 |
||||
|
||||
|
||||
|
||||
Previous Article | Next Article
The Journal of Neuroscience, March 15, 2001, 21(6):2015-2027
Acetylcholine Becomes the Major Excitatory Neurotransmitter in the Hypothalamus In Vitro in the Absence of Glutamate Excitation
Andrei B. Belousov1, Bruce F. O'Hara2, and Janna V. Denisova1 1 Department of Cell and Molecular Biology, Tulane University, New Orleans, Louisiana 70118, and 2 Department of Biological Sciences, Stanford University, Stanford, California 94305
ABSTRACT
TOP
ABSTRACT
INTRODUCTION
Glutamate and GABA are two major fast neurotransmitters (excitatory and inhibitory, respectively) in the CNS, including the hypothalamus. They play a key role in the control of excitation/inhibition balance and determine the activity and excitability of neurons in many neuronal circuits. Using neuronal cultures, whole-cell recording, Ca2+ imaging, and Northern blots, we studied the compensatory regulation of neuronal activity during a prolonged decrease in glutamate excitation. We report here that after a chronic (6-17 d) blockade of ionotropic glutamate receptors, neurons in hypothalamic cultures revealed excitatory electrical and Ca2+ synaptic activity, which was not elicited in the control cultures that were not subjected to glutamate blockade. This activity was suppressed with acetylcholine (ACh) receptor antagonists and was potentiated by eserine, an inhibitor of acetylcholinesterase, suggesting its cholinergic nature. The upregulation of ACh receptors and the contribution of ACh to the control of the excitation/inhibition balance in cultures after a prolonged decrease in glutamate activity were also demonstrated. Enhanced ACh transmission was also found in chronically blocked cerebellar but not cortical cultures, suggesting the region-specific character of glutamate-ACh interactions in the brain. We believe that in the absence of glutamate excitation in the hypothalamus in vitro, ACh, a neurotransmitter normally exhibiting only weak activity in the hypothalamus, becomes the major excitatory neurotransmitter and supports the excitation/inhibition balance. The increase in excitatory ACh transmission during a decrease in glutamate excitation may represent a novel form of neuronal plasticity that regulates activity and excitability of neurons during the glutamate/GABA imbalance.
Key words: acetylcholine; glutamate; GABA; hypothalamus; plasticity; excitation/inhibition balance
INTRODUCTION
Synaptic excitation/inhibition imbalance may occur in neuronal circuits under normal conditions such as during an increase or decrease in activity of glutamate excitatory or GABA inhibitory inputs to neurons during development. It may also occur under pathological conditions (e.g., during the degeneration of glutamate or GABA neurons and terminals) or under conditions of pharmacological blockade of glutamate or GABA receptors used for clinical purposes. Such imbalance between the excitation and inhibition may change the activity and excitability of neurons in a circuit or may also disturb circuit function and viability. It has been known for many years that a relative increase in glutamate excitation or decrease in GABA inhibition in slices or cultures obtained from different regions of the CNS lead to glutamate-dependent neuronal hyperexcitability, which, if sustained, causes cell death (Mody et al., 1992
; Choi, 1994
The hypothalamus is the crucial part of the brain that regulates homeostasis throughout the body. It contains >20 active substances that could be released synaptically within this brain structure, including acetylcholine (ACh), dopamine, and several other neurotransmitters and neurohormones. Glutamate and GABA neurons and receptors are also distributed within the hypothalamus, where they control the release of neurohormones, circadian activity, and other hypothalamic functions (van den Pol et al., 1990
MATERIALS AND METHODS
Tissue cultures. Neuronal cultures were prepared from the embryonic (day 18-19) medial hypothalamus, cerebellum, and cortex obtained from Sprague Dawley rats as described (Belousov and van den Pol, 1997a
-D-arabinofuranoside (1 µM). Two groups of cultures were used in most experiments: (1) cultures subjected to a chronic (14-17 d) blockade of NMDA and non-NMDA (AMPA and kainate) ionotropic glutamate receptors with D,L-2-amino-5-phosphonovalerate (AP5; 100 µM) and 6-cyano-7-nitroquinoxaline-2,3-dione (CNQX; 10 µM) ("blocked cultures") and (2) sister control cultures not subjected to a glutamate receptor blockade. In most experiments, AP5 and CNQX were added to the incubation medium of the first group of cultures beginning 4 DIV, and neurons were tested after 14-17 d in block (DIB). Some cultures were chronically incubated in the presence of other receptor antagonists or 20 mM KCl as described in the text. Tissue culture medium was changed twice a week. Only healthy looking cultures were used in experiments; unhealthy cultures were discarded.
Electrophysiology. Standard bathing solution contained (in mM): 158.5 NaCl, 2.5 KCl, 2 CaCl2, 10 HEPES, 10 glucose, and 1 × 10
3 glycine, pH 7.3, 325 mOsm (room temperature, 20-22°C). To perfuse the cells with solutions containing different agonists and antagonists for receptors, a flow pipe perfusion system was used (Belousov and van den Pol, 1997a
The whole-cell current-clamp or voltage-clamp recordings were made with an Axoclamp-2B amplifier (Axon Instruments, Foster City, CA). Glass pipettes were pulled from borosilicate glass capillaries of 2 mm diameter and 0.2 mm wall thickness and filled with an internal solution that included (in mM): 145 potassium methylsulfate, 10 HEPES, 5 MgCl2, 1.1 EGTA, 4 Na-ATP, 0.5 Na-GTP, pH 7.2, 310 mOsm. After they were filled, the electrodes had a resistance of 2-5 M
. The seal resistances were 8-10 G
60 mV. IPSCs were recorded at a holding potential of
Fura-2 Ca2+ digital imaging. Cells were loaded for 30 min at 37°C with 5 µM fura-2 acetoxymethyl ester (Molecular Probes, Eugene, OR) in a standard perfusion solution containing (in mM): 137 NaCl, 25 glucose, 10 HEPES, 5 KCl, 1 MgCl2, 3 CaCl2, 1 × 10
Fura-2 data were calibrated using a commercially available kit (Molecular Probes fura-2 Calcium Imaging Calibration Kit; F-6774). Solutions containing 50 µM fura-2, 100 mM KCl, 10 mM MOPS, and defined Ca2+ concentrations ranging from 0 to 39 µM were placed between two glass coverslips and imaged. Ratios for zero Ca2+ (Rmin) and saturating Ca2+ (Rmax) were determined after background fluorescence from a fura-2-free solution had been subtracted. The values of Rmin and Rmax were substituted into Equation 5 of Grynkiewicz et al. (1985), along with the value Sf/Sb (zero calcium fluorescence at 380 nm divided by saturating calcium fluorescence at 380 nm) and the dissociation constant of fura-2 (Kd = 224 nM). Calibrated Ca2+ data were transferred to a Power Macintosh G3 computer and analyzed with Igor Pro and InStat software.
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
Northern blots. Total RNA was extracted from cultures, and Northern analysis was performed for three nicotinic ACh receptor (nAChR) subunits (
4,
Toxicity assay. Glutamate and ACh excitotoxicity were estimated using a LIVE/DEAD Kit (Molecular Probes). Cultures were stained for 30 min with calcein AM (1 µM), which labels only live cells, and ethidium homodimer-1 (2 µM), which labels only dead cells. The coverslips were then washed and exposed to the appropriate excitation wavelengths: 490 nm (FITC filter) for the analysis of living cells and 580 nm (Texas Red filter) for the analysis of dead cells. A Nikon inverted microscope with a Nikon Super Fluor 20× objective was used for the staining analysis. The fluorescent colors of living (green) and dead (red) cells did not overlap (Fig. 1). The number of live neurons was counted in 36 randomly chosen fields in three coverslips for each test.
View larger version (25K):[in this window]
[in a new window]
Figure 1. Staining of live and dead cells in hypothalamic neuronal cultures. The picture contains two superimposed images that were taken from the same microscope field using two different filters. The fluorescent colors of live (green) and dead (red) cells did not overlap. Live neurons were also recognized by the characteristic round or oval cell body and processes. An arrow points at the dead cell. Scale bar, 20 µm.
Drugs and chemicals. AP5, CNQX, NMDA, cytosine
Data analysis. In electrophysiological and Ca2+ experiments, characteristics of all neuronal responses (e.g., amplitude, frequency, etc.) to pharmacological agents were measured between 45 and 60 sec after the beginning of agent application. Statistical analysis of the experimental data was performed using a Power Macintosh G3 computer and InStat software. Data in all experiments were compared by Student's t test, using paired data when possible. All data are reported as mean ± SEM for the number of neurons indicated.
RESULTS
Neuronal activity in control hypothalamic cultures
High-frequency spontaneous EPSPs were detected in current-clamp recordings from 24 of 25 neurons (96%) in hypothalamic cultures (18-21 DIV). All EPSPs were suppressed by the joint application of ionotropic glutamate receptor antagonists AP5 (100 µM) and CNQX (10 µM) (n = 24 of 24 cells) (Figs. 2a, 3a) and were glutamatergic. EPSPs were not affected by mAChR antagonist atropine and nAChR antagonist mecamylamine (100 µM each; n = 5 neurons tested) (Fig. 3a). The excitatory activity was potentiated by the synaptic disinhibition with the GABAA receptor antagonist bicuculline (50 µM) (Fig. 3a). The frequency of action potentials was 0.5 ± 0.2 spikes/sec before and 4.8 ± 0.6 spikes/sec after the application of bicuculline (n = 5; p < 0.001). In three of five cells, bicuculline elicited a depolarization of the membrane potential (Vm) from
View larger version (33K):[in this window]
[in a new window]
Figure 2. Spontaneous neuronal activity in the control cultures and cultures subjected to a chronic glutamate receptor blockade. Representative current-clamp (a, c, d) and voltage-clamp (b) recordings from two control (a, b) and two blocked (c, d) cells are shown. a, Glutamate-mediated activity recorded in this control neuron was suppressed with glutamate receptor antagonists (AP5/CNQX, 100 and 10 µM) and recovered after washout. b, GABA-mediated activity in another control cell was suppressed with bicuculline (BIC, 50 µM). c, d, ACh-dependent activity recorded in these two chronically blocked neurons was suppressed with either atropine (c, ATR, 10 µM) or mecamylamine (d, MEC, 10 µM). The following applies in all figures: Background, recording made before the applications of antagonist(s); Recovery, recording made after the antagonist(s) washout. AP5 (100 µM) and CNQX (10 µM) were in all mediums in b-d. Calibration bars: 1 sec (horizontal; a-d), 5 mV (vertical; a, c, d), 30 pA (vertical; b).
View larger version (37K):[in this window]
[in a new window]
Figure 3. ACh-mediated electrical activity after a chronic glutamate receptor blockade. Current-clamp recordings are shown. a, This control cell revealed glutamate-dependent EPSPs after the removal of AP5/CNQX (100 and 10 µM) from the incubating medium. This activity was not affected by AChR antagonists (ATR/MEC; 100 µM each) but was potentiated by synaptic disinhibition (BIC; 50 µM). No response to bicuculline was detected in the cell in the presence of glutamate receptor antagonists. b, In this chronically blocked neuron, bicuculline revealed the sustained depolarization and increase in activity that was suppressed with atropine (100 µM). c, Large spontaneous EPSPs induced by bicuculline in another chronically blocked neuron (c-2) were suppressed with atropine (c-3) and recovered after washout (c-4). After bicuculline washout, activity recovered (c-5) to its initial low level (c-1). d, Bicuculline-induced activity in this blocked neuron (d-1) was not affected by atropine (100 µM; d-2) but was suppressed by mecamylamine (100 µM; d-3) or TTX (1 µM; d-4) and recovered after washout (d-5). Applications of antagonists in a and b are indicated by the bars above the recordings. Dashed lines in a also indicate the time periods of AP5 and CNQX application. The dotted line in b represents the background Vm level. The dashed line in d is the beginning of antagonist(s) application. AP5 (100 µM) and CNQX (10 µM) were in all solutions in b-d. Calibration bars are shown below the recordings.
Ca2+ imaging experiments revealed spontaneous glutamate-mediated intracellular Ca2+ oscillations in 11 of 200 (6%) control hypothalamic neurons (Fig. 4a). The frequency of oscillations was 0.051 ± 0.003 Hz (n = 11; range 0.04-0.06 Hz). The average amplitude was 81.5 ± 10.2 nM Ca2+ (n = 11), as measured from the background Ca2+ level that usually was 50-75 nM in different cells. Ca2+ rises were completely suppressed in the presence of glutamate receptor antagonists (AP5 and CNQX, 100 and 10 µM, respectively; n = 11). They were not affected, however, by atropine and mecamylamine (100 µM each) (Fig. 4a). After the joint application of AChR antagonists, the frequency of Ca2+ oscillations was 0.052 ± 0.003 Hz and the amplitude was 92.4 ± 12.6 nM Ca2+ (n = 11; no significant difference from the control). The level of intracellular Ca2+ increased dramatically in almost all tested neurons (n = 196 of 200; 98%) in the presence of bicuculline (50 µM) (Fig. 4a). In these cells, the Ca2+ increase was 194.2 ± 5.9 nM from the initial background level. Oscillations of intracellular Ca2+, synchronous between many cells, were common in these conditions. Because the bicuculline-mediated Ca2+ increases were not detected in the presence of glutamate receptor antagonists (Fig. 4a) (n = 436 cells tested), they likely represented the disinhibition of glutamate excitatory activity in cultured neurons.
View larger version (30K):[in this window]
[in a new window]
Figure 4. ACh-mediated Ca2+ activity in hypothalamic cultures after a chronic glutamate receptor blockade. Ca2+ digital imaging data obtained from one control (a) and six chronically blocked (b-f) neurons are presented. a, Typical control cell revealed glutamate-dependent Ca2+ rises after the removal of AP5/CNQX (100 and 10 µM) from the incubating medium. This activity was not affected by AChR antagonists (ATR/MEC; 100 µM each) but was potentiated by synaptic disinhibition (BIC; 50 µM). No response to bicuculline was detected in the cell in the presence of glutamate receptor antagonists. b, In this blocked neuron, bicuculline increased the level of intracellular Ca2+ that was suppressed by atropine and mecamylamine (ATR/MEC; 10 µM each) applied jointly. c, Two cells in one culture responding differently to the separate application of atropine and mecamylamine (100 µM each). d, Ca2+ activity was blocked with TTX (1 µM). e, The activity was not affected by metabotropic glutamate receptor antagonists (mGluR-A; see Results for details; ATR; 10 µM). f, Ca2+ activity was also induced by eserine (10 µM). AP5 and CNQX were in all solutions in b-f. Bicuculline (1 µM) was also in all solutions in f. Calibration bars: 2 min (horizontal) and 75 nM Ca2+ (vertical) for all recordings.
Inhibitory neuronal activity was also detected in the control hypothalamic cultures. Voltage-clamp recordings revealed spontaneous IPSCs in 12 of 20 neurons (60%) that were completely suppressed by bicuculline (50 µM) and therefore were GABA mediated (Fig. 2b) (AP5 and CNQX were in all solutions; holding potential was
Upregulation of ACh transmission in hypothalamic cultures after a chronic glutamate receptor blockade
Chronic (14-17 d) glutamate receptor blockade with AP5 (100 µM) and CNQX (10 µM) was maintained in some hypothalamic cultures starting 4 DIV ("blocked cultures"). Cultures of the same age (18-21 DIV) not subjected to a glutamate receptor blockade served as control. When characteristics of neuronal activity were measured and compared between blocked and control cultures, no significant difference was detected in the following variables: Vm was
Surprisingly, however, compared with the control cultures in which all excitatory activity was suppressed in the presence of AP5/CNQX (see above), recordings revealed low-amplitude (3-10 mV) spontaneous EPSPs at a frequency of 0.5-5 Hz in 54% of blocked neurons (n = 45 of 84) (Fig. 2c,d). The EPSPs were completely and reversibly suppressed by the mAChR antagonist atropine (10-100 µM) in 8 of 10 neurons (Fig. 2c) or by the nAChR antagonist mecamylamine (10 µM) in 2 of 10 cells (Fig. 2d). This suggested that they were cholinergic in nature. Additionally, although no excitatory electrical activity was seen in control cultures after synaptic inhibition was blocked with bicuculline (50 µM) when AP5 and CNQX were present in the medium (n = 19) (Fig. 3a), large amplitude EPSPs, sustained depolarization, and a dramatic increase in action potentials were detected after synaptic disinhibition in 29 of 32 chronically blocked neurons (Fig. 3b-d). The frequency of action potentials was 0.3 ± 0.1 and 3.9 ± 0.5 spikes/sec before and after the application of bicuculline, respectively (n = 29; p < 0.0001). In the presence of bicuculline, spontaneous large EPSPs (16-30 mV; 0.2-0.25 Hz) were found in 17 of 29 (59%) of neurons (Fig. 3c). The rest of the cells (41%) revealed a sustained depolarization of Vm from the background level of
The bicuculline-induced electrical activity in blocked neurons was associated with a dramatic increase in the level of intracellular Ca2+ (Fig. 4b-e), which was not detected in control neurons in the AP5/CNQX-containing medium (Fig. 4a). In chronically blocked cultures, bicuculline increased intracellular Ca2+ level in 1938 of 2335 neurons (83%). The amplitude of Ca2+ increase was in the range of 24-770 nM (average 203.2 ± 2.0 nM; n = 1938) as measured relative to the initial background Ca2+ level. During bicuculline application, many of these neurons (52% of 1938) revealed Ca2+ oscillations that were usually regular and synchronized between all cells in the microscope field and were in the range of 0.02-0.06 Hz (Fig. 5). Ca2+ increase in other cells (48% of 1938) was steady, as shown in Figure 4b. The bicuculline-induced Ca2+ activity was blocked or significantly (>50%) suppressed with the joint application of atropine and mecamylamine in 67% of 235 neurons at concentrations of 10 µM and in 94% of 161 neurons at concentrations of 100 µM (Fig. 4b, Table 1). Atropine alone blocked or reduced the Ca2+ activity in 39% (of 144; 10 µM) and 69% (of 269; 100 µM) of neurons, and mecamylamine alone did that in 24% (of 90; 10 µM) and 48% (of 65; 100 µM) of cells. When atropine (100 µM) and mecamylamine (100 µM) were separately applied to the same cells, activity in 5 of 65 neurons (8%) was suppressed with mAChR but not nAChR antagonist (Fig. 4c-1), whereas in 3 of 65 cells (5%) the effect was the opposite (Fig. 4c-2).
View larger version (43K):[in this window]
[in a new window]
Figure 5. ACh-mediated synchronous intracellular Ca2+ oscillations in blocked hypothalamic neurons. The activity of all seven cells was recorded simultaneously from one microscope field. BIC, 50 µM; ATR, 10 µM; MEC, 10 µM. AP5 and CNQX were in all solutions. Calibration bars: 2 min and 100 nM Ca2+ for all recordings.
View this table: [in this window]
[in a new window]
Table 1. Effect of some pharmacological agents on bicuculline-mediated Ca2+ increases in hypothalamic cultures subjected to a chronic glutamate receptor blockadea The data described above supported the idea that the bicuculline-induced Ca2+ activity in blocked cultures was of a cholinergic origin. Importantly, the activity was also suppressed by TTX (1 µM) in all electrical recordings (n = 5 neurons) (Fig. 3d-4) and in most of the Ca2+ imaging experiments (94% of 228 cells) (Fig. 4d), suggesting the synaptic release of ACh from cultured neurons. The activity was not significantly affected, however, by the antagonists of metabotropic glutamate receptors (AIDA, 100 µM; EGLU, 100 µM; and MSPG, 100 µM; applied together; decrease only in 3% of 303 cells) (Fig. 4e), P2 purinoreceptors (suramin, 100 µM; and PPADS, 100 µM; 0% of 122 cells; data not shown), and
Eserine (10 µM), an inhibitor of acetylcholinesterase (enzyme that hydrolyzes ACh), also induced Ca2+ increases in 57 of 106 (54%) chronically blocked cells (Fig. 4f). The average amplitude of Ca2+ increase during eserine application was 121.2 ± 10.0 nM, as measured from the initial background level (n = 57; p < 0.0001, calculated relative to zero; 1 µM bicuculline was in all solutions). In the meantime, in the control cultures, no Ca2+ increase was detected in neurons (n = 0 of 132; data not shown) under similar conditions. These data provided further evidence for the upregulation of ACh transmission during a chronic decrease of glutamate excitation.
ACh regulation in cerebellar and cortical cultures
In our experiments, we also tested the expression of ACh-mediated Ca2+ activity in the control (18 DIV) and chronically blocked (14 DIB/18 DIV) cultures obtained from the cerebellum and cortex. Almost all (98%; n = 248 of 253) of the chronically blocked neurons but none (0%; n = 0 of 211) of the control cerebellar cells increased the level of intracellular Ca2+ (by ~190 nM) or revealed synchronous intracellular Ca2+ oscillations (Fig. 6a) during the application of bicuculline (AP5 and CNQX were in all mediums). This bicuculline-induced activity was blocked or suppressed (>50%) in 81% (n = 201 of 248) neurons by the joint application of atropine and mecamylamine (10 µM each) (Fig. 6a). In cortical cultures, however, bicuculline-induced activity was detected in neither control neurons (n = 237 tested) nor neurons after a chronic glutamate receptor blockade (n = 254) (Fig. 6b).
View larger version (26K):[in this window]
[in a new window]
Figure 6. Region-specific character of glutamate-dependent regulation of ACh transmission. a, ACh-mediated synchronous Ca2+ oscillations were detected in many cerebellar neurons after a chronic decrease in glutamate excitation. Activity of two representative cerebellar cells in a (a-1, a-2) were recorded simultaneously from the same microscope field. ATR, 10 µM; MEC, 10 µM. b, No excitatory activity was detected in chronically blocked cortical neurons. AP5 and CNQX were in all solutions except for the NMDA-containing (10 µM) but Mg2+-free solution in b that was used to confirm that the cell was healthy and responsive. Calibration bars: 1 min and 75 nM Ca2+ for all recordings.
Upregulation of ACh receptors
Two major types of AChRs have been determined previously in the mammalian CNS, including the hypothalamus: ionotropic nAChRs (directly coupled to an ionic channel) and metabotropic mAChRs (coupled to GTP-binding proteins) (Seguela et al., 1993
View larger version (54K):[in this window]
[in a new window]
Figure 7. Upregulation of AChRs after a chronic glutamate receptor blockade. Data from Ca2+ imaging (a, b, e, f) and Northern blot analyses (c, d) are presented. a, b, Responses of two representative control (a) and chronically blocked (b) neurons to nicotine (NIC; 10 µM) and muscarine (MUSC; 10 µM). c, d, Expression of mRNAs encoding five mAChRs (c) and three subunits of nAChRs (d) in the control (open bars; n = 4 coverslips) and chronically blocked (filled bars; n = 5 coverslips) cultures. Each bar shows the mean and SE. Significance of differences (Student's t test) relative to the control: *p < 0.05, **p < 0.001, ***p < 0.0001. e, f, Bicuculline-induced activity in two chronically blocked neurons was suppressed by pirenzepine (e, PIRENZ; 5 µM) or
The significant upregulation of mRNAs encoding two of five subtypes of mAChRs (Fig. 7c, m1 and m3) and two of three subunits of nAChRs (Fig. 7d,
Some insights on the mechanisms of glutamate-dependent ACh regulation
To further characterize the mechanisms responsible for the development of ACh activity in hypothalamic cultures during chronic glutamate receptor blockade, we used the partial blockade of NMDA and non-NMDA receptors with 20 µM AP5 and 2 µM CNQX for 2 weeks. Although in control cultures such concentrations of AP5 and CNQX were not sufficient to completely block the excitatory effects of externally applied glutamate (~50% blockade; n = 78 cells) (Fig. 8a), ACh-mediated Ca2+ activity was still detected in 34% of 80 cells chronically incubated in the presence of 20 µM AP5 and 2 µM CNQX (Fig. 8b, Table 2). A chronic (14 d) blockade of only NMDA glutamate receptors was sufficient for the upregulation of ACh activity in neurons: 73% of 90 cells and 69% of 106 cells expressed ACh-mediated activity in cultures subjected to NMDA receptor antagonists AP5 (100 µM) (Fig. 8c) and CPP (5 µM) (data not shown), respectively. A blockade of only non-NMDA receptors with 10 µM CNQX was not critical for the development of this activity (Fig. 8d) (only 1% of 233 neurons in such cultures revealed ACh excitation). Elevated potassium (20 mM), which is known to increase Ca2+ influx to cells through L-type voltage-gated Ca2+ channels (Bessho et al., 1994
View larger version (51K):[in this window]
[in a new window]
Figure 8. Development of ACh activity in hypothalamic neuronal cultures. Representative recordings from Ca2+ imaging experiments are shown. a, Amplitude of responses of control hypothalamic neurons to externally applied glutamate (10 µM) was reduced by ~50% in the presence of 20 µM AP5 and 2 µM CNQX (20/2). All glutamate-mediated activity recorded in the absence of glutamate receptor antagonists (0/0) was suppressed by 100 µM AP5 and 10 µM CNQX (100/10). Arrows indicate the time of antagonist introduction. b, ACh activity in culture grown in the presence of 20 µM AP5 and 2 µM CNQX. c, ACh activity in culture grown in the presence of 100 µM AP5 alone. d, Culture grown in the presence of 10 µM CNQX alone. e, Neurons grown in the presence of AP5, CNQX, and KCl (20 mM) did not reveal ACh activity, but they did express low amplitude (15-20 nM) Ca2+ oscillations synchronized between all cells in the microscope field (two representative cells are shown). f, Culture grown in the presence of 1 µM TTX. g, ACh-mediated Ca2+ activity in cultures at different levels of glutamate excitation (see Results for explanation). h, Culture after 10 DIV/6 DIB. i, Culture after 74 DIV/14 DIB. All recordings in b-i were done in the presence of AP5 (100 µM) and CNQX (10 µM) in all solutions. NMDA (10 µM) solution in d, f, g-1, and g-3 did not contain Mg2+, AP5, and CNQX. b-i, BIC, 50 µM; ATR, 100 µM; MEC, 100 µM. Calibration bars: 2 min and 75 nM Ca2+ for all recordings.
View this table: [in this window]
[in a new window]
Table 2. Acetylcholine-dependent Ca2+ activity in hypothalamic cultures raised under various cell culture conditionsa We tested the expression of ACh transmission in cultures at different levels of glutamate activity (Fig. 8g). Cultures in six coverslips were subjected for 14 d to glutamate receptor blockade with AP5 and CNQX. Two of those six cultures were tested and revealed ACh-mediated Ca2+ rises in 85% of 169 cells (Fig. 8g-2). In the remaining four coverslips, the concentrations of both AP5 (100 µM) and CNQX (10 µM) in the culture medium were then decreased by 20% each successive day. After 5 d, AP5 and CNQX were removed completely, and cells were cultured for 10 additional days in the absence of the glutamate receptor antagonists. When cultures in two of those four coverslips were tested, they expressed spontaneous glutamate-mediated EPSPs (n = 3 cells; data not shown) and did not reveal any ACh-dependent Ca2+ activity (n = 119) (Fig. 8g-3), as did their sister control cultures that had never been subjected to chronic glutamate receptor blockade (174 cells in two coverslips; 33 DIV) (Fig. 8g-1). Cultures in the other two of four coverslips were again subjected to a chronic glutamate receptor blockade for the next 10 d. When tested, ACh-mediated Ca2+ activity was detected in these cultures, although in a lower percentage of neurons (54% of 105 neurons) (Fig. 8g-4). This showed the dynamic bi-directional plasticity of ACh transmission at different levels of glutamate excitation.
The chronic presence of glutamate receptor antagonists in the incubating medium is known to slow the rate of natural cell death and to increase neuronal survival in cell cultures, which is attributable most likely to a decrease in glutamate excitotoxicity (Obrietan and Van den Pol, 1995
We tested whether ACh excitation can develop in cultures after their maturation or whether only immature cultures develop this activity. In these experiments, hypothalamic cultures were subjected to a chronic (14 d) glutamate receptor blockade starting at 60 DIV [i.e., well after maturation of synaptic connections in rat hypothalamic cultures (van den Pol et al., 1998
ACh supports the excitation/inhibition balance
A decrease in GABA activity in neuronal cultures or cultured slices obtained from different regions of the CNS can cause an imbalance in synaptic excitation/inhibition and, if sustained, cell death. Application of the GABAA receptor antagonists bicuculline (100 µM) and picrotoxin (500 µM) for 3 d to cultured hippocampal slices was found previously to cause a glutamate-dependent neurodegeneration that could be prevented with the NMDA and non-NMDA receptor antagonists MK-801 and CNQX (Thompson et al., 1996
View larger version (53K):[in this window]
[in a new window]
Figure 9. In the absence of glutamate excitation, ACh supports the excitation/inhibition balance in hypothalamic cultures. The number of live neurons per microscope field (p.m.f.) was calculated in control hypothalamic cultures (a, c), in blocked hypothalamic cultures (b, d), and in blocked cortical cultures (e). The first column in all graphs represents no treatment. a-2, b-2, e-2, Cultures treated for 4 d with bicuculline (100 µM) and picrotoxin (500 µM). a-3, b-3, e-3, Cultures treated for 4 d with bicuculline (100 µM), picrotoxin (500 µM), atropine (100 µM), and mecamylamine (100 µM). c-2, d-2, Cultures treated for 4 d with carbachol (50 µM). c-3, d-3, Cultures treated for 4 d with carbachol (50 µM), atropine (100 µM), and mecamylamine (100 µM). All control cultures were stained on 22 DIV; blocked cultures were stained on 22 DIV/18 DIB. Significance of differences relative to the corresponding group of cells that were not treated: ***p < 0.0001, **p < 0.001.
We then tested whether after a chronic glutamate receptor blockade ACh supports the excitation/inhibition balance and, namely, whether a prolonged imbalance between ACh excitation and GABA inhibition also leads to cell death. To test this hypothesis, bicuculline (100 µM) and picrotoxin (500 µM) were added for 4 d to some chronically blocked hypothalamic cultures starting 18 DIV/14 DIB. Other blocked cultures were not treated with GABAA receptor antagonists. In 4 d (22 DIV/18 DIB), all cultures were stained using the Live/Dead Kit, and the number of live cells was analyzed. In cultures not treated with GABA antagonists, 54.4 ± 3.8 live neurons p.m.f. were found (Fig. 9b, b-1). Treatment of neurons for 4 d with bicuculline and picrotoxin significantly decreased the number of live neurons to 19.5 ± 3.4 p.m.f. (p < 0.0001) (Fig. 9b, b-2). The neurotoxicity mediated by GABAA receptor antagonists was prevented with atropine and mecamylamine (100 µM each; 61.5 ± 4.5 cells p.m.f) (Fig. 9b, b-3), suggesting its cholinergic nature. In another experiment, when the AChR agonist carbachol (50 µM) was applied to blocked cultures for 4 d, it decreased the number of live neurons from 52.8 ± 2.4 to 17.3 ± 2.2 p.m.f. (p < 0.0001) (Fig. 9d, d-1, d-2). This was prevented with atropine and mecamylamine (100 µM each; 51.7 ± 3.5 p.m.f) (Fig. 9d, d-3). Meanwhile, in control cultures (not subjected to a chronic glutamate receptor blockade) no effect of carbachol on cell survival was detected: 44.9 ± 3.43, 46.8 ± 3.09, and 45.9 ± 3.0 live neurons p.m.f. were found in nontreated cultures, carbachol-treated cultures, and cultures treated with carbachol and AChR antagonists, respectively (Fig. 9c).
In these experiments, we also tested cortical cultures subjected to a chronic glutamate receptor blockade. GABAA receptor antagonists applied to these cultures for 4 d did not elicit neurodegeneration. The number of live neurons p.m.f. was 32.2 ± 2.1, 33.1 ± 2.3, and 34.4 ± 2.0 in nontreated cultures, cultures treated with GABAA receptor antagonists, and cultures treated with AChR and GABAA receptor antagonists, respectively (Fig. 9e). The data suggest that in the absence of glutamate excitatory activity, ACh supports the excitation/inhibition balance in hypothalamic but not cortical cultures.
DISCUSSION
Taken together, our experiments demonstrated glutamate-dependent bi-directional regulation of cholinergic transmission in cultured neurons: a dramatic increase in excitatory ACh activity and upregulation of AChRs after a chronic blockade of glutamate neurotransmission and a decrease in cholinergic excitation in the presence of glutamate activity. Upregulation of ACh excitation after a decrease in glutamate activity was detected in hypothalamic and cerebellar cultures but not in cultures obtained from the cortex. These data parallel previous observations on rat pups in vivo (Facchinetti et al., 1993
Several different factors can cause a decrease in glutamate excitatory transmission in the CNS. Decreased glutamatergic functions in the cortex and striatum have been postulated to be a significant factor in the pathophysiology of schizophrenia (Riederer et al., 1992
In neuronal circuits, a decrease in glutamate activity usually triggers the compensatory mechanisms that are intended to increase glutamate functions of neurons: upregulation of postsynaptic glutamate receptors, regeneration and sprouting of glutamate terminals, etc. (Follesa and Ticku, 1996
It is important to note that neuronal regeneration and development share many mechanisms and regulatory molecules that regulate axonal outgrowth and pathfinding, formation of synaptic connections, trophic interactions between synapses and target cells, and changes in neurotransmitter release and reception (Nicholls et al., 1992
Importantly, our experiments also revealed the upregulation of excitatory cholinergic neurotransmission even after a partial (~50%) decrease in glutamate activity. Such culture conditions are probably more relevant to pathological or developmental conditions in vivo than to a total decrease in glutamate excitation.
Several possible mechanisms may potentially account for the upregulation of ACh transmission after a decrease in glutamate excitation. One possibility is a direct Ca2+-dependent regulation of cholinergic gene expression described previously for excitable cells (Walke et al., 1994
In conclusion, the upregulation of excitatory ACh transmission detected in our experiments on the hypothalamus in vitro appears to represent a novel form of neuronal plasticity that regulates the activity and excitability of neurons and supports the excitation/inhibition balance in many neuronal circuits in the CNS during a long-term decrease in glutamate excitation.
FOOTNOTES Received June 18, 2000; revised Dec. 20, 2000; accepted Jan. 4, 2001.
This research was supported by Tulane University funds (A.B.B.), Board of Regents Support Fund (A.B.B.), and National Institutes of Health Grant DA00187 (B.F.O.). We thank Dr. Anthony N. van den Pol and Dr. Hilary Srere for helpful discussions in the early phases of this project, and Steve Wiler and Vinh Cao for technical support.
Correspondence should be addressed to Dr. Andrei B. Belousov, Department of Cell and Molecular Biology, 2000 Percival Stern Hall, Tulane University, New Orleans, LA 70118. E-mail: belousov{at}tulane.edu.
REFERENCES
