Disruption of neurogenesis by amyloid β-peptide, and perturbed neural progenitor cell homeostasis, in models of Alzheimer's disease

Authors

  • Norman J. Haughey,

    1. *Laboratory of Neurosciences, National Institute on Aging Gerontology Research Center, Baltimore, Maryland, USA
      Departments of †Neurology and ‡Neuroscience, Johns Hopkins University School of Medicine, Baltimore, Maryland, USA
      §Department of Neurology, University of Kentucky, Lexington, Kentucky, USA
    Search for more papers by this author
  • Avi Nath,

    1. *Laboratory of Neurosciences, National Institute on Aging Gerontology Research Center, Baltimore, Maryland, USA
      Departments of †Neurology and ‡Neuroscience, Johns Hopkins University School of Medicine, Baltimore, Maryland, USA
      §Department of Neurology, University of Kentucky, Lexington, Kentucky, USA
    Search for more papers by this author
  • Sic L. Chan,

    1. *Laboratory of Neurosciences, National Institute on Aging Gerontology Research Center, Baltimore, Maryland, USA
      Departments of †Neurology and ‡Neuroscience, Johns Hopkins University School of Medicine, Baltimore, Maryland, USA
      §Department of Neurology, University of Kentucky, Lexington, Kentucky, USA
    Search for more papers by this author
  • A. C. Borchard,

    1. *Laboratory of Neurosciences, National Institute on Aging Gerontology Research Center, Baltimore, Maryland, USA
      Departments of †Neurology and ‡Neuroscience, Johns Hopkins University School of Medicine, Baltimore, Maryland, USA
      §Department of Neurology, University of Kentucky, Lexington, Kentucky, USA
    Search for more papers by this author
  • Mahendra S. Rao,

    1. *Laboratory of Neurosciences, National Institute on Aging Gerontology Research Center, Baltimore, Maryland, USA
      Departments of †Neurology and ‡Neuroscience, Johns Hopkins University School of Medicine, Baltimore, Maryland, USA
      §Department of Neurology, University of Kentucky, Lexington, Kentucky, USA
    Search for more papers by this author
  • Mark P. Mattson

    1. *Laboratory of Neurosciences, National Institute on Aging Gerontology Research Center, Baltimore, Maryland, USA
      Departments of †Neurology and ‡Neuroscience, Johns Hopkins University School of Medicine, Baltimore, Maryland, USA
      §Department of Neurology, University of Kentucky, Lexington, Kentucky, USA
    Search for more papers by this author

Address correspondence and reprint requests to Mark P. Mattson, NIA Gerontology Research Center, 4F02, 5600 Nathan Shock Drive, Baltimore, MD 21224, USA. E-mail: mattsonm@grc.nia.nih.gov

Abstract

Neurogenesis occurs in the adult mammalian brain and may play roles in learning and memory processes and recovery from injury, suggesting that abnormalities in neural progenitor cells (NPC) might contribute to the pathogenesis of disorders of learning and memory in humans. The objectives of this study were to determine whether NPC proliferation, survival and neuronal differentiation are impaired in a transgenic mouse model of Alzheimer's disease (AD), and to determine the effects of the pathogenic form of amyloid β-peptide (Aβ) on the survival and neuronal differentiation of cultured NPC. The proliferation and survival of NPC in the dentate gyrus of the hippocampus was reduced in mice transgenic for a mutated form of amyloid precursor protein that causes early onset familial AD. Aβ impaired the proliferation and neuronal differentiation of cultured human and rodent NPC, and promoted apoptosis of neuron-restricted NPC by a mechanism involving dysregulation of cellular calcium homeostasis and the activation of calpains and caspases. Adverse effects of Aβ on NPC may contribute to the depletion of neurons and cognitive impairment in AD.

Abbreviations used

amyloid β-peptide

AD

Alzheimer's disease

APP

amyloid precursor protein

BrdU

bromo-2′-deoxyuridine

[Ca2+]c

cytosolic calcium concentration

[Ca2+]i

intracellular calcium concentration

NPC

neural progenitor cells

PBS

phosphate-buffered saline

The brains of adult mammals contain small populations of cells that are capable of dividing and differentiating into neurons and glial cells (Gage 2000). Such neural progenitor cells (NPC) can respond to environmental demands, including brain injury, increased mental and physical activity, and dietary manipulations, by increasing their proliferation and/or survival (Kempermann et al. 1997; Liu et al. 1998; van Praag et al. 1999; Young et al. 1999; Lee et al. 2000a). The signals that influence neural progenitor cell fate are being identified and include growth factors and cytokines such as basic fibroblast growth factor, brain-derived neurotrophic factor and leukemia inhibitory factor (Cameron et al. 1998; Pincus et al. 1998; Fricker et al. 1999; Rao 1999; Lee et al. 2000a). The expression of ligand- and voltage-gated ion channels change in NPC in association with changes in their proliferation and differentiation (Sah et al. 1997), although roles for ion fluxes in controlling NPC cell fate have not been established.

Recent findings suggest that the formation of some types of memory may depend on the continued production of new neural cells. As an example, in response to hippocampal-dependent but not hippocampal-independent learning, the survival of newly generated dentate granule neurons is enhanced (Gould et al. 1999), and when NPC proliferation is suppressed with an antimitotic agent, hippocampal-dependent trace conditioning is impaired (Shors et al. 2001). Age is associated with memory impairment, and a reduction in the proliferation rate of NPC (Kuhn et al. 1996). Transplantation of NPC can reverse age-associated memory impairment (Hodges et al. 2000; Qu et al. 2001) suggesting that a slowing of NPC proliferation with age may play a role in age-associated memory impairment. These findings prompted us to ask the question of whether abnormalities in the proliferation, differentiation and/or survival of NPC play a role in the memory impairment that occurs in neurodegenerative disorders such as Alzheimer's disease (AD).

Abnormal proteolytic processing of the amyloid precursor protein (APP), resulting in increased production of a self-aggregating form of Aβ is thought to be a seminal event in the pathogenesis of AD (Yankner 1996). Considerable evidence, obtained from studies of patients, and cell culture and animal models, suggests that Aβ promotes synaptic dysfunction and death of mature neurons in AD (Mattson 2000). Thus, mutations in APP and presenilins that cause early onset inherited forms of AD uniformly result in excessive production of a highly amyloidogenic form of amyloid β-peptide (Aβ1−42), and massive accumulation of plaques and associated neurofibrillary pathology in the brain (Selkoe 2001). Impaired learning and memory occurs in APP mutant mice that exhibit amyloid deposition (Hsiao et al. 1996), and chronic influsion of Aβ into the lateral ventricles impairs spatial learning in mice (Nitta et al. 1997). The mechanism whereby Aβ disrupts synaptic function and promotes neuronal degeneration involves induction of cellular oxidative stress and dysregulation of cellular calcium homeostasis (Mattson 1997). Aβ is deposited in brain regions where NPC are known to reside including the hippocampus and subventricular zone (Morys et al. 1994; Irizarry et al. 1997; Buxbaum et al. 1998; Su and Ni 1998; Shoji et al. 2001), and it is therefore possible that Aβ could adversely affect those cells. Previous studies have shown that Aβ can adversely affect mitotic cells including the inhibition of proliferation of astrocytes (Kerokoski et al. 2001) and PC12 cells (Luo et al. 1996), and the induction of apoptosis in vascular endothelial cells (Blanc et al. 1997) and microglia (Korotzer et al. 1993). In the present study we document striking adverse effects of Aβ on human NPC in vitro and on mouse NPC in vivo.

Materials and methods

Mice and experimental treatment

Male mice (12–14 months-old) overexpressing a mutant form of APP (Borchelt et al. 1996) and age-matched male non-transgenic littermate control mice (12 mice of each genotype) were maintained on a 12-h light/12-h dark cycle with free access to food and water. This line of mice exhibits increased levels of soluble Aβ and develops amyloid deposits in an age-dependent manner with diffuse deposits first appearing at approximately 12 months of age and plaque-like deposits developing later, typically by 18–22 months of age. To determine proliferation and survival of NPC in the dentate gyrus, mice were given five daily injections of 5-bromo-2′-deoxyuridine (BrdU; 50 mg/kg, i.p). Six mice of each genotype were killed 24 h after the last BrdU injection and the remaining six were killed 12 days after the last BrdU injection. In an additional study, 3-month-old APP mutant mice and four non-transgenic mice (n = 4 mice of each genotype) were killed 24 h after the last BrdU injection. Mice were killed by anesthesia overdose and perfused transcardially with saline followed by ice-cold phosphate-buffered 4% paraformaldehyde. Brains were cryoprotected in a 30% sucrose solution in phosphate buffered saline and stored at −80°C.

Stereology and immunohistochemistry

For stereology, coronal brain sections were cut at 30 µm on a freezing microtome and DNA was denatured using a 2% HCl solution. Free-floating sections were incubated for 2 h with 2% normal horse serum in phosphate-buffered saline (PBS) containing 0.2% Triton X-100, and then incubated overnight at 4°C in the presence of a 1 : 200 dilution of monoclonal BrdU antibody (Becton Dickinson, San Jose, CA, USA) in PBS. Sections were then washed with PBS, incubated for 2 h in the presence of biotinylated anti-mouse secondary antibody (Vector Laboratories, Burlingame, CA, USA), washed with PBS and incubated for 1 h in the presence of streptavidin-peroxidase complex. Sections were washed in water and then incubated for 5 min in the presence of diaminobenzidine. Sections were mounted on slides and numbers of BrdU-positive cells in the subgranular zone of the dentate gyrus were quantified using unbiased stereological methods. BrdU-positive cells were counted in a one-in-five series (150-µm apart; 12 sections in total), using a 63× oil immersion objective from the most rostral portion of the dentate gyrus, bregma −1.0 mm, to the caudal point, bregma −2.80 mm. The same area and number of sections were counted for each group. We used the optical dissector technique with 1 µm as the guard height and a dissector frame area of 70 µm2. Counts were made by an investigator (NJH) unaware of the genotype and experimental treatment history of the mice. Immunostaining for confocal analysis was performed on 30-µm coronal brain sections as follows. Sections were incubated for 1 h in a solution containing 2% normal horse serum and 0.2% Triton X-100 in PBS, and then incubated overnight at 4°C in PBS containing 2% horse serum plus monoclonal antibody against E-NCAM (1 : 250 dilution; Mujtaba et al. 1999) and a monoclonal antibody against BrdU (1 : 200 dilution; Becton Dickinson). Sections were washed in PBS and incubated for 1 h in the presence of anti-mouse IgM conjugated to AlexaFluor-488 and anti-mouse IgG conjugated to AlexaFluor-546 (Molecular Probes, Eugene, OR, USA). Sections were then washed with water, mounted on slides, and imaged using a Zeiss Confocal microscope (Carl Zeiss Inc., Thornwood, NY, USA). For the quantification of neurogenesis, immunopositive cells in the dentate gyrus from three sections in each of six APP mutant mice and six non-transgenic mice were counted and E-NCAM/Brdu-positive cells were expressed as a ratio to the total E-NCAM-positive population.

Cell cultures and experimental treatments

Neurosphere cultures were established either from the neural tube and telencephalon of 8-day-old mouse embryos (C57Bl6) or from cortical tissue from 8–10-week-old human fetuses using methods similar to those described previously (Caldwell et al. 2001). Neurospheres were maintained in suspension cultures in uncoated plastic flasks containing either DMEM/F12 (mouse cells) or Neurobasal (human cells) culture media (Gibco BRL, Rockville, MD, USA) supplemented with epidermal growth factor (EGF) and basic fibroblast growth factor (bFGF) (20 ng/mL each; Sigma, St Louis, MO, USA). In order to induce differentiation of NPC, the neurospheres were plated on polyethyleneimine-coated glass coverslips in medium lacking bFGF and EGF and containing 1% fetal bovine serum. Experiments were performed on differentiating neurospheres during a 5-day period. The C17.2 NPC line (a kind gift from C. Cepko, Harvard University) was maintained in DMEM with 10% fetal bovine serum and 5% horse serum (Gibco BRL) as described previously (Snyder et al. 1992). C17.2 cells were subcultured onto polyethyleneimine-coated glass coverslips 24–48 h prior to experimentation. Experimental treatments included: Aβ1−42 (pre-aggregated overnight at 37°C; Bachem, Torrance, CA, USA), the caspase inhibitor zVADfmk (Calbiochem, San Diego, CA, USA), calpain-1 and -2 inhibitors (Boehringer Ingelheim, Petersburg, VA, USA), IP3-receptor antagonist xestospongin-C (Calibochem), ryanodine (at 10 µm, a concentration known to lock the ryanodine receptor in a closed state; Calbiochem), sarco-endoplasmic pump inhibitor thapsigargin (Molecular Probes), glutamate and ATP (Sigma). All experimental treatments were prepared as concentrated 200–1000× stocks in either water or dimethylsulfoxide, as appropriate, and were added to cultures 1 h before Aβ.

Quantification of cell proliferation, differentiation and apoptosis

To measure cell proliferation, human neurospheres were treated with [3H]thymidine (2 µCi/mL; Amersham Pharmacia Biotech, Piscataway, NJ, USA) for 24–48 h before harvest onto glass-fiber filters using a cell harvester (Pharmacia, Peapak, NJ, USA). Radioactivity was counted using a Betaplate liquid scintillation counter (Pharmacia), and values were expressed as cpm/mg protein. As one measure of cell differentiation in neurospheres, we quantified the area of cell migration across the growth substrate. Photographs of individual spheres were taken every 24 h for 5 days. Area was calculated as the total area minus the area of the central sphere using NIH Image software. We quantified the numbers of NeuN-positive cells migrating from plated neurospheres by counting four 40× microscope fields per neurosphere; each field was located adjacent to the central cell mass. Quantification of apoptosis was performed in cells stained with the fluorescent DNA binding dye Hoescht 33342 using methods described previously (Kruman et al. 1997). Nuclei were visualized and photographed under epilfuorescence illumination (340-nm excitation and 510-nm barrier filter) using a 40× oil immersion objective. Two-hundred cells from three fields in three separate cultures per experimental condition were counted without knowledge of experimental condition. Cells in which nuclear staining was diffuse were considered viable and cells where nuclear staining was condensed or fragmented were considered ‘apoptotic’.

Immunocytochemistry

For immunostaining cultured neurospheres, the cultures were fixed for 30 min in a solution of 4% paraformaldehyde in PBS, and membranes were permeabilized by incubation for 10 min in a solution of 0.2% Triton X-100 in PBS. Cultures were then incubated for 1 h in blocking solution (2% normal goat serum and 2% normal horse serum in PBS) and then primary antibodies were added and the cultures were incubated overnight at 4°C. The primary antibodies included: a rabbit polyclonal antibody against glial fibrillary acidic protein (GFAP) (1 : 200; Sigma), a mouse monoclonal antibody against the neuron-specific nuclear antigen NeuN (1 : 500; Chemicon, Temecula, CA, USA), a mouse monoclonal antibody against βIII-tubulin (1 : 1000; Chemicon), a mouse monoclonal antibody against psa-NCAM (1 : 1000; Chemicon) a mouse monoclonal antibody against E-NCAM (1 : 250; provided by M. Rao) and a mouse monoclonal antibody against human nestin (1 : 500; provided by C. Messam and E. Major, National Institutes of Health, Bethesda, MD, USA). Cultures were washed with PBS and then incubated for 1 h in the presence of appropriate fluorescently tagged anti-mouse and anti-rabbit secondary antibodies (AlexaFluor 633, 546 and 488; 1: 2000 dilution; Molecular Probes). In some cases cultures were further stained with propidium iodide or Hoescht 33342. Confocal images were acquired using a Zeiss 510 CSLM microscope.

Measurement of intracellular calcium concentration

Cytosolic calcium concentration ([Ca2+]c) was determined by imaging of the fluorescent probe Fura-2/AM using the methods described previously (Haughey et al. 1999). Cells were incubated for 25 min at 37°C in Locke's buffer (154 mm NaCl, 3.6 mm NaHCO3, 5.6 mm KCl, 1 mm MgCl2, 5 mm HEPES, 2.3 mm CaCl2, 10 mm glucose; pH 7.4) with 10% pluronic and 2 µm Fura-2/AM. Cells were washed with Locke's buffer to remove extracellular Fura-2 and were incubated at 37°C for 20 min to allow complete de-esterfication of the probe. Cultures were transferred to the stage of a Zeiss Axiovert microscope coupled to a CCD camera and a Zeiss AttoFluor calcium imaging system and cellular fluorescence was imaged using a 40× oil objective. The average intracellular calcium concentration ([Ca2+]i) in individual cell bodies was determined from the ratio of the fluorescence emissions obtained using two different excitation wavelengths (340 nm and 380 nm). The system was calibrated by curve fitting using reference standards containing 0–10 mm calcium (Molecular Probes).

Immunoblot analyses

Cells were lysed in ice-cold buffer consisting of 62 mm Tris, 2 mm EDTA, 2 mm EGTA, 2% sodium dodecyl sulfate, 10% glycerol and a protease inhibitor cocktail (Sigma, St Louis, MO, USA), pH 6.0. Proteins (50 µg/lane) were separated by sodium dodecyl sulfate–polyacrylamide gel electrophoresis (10–15% acrylamide) and transferred to a nitrocellulose membrane. Membranes were incubated for 30 min in the presence of 5% non-fat milk and incubated overnight at 4°C with the primary antibodies P2X1, P2X3, P2X4, P2Y2, P2Y4 (1 : 500; Chemicon) or P35/Cdk5 (1 : 1000; Oncogene, Madison, WI, USA). Membranes were exposed to horseradish peroxidase-conjugated secondary antibody (1 : 3000; Jackson Immunological Research Laboratories Inc., West Grove, PA, USA) and immunoreactive proteins were visualized using a chemiluminescene-based detection kit according to the manufacture's protocol (ECL kit; Amersham Corp., Arlington Heights, IL, USA).

Results

Neurogenesis is impaired in the dentate gyrus of APP mutant mice

Transgenic mice overexpressing a mutant form of APP that causes autosomal dominant familial AD in humans develop progressive plaque-like deposits of Aβ in their hippocampus and cerebral cortex (Sturchler-Pierrat et al. 1997), and learning and memory deficits (Chen et al. 2000). We employed one such line of APP mutant mice (Borchelt et al. 1996) to determine whether the proliferation and/or survival of NPC were altered by Aβ. APP mutant mice (12–14 months old) with diffuse Aβ deposits, and age-matched non-transgenic control mice, were administered the thymidine analog bromo-deoxyuridine (BrdU; 50 mg/kg, i.p) once daily for 5 days. Mice were killed either 1 day or 12 days after the last BrdU injection and BruU-labeled cells were identified in the proliferative subgranular zone of the dentate gyrus by immunostaining (Fig. 1a). As expected, numerous amyloid deposits were present in the hippocampus of APP mutant mice, including the dentate gyrus, whereas no such deposits were found in non-transgenic mice (Fig. 1b). We quantified numbers of BrdU-positive cells in the entire dentate gyrus using unbiased stereology methods (Lee et al. 2000a). One day after the final injection (a measure of proliferation) the number of BrdU-positive cells in the dentate gyrus was decreased significantly in APP mutant mice compared with non-transgenic mice (Fig. 1c). One day after BrdU administration, BrdU immunopositive cells did not label with antibodies against NeuN or GFAP (Fig. 1d), indicating that the newly generated cells had not yet differentiated into neurons or astrocytes. Twelve days after the final injection (a measure of the survival of newly generated cells) there was a greater decrease (approximately 55%) in BrdU-labeled cells in the APP mutant mice compared with the control mice (25% decrease; Fig. 1c). Thus, both the proliferation and survival of NPC was decreased in the dentate gyrus of APP mutant mice with amyloid deposits. In contrast to the 12-month-old APP mutant mice, younger (3-month-old) APP mutant mice, which had not yet developed amyloid deposits, did not exhibit a decrease in the number of BrdU-immunopositive cells: the total number of BrdU-positive cells in the dentate gyrus of 3-month-old APP mutant mice was 3229 ± 369 compared with 3363 ± 124 BrdU-positive cells in the dentate gyrus of age-matched non-transgenic mice (n = 4 animals per group; mean ± SD).

Figure 1.

Neural progenitor cell proliferation and survival are decreased in transgenic APP mutant mice. (a) Micrographs showing BrdU immunopositive cells in the subgranular layer of the dentate gyrus in non-transgenic (Non-tg) and APP mutant transgenic mice (mAPP) visualized using nickel-enhanced diaminobenzidine as the chromagen that stains positive nuclei black (arrows). (b) Congo red staining showing deposits of amyloid (arrows) in the dentate gyrus of APP mutant mice. (c) Results of quantification of BrdU immunopositive cells in the dentate gyrus of non-transgenic and APP mutant mice at either 1 or 12 days after BrdU administration. Values are the mean ± SEM (n = 6 mice per group). **p < 0.01; ***p < 0.001 compared with the corresponding value for non-transgenic mice (unpaired t-test). (d) Confocal image of the dentate gyrus of an APP mutant mouse showing that BrdU immunopositive cells (red) do not colocalize with NeuN-positive neurons (green) or GFAP-positive astrocytes (white) 1 day after BrdU administration. Scale bar represents 50 µm.

To determine whether Aβ affected the population of NPC that differentiate into neurons, we immunostained brain sections from mice killed 12 days after BrdU administration with an antibody against E-NCAM. In our previous studies we showed that E-NCAM (high polysialic-acid NCAM) labels neuron-restricted NPC and their differentiated progeny (Mujtaba et al. 1999). E-NCAM-immunoreactive cells can generate multiple kinds of neurons but not oligodendrocytes or astrocytes (Mayer-Proschel et al. 1997; Kalyani et al. 1998), although this form of NCAM may not be present in all cells committed to a neuronal phenotype (Kuhn et al. 1996; Vitry et al. 2001). The E-NCAM antibody labeled a subpopulation of dentate granule cells and their axonal projections in region CA3 (mossy fibers) (Fig. 2a). The latter observation is consistent with localization of E-NCAM to axons and growth cones in the developing spinal cord (Joosten et al. 1996). In additional sections we double-stained cells with antibodies against BrdU and E-NCAM (Fig. 2b). Numbers of E-NCAM immunoreactive cells in the dentate gyrus were decreased to 143 ± 45 in APP mutant mice compared with 211 ± 26 in control mice (mean ± SD; Fig. 2b). The number of BrdU+/E-NCAM-positive cells was decreased to 12 ± 6 in the APP mutant mice compared with 71 ± 11 in the non-transgenic mice. When expressed as a percentage of the total number of E-NCAM-positive cells there were fewer BrdU+/E-NCAM-positive cells, demonstrating that fewer NPC differentiate into neuron-restricted cells in APP mutant mice (Fig. 2b).

Figure 2.

Neurogenesis is decreased in transgenic APP mutant mice. (a) Confocal images showing E-NCAM immunoreactivity (neuronal-progenitor-cells) in the dentate gyrus and region CA3 of hippocampus in non-transgenic (Non-tg) and APP mutant transgenic mice (mAPP). Scale bars represent 50 µm (40× magnification) and 200 µm (10× magnification). Arrows indicate rostral orientation. (b) Confocal images showing BrdU (red), E-NCAM (green) and cells-immunoreactive for both BrdU and E-NCAM (yellow) in the dentate gyrus. Scale bar represents 10 µm. Dashed line indicates subgranular zone margin. The number of newly generated neuronal progenitor cells (dual labeled BrdU/E-NCAM cells) were significantly decreased in APP mutant mice. Values are expressed as the percentage of all E-NCAM-positive cells in the dentate gyrus that were dual labeled with BrdU and E-NCAM. ***p < 0.001 compared with the corresponding value for non-transgenic mice (n = 6; unpaired t-test). (c) Confocal images showing a neuronal progenitor that is immunopositive for both βIII-tubulin (red) and E-NCAM (green) in a non-transgenic mouse. Overlapping areas appear in yellow on the merge. Images are merged 3D reconstructions of serial z-sections.

Amyloid β-peptide inhibits neurogenesis of human NPC and induces apoptosis of neuron-restricted progenitors

We next employed human cortical neurospheres to more directly determine the effects of Aβ on neural progenitor cell proliferation, differentiation and survival. We chose to study human progenitor cells because AD is a human disorder and it is therefore critical to establish the effects of Aβ on human NPC. Similar to previous reports (Uchida et al. 2000), we found that when embryonic human neurosphere cultures are maintained in suspension in the presence of EGF and bFGF, the vast majority of cells are immunoreactive with a nestin antibody (Figs 3a and b), and are not immunoreactive with markers of neuronal progenitor cells (psa-NCAM, E-NCAM), neuron-restricted progenitor cells (βIII tubulin), astrocytes (GFAP), or differentiated neurons (NeuN) (data not shown), suggesting that neurospheres consist mainly of pluripotent stem cells. However, within 24 h of the removal of growth factors and attachment to a polyethyleneimine substrate, cells at the outer margin of the sphere were immunopositive for both nestin and E-NCAM (Fig. 3a). NPC progressively differentiated and migrated in a radial pattern from the core of the sphere (Fig. 3b). Two morphologically and antigenically distinct differentiated cell types were observed, namely, βIII-tubulin-positive cells (∼40% of cells) with a neuron-like morphology and GFAP-positive cells (∼50% of cells) with an astrocyte-like morphology (Fig. 3c).

Figure 3.

Aβ disrupts cell differentiation in cultured human neurospheres. (a) Confocal images of a human neurosphere 24 h after attachment to substrate showing nestin immunoreactive cells in the core of the sphere (green) and cells dual labeled with nestin and E-NCAM (red; dual labeled cells are yellow) at the outer margin of the sphere. Images are merged 3D reconstructions of serial z-secitons. Scale bars represent 200 µm (10× magnification) and 50 µm (40× magnification). (b) Confocal images of a human neurosphere 5 days after withdrawal of EGF and bFGF showing triple-label staining for GFAP (white), nuclei (propidium iodide; red) and βIII-tubulin (green). The micrograph at the lower right is a merged image of a 3D reconstruction of serial z-sections. (c) Phase-contrast micrographs of neurospheres in a control culture and an Aβ-treated culture 5 days after plating (Aβ was added to the culture at a concentration of 5 µm 3 h after plating). (d) The area that NPC migrated from the core of the neurosphere decreased significantly when concentrations of Aβ of 1 µm or greater were added to cultures. **p < 0.01, ***p < 0.001 compared with corresponding control value (n = 3 cultures per condition; 4–10 spheres analyzed/culture); anova with Tukey's post-hoc tests). (e) The percentage of dissociated human NPC that attached to substrate decreased significantly in cultures exposed to Aβ at concentrations of 1 µm or greater. **p < 0.01, ***p < 0.001 (n = 4–6; anova with Tukey's post-hoc tests).

To determine the effect of Aβ on the differentiation and migration of NPC, we exposed human neurospheres to Aβ1−42 3 h after plating. Aβ decreased the area of radial outgrowth of cells from neurospheres in a concentration-dependent (0.1–5 µm) manner (Figs 3c and d). At the highest concentration of Aβ tested (5 µm) there was a dramatic disruption in the migration of cells from neurospheres. Clumping of rounded cells near the border of spheres treated with high concentrations of Aβ suggested that cells did not migrate because of reduced adhesion to the substrate (Fig. 3c). To determine whether Aβ decreases the radial outgrowth of cells from neurospheres by reducing cell adhesion to the growth substrate, we dissociated neurospheres and plated single cells in the presence of Aβ. The number of cells that attached to substrate was reduced significantly by concentrations of Aβ of 1 µm or greater, but not by lower concentrations (Fig. 3e). Thus, Aβ reduced the migration of cells from neurospheres NPC by interfering with attachment to substrate.

Amyloid-β peptide is toxic to NPC and neuron-restricted progenitors

To determine whether the disruption of NPC involves an apoptotic process, we exposed human neurospheres that had been maintained for 5 days in culture to Aβ for 24 h and counted the numbers of cells with condensed or fragmented (apoptotic) nuclei. Aβ increased the percentage of apoptotic cells by three-fold and many of the apoptotic cells were E-NCAM positive (Fig. 4a). Thus, Aβ is toxic to NPC in human neurosphere cultures.

Figure 4.

Aβ induces death of NPC by an apoptotic mechanism involving caspases and calpains. (a) The micrographs show human neurosphere cultures that had been exposed for 24 h to vehicle (Control) or 5 µm Aβ and then stained with the fluorescent DNA-binding dye Hoechst 33342. There was a significant increase in the number of cells with apoptotic (condensed and fragmented) nuclei in neurospheres that had been exposed to Aβ. Color panel shows that cells with apoptotic nuclei were also immunopositive from E-NCAM. Values are the mean ± SEM of determinations made in three separate cultures (***p < 0.001; paired t-test). (b) Fluorescence image of a neurosphere culture exposed to 5 µm Aβ for 24 h with an arrow pointing to an E-NCAM immunospositive cell (green) with a condensed and fragmented nucleus consistent with apoptosis (Hoescht 33342 stain shown in blue). (c) and (d), C17.2 cells (c) and mouse NPC cultures (d) were exposed for 24 h to the indicated concentrations of Aβ or glutamate (Glut) and the percentage of cells with apoptotic nuclei in each culture were quantified. Values are the mean ± SEM of between three and six cultures (***p < 0.001 compared with control vaue; anova with Tukey's post-hoc tests). (e) Inhibitors of caspases and calpains protect NPC against Aβ-induced death. Cultures of C17.2 cells were pre-treated for 1 h with the caspase inhibitor zVAD-fmk (50 µm), calpain inhibitor-1 (CP1, 30 µm) or calpain inhibitor-2 (CP2, 30 µm). Cultures were then exposed for 24 h to 5 µm Aβ and the percentage of cells with apoptotic nuclei in each culture was determined. Values are the mean ± SEM of determinations made in three cultures. (***p < 0.001 compared with each of the other values, ap < 0.001 compared with Aβ; anova with Tukey's post-hoc tests).

It is possible that the toxic effects of Aβ on NPC in neurosphere cultures is indirect and requires the release of toxic factors from underlying glia or other cells. To determine whether Aβ has direct effects on survival in NPC we used disassociated neurospheres and C17.2 cells (a neural progenitor cell line; Snyder et al. 1992). We chose C17.2 cells for analysis because they are a pure clonal cell type that allows unequivocal demonstrations of direct effects of Aβ on their proliferation and survival. Primary mouse NPC and C17.2 cells were equally susceptible to the toxic effect of Aβ, suggesting a direct effect of Aβ on NPC (Figs 4b and c). Neither NPC nor C17.2 cells were affected by the excitotoxin glutamate (Figs 4b and c), consistent with results showing that glutamate receptors are not expressed until NPC begin to differentiate into neurons (Maric et al. 2000).

We confirmed that the cell death caused by Aβ was indeed apoptosis by showing that C17.2 cells pre-treated with the caspase inhibitor zVAD-fmk were resistant to Aβ-induced death (Fig. 4d). This mode of cell death in NPC is consistent with previous studies showing that Aβ can induce caspase-mediated apoptosis in differentiated neurons and tumor cell lines (Jordan et al. 1997; Guo et al. 1998, 1999; Mattson et al. 1998; Troy et al. 2000). We also found that an inhibitor of the calcium-activated protease calpain-1 was very effective in protecting NPC against Aβ-induced death (Fig. 4d), which prompted further studies to address the role of calcium in the adverse effects of Aβ on NPC.

The adverse effects of Aβ on NPC involve dysregulation of cellular calcium homeostasis

Aβ is known to disrupt calcium homeostasis in mature neurons (Mattson et al. 1992). NPC in culture did not respond with increased levels of [Ca2+]c when challenged with either glutamate (10–200 µm) or KCl (50–100 mm), and we were unable to detect either the glutamate receptor subunit NR2A or the α1-subunit of l-type voltage-dependent calcium channels by immunoblot analysis (data not shown), consistent with results showing that glutamate receptors are not expressed until NPC begin to differentiate into neurons (Maric et al. 2000). However, immunoblot analysis showed that NPC express several purinergic receptor subunits including P2X4, P2Y2, P2Y4 but not P2X1 (data not shown) and NPC showed a robust increase in [Ca2+]c when exposed to ATP (Fig. 5a), an agent that stimulates purinergic-receptor mediated calcium influx and release from IP3-sensitive internal stores (Fields and Stevens 2000; Khakh 2001). To determine whether Aβ can alter calcium homeostasis in NPC, we exposed C17.2 cells and primary mouse NPC to Aβ and then measured [Ca2+]c prior to and after exposure to ATP. Exposure of C17.2 cells to Aβ increased the basal [Ca2+]c and increased the peak ATP-triggered [Ca2+]c response (Figs 5a and b). The effects of Aβ on basal [Ca2+]c and peak responses to ATP were concentration-dependent (Fig. 5c) and were similar in primary mouse NPC and C17.2 cells (Figs 6c and d). When cells were pretreated with agents that selectively block the release of calcium from either IP3-sensitive stores (xestospongin; Gafni et al. 1997) or ryanodine-sensitive stores (ryanodine), they were resistant to Aβ-induced death (Fig. 5e). Thus, calcium release is an essential event in the death-inducing effect of Aβ on primary NPC and C17.2 cells.

Figure 5.

Aβ increases the basal and ATP-triggered levels of cytosolic calcium in NPC. (a) Pseudocolor images showing intracellular free calcium concentrations in C17.2 cells prior to (basal) and after (peak) exposure to 100 µm ATP. Insets show magnification of a single cell from the corresponding image. Cells had been pre-treated for 12 h with vehicle (Control) or 5 µm Aβ. (b) Tracings of cytosolic calcium concentrations showing ATP-stimulated response in control cells and cells pre-treated with 5 µm Aβ (each data point is the mean ± SEM of 46–52 cells from three separate experiments). (c) C17.2 cells were pre-treated for 12 h with the indicated concentrations of Aβ, and cystosolic calcium levels were measured prior to (basal; white bars) and after (peak; black bars) exposure to 100 µm ATP. Values are the mean ± SEM of measurements made in either three or four separate experiments (***p < 0.001 compared with corresponding control value, ap < 0.05 compared with the corresponding control value; anova with Tukey's post-hoc tests). (d) Primary NPC isolated from E9 mouse embryos were pre-treated for 12 h with 5 µm Aβ, and cystosolic calcium levels were measured prior to (basal; white bars) and after (peak; black bars) exposure to 100 µm ATP. Values are the mean ± SEM of measurements made in either three or four separate cultures (*p < 0.05, ***p < 0.001 compared with corresponding control value; anova with Tukey's post-hoc tests). (e) Cultures of C17.2 cells were pre-treated for 1 h with xestospongin (1 µm) or ryanodine (1 µm), and were then exposed for 24 h to 5 µm Aβ. The percentage of cells with apoptotic nuclei in each culture was determined (mean ± SEM; n = 3 cultures). ***p < 0.001 compared with the control value, ap < 0.001 compared with the Aβ value (anova with Tukey's post-hoc tests).

Figure 6.

Aβ enhances ATP-triggered calcium flux in neuronal stem cells by increasing ER calcium load. (a) Cultured C17.2 cells were pre-treated for 12 h with 5 µm Aβ. The culture medium was then replaced with Ca2+-free medium and the cytosolic calcium concentration was measured prior to (basal; white bars) and after (peak; black bars) exposure to 100 µm ATP. Values are the mean ± SEM of determinations made in between four and six cultures (***p < 0.001 compared with the corresponding value for cultures not treated with Aβ; anova with Tukey's post-hoc tests). (b) C17.2 cells were pre-treated for 12 h with vehicle (white bars) or 5 µm Aβ (black bars), and the peak cytosolic calcium concentration after exposure to 1 µm thapsigargin was determined (left). Additional cultures were exposed to 1 µm thapsigargin for 10 min to deplete ER calcium stores, and were then exposed to 100 µm ATP for 1 min (right). Values are the mean ± SEM of measurements made in at least four cultures/condition (***p < 0.001 compared with control value; unpaired t-test). (c) Immunoblots showing levels of P2Y2 and P2Y4 proteins in lysates of cells that had been exposed to 5 µm Aβ for the indicated time periods. (d) C17.2 cells were pre-treated for 12 h with 5 µm Aβ, and cytosolic calcium levels were measured prior to and after exposure to thapsigargin in calcium-free buffer. Each trace represents the mean of measurements made in 36 (Control) and 54 (Aβ) cells. (e) C17.2 cells were pre-treated for 12 h with vehicle (Control) or 5 µm Aβ. The cytosolic calcium concentration was then measured prior to and after exposure to thapsigargin and the total thapsigargin-releasable pool of calcium was estimated by determining the area under the curve. Values are the mean ± SEM of determinations made in three separate cultures (24–35 cells analyzed per culture (***p < 0.001; paired t-test). (f) C17.2 cells were pre-treated for 12 h with vehicle (Control) or 5 µm Aβ. Xestospongin (1 µm) and ryanodine (1 µm) were then added to the cultures and the cytosolic calcium concentration was measured prior to (basal; white bars) and after (peak; black bars) exposure to 100 µm ATP. Values are the mean ± SEM of determinations made in between four and six cultures (basal, **p < 0.01, ***p < 0.001 compared with control value; peak, ***p < 0.001 compared with control value; anova with Tukey's post-hoc tests).

To determine whether the increased calcium response to ATP caused by Aβ was the result of increased calcium release from endoplasmic reticulum stores and/or influx through plasma membrane channels, we removed extracellular calcium prior to examining the response to ATP exposure. In the absence of extracellular calcium, Aβ did not alter the basal level of [Ca2+]c, however, the peak ATP-triggered level of [Ca2+]c was increased to a degree similar to that when extracellular calcium was present (Fig. 6a). These findings suggest that Aβ increases basal levels of [Ca2+]c by increasing the permeability of the plasma membrane to calcium, and that Aβ enhances calcium ATP-triggered release of calcium from internal stores but not calcium influx. To determine whether Aβ increased the endoplasmic reticulum (ER) calcium load we inhibited sarco-endoplasmic calcium pump activity with thapsigargin and measured the accumulation of calcium in the cytosol. The calcium response to thapsigargin was increased in cells that had been pre-treated with Aβ (Fig. 6a), suggesting that Aβ increases the size of the ER calcium pool. When ER calcium pools were depleted by prolonged exposure to thapsigargin, the calcium response to ATP was reduced and was not different in control and Aβ-treated cells (Fig. 6b). Exposure of NPC to Aβ did not affect levels of the IP3-coupled purinergic receptors P2Y2 and P2Y4, as determined by immunoblot analyses (Fig. 6c), suggesting that the mechanism by which Aβ increases ATP-triggered ER calcium release does not involve up-regulation of IP3-linked purinergic receptors. To confirm that Aβ increases ER calcium load, we determined the total amount of ER calcium in Aβ-treated and control NPC by treating the cells with thapsigargin in the absence of extracellular calcium to remove the influence of capacitative calcium entry. The area under the curve (a measure ER calcium content) was greater in Aβ-treated NPC (Figs 6d and e). The increase of basal [Ca2+]c following Aβ treatment was not affected by inhibition of calcium-induced calcium-release with ryanodine or by inhibition of IP3 receptors with xestospongin (Fig. 6f), further supporting an effect of Aβ on basal [Ca2+]c mediated by increased influx through plasma membrane channels similar to that reported for mature hippocampal neurons (Mattson 1997). We found that xestospongin markedly attenuated the peak ATP-triggered calcium response, whereas ryanodine was less effective (Fig. 6f) consistent with the ATP-triggered release of calcium from IP3-sensitive ER calcium stores. Collectively, our data indicate that ER calcium content is increased by Aβ and results in an exaggerated ATP-triggered release of calcium from IP3-regulated ER stores.

Subtoxic concentration of amyloid-β peptide decreases NPC proliferation and neuronal differentiation

Whereas Aβ at concentrations of 10 and 100 nm did not reduce the total area of cell migration, cell density was lower within the outgrowth area. These observations prompted us to determine whether subtoxic quantities of Aβ affect the proliferation and differentiation of NPC. When expanding neurospheres were exposed to Aβ (100 nm and 1 µm), the incorporation of [3H]thymidine into DNA was significantly decreased (control, 659 ± 60 cpm/mg protein; 100 nm Aβ, 481 ± 32 cpm/mg protein, p < 0.01; 1 µm Aβ, 447 ± 56 cpm/mg protein, p < 0.01; mean ±SEM, n = 4) demonstrating an inhibitory effect of Aβ on NPC proliferation.

To determine the effects of Aβ on NPC differentiation, we exposed neurospheres to low concentrations of Aβ (1–100 nm) 3 h after plating and allowed cell migration and differentiation to continue for 5 days. After 5 days, an antibody to GFAP labeled astrocytes in a typical whispy pattern and E-NCAM-positive cells were healthy in appearance (Fig. 7a). Neurosphere cultures exposed to Aβ (100 nm), astrocytes assumed a stellate morphology and the processes of E-NCAM-positive cells appeared fragmented (Fig. 7b) suggesting that neuronal differentiation may be suppressed. In support of this hypothesis we observed a concentration-dependent decrease in the number of NeuN-positive cells that was significant with concentrations of Aβ greater than or equal to 10 nm (Fig. 7b). The 10 nm concentration of Aβ is 100 times lower than the minimally toxic concentration to NPC- or neuron-restricted precursors. Together these results suggest that Aβ can disrupt neuronal differentiation at concentrations that are not directly toxic to NPC.P35/CDK 5 kinase is involved in cell fate decisions, and cleavage of p35 to p25 is involved in Aβ-induced death of neurons (Kwon and Tsai 1998; Alvarez et al. 1999; Lee et al. 2000b). We found that p35/CDK 5 levels were decreased in NPC exposed to Aβ (Fig. 7c), suggesting that Aβ-induced cleavage of p35 was involved in the reduction of differentiated neurons. Thus, in human neurospheres, Aβ can inhibit cell proliferation, alter the morphology of newly generated astrocytes, decrease the numbers of neuron-restricted progenitors, and decrease the number of differentiated neurons. Together these results suggest that Aβ can disrupt neuronal differentiation at concentrations that are not directly toxic to NPC.

Figure 7.

Aβ inhibits neurogenesis in human neurospheres. (a) Confocal images of human neurospheres that had been exposed for 5 days to either vehicle (Control) or 100 nm Aβ. The neurospheres were triple-stained for GFAP (white), propidium iodide (red) and E-NCAM (green); the lower right panel is the merged image. The morphology of astrocytes was changed to a more activated state, and E-NCAM-positive cells appeared damaged. Scale bar represents 50 µm Aβ. (b) Confocal images showing Neu-N (green) and GFAP (white) immunoreactivity in neurospheres that had been exposed for 5 days to the indicated concentrations of Aβ. Numbers of Neu-N-positive cells were quantified and the results are shown. **p < 0.01, ***p < 0.001 compared with the control value (anova with Tukey's post-hoc tests; n = 4 cultures; from four to ten spheres/culture). Scale bar represents 200 µm. (c) Immunoblot analysis of p35/CDK5 kinase in human neurosphere cultures that had been exposed to 5 µm Aβ for the indicated times.

Discussion

Prior studies of the pathogenesis of AD have focused on the degeneration of existing neuronal circuits, and have not explored the possibility of a defect in production of new neurons from NPC. Subventricular zone NPC give rise to olfactory neurons (Luskin 1998), and olfactory deficits occur very early in course of AD (Devanand et al. 2000; Kovacs et al. 2001).

Our findings of decreased proliferation and migration of NPC in the subventricular zone (Haughey et al. 2002) suggest that impaired neurogenesis may contribute to the depletion of olfactory neurons in AD patients, and prompted us to explore the possibility that similar deficits of NPC in the dentate gyrus may occur in AD. We found that Aβ inhibits neurogenesis of human NPC in culture and of mouse NPC in vivo, and that neurogenesis is also impaired in mice overexpressing a mutant form of APP that causes early onset autosomal dominant AD in humans. These findings do not link the diffuse cortical and hippocampal pathology seen in late-stage AD to dysfunctions of NPC, but rather present a previously undocumented pathology in AD that could affect memory prior to massive cell loss.

Neurogenesis involves the proliferation of NPC, and the differentiation and survival of neurons arising from the NPC. Our data demonstrate that Aβ can adversely affect each of these three critical steps of neurogenesis. NPC from both the dentate gyrus of the hippocampus and the subventricular zone (Haughey et al. 2002) were affected by Aβ, suggesting that the functions of brain regions supplied by these two NPC populations may be compromised by Aβ-mediated impairment of neurogenesis. The hippocampus is a major focus of pathology in AD, with massive loss of pyramidal neurons and disruption of perforant path connections with dentate gyrus being particularly prominent. The major progeny of dentate NPC (dentate granule neurons) are not markedly depleted in AD; instead, CA1 pyramidal neurons are selectively vulnerable (Thal et al. 2000) and degenerate in some lines of APP mutant transgenic mice (Staufenbiel et al. 1998). The hippocampus has a complex neuronal circuitry. Dysfunctions of cells in the perforant pathway that synapse with dendrites of dentate granule neurons or the loss of subpopulations of cells in the hippocampus are known to result in aberrant dendritic sprouting, epilepsy and cell death of other neuronal populations (Golarai et al. 2001). New neurons in the dentate gyrus are functionally incorporated into the hippocampus (van Pragg et al. 2002) and may participate in the formation of hippocampal-dependent memory (Shors et al. 2001). Alternatively, these new neurons may participate in the clearance of old memory traces after cortical memory consolidation, leaving the hippocampus available to process new memories (Feng et al. 2001). Environmental enrichment and exercise enhance neurogenesis and learning and memory in rodents, which is associated with enhanced long-term potentiation of synaptic transmission (a cellular correlate of learning and memory) in region CA1 (Gould et al. 1999; Nilsson et al. 1999; van Praag et al. 1999). A reduction of new dentate granule neurons in AD could thus disrupt hippocampal circuitry critical for learning and memory.

Although the effects of Aβ on NPC have not been studied previously, the mechanism whereby Aβ can damage and kill mature neurons has been examined in considerable detail. Abnormal processing of amyloid precursor protein that results in the formation of Aβ1–42 is thought to be a seminal event in the pathogenesis of AD (Yankner 1996). Monomeric forms of Aβ are considerably less toxic to neurons than aggregated forms. When in the process of self-aggregation, Aβ induces oxidative stress resulting in the impairment of membrane ion and glucose transporters and destabilization of cellular calcium homeostasis (Mattson 1997). Subtoxic levels of Aβ can impair synaptic transmission (Chapman et al. 1999; Larson et al. 1999), learning and memory (Nabeshima and Nitta 1994; Hsiao et al. 1996), and can alter calcium responses to glutamate (Mattson et al. 1992), suggesting that perturbed calcium homeostasis is an early and pivotal event in the pathogenesis of AD and can enhance neurons susceptibility to excitotoxic insult. Unlike neurons, NPC do not express glutamate or voltage-sensitive ion channels and are insensitive to glutamate toxicity. We found that Aβ increased the basal concentration of cytosolic calcium in NPC by increasing the permeability of the plasma membrane to calcium, and also increased the ER calcium load. Agents that block either IP3 or ryanodine receptors attenuated the toxic effects of Aβ on NPC, demonstrating a major role for perturbed calcium homeostasis in the disruption of NPC by Aβ.

Our data are the first to demonstrate a role for calcium in modulating neural progenitor cell fate. However, previous studies have established the calcium ion is an intracellular messenger that regulates neurite outgrowth and cell survival during the development of the nervous system (Mattson 1992; Spitzer 1995). Calcium signaling can also regulate the proliferation and differentiation of hematopoietic (Podesta et al. 2000) and oligodendrocyte (Johnson et al. 2000) progenitor cells. Our findings suggest roles for calcium release and influx stimulated by ATP and other signals in the control of neurogenesis in the adult mammalian brain. Although effects of ATP on NPC have not previously been described, it was reported that ATP can affect the proliferation of multipotent hematopoeitic stem cells (Whetton et al. 1988). ATP can be released from neurons and astrocytes in response to activity in neural circuits (Sawynok et al. 1993; Santos et al. 1999) suggesting the possibility that ATP plays a role in activity-dependent regulation of NPC proliferation and survival (Kempermann et al. 1997; van Praag et al. 1999). Our data also suggest that dysregulation of calcium homeostasis in NPC may occur in AD. The latter possibility is consistent with the established effects of Aβ on calcium homeostasis in neurons, and with emerging data suggesting a widespread abnormality in calcium release resulting from mutations in presenilin-1 that cause early onset inherited AD (Mattson et al. 2001). Indeed, neuronal differentiation is premature in presenilin-1 deficient mice Handler et al. 2000) and enrichment-induced neurogenesis is impaired in adult mice where the presenilin-1 gene was conditionally knocked out (Feng et al. 2001). Mutations that are associated with early onset forms of AD and result in altered calcium handling could thus have important consequences on NPC cell fate decisions.

We found that inhibitors of calpains and caspases can protect NPC against Aβ-induced death. Previous studies have established the involvement of these two types of cysteine proteases in calcium-mediated apoptosis of various types of cells (Takadera and Ohyashiki 1997; Muller et al. 1999; Nakagawa and Yuan 2000). Interestingly, calcium (Mattson 1990), calpains (Grynspan et al. 1997) and caspases (Chan et al. 1999; Rohn et al. 2001) have each been implicated in neurofibrillary degeneration of neurons in AD, and it has been proposed that Aβ initiates a neurodegenrative cascade that involves over-activation of glutamate and voltage-sensitive calcium channels (Mattson 1997; Selkoe 2001). Our findings suggest that Aβ can cause a similar destabilization of calcium homeostasis and activation of death proteases in neural progenitor cells, despite a lack of glutamate or voltage-sensitive calcium channels on these cells. However, we cannot rule out indirect effects of Aβ on NPC. Subtoxic concentrations of Aβ altered the morphology of glia and reduced the numbers of differentiated neurons. Aβ is known to alter glia cell function (Gitter et al. 1995; Hu et al. 1998), and a recent report suggests that astrocytes can regulate neurogenesis (Song et al. 2002). It is therefore possible that concentrations of Aβ that are not directly toxic to NPC disrupt the normal function of glia and suppress neurogenesis.

The ability of Aβ to alter proliferation and differentiation of NPC suggests a role for perturbed NPC behavior in the pathogenesis of AD. A better understanding of the signals that regulate the NPC proliferation, differentiation and survival may lead to the development of novel approaches for preserving or stimulating neurogenesis in AD and other neurodegenerative disorders.

Ancillary