Neural Control of Heart Beat in the African Giant Snail, Achatina Fulica Férussac: I. Identification of the Heart Regulatory Neurones

It is well known that the molluscan heart’ is myogenic but that its beating is regulated by the central nervous system. Until now, this neural regulation has been studied mainly in two species (Aplysia californica and Helix pomatia). In Aplysia, two heart excitors, two heart inhibitors and higher order interneurones have been identified (Mayeri et al. 1974; Koester, Mayeri, Liebeswar & Kandel, 1974). In Helix, two heart inhibitors, four heart excitors and one interneurone have been found as well as many neurones which receive synaptic inputs from the cardio-renal system (S.-Rózsa & Salánki, 1973; S.-Rózsa, 1979a, 1981). The heart regulatory network in the African giant snail, Achatina fulica Férussac, has also been studied (S.-Rózsa, 1979b), but the heart excitatory motoneurones have not been identified.

In the present study, heart regulatory neurones in the central ganglia of Achatina fulica Férussac were identified as a first step towards understanding the neural regulation of the heart.

The African giant snail, Achatina fulica, captured in Okinawa and transported by air to Hiroshima, was bred in our laboratory at 24°C. The preparation consisted of the cerebral ganglia, suboesophageal ganglia, intestinal nerve and heart. A branch of intestinal nerve which goes to the heart further bifurcates near the heart. The larger branch goes to the auricle along the pericardium and the smaller one goes to the aorta. In the present study, the pericardium was cut around the base of the ventricle and the heart was exposed to record the heart beat efficiently. This treatment involved transection of the smaller nerve branch going to the aorta. Thus, the possible innervation of the heart muscle by this branch is neglected in the present experiments.

The preparation was pinned to the bottom of a chamber covered with a silicone resin (Fig. 1A). The recording chamber was separated into two compartments by a partition with a slit through which the intestinal nerve was led. The slit was sealed with silicone grease, which enabled independent perfusion of the compartments (ganglia compartment and heart compartment). The heart was perfused by means of a cannula inserted into the vein. Usually, the outer thick connective capsule but not the inner thin sheath covering the ganglia was removed by dissection. In some experiments, the inner thin sheath was also carefully removed to expose the nerve cells.

The heart beat and the neuronal activity were recorded simultaneously. Intracellular recording and stimulation of neurones were carried out using microelectrodes filled with 3 mol l⁻¹ potassium acetate, of 5–10 MΩ resistance. In some cases, a second microelectrode was inserted into a cell for current injection. Heart beat was monitored using a strain gauge connected to the aorta by fine thread. The aorta was cut open between the ligating point and the ventricle to allow exudation of the perfusate. In a few cases, an extracellular recording of the intestinal nerve activity was made at the point just before it entered the pericardium using a Ag–AgCl bipolar electrode. This electrode was also used for stimulation. The data were stored on an FM tape recorder for later analysis and permanent records were produced using a pen-recorder.

Intracellular staining was performed by pressure injection of 5 % Lucifer Yellow CH (Sigma Chemical Co.) based on the standard method of Stewart (1978). The ganglia were fixed in 4 % formaldehyde for 12 h at 4°C, and then dehydrated, cleared in methylbenzoate, and viewed in wholemount using a fluorescence microscope. The stained neurone was photographed at several depths and reconstructed.

The composition of normal physiological solution was as follows (in mmol I^-1): NaCl, 61; KC1, 3·3; CaC1₂, 10·7; MgC1₂, 13; glucose, 5; Hepes, 10 (pH adjusted to 7·5 by titration with NaOH). To block the chemical synapses in the ganglia, the ganglia compartment was perfused with Ca²⁺-free solution of the following composition (in mmol l⁻¹): NaCl, 38; KC1, 3·3; MgCl₂, 39; glucose, 5; Hepes, 10 (pH adjusted to 7·5 by titration with NaOH). In this solution, [Mg²⁺] was increased to three times that in normal solution and [NaCl] was decreased to adjust the osmolarity.

All experiments were performed at room temperature (20–25°C).

The cell bodies of neurones which were identified as members of the heart regulatory network in the ganglia were typically located as shown in Fig. IB. All neurones except VG1 have been described previously (Takeuchi, Yokoi, Mori & Kohsaka, 1975; Ku & Takeuchi, 1983; Ku, Isobe & Takeuchi, 1985; Boyles & Takeuchi, 1985; Matsuoka, Goto, Watanabe & Takeuchi, 1986). Each neurone except VG1 could be identified easily by its cell size, location and colour. The location of VG1 is rather variable from preparation to preparation and there are two other similar-sized cells near VG 1. Of these three cells, however, VG 1 is the only one which has its axon in the intestinal nerve. Thus, VG1 could be identified by the demonstration of antidromic action potentials in response to stimulation of the intestinal nerve.

PON can be identified as a bursting neurone in the isolated ganglia preparation (Takeuchi et al. 1975). However, in the present, more intact, preparation, PON usually showed irregular firing and received numerous inhibitory inputs (Fig. 2A).

When the intestinal nerve was cut and the ganglia were isolated from the heart, PON began to show periodic bursting (Fig. 2B). This suggests that the bursting activity of PON is usually depressed by inhibitory inputs originating from the heart.

The spontaneous activity of PON was correlated with an increase in peak tension and in beating frequency of the heart. When PON was inhibited, the heart beat was also depressed (Fig. 3A). If PON was driven to fire by depolarizing current injection, the heart rate and beating amplitude increased dramatically (Fig. 3B). The heart excitatory action of PON was so strong that only 5–6 action potentials at a frequency of 1 Hz were enough to produce positive inotropic and chronotropic actions. Usually, 10–20 spikes at a frequency of 1–2 Hz produced 50–70% increase in beat amplitude and 30–70% increase in heart rate. These results suggest that PON is one of the heart excitatory neurones.

Recent morphological investigations (Goto, Ku & Takeuchi, 1986; Matsuoka et al. 1986) and our own observations (data not shown) show that PON, TAN, TAN-2 and TAN-3, but not VIN, have axons in the intestinal nerve, and PON has multiple axons only in this nerve. Unfortunately, however, the termini of these neurones could not be stained, because the distance from the suboesophageal ganglia to the heart is rather great (3–4 cm). Thus, electrophysiological techniques were used to examine whether the axons in the intestinal nerve go to the heart.

When the intestinal nerve was stimulated at the point just before entering the pericardium (see Fig. 1A), an antidromic action potential was recorded at the soma of PON (Fig. 4Ai). If the somatic membrane was hyperpolarized, separation of the axonal spike from the somatic spike was seen as a hump on the rising phase of the antidromic action potential (Fig. 4Aii). The somatic spike was inhibited by further hyperpolarization (Fig. 4Aiii) and the invading axonal spike was also depressed (Fig. 4Aiv). EPSP-like potentials were not seen. Fig. 4B shows the simultaneous measurement of PON membrane potential and extracellularly recorded spikes of the intestinal nerve at the point just before it enters the pericardium. Action potentials of PON were correlated with the largest spikes in the intestinal nerve (see Fig. 4Bii). The conduction velocity of PON action potentials was about 25 cms⁻¹. These results indicate that the axon of PON extends to the heart.

To investigate whether the heart excitatory actions of PON were produced by the actions of PON upon another neurone, the excitation was examined before and after blockage of chemical synapses by perfusion of the ganglia with Ca²⁺-free solution. This treatment was seen to block synapses, because the large inhibitory input received by PON when the cerebropleural connective was stimulated (Fig. 5Aii) was almost completely absent in Ca²⁺-free solution (Fig. 5Bii). However, the heart excitatory action of PON was not blocked by this treatment (Fig. 5Bi). Consequently, PON is considered to be a heart excitatory motoneurone.

TAN, TAN-2 and TAN-3 fire tonically at a frequency of 0·5–2 Hz. They have similar sensitivities to several neurotransmitter candidates, and have similar morphology (Matsuoka et al. 1986). These neurones have axons in several nerve bundles, including the intestinal nerve, and were found to have an excitatory effect upon the heart. The effect was less pronounced than that of PON. When the spontaneous activity of these neurones was stopped by hyperpolarizing current injection, the heart beat was decreased to 70–90% (Fig. 6A). When the firing rate was increased, heart activity was increased (Fig. 6B); a burst of spikes (20 s) produced 20–30 % increase in beat amplitude and heart rate. In a given preparation, one of these neurones (usually TAN) generally showed a stronger action than the other two.

Stimulation of the branch of the intestinal nerve that entered the pericardium produced antidromic action potentials in TAN, TAN-2 and TAN-3 (Fig. 6C), indicating that these neurones had axons in this branch. The conduction velocity of these spikes was less than that of PON, about 15cms⁻¹.

Fig. 6D shows that the heart excitatory action of TAN was not blocked by perfusion of the ganglia with Ca²⁺-free solution.

Although many features of these three neurones were quite similar, their spontaneous activity was not usually coordinated. To examine whether any coupling exists among these cells, two cells were impaled simultaneously and the effect of current injection in one cell on the other was examined. Although not very strong, there were electrical couplings among these neurones (e.g. Fig. 6E).

Two cerebral ganglion cells, d-RCDN and d-LCDN, are symmetrically situated on the dorsal surface of the ganglia (Ku et al. 1985). As shown in Fig. 7, these neurones send their axons to the contralateral ganglion, where they bifurcate near the origin of the cerebropleural connectives. The two branches go into the contralateral cerebropleural connectives. No axons of these cerebral cells could be traced in the suboesophageal ganglia. The absence of these processes might be due to the failure of the dye to penetrate the more distal processes of the neurones.

These neurones were usually silent or firing at low frequencies (<0·3 Hz). A burst of spikes (less than 20 s) in these cells produced a 20–40 % increase in beat amplitude and heart rate. As shown in Fig. 8 for d-RCDN, the excitation of the heart produced when these neurones were made to fire (Fig. 8A) was not blocked when the ganglia were perfused with Ca²⁺-free solution.

Stimulation of the intestinal nerve entering the pericardium produced antidromic action potentials in d-RCDN and d-LCDN (Fig. 8B). EPSP-like potentials were not seen following stimulation of the intestinal nerve when the cell membrane was hyperpolarized. The conduction velocities of the antidromic action potentials were comparable to that of TAN. There is some doubt about the axonal projections of these cerebral neurones. Out of more than 40 d-RCDN and d-LCDN examined, 40 % of these neurones did not show antidromic action potentials following stimulation of the intestinal nerve. This does not seem to be the result of the poor state of the preparations, as in a given preparation the loss of antidromic action potentials was seen in d-RCDN but not in d-LCDN, and vice versa. Also, in some cases, the antidromic action potential was recorded following stimulation of the proximal portion of the intestinal nerve, but not recorded following stimulation of the branch going to the heart.

Although these bilaterally symmetrical neurones did not necessarily show coordinated firing patterns, it was found that they received common inhibitory (Fig. 8C) and excitatory (Fig. 8D) inputs. At present, the source of these inputs is not known, but they originate centrally as they were also seen in the isolated brain preparation. There was no electrical coupling between these neurones.

VG1 has not previously been described. As shown in Fig. 9, the soma lies near the middle of the dorsal surface of the visceral ganglion. This neurone has two main axons and some thin collaterals which go into the intestinal nerve. There are also other thin collaterals going to the right posterior pallial nerve. Whether all axons in the intestinal nerve go to the heart is not known at present.

The action of this neurone on the heart activity was somewhat variable. A burst of spikes produced in this cell by current injection usually induced a slight increase in peak tension of the heart beat but not an increase in heart rate (Fig. lOAi). When the firing frequency of this neurone was increased by larger current injections, the heart rate was decreased (Fig. 10Aii,iii). In a few preparations, only an increase in beat amplitude (Fig. 10B) or an inhibition of heart rate and decrease in beat amplitude (Fig. IOC) were recorded. This neurone had an axon in the branch of the intestinal nerve going to the pericardium, as demonstrated by the antidromic action potential (Fig. 10D). No EPSP-like potential was seen following stimulation of the intestinal nerve when the membrane of this cell was hyperpolarized. Conduction velocity of this spike was much faster than that of the other neurones described, about 40 cms^{− 1}.

In the present study, seven neurones (PON, TAN, TAN-2, TAN-3, d-RCDN, d-LCDN and VG1) were identified as being concerned with regulation of the heart of the African giant snail, Achatina fulica. Among these neurones, six cells had an excitatory effect on the heart.

PON, previously described as RPal (S.-Rózsa, 19796), is a bursting neurone in the isolated ganglia preparation. From its position in the ganglia, and its bursting characteristics, it is considered to be a homologous neurone to R15 in Aplysia, Fl in Helix aspersa, RPal in Helix pomatia and Cell 11 in Otala lactea (Chase & Goodman, 1977; Rittenhouse & Price, 1985; Kai-Kai & Kerkut, 1979; S.-Rózsa, 1979a ; Gainer, 1972). All these neurones have axons in the nerve bundle which goes to the heart. Helix Fl contains a cardioactive peptide whose structure has been reported (Price et al. 1985). This peptide as well as the homogenate of isolated Fl shows excitatory actions on the heart (Cottrell, Price & Greenberg, 1981 ; Price et al. 1985). However, the effects of stimulation of Aplysia R15 or Helix Fl on heart activity have not been reported, and negative results have been reported in the cases of Helix RPal, Otala cell 11 and Achatina RPal (S.-Rózsa, Salánki & Sakharov, 1983; Gainer, 1972; S.-Rózsa, 1979b). S.-Rózsa (1979b) reported that Achatina RPal (i.e. PON) received inputs from the heart but was not involved in heart regulation. Although the difference between the present results and those reported for other snails may reflect the difference in species, the inconsistency between our results and those of S.-Rózsa is not easily reconciled because PON has the strongest excitatory action on the heart among the identified heart excitors in the present experiments. Indeed, its action was so strong that frequent activity of this cell following injury masked the excitatory action of the other heart excitors. One possible reason may be the difference of experimental conditions. In S.-Rózsa’s experiments, the pericardium was not removed and the heart activity was monitored by a photo-optic method. In our experience, if the pericardium was left intact, the heart of Achatina did not show regular beating after it had been dissected out from the animal. Under this condition, it was very difficult to examine the action of PON on the heart.

In the present experiments, PON showed irregular firing interposed by many inhibitory inputs. Although direct presynaptic inhibitory neurones were not identified, these inhibitory inputs were suggested to originate from the heart activity. The bursting activities of Helix RPal and Aplysia R15 are also reported to be influenced by synaptic inputs (Salánki, S.-Rózsa & Vadász, 1979; Koester et al. 1974).

TAN, TAN-2 and TAN-3 are tonically firing neurones and have several output axons in many nerve bundles arising from the suboesophageal ganglia (Matsuoka et al. 1986). This indicates that they may have multiple functions. One clear function is the modulation of heart beat. The tonic heart excitatory action of these neurones is comparable to that of RB_HE in Aplysia (Mayeri et al. 1974).

In Aplysia, identified heart regulatory motoneurones are located in a rather restricted region of the abdominal ganglion (Mayeri et al. 1974). However, in Achatina, two cerebral ganglion cells, d-RCDN and d-LCDN, were shown to have axons which go to the heart, although the occurrence of axons in the heart nerve was variable. When these axons were present, the heart excitatory action of these neurones was always unaffected by perfusion of the ganglia with Ca²⁺-free solution. This result suggests that the action of these cerebral neurones is not dependent on other neurones. Thus, at least in some preparations, d-RCDN and d-LCDN may also act as heart excitors like PON, TAN, TAN-2 and TAN-3. In our preparation, these two cerebral cells did not necessarily show coordinated activity. However, as common excitatory and inhibitory inputs were seen, these cells may be, in some cases, driven to fire at the same time in intact animals.

There is no obvious reason for the variable occurrence of the d-RCDN and d-LCDN axon in the heart nerve. There may be a seasonal variation because most of the experiments were carried out during winter to early spring (1985– 1986). In its natural habitat (Okinawa, Japan), this snail hibernates during this period. The hibernation would produce a decrease of body fluid and this may induce the sprouting or retraction of axons as indicated in Helisoma neurones (Bulloch, 1984; Maetzold & Bulloch, 1986). Further investigations are needed to clarify this problem.

The four heart excitatory neurones reported for Helix pomatia (S.-Rózsa, 1981) are not considered to be homologous to any heart excitors described in this paper. The heart inhibitory motoneurones, V12 and V13, the heart regulatory neurone, V21, and RPal of Helix pomatia (S.-Rózsa & Salánki, 1973; S.-Rózsa, 1979a) have been considered to have homologues in Achatina (S.-Rózsa, 19796). In the present experiments, homologous neurones to V12 and VI3 were not found. This discrepancy may arise from the present preparation which destroyed the small nerve branch going to the aorta.

The newly identified VG1 may be homologous to V21 of Helix pomatia which also has an axon in the intestinal nerve and whose tonic activity arrested the heart beat (S.-Rózsa, 1979a,b), although VG1 usually increased beat amplitude at low firing frequency (Fig. 10). Since the actions of VG1 were variable, this neurone is not considered to be a motoneurone. Again, the contrast between the present observation and the earlier one may reflect the difference of experimental conditions. In the earlier experiments, heart activity was monitored by a photo-optic method without dissecting the pericardium; in this condition, the tension imposed on the heart is considered to be lower than that in the present experimental condition. In our experience, it seems that the more the tension of the heart was increased, the less explicit the heart inhibitory action of VG1 became.

Direct inhibitory neurones were not identified in the present experiments. However, spontaneous depression of the heart activity was frequently observed in our preparation, which was not usually seen in the isolated heart preparation. Moreover, the high firing frequency of VG1 induced depression of the heart activity in most cases. So, the failure to identify inhibitory neurones is not considered to be due to the absence of the small nerve branch going to the aorta.

In this paper, the properties of seven identified heart regulatory neurones have been described. The accompanying paper (Furukawa & Kobayashi, 1987) examines the interconnections among these neurones and also gives an account of the other central neurones which influence their activity.

The authors thank Mr T. Shirane for his help in the morphological examination of the neurones.