The pharmacology of an insect ganglion: actions of carbamylcholine and acetylcholine

The overall aim of the present series of experiments is to determine the properties of the acetylcholine receptors in an insect central ganglion. The two main experimental approaches to this problem are biochemical and pharmacological. A biochemical study of the acetylcholine receptor(s) isolated from the ventral nerve cord and the sixth abdominal ganglion of the cockroach Periplaneta americana is currently being undertaken. The present series of pharmacological experiments constitutes a parallel investigation of cholinergic transmission at the cereal nerve, giant fibre synapse located in the sixth abdominal ganglion of Periplaneta americana.

Following the demonstration by Pumphrey & Rawdon-Smith (1937) of a monosynaptic link between fibres in the cereal nerve and giant interneurones in the sixth ganglion of Periplaneta americana, a number of pharmacological studies have been carried out on this synapse (Roeder & Roeder, 1937; Roeder, 1948; Twarog & Roeder, 1957; Yamasaki & Narahashi, 1958, 1960; Callec & Boistel, 1967, 1971; Flattum & Shankland, 1971; Shankland, Rose & Donniger, 1971; Callec, 1972). This work has recently been reviewed, and the overall evidence for cholinergic transmission is compelling when consideration is given to the physical and chemical barriers which can reduce the activity of acetylcholine and drugs applied to the ganglion (Callec, 1972, 1974).

Recent advances in techniques for recording synaptic events at the cereal nerve, giant fibre synapse of the cockroach have facilitated the interpretation of results of pharmacological experiments (Pichon & Callec, 1970; Callec & Sattelle, 1973; Sattelle, 1976). In this investigation, the mannitol-gap technique (Callec & Sattelle, 1973) has been employed to compare various ways of expressing the dose response relationship of a cholinergic agonist. In addition, the actions of carbamylcholine (carbachol) and acetylcholine have been investigated.

The mannitol-gap technique employed to record synaptic potentials from the sixth abdominal ganglion of Periplaneta americana was as described in a previous report (Callec & Sattelle, 1973). Most of the experiments reported here were performed on partly desheathed ganglia in which the sheath was split in several places using fine stainless-steel needles. This procedure improved the accessibility of synaptic elements to drugs introduced into the bathing medium. The normal Ringer solution employed in these experiments had the following composition: NaCl, 208 · 6 mm; KC1, 3·1 mm; CaCl₂, 5·4 DIM; NaHCO₃, 2·0 mm; pH 7·0. Acetylcholine chloride, carbamylcholine chloride and eserine sulphate were dissolved in this Ringer solution.

When carbamylcholine chloride was added to the Ringer solution at concentrations of 10⁻⁴ M and higher, a rapid reduction in the monosynaptic EPSP was noted. This was accompanied by a rapid ganglionic depolarization. The time-course of the EPSP changes and the depolarization were followed for several preparations during the application of 5·10⁻³ M carbamylcholine chloride. At this concentration the EPSP was quickly blocked and a depolarization of 25–35 mV was noted. On subsequent reexposure to normal Ringer, synaptic transmission recovered after some delay, the EPSP and ganglionic polarization being restored (Fig. 1).

In several experiments in which low concentrations (around 10⁻⁵ M) of carbamylcholine were applied, although no change in polarization was noted, a distinct, reversible potentiation of the EPSP was detected (Fig. 2). When longer exposures to higher concentrations of carbamylcholine were studied, repolarization of the ganglion was noted in the presence of the drug. This attenuation of the ganglionic response (desensitization) was clearly seen when 10⁻³ M carbamylcholine was applied continuously. Complete repolarization was noted within about an hour but not the return of the EPSP (Fig. 3).

The response of the ganglion to carbamylcholine was used to evaluate various ways in which the dose-response curve for the ganglionic actions of cholinergic agonists could be usefully evaluated. The curves in Fig. 4 were constructed using data obtained from eighteen ganglia. For Fig. 4 a all the values for depolarization at a particular concentration were utilized whether the ganglion was challenged once or up to six times with the drug. The mean depolarization and the standard error were calculated for each concentration.

Some problems of utilizing the data in this way to produce a dose-response curve emerged when the effects of repeated doses on the same preparation were examined. In the lower curve of Fig. 4,b increasing doses of carbamylcholine were added successively to the same ganglion. After each application, the ganglionic polarization was allowed to recover in Ringer to the initial level before a subsequent (higher) concentration was added. Under these conditions the maximum level of depolarization achieved was usually much less than when a ganglion was challenged only once by a high concentration of carbamylcholine. Similar findings were noted with other drugs (Callec, Dowson & Sattelle, 1975). This attenuation of the response by desensitization could be avoided by challenging each preparation with only one test concentration of carbamylcholine. The upper curve in Fig. 4 b was obtained by plotting data from thirteen ganglia which were tested once or twice only. In cases where the preparation was challenged twice, a concentration giving a minimal response was initially employed, followed by a higher concentration giving a maximal or near maximal response. Under these conditions, the response of the ganglion was not noticeably depressed by the application of the first, low, concentration. The resultant curve was steeper and the experimental points were less scattered than in Fig. 4 a.

A dose-response curve was also computed from a twin-pulse experiment on a single ganglion following a method previously applied to the response of the isolated rat superior cervical ganglion to acetylcholine in the presence of eserine (Cuthbert & Dunant, 1970; see also Appendix). Such a curve is illustrated in Fig. 4,c (solid line). The dose-response curve computed in this way was compared with data points representing thirteen first responses on thirteen separate ganglia (Fig. 4c). The experimental points showed at best only an approximate fit to the computed curve.

For the remainder of this study and in subsequent experiments in this series (see also Callec et al. 1975), dose-response curves have been produced in the manner shown in Fig. 4 b (upper curve) using only first responses or the results of twin-pulse experiments in which the test concentrations were widely separated, the first pulse producing a near minimal response. By adopting this method in preference to the other two described, the effects of desensitization were made negligible and a doseresponse curve was produced which minimized the scatter of experimental points, allowing a ready comparison of ganglionic sensitivity to a range of applied drugs.

The action of acetylcholine was described previously when it was noted that the application of acetylcholine at 4·10⁻² M to a desheathed ganglion resulted in depolarization accompanied by abolition of the EPSP (Callec & Sattelle, 1973).

A dose-response curve was obtained for acetylcholine using the method outlined for carbamylcholine in which primarily first responses (and first and second responses in the case of a low dose followed by a high dose) were used (Fig. 5). This curve was of similar shape to that described for carbamylcholine but was shifted to the right on the log concentration axis (Fig. 5).

In a number of experiments the desheathed ganglia were pretreated for 10–15 min. with the anticholinesterase eserine at a concentration of 10⁻⁶ M. The subsequent addition of acetylcholine 10⁻⁶ M and higher concentrations, in the presence of eserine, resulted in a rapid depolarization and reduction in amplitude of the EPSP. The doseresponse curve shown in Fig. 5 was obtained. Only a brief pretreatment with the anticholinesterase was possible, since prolonged (20–25 min.) exposure to eserine at this concentration produced depolarization.

The mannitol-gap technique has proved to be a useful way of rapidly obtaining dose-response data for the actions of pharmacological agents on synaptic transmission in the terminal abdominal ganglion of the cockroach (Periplaneta americana). In this report, we have followed changes in the monosynaptic EPSP (recorded from the cereal nerve, giant fibre connexions) and ganglionic polarization (cf. Callec & Sattelle, 1973), during the actions of acetylcholine and carbamylcholine. Clearly the response to a drug recorded in this way is several physiological steps away from the primary drugreceptor interaction (cf. Triggle, 1971). Other problems in the interpretation of the results of pharmacological experiments on a central nervous ganglion include multiple sites of action and differential accessibility of structurally related drugs. In intact and desheathed ganglia and connectives, problems of drug accessibility may be severe. For example, both the lipid solubility and charge of the molecule are important factors in accessibility (O’Brien & Fisher, 1958; O’Brien, 1967; Thomas, 1976a, b).

In the case of acetylcholine and closely related chemicals, the presence of acetyl-cholinesterase and non-specific choline esterases may profoundly affect the access of the drug (Colhoun, 1958, 1963 ; Smith & Treherne, 1965). An added complication is that the amplitude of the response to a particular dose of the drug should strictly be assessed under equilibrium conditions. It is nevertheless common practice to record the maximum response following a single dose.

With these limitations in mind, ways of evaluating dose-response data have been considered for the cholinergic agonist carbamylcholine. This drug is highly resistant to hydrolysis by either acetylcholinesterase or non-specific cholinesterase. Data obtained in experiments with carbamylcholine have been plotted in various ways in order to establish criteria for the production of a useful dose-response relationship. One method has been to pool data from several experiments on different ganglia (Fig. 4 a). This can distort the dose-response curve for two main reasons. First, averaging the responses at a particular dose-level will flatten the ‘mean’ curve in comparison with the curve that would be obtained for a single preparation. Secondly, any ‘desensitization’ of the preparation during the application of repeated doses of agonist will also flatten the dose-response curve. The return to the initial level of polarization in normal Ringer following achievement of a maximum depolarization by a particular dose of agonist is no guarantee of a recovery of full sensitivity (Fig. 4 b, lower curve).

In another method of obtaining a dose-response curve it is assumed that the rate of diffusion of carbamylcholine to the receptor governs the time-course of depolarization. Under these conditions a dose-response curve can be computed (based on diffusion equations) using data from the application of a supramaximal and a submaximal dose to a single preparation (see Cuthbert & Dunant, 1970; see also Appendix). A comparison of the curve computed in this way with data points that represent first responses only reveals at best an approximate fit (Fig. 4c). This method has been used to derive a dose-response relationship which is a good fit to data for the response of the isolated rat superior cervical ganglion to acetylcholine in the presence of eserine (Cuthbert & Dunant, 1970). It is not immediately obvious why the assumptions made for the vertebrate ganglion do not appear to hold so well for the insect preparation. However, the successful application of equation 1 (Appendix) demands that the thickness of the stationary layer is small compared to the dimensions of the preparation. This may well not be the case in the insect ganglion with its long and convoluted system of extracellular spaces.

Finally, we have obtained a dose-response curve based on first responses from ganglia and the results of twin-pulse experiments in which the test concentrations are widely separated (Fig. 4 b, upper curve). Since this procedure minimizes the effects of desensitization and considerably reduces the scatter of the experimental points by comparison with the other two methods it has been adopted as a routine method for producing dose-response curves in this survey.

The application of a relatively low concentration of carbamylcholine (around 10⁻⁵ M) induces a slight potentiation of the EPSP; higher concentrations give rise to ganglionic depolarization accompanied by either a decrease or block of the EPSP, according to the dose. These last mentioned changes can be accounted for in terms of postsynaptic actions, and the results for carbamylcholine closely parallel those for acetylcholine reported in an earlier paper (Callec & Sattelle, 1973). Presynaptic effects of this drug have been noted in the sympathetic ganglion (Voile & Koelle, 1961). In this vertebrate preparation it is proposed that carbamylcholine stimulates the release of acetylcholine. It may be that in the insect ganglion presynaptic, as well as postsynaptic, effects of carbamylcholine contribute to the potentiation of the EPSP at low concentrations of carbamylcholine.

In an earlier paper in which the mannitol-gap technique was employed (Callec & Sattelle, 1973) and in all previous studies on the cereal nerve, giant fibre synapse (cf. Callec, 1974), the relative insensitivity of even the desheathed preparation to acetylcholine was noted. We note that eserinization (10⁻⁶ M) shifts the dose-response curve for acetylcholine considerably (Fig. 5) so that effects on polarization and EPSP are noted at concentrations of io⁻⁶ M and above. This finding is consistent with the observations of Shankland et al, (1971) who, using qualitative changes in overall spike activity, note that 10⁻⁸ M is the lowest concentration of acetylcholine which reliably gives a ‘moderate response’ to acetylcholine. It should be noted, however, that these authors also pre-treated ganglia with hemicholinium-3 in order to prevent transmitter synthesis, and that in some cases responses were noted at concentrations as low as 10⁻⁷ M and 5·10⁻⁷ M acetylcholine.

The main findings of the present report are the development of a routine procedure for presenting dose-response data for the synaptic actions of cholinergic agonists and a quantitative demonstration of the increased sensitivity to acetylcholine of synaptic transmission at the cereal nerve, giant fibre synapse following eserinization of the desheathed ganglion.

The authors thank Dr J. F. Donnellan and Dr J. M. Young for helpful discussions during the course of this work.

In this appendix a brief account is given of the theory and method employed to compute the dose-response curve shown in Fig. 4 c. For a more detailed account of this method and a wider discussion of its applications, the reader should consult the original paper of A. W. Cuthbert and Y. Dunant (1970).

Drugs must traverse a stationary fluid layer by diffusion in order to reach the receptors located on the nerve membranes. If the interaction of the drug and the receptor is rapid, then diffusion can be the rate-limiting process determining the time-course of the reaction and possibly also that of the response. If the time-course of the response (e.g. depolarization) produced by an applied drug is diffusion limited, then it is possible to calculate the concentration of drug existing at the membrane at any time following application and thereby relate concentration to response. To achieve this the following assumptions must be made: (i) only a small fraction of nerve cell membrane is occupied by receptors and negligible amounts of the drug are immobilized by the receptors compared to the total amount of drug reaching the cell surface; (ii) all receptors involved in the response are equidistant from the bathing solution (i.e. the stationary layer has a constant thickness l); (iii) the interaction of drug and receptor and the consequent cellular events are very rapid; (iv) the diffusion coefficient of carbamylcholine is assumed to be constant in the stationary layer, which in this case consists largely of remnants of the neural lamella and the extracellular spaces of the ganglion.