A Quantitative Study of the Ionic Basis of Extraneuronal Potential Changes in the Central Nervous System of the Cockroach (Periplaneta Americana L.)

Earlier studies carried out in this laboratory have demonstrated the existence of rapid and relatively large extraneuronal potential changes following elevation of the potassium concentration in the medium bathing intact cockroach connectives (Pichon & Treherne, 1970; Treherne et al. 1970). This phenomenon appeared to be associated with a restricted access of potassium ions to the fluid bathing the axon surfaces, for the extraneuronal positivation^† observed with intracellularly located micro-electrodes was not accompanied by an equivalent reduction in the amplitude of the recorded action potentials. It was also shown that this effect was reduced or abolished if tension was applied to the nerve or if the connectives were dried and briefly exposed to air (Pichon & Treherne, 1970). In such preparations only the continuous and relatively slow potential changes were observed, corresponding to the direct depolarization of the giant axons. It was concluded that this latter electrical response resulted from an increased access of potassium ions into the extracellular system, most probably as a result of the disruption of intercellular occlusions at the inner margin of the perineurium (Maddrell & Treherne, 1967; Lane & Treherne, 1970).

The previous investigations have not elucidated the nature of the extraneuronal potential changes resulting from increased potassium concentration in the bathing medium. Neither is it possible to predict the effects of other ions on these changes. The present study was therefore carried out in an attempt to throw some further light on the ionic basis of the extraneuronal potential changes in intact cockroach connectives. We have tested the effects of monovalent inorganic and organic cations (Li⁺, Cs⁺, Rb⁺, K⁺, TEA⁺, tris and choline), divalent cations (Ca²⁺ and Mg²⁺) and the effects of replacement of Cl⁻ by ⁠. An attempt is made to elucidate the structural basis of the phenomena by comparison of the experimental results with those derived from four possible theoretical model systems.

The potential changes described in this paper were measured using the ‘sucrose gap’ technique employed in a previous investigation (Pichon & Treherne, 1970). The penultimate connective was passed through petroleum jelly seals between three parallel compartments, the right-hand one being continuously perfused with the experimental solution. The central compartment contained flowing mannitol (483·0 mm/1) solution and the left-hand one contained normal Ringer solution. The potential changes resulting from alterations in the ionic concentration in the right-hand compartment were recorded via a saline-filled agar bridge connected to a high-impedance amplifier, the left-hand compartment being connected to the indifferent electrode via a second saline-agar bridge. Perfusion of the compartments was achieved using the gravity-fed system previously described.

The normal physiological solution used was that devised by Yamasaki & Narahashi (1959): 210·0 mm/1 Na+, 3·1 mm/1 K⁺, 1·8 mm/I Ca²⁺, 216·9mm/1 Cl⁻, 0·2 mm/1 and 1·8 mm/1 ⁠. Variation of the ionic composition of the experimental solutions was achieved by substituting various ions for those of sodium.

Fig. 1 illustrates the effects of solutions in which the sodium ions were replaced by those of potassium and subsequently by those of choline or tris (2-Amino-2(hydroxyl-methyl)-propane-1,3-diol) on the potential changes measured using the sucrose-gap, technique in intact preparations. The rapid positivation produced by the high-potassium solution was followed by smaller negative-going potential changes in preparations exposed to solutions in which the normal concentration of sodium ions was replaced by the organic cations.

A somewhat variable response was obtained to the solutions in which sodium ions were replaced by those of choline or tris. The potential changes produced by choline solutions varied between 4·0 and 11·0 mV with a mean value of 7·0 ± 0·75 (s.E.) mV. These values are expressed relative to the potential changes produced by the high-potassium solution in Fig. 12. The potential changes produced by replacement of sodium ions with those of tris resulted in potential changes of between 5·0 and 15·0 mV, with a mean value of 9·3 ± 1·1 (S.E.) mV.

Replacement of sodium ions in the bathing solution with those of lithium was found to produce only slight reversal of potential of the form illustrated in Fig. 2. The small and transient positivation obtained on change of solutions was a consistent feature of the results obtained with solutions in which sodium ions were replaced by those of lithium. The mean value of this change was 1·8 ± 0·37 (S.E.) mV.

Fig. 3 illustrates the effect of a solution in which sodium ions were replaced by those of TEA. Incorporation of this organic cation in the bathing solution resulted in slight positivation (1·9 ± 0·24 mV).

Replacement of sodium ions by those of caesium in the bathing solution resulted in an appreciable positivation, averaging 7·9 ± 0·24 (S.E.) mV (Fig. 4).

Fig. 5 illustrates the effects of 214·0 mm Rb⁺ in place of Na⁺ in the bathing medium. It will be seen that rubidium produced a positivation which approached that recorded with the high-potassium solution. These results also show that application of a pulse of potassium during an extended exposure to the rubidium solution or a pulse of rubidium during a prolonged exposure to high-potassium resulted in only small and transient potential changes. The positive potential change produced by the rubidium solution averaged 29·1 ± 2·1 (S.E.) mV which, expressed as a fraction of the potassium positivation, corresponds to a relative change of + 0·98 ± 0·07 (S.E.).

Fig. 6 shows the effects of pulses of solutions of various potassium concentrations on the potentials recorded in an intact connective. In these experiments sodium ions were replaced by those of choline, the concentrations being appropriately reduced with the increased potassium concentration. A semilogarithimic plot of the results is shown in Fig. 7. It will be seen that this relation shows a slight departure from linearity, an effect which was observed with all the preparations tested.

Elevation of the concentration of these divalent cations (to replace those of sodium in the bathing medium produced complex potential changes. Fig. 8 shows the potential changes recorded in an experiment in which the intact connective was exposed to two short pulses of a high-potassium solution, followed by a prolonged exposure to a high-calcium solution. The rapid positivations produced by the high-potassium solution were followed (with high-calcium solution) by a relatively rapid negative-going potential (1) which decayed rather more slowly (2) and then showed a slow negative component (3). Only a small positive potential change (4) was recorded on return to normal solution. A final pulse of high-potassium solution showed a change in the form of the recorded positivation which tended to develop more slowly.

Elevation of the magnesium concentration in the bathing medium resulted in a pronounced potential change of opposite polarity to that produced by high-potassium solution (Fig. 9). Return to normal solution was associated with a relatively slow decrease in potential following a small transient positivation. A second exposure to high-magnesium resulted in a similar potential change to that observed in the first exposure, except that the absolute level attained was lower. This effect is more clearly seen in Fig. 10 when the successive pulses of high-magnesium were associated with roughly equivalent potential changes which showed a progressive decline in the absolute level attained. These results contrast with those for calcium in which, as will be seen from Fig. 10, there was a progressive decrease in the magnitude of the change but a return to the potential level initially obtained in normal solutions.

To test the effect of the anion species present on the extraneuronal positivation produced by high external potassium concentrations, experiments were carried out in which sulphate was substituted for chloride in the bathing solutions, before replacing the sodium ions by potassium. As will be seen from Fig. 11 there was very little effect of substitution of sulphate for chloride ions on the rapid initial positivation recorded with 214·0 mm/1 potassium in the bathing solution.

It has recently been established that the extraneuronal positivation (measured using microelectrodes or sucrose-gap technique), produced by elevation of the external concentration of potassium ions, is correlated with a restricted access of the cation to the fluid bathing the surfaces of the giant axons in intact cockroach connectives (Treherne et al. 1970; Pichon & Treherne, 1970). It was tentatively suggested, in earlier publications from this laboratory, that the observed restriction upon the intercellular movements of cations from the bathing medium was associated with occluded regions, containing tight junctions and septate desmosomes (Maddrell & Treherne, 1967), at the inner ends of the tortuous intercellular clefts which traverse the perineurium, for the inward movement of peroxidase molecules has been shown to be restricted in this region (Lane & Treherne, 1970).

The postulation of a severe restriction upon the inward diffusion of cations at the inner end of the perineurial clefts provides a basis for the explanation of the extraneuronal potential changes produced by the substitution of potassium, and other cations, for sodium ions in the bathing medium. It can readily be envisaged, for example, that in the presence of high external concentrations of potassium ions there would be a depolarization of the outwardly directed perineurial membrane, the equivalent depolarization of the inwardly directed one being effectively prevented by the presence of the occluded regions at the inner ends of the perineurial clefts. Such a system would produce rapid potential changes, measured using both extracellularly and intracellularly located microelectrodes, without an equivalent direct depolarization of the axonal membranes.

The various extraneuronal potential changes described above for other alkali metal and organic cations can also be interpreted in terms of their effects on the outer perineurial membrane. The appreciable negative potential change produced by tris and choline can, for example, be postulated to result from an efflux of extracellular sodium ions induced in the absence of this cation in the bathing medium. This effect would thus be analogous to the hyperpolarization produced by the replacement of sodium ions by those of tris in the solution bathing gastropod neurones (Moreton, 1968 a; Sattelle, 1970). In the case of Helix neurones this effect can be reasonably inferred from application of the ‘constant-field’ theory which shows that the observed resting potential is considerably less negative than the potassium equilibrium potential largely as a result of the significant sodium permeability of the resting cell membrane (Moreton, 1968b). The appreciable extraneuronal positivation induced by externally applied rubidium and caesium ions also parallels the effects of these cations on a Variety of excitable cell membranes in which their action resembles, in varying degrees Fhat of potassium ions (cf. Sjodin, 1959).

From Table 1 it is apparent that the polarity of the extraneuronal potential can be related to the radii, either crystal or hydrated, of the inorganic monovalent species, only those of smaller hydrated radii than Na⁺ producing a positivation. The very small positivation produced by TEA⁺ also accords with the observation that this ion can partly replace sodium in nerve membranes (Binstock & Lecar, 1969; Y. Pichon, unpublished result) and is similar enough to potassium with a single shell to be accepted by a potassium ‘site’ but unsymmetrical enough to block the ‘channel’ (Armstrong, 1966).

The perineurial system proposed here differs, however, from the frog skin in the small extent of the potential changes produced by alteration in the potassium concentration of the bathing medium. Koefed-Johnsen & Ussing (1958) showed that the inner epidermal surface of the frog skin approximated to a potassium electrode. It is apparent from Fig. 7 that the measured extraneuronal potentials in the cockroach connective depart very markedly from the 57 mV slope for decade change in external potassium concentration, the equivalent slope being only about 17 mV.

It could be envisaged that this extreme departure from the relationship predicted by the conventional Nernst equation might result from the appreciable permeability of the outward-facing perineurial surface to other ion species. It is presumably not due to attenuation of the potentials as measured by the sucrose gap technique, since the potassium-dependence of the resting potential of the axons in de-sheathed preparations was found to show the same 42 mV slope, when measured by the sucrose-gap technique, as when measured directly with an intracellular microelectrode (Schofield & Treherne, unpublished observations). Such a system might, therefore, reasonably be supposed to be described by constant-field equations (Goldman, 1943; Hodgkin & Katz, 1949).

The potential changes resulting from elevated concentrations of monovalent cations other than sodium or potassium are similarly interpretable in terms of the selective permeabilities of the perineurial barrier. The extraneuronal potentials resulting from elevated concentrations of divalent ions, however, were complex in form and are consequently more difficult to interpret. It would seem reasonable to suppose that the initial rapid negative potential changes produced by high concentrations of calcium and magnesium ions were similar in nature to those produced by choline and tris, that is, they resulted from a significant outward movement of sodium ions across the outer perineurial membrane in the presence of relatively non-penetrant ions in the bathing medium. The subsequent potential changes in the presence of high external concentrations of calcium ions might then reasonably be expected to result from the specific effects of this cation, either on the external perineurial membrane, or on the tight junctions.

Similarly the progressive reduction in the potential level observed in successive exposures to pulses of high-magnesium solution could result from a progressive decrease in the sodium permeability of the barrier, so that the sodium efflux during successive exposures to a high concentration of magnesium is gradually reduced.

This section describes an attempt to analyse the situation in the intact connective theoretically, using the constant-field theory, as developed by Moreton (1968b). Assuming that the behaviour of the extracellular potential changes observed in the experiments are due to the presence in the intact, unstretched connective of a peripheral diffusion barrier, the appropriate model system must consist of such a barrier, in series with a long, narrow channel, at the inner end of which is the giant axon (Treherne et al. 1970). The barrier will then presumably represent either the cell membranes of the perineurial cells, or the tight junctions at the base of the perineurium, or a combination of the two. The long channel will represent in a simplified form the mesaxon channel, in series with the network of intercellular spaces between the glial cells. In the first instance, it is assumed that the system of channels is closed (i.e. that significant exchange of ions with the intracellular compartment does not occur). This simplification is probably not justified – the effects of exchange with the intracellular compartment will be discussed below – but it forms a useful starting point for the investigation. Clearly, the detailed behaviour of such a grossly simplified model can resemble only qualitively that of the real tissue ; nevertheless, varying degrees of elaboration of the model are likely to produce only quantitative changes in its theoretical behaviour, so that the analysis of its behaviour may be expected to throw at least some light on the mechanisms responsible for the behaviour of the extraneuronal potential.

The first stage in the analysis is to investigate the electrical and permeability properties which must be attributed to the perineurial barrier, in order to reproduce the experimental behaviour of the extraneuronal potential. These properties can then be used to investigate the movements of ions in the model system, occurring in response to changes in the composition of the bathing medium, and hence to predict some aspects of the behaviour of the giant axons under these conditions. The most complete set of results available is that shown in Fig. 7, relating the extraneuronal potential to the concentration of potassium ions in the external medium. These results will be used as the principal criterion with which to test the suitability of the chosen parameters of the barrier.

A single barrier in parallel with a leakage channel of arbitrary ion-selectivity, such as might be formed by cell membranes of the perineurial cells in parallel with an intercellular pathway, the latter partially restricted by tight junctions between cells, is similarly insufficient. This can readily be seen, since equation (1) is derived by addition of the individual ionic currents across the barrier, without reference to the morphological location of the ‘channels’ through which they move. Provided, that the concentrations of the ions are uniform over each of the two faces of the system, addition of the currents will always produce an equation of the form (1), in which the effective permeability to each ion is the sum of the permeabilities of the individual ‘channels’ through which it can pass.

Qualitatively, Fig. 14 shows that the concentration of potassium ions immediately inside the barrier, curve (a) is expected to rise rapidly at first, subsequently increasing more slowly, towards the value at the outer face. The potential difference across the barrier, as derived from equation (3), curve (c) thus falls rapidly at first, from an initially high value, and then more slowly. The initial high value would be unlikely to be attained in practice, since the raised concentration of ions at the outer surface of the barrier would not be established instantaneously owing to the presence of unstirred layers in the bathing medium. The value reached by the potential when it has begun to fall more slowly may be interpreted as representing the extraneuronal potential level observed in the experiments. Curve (b) in Fig. 14 shows the corresponding behaviour of the potassium concentration at the inner, axonal end of the extracellular cleft, which rises initially more slowly, subsequently following a parallel course to that at the outer end, but with a lag time comparable with that required for diffusion in the stretched connective (Treherne et al. 1970).

The detailed behaviour of the concentration is likely to be affected significantly by exchange of ions with the intracellular (glial) compartment. This is in contrast with the situation in the stretched connective (Treherne et al. 1970), where movements of ions in the extracellular spaces were shown to be much more rapid than exchange across the cell membranes, so that the latter could be neglected. In the present case, the flux of ions across the peripheral barrier is comparatively slow, so that it can be shown, for example, that movement in the extracellular system will be significantly affected by exchange across the glial cell membranes, even if the latter have a potassium permeability as low as 10⁻⁸ cm s⁻¹. Since the rate of uptake by the glia would presumably increase with the extracellular potassium concentration, the effect would be to accentuate the difference between the initial rapid rise and the subsequent slow approach to equilibrium (Fig. 14). This is in accord with the experimental results, which show that the extraneuronal potential does not decline appreciably during several minutes exposure to a high-potassium solution, showing that the concentration at the inner face of the barrier has not increased significantly beyond the initial, rapidly established level during this time. It would also explain the observations of Treherne et al. (1970), that the axonal action potential is not appreciably attenuated after 10 min exposure to a high-potassium solution, showing that the extra-axonal concentration increases only very slowly under these conditions. Diffusion in the extracellular system is thus presumably much retarded by uptake of potassium ions by the glia.

The potassium permeability, used to obtain curves (a) and (b) in Fig. 14, was chosen to give a concentration of 15·6 mm/1 at the inner face of the barrier after 5 min. The calculated potential difference across the barrier at this time will thus agree with the experimental figure of 26 mV. Corresponding calculations for other values of the internal potassium concentration produce results in broad agreement with the experimental figures, although in view of the considerations discussed above it is unreasonable to expect more than a qualitative agreement.

The effect on the potential difference across the barrier of replacing sodium in the bathing medium by other cations is clearly dependent on the ability of the latter to penetrate the barrier. Ions which penetrate more readily than sodium (potassium, rubidium, caesium, TEA) cause a positive-going potential change ; ions which penetrate less readily than sodium cause a negative-going change, since the principal effect is to cause sodium efflux from the system. In the case of divalent ions, particularly calcium, there appears also to be a secondary effect, which could be interpreted in terms of changes in the permeability properties of the barrier. Thus, the initial, transient negative potential caused by high calcium (Fig. 8) is presumably due to sodium efflux, which is subsequently reduced, possibly owing to a fall in the sodium permeability of the barrier, caused by the high concentration of calcium ions. There is, however, insufficient evidence to form a basis for more detailed investigation.

= Concentrations^* of potassium, choline, calcium and chloride ions in the external medium (mm/1).

⁠, etc. = Concentrations of the respective ions at the inner face of the peripheral diffusion barrier (mm/1).

P_K, etc. = Permeabilities of the barrier to the respective ions (cm s⁻¹).

V = Potential difference across the barrier (mV).

= ‘Extraneuronal potential.’

F = The Faraday (coulombs mole^{− 1}).

R = Gas constant (joules mole^{− 1} °K^{− 1}).

T = Absolute temperature (°K).

M_K, etc. = Inward net fluxes of the respective ions across the barrier (mole cm^{− 2} s^{− 1}).

δ = P_ch/P_k.

α = P_Na/P_k.

β = P_Cl/P_k.

Y = (mm/1).

μ = (Change in /Change in ⁠.)

D = Diffusivity of potassium ions (= 1·8 × 10⁻⁵ cm² s⁻¹).^*

C = Local concentration of potassium ions in the extracellular system (mm/1).

C₀ = Concentration of potassium ions in the bathing medium (mm/1).

x = Distance along the extracellular pathway, measured from the peripheral barrier (cm).

l = Total length of extracellular pathway, from peripheral barrier to axon (cm).

t = Time, measured from the time when the bathing solution is changed (s). Other symbols are as defined in the text.

We gratefully acknowledge the expert technical assistance of Mr P. K. Schofield and the help of Mr J. Rodford in preparing the diagrams.