HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS		QUICK SEARCH:   [advanced] Author: Keyword(s): Year:  Vol:  Page:

This Article Abstract Full Text (PDF) All Versions of this Article: 97/6/3986    most recent 00141.2007v1 Alert me when this article is cited Alert me if a correction is posted Citation Map Services Email this article to a friend Similar articles in this journal Similar articles in ISI Web of Science Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Citing Articles Citing Articles via HighWire Citing Articles via ISI Web of Science (2) Citing Articles via Google Scholar Google Scholar Articles by Gittis, A. H. Articles by du Lac, S. Search for Related Content PubMed PubMed Citation Articles by Gittis, A. H. Articles by du Lac, S.

Firing Properties of GABAergic Versus Non-GABAergic Vestibular Nucleus Neurons Conferred by a Differential Balance of Potassium Currents

University of California, San Diego Graduate Program in Neuroscience, The Salk Institute for Biological Studies, Howard Hughes Medical Institute, La Jolla, California

Submitted 7 February 2007; accepted in final form 21 March 2007

	ABSTRACT

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION GRANTS ACKNOWLEDGMENTS REFERENCES

 
Neural circuits are composed of diverse cell types, the firing properties of which reflect their intrinsic ionic currents. GABAergic and non-GABAergic neurons in the medial vestibular nuclei, identified in GIN and YFP-16 lines of transgenic mice, respectively, exhibit different firing properties in brain slices. The intrinsic ionic currents of these cell types were investigated in acutely dissociated neurons from 3- to 4-wk-old mice, where differences in spontaneous firing and action potential parameters observed in slice preparations are preserved. Both GIN and YFP-16 neurons express a combination of four major outward currents: Ca²⁺-dependent K⁺ currents (I_KCa), 1 mM TEA-sensitive delayed rectifier K⁺ currents (I_1TEA), 10 mM TEA-sensitive delayed rectifier K⁺ currents (I_10TEA), and A-type K⁺ currents (I_A). The balance of these currents varied across cells, with GIN neurons tending to express proportionately more I_KCa and I_A, and YFP-16 neurons tending to express proportionately more I_1TEA and I_10TEA. Correlations in charge densities suggested that several currents were coregulated. Variations in the kinetics and density of I_1TEA could account for differences in repolarization rates observed both within and between cell types. These data indicate that diversity in the firing properties of GABAergic and non-GABAergic vestibular nucleus neurons arises from graded differences in the balance and kinetics of ionic currents.

	INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION GRANTS ACKNOWLEDGMENTS REFERENCES

 
Diverse classes of neurons have evolved to perform specific functions in complex circuits, requiring specialization of their ability to process inputs into meaningful patterns of firing. The ability of a neuron to processes and transmit information depends on its intrinsic ionic currents. An emerging question is how these currents are regulated to produce the appropriate pattern of output. Recent studies in invertebrates have shown that the level of channel expression can vary considerably within the same cell from animal to animal, but the output pattern is kept constant through correlated channel expression that maintains a target balance of currents that are unique to a particular cell type (Prinz et al. 2004; Schulz et al. 2006).

Spontaneously firing neurons in the medial vestibular nuclei (MVN) respond linearly over a wide dynamic range and are capable of sustaining firing rates of hundreds of spikes/second (Sekirnjak and du Lac 2002, 2006; Sekirnjak et al. 2003; Smith et al. 2002). Action potential and firing properties form a continuum across MVN neurons (du Lac et al. 1995; Sekirnjak and du Lac 2002; Straka et al. 2005). Initial studies subdivided the continuum into two broad types defined by canonical properties of action potentials at the extremes (Johnston et al. 1994; Serafin et al. 1991). Subsequent studies combining electrophysiological recordings with anatomical (Sekirnjak and du Lac 2006; Sekirnjak et al. 2003) or molecular (Takazawa et al. 2004) analyses revealed a diversity of cell types with graded differences in firing properties. Experience-dependent changes in intrinsic excitability of MVN neurons can be evoked by synaptic inhibition (Nelson et al. 2003) or by unilateral labyrinthectomy, the vestibular equivalent of monocular deprivation (Beraneck et al. 2003, 2004; Cameron and Dutia 1997; Guilding and Dutia 2005; Him and Dutia 2001). Progress in dissecting the mechanisms and functional consequences of such intrinsic plasticity, however, has been hampered by a lack of knowledge about the ionic currents expressed in specific cell types.

Recently, two lines of transgenic mice have been identified that label different classes of MVN neurons: GIN mice (Oliva et al. 2000) express GFP in GABAergic neurons, and YFP-16 mice (Feng et al. 2000) express YFP in non-GABAergic, glutamatergic, and glycinergic neurons (Bagnall et al. 2007). YFP-16 neurons have narrower action potentials and can sustain higher firing rates than GIN neurons. These differences in firing properties could be achieved via a number of alternative mechanisms, including expression of distinct ionic currents, as observed in regular spiking pyramidal cells versus fast-spiking interneurons in the cortex (Martina et al. 1998), differences in dendritic morphology (Mainen and Sejnowski 1996), or variations in the ratio of current expression, as in somatogastric ganglion neurons (Schulz et al. 2006).

To investigate mechanisms that underlie differences in firing properties between GIN and YFP-16 neurons, somatic whole cell currents were measured in an acutely dissociated cell preparation that preserves both spontaneous firing and differences in action potential properties between the two cell classes. The results suggest that graded differences in the balance of ionic currents underlie the continuous variations in firing properties of GABAergic and non-GABAergic vestibular nucleus neurons.

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION GRANTS ACKNOWLEDGMENTS REFERENCES

 
Cell preparation

Coronal slices (350–400 µM) through the rostral 2/3 of the MVN were prepared as described in Sekirnjak et al. (2003) from 24- to 39-day-old mice (average = 29 ± 4 days; mean ± SD), c57bl6 wild-type, GIN (Oliva et al. 2000), or YFP-16 (Feng et al. 2000) lines of mice both in c57bl6 backgrounds. Neurons were enzymatically dissociated at 30°C for 10 min in a solution of 9.4 mg/ml MEM powder (Gibco), 10 mM HEPES, 0.2 mM cysteine, and 40 U/ml papain (Worthington), pH 7.2. The vestibular nuclei were dissected out in a similar ice-cold solution in which papain was replaced by 1 µg/ml BSA and 1 µg/ml Trypsin inhibitor. The nuclei were triturated with fire polished Pasteur pipettes of decreasing diameter in 500 µl Tyrode's solution (see Electrophysiological recording) and plated on the glass slide of the recording chamber. The cells were allowed to settle for 10 min, then were continuously perfused with oxygenated Tyrode's solution for the duration of the recording (2–3 h).

Electrophysiological recording

Whole cell patch recordings were made at room temperature under continuous perfusion with oxygenated Tyrode's solution (in mM: 150 NaCl, 3.5 KCa, 2 CaCl₂, 1 MgCl₂, 10 HEPES, 10 glucose). Borosilicate pipettes (2–4 M) were filled with a KMeSO₄-based intracellular solution (in mM: 140 KMeSO₄, 8 NaCl, 10 HEPES, 0.02 EGTA, 2 Mg₂-ATP, 0.3 Na₂-GTP, and 14 Tris-creatine PO₄). The measured liquid junction potential was +15 mV and was corrected off-line. Data were collected and analyzed using IGOR software with a MultiClamp 700B amplifier (Axon Instruments) and an ITC-16 interface (Instrutech).

Action potentials recorded in current-clamp mode were filtered at 10 kHz and digitized at 40 kHz. Action potential width, rate of repolarization, afterhyperpolarization (AHP), and afterdepolarization (ADP) were calculated from the average action potential over a 5-s window during which the cell was made to fire at 5 ± 2 spike/s with DC current injection. For experiments in which the action potential was measured in different drug conditions, the firing rate of the neuron was maintained at 5 spike/s by adjusting the level of DC current injection as needed. Cells included for analysis had action potential heights >50 mV (average = 71.8 ± 7.9 mV) and could fire spontaneously. Action potential threshold was defined as the voltage at which the rate of change exceeded 10 V/s. Action potential height was calculated as the change in voltage from threshold to the peak of the action potential. Action potential width was measured half way between action potential threshold and peak.

The rate of repolarization was measured as the greatest rate of change (minimum derivative V/s) during the falling phase of the action potential. The amplitude of the AHP was measured as the peak drop in membrane voltage (V_m) below action potential threshold. The ADP was calculated as the maximum derivative of Vm within 3 ms of action potential repolarization below threshold.

After action potentials were collected in current clamp, the amplifier was switched into voltage-clamp mode. Recordings of whole cell currents were made in voltage-clamp mode with a 6 kHz filter and digitized at 20 kHz. Whole cell capacitance was compensated through the amplifier and series resistance (R_series) was compensated at 70%. The average series resistance read off the dial was 9 ± 3 M, and cells were excluded if they had a series resistance >20 M. The capacitance was measured by integrating the area of the transient following a step from –65 to –95 mV with whole cell capacitance and series resistance compensation turned off.

To evaluate stability in currents during the recording, the waveform of each component current was added together and compared with the outward current measured in TTX at the beginning of the experiment at nominal +15 mV. The error between the summed wave and measured wave was calculated by dividing the integral of the summed wave by the integral of the measured wave. The average error was 2%, and cells with >3% error were excluded.

Corrections for voltage errors

MVN neurons had large whole cell outward currents, often reaching 10 nA or more in response to a +15-mV command potential. The actual voltage experienced by the cell deviated from the command voltage of the amplifier by the product of the amplitude of the evoked current and the uncompensated series resistance, which averaged 2 ± 1 M. In response to the highest nominal voltage command used in this study (+15 mV), the average evoked current was 10 ± 3 nA, so the actual voltage used to evoke I_total and I_Kca deviated from the command voltage by 20 ± 6 mV. As drugs were applied and currents got smaller, this voltage error got smaller; the voltage error for I_1TEA was 11 ± 6 mV and for I_10TEA and I_A was <5 mV. I_total and R_series did not differ significantly between GIN and YFP-16 neurons, enabling comparison of the balance of currents in response to the same nominal (+15 mV) voltage step.

Boltzmann fits for I_total, I_KCa, and I_1TEA were corrected for errors in voltage due to uncompensated R_series, resulting in shifts of 3 mV on average in V_1/2 and 3 mV on average in the slope. The remaining currents, I_A and I_10TEA, were small enough that the voltage error was typically <5 mV different from the command voltage, and therefore no corrections were made to their Boltzmann fits.

Pharmacology

GIN and YFP-16 neurons were targeted for recording using fluorescence. After formation of a gigohm seal, the cell was lifted off the bottom of the recording chamber and positioned directly in front of a small piece of tubing through which pharmacological solutions were delivered to isolate ionic current components of the TTX-insensitive current in the cell (I_total). Solutions were rapidly exchanged using a gravity-driven, VC-6 perfusion valve control system (Warner) and were applied in the following order: 1) Tyrode's, 2) Tyrode's +300 nM TTX, 3) 0 Ca²⁺ Tyrodes (Tyrode's in which 2 mM CaCl₂ was replaced with 1.7 mM MgCl₂ and 0.3 mM CdCl₂) + 300 nM TTX, 4) 0 Ca²⁺ Tyrode's +300 nM TTX +1 mM TEA, 5) 0 Ca²⁺ Tyrode's +300 nM TTX +10 mM TEA; 6) In some neurons, a sixth solution was applied containing 0 Ca²⁺ Tyrode's + 300 nM TTX +10 mM TEA +5 mM 4-AP.

The transient Na current (I_NaT) was measured by subtracting the currents between solutions 1 and 2. The Ca²⁺-dependent K⁺ current (I_KCa) was measured as the difference current between solutions 2 and 3. The 1 mM TEA-sensitive current was measured as the difference current between solutions 3 and 4, and the 10 mM TEA-sensitive current was measured as the difference current between solutions 4 and 5. I_A was insensitive to 10 mM TEA and was isolated as the difference current inactivated by a predepolarizing step to –45 mV compared with a prehyperpolarizing step to –75 mV. This divided the 10-mM TEA-insensitive current into I_A and a small current that could not be resolved further, termed I_other that made up <5% I_total. The A current isolated in this manner was similar in amplitude and kinetics to the 4-AP-sensitive current.

In some cells, I_KCa was further divided into I_BK and I_SK. In these cells, the following solutions were applied: 1) Tyrode's, 2) Tyrode's +300 nM TTX, 3/4) Tyrode's +120 nM iberiotoxin (IBTX), 3/4) Tyrode's +100 nM apamin, 5) 0 Ca²⁺ Tyrode's +300 nM TTX.

IBTX was slower to block current compared with other drugs used in this study, so IBTX was applied to the cell for 2 min while looping the voltage protocol three times. The current remaining during the last voltage protocol was subtracted from I_total measured in TTX to calculate I_BK. A similar protocol was used to measure I_SK. In some cells, both IBTX and apamin were applied. In all of these cells, 0.3 mM CdCl₂ blocked outward current that had not been previously blocked by IBTX or apamin.

The solutions for measuring Ca²⁺ currents were adapted from Swensen and Bean (2005). The solution consisted of (in mM) 50 NaCl, 3.5 KCl, 2 CaCl₂, 1 MgCl₂, 10 HEPES, 100 TEA-Cl, 300 nM TTX, 10 glucose +5 4-AP, and 100 nM apamin. Ca²⁺ currents were isolated by subtraction following application of a similar solution in which 2 mM CaCl₂ was replaced by 2 mM MgCl₂.

TTX and IBTX were purchased from Tocris. TEA, CdCl₂, 4-AP, and apamin were from Sigma. Stock solutions were diluted in water and stored at 4°C except 4-AP, IBTX, and apamin, which were stored at –20°C.

Calculations and statistics

To control for the different soma sizes of MVN neurons, current magnitudes were compared across cells in terms of current density. Current density was calculated by dividing the current amplitude (pA) by the cell capacitance (pF).

Because the data were not normally distributed, statistical differences were tested with the nonparametric Wilcoxon test for unpaired data with the exception of changes in action potential shape following CdCl₂ or 1 mM TEA application where a Wilcoxon test for paired data was used. The strength of a correlation was measured with the Pearson correlation (r) and was tested for significance against the critical values on a two-tailed test. Errors reported in text are SDs.

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION GRANTS ACKNOWLEDGMENTS REFERENCES

 
Firing properties of dissociated MVN neurons

Acutely dissociated MVN neurons were isolated from mice, age 24–39 days old (average = 29 ± 4 days). At this age, the intrinsic firing dynamics of MVN neurons are mature (Dutia et al. 1995; Johnston and Dutia 1996; Murphy and du Lac 2001). Soma sizes ranged from 15 to 30 µm along the long axis, with short, proximal processes (<30 µm; Fig. 1A). The neurons had an average input resistance (R_input) of 1662 ± 602 M and capacitance of 7.9 ± 2.3 pF (n = 129).

View larger version (18K): [in this window] [in a new window]   FIG. 1. Acutely dissociated medial vestibular nucleus (MVN) neurons fire spontaneously and exhibit differences in action potential parameters between GIN and YFP-16 neurons. A: DIC images of dissociated MVN neurons. Typically, processes did not extend beyond 30 µm, and soma sizes ranged from 15 to 30 µm in diameter. B: example of the spontaneous, tonic firing pattern recorded from a dissociated MVN neuron. C: histogram of the spontaneous firing rates across the population of dissociated MVN neurons (average = 12.7 ± 6.8 spike/s, n = 126). D: representative action potentials recorded from GIN (left) and YFP-16 neurons (right). dashed line, –55 mV. E: peak afterhyperpolarization (AHP) amplitude is plotted vs. action potential width (defined at half-height, see METHODS) for GIN neurons (), YFP-16 neurons (), and a population of unidentified MVN neurons (). GIN neurons tended to have wider action potentials and larger AHPs than YFP-16 neurons. Action potential parameters from GIN and YFP-16 neurons overlap with those from unidentified neurons, forming a continuum.

To specifically target different cell types in this study, recordings were made from fluorescently labeled neurons from GIN (Oliva et al. 2000) and YFP-16 lines of transgenic mice (Feng et al. 2000), which in the MVN label GABAergic and non-GABAergic neurons, respectively (Bagnall et al. 2007) YFP-16 and GIN neurons fired spontaneous action potentials, with YFP-16 neurons exhibiting somewhat higher firing rates than GIN neurons (Table 1). R_input, capacitance, and R_series did not differ significantly between GIN and YFP-16 neurons, indicating that cell size and recording quality were equivalent (Table 1).

Inward and outward whole cell currents

The preservation of the intrinsic differences in action potentials between GIN and YFP-16 neurons in the dissociated preparation implies differences in the underlying somatic currents. To identify these differences, whole cell somatic currents were elicited from dissociated neurons with 150-ms voltage steps from –55 to +15 mV from a holding potential of –65 mV, and pharmacology was used to isolate multiple currents within each neuron.

The transient Na current (I_NaT) was defined as the large, fast inward current isolated by subtraction after application of 300 nM TTX (Fig. 2A). The voltage at which I_NaT reached its peak varied across cells but tended to occur between –45 and –35 mV and was not significantly different between GIN (–36 ± 8.9 mV, n = 35) and YFP-16 neurons (–36 ± 8.8 mV, n = 39). Although I_NaT density, measured at –35 mV, tended to be larger in YFP-16 neurons, this difference was not significant, P = 0.10 (Fig. 2B).

View larger version (21K): [in this window] [in a new window]   FIG. 2. Whole cell inward and outward currents are similar in GIN and YFP-16 neurons. A: family of TTX-sensitive transient Na+ currents (INaT) recorded from a GIN neuron. The voltage protocol used to evoke the currents is illustrated: neurons were held at –65 mV and stepped for 150 ms to voltages between –55 and +15 mV in 10 mV increments. B: plot of the peak current density of INaT in GIN () and YFP-16 neurons () at each voltage. There was no difference in INaT density between GIN (956 ± 313 pA/pF; n = 35) and YFP-16 neurons (1,085 ± 353 pA/pF; n = 40) at –35 mV (P = 0.10). C: TTX-insensitive current (Itotal) recorded from the same neuron as in A. D: plot of the current density of Itotal in GIN () and YFP-16 neurons () at each voltage, measured as the average steady-state current density over the last 50 ms of the trace. The density of Itotal did not differ significantly between GIN (1,239 ± 524 pA/pF) and YFP-16 (1,423 ± 532 pA/pF) neurons at +15 mV, P = 0.29. Error bars represent SE.

I_total had a similar time course between GIN and YFP-16 neurons and a similar rate of activation, calculated as the peak derivative over the rising phase of the current (GIN = 6,090 ± 2,553 pA/ms, n = 19; YFP-16 = 6,640 ± 2,728 pA/ms, n = 20). The density of I_total varied by >2.5-fold across cells but did not differ significantly between GIN and YFP-16 neurons, P = 0.29 (Fig. 2D). The voltage dependence of I_total was measured by its voltage of half-maximal activation (V_1/2) and the steepness of its voltage dependence (k), measured by fitting the normalized conductance graph with a Boltzmann fit. The V_1/2 and slope (k) values of I_total in GIN and YFP-16 neurons were similar (Table 2), suggesting that differences in firing properties between the two cell types arise from differences either in specific current subtypes or in the balance of currents.

The Ca²⁺-dependent K⁺ current (I_KCa) was measured by subtraction after replacing extracellular Ca²⁺ with a mixture of Mg²⁺ (1.7 mM) and Cd²⁺ (0.3 mM). I_KCa varied in time course and magnitude across the population of recorded neurons. In most neurons, I_KCa had a prominent transient component that decayed within the first 10 ms, revealing a steady-state sustained component (Fig. 3A₁). The rate of activation of I_KCa, described as the maximum derivative during the rising phase of the current was not different between the cell types [3,541 ± 1,825 pA/ms (GIN) vs. 3,030 ± 1,240 pA/ms (YFP-16), P = 0.38]. Although both components of I_KCa tended to be larger in GIN neurons, the current density was not significantly different between cell types (Fig. 3, B and C).

View larger version (20K): [in this window] [in a new window]   FIG. 3. Ca2+-dependent K+ currents (IKCa) in GIN and YFP-16 neurons. (A, 1–3) Family of Cd-sensitive IKCa, divided into IBTX-sensitive IBK (A2) and apamin-sensitive ISK (A3) components from the same YFP-16 neuron. B: current density of the transient component (peak amplitude within the 1st 10 ms) of IKCa at each voltage in GIN () and YFP-16 neurons (). IKCa density of the transient component was not different between GIN (678 ± 351 pA/pF; n = 19) and YFP-16 neurons (585 ± 243 pA/pF; n = 20) at +15 mV, P = 0.34. C: current density of the sustained component (average over the last 50 ms) of IKCa at each voltage in GIN () and YFP-16 neurons (). IKCa density of the sustained component was not different between GIN (616 ± 404 pA/pF) and YFP-16 neurons (455 ± 221 pA/pF) at +15 mV, P = 0.42. D: bar graph showing fraction of IKCa that was sensitive to IBTX and apamin in YFP-16 (0.62 ± 15; 0.20 ± 23, ) and GIN neurons (0.63 ± 18; 0.11 ± 14, ) at +15 mV, calculated as the average current density over the entire 150 ms step. Error bars represent SE.

The relative contribution of I_BK and I_SK to I_KCa varied considerably across neurons but did not differ between GIN and YFP-16 neurons. In GIN neurons, 63 ± 18% of I_KCa was sensitive to IBTX and 11 ± 14% was sensitive to apamin; in YFP-16 neurons, 62 ± 15% of I_KCa was sensitive to IBTX and 20 ± 23% was sensitive to apamin (Fig. 3D). Although more than half of I_KCa in MVN neurons tended to be IBTX sensitive, this proportion ranged from 38 to 85% in GIN neurons and from 34 to 73% in YFP-16 neurons. This high variability could reflect differences in channel expression or in the relative insensitivity to IBTX conferred by some BK channel -subunits (Brenner et al. 2005; Meera et al. 2000). The remaining Ca²⁺-sensitive K⁺ current could reflect a combination of BK and SK insensitive to IBTX and apamin (Brenner et al. 2005; Coetzee et al. 1999; Meera et al. 2000) but suggests there is likely to be a third type of I_KCa in MVN neurons as described in other cell types (Joiner et al. 1998; Limon et al. 2005; Sah and Faber 2002; Vergara et al. 1998). Taken together, these results suggest that BK is the dominant somatic current that contributes to I_KCa in MVN neurons.

Boltzmann fits revealed differences in the gating properties of I_KCa current between GIN and YFP-16 neurons. The V_1/2 was more hyperpolarized in YFP-16 neurons than GIN neurons (P = 0.006) and showed a steeper voltage dependence compared with GIN neurons (P = 0.0002; Table 2). This could reflect differences in the channel subunits contributing to I_KCa or differences in Ca²⁺ currents. Because the Ca²⁺ concentration influences the probability of opening of BK and SK channels, the voltage-dependent properties of I_KCa should be related to the voltage dependence of I_Ca. The Ca²⁺ current reached its peak voltage at –15 mV in 3/5 YFP-16 neurons and at –5 mV in 7/9 GIN neurons, consistent with the lower V_1/2 of YFP-16 neurons compared with that of GIN neurons.

TEA K⁺ currents

A subset of K⁺ currents can be identified based on their high sensitivity to tetraethylammonium (TEA, 1 mM), including BK, Kv1, and Kv3 currents (Coetzee et al. 1999). In dissociated MVN neurons, in the presence of CdCl₂ (which blocks I_BK), 1 mM TEA blocked a noninactivating current with a depolarized V_1/2 that activated between –25 and –15 mV (Fig. 4B, Table 2). In 14/14 cells, this current was insensitive to 50 nM dendrotoxin, a specific blocker of noninactivating, Kv1-containing channels (Gamkrelidze et al. 1998; Grissmer et al. 1994; Khavandgar et al. 2005). Based on its voltage dependence and pharmacology, the 1 mM TEA-sensitive current in MVN neurons likely represents current through Kv3-containing channels (Coetzee et al. 1999).

View larger version (18K): [in this window] [in a new window]   FIG. 4. TEA-sensitive K+ currents in GIN and YFP-16 neurons. A: family of 1 mM TEA-sensitive I1TEA with typical kinetics recorded from a YFP-16 neuron. B: plot of the current density of I1TEA in GIN (open circles) and YFP-16 neurons (filled triangles) at each voltage, measured as the average steady-state current density over the last 50 ms of the trace. There was more I1TEA in YFP-16 (634 ± 361 pA/pF; n = 20) than in GIN neurons (406 ± 186 pA/pF; n = 19) at +15 mV, P = 0.05. C: average current waveforms at +15 mV from 19 GIN neurons (gray) and 20 YFP-16 neurons (black). I1TEA from GIN neurons was scaled by a factor of 1.6 to match the steady-state value of I1TEA from YFP-16 neurons. Note the faster kinetics of I1TEA in YFP-16 neurons. D: quantification of faster rise time of I1TEA in YFP-16 neurons. The rate of rise was calculated as the maximum derivative over the rising phase of the current (1st 25 ms) at +15 mV (not scaled). Average time of max derivative occurred at 1 ± 0.19 ms after onset of the voltage step in both cell types. The rate of rise of I1TEA was faster in YFP-16 (2,121 ± 1444 pA/ms, filled triangles) than in GIN neurons (999 ± 328 pA/ms, open circles), P = 0.01. E: family of the 10 mM TEA-sensitive I10TEA from a YFP-16 neuron. I10TEA displayed similar kinetics in GIN and YFP-16 neurons. F: plot of the current density of I10TEA in GIN (open circles) and YFP-16 neurons (filled triangles) at each voltage, measured as the average steady-state current density over the last 50 ms of the trace. There was more I10TEA in YFP-16 (201 ± 115 pA/pF; n = 20) than in GIN neurons (123 ± 42 pA/pF; n = 19) neurons at +15 mV, P = 0.05. Error bars represent SE.

The second component of the delayed rectifier current in MVN neurons was measured by subtraction following application of 10 mM TEA (I_10TEA; Fig. 4E). YFP-16 neurons expressed a greater density of I_10TEA than GIN neurons, P = 0.05 (Fig. 4F), but the current did not exhibit different activation rates (351 ± 189 pA/ms, n = 20, YFP-16 vs. 283 ± 167 pA/ms, n = 19, GIN) or different voltage dependences (Table 2) between the cell types. The depolarized activation voltage of I_10TEA (between –25 and –15 mV) and depolarized V_1/2 are consistent with values reported for Kv2-containing channels (Coetzee et al. 1999; Kerschensteiner and Stocker 1999; Murakoshi and Trimmer 1999; Murakoshi et al. 1997).

TEA K⁺ currents

The remaining current in MVN neurons was insensitive to 0.3 mM CdCl₂ and 10 mM TEA. A portion of this current inactivated rapidly on depolarization, was blocked by 5 mM 4-AP (n = 25), and had the classic fast inactivation kinetics of an A current (I_A). Because of the unique voltage-dependent properties of I_A, it was possible to isolate the current without pharmacology. I_A was maximally activated with a 500-ms prehyperpolarizing step to –75 mV then inactivated with a predepolarizing step to –45 mV. The current obtained by subtraction between these two protocols was I_A (Fig. 5A), and the remaining current, not blocked by depolarization to –45 mV, was referred to as I_other.

View larger version (21K): [in this window] [in a new window]   FIG. 5. TEA-insensitive K+ currents in GIN and YFP-16 neurons. A: family of IA from a GIN neuron isolated as the difference current inactivated by a 500-ms predepolarizing step to –45 mV, voltage protocol shown. B: plot of the current density of IA at each voltage in GIN () and YFP-16 neurons (), measured as the peak amplitude over the 1st 50 ms. There was no difference in IA density between GIN (468 ± 179 pA/pF, n = 19) and YFP-16 neurons (400 ± 234 pA/pF; n = 20) at +15 mV, P = 0.25. C: representative trace of Iother from a YFP-16 neuron. D: plot of the current density of Iother in GIN and YFP-16 neurons measured as the average steady-state current density over the last 50 ms of the trace. The density of Iother was greater in YFP-16 (91 ± 48 pA/pF; n = 20, ) than in GIN neurons (59 ± 20 pA/pF; n = 19, s) at +15 mV, P = 0.01. Error bars represent SE.

I_other was insensitive to CdCl₂, TEA, and 4-AP, did not inactivate upon depolarization, and contributed 5% to I_total (Fig. 5, C and D). The expression of this current was not significantly different between GIN and YFP-16 neurons, P = 0.22, and given its small size and lack of specific identification, it was not analyzed further in this study

Balance of outward currents differs between GIN and YFP-16 neurons

GIN and YFP-16 neurons express the same outward currents yet exhibit different action potential and firing properties. The analyses thus far have only considered the absolute levels of current density expression and demonstrate that YFP-16 neurons express more I_1TEA and I_10TEA than GIN neurons. However, these analyses do not address the relative expression levels of currents within neurons, which might better distinguish cell types than the absolute expression level of any individual current.

To determine whether the relative expression levels of outward currents differed between GIN and YFP-16 neurons, the ratios of I_KCa, I_1TEA, I_10TEA, and I_A were compared across the two populations. The balance of currents in each neuron was quantified by normalizing each isolated outward current by I_total, obtained at nominal +15 mV (see METHODS). Currents were quantified using the integral rather than the peak over the first 30 ms to compare currents with different time courses. Qualitatively similar results were also observed over the first 10 ms and the first 50 ms.

The balance of currents varied considerably within and between cell types (Fig. 6). Overall, I_KCa and I_1TEA dominated; however, some neurons had a prominent contribution from I_A. I_10TEA was a small fraction of I_total in all neurons. Although there was a high degree of variability, the expression pattern of currents differed significantly in GIN neurons compared with YFP-16 neurons. GIN neurons had proportionately more I_KCa (P = 0.006) and I_A (P = 0.03), whereas YFP-16 neurons had proportionately more I_1TEA (P = 0.004) and I_10TEA (P = 0.04; Fig. 6A). GIN and YFP-16 neurons were best distinguished by the ratio of I_KCa:I_1TEA. In 84% (16/19) of GIN neurons, the I_KCa:I_1TEA ratio was >1.6 (0.55–3.6, average = 2.3 ± 0.9) and in 86% (18/21) of YFP-16 neurons, the I_KCa:I_1TEA ratio was <1.6 (0.07–4.1, average = 1.2 ± 1.0).

View larger version (12K): [in this window] [in a new window]   FIG. 6. The balance of outward current expression is different in GIN and YFP-16 neurons. The charge density of each current was calculated by integrating the current at +15 mV over the 1st 30 ms of the trace. The charge densities of the 4 major current components were normalized by the charge density of Itotal to control for differences in Itotal density across cells. All 4 currents were found in GIN () and YFP-16 neurons () but GIN neurons had proportionately more IKCa and IA and YFP-16 had proportionately more I1TEA and I10TEA. *P < 0.05, **P < 0.01, ***P < 0.005.

View larger version (18K): [in this window] [in a new window]   FIG. 7. The expression of currents is correlated in GIN and YFP-16 neurons. A: charge density of IKCa over the 1st 30 ms was correlated with the charge density of IA, weakly in GIN neurons (r2 = 0.22, n = 19, P = 0.04) and strongly in YFP-16 neurons (r2 = 0.64, n = 20, P < 0.0001). B: charge density of I1TEA over the 1st 30 ms was correlated with the charge density of I10TEA, strongly in GIN neurons (r2 = 0.62, n = 19, P < 0.0001) and weakly in YFP-16 neurons (r2 = 0.29, n = 20, P = 0.01). C: charge density of IKCa over the 1st 30 ms was correlated with the charge density of I1TEA in GIN (r2 = 0.26, n = 19, P = 0.03) but not YFP-16 neurons (r2 = 0.17, n = 20, P = 0.46). D: charge density of IKCa over the 1st 30 ms was correlated with the charge density of I10TEA in GIN (r2 = 0.42, n = 19, P = 0.001) but not YFP-16 neurons (r2 = 0.01, n = 20, P = 0.71). , GIN; , YFP-16.

How does the differential balance of currents in GIN versus YFP-16 neurons relate to differences in action potential and firing properties of these two cell classes? The AHP in MVN neurons influences firing response gain and depends predominantly on Ca²⁺-dependent K⁺ currents (Johnston et al. 1994; Smith et al. 2002) (Fig. 9, A and B) with additional contributions from 4-AP and TEA-sensitive currents (Johnston et al. 1994). The relatively larger contribution of I_KCa and I_A in GIN versus YFP-16 neurons is consistent with relatively larger AHP in GIN neurons. Across individual neurons, however, neither the amplitude nor density of I_KCa (or any of the other outward current measured in this study) correlated with the magnitude or integral of the AHP (data not shown). The lack of correlations with individual outward currents are consistent with the AHP waveform depending on voltage-dependent interactions of multiple currents, including potentially critical contributions from sodium currents (Akemann and Knopfel 2006; Swensen and Bean 2005).

View larger version (12K): [in this window] [in a new window]   FIG. 8. The rate of action potential repolarization correlated with the rate of rise of I1TEA in both GIN () and YFP-16 neurons (). Rate of rise was calculated as the maximum derivative over the rising phase of the current (see Fig. 4C). This correlation was stronger in GIN neurons (r2 = 0.85, n = 19, P < 0.001) than in YFP-16 neurons (r2 = 0.28, n = 20, P = 0.02).

View larger version (19K): [in this window] [in a new window]   FIG. 9. Variability in I1TEA underlies diversity in action potential repolarization in GIN and YFP-16 neurons. A: action potential from an example GIN neuron (left) in control solution (thick, black line); in 0 Ca2+ 0.3 mM CdCl2 (dotted line); and in 1 mM TEA (gray line). B: action potential from an example YFP-16 neuron (right) in control solution (thick, black line); in 0 Ca2+ 0.3 mM CdCl2 (dotted line); and in 1 mM TEA (gray line). C: rate of action potential repolarization was not slowed in the presence of CdCl2 in either GIN (open circles) or YFP-16 neurons (filled triangles). Solid line represents no change in repolarization rate between Tyrode's and CdCl2. D: rate of action potential repolarization was slowed in the presence of 1 mM TEA in both GIN (open circles) and YFP-16 neurons (filled triangles). Solid line represents no change in repolarization rate between control and 1 mM TEA solutions. Before application of 1 mM TEA, YFP-16 neurons repolarized more quickly than GIN neurons (162 ± 30, n = 7; 119 ± 42, n = 7; P = 0.05), but in 1 mM TEA, GIN and YFP-16 neurons repolarized at the same rate (68 ± 16, n = 7; 61 ± 23, n = 7; P = 0.46). E: extent to which 1 mM TEA slowed action potential repolarization was correlated with the initial rate of repolarization. Neurons that repolarized faster in 1 mM TEA were more strongly affected by 1 mM TEA than those that initially repolarized slower (r2 = 0.79, P < 0.001). Open circles, GIN; filled triangles, YFP-16. F: summary of the effects of 1 mM TEA and CdCl2 on the rate of action potential (AP) rise and fall. Blocking neither IKCa nor I1TEA affected the rate of action potential rise. The action potential fall rate was significantly slowed in 1 mM TEA (P = 0.0002) but not CdCl2 (P = 0.17), indicating the effect of 1 mM TEA on action potential fall rate is specific to I1TEA.

To directly test the contributions of potassium currents to action potentials, neurons were allowed to fire in current clamp and action potentials were compared in control solution and in the presence of pharmacological blockers of the two dominant currents, I_KCa and I_1TEA. As exemplified in Fig. 9, A and B, blocking I_KCa with CdCl₂ (0.3 mM) had no effect on the width of the action potential or rate of repolarization (Fig. 9C) but significantly reduced the magnitude of the AHP in both GIN (by 7.2 ± 2.8 mV; n = 7; P = 0.004) and YFP-16 neurons (by 6.7 ± 2.2 mV; n = 7; P = 0.008). In contrast, blockade of I_1TEA with 1 mM TEA broadened the action potential, increasing action potential width in both GIN (by 0.5 ± 0.3 ms, n = 7; P = 0.03) and YFP-16 neurons (by 0.4 ± 0.2 ms, n = 7; P = 0.008). The increase in action potential width was due specifically to slower repolarization (Fig. 9D) as TEA had no effect on action potential rise rates (Fig. 9F). Interestingly, the effects of TEA on repolarization rate were well correlated with initial repolarization rate (Fig. 9E). Furthermore, 1 mM TEA abolished the differences in repolarization rates of YFP-16 and GIN neurons (P = 0.26). Taken together, these results demonstrate that differences in density and kinetics of I_1TEA account for differences in repolarization rates between GIN and YFP-16 neurons and that I_1TEA affects repolarization rates in a graded manner across MVN neurons.

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION GRANTS ACKNOWLEDGMENTS REFERENCES

 
In this study, recordings from dissociated vestibular nucleus neurons from transgenic mouse lines indicate that heterogeneity in the firing properties of MVN neurons reflects variations in the balance of potassium currents. GABAergic neurons recorded in GIN mice tended to have wider action potentials, deeper AHPs, and lower spontaneous firing rates than did non-GABAergic neurons recorded in YFP-16 mice. GABAergic and non-GABAergic neurons expressed the same four major potassium currents, but GABAergic neurons expressed relatively more I_KCa and I_A and relatively less I_1TEA and I_10TEA than did non-GABAergic neurons. Variations in the expression of I_1TEA, which likely corresponds to current carried through Kv3 channels, accounted for differences in action potential repolarization rates both within and between the two classes of neurons. Shifts in the balance of currents in vestibular nucleus neurons could mediate activity-dependent changes in intrinsic excitability that have been observed in response to acute synaptic inhibition (Nelson et al. 2003) or after loss of peripheral vestibular function (Beraneck et al. 2003, 2004; Cameron and Dutia 1997; Guilding and Dutia 2005; Him and Dutia 2001).

Intrinsic firing properties are preserved in dissociated MVN neurons

Although acutely dissociated MVN neurons had slower action potential kinetics and deeper AHPs than those recorded in slice at room temperature (Bagnall et al. 2007) differences in spontaneous firing rates and action potential waveforms observed in slice between GIN and YFP-16 neurons were largely preserved in the dissociated preparation. Together with the observation that YFP-16 neurons have more dendrites and lower input resistances than GIN neurons in slice (Bagnall et al. 2007) but not in dissociated neurons, these results indicate that dendritic currents contribute to action potential repolarization but that the predominant currents underlying the differences in the firing properties of GIN and YFP-16 MVN neurons are located on the soma and proximal processes (and are not greatly altered by enzymatic or mechanical stress during the dissociation processes). In support of this, the voltage-dependence of the currents, measured by the V_1/2 and slopes of the Boltzmann fits were within the range of reported values for each current with the exception of I_A, the voltage of activation and V_1/2 of which was slightly depolarized compared with more commonly reported values (Bekkers 2000; Molineux et al. 2005; Sacco and Tempia 2002; Song et al. 1998; but see also Martina et al. 1998). This could reflect the presence of CdCl₂ in the perfusion solution, which has been shown to shift the voltage dependence of I_A (Song et al. 1998).

Variations in I_1TEA density and kinetics underlie differences in action potential repolarization rates

A previous model of firing mechanisms in MVN neurons indicated that differences between action potential waveforms across neurons could be attributed predominantly to differences in the kinetics of a fast voltage-gated K⁺ current (Quadroni and Knopfel 1994). Consistent with the predictions of this model, YFP-16 neurons expressed more I_1TEA and I_10TEA than GIN neurons. Variations in I_1TEA accounted entirely for differences in action potential repolarization rates between and within cell types. Based on its depolarized voltage of activation, Boltzmann parameters, and insensitivity to CdCl₂ and dendrotoxin, the 1 mM TEA-sensitive current in MVN neurons likely flows through Kv3-containing channels.

Kv3 currents are expressed in neurons specialized for high-frequency firing (Akemann and Knopfel 2006; Erisir et al. 1999; Hernandez-Pineda et al. 1999; Martina et al. 1998; Massengill et al. 1997; McDonald and Mascagni 2006; McKay and Turner 2004; Perney et al. 1992; Rudy and McBain 2001; Song et al. 2005; Weiser et al. 1995) and their voltage dependence and fast decay kinetics are precisely tuned to facilitate the rapid repolarization of action potentials (Erisir et al. 1999; Raman and Bean 1999). In the cortex, the presence of I_Kv3 distinguishes fast-spiking GABAergic interneurons from excitatory pyramidal cells (Martina et al. 1998; Massengill et al. 1997). In the cerebellar and vestibular nuclei, in contrast, both GABAergic and non-GABAergic neurons are capable of sustaining very fast firing rates (Bagnall et al. 2007; Uusisaari et al. 2007), indicating a role for Kv3 in both cell types in these nuclei. Each of the four major Kv3 family subunits are expressed in MVN neurons (Weiser et al. 1994, 1995). Differences in kinetics across MVN neurons might arise from the expression of different subunits (Baranauskas et al. 2003; Lewis et al. 2004; McCrossan et al. 2003; Murakoshi and Trimmer 1999; Murakoshi et al. 1997), regulation of mRNA transcript levels (Schulz et al. 2006), or different phosphorylation states (Song et al. 2005).

Although I_Kca, was expressed strongly in MVN neurons and exhibited rapid activation during voltage steps, action potential repolarization was dominated by I_1TEA. Results from MVN neuronal recordings in brain slices similarly indicate a prominent role for I_Kca in generation of the AHP but not in action potential repolarization (Smith et al. 2002). Analysis of currents regulating burst firing in Purkinje neurons showed that calcium influx occurs during the falling phase of the action potential and that the peak of I_KCa is delayed compared with the peak of TEA-sensitive repolarizing currents (Swensen and Bean 2003), consistent with pharmacological results in MVN neurons. Thus although voltage step protocols can provide valuable information about current expression levels, assessing how specific currents interact to shape neuronal excitability is facilitated by the use of more natural stimuli, such as action potential waveforms (Raman and Bean 1999; Swensen and Bean 2003).

Coregulation of currents in MVN neurons

Given that firing properties are shaped by a the interplay of intrinsic currents, neurons must be able to actively monitor and adjust the balance of currents accordingly. In support of such a mechanism in MVN neurons, correlations were observed in charge densities of several currents. In both GIN and YFP-16 neurons, significant positive correlations existed between I_1TEA and I_10TEA and between I_KCa and I_A. Additional correlations between I_KCa and I_1TEA and between I_KCa and I_10TEA were observed in GIN neurons but not in YFP-16 neurons. These data suggest that outward currents are functionally coregulated in MVN neurons but that the rules for this coregulation differ across cell types.

Noninactivating Na⁺ currents are likely to play a prominent role in shaping the firing properties of MVN neurons as is the case for cerebellar neurons (Khaliq et al. 2003; Raman and Bean 1997; Raman et al. 2000). Although no correlations were observed in MVN neurons between transient I_NaT and I_1TEA kinetics or densities, action potential rise and fall rates were well-matched (Fig. 9D), suggesting an interaction between Na⁺ and K⁺ currents that might only be revealed during natural spiking behavior (Akemann and Knopfel 2006; Swensen and Bean 2005). Coregulation of Na⁺ and K⁺ currents have been observed in neurons of the electric fish, where the kinetics of the currents co-vary as a function of the neuronal output properties (McAnelly and Zakon 2000). Coregulation of currents might occur at the transcriptional level (MacLean et al. 2005; Schulz et al. 2006), by posttranslational modifications (Park et al. 2006; Song et al. 2005), or by functional interactions via voltage dependence of the currents themselves (Akemann and Knopfel 2006; Swensen and Bean 2005).

Implications for plasticity

MVN neurons express a novel form of intrinsic plasticity, termed firing rate potentiation (FRP), which produces increases in spontaneous and evoked firing rates via decreases in the AHP (Nelson et al. 2003). FRP is accompanied by a decreased sensitivity to IBTX and is occluded by blockade of CaMKII, which reduces BK currents (Nelson et al. 2005). Most neurons in the MVN have the capacity to express FRP, but its expression varies across neurons (Nelson et al. 2003, 2005). This finding is better understood in light of the variability of I_KCa expression across the population of MVN neurons. The presence of other currents, such as I_A, might compensate functionally for the loss of BK currents in some neurons.

The finding that GABAergic and non-GABAergic neurons possess the same major outward currents is significant because it suggest that both cell types express currents required for a broad range of firing properties. Long-term changes in the ratio of "type A" and "type B" neurons in the MVN, which appear to correspond to GIN and YFP-16 neurons in slice, respectively (Bagnall et al. 2007), have been reported during recovery from unilateral labyrinthectomy (Beraneck et al. 2003, 2004). The data from the present study would suggest that a GIN neuron with a type A action potential shape could adopt a type B action potential shape if there were a shift in the I_KCa:I_1TEA ratio, induced either by a downregulation of BK currents, which occurs during FRP, or an increase in the kinetics or expression of I_1TEA. Rapid shifts in the balance of currents, and by extension in firing properties, could be induced by phosphorylation-dependent changes in current kinetics or channel conductances as has been shown for each of the predominant potassium currents expressed in MVN neurons (Jerng et al. 2004; Koh et al. 1999; Liu and Kaczmarek 1998; Nelson et al. 2005; Park et al. 2006; Sansom et al. 2000; Sergeant et al. 2005; Smith et al. 2002; Song et al. 2005). Regulation of the firing properties of neurons on a fast time scale might be especially important in systems with high levels of activity, such as the vestibular system in which neurons fire at rates of hundreds of action potentials/second in vivo. Activity-dependent shifts in the balance of currents would provide rapid, on-line regulation of firing properties, maintaining the balance of activity across the network.

	GRANTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION GRANTS ACKNOWLEDGMENTS REFERENCES

 
This work was supported by the Howard Hughes Medical Institute, National Eye Institute Grant EY-11027, and the Legler-Benbough Foundation

	ACKNOWLEDGMENTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION GRANTS ACKNOWLEDGMENTS REFERENCES

 
We thank I. Raman for technical advice and P. Slesinger for helpful discussions.

	FOOTNOTES

 
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.

Address for reprint requests and other correspondence: S. du Lac, Systems Neurobiology Laboratories, The Salk Institute for Biological Studies, 10010 N. Torrey Pines Rd., La Jolla, CA 92037 (E-mail: sascha{at}salk.edu)

	REFERENCES

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION GRANTS ACKNOWLEDGMENTS REFERENCES

 
Akemann W, Knopfel T. Interaction of Kv3 potassium channels and resurgent sodium current influences the rate of spontaneous firing of Purkinje neurons. J Neurosci 26: 4602–4612, 2006.[Abstract/Free Full Text]

Bagnall MW, Stevens RJ, du Lac S. Transgenic mouse lines subdivide medial vestibular nucleus neurons into discrete, neurochemically distinct populations. J Neurosci 27: 2318–2330, 2007.[Abstract/Free Full Text]

Baranauskas G, Tkatch T, Nagata K, Yeh JZ, Surmeier DJ. Kv3.4 subunits enhance the repolarizing efficiency of Kv3.1 channels in fast-spiking neurons. Nat Neurosci 6: 258–266, 2003.[CrossRef][ISI][Medline]

Bekkers JM. Properties of voltage-gated potassium currents in nucleated patches from large layer 5 cortical pyramidal neurons of the rat. J Physiol 525: 593–609, 2000.[Abstract/Free Full Text]

Beraneck M, Hachemaoui M, Idoux E, Ris L, Uno A, Godaux E, Vidal PP, Moore LE, Vibert N. Long-term plasticity of ipsilesional medial vestibular nucleus neurons after unilateral labyrinthectomy. J Neurophysiol 90: 184–203, 2003.[Abstract/Free Full Text]

<