| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
MS 9129 Received 4 January 1999; accepted after revision 19 February 1999.
 |
ABSTRACT |
|---|
 TopAbstract IntroductionMethodsResultsDiscussionReferences |
|---|
|
INTRODUCTION |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionReferences |
|---|
In the central nervous system rhythmic oscillations are an important property for diverse functions such as cognitive activities (Jefferys et al. 1996) or movements (Rossignol & Dubuc, 1994; Arshavsky et al. 1997), and rely on a repertoire of cellular and network characteristics often difficult to unravel in complex neural structures. One useful model for studying such oscillatory properties is the isolated mammalian spinal cord which contains a network (termed central pattern generator; CPG) generating rhythmic activity (for example, that responsible for the locomotor programme) even in the absence of external input or feedback (Rossignol & Dubuc, 1994; Arshavsky et al. 1997). In this preparation such rhythmic patterns are induced by N-methyl-D-aspartate (NMDA) (Kudo & Yamada, 1987), 5-hydroxytryptamine (5-HT) (Cazalets et al. 1992; Beato et al. 1997), or even by raising the extracellular [K+] (Bracci et al. 1998); all these conditions activate CPG interneurones to drive motoneurones regarded as output elements of the system (Rossignol & Dubuc, 1994; Arshavsky et al. 1997). A fundamental question is whether the wiring properties of the network endow it with the ability to express a certain pattern or whether in the CPG there are distinct oscillatory cells which trigger this type of activity. While recent reports have found cells in the ventral horn (Kiehn et al. 1996) or around the central canal (Hochman et al. 1994) displaying intrinsic membrane oscillations, it would be useful to study the cellular properties of CPG interneurones grown in tissue culture as this approach could also provide information concerning any developmentally regulated change in CPG activity, which is known to take place in embryonic life (Kudo et al. 1991).
In an attempt to provide a suitable model for such studies, we examined the rhythmogenic properties of spinal interneurones in the organotypic slice culture from the rat spinal cord (Streit et al. 1991; Spenger et al. 1991; Ballerini & Galante, 1998). This preparation allows visualization of interneurones in a structure which maintains the basic cytoarchitecture of a spinal segment. Once synaptic inhibition is pharmacologically blocked, ventral interneurones in organotypic culture express spontaneous rhythmic bursts (Ballerini & Galante, 1998) analogous to those found in the isolated spinal cord (Bracci et al. 1996a, b), implying the presence of a CPG mechanism. The present study reports for the first time that, by raising neuronal excitability with high-K+ solutions, ventral interneurones in the organotypic spinal slice generated a rhythmic pattern widely distributed within the ventral horn area and manifested as bursts of synaptic currents (or potentials) due to activation of receptors for glutamate, GABA and glycine. Under current clamp conditions, depolarizing potentials were usually followed by hyperpolarizing potentials as described in the isolated spinal cord of the rat (Cazalets et al. 1996; Raastad et al. 1997), indicating temporal patterning of excitatory and inhibitory signals.
|
METHODS |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionReferences |
|---|
Culture preparation
Organotypic cultures of spinal cord and dorsal root ganglia (DRG) were obtained from rat embryos at days 13-14 of gestation (E13-14) following a procedure described by Braschler et al. (1989) and Ballerini & Galante (1998). The fetuses were delivered by Caesarian section from anaesthetized rats (10·5 % chloral hydrate, 0·4 ml (100 g)-1 I.M.) subsequently killed by an intracardiac injection (2 ml) of chloral hydrate. This procedure is in accordance with the regulations of the Italian Animal Welfare Act and is approved by the local authority veterinary service. After decapitation of the fetuses their backs were isolated and cut into 270 µm thick transverse slices by means of a tissue chopper. The spinal cord with its DRG was then separated from the rest of the slice and fixed on a glass coverslip with reconstituted chicken plasma (Cocalico, Reamstown, PA, USA) clotted with thrombin (Sigma). The coverslips were inserted into plastic tubes containing 1·5 ml of medium. These tubes were kept in a roller drum rotating at 120 r.p.h. in an incubator at 36·5°C in the presence of dry atmosphere with 5·2 % CO2. The medium contained 82 % Dulbecco's modified Eagle's medium (D-MEM; Gibco), 8 % sterile water for tissue culture (Gibco), 10 % fetal calf serum (Gibco) and 5 ng ml-1 nerve growth factor (Alomone Laboratories, Jerusalem, Israel); osmolarity, 300 mosmol l-1, pH 7·35.
Experimental procedure and drug solutions
The experiments were performed on slices grown for 12-18 days in culture (days in vitro; DIV). A coverslip with the culture was positioned in a Perspex chamber mounted on an inverted microscope (Nikon TE200) and superfused with normal Krebs solution containing (mM): NaCl, 156; KCl, 4; MgCl2, 1; CaCl2, 2; Hepes, 10; glucose, 10. The pH was adjusted to 7·4 using NaOH (osmolarity, 305 mosmol l-1). In the experiments with high [Mg2+]o and low [Ca2+]o ([Mg2+]/[Ca2+] = 5/1), the [CaCl2]o was lowered to 0·5 mM while [MgCl2]o was increased to 2·5 mM. In Mg2+-free solution, MgCl2 was omitted and the concentration of NaCl was increased to 157 mM or CaCl2 to 3 mM. All the experiments were performed at room temperature (22 ± 2°C). High-K+ solution was obtained by adding the required amount of KCl (1 M stock) to the normal Krebs solution. All agents were bath applied; these included 3-((RS)-2-carboxy-piperazine-4-yl)-propyl-1-phosphonate (CPP; Tocris), 6-cyano-7-nitroquinoxaline-2,3-dione (CNQX; Tocris), 
-amino-3-hydroxy-5-methyl-4-isoxazolepropionate (AMPA; Tocris), NMDA (Tocris), bicuculline methiodide (Sigma), strychnine nitrate (Sigma), QX-314 (Alomone Laboratories) and tetrodotoxin (TTX; Affiniti Research Products, Exeter, UK).
Electrophysiological recordings
Voltage clamp recordings in the whole-cell configuration were obtained from ventrally located spinal neurones using low-resistance glass pipettes (4-6 M
) filled with a solution of the following composition (mM): potassium gluconate, 120; KCl, 20; Hepes, 10; EGTA, 10; MgCl2, 2; Na2ATP, 2 (pH 7·35, adjusted with KOH). Ventral interneurones were identified on the basis of criteria described previously (Spenger et al. 1991; Streit et al. 1991; Ballerini & Galante, 1998). Visual identification of these cells was aided by Nomarski optics coupled with an infrared microscopy system mounted on the Nikon microscope. Series resistance compensation was not routinely employed as series resistance values were consistently below 15 M. Previous experiments have shown that these experimental conditions allow recording of fast synaptic events without apparent distortion (Ballerini & Galante, 1998). Responses were acquired with an EPC-7 amplifier (List, Germany) at 10 kHz, stored on videotape for further analysis, digitized at 10 kHz with pCLAMP software (version 6.2; Axon Instruments), and displayed on a chart recorder. Extracellular field potential recordings were simultaneously performed with patch clamp recordings by placing a 2 M NaCl-filled microelectrode (2-4 M) in a certain region of the culture. Responses were amplified (Axoclamp-2B; Axon Instruments), stored on videotape and displayed on a chart recorder. Extracellular electrical stimulation of the DRG cells was performed with a low-resistance patch pipette containing external bath solution. Short voltage pulses (100 µs) of various amplitude (5-50 V) were delivered by a Digitimer stimulator. For current clamp recording from interneurones we used patch pipettes filled with (mM): potassium gluconate, 140; Hepes, 10; EGTA, 10; MgCl2, 2; Na2ATP, 2; QX-314, 0·5 (used to minimize voltage-activated Na+ currents). The pH of the solution was adjusted to 7·35 by addition of KOH. Responses (obtained in bridge mode after negative capacity compensation and with continuous monitoring of bridge balancing) were amplified by an Axoclamp-2B unit and processed in the same way as those collected under voltage clamp conditions. Current clamp records indicated that interneurones had spontaneous resting potentials of -58 ± 7 mV (n = 7 neurones), values thus close to that used for holding cells under voltage clamp. Cells readily fired action potential trains with minimal accommodation (see Fig. 6C) when the recording electrode did not contain QX-314, indicating their healthy condition. To estimate the Cl- reversal potentials we measured, in the presence of TTX (1 µM), the value of the reversal of the responses (n = 3) evoked by exogenously applied GABA (0·1 mM) in voltage or current clamp mode: these values were -52 mV (24 mM internal Cl-; voltage clamp) and -63 mV (4 mM internal Cl-; current clamp). Membrane potential values were systematically corrected for liquid junction potentials and other offset potentials (see Marty & Neher, 1983).
Statistical analysis
Results are presented as means ± S.D.; n, number of neurones. ANOVA was used to assess statistical significance of the data. Rhythmic activity was characterized by cycle period, measured as the time between the onset of two oscillations, and by burst duration, measured as the time between the onset of one oscillation and 90 % of baseline recovery. The rhythm regularity was quantified by the coefficient of variation of cycle period (CVcp) and the coefficient of variation of burst duration (CVd), expressed as a percentage. I-V curves were constructed by plotting the peak amplitude of bursts (from baseline) versus the level of imposed holding potential.
Histology
The microscopic structure of the organotypic slice cultures used for the present study was investigated to establish the approximate degree of multilayer organization. When the slice was viewed with epifluorescence microscopy in its normal plane after propidium iodide (1 mM in phosphate-buffered saline (PBS) supplemented with 0·4 mg ml-1 RNAse A) labelling, it could be seen that, after 14 days in culture, the slice contained three to four cell layers, while the DRG cells were organized as a monolayer. Propidium iodide fluorescence was elicited by excitation with an Ar-Kr laser at 568 nm, using a rhodamine filter set, and detected by the photomultiplier tube of a Multi Probe 2001 confocal laser scanning microscope (Molecular Dynamics, Sunnyvale, CA, USA). A vertical scan of the slice culture was performed, and a series of optical sections obtained (one section every 3 µm). The ImageSpace program (Molecular Dynamics) was used to analyse the peak of cell nuclei fluorescence, allowing the three to four cell layer organization of the slice to be determined.
Previous experiments using Lucifer Yellow staining of single ventral horn interneurones have indicated the main morphological features of these cells (Ballerini & Galante, 1998). In the present study, further checks were undertaken in order to exclude the possibility that recordings were from motoneurones. For this purpose, organotypic slices were stained using the choline acetyltransferase (ChAT) immunocytochemical technique (Phelps et al. 1991) to localize motoneurones. Cultures grown for 2 weeks in vitro were fixed for 1 h with 4 % paraformaldehyde in PBS, then washed in PBS and treated with 2·3 % sodium metaperiodate in H2O for 5 min at room temperature. After a rapid wash in H2O, cultures were maintained in 1 % sodium borohydride in 100 mM Tris-HCl (pH 7·5) for 10 min followed by a wash in PBS plus 0·1 % (v/v) Tween 20 (PBST) and incubation overnight at 4°C with goat anti-ChAT polyclonal antibodies (Chemicon Int., Temecula, CA, USA) diluted 1 : 200 in 10 % fetal calf serum (FCS) in PBST. After several washes in PBST, cultures were incubated in biotinylated anti-goat IgG antibody (Vector, Burlingame, CA, USA) diluted 1 : 100 in 10 % FCS in PBST. Two hours later the reaction was revealed using the ABC kit (Vector). Figure 1A and B shows an example of ChAT-reactive cells in the slice preparation. The low power microphotograph (A) shows their sparse distribution clustered around the central fissura (asterisk; see Streit et al. 1991) while at higher magnification (B) these stained cells (e.g. the cell indicated by the arrow) appeared as large diameter neurones (> 20 µm diameter) with a few processes. Only in rare instances were the ChAT-stained cells found to give processes which clustered into a multifibre projection exiting the spinal slice. No apparent organization of the organotypic slice culture in Rexed laminae was found even after 21 DIV; this is not surprising since the tissue had been obtained at E13-14 when the architectural arrangement in the spinal cord is poor and no direct monosynaptic conections to motoneurones from dorsal root afferents are established (Ziskind-Conhaim, 1988). However, the majority of cells in this area remained unstained suggesting that they were ChAT-negative interneurones. Further experiments were carried out to assess the localization within the same slice culture area of interneurones and motoneurones. For this purpose single putative interneurones were patch clamped and filled intracellularly with biocytin (0·2 %) for about 60 min; the stain was developed using the ABC kit. The slice was then subsequently processed for ChAT immunohistochemistry as described above, except that the second antibody was directly conjugated with alkaline phosphatase (Sigma). Figure 1C shows that in the same ventral area of the slice culture there was only one cell labelled with biocytin which corresponded to a round-shaped monopolar neurone distinct from a triangle-shaped ChAT-labelled cell presumably corresponding to a motoneurone. Similar observations were obtained in eight slice preparations. Further, functional identification of interneurones was obtained by observing the apparent lack of spike accommodation properties of these cells following a depolarizing current pulse (see example in Fig. 6C); this phenomenon contrasts with the well-known spike accommodation displayed by rat motoneurones (Fulton & Walton, 1986). Note that all cells presumed to be interneurones were found to possess the same firing pattern, suggesting that they represented a relatively homogeneous population. For the present experiments, routine electrophysiologial recordings were obtained from cells (< 20 µm somatic diameter) which fulfilled the criteria for interneuronal identification on the basis of their shape and/or functional properties and were directly visualized in the ventral area at some distance from the ventral fissure.
|
RESULTS |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionReferences |
|---|
Results were obtained from 101 neurones voltage clamped at a membrane potential (Vh) of -56 mV and seven neurones recorded under current clamp conditions. Cells were identified as interneurones by several criteria: histological staining (see Methods and Fig. 1), firing properties (see Methods and Fig. 6C) and typical appearance of small diameter monopolar cells directly visualized with Nomarski optics coupled with infrared microscopy. Previous work (Ballerini & Galante, 1998) has shown that single ventral interneurones possessing the characteristics outlined above generate spontaneous rhythmic activity following coapplication of pharmacological blockers of synaptic inhibition. The present study investigated whether rhythmic activity could also be induced by increasing excitability (via a rise in extracellular [K+]) instead of suppressing inhibition.
 |
View larger version [in this window] [in a new window] |
|
|
A, localization of ChAT-positive cells (arrow) in a 14 DIV spinal cord culture. The central fissura (asterisk) is a topographical marker of the ventral area. Note the ventral, bilateral distribution of positive cells. Calibration, 200 µm (× 5 objective). B, ventral motoneurones visualized by ChAT immunostaining. Same slice as in A (arrow). Note the large somata and the multipolar dendrites of these cells. The central fissura (asterisk) is outlined with a dashed line. Calibration (see bar in A), 75 µm (× 20 DIC (differential interference contrast) objective). C, double staining in a different spinal cord culture (14 DIV). Note the ChAT-positive neurone (arrow) and the ChAT-negative neurone (arrowhead). Calibration (see bar in A), 50 µm (× 20 objective). D, ChAT-positive neurone; note the large and triangle-shaped soma. Calibration (see bar in A), 25 µm (× 40 objective). E, ChAT-negative neurone; note the small dimension of the cell soma. The long time taken for the cell to be filled with biocytin from the patch pipette (the position is indicated by the arrowhead; see Methods) resulted in the soma being stained dark brown. Calibration (see bar in A), 25 µm (× 40 objective). | ||
Effects of increased extracellular [K+]
In all ventral interneurones tested under voltage clamp conditions (n = 43) an increase in K+ concentration from 4 to 7 mM induced a slow inward current as exemplified by the continuous trace in Fig. 2A (middle; the arrow indicates the start of 7 mM K+ superfusion) which on average peaked 30 ± 10 s from the onset of application, reached a mean amplitude value of -36·4 ± 20·2 pA (measured at mid-point noise level), and was associated with a large enhancement in spontaneous synaptic activity (as indicated by a much thicker baseline in Fig. 2A, middle, even when the gain was 2-fold lower). Spontaneous events identified as activity-dependent synaptic currents (see Ballerini & Galante, 1998) were also observed in control solution (Fig. 2A, left), but were less frequent and lacked periodicity. The K+-induced inward current gradually declined from its peak despite the continuous application of this modified solution and eventually reached a plateau level while the superimposed synaptic activity continued and, in 35/46 neurones, displayed periodic patterning (as exemplified in Fig. 2B, middle; note 10-fold faster speed) which, despite the gradual decline in the baseline inward current, persisted 5 min later (Fig. 2A, right). The latter periodic events were then averaged (20 events) to reconstruct a mean current burst as shown in Fig. 2B (left) with 26 ms rise time and 64 ms monoexponential decay. The induced patterned activity was readily abolished by washout to control solution, which was always accompanied by a return to baseline current (not shown).
 |
View larger version [in this window] [in a new window] |
|
|
A: left, spontaneous synaptic activity in control solution. Middle, increased extracellular K+ (7 mM; arrow indicates the start of application) induced a slow inward current which reached a peak value of -65 pA 40 s later (note smaller gain). This response was accompanied by a large increase in spontaneous synaptic activity and a gradual decline of the inward current towards baseline during continuous K+ superfusion. In the presence of high K+, spontaneous synaptic activity turned into a patterned activity within 2 min from the start of K+ application (expanded trace in B, middle), consisting of rapid bursts of inward current. Right, this pattern persisted with a regular although slower rhythm at 5 min of K+ superfusion. B, left trace depicts an average of 20 events taken at 5 min of K+-induced bursting (same cell as in A). Right trace, in the presence of K+ spontaneous bursting was suppressed by CNQX (same cell as in A). C, graphs showing burst cycle period (left) and duration (right) plotted against three different concentrations of extracellular K+. Cycle period and burst duration values were collected at 2 min ( | ||
The rhythm observed during the initial 2 min of 6-7 mM K+ application consisted of events occurring at a cycle period of 1·58 ± 1·20 s (Fig. 2C, left; filled squares) and with a mean duration of 0·65 ± 0·21 s (Fig. 2C, right; filled squares). Whereas in a small minority of bursting neurones (6/35 cells) this rhythmicity disappeared within 5 min from the start of high-K+ application, in the majority of cells (n = 29) this pattern persisted as long as this high-K+ solution was applied (> 25 min) and stabilized at a period value which was not significantly different from that observed after 2 min application, as shown in Fig. 2C (left; filled triangles). In the example shown in the right panel of Fig. 2A (note higher gain than in middle panel as event amplitude had decreased), the period value was 2·8 ± 1·5 s at 5 min. Burst duration was also similar at 2 or 5 min application as indicated by the plot of mean values shown in Fig. 2C (right; filled triangles). Regular occurrence and duration of persistent rhythmic activity was indicated by low values of averaged CVcp and CVd (20 ± 5 and 19 ± 10 %, respectively).
The threshold concentration of K+ sufficient to reliably induce rhythmic activity was 6 mM (n = 5). No significant (P > 0·05) difference in cycle period or burst duration was detected between 6 or 7 mM K+ measured at both 2 and 5 min from application (Fig. 2C). With higher concentrations of K+ (8 mM; n = 3) a significant (P < 0·001) shortening of cycle period and burst duration was observed; both parameters were similar at 2 or 5 min from the start of K+ superfusion (Fig. 2C). In three interneurones in which 6-8 mM K+ was effective in inducing bursting activity, the effects of 10 mM K+ were also investigated (not shown); in these cells this high-K+ solution evoked a large inward current with a reversible decrease in spontaneous synaptic activity which lacked any rhythmicity.
Ventral horn distribution of rhythmicity
The high-K+-induced activity appeared to be widely distributed within the organotypic slice. This was observed by simultaneously recording the activity of a single cell under voltage (patch) clamp and the network responses with an extracellular field electrode (see schematic drawing in Fig. 3B and the slice microphotograph in Fig. 1A). In this example the extracellular electrode (Extra) was positioned in the contralateral ventral horn with respect to the patched ventral interneurone (WCR). The 7 mM K+ concentration induced a patterned activity of 0·98 ± 0·4 s cycle period and 0·29 ± 0·1 s burst duration at 2 min from the application in the recorded interneurone (Fig. 3Ab), which at 5 min stabilized at values of 1·2 ± 0·1 and 0·19 ± 0·04 s, respectively (Fig. 3Bb). The field recording (Fig. 3Aa and Ba) showed comparable rhythmic events (of different polarity) in the controlateral ventral horn occurring 50-100 ms after the large events recorded from the patched neurone. Similar results were obtained in four different organotypic slices. The widely observed activity found in the ventral horn area was not present in DRG cells. In fact, in two different culture dishes, after the regular pattern of activity had been recorded from ventral interneurones in the presence of 7 mM K+, a DRG cell was also patch clamped during K+ superfusion. In both tests, no spontaneous activity was detected under control conditions and no bursting appeared after K+ application, although an inward current was present, similar to that induced by K+ in ventral interneurones (not shown). Systematic mapping of the distribution of rhythmic activity within ventral and dorsal subfields of the slice culture will require further investigation. However, the temporal phasing of the bursting pattern on either side of the slice was studied by simultaneous, extracellular recording of bursting activity from two contralateral areas of the same slice culture once rhythmic patterns were established in high-K+ solution. In these cases, the delay in the burst onset between the two ventral areas was 37 ± 15 ms (5 slices), a value relatively close to the latency (35 ± 4 ms; 32 recordings, in 4 different preparations) of the excitatory inward current evoked by focal electrical stimulation of a DRG on a homolateral ventral interneurone.
 |
View larger version [in this window] [in a new window] |
|
|
A, simultaneous field (a) and whole-cell recordings (b). Perfusion of K+ (arrowhead) induced an inward current in the recorded interneurone (b) with enhanced spontaneous synaptic activity, shown on a slow time base. In the presence of K+ this activity developed into a regular pattern, which could be detected with both electrodes although with opposite polarity, as more clearly shown by the faster records on the right. Note the similarity in cycle period and burst duration between the two recordings. The right-hand time calibration in b also applies to a. B: left, schematic drawing of the position of the field (Extra) and patch (WCR) electrodes within the organotypic slice. Note that the two electrodes were placed in the two contralateral ventral horns. In the presence of high K+, spontaneous rhythmic activity stabilized at 5 min in both ventral horns as shown by the two tracings (field (a) and patch (b) recordings). Note that the extracellular recorded events occurred with a delay from the patch recorded ones. The current calibration bar in B also applies to A. | ||
Sensitivity of K+-induced activity to pharmacological agents
Pharmacological tests were initiated to find out the nature of the rhythmic activity evoked by high K+. The non-NMDA glutamate receptor blocker CNQX (10 µM) fully abolished the rhythmic pattern evoked by 7 mM K+ superfusion, as can be demonstrated by comparing the trace in Fig. 2A (right) with that in Fig. 2B (right) in which a very high gain was used to detect even small responses. However, the slow inward current generated by K+ application was unaffected (not shown). CNQX completely blocked K+-induced bursts and strongly decreased spontaneous synaptic activity in four of four cells. Attempts to replicate the bursting activity in high-K+ solution with the glutamate receptor agonist AMPA met with limited success. In fact, AMPA (1 µM) induced a slow inward current with superimposed intense synaptic activity, which only transiently (after about 2-3 min) became organized into a rhythmic pattern (3/4 neurones) lasting at most 1-2 min. The evanescent nature of this response prevented further studies of its characteristics, as was also found in the case of the rat isolated spinal cord (E. Bracci, M. Beato & A. Nistri, unpublished observations).
In Fig. 4A and B, the effects of NMDA receptor block are shown. In Fig. 4B (left), 7 mM K+ induced rhythmic activity with cycle period and burst duration of 1·9 ± 0·5 and 0·7 ± 0·2 s, respectively, in a representative cell. In this neurone, superfusion with the NMDA receptor antagonist CPP (10 µM) reduced burst duration (to 45 ± 5 % of control) without affecting cycle period, which was 83 ± 10 % of control. Figure 4A (right) shows the superimposed average of five bursts, in the absence and presence of CPP (same cell as in Fig. 4B); when CPP was present, bursts had reduced mean amplitude (-136 and -95 pA, respectively) with slower rise time (50 %) and a faster decay time constant (55 ms in the presence of the drug versus 65 ms in control). The plot in Fig. 4A (left) summarizes results obtained from five ventral interneurones, in which CPP significantly (P < 0·001) reduced burst duration (to 60 ± 22 %) while it did not change the mean cycle period (105 ± 50 %). Application of TTX (2 µM) completely blocked bursts evoked by 7 mM K+ while it did not abolish the slow inward current associated with this treatment and strongly reduced the amplitude and frequency of background spontaneous activity (n = 3; not shown).
 |
View larger version [in this window] [in a new window] |
|
|
The NMDA receptor antagonist CPP was superfused after the K+-induced rhythm had stabilized. A: left, histograms of cycle period and burst duration recorded after 15 min application of CPP (10 µM) in 5 interneurones, expressed as a percentage of the rhythm observed in high-K+ solution. A significant reduction (** P < 0·001) in burst duration was observed. Right, averaged traces of 5 consecutive burst events in the absence or in the presence (arrow) of CPP are superimposed. Note the reduction in amplitude and duration brought about by CPP. B: left, example of 7 mM K+-induced rhythmic activity. Right, the burst duration and amplitude were reduced by CPP. Different cell from A (right). | ||
Voltage dependence of rhythmic activity
This issue was explored using either voltage or current clamp experiments. In the first case, after > 5 min application of 6 or 7 mM K+ Vh was changed within the range -66 to +26 mV (each value being maintained for > 2 min) as indicated in Fig. 5A (6 mM K+ solution); at a Vh of -66 mV bursts appeared as multiphasic inward currents which at -36 mV were reduced in amplitude (while their duration and periodicity were preserved) and could not be fractionated into discrete components. These responses disappeared at -18 mV (see data point in the I-V plot in Fig. 5B, inset), and returned with opposite polarity at +16 mV (Fig. 5A, bottom). Figure 5B shows that in a sample of six interneurones neither cycle period nor burst duration was affected by changes in Vh. In the negative membrane potential range the burst I-V relation was non-linear, indicating that these events had a complex origin.
 |
View larger version [in this window] [in a new window] |
|
|
A, bursting currents were inward at a Vh of -66 mV (top), decreased in amplitude at -36 mV (middle) and were outward at a Vh of +16 mV (bottom). Traces were digitized at 5 kHz. The current calibration for the middle trace also applies to the top trace. B, data shown are from 6 interneurones for which cycle period (top) and burst duration (bottom) are plotted against Vh. Values are expressed as a percentage of cycle period and burst duration recorded in each neurone at a Vh of -56 mV. Note that these values were not affected by changes in Vh. The inset shows the I-V curve obtained by plotting burst amplitude against Vh. Note the non-linearity in the negative potential range and that the calculated reversal potential was -18 mV. | ||
In fact, the -18 mV reversal potential of the K+-evoked bursts suggested a mixed contribution by glutamatergic as well as inhibitory, glycine- and GABA-mediated Cl- events, particularly because the raised intracellular Cl- concentration (due to the voltage clamp pipette solution) would have set the inhibitory transmitter reversal potential at a relatively less negative level (this value was actually found to be -52 mV; see Methods). It should be noted that the outward rectification of the I-V plot occurred at around -50 mV, even if the inhibitory currents became inward and would thus tend to oppose it. To observe whether excitatory and inhibitory currents made discrete contributions to the rhythmic pattern we therefore pursued a different approach by performing a number of current clamp experiments in which the intracellular solution contained a low Cl- concentration (to keep inhibitory reversal potentials further away from excitatory potentials) and the Na+ channel blocker QX-314 to suppress spiking generated during bursting (which would otherwise obscure identification of synaptic components). Consequently this approach should allow not only a better separation between excitatory and inhibitory events but also validation that the observed responses were not artefactually generated by inadequate space clamping of neurones with a widely distributed dendritic arbor. Figure 6 shows an example of such experiments. At -42 mV (see Fig. 6A, middle trace) bursting (induced by 7 mM K+) was expressed as a series of events with an initial depolarizing component (filled triangles) followed by complex hyperpolarizing events (open triangles) intertwined with smaller amplitude depolarizing components. Shifting the membrane potential to -66 mV (top trace in Fig. 6A) by intracellular injection of -0·04 nA maintained a similar rhythmicity in which a large depolarizing response was followed by slower components of similar polarity with superimposed faster events (filled triangles) without change in periodicity. When the cell was depolarized to -12 mV by +0·07 nA (Fig. 6A, bottom trace) each burst appeared to comprise mainly a large inhibitory component (open triangles) with unchanged periodicity. These data indicate that rhythmicity consisted of a series of alternating excitatory postsynaptic potentials (EPSPs) followed by inhibitory postsynaptic potentials (IPSPs). It seemed interesting to examine whether there was any tight relation between the temporal patterning of EPSPs and IPSPs. This possibility was tested at a membrane potential of -42 mV by correlating the cycle period of EPSPs and IPSPs; any spontaneous fluctuation in burst periodicity which takes place during a persistent period of bursting should similarly affect excitatory and inhibitory events as long as the two remain temporally correlated. Figure 6B shows that there was a linear relation between the period of the EPSPs (abscissa) and that of the IPSPs (ordinate) with a Pearson correlation coefficient of 0·998 (P < 0·0005). Similar results were obtained in five interneurones while in two cells comparable shifts to depolarized potentials greatly attenuated the depolarizing events to reveal a tonic background of hyperpolarizing events. In five cells, a plot of the peak amplitude of the burst events against membrane potential indicated that their estimated reversal potential was -30 mV.
 |
View larger version [in this window] [in a new window] |
|
|
A, rhythmic episodes evoked by 7 mM K+ recorded at three different levels of resting membrane potential (indicated by each trace) from the same cell patch clamped with an intracellular solution containing 4 mM Cl- and 0·5 mM QX-314 (see Methods). At a resting membrane potential of -42 mV, depolarizing potentials ( | ||
These observations thus indicate that the K+-induced rhythmic activity was made up of large, mixed synaptic events originating at network level and was expressed as a series of EPSPs usually followed by IPSPs or, in a minority of cells, superimposed upon tonic inhibition. The possibility of distinguishing these components was aided by recording them at different potentials and with a different intracellular solution. The membrane potential of the recorded cell thus determined their polarity but not their origin or temporal patterning.
Effects of increased K+ concentration on strychnine- and bicuculline-induced bursts
Our observation of IPSPs following EPSPs during each burst episode suggested that excitation was apparently accompanied by inhibition but the nature of those inhibitory components could not be identified. The latter could not be studied in isolation since blocking excitatory transmission with CNQX suppressed high-K+-evoked bursting. We thus opted to inhibit glycine or GABA receptors to ascertain whether they were responsible for the Cl--dependent inhibition recorded under the present experimental conditions (Ballerini & Galante, 1998). In this case a slower type of bursting develops spontaneously and is primarily supported by glutamate receptor activation (Ballerini & Galante, 1998). Accordingly, in the present study under voltage clamp conditions, concomitant block of glycine and GABAA receptors by coapplication of strychnine (1 µM) and bicuculline (20 µM) generated rhythmic bursting activity as exemplified in Fig. 7A (left). These bursts were different from those induced by high K+ since they consisted of repeated large plateaux of inward current (lasting several seconds) with superimposed intense synaptic activity (see expanded record in Fig. 7B, left), and had a slower period and longer duration. Application of 8 mM K+ reduced (with a latency of 2-3 min) the cycle period of spontaneous bursting episodes evoked by strychnine and bicuculline (see Fig. 7A, right; to a mean of 47 ± 18 % of control, as shown in Fig. 7B, middle; P < 0·0001, n = 5 neurones). Within each episode the high K+ decreased the burst duration of the rhythmic activity to 71 ± 15 % of control (see histogram in Fig. 7B, right). This effect of high-K+ solution was associated with a slow inward current (not shown). {define style "Single column legend";
 |
View larger version [in this window] [in a new window] |
|
|
A: left, strychnine and bicuculline induced sustained rhythmic bursting activity of a ventral horn interneurone. Right, an increased K+ concentration in the presence of strychnine and bicuculline shortened both cycle period and duration of bursting. B: left, expanded record of an individual event elicited by strychnine and bicuculline (segment of trace shown at higher gain and speed is indicated by dashed lines). Middle, plot of cycle period values observed in strychnine plus bicuculline solution (S + B) and after the subsequent increase in extracellular K+ to 8 mM, which significantly (** P < 0·0001; n = 5 cells) reduced them. Right, plot of burst duration values calculated for the same events under the same experimental conditions in which 8 mM K+ also decreased burst duration (** P < 0·0001). Data in 8 mM K+ were normalized with respect to those obtained in strychnine plus bicuculline solution. C: left, burst amplitude in the presence of strychnine and bicuculline plotted against Vh. I-V relations before ( | ||
During strychnine- and bicuculline-induced bursts, Vh was changed in a stepwise fashion in the range -66 to +36 mV, as described previously. In Fig. 7C (left) the strychnine- and bicuculline-evoked burst amplitude decreased with membrane depolarization, disappeared at 0 mV and returned with opposite polarity at +36 mV (filled circles). Application of K+ (8 mM) reduced burst amplitude uniformly throughout the voltage range although the same reversal potential value was retained (Fig. 7C, left; filled triangles). Note, however, that the burst I-V relation remained non-linear at negative potentials despite the pharmacological suppression of GABA- and glycine-mediated components. One possibility is that voltage-dependent block of NMDA receptors might have contributed to such non-linearity. This was explored (see Fig. 7C, right) by measuring the I-V curve and reversal potential of strychnine and bicuculline bursts before (open squares) and after (open triangles) application of the NMDA antagonist CPP. The plot became linear with a steeper slope after application of CPP; for example, in the range -46 to -16 mV the slope value was 0·007 in strychnine plus bicuculline solution (by linear regression analysis) and became 0·3 in the presence of CPP. Similar results were obtained in three cells. These observations thus indicate that after pharmacological block of glycine and GABAA receptors spontaneous inward currents developed with a reversal consistent with that of glutamate-mediated responses in the spinal cord (Mayer & Westbrook, 1987) and which were readily accelerated by external K+. When the cell membrane was depolarized, a more consistent contribution by NMDA receptor activity to burst amplitude was manifested. These data thus expand those obtained under current clamp conditions and provide additional evidence about the multicomponent (glutamatergic, glycinergic and GABAergic) nature of bursting currents.
Sensitivity of K+-induced activity to changes in external concentration of Mg2+ and Ca2+
While the available data suggested a network origin of the rhythmic patterns, it seemed necessary to study whether the high-K+ solution actually facilitated transmitter release from network cells. This issue was investigated indirectly by comparing K+-induced rhythmic bursting in control and modified ([Mg2+]/[Ca2+] = 5/1, see Methods) solution which should suppress synaptic transmission. Synaptic currents evoked by DRG electrical stimulation (see Methods) were also monitored in these experiments to check for the effectiveness of synaptic transmission block (see for instance average of 5 evoked responses recorded from a ventral interneurone in control solution in Fig. 8C, top). Figure 8A shows that 6 mM K+ changed spontaneous synaptic activity into bursting activity in a ventral neurone. After washout and disappearance of rhythmic bursting (not shown), subsequent superfusion with 5/1 solution largely reduced spontaneous synaptic activity, and fully prevented the K+ (6 mM)-induced rhythmic pattern (Fig. 8B). Under the same conditions, evoked responses were fully blocked (Fig. 8C, middle) and recovered after washout (Fig. 8C, bottom).
 |
View larger version [in this window] [in a new window] |
|
|
A, spontaneous synaptic activity under control conditions (left) developed into a patterned activity in the presence of high K+ (5 min; right). B, in the same cell spontaneous synaptic activity was greatly reduced in the presence of external solution with a 5/1 ratio of [Mg2+]/[Ca2+] (left; see Methods). Under these conditions addition of 6 mM K+ failed to induce rhythmic bursting (right). As these traces were not DC mounted, the steady inward current elicited by 6 mM K+ is not shown. C, synaptic currents evoked by DRG stimulation (see Methods) under control conditions (top; same cell as in A and B) disappeared in the presence of the 5/1 external solution (middle) and recovered after 10 min washout in control solution (bottom). Each panel is an average of 5 consecutive evoked responses. | ||
Mg2+-free-induced activity
In five neurones, NMDA (2-4 µM, a typical concentration range used to evoke fictive locomotion in the rat spinal cord; Kudo & Yamada, 1987; Cazalets et al. 1992) did not produce any consistent change in spontaneous activity or in baseline current suggesting that NMDA receptors might have been functionally blocked perhaps by ambient Mg2+ in a voltage-dependent fashion (Mayer & Westbrook, 1987) as also suggested by the non-linearity of the burst I-V relation. Addition of 5-HT (30 µM) to the NMDA solution did not help to elicit rhythmic bursting (data not shown). The insensitivity to NMDA was not a matter of insufficient concentration of this agent since 10 µM (n = 10 cells) or 20 µM (n = 3) NMDA also failed to elicit rhythmic patterns of activity.
To reduce the extent of any Mg2+ block of NMDA glutamate receptors we used a Mg2+-free solution (see Methods). Removal of external Mg2+ evoked an inward current of -14 ± 3 pA (in 8/13 cells) which peaked at 30 s. After 5-6 min from the onset of application of the Mg2+-free solution, at a Vh of -56 mV a regular rhythmic pattern of activity developed and persisted as long as this divalent cation was absent from the bathing solution (Fig. 9A, left). The Mg2+-free-induced pattern of activity was detected in all cells (n = 13), and was characterized by a cycle period of 4 ± 1 s and burst duration of 1·5 ± 0·7 s. The CVcp and the CVd were 21 ± 6 and 21 ± 2 %, respectively, indicating the regular occurrence of these bursts. The bursting activity evoked by omission of Mg2+ from the bathing solution was not a mere consequence of increased membrane excitability due to a decreased divalent cation concentration. In fact, in three cells in which the total extracellular divalent cation concentration was kept constant by raising external Ca2+ to 3 mM, an analogous bursting pattern appeared when a Mg2+-free solution was applied.
 |
View larger version [in this window] [in a new window] |
|
|
A: left, in the presence of Mg2+-free solution rhythmic bursting developed spontaneously after 5-10 min. Right, bursting currents were inward at a Vh of -66 mV, and outward at -10 mV. B, I-V relation of burst amplitude against Vh (same cell as in A). Note that the reversal potential was -20 mV. C, rhythmic bursting induced by superfusion with Mg2+-free solution (left) was fully blocked by CPP application (top right) but persisted in the presence of CNQX (bottom right), although with reduced amplitude and duration. In this cell, cycle period was not affected by CNQX. Different cell from A. | ||
During the Mg2+-free-induced patterned activity variations in Vh were imposed within the range -66 to +26 mV with a procedure similar to that described previously. In Fig. 9A, right (same cell depicted in the left panel at -56 mV) an example of this is shown; each burst appeared as an inward current at -66 mV while it was outwardly directed at -10 mV. The graph of Fig. 9B shows the I-V relation for burst amplitude with inversion of polarity at -20 mV. In this case the calculated slope for the -46 to -10 mV range was 1·6, indicating a greater degree of steepness when compared with that in standard Mg2+ solution (see Fig. 7C). In a sample of six cells, cycle period and burst duration were not affected by changes in Vh (data not shown).
In five cells, 10 µM CPP was applied during Mg2+-free-induced rhythmic activity. In three of five cells this agent fully blocked the on-going bursting activity, as depicted in the example of Fig. 9C (top right). In two of five neurones rhythmic activity was disrupted but irregular bursts of reduced amplitude still occurred (not shown). CNQX (10 µM) was added to the Mg2+-free solution in seven neurones. Bursting activity persisted in six of seven cells in the presence of CNQX, although it was reduced in amplitude (to 28·5 ± 7 % of control; see example in Fig. 9B, bottom right). On average, cycle period and burst duration were changed to 166 ± 44 and 135 ± 25 % of control, respectively.
|
DISCUSSION |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionReferences |
|---|
The principal finding of the present study is the novel demonstration that ventral interneurones of the organotypic spinal slice culture reliably generated a rhythmic pattern following relatively small increases in extracellular [K+] or the use of Mg2+-free solution. These results show that even such a simplified structure of the mammalian spinal cord possessed rhythmogenic ability.
Recording from organotypic culture interneurones
In vitro spinal preparations from the lamprey (Grillner et al. 1991), chick (Ho & O'Donovan, 1993) or tadpole (Dale & Kuenzi, 1997a) have contributed to our understanding of motor pattern generation mechanisms. In particular, extensive work carried out on the locomotor network of the Xenopus embryo spinal cord (see Dale & Kuenzi (1997b) for a recent review) has revealed the complex interplay between conductances operated by transmitters or by changes in membrane voltage to determine the locomotor pattern; this empirical work has also been validated by realistic modelling of CPG activity (Dale & Kuenzi, 1997a, b). To date a similarly advanced understanding of motor pattern mechanism in mammals is lacking, a situation which would thus benefit from the availability of simple models to which experimental studies analogous to those successfully employed on lower vertebrates could be applied.
Such a reductionist approach seems feasible since experiments on the mudpuppy spinal cord have indicated that the locomotor network is divisible into individual 'modular pattern generators', each one capable of expressing regular oscillatory activity (Cheng et al. 1998). In the rat spinal cord a ventral quadrant comprising only two to three segments already contains a simple circuit for rhythmic bursting (Bracci et al. 1996b) although co-ordinated locomotor activity requires more extensive connections (Kjaerulff & Kiehn, 1996). The present study used an organotypic culture of the rat spinal cord in which the possible participation of ventral horn interneurones in rhythmogenesis could be investigated by recording from visually identified cells.
Characteristics of high-K+-evoked bursting
In our experiments small increases in external [K+] triggered rhythmic discharges which lagged behind the inward current onset induced by K+, implying slow activation of a network mechanism. Bursts had variable amplitude, regular periodicity and duration, and were synaptically generated since they were blocked by CNQX, TTX or low-Ca2+ solutions and reversed in polarity at depolarized potentials. The latter property makes it unlikely that their origin is via activation of voltage-dependent intrinsic conductances in the recorded interneurone, as burst period and duration were insensitive to membrane potential changes. Burst polarity inversion with strong depolarization is not readily compatible with their generation via extensive electrotonic coupling as in the chick spinal cord (Chub & O'Donovan, 1998). Furthermore, the delay between patch recorded bursts and rhythmic activity detected extracellularly in the contralateral area was longer than that expected for gap junction-mediated mechanisms. The sensitivity to CNQX indicates that they were mainly mediated by non-NMDA receptors although NMDA receptors also contributed to them at depolarized potentials.
In high-K+ solution under voltage clamp the burst reversal potential was negative. To rule out the possibility that part of these events was a voltage clamp artefact, we also recorded them under current clamp conditions and found them to reverse at a more negative level as the internal Cl- concentration was reduced (4 mM). With such lower internal [Cl-] and at values near the resting potential, individual bursts comprised EPSPs usually followed by IPSPs. With the higher intracellular Cl- concentration typically employed for voltage clamp experiments, the segregation of these events at -56 mV was absent as Cl--mediated events became inward currents mixed with the excitatory currents. Further support for this notion stems from the observation that the burst episode reversal (under voltage clamp) became 0 mV after block of glycine and GABAA receptors. The synaptic currents detected in the present study therefore consisted of multicomponent responses generated primarily by glutamate, glycine and GABA.
Pharmacology of bursting
Attempts to induce bursts with AMPA or NMDA (even at relatively high concentrations or together with 5-HT) were of limited success. In the first case we suspect that AMPA receptor desensitization might have prevented sustained rhythmic patterns, as also shown by the intense bursting evoked by block of AMPA receptor desensitization (Ballerini et al. 1995). NMDA receptors were functionally blocked by extracellular Mg2+ as Mg2+-free solutions elicited rhythmic bursting antagonized by CPP and sharing an analogous reversal potential with that of the high-K+ rhythm.
No rhythmic activity was optically recorded from striated muscle fibres present in a similar preparation bathed with a Mg2+-free solution (Streit, 1996). Nevertheless, in that study lack of continuous superfusion makes the extent of Mg2+ removal uncertain. In the present investigation the Mg2+-free-induced rhythm was depressed in amplitude and rate by CNQX, indicating that it also required non-NMDA receptor operation. Its insensitivity to NMDA and 5-HT, agents routinely used to induce fictive locomotion in the intact preparation, can be explained by ambient Mg2+ block of NMDA receptors and perhaps by inadequate 5-HT receptor expression especially as the 5-HT projection cells of the brainstem were removed during the preparation of this culture. It is worth noting that lowering the extracellular [Mg2+] evoked spontaneous epileptiform activity in organotypic slice cultures of the hippocampus (Scanziani et al. 1994).
The rhythmic activity which developed in spinal cultures following exposure to Mg2+-free solution was qualitatively similar to that generated by high K+ in terms of burst structure, periodicity and reversal potential. However, the Mg2+-free activity was only partially sensitive to CNQX antagonism as it presumably relied more strongly on the NMDA receptor activation (as validated by CPP antagonism) than the comparable situation induced by K+ which mainly depended on AMPA/kainate receptor activation. This phenomenon suggests that, as in the case of the network operating in the isolated spinal cord of the rat (Beato et al. 1997), the organotypic slice preparation could generate apparently analogous rhythms using either class of glutamate receptors. Such rhythms were quite different from those induced by pharmacological block of synaptic inhibition via coapplication of strychnine and bicuculline (Ballerini & Galante, 1998), since the latter had a different burst structure, longer burst duration, much lower periodicity and zero reversal potential. Nevertheless, the bursting activity which developed after block of synaptic inhibition could be accelerated by high K+ as in the case of the rat spinal cord in vitro (E. Bracci, M. Beato, L. Ballerini & A. Nistri, unpublished results).
Comparison of organotypic slice bursting with that observed in the rat spinal cord
An interesting issue is why a rhythmic pattern was consistently observed in organotypic slice cultures (which has a three-layered structure only) but has not been reported in an acute slice preparation of the spinal cord. One possibility is that rhythmic bursting requires neurones from other spinal segments severed in the acute slice but preserved in the organotypic culture. Alternatively, development of more extensive connections between neurones in organotypic slices might have facilitated their ability to generate spontaneous synchronized activity. In organotypic slice cultures of the hypothalamus it has recently been observed that a rhythmic pattern is reliably generated with characteristics analogous to those found in vivo while slice preparations of the same area fail to express it (Jourdain et al. 1998). While remodelling might have taken place in culture so as to induce a rhythm not normally found in vivo, it seems difficult to reconcile this view with the fact that the pattern of bursting is so similar to that seen in the intact preparation of the hypothalamus (Jourdain et al. 1998) or the spinal cord (Bracci et al. 1998). Thus, any comparison of patterned activity of the organotypic slice culture must be made with the isolated spinal cord rather than the acute slice preparation.
Fictive locomotor patterns of the rat spinal cord in vitro typically alternate between flexor and extensor motor pools and between the left and right sides of the spinal cord with a 0·5-1 s period (Kudo & Yamada, 1987; Cazalets et al. 1992; Kjaerulff & Kiehn, 1996; Beato et al. 1997; Bracci et al. 1998). The method for preparing the organotypic spinal slice does not allow separate motor pools for flexor and extensor muscles to be obtained nor, of course, are there multiple ventral roots and motor nerves to record such an activity. Hence, the present model does not possess the functional alternation of motor pattern seen in the whole spinal tissue. As far as the left-right alternation is concerned, the slice culture was obtained from E14 embryos, that is at a prenatal period when left-right alternation is clearly absent in the isolated embryo spinal cord, in which it appears at E19 (Kudo et al. 1991). Note that in the rat embryo spinal cord even spontaneous electrical discharges from motor pools lack alternation up to E17·5 (Nishimaru et al. 1996). Even allowing for some in vitro maturation, the organotypic slice culture cannot thus be expected to show side alternation. This prediction was borne out by recordings made with two extracellular electrodes from both sides of the slice preparation; in this case there was only a modest delay between bursting events which were not in antiphase to each other. It appears that the slice culture is thus intrinsically devoid of the alternating rhythmicity of the fictive locomotion typically studied in the postnatal spinal cord. In this sense the present model is of limited value but it does also offer certain advantages such as the possibility of demonstrating regular rhythmicity in visually identified cells of a three-layer structure. Such a network had evidently developed a complex circuit exhibiting a regular rhythm with periods comparable to those elicited by high K+ in motoneurones of the rat isolated spinal cord (Bracci et al. 1998) and with temporal patterning of EPSPs and IPSPs reminiscent of the comparable responses recorded intracellularly from interneurones (Raastad et al. 1997) and motoneurones (Cazalets et al. 1996) of the rat isolated spinal cord during fictive locomotion. In another brain area typically endowed with rhythmic activity, namely the respiratory neurones of the ventrolateral medulla of the brainstem (Paton & Richter, 1995; Paton, 1997), periodic sequences of excitatory and inhibitory synaptic events have also been observed.
It should also be noted that, even under the more physiological conditions of recording from the isolated spinal cord preparation, rhythmic alternation of excitation and inhibition for each cycle of fictive locomotion is far from being a straightforward observation. In fact, revealing inhibitory synaptic events in rat interneurones requires extensive averaging of raw data with template subtractions of certain events (Raastad et al. 1996, 1997). Even under such conditions, not all interneurones in the intermediate grey matter show alternation of excitation and inhibition, leaving the possibility that parts of the network could actually give rhythmic signals through either IPSPs or EPSPs alone (Raastad et al. 1997). A similar complexity was found recently even when recordings were obtained from single rat motoneurones during fictive locomotion (Cazalets et al. 1996; Hochman & Schmidt, 1998). A systematic study of the CPG in the mudpuppy spinal cord (Wheatley et al. 1994) has indicated that the majority of spinal interneurones involved in the locomotor rhythm actually fire during the transition between excitation and inhibition of motor pools. These observations thus suggest that the tight correlation between EPSPs and IPSPs observed in some interneurones of the organotypic slice preparation is compatible with the operation of a CPG but that also the other pattern of excitation superimposed upon tonic inhibition does not exclude their role in CPG mechanisms.
These data suggest that some basic mechanisms underlying rhythmic patterns in the isolated spinal cord were also present in the organotypic culture and that the rhythm recorded from ventral horn interneurones was the expression of an oversimplified network programme executed (albeit without the expected alternation due to tissue immaturity and/or lack of appropriate motor pools) as long as some minimal wiring connections had been established.
Mechanisms of rhythmic bursting
In the spinal cord the CPG operation (especially its periodicity) is proposed to depend on a subset of oscillatory interneurones ('conditional bursters'; Arshavsky et al. 1997). Conditional bursters are supposed to oscillate via transient relief of Mg2+-dependent block of their NMDA receptors. Searching for conditional bursters in the rat spinal cord has not been easy as only two studies have found a few cells with suitable properties (Hochman et al. 1994; Kiehn et al. 1996). This is not surprising as studies on the hippocampus have suggested that even detecting a so-called ancestral network-driving cell is difficult as they may represent only 0·2 % of the population (Traub & Dingledine, 1990). Amongst ventral horn interneurones we failed to find any intrinsically oscillating cell. While it is possible that such cells were located in different areas, regular rhythmicity was, however, maintained despite NMDA receptor antagonism. It has also been noted that transmitter-induced locomotor activity can develop in the absence of extracellular Mg2+ (Kiehn et al. 1996). Since NMDA receptor antagonism does not prevent rhythmicity in the rat isolated spinal cord (Bracci et al. 1996a; Beato et al. 1997), it is suggested that either conditional bursters did not use NMDA receptor-dependent mechanisms or the presence of conditional bursters was not mandatory for this activity. Similar conclusions were reached for bursting observed in long-term hypothalamic cultures (Müller & Swandulla, 1995).
The cellular mechanism responsible for high-K+-evoked bursting of organotypic slices is unclear. In hippocampal CA3 cells high K+ (up to 8·5 mM) has modest depolarizing activity but significantly lowers the Cl- reversal potential for GABAA receptor-mediated inhibition leading to partial disinhibition and gradual build-up of excitatory currents which trigger bursting synchronized through recurrent collaterals (Chamberlin et al. 1990; Traub & Dingledine, 1990). As large inhibitory potentials were observed during bursting (in current clamp conditions) this explanation is unlikely to apply to the present situation. Furthermore, this proposal cannot explain why in the organotypic slice culture high K+ could still accelerate bursting after block of synaptic inhibition with strychnine and bicuculline and no further disinhibition was expected. Finally, why should Mg2+-free solution have had an effect qualitatively similar to that of high K+ when synaptic inhibition was apparently undisturbed? We suggest that, in the organotypic spinal slice, as in the rat spinal cord (Bracci et al. 1998), rhythmicity crucially depended on an increase in excitability of CPG interneurones. This phenomenon could be elicited by high K+ through depolarization of network cells or by Mg2+ removal through enhancement of glutamatergic transmission. With either condition, the spinal CPG present in the organotypic slice culture expressed a pattern of rhythmic activity as long as its excitability was raised.
|
REFERENCES |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionReferences |
|---|
| Arshavsky, Y. I., Deliagina, T. G. & Orlovsky, G. N. (1997). Pattern generation. Current Opinion in Neurobiology 7, 781-789 | [Medline] |
| Ballerini, L., Bracci, E. & Nistri, A. (1995). Desensitization of AMPA receptors limits the amplitude of EPSPs and the excitability of motoneurons of the rat isolated spinal cord. European Journal of Neuroscience 7, 1229-1234 | [Medline] |
| Ballerini, L. & Galante, M. (1998). Network bursting by organotypic spinal slice cultures in the presence of bicuculline and/or strychnine is developmentally regulated. European Journal of Neuroscience 10, 2871-2879 | [Medline] |
| Beato, M., Bracci, E. & Nistri, A. (1997). Contribution of NMDA and non-NMDA glutamate receptors to locomotor pattern generation in the neonatal rat spinal cord. Proceedings of the Royal Society B 264, 877-884 | [Medline] |
| Bracci, E., Ballerini, L. & Nistri, A. (1996a). Spontaneous rhythmic bursts induced by pharmacological block of inhibition in lumbar motoneurons of the neonatal rat spinal cord. Journal of Neurophysiology 75, 640-647 | [Medline] |
| Bracci, E., Ballerini, L. & Nistri, A. (1996b). Localization of rhythmogenic networks responsible for spontaneous bursts induced by strychnine and bicuculline in the rat isolated spinal cord. Journal of Neuroscience 16, 7063-7076 | [Abstract/Full Text] |
| Bracci, E., Beato, M. & Nistri, A. (1998). Extracellular K+ induces locomotor-like patterns in the rat spinal cord in vitro: comparison with NMDA or 5HT induced activity. Journal of Neurophysiology 79, 2643-2652 | [Abstract/Full Text] |
| Braschler, U. F., Iannone , A., Spenger, C., Streit, J. & Lüscher, H. R. (1989). A modified roller tube technique for organotypic cocultures of embryonic rat spinal cord, sensory ganglia and skeletal muscle. Journal of Neuroscience Methods 29, 121-129 | [Medline] |
| Cazalets, J. R, Borde, M. & Clarac, F. (1996). The synaptic drive from the spinal locomotor network to motoneurons in the newborn rat. Journal of Neuroscience 16, 298-306 | [Abstract] |
| Cazalets, J. R., Sqalli-Houssaini, Y. & Clarac, F. (1992). Activation of the central pattern generators for locomotion by serotonin and excitatory amino acids in neonatal rat. The Journal of Physiology 455, 187-204 | [Abstract] |
| Chamberlin, N. L., Traub, R. D. & Dingledine, R. (1990). Role of EPSPs in initiation of spontaneous synchronized burst firing in rat hippocampal neurons bathed in high potassium. Journal of Neurophysiology 64, 1000-1008 | [Medline] |
| Cheng, J., Stein, R. B., Jovanovic, K., Yoshida, K., Bennett, D. J. & Hun, Y. (1998). Identification, localization and modulation of neural networks for walking in the mudpuppy (Necturus maculatus) spinal cord. Journal of Neuroscience 18, 4295-4304 | [Abstract/Full Text] |
| Chub, N. & O'Donovan, M. J. (1998). Blockade and recovery of spontaneous rhythmic activity after application of neurotransmitter antagonists to spinal networks of the chick embryo. Journal of Neuroscience 18, 294-306 | [Abstract/Full Text] |
| Dale, N. & Kuenzi, F. M. (1997a). Ion channels and the control of swimming in the Xenopus embryo. Progress in Neurobiology 53, 729-756 | [Medline] |
| Dale, N. & Kuenzi, F. M. (1997b). Ionic currents, transmitters and models of motor pattern generators. Current Opinion in Neurobiology 7, 790-796 | [Medline] |
| Fulton, B. P. & Walton, K. (1986). Electrophysiological properties of neonatal rat motoneurons studied in vitro. The Journal of Physiology 370, 651-678 | [Abstract] |
| Grillner, S., Wallen, P. & Brodin, L. (1991). Neuronal network generating locomotor behaviour in lamprey. Annual Review of Neuroscience 14, 169-199 | [Medline] |
| Ho, S. & O'Donovan, M. J. (1993). Regionalization and intersegmental coordination of rhythm-generating networks in the spinal cord of the chick embryo. Journal of Neuroscience 13, 1354-1371 | [Abstract] |
| Hochman, S., Jordan, L. M. & MacDonald, J. F. (1994). N-Methyl-D-aspartate receptor-mediated voltage oscillations in neurons surrounding the central canal in slices of rat spinal cord. Journal of Neurophysiology 72, 565-577 | [Medline] |
| Hochman, S. & Schmidt, B. J. (1998). Whole cell recordings of lumbar motoneurones during locomotor-like activity in the in vitro neonatal rat spinal cord. Journal of Neurophysiology 79, 743-752. | [Abstract/Full Text] |
| Jefferys, J. G. R., Traub, R. D. & Whittington, M. A. (1996). Neuronal networks for induced '40 Hz' rhythms. Trends in Neurosciences 19, 202-208 | [Medline] |
| Jourdain, P., Israel, J. M., Dupouy, B., Oliet, H. R., Allard, M., Vitiello, S., Theodosis, D. T. & Poulain, D. A. (1998). Evidence for a hypothalamic oxytocin-sensitive pattern-generating network governing oxytocin neurons in vitro. Journal of Neuroscience 18, 6641-6649 | [Abstract/Full Text] |
| Kiehn, O., Johnson, B. R. & Raastad, M. (1996). Plateau properties in mammalian spinal interneurons during transmitter-induced locomotor activity. Neuroscience 75, 263-273 | [Medline] |
| Kjaerulff, O. & Kiehn, O. (1996). Distribution of networks generating and coordinating locomotor activity in the neonatal rat spinal cord in vitro: a lesion study. Journal of Neuroscience 16, 5777-5794 | [Abstract/Full Text] |
| Kudo, N., Ozaki, S. & Yamada, T. (1991). Ontogeny of rhythmic activity in the spinal cord of the rat. In Neurobiological Basis of Human Locomotion, ed. Shimamura, M., Grillner, S. & Edgerton, V. R., pp. 127-136. Japan Scientific Societies Press, Tokyo. | |
| Kudo, N. & Yamada, T. (1987). N-Methyl-d,l-aspartate-induced locomotor activity in a spinal cord-hindlimb muscles preparation of the newborn rat studied in vitro. Neuroscience Letters 75, 43-48 | [Medline] |
| Marty, A. & Neher, E. (1983). Tight-seal whole-cell recording. In Single-Channel Recording, ed. Sakmann, B. & Neher, E., pp. 107-122. Plenum Press, New York and London. | |
| Mayer, M. L. & Westbrook, G. L. (1987). The physiology of excitatory amino acids in vertebrate central nervous system. Progress in Neurobiology 28, 197-276 | [Medline] |
| Müller, W. & Swandulla, D. (1995). Synaptic feedback excitation has hypothalamic neural networks generate quasirhythmic burst activity. Journal of Neurophysiology 73, 855-861 | [Medline] |
| Nishimaru, H., Iizuka, M., Ozaki, S. & Kudo, N. (1996). Spontaneous motoneuronal activity mediated by glycine and GABA in the spinal cord of rat fetuses in vitro. The Journal of Physiology 497, 131-143 | [Abstract] |
| Paton, J. F. R. (1997). Rhythmic bursting of pre- and post-inspiratory neurones during central apnoea in mature mice. The Journal of Physiology 502, 623-639 | [Abstract] |
| Paton, J. F. R. & Richter, D. W. (1995). Role of fast inhibitory synaptic mechanisms in respiratory rhythm generation in the maturing mouse. The Journal of Physiology 484, 505-521 | [Abstract] |
| Phelps, P. E, Barber, R. P. & Vaughn, J. E. (1991). Embryonic development of choline acetyltrasnferase in thoracic spinal motor neurons: somatic and autonomic neurons may be derived from a common cellular group. Journal of Comparative Neurology 307, 77-86 | [Medline] |
| Raastad, M., Johnson, B. R. & Kiehn, O. (1996). The number of postsynaptic currents necessary to produce locomotor-related cyclic information in neurons in the neonatal rat spinal cord. Neuron 17, 729-738 | [Medline] |
| Raastad, M., Johnson, B. R. & Kiehn, O. (1997). Analysis of EPSCs and IPSCs carrying rhythmic, locomotor-related information in the isolated spinal cord of the neonatal rat. Journal of Neurophysiology 78, 1851-1859 | [Abstract/Full Text] |
| Rossignol, S. & Dubuc, R. (1994). Spinal pattern generation. Current Opinion in Neurobiology 4, 894-902 | [Medline] |
| Scanziani, M., Debanne, D., Müller, M., Gähwiler, B. H. & Thompson, S. M. (1994). Role of excitatory amino acids and GABAB receptors in the generation of epileptiform activity in disinhibited hippocampal slice cultures. Neuroscience 61, 823-832 | [Medline] |
| Spenger, C., Braschler, U. F., Streit, J. & Lüscher, H. R. (1991). An organotypic spinal cord-dorsal root-ganglion skeletal muscle coculture of embryonic rat. I. The morphological correlates of the spinal reflex arc. European Journal of Neuroscience 3, 1037-1053. | |
| Streit, J. (1996). Mechanisms of pattern generation in co-cultures of embryonic spinal cord and skeletal muscle. International Journal of Developmental Neuroscience 14, 137-148 | [Medline] |
| Streit, J., Spenger, C. & Lüscher, H. R. (1991). An organotypic spinal cord-dorsal root-ganglion skeletal muscle coculture of embryonic rat. II. Functional evidence for the formation of spinal reflex arcs in vitro. European Journal of Neuroscience 3, 1054-1068. | |
| Traub, R. D. & Dingledine, R. (1990). Model of synchronized epileptiform bursts induced by high potassium in CA3 region of rat hippocampal slice. Role of spontaneous EPSPs in initiation. Journal of Neurophysiology 64, 1009-1018 | [Medline] |
| Wheatley, M., Jovanovic, K., Stein, R. B. & Lawson, V. (1994). The activity of interneurons during locomotion in the in vitro Necturus spinal cord. Journal of Neurophysiology 71, 2025-2032 | [Medline] |
| Ziskind-Conhaim, L. (1988). Electrical properties of motoneurons in the spinal cord of rat embryos. Developmental Biology 128, 21-29 | [Medline] |
This work was supported by grants from the Telethon Foundation (project no. 823), Istituto Nazionale di Fisica della Materia and MURST (co-finanziamento ricerca). We thank Drs Enrico Tongiorgi for motoneurone ChAT staining and Silvia Di Angelantonio for confocal imaging.
Corresponding author
L. Ballerini: SISSA, Via Beirut 4, 34014 Trieste, Italy.
Email: ballerin{at}sissa.it
This article has been cited by other articles:
|
|
 |
|
  F. Furlan, G. Taccola, M. Grandolfo, L. Guasti, A. Arcangeli, A. Nistri, and L. Ballerini ERG Conductance Expression Modulates the Excitability of Ventral Horn GABAergic Interneurons That Control Rhythmic Oscillations in the Developing Mouse Spinal Cord J. Neurosci., January 24, 2007; 27(4): 919 - 928. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 P. A. Guertin and J. Hounsgaard Conditional Intrinsic Voltage Oscillations in Mature Vertebrate Neurons Undergo Specific Changes in Culture J Neurophysiol, March 1, 2006; 95(3): 2024 - 2027. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
P. Darbon, A. Tscherter, C. Yvon, and J. Streit Role of the Electrogenic Na/K Pump in Disinhibition-Induced Bursting in Cultured Spinal Networks J Neurophysiol, November 1, 2003; 90(5): 3119 - 3129. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
) or 5 min (
) of the K+-induced patterned activity. Note the significant (** P < 0·001) shortening in cycle period and burst duration with 8 mM K+. Data were pooled from a population of ventral interneurones.
) preceded hyperpolarizing components (
). At -66 mV, only large depolarizing events were manifested while at -12 mV events consisted mainly of large hyperpolarizing potentials. B, EPSP cycle period versus IPSP cycle period during rhythmic bursting evoked by K+ at a resting membrane potential of -42 mV. Data are from the cell shown in A. C, representative current clamp trace of spike activity recorded from a ventral interneurone (at -56 mV) following intracellular injection of a 500 ms depolarizing current pulse. In this case, QX-314 was omitted from the intracellular solution. Note that spike activity was maintained throughout the train without apparent accommodation.
) and after (
) and after (