Address correspondence and reprint requests to Dr W. P. Gray, Division of Clinical Neurosciences, University of Southampton, Room 6007, Level 6, Biomedical Sciences Building, Bassett Crescent East, Southampton SO16 7PX, UK. E-mail: email@example.com
New neurones are produced in the adult hippocampus throughout life and are necessary for certain types of hippocampal learning. Little, however, is known about the control of hippocampal neurogenesis. We used primary hippocampal cultures from early post-natal rats and neuropeptide Y Y1 receptor knockout mice as well as selective neuropeptide Y receptor antagonists and agonists to demonstrate that neuropeptide Y is proliferative for nestin-positive, sphere-forming hippocampal precursor cells and β-tubulin-positive neuroblasts and that the neuroproliferative effect of neuropeptide Y is mediated via its Y1 receptor. Immunohistochemistry confirmed Y1 receptor staining on both nestin-positive cells and β-tubulin-positive cells in culture and short pulse 5-bromo-2-deoxyuridine studies demonstrated that neuropeptide Y has a proliferative effect on both cell types. These studies suggest that the proliferation of hippocampal neuroblasts and precursor cells is increased by neuropeptide Y and, therefore, that hippocampal learning and memory may be modulated by neuropeptide Y-releasing interneurones.
Neuropeptide Y (NPY) is widely expressed in neural tissue of the peripheral and central nervous systems. Within the brain, this highly conserved 36 amino acid polypeptide is involved in aspects of feeding behaviour, circadian rhythm, memory processing, affective disorders and has a critical role in seizure control where it may act as an endogenous anticonvulsant (Baraban et al. 1997; Vezzani et al. 1999). Of particular interest is the demonstration of effects on cognitive function in association with learning and memory after intracerebroventricular injections of NPY fragments (Dumont et al. 1992) and impaired spatial learning in transgenic rats with hippocampal NPY overexpression (Thorsell et al. 2000). In the mammalian hippocampus, NPY is localized to GABAergic interneurones of the dentate hilus and hippocampus proper and its activity is predominantly mediated by G-protein-coupled Y1, Y2 or Y5 receptors (Michel et al. 1998).
A recent addition to the physiological effects of NPY has been the demonstration of NPY-induced proliferation of olfactory neuroblasts within the olfactory epithelium mediated via the Y1 receptor (Hansel et al. 2001). This work highlighted the possibility that NPY, the predominant neuropeptide of the hippocampus, may have a similar neuroproliferative action on hippocampal precursor cells, especially those in the dentate gyrus. To address this issue we have used cell cultures prepared from young rat hippocampi and examined the proliferative potential of NPY and the phenotypic fate of this cell proliferation.
Materials and methods
Generation of progenitor cell cultures from the early post-natal rat
All animal work was performed under UK Home Office licence and was approved by the University Biomedical Ethics Committee. Briefly, either Wistar rat pups or NPY Y1 receptor knockout mice (8–10 days old) were killed by decapitation and the hippocampus rapidly dissected out in Gey's balanced salt solution (Life Technologies, Rockville, MD, USA) at 4°C. The corpus callosum was bisected above the thalamus and the posterior margin of each cortical hemisphere rolled back. Each hippocampus was separated from the overlaying cortical white matter along the natural separating line of the alveus hippocampus. Care was taken in removing as much white matter as possible, including the subiculum, although some white matter remained. Contamination by the loosely adherent subependymal layers of the hippocampal arch and posterior lateral ventricle could not be ruled out. Transverse sections were cut at 400 µm on a McIlwain tissue chopper and placed in Neurobasal A® supplemented with 2% B27 (Life Technologies) (NB/B27) and 0.5 mm glutamine (Sigma, St Louis, MO, USA) for 5 min. Slices were transferred to pre-warmed papain (22.0 U/mg, 2 mg/mL; Sigma) and incubated for 30 min at 30°C. The papain was then aspirated, replaced with 2 mL of NB/B27 and glutamine and the tissue triturated to separate clumps. Partial purification of cells from debris was achieved by centrifugation through a two-step density gradient (OptiPrep; Axa-Shields, Oslo, Norway) for 15 min at 400 g. The viable cell fraction was collected, washed free of OptiPrep and resuspended in NB/B27 and glutamine to yield 100 000 viable cells (trypan blue dye-excluding cells) per mL. Aliquots were plated on poly-l-lysine (Sigma) coated sterile glass slides for confocal imaging (13 or 19 mm diameter; Chance Proper brand; Fisher Scientific, Loughborough, Leics., UK), or directly onto poly-l-lysine-coated 24- or 96-well plates (Co-star brand; Fisher Scientific) for cell counting; 2 h after plating the cells were rinsed and replenished with fresh media. All media included a combined antibiotic/antimycotic (Pen/Strep and Fungizone; Life Technologies).
Bromodeoxyuridine incorporation and cell fixation
Cell proliferation was measured by the incorporation of the thymidine analogue 5-bromo-2-deoxyuridine (BrdU) (Sigma) which is incorporated into the DNA of dividing cells in immunohistochemically detectable quantities during the S phase of cell division (Cameron and McKay 2001). The BrdU was added at a final concentration of 10 µm to the growth media at various time points before cell fixation. Cells were rinsed once in phosphate-buffered saline before being fixed in 4% paraformaldehyde (Sigma) for 20 min.
The paraformaldehyde-fixed cell cultures were washed three times in Tris-buffered saline (TBS) (0.15 m NaCl, 0.05 m Tris, pH 7.4). For immunohistochemical analysis of BrdU incorporation cultures were incubated in 2 m HCl for 30 min at 37°C. Cells were then washed in TBS. Non-specific antibody-binding sites were blocked with TBS containing 0.1% Triton and 5% pre-immune donkey serum (Sigma) at room temperature for 30 min and cells were then incubated in TBS-0.1% Triton with appropriate dilutions of primary antibodies overnight at 4°C. Cells were washed in TBS before the addition of secondary antisera conjugated to cyanin (Cy) 2, Cy3, Cy5 or FITC (Jackson Immuno-Research, West Grove, PA, USA) at 1 : 200 in TBS-0.1% Triton for 2 h at room temperature. Samples were subsequently counterstained with 4′,6-diamidino-2-phenylindole (DAPI; 10 µg/mL) (Sigma) for 15 min. Primary antibodies raised in mouse and rat were used at the following concentrations: rat anti-BrdU (1 : 200, clone BU1/75; Harlan Sera Laboratory, Loughborough, UK), mouse anti-class III β-tubulin (1 : 200, TUJ1 clone; Babco, Berkley, CA, USA), mouse anti-nestin (1 : 200; BD Pharmingen, San Diego, CA, USA) and mouse anti-glial fibrillary acidic protein (GFAP) (1 : 400; Sigma).
Immunofluorescence for confocal imaging
The paraformaldehyde-fixed cell cultures grown on coverslips were washed three times in TBS before the addition of a pre-immune block of 0.1% bovine serum albumin (Sigma) in TBS for 30 min at room temperature. Samples were then incubated in TBS-0.1% Triton with dilutions of primary antibodies for 2 h at room temperature. The primary antibodies were: mouse anti-class III β-tubulin (1 : 1000, clone Tuj-1; Babco), mouse anti-MAP2 (1 : 1000; Chemicon, Temecula, CA, USA), mouse anti-nestin (1 : 1000; BD Pharmingen), sheep anti-NPY-Y1 receptor (1 : 200, raised against C-terminus residues 365–378; Biogenesis, Poole, UK) and mouse anti-oligodendrocyte (1 : 200 000, clone RIP; Chemicon). Cells were then washed in TBS before the addition of secondary antisera conjugated to Cy2 or Cy3 (Jackson Immuno Research) at 1 : 200 dilution in TBS-0.1% Triton for 1 h at room temperature. Coverslips were washed in TBS, rinsed briefly in distilled water to remove salts and mounted ‘face down’ in Moviol aqueous mount (Harco, Harlow, UK).
Secondary antibody only control coverslip cultures were processed simultaneously using an identical protocol except that dilution solutions were devoid of primary antibodies. All secondary controls were negative for staining.
Cell counting and statistical analysis
Fluorescence cell counting was performed on an inverted DM IRB microscope (Leica Microsystems UK Ltd, Milton Keynes, UK). The area of a 20× field was measured using a 25-µm grid graticule slide (Microbrightfield, Williston, VT, USA). Cell counting was performed on four to eight random 20× fields per well using the Open Lab image-capturing system version 2.1 (Improvision, Lexington, MA, USA). Raw data from the 20× field counts were averaged and plotted ± SEM and expressed as cells/mm2 per well or as percentage of control, based on a sample of four to 10 wells per condition per experiment from a minimum of three separate experiments. One experiment consisted of four hippocampi from two animals, pooled and prepared as described above. Data points were plotted using prism data analysis software (GraphPad Software Inc., San Diego, CA, USA) and means compared using anova with Dunnett's multiple comparison post-hoc test, Bonferonni's post-hoc test or the unpaired Student's t-test as appropriate.
Confocal images were taken on an LSM 510 meta system (Carl Zeiss, Oberkochen, Germany) and images were captured by sequential scanning of each channel using appropriate filters and later merged using LSM v. 5 software (Leica Microsystems UK Ltd) or Adobe PhotoShop 5.01 (Adobe Systems, San Jose, CA, USA).
Culture characterization and time course experiments
Two hours after final plating the cultures were either replenished with fresh medium (controls) or fresh medium with NPY (Peptide Synthesis Unit, University of Southampton, UK) at a final concentration of 1 µm. The culture medium was changed every 3 days aspirating two-thirds of the medium volume and replacing it with either control medium or medium containing NPY at 1 µm concentration. The BrdU was added to all culture media at a concentration of 10 µm 24 h before fixation, to assess entry into the S phase of cell division. Control and NPY-exposed cultures were fixed at 1, 3, 5 and 7 days, processed for immunohistochemistry against BrdU, nestin, GFAP and class III β-tubulin and counterstained with DAPI. Cell counts, as detailed above, were determined in relation to cells cultured under control conditions (NB/B27 and glutamine only). Cell death was quantified by counting the number of pyknotic/condensed nuclei on DAPI staining within the same high power fields as the BrdU and phenotypic marker counts under control and NPY conditions.
In a separate set of experiments cells were cultured for 5 days under control conditions and similarly processed and analysed to accurately characterize the proportions of different cell phenotypes present.
Comparing the response of cultures to neuropeptide Y and basic fibroblastic growth factor
Cultures were similarly generated and allowed to grow under control conditions and in the presence of either 1 µm NPY or 20 ng/mL of basic fibroblastic growth factor (bFGF) (Sigma) for 5 days. 5-Bromo-2-deoxyuridine (10 µm) was added for the final 24 h and cells were immunostained for BrdU and nestin and counterstained with DAPI.
Short pulse 5-bromo-2-deoxyuridine and neuropeptide Y experiments
To determine the cell types upon which NPY has a proliferative effect we grew cultures for 3 days under control conditions (NB/B27 and glutamine) and then exposed one group to 10 µm BrdU for 4 h and another group to both 10 µm BrdU and 1 µm NPY for 4 h after which all cultures were immediately fixed. The cell cycle time of dividing hippocampal precursor cells is 16 h (Lewis 1978; Nowakowski et al. 1989) and, therefore, nuclear BrdU staining will accurately reflect cellular entry into the S phase of cell division over the 4-h exposure period. Cultures were immunostained for either BrdU and nestin or BrdU and class III β-tubulin and the proportions of each cell phenotype incorporating BrdU under NPY and control conditions were calculated from raw data counts.
Determination of the receptor mediation of the neuropeptide Y proliferative effect
Monolayer cell cultures were grown either under control conditions, with 1 µm NPY or with specific Y1, Y2 or Y5 receptor agonists (1 µm) for 3 days after which they were fixed. 5-Bromo-2-deoxyuridine was again added to all cultures 24 h prior to fixation. Cultures were immunostained for BrdU, class III β-tubulin and counterstained with DAPI for total cell, BrdU- and β-tubulin-positive cell counting. In a separate set of experiments cells were grown for 3 days either under control conditions or with NPY (1 µm) alone, the specific Y1 antagonist BIBP326 (1 µm) alone or both NPY and BIBP3226 to examine the effect of blocking the Y1 receptor on the NPY effect. Finally, to confirm a Y1 effect, cultures were generated from eight P8 Y1–/– mice and eight Y1lox/lox controls and grown under control conditions or in 1 µm NPY for 5 days, with 10 µm BrdU added for the final 24 h prior to fixation. Cultures were processed as above and total cell counts and BrdU-positive cell counts generated.
Neuropeptide Y Y1 receptor expression on 5-bromo-2-deoxyuridine-positive cells
To determine the proportion of proliferating cells expressing the NPY Y1 receptor, hippocampal cultures were grown on sterile glass coverslips for 3 days (as described above) in the presence of 1 µm NPY and pulsed with BrdU for 4 h prior to cell fixation. Wells were then processed for BrdU and NPY Y1 immunoreactivity. Cells were counted by positioning the field of view (630× magnification) at a random point near the top of the coverslip and systematically moving the stage stepwise transversely and downwards in non-overlapping fields until 100 BrdU-positive cells were counted and each cell was scored for NPY Y1 immunoreactivity. One hundred cells per well, from four separate experiments, were quantified and presented as mean percent ± SE.
Neuropeptide Y receptor-subtype specific agonists and antagonist
We used the specific NPY receptor-specific agonists [F7, P34] NPY against the Y1 receptor (Soll et al. 2001), Ahx[5–24]NPY against the Y2 receptor (Beck-Sickinger et al. 1992) and [cPP1–7, NPY19–23, A31, Aib32, Q34]hPP against the Y5 receptor (Cabrele et al. 2000) (Dr Annette Beck-Sickinger, University of Leipzig, Germany) all at a final concentration of 1 µm. The Y1 antagonist BIBP3226 (Bachem Ltd, St. Helens, UK) was also used at 1 µm. All compounds were present in the media for the duration of the experiment.
Neuropeptide Y Y1 knockout mice
A targeting vector for the Y1 receptor gene has been designed that allows the production of conditional (floxed, Y1lox/lox) and germline (Y1–/–) knockout mice in which the entire coding region of the Y1 receptor is removed. Briefly, a 129/SvJ mouse genomic BAC library (GenomeSystems, St Louis, MO, USA) was screened under low stringency for the Y1 receptor gene using a human Y1 cDNA probe. A 4-kb SacI and a 6- and 1.5-kb PstI fragment were subcloned into pBluescript and used to generate the targeting construct. A loxP-flanked Neo cassette was inserted into the Hind III site upstream of the Y1 coding sequence and a third loxP sequence was introduced by cloning two complementary 46 mer oligonucleotides into an XbaI site 1 kb downstream of the Y1 gene. A 0.8-kb EcoRI/PstI fragment 5′ to the Y1 targeting construct was used to screen for positively targeted ES cell clones. Two of these clones were injected into C57BL/6 blastocysts. Chimeric offspring were crossed with oocyte-specific Cre-recombinase-expressing C57BL/6 mice (Schwenk et al. 1995) in order to obtain either heterozygotes carrying the floxed gene (Y1lox/+ mice) or heterozygotes carrying the Cre-recombinase gene with the floxed gene already deleted (Y1+/– mice). Homozygous lines, both Y1–/– and Y1lox/lox, were generated by crossing the respective heterozygous animals. All further mice were maintained on this mixed C57BL/6–129SvJ background and, in the case of the Y1–/– line, mice were selected that no longer contained the Cre-transgene.
To confirm that introduction of the loxP sites and Neo cassette did not alter expression of the Y1 receptor gene in our conditional line, we performed radioligand binding experiments on brain slices from Y1+/+, Y1lox/lox and germline Y1–/– animals as described previously (Sainsbury et al. 2002). Germline Y1–/– mice had negligible binding to the Y1-specific antagonist 125I-1229U91, with the levels of binding in non-induced conditional Y1lox/lox mice comparable to those of Y1+/+ mice (Fig. 5c). Southern blot analysis of genomic DNA confirmed the absence of Y1 receptor DNA (Fig. 5b). Therefore, non-induced Y1lox/lox mice derived from the same founder line on the same mixed C57BL/6–129SvJ background as germline Y1–/– were used as controls in all experiments and are termed this way throughout this work.
Generating sphere-forming cultures
Hippocampal cells were isolated and prepared as described. Cells prepared from two animals were plated onto untreated, sterile, 12-well tissue culture plates at 100 000 viable cells/mL and incubated in NPY-containing media or control media for 8 days. Quantification of sphere numbers and size was performed on four 5× fields within six control and six NPY culture wells. Only neurospheres with a diameter greater than 50 µm were scored.
To determine the cell lineages that can be generated from spheres cultured in NPY, we transferred spheres to NPY-deficient, differentiating conditions (Vescovi et al. 1993; Gage et al. 1998). Spheres (8 days old), grown in 1 µm NPY, were pelleted by gentle centrifugation at 130 g for 1 min, the NPY-containing supernatant fluid was aspirated and the pellet was resuspended in control media alone (NB/B27 and glutamine). This was repeated three times before replating on sterile, poly-l-lysine-coated glass slides in a 24-well tissue culture plate. Spheres were allowed to attach and differentiate for 7 days in vitro (DIV) before fixation and processing for fluorescent immunochemistry. Neurospheres from two separate experiments (two animals per experiment) were processed and imaged.
Confocal imaging of intact spheres was performed on 8-day-old cultures. Spheres were exposed to BrdU for 24 h between days 7 and 8. Sphere cultures were carefully aspirated and transferred to poly-l-lysine-coated 13-mm diameter glass slides within a 24-well plate to allow the cells to adhere before paraformaldehyde fixing and subsequent processing for BrdU and nestin immunoreactivity and confocal imaging.
Characterization of dissociated cultures from rat pups at post-natal days 8–10
Hippocampal cultures contained cells immunopositive for the stem/precursor cell marker nestin, the astrocytic marker GFAP and the neuronal specific marker class III β-tubulin. After 5 days in culture the proportions of each cell type and the proportion of cells incorporating the proliferation marker BrdU on day 5 are given in Table 1.
Table 1. Cell phenotype after 5 days in culture under control conditions
The proportion of cells incorporating 5-bromo-2-deoxyuridine (BrdU) or presenting a specific phenotype following 5 days growth under control (Neurobasal A® supplemented with 2% B27) conditions are shown. Cells incorporating BrdU were assessed over the final 24 h of the experiment.
n, No. of wells sampled from at least four separate experiments; GFAP, glial fibrillary acidic protein.
Proliferative effects of neuropeptide Y on hippocampal cell cultures
We followed the progression of cell number, rates of division and phenotypic fate [class III β-tubulin (neuronal), GFAP (glial) and nestin (stem/precursor cell)] over 7 days in culture. 5-Bromo-2-deoxyuridine was given for the final 24 h before each time point. No significant effects of NPY could be seen at 24 h in vitro(Figs 1a–d). Subsequent time-points revealed significant increases in cell number, BrdU incorporation, nestin immunopositivity and class III β-tubulin immunopositivity relative to controls (Figs 1a–d). Increases in total cell number, BrdU incorporation and nestin immunopositivity all followed a similar exponential curve over time. The cell types that showed the greatest increase in number and BrdU incorporation were small unipolar or bipolar cells with phase-bright cell bodies (Fig. 1e) and which immunostained for nestin, typical of the hippocampal stem/precursor cells previously described by Palmer et al. (1997). The time course of increases relative to controls in β-tubulin-expressing cells was different with a peak at 3 and 5 DIV after which the numbers plateaued. The number of GFAP-positive cells was not increased at 5 DIV compared with controls (34.7 ± 2.5 cells/mm2 in controls vs. 33.6 ± 2.8 cells/mm2 in NPY).
The NPY-induced increase in cells with a neuronal phenotype could be a proliferative and/or a neurotrophic effect. To address this question we quantified cell death, identified by nuclear condensation, using DAPI staining in the same cultures. Cell death was maximal on day 1 in vitro but there was no difference in cell death between control cultures and cultures exposed to NPY at any time point examined (Fig. 1f). We confirmed a proliferative effect of NPY on neuroblasts (β-tubulin- and BrdU-coexpressing cells) and putatively more primitive precursor cells (nestin- and BrdU-coexpressing cells) by exposing 3-DIV cultures to both 1 µm NPY and BrdU for 4 h showing a significant increase in the proportion of both β-tubulin cells (Fig. 1g) and nestin cells (Fig. 1h) coexpressing BrdU, implying a proliferative effect of NPY on both cell phenotypes.
Proliferative effects of neuropeptide Y at concentrations as low as 1 nm
The total cell number and BrdU incorporation were exquisitely sensitive to NPY, responding to doses as low as 1 nm(Figs 2a and b), while counts of cells immunopositive for either nestin or class III β-tubulin were increased at 10 nm (Figs 2c and d).
Neuropeptide Y and basic fibroblastic growth factor have similar effects on nestin-positive cells in culture
In a separate set of experiments we compared the absolute number and phenotypic fate of cells grown for 5 days in the presence of 1 µm NPY, bFGF (20 ng/mL) or under control conditions (NB/B27 and glutamine). The rate of BrdU incorporation on day 5 was measured using a single pulse of BrdU 24 h prior to fixation. Five days growth in the presence of 1 µm NPY yielded significant increases in total cell number, BrdU incorporation and the number of cells immunopositive for nestin as before (Table 2). Interestingly, the magnitude of the NPY effect was almost identical to that of bFGF, which is proliferative for hippocampal stem cells in culture (Palmer et al. 1995, 1999).
Table 2. Comparison of the effects of neuropeptide Y (NPY) and basic fibroblastic growth factor (bFGF)
Cells cultured in NPY for 5 days demonstrate an increase in total cells [4′,6-diamidino-2-phenylindole (DAPI)], 5-bromo-2-deoxyuridine (BrdU) incorporation and nestin expression similar to that seen in cultures grown in the presence of bFGF. All values are expressed as percent of control values (growth in Neurobasal A® supplemented with 2% B27 only). **p < 0.01; ***p < 0.001.
Neuropeptide Y increases sphere formation and sphere size in culture
To investigate the effect of NPY on putative stem cells we cultured cells in the absence of a poly-l-lysine coating to encourage sphere formation, a characteristic of neural stem cells. We found spontaneous sphere formation under control conditions (Table 3). Confocal imaging of spheres exposed to BrdU for 24 h at 8 days showed the spheres to consist largely of nestin cells, many of which had incorporated BrdU (Fig. 3a). Exposure to 1 µm NPY for 8 days increased sphere formation and average sphere size compared with controls (Table 3).
Table 3. Neuropeptide Y (NPY) supports the generation of neurospheres
Control ± SEM
1 µm NPY ± SEM
8 days growth in 1 µm NPY almost doubles the number of neurospheres in comparison to those grown in control media and also results in the formation of a greater number of larger spheres. Sphere counts shown are per 5×field.
Total spheres per 5× field
10.6 ± 1.4
19.2 ± 0.7 p < 0.001
50–75 µm spheres
7.8 ± 0.9
13.2 ± 0.6 p < 0.001
75+ µm spheres
2.8 ± 0.6
6.0 ± 0.8 p < 0.02
Spheres cultured in neuropeptide Y give rise to neurones, astrocytes and oligodendrocytes under differentiating conditions
Spheres grown for 8 days with 1 µm NPY were harvested, re-plated on poly-l-lysine-coated slides and allowed to differentiate in control medium (Vescovi et al. 1993; Gage et al. 1998) for 7 days. Confocal examination of immunostained tissue confirmed the generation of neurones (MAP2-positive cells and class III β-tubulin-positive cells), astrocytes (GFAP-positive cells) (Figs 3b–d) and oligodendrocytes [oligodendrocyte (RIP) positive cells; not shown].
Proliferative effects of neuropeptide Y are mediated via the Y1 receptor subtype
To determine if the activity of NPY is mediated via a single receptor subtype or through a mixed population we used Y1, Y2 and Y5 selective peptide agonists in an effort to mimic the effects of NPY (Fig. 4a). The Y1 receptor selective [F7, P34] NPY was uniquely able to mimic the effects of 1 µm NPY on total cell counts, cells incorporating BrdU and the increase in class III β-tubulin-positive cells. To confirm that the activity of NPY is via the Y1 subtype, we exposed whole hippocampal cultures to 1 µm NPY with the non-peptide Y1 selective antagonist, BIBP3226 for 3 days. BIBP3226 completely abolished the proliferative activity of NPY to basal levels and also abolished the NPY-induced increase in class III β-tubulin-positive cells in comparison to control counts (Fig. 4b). Incubation with BIBP3226 alone had no effect.
Using NPY-Y1 receptor subunit knockout mice we further attempted to corroborate the proliferative activity of NPY as a Y1 receptor-mediated event. Age-matched wild-type and Y1 knockout mice were prepared and cultured as for rat cell cultures. Cell number and BrdU incorporation were significantly increased over controls (p < 0.05) in cultures from age-matched Y1lox/lox animals exposed to 1 µm NPY (Fig. 4c). Cultures from age- and generation-matched Y1–/– animals showed no change in total cell counts or in the number of cells incorporating BrdU in response to NPY (Fig. 4c).
5-Bromo-2-deoxyuridine-, β-tubulin- and nestin-positive cells immunostain for the Y1 receptor
Immunostaining 5-day-old cultures grown in 1 µm NPY against the Y1 receptor showed three patterns of Y1 antibody staining. One pattern of staining was clearly neuronal (Fig. 4d) and cells with this morphology almost always stained positively for the early neuronal marker class III β-tubulin (Lee et al. 1990) (Figs 4e and f). The second pattern of Y1 antibody staining was flat rounded cells with short simple processes (Figs 4d and g). These cells did not stain for class III β-tubulin (Fig. 4f) but did stain for the stem/precursor cell marker nestin (Figs 4h and i). These cells were typically much smaller than GFAP staining astrocytes and had a different shorter pattern of processes (Figs 4d and g). Interestingly, Y1 imunostaining cells with a neuronal morphology were negative for nestin (Figs 4g–i). The final pattern of Y1 immunostaining was small uni- or bipolar cells which were nestin positive (Figs 4j–l).
A total of 83.5 ± 6.6% (n = 400 cells) of nestin immunopositive cells were immunoreactive for Y1 receptor antibody staining at 3 DIV and 100% (n = 400 cells) of BrdU-positive cells after a single 4-h pulse were positive for Y1 receptor antibody staining.
In this study we have demonstrated that NPY is a proliferative factor both for hippocampal nestin-positive precursor cells and β-tubulin-expressing neuroblasts in vitro and that the proliferative effect is mediated via Y1 receptors. To our knowledge this is the first report of a proliferative effect of NPY on hippocampal cells and extends the previously reported proliferative effect on olfactory neuroblasts (class III β-tubulin cells incorporating BrdU) (Hansel et al. 2001) to nestin-positive sphere-forming cells which appear to have stem cell-like properties.
The effect of NPY in increasing cell numbers in culture is clearly proliferative as there was no difference in cell death between cultures grown under control and NPY conditions (Fig. 1f) and the increases in total cell numbers paralleled the increases in 24 h BrdU incorporation rates over 7 days in culture (Figs 1a and b). As the cell cycle time of hippocampal precursor cells from early post-natal rats is 16 h (Lewis 1978; Nowakowski et al. 1989), the increased proportion of class III β-tubulin-staining cells incorporating BrdU and of nestin-positive cells incorporating BrdU after just 4 h coexposure to 1 µm NPY, therefore, implies that NPY had a proliferative action on both cell types. That this increase in BrdU incorporation is translated into increased cell proliferation is supported by our finding that significant increases in BrdU incorporation were associated with significant increases in cell numbers in both nestin-positive cells and β-tubulin-positive cells at 3 DIV (Figs 1a–d). We, therefore, conclude that NPY has separate proliferative effects on cells immunostaining for nestin and for class III β-tubulin.
The exponential-like increase in nestin-positive cells during NPY exposure over time was not mirrored by a parallel increase in class III β-tubulin-positive cells (Figs 1c and d) which is expressed in dividing neuroblasts and mature neurones (Lee et al. 1990). This implies a lack of a differentiating effect towards a neuronal lineage of NPY on nestin-positive cells. While the nestin-positive cell population appeared to retain its ability to proliferate in response to NPY, the class III β-tubulin-positive cell population did not, with cell numbers of this phenotype increasing between days 1 and 3 in culture but plateauing thereafter (Fig. 1d). This was corroborated by our finding that short pulse (4 h) coexposure of BrdU and NPY to cultures older than 3 days showed no increase in the proportion of class III β-tubulin-positive cells incorporating BrdU (data not shown). These results are consistent with NPY causing committed neuroblasts to divide and become terminally differentiated neurones whilst having a separate and purely proliferative effect on a more primitive nestin-positive precursor cell.
It is possible that there was an overlap in the cell phenotype staining for nestin and class III β-tubulin in our culture system; however, a number of our findings would argue against this being a significant population. Firstly, data from the time course experiments showed that, at 5 DIV, there were a mean of 140 nestin-positive cells/mm2 compared with 80 class III β-tubulin-positive cells/mm2 with an even greater difference at 7 DIV with NPY clearly having an effect on a population of nestin-positive cells that were not class III β-tubulin positive. Secondly, the significant increase in nestin-positive cells over time was not mirrored by an increase in class III β-tubulin-positive cells after day 3 in vitro. Thirdly, although cells that immunostained for the presence of the Y1 receptor had varying morphologies, those with a neuronal morphology did not stain for nestin but did stain for the neuronal marker class III β-tubulin (Figs 4d–f).
There was no evidence of either a proliferative or differentiating effect towards an astrocytic phenotype as evidenced by a lack of increase in GFAP immunostaining cells after exposure to NPY.
Consistent with previous reports (Brewer 1999) we found that hippocampal cell cultures showed spontaneous proliferation in defined serum-free conditions with B27 supplement. This, however, does not detract from the proliferative effect of NPY as proliferation under NPY conditions was compared with time-matched control cultures taken from the same animals. Indeed, this was the reason we performed time- and animal-matched control cultures. Also, given the specificity of the Y1 receptor agonist and antagonists used to mimic the effect of NPY and the corroboration of our results in cultures from Y1 knockout animals, the proliferative effect of NPY is not in doubt.
The NPY-responsive nestin-positive cells in our study shared many of the characteristics of hippocampal stem cells described by Palmer et al. (1995, 1997), being small uni- or bipolar cells, with phase-bright cell bodies (Fig. 1e), that are nestin positive (Fig. 3a), proliferate in response to bFGF (Table 2), show spontaneous sphere formation (Table 2 and Fig. 3a) and generate neurones, astrocytes and oligodendrocytes after growth factor withdrawal (Figs 3b–d). The spontaneous sphere formation we observed in culture is most likely to have arisen from hippocampal precursor cells and/or subependymal cells as both of these cell types would have been present in the whole hippocampal preparation. It is also possible that some of these precursor cells may have arisen from the dentate gyrus, as region-specific dentate cultures from post-natal day (P)10 rats generate spheres spontaneously (Seaberg and van der Kooy 2002) consistent with the early post-natal generation of the granule cell layer. Preliminary studies suggest that the proliferative effect of NPY on nestin-positive cells is separate to that of bFGF and further studies addressing this and the proliferative effect of NPY on region-specific cultures from the hippocampus proper, subependymal layers and dentate gyrus are in progress.
Using specific Y1 receptor agonists and antagonists we have shown that the proliferative effect of NPY is mediated via its Y1 receptor. We demonstrated a loss of proliferative effect of NPY in cultures from Y1 receptor knockout mice with preservation of the proliferative effect in Y1lox/lox controls, confirming that NPY-induced proliferation is mediated via the Y1 receptor. In addition, we have demonstrated that 100% of BrdU-positive cells after a 4-h pulse were Y1 receptor antibody positive. We have also demonstrated that the neuroproliferative effect of NPY is specifically mediated via the Y1 receptor by confirming this effect using a specific Y1 agonist, abolishing the neuroproliferative effect with a Y1 antagonist and showing an absence of a neuroproliferative effect with specific Y2 and Y5 agonists. In addition, immunohistochemistry against the NPY Y1 receptor confirmed Y1 receptor staining on both nestin-positive immature precursor-like cells and on class III β-tubulin-positive cells with a neuronal morphology (Figs 4f, i and l).
We chose the early post-natal period for generating hippocampal cultures as the granule cell layer is predominantly formed between P5 and P20 in rats (Altman and Bayer 1990) allowing for the greatest yield of neuroblasts and other precursor cells for culture. Studies of basic helix loop proteins expression in cells in the SGZ of the hippocampus and during development have shown that precursor cells in the mature dentate gyrus share features with precursor cells found in development, specifically expression of MASH1 and NeuroD (Pleasure et al. 2000). Specifically relevant to this discussion is the comparison between the tertiary matrix (present up to P14) which gives rise to the SGZ (present from P5) which is the site of neurogenesis throughout adulthood. Pleasure et al. (2000) have shown that MASH1 is expressed in the tertiary matrix and SGZ at P8 and is limited to the SGZ in adulthood and that NeuroD is expressed in newly born granule cells at all stages of development, i.e. the molecular profiles of precursor cells in the post-natal period were identical to that of precursors in adulthood. The neuroblast cells (class III β-tubulin positive incorporating BrdU) upon which NPY had a proliferative effect most probably originated from the dentate gyrus as this contains the predominant neurogenic zone in the post-natal hippocampus and is undergoing rapid proliferation at this time. Kopp et al. (2002) have recently demonstrated that the greatest concentration of hippocampal Y1 receptors is in the molecular and granule cell layers of the dentate gyrus, including a distinct band of staining on cell bodies in the SGZ, the neuroproliferative area of the dentate gyrus. The morphology of these cells is reminiscent of the small rounded BrdU-positive SGZ precursor cells found in vivo. It is intriguing to speculate that, given the large concentration of NPY Y1 receptors in the neurogenic SGZ of the dentate, the abundance of NPY interneurones in the dentate hilus adjacent to the SGZ and the proliferative effect of NPY on stem-like precursor cells and neuroblasts in hippocampal culture, NPY is a key modulator of hippocampal neurogenesis in vivo.
Neuropeptide Y, neurogenesis and cognitive function
Hippocampal learning (Gould et al. 1999b), as well as lesioning of the perforant path or NMDA receptor blockade, increases dentate neurogenesis (Cameron et al. 1995) but the mechanisms underlying these effects are unknown. The NPY interneurones in the dentate hilus 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 interneurones) (Freund and Buzsaki 1996). In addition to their role in modulating granule cell excitability, NPY interneurones are thus ideally positioned to sample patterns of afferent and efferent activity and thus modulate neurogenesis via NPY release on SGZ precursors.
Neuropeptide Y is one of the most conserved peptides in evolution (Larhammar 1996), suggesting an important role in the regulation of physiological function. Intracerebroventricular injections of NPY fragments have been shown to alter cognitive function in association with learning and memory (Flood et al. 1987; Dumont et al. 1992; Whittaker et al. 1999). Transgenic rats with hippocampal NPY overexpression have impaired spatial learning, which is associated with decreased NPY–Y1 binding within the hippocampus (Thorsell et al. 2000). This finding is particularly intriguing given that blockade of dentate neurogenesis reduces a hippocampal-based learning task (Shors et al. 2001). Some of the effects of NPY on learning and memory may thus be mediated via the effect of NPY on dentate neurogenesis and further studies on NPY, neurogenesis, learning and memory are warranted.
Acute stress reduces dentate neurogenesis, an effect mediated via adrenal steroids (Gould et al. 1997) and reduced dentate neurogenesis has been proposed as a cause of clinical depression (Jacobs et al. 2000; Kempermann 2002). Neuropeptide Y is important for habituation to repeated stress (Thorsell et al. 2000) and the mechanism of the effect of NPY on stress habituation may be by counteracting the direct inhibitory effect of steroids on neurogenesis.
Neuropeptide Y, neurogenesis and affective disorders
Several avenues of research have suggested that decreased hippocampal neurogenesis may be important in the pathogenesis of depression (for review see D'Sa and Duman 2002). Stress, often a requisite for depression, reduces hippocampal neurogenesis (Gould et al. 1997). Chronic administration of antidepressants increases hippocampal neurogenesis (Malberg et al. 2000) and chronic tianeptine treatment reverses the stress-induced decrease in hippocampal neurogenesis (Czeh et al. 2001). Interestingly, clinical studies have reported reduced NPY in the cerebrospinal fluid and plasma from depressed patients (Westrin et al. 1999). Animal models of depression, such as the Flinders sensitive line rat (Caberlotto et al. 1999) and rats exposed to early post-natal stress (Husum and Mathe 2002), have reduced hippocampal levels of NPY and altered Y1 binding levels as adults. Intriguingly, the Y1 receptor has recently been identified as the receptor subtype mediating an antidepressant-like activity of NPY (Redrobe et al. 2002). These studies are consistent with our hypothesis that NPY modulates hippocampal neurogenesis.
This work was supported by grants from the Wessex Medical Trust Hope, the University of Southampton and the MRC under the JREI scheme to WPG and the HFSP to HES, HH and AB-S. The authors are grateful to Dr Matt Cuttle for his excellent technical assistance with the confocal microscopy.