Abstract
The modification of the discharge pattern of subthalamic nucleus (STN) neurons from single-spike activity to mixed burst-firing mode is one of the characteristics of parkinsonism in rat and primates. However, the mechanism of this process is not yet understood. Intrinsic firing patterns of STN neurons were examined in rat brain slices with intracellular and patch-clamp techniques. Almost half of the STN neurons that spontaneously discharged in the single-spike mode had the intrinsic property of switching to pure or mixed burst-firing mode when the membrane was hyperpolarized from −41.3 ± 1.0 mV (range, −35 to −50 mV; n = 15) to −51.0 ± 1.0 mV (range, −42 to −60 mV; n = 20). This switch was greatly facilitated by activation of metabotropic glutamate receptors with 1S,3R-ACPD. Recurrent membrane oscillations underlying burst-firing mode were endogenous and Ca2+-dependent because they were largely reduced by nifedipine (3 μm), Ni2+ (40 μm), and BAPTA-AM (10–50 μm) at any potential tested, whereas TTX (1 μm) had no effect. In contrast, simultaneous application of TEA (1 mm) and apamin (0.2 μm) prolonged burst duration. Moreover, in response to intracellular stimulation at hyperpolarized potentials, a plateau potential with a voltage and ionic basis similar to those of spontaneous bursts was recorded in 82% of the tested STN neurons, all of which displayed a low-threshold Ni2+-sensitive spike. We propose that recurrent membrane oscillations during bursts result from the sequential activation of T/R- and L-type Ca2+ currents, a Ca2+-activated inward current, and Ca2+-activated K+ currents.
- tonic and bursting activities of STN neurons in slices
- burst ionic mechanisms
- low-threshold spike
- Ca2+-dependent plateau potential
- intracellular and patch-clamp recordings
- Parkinson’s disease
The subthalamic nucleus (STN) is composed of glutamatergic neurons that control the circuitry of the basal ganglia by modulating the activity of the two principal output structures of the network: the internal pallidal segment and the substantia nigra pars reticulata (for review, see Albin et al., 1989;DeLong, 1990; Parent and Hazrati, 1995; Mink, 1996; Féger et al., 1997). The importance of this control is exemplified by the various consequences of STN lesion in both control animals and animal models of Parkinson’s disease.
Electrolytic lesions of the STN in normal monkeys produce a hyperkinetic syndrome (Whittier and Mettler, 1949). This has also been reproduced by toxic lesions restricted to the STN, sparing the fibers of passage (Hammond et al., 1979; Hamada and DeLong, 1992), transient pharmacological blockade of STN activity (Crossman et al., 1984), and high-frequency STN stimulation (Beurrier et al., 1997). All these observations reflect the importance of the control exerted by the STN in control animals and provide an explanation for the violent, involuntary movements of the contralateral limbs (termed “hemiballism”) that occur in STN-lesioned humans (Martin, 1927;Bathia and Marsden, 1994). Manipulating STN neurons in animal models of Parkinson’s disease leads to a very different consequence. In monkeys treated with the neurotoxic 1-methyl-4-phenyl-1,2,3,6-tetrahydropyridine (MPTP), STN lesion (Bergman et al., 1990; Aziz et al., 1991; Guridi et al., 1996), pharmacological blockade of the subthalamopallidal pathway (Graham et al., 1990; Brotchie et al., 1991), or high-frequency stimulation of the STN (Benazzouz et al., 1993), produce a reduction in motor impairments. These results suggest that STN constitutes a good therapeutic target for the treatment of Parkinson’s disease. For this reason, high-frequency stimulation of the STN is being performed in several patients suffering from severe parkinsonism and gives very consistent results (Limousin et al., 1995).
To understand how the activity of STN neurons can regulate the operational mode of basal ganglia, it is essential to determine in detail the electrical properties of STN neurons and the underlying ionic mechanisms in physiological conditions in vitro. Previous studies have described the responses of STN neurons to intracellular current pulses (Nakanishi et al., 1987; Overton and Greenfield, 1995; Overton et al., 1995; Plenz et al., 1997). However, their ionic basis, as well as those of spontaneous firing patterns, have not been fully characterized. We now report, with the use of intracellular and patch-clamp techniques in rat brain slices, that a substantial proportion of STN neurons can shift from a regular single-spike mode to a burst-firing mode. We have analyzed the intrinsic membrane properties underlying this property and propose that this electrical behavior provides a cellular substrate for the functional role of the STN in controlling movements under normal and altered conditions.
MATERIALS AND METHODS
Slice preparation. Experiments were performed on STN neurons in slices obtained from 20- to 28-d-old male Wistar rats. Rats were anesthetized with ether and decapitated. The brain was quickly removed, and a block of tissue containing the STN was isolated on ice in a 0–5°C oxygenated solution containing (in mm): 1.15 NaH2PO4, 2 KCl, 26 NaHCO3, 7 MgCl2, 0.5 CaCl2, 11 glucose, and 250 saccharose, equilibrated with 95% O2 and 5% CO2, pH 7.4. This cold solution, with a low NaCl and CaCl2 content, improved tissue viability. In the same medium, 300- to 400-μm-thick coronal slices were prepared using a Vibratome (Campden Instruments LTD, Loughborough, UK) and were then incubated at room temperature in a Krebs’ solution containing (in mm): 124 NaCl, 3.6 KCl, 1.25 HEPES, 26 NaHCO3, 1.3 MgCl2, 2.4 CaCl2, and 10 glucose, equilibrated with 95% O2 and 5% CO2, pH 7.4. After a 2 hr recovery period, STN slices were transferred one at a time to an interface-type recording chamber, maintained at 30 ± 2°C and continuously superfused (1–1.5 ml/minute) with the oxygenated Krebs’ solution.
Electrophysiological recordings. Slices were visualized using a dissecting microscope, and the recording electrode was precisely positioned in the STN. Electrophysiological recordings of STN neurons were performed in current-clamp mode using the intracellular or patch-clamp technique. Signals were recorded using an Axoclamp 2A (Axon Instruments, Foster City, CA) in bridge or continuous single-electrode voltage-clamp mode for intracellular and patch-clamp experiments, respectively.
For intracellular recordings, microelectrodes were pulled from filamented borosilicate glass (BF-100–50-10; Sutter Instruments, Novato, CA) on a horizontal Flaming–Brown micropipette puller (P-87; Sutter Instruments). They had a resistance of 150–200 MΩ when filled with 2 m potassium acetate. For patch-clamp experiments, recordings were made using the blind patch-clamp technique in the cell-attached or whole-cell configuration. Patch electrodes were pulled from filamented borosilicate thin-wall glass capillaries (GC150F-15; Clarck Electromedical Instruments, Pangbourne, UK) with a vertical puller (LM-3P-A; List Instruments, Darmstadt-Eberstad, Germany) and had a resistance of 10–12 MΩ when filled with (in mm): 120 Kgluconate, 10 KCl, 10 NaCl, 10 EGTA, 10 HEPES, 1 CaCl2, 2 MgATP, and 0.5 GTP, pH 7.25.
Drugs. All drugs were purchased from Sigma (St. Louis, MO), except 1S,3R-1-aminocyclopentane-1,3-dicarboxylate (1S,3R-ACPD) purchased from Tocris Cookson (Bristol, UK) and tetrodotoxin (TTX) and apamin purchased from Latoxan (Rosans, France). Biocytin and 1,2-bis(2-aminophenoxy)-ethane-N,N,N′,N′-tetraacetic acid (BAPTA) were diluted in the pipette solutions. All other drugs were diluted in the oxygenated Krebs’ solution and applied through this superfusion medium. Nifedipine and BAPTA-AM were dissolved in dimethylsulfoxide (final concentration, 0.03–0.5%). For experiments with cobalt (Co2+), calcium (Ca2+)-free (Ca2+ was substituted for equimolar concentration of Co2+) or low Ca2+ solutions (0.4 mm with 2 mm Co2+) were used.
Histology. In some experiments, recordings were performed with pipette solution containing biocytin (0.5–1%) to allow the subsequent identification and morphological analysis of the recorded neurons. The avidin–biotinylated horseradish peroxidase (ABC complex) reaction was used to visualize the biocytin-filled neurons. After recording, slices were fixed for 1–2 d in a solution containing 4% paraformaldehyde and 0.15% picric acid in phosphate buffer (0.1m, pH 7.4) at 4°C. After several rinsings in Tris-buffered saline (TBS; 0.05 m, pH 7.4), the sections were treated with a mixture of methanol and H2O2 for 30 min, rinsed again in TBS, and processed for 2 or 3 d with the standard ABC complex (Vectastain ABC kit, Vector laboratories, Burlingame, CA) at 4°C. After several washes in TBS, sections were treated with diaminobenzidine as a chromogen and H2O2 for 5–10 min (DAB substrate peroxidase kit; Vector Laboratories). The sections were then rinsed several times in TBS, dehydrated and rehydrated in graded ethanol, stained with cresyl violet, dehydrated again in graded ethanol, cleared in xylene, and mounted in Eukitt (053–47505; Labonord, Villeneuve d’Ascq, France).
Data analysis. Current and voltage outputs were displayed simultaneously on a storage oscilloscope and a four-channel chart recorder (Gould Instruments, Longjumeau, France), digitized (DR-890; NeuroData Instruments, New York, NY), and stored on a videotape for subsequent off-line analysis with pClamp6 software (Axon Instruments). Values are expressed as mean ± SEM. Statistical significance was assessed using the Student’s t test (unpaired data). Parameters of the single-spike and bursting modes were quantified from intracellular and patch-clamp recordings, respectively.
RESULTS
Morphology and passive membrane properties of STN neurons
Results were obtained from 141 STN neurons. The soma of biocytin-filled recorded neurons (n = 5) were all located within the boundaries of the STN identified with the cresyl violet staining technique (Fig.1A). Soma had diameters of 10–25 μm and gave rise to four or five dendritic trunks (Fig. 1B). Axons gave rise to numerous collaterals. Only neurons with an input membrane resistance >100 MΩ, firing action potentials with an amplitude of at least 50 mV, and an afterhyperpolarizing potential (AHP) at a threshold of −50 mV, were included in the present study. The resting potential of spontaneously firing neurons was difficult to establish because of the absence of a stable membrane potential. The input resistance of STN neurons was 200.2 ± 6.8 MΩ (n = 88) and not significantly different when measured with both the patch-clamp and intracellular techniques (p = 0.36; n = 88).
Microphotographs of a biocytin-filled STN neuron at two magnifications. The labeled neuron is located within the boundaries of the STN (top) and presents a dense dendritic arborization (top) and numerous spines on dendrites (bottom).
Firing patterns of STN neurons: single-spike and burst-firing modes
All tested STN neurons (n = 83) displayed a tonic discharge of single spikes (Fig. 2, single-spike mode) that totally disappeared in the presence of TTX (1 μm, n = 33 of 33, data not shown). Among STN neurons, 46% (n = 38 of 83) were also able to fire in bursts (burst-firing mode, Fig. 3). These STN neurons switched from one mode to the other depending on membrane potential (Fig. 4). Single-spike mode was recorded at membrane potentials between −35 and −50 mV (−41.3 ± 1.0 mV; n = 15), whereas burst firing was present in the membrane potential range of −42 to −60 mV (−51.0 ± 1.0 mV; n = 20) (Fig. 4). These values represent the threshold potential of spikes in the single-spike mode (Table 1, parameter 3) and that of bursts in the burst-firing mode (Table 2, parameter 1). At membrane potentials more hyperpolarized than −60 to −70 mV, most STN neurons were silent (Fig. 4). At membrane potentials more depolarized than −30 mV, spike amplitude decreased, and spike frequency increased, leading rapidly to a blockade of STN tonic activity. Activation of group I and II metabotropic glutamate receptors (I-IImGluRs) by 1S,3R-ACPD (25 μm) (Nakanishi, 1994; Pin and Duvoisin, 1995) induced burst firing in some STN neurons that bursted poorly in control (n = 16 of 20, data not shown).
Single-spike mode. A, Tonic and regular activity (single-spike mode) of an STN neuron, recorded with intracellular techniques at resting membrane potential and (B) corresponding interspike interval histogram (mean interval, 66.1 ± 15.6 msec; bin width, 12.5 msec).C, Discharge frequency histogram of single-spike mode recorded in 41 STN neurons (mean frequency, 22.3 ± 1.5 Hz; bin width, 10 Hz). Spikes in A are truncated.
Burst-firing mode. Two types of burst mode recorded with patch-clamp techniques in two different STN neurons: pure burst mode (A) and mixed burst mode (B), consisting of long bursts (⋆) separated by sequences of short bursts (○). C, Spontaneous bursts recorded in the cell-attached configuration in voltage-clamp mode.
Switch of firing mode according to membrane potential. Pure burst mode (⋆) was triggered at resting membrane potential (I = 0, dotted line, middle) in a whole-cell-recorded neuron that displayed the single-spike mode (⋄) at a more depolarized membrane potential (I = +0.2 nA,left). At a more hyperpolarized potential (I = −0.6 nA, right), the cell became silent. The two bottom traces are taken from the above records and displayed at an expanded time scale. Spikes of the single-spike mode in the left part are truncated.
Quantitative parameters of the single-spike mode
Quantitative parameters of the burst-firing mode
Quantitative characteristics of the two firing modes are summarized in Tables 1 and 2. The single-spike mode was characterized by an extreme regularity and a rather high frequency (22.3 ± 1.5 Hz;n = 41; Fig. 2). Burst-firing mode could be divided into “pure burst mode” consisting of long-lasting bursts of even duration (Fig. 3A) and “mixed burst mode” alternating bursts of long and short duration (Fig. 3B). Discharge frequency of long bursts in the pure burst mode was highly variable (15.0 ± 1.4 bursts/min; range, 7–29 bursts/min;n = 21). Long bursts (lasting >800 msec) gave rise to numerous spikes with a “crescendo–decrescendo” frequency sequence: the first spike was followed by a rapid increase in spike frequency that reached a maximum before the end of the burst. Frequency then decreased as membrane potential slowly repolarized during bursts (Fig.4, right bottom trace). Table 2 recapitulates the quantitative characteristics of the burst-firing mode. On the basis of the three parameters studied, there was no statistical difference between long bursts taken from “pure burst” or from “mixed burst” firing modes (Table 2).
Burst-firing mode was recorded using both patch-clamp and intracellular techniques. Noteworthy, spontaneous bursts were also observed in the cell-attached configuration in patch-clamp experiments (Fig.3C). However, the relative percentage of bursting neurons varied according to the recording technique, because burst firing was more often observed in patch-clamp recordings (48%) than in intracellular recordings (36%). The input resistance of bursting cells was not significantly different from that of cells that did not burst (195.2 ± 9.4 MΩ and 225.2 ± 15.0 MΩ, respectively;p = 0.1; n = 48; whole-cell recordings).
Ionic basis of burst-firing mode
Pharmacological studies were performed in the whole-cell configuration. TTX (1 μm) suppressed action potentials, but spared rhythmic membrane oscillations underlying bursts whose duration was increased by 358 ± 76% as compared with control (n = 5; Fig.5A). Knowing that Ca2+-dependent mechanisms are often involved in burst generation, we tested several drugs known to interfere with Ca2+ entry or intracellular free Ca2+ ions. They all had an inhibitory effect on burst firing. Bath application of nifedipine (3 μm), an L-type Ca2+ channel blocker, largely reduced the duration of bursts and even suppressed burst firing (n= 9) at any potential tested (n = 4; Fig.5B). This effect did not reverse throughout the experiment (1–2 hr). Nickel (Ni2+) at a concentration that preferentially blocks T/R-type Ca2+ channels (40 μm, data not shown) (Fox et al., 1987) had a similar, but reversible effect: it decreased the duration of bursts (n = 8) at any potential tested (n = 3). Finally, bath application of the permeable form of the Ca2+ chelator BAPTA (BAPTA-AM, 10–50 μm) largely reduced the duration of bursts and even suppressed burst firing after a delay of ∼40 min (n = 7) at any potential tested (n = 4; Fig.6A).
Ionic basis of the burst-firing mode I.A, TTX (1 μm) totally suppressed action potentials evoked during bursts while sparing the rhythmic oscillations of the membrane potential that underlie bursts. Note the increase in the duration of membrane oscillations in the presence of TTX (from 5.0 ± 0.0 sec to 16.4 ± 3.1 sec). B, The duration of bursts was irreversibly decreased by an application of nifedipine (3 μm, bottom trace) at all tested potentials, whereas the single-spike mode was unaffected.Traces in A and B were obtained from two different STN neurons recorded with patch-clamp techniques in whole-cell configuration. Spikes from the single-spike mode in B are truncated. Calibration is the same fortraces of each section.
Ionic basis of the burst-firing mode II.A, Bath application of BAPTA-AM (50 μm) decreased burst duration at any potential tested, although it spared the single-spike activity (left column).B, Simultaneous application of TEA (1 mm) and apamin (0.2 μm) prevented burst repolarization and locked membrane potential at −30 mV. Repolarizations were obtained by injecting brief hyperpolarizing current pulses (−80 pA, 100 msec). After each repolarizing command, the membrane spontaneously depolarized again (middle trace). As drugs washed out, bursts reappeared, but with a longer duration (2.9 ± 0.1 sec vs 1.7 ± 0.2 sec, bottom trace). Traces inA and B were obtained from two different STN neurons with patch-clamp recordings in whole-cell configuration. Spikes from the single-spike mode in A (on theleft) are truncated.
To analyze whether Ca2+-activated K+ currents play a role in burst repolarization, tetraethylammonium (TEA; 1 mm) was applied at a concentration that blocks the big conductance Ca2+-dependent K+ current, but not the delayed rectifier one. The TEA was applied in combination with apamin (0.2 μm), the selective blocker of the small conductance Ca2+-dependent K+current (for review, see Sah, 1996). TEA and apamin totally prevented burst repolarization and suppressed spikes, except a few early ones (Fig. 6B; n = 4). The sustained depolarization of the membrane, resulting from the blockade of Ca2+-activated K+ currents suppressed spikes, probably by inactivating Na+-channels as already described in cortical neurons (Prince and Connors, 1986). Short hyperpolarizing current pulses (100 msec, −80 pA) were needed to cut off bursts (Fig.6B, middle traces). As the drugs washed out, bursts spontaneously repolarized, although after a longer duration than in control, and spikes reappeared (Fig.6B, bottom trace).
To better understand the Ca2+ and Ca2+-activated currents present in STN neurons, their responses to current pulses were then analyzed.
Ionic basis of responses to intracellular current pulses
Injection of depolarizing or hyperpolarizing current pulses in STN neurons triggered two kinds of responses: a long depolarization that outlasted the current pulse (Figs. 7,8, plateau potentials) and/or a short depolarizing rebound [low-threshold spike (LTS); Fig.9].
Plateau potentials. A, A short depolarizing (100 pA, 100 msec, left trace) or hyperpolarizing (−100 pA, 100 msec, right trace) current pulse triggered both a plateau potential (1.8 and 2.6 msec duration, respectively) that considerably outlasted the duration of the stimulus and was followed by a prominent AHP (28 and 17 mV amplitude, respectively). B, TTX (1 μm) revealed the presence of two different phases in the plateau potential: a slow depolarization triggered by the depolarizing current pulse (50 pA, 200 msec) and an afterdepolarization (267 msec) triggered at the break of the current pulse. C, Long duration plateau potential terminated by a short hyperpolarizing current pulse (−20 pA, 100 msec). D, Amplitude and duration of the plateau potential according to membrane potential. In the presence of TTX (1 μm), the same depolarizing current pulse (100 pA, 100 msec) evoked a plateau potential in the membrane potential range of −60 to −70 mV. At more depolarized (−40 mV, extreme left) or hyperpolarized (−80 mV, extreme right) potentials, the amplitude and duration of the plateau potential was considerably reduced. Traces in A,C, and D were obtained with patch-clamp recordings (whole-cell configuration), and traces inB were obtained with intracellular recordings. All spikes are truncated.
Ionic basis of the plateau response.A, TTX (1 μm) suppressed action potentials but not the plateau potential (middle trace) evoked by a 100 pA, 100 msec current pulse (left trace). In the presence of TTX, bath application of nifedipine (3 μm) suppressed the plateau potential (right trace).B, The duration of the plateau potential was also decreased by bath application of BAPTA-AM (50 μm). Note that the cell fired some action potentials during the current pulse (100 pA, 100 msec). C, In contrast, the plateau potential was significantly increased by simultaneous application of TEA (1 mm) and apamin (0.2 μm) from 0.5 sec (left trace) to 5.5 sec (right trace). All traces were obtained with patch-clamp recordings (whole-cell configuration).
Voltage dependency and pharmacological properties of LTS. A, The three superimposed voltage traces show that LTS was recorded at the break of a hyperpolarizing current pulse (−300 pA, 200 msec) applied at a membrane potential of −63 mV. Its amplitude and rise time depended on the value of membrane potential at the end of the negative current pulse: at −78 mV, LTS was not evoked (bottom trace), whereas at a slightly more depolarized potential (approximately −72 mV, middle trace) it was present and gave rise to a spike. With increased depolarization, LTS amplitude increased, and spike delay decreased (top trace). B, Three superimposed responses to hyperpolarizing current pulses of increasing amplitude (−100, −250, and −350 pA) and fixed duration (300 msec) from Vm = −53 mV. LTS was only evoked when the membrane was held for 300 msec at a potential more hyperpolarized than −78 mV for 300 msec. C, Three superimposed voltage traces in response to hyperpolarizing current pulses of fixed amplitude (−150 pA) and increasing duration (40, 80, and 120 msec). LTS was evoked in a neuron maintained at −64 mV when the membrane was held at −85 mV for at least 80 msec during the application of a hyperpolarizing current pulse. D, In the presence of TTX (1 μm), LTS evoked in response to a hyperpolarizing current pulse (−130 pA, 200 msec) from Vm = −70 mV was not affected (Control and Wash), whereas it was reversibly suppressed by the concomitant application of Ni2+(40 μm). Traces in A,C, and D were obtained with intracellular recordings, and traces in B were obtained with patch-clamp recordings (whole-cell configuration). All spikes are truncated.
The plateau potential
Plateau potentials, as burst-firing mode, were more often observed in whole-cell patch-clamp recordings than in intracellular recordings (93 vs 62%, respectively). In response to depolarizing or hyperpolarizing current pulses (100 pA, 100 msec), 86 STN neurons of the 106 tested (82%) generated long-lasting plateau potentials (mean duration, 1043.7 ± 69.8 msec; range, 300–2500 msec;n = 52) that gave rise to numerous action potentials (Fig. 7A). Plateau potential duration was measured from the beginning of the current pulse to the peak of the AHP. During plateaus, spike frequency increased until the end of the depolarizing current pulse and then gradually decreased during the rest of the plateau phase, thus showing some adaptation (Fig. 7A). Two different phases were easily identified in the presence of TTX (1 μm): a ramp-like, slow-rising depolarization that corresponded to the duration of the depolarizing current pulse, and an afterdepolarization that outlasted the current pulse (Fig.7B). Plateau potentials spontaneously ended with an AHP of 20.6 ± 0.8 mV amplitude (range, 10–35 mV; n = 52) and could also be terminated by a short hyperpolarizing current pulse (Fig. 7C). AHP amplitude was measured in the same way as in burst-firing mode (Table 2). Plateau potentials were triggered within a narrow range of membrane potentials, between −50 and −75 mV (n = 43; Fig. 7D), thus showing that they resulted from the activation of voltage-dependent conductances.
Ionic currents underlying plateau potentials were analyzed in whole-cell recordings with the use of channel blockers. TTX (1 μm) suppressed Na+-dependent spikes (n = 19 of 19) (Fig. 8A, middle trace) but did not affect the plateau response. A low Ca2+ (0.4 mm) extracellular solution containing Co2+ (2 mm) completely suppressed plateau potentials (n = 5, data not shown). To determine the type of voltage-gated Ca2+ currents involved, Ca2+ channel blockers were tested. Nifedipine (3 μm) suppressed plateau potentials (n = 8; Fig. 8A), whereas Ni2+ (40 μm) had no effect (n = 5; data not shown). Chelation of intracellular Ca2+ with either BAPTA-AM (10–50 μm;n = 4; Fig. 8B) or BAPTA in the pipette solution (20 mm; n = 2; data not shown) suppressed or decreased the plateau potential duration after ∼30 and 20 min, respectively. These results suggest that both Ca2+ entry through L-type Ca2+channels and intracellular free Ca2+ ions are involved in the generation of plateau potentials. Simultaneous application of TEA (1 mm) and apamin (0.2 μm) increased the duration of plateau potentials by 610 ± 297% (n = 7; Fig. 8C). In some cells (n = 2), spikes were suppressed probably as a result of the sustained membrane depolarization and the consequent Na+-channels inactivation. Similarly, 1S,3R-ACPD (25 μm) increased plateau potential duration by 252 ± 37% (n = 10; data not shown).
The low-threshold spike
A small, transient depolarization triggering a few spikes was observed at the break of a short hyperpolarizing current pulse in 71% of the STN neurons tested (n = 66 of 93; Fig. 9). In the remaining 29% (n = 27), the small depolarizing rebound was masked by a plateau potential response. This postinhibitory rebound, previously described in STN neurons (Nakanishi et al., 1987;Overton et al., 1995) and in other preparations (for review, seeHuguenard, 1996), has been called an LTS, because of its negative threshold compared with that of Na+-dependent spikes. As illustrated in Figure9A, LTS amplitude increased with membrane depolarization at the end of the pulse. It is worthwhile noting the depolarizing sag of the membrane potential, a typical sign of the presence of the hyperpolarization-activated cation current (Ih) previously described in STN neurons (Nakanishi et al., 1987; Overton et al., 1995). Currents underlying LTS inactivated with depolarization as LTS was triggered, but only after maintaining the membrane at potentials more hyperpolarized than −84 mV for 300–400 msec (n = 9; Fig. 9B). LTS inactivation was also time-dependent, as illustrated in Figure9C, in which membrane potential had to be maintained at −85 mV for at least 80 msec to trigger an LTS. LTS was unaffected by TTX (1 μm; n = 6; Fig. 9D) but completely disappeared in a Ca2+-free Co2+-containing (2.4 mm) external solution (n = 3; data not shown), or in the presence of a low concentration of Ni2+ (40 μm;n = 5; Fig. 9D). All these results suggest that a rapid voltage-inactivating, Ni2+-sensitive current such as the low-threshold T/R-type Ca2+current underlies LTS.
DISCUSSION
The main result of our study is that approximately half of the STN neurons in the slices have the intrinsic property of switching from single-spike activity to burst-firing mode. To the best of our knowledge, this type of functional property is unknown in neurons of the various basal ganglia nuclei. It may emphasize the role of STN neurons in normal and parkinsonian states.
The cascade of currents underlying burst-firing mode
In tonic and bursting modes, STN neurons fire Na+-dependent action potentials. In burst-firing mode, neurons display cycles of membrane oscillations (Fig.10,b–a) separated by slow membrane depolarizations (Fig.10,a–b). We have shown that the three phases of bursts: depolarization (b–c), slowly declining plateau (c–d), and repolarization to the AHP (d–a), are dependent on Ca2+entry through voltage-sensitive Ca2+ channels. We propose that the depolarization phase (b–c) results from a low-threshold T/R-type Ca2+ current (IT/R) that depolarizes the membrane to the threshold potential of the nifedipine-sensitive L-type Ca2+ current (IL) and then inactivates. The slowly inactivating ILdepolarizes the membrane to the plateau phase of bursts during which spikes are evoked (c–d). Spikes amplify Ca2+ entry by activating moreIL and, possibly, other types of high-threshold Ca2+ currents (Song et al., 1997). The resulting increase in intracellular Ca2+ concentration activates the TEA- and apamin-sensitive Ca2+-activated K+ currents (IK,Ca). The balance between depolarizing (IL) and hyperpolarizing (IK,Ca) currents, slightly in favor of the latter, explains the gradual decline of the plateau (decrescendo phase). When membrane potential has declined to a certain level, it suddenly repolarizes (d–a) because of rapidIL deactivation (Reuveni et al., 1993) and stronger IK,Ca activation. This leads to the peak of AHP, during which IT/R deinactivates. The membrane then spontaneously depolarizes (a–b) as IK,Ca decays because of Ca2+ clearance mechanisms. Depolarization to the threshold potential of IT/R initiates a new cycle. Because BAPTA suppressed burst firing, instead of blocking burst repolarization by preventing IK,Caactivation, this suggests the participation of a Ca2+-dependent inward current, such as the nonspecific cationic current (ICAN) in the plateau phase of bursts (c–d). The progressive activation of this inward current (together with high voltage-activated Ca2+ currents) may underlie spike acceleration (crescendo phase) during burst and plateau potential as previously described for bursts of rat thalamic reticular neurons (Huguenard and Prince, 1992). Finally, Ih(Nakanishi et al., 1987; Overton et al., 1995; present study) recorded in some STN neurons (Fig. 9A) may also participate in the slow depolarization between consecutive bursts.
The hypothetical cascade of currents underlying the different phases of burst-firing mode. See Discussion for explanation.
Burst-firing mode was observed in whole-cell recordings as well as in recording configurations in which the intracellular medium was left intact, such as cell-attached or intracellular recordings. These results showed that burst firing is a physiological firing mode of STN cells. This mode was, however, more easily obtained when K+ currents were decreased because of the presence of gluconate in the pipette solution (Velumian et al., 1997), or when I-IImGluRs were activated with 1S,3R-ACPD (Nakanishi, 1994; Pin and Duvoisin, 1995). The activation ofIL (Chavis et al., 1995; Russo et al., 1997;Svirskis and Hounsgaard, 1998) or ICAN(Crépel et al., 1994; Guérineau et al., 1995; Congar et al., 1997), and/or the inhibition of K+ currents (Charpak et al., 1990; Guérineau et al., 1994; Schrader and Tasker, 1997) by the stimulation of mGluRs may account for the triggering of burst firing by 1S,3R-ACPD. Indeed, the presence of mGluRs (subtypes 2, 3, and 1a) has been reported in the STN neuropile (Martin et al., 1992; Testa et al., 1998) as well as that of mGluR2 mRNA in STN neurons (Testa et al., 1994).
The presence of a plateau potential response is a characteristic of bursting STN neurons
We have shown in the present study that the ionic conductances underlying bursts and plateau potentials are the same: (1) bursts and plateau potentials shared the same pharmacological sensitivity; (2) the same promoting effect of intracellular gluconate and 1S,3R-ACPD was observed in the occurence of bursts and plateau potentials, and (3) only cells exhibiting a plateau potential were able to generate bursts, either spontaneously or in response to 1S,3R-ACPD, and the reverse also holds true: STN neurons in which plateau potentials could not be triggered were not able to burst. However, this was not the only prerequisite for bursting, because 76% of cells responding with a plateau potential never bursted. The only difference is the lack of effect of a low concentration of extracellular Ni2+on plateau potentials. This can be explained by the fact that the depolarizing current pulse applied to trigger the plateau potential replaced the Ni2+-sensitive T/R-type Ca2+ current that normally supported the slow depolarization between bursts. The other type of triggered response, LTS, was present in all recorded STN neurons. Therefore, we propose to distinguish between two populations of STN neurons in vitro, those that are able to burst and generate LTS and plateau potentials and those that are not able to burst and only respond with an LTS. We propose that the difference between these two populations is the presence of the inward currents that underlie the plateau phase of bursts or the plateau potentials i.e., mainly ILand ICAN.
Single-spike modes with frequencies comparable to the ones reported in the present study (Nakanishi et al., 1987; Overton et al., 1995) have already been described in STN neurons in acute slices. To our knowledge, burst-firing mode had only once been recorded in vitro, (in organotypic cultures of STN neurons with intracellular recording techniques), but their ionic mechanisms were not analyzed (Plenz et al., 1997). In contrast, evoked responses similar to the ones reported here have been previously described. The “slow depolarizing potentials” mentioned by Nakanishi et al. (1987) are similar to the plateau potentials we described, which were triggered when the membrane potential was in the −50 to −60 mV range. Likewise, the “slow action potentials” (Nakanishi et al., 1987), “strong rebound bursts” (Plenz et al., 1997), and the “LTS” (Overton et al., 1995) correspond to the LTS described in the present study. Because these different studies did not include precise pharmacological characterization, this similarity is based only on the voltage dependence of the responses. In vivo, both tonic and bursting activities have been recorded in rat and monkey STN (Hollerman and Grace, 1992; Fujimoto and Kita, 1993; Bergman et al., 1994; Overton and Greenfield, 1995; Kreiss et al., 1997). However, the switch from one mode to the other, and the ionic mechanisms of the bursting mode, have not been observed or analyzed in either of the preparations.
Functional implications
In a normal in vivo situation, the great majority of rat and monkey STN neurons present a tonic activity with a frequency varying from 5 to 65 Hz, and few neurons discharge in bursts (Matsumara et al., 1992; Wichmann et al., 1994; Overton and Greenfield, 1995). In relation with conditioned arm (Georgopoulos et al., 1983; Miller and DeLong, 1987; Wichmann et al., 1994) or saccadic eye (Matsumara et al., 1992) movements, a burst of high-frequency spikes lasting ∼200–300 msec is usually recorded after the onset of the movement. Because we showed in the present study that bursts or plateau potentials are triggered by membrane hyperpolarization, the movement-correlated burst of action potentials may result from the activation of inhibitory afferents to the STN.
The increase in the percentage of bursts in the discharge of STN neurons is noteworthy after a lesion of the substantia nigra pars compacta in rats and monkeys in vivo (Hollerman and Grace, 1992; Bergman et al., 1994; Hassani et al., 1996) and in parkinsonian patients (Benazzouz et al., 1996; Rodriguez et al., 1997). The origin of this modification in STN activity in a pathological situation is still under debate (for review, see Chesselet and Delfs 1996; Levy et al., 1997), although a disinhibition mechanism and increased activity of glutamatergic STN afferents seem to be crucial, according toDeLong’s (1990) model. We have in fact shown that the activation of metabotropic glutamate receptors that may occur because of increased activity of glutamatergic STN afferents, strongly favored the bursting mode, as also observed in the hippocampus (Bianchi and Wong, 1995). Bursting STN neurons may drive target neurons in the internal segment of the pallidum and in substantia nigra pars reticulata, where an oscillatory activity or an increase in cytochrome oxidase activity, a marker of metabolic activity, have been recorded in animal Parkinson models (Miller and DeLong, 1987; Filion et al., 1988; Bergman et al., 1994; Nini et al., 1995; Vila et al., 1997; Bergman et al., 1998) and in parkinsonian patients (Hutchison et al., 1997; Vila et al., 1997). This suggests that the strong regulation exerted on STN neurons in control animals is disorganized in animal models of parkinsonism. This may indeed account for the switch from a tonic to a bursting mode.
Footnotes
This work was supported by grants from Centre National de la Recherche Scientifique, Fondation pour la Recherche Médicale, Conseil Régional d’Aquitaine, and Université de Bordeaux II. C.B. has a scholarship from Ministère de l’Enseignement Supérieur et de la Recherche (French Ministry of Research and Higher Education) and from the Lilly Institute. We thank S. Olliet and A. Taupignon for their comments on this manuscript, J. Audin, R. Bonhomme, J. M. Calvinhac, and G. Gaurier for technical assistance. Outstanding thanks to J. M. Israel.
Correspondence should be addressed to Corinne Beurrier, Université de Bordeaux II, Centre National de la Recherche Scientifique Unité Mixte de Recherche 5543, 146 rue Léo Saignat, 33 076 Bordeaux cédex, France.
