INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

The piriform region comprises both the olfactory cortex and associative structures implicated in olfactory functions. The latter include a nucleus localized in between the piriform cortex (PC) proper and the caudatus-putamen, named endopiriform nucleus (EPN; Fig. 1A), which has been considered by some authors as the fourth layer of the PC (O'Leary 1937; Valverde 1965), and by others as an anatomically independent structure (Haberly 1998; Loo 1931). The EPN contains multipolar cells of medium to large size (Tseng and Haberly 1989a) with long axons that form a net of intrinsic connections and project over long distances to the olfactory and limbic cortices (Behan and Haberly 1999; Haberly and Price 1978; Krettek and Price 1978; Luskin and Price 1983a,b). From a functional point of view, the importance of EPN is believed to consist in its participation in both olfactory processing and pathological events, notably epileptogenesis (Demir et al. 1999; Hoffmann and Haberly 1993, 1996; Piredda and Gale 1985). It has been proposed that such roles of EPN depend, at least in part, on the ability of its neurons to generate sharp, low-threshold, regenerative depolarizing events and burst firing (Tseng and Haberly 1989a,b) that would synchronize large populations of cortical neurons via the widely distributed EPN projections to the olfactory cortices (Behan and Haberly 1999).

View larger version (66K): [in this window] [in a new window]   Fig. 1. Location of the endopiriform nucleus (EPN) and exemplary voltage-gated Ca2+ currents in EPN neurons. A, top: schematic drawing of a coronal section of a guinea pig brain hemisphere. The section level is approximately 12.5 mm rostral to lambda. The EPN is depicted in gray. Other structures shown: NC, neocortex; Cl, claustrum; CPu, caudatus-putamen; ml, corpus callosum; rs, rhinal sulcus; lot, lateral olfactory tract; PC, anterior piriform cortex; I, II, III, PC layers I-III; OTu, olfactory tubercle. In the bottom panel the region of piriform cortex and EPN (corresponding to the rectangle outlined in the top panel) of a thionin-stained slice is shown at a higher magnification. B: whole cell currents recorded in a representative EPN neuron (cell B9525), in response to the voltage-clamp protocol shown in B1, in the presence of either 5 mM extracellular Ca2+ (B2) or Ba2+ (B3). Calibration bars: 300 pA, 50 ms. The inset in B2 shows the current-voltage (I-V) relationships derived from the same current families (, Ca2+; , Ba2+). In the inset of B3, the currents traces (arrows) recorded at the test potentials of 30 mV in Ca2+ (asterisk) and 40 mV in Ba2+ are superimposed and shown at a higher gain. Calibration bars: 20 pA, 50 ms.

The intrinsic membrane properties of EPN potentially involved in such membrane potential events are not completely understood yet. Some differences in membrane properties of deep multipolar neurons in comparison to pyramidal cells in the overlying PC have been reported. Such differences include a less negative resting membrane potential, a higher time constant, and the ability to generate sustained, Ca²⁺-dependent depolarizing spikes within an unusually negative window of membrane voltages (Tseng and Haberly 1989b). It has also been shown that, in contrast with most layer II pyramidal neurons, pyramidal and multipolar neurons dissociated from both deep PC and EPN consistently express a robust low-voltage-activated, T-type Ca²⁺ current (Magistretti and de Curtis 1998). High-voltage-activated (HVA) Ca²⁺ currents have been extensively characterized in layer II neurons (Magistretti et al. 1999, 2000), but so far no exhaustive data are available for EPN neurons. Finally, it has been demonstrated that the transient potassium current, I_A, in EPN cells has biophysical properties that could contribute to the generation of sustained depolarizations (Banks et al. 1996).

Since voltage-dependent Ca²⁺ channels are determinant for shaping firing patterns and promoting intracellular propagation of excitation in central neurons (Llinás 1988) and in the EPN in particular (Tseng and Haberly 1989b), we undertook a study of HVA Ca²⁺ currents in EPN multipolar neurons to identify possible candidate currents responsible for the regenerative depolarizing events these neurons can generate.

Part of the present data has been reported in abstract form (Brevi et al. 2000).

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Cell preparation

Female Hartley guinea pigs (7-21 day old) were anesthetized with an intraperitoneal injection of pentothal sodium (20 mg/kg) and decapitated, according to a procedure approved by the Institute's Ethical Committee and international regulations on animal research. After extracting the brain under hypothermic conditions, each hemisphere was cut into 500-µm-thick slices normal to the main axis of the lateral olfactory tract using a McIlwan tissue chopper (Mickle, Gomshall, UK). The EPN underlying the anterior PC (see Fig. 1A) was dissected from each slice under microscopic control, and its neurons were isolated by an enzymatic and mechanical dissociation procedure described elsewhere (Magistretti and de Curtis 1998; Magistretti et al. 1999). During extraction and dissection the tissue was submerged in an ice-cold solution containing (in mM) 115 NaCl, 3 KCl, 3 MgCl₂, 0.2 CaCl₂, 20 PIPES-Na, and 25 D-glucose (pH 7.4 with NaOH), and bubbled with pure O₂. Dissociated cells were seeded in the recording chamber and left to settle for 15 min before starting the recording.

Patch-clamp recordings

The recording chamber was mounted on the stage of an Axiovert 100 microscope (Zeiss, Oberkochen, Germany). Cells were perfused at about 0.5 ml/min with an oxygenated extracellular solution suitable for isolating Ba²⁺ currents flowing through Ca²⁺ channels containing (in mM) 5 BaCl₂, 88 choline-Cl, 40 tetraethylammonium (TEA)-Cl, 3 KCl, 2 MgCl₂, 3 CsCl, 10 HEPES, 5 4-aminopyridine, and 25 D-glucose, pH 7.4 with HCl. Cells were observed at ×400 magnification.

Patch pipettes fabricated from thick-wall borosilicate glass capillaries (GC 150-7.5; Clark Electromedical Instruments, Reading, UK) were filled with a solution containing (in mM) 78 cesium methanesulphonate (CsMeSO₃, obtained by neutralizing CsOH with equimolar methanesulphonic acid), 40 TEA-Cl, 10 HEPES, 10 EGTA, 20 phosphocreatine (di-Tris salt), 2 ATP (Mg²⁺ salt), 1 adenosine 3'-5' cyclic monophosphate, and 20 U/ml creatine phosphokinase, pH adjusted to 7.2 with TEA-OH. Pipette input resistance was 5-8 M. Tight seals (>10 G) and the whole cell configuration were obtained according to the standard technique (Hamill et al. 1981). Voltage-clamp recordings were performed at room temperature (~22°C) by means of an Axopatch-1D amplifier (Axon Instruments, Foster City, CA). Series resistance (R_s) and cell capacitance (C_m) were evaluated on-line by canceling the whole cell capacitive transients evoked by 10-mV voltage square pulses with the amplifier's compensation section, and reading out the corresponding values. R_s (18.3 ± 0.7 M, mean ± SE, n = 61) was always compensated by ~80% with the amplifier's built-in circuitry, and continuously monitored during the experiment. The lag potentiometer of the compensation section was set to the lowest values compatible with ringing avoidance (normally about 20 µs). C_m averaged 6.8 ± 0.3 pF (n = 61; extreme values: 3.5 and 16.0 pF). Voltage protocols were commanded using the Clampex program of the pClamp 6.0.3 software package (Axon Instruments), and current signals were acquired with a Pentium personal computer interfaced to an Axon DigiData 1200 converter. In all recordings the general holding potential was 70 mV. Current signals were filtered at 5 kHz, digitized at 10 or 20 kHz, and on-line leak subtracted via a P/4 protocol.

Drugs were applied via a laminar solution flux generated by a local-perfusion system consisting of a multibarrel pipette with a tip diameter of 150 µm, connected to six perfusion channels individually operated by electrovalves. The tip of the perfusion pipette was positioned in close proximity of the recording site. Concentrated stock solution of drugs were prepared in small aliquotes stored at 20°C. Nifedipine (Sigma) was dissolved in dimethylsulfoxide (DMSO) at 10 mmol/l; -conotoxin GVIA (-CTx GVIA; Alomone Labs, Jerusalem, Israel) and -conotoxin MVIIC (-CTx MVIIC; Bachem, Bubendorf, Switzerland) were dissolved in pure water at 500 µmol/l and 1 mmol/l, respectively. The aliquots were then diluted to final concentrations in the recording solution at 10 µmol/l, 500 nmol/l, and 1 µmol/l, respectively. -CTx MVIIC was dissolved in the presence of lysozime (Sigma, 1 mg/ml) to minimize aspecific binding of the peptide to recipient walls. All additional substances used for each drug (DMSO and lisozyme) were also added, in the same amounts, to the control solution and other drug-containing solutions. Because of its light sensitivity, nifedipine was prepared and stored in the dark, and its perfusion channel was light shielded with aluminum foil.

Data analysis

Current traces were analyzed by means of the Clampfit program of pClamp 6.0.3. Currents were normally refiltered off-line at 1-2 kHz. Maximal voltage error due to series resistance (V_Rs) was estimated in each cell as V_Rs = I_peak · R_s · (1 - f), where I_peak is the maximal current amplitude at the peak of the current-voltage (I-V) relationship, and f is the fractional compensation of R_s. In 61 cells, V_Rs averaged 3.5 ± 0.3 mV. Fractional current persistence during 300-ms test voltage pulses was measured as the ratio between the current amplitude at t = 300 ms and the peak current amplitude (R_300/p). Ba²⁺ permeabilities (P_Bas) were calculated from peak current amplitudes (I_Bas) by applying the constant-field equation in the form

in which the nominal intra- and extracellular Ba²⁺ concentration values (0 and 5 mM, respectively) were introduced. Data were fitted with exponential functions, I =  A_i · exp(-t/_i) + C, using Clampfit, and with Boltzmann functions, P_Ba = P_Ba(max)/{1 + exp[(V - V_1/2)/k]}, and linear functions using Origin 3.06 (MicroCal Software, Northampton, MA). In linear fittings, the parameter P indicates the probability of the correlation coefficient to be zero. Average values are expressed as means ± SE. Statistical significance was evaluated by means of the two-tail Student's t-test for unpaired data.

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Voltage-dependent Ca²⁺ currents were recorded in 72 large multipolar neurons acutely dissociated from the anterior part of the EPN (see Fig. 1A). Currents from 61 neurons were selected for analysis based on low and stable series resistance levels, high-input resistance levels, and high signal-to-noise ratio. To evoke both low-voltage-activated (LVA) and HVA currents, 300-ms depolarizing test pulses at various voltage levels (90 to +30 mV) were delivered after a 2-s conditioning prepulse at 100 mV. A typical example of the currents recorded in response to this protocol in extracellular Ca²⁺ (5 mM; n = 8) is shown in Fig. 1B1. Ca²⁺ currents had a "threshold" at about 50/40 mV, and peaked at +10/+20 mV (Fig. 1B1, inset). Replacement of extracellular Ca²⁺ with equimolar Ba²⁺ always resulted in the following effects: 1) a ~10-mV shift in the negative direction of the I-V relationship (Fig. 1B1, inset), which is consistent with the lower surface-charge shielding effect of Ba²⁺ as compared with Ca²⁺ (see Hille 1992); 2) a prominent decrease of time-dependent current inactivation at all test voltage levels (Fig. 1B2), which is consistent with the removal of Ca²⁺-dependent inactivation of HVA currents; and 3) a marked increase in current maximal amplitude (by 1.85 ± 0.18 times at the peak of the I-V relationship; n = 8). Close to current threshold, however, current amplitude was actually decreased, rather than increased, by the substitution of Ca²⁺ with Ba²⁺ (Fig. 2B2, inset). This is in agreement with our previous observation that EPN neurons consistently express a LVA, T-type current (Magistretti and de Curtis 1998), since T-type channels, in contrast to most HVA channels, are known to be less permeable to Ba²⁺ than Ca²⁺ (see Carbone and Swandulla 1989; Hille 1992).

View larger version (28K): [in this window] [in a new window]   Fig. 2. Isolation and properties of low-voltage-activated (LVA), T-type currents carried by Ba2+ in EPN neurons. A: IBas recorded in a representative EPN neuron (cell A9517) in response to the voltage-clamp protocol shown in A1. The currents in A2 were evoked by pulses (at the voltage levels indicated) preceded by a prestep at either 100 mV (asterisk) or 60 mV; those in A3 were obtained by subtracting the latter from the former. Calibration bars: 50 (A2) or 20 (A3) pA, 50 ms. The inset in A3 shows the average, normalized I-V relationship of the IBaTs isolated with the above subtraction procedure (n = 8). The I-Vs derived from individual cells were normalized for their peak values and averaged among cells, and the average I-V thus obtained was further normalized to its peak value. Open symbols represent data points measured from unsubtracted traces (prestep at 100 mV, test pulses at 80 to 60 mV). B: lack of effect of 50 µM Ni2+ on EPN IBaT. B1 shows the IBas recorded in another neuron (cell B9528) at the test potential of 20 mV, starting form a conditioning potential of either 100 (asterisks) or 60 mV, both in the absence and in the presence of 50 µM Ni2+ (calibration bars: 50 pA, 50 ms). B2 shows the IBaTs isolated by subtraction in the 2 conditions, superimposed to emphasize the lack of effect of Ni2+ (calibration bars: 25 pA, 50 ms).

From this point on, the analysis will be limited to currents recorded in 5-mM extracellular Ba²⁺ (I_Bas).

Ni²⁺-resistant T-type current in EPN neurons

We have previously shown that the T-type current recorded in EPN neurons in 5-mM extracellular Ca²⁺ is steady-state inactivated by more than 90% at 60 mV (Magistretti and de Curtis 1998). Given the ~10-mV negative shift caused on I-V dependence by substitution of Ca²⁺ with Ba²⁺, this voltage level was even more effective in inactivating the T-type current recorded in extracellular Ba²⁺ (I_BaT). Indeed, I_BaT could be properly isolated by subtracting I_Bas elicited after a 2-s conditioning prepulse at 60 mV from I_Bas elicited from 100 mV, according to the protocol illustrated in Fig. 2A1. Typical currents obtained in this way are shown in Fig. 2, A2 and A3. The I-V relationship of I_BaT had a "threshold" negative to 50 mV and peaked at about 20 mV (Fig. 2A3, inset). These and other biophysical properties of I_BaT will be compared with those of R-type I_Bas expressed by the same EPN neurons in a later section of this paper.

Ni²⁺ ions have been reported to exert a potent, relatively selective blocking action on T-type currents in some neuronal populations, but not in others (see Carbone and Swandulla 1989). We tested the effects of a moderate concentration of Ni²⁺ (50 µM) on I_BaT in the neurons under study. Although 50 µM Ni²⁺ reduced the amplitude of total HVA I_Bas by an average of 43.7 ± 19.4% at the peak of the I-V relationship (n = 5; not shown), it had no detectable effects on I_BaT (Fig. 2B; n = 5). Hence the T-type current expressed by EPN neurons must be classified as "Ni²⁺ resistant."

Properties of total HVA I_Bas

The properties of HVA I_Bas expressed by EPN neurons were then investigated. For isolation of HVA currents, 300-ms depolarizing test pulses were preceded by a 2-s conditioning prepulse at 60 mV. The average current density of I_Bas evoked in this way was 169.8 ± 10.9 pA/pF at the peak of the I-V relationship (n = 61). I_Bas were completely abolished by application of 200 µM CdCl₂, after which significant outward-current contaminants were occasionally observed only at test potentials positive to +20 mV (Fig. 3A3). This confirms that the currents under study were purely mediated by Ca²⁺ channels.

View larger version (42K): [in this window] [in a new window]   Fig. 3. Kinetic diversity of total high-voltage-activated (HVA) IBas in EPN neurons. A: total IBas recorded at various test potentials from a prestep level of 60 mV (A1) in 3 different neurons (A2, cell A9517; A3, cell A9907; A4, cell B9528). The inset in A2 shows the normalized IBa-V relationships derived from the same 3 cells. The bottom traces in A3 were recorded after application of 200 µM Cd2+. Calibration bars: 300 pA, 50 ms. B: frequency-distribution diagram of the index of current persistence, R300/p (see text for explanations), for the whole population of neurons recorded. The inset is the scatter plot of R300/p as a function of animal postnatal age (P) for the same cell population. The straight line is the best linear fit to data points (slope factor = +0.0136 day-1, with P = 0.027). C: normalized plots of Ba2+ permeabilities (PBa, derived from peak IBas as explained in the text) as a function of test potential for the same 3 cells as illustrated in A. The smooth lines are Boltzmann fittings to data points, with the following fitting parameters: V1/2 = 3.2 mV, k = 6.6 mV (); V1/2 = 1.8 mV, k = 6.4 mV (); V1/2 = +6.4 mV, k = 6.8 mV (). D: scatter plot of PBa half-activation potential (V1/2) as a function of R300/p for the whole neuronal population recorded. The straight line is the best linear fit to data points (slope factor = +14.2 mV, with P = 0.0015).

Total I_Bas recorded in different cells displayed considerable variability in their kinetic behavior. The current decay of total I_Bas could be either very slow (Fig. 3A4) or fast (Fig. 3A2), or show an intermediate kinetic behavior (Fig. 3A3). To quantify the variability of current inactivation rate, we used an inverse index of current decay, R_300/p, which represents the ratio between the current amplitude at the end of a 300-ms depolarizing test pulse and the peak current amplitude (see Magistretti et al. 1999). In each cell, the current trace corresponding to the peak of the I-V relationship was used for deriving R_300/p. This parameter approached 0 in high-decay I_Bas, and 1 in low-decay currents. As illustrated in the frequency-distribution diagram of Fig. 3B, R_300/p values were dispersed between 0.1 and 1, although a clear peak corresponding to 0.6-0.7 was observed, and in an additional, substantial cell fraction, R_300/p was comprised between 0.2 and 0.6. A slight, positive correlation was found between R_300/p and increasing animal postnatal ages, within the age window considered in this study [postnatal days 7-21 (P7-P21); see Fig. 3B, inset]. Conversely, no correlation was observed between R_300/p and cell capacitance [the regression coefficient obtained from the R_300/p (C_m) scatter plot was 0.0022 pF^-1, with P = 0.824; not shown], thus indicating that I_Ba kinetic diversity did not depend on the amount of cell membrane available in each dissociated neuron.

Long-lasting (5 s) depolarizations allowed for better appraisal and quantification of total I_Ba decay kinetics. Figure 4A illustrates the currents recorded in two representative neurons in response to 5-s square pulses at 0 mV. In general, the decay phases of currents elicited in this way could be fitted with triple exponential functions. Panels B1 and B2 of Fig. 4 are scatter plots of the values of time constants (_ina) and normalized amplitude coefficients returned by triple-exponential fittings of the currents recorded in seven neurons. The fast _ina ranged from ~50 to ~130 ms, the intermediate _ina from ~225 to ~820 ms, and the slow _ina from ~1.1 to ~6.0 s. Different combinations of _ina and amplitude-coefficient values determined the specific decay behavior of different currents. For instance, in the low-decay current of A2 (R_300/p = 0.76), the relative amplitude of the fast exponential component was zero, whereas in the faster decaying current of A2 (R_300/p = 0.56), the fast exponential component (_ina = 129.4 ms) had a relative amplitude-coefficient value of 0.25. 

View larger version (16K): [in this window] [in a new window]   Fig. 4. Inactivation kinetics of total high-voltage-activated (HVA) IBas. A: total IBas recorded in response to 5-s depolarizing steps at 0 mV in 2 representative neurons (A1, cell C9831; A2, cell A9903). Calibration bars: 200 (A1) or 400 (A2) pA, 1 s. The enhanced lines are exponential fittings to IBa decay phase, with the following fitting parameters: 1 = 129.4 ms, A1 = 180.9 pA, 2 = 612.5 ms, A2 = 255.8 pA, 3 = 3,642.9 ms, A3 = 273.9 pA, C = 0 pA (A1); A1 = 0 pA, 2 = 389.8 ms, A2 = 529.1 pA, 3 = 2,296.0 ms, A3 = 944.2 pA, C = 158.7 pA (A2). B: scatter diagrams of the fitting parameters returned by exponential fitting of IBa decay in 7 cells. Fittings were obtained from 5-s current traces at Vtest = 0 mV, as illustrated in A. B1: time constants (note the semilogarhythmic scale). B2: normalized amplitude coefficients (A1 + A2 + A3 + C = 1). The crosses are average values of each data set.

The voltage dependence of activation of total I_Bas was then investigated. In currents characterized by fast inactivation kinetics, both activation threshold and peak of the I-V relationship were often observed at more negative voltage levels than in slower currents (see inset in Fig. 3A2). Since a correlation between inactivation kinetic and voltage dependence of activation in total I_Bas was demonstrated for PC layer-II neurons (Magistretti et al. 1999), the existence of such a correlation was also investigated in EPN neurons. Ba²⁺ permeability values (P_Bas) were derived from peak current amplitudes (I_Bas) by applying the constant-field equation as explained in METHODS. For each cell a plot of P_Ba as a function of test potential [P_Ba(V)] was constructed, and data points were fitted with single Boltzmann functions (Fig. 3C). The half-activation voltage (V_1/2) returned by each fitting was then plotted as function of the R_300/p value of the corresponding cell, in a total of 61 cells (Fig. 3D). The best linear fit to the data points returned a slope coefficient equal to 14.2 mV, indicating the existence of a significant correlation between the two parameters. By contrast, virtually no correlation was observed between the voltage error due to series resistance (V_Rs: see METHODS) and R_300/p, nor between V_Rs and V_1/2. Indeed, linear fittings of the V_Rs(R_300/p) and V_Rs(V_1/2) scatter plots returned extremely low slope coefficients (1.84 mV and +0.037, respectively, with P = 0.199 and 0.336, respectively; not shown). It can be concluded that currents characterized by fast kinetics also activate in a more negative range of membrane voltages than slower-kinetics currents.

Pharmacological analysis of HVA I_Bas: properties of L-, N-, and P/Q-type currents

To elucidate the nature of the Ca²⁺ channels that underlie total I_Bas and their biophysical diversity in EPN neurons, we investigated the effects of Ca²⁺-channels blockers selectively active on specific Ca²⁺-channel types. We used the L-type channel blocker, nifedipine, the N-type channel blocker, -conotoxin GVIA (-CTx GVIA), and the N- and P/Q-type channel blocker, -conotoxin MVIIC (-CTx MVIIC), at saturating concentrations (10, 0.5, and 1 µM, respectively). To avoid the development of cumulative channel inactivation during repetitive delivery of long-lasting depolarizing test pulses, the effects of drug application were monitored on I_Bas evoked by fast (2.6 mV/ms) depolarizing ramps, which returned "instantaneous" I-V curves satisfactorily overlapping with those constructed from step protocols (see Fig. 5A, inset). Single, standard I-V step protocols were commanded after the effect of each drug had reached a steady state. Nifedipine blocked a fraction of total I_Bas equal on average to 24.7 ± 5.4% (n = 11), which was identified as L type. Its effect was readily and fully reversible (see Fig. 5A). -CTx GVIA blocked a fraction of total I_Bas equal on average to 27.1 ± 3.4% (n = 24), identified as N type. Its effect was largely irreversible over the time scale of our experiments (up to 10 min; not shown). -CTx MVIIC, which was always applied after the block of the N-type current fraction with -CTx GVIA (see Fig. 5A), abolished a fraction of total I_Bas equal on average to 22.2 ± 2.4% (n = 16), identified as P/Q type. The effects of conotoxins and that of nifedipine were largely additive (n = 3; see Fig. 5A).

View larger version (26K): [in this window] [in a new window]   Fig. 5. Effects of L-, N-, and P/Q-type channel blockers on HVA IBas. A: effects of sequential application of nifedipine (Nif., 10 µM), -CTx GVIA (0.5 µM), and -CTx MVIIC (1 µM) on IBa peak amplitude in a representative neuron (cell B0105). Currents were evoked by fast (2.6 mV/ms) depolarizing ramps delivered every 7 s. The overlapping of a ramp-evoked current (recorded in control conditions) with the I-V relationship constructed from a step protocol in the same cell is illustrated in the inset. B: IBa traces recorded within I-V step protocols (Vtest = +10 mV) commanded during steady-state effects of drug or control-solution application, at the time points indicated by numbers and arrows. Calibration bars: 150 pA, 50 ms.

The biophysical properties of individual blocker-sensitive current components were then analyzed. Drug-sensitive currents obtained by subtraction in some representative neurons are illustrated in Fig. 6. Nifedipine-sensitive currents consistently displayed low tendency to inactivate over the 300 ms of the routinely applied depolarizing test pulses, regardless of the kinetic behavior of the corresponding total I_Bas (Fig. 6, A1-A3 and D1 and D2). In these currents, R_300/p was always higher than 0.65 (Fig. 7A, ), averaged 0.77 ± 0.02 (n = 7), and was significantly higher than in the corresponding total I_Bas. Indeed, the ratio between R_300/p in nifedipine-sensitive currents and R_300/p in control currents ("R_300/p ratio") averaged 1.30 ± 0.17. -CTx GVIA-sensitive currents showed a larger variability in their decay kinetics (extreme values for R_300/p = 0.30 and 0.87; see Fig. 7A, ), although typically a low-to-moderate tendency to inactivate was observed (average R_300/p = 0.60 ± 0.06, n = 14; Fig. 6, B1-B3 and E1 and E2). Average R_300/p ratio (with respect to the corresponding total I_Bas) was also >1 in -CTx GVIA-sensitive currents (1.23 ± 0.09). As to currents selectively sensitive to -CTx MVIIC, in most cases (n = 6 of 10) they displayed low tendency to inactivate (Fig. 6, C1-C3 and F2), with an average R_300/p equal to 0.68 ± 0.04. A slow kinetic behavior has been considered to be typical of "P-type" currents proper (Hilaire et al. 1996; Randall and Tsien 1995; Teramoto et al. 1995). In four other cells, characterized by relatively high-decay total I_Bas (0.41 ± 0.05), -CTx MVIIC-sensitive currents displayed a clearly higher tendency to inactivate (Fig. 6F1), with an average R_300/p of 0.34 ± 0.05. Higher inactivation speed, as compared with P-type currents, has been considered as a distinctive property of "Q-type" currents proper (Hilaire et al. 1996; Randall and Tsien 1995; Teramoto et al. 1995; Wang et al. 1997).

View larger version (42K): [in this window] [in a new window]   Fig. 6. Isolation and decay properties of single blocker-sensitive HVA current components. A-C: isolation of nifedipine- (A), -CTx GVIA- (B), and -CTx MVIIC- (C) sensitive current components in 3 representative neurons (A, cell B0105; B, cell A9D28; C, cell C9618). The voltage protocol applied was the same as illustrated in Fig. 3A1. Currents in row 3 (blocker-sensitive currents) were obtained by subtracting currents in row 2 (recorded during drug application) from those in row 1 (control condition). Calibration bars: 200 pA, 50 ms. D-F: comparison of blocker-sensitive currents with original, total IBas. Total IBas recorded at the peak of the I-V relationship and characterized by either low (row 1) or high (row 2) R300/p values are shown superimposed to blocker-sensitive components (asterisks) obtained by subtraction at the same voltage level. All of the currents are normalized to their peak values. Recordings are from 5 different neurons (D1 and E1, D9D09; F1, C9D09; D2, B9D30; E2, C9910; F2, C9901).

View larger version (20K): [in this window] [in a new window]   Fig. 7. Inactivation and voltage dependence parameters of single blocker-sensitive HVA current components. A: scatter diagram of R300/p values in nifedipine-, -CTx GVIA-, and -CTx MVIIC-sensitive current components (L-, N-, and P/Q-type, respectively), and blocker-resistant currents (R-type). Open vs. filled triangles distinguish high-decay (putative "Q-type") vs. low-decay (putative "P-type") -CTx MVIIC-sensitive currents, respectively. The crosses are average values of each data set. B: scatter plots of PBa half-activation potential (V1/2) as a function of R300/p for N-type (B1) and P/Q-type (B2) current components. The straight lines are the best linear fits to data points (slope factor = +7.6 mV in B1; +30.8 mV in B2, with P = 0.265 and 0.005, respectively).

The voltage-dependent properties of each pharmacological current component were also investigated. Voltage dependence of activation was analyzed by constructing P_Ba(V) plots for the single pharmacological components. Boltzmann best fittings to data points returned V_1/2 and k values of 4.6 ± 0.6 mV and 7.2 ± 0.5 mV, respectively, for nifedipine-sensitive currents (n = 5); 3.4 ± 1.3 mV and 6.9 ± 0.7 mV, respectively, for -CTx GVIA-sensitive currents (n = 13); +2.9 ± 2.1 mV and 9.1 ± 1.4 mV, respectively, for low-decay ("P-type") -CTx MVIIC-sensitive currents (n = 6); and 8.7 ± 1.7 mV and 7.4 ± 0.9 mV, respectively, for high-decay (Q-type) -CTx MVIIC-sensitive currents (n = 4). Similarly to total I_Bas, the possible correlation between voltage dependence of activation and decay kinetics was also analyzed in -CTx GVIA- and -CTx MVIIC-sensitive currents. The V_1/2(R_300/p) plot revealed poor correlation for -CTx GVIA-sensitive currents (Fig. 7B1). On the contrary, the correlation was high for -CTx MVIIC-sensitive currents (Fig. 7B2). This finding supports the tentative subdivision of -CTx MVIIC-sensitive currents expressed by EPN neurons into functionally distinct P-type currents (filled triangles in Fig. 7, A and B2) and Q-type currents (open triangles in Fig. 7, A and B2).

Pharmacological analysis of HVA I_Bas: properties of R-type currents

The simultaneous application of the saturating concentrations of nifedipine, -CTx GVIA, and -CTx MVIIC blocked a current fraction equal on average to 68.5 ± 6.6% (n = 10). This value matches reasonably with the sum of the average current fractions individually blocked by nifedipine, -CTx GVIA and -CTx MVIIC (74.0%), as determined in a larger sample of cells. The residual currents insensitive to the combined action of these blockers were identified as R-type (I_R).

Although I_Rs also showed some kinetic variability, their decay speed was consistently faster than in the corresponding, total I_Bas, with an average R_300/p ratio of 0.51 ± 0.11 (n = 8). In three cells, I_Rs displaying moderate tendency to inactivate during the routinely applied 300-ms depolarizing pulses (R_300/p between 0.47 and 0.61) were observed. An example of such currents is illustrated in Fig. 8. The analysis of voltage dependence of these I_Rs returned an average half-activation potential of 6.6 ± 2.7 mV, with k = 5.7 ± 0.2. 

View larger version (21K): [in this window] [in a new window]   Fig. 8. R-type currents with relatively slow decay kinetics. A: currents recorded in a representative neuron (cell A0324) either under control conditions (A1) or during the application of 10 µM nifedipine + 0.5 µM -CTx GVIA + 1 µM -CTx MVIIC (A2). The voltage protocol applied was the same as illustrated in Fig. 3A1. Calibration bars: 200 pA, 50 ms. B: control and blocker-resistant (asterisk) currents (from the same cell as above) recorded at the peak of the I-V relationship, normalized to their peak amplitudes and superimposed.

In the majority of cases (5 of 8), however, the fractional decay of I_R was particularly high, with an average R_300/p of 0.16 ± 0.05 at the peak of the I-V relationship. Two typical examples of these fast-decaying I_Rs are illustrated in Fig. 9, A and B. Besides displaying a higher inactivation speed than the corresponding total I_Bas (Fig. 9, A3 and B3) and the global drug-sensitive current fraction, fast-decaying I_Rs also consistently activated at more negative voltage levels. This is illustrated in Fig. 9C, which compares the normalized current amplitudes at the peak of the I-V relationship and at the voltage levels of 40 and 30 mV for both the fast-decaying I_R and the corresponding, global blocker-sensitive current fraction (average currents from 5 cells). It is apparent that at 40 mV a sizeable I_R was activated as compared with no detectable blocker-sensitive I_Ba, and at 30 mV the fractional activation was much higher in the fast-decaying I_R than in the blocker-sensitive I_Ba. Figure 9D shows the average, normalized I-V relationships for both fast-decaying I_Rs and the corresponding, global blocker-sensitive I_Bas. The "threshold" of activation of fast-decaying I_Rs was more negative by about 10 mV than that of blocker-sensitive I_Bas. P_Ba(V) plots for fast-decaying I_Rs were also derived, and Botzmann fittings to data points returned an average V_1/2 of 9.0 ± 1.7 mV, with k = 8.4 ± 0.5 mV. The average V_1/2 value obtained for fast-decaying I_Rs was significantly lower as compared with those measured in L-, N-, and P-type current components (see above; P < 0.05 in all cases).

View larger version (28K): [in this window] [in a new window]   Fig. 9. Fast-decaying R-type currents (IRfi) and their voltage dependence. A and B: currents recorded in 2 representative neurons (A, cell C9D09; B, cell A9D30) either under control conditions (column 1) or during the application of 10 µM nifedipine + 0.5 µM -CTx GVIA + 1 µM -CTx MVIIC (column 2). The voltage protocol applied was the same as illustrated in Fig. 3A1. Calibration bars: 200 (A) or 100 (B) pA, 50 ms. Column 3 shows control and blocker-resistant (asterisks) currents recorded at the peak of the I-V relationships, normalized to their peak amplitudes and superimposed. C: average, normalized IRfis and global blocker-sensitive currents. IRfis and the corresponding global blocker-sensitive currents from 5 cells were normalized to the amplitude value observed at the peak of the I-V relationships and averaged. Average IRfis (enhanced lines) and blocker-sensitive currents (thinner lines) recorded at 3 different test potentials are shown superimposed. D: average, normalized I-V relationships of IRfi () and the corresponding global blocker-sensitive currents (; n = 5). The I-Vs derived from individual cells were normalized to their peak values and averaged among cells, and the average I-Vs thus obtained were further normalized to their peak values. The dotted line is the average I-V relationship of the IBaT found in EPN neurons (same as in Fig. 2A3), shown here for comparison.

The inactivation kinetics of fast-decaying I_Rs were further characterized by performing exponential fittings of current decay phase. Biexponential inactivation kinetics were consistently found (Fig. 10A), with a fast time constant ranging from ~100 to ~20 ms, and a slower time constant ranging from ~230 to ~115 ms, depending on the voltage level (Fig. 10B1).

View larger version (26K): [in this window] [in a new window]   Fig. 10. Inactivation kinetics of IRfi. A: currents recorded in a representative neuron (cell D9D09) during the application of 10 µM nifedipine + 0.5 µM -CTx GVIA + 1 µM -CTx MVIIC, at the test potentials of 20 and 10 mV. Enhanced lines are biexponential fittings to current decay phase, with the following fitting parameters: 1 = 33.8 ms, A1 = 53.0 pA, 2 = 214.3 ms, A2 = 89.7 pA, C = 0 pA (20 mV); 1 = 26.1 ms, A1 = 130.8 pA, 2 = 200.0 ms, A2 = 142.5 pA, C = 5.5 pA (10 mV). Calibration bars: 50 pA, 50 ms. B: average fitting parameters returned by biexponential fittings of IRfi decay phase (n = 5). B1: fast () and slow () time constants (note the semilogarhythmic scale) as a function of test potential. B2: relative amplitude coefficient of the fast exponential component as a function of test potential. C: near-threshold kinetic behavior of IRfi (C1) and IBaT (C2) recorded in the same neuron (cell C9618). Test-potential levels are indicated close to each trace. Calibration bars: 20 pA, 50 ms. The thick, horizontal bars remark the time span over which data points were averaged to obtain a measurement of current amplitude at the end of the 300-ms test pulses, for the determination of R300/p (see next panel). D: average plots of R300/p as a function of test potential for both IRfi (; n = 5) and IBaT (; n = 8).

Due to its insensitivity to L-, N-, and P/Q-type channel blockers, fast decay kinetics, and relatively more negative threshold of activation, the R-type current here described closely resembles a current expressed by PC layer-II pyramidal neurons, for which we proposed the name of I_Rfi (R-type, fast-decaying, intermediate-threshold) (see Magistretti et al. 2000). The same name, therefore will be applied here to fast-decaying I_Rs of EPN neurons.

The above pharmacological, kinetic, and voltage-dependent properties that distinguish I_Rfi from other HVA currents could be somewhat reminiscent of those typical of LVA, T-type Ca²⁺ currents. To assess the possible relationship of I_Rfi with typical T-type currents, we directly compared the biophysical properties of the I_Rfi found in EPN neurons with those of the I_BaT expressed by the same cell population. The two currents turned out to be functionally distinct in various respects. First of all, the voltage range of activation of I_Rfi was shifted by about 20 mV in the positive direction as compared with that of I_BaT. In particular, the activation "threshold" of I_Rfi was at approximately 40 mV versus a level negative to 50 mV for I_BaT, and the peak of the I-V relationship was at about 10 to 0 mV for I_Rfi versus 20 mV for I_BaT (see Fig. 9D). Second, in spite of its fast-decaying nature close to threshold, I_Rfi behaved as a relatively sustained current (Fig. 10C1), with average R_300/p values of 0.68 and 0.44 at 40 and 30 mV, respectively (Fig. 10D). On the contrary, I_BaT inactivated almost completely at the end of the routinely applied 300-ms depolarizing test pulses even close to its threshold (Fig. 10C2), with average R_300/p values of 0.13 and 0.07 at 50 and 40 mV, respectively (Fig. 10D). Finally, activation speed was remarkably higher in I_Rfi than in I_BaT. For a quantitative analysis of current activation kinetics, exponential fittings of current activation phases were performed. To improve the reliability of activation-kinetics data, both activation and inactivation phases were simultaneously fitted with triple- (I_Rfi) or double- (I_BaT) exponential functions, the fastest of which described the activation process. Figure 11, A-C, illustrates the results obtained for both I_Rfi and I_BaT recorded in a single, representative neuron. Activation time constants (_as) were remarkably faster in I_Rfi than in I_BaT, especially close to current threshold. I_{BaT a}s were markedly voltage dependent and decreased monotonically from 50 to 10 mV, where they averaged ~6.8 ms and ~1.1 ms, respectively (Fig. 11D). On the contrary, _a values derived for I_Rfi consistently displayed a bell-shaped voltage dependence, and ranged from ~1.14 ms (at 20/10 mV) to ~550 µs (at +30 mV) on average. Figure 11C highlights the difference in I_Rfi and I_BaT activation speeds in two normalized current traces recorded at approximately the same voltage level relative to current threshold. It can be concluded that the activation and inactivation properties of the I_Rfi expressed by EPN neurons are not consistent with those, typical of T-type currents, of the I_T present in the same cells.

View larger version (34K): [in this window] [in a new window]   Fig. 11. Activation kinetics of IRfi. A and B: detail of the activation phase of IRfi (A) and IBaT (B) recorded in the same neuron (cell C9618) at various test potentials (specified close to each trace). Calibration bars: 40 (A) or 50 (B) pA, 15 ms. Enhanced lines are triple- (A) or double- (B) exponential fittings. The values of the fast (activation) time constant (a) are indicated nearby. The inset in A highlights, over a further expanded time scale, the activation phase of IRfi recorded at 10 mV (calibration bar: 1 ms). C: IRfi (thinner line) and IBaT (enhanced line) recorded at 0 and 20 mV, respectively, are shown normalized to their peak amplitudes and superimposed, to highlight the difference in activation speed. D: average plots of a as a function of test potential for both IRfi (; n = 5) and IBaT (; n = 11). The inset is a detail, over an expanded y scale, of the plot relevant to IRfi.

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

The present study provides the first biophysical and pharmacological analysis of voltage-gated Ca²⁺ currents in EPN neurons. Similarly to most neuronal populations of the CNS so far studied, these cells were found to express, in their somatic/proximal dendritic compartment, several types of functionally and pharmacologically different Ca²⁺ currents, including both LVA and multiple HVA currents. The pattern of Ca²⁺-current expression in EPN neurons showed similarities but also remarkable differences as compared with pyramidal neurons of neighboring PC layer II (Magistretti et al. 1999), which is in line with the specific roles recognized or proposed for the EPN in both physiological and pathological processes taking place within the piriform region.

We confirmed our previous finding that a LVA, T-type Ca²⁺ current (I_T) is robustly expressed by practically all EPN neurons (Magistretti and de Curtis 1998). We found that this current was virtually insensitive to moderate concentrations of Ni²⁺. T-type Ca²⁺ channels have been considered to be highly and specifically sensitive to Ni²⁺ block in several neuronal populations (Allen et al. 1993; Carbone and Swandulla 1989; Crunelli et al. 1989; Fox et al. 1987; Lucaj and Fujii 1994; Ozawa et al. 1989). Three different ₁ subunits accounting for T-type channels (_1G, _1H, _1I) have been isolated so far (Lee et al. 1999a; Perez-Reyes 1998). Each of them displays a remarkably specific pattern of Ni²⁺-blockade potency (Lee et al. 1999b), with the highest potency for _1H (K_D = 5.7-12 µM), and the lowest for _1G (K_D = 167-250 µM). Our finding that the T-type conductance present in EPN neurons is insensitive to 50 µM Ni²⁺ is in complete agreement with in situ hybridization data showing that, of the three subunits forming T-type channels, basically only _1G is intensely expressed in the EPN, whereas _1H is virtually absent (Talley et al. 1999).

HVA currents were found to display considerable biophysical diversity among individual EPN neurons. Similarly to what has been already reported from PC layer-II neurons (Magistretti et al. 1999), the kinetic variability of total HVA I_Bas was paralleled by differences in voltage dependence: on average, the voltage range of activation of fast-decaying currents was significantly shifted in the negative direction as compared with that of slower-decaying currents. The above differences may not be surprising in light of the fact that EPN neurons constitute a heterogeneous population (Haberly 1983, 1998; Haberly et al. 1987; Hoffman and Haberly; Kubota and Jones 1993). The functional diversity of total HVA I_Bas expressed by EPN neurons turned out to be underlain by differential expression of multiple pharmacological current components characterized by different kinetic and voltage-dependent properties. Nifedipine-sensitive, L-type currents were consistently slowly decaying, and activated within a relatively positive voltage range. By contrast, R-type currents showed consistently faster decay than the total I_Bas, and most of them (named I_Rfi on the analogy of a similar current present in PC layer-II neurons) (Magistretti et al. 2000) displayed very fast inactivation kinetics as well as activation threshold within an "intermediate" range of membrane voltages. -CTx GVIA-sensitive, N-type currents showed higher variability with respect to their inactivation kinetics, but not to voltage dependence of activation. Finally, the majority of -CTx MVIIC-sensitive currents were slowly decaying and high-threshold activated, similarly to typical P-type currents present in other CNS neuronal populations (e.g., Lorenzon and Foehring 1995; Mintz et al. 1992; Randall and Tsien 1995). However, a significant proportion of -CTx MVIIC-sensitive currents displayed considerably fast decay kinetics, which has been considered to be a distinctive feature of Q-type currents in other cell systems (Hilaire et al. 1996; Randall and Tsien 1995; Teramoto et al. 1995; Wang et al. 1997). Moreover, the threshold of activation of this presumptive Q-type current was found to be clearly lower than that of slowly decaying -CTx MVIIC-sensitive currents. Although differences in voltage dependence of activation between P- and Q-type currents have not been previously reported in other native neuronal preparations, it should be noted that possible molecular correlate(s) of our finding have been actually described in heterologous-expression systems. Indeed, co-expression of different  subunits with the P/Q-type-channel main subunit, the _1A subunit, results in concomitant modulation of both current kinetics and voltage dependence. For instance, _1A-₃ channels have been found to generate currents that are both faster decaying and lower-threshold activated than those produced by _1A-₄ channels (De Waard and Campbell 1995; Stea et al. 1994).

An R-type current endowed with interesting functional properties is the above mentioned I_Rfi (Magistretti et al. 2000), expressed by in most, although not all EPN neurons. Its resistance to dihydropyridines and Conus venom toxins, as well as its peculiar biophysical features, including fast decay kinetics and relatively low threshold of activation, raised the question of its possible relationship with LVA, T-type currents. The expression, within the same EPN neuronal population, of both I_Rfi and I_T, enabled us to carry out a direct comparison between the two currents. The relatively slow activation kinetics and the full decay within <300 ms, over the whole voltage range of activation, observed in EPN I_T (and expected for a classical T-type current), were in clear contrast with the kinetic behavior of I_Rfi. Indeed, this current consistently displayed fast activation speed and incomplete inactivation during 300-ms depolarizations close to threshold. On these bases, it can be concluded that EPN I_Rfi is clearly different from typical I_Ts. The fact that fast-decaying I_Rfi is actually a relatively sustained current at near-threshold voltage levels in not incompatible with a homogeneous channel population (discussed in Magistretti et al. 2000). This feature is actually shared by the currents resulting from heterologous expression of the _1E subunit (i.e., Parent et al. 1997; Schneider et al. 1994; Williams et al. 1994), which is the most likely molecular correlate of I_Rfi (discussed in Magistretti et al. 2000). In PC layer II, indeed, the presence of prominent R-type currents, and in particular I_Rfi (Magistretti et al. 2000), is paralleled by high levels of _1E-subunit mRNA (Soong et al. 1993; Wakamori et al. 1994; Williams et al. 1994).

Depolarizing events dependent on voltage-gated Ca²⁺ conductances have been shown to significantly influence the intrinsic electroresponsiveness of EPN neurons. Current-clamp recordings obtained from multipolar EPN cells and deep PC neurons in rat brain slices revealed the existence of different intrinsic, regenerative potentials insensitive to TTX and blocked by Co²⁺ (Tseng and Haberly 1989b), including a low-threshold Ca²⁺ spike, and a more sustained depolarizing potential evoked at a threshold of 50 to 35 mV. Although no detailed pharmacological characterization of such membrane potential events has been carried out yet, the two types of Ca²⁺-dependent potentials were ascribed to the existence of both a LVA current and HVA Ca²⁺ current(s) activated at relatively negative potentials, respectively. These current-clamp data are in agreement with our demonstration of the co-existence, in EPN neurons, of a typical I_T and non-T-type currents (including I_Rfi and the "Q-type" current) elicited within an "intermediate" range of membrane voltages (starting at less than or equal to 40 mV). The observation that fast-decaying I_Rfi inactivates more slowly, close to its threshold, as compared with I_T is also compatible with the duration (up to a few hundreds of ms) of the sustained, intermediate-threshold Ca²⁺ potential described in slices (Tseng and Haberly 1989b). The presence of regenerative Ca²⁺ spikes has been postulated to favor a sustained firing in EPN neurons, which might promote prolonged synaptic depolarizing potentials in target cells within the EPN and the PC (Tseng and Haberly 1989a). Such activity is thought to represent a synchronizing element in the piriform region and has been hypothesized to be involved in epileptogenesis (de Curtis et al. 1999; Demir et al. 1999; Forti et al. 1997; Tseng and Haberly 1989a; de Curtis, unpublished observations). To confirm this view, specific pharmacological manipulations during current-clamp recordings of interictal spike discharge will be necessary.