1The effect of a brief train of electric stimuli in the dorsolateral funiculus on the intrinsic response properties of turtle motoneurones was investigated in transverse sections of the spinal cord in vitro.
2Even when glutamatergic, GABAergic and glycinergic ionotropic synaptic transmission was blocked by antagonists of AMPA, NMDA, glycine and GABA receptors, dorsolateral funiculus (DLF) stimulation induced a facilitation of plateau potentials during current clamp and the underlying inward current in voltage clamp. This facilitation lasted more than 10 s.
3The plateau potential and the facilitation by DLF stimulation was absent in the presence of 10 μm nifedipine. The DLF-induced facilitation was reduced by antagonists of 5-HT1a, group 1 metabotropic glutamate receptors and muscarine receptors.
4These findings suggest that the intrinsic properties of spinal motoneurones are dynamically regulated by afferent synaptic activity. These afferents can be of spinal and extraspinal origin. Continuous regulation of intrinsic response properties could be a mechanism for motor flexibility.
The diversity of movements seen in most vertebrates requires a rich repertoire of response patterns from individual motoneurones. Sensory afferents, segmental interneurones and descending pathways are obvious and well-established contributors to this repertoire. Here we examine if synaptic regulation of the intrinsic properties of motoneurones could be an additional source of functional flexibility.
In spinal motoneurones, firing patterns are kept in a range that fits the biomechanical properties of muscle fibres by a distinct set of calcium and potassium channels. In addition, intrinsic bistability is a latent property mediated by L-type calcium channels (Hounsgaard, Hultborn, Jespersen & Kiehn, 1988; Hounsgaard & Mintz, 1988). This property is known to be uncovered by activation of a number of agonists for metabotropic receptors (G. Svirskis & J. Hounsgaard, in preparation) and has been implicated in some forms of motor activity (Eken & Kiehn, 1989). It is not known, however, if and how afferent synaptic activity regulates plateau properties in motoneurones. Using transverse slices of the turtle spinal cord, we now show that a brief train of electric stimuli in the dorsolateral funiculus (DLF), a major pathway for locomotor control (Grilmer, 1976; Lennard & Stein, 1977), increases motoneuronal excitability for more than 10 s, even in the absence of ionotropic transmission.
Our analysis shows that the increased excitability is caused by facilitation of a postsynaptic dihydropyridine-sensitive inward current. This facilitation is due to activation of metabotropic receptors for glutamate, acetylcholine and 5-HT.
Our study demonstrates that the response properties of spinal motoneurones are synaptically regulated on a time scale comparable to the duration of movements. By this mechanism the pattern and weight with which individual motoneurones contribute to movements may be changed dynamically.
Transverse slices (2–3 mm thick) were obtained from the lumbar enlargement of adult turtles (Pseudemys scripta) anaesthetized by intraperitoneal injection of 100 mg sodium pentobarbitone. Experiments were performed at room temperature (20–22 °C) in a solution containing (mM): 120 NaCl, 5 KC1, 15 NaHCO3, 2 MgCl2, 3 CaCl2 and 20 glucose, saturated with 98% O2 and 2% CO2 to obtain a pH of 7.5.
Intracellular recordings in current- and voltage-clamp mode were performed with an AxocIamp-2A amplifier. Pipettes for voltage-clamp (40 MΩ) and current-clamp (50–60 MΩ) were filled with a solution containing 0.5 m potassium acetate + 1.5 m KC1 and 1 m potassium acetate, respectively. In some experiments action potentials were eliminated by QX-314 (2(triethylamino)-N-(2,6-dimethylphenyl)acetamide; Research Biochemicals, Inc. (RBI)) added to the pipette solutions (0.2m final concentration). Moto-neurones were selected for study if they had a stable membrane potential of more than –60 mV. Extracellular stimulation in the DLF was performed by applying thirty constant current pulses (0.2 ms duration) at 20 Hz by means of a bipolar wire electrode. Depolarizing current pulses lasting from 1 to 3 s were applied before and after DLF stimulation in current-clamp recordings. Voltage-clamp recordings were performed in discontinuous service mode at a sample rate of 5–12 kHz, anti-alias filter of 5–10 μS, gain of 0.7–1.5 nA mV−1 and low-pass filter of 0.1 kHz. The use of ramp-clamp has been detailed elsewhere (Svirskis & Hounsgaard, 1997). Briefly, a triangular voltage waveform command (4 s duration) was used to depolarize the motoneurones from the resting potential before and after DLF stimulation. The data were sampled at 16.6 kHz (current-clamp recordings) and 6.6 kHz (voltage-clamp recordings) with a 12-bit analog-to-digital converter (DIGIDATA 2000 from Axon Instruments) and displayed by means of Axoscope software and stored on a hard disk for later analysis. The motoneurones in which DLF stimulation increased the excitability (56/82) were encountered more than 100 μm from the surface of the slice.
Synaptic potentials were eliminated by adding the following drugs to normal medium: 6-cyano-7-nitroquinoxaline-2,3-dione (CNQX, 20 μM); (±)-2-amino-5-phosphonopentanoic acid (AP5, 50 μm) (both from RBI); strychnine (10 μm); bicuculhne (10–50 μM); and kynurenic acid (0.50 mm) (all from Sigma). In different experiments one of the following antagonists of metabotropic receptors was added to the medium: pindobind-5-HT1a (10 μM) (RBI) and 1-(2-methoxyphenyl)-4-(4-phthahmidobutyl) piperazine HBr (NAN-190, a 5-HT1A receptor antagonist, 12.5-25 μm); (RS)-1-aminoindan-1,5-dicarboxyhc acid (ADA, a group I metabotropic glutamate receptor antagonist, 20 μm); (S)-α-methyl-4-carboxyphenylglycine (MCPG, a non-selective metabotropic glutamate receptor antagonist, 0.5 mm); atropine (a cholinergic receptor antagonist, 0.1 μm) (all from Tocris); and tetrodotoxin (TTX, 1 μm) (Sigma).
During intracellular recording from motoneurones in transverse slices of the turtle spinal cord, the response to depolarizing current pulses, applied through the recording electrode, served to monitor excitability. We found that this response was facilitated when preceded by a brief train of stimuli applied in the ipsilateral dorsolateral funiculus (DLF; Fig. 1A). The effect of the train was not mediated by ionotropic synaptic receptors since it persisted when synaptic potentials were eliminated by antagonists of AMPA/kainate, NMDA, glycine and GABAa receptors (Fig. 1B). When action potentials in the cell recorded from, were suppressed by applying QX-314 intracellularly from the recording electrode, DLF stimulation was found to promote a plateau potential during the depolarizing test pulse (Fig. 1C; n= 9). In I–V plots obtained from the current generated during triangular voltage commands (Fig. 1D), the plateau current, induced by a train of DLF stimuli, gave rise to a clockwise hysteretic configuration (Fig. 1E; n= 53). DLF stimulation did not change the current at the holding potential or during the rising phase of the command ramp (Fig. 1E), suggesting that the plateau current was the result of an enhanced inward current rather than elimination of a leak conductance.
We examined the origin and mechanism of the DLF-promoted plateau potential. Turtle motoneurones project dendrites to the DLF region (Ruigrok, Crowe & Donkelaar, 1984). The possibility that the plateau current was induced by direct activation of dendrites was eliminated by the finding that all effects of DLF stimulation were blocked in medium containing 1μM TTX (Fig. 2A; n= S). DLF stimulation never increased the excitability in superficial motoneurones (0–100 μm depth) in which even high stimulus strength (50 × threshold) only induced weak synaptic responses in normal medium. Motoneurones in which DLF stimulation evoked robust synaptic responses in normal medium and increased excitability were located more than 100 μm from the surface.
In the spinal cord of the turtle, plateau potentials in motoneurones and interneurones are sensitive to dihydro-pyridines (Hounsgaard & Mintz, 1988; Hounsgaard & Kjærulff, 1994; Russo & Hounsgaard, 1994). The plateau potential facilitated by DLF stimulation was reduced or eliminated in medium containing 10 μm nifedipine (Fig. 2B; n= 6). As previously shown in the dorsal horn (Russo & Hounsgaard, 1994) this is not due to a general suppression of transmitter release from presynaptic terminals. In normal medium, nifedipine eliminated the facilitatory effect of DLF stimulation without affecting the ionotropic PPSP in response to single stimuli (Fig. 2C) or during a stimulus train (n= 6). These findings suggested that activation of presynaptic axons in the DLF resulted in facilitation of a voltage-sensitive inward current, mediated by L-type calcium channels in motoneurones.
We next tested the possible links between the activated presynaptic axons and facilitation of the plateau current in motoneurones. Plateau potentials in motoneurones are facilitated by agonists of certain receptors for 5-HT, glutamate and muscarine when applied to the bathing solution (Hounsgaard & Kiehn, 1989; G. Svirskis & J. Hounsgaard, in preparation). In the present experiments we found that the DLF-induced facilitation was eliminated or reduced by antagonists of these receptors. As illustrated in Fig. 3 the facilitation of the plateau current (Fig. 3A) was reduced in medium containing 12.5–25 μm NAN-190 (n= 7; Fig. 3B) or 10 μm pindobind-5-HT1A (n= S), selective antagonists of 5-HT1A receptors (Glennon, Naiman, Pierson, Lyon & Weisberg, 1988; Liau, Sleight, Pitha & Peroutka, 1991). Similar reduction of DLF-induced facilitation was found in medium containing 0.5 him MCPG (n= 6), a selective antagonist of metabotropic glutamate receptors (Pin & Duvoisin, 1995) or 25–50 μM ADA (n= 7), a specific antagonist of phospholipase C-linked mGluEs (Pellicciari et al. 1995) or 0.1 μM atropine (n= 6), a selective antagonist of muscarinic acetylcholine receptors. These results are illustrated in Fig. 4.
Voltage-sensitive ion channels in cell bodies and dendrites provide neurons with non-linear, intrinsic response properties (Llinas, 1988). If these properties favour functionally useful axonal impulse patterns (Midtgaard & Hounsgaard, 1989), they must be adjusted to perform optimally with varying behavioural needs. The present findings show that the intrinsic response properties of spinal motoneurones, provided by L-type calcium channels, are regulated by multiple metabotropic synaptic receptors.
The elimination of modulation by antagonists of glutamatergic, cholinergic and serotonergic receptors suggests that the DLF stimulus activates a heterogeneous population of afferent fibres. The origin of these pathways may be both intraspinal, as in the case of glutamatergic and cholinergic connections and extraspinal, as in the case of serotonergic and glutamatergic connections. This is supported by the fact that the DLF in the turtle includes excitatory axons terminating on motoneurones, and that monosynaptic excitatory synaptic potentials, evoked by the DLF stimulus, are eliminated by CNQX and AP5, antagonists of ionotropic glutamate receptors. The DLF also contains cholinergic axons, probably originating from neurones in the dorsal horn (J. Damsgaard and J. Hounsgaard, unpublished histochemical observations). Finally, the DLF in the turtle is a major pathway for descending serotonergic fibres (Kiehn, Eostrup & Moeller, 1992).
Bath-applied agonists for metabotropic glutamate receptors and muscarine receptors not only produce an increase in voltage-sensitive inward current in motoneurones, but also induce depolarization and increase membrane resistance (G. Svirskis & J. Hounsgaard, in preparation). DLF stimulation, however, only facilitated activation of the inward current. It is possible that the synaptic input from this pathway only affects a small area of dendritic membrane. If so, the change in the membrane conductance may be too small to detect.
The spinal mechanisms that provide for flexible generation of movements are largely unknown. In general, ionotropic synaptic connections allow neurones in central pattern generators to interact on a time scale not extending beyond a few hundred milliseconds. Motor patterns, however, often involve co-ordinated activity on a time scale of seconds. Non-linear response properties, which bridge a time scale from milliseconds to tens of seconds in postsynaptic neurones, may fill the gap between short-lived synaptic potentials and the long-term changes involving long-term potentiation and gene regulation. Metabotropic regulation of intrinsic response properties may facilitate flexible generation of motor patterns at two levels. The response properties of individual neurones can be tuned continuously to bias for functionally useful outcomes of synaptic integration in terms of response duration and impulse pattern. At the network level, neurones may be recruited to or excluded from a functional network by metabotropic regulation of their excitability. Recent complementary findings in dorsal horn interneurones (Russo, Nagy & Hounsgaard, 1997) and elsewhere (Katz & Frost, 1996) suggest that metabotropic regulation of the kind illustrated here may be widespread in the nervous system.
This work was kindly funded by the European Economic Community, The Danish MBC, The Lundbeck Foundation, The NOVO-Nordisk Foundation, CINVESTAB and CONACYT.