INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

In the CNS, three major conditions are characterized by a low level of glutamate synaptic activity. The first condition is embryonic development. During prenatal development, glutamate activity is not yet manifested in many regions of the CNS, although the activity of other types of neurotransmitters is detected [e.g., GABA in rat hypothalamus (Chen et al. 1995) and ACh in chick spinal cord (Milner and Landmesser 1999) and rat retina (Feller et al. 1996)]. The second condition is selective degeneration of glutamatergic neurons or projections found during a number of neurodegenerative diseases. This includes degeneration of the hippocampal and cortical glutamate-secreting projections during epilepsy (Babb 1997) and Alzheimer's disease (Lawlor and Davis 1992; Simpson et al. 1988). Such degeneration may deprive target regions of excitatory glutamate inputs and reduce glutamate synaptic transmission in those regions. The third condition is the administration of anti-glutamate receptor drugs for therapeutic purposes. Agents that reduce or completely suppress the activity of N-methyl-D-aspartate (NMDA) glutamate receptors (e.g., eliprodil, amantadine, dextromethorphan, memantine, CGS 19755, etc.) are considered major drug candidates for neuroprotection in epilepsy, stroke, multiple sclerosis, Huntington's chorea, Parkinson's disease, Alzheimer's disease, and several other disorders (http://clinicaltrials.gov/). These drugs are used for the chronic treatment of patients, and they reduce glutamate transmission in the CNS.

Data indicate an increased ACh function in the CNS during each of these three conditions. ACh is considered to be the major excitatory neurotransmitter in some CNS regions (spinal cord, retina) during early stages of development before glutamate begins to play this role at later stages (Feller et al. 1996; Milner and Landmesser 1999). ACh hyperactivity, hypersensitivity, and the sprouting of subcortical cholinergic neurons have been found to accompany the degeneration of glutamate projections during epilepsy (Correia et al. 1998; Holtzman and Lowenstein 1995; Kish et al. 1988) and early stages of Alzheimer's disease (Geddes et al. 1985; Olney et al. 1997). NMDA antagonists used for therapeutic purposes have potentially serious side effects including motor disturbances, weight loss, psychotic symptoms, memory impairments, and morphological damage of neurons in the cerebral cortex (Facchinetti et al. 1993; Kornhuber and Weller 1997; Olney 1994). Some of these side effects (e.g., morphological damage of cortical neurons) are prevented with ACh receptor (AChR) antagonists and are presumably due to the hyperactivation of subcortical cholinergic projections to the cortex (Olney et al. 1991). Observations on rat pups chronically subjected to NMDA receptor blockade indicated a dramatic increase in the catalytic activity of choline acetyltransferase (ChAT, an ACh biosynthetic enzyme) in some regions of the CNS (cerebellum, spinal cord, striatum, etc.) but not in others (cortex, hippocampus) (Facchinetti et al. 1993, 1994; Virgili et al. 1994). Our recent experiments in vitro (Belousov et al. 2001) also revealed a dramatic increase in the expression of ACh synaptic transmission during a chronic blockade of ionotropic glutamate receptors. This was detected in neuronal cultures obtained from the hypothalamus and cerebellum but not from the cortex. The data also implied that a decrease of Ca²⁺ influx into cells through NMDA glutamate receptors was the principal component responsible for the cholinergic upregulation in hypothalamic cultures (Belousov et al. 2001). Here we study Ca²⁺-dependent mechanisms of the regulation of ACh transmission and cholinergic cell phenotype in the hypothalamus in vitro. We demonstrate that a decrease in Ca²⁺ influx into cells, a blockade of cytoplasmic Ca²⁺ fluctuations, or an inactivation of intracellular Ca²⁺-dependent signaling pathways increase the number of cholinergic neurons and induce excitatory cholinergic activity in hypothalamic neuronal cultures. Together, these data support our original hypothesis (Belousov et al. 2001) that during a long-term decrease in glutamate excitation in the hypothalamus in vitro, another less predominant excitatory neurotransmitter, ACh, plays the role of the major excitatory neurotransmitter and supports the excitation/inhibition balance.

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Tissue cultures

Cultures were prepared from the embryonic (day 18-19) medial hypothalamus or cerebral cortex obtained from Sprague-Dawley rats as described (Belousov et al. 2001). After disaggregation into suspension of single cells (using papain, 10 units/ml), neurons were plated on glass coverslips (on a polylysine substrate) and maintained in a Napco 5430 incubator at 37°C with 5% CO₂. Cells were raised in glutamate- and glutamine-free MEM (Life Technologies, Rockville, MD) supplemented with 10% fetal bovine serum, 5 mg/100 ml gentamicin, and 6 g/l glucose. After 2 days in vitro, cell proliferation was inhibited by the application of cytosine -D-arabinofuranoside (AraC; 1 µM), the selective inhibitor of DNA (but not RNA) synthesis (Furlong and Gresham 1971). In these experiments, we used the following culture conditions: 1) cultures chronically treated (for 14-17 days; starting day 4 in vitro) with an agent blocking specific signaling pathways or cell functions (nifedipine, nimodipine, W13, KN-62, GF 109203X, or Rp-cGMPS), 2) cultures subjected (on day 4 in vitro) to 1,2-bis(2-aminophenoxy)ethane-N,N,N',N'-tetraacetic acid tetrakis(acetoxy-methyl) ester (BAPTA AM, 2.5 µM) for 30 min with subsequent washing (3 times) followed by incubation in the control medium for 14-17 days, 3) cultures chronically (for 14-17 days; starting day 4 in vitro) treated with 20 mM KCl and some blocking agents (nimodipine, W13, or KN-62), 4) cultures chronically subjected to blockade of NMDA and/or non-NMDA [-amino-3-hydroxy-5-methyl-4-isoxazolepropionic acid (AMPA) and kainate] glutamate receptors with, respectively, D, L-2-amino-5-phosphonovalerate (AP5, 100 µM) and 6-cyano-7-nitroquinoxaline-2,3-dione (CNQX, 10 µM), and 5) cultures not subjected to any treatment (control cultures; 18-21 days in vitro; in some experiments, older control cultures (7 wk in vitro) were used). Tissue culture medium was changed twice a week. Only healthy looking cultures were used in experiments; unhealthy cultures were discarded.

Ca²⁺ current recordings

Bathing solution contained (in mM): 148.5 NaCl, 2.5 KCl, 2 BaCl₂, 10 HEPES, 10 glucose, 1 × 10³ tetrodotoxin, 20 tetraethylammonium, 2 4-aminopyridine, and 1 × 10³ glycine (pH 7.3, 325 mosM, flow rate of 1.5 ml/min, room temperature). The internal solution contained (in mM): 145 CsCl, 10 HEPES, 5 MgCl₂, 1.1 EGTA, 4 Na-ATP, and 0.5 Na-GTP (pH 7.2, 310 mosM). Single-electrode continuous voltage-clamp mode (Axoclamp-2B; Axon Instruments, Foster City, CA) was used to measure Ca²⁺ currents activated by a step of depolarization (90 mV, 130 ms) from a holding potential of 90 mV. A flow pipe perfusion system (Belousov et al. 2001) was used for the cell perfusion and nimodipine and CdCl₂ application. Data were monitored using a Dell Pentium II XPS R400 MHz computer and pCLAMP7 software (Axon Instruments) and analyzed off-line with Igor Pro (WaveMetrics, Lake Oswego, OR) and InStat 2.03 (GraphPad Software, San Diego, CA). Leak currents were subtracted. Measurements of the amplitude of Ca²⁺ currents were made at the peak of current.

Fura-2 Ca²⁺ digital imaging

Cells were loaded with fura-2 acetoxymethyl ester (5 µM; Molecular Probes, Eugene, OR) for 30 min and studied on a Nikon inverted microscope with a Nikon Super Fluor ×20 objective as described (Belousov et al. 2001). Conventional dual wavelength ratios were obtained during sequential recordings at 340- and 380-nm excitation. Switching excitation filters was performed by a Sutter DG-4 optical filter changer (Sutter Instrument, Novato, CA). A Dell Pentium II XPS R400 MHz computer and Axon Imaging Workbench software were used to control peripheral devices. Cells were imaged in a laminar-style chamber and bathed in a perfusion solution containing (in mM): 137 NaCl, 25 glucose, 10 HEPES, 5 KCl, 1 MgCl₂, 3 CaCl₂, and 1 × 10³ glycine (pH 7.4, flow rate of 1.5 ml/min, room temperature). Images were taken every 4 s with a SensiCam Digital CCD camera. Calcium standards from Molecular Probes (calcium imaging calibration kit, F-6774) were used to calibrate the imaging system as described previously (Belousov et al. 2001). Calibrated Ca²⁺ data were transferred to a Power Macintosh G3 computer and analyzed with Igor Pro and InStat software. No significant difference in the background Ca²⁺ level was detected in neurons from various cell culture conditions; this was usually in the range of 50-80 nM in different cells. The expression of excitatory ACh-mediated Ca²⁺ activity was tested in cultured neurons during application of the GABA_A receptor antagonist bicuculline (50 µM) in the presence of 100 µM AP5 and 10 µM CNQX. In hypothalamic cultures, AP5 and CNQX in such concentrations effectively suppress glutamate-mediated neuronal responses (Belousov et al. 2001). To confirm the cholinergic origin of bicuculline-mediated Ca²⁺ activity, AChR antagonists atropine and mecamylamine (100 µM each; applied jointly) were used. Previous experiments (Belousov et al. 2001) revealed that at this concentration atropine and mecamylamine exert only specific, receptor-mediated effects and do not affect isolated glutamate-dependent activity.

Only Ca²⁺ changes in cell bodies were recorded. Previous Ca²⁺ imaging experiments revealed that all cultured hypothalamic neurons are NMDA-sensitive (Obrietan and van den Pol 1995). Therefore in our experiments, neurons were recognized by their responsiveness to the application of 10 µM NMDA in a Mg²⁺-free solution and by their "phase bright" appearance. The responsiveness of neurons to NMDA was also used in some experiments to confirm that cells were healthy and responsive. A neuron was considered as responding to a pharmacological agent (e.g., bicuculline, NMDA, etc.) if during the agent application Ca²⁺ increased by more than 10 nM from the initial background level and if the level of Ca²⁺ decreased to the background after the agent washout. As in the preceding work (Belousov et al. 2001), during neuronal disinhibition with bicuculline, both sustained and oscillatory Ca²⁺ increases were detected in different neurons from various cell culture conditions. Ca²⁺ increases lasted as long as bicuculline was applied (usually 3-5 min). The amplitude of Ca²⁺ increase was measured either at the peak of response (in cells with sustained Ca²⁺ increase) or by averaging peaks of several oscillations (in oscillating cells).

ChAT immunocytochemistry

Coverslips containing neurons were fixed in 4% paraformaldehyde and 0.3% glutaraldehyde in 0.01 M phosphate buffer for 15 min. After three buffer washes, the fixed neurons were treated with 0.2% Triton X-100 to permeabilize the membranes, and then incubated in mouse anti-ChAT monoclonal antibody (Chemicon International, Temecula, CA) at a dilution of 1:1500 for 1 h at room temperature and then for 5 days at 4°C. After washing (3 times), the neurons were incubated for 30 min in biotinylated goat anti-mouse secondary antibody (Jackson ImmunoResearch Laboratories, West Grove, PA) diluted in phosphate buffer at 1:150. Cells were then rinsed three times in phosphate buffer and incubated in ABC reagent (Vector Laboratories, Burlingame, CA) for 30 min at room temperature. For visualization of the immunoreaction, DAB substrate was used as a chromogen (Vector Laboratories). Stained (ChAT-immunopositive) cells were distinguished as dark-colored cells as compared with light-colored or noncolored cells that were considered as nonstained (ChAT-immunonegative). The number of ChAT-positive and -negative neurons was counted in 10 randomly chosen microscope fields in each coverslip using a Nikon inverted microscope (×20 objective). Each test was done two to four times on independent culture preparations. Negative controls included staining for the secondary antibody without preincubation with the primary antibody (10 coverslips total). The staining analysis was performed blindly. The experimenter did not know the treatment condition of the cultures that were labeled with randomly chosen numbers by another person.

Bromodeoxyuridine labeling

Three groups of hypothalamic cultures were used in these experiments. The first group included six coverslips with control cultures incubated in the absence of an antimitotic drug, AraC ("control with no AraC"). 5-Bromo-2'-deoxyuridine (BrdU; 10 µM) was chronically present in these cultures starting day 2 in vitro. Cells were taken for BrdU staining on day 6 in vitro. The second group included six coverslips with cells that were incubated in the constant presence of AraC (1 µM; starting day 2 in vitro) and the NMDA receptor blocker, AP5 (100 µM; starting day 4 in vitro) ("AP5- and AraC-treated" cultures). BrdU (10 µM) was added to these cultures starting day 4 in vitro. Cells from this group were taken for BrdU staining (3 coverslips) and ChAT staining (3 coverslips) on day 18 in vitro. The first group of cultures was taken for staining earlier than the second group because in the absence of AraC, glial cells quickly overgrow and unfavorably affect neurons. The third group included three coverslips with control cultures incubated in the presence of 1 µM AraC only (starting day 4 in vitro) ("control with AraC"). Cells of this group were taken for ChAT staining on day 18 in vitro.

For BrdU staining, cells were fixed and processed as described above for ChAT staining. An additional step, performed before the application of the primary antibody, included an incubation of cells in the presence of 2N HCl for 40 min followed by wash out (3 times) with the phosphate buffer. A monoclonal primary mouse anti-BrdU antibody (Sigma-RBI, St. Louis, MO) was used for immunostaining (1:1000 dilution, 2 h incubation at room temperature). As with ChAT staining, biotinylated goat anti-mouse secondary antibody (Jackson ImmunoResearch Laboratories; 1:150 dilution; 30 min incubation) and DAB substrate were used to complete BrdU staining. Staining was analyzed in 10 randomly chosen microscope fields in each coverslip (×20 objective). BrdU-positive cells appeared as black nuclei in the grayscale contrast field. The percentage of BrdU-positive cells was quantified from a ratio of BrdU-labeled nuclei to the total number of neurons. Staining analysis was performed blindly. Negative controls included staining for the primary and secondary antibodies in cultures incubated without BrdU (3 coverslips) and staining for the secondary antibody without preincubation with the primary antibody in BrdU-treated cultures (3 coverslips).

Drugs and chemicals

All drugs used in this research were obtained from Sigma-RBI (St. Louis, MO).

Statistical analysis

Statistical analysis of data were performed using InStat software. Data in all experiments were compared by Student's t-test, using paired data when possible. All data are reported as mean ± SE for the number of neurons indicated.

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Ca²⁺ mechanisms of ACh regulation in the hypothalamus in vitro

NMDA glutamate receptors and L-type voltage-gated Ca²⁺ channels are two major sites for Ca²⁺ entry into neurons that transmit signals to the nucleus and regulate gene transduction through intracellular Ca²⁺-dependent signaling pathways (Bading et al. 1993; Hardingham et al. 1999). In our previous experiments in hypothalamic neuronal cultures (Belousov et al. 2001), a chronic blockade of NMDA receptors induced excitatory ACh-dependent activity that was detected in many neurons. Here, we tested whether a chronic blockade of L-type Ca²⁺ channels mimics NMDA receptor blockade and triggers the development of ACh activity.

First, the amplitude of the total voltage-dependent Ca²⁺ current was measured in cultured hypothalamic neurons (18-21 days in vitro; control culture) and was 375.3 ± 43.8 pA (n = 9 neurons; Fig. 1A). Nimodipine (5 µM), an L-type Ca²⁺ channel blocker, decreased this current to 194.4 ± 28.6 pA (n = 9; P < 0.0001). The current was completely abolished by CdCl₂ (500 µM). This suggested that L-type Ca²⁺ channels contribute to Ca²⁺ influx into hypothalamic neurons.

View larger version (23K): [in this window] [in a new window]   Fig. 1. ACh-mediated excitatory activity in hypothalamic neuronal cultures subjected to Ca2+ channel blockade. Representative recordings from voltage-clamp (A) and Ca2+ imaging (B-F) experiments on hypothalamic (A-E) and cortical (F) neurons are shown. A: voltage-gated Ca2+ current recorded in this control neuron was reduced by nimodipine (5 µM) and completely suppressed by CdCl2 (Cd2+; 500 µM). B: typical control hypothalamic cell did not respond to bicuculline (BIC; 50 µM) in the presence of glutamate receptor antagonists. C and D: hypothalamic neurons chronically subjected to nimodipine (20 µM; C) or nifedipine (20 µM; D) expressed excitatory Ca2+ activity during neuronal disinhibition. The activity was suppressed by ACh receptor (AChR) antagonists atropine and mecamylamine (ATR/MEC; 100 µM each). E: neurons in hypothalamic cultures grown in the presence of nimodipine (20 µM) and KCl (20 mM) did not reveal ACh-dependent activity, but they did express low-amplitude (10-30 nM) Ca2+ oscillations synchronized between all cells in the microscope field (2 representative cells are shown). F: cultures obtained from the cortex that were incubated in the presence of nimodipine (20 µM) did not reveal excitatory Ca2+ activity. D,L-2-amino-5-phosphonovalerate (AP5, 100 µM) and 6cyano-7-nitroquinoxaline-2,3-dione (CNQX, 10 µM) were in all solutions in Ca2+ experiments except for the N-methyl-D-aspartate (NMDA)-containing (10 µM) but Mg2+-free solution in B, E, and F that was used to confirm that the cells were healthy and responsive. Here and in the following figure, arrows below the recordings indicate the time of AP5/CNQX application. Applications of other drugs are indicated by the bars above the recordings. Asterisks in E indicate Ca2+ spikes synchronized in 2 presented neurons. Calibration bars: 50 ms/200 pA in A and 1 min/75 nM Ca2+ for all Ca2+ recordings.

Further, Ca²⁺ imaging experiments were conducted in three groups of cultures. Two groups were incubated chronically (for 14-17 days; starting day 4 in vitro) in the presence of nimodipine (20 µM) or nifedipine (20 µM; another L-type Ca²⁺ channel blocker). The third group of cultures was not subjected to Ca²⁺ channel blockade and served as the control. Cells were tested during application of the GABA_A receptor antagonist bicuculline (50 µM) in the presence of NMDA and non-NMDA glutamate receptor antagonists AP5 (100 µM) and CNQX (10 µM) in the incubating medium. No Ca²⁺ increases were detected in control neurons under these test conditions (n = 285; Fig. 1B; Table 1). In contrast, when neurons chronically subjected to Ca²⁺ channel blockers were tested during synaptic disinhibition in the presence of glutamate receptor antagonists, they revealed a significant increase in intracellular Ca²⁺ levels. Ca²⁺ increases were detected in 99 of 154 neurons in nimodipine-treated cultures (Fig. 1C) and in 89 of 124 cells in nifedipine-treated cultures (Fig. 1D). The muscarinic AChR antagonist, atropine, and the nicotinic AChR antagonist, mecamylamine (100 µM each; applied jointly), blocked or significantly (>50%) suppressed the bicuculline-mediated Ca²⁺ rises in most of neurons in both of these cultures (Fig. 1, C and D). Data suggested that 55% (of 154) and 53% (of 124) neurons in cultures treated with nimodipine and nifedipine, respectively, expressed excitatory ACh-mediated Ca²⁺ responses (Table 1).

BAPTA is a Ca²⁺-chelating agent that prevents increase in intracellular Ca²⁺ concentration and blocks cytoplasmic Ca²⁺ fluctuations in neurons (Kao 1994). We tested whether such a Ca²⁺ blockade with BAPTA upregulates ACh activity. Cultures (4 days in vitro) were incubated for 30 min in the presence of BAPTA AM (2.5 µM) in the cell culture medium with the subsequent wash out of the drug. For the following 14 days, cultures were incubated in the normal culture medium. When tested, 46% of 186 neurons in these cultures revealed ACh-dependent Ca²⁺ activity (Fig. 2A; Table 1).

View larger version (33K): [in this window] [in a new window]   Fig. 2. Calcium mechanisms of ACh regulation in hypothalamic neurons. Representative Ca2+ imaging recordings from hypothalamic neurons are shown. A: ACh-mediated Ca2+ activity was detected in this neuron 2 wk after the 30 min treatment with 1,2-bis(2-aminophenoxy)ethane-N,N,N',N'-tetraacetic acid tetrakis(acetoxy-methyl) ester (BAPTA AM, 2.5 µM). B-D: ACh activity was detected in hypothalamic cultures chronically incubated in the presence of W13 (5 µg/ml, B), KN-62 (2.5 µM, C), or GF 109203X (500 nM, D). E: no excitatory activity was detected in most of the neurons in cultures subjected to Rp-cGMPS (5 µM).

Calmodulin (Ca²⁺-sensing protein) and Ca²⁺/calmodulin-dependent protein kinases (CaMK) I, II, and IV appear to play an important role in mediating many of the second-messenger actions of Ca²⁺ in neurons (Picciotto et al. 1996; Soderling 1996). CaMK II and IV are involved in the mechanisms of neuronal plasticity in the CNS through the modulation of gene expression (Ahn et al. 1999; Deisseroth et al. 1998; Siegel et al. 1999). We tested whether inactivation of calmodulin or CaMK II/IV mimics the blockade of NMDA receptors and induces ACh activity in cultures. Cultures were incubated chronically in the presence of either W13 (5 µg/ml; selective calmodulin antagonist) or KN-62 (2.5 µM; blocker for CaMK II/IV). ACh-mediated excitatory activity was detected in 60% of 96 neurons after calmodulin blockade (Fig. 2B) and 40% of 129 neurons after the blockade of CaMK II/IV (Fig. 2C).

The contribution of protein kinase C (PKC) to the regulation of AChRs in chick muscles has been demonstrated previously (Klarsfeld et al. 1989). Here, we tested the contribution of PKC to the regulation of ACh excitation in the hypothalamus in vitro. Cultures were incubated chronically in the presence of 500 nM GF 109203X (specific antagonist for PKC). ACh-dependent Ca²⁺ responses were detected in 41% of 254 neurons in these cultures (Fig. 2D). In contrast, most of cells in cultures that were subjected chronically to the blockade of cGMP-dependent protein kinase (PKG) with the specific antagonist Rp-cGMPS (5 µM) did not reveal excitatory Ca²⁺ increases (Fig. 2E). Only 5% of 331 neurons in these cultures revealed ACh-dependent responses in Ca²⁺ recordings (Table 1).

Cell depolarization in neuronal cultures with elevated KCl has been shown to increase Ca²⁺ influx into neurons through voltage-gated Ca²⁺ channels and to prevent the effect of NMDA glutamate receptor blockade on neurite motility (Lin and Constantine-Paton 1998). We have shown previously that the co-incubation of neurons with KCl (20 mM) prevented the development of ACh activity in cultures subjected to a chronic glutamate receptor blockade with AP5 and CNQX (Belousov et al. 2001). Here, we tested whether cell depolarization (and Ca²⁺ influx to cells) also prevents the development of ACh activity in neurons under other treatment conditions. KCl (20 mM) was added chronically to cultures incubated in the presence of nimodipine (20 µM), W13 (5 µg/ml), or KN-62 (2.5 µM). After 2 wk of the treatment, neurons in all three groups were sensitive to NMDA (10 µM; applied in Mg²⁺-free solution; Fig. 1E) and were, therefore, healthy. The neurons, however, did not reveal excitatory responses, although they showed low-amplitude (10-30 nM) spontaneous fluctuations of intracellular Ca²⁺ synchronized among nearly all cells in a given microscope field (Fig. 1E; Table 1).

In the experiments, some bicuculline-mediated excitatory responses were not affected by AChR antagonists and were not cholinergic in nature. The percentage of neurons with noncholinergic responses varied among groups and was 10% (15 of 154 cells) in nimodipine-treated cultures, 19% (23 of 124 cells) in nifedipine-treated cultures, 3% (5 of 186) in BAPTA-AM-treated cultures, 8% (8 of 96) in W13-treated cultures, 9% (11 of 129) in cultures treated with KN-62, 19% (47 of 254) in cultures treated with GF 109203X, and 5% (16 of 331) in Rp-cGMPS-treated cultures.

Cholinergic neurons

In our previous experiments during neuronal disinhibition with bicuculline (Belousov et al. 2001), >80% of neurons expressed ACh-mediated Ca²⁺ responses in hypothalamic cultures chronically subjected to glutamate receptor blockade with AP5 and CNQX. An increase in ACh-dependent activity in those cultures, as well as in cultures under the experimental conditions described in the preceding text in this article, may be associated with an increase in the number of cholinergic neurons. To test this possibility, we used immunocytochemistry for ChAT. Only 0.5% of neurons in control hypothalamic cultures were ChAT-immunopositive and were, therefore cholinergic (Fig. 3A; Table 2). The number of ChAT-positive neurons increased to 35.5% in cultures after a chronic glutamate receptor blockade (with 100 µM AP5 and 10 µM CNQX; Fig. 3B). In other culture conditions, the percentage of ChAT-positive neurons was 22.2% in nimodipine-treated cultures, 17.8% in cultures treated with BAPTA AM (30 min treatment), 21.3% in W13-treated cultures, 23.0% in KN-62-treated cultures (Fig. 3C), 24.2% in cultures treated with GF 109203X (Fig. 3 days), and 5.0% in cultures treated with Rp-cGMPS (Table 2).

View larger version (50K): [in this window] [in a new window]   Fig. 3. ChAT immunostaining in hypothalamic cultures. a: most of neurons in the control cultures were not stained for ChAT (gray arrows). b-d: ChAT-immunopositive (black arrows) and ChAT-immunonegative (gray arrows) neurons were found in cultures subjected to a chronic blockade of glutamate receptors (with AP5 and CNQX, 100 and 10 µM; b), CaMK II/IV (with KN-62, 2.5 µM; c), or PKC (with GF 109203X, 500 nM; d).

In previous experiments, ACh activity was detected in many neurons (73%) in cultures chronically treated with AP5 alone and only in 1% of neurons subjected to CNQX alone. Here, the number of ChAT-positive cells was also increased in cultures after chronic AP5 treatment (15.1%) and was low (1.7%) in cultures treated with CNQX (Table 2).

Additionally, we tested whether the number of ChAT-positive neurons changes in older control cultures. These cultures were grown for 7 wk without any treatment. No change in the number of ChAT-positive neurons in such cultures was observed as compared with the younger controls (18-21 day in vitro; Table 2).

BrdU labeling

As described in detail in METHODS, three groups of cultures were used in these experiments. High BrdU staining was found only in the first group of cultures, "control with no AraC" (Fig. 4A). The ratio of BrdU-positive nuclei/total number of neurons was 1,228/10,497 (11.7%) as found in 60 microscope fields from six coverslips. Practically all BrdU-labeled cells had nonneuronal morphological appearance; they did not have "phase bright" round or oval cell bodies and exhibited flattened morphology. These cells were likely astrocytes. In some instances, when such BrdU-labeled cells also had slight background staining of the cell body, glial rather than neuronal cell bodies were seen (Fig. 4B). In three coverslips from the second group of cultures, "AP5- and AraC-treated", very sparse BrdU staining was found (Fig. 4C). The ratio of BrdU-positive nuclei/total number of neurons was 11/1,547 (0.7%) as found in 30 microscope fields from three coverslips. These 11 BrdU-labeled cells also had nonneuronal appearance. Three other coverslips from the second experimental group were stained for ChAT. The staining revealed an increased number of ChAT-positive neurons (232 of 1,439 cells, 16.1%; Fig. 4 days) as compared with control cultures from group three ("control with AraC"; 4 ChAT-positive cells of 1198 neurons, 0.3%; 30 microscope fields from 3 coverslips were analyzed).

View larger version (54K): [in this window] [in a new window]   Fig. 4. Assessment of cell proliferation in cultures using 5-bromo-2'-deoxyuridine (BrdU) incorporation. a: high BrdU staining was found in control cultures incubated in the absence of cytosine -D-arabinofuranoside (AraC). BrdU-labeled cells appeared as black nuclei (black arrows) in the grayscale background. Nonlabeled neurons had "phase bright" round or oval cell bodies (gray arrows). b: the body of this BrdU-labeled cell also has slight background staining and has rather glial than neuronal morphology (the cell body is indicated by black arrows). Gray arrow in b indicates typical nonstained neuron. c: no BrdU-staining was normally found in AP5- and AraC-treated cultures. d: cultures in this group, however, showed high ChAT-immunostaining. In d, ChAT-immunopositive neurons are indicated by black arrows, ChAT-immunonegative neurons are indicated by gray arrows. Cultures in a and b are 6 days in vitro (scale bar: 10 µm). Cultures in c and d are 18 days in vitro (scale bar: 20 µm).

Cortical cultures

In the preceding work (Belousov et al. 2001), no ACh activity was detected in cultures obtained from the cerebral cortex, either in controls or in cultures subjected to a chronic glutamate receptor blockade. Here no excitatory Ca²⁺ activity was found during synaptic disinhibition in the control cortical neurons (n = 0 of 158) or in cortical neurons chronically subjected to nimodipine (20 µM; n = 0 of 114; Fig. 1F), nifedipine (20 µM; n = 0 of 97), W13 (5 µg/ml; n = 0 of 192), or KN-62 (2.5 µM; n = 0 of 233; all cells were tested in the presence of glutamate receptor antagonists, AP5 and CNQX; Table 1). No change in the number of cholinergic neurons in these cultures was detected; <1% of cells were stained for ChAT both in the control and after various treatment conditions (Fig. 5; Table 2). These data support the idea of the region-specific character of glutamate- and Ca²⁺-dependent regulation of ACh transmission in the CNS.

View larger version (25K): [in this window] [in a new window]   Fig. 5. ChAT immunostaining in cortical cultures. a and b: no change in the number of cholinergic neurons was detected in cortical cultures after a chronic glutamate receptor blockade (b) as compared with the control (a). Most of the cells in both culture conditions were not stained for ChAT. Only a small number of cells were ChAT-immunopositive (black arrow in b).

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Ca²⁺-dependent mechanisms

We have previously described (Belousov et al. 2001) that a long-term blockade of ionotropic glutamate receptors in hypothalamic neuronal cultures dramatically increased the expression of excitatory ACh transmission. The data indicated that during a decrease in glutamate excitation, ACh, which normally exhibits weak activity in the hypothalamus, begins to play the role of the major excitatory neurotransmitter and to support the excitation/inhibition balance. Our findings in hypothalamic cultures described in this article indicate that the blockade of NMDA glutamate receptors, L-type Ca²⁺ channels, or Ca²⁺-dependent signal transduction pathways (CaMK II/IV and PKC) induce cholinergic phenotypic properties (ChAT immunoreactivity) in neurons. These cultures also reveal the expression of ACh activity. In the meantime, the blockade of Ca²⁺-independent intracellular signaling pathways (e.g., PKG) had only negligible effect on the expression of cholinergic functions.

BAPTA is a Ca²⁺-chelating agent that blocks cytoplasmic Ca²⁺ fluctuations (Kao 1994). BAPTA AM is a cell permeable derivative of BAPTA. The AM compound is labile to enzymatic hydrolysis, and once BAPTA AM crosses the cell membrane and enters the neuron, cytoplasmic esterases cleave the AM group. This leads to the formation of the cell impermeable BAPTA that is trapped and accumulated in the cell (Kao 1994). Previous work (Lin and Constantine-Paton 1998) indicated that a short-term (30 min) pretreatment with BAPTA AM had a long-term effect on neurite sprouting and elongation in cultured neurons. In our experiments, increased ACh properties also were found in neurons 2 wk after a 30-min treatment of hypothalamic cultures with BAPTA AM. The data suggested that a short-term pretreatment with BAPTA AM has a prolonged effect in the blockade of intracellular Ca²⁺ increases and induces cholinergic properties in neurons.

Cell depolarization increases Ca²⁺ influx into neurons through both NMDA channels (in the presence of glutamate) and voltage-gated Ca²⁺ channels (Purves et al. 1997). In cell cultures, depolarization is often achieved by the use of a high concentration of KCl in the culture medium. As reported (Bessho et al. 1994; Ciani et al. 1997; Gallo et al. 1987), a chronic treatment of cultures with 20-40 mM KCl does not produce unhealthy neurons, but rather it promotes the survival of cultured cells and is considered as a "trophic condition." In our experiments, all neurons chronically pretreated with KCl were healthy and NMDA-responsive. The depolarization-mediated intracellular Ca²⁺ increase in these neurons compensated for the inactivation of L-type Ca²⁺ channels, calmodulin, and CaMK II/IV and prevented the upregulation of ACh transmission in hypothalamic cultures (Fig. 1E). Additionally, data obtained in the previous research (Belousov et al. 2001) and in this work indicate that although the blockade of non-NMDA receptors itself does not significantly affect ACh transmission, it potentiates the effect of NMDA receptor blockade on the cholinergic upregulation in cultures. This suggests the importance of not only NMDA but also non-NMDA receptors in the control of cholinergic phenotypic properties in hypothalamic neurons. Moreover, the joint blockade of NMDA and non-NMDA receptors more efficiently induces cholinergic properties than does inactivation of only L-type Ca²⁺ channels, intracellular Ca²⁺ fluctuations, calmodulin, CaMK II/IV, or PKC. These data indicate that multiple Ca²⁺-dependent pathways may be involved in ACh regulation during a decrease in glutamate excitation.

Ca²⁺-dependent protein kinases, CaMK II/IV and (presumably) PKC, are involved in the regulation of gene expression in neurons through Ca²⁺/cAMP response element binding protein and Ras/MAPK signaling pathways (Bito et al. 1997; Dolmetsch et al. 2001; Greenberg and Ziff 2001). Therefore the effect of the blockade of calmodulin, CaMK II/IV, and PKC on cholinergic functions in hypothalamic neurons may be due to direct regulation of cholinergic genes by these signal transduction pathways. Additionally, CaMK II/IV and PKC can modulate the activity of voltage-gated Ca²⁺ channels through their phosphorylation (Dzhura et al. 2000; Wu et al. 2001; Yang and Tsien 1993). Therefore a decrease in Ca²⁺ influx into cells through Ca²⁺ channels during inactivation of calmodulin, CaMK II/IV, and PKC may also contribute to the regulation of cholinergic properties in neurons.

Altogether, our data suggest that in cultured hypothalamic neurons, Ca²⁺ influx and intracellular Ca²⁺ oscillations normally prevent the development of cholinergic properties presumably through tonic suppression of molecular mechanisms responsible for the expression of cholinergic functions. However, a decrease in Ca²⁺ influx into cells (through NMDA glutamate receptors or L-type Ca²⁺ channels), inactivation of intracellular Ca²⁺ fluctuations, or downregulation of Ca²⁺-dependent signal transduction pathways (CaMK II/IV and PKC) remove the tonic inhibition and trigger the development of cholinergic phenotype. Ca²⁺-independent PKG signaling pathway does not appear to contribute substantially to the cholinergic regulation in the hypothalamus in vitro. More interestingly, the Ca²⁺-dependent mechanisms that regulate ACh phenotype are region-specific and are not expressed in the cortex in vitro. In our experiments, neither increase in cholinergic activity nor change in the number of cholinergic neurons were detected in cortical cultures under all experimental conditions.

The importance of Ca²⁺ for neuronal differentiation has been demonstrated previously in some regions of the nervous system. Increase in cytoplasmic Ca²⁺ during cell depolarization increased the number of GABAergic neurons (from 10 to 43%) and the level of mRNAs encoding a GABA-synthetic enzyme, glutamic acid decarboxylase 67, in Xenopus spinal cord cultures (Watt et al. 2000). Influx of Ca²⁺ through L-type Ca²⁺ channels dramatically increased the number of dopaminergic neurons in cultured rat sensory ganglia (Brosenitsch et al. 1998). Ca²⁺ increase during cell depolarization also potentiated adrenergic differentiation of cultured sympathetic neurons (Raynaud et al. 1987; Walicke and Patterson 1981). In striking contrast, Ca²⁺ influx decreased the background ChAT activity in rat sympathetic cultures (Raynaud et al. 1987) and prevented cholinergic differentiation in sympathetic neurons cultured in the ACh-inducing conditioned medium (Walicke and Patterson 1981). Moreover, a chronic decrease in Ca²⁺ influx into cultured sympathetic neurons (with 20 mM MgCl₂ or 1 µM D600) increased ACh synthesis by five- to eightfold and raised ACh to catecholamine ratio by three- to fourfold (Walicke and Patterson 1981). Our data on Ca²⁺-dependent regulation of ACh phenotype in the hypothalamus in vitro correlate with those previous observations on cultured sympathetic neurons.

Cholinergic neurons

Our findings demonstrate that only a small number (0.5-0.6%) of ChAT-positive (cholinergic) neurons are present in hypothalamic cultures, both younger (18-21 days in vitro) and older (7 wk in vitro). These data correlate with other observations that also indicated the low expression of cholinergic neurons in the hypothalamus in vitro (Wahle et al. 1993) and in vivo (Arvidsson et al. 1997; Tago et al. 1987; Tinner et al. 1989). Further, no spontaneous or evoked ACh activity is usually detected in control hypothalamic cultures or slices (Belousov and van den Pol 1997; Belousov et al. 2001). The experiments described in this article reveal an increase in the number of ChAT-positive neurons under various treatment conditions that inactivate cellular Ca²⁺ functions or pathways (Table 2). Because cholinergic neurons are almost absent in the hypothalamus in vivo and in vitro, the possibility that Ca²⁺ downregulation increases the expression of ChAT in already committed (existing) cholinergic cells seems unlikely. Noncholinergic cells may potentially express low (undetectable) levels of ChAT. They would not, however, be considered as cholinergic cells if they did not secrete ACh.

An increase in the number of ChAT-positive neurons under these treatment conditions is also unlikely to be due to selective death of noncholinergic cells and survival of cholinergic cells. If this was the case, such an increase in the percentage of cholinergic neurons would have been accompanied by a decrease in the total number of neurons following the treatment. However, the number of neurons in treated cultures (e.g., following glutamate receptor blockade) is not less than but even higher than in the control cultures due most likely to a decrease in glutamate excitotoxicity (Belousov et al. 2001; Obrietan and van den Pol 1995).

Additionally, an increase in the number of ChAT-positive neurons is not likely due to the formation of new cholinergic neurons from progenitor cells that may be present in cultures obtained from the embryonic brain. In our work, the presence of dividing cells was demonstrated using immunostaining for the brominated analogue of thymidine, BrdU. During cell divisions, BrdU is selectively incorporated into newly formed DNA and therefore labels dividing cells (Boccadoro et al. 1986). This may include progenitors, glia, and others. In our experiments, high BrdU incorporation was found only in the first group of cultures that were incubated in the absence of the antimitotic drug, AraC (Fig. 4). Most of these cells were apparently glial cells. In contrast, the second group of cultures that was incubated in the presence of AraC demonstrated little BrdU staining. This group also was chronically treated with AP5, an NMDA receptor antagonist, and exhibited an increased number of cholinergic neurons. Because in the present and previous experiments (Belousov et al. 2001) AraC was always added to all cultures (both control and treated), the upregulation of ChAT is likely due to the induction of cholinergic phenotypic properties in postmitotic noncholinergic neurons and/or due to the formation of ACh neurons from neuroblasts. Moreover, because only a portion of hypothalamic neurons become ChAT-positive under our treatment conditions, the induction of cholinergic properties likely occurs in selective type(s) of neurons. What type of hypothalamic cells becomes cholinergic and whether the phenotypic switch is complete (cells secrete only synaptic ACh) or partial (cells secrete both ACh and an original neurotransmitter) still needs to be determined.

Functional aspects

Increased ACh transmission during a decrease in glutamate excitation may reflect three possible aspects of ACh-glutamate interactions in the CNS. The first aspect is developmental. ACh is considered the major excitatory neurotransmitter in the spinal cord and retina during early development before glutamate begins to play this role during later stages (Feller et al. 1996; Milner and Landmesser 1999). Indirect evidence also suggests high levels of cholinergic activity in the hypothalamus of rodents on embryonic days 15-18 (Naeff et al. 1992; Schambra et al. 1989; Zoli et al. 1995) while glutamate activity in this region is not yet manifested (Chen et al. 1995). The second aspect is compensatory. The upregulation of ACh transmission may represent the re-establishment of excitatory pathways in neuronal circuits that allows neurons to continue excitatory communication even with reduced or absent glutamate activity, which normally is responsible for the fast excitatory communication in the CNS. Consistent with this idea is ACh hyperactivity, hypersensitivity, and the sprouting of cholinergic projections in the brain that accompany the degeneration of glutamate neurons during epilepsy (Correia et al. 1998; Holtzman and Lowenstein 1995; Kish et al. 1988) and early stages of Alzheimer's disease (Geddes et al. 1985; Olney et al. 1997). However, these disease-related changes may also represent the third aspect, pathological, because ACh is believed to contribute to epileptogenesis (Kish et al. 1988) and early stages of Alzheimer (Olney et al. 1997).

Additionally, NMDA receptor antagonists that are used for therapeutic purposes have serious side effects. One such side effect is neurodegeneration in the cerebral cortex (Olney et al. 1991). The neurodegeneration is prevented with AChR antagonists and presumably is due to the hyperactivation of subcortical cholinergic projections to the cortex. In line with this observation is the finding by another group of researchers (Facchinetti et al. 1993, 1994; Virgili et al. 1994) indicating an increase in the catalytic activity of ChAT in some regions of the CNS (cerebellum, spinal cord, etc.) but not in others (cortex, hippocampus) during a chronic administration of NMDA receptor antagonists to rats in vivo. Our experiments in the hypothalamus in vitro demonstrated that glutamate receptor blockade increases ACh properties in hypothalamic and cerebellar cultures but not in cortical cultures (Belousov et al. 2001; this article). The data also suggested that synaptically released ACh becomes excitotoxic when cholinergic transmission increases during a chronic decrease in glutamate activity in the hypothalamus in vitro (Belousov et al. 2001).

In this work, to study the mechanisms of ACh regulation, we used neuronal cultures and pharmacological blockade of either ionotropic glutamate receptors or specific Ca²⁺-dependent signaling pathways. This approach can be considered as a simplified model of the conditions that exist in neuronal circuits at low levels of glutamate excitatory activity. Particularly, the Ca²⁺-dependent mechanisms described here may underlie the regulation of excitatory cholinergic activity in some regions of the CNS during early developmental stages in vivo. They may also contribute to the pathogenesis of neurodegenerative diseases. Additionally, they may be responsible for ACh-dependent side effects mediated by anti-glutamate drugs used for the treatment of a number of neurological disorders.