ABSTRACT
The nature of the transmitter substance released at the lateral-line efferent synapses was investigated by histochemical techniques in Xenopus and Acerina and by pharmacological methods in Xenopus.
The bases of lateral-line hair-cells and fine fibres in the lateral-line nerves reacted positively for acetylcholinesterase in Acerina and to a lesser extent in Xenopus.
Acetylcholine (10−6M, and acetyl-β-methyl choline (5×10−6M), which has a muscarinic action, caused strong reversible inhibition of afferent impulses when pipetted on to the undersides of lateral-line stitches. Carbachol (5 × 10−4 to 5 × 10−6M) caused a smaller reversible inhibition of spontaneous afferent impulses, but other nicotinic substances (propionylcholine and buterylcholine) had no effect.
Physostigmine (5 × 10−8M) prolonged inhibition of afferent impulses by electrical stimulation of efferent fibres, but atropine (5 × 10−6M) blocked it.
Calcium and magnesium interact at the efferent synapses in a way similar to that found at the amphibian neuromuscular junction.
Arguments are put forward to support the hypothesis that acetylcholine is released at lateral-line efferent synapses.
INTRODUCTION
The sensory hair-cells of the acoustico-lateralis system of sense organs have been shown, in the majority of cases, to receive efferent innervation. Studies on the ultrastructure of the sensory epithelium in the inner ear (Wersäll, 1956; Engström, 1958) gave the first clear demonstration of vesiculated nerve terminals in synaptic contact with the membranes of hair-cells. From their similarity with axon-dendritic synapses in the central nervous system they were thought to be efferent (Engström, 1958). Degeneration experiments have confirmed this (Engström & Fernandez, 1961). Subsequent ultrastructural studies on the hair-cells of other members of the acoustico-lateralis system including the lateral-line receptors (Hama, 1962; Flock, 1965, 1967) have also revealed vesiculated nerve terminals in synaptic contact with their membranes.
Electrical stimulation of efferent fibres to the cochlea (Galambos, 1956; Fex, 1962) inhibits impulse activity in afferent fibres in the eighth nerve. Recently the inhibitory nature of efferent fibres to the lateral-line organs has also been demonstrated in the clawed toad Xenopus laevis (Russell, 1968, 1970) and in the Japanese sea eel (Katsuki, Haskimoto & Yanagisawa, 1968). In Xenopus, however, stimulation of lateral-line efferent fibres, in the absence of calcium, failed to inhibit afferent impulses. Inhibition was also attenuated or blocked by the presence of D-tubocurarine chloride. These observations have promoted this investigation into the nature of the transmitter released at lateral-line synapses. Its object is to discover what evidence there may be that the transmitter is acetylcholine.
METHODS
Histochemistry
Selection of animals
This investigation was carried out on two species of animals: the South African clawed toad, Xenopus laevis, and a freshwater teleost, Acerina cernua. The investigations were not confined to Xenopus for the following reasons:
The area of sensory epithelium of individual neuromasts of the lateral-line receptors of Xenopus laevis is small. It occupies an area some 25 μm by 10 μm and containing on average thirty hair-cells. The elongate lateral-line receptors in Xenopus have been called stitches (Harris & Milne, 1966). This term will be used here.
Not all lateral-line stitches of Xenopus receive efferent innervation (Gömer, 1968), thus adding a complication to the interpretation of results. Acerina was chosen because of its large lateral-line organs, particularly those in the supra-orbital canal, which have sensory epithelia occupying an area of 1 mm2. Furthermore, a preliminary neurophysiological investigation of its lateral-line innervation demonstrated the presence of considerable efferent nervous activity.
Preparation of animals and histological methods
Young Xenopus (20–30 g) and Acerina (20 g) were pithed, and appropriate pieces of tissue containing lateral-line organs were removed. In the case of Xenopus these were pieces of skin containing single stitches, and in Acerina the sub-orbital bones containing the sub-orbital canals were removed and cut into short lengths. Each length contained a portion of the canal and a single canal organ. The pieces of tissue were washed for 1 h in isotonic sodium sulphate to displace as much chloride from the tissue as possible, since chloride interferes with the subsequent staining operations (Lewis & Shute, 1961). Frozen sections of the tissue were made by initially freezing the tissue in a mixture of solid CO2 and methanol at – 70°C. Subsequently the tissue was sectioned at 8 μm or 16 μm at – 20°C with an International Cryostat. The sections were picked up directly on to slides, thawed in formalin vapour, and fixed in 10% neutral formalin for 1–2 h. They were then washed in distilled water and treated for cholinesterase activity.
The cholinesterase method used was the modification of the Koelle & Friedenwald (1949) technique discussed by Lewis & Shute (1961). Serial sections of lateral-line organ were placed singly on consecutive slides numbering 1–5, and one slide was placed into each of five different incubating solutions for periods of 4–6 h. The first incubating solution (Table 1) contained 6 mm acetylthiocholine iodide, 9 mm copper, and 16 mm glycine at a pH of 5·5. The other incubating media (Table 1) contained the same ingredients as the first together with a specific inhibitor.
After incubation the slides were removed, washed for 5 min in several changes of distilled water, and placed for 2 min in a solution containing 1 g of sodium sulphide dissolved in 45 ml of N/5 acetic acid at pH 5·5. The sections were removed, rinsed in distilled water, dehydrated and mounted in Canada balsam.
Electrophysiology
The animals used in these experiments were Xenopus. They were anaesthetized in a 0·01 % solution of MS-222 (Tricaine), the brain anterior to the cerebellum was removed and the spinal cord was cut posterior to the second spinal nerve. Methods used for recording afferent impulses from single lateral-line receptors and for stimu-lating lateral-line efferent nerve fibres have already been described (Russell, 1970).
All pharmacological agents used in these experiments were made up in a saline solution (Russell, 1968) and were pipetted on to the undersides of the receptors and removed by suction. The amount of solution used in each experimental application was 100 ml. The following pharmacological agents were used: acetylcholine chloride, physostigmine salycate, acetyl-β-methyl choline chloride, propionylcholine chloride, butyrylcholine chloride, carbocholamine chloride, atropine, and D-tubocurarine chloride.
In experiments where the calcium concentration of the saline was altered or magnesium was added, the osmotic strength of the saline was maintained constant by appropriate alterations to the sodium chloride concentration.
RESULTS
Histochemistry
Acerina
Sections incubated with ethapropazine
Hair-cells and fine fibres of the lateral-line nerve stained intensely (Plate 1, figs. 1,2). Hair-cells in sections incubated for 4 h or more stained diffusely throughout the cell (Plate 1, fig. 1), whereas in sections incubated for only 2 h staining was confined to discrete points within or on the cell surface (Plate 1, fig. 3). No apparent difference existed between sections incubated with and without ethapropazine. Acetylcholinesterase staining was confined to the hair-cells and nerve fibres only, but bone in sections incubated for 3 h or longer stained un-specifically for cholinesterase. This effect was not inhibited by ethapropazine.
Sections incubated with 62C47 and neostigmine
Xenopus
A staining reaction for cholinesterase was confined to fine fibres in the trunk and branches of the lateralis vagus at the base of hair-cells in sections incubated in a solution containing ethapropazine (Plate 1, figs. 5, 6). It was observed that staining was absent from nerve fibres in sections incubated in a solution containing 62C47.
Pharmacology
The action of acetylcholine and related compounds on afferent activity
The frequency of spontaneous impulses from afferent lateral-line fibres was unaffected when acetylcholine in concentrations less than 10−4M was pipetted on to the undersides of the stitches which they innervated. But acetylcholine reversibly inhibited afferent impulses if an anticholinesterase, namely physostigmine, was pipetted on to the stitches 10–20 min before and during the addition of acetylcholine. Total inhibition of afferent impulses by 10−8M acetylcholine in the presence of 5 × 10−6M physostigmine is shown in Text-fig. 1. Physo-stigmine in the absence of acetylcholine occasionally caused a slight decrease in the frequency of afferent impulses (Text-fig. 1). Text-fig. 2 shows that the same result was obtained when the experiment was repeated with acetyl-β-methyl choline, which is a cholinergic transmitter substance with a high affinity for muscarinic receptors.
Butyrylcholine and propionylcholine, which are both cholinergic transmitter substances with affinities for nicotinic receptors, had no apparent influence on afferent impulses when added to lateral-line stitches even in concentrations as high as 10−4M and in the presence of physostigmine (Text-figs. 3, 4).
Carbocholamine chloride, a cholinergic compound with a high affinity for nicotinic receptors, which is not inactivated by acetylcholinesterase, was pipetted on to the undersides of lateral-line stitches. At a concentration of 5 × 10−4M it caused a reversible inhibition of spontaneous afferent impulses, but at 5 × 10−6M this inhibition was less marked (Text-fig. 5).
The action of acetylcholine on the afferent impulses was characterized still further by observing the influence of the cholinergic blocking drugs atropine and D-tubo-curarine on the pharmacologically induced inhibition. Atropine, which specifically blocks the action of acetylcholine at muscarinic synapses, blocked the action of acetyl-, choline on afferent impulses when it was applied to the lateral-line stitches 10–20 min beforehand in concentrations of 5×10−6M (Text-fig. 6). D-Tubocarine chloride, which tends to block the action of acetylcholine more easily at nicotinic than at muscarinic synapses, blocked the action of acetylcholine on afferent impulses in concentrations exceeding 5×10−5M (Text-fig. 7).
Impulse traffic in the lateral-line afferent fibres took some 7–10 min to respond to the action of the cholinergic drugs. This time delay was interpreted as a reflection of the relative inaccessibility of the lateral-line stitches to the drugs. The response time was noticeably shorter in young animals with thin skins.
The influence of physostigmine on inhibition
In order to examine the action of physostigmine on the inhibition of spontaneous activity of single lateral-line afferent fibres caused by the electrical stimulation of efferent lateral-line nerve fibres, it was decided to measure a parameter which would indicate the duration of the inhibitory effect. Physostigmine (an anti-cholinesterase) might be expected to increase the length of inhibition if the transmitter released from the efferent terminal was acetylcholine. The parameter chosen was the period d shown in Text-fig. 8, which is the time between the last of the efferent compound action potentials producing inhibition and the first of the spontaneous afferent impulses to reappear after the cessation of efferent inhibition.
Spontaneous afferent activity in nerve fibres from individual stitches was inhibited by trains of efferent impulses, each 1 s long, separated by 3 s, and containing impulses at a frequency of 120 s−1. A succession of ten consecutive trains of efferent impulses was delivered to each stitch once every 2 min. The period d was measured after each period of inhibition.
The addition of physostigmine to lateral-line stitches in concentrations above 1×10−5M increased the period d by factors up to 80%. For example, in the experiment illustrated in Text-fig. 8 the period d increased from 45 ms in saline solution to above 70 ms when physostigmine in concentrations of 5 ×10−5M was added. Throughout each experiment the resting frequency of afferent impulses remained constant and apparently unaltered by the addition of physostigmine.
Atropine and the inhibitory action of lateral-line efferent fibres
Spontaneous afferent impulses from single stitches were inhibited by trains of efferent impulses having the same parameters described in the previous section.
To assess the influence of atropine on efferent inhibition, the numbers of afferent impulses returning during each period of inhibition were counted before, during and after the addition of atropine in the saline solution bathing the underside of each stitch. Again a succession of ten consecutive trains of efferent impulses was delivered to each stitch once every 2 min for the duration of the experiment. The application of atropine in concentrations between 5×10−6M and 5×10−5M to the underside of the stitch caused an increase in the number of afferent impulses to appear during the inhibitory period, until eventually stimulation of the efferent lateral-line nerve fibres ceased to exert an inhibitory influence on the spontaneous afferent activity (Text-fig. 9). It was noticed that the block exerted by atropine was usually reversible, although recovery was slow and rarely complete (Text-fig. 9). Atropine also caused a slight increase in the frequency of spontaneous afferent activity.
Calcium — magnesium interaction
It was found that afferent impulses from lateral-line stitches were inhibited when the efferent fibres were electrically stimulated providing the stitches were bathed in a saline containing no magnesium but 1 · 8 mm calcium (Text-fig. 10a). Inhibition failed to occur when saline solutions containing 0 · 9 mm were used (Text-fig. 10b). These conditions were used in equation (4) to plot concentrations of magnesium and calcium in which stimulation of efferent fibres may be predicted to inhibit and not inhibit respectively (Text-fig. 11). The two curves were tested at points a, b, c, d, and a1, b1, c1,d1 and were found to be in agreement with the prediction.
In comparison with the efferent synapses, the afferent synapses of lateral-line hair-cells in Xenopus are insensitive to the presence of calcium or magnesium. In the experiments described above, spontaneous afferent impulses continued under conditions in which electrical stimulation of efferent fibres failed to inhibit afferent impulses (Text-fig. 11). Furthermore, under these conditions, the stitches were mechanically sensitive.
Impulses were recorded from afferent fibres to a stitch, and efferent fibres to the same stitch were electrically stimulated with trains of pulses in the usual way. The trains were delivered once every 2 s and caused inhibition. The stitch was then bathed in a saline solution containing no calcium or magnesium but 1 mm EDTA. Inhibition was soon reversibly blocked (Text-fig. 12) and the frequency of spontaneous afferent impulses was increased. Eventually the afferent nerve fibre itself began to produce bursts of impulses spontaneously at frequencies up to 600 Hz; but between bursts, spontaneous impulses of synaptic origin still occurred. Furthermore, the fibres remained sensitive to mechanical stimulation of the hair-cells. Both high calcium (10 mm) or high magnesium (10mm) concentrations (Text-fig. 13) reduced spontaneous afferent activity, but they did not abolish the responses of the afferent fibres to mechanical stimulation of the hair-cells.
DISCUSSION
Several criteria must be satisfied before a substance can be accepted as a synaptic transmitter. These criteria have been extensively reviewed (Paton, 1958; Curtis, 1961; McLennan, 1963). It is essential that the substance presumed to be the transmitter is released on nerve stimulation and that its pharmacology and its post-synaptic action are the same as those of the natural transmitter. Other criteria include evidence for the presence at the synapse of enzymes capable of the synthesis and inactivation of the presumed transmitter.
In identifying the transmitter released at the lateral-line efferent synapse several important criteria remain to be tested, but features of the lateral-line efferent synapse, which have so far been observed are characteristic of cholinergic synapses.
Acetylcholinesterase has been located at the base of hair-cells and in the fine nerve fibres of the lateral-line nerve. Although the presence of acetylcholinesterase is a requisite of cholinergic synapses, its presence at a synapse is not irrefutable evidence that the synapse is cholinergic (Lewis, Shute & Silver, 1967). Acetylcholinesterase has, however, been located on the pre-synaptic membrane of efferent synapses on cochlear hair-cells (Hilding & Wersäll, 1962) where acetylcholine is the presumed transmitter of the olivo-cochlear bundle (Fex, 1968). Localization of acetylcholinesterase in the cochlea was achieved using electron-microscope techniques. This degree of localization was not achieved with the light-microscopical techniques used here, but the localization of acetylcholinesterase in fine nerve fibres is in agreement with the known physiology, namely that efferent fibres conduct more slowly than lateral-line afferent fibres.
The inhibition of spontaneous afferent impulses by the addition of low concentrations of acetylcholine to the lateral-line receptors may be due to one of three causes. The excitability of the hair-cell membrane may be reduced, resulting in a drop in spontaneous afferent activity. The threshold for impulse generation at the afferent nerve terminal may be increased. Alternatively acetylcholine may be acting on specific receptor sites at the efferent synapse. High concentrations of acetylcholine are needed before any influence can be exerted on membranes without receptor sites specific for acetylcholine. Acetylcholine and carbachol in concentrations of 3 × 10−8 to 3 × 10−4 g ml−1 slightly decrease the amplitude of compound action potentials from mammalian myelinated nerve fibres (Armett & Ritchie, 1961). Complete inhibition of spontaneous afferent impulses is achieved in lateral-line nerves by bathing the undersides of the receptors in solutions containing 10−6M (2 × 10−7 g ml−1) acetylcholine or 10−6M acetyl-β-methyl choline. These concentrations are about three orders of magnitude less than those needed by Armett and Ritchie to produce a very small effect in conduction in unmyelinated nerve fibre. It is therefore proposed that acetylcholine is acting at specific sites on the hair-cell membrane. Furthermore, the specific sites have a higher affinity for muscarinic substances such as acetyl-β-methyl choline than for nicotinic substances such as butyrylcholine or carbachol. The specificity of action of acetylcholine is reflected in the actions of D-tubocurarine chloride and atropine on the receptor. The inhibition of spontaneous afferent impulses, caused either by tetanic electrical stimulation of efferent fibres or by bathing the receptor in low concentrations of acetylcholine, is blocked by lower concentrations of atropine (5×10−6M) than of D-tubocurarine (5×10−5M). Thus the actions of atropine and curare on the synaptic inhibition and on inhibition by applied acetylcholine are similar. The argument may again be put forward that atropine and D-tubocurarine chloride reduce the excitability of the efferent nerve terminal or, in the experiments where acetylcholine was applied to receptors, that they interact non-specifically with the acetylcholine. Atropine in concentrations of 10−3 g ml−1, but not 10−4 g ml−1 reduced the effect of acetylcholine on the compound action potentials elicited from mammalian unmyelinated fibres (Armett & Ritchie, 1961), whereas D-tubocurarine even in concentrations of 10−2 g ml−1 had no influence at all. It is unlikely that in the concentrations used, atropine and acetvl-choline were acting non-specifically on the lateral-line receptors. Instead it is proposed that D-tubocurarine and atropine are interfering with the transmitter mechanism of the lateral-line efferent synapses, either pre-synaptically or post-synaptically or both.
Atropine slightly increases the frequency of spontaneous afferent activity (Text-fig. 9) and it may be that there is a constant leakage of transmitter from the efferent synapse which reduced the frequency of spontaneous afferent activity. Atropine is also known to reduce the quantal content of end-glate potentials and miniature end-plate potentials at the neuromuscular junction (Beranek & Vyskocil, 1968) but in concentrations several hundred times greater than here. It is suggested therefore that in experiments described in this paper, the action of atropine is probably to interact competitively with the transmitter rather than to interfere with transmitter release.
The reason why it is essential to pre-treat each preparation with physostigmine before the effects of acetylcholine on lateral-line activity can be observed may well be due to the large amounts of non-specific cholinesterase present in the skin. In this connexion it is worth noting that before treatment with ethapropazine it was observed that blood vessels and many cells in the skin stained strongly for cholinesterase. It seems probable therefore that in the absence of an anti-cholinesterase small amounts of acetylcholine added to the preparation are hydrolysed before reaching the lateralline hair-cells. Physostigmine also prolongs inhibition of afferent impulses caused by electrical stimulation of efferent fibres. This might be expected if, as proposed, the transmitter is acetylcholine.
In one other aspect so far examined, the lateral-line efferent synapse is similar to cholinergic synapses. Synaptic transmission across the lateral-line efferent synapse is dependent upon calcium, and inhibition is incomplete or fails to manifest itself either in saline solutions containing low calcium or with magnesium added. At the neuromuscular junction magnesium and calcium compete pre-synaptically in such a way that three magnesium ions compete with one calcium ion for the same site (Dodge & Rahamimoff, 1967). As a guide for experimental procedure the ratio of 3 magnesium ions to i calcium ion was tried here with success. The possibility exists that magnesium blocks conduction in the efferent nerve terminal. Calcium and magnesium in high concentrations depress the excitability of muscle-fibre membranes (del Castillo & Engb ä ck, 1954; Katz & Miledi, 1964). High calcium concentrations, which abolish spontaneous afferent impulses in the lateral-line nerve, do not block the action of the efferent fibres on lateral-line organs. They still permit the occurrence of after-discharges (Text-fig. 13). It is unlikely therefore that concentrations of magnesium which do not block spontaneous afferent impulses should block conduction in the efferent nerve terminals. Instead it is believed that calcium and magnesium competitively interact for the release of transmitter at the efferent synapse, but the site of interaction is unknown.
Afferent impulses, both spontaneous and caused by mechanical stimulation of the hair-cells, are insensitive to concentrations of magnesium and calcium, which prevent transmission across lateral-line efferent synapses. The afferent impulses are believed to be pre-synaptic in origin, e.p.s.p.s have been recorded intracellularly from afferent nerve terminals in the vestibular apparatus of goldfish (Furukawa & Ishii, 1967). Furthermore, from electron-microscopical studies, afferent synapses are chemically transmitting (Flock, 1965, 1967). The transmitter has not, however, been identified. It may be that the calcium-free and high-magnesium salines did not reach the afferent synapses. This is most unlikely because the closely adjacent efferent synapses were blocked by the same salines (Text-fig. 12). Thus it appears that the afferent trans’ mitter is released under conditions unfavourable for the release of transmitter at the lateral-line efferent synapse or neuromuscular junction.
The small size of the hair-cells in lateral-line receptors of Xenopus and the small numbers of the efferent synapses have prevented the examination of two important questions. These are: (1) Is acetylcholine released by efferent nerve stimulation? (2) Are the post-synaptic events which accompany electrical stimulation of the efferent fibres and application of acetylcholine to the hair-cell the same?
In order to examine these two questions animals must be selected which have large numbers of efferent synapses and hair-cells capable of penetration by micro-electrode.
Several characteristics of the efferent synapses of the related cochlear hair-cells have raised the problem of whether one or more transmitter substances are involved. Fex (1968) has evidence that the synapses are cholinergic, but earlier work (Fex, 1962; Desmedt and Monaco, 1962) shows that strychnine and brucine, but not a related drug picrotoxin, block synaptic inhibition of the olivo-cochlear fibres. A theory to reconcile these two findings has been presented (McKinstry & Koelle, 1967), namely that more than one transmitter is released pre-synaptically. The influence of strychnine and related drugs on lateral-line efferent synapses has yet to be described.
ACKNOWLEDGEMENT
I am most grateful to Dr H. W. Lissmann for his encouragement and interest in this work and to Dr S. H. P. Maddrell and Dr R. W. Meech for advice and comments on the manuscript. This work was supported by a Research Studentship from the Science Research Council, and a Research Studentship from Trinity Hall, Cambridge.
References
EXPLANATION OF PLATE
The histochemical localization of acetylcholinesterase.