Synaptic Release of Serotonin Induced by Stimulation of the Raphe Nucleus Promotes Plateau Potentials in Spinal Motoneurons of the Adult Turtle

Stimulation of the raphe nucleus promotes plateau potentials in motoneurons

We tested the excitability of spinal motoneurons recorded in the brainstem spinal cord preparation by intracellular injection of depolarizing current pulses. In the example shown in Figure 1B, the motoneuron responded with a train of action potentials with a frequency that slightly increased during the pulse. Electrical stimulation of the raphe nucleus (20 shocks, 10 Hz) applied 3 s before the depolarizing current pulse facilitated a plateau potential (Fig. 1C). The frequency of action potentials increased more than in control conditions. In addition, an afterdepolarization that was sufficient to trigger action potentials was present after the current pulse was turned off (Fig. 1C, arrow).

It was originally suggested that the promotion of plateau potentials by 5-HT is attributable to the inhibition of the afterhyperpolarization (AHP) following action potentials (Hounsgaard and Mintz, 1988; Hounsgaard and Kiehn, 1989). We therefore tested whether stimulation of the raphe nucleus inhibited the AHP. Figure 1D shows the recording of an action potential evoked by a strong depolarizing current pulse (4 nA, 4 ms) in a spinal motoneuron. Under control conditions, the spike was followed by a fast and a slower AHP that had an amplitude of –1.5 mV (Fig. 1D). When the raphe nucleus was stimulated by a single electrical shock applied 300 ms before the action potential, the amplitude of the slow component of the AHP was reduced to –0.9 mV (i.e., a decrease of 40%) (Fig. 1E). Inhibition of the slower component of the AHP by raphe stimulation was recorded in six motoneurons (mean inhibition, 21.5 ± 11.3%; p < 0.01).

To test whether the facilitation of plateau potentials occurred independently of the inhibition of the AHP, we adjusted the amplitude of depolarizing current pulses to the threshold for action potentials. In the example illustrated in Figure 2A, a control depolarizing current pulse generated no action potentials. However, after stimulation of the raphe nucleus (10 shocks at 10 Hz), the same depolarizing current pulses applied 0.5, 4.5, and 8.5 s after the stimulation generated seven, eight, and six action potentials, respectively. A similar increase in excitability induced by raphe nucleus stimulation was recorded in 11 motoneurons (Fig. 2D, left). The increase in excitability induced by stimulation of the raphe nucleus peaked during the first current pulse and gradually decreased after the stimulation. It remained statistically significant after 8.5 s and was back to normal after 13.5 s (±3.6; n = 10; data not shown). Because control pulses were applied at the threshold level for action potentials, it is likely that the facilitation was not only caused by inhibition of the AHP.

• Download figure
• Open in new tab
• Download powerpoint

Figure 1.

Stimulation of the raphe nucleus promotes plateau potentials in spinal motoneurons. A, Scheme of the brainstem spinal cord preparation. A catheter inserted between the backbone and the brainstem was continuously perfusing the preparation with Ringer's solution (arrows). A bipolar electrode was positioned at the ventral third of the midline of the brainstem section (i.e., at the level of the raphe nucleus) (see Materials and Methods). Intracellular recording of motoneurons was performed with a microelectrode positioned in the lateral part of the ventral horn of the spinal cord section. B, Control response to a depolarizing current pulse. C, After stimulation of the raphe nucleus, the same current protocol induced a train of action potentials occurring at a higher frequency and followed by an afterdepolarization with a superimposed afterdischarge (arrow), indicative of an underlying plateau potential. For B and C: bottom trace, current injected intracellularly; middle trace, membrane potential; top plot, instantaneous spike frequency. D, Average response of a motoneuron to depolarizing current pulses in control conditions (average of 20 consecutive traces; SD in gray). E, Average response of the same motoneuron when the depolarizing current pulse was preceded by stimulation of the raphe nucleus 300 ms before the depolarizing current pulse. Note the inhibition of the AHP. The action potential was cut to fit in the frame of the figure.

Facilitation of plateau potentials involves 5-HT₂ receptors

In the preceding paragraph, we showed that synaptic release of transmitter, induced by the activity of the raphe nucleus, generates an increase in excitability that lasts several seconds. However, besides 5-HT, raphe neurons also release other neuromodulators such as enkephalin or substance P (Uhl et al., 1979; Inagaki et al., 1981; Schmidt and Jordan, 2000). In addition, although the stimulus electrode was positioned in the region of the raphe nucleus, we cannot exclude that some of the recruited fibers were not serotonergic. To ensure that the increase in excitability recorded in motoneurons was actually attributable to synaptic release of 5-HT, we performed pharmacological experiments by applying an antagonist at the cut end of the spinal cord through a local superfusion system (Fig. 2B). Because 5-HT promotes plateau potentials in turtle motoneurons by activating 5-HT₂ receptors (Perrier and Hounsgaard, 2003), we tested the effect of SB-206553, a selective antagonist for these receptors. Figure 2C shows a recording obtained from the same motoneuron as in Figure 2A, but during the superfusion of SB-206553. Depolarizing current pulses of the same amplitude as in normal medium generated zero or one action potential (Fig. 2C, first two pulses), showing that the threshold for action potentials remained unchanged. However, stimulation of the raphe nucleus no longer potentiated the response induced by depolarizing current pulses. We observed the same effect in four of four motoneurons from four different preparations (Fig. 2D, right) (paired t test; p > 0.05). This experiment confirms that synaptic release of 5-HT induced by electrical stimulation of the raphe nucleus facilitates plateau potentials in motoneurons. Moreover, it demonstrates that 5-HT₂ receptors are involved in the facilitation.

• Download figure
• Open in new tab
• Download powerpoint

Figure 2.

The increase in excitability induced by raphe stimulation involves 5-HT₂ receptors. A, Response of a motoneuron before (Ctrl) and after stimulation of the raphe nucleus (1, 2, 3). The excitability increased for >8 s. B, Scheme of the preparation. The superfusion system applying SB-206553 at the cut level of the spinal cord (black dotted arrows) is shown. C, The same motoneuron as in A, recorded during superfusion of SB-206553. No additional potentiation of the response is observed. D, Summary plot of the number of spikes induced by each depolarizing current pulse in normal medium (n = 11) and during the superfusion of SB-206553 for the different motoneurons recorded (n = 4). Statistical difference before and after raphe stimulation in normal medium (*p < 0.05; **p < 0.01) but not in the presence of SB-206553 is shown. Ctrl, Control.

Stimulation of the DLF reproduces the effect induced by stimulation of the raphe nucleus

The results obtained with the integrated preparation can be interpreted in two different ways: SB-206553 inhibited the facilitation either by blocking 5-HT₂ receptors located on motoneurons or by inhibiting receptors located at the premotor level. To circumvent this limitation of the preparation, we pursued our investigation in a slice preparation from the spinal cord. The raphe spinal pathway was stimulated by means of a microelectrode located on the DLF (Fig. 3A), a major pathway for descending serotonergic fibers in reptiles (Ten Donkelaar, 1976) including turtles (Kiehn et al., 1992). Fast synaptic potentials were blocked by a mixture of antagonists for ionotropic receptors (see Materials and Methods). Motoneuron excitability was tested as in the integrated preparation, by intracellular injection of depolarizing current pulses with an amplitude adjusted to the level for spiking threshold. Figure 3B shows the recording from a motoneuron obtained with this set-up. A depolarizing current pulse generated no action potential. After stimulation of the DLF (30 shocks at 25 Hz), the same depolarizing current pulse applied 0.5, 4.5, and 8.5 s after the stimulation generated seven, one, and zero action potentials, respectively. Similar increase in excitability induced by DLF stimulation was recorded in 11 of 11 motoneurons tested (Fig. 3D, left). The increase in excitability gradually decreased after stimulation of the DLF but remained statistically significant up to 8.5 s after.

• Download figure
• Open in new tab
• Download powerpoint

Figure 3.

Stimulation of the DLF on a slice preparation increases motoneuron excitability. A, Scheme of the preparation for stimulation and recording. B, Response of a motoneuron before (Ctrl) and after stimulation of the DLF (1, 2, 3). The excitability increased for >8s. C, The same motoneuron recorded after the addition of SB-206553. No additional potentiation of the response was observed. D, Summary plot of the number of spikes induced by each depolarizing current pulse in normal medium (n = 11) and after the addition of SB-206553. Statistical differences before and after DLF stimulation in normal medium (*p < 0.05; ***p < 0.001) but not in the presence of SB-206553 (n = 5) is shown. For all recordings, fast synaptic potentials were blocked by CNQX (25 μm), AP-5 (50 μm), strychnine (10 μm), and bicuculline (20 μm). Ctrl, Control.

To test whether the stimulus-induced increase in excitability in motoneurons was attributable to activation of 5-HT₂ receptors, we added SB-206553 to the extracellular medium. Figure 3C shows a recording obtained from the same motoneuron as in Figure 3B, but in the presence of the 5-HT₂ receptor antagonist. Electrical stimulation of the DLF no longer increased the excitability of the motoneuron. Consistent results were obtained in five of five motoneurons tested (Fig. 3D, right). In the presence of SB-206553, DLF stimulation no longer increased the number of spikes generated by a depolarizing current pulse (p > 0.05; paired t test). These results show that synaptic release of 5-HT increases the excitability of motoneurons by activating 5-HT₂ receptors.

Synaptic activation of 5-HT₂ receptors facilitates a low-threshold, voltage-sensitive inward current

A previous study ascribed the increase in excitability induced by DLF stimulation to the facilitation of a plateau potential mediated by L-type Ca²⁺ channels because it was abolished by nifedipine (Delgado-Lezama et al., 1997). To test whether the increase in excitability induced by synaptic activation of 5-HT₂ receptors also involved L-type Ca²⁺ channels, we performed voltage-clamp experiments. The protocol consisted of 20 mV triangular voltage depolarizations applied from –60 mV (Fig. 4A1). Stimulation of the DLF facilitated a voltage-sensitive, slowly activating inward current (Fig. 4A1, gray trace) responsible for the clockwise hysteresis configuration in I–V plots (n = 7 of 7) (Fig. 4A2). Motoneuron input resistance measured as the slope of I–V plots during the first second of voltage ramps was not significantly affected by DLF stimulation (Fig. 4D) (paired t test; p > 0.05; n = 7). In the presence of SB-206553, the response recorded before stimulation of the DLF was unaffected (Fig. 4A1,B1, compare black traces). However, the facilitation of the inward current induced by DLF stimulation and consequently the hysteresis in the I–V plot were not as substantial as in normal medium (n = 5of5) (Fig. 4B1,B2, gray traces). The current facilitated by the DLF and inhibited by SB-206553 was calculated by subtracting the current recorded after DLF stimulation in the presence of SB-206553 to the current recorded after DLF stimulation in normal medium (difference between gray traces in Fig. 4A1,B1). As illustrated in Figure 4C, SB-206553 only inhibited the inward current responsible for the hysteresis in I–V plots. To quantify the effect of SB-206553, we compared the current facilitated by DLF stimulation (calculated as the difference of integrated current before and after DLF stimulation) before and after the addition of the 5-HT₂ antagonist. SB-206553 significantly inhibited the current facilitated by the DLF (paired t test; p < 0.05; n = 5). Because the clockwise hysteresis corresponds to the activity of low-threshold L-type Ca²⁺ channels (Svirskis and Hounsgaard, 1997, 1998; Delgado-Lezama et al., 1997; Carlin et al., 2000) and because DLF stimulation did not affect the input resistance of motoneurons, we concluded that synaptic release of 5-HT activated 5-HT₂ receptors, which were specifically facilitating low-threshold L-type Ca²⁺ channels (Ca_V1.3).

• Download figure
• Open in new tab
• Download powerpoint

Figure 4.

DLF-induced plateau currents are reduced by 5-HT₂ receptor antagonists. A1, Current recorded during triangular voltage command (bottom trace) before (black trace) and after (gray trace) stimulation of the DLF. DLF stimulation facilitated a slowly activating voltage-sensitive inward current. A2, I–V plot of the recording. Note the facilitation of the clockwise hysteresis (arrows). B1, The same motoneuron recorded in the presence of SB-206553. The facilitation induced by the DLF was reduced. B2, I–V plot. Note the lesser facilitation of the hysteresis. C, Current facilitated by the DLF and inhibited by SB-206553. The current blocked by SB-206553 corresponds to the one responsible for the hysteresis. D, Mean input resistance of motoneurons before (black) and after (gray) DLF stimulation was not statistically different. For all recordings, fast synaptic potentials were blocked by CNQX (25 μm), AP-5 (50 μm), strychnine (10 μm), and bicuculline (20 μm).

Specificity of the effects induced by stimulation of the raphe nucleus

A previous study performed in the cat showed that motoneuronal excitability is increased in response to raphe stimulation (Fung and Barnes, 1989). In agreement, we found that electrical stimulation in the ventral midline of the cut surface of the brainstem increased the excitability of motoneurons by facilitating plateau potentials. Although the stimulation electrode was positioned at the level of the nucleus raphe inferior (Cruce and Nieuwenhuys, 1974), one can argue that this does not ensure facilitation to be mediated by synaptically released 5-HT. First, serotonergic neurons also contain other transmitters such as enkephalins, substance P, or GABA (Schmidt and Jordan, 2000). Second, it is likely that the bipolar electrode recruited non-serotonergic neurons in the raphe nucleus (Schmidt and Jordan, 2000) or in its immediate surrounding. It was therefore necessary to perform pharmacological experiments to show that the facilitation induced by stimulation of the raphe nucleus was actually induced by 5-HT. The fact that SB-206553 completely abolished the facilitation demonstrates that the main effect induced by the stimulation was caused by synaptic release of 5-HT acting on 5-HT₂ receptors. Finally, the experiments performed with SB-206553 in a slice preparation in which ionotropic receptors were blocked strongly suggest that the 5-HT₂ receptors responsible for the facilitation are located on motoneurons. An alternative possibility would be that 5-HT, acting on presynaptic terminals, induces the release of transmitters such as glutamate or acetylcholine, which are known to promote plateau potentials in motoneurons (Delgado-Lezama et al., 1997; Svirskis et al., 1998).

Timing of the effects induced by stimulation of the raphe nucleus

The effects on the AHP and on the excitability of motoneurons were tested and detected 300–500 ms after stimulation of the raphe nucleus. To be valid, a test on the amplitude of the AHP or on excitability of motoneurons has to be performed from the same membrane potential as in the control condition. However, in most cases, stimulation of the raphe nucleus also induced excitatory synaptic potentials in motoneurons (data not shown), as reported by Fung and Barnes (1989). We ascribed these EPSPs to the recruitment of non-serotonergic fibers. The delay of intracellular test pulses in motoneurons was adjusted in order for the membrane potential not to be contaminated by any EPSP. Therefore, a value of 300 ms is probably an overestimation for the latency of the effects induced by the raphe nucleus.

The increase in excitability peaked 0.5 s after stimulation of the raphe nucleus and was still detectable after 8.5 s. The excitability was back to its control level after 13.5 s. Because it is unlikely that 5-HT remains in the extracellular space for >1.5 s (Hille, 2001), the increase in excitability can either be ascribed to a postsynaptic mechanism or to a prolonged firing of raphe neurons. This latter hypothesis can be discarded, because in slice preparations, stimulation of the DLF induced an increase in excitability of motoneurons with a similar time course.

Multiple effects of 5-HT

In motoneurons, 5-HT induces multiple effects including depolarization (Larkman and Kelly, 1992; Hsiao et al., 1997; Perrier et al., 2003), inhibition of the medium AHP following action potentials (Berger et al., 1992; Hsiao et al., 1997; Grunnet et al., 2004), and facilitation of plateau potentials (Hounsgaard et al., 1988a; Hounsgaard and Kiehn, 1989; Bennett et al., 2001; Perrier and Hounsgaard, 2003; Heckman et al., 2004). It was originally suggested that the facilitation of plateau potentials was a consequence of the inhibition of the AHP (Hounsgaard and Mintz, 1988; Schmidt and Jordan, 2000). Here, we show that synaptic release of 5-HT increases the excitability of motoneurons during a few seconds, both by inhibiting the AHP and by facilitating plateau potentials. Several lines of evidence suggest, however, that the two effects are attributable to different mechanisms occurring in parallel. First, a depolarizing current that did not generate action potentials before stimulation of the raphé spinal pathway induced action potentials afterward (Figs. 2B, 3B). Inhibition of the AHP cannot explain this result. Second, voltage-clamp experiments showed that the main current modulated by 5-HT₂ receptors after DLF stimulation is the one responsible for the hysteresis in I–V plots (Fig. 4). This suggests that a specific pathway links 5-HT₂ receptors with the current responsible for the plateau potential. Third, the inhibition of the AHP is induced by the activation 5-HT_1A receptors in different types of motoneurons (Bayliss et al., 1995; Wikstrom et al., 1995) including turtle motoneurons (Grunnet et al., 2004). Thus, inhibition of the AHP and facilitation of plateau potentials are induced by two different subtypes of 5-HT receptors.

Other studies have ascribed the facilitation of plateau by 5-HT to the inhibition of a leak conductance that would enhance the persistent inward current responsible for plateau potentials (Heckman et al., 2003). In support of this idea, activation of 5-HT_1A receptors increases the input resistance of turtle motoneurons by inhibiting a leak conductance (Perrier et al., 2003). However, the experiments performed in the slice preparation show that stimulation of the DLF essentially modulates the inward current responsible for the hysteresis in I–V plots (Fig. 4A) without significantly affecting the input resistance of motoneurons (Fig. 4D). This result implies a specific pathway between serotonergic receptors and L-type Ca²⁺ channels. This apparent discrepancy may be explained by the fact that, in the present study, the excitability of motoneurons was always tested several hundreds of milliseconds after synaptic release of transmitter, whereas in other studies, the increase in input resistance was observed during the continuous presence of 5-HT agonists in the extracellular medium. It is also possible that stimulation of the DLF or of the raphe nucleus was too brief to inhibit a leak conductance.

The persistent inward current recorded in slices was measured in the presence of the intracellular blocker of Na⁺ channels, QX-314, applied through the microelectrode. This drug also reduces Ca²⁺ currents (Talbot and Sayer, 1996). However, stimulation of the DLF promoted a hysteresis, showing that an effect on Ca_V1.3 current was minor. Nevertheless, we cannot exclude that QX-314 was partly responsible of the reduction of the hysteresis observed in the presence of the SB-206553.

5-HT receptor subtype involved in the facilitation of plateau potentials

Previous studies demonstrated that extracellular addition of the 5-HT_2A/C receptor agonist (±)-1-(2,5-dimethoxy-4-iodophenyl)-2-aminopropane hydrochloride (DOI) (Ichikawa and Meltzer, 1995) promotes plateau potentials in spinal motoneurons (Perrier and Hounsgaard, 2003). In the present study, we show that plateau potentials promoted by 5-HT are inhibited by SB-206553, which is an antagonist for 5-HT_2B/C receptor subtypes (Forbes et al., 1995). Therefore, it seems plausible that 5-HT_2C receptors are responsible for the facilitation of plateau potentials in motoneurons.

Functional considerations

By showing that synaptic release of 5-HT induced by stimulation of the raphe nucleus promotes plateau potentials in motoneurons, we have established a mechanism by which the brain can regulate the excitability of motoneurons. In rats with complete spinal transections, activation of 5-HT_2C receptors increases the ability of animals to perform weight-supported locomotion (Kim et al., 2001). Moreover, the 5-HT_2A/C agonist DOI upregulates of the excitability of extensor motoneurons in acute spinal cats (Miller et al., 1996). Both observations could be explained by the facilitation of plateau potentials in extensor motoneurons. This hypothesis could be linked to the fact that plateau potentials promoted by 5-HT contribute to the shaping of the discharge pattern of motoneurons during withdrawal behaviors in frogs (Perrier and Tresch, 2005).