| The realization that different behavioral and perceptual states of the brainare associated with different brain rhythms has sparked growing interest in theoscillatory behavior of neurons. With the emergence of new techniques, studieshave shown that temporally and spatially organized activity among distributedneuronal populations often takes the form of synchronous oscillations. Thesesynchronous oscillations may evolve to dynamically control the grouping ofneurons into organized assemblies. Recent research has revealed a closeassociation between electrical oscillations and resonance in neurons. It has beensuggested that resonance underlies oscillatory behavior in neural networks.Resonance is the ability of neurons to respond selectively to inputs at preferredfrequencies, which can serve as a substrate for coordinating network activityaround a particular frequency. Resonance dramatically affects the ability ofnetworks to produce oscillatory patterns of activity. Gimbarzevsky has studiedthe property of resonance since the1980s. There is accumulating evidence thatneurons in different brain areas have prominent resonant properties, such as inthe neocortical cortex, entorhinal cortex, prefrontal cortex, hippocampus and thalamus.Studies in recent years have shown that there are several types ofoscillatory phenomena in the basal ganglia. Investigations in humans andanimals have demonstrated the existence of a range of oscillatory activity in thevarious nuclei of the basal ganglia, and the oscillatory activity is believed toplay an important role in both the physiology and pathophysiology of thissystem. Studies using electroencephalography and magnetoencephalographyalso support the theory that basal ganglia engage in synchronous oscillatoryactivity. However, little is known about the actual mechanisms underlying thegeneration of neuronal oscillatory activity in basal ganglia. Now, it is possible toimage the resonance underlying oscillatory behavior in basal ganglia.A critical role of the subthalamic nucleus (STN) in the control ofmovement has been proposed based in part on the observation that its lesion orhigh-frequency electrical stimulation is highly effective in alleviating theakinesia, rigidity and tremor of Parkinson’s disease (PD). PD, which is causedby the progressive loss of dopaminergic neurons in the substantia nigra, isassociated with hyperactivity of neurons in the STN as detected by metabolicand electrophysiological studies in parkinsonian patients and in parkinsonianrats and monkeys. In this paper, we studied the resonance characteristics of STNneurons and its correlation with the dopamine receptor using whole-cellpatch-clamp recordings in rat brain slices.MethodsTissue preparationCoronal slices containing STN (350-400μm thick) were prepared frommale young (13-18days of age) Sprague-Dawley rats as described previously.Rats were euthanized with isoflurane anesthesia in accordance with the Principles of Medical Laboratory Animal Care issued by the NationalMinistry of Health. The brain was removed rapidly and slices were cut with avibratome in ice-cold artificial cerebrospinal fluid (ACSF) solution of thefollowing composition (in mM): NaCl (126), KCl (2.5), CaCl2(2.4), MgCl2(1.2), NaH2PO4(1.2), NaHCO3(19), and glucose (11), gassed with95%O2and5%CO2(pH7.35-7.45). The slices were incubated in95%O2/5%CO2-equilibrated ACSF and kept at34°C for at least one hour. A slice was thenplaced on the recording chamber and submerged in a continuously flowingACSF (2ml/min) gassed with95%O2and5%CO2and heated to34°C, exceptin a few experiments performed to test the temperature dependence of resonance,when the medium was heated from30°C to34°C and38°C. Using a10×objective for visual guidance, STN was readily identified as ovoid gray matterimmediately medial to the cerebral peduncle.Electrophysiological recordingIn some experiments, tetrodotoxin (TTX,1μM) was applied to blockaction potentials. Individual neurons were visualized (40×water immersionobjective) using differential interference contrast infra-red microscopy. STNneurons were not consciously selected on the basis of somatic size or shape.Whole-cell recordings were made with borosilicate glass pipettes of3-7MΩresistance containing (in mM): potassium gluconate (125), NaCl (10), MgCl2(2.0) CaCl2(1.0), EGTA (10), HEPES (10), ATP (2.0), GTP (0.3)(pH7.3,osmolarity285-295). Recorded electrical signals were amplified with anAxopatch-700B amplifier. Data were acquired using a computer with thedigidata1440A acquisition system. All potentials were corrected online for thejunction potential using the multiclamp700B commander software. Onlyneurons with a stable resting membrane potential more negative than-55mV and stable action potential amplitudes were used for recording.Recording and analysis of electrical resonanceThe impedance amplitude profile (ZAP) method was used to characterizethe electrical resonance behavior of the neurons as previously described. TheZAP current waveform was a swept-sine-wave current with constant amplitudeand linearly increasing frequency (0-18Hz for20s) generated by computer. Adirect current was firstly used to hold neurons near a desired membranepotential, then the ZAP current was injected, and the resulting voltage responsewas recorded. The amplitude of the ZAP current was adjusted to keep theperturbation of the membrane potential close to±10mV (peak to peak) to avoidfiring of action potentials. Resonance was manifest as a distinct andreproducible peak in the voltage response at a certain frequency. The Q value,the ratio of the impedance at the resonance peak to the impedance at0.5Hz, wasused to quantify the strength of the resonance. To plot the magnitude of the cellimpedance as a function of frequency, data were transformed into the frequencydomain through fast Fourier transformation (FFT). Impedance (Z) is a complexnumber defined as Z=FFT (V)/FFT (I). The magnitude of impedance wasplotted against frequency to give an impedance curve. The resonance frequency(fres) could be read at the peak of the impedance profile.Morphology of electrophysiologically characterized neuronsSome STN neurons were labeled by adding Lucifer Yellow (0.15%) in thepipette solutions. During the course of recording, Lucifer Yellow diffused fromthe pipette into the cell. A period of30min of simple diffusion was sufficient toobtain complete labeling. Stained neurons were visualized with a laser scanningconfocal microscope. Drug applicationAll drugs were dissolved in aqueous stock solutions with the exceptions ofclozapine and haloperidol, which were dissolved in dimethyl sulfoxide, andsulpiride which was dissolved in dehydrated alcohol. Each stock solution wasdiluted at least1:1000in perfusate immediately prior to its use. Dimethylsulfoxide and dehydrated alcohol, diluted1:1000in ACSF, had no side effect.Most drugs were applied by superfusion. Approximately30s were required forthe drug solution to enter the recording chamber; this delay was due to passageof the perfusate through a heat exchanger. Complete exchange of the bathsolution occurred within3min. In some experiments, drugs (Lucifer Yellow,GDP-β-S and GTP-γ-S) were added directly to the pipette internal solution fromwhich they diffused spontaneously into the cell.ResultsMembrane resonance of STN neuronsSTN neurons were characterized by obvious voltage sags in response to aseries of hyperpolarizing current injections and anodal break rebounddepolarization. Most of the STN neurons tested exhibited membrane resonance,which manifested as a distinct and reproducible hump in the voltage response toZAP current injection. This resonant hump occurred at2.67±0.29Hz at34°Cwhen the holding potential was at-70mV, and the Q value was1.093±0.0202at the same condition.To test whether the resonant humps reflectedfrequency-dependent as opposed to time-dependent membrane properties, aninverted ZAP current with the same amplitude and linearly decreasing frequency(from18to0Hz for20s) was injected. The fresand Q values yielded from theinverted protocol were2.64±0.31and1.090±0.0213, respectively (n=6). Thevalues were similar to those generated from the standard ZAP protocol in the same neurons (P>0.05and0.05, paired t-test), suggesting that the resonancedid not depend on the input waveform.Temperature dependence of resonance in STN neuronsTo test the temperature dependence of the resonance, the temperature ofperfusion solutions in the recording chamber was raised from30°C to34°C and38°C at a holding potential of-70mV. The resonant hump shifting as thetemperature rises. The fresshifted from1.99±0.36Hz at30°C, to2.67±0.29Hz, and4.07±0.40Hz at34and38°C, respectively. Higher temperaturessignificantly shifted the resonant hump to a higher frequency (n=8; P <0.05,one-way ANOVA).Voltage dependence of resonance in STN neuronsTo determine the voltage dependence of the electrical resonance, cells wereheld at different membrane potentials from-50to-90mV (10mV increment).TTX was added to perfusion solutions to avoid depolarization-evokedspikes.The voltage responses to the ZAP current injections exhibited distinctmembrane resonance at holding potentials from-60to-90mV. At-50mV, nosignificant resonance was observed. Both the fresand Q values werevoltage-dependent. The freswas2.01±0.25Hz at-60mV, and increased to3.09±0.32Hz at-90mV (n=10). The Q value was1.009±0.015at-50mV,suggesting a very weak resonance occurs at this potential in a handful ofneurons. The Q value reached its peak of1.092±0.021at-70mV (n=10).These results indicate voltage-dependent membrane resonance of STN neurons.Hyperpolarization-activated cation current (Ih) in STN neurons and itsrole in resonanceSTN neurons exhibited significant Ihin voltage-clamp mode or distinctvoltage sags in response to hyperpolarizing current injections in current-clamp mode. Ihand voltage sags could be completely blocked by ZD7288(20μM)(n=12; P <0.05, two-way ANOVA). Therefore, ZD7288was used to test the roleof Ihin the resonance of STN neurons. The ZAP currents were injected beforeand after application of20μM ZD7288at-70mV. The resonant hump wasdistinct in the absence of ZD7288and completely abolished in the presence ofZD7288. The results suggest that Ihis essential for the resonance in STNneurons.Resonance-mediated frequency-selective coupling of inputs and firingLarge-amplitude ZAP current was injected to evoke action potentials atholding potentials of-70mV. Action potentials fired most readily as the ZAPinput swept through frequencies near fres. We also used single-frequency sinewave current to evoke firing at the same potentials. Neurons preferred to firewhen the input frequency was near fres. After application of ZD7288(20μM),the spikes evoked by both ZAP current and single-frequency sine waves currentarose readily at the lowest frequencies. In summary, the frequency preference ofneurons resulted in a preferential coupling at frequencies near fresbetween inputsand firing, which could be blocked by ZD7288.Dopamine receptor pharmacologyThe ZAP currents were injected before and after application of clozapine(50μM) or haloperidol (30μM), the effective antagonist of D1and D2. Theresonant hump was distinct in the absence of clozapine or haloperidol andcompletely abolished in the presence of clozapine or haloperidol. The effect ofclozapine was found to be mimicked by sulpiride, a D2antagonist, but not bythe D1antagonist SCH-23390. So we can conclude that the D2antagonistsulpiride, but not the D1antagonist SCH-23390, blocked resonance of STNneurons. Effect of sulpiride on the frequency preference and Ihof STN neuronsLarge-amplitude ZAP current was injected to evoke action potentials atholding potentials of-70mV. Action potentials fired most readily as the ZAPinput swept through frequencies near fres. We also used single-frequency sinewave current to evoke firing at the same potentials. Neurons preferred to firewhen the input frequency was near fres. After application of sulpiride (1μM), thespikes evoked by both ZAP current and single-frequency sine waves currentarose readily at the lowest frequencies. Ihof STN neurons could be significantlyinhibited by sulpiride (1μM)(n=10; P <0.05, two-way ANOVA)G-protein is involved in the effect of sulpiride on STN neurons.In order to test for involvement of G proteins in effect of sulpiride onmembrane resonance of STN neurons. we compared the voltage response toZAP current injection with patch pipettes that contained either0.5mM GDP-β-Sor GTP-γ-S in the absence or presence of sulpiride. When the pipette solutioncontained the non-hydrolyzable G-protein inhibitor GDP-β-S and the ACSF donot contained the sulpiride, the resonant hump was completely abolished. Incontrast, when the pipette solution contained the GTP analogue GTP-γ-S and theACSF contained the sulpiride, the resonant hump was distinct. On the otherhand, similar to that on resonance, the effect of sulpiride on frequencypreference of STN neurons firing was mimicked in the presence of GDP-β-S inthe pipette solution, and was completely abolished in the presence of GTP-γ-S inthe pipette solution. These data suggest that G-protein is involved in effect ofsulpiride on membrane resonance and frequency preference of STN neurons.ConclusionThere is a θ-frequency resonance in STN neurons. It is mediated by Ih.The resonance characteristics are temperature-and voltage-dependent. Theresonance mediates a frequency-selective coupling between inputs and firing.The D2antagonist sulpiride, but not the D1antagonist SCH-23390, blockedresonance and Ihof STN neurons. G-protein-coupled D2-like receptor isinvolved in the effect of sulpiride on STN neurons. |
PDF Full Text Request