Transplantation of Melanomas from Platyfish-Swordtail Hybrids Into Embryos of Swordtails, Platyfish and Their Hybrids

The actions of the macromelanophore genes of the playtfish in the hybrid are manifested in two ways. First, they enable the developing fish to produce macromelanophores and, secondly, they limit the distribution of these cells to relatively localized areas of the fish.

The growth-modifying genes contributed to the hybrid by the swordtail parent may alter the genotype of the pigment cell in such a way that its physiology is basically disturbed, thereby resulting in an overgrowth of these cells; or the effect may be an indirect one on the environment of the growing cells, resulting in hyperpigmentation.

The above suggestions are in certain cases susceptible to experiment. The physiology of the malignant pigment cell has been compared with that of the normal pigment cell in the platyfish in terms of oxidation against selected substrates. Some results obtained by this method have already been reported (Humm & Clark, 1955). In the experiments reported here the environment of the tumour cells has been changed by transplanting melanomatous tissue of the hybrid into hosts of different genetic constitutions and the extent, type, and duration of cell growth in the various hosts have been analysed.

Marcus & Gordon (1954) employed this method and analysed the growth of transplanted Sc melanoma cells when implanted under the scales of adult fish of closely related strains of swordtails and hybrids. They compared results from auto-, homo- and hetero-transplantation and reported varying degrees of success in accordance with the genetic relationship between donor and hosts. Except in autotransplants the grafts gradually decreased in size and disappeared. It seemed to us that the failure of homo- and hetero-grafts might be attributed to an immunological reaction on the part of the hosts, rendering it difficult for the tumour cells to maintain themselves in the host tissues. Accordingly, attention was turned to the embryo of the same groups of fish in the hope that grafting at an early stage, before the animals had developed the ability to react immunologically, might allow the tumour to become so firmly established on the host that it would survive after the animal hatched.

The terminology of pigmented cells which will be used here follows that which was adopted during the Third Conference on the Biology of Normal and Atypical Cell Growth (Gordon, 1953).

The swordtails (Xiphophorus helleri) and the platyfish (X. maculatus) and their hybrids used in these experiments were bred from an original stock supplied by the Genetics Laboratory of the Aquarium, New York Zoological Society.

Three types of host embryos were used : swordtails, platyfish, and backcrossed swordtail-platyfish hybrids (Text-fig. 1). All were removed by Caesarean section from gravid females.

These were obtained from swordtails which had been inbred for several generations. Their normal pigment pattern results from many widely distributed micro-melanophores on their bodies.

These were obtained from gravid platyfish females of inbred stocks. They, too, were uniformly pigmented by micromelanophores; but in addition had macro-melanophores. As Gordon (1948) has shown, one sex-linked allele Sd is responsible for the development of macromelanophores in the dorsal fin, another allele Sp accounts for the presence of large black pigment cells on the flanks.

The backcross Sd embryos (B.C.) used were obtained from female swordtails which had been backcrossed with an F₁(PS) male (Text-fig. 1). Only 50% of the backcross hybrid embryos may be expected to develop melanomas because they are the product of a mating between one parent which is heterozygous for the Sd allele for macromelanophores and one parent which is recessive for that allele. About 50 % of the embryos would be expected to develop into swordtail-like fish and were denoted as B.C. (+).

The host embryos were selected at stage 20 using the criteria of Tavolga (1949). The methods of staging, of removing the host embryos from the mother, and decapsulation have been described in a previous paper by Humm & Young (1956).

Melanoma tissue was obtained almost exclusively from B.C.₂Sd hybrids. The donor was placed in 1: 10,000 merthiolate (Lilly) for 30 min., and a small piece of melanoma removed by means of iridectomy scissors. Under the dissecting microscope the tissue was cut into small pieces and maintained until used in 3 × Niu’s solution. Tissue fragments kept for 60 min. in Niu’s solution were just as viable as those kept for shorter periods.

In the earlier experiments, a small notch was made with glass needles through the ectoderm of the host, just anterior to its dorsal fin, and a 0-1 mm.³ piece of tumour tissue was immediately laid into the notch. Attachment was good and the graft healed well into place. There was little tendency for the host tissue to heal over the graft and often part of the tumour fragment was exposed to Niu’s solution. This exposure resulted in the disintegration of the exposed portion within 3 or 4 days.

Later, melanoma fragments were placed under the extra-embryonic membranes of the embryos. Turner (1937) and Tavolga (1949) have shown that in the early stages of development the serosa grows up and over the head of the embryo very much like the extra-embryonic membranes of the chick. The serosa consists of a double membrane which is heavily vascularized and thus presents an ideal site for melanoma cells. Just lateral to the eye the inner membrane remains closely applied to the embryo, while the outer membrane bellies out slightly. After making a small cut a piece of tumour tissue was inserted between the two membranes and by means of a gentle stroking with a ball-tipped glass needle the tissue was moved to the dorsal aspect of the embryo’s head. Here the melanoma fragment is protected and comes into intimate contact with an abundant circulation. Grafts placed in this position apply themselves to the serosa’s upper and lower membranes and spread out in a sheet (Pl. 8, figs. 2-4). These grafts are essentially chorio-allantoic. Continuous microscopic observation and photography of these grafts is facilitated since the fragment is always uppermost in the embryo because its relatively heavy yolk prevents the embryo from turning over. The cellular migrations from the tumour implants are clearly traceable.

The life of such a chorio-allantoic graft was limited. As the extra-embryonic membranes regressed at the approach of hatching the tumour cells growing in the membranes also regressed. Only a few melanoma cells moved away from the membranes and became attached directly to the host.

Two modifications in grafting technique were developed to improve the chances that tumour tissue would attach to the embryo proper and remain after the serosa regressed. (1) A small cut was made through the extra-embryonic membrane just over the opercular region of the embryo. With a pair of Dumont forceps this hole was held open while a small piece of tumour tissue was pressed through the hole and under the operculum of the developing fish. The operation is dangerous since it impinges on one of the larger veins of the extra-embryonic circulation, the rupture of which almost always causes death of the embryo. While the graft did not grow as well or as immediately in the subopercular position some tumour cells did attach to the host and spread. (2) With the embryo held firmly in an agar depression, the tips of the Dumont forceps were pressed under the rear margin of the extra-embryonic membrane, lifting both membranes away from the embryo. A piece of tumour tissue inserted between the slightly spread points of the forceps came to lie against the side of the skull of the embryo where it came into contact with the abundant serosal circulation. Such grafts grew well and as a result large numbers of tumour cells became attached to the surface of the embryo.

The proper maintenance of the post-operative embryos contributed greatly to the success of the experiments. After the insertion of the graft, the embryo was placed in a stendor dish with a minimum volume of growing medium which had the following composition: 3 × Niu’s solution to which had been added—10 i.u. penicillin G per c.c., 0-25 mg. dehydrostreptomycin per c.c. ; casein hydrolysate (enzymic) to a final concentration of 25 mg.%; 0-02 ml./ml. embryonic extract, i : 50 (beef serum ultrafiltrate gives at least as good results).

Formerly, glucose was added to the nutrient medium (Humm & Young, 1956), but its use was discontinued since it appeared to have a slightly deleterious effect on the survival of the embryos.

Despite the presence of penicillin and streptomycin in the culture medium, frequent infections occurred during observations of the embryos which required removal of the stendor dish cover. This hazard was eliminated by sealing a sterile 2 × 3 in. microscope slide to the ground surface of the stendor dish by means of a sterilized mixture of beeswax and vaseline, 1:3. This made an airtight, sterile and optically flat seal through which the embryo could be examined under a dissecting microscope. It was also possible to photograph it with a 32 mm. Micro Tessar lens using paired electronic flash guns as light sources.

After about 30 days, when the yolk mass of the embryo had decreased considerably, the cover-glass was taken off and the hatchling was serially exposed to a decreasing concentration of salts. Several changes of decreasing salt concentration were usually required before the fry could be safely released into tank water. As soon as possible the small fish were fed with brine shrimp nauplii.

All melanoma grafts, regardless of the site of implantation, gave rise to three types of pigmented cells : melanophages, melanocytes and macromelanophores. Of these the first cells to appear on the embryo’s surface, singly or in groups, were melanophages which made their appearance around the margin of the graft 1-5 days after the operation. They were small cells, roughly triangular, containing much pigment. These cells migrated very actively, frequently traversing the entire length and surface of the embryo. They finally disintegrated and were eliminated in a manner described by Marcus & Gordon (1954), who studied the fate of these cells after a melanoma fragment was implanted under scales of adult fish. It is presumed that melanin-laden macrophages erupt through peripheral tissues and are lost—a process also described earlier by Gordon & Lansing (1943).

Two types of melanocytes begin to migrate from the margin of the graft during the 1st-5th day after transplantation. Those which first emerge are characterized by having long, slender processes, and by being quite dark and heavily laden with pigment. These melanocytes were similar in appearance and motility to those studied in tissue cultures of fish melanoma by Grand, Gordon & Cameron (1941). The second type of melanocyte, which appears a little later, is light grey in colour, has broader processes and contains local concentrations of melanin around the nucleus and at the periphery of the cell. Both types of melanocytes migrate slowly, spread widely but remain in contact with one another.

The shape of the two types of melanocytes can differ tremendously, depending upon their immediate environment. When crowded the melanocytes assume a long thin condensed appearance, whereas when they are migrating upon the surface of the embryo or in an uncrowded portion of the yolk they will assume the serrated shape typical of broad melanocytes. Plate 8, fig. 1, is a photograph of a whole mount of a melanoma graft and shows thin melanocytes, broad melanocytes and melanophages.

The macromelanophore is a much larger polymorphic cell of 300—470/μ, with lobulated or dendritic processes. It has been described by Grand et al. (1941), Ermin & Gordon (1955) as having oval or polymorphic nuclei that are 15-20μ in diameter; it may contain one or two nucleoli and a loose network of chromatin.

The grafts of melanoma tissue into a dorsal notch of thirty-five swordtail embryos suffered two distinct fates. Either the graft broke down immediately into isolated clumps of pigmented tissue and was then attacked by macrophages, or after a portion of the graft had broken down and the cellular debris was removed, macromelanophores appeared, usually under the epidermis; but melanocytes were never seen. After an initial period of spreading, which seldom exceeded a week, the large pigment cells remained static for relatively long periods of time, and then gradually disappeared. Since the cells implanted in this manner were in an ideal position to invade the host tissues and occasionally did, the degenerative changes in the grafted cells cannot be assumed to be due to a poor nutritional environment. Similar cells in the same position on the back of a hybrid fish invade and destroy muscle and connective tissue. It can be concluded that the failure of these cells to grow represents evidence for a basic difference in the tissue environment of the cells. The wild-type swordtail embryo apparently does not present as favourable a growth environment for the pigment cells as do the hybrid tissues.

The sixty-six grafts placed on the extra-embryonic membranes of the swordtail embryo presented quite a different cytological picture. A high percentage (68%) of the grafts were successful and gave rise to typical melanocytes, usually but not invariably accompanied by macrophages. The growth of these melanocytes continued for a week or more with an occasional case in which it continued for several weeks. After this growth period the graft, although now spread out in a thin sheet on the surface of the membranes, became static. See table 1.

The survival of swordtail embryos bearing tumour grafts depended to a considerable extent upon the size of the graft. The protocols indicated that a distinct antagonism existed between the graft and the host which was manifested in two ways. When the graft was large, it spread out on the host, and within 4–7 days caused a circulatory failure in the host characterized by extensive engorgement and stasis of the blood vessels of the yolk followed by apparent haemolysis and clotting of the blood. This was judged to be an immunological response, and Hilde-mann (1957) has described a similar circulatory phenomenon resulting after the homotransplantation of scales in goldfish.

When the graft was small, the host destroyed the cells and engulfed the graft ; the host embryo continued to grow and eventually hatched. When the graft was neither too large nor too small, the graft initially grew successfully and the host embryo did not suffer undue reaction, but during a period of up to 2 months the tumour cells gradually and completely disappeared. The swordtails continued their development to maturity.

The effect of the grafting procedure on the host was evaluated by sham operations. A hole was made in the embryo’s membrane but no piece of tissue was inserted. About 80% survived but were distinctly retarded in their subsequent growth, as compared with the normal development time in the ovary, probably attributable to a considerable extent to the inadequacy of the growth solutions used.

Fragments of melanoma were transplanted to the extra-embryonic membrane of sixty platyfish embryos carrying either the Sd (spotted-dorsal) or Sp (spot-sided) allele or both. Forty platyfish embryos died from circulatory failure between 4 and 7 days following the operation. Eleven of twenty, or approximately one-half, of the host platyfish embryos survived to hatching time, whereas only eight of thirty-two, or one-quarter, of the host swordtail embryos survived under comparable conditions. These results suggest that the platyfish embryo tissues offered a considerably more favourable genetic environment for the growth of melanoma cells than the tissues of the swordtail embryo.

Even more striking evidence of the above was obtained from other details in the protocols. After tumour tissue was grafted onto swordtail embryos, some cells spread out in a thin sheet, but they never seemed to multiply. However, in both the platyfish embryos and, as will be seen subsequently, in the hybrid embryos, the tumour cells definitely proliferated. Several of the host platyfish embryos died from the proliferation of tumour cells which invaded the pericardial area and physically impeded the action of the heart.

The most striking differences observed between the host hybrid embryos and host platyfish and the swordtails were found in the type of tumour cell migration, proliferation and invasiveness. This may be illustrated by comparing Pl. 8, fig. 4, which shows a graft 9 days after transplantation on to a hybrid embryo, and Figs. 2 and 3 of Pl. 8, which show swordtail and platyfish embryos respectively, bearing malignant grafts of comparable age. The pigment cells in the hybrid show broad processes typical of rapidly migrating melanocytes. In the platyfish and swordtail embryos the pigment cells have barely become established and only thin projections of melanocytes may be seen under the edges of the grafts. The tremendous activity of the melanoma cells killed many hybrid embryos early. In the hybrid shown in fig. 4, the tumour cells multiplied so profusely that they filled the entire pericardial cavity in 7 days and finally killed the host.

A further significant difference between the hybrid embryo hosts and those of others was found in the length of time tumour cells exhibited active migration. Melanoma cells on one hybrid embryo (Pl. 9, fig. 3) after 30 days had proliferated abundantly, migrating and spreading over most of the head and yolk sac. For comparison, in a platyfish embryo of the same age, it may be seen in Pl. 9, fig. 2, that the tumour cells have not proliferated and migrated as extensively as in the hybrid. Compare also the swordtail embryo with a graft of the same age (Pl. 9, fig-1).

Despite their luxuriant growth in the embryos the melanoma cells did not gain a permanent foothold on the young fry. One graft which had grown very abundantly on the hybrid embryo (Pl. 9, fig. 3), and had come to cover the anterior portion of the body on the left side, continued to grow for 43 days ; then it disintegrated quite rapidly as shown in Pl. 9, figs. 4-6. The graft had consisted (fig. 4) of many macromelanophores which were typically chocolate brown and powdery in appearance ; this is characteristic of the intact melanoma overgrowth in adult hybrids. Retrogression of the graft coincided with the appearance of six or eight melanophages at the anterior edge of the graft ; then the number of melanophages increased rapidly in the next few days. In the meantime individual pigment cells and groups of cells disintegrated, leaving small clumps of pigment granules behind. In about 8 days the whole graft (Pl. 9, fig. 5) consisted of clumps of melanophages and pigment detritus. The disappearance of all graft pigment was complete in 13 days after the first sign of degeneration.

Although 101 transplantations of melanoma tissue to swordtail embryos were made, significant growth never occurred in any of them. This observation is in accord with that made by Marcus & Gordon (1954), who reported little or no success after Sc melanoma fragments from hybrids were transplanted under the scales of adult swordtails. Thus it would appear that the genetic constitution of swordtails lacks the proper environmental qualities suitable for the growth of transplanted hybrid melanocytes and macromelanophores. The growth of grafted tissue on the platyfish embryo was significantly better than on swordtails but inferior to the melanocyte growth on the backcross hybrid embryos. In ten grafts on hybrid embryo hosts, the rapid and excessive growth of the tissue actually killed the animal by growing into the pericardium and stopping the heart. Only two of the platyfish embryos were killed in this manner, but none of the swordtails. The basic environmental requirements for the survival of malignant melanocytes were present in both the platyfish and hybrid embryos, while the swordtail’s environment lacked almost completely the competence to support growth.

The better growth of tumour tissue on the platyfish and hybrid embryos may be accounted for by a consideration of the genetic make-up of the hybrid donor and its tumour as compared with the constitutions of the swordtail, the platyfish, and the hybrid embryos. The hybrid tumour tissues must of necessity possess at least one platyfish chromosome that carries the macromelanophore allele. But owing to the fact that the hybrid embryos were the product of at least one backcross to pure swordtail (Text-fig. 1) the chances are that the backcross hybrid embryos had more swordtail than platyfish chromosomes. But since embryos with a complete complement of 48 swordtail chromosomes do not provide a suitable medium for the growth of platyfish-swordtail hybrid melanomas, it appears that a certain quantity of platyfish heritable material is necessary, probably at least the chromosome carrying the macromelanophore allele. This interpretation is consistent with that of Snell (1953), according to whom the best growth of transplanted tumour tissue should have been obtained (among those hosts employed) in the backcrossed fish. These fish possess a much wider spectrum of inherited cellular antigens than either parent, and the graft by virtue of its own heterogeneity would appear less foreign to this host and be less likely to provoke a transplantation immunity reaction.

In this respect it is interesting to note that the tumour grew much better in the platyfish than in the swordtail. Since the tumour is made up primarily of pigment cells of one type or another, it is interesting to speculate that the possession of the macromelanophore gene ‘conditions ‘the platyfish host somewhat to the reception of the tumour. The logical control, namely the implantation of tumours into platyfish embryos lacking melanophore genes, although projected, has not been performed because of the lack of this strain in our colony.

From other evidence the importance of the platyfish macromelanophore-carrying chromosome has been indicated; thus Marcus & Gordon (1954) had better success in transferring fragments of the Sc melanoma into Sd dominant montexumae-helleri hybrids than to those siblings which were recessive (+). Kallman & Gordon (1957), in experiments in which whole fins were transplanted, concluded that genetic correspondences between donor and host were most important in determining the successful transplantation.

Concerning the effect of hybridization upon the macromelanophore, it appears from other work done in this laboratory (Humm & Clark, 1955 ; Humm & Humm, unpublished) that an admixture of swordtail genes definitely changes the physiological constitution of the macromelanophore. These changes are manifested by a greatly augmented oxidative capacity in the tumour melanocyte. Whether or not these changes in the metabolism of the cell are causative in the production of hyperpigmentation and eventual melanoma formation will be the subject of further investigation.

Marcus & Gordon (1954), Ermin & Gordon (1955) and Greenberg, Kopac & Gordon (1956) have reported the conversion of a cell having the shape and characteristic pigment pattern of a melanocyte into a pigment cell having the characteristics of a macromelanophore. We have observed similar cellular changes repeatedly. It would appear, however, that the physical character of the substratum is an important determinant of the shape of these cells. When the first migration from a graft occurs it is characterized by the appearance of very long thin dark processes. As these cells migrate out from the graft they begin to spread out and usually fill the space between two small blood vessels on the serosa. The anterior edge of such a cell possesses a fringe of pigment, but the main body of the cell is only lightly pigmented. This type of cell, the broad melanocyte, has been described as a cell type different from the small dark melanocyte; however, we believe that the difference lies in the type of substratum upon which the cell is moving.

When the migration of the graft has continued to the point where the pigment cells have spread out in a single-layered sheet, migration appears to cease. The most spectacular case of this indicates that the cells may remain static for several months. Under these circumstances the pigment cells tend to assume a highly dendritic rounded appearance very typical of the macromelanophore. Accordingly, it seems quite possible that the shape of the three ‘types ‘of pigment cells reported from the melanoma of the fish may be a result of crowding and the substrate upon which they are migrating at the time of observation.

Small pieces of melanoma tissue were obtained from swordtail-platyfish hybrids and grafted into the embryos of swordtails, platyfish and hybrid embryos.

Several types of cells appeared in all the grafts: melanophages, two kinds of melanocytes, and macromelanophores were observed and described.

It was found that the environment of the grafted tissue played a predominant role in the survival of the graft. Grafts to swordtail embryos survived only a short time, whereas grafts to platyfish embryos bearing the gene for macromelanophores survived significantly longer. It was found that hybridization, with or without the macromelanophore gene, presented a still more favourable environment for the growth and migration of pigmented cells.

In no case was it possible to establish and maintain a permanent strain of tumour cells by embryonic transplantation.

This research was supported in part by United States Public Health Service Grant No. 1717 from the National Cancer Institutes and a grant from The American Cancer Society.