We employed whole cell patch-clamp recordings to establish the effect of Zn2+ on the gating the brain specific, T-type channel isoform CaV3.3 expressed in HEK-293 cells. Zn2+ (300 μM) modified the gating kinetics of this channel without influencing its steady-state properties. When inward Ca2+ currents were elicited by step depolarizations at voltages above the threshold for channel opening, current inactivation was significantly slowed down while current activation was moderately affected. In addition, Zn2+ slowed down channel deactivation but channel recovery from inactivation was only modestly changed. Zn2+ also decreased whole cell Ca2+ permeability to 45% of control values. In the presence of Zn2+, Ca2+ currents evoked by mock action potentials were more persistent than in its absence. Furthermore, computer simulation of action potential generation in thalamic reticular cells performed to model the gating effect of Zn2+ on T-type channels (while leaving the kinetic parameters of voltage-gated Na+ and K+ unchanged) revealed that Zn2+ increased the frequency and the duration of burst firing, which is known to depend on T-type channel activity. In line with this finding, we discovered that chelation of endogenous Zn2+ decreased the frequency of occurrence of ictal-like epileptiform discharges in rat thalamocortical slices perfused with medium containing the convulsant 4-aminopyridine (50 μM). These data demonstrate that Zn2+ modulates CaV3.3 channel gating thus leading to increased neuronal excitability. We also propose that endogenous Zn2+ may have a role in controlling thalamocortical oscillations.
T-type channels—which possess peculiar kinetic properties along with differences in ionic permeation when compared with high-voltage activated Ca2+ channels (Cataldi et al. 2002; Tsien et al. 1988)—are crucial in regulating the periodic discharge of cells in relay thalamic nuclei and in nucleus reticularis thalami (nRT). These neurons are an integral component of the thalamocortical system that is known to be involved in physiological and pathological conditions such as absence epilepsy (Destexhe and Sejnowski 2003; Perez-Reyes 2003; Steriade 2005; Steriade et al. 1993).
At least three different T-type channel isoforms are expressed in neurons as well as in different compartments of the same neuron (Joksovic et al. 2005a; Talley et al. 1999). These include CaV3.1 (Perez-Reyes et al. 1998) and CaV3.2 (Cribbs et al. 1998) channels—which carry the classical fast inactivating T-type current (Carbone and Lux 1984; Fox et al. 1987; Kostyuk et al. 1988)—and the CaV3.3 channels (Lee et al. 1999), carrying more slowly inactivating currents such as those recorded in nRT cells (Huguenard and Prince 1992). As these different isoforms have specific kinetic properties, neuron behavior will depend on the type of T-current it expresses. Specifically, because slow inactivating CaV3.3 channels are required to fire action potentials at a high frequency and for prolonged periods of time (Kozlov et al. 1999), their presence in nRT cells may be essential for periodic spiking (Destexhe and Sejnowski 2003).
Although point mutations affecting channel gating have been identified in absence epilepsy (Khosravani et al. 2004; Vitko et al. 2005), it is unknown how endogenous neurotransmitters or neuromodulators influence T-type-dependent spiking. Here we explore the hypothesis that CaV3.3 channel gating could be affected by Zn2+, which is stored with glutamate in the terminals of a subset of cells, also called the gluzinergic neurons, and coreleased with this excitatory neurotransmitter (Frederickson et al. 2000, 2005; Howell et al. 1984; Qian and Noebels 2005). That Zn2+ may affect T-type channel gating is suggested by the evidence that this metal does affect the gating kinetics of other voltage-gated ion channels (Harrison and Gibbons 1994; Mathie et al. 2006) and greatly increases neuronal excitability in the thalamus, thus suggesting its involvement in absence epilepsy (Gibbs et al. 2000; Wu et al. 2004). Therefore we explored here the hypothesis that Zn2+ influences the slowly inactivating T-type current expressed in nRT neurons—which is encoded by the CaV3.3 gene (Talley et al. 1999)—by evaluating the effects of exogenous Zn2+ on CaV3.3 channels heterologously expressed in stably transfected HEK-293 cells as their currents are identical to those recorded in nRT neurons (Joksovic et al. 2005b). This experimental approach allows study of this channel type in isolation without any interference from other ion channels. By doing so, we found that this ion slowed down the gating of CaV3.3 channel heterologously expressed in HEK-293 cells. We also discovered that when the changes in gating were translated into simulated nRT neurons by using the NEURON software interface (Hines and Carnevale 1997), the frequency and duration of T-type-dependent action potential trains increased. Finally, we established in rat thalamocortical slices maintained in vitro that chelating endogenous Zn2+ decreased the frequency of occurrence of ictal-like epileptiform discharges generated in the presence of the K+ channel blocker 4-aminopyridine (4AP).
Stably transfected HEK-293 cells (courtesy of Dr. Perez-Reyes), expressing the rat CaV3.3a T-type Ca2+ channel isoform (CaV3.3 cells) (Lee et al. 1999), were cultured in a humidified 5% CO2 atmosphere using Dulbecco's modified Eagle's medium supplemented with 5% fetal calf serum, 100 IU/ml penicillin, 100 μg/ml streptomycin, and nonessential amino acids; they were kept under constant selection with 1 g/l geneticin. For electrophysiological recordings, cells were plated on poly-l-lysine (30 μg/ml) precoated glass coverslips and used 24–48 h after plating.
Whole cell patch-clamp recordings
Experiments were performed using the whole cell configuration of the patch-clamp technique as previously reported (Cataldi et al. 2004). Briefly, the coverslips used for cell culture were placed into a laminar flow chamber (Warner Instrument, Hamden, CT) mounted on the stage of a Axiovert 25 Zeiss inverted microscope. The cells were continuously superfused by a gravity-fed multilane system controlled by microcomputer-operated electro-valves (Cell Microcontrol, Norfolk, VA) that allowed the rapid exchange of the perfusing solution. Whole cell recordings were performed at room temperature (25–26°C).
Ruptured patches were obtained by suction using fire-polished borosilicate electrodes having a final resistance of 3–5 MΩ and back filled with a CsCl-based internal solution containing (in mM) 110 CsCl, 30 tetraethyl ammonium chloride (TEA-Cl), 10 EGTA, 2 MgCl2, 10 HEPES, 8 glucose, 15 phosphocreatine di-sodium salt, 5 ATP di-sodium salt, and 1 cAMP sodium salt (pH 7.4 adjusted with CsOH). Unless otherwise specified, the external solution contained (in mM) 125 N-methyl-d-glucamine (NMDG), 2 CaCl2, 1 MgCl2, 5 CsCl, 10 glucose, and 10 HEPES (pH 7.4 adjusted with HCl, yielding a final Cl− concentration of ∼110 mM, as determined by silver nitrate titration). Under these experimental conditions, Ca2+ currents are presumably recorded in isolation because K+ currents are blocked by CsCl present in both internal and external solution and by TEA-Cl contained in the internal solution; in addition, no detectable Na+ current is expected to flow through the membrane because of the lack of Na+ channels in HEK-293 cells and because of the substitution of external Na+ with NMDG. The osmolarity of the external solution was adjusted to 300 mosM by adding an appropriate amount of sucrose. All the voltages reported in the text or shown in figures represent the values obtained after correction for junction potential (∼7 mV in accordance with the estimate provided by the junction potential calculator included in the pClamp8 software) and charge screening (∼4 mV).
Test pulses were generated and the ensuing currents were collected with an Axon 200 B patch-clamp amplifier (Molecular Devices, Foster City, CA) driven by the pClamp6 software running on a personal computer. Currents were filtered at 2 kHz with the amplifier built-in Bessel filter, and leak currents were subtracted on-line with a P/4 protocol. Using the specific commands of the amplifier, on-line correction of membrane capacitance and series resistance was routinely performed. Data were stored onto the hard disk of the personal computer. Off-line analyses were then performed with the Clampfit 8.0 (Molecular Devices) and Sigmaplot 5.0 (SPSS, Chicago, IL) softwares.
Patch-clamp data analysis and computer modeling
The voltage-dependence of activation constants, V1/2m (the membrane potential at which the current is half-maximal) and km (the voltage required to change the conductance, g, e-fold), were calculated fitting the I-V data to the Goldman and Hodgkin current equation as reported by McCormick and Huguenard (1992) (1) where ICa is the current carried by Ca2+ ions (in A) P is the maximum permeability in cm3/s, z is 2 (the valence of Ca2+), V is the applied voltage in V, F is the Faraday constant, R is the gas constant, T is the absolute temperature in Kelvin degrees and ĝ represents fractional conductance (2)
Because (3) for fitting purposes, Eq. 2 was expressed in the form (4)
The values of activation (τm) and inactivation time constants (τh) of the inward Ca2+ currents evoked by membrane depolarization to suprathreshold voltages were obtained by fitting current data to the Hodgkin and Huxley equation (5) where V is the applied voltage, ECa is the reversal potential of Ca2+ currents, and the fraction of open activation (m) and inactivation (h) gates are calculated from the relations (6) (7) and (8) (9) where V1/2m and V1/2h are respectively the half-voltage of activation and of inactivation and km and kh are the slopes respectively of activation and inactivation.
Data fitting was performed using the NEUROFIT software (developed by Allan Willms, Gregory Ewing, and Howie Kuo at the Department of Mathematics and Statistics of the University of Guelph, Guelph, Canada and freely downloadable at the internet address www.uoguelph.ca/∼awillms/neurofit/). Tail currents appearing during channel deactivation were fitted to a bi-exponential equation with fast (τ1) and slow (τ2) time constants. For fitting purposes, the kinetics of tail relaxation was expressed by the single kinetic parameter τ weighted (τw) that was calculated using the equation (10) where A1 and A2 represent the relative amplitude of the fast (A1) and of the slow (A2) decay component of the tail current.
Simulations of current-clamp recordings in modeled reticular thalamic cells were performed using the NEURON 5.8 software (Hines and Carnevale 1997), freely downloadable at www.neuron.yale.edu/neuron. The IT2 model file required to run the NEURON simulation was obtained by entering the kinetics parameters—obtained from the experiments described in the results section—into the model file developed by Murbartian et al. (2004), which was kindly provided by Dr. E. Perez-Reyes (Department of Pharmacology, University of Virginia, Charlottesville, VA). This mod file used for CaV3.3 cells is a modified version of the model file originally developed by Destexhe et al. (1996) for thalamic reticular cells (downloadable at Dr. Desthexe web site, http://cns.iaf.cnrs-gif.fr/Main.html). The NEURON software extrapolates the changes in membrane voltage evoked by the injection of current into the cells by applying an m2 Hogdkin Huxley formalism according to Eq. 5. To get this software running, the fitting parameters needed to calculate m∞ and h∞, and the kinetic constants τm and τh have to be entered into the program. The values of the kinetic constants of activation (V1/2m and km) were obtained fitting the I-V data to the Goldman and Hodgkin current equation as described in the preceding text, whereas those for inactivation (V1/2h and kh) were obtained by fitting the steady-state inactivation data to the Boltzmann equation (11)
Both the voltage dependence of the activation time constant (τm) and that of the inactivation time constant (τh) were specified by the equation (12) where V1, V2 and k1, k2 represent the half-voltages (V1/2) and slopes of increasing and decreasing components of the bell shaped curves generated by plotting τm or τh data obtained at the different tested voltages as detailed in results.
Using Eq. 5, the NEURON software calculates ICa as a function of Ca2+ conductance gCa (5). In our simulation of nRT neurons bathed with 2 mM Ca2+ and no Zn2+, we used the gCa value of 0.0008 (Siemens/cm2) reported in the original nRT model developed by Destexhe et al. (1996) and adapted to HEK cells by Murbartian et al. (2004). Conversely, when the addition of Zn2+ was simulated, to reproduce the 55% decrease in Ca2+ permeability that we observed in our whole cell recordings, we lowered the above-reported gCa value by introducing a multiplication factor of 0.455, according to the equation (13)
The entire simulation was run using the option “burst behavior in a three-compartment model” keeping the resting potential at −70 mV, a value close to those observed in thalamic neurons in vivo during sleep (Fuentealba et al. 2005).
Field potential recordings in talamocortical slices
Thalamocortical slices were obtained using the modifications proposed by Biagini et al. (2001) of the original procedures described by Agmon and Connor (1991). Procedures were approved by the Canadian Council of Animal Care. Briefly young male Sprague-Dawley rats (15–28 days old) were deeply anesthesized with isoflurane and killed by decapitation. The skull was quickly opened, and the brain was removed, placed into an ice-cold oxygenated sucrose-containing solution (in mM: 206 sucrose, 3.5 KCl, 1 CaCl2, 2 MgSO4, 1.25 NaH2PO4, 26 NaHCO3, 1 MgCl2, 10 glucose, 0.4 ascorbic acid, and 1 kynurenic acid) and left to refrigerate into this solution for 3–4 min. Then the brain was transferred onto a Plexiglas ramp the surface of which makes a 10° angle with the horizontal plane. Brain stem and cerebellum were removed with a razor blade by a vertical cut. Then the frontal pole was vertically cut along a plane describing a 45° angle with the sagittal plane. During the entire procedure, the brain was repeatedly irrigated with ice-cold cutting solution. The cut surface of the frontal pole was glued onto the chamber of a Leica VT1000S vibratome (Richmond Hill, Ontario, Canada). Then this chamber was filled with an ice-cold cutting solution, which was continuously oxygenated via an oxygen line, and the brain was cut in 450-μm-thick slices. Only the slices clearly showing bundles of fibers connecting cortex and thalamus were chosen for electrophysiological recordings. On average three to four slices per animal showed these characteristics. Chosen slices were transferred onto the surface of a custom-made interface perfusion chamber and let equilibrate for 1–1.5 h while continuously superfused with oxygenated artificial cerebrospinal fluid (ACSF) (containing in mM: 124 NaCl, 2 KCl, 1.25 KH2PO4, 0.5 MgSO4, 2 CaCl2, 26 NaHCO3, and 10 glucose) at 32°C. At the end of this equilibration period, the K+ channel blocker 4AP (50 μM) was added to the medium to induce spontaneous epileptiform activity (Biagini et al. 2001), and the slices were perfused with this 4AP containing ACSF for an additional hour. ACSF-filled borosilicate electrodes (resistance = ∼10 MΩ) were placed under visual control with the help of a stereomicroscope into the parietal cortex, the ventrobasal nucleus (VB) of the thalamus and the nRT. Field potentials were recorded with preamplifier probes connected with a Cyberamp 300 signal conditioner (Axon Instruments, Molecular Devices), digitized at 5 kHz with a Digidata 1200 A/D converter and stored as pClamp8 files (Axon Instruments, Molecular Devices) for off-line analysis. After recording the baseline activity for >30 min, the cell membrane–permeant Zn2+ chelator N,N,N′,N-tetrakis(2-pyridylmethyl) ethylenediamine (TPEN) was added to the solution and data were recorded in the presence of this drug for additional 30–45 min. Finally, TPEN was washed out and data were acquired for 30 more min. Data analysis was performed off-line with the Clampfit 8 software (Axon Instruments, Molecular Devices).
Geneticin and ATP-sodium salt were obtained from Calbiochem (La Jolla, CA), whereas CsOH was from Aldrich Chemicals (Milan, Italy). DMEM, FCS, penicillin, streptomycin, and nonessential amino acids were from Invitrogen (San Giuliano Milanese, Italy). All the other chemicals were of analytical grade and were from Sigma (Milan, Italy).
All data are reported as means ± SE. When comparing two data sets, the Student's t-test for paired or unpaired data were used as appropriate whereas, repeated-measure ANOVA was used for multiple group comparison. Statistical comparisons were carried out using the Graph-PAD 2.04 software suite (GraphPad Software, San Diego, CA). Curve fitting was performed with the Sigma-plot 5.0 (SPSS) or the N-fit (The University of Texas, Medical Branch at Galveston, Galveston, TX) software.
Effect of Zn2+ on steady-state CaV3.3 channel activation and inactivation
To study the effect of Zn2+ (300 μM) on the voltage dependence of activation, current versus voltage (I-V) curves were obtained from CaV3.3 cells. Briefly, repetitive step depolarizations were delivered at progressively more positive voltages both in the presence and in the absence of this metal (Fig. 1A). The values of half voltage (V1/2m) and slope (km) of the channel activation were extrapolated by fitting the data to the Goldman and Hodgkin current equation reported in methods (Fig. 1B). The values obtained with this approach were not significantly different in the two experimental groups neither in the case of V1/2m (−48.36 ± 0.73 vs. −48.77 ± 1.21 mV, before and after Zn2+ addition, respectively) nor in the case of km (8.26 ± 0.2 vs. 7.25 ± 0.63, before and after Zn2+ addition, respectively). Curve fitting also provided an estimate of the whole cell Ca2+ permeability in the presence and in the absence of Zn2+, thereby showing that this metal reduces cell permeability by 55.6% (from 0.036 ± 0.004 to 0.0237 ± 0.003 cm3/s, P < 0.001). This reduction is close to that of inward Ca2+ current measured either with the protocol shown in Fig. 1, A and B (56.7 ± 7.4%) or by repetitive step depolarization up to −11 mV (56.7 ± 6.9%; Fig. 1C) (cf., Jeong et al. 2003). These findings suggest that even in the presence of high micromolar concentrations of Zn2+ CaV3.3 channel activity is not totally blocked and, therefore its presence in the extracellular solution does not hamper the flow of Ca2+ through CaV3.3 channels.
To study the effect of Zn2+ on the voltage dependence of inactivation, 10-s prepulses of increasing voltages (from –121 up to −11 mV with 10-mV increments) were delivered to two different groups of CaV3.3 cells: one perfused with 2 mM Ca2+ and the other with the same solution plus 300 μM Zn2+; then, after setting back membrane potential to –101 mV for 5 ms, cells were step depolarized up to −11 mV for 75 ms (Fig. 2A). Steady-state inactivation curves were generated by plotting the maximal amplitude of the inward currents evoked by each step, as a function of prepulse voltages (Fig. 2B). Mean half voltages (V1/2h) of inactivation obtained by fitting the data to a charge-voltage Boltzman Eq. 11 were not significantly different in the two groups of CaV3.3 cells (V1/2h = −79.2 ± 2.5 vs. −79.6 ± 2.3 mV, in Ca2+ and Zn2+ groups, respectively). Instead, slope (Kh)values were significantly higher in the presence of Zn2+ (6.1 ± 0.6 vs. 9.2 ± 0.9, P < 0.05; Fig. 2B) in Ca2+ and Zn2+ groups, respectively.
Effect of Zn2+ on CaV3.3 activation and inactivation kinetics
Accordingly to the Hodgkin and Huxley formalism (1952), channel opening and inactivation do not occur instantaneously but take time to appear. Therefore the amount of current that goes through a voltage-gated ion channel at each given time after a depolarization is determined not only by membrane voltage and by the steady-state activation and inactivation of the channels but also by the kinetics of the activation and inactivation gates, which are both intrinsically voltage dependent and can be described with appropriate activation (τm) and inactivation (τh) time constants. This strict voltage dependence of activation and inactivation kinetics is know to apply also to T-type channels as originally described by Coulter et al. (1989) in relay thalamic neurons and may have important functional implications when rapid phenomena such as action potential generation at the postsynaptic membrane of the gluzinergic synapse, occur. Accordingly, we determined the effect of Zn2+ on CaV3.3 channel activation and inactivation gate kinetics at different membrane voltages. For values of membrane potential more positive than the threshold for CaV3.3 channel opening, activation (τm) and inactivation (τh) time constants were determined by fitting the current traces elicited by the step depolarization protocol reported in Fig. 1 to the Hodgkin and Huxley m2h equation described by McCormick and Huguenard (1992). Using this approach we found that Zn2+ significantly slowed both the activation time constant (τm) and the inactivation time constant (τh; Fig. 3A). When CaV3.3 cells were exposed to 100 μM Zn2+ a significant slowing down of current inactivation was observed (τh percent increase 363,6 ± 117,0% of control values, P < 0.05, repeated-measure ANOVA), whereas channel activation was not affected (Fig. 3B). Conversely, Zn2+ was ineffective at concentrations of ≤30 μM (Fig. 3B).
At voltages more negative than the threshold for CaV3.3 channel opening, the values of the time constant of the activation gate τm values were obtained by measuring the decay kinetics of the tail currents that appeared when the membrane was repolarized at different voltages after a step depolarization up to −11 mV (Fig. 4A). These tails relaxed with a bi-exponential kinetics characterized by fast (τ1) and slow (τ2) deactivation kinetic constants. By comparing the tail current τ values in the presence and in the absence of 300 μM Zn2+, we found that this metal slowed down the deactivation kinetics at all the tested voltages in the interval between −141 and −71 mV (Fig. 4, B and C). It is worth noticing that in the range of the voltages around −70 mV, i.e., close to resting potential in nRT neurons, τ1 values were similar in the presence and in the absence of Zn2+. Importantly, Zn2+ effect on deactivation kinetics was also observed when CaV3.3 cells were exposed to a concentration of 100 μM of this metal (τ2 percent increase 568.7 ± 140.4% of control values, P < 0.05) but not when concentrations of ≤30 mM were used (Fig. 3B).
At voltages more negative than the threshold for CaV3.3 channel opening, the values of the time constant of the inactivation gate (τh) were obtained by measuring the kinetics at which CaV3.3 channels recovered from inactivation. To this aim, CaV3.3 channels were first inactivated by applying a 10-s inactivating prepulse to −51 mV. Then the membrane was repolarized and step depolarizations up to −11 mV were delivered at progressively longer time intervals from the end of the prepulse (Fig. 5A). The repolarization voltage was set at different values (from –111 to −71 mV) in different groups of cells. Using this approach, we found that Zn2+ caused a significant slowing of recovery from inactivation when the repolarization voltage was set at −71 mV (Fig. 5B).
A bell-shaped voltage-dependence curve similar to what described by Hodgkin and Huxley (1952) for Na+ and K+ currents and by McCormick and Huguenard (1992) for native T-type currents was obtained when the values of the time constant of the activation gate τm values, plotted as a function of voltage, were fitted to the Eq. 12 reported in methods. The curve obtained by plotting the τm values in the presence of 300 μM Zn2+ was significantly larger at all the voltages tested in the range between −131 and −1 mV, reflecting both the slowing of channel activation at suprathreshold voltages and the slowing of channel deactivation at subthreshold voltages (Fig. 6A). Similarly, a bell-shaped curve was generated when the values of the time constant of the inactivation gate τh data were plotted as a function of the voltage and fitted to the equation 12 reported in methods (Fig. 6B). In the voltage range between −91 and −41 mV, the curve obtained in the presence of 300 μM Zn2+ was significantly larger than in control, reflecting the slowing effect of this ion on recovery from inactivation at subthreshold voltages and on current inactivation at suprathreshold voltages.
Effects of Zn2+ on the currents evoked by mock action potentials in CaV3.3 cells
To explore how Zn2+-induced modifications in CaV3.3 channel gating could affect the ability of CaV3.3 channels to generate action potentials, we measured the Ca2+ currents elicited by mock action potentials delivered to CaV3.3 cells through the patch pipette (Kozlov et al. 2005). Cells were first depolarized by delivering an ascending ramp from −101 up to +39 mV in 3 ms and, then, immediately repolarized by a second descending ramp bringing the potentials down to –101 mV in 6 ms. This protocol elicited large inward currents in cells perfused with 2 mM Ca2+ (Fig. 7). When the same cells were perfused with 2 mM Ca2+ +300 μM Zn2+, the currents were significantly smaller, peaked significantly later, and persisted for a significantly longer period of time because their relaxation kinetics was slowed (Fig. 7). These data indicate that, in the presence of Zn2+, Ca2+ influx through CaV3.3 channels is maintained for a significantly longer time interval than under control condition, thus suggesting that Zn2+ can influence excitability in cells expressing CaV3.3 channels.
Zn2+ effects on the frequency and duration of action potentials in simulated nRT cells
To establish whether Zn2+ modifies the frequency and/or the duration of neuronal action potential trains, we inserted Zn2+-induced changes in CaV3.3 channel gating parameters into the NEURON simulated model developed by Destexhe et al. (1996). This model reproduces the electrophysiological behavior of nRT neurons, a cell type where T-type channels participate in depolarization-induced action potentials and are required for triggering rebound burst firing. To run the nRT simulation, we introduced the kinetic parameters obtained from CaV3.3 cells exposed or not to Zn2+ into the IT2 model file, which provides to the NEURON program the parameters needed to build up the simulated T-type current (see methods). Conversely, the parameters describing the contribution of the other ion channels, i.e., voltage-gated Na+ and K+ channels, included in the NEURON model were not modified.
When the injection of a 200-ms hyperpolarizing current (0.3 nA) pulse was simulated, we observed under control conditions the appearance of a burst firing with the crescendo-decrescendo pattern typical of nRT cells; in contrast, this protocol was characterized by a dramatic increase in burst firing frequency in the presence of Zn2+ (Fig. 8A). It is worth noticing that the burst firing onset in the presence of Zn2+ was significantly delayed and that the simulated neuron returned to baseline condition in ∼300 ms from the beginning of burst firing. Similarly, when we simulated the injection of a 200-ms depolarizing current (0.3 nA) pulse, we observed an increase in action potential frequency in the presence of Zn2+ (Fig. 8B). To reproduce the Zn2+-induced blockade of CaV3.3 channels, both the simulation involving the injection of a negative current pulse and that involving the injection of a positive current pulse were run introducing into the model a value for Ca2+ whole cell conductance 55.5% lower that that used in control cell recording. Therefore these data suggest that, despite the reduction in the amplitude of the currents flowing through CaV3.3 channels by Zn2+, this metal caused a substantial increase in neuronal excitability.
Effects of Zn2+ chelation on spontaneous epileptiform activity in thalamocortical slices
Overall, the data obtained in vitro from HEK CaV3.3 cells and from the nRT computer model indicate that Zn2+ may exert an excitatory effect on thalamic activity. To further test this hypothesis, we explored the consequences of Zn2+ chelation in an acute talamocortical slice preparation. Spontaneous field activity was induced by perfusing these slices with 50 μM 4AP (Biagini et al. 2001). As illustrated in Fig. 9A (control panel), two different types of epileptiform field activity were observed: prolonged (>8 s) ictal-like discharges and short (0.5–4 s) interictal-like events. As expected with preserved thalamocortical connectivity, the majority of discharges occurred synchronously in the cortex, VB and nRT, with a few tens of ms delay between these structures (cf., Biagini et al. 2001). After a 15- to 30-min exposure to the Zn2+ chelator TPEN (300 μM), ictal discharges significantly decreased in frequency (Fig. 9A, TPEN panel) in all structures while interictal discharges remain unchanged. These data are summarized in Fig. 9, B and C. It should be emphasized that TPEN did not induce any significant change in the mean amplitude and duration of either ictal or interictal discharges and that TPEN-induced effect on ictal discharge occurrence was not reversible on drug washout.
The present paper demonstrates that high micromolar concentrations of Zn2+ slow down the gating kinetics of CaV3.3 channels while only partially reducing their Ca2+ permeability. We also provide evidence that Zn2+ induces an increase in the frequency of CaV3.3-dependent burst firing as indicated by incorporating the Zn2+-induced changes in gating kinetics into computer simulations of nRT neuron activity. Finally, this study shows that chelation of endogenous Zn2+ reduces the occurrence of ictal-like epileptiform events in thalamocortical slices maintained in vitro. Although it is well established that Zn2+ can influence the gating properties of several voltage-gated ion channels (Harrison and Gibbons 1994; Mathie et al. 2006), this is the first report to explore the effect of Zn2+ on T-type channel-dependent firing in simulated nRt neurons and the first to establish a role for endogenous Zn2+ in the activity of the thalamocortical system.
Surface charge screening by Zn2+ represents an obvious possible explanation for the effects induced by this ion on channel gating. In line with this view, it has been reported that like other metal ions, Zn2+ effectively binds to and neutralizes negative surface charges on the membrane, thus modifying the local electrical field around ion channels (Hille 1992). Even though the shielding effect of high concentrations of Zn2+ may have major consequences on channel gating, this mechanism does appear to be relevant when CaV3.3 cells were exposed to 300 μM Zn2+. In fact, while the Gouy-Chapman theory (Hille 1992) predicts that the half voltages for channel activation and inactivation and the τm and τh curves would be shifted rightward owing to surface charge screening, we did not find any significant shift suggesting that the concentration used in this study (i.e., 300 μM) was too low to significantly shield surface charges. Interestingly, it has been proposed that metal ions, such as Ni2+ or La3+—which are known to bind to ion channels—affect channel gating in a way that cannot be explained merely on the basis of the Gouy-Chapman theory of surface screening (Armstrong and Cota 1990; Elinder et al. 1996; Hille 1992). This could also be the case of Zn2+ that binds to the channel pore of voltage-dependent Ca2+ channels. In this context, it is tempting to speculate that Zn2+, while standing into the pore of CaV3.3 channels, may cause structural rearrangements of this channel region that contribute to the gating process. In line with this view, a crucial role in the gating process has been demonstrated for the pore region of T-type channels, the gating of which has been shown to be influenced by changes in the nature of the permeant ion (Shuba et al. 1991) or by point mutations in the selectivity filter of the channel pore (Talavera et al. 2003).
As the presence of Zn2+ dramatically slowed down CaV3.3 channel inactivation and deactivation kinetics, we believe that exposure of CaV3.3 to Zn2+ alters the function of this class of T-type channels during physiological processes that are characterized by rapid changes in membrane voltage. This is the case, for instance, of action potential generation and propagation. Therefore we hypothesize that Zn2+ would modify the response of CaV3.3 channels to incoming action potentials and/or their ability to generate hyperpolarization-induced low-threshold spikes, which entrain Na+-dependent action potential bursts (Perez-Reyes 2003). Our data presented support this idea. Indeed by exploring the ability of CaV3.3 channels to open in response to mock action potential, we found that the shape of Ca2+ currents was significantly different during Zn2+ application. Thus in the presence of this ion, the currents elicited by mock action potential were smaller but more persistent than under control conditions, suggesting that when Zn2+ is released in the extracellular compartment, Ca2+ influx through T-type channels lasts longer than under control conditions. Accordingly, using the NEURON computer simulation of nRT neurons (Dexthexe et al. 1996; Murbartian et al. 2004), we found that Zn2+ increases the frequency of burst firing. This type of computer simulation has already been adapted to CaV3.3 cells (Murbartian et al. 2004), and simulations in HEK-293 cells have been used to infer conclusions on neuronal T-type channel activity in physiological (Kozlov et al. 1999) and pathophysiological conditions (Vitko et al. 2005).
Assuming that our simulation data can be extrapolated to real nRT cells, the ability of Zn2+ to affect CaV3.3-dependent burst firing may have relevant implications in pathophysiological conditions in which the extracellular concentration of Zn2+ released from the terminals of the gluzinergic fibers, which impinge onto nRT neurons (Mengual et al. 2001), reaches values close to what used in our experiments (300 μM). The extracellular concentrations of Zn2+ have not been clearly established to date (Frederickson et al. 2006a; Kay 2003). It is, however, believed that Zn2+ release is negligible under basal conditions (Frederickson et al. 2006b), whereas it can increase up to 30–100 μM during ischemia or seizures (Thompson et al. 2002; Ueno et al. 2002). In addition, much higher concentrations of this ion are presumably reached in restricted areas around the site of release of Zn2+-containing vescicles. It should be emphasized that we observed relevant kinetic effects also in cells exposed to 100 μM Zn2+. In addition, it should also be noticed that, concomitantly with intense synaptic stimulation, as in epilepsy (Heinemann and Louvel 1983; Heinemann et al. 1977) or ischemia (Harris et al. 1981), extracellular Ca2+ concentration is expected to drop as a consequence of neuronal sink action. This could significantly shift rightward the Zn2+ effect-concentration curve assuming a binding competition between Ca2+ and Zn2+ as should happen if, as we hypothesized, the pore of CaV3.3 could represent the site for Zn2+ action on CaV3.3 channel kinetics.
On the basis of these considerations, we believe that our data have implications in brain ischemia or in epilepsy, two neurological conditions in which extracellular Zn2+ concentrations reach high micromolar values. This hypothesis is supported by the results obtained from acute thalamocortical slices exposed to the convulsant drug 4AP, which represents an in vitro model of epileptic discharge (Biagini et al. 2002; D'Arcangelo et al. 2002). We found that exposure to the Zn2+ chelator TPEN decreased the occurrence of 4AP-induced ictal-like discharges in nRT and cortex. TPEN presumably chelated the intracellular releasable Zn2+ as suggested by the ability of this drug to deplete brain terminals from Zn2+ (Cuajungco and Lees 1996). Despite its ability to chelate also iron and copper TPEN is usually considered a selective Zn2+ chelator, and it has been used to assess the role of this ion in neurotransmission in the hippocampus (Matias et al. 2006).
The ability of exogenous Zn2+ to increase the frequency of thalamocortical discharges in brain slices has been reported by Gibbs et al. (2000). However, no information is available on the physiological meaning of this Zn2+ action in the thalamocortical system; in addition, it was unknown whether endogenous Zn2+ could exert similar effects. At the best of our knowledge, our data represent the first evidence in favor of the idea that endogenous Zn2+ plays a role in thalamocortical excitability as shown by the dramatic consequences of its chelation. This raises the question of what are the sources of the endogenous Zn2+ affecting thalamocortical discharges. The rat cortex is rich in Zn2+, which mostly originates into the cortex itself as cortico-cortical connections are enriched in synaptic Zn2+ (Casanovas-Aguilar et al. 1998, 2002). In contrast, projecting thalamocortical fibers connecting the VB with cortex are not zincergic (Brown and Dyck 2005). However, dense zincergic terminals are concentrated in specific thalamic regions such as the dorsal portion of nRT, where we positioned our recording electrode; in addition, both anterodorsal and lateral dorsal thalamic nuclei display intense Timm staining (Mengual et al. 2002). Finally, zincergic terminals into the thalamus appear to have a cortical origin, thus identifying the zincergic thalamic afferents as a subset of glutamatergic corticothalamic fibers. As the ventrobasal thalamus is not significantly innervated by zincergic fibers and exogenous Zn2+ does suppress the activity of thalamic relay neurons (Noh and Chung 2003), it is unlikely that the decrease in VB ictal firing induced by Zn2+ chelation could have been determined by a direct effect on thalamic relay neurons. Conversely, the most likely explanation for TPEN effect is a decrease in the excitatory drive by afferent fibers of nRT and/or cortical origin. In line with this hypothesis comes a study, which was published while the present paper was under review, showing that Zn2+ does affect the gating of the nRT-specific T-type channel isoform CaV3.3 but not that of the T-type channel isoform CaV3.1 that is preferentially expressed in thalamic relay neurons (Traboulsie et al. 2007).
A limitation of our study is that we did not include in our computer simulation the effect of Zn2+ on other ion channels such as voltage-gated Na+, Ih, and type A K+ channels or ligand-gated GABAA channels, which are expressed in nRT neurons and have a role in determining their electrical properties (Abbass 2006; Destexhe and Sejnowski 2003; Yue and Huguenard 2001). Evidence reported in the literature suggests that some of these mechanisms could cooperate in a determining the pro-excitatory effect of this metal ion. For instance, it has been reported that Zn2+ is ineffective on voltage-gated Na+ channels or induce only a marginal decrease in Na+ currents (Harrison and Gibbons 1994; Horning and Trombley 2001), whereas the effect on type A K+ channels is strictly voltage dependent as these currents are enhanced by Zn2+ at membrane potentials more positive than −50 mV and inhibited at more negative potentials, such as those applied by us in the simulation experiments (Horning and Trombley 2001). As far as Ih channels are concerned, very few data are available on Zn2+ effect on these channel type, which is inhibited only by millimolar Zn2+ concentrations in neuroendocrine cells (Raymond and Lapied 1999). Finally, GABAA receptor blockade in the nRT has been reported to contribute to the pro-excitatory action of exogenous Zn2+ (Gibbs et al. 2000). However, it is unlikely that this can be the only factor because we have found that exogenously added Zn2+ increases the frequency of thalamocortical epileptiform discharges in the presence of saturating concentrations of the GABAA receptor antagonist picrotoxin (Cataldi and Avoli, unpublished data).
In conclusion, endogenous Zn2+ controls thalamocortical firing, and this effect can be at least in part determined by its ability to slow down the gating of CaV3.3 channes, the major T-type channel isoform expressed in nRT neurons.
This study was supported by the following grants: Programma Speciale art. 12bis comma 6, D. Lgs. 229/99, Special Project “Alzheimer 2001/2004” to L. Annunziato from the Italian Ministry of Health and Regione Campania; COFIN 2004 to M. Cataldi and L. Annunziato from the Italian Ministry of Education, University and Research; PNR-FIRBRBNE01E7YX_007 2001 to L. Annunziato from the Italian Ministry of Education, University and Research; Ricerca Finalizzata, legge 502/92 “Geni di vulnerabilità e di riparazione DNA” to L. Annunziato from the Italian Ministry of Health; Legge regionale 28/5/02, Finanziamento 2003 to L. Annunziato from Regione Campania; Programma Operativo Regionale (POR), Centro regionale di competenza GEAR to L. Annunziato from Regione Campania; 12th Italian-Chinese Executive program for scientific and technological Cooperation for the period 2006–2009 to L. Annunziato from the Italian Foreign Ministry; the CIHR (MOP 8109) and the Savoy Foundation to M. Avoli. V. Lariccia was supported by a fellowship from the Centro Regionale di Competenza di Genomica Funzionale. M. Cataldi was supported by a travel grant from the Federico II University of Naples, “ Programma di scambi internazionali con universitá ed istituti di ricerca stranieri per la mobiltá di breve durata di docenti, ricercatori e studiosi.”
The authors are indebted to Dr. E. Perez-Reyes for kindly providing the stably transfected HEK-293, to Dr. G. Panuccio for friendly help, to V. Grillo and C. Capitale for technical support, and to Dr. P. Merolla for editorial help.
The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
- Copyright © 2007 by the American Physiological Society