Synaptogenesis by Cholinergic Stimulation of Astrocytes

Astrocytes release numerous factors known to contribute to the process of synaptogenesis, yet knowledge about the signals that control their release is limited. We hypothesized that neuron-derived signals stimulate astrocytes, which respond to neurons through the modulation of astrocyte-released synaptogenic factors. Here we investigate the effect of cholinergic stimulation of astrocytes on synaptogenesis in co-cultured neurons. Using a culture system where primary rat astrocytes and primary rat neurons are first grown separately allowed us to independently manipulate astrocyte cholinergic signaling. Subsequent co-culture of pre-stimulated astrocytes with naïve neurons enabled us to assess how prior stimulation of astrocyte acetylcholine receptors uniquely modulates neuronal synapse formation. Pre-treatment of astrocytes with the acetylcholine receptor agonist carbachol increased the expression of synaptic proteins, the number of pre- and postsynaptic puncta, and the number of functional synapses in hippocampal neurons after 24 h in co-culture. Astrocyte secretion of the synaptogenic protein thrombospondin-1 increased after cholinergic stimulation and inhibition of the receptor for thrombospondins prevented the increase in neuronal synaptic structures. Thus, we identified a novel mechanism of neuron-astrocyte-neuron communication, where neuronal release of acetylcholine stimulates astrocytes to release synaptogenic proteins leading to increased synaptogenesis in neurons. This study provides new insights into the role of neurotransmitter receptors in developing astrocytes and into our understanding of the modulation of astrocyte-induced synaptogenesis.

The fact that astrocytes are integral to neuronal development and that they express ACh receptors (AChRs) and respond to synaptic release of ACh suggests that stimulation of astrocyte AChRs may promote neuronal development by modulating the expression and release of factors that create an environment conducive to neuronal development. Here we test the hypothesis that cholinergic stimulation of astrocytes increases synaptogenesis through the release of factors that potentiate the formation of synaptic structures in co-cultured neurons. Our results show that pre-treatment of astrocyte cultures with a cholinergic agonist leads to enhanced expression of neuronal synaptic proteins, an increase in the number of pre-and postsynaptic puncta, and an increase in the number of structural and functional synapses in cocultured hippocampal neurons. This enhanced synaptogenic function of astrocytes is prevented by the inhibition of TSP receptors in the co-culture system and mimicked by exogenous TSP1, an extracellular matrix protein whose release is increased after cholinergic stimulation of astrocytes. The elucidation of the signals that regulate changes in astrocytesecreted proteins involved in synaptogenesis is a key step toward the understanding of normal and pathological neurodevelopmental processes.

Animals
Time-pregnant Sprague-Dawley rats were purchased from Charles River (Wilmington, MA) and singly housed. All animals were housed in the animal facility of the Department of Environmental and Occupational Health Science at the University of Washington. All animal procedures were performed in accordance with the National Institute of Health Guidelines for the Care and Use of Laboratory Animals and as described in an animal protocol approved by the University of Washington Institutional Animal Care & Use committee (IACUC).

Hippocampal Neuron Preparation and Culture
Primary hippocampal neurons from E21 Sprague Dawley rats were prepared as previously described [48,49]. Briefly, cells were plated in poly l-ornithine hydrobromide (PLO,15 µg/mL) pre-coated 6-well plates (1.5 × 10 6 /well) for protein expression experiments, or on PLO pre-coated glass coverslips (12 mm circles) with manually adhered paraffin spacers for co-culture experiments (0.08 × 10 6 neurons/coverslip) placed into 24 well plates (1 coverslip/well). Neurons were maintained in Neurobasal-A medium supplemented with B-27 neuronal survival and growth factors (1%), 3 mM GlutaMax, 30 mM D-(+) glucose solution, 0.5% fungizone, and 100 µg/mL gentamicin and maintained at 37˚C for 12-13 days prior to co-culture. To hinder astrocyte proliferation, ARAC was added at a final concentration of 2.5 µM to each well after three days in culture. One-third of the medium was changed every three to four days post ARAC treatment. On day in culture (DIC) 12-13, neurons were either co-cultured with primary astrocytes or treated with human TSP1 for 24 h (see below).

Astrocyte Treatments and Neuronal Co-Culture
Primary cortical astrocytes were trypsinized from established cultures and sub-cultured in 24 well plates (250,000 cells/mL) for synaptogenesis and electrophysiology experiments, or on the underside of 0.4 μm mesh 6-well plate inserts for protein expression experiments. Wells and inserts were coated with 40 µg/mL PDL. Forty-eight hours prior to co-culture with neurons, astrocytes were serum deprived (DMEM + 0.1% bovine serum albumin and P/S) for 24 h, and then treated with or without carbachol (0, 0.01, 0.10, 1 mM) for an additional 24 h. Astrocytes were washed with PBS and incubated in serum-free medium for three hours prior to co-culture with neurons. In some experiments, 30 min prior to the addition of carbachol, astrocytes were treated with 10 µM of the acetylcholine receptor antagonists, mecamylamine, gallamine, or 4-DAMP. After treatment washout and medium conditioning, primary hippocampal neurons (12)(13), grown on glass coverslips were inverted over the pre-treated astrocyte monolayer for immunocytochemistry or electrophysiology experiments. Neurons were never in direct contact with astrocytes, nor were they exposed to astrocyte treatments. For Western blot experiments, astrocytes plated on the underside of porous inserts and their medium were transferred to 6-well plates containing primary hippocampal neurons.
To block TSP1 signaling in neurons, after astrocyte carbachol pre-treatment and washout, neurons and astrocytes were pre-treated for 30 min with gabapentin (15 or 30 µM) prior to co-culturing; gabapentin remained throughout the co-culture incubation. For all co-culture experiments, astrocytes and neurons were co-incubated for 24 h.

Immunocytochemistry, Image Acquisition and Analysis
Neurons were fixed in 4% PFA for 20 min at 37 °C, blocked and permeabilized for 30 min (0.1% Triton-X, 5% bovine serum albumin in PBS) and co-labeled with primary antibodies against synaptophysin (1:250) and PSD-95 (1:200), or synaptotagmin (1:250) and NR2B (1:200) for at least 18 h at 4 °C under slow rocking. Coverslips were then incubated for 1 h at room temperature with fluorescent secondary donkey anti-rabbit Alexa 488, and donkey anti-mouse Alexa 555 antibodies (1:500) and the nuclear dye Hoechst 33,342 (1 µg/mL). Coverslips were mounted on glass slides using Vectashield, topped with a cover glass, and sealed with nail polish. Individual, healthy neurons, identified with Hoechst stain, and located at least two cell bodies from their nearest neighbor, were imaged using confocal microscopy (Olympus Fluoview-1000) Confocal images were acquired using a 60X oil immersion objective, producing a 1024 × 1024 image size with a 2X optical zoom, yielding a 0.103 μm/ pixel resolution. 12-18 images per neuron were acquired through the z-plane, using a step size of 0.30 μm from 5 neurons per coverslip. To capture the entire range of fluorescent signal, the slide from the treatment group with the greatest intensity was used to set the imaging parameters for each channel and the settings remained constant during acquisition of images from each treatment groups within an experiment. Images were deconvolved using Huygens Professional software (Scientific Volume Imaging). For object analysis, the surface of synaptic puncta was rendered three-dimensionally based on a size and intensity threshold which was determined by obtaining the mean intensity of 30 manually selected, size appropriate puncta from 15 neuronal fields of each channel (350-450 total) using Image J, from the treatment group used to set confocal imaging parameters. Channel thresholds were held constant for all treatment groups within an experiment. The number of individual pre and postsynaptic objects and those overlapping between channels were automatically calculated and recorded using Huygens Object Analysis software.
For immunocytochemical labeling of TSP1 on the astrocyte surface, astrocytes were plated on PDL-coated glass coverslips at a density of 250,000 per well, in 24 well plates. Control and carbachol-treated astrocytes were incubated for 24-hours, washed in PBS, and fixed in 4% PFA for 20 min at 36 °C, but not permeabilized in order to label only cellsurface TSP1. Astrocytes were then blocked in 10% goat serum and incubated in anti-mouse TSP1 primary antibody (1:80 in PBS) for 3 h at room temperature, followed by donkey anti-mouse Alexa 488 (1:75, PBS, 3% BSA) secondary antibody and the nuclei stained with Hoechst 33,342 (1:2000 in PBS). Cells were mounted on slides using Vectashield and imaged using confocal imaging with parameters held constant between treatment groups. Confocal images were acquired using a 40X oil immersion lens, at a step-size of 0.50 μm for a total of 18-21 planes per field. Nine to ten fields were imaged per treatment group from two coverslips. For each field, the integrated optical density per plane was determined using Metamorph software (Molecular Devices), totaled, normalized to cell number and averaged for each treatment group. For presentation of using Pierce concentrators (MWCO 9 kDa) centrifuged at 4000 x g for 25 min at 25°C. After the addition of protease inhibitors, reducing reagents, and sample buffer, samples were denatured by heating (70 °C for 10 min) and equal volumes (30 µl) of concentrated sample were loaded into 3-8% Tris-Acetate pre-cast gels. After transfer, PDVF membranes were blocked in TBST (3% BSA) for 1-3 h and incubated overnight in goat anti-mouse TSP1 (1:500). Band densitometry was determined using Image J software and normalized to corresponding cell lysate protein concentrations.

qRT-PCR Analysis of Thbs1 Expression
The effect of carbachol on Thbs1 gene (encoding for TSP1) expression was quantified by qRT-PCR as previously described [50]. Primers for rat Thbs1 (Forward: AAGAC-GTCGACGAGTGCAAA; Reverse: CAGGCAGTTG-TAGCCAGGAT) were used with 5 ng of RNA and the Luna Universal One-Step RT-qPCR Kit (NEB) with SYBR Green detection on the CFX96 Real-Time System (Bio-Rad). Cycle-threshold data were normalized to total RNA using RiboGreen (ThermoFisher) and expressed as fold/ time-matched control with significance determined by Student's t-test.

Electrophysiology
Spontaneous miniature excitatory postsynaptic currents (mEPSCs) were recorded at room temperature (21.5-23.5˚C) from hippocampal neurons after co-culture with carbachol (1 mM) pre-treated astrocytes, or untreated astrocyte controls using whole cell patch clamp techniques. Currents were recorded using a MultiClamp 700B by Axon Instruments; cells were voltage clamped at a holding potential of -70 mV. Patch pipettes were pulled from borosilicate capillary glass (2.4-5.9 MΩ). Dissociated hippocampal neurons were bathed in artificial cerebral spinal fluid (ACSF) (119 mM NaCl, 2.5 mM KCl, 4 mM CaCl2, 4 mM MgCl2, 26 mM NaHCO3, 1 mM NaH2PO4, 11 mM glucose at pH of 7.4 and gassed with 5% CO2 and 95% O2). Tetrodotoxin (1 μm) was added to the circulating bath to isolate spontaneous miniature postsynaptic currents. Internal solution contained: 115 mM CsMeSO 4 , 20 mM CsCl, 2.5 mM MgCl 2 , 10 mM HEPES, 4 mM Na 2 ATP, 0.4 mM Na 3 GTP, 10 mM Na-phosphocreatine, and-0.6 mM EGTA, pH 7.25 (CsOH). Activity was recorded for up to ten minutes and events were manually selected using Synaptosoft Mini Analysis software. Traces from individual neurons of each treatment group were combined into arrays and the inter-event interval (IEI), amplitude, and time to decay, cumulative distributions were plotted (7-9 neurons per treatment group array from 3 independent experiments). The frequency of events, a representative image, the maximum intensity Z-projection was obtained from confocal images using Image J, and channels were merged. Negative control coverslips containing astrocytes at the same density as the experimental coverslips were fixed and incubated in the presence of the secondary antibody alone (donkey anti-mouse Alexa 488 at the dilution of 1:75 in PBS supplemented with 3% BSA); images of these coverslips were taken using the same imaging parameters used for experimental coverslips and did not show any specific immunostaining (not shown).

Western Blot
After 24 h co-culture with carbachol-treated or control astrocytes, neurons were lysed in 1% SDS lysis buffer, sonicated twice at 3.5 power for 5 s and the protein quantified using the bicinchoninic acid assay (BCA). After the addition of protease inhibitors, reducing reagents, and sample buffer, samples were denatured by heating (70 °C for 10 min), and equal amounts of proteins (30 µg proteins/lane) were loaded into freshly cast bis-trisacrylamide gels (10%). Proteins were separated by gel electrophoresis and transferred to PVDF membranes. Membranes were blocked in TBST (5% milk) for 1 h, then probed overnight at 4˚C with primary antibodies for synaptophysin (1:500), or PSD-95 (Neuromab, 1:500). Membranes were incubated for 1 h with HRPconjugated secondary antibodies (1:1000) and developed. Band densitometry was determined using Image J software, and results normalized to total protein as determined by Coomassie blue protein stain.
For measurement of astrocyte TSP1 levels, astrocytes were plated on PDL (40 µg/mL) pre-coated 100 mm plates at a density of 2.5 × 10 6 cells per plate and cultured for 4 days, serum-deprived for 24 h, then treated for 24 h with carbachol (1 mM) prepared in serum free medium. Medium was collected and cells were lysed in 1% SDS lysis buffer from treated and untreated plates immediately after the 24 h treatment (time 0), and 6 and 24 h after treatment washout. Intracellular protein was quantified from lysate samples using the BCA method. After the addition of protease inhibitors, reducing reagents, and sample buffer, samples were denatured by heating (70 °C for 10 min) and equal amounts of proteins (30 µg proteins/lane) were loaded into 3-8% Tris-Acetate gels and separated by gel electrophoresis. After transfer, PDVF membranes were blocked in TBST (3% BSA) for 1 h and probed overnight in goat anti-mouse TSP1 (1:1000), then incubated for 1 h with HRP-conjugated secondary antibodies (1:1000) and developed. TSP1 was detected by chemiluminescence and normalized to β-actin levels using densitometric analysis. For measurements of TSP1 in the astrocytic medium, 7 mL of media from each time point was concentrated to 200 µl the overlapping puncta shown in yellow. Pre-treatment of astrocytes with carbachol increased the number of individual presynaptic synaptophysin (Fig. 1B) and postsynaptic PSD-95 (Fig. 1C) puncta in neurons in a dose-dependent manner. An almost two-fold increase in synaptophysin puncta and a 2.2-fold increase in PSD-95 puncta were observed in neurons after co-culture with astrocytes pre-treated with 0.1 mM carbachol, when compared to control astrocytes. Astrocyte pre-treatment with 1 mM carbachol induced a 1.6-and 2.3-fold increase in the number of individual synaptophysin and PSD-95 puncta, respectively (Fig. 1B, C). Because synapses are structures consisting of aligned pre and postsynaptic specializations, we used the number of overlapping synaptophysin and PSD-95 puncta as a measure of the number of structural synapses in each treatment group. Stimulation of astrocytes with carbachol induced the formation of a greater number of synapses than those induced by untreated astrocytes. Pre-treatment of astrocytes with 1 mM carbachol caused a 3.2-fold increase in the number of aligned pre and postsynaptic puncta (structural synapses) above the number of synapses observed when neurons were co-cultured with unstimulated astrocytes (Fig. 1D). Together, these results indicate that astrocyte-induced hippocampal synaptic structures are increased when neurons are co-cultured with astrocytes previously stimulated with the cholinergic agonist carbachol.
To determine the cholinergic receptor subtype(s) involved in carbachol-stimulated astrocyte-induced synaptogenesis, astrocyte cultures were pre-treated with carbachol (1 mM) in the presence of muscarinic and nicotinic AChR antagonists for 24 h before co-culturing the astrocytes with hippocampal neurons. As rat cortical astrocytes do not express M1 or M4 muscarinic receptors [39], we investigated the role of M2 and M3 muscarinic receptors. Representative deconvolved images of hippocampal neurons after co-culture for 24 h with astrocytes pretreated with 1 mM carbachol in the presence or absence of cholinergic antagonist or control astrocytes are shown in Fig. 2A. Inhibiting the M2 muscarinic AChR with gallamine (10 µM) had no effect on carbachol-induced increase in synapse number, while the M3 muscarinic antagonist 4-DAMP reduced the increase in the number of structural synapses observed; the effect of carbachol-treated astrocytes on synapse formation was also inhibited by the nicotinic antagonist mecamylamine (10 µM) (Fig. 2B). These data suggest that cholinergic signaling through both astrocytic M3 muscarinic and nicotinic receptors were responsible for the observed increase in hippocampal neuron synapse number.
Nest, we examined the total expression of the pre-and post-synaptic proteins synaptophysin and PSD-95 using Western blot. Neurons co-cultured with astrocytes pretreated with 1 mM carbachol showed a robust increase in as measured by inter-event-intervals, was used to represent the relative number of functional synapses.

Statistical Analysis
Student's t-test was used for statistical comparison of the means in experiments containing two treatment groups and One-way ANOVA with Dunnett's Multiple Comparison test was used where multiple concentrations were assessed relative to control. To compare multiple treatments, oneway ANOVA followed by Bonferroni's Multiple Comparison tests was performed. Electrophysiology analysis was performed using Synaptosoft Mini Analysis software and Graph Pad Prism. Kolmogorov-Smirnov Two Sample Analysis was used to compare the cumulative distributions from arrayed electrophysiology recordings. In figures where individual data points are presented using dot plots, the means and standard error of the means (SEM) are shown. For experiments with a high number of data points, the distribution of the data, the median and the 25th and 75th percentiles are presented using violin plots and significant differences of the means are indicated.

Cholinergic-Stimulation of Astrocytes Enhances Astrocyte-Induced Synapse Formation
To test whether cholinergic stimulation of astrocytes affects synapse formation we treated astrocyte cultures for 24 h with increasing concentrations of the acetylcholine receptor (AChR) agonist carbachol (0.01, 0.1, or 1 mM). At treatment end, carbachol was removed and fresh treatment-free medium was added. Primary hippocampal neurons, grown separately on glass coverslips, were then inverted over the pre-treated astrocytes for an additional 24 h. In this co-culture system neurons were neither directly exposed to carbachol nor were they in contact with the astrocytes, limiting any effect on neuronal synapse formation to the modulation of astrocyte-secreted factors.
To quantify the effect of astrocyte cholinergic stimulation on the localization of synaptic proteins into excitatory synaptic structures, we generated three-dimensional surface renderings of the clustered synaptophysin and PSD95 protein puncta from deconvolved confocal images and the number of pre and postsynaptic specializations were automatically counted for each neuronal field. Representative deconvolved images of hippocampal neurons after co-culture for 24 h with astrocytes pretreated with 1 mM carbachol or control astrocytes are shown in Fig. 1A with synaptophysin shown in green, PSD-95 shown in red, and mM carbachol or control astrocytes are shown in Fig. 4A with synaptotagmin shown in green, NR2B shown in red, and the overlapping puncta shown in yellow. We observed a 2.3-fold increase in synaptotagmin puncta (Fig. 4B), a 2.8-fold increase in the NR2B puncta (Fig. 4C), and a 4.6fold increase in synaptotagmin/NR2B overlapping puncta (Fig. 4D) in neurons co-cultured with carbachol-pre-treated astrocytes, compared to untreated astrocytes. Together these data indicates that cholinergic stimulation of astrocytes enhances synapse formation in co-cultured hippocampal neurons.
synaptophysin and PSD-95 protein levels (Fig. 3A, B). These findings suggest that the increase in puncta may be due, at least in part, to increased expression of synaptophysin and PSD95 expression in neurons.
To further verify the increase in synaptic structures induced by the cholinergic stimulation of astrocytes, we immunolabeled neurons with two other markers of synaptic structures, the presynaptic protein synaptotagmin, and the postsynaptic NMDA receptor NR2B subunit. Representative deconvolved images of hippocampal neurons after co-culture for 24 h with astrocytes pretreated with 1 Fig. 1 Cholinergic stimulation of astrocytes potentiates synaptic structure formation. Astrocytes were treated with increasing concentrations of carbachol (0, 0.01, 0.10, and 1 mM) for 24 h. After treatment washout, primary hippocampal neurons, grown on glass coverslips, were inverted over astrocytes and co-cultured for 24 h. Neurons were immunocytochemically labeled for the pre and postsynaptic proteins synaptophysin and PSD-95 and imaged using confocal microscopy. Synaptic structures were assessed using three-dimensional object analysis. (A) Representative deconvolved images of hippocampal neurons after coculture for 24 h with astrocytes pretreated with 1 mM carbachol or with control astrocytes. Synaptophysin is shown in green pseudocolor; PSD-95 is shown in red; overlapping puncta are shown in yellow (scale bar = 10 μm). Quantification of the number of synaptophysin puncta (B), PSD-95 puncta (C), and overlapping puncta (D) in neurons exposed for 24 h to control astrocytes or astrocytes pre-treated with increasing concentrations of carbachol. Control and 1 mM treatments: n = 55-66 neurons from 6 independent experiments; 0.010 and 0.100 mM treatments: n = 21-34 neurons from 3 independent experiments. The bold, dashed line represents the median; the upper and lower dotted lines represent the 75th and 25th percentile, respectively. One-way ANOVA followed by Dunnett's Multiple Comparison test was used for statistical comparison of the means. *: p < 0.05, **: p < 0.01, ***: p < 0.001 vs. control

Astrocyte-released TSP1 is Involved in the Increase in Synaptic Structures After Cholinergic Stimulation of Astrocytes
As TSP1 has been involved in the formation of synaptic structures, we investigated whether astrocytes induce protracted expression and secretion of TSP1 after carbachol pre-treatment. Astrocytes were treated with carbachol for 24 h followed by treatment washout and the subsequent replacement with treatment-free culture medium. We collected astrocyte medium and cells lysate at the end of the 24 h incubations (time 0), and 6 and 24 h after treatment wash-out; levels of secreted and total cellular TSP1 were quantified by Western blot.
Medium from astrocytes treated with carbachol showed a 2.4-fold increase in TSP1 compared to medium from untreated astrocytes at the end of the 24 h incubation with carbachol (Fig. 6A, B) confirming our previous findings [52]. Six hours after replacing the treatment medium with carbachol-free medium, TSP1 was detectable, though at lower levels, suggesting continuous release of TSP1. Importantly, 6 h after treatment wash-out, but not after 24 h, the amount of TSP1 released by carbachol-treated astrocytes was still greater than that released by untreated astrocytes

Cholinergic Stimulation of Astrocytes Increases the Number of Functional Synapses
To investigate whether carbachol-stimulated astrocytes increased functional synapses, we measured the frequency of spontaneous miniature excitatory postsynaptic currents (mEPSCs) [51] in neurons co-cultured with astrocytes pretreated with carbachol compared to neurons co-cultured with control astrocytes. We found a greater frequency of mEPSCs, as measured by the mEPSCs inter-event interval (Fig. 5A), indicating an increased number of functional synapses in neurons co-cultured with carbachol-stimulated astrocytes. A slight, but statistically significant increase in the amplitude was also observed (Fig. 5B), suggesting a potential increase in the number of postsynaptic receptors in neurons co-cultured with carbachol-treated astrocytes. There was no significant difference in the mean time to half decay between treatment groups indicating no changes in the kinetics of the response (Fig. 5C). Overall, these results indicate that astrocytes pre-treated with the AChR agonist carbachol enhance the formation of functional synapses. in total cellular TSP1. It is possible, however, that additional mechanisms, such as decreased degradation of TSP1, may be involved, since the increased levels of extracellular TSP1 last for several hours after the removal of carbachol treatments ( Fig. 6A-D), while the increased expression of Thbs1 gene by carbachol observed after 6 h, returns to control levels after 24 (Fig. 6G). In a previously published study we have reported that carbachol increases the levels of two other extracellular matrix proteins, laminin and fibronectin, by decreasing their proteolysis by plasmin [2]; a similar mechanism, in combination with the initial increase in gene expression, may be also involved. The lasting increase in extracellular TSP1 levels may be responsible, in part, for carbachol-treated astrocyte-induced synaptogenesis.
To test whether TSP1 alters synaptogenesis directly, hippocampal neurons were treated with TSP1 (5 or 10 µg/ mL) for 24-hours. The number of synaptophysin and PSD-95 puncta was not increased by treatment of neurons with TSP1 (Fig. 7A, B). However, the number of co-localized ( Fig. 6A, B). These post-treatment time points are within the time-frame of astrocyte-neuron co-cultures (see Figs. 1, 2 and 3). We used immunocytochemistry and confocal imaging to measure the surface expression of TSP1 in non-permeabilized cells. Twenty-four hours after carbachol treatments, astrocytes showed a 3.35-fold increase in TSP1 surface fluorescence relative to untreated astrocytes (Fig. 6C, D). Total TSP1 levels, measured by Western blot in the astrocyte cell lysate, were not altered (Fig. 6E, F). We also investigated the effect of carbachol on the expression of the Thbs1 gene encoding for TSP1 and found that incubation of astrocytes with carbachol increases Thsb1 RNA expression after 6 h; Thsb1 RNA expression returns to control levels after 24 h incubation with carbachol and remains at control levels 6 h after treatment removal following 24 h incubation with carbachol (Fig. 6G). Taken together, these results indicate that cholinergic stimulation of astrocytes enhances the secretion of TSP1, an effect that may be due to the transient increase in Thbs1 gene expression, since no changes were observed Control is represented as dotted line. Shown are means ± SEM. Statistical analysis was performed using one-way ANOVA followed by Dunnett's Multiple Comparison test (***: p < 0.001); n = 5 coverslips for 10 and 100 µM carbachol; n = 14-15 coverslips for control and 1 mM carbachol secretion by astrocytes is increased by carbachol treatment, is involved in the increased synaptogenesis observed after co-culture of neurons with carbachol-stimulated astrocytes.

Discussion
Increasingly, the study of CNS function in health and disease has focused on the bi-directional communication and signaling between glia cells and neurons [10,26,[54][55][56][57][58][59]. In particular, astrocytes have been shown to respond to external signals, including neurotransmitters and neuroactive substances [58], to modulate gene expression [60], and to synaptophysin/PSD-95 puncta increased up to 2-fold relative to untreated neurons (Fig. 7C) indicating an increase in the formation of apposed pre-and postsynaptic structures.
To test whether TSP1-activated signaling in neurons is necessary for the enhanced synaptogenic effect of carbachol pre-treated astrocytes, we incubated astrocyte-neuron co-cultures with gabapentin, an inhibitor of the α2δ1 subunit of neuronal voltage-gated calcium channels that mediates the synaptogenic effects of TSPs on retinal ganglion cells [18,53]. The addition of gabapentin (15 and 30 µM) to the co-culture system containing both neurons and carbachol pre-treated astrocytes blocked the upregulation of synaptic puncta induced by carbachol-treated astrocytes (Fig. 7D). Together, these results indicate that TSP1, whose Fig. 4 Carbachol-stimulated astrocytes increase the number of synaptotagmin and NR2B pre and postsynaptic puncta and their overlap. Hippocampal neurons were co-cultured for 24 h with astrocytes pre-treated with 1 mM carbachol. Neurons were immunolabeled for the presynaptic protein synaptotagmin and the postsynaptic protein NR2B. (A) Representative deconvolved images of hippocampal neurons after co-culture for 24 h with astrocytes pretreated with 1 mM carbachol or with control astrocytes. Synaptotagmin is shown in green pseudocolor; NR2B is shown in red; overlapping puncta are shown in yellow (scale bar = 10 μm). Synaptotagmin (B) and NR2B (C) puncta and their overlap (D) were quantified after confocal imaging using three-dimensional object analysis. The bold, dashed line represents the median; the upper and lower dotted lines represent the 75th and 25th percentile, respectively; n = 19-24 neurons from 6 coverslips. Statistical comparison of the means was performed using the Student's t-test. ***: p < 0.001 vs. control Several studies have focused on the role of ACh in brain development independent of its function as neurotransmitter as nicotinic and muscarinic ACh receptors and acetylcholinesterase are expressed in the brain before synapses are formed [43][44][45]. Neuronal acetylcholine release is known to influence neuronal migration, differentiation, synapse formation and function [32,[38][39][40][41][42], as well as oligodendrocyte development and function [56]. However, this is the first study to report that cholinergic stimulation in astrocytes directly modulates synaptogenesis in neurons.
In this study we characterized the effects of cholinergic stimulation in astrocytes on the formation of active synapses. We report that cholinergic stimulation of astrocytes potentiates the formation of synapses (Figs. 1, 2, 3, 4 and 5), in part by enhancing astrocyte secretion of the synaptogenic factor TSP1, which persists after the ACh receptor agonist has been removed (Fig. 6), and that the inhibition of TSP signaling by gabapentin abolishes the synaptogenic effects of carbachol-stimulated astrocytes (Fig. 7).  24 h incubation with carbachol and 24 h incubation with carbachol followed by treatment wash out and 6 h incubation in treatment-free medium did not alter Thbs1 expression indicating that carbachol transiently increases Thbs1 expression in astrocytes. Statistical analysis was performed using the Student's t-test at each time point relative to time-matched control; *: p < 0.05; ***: p < 0.001 slower desensitization time course than ionotropic receptors [61]. Somewhat surprisingly, the nicotinic receptor antagonist mecamylamine also attenuated the carbachol effect (Fig. 2). While ionotropic receptor signaling in neurons is We found that synapse formation induced by carbachol-stimulated astrocytes is inhibited by the M3-AChR antagonist 4-DAMP (Fig. 2). Signaling through GPCRs in astrocytes is typically long-lasting because GPCRs have a control. (D): Astrocytes were treated for 24 h with 1 mM carbachol or left untreated; after carbachol treatment washout astrocytes were cocultured with hippocampal neurons for 24 h in the presence or absence of gabapentin (15, 30 µM) to block TSP1 function in neurons. Presynaptic (synaptophysin) and postsynaptic (PSD95) puncta overlap was quantified as previously described. The distribution of the data is shown (the bold, dashed line represents the median; the upper and lower dotted lines represent the 75th and 25th percentile, respectively; n = 25-43 neurons from 4 independent experiments). Statistical analysis was performed using one-way ANOVA followed by Bonferroni's Multiple Comparison test. **: p < 0.01 vs. control; ##: p < 0.01, ###: p < 0.001 vs. carbachol which inhibits the activation of the extracellular proteolytic enzyme plasmin, leading to decreased proteolysis of neuritogenic laminin and fibronectin in the extracellular space [2,73]. The present study reports on a novel function of cholinergic signaling in astrocytes: i.e., synaptogenesis induction. We have not investigated which of the intracellular signaling pathways activated by carbachol in astrocytes are responsible for synaptogenesis induction; this is a limitation of the present study, and an opportunity for future investigations.
In a previous proteomic study we reported that carbachol increased the release of TSP1 from astrocytes [52]. In this study, we confirmed and expanded this observation (Fig. 6) by showing that TSP1, in the absence of astrocytes, increased synapse numbers (i.e., the colocalization of presynaptic and postsynaptic puncta), but did not increase the number of pre-or postsynaptic puncta (Fig. 7), suggesting that TSP1 may facilitate the alignment of pre and postsynaptic puncta into structural synapses. Indeed, it has been proposed that the TSPs, acting through the postsynaptic α2δ1 receptor, may exert their synaptogenic effects through the recruitment of cell adhesion and scaffolding proteins, promoting the assembly of apposing pre-and postsynaptic structural synapses [14,17,18,53].
Consistent with these findings, we report that inhibition of α2δ1 signaling by gabapentin prevents carbacholtreated-astrocyte-induced increased synapses (Fig. 7). TSP1 is expressed [60] and released by astrocytes during synaptogenesis [3] and exogenous addition of TSP1 and TSP2 to retinal ganglion cells increased the number of synapses which are presynaptically active, but postsynaptically silent usually localized and short-lived, the activation of astrocyte-expressed α7 ionotropic nAChRs has been shown to increase permeability to Ca 2+ , which in turn causes a prolonged release of Ca 2+ from intracellular stores [23] and plays a role in synapse function by recruiting AMPA receptors to the postsynaptic surface [62].
We also found that cholinergic stimulation of astrocytes increased the frequency and amplitude of mEPSCs (Fig. 5), indicating an increase in the number of functional synapses [51]. Importantly, we found that the number of puncta containing the NR2B-subunit of the NMDA receptor increased when neurons where co-cultured with carbachol pre-treated astrocytes (Fig. 4). NMDARs containing this subunit are expressed during early development [63], and have been shown to be required for synaptic plasticity and synaptogenesis [64].
Previous studies carried out in our laboratory characterized the signaling pathways activated by carbachol in astrocytes via M3 muscarinic receptors. We have shown that carbachol, through the activation of phospholipase C (PLC), PLD, and PI3K and the formation of inositol 1,4,5-trispophate (IP 3 ) increases the release of Ca ++ from intracellular stores and induces activation of PKCε, PKCζ, NFκB, MAPK, and p70S6Kinase [65][66][67][68][69][70][71][72]. We also have reported that this signaling pathway is involved in carbachol-treated astrocyte-induced neurite outgrowth in co-cultured hippocampal neurons at an early stage of development before synaptogenesis has started. Carbachol-stimulated astrocyte-mediated neurite outgrowth is the result of increased expression and release of plasminogen activator inhibitor 1, This effect is in part mediated by an increased release of TSP1 from astrocytes, which increases the alignment of pre and postsynaptic puncta into structural synapses [3]. TSP1 also increases the speed of synapse formation in hippocampal neurons [74]. Additionally, ATP activation of astrocytes results in increased expression of TSP1 in a time and concentration-dependent manner [75]. Because we found that carbachol-treated astrocytes increase the formation of functional synapses (Fig. 5), additional factors involved in the formation of functionally active synapses must be released by the cholinergic stimulation of astrocytes. Future studies will investigate other astrocytereleased factor induced by cholinergic stimulation that are involved in the formation or in the inhibition of functional synapses, such as hevin, TGF-β and TGF-β1, GPC 4 and 6, Chrdl1, PTX3, and SPARC [3,6,[9][10][11][12][13][14][15][16][17][18]. Interestingly, our proteomic study carried out on astrocyte conditioned medium reported that astrocyte cultures release several of these factors including GPC4, hevin, Chrdl1, and SPARC [52].
In conclusion our study described a new mechanism of astrocyte-neuron communication by which a signal from neurons, ACh, through the stimulation of a robust signaling pathway in astrocytes [73], leads to the release of synaptogenic proteins and increased formation of functional synapses as depicted in schematic form in Fig. 8. This study expands our understanding of the signals that regulate the release of synaptogenic proteins from astrocytes and contributes a novel mechanism by which the cholinergic system is involved in brain development.