Reprints Available Directly from the Publisher Photocopying Permitted by License Only Expression of Class Ii Major Histocompatibility Complex Antigens on Adult T Cells in Xenopus Is Metamorphosis- Dependent

Class II major histocompatibility complex (MHC) antigens are expressed predominantly on B lymphocytes and macrophages of tadpoles of the South African clawed frog, Xenopus laevis, as is the pattern in lymphocyte populations of most mammals. However, unlike most mammals, young postmetamorphic frogs show expression of class II MHC antigens on a high proportion of thymocytes and most peripheral T and B lymphocytes. Using the J-strain of Xenopus and the anticlass II monoclonal antibody, 14A2, we have studied, by indirect immunofluorescence, whether inhibition of metamorphosis would alter the pattern of expression of class II antigens during ontogeny. In control animals, class II antigens were virtually absent from thymic lymphocytes and peripheral T cells of normal untreated larvae, but could be found in increasing numbers in both populations after metamorphosis (10-12 weeks of age). In contrast, larvae, whose metamorphosis was inhibited by treatment with sodium perchlorate, had relatively few class II thymic lymphocytes throughout the 6-month period of study, and the proportion of class II splenic lymphocytes was approximately equal to that of IgM B lymphocytes. Thus, perchlorate-treated animals retained the larval pattern of class II epression, suggesting that emergence of class II T cells is dependent on metamorphosis.


INTRODUCTION
Xenopus laevis is the most primitive vertebrate with a defined MHC (reviewed by Du Pasquier et al., 1989). As an amphibian, it is also unique among vertebrates because it undergoes metamorphosis. With this reorganization of nearly every major physiological system comes the activation of new gene programs and the appearance of new adultspecific molecules (Just et al., 1977(Just et al., , 1980May and Knowland, 1980). As a result of metamorphosis, the immune system also undergoes a number of changes (reviewed in Flajnik et al., 1987, andDu Pasquier et al., 1989). Among them are changes in the pattern of expression of MHC molecules. Reagents that immunoprecipitate class I MHC molecules from adult cells reveal no classical class molecules on tadpole cells (Flajnik et al., 1986). The class II *Corresponding author. molecules expressed by tacipoles and adults are the same (Flajnik et al., 1986), but they are expressed on different subpopulations of lymphocytes. In the tadpole, class II antigens are expressed on B lymphocytes and macrophages, whereas in the adult, they are expressed on some thymic lymphocytes and virtually all peripheral B and T lymphocytes (Du Pasquier and Flajnik, 1990;Flajnik et al., 1990). We have hypothesized that one way by which tadpoles could accommodate new adult-specific antigens would be to largely discard the larval immune system and develop a new one after metamorphosis. That is, metamorphosis would be characterized by a major loss of larval-type lymphocytes with new lymphopoiesis and expansion of the adult-type population after metamorphosis. This is a reasonable hypothesis in view of the well-studied changes in erythrocyte populations that occur at metamorphosis in anuran amphibians. Larval and adult erythrocytes differ in morphology and express different hemo-98 L.A. ROLLINS-SMITH AND P. BLAIR globins (reviewed by Broyles, 1981). In both Rana catesbeiana and Xenopus laevis, adult hemoglobin appears to be expressed at metamorphosis only in newly differentiating erythrocytes. Larval erythrocytes become biosynthetically inactive and soon disappear from circulation (Just et al., 1980;Dorn and Broyles, 1982;Flajnik and Du Pasquier, 1988). Here we present evidence that the adult-type class II / population of T lymphocytes that normally arises in young postmetamorphic frogs does not appear in metamorphosis-inhibited permanent larvae. This suggests that development of this population is metamorphosis-dependent. stages 62-66. By 3-4 months of age (approximately 1-2 months postmetamorphosis), the number of thymocytes expressing class II determinants had increased to about 19%. At 5-6 months of age, 32% of thymocytes expressed class II antigens. In contrast, the proportion of thymocytes staining positive for surface IgM remained low throughout ontogeny ( Fig. 1). Thus, most class II / cells in the thymus are not B lymphocytes. Some macrophages in the thymus (by morphological criteria) are also class II/, but they represent only 1-3% of the larval or adult thymocyte population by esterase staining (data not shown). Thus, the vast majority of class II / cells in the thymus are maturing T cells. (4-7 wks) (8-9 wks) FIGURE 1. Ontogeny of expression of class II MHC antigens on thymocytes of J-,strain frogs. Thymocytes were stained for class II antigens (O--O) using the monoclonal antibody designated 14A2 and surface IgM (OO) using the monoclonal antibody designated 6.16. Both antibodies were used at a concentration of approximately 100-200/zg/ml and were followed by staining with a fluorescein-conjugated polyvalent goat antimouse Ig used at 100/g/ml. Each data point represents the mean 4-SE of 5-7 individuals or pools of animals. Metamorphosis occurred in these frogs at about 10-12 weeks of age.

99
In the spleen, between 35-40% of lymphocytes expressed class II antigens during larval stages 53-57 and metamorphic stages 62-66. These percen-. tages were virtually identical to the percentage of cells expressing surface IgM. This confirms the observation of Du Pasquier and Flajnik (1990) that class II antigens are expressed on B lymphocytes during larval life. At 5-6 months of age (2-3 months postmetamorphosis), the percent of class II + splenocytes had increased to about 72% while IgM + cells remained at about 42%. This suggests that class II + T lymphocytes had become a significant part of the splenic, lymphocyte population ( Fig. 2), as shown previously by Flajnik et al. (1990). In contrast to the pattern shown in Figs. 1 and 2 for normal animals, perchlorate-treated frogs showed a much different pattern of expression of class I! antigens during ontogeny. Thymocytes from perchlorate-treated frogs at 5-9 weeks of age, when normal control frogs are at larval and metamorphic stages, contained about the same number of class II / lymphocytes as the normal J-strain controls of the same age. By 3-4 months of age, the percent of class II / thymocytes had increased to about 11%, but did not increase further at 5-6 months of age, when normal controls contained about 32% class II / thymocytes (Fig. 3). In the spleen of perchloratetreated frogs, class II / lymphocytes comprised about 25-40% of the total throughout the 6-month period    When preparing cells for staining, we recorded the total number of leukocytes recovered from spleen and bilateral thymuses of normal and perchloratetreated frogs of various ages. Included in these studies were several groups of outbred larvae that developed normally or were perchlorate-treated in the same way as the J-strain animals were treated. Total leukocyte numbers in both organs in perchlorate-treated animals were not different from normal controls .during larval and metamorphic stages (6-11 weeks of age). After metamorphosis, however, the number of cells in thymus and spleen of normal control frogs increased very dramatically while those in perchlorate-treated frogs remained the same or only slightly higher than those of younger larvae. At 5-6 months of age, control thymuses held six-fold more cells while control spleens held three-fold more cells (Fig. 5).

DISCUSSION
These experiments were designed to demonstrate whether the emergence of class II / T lymphocytes in postmetamorphic frogs (Du Pasquier and Flajnik, 1990;Flajnik et al., 1990) is a metamorphosisdependent Phenomenon. Clearly, in larvae whose metamorphosis is inhibited, this population of lymphocytes does not expand. The expression of class II antigens on 8-11% of thymocytes in 3-6month-old perchlorate:blocked larvae may suggest that in these older larvae some immature class II + thymocytes develop, but the thymic environment is not appropriate for their selection, ,expansion, and export. Alternatively, we cannot presently exclude the possibility that the positive cells are epithelial cells. Because most class II / cells in the spleen of these blocked animals are IgM/, it is clear that very few class II/T cells reach the spleen. The appearance of class II antigens on Iglymphocytes could be due to a change in the pattern of expression of class II genes on existing cells after metamorphosis or due to the appearance of a new set of cells due to lymphopoiesis after metamorphosis. The observation of Turpen and Smith (1989) that the lymphocytes of a diploid thymus implanted in a triploid host before metamorphosis were largely replaced by host cells after metamorphosis suggests a new wave of stem-cell immigration and expansion in the thymus at metamorphosis. Our comparison of the total recoverable leukocytes in untreated and perchlorate-treated frogs (Fig. 5) suggests that prevention of metamorphosis inhibits expansion of the thymic lymphocyte population that normally occurs in postmetamorphic frogs (Du Pasquier and Weiss, 1973;Rollins-Smith et al., 1984). This observation is consistent with the hypothesis that emergence of the class II / T-cell population occurs as a result of the generation of a new set of adult-type lymphocytes because of a maturational change in the thymus at metamorphosis. There are a number of structural changes observed in the thymus of Xenopus during metamorphosis, and failure of animals to metamorphose results in retention of larval histological features seen at the light-and electronmicroscopic levels (Clothier and Bails, 1985). The time frame in which the new adult-type T cells emerge appears to be quite broad. Although metamorphosis is generally complete by 10-12 weeks of age in our laboratory, a high proportion of class II / T cells is not observed until 5-6 months of age. Thus, it would appear that there is a period of at least two months after metamorphosis in which larval lymphocytes persist while adult-type lymphocytes are maturing. Because thymectomy just prior to metamorphosis does not significantly impair the ability of 2-month postmetamorphic frogs to reject skin allografts (Barlow and Cohen, 1983), persisting larval cells must be functional. This pattern of development may ensure that some functional tadpole cells remain to protect against pathogens, as suggested by Flajnik and Du Pasquier (1988). It raises, however, the question of how these functional tadpole cells tolerate the array of new adult-specific antigens without reacting against self. There may be suppressive mechanisms that prevent an antiself response while allowing a protective response to perchlorate-treated (@ @) frogs during ontogeny. Untreated controls and perchlorate-treated animals Were raised under identical conditions of animal density, temperature, nutrition, etc.. Each data point represents the mean+ SE of 5-12 individuals or pools of animals. pathogens to occur. The phenomenon of immunonens and ; DiMarzo and Cohen, suppression around the time of metamorphosis is 1982a, b; Obara et al., 1983;Rollins-Smith et al., well documented (reviewed in Du Pasquier et al., 1988), and development of an adult-type antibody 1989) and some type of suppressor cell appears to repertoire (Du Pasquier and Haimovich, 1976; Du be involved (Du Pasquier and Bernard, 1980;Barlow Pasquier et al., 1979;Hsu and Du Pasquier, 1984b). and Cohen, 1983). When metamorphosis was prevented by goitrogen-Our observation that development of a class II / treatment, the adult-type antibody pattern did not T-cell population is metamorphosis-dependent lends develop (Hsu and Du Pasquier, 1984b). Some other support to the general hypothesis that maturation of features of adult-type immunity appear to develop the immune system in Xenopus requires a normal independently of thyroid hormone influences. metamorphosis. Other features or functions of the Examples of this are development of high-titer IgY immune system that depend on normal metamor-antibodies to a T-dependent antigen by perchloratephosis are adult-type allograft rejection (Chardon-blocked permanent larvae, whereas very young larvae make IgY antibodies poorly (Hsu and Du MonoclonalAntibodies Pasquier, 1984a, b), and expression of class MHC antigens by perchlorate-blocked larvae (Flajnik et al., The monoclonal antibodies, 6.16, specific for the 1986;Flajnik and Du Pasquier, 1988). Because / chain of Xenopus IgM (Bleicher and Cohen, 1981), and 14A2, specific for a polymorphic determinant on expression of class antigens is not prevented when the Xenopus class II molecule (Flajnik et al., in metamorphosis is blocked, but expression of class II antigens on adult T cells is prevented, it seems clear press), were used for staining surface determinants on freshly prepared thymic and splenic lymphothat expression of each set of MHC antigens in adult lymphocytes is independently regulated. It cytes. Ammonium sulfate precipitated 6.16 was used at approximately 100/g/ml in amphibian phosphate seems likely, from the evidence so far accumulated, that the normal development of adult-type buffered saline (APBS) (6.6 g NaC1, 1.15 g Na2HPO4, immunity in frogs requires a postmetamorphic and 0.2 g KH2PO in 1000 ml of glass distilled water, expansion of both B and T lymphocytes. This expan-pH adjusted to 7.4) containing 2% fetal calf serum sion results in a different .allograft-rejection capa-and 0.2% sodium azide (PFA). Ammonium sulfate bility and a different antibody repertoire. If meta-precipitated 14A2 or culture supernatants from the morphosis is inhibited, the lymphocyte expansion is .hybridoma cell line that produces it were used at approximately 200/g/ml in PFA. greatly retarded and the resulting response capability reaches a more mature larval state that does not equal the adult. C'learly, metamorphosis Indirect Immunofluorescence presents a variety of problems for the developing immune system. We are beginning t understand Approximately 5 xl05 thymocytes or splenocytes how the frog solves these problems, but a greater were stained with 20/zl of one of the monoclonal understanding awaits further study, antibodies for 20 min at 4C. After a wash with 0.5 ml of cold (4C) PFA, the cells were stained with fluoresceinated goat antimouse immunoglobulin (polyvalent) antisera (Southern Biotechnology Asso-MATERIALS AND METHODS ciates, Inc., Birmingham, AL) at 100/g/ml in PFA for an additional 20 min at 4C. After the second Frogs antibody was washed out, the cells were resuspended at I xl06/ml in PFA and I xl05 were MHC homozygous J-strain (Tochinai and Katagiri, centrifuged onto microscopic slides with a cyto-1975; DiMarzo and Cohen, 1982b) males and females centrifuge (Shandon Southern Instrument, Inc., were induced to breed by injection of human Sewickley, PA). After air drying, the adherent cells chorionic gonadotropin according to standard prowere fixed for 20 min in a very cold (-16C) 95% cedures. Larvae were reared at approximately 8-10 ethanol, 5% glacial acetic acid solution. After rehytadpoles per 4 liters of dechlorinated tap water, dration in APBS and mounting, the slides were Water was changed and the animals were fed examined at 500-800x magnification with a Leitz powdered nettle leaf three times weekly. Post-Orthoplan fluorescent microscope with epilluminametamorphic frogs were fed freeze-dried tubifex tion. The percentage of cells with positive immunoworms (Wardley Product Company, Inc., Secaucus, fluorescence was determined after counting approxi-NJ) and small pieces of commercial trout chow mately 100-500cells of lymphocyte morphology. (Purina, Inc., St. Louis, MO). Larval stages were determined according to the Normal Table of Nieuwkoop and Faber (1967).

Flow Cytometry
Cells were prepared and stained as described before and fixed for at least I h in PFA containing 1% Inhibition of Metamorphosis formalin. Cells were analyzed on an EPICS 753 flow Beginning at 21-27 days postfertilization, some cytometer (Coulter Electronics, Hialeah, FL) using larvae were reared in dechlorinated tap water conthe 488-nm line of an argon laser operating at 25 A taining I g/1 of sodium perchlorate (Sigma Chemical with 500-mW output. Fluorescent signals passed Co., St. Louis, MO). Water was changed and the through a 457-502-nm laser-blocking filter, 515-nm animals were fed three times weekly, absorbance filter, and a 525-nm bandpass filter prior to amplification by an RCA 4526 photomultiplier tube (PMT). Viability was determined by propidium iodide dye exclusion of an unstained cell suspension. Electronic gates were defined on forward and 90 light-scatter parameters so as to select 5000 viable cells for subsequent immunofluorescence analyses.

Dual Fluorescence
The monoclonal antibodies, 6.16 and 14A2, were directly conjugated with fluorescein and rhodamine, respectively, according to standard protocols (Mishell and Shiigi, 1980). Each antibody stained with very dim but detectable fluorescence. To enhance the staining for microscopic examination, cells were indirectly stained with rhodamine-conjugated goat antimouse kappa light-chain-specific antibody (to detect the 14A2 monoclonal) and fluorescein-conjugated goat antimouse lambdachain-specific antibody (to detect the 6.16 monoclonal). By this method, we were able to detect cells positive for both IgM and class II antigens.