Plasticity of spontaneous excitatory and inhibitory synaptic activity in morphologically defined vestibular nuclei neurons during early vestibular compensation

Abstract

After unilateral peripheral vestibular lesions, the brain plasticity underlying early recovery from the static symptoms is not fully understood. Principal cells of the chick tangential nucleus offer a subset of morphologically defined vestibular nuclei neurons to study functional changes after vestibular lesions. Chickens show posture and balance deficits immediately after unilateral vestibular ganglionectomy (UVG), but by 3 days most subjects begin to recover, although some remain uncompensated. With the use of whole cell voltage-clamp, spontaneous excitatory and inhibitory postsynaptic currents (sEPSCs and sIPSCs) and miniature excitatory and inhibitory postsynaptic currents (mEPSCs and mIPSCs) were recorded from principal cells in brain slices 1 and 3 days after UVG. One day after UVG, sEPSC frequency increased on the lesion side and remained elevated at 3 days in uncompensated chickens only. Also by 3 days, sIPSC frequency increased on the lesion side in all operated chickens due to major increases in GABAergic events. Significant change also occurred in decay time of the events. To determine whether fluctuations in frequency and kinetics influenced overall excitatory or inhibitory synaptic drive, synaptic charge transfer was calculated. Principal cells showed significant increase in excitatory synaptic charge transfer only on the lesion side of uncompensated chickens. Thus compensation continues when synaptic charge transfer is in balance bilaterally. Furthermore, excessive excitatory drive in principal cells on the lesion side may prevent vestibular compensation. Altogether, this work is important for it defines the time course and excitatory and inhibitory nature of changing spontaneous synaptic inputs to a morphologically defined subset of vestibular nuclei neurons during critical early stages of recovery after UVG.

normal brain function requires a balance between convergent excitatory and inhibitory inputs onto neurons in the central nervous system, with imbalance characteristic of major neurological disorders (e.g., Naylor 2010; Gajcy et al. 2010). Unilateral vestibular-nerve deafferentation destroys the balance in vestibular nuclei neurons by removing a major excitatory input. The resultant static symptoms include profound deficits in posture, balance, and eye movements, which diminish or disappear in days to a week due to a process called vestibular compensation (for review, see Dutia 2010; Halmagyi et al. 2010). Changes in passive membrane properties, spontaneous and evoked spike firing, and sensitivity to inhibitory neurotransmitters are recorded in vestibular nuclei neurons from brain slices of rat (e.g., Yamanaka et al. 2000; Him and Dutia 2001; Olabi et al. 2009) and guinea pig after unilateral labyrinthectomy (UL; e.g., see Vibert et al. 2000; Beraneck et al. 2003, 2004), and chicken after unilateral vestibular ganglionectomy (UVG; see Shao et al. 2009).

Presently, vestibular compensation is thought to involve plasticity of nonvestibular afferent inputs, with vestibular nuclei neurons a prime target. In addition to vestibular afferents, vestibular nuclei neurons are sites for convergence and integration of excitatory inputs from the spinal cord (Highstein and Holstein 2006) and contralateral vestibular nuclei (Malinvaud et al. 2010), as well as inhibitory inputs from cerebellar neurons (e.g., see Cox and Peusner 1990; Bagnall et al. 2009), contralateral vestibular nuclei (e.g., see Precht et al. 1973), and local interneurons (e.g., see Holstein et al. 1999). After vestibular deafferentation, there is evidence for changes in the excitatory and inhibitory commissural inputs on the lesion and intact sides and excitatory inputs from spinal cord on the lesion side (Vibert et al. 1999; Dieringer 2003). Also, significant numbers of new GABAergic interneurons appear in the cat vestibular nuclei on the lesion side after UVG (Tighilet et al. 2007) but not after bilateral vestibular ganglionectomy performed on the rat (Zheng et al. 2010). Altogether, the sequence of events for changing excitatory and inhibitory inputs to a specific subset of vestibular nuclei neurons has not been established during early stages of vestibular compensation, which is the main objective of this study. Accordingly, here we investigated spontaneous excitatory and inhibitory synaptic activity in vestibular nuclei neurons, which provides a good estimate of overall synaptic inputs from various sources and their influence on neuronal responses after lesions, which are key to understanding the functional state of the network (Hanganu et al. 2001; Greenhill and Jones 2007).

To characterize changes in signal processing after lesions, it is helpful to select a neuron subset with a known output target, since vestibular nuclei neurons with different axonal projections show diverse spike discharge patterns (Goldberg 2000; Sekirnjak and du Lac 2006; Gottesman-Davis et al. 2011; Kolkman et al. 2011) and neurotransmitter phenotypes, with glutamate, glycine, and GABA predominating (Spencer et al. 1989; Takazawa et al. 2004; Gittis and du Lac 2007). Furthermore, in other sensory systems, the response of morphologically defined neuron subsets to deafferentation depends on their role in the network (e.g., Francis and Manis 2000; Li et al. 2009; Zhou et al. 2009). Accordingly, here we propose that the vestibular nuclei neuron subsets selected for study should be distinguished by key morphological features, including axonal output, neurotransmitter phenotype, and/or neuron morphology and that these subsets should be analyzed as separate independent groups after peripheral vestibular lesions. This approach has been avoided in most vestibular compensation studies due to the lack of architectural organization in the vestibular nuclei, the rather large catalog of functional subsets of vestibular nuclei neurons, and the absence of distinctive neuron morphology for vestibular nuclei neurons linking them to axonal outputs, as described for cortical pyramidal cells (Romand et al. 2011) and cochlear nuclei neurons (for review, see Hackney 1987).

Unlike most vestibular nuclei neurons, principal cells of the chick tangential nucleus are morphologically striking in simple stains such as MAP2 (Popratiloff and Peusner 2007). Principal cells are characterized by large oval, glutamatergic cell bodies (30–35 μm) aligned in rows between the incoming primary vestibular fibers in the lateral medulla oblongata (Peusner and Morest 1977; Peusner and Giaume 1997; Popratiloff and Peusner 2011). Moreover, the distinctive morphology of principal cells is retained after UVG (Aldrich and Peusner 2002). The axons of most principal cell course medially without producing collaterals in the tangential nucleus and enter the contralateral medial longitudinal fasciculus to contact neurons in the oculomotor, trochlear, or abducens nuclei (Wold 1978; Labandeira-Garcia et al. 1989; Petursdottir 1990; Evinger and Erichsen 1986; Cox and Peusner 1990; Gottesman-Davis and Peusner 2010). Axons of principal cells projecting to the abducens nucleus may produce collaterals to cervical spinal cord, while other principal cells send axons directly to cervical spinal cord (Cox and Peusner 1990). Thus all principal cells are morphologically distinctive, glutamatergic vestibular reflex projection neurons (Popratiloff and Peusner 2011), which may differ in their projections within the vestibuloocular, vestibulocollic, and/or vestibuloocular collic pathways.

The bipedal chicken offers an important model to study vestibular compensation, since, like patients, most chickens start to recover from the static symptoms shortly after UVG (3 days; Aldrich and Peusner 2002), while a significant fraction fails to compensate (Shao et al. 2009). In fact, ∼20% of patients fail to recover after surgical unilateral vestibular deafferentation, experiencing persistent postural imbalance and dynamic reflex deficits (for review, see Badke et al. 2002; Halmagyi et al. 2010). At present, it is uncertain what factors underlie the failure to compensate. Accordingly, here we investigated spontaneous synaptic activity in control, operated chickens 1 day after UVG, chickens starting to compensate by 3 days, and those that failed to compensate by 3 days (uncompensated chickens). Taken together, the present study demonstrates that UVG affects both the excitatory and inhibitory inputs on a morphologically defined subset of vestibular nuclei neurons at different times during recovery. Significant fluctuations occurred in frequency and kinetics of these spontaneous synaptic events. By calculating the synaptic charge transfer, defined as the integration of all deviations from baseline due to spontaneous synaptic events for a set period, a major change was found only in the excitatory synaptic drive in principal cells on the lesion side of uncompensated chickens 3 days after UVG. Thus compensation may be linked to balancing excitatory and inhibitory synaptic drive in vestibular nuclei neurons bilaterally, with failure to balance associated with uncompensated subjects. Moreover, here we defined the time course and nature of the synaptic inputs to these morphologically defined vestibular nuclei neurons during early recovery.

MATERIALS AND METHODS

Experimental animals

Cold, freshly fertilized, white Leghorn chick embryo eggs (Gallus gallus) were purchased from CBT Farm (Chesterton, MD) or Charles River Spafas (North Franklin, CT) and maintained until hatching in an egg incubator equipped with circulating air, egg rotation unit, and temperature and humidity controls (model 1502; G.Q.F. Manufacturing, Savannah, GA). Chickens were identified as 1-day-old hatchlings (H1) 24 h after hatching. Hatchlings were housed in cages equipped with heating lamps at the University's Animal Research Facility until the appropriate ages for the experiments (H4–H7). Control and operated chickens were housed under identical conditions. Animal protocols were approved by the Institution Animal Care and Use Committee of the George Washington University. The experiments conformed to the guidelines for the care and use of animals in research (National Research Council 2003), and considerable effort was made to minimize the number of animals used and their suffering.

UVG

UVG was performed on H4 chickens that were killed 1 or 3 days later. The surgical protocol was adapted from previous studies (Aldrich and Peusner 2002; Pollack et al. 2004; Shao et al. 2009). After general anesthesia was obtained using ketamine (100 mg/kg; Fort Dodge Animal Health, Fort Dodge, IA) and xylazine (20 mg/kg; Akorn, Decatur, IL), an incision was made in the skin on the left side of the head posterior to the external auditory meatus. The exoccipital and squamosal bones, lateral part of the bony vestibule, membranous labyrinths, and medial wall of the bony vestibule were removed sequentially to expose the dura mater overlying the vestibular ganglion. After the dura mater was cut, both anterior and posterior parts of the vestibular ganglion and auditory nerve were exposed. The vestibular nerve was cut between the vestibular ganglion and lateral brain surface using a curved, flame-sharpened tungsten wire. The survival rate for UVG surgery was 95%. On the first day after surgery, the operated chickens received an intramuscular injection of 0.25 ml penicillin-G prophylactically (Bimeda, Le Sueur, MN). Starting on the second day, all operated chickens received the antibiotic Clavamox (Pfizer, New York, NY; 62.5 mg/ml, 1 ml) twice a day via gavage. Dehydration was treated by subcutaneous injection of sterile, lactated Ringer solution (Braun Medical, Irvine, CA), starting 1–2 h after surgery when the animals recovered consciousness. The volume of replacement fluid was determined by the amount of weight loss (1 ml Ringer solution/1 g weight loss, injected under the skin on the lower extremity; National Research Council 2003). Also, on the second day after UVG, chickens that were unable to feed themselves were administered by gavage baby rice cereal dissolved in water (Gerber, Fremont, MI) every 3 h, four times a day until killed.

One day after UVG, all operated chickens showed major postural deficits, including severe head tilt towards the lesion side, flexion of the lower extremity on the lesion side, extension of the lower extremity on the intact side, and inability to stand, feed, or drink independently. However, by 3 days after UVG, most operated chickens could stand, feed, and drink independently, so they were considered “compensating chickens,” whereas other operated chickens could not perform these activities so they were called “uncompensated chickens” (see Shao et al. 2009). Indeed, after peripheral vestibular surgery, patients may fail to compensate (Halmagyi et al. 2010). Compensating and uncompensated chickens were indistinguishable 1 day after UVG. However, no brain damage or hemorrhage into the brainstem was observed in the compensating or uncompensated chickens.

Brain Slice Preparation and Solutions

Chickens were decapitated and the part of the cranium containing the brainstem and the cerebellum was pinned to sylgard (World Precision Instruments, Sarasota, FL) that was placed on the bottom of a dissecting dish which was filled with oxygenated, sucrose-containing, artificial cerebrospinal fluid (ACSF) at 1–2°C. Under a dissecting microscope, the cerebellum, choroid plexus, periotic capsule, and vestibular ganglia were removed from the brainstem, which was glued (Cyanoacrylic glue, Crazy Glue; Elmer's Products, Columbus, OH) to a vibratome chuck, placed in a bath containing sucrose ACSF (1–2°C) and mounted on a vibratome (VT1000S; Leica Instruments, Bannockburn, IL). Brain slices were cut at 240-μm thickness using a sapphire knife (Delaware Diamond Knives, Wilmington, DE). Those slices containing the tangential nucleus were placed in heated ACSF (37°C) for ∼40 min and then cooled to room temperature for ≥20 min before being placed in the recording chamber. Sucrose ACSF contained the following (in mM): 210 sucrose, 2.5 KCl, 7 MgSO₄, 1.5 NaH₂PO₄, 1 CaCl₂, 26 NaHCO₃, and 10 d-glucose. ACSF contained the following (in mM): 126 NaCl, 3.5 KCl, 1.2 NaH₂PO₄, 1.3 MgCl₂, 2 CaCl₂, 26 NaHCO₃, and 10 d-glucose. The pH of sucrose ACSF and ACSF was 7.2–7.4 after saturation with 95% O₂-5% CO₂ at room temperature. The osmolarity of both solutions was 310–320 mosM.

All drugs, except for 6-cyano-7-nitroquinoxaline-2, 3-dione (CNQX), were dissolved in distilled water and then added to the ACSF inflow line to achieve the final concentration. Tetrodotoxin (1 μM; Alomone Labs, Jerusalem, Israel) blocked voltage-dependent sodium channels and revealed miniature synaptic events. Bicuculline methochloride (10 μM; Tocris, St. Louis, MO) blocked GABA_A receptor-mediated events, and strychnine (1 μM; Sigma) blocked glycine receptor-mediated events. CNQX (RBI, Natick, MA) was dissolved in dimethyl sulfoxide (20 mM; Fisher Scientific, Fair Lawn, NJ) and then added to the ACSF inflow line to achieve the final concentration (10 μM). CNQX blocked AMPA/KA receptor-mediated events.

Electrophysiological Techniques

The recording chamber (180 μl; Warner Instruments, Hamden, CT) was superfused with preheated ACSF at a rate of 2–3 ml/min and maintained at 30°C using a temperature controller (TC324B; Warner Instruments). Brain slices were observed on a fixed-stage, upright light microscope (Zeiss Axioskop FS-1, Jena, Germany), equipped with differential interference contrast optics and a ×40 water-immersion objective (NA, 0.75). Visualization of the recorded neuron and recording pipet was achieved using an infrared light source (filter, 770 nm) detected by an infrared-sensitive tube camera (Vidicon C2400–01; Hamamatsu, Hamamatsu City, Japan) and observed on a video monitor (Sony, Tokyo, Japan). A ×4 lens was inserted between the microscope and camera, and image contrast and shading were adjusted using a camera controller (C2400; Hamamatsu).

Recording pipets were moved using a piezoelectric manipulator (PCS-5000; Burleigh Instruments, Fishers, NY). Recording pipets (2–5 MΩ) were pulled from thin-walled (1.5-mm outer diameter; 1.12-mm inner diameter) borosilicate glass tubing (World Precision Instruments) on a Brown/Flaming horizontal pipet puller (P-87; Sutter Instruments, Novato, CA). Cesium gluconate pipet solution contained the following (in mM): 135 Cs-gluconate, 1 EGTA, 10 HEPES, 0.1 CaCl₂, and 2 MgCl₂. The pH was adjusted to 7.2 with CsOH (Shao et al. 2003), and the osmolarity was adjusted to 270–290 mosM. Biocytin (0.5%; Sigma) was dissolved in the pipet solution weekly and kept frozen between experiments.

Electrophysiological Data Acquisition and Analysis

Electrophysiological signals were acquired and processed using Axopatch-200B amplifier or Multiclamp 700B (Molecular Devices, Foster City, CA) and Clampex 9.2 or 10.2 computer program (Molecular Devices). Signals were digitized at 10 kHz with Digidata 1320 or Digidata 1440 (Molecular Devices) and low-pass filtered at 2 kHz.

With the use of Cs-gluconate pipet solution, spontaneous spike activity was measured in the cell-attached configuration before the whole-cell voltage-clamp configuration was established, since this is an effective and efficient approach to acquire data on spontaneous spike activity before obtaining the whole cell mode (Perkins 2006). Spontaneous spike discharge rate was expressed as the inverse of the mean interspike interval and obtained by averaging the activity from spontaneous spike firing cells only in each experimental group. After obtaining the whole-cell configuration, input resistance and series resistance were determined by injecting +5-mV pulses using the seal test function in voltage-clamp mode. During the recordings, if the input resistance changed >25%, the neuron was excluded from the study. Series resistance compensation was set at 80%. Based on reversal potential, excitatory postsynaptic currents (EPSCs) were recorded at a holding potential of −60 mV, and inhibitory postsynaptic currents (IPSCs) were recorded at +10 mV (Shao et al. 2003). No corrections were made for the voltage offset observed after withdrawing the pipet from the cell (no more than ±3 mV), or liquid junction potential of the pipet solution (+3 mV). Synaptic currents were analyzed offline using Minianalysis 6.0.3 program (Synaptosoft, Decatur, GA). The threshold for acceptable events was set at 3 × root mean square value of the baseline noise recorded at −60 mV (2–4 pA) and at +10 mV (4–7 pA). Postsynaptic currents were analyzed for frequency and kinetics, including peak amplitude, rise time (from 10 to 90% peak current), and decay time (from 90 to 37% amplitude return to baseline). Synaptic charge transfer was calculated by integrating the area of miniature (m)EPSCs or mIPSCs waveforms for 1 min (Gao et al. 2011).

All data are presented as means ± SE. Two group comparisons were performed using a nonparametric Mann-Whitney test, and multiple group comparisons were performed using a nonparametric one-way ANOVA (Kruskal-Wallis) test with Dunn's posttest. Statistical significance was set at P < 0.05. In all experiments, the data from H5 and H7 controls were not significantly different, so they were pooled and presented together.

Visualization of Principal Cells Injected with Biocytin During Recordings in Brain Slices

All recorded neurons were filled with biocytin passively during the recordings. After the recordings, the brain slices were processed for visualization of the biocytin-injected principal cells (Shao et al. 2003). Brain slices were fixed in 4% paraformaldehyde in 0.1 M phosphate buffer (overnight, refrigerated), rinsed with 0.1 M PBS containing 0.1% Triton X-100 (PBS-T), incubated in 50% ethanol in PBS-T, and placed in streptavidin Alexa Fluor 488 or 647 (1:200) (Molecular Probes, Eugene, OR) in PBS-T (overnight, room temperature). After being rinsed in PBS, the brain slices were mounted on gelatin-coated slides, and coverslipped with Fluoromount G (Electron Microscopy Sciences, Hatfield, PA) and refrigerated.

Biocytin-labeled neurons were visualized on a Zeiss LSM 710 confocal laser scanning microscope (Zeiss Instruments) equipped with ×10 (NA, 0.45) and ×20 PlanApo (NA, 0.8) objectives. Excitation for Alexa Fluor 488 was achieved using an Argon laser emitting at 494–600 nm, while excitation for Alexa Fluor 647 was achieved using a HeNe633 laser emitting at 637–755 nm. Recorded neurons were confirmed as principal cells by evaluating somatic size and shape and its position within the tangential nucleus (Fig. 1).

Fig. 1.

Biocytin-injected and recorded principal cells in brain slices. Top: principal cell located in the posterior-to-central tangential nucleus from H5 control. Center: principal cell located in the central tangential nucleus on the lesion side 1 day after unilateral vestibular ganglionectomy (UVG). Bottom: principal cell located in the anterior-to-central tangential nucleus on the intact side 1 day after UVG. Arrows point to the axon of the principal cell. Dotted line indicates the surface of the brainstem. Dashed lines indicate the boundaries of the tangential nucleus, when present in the image. RB, restiform body; D, dorsal; V, ventral; M, medial; L, lateral.

RESULTS

The experiments were performed on a total of 120 chickens in 4 experimental groups: normal controls (n = 29), operated chickens 1 day after UVG (n = 31), operated chickens 3 days after UVG that started to compensate (n = 33), and operated chickens which failed to show signs of compensation by 3 days (uncompensated) (n = 27). Fifty-five percent (33/60) of the operated chickens started to compensate by 3 days after UVG, in keeping with the recovery rate previously reported (Shao et al. 2009). Whole cell voltage-clamp recordings were obtained from a total of 179 morphologically identified principal cells, including those from controls (n = 38), the lesion (n = 25) and intact side (n = 24) of operated chickens 1 day after UVG, the lesion (n = 26) and intact side (n = 27) of compensating chickens 3 days after UVG, and the lesion (n = 21) and intact side (n = 18) of uncompensated chickens 3 days after UVG. The morphology of the biocytin-injected and recorded principal cells from controls and operated chickens was similar to those described in a previous study (Fig. 1) (Shao et al. 2009).

Spontaneous Spike Activity in Cell-Attached Configuration

In controls, 47% of the principal cells generated spontaneous spike activity (Table 1). One day after UVG, the percentage of spontaneous spike firing principal cells decreased on the intact side compared with controls. By 3 days after UVG, the percentage of spontaneous spike firing principal cells had decreased significantly on the intact side compared with the lesion side and controls in all the operated chickens. In controls, the mean spontaneous spike discharge rate was 20 ± 3.2 spikes/s, with a range of 7.9–50.0 spikes/s. The spike discharge rate did not change in principal cells after UVG (Table 1).

Table 1.

Spontaneous spike activity and spontaneous synaptic events in principal cells before and after UVG

	Control	One Day After UVG		Compensating 3 Days After UVG		Uncompensated 3 Days After UVG
		Lesion side	Intact side	Lesion side	Intact side	Lesion side	Intact side
SSF
%SSF cells	47%a,b	40%	17%a	58%b	11%b	52%b	6%b
	(18/38)	(10/25)	(4/24)	(15/26)	(3/27)	(11/21)	(1/18)
Firing rate, spikes/s	20 ± 3.2	21.4 ± 5	29 ± 10	29 ± 5	20.7 ± 4.7	16.3 ± 3.7	16.3
n	18	10	4	15	3	11	1
Input resistance, MΩ	149 ± 10c	177 ± 15	126 ± 13	267 ± 20c	138 ± 13c	259 ± 23c	122 ± 16c
n	38	25	24	26	27	21	18
sEPSCs
Frequency, Hz	1.7 ± 0.4d	4.8 ± 1.1d	2.3 ± 0.9d	1.9 ± 0.2	2.6 ± 0.5	5.3 ± 1.6d	1.7 ± 0.3d
n	31	24	21	20	26	15	17
sIPSCs
Frequency, Hz	6.4 ± 2.4e	8 ± 2.5	6.4 ± 2.9	19.4 ± 6.5e	4 ± 1e	10.1 ± 2.1f	2.4 ± 0.6f
n	32	23	24	26	24	20	18

Data are means ± SE. Significant difference was set at P < 0.05. Note, neurons were more stable at +10 mV than at −60 mV with Cs-gluconate pipet solution, resulting in more neurons recorded at the former than the latter voltage. ^aPercent spontaneous spike firing (%SSF) cells was significantly lower on the intact side 1 day after unilateral vestibular ganglionectomy (UVG) compared with controls. ^b%SSF cells was significantly lower on the intact side compared with the lesion side and controls in operated chickens 3 days after UVG. ^cInput resistance was significantly higher on the lesion side compared with the intact side and controls in operated chickens 3 days after UVG. ^dSpontaneous excitatory postsynaptic current (sEPSC) frequency was significantly higher on the lesion side compared with the intact side and controls in operated chickens 1 day after UVG and uncompensated chickens 3 days after UVG. ^eSpontaneous inhibitory postsynaptic current (sIPSC) frequency was significantly higher on the lesion side compared with the intact side and controls in compensating chickens 3 days after UVG. ^fsIPSC frequency was significantly higher on the lesion side compared with the intact side in uncompensated chickens 3 days after UVG.

Input Resistance

The input resistance of the principal cells from controls was 149 ± 10 MΩ and increased significantly on the lesion side compared with the intact side and controls in all the operated animals 3 days after UVG (Table 1). Similar changes in input resistance were reported in another study of the principal cells performed before and after UVG (Shao et al. 2009).

Frequency of Spontaneous Synaptic Events

Spontaneous (s)EPSCs and sIPSCs were recorded in the same principal cell at different holding potentials without adding drugs. At −60 mV, all the currents generated in the principal cells from all experimental groups were inward currents that were abolished on exposure to CNQX (n = 17), suggesting that they were AMPA/kainate receptor-mediated events. This result suggests that after loss of the glutamatergic primary vestibular fiber inputs, no new nonglutamatergic, excitatory inputs appeared (e.g., nicotinic receptor-mediated). At +10 mV, all the currents generated in the principal cells were outward currents, which were abolished completely by bicuculline and strychnine (n = 140), indicating that they were GABA_A- and/or glycine-receptor-mediated events. No significant difference was detected in either sEPSC or sIPSC frequency for spontaneous spike firing or silent principal cells from the same experimental group, suggesting that the ability of the principal cells to generate spontaneous spike activity was not related to the frequency of these spontaneous events. This conclusion is supported by finding that on exposure to a cocktail of excitatory and inhibitory neurotransmitter antagonists, principal cells in H5 controls do not change their spontaneous spike firing rates (Shao et al. 2006).

In all experimental groups, sEPSC frequency varied appreciably. For example, in controls the mean sEPSC frequency of the principal cells was 1.7 ± 0.4 Hz with a range of 0–13.1 Hz (Fig. 2A₁; see Fig. 4A₁). sEPSC frequency increased significantly in the principal cells on the lesion side compared with the intact side and controls in operated chickens 1 day after UVG (Fig. 2B₁; see Fig. 4A₁) and in uncompensated chickens 3 days after UVG (Fig. 2C₁; see Fig. 4A₁; Table 1). These findings are consistent with those reported in a previous study of the principal cells before and after UVG using KMeSO₄ pipet solution (Shao et al. 2009). Thus an important indicator of compensation from the static deficits may be the reappearance of balanced sEPSC frequencies in principal cells on the lesion and intact sides 3 days after UVG. The mean frequency of sIPSCs in principal cells from controls was 6.4 ± 2.4 Hz (Figs. 3A₁ and 4A₂). Three days after UVG, sIPSC frequency increased significantly in principal cells on the lesion side compared with the intact side and controls in compensating chickens (Figs. 3B₁ and 4A₂). In uncompensated chickens, sIPSC frequency also increased significantly in principal cells on the lesion side compared with the intact side but not compared with controls (Fig. 3C₁ and 4A₂) (Table 1). Thus, in all operated chickens 3 days after UVG, the principal cells showed a significant imbalance in the frequency of inhibitory events bilaterally to a greater extent in the compensating chickens.

Fig. 2.

Excitatory synaptic events in principal cells before and after UVG. Spontaneous excitatory postsynaptic currents (sEPSCs) and miniature (m)EPSCs frequencies were higher on the lesion side one (B₁ and B₂) and 3 days after UVG in the uncompensated chickens (C₁ and C₂) compared with controls (A₁ and A₂). D: superimposed and scaled mEPSC waveforms show that the principal cells had slower mEPSC decay times on the lesion side compared with the intact side in all the operated animals 3 days after UVG. For each experimental group, the traces were taken from the same principal cell (n, number of events averaged to obtain the waveform).

Fig. 4.

Spontaneous and miniature synaptic events in principal cells before and after UVG. A₁ and B₁: sEPSC and mEPSC frequencies increased significantly in principal cells on the lesion side 1 day after UVG compared with the intact side and controls and remained elevated only in the uncompensated chickens 3 days after UVG. A₂ and B₂: sIPSC and mIPSC frequencies increased significantly in principal cells on the lesion side compared with the intact side in all operated animals 3 days after UVG, with a greater increase in the compensating chickens. *Significant difference between 2 columns at P < 0.05.

Fig. 3.

Inhibitory synaptic events in principal cells before and after UVG. Spontaneous inhibitory postsynaptic currents (sIPSCs) and miniature (m)IPSC frequencies were higher on the lesion side in compensating (B₁ and B₂) and uncompensated chickens (C₁ and C₂) 3 days after UVG compared with controls (A₁ and A₂). D: superimposed and scaled mIPSC waveforms show that principal cells have faster mIPSC decay times on the lesion side compared with the intact side in all operated animals 3 days after UVG. For each experimental group, the traces were recorded from the same principal cell (n, number of events averaged to obtain the waveform).

Frequency of Miniature Synaptic Events

While sEPSC and sIPSC frequencies depend on the number and firing rate of presynaptic neuron cell bodies retained in the brain slice, the frequency of miniature synaptic events depends mainly on the number of presynaptic terminals contacting the recorded neuron and the probability of neurotransmitter release at each terminal. When tetrodotoxin was added to reveal the action potential-independent miniature events in the principal cells, ∼57% of sEPSCs were mEPSCs and 74% of sIPSCs were mIPSCs in principal cells from controls. These percentages did not change after UVG (Table 2). However, in all experimental groups, the percentage of mEPSCs in principal cells was significantly lower than the percentage of mIPSCs, indicating there were either more functionally active, presynaptic excitatory neurons than inhibitory neurons contacting the principal cells in the brain slices, or that the excitatory neurons fired spontaneous spikes at higher rates, and/or the probability of vesicular release at a single synapse was higher at excitatory than at inhibitory synapses.

Table 2.

Miniature synaptic events in principal cells before and after UVG

		One Day After UVG		Compensating 3 Days After UVG		Uncompensated 3 Days After UVG
	Control	Lesion side	Intact side	Lesion side	Intact side	Lesion side	Intact side
mEPSCs
%mEPSC	57%	44%	55%	54%	46%	47%	58%
Frequency, Hz	1 ± 0.3a	2.5 ± 0.7a	0.8 ± 0.1a	1 ± 0.2	1 ± 0.2	3.4 ± 1.2a	0.9 ± 0.2a
Amplitude, pA	15 ± 1.4	15 ± 1.3	15.1 ± 0.9	15 ± 1.1	16.3 ± 1	18.1 ± 1.4	16.1 ± 1.3
Rise time, ms	0.35 ± 0.02	0.38 ± 0.02	0.33 ± 0.02	0.36 ± 0.02	0.31 ± 0.02	0.34 ± 0.02	0.31 ± 0.03
Decay time, ms	1.5 ± 0.3c	1.5 ± 0.2	1.3 ± 0.2	1.6 ± 0.2b	1.1 ± 0.1b	1.2 ± 0.1b	0.8 ± 0.1bc
Synaptic charge transfer, pA · ms	1,137 ± 297	3,853 ± 1,395	878 ± 145	1,608 ± 366	1,219 ± 289	2,665 ± 895d	720 ± 138d
n	23	20	16	19	19	15	10
mIPSCs
%mIPSC	74%	65%	72%	69%	67%	70%	74%
Frequency, Hz	4.2 ± 1.7e	5.1 ± 1.9	5.7 ± 2.7	12.2 ± 2.5e	2.9 ± 0.9e	6.1 ± 1.9f	1.8 ± 0.5f
Amplitude, pA	46 ± 4.6g	39.8 ± 3.9	42.7 ± 6.1	46.3 ± 8.9	54.1 ± 6	46.3 ± 4.6	70.9 ± 9.3g
Rise time, ms	0.79 ± 0.03	0.82 ± 0.05	0.8 ± 0.03	0.75 ± 0.04	0.83 ± 0.05	0.82 ± 0.05	0.71 ± 0.04
Decay time, ms	8.8 ± 1.0	7.5 ± 1	8.5 ± 1.0	6.2 ± 0.8h	10.3 ± 1.7h	6.2 ± 0.7h	10.9 ± 1.6h
Synaptic charge transfer, pA · ms	28,226 ± 5,011	53,765 ± 20,370	48,765 ± 19,122	184,946 ± 111,922	120,685 ± 84,107	54,615 ± 14,344	36,131 ± 16,455
n	32	21	23	25	23	18	16
GABAergic mIPSCs
Frequency, Hz	3.1 ± 1.5	1.7 ± 0.7	3.3 ± 2.8	11.9 ± 4.3i	1.8 ± 0.7i	4.1 ± 1.6	1.4 ± 0.9
Amplitude, pA	24.8 ± 2.3	24.6 ± 2.9	26.9 ± 2.2	29.1 ± 3.2	24.4 ± 1.5	25.4 ± 3.5	26.6 ± 3.5
Rise time, ms	0.87 ± 0.06	0.93 ± 0.09	0.8 ± 0.09	0.74 ± 0.06j	1 ± 0.06j	0.8 ± 0.04	0.84 ± 0.11
Decay time, ms	7.0 ± 0.7k	6.8 ± 0.5	7.3 ± 1.0	5.1 ± 0.8h	8.8 ± 1.5h	6.2 ± 0.7h	10.4 ± 1.6h,k
n	13	12	10	12	10	8	5
Glycinergic mIPSCs
Frequency, Hz	0.6 ± 0.1	1 ± 0.3	0.6 ± 0.2	0.8 ± 0.3	0.8 ± 0.4	0.9 ± 0.3	1 ± 0.3
Amplitude, pA	65.7 ± 9.3	46.8 ± 7	78.6 ± 13.8	44.9 ± 5.2l	67.2 ± 6.8l	54.2 ± 7	80 ± 13.2
Rise time, ms	0.8 ± 0.06	0.81 ± 0.05	0.93 ± 0.12	0.83 ± 0.09	0.82 ± 0.07	0.97 ± 0.09	0.77 ± 0.05
Decay time, ms	10.5 ± 2.1	5.8 ± 1.2m	14.1 ± 3.6m	5.7 ± 0.9h	10.3 ± 2.6h	6.0 ± 1.1h	12.4 ± 2.2h
n	13	9	9	11	11	9	9

Data are means ± SE. Significant difference was set at P < 0.05.

^aMiniature (m)EPSC frequency was significantly higher on the lesion side compared with the intact side and controls in operated chickens 1 day after UVG and uncompensated chickens 3 days after UVG.

^bmEPSC decay time was significantly slower on the lesion side compared with the intact side in all operated chickens 3 days after UVG.

^cmEPSC decay time was significantly faster on the intact side in uncompensated chickens 3 days after UVG compared with controls.

^dSynaptic charge transfer for mEPSCs was significantly larger on the lesion side compared with the intact side in uncompensated chickens 3 days after UVG.

^emIPSC frequency increased significantly on the lesion side compared with the intact side and controls in compensating chickens 3 days after UVG.

^fmIPSC frequency increased significantly on the lesion side compared with the intact side in uncompensated chickens 3 days after UVG.

^gmIPSC amplitude was significantly larger on the intact side in uncompensated chickens 3 days after UVG compared with controls.

^hmIPSC, GABAergic, and glycinergic mIPSC decay time was significantly faster on the lesion side compared with the intact side in all operated chickens 3 days after UVG.

ⁱGABAergic mIPSC frequency increased significantly on the lesion side compared with the intact side in compensating chickens 3 days after UVG.

^jGABAergic mIPSC rise time was significantly faster on the lesion side compared with the intact side in compensating chickens 3 days after UVG.

^kGABAergic mIPSC decay time was significantly slower on the intact side of uncompensated chickens 3 days after UVG compared with control.

^lGlycinergic mIPSC amplitude was significantly smaller on the lesion side compared with the intact side in compensating chickens 3 days after UVG.

^mGlycinergic mIPSC decay time was significantly faster on the lesion compared with intact side in operated chickens 1 day after UVG.

Since the percentage of excitatory and inhibitory miniature synaptic events relative to spontaneous synaptic events did not change in principal cells from any experimental group, the frequency of miniature events followed the same pattern described for spontaneous excitatory (Figs. 2, A₂–C₂, and 4B_1; Table 2) and spontaneous inhibitory (Figs. 3, A₂–C₂, and 4B_2; Table 2) synaptic activities.

Frequency of GABAergic and Glycinergic mIPSCs

The frequency and percentage of GABAergic and glycinergic events were calculated after adding strychnine and bicuculline, respectively. The frequency of GABAergic mIPSCs was 3.1 ± 1.5 Hz in controls and increased significantly in the principal cells on the lesion side compared with the intact side in the compensating chickens 3 days after UVG (Fig. 5A; Table 2). However, GABAergic mIPSCs did not change in other experimental groups. Moreover, glycinergic mIPSC frequency was similar in the principal cells from all experimental groups (Fig. 5B; Table 2). Thus increased mIPSC frequency in principal cells on the lesion side of compensating chickens 3 days after UVG was due to increased GABAergic mIPSC frequency. In controls, GABAergic mIPSCs composed 63% of the synaptic events in the principal cells, while glycinergic mIPSCs composed 39%. Similar percentages of GABA and glycine mIPSCs were recorded in principal cells from all experimental groups (Fig. 5, C and D), with the exception of the principal cells on the intact side of the uncompensated chickens, where a large subset of principal cells (6/14) exhibited only glycinergic inhibitory events (Fig. 5D, grey bar). In fact, a few random principal cells in other experimental groups diverged from the general pattern and exhibited purely GABAergic inhibitory events (12/127).

Fig. 5.

GABAergic and glycinergic mIPSCs in principal cells before and after UVG. A: GABAergic mIPSC frequency increased significantly in principal cells on the lesion side of compensating chickens 3 days after UVG compared with the intact side. B: glycinergic mIPSC frequency did not change significantly in principal cells from all experimental groups. C: percentage of GABAergic mIPSCs was similar in principal cells from all experimental groups. D: in contrast, many principal cells on the intact side had purely glycinergic mIPSCs in uncompensated chickens 3 days after UVG (grey bar; n = 6). *Significant difference between 2 columns at P < 0.05.

Kinetics of Miniature Events

When mEPSC kinetics were compared between controls and the different experimental groups, the amplitude and rise time did not change significantly before or after UVG, but the decay time was significantly slower in those principal cells on the lesion side compared with the intact side in both compensating and uncompensated chickens 3 days after UVG (Figs. 2D and 6, A and B; Table 2). In addition, mEPSC decay time was significantly faster in principal cells on the intact side of the uncompensated chickens 3 days after UVG compared with controls.

Fig. 6.

mEPSC and mIPSC kinetics in principal cells before and after UVG. A and B: mEPSC decay time was significantly slower in principal cells on the lesion side compared with intact side in all experimental animals 3 days after UVG. C and D: GABAergic mIPSC decay time was significantly faster in principal cells on the lesion side compared with the intact side in all experimental animals 3 days after UVG. E and F: glycinergic mIPSC decay time was significantly faster in principal cells on the lesion side compared with the intact side in all experimental animals 3 days after UVG.

Concerning mIPSC kinetics, the amplitude was significantly larger in principal cells on the intact side of the uncompensated chickens 3 days after UVG compared with controls (Table 2), which may be due to the larger percentage and larger amplitude of glycinergic events composing the mIPSCs in these principal cells. (see below). On the lesion side, principal cells had faster mIPSC decay time compared with those on the intact side in both compensating and uncompensated chickens 3 days after UVG (Fig. 3D; Table 2).

In all experimental groups, GABAergic mIPSCs showed significantly smaller amplitudes compared with the glycinergic mIPSCs (Table 2). In controls, rise time and decay time were not significantly different between GABAergic and glycinergic mIPSCs, although the GABAergic mIPSCs tended to show faster decay times. Indeed, after UVG, GABAergic mIPSCs displayed faster decay times in the principal cells on the lesion side of all experimental animals 3 days after UVG compared with the intact side (Fig. 6, C and D). In addition, the decay time of GABAergic mIPSC was significantly slower in principal cells on the intact side of uncompensated chickens 3 days after UVG than in controls (Table 2).

Glycinergic mIPSCs had smaller amplitudes in the principal cells on the lesion side compared with those recorded on the intact side in the compensating chickens 3 days after UVG (Table 2). Similar to GABAergic mIPSCs, glycinergic mIPSCs displayed faster decay times in principal cells on the lesion side in all experimental animals 3 days after UVG compared with the intact side (Fig. 6, E and F). Moreover, the decay time of glycinergic mIPSCs first became significantly faster in principal cells on the lesion compared with the intact side 1 day after UVG (Table 2) and thus represented the first evidence of a change in the postsynaptic response to a neurotransmitter. Thus, overall decrease in mIPSC decay time in principal cells on the lesion side compared with the intact side 3 days after UVG was due to modifications in both GABAergic and glycinergic kinetics in the same direction.

Synaptic Charge Transfer of Miniature Events

Since major fluctuations in the frequency and decay time of synaptic events were recorded from principal cells in different experimental groups at various times after UVG, the question was raised as to whether these fluctuations changed the overall excitability. For example, the effect of significantly increased excitatory frequency could be countered by faster decay time, but to what extent? To address this issue, synaptic charge transfer was calculated by integrating all the spontaneous synaptic events from baseline, with excitatory and inhibitory events calculated independently in each experimental group. This measurement is a good indicator of overall neuron excitability (Dani et al. 2005; Han and Stevens 2009; Gao et al. 2011). At 1 day after UVG, mEPSC synaptic charge transfer tended to be higher on the lesion side compared with the intact side, although it did not reach statistical significance (P = 0.058). However, synaptic charge transfer was significantly increased on the lesion compared with the intact side of uncompensated chickens 3 days after UVG (Fig. 7A; Table 2), when mEPSCs recorded from principal cells on the lesion side exhibited both higher frequency and slower decay time compared with the intact side.

Fig. 7.

Synaptic charge transfer for mEPSCs and mIPSCs in principal cells before and after UVG. A: synaptic charge transfer for mEPSCs was significantly larger on the lesion side compared with the intact side in the uncompensated chickens 3 days after UVG. B: no significant difference was found for mIPSC synaptic charge transfer among all the experimental groups. *Significant difference between 2 columns at P < 0.05.

In contrast, our findings on mIPSC frequency and decay time after UVG did not drive spontaneous synaptic inhibition in the same direction. For example, considerably higher frequency on the lesion side 3 days after UVG in compensating animals would tend to increase the inhibitory synaptic drive, but the faster decay time would tend to decrease the synaptic drive. According to measurements of synaptic charge transfer for mIPSCs, no significant difference was detected in principal cells bilaterally from all the experimental groups (Fig. 7B; Table 2). In summary, synaptic charge transfer was found to differ only for mEPSCs in the uncompensated animals 3 days after UVG. Thus symmetric mEPSC and mIPSC synaptic charge transfer between principal cells on the lesion and intact sides is a useful parameter to define the process of vestibular compensation.

DISCUSSION

For the first time, this work provides direct evidence for changes in frequency, kinetics, and synaptic charge transfer of spontaneous excitatory and inhibitory synaptic inputs in a morphologically defined subset of vestibular nuclei neurons during early stage of vestibular compensation. Three days after UVG, excitatory synaptic charge transfer was symmetric bilaterally in the compensating chickens but was significantly higher on the lesion side compared with the intact side in the uncompensated chickens. Increased mEPSC synaptic charge transfer on the lesion side of uncompensated chickens was due to combined higher frequency, which is under the control primarily of presynaptic factors, and slower decay time, which is regulated mainly by postsynaptic factors (see below). In contrast, inhibitory synaptic charge transfer was not significantly different in principal cells on the lesion and intact sides in the compensating and uncompensated chickens 3 days after UVG, because higher mIPSC frequency was offset by faster decay time. Thus a major difference between compensating and uncompensated chickens 3 days after UVG is the failure to maintain bilateral symmetry in excitatory synaptic charge transfer in principal cells from uncompensated chickens. In addition, this study provides the time course for changes in spontaneous synaptic inputs on a morphologically defined subset of vestibular nuclei neurons and identifies the excitatory or inhibitory nature of the inputs during early recovery after UVG.

The percentage of spontaneous spike firing principal cells and spike firing rates was generally similar to those obtained from an earlier whole-cell recording study of principal cells using potassium methylsulfate pipet solution before and after UVG (Shao et al. 2009). However, a significantly lower percentage of spontaneous spike firing principal cells was recorded here on the intact side of compensating chickens (11%, n = 27, vs. 67%, n = 15). Since no other differences were detected, the reason for this change is not apparent.

Role of Excitatory Inputs in Vestibular Compensation

Approximately 24% of synaptic terminals in the tangential nucleus originate from the primary vestibular fibers (Popratiloff et al. 2004). Large, axosomatic “spoon terminals,” formed by the primary vestibular fibers on the principal cell bodies, degenerate 1–3 days after UVG, while the primary vestibular fibers themselves degenerate by 7 days (Aldrich and Peusner 2002). Glutamate concentration decreases on the lesion side of rat vestibular nuclei 2 days after UVG (Li et al. 1996a). In the present study, increased EPSC frequency was detected in principal cells at 1 day after UVG and remained significantly increased only in the uncompensated chickens (Fig. 4B₁). Slower decay time was recorded in principal cells on the lesion side by 3 days in both compensating and uncompensated chickens (Fig. 6, A and B). In fact, excitatory synaptic charge transfer was significantly higher in principal cells on the lesion side of the uncompensated chickens compared with the intact side (Fig. 7A) and tended to be higher in principal cells on the lesion side at 1 day. Failure to detect significant differences in synaptic charge transfer on the lesion and intact side at 1 day may be due to the mixture of compensating and uncompensated chickens at this stage, which were indistinguishable by behavior. Thus increased frequency of excitatory spontaneous synaptic activity may be the first step to counter the effect of loss of vestibular afferents on the lesion side after vestibular deafferentation. Apparently, the second step for compensation to continue is to achieve balanced excitatory synaptic drive bilaterally, as recorded here in the compensating chickens.

EPSC frequency.

The most important presynaptic factors capable of generating increased EPSC frequency include the following: 1) increased glutamatergic terminal number due to the sprouting of new terminals, 2) increased probability of glutamate release from existing presynaptic terminal, and/or 3) increased spike firing of presynaptic neurons. The most important postsynaptic factor influencing EPSC frequency is the density of AMPA receptor subunits at synapses (Petralia et al. 1999). In rat vestibular nuclei, no change in glutamatergic terminal density is detected after UL using immunolabeling for the glutamate transporter VGLUT2 (Eleore et al. 2007). Also, no new terminals are detected on the principal cell bodies 1–56 days after UVG (Aldrich and Peusner 2002), although terminals in the neuropil were not quantified. Increased EPSC frequency in principal cells could be related to the decreased Kv1.2 expression found in terminals on the principal cell bodies on the lesion side of uncompensated chickens 3 days after UVG (Shao et al. 2009). In other systems, decreased Kv1.2 expression has been linked to increased neurotransmitter release (Dodson et al. 2003). Future studies should determine the neurotransmitter phenotype of the synaptic terminals on the principal cells undergoing decreased Kv1.2 expression 3 days after UVG and whether Kv1.2 expression decreases in these terminals 1 day after UVG. The latter result could provide an early marker to distinguish compensating from uncompensated chickens by 1 day. So far, there is no evidence for increased glutamate release from the remaining terminals on the principal cell or increased density of postsynaptic AMPA receptor subunits during vestibular compensation (Li et al. 1996b; Horii 2001).

Kinetics of mEPSC.

The composition of AMPA receptor subunits in the postsynaptic membrane is a major factor influencing decay time (Jonas 2000). GluR3 and GluR4 are the predominant AMPA receptor subunits in the hatchling principal cells, with GluR1 and GluR2 expressed in low amounts (Popratiloff et al. 2004). Since GluR2 subunits show relatively slow gating and GluR4 fast decay (Geiger et al. 1995), the slower EPSC decay time in the principal cells on the lesion side could be due to increased GluR2 expression and/or a decreased GluR4 expression, and vice versa on the intact side. Moreover, since AMPA receptors with abundant GluR2 subunit are calcium impermeable (Geiger et al. 1995), increased GluR2 expression on the lesion side could protect the principal cells against glutamate-induced cell death resulting from calcium loading. In fact, none of the principal cells degenerate up to 56 days after UVG (Aldrich and Peusner 2002). Decreased calcium loading could also increase the trafficking of AMPA receptor subunits by activating a transcription factor which increases subunit accumulation at synaptic sites (Turrigiano 2008).

Role of Inhibitory Inputs in Vestibular Compensation

In addition to the critical role for excitatory events in recovery of function after UVG, inhibitory events play an important role but with a delayed time course. Indeed, IPSC frequency increased significantly by 3 days in principal cells on the lesion side in all experimental groups. However, in all cases, the higher frequency was offset by significantly faster decay time so that IPSC synaptic charge transfer remained in balance bilaterally

IPSC frequency.

Increased IPSC frequency was primarily due to increased GABAergic events. GABA and glycine-positive terminals show distinct subcellular termination patterns in the tangential nucleus (Popratiloff and Peusner 2011). More than 50% of synaptotagmin-1-labeled terminals in the tangential nucleus are GABA-positive, while glycine is expressed in about one-third. Little colocalization of GABA and glycine is detected, supporting our finding here that the total percentage of GABAergic and glycinergic spontaneous synaptic events is ∼100% in principal cells.

The same presynaptic factors involved in increasing EPSC frequency can increase IPSC frequency. The neurons responsible for increased GABAergic inputs to the principal cells on the lesion side have not been identified. In the cat, increased numbers of GABA-positive neurons and GABA-labeled terminals are found on the lesion side 3 days after UVG (Tighilet et al. 2007), which could contribute to the increased GABAergic IPSCs. In the rat, the release of GABA neurotransmitter from commissural inputs to medial vestibular nucleus (MVN) neurons increases on the lesion side shortly after UL (Bergquist et al. 2008). Finally, it is interesting that GABAergic mIPSCs were absent from half the principal cells recorded on the intact side of uncompensated chickens. At present, there is no evidence for degeneration of GABAergic terminals in the intact tangential nucleus after UVG. Unlike GABAergic events, glycinergic mIPSC frequency did not change significantly in principal cells from any experimental group after UVG.

Kinetics of mIPSCs.

Changes in kinetics were recorded for both GABAergic and glycinergic events. Faster decay time was first detected on the lesion side in glycinergic IPSCs 1 day after UVG and in GABAergic IPSCs by 3 days. Faster decay time for GABAergic and glycinergic mIPSCs in principal cells is consistent with decreased sensitivity to GABA_A and glycine receptor agonists reported in mammalian MVN neurons on the lesion side after UL (Vibert et al. 2000; Yamanaka et al. 2000). However, no changes have been detected in GABA and glycine concentrations in presynaptic terminals or postsynaptic receptors of rat MVN neurons shortly after UL (de Waele et al. 1994; Eleore et al. 2004, 2005, 2007).

The few electrophysiological studies on inhibitory synaptic transmission during vestibular compensation have resulted in contrasting findings. On stimulating inhibitory commissural inputs of randomly selected MVN neurons bilaterally, response amplitudes decrease 3 days after UL in the guinea pig (Vibert et al. 1999). In mouse MVN neurons, the frequency of glycinergic mIPSCs increases bilaterally shortly after UL, with increased amplitude detected only in those neurons on the intact side (Lim et al. 2010). These differences in responses underscore the necessity of studying morphologically defined subsets of vestibular nuclei neuron, since different neuron subsets may have different neurotransmitter phenotypes and axonal connections, which can affect their responses to peripheral vestibular lesions. For example, after whisker removal, only one of four types of inhibitory neurons in layer 4 of mouse barrel cortex exhibits changes in IPSC kinetics (Li et al. 2009). Finally, these studies underscore the importance of calculating synaptic charge transfer to weigh the collective influence on synaptic drive of fluctuations in frequency and kinetics of spontaneous synaptic events.

In summary, our present results indicate that both spontaneous excitatory and inhibitory synaptic events in principal cells undergo major changes after UVG according to a set time course. The next step is to identify the source of the changing inputs, so that they can be stimulated selectively to determine how their spatiotemporal integration regulates principal cell output before and after UVG.

GRANTS

This work was supported by National Institute on Deafness and Other Communication Disorders Grant R01-DC-000970 and Shared Instrumentation Division of Research Resources Grant S10-RR-025565.

DISCLOSURES

No conflicts of interest, financial or otherwise, are declared by the author(s).

AUTHOR CONTRIBUTIONS

M.S., J.C.H., and K.D.P. conception and design of research; M.S. performed experiments; M.S., J.C.H., and K.D.P. analyzed data; M.S., J.C.H., and K.D.P. interpreted results of experiments; M.S., J.C.H., and K.D.P. prepared figures; M.S., J.C.H., and K.D.P. drafted manuscript; M.S., J.C.H., and K.D.P. edited and revised manuscript; M.S., J.C.H., and K.D.P. approved final version of manuscript.