Adaptation of chicken vestibular nucleus neurons to unilateral vestibular ganglionectomy

Neuroscience 161 (2009) 988 –1007

ADAPTATION OF CHICKEN VESTIBULAR NUCLEUS NEURONS TO UNILATERAL VESTIBULAR GANGLIONECTOMY M. SHAO,1 A. POPRATILOFF,1 J. YI, A. LERNER, J. C. HIRSCH AND K. D. PEUSNER*

principal cells in uncompensated chickens displayed gross asymmetry in these properties bilaterally. © 2009 IBRO. Published by Elsevier Ltd. All rights reserved.

Department of Anatomy and Regenerative Biology, George Washington University Medical Center, 2300 I Street Northwest, Washington, DC 20037, USA

Key words: plasticity, brain slice, sodium and potassium channels.

Abstract—Vestibular compensation refers to the behavioral recovery after a unilateral peripheral vestibular lesion. In chickens, posture and balance deficits are present immediately following unilateral vestibular ganglionectomy (UVG). After three days, most operated chickens begin to recover, but severe deficits persist in others. The tangential nucleus is a major avian vestibular nucleus whose principal cells are vestibular reflex projection neurons. From patch-clamp recordings on brain slices, the percentage of spontaneous spike firing principal cells, spike discharge rate, ionic conductances, and spontaneous excitatory postsynaptic currents (sEPSCs) were investigated one and three days after UVG. Already by one day after UVG, sEPSC frequency increased significantly on the lesion side, although no differences were detected in the percentage of spontaneous spike firing cells or discharge rate. In compensated chickens three days after UVG, the percentage of spontaneous spike firing cells increased on the lesion side and the discharge rate increased bilaterally. In uncompensated chickens three days after UVG, principal cells on the lesion side showed increased discharge rate and increased sEPSC frequency, whereas principal cells on the intact side were silent. Typically, silent principal cells exhibited smaller persistent sodium conductances and higher activation thresholds for the fast sodium channel than spiking cells. In addition, silent principal cells on the intact side of uncompensated chickens had larger dendrotoxin-sensitive potassium conductance, with a higher ratio of Kv1.1 surface/cytoplasmic expression. Increased sEPSC frequency in principal cells on the lesion side of uncompensated chickens was accompanied by decreased Kv1.2 immunolabeling of presynaptic terminals on principal cell bodies. Thus, both intrinsic ionic conductances and excitatory synaptic inputs play crucial roles at early stages after lesions. Unlike the principal cells in compensated chickens which showed similar percentages of spontaneous spike firing cells, discharge rates, and sEPSC frequencies bilaterally,

Vestibular compensation is the popular term for the behavioral recovery which occurs following unilateral labyrinthectomy, or unilateral vestibular ganglionectomy (UVG) (for review, see Straka et al., 2005). Most vestibular ganglion cells do not recover their spike activity after labyrinthectomy (Sirkin et al., 1984), and the primary vestibular fibers degenerate centrally after UVG (Li et al., 1995; Aldrich and Peusner, 2002). Accordingly, the mechanisms underlying adaptation to vestibular deafferentation are thought to reside within the central nervous system. The cellular mechanisms underlying vestibular compensation have been investigated most extensively on medial vestibular nucleus (MVN) neurons from rat and guinea pig using both in vivo and in vitro preparations. In studies performed on whole animals, the immediate response to vestibular deafferentation is decreased spontaneous spike discharge rate in neurons on the lesion side, while the rate increases in neurons on the intact side (Smith and Curthoys, 1988a,b). Overall, behavioral recovery after labyrinthectomy coincides with the restoration of symmetric discharge rates in MVN neurons bilaterally, despite a certain degree of dissociation (see Ris et al., 1997). In brain slice preparations, the spontaneous spike firing rate increases on the lesion side within hours after labyrinthectomy (Cameron and Dutia, 1997; Vibert et al., 1999; Beraneck et al., 2003), whereas the discharge rate remains unchanged, or decreases on the intact side (Cameron and Dutia, 1997; Beraneck et al., 2004). Like in vivo studies, MVN neurons in brain slice preparations show a gradual recovery of symmetric spontaneous discharge rate bilaterally (Cameron and Dutia, 1997). In vitro studies indicate that vestibular nuclei neurons on the lesion side exhibit increased subthreshold potentials, which could explain the increased spontaneous spike discharge recorded (Him and Dutia, 2001). Furthermore, the relative proportions of the two main MVN neuron subclasses change after labyrinthectomy, with type B MVN neurons increasing on the intact side, while type A neurons increases on the lesion side (Beraneck et al., 2003, 2004). MVN neuron types are classified according to their spike waveform, without knowledge of their axonal output as projection, commissural, interneuron, vestibulocerebellar, vestibulothalamic, or a combination of these functional subtypes. Since neurons in

1 These authors contributed equally to the paper, each in his own field. *Corresponding author. Tel: ⫹1-202-994-3489; fax: ⫹1-202-994-8885. E-mail address: [email protected] (K. D. Peusner). Abbreviations: ACSF, artificial cerebrospinal fluid; AHP, afterhyperpolarization; CNQX, 6-cyano-7-nitroquinoxaline-2, 3-dione; CV, coefficient of variation; DTX, dendrotoxin; H, hatching day; IDS, dendrotoxinsensitive current; IgG, antibody; INa, fast sodium current; INaP, persistent sodium current; ISI, interspike interval; Kv1.1, Kv1.2, sustained potassium channel subunit with high sensitivity to dendrotoxin; MAP2, microtubule-associated protein 2; mEPSC, miniature excitatory postsynaptic current; MVN, medial vestibular nucleus; NA, numerical aperture; PBS, phosphate-buffered saline; sEPSC, spontaneous excitatory postsynaptic current; TTX, tetrodotoxin; UVG, unilateral vestibular ganglionectomy; VOC, vestibuloocular collic; VOR, vestibuloocular reflex.

0306-4522/09 $ - see front matter © 2009 IBRO. Published by Elsevier Ltd. All rights reserved. doi:10.1016/j.neuroscience.2009.04.027

988

M. Shao et al. / Neuroscience 161 (2009) 988 –1007

other sensory systems respond differentially to deafferentation depending on the neural circuitry in which they participate (e.g. see Francis and Manis, 2000), it is conceivable that vestibular neuron subtypes also respond differentially to vestibular deafferentation. Thus, it could be informative to record from a single functional class of vestibular nucleus neurons after vestibular deafferentation. The tangential nucleus is a major avian vestibular nucleus in the chicken. Its principal cells represent a morphologically distinctive class of glutamatergic (unpublished observations, Popratiloff and Peusner) vestibular reflex projection neurons, which receive monosynaptic input from the primary vestibular fibers (Peusner and Giaume, 1994). While some principal cells are vestibulo-ocular reflex (VOR) neurons which project to the oculomotor (Wold, 1978; Labandeira-Garcia et al., 1989, 1991; Petursdottir, 1990), trochlear (Evinger and Erichsen, 1986), or abducens nucleus (Gottesman-Davis and Peusner, 2008), other principal cells are vestibulo-ocular collic (VOC) neurons, whose axons terminate in the abducens nucleus with collaterals descending to high cervical spinal cord (Cox and Peusner, 1990a). Thus, all principal cells are VOR or VOC neurons. The distribution of VOR and VOC neuronal classes among the vestibular nuclei in chicken differs from mammals, since most VOR and VOC neurons are situated more laterally within the chicken medulla oblongata, with heavy involvement of neurons in the tangential, ventrolateral vestibular, and descending vestibular nuclei, and, unlike mammals, only a modest contribution from the MVN (Gottesman-Davis and Peusner, 2008). It will be important to determine how spontaneous spike discharge rate in principal cells is affected by UVG, and compare this to what was found after averaging responses from different MVN neuron classes after labyrinthectomy. Most important, recording from a homogeneous class of VOR and VOC neurons will broaden our understanding of the specific role of this neuron class in adaptation to vestibular deafferentation. The adult vestibular system exhibits plasticity in response to changing environmental demands (e.g. Mandl et al., 1981). However, the juvenile vestibular system is characterized by a more complete and rapid functional recovery after injury and lesions than the adult system (e.g. Kaga, 1999). Accordingly, here UVG was performed on four day-old hatchling chickens (hatching day [H] 4), which were sacrificed one (H5) or three days (H7) after surgery to trace the sequence of cellular events during early recovery. In preliminary experiments, some operated chickens did not begin to recover by three days, despite similar surgical and postsurgical treatment for all operated animals (see Curthoys and Hamalgii, 2007). Therefore, this new group of uncompensated animals was included in the present study. To identify which intrinsic membrane conductances underlie changes in the spontaneous spike discharge rate after UVG, we focused on the sodium and potassium channels, INaP, INa, and IDS, which are critically involved in the emergence of spontaneous spike discharge in developing principal cells (Gamkrelidze et al., 1998, 2000; Popratiloff et al., 2003; Shao et al., 2006a,b). In addition,

989

spontaneous excitatory postsynaptic currents (sEPSCs), which are crucial events in the assembly of the neural circuitry, were investigated (Shao et al., 2003, 2004). This work represents the first voltage-clamp study of sEPSCs, and sodium and potassium conductances, in combination with immunolabeling of potassium channels in vestibular nuclei neurons after UVG. Part of this work was presented in abstract form (Peusner et al., 2007; Popratiloff et al., 2007; Shao et al., 2007).

EXPERIMENTAL PROCEDURES Experimental animals Chick embryo eggs (Gallus gallus) were purchased from CBT Farm (Chesterton, MD, USA), and placed until hatching in an egg incubator equipped with circulated air, egg rotation unit, and temperature and humidity controls (model 1502, G.Q.F. Manufacturing Co., Savannah, GA, USA). Twenty-four hours after hatching, the chicken was identified as a one day-old hatchling (H1). Chickens were housed in cages equipped with heating lamps at the University’s Animal Research Facility until the appropriate ages (H4 –H7). 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 all efforts were made to minimize the number of animals used and their suffering.

UVG UVG was performed on H4 chickens, which were sacrificed one or three days later. The surgical protocol was adapted from previous studies (Aldrich and Peusner, 2002; Pollack et al., 2004). Briefly, under general anesthesia (ketamine, 100 mg/kg, Fort Dodge Animal Health, Fort Dodge, IO, USA; xylazine, 20 mg/kg; Akorn Inc., Decatur, IL, USA), 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 were removed to expose the lateral part of the bony vestibule of the inner ear, which after removal exposed the membranous labyrinths. The membranous labyrinths and bony medial wall of the vestibule were removed to expose the dura mater overlying the vestibular ganglion. The dura mater was cut, so that both the anterior and posterior portions of the vestibular ganglion and the auditory nerve were revealed and extirpated. A curved, flame-sharpened tungsten wire was used to cut the vestibular nerve between the vestibular ganglion and lateral brain surface. After achieving deep anesthesia in sham-operated chickens, the skin was cut on the left side of the head posterior to the external auditory meatus, several blood vessels in the fascia were coagulated, and the skin was sutured closed. Control, operated, and sham-operated chickens were housed under the same conditions. After UVG, the survival rate was about 90%. Operated chickens received daily i.m. injections of 0.25 ml penicillin-G prophylactically (Bimeda Inc., Le Sueur, MN, USA). Dehydration was treated by s.c. injections of sterile, lactated Ringer’s solution (Braun Medical, Inc., Irvine, CA, USA), which was started about 1–2 h after surgery when the operated chickens regained consciousness. The volume of replacement fluid was determined by the weight loss (National Research Council, 2003). On the second day after surgery, chickens which could not eat independently were given rice cereal for babies (Gerber, Fremont, MI, USA) via gavage every 3 hours, four times a day, until sacrificed. Operated chickens, which could stand, drink, and eat on their own three days after UVG were considered compensated, while those which could not perform these activities were considered uncompensated.

990

M. Shao et al. / Neuroscience 161 (2009) 988 –1007

Brain slice preparation and solutions The chickens were decapitated without anesthesia, and the brainstem and cerebellum were placed in a dissecting dish containing oxygenated, high sucrose, artificial cerebrospinal fluid (ACSF) at 1–2 °C. The cerebellum, choroid plexus, periotic capsule, and vestibular ganglia were removed from the brainstem, which was sectioned transversely in sucrose ACSF (1–2 °C) at 240 ␮m thickness using a vibratome (VT1000S; Leica Instruments, Bannockburn, IL, USA) and sapphire knife. Before and during slicing, the brainstems from chickens which underwent UVG were examined under a dissecting microscope for signs of hemorrhage or edema. Small amounts of blood were seen routinely at the lateral surface of the brain, but did not spread internally. If signs of internal brain hemorrhage were detected, the operated animals were excluded from the experiments (n⫽4). The brain slices were incubated in normal ACSF at 37 °C for 40 min, and then maintained in room temperature ACSF (21–22 °C) for at least 20 min before transferring to the recording chamber. Preheated ACSF was superfused through the recording chamber (volume, 180 ␮l; Warner Instruments, Hamden, CT, USA) at a rate of 2–3 ml/min, while ACSF in the recording chamber was maintained at 30 –31 °C using a temperature controller (TC324B, Warner Instruments). Sucrose ACSF contained (in mM): 210 sucrose, 2.5 KCl, 7 MgSO4, 1.5 NaH2PO4, 1 CaCl2, 26 NaHCO3, and 10 D glucose. ACSF contained (in mM): 126 NaCl, 3.5 KCl, 1.2 NaH2PO4, 1.3 MgCl2, 2 CaCl2, 26 NaHCO3, and 10 D glucose. Both sucrose ACSF and ACSF were bubbled with 95% O2/5% CO2 to maintain the pH at 7.2–7.4. The osmolarity was 310 –320 mOsm. Tetrodotoxin (TTX; RBI, St. Louis, MO, USA) was applied to block the voltage-dependent, fast (INa) and persistent sodium currents (INaP). Dendrotoxin (␣-DTX, Sigma, St. Louis, MO, USA) was used to block DTX-sensitive low threshold, potassium currents (IDS). Bicuculline methochloride (Tocris, Ellisville, MI, USA), strychnine (Sigma), 6-cyano-7-nitroquinoxaline-2, 3-dione (CNQX) (RBI) and DL-2-amino-5-phosphovaleric acid (AP-5; Sigma) were applied to block GABAA, glycine, AMPA and N-methyl-D-aspartate receptor-mediated events, respectively.

Electrophysiological techniques Brain slices were viewed on a fixed-stage, upright microscope (Zeiss Axioskop, FS-1, Jena, Germany) equipped with differential interference contrast optics and a 40⫻ water-immersion lens (numerical aperture [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 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). Microelectrodes (2–5 M⍀) were pulled from thin-walled (1.5 mm OD; 1.12 mm ID) borosilicate glass tubing (World Precision Instruments, Sarasota, FL, USA) using a Brown/Flaming horizontal puller (P-87, Sutter Instruments, Novato, CA, USA). Potassium methylsulfate (KMeSO4) pipet solution contained (in mM): 120 KMeSO4, 20 KCl, 10 Hepes, 0.5 EGTA, 2 MgCl2, 4 Na2ATP, 0.3 Tris GTP. The osmolarity was 270 –290 mOsm, and the pH was adjusted to 7.2 with KOH. The liquid junction potential (⫹6 mV) was not corrected. Series resistance was compensated at 80% (lag 10 ␮s). Biocytin (0.5%) (Sigma) was dissolved in the pipet solution weekly and kept frozen between experiments.

Electrophysiological data acquisition and analysis Data were acquired using Clampex 9.2 (Axon Instruments, Foster City, CA, USA). Immediately after obtaining the whole-cell configuration, the input resistance was determined by giving a ⫹5 mV

pulse using the seal test function in voltage-clamp mode. The resting membrane potential was read from the amplifier panel meter in current-clamp mode. Signals were amplified using Axopatch-200 B in normal mode (Axon Instruments), as recommended when using ⬍10 M⍀ recording pipets (Axopatch 200 B Theory and Operation Manual, Axon Instruments, 1999; Shao et al., 2006a). Signals recorded in current-clamp mode were digitized at 20 Hz and filtered at 5 Hz, while signals recorded in voltage-clamp were digitized at 10 kHz and filtered at 2 kHz. Spontaneous spike activity was measured first in the cellattached mode at ⫺60 mV for ⬍1 min before establishing the whole-cell current-clamp configuration and recording the activity at resting membrane potential for 2–3 min. Silent cells were tested for spike firing by driving the membrane potential to ⫺50 mV. The spontaneous spike discharge rate was expressed as the inverse of the mean interspike interval (ISI). The coefficient of variation (CV) was defined as the standard deviation of the ISI divided by the mean ISI (Shao et al., 2006a). Spike discharge rate was obtained for an experimental group by averaging the activity from the spontaneous spike firing cells only. The kinetics for spontaneous action potentials were analyzed from averages of 15 spikes/ cell. The kinetics for evoked action potentials were calculated from the spikes generated on depolarization using the lowest intensity injected current. Spike threshold was defined either visually as the point in the voltage trace where a sharp inflection occurred just before generating the action potential, or using Minianalysis program (6.0.3, Synaptosoft Inc., Decatur, GA, USA), which determines the spike threshold based on the value of dV/dt (Shao et al., 2006a). Spike amplitude was defined as the potential difference between threshold and peak amplitude. Rise time was defined as the duration between threshold and peak amplitude during depolarization. Decay time was defined as the duration between peak amplitude and threshold during repolarization. Action potential width was measured as the duration between 50% spike amplitude during depolarization and repolarization. AHP amplitude was defined as the difference in voltage between the spike threshold and peak negativity after a spike. The gain for the evoked action potential (m) was defined as the ratio of the firing rate relative to the injected current (Smith et al., 2002; Shao et al., 2006a). INa, INaP, and IDS were recorded in voltage-clamp mode. INa was induced using 100 ms voltage steps from ⫺100 to ⫺10 mV, in 10 mV increments. INaP was induced using 800 ms voltage steps from ⫺80 to ⫺47.5 mV in 2.5 mV increments. Outward currents were induced using 600 ms voltage steps from ⫺80 to ⫹10 mV, with 5 mV increments, preceded by an 800 ms prestep to ⫺100 mV. INa was measured as the difference between baseline and the inward peak by subtracting the traces before and after TTX. INaP was measured 100 ms after initiating the voltage step by subtracting the traces before and after TTX. IDS was measured 300 ms from the onset of the voltage step, and was determined by subtracting the outward current traces before and after DTX exposure. To study the outward current, ACSF was modified to contain 0.5 mM CaCl2 and 2.8 mM MgCl2 to minimize the Ca2⫹dependent currents, and 1 ␮M TTX was added to block the voltage-gated sodium currents. sEPSCs were recorded at a holding potential of ⫺60 mV in voltage-clamp mode, and analyzed off-line using Minianalysis program (Synaptosoft Inc.). The minimal acceptable amplitude was set at 15 pA, which was determined from the root mean square (rms) of the baseline noise (3– 4 pA). Values were presented as mean⫾SEM. Differences were considered significant when P⬍0.05 (independent Student’s ttest). Differences in percentages were considered significant with P⬍0.05 (Fisher’s exact test) (GraphPad InStat, GraphPad software, La Jolla, CA, USA). In all experiments, the data from H5 and H7 controls were not significantly different, so they were pooled and presented together.

M. Shao et al. / Neuroscience 161 (2009) 988 –1007

Visualization of principal cells injected with biocytin in brain slices After the recordings, the brain slices were processed to visualize the biocytin-injected neurons using streptavidin Alexa Fluor 647 (Molecular Probes, Eugene, OR, USA) (see Shao et al., 2003). Biocytin-labeled neurons were observed on an Olympus IX70 inverted microscope equipped with Bio-Rad confocal hardware (MRC 1024 ES, Bio-Rad laboratory, Hercules, CA, USA) and 10⫻ UPlanFl (NA, 0.3) and 20⫻ PlanApo (NA, 0.70) objectives. Excitation for Alexa Fluor 647 was achieved using a standard protocol for Cy5. Recorded neurons were confirmed as principal cells by studying their somatic size, shape, and position in the tangential nucleus relative to the primary vestibular fibers in 10⫻ confocal images. In addition, 20⫻ confocal images were used to measure the longest axis of the cell body and axis perpendicular to this, count the primary dendrites, and measure the extension of the dendrites in the dorsoventral and mediolateral axes.

Animal perfusion and fixation protocols for immunolabeling experiments To improve immunolabeling of membrane-integrated proteins, the standard fixation protocol was lowered to 1% paraformaldehyde fixative from the 4% paraformaldehyde used in developmental studies (Popratiloff et al., 2003). The application of 1% paraformaldehyde fixation was important for visualizing Kv1 membraneintegrated proteins. Originally, 1% fixative was designed to reveal AMPA receptors integrated into synaptic membranes (Burette et al., 2001; Melone et al., 2005), since the standard 4% paraformaldehyde fixation creates strong bonds between proteins, which can block epitope binding. However, the 1% paraformaldehyde fixation can decrease the tissue integrity, so overnight immersion fixation was added to the fixation protocol. Throughout the study, the quality of immunolabeling was weighed against maintaining adequate tissue preservation.

991

Briefly, under deep anesthesia, the chickens were perfused via the ascending aorta with a brief flush of 0.1 M phosphate buffered saline (PBS) (1 min), followed by 20 min with 1% paraformaldehyde in 0.1 M phosphate buffer. After perfusion, the chickens were decapitated and the heads were immersed in 1% paraformaldehyde fixative overnight (12 h). On the following day, the brains were dissected out of the cranium and immersed in PBS just before sectioning. To evaluate the effect of fixative strength on Kv1.1 and Kv1.2 immunolabeling, preliminary experiments were performed to test weaker and stronger variations of the 1% fixation protocol. The weak fixation protocol used 1% paraformaldehyde perfusion for 1–2 min (n⫽2), while the strong fixation protocol used 1% paraformaldehyde perfusion for 1 h (n⫽2). After using different fixation protocols, all the tissues were processed according to the standard procedure given above, unless otherwise stated. The strong and weak fixation protocols produced striking differences in Kv immunolabeling from the standard protocol (Fig. 1A). The weak fixation protocol produced deteriorated cellular morphology (i.e. loss of principal cell body shape), less Kv1.1 cytoplasmic immunolabeling in the principal cell bodies (Fig. 1B), and weaker Kv1.2 pericellular labeling (not shown). The brainstems which underwent strong fixation lost the characteristic patchy Kv1.1 immunolabeling in the principal cell bodies, which instead appeared punctate on the cell surface and in the cell body (Fig. 1C), and Kv1.2 pericellular labeling was weaker (not shown). In addition, after using the strong fixation protocol in a few experiments, the tissue sections were maintained at 4 °C for 1 week. In these experiments, the background immunolabeling increased significantly, resulting in decreased signal/noise ratio (not shown). Finally, line scans drawn through the principal cell bodies confirmed that the tissue processed by the weak or strong fixation protocols produced lesser pixel brightness for the principal cell cytoplasm and cell surface than from the standard protocol (Fig. 1D–F). Accordingly, all data presented here were obtained from the tissue fixed using the standard protocol.

Fig. 1. Effect of fixation strength on Kv1.1 immunolabeling in principal cell bodies from H7 controls. (A) Using the routine fixation protocol, large clusters of Kv1.1 immunolabeling appeared in the principal cell bodies. (B) Using the weak fixation protocol, some principal cells were disfigured (arrow), and Kv1.1 cytoplasmic labeling was decreased. (C) Using the strong fixation protocol, Kv1.1 labeling was decreased. (D–F) Graphs representing pixel brightness levels along lines drawn through principal cells in A–C, respectively. Principal cells in B and C showed decreased signal/noise ratio, as plotted in E and F, respectively. In all of the principal cells whose Kv1 immunolabeling was analyzed in the present study, the routine fixation protocol was applied.

992

M. Shao et al. / Neuroscience 161 (2009) 988 –1007

Tissue processing for immunolabeling experiments All the operated chickens processed for immunolabeling were checked before sectioning for the complete removal of the vestibular ganglion on the lesion side, and the absence of gross brain damage, edema, and hemorrhaging. Despite no signs of brain damage after UVG, all of the operated chickens used for the immunolabeling experiments remained uncompensated after UVG. The brainstem containing the tangential nucleus was glued to the stage of a vibratome (Oxford), and 50 ␮m transverse sections were cut in ice cold PBS. Sections containing the tangential nucleus were identified under a dissecting microscope and processed for immunolabeling immediately, or were stored in the refrigerator overnight.

Antibodies Anti Kv1.1 and Kv1.2 were rabbit antibodies against fusion proteins corresponding to residues 416 – 495 and 417– 499 of the C-terminus of native mouse Kv1.1 and Kv1.2, respectively (Alomone Laboratories, Jerusalem, Israel). Both antibodies were used at concentrations of 1:1000. A monoclonal antibody against microtubule-associated protein 2 (MAP2 recognizes both MAP2A and MAP2B) (MAB 3418, Chemicon, Billerica, MA, USA) was used at a concentration of 1:500. Mouse antibody for synaptotagmin, a specific synaptic vesicle protein, was used at a concentration of 1:400 (MAB 5200; Chemicon).

Immunolabeling procedure The immunolabeling procedure was modified from a previous study (Popratiloff et al., 2003) to permit simultaneous acquisition of two immunolabels on the confocal microscope using the 488 and 647 lines of the Kr/Ar laser. Both Kv1.1 and Kv1.2 were revealed with secondary antibodies conjugated to Alexa Fluor 488 (instead of Alexa Fluor 594), while MAP2 and synaptotagmin were revealed with secondary antibodies conjugated to Alexa Fluor 647 (instead of Alexa Fluor 488). Brainstem sections containing the tangential nucleus were subdivided into two groups, which were immunolabeled simultaneously with either rabbit anti-Kv1.1 antibody (IgG) (Alomone Laboratory) and mouse anti-MAP2 IgG (Chemicon) to label the neurons, or rabbit anti-Kv1.2 IgG (Alomone Laboratory) and mouse synaptotagmin IgG (Chemicon) to label the synaptic terminals. Since low concentration 1% paraformaldehyde fixation was used, tissue permeabilization with 50% alcohol or Triton X-100 was unnecessary. All steps in tissue processing were performed in PBS. Briefly, after extensive rinsing with PBS, sections were incubated in 10% normal goat serum in PBS (30 min), followed by incubation in a cocktail of primary rabbit IgG specific for Kv1.1 (1:1000 in PBS) and mouse IgG specific for MAP2 (1:500 in PBS) for 12–14 h. Alternate sections from each brainstem were incubated in a cocktail of rabbit IgG specific for Kv1.2 (1:1000 in PBS) and mouse IgG marker specific for synaptotagmin (1:400 in PBS). On the next day, the sections were rinsed with PBS and incubated in 2% normal goat serum in PBS (10 min). Then, the sections were incubated (2 hours) in a cocktail of secondary goat anti-rabbit IgG conjugated to Alexa Fluor 488 and goat anti-mouse IgG conjugated to Alexa Fluor 647 (Molecular Probes, Carlsbad, CA, USA). Finally, all the tissue sections were rinsed and mounted on glass slides, air dried, and coverslipped with anti-fading, water-soluble medium (Fluoromount G; Electron Microscopy Sciences). Two controls for antibody specificity were performed. First, the primary antibodies to Kv1.1 and Kv1.2 were pre-incubated with the corresponding fusion protein (1 h) before performing the immunolabeling protocol as described above using the pre-absorbed primary antibodies (Popratiloff et al., 2003). A second

control omitted the primary antibodies. In both control experiments, Kv.1.1 and Kv1.2 immunolabeling was not detected.

Confocal imaging and analysis Confocal images were captured on an Olympus IX70 microscope equipped with a 60⫻ (NA, 1.4) objective and Bio-Rad MRC 1024 ES confocal hardware. Alexa Fluor 488 was displayed on the red channel, while Alexa Fluor 647 was displayed on the green channel. Confocal stacks were collected from selected tissue sections based on MAP2-labeling of randomly selected principal cells. Kv1.1 and Kv1.2 immunolabeling was adjusted within the image dynamic range using the gain and offset controls of the photomultiplier. To maintain consistency between images, a minimum number of pixels were saturated at 0 and at the maximum grayscale value of 255. Thus, routinely the dark current overlapped with pixels at the lower dynamic range of the image, and immunolabeling was represented by a Bolzmann distribution. Images were analyzed using ImageJ software (National Institutes of Health). By viewing separately the MAP2 and synaptotagmin labeling, each identified principal cell body and the synaptic profiles on the principal cell surface could be outlined. Then, the outline was transferred to the confocal image depicting Kv1.1 or Kv1.2 labeling, so that the mean pixel brightness of cell bodies (Kv1.1) or synaptic terminals (Kv1.2) was obtained. A similar protocol was applied to extract the signal from the cell nucleus, which was used to set the background level in the confocal stack. Raw data, representing mean pixel brightness of principal cell bodies or synaptic profiles, were collected from controls, shamoperated, and the lesion and intact sides of uncompensated chickens three days after UVG. In all experiments, the raw number for each principal cell cytoplasmic value was normalized by subtracting the value for the principal cell nucleus. Data sets were tested for significance using Kruskal–Wallis non-parametric analysis, or an independent sample t-test, with significance set at P⬍0.05. Data are presented as mean⫾SEM. To evaluate Kv1.1 surface expression, one pixel thick lines (0.31 ␮m at 60⫻ magnification; NA, 1.4) were drawn along the entire surface of the principal cell bodies identified by the presence of MAP2 (Fig. 8). The total length of the surface lines was 1521 ␮m (n⫽21 neuronal profiles), 1971 ␮m (n⫽21), 1742 ␮m (n⫽20) from principal cell bodies in controls, and the lesion and intact sides of uncompensated chickens three days after UVG, respectively. The lines were transferred to the Kv1.1 channel where pixel brightness along the lines was calculated and plotted. In addition, lines were drawn at random within the cytoplasm and cell nucleus for each principal cell body. The mean values of the line scans from both the surface and cytoplasm were normalized by subtracting the value of line scan from the cell nucleus.

RESULTS The electrophysiological experiments were performed on control animals at H5 (n⫽15 animals) and H7 (n⫽35) (Fig. 2A), hatchlings which underwent UVG at H4 and were sacrificed one day later (H5, n⫽45 uncompensated chickens) (Fig. 2B), hatchlings which could stand eat, and drink three days after UVG (H7, n⫽30 compensated chickens) (Fig. 2C), and hatchlings which were unable to stand, feed or drink independently three days after UVG (H7, n⫽24 uncompensated chickens) (Fig. 2D). Immunolabeling experiments were performed on H7 control animals (n⫽2), animals which underwent sham UVG at H4 and were sacrificed at H7 (n⫽5), and uncompensated chickens three days after UVG (n⫽4).

M. Shao et al. / Neuroscience 161 (2009) 988 –1007

993

Fig. 2. Behavior before and after UVG. (A) Normal H7 chicken. (B) One day after UVG, all of the chickens were uncompensated, since they could not stand and exhibited roll head tilt 90 –180° toward the side of the lesion, flexion of the ipsilateral leg and extension of the contralateral leg. (C) Three days after UVG, 52% of the chickens were compensated. These chickens could stand, eat, and drink independently, but retained a 30° head tilt. (D) Three days after UVG, 48% of the chickens were uncompensated. The chickens could not stand, feed, and drink independently, and retained a 90 –180° head tilt.

Behavioral deficits after UVG One day after the surgery, none of the operated chickens could stand, eat, or drink, and all of them exhibited roll head tilt from 90 to 180° toward the side of the lesion, flexion of the ipsilateral lower extremity, and extension of the contralateral lower extremity (Fig. 2B). However, three days after UVG, about half of the operated chickens (30/ 58; 52%) had recovered from most of the early static deficits in posture and balance, except for a 30° roll head tilt toward the side of the lesion (Fig. 2C), which has been reported to persist for 56 days after the lesion (Aldrich and Peusner, 2002). The remaining operated chickens still could not stand, feed or drink independently, and exhibited roll head tilt from 90 to 180° toward the side of the lesion. However, muscle tone in their lower extremities recovered to some extent (Fig. 2D). Morphology of biocytin-injected principal cells In controls, the biocytin-injected principal cells with their large, oval cell bodies measuring 24⫻16 ␮m, on average, had axons which originated from the medial side of the soma, and three primary dendrites which extended most often dorsally and laterally (Fig. 3A, Table 1). Routinely, only short segments of the axon were present in the brain slice (Fig. 3, arrowheads). Nonetheless, most principal cells generated spontaneous spike activity, which suggests that their spike initiation site was still functional in the brain slices, as found for neurons in other systems (Khaliq and Raman, 2006). In all experimental groups, the principal cell body size and number of primary dendrites did not change (Table 1). The primary dendrites of the principal cells either extended for long distances with few branches (Fig. 3C), like typical VOR neurons (e.g. McCrea et al., 1987), or were shorter with increased dendritic branching (Fig. 3B). No correlation was found between the dendritic pattern and ability to generate spontaneous spike activity (n⫽31). In controls, the dorsoventral dendritic extension was 246 ␮m, while the mediolateral dendritic extension was 313 ␮m, on average (n⫽34). In all experimental groups, some of the dendrites of the principal cells exhibited beading (Fig. 3, arrows). In all the operated chickens, the mediolateral dendritic extension decreased signifi-

cantly on the lesion side compared to controls (Table 1), as well as on the lesion side compared to the intact side at one day after UVG. Also, the dorsoventral dendritic extension on the lesion side decreased compared to controls one day after UVG. Dendritic shrinkage could be related to the degeneration of the primary vestibular fibers centrally, which starts one day after UVG (Aldrich and Peusner, 2002). Passive membrane properties Whole-cell patch-clamp recordings were obtained from 154 morphologically identified principal cells. The experimental groups included principal cells from controls (H5, n⫽15; H7, n⫽26), lesion (n⫽25) and intact side (n⫽28) one day after UVG, lesion (n⫽16) and intact side (n⫽15) of compensated chickens three days after UVG, and the lesion (n⫽12) and intact sides (n⫽17) of uncompensated chickens three days after UVG. In controls, the average resting membrane potential of the principal cells was ⫺59.4⫾0.9 mV, with no difference between the spontaneous spike firing and silent cells, or that reported for principal cells at H5 (⫺60⫾4 mV; liquid junction potential not corrected) (Shao et al., 2006a). After UVG, the resting membrane potential of the principal cells fluctuated. One day after UVG, the principal cells on the lesion side were depolarized compared to those on intact side and from controls, whereas three days after UVG, the principal cells on the intact side were depolarized compared to those on the intact side one day after UVG, regardless of whether the animals were compensated or uncompensated (Fig. 4A). In controls, the average input resistance of the principal cells was 144⫾9.4 M⍀, with no difference detected between spontaneous spike firing and silent cells. Three days after UVG, the input resistance of the principal cells was significantly higher on the lesion side compared to the intact sides from all operated chickens (Fig. 4B). Increased input resistance on the lesion side may be related to dendritic shrinkage (Carrascal et al., 2005). Altogether, after UVG, the principal cells on both the lesion and intact sides were susceptible to changes in their passive membrane properties.

994

Cell body size (␮m) Number of primary dendrites Mediolateral dendritic extension (␮m) Dorsoventral dendritic extension (␮m)

Controls (n⫽34)

1 day after UVG Lesion side (n⫽20)

24⫾0.8⫻16⫾0.7 3⫾0.2 313⫾17a 246⫾13c

24⫾1.1⫻17⫾0.8 2⫾0.3 221⫾22a,b 192⫾19c

3 days after UVG, compensated

3 days after UVG, uncompensated

Intact side (n⫽26)

Lesion side (n⫽14)

Intact side (n⫽12)

Lesion side (n⫽14)

Intact side (n⫽12)

25⫾1.3⫻18⫾0.7 3⫾0.3 307⫾20b 232⫾17

25⫾2⫻18⫾1.4 3⫾0.4 247⫾34a 235⫾30

25⫾1.4⫻17⫾0.2 3⫾0.5 300⫾31 258⫾33

24.4⫾2.1⫻15.4⫾1.3 2.5⫾0.2 238.1⫾37a 223.1⫾40

22.3⫾0.9⫻15.4⫾0.8 2.1⫾0.3 270.1⫾33.6 240.4⫾41.8

Values are presented as mean⫾SEM. There was a significant difference in the mediolateral dendritic extension between the control and lesion sides in all operated animals (P⬍0.05). b There was a significant difference in the mediolateral dendritic extension between the lesion and intact sides one day after UVG (P⬍0.05). c There was a significant difference in the dorsoventral dendritic extension between the control and lesion side one day after UVG (P⬍0.05). a

M. Shao et al. / Neuroscience 161 (2009) 988 –1007

Fig. 3. Biocytin-injected principal cells recorded in brain slices. (A) H7 control. (B, C) Lesion side and intact sides of compensated chicken three days after UVG. (D, E) Lesion and intact sides of uncompensated chicken three days after UVG. Arrows point to beaded dendrites. Arrowheads point to axons of the principal cells. D, Dorsal; V, ventral; L, lateral; M, medial. Table 1. Morphometric analysis of principal cells before and after UVG

M. Shao et al. / Neuroscience 161 (2009) 988 –1007

995

significant change was the increased percentage of spontaneous spike firing cells on the lesion side of compensated chickens three days after UVG compared to controls. Since there are no signs of neuron loss in the tangential nucleus from one to 56 days after UVG (Aldrich and Peusner, 2002), the spontaneous spike firing and silent cells must be capable of transforming bidirectionally after vestibular deafferentation. To further define the properties of the spontaneous spike firing and silent principal cells, their intrinsic membrane conductances and synaptic inputs were recorded from the different experimental groups (see below).

Fig. 4. Passive membrane properties of principal cells before and after UVG. (A) One day after UVG, the resting membrane potential of the principal cells on the lesion side (⫺55.6⫾1 mV, n⫽25) was depolarized compared to controls (⫺59.4⫾0.9 mV, n⫽41) and the intact side (⫺61.6⫾1.1 mV, n⫽28). Also, the resting membrane potential of the principal cells on the intact side was depolarized three days after UVG in both compensated (⫺56.9⫾1.1 mV, n⫽15) and uncompensated (⫺57.8⫾1.2 mV, n⫽17) chickens compared to one day after UVG. However, no significant difference was found in the principal cells on the lesion side in both compensated (⫺57⫾1.2 mV, n⫽16) and uncompensated (⫺56.7⫾1.2 mV, n⫽12) chickens three days after UVG compared to one day after UVG. (B) One day after UVG, the input resistance did not change for principal cells on the lesion (162.8⫾17.5 M⍀, n⫽25) or intact sides (135⫾13.2 M⍀, n⫽28) compared to controls (144⫾9.4 M⍀, n⫽41). Three days after UVG, the input resistance of the principal cells was significantly higher on the lesion side (178.9⫾10.2 M⍀, n⫽16) compared to the intact side (126⫾13.7 M⍀, n⫽15) in the compensated and also the uncompensated chickens (lesion side: 176.3⫾16.6 M⍀, n⫽12; intact side: 120.8⫾9.7 M⍀, n⫽17) (P⬍0.05).

Spontaneous spike firing pattern According to their ability to generate spikes at resting membrane potential, the principal cells were divided into two classes, spontaneous spike firing cells and silent neurons (Fig. 5A1, A2). The majority of silent cells (68/72) did not fire spikes, even when the membrane potential was driven up to ⫺50 mV using steady depolarizing current. However, on further depolarization, all the silent cells could fire action potentials (see below). The analysis of spontaneous spike firing cells included: (1) percentage of recorded neurons, (2) spike discharge rate, and (3) spike discharge regularity (CV). Percent of spontaneous spike firing cells. The percentage of spontaneous spike firing cells was 56% (23/41) in controls, 48% (12/25) on the lesion side and 43% (12/ 28) on the intact side one day after UVG, 94% (15/16) on the lesion side and 67% (10/15) on the intact side of the compensated chickens three days after UVG, and 83% (10/12) on the lesion side and 0% (0/17) on the intact side of the uncompensated chickens (Fig. 5B). It was most striking that all the principal cells on the intact side of the uncompensated chickens were silent (P⬍0.01). Another

Spike discharge rate and regularity. Spontaneous spike discharge rate was measured from recordings of principal cells using either cell-attached or whole-cell configuration. The mean discharge rate for the spontaneous spike firing principal cells was 25.4⫾3 spikes/s under cellattached (n⫽64), and 25.2⫾2.7 spikes/s in the whole-cell configuration (n⫽64). Thus, spontaneous spike discharge rate was not altered appreciably in the whole-cell configuration due to rupture of the membrane or exchange of the intracellular media. In controls, the spontaneous spike discharge rate of the principal cells was 17.5⫾2.3 spikes/s (CV, 0.6⫾0.1) (Fig. 5C). One day after UVG, the discharge rate was 19.2⫾3.8 spikes/s (CV, 0.4⫾0.1) for the principal cells on the lesion side, and 12.2⫾3.2 spikes/s (CV, 0.7⫾0.1) for the principal cells on the intact side. There was no significant difference in both firing rate and regularity of the principal cells on the lesion and intact side one day after UVG compared to controls. In addition, the spontaneous spike firing for principal cells on the lesion (n⫽2) and intact sides (n⫽2) was not abolished one day after UVG by a cocktail of synaptic transmission antagonists (10 ␮M bicuculline, 1 ␮M strychnine, 10 ␮M CNQX, and 30 ␮M APV) (data not shown), suggesting that the spontaneous spike firing was generated primarily by intrinsic membrane conductances, like in H5 controls (Shao et al., 2006a). Finally, in these experiments, the spontaneous spike discharge rate increased from 14⫾2.4 spikes/s to 21⫾4.6 spikes/s (n⫽4). Three days after UVG, the spontaneous spike discharge rate for principal cells in the compensated chickens was 32.3⫾5.4 spikes/s (CV, 0.2⫾0.02) on the lesion side, and 34.6⫾9.9 spikes/s (CV, 0.4⫾0.1) on the intact side (Fig. 5C). In the uncompensated chickens, the discharge rate for principal cells on the lesion side was 37.8⫾7.5 spikes/s (CV, 0.2⫾0.06), which contrasted to the absence of spike discharge for the silent principal cells on the intact side. Indeed, in all of these experimental groups containing spontaneous spike firing cells, the discharge rate increased significantly compared to the principal cells from controls and in operated chickens one day after UVG (Fig. 5C). In addition, the CV for principal cells on the lesion side of both compensated and uncompensated chickens decreased significantly compared to those from controls, suggesting a more regular firing pattern was characteristic for principal cells on the lesion side.

996

M. Shao et al. / Neuroscience 161 (2009) 988 –1007

Fig. 5. Spontaneous spike firing pattern in principal cells before and after UVG. (A1) Spontaneous spike firing cell from H7 control. (A2) Silent cell from H7 control. Note the sEPSCs in the trace (*). (B) Percentage of spontaneous spike firing cells before and after UVG. Three days after UVG in compensated chickens, a significantly higher percentage of principal cells fired spontaneously compared to controls (P⬍0.05). Moreover, on the intact side of uncompensated chickens three days after UVG, all the recorded principal cells were silent, which was highly significantly different compared to all the other groups (P⬍0.01). (C) Firing rate for spontaneous spike firing cells. Three days after UVG, spike discharge rate was significantly higher on the lesion and intact sides in compensated chickens, as well as on the lesion side of uncompensated chickens compared to controls and one day after UVG (P⬍0.05).

Spike discharge evoked by depolarization In response to 400 ms depolarizing current pulses from resting membrane potential, most principal cells (38/41) in controls fired spikes repetitively, with the few remaining neurons generating single spikes (not shown). Routinely, the principal cells in controls generated repetitive spike firing in response to 0.2 nA or less depolarizing current injection. One day after UVG, most principal cells fired spikes repetitively on depolarization on the lesion (20/23) and intact sides (24/28), whereas three days after UVG, all the principal cells fired repetitively on depolarization on the lesion (n⫽16) and intact sides (n⫽15) of the compensated chickens, and on the lesion side of the uncompensated chickens (n⫽12). However, on the intact side, which contained only silent principal cells, repetitive spike firing on depolarization was recorded in only 54% (7/13) of these neurons after injecting 0.2 nA of current. The remaining neurons required current injections ⬎0.6 nA to fire repetitively (n⫽3), or fired only single spikes at all injected currents (n⫽3). These data indicated that the principal cells on the intact side had difficulty generating spikes on depolarization three days after UVG in uncompensated chickens. In principal cells capable of firing spikes repetitively on depolarization, there was a linear relationship between the injected current and evoked firing rate (gain). In controls, the gain was 343⫾22 spikes/s/nA (n⫽24). In addition, there were no significant changes in the gain for the principal cells one day (lesion side: 277⫾23 spikes/s/nA, n⫽17; intact side: 321⫾25 spikes/s/nA, n⫽21), three days

after UVG in compensated chickens (lesion side: 275⫾22 spikes/s/nA, n⫽15; intact side: 364⫾46 spikes/s/nA, n⫽13), or in the uncompensated chickens (lesion side: 326.9⫾44.4 spikes/s/nA, n⫽11; intact side: 375.9⫾53.3 spikes/s/nA, n⫽7). Thus, the firing gain of the principal cells on the lesion and intact sides was not modified significantly before or after UVG. However, the number of principal cells capable of firing repetitively on the intact side of the uncompensated chickens decreased, so that the global output of the tangential nucleus was likely decreased in response to receiving major depolarizing stimuli from the primary vestibular fibers, for example. Spontaneous and evoked spike waveforms The action potentials generated by spontaneous spike firing principal cells had similar waveforms, regardless of whether they were produced spontaneously or evoked by small intensity current injection (not shown). Typically, the action potential was followed by two afterhyperpolarizations (AHPs) (Table 2), but the second AHP could fail on larger current injection (Shao et al., 2006a). In contrast, silent principal cells from all experimental groups generated action potentials at higher thresholds and with smaller spike amplitudes compared to the spontaneous spike firing cells. However, no differences were detected in the spike rise time, decay time, half width, or amplitude of the AHPs from spontaneous spike firing and silent cells (Table 2). Thus, the major distinction between action potentials generated by spontaneous spike firing and silent cells involved the spike threshold and spike amplitude, not the kinetics.

M. Shao et al. / Neuroscience 161 (2009) 988 –1007

997

Table 2. Action potential waveform and AHPs in principal cells

Control Lesion side, 1 day after UVG Intact side 1 day after UVG Lesion side, compensated, 3 days after UVG Intact side, compensated, 3 days after UVG Lesion side, uncompensated, 3 days after UVG Intact side, uncompensated, 3 days after UVG

Firing pattern

Spike threshold (mV)

Spike amplitude (mV)

Spike rise time (ms)

Spike decay time (ms)

Spike width (ms)

First AHP (mV)

Second AHP (mV)

n

Silent (n⫽15) SSF (n⫽23) Silent (n⫽11) SSF (n⫽12) Silent (n⫽15) SSF (n⫽12) Silent (n⫽1) SSF (n⫽15) Silent (n⫽5) SSF (n⫽10) Silent (n⫽2) SSF (n⫽9) Silent (n⫽11) SSF (n⫽0)

⫺34.5⫾1 ⫺41.9⫾1.2 ⫺33.3⫾0.1 ⫺43.9⫾1.6 ⫺35.2⫾1.1 ⫺45.2⫾1.1 N/A ⫺42.2⫾1.2 ⫺36.8⫾1.9 ⫺44⫾1.1 N/A ⫺40.5⫾0.9 ⫺34.2⫾1.7 N/A

45.5⫾1.8 61.5⫾1.6 44.3⫾1.6 50.2⫾2.2 47.9⫾2.7 62.6⫾2.3 N/A 58⫾2.5 50.4⫾0.8 61.9⫾3.2 N/A 53.2⫾1.7 43.4⫾2.2 N/A

0.38⫾0.01

0.29⫾0.01

0.32⫾0.01

31.3⫾0.9

16.1⫾0.8 (n⫽35)

38

0.39⫾0.03

0.29⫾0.02

0.34⫾0.02

32.6⫾1.1

19.6⫾0.9

23

0.36⫾0.02

0.29⫾0.02

0.34⫾0.02

32.7⫾1.5

18.4⫾0.9 (n⫽25)

27

0.37⫾0.03

0.30⫾0.02

0.34⫾0.02

29.8⫾0.9

16.9⫾1

16

0.37⫾0.02

0.32⫾0.02

0.36⫾0.02

31.2⫾1.6

15.5⫾1.6 (n⫽13)

15

0.34⫾0.02

0.27⫾0.02

0.31⫾0.02

32.7⫾2.3

15.1⫾1.9

11

0.43⫾0.03

0.32⫾0.03

0.36⫾0.03

30.5⫾1.7

18.3⫾1.7

11

Values are presented as mean⫾SEM. No significant differences were found in SSF cells from all experimental groups. There were significant differences in spike threshold and spike amplitude between SSF and silent cells in each group (if applicable). However, no significant differences were found in spike rise time, decay time, width, and AHP amplitude between SSF and silent cells, so the data were pooled. Also, no significant differences were found in spike threshold and amplitude for SSF and silent cells, regardless of experimental group. Due to unbalanced bridge in some recordings (n⫽9), the data were excluded from analysis.

Sodium channels On exposure to 1 ␮M TTX, all spike activity in the spontaneous spike firing cells was abolished, indicating that sodium channels contributed to spike generation (not shown). Two types of sodium currents, INaP and INa, were recorded according to their voltage dependencies. INaP. On exposure to 800 ms duration depolarizing voltage steps from ⫺60 to ⫺47.5 mV, a persistent inward current was induced in the principal cells (Fig. 6A1). The inward current was abolished by TTX, indicating that it was sodium-dependent (Fig. 6A2). When the current traces were subtracted before and after TTX exposure, a slowlyactivating, inward current was revealed (Fig. 6A3). The I/V curve demonstrated that this current activated around ⫺60 mV and became larger with more depolarized voltages (Fig. 6B). The inward current has all of the characteristics of INaP, including a low-activation threshold, slow inactivation, and TTX sensitivity (Llinas, 1988). When taken together from all experimental groups, the spontaneous spike firing principal cells had steeper I/V curves for INaP and larger INaP at ⫺47.5 mV than the silent cells (Fig. 6C, D, Table 3). Thus, INaP amplitude distinguished the spontaneous spike firing and silent principal cells. INa. On exposure to short duration depolarizing voltage steps (100 ms) from ⫺60 to ⫺40 mV, the principal cells exhibited a fast-rising, fast-recovering inward current. The inward current was sensitive to 1 ␮M TTX (not shown), indicating it was an inward INa. INa was activated at ⫺50 mV in most spontaneous spike firing cells (26/45) (Fig. 6D, arrow), but was not activated in most silent principal cells (30/33) until the membrane was depolarized to ⫺40 mV (Fig. 6E, arrow). Accordingly, spontaneous spike firing and silent principal cells displayed different I/V curves, with the former exhibiting significantly larger INa at ⫺50 mV (Fig.

6F). However, once INa was activated around ⫺40 mV in silent cells, no significant difference in INa was detected between these two neuron classes. The lower INa activation threshold for the spontaneous spike firing cells made them more capable of generating action potentials with less membrane depolarization than required for silent cells. Regardless of experimental group, the spontaneous spike firing and silent cells were distinguished by their sodium conductances. In summary, INaP and INa retained their signature amplitudes in the spontaneous spike firing and silent cells, despite UVG. Potassium channels IDS. In controls, the outward current induced in principal cells could be blocked partially by the application of 200 nM DTX (Fig. 7A1, A2). Subtraction of the outward current before and after DTX exposure revealed IDS (Fig. 7A3). The I/V curve for IDS showed that this current activated around ⫺60 mV (Fig. 7B), as reported previously (Gamkrelidze et al., 1998). No significant difference in IDS, or its percentage of the total outward current, was detected in the spontaneous spike firing and silent cells from different experimental groups, so the data were pooled. The I/V curve for the principal cells on the intact side of uncompensated chickens three days after UVG was steeper compared to principal cells from all other experimental groups (Fig. 7C). In addition, IDS amplitude and its percentage of the outward current were compared at ⫺30 mV, where IDS is sufficiently large to permit comparisons with minimal interference from calcium-dependent potassium currents. In the uncompensated chickens three days after UVG, IDS amplitude and its percentage of the total outward current increased significantly in the silent principal cells on the intact side compared to principal cells from all other experimental groups (Fig. 7D, E).

998

M. Shao et al. / Neuroscience 161 (2009) 988 –1007

Fig. 6. INaP and INa in principal cells before and after UVG. (A1) Voltage-clamp recording showing a slowly-inactivated inward current when the holding potential was switched from ⫺60 to ⫺47.5 mV. (A2) The inward current was abolished by 1 ␮M TTX. (A3) Subtracting A2 from A1 revealed INaP. (B) I/V curve from principal cell in (A) showing that INaP was activated around ⫺60 mV and its amplitude increased with more depolarizing voltage commands. (C) I/V curves for INaP in principal cells from different experimental groups (see inset). * Regardless of experimental group, silent principal cells exhibited smaller amplitude INaP compared to spontaneous spike firing principal cells (P⬍0.05). (D, E) INa in spontaneous spike firing and silent principal cells. Note that in D (arrow) a large fast inward current was evoked at ⫺50 mV, but was absent in the silent cell in E (arrow). (F) At ⫺50 mV (*), there was a significant difference in INa amplitude in spontaneous spike firing and silent principal cells. SSF, spontaneous spike firing.

Kv1.1 cytoplasmic expression in MAP2-labeled principal cell bodies. Since silent principal cells on the intact side in uncompensated chickens three days after UVG showed significantly increased IDS, Kv1.1, a potassium channel subunit related to IDS, was tested for its expres-

sion. In H7 controls, Kv1.1 immunolabeling of the principal cells was similar to that reported at H9 (Popratiloff et al., 2003). The mean pixel brightness for Kv1.1 in the principal cell body cytoplasm was 39.7⫾1.5 (n⫽98 profiles) (Fig. 8A–C, J) in controls, 36.7⫾1.1 (n⫽134 profiles) on the

M. Shao et al. / Neuroscience 161 (2009) 988 –1007 Table 3. INaP amplitude in principal cells at ⫺47.5 mV

Control Lesion side, 1 day after UVG Intact side, 1 day after UVG Lesion side, compensated 3 days after UVG Intact side, compensated 3 days after UVG Lesion side, uncompensated 3 days after UVG Intact side, uncompensated 3 days after UVG

Firing pattern

INaP amplitude (pA)

n

Silent SSF Silent SSF Silent SSF SSF

37.5⫾12.5 162.7⫾17.2 32.3⫾11.6 102.3⫾14.4 38.9⫾11.4 182.7⫾51.3 138.7⫾34.3

14 18 6 9 9 7 9

Silent SSF SSF

56.3⫾9 117⫾18.1 145.4⫾27.7

4 8 9

Silent

27.1⫾8.5

7

Value are presented as mean⫾SEM. Significant differences were found between SSF and silent principal cells from all experimental groups, when applicable (P⬍0.05).

lesion side, and 38.3⫾1.0 (n⫽120 profiles) on the intact side of sham-operated chickens, with no significant differences (Fig. 8J). Three days after UVG, the cytoplasmic signal for Kv.1.1 immunolabeling in the principal cell bodies was significantly stronger on the lesion (59.2⫾1.7; n⫽59 profiles) and intact sides (59.7⫾2.8; n⫽34 profiles) of uncompensated chickens compared to controls (P⬍ 0.01; Kruskal–Wallis test) (Fig. 8D–J), with no significant difference between the two sides. The lack of detectable changes in the sham-operated chickens suggested that the findings after UVG were related to vestibular deafferentation, and not due to the effects of anesthesia or stress related to surgery. Using conventional fluorescence imaging in normal H9 chickens, differential Kv1.1 immunolabeling of individual principal cell bodies is apparent, with most principal cells displaying weaker immunolabeling, but a subset showing stronger immunolabeling (Popratiloff et al., 2003). A similar pattern of distribution for Kv1.1 immunolabeling was plotted from the present sample of principal cell bodies from controls, using the higher detection sensitivity provided by confocal imaging (Fig. 8K). However, the distribution of principal cell body labeling with Kv1.1 on the lesion side of uncompensated chickens followed a different pattern, with uniform labeling concentrated at the peak for the higher pixel brightness, whereas the principal cell bodies from the intact side preserved the dual distribution seen in the principal cell bodies from controls, with a shift of the peaks toward higher pixel brightness (Fig. 8K). Thus, like principal cell bodies on the intact side, those on the lesion side showed increased pixel brightness in the cytoplasm, but also showed a major shift in the distribution pattern of pixel brightness from a bimodal to a unimodal distribution. Kv1.1 surface expression in MAP2-labeled principal cell bodies. While the present analysis provided a good estimate of the overall cytoplasmic expression of Kv1.1 in the principal cell bodies, Kv1.1 surface expression could not be evaluated from these data. To test the surface

999

expression of Kv1.1 in the principal cell bodies, equal numbers of principal cell bodies were selected randomly for measurements from controls (n⫽21), lesion (n⫽21), and intact sides (n⫽20) of the uncompensated chickens (Fig. 8L–O). To measure pixel brightness, a line of one pixel thickness (0.31 ␮m) was placed along the surface of a MAP2-labeled principal cell body and then transferred to the corresponding Kv1.1 image (Fig. 8L). In addition, line scans were drawn at random in the cytoplasm and in the nucleus. Each line scan of the cell surface showed large variability in pixel brightness, with prominent peaks corresponding to large immunolabeled clusters, or smaller peaks corresponding to small immunolabeled patches (Fig. 8M). Mean pixel brightness of the surface line scans was calculated and normalized by subtracting the mean pixel brightness of the line scans from the nucleus. The mean pixel brightness of the line scans from the principal cell body surfaces was 26⫾2 in controls, 39⫾3 on the lesion side, and 49⫾4 on the intact side of the uncompensated chickens (Fig. 8N). Highly significant differences were detected on comparing the cell surface line scans from principal cells in controls to those obtained on the lesion and intact sides of the uncompensated chickens (P⬍0.001, independent sample t-test), and significant differences were found between the lesion and intact sides (P⬍0.05, independent sample t-test). To obtain a measure of the efficiency of Kv1.1 surface expression from the cytoplasmic synthetic pool, the ratio between the line scans taken from the surface and cytoplasm was compared (Fig. 8O). The principal cell bodies from the lesion side had similar surface/cytoplasm labeling ratio as controls (0.64⫾0.01 and 0.67⫾0.01, respectively). In contrast, the principal cell bodies from the intact side of the uncompensated chickens showed significantly higher surface/cytoplasm labeling ratio (0.79⫾0.02) compared to both the lesion side and controls (P⬍0.001, independent sample t-test). Finally, like in the normal H9 chickens, Kv1.1 expression decreased to background levels outside the principal cell bodies, which was defined by the absence of MAP2-labeling. Kv1.1 expression in the axonal initial segment was not quantified because the axon initial segment in these preparations was not revealed reliably due to the gradual disappearance of MAP2 signal within the axons. Excitatory synaptic transmission sEPSC and mEPSC frequencies. Using KMeSO4 pipet solution at ⫺60 mV, all the spontaneous inward currents generated in the principal cells were blocked by 10 ␮M CNQX, suggesting that they were AMPA/KA receptors-mediated sEPSCs. In controls, sEPSC frequency from principal cells was 1.2⫾0.2 Hz (n⫽34), which is similar to that reported at H1 (1.45⫾0.22 Hz) (Shao et al., 2004). However, there was a tendency for the silent cells to show higher sEPSC frequency (1.82⫾0.6 Hz, n⫽12) than the spontaneous spike firing cells (0.86⫾0.1 Hz; n⫽22) (P⫽0.06). TTX (1 ␮m) was applied to determine what percentage of sEPSCs was evoked by presynaptic firing of actions potentials. TTX-

1000

M. Shao et al. / Neuroscience 161 (2009) 988 –1007

Fig. 7. IDS in principal cells before and after UVG. (A1) Outward currents induced in H7 control principal cell, which fired spikes spontaneously. ACSF contained low calcium (0.05 mM) and 1 ␮M TTX. (A2) Outward currents were abolished partially on exposure to 200 nM ␣-DTX. (A3) Subtracting A2 from A1 revealed IDS. (B) I/V curves obtained from cell in A. (C) I/V curves for IDS in principal cells. (D) IDS amplitude at ⫺30 mV. (E) IDS percentage of total outward current at ⫺30 mV. IDS amplitude and its percentage increased significantly on the intact side in uncompensated chickens three days after UVG compared to all the other experimental groups (P⬍0.05).

resistant, miniature excitatory postsynaptic currents (mEPSCs) composed 39%⫾16% (n⫽3) of sEPSCs, which is similar to that reported for principal cells at H1 (32%⫾5%). Thus, 61% of sEPSCs were generated by presynaptic firing of action potentials, indicating the presence in the brain slices of sig-

nificant numbers of presynaptic neurons capable of generating spikes in principal cells. One day after UVG, sEPSC frequency in the principal cells on the lesion side was 1.91⫾0.3 Hz (n⫽19) and 1.03⫾0.2 Hz on the intact side (n⫽20), with a significant

M. Shao et al. / Neuroscience 161 (2009) 988 –1007

1001

Fig. 8. A–C, High power (60⫻) confocal images of principal cells from control chicken (H7), and the lesion (D–F) and intact sides (G–I) from an uncompensated chicken three days after UVG. (A, D, G) MAP2-labeled principal cells. (B, E, H) Kv1.1-labeled principal cells. (C, F, I) Digitally-merged channels showing principal cells labeled with MAP2 and Kv1.1. (J) Mean pixel brightness (⫾SEM) for Kv1.1 in principal cell bodies from controls, sham-operated and lesion and intact sides of uncompensated chickens. Kv1.1 labeling was higher in principal cells on the lesion and intact sides after UVG, compared to controls and sham-operated chickens (P⬍0.01). No difference was detected in Kv1.1 labeling for cells on the lesion and intact sides. (K) Pixel brightness for Kv1.1 labeling in the cytoplasm of principal cell bodies, plotted into bins from controls and the lesion and intact sides of uncompensated chickens three days after UVG. The distribution of Kv1.1 labeling in principal cells from controls showed a nearly normal distribution with a slight shoulder representing a cluster of brighter cells at the higher end of pixel brightness. The distribution of Kv1.1 labeling in principal cells from the lesion side was unimodal with a peak at a higher pixel brightness level than in controls. Kv1.1 distribution was more complex on the intact side of operated chickens compared to controls, showing two distinct peaks with greater pixel brightness. (L) Confocal image of a typical principal cell body used to measure Kv1.1 cell surface expression from H7 controls. (M) Plot of pixel brightness along lines drawn in the cytoplasm, cell surface, and nucleus of the principal cell in L. (N) Mean pixel brightness from line scans drawn along the cell surface of principal cell bodies from controls, and intact and lesion sides of uncompensated chickens three days after UVG. (O) Ratio for Kv1.1 cell surface/cytoplasmic expression in principal cells from controls, and the intact and lesion sides of uncompensated chickens three days after UVG. Kv1.1 cell surface expression increased significantly in principal cells on the intact side (P⬍0.05).

1002

M. Shao et al. / Neuroscience 161 (2009) 988 –1007

Fig. 9. (A) sEPSC frequency in principal cells from controls, and the lesion and intact sides one day after UVG, and three days after UVG for both compensated and uncompensated chickens. There was a significant increase in sEPSC frequency on the lesion sides compared to controls in uncompensated chickens after UVG. (B) Kv1.2 immunolabeling in synaptotagmin-positive terminal profiles on principal cell bodies from controls (B–D), and the lesion (E–G) and intact sides (H–J) of uncompensated chickens three days after UVG. Kv1.2 immunolabeling decreased significantly in the perisomatic terminal profiles on the lesion side compared to controls (P⬍0.001) and intact side (P⬍0.05). (K) Mean pixel brightness (⫾SEM) for Kv1.2 in synaptotagmin-positive terminal profiles on principal cell bodies from controls, sham-operated, and uncompensated chickens three days after UVG. Kv1.2 decreased significantly in the profiles on the lesion side three days after UVG in the uncompensated chickens. (L) Histogram plotting Kv1.2 pixel brightness in synaptotagmin-positive terminal profiles contacting the principal cell bodies in controls, lesion, and intact side of uncompensated chickens three days after UVG. The distribution of Kv1.2 immuno labeling in the terminal profiles contacting the principal cell bodies in controls showed a normal distribution, as found on the intact side. However, on the lesion side, most terminals contained significantly lower Kv1.2 expression.

difference between the two sides (P⬍0.05) (Fig. 9A). By three days after UVG, sEPSC frequency in the principal cells on the lesion side of the compensated chickens was 1.75⫾0.40 Hz (n⫽15) and 1.61⫾0.6 Hz on the intact side (n⫽14), with no significant difference. In contrast, in the

uncompensated chickens three days after UVG, sEPSC frequency in the principal cells on the lesion side was 6.8⫾2.7 Hz (n⫽8), while the frequency on the intact side was 0.7⫾0.2 Hz (n⫽10) (Fig. 9A). mEPSCs composed 57%⫾2% of sEPSCs (n⫽4) on the lesion side. Since the

M. Shao et al. / Neuroscience 161 (2009) 988 –1007

sEPSC frequency was low on the intact side, no measurement of mEPSC frequency was made (n⫽10). Accordingly, presynaptic glutamate release decreased on the intact side one day after UVG, but recovered by three days to the level recorded in the principal cells from controls and the lesion side in the compensated chickens. In uncompensated animals three days after UVG, it was interesting that the frequency of both sEPSCs and mEPSCs in the principal cells on the lesion side increased significantly compared to controls and the intact side, indicating increased presynaptic glutamate release. Kv1.2 expression in synaptotagmin-positive profiles contacting the principal cell bodies. In other systems, Kv1.2 plays a critical role in presynaptic transmitter release (Dodson et al., 2003), raising the possibility that changes in Kv1.2 expression could contribute to the increased sEPSC frequency recorded in the principal cells on the lesion side of uncompensated chickens three days after UVG. Synaptotagmin labeled the terminal profiles contacting the principal cell bodies in controls, sham-operated, and on the lesion and intact sides of the uncompensated chickens. In H7 controls, the pattern of Kv1.2 and synaptotagmin immunolabeling was similar to that reported for H9 chickens (Popratiloff et al., 2003). Kv1.2 immunolabeling was concentrated around the principal cell bodies and co-localized with the synaptotagmin-positive profiles (Fig. 9B–D). Like H9 chickens, Kv1.2 immunolabeling in the principal cell bodies was present at background levels (Fig. 9C). In addition, Kv1.2 immunolabeling was detected in distinct clusters at the juxtaparanodes of axons coursing through the tangential nucleus (Fig. 9C). In controls, there was less overlap between Kv1.2 immunolabeling and synaptotagmin-labeled terminal profiles in the tangential nucleus neuropil compared to that found around principal cell bodies (Fig. 9D). Specifically, the mean pixel brightness for Kv1.2 in the synaptotagmin-labeled terminal profiles in the neuropil from controls was 18⫾0.9 (n⫽408 profiles), with about 35% of the terminal profiles containing no signal (Fig. 9D). Overall, there was a striking decrease in Kv1.2 immunolabeling of the juxtaparanodes from fibers coursing through the neuropil of the tangential nucleus on the lesion side (Fig. 9E–G). Since no change was detected on the intact side (Fig. 9H–J), this decrease could be due to

1003

degenerative changes occurring within the primary vestibular fibers whose ganglion cell bodies of origin were removed (Aldrich and Peusner, 2002). Kv1.2 expression was quantified in the synaptotagminpositive profiles contacting the principal cell bodies by outlining the individual profiles and measuring the mean pixel brightness for Kv1.2 in each profile (Fig. 9K, L). Mean pixel brightness for these profiles was 33⫾0.6 (n⫽531 profiles) in controls, 32⫾0.5 (n⫽661 profiles) for profiles on the lesion side of sham-operated chickens, and 33⫾0.5 (n⫽552 profiles) for profiles on the intact side of shamoperated chickens, with no significant differences (Kruskal– Wallis, P⬎0.05). However, in the uncompensated chickens three days after UVG, the mean pixel brightness for Kv1.2 immunolabeling in the synaptotagmin-positive profiles was 28⫾0.7 (n⫽425 profiles) for profiles on the lesion side and 33⫾0.5 (n⫽778 profiles) for profiles on the intact side. No significant difference was detected between the profiles on the intact side and controls (Kruskal–Wallis, P⬎0.05). However, there was a highly significant difference between the profiles on the lesion side and controls (Kruskal–Wallis, P⬍0.001). Accordingly, the lesion side contained a population of synaptotagmin-positive profiles with lower Kv1.2 levels.

DISCUSSION Profound changes were recorded in the spontaneous spike activity, ionic conductances and/or sEPSC frequency of tangential principal cells on the lesion and intact sides at the selected stages (Fig. 10). Furthermore, using immunolabeling and confocal imaging, the DTX-sensitive potassium channel, IDS, whose conductance critically changed in neurons from uncompensated chickens, exhibited major changes in the level of subunit expression. Is there a direct link between behavioral recovery and electrophysiological properties of tangential principal cells after vestibular deafferentation? Like MVN neurons from rat and guinea pig studied in vitro after labyrinthectomy (Him and Dutia, 2001; Beraneck et al., 2003), the principal cells increased their input resistance and spontaneous spike discharge rate on the lesion side in both compensated and uncompensated chickens

Fig. 10. Summary diagram showing the major changes found in percentage of spontaneous spike firing (SSF) principal cells, firing pattern, spike discharge rate for spontaneous spike firing cells only, and sEPSC frequency for principal cells from controls, one days after UVG, and three days after UVG in compensated and uncompensated chickens.

1004

M. Shao et al. / Neuroscience 161 (2009) 988 –1007

three days after UVG, at the same time that behavioral recovery was initiated. However, unlike MVN neurons in rat and guinea pig (Cameron and Dutia, 1997; Beraneck et al., 2004), the principal cells on the intact side of compensated chickens increased their spontaneous spike discharge rates. Like rat MVN neurons one day after labyrinthectomy (Guilding and Dutia, 2005), the spontaneous spike discharge rate in principal cells one day after UVG was not abolished on exposure to a cocktail of antagonists for synaptic transmission. This suggests that the spontaneous spike firing is relatively independent of synaptic inputs at this early stage. The increased firing rate observed in the principal cells could be due to a bicuculline block of the SK channel (Seutin and Johnson, 1999; Shao et al., 2006a). Despite partial coincidence in the timing of behavioral recovery and restoration of spontaneous spike firing in MVN neurons in rodents studied in vivo, a time-locked sequence of events does not exist (Ris et al., 1997; for review, see Straka et al., 2005). Indeed, behavioral adaptation after vestibular deafferentation is not explained exclusively by changes in vestibular nucleus neurons. At present, comparisons between in vivo and in vitro properties of the tangential principal cells after vestibular deafferentation cannot be made, due to lack of in vivo electrophysiological recordings. However, in the chicken, a dissociation between behavior and in vitro electrophysiological changes in the principal cells was observed. For example, one day after UVG, the spontaneous spike firing was symmetric on both sides, and the operated chickens exhibited severe postural deficits. In contrast, three days after UVG in the uncompensated chickens, the principal cells exhibited gross asymmetry in spike discharge rates bilaterally, but their head position was similar to that found one day after UVG. The interactions between behavior and neuronal excitability are likely reciprocal. For example, after a lesion, changes in signal processing within the central vestibular circuitry could modify the observed behavior, which in turn could change the level of stimuli received by the vestibular circuitry. This work provides a model for studying the role of VOR and VOC neurons during vestibular compensation. To determine the exact role of the principal cells in the recovery process, further identification of their neural circuitry is essential. In addition, other vestibular reflex projection neurons participating in the VOR and VOC circuitries independent of the principal cells, as well as brainstem, spinal cord, and cerebellar neurons which modulate the various vestibular circuitries must be identified under normal conditions and after lesions. Spontaneous spike firing and silent cells Although some principal cells first acquire spontaneous spike firing at hatching, others remain silent (Shao et al., 2006a). In the present study, the silent cells were not a product of hyperpolarized resting membrane potential, since most silent principal cells were incapable of firing action potentials, despite driving their membrane potential up to ⫺50 mV. Both spontaneous spike firing and silent principal cells had resting membrane potentials and input

resistances in the range of healthy neurons. The silent cells were not a by-product of preparing the brain slices, since all the principal cells on the intact side of the uncompensated chickens were silent three days after the lesion. Finally, silent vestibular nuclei neurons are recorded under diverse experimental conditions, including from whole adult cats in vivo (Shimazu and Precht, 1965), whole adult guinea-pig brains in vitro (Babalian et al., 1997), and vestibular deafferented whole adult guinea pig in vivo (Ris et al., 1995, 1997). Moreover, silent cells are recorded from brain slices of cerebellum (McKay and Turner, 2005) and auditory nuclei (Kim and Trussell, 2007), to name a few. Since no signs of degenerating principal cells were seen in the present or a previous study (Aldrich and Peusner, 2002), shifts in the percentage of spontaneous spike firing and silent cells must be due to transformation of the existing principal cell pools. Indeed, changes in the percentage of silent vestibular nucleus neurons are recorded after labyrinthectomy performed on adult animals in vivo (Precht et al., 1966; Ris et al., 1997) and in vitro (Beraneck et al., 2004). Thus, detecting both spontaneous spike firing and silent cells within the same neuron class in normal subjects suggests that neuron subsets coexist under normal conditions and can transform bidirectionally due to changing stimuli and/or under pathophysiological conditions (for review, see Darlington et al., 2001). What determines whether a neuron fires spontaneously or remains silent? Spontaneous spike activity is generated by the spatiotemporal integration of excitatory and inhibitory synaptic inputs and/or intrinsic membrane conductances (Shao et al., 2006a). Here, spontaneous spike firing and silent principal cells differed in INaP and the activation threshold for INa. The higher INa activation threshold, which is consistent with a depolarized threshold for action potential generation, together with the smaller amplitude of INaP, decreased the probability of firing spikes in the silent cells. In other systems, INaP plays a critical role in modulating locomotion (Tazerart et al., 2008) or respiratory rhythms (Del Negro et al., 2005). Moreover, sodium channelopathies occur when normal sodium channels are expressed at abnormal levels, resulting in diverse symptoms, including seizures and ataxia (Wada, 2006). While no changes were detected in sodium channels at the transcriptional level in MVN neurons after labyrinthectomy (Patkó et al., 2003; Vassias et al., 2003), sodium channels integrated into the surface membranes were not studied. Perhaps a simple cellular mechanism, such as changes in the sodium channel density and/or subunit composition in the soma or in the axon, could be responsible for transforming silent and spontaneous spike firing cells (Khaliq and Raman, 2006). While at present it is uncertain what role silent cells play in central vestibular signal processing, clearly spontaneous spike firing and silent cells cooperate in signal processing after lesions. Unique properties of principal cells recorded from uncompensated chickens Compared to controls and compensated chickens, the principal cells in the uncompensated chickens showed

M. Shao et al. / Neuroscience 161 (2009) 988 –1007

several unique properties. First, all the principal cells on the intact side were silent. Second, besides sodium conductances, potassium channels were critically involved in changing the evoked spike firing properties of these neurons. Third, sEPSCs frequency increased almost fourfold in principal cells on the lesion side of the uncompensated chickens. Silent principal cells on the intact side of the uncompensated chickens had dramatically larger IDS compared to both spontaneous spike firing and silent principal cells in all other experimental groups. IDS is activated at a subthreshold voltage for spike generation, and influences the ability of neurons to generate action potentials (Bean, 2007). In the uncompensated chickens three days after UVG, half of the principal cells on the intact side exhibited either a single spike on depolarization or had difficulty generating spikes on depolarization. The inability of principal cells to fire spikes on depolarization is reminiscent of developing principal cells at E16, when most of them do not generate repetitive spike firing on depolarization and display very large IDS (Gamkrelidze et al., 1998). Not surprising, the silent principal cells on the intact side had larger IDS. In support of this, confocal imaging of Kv1.1 immunolabeling showed that the surface/cytoplasmic expression ratio increased significantly for principal cells on the intact side compared to principal cells in controls and on the lesion side of uncompensated chickens. Subcellular compartmentalization of Kv1.1 immunolabeling was preserved after UVG, but the distribution of neuronal labeling intensities switched from bimodal to unimodal on the lesion side (Fig. 8B). Accordingly, bimodal distribution of neuronal labeling intensities for Kv1.1 among principal cells could reflect the topographic distribution of the primary vestibular fibers in the tangential nucleus (Cox and Peusner, 1990a; Popratiloff and Peusner, 2007). Three days after UVG in uncompensated chickens, Kv1.1 surface expression increased bilaterally in the principal cell bodies, but the surface/cytoplasmic ratio increased only on the intact side, suggesting that different signaling mechanisms governed Kv1.1 expression bilaterally. The composition of Kv1 subunits is relevant for determining the efficiency of its surface expression. For example, Kv1.1 surface expression is low if Kv1.1 homomers are transfected to COS-1 cells, but its expression increases if the cells are co-transfected with Kv1.2 or Kv.1.4 subunits (Manganas and Trimmer, 2000, 2004; Manganas et al., 2001). Since in the present study Kv1.2 immunolabeling was weak in the principal cell bodies, it is possible that other Kv1 subunits may be involved in generating the higher efficiency of Kv1.1 surface expression observed in principal cells on the intact side. Increased sEPSC frequencies suggest increased presynaptic glutamate release at the principal cells. From immunolabeling and confocal imaging, Kv1.2 expression decreased significantly in the synaptic terminal profiles surrounding the principal cell bodies on the lesion side in the uncompensated chickens. The perisomatic Kv1.2 immunolabeling co-localized with synaptotagmin-labeled small terminals contacting the principal cell bodies rather

1005

than the large spoon terminals. The numerous small, synaptotagmin-positive terminals on the principal cell bodies most likely originate from neurons in high cervical spinal cord (Gross, 1985; Cox and Peusner, 1990b), rather than the primary vestibular fibers, which are distributed primarily on dendrites of neurons in the tangential nucleus (Popratiloff et al., 2004). Indeed, increased presynaptic glutamate release likely originates from the existing terminals, since no new terminals were detected on the principal cell bodies from one to 56 days after UVG (Aldrich and Peusner, 2002). Presynaptic Kv1.2 could play a pivotal role in tuning the excitability of the existing terminals in the tangential nucleus. For example, blocking Kv1.2 channels on the calyx of Held increases its excitability and results in aberrant spike-firing and multiple responses in the postsynaptic neurons (Dodson et al., 2003; Ishikawa et al., 2003). Accordingly, decreased Kv1.2 expression in the presynaptic terminals contacting the principal cell bodies on the lesion side could be related to the dramatically increased sEPSC frequency recorded.

CONCLUSION In conclusion, after a one-sided vestibular deafferentation, behavioral recovery coincided with the presence of symmetric spontaneous spike discharge rates bilaterally and the restoration of symmetric sEPSC frequencies bilaterally within a single class of vestibular nucleus neurons. In contrast, lack of symmetry in these properties bilaterally was found in the uncompensated chickens. Certain intrinsic ionic conductances crucial for the maturation of spike firing pattern played important roles during this period of early adaptation to vestibular deafferentation. Immunolabeling for Kv1 subunits strongly supported the electrophysiological findings. Identifying the relationship between increased sEPSC frequency and changes in the spontaneous spike discharge rate of vestibular nucleus neurons at specific stages of vestibular compensation is a fundamental step for understanding why recovery occurs. Acknowledgments—This work was supported by NIH grant R01 DC00970. We would like to thank Mr. Xudong Cai for his excellent technical assistance. J. Yi and A. Lerner received support from GWUMC in the form of Gill Fellowships to work on these studies during the summer of 2006.

REFERENCES Aldrich EM, Peusner KD (2002) Vestibular compensation after ganglionectomy: ultrastructural study of the tangential vestibular nucleus and behavioral study of the hatchling chick. J Neurosci Res 67:122–138. Babalian A, Vibert N, Assie G, Serafin M, Mühlethaler M, Vidal PP (1997) Central vestibular networks in the guinea-pig: functional characterization in the isolated whole brain in vitro. Neuroscience 81:405– 426. Bean BP (2007) The action potential in mammalian central neurons. Nat Neurosci 8:451– 465. Beraneck M, Hachemaoui M, Idoux E, Ris L, Uno A, Godaux E, Vidal PP, Moore LE, Vibert N (2003) Long-term plasticity of ipsilesional medial vestibular nucleus neurons after unilateral labyrinthectomy. J Neurophysiol 90:184 –203.

1006

M. Shao et al. / Neuroscience 161 (2009) 988 –1007

Beraneck M, Idoux E, Uno A, Vidal PP, Moore LE, Vibert N (2004) Unilateral labyrinthectomy modifies the membrane properties of contralesional vestibular neurons. J Neurophysiol 92:1668 –1684. Burette A, Khatri L, Wyszynski M, Sheng M, Ziff EB, Weinberg RJ (2001) Differential cellular and subcellular localization of AMPA receptor-binding protein and glutamate receptor interacting protein. J Neurosci 21:495–503. Carrascal L, Nieto-Gonzales JL, Cameron WE, Torres B, NunezAbades PA (2005) Changes during the postnatal development in physiological and anatomical characteristics of rat motoneurons studied in vitro. Brain Res Brain Res Rev 49:377–387. Cameron SA, Dutia MB (1997) Cellular basis of vestibular compensation: changes in intrinsic excitability of MVN neurones. Neuroreport 8:2595–2599. Cox RG, Peusner KD (1990a) Horseradish peroxidase labeling of the efferent and afferent pathways of the avian tangential vestibular nucleus. J Comp Neurol 296:324 –341. Cox RG, Peusner KD (1990b) Horseradish peroxidase labeling of the central pathways in the medulla of the ampullary nerves in the chicken, Gallus gallus. J Comp Neurol 297:564 –581. Curthoys IS, Halmagyi GM (2007) Vestibular compensation: clinical changes in vestibular function with time after unilateral vestibular loss. In: Vestibular rehabilitation, 3rd ed (Herdman SJ, ed). Philadelphia, PA: F.A. Davis Company. Darlington CL, Dutia MB, Smith PF (2001) The contribution of the intrinsic excitability of vestibular nucleus neurons to recovery from vestibular damage. Eur J Neurosci 15:1719 –1727. Del Negro CA, Morgado-Valle C, Hayes JA, Mackay DD, Pace RW, Crowder EA, Feldman JL (2005) Sodium and calcium currentmediated pacemaker neurons and respiratory rhythm generation. J Neurosci 25:446 – 453. Dodson PD, Billups B, Rusznák Z, Szûcs G, Barker MC, Forsythe ID (2003) Presynaptic rat Kv1.2 channels suppress synaptic terminal hyperexcitability following action potential invasion. J Physiol 550: 27–33. Evinger C, Erichsen JT (1986) Transsynaptic retrograde transport of fragment C of tetanus toxin demonstrated by immunohistochemical localization. Brain Res 380:383–388. Francis HW, Manis PB (2000) Effects of deafferentation on the electrophysiology of ventral cochlear nucleus neurons. Hear Res 149:91–105. Gamkrelidze G, Giaume C, Peusner KD (1998) The differential expression of low-threshold sustained potassium current contributes to the distinct firing patterns in embryonic central vestibular neurons. J Neurosci 18:1449 –1464. Gamkrelidze G, Giaume C, Peusner KD (2000) Firing properties and dendrotoxin-sensitive sustained potassium current in vestibular nuclei neurons of the hatchling chick. Exp Brain Res 134:398 – 401. Gottesman-Davis A, Peusner KD (2008) Identification of vestibuloocular projection neurons in chick MVN. Soc Neurosci Abst. Gross GH (1985) Innervation of the complexus (“hatching”) muscle of the chick. J Comp Neurol 232:180 –189. Guilding C, Dutia MB (2005) Early and late changes in vestibular neuronal excitability after deafferentiation. Neuroreport 16:1415– 1418. Him A, Dutia MB (2001) Intrinsic excitability changes in vestibular nucleus neurons after unilateral deafferentation. Brain Res 908: 58 – 66. Ishikawa T, Nakamura Y, Saitoh N, Li WB, Iwasaki S, Takahashi T (2003) Distinct roles of Kv1 and Kv3 potassium channels at the calyx of Held presynaptic terminal. J Neurosci 23:10445–10453. Kaga K (1999) Vestibular compensation in infants and children with congenital and acquired vestibular loss in both ears. Int J Pediatr Otorhinolaryngol 49:215–224. Khaliq ZM, Raman IM (2006) Relative contributions of axonal and somatic Na channels to action potential initiation in cerebellar Purkinje neurons. J Neurosci 26:1935–1944.

Kim Y, Trussell LO (2007) Ion channels generating complex spikes in cartwheel cells of the dorsal cochlear nucleus. J Neurophysiol 97:1705–1725. Labandeira-Garcia JL, Guerra-Seijas MJ, Labandeira-Garcia JA, Jorge-Barreiro FJ (1989) Afferent connections of the oculomotor nucleus in the chick. J Comp Neurol 282:523–534. Labandeira-Garcia JL, Guerra-Seijas MJ, Castroviego-Bolivar MA, Jorge-Barreiro FJ (1991) A comparative study of the tangentialoculomotor projection. J Hirnforsch 32:203–209. Li H, Godfrey DA, Rubin AM (1995) Comparison of surgeries for removal of primary vestibular inputs: a combined anatomical and behavioral study in rats. Laryngoscope 105:417– 424. Llinas RR (1988) The intrinsic electrophysiological properties of mammalian neurons: insights into central nervous system function. Science 242:1654 –1664. Mandl G, Melvill Jones G, Cynader M (1981) Adaptability of the vestibulo-ocular reflex to vision reversal in strobe reared cats. Brain Res 209:35– 45. Manganas LN, Trimmer JS (2000) Subunit composition determines Kv1 potassium channel surface expression. J Biol Chem 275:29685– 29693. Manganas LN, Trimmer JS (2004) Calnexin regulates mammalian Kv1 channel trafficking. Biochem Biophys Res Commun 322:577–584. Manganas LN, Wang Q, Scannevin RH, Antonucci DE, Rhodes KJ, Trimmer JS (2001) Identification of a trafficking determinant localized to the Kv1 potassium channel pore. Proc Natl Acad Sci U S A 98:14055–14059. McCrea RA, Strassman A, Highstein SM (1987) Anatomical and physiological characteristics of vestibular neurons mediating the vertical vestibulo-ocular reflexes of the squirrel monkey. J Comp Neurol 264:571–594. McKay BE, Turner RW (2005) Physiological and morphological development of the rat cerebellar Purkinje cell. J Physiol 567:829 – 850. Melone M, Burette A, Weinberg RJ (2005) Light microscopic identification and immunocytochemical characterization of glutamatergic synapses in brain sections. J Comp Neurol 492:495–509. National Research Council (2003) Guidelines for the care and use of mammals in neuroscience and behavioral research. Washington, DC: National Academy Press. Patkó T, Vassias I, Vidal PP, De Waele C (2003) Modulation of the voltage-gated sodium- and calcium-dependent potassium channels in rat vestibular and facial nuclei after unilateral labyrinthectomy and facial nerve transsection: an in situ hybridization study. Neuroscience 117:265–280. Petursdottir G (1990) Vestibuloocular projections in the 11-day chicken embryo: pathway specificity. J Comp Neurol 297:283–297. Peusner KD, Giaume C (1994) The first developing “mixed” synapses between vestibular sensory neurons mediate glutamate chemical transmission. Neuroscience 58:99 –113. Peusner KD, Shao M, Hirsch JC (2007) Spontaneous spike-firing in chick vestibular nucleus neurons after unilateral vestibular-nerve deafferentation. Soc Neurosci Abstr 37:16, 180. Pollack SM, Poratiloff A, Peusner KD (2004) Vestibular ganglionectomy and otolith nerve identification in the hatchling chicken. J Neurosci Methods 138:149 –155. Popratiloff A, Giaume C, Peusner KD (2003) Developmental change in expression and subcellular localization of two shaker-related potassium channel proteins (Kv1.1 and Kv1.2) in the chick tangential vestibular nucleus. J Comp Neurol 461:466 – 482. Popratiloff A, Peusner KD (2007) Otolith fibers and terminals in chick vestibular nuclei. J Comp Neurol 502:19 –37. Popratiloff A, Wang YX, Narvid J, Petralia RS, Giaume C, Peusner KD (2004) AMPA receptor subunit expression in chick vestibular nucleus neurons. J Neurosci Res 76:662– 677. Precht W, Shimazu H, Markham CH (1966) A mechanism of central compensation of vestibular function following hemilabyrinthectomy. J Neurophysiol 29:996 –1010.

M. Shao et al. / Neuroscience 161 (2009) 988 –1007 Ris L, deWaele C, Serafin M, Vidal P-P, Godaux E (1995) Neuronal activity in the ipsilateral vestibular nucleus following unilateral labyrinthectomy in the alert guinea pig. J Neurophysiol 74:2087–2099. Ris L, Capron B, de Waele C, Vidal P-P, Godaux E (1997) Dissociations between behavioral recovery and restoration of vestibular activity in the unilabyrinthectomized guinea-pig. J Physiol 500:509 – 522. Seutin V, Johnson SW (1999) Recent advances in the pharmacology of quaternary salts of bicuculline. Trends Pharmacol Sci 20:268 – 270. Shao M, Hirsch JC, Giaume C, Peusner KD (2003) Spontaneous synaptic activity is primarily GABAergic in vestibular nucleus neurons of the chick embryo. J Neurophysiol 90:1182–1192. Shao M, Hirsch JC, Giaume C, Peusner KD (2004) Spontaneous synaptic activity in chick vestibular nucleus neurons during the perinatal period. Neuroscience 127:81–90. Shao M, Hirsch JC, Peusner KD (2006a) Maturation of firing pattern in chick vestibular nucleus neurons. Neuroscience 141:711–726. Shao M, Hirsch JC, Peusner KD (2006b) Emergence of action potential generation and synaptic transmission in vestibular nucleus neurons. J Neurophysiol 96:1215–1226. Shao M, Hirsch JC, Peusner KD (2007) Intrinsic membrane properties of chick vestibular nucleus neurons after unilateral vestibular-nerve deafferentation. Soc Neurosci Abstr 37:17, 180. Shimazu H, Precht W (1965) Tonic and kinetic responses of cat’s vestibular neurons to horizontal angular acceleration. J Neurophysiol 28:1014 –1028. Sirkin DW, Precht W, Courjon JH (1984) Initial, rapid phase of recovery from unilateral vestibular lesion in rat not dependent on survival of central portion of vestibular nerve. Brain Res 302:245–256.

1007

Smith PF, Curthoys IS (1988a) Neuronal activity in the ipsilateral medial vestibular nucleus of the guinea pig following unilateral labyrinthectomy. Brain Res 444:308 –319. Smith PF, Curthoys IS (1988b) Neuronal activity in the contralateral medial vestibular nucleus of the guinea pig following unilateral labyrinthectomy. Brain Res 444:295–307. Smith MR, Nelson AB, Du Lac S (2002) Regulation of firing response gain by calcium-dependent mechanisms in vestibular nucleus neurons. J Neurophysiol 87:2031–2042. Straka H, Vibert N, Vidal PP, Moore LE, Dutia MB (2005) Intrinsic membrane properties of vertebrate vestibular neurons: function, development and plasticity. Prog Neurobiol 76:349 –392. Tazerart S, Vinay L, Brocard F (2008) The persistent sodium current generates pacemaker activities in the central pattern generator for locomotion and regulates the locomotor rhythm. J Neurosci 28: 8577– 8589. Vassias I, Patko T, Vidal PP, de Waele C (2003) Modulation of the beta1-3 voltage-gated sodium channels in rat vestibular and facial nuclei after unilateral labyrinthectomy and facial nerve section: an in situ hybridization study. Brain Res Mol Brain Res 120:73–78. Vibert N, Bantikyan A, Babalian A, Serafin M, Mühlethaler M, Vidal PP (1999) Post-lesional plasticity in the central nervous system of the guinea-pig: a “top-down” adaptation process? Neuroscience 94:1–5. Wada A (2006) Roles of voltage-dependent sodium channels in neuronal development, pain, and neurodegeneration. J Pharmacol Sci 102:253–268. Wold JE (1978) The vestibular nuclei in the domestic hen (Gallus domesticus) III. Ascending projections to the mesencephalic eye motor nuclei. J Comp Neurol 179:393– 405.

(Accepted 10 April 2009) (Available online 15 April 2009)