A Histochemical Study of the Pigment Cells of the Leech, Glossiphonia complanata

Many authors since Cuénot (1891) have described the large pigment cells which occur deep in the connective tissue of Glossiphonia complanata and other rhynchobdellid leeches. Detailed studies have been published by Graf (1899), Scriban (1910), Abeloos (1925), and Bobin (1950), but a histochemical study by modern methods does not seem to be available.

It is the purpose of this paper to give the results of such a study, both of the large deep-seated pigment cells, and of the smaller pigment cells which are found in the connective tissue immediately below the epidermis.

Specimens of G. complanata were collected in the Oxford region and kept in the laboratory until required. Much of the work was done on sections of fixed material because of the difficulties in obtaining separate living cells (Bradbury, 19566); details of the histochemical tests are not given in the paper but are set out in an appendix. As well as in situ histochemistry, chemical and spectroscopical studies were made of pigment solutions ex-tracted from the leeches.

These cells occur in the deeper layers of the connective tissue. Bobin (1950) has shown that they are found in the region between the base of the pharynx and the ventral sucker, with especial concentrations around the caeca of the stomach and intestine. The cells appear round or oval in section, often with a diameter of over 100 p. Bobin succeeded in isolating some of these cells and showed that they were globular or pear-shaped. During the present study it proved possible, by the use of gentle maceration, to separate the gut and much of the adjacent connective tissue and mount them on a slide. The large pigment cells were clearly visible and their shape and size agreed with the account given by Bobin.

In many of the sections studied, very little ground cytoplasm was seen in these cells, as the whole body of the cell appeared packed with very many pigmented spheres (fig. 1, A, B). The spheres are very regular; they mostly have a diameter of about 3 p. The pigment contained in or on these spheres is yellowish-brown or green when seen in unstained sections. The cells rarely show a nucleus, but when one is included in the plane of section it is seen to be oval, about 25/4 by 12 p, with a very prominent nucleolus approximately 5 p in diameter. Occasionally a nucleus is found with a very irregular outline ; it appears to be pushed up against one of the cell-walls by the pigmented spheres. It is very probable that these represent the ‘secondary nuclei’ noted and figured by Bobin in her recent paper. From the study of serial sections she concludes that the adipose cells may under some circumstances transform into these pigment cells by the gradual accumulation of the pigmented spheres in their cytoplasm. Bobin gives a schematic representation of such a cell cut in perfect longitudinal section, showing the transformation of the cytoplasmic structures typical of an adipose cell to those characteristic of the pigment cell. In the present work some sections cut longitudinally and coloured with Sudan black showed pigment cells with a very similar appearance to this ideal. At one end were large fat drops coloured intensely by the Sudan black; these graded into an intermediate zone with fat drops among pigmented spheres, whilst at the other end of the cell only pigmented spheres were seen. One such cell is shown in fig. 1, c. These observations may be interpreted as supporting Bobin’s idea that there is no sharp distinction between the adipose cells on the one hand, and the pigment cells on the other.

The presence of ‘residual lipid’ as noted by the same author was confirmed. This lipid will colour with Sudan black and often appears as irregular masses. Mitochondrial techniques show the mitochondria to be rod-like, with a length of about 7 /x and a diameter of 1 p. They tended to be scattered around the periphery of the cell, but were not very numerous.

Some cells were noted beneath the dorsal surface of the body which resembled the large pigment cells but were smaller and contained many extremely small pigment granules. The latter appeared rather yellowish in colour and were never larger than ip in diameter. Few observations were made on these cells, but they seem to correspond to the ‘cellules blanches’ noticed by Bobin (1950, pp. 73, 78). It would appear from the histochemical studies that this pigment differs significantly from that in the larger and more deeply situated pigment cells, but closely resembles that in the subcutaneous pigment cells.

In most standard histological preparations, the pigmented spheres in this cell do not take up either acid or basic dyes, but appear in their natural yellowish colour. After fixation in Zenker’s fluid, however, the spheres will colour feebly with dye lakes such as that of haematein, thus suggesting that they possess some basiphil components ; this property was noticed by earlier workers. The basiphilia may be due to the pigment itself, or more probably is a property of some substrate to which the pigment is bound.

Lipids. Details of the tests for lipids which were used in this work are given in the appendix. It is seen that no positive result could be obtained with Sudan black or Sudan IV, either at room temperature or at 6o° C. This makes it seem very unlikely that there is any free lipid in the pigmented spheres. Positive reactions were obtained in some instances from small, rather irregular masses situated among the pigmented spheres (the ‘residual lipid’ noted by Bobin, 1950). By the use of the acid haematein test (Baker, 1946) it can be shown that this residual lipid is a phospholipid.

A positive reaction to the Sudan black test was also given by the peripheral cytoplasm; this appearance exactly paralleled that of the cytoplasm of the adipose cell of the same animal. In the latter cell it was found that a positive reaction to Sudan black was given by the fat drop, the so-called ‘surround’, and also to a considerable extent by the ground cytoplasm (Bradbury, 19566). This observation taken alone is not significant, but when considered in the light of all the other pieces of evidence which will be given in the paper, it seems to lend further support to Bobin’s views on the origin of these pigment cells.

There appeared to be a possibility that the pigment represented some form of modified lipid, i.e. it was one of the pigments generally termed a ‘lipochrome’. If this were the case, it seemed likely that free lipid would be detected in the pigment spheres of some of these cells, or, alternatively that some lipid material might be revealed by unmasking techniques. In none of the cells studied was there any evidence for the existence of free lipid in the pigmented spheres. As recommended by Ciaccio (1926), unmasking was attempted both at the time of fixation by the use of a fixative such as Da Fano’s fluid, or by the subsequent action of a 1% phenol solution for 24 h in the oven at 37^o C. These treatments were followed by coloration with Sudan black in the normal way. It was not possible to show the presence of any lipid in the pigmented spheres, even after the unmasking with phenol ; this makes it unlikely that the pigment may be considered to be a lipochrome.

A positive result was obtained with the performic acid / Schiff (PFAS) technique. This normally suggests the presence of unsaturated lipids, but since no lipid material was demonstrable in the pigmented spheres, it seems that this reaction must be the consequence of unsaturation in some other component of the pigmented spheres.

The tests for carbohydrates were negative with the pigmented spheres in this cell and it may be concluded that they contain no free carbohydrate. A positive PAS reaction was, however, obtained from the little residual cytoplasm situated round the edge of the cell. This reaction persisted after slides had been incubated in saliva at 37^o C for one hour, so that it is not due to the presence of glycogen. Further, no metachromasy was seen when sections were stained in a solution of toluidine blue, but after sulphation with concentrated sulphuric acid for a very short time, as recommended by Lison (1953), the PAS-positive material became intensely metachromatic. This seems to suggest that the reaction is due to the presence of some neutral mucopolysaccharide.

The large size of the pigmented spheres in this cell, together with their perfect regularity, suggest that there is some kind of supporting framework to which the pigment is bound. This view is supported by the fact that it is possible to bleach the pigment by a prolonged exposure either to strong hydrogen peroxide or to acids and alkalis, and if unstained sections are examined after such a treatment, the large pigment cells are seen to contain colourless spheres of a similar size to the pigmented ones which previously filled the cells. These colourless spheres may be supposed to represent the supporting framework of the pigment, and it is on such bleached preparations that all the histochemistry of amino-acids was done. When the Sakaguchi test as adapted to histochemical use by Baker (1947) was tried, a fairly strong positive result was obtained, the spheres colouring much more strongly than the ground cytoplasm. A similar result was given by the same author’s Hg / nitrite test for phenols (Baker, 1956). It thus seems possible to conclude that the structural basis of the spheres is protein, and that it contains appreciable quantities of both arginine and tyrosine.

The ‘coupled tetrazonium’ reaction (Danielli, 1947; Pearse, 1954), together with the blocking reactions suggested by Pearse, was applied to this material. A strongly positive reaction was given by the spheres in this cell, with or

Pigment Cells of the Leech, Glossiphonia complanata 305 without pre-treatment with performic acid. After benzoylation, the reaction was completely negative, but after treatment with dinitrofluorobenzene there was a positive reaction, though it was much reduced in intensity when compared with slides which had not been treated with this reagent. This reduction in intensity may perhaps indicate that a partial blockage was taking place. These results suggest that there are appreciable amounts of histidine and tyrosine present in the spheres, together with some arginine. It may this be supposed that the spheres are primarily protein, to which the yellowish brown pigment is attached.

When the pyronin / methyl green technique (Jordan and Baker, 1955) was used there was no coloration of the pigmented spheres, which appeared in their natural colour, but the ground cytoplasm around the edge of the cell was strongly stained with the pyronin. By the use of the treated saliva technique (Bradbury, 1956a) this coloration could be completely prevented, so that this staining with pyronin may be considered to be due to the presence of RNA. The nuclei in the large pigment cells often appeared unusual, in that they coloured to a much greater extent than normal with pyronin ; this could perhaps be due to their low content of DNA. The Feulgen reaction gave a strongly positive result for the nuclei which appeared among the pigmented spheres, but those which were found near the cell periphery coloured only weakly, and showed much vacuolation. It was these nuclei which were abnormal in their reaction to the pyronin ; they may well be the degenerating nuclei noted by Bobin (1950), which have lost the greater part of their DNA.

The histochemistry of the pigments is still in a rather unsatisfactory condition. This is due largely to the lack of knowledge of the exact chemical composition of many of the pigments. As a result, it is not easy to obtain a definite diagnosis of the nature of a pigment by histochemical methods alone. Certain results can be obtained by using histochemical tests on the pigment in its original position and these, when taken in conjunction with studies of the extracted material, enable one to obtain some idea of its nature.

Unstained sections of the material were mounted and examined by ultraviolet microscopy. The pigmented spheres showed a strong brownish fluorescence which contrasted vividly with the bluish-white fluorescence of the background. This test excludes the possibility that the pigment of the spheres may be a porphyrin, as these compounds show a characteristic red fluorescence.

A further means of identifying pigments is by a study of their absorption spectra. This was attempted in the first instance by the use of the eyepiece spectroscope. It was found that the image of the large pigment cell could be made to occupy the whole of the slit of the instrument ; this allowed its spectrum to be examined. The pigment did not show any absorption bands in the visible region of the spectrum, though there was much general absorption in the blue and violet region below about 500 m/x. These observations were checked by using solutions of the pigment and measuring the absorption at different wavelengths with the spectrophotometer. These results will be considered in the next section.

The solubility of the pigment was tested on formalin-fixed material by immersing unstained sections in different solvents and examining the sections at regular intervals. It was found that water, ether, and chloroform had no effect at all, even after prolonged immersion; 70% alcohol and carbon disulphide each showed some slight extraction of the pigment after about 48 h, whilst glacial acetic acid or a o-i N solution of sodium hydroxide removed all the pigment in less than 24 h. In the slide which was acted upon by the alkali, most of the pigment was extracted in about 1 h.

Hydrogen peroxide was effective in decolourizing the pigmented spheres, but in this case the action might be due either to a solvent effect or to an oxidation of the pigment to some colourless compound. These solubilities could be explained by postulating that the pigment was in fact a bile pigment, a supposition put forward by Abeloos (1925) and Juga (1931). Abeloos found that the pigment in Glossiphonia had solubilities very similar to those listed above. Verne (1926) in his book on pigments remarks that bile pigments are soluble in chloroform and also, to a certain extent, in carbon disulphide. Biliverdin would also, according to this author, be soluble in alcohol. The other solubilities (water, acids, alkalis, ether) are in accord with those of a bile pigment. He states that bile pigments are characterized by the fact that they are ‘bien soluble … dans l’acide acétique avec une teinte bleu-vert’. This is also true for alkalis and corresponds exactly to the results obtained in the experiments with whole animals to be considered on p. 307.

As a bile pigment was suspected, the two available confirmatory tests for bile pigments were used. The tests, which are said to be specific, are the Gmelin reaction (Tiedmann and Gmelin, 1826) and the Fouchet reaction quoted by Juga, 1931 ; Cole, 1955). In the case of Gmelin’s test the chemical basis has been studied in some detail (With, 1954; Lemburg and Legge, 1949), and a positive result with this test may be taken as diagnostic for the presence of bile pigment. Both Abeloos (1925) and Juga (1931) obtained a positive Gmelin reaction from the pigmented spheres studied in sections; this was tried many times in the course of the present study, but in no case did the reaction work on sections. Fouchet’s test, on the other hand, when tried on gently macerated material, gave a very marked blue-green colour which was localized in the pigment cells. This colour may be considered as a strongly positive reaction.

Pearse (1954) mentions a further reaction for bile pigments. This is Stein’s test, based on the oxidation of the pigment to green biliverdin by means of a dilute iodine solution. This reaction is capricious, and no definite results were given by the pigment cells.

There was a possibility that the pigment might be a melanin, though in view of its solubilities this seemed rather unlikely. The melanins are characterized by their general inertness, but one of the characters which serves to differentiate them from other pigments is their capacity for reducing silver in alkaline solution. This test was carried out according to the directions given in Pearse (1954), with the result that the large pigmented spheres were found to be non-reducing. The small pigmented spheres in the ‘cellules blanches* already mentioned were found to have fairly strong powers of reducing alkaline silver solutions, which seems to indicate that this particular pigment is very different from that in the large pigment cells.

Abeloos and Juga both stated definitely that there was no iron in these pigmented spheres. This was checked during the present work by the use of both Peris’s technique and by microincineration. In a previous paper (Bradbury, 1955) it was stated that the pigmented spheres gave a weak Peris’s reaction; this was again noticed in several animals, though one or two gave a much more marked reaction than the rest. If the test were preceded by a treatment with very strong (too vol) hydrogen peroxide, this positive reaction was much increased, suggesting that perhaps the iron is normally present in a masked form. Microincineration gave a positive result in all the sections examined, and as this test is independent of the state of the iron, it seems possible to say that iron is generally present in the pigmented spheres, though usually in a form which is not easily demonstrable with the standard chemical reactions. This conclusion is opposed to that reached by Abeloos and Juga, though as far as can be ascertained, neither of these workers used the microincineration technique.

When it was found that glacial acetic acid and caustic soda would remove the pigment from this cell, it was decided to attempt extraction on a larger scale in order to obtain sufficient of the pigment for spectrophotometric study. At the same time it was thought worth while to check the solubility in various solvents when large pieces of fresh tissue were used.

When fresh material was boiled with the solvents in a simple reflux apparatus, it was found that neither water, 70% alcohol, chloroform, nor a chloroform / methyl alcohol mixture extracted any pigment even after refluxing for 48 h. When the tissue was subsequently embedded and sectioned, it was found that the pigment was still present both in the large deep-lying cells containing the pigmented spheres and in the subcutaneous pigment cells. With carbon disulphide a yellowish solution resulted after refluxing for 24 h, but it was doubtful whether this represented a pigment from the pigmented spheres, because when sections of this extracted material were examined, these cells appeared to show no change in colour. When glacial acetic acid or ci N sodium hydroxide was used as the solvent, the pigment passed into solution in less than an hour. These solutions were filtered to clarify them and then studied in the spectrophotometer. The absorption in the visible region of the spectrum is shown in fig. 2. A further sample was examined with a Unicam ultra-violet spectrophotometer so that the absorption in the region 420680 m/x could be studied (fig. 3). Similar curves for a solution of mammalian bile are drawn on the same figures for comparison ; the similarity of the two is obvious. With (1954) in his monograph on the bile pigments points out that there are no pronounced absorption bands in the visible spectrum, but that bilirubin in alkaline solution shows a maximum absorption at 420 m μ, whilst biliverdin has its maximum between 300 and 370 m μ,. It is seen in the figure that both curves show a maximum around 240 m μ, but the strong maximum of vertebrate bile at 290 m μ is not obvious in the extract. There are, however, suggestions of irregularities in the absorption curve of the extract at 255 m μ and 270 m μ. These may well be ill-defined absorption peaks which are not more obvious because of the very dilute nature of the extract. They may, however, be due to the presence of some impurity in the pigment solution.

The Gmelin and Fouchet tests were performed on this aqueous alkaline extract. Positive results were obtained in both cases. As a result of the tests listed above, it seems possible to arrive at a fairly certain identification of the pigment in the large pigment cells. It does not belong to the group of melanins, as it is fairly soluble in several different liquids and it does not have the power to reduce alkaline solutions of silver. Similarly, no trace of lipids could be found in the pigmented spheres, even after the most vigorous unmasking techniques, so that it seems unlikely that the pigment belongs to the lipochrome or lipid-derived pigments.

The solubilities and absorption spectra suggest that Abeloos and Juga were correct in thinking that this pigment is one of the tetra-pyrrol derivatives, probably resembling the vertebrate bile pigments. In addition it must be noted that the ultra-violet fluorescence of this pigment is similar to that of vertebrate bile pigment; furthermore there was an association with a proteinaceous substrate which, as With (1954) points out, is always true of bile pigments. There was also a positive Gmelin and Fouchet’s reaction from extracted samples of the pigment. On the other hand, the pigment was not soluble in chloroform or alcohol, which seems to show that it is not identical in every respect with a bile pigment. Stein’s reaction and Gmelin’s reaction when applied to sections did not give a positive result. As Pearse (1954) emphasizes, a negative result with these tests does not signify anything, as they are known to be capricious. Extracts did give a positive Gmelin’s reaction, and as the chemistry of this reaction is known with some certainty (Lemburg and Legge, 1949; With, 1954), it may be assumed that a positive result is diagnostic of a bile pigment.

It thus seems possible, after considering all the evidence, to draw the following conclusions. The pigmented spheres in the large pigment cells have a proteinaceous substrate; this contains arginine, tyrosine, and possibly some histidine. The pigment in these spheres is almost certainly a tetra-pyrrol compound, resembling, though not identical with, the bile pigment of vertebrates. It has been possible to show that the pigmented spheres may in some cases contain iron present in a ‘masked’ condition.

These cells are found in the connective tissue immediately below the epidermis. They are much more numerous beneath the dorsal surface, but isolated groups of these cells occur below the epidermis of the ventral side. The cells are arranged to form a single layer and are so close together that their processes interlace (fig. i, D, E). It seems that these cells are responsible for the general body coloration.

The actual cells are irregular in shape ; they could perhaps be described as stellate. The body of the cell is about 20 p in diameter and both it and the numerous branching processes are entirely filled with pigment granules. It is noticeable that the finest processes of these cells may have pigment granules located in swellings along their length and often the process terminates in one of these swellings (fig. 4). The cells appear to contain only one nucleus which is oval and about 12 μ long. The pigment granules are extremely numerous and are dark brown or yellowish brown ; they are very much smaller than the pigmented spheres in the large pigment cell, for in no instance were they found to exceed 1 p in diameter, and usually the majority were less than 0-5 p, across.

It was found that the pigment granules in these cells were insoluble in water, o-i N NaOH, glacial acetic acid, 70% alcohol, ether, carbon disulphide, and chloroform; with hydrogen peroxide the pigment was bleached in about 48 h.

The pigment granules in this cell reduced alkaline silver solutions very strongly ; this is one of the characteristics of melanins and was used for their identification by Bizzozero (1908). The granules also possess the power of reducing ferricyanide to ferrocyanide and so forming Prussian blue in the presence of ferric salts. This reaction is the Schmorl reaction and is characteristic of lipochromes, argentaffin cell granules, and melanins (Pearse, 1954).

All other tests applied were found to give negative results with the pigment granules of this cell (see appendix). It thus seems that this brownish pigment is characterized by an extreme insolubility and chemical inertness. Its properties seem to place it in the class of melanins.

Several suggestions have been put forward during the last 50 years to account for the origin and function of these large pigment cells in Glossiphonia and in other rhynchobdellid leeches. Cuénot (1891) considered that they were excretory, and acted as a ‘kidney of accumulation’; Graf (1899), on the other hand, termed them Stapelzellen or reserve cells. He considered that they formed one of the functional stages in the progressive development of the ‘excretophores’ from the acidic cells which line the coelomic sinuses. Juga (1931) supported this idea of their derivation, including them among her category of ‘chromatocytes’, but she seems to think that these cells are specifically excretory. Both of these authors considered that the dermal pigment cells are derived from the large deeper-lying pigment cells.

Scriban (1910) and Abeloos (1925) both reject Graf’s idea of the origin of the large pigment cells, because they did not find that the ‘acid’ cells possessed any amoeboid powers, as suggested by Graf and Juga; instead, the large pigment cells were considered to arise by independent differentiation from an embryonic connective tissue cell. Abeloos also came to the conclusion that these cells acted as a kidney of accumulation.

Bobin (1950) puts forward a very strong case for supposing that the pigment cells have their origin from some of the large adipose cells which form such a prominent feature of the connective tissue of Glossiphonia, but she does not seem to put forward any hypothesis as to their function. From results obtained during the course of the present study, it seems that Bobin is correct in supposing the pigment cells to arise from adipose cells, as stages resembling those figures in her paper have been seen in my preparations.

The pigment in these cells seems almost certainly to be a tetra-pyrrol compound closely related to the vertebrate bile pigments. This would accord with the views of Cuénot and Abeloos that the pigmented spheres are the products of haemoglobin metabolism. It must be supposed that the protein of the ingested blood would be freed from the haem prosthetic group, which would then be further degraded to give a porphyrin and free iron. The porphyrin in turn would be broken down to a linear tetra-pyrrol compound which would accumulate as the pigment in the coloured spheres. Wigglesworth (1943) found that in the blood-sucking bug Rhodnius some of the ingested blood is denatured to give biliverdin which is subsequently either excreted through the gut, or stored in the pericardial cells which presumably act as a kidney of accumulation.

It has been shown (Bradbury, 1955) that free iron can be detected in great quantities in the connective tissue of Glossiphonia after a meal of blood haemoglobin. Most of this iron is localized in the adipose cells, but some does appear in other regions of the body. It might reasonably be supposed that the actual breakdown and metabolism of the haemogoblin is taking place in the adipose cell. Bobin’s conclusion that the pigmented spheres have their origin in this cell accords very well with the present hypothesis. These bodies would be regarded as the end products of the digestion of haemoglobin which are of no further use to the animal and accumulate in the cytoplasm of this cell. Eventually, when they are present in great quantity in any one cell, the cytoplasm degenerates, leaving an ‘envelope’ enclosing a mass of pigmented spheres. It seems that the view proposed so long ago by Cuénot is perhaps correct and the pigment cells are acting as a kidney of accumulation. .

The failure to detect porphyrin in any of these cells could be explained by assuming that this stage in the breakdown of haemoglobin was very rapid, so that these compounds would only be present in the cells for a very short time. The presence of some ‘masked’ iron in the coloured spheres might be expected if they represent the metabolic waste-products of a process involving iron-containing compounds ; it might be that the process of breakdown was not complete in a particular animal at the time of killing. This could also explain why Juga and Abeloos were unable to detect any iron in these cells.

In the paper already quoted (Bradbury, 1955) it was pointed out that the ‘acidic’ cells of the coelomic epithelium contain very little iron, even after the leech has been fed ; this is in contrast to the adipose cells, which contain very large amounts of iron. This observation seems to furnish further indirect evidence against the views of Graf and Juga that the coelomic epithelium gives rise to the pigment cells. If the latter are concerned in the excretion or accumulation of unwanted iron-containing, pigment and originate from coelomic epithelium, then it might reasonably be supposed that the coelomic cells would show incipient pigment formation or accumulations of iron ; these have not been observed in the present work. On the other hand, if Bobin’s idea is correct and the pigment cells are closely related to the adipose cells, then there is a very good correlation between the presence of iron and the pigment, both of which are possible waste-products of the metabolism of haemoglobin and similar compounds.

It will be difficult to obtain definite proof of this, but it is hoped that further experimental work which is planned will help to provide a better understanding of the functions of both the adipose cell and the pigment cell.

One further point which is app.arent from the present work is that there does not seem to be any direct relationship between the large, deep-lying pigment cells and the subcutaneous pigment cells. They differ both in structure, and so far as can be ascertained, in chemical composition of the pigment granules.

I wish to acknowledge the help I have received from Dr. J. R. Baker, under whose direction this work was carried out. My thanks are due also to Professor A. C. Hardy, F.R.S., for facilities in his department, to Professor Krebs, Dr. Cecil, and Miss M. Lunt of the Biochemistry Department for help with the measurements of ultra-violet absorption spectra. The work was carried out during the tenure of a Senior Hulme Scholarship of Brasenose College, Oxford, and of a grant from the Department of Scientific and Industrial Research.