The monosynaptic connections between the serotonin-containing LP3 and RPas neurones in Helix are serotonergic

In a number of molluscan species injection of the serotonin analogues 5,6 or 5,7-dihydroxytryptamine (5,6- or 5,7-DHT) into intact animals results in selective accumulation of dark pigment granules in the somata of serotonin-containing neurones (Helix;Balaban et al. 1985; S.-Rózsa et al. 1986; Aplysia:Jahan-Parwar et al. 1987; Lymnaea ; Kemenes et al. 1989). However, the pharmacological sensitivity and membrane properties of the pigment-labelled neurones are not affected by the presence of these pigment granules in the soma. The serotonin (5-HT) content of the axonal elements is only transiently depleted for a few days about 2 weeks after injection with 5,6- or 5,7- DHT and during these few days the ability of the serotonergic neurones to activate follower cells is abolished (Gadotti et al. 1986; Kemenes and S.-Rózsa, 1987; Jahan-Parwar et al. 1987; Vehovszky et al. 1988; Kemenes et al. 1988, 1990). Following recovery from the transient effects of these neurotoxins, the 5-HT neurones containing 5,6- or 5,7-DHT-induced pigment are easily visible and so they make good targets for electrophysiological investigations in both isolated and semi-intact preparations.

Using this novel approach, we have previously described monosynaptic connections between pigment-labelled neurones in the snail Helix pomatia (Vehovszky et al. 1989). The excitatory synaptic effects between the giant left pedal LP3 cell and its followers in the right parietal ganglion (RPa_s cells) are chemically mediated. The presence of dark pigment in both LP3 and its followers in 5,6-DHT-treated snails suggested that these neurones contained serotonin (S.-Rózsa et al. 1986). This was later supported by immunocytochemistry (Hernádi et al. 1989). However, evidence for the serotonergic nature of neurotransmission between LP3 and the follower cells was lacking. Without this it could not be excluded that LP3 uses not 5-HT but a different neurotransmitter which might coexist with serotonin in the cell body.

In the present paper we provide pharmacological evidence to show that the monosynaptic connections between the pedal LP3 cell and its followers in the right parietal ganglion are serotonergic. In addition, we give a morphological description of the axonal branching pattern of LP3 and the follower cells, and suggest that through their widespread connections these neurones have regulatory roles in a variety of serotonergic functions in Helix.

Adult Helix pomatia L. were collected locally on the Tihany peninsula (Hungary). The snails were maintained in an active state in the laboratory for several weeks prior to the experiments.

We used conventional microelectrophysiological techniques to make pairwise intracellular recordings from LP3 and its follower cells (RPa_s neurones) in isolated preparations of the central nervous system (Vehovszky et al. 1989). The experimental chamber was perfused with normal Helix saline: NaCl, 80mmoll^-1; KC1, 4 mmol I^-1; CaC1₂.2H₂O, 10 mmol l^-1; MgC1₂.6H₂O, 5 mmol l^-1; Tris, 4 mmol l^-1 (pH7.4).

For local application, serotonin was ejected ionophoretically onto the surface of the cell body of neurones. The microelectrode used for application was filled with a 0.1 mol l^-1 solution of serotonin creatinine sulphate (Sigma) and the tip was positioned adjacent to the surface of the cell body. Ionophoretic application of 5-HT was carried out by passing 0.5–1 s, 5–20 nA positive current pulses through the electrode. To prevent the possible desensitizing effect of the transmitter leaking from the tip of the electrode, prior to and between the applications a constant 0.5 nA negative retention current was used.

In each test we first injected depolarizing current into LP3 to evoke reliable postsynaptic responses in the followers. 10 s after the test with presynaptic activation we applied 5-HT ionophoretically onto the same postsynaptic RPa_s neurone. Three replicate presynaptic stimuli and 5-HT applications were used to establish control responses in normal saline. The same procedure was then repeated in the presence of 10⁻⁷ and 10⁻⁶moll^-1 5-HT or various serotonin antagonists in the bath. Within the same experiment both the duration and frequency of the presynaptic bursts and the ionophoretic current used to apply 5-HT were the same before and after perfusion with antagonists.

The following drugs were used: MDL 72222-EF02 (Merrell), cinanserin HC1 (Squibb), 7-methyltryptamine (Koch Light), tryptamine hydrochloride (Sigma), bufotenine-hydrogenoxalate (Fluka AG). Bufotenine, tryptamine and 7-methyltryptamine are known serotonin antagonists in the gastropod central nervous system (reviewed by Walker, 1986). MDL 72222-EFO2 and cinanserin (known serotonin antagonists in vertebrates: Richardson and Engel, 1986; Fozard, 1987) were also found to be selective 5-HT receptor antagonists on Helix neurones (Vehovszky and Walker, 1991).

The drugs were tested in the concentration range from 10⁻⁷ to 10⁻⁴ mol1^-1 (made up in normal saline). Prior to testing, 5 min rest periods allowed the drugs to equilibrate in the perfusion chamber. The reversibility of the antagonist effect was tested by washing out with normal saline. The effect of each drug on the synaptic and 5-HT responses was tested on 5–8 different preparations. To exclude changes in the postsynaptic response due to changes in the membrane potential, we set the membrane potential of the parietal neurones to —100 mV prior to testing either the synaptic or the serotonin response.

After electrophysiological experiments the LP3 and RPa_s neurones were filled intracellularly with Ni²⁺-lysine solution (Fredman, 1987). To develop the chemical reaction for staining, rubeanic acid solution was used according to the method employed by Quicke and Brace (1979). After dehydration in graded alcohols and clearing in methyl salicylate, the suboesophageal ganglia were mounted on slides in Canada balsam. Wholemount preparations were photographed or stained cells were traced by using a drawing apparatus attached to a stereomicroscope.

The LP3 neurone is located on the dorsal surface of the medio-rostral lobe of the left pedal ganglion (S.-Rózsa and Logunov, 1981 ; Vehovszky et al. 1989 and Fig. 1). The cell body has a diameter of 100–120 μm and often contains a yellowish pigment granule.

The LP3 cell has a pseudo-bipolar shape with a thick axon trunk dividing into two main branches very close to the soma (Fig. 1). One of the main branches runs in the pleuro-pedal connective, sending collaterals to the cerebro-pedal connective and the pedal neuropile. This latter branch leaves the pedal ganglion through pedal nerves V and VI. The fine branch in the pleuro-pedal connective further divides into two collaterals, one running in the very thin pharyngeal retractor muscle nerve (n.m.ph.: nervus musculi retractoris pharyngealis, Schmalz, 1914), the other projecting to the visceral and right parietal ganglia (Fig. 1).

The other branch of the main axon trunk enters the contralateral pedal ganglion where its axonal branching pattern is virtually symmetrical with that on the ipsilateral side. The two main branches of the LP3 axon run through the neuropile of each of the main ganglia and seem to connect all the ganglia in the suboesophageal complex in a ring-like manner (Fig. 1B).

The follower cells of LP3 in the right parietal ganglion are located in a single cluster of 4–5 cells on the dorsal side of the ganglion and close to its medial border. Besides the similarities in their electrophysiological properties, synaptic inputs and chemical sensitivity, they are also similar in their morphology. Their somata have diameters of 100–130 μm and a similar unipolar shape. They send axonal collaterals through peripheral nerves (right pallial, anal and intestinal nerves) as well as to the visceral and left parietal ganglia (Fig. 1B). Because of the morphological and physiological similarities of these neurones they were treated as members of a homogeneous cluster and named RPa_s (serotonin-containing right parietal cells). Axon branches from LP3 and the RPa_s cells run in close proximity in the visceral ganglion (Fig. 1B).

Serotonin, when applied to the membrane of the RPa_s cell body from a pipette, produced an excitatory response very similar to that evoked by stimulating LP3 (Fig. 2A,B). After hyperpolarizing the membrane of the RPa_s cells, the 5-HT-evoked depolarizing effect was similar to the summated excitatory postsynaptic potentials (EPSPs) following a series of action potentials (APs) in LP3 (Fig. 2C). Serotonin also depolarized LP3 when applied locally to its soma membrane. The depolarization was sufficient to generate a burst of spikes in LP3, which then excited the RPa_s followers (Fig. 2B). The similarity of the synaptic and serotonin-evoked excitation recorded in the RPa_s neurones allowed us to compare the effects of drugs on both types of responses.

To test the serotonergic nature of the synaptic connection between LP3 and the RPa_s neurones, we first compared the desensitizing effect of bath application of serotonin on the responses of the RPa_s neurones to LP3 stimulation and somatic ionophoresis of serotonin. Low concentrations (10⁻⁷ and 10⁻⁶moll^-1) of serotonin in the bath reduced or inhibited the responses of the RPa_s neurones to ionophoretic serotonin (Fig. 3). The same concentration of bath-applied serotonin also reduced the size of the compound EPSP in the RPa_s cells, but did not prevent the presynaptic spike discharge caused by intracellular electrical stimulation of the LP3 neurone (Fig. 3B,C).

In this second series of tests we examined the synaptic and serotonin-evoked responses of the RPa_s cells in the presence of known serotonin antagonist drugs.

All the drugs except tryptamine had some initial general excitatory effect on both the pre- and postsynaptic neurones. The general excitability of the cells increased in the presence of the drugs, resulting in more frequent firing and an increased number of spontaneous synaptic potentials (see Figs 6, 7). However, all the drugs tested antagonized the effect of ionophoretically applied serotonin on the RPa_s cells, and all but one antagonized the synaptic response as well. This was found in a minimum of four replicate preparations for each drug.

Tryptamine (10⁻⁵moll^-1) abolished the serotonin response (Fig. 4B), while 10⁻⁴moll^-1 tryptamine blocked both the synaptically evoked and serotonin responses (Fig. 4C). The blocking effect of tryptamine was reversible (Fig. 4D).

The serotonin analogue 7-methyltryptamine had a weaker effect than tryptamine but still blocked both the synaptic and the 5-HT-evoked responses (Fig. 5). After washing, the summated EPSPs reappeared but the serotonin response was only partially restored (Fig. 5D). ‘

Bufotenine had the strongest general excitatory effect on both the LP3 and RPa_s neurones. In the presence of 10⁻⁶moll^-1 bufotenine the spontaneous activity of LP3 and the number and amplitude of spontaneous EPSPs from other neurones seen in RPa_s cells both increased, as did the amplitude of excitatory potentials in the RPa_s neurones caused by LP3 activity (Fig. 6). In contrast, the depolarization evoked by locally applied serotonin decreased at the same time (Fig. 6B). After a longer time [or in a higher concentration of bufotenine (10⁻⁵ mol l^-1)] both the synaptically and pharmacologically evoked responses were blocked (Fig. 6B). After washing, the summated EPSPs reappeared again, but the depolarization evoked by serotonin was only partially restored (Fig. 6C).

10⁻⁵mol I^-1 MDL 72222-EFO2 had a general enhancing effect on the EPSPs similar to that of bufotenine. At a higher concentration (5×10⁻⁵ mol l^-1), this drug had a blocking effect on both the synaptic and 5-HT-evoked responses, which was irreversible (Fig. 7A), unlike that of the previous drugs.

Cinanserin (5×10⁻⁵ mol l^-1) strongly and irreversibly blocked the 5-HT-evoked response of the RPa_s cells, but had no effect on the synaptic response evoked by presynaptic discharges of LP3 (Fig. 8).

We have shown that the monosynaptic excitatory connection from the serotonin-containing left pedal LP3 neurone to its right parietal followers in Helix pomatia (described by Vehovszky et al. 1989) is mediated by serotonergic mechanisms. This is important because a variety of transmitters can coexist in neurones, and the observation that a cell contains serotonin does not necessarily mean that it uses it as a transmitter at a particular synapse.

Serotonin (5-HT) applied locally onto the cell body of RPa_s neurones depolarized these cells (Figs 2–8) and so mimicked the excitatory effect of the electrical stimulation of the presynaptic LP3 cell. The desensitizing effect of serotonin in the bath on both the synaptically evoked and the serotonin responses (Fig. 3) suggests that the same receptors are involved in the mediation of both types of responses. Furthermore, the excitatory effect of locally applied 5-HT and the synaptic responses evoked by presynaptic stimulation could be similarly blocked by bath application of serotonin antagonists at concentrations of 10^_6-10⁻⁴moll^-1. All but one of the drugs also had a general excitatory effect on both LP3 and the RPa_s cells. We suggest that this may be due to an initial serotonin agonist effect of the bath-applied antagonists (Vehovszky and Walker, 1991; Walker, 1985, 1986).

In Helix aspersa, the parietal F4, F5 and F6 cells (Kerkut et al. 1975) are homologous with the RPa_s neurones of Helix pomatia. Exogenously applied serotonin evoked a Na⁺- and Ca²⁺-dependent depolarization of the H. aspersa neurones (Wright and Walker, 1984; Bokisch and Walker, 1986) similar to that seen in the RPa_s cells following either the application of 5-HT or stimulation of LP3 (this study). The serotonin-evoked responses of the F cells were blocked by the application of the same serotonin antagonists that we used (Wright and Walker, 1984; Walker, 1985). In voltage-clamp studies of the Helix aspersa neurones, the two-component inward current evoked by serotonin could also be blocked with 7-methyltryptamine (Paupardin-Tritsch et al. 1981).

The serotonin analogue tryptamine had a highly selective antagonist effect on the serotonin-evoked excitatory responses (Gerschenfeld and Paupardin-Tritsch, 1974; Vehovszky and Walker, 1991), and in our study proved to be an effective blocker of both the serotonin and synaptically evoked responses of RPa_s neurones.

MDL 72222-EFO2 and cinanserin, often used as serotonin antagonists in vertebrates (Richardson and Engel, 1986; Fozard, 1987) are selective but nearly irreversible 5-HT receptor antagonists in Helix neurones (Walker and Vehovszky, 1989). MDL 72222-EFO2 had a similarly irreversible antagonist effect on the synaptically evoked responses of right parietal neurones after LP3 stimulation.

The only antagonist which at a concentration higher than 10⁻⁵moll^-1 inhibited the pharmacological effect of locally applied serotonin but did not affect the synaptic response on the same cell was cinanserin (highest concentration tested 5×10⁻⁵ mol 1^-1). However, the threshold concentrations of the other effective antagonists such as MDL 72222-EFO2, tryptamine, 7-methyltryptamine and bufotenine were also higher for blocking the synaptic response than for reducing or abolishing the serotonin-evoked depolarization. This difference could be due to differences in the concentration of the exogenously applied and synaptically released 5-HT but is more likely to be due to the different localization of the postsynaptic and soma receptors, the latter being more accessible for exogenously applied drugs.

The only other identified serotonergic monosynaptic connections in Helix are made by the giant serotonergic cells (GSCs) of the cerebral ganglion onto followers in the buccal ganglia. These cells and their connections have homologues in several molluscan species (reviewed by Pentreath et al. 1982). In many respects the pedal LP3 neurone shares common features with the cerebral GSCs in Helix. Both cells contain serotonin, have similar electrophysiological characteristics and are excited by serotonin themselves (see Cottrell, 1982). Serotonin applied exogenously to either the RPa_s or the buccal M cells mimicked the effect of the stimulation of the presynaptic cell on both cell types. In addition, in both the RPa_s cells and the buccal follower neurones of the GSCs the fast excitatory effect of 5-HT could be blocked by bath application of bufotenine, tryptamine and 7-methyltryptamine (reviewed by Walker, 1986). However, there is one major difference between the two types of serotonergic connection: in contrast to the buccal followers of the serotonergic GSCs, the RPa_s cells themselves (like their presynaptic LP3 neurone) contain serotonin (Vehovszky et al. 1989; Hernádi et al. 1989). This means that in their connections with other neurones the RPa_s cells can also act as presynaptic serotonergic cells.

The pedal LP3 cell has a widespread axonal arborization with processes running in peripheral nerves and in the neuropile of central ganglia. The most likely site where the monosynaptic connections between LP3 and the RPa_s cells are made is in the visceral ganglion where the LP3 and RPas axons run in close proximity.

The identified monosynaptic connections between the serotonergic pedal LP3 neurone and its serotonin-containing right parietal follower cells could be active in a variety of serotonin-dependent functions, such as feeding, locomotion and excitation of the heart. In this way the pedal LP3 neurone or other LP3-type neurones may link several neural networks controlling different motor functions.

This work was supported by an OTKA grant to Katalin S.-Rózsa from the Hungarian Governmental Grant Commitee. We thank Dr C. J. H. Elliott for reading the manuscript.