ABSTRACT
The effect of nerve lesions upon the dendritic branching of the metathoracic ‘fast’ coxal depressor motoneurone in the adult cockroach, Peri-planeta americana has been studied by intracellular cobalt injection. Dendrites may sprout after several nerve trunks entering the metathoracic ganglion are lesioned. Direct surgical damage to the neurone is not a pre-requisite for sprouting. New sprouts may extend into regions of the ganglion or into nerve trunks which the neurone would not normally occupy. The developing sprouts appear to grow preferentially within regions containing degenerating nerve fibres. The nerve sprouts frequently have varicosities along their length and at their ends.
INTRODUCTION
Knowledge of the cellular changes which occur after nerve damage may provide a deeper understanding of the events which lead to successful regeneration.
Several studies on populations of vertebrate neurones indicate that the extent of branching and diameter of dendrites can become reduced after axotomy (Cerf & Chacko, 1958; Sumner & Watson, 1971; Sumner & Sutherland, 1973); similar changes also occur after axotomy of identified neurones of the lamprey (Cohen, 1976). By contrast, axotomy can produce marked dendritic sprouting in an identified snail neurone (Murphy & Kater, 1978; 1980), while no dendritic changes were observed in axotomized motoneurones of adult cockroaches (Tweedle, Pitman & Cohen, 1973).
Removal of synaptic input, like axotomy can reduce the branching and diameter of mammalian neurones (Cowan, 1970; Jones & Thomas, 1962; Matthews & Powell, 1962). Denervation if performed early in postembryonic development, can also reduce the growth of specific dendrites of identified cricket interneurones (Murphey, Mendenhall, Palka & Edwards, 1975; Murphey, Matsumoto & Mendenhall, 1976). More recently, Hoy, Casaday & Rollins (1978) (see Anderson, Edwards & Palka, 1980) have shown that postembryonic denervation of cricket auditory interneurones not only causes reduced growth of one dendrite, but can also cause another dendrite to branch into the neuropile of the contralateral side of the ganglion, instead of remaining confined to the ipsilateral neuropile as it would in unoperated animals.
Wine (1973), however, found that crayfish motoneurone dendritic branches remained unaltered for well over a year after destruction of a presynaptic giant interneurone, and Tweedle et al. (1973) observed no dendritic changes up to 7 weeks after denervation of adult cockroach giant interneurones.
Thus, after axotomy or denervation, dendritic branching of neurones from a number of species may be altered: dendrites may sprout or retract. In at least some adult insects, however, dendritic branches have appeared to show relative stability; in this paper, we demonstrate that extensive dendritic sprouting can occur in an adult neurone following certain lesions affecting its neural environment.
MATERIALS AND METHODS
To produce experimental lesions, young adult male cockroaches Periplaneta americana L. were taken from the laboratory culture and immobilised with 100% CO2. Dissection tools were initially dipped in alcohol and flamed; the tips of scissors and forceps were dipped in gentamicin in insect saline (800 μg/ml) immediately before each use. To make lesions, a flap of cuticle over the metathoracic ganglion was lifted to allow selected nerve trunks to be sectioned; the cuticle was sealed down with low-melting-point wax after the operation.
At intervals after operation animals were decapitated; the nerve cord was removed and the desheathed metathoracic ganglion immersed in insect saline in preparation for intrasomatic injection of cobalt ions into the ‘fast’ coxal depressor motoneurone (numbered cell 28 by Cohen & Jacklet (1967) and abbreviated Df by Pearson & Iles (1970)). Occasionally, the experimental chamber was washed with destaining solution (Pitman, 1979) prior to an experiment; this minimized non-specific staining of ganglia that could result from contaminants such as silver (from electrodes) which accumulate in the chamber. Microelectrodes were filled with 100 mm cobalt chloride dissolved in 100 HIM potassium chloride and had a resistance of 10–20 MΩ). Cell bodies were impaled under visual control. After injection, preparations were treated with ammonium sulphide, fixed in 70% ethanol, immersed in destaining solution, washed in sodium thiosulphate and then block intensified as described previously (Pitman, 1979). Introduction of the destaining procedure before intensification had the advantage that it had no detectable effect on the stained neurone, but prevented the remainder of the tissue darkening during intensification.
RESULTS
The ‘fast’ coxal depressor motoneurone has been studied throughout these experiments and was identified by the following criteria, (a) The size and position of the cell body; this neurone is one of the largest in the metathoracic ganglion and has a characteristic position relative to tracheal landmarks. (b) The pattern of its dendritic tree; although new sprouts could alter the overall dendritic pattern, it is normally possible to discern the characteristic branches of this neurone, (c) The size and course of its axon; the axon of Df is one of the largest in the 5th nerve and runs in a characteristic position in the nerve trunk, finally passing into nerve 5r1 (Pipa & Cook, 1959). (This criterion was of limited use in experiments in which the motoneurone had been axotomized). Application of these criteria would bias the sample against those neurones which show the most dramatic responses to lesions, since such cells would be impossible to identify with certainty.
Fig. 1 A shows a camera lucida drawing of Df from a control animal. These animals were immobilized with carbon dioxide, opened up and the tracheal supply to the metathoracic ganglion partially disrupted to test whether surgery or damage to the tracheae was responsible for dendritic sprouting. The dendritic branching pattern is indistinguishable from that seen in normal, unoperated animals (Tweedle et al. 1973). Fig. 1 B shows the branching pattern of Df stained 21 days after nerves to the metathoracic ganglion had been lesioned as shown in the inset; note that the axon of this neurone had not been severed. Branches of this neurone can be seen passing into the ipsilateral anterior connective.
Dendritic sprouts observed in neurones from lesioned animals showed the following features :
(A) Origin of outgrowth
New sprouts usually originate at or near the ends of dendrites which form part of the normal dendritic tree of the cell, although in some cases, abnormal sprouts were seen originating from the neurite (the nerve process which connects the cell body to the axon).
(B) Extension outside normal territory
Supernumerary branches can travel into damaged connectives (Fig. 1 B) or segmental nerve trunks (Fig. 1 C) which would not normally contain processes of this neurone. Most of the new branches are normally confined to the neuropile ipsilateral to the neurone cell body, but occasionally cross the midline and extend some distance into the contralateral neuropile (Fig. 1 D).
(C) Tortuous processes
Nerve sprouts often travel in a more or less convoluted manner close to their origin, sometimes forming loops or doubling back upon themselves (Fig. 2A).
(D) Straight processes
Sprouts can quite suddenly begin to travel relatively directly towards a connective or segmental nerve trunk (Fig. 2A). One possibility is that the growing sprouts may meet and follow tracts of degenerating nerve fibres. When this occurs, a number of fibres may run more or less parallel to each other and branch along their length (Fig. 2B).
(E) Varicosities
New neuronal processes frequently have varicosities along their length or at their ends (Fig. 2B and 2C). Those along the length of nerve sprouts frequently give a beaded appearance, while those at the ends are larger and more discrete. At the present time we do not know the functional significance of these enlargements but they may be associated with the dynamic state of the neurone since dendrites of mammalian (Purves, 1975), molluscan (Murphy & Kater, 1978), insect (Clark, 1976; Truman & Reiss, 1976) and leech (Miyazaki, Nicholls & Wallace, 1976) neurones show an increase in the incidence of varicosities under conditions of physiological change.
Table 1 summarizes the frequency with which dendritic sprouting was observed after different lesions.
DISCUSSION
At present the stimuli for dendritic sprouting are unknown. The most obvious candidates are : (a) loss of presynaptic or postsynaptic contacts, (b) materials released into the local environment by other damaged neurones, or (c) glial changes associated with damage. Whether one or more of these factors are responsible for stimulating this sprouting, our observations indicate that the extent of new outgrowth is dependent on the overall degree of damage in the local neural environment rather than upon direct surgical damage to the neurone itself; sprouting was greatest when a number of ganglionic nerve trunks had been damaged, and was almost invariably absent after section of the 5th nerve alone (Tweedle, et al. 1973). Neurones of isolated ganglia maintained in organ culture showed an even higher degree of sprouting than those in ganglia which had been lesioned in the manner described above (Pitman & Rand, 1981).
Changes in dendritic branching of adult insect neurones have been reported in one investigation; supernumerary dendritic branches were observed on parts of a neurone isolated from the soma (Clark, 1976). It was suggested that this sprouting might result from loss of a somatic influence which normally represses dendritic sprouting. The observations reported here suggest that supernumerary sprouting in soma-less dendritic trees may instead result from general neural damage.
The results presented here show that dendrites of an adult insect motoneurone retain the ability to sprout after central nervous system lesions. Such sprouting occurs even after lesions which do no surgical damage to any part of the neurone. Branches can grow into regions of the nervous system not normally occupied by that neurone. Experiments are in progress at present to determine whether supernumerary dendritic sprouts persist once formed or whether they will eventually retract to restore the normal branching pattern of the neurone.
ACKNOWLEDGEMENTS
We thank Dr J. F. Aiton for reading previous drafts of this manuscript.