Thymocyte/stromal cell chimaerism in allothymus-grafted Xenopus: developmental studies using the X. borealis fluorescence marker

Although the mammalian thymus is thought to play a key role in eliminating potentially autodestructive T-lineage lymphocytes, the thymic stromal cell types responsible for effecting tolerance induction to MHC antigens remain uncertain: thymic epithelial cells (Jordan, Robinson, Hopkinson, House & Bentley, 1985) and thymic dendritic cells (Ready, Jenkinson, Kingston & Owen, 1984; Jenkinson, Jhittay, Kingston & Owen, 1985) have both been implicated. The ease with which early thymectomy (Horton & Manning, 1972; Tochinai & Katagiri, 1975) and subsequent thymus implantation (Horton & Horton, 1975; Du Pasquier & Horton, 1982; Gearing, Cribbin & Horton, 1984) can be performed in Xenopus makes this anuran amphibian model system a useful alternative for exploring the role of the thymus microenvironment in self- and allo-tolerance induction (see, for example, Nagata & Cohen, 1984; Arnall & Horton, 1986). We are currently attempting to refine the thymus transplantation technique in Xenopus, initially by using lymphocyte-depleted thymus implants, so that the role of an allogeneic thymic stromal compartment in the development of host-derived T cells can eventually be assessed without the complication of influence by donor-derived lymphocytes. In a companion paper, we have described the morphological effects of γ-irradiation on the froglet thymus (Russ & Horton, 1987) and have shown that this treatment, in the short term, is successful in depleting thymic lymphocytes, while many stromal cell types remain intact.

The derivation of lymphocytes developing within the thymus and in the periphery of allothymus-grafted, early-thymectomized (Tx) Xenopus has, to date, only been explored following transfer of fully lymphoid thymus to larval (Gearing et al. 1984) or adult (Nagata & Kawahara, 1982; Nagata & Cohen, 1984) recipients and, furthermore, the question of whether thymic stromal cells remain of donor or host origin was not examined in these ploidy-labelling experiments. Irradiated larval thymus implants are known to become lymphoid within a month of transplantation; however, the origin of these thymocytes was not assessed (Gearing et al. 1984).

In amphibians, triploid embryos can be produced experimentally by routine procedures, and this has provided an important and widely used cell marker system (Turpen, Cohen, Deparis, Jaylet, Tompkins & Volpe, 1982; Turpen & Smith, 1986). Until recently, the ploidy of chimaeric tissues was determined by laborious or rather imprecise techniques – e.g. evaluating cell size, estimating nucleolar number, counting the number of chromosomes in metaphase plates or microdensitometry of Feulgen-stained cells (reviewed in Flajnik, Horan & Cohen, 1984). However, a far more accurate, fairly rapid method of detecting diploid and triploid cells has now been developed; this employs flow cytometry of mithra-mycin-stained cells (Nagata & Cohen, 1984; Flajnik et al. 1984). A recently described marker technique, that holds considerable promise for those working with the Xenopus immune system, is based on the differential fluorochrome staining of the nuclei of different Xenopus species, e.g. X. laevis and X. borealis (Thiébaud, 1983). When stained with the fluorescent dye quinacrine the nuclei of X. borealis cells display (under u.v. illumination) a number of bright fluorescent spots on a homogeneous background, whereas X. laevis quinacrine-stained nuclei have no bright spots. Some workers believe that the bright fluorescent spots may be associated with AT-rich sequences of DNA along the chromosomes (Weisblum & de Haseth, 1972; Brown, Dawid & Reeder, 1977). However, there is also evidence that AT-rich DNA and quinacrine-stained fluorescent spots may not be so strictly related in situ (Comings, 1978). With the quinacrine-staining technique, the origins of a variety of cell types in X. borealis/laevis chimaeras are clearly visible (Thiébaud, 1983) although lymphocytes were not examined in her study. As an added advantage (compared with ploidy-marking), the quinacrine fluorescence is applicable to sectioned material as well as smears.

The quinacrine marker system (along with toluidine-blue stained sections) is here used to probe the timing of host lymphoid cell input to normal and lymphocyte-depleted foreign thymus implants, to identify the origins of stromal and lymphoid cells within the implants, and to assess the extent to which donor cells persist in the host blood and spleen of the early-thymectomized hosts.

X. borealis adults carrying the quinacrine fluorescent cell marker were a generous gift from Dr R. H. Clothier, University of Nottingham, UK. The parents of his X. borealis colony had originally been purchased several years ago from Xenopus Ltd, who had field collected them from Lake Nukuru in Southern Kenya. Putative X. borealis adults were also purchased by us directly from Xenopus Ltd. Apparently this stock had originally been collected from the NW Province of Kenya (by Mr M. P. Simmonds, School of Biological Sciences, Queen Mary College, London). We have found these animals do not possess the quinacrine-spotting nuclear pattern. We have not formally identified these quinacrine-negative Xenopus, although the males do possess the same distinctive mating call as the quinacrine-positive X. borealis. The quinacrine-positive and -negative ‘borealis’ are here referred to as X. borealis Q+ve and X. borealis Q−ve respectively. X. borealis, J strain X. laevis and the isogeneic X. laevis/X. gilli hybrids (clone LG15) were bred and reared in the laboratory at 23°C as previously described (Horton & Manning, 1972; Kobel & Du Pasquier, 1975).

Thymectomy was carried out using the microcautery technique of Horton & Manning (1972). Thymuses were removed when larvae were 7 days old (stage 47 of Nieuw-koop & Faber, 1956). Single thymus implants were subcutaneously transplanted to Tx larvae when 4–6 weeks old (stages 56–58, the beginning of metamorphosis), as described in detail elsewhere (Russ, 1986). Thymuses were either from normal froglet donors or removed from donors 8 days post-3000 rad irradiation. Donor/host combinations and ages of donors used are shown in the Tables. All ages given in this paper are postfertilization. Thymus donors were froglets, rather than larvae, since we wished to implant thymuses that would be likely to express both class II and class I donor MHC antigens – this occurs only after metamorphosis in Xenopus (Flajnik, Kaufman & Du Pasquier, 1985).

Fine glass pipettes were used to draw blood from the heart and blood smears thus obtained. Thymuses were teased apart in amphibian L15 (see Russ & Horton, 1987) and single cell suspensions made by repeated pipetting to dissociate clumps. Thymocyte cytospin preparations were prepared as described elsewhere (Williams, 1981). Blood and thymocyte smears were air dried for 30min, then fixed in methanol for 10min, prior to quinacrine staining.

Thymus implants and spleens were embedded in historesin (historesin kit 2218/500, from LKB), which proved superior to paraffin wax for preparing 4 μm sections, the thickness most suitable for subsequent quinacrine staining. Tissues were fixed in Carnoy’s for 30 min, then transferred to 95 % ethanol. Dehydrated specimens were first immersed in intermediate infiltration solution for ⁠, then in pure infiltration solution until specimens appeared slightly translucent. Tissues were then placed in embedding medium in gelatin capsules and left to polymerize for 40–120 min at room temperature. 4 μm sections were stained in toluidine blue or quinacrine.

Slides were first placed in buffer (18% citric acid 0·1 M; 82% Na₂HPO₄ 0·2 M, pH 7·0) for 10min. They were then stained with fluorescent dye (0·5 % solution of quinacrine dihydrochloride (Sigma) in buffer) for 40 min using a foil-covered staining dish. Slides were then passed through three changes of buffer, sections or smears being left for 20 min in the last buffer wash, then mounted in 50% sucrose: the slides were finally sealed with wax (50:50 wax/vaseline) and stored in the dark until observation with a Nikon fluorescence microscope, using a BV (blue/violet) filter.

(i) Blood smears from X. borealis were stained with quinacrine to confirm the nuclear spotting pattern of fluorescence. All the F₁ offspring (larvae and frog-lets) from the X. borealis Q+ve adults given by Dr R. Clothier had such ‘spotted’ nuclei, whereas all the F₁s of the X. borealis Q−ve adults were quinacrine negative.

(ii) Thymus cytospins were prepared from X. borealis Q+ve (Fig. 1A) froglets (4 months old) and also from similar-aged J strain Xenopus laevis (Fig. IB) to confirm the quinacrine-staining properties of thymocytes. The usefulness of the quinacrine fluorescence to detect the origin of individual cells was indicated by preparing 50:50 in vitro mixes of X. borealis Q+ve and J strain Xenopus thymocytes (Fig. 1C). Unfortunately, in the preparation of thymocyte smears, the majority of epithelial and other stromal cells, e.g. the adherent cells such as dendritic antigen-presenting cells and macrophages (Turpen & Smith, 1986) are lost, and the majority of cells examined are thymic lymphocytes. Therefore it was necessary in the planned studies to examine quina-crine-stained sectioned material, which would be able to reveal far more about the origin of various cell types in the implanted thymuses.

(iii) Quinacrine-stained historesin thymus sections of 4 μm thickness proved optimal for observing the spotted appearance of X. borealis Q+ve nuclei and the lack of spotting in X. ‘laevis nuclei. Thinner sections revealed too few spots for fluorescence analysis and thicker sections gave rather high background fluorochrome staining.

These experiments were designed to determine when a more detailed kinetic study on lymphoid repopulation should take place. Thymus grafts were examined at 2 and 16 weeks postimplantation and host spleens at 16 weeks. The outcome of this quinacrine-staining experiment is shown in Table 1. Nonirradiated thymus implants were mainly composed of donor cells at 2 weeks postimplantation, although a few host cells were seen, mainly in the medulla. By 16 weeks, the vast majority of lymphocytes in the thymus cortex were now host-derived, whereas stromal cells throughout the thymus appeared to be predominantly of donor origin. In the medulla some donor lymphocytes were still found alongside host lymphocytes. Spleens removed from LG Tx animals given X. borealis Q+ve normal thymus implants 16 weeks previously revealed an occasional thymusimplant-derived cell both in the perifollicular regions (see Manning & Horton, 1982 for detail of Xenopus spleen structure) and in the red pulp.

Within 2 weeks of grafting a 3000 rad-irradiated thymus, the majority of thymocytes found in the implants were already of host origin (see Table 1). By 16 weeks, irradiated thymus implants were no longer found: spleens were not examined.

A detailed kinetic study at 1, 2, 3 and 8/9 weeks postimplantation was carried out. A summary of the data on thymus implants is given in Table 2. Sections of spleen and blood smears were also examined in this experiment, which assessed both quinacrine- and toluidine-blue-stained material.

Thymuses were healthy (Fig. 2E) with relatively normal lymphoid cortex and medulla. Host polymorphs (particularly neutrophils) were frequently seen adjacent to the implants. ‘Spotted’ (host) cells were seen in the cortex (both in blood vessels and in the thymic parenchyma), although the majority of cortical cells, including mitotic cells, were still donor in origin. Overall, the medulla was far more of a chimaeric cellular population than the cortex – i.e. both host and donor cells were found interspersed. Host cells included lymphocytes, polymorphs and cells with pale, irregular-shaped nuclei. The identity of the latter was uncertain.

Irradiated thymus implants were greatly reduced in size (Fig. 2A) and contained only scattered lymphocytes. Much adipose tissue surrounded these implants. Neutrophils were frequent around and within the irradiated thymus; these cells were of host origin and may well be involved in the phagocytosis of dead or dying irradiated thymic cells. Epithelial cells of the ‘cortical’ region were clearly of donor (‘nonspotted’) origin whereas the few lymphocytes found emanated from both donor and host. The ‘medulla’ was heavily infiltrated with a variety of host cell types, including haemocytoblasts (large cells with prominent nucleoli and basophilic cytoplasm), lymphocytes and cells with large, irregularly shaped nuclei (see Fig. 3B).

These thymuses were comparable to the normal Q− ve implants described above. This donor/ host combination proved useful for tracing donor-derived (‘spotted’) cells in the periphery. Quinacrine-stained blood smears revealed a significant number of donor lymphocytes, although the precise proportion of donor-derived PBL was not calculated. The spleens from these J recipients contained a high level of thymus donor-derived cells. These were mostly noticeable in the red pulp and perifollicular zones, but a few donor lymphocytes were also seen within the white pulp.

These thymus implants were less lymphoid than at 1 week. The cortex was particularly reduced in lymphoid content and some necrosis was visible in the medulla. The vast majority of cells in both cortex and medulla were shown by quinacrine staining to be of donor origin.

Compared with week 1, irradiated thymus implants contained noticeably more lymphocytes within the cortical region, but the thymuses were still very small (Fig. 2B); cysts, mucous cells and myoid cells frequented the medulla. In quinacrine-stained preparations, clusters of host-derived lymphocytes, occasionally seen in mitosis, were now seen in the cortex (Fig. 3A); the medulla also contained a variety of host-derived cells, as described for week 1 (Fig. 3B).

Thymus implants were comparable to those removed at 1 week postimplantation. Thymus donor-derived lymphocytes were again frequently seen in the blood smears prepared (Fig. 4C). Spleens had distinct white pulp areas bordered by a double layer of boundary cells. Lobular cells, thought to be antigen-presenting cells (Baldwin & Cohen, 1981), were evident in the white pulp (see Fig. 4B). Spleens appeared well populated with lymphocytes, many of which are thymus donor-derived (Fig. 4A). The red pulp and perifollicular zones were especially rich in spotted, thymus-derived lymphocytes. Donor cells were also found scattered within the white pulp.

Nodules of small lymphocytes were found in the thymus cortex, but the centre of the organ appeared necrotic; very few host cells were recorded.

These thymus implants have now a greatly restored lymphoid population (Fig. 2C). The subcapsular cortex contained many lymphoblasts with frequent mitotic figures visible (Fig. 2D). Quinacrine staining revealed that the cortex was rich in host-derived lymphoid cells. Stromal elements were predominantly of donor origin; these included cystic cells, mucous and myoid cells.

Thymus implants in J recipients had only minimal infiltration by host cells, whereas a mixture of donor and host-derived thymocytes was seen in the LG recipients. In the latter, epithelial cells and reticuloendothelial cells (lining blood capillaries of the thymus) appeared predominantly to be of donor origin, although some host-derived reticuloendothelial cells were noted. In both groups, thymus-derived cells were frequent in the splenic perifollicular and red pulp regions. The blood was also chimaeric, with donor-derived lymphocytes representing approximately 2– 3 % of PBL.

This combination highlighted the persistent donor stromal cell types remaining in the thymus implants (see Fig. 3C,D). Donor-derived cell types frequently seen were as follows: epithelial cells, reticuloendothelial cells, myoid and mucous cells, melanin-containing cells and cells forming the capsule. X. borealis Q+ve cells were also occasionally seen amidst the cellular debris within cysts, the latter being frequently found in the irradiated thymus (Fig. 3D). The thymic lymphocyte population appeared to be almost entirely of host origin.

No donor-derived lymphocytes were seen over the two blood smears examined (although one fluorescent erythrocyte was found!) and the spleen was also devoid of Q+ve cells, despite having restored lymphoid perifollicular regions.

These thymuses (11 months old when transplanted) had by now decreased in size and, although still containing many lymphocytes, had little suggestion of cortex and medulla. Cystic spaces were frequently seen. There were still very few host-derived lymphoid cells within these thymuses; those that had immigrated were scattered throughout the organ.

In contrast to 3 weeks postimplantation, these thymuses now appeared very reduced in lymphoid content and seemed to be degenerating. Stromal elements were of donor origin, whereas lymphoid cells remaining were host-derived.

In contrast to the situation above, these nonirradiated thymuses – from younger (6 or 8 months old) donors – were very healthy, with masses of lymphocytes and cortex/medulla differentiation. Quinacrine staining revealed that the vast majority of lymphocytes were, at 8-9 weeks postimplantation, host-derived. Thymic stromal elements were predominantly of donor origin. With these thymuses, then, we see a dramatic development of host-derived lymphocytes in the thymus soon after metamorphosis. Donor-derived lymphocytes were found scattered in the spleen (mainly in the red pulp) and also very occasionally found in the blood.

These thymuses, although still very much smaller than nonirradiated implants, were extremely lymphoid and showed no signs of the degeneration seen in the X. borealis Q−ve irradiated thymus noted above. Thymic lymphocytes were host-derived, whereas stromal cells appeared mostly of donor origin. No donor-derived cells were found in spleen or blood.

In this paper, the X. borealis marker system has proved to be useful for studying cell traffic to and from thymuses implanted to early-Tx Xenopus. The identity of quinacrine-positive/negative cells is especially clear in smears and cytospin preparations, since the whole of each cell is seen and hence the fluorescence is bright. Quinacrine staining can also be successfully used with resin-embedded material, sectioned at 4 μm thickness, even though the fluorescence intensity is naturally lower, as only part of each cell is visible. This success with tracing cells in sectioned material means that the origin of thymic (and splenic) lymphoid and nonlymphoid cells can readily be examined in situ, a distinct advantage over previous ploidy-labelling studies.

The results presented here with 4- to 8-month-old MHC-incompatible thymus implants confirm our recent preliminary ploidy-labelling data (Russ, 1986), where MHC-compatible, normal or irradiated froglet (6 months old) thymus grafts applied to control or Tx larval Xenopus contained large numbers of host-derived lymphocytes within 7 weeks. Furthermore, they extend similar observations of others on lymphoid thymuses implanted to Tx Xenopus (Nagata & Kawahara, 1982; Nagata & Cohen, 1984; Gearing et al. 1984) by revealing for the first time that many stromal components of the thymus remain donor-derived. Chimaeric Xenopus, produced by joining the anterior portion of one 24 h embryo (containing the thymic epithelial anlagen) to the posterior portion of an MHC-incompatible embryo (from which the haemopoietic stem cells, including lymphocytes, arise) have recently been shown, by use of alloantibodies and fluorescence microscopy, to possess thymus glands composed of epithelium of one MHC and thymic lymphocytes of another type (Flajnik, Du Pasquier & Cohen, 1985). We have shown here that, in addition to epithelial cells, a whole range of stromal cell types persist after introducing a foreign thymus to a Tx larva. These various cell populations must therefore be considered as possible candidates for influencing the differentiation of developing T-lineage lymphocytes. It is important that further studies are performed to examine the stroma in more detail, i.e. to assess MHC antigen expression on stromal cells and to determine the origin of phagocytic and nonphagocytic (dendritic) accessory cells (thought to enter the thymus early in embryogenesis – see Turpen & Smith, 1986) in thymus implants, since these may play a critical role in the development and differentiation of neighbouring thymic lymphocytes, as noted in the Introduction. The nature of host cells, with irregularly shaped nuclei, that we frequently saw within normal and irradiated thymus implants soon after transplantation is uncertain. These could well be one population of thymic accessory cells.

Our kinetic investigations reveal that when hosts are still in the final stages of transformation (i.e. 3 weeks post-thymus implantation), their lymphoid thymus implants are generally still predominantly of donor origin. However, by nine weeks the picture has, with the exception of thymus implants from 11-month-old donors, dramatically changed to a mainly host-derived thymocyte population. Interestingly, lymphoid thymuses allotransplanted to early-Tx Xenopus in adult life are also infiltrated by host cells, with timing comparable to that seen here (Nagata & Kawahara, 1982). Although stem cells are known to enter the larval Xenopus thymus at 3-4 days of age (Tochinai, 1978; Kurihara & Kato, 1986), details about later windows of input to the in situ thymus remain fragmentary. Slow repopulation of lymphoid thymus grafts seen in the present study appears not to be caused by lack of available stem cells, since 3000rad-irradiated grafts become rapidly recolonized by host cells, their lymphoid complement being almost exclusively host type within 2 weeks of implantation. Interestingly, immigration of precursor cells into the irradiated murine thymus is also rapid compared with immigration to the normal thymus (Scollay, 1985).

It seems that 3000 rad irradiation may cause irreparable, long-term damage to thymic stromal cells, since although irradiated thymus grafts initially become repopulated with host lymphocytes etc., they fail to grow in the froglet and, usually, eventually disappear. This disappearance occurs even when thymus implants from J strain Xenopus donors bear only minor histocompatibility differences to their larval hosts (Russ, 1986). Disappearance is unlikely to be due to rejection, since Tx animals given irradiated thymus grafts, although restored to 3rd party MHC antigens, tolerate skin of the thymus donor type (Arnall & Horton, 1986). Irradiation is known to affect the ability of cells to divide and so the relatively slow cell turnover rate of thymic stromal cells (compared with lymphocytes) means that the effect of irradiation on the stromal cells takes many weeks/months to manifest itself. These findings indicate a critical role of donor stromal cells in the longterm survival of implanted thymuses. When donor stromal cells die (possibly within thymus cysts – see Fig. 3D), the thymus degenerates, presumably because the essential (epithelial?) cells cannot be supplied by the host.

Tx J strain animals given X. borealis Q+ve normal thymuses displayed extensive numbers of donor cells in both blood and spleen soon after implantation; low-level chimaerism of peripheral lymphocytes was still apparent several months postimplantation. (Graft-vers-us-host reactivity was never noticed in these experiments with fully lymphoid, xenogeneic, thymus implants.) Chimaerism (lasting for at least a year in some cases) has also been revealed by mithramycin staining following transfer of MHC-matched or mismatched, ploidy-labelled thymuses to postmetamorphic, early-Tx Xenopus (Nagata & Cohen, 1984). These authors revealed the proportion of persistent donor cells, 6–7 months postimplantation, ranged from 0–46 % for thymocytes and from 0–32% for splenocytes, depending on donor/host combination used. We show here that thymus donor lymphoid cells were mainly located in the marginal zone and red pulp lymphoid tissue of the spleen - known to be T-dependent lymphoid regions (see Horton & Manning, 1974; Obara, 1982). However, some cells of thymic origin were also found in the splenic white pulp, a B-cell-rich zone (Baldwin & Cohen, 1981). Perhaps the thymic-derived cells in the white pulp seen here represent the B cell population known to exist in the Xenopus thymus (Williams, Cribbin, Zettergren & Horton, 1983; Hsu, Julius & Du Pasquier, 1983). Few, if any, implant-derived lymphocytes were to be seen in the periphery of Tx animals given irradiated thymus implants. Persistent tolerance towards allogeneic skin coming from the same strain as the irradiated thymus implant (Arnall & Horton, 1986) must therefore not be dependent on peripheralized, thymus graft-derived cells. Our recent experiments (unpublished) with Tx LG15 Xenopus, implanted in larval life with adult LM3 (X. laevis/X. muelleri isogeneic hybrids) in v/vo-irradiated (3000 rad) thymus, confirm Arnall & Horton’s (1986) findings on skin graft tolerance towards thymus donor MHC. However, splenic lymphocytes of such animals could still respond well in in vitro MLR to LM3 stimulator splenocytes. Studies are in progress to assess further the alloimmune system of Tx Xenopus implanted with lymphocyte-depleted thymus, in an attempt to understand better the influence of various thymic stromal cell types on T cell development.

The so-called X. borealis Q-ve’ animals used in some of the experiments here were raised from adults which we believe to have been originally collected from NW Kenya. Those Xenopus morphologically appeared to be true X. borealis (M. P. Simmonds, personal communication). However, it is known that considerable hybridization between Xenopus species occurs - this is certainly so in the NW Province of Kenya, where hybridization of X. borealis with X. laevis victorianus is suspected (M. P. Simmonds, pers. comm.). Perhaps such hybridization has ‘diluted out’ the quinacrine-specific marker. F] hybrids (produced in our laboratory) between X. borealis Q+ve and J strain X. laevis still carry the fluorescent marker, but this is less intense than in X. borealis Q+ve animals. It seems unlikely to us that our ‘X. borealis Q−ve’ animals are actually X. muelleri, although these two species of Xenopus have apparently often been mistaken for one another in the past (Brown et al. 1977 ; Thiébaud, 1983). The origin of our quinacrine-nega-tive animals (the highlands of NW Kenya near to the town of Eldoret, between Kisumu and Tambach), is believed to be outside of the distribution of X. muelleri (M. P. Simmonds, pers. comm.). Thiébaud (1983) reveals that X. borealis show extremely clear quinacrine fluorescence, while X. muelleri have been found to be entirely negative. The present study suggests that some X. borealis populations may also not carry this fluorescent cell marker.

This work was supported by research grants from the Science and Engineering Research Council and the North of England Cancer Research Campaign (to J.D.H.) and a SERC Research Studentship to J.H.R. We thank David Hutchinson for help with the photography and Mark Simmonds for valuable information about X. borealis.