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

This Article Abstract Full Text (PDF) Supplemental Data All Versions of this Article: 572/2/407    most recent jphysiol.2005.100412v1 Services Email this article to a friend Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Citing Articles Citing Articles via HighWire Citing Articles via Google Scholar Google Scholar Articles by Sharifullina, E. Articles by Nistri, A. Search for Related Content PubMed PubMed Citation Articles by Sharifullina, E. Articles by Nistri, A. Related Collections Neuroscience

¹ Neurobiology Sector and CNR-INFM Democritos National Simulation Center, International School for Advanced Studies (SISSA), 34014 Trieste, Italy

	Abstract

Top Abstract Introduction Methods Results Discussion Supplemental material References

 
In the brain the extracellular concentration of glutamate is controlled by glial transporters that restrict the neurotransmitter action to synaptic sites and avoid excitotoxicity. Impaired transport of glutamate occurs in many cases of amyotrophic lateral sclerosis, a devastating motoneuron disease. Motoneurons of the brainstem nucleus hypoglossus are among the most vulnerable, giving early symptoms like slurred speech and dysphagia. However, the direct consequences of extracellular glutamate build-up, due to uptake block, on synaptic transmission and survival of hypoglossal motoneurons remain unclear and have been studied using the neonatal rat brainstem slice preparation as a model. Patch clamp recording from hypoglossal motoneurons showed that, in about one-third of these cells, inhibition of glutamate transport with the selective blocker DL-threo-ß-benzyloxyaspartate (TBOA; 50 µM) unexpectedly led to the emergence of rhythmic bursting consisting of inward currents of long duration with superimposed fast oscillations and synaptic events. Synaptic inhibition block facilitated bursting. Bursts had a reversal potential near 0 mV, and were blocked by tetrodotoxin, the gap junction blocker carbenoxolone, or antagonists of AMPA, NMDA or mGluR1 glutamate receptors. Intracellular Ca²⁺ imaging showed bursts as synchronous discharges among motoneurons. Synergy of activation of distinct classes of glutamate receptor plus gap junctions were therefore essential for bursting. Ablating the lateral reticular formation preserved bursting, suggesting independence from propagated network activity within the brainstem. TBOA significantly increased the number of dead motoneurons, an effect prevented by the same agents that suppressed bursting. Bursting thus represents a novel hallmark of motoneuron dysfunction triggered by glutamate uptake block.

(Received 19 October 2005; accepted after revision 1 February 2006; first published online 2 February 2006)
Corresponding author A. Nistri: Neurobiology Sector and CNR-INFM Center, International School for Advanced Studies (SISSA), 34014 Trieste, Italy. Email: nistri{at}sissa.it

	Introduction

Top Abstract Introduction Methods Results Discussion Supplemental material References

 
Amyotrophic lateral sclerosis (ALS) is a devastating neurodegenerative disease primarily affecting motoneurons (Bruijn et al. 2004). A large number of ALS patients show a deficit in the transport process of the excitatory transmitter glutamate that builds up extracellularly to produce excitotoxic damage to motor cells (Rothstein et al. 1992; Cleveland & Rothstein, 2001; Rao & Weiss, 2004). Because in the vast majority of cases the disease is sporadic and associated with normal synthesis of the transporter proteins (Rao & Weiss, 2004), various environmental factors are suspected to generate this condition by inhibiting glutamate transport in vulnerable brain regions (Bruijn et al. 2004).

One important form of ALS (termed bulbar) is clinically manifested as severe degeneration of brainstem motoneurons, although some motor nuclei are more vulnerable than others (Rowland & Shneider, 2001). In particular, the nucleus hypoglossus, that exclusively innervates tongue muscles, is among the most strongly involved in ALS (Krieger et al. 1994; Lips & Keller, 1999; Laslo et al. 2001), producing slurred speech, difficulty in mastication, swallowing and breathing. While the early damage of hypoglossal motoneurons (HMs) may be related to their characteristic intracellular Ca²⁺ homeostasis (Ladewig et al. 2003) and expression of Ca²⁺-permeable AMPA receptors (Del Cano et al. 1999; Laslo et al. 2001; Essin et al. 2002), it is also suggested that vulnerable motor nuclei normally possess distinctive properties of glutamate uptake to protect them against the risk factor of excitoxocity (Medina et al. 1996).

Previous studies have indicated that excitatory transmission on HMs is mediated by glutamate via AMPA-sensitive receptors (Rekling et al. 2000a; Essin et al. 2002), while glycine and GABA act as inhibitory transmitters (Donato & Nistri, 2000; Marchetti et al. 2002). Pharmacological block of glutamate uptake potentiates glutamatergic transmission and has been used as a model to investigate excitotoxicity and to devise prevention treatments (Danbolt, 2001; Huang & Bergles, 2004; Shigeri et al. 2004).

However, little is known about the functional consequences of glutamate uptake block and build-up of extracellular glutamate on HMs and their surrounding network. For instance, AMPA receptors may become inactive in the continuous presence of their agonist (Mayer & Armstrong, 2004). It might then be predicted that lingering glutamate (Cavelier et al. 2005) should rather activate NMDA receptors (Campbell & Hablitz, 2004; Huang & Bordey, 2004) and glutamate metabotropic receptors (mGluRs, Brasnjo & Otis, 2001; Huang et al. 2004) to stimulate firing of HMs.

In the present study, based on electrophysiological recording, and intracellular Ca²⁺ imaging from HMs of the rat brainstem slice preparation, we used the very selective glutamate transport inhibitor DL-threo-ß-benzyloxyaspartate (TBOA; Shigeri et al. 2004) to explore how it may change synaptic transmission, and its consequences on HM survival estimated with histochemical methods. Even after a short period of uptake block, we discovered the emergence of a novel type of bursting with significant neurotoxic damage to HMs.

	Methods

Top Abstract Introduction Methods Results Discussion Supplemental material References

 
Ethical standards

In accordance with the regulations of the Italian Animal Welfare Act following the European Community directives and approved by the local authority veterinary service, neonatal Wistar rats (1–5 days old; P1–5) used for this study were anaesthetized with I.P. urethane (2 g (kg body weight)^–1) and quickly decapitated.

Slice preparation

The brainstem was removed and transverse slices were cut, as recently described (Sharifullina et al. 2004, 2005; Pagnotta et al. 2005). For this purpose, the brainstem was horizontally fixed to an agar block and sectioned with a vibroslicer (starting from the caudal end) while submerged in Krebs solution (see below) at 4°C. Slices (usually five) containing the nucleus hypoglossus were cut at 200 µm intervals. The presence of this nucleus in each slice was immediately confirmed by viewing it under light microscopy. Thereafter, slices were continuously superfused (2–3 ml min^–1) at room temperature (24–25°C) with Krebs solution (gassed with 95% O₂–5% CO₂) containing (mM): NaCl 130, KCl 3, NaH₂PO₄ 1.5, CaCl₂ 1.5, MgCl₂ 1, NaHCO₃ 25, glucose 15 (pH 7.4; 300–320 mosmol). For further details see Sharifullina et al. (2005).

Ablation of the reticular formation

To remove the reticular formation adjacent to each nucleus hypoglossus, slightly thicker (300 µm) slices (n= 14) were cut due to the frailty of the tissue. Before patching, the lateral areas of slices were sectioned off under microscopic control as shown in the scheme of Fig. 2E. The ablation boundary was the lateral margin of the hypoglossus nucleus to remove a major excitatory input to this nucleus (Sharifullina et al. 2004, 2005). After this procedure, slices were used as for standard electrophysiological experiments. As indicated in the results, their bursting characteristics were identical to those of 200 µm slices.

View larger version (34K): [in this window] [in a new window]   Figure 2.  Characteristics of TBOA-evoked bursts in strychnine and bicuculline solution A, in the continuous presence of strychnine (0.4 µM) and bicuculline (10 µM), TBOA produces a slow inward current and repeated bursts (–60 mV holding potential). In the interburst intervals composite glutamatergic sPSCs (see inset for averaged responses before (n= 214 events) and after adding TBOA (n= 600 events)) are largely increased and prolonged. B, example of a non-bursting HM after TBOA application (filled arrow) that generates bursts after adding strychnine and bicuculline (open arrow). One burst is shown at higher gain. C, scatter plots of burst period or durations versus burst amplitude. Note lack of correlation. D, in the presence of strychnine and bicuculline (–60 mV holding potential), TBOA induced bursting (top trace) with repeated oscillations (middle record shows, at faster speed, one burst indicated by horizontal bar) that are more evident after application of D-APV (50 µM; bottom trace). E, the example shows that removal of the lateral regions of the brainstem slice (see scheme) does not prevent TBOA-induced bursting in strychnine and bicuculline solution (holding potential =–60 mV). All records were obtained with CsCl- and QX-314-filled electrodes.

Whole-cell patch clamping was used to record HM responses under voltage- or current-clamp conditions. HMs were visually identified with infrared microscopy (Ladewig et al. 2003). Patch pipettes were filled with intracellular solution containing (mM): CsCl 130, NaCl 5, MgCl₂ 2, CaCl₂ 1, Hepes 10, BAPTA 10, ATP–Mg 2, sucrose 2 (pH 7.2 with CsOH; 280–300 mosmol). This pipette solution was used for 67 HMs, 50 of which were recorded after adding QX-314 (300 µM) to the patch solution to block voltage-activated Na⁺ currents and the hyperpolarization-activated current I_h (Marchetti et al. 2002). This drug preserved recording stability that allowed TBOA-induced bursting to be observed for at least 35 min. Ninety-four HMs were recorded with a patch solution in which 130 mM KCl had replaced CsCl; within this group, 33 HMs were recorded with QX-314 added to the solution. Since there was no difference in TBOA-evoked bursting characteristics between KCl- and CsCl-recorded HMs, these data were pooled together for statistical analysis. For 16 HMs recorded with CsCl and QX-314 we added 20 mM BAPTA (tetrapotassium salt) to the patch pipette, to enhance intracellular Ca²⁺ buffering. Cells were clamped at –60 or –65 mV holding potential, while series resistance (5–25 M) was routinely monitored. All recorded currents were filtered at 3 kHz and sampled at 5–10 kHz. Postsynaptic currents, electrical oscillations and bursts were quantified as previously reported (Sharifullina et al. 2005) using a template search protocol (pClamp 9.0; Axon Instruments, Molecular Devices, Union City, CA, USA) applied to at least 5 min-long consecutive records.

Intracellular Ca²⁺ imaging

Ca²⁺ imaging was carried out according to the method recently described (Fabbro et al. 2004). In brief, slices were loaded with the fluorescent Ca²⁺ dye Fluo-3 AM (20 µM; Molecular Probes, Eugene, OR, USA) for 40 min in oxygenated standard saline solution. After a 20 min wash, slices were transferred to the recording chamber and Ca²⁺ transients were visualized with a fast CCD camera (Coolsnap HQ; Roper Scientific, USA). Because of the need for continuous, long-lasting (about 1 h) recording, Ca²⁺ transients (usually 30 s long) from single cells within the nucleus hypoglossus were acquired at 1 Hz to minimize photobleaching as indicated by a stable baseline. In each slice, 10 randomly distributed motoneurons were identified as such because their somatic diameter was > 20 µm, and were analysed by placing a small region of interest over the cell body using the Metafluor software (Metafluor Imaging Series 6.0, Universal Imaging Corporation, USA). Ca²⁺ transients were expressed as amplitude fractional increase (F/F₀, where F₀ is the baseline fluorescence level and F is the rise over baseline). Cells with very bright baseline Ca²⁺ fluorescence were not analysed on the assumption they were already damaged. To maximize the detection of TBOA-induced rhythmic Ca²⁺ transients, 0.4 µM strychnine and 10 µM bicuculline were pre-applied to slices for 10 min prior to the start of 50 µM TBOA application, and maintained thereafter. Data were obtained from 16 slices from P4–6 rats (n= 6). In each slice 10 motoneurons were analysed; synchronicity of Ca²⁺ signals (within the temporal resolution of 1 s) was determined by cross-correlation analysis (Sharifullina et al. 2005) using the cross-correlation function (CCF) of Clampfit software (9.2 version; Axon Instruments, Molecular Devices, Union City, CA, USA). The same software was also used for analysing responses obtained with double patch recordings.

Drugs

Drugs were applied via the bathing solution with the exception of AMPA delivered by focal pressure pulses (10 ms; 6 p.s.i.; Pagnotta et al. 2005; Sharifullina et al. 2005) via a pipette. In this case the puffer pipette was filled with AMPA (0.1 mM) dissolved in standard Krebs solution and positioned approximately 20–50 µm away from the soma of the recorded cell. AMPA was applied once every 45 s to minimize desensitization. A number of experiments (n= 112 HMs) were performed in the continuous presence of bicuculline (10 µM) and strychnine (0.4 µM) in the bathing solution to block GABA- and glycine-mediated transmission (Donato & Nistri, 2000; Marchetti et al. 2002; Sharifullina et al. 2004) so that glutamatergic transmission could be studied in isolation. The following drugs were used: DL-threo-ß-benzyloxyaspartate (TBOA), L-trans-pyrrolidine-2,4-dicarboxylic acid (PDC), 6-cyano-7-nitroquinoxaline-2,3-dione (CNQX), D-amino-phosphonovalerate (APV), 7-(hydroxyimino) cyclopropan[b]chromen-1a-carboxylate ethyl ester (CPCCOEt; selective antagonist for mGlu1 receptors), (RS)--amino-3-hydroxy-5-methyl-4-isoxazolepropionic acid (AMPA), N-2,6-dimethylphenylcarbamoyl-methyl triethylammonium bromide (QX-314), (±)-4-(4-aminophenyl)-1,2-dihydro-1-methyl-2-propylcarbamoyl-6,7-methylenedioxyphthalazine (SYM 2206; selective antagonist for AMPA receptors), and the high-threshold Ca²⁺ channel blocker nifedipine were purchased from Tocris (Bristol, UK); bicuculline methiodide (bicuculline), strychnine hydrochloride (strychnine), 1,2-bis(2-aminophenoxy)ethane-N,N,N',N'-tetraacetic acid tetrakis(acetoxymethyl ester) (BAPTA AM), and carbenoxolone (disodium salt) were from Sigma (Milan, Italy). Tetrodotoxin (TTX) was obtained from Latoxan (Valence, France).

HM identification and estimate of their excitotoxic damage

In the neonatal as well as in the adult rat (Nunez-Abades & Cameron, 1995), HMs comprise the largest population (about 90%; Viana et al. 1990) of neurons within the XII nucleus, are large cells (25–50 µm), multipolar in shape and distributed throughout this nucleus (Kitamura et al. 1983; Boone & Aldes, 1984). Conversely, interneurons (10–18 µm somatic diamater) are round- to oval-shaped neurons and are much less numerous (Boone & Aldes, 1984). Thus, within the hypoglossus nucleus, in control as well as in drug-treated slices, we counted only cells with somatic diameter > 20 µm, to make sure that we analysed HMs. Control experiments were done to obtain further confirmation of the identification of large cells as motoneurons via immunocytochemical staining of brainstem slices with an antibody against the acetylcholine synthetic enzyme (ChAT; see also Donato & Nistri, 2000), and by the use of the monoclonal antibody against the non-phosphorylated form of neurofilament H (SMI32; a motoneuron marker; see Jacob et al. 2005; Raoul et al. 2005). For these experiments the brainstem was quickly removed from P3–5 rats, fixed in 4% paraformaldehyde containing 30% sucrose in phosphate buffer pH 7.4 (PBS) for 24 h at 4°C. Microtome sections (40 µm thick) were used for free-floating immunostaining as previously described (Pagnotta et al. 2005). Briefly, slices were treated with a blocking solution (5% bovine serum albumin, 4% fetal calf serum, 0.1% Triton X100 in phosphate-buffered saline, pH 7.4) for 60 min at room temperature. Slices were incubated overnight at 4°C with mouse monoclonal antibodies against ChAT (1 : 100; kindly provided by Dr L. Domenici, SISSA, Trieste; Umbriaco et al. 1994), or against SMI32 (1 : 1000; Sternberger monoclonals, Covance Research Products Inc., Berkeley, CA, USA) in the same blocking buffer at 4°C. The secondary antibodies used were AlexaFluor 488 (1 : 500 dilution; Molecular Probes, Invitrogen, San Giuliano Milanese, Italy) for 2 h at room temperature. Slices stained with secondary antibody only showed no immunostaining. Measurements were obtained with ImagePro software (Hamamatsu srl, Arese, Italy).

Intracellular markers such as the cytosolic enzyme ChAT may be lost when cell membranes are damaged as a result of toxicity and death. Moreover, since ChAT is also present in a number of interneurons and afferent terminals to HMs (Ichikawa & Hirata, 1990; Ichikawa & Shimizu, 1998) and many HMs only show moderate ChAT staining (Ichikawa & Hirata, 1990), counting motoneurons killed by a certain drug treatment on the basis of ChAT staining may produce confusing results. Hence, to evaluate the number of HMs surviving after pharmacological treatment, we counted only cells (within the nucleus hypoglossus) with somatic diameter > 20 µm after applying the cell-permeable dye Hoechst 33342 (10 mg ml^–1 stock from Molecular Probes, Invitrogen; final dilution was 1 : 500 in standard Krebs solution) that, once bound to DNA, emits blue fluorescence. Only cell profiles with a clearly outlined nucleus at the same focal plane were analysed. This method provided the global number of motoneurons (surviving plus damaged) in each tissue section. To investigate the number of HMs killed by glutamate excitotoxicity, in analogy with previous studies (Kristensen et al. 2001; Babot et al. 2005; Bosel et al. 2005) including those with TBOA (Bonde et al. 2003), we then stained dead cells with propidium iodide (PI) solution (1.0 mg ml^–1 stock from Sigma, Milan, Italy; 1 : 3000 in Krebs solution), a cell-impermeable DNA dye which can bind DNA and emits red fluorescence exclusively when cell and nuclear membranes have been severely damaged. For each experiment on excitotoxicity, brainstem slices were separated into various groups: control, TBOA (50 µM), and TBOA plus an antagonist. Slices were incubated at room temperature in the corresponding solution, continuously oxygenated for 1 h, then rinsed with Krebs solution, and placed in Krebs containing Hoechst (1 : 500) and PI (1 : 3000) for at least 45 min. Thereafter, slices were transferred to glass coverslips (without any fixation) and examined under a fluorescence microscope (x5). An 18-square grid was applied over each hypoglossal nucleus (see scheme in Fig. 7D). For each protocol at least six brainstem slices (from 24 rats) were used. Cells were counted at the same level in each slice.

View larger version (76K): [in this window] [in a new window]   Figure 7.  Excitotoxic damage by TBOA application A, example of immunocytochemical identification of HMs with the SMI32 motoneuron marker (left) and the ChAT immunopositivity (middle). PI(+) HMs are shown on the right after 1 h TBOA application. In each panel the calibration bar is 100 µm. B, examples of fluorescent cell counts after staining with PI (top row) or Hoechst 33342 (bottom row) in various experimental protocols ranging from control Krebs (left) to TBOA (middle), and TBOA + SYM 2206 (right). In each panel PI(+) cells indicate dead motoneurons. Calibration bars are 100 µm (bottom rows) and apply to top panels as well. C, histograms quantifying the number of dead neurons (PI(+)) in various experimental protocols. TBOA significantly increases the number of PI(+) HMs, an effect absent when it is incubated together with D-APV, SYM 2206, carbenoxolone, or CPCCOEt, as the values are not different from control values that reflect cells damaged by the brain-slicing preparation. For each histogram the range of counted neurons was 239–388. Data are expressed as percentage of PI(+) cells with respect to those stained with Hoechst 33342. ***P < 0.0005 D, schematic diagram of the grid arrangement positioned over the nucleus hypoglossus to count HMs for histological purposes. Error bars are the standard error of the means.

The electrophysiological database of the present study comprises 193 HMs. Results were expressed as means ±S.E.M. where n refers to the number of cells. For immunohistochemical analysis, data with PI staining were expressed as a percentage of those labelled with Hoechst 33342 (taken as 100%). Statistical significance was assessed with Student's paired t test applied to parametric raw data only, or for non-parametric values with ANOVA followed by the Tukey test. Two groups of data were considered statistically different if P < 0.05.

	Results

Top Abstract Introduction Methods Results Discussion Supplemental material References

 
Bursting induced by glutamate uptake blocker

As shown in Fig. 1A, bath-application of TBOA (50 µM) induced a slowly developing inward current (–44 pA) which stabilized at a plateau after approximately 4 min. In 33/103 HMs (32%) bathed in standard Krebs solution, the TBOA-induced inward current was accompanied by the emergence of bursting activity (Fig. 1A) characterized by large inward episodes (open arrow points to expanded trace of a single burst) with superimposed fast (9.5 Hz) stereotypic discharges intermingled with much slower, composite spontaneous postsynaptic currents (sPSCs; see expanded timebase averages during one burst shown in Fig. 1A). Since the fast discharges possessed all the characteristics of action potentials generated in nearby HMs and transmitted to the voltage-clamped HM via gap junctions (Sharifullina et al. 2005), they were referred to as spikelets (Long et al. 2004, 2005). During the interburst interval, sPSCs had higher frequency (2.3 ± 0.3 Hz versus 1.2 ± 0.2 Hz; n= 33; P < 0.005) and larger amplitude (–69 ± 11 pA versus–40 ± 4 pA; n= 33; P < 0.05) than in control, although the cell input resistance did not change significantly (150 ± 14 Mversus 177 ± 21 M in control; n= 33; P > 0.05).

View larger version (42K): [in this window] [in a new window]   Figure 1.  Bursting induced by TBOA (50 µM) application A, sample record of HM voltage clamped at –60 mV (patch pipette containing Cs and QX-314), demonstrating emergence of bursts over a slow inward current. One burst (open arrow) is shown at higher gain and faster speed to depict its components that include averaged spikelets and composite sPSCs. Vertical calibration for sPSCs applies also to spikelets. B, scatter plots of burst period (top) and duration (bottom) for varying burst amplitudes. Note lack of correlation between these parameters. C, TBOA-induced bursting has a similar time course under current- (top) or voltage- (bottom) clamp conditions at the same membrane potential (two different cells both recorded with KCl-filled electrodes). D, voltage dependence of TBOA-evoked bursts (same cell as in A), indicating that burst currents became outwards at positive potentials. The inset shows average current amplitude–voltage plots for bursts recorded from 11 HMs. E, example of five averaged inward currents induced by puffer application (10 ms) of AMPA to HMs in control solution or in the presence of TBOA in the interval between bursts. The histograms summarize the AMPA response amplitude in TBOA solution as percentage of control (in Krebs solution containing strychnine and bicuculline to avoid AMPA-mediated release of glycine and GABA) and after washout of TBOA. No significant difference was found (n= 6 HMs). Error bars are the standard error of the means.

Bursts induced by TBOA were sensitive to changes in cell membrane potential between –100 and +40 mV (Fig. 1D). They reversed near 0 mV and became outward currents at positive values (Fig. 1D), suggesting that burst suppression near 0 mV was not due to deactivation of HM voltage-dependent conductances. Figure 1D (bottom) presents the average current–voltage relation for bursts which had a null potential at +10 mV.

Since certain glutamate uptake blockers can have agonist action on glutamate receptors (Danbolt, 2001), we explored whether TBOA could alter currents elicited by brief puffer applications of the non-transportable glutamate agonist AMPA. As shown in Fig. 1E, the amplitude of the AMPA-induced currents was not changed in the presence of TBOA, thus indicating that the uptake blocker did not bind to AMPA-sensitive receptors.

In 70 HMs, despite the fact that TBOA elicited an inward current (–47 ± 6 pA) and increased the frequency (2.9 ± 0.2 versus 1.7 ± 0.2 Hz; P < 0.005) and amplitude (–98 ± 9 versus–47 ± 3 pA; P < 0.005) of sPSCs, no bursting was apparent for at least 20 min continuous application of this agent. These cells were therefore regarded as non-bursters. Unlike bursters, non-bursters showed a significant fall in input resistance in the presence of TBOA (132 ± 9 versus 169 ± 10 M in control, n= 34; P < 0.002). Seven of these cells did, however, generate burstlets similar to those evoked by application of an mGluR agonist (Sharifullina et al. 2005).

Bursting was facilitated by synaptic inhibition block

Because bursting due to glutamate uptake block was present in a minority of HMs only, it seemed possible that synaptic inhibition mediated by glycine and GABA might have curtailed the onset of this activity. When synaptic inhibition was suppressed by strychnine and bicuculline (Donato & Nistri, 2000; Marchetti et al. 2002), subsequent application of TBOA produced bursts in 47% of tested cells (23/49), as exemplified in Fig. 2A. In this condition, it was possible to observe pharmacologically isolated glutamatergic currents (see example of averaged events in the inset to Fig. 2A) which exhibited larger amplitude and slower decay in the presence of TBOA (full details about the properties of these events are in Table 1). Note that pharmacologically isolated synaptic events had much slower kinetics than spikelets and displayed variable amplitude in contrast with the stereotypic size of spikelets (Sharifullina et al. 2005). In the presence of strychnine and bicuculline, the baseline inward current evoked by TBOA on bursting HMs was –57 ± 9 pA (n= 23).

On seven HMs we also tested the glutamate uptake inhibitor PDC (100 µM; a concentration fully inhibiting glutamate transporters; Shigeri et al. 2004) that evoked a slow inward current (–49 ± 8 pA) without triggering burst activity.

Ablation of the reticular formation

To establish if bursting required an extensive circuitry comprising the reticular formation, we resected regions of the slice immediately lateral to the nucleus hypoglossus (scheme in Fig. 2E). On 14 HMs (one in each reduced slice), TBOA did not evoke bursting activity. Subsequent application of bicuculline and strychnine turned six cells into bursters (43% of all tested cells; see Fig. 2E, lower panel). Bursts had –160 ± 22 pA amplitude, 120 ± 17 s period (22 ± 8% CV) and 23 ± 1 s duration that was significantly shorter than in intact slices (P= 0.017).

Role of Ca²⁺ in TBOA-evoked bursting

It seemed useful to explore the spatiotemporal distribution of bursting HMs within the TBOA-treated slices. For this purpose we studied changes in intracellular Ca²⁺ ([Ca²⁺]_i) after loading slices with the membrane-permeable fluorescent Ca²⁺ indicator Fluo-3 AM and maximizing bursting occurrence with application of strychnine and bicuculline. Figure 3A shows an example of these experiments in which [Ca²⁺]_i changes were recorded from 10 bursting HMs, the distribution of which in the slice is shown in Fig. 3B. TBOA induced a gradual rise in HM baseline [Ca²⁺]_i, and triggered the onset of repetitive Ca²⁺ signals. Despite their widespread topography within the slice (Fig. 3B), bursting HMs generated transients strongly correlated as indicated by the cross-correlation plot (Fig. 3C) for the 10 HMs of Fig. 3A. For example, the cell labelled as ‘1’ (Fig. 3A) was 287 µm apart from cell 2, yet they both generated synchronous [Ca²⁺]_i transients. In each cell, bursting developed after the baseline rise had reached an apparent plateau. While the amplitude of [Ca²⁺]_i transients was variable because of dissimilar dye loading among HMs, their period was 52 ± 3 s (CV = 13 ± 2%) with average duration of 29 ± 1 s (CV = 18 ± 1%; n= 6 slices; 60 HMs in total). Hence, TBOA application induced rhythmic changes in [Ca²⁺]_i, which occurred synchronously.

View larger version (43K): [in this window] [in a new window]   Figure 3.  Rhythmic changes in [Ca2+]i evoked by application of TBOA A, example of records of [Ca2+]i changes taken from 10 HMs distributed within the same slice preparation (B, where each circle indicates a different cell). Note that, after a slow rise in baseline, all 10 motoneurons developed rhythmic changes in [Ca2+]i Bar = 100 µm. C, plot of superimposed cross-correlation functions for the 10 HMs shown in A and B. Note strong synchrony of [Ca2+]i signals. Recording made in the continuous presence of bicuculline and strychnine.

Pharmacology of bursting

Once bursting was established, it was always fully suppressed by TTX (1 µM; n= 17; Fig. 4A) while the steady inward current persisted (–31 ± 7 pA). Table 2 shows that miniature glutamatergic currents (mPSCs) in TBOA solution had slower decay with otherwise unchanged characteristics. Figure 4B indicates that the gap junction blocker carbenoxolone (100 µM) completely abolished bursting activity and associated spikelets (compare responses in the inset to Fig. 4B). However, on average the baseline inward current evoked by TBOA was not significantly changed (–5 ± 11 pA; n= 9) by carbenoxolone, the action of which was present even when outward burst currents were observed at +40 mV membrane potential, or in the presence of strychnine and bicuculline (n= 4). Further support for the role of electrical coupling in bursting was obtained by simultaneous double patch recording from two HMs as exemplified in Fig. 5A, in which, in the presence of TBOA (in strychnine plus bicuculline solution), one cell was recorded under voltage clamp at –61 mV (top) and the neighbouring one under current clamp at –48 mV (bottom). The inward currents (–40 pA) were associated with depolarizing bursts in the nearby motoneuron which generated repeated firing. Taking the records during bursts and analysing them for their cross-correlation (Fig. 5B) demonstrated event synchronicity.

View larger version (37K): [in this window] [in a new window]   Figure 4.  Pharmacology of TBOA-induced bursting A, in the presence of strychnine and bicuculline, TBOA induces bursts that are suppressed by TTX (1 µM). KCl-filled patch pipette. B, in Krebs solution TBOA-evoked bursts are blocked by carbenoxolone (100 µM). The two horizontal bars show one single burst before carbenoxolone with fast discharges (spikelets) and the last burst during carbenoxolone when fast discharges are absent. CsCl-filled patch pipette. C, in Krebs solution (KCl-filled electrodes), TBOA-dependent bursts are blocked by the AMPA antagonist SYM 2206 (20 µM) leaving a background of high-frequency activity either in current (top trace) or in voltage (bottom trace) clamp (same membrane potential; –67 mV). D, in the presence of strychnine and bicucuculline, TBOA-elicited bursts are reversibly suppressed by applying the NMDA receptor antagonist D-APV (50 µM). KCl-filled patch pipette. E, top, in the presence of strychnine and bicuculline TBOA-induced bursts are blocked by the selective mGluR1 antagonist CPCCOEt (100 µM). Bottom, in a different cell bathed in standard solution, CPCCOEt (100 µM) blocks TBOA bursting even when the cell is held at +40 mV, and displays outward burst currents. CsCl and QX-314-filled patch pipette.

View larger version (24K): [in this window] [in a new window]   Figure 5.  Double recording from HMs of TBOA-evoked bursting A, sample traces in the presence of TBOA (in strychnine and bicuculline solution) showing recording from two adjacent HMs in the same slice. One cell (top) is under voltage clamp (–61 mV), while the other one is under current clamp (–48 mV). Bursting is observed to occur concomitantly as large burst currents (top) in association with depolarizing bursts with superimposed repeated firing of action potentials (bottom). B, cross-correlogram analysis of records taken during each burst event to demonstrate synchronicity between these two cells.

Time course of bursting

Although the effect of TBOA was reversible if applied for less than 20 min, longer exposure to this agent was deleterious for HMs as shown in the example of Fig. 6A depicting the temporal evolution of the action of TBOA in strychnine and bicuculline solution. After 35 min application, the HM expressed a large inward current and became leaky, indicating cell deterioration without recovery. Monitoring the time profile of [Ca²⁺]_i changes simultaneously in several HMs allowed us to discover whether any sign of cellular damage, like an irreversible large rise in [Ca²⁺]_i, was a common phenomenon. To this end, we performed continuous, long-lasting imaging (60 min) as exemplified in Fig. 6B, in which 10 HMs were recorded during sustained application of TBOA plus strychnine and bicuculline (their location in the slice is shown in Fig. 6C). HMs labelled 1–3 had minimal [Ca²⁺]_i rise and minimal bursting after applying TBOA, while HMs 4–6 displayed a short cycle of low-amplitude bursts. HMs 7–10 showed a large, irreversible increment of [Ca²⁺]_i baseline together with bursting activity. We analysed the cross-correlation of [Ca²⁺]_i transients at the early (8–14 min) and late (26–32 min) stages of TBOA application. The early transients had good cross-correlation for HMs 4–10 (CCF = 0.59 ± 0.08), although the CCF later fell to 0.25 ± 0.03, indicating disorganized discharges within the slice. The implication that sustained application of TBOA caused excitotoxic damage to HMs was then quantitatively evaluated with histochemical techniques.

View larger version (40K): [in this window] [in a new window]   Figure 6.  Time course of TBOA-induced bursting A, example of HM (recorded with CsCl and QX-314-filled pipette; bathing solution contains strychnine and bicuculline) under voltage clamp showing the action of TBOA with early inward current, bursting and final deterioration of the cell (35 min later) with onset of very large inward current and loss of membrane resistance. B, simultaneous Ca2+ imaging from 10 HMs (indicated in C with corresponding numbers) following TBOA application in strychnine and bicuculline solution to show evolution and topography of bursting over 68 min of continuous recording. Bar = 100 µm.

For quantitative analysis of motoneuron survival after TBOA application, we first wished to validate that, within the nucleus hypoglossus, large (> 20 µm cell body diameter) cells were indeed HMs. Hence, we used an antibody against ChAT to map HMs, as these are prototypical cholinergic neurons. In accordance with previous studies, as shown in Fig. 7A (middle), ChAT-positive neurons were large cells (with a large nucleus of 108 ± 20 µm² area; n= 50) densely packed within the nucleus hypoglossus. The identification of large cells as motoneurons was confirmed by staining them with the motoneuron marker SMI32 (Fig. 7A, left). However, both histochemical methods relied on fixed tissue and permeabilized cells to enable intracellular antibody-based staining. We wished to quantify motoneuron excitotoxicity in unfixed specimens using a standard method based on staining of dead cells with the nuclear dye PI, which penetrates inside cells only after their membrane disruption (Kristensen et al. 2001; Bonde et al. 2003; Babot et al. 2005; Bosel et al. 2005). Figure 7A (right) shows cells stained with PI within the nucleus hypoglossus. By counting just cells > 20 µm diameter within the area indicated in Fig. 7D, we therefore restricted our measurements to damaged HMs. The global number of cells (intact as well as damaged) within each section was estimated with the cell-permeable dye Hoechst 33342. As depicted in Fig. 7B (top left), after incubation in Krebs solution a few HMs were PI(+), in keeping with the inevitable consequence of tissue slicing. After 1 h exposure to TBOA the number of PI(+) cells was larger (top middle), an effect minimized when TBOA was applied in the presence of SYM 2206 (top right). Figure 7C quantifies these observations: whereas in Krebs solution about one-third of HMs were PI(+) (34 ± 1% of the Hoechst 33342-stained neurons within selected area, n= 6 slices), there was a significant increase in the number (46 ± 2%; n= 6) of PI(+) HMs with TBOA application, an effect absent when D-APV (50 µM; 33 ± 1%, n= 7), SYM 2206 (20 µM; 37 ± 1%, n= 5), carbenoxolone (100 µM; 36 ± 2%, n= 6) or CPCCOEt (100 µM; 30 ± 1%, n= 6) were incubated in the presence of TBOA.

	Discussion

Top Abstract Introduction Methods Results Discussion Supplemental material References

 
One unexpected finding of the present study was that the glutamate uptake blocker TBOA triggered a novel type of electrical bursting of HMs due to the complex interplay of various classes of glutamate receptor, and apparently supported by gap junctions. TBOA also produced excitotoxicity in HMs which was prevented by the same drugs that blocked bursting. Our data thus suggest that this form of bursting was the hallmark of runaway excitation leading to motoneuron death.

Glutamate transporters: excitation and excitotoxicity

In the brain the extracellular concentration of the excitatory neurotransmitter glutamate is tightly controlled by a family of membrane transporters predominantly expressed by glia and generating electrogenic signals due to cotransport of ions (Danbolt, 2001; Huang & Bergles, 2004; Shigeri et al. 2004). Their role is to regulate the large amount of glutamate released at synapses and to prevent its spillover activating extrasynaptic receptors (Cavelier et al. 2005). Because excessive activation of glutamate receptors is excitotoxic to neurons, change in the function and expression of transporters can have important pathogenetic roles in neurodegenerative diseases like ALS (Cleveland & Rothstein, 2001; Rao & Weiss, 2004).

While the full cycle of the transporter operation is usually slower that the duration of glutamatergic synaptic events (Danbolt, 2001), it appears that rapid glutamate binding to the transporter sites is already sufficient to buffer the transmitter synaptic concentration (Bergles et al. 1999). To prevent glutamate binding to the transporter and to investigate the impact of glutamate uptake systems on neuronal network functions, it is thus useful to apply a non-transportable, broad-spectrum blocker like TBOA (Danbolt, 2001; Shigeri et al. 2004). Conversely, transportable inhibitors like PDC can release endogenous glutamate from neuronal and glial pools (Danbolt, 2001), thus generating complex effects that are quite different from those observed with selective uptake blockers (like TBOA) which are non-competitive as well as non-transportable inhibitors (Danbolt, 2001). It is therefore not unexpected that, unlike TBOA, application of PDC could not evoke rhythmic bursting even though it generated an inward current comparable to the one observed with TBOA administration. It seems likely that widespread release of glutamate at network level was incompatible with burst generation. In support of this notion, we found that bath-applied 100 µM glutamate did not evoke bursting (Marchetti et al. 2002; authors' unpublished observation).

The nucleus hypoglossus as an in vitro model for studying the role of glutamate uptake

The brainstem nucleus hypoglossus is a compact region in which HMs make up at least 90% of its cell population (Viana et al. 1990). HMs are at special risk of glutamate-evoked Ca²⁺-dependent excitotoxicity because of their expression of Ca²⁺-permeable AMPA receptors (Del Cano et al. 1999; Laslo et al. 2001; Essin et al. 2002) and because of the large amount of intracellular free Ca²⁺ (Ladewig et al. 2003). While the nucleus hypoglossus expresses a higher level of glutamate transporter than any other brainstem nucleus (Medina et al. 1996), presumably to reduce these risk factors, it is important to note that deficit in glutamate uptake has been demonstrated in many cases of ALS (Rothstein et al. 1992; Cleveland & Rothstein, 2001; Rao & Weiss, 2004), and that the bulbar form of this insidious disease usually starts with early symptoms of tongue muscle impairment (slurred speech, dysphagia, tongue biting, etc.). For these reasons it seemed helpful to study the impact of glutamate transporters on HMs using a brainstem slice preparation as a model, bearing in mind that experimental conditions (motoneuron viability and stability in vitro) required the use of a neonatal preparation with intrinsic properties due to the developmental characteristics. With the present model it was unnecessary to seek long-range interaction among various circuits to explain bursting activity, because this phenomenon could be observed even with a reduced slice preparation from which most reticular formation inputs had been removed. This realization prompted us to focus our attention on local mechanisms underlying bursting.

Network origin of TBOA-induced bursting

In our conditions TBOA elicited a novel type of bursting from a number of HMs even when, on all recorded cells, there was an early, steady inward current and changes in synaptic event kinetics. In general, neuronal bursting is by no means a necessary result of uptake block because in other brain areas either synaptic depression (Iserhot et al. 2004) or massive depolarization shift (Tsukada et al. 2005) develops.

Several features of the TBOA-evoked bursts were analogous to those of the mGluR-dependent bursting we have recently reported following activation of group I mGluRs (Sharifullina et al. 2005). Hence, under voltage clamp, each burst was the expression of the depolarization (mainly due to glutamate receptor activity) affecting multiple, interconnected HMs (via gap junctions), and was accompanied by superimposed spikelets and synaptic events. The burst reversal potential near 0 mV, the burst period insensitivity to membrane voltage, and the observation that, under voltage clamp, bursts became outward currents, all indicate that these responses were summated currents mediated by the activation of glutamate receptors on premotoneurons and motoneurons (cf. pharmacological antagonism) rather than of voltage-dependent conductances intrinsic to HMs. These characteristics closely resemble those of disinhibited bursting of rat spinal motoneurons which has a clear network origin (Bracci et al. 1996), and are in accordance with the properties of rhythmogenic motor networks (Marder & Calabrese, 1996), to which the nucleus hypoglossus belongs. Burst suppression by TTX also suggests dependence on network-propagated activity. Further support for the network origin of these bursts is supplied by the fact that these phenomena were present even when HMs were recorded with a patch solution containing the Na⁺ (and I_h) channel blocker QX-314. Since a large concentration of BAPTA in the recording pipette could not inhibit bursting, it appears that bursts indeed originated from extensive network excitation which this Ca²⁺ chelator (applied to a single HM) could not switch off. Validation of the network origin of the TBOA-evoked bursts came from imaging [Ca²⁺]_i. HMs generated synchronous, rhythmic changes in [Ca²⁺]_i riding on a slowly rising basal Ca²⁺. Nevertheless, because imaging was restricted to superficial HMs in each slice, we could not use this method to quantify the number of bursting HMs.

The origin of Ca²⁺ waves recorded from motoneurons presumably comprised rhythmic intracellular release from internal stores subsequent to persistent activation of mGluRs (Schoepp et al. 1999), plus depolarization-dependent Ca²⁺ influx of an oscillatory nature due to the particular properties of HM bursts mediated by gap junctions and recurrent activation of certain K⁺ conductances (Sharifullina et al. 2005). Of course, transmembrane influx of Ca²⁺ via activated AMPA and NMDA receptors is likely to have contributed to the intracellular Ca²⁺ rises. The heterogeneous origin of [Ca²⁺]_i signals might explain why rhythmic changes in [Ca²⁺]_i had a shorter period than electrical bursts, because the dynamics of [Ca²⁺]_i rise and buffering were possibly controlled by processes different from electrical oscillations, and actually contributed to the composite shape of electrical bursts. Confirmation of the network origin of bursting and its Ca²⁺ dependence was obtained with experiments based on the application of the cell-permeable BAPTA AM to the slice preparation. In this case, after long exposure to this agent to enable intracellular loading and Ca²⁺ buffering, bursting was, in fact, suppressed. Furthermore, simultaneous patch clamp recording from two adjacent HMs showed synchronicity of bursting in keeping with a network origin of this phenomenon.

Gap junctions and synaptic inhibition have opposite effects on TBOA-induced bursting

A process probably required for bursting (and indeed excitotoxicity) was the presence of gap junctions among HMs. These have been demonstrated in vivo (Mazza et al. 1992) and in vitro (Rekling et al. 2000b; Sharifullina et al. 2005) to involve about 40% HMs, approximately the same proportion of cells showing TBOA-dependent bursting. Although the incidence of gap junctions decreases with developmental maturation (Mazza et al. 1992), certain connexins responsible for electrical coupling among motoneurons are strongly expressed even in adult brainstem motoneurons (Rekling et al. 2000b; Honma et al. 2004), suggesting that motoneuron coupling can occur even after development is complete. Further studies will be necessary to understand if TBOA-induced bursting occurs in adult HMs with properties analogous to those found in neonatal motoneurons, and whether gap junctions contribute to it.

This type of bursting displayed certain features of gap junction-dependent activity such as appearance of spikelets, and block by carbenoxolone. In the neonatal rat spinal cord, gap junctions among motoneurons have been recently shown to mediate synchronicity of motor discharges (Tresch & Kiehn, 2000). It seems likely that gap junctions among HMs not only had a similar action, but were also responsible for spreading glutamate-dependent excitation once its uptake was inhibited. In fact, Ca²⁺ imaging showed that synchronous bursting could develop between HMs at distant locations within the same slice preparation. It is feasible that in such cases, synchronicity was due to coactivation of local circuits rather than direct electrical coupling between remote HMs. Nevertheless, gap junctions played a major role because their pharmacological block suppressed bursting. Bursting was partly suppressed also by GABA- and glycine-mediated synaptic inhibition because blocking glycinergic and GABAergic transmission increased the likelihood of observing bursting, though the majority of HMs remained non-bursters.

Mechanisms underlying TBOA-induced bursting

Gap junctions were not the only contributors to bursting. AMPA, NMDA and mGluR1 receptors also played an important role because blocking any one of these systems suppressed bursting and excitotoxicity. Previous studies have shown how glutamate uptake block enables activation, by ambient glutamate, of NMDA receptors (Campbell & Hablitz, 2004; Huang & Bordey, 2004) and mGluRs (Brasnjo & Otis, 2001; Huang et al. 2004) not normally accessible to synaptic glutamate. We suggest that, through HM gap junctions that allow spread and synchronization of excitation (Sharifullina et al. 2005), each glutamate receptor class brought its distinctive, yet complementary contribution. The coincidence of all of them (plus the presence of gap junctions) was the condition necessary for the onset of bursting. A mechanistic hypothesis to account for bursting is provided in Fig. 8 with a schematic diagram to summarize the role of various types of glutamate receptors. Electrical coupling among HMs is proposed to be important for recruiting these cells to bursting and to synchronize their discharges. AMPA receptors were probably responsible for supporting glutamatergic transmission to mediate depolarization of HMs and local network neurons. Although the actual concentration of free glutamate after application of TBOA was unknown, it is likely that it was insufficient to largely desensitize AMPA receptors that have lower affinity than NMDA receptors for glutamate (Danbolt, 2001). When AMPA-mediated depolarization was pharmacologically inhibited by SYM 2206 or CNQX, high-frequency synaptic events presumably caused by activation of glycine and GABA receptors (Donato & Nistri, 2000; Marchetti et al. 2002) persisted, although they were apparently inadequate to activate Mg²⁺-blocked NMDA receptors. Likewise, mGluRs could not relieve blocked NMDA receptors because they mediate a relatively small inward current even when saturating agonist concentrations are used (Sharifullina et al. 2004).

View larger version (26K): [in this window] [in a new window]   Figure 8.  Idealized diagram to account for the mechanism of HM bursting evoked by TBOA Network premotoneurons (expressing receptors for AMPA, NMDA and mGluR1s) release glutamate onto motoneurons to activate AMPA and NMDA receptors. Astrocytes in close proximity to synapses mop up released glutamate, minimizing its spillover to extrasynaptic mGluR1s. Some HMs are electrically coupled via gap junctions. When astrocyte uptake is blocked by TBOA, build-up of extracellular glutamate activates premotoneurons and motoneurons, an effect amplified by mGluR1 activation, which increases cell input resistance and thus excitability. Gap junctions enhance excitation among coupled HMs and facilitate synchronous bursting. Coincidence of strong glutamate receptor activity with gap junctions therefore enables rhythmic bursting. Non-coupled HMs are much less prone to bursting. The scheme is highly simplified and omits inhibitory interneurons. EAAT2; excitatory amino acid transporter 2.

The third major contributor to busting in the nucleus hypoglossus was the mGluR1 receptor class. Our previous work has demonstrated that activation of such receptors triggers sustained oscillations of rat HMs via a combination of effects including increased release of glutamate, enhanced neuronal resistance and membrane depolarization (Sharifullina et al. 2004, 2005). Mere activation of such receptors by an exogenous agonist in Krebs solution is, however, insufficient for this type of bursting, possibly because they generate moderate membrane depolarization only and glutamate uptake protects the network from widespread excitation. When uptake was blocked by TBOA, glutamate acting on these receptors not only facilitated further glutamate release, but it also increased membrane resistance to render neurons electrotonically more compact and thus more sensitive to excitatory inputs. Since strong bursting of rat spinal motoneurons is controlled by processes like cyclic operation of the electrogenic Na⁺–K⁺ pump and synaptic fatigue (Rozzo et al. 2002), it seems feasible that analogous mechanisms were responsible for the termination of single bursts in HMs. A further contributor might be pulsatile release of glutamate during uptake block as suggested for bursts generated in the cerebral cortex (Demarque et al. 2004).

Correlation between bursting and excitotoxicity

Continuous electrical recording or Ca²⁺ imaging showed that, after a period of bursting, an irreversible large inward current or baseline [Ca²⁺]_i rise developed. [Ca²⁺]_i imaging indicated that bursting became disorganized and asynchronous among HMs. These data thus suggested HM damage. Not all imaged HMs displayed these effects, indicating that, at least within the framework of the present study, there was no generalized excitotoxic death of HMs. Previous experiments with purified cultures of motoneurons have also shown that glutamate excitotoxicity is not a universal phenomenon despite a rise in [Ca²⁺]_i, and that some subsets of motoneurons are more sensitive than others (Fryer et al. 1999). It is also interesting that, in ALS, motoneuron degeneration shows patchy disease progression (Rowland & Shneider, 2001; Bruijn et al. 2004). While a long exposure to glutamate uptake blockers has been used to generate excitotoxicity in cell cultures (Bonde et al. 2003), our report shows significant excitotoxic damage to HMs after applying an uptake blocker for 1 h only.

Using vital staining of HMs to quantify their excitotoxic damage, it was apparent that the number of PI(+) HMs (thus with severe membrane lesion) reached nearly half of the counted ones. On average, this amounted to a significant 43% increase. We propose that the strong, repeated excitation due to bursting was the functional substrate promoting the intracellular metabolic dysfunction leading to the excitotoxic lesion (Rao & Weiss, 2004). In accordance with our model (Fig. 8), we observed that each one of the pharmacological antagonists that blocked bursting also prevented excitotoxicity. Hence, the early inward current of relatively small amplitude induced by TBOA on all HMs was per se inadequate to produce HM damage during the time protocol of the present study, because drugs that prevented HM excitotoxicity did not abolish this current. Such a current was probably due to ambient glutamate directly affecting receptors on the HM membrane and did not differ among bursting and non-bursting HMs.

Since the present model was obtained from data collected using neonatal brain slices, it remains to be established whether analogous mechanisms might be applicable to adult brainstem neurons. In principle, this seems likely because glutamate uptake is already expressed in the neonatal brain and readily demonstrated even in primary cultures (Danbolt, 2001). In addition, TBOA neurotoxicity is reported to occur when this agent is bath-applied to neuronal cultures from neonatal animals as much as when it is microinjected into the rat adult brain in vivo (Selkirk et al. 2005), indicating a widespread potential for neurotoxicity when glutamate uptake is blocked.

A recent theory (Cleveland & Rothstein, 2001) to account for the timing and selectivity of motoneuron killing in neurodegenerative diseases proposes that this process arises from the unfortunate convergence of a series of factors (genetic, environmental, metabolic), all of which are necessary to place motoneurons at risk, whereas each one in isolation is insufficient (the ‘convergence model’). Our results suggest that, in the case of vulnerable HMs, there might be additional convergence of molecular (glutamate receptors) and histological (gap junctions) factors that amplify the likelihood of motoneuron death when the uptake of glutamate is impaired.

	Supplemental material

Top Abstract Introduction Methods Results Discussion Supplemental material References

 
The online version of this paper can be accessed at: DOI: 10.1113/jphysiol.2005.100412
http://jp.physoc.org/cgi/content/full/jphysiol.2005.100412/DC1
and contains supplemental material consisting of a figure and legend entitled: Patch clamp recording (under voltage clamp at –65 mV) from HMs in the presence of TBOA (50 µM) in strychnine and bicuculline solution.

This material can also be found as part of the full-text HTML version available from http://www.blackwell-synergy.com

	References

Top Abstract Introduction Methods Results Discussion Supplemental material References

 
Babot Z, Cristofol R & Sunol C (2005). Excitotoxic death induced by released glutamate in depolarized primary cultures of mouse cerebellar granule cells is dependent on GABA_A receptors and niflumic acid-sensitive chloride channels. Eur J Neurosci 21, 103–112.[CrossRef][Medline]

Bergles DE, Diamond JS & Jahr CE (1999). Clearance of glutamate inside the synapse and beyond. Curr Opin Neurobiol 9, 293–298.[CrossRef][Medline]

Bonde C, Sarup A, Schousboe A, Gegelashvili G, Zimmer J & Noraberg J (2003). Neurotoxic and neuroprotective effects of the glutamate transporter inhibitor DL-threo-beta-benzyloxyaspartate (DL-TBOA) during physiological and ischemia-like conditions. Neurochem Int 43, 371–380.[CrossRef][Medline]

Boone TB & Aldes LD (1984). The ultrastructure of two distinct neuron populations in the hypoglossal nucleus of the rat. Exp Brain Res 54, 321–326.[Medline]

Bosel J, Gandor F, Harms C, Synowitz M, Harms U, Djoufack PC et al. (2005). Neuroprotective effects of atorvastatin against glutamate-induced excitotoxicity in primary cortical neurones. J Neurochem 92, 1386–1398.[CrossRef][Medline]

Bracci E, Ballerini L & Nistri A (1996). Spontaneous rhythmic bursts induced by pharmacological block of inhibition in lumbar motoneurons of the neonatal rat spinal cord. J Neurophysiol 75, 640–647.[Abstract/Free Full Text]

Brasnjo G & Otis TS (2001). Neuronal glutamate transporters control activation of postsynaptic metabotropic glutamate receptors and influence cerebellar long-term depression. Neuron 31, 607–616.[CrossRef][Medline]

Bruijn LI, Miller TM & Cleveland DW (2004). Unraveling the mechanisms involved in motor neuron degeneration in ALS. Annu Rev Neurosci 27, 723–749.[CrossRef][Medline]

Campbell SL & Hablitz JJ (2004). Glutamate transporters regulate excitability in local networks in rat neocortex. Neuroscience 127, 625–635.[CrossRef][Medline]

Cavelier P, Hamann M, Rossi D, Mobbs P & Attwell D (2005). Tonic excitation and inhibition of neurons: ambient transmitter sources and computational consequences. Prog Biophys Mol Biol 87, 3–16.[CrossRef][Medline]

Cleveland DW & Rothstein JD (2001). From Charcot to Lou Gehrig: deciphering selective motor neuron death in ALS. Nat Rev Neurosci 2, 806–819.[CrossRef][Medline]

Danbolt NC (2001). Glutamate uptake. Prog Neurobiol 65, 1–105.[CrossRef][Medline]

Del Cano GG, Millan LM, Gerrikagoitia I, Sarasa M & Matute C (1999). Ionotropic glutamate receptor subunit distribution on hypoglossal motoneuronal pools in the rat. J Neurocytol 28, 455–468.[CrossRef][Medline]

Demarque M, Villeneuve N, Manent JB, Becq H, Represa A, Ben Ari Y & Aniksztejn L (2004). Glutamate transporters prevent the generation of seizures in the developing rat neocortex. J Neurosci 24, 3289–3294.[Abstract/Free Full Text]

Donato R & Nistri A (2000). Relative contribution by GABA or glycine to Cl^–-mediated synaptic transmission on rat hypoglossal motoneurons in vitro. J Neurophysiol 84, 2715–2724.[Abstract/Free Full Text]

Essin K, Nistri A & Magazanik L (2002). Evaluation of GluR2 subunit involvement in AMPA receptor function of neonatal rat hypoglossal motoneurons. Eur J Neurosci 15, 1899–1906.[CrossRef][Medline]

Fabbro A, Skorinkin A, Grandolfo M, Nistri A & Giniatullin R (2004). Quantal release of ATP from clusters of PC12 cells. J Physiol 560, 505–517.[Abstract/Free Full Text]

Fryer HJ, Knox RJ, Strittmatter SM & Kalb RG (1999). Excitotoxic death of a subset of embryonic rat motor neurons in vitro. J Neurochem 72, 500–513.[CrossRef][Medline]

Honma S, De S, Li D, Shuler CF & Turman JE Jr (2004). Developmental regulation of connexins 26, 32, 36, and 43 in trigeminal neurons. Synapse 52, 258–271.[CrossRef][Medline]

Huang YH & Bergles DE (2004). Glutamate transporters bring competition to the synapse. Curr Opin Neurobiol 14, 346–352.[CrossRef][Medline]

Huang H & Bordey A (2004). Glial glutamate transporters limit spillover activation of presynaptic NMDA receptors and influence synaptic inhibition of Purkinje neurons. J Neurosci 24, 5659–5669.[Abstract/Free Full Text]

Huang YH, Sinha SR, Tanaka K, Rothstein JD & Bergles DE (2004). Astrocyte glutamate transporters regulate metabotropic glutamate receptor-mediated excitation of hippocampal interneurons. J Neurosci 24, 4551–4559.[Abstract/Free Full Text]

Ichikawa T & Hirata Y (1990). Organization of choline acetyltransferase-containing structures in the cranial nerve motor nuclei and lamina IX of the cervical spinal cord of the rat. J Hirnforsch 31, 251–257.[Medline]

Ichikawa T & Shimizu T (1998). Organization of choline acetyltransferase-containing structures in the cranial nerve motor nuclei and spinal cord of the monkey. Brain Res <