Histological and electrophysiological Studies On The Giant Axons Of The Cockroach Periplaneta Americana

In their studies on the organization of the nervous system of the American cockroach, Periplaneta americana, Pumphrey & Rawdon-Smith (1937) concluded that some of the ascending giant fibres (AGF) in the ventral cord originate in the cerci and ascend to the suboesophageal ganglion (S₀) without any synapse. However, Roeder (1948) found that the cereal sensory nerves synapse at the 6th abdominal ganglion with the AGF. The latter then ascend uninterruptedly to their termination in the metathoracic ganglion (T8). From here, a polysynaptic pathway continues to the head. Hess (1958) confirmed these results by degeneration experiments. Sectioning the cereal nerves did not cause degeneration of abdominal giant axons, whereas severance of the cord at the connective between the 5th and 6th abdominal ganglion (A₅-A₀) resulted in the degeneration of the abdominal GFs up to the level of T3. However, two of the AGFs did not degenerate even after 30 days. Hess therefore suggested that these two axons may subserve descending impulses or may even represent a common ascending/descending pathway. No degeneration of GFs was detected by Hess in the thoracic portion of the nerve cord. Pipa, Cook & Richards (1959) showed in their histological studies that the ventral giant fibres could be traced along the thoracic connectives. However, they did not perform degeneration experiments and hence did not exclude the possibility of a polysynaptic chain.

Hess (1958) also identified small degenerated fibres both in abdomen and thorax. This led to the assumption that small fibres may ascend continuously from A₆ to the head.

In the first paper of this series (Spira, Parnas & Bergmann, 1969) it was shown that the giant fibres conduct between A₆ and S₀ bi-directionally. Two schemes were considered to explain these electrophysiological findings :

• (1) The GFs are continuous from A₆ to S₀. In this case, the descending potentials might represent an unphysiological antidromic conduction. This model does not necessarily contradict the results of Hess (1958) since a ‘continuous* pathway may be composed of several cells, connected by septa. Such an arrangement would permit bi-directional conduction, but would at the same time explain the absence of degeneration, rostrally to T3.

• (2) At each thoracic ganglion, a double axon/axonal synapse is found, one sub-serving ascending conduction, and the second for descending impulses. Here again, the same pathway would conduct in both directions and the degeneration process, following abdominal cuts, would not extend beyond T3.

Although the second scheme can account for the marked susceptibility of the thoracic portion of the pathway to the blocking action of nicotine, it does not explain why the alkaloid affected ganglionic conduction in the order T1-T2-T3 (Spira et al. 1909).

In the present experiments, the validity of either model was studied with histological and electrophysiological techniques in normal and degenerated preparations. It will be shown that only the first model fits the experimental results of the present paper and that the ‘ganglionic blockade’ by nicotine can be explained in an entirely different way.

The methods of preparation and recording were described in a previous paper (Spira et al. 1969). For degeneration studies adult cockroaches of either sex were anaesthetized with ether and the nerve cord was cut with fine scissors between the 5th and 6th abdominal ganglia. The animals were dissected after different time lapses and the ventral nerve cord was isolated for electrophysiological studies or fixed in Bouin-Holland sublimate solution, embedded in paraffin, sectioned serially at 7 µ, and stained with haematoxylin-eosine.

As described by Roeder (1948) and Hess (1958), the giant axons are arranged mainly in two groups. In each connective there is a dorsal group, composed of three large axons diameter (25–30µ) and a smaller one of 15 µ, and a ventral group of 4-5 axons (ranging from 15 to 60 µ), arranged in a typical way (Plate 1 a). Separation of the two groups is constant all along the abdominal connectives.

The separation of the axons into dorsal and ventral groups is even more pronounced in the abdominal ganglia. Plate 1b shows that the axons, as they pass through the ganglion, narrow to 5µ, as already described by Roeder (1948) and Hess (1958). The change of the diameter was more or less the same in each of the abdominal ganglia.

It is possible to follow the course of the dorsal and ventral groups as they enter the metathoracic ganglion. The separation of the two groups is quite clear in the posterior part of the ganglion (Plate 1 c). However, in central and anterior sections of the same ganglion (see Plate id) the identification of these two groups is more difficult, but nevertheless feasible.

In accordance with Pipa et al. (1959), 4-5 ventral GFs (7-27 µ) could be clearly distinguished in the thoracic connectives. This arrangement is especially clear in the posterior sections of the T₂-T₃ connective (Plate 2 a) but becomes less distinct in anterior sections of the same connective (Plate 2b).

On the other hand, it is impossible to identify any longer the dorsal group of GFs, because many more dorsal giant fibres are present in the thorax than in the abdominal parts of the nerve cord. It is thus quite possible that the dorsal GF group of the abdomen terminates sharply at T₃ or that these axons narrow sharply at ganglion T₃.

Upon first inspection of Plate 2 c, it may be concluded that the ventral GFs have disappeared. However, in the ventral area of the cord a distinct group of fibres shows the same arrangement as the ventral GFs in more caudal parts of the cord, while their diameter is greatly reduced (7–15 µ instead of 15-60 µ as in the abdomen).

GF degeneration could be observed as early as 4 days after division of the nerve cord at A₅-A₆, but not all axons collapsed at the same rate. Thus in Plate 2d the connective on the right side of the picture shows complete absence of the dorsal group and survival of two ventral fibres, while on the left side only a single giant axon is still present. This shows that the degeneration process is not strictly synchronized and sometimes proceeds more rapidly on one side. However, after 10–15 days all abdominal GFs had disappeared (Plate 3 a, right connective).

Here the degeneration process assisted greatly in the identification of fibres that ascend from the abdominal cord. In view of the variability of the localization of the ventral group in thoracic sections it was expected that unequivocal localization of the ventral GFs would be greatly facilitated by unilateral degeneration. However, cutting of only one connective at A₅-A₆ presented considerable technical difficulties. A very fine blade was inserted laterally to the abdominal nerve cord and was moved cautiously towards the ventral mid-line. After may trials, the operation was finally successful in one animal. When this specimen was dissected 60 days after the operation, it was found that both dorsal and ventral abdominal GFs had degenerated on one side only (Plate 3 a). In addition, sections at T₂-T₃ (Plate 36) and at T₂-T₃ (Plate 3 c) clearly showed an obliterated strand on one side (single-headed arrow) and thus permitted identification of the ventral GFs in the intact connective. Furthermore, a similar degenerated band was present in the dorsal part (double-headed arrow) indicating that here too continuous fibres ascended from the abdomen, although the corresponding contralateral GFs could not be recognized beyond doubt. Thus Plate 3a—c furnishes evidence that at least the ventral GFs all ascend uninterruptedly from Ae to Tr Based on this experience one can easily recognize degeneration. In such preparations, collapsed axons could be identified in the thorax as early as 2 days after the operation (Plate 3d). Plate 3a-c reveals that unilateral transection leads only to degeneration on the same side. This serves as evidence that no crossing of giant fibres takes place along the whole nerve cord, a conclusion in accordance with the electrophysiological experiments described in the first paper of this series (Spira et al. 1969).

Four to 15 days after the connectives were cut at A₅-A₆, the nerve cord was isolated and fitted with stimulating electrodes at A₄-A₆ and with recording electrodes from A₁-T₃ to SO-T₁, as illustrated by the inset in Text-fig. 1. Control experiments were carried out with nerve cords from intact animals. In all cases, conduction time was plotted as function of electrode distance.

The curves, characteristic for conduction in intact fibres (Text-fig. 1, full lines), are slightly bent upwards, indicating a progressive decrease in conduction velocity. This may be explained by the decrease of fibre diameter towards the head. The broken fines, representing degenerated preparations up to 15 days after transection, show a marked delay at the thoracic recording electrodes, while for the abdominal portion the curves do not differ significantly from the intact connectives. Since degenerated fibres cannot conduct at all, the data of Text-fig. 1 suggest that the signals recorded above T₃ originated from small fibres with a much slower conduction velocity. This assumption is in accordance with the observation that 20 days or more after cutting at A₅-A₆, the conduction in the abdominal cord also became much slower while the amplitude of the signals decreased (Text-fig. 1, crosses). Here again it appears probable that after complete degeneration of the GFs the impulses were conducted by small fibres.

Roeder (1948) has shown that stimulation of the cereal nerves causes repetitive firing at the crural nerves. He therefore assumed that the ascending giant axons are connected through at least one synapse to the crural motor neurones. In confirmation, when the nerve cord was cut in the abdomen, no signs of degeneration were found in the crural nerves (Hess, 1958). The same holds true for the abdominal giant fibres when the cereal nerves are cut. On this basis Hess excluded the possibility of transneuronal degeneration.

In the present study we observed that when the nerve cord was cut at A₅-A₆, the dorsal giant groups degenerated first. However, unlike Hess (1958), we found that all ventral fibres degenerated, although two of them did so more slowly. Since these were the fibres of largest diameter, the rate of degeneration may be related to fibre diameter. Similar results were obtained recently by R. Farley & N. Milbum (personal communication).

The present findings clearly demonstrate the continuity of the ventral group from A₆ to T₁. The giant fibres retain their large diameter (15–60µ) throughout the abdomen, but in the thorax they narrow gradually until they reach 6–15 µ at T₁-T₂. On the other hand, unequivocal identification of the continuation of the abdominal dorsal GFs in the thorax was impossible. It cannot be excluded that some of them terminate at T₃, but from the dorsal degeneration band in thoracic sections (Plate 3 b–d) it is clear that at least part of the dorsal GFs run uninterruptedly through the thoracic cord.

In the first paper of this series (Spira et al. 1969) two models were suggested to explain the interaction of spike responses, evoked from S₀-T₁ and from the cereal nerves. The present degeneration studies contradict the model, based on the assumption of double axon/axonal synapses, since such a structure would not permit degeneration to proceed beyond T₃. We thus arrive at the conclusion that all experimental data are best explained by the continuity of the GFs from A₆ to S₀. The question arises whether the bi-directional conduction in these fibres is an artificial phenomenon, the descending potentials representing in fact antidromic impulses which cannot occur under physiological conditions.

If the pathway from A₆ to S₀ is continuous it is difficult to understand why GFs at the thorax are blocked by very low concentrations of nicotine, similar to those inhibiting synaptic transmission at Ag (Spira et al. 1969). A possible answer to this problem may be derived from the fact that the abdominal giants can also be blocked by nicotine, but in much higher concentrations. The vulnerability of fibres to the alkaloid presumably depends on fibre diameter. The ascending GFs have their smallest diameter at S₀-T₁ and therefore it is here that nicotine becomes effective most rapidly, while the connectives T₁-T₂ and T₂-T₃ are blocked later. In view of these considerations, it may be asked why nicotine has no influence on the narrow parts of the GFs in the abdominal ganglia. Apparently, the drug penetrates only with difficulty through the compact structure of the ganglia. If these views are correct, then conduction block by nicotine cannot be used to localize synapses in the nervous system of insects.