| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Laboratoire de Neurobiologie des Réseaux Sensorimoteurs, Centre National de la Recherche Scientifique-Université Paris 5, Unite Mixte de Recherche 7060, 75270 Paris Cédex 06, France
Submitted 18 February 2004; accepted in final form 10 May 2004
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMETHODSRESULTSDISCUSSIONGRANTSREFERENCES |
|---|
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONGRANTSREFERENCES |
|---|
; Darlington and Smith 1996; Dieringer 1995; Smith and Curthoys 1989). This syndrome includes a spontaneous ocular nystagmus and massive postural distortions observed in the absence of movement, and abnormalities of the vestibulo-ocular and vestibulo-spinal synergies observed during head movements. Whereas the static disturbances disappear in a few days in most vertebrate species, the dynamic reflexes remain impaired indefinitely. They improve over several weeks, but this recovery is limited to low acceleration and the low to middle frequency range of head movements, both in animals (Broussard et al. 1999b; Gilchrist et al. 1998; Lasker et al. 2000; Murai et al. 2003; Vibert et al. 1993) and humans (Allum and Ledin 1999; Brandt 2000; Curthoys and Halmagyi 1999).
The vestibular system works bilaterally, using sensory information from both labyrinths simultaneously for gaze and posture stabilization. Most subdivisions of the vestibular nuclei are linked by commissural fibers that interconnect the 2 vestibular complexes located on each side of the brain stem through monosynaptic or disynaptic pathways. Commissural connections are particularly extensive at the level of the medial and superior vestibular nucleus (Epema et al. 1988; Gacek 1978; Ito et al. 1985; Newlands et al. 1989). These commissural pathways are predominantly inhibitory but also include excitatory connections (Babalian et al. 1997; Highstein et al. 1987). The role of the contralesional, intact labyrinth and of commissural pathways in the rapid disappearance of the static symptoms triggered by unilateral labyrinthectomy is controversial (Cartwright and Curthoys 1996; Graham and Dutia 2001; see Curthoys and Halmagyi 1999; Dieringer 1995; Smith and Curthoys 1989). In contrast, the contralesional labyrinth is obviously essential for the recovery of the dynamic vestibular reflexes. Indeed, it provides all the remaining sensory vestibular information related to linear and angular head movements.
Few in vivo data are available on how the contralesional vestibular neurons are modified after unilateral labyrinthectomy, but they all suggest that substantial adaptive mechanisms are triggered during vestibular compensation. Immediately after the lesion, the spontaneous discharge of contralesional MVNn recorded in awake guinea pigs increases (Ris and Godaux 1998). This is probably attributable to the loss of the commissural inhibition normally provided by the ipsilesional MVNn, which are deprived of their vestibular excitatory drive (Curthoys and Halmagyi 1995; Darlington and Smith 1996; Dieringer 1995; Smith and Curthoys 1989). The spontaneous discharge of contralesional MVNn returns to normal within 1 wk. After several months, the contralesional vestibular neurons show structural changes (Gacek et al. 1998). In patients, caloric stimulation of the intact labyrinth reveals first a reduction of excitability during the first months after the lesion, then a recovery of normal values in about 1 yr, which afterward develops into a hyperexcitability during the next year (Fisch 1973). In monkeys, caloric response of the contralesional labyrinth to ice water is increased after 3 mo of vestibular compensation (Fetter and Zee 1988).
Intracellular recordings in brain stem slices led to the identification of 2 types of MVNn, type A and type B neurons, based primarily on different action potential profiles (Gallagher et al. 1985; Johnston et al. 1994; Serafin et al. 1991a, b). Previous studies (Babalian et al. 1997; Ris et al. 2001) have shown that type A MVNn correspond to the tonic, regular vestibular neurons identified in vivo, and have a more stable spontaneous activity and more linear response dynamics. In contrast, type B MVNn correspond to the phasic, irregular vestibular neurons identified in vivo, and show more irregular spontaneous activity and a higher sensitivity to applied current. Recently, we showed that 1 mo after a unilateral labyrinthectomy, the intrinsic firing properties of ipsilesional MVNn were greatly modified (Beraneck et al. 2003a). The ipsilesional neurons were depolarized by 6–10 mV, and the static and dynamic membrane properties of type B neurons were more similar to those of type A neurons than in control slices. Altogether, the neurons on the ipsilesional side showed a more tonic behavior, which should increase the stability and regularity of the resting discharge that is restored in these MVNn during vestibular compensation.
In view of the prominent role of the contralesional labyrinth in the generation of vestibular synergies after unilateral labyrinthectomy, additional experiments were done on the static and dynamic membrane properties of contralesional MVN neurons after 1 mo of compensation. Statistical comparisons were made with identical measurements obtained on slices from normal animals, and on the ipsilesional side of slices taken 1 mo after unilateral labyrinthectomy (Beraneck et al. 2003a; Ris et al. 2001). Some of these data were previously published in abstract form (Beraneck et al. 2003b).
|
METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONGRANTSREFERENCES |
|---|
Experiments were carried out on pigmented guinea pigs of both genders (Elevage de la Garenne, Saint-Pierre d'Exideuil, France). The animals were handled in accordance with the European Communities Council Directive of November 24, 1986, and followed the procedures issued by the French Ministère de l'Agriculture.
Unilateral labyrinthectomies were performed under halothane anesthesia as described in Vibert et al. (1999a, b). Guinea pigs were allowed to compensate in a normal visual environment until their brain was removed after about 1 mo of compensation (mean of 32 days, range: 27 to 41 days) to prepare the slices. The 28 animals used in this experiment were 8–10 wk of age, and their weight ranged from 250 to 450 g (mean about 350 g). These parameters were similar to those of the animals used in our previous experiments (Beraneck et al. 2003a).
Intracellular electrophysiological recordings
After nembutal (pentobarbital) anesthesia, thick (500 µm), coronal brain stem slices were cut from previously labyrinthectomized animals and maintained using standard techniques (Gallagher et al. 1985; Serafin et al. 1991a; Vibert et al. 1999b). In most experiments, a cold (4°C) sucrose-containing artificial cerebrospinal fluid (sucrose ACSF) was used during the dissection and preparation of the slices (Devor et al. 2001; for details see Uno et al. 2003). During the recordings, however, the slices were superfused with normal ACSF maintained at 31–32°C (Serafin et al. 1991a; Vibert et al. 1999b). Use of the sucrose solution during the preparation of the slices increased the number of viable neurons recorded with microelectrodes. Because we obtained a few contralesional neurons (9 out of 74) without using the sucrose ACSF, we verified that neither their static nor their dynamic membrane properties were different from those of the contralesional MVNn obtained using the sucrose ACSF (see Uno et al. 2003).
Intracellular electrophysiological recordings were obtained with sharp, 3 M potassium acetate-containing glass microelectrodes from neurons within the contralesional medial vestibular nucleus (MVN). As in Beraneck et al. (2003a), recordings were restricted to the two 500-µm coronal slices corresponding to the middle third of the guinea pig MVN, at the level of the cerebellar peduncles. A few cells (<10%) were recorded in more caudal slices taken from the last third of the MVN. About 20% of the experiments were performed blind to the recording side to check for any experimenter-linked bias. It is likely that a large majority (>80%) of the MVNn recorded in slices are second-order vestibular neurons, given that 80–85% of the central vestibular neurons recorded in the MVN area of the isolated whole brain of guinea pig with similar electrodes were identified as second-order neurons (Babalian et al. 1997).
All measurements were done with an Axoclamp 2A system (Axon Instruments, Union City, CA) in the bridge mode, current-clamp configuration. The electrode resistance varied from 80 to 150 M
. Both series resistance (bridge balance) and capacitance compensation were checked throughout the recording of each individual neuron (Ris et al. 2001). The current stimulation and data acquisition were done with a PC-compatible computer using the Acquis 1 program (version 4.0, Bio-logic S.A., Gif-sur-Yvette, France) or MATLAB 6.5 (The MathWorks, Natick, MA). The data were analyzed using program scripts with Mathematica 4.0 (Wolfram Research, Champaign, IL) or MATLAB 6.5 (The MathWorks).
Basic membrane and firing properties of MVNn
Because most MVNn are spontaneously active on slices, the membrane potential was filtered with a 1-Hz low-pass filter to obtain an estimate of a "mean resting membrane potential" for each neuron. As in Beraneck et al. (2003a), all cells that had a resting potential more negative than –50 mV and a spike amplitude >50 mV were analyzed. We also analyzed the MVNn whose membrane potential ranged from –50 to –40 mV if their spike amplitude was 50 mV or more, with a normal spike width ranging from 0.7 to 2.2 ms (see Beraneck et al. 2003a).
Recordings of the neurons at rest were used to determine their mean spontaneous firing rate, the coefficient of variation (CV), and the amplitude of the spike. For each neuron, an average of the spike shape and following interspike interval was obtained by averaging successive spikes taken either at rest or during slight depolarizations, if the neurons were silent at rest. The spikes were synchronized to their thresholds, defined as the time the positive slope of the action potential reached 10 V/s (Krawitz et al. 2001). The averaged spike shape was used to determine the amplitude of the afterhyperpolarization (AHP) and the width of the spike at threshold (see Beraneck et al. 2003a).
The cell's firing threshold (i.e., the membrane potential for which the cell begins to fire action potentials) was measured as the potential reached by the neuron at the threshold of the first spike triggered by a slow, depolarizing current ramp (0.06 nA/s). In each cell, the presence and duration of long-lasting, subthreshold plateau potentials were determined. These plateau potentials were elicited by low-amplitude (0.1–0.2 nA), short-duration (10 ms) current pulses, while the neurons were hyperpolarized just below their firing threshold (Babalian et al. 1997; Serafin et al. 1991a,b).
Quantitative determination of the neuronal type
In previous publications, MVN neurons were classified as type A, B, B+LTS, and C neurons using only qualitative criteria. To assess reliably the long-term effects of unilateral labyrinthectomy on ipsi- and contralesional MVNn, quantitative, objective criteria were developed to classify intracellularly recorded MVN neurons as briefly explained below (for a complete description see Beraneck et al. 2003a).
The first derivative of the averaged spike profile of each neuron was used to assess (in V/s) the A-like rectification and double AHP, the 2 main criteria previously used for the qualitative classification. All MVNn showing an A-like rectification with an amplitude <0.15 V/s were classified as type B neurons. The type B+LTS MVNn formed a subtype of these type B neurons showing low-threshold calcium spikes when released from a strong hyperpolarization. The MVNn that had an A-like rectification stronger than 0.15 V/s and no double AHP were classified as type A neurons. The few MVNn that showed both a double AHP and an A-like rectification >0.15 V/s were considered as type C neurons.
These criteria were previously used by Beraneck et al. (2003a) to classify and compare the neurons recorded on slices taken from normal animals (control slices), and on the ipsilesional side of slices taken from animals labyrinthectomized 1 mo before. The MVNn recorded in control slices included 47% type A neurons, 50% type B neurons (of which 9% were type B+LTS neurons), and 3% type C neurons.
Measurement of the input resistance of MVNn using current steps
The passive input resistance of each neuron was assessed using a series of hyperpolarizing current steps (1-s duration) of decreasing amplitudes. The cell was maintained by a steady-state hyperpolarization at a few mV (0–10) below its threshold for discharge. The whole cell resistance for each MVNn (input resistance = voltage deflection/current input) was estimated from the final steady-state amplitude of the hyperpolarizing steps.
Stimulation with depolarizing ramp currents
Increasing ramp currents of 0.3 nA amplitude were applied at 5 different slopes up to a final steady-state value, leading to a proportionate increase in the firing rate above the resting spontaneous activity (for details see Beraneck et al. 2003a). The slope of increase of the instantaneous firing rate of the cell (kIF in spikes · s–1 · nA–1) was estimated during the depolarizing, ramplike portion of the applied current (Fig. 1A). The difference between the firing rate reached at the end of the applied depolarizing current and the final, stable discharge rate reached at the end of the plateau was measured as an overshoot in spikes/s (Fig. 1A). To assess how the level of polarization influenced the responses, the whole sequence of ramp stimulations was repeated from a hyperpolarized level of about 10 mV below the firing threshold. Of the 5 ramps applied to each cell, the 600-ms (slope of 0.5 nA/s) and 200-ms ramps (slope of 1.5 nA/s) gave the most significant results and were taken as the main indices of the response of MVNn to ramplike currents.
|
A third series of stimuli consisted of current sine waves applied for 5,000 ms at frequencies ranging from 0.2 to 50 Hz (du Lac and Lisberger 1995; for details see Ris et al. 2001). The amplitude of the stimulus (
I, Fig. 1B) was adjusted to keep the membrane potential variation around 10 mV peak to peak. The sinusoidal currents applied in the presence of resting spontaneous activity led to a significant modulation of the firing rate. For each frequency of stimulation below or equal to about one third of the neuron's resting discharge, the modulation of the instantaneous firing rate (IF) of MVNn was fitted with a sine wave that was used to calculate the amplitude and phase of the IF modulation (IF, Fig. 1B). When the frequency of stimulation passed a third of the neuron's firing rate, the amplitude of the IF modulation was calculated empirically as the difference between the minimum and maximum IF reached by the neuron during the stimulation. No phase measurements were made in this situation. The underlying membrane potential excursion (V) was computed by a Fourier analysis of the total membrane potential response (Fig. 1B). The magnitude of the Fourier component corresponding to the stimulation frequency was taken as the potential response. IF and I were used to evaluate at 0.4 Hz the cell sensitivity to current injection by dividing IF by the amplitude of the injected current (IF/I in spikes · s–1 · nA–1). The sensitivity of the firing rate of the cell to variations of the mean membrane potential IF/V was also measured, in spikes · s–1 · mV–1. The "active" impedance Z of the cell was calculated as the amplitude of the membrane potential change obtained for the 0.4-Hz stimulus divided by the amplitude of the injected current (V/I in M). In general, a spike rate transfer function can be defined for each neuron as the ratio IF/I versus the stimulation frequency. In most experiments, a similar series of sinusoidal stimuli was given while the cell was maintained at a depolarized membrane potential by a steady-state current stimulation of 0.15–0.25 nA.
Most of the cells were also submitted to the same series of sinusoidal current injections while they were maintained at 10 to 20 mV below their disharge threshold, so that no spike was evoked by the stimulation. The amplitude and phase of the membrane potential change (Vh) was computed for each frequency, and the response to the 0.4-Hz stimulus was used to evaluate the impedance Zh of the cell maintained under a steady-state hyperpolarization (Zh = Vh/I in M). An impedance transfer function can be defined for each neuron as the ratio Vh/I versus the stimulation frequency.
For each MVNn, the modulation of the firing rate (i.e., the amplitude of the spike rate transfer function) increased with stimulation frequency to reach a maximum at the peak frequency of resonance. Then, the modulation progressively decreased to lower levels. The "amplitude" of the resonance was defined as the ratio between the amplitude of the firing rate modulation at the peak frequency of resonance and the amplitude obtained at the lowest frequency (0.2 Hz). When the neurons were hyperpolarized to suppress action potentials, the amplitude of the resonance was measured for the membrane potential response. Both the impedance and spike rate transfer functions showed a small phase lead with respect to the injected current at the lowest frequencies of stimulation, which decreased to zero and became a phase lag at higher frequencies (Beraneck et al. 2003a). For all MVNn, we could therefore define as the "zero-crossing frequency" the lowest frequency for which a phase lag was measured.
Statistical analysis
Data presented in this study were obtained from a database of 74 MVNn recorded on the contralesional side of slices taken from animals labyrinthectomized about 1 mo before (contralesional neurons). All mean values are presented with their SD. Statistical analysis was carried out using the Systat 8.0 software (SPSS, Chicago, IL). For each parameter, normality of the distributions was assessed using one-sample Kolmogorov–Smirnov tests, with significance set at P 
0.05. Statistical comparisons were achieved through either parametric (for normal distributions including a minimum of 15 samples) or otherwise nonparametric tests, with the threshold for significance set at P 0.05. Type B + LTS neurons were pooled together with the other B neurons for analysis. ANOVA or the nonparametric Kruskal–Wallis ANOVA was first performed to search for differences between the average values obtained for type A and B neurons recorded in control slices, and on both sides of slices taken from labyrinthectomized animals (which defined 6 categories of neurons). Comparisons (2 x 2) between the cell groups were then performed using Student's t-test or the nonparametric Mann–Whitney U-tests. Paired parametric (ANOVA followed by paired t-test) or nonparametric tests (Friedman ANOVA followed by Wilcoxon signed-rank tests) were used to compare the responses evoked by ramps of different slopes, and to determine how the responses to ramps and sinusoidal currents were modified by the polarization level of the neurons.
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONGRANTSREFERENCES |
|---|
As described in METHODS, quantitative criteria were developed to categorize MVNn according to the measure of the A-like rectification and/or double AHP shown by each neuron (see METHODS; Beraneck et al. 2003a). Among the 74 contralesional MVNn we recorded (Fig. 2A), there were 20 type A neurons (27.0%) and 53 type B neurons (71.6%). Four of the type B neurons (7.5%) displayed low-threshold calcium spikes and were B+LTS MVNn. Only 1 contralesional MVNn (1.4%) was a type C neuron.
|
In addition, the parameters used to categorize MVNn, that is, the A-like rectification and the double AHP (Table 1) also confirmed these opposite changes between the 2 sides after 1 mo of compensation. The average measure of the double AHP of contralesional MVNn was multiplied by more than 2 compared with control MVNn (P = 0.01, Fig. 2A). In contrast, the double AHP of ipsilesional MVNn was divided by more than 2 compared with control conditions (P = 0.01; Beraneck et al. 2003a). Whereas the double AHP of contralesional MVNn was increased, the average measure of their A-like rectification tended to decrease (Table 1) compared with control MVNn (P = 0.08), and was lower than that on the ipsilesional side (0.42 ± 0.58 V/s, P = 0.02).
|
), that is, an increased proportion of type A neurons and a smaller double AHP. Membrane potential and firing rate of contralesional MVNn
Contralesional MVNn had a mean resting membrane potential of –56.3 ± 5.4 mV (Table 1) and were slightly but significantly depolarized compared with control MVNn (–58.7 ± 8.5 mV, P = 0.048). This change was restricted to type B MVNn, whose resting membrane potential was depolarized by almost 5 mV (P = 0.004) compared with control type B neurons. The depolarization of contralesional type B neurons was accompanied by a significant depolarization of their firing threshold (P = 0.003, Table 1).
The mean resting discharge displayed by contralesional MVNn (24.8 ± 19.9 spikes/s) was similar to the one of control MVNn (P = 0.56, Table 1). Type A and B contralesional MVNn had the same mean resting discharge (P = 0.68). The increase in firing threshold that accompanied the depolarization of contralesional type B MVNn was consistent with the lack of an increase in their average firing rate. Interestingly, the distribution of the resting discharges of contralesional MVNn was modified compared with control slices (Fig. 3A). Specifically, 17.4% of the contralesional neurons were silent at rest (Fig. 3B), whereas silent neurons represented only 5.3% of MVNn in control conditions and 3.6% of ipsilesional MVNn. The proportion of silent neurons was 22.2% for contralesional type A MVNn and 14% for contralesional type B MVN neurons. Those silent contralesional neurons had more hyperpolarized resting membrane potentials (–60.5 ± 6.1 mV, P = 0.008), and more depolarized firing thresholds (–52.1 ± 3.9 mV, P < 0.001) than the other contralesional MVNn. In contrast, the proportion of neurons whose spontaneous firing rate ranged from 10 to 20 spikes/s was 11.6% among contralesional MVNn versus 23.7% among control MVNn (Fig. 3, A and B). The regularity of the spontaneous discharge of contralesional MVNn assessed by their average CV was not different from that of normal MVNn (P = 0.31, Table 1). There was no difference in the CV between type A and type B contralesional MVNn (P = 0.78), whereas in control slices the regularity of type A MVNn was significantly greater than that of type B MVN neurons. A similar disappearance of this difference in regularity was previously observed for ipsilesional neurons (Beraneck et al. 2003a).
|
The afterhyperpolarization (AHP) of MVNn recorded on the contralesional side of slices after 1 mo of compensation had an average amplitude of 14.8 ± 4.6 mV (Table 1), and was significantly smaller than the AHP of MVNn recorded in control slices (P = 0.016). This decrease was related in part to the higher proportion of type B MVNn among contralesional neurons. Indeed, as on control slices, the AHP of contralesional type B MVNn was significantly smaller than that of type A neurons (P < 0.001). The decrease in the size of the AHP of contralesional MVNn was observed for both cell types, but was not significant when each type was considered individually (P = 0.12 for type A and P = 0.15 for type B MVNn). This decrease of the AHP of contralesional MVNn contrasts again with observations on the ipsilesional side where the size of the AHP of MVNn was higher than that in control slices (Beraneck et al. 2003a). In contrast, neither the amplitude nor the width of the spikes was modified compared with control MVNn (Table 1).
Subthreshold plateau potentials could be triggered in 23% of contralesional type A MVNn and 83% of contralesional type B MVN neurons. Neither the proportions of MVNn exhibiting subthreshold plateau potentials nor the duration of these plateau potentials was different from control values (Table 1).
Finally, the input resistance of contralesional MVNn measured with steps delivered in the presence of a steady-state hyperpolarization was 86.2 ± 33.8 M (Table 1). It was similar to the one of control MVNn, but lower than that of ipsilesional MVNn (P = 0.001; Beraneck et al. 2003a).
Responses of contralesional MVNn to ramp currents
The basis of the change in the neuronal populations after a labyrinthectomy clearly lies in the intrinsic voltage-dependent conductances of vestibular neurons. Changes in these membrane properties can be demonstrated by probing the neurons with a range of ramp stimuli having different slopes providing low- and high-frequency stimulation. Contralesional type A MVNn showed significantly increased overshoots (Fig. 4A) and slopes of firing rate increase kIF (Fig. 4B) in response to ramps delivered from rest. For the 600-ms ramp, the overshoot of contralesional type A MVNn increased nearly a factor of 3 compared with control type A MVNn (P = 0.01), and the kIF increased by 33% (P = 0.03). Because the overshoots and slopes of contralesional type B MVNn did not increase significantly compared with control type B MVNn (Fig. 4), the significant difference observed between the overshoots of type A and type B MVNn in control slices disappeared. The kIF measured on the 600-ms ramp became greater in type A than in type B contralesional MVNn (P = 0.042), which is the reverse of that observed in control slices where the kIF of type A MVNn was lower than that of type B MVNn.
|
). This reveals a strong and unusual dependency of the sensitivity of contralesional type A MVNn on their membrane potential as revealed by high-amplitude ramp current stimulation (Fig. 5).
|
The increased overshoot displayed by contralesional MVNn in response to ramps is similar to that observed on the ipsilesional side (Beraneck et al. 2003a). However, the strong and selective increase of the kIF of contralesional type A MVN for the ramps delivered from rest is specific to the contralesional side.
Impedance functions of contralesional MVNn measured with sinusoidal currents delivered during steady-state hyperpolarization and in the absence of action potentials
When measured at 0.4 Hz, the impedance of hyperpolarized contralesional MVNn decreased by 31% compared with control MVNn (P = 0.001, Table 2). This decrease was significant for all frequencies of sinusoidal stimulation from 0.2 to 50 Hz (Fig. 6A1). This reflected a selective decrease of the impedance of contralesional type B MVNn, which was significantly lower than that of control type B MVNn (Fig. 7B1) as well as contralesional type A MVNn over the whole range of frequencies tested (Table 2, Fig. 7). The decreased impedance of contralesional type B MVNn was not associated with any change in the peak frequency or amplitude of the resonance (Table 2). However, there was a selective increase of their zero-crossing frequency (Table 2), which reached a median value of 0.8 Hz instead of 0.3 Hz for control type B MVNn (P = 0.02). In accordance with this increase, contralesional type B MVNn displayed significantly bigger phase leads at low frequencies (P = 0.035 at 0.6 Hz) and smaller phase lags at intermediate frequencies (P = 0.02 at 2 Hz). In contrast, neither the phase nor the amplitude function of contralesional type A MVNn was modified compared with control conditions. These findings suggest that the marked decrease in the impedance of contralesional type B MVNn could be a result of the potentiation of a current activated during hyperpolarization.
|
|
|
Impedance and spike rate transfer functions of contralesional MVNn: amplitude functions
The active impedance, Zrest, of contralesional MVNn measured at their resting membrane potential (45.8 ± 35.8 M at 0.4 Hz) was decreased by 22% compared with control MVNn (58.5 ± 33.9 M, P = 0.01, Table 3). This decrease was observed at rest for both type A (–30%, P = 0.02) and type B (–24%) contralesional MVNn, in contrast to being restricted to type B neurons at hyperpolarized potentials (Table 3). During steady-state depolarization, the active impedance of contralesional MVNn decreased by 19% compared with control MVNn (but P = 0.13 only, Table 3).
|
IF/I of the discharge of contralesional MVNn to applied current increased compared with control MVNn, both at rest and during depolarization. This increase of the average sensitivity of MVNn to current injection was either significant (P 0.05) or visible as a strong trend (P values ranging from 0.05 to 0.1) for all frequencies of stimulation between 0.2 and 30 Hz at rest (Fig. 6B1) and between 0.2 and 40 Hz during depolarization (Fig. 6C1). As discussed in METHODS, this sensitivity is a spike rate transfer function of the MVNn firing rate modulation relative to the current input for the entire range of stimulation frequencies used. As a rule, the sensitivity of the discharge of contralesional MVNn to applied current increased more for type A than type B MVNn, particularly at low frequencies (Table 3, Fig. 7). Indeed, the sensitivity of type A MVNn to applied current was significantly increased for most frequencies between 0.2 and 20 Hz at rest (Fig. 7A2), and between 4 and 50 Hz during depolarization. In contrast, the sensitivity of contralesional type B MVNn was increased only for high frequencies of stimulation ranging from 12 to 20 Hz at rest, and from 12 to 40 Hz during depolarization (Fig. 7B2). Altogether, the significant difference in sensitivity to current that was observed between type A and type B control MVNn (Beraneck et al. 2003a) disappeared.
The increased sensitivity of the discharge of contralesional MVNn (IF/I) to current injection combined with the decrease in their active impedance (V/I) means that there was a strong increase in the sensitivity of their discharge relative to the membrane potential IF/V, both at rest and during depolarization. At 0.4 Hz, IF/V increased by 51% at rest (P = 0.004, Table 3), and by 57% during steady-state depolarization (P = 0.025). This increased sensitivity of contralesional neurons to membrane potential was observed for both types of MVNn (Table 3).
The amplitude of the resonance shown by contralesional MVNn was not modified compared with control values, either at rest or during steady-state depolarization (Table 3), nor was the peak frequency of resonance of contralesional type A MVNn. However, the median peak frequency of resonance of contralesional type B MVNn tended to increase (Table 3), and the significant difference between the peak frequency of resonance of type A and B MVNn observed in control slices (Beraneck et al. 2003a) was no longer present.
Interestingly, data from individual neurons demonstrated that most contralesional MVNn did not show the gradual decrease of the modulation of their firing rate by current injection usually observed for control and ipsilesional MVNn after the peak frequency of resonance (Beraneck et al. 2003a). In contrast, the modulation of spontaneous firing stopped rather abruptly, given that at higher frequencies of stimulation contralesional MVNn tended to synchronize their firing with the depolarizing phase of sinusoidal current injections. This phenomenon, which was previously described for lateral vestibular nucleus neurons (Uno et al. 2003), was observed for both types of contralesional MVNn.
Spike rate transfer functions of contralesional MVNn: phase functions
Just as active conductances can manifest themselves as a resonance in amplitude functions, phase functions show a phase lead at low frequencies, then cross zero near the peak resonant frequency, and finally show a phase lag that increases with higher frequencies. Compared with control MVNn, the phase functions of the firing rate modulation displayed by contralesional type A and type B MVNn were significantly modified. The median zero-crossing frequency of contralesional MVNn reached 3 Hz instead of 1 Hz at rest and 4 Hz instead of 1 Hz under steady-state depolarization (Table 3, P < 0.001 in both cases). Significant phase shifts toward higher phase leads and smaller phase lags were associated with this shift of the zero-crossing frequency. For the whole sample of contralesional MVNn, there was a significant change of the phase function for most frequencies of stimulation between 0.4 and 10 Hz at rest (Fig. 6B2) and for all frequencies between 4 and 20 Hz under steady-state depolarization (Fig. 6C2). The changes in phase functions were more important in type A than in type B MVNn. Indeed, the zero-crossing frequency of contralesional type A MVNn reached 6 Hz instead of 1 Hz in control slices at rest (P = 0.001) and 7 Hz instead of 2 Hz during steady-state depolarization (P = 0.009). As a result, the zero-crossing frequency of contralesional type A MVNn was higher than that of contralesional type B MVNn (Table 3), whereas the phase functions of control type A and type B MVNn were similar. Contralesional type A MVNn also displayed significantly larger phase leads and smaller lags than those of contralesional type B MVNn for frequencies of stimulation between 2 and 20 Hz at rest, and 2 and 30 Hz during depolarization.
Responses of contralesional MVNn to sinusoidal currents delivered in the presence of action potentials: comparison with the ipsilesional side
The increased sensitivity of the firing rate of contralesional MVNn to current was similar to that observed on the ipsilesional side (Beraneck et al. 2003a), but was associated with a stronger increase of the sensitivity of the discharge of contralesional MVNn to membrane potential (IF/V). Although a shift of the zero-crossing frequency was not observed on the ipsilesional side, ipsilesional MVNn did show a slight shift of their phase function and had larger phase leads and smaller phase lags compared with those of control MVN neurons. In general, the modifications of the phase functions were small for ipsilesional MVNn compared with those of contralesional MVNn. In addition, the most pronounced phase effects were on type A MVNn on the contralesional side, whereas they were practically restricted to type B MVNn on the ipsilesional side (Beraneck et al. 2003a).
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONGRANTSREFERENCES |
|---|
The plasticity of intrinsic membrane properties is strikingly demonstrated in the nearly opposite changes observed in the excitability of ipsilesional and contralesional vestibular neurons after a unilateral labyrinthectomy. These complementary modifications are even apparent in the distributions of particular populations of neurons based solely on action potential profiles, that is, type A and B neurons described previously (Beraneck et al. 2003a). After 1 mo of vestibular compensation (Fig. 8), there was a strong increase in the proportion of type B MVNn and a concomitant decrease in the proportion of type A MVNn among contralesional neurons. The AHP of contralesional MVNn decreased, in contrast to the increase observed on the ipsilesional side. The dynamic responses of contralesional MVNn (i.e., spike rate transfer functions) were also modified. Although the active impedance of contralesional MVNn was decreased, the amplitude of the spike rate modulation increased. These findings are consistent with an increase in active voltage-dependent conductances manifested by an increase in the sensitivity of the discharge of contralesional MVNn to membrane potential, IF/V. There was also a significant phase shift toward greater phase leads and smaller lags over the observed frequency range. In general, the dynamic responses of contralesional type A MVNn were more modified than those of type B MVNn, in contrast to that observed on the ipsilesional side. In both cases, the response dynamics of MVN neurons became more homogeneous than in the control situation (Fig. 8).
|
There is a wide range of examples in the literature in which unilateral peripheral lesion produces bilateral effects (for review see Koltzenburg et al. 1999). The contralesional effects are usually qualitatively similar to those that occur on the ipsilesional side but often of different magnitude. The putative neuronal mechanisms include the release of trophic signals during stressful conditions, or altered patterns of neuronal activity attributed to the compensatory behavior of the animal. Gacek et al. (1996) demonstrated in cats that contralateral labyrinthectomy was followed by significant morphological changes of the superior vestibular nucleus neurons, including a decrease in cell size. Whether these morphological changes occur in the guinea pig and are linked to the modifications of the membrane properties we observed remains an open question. Supporting the trophic factor hypothesis, other studies by the same authors using knock-out mice suggest that neurotrophin 3 might be involved (for review see Gacek et al. 1998).
Static membrane and firing properties of contralesional MVNn
A striking result of this study is the finding of an increase in the proportion of type B neurons among contralesional MVNn and a concomitant decrease in the proportion of type A neurons (Fig. 8). Because identical criteria were used to classify all contralesional, ipsilesional, or control MVNn, this change could not be the result of differences in the subjective assessment of neurons. The 2 experimenters who performed the recordings obtained similar ratios of type A to type B neurons. The proportion of cells that were silent at rest was similar among type A and B neurons, which indicate that there was no specific silencing of type A MVNn on the contralesional side.
Interestingly, on the ipsilesional side of slices taken after 1 mo of compensation, the proportion of type A neurons among MVNn tended to increase, whereas the remaining type B MVNn were modified and displayed more "type A–like" membrane properties (Beraneck et al. 2003a). In particular, the amplitude of the AHP of ipsilesional type B MVNn was selectively increased. Complementary modifications between the 2 sides were observed for the 2 parameters used to categorize MVNn (Fig. 8): A-like rectification (that characterizes type A MVNn) and double AHP (typical of B MVNn).
Serafin et al. (1991a), who used qualitative criteria to categorize MVNn in slices taken from normal guinea pig, obtained about 30–35% of type A, 50% of type B, and 15–20% of type C MVN neurons. The type C MVNn corresponded to those that showed "intermediate properties." Using the quantitative method of classification, most of the neurons previously considered as type C MVNn had no double AHP and an A-like rectification >0.15 V/s, and thus fell into the type A category. This led to an almost equal division of control MVNn between type A (47%) and B (50%) neurons (Beraneck et al. 2003a). We suggest that the changes in the ratio of type A to B MVNn observed on each side of the brain stem are linked mainly to modifications of the properties of these 20% of "intermediate" neurons that were qualitatively classified by Serafin et al. (1991a) as type C cells. On the ipsilesional side, the shift of MVNn toward more A-like properties did not significantly modify the ratio of type A to type B MVNn because most of the "intermediate" neurons that could have switched categories were already classified as type A MVNn in control slices. On the contralesional side in contrast, the shift of MVNn toward more B-like membrane properties, and in particular the decrease of the A-like rectification, led to a reclassification of the "intermediate" MVNn as type B neurons.
Thus our data suggest that during long-term postlesional plasticity, the intrinsic membrane properties of guinea pig MVNn can change in different ways, leading to a shift of the MVNn population toward either "A-like" (on the ipsilesional side) or "B-like" (on the contralesional side) properties. This is consistent with the "continuum theory" proposed by Du Lac and Lisberger (1995), which states that the membrane properties of MVNn are distributed between those of the 2 canonical A and B neuronal types. This "population shift" in the membrane properties that leads to substantial modifications of the ratio of type A to type B MVNn could depend on the species and/or the exact criteria used to categorize MVNn. In the rat and mouse, for instance (Dutia and Johnston 1998; Him and Dutia 2001; Johnston et al. 1994), contrary to the guinea pig, the type B MVNn make up the vast majority of the population and all show a double AHP during spontaneous firing.
Although long-term vestibular compensation induced a significant depolarization of the average resting membrane potential of ipsilesional MVNn, this effect was slight for contralesional MVNn and was restricted to type B cells (Fig. 8). In the rat (Him and Dutia 2001), an early depolarization of the ipsilesional, deafferented type B MVNn appears within the first week of compensation. This depolarization develops and extends progressively to type A neurons during the first month of compensation, consistent with findings in the guinea pig (see DISCUSSION in Beraneck et al. 2003a). Hence, the average resting membrane potential of both ipsilesional and contralesional type B MVNn appears to be more sensitive to unilateral labyrinthectomy (i.e., more dependent on the permanent activity of the vestibular sensory afferents than the resting potential of type A neurons).
The mean resting activity of contralesional MVNn was not modified compared with control slices, and tended to be less than that on the ipsilesional side (Fig. 8). The absence of a strong decrease in the spontaneous firing rate of the whole sample of MVNn contrasts with that found at the same stage of compensation using extracellular recordings (Vibert et al. 1999b). This discrepancy may be ascribed to the use of sharp electrode intracellular recording techniques, which may provoke an artificial increase of the neuronal firing rate (for a detailed discussion of that point see Beraneck et al. 2003a). However, it is noteworthy that the proportion of neurons silent at rest reached 17.5% among contralesional MVNn, and was thus higher than that obtained in control slices (5%) or on the ipsilesional side (3%). Accordingly, the proportion of contralesional MVNn whose spontaneous firing rate ranged from 10 to 20 spikes/s was half that of normal MVNn. This is consistent with both our extracellular study and that of Darlington et al. (1989), who observed in the guinea pig a marked decrease of the number of spontaneously active neurons on the contralesional side of slices 2 mo after the lesion.
After unilateral labyrinthectomy, the spontaneous discharge of contralesional MVNn recorded in vivo increases, before returning to normal in 1 wk (Ris and Godaux 1998). At the cellular level, the intrinsic depolarization of neurons that develops with time on the side of the lesion might partially sustain the spontaneous discharge that has been restored in deafferented MVNn (Beraneck et al. 2003a; Him and Dutia 2001). The increased proportion of silent and slowly discharging neurons in the MVNn could be the contralesional mirror of this phenomenon. The long-term plasticity of firing properties on the contralesional side might contribute to the decrease of the spontaneous discharge rate of contralesional MVNn observed in vivo, which should act to maintain the new balance of activity that is reached between both vestibular complexes 1 wk after the lesion (Ris and Godaux 1998).
The amplitude of the AHP of the whole population of contralesional MVNn decreased compared with control MVNn (Fig. 8), which should decrease the stability and regularity of their resting discharge. This and the increased sensitivity to membrane potential variations (see following text) should render contralesional MVNn more responsive to synaptic inputs (Babalian et al. 1997). On the ipsilesional side in contrast, there was an increase in the amplitude of the AHP of type B MVNn compared with control MVN neurons, which would contribute to the stability of the spontaneous discharge that is restored in the deafferented MVNn during compensation (for discussion see Beraneck et al. 2003a). This suggests that after 1 mo of compensation, ipsilesional MVNn develop properties that maintain stable tonic activity, whereas contralesional MVNn become more sensitive to modulation of their activity by external inputs.
Dynamic response properties of contralesional MVNn
In contrast with the differential changes observed for the static membrane properties, the contralesional and ipsilesional MVNn showed rather similar changes of their responses to ramps of current compared with control MVNn (Fig. 8). The overshoot, which was higher in type B than in type A MVNn in control slices (Beraneck et al. 2003a; Ris et al. 2001), increased for both types of contralesional as well as ipsilesional MVNn for the ramps delivered during hyperpolarization. For the ramps delivered from rest, the increase of the overshoot occurred mainly in type A MVNn on the ipsilesional side, and was restricted to them on the contralesional side. The increased overshoot was associated with an augmentation of the slope of firing rate increase during the ramps. Thus on both sides, the MVNn recorded after 1 mo of compensation showed a higher sensitivity to large-amplitude current injections. Type A MVNn developed responses to ramp currents that resembled those normally seen in type B neurons (Fig. 5).
Consistent with these observations, we reported an increase, on both sides of the brain stem, in the sensitivity of the firing rate of neurons IF/I to injection of sinusoidal currents (Fig. 8). On the ipsilesional side, this increased sensitivity was associated with a slight, nonsignificant increase in the sensitivity of the discharge of MVNn IF/V to membrane potential. However, on the contralesional side, there was a marked and significant increase in the sensitivity of their discharge to membrane potential IF/V, which was presumably attributable to an increase of their voltage-dependent conductances, as indicated by the decrease in the active impedance V/I.
On both sides, the increased IF/I was associated with a general phase shift of the responses toward greater phase leads and smaller phase lags, also suggesting a major involvement of voltage-dependent conductances (Fig. 8). On the ipsilesional side, this more active behavior might result from the fact that the resting membrane potential of MVNn was more depolarized than in control slices. However, this is not likely on the contralesional side, where only type B neurons were slightly depolarized. The fact that the phase shift was stronger on the contralesional side than on the ipsilesional side correlates with the greater dependency of the firing rate of contralesional neurons on the membrane potential (IF/V).
Although the neuronal properties at rest of contralesional type A MVNn were not different from those of control type A MVNn, their evoked responses were strongly modified. Interestingly, the amplitude of these changes depended on the level of their membrane potential. During steady-state hyperpolarization, contralesional and control type A MVNn showed very similar responses to sinusoidal current injections, whereas the active impedance of contralesional type B MVNn was already lower than normal. At rest or under steady-state depolarization, in the presence of action potentials, the sensitivity of type A contralesional MVNn to applied current was strongly increased and became very similar to that of type B neurons. Thus type A contralesional MVNn displayed enhanced dynamic properties and behaved like the phasic type B MVNn as soon as they discharged spikes. Interestingly, Uno et al. (2003) previously observed that, although the lateral vestibular nucleus neurons had membrane properties similar to a composite of type A and type B MVNn, they showed very different dynamic responses. Thus it appears that subtle changes in the membrane properties of central neurons are sufficient to strongly modify their dynamic responses.
When the amplitude of the firing rate modulation of contralesional MVNn by sinusoidal currents to increasing stimulation frequency reached the modulation limit imposed by their essentially constant spontaneous discharge rate, they tended to synchronize their discharge with high-frequency inputs more than control or ipsilesional MVN neurons. Interestingly, the control type B MVNn also show a better synchronization than that of the control type A MVNn (Ris et al. 2001).
Altogether, these data demonstrate that long-term plasticity after sensory deafferentation can lead to bilateral modifications of the membrane properties of central neurons. Most of the voltage-sensitive conductances expressed by neurons are potentially subject to persistent use-dependent modulation. In a recent review on intrinsic neuronal properties, Zhang and Linden (2003) suggested several specific K+, Ca2+, and Na+ conductances as possible molecular substrates of long-term plasticity. In MVNn, recent studies by Du Lac and colleagues (Nelson et al. 2003; Smith et al. 2002) proposed that the level of intracellular calcium and the calcium-activated potassium conductances are major determinants of the excitability of MVN neurons.
Functional implications of the long-term changes of the membrane properties of contralesional MVNn during vestibular compensation
In vivo experiments on guinea pigs demonstrated that after unilateral labyrinthectomy, there is at first a similar decrease of the vestibulo-ocular responses triggered by head rotations directed toward either the lesioned or intact side (Gilchrist et al. 1998; Vibert et al. 1993). On the contralesional side, the decrease arises from the loss of the disinhibition normally coming from the deafferented side through commissural pathways. In guinea pigs, as in other vertebrates (Broussard et al. 1999b; Gilchrist et al. 1998; Lasker et al. 2000; Murai et al. 2003; Vibert et al. 1993), the vestibulo-ocular reflex triggered by rotations toward the contralesional side largely recovers within several weeks. In contrast with the behavior on the ipsilesional side, the quality of the recovery does not decrease with the amplitude of the movement, and high acceleration impulses directed toward the contralesional side trigger high-frequency responses similar to those obtained in intact animals. Many of the modifications of the membrane properties of contralesional MVNn reported here might be involved in this recovery. The increased sensitivity of the firing rate of neurons to membrane potential modulation and their smaller AHP should lead to more efficient responses of contralesional neurons to synaptic inputs. The increased sensitivity of the firing rate of contralesional MVNn to both ramplike and sinusoidal current stimulation was associated with a greater involvement of active conductances, which should increase both the efficacy and frequency range of the neuronal responses. Altogether, the contralesional MVNn displayed a more phasic behavior than that of control MVNn. The proportion of the more phasic type B neurons among MVNn was increased, and the dynamic properties of the remaining tonic type A MVNn were strongly modified and resembled more those of type B neurons than in control slices. There was also an amplification of the nonlinear properties of MVNn such as the overshoot and synchronization capabilities. Contrary to changes on the ipsilesional side, where the type B MVNn tended to acquire the AHP and response dynamics of type A neurons (see DISCUSSION in Beraneck et al. 2003a), the ability of type B contralesional MVNn to respond to high-frequency, high-amplitude stimuli was not impaired.
In 1999 Minor et al. proposed a mathematical model of the vestibulo-ocular reflex dynamics where the system includes distinct linear and nonlinear pathways that would be mediated by different sensory vestibular afferents and central vestibular neurons. Based on their observations in the monkey and those of Broussard et al. (1999a,b) in the cat, Lasker et al. (2000) suggested that an extension of the nonlinear component of vestibular reflexes was responsible for the increase in contralesional gain after unilateral labyrinthectomy or canal plugging. This should apply to guinea pigs because the dynamics of the vestibulo-ocular reflex and the way they are affected by labyrinthectomy are similar in all m
