Role of the cholinergic system in regulating survival of newborn neurons in the adult mouse dentate gyrus and olfactory bulb

Authors

  • Naoko Kaneko,

    1. Department of Physiology, Keio University School of Medicine, Tokyo 160-8582, Japan
    2. Bridgestone Laboratory of Developmental and Regenerative Neurobiology, Keio University School of Medicine, Tokyo 160-8582, Japan
    3. Core Research for Evolutional Science and Technology (CREST), Japan Science and Technology Agency, Saitama 332-0012, Japan
    4. Interdisciplinary Graduate School of Medicine and Engineering, University of Yamanashi, 1110 Tamaho-cho, Nakakoma-gun, Yamanashi, 409-3898, Japan
    Search for more papers by this author
  • Hideyuki Okano,

    Corresponding author
    1. Department of Physiology, Keio University School of Medicine, Tokyo 160-8582, Japan
    2. Core Research for Evolutional Science and Technology (CREST), Japan Science and Technology Agency, Saitama 332-0012, Japan
    Search for more papers by this author
  • Kazunobu Sawamoto

    Corresponding author
    1. Department of Physiology, Keio University School of Medicine, Tokyo 160-8582, Japan
    2. Bridgestone Laboratory of Developmental and Regenerative Neurobiology, Keio University School of Medicine, Tokyo 160-8582, Japan
    3. Core Research for Evolutional Science and Technology (CREST), Japan Science and Technology Agency, Saitama 332-0012, Japan
    Search for more papers by this author

  • Communicated by: Noriko Osumi

* Correspondence: E-mail: sawamoto@sc.itc.keio.ac.jp or hidokano@sc.itc.keio.ac.jp

Abstract

Neurogenesis in the subgranular zone of the hippocampal dentate gyrus and olfactory bulbs continues into adulthood and has been implicated in the cognitive function of the adult brain. The basal forebrain cholinergic system has been suggested to play a role in regulating neurogenesis as well as learning and memory in these regions. Herein, we report that highly polysialylated neural cell adhesion molecule (PSA-NCAM)-positive immature cells as well as neuronal nuclei (NeuN)-positive mature neurons in the dentate gyrus and olfactory bulb express multiple acetylcholine receptor subunits and make contact with cholinergic fibers. To examine the function of acetylcholine in neurogenesis, we used donepezil (Aricept), a potent and selective acetylcholinesterase inhibitor that improves cognitive impairment in Alzheimer's disease. Intraperitoneal administrations of donepezil significantly enhanced the survival of newborn neurons, but not proliferation of neural progenitor cells in the subgranular zone or the subventricular zone of normal mice. Moreover, donepezil treatment reversed the chronic stress-induced decrease in neurogenesis. Taken together, these results suggest that activation of the cholinergic system promotes survival of newborn neurons in the adult dentate gyrus and olfactory bulb under both normal and stressed conditions.

Introduction

Neurogenesis, the production of neurons, continues in two regions of the adult mammalian brain: the subgranular zone (SGZ) of the hippocampal dentate gyrus (DG) and the subventricular zone (SVZ) of the lateral ventricle (Altman 1969; Kaplan & Hinds 1977; Cameron et al. 1993; Doetsch & Alvarez-Buylla 1996; Eriksson et al. 1998). In the DG, newborn neurons are generated in the SGZ, then migrate and differentiate into mature neurons in the granule cell layer (GCL), where they are functionally integrated into the neural circuit in the DG (van Praag et al. 2002). Neural stem cells in the SVZ generate neuroblasts, which migrate through the rostral migratory stream (RMS) towards the olfactory bulb (OB), where they differentiate into interneurons (Luskin 1993; Doetsch & Alvarez-Buylla 1996). Learning enhances neurogenesis (Gould et al. 1999; Rochefort et al. 2002). Conversely, neurogenesis in the DG and OB is involved in memory and learning tasks (Gheusi et al. 2000; Madsen et al. 2003). Therefore, it is likely that neurogenesis and the function of newborn neurons are regulated by common molecular mechanisms.

Acetylcholine (ACh) is one of the most important neurotransmitters in learning and memory (Zaborszky et al. 1986; McGurk et al. 1991). The two neurogenic regions, the DG and the OB, are extensively innervated with basal forebrain cholinergic (BFC) neurons (Wenk et al. 1977; Zaborszky et al. 1986; Fibiger 1991; Changeux et al. 1998). Neurons in the DG and OB abundantly express nicotinic acetylcholine receptors (nAChRs) (Wada et al. 1989; Le Jeune et al. 1995; Quik et al. 2000), a family of ligand gated ion channels composed of five polypeptides (Changeux et al. 1998), and metabotropic muscarinic acetylcholine receptors (mAChRs) (Levey et al. 1995; Rouse et al. 1999), a family of G-protein coupled receptors classified into five subtypes (m1–m5) (Caulfield & Birdsall 1998). Blocking cholinergic input significantly impairs memory and learning tasks (McGurk et al. 1991; Ravel et al. 1994; Leanza et al. 1995), indicating the cholinergic system to play an important role in memory and learning in neurogenic regions. Decreased cholinergic system function also impairs neurogenesis (Cooper-Kuhn et al. 2004; Harrist et al. 2004). In vitro studies indicate that cholinergic stimulation modifies the proliferation and survival of neural precursor cells (Berger et al. 1998; Coronas et al. 2000; Ma et al. 2000). Thus, it is possible that cholinergic input is involved in control of neurogenesis as well as the activity of neurons generated in the OB and DG. However, direct connections of cholinergic fibers to newborn neurons have not been demonstrated.

Herein, we report that cholinergic fibers innervate both the OB and the DG, where neuronal progenitor cells and immature neurons express various subtypes of acetylcholine receptors (AChRs). The data suggest the progenitors and immature neurons to be directly controlled by cholinergic input. To study the effect of enhanced cholinergic neurotransmission on neurogenesis, we used donepezil, a potent acetylcholinesterase inhibitor (ChEI), which is widely used for the treatment of Alzheimer's disease to ameliorate cognitive impairment and memory disturbance. Based on the present results, we demonstrate that cholinergic stimulation promotes survival of newborn neurons in the adult DG and OB.

Results

New neurons express ACh receptors in the adult mouse DG and OB

Previous studies have revealed that various types of nicotinic and muscarinic AChRs are expressed on mature granule cells in the DG and OB (Wada et al. 1989; Le Jeune et al. 1995; Levey et al. 1995; Rouse et al. 1999; Quik et al. 2000). However, little is known about their expression on immature neurons. As a first step to understanding the role of the cholinergic system in adult neurogenesis, we investigated the distribution of major AChRs using antibodies specific for the β2 subunit of nAChR and muscarinic receptors (m1, m2 and m4), and a rhodamine-conjugated ligand to α7nAChR (α-bungarotoxin) in combination with antibodies against NeuN, a marker of mature granule cells, and polysialylated neural cell adhesion molecule (PSA-NCAM), a marker of newborn immature granule cells in the DG and migrating neuroblasts in the SVZ, RMS and OB (Bonfanti et al. 1992; Seki & Arai 1993, 1995). Localization of each AChR on immature and mature neurons was analyzed by confocal laser microscopy. In accordance with previous reports (Wada et al. 1989; Le Jeune et al. 1995; Levey et al. 1995; Rouse et al. 1999; Quik et al. 2000), α7 and β2 nicotinic AChRs and m1 and m4 AChRs were detectable on most NeuN-positive mature granule cells in the DG (Fig. 1F–I) and OB (Fig. 1P–S). In contrast, m2-positive cells were sparse in the GCL in the DG and OB, both of which were negative for NeuN (Fig. 1J,T) and PSA-NCAM (Fig. 1E,O). PSA-NCAM-positive immature neurons in the DG were colabeled with α7 (Fig. 1A; 72.5%, n = 138), β2 (Fig. 1B; 82.7%, n = 104), m1 (Fig. 1C; 57.9%, n = 133) and m4 (Fig. 1D; 35.0%, n = 103). Similarly, PSA-NCAM-positive cells in the GCL of the OB were colabeled with β2 (Fig. 1L; 95.2%, n = 124), m1 (Fig. 1M; 45.8%, n = 107) and m4 (Fig. 1N; 62.0%, n = 158), but rarely with α7 (Fig. 1K; 19.6%, n = 270). Thus, immature neurons in the DG and OB, as well as mature granule cells, expressed multiple types of AChRs, suggesting that newborn neurons can receive cholinergic stimulation directly.

Figure 1.

Expression of AChRs in the DG (A–J) and OB (K–T). Laser scanning confocal images of AChRs (red) labeled with fluorescence-conjugated aBTX (a ligand for α7nAChR) (A,F,K,P) or antibodies for β2nAChR (B,G,L,Q), m1 (C,H,M,R), m4 (D,I,N,S) or m2 (E,J,O,T), in combination with neuronal markers (PSA-NCAM for immature neurons and NeuN for mature neurons) (green) and nuclear staining (Hoechst, blue). Higher magnification views of the regions indicated by rectangles are shown to the right. Co-localization of AChRs and neuronal markers (PSA-NCAM or NeuN) is indicated by arrows. Color sets of high magnification views of nicotinic AChR-positive cells are shown on the lower side (A,B,F,G,K,L,P,Q). Scale bars = 20 µm.

Cholinergic fibers innervate adult neurogenic regions

To study the distribution of cholinergic fibers in adult neurogenic regions, mouse brain sections were stained for ChAT, a marker of cholinergic neurons, and PSA-NCAM. ChAT-positive cholinergic fibers were widely observed throughout the DG (Fig. 2A), consistent with previous reports (Frotscher & Leranth 1985, 1986). Notably, relatively higher densities of the cholinergic fibers were observed in the inner GCL and SGZ, the regions in which immature cells localize (Fig. 2A,B,D), in addition to the molecular layer. In the GCL, cholinergic fibers were seen to be in contact with PSA-NCAM-positive postmitotic neurons (Fig. 2D,E, arrows) as well as PSA-NCAM-negative mature granule neurons. ChAT-positive fibers were also recognized on PSA-NCAM-positive cells in the SGZ (Fig. 2B,C, arrows). Similarly, cholinergic fibers were frequently observed to be in contact with PSA-NCAM-positive neuroblast in the OB (Fig. 2F–H, arrows). Quantitative analysis revealed that about 80% (350 cells/436 cells) of PSA-NCAM-positive neuroblasts in the DG and 40% (77 cells/181 cells) of those in the OB had these contacts. We found no ChAT-positive fibers in the SVZ (Fig. 2I,J), although there were abundant ChAT-positive fibers in the striatum adjacent to the SVZ. In the RMS, which contains neuroblasts migrating from the SVZ towards the OB, ChAT-positive fibers were rarely observed between the cells (Fig. 2K,L). Thus, cholinergic fibers innervate and contact immature neurons in the OB and DG.

Figure 2.

Cholinergic innervation of neurogenic regions in the adult mouse brain. Double-immunofluorescence of ChAT (green), a marker for cholinergic neurons, and PSA-NCAM (red), a marker for immature neurons in the DG (A–E), OB (F–H), SVZ (I,J) and RMS (K,L). All nuclei are labeled with Hoechst (blue). (A) ChAT-positive fibers are distributed throughout the DG including the GCL and SGZ. (B,D) High magnifications of the GCL and SGZ in which immature neurons reside. ChAT-positive fibers are dispersed adjacent to or in contact with PSA-NCAM-positive immature neurons in the SGZ (B, arrows) and relatively mature PSA-NCAM-positive cells in the GCL (D, arrow). (C,E) Higher magnifications of (B) and (D). Contacts between ChAT-positive fibers and PSA-NCAM-positive cells are indicated with arrows. (F) ChAT-positive fibers are widely distributed throughout the OB, especially prominently in the outer GCL and mitral cell layer. (G) Higher magnification of the GCL. Numerous ChAT-positive fibers in contact with PSA-NCAM negative and positive cells in the GCL can be seen. (H) A higher magnification of (G). Contacts between ChAT-positive fibers and PSA-NCAM-positive cells are indicated with arrows. (I,J) ChAT-positive fibers are abundant in the striatum, but undetectable in the SVZ where PSA-NCAM positive cells are present. (K,L) In the RMS, a few ChAT-positive fibers are observed but they are not in contact with PSA-NCAM-positive neuroblasts. Scale bars = 100 µm (A,D); 50 µm (G,I,K); 20 µm (B,E,H,J,L); 5 µm (C,F). gcl, granule cell layer; sgz, subgranular zone; ml, molecular layer; mcl, mitral cell layer; rms, rostral migratory stream; LV, lateral ventricle.

ChEI-treatment promotes survival of newborn neurons in the adult brain

To study the function of ACh in regulating adult neurogenesis, we examined the effects of long-term treatment with donepezil (Aricept), a selective, non-competitive and long-lasting inhibitor of acetylcholinesterase, which continuously increases extracellular ACh in the brain and ameliorates cognitive impairments in Alzheimer's disease (Rogers & Friedhoff 1996; Rogers et al. 1998). To investigate the effect of an increased ACh level on cell proliferation, BrdU was administered at the completion of ChEI-treatment, and numbers of BrdU-labeled cells were then counted 24 h later (Fig. 3A). There was no significant difference in the number of BrdU-positive cells labeled 24 h before fixation in the DG, SVZ, RMS and OB between the control and ChEI-injected groups (P > 0.05) (Fig. 3B–G), indicating that ChEI-treatment does not affect cell proliferation. To investigate the effect of ChEI-treatment on the survival of newborn cells, BrdU was injected at the beginning of ChEI-treatment, and numbers of remaining BrdU-labeled cells were then counted (Figs 4A and 5A). Numbers of BrdU-positive cells in the DG, labeled at the beginning of treatment, were significantly increased after 1 week of treatment with 1.0 mg/weight (kg)/day of donepezil (P = 0.0232), and 4-week treatment at doses of 1.0 and 2.0 mg/weight (kg)/day (P = 0.0420) (Fig. 4B–D). They were also increased in the OB after 4 weeks (P < 0.0001) (Fig. 5B,C,E, right), but not 1 week, of ChEI-treatment at all doses used (P > 0.05) (Fig. 5D, right). The 4-week ChEI-treatment also slightly increased the number of BrdU-positive cells in the RMS (Fig. 5E, middle). However, no significant difference was observed between these groups in the SVZ (P > 0.05) (Fig. 5E, left). These data suggest that ChEI-treatment promotes the survival of newborn cells in both the DG and the OB, and the finding that 1-week treatment increased the survival of newborn cells in the DG implies that this treatment affects not only mature but also immature cells. Nuclear staining revealed that 4 weeks of ChEI-treatment decreased the number of pyknotic cells in the DG (P = 0.0001) (Fig. 4E) and OB (P = 0.0006) (Fig. 5F). There were no significant differences in the number of Ki67-positive proliferating cells with 4 weeks treatment in any of the regions examined (P > 0.05, respectively) (Figs 4F and 5G). Thus, ChEI-treatment promotes the survival of newborn cells without affecting the proliferative capacity.

Figure 3.

ChEI does not affect cell proliferation. (A) Experimental procedure for chronic ChEI-treatment. Mice were intraperitoneally administered either saline (Sal, n = 6), as a control, or ChEI (donepezil, 0.5, 1.0 or 2.0 mg/weight (kg)/day, n = 5) once a day for 2 consecutive weeks, injected with BrdU at the completion of the ChEI-treatment, and then transcardially perfused 24 h later. (B,C,E,F) BrdU labeled cells in the DG (B,C) and SVZ (E,F) of mice injected with (B,E) saline or (C,F) ChEI (donepezil, 1.0 mg/weight (kg)/day). (D,G) There are no significant differences in the number of BrdU-positive proliferating cells in the DG (D), SVZ (G, left), RMS (G, middle) or OB (G, right) between the saline- and ChEI-injected groups (P > 0.05, respectively). LV, lateral ventricle. Scale bars = 100 µm.

Figure 4.

ChEI promotes cell survival in the DG. (A) Experimental procedure. The animals received BrdU injections at the beginning of the procedure, followed by either 1 or 4 weeks of treatment (i.p) with saline (Sal, n = 5) or ChEI (donepezil, 0.5, 1.0 or 2.0 mg/weight (kg)/day, n = 4 or (5) before perfusion. (B,C) Photographs of BrdU-labeled cells in the DG of mice injected with saline (B) or ChEI (C) for 4 weeks. Scale bars = 100 µm. (D) Number of BrdU-labeled cells in the DG. In the DG, 1 week of treatment with 0.5 or 1.0 mg/weight (kg)/day (left), and 4 weeks of treatment with 1.0 mg/weight (kg)/day of donepezil (right) significantly increased BrdU-labeled cells (1 week; P = 0.0232, 4 weeks; P = 0.0420). (E) The density of pyknotic cells in the DG was significantly decreased in mice treated with ChEI for four weeks (P = 0.0001). (F) The number of cells expressing Ki67, a marker of proliferating cells, in the DG of mice injected with ChEI or saline for 4 weeks. There is no significant difference between the groups (P > 0.05).

Figure 5.

ChEI promotes cell survival in the RMS and OB. (A) Experimental procedure. The animals received BrdU injections at the beginning of the procedure, followed by either 1- or 4-week treatment (i.p.) with saline (Sal, n = 5) or ChEI (donepezil, 0.5, 1.0 or 2.0 mg/weight (kg)/day, n = 4 or 5) before perfusion. (B,C) Photographs of BrdU-labeled cells in the OB of mice injected with (B) saline or (C) ChEI for 4 weeks. Scale bars = 100 µm. (D) Numbers of BrdU-labeled cells in the SVZ-OB of mice treated for 1 week. There were no significant differences between groups in any of the regions examined (P > 0.05). (E) Numbers of BrdU-labeled cells in the SVZ-OB of mice treated with ChEI or saline for 4 weeks. The numbers of BrdU-labeled cells were significantly increased in the RMS (middle, P = 0.0198) and OB (right, P < 0.0001) by 4 weeks of treatment with ChEI. (F) The number of pyknotic cells in the OB was significantly decreased in mice treated with ChEI for 4 weeks (P = 0.0006). (G) The number of cells expressing Ki67, a marker of proliferating cells, in the SVZ (left), RMS (middle) and OB (right) of mice injected with ChEI or saline for 4 weeks. There is no significant difference between the groups (P > 0.05).

To further investigate the impact of ChEI-treatment on neurogenesis, we examined the effect of 4 weeks of ChEI-treatment (Figs 4A and 5A) on the numbers of NeuN-positive mature neurons labeled with BrdU (Fig. 6). In both the DG and the OB, numbers of BrdU+NeuN+ neurons were significantly increased by ChEI-treatment (DG, Fig. 6A–D, left, P = 0.0392, OB, Fig. 6E–G,H, left, P = 0.0059). There was no significant difference in the percentage of NeuN+ cells in the BrdU+ cell population between the control and ChEI-treated groups (DG, Fig. 6D, right; OB, Fig. 6H, right; P > 0.05, respectively). These results indicate that ChEI-treatment increases the supply of new neurons in the DG and OB without affecting NeuN expression in newborn cells. Taken together, the ChEI-treatment increased neurogenesis by promoting the survival of newborn cells.

Figure 6.

ChEI increases the number of newborn neurons in the DG and OB. (A,B,E,F) Double-immunofluorescence for BrdU (red) and NeuN (green) in the DG (A,B) and OB (E,F) of mice injected with saline (Sal) or ChEI (donepezil, 1.0 mg/weight (kg)/day) for 4 weeks (see Figs 4A or 5A), and higher magnifications of double-labeled cells in (C) the DG and (G) the OB. Scale bars = 50 µm (A,B,E,F); 10 µm (C); 5 µm (G). (D) In the DG, the number of newborn neurons labeled with BrdU and NeuN was significantly increased in mice injected with ChEI (donepezil, 1.0 mg/weight (kg)/day) for 4 weeks (left, P = 0.0392). There is no significant change in the proportion of NeuN-positive neurons in the BrdU-labeled cell population (right, P > 0.05). (H) In the OB, the number of newborn neurons labeled with BrdU and NeuN was significantly increased in mice injected with ChEI (left, P = 0.0059). There is no significant change in the proportion of NeuN-positive neurons in the BrdU-labeled cell population (right, P > 0.05).

ChEI-treatment reverses the stress-induced decrease in adult neurogenesis

It has been reported that exposure to chronic stress impairs hippocampal neurogenesis as well as hippocampus-dependent memory and learning (Luine et al. 1994; Bodnoff et al. 1995; Nishimura et al. 1999; Czeh et al. 2002; Pham et al. 2003). Acute-stress-induced increases in glucocorticoids, in which the BFC afferent to the hippocampus functions as an important negative-feedback system (Sithichoke & Marotta 1978; Han et al. 2002; Helm et al. 2002), are reportedly involved in the stress-induced decrease in cell-proliferation (Cameron et al. 1995; Gould et al. 1997; Tanapat et al. 2001). However, the mechanisms underlying the inhibition of neurogenesis caused by chronic stress remain unknown. To investigate the effects of cholinergic stimulation on the stress-induced decrease in neurogenesis, mice were exposed to chronic restraint while receiving ChEI-treatment. We chose 1.0 mg/weight (kg)/day of donepezil, as this dose showed the maximal promoting effect on the survival on newborn cells in normal mice (Figs 4–6). Following 1 week of treatment with saline (i.p., for mice in the control and the restraint group) or ChEI (i.p., for mice in the ChEI-treated group), BrdU was injected. Mice were then placed in a close-fit restrainer for 6 h per day for 4 consecutive weeks (Fig. 7A), and were treated immediately before each restraint session with saline or ChEI. Plasma corticosterone levels measured at the end of the experimental period, before the last restraint session, did not differ significantly between the groups (P > 0.05) (Fig. 7B). Consistent with a previous study (Tanapat et al. 2001), 4 weeks of restraint stress decreased Ki67-positive proliferating cells in the DG (Fig. 7C, P = 0.0009), but not the SVZ, RMS or OB (Fig. 7D, P > 0.05). The ChEI-treatment did not affect the number of Ki67-positive cells (Fig. 7C,D). The number of BrdU+NeuN+ newborn neurons in the DG and OB were decreased by restraint stress, indicating that stress reduces the survival of young neurons. However, ChEI-treatment almost completely reversed the suppressive effect of stress on neuronal survival (Fig. 7E–L) (DG, P = 0.0029, OB, P = 0.0161). There was no significant difference in the percentages of NeuN+ neurons in the BrdU+ cell population between these groups (P > 0.05) (Fig. 7H,L). Taken together, these data indicate that chronic ChEI-treatment reverses the stress-induced decrease in the survival of newborn cells, while having no effect on proliferation.

Figure 7.

ChEI restores neurogenesis in the DG and OB of mice exposed to chronic restraint. (A) The experimental procedure for the study of the effect of ChEI-treatment on mice exposed to chronic restraint. Following 1 week of administration of saline to habituate the mice to intraperitoneal injections, the animals were divided into three groups; during the week following saline administration, mice were injected with BrdU twice with a 2 h-interval. Subsequently, the mice in the control group (Cont) were left in their home cages without treatment for 4 weeks, while in the restraint group (Rest), mice were restrained for 6 h every day for 4 weeks. In the restraint group with ChEI-treatment (ChEI), mice were pretreated with ChEI (donepezil, 1.0 mg/weight (kg)/day), followed by BrdU injection as described above, and then restrained for 4 weeks. (B) Corticosterone levels in plasma samples taken before restraint on the final day of the experiments. There is no significant difference in plasma corticosterone levels between the groups (P > 0.05). (C) The number of Ki67-positive proliferating cells in the DG was significantly decreased by chronic restraint but was not affected by ChEI-treatment (P = 0.0009). (D) The numbers of Ki67-positive cells in the SVZ (left), RMS (middle) and OB (right). None of the regions examined showed any significant difference in the number of Ki67-positive cells between the groups (P > 0.05). (E–L) Double immunofluorescence for BrdU (red) and NeuN (green) in the DG (E, Cont; F, Rest; G, ChEI) and OB (I, Cont; J, Rest; K, ChEI). Chronic restraint significantly decreased the number of newborn neurons labeled with BrdU in the DG (H, upper) and OB (L, upper). The decrease in neurogenesis was completely reversed by chronic ChEI-treatment (DG, P = 0.0029; OB, P = 0.0161). The proportions of NeuN-positive neurons in the BrdU-positive cell population in the DG (H, lower) and OB (L, lower) were not affected by either restraint or ChEI-treatment (P > 0.05). Scale bars = 100 µm.

Discussion

Previous studies have indicated that approximately half of newborn neurons die before reaching maturity in the DG and OB (Petreanu & Alvarez-Buylla 2002; Dayer et al. 2003), but the mechanisms regulating the apoptosis of these neurons are largely unknown. Herein, we report that the cholinergic system promotes the survival of newborn neurons in the DG and OB. Cholinergic neurons innervating the DG and OB play an important role in learning and memory formation (Wenk et al. 1977; Zaborszky et al. 1986; McGurk et al. 1991; Leanza et al. 1995). Alterations in these processes affect neurogenesis in the DG and OB (Gould et al. 1999; Gheusi et al. 2000; Shors et al. 2001; Rochefort et al. 2002; Madsen et al. 2003), raising the possibility that the cholinergic system may act on neurogenesis directly and/or indirectly. An indirect effect of cholinergic input on neurogenesis has been proposed based on the observation that lesions in the BFC system reduce expression of brain-derived neurotrophic factor (BDNF) (Hefti 1986; Williams et al. 1986; Berchtold et al. 2002), which promotes neurogenesis (Zigova et al. 1998; Lee et al. 2002). Our results indicate that immature neurons in the adult mouse DG and OB express various ACh receptor subtypes (Fig. 1), like mature neurons in the DG and OB (Wada et al. 1989; Le Jeune et al. 1995; Levey et al. 1995; Rouse et al. 1999; Quik et al. 2000). The immature neurons in the DG and OB are in contact with cholinergic fibers (Fig. 2), and are increased by ChEI-treatment (Figs 3–6). In contrast, the SVZ was found to be devoid of cholinergic innervation (Fig. 2) and insensitive to ChEI-treatment in our experiments (Figs 3 and 5). These results suggest that there is functional cholinergic transmission in the immature neurons in the DG and OB, which is involved in regulating adult neurogenesis.

Using PSA-NCAM as a marker for immature neurons and neuroblasts in combination with specific antibodies or ligands for various AChRs, we comprehensively studied the distributions of major AChRs in the DG and OB (Fig. 1). m1 and m4 were recently reported to be expressed by immature DG cells labeled with BrdU 24 h after injection (Mohapel et al. 2005). Consistently, a subpopulation of PSA-NCAM–positive cells in the DG was positive for m1 and m4 AChRs. Moreover, these cells also expressed nicotinic (α7 and β2) AChRs. We further demonstrated that PSA-NCAM–positive migrating neuroblasts in the OB similarly expressed several types of nicotinic and muscarinic receptors. These findings suggest that immature neurons can receive cholinergic stimulation directly. In fact, these AChR subtypes play roles in various regulatory processes for neural progenitor cells. α7, a major nAChR subunit most abundantly distributed in the DG (Seguela et al. 1993; Quik et al. 2000), is involved in neuroprotection against excitotoxic stimulation (Kihara et al. 1997; Shimohama et al. 1998; Prendergast et al. 2001) and participates in neurite outgrowth in olfactory cultures (Coronas et al. 2000), which is important for the survival of newborn neurons. In fact, previous studies have demonstrated several molecular mechanisms involving nicotine, via the α7 nicotinic receptor, which promote the survival of cultured neurons. Nicotine attenuates the expression of pro-apoptotic proteins such as caspase3, 8 and 9, and increases the levels of anti-apoptotic factors including phosphorylated Akt and Bcl-2 (Garrido et al. 2001; Kihara et al. 2001), possibly involved in the effect of ChEI in promoting the survival of newborn neurons as demonstrated in this study. Herein, α7nAChRs were detectable in almost all mature granule cells in both the OB and the DG. The proportion of α7-positive cells among PSA-NCAM-positive immature neurons was much lower in the OB (less than 20%) than that in the DG (over 70%), probably because most new neurons generated in the SVZ do not express α7 until they have differentiated into mature interneurons. Indeed, a previous study with retroviral labeling indicated that neuroblasts take about 1 week to reach the deeper GCL in the OB (Petreanu & Alvarez-Buylla 2002). Consistently, immature neurons in the OB showed delayed reactivity to ChEI-treatment as compared to those in the DG; 1 week treatment promoted the survival of newborn cells only in the DG, while longer treatment (for 4 weeks) affected both the DG and the OB. Previous reports indicate that AChRs are also involved in regulating cell-proliferation (Ma et al. 2000; Abrous et al. 2002; Jang et al. 2002), although we found no significant differences in the numbers of proliferating cells among our ChEI-treated animals. Thus, immature neurons express multiple AChRs that have different effects on neurogenesis. ChEI-treatment increases the extracellular ACh concentration, which should stimulate all of the AChRs available on immature neurons and neuroblasts. Therefore, the results of ChEI-treatment experiments may reflect effects on multiple receptors with distinct functions expressed by immature neurons.

Our results indicate that chronic cholinergic stimulation with ChEI-treatment promotes the survival of newborn neurons in the DG and OB (Figs 4–6), but does not affect the proliferation of progenitor cells (Fig. 3). In contrast, acute treatment with another ChEI (physostigmine) was reported to increase cell proliferation, but not to affect the long-term survival of newborn cells in the DG (Mohapel et al. 2005). We used donepezil, a potent ChEI widely used for treatment of Alzheimer's disease, which has a longer-lasting action and greater specificity for acetylcholinesterase than any other ChEIs currently available, while exerting a minimal impact on the peripheral nervous system (Kosasa et al. 1999). These pharmacological properties make donepezil highly suitable for studying adult neurogenesis, considering that changes in physical conditions such as voluntary activity and nutrition reportedly modify neurogenesis (van Praag et al. 1999; Brown et al. 2003; Mattson et al. 2003). The difference in treatment duration may also account for the discrepancy between the results of these two studies. ChEI rapidly increases the extracellular ACh concentration, within several hours after a single administration (Kosasa et al. 1999). However, it often takes several months to exert clinical effects such as memory and cognitive improvement (Rogers et al. 1998). Although the precise mechanisms underlying this long-term effect of ChEIs have not been demonstrated, chronic ChEI-treatment up-regulates nicotinic AChRs and down-regulates muscarinic AChRs (Nilsson-Hakansson et al. 1990; Barnes et al. 2000), which is likely to result in changes in the amount and/or balance of signals mediated via each AChR. Thus, the protective effect of long-term ChEI-treatment on immature neurons observed in this study may be highly relevant to the clinical efficacy of ChEIs.

The extracellular ACh level in the brain varies depending on various changes which occur under physiological conditions. Learning, motor activity and estrogen treatment increase ACh release in the cortex and hippocampus (Day et al. 1991; Mizuno et al. 1991; Gabor et al. 2003). All of these conditions also promote the survival of new neurons and enhance neurogenesis in the DG and/or SVZ (Gould et al. 1999; van Praag et al. 1999; Banasr et al. 2001; Smith et al. 2001), suggesting the increased ACh level to be at least partially responsible for the increased neurogenesis. Conversely, selective lesioning of the BFC system reduces neurogenesis in the adult rat brain (Calza et al. 2003; Cooper-Kuhn et al. 2004). Therefore, it is possible that the cholinergic system is among the physiologically important mechanisms regulating adult neurogenesis. Moreover, nAChRs in these regions are reduced in response to various insults including chronic stress, aging, traumatic injury and degenerative disorders such as Alzheimer's disease (Banerjee et al. 2000; Burghaus et al. 2000; Rei et al. 2000; Verbois et al. 2002, 2003; Gahring et al. 2005). BFC neurons are vulnerable to inflammatory processes induced by these lesions (Wenk et al. 2003). Thus, the cholinergic system can be impaired in various situations causing a decrease in neurogenesis. Therefore, increasing the ACh level using ChEI is likely to be a promising strategy for reversing the decrease in neurogenesis resulting from these insults. Indeed, our data indicate that an activated BFC system is capable of restoring neurogenesis in a restraint-induced stress model. ChEI-treatment improves the survival of newborn neurons in the DG and OB (Fig. 7). If the increased survival of newborn neurons results in functional recovery, ChEI-treatment may be useful for increasing neuronal plasticity and facilitating neuronal regeneration in injury/aging as well as under normal conditions.

Experimental procedures

Animals

Ten week-old male C57BL6 mice (SLC, Shizuoka, Japan) were group-housed under controlled conditions (21 °C room temperature, 12 h light/12 h dark cycle) with food and water available ad libitum. All animal experiments were approved by the Laboratory Animal Care and Use Committee of Keio University School of Medicine.

ChEI-treatment

Donepezil (Aricept) was generously provided by Eisai Co., Ltd, Japan. Mice were intraperitoneally injected with saline (control, n = 5 or 6) or donepezil (0.5, 1 or 2 mg/weight (kg)/day, n = 4 or 5) for 1, 2 or 4 weeks consecutively (see Figs 3–5 for details).

Chronic stress protocol

Mice were separated into three groups. In the restraint group, mice were injected (i.p.) with either ChEI (donepezil 1.0 mg/weight (kg)/day, n = 4) or saline (n = 5) for 1 week, injected twice with bromodeoxyuridine (BrdU, Sigma, St. Louis, MO, USA) (50 mg/kg, i.p.) with a 2-h interval, and then placed in a close-fit cylindrical restrainer (acrylic tube 2.8 cm in diameter with ventilating holes on the top and bottom to avoid elevating body temperature) inside their home cage for 6 h per day for 4 consecutive weeks. To decrease effects on the motor activities and food intakes of the mice, the restraint was performed during the light period, while the mice were inactive. ChEI or saline was injected immediately before each restraint period. Mice in the control group (n = 5) were treated with saline and BrdU for 1 week as described above, and then left undisturbed in their home cages for another 4 weeks. Before the final restraint, blood was sampled for measurement of the plasma corticosterone level using an enzyme immunoassay kit (Assay Designs, Ann Arbor, MI, USA).

BrdU labeling

To assess the effects of the donepezil treatments on cell-proliferation in the DG, SVZ, RMS and OB, mice were intraperitoneally injected with BrdU (50 mg/kg) immediately after 2 weeks of donepezil treatment, and allowed to survive for another 24 h before perfusion. To assess the differentiation and survival of the newborn cells, mice were injected with BrdU (50 mg/kg) twice with a 2-h interval at the beginning of the donepezil treatment or restraint stress exposure.

Immunohistochemistry

Mice were deeply anesthetized with diethyl ether and transcardially perfused with PBS, followed by 4% paraformaldehyde in 0.1 m phosphate buffer. The brains were removed and postfixed in the same fixative for 24 h at 4 °C. Fifty-micrometer serial sections were coronally cut on a vibratome (VT1000S, Leica, Heidelberg, Germany) and collected in PBS.

For AChR-immunostainings (Fig. 1), sections were incubated with antibodies against muscarinic AChR m1 (m1AChR, rabbit, 1 : 300, Sigma), m2 (m2AChR, rabbit, 1 : 300, Alomone labs, Israel) or m4 (m4AChR, rabbit, 1 : 300, Santa Cruz Biotechnology, Santa Cruz, CA, USA), or the β2 subunit of nicotinic AChR (β2nAChR, rabbit, 1 : 100, Santa Cruz) combined with antibodies, against the highly polysialylated neural cell adhesion molecule (PSA-NCAM) or neuronal nuclei (NeuN), at 4 °C overnight. After washing with 0.1% TritonX-100 in PBS (PBS-T), the sections were incubated with biotinylated anti-rabbit IgG antibody (goat, 1 : 500, Jackson) for 2 h, followed by signal amplification with the avidin-biotin-complex (ABC)-system (Vectastain ABC Elite kit, Vector Laboratories, Burlingame, CA, USA) for 1 h, and visualized with rhodamine-labeled thyramide (PerkinElmer, Boston, MA, USA). The sections were then incubated in Alexa Fluor 488-conjugated anti-mouse IgM or IgG antibody (goat, 1 : 500, Molecular Probes) for 2 h. To detect α7nAChR, sections were incubated in rhodamine-conjugated α-bungarotoxin (α-BTX) for 2 h at 37 °C, followed by incubation in antibodies for PSA-NCAM or NeuN overnight at 4 °C. After washing, the sections were incubated with Alexa Fluor 488-conjugated anti-mouse IgM or IgG antibodies (goat, 1 : 500, Molecular Probes) mixed with Hoechst 333442 (Sigma) for 2 h.

For choline acetyltransferase (ChAT) immunohistochemistry (Fig. 2), sections were pretreated in cold (−20 °C) methanol for 6 min, and then boiled in 0.01 m citric acid (pH 6.0) for 30 s. Following treatment with 1% H2O2 for 1 h, the sections were preincubated in TNB blocking reagent (PerkinElmer) for 2 h at room temperature, and incubated overnight in anti-ChAT antibody (goat, 1 : 100, Chemicon International, Temacula, CA, USA) mixed with anti-PSA-NCAM antibody (mouse, 1: 2000, kindly provided by Dr Tatsunori Seki) (Seki & Arai 1993) at 4 °C. After washing with PBS-T, sections were incubated with biotinylated anti-goat IgG (donkey, 1 : 500, Jackson Immuno Research Laboratories, West Grove, PA, USA) for 2 h. The ChAT signal was amplified using an ABC-system (Vector Laboratories) and visualized with fluorescein-labeled thyramide (TSA Fluorescence Systems, PerkinElmer). These sections were then incubated with Alexa Fluor 568-conjugated anti-mouse IgM or anti-mouse IgG antibody (goat, 1 : 500, Molecular Probes, the Netherlands), and Hoechst 33342 (Sigma) to stain the nuclei for 2 h at room temperature.

For BrdU staining (Figs 3–5), a series of floating sections, with 250-µm intervals, from each animal were incubated in 2 N HCl for 30 min at 60 °C, and then incubated with anti-BrdU antibody (mouse, 1 : 200, Becton Dickinson, Franklin Lakes NJ, USA) overnight at 4 °C. After washing with PBS-T, the sections were incubated in biotinylated anti-mouse IgG antibody (1 : 500, Jackson). The signal was visualized using an ABC Elite kit (Vector Laboratories) and diaminobenzidine (DAB).

For double-labeling with BrdU and NeuN (Figs 6 and 7), following the incubation in HCl, sections were incubated in antibodies for BrdU (rat, 1 : 200, Abcam Ltd, Cambridge, UK) and NeuN (mouse, 1 : 200, Chemicon) overnight at 4 °C, then incubated in biotinylated anti-rat IgG antibody (donkey, 1 : 500, Jackson) and finally Alexa Fluor 488-conjugated anti-mouse IgG antibody (goat, 1 : 500, Molecular Probes) for 2 h at room temperature. The BrdU signal was amplified with an ABC system, and visualized by rhodamine-labeled thyramide (PerkinElmer).

For Ki67 staining (Figs 5 and 7), sections were incubated in anti-Ki67 antibody (rabbit, 1 : 800, Novocastra, Newcastle, UK) overnight at 4 °C, and then in Alexa Fluor 488-conjugated anti-rabbit antibody (goat, 1 : 500, Molecular Probes) for 2 h.

Image processing and quantification

All the slides were coded before quantitative analysis, and the code was not broken until the analysis was complete. Cell counting was consistently performed by the same investigator (NK) blind to the group identification of each section. BrdU-labeled cells in the GCL and SGZ in the DG were counted under 400× magnification (Axioplan 2, Carl Zeiss, Germany). Cell numbers were counted using six or seven 50 µm-thick sections (250 µm apart) per animal. Therefore, the total number of cells in the DG was calculated by multipling the number of cells counted in these sections by six. For the SVZ, RMS and OB, absolute cell numbers were counted and presented. BrdU-labeled cells and Ki67 positive cells in the DG (SGZ and GCL), SVZ, RMS and OB were counted at 400× magnification under a microscope (Axioplan 2, Carl Zeiss, Germany). Pyknotic cells were identified and counted using DG and OB sections stained with Cresyl violet. The areas of the GCL and SGZ in each section were measured using Photoshop (Adobe) to calculate the density of pyknotic cells In the DG. BrdU-labeled cells in the DG and five randomly chosen areas scanned in the OB were captured by confocal laser microscopy (Carl Zeiss, LSM510, C-apochromat 40×/1.20 W objective and C-apochromat 63x/1.20 W objective lenses) to confirm double-labeling with NeuN. Cells were considered double-labeled if colabeling was seen throughout the extent of the nucleus on 1 µm optical sections. Co-localizations of AChR signals with neuronal markers (PSA-NCAM and NeuN) were analyzed on 1.0 µm optical sections, and contacts of ChAT-positive cholinergic fibers on to PSA-NCAM-positive immature neurons were analyzed on 0.4 µm optical sections using confocal laser microscopy (LSM510; Carl Zeiss).

Statistical analysis

The obtained data were expressed as mean ± SE. Differences between means were determined by one-way anova, followed by the Bonferroni post hoc multiple comparison test. Differences were regarded as statistically significant when P < 0.05.

Acknowledgements

We are grateful to T. Yamauchi, T. Hashimoto and H. Ogura (Eisai Company Ltd.) for valuable discussions and providing donepezil and AChR antibodies, Tatsunori Seki (Juntendo University) for PSA-NCAM antibody, Jose Manuel Garcia-Verdugo (University of Valencia) for comments on the figures, and members of our laboratories for helpful advice and encouragement. This work was supported by Bridgestone Corporation and grants from The Ministry of Education, Culture, Sports, Science and Technology of Japan, The Ministry of Health, Labour, and Welfare of Japan, Mitsui Life Social Welfare Foundation, and the Japan Science and Technology Agency (CREST).

Ancillary