Nitric oxide and cyclic nucleotide signal transduction modulates synaptic vesicle turnover in human model neurons

Authors

  • Million Adane Tegenge,

    1. Division of Cell Biology, Institute of Physiology, University of Veterinary Medicine Hannover, Hannover, Germany
    2. Center for Systems Neuroscience (ZSN), Hannover, Germany
    Search for more papers by this author
  • Michael Stern,

    1. Division of Cell Biology, Institute of Physiology, University of Veterinary Medicine Hannover, Hannover, Germany
    Search for more papers by this author
  • Gerd Bicker

    1. Division of Cell Biology, Institute of Physiology, University of Veterinary Medicine Hannover, Hannover, Germany
    2. Center for Systems Neuroscience (ZSN), Hannover, Germany
    Search for more papers by this author

Address correspondence and reprint requests to Gerd Bicker, Division of Cell Biology, Institute of Physiology, University of Veterinary Medicine Hannover, Bischofsholer Damm 15, D-30173 Hannover, Germany. E-mail: gerd.bicker@tiho-hannover.de

Abstract

The human Ntera2 (NT2) teratocarcinoma cell line can be induced to differentiate into post-mitotic neurons. Here, we report that the human NT2 neurons generated by a spherical aggregate cell culture method express increasing levels of typical pre-synaptic proteins (synapsin and synaptotagmin I) along the neurite depending on the length of in vitro culture. By employing an antibody directed against the luminal domain of synaptotagmin I and the fluorescent dye N-(3-triethylammoniumpropyl)-4-(4-(dibutylamino)styryl)pyridinium dibromide, we show that depolarized NT2 neurons display calcium-dependent exo-endocytotic synaptic vesicle recycling. NT2 neurons express the neuronal isoform of neuronal nitric oxide synthase and soluble guanylyl cyclase (sGC), the major receptor for nitric oxide (NO). We tested whether NO signal transduction modulates synaptic vesicle turnover in human NT2 neurons. NO donors and cylic guanosine-monophosphate analogs enhanced synaptic vesicle recycling while a sGC inhibitor blocked the effect of NO donors. Two NO donors, sodium nitroprusside, and and N-Ethyl-2-(1-ethyl-2-hydroxy-2-nitrosohydrazino) ethanamine evoked vesicle exocytosis which was partially blocked by the sGC inhibitor. The activator of adenylyl cyclase, forskolin, and a cAMP analog induced synaptic vesicle recycling and exocytosis via a parallel acting protein kinase A pathway. Our data from NT2 neurons suggest that NO/cyclic nucleotide signaling pathways may facilitate neurotransmitter release in human brain cells.

Abbrevations used:
8-Br-cAMP

8-Bromoadenosine 3′, 5′-cyclic adenosine monophosphate, sodium salt

cGMP

cylic guanosine-monophosphate

cGMP-IR

cGMP-immunoreactive

DIV

days in vitro

FM1-43

N-(3-triethylammoniumpropyl)-4-(4-(dibutylamino)styryl)pyridinium dibromide

FM1-43Fx

fixable analog of FM1-43

H-89

N-[2-((p-Bromocinnamyl) amino) ethyl]-5-isoquinolinesulfonamide, 2HCl

KRH

Krebs-Ringer’s – HEPES

nNOS

neuronal nitric oxide synthase

NO

nitric oxide

NOC-12

N-Ethyl-2-(1-ethyl-2-hydroxy-2-nitrosohydrazino) ethanamine

NT2

Ntera2

ODQ

1H-[1,2,4]-oxadiazolo[4,3-a]quinoxalin-1-one

PBS

phosphate-buffered saline

PFA

4% paraformaldehyde in PBS

PKA

protein kinase A

PKG

protein kinase G

Rp-cAMP

adenosine 3′,5′-cyclic monophosphorothioate, Rp-isomer

sGC

soluble guanylyl cyclase

SNP

sodium nitroprusside

The gaseous messenger nitric oxide (NO) which is produced by the enzyme nitric oxide synthase has been extensively studied as a retrograde messenger that modulates synaptic plasticity and memory formation (Boehning and Snyder 2003; Garthwaite 2008; Taqatqeh et al. 2009). The major physiological target of NO is the cylic guanosine-monophosphate (cGMP)-synthesizing enzyme soluble guanylyl cyclase (sGC). Numerous studies indicate that NO can induce transmitter release at pre-synaptic terminals (Hawkins et al. 1994; Arancio et al. 1996; Meffert et al. 1996; Sporns and Jenkinson 1997; Wildemann and Bicker 1999; Li et al. 2002; Nickels et al. 2007).

Mechanisms by which NO induces vesicle release are gradually being discovered. Enhancement of transmitter release via the cGMP pathway which certainly involves phosphorylation via protein kinase G (PKG) has been suggested as a major mechanism (Arancio et al. 1995, 2001; Lu et al. 1999; Wildemann and Bicker 1999; Li et al. 2004; Garthwaite 2008). Moreover, post-translational modification of ion channels and proteins involved in vesicle docking/fusion processes by S-nitrosylation has also been reported (Meffert et al. 1996; Ahern et al. 2002). The NO/cGMP pathway has been shown to regulate synaptic vesicle endocytosis and recycling by increasing pre-synaptic phosphatidylinositol 4, 5-bisphosphate (Micheva et al. 2003). In addition to its role as a retrograde messenger, NO has been implicated to play a critical role in early neuronal development including cell proliferation, migration, and synaptogenesis (Enikolopov et al. 1999; Bicker 2005, 2007). However, the limited availability of in vitro models for the developing human brain has restricted the analysis of potential roles for NO signals in vesicle turnover, a pathway that may also contribute to activity-dependent nerve cell development.

In this study, we used the well-characterized human teratocarcinoma cell line Ntera2 (NT2) that can be induced to differentiate into neurons upon treatment with retinoic acid as surface-attached adherent monolayer culture (Andrews 1984; Pleasure et al. 1992). In recent years, this rather lengthy differentiation method was significantly reduced by employing a cell aggregate culture method (Paquet-Durand et al. 2003; Podrygajlo et al. 2009). Despite their common clonal origin, differentiated NT2 neurons comprise a heterogeneous population of cells expressing several neurotransmitters in vitro (Guillemain et al. 2000; Podrygajlo et al. 2009). Electron microscopical and immunocytochemical studies revealed the presence of pre-synaptic vesicles and synaptic vesicle-associated proteins such as synaptobrevin, synaptophysin, and synapsin (Sheridan and Maltese 1998; Hartley et al. 1999; Guillemain et al. 2000; Podrygajlo et al. 2009). NT2 neurons also express voltage-gated Na channels (Matsuoka et al. 1997) and L, N, P/Q and R-type Ca channels that are recruited for neurotransmission (Gao et al. 1998; Neelands et al. 2000).

Here, we characterized in aggregate culture generated human NT2 neurons by monitoring luminal synaptotagmin I immunoreactivity and imaging of synaptic vesicle recycling with the fluorescent dye N-(3-triethylammoniumpropyl)-4-(4-(dibutylamino)styryl)pyridinium dibromide (FM1-43). Both methods revealed that depolarized NT2 neurons display calcium-dependent synaptic vesicle recycling. By employing immunocytochemical methods, we showed the expression of neuronal nitric oxide synthase (nNOS) and functional sGC in subpopulations of NT2 neurons. Finally, we used the human NT2 neurons as a model to demonstrate the involvement of NO/cGMP and cAMP/protein kinase A (PKA) signal transduction in vesicle recycling and pre-synaptic vesicle exocytosis.

Materials and methods

The NO donor N-Ethyl-2-(1-ethyl-2-hydroxy-2-nitrosohydrazino) ethanamine (NOC-12), the PKA antagonists N-[2-((p-Bromocinnamyl) amino) ethyl]-5-isoquinolinesulfonamide, 2HCl (H-89) and adenosine 3′, 5′-cyclic monophosphorothioate, Rp-isomer, triethylammonium salt were purchased from Calbiochem (Darmstadt, Germany). 8-bromoguanosine-3′, 5′-cyclic monophosphate, sodium salt and 8-bromoadenosine 3′, 5′-cyclic monophosphate, sodium salt (8-Br-cAMP) were purchased from Alexis Biochemicals (Lörrach, Germany). All other materials were obtained from Sigma (Taufkirchen, Germany) unless otherwise noted. Buffers with the following compositions were prepared: phosphate-buffered saline (PBS; 10 mM sodium phosphate, 150 mM NaCl, pH 7.4), Krebs-Ringer’s – HEPES (KRH) (in mM; 115 NaCl, 5 KCl, 1 MgCl2 × 6H2O, 24 NaHC03, 2.5 CaCl2.2H2O, 25 Glucose, 25 HEPES, pH 7.4), calcium free KRH that contains 2.5 mM EGTA instead of CaCl2 × 2H2O (Ca2+ free), KRH buffer that contains 60 mM of KCl, corrected for osmolarity by reduction of NaCl (high K+) and high K+ without CaCl2 × 2H2O (Ca2+ free high K+). All chemicals were diluted in KRH at the following final concentrations: sodium nitroprusside (SNP, 1 mM), NOC-12 (100 μM), 1H-[1, 2, 4]-oxadiazolo [4, 3-a] quinoxalin-1-one (ODQ, 50 μM), forskolin (50–100 μM), 8-bromoguanosine-3′, 5′-cyclic monophosphate, sodium salt (100–1000 μM), 8-Br-cAMP (100–1000 μM), adenosine 3′, 5′-cyclic monophosphorothioate, Rp-isomer (Rp-cAMP) (10 μM) and H-89 (10 μM). 3-Isobutyl-1-methylxanthine, ODQ, and forskolin were prepared from stocks in dimethylsulfoxide, which were further diluted in KRH to result in a maximum concentration of 0.25% dimethylsulfoxide. Pathway inhibitors were pre-incubated for 30 min with the NT2 neurons prior to the experiments while activators were applied for 5 min during loading of FM1-43.

Cell culture

The human NT2/D1 cell line was obtained from American Type Culture Collection, Manassas, VA, USA. NT2 precursor cells were maintained and cultivated in Dulbecco’s modified Eagle medium/F12 (Gibco-Invitrogen, Karlsruhe, Germany) supplemented with 10% fetal bovine serum (Gibco-Invitrogen) and 1% penicillin/streptomycin (Gibco-Invitrogen) in an atmosphere of 5% CO2 at 37°C (Andrews 1984). Post-mitotic NT2 neurons were obtained by free-floating spherical aggregate methods (Paquet-Durand et al. 2003; Podrygajlo et al. 2009). Briefly, NT2 precursor cells were seeded in 95 mm bacteriological grade Petri dishes (Greiner, Hamburg, Germany) for 24 h. Dulbecco’s modified Eagle medium/F12 medium containing 10% fetal bovine serum and 10 μM retinoic acid was used up to 1 week. The aggregates were mechanically dispersed and treated in RA (retinoic acid) for one additional week in T75 flasks as adherent culture. Finally, cells were trypsinized and seeded in T75 flasks and supplied with culture medium containing mitotic inhibitors (1 μM 1-6-d-arabinofuransylcytosine, 10 μM 2′-deoxy-5-fluorouridine, 10 μM 1-β-d-ribofuranosyluracil). After 7–10 days, fully differentiated neurons were selectively trypsinized and were plated on poly-d-lysine and Matrigel (Becton-Dickinson, Bedford, MA, USA) coated 12 or 25 mm cover glasses at a density of 20 000 or 100 000 cells/cover glass for further experimentation.

Immunofluorescence

Ntera2 neurons were washed with PBS and fixed for 30 min at 20°C with 4% paraformaldehyde in PBS (PFA). Immunofluorescence detection of microtubule associated protein 2 (MAP2) (1 : 1000; Sigma), Tau1 (1 : 500; Millipore International), synapsin (1 : 500; Synaptic Systems, Göttingen, Germany), synaptotagmin I directed to cytoplasmic domain (1 : 10, mAb30; Developmental Studies Hybridoma Bank, Iowa, USA), and nNOS (1 : 200; Sigma) was performed as previously described (Tegenge and Bicker 2009). For the detection of cGMP-immunoreactive (cGMP-IR) cells, NT2 neurons were pre-incubated for 20 min at 20°C with 1 mM SNP as an NO donor, 20 μM 3-(50-Hydroxymethyl-20-furyl)-1-benzyl indazole as an enhancer of NO-induced activity of sGC, and 1 mM 3-isobutyl-1-methylxanthine as a phosphodiesterase inhibitor in PBS. Cultures were washed once with PBS and fixed with 4% PFA for 30 min. The polyclonal sheep cGMP antiserum (1 : 10 000; a kind gift from Dr. J. de Vente, Maastricht University, the Netherlands) was used as primary antibody to detect the level of cGMP (Tegenge and Bicker 2009).

Luminal synaptotagmin I immunostaining

Ntera2 neurons were assayed for synaptic vesicle recycling using polyclonal antibody directed against the luminal domain of synaptotagmin I (Synaptic Systems). Cultures were incubated for 20 min with the antibody diluted 1 : 100 in high K+. For comparison, incorporation of anti-luminal synaptotagmin I was performed in normal KRH (basal), Ca2+ free high K+, and SNP or SNP + ODQ diluted in KRH. Cells were then fixed with 4% PFA and the incorporated antibody was detected by immunofluorescence.

Western blotting

The protein content of cell lysate was estimated by the BCATM protein assay kit (Pierce, Rockford, IL, USA). Equal amount of protein (50 μg) from NT2 cells and NT2 neurons was used. The monoclonal anti-nNOS (Sigma) at 1 : 1000 dilution and anti-acetylated-α-tubulin (Sigma) diluted 1 : 10 000 were used for western blotting as previously described (Tegenge and Bicker 2009).

FM1-43 imaging

The uptake and release of FM1-43 dye was performed as described by Gaffield and Betz (2006). Briefly, for uptake experiments, NT2 neurons cultured for 3–5 weeks were washed with PBS and stimulated for 5 min with high K+ buffer in the presence of 10 μM fixable analog of FM1-43 (FM1-43Fx; Molecular Probes, Eugene, OR, USA) and compared with the uptake of FM1-43 in normal KRH (basal), Ca2+ free high K+ and chemicals. After FM1-43Fx was loaded into pre-synaptic terminals, NT2 neurons were washed twice in KRH followed by additional twice washing in Ca 2+ free buffer. Neurons were then fixed with 4% PFA and images were taken immediately. The release of normal FM1-43 was followed in real time after loading the NT2 neurons with 10 μM FM1-43 (Molecular Probes) in high K+ buffer for 5 min. The culture was transferred into a perfusion chamber, mounted on a Zeiss Axiovert 200 microscope and maintained at 37°C. Cultures were continuously washed for 10 min in normal KRH in a perfusion chamber at flow rate of 2–2.5 mL/min. Cultures were washed for additional 5 min in Ca2+ free buffer. The unloading of FM1-43 was followed upon stimulation of the culture with high K+, Ca2+ free high K+, normal KRH, or chemicals diluted in normal KRH.

Microscopy and data analysis

Preparations were viewed with a Zeiss Axiovert 200 (Göttingen, Germany), equipped with a CoolSnap camera (Photometrics, Tucson, AZ, USA) and MetaMorph software (Molecular Devices, Sunnyvale, CA, USA). Confocal images of NT2 neurons loaded with FM1-43Fx and co-stained against cytoplasmic domain of synaptotagmin I were prepared using a Leica TCS-SP5 spectral laser scanning confocal microscope (Leica Mikrosysteme Vertrieb GmbH, Wetzlar, Germany) with LAS AF software and under 63× oil immersion objective. For quantification of synapsin and synaptotagmin levels, stained synaptic puncta were counted along the length of a neurite and expressed as number of synaptic puncta per number of neurons obtained from 4′,6-diamidino-2′-phenylindol-dihydrochloride staining. To obtain an estimate of cGMP levels, we counted cGMP-IR cell bodies. Data were presented as mean ± SEM of at least eight fields of view (220 × 165 μm) from at least three independent experiments. For FM1-43Fx uptake experiments, fluorescence intensity of individual puncta (numbers are given in the figure legends) was measured using Simple PCI software (C-Imaging Systems, Sewickley, PA, USA). Background fluorescence from a selected blank area was subtracted from individual puncta. For FM1-43 unloading experiments, fluorescence intensity was measured as above every 15 s for about 5 min. The percent fluorescence intensity remaining at the end of stimulation was calculated. Data were presented as mean ± SEM. Statistical analysis was performed using unpaired Student’s t-test. Levels of significance were: *< 0.05, **< 0.01, ***< 0.001.

Results

NT2 neurons undergo pre-synaptic maturation

We used the aggregate culture method (Paquet-Durand et al. 2003; Podrygajlo et al. 2009) to generate NT2 neurons expressing MAP2 and Tau, typical neuronal cytoskeletal markers of dendrites and axons, respectively. NT2 neurons cultured for 7 days in vitro (DIV) after differentiation displayed MAP2 staining on short processes originating from the cell somata (Fig. 1a). Only weak Tau-staining appeared on long-neuronal processes after 7 DIV (Fig. 1b). NT2 neurons cultured for 28 DIV were stained strongly for both MAP2 and Tau (Fig. 1e and f). To monitor the progress of pre-synaptic maturation, fully differentiated NT2 neurons were cultured in vitro for about 6 weeks and stained over time for the pre-synaptic proteins synapsin and synaptotagmin I. After 7 DIV culture, NT2 neurons expressed very little punctate synapsin staining that appeared mainly around cell somata (Fig. 1c). However, within 14–28 DIV intense punctate synapsin staining appeared on long neurites and around cell somata (Fig. 1g). Similarly, only little punctate synaptotagmin I staining was confined to the cell bodies of NT2 neurons cultured for 7 DIV (Fig. 1d). Intense synaptotagmin I staining appeared along the neurites within 14–28 DIV cultures (Fig. 1h). We quantified the levels of synapsin and synaptotagmin I as number of synaptic puncta per neuron. The level of both proteins increased significantly within 14–28 DIV culture as compared with 7 DIV (Fig. 2a) indicating that human NT2 neurons undergo pre-synaptic maturation. In NT2 neurons cultured for about 6 weeks, the level of synapsin and synaptotagmin did not increase compared with 28 DIV old cultures (data not shown).

Figure 1.

Human NT2 neurons express typical cytoskeletal neuronal marker and pre-synaptic proteins. The maturation of NT2 neurons generated by aggregate culture was followed by immunocytochemical staining for (a, f) MAP2, (b, g) Tau, (c, h) synapsin, (d, i) cytoplasmic, and (e, j) luminal synaptotagmin I. After 7 days in vitro (DIV), NT2 neurons were (a) intensely stained for the neuronal marker MAP2, (b) weakly stained for the axonal marker, Tau. (c, d) Only little synapsin and synaptotagmin puncta staining appeared on NT2 neurons cultured for 7 DIV. After 28 DIV culture, NT2 neurons were intensely stained for (f) MAP2 and (g) Tau. (h, i) Intense synapsin and synaptotagmin staining appeared along the neurites of NT2 neurons cultured for 28 DIV. (e, j) Pre-synaptic terminals of NT2 neurons cultured for 28 DIV were labeled with luminal synaptotagmin I compared with 7 DIV culture. Arrow heads (h, i and j) indicate representative punctate staining along the neurites. Blue (4′,6-diamidino-2′-phenylindol-dihydrochloride) indicates nuclear counter-staining. Scale bar: 200 μm (a, b, f and g), 100 μm (c, d, h and i), 50 μm (e and j). [Correction added after online publication 9 November 2009; panel labelling changed in figure legend].

Figure 2.

Human NT2 neurons undergo pre-synaptic maturation. The levels of synapsin and synaptotagmin I were quantified by counting the number of synaptic puncta along the neurites in at least eight fields of view taken from at least three independent experiments. (a) The number of puncta staining for synapsin and synaptotagmin were significantly increased within 14–28 DIV culture as compared with 7 DIV. (b) Synaptic vesicle recycling was determined by antibody directed to the luminal domain of synaptotagmin I during depolarization of human NT2 neurons by high K+. The extent of synaptic vesicle recycling increased significantly during the maturation of NT2 neurons within 14–28 DIV. (c) The incorporation of luminal synaptotagmin I depends on both depolarization and presence of calcium.

Depolarized human NT2 neurons display synaptic vesicle recycling in a calcium-dependent manner

Antibodies directed to the luminal domain of synaptotagmin I have been used to label pre-synaptic vesicles that undergo exocytosis (Matteoli et al. 1992; Kraszewski et al. 1995; Verderio et al. 2007). We applied the luminal synaptotagmin I antibody in the presence of a depolarizing high K+ (60 mM) which causes multiple round of synaptic vesicle recycling and hence incorporation of the antibody into the nerve terminals. Upon stimulation with high K+, NT2 neurons (28 DIV) were intensely labeled with synaptotagmin (Fig. 1j) indicating that the neurons perform synaptic vesicle recycling. The immunoreactivity appeared as punctate staining along the neurites and cell somata (Fig. 1j). We found on average higher values of immunoreactive synaptic puncta in NT2 neurons cultured for about 28 DIV than for 7 DIV (Figs 1i, j and 2b). Thus, similar to the expression of the pre-synaptic proteins, the activity-dependent immunostaining of the neurons critically depended on the length of in vitro culture. Quantification of the number of synaptic puncta revealed that depolarization by high K+ resulted in significantly higher values of luminal synaptotagmin labeling compared with the basal level (Fig. 2c). Application of anti-synaptotagmin antibody under high K+ stimulation in the absence of Ca2+ resulted in significantly lowered numbers of synaptic puncta (Fig. 2c). These results clearly demonstrate that depolarized NT2 neurons show Ca2+-dependent synaptic vesicle recycling.

To directly visualize synaptic exo- and compensatory endocytosis, we used fluorescent FM-dyes that are trapped in retrieved vesicle membranes in an activity-dependent manner. Firstly, we performed vesicle uptake experiments with the fixable dye FM1-43Fx as a monitor of synaptic vesicle turnover (Fig. 3a–c). Upon stimulation with high K+, NT2 neurons were successfully loaded with FM1-43Fx as compared with the basal level (Fig. 3a–c). Furthermore, the loading of FM1-43Fx was significantly lowered when it was applied with high K+ in the absence of Ca2+ (Fig. 3c). Then, we imaged exocytosis by monitoring the dimming of fluorescence when dye trapped in the synaptic vesicles is being released. The unloading of FM1-43 was followed in real time during second period of stimulation with high K+ after dye loading and washing. Upon stimulation, the NT2 neurons released the dye within 5 min (Fig. 3d–g) indicating the pre-synaptic vesicle exocytosis. The unloading of FM1-43 also depended on the presence of calcium in the stimulation buffer (Fig. 3g and h). To confirm that the FM1-43Fx loaded punctate staining indicates pre-synaptic structure, we performed co-staining of FM1-43Fx loaded neurons with the synaptotagmin I antibody (Fig. 4a–c). Even though the staining for FM1-43Fx weakened during permeabilization and washing of the culture, we found that most of the FM-labeled puncta were synaptotagmin I-immunoreactive (Fig. 4c).

Figure 3.

Human NT2 neurons display synaptic vesicle recycling and exocytosis. NT2 neurons were cultured for 28 DIV and labeled with the fixable analog of FM1-43 (FM1-43Fx) and normal FM1-43. (a, b) FM1-43Fx is loaded into recycling synaptic vesicle of NT2 neurons upon stimulation with high K+ compared to basal (KRH). Photomicrographs at the bottom of (a, b) represent the respective phase contrast images of NT2 neurons. (c) Quantification of FM1-43Fx uptake as fluorescent intensity of the basal level indicates significantly higher uptake of the dye into NT2 neuron terminals upon stimulation by high K+ as compared with KRH and Ca2+ free high K+. (d–h) The unloading of normal FM1-43 from NT2 neurons was followed upon stimulation with high K+ with or without Ca2+. (d) FM1-43 loaded punctate staining after washing with KRH (10 min) and calcium free buffer (5 min). (e, f) FM1-43 loaded terminals undergo destaining within 5 min upon stimulation with high K+. Arrow heads in (d–f) indicates representative puncta that undergo destaining. (g) A representative FM1-43 destaining curve upon stimulation by high K+ with or without calcium. (h) The percent fluorescence intensity remaining at the end of stimulation with high K+ was significantly lower than Ca2+ free high K+. Data represent mean ± SEM of at least 100 (c) and 20 (h) synaptic puncta from three independent experiments. Scale bars: 50 μm.

Figure 4.

FM1-43Fx loaded synaptic puncta co-localize with synaptotagmin I. NT2 neurons (21 DIV) were loaded with (a) FM1-43Fx, fixed and stained against (b) cytoplasmic domain of synaptotagmin I. (c) FM1-43Fx loaded synaptic puncta were colocalized with synaptotagmin I. Scale bar: 25 μm.

NO and cyclic GMP modulates synaptic vesicle recycling and exocytosis

The expression of nNOS by human NT2 neurons was detected by western blotting which resolved a major protein band around 155 kDa (Fig. 5a). This is in line with previous reports (Lee et al. 2001; Tegenge and Bicker 2009) for the expression of nNOS in the NT2 precursor cells. Subpopulations of NT2 neurons positively stained for the nNOS monoclonal antibody (Fig. 5b). The immunofluorescence labeling for nNOS appeared mainly around the cell somata and weakly along the neurites of NT2 neurons (Fig. 5b). To reveal the presence of the functional NO receptor sGC, we used an antibody against cGMP (De Vente et al. 1987). In the absence of an exogenous source of NO, on average 6.7 ± 1% cGMP-IR cells were detected (Fig. 5c). The number of cGMP-IR cells increased significantly up to 49.4 ± 4% upon stimulation with the NO donor SNP (Fig. 5d). The number of NO-induced cGMP-IR cells reduced to 23.6 ± 4% when SNP stimulation was accompanied by the sGC inhibitor, ODQ (Fig. 5e). Our data clearly demonstrate the expression of a NO-sensitive sGC in subpopulations of NT2 neurons.

Figure 5.

NT2 neurons express nNOS and functional sGC. (a) Western blotting of cell lysate from NT2 precursor cells and NT2 neurons. The antibody against nNOS recognizes a protein band of apparent molecular weight at 155 kDa and the lower band represents acetylated α-tubulin. (b) Immunofluorescence detection of nNOS in cell somata of NT2 neurons and around the neurites. (c–e) cGMP-IR of NT2 neurons for the detection of functional sGC. NT2 neurons cultured for 28 DIV was exposed for 20 min to SNP (1 mM) or SNP + ODQ (50 μM) together with 1 mM 3-isobutyl-1-methylxanthine (IBMX), as a phosphodiesterase inhibitor and 20 μM of 3-(50-Hydroxymethyl-20-furyl)-1-benzyl indazole (YC-1), as an enhancer of NO-induced activity of sGC. Controls were also treated with IBMX and YC-1. Cultures were fixed with 4% paraformaldehyde in PBS and stained against cGMP. (c) Under control conditions, NT2 neurons showed little cGMP-IR. (d) The addition of the NO donor SNP increased the cGMP-IR. (e) Application of SNP together with ODQ reduced NO-induced cGMP-IR. Blue in (c–e) indicates nuclear counter-staining (4′,6-diamidino-2′-phenylindol-dihydrochloride). Scale bar: 50 μm (b), 100 μm (c–e).

To investigate the involvement of the NO-cGMP signal transduction pathway in pre-synaptic neurotransmitter release, we exposed NT2 neurons to the NO donor SNP. This resulted in a significantly higher incorporation of the luminal synaptotagmin I antibody similar to depolarization-induced labeling by high K+ (Fig. 6a). The incorporation of anti-luminal synaptotagmin by SNP was blocked when it was applied together with the sGC inhibitor, ODQ (Fig. 6a). Likewise, the uptake of FM1-43Fx was enhanced by SNP (Fig. 6b) and a membrane-permeable analog of cGMP (Fig. 6c). Pre-incubation with ODQ resulted in blocking of SNP-induced uptake of the dye (Fig. 6b). These experimental results suggest that NO/cGMP modulates synaptic vesicle recycling in NT2 neurons.

Figure 6.

NO and cGMP analog facilitate synaptic vesicle recycling. (a) The NO donor, SNP (1 mM) significantly increased the incorporation of luminal synaptotagmin I into NT2 neuron that was blocked by the sGC inhibitor, ODQ (50 μM). (b) The uptake of FM1-43Fx was significantly increased by SNP (1 mM). Pre-incubation with sGC inhibitor (ODQ, 50 μM) blocked SNP-induced uptake of FM1-43Fx. (c) The cell membrane-permeable analog of cGMP, 8-bromoguanosine-3′, 5′-cyclic monophosphate, sodium salt enhanced the uptake of FM1-43Fx. Data represent mean ± SEM of percent fluorescence intensity of at least 80 synaptic puncta from three independent experiments.

The unloading of FM1-43 was followed in real time by monitoring the dimming of fluorescence intensity after stimulation with two NO donors, SNP, and NOC-12. Both NO donors induced the release of the dye with a destaining profile of FM1-43 similar to that seen in high K+ (Fig. 7a). At the end of 5 min lasting stimulation with SNP or NOC-12, the percentage of fluorescence remaining in NT2 nerve terminals was significantly lower than in normal KRH, indicating that NO-induced pre-synaptic vesicle exocytosis (Fig. 7b). After FM1-43 was loaded into the neurons, the culture was incubated with ODQ for 10 min before the unloading of FM1-43 was initiated with SNP or NOC-12. Pre-incubation with ODQ significantly blocked SNP (Fig. 7b) and NOC-12 (Fig. 7c) induced FM1-43 unloading. These results show that NO induces vesicle exocytosis in NT2 neurons via the cGMP pathway.

Figure 7.

NO induces pre-synaptic vesicle exocytosis via the cGMP pathway. (a) Representative destaining profiles of FM1-43 after stimulation by the NO donors SNP (1 mM) and N-Ethyl-2-(1-ethyl-2-hydroxy-2-nitrosohydrazino) ethanamine (NOC-12) (100 μM) in comparison with high K+. (b) The percent fluorescence intensity remaining after 5 min stimulation with SNP was comparable with 60 mM KCl. Pre-incubation with ODQ (50 μM) for 10 min and stimulation with SNP + ODQ for 5 min resulted in blocking of SNP-induced FM1-43 unloading. (c) NOC-12 significantly induced unloading of FM1-43 which was blocked by ODQ. Data represent mean ± SEM of percent fluorescence intensity remaining after 5 min stimulation from at least 15 synaptic puncta.

cAMP/PKA signal transduction modulates synaptic vesicle recycling and exocytosis

To reveal a possible involvement of the cAMP signaling pathway, we used activators and inhibitors of this cascade in FM1-43 uptake experiments. The uptake of FM1-43Fx was significantly enhanced in the presence of the adenylyl cyclase stimulator forskolin (Fig. 8a) and a cell membrane permeable analog of cAMP (8-Br-cAMP) (Fig. 8b). Pre-incubation of NT2 neurons with PKA inhibitors (Rp-cAMP or H-89) significantly reduced forskolin-induced (Fig. 8c) and high K+-induced (data not shown) uptake of FM1-43Fx indicating that cAMP/PKA pathway positively regulates synaptic vesicle recycling. Additionally, we followed the release of FM1-43 upon stimulation with forskolin. Stimulation of NT2 neurons with forskolin resulted in rapid release of FM1-43 which indicates that forskolin induced vesicle exocytosis (Fig. 8d). Pre-incubation with H-89 significantly reduced the release of FM1-43 indicating the participation of PKA in forskolin induced vesicle exocystosis (Fig. 8d).

Figure 8.

The cAMP/PKA pathway modulates synaptic vesicle recycling in NT2 neurons. The uptake of FM1-43Fx was significantly increased in the presence of (a) forskolin and (b) cAMP analog (8-Br-cAMP). (c) Pre-incubation of NT2 neurons for 30 min with protein kinase A antagonist (10 μM RPcAMP or H-89) significantly reduced forskolin induced FM1-43Fx uptake. (d) Real time imaging of vesicle exocytosis upon stimulation with forskolin (100 μM) indicates fast release of FM1-43 and the percent fluorescence intensity remaining after 5 min was comparable with that of high K+. Pre-incubation with H-89 (10 μM) for 10 min and stimulation with forskolin (100 μM) resulted in marked reduction of the release of FM1-43. Data represent mean ± SEM at least 80 (a–c) and 10 (d) synaptic puncta.

Discussion

Human NT2 neurons display synaptic vesicle recycling and pre-synaptic release

Recently, we have shown that the human NT2 cells proliferate during retinoic acid treatment as spherical aggregate culture and cells migrate out of the aggregate to acquire fully differentiated neuronal phenotypes (Tegenge and Bicker 2009). In this study, we followed the progress of NT2 neuronal maturation after the migratory phase when neuronal networks are formed. After 7 DIV, intense MAP2 staining indicated full differentiation of NT2 neurons but staining for phosphorylated Tau, indicative of axonogenesis, was weak (Fig. 1a and b). However, after 28 DIV, intense Tau-staining appeared on long neurites, whereas MAP2 staining did not change in the cell cultures (Fig. 1e and f).

Immunocytochemical evidence for pre-synaptic maturation of NT2 neurons was monitored by staining for the pre-synaptic proteins, synapsin I, and synaptotagmin I. Synapsin I is a member of the synapsin family specifically associated with synaptic vesicles which has been implicated in the synapse development and regulation of neurotransmitter release (reviewed by de Camilli et al. 1990; Hilfiker et al. 1999; Ferreira and Rapoport 2002). Synaptotagmin I is the best characterized isoform of the synaptotagmin family of vesicle-associated proteins, thought to function as a calcium sensor for fast neurotransmitter release at the synapse (reviewed by Yoshihara and Montana 2004; Chapman 2008).

Immunofluorescence staining and quantification of synaptic puncta showed that the level of synapsin and synaptotagmin I increased significantly with the length of in vitro culture. Synapsin immunoreactivity became more intense and appeared in the neuronal process within 14–28 DIV (Fig. 1g). Previously, for NT2 neurons generated by the classical methods, intense immunoreactivity to synapsin was observed upon co-culture with rat astrocytes (Hartley et al. 1999). Transcriptional up-regulation of synapsins during retinoic acid-induced differentiation of NT2 cells has been reported (Leypoldt et al. 2002). The immunostaining of synaptotagmin I that appeared as punctate staining (Fig. 1h) could indicate local accumulation of the protein at presumptive pre-synaptic sites for participation in fast neurotransmitter release. The expression of other pre-synaptic proteins such as synaptobrevin, synaptophysin, and SNAP25 has also been reported for NT2 neurons (Sheridan and Maltese 1998).

However, it remains unclear whether these pre-synaptic proteins expressing immunopositive puncta represent functional synaptic terminals. In this study, we showed for the first time that human NT2 neurons display synaptic vesicle recycling and exocytosis by two independent approaches. Firstly, we used functional immunofluorescence methods to label pre-synaptic terminals of NT2 neurons during synaptic vesicle recycling. For this purpose, an antibody directed to the luminal domain of synaptotagmin I was used in the presence of high K+. It is well known that high K+ induces multiple rounds of exocytosis leading to labeling of the entire pool of recycling synaptic vesicles (Klingauf et al. 1998; Sara et al. 2002; Menegon et al. 2006). In NT2 neurons, the labeling of synaptic vesicles by anti-luminal synaptotagmin I depends on depolarization induced by high K+ and presence of calcium in the stimulation buffer (Fig. 2c). Our data also show that the extent of synaptic vesicle recycling depends on the length of in vitro culture corresponding to the expression of the pre-synaptic proteins (Figs. 1i, j and 2b).

In the second approach, synaptic vesicle recycling and exocytosis was analyzed by employing the fluorescent dye, FM1-43. The protocol for labeling pre-synaptic vesicles that display synaptic vesicle recycling by FM1-43 is well established (Cochilla et al. 1999; Gaffield and Betz 2006). The uptake of FM1-43Fx was significantly increased upon stimulation with high K+ in the presence of calcium. The little punctate staining that appeared upon stimulation with KRH (basal) during the incorporation of luminal synaptotagmin I antibody (Fig. 2c) and FM1-43Fx (Fig. 3a) could indicate spontaneous activity of the neurons. Intriguingly, confocal images of NT2 neurons loaded with FM1-43Fx and double-labeled for synaptotagmin I display clear co-localization of the punctate staining (Fig. 4a–c). Thus, both methods indicate that human NT2 neurons undergo synaptic vesicle recycling in a depolarization and calcium-dependent manner.

Unloading of FM1-43 which monitors pre-synaptic vesicle exocytosis revealed rapid release of the dye in a both depolarization and calcium-dependent manner (Fig. 3d–h). This pre-synaptic vesicle exocytosis induced by high K+ presumably indicates neurotransmitter release at nerve terminals. Recently, we have shown that NT2 neurons express markers for glutamate, GABA, and for cholinergic transmission (Podrygajlo et al. 2009) suggesting that these neurotransmitters are putatively released by high K+. At the post-synaptic site, expression of functional receptors including GABAA-receptors (Matsuoka et al. 1997; Neelands et al. 1998, 1999;.), NMDA-type (Rootwelt et al. 1998; Paquet-Durand and Bicker 2004;Paquet-Durand et al. 2006; Garcia de Arriba et al. 2006), and α-amino-3-hydroxy-5-methylisoxazole-4-propionate (AMPA)-type glutamate receptors (Rootwelt et al. 1998) have been reported. Nevertheless, functional synapses between NT2 neurons have been demonstrated electrophysiologically only when cultured together with rat astrocytes (Hartley et al. 1999). Spontaneous but uncorrelated firing patterns of NT2 neurons have been recorded on microelectrode arrays (Görtz et al. 2004). Here, we showed pre-synaptic vesicle release in NT2 neurons in the absence of glia cells. Since we could demonstrate that the neurons display increasing levels of an axonal marker (Tau), pre-synaptic proteins (synapsin and synaptotagmin I), and synaptic vesicle recycling with time in culture, NT2 neurons seem to undergo pre-synaptic maturation processes. Therefore, NT2 neurons can be utilized as a model to investigate the cellular and molecular basis of synaptic vesicle recycling and pre-synaptic vesicle release in human nerve cells.

NO signal transduction and pre-synaptic vesicle release

Nitric oxide synthase has been shown to be expressed during neural development and synaptogenesis (Williams et al. 1994; Ogilvie et al. 1995; Sporns and Jenkinson 1997; Gibbs and Truman 1998) indicating that NO is involved in synaptic maturation processes. This study demonstrates for the first time that NO and cyclic nucleotides regulate vesicle recycling and pre-synaptic vesicle release in human model neurons. Subpopulations of NT2 neurons positively stained for the nNOS monoclonal antibody (Fig. 5b). Thus, similar to other neurotransmitter phenotypes (Guillemain et al. 2000; Podrygajlo et al. 2009), the NT2 neurons are also heterogeneous with respect to nNOS expression. These neurons expressing nNOS could serve as endogenous source of NO. The presence of functional sGC was demonstrated by exogenous stimulation of the neurons with NO donor and subsequent detection of cGMP by immunofluorescence methods (Fig. 5c–e). NO induced cGMP-IR was reduced in the presence of a sGC inhibitor.

The NO donor, SNP significantly increased the labeling of NT2 neurons by luminal synaptotagmin I which was blocked by the sGC inhibitor ODQ indicating that NO facilitates synaptic vesicle recycling via the cGMP pathway. SNP enhanced the uptake of FM1-43Fx similar to stimulation with high K+. The effect of exogenous NO application was blocked by ODQ. Moreover, a membrane permeable analog of cGMP significantly increased the uptake of FM1-43Fx which further confirms the participation of cGMP pathway in NO-mediated synaptic vesicle recycling. These experimental results suggest that the NO/cGMP signaling pathway could facilitate synaptic vesicle recycling in human neurons. A recent study demonstrates that cGMP reduces the cycle time for synaptic vesicles through the enhancement of vesicular traffic rate from the recycling pool to the readily releasable pool and accelerates fast endocytosis (Petrov et al. 2008). Pre-synaptic exocytosis was followed in real time during stimulation of NT2 neurons by two NO donors, SNP and NOC-12. Both SNP and NOC-12 induced the unloading of FM1-43 similar to high K+ (Fig. 7a–c). The unloading of FM1-43 by SNP and NOC-12 was blocked by ODQ which indicates that NO induces vesicle release via the cGMP pathway. Our data are in agreement with several reports that demonstrated increased pre-synaptic transmitter release by the NO/cGMP pathway (Arancio et al. 1995, 2001; Lu et al. 1999; Wildemann and Bicker 1999; Li et al. 2004; Nickels et al. 2007). To analyze whether the NO/cGMP induced enhancement of vesicle exocytosis is caused by excitatory effects at the level of the membrane or by directly facilitating the calcium-dependent release machinery, electrophysiological techniques will be required. Moreover, in hippocampal neurons and synaptosomes, NO has been shown to induce neurotransmitter release independent of intracellular calcium levels (Meffert et al. 1994; Sporns and Jenkinson 1997).

One potential intracellular downstream effector protein for the cGMP signaling pathway is PKG. Enhanced transmitter release via phosphorylation of pre-synaptic voltage-gated K+ channels by PKG has recently been implicated (Yang et al. 2007). Additionally, formation of new cluster of pre- and post-synaptic proteins which involves phosporylation of PKG substrate proteins, vasodilator-stimulated phosphoprotein, and RhoA have been reported (Wang et al. 2005). A direct action of NO without the involvement of cGMP and PKG via S-nitrosylation on the exo–endocytotic machineries has been also implicated (Meffert et al. 1996; Ahern et al. 2002). Even though our data from both immunofluorescence and FM imaging strongly support the participation of the cGMP pathway, we could not completely exclude the direct action of NO since the sGC inhibitor, ODQ did only partially block the unloading of FM1-43 induced by NOC-12 (Fig. 7c).

Pre-synaptic activation of calcium-sensitive adenylyl cyclase that leads to a rise in the level of cAMP and consequent activation of PKA has been implicated to enhance the probability of neurotransmitter release (Siegelbaum et al. 1982; Chavez-Noriega and Stevens 1994; Trudeau et al. 1996; Evans and Morgan 2003). The proteins involved in pre-synaptic vesicle exocytosis have been suggested as major substrates for PKA-dependent phosphorylation (reviewed by Leenders and Sheng 2005). Phosphorylation of synapsin by the cAMP/PKA signaling which leads to release of vesicles from the reserve pool has been suggested to enhance pre-synaptic vesicle releases (Fiumara et al. 2004, 2007; Bonanomi et al. 2005; Menegon et al. 2006).

Here, we showed that stimulation of adenylyl cyclase by forskolin induces vesicle exocytosis and facilitates vesicle recycling while inhibition of PKA by Rp-cAMP or H-89 blocks the effect of forskolin indicating that the cAMP/PKA pathway modulates vesicle release and synaptic vesicle recycling in NT2 neurons. Our result is in line with the increased neurotransmitter release reported for hippocampal neurons upon activation of cAMP/PKA signaling (Trudeau et al. 1996, 1998). However, we do not know at present which target proteins are modified by NO and cyclic nucleotide signaling in NT2 neurons. In our readily accessible human model neurons, we can now investigate the candidate proteins modulated via NO and cyclic nucleotide signal transduction.

In summary, we have presented a model of human neurons that express increasing levels of pre-synaptic proteins and display synaptic vesicle recycling. Our data also revealed for the first time that the NO and cyclic nucleotide signal transduction modulates synaptic vesicle recycling and exocytosis in human model neurons.

Acknowledgements

We thank Dr J. de Vente for his kind gift of the cGMP antiserum, Dr S. Knipp for the discussion during the preparation of the manuscript and S. Tan for technical support. This research was in part supported by BMBF grant 0313732 to G. Bicker and by a DFG grant (FG 1103, BI 262/16-1). M. A. Tegenge received a Georg-Christoph-Lichtenberg scholarship from the Ministry for Science and Culture of Lower Saxony.

Ancillary