Necdin Is Required for Terminal Differentiation and Survival of Primary Dorsal Root Ganglion Neurons

Experimental Cell Research 277, 220 –232 (2002) doi:10.1006/excr.2002.5558

Necdin Is Required for Terminal Differentiation and Survival of Primary Dorsal Root Ganglion Neurons Risa Takazaki, 1 Isao Nishimura, 1 and Kazuaki Yoshikawa 2 Division of Regulation of Macromolecular Functions, Institute for Protein Research, Osaka University, Yamadaoka 3-2, Suita, Osaka 565-0871, Japan

Necdin is expressed predominantly in postmitotic neurons and serves as a growth suppressor that is functionally similar to the retinoblastoma tumor suppressor protein. Using primary cultures of dorsal root ganglion (DRG) of mouse embryos, we investigated the involvement of necdin in the terminal differentiation of neurons. DRG cells were prepared from mouse embryos at 12.5 days of gestation and cultured in the presence of nerve growth factor (NGF). Immunocytochemistry revealed that necdin accumulated in the nucleus of differentiated neurons that showed neurite extension and expressed the neuronal markers microtubule-associated protein 2 and synaptophysin. Suppression of necdin expression in DRG cultures treated with antisense oligonucleotides led to a marked reduction in the number of terminally differentiated neurons. The antisense oligonucleotide-treated cells did not attempt to reenter the cell cycle, but underwent death with characteristics of apoptosis such as caspase-3 activation, nuclear condensation, and chromosomal DNA fragmentation. Furthermore, a caspase-3 inhibitor rescued antisense oligonucleotidetreated cells from apoptosis and significantly increased the population of terminally differentiated neurons. These results suggest that necdin mediates the terminal differentiation and survival of NGF-dependent DRG neurons and that necdin-deficient nascent neurons are destined to caspase-3-dependent apoptosis. © 2002 Elsevier Science (USA) Key Words: necdin; cell cycle; postmitotic neurons; nerve growth factor; terminal differentiation; apoptosis; caspase-3; antisense oligonucleotide; Prader–Willi syndrome.

INTRODUCTION

In the neural tube of vertebrate embryos, neuroepithelial stem cells proliferating in the ventricular zones 1

R.T. and I.N. contributed equally to this work. To whom correspondence and reprints requests should be addressed. Fax: ⫹81-66879-8623. E-mail: [email protected]. 2

0014-4827/02 $35.00 © 2002 Elsevier Science (USA) All rights reserved.

commit to a neural fate and undergo terminal mitosis. Thereafter neurons withdraw permanently from the cell cycle and enter the postmitotic state to express a terminally differentiated phenotype. Although such permanent exit from the cell cycle is the most fundamental phenomenon displayed by neurons, little is known about the molecular mechanisms by which neurons enter the postmitotic state and become terminally differentiated. The murine embryonal carcinoma P19 cells differentiate into postmitotic neurons in response to retinoic acid treatment [1]. Necdin is a 325-aminoacid-residue protein encoded in a cDNA sequence isolated from a subtraction library of neurally differentiated P19 cells [2]. The necdin gene is expressed in most of the terminally differentiated neurons, although its expression levels vary among neuronal cell types, being most abundant in the hypothalamus [3, 4]. Necdin shows a slight homology to the MAGE family proteins which are expressed in melanoma cells and act as antigens recognized by cytolytic T lymphocytes[5]. Most of the MAGE family genes are located on the X chromosome, whereas the human necdin gene NDN is mapped to chromosome 15q11.2– q12, a region deleted in Prader–Willi syndrome (PWS) [6 – 8]. PWS is a neurogenetic disorder related to genomic imprinting, and its major symptoms such as feeding problems, gross obesity, and hypogonadism are consistent with a hypothalamic defect. NDN is maternally imprinted and transcribed only from the paternal allele [6, 7, 9]. Necdin is not expressed in the cells prepared from PWS patients whose chromosome 15q11.2– q12 region in the paternal allele is deleted. Disruption of the mouse necdin gene results in early postnatal lethality [10], reduction in specific hypothalamic neurons such as LHRH- and oxytocin-containing neurons, and behavioral alterations, which are reminiscent of human PWS [11]. These observations suggest that necdin is involved in the differentiation and development of neurons in the brain, especially in the hypothalamus. Several lines of evidence have suggested that necdin is a growth suppressor expressed in postmitotic neurons. Ectopic expression of necdin suppresses the proliferation of several cell lines [12–14]. Furthermore,

220

NECDIN IN NEURONAL TERMINAL DIFFERENTIATION

necdin interacts with viral oncoproteins such as SV40 large T antigen, adenovirus EIA, and the transcription factor E2F1 [13]. These characteristics of necdin resemble those of the retinoblastoma gene product (Rb), a major growth suppressor protein that is mutated in various cancer cells. Rb is believed to play a pivotal role in terminal mitosis and subsequent differentiation during neurogenesis [15]. These findings prompted us to study whether necdin is also involved in neuronal differentiation. Using primary dorsal root ganglion (DRG) cells and antisense oligonucleotides, we here demonstrate that necdin is involved in neuronal terminal differentiation and survival dependent on nerve growth factor (NGF). MATERIALS AND METHODS Culture of DRG cells. Primary dissociated neuronal cultures of DRG were prepared from ICR mice (SLC, Shizuoka, Japan) at embryonic day 12.5 (E12.5) using the method reported previously [16]. The ganglia were dissected under aseptic conditions, placed in L-15 medium (Life Technologies), and incubated for 5 min at 37°C with 0.05% trypsin in Ca 2⫹/Mg 2⫹-free Hank’s balanced salt solution. After removal of the trypsin solution, the ganglia were washed with Ham’s F-14 medium (Imperial Laboratories, Andover, U.K.) containing 10% fetal calf serum (FCS) and were gently triturated with a fire-polished Pasteur pipette to give a single-cell suspension. The cells were plated at a density of 4 ⫻ 10 4 cells per well in Lab-tek II 8-well Chamber Slides (Nunc) precoated with poly-DL-ornithine and laminin. DRG cells were cultured at 37°C in a humidified 5% CO 2 incubator in a defined medium consisting of Ham’s F-14 medium and supplements reported previously [16]. DRG cells were incubated in the presence of 2.5S mouse NGF (50 ng/ml) (Invitrogen) which was added 1 h after plating. Neurite extension was analyzed by differential interference contrast (DIC) microscopy. Fluorescence immunocytochemistry. Anti-necdin polyclonal antibody (MNF) was raised in a New Zealand rabbit against purified poly-histidine (His)-tagged necdin protein [17]. DRG cells were fixed with 4% formaldehyde in phosphate-buffered saline (PBS) (pH 7.4) at 4°C for 20 min and permeabilized with acetone at ⫺20°C for 20 min. Fixed cells were incubated with the primary antibodies MNF (1:1000), mouse monoclonal anti-microtubule-associated protein 2 (MAP2) antibody (Chemicon) (1:500), and mouse monoclonal antisynaptophysin antibody (Sigma) (1:1000) in PBS containing 0.05% Tween 20 and 5% normal goat serum at 4°C overnight. The cells were then incubated at room temperature for 90 min with anti-rabbit and anti-mouse IgGs conjugated with fluorescein isothiocyanate (FITC) (Cappel) (1:500) and rhodamine (Cappel) (1:500), respectively. For chromosomal DNA staining, cells were incubated with 5 ␮M Hoechst 33342 (Sigma) in PBS at room temperature for 10 min. Fluorescent images were observed with a fluorescence DIC microscope (BX60-34-FLAD1, Olympus). Images were taken by CCD camera system (M-3204C, Olympus) and processed using Adobe Photoshop 5.0 software. For microscopic photometry of nuclear necdin immunoreactivity, 12-bit digital monochrome images of necdin immunoreactivity and nuclear DNA were captured with a CCD camera (CoolSNAP monochrome; Nippon Roper, Chiba, Japan) and fluorescence intensity of necdin in the nuclear area detected by DNA staining was measured using fluorescence image analysis software (Fluoroimage Cool V, Mitani Co., Osaka, Japan). Antisense oligodeoxynucleotide treatment. Phosphorothioate oligodeoxynucleotides corresponding to mouse necdin cDNA subsequences (nucleotide positions ⫺14 to ⫹11 for Oligo A, ⫹444 to ⫹463 for Oligo B; translation start position ⫽ ⫹1) [2] were synthesized

221

by Sawady Technology (Tokyo, Japan): antisense Oligo A, 5⬘TGTTCCGACATGTCTGCGCCTTCGG-3⬘, sense Oligo A, 5⬘CCGAAGGCGCAGACATGTCGGAACA-3⬘; antisense Oligo B, 5⬘-ACACTCTGGCGAGGATGACG-3⬘, sense Oligo B, 5⬘-CGTCATCCTCGCCAGAGTGT-3⬘. Oligonucleotides (1 ␮M) were mixed with 1 ␮l of OligofectAMINE (invitrogen) in 40 ␮l medium and sat for 15 min at room temperature. DRG cells were trypsinized and dissociated cells were suspended at 4 ⫻ 10 4 cells in 160 ␮l medium. The cell suspension was mixed with the oligonucleotide solution and remained on ice for 30 min (the final concentration of each oligonucleotide was 200 nM). The cells were plated on the culture slides, and 1 h later NGF (50 ng/ml) was added to the cultures. DRG cells were incubated for 24 h and examined by immunocytochemistry. Cell cycle reentry analyses. For the bromodeoxyuridine (BrdU) incorporation assay, DRG cells were treated with oligonucleotides, incubated in the presence of 10 ␮M BrdU over the culture period (24 h), and triply stained for BrdU with an anti-BrdU antibody (BrdU labeling and detection Kit II; Boehringer Mannheim), DNA with Hoechst 33342, and MAP2. For immunocytochemical detection of G1 cyclins, DRG cells were treated with antisense oligonucleotide and NGF, fixed 12 h later, and immunostained with mouse monoclonal anti-cyclin cyclin D1 antibody (A-12, Santa Cruz) (1:500) and rabbit polyclonal anti-cyclin E antibody (M-20, Santa Cruz) (1:500) as described above. Primary cultures of mouse embryonic fibroblasts (MEF) were prepared from mouse E12.5 embryo, fixed during the exponential growth, and immunostained as a positive control for G1 cyclins. Cell death analyses. DRG cells were treated with oligonucleotides, and 24 h later the cells were incubated with ethidium homodimer (EthD) (Molecular Probes, Eugene, OR) diluted 1:2000 in PBS at 37°C for 30 min. After fixation, the cells were incubated with 5 ␮M Hoechst 33342 (Sigma) in PBS at room temperature for 10 min. EthD-positive cells among 100 –200 Hoechst 33342-positive cells were counted by choosing 10 random fields per each well and expressed as a percentage [18]. Nuclear DNA fragmentation was analyzed by terminal deoxynucleotidyl transferase-mediated dUTP-biotin nick-end labeling (TUNEL) [19, 20] visualized with Texas red. The number of TUNEL-positive cells was expressed as a percentage of the cells stained with Hoechst 33342. Cells were immunocytochemically stained for activated (cleaved) caspase-3 p20/p17 subunits with ACP3 [18]. DRG cells were cultured in the absence of NGF and used as an undifferentiated control for activated caspase-3 immunostaining. For detection of activated caspase-3 in apoptotic neurons, DRG cells were treated with NGF for 36 h, deprived of NGF, and further incubated in the absence of NGF for 24 h. For caspase-3 inhibition analysis, DRG neurons were incubated in the presence of 150 ␮M carbobenzoxy-Asp-Glu-Val-Asp-fluoromethylketone (ZDEVD-FMK) (Caspase-3 Inhibitor II; Calbiochem) or DMSO vehicle (control) in combination with Oligo A or OligofectAMINE (for control), cultured for 24 h, and examined for neuronal differentiation and cell death. Statistical tests. One-way analysis of variance (ANOVA) was used to test for overall statistical significance. If a significant overall effect of treatment was observed, post hoc comparisons between group means were made with Fisher’s Least Significant Difference Test. A significance of P ⬍ 0.05 was required for rejection of the null hypothesis.

RESULTS

Necdin Accumulates in the Nucleus of Terminally Differentiated Neurons in Response to NGF Treatment We have previously reported that necdin mRNA is expressed in postmitotic neurons that are distributed

222

TAKAZAKI, NISHIMURA, AND YOSHIKAWA

throughout the central and peripheral nervous system [4]. By in situ hybridization histochemistry, we detected high levels of necdin mRNA in DRG cells in vivo at E12.5, a period when many neurons in DRG undergo terminal differentiation and death (data not shown). Therefore, we examined the cellular and subcellular distribution of the necdin protein in DRG cells using primary cultures isolated from E12.5 mouse embryos. Endogenous levels of necdin immunoreactivity and its localization were analyzed during the course of differentiation induced by NGF (Fig. 1). Most of the dissociated DRG cells were round in shape and had no processes prior to NGF treatment (O h)(Fig. 1A). NGFtreated cells extended their processes and became neuron-like cells. These cells were relatively small, and their somata were 10 –20 ␮m in diameter. The time course study revealed that ⬃60% of NGF-treated cells had neurites with a length ⬎5 times their cell diameter during the 24 to 48-h incubation period (Fig. 1B). We then examined necdin expression by immunocytochemistry using an antibody (MNF) against full-length necdin. This antibody positively stained the nucleus of neurally differentiated P19 cells, a cell line from which necdin was isolated [2], and necdin cDNA-transfected transformed cell lines (data not shown). Low levels of necdin immunoreactivity were present in the round cells at 0 h, but the immunoreactivity increased gradually and accumulated in the nucleus of neurons carrying long processes (Fig. 1A, b, e, h, and k). In the absence of NGF, neither neurite extension nor nuclear accumulation of necdin immunoreactivity was observed (data not shown). We quantified necdin immunoreactivity in the nucleus of differentiated DRG neurons by measuring the nuclear fluorescence intensity (Fig. 1C). The immunoreactivity was gradually increased to the maximum of ⬃6 times the 0-h level during the course of NGF-dependent terminal differentiation. We next examined the expression of neuronal marker proteins such as MAP2 and synaptophysin (Fig. 2). DRG cells initially contained very low levels of MAP2 and synaptophysin immunoreactivities (Figs. 2A and 2G), both of which increased in perikarya and neurites of differentiated neurons by NGF treatment for 24 h (Figs. 2D and 2J). The cells that strongly expressed these neuronal markers contained high levels of necdin in the nucleus (Figs. 2E and 2K). Thus, accumulation of necdin in the nucleus of postmitotic neurons may be associated with terminal differentiation. Necdin Antisense Oligonucleotides Abrogate NGFDependent Terminal Differentiation We examined the effects of necdin antisense oligonucleotides on differentiation of DRG neurons. Antisense

and sense oligodeoxynucleotides corresponding to two different regions, ⫺14 to ⫹11 (Oligo A) and ⫹444 to ⫹463 (Oligo B), of the mouse necdin cDNA sequence were synthesized (Fig. 3A). When DRG neurons were induced to differentiate by NGF in the presence of antisense Oligo A, necdin immunoreactivity in the nucleus was markedly diminished at 24 h (Fig. 3B, d). The antisense oligonucleotide-treated cells lost morphological features of differentiated neurons and had a round and shrunken soma without neurites (Fig. 3B, f). The nuclei of these cells were slightly shrunken and condensed, but neither severe condensation nor fragmentation of the nucleus was observed (Fig. 3B, e). These morphological changes were also observed in DRG neurons treated with antisense Oligo B (data not shown). The number of cells displaying neurite outgrowth was markedly reduced in cultures treated with antisense Oligo A or antisense Oligo B (Fig. 3C). We then examined whether antisense Oligo A treatment abrogates NGF-induced neuronal differentiation by immunocytochemistry of MAP2 and synaptophysin (Fig. 4A). Immunoreactivities of MAP2 and synaptophysin in antisense Oligo A-treated cells were very low even in the presence of NGF, suggesting that reduction of endogenous necdin expression blocks NGF-induced differentiation. The number of differentiated neurons was significantly reduced by antisense Oligo A as judged by the expression of MAP2 and synaptophysin (Fig. 4A, b and h). A similar reduction in the expression of these proteins was observed in the DRG cells treated with antisense Oligo B (data not shown). Quantification by immunocytochemistry revealed that both antisense Oligo A and antisense Oligo B significantly reduced the number of DRG cells immunoreactive for MAP2 and synaptophysin (Fig. 4B). Rb-deficient DRG cells, which fail to differentiate into neurons, reenter S phase prior to apoptosis [21]. We thus examined whether these antisense Oligo Atreated cells reenter the cell cycle. In control and sense Oligo A-treated cultures incubated in the presence of BrdU throughout the 24-h incubation period, BrdU was incorporated into some proliferating nonneuronal cells, but not into MAP2-immunopositive neurons (Fig. 5A, a, c, d, f ). Under the same conditions, antisense Oligo A-treated undifferentiated cells did not incorporate BrdU (Fig. 5A, b). We examined whether antisense Oligo A-treated cells express cyclin D1 and cyclin E, both of which are G1 cyclins and serve as early markers for cell cycle reentry from G0 phase. Under the conditions in which proliferative fibroblasts were positively stained for these cyclins, neither cyclin D1 nor cyclin E was detected in antisense Oligo A-treated cells (Fig. 5B). These results suggest that DRG neurons do not attempt to enter the cell cycle even when endogenous necdin expression is repressed.

NECDIN IN NEURONAL TERMINAL DIFFERENTIATION

223

FIG. 1. Nuclear accumulation of necdin immunoreactivity in differentiating DRG neurons in vitro. (A) Fluorescence immunocytochemistry of necdin in DRG cells during terminal differentiation. DRG cells prepared from E12.5 mouse embryo were cultured for 48 h in the presence of NGF, fixed at the indicated times, and observed by DIC microscopy (DIC; a, d, g, j). Fixed cells were double-stained for necdin (Necdin; b, e, h, k) and DNA with Hoechst 33342 (DNA; c, f, i, l). Arrows point to differentiating DRG cells. Scale bar, 10 ␮m (in a for a–l). (B) Neurite outgrowth of DRG neurons induced by NGF. Cells with long neurites (⬎5 times the cell diameter) were counted among ⬎100 cells stained with Hoechst 33342. Each value (%) is the mean ⫾ SEM of three separate experiments. (C) Nuclear accumulation of necdin. Necdin was detected by fluorescence immunocytochemistry, and intensity of necdin immunoreactivity within the nuclear area was measured. Each point (relative intensity) represents the mean ⫾ SEM (⬎50 cells at each time point).

Necdin Antisense Oligonucleotide Induces Cell Death To test the viability of antisense oligonucleotidetreated cells, we used ethidium homodimer (EthD) uptake analysis. We found that a large number of anti-

sense Oligo A-treated cells retained EthD (Fig. 6A, left), which is excluded from viable cells. The number of dead cells in the antisense Oligo A-treated group was significantly larger than that of the sense Oligo A-

224

TAKAZAKI, NISHIMURA, AND YOSHIKAWA

larger than that of the sense Oligo A-treated group (Fig. 6C, right). Failure of Terminal Differentiation of NecdinDeficient Cells Induces Caspase-3-Dependent Apoptosis

FIG. 2. Accumulation of necdin in terminally differentiated DRG neurons. DRG cultures prepared from E12.5 mouse embryos were incubated in the presence of NGF for 24 h (24 h), fixed, and stained for MAP2 (MAP2; A, D) or synaptophysin (SYP; G, J) in combination with necdin (Necdin; B, E, H, K). Control cells (0 h) were fixed before NGF addition. Cells were visualized by DIC microscopy (DIC; C, F, I, L). Arrows point to undifferentiated (A–C for MAP2, G–I for SYP) and differentiated (D–F for MAP2, J–L for SYP) DRG neurons. Scale bar, 10 ␮m (in A for A–L).

treated group (Fig. 6B, right). Nuclear morphology showed no typical apoptotic characteristics such as pyknosis and fragmentation. We therefore tested TUNEL for nuclear DNA fragmentation, a typical DNA change seen in apoptosis. The nuclei of antisense Oligo A-treated cells were positively stained for TUNEL (Fig. 6B, left). The number of TUNEL-positive cells in the antisense Oligo A-treated group was much larger than that of TUNEL-positive cells in the sense Oligo Atreated group (Fig. 6B, right). Caspases are cystein proteases that mediate apoptosis. Of caspase family members, caspase-3 is a major effector caspase to execute apoptosis. Thus, we examined whether caspase-3 is activated in antisense Oligo A-treated DRG cells (Fig. 6C). Owing to the limited amounts of cultured cells, it was difficult to demonstrate the cleavage-dependent activation of caspase-3 by Western blotting. We alternatively used an end-specific antibody against cleaved procaspase-3 (Fig. 6C, left). The population of activated caspase-3-positive cells treated with antisense Oligo A was ⬃30%, which was significantly

We analyzed MAP2 expression in caspase-3-activated DRG cells by double immunostaining. In antisense Oligo A-treated DRG cells, intensely caspase-3immunopositive cells had condensed nuclei and contained little or no MAP2 immunoreactivity (Fig. 7A, a– c), whereas MAP2 was abundantly expressed in sense Oligo A-treated cells (Fig. 7A, d–f ). MAP2 immunoreactivity in activated caspase-3-immunopositive cells was also undetectable in undifferentiated DRG cells without NGF treatment (Fig. 7A, g–i). Thus, MAP2 expression and caspase-3 activation in antisense Oligo A-treated DRG cells are similar to those in undifferentiated DRG cells. To exclude the possibility that reduced MAP2 levels in antisense Oligo A-treated DRG cells are a consequence of proteolysis by caspase-3 activation and apoptosis, we examined MAP2 expression and caspase-3 activation in MAP2positive DRG neurons deprived of NGF, a typical model of neurotrophin deprivation-induced neuronal apoptosis. Following NGF deprivation, DRG neurons contained activated caspase-3 and a fragmented nucleus, but still retained large amounts of MAP2 immunoreactivity (Fig. 7B, a– c), similar to the levels of control DRG neurons incubated in the presence of NGF (Fig. 7B, d–f ). This suggests that MAP2 is resistant to proteolytic degradation in apoptotic DRG neurons and that the absence of MAP2 in antisense Oligo A-treated DRG cells is not a consequence of its proteolytic degradation. Suppression of Apoptosis Restores Neuronal Differentiation in Necdin-Deficient DRG Neurons We then quantified the DRG cells containing MAP2 and activated caspase-3 during the course of oligonucleotide treatment. The number of MAP2-immunopositive cells treated with antisense Oligo A remained very small during the 6- to 24-h incubation period, whereas the number of sense Oligo A-treated cells containing MAP2 markedly increased (Fig. 8A, left). In antisense Oligo A-treated cultures, only a few cells contained activated caspase-3 during the period 0 – 6 h but the population was markedly increased to ⬃50% at 12 h (Fig. 8A, right). These findings suggest that failure of differentiation triggers the caspase pathway in antisense Oligo A-treated cells. We next examined whether antisense Oligo A-treated cells stay undifferentiated or terminally differentiated when apoptosis is blocked by a caspase-3 inhibitor (Fig. 8B). In the absence of the cell-permeable caspase-3 inhibitor

NECDIN IN NEURONAL TERMINAL DIFFERENTIATION

225

FIG. 3. Morphological changes in necdin-deficient DRG cells. (A) Antisense oligonucleotides used for suppression of endogenous necdin expression. The positions and sequences of antisense Oligo A and Oligo B are shown. S, transcription start site; poly(A), polyadenylation site. (B) Morphological changes of DRG cells after treatment with antisense Oligo A. Suspended DRG cells were treated with OligofectAMINE alone for control (CT; a, b, c), antisense Oligo A (AO; d, e, f ), or sense Oligo A (SO; g, h, i), plated onto culture slides, and incubated in the presence of NGF for 24 h. Cells were double-stained for necdin (Necdin; a, d, g) and DNA with Hoechst 33342 (DNA; b, e, h). Cells were also visualized by DIC microscopy (DIC; c, f, i). Arrows point to representative DRG neurons. Scale bar, 10 ␮m (in a for a–i). (C) Quantification of cells with extended neurites after treatment of antisense oligonucleotides. DRG cells treated with oligonucleotides were incubated in the presence of NGF as above, and cells with extended neurites (⬎5 times the cell diameter) among ⬎100 cells (stained with Hoechst 33342) were counted. The value of the SO or AO group was divided by that of the CT group. Each value (%) represents the mean ⫾ SEM of three separate experiments. *P ⬍ 0.0005.

Z-DEVD-FMK, many antisense Oligo A-treated cells were positively stained for activated caspase-3 (Fig. 8B, left) and had TUNEL-positive nuclei (Fig. 8B, right). In the presence of the inhibitor, the number of cells containing activated caspase-3 or TUNEL-positive nuclei was significantly reduced. The caspase inhibitor did not alter the number of activated caspase-

3-positive or TUNEL-positive cells in the sense Oligo A-treated group. On the other hand, the caspase-3 inhibitor significantly increased the population of differentiated neurons that exhibited neurite extension (Fig. 8C, left) and MAP2 expression (Fig. 8C, right). These data suggest that suppression of apoptosis restores neuronal differentiation in necdin-deficient DRG cells.

226

TAKAZAKI, NISHIMURA, AND YOSHIKAWA

FIG. 4. Impaired neuronal differentiation of necdin-deficient DRG cells. (A) Expression of MAP2 and synaptophysin in oligonucleotidetreated DRG cultures. DRG cells treated with OligofectAMINE alone for control (CT; a, d, g, j), antisense Oligo A (AO; b, e, h, k) or sense Oligo A (SO; c, f, i, l), plated onto culture slides, and incubated in the presence of NGF for 24 h. Cells were stained for MAP2 (MAP2; a– c) or synaptophysin (SYP; g–i) and visualized by DIC microscopy (DIC; d–f for MAP2, j–l for SYP). Arrows point to representative DRG neurons. Scale bar, 10 ␮m (in a for a–l). (B) Quantification of MAP2- and synaptophysin-immunopositive cells after oligonucleotide treatment. DRG cells treated with oligonucleotides were incubated in the presence of NGF for 24 h as above, and MAP2-positive cells (left) or synaptophysin (SYP)-positive cells (right) among ⬎100 cells (stained with Hoechst 33342) were counted. The value of the SO or AO group was divided by that of the CT group. Each value (%) represents the mean ⫾ SEM of three separate experiments. *P ⬍ 0.005.

DISCUSSION

The present study has provided direct evidence for the involvement of necdin in neuronal terminal differentiation. The two antisense oligonucleotides corresponding to distinct necdin subsequences blocked NGF-dependent terminal differentiation of DRG sensory neurons. Necdin functions as a strong growth suppressor that interacts with E2F1 and suppresses transcription driven by E2F1 [13]. Necdin also functions as a transcription factor that directly binds to a specific G-rich DNA motif [17]. Several transcription

factors involved in the differentiation of specific cell types suppress E2F1 activity. For example, Rb is a typical E2F1-interacting growth suppressor involved in terminal mitosis during neurogenesis [22, 23]. In addition, the lineage-instructive transcription factor C/EBP␣ suppresses E2F1 activity and induces both growth arrest and terminal differentiation of adipocytes and granulocytes [24]. These findings suggest that terminal differentiation induced by these growth suppressors is associated with their suppressive effects on E2F activity. Thus, it is possible that necdin mediates terminal differentiation through suppression of

NECDIN IN NEURONAL TERMINAL DIFFERENTIATION

FIG. 5. Failure of cell cycle reentry of necdin-deficient cells. (A) BrdU labeling analysis. Suspended DRG cells were treated with OligofectAMINE alone for control (CT; a, d, g, j), antisense Oligo A (AO; b, e, h, k), and sense Oligo A (SO; c, f, i, l), plated onto culture slides, and incubated for 24 h in the presence of NGF and BrdU. Cells were fixed and stained for BrdU (BrdU; a– c), MAP2 (MAP2; d–f ), and DNA (DNA; g–i) and observed by fluorescence DIC microscopy (DIC; j–l). In the CT and SO groups, differentiated MAP2-positive DRG neurons (arrows in d and f ) are BrdU-negative (arrows in a and c), whereas some MAP2-negative nonneuronal cells (arrowheads in d and f) are BrdU-positive (arrowheads in a and c). In the AO group, undifferentiated shrunken cells (arrows in e, h, and k) are BrdU-negative (arrows in b). Scale bar, 20 ␮m (in a for a–l). (B) Immunocytochemical analysis of G1 cyclins. DRG cells treated with antisense Oligo A (AO-treated DRG cells) for 12 h and mouse embryonal fibroblasts (MEF) were immunostained for cyclin D1 or cyclin E. These cyclins, which are highly expressed in proliferating MEF (arrows in a, d), are absent from AO-treated DRG cells (arrows in b, c, e, f ). Scale bar (in a); 20 ␮m (for a, d), 10 ␮m (for b, c, e, f ).

227

E2F1 activity. Another possibility is that necdin induces terminal differentiation by a mechanism distinct from that of growth suppression. Necdin binds to several proteins located in both the cytoplasm and the nucleus of differentiated neurons [13, 14, 25, 26], indicating that necdin maintains terminal differentiation by interacting with these proteins. NGF acts on nascent neurons expressing the highaffinity NGF receptor TrkA, which is expressed in postmitotic neurons differentiating at late stages of DRG development [27]. By the BrdU incorporation assay, we confirmed that NGF-dependent neurons are mitotically inactive (Fig. 5). Thus, NGF promotes terminal differentiation of nascent postmitotic neurons in the present in vitro system. Our present data that necdin accumulated in such growth-arrested neurons (Fig. 2) indicate that necdin contributes to the stabilization of mitotic quiescence of neurons. On the other hand, several lines of evidence have suggested that Rb is involved in terminal mitosis during neurogenesis [28 – 30]. The absence of Rb during in vivo neurogenesis in DRG induces neuronal apoptosis shortly after their entrance into S phase and causes a marked reduction in the expression of the neurotrophin receptors TrkA and TrkB and the low-affinity receptor p75 [21]. These results suggest that Rb regulates both the exit from the cell cycle of neuronal progenitors and the differentiation of DRG neurons. We infer that mitotic quiescence displayed by postmitotic DRG sensory neurons is accomplished by the cooperation of multiple growth suppressors such as necdin, Rb, and p130 [31]. This may account for the present finding that suppression of endogenous necdin alone by the antisense oligonucleotide failed to induce their aberrant entry into the cell cycle in nascent neurons (Fig. 5). Recently it was reported that NRAGE, a necdin homologous MAGE family protein, interacts with p75 [32], which has an important role in the development and function of sensory neurons [33]. Intriguingly, NRAGE (Dlxin-1), like necdin, exerts growth suppressive effects [32, 34]. Thus, we speculate that necdin and its homologous proteins mediate growth arrest and terminal differentiation of sensory neurons. The present study has also shown that suppression of endogenous necdin expression by antisense oligonucleotides induces apoptosis, indicating that failure of terminal differentiation is closely associated with cell death. Necdin binds to two major transcription factors, E2F1 and p53, to repress their transcriptional activities [13, 14]. Previous studies have suggested that E2F1 and p53 serve as apoptotic inducers in differentiated neurons [35–39]. Thus, it is possible that necdin suppresses apoptosis through inhibition of these apoptosis-inducing transcription factors. Further studies on the caspase-3 activation in necdin-deficient cells may

228

TAKAZAKI, NISHIMURA, AND YOSHIKAWA

FIG. 6. Death of necdin-deficient DRG cells. (A) EthD assay. Suspended DRG cells were treated with antisense Oligo A (AO; a, c) or sense Oligo A (SO; b, d), plated onto culture slides, and incubated for 24 h in the presence of NGF. Cell death was analyzed by EthD retention (EthD; a, b) and DIC (DIC; c, d) (left). Arrows point to positive cells. Scale bar, 10 ␮m (in a for a– d). Positive cells among ⬎100 cells (stained with Hoechst 33342) were counted in each group (right). Each value (%) represents the mean ⫾ SEM of three separate experiments. *P ⬍ 0.01. (B) TUNEL analysis. (C) Activated caspase-3 immunocytochemistry. Cells were treated with antisense Oligo A (AO; a, c) or sense Oligo A (SO; b, d), incubated in the presence of NGF for 24 h, and analyzed by the TUNEL method (B) and activated caspase-3 immunocytochemistry (C). The data are presented as in A. **P ⬍ 0.002 for B, C (right).

elucidate the mechanism that links terminal differentiation and survival of DRG neurons. We found that MAP2 expression was undetectable in antisense-treated DRG cells, in which caspase-3 was activated (Fig. 7). This raises the possibility that MAP2 undergoes proteolytic degradation in apoptotic cells in which caspases and other proteases can be activated. However, we have previously found that MAP2 is resistant to proteolytic degradation in apoptotic neurons and serves as a useful marker of neuronal apoptosis [18, 20]. In the present study, we confirmed that NGFdeprived apoptotic DRG neurons contained high MAP2 levels (Fig. 7B). Therefore, undetectable MAP2 in an-

tisense oligonucleotide-treated DRG cells may not be a consequence of apoptosis, but may be due to suppression of endogenous MAP2 expression in undifferentiated cells. Furthermore, the time course analysis of MAP2 expression and caspase-3 activation revealed that MAP2 expression was impaired prior to caspase-3 activation (Fig. 8A). This also supports the view that failure to differentiate is not a consequence of apoptosis. On the other hand, caspase-3 inhibition rescued necdin-deficient DRG cells from apoptosis and increased the population of differentiated neurons (Figs. 8B and 8C). This suggests that cells escaping from apoptosis are incapable of staying in an undifferenti-

NECDIN IN NEURONAL TERMINAL DIFFERENTIATION

229

FIG. 7. MAP2 expression in apoptotic DRG cells. (A) Cleavage-dependent activation of caspase-3 in oligonucleotide-treated DRG cells. DRG cells were treated with antisense Oligo A (AO; a, b, c) or sense Oligo A (SO; d, e, f), plated onto culture slides, and incubated for 24 h in the presence of NGF. DRG cells were also incubated for 12 h in the absence of NGF (undifferentiated control) (UD). Cells were triply stained for MAP2 (MAP2; a, d, g), activated caspase-3 (Casp3 act; b, e, h), and DNA (DNA; c, f, i). Arrows in AO (a– c) and UD (g–i) groups point to MAP2-negative, Casp3 act-positive cells. In the SO group (d–f), arrows point to MAP2-positive, Casp3 act-negative cells. Scale bar, 10 ␮m (in a for a–i). (B) MAP2 expression in apoptotic DRG neurons deprived of NGF. DRG cells were induced to differentiate by incubation with NGF for 36 h, deprived of NGF, and incubated for an additional 24 h in the presence (NGF⫹) or absence (NGF⫺) of NGF. Cells were triply stained as in A. Arrows in the NGF- group (a– c) point to MAP2-positive, Casp3 act-positive cells. In the NGF⫹ group (d–f ), arrows point to MAP2-positive, Casp3 act-negative cells. Scale bar, 10 ␮m (in a for a–f ).

ated state and that inhibition of apoptosis activates necdin-independent compensatory mechanisms to restore differentiation. This is similar to the previous report that DRG neurons in Rb-knockout mice undergo apoptosis but caspase-3 knockout in these mutant mice rescues DRG neurons from death and normalizes neuronal differentiation [40]. We infer that nascent neurons are destined to either terminal differentiation or death and that they cannot remain undifferentiated at their terminal stage. This feature is distinct from that of transformed neuronal cell lines (e.g., neuroblastoma and pheochromocytoma), which are normally viable in an undifferentiated state. We have attempted to examine the effects of necdin

antisense oligonucleotides on the differentiation and viability of primary cultured cells exhibiting neurogenesis such as cortical neuronal progenitors and neuroepithelial stem cells. However, we failed to demonstrate appreciable effects of necdin on these rapidly proliferating cells (K.Y., unpublished observations). The present system using DRG neurons prepared from E12.5 mouse embryo is suitable for the analyses of necdin’s functions in neuronal terminal differentiation and survival, because terminal differentiation and nuclear necdin accumulation show similar kinetics in response to NGF. Moreover, these DRG sensory neurons are less vulnerable to oligonucleotide treatment than neuronal progenitors and stem cells.

230

TAKAZAKI, NISHIMURA, AND YOSHIKAWA

FIG. 8. Effects of caspase-3 inhibition on death and differentiation of necdin-deficient DRG cells. (A) Time course analysis of MAP2- and activated caspase-3-positive cell populations. DRG cells were treated with antisense Oligo A (AO) or sense Oligo A (SO), plated onto culture slides, and incubated in the presence of NGF. Cells were fixed at the indicated times and stained for MAP2 and activated caspase-3 (Casp3 act). MAP2-positive cells (left) and Casp3 act-positive cells (right) among ⬎100 cells stained with Hoechst 33342 were counted. Each value (%) represents the mean ⫾ SEM of three separate experiments. (B) Effects of caspase-3 inhibition on the numbers of Casp3 act-positive and TUNEL-positive cells. DRG cells were treated with antisense or sense Oligo A, incubated in the presence (Inhibitor⫹) or absence (Inhibitor⫺) of caspase-3 inhibitor Z-DEVD-FMK in combination with NGF, and fixed at 12 and 24 h for Casp3 act and TUNEL, respectively. Casp3 actpositive cells (left) and TUNEL-positive cells (right) among ⬎100 cells were counted. Each value (%) represents the mean ⫾ SEM of three separate experiments. *P ⬍ 0.005. (C) Effects of caspase-3 inhibition on neuronal differentiation. DRG cells were treated as above and fixed at 12 h for analyses of neurite extension and MAP2. Cells with extended neurites (⬎5 times the cell diameter) (left) and MAP2-positive cells (right) among ⬎100 cells were counted. Each value (%) represents the mean ⫾ SEM of three separate experiments. *P ⬍ 0.02.

Disruption of the mouse necdin gene in the paternal allele results in postnatal lethality [10]. Another model of necdin-deficient mouse exhibits the reduction in specific hypothalamic neurons such as LH-RH- and oxytocin-containing neurons [11]. A higher incidence of abnormal phenotypes is observed only in the C57BL/6

background. The strain-dependent abnormalities of these necdin-deficient mice suggest the existence of necdin modifier genes. Although no morphological abnormalities of DRG neurons in vivo in these necdin knockout mice have been reported to date, it might be interesting to analyze the differentiation in vitro of

NECDIN IN NEURONAL TERMINAL DIFFERENTIATION

primary DRG neurons prepared from these mutant mice. This may help establish whether the events documented in this study are indeed attributable to the effect of necdin deficiency or additional perturbations required for apoptosis. Furthermore, primary DRG neurons prepared from the knockout mice may be useful for identifying the modifier genes that compensate for necdin deficiency. These studies using DRG neurons prepared from necdin knockout mice are under way in our laboratory. The present study raises the possibility that the absence of necdin expression in patients with PWS causes abnormal development of DRG sensory neurons. It has previously been reported that impaired peripheral somatosensory functions due to a reduced number of normal axons are observed in children with PWS [41]. The system of primary NGF-dependent DRG sensory neurons represents a useful model for the analyses of the molecular mechanisms of necdin-associated neuronal differentiation and of its abnormality in PWS.

231

neuronal growth suppressor, in the Prader–Willi syndrome deletion region. Gene 213, 65–72. 9.

Sutcliffe, J. S., Han, M., Christian, S. L., and Ledbetter, D. H. (1997). Neuronally-expressed necdin gene: An imprinted candidate gene in Prader–Willi syndrome. Lancet 350, 1520 –1521.

10.

Gerard, M., Hernandez, L., Wevrick, R., and Stewart, C. L. (1999). Disruption of the mouse necdin gene results in early post-natal lethality. Nat. Genet. 23, 199 –202.

11.

Muscatelli, F., Abrous, D. N., Massacrier, A., Boccaccio, I., Moal, M. L., Cau, P., and Cremer, H. (2000). Disruption of the mouse Necdin gene results in hypothalamic and behavioral alterations reminiscent of the human Prader–Willi syndrome. Hum. Mol. Genet. 9, 3101–3110.

12.

Hayashi, Y., Matsuyama, K., Takagi, K., Sugiura, H., and Yoshikawa, K. (1995). Arrest of cell growth by necdin, a nuclear protein expressed in postmitotic neurons. Biochem. Biophys. Res. Commun. 213, 317–324.

13.

Taniura, H., Taniguchi, N., Hara, M., and Yoshikawa, K. (1998). Necdin, a postmitotic neuron-specific growth suppressor, interacts with viral transforming proteins and cellular transcription factor E2F1. J. Biol. Chem. 273, 720 –728.

14.

Taniura, H., Matsumoto, K., and Yoshikawa, K. (1999). Physical and functional interactions of neuronal growth suppressor necdin with p53. J. Biol. Chem. 274, 16242–16248.

We thank Dr. Y. Enokido for DRG cultures, Ms. T. Hara for fibroblast preparation, and Drs. T. Uetsuki and H. Taniura for discussions. This work was supported by grants (Research for the Future, and Scientific Research) from the Japan Society for the Promotion of Science (to K. Yoshikawa).

15.

Lipinski, M. M., and Jacks, T. (1999). The retinoblastoma gene family in differentiation and development. Oncogene 18, 7873– 7882.

16.

Davies, A. M., Lee, K. F., and Jaenisch, R. (1993). p75-deficient trigeminal sensory neurons have an altered response to NGF but not to other neurotrophins. Neuron 11, 565–574.

REFERENCES

17.

Matsumoto, K., Taniura, H., Uetsuki, T., and Yoshikawa, K. (2001). Necdin acts as a transcriptional repressor that interacts with multiple guanosine clusters. Gene 272, 173–179.

18.

Nishimura, I., Uetsuki, T., Kuwako, K., Hara, T., Kawakami, T., Aimoto, S., and Yoshikawa, K. (2002). Cell death induced by a caspase-cleaved transmembrane fragment of the Alzheimer amyloid precursor protein. Cell Death Differ. 9, 199 –208.

19.

Gavrieli, Y., Sherman, Y., and Ben-Sasson, S. A. (1992). Identification of programmed cell death in situ via specific labeling of nuclear DNA fragmentation. J. Cell Biol. 119, 493–501.

20.

Uetsuki, T., Takemoto, K., Nishimura, I., Okamoto, M., Niinobe, M., Momoi, T., Miura, M., and Yoshikawa, K. (1999). Activation of neuronal caspase-3 by intracellular accumulation of wild-type Alzheimer amyloid precursor protein. J. Neurosci. 19, 6955– 6964.

21.

Lee, E. Y., Hu, N., Yuan, S. S., Cox, L. A., Bradley, A., Lee, W. H., and Herrup, K. (1994). Dual roles of the retinoblastoma protein in cell cycle regulation and neuron differentiation. Genes Dev. 8, 2008 –2021.

22.

Weinberg, R. A. (1995). The retinoblastoma protein and cell cycle control. Cell 81, 323–330.

23.

Yoshikawa, K. (2000). Cell cycle regulators in neural stem cells and postmitotic neurons. Neurosci. Res. 37, 1–14.

24.

Porse, B. T., Pedersen, T. A., Xu, X., Lindberg, B., Wewer, U. M., Friis-Hansen, L., and Nerlov, C. (2001). E2F repression by C/EBPalpha is required for adipogenesis and granulopoiesis in vivo. Cell 107, 247–258.

25.

Taniguchi, N., Taniura, H., Niinobe, M., Takayama, C., Tominaga-Yoshino, K., Ogura, A., and Yoshikawa, K. (2000). The postmitotic growth suppressor necdin interacts with a calciumbinding protein (NEFA) in neuronal cytoplasm. J. Biol. Chem. 275, 31674 –31681.

1.

2.

3.

4.

5.

6.

7.

8.

McBurney, M. W., Reuhl, K. R., Ally, A. I., Nasipuri, S., Bell, J. C., and Craig, J. (1988). Differentiation and maturation of embryonal carcinoma-derived neurons in cell culture. J. Neurosci. 8, 1063–1073. Maruyama, K., Usami, M., Aizawa, T., and Yoshikawa, K. (1991). A novel brain-specific mRNA encoding nuclear protein (necdin) expressed in neurally differentiated embryonal carcinoma cells. Biochem. Biophys. Res. Commun. 178, 291–296. Aizawa, T., Maruyama, K., Kondo, H., and Yoshikawa, K. (1992). Expression of necdin, an embryonal carcinoma-derived nuclear protein, in developing mouse brain. Brain Res. Dev. Brain Res. 68, 265–274. Uetsuki, T., Takagi, K., Sugiura, H., and Yoshikawa, K. (1996). Structure and expression of the mouse necdin gene. Identification of a postmitotic neuron-restrictive core promoter. J. Biol. Chem. 271, 918 –924. van der Bruggen, P., Traversari, C., Chomez, P., Lurquin, C., De Plaen, E., Van den Eynde, B., Knuth, A., and Boon, T. (1991). A gene encoding an antigen recognized by cytolytic T lymphocytes on a human melanoma. Science 254, 1643–1647. Jay, P., Rougeulle, C., Massacrier, A., Moncla, A., Mattei, M. G., Malzae, P., Roeckel, N., Taviaux, S., Lefrane, J. L., Cau, P., Berta, P., Lalande, M., and Muscatclli, F. (1997). The human necdin gene, NDN, is maternally imprinted and located in the Prader–Willi syndrome chromosomal region. Nat. Genet. 17, 357–361. MacDonald, H. R., and Wevrick, R. (1997). The necdin gene is deleted in Prader–Willi syndrome and is imprinted in human and mouse. Hum. Mol. Genet. 6, 1873–1878. Nakada, Y., Taniura, H., Uetsuki, T., Inazawa, J., and Yoshikawa, K. (1998). The human chromosomal gene for necdin, a

232 26.

27.

28.

29.

30.

31.

32.

33.

34.

TAKAZAKI, NISHIMURA, AND YOSHIKAWA Taniura, H., and Yoshikawa, K. (2002). Necdin interacts with the ribonucleoprotein hnRNP U in the nuclear matrix. J. Cell. Biochem. 84, 545–555. Ma, Q., Fode, C., Guillemot, F., and Anderson, D. J. (1999). Neurogeninl and neurogenin2 control two distinct waves of neurogenesis in developing dorsal root ganglia. Genes Dev. 13, 1717–1728. Clarke, A. R., Maandag, E. R., van Roon, M., van der Lugt, N. M., van der Valk, M., Hooper, M. L., Berns, A., and te Riele, H. (1992). Requirement for a functional Rb-1 gene in murine development. Nature 359, 328 –330. Jacks, T., Fazeli, A., Schmitt, E. M., Bronson, R. T., Goodell, M. A., and Weinberg, R. A. (1992). Effects of an Rb mutation in the mouse. Nature 359, 295–300. Lee, E. Y.-H. P., Chang, C.-Y., Hu, N., Wang, Y.-C. J., Lai, C.-C., Herrup, K., Lee, W.-H., and Bradley, A. (1992). Mice deficient for Rb are nonviable and show defects in neurogenesis and haematopoiesis. Nature 359, 288 –294. LeCouter, J. E., Kablar, B., Whyte, P. F., Ying, C., and Rudnicki, M. A. (1998). Strain-dependent embryonic lethality in mice lacking the retinoblastoma-related p130 gene. Development 125, 4669 – 4679. Salehi, A. H., Roux, P. P., Kubu, C. J., Zeindler, C., Bhakar, A., Tannis, L. L., Verdi, J. M., and Barker, P. A. (2000). NRAGE, a novel MAGE protein, interacts with the p75 neurotrophin receptor and facilitates nerve growth factor-dependent apoptosis. Neuron 27, 279 –288. Lee, K. F., Li, E., Huber, L. J., Landis, S. C., Sharpe, A. H., Chao, M. V., and Jaenisch, R. (1992). Targeted mutation of the gene encoding the low affinity NGF receptor p75 leads to deficits in the peripheral sensory nervous system. Cell 69, 737–749. Shiras, A., Sengupta, A., and Shepal, V. (2001). Cloning and tissue-specific gene expression studies with dlxin-1, a newly

Received December 31, 2001 Revised version received March 25, 2002

identified transcriptional activator. Mol. Cell. Biol. Res. Commun. 4, 313–319. 35.

Azuma-Hara, M., Taniura, H., Uetsuki, T., Niinobe, M., and Yoshikawa, K. (1999). Regulation and deregulation of E2F1 in postmitotic neurons differentiated from embryonal carcinoma P19 cells. Exp. Cell Res. 251, 442– 451.

36.

Hou, S. T., Callaghan, D., Fournier, M. C., Hill, I., Kang, L., Massie, B., Morley, P., Murray, C., Rasquinha, I., Slack, R., and MacManus, J. P. (2000). The transcription factor E2F1 modulates apoptosis of neurons. J. Neurochem. 75, 91–100.

37.

Slack, R. S., Belliveau, D. J., Rosenberg, M., Atwal, J., Lochmuller, H., Aloyz, R., Haghighi, A., Lach, B., Seth, P., Cooper, E., and Miller, F. D. (1996). Adenovirus-mediated gene transfer of the tumor suppressor, p53, induces apoptosis in postmitotic neurons. J. Cell Biol. 135, 1085–1096.

38.

Eizenberg, O., Faber-Elman, A., Gottlieb, E., Oren, M., Rotter, V., and Schwartz, M. (1996). p53 plays a regulatory role in differentiation and apoptosis of central nervous system-associated cells. Mol. Cell. Biol. 16, 5178 –5185.

39.

Jordan, J., Galindo, M. F., Prehn, J. H., Weichselbaum, R. R., Beckett, M., Ghadge, G. D., Roos, R. P., Leiden, J. M., and Miller, R. J. (1997). p53 expression induces apoptosis in hippocampal pyramidal neuron cultures. J. Neurosci. 17, 1397– 1405.

40.

Simpson, M. T., MacLaurin, J. G., Xu, D., Ferguson, K. L., Vanderluit, J. L., Davoli, M. A., Roy, S., Nicholson, D. W., Robertson, G. S., Park, D. S., and Slack, R. S. (2001). Caspase 3 deficiency rescues peripheral nervous system defect in retinoblastoma nullizygous mice. J. Neurosci. 21, 7089 –7098.

41.

Brandt, B. R., and Rosen, I. (1998). Impaired peripheral somatosensory function in children with Prader–Willi syndrome. Neuropediatrics 29, 124 –126.