Potassium Movements in A Central Nervous Ganglion of Limnaea Stagnalis (L.) (Gastropoda : Pulmonata)

The processes and pathways involved in the exchange of inorganic ions and small water-soluble molecules between the fluid bathing the surface of the central nervous system and the extra-neuronal fluid have been studied in few invertebrate species (cf. Treherne & Moreton, 1970). Detailed investigations have been performed on ventral nerve cord preparations from the central nervous systems of the leech Hirudo medicinalis (Nicholls & Kuffler, 1964) and the insect Periplaneta americana (Treherne et al. 1970). In these studies membrane potentials of resting and active nerve are used to monitor the concentrations of potassium and sodium in the immediate extraneuronal fluid. Changes in the composition of this fluid can then be studied following changes in the concentrations of these ions in the bathing medium. Experiments of this type have been carried out on preparations that have undergone varying degrees of microsurgery enabling some assessment to be made of the potential sites of restriction of ionic movements.

In a central nervous preparation of Hirudo medicinalis which has undergone no further dissection other than that necessary for isolation, resting potentials and action potentials rapidly reflect changes in the ionic composition of the bathing medium (Nicholls & Kuffler, 1964). For example, 50% of the sodium at the neuronal surface is exchanged in less than 12 sec when either sucrose or choline replace this cation in the Ringer solution perfusing the experimental chamber. During the movements of sodium sucrose and choline through the leech central nervous system the membrane potentials of glial cells do not change by more than a few millivolts and the intracellular potassium concentration remains constant. Also, cooling the preparation from 23 °C to between 1 and 4 °C produces only a small delay in the time courses for the exchange of sodium and sucrose. Such findings strongly suggest that these substances do not pass from the external solution to the neuronal surface via the glial cytoplasm. A comparison by these authors of the observed half-times for ion exchanges and the calculated half-times assuming diffusion along the system of narrow intercellular clefts clearly demonstrates that these channels can provide a rapid and effective pathway for the movements of ions and small water-soluble molecules. Microelectrodes in the extracellular spaces of the isolated (intact) preparation record a potential indistinguishable from that obtained with the electrode tip in the bathing medium, confirming the absence of any regulation of the extra-neuronal ionic microenvironment.

This contrasts with the situation observed in similarly isolated intact preparation from the central nervous system of Periplaneta americana where microelectrodes in the extracellular spaces record a potential some 10–15 mV positive with respect to the bathing medium (Pichon & Boistel, 1967). Electrophysiological evidence of this nature supports the view expressed earlier (cf. Treheme, 1966) that the fluid bathing the nerve cells of Periplaneta differs in ionic composition from the haemolymph of this insect. Recently, large extra-neuronal potential changes have been observed following elevation of the potassium concentration of the external medium (Treheme et al. 1970; Pichon & Treheme, 1970). The apparent depolarization recorded with intracellular microelectrodes under these conditions is not associated with any equivalent reduction in the amplitude of the action potentials. On the basis of these results it is suggested that there is appreciable restriction to the intercellular diffusion of potassium ions in intact preparations and that this is represented by the tight junctions at the inner margin of the intercellular clefts traversing the perineurium (cf. Treherne & Pichon, 1972). The extra-neuronal potential changes observed in this insect preparation appear to result from a depolarization of the outwardly directed perineurial membrane, access to the inwardly directed membrane being reduced by the presence of the tight junctions (Pichon, Moreon & Treherne, 1971). A considerable degree of control over the ionic environment of nerve cells has also been demonstrated in the central nervous systems of the phytophagous insects Carausius morosus (Treherne, 1967; Treherne & Maddrell, 1967) and Manduca sexta (Pichon, Sattelle & Lane, 1972). Such marked restriction upon ionic movements in the insect central nervous system constitutes an effective blood-brain barrier, a situation apparently unique amongst invertebrate animals (cf. Treherne & Moreton, 1970; Treherne & Pichon, 1972).

Few comparable studies are available for molluscan preparations. In the squid, for example, movement away from the giant axon surface of potassium ions extruded during activity can be followed by observing the rate of decline of the positive phase of the action potential during a train of impulses (Frankenhaeuser & Hodgkin, 1956). From such experiments it has been estimated that the space around the axon in which the potassium ions accumulate is about 300 Å wide, a figure comparable with the dimensions of the spaces observed between the axolemma and the investing glial cells by electron microscopy (Geren & Schmitt, 1954; Villegas & Villegas, 1960). Diffusion along such intercellular channels adequately accounts for the exchange of inorganic ions between the extra-axonal fluid and the bathing medium (Frankenhaeuser & Hodgkin, 1956). In the lamellibranch Anodonta cygnea the accessibility of the bulk of the axons of the cerebro-visceral connectives to inorganic ions, small water-soluble molecules and the exogenous tracer molecule macroperoxidase (M.W. 40000) has been demonstrated (Mellon & Treheme, 1969; Treherne, Mellon & Carlson, 1969; Lane & Treherne, 1972). Rapid exchanges of inorganic ions and small water-soluble molecules are also reported for the pleural-supraintestinal connective of the prosobranch gastropod Viviparus contectus (Sattelle, 1972, 1973 a; Sattelle & Lane, 1972). Ultrastructural studies on these species reveal the absence of any restriction of the intercellular spaces (Gupta, Mellon & Treherne, 1969; Sattelle & Lane, 1972) other than the macula adhaerens desmosomal junction between glial cells of Anodonta. This type of cell connexion is also found in epithelia, where it does not retard the movements of small ions and molecules through the intercellular clefts (Farquhar & Palade, 1963).

The rapid movements of such substances through the central nervous tissues of Anodonta and Viviparus can consequently be accounted for in terms of diffusion through the extracellular system.

Recently, the giant nerve cell bodies of pulmonate gastropods have received a good deal of attention as suitable preparations for the investigation of the ionic basis of excitation (cf. Kostyuk, 1968; Treherne & Moreton, 1970). Most of these studies utilize exposed neurones from ganglia which have undergone considerable surgery and therefore provide little information concerning either the routes or mechanisms for the exchange of ions and small molecules through the central nervous tissues of these animals. The work reported here complements a previous investigation on a central nervous ganglion of Limnaea stagnalis in which ultrastructural and preliminary electrophysiological evidence has been obtained suggesting that a rapid exchange of potassium ions takes place between the blood (bathing medium) and the extraneuronal fluid (Sattelle & Lane, 1972). These findings are consistent with the observation of Pentreath & Cottrell (1970) that the nerve sheath of Helix pomatia is permeable to ferritin. Also, it has been clearly demonstrated by Moreton (1972) that in de-sheathed ganglia of Helix aspersa potassium exchange takes place via the extracellular channels and is not restricted by any discrete diffusion barriers. This author has shown in addition that, with the inner of the two ganglionic sheaths intact, the diffusion time is increased and that this is attributable to an increased diffusion path length rather than to any specific barrier effect of the sheath. Such findings conflict with the views, based on earlier ultrastructural studies on the central nervous systems of Helix aspersa and Cryptomphallus aspersa, that the gastropod nerve sheath is either involved in the active transport of ions (Fernandez, 1966) or acts as a barrier to diffusion (Rogers, 1969; Sanchis & Zambrano, 1969). This paper reports a study on the movements of potassium ions in a central nervous ganglion of Limnaea stagnalis. The way in which the resting potential of de-sheathed (exposed) giant neurones of this pulmonate gastropod varies with the external potassium concentration is known (Sattelle & Lane, 1972). In this investigation the resting potential is employed to monitor the potassium concentration at the cell surface. With this technique an attempt is made to determine, first, whether or not the ionic composition of the extra-neuronal fluid resembles that of the bathing medium and, secondly, the processes and.pathways by which potassium ions move through a central nervous ganglion of Limnaea stagnalis.

Specimens of Limnaea stagnalis, obtained from L. Haig and Co. Ltd, were maintained in large glass aquaria under continuous flow conditions (cf. van der Steen, van den Hoven & Jager, 1969) until required for experimentation. Experiments were performed using either giant cell bodies of neurones from the right parietal ganglion, or isolated lengths of the osphradial nerve. For extracellular recordings the latter preparation was used in the apparatus described for use with connectives of Viviparus contectus (Sattelle, 1972). For intracellular studies the circumoesophageal ring of ganglia was dissected and removed (except for in situ preparations discussed below) from an unanaesthetized animal under Ringer and secured by inserting through the centre of the ring a nylon tube, the ends of which were held against a small glass plate by rubber bands. Such preparations were transferred to a perspex experimental chamber (see Fig. 2) which was perfused at a rate of 15 ml/min by a system of gravity feed saline reservoirs. A multiway non-return valve was situated close to the experimental chamber facilitating a rapid change to any of seven test solutions whilst maintaining a very small dead-space. Details of the construction and performance of this device have been described elsewhere (Holder & Sattelle, 1972). All experiments were performed under conditions of continuous perfusion of saline, and experiments with coloured dyes indicated that, following a change of test solution, the fluid in the experimental chamber was completely replaced in less than 7 sec. A reduction in temperature of the test solution from 23 to 8 °C was achieved by the application of an ice-filled polystyrene sleeve to the saline reservoir.

Glass microelectrodes for intracellular recording were drawn from cleaned, grease-free tubing on a vertical micropipette puller and filled by the method of boiling under reduced pressure in methanol at 40 °C for 8–10 min until air was eliminated. Thirty minutes of boiling under reduced pressure in distilled water at the same temperature was followed by soaking the electrodes overnight in 3 M potassium chloride after the method of Tasaki, Polley & Orrego (1954). Electrodes prepared in this way with resistance values of 10–30 MΩ were normally used for recording and those with tip potentials of 5 mV or more were rejected (cf. Adrian, 1956). An electrode-mounting device was constructed, based on a design by Moreton (1967), enabling simultaneous penetration of a cell by two electrodes. Membrane potentials obtained with such electrodes were led via a unity-gain, high-impedance amplifier to a Tektronix 502 or 561 oscilloscope and filmed on either a Cossor or a Nihon-Kohden oscilloscope camera. Ringer and potassium chloride bridges linked the preparation to an indifferent electrode consisting of a silver-silver chloride plate. The Ringer compartment, interposed between the recording chamber and the chamber containing 3 M potassium chloride (see Fig. 1), served to reduce junction potentials (Moreton, 1968). A bridge circuit was used during the application of stimuli via the recording electrode (Araki & Otani,1955). Stimuli consisted of rectangular current pulses applied via an R.F. isolation unit of low output impedance.

Electrodes were lowered, under visual control, to the surface of the ganglion or desheathed cell. Tapping the base of the micromanipulator, to which the electrode mounting device was rigidly attached, effected impalement of neurones. As a precaution against poor impalements and records from damaged cells experiments were not started until 20 min after the neurones had been impaled, during which time resting potential and activity patterns were followed. Three classes of preparation were used throughout these studies.

A few experiments were performed on giant neurones from animals in which the body wall had been cut and the organs reflected to expose the circum-oesophageal ring, the peripheral nerves and connectives remaining intact. Intracellular impalements were only maintained for short periods because of muscular contractions.

Many experiments were carried out using the isolated circumoesophageal ring of ganglia, mounted without further dissection. Peripheral nerves were severed as far from the ganglia as possible. Impalement of neurones was achieved through the ganglionic sheath without difficulty.

Experiments were also conducted on neurones from which the overlying ganglionic sheath had been removed. This was achieved using electrolytically sharpened tungsten needles. In addition to the sheath, a portion of the underlying glial layer was undoubtedly removed, though the extent of glial damage was not estimated.

The normal Ringer solution employed was based upon the recent analyses of the blood ionic composition (van der Borght, 1962; Chaisemartin, Mouzat & Sourie, 1967; Greenaway, 1970) and consists of: 50·0 mm/1 NaCl; 1·6mm/l KC1; 2·0mm/l MgCl₂; 4·0 mm/1 CaCl₂; pH 7·4. High-potassium Ringer was made by increasing the potassium chloride concentration at the expense of sodium chloride, to maintain osmolarity. Low-potassium Ringer was produced by simply reducing the concentration of potassium chloride.

Resting potentials and action potentials recorded from in situ, intact and de-sheathed preparations (see above) have been summarized in Fig. 2. Neurones in intact preparations exhibited only slightly higher values of resting potential and action potential overshoot than de-sheathed cells. This may have been due to damage incurred during the de-sheathing process. Membrane potentials from the few cells investigated in situ closely resembled those obtained from both intact and de-sheathed neurones. There were in fact no large discrepancies in values of resting potential and action-potential overshoot recorded from the three preparations studied, resembling the situation in the leech, Hirudo medicinalis (Nicholls & Kuffler, 1964). The shape of the action potential (cf. Fig. 2) also differed little between the three preparations, though some variation was observed in the shape of the action potentials recorded from different cells of the right parietal ganglion.

Resting potentials of intact neurones were recorded for different values of external potassium concentration, [K⁺]₀, over the range 0·5–50·0 mM/1. Although the neuronal membrane was relatively insensitive to changes in [K⁺]₀ at concentrations in the physiological range, at higher concentrations of potassium the resting potential showed a decline which could be represented as a straight line on a logarithmic scale (Fig. 3). Over the straight-line section of the graph a tenfold change in [K+]₀ corresponded to a 34 mV potential change for intact neurones. In de-sheathed cells the resting potential followed a similar relation to [K⁺]₀ but potentials throughout were a few millivolts lower.

In all these experiments 5 min were allowed for the membrane potential to equilibrate to a new external potassium concentration. The upper limit of potassium concentrations used was determined by the ability of the membrane potential to return to its normal resting level following depolarization. The consistent discrepancy of a few millivolts between the two curves was probably attributable to the process of de-sheathing.

An increase in the external potassium concentration bathing an intact ganglion from 1·6 mm/1 (Ringer concentration) to 10·0 mm/1 resulted in the onset of a rapid potential change in the direction of depolarization^* (Fig. 4). In depolarizations on five intact cells equilibrium was reached within 2-3 min and the mean half-time for depolarization was 34·2 (S.E. ± 5·9) sec. Recovery from exposure to high-potassium Ringer was somewhat slower and the effects of repeating such depolarization/repolarization experiments were investigated. The time-course in both cases was unaltered provided that time was allowed for equilibration to the original resting potential between experiments (Fig. 4). Depolarizations were also followed for five de-sheathed neurones during an elevation of the external potassium concentration from 1·6 to 10 mM/1. Desheathing clearly decreased the time required to reach equilibrium. The half-time was reduced from 34·2 (S.E. ± 5·9) sec (intact) to 10·0 (S.E. ± 2·8) sec (de-sheathed).

To check the possibility that the microelectrode was creating an artificial, shorter diffusion pathway through the central nervous system and thereby linking the extraneuronal fluid directly with the bathing medium, extracellular recordings were made of compound action potentials from the isolated osphradial nerve of Limnaea. Within 1 min of exposure to high-potassium (25·0 mM/I K⁺) Ringer, conduction block was almost complete, indicating a rapid exchange of this cation between the bathing medium and the axonal surface in this peripheral nerve (Fig. 5). The rate of potassium depolarization in peripheral nerve, determined by a technique which did not involve puncturing the nerve sheath, appeared to closely resemble that obtained using intracellular electrode recordings from giant neurones of an intact central nervous ganglion. Such results suggested that the resting potentials recorded in intact ganglionic preparations were monitoring extraneuronal concentrations of potassium ions.

The studies so far have not eliminated the possibility that an active mechanism, situated for example in the glia, operates in the exchange of ions between the blood (bathing medium) and the neuronal surface. Such a system would be profoundly affected by changes in temperature. A drop in temperature from 23 to 8 °C, although causing some retardation, did not greatly affect the time course of depolarization and repolarization in an intact neurone (Fig. 6).

The potential profile was continuously recorded during the impalement of an intact neurone (Fig. 7). At (A) the microelectrode measured the zero potential in the saline. As the electrode depressed the ganglionic sheath the potential fell to the level at (B). By tapping the base of the micromanipulator, the electrode was forced through the sheath, presumably into an extracellular position, and the potential returned to (C) which approximated to the zero potential. Further tapping of the micromanipulator resulted in impalement of the cell body (D). The initial resting potentials obtained from all cells were rather low (25–35 mV) but increased steadily after impalement (Fig. 7) and consistent resting potentials and activity patterns were usually observed within 5–10 min.

Using the method applied to axons in the central nervous system of the insect Periplaneta americana (Treheme et al. 1970) and the giant neurones in a central nervous ganglion of the mollusc Helix aspersa (Moreton, 1972), it was possible to follow the changes of potassium concentration at the surface of an intact neurone during elevation of [K⁺]₀. Combining data from Fig. 3 and data from depolarization experiments, it appeared that the potassium concentration at the cell surface varied exponentially with time (Fig. 8). This indicates that potassium movements take place by a first-order diffusion process in a system with no discrete diffusion barriers.

Resting potentials and action potentials recorded from in situ, intact and de-sheathed neurones from the right parietal ganglion of Limnaea stagnalis are of similar magnitude (cf. Fig. 2). Experiments of this nature on ganglia of the leech Hirudo medicinalis have also indicated that removal of the nerve sheath and part of the glial cytoplasm does not appreciably affect the membrane potentials recorded from the underlying neurones (Nicholls & Kuffler, 1964). In the nerve cord of the insect Periplaneta americana, however, resting potentials are several millivolts lower in intact than in de-sheathed preparations and action potentials are several millivolts higher in intact than in de-sheathed preparations (Pichon & Boistel, 1967). Potential profiles have been continuously recorded during the impalement of neurones in intact ganglia of Limnaea. These reveal that the potential obtained with the microelectrode tip in an apparently extracellular position is very similar to the zero potential recorded in the bathing medium. A positive sheath potential, corresponding to that observed in the intact nerve cord of Periplaneta, has not been demonstrated in the central nervous system of Limnaea. These findings suggest that the ionic composition of the extraneuronal fluid in the right parietal ganglion of Limnaea closely resembles that of the blood or bathing medium. This is comparable with the situation reported for the central nervous system of Hirudo (Nicholls & Kuffler, 1964) but contrasts with the demonstration of a degree of regulation of the ionic composition of the extra-axonal fluid in the nerve cord of Periplaneta (cf. Treherne & Pichon, 1972).

A 34 mV decade potassium slope has been obtained for the resting potential of de-sheathed neurones of Limnaea (Sattelle & Lane, 1972). This departs considerably from the behaviour of an ideal potassium electrode for which a 58 mV decade potassium slope could be predicted by the Nernst equation. An attempt is made else where to account for this departure from the ideal situation (Sattelle, 1973 b). It is of interest to note, however, that a 34 mV decade potassium slope is also obtained from intact neurones. Resting potentials of de-sheathed neurones are consistently 2-3 mV lower than those of intact cells, which can probably be attributed to the process of de-sheathing. It is nevertheless clear that, in intact preparations, changes in the potassium concentration in the extra-neuronal fluid follow changes in the bathing medium. Experiments with elevated potassium concentrations indicate that changes in the concentration of this ion in the bathing medium are rapidly transmitted to the extraneuronal fluid, as reflected by the rapid changes in resting potential. The time-course of depolarization of the neurones is only slightly affected by the de-sheathing process, the half-time being reduced from 34-2 (S.E. ± 5·9) sec for intact cells to 10·0 (S.E. ± 2·8) sec for de-sheathed cells. These results establish that the tissues removed by desheathing, namely the nerve sheath and an unknown proportion of the underlying glia, do not act as a major barrier to the movements of potassium ions in this central nervous ganglion of Limnaea. It has been noted that repolarization of the neuronal membrane following exposure to high-potassium Ringer and a subsequent return to normal Ringer takes somewhat longer than the initial depolarization. An explanation for this is provided in a discussion of similar findings for neurones of Helix aspersa (Moreton, 1972). This author suggests that the asymmetry may best be accounted for by the fact that the concentrations at which depolarization/repolarization experiments are initiated or completed refer to different sections of the potential/concentration curve and hence represent points of different membrane sensitivity to external potassium. The observed differences in the rates of these potential changes can therefore be regarded as a property of the experimental procedure.

It is possible that an active process, situated for example in the glia, is involved in potassium exchange, which may account foi the rapid changes in the potassium concentration of the extra-neuronal fluid following the exposure of intact preparations to a high-potassium saline. Since it has been shown that depolarization takes place faster in de-sheathed than in intact neurones, such a view demands that the de-sheathing procedure, although undoubtedly damaging the underlying glia, does not affect thq active mechanism. The rate of potassium movement by such an intracellular route would be profoundly affected by changes in temperature, but a reduction in temperature from 23 to 8 °C only slightly increases the time taken for the depolarization and repolarization of an intact neurone. In the apparent absence of any active mechanism for the exchange of potassium between the bathing medium and the extra-neuronal fluid it seems likely that extracellular channels represent the pathway for potassium movements. In this context it is relevant to note that examination of the extracellular spaces observed in the right parietal ganglion of Limnaea stagnalis reveals the absence of any restrictions such as intercellular tight junctions which are likely to hinder the passage of ions (Sattelle & Lane, 1972). Also, changes in potassium concentration at the surface of intact cells, in response to an elevation of the potassium concentration of the bathing medium, follow an exponential time-course (cf. Fig. 8). It appears therefore that potassium movements can be accounted for in terms of a first-order diffusion process. These findings point to the unrestricted extracellular channels as the pathway for potassium exchange in this central nervous ganglion of Limnaea. This being so, the reduction in the half-time for depolarization in de-sheathed compared to intact cells is readily explained in terms of a shortening of the extracellular diffusion pathway in the de-sheathed preparations.

It is established for the right parietal ganglion of Limnaea stagnalis that the extraneuronal fluid and the external bathing medium (blood) are of similar ionic composition, and it appears that exchanges of potassium ions between these two fluid compartments take place by diffusion along the extracellular channels. The results presented here accord well with those reported for the central nervous system of the leech Hirudo medicinalis (Nicholls & Kuffler, 1964) and the gastropod mollusc Helix aspersa (Moreton, 1972).

I am indebted to Dr J. E. Treherne for his advice and constructive criticism throughout the course of this work. I also thank Dr R. B. Moreton for helpful discussions and for permission to refer to his unpublished observations.