Voltage-dependent Ca2+ conductances were investigated in Purkinje cells in rat cerebellar slice cultures using the whole-cell and cell-attached configurations of the patch-clamp technique. In the presence of 0.5 mm Ca2+ in the extracellular solution, the inward current activated with a threshold of −55 ± 1.5 mV and reached a maximal amplitude of 2.3 ± 0.4 nA at −31 ± 2 mV. Decay kinetics revealed three distinct components: a fast (24.6 ± 2 msec time constant), a slow (304 ± 46 msec time constant), and a nondecaying component. Rundown of the slow and sustained components of the current, or application of antagonists for the P/Q-type Ca2+ channels, allowed isolation of the fast-inactivating Ca2+ current, which had a threshold for activation of −60 mV and reached a maximal amplitude of 0.7 nA at a membrane potential of −33 mV. Both activation and steady-state inactivation of this fast-inactivating Ca2+ current were described with Boltzmann equations, with half-activation and inactivation at −51 mV and −86 mV, respectively. This Ca2+ current was nifedipine-insensitive, but its amplitude was reduced reversibly by bath-application of NiCl2 and amiloride, thus allowing its identification as a T-type Ca2+ current. Channels with a conductance of 7 pS giving rise to a fast T-type ensemble current (insensitive to ω-Aga-IVA) were localized with a high density on the dendritic membrane. Channel activity responsible for the ensemble current sensitive to ω-Aga-IVA was detected with 10 mm Ba2+ as the charge carrier. These channels were distributed with a high density on dendritic membranes and in rare cases were also seen in somatic membrane patches.
- T-type calcium channels
- Purkinje cells
- patch clamp
- dendritic recordings
- cerebellar slice cultures
Voltage-activated Ca2+ channels in mammalian neurons have been shown to control many physiological processes, such as transmitter release, neuronal integration, and the generation of neuronal firing patterns. The existence of different types of voltage-activated Ca2+ channels accounts for the variety of these cellular functions (for review, see McCleskey, 1994), and six classes of Ca2+ channels (termed T, L, N, P, Q, and R) have been distinguished kinetically, pharmacologically, and molecularly (for review, see Bean, 1989; Tsien et al., 1991; Snutch and Reiner, 1992; McCleskey, 1994).
There is overwhelming physiological (Llinás et al., 1989; Usowicz et al., 1992) and molecular (Stea et al., 1995) evidence that Purkinje cells (PCs) express high levels of high-threshold P/Q-type Ca2+ channels. Whether functional T-type Ca2+channels exist on cerebellar PCs remains controversial. Although some groups found no evidence for the presence of these channels in adult guinea pigs and young rats (Llinás and Sugimori, 1980a,b; Mintz et al., 1992a,b; Usowicz et al., 1992), others found T-type Ca2+ channels in PCs of young and adult rats (Bossu et al., 1989a,b; Hirano and Hagiwara, 1989; Kaneka et al., 1990; Regan, 1991). In the present study, we investigated both the presence and localization of Ca2+ channels in PCs by using cerebellar slice cultures and Ca2+ or Ba2+ as charge carriers.
MATERIALS AND METHODS
Preparation of cerebellar slice cultures
Organotypic cultures were prepared from cerebella removed from newborn rat pups (0–1 d old) and cultured as described previously (Gähwiler, 1981). In short, the cerebellum was dissected under aseptic conditions, and parasagittal slices of 400 μm thickness were cut with a McIlwain tissue chopper. Individual slices were embedded in clotted chicken plasma on glass coverslips and cultured by means of the roller tube technique at 36°C. The cultures were fed once a week with a medium consisting of 25% heat-inactivated horse serum, 50% Eagle’s basal medium, and 25% HBSS, containing 33.3 mm d-glucose and 0.1 mm glutamine. Electrophysiological recordings were made after a period of >2 weeksin vitro.
Cerebellar slice cultures were transferred to a recording chamber mounted on the stage of an inverted microscope (Axiovert 35M, Zeiss, Jena, Germany), and all experiments were performed at room temperature (25°C). Patch-clamp recordings were carried out under voltage clamp in both the whole-cell recording (WCR) and cell-attached configurations, using an Axopatch 200A amplifier (Axon Instruments, Foster City, CA).
WCR configuration.For the WCR experiments, electrodes were pulled with a Flaming/Brown micropipette puller (Model P-87, Sutter Instruments, San Rafael, CA) using hematocrit glass tubes (plain micro-hematocrit tubes, no. 564; Assistent). Electrodes were filled with a solution containing (in mm): 3 CaCl2, 2 MgCl2, 20 tetraethylammonium chloride (TEACl), 100 HEPES, 30 EGTA (pCa2+ 10−8m), 2 MgATP, and 0.5 GTP. The pH was adjusted to 7.2 with CsOH. Electrodes had a final tip resistance of 1.4 ± 0.15 MΩ. The extracellular solution had the following composition (in mm): 120 trichloroacetic acid (TCA), 0.5 or 1 CaCl2, 3 MgCl2, 20 TEACl, 10 HEPES/CsOH, phenol red (7%). TCA was substituted for NaCl to remove both Na+- and Ca2+-activated Cl− currents (Bossu et al., 1991). The pH of the TCA stock solution (250 mm) and the extracellular solution were adjusted to 7.4 with Tris-base buffer (2.5 m).
PCs in organotypic cultures display an extensive dendritic arborization that could produce voltage-clamp artifacts. Therefore, recordings from PCs that had leak current exceeding 10 pA or recordings showing notable space-clamp artifacts were excluded from this study. The remaining cells had an average membrane capacitance of 61.6 ± 23 pF and a series resistance of 4 ± 2.5 MΩ (n = 20). Whole-cell and series-resistance compensation were applied.
Macroscopic Ca2+ currents were studied by using the software pClamp 5.5 (Axon Instruments) and digitized at 10 kHz (Digidata 1200A, Axon Instruments) before storage on a computer hard disk for off-line analysis (Clampfit 6, Axon Instruments).
Cell-attached configuration. For the experiments performed using the cell-attached configuration, the electrodes were pulled with a vertical puller (L/M-3P-A, List Medical, Darmstadt, Germany) using hematocrit glass tubing (Clark Electromedical Instruments, Reading, UK). To reduce the capacitance of the glass, the tip of the electrode was coated with a thick layer of a synthetic polymer (RTV 141, Rhône Poulenc, Lyon, France). Electrodes were filled with a solution containing (in mm): 120 TEACl, 10 CaCl2, 2 MgCl2, 1 CsCl, 10 HEPES, 1 4-aminopyridine. The pH was adjusted to 7.2 with a Tris-base solution (2.5 m). In some experiments CaCl2 was replaced with BaCl2, as indicated in the text and figure legends. The final tip resistance of the electrodes was 5–10 MΩ. The extracellular solution had the following composition (in mm): 137 NaCl, 2.7 KCl, 2.8 CaCl2, 2 MgCl2, 11.6 NaHCO3, 0.4 NaH2PO4, 5.6 glucose. The pH was fixed at 7.4 by bubbling with 95% O2/5% CO2. Tetrodotoxin (5 × 10−7m) was added to the bathing solution to block propagation of action potentials and to reduce release of neurotransmitter, which may change the resting potential (RP) of the recorded cell.
Voltage steps were elicited with a stimulus generator (Model PG 4000A, Neuro Data Instruments, Delaware Water Gap, PA), and the resulting current traces were filtered at 10 kHz and digitized at 47.2 kHz using a digital data recorder (VR-10B, Instrutech, Great Neck, NY) before storage on a video recorder (Panasonic NV-SD 30). For off-line analysis, data were sampled at 5 kHz and filtered with a cut-off frequency of 1.5 kHz (PClamp 6, Axon Instruments). Individual current traces were leak-subtracted before they were averaged. The leak subtraction was performed using averaged episodes that did not show channel openings. Numerical values are given as mean ± SEM.
Chemicals. ω-Conotoxin-MVIIC was purchased from Latoxan (Rosans, France), and ω-Aga-IVA was a gift from Dr. N. A. Saccomano (Pfizer, Groton, CT). The two drugs were prepared as 100 μm stock solutions in distilled water and stored at −80°C. Before use, aliquots were diluted to 10 μm in extracellular solution containing 1 mg/ml cytochrome C (Sigma, St. Louis, MO). (−)β-Baclofen (β-p-chlorophenyl-GABA) was provided by Ciba Geigy (Basel, Switzerland). Amiloride and nifedipine were purchased from Sigma. All drugs were diluted to their final concentration just before bath application.
Somatic whole-cell and somatodendritic cell-attached recordings were performed from PCs in cerebellar slice cultures maintained for 15–31 d in vitro. Living PCs were identified by the size and shape of their soma and dendritic arborization and by their location within the culture (Gähwiler, 1981).
Characteristics of the macroscopic Ca2+ current
PCs were maintained at a holding potential of −70 mV. Hyperpolarizing voltage steps to −90 mV (300 msec) were first applied to remove possible steady-state inactivation of Ca2+currents, followed by depolarizing voltage steps (500 msec) of increasing amplitude to elicit voltage-dependent Ca2+currents (see voltage protocol in Fig. 1 A). With an external Ca2+ concentration ([Ca2+]e) of 0.5 mm, depolarizing voltage steps elicited an inward current displaying complex decay kinetics, including both transient and sustained components (Fig.1 A). Kinetic analysis using a multi-exponential function was performed on the current at maximal amplitude (n = 16). This type of analysis revealed the presence of three distinct components for the decay: a fast (25 ± 2 msec time constant), a slow (304 ± 46 msec time constant detected in 64% of the cells analyzed), and a nondecaying or sustained component. To compare the voltage dependency for activation of transient and sustained current components, as well as their relative contribution to the total current, current–voltage relationships were obtained by measuring the amplitude of the current at its peak value (including the fast, slow, and sustained components), and at the end of the depolarizing voltage step (500 msec), to measure the sustained component alone (Fig. 1 B). The peak current had a threshold for activation of −55 ± 1.5 mV, whereas the sustained component activated at a more depolarized potential (−43 ± 1 mV). Both components, however, reached a maximal amplitude at a similar membrane potential (−31 ± 2 mV and −29 ± 2 mV for the peak and the sustained current, respectively). The maximal peak current amplitude was −2.3 ± 0.4 nA, and the amplitude of the sustained current was −1.31 ± 0.3 nA.
The [Ca2+]e was raised to 1 mm in the remaining experiments to increase the amplitude of the fast-inactivating Ca2+ current.
Isolation of the fast-inactivating Ca2+ current after rundown of both the slow and the sustained components
In the absence of ATP and GTP in the electrode solution, a decline of the Ca2+ current was observed once the WCR configuration had been established. The amplitude of the peak and the sustained currents were measured every 20 sec during this rundown process. The sustained current clearly decreased much faster (92.4 ± 3.2% reduction, measured after 17 min) than the peak current (51.5 ± 3.7% reduction). This procedure thus allowed the isolation of the fast-decaying transient current, defined as the difference of amplitude between the peak and the sustained currents (Fig. 2). Its amplitude remained relatively stable during at least an additional 30 min period of time (data not shown). This fast-inactivating Ca2+current displayed a threshold for activation of −59 ± 2 mV and reached a maximal amplitude of −0.76 ± 0.14 nA at a membrane potential of −33 ± 2.7 mV (n = 9).
Isolation of the fast-inactivating component with toxins against the P/Q-type Ca2+ currents
In the presence of ATP and GTP in the pipette solution, bath-application of ω-Aga-IVA (200 nm, n= 5), a selective blocker of P/Q-type Ca2+ channels, reduced the amplitude of the peak current (60 ± 12%) and almost abolished the sustained component of the Ca2+ current (85 ± 10%), thereby unmasking a component that inactivated during the first 100 msec of the depolarizing voltage step (n = 5). Similarly, treatment of cerebellar slice cultures for 30 min with ω-conotoxin-MVIIC (5 μm,n = 5), another peptide that blocks P/Q-type Ca2+ channels (Hillyard et al., 1992), or with ω-Aga-IVA (200 nm) allowed isolation of the fast-inactivating Ca2+ current (Fig. 3, A and B, respectively). Its threshold for activation was −62 ± 0.8 mV, and the maximal amplitude was 0.62 ± 0.06 nA, achieved at a membrane potential of −33 ± 1.7 mV. These values are not significantly different from those obtained after rundown of the high-threshold Ca2+ currents. We thus pooled together the results obtained in both conditions (n = 19). The fast-inactivating Ca2+ component activated at −60.3 ± 1.1 mV and displayed a maximal amplitude of −0.69 ± 0.08 nA at a membrane potential of −33 ± 1.7 mV (Fig.4 A).
Taken together, our results demonstrate the presence of a fast-inactivating Ca2+ current in PCs that could be isolated after either rundown or pharmacological block of the slowly inactivating and sustained components of the total Ca2+current.
Kinetic properties of the isolated low-threshold fast-inactivating Ca2+ current
The kinetics of the low-threshold Ca2+ current were studied at potentials varying between the threshold for activation (−60 mV) and the membrane potential showing the maximum amplitude of the current (−30 mV). The activation and inactivation of the low-threshold Ca2+ current could be well fitted as the sum of two exponentials at every potential tested (n = 19). The activation time constant decreased progressively from 8.7 ± 0.7 msec at −60 mV to 5.8 ± 0.5 msec at −30 mV (Fig.4 B, open circles). The inactivation time constant increased progressively from 24.5 ± 3.3 msec at −60 mV to 14 ± 0.9 msec at −30 mV (Fig. 4 B,filled circles).
Activation and inactivation properties of isolated low-threshold fast-inactivating Ca2+ current
Step depolarizations (500 msec) to potentials between −65 and −20 mV were preceded by hyperpolarizing prepulses to −100 mV to determine the activation of the fast-inactivating Ca2+current (Fig. 5 A, left traces). The relative conductance as a function of the test potential could be described with a Boltzmann equation having a half-maximal conductance at −51 mV and a slope factor of 4.6 (Fig. 5 A, right trace). The steady-state inactivation of the transient Ca2+ current was examined with step hyperpolarizing prepulses to potentials between −110 and −65 mV, followed by depolarizing voltage steps eliciting the maximal amplitude of the current (Fig. 5 B, left traces). The relative amplitude of the current as a function of the hyperpolarizing prepulses was fitted with a Boltzmann equation having a half-maximal inactivation at −86 mV and a slope factor of 4.9 (Fig. 5 B, right trace). The two Boltzmann functions did not overlap, indicating that the low-threshold Ca2+ current was not tonically activated at the resting membrane potential.
The kinetic properties of the low-threshold Ca2+ current described above thus closely resemble those reported for T-type Ca2+ currents in other cell types (Bean, 1989).
Pharmacology of the low-threshold Ca2+ current
The low-threshold Ca2+ current in PCs was further characterized pharmacologically, after blocking P/Q-type Ca2+ currents by incubating the cultures for 30 min in saline containing 200 nm ω-Aga-IVA. Inactivation of the low-threshold Ca2+ current was removed with hyperpolarizing prepulses to −100 mV (300–500 msec). Bath application of 10 μm Cd2+ reduced the current by 90 ± 2% (n = 5), demonstrating that it was mediated by Ca2+ channels. Four different compounds reported to reduce T-type Ca2+ currents were tested. The inorganic cation Ni2+ (NiCl2, 100 μm;n = 5) reduced the amplitude of the current by 19 ± 2.7% (Fig. 6 A). Amiloride (250 μm,n = 5) was the most effective drug tested, reversibly decreasing the amplitude of the low-threshold Ca2+ current by 53 ± 4% (Fig. 6 B). Nifedipine (10 μm, n = 5) and baclofen (100 μm, n = 5) only weakly affected the amplitude of the fast-inactivating Ca2+ current (5.5 ± 1.3% and 4.6 ± 2%, respectively) (Fig. 6 C). These pharmacological characteristics are consistent with the identification of the low-threshold fast-inactivating Ca2+ current present in PCs as a T-type Ca2+ current.
Characteristics of the Ca2+ channels
The cell-attached configuration of the patch-clamp technique was used to localize the channels underlying the macroscopic T-type Ca2+ current. Recordings were performed from either somata or dendrites of PCs, using Ca2+ (10 mm) as the charge carrier. The dendritic recordings were carried out from proximal primary dendrites (distance from the cell bodies, 20–50 μm) or from more distal secondary dendritic branches (distance from the cell bodies, 100–200 μm). A high density of Ca2+ channels producing ensemble averages with a low threshold of activation and fast inactivation were seen in 70% of dendritic recordings (n = 10). In those experiments, depolarizing voltage steps from −20 mV relative to the RP to +40 mV relative to RP elicited inward ensemble currents that completely inactivated during the first 100 msec of the depolarizing voltage step (Fig. 7 A,left panel). The decay of 30 averaged current traces could be fitted with a single exponential, having a time constant of 25 ± 6 msec (n = 6). The amplitude of the inward ensemble current was plotted as a function of the pipette potential (Fig. 7 A, right panel). The fast-inactivating ensemble current had a threshold for activation of +10 mV relative to RP and reached a maximal amplitude of −7.9 ± 2.1 pA (n = 6).
Similar channel activity has also been observed with somatic recordings but in only 25% of the cells tested (n = 20, not illustrated). The maximal amplitude of the transient inward ensemble current (−1.1 ± 0.9 pA) was smaller than at dendritic sites, suggesting a lower density of channels on somatic membranes. By plotting the amplitude of individual single-channel currents as a function of the pipette potential, a mean conductance of 7 ± 1 pS was determined for the somatic and dendritic channels (n = 5). In another set of experiments, the pharmacology of the ensemble currents in somatic and dendritic recordings was tested with the P/Q-type Ca2+ channel blocker ω-Aga-IVA. When ω-Aga-IVA (200 nm) was added to both the pipette and the extracellular solutions, depolarizing voltage steps from −20 to +30 mV (500 msec) relative to RP evoked channel activity whose ensemble averages completely inactivated during the pulse in all cells tested (Fig. 7 B, left panel). The threshold for activation and the maximal amplitude of the ensemble current were similar to the values obtained in control conditions (Fig. 7 B). These results demonstrate that the P/Q-type Ca2+ channels contribute relatively little current in such cell-attached recordings. Taken together with the activation threshold and inactivation kinetics, we suggest rather that T-type Ca2+ channels underlie the bulk of the current under the present, relatively physiological, conditions.
No detectable channel activity resulting in a high-threshold sustained ensemble current was apparent in any of the dendritic or somatic membrane patches. In some dendritic recordings, however, current traces with a high level of noise, in addition to the transient inward ensemble current, were observed when the patch was depolarized to +50 mV relative to RP (data not shown). Because high-threshold Ca2+ channels have a greater permeability for Ba2+ than for Ca2+ (Hagiwara and Byerly, 1981), we replaced 10 mm CaCl2 with 10 mmBaCl2 in the pipette solution. Under these conditions, voltage steps (500 msec) from −20 mV relative to RP to test potentials ranging from +10 to +80 mV relative to RP elicited channel openings in all dendritic recordings (Fig. 8). This activity consisted of the very brief opening of many channels and gave rise to ensemble averages with both transient and sustained components. The maximal amplitude of these components was −28.5 ± 9.5 pA and −12.4 ± 2.4 pA, respectively, for a depolarization of +50 mV relative to RP (n = 8).
Such sustained channel activity was not recorded in the presence of ω-Aga-IVA (200 nm, n = 3) added to both the pipette and extracellular solutions, suggesting that P/Q-type channels underlie the high-threshold sustained ensemble Ba2+ current (data not shown). High-threshold channel openings could be also detected on the somatic membrane of PCs but in only 19% of the cells tested (n = 32). In three of these six cells, the channel activity was sufficient to give rise to a sustained ensemble current with a maximal amplitude of −11 ± 6 pA, for a membrane potential of +50 mV relative to RP (data not shown).
PCs in cerebellar slice cultures display a fast-inactivating Ca2+ current that can be recorded in isolation after either rundown or pharmacological blockade of the sustained components of the macroscopic Ca2+ current. Several lines of evidence indicate that the fast-inactivating Ca2+ current is a T-type Ca2+ current.
First, the threshold for activation of the transient current was −60 ± 1.1 mV, close to the value reported for T-type Ca2+ currents (Bean, 1989). The decay time constant (14 ± 0.9 msec at −30 mV) and half-maximal activation and inactivation (−51 and −86 mV, respectively) are similar to those of T-type Ca2+ currents in sensory neurons (Carbone and Lux, 1984; Bossu et al., 1985; Fox et al., 1987a), mammalian thalamic neurons (Coulter et al., 1989; Huguenard and Prince, 1992), and hippocampal pyramidal cells (O’Dell and Alger, 1991; Thompson and Wong, 1991). Second, its amplitude was affected neither by the rundown process (Bossu et al., 1985; Fedulova et al., 1985) nor by application of nifedipine, as reported for sensory neurons (Boll and Lux, 1985; Fox et al., 1987a). The transient Ca2+ current, however, was reduced by amiloride, a selective blocker of T-type Ca2+channels (Tang et al., 1988), and to a lesser extent by Ni2+, as originally reported by Carbone et al. (1987). The weak effect of Ni2+ (100 μm) may reflect a particular property of LVA Ca2+ currents in PCs in cerebellar slice cultures, because Ni2+ (25–100 μm) strongly reduced LVA Ca2+ current in embryonic hippocampal neurons (Ozawa et al., 1989) and DRG neurons (Fox et al., 1987a) as well as various thalamic cells (Coulter et al., 1989;Huguenard and Prince, 1992). On the other hand, higher concentrations (200–600 μm) were needed in other neuronal cells (Akaike et al., 1989; Crunelli et al., 1989; Barish, 1991).
The data obtained with the cell-attached recording configuration with Ca2+ as the charge carrier revealed the existence of a voltage-activated channel activity that gave rise to a transient ensemble Ca2+ current that was insensitive to ω-Aga-IVA. The threshold for activation of the transient ensemble Ca2+current was 0 to +10 mV relative to RP, a value that is compatible with the threshold of the T-type Ca2+ current (−60 ± 1.1 mV), given that the RP of PCs ranges between −70 and −60 mV in sharp electrode recordings (D. Mouginot, unpublished observations). In addition, the transient ensemble current displayed fast inactivation kinetics with a time constant of 25 ± 6 msec, comparable to the time constant of 24.6 ± 2 msec for the fast component of the macroscopic Ca2+ current. Finally, the conductance of the Ca2+ channels in PCs was 7 ± 1 pS, a value that is similar to that of T-type Ca2+ channels in primary cultures of dissociated PCs (Bossu et al., 1989b), sensory neurons (Carbone and Lux, 1984; Nowycky et al., 1985), and hippocampal neurons (Fisher et al., 1990; O’Dell and Alger, 1991), when Ba2+ (110 mm) was used as the charge carrier.
Our data indicate that the sustained Ca2+ current is a member of the P/Q-type Ca2+ current family. First, the sustained Ca2+ current had a threshold for activation of −43 ± 1 mV and showed weak inactivation during depolarizing voltage steps, a property that characterizes these Ca2+currents (Bean, 1989; Zhang et al., 1993). Maximal current amplitudes were reached at approximately −30 mV, which supports the view that high-threshold Ca2+ currents in PCs peak at more negative potentials than they do in most other neurons (Regan et al., 1991). In the absence of ATP and GTP in the pipette solution, the sustained Ca2+ current underwent a time-dependent reduction of its amplitude, as observed consistently during WCRs of high-threshold currents (Bossu et al., 1985; Fedulova et al., 1985). Second, the sustained Ca2+ current described in this study was reduced markedly by ω-Aga-IVA and by ω-conotoxin-MVIIC, two potent inhibitors of the P/Q-type Ca2+ current (Hillyard et al., 1992; Mintz et al., 1992a,b). Third, sustained channel activity was detected readily in cell-attached recordings in which Ba2+was used as the charge carrier. This channel activity presumably underlies the sustained Ca2+ current seen in macroscopic recordings, because it was not apparent when ω-Aga-IVA was present in both the pipette and the extracellular solutions. The sustained channel activity could not be resolved sufficiently when the charge was carried by Ca2+, indicating that the conductance of the P/Q-type channels is greater for Ba2+ than for Ca2+, as reported for P/Q-type Ca2+ channels isolated from squid optic lobe and incorporated into artificial lipid membranes (Llinás et al., 1989). In addition, these Ca2+channels may be highly sensitive to Ca2+-dependent inactivation induced by Ca2+ influx through these channels.
The cell-attached configuration of the patch-clamp technique allowed the localization of the channels underlying both T-type and P/Q-type Ca2+ currents in PCs. T-type Ca2+ channel activity was recorded from both dendrites and somata of PCs. T-type channel activity was uncommon (25% of the cells tested) in somatic recordings, however, and the amplitude of the transient ensemble Ca2+ current in these patches was smaller than in dendritic patches. A dendritic localization of T-type Ca2+ currents has been reported for PCs in dissociated cultures derived from young rats (Bossu et al., 1989b), in acutely dissociated hippocampal neurons (O’Dell and Alger, 1991), and in CA1 pyramidal cells in hippocampal slices (Karst and Wadman, 1993). In contrast, a somatic location of T-type Ca2+ channels was suggested by data obtained with intracellular recordings from inferior olive neurons (Llinás and Yarom, 1981), as well as by single-channel recordings from acutely isolated cell bodies of thalamic neurons (Suzuki and Rogawski, 1989), from sensory neurons (Fox et al., 1987b), and from somata of CA3 pyramidal cells (Fisher et al., 1990). These studies, however, do not exclude the presence of T-type Ca2+ channels on dendritic membranes.
In cerebellar slice cultures, P/Q-type Ca2+ channel activity was detected mainly on dendritic membranes of PCs. These results are in agreement with a previous single-channel study carried out on acute cerebellar slices from guinea pigs (Usowicz et al., 1992), as well as with immunohistochemical data obtained from rat cerebellar PCs (Hillman et al., 1991). Similarly, the K+-induced increase in the intradendritic free Ca2+ concentration in cultured rat cerebellar PCs was blocked by ω-Aga-IVA (Bindokas et al., 1993). Although only 19% of somatic patches displayed high-threshold channel activity in cerebellar slice cultures, the amplitude of the sustained ensemble current, when present, was comparable in somatic and dendritic locations. These results therefore support the hypothesis that P/Q-type Ca2+ channels are present in high densities at hot spots of somatic membranes, as also suggested by recent Ca2+ imaging data from somatic membranes of PCs (Kano et al., 1995).
Although there is general agreement about the existence of P/Q-type Ca2+ channels in cerebellar PCs, there is no consensus with respect to the presence of T-type Ca2+ channels. Macroscopic T-type Ca2+ currents have been identified in cultures of dissociated PCs derived from embryonic rat pups (Hirano and Hagiwara, 1989), in young and adult rats (Kaneka et al., 1990; Regan, 1991), and in cultured PCs from newborn rats (Bossu et al., 1989a,b). Furthermore, a recent single-channel study on rat cerebellar slices reported the presence of multiple types of Ca2+ channels on somata and dendrites of PCs. One of these channels activates at low threshold and shows marked inactivation. Although not pharmacologically identified, this channel may thus represent T-type Ca2+channels of young adult rat PCs (Dupere and Usowicz, 1996).
On the other hand, other researchers working with dissociated PCs from young rats (Mintz et al., 1992a,b) or with acute cerebellar slices of adult guinea pigs (Usowicz et al., 1992) failed to find a low-threshold-activated Ca2+ current. One possible explanation for these seemingly contradictory data is that T-type Ca2+ channels may be more prominent in immature neurons, as shown for embryonic spinal cord neurons of Xenopus (Barish, 1991) and cultures of embryonic hippocampal (Meyers and Barker, 1989;O’Dell and Alger, 1991) and sensory neurons (Gottmann et al., 1988). The density of the T-type Ca2+ channels decreases with development in hippocampal pyramidal cells (Thompson and Wong, 1991; but see Karst and Wadman, 1993), in Xenopus spinal neurons (Gu and Spitzer, 1993), and in chick motoneurons (McCobb et al., 1989). The failure to detect T-type Ca2+ currents also may be attributable to the fact that these channels were partially inactivated. Indeed, PCs were often held at membrane potentials of −70 to −80 mV (Llano et al., 1994; Kano et al., 1995), and our data showed that half-maximal inactivation occurs at −86 mV. In other studies, T-type Ca2+ currents may not have been recorded, because PCs in these preparations lacked extended dendrites (e.g., Mintz et al., 1992b).
It will be of interest to determine the physiological implications of Ca2+ entry through low-threshold Ca2+ channels in PCs. It should be noted that our results were obtained from relatively young tissue (2–4 weeks in vitro, derived from neonatal rat pups), and therefore these properties may not reflect the physiology of mature PCs. Preliminary evidence indicates that these Ca2+ channels do not play a role in synaptic transmission between PCs and deep nuclei neurons in cerebellar slice cultures. Indeed, intracellular recordings from deep cerebellar nuclear neurons indicated that synaptic potentials induced by stimulation of PCs are fully blocked by antagonists of the P/Q-type Ca2+ channels (D. Mouginot, unpublished observations). Similar conclusions were drawn from studies with acute cerebellar slices in which IPSCs were abolished almost completely with P- and N-type Ca2+ channel antagonists (Takahashi and Momiyama, 1993). The preferential localization of T channels on dendrites, however, suggests a role in synaptic integration, for example, by producing Ca2+oscillations that could modulate glutamate receptor function. Because T-type channels may be expressed preferentially in immature PCs, they may be implicated in the development of autorhythmic behavior necessary for the generation of synaptic connections (Llinás, 1987).
This study was supported by the Human Frontier Science Program (Grant RG 67/92 B) and the Swiss National Science Foundation (Grant 31-42174.94). We thank L. Rietschin and L. Heeb for the preparation of slice cultures and E. Hochreutener and R. Schöb for technical assistance. We also thank Dr. N. A. Saccomano for the gift of ω-Aga-IVA and Drs. S. M. Thompson and Q. J. Pittman for critical reading and correction of this manuscript. Many thanks to J. C. Poncer and Dr. S. B. Kombian for their assistance with computer software.
Correspondence should be addressed to Dr. Didier Mouginot at Health Science Center, Neuroscience Research Group, Department of Medical Physiology, 3330 Hospital Drive NW, Calgary, Alberta, Canada T2N 4N1.
Jean-Louis Bossu’s present address: Laboratoire de Neurobiologie Cellulaire, Centre National de la Recherche Scientifique, Centre de Neurochimie, 5 Rue Blaise Pascal, F-67084 Strasbourg Cedex, France.