Neuropeptide Y stimulates neuronal precursor proliferation in the post-natal and adult dentate gyrus

Authors


Address correspondence and reprint requests to Dr W. P. Gray, Division of Clinical Neurosciences, Room 6207, Level 6, Biomedical Sciences Building, Bassett Crescent East, Southampton SO16 7PX, UK. E-mail: w.p.gray@soton.ac.uk

Abstract

Adult dentate neurogenesis is important for certain types of hippocampal-dependent learning and also appears to be important for the maintenance of normal mood and the behavioural effects of antidepressants. Neuropeptide Y (NPY), a peptide neurotransmitter released by interneurons in the dentate gyrus, has important effects on mood, anxiety-related behaviour and learning and memory. We report that adult NPY inline image receptor knock-out mice have significantly reduced cell proliferation and significantly fewer immature doublecortin-positive neurons in the dentate gyrus. We also show that the neuroproliferative effect of NPY is dentate specific, is Y1-receptor mediated and involves extracellular signal-regulated kinase (ERK)1/2 activation. NPY did not exhibit any effect on cell survival in vitro but constitutive loss of the Y1 receptor in vivo resulted in greater survival of newly generated neurons and an unchanged total number of dentate granule cells. These results show that NPY stimulates neuronal precursor proliferation in the dentate gyrus and suggest that NPY-releasing interneurons may modulate dentate neurogenesis.

Abbreviations used:
BrdU

5-bromo-2deoxyuridine

DAPI

6-diamidino-2-phenylidole

DIV

days in vitro

ECT

electroconvulsive therapy

ERK

extracellular signal-regulated kinases

GFAP

glial fibrillary acidic protein

i.p.

intraperitoneal

MAPK

mitogen-activated protein kinases

NPY

neuropeptide y

PFA

paraformaldehyde

SDS

sodium dodecyl sulfate

SGZ

subgranular zone

TBS

tris-buffered saline

Neuropeptide Y (NPY), a 36-amino-acid polypeptide, is one of the most highly conserved neuropeptides in evolution (Larhammar 1996) and is widely expressed in neural tissue of the central and peripheral nervous systems. NPY exerts a multiplicity of biological actions, including modulation of anxiety (Thorsell et al. 2000), social and feeding behaviour (de Bono and Bargmann 1998; Sokolowski 2003; Wu et al. 2003), circadian rhythm (Albers and Ferris 1984; Albers et al. 1984) and seizure control (Vezzani et al. 1999), all mediated via G-protein coupled receptors designated as Y1, Y2, Y4, Y5 and Y6 (Michel et al. 1998). Of particular interest are studies of NPY on cognitive function where NPY infusion enhances memory retention and recall (for review see Redrobe et al. 1999). NPY has also been implicated in the pathogenesis of depression (Caberlotto et al. 1999; Husum and Mathe 2002) and the antidepressant-like activity of NPY has recently been identified as being mediated via its Y1 receptor (Redrobe et al. 2002).

The hippocampus has been repeatedly implicated in the modulation of cognition in association with learning and memory, the dentate gyrus of which shows particularly high NPY-like immunoreactivity (Dumont et al. 1992). The dentate subgranular zone (SGZ) of the hippocampus is also one of the few restricted sites for neurogenesis in the adult brain (Altman and Das 1965) and the demonstration of functional integration of newly born granule cells into hippocampal circuitry (Scharfman et al. 2000; van Praag et al. 2002) has led to the hypothesis that new neurons may be important in learning and memory formation (Gould et al. 1999; Shors et al. 2001; Snyder et al. 2001). Several avenues of research have suggested that reduced hippocampal neurogenesis may also be important in the pathogenesis of depression (Gould et al. 1997; Jacobs et al. 2000; Kempermann 2002) (for review see D'Sa and Duman 2002).

Given the established role of NPY and the emerging role of adult neurogenesis in mood control and hippocampal learning, combined with the demonstration that NPY is proliferative for neuroblasts in the olfactory epithelium of adult mice (Hansel et al. 2001), we have investigated a possible role for NPY in hippocampal neurogenesis. We have previously demonstrated that NPY is proliferative in vitro for nestin-positive cells and neuroblasts in cultures derived from the whole hippocampus of perinatal rat pups via the NPY Y1 receptor (Howell et al. 2003). However, subhippocampal selective culture studies have shown that precursor cells in the dentate gyrus differ from those in the hippocampus proper and the adjacent subventricular zone in their ability to generate neurons and their responses to growth factors (Seaberg and van der Kooy 2002; Becq et al. 2004). We wished to see if NPY was proliferative for neuroblasts from the dentate gyrus in perinatal animals and also if NPY had any effect on dentate proliferation in the adult animal in vivo. We therefore used a combination of in vitro studies using dentate gyrus specific cultures dissected from perinatal rat hippocampus and in vivo studies in adult mice lacking the Y1 receptor or lox/lox controls to establish the effect of NPY on dentate gyrus neurogenesis.

Materials and methods

Animals

All animal experimentation was approved by a University Bio-Ethics Committee and performed under UK Home Office or NY State Department of Health Guidelines. The NPY Y1 knockout mice and the age, sex and generation matched Y1lox/lox controls used in these experiments have been previously reported (Howell et al. 2003; Karl et al. 2004).

Micro-dissected dentate and hippocampus cultures

Fifty-six Wistar rat pups (8–10 days old) were decapitated, the hippocampus dissected out in cold Gey's salt solution (Life Technologies, Paisley, UK) with 4.5 mg/mL glucose at 4°C. Using a pair of plastic spatulas, the corpus callosum was bisected and each cortical hemisphere rolled back. Each hippocampus was separated from the overlying cortical white matter along the natural line of the alveus hippocampus. Transverse sections (1 mm) of hippocampus were then cut on a McIlwain tissue chopper and placed in 5 mL of Gey's solution in a sterile Petri dish. A microdissection method (Seaberg and van der Kooy 2002) was used to dissect the dentate gyrus from the slices. Under a dissecting microscope, the dentate gyrus was dissected free from the hippocampus using a tungsten needle and fine scalpel. By holding a tissue slice in place with the needle, an incision was made perpendicular to the hippocampal fissure in a line at the lateral extreme of the upper and lower blades of the dentate granule cell layer to the level of the hippocampal fissure. Peeling the dentate along the natural separation line of the fissure carefully isolated the dentate gyrus and small amount of hilar tissue. Finally, the dentate apex was removed to avoid contamination with third ventricular subependymal cells. Eight to ten slices were recovered from two hippocampi and dentate tissue and the dentate-free hippocampal tissue (containing a small amount of loosely adherent subependymal material of the hippocampal arch and posterior lateral ventricle) were then cultured in parallel. Cell culturing was as described previously (Howell et al. 2003) with trituration in papain (22.0 U/mg, 2 mg/mL, Sigma, Poole, UK), partial cell purification by centrifugation through a two-step density gradient (OptiPrep; Axa-Shields, Oslo, Norway) and plating at 100 000 viable cells (Trypan Blue dye exclusion) per ml of control medium (Neurobasal A/B27 and l-glutamine, Life Technologies) on poly-l-lysine coated sterile glass slides, or onto poly-l-lysine coated 24- or 96-well plates (Co-star brand, Fisher Scientific, Pittsburgh, PA, USA).

Comparing cell division and neuronal counts in microdissected primary cell cultures

Two hours after final plating, the cultures were either replenished with fresh medium (controls) or fresh medium with 1 μm NPY (Peptide Synthesis Unit, University of Southampton). In initial experiments, BrdU (10 μm) (Sigma), a thymidine analogue incorporated into DNA during division (Cameron and McKay 2001), was added to the culture medium on day 4 for 24 h before paraformaldehyde (PFA) fixation. Cultures were then processed for BrdU and class III β-tubulin immunohistochemistry. To determine the cells on which NPY had a proliferative effect, we exposed 3-day-old control cultures to 10 μm BrdU with or without 1 μm NPY for 4 h, followed by immediate fixation. Cultures were immunostained and the proportions of neurons (class III β-tubulin) or astrocytes (GFAP) incorporating BrdU under NPY and control conditions calculated.

Fluorescence immunocytochemistry

Standard BrdU immunohistochemical staining was performed as previously reported (Howell et al. 2003). Cells were counter-stained with DAPI and then counted. Cultures grown on coverslips were similarly processed and finally mounted on glass slides in Moviol (Harco, Harlow, UK). Primary antibodies used were: rat anti-BrdU (1 : 200, Harlan Sera Laboratory, Loughborough, UK), mouse anti-class III β-tubulin (1 : 200, Babco, Berkeley, CA, USA), rabbit anti-GFAP (1 : 400, P. Steer, University of Southampton), guinea pig anti-doublecortin (1 : 1000, Chemicon, Temecula, CA, USA), sheep anti-Y1 (1 : 200, Biogenesis, Poole, UK) and Alexa 488 pre-conjugated mouse anti-Neu-N (1 : 200, Chemicon). Secondary antibody-only controls showed negative staining.

Examining the role of the Y1 receptor and the extracellular signal-regulated kinase (ERK) 1/2 cascade in neuroproliferation mediated by NPY

Cultures were grown in either control medium alone or with 1 μm NPY, BIBP3226 (Bachem, St Helens, UK), NPY and BIBP3226 or in the selective Y1 receptor agonist (F7, P34) NPY (Dr Beck-Sickinger, University of Leipzig) for 3 days. In a separate set of experiments, cells were plated at a higher density of 200 000 cells per ml and control and 1 μm NPY-treated cultures were grown for 3 days in the absence or presence of U0126 (1 μm, Sigma), an inhibitor of MEK activation of ERK1/2. BrdU was added to all cultures for the final 4 h prior to fixation. Cultures were then immunostained for BrdU and class III β-tubulin cell counting.

ERK1/2 activation

Dentate-derived cells were plated at a density of 106 cells per ml and cultures were grown for 3 days in control conditions before treatment with 1 μm NPY or 1 μm NPY and 1 μm U0126 for 4 min. After washing in cold Tris-buffered saline (TBS), cells were lysed in buffer (25 mm NaCl, 2 mm EDTA, 0.5 mm dithiothreitol, 20 mm HEPES, 20 mm glycerophosphate, 50 mm NaF, 1 mm NaVO3 and 0.1% Triton X-100) containing protease inhibitor mixture (Roche Diagnostics, Basel, Switzerland) per 10 mL, centrifuged and the supernatant stored as frozen aliquots. Similar protein content between samples was ensured using the Bio-Rad DC protein assay system and appropriate dilution in lysis buffer. Samples denatured by boiling in sodium dodecyl sulfate (SDS) buffer (0.125 m Tris-HCl, 4% SDS, 20% glycerol, 10% 2-mercaptoethanol and 200 μL of 10 mg/mL bromophenol blue) were loaded at 6 μg of protein per well (including a lane of molecular weight markers; Amersham Life Sciences, Bucks, UK) and separated electrophorectically on acrylamide gel. Protein was transferred to nitrocellose (BDH, Poole, UK) by semidry transfer and probed with rabbit anti-phopho p44/42 (1 : 1000, New England Biolabs, Hitchin, Herts, UK) in TBS, 0.1% Tween and 5% dried milk powder. Blots were incubated for 1 h at room temperature (23°C) with a secondary antibody and visualised with the Amersham ECL plus chemiluminescent system. Blots were then treated in a solution of 100 mm 2-mercaptoethanol, 2% SDS, 62.5 mm Tris-HCl pH 6.7 for 30 min at 50°C to remove all primary and secondary antibodies. Following extensive washing, blots were processed for total p44/42 signal (New England Biolabs).

Granule cell number, cell proliferation and neurogenesis in NPY inline image and inline image mice

Five 35-day-old male NPY inline image and five age- and generation-matched male inline image controls (Howell et al. 2003) were given one intraperitoneal (i.p.) dose of BrdU (50 mg/kg) each day for 5 days. Animals were killed by terminal anaesthesia, saline and PFA perfused and the brains post-fixed in 4% PFA. Coronal sections containing hippocampus were cut at 40 μm on a Leica VT100M Vibratome. Every 6th (BrdU; n = 8) and 12th (doublecortin DcX; n = 3) systematically randomly sampled section was stained. Immunopositive cells were counted in the subgranular zone using a Leitz Dialux22 microscope with a 63 × oil objective and the investigator blind to the experimental group.

For the proliferation rate and survival studies, ten 70-day-old male NPY inline image and nine age- and generation-matched male inline image controls (Howell et al. 2003) were given a single i.p. dose of BrdU (50 mg/kg). Four NPY inline image animals and four inline image controls were killed and perfused 2 h after the i.p. BrdU injection for proliferation rate studies. Six NPY inline image animals and five inline image controls were killed 21 days after BrdU injection and sections processed for BrdU and Neu-N immunohistochemistry. Five systematically randomly sampled sections (every 12th section) were stained from each animal and 200 BrdU-positive nuclei (Cy3 flourophor, Jackson Immuno-Research, West Grove, PA, USA) were examined for Neu-N immunopositivity (Alexa 488 fluorophor, Molecular Probes, Eugene, OR, USA) using confocal laser scanning microscopy. Sequential 0.4 μm optic sections were reconstructed and viewed in orthogonal planes and BrdU/Neu-N-positive nuclei in the granule cell layer were counted by an investigator blinded to the experimental group.

Total granule layer cell counts and granule cell layer volume were estimated using the optical fractionator (West et al. 1991) (StereoInvestigatorTM; MicroBrightfield). Eight systematically randomly sampled sections were DAPI stained, and mounted in DPX. The left dentate granule cell layer area was traced at low magnification, a randomly orientated sampling grid (150 × 150 μm) with a 10 × 10 μm counting frame was applied and permitted cell counts were made with a guard height of 1.2 μm under a 100 × oil-immersion objective. Total granule cell numbers in 90-day-old mice for the BrdU survival studies were generated using a 62 × 220 μm sampling grid using five systematically randomly sampled sections per dentate. Total granule layer cell counts were calculated from the mean count per frame using the Cavalieri principle. All data was collected with the investigator blinded to the animals' status.

Cell counting and statistical analysis

Fluorescence cell counting was performed on an Leica DM IRB microscope (Leica Microsystems UK Ltd, Milton Keynes, UK) linked to an Open Laboratory image-capturing system (Improvision Inc., Lexington, KY, USA) as described previously (Howell et al. 2003). Cell counting was performed on three to four random 20 × fields per well and expressed as cells/mm2 per well based on the averaged well count from four to 12 wells per condition per experiment from a minimum of three separate experiments. One experiment consisted of four hippocampi pooled from two animals. Data was plotted using Prism (GraphPad Software Inc., San Diego, CA, USA) and means compared using anova with Dunnetts, student Newman–Keuls, Bonferonni's or unpaired students t-test as appropriate. Confocal images were taken on a Zeiss LSM 510 meta system (Carl Zeiss Ltd. Oberkochen, Germany) by sequential scanning of each channel using appropriate filters and later manipulated using Leica software (LSM v5).

Results

NPY increases total cell counts, cell proliferation and β-tubulin expressing cells in dentate gyrus cultures

Five days' culture in 1 μm NPY significantly increased cell number and cells incorporating BrdU on day 5 in cultures derived from both dentate gyrus and hippocampus-only cultures (Table 1). The number of neurons (class III β-tubulin expressing cells) was greater in control dentate versus hippocampus-only cultures, as expected. However, dentate cultures showed a selective increase in the number of neurons in response to 1 μm NPY, whereas hippocampal cultures did not (Table 1). This was not as a result of a neurotrophic effect as both control and NPY-treated dentate cultures had similar proportions of cell death up to day 3 (Fig. 1). Four-hour exposure of BrdU and 1 μm NPY to day 3 cultures showed a proliferative effect of NPY on class III β-tubulin expressing cells by increasing the proportion of β-tubulin cells incorporating BrdU (Fig. 1). NPY had no effect on GFAP cell numbers up to 3 days and no proliferative effect on GFAP-positive cells at 3 days in culture (Table 2).

Table 1.  NPY increases the number of neurons in dentate cultures
 Dentate cultures
(cells/mm2)
 Hippocampal cultures
(cells/mm2)
 
  1. Cultures were established from dissected dentate tissue, including the hilus, and from remaining hippocampal tissue, and plated at 200 000 cells/mL. Cells were grown in control media (neurobasal A/2% B27/2 mm l-glutamine) or in the presence of 1 μm NPY for 5 days and included the addition of 10 μm BrdU for the final 24 h. Total cell counts as assessed by DAPI nuclear staining and cells that had incorporated BrdU were significantly increased after NPY in respect to control cell counts in both dentate and hippocampal cell cultures. After NPY exposure, the number of cells expressing the neuronal marker class III β-tubulin were significantly increased in dentate cultures, but not in hippocampal cultures. The number of β-tubulin-positive cells were also significantly different between dentate and hippocampal cell preparations (p < 0.0001) under control conditions. Data is mean ± SE of 8–12 wells per condition from a total of four experiments. Students t-test; **p < 0.01, ***p < 0.001.

Cell markerControl1 μm NPYControl1 μm NPY
Total cell count432 ± 44719 ± 42**426 ± 53696 ± 41**
BrdU140 ± 23277 ± 11***223 ± 22.3359 ± 34**
β-tubulin108 ± 10181 ± 14**27 ± 4.532 ± 4.1
Figure 1.

A short pulse of NPY stimulates neuroblast proliferation whilst having no effect upon cell survival. (a, b) Cells from dentate tissue grown in 1 μm NPY for 24, 48 and 72 h after plating. Viable cell counts (a) were significantly increased after 48- and 72-h NPY compared with the same time points under control conditions. Cell death was similar between control and NPY groups (b). Cell death was assessed by quantifying chromatin condensation (cells displaying a pyknotic nucleus) and expressed as a proportion of total cells counted per mm2. One-way anova with Bonferroni's post-test (**p < 0.01). (a, b) Values represent mean ± SE based on a sample of eight wells per condition from two separate experiments. (c) A representative field of DAPI stained cells at 72 h in culture. Solid arrowheads indicate nuclei with a condensed morphology. Open arrow indicates a normal healthy nucleus. Scale bar = 50 μm. (d) Cells grown for 3 days in vitro (DIV) in control conditions were exposed to BrdU and/or NPY for 4 h. The proportion of β-tubulin cells incorporating BrdU over a 4-h period is increased with a 4-h exposure to NPY implying a proliferative effect. Values are mean ± SE of 20–24 wells per condition from four separate experiments analysed using student's t-test (**p < 0.01).

Table 2.  NPY does not affect the number of GFAP-positive astrocytes in culture
Cell markerControl
(cells/mm2)
1 μm
NPY (cells/mm2)
  1. Cells from dissected dentate tissue were cultured for 3 days in control or 1 μm NPY conditions to assess effects upon glial fibrillary acidic protein (GFAP) cell numbers. Cultures were pulsed with BrdU for 4 h prior to fixation. The numbers of BrdU-incorporating cells were significantly increased in NPY conditions, whilst numbers of GFAP immuno-positive cells or those proportion positive for both GFAP and BrdU were unchanged following 3 DIV in NPY. Data expressed as mean ± SE of 18–24 wells from four separate experiments. Analysed using students t-test, **p < 0.01.

BrdU3.6 ± 0.365.1 ± 0.38**
GFAP16.3 ± 2.1516.0 ± 1.98
GFAP/BrdU0.8 ± 0.191.0 ± 0.19

The neuroproliferative effect of NPY is Y1-receptor mediated and requires ERK1/2 activation

NPY (1 μm) significantly increased the number of BrdU-incorporating cells, β-tubulin expressing cells and neuroblasts (β-tubulin-positive cells incorporating BrdU) after 3 days in culture (Fig. 2). This neuroproliferative effect was abolished by the Y1 antagonist BIBP3226 and was mimicked by the selective agonist (F7, P34) NPY (Fig. 2). BIBP3226 alone had no effect. The ERK kinases are a group of mitogen-activated protein kinases (MAPK), important in control of cellular proliferation (Lopez-Ilasaca 1998), including neural progenitors (Li et al. 2001). Inhibition of ERK1/2 phosphorylation by UO126 abolished the neuroproliferative effect of NPY (Fig. 2). Confocal imaging confirmed co-localisation of Y1 antibody staining with BrdU and β-tubulin-positive cells (Figs 2f and g). Counting of 100 BrdU cells from two experiments showed 95 ± 1.1% of BrdU positive and 79 ± 3% of β-tubulin positive cells co-localised for Y1 receptor staining.

Figure 2.

NPY mediates neuroproliferation via the NPY Y1 receptor and requires ERK1/2 activation. (a) NPY significantly increases the number of cells immunopositive for BrdU following a 4-h pulse of BrdU. This effect was blocked by the specific Y1 receptor antagonist BIBP3226 and mimicked by the specific Y1 receptor agonist (F7, P34) NPY while BIBP3226 alone had no effect. This pattern was repeated for β-tubulin-positive cells (b) and neuroblasts (BrdU + β-tubulin positive cells) (c). Experiments were performed on 12–16 wells per condition from four separate experiments. anova with Dunnett's post-test (comparing control wells to all other conditions), **p < 0.01. (d) To assess if the MAP kinases, ERK1/2 activation is required for NPY-mediated neuroproliferation, cells were grown in the absence or presence of U0126, an inhibitor of MEK1 and MEK2 (MAP or ERK kinases). BrdU was added for the final 4 h. U0126 (1 μm) totally abolished any increase in the number of β-tubulin/BrdU-positive cells in comparison with NPY (anova with Dunnett's post-test, **p < 0.01). Values represent mean ± SE from three separate experiments. (e) NPY phosphorylation of ERK1/2 was confirmed by probing immunoblots of cell lysates from 3 DIV cultures exposed to NPY or NPY plus U0126 for 4 min. Levels of activated ERK1/2 were increased upon the addition of NPY, whilst NPY plus U0126 had low levels of activated protein. Total ERK levels in cell lysates were comparable. (f, g) Confocal images showing BrdU (red) and Y1 antibody staining (green) (f) and β-tubulin (red), DAPI (blue) and Y1 antibody (green) co-localisation (g) in cells cultured from the dentate gyrus and exposed to BrdU for 4 h prior to fixation. Scale bar = 20 μm.

inline image mice have reduced subgranular zone BrdU incorporation and doublecortin-positive cells

Stereological estimation of subgranular zone cells labelled with BrdU over five consecutive days in 35-day-old inline image mice revealed an approximate 40% decrease in the number of BrdU positive cells in comparison with age- and generation-matched inline image controls (6685 ± 729 vs. 4176 ± 217 cells per subgranular zone for control and inline image groups, respectively; one-way anova with student Newman–Keuls multiple comparison post-test; p < 0.01) (Figs 3a and b). There was no difference in BrdU cell counts from the hilus or areas CA1-3, demonstrating no difference in BrdU bioavailability between the inline image mice and controls. The reduction in SGZ proliferation appeared to translate into a reduction in neurogenesis as staining for the immature neuronal marker doublecortin showed a reduction of similar magnitude (Figs 3c and d). Interestingly, estimates of total number of cells in the granule cell layer of inline image mice (375 800 ± 50 050 cells) did not differ from inline image controls (363 200 ± 33 270 cells) (Table 3).

Figure 3.

Y1 homozygous knock-out animals have reduced subgranular zone cell proliferation and lower numbers of immature doublecortin-positive neurons. (a) Five inline image controls and five inline image 35-day-old mice were administered BrdU (50 mg/kg) intraperitoneally once per day for 5 days and killed 5 h after the final injection. Immunohistochemistry illustrates cells that incorporated BrdU in the subgranular zone. (b) Eight systematically random sampled sections per animal from the rostro-caudal extent of the hippocampus were scored for BrdU-positive cell nuclei at high magnification and an unbiased stereological estimate of the total number of BrdU-positive cells per dentate generated using the Cavalieri method. inline image animals had a significant reduction in the number of BrdU-positive cells counted in the SGZ (anova with student Newman–Keuls post-test; **p < 0.01). There was no significant difference in BrdU cell counts from the hilus or areas CA1-CA3 of the hippocampus. (c) Doublecortin immuno-reactivity (brown DAB stain) showing new neurons in the subgranular zone. Granule cell layer counterstained with thionine (blue). (d) The number of doublecortin-positive cells is significantly reduced in the dentate of inline image animals in comparison with inline image controls (n = 5 animals per group). Data was compiled from three systematically randomly sampled sections from the full extent of the hippocampus from the same animals as section (b). Students t-test, **p < 0.01. (e) The number of BrdU-positive cells in the SGZ was significantly reduced in inline image mice compared with Y1lox/lox controls 2 h after a single i.p. injection of BrdU (50 mg/kg) (p < 0.001: anova with Bonferronni's post-test, n = 4 animals per group and five sections sampled from each dentate). However, there was no significant difference in BrdU counts 21 days after BrdU injection (n = 6 animals per group and five sections sampled from each dentate). (f) Representative projection image of a 3-D confocal Z stack through the granule cell layer and subgranular zone (SGZ) of a inline image knock-out animal showing Neu-N positive nuclei (green) and a single BrdU-positive nucleus (red) co-localising for Neu-N (white arrow) in the SGZ. Scale bar = 10 μm.

Table 3.  The numbers of dentate granule layer cells are similar between wild type Y1lox/lox controls and homozygous Y1−/− animals
 inline imageinline image
  1. The total number of granule cells of the left hippocampus of 40-day-old male inline image and inline image mice were determined using the optical fractionator technique. Eight 40-μm sections corresponding to every 6th serial section were quantified along the entire length of the hippocampus in 35-day-old animals and five 40-μm sections corresponding to every 12th serial section were quantified along the entire length of the hippocampus in 90-day-old animals. Absence of the NPY Y1 receptor resulted in no difference in the total number of cells at either age. Data is mean ± SE, n = 6 40-day-old animals and 5 90-day-old animals.

35-day-old mice
Granule layer cell number363 200 ± 33 270375 800 ± 50 050
90-day-old mice
Granule layer cell number342 600 ± 13 700346 200 ± 11 690

inline imagemice have a reduced subgranular zone proliferation rate but an increased survival of newly generated neurons

Adult 90-day-old inline image mice showed a lower cell proliferation rate in the subgranular zone (1086 ± 96.06 cells/SGZ) compared with age- and generation-matched inline image controls (1590 ± 141.0 cells/SGZ; p < 0.001: anova with Bonferroni's post-hoc test) as judged by the number of cells incorporating BrdU 2 h after a single i.p. injection (Fig. 3e). However, the number of BrdU-positive cells in inline image mice killed 21 days after a single BrdU injection was not significantly different from that in the inline image controls (Fig. 3e), indicating a relatively greater survival of newly generated cells in the dentate gyrus of inline image animals. There was no significant difference between the proportion of BrdU-positive cells that co-localised for the mature neuronal marker Neu-N at 21 days after a single BrdU injection between inline image mice (98%) and inline image controls (96%). The Neu-N antibody was pre-conjugated to the Alexa 488 fluorophor and no non-specific antibody staining was seen. This lack of difference in new neuron incorporation by 21 days was confirmed by the absence of any significant difference in total granule cell layer cell counts between 90-day-old constitutive inline image mice (346 200 ±11 690 n =  6) and inline image controls (342 600 ± 13 700 n =5) (Table 3).

Discussion

Here, we show that NPY is a proliferative factor for neuroblast precursor cells in the early post-natal dentate gyrus in vitro and the adult dentate gyrus in vivo. We further show that the neuroproliferative effect is Y1-receptor mediated and requires ERK1/2 activation (Fig. 2).

Our in vitro studies demonstrated a direct proliferative effect of NPY on neuroblasts cultured from the early post-natal dentate gyrus (Fig. 1d), but no effect on neuronal number cultured from the rest of the hippocampus (Table 1) where neurogenesis is also ongoing, albeit at a much lower rate (Rietze et al. 2000). This was not a culture artefact, as the plating densities of hippocampus-derived and dentate-derived cultures were equal and there was robust cell proliferation of β-tubulin negative cells in the rest of hippocampus cultures (Howell et al. 2003). In agreement with our previous study on whole hippocampus-derived cultures (Howell et al. 2003), the proliferative effect of NPY on dentate β-tubulin-positive neuroblasts is Y1-receptor mediated and requires ERK1/2 activation (Fig. 2), consistent with a previous report of a neuroproliferative effect of NPY in the olfactory epithelium (Hansel et al. 2001). The ERK are a group of MAPK, important in control of cellular proliferation (Lopez-Ilasaca 1998), including neural progenitors (Li et al. 2001). ERK activation-dependant pathways are involved in neurogenesis in the embryonic (Harada et al. 2004) and adult hippocampus (Rueda et al. 2002; Persson et al. 2003). It is worth noting that blocking activation of either the Y1 receptor or the MEK/ERK pathway selectively abolished the proliferative effect of NPY in vitro but had no effect on neuroproliferation under control conditions, implying absence of an endogenous NPY effect in vitro and the presence of alternate pathways driving proliferation under control conditions. The latter was confirmed in vivo, with significant residual SGZ cell proliferation in mice lacking the Y1 receptor (Fig. 3e).

We confirmed a neuroproliferative effect of NPY on SGZ precursor cells in vivo, showing a significant reduction in SGZ proliferation rate and doublecortin-expressing cells in inline image mice (Fig. 3). The production of new neurons in the SGZ exceeds the number that survive (Dayer et al. 2003) and doublecortin immunoreactivity reliably estimates immature neurons in the first 12 days after neuronal birth, while Neu-N staining identifies a more mature subset that survives and integrates into hippocampal circuitary (Rao and Shetty 2004). The paradoxical finding of a lower rate of neurogenesis, yet similar total granule layer cell counts in inline image mice compared with controls at both 1 and 3 months of age, is explained by the compensatory increase in survival of newly generated neurons in constitutive inline image animals, such that similar numbers of newly born neurons result by 21 days (Fig. 3c). This is not to say that NPY has no effect on net neurogenesis in vivo, but rather that compensatory mechanisms may have developed in a constitutive Y1 receptor knockout mouse model to compensate for loss of the Y1-mediated proliferative effect. Indeed, reducing precursor proliferation with methylazoxymethanol (MAM) enhances the survival of newly born granule cells (Ciaroni et al. 2002b), suggesting the presence of such compensatory mechanisms under normal conditions (Ciaroni et al. 2002a). This would also explain why we found no evidence for a trophic effect on survival of NPY in vitro yet loss of the Y1 receptor results in increased survival of newly born granule cell neurons in vivo. Given the reduction in proliferation and neurogenesis seen in naive inline image animals (Fig. 3), it is clear that NPY is important for the maintenance of baseline neurogenesis. It is also clear from the residual level of precursor proliferation, as well as the compensatory changes in constitutive inline image mice, that many other factors dynamically influence dentate neurogenesis in vivo.

Because exercise is neuroproliferative (van Praag et al. 1999), knock-out of NPY Y1 receptors causing reduced locomotor activity (Pedrazzini et al. 1998) could explain the reduced dentate neurogenesis. However, this mechanism is unlikely, as we have demonstrated a direct neuroproliferative effect on dentate neuroblasts in vitro (Fig. 1). The proliferative effect of NPY might also be explained by an NPY-induced reduction in synaptic glutamate release, which is known to increase SGZ proliferation (Cameron et al. 1995), however, this is a Y2 receptor-mediated effect (Greber et al. 1994; McQuiston and Colmers 1996) and we have previously shown a lack of any Y2 effect in vitro (Howell et al. 2003).

Based on our in vitro and in vivo data, we conclude that NPY is neuroproliferative for post-natal and adult dentate precursors. This shared precursor cell responsiveness across development and adulthood is consistent with ongoing neurogenesis in the dentate throughout life (Pleasure et al. 2000).

NPY control of dentate neurogenesis

The most likely origin of NPY for modulating dentate neurogenesis is from NPY releasing interneurons in the dentate hilus as granule cell production of NPY is only induced after seizures (Gruber et al. 1994) or perforant path stimulation (Causing et al. 1996). These NPY interneurons modulate granule cell excitability mainly via feedback circuits from mossy fibre collaterals [HIPP cells (Freund and Buzsaki 1996)] but also via feed forward circuits from entorhinal cortex and contralateral hippocampal inputs (basket cell interneurons) (Freund and Buzsaki 1996). The neurogenic SGZ and the molecular layer of the dentate gyrus are rich in Y1 receptors (Kopp et al. 2002), the functional significance of which has heretofore been unclear. NPY interneurons are thus ideally positioned to sample patterns of afferent and efferent activity and thus modulate neurogenesis via NPY release on Y1 receptor positive subgranular zone precursors. Volume transmission by released NPY acting via rapidly internalising Y1 receptors (Fabry et al. 2000) is a well-recognised mechanism of action of NPY and is particularly attractive as a means of orchestrating the proliferation of populations of precursors in the dentate SGZ. The other possible origin of NPY is from the sympathetic nervous system releasing NPY into the circulation (Lundberg et al. 1985; Pernow et al. 1986) and crossing the blood–brain barrier (Kastin and Akerstrom 1999) in the neurogenic SGZ. Indeed, exercise-induced systemic NPY release (Lundberg et al. 1985) may also play a role in mediating the exercise-induced increase in dentate neurogenesis (van Praag et al. 1999).

NPY and behaviour, learning and memory

Acute stress reduces dentate neurogenesis, an effect mediated via adrenal steroids (Gould et al. 1997). Stress is often a requisite for depression and decreased hippocampal neurogenesis may be important in the pathogenesis of depression (Jacobs et al. 2000; D'Sa and Duman 2002). Indeed, antidepressant-induced increases in dentate neurogenesis (Malberg et al. 2000) may be required for the behavioural effects of antidepressants (Santarelli et al. 2003). Both laboratory (Heilig et al. 1989) and clinical studies (Morgan et al. 2000, 2002) have implicated NPY in the regulation of anxiety-related behaviours and the central anxiolytic effect of NPY appears to be Y1-receptor mediated (Wahlestedt et al. 1993; Sajdyk et al. 1999). Clinical studies have reported reduced NPY in the cerebrospinal fluid and plasma from depressed patients (Westrin et al. 1999). Animal models of depression (Caberlotto et al. 1999; Husum and Mathe 2002) have reduced hippocampal levels of NPY and altered Y1 binding levels and the antidepressant-like activity of NPY appears to be mediated via the Y1 receptor (Redrobe et al. 2002). NPY has been shown to be involved in the antidepressant actions of lithium, electroconvulsive therapy (ECT) and citalopram (Husum et al. 2000), all of which increase neurogenesis (Chen et al. 2000; Madsen et al. 2000; Malberg et al. 2000), and the NPY gene is one of the few genes whose expression is increased only with chronic ECT (Altar et al. 2004).

Dentate neurogenesis also appears to be necessary for some forms of hippocampal-dependant learning (Shors 2004). NPY appears to have a role in learning and memory, albeit less established than in anxiety. Intracerebroventricular injections of NPY fragments have been shown to enhance memory retention, recall and prevent scopolamine-induced amnesia (Flood et al. 1987), especially when injected into the rostral hippocampus (Flood et al. 1989). NPY attenuates learning impairments induced by MK801 (Bouchard et al. 1997) and the Y2 receptor also appears to be important in learning and memory processing (Redrobe et al. 2004). Studies in hippocampal NPY over-expressing rats have shown impaired spatial learning associated with decreased NPY–Y1 binding in young animals (Thorsell et al. 2000), but normal spatial learning in old animals (Carvajal et al. 2004).

Although Y1 receptors are found throughout the hippocampus and are not just confined to the neurogenic dentate subgranular zone (Kopp et al. 2002), it is intriguing to speculate that some of the cognitive and behavioural effects of NPY on learning and memory may be mediated via the effect of NPY on dentate neurogenesis, although clearly this hypothesis is currently speculative and will need to be tested by experiment.

Conclusion

Here, we show that NPY is a specific neuroproliferative factor for adult dentate gyrus neurogenesis acting via the ERK signalling pathway and has a role in the maintenance of baseline proliferation in vivo, raising the attractive hypothesis that dentate neurogenesis is activity modulated by NPY-releasing interneurons.

Acknowledgements

The authors are grateful to Dr M. Cuttle for assistance with confocal microscopy. This work was supported by grants from the Wessex Medical Trust Hope, the University of Southampton and the MRC to WPG and the Human Frontiers Science Program to HES, HH and AB-S.

Ancillary