The Distribution and Exchange of Inorganic Ions in the Central Nervous System of the Stick Insect Carausius Morosus

The ionic composition of the haemolymph of some groups of largely phytophagous insects is known to be very different from that of the body fluids of most of the other animals which have been studied (Boné, 1944, 1947; Duchâteau, Florkin & Leclercq, 1953 ; Sutcliffe, 1963). In these insects the haemolymph is characterized by very low concentrations of sodium ions relative to those of potassium and magnesium. Such an ionic balance presents certain difficulties in the interpretation of the processes involved in nervous transmission in these insects, for in all other animals which have been studied the propagation of nerve impulses has been shown to depend upon high external concentrations of sodium ions (cf. Hodgkin, 1951, 1958; Shanes, 1958,a, b). Thus, for example, a solution approximating to the ionic composition of the haemolymph of Carausius resulted in a rapid loss of conduction when applied to the abdominal nerve cord of an insect such as Periplaneta americana (Treherne, 1965).

Previous studies on the exchanges of ions and molecules in the insect central nervous system have been carried out using the cockroach, which has a haemolymph containing a normal balance of the common cations (Treherne, 1961 a, b, c, d, 1962). In these investigations it was shown that the ionic exchanges between the haemolymph and the central nervous tissues occurred relatively rapidly. These exchanges took the form of a rapid movement between the haemolymph and the extracellular fluid, together with slower movements of ions between the extracellular fluid and the cells of the central nervous system. The high level of cations, relative to those of chloride, in the rapidly exchanging extracellular fraction appeared to be governed very largely by a Donnan equilibrium with the haemolymph.

Preliminary experiments on the effects of cations on conduction processes in the nerve cord of Carausius morosus showed that, despite their very low concentration in the haemolymph, the presence of sodium ions was essential for the propagation of nerve impulses in this preparation (Treherne, 1965). The presence of magnesium was also essential, for there was a rapid loss of conduction when the nerve cord was bathed in solutions containing reduced concentrations of this cation. Transmission of impulses continued when substantial portions of the nerve sheath were removed from abdominal ganglia, which suggested that the function of the axons in this insect did not depend upon the presence of a peripheral diffusion barrier. These results taken with the earlier work on the nerve cord of Periplaneta throw very little light on the ionic regulation or on the possible mechanisms involved in nervous conduction in the central nervous system of an insect like Carausius. The present investigation was therefore undertaken in an attempt to elucidate some of the problems encountered in the neurophysiology of species with such abnormal balances of ions in the haemolymph.

Many of the techniques used in this investigation were essentially similar to those employed in previous investigations on the central nervous system of Periplaneta americana J.,. (Treherne, 1961 a, d, 1962). Detailed descriptions will, therefore, only be given for techniques which differ from those used in the previous published accounts.

The physiological solution used in this investigation was that devised by Wood (1957) and had the following ionic composition: Na, 15·0 mM./l.; K, 18·0 mM./l.; Ca, 7·5 mM./l.; Mg, 50·0 mM./l.; H₂PO₄, 6·0 mM./l.; HPO₄, 4·5 mM./l.; Cl, 133 mM/1. The pH of this solution was approximately 6-6. The osmotic concentration of the fluid was maintained by the addition of disaccharide, in this case 69-9 g. trehalose/1. in place of the sucrose used by Wood.

The concentrations of sodium, potassium, calcium and magnesium in the nerve cord tissues and in the haemolymph were measured using the Unicam SP 900 flame spectrophotometer. The tissue samples were previously ashed in a muffle furnace on pieces of platinum foil at temperatures of 460–480° C. The direct flame photometric determination of inorganic ions in biological materials is sometimes complicated by the presence of other elements which have various interference effects. A standard check was, therefore, made to ensure that quantitative recoveries of added amounts were obtained for each particular ion (e.g. Fig. 1).

In these experiments the whole ventral nerve cord of Carausius was used instead of the abdominal nerve cord preparation employed in the previous investigations with Periplaneta (Treheme, 1961a, b, d, 1962). All experiments were carried out at a temperature of about 18·0° C.

The concentrations of the four main cations in the haemolymph and the tissues of the freshly dissected nerve cord are summarized in Table 1. In each case the water content of the tissues was measured, the individual values being used to calculate the concentrations of the ions in the tissue water.

The values for the ionic composition of the haemolymph are in approximate agreement with previous measurements which have been made with this insect (Boné, 1944; Duchâteau et al. 1953 ; Ramsay, 1955 ; Wood, 1957). The value for the sodium concentration is nearer to the figure of 21 mM./l. published by Boné than that of 9·0 mM./l. given by Duchâteau et al. The potassium concentration shown here is higher than in the previous investigations. These variations can presumably be attributed to dietary and genetic differences in the various stocks used.

These measurements were made using ¹⁴C-labelled inulin molecules. The radio-activity associated with the nerve cord tissues was compared as between excised nerve cords exposed to a 1·0% ¹⁴C-inulin solution for approximately 1·0 sec. and for 10 hr. respectively. These results were also compared with those obtained 3 hr. after the injection of ¹⁴C-inulin into the haemolymph (Table 2). It will be seen that in the in vitro experiments the radioactivity of nerve cords given the brief exposure to the solution was much less than those soaked for 1 hr. The figure of 12·5 ml./kg. tissue can be used as a measure of the surface contamination by the inulin molecules. The in vivo experiments showed an apparent inulin space of 104·3 ml./kg. tissue. Applying the figure for the brief exposure as a correction factor for the in vivo experiments this would correspond to an estimated extracellular inulin space of 91·8 ml./kg. tissue or 151·6 ml./l. of nerve cord water. It is apparent from the higher figure obtained with the in vitro preparation that some increase in extracellular space must have occurred as a result of this treatment.

Figure 2 illustrates the changes in specific activity which occur in the tissues of the central nervous system following the injection of 100μ. of physiological solution containing ²²Na. The experimental values approached complete exchange after a period of about 1·0 hr. The changes in the ratio of the specific activities in the nerve cord and the haemolymph occurred at a similar rate to those taking place in the cockroach central nervous system following the injection of ²⁴Na (Treherne, 1961 a).

In these experiments 100μl. of solution containing ²²Na was injected into each insect. After 4 hr. the whole ventral nerve cord was removed, carefully blotted on filter-paper and then washed in successive 0·2 ml. volumes of physiological solution. The radioactivities of these washings were measured in a liquid scintillation counter and the amount of radioisotope remaining in the nerve cord at varying times was calculated. Figure 3 is an example of the results obtained in this way. As in previous investigations on the central nervous system of Periplaneta this efflux can be represented as a two-stage process, an initial rapid component giving way to a second slow phase.

Table 3 summarizes the results of the experiments on the efflux of ²²Na from the nerve cord on washing in inactive physiological solution. It will be seen that approximately a third of the exchangeable sodium was contained in the rapidly exchanging fraction. This can be calculated to be equivalent to 19·5 mM./kg. tissue. These results also show that the rapid fraction in sodium efflux was not affected by the presence of 2,4-dinitrophenol, but that the slow escape of these ions from the nerve cord was significantly reduced in the presence of the poison. These results are similar to those obtained in previous investigations on Periplaneta in which the rapid fraction was identified as the extracellular ion fraction (Treherne, 1961 d, 1962). If the previously obtained figure of 91·8 ml./kg. tissue is used as a measure of the extracellular fluid in the nerve cord then it can be calculated that the sodium concentration of the rapidly exchanging fraction was 212·4 mM./l. This is roughly ten times greater than the measured sodium concentration of the haemolymph. The extracellular concentration also exceeds that of the cellular fraction, which on the basis of a slowly exchanging fraction of 69·5 % can be calculated to be in the region of 86·3 mM./l.

The above calculation on the concentrations of the rapidly exchanging sodium in the nerve cord is based on the assumption that the surface contamination does not contribute appreciably to this fraction. This assumption seems to be justified for the abdominal nerve cord of the cockroach (Treherne, 1962), but there is no evidence that this is the case for Carausius. An attempt was made to estimate the extent of the surface contamination by ²²Na in the nerve cord of this insect by comparing the amount of the isotope picked up during an approximately 1·0 sec. exposure with that contained in the rapid fraction after a prolonged exposure to the isotope. The results shown in Table 4 give a mean value of 1·04 mM./kg. nerve cord tissue as the amount of radioisotope associated with the cord following a brief exposure to the radioactive solution. This represents only about 5 % of the rapidly exchanging radiosodium so that it is reasonable to assume that the greater part of this fraction is contained within the nerve cord.

In these experiments the uptake of ²²Na by isolated nerve cords was compared as between those soaked in normal physiological solution and those exposed in addition to 0·5 mM./l, 2:4-dinitrophenol. The excised nerve cords were pre-treated for 5 min. in normal or poisoned non-radioactive solution. Figure 4 illustrates the specific activity changes of the sodium within the nerve cord relative to that in the external solution. Despite the considerable individual variation these results indicate that the rate of mixing of labelled sodium ions in the nerve cord tissues took place more slowly in the presence of the poison molecules.

In view of the variation experienced in the above experiments some direct measurements were made on the uptake of radiosodium by isolated nerve cords. For this purpose the excised nerve cords were pre-treated for 5 min. by soaking in non-radio-active solutions. In half these experiments this solution also contained 0·5 mM./l. 2,4-dinitrophenol. The nerve cords were then transferred to solutions in which the sodium was labelled with ²⁴Na. At varying intervals after this the cords were removed from the radioactive solution, quickly washed in non-radioactive solution, blotted and then their radioactivity was measured with a GM tube. The isolated cords were then returned to the radioactive solution before being assayed again. The effect of the metabolic inhibitor on this process is illustrated in Fig. 5. These results show that the effect of the poison is on the initial rapid uptake of the isotope, although the slope of the line for the subsequent uptake appears to be steeper in the normal as compared with the poisoned preparations.

The efflux of ⁴²K was studied in experiments in which the nerve cord was made radioactive by the injection into the haemolymph or by soaking isolated cords in a solution containing labelled potassium ions. Figure 6 shows that, as for sodium, the efflux of ⁴²K can be represented as a two-stage process. The results of the in vivo experiments showed some variation in the percentage of the rapidly exchanging potassium fraction (Table 6).

The above variation resulted partly from the low specific activity of the ⁴²K used in these experiments. The further experiments, summarized in Table 7, were, therefore, carried out by soaking the isolated nerve cords in solutions containing radioactive potassium, thus avoiding the dilution of the isotope obtained on injection into the haemolymph. In some experiments isolated nerve cords were soaked for 1·0 hr. and then effluxed in normal solution ; in the second batch the nerve cords were soaked for 1·0 hr. and then effluxed in a solution containing 0· 5 mM./l. 2,4-dinitrophenol and 2·0 mM./l. iodoacetic acid. It will be seen from Table 7 that the rate of efflux of both the rapid and the slow ion fractions was unaffected by the presence of the poisons. The concentration of the ⁴²K in the rapid fraction was also not affected by the presence of the poison molecules. The concentration of the ⁴²K which had been accumulated in the slow fraction during the experimental period was, however, reduced by the presence of the 2,4-dinitrophenol and the iodoacetic acid.

The results of the experiments illustrated in Fig. 7 show that the presence of 0·5 mM./l. 2,4-dinitrophenol and 2·0 mM./l. iodoacetic acid did not significantly affect the initial stages of the uptake of ⁴²K in these isolated preparations.

Isolated nerve cords were loaded with ⁴⁵Ca by soaking them for i hr. in a solution containing this radioisotope. The subsequent efflux of the labelled ions was studied by washing them in inactive solution (Fig. 8). In one batch of radioactive nerve cords the efflux was carried out in the presence of 0·5 mM./l. 2,4-dinitrophenol, whilst in another batch both the loading and the efflux were carried out in the presence of this poison. These results are summarized in Table 8. The presence of the metabolic inhibitor did not produce any dramatic effects either on the rate of efflux or on the concentrations of the rapid or slowly exchanging calcium fractions in the nerve cord.

The escape of magnesium ions from the nerve cord was studied by loading the isolated preparations for i hr. in physiological solution containing ²⁸Mg and studying the efflux in non-radioactive solution. The data from these experiments are summarized in Table 9. The mean concentration of the magnesium ions in the fast fraction was 13·8 mM./kg. tissue.

Figure 9 illustrates the efflux of ³⁶CI from a nerve cord which had been soaked in a solution containing the radioisotope for 1 hr. The mean concentration of chloride in the rapidly exchanging fraction was found to be 31·4 mM./kg. nerve cord tissue (Table 10). These results should be considered in relation to those carried out on the uptake of ³⁶C1 obtained in approximately 1·0 sec. exposures to the physiological solution (Table 11). Unlike those carried out with ³⁶Na these results with radio-chloride showed that an appreciable proportion, 19·6 mM./kg., was associated with the nerve cord following a brief exposure to the experimental solution. If this effect resulted entirely from surface contamination then the true rapidly exchanging fraction within the cord would only be about 11·8 mM./kg. tissue. It is thus only possible to make the rather unsatisfactory statement that the true value for the concentration of the rapidly exchanging chloride fraction lies between 11·8 and 31·4 mM./kg. nerve cord tissue.

Experiments were also carried out to investigate the effects of metabolic inhibitor on the initial uptake of ³⁶C1 into the nerve cord. Isolated nerve cords were soaked for 5 min. in normal physiological solution, or in one containing 0·5 mM./l. 2,4-dinitrophenol, before being transferred to the ³⁶Cl-labelled solution for a period of 2 min. The radioactive nerve cords were quickly washed in normal solution and then blotted before the ³⁸C1 content was determined. The results, which are summarized in Table 12, indicate that the initial uptake of ³⁶C1 was reduced by the presence of the metabolic inhibitor (P < 0·05).

The measurements on the ionic content of the central nervous system showed that the concentration of sodium in the nerve cord of Carausius (102·4 mM./l.) was very close to that of Periplaneta americana (103·2 mM./l. ; Treherne, 1961 a) despite the fact that the concentration of this ion in the haemolymph is almost an order greater in the latter insect. The potassium content of the nerve cord tissues in Carausius, on the other hand, greatly exceeded that of the cockroach (510·3 as against 180·2 mM./l.). This difference can be related to the haemolymph concentration of this ion in the phytophagous insect, for the potassium level in the haemolymph exceeded that of the cockroach by a factor of 2·7 (33·7 as compared with 12·3 mM./l.). With the divalent cations it should be noted that the tissue concentration of magnesium was low whilst that of calcium was high relative to their haemolymph concentrations.

The experiments on the in vivo uptake of ²²Na by the nerve cord of Carausius showed that the changes in the specific activity of the ion within nervous tissues occurred at approximately the same rate as in the cockroach central nervous system, despite the very low level of this ion in the haemolymph of Carausius. As the nerve cords of these two insects are of the same order of size these results must indicate a changed permeability to this cation at some stage in its entry into the tissues of the central nervous system of the latter species.

The existence of an appreciable extracellular fraction in the cockroach central nervous system was inferred from the following evidence (Treherne, 1961 d, 1962):

1. The efflux of labelled sodium from the isolated nerve cord was found to occur as a two-stage process.

2. The escape of the rapidly exchanging fraction was unaffected by the presence of metabolic inhibitors, although the efflux of the slowly exchanging sodium (identified as that taking place from the cellular fraction) was reduced under these circumstances.

3. The rapidly exchanging component was present in desheathed preparations, and was, therefore, not associated with the nerve sheath.

4. 18·2% of the nerve cord water was found to be accessible to ¹⁴C-inulin molecules.

5. The efflux of triated water occurred as a two-stage process, the rapidly exchanging fraction being of a similar magnitude to the demonstrated inulin space.

6. Electron micrographs of the nerve cord demonstrated the existence of an appreciable extracellular system, sufficient to accommodate the inulin molecules used in the previous experiments (Smith & Treherne, 1963).

The present experiments on the nerve cord of Carausius also demonstrated that the efflux of labelled sodium occurred as a two-stage process : an initial rapid escape, with a half-time of 19·7 sec., giving way to a slower leakage with a half-time of 649·0 sec. As with the nerve cord of Periplaneta the slow component of sodium efflux was significantly reduced by the presence of 0·5 mM./l. 2,4-dinitrophenol, the initial rapid escape being unaffected by this metabolic inhibitor. Similarly, only a relatively small proportion of the rapidly exchanging fraction appeared to be associated with the nerve sheath. The measured inulin space of 15·2% of the nerve cord water was lower than for the cockroach central nervous system. These results would, however, seem to indicate an essentially similar state of affairs, as far as the efflux of sodium is concerned, to that prevailing in the cockroach. Thus it seems reasonable to conclude that the slow component in the Carausius nerve cord represented the secretion of sodium from the cellular fraction, the initial escape being that of the ions from the extracellular fluid. It has been pointed out that the extrusion of sodium ions from the cellular fraction of the central nervous system of Periplaneta appears to be dependent on the presence of external potassium (Treherne, 1961,b, c) and it was suggested that as in the squid giant axon (cf. Hodgkin, 1958) the efflux is effected by some sort of linked sodium pump. A similar state of affairs may also exist in the nerve cord of Carausius, for it has been demonstrated that uptake of ⁴²K into the slowly exchanging fraction is reduced in the presence of metabolic inhibitors (Table 7).

If the concentrations of the rapidly exchanging ion fractions are related to the measured inulin space in the ventral nerve cord of Carausius (Table 2) then it is possible to arrive at an approximate estimate of the extracellular concentrations of these ions (Table 13). In the cockroach it was suggested that the relatively high concentrations of cations in the extracellular fluid resulted from a Donnan equilibrium with the haemolymph (Treherne, 1962). It is evident that no such simple relation appears to exist for the distribution of the ions between the extracellular fluid and the haemolymph in Carausius morosus.

The most obvious departure from a simple Donnan equilibrium is the high concentration of sodium ions in the extracellular fraction relative to the haemolymph (the ratio being 10·6) as compared with the other cations. In the in vitro preparation the ratio of the extracellular to the external sodium concentrations fell to 5·2, due partly to enlargement of the inulin space in the isolated preparation. In the presence of 2,4-dinitrophenol the ratio fell still further to 1·2. The obvious dependence of the high extracellular sodium concentration on aerobic metabolism clearly suggests some sort of secretion of sodium ions into the extracellular fluid. This supposition was confirmed by the experiments demonstrating the effects of the metabolic inhibitor on the initial rapid uptake of ²²Na and ²²Na by isolated nerve cords (Figs. 4 and 5). These experiments also help to explain the observation, which was discussed earlier, that changes in the specific activity of ²²Na within the nerve cord of Carausius took place as rapidly as in Periplaneta although the concentration gradient between the haemolymph and the nerve cord tissues was much steeper in the former species.

The rapid initial uptake of chloride ions by the isolated nerve cord was shown to be reduced in the presence of 0·5 mM./l. 2,4-dinitrophenol (Table 12). This effect could have resulted from the poisoning of the process involved in the uptake of sodium ions, for it might be expected that chloride ions would passively follow the active secretion of the sodium into the extracellular fluid. Such an uptake of chloride ions into an extracellular system containing fixed or indiffusible anion groups would again tend to obscure the Donnan effect with the haemolymph.

The extracellular compartment of the nerve cord of Carausius appears, from the relatively rapid exchanges taking place with the haemolymph, to be a rather ‘leaky ‘system. The pumping of sodium ions into the extracellular fluid must, therefore, be a very rapid process. This process is, for example, of a different order of magnitude from the outward secretion of sodium ions from the cells into the extracellular fluid of the central nervous system. It is relevant to compare this process with other sodium-transporting systems in insects. The most detailed investigations of sodium transport in insects appear to be those carried out on the anal papillae of the larvae of Aedes aegypti (Treherne, 1954; Stobbart, 1959, 1960). The calculated permeability constant, k_out (in the direction haemolymph → external medium), for fed larvae maintained in 2·0 mM./l. Na was 0·0761 hr.⁻¹ or 2·11 × 10⁻⁸ sec.⁻¹ (Stobbart, 1959). To estimate Kom, the true permeability constant of the anal papillae, it is necessary to know the effective surface/volume ratio of the system. A rough estimate of this ratio can be made by comparing the haemolymph volume of 4th-stage larva (measured by weighing that squeezed from individual insects) with the surface area of the papillae (in which these organs were considered as cylindrical structures). The mean result of 2·053 mm.² for 128 μl. of haemolymph gave an approximate surface volume/ratio of 1·604 cm.⁻¹. It seems legitimate to use the haemolymph for the volume in this ratio, for it has been shown that the exchanges between it and the external medium approximate to those taking place in a two-compartment system (Treherne, 1954). The constant K_out then becomes 1·31 × 10⁻⁸ cm. sec.⁻¹ and, with a haemolymph concentration of 116·0 mM./l. Na (Stobbart, 1959), K_In is approximately 7·60 × 10⁻⁴ cm. sec.⁻¹. K_ln for the exchanges taking place in the nerve cord of Carausius has already been estimated to be about 9·86 × 10⁻⁶ cm. sec.⁻¹. It is apparent, therefore, that the transport of sodium ions into the extracellular fluid of the central nervous system of Carausius is taking place at less than the rate demonstrated for the uptake of these ions by the anal papillae of Aedes larvae.

The demonstrated transport of sodium ions into the extracellular fluid presumably results from the activities of the cellular layer, or perineurium, which is associated with the connective tissue sheath of the insect central nervous system (cf. Smith & Treherne, 1963). The ability of the nerve of Carausius to continue functioning when portions of the sheath were removed (Treherne, 1965) suggests that other glial cells may also be involved in sodium movements into the extracellular system, for in the cockroach desheathing has been shown to cause removal or damage to the perineurium (Twarog & Roeder, 1956).

There has been in the past some speculation as to the means by which the axons of phytophagous insects are able to function in haemolymph containing very low concentrations of sodium ions and relatively high levels of other cations. The present investigation has shown that the cells of the central nervous system of Carausius function in a medium with a relatively high sodium content. Table 14 contains the estimated extracellular and cellular concentrations of the various cations used in this investigation. It should be emphasized that the figures for potassium, calcium and magnesium are only approximate for they involve the extrapolation of data from in vitro studies to those made on the intact nerve cord. The resting potential of many excitable cells has been shown to be related to the external and internal potassium concentrations by the Nernst equation:

The conduction processes in the nerve cord of Carausius have been shown to be dependent upon the presence of external magnesium ions as well as sodium ions Treherne, 1965). Such an effect could result from some secondary effects of the divalent ions on the axon membranes within the nerve cord. It is also possible, however, as with the muscle fibres of this insect (Wood, 1957) and of the crab (Fatt & Katz, 1953; Fatt & Ginsborg, 1958), that the action potential could result from an inward movement of the divalent ion. In the event of such an influx of both sodium and magnesium ions it can be calculated (using the Nemst equation, and from the ratio of 11·0 shown in Table 14) that the magnesium ions alone would only contribute an equilibrium potential of about 29·6 mV. The involvement of the divalent ions would thus only produce a slight increase in the mean potential as compared with that resulting from the movement of sodium ions alone. The increased magnesium conductance would, however, contribute appreciably to the total ionic current carried during the action potential. These estimates of the possible effects of the high concentration of magnesium ions in the extracellular fluid are made on the assumption that the activity coefficients are the same in this phase and within the axons. Any reduction in the activity of the cation in either phase would result in a departure from the estimated magnesium equilibrium potential. If, as is the case for calcium ions which move in the axoplasm at less than 1/30 the rate in free solution (Hodgkin & Keynes, 1957), there is a reduced activity within the axons in this insect then it is possible that the potential resulting from any inward movement of magnesium ions could be greater than that indicated above.

I am indebted to Dr R. H. Adrian and Prof. A. L. Hodgkin for helpful discussions during the course of this work.