajima. Texas Red-conjugated Phalloidin was provided by Sigma. Alexa Fluor 488 goat anti-mouse immunoglobulin G (IgG) (H + L) and Alexa Fluor 488 F(ab′)2 fragment of goat anti-rabbit IgG (H + L) were from Molecular Probes. Goat anti-mouse–HRP and rabbit anti-goat–HRP secondary antibodies used in WB were purchased from DAKO (Glostrup, Denmark); goat anti-rabbit–HRP was from Sigma.
Protein G-Sepharose 4B Fast Flow was purchased from Sigma and Glutathion-Sepharose 4B beads from Amersham (Little Chalfont, UK). rhBDNF was from Promega (Madison, WI). All inhibitors used in the experiments were purchased from Calbiochem (San Diego, CA): PP2, PD 98059, and LY 294002.
Reelin Production and Neuronal Primary Culture Treatment
Reelin production was prepared as described (Gonzalez-Billault and others 2005). Reelin working concentration was around 2 ng/ml. Telencephalic neurons were obtained from E16 OF1 mouse embryos (Charles River Laboratories, Wilmington, MA) (Gonzalez-Billault and others 2005). The experiments were carried out in accordance with the European Community Council Directive and the National Institute of Health guidelines for the care and use of laboratory animals. Prior to treatment, neuronal cultures were starved. Reelin or control mock-conditioned media were then used for treatment at working concentration. Partially purified Reelin was a gift from Drs T. Curran and D. Benhayon (Keshvara and others 2001).
For pharmacological inhibition of SFK, PI3K, or mitogen activated protein kinase kinase (MEK), 1 h before treatment, cultures were supplemented with 10 μM of PP2, 50 μM of LY 294002, or 50 μM of PD 98059. Cultures from mdab1 −/− mice were similarly obtained from E16 mouse embryos.
For blocking experiments, Reelin and control supernatants were preincubated for 1 h with the CR-50 antibody or with control mouse IgGs (0.1 mg/ml) at 4 °C. rhBDNF treatments were performed at 25 μg/ml.
Stimulated cortical neurons were lysed in radioimmunoprecipitation (RIPA) buffer (50 mM Tris–Cl [pH 8.0], 150 mM NaCl, 1.0% Nonidet P-40, 0.5% sodium deoxycholate, and 0.1% sodium dodecyl sulfate [SDS]). Lysates were immunopricipitated with an agarose-conjugated antibody directed against Erk2 (Santa Cruz Biotechnology). Reactions were initiated by resuspending immunoprecipitates in assay buffer (25 mM Tris–Cl [pH 7.4], 5 mM b-glycerophosphate, 2 mM dithiothreitol, 0.1 mM Na3VO4, 10 mM MgCl2, and 1 mM phenylmethylsulfonyl fluoride) containing 20 μM unlabeled adenosine triphosphate, 5 μCi of [γ-32P] ATP, and 5 μg of myelin basic protein (MBP; UBI). Reaction mixtures were incubated at 30 °C for 15 min and were terminated by the addition of 7 μl of 5× Laemmli sample buffer. Reaction products were electrophoresed and transferred to an Immobilon membrane. Membranes were dried, exposed overnight to a phosphorimager screen (Fuji), and 32P-MBP levels were quantitated using a Fuji BAS1000 phosphorimager and PCBAS 2.0 software (Fuji Photo Film Co. Ltd.).
Immunoprecipitation, Pull-Down Assays, and WB
For WBs, lysates were collected in loading buffer (75 mM Tris [pH 6.5], 0.5 mM β-mercaptoethanol, 0.5% SDS, 5% glycerol, and 0.0125% bromophenol blue) and boiled. For immunoprecipitation or activity assays, cells were collected in Lysis Buffer (4-2-hydroxyethyl-1-piperazine ethanesultanic acid 50 mM [pH 7.5], 150 mM sodium chloride, 1.5 mM magnesium chloride, 1 mM ethyleneglycol-bis(2-aminoethylether)-N,N,N′,N′-tetra acetic acid, 10% glycerol, and 1% Triton X-100) containing a protease inhibitor cocktail (Roche) and phosphatase inhibitors (10 mM tetra-sodium pyrophosphate, 200 μM sodium orthovanadate, and 10 mM sodium fluoride). For immunoprecipitation, samples were incubated with the primary antibody overnight (o/n) at 4 °C (4 μg/sample). Protein G-Sepharose beads were added for 90 min at 4 °C, recovered, washed, and boiled in Loading Buffer before WB. Pull-down assays of activated Ras were performed using purified RBD-GST (GST-conjugated Ras Binding Domain of Raf1) as described (de Rooij and Bos 1997). Immunoprecipitates, pull-down samples, and cell lysates were resolved by SDS-polyacrylamide gels, transferred onto nitrocellulose membranes and developed with the enhanced chemiluminescence system (ECL) + system (Amersham).
Semiquantitative Reverse Transcription–Polymerase Chain Reaction
cDNA was reverse transcribed from RNA extracted from neuronal cultures treated with Reelin or control mock supernatants using Oligo (dT)15 (Promega cat. C1101). Care was taken to arrest amplification during the linear phase.
Multiplex polymerase chain reaction was performed by coamplification of Egr-1 and the house-keeping gene Gapdh. Primers used to amplify 505 and 435 bp specific fragments corresponding to Egr-1 and Gapdh, respectively, are described below: Egr-1-F: 5′-ATCCCAGCCAAACGACTCG-3′, Egr-1-R: 5′-GTGGAGTGAGCGAAGGCTGCT-3′, GAPDH-(+): 5′-GGCCCCTCTGGAAAGCTGTGG-3′, GAPDH-(−): 5′-CCTTGGAGGCCATGTAGGCCAT-3′.
Pictures were taken using a Gene Genius Bio Imaging System, and band intensities were studied using GeneTools software, both from Syngene.
SVZ Explant and Primary Cultures
Explants from the SVZ were dissected from P5 OF1 mice (Charles River Laboratories, Wilmington, MA) and cultured for 2–3 days in Matrigel matrix (Bioscience, San Jose, CA), as described (Wichterle and others 1997). Explants were maintained in serum-free Neurobasal-based medium and B27 supplemented (GibcoBRL). SVZ explants were cultured with control mock or Reelin-containing media. Some cultures were incubated with 25 or 50 μM of PD 98059. LY 294002 was used at 50 μM, and SFK inhibitor was used at 10 μM. Other SVZ explants were cultured with 25 μg/μl brain-derived neurotrophic factor (BDNF). Cultures were fixed and immunolabeled for β-III-tubulin.
For SVZ dissociated cultures, cells were cultured in coverslips coated with Poly-D-Lysine and Matrigel at a density of 70 000 cells per coverslip.
Quantification of detached cells was done as in Hack and others (2002). Thus, random fields of control mock-treated, Reelin-treated, or untreated SVZ explants were taken, detached cells were harvested at the microscope (40× objective), and percentages of detached cells were calculated. Five independent experiments were counted (21–28 explants per group).
Cultured neurons or explants in vitro were fixed with 4% paraformaldehyde and blocked with 0.2% gelatin and 10% Normal Goat Serum for 1 h at room temperature. Anti-β-III-tubulin (TUJ1), anti-phospho-p44/42MAP Kinase, anti-Egr-1, anti-mDab1, anti-pan-ERK, anti-pan-Ras antibodies, or phalloidin-Texas Red–conjugated were then used at 1:3000, 1:500, 1:200, 1:300, 1:500, 1:100, or 1:500 dilution, respectively. Secondary Alexa Fluor 488-tagged antibodies were incubated at a 1:500 dilution for 2 h. Cultures were mounted with Mowiol (Calbiochem) and visualized under an Olympus Fluoview FV300 confocal microscope.
Statistical significances were tested using the Statgraphics Plus 4.0 (including an analysis of variance test).
Reelin Induces Erk1 and Erk2 Phosphorylation through MEK
To test whether Reelin induces the activation of the ERK pathway, primary neuronal cultures from telencephalon were cultured for 4 days and treated with either recombinant Reelin supernatant or supernatant from control mock-transfected cells. As reported (Howell and others 1999), Reelin treatment for 15 min led to mDab1 phosphorylation, as detected by WB with anti-phosphotyrosine antibody after mDab1 immunoprecipitation (Fig. 1A). Reelin-treated cultures exhibited increased Akt1 phosphorylation, in agreement with Beffert and others (2002). Cultures treated with control mock supernatant did not exhibit phosphorylation of either the mDab1 or the Akt1 protein (Fig. 1A).
We next focused on the activation of p38, Jun N-terminal kinase (JNK), Erk1, and Erk2 proteins. Although we did not detect activation of p38 and JNK upon Reelin treatment (Fig. 1A), strong Erk1/2 phosphorylation was found when neuronal cultures were treated with Reelin for 15 min, compared with the control mock treatment (Fig. 1B). Similar findings were observed when cultures were treated with partially purified Reelin (2 ng/ml) (Fig. 1C). Reelin-dependent Erk1/2 activation reached maximum levels between 5 and 15 min of treatment and returned to basal activation states by 2 h (Fig. 1D). This short-term profile of Erk1/2 activation occurs after stimulation with numerous trophic factors, including glial cell derived neurotrophic factor, BDNF, and nerve grow factor growth factors and with axonal guidance cues such as Netrin-1 (Trupp and others 1999; Harada and others 2001; Forcet and others 2002; Gavalda and others 2004).
To confirm that Erk1/2 phosphorylation was dependent on Reelin, we used the CR-50 Reelin antibody (Nakajima and others 1997). Reelin and control mock supernatants were preincubated for 1 h with the CR-50 antibody or with control mouse IgGs (0.1 mg/ml). The Reelin supernatant preincubated with control IgGs led to Erk1/2 activation, whereas the CR-50–treated supernatants failed to induce the phosphorylation of these kinases (Fig. 1E). We conclude that preincubation with the CR-50 antibody, but not with control IgGs, blocks Reelin-induced Erk1/2 activation. Taken together, these results indicate that Reelin signaling activates MAPK Erk1/2 but not JNK or p38.
To determine whether Reelin-dependent Erk1/2 phosphorylation is mediated through the classical MAPK activator MAPK kinase (MEK, MAPKK), we used the MEK-specific inhibitor PD 98059. Pharmacological blockade of MEK inhibited Reelin-induced Erk1/2 phosphorylation (Fig. 2A). Moreover, we also determined the activity of Erk1/2 on an artificial substrate, MBP, by the addition of γ-P32 (Aronica and others 1997). A 2.1-fold increase in activity was detected in Reelin-treated samples in comparison with untreated, control mock-treated, or PD 98059-incubated samples (Fig. 2B). These results show that Erk1/2 phosphorylation corresponds to an increased enzymatic activity and that the Reelin-induced activation of Erk1/2 is performed through MEK.
Reelin-Dependent Erk1/2 Phosphorylation Requires SFK, mDab1, and PI3K Activation
Erk1/2 induction by many extracellular factors requires Ras activation. We thus studied whether this protein is activated by the Reelin-signaling pathway. Reelin led to a weak activation of Ras at 5 min (1.5-fold) that was not significant in comparison with controls (P > 0.05) (Fig. 3A). No Ras activation was found at longer treatment times (15 min, unpublished data), indicating that this activation may not be the only signaling pathway through which Reelin induces Erk1/2 phosphorylation. We compared the activation of Ras and the phosphorylation levels induced by Reelin on Akt1 and Erk1/2 proteins with the activation levels induced by a classical factor, BDNF (Huang and Reichardt 2003). BDNF-treated cultures exhibited robust Ras activation after 5 min (9-fold) (Fig. 3A). Consistent with this, the phosphorylation of Akt1 and Erk1/2 proteins was notably stronger in BDNF-treated samples than in Reelin-stimulated cultures (Fig. 3E). Interestingly, the effect of Reelin on Akt1 and Erk1/2 phosphorylation levels was similar. These results indicate that Reelin induces similar, but not maximal, Akt1 and Erk1/2 activation levels and that the contribution of Ras to Reelin-dependent Erk1/2 phosphorylation is limited.
Ras-independent Erk1/2 phosphorylation is involved in a number of events (York and others 1998; Takeda and others 1999; Yart and others 2002; Schmidt and others 2004; Wandzioch and others 2004). To analyze the possible cross talk between downstream Reelin-signaling events and Erk1/2 activation, we treated neuronal cultures with Reelin in the presence of pharmacological inhibitors for SFK, MEK, or PI3K (see Fig. 7). As expected, the blockade of SFK with PP2 inhibited Reelin-dependent mDab1 phosphorylation, which was not inhibited by the MEK inhibitor PD 98059 or by the PI3K inhibitor LY 294002 (Fig. 3C). Similarly, activation of the downstream Reelin effector Akt1 was inhibited by SFK and PI3K inhibitors (Fig. 3D, E), but not by PD 98059 (Fig. 3B). These results are consistent with previous studies showing that mDab1 phosphorylation requires the SFK Fyn and Src (Arnaud and others 2003; Bock and others 2003) and indicate that the PI3K and ERK pathways may be 2 coordinated Reelin-dependent cascades downstream of mDab1.
We analyzed the effect of pharmacological kinase blockade on Erk1/2 activation. Treatment of cultures with PP2 blocked Reelin-dependent Erk1/2 phosphorylation (Fig. 3D, left panel), indicating that SFK-dependent mDab1 phosphorylation is required for ERK activation. To confirm these findings, we cultured neurons from mDab1-deficient mice, and the phosphorylation of Akt1 and Erk1/2 proteins was analyzed after Reelin stimulation. In agreement with other studies, Reelin did not phosphorylate Akt1 in mDab1 −/− neuronal cultures (Fig. 3D, right panel) (Beffert and others 2002). Similarly, Reelin did not induce the phosphorylation of Erk1/2 in cultures lacking the mDab1 protein (Fig. 3D, right panel).
Interestingly, the blockade of PI3K by LY 294002 also inhibited Reelin-induced Erk1/2 phosphorylation, as occurred with PD 98059, implying that PI3K activation is required for Reelin-induced Erk1/2 activation (Fig. 3E). In contrast, LY 294002 did not affect the Erk1/2 phosphorylation induced by BDNF (Fig. 3E), indicating that the cross talk of the PI3K and ERK pathways is specific for the Reelin-signaling cascade. Altogether, these results show that Reelin leads to the sequential activation of SFK and mDab1, and of the PI3K, which in turn activates in parallel both the Akt1 and the ERK pathway.
Reelin-Induced Erk1/2 Activation Facilitates the Detachment of SVZ Cells
Next, we analyzed the involvement of ERK in a Reelin-regulated migratory process, that is, the detachment of neurons migrating from the RMS. Explants from postnatal SVZ were dissected out and cultured in Matrigel matrix. In these conditions, neuronal precursors migrate in chains, as occurs in vivo (Wichterle and others 1997). As described, the neurons migrated out of the explants and formed neuronal chains, which became disorganized after 3 days in vitro (DIV) (Fig. 4A). We first studied whether proteins of the Reelin-signaling pathway were expressed by SVZ neurons in vitro. Using immunofluorescence analysis of neuronal cultures, we found that chains of neurons migrating from the SVZ expressed mDab1, Erk, Ras (Supplementary Fig. 1) as well as Akt1 (unpublished data). We used PD 98059 to analyze the effect of ERK blockade on the detachment of SVZ cells in vitro. This treatment did not inhibit the migration of SVZ cells or the formation of neuronal chains. In contrast, PD 98059 inhibited the detachment of the SVZ cells at 3 DIV (Fig. 4A). Interestingly, PI3K inhibition with LY 294002 also blocked the detachment of the cultured SVZ neurons after 3 DIV (Fig. 4A).
In agreement with a previous study, the incubation of SVZ explants with Reelin caused a dramatic detachment of SVZ neurons already after 2 DIV. (Fig. 4B, C) (Hack and others 2002). We also found that explants obtained from reeler mice showed a response to Reelin incubation that was indistinguishable from that of wild-type SVZ explants, indicating that wild-type and reeler SVZ-cultured neurons show a similar response when exposed to Reelin (unpublished data). Incubation of Reelin-treated SVZ explants with the MEK inhibitor PD 98059 prevented the Reelin-induced neuronal detachment of SVZ chains. Instead, neuronal chains treated with Reelin 50 μM PD 98059 showed a compact appearance with virtually no detached neurons (Fig. 4B). Similar effects were observed with 25 μM PD 98059, as shown by the lack of detached cells in Reelin-treated cultures, although the thickness of neuronal chains was thinner than in cultures incubated with 50 μM PD 98059 (Supplementary Fig. 2). These findings indicate that ERK activation is required for Reelin-induced detachment of neurons migrating from the SVZ.
Blockade of SFK using PP2 also inhibited the Reelin-induced detachment of SVZ neurons, indicating that mDab1 phosphorylation may be crucial in this migration process (Fig. 4D). To address this issue, SVZ explant cultures were prepared from mDab1 −/− mice and were cultured with control mock or Reelin-containing supernatants for 2 DIV. Reelin did not induce neuronal detachment in the absence of mDab1 protein (Fig. 4D).
Finally, we also examined whether an independent activator of ERK (BDNF) induces the detachment of SVZ neuronal chains. Incubation with BDNF (25 μg/ml) did not cause detachment, either in the absence or in the presence of LY 294002 (Supplementary Fig. 3), suggesting that the Reelin-induced detachment of SVZ neurons occurs upon specific activation of the Reelin transduction cascade. Taken together, our results indicate that Reelin induces the detachment of migrating neurons from SVZ explants through activation of the Reelin-signaling cascade, which includes SFK, mDab1, PI3K, and Erk1/2.
Reelin-Dependent Erk1/2 Activation Upregulates the Expression of the Transcription Factor Egr-1
Activated Erk1/2 translocates to the nucleus and phosphorylates substrates like the transcription factor Ets-like transcription factor 1 (Elk1), which in turn enhances the expression of genes containing Serum Response Element (Shaw and Saxton 2003; Buchwalter and others 2004). We next explored the possible regulation of ERK target genes by Reelin.
To test whether the ERK/Elk1 downstream effector, Egr-1 (Hodge and others 1998), is regulated by Reelin, we analyzed Egr-1 expression in neuronal cultures. First, Reelin-treated cultured neurons (15 min) were incubated with anti-Erk1/2 phospho-specific antibodies. Reelin induced increased levels of phospho-Erk1/2 immunolabeling and translocation of phospho-Erk1/2 to the nucleus, both in telencephalic primary cultures (Fig. 5A) and in dissociated SVZ neuronal cultures (Fig. 5B). Similarly, incubation with Reelin for 1 h increased the expression of Egr-1 protein, as detected by immunofluorescence in neuronal cultures, which showed a nuclear localization (Fig. 5C).
WB performed after Reelin treatment of neuronal cultures showed that Egr-1 was upregulated in comparison with control mock-treated or untreated cultures (Fig. 6A). As expected for an early gene, increased protein levels were already observed at 30 min, reaching a maximal plateau at 60–90 min, decreasing thereafter (Fig. 6B).
To confirm that increased Egr-1 protein levels were due to transcriptional activation of Egr-1, we performed a semiquantitative reverse transcription–polymerase chain reaction analysis. Neuronal cultures showed a marked increase in Egr-1 mRNA after Reelin treatment for 30 min compared with cultures treated with control mock supernatant (Fig. 6C).
Moreover, pharmacological experiments showed that Reelin-induced Egr-1 protein levels were not upregulated when PD 98059 or PP2 was present, thereby confirming that Egr-1 expression depends on Reelin-dependent SFK and ERK activation (Fig. 6D). Finally, Egr-1 upregulation was also blocked by LY 294002 (Fig. 6D), thus reinforcing the notion of cross talk between the PI3K and ERK pathways in the Reelin-signaling cascade.
Here we demonstrate that, in addition to the PI3K/Akt1 pathway, Reelin activates the ERK pathway. Moreover, Reelin regulates the transcriptional activation of Egr-1 and cell adhesion of SVZ neurons during neuronal chain migration in an ERK-dependent manner. We also report a model in which SFK, and the PI3K and ERK pathways, cooperate to transduce Reelin signaling, thus involving Reelin receptors and the adapter protein mDab1. These findings show a novel Reelin-signaling mechanism (Fig. 7).
Reelin Stimulates the ERK Pathway through Activation of the PI3K Signaling Cascade
A previous study reports that Reelin does not stimulate the MAPK pathway (Ballif and others 2003). The differences with our findings may result from the distinct cell preparations tested or the different Reelin preparations used. In our hands, however, Erk1/2 activation of telencephalic cultured neurons was observed after treatment with 2 Reelin preparations (Reelin-containing supernatant and partially purified Reelin) and was specifically inhibited by preincubation with the CR-50 blocking antibody and by SFK inhibitors, which block mDab1 phosphorylation. Similarly, Erk1/2 activation was dependent on mDab1, which is an essential transducer of the Reelin signal. Moreover, although Reelin-induced Erk1/2 phosphorylation was not maximal (compared with BDNF treatment), it was comparable with the levels of Akt1 phosphorylation induced by Reelin. Finally, we show that Reelin induces nuclear translocation of phospho-Erk1/2 and that both Reelin-dependent transcriptional activation of Egr-1 and cell detachment of SVZ cells require Erk1/2 activation. Altogether, our findings support the view that ERK activation is a crucial event in the signaling cascade triggered by Reelin.
Our findings indicate that Reelin leads to weak Ras activation. Ras activation could be due to the classical Shc/Grb2/Sos induction (Hunter 2000). However, we were unable to find consistent association of Shc with mDab1 after Reelin treatment (unpublished data). Thus, although the participation of Shc in the Ras stimulation induced by Reelin cannot be discarded, we favor other mechanisms. For instance, similar to mDab1, the phosphotyrosine-binding domain of Shc may bind to the NPXY motifs of intracellular receptor tails, which would suggest a direct VLDLR/ApoER2 interaction with Shc, as occurs with the low density lipoprotein receptor–related protein 1 (LRP1), another NPXY-containing lipoprotein receptor (Barnes and others 2001). This process may be facilitated by receptor multimerization because receptor clustering is required for Reelin signaling and mDab1 phosphorylation and for the activation of the SFK and PI3K pathways (Strasser and others 2004).
In neurons, the ERK pathway can be activated via Rap1, in a Ras-independent manner, leading to sustained ERK activity (York and others 1998; Stork 2005). Although it is known that Reelin weakly activates Rap1 (Ballif and others 2004), we did not observe a sustained ERK activation upon Reelin stimulation. Our pharmacological experiments with the SFK inhibitor PP2 and the PI3K inhibitor LY 294002 indicate that the sequential SFK/mDab1/PI3K activation is the most relevant pathway that leads to MAPK activation and phosphorylation of Erk1/2. Consistent with this view, similar cross talk between the PI3K and the ERK pathways has been reported (e.g., in lysophosphatidic acid and Erythropoietin signaling) (Takeda and others 1999; Yart and others 2002; Schmidt and others 2004). Multiple models of cross talk between the PI3K and ERK pathways have been observed, including a Ras-independent pathway leading to ERK activation mediated by PI3K and protein kinase C (PKC) (Takeda and others 1999). The precise mechanism used by PI3K to activate the ERK pathway after treatment with Reelin remains to be clarified.
Reelin-dependent Activation Regulates the Transcription of Egr-1
Here we report that Reelin regulates transcriptional activation. This opens the possibility of new mechanisms for Reelin action in neural development. A number of genes show altered expression in reeler mice (Kuvbachieva and others 2004). Several extracellular proteins, for instance, Fibronectin or Tenascin-C, modulate gene expression (Ogawa and others 2002; Ruiz and others 2004). Interestingly, Netrin-1, another neuronal cue involved in migration and axonal growth, has been shown to activate the transcription factor nuclear factor of activated-T-cells (Graef and others 2003). Here we demonstrate that Reelin treatment of cultured neurons results in the upregulation of another immediate early gene, the Egr-1 gene, through activation of the ERK pathway. Egr-1 target genes include a variety of factors such as other transcriptional regulatory proteins, the p35 neuron-specific activator of Cdk5, extracellular matrix proteins such as Collagen III and Fibronectin, the phosphatase and tensin homolog, and the tumor suppressor protein p53 (Nair and others 1997; Harada and others 2001; Virolle and others 2001; Fu and others 2003). Moreover, constitutive expression of Egr-1 mRNA in brain is detected in several areas, including the neocortex, hippocampus, entorhinal cortex, and cerebellum, all of which are severely targeted in reeler and reeler-like mutants (Rice and Curran 2001; Bozon and others 2002).
However, neither the Egr-1 nor the Erk1 mutants appear to display migratory abnormalities. In fact, mutations for several components of the Reelin pathway, such as Src, p39, Nckβ, or Fyn, also do not display these abnormalities. This observation has been attributed either to functional redundancy or to the downstream position of these components in the pathway (Stein and others 1994; Ko and others 2001; Bladt and others 2003).
Egr-1 transcription in neurons is dynamically regulated by a variety of pharmacological and physiological stimuli (Beckmann and Wilce 1997). Interestingly, this gene is essential, in an ERK-dependent manner, for long-term potentiation (LTP) and long-term memory, as well as for memory reconsolidation (Jones and others 2001; Lee and others 2004). On its own, Erk1 also participates in the regulation of long-term adaptive changes, as reported in studies of the Erk1-deficient mouse (Pages and others 1999; Mazzucchelli and others 2002). Moreover, Egr-1 is strongly downregulated in memory-deficient doubly transgenic mice over-expressing amyloid precursor protein and presenilin 1, a mouse model for Alzheimer's disease (Dickey and others 2003). Similarly, Reelin and the Reelin receptors VLDLR/ApoER2 are also involved in memory formation and LTP (Weeber and others 2002; Larson and others 2003; Beffert and others 2005). We propose that the activation of the ERK/Egr-1 pathway is one of the signaling events by which Reelin modulates memory formation in the mature brain.
Reelin-Dependent ERK Activation Is Required for the Detachment of Neurons Migrating from the SVZ
Reelin induces the detachment of neurons migrating from the SVZ in vitro (Hack and others 2002). This finding agrees with the observations in vivo showing that the lack of Reelin (Hack and others 2002) or the adapter mDab1 (unpublished data) leads to the accumulation of SVZ-derived olfactory neurons before entry to the OB. These data have led to the notion that Reelin in the OB controls the change of neuronal chain migration to radial glia-guided migration and the subsequent detachment of migrating neurons. Here we have shown that the Reelin-induced detachment of migrating neurons in SVZ-derived explants is induced through ERK activation and can be prevented by ERK inhibitors, indicating that MAPK activation participates in the biological functions of Reelin.
The observation that the addition of ERK inhibitors alone also increases the thickness of migrating neuronal chains and alters their adhesion implies that ERK is involved in the regulation of neuron-to-neuron adhesion during neuronal migration, at least in neurons fated to the OB. Our pharmacological and genetic experiments also show that the Reelin-induced detachment of SVZ neurons also involves mDab1 and PI3K phosphorylation. Thus, together with a previous study showing that SFK and activation of atypical PKC are necessary for Reelin-dependent neuronal migration in an in vitro model of corticogenesis (Jossin and others 2003), our results indicate that a complex scenario of signaling cascades is required for Reelin-dependent neuronal migration, which includes ERK activation, at least in the SVZ.
Supplementary figures can be found at: http://www.cercor.oxfordjournals.org/.
We thank Drs J. A. Cooper, K. Nakajima, T. Curran, D. Benhayon, and A. M. Goffinet for generously providing the materials used in this study, M.J. Barallobre for help in preliminary experiments, C. Solé for scientific assistance, and T. Yates for technical assistance. LP, SS, and MFS hold postgraduate fellowships from the Spanish Ministry of Education and Science. JMU is a recipient of a “Ramón y Cajal” contract from the Spanish Ministry of Education and Science. This work was supported by grants from La Caixa Foundation, the Pfizer Foundation, and the Spanish Ministry of Education and Science (SAF2001-3290, SAF2004-07929, FIS-PI042280 and BFI2003-03594) to ES, JAD, and JMU and from the Spanish Ministry of Health (FIS), La Caixa Foundation (Convocatòria de Malalties Neurodegeneratives), and the Government of Catalonia (Suport als Grups de Recerca, and Distinció Joves Investigadors) to JXC. Conflict of Interest: None declared.