Serotonin Depresses the After-Hyperpolarization Through the Inhibition of the Na⁺/K⁺ Electrogenic Pump in T Sensory Neurones of the Leech

In each segmental ganglion of the leech nervous system, three types of mechanosensory neurones have been identified: touch (T), pressure (P) and nociceptive (N) cells (Nicholls and Baylor, 1968). The firing discharge of these neurones produces an after-hyperpolarization (AHP) which is involved in important physiological functions: it leads to the inhibition of the previously activated pathway, to an elevation of the volley threshold, to a decrease of synaptic efficacy at the axon terminals and, primarily, to a conduction block at the branching points where small axonal processes join the main neurite (Baylor and Nicholls, 1969; Jansen and Nicholls, 1973; Yau, 1976; Van Essen, 1973).

Existing electrophysiological research, analyzing the cellular mechanism of non-associative learning in the swimming (Brunelli et al. 1985a, 1986; Debski and Friesen, 1987; Nusbaum and Kristan, 1986) and shortening (Bagnoli et al. 1973; Belardetti et al. 1982; Frank et al. 1975) behaviour patterns, triggered by a light touch to the skin, has made it possible to demonstrate a clear effect of the endogenous monoamine serotonin (5-HT) (Lent and Frazer, 1977; Lent et al. 1979) on the AHP of T neurones. A previous paper (Belardetti et al. 1984) has shown that AHP amplitude can be depressed for prolonged periods either by the activation of serotonergic Retzius cells (a pair of giant neurones located in each segmental ganglion) or by 5-HT application. The effect is dose-dependent and is partially blocked by the 5-HT antagonist methysergide.

In this paper we have investigated the mechanism of action of 5-HT on the AHP of T sensory neurones. Since the AHP in these cells is partly due to the activation of the electrogenic Na⁺/K⁺-ATPase and partly to a Ca²⁺-dependent K⁺ conductance (gK.ca) (Baylor and Nicholls, 1969; Jansen and Nicholls, 1973; Van Essen, 1973), we performed two series of experiments: in the first, we tested the effect of 5-HT on the residual AHP after inhibiting g_K,ca; in the second, we investigated the effect of 5-HT on T cells in which the activity of the Na⁺/K⁺-ATPase was modified.

The data indicate that 5-HT may act through the inhibition of the Na⁺/K⁺ electrogenic pump.

Adult leeches of the species Hirudo medicinalis were purchased from a local supplier and kept in a well-aerated aquarium at 15 °C. The animals were anaesthetized with 10% ethanol in water and pinned, ventral side up, to the paraffin wax floor of a plastic chamber. The ventral nerve cord was exposed by cutting the body wall along the midline and opening the ventral sinus. A short chain of ganglia was isolated at midbody level and pinned to the bottom of a small recording chamber coated with Sylgard (Dow Corning).

Unless otherwise stated, the following leech saline was used: 115 mmol l⁻¹ NaCl, 4mmoll⁻¹ KC1, 1.8mmoH⁻¹ CaCL, 10mmoll⁻¹ glucose, buffered to pH7.4 with Tris-maleate.

In some experiments trains of depolarizing pulses (200 ms, 2–3 s⁻¹, 25 s duration) were injected intracellularly into T cells; the discharge frequency of each series of trials was kept constant by adjusting the amount of current injected. Microelectrodes filled with 3–4 mol l⁻¹ potassium acetate with resistances ranging from 40 to 100 MΩ were used for intracellular recordings and stimulations.

In one group of experiments the temperature was lowered by circulating a mixture of acetone and dry ice on the external surface of the recording chamber. The temperature of the bath was constantly monitored with an electronic thermal probe. The shift produced by temperature variation was detected by a second microelectrode placed in the recording chamber and algebraically subtracted from the voltage measured by the microelectrode in the cell. The earth wire was located far from the recording chamber, using an agar/salt bridge. Serotonin (serotonin creatinine sulphate) (Sigma) was freshly dissolved in saline.

Only T cells with a resting potential of at least 40–50 mV, an input resistance greater than 20 MΩ and an action potential of 60–80 mV were selected. The AHP is very sensitive to the quality of the impalement: therefore, only cells with an AHP of at least 10–15 mV were used for the experiments.

During the recording session, the ganglia were perfused at a rate of 1.5 ml min⁻¹ using a peristaltic pump. All the experiments were displayed on the screen of a storage oscilloscope and collected on a video recorder connected to a pulse code modulator (PCM501 ES) (Sony).

We measured the effects of different concentrations of 5-HT on some electrophysiological parameters of T sensory neurones. As pointed out in a previous paper (Belardetti et al. 1984), although a dose-dependent reduction of the AHP occurred, we have never seen consistent changes in the shape of the action potential (see Fig. 4A) or in the input resistance (see Fig. IB), or a modification of the resting potential (see Fig. 1C). When 50μmoll⁻¹ 5-HT was applied, the maximal effect on AHP amplitude occurred 10 min after the removal of 5-HT by a saline wash; with 10μmoll⁻¹ 5-HT the maximal effect occurred 10min after 5-HT application (Fig. 1).

Different approaches were used to investigate the mechanisms underlying 5-HT modulation of the AHP recorded in T neurones. Responses to 5-HT were first examined after the addition of g_K,Ca blockers (BaCl₂ or CdCl₂), which eliminated the component that contributes to the generation of the AHP. The effects of 5-HT were then studied after increasing or decreasing the activity of the Na⁺/K⁺-ATPase.

In a preparation in which the AHP had been generated by trains of action potentials elicited by intracellular depolarizing pulses, CaCl₂ was completely replaced by BaCl₂ (1.8 mmol l⁻¹) to block g_K.Ca from the intracellular side of the Ca²⁺ channel (Hagiwara, 1981; Meech, 1978; Paupardin-Tritsch et al. 1981). A 5 min application of BaCl₂ produced a reversible reduction of AHP amplitude to 60% of the initial response (taken as 100%). Prolonged applications (up to 30 min) did not reduce the AHP to a greater degree (Fig. 2B). During the 5 min BaCl₂ application the spike shape changed dramatically: the repolarization slowed and the undershoot disappeared (see Fig. 4B). After the initial 5 min of BaCl₂ application, perfusion with the blocking agent was continued and 50 or 10 μmol l⁻¹ 5-HT was applied for 10min. A further clear-cut reduction of AHP amplitude (to 25 % of the initial response with 50μmol l⁻¹ 5-HT or 40% with 10μmol l⁻¹ 5-HT) was observed (Fig. 2), which agreed with the kinetics shown by 5-HT alone (Fig. 1A). The effect of the monoamine reversed after a 15 min wash in saline with BaCl₂, and an additional 15 min of perfusion with normal physiological solution reversed the effect of BaCl₂, restoring, almost completely, the AHP amplitude. The effects are illustrated and summarized in Fig. 2B.

A similar experiment was performed using 0.1 mmol l⁻¹ CdCl₂ in normal saline to block g_K,Ca from the extracellular side of the Ca²⁺ channel (Hagiwara, 1981; Madison and Nicoll, 1982; Meech, 1978) (Fig. 3). The same results were obtained as in the previous experiment. In this case the trace was less noisy with no spontaneous spikes after the testing train: this is probably attributable to the fact that, in the experiment with BaCl₂, Ca²⁺-free solution was used and consequently cellular excitability increased. A 5 min application of CdCl₂ induced a decrease of AHP amplitude to 60% of the initial response, with no further reduction when application was continued for up to 30 min (Fig. 3B). A 5 min application of CdCl₂ caused an increase in spike duration and a disappearance of the undershoot (Fig. 4C). A 10 min incubation with 5-HT led to a further depression of the AHP amplitude, to approximately 30 % of the initial response with 50 μmol l⁻¹ 5-HT or approximately 45% with 10μmoll⁻¹ 5-HT (Fig. 3). The effect was partially restored after washing out the 5-HT (Fig. 3B).

When T neurones were impaled with a 3 mol l⁻¹ sodium acetate microelectrode a clear hyperpolarization (about 15 %) could be observed (Fig. 5A). It is assumed that this effect is due to the activation of the Na⁺/K⁺ electrogenic pump following Na⁺ leakage from the microelectrode into the cell (Jansen and Nicholls, 1973). When potassium acetate microelectrodes were used, no variation of membrane voltage (V_m) could be detected (Fig. 5B).

In some experiments we perfused the ganglia with 50μmoll⁻¹ 5-HT for 10 min before penetrating T cells with sodium acetate electrodes and we obtained a depolarization of membrane voltage of about 15% (Fig. 5C). This effect may be explained by an inhibitory action of 5-HT on the Na⁺/K⁺-ATPase so that Na⁺ leaking into T cells cannot be pumped out and a depolarization results. Fig. 5D clearly illustrates the comparative membrane voltages of T neurones after penetration with 3 mol l⁻¹ potassium acetate or 3 mol l⁻¹ sodium acetate with or without the application of 5-HT.

In experimental preparations in which the ganglia were perfused with K⁺-free saline, we obtained a biphasic variation of membrane voltage. An early hyperpolarization phase, probably due to a greater K⁺ outflow caused by the rise of the chemical gradient, was followed by a delayed depolarization phase which was sensitive to ouabain and, therefore, mainly due to the inactivation of the electrogenic pump, which is known to be sensitive to the external K⁺ concentration (Fig. 6A). These results agree with those found in P and N cells by Schlue and Deitmer (1984).

If, at the end of the depolarization phase, we perfused the ganglia with K⁺-free saline containing 4 mmol l⁻¹ Cs⁺, a Na⁺/K⁺-ATPase activator (Skou, 1960,1965), we observed a membrane repolarization (Fig. 6C). Alternatively, if 50μmoll⁻¹ 5-HT was applied 10 min before CsCl, the repolarization was greatly reduced (by approximately 30%) and delayed (Fig. 6B).

The V_m of T neurones was monitored during the reversible cooling of the preparation from 18 to 10°C. During cooling a depolarization of about 15 % was obtained (Fig. 7). When the temperature was restored to 18°C a repolarization was observed; this was sensitive to ouabain and strophantidin and often rose to 105–110 % of the initial V_m. This effect is due to the inhibition and reactivation of the Na⁺/K⁺-ATPase (Baylor and Nicholls, 1969; Kuba and Koketsu, 1979).

The same procedure was adopted after a 10 min application of 50 μmol l⁻¹5-HT, producing a larger depolarization phase of about 25–30% and a smaller repolarization phase reaching 90% of the initial voltage (Fig. 7).

The Na⁺/K⁺-ATPase was blocked with 2×10⁻⁴ mol l⁻¹ ouabain or strophantidin in order to study the effect of 5-HT on the residual AHP; however, the remaining AHP was often very small (Baylor and Nicholls, 1969; Jansen and Nicholls, 1973) and it was difficult to maintain a constant discharge frequency, probably because of an excess of Na⁺ in the T cell that cannot be pumped out. In a few cases in which a residual AHP could be detected, 5-HT was unable to provoke a further reduction of the AHP amplitude (data not shown).

In a previous paper (Belardetti et al. 1984) it was shown that 5-HT perfusion or the electrical stimulation of the serotonergic Retzius cells produced a long-lasting clear-cut reduction of AHP amplitude in T sensory neurones. This effect depended on the concentration of the monoamine, was inhibited by the antagonist methysergide and was not blocked by bathing the ganglion with high-Mg²⁺ saline. In agreement with the observations of Belardetti et al. (1984), we have never seen consistent changes of T cell resting potential, input resistance or action potential shape with 5-HT perfusion (Figs 1, 4A).

From these preliminary observations, it was difficult to ascribe the mechanism of action of serotonin to an inhibition of g_K,Ca, which is known to make a partial contribution to the AHP (Baylor and Nicholls, 1969; Jansen and Nicholls, 1973); in fact, no increase of input resistance was observed. A simple short-circuit of the currents generating the AHP probably also fails to explain the action of 5-HT; a 70–75% decrease of the AHP amplitude, obtained with 50μmoll⁻¹ 5-HT perfusion for 10 min, should result from an identical reduction of the input resistance. A possible shunting action of 5-HT is also unlikely because it would affect the AHPs identically, irrespective of their amplitudes, whereas large AHPs are more depressed than the small ones (Belardetti et al. 1984).

To clarify the mechanism of action of 5-HT on T neurones, we performed two series of experiments. In the first series, we blocked g_K,Ca with BaCl₂ or CdCl₂ and demonstrated that 5-HT was still able to reduce AHP, even after g_K,Ca inhibition. We also observed that Ba²⁺ or Cd²⁺ application caused an unexpectedly large reduction (about 40%) of the AHP amplitude, which is known to depend greatly on the activation of the electrogenic pump in T neurones. This may be caused by the reduction of input resistance that accompanies BaCl₂ or CdCl₂ application (about 16% and 11 %, respectively), probably due to an increase of membrane permeability following variation of Ca²⁺ concentration. Moreover, an interaction between g_K,ca and electrogenic pump activity cannot be ruled out. An involvement of 5-HT in g_K,Ca is made unlikely by the finding that, in experiments in which it was possible to get a residual AHP following ouabain or strophantidin treatment, 5-HT was ineffective in provoking a further reduction of the AHP amplitude.

We also tried to study the effect of 5-HT by holding T cells at a membrane potential close to the K⁺ reversal potential, but the long projections of the neurite do not allow good space-clamping even with double-electrode voltage-clamp techniques (Jansen and Nicholls, 1973; Stewart et al. 1989).

In the second series of experiments, it has been possible to collect evidence for a direct inhibition of the Na⁺/K⁺-ATPase by 5-HT. Injection of Na⁺ into T neurones after perfusion with 5-HT led to a depolarization instead of the normal, ouabain-sensitive hyperpolarization. This effect may be explained by an inhibition of the electrogenic pump by 5-HT and by leakage of positive ions into the cell. These ions cannot then be pumped out.

Moreover, 5-HT caused a reduction of the repolarization normally obtained with the Na⁺/K⁺ pump activator Cs⁺ after K⁺-free perfusion. Cs⁺ has a permeability of less than 8 % for the K⁺ channel (taking the permeability of K⁺ for its channel as 100%) (Stryer, 1982) and, therefore, its effect on the Na⁺/K⁺ electrogenic pump is not masked by the current through the K⁺ channel, as happens when K⁺ itself is used to reactivate the pump.

In addition, 5-HT increased the depolarization phase, during cooling of the ganglia and produced a considerable reduction in speed of repolarization when the temperature was returned to 18°C and, consequently, the electrogenic pump should have been reactivated. The inhibitory action of 5-HT on the electrogenic pump is a surprising result since, although examples of positive modulation of the Na⁺/K⁺-ATPase caused by neurotransmitters have been described (Kuba and Koketsu, 1979; Phillis and Wu, 1981; Thomas, 1972), inhibitory effects have rarely been found and have been investigated only with biochemical methods (Lingham and Sen, 1983; Mourek, 1988).

A further observation is that 5-HT has a reduced effect on the AHP amplitude of P sensory cells and no effect at all on that of N neurones (Brunelli et al. 1985b), the AHP of which is more dependent on g_K,Ca (Baylor and Nicholls, 1969; Jansen and Nicholls, 1973; Van Essen, 1973).

A possible effect of 5-HT on the Na⁺/K⁺-ATPase might be to change the transport coupling ratio, making the pump work with a less electrogenic mechanism. Through the inhibition of the electrogenic pump, 5-HT may have a role in removing the conduction blocks caused by the AHP. This potentiating effect may, at least in part, explain the behavioural sensitization and dishabituation induced by 5-HT and observed in experiments on the induction of swimming (Brunelli et al. 1985a, 1986; Debski and Friesen, 1987; Nusbaum and Kristan, 1986) and on the ‘fast conducting system’ (Belardetti et al. 1982; Frank et al. 1975).

It will be interesting to investigate whether a similar inhibitory action of the electrogenic pump can be found in other leech neurones, in other nervous systems or in non-excitable tissues.

This work was supported by the Ministero della Pubblica Istruzione: 40–60%, 1987–1988. Additional funding was provided by Regione Toscana, Italy, grant number 3342, 1987.