Secretoneurin promotes pertussis toxin-sensitive neurite outgrowth in cerebellar granule cells

Authors


Address correspondence and reprint requests to Alois Saria, Department of Neurochemistry, University Hospital of Innsbruck, Anichstrasse 35, 6020 Innsbruck. E-mail: alois.saria@uibk.ac.at

Abstract

The neuropeptide secretoneurin (SN) is an endoproteolytic product of the chromogranin secretogranin II. We investigated the effects of SN on the differentiation of immature cerebellar granule cells derived from the external granular layer (EGL). Secretoneurin caused concentration-dependent increases in neurite outgrowth, reflecting differentiation. The maximum effect was reached at a concentration of 100 nm SN. Secretoneurin immunoneutralization using specific antiserum significantly decreased neurite outgrowth; however, neurite morphology was altered. An affinity chromatography-purified antibody significantly inhibited the outgrowth response to SN (p < 0.001) without altering the morphology. Binding studies suggest the existence of specific G-protein-coupled receptors on the surface of monocytes that recognize SN. Assuming that SN promotes neurite outgrowth in EGL cells by acting through a similar G-protein-coupled mechanism, we treated SN-stimulated EGL cultures with pertussis toxin. Exposure to pertussis toxin (0.1 µg/mL) showed a significant inhibition of the SN-induced outgrowth. To establish a second messenger pathway we used the protein kinase C inhibitor staurosporine. We found that EGL cell viability was not enhanced following chronic SN treatment for 24 h. These data indicate that SN is a novel trophic substance that can affect cerebellar maturation, primarily by accelerating granule cell differentiation through a signalling mechanism that is coupled to pertussis toxin-sensitive G-proteins.

Abbreviations used
DAB

diaminobenzidine

EGL

external granular layer

GFAP

glial fibrillary acidic protein

MAP

microtubule-associated protein

PBS

phosphate-buffered saline

PKC

protein kinase C

PTx

pertussis toxin

SgII

secretogranin II

SN

secretoneurin

Secretoneurin (SN), a neuropeptide generated by endoproteolytic processing of secretogranin II (SgII) (Vaudry and Conlon 1991; Kirchmair et al. 1993), is found in large dense core vesicles in endocrine and nervous tissues (Fischer-Colbrie et al. 1995; Cozzi et al. 1989). In the bovine and rat brain, 90% of SgII is proteolytically processed (Kirchmair et al. 1993; Marksteiner et al. 1993a). Secretoneurin appears early in embryonic life (Leitner et al. 1997; Saria et al. 1997). In rat, SgII is fully processed to the free peptide SN from gestational day 13 to adulthood. The SgII mRNA is widely but distinctly distributed in brain (Kirchmair et al. 1993) (Mahata et al. 1992) including the cerebellum (Cozzi et al. 1989). A number of biological effects of SN have been reported, such as transmitter release (Saria et al. 1993; Agneter et al. 1995), chemotaxis (Reinisch et al. 1993; Kähler et al. 1996; Dunzendorfer et al. 1998; Kong et al. 1998), proliferation (Kähler et al. 1997a,b), gonadotropin release from the pituitary (Blazquez et al. 1998) and survival of cerebellar granule cells (Fujita et al. 1999). Secretoneurin-like immunoreactivity showed a widespread distribution (Marksteiner et al. 1993a,b) in endocrine (Kirchmair et al. 1993; Schmid et al. 1995) and nervous tissues (Kirchmair et al. 1994; Dun et al. 1997) of various species. Cerebellar granule neurones are the most numerous in mammalian brain. In the mouse, precursors first appear in the neuroepithelium and migrate to the pial surface of the cerebellum (Hatten 1985; Hynes et al. 1986). These precursors spread across the surface to form a layer of proliferating cells, the external granular layer (EGL). As cell precursors exit the cell cycle, they migrate along Bergmann glia processes, through the molecular and Purkinje cell layers into the internal granular layer, where they reside permanently (Hatten 1999).

External granular layer cells are frequently used for studies of potential neurotrophic substances, because they can be isolated to relative homogeneity to yield highly enriched populations of neurones. Furthermore, because the EGL contains neuroblasts at the earliest stages of granule cell development, proliferation, differentiation, migration and programmed cell death can be subsequently assayed as these cells mature. Neurites elongate rapidly in this system and, therefore, effects of neurotrophic substances can be studied in short-term cultures. A specific SN receptor has not yet been identified. However, specific high affinity binding sites for SN have been characterized on human monocytes as well as two monocytic cell lines (Kong et al. 1998; Schneitler et al. 1998). Treatment of monocytes with SN increased intracellular calcium levels and SN-induced chemotaxis was inhibited by pertussis toxin (PTx) as well as by a protein kinase C (PKC) inhibitor (Schratzberger et al. 1996), suggesting that SN might act via a classical G-protein-coupled neuropeptide receptor, as with other neuropeptides fully characterized to date.

Therefore, we investigated a possible influence of SN on neurite outgrowth and its underlying mechanisms in cerebellar granule cells. The SN specificity was evaluated with a SN antibody and the signalling with PTx and staurosporin.

Materials and methods

Tissue preparation, cell culture and treatment

Post-natal day 5 or 6 mice pups (Balb/c) were killed by decapitation and the cerebellum was dissected under aseptic conditions. Experiments were carried out in accordance with European Communities Council Directive of 24 November 1986 (86/609/EEC) and with the Austrian legislation on animal experiments. The neuronal precursor cells of the EGL were isolated and purified by the protocol of Hatten (1985) modified by Hauser et al. (2000). Briefly, whole cerebella were washed in Ca2+- and Mg2+-free Tyrode's solution, chopped twice in a tissue chopper and dissolved in DNAse I (Roche, Mannheim, Germany). After a short centrifugation, cells were resuspended in DNAse again and passed through a 35-µm nylon mesh filter (Falcon®; Becton-Dickinson, Franklin Lakes, NJ, USA). Cerebellar cells were separated from glia and remaining clusters by centrifugation by layering cells onto a step density gradient of Percoll (35 and 60% in Ca2+- and Mg2+-free Tyrode's solution containing 2 mm EDTA; Amersham Pharmacia Biotech, Uppsala, Sweden). The gradients were centrifuged at 2100 g for 15 min and one of the two resulting bands, at the interface between the layers of 35 and 60% Percoll, was removed with a fire-polished Pasteur pipette, diluted with Ca2+- and Mg2+-free Tyrode's solution, pelleted at 600 g for 7 min and resuspended in medium (Basal Medium Eagle; GibcoTM, Paisley, UK) containing glutamine, glucose, 10% horse serum and 5% fetal calf serum. Cells were plated on poly-d-lysine culture flasks (Falcon®, Becton Dickinson Labware Europe, La Pont de Clair, France) for 15 min, where remaining glia attached to the ground. The EGL precursors were washed, counted and plated on an uncoated 96-well tissue culture dish (Falcon®) at a concentration of 3 × 105 cells/well. Plates were incubated at 37°C/5% CO2 for 16–24 h to form neurospheres with cluster sizes of 20–50 cells. Neurospheres were centrifuged, resuspended in serum-free medium (Basal Medium Eagle) containing the supplement factors B27 (2% v/v) and N2 (1% v/v; Invitrogen, Life Technologies, Karlsruhe, Germany) and plated on polylysine-coated 24-well dishes (Costar®, Microtest™, Becton Dickinson Labware Europe) with a centrally positioned 15-mm diameter glass.

Cells were exposed to different concentrations of SN (10, 30, 50, 100 and 300 nm; Neosystem, Strasbourg, France), PTx (0.1 µg/mL; Calbiochem, Dermstadt, Germany), staurosporin (50 nm; Roche Molecular Biochemicals, Indianapolis, IN, USA), SN antiserum and affinity chromatography-purified SN antibody. Antiserum and antibodies were pre-incubated with SN for 24 h at 4°C and PTx and staurosporine were added to cell cultures 1 h before SN exposure.

Purification of secretoneurin antiserum by affinity chromatography

The SN antiserum (polyclonal, rabbit) was generated as described previously (Kirchmair et al. 1993). Protein A-Sepharose CL-4B bead (Pharmacia Biotech, Uppsala, Sweden) columns were prepared according to the manufacturer's instructions. The column was washed with phosphate-buffered saline (PBS; pH 7.4) until base line was achieved. The SN antiserum was loaded into the column at a flow rate of 0.4 mL/min. Bound immunoglobulins were eluted with acetic acid (0.58% HAc/0.15 m NaCl). The IgG fraction was dialysed against serum-free medium (Basal Eagle Medium) containing 2 mm glutamine Ig overnight at 4°C.

Immunocytochemistry

Cells were fixed in 4% paraformaldehyde for 10 min at 4°C and rinsed in 100 mm PBS, pH 7.4. Cultures were incubated in 100 mm PBS/0.3% Triton X-100 for 30 min and the endogenous peroxidase was blocked with 1% H2O2, 5% methanol in PBS. After EGL cells were rinsed with 100 mm PBS, they were incubated in 20% horse serum (Gibco BRL, Karlsruhe, Germany)/100 mm PBS/0.3%Triton X-100 for 1 h at room temperature. Cultures were incubated in microtubule-associated protein (MAP-2; Sigma Immunochemicals, St Louis, MO, USA) or anti-glial fibrillary acidic protein (GFAP; Sigma Immunochemicals) antibodies diluted 1 : 200 at room temperature on an orbital shaker for 24 h. Secondary biotinylated anti-mouse/anti-rabbit antibodies in 100 mm PBS/0.3% Triton X-100/0.2% bovine serum albumin were used as directed (Vectastain Elite ABC kit; Vector Laboratories, Burlingame, CA, USA) to detect MAP-2 and GFAP antibodies. As a substrate for peroxidase, diaminobenzidine (DAB) solution consisting of 0.05% DAB and 0.3% H2O2 in 50 mm Tris/HCl, pH 7.6, was used. The reaction was observed under the microscope and stopped by removing the DAB solution and by washing with 50 mm Tris/HCl. The plate was kept at 4°C until measurement.

Quantification of neurite growth

Neurospheres were visualized using a phase contrast microscope (Leitz, Wetzler, Germany) equipped with a × 20 objective and connected to a computer system. Neuritic length was determined using Image-Pro Plus software (Media Cybernetics, Gleichen, Germany). Measurement operations are performed in terms of screen pixel positions, e.g. the number of pixels along the line was used to determine length. Numbers of pixels were converted to µm. All neurites longer than 10 µm in each culture well were measured to obtain an average length.

Neurites were systematically excluded if they overlapped with other neurites, or an EGL cell or cell cluster or if the neurite did not have a defined cell of origin or ending.

Neurone viability

Neurone viability was assessed at 24 h following continuous SN (100 nm) treatment. Viability was established using a commercially available kit (Live-Dead Assay; Molecular Probes, Eugene, OR, USA) as previously reported (Hauser et al. 2000). Briefly, cultures were incubated in basal cell culture media containing ethidium homodimer (3.5 µm) and calcein-AM (4 µm; Molecular Probes) for 40 min at 35°C in 5% CO2/95% air. Red ethidium fluorescence (dead cells) was detected at 535 nm excitation and 590 nm emission wavelengths and green calcein fluorescence (live cells) was detected at 485 nm excitation and 530 nm emission wavelengths. Determinations were made on (n = 4) independent samples of cells derived from different animals.

Results

Immunocytochemistry (anti-glial fibrillary acidic protein and microtubule-associated protein-2)

The results from immunocytochemical experiments are shown in Fig. 1. Figure 1(a) represents a control experiment without a primary antibody. Cultured EGL cells did not show any GFAP (Fig. 1b) immunoreactivity, suggesting the absence of contaminating astroglia. Immunocytochemical detection of MAP-2, which is confined to neuronal cell bodies and dendrites in the CNS, clearly identified neurone clusters and developed processes (Fig. 1c).

Figure 1.

Immunocytochemical analysis of cerebellar EGL cells fixed after 48 h. (a) Control culture treated with the biotinylated anti-mouse antibody only. (b) Staining of SN-treated clusters with GFAP-antibody. (c) Staining of SN-treated clusters with MAP-2 antibody (affinity to cell bodies and dendrites). Scale bars = 100 µm.

Dose-dependent effect of secretoneurin on neurite outgrowth

The EGL cells were cultured for 48 h in a neurotrophin-free, defined medium for controls, or with SN added at different concentrations (10, 30, 50, 100 and 300 nm). Neurites elongate rapidly in the reaggregated cell cultures and, therefore, effects of factors promoting growth can be studied in short-term cultures. Prior to 20 h in serum-free medium, neither control nor SN-treated EGL cell cultures showed a measurable neurite outgrowth. For this reason, neurite outgrowth was analysed at 24 and 48 h (Figs 1 and 2).

Figure 2.

Dose-dependent effect of SN on neurite outgrowth of EGL cells after 24 and 48 h. The results show the mean neurite length (in µm) ± SEM from four to ten different experiments (SN 10 nm, n = 4; SN 30 nm, n = 4; SN 50 nm, n = 7; SN 100 nm, n = 10; SN 300 nm, n = 5). Statistical analysis was carried out with a two-tailed unpaired t-test. As a total of eight comparisons were performed, the level of statistical significance for each individual comparison was adjusted to p = 0.0062 (Bonferroni's correction). *Significantly different from control, p < 0.0062.

Control groups incubated only in serum-free medium alone showed a mean neurite outgrowth of 54 ± 2 µm at 24 h and 89 ± 2 µm at 48 h, corresponding to an increase of 64.8%. Concentrations of 10 nm SN (24 h, 52 ± 2 µm; 48 h, 82 ± 5 µm) and 30 nm SN (24 h, 58 ± 7 µm; 48 h, 71 ± 7 µm) did not influence EGL differentiation, in contrast to higher concentrations, e.g. 50 nm after 24 h (86 ± 2 µm), 100 nm SN (24 h, 114 ± 3 µm, 2.1-fold increase; 48 h, 168 ± 5 µm, 1.9-fold increase) or 300 nm SN (24 h, 109 ± 3 µm; 48 h, 157 ± 8 µm) (Fig. 3).

Figure 3.

SN dose dependently stimulates neurite outgrowth of EGL clusters after 24 and 48 h. Photographs of clusters in serum-free medium are taken at two time points: (a) no added SN after 24 h, n = 10; (b) no added SN after 48 h, n = 10; (c) 100 nm SN after 24 h, n = 10; (d) 100 nm SN after 48 h, n = 10. Scale bars = 100 µm.

Effects of antisera on secretoneurin-induced neurite outgrowth

In the course of conducting inhibition measurements with SN and rabbit SN antiserum, we noticed a reproducible effect of SN antiserum on the pattern of neurite outgrowth. Instead of individual processes, a dense pattern of neurites within a halo was found. Quantification as described previously was no longer feasible because fasciculation made it impossible to discern individual processes. The average radius of the halo of processes extending from the aggregate was diminished compared with SN-treated neurites, but these data were not comparable with our method of quantification (Fig. 4).

Figure 4.

Effect of SN antiserum on neurite outgrowth in EGL clusters in the presence and absence of SN. Please note the morphological changes in photographs (b)–(d) versus (a). (a) EGL clusters under control conditions with no added SN or antiserum, n = 8; (b) with 100 nm SN + SN antiserum after 24 h, n = 8; (c) with 100 nm SN + SN antiserum after 48 h, n = 8; (d) with rabbit pre-immunserum only. Scale bars = 100 µm.

Granule cell reaggregates were pre-treated for 1 h with the affinity chromatography-purified SN antibody. Cells were then exposed to 100 nm SN, the dose with the maximum effect on neurite outgrowth. In comparison to neurospheres treated with 100 nm SN only neurite outgrowth was significantly decreased (p < 0.0001). Other parameters of neurite outgrowth, including number of neurites and neurite fasciculation, were not qualitatively different among the treatments. Cultures treated with purified SN antibody only were not significantly different from control cultures (Fig. 5).

Figure 5.

Effect of the purified SN antibody on neurite elongation in EGL clusters in the presence and absence of SN. Length of neurites is measured after 24 and 48 h. The results show the mean neurite length (in µm) ± SEM from six different experiments. Statistical analysis was carried out with a two-tailed unpaired t-test. As a total of six comparisons were performed, the level of statistical significance for each individual comparison was adjusted to p = 0.008 (Bonferroni's correction). **p < 0.0001, *p < 0.001.

Neurite outgrowth response to secretoneurin is inhibited by pertussis toxin

The effect of PTx is demonstrated in Figs 6and 7. We noticed that neurospheres exposed to PTx in comparison to control groups showed a significant decrease in neurite elongation (p < 0.0001). The ability of SN to induce neurite outgrowth was blocked by pre-treatment of neurospheres with PTx. The toxin significantly inhibited outgrowth at a concentration of 1 µg/mL compared with SN-treated cultures (p < 0.0001).

Figure 6.

Pertussis toxin inhibits SN-promoted neurite outgrowth in EGL neurospheres. The length of neurites was measured after 24 and 48 h. Note the significantly diminished length of neurites in cultures exposed to the PTx (n = 7) cultures. Statistical analysis was carried out with a two-tailed unpaired t-test. As a total of six comparisons were performed, the level of statistical significance for each individual comparison was adjusted to p = 0.008 (Bonferroni's correction). **p < 0.0001. The mean of results from seven different experiments is shown. Error bars represent each SEM.

Figure 7.

Exposure of EGL neurospheres to SN and PTx in vitro. Treatment with (a) 0.1 µg/ml PTx only and (b) pretreatment with PTx before incubation with 100 nm SN over a time course of 48 h. Scale bars = 100 µm.

Neurite outgrowth response to secretoneurin is inhibited by staurosporine

Neurospheres exposed to staurosporine showed a significant decrease in neurite elongation (p < 0.0001) compared with control groups (Fig. 8). The ability of SN to induce neurite outgrowth was blocked by pre-treatment of neurospheres with staurosporine. The PKC inhibitor significantly decreased neurite outgrowth at a concentration of 50 nm compared with SN-treated cultures (p < 0.0001).

Figure 8.

Staurosporine inhibits SN-promoted neurite outgrowth in EGL neurospheres in vitro. Length of neurites is measured after 24 and 48 h. Statistical analysis was carried out with a two-tailed unpaired t-test. As a total of six comparisons were performed, the level of statistical significance for each individual comparison was adjusted to p = 0.008 (Bonferroni's correction). **p < 0.0001, *p < 0.001.

Neurone viability is not affected by secretoneurin

After observation of the survival of immature granule cells and/or their precursors following 24 h continuous exposure, we determined that treatment with SN (100 nm) did not affect the proportion of dying cells. Controls, treated with vehicle, showed a similar number of non-viable neurones (9.4 ± 0.6%) to SN-treated granule cells (10.0 ± 1.2; Fig. 9).

Figure 9.

Effect of SN on the survival of immature granule cells and/or their precursors following 24 h of continuous exposure. (a–b) Fluorescence microscopic images of living (green) and non-viable (red) cells in vehicle-treated control (a) and SN-treated (b) cultures. Exposure to SN (100 nm) did not affect the proportion of dying cells. Scale bar = 20 μm.

Discussion

Recent immunohistochemical (Marksteiner et al. 1993b) as well as functional (Saria et al. 1993) data suggest that SN, the cleavage product of Sg II (Kirchmair et al. 1993), may fulfil a role as neuropeptide transmitter or modulator in the brain. We have utilized reaggregated EGL cells to study the influence of SN on differentiation and to investigate the underlying mechanisms.

The present results show that SN promotes neurite outgrowth in cerebellar granule cells in vitro. The SN-evoked differentiation was dose dependent, the dose curve reaching a maximum value at a concentration of 100 nm SN. This suggests the involvement of a receptor with an affinity that is comparable to other neuropeptide seven-transmembrane receptors or to receptors for trophic factors such as glial cell line-derived neurotrophic factor (Choi-Lundberg and Bohn 1995) or brain-derived neurotrophic factor (Segal et al. 1995). However, the slope of the dose–response curve is steeper than that obtained with single ligand–receptor interaction. This indicates the involvement of multiple receptors or additional mechanisms in the response to SN.

The maximum effect of SN in the present study was a 1.9- to 2.1-fold increase in neurite elongation at a dose of 100 nm SN as a concentration of 300 nm SN did not further increase neurite growth within 48 h. This is comparable to other substances effecting neurite outgrowth in the developing cerebellum, e.g. exposure to somatostatin resulted in a statistically significant dose-dependent increase in the neurofilament content as an indicator of neurite outgrowth (Taniwaki and Schwartz 1995). As SN in our experiments, somatostatin had a maximum effect at 100 nm. Hatten et al. (1988) found a dose-dependent effect of basic fibroblast growth factor on granule cell neurite outgrowth. Outgrowth was stimulated markedly in the presence of 0.05–1.4 nm basic fibroblast growth factor, but effects were not seen below 0.05 or above 2.8 nm.

An SN antiserum was utilized to evaluate the SN specificity on cerebellar granule cells. Exposure of SN-treated cultures to an SN antiserum caused a different morphology, which made quantification and comparisons with untreated cultures impossible. Possibly, SN antiserum contains different growth-promoting and -inhibiting substances which influence neurite outgrowth. The pattern of fasciculation was similar to the growth pattern described in neurotrophin-3-treated cerebellar granule cell cultures (Segal et al. 1995).

Addition of SN, pre-incubated with an affinity-purified antibody to reaggregated granule cells, severely impaired process extension by granule neurones which indicates that SN selectively enhances neurite elongation. Fujita et al. (1999) determined SgII mRNA in cerebellar granule cell cultures which indicates the existence of endogenous SN. However, the magnitude of neurite outgrowth in cultures treated with immuno-neutralized SN did not differ significantly (either increased or decreased) from controls. This suggests that endogenous SN, assumed to be located in large dense core vesicles (Kirchmair et al. 1993), is not accessible to the exogenously applied antibody and release of SN does not occur under our experimental conditions.

The mechanism of signal transduction following SN treatment has not been completely understood until now. Previous studies demonstrated that SN elevates intracellular calcium concentrations in human monocytes and that the chemotactic activity of SN can be blocked by treatment of cells with PTx (Schratzberger et al. 1997; Kong et al. 1998). In addition, SN was shown to interact with specific cell surface binding sites on human monocytes (Kong et al. 1998). In the present study, SN-evoked neurite outgrowth could be abolished by PTx, indicating the existence of a specific SN receptor on cerebellar granule cells coupled to Gi/o-protein. This theory is supported by the observation of a staurosporine-dependent decrease in neurite outgrowth which indicates the involvement of the PKC. Alternatively, SN could act through activation of another neurite outgrowth-promoting substance using a G-protein-coupled mechanism. One possibility is somatostatin, which increased neurite outgrowth in PC12 cells and enhanced neurite outgrowth after nerve growth factor exposure, as this effect was inhibited by somatostatin antibody and PTx (Ferriero et al. 1994). The PTx inhibited spontaneous growth of neurites in the absence of SN. This indicates that other G-protein-coupled factors are involved in the spontaneous neurite outgrowth of EGL cultures.

A number of other factors, which exert their neurite outgrowth-promoting activity on cerebellar granule cells, can be excluded as being involved in the SN-dependent neurite elongation because of their Gi/o-protein absent signal transduction. It has been shown that cerebellar neurones are responsive to neurotrophins (Segal et al. 1992; Lindholm et al. 1993; Leingartner et al. 1994), a family of related protein factors acting through the Trk receptors. Endogenous glutamate, acting through the PKC-coupled N-methyl-d-aspartate receptor (Cambray-Deakin and Burgoyne 1990; 1992), as well as pituitary-adenylate-cyclase-activating polypeptide that stimulates the adenylyl-cyclase and phospholipase C transduction pathways (Vaudry et al. 1998a), are known to promote neurite outgrowth in immature cerebellar granule cells.

It is known that proliferative granule neurones undergo apoptosis more frequently than post-mitotic granule neurones in EGL of the developing cerebellum. This suggests that there are developmental stage-specific mechanisms of apoptosis of cerebellar granule neurones and a direct relation between apoptosis and proliferation (Tanaka and Marunouchi 1998). Therefore, an influence of apoptosis on the time course of differentiation cannot be excluded. In our experiments SN did not enhance apoptosis; in consequence we suggest that the observed SN-promoted neurite outgrowth is an independent trophic effect. Therefore, SN belongs to the kind of trophic factors which have an exclusive effect on neurite outgrowth without influencing programmed cell death, e.g. neuregulin (Rieff et al. 1999) and somatostatin (Taniwaki and Schwartz 1995), in contrast to substances such as brain-derived neurotrophic factor (Segal et al. 1992; Segal et al. 1995), pituitary-adenylate-cyclase-activating polypeptide (Vaudry et al. 1998b) or basic fibroblast growth factor (Hatten et al. 1988) which promote both neurite outgrowth and survival. The present results demonstrate that SN promotes PTx-sensitive neurite outgrowth in cultured immature granule cells and suggest that SN may be a novel trophic factor involved in the differentiation of cerebellar granule cells.

Acknowledgements

This work was supported by the Austrian Science Fund SFB F00206. We thank Prof. Christian Humpel for help with the immunocytochemistry. The authors acknowledge the excellent help of Iris Berger and Margit Auffinger.

Ancillary