The Thoracic Gland in Rhodnius Prolixus (Hemiptera) and Its Role in Moulting

The source of these active secretions appeared to differ in different insects. The original view of Kopeć (1922) that in Lepidoptera the secretion came from the brain was confirmed in Rhodnius (Wigglesworth, 1940) when it was shown that the neurosecretory cells of the pars intercerebralis were the apparent source of the ‘moulting hormone’. But about the same time Fukuda (1940, 1941, 1944) found that in both the larva and pupa of the silkworm the prothoracic gland was the immediate source. So far as the Lepidoptera are concerned this controversy has been resolved by Williams (1947, 1948 a), who has proved that secretions from the neurosecretory cells in the brain are necessary to activate the prothoracic glands.

In cyclorrhaphous Diptera it has long been recognized that the hormone causing pupation is produced in the ring gland of Weismann (Burtt, 1937, 1938; Hadorn, 1937) and almost certainly in the large lateral cells of this gland (Vogt, 1943) which are now commonly regarded as homologous with the ‘pericardial glands’ (or prothoracic glands) of other insects (Thomsen, 1941). It has recently been proved by Possompés (1950) that these lateral cells (called by him the ‘peritracheal gland ‘) are induced to secrete the pupation hormone by a factor liberated from the brain.

In Dixippus, ‘ventral glands’ and ‘pericardial glands’, again perhaps homologous with the prothoracic glands (Williams, 1948b, 1949; Pflugfelder, 1949), have been thought on histological grounds to be concerned in the control of moulting (Pflugfelder, 1947). Experiments on Odonata make it appear probable that the ‘ventral glands’ or ‘intersegmental glands’ are concerned (Deroux-Stralla, 1948). And in Sialis the moulting hormone appears to come from some centre in the thorax (Geigy & Ochsé, 1940). In Gryllus the brain is necessary for moulting (Sellier, 1949), but the histological changes in the prothoracic glands (Sellier, 1951) show that these also are probably concerned. In Periplaneta the immediate source of the ‘moulting hormone’ is the prothoracic gland (Bodenstein, 1953).

In the light of all these observations it has seemed probable for some time that a ‘two-stage’ control of moulting, neurosecretory cells activating a thoracic gland, might well prove general among insects. The purpose of this communication is to confirm the existence of such a system in Rhodnius prolixus^*

It was shown earlier (Wigglesworth, 1940) that 4th-stage larvae of Rhodnius decapitated at 24 hr. after feeding can be induced to moult by implanting into the abdomen the dorsal region of the brain from larvae (either 4th stage or 5th stage) which have just passed the critical period, that is, at about 10 days after feeding in the 5th stage, about 7 days after feeding in the 4th stage (Fig. 1B). In the present series of experiments this result was obtained in sixteen (80%) out of twenty survivors.

Similar implantations were made into larvae 24 hr. after feeding, and the abdomen was then isolated by tying a firm ligature round the metathorax; the anterior part of the body was removed and the wound sealed with paraffin wax (Fig. 1C). The solated abdomen treated in this way fails to moult; there were no positive results in twenty-three experiments. Many of the body fragments remained alive for more than a month, but in none of them was there any sign of moulting even beginning. That is, there were none of the nuclear changes of ‘activation ‘and mitosis which precede the deposition of a new cuticle (Wigglesworth, 1933).

It would appear, therefore, that the neurosecretory cells in the brain probably act, as in Lepidoptera, through the mediation of a secretory centre in the thorax.

The presence of a thoracic gland in Rhodnius can be demonstrated as follows. The tergites of the mesothorax and prothorax of the 4th-stage or 5th-stage larva are removed by cutting along the margins and severing the muscles. The arrangement shown in Fig. 2A is then seen, with the cut muscles on either side, the green dorsal vessel in the mid-line and a mass of fat-body lying longitudinally below it.

The dorsum of the metathorax and anterior segments of the abdomen are now cut away. The mass of fat-body can then be teased apart and is seen to consist of outer and inner lobes (Fig. 2B). The outer lobes are poorly supplied with tracheae and are extremely friable. The inner lobes are spindle-shaped, tapering to a point behind. They are richly supplied with tracheae which spread over their lateral aspects. The entire structure is much less friable than the outer lobe and can readily be dissected out without breaking up.

The tracheal supply comes from the anterior extremity, from the ventro-lateral aspect about one-third of the distance from the neck, and particularly from the tapering posterior extremity which is tied by tracheae to the salivary glands, to the narrow commencement of the mid-gut and to the ‘stomach’. The ‘retort-shaped’ organs of the stylets extend for a variable distance into the base of the inner lobes.

It is not possible to recognize the thoracic gland in such dissections; but if the inner lobe of the thoracic fat-body is isolated and stained with haematoxylin it can be seen that there is a network of cells with very large nuclei spread out over its lateral aspect (Fig. 3). This is clearly the thoracic gland ; it has characters resembling those of the prothoracic gland in Lepidoptera (Williams, 1948b). It is abundantly supplied with tracheae ; the rich tracheal network is limited to the superficial layer of the inner lobe and to those parts of it on which the thoracic gland is spread. Fig. 4 shows a part of the gland stained with haematoxylin after injection of the tracheal system with cobalt sulphide (Wigglesworth, 1950). The extremely rich tracheal supply to the gland is in striking contrast with the very sparse supply to the cells of the fat-body. (During moulting the ‘retortifonn organs ‘of the developing stylets extend backwards and enlarge until the inner lobes of the fat-body become thin-walled sacs in which they are enclosed. The retortiform organs will thus benefit from the rich tracheal supply to the surface of the lobe ; but it seems clear that this supply is connected primarily with the thoracic gland, for it runs only to that part of the surface that is occupied by the gland, and this comprises no more than one-sixth or less of the total surface when the lobe is fully distended.)

Fig. 5 A shows a transverse section through the thorax of the 4th-stage larva at the level of the mesothoracic spiracles to show the relation of the thoracic gland cells, deeply embedded in the fat-body, to the other organs. Fig. 5B shows the detail of the gland cells.

The prothoracic gland in Lepidoptera is well supplied with nerves (Lee, 1948; Williams, 1948 b), and that of the cockroach Leucophaea is innervated from the prothoracic ganglion (Scharrer, 1948a). No nerves could be seen in sections or dissections of the thoracic gland in Rhodnius. A number of 4th-stage and 5th-stage larvae were injected with methylene blue or methylene blue and rongalite, but even where the nerves to the adjacent muscles, to the salivary glands and to the gut were well stained no trace of any nerve supply to the thoracic gland could be seen. It is still possible, however, that the nerves have been overlooked.

The thoracic gland has been studied in sections and in whole mounts throughout the moulting cycle in 4th-stage and 5th-stage larvae. In the unfed larva (Fig. 6A) the cells are shrunken and pale-staining, the nuclei are elongated and smooth in outline, and there are a few scattered haemocytes present.

After feeding, the cytoplasm gradually increases in extent and in depth of staining, and the nuclei become greatly enlarged and lobulated, and great numbers of haemocytes are applied to the surface of the gland. These changes reach their peak during the ‘critical period’ 7-12 days after feeding (Fig. 6B). Thereafter the cells and nuclei slowly diminish in size, and by the time of moulting they have almost reverted to the resting condition.

When the 4th-stage larva has moulted to the 5 th stage the thoracic gland remains in the resting condition until the new moulting cycle is initiated by a meal of blood. In the adult insect, on the other hand, the gland cells quickly break down and disappear.

In the newly moulted adult the cells are shrunken but the nuclei are intact. At 24 hr. after moulting almost no healthy nuclei remain; everywhere they are breaking down with the liberation of innumerable chromatic droplets (just like those in the epidermis during moulting (Wigglesworth, 1942)) and among them are numerous haemocytes (Fig. 6C). At 2 days after moulting the chromatic droplets are reduced in number ; and at 3 days after moulting they have virtually disappeared, no vestiges of the thoracic gland remain, and the haemocytes are relatively sparse.

Implantation of the fat-body, removed from Rhodnius larvae just after the critical period, into larvae decapitated at 24 hr. after feeding, will not induce moulting (Wigglesworth, 1940). This has been confirmed. But if the inner lobes of the thoracic fat-body carrying the thoracic gland are removed from 4th-stage or 5th-stage larvae which have just passed the critical period and implanted into the abdomen of 4th-stage larvae decapitated at 24 hr., these are caused to moult. In the first trial there were six positive results in eighteen experiments.

If these lobes are implanted into larvae ligatured through the metathorax (Fig. 1 C) the isolated abdomen is likewise caused to moult within about 1 month. There were seven positive results among thirteen survivors.

As in Lepidoptera the thoracic gland appears to be the source of the secretion immediately responsible for initiating growth and moulting. Implantation of the dorsum of the brain is effective in inducing moulting only if thoracic glands are present.

Adult Rhodnius can be made to moult by joining them to 5th-stage larvae which have passed the critical period (Wigglesworth, 1940). If the above interpretation is correct the adult should not be caused to moult by implantation of the brain, since it lacks thoracic glands. That has proved to be the case : ten adult female Rhodnius (about 2 weeks after moulting, 24 hr. after feeding) each received implants of two brains removed from 5th-stage larvae 10 days after feeding. None moulted; egg development and oviposition occurred normally.

On the other hand, of ten adult Rhodnius each receiving implants of thoracic glands removed from 5th-stage larvae 9 days after feeding, four moulted ; that is, they developed a new cuticle preliminary to moulting. This new cuticle was in all cases exceedingly thin and delicate; but this was probably because the development of eggs in the females and of the accessory glands in the males was not inhibited by the process of moulting; consequently, the food material which might have been used for cuticle formation was diverted for reproductive purposes.

Another series of eleven adults into which the thoracic glands from 5th-stage larvae had been implanted were therefore decapitated in order to suppress the activity of the reproductive system (Wigglesworth, 1936). Of these, four moulted and three of them laid down cuticles of normal thickness.

In order to see whether secretion from the brain of the adult Rhodnius would serve to activate the thoracic gland if such were present, twenty adult Rhodnius of mixed sexes, about 2 weeks after moulting and 24 hr. after feeding, received implants of thoracic glands from unfed 5th-stage larvae. Mating and oviposition occurred normally; none showed any signs of moulting. It would appear, therefore, that the absence of moulting in the normal adult Rhodnius is due both to the absence of the thoracic glands and the failure of the appropriate activity in the secretory cells of the brain.

It is to be noted that when the thoracic glands of the moulting 5th-stage larva are implanted into the adult both moulting and egg development take place. On the other hand, it was previously observed (Wigglesworth, 1940) that when the adult female Rhodnius is caused to moult by joining it to two moulting 5th-stage and two moulting 4th-stage larvae with brain, corpus cardiacum and corpus allatum intact, the formation of eggs was arrested.

This result has been confirmed in further experiments in which nine adult females with their heads intact were caused to moult by joining to each of them two 5th-stage larvae with the tip of the head removed at 10 days after feeding. None of them showed any egg development, in spite of the fact that ‘juvenile hormone’ was being secreted by the corpus allatum of the adult females themselves (Wigglesworth, 1948), and caused the 5th-stage larvae to develop partial larval characters when they moulted. This suppression of egg development in the presence of the intact brain and retrocerebral organs of the 5th-stage larva is at present unexplained.

Scharrer (1948a) and Pflugfelder (1949) have suggested that the ‘tracheal organs’ described by Hamilton (1931) (among others) in Nepa may correspond with the prothoracic glands of other insects. But these are usually regarded as vestigial flight muscles ; they do not consist of glandular cells ; and they persist in the adult. A reexamination of Nepa has revealed the presence of a very distinctive glandular organ in the thorax of the larva, not mentioned by Hamilton, which is completely absent in the adult. This organ, and the thoracic glands in certain other Hemiptera, will be described elsewhere.

The only other reference to thoracic glands in Hemiptera is by Williams (1949), who records that they have been observed by Edwards (unpublished) in the Lygaeid Oncopeltus.

In the light of the experiments described in this paper and the accumulated evidence from other groups of insects as set out in the introduction, it appears that in many, perhaps all insects, the ‘moulting hormone’ is composite and consists at least of (a) an ‘activating factor ‘from secretory cells in the brain and (b) a ‘moulting factor ‘from the thoracic gland (prothoracic gland, ventral gland, pericardial gland, peritracheal gland).

No generally accepted terminology for these factors yet exists. In earlier papers dealing with Rhodnius the complex has been referred to as the ‘moulting hormone ‘because the terminal event in the process of growth which it brings about is the moulting of the old skin. Scharrer (1948a, b) refers to it as the ‘growth and differentiation hormone’, since she conceives this complex as being responsible for the differentiation of imaginal (or pupal) characters. The same point of view is adopted by Piepho (1942).

Certainly these secretions are ‘growth hormones’—though I am inclined to use this expression as a comprehensive term embracing all those humoral factors concerned in regulating growth (‘activating factor’, ‘moulting factor’, ‘juvenile’ or ‘inhibitory’ hormone, etc.).

On the other hand, it is perhaps misleading to describe this complex as a ‘differentiation hormone’. Differentiation is the realization of potential capacities for growth hitherto latent in the different body cells. Looked at from the present point of view, insects have dual potentialities : larval and imaginal. The choice between these is controlled by the amount of the juvenile hormone secreted by the corpus allatum. The insect cannot moult without differentiating either in the larval or the imaginal direction. It is characteristic of the hormone from the thoracic gland that it merely initiates growth and moulting—the type of differentiation is determined by other factors, partly humoral, partly inherent in the cells.^*

One further point which calls for comment is the very rapid disintegration of the thoracic gland immediately the 5th-stage Rhodnius becomes adult. There must be some abrupt change in the internal environment which is responsible for this breakdown, but the cause of it has not yet been discovered. The similar though more gradual breakdown of the prothoracic gland in the adult Periplaneta has been shown by Bodenstein (1953) to depend upon a change in the corpus cardiacum of the adult.

The ‘moulting hormone ’in Rhodnius is composite. The factor secreted in the dorsum of the brain activates a gland in the thorax which then produces the factor initiating growth and moulting. Implantation of the thoracic gland will induce moulting in the isolated abdomen ; implantation of the brain is effective only if the thorax is intact.

This system agrees with that described in Lepidoptera and Díptera and is probably widespread in insects.

The thoracic gland in Rhodnius consists of a loose network of very large cells, richly supplied with tracheae, spread as a single diffuse layer over the surface of the inner lobes of the thoracic fat-body. These cells go through a cycle of secretory activity which reaches its peak during the critical period. They break down and disappear within 2 days after the insect becomes adult.

The adult Rhodnius is caused to moult by implantation of the thoracic gland from a moulting larva; it is not caused to moult by implantation of the brain.