Restoration of the antibody response to sheep erythrocytes in thymectomized Xenopus implanted with MHC-compatible or MHC-incompatible thymus

The anuran amphibian, Xenopus, possesses a major histocompatibility complex (MHC) that appears to display structural and functional homologies with the mammalian MHC (Du Pasquier, Chardonnens & Miggiano, 1975; Flajnik, 1983; Flajnik, Kaufman, Hsu & Du Pasquier, 1984). As in mammals, MHC antigens in Xenopus appear to be involved in restriction of T-B collaborative responses. Thus carrier-primed T cells and hapten-primed B cells cooperate in vitro to produce 7S (IgY (Hadji-Azimi, 1979); also referred to as Ig‘G’ - see Bernard Bordman, Blomberg & Du Pasquier, 1981) anti-dinitrophenyl (DNP) antibody only if taken from Xenopus laevis/X. gilli clones (Kobel & Du Pasquier, 1975) having either one or two MHC haplotypes in common (Bernard et al. 1981). If the carrier-primed T cells were histo-incompatible with the DNP-primed B cells, then no 7S antibody response was observed, although occasional low affinity IgM antibody responses took place.

Xenopus can be thymectomized (Tx) very early in life, at a rudimentary stage of thymus differentiation (Horton & Manning, 1972; Tochinai & Katagiri, 1975), and subsequently are deficient in humoral antibody production to an array of T-dependent antigens (see Cribbin, 1984, for review). Therefore these animals should prove to be a useful comparative model for investigating the extent to which the thymus (rather than a peripheral site) is involved in MHC restriction of T helper cell responses. This issue remains controversial in mammals (see Singer, Hathcock & Hodes, 1982). Recently Du Pasquier & Horton (1982) have shown that Tx Xenopus laevis/X. gilli and X. laevis/X. muelleri clones implanted with genetically defined MHC-compatible or MHC-incompatible larval thymuses can respond normally to DNP-keyhole-limpet haemocyanin. The IgM antibody produced was similar in quantity, affinity and specificity to that of non-operated controls; and 7S antibody was of host spectrotype.

The experiments reported here were designed to examine the ability of MHC-incompatible larval thymus (irradiated or non-irradiated) and ‘adult’ thymus implants (MHC-compatible or MHC-incompatible) to restore the cellular and serum antibody response of Tx Xenopus laevis to sheep red blood cells (SRBC). ‘Adult’ thymuses were used because it is only during or after metamorphosis that ‘adult-type’ MHC antigens are found on Xenopus cells (Du Pasquier, Blomberg & Bernard, 1979; Flajnik et al. 1984). In order to reveal whether our thymectomy/ thymus implant system would allow host cells to migrate to and develop within the foreign thymus environment, we have also investigated the origin of implant thymocytes, by ploidy-labelling experiments.

Outbred adult Xenopus laevis were purchased commercially (Xenopus Ltd.). Inbred (G-line) adult Xenopus laevis were à gift from C. Katagiri (Hokkaido University, Sapporo, Japan). G-line Xenopus are MHC-identical (Katagiri,1978) and have recently been designated JJ to reflect their homozygosity at the MHC (DiMarzo & Cohen, 1982); these toads display minor histocompatibility disparities (DiMarzo & Cohen, 1982) and reject skin grafts in ‘sub-acute’ (30–50 days) or ‘chronic’ (> 50 days) fashion (J. C. Arnall & J. D. Horton, in preparation). A few of the animals used as thymus donors were isogenic hybrids - Xenopus laevis/Xenopus gilli, clone LG5 (MHC haplotypes ad). The LG5 adults were a gift from L. Du Pasquier, Basel Institute for Immunology, Switzerland. X. laevis and LG5 offspring were produced as previously described (Horton & Manning, 1972; Kobel & Du Pasquier, 1975 respectively). Animals were kept at 23 °C until immunization.

Thymectomy was carried out using the microcautery method of Horton & Manning (1972). Thymuses were removed when larvae were 7 days old (stage 47). Larvae were checked for absence of thymus regeneration before metamorphosis and also at post-mortem.

Single thymus implants were transplanted to thymectomized larvae when 4–6 weeks old (stages 54–57); implants were placed under the skin, medial to the eye. The donor thymuses used were either from larvae of similar stage to the host, or suitably sized fragments of thymuses (usually half a thymus) from 4- to 5-month-old toadlets:these toadlet thymuses are referred to as ‘adult’. All reconstitutions with larval thymuses were performed using MHC-incompatible thymuses, some of which were irradiated before use: they were given a dose of either 1000 rads (dose rate 200 rads/min) or 5000 rads (1000 rads/min) by exposure to a cobalt-60 source. Adult thymuses were either MHC-compatible or MHC-incompatible. Some larval thymus implants were examined histologically at 4 and 30 days after transplantation. For this study alone Tx animals were implanted (on opposite sides of the head) with both a control and an irradiated (either 1000 or 5000 rads) thymus. This allowed direct comparison of the effect of irradiation on thymus implant development, avoiding individual host differences.

Animals used in the MHC-incompatible reconstitution experiments include: (i) outbred thymectomized hosts reconstituted with thymus from outbred non-siblings; (ii) outbred thymectomized hosts reconstituted with thymus from inbred G-line animals; and (iii) inbred G-line thymectomized hosts reconstituted with thymus from cloned LG5 animals (see Tables 1–4). Where outbred/outbred combinations were used, the donor and host were Ffs of different parentage and differed from each other by at least one, and probably by two MHC alleles, as controls (from the same families as used here) were shown always to reject each other’s skin grafts in an ‘acute’ fashion (18–21 days at 24–26 °C - data not shown). The same is true for the outbred/G-line G-line (jj) and LG5 (ad) reject each other’s skin grafts in 21–23 days (J. C. Arnall & J. D. Horton, in preparation). For the MHC-compatible reconstitution experiments, both Tx hosts and thymus donors were inbred, G-line animals.

Triploid, Tx larvae were implanted with diploid thymuses and thymic lymphocytes were assayed for ploidy at 6–8 months of age. Triploid (3N) Xenopus were produced by the cold-shock technique of Kawahara (1978). Ploidy of lymphocytes was assessed by silver staining of nucleoli (Olert, 1979).

Sterile sheep red blood cells (SRBC) (Tissue Culture Services) were washed three times and resuspended in saline. 6- to 12-month-old toadlets were immunized by injection of 0·05 ml 10 % SRBC/gm body wt., intraperitoneally. Toadlets received either a single SRBC injection and were tested for plaque-forming cells (PFC) 6 days later, or three injections given 3 days apart (the ‘multiple-injection’ schedule: see Results) and tested for splenic PFC and serum antibody production 2 weeks after the final injection. Immunized animals were kept at 26 °C.

Spleens were removed aseptically from MS222 (Sandoz)-anaesthetized animals. Spleen cell suspensions were prepared, chilled on ice, as described in detail elsewhere (Cribbin, 1984) using amphibian strength L15 medium supplemented with 1 % heat-inactivated foetal calf serum (Flow). Splenocytes were adjusted to a maximal concentration of 5×10⁶ lymphocytes/ml. Samples (160 μl) of splenocytes were mixed with 12 μl 25 % SRBC, and 40 μl 1: 10 (SRBC-absorbed) guinea pig serum (Welcome) as a source of complement. Assay mixtures were allowed to warm to room temperature and then gently pipetted into 90 μl (double microscope slide) chambers. Two to four chambers were set up per sample. After 2 h incubation at 30 °C, PFC were counted - only when a central Xenopus lymphocyte could be seen in the plaque was a PFC scored. PFC were expressed as the number/10⁶ leucocytes.

Blood was collected by cardiac puncture and serum removed. 25 μl of serum were serially diluted in V-well microtitration plates (Flow) in mammalian-strength phosphate-buffered saline (PBS) or 0·1 M-2-mercaptoethanol (2-ME) diluted in PBS. 2-ME-treated plates were incubated at 37 °C for 1 h: this removes IgM antibody activity, but leaves IgY antibody intact (Turner & Manning, 1974). 5 μl 5 % SRBC and 20 μl 1: 10 guinea pig complement (absorbed with SRBC) were then added. The contents of the wells were mixed by gentle agitation and the plates then incubated for 2 h at room temperature. The maximum dilution of serum which caused lysis of the SRBC was taken as the titre of haemolytic antibody, and was expressed as — log₂ titre.

Antibody responses were compared using Student’s t-test.

Controls, Tx and Tx animals reconstituted with a larval allothymus showed neither background cellular production of antibody nor 2-Me-resistant serum antibody (see Table 1). Three out of eight controls, two out of three Tx and two out of two thymus-implanted Tx animals did give a positive background total serum antibody titre; levels of antibody were comparable (P > 0·1) between groups.

The 6-day splenic PFC responses given by control, Tx and Tx animals implanted with MHC-incompatible larval thymus are shown in Table 2. Control animals gave a response of 215 ± 180PFC/10⁶ leucocytes (mean ± S.D.) whereas Tx animals did not respond to SRBC. However, implantation of an MHC-incompatible larval thymus restored the abilty of Tx animals to produce plaques against SRBC. The number of PFC obtained was 108 ± 98 PFC/10⁶ leucocytes which was not significantly different from the control response (P > 0·1). (Serum antibody production was not tested in animals receiving a single SRBC injection.)

The responses of control, Tx and Tx animals implanted with non-irradiated larval or irradiated larval MHC-incompatible thymus are shown in Table 3, and the response of Tx animals reconstituted with MHC-compatible or MHC-incompatible adult thymus are shown in Table 4. .

Control animals all responded to SRBC, with a cellular antibody level of 85 ± 43 PFC/10⁶ leucocytes and a total serum antibody titre of 7·4 ± 1·2. After treatment of the immune sera with 2-ME, the IgY antibody titre was 4·0 ± 2·1.

Tx animals did not respond to SRBC. None of the nine animals tested gave any anti-SRBC PFC. Six out of nine animals did have anti-SRBC serum antibody, but the mean titre of 3·1 ± 2·5 was similar (P > 0·1) to the titre given by non-immunized Tx animals (see Table 1). This value of 34 was significantly lower than the titre displayed by immunized controls (P < 0·001). The immunized Tx animals produced no 2-ME-resistant antibody.

An MHC-incompatible larval thymus implant restored the ability of Tx animals to respond to SRBC. However, the splenic cellular antibody response was only 38 ± 36 PFC/10⁶ leucocytes, which was significantly different from the control level (P<0-02). The total serum antibody titre obtained was 7·4 ± 1·8 and the 2-ME-resistant antibody titre was 3·5 ±1·2. These values were not significantly different from the control levels (P>0T).

Irradiated MHC-incompatible larval thymus implants were able to restore the ability of Tx animals to mount a PFC response and total serum antibody response to SRBC; thymuses given 1000 or 5000 rads were equally effective. The cellular antibody response (87 ± 19 and 118 ±92 PFC) and total serum antibody titres (7·0 ± 1 and 6·7 ± 1*5) are comparable to control levels, in both 1000 and 5000 rad thymus-reconstituted animals respectively (P>0·l). The 2-ME-resistant antibody titres for animals reconstituted with irradiated thymuses appeared lower than control levels, but this difference was not significant (P>0·05, but <0·l). However, more experiments are needed to confirm that 2-ME-resistant antibody titres are restored to control levels.

Both MHC-compatible and MHC-incompatible ‘adult’ thymus implants were able to restore SRBC responses of Tx animals, but to different extents. Animals implanted with an MHC-compatible thymus gave a cellular antibody response of 137 ± 102 PFC and a total serum antibody titre of 7 ± 3. Two out of three animals produced 2-ME-resistant antibody, with a mean level of 2·3 ±2·1. Reconstitution with an MHC-incompatible thymus enabled Tx animals to mount a cellular antibody response of 25 ± 18 PFC, and a mean total serum antibody titre of 5·3 ± 1·8. Only two out of eight animals were able to produce 2-ME-resistant antibody, with a mean level of 0·8 ± 1·4. All the responses given by animals reconstituted with MHC-compatible thymus were comparable to control levels (P>0·l). In′ contrast, in the animals given an ‘adult’ MHC-incompatible thymus, the cellular antibody response and both total and 2-ME-resistant serum antibody titres were all significantly lower than controls (P< 0·001, P<0·01 and P<0·01 respectively).

Non-irradiated larval allothymus implants appeared normal in structure within 4 days of implantation (Fig. 1A). In contrast, both 1000 rad and 5000 rad (Fig. 1B) irradiated implants were smaller and predominantly epithelial, with only a few scattered lymphocytes. The difference in size between control and irradiated thymuses was pronounced at 30 days post-implantation (Fig. 1C) when animals have begun postmetamorphic life. However, both 1000 and 5000 rad thymuses did contain lymphocytes and displayed cortex/medulla differentiation. Although histological studies were not continued here after 30 days, external observations revealed that non-irradiated implants remained readily detectable under the skin, medial to the eye (Fig. ID). (Long-term retention of normal histology of larval allothymus implants has been shown by Horton & Horton, 1975.) Some 1000 rad-treated implants were just visible at postmortem, whereas others had disappeared, as had all the 5000 rad-irradiated thymuses. Both MHC-compatible and -incompatible ‘adult’ thymus implants remained readily visible throughout the experiments.

Table 5 shows the results of ploidy-labelling experiments investigating the origin of lymphocytes in larval (non-irradiated) allothymus implants. In 3N control animals 61 ± 7 % (mean ± S.D.) of thymic lymphocytes were recorded with three nucleoli, whereas in 2N controls only 4% were recorded with three nucleoli; now 76±9% had just two nucleoli. In triploid Tx animals implanted with diploid thymuses, thymus implant lymphocyte nucleolar number appeared directly comparable to triploid controls; 63 ± 8 % were recorded as having three nucleoli and under 30% were recorded with two nucleoli. Thus the implants have been heavily infiltrated by host-derived cells. No observations were made here on irradiated larval or adult thymus implants.

Non-irradiated, MHC-incompatible larval thymus implants display a normal histology and become repopulated by host lymphocytes, although precise, quantitative, measurements of the extent of this repopulation cannot be made with the silver-staining technique used (see below). Anti-SRBC antibody responses of thymectomized animals given such a thymus are restored, although the splenic PFC response following multiple SRBC injections was lower than in controls. Restoration was not due simply to a non-specific ‘allogeneic effect’, since non-SRBC-injected Tx toadlets reconstituted with larval allothymus displayed only background anti-SRBC reactivity. Experiments with nude mice have shown that an implanted, neonatal, allothymus also becomes repopulated with host cells (Loor & Kindred, 1973). However, the ability of such ‘athymie’ mice to mount an antibody response to SRBC is only partially restored (Kindred & Loor, 1974). In Xenopus irradiated, MHC-incompatible, larval thymus implants initially contained few lymphocytes, but became lymphoid within a few weeks, although the origin of these thymocytes has yet to be examined. Irradiated implants remain very much smaller than non-irradiated implants and tend to disappear after metamorphosis: thus long-term lymphoid repopulation appears not to occur. In nude mice, irradiated, allogeneic perinatal thymus implants remain small and largely epithelial (Loor & Hägg, 1977). Such irradiated thymus implants fail to restore responses of nude mice to T cell-dependent antigens (Kindred, 1978), whereas in Xenopus, irradiated implants restored antibody responses to SRBC.

MHC-incompatible adult thymus implants (in contrast to MHC-compatible ones) fail to restore humoral responses to T-dependent antigens above background levels in nude mice (Radov, Sussdorf & McCann, 1975) and adult thymus grafts are inferior (with respect to restoring immune function) to neonatal thymuses in neonatally-thymectomized mice (discussed by Loor & Hâgg, 1977). In Xenopus ‘adult’ alloimplants seem partially to restore both the cellular antibody response in the spleen and 2-ME-sensitive serum antibody production. However, 2-ME-resistant antibody (presumably IgY) was not recorded in the majority of Tx animals given MHC-incompatible ‘adult’ thymus. Future thymus reconstitution experiments with Xenopus will concentrate on the use of ‘adult’ thymus implants and determine whether there are real differences in IgY antibody production between (genetically defined) Tx animals given MHC-compatible and MHC-incompatible (2 haplotypedisparate) thymus implants. Monoclonal antibodies against Xenopus IgY are now available, and these will be used in an enzyme-linked immunosorbent assay or radio-immunoassay.

Overall, our restoration experiments using SRBC, and others (Du Pasquier & Horton, 1982) using an antigen known to require MHC-restricted T-B collaboration in vitro (Bernard et al. 1981), would tend to lead to the conclusion that host-derived T cells, that have developed in the foreign thymus, can repopulate the periphery and there collaborate perfectly well with hostderived B lymphocytes (and antigen-presenting cells) in primary in vivo antibody responses. The amphibian thymus may, at first glance, seem not to be centrally involved in self-restriction of helper T cells. However, the involvement of donor-derived cells in effecting the antibody responses observed in our experiments cannot be ruled out, particularly where non-irradiated thymus implants were used. In this respect Nagata & Cohen (1984) have recently found, using flow cytometry on mithramycin-stained cells, either no donor-derived cells or up to 46 % or 32 % in thymus or spleen respectively (the proportion depending on the hostμonor combinations used) 6-13 months after Tx Xenopus were implanted (as toadlets) with ‘adult’ MHC-compatible or MHC-incompatible thymuses. (Interestingly, these authors reveal that MHC-incompatible thymuses also fully restore the splenic PFC response to SRBC).

The experiments with irradiated thymus reported here and elsewhere (Du Pasquier & Horton, 1982) were set up to remove the possibility of donor B cell contamination, as both T and B lymphocytes exist in the Xenopus thymus (Williams, Cribbin, Zettergren & Horton, 1983; Hsu, Julius & Du Pasquier, 1983). B cell activity in Xenopus is eliminated by a dose of 500 rads in vitro (Blomberg, Bernard & Du Pasquier, 1980) and so the irradiated thymuses used here should be free of B cells. The possibility that radiation-resistant (see Bernard et al. 1981) donor-derived helper T cells, rather than host-derived T cells, might be involved in the immune response of restored Xenopus still cannot be excluded. Moreover, involvement of radiation-sensitive donorderived suppressor T cells could account for the higher mean level of splenic PFC’s recorded in Tx animals implanted with irradiated implants than in animals given non-irradiated thymus - see Table 3. Although a dose of 3000 rads results in organ-culture thymuses becoming virtually devoid of lymphocytes within 2 weeks (J. H. Russ, personal communication), on the other hand Cribbin (1984) has evidence that 3000 rad-irradiated lymphocytes (from spleen/peripheral blood lymphocyte mixtures) of one genotype can, when injected alongside immunizing SRBC into a thymectomized host of another genotype, cooperate with the B cells of that host to effect a normal 6 day primary anti-SRBC PFC response. Whether or not a long-term restoration (i.e. comparable to the interval used here between thymus implantation and SRBC injection) occurs was not resolved in Cribbin’s experiments. (A long-lasting T-B cooperation does not occur in anti-SRBC responses of nude mice injected with MHC-incompatible thymocytes or splenocytes (Kindred & Weiler, 1972).) The outcome of such experiments in Xenopus should allow a clearer interpretation of the red cell studies with MHC-incompatible thymusreconstituted animals reported here. Lack of involvement of donor implant lymphocytes in restoration is suggested by our unpublished observations that neither larval spleen nor larval liver alloimplants reconstitute anti-SRBC responses of early thymectomized Xenopus.

The failure to demonstrate a role of the thymus in MHC restriction of T cells could conceivably be obscured in these thymectomy-reimplantation studies by early seeding of immature, but already host-MHC-restricted, T-axis lymphocytes, prior to removal of the thymus at 7 days of age. These lymphocytes could then mature into functional T cells under the hormonal influence (see Dardenne, Tournefier, Charlemagne & Bach, 1973) of the thymus implant: extra-thymic differentiation of T precursor cells could well be the restorative mechanism in animals given irradiated thymus implants, where implants failed to grow well. Experiments have recently been performed by Flajnik, Du Pasquier & Cohen (1984) which address this issue. They have used chimaeric Xenopus in which the anterior ‘half’ of an embryo is joined to a posterior ‘half’ of an MHC-disparate embryo. The anterior portion contains the thymic anlagen of one MHC haplotype and the posterior portion contains the haemopoietic stem cell source of a different MHC type. Hence during ontogeny of these chimaeras, lymphocyte precursors entering the thymus exclusively differentiate in an MHC-incompatible thymus epithelial environent. It was found that cellular antibody responses to SRBC were normal in some chimaeric combinations, whereas they did not occur in others. Chimaeric animals were still able to produce both IgM and IgY antibodies to DNP-keyhole-limpet haemocyanin, but the IgY response showed delayed kinetics and lower antibody titres. For the moment it appears that the thymus may play some part in the education of Xenopus helper T cells, but its influence is by no means critical.

In none of the experiments investigating the role of the Xenopus thymus in MHC restriction is it known whether the host T cells developing are biased towards interactions with cells of the thymus donor type rather than with those of the genotype of the Tx host or stem cell source in the chimaeric animal. Work in this laboratory has begun to examine this issue (see Lallone, 1984) by studying the potential of T cells from control animals and thymus-implanted Tx Xenopus to cooperate with B cells of various MHC types in a primary in vitro antibody assay (to avoid possible alteration in T helper cell bias induced by in vivo priming). Finally, it will be important to consider more closely the actual MHC-restricting elements (epithelium, etc.) within the amphibian thymus. It has been suggested that self-MHC restriction is learned in mammalian thymus by interaction of developing T cells with MHC products on thymic antigenpresenting cells; the latter are able to migrate into the thymus from the bone marrow (Longo & Schwartz, 1980). Thus host-derived, MHC-restricting, antigen-presenting cells may well migrate into an MHC-incompatible (amphibian) thymus, along with lymphoid stem cells, and so partially obscure the organ’s role in T cell education.

Supported by MRC Project Grant G80/0720/2CB (to J.D.H.) and a S.E.R.C. Studentship (to F.A.C.). Thanks go to Mrs T. L. Horton and Mrs J. Mather for help in preparation of the manuscript, and to Mr D. Hutchinson for the photography.

 
• 2-ME

  2-mercaptoethanol

 
• MHC

  major histocompatibility complex

 
• PFC

  plaque-forming cell

 
• SRBC

  sheep red blood cell