*Author to whom all correspondence should be addressed. Graduate School of Science, Niigata University, Niigata 950-2181, Japan. Email: firstname.lastname@example.org
Results from previous studies using an inbred strain of Xenopus laevis have led to the proposition that metamorphosis includes the events by which the newly differentiating adult immune system, including T lymphocytes, recognizes and eliminates larval skin cells as ‘non-self’. More recently, a larval antigen targeted by adult T cells was identified as a 59 kDa protein with a specific peptide sequence. Using antisera directed against the larval antigen and the peptide, immunohistochemistry and western blotting were done to examine expression of the 59 kDa larval antigen in the skin during larval and metamorphic periods. There was no expression before Nieuwkoop and Faber stage 53. Expression was first seen at the beginning of metamorphic stage 54, when hind limbs appear, and increased thereafter, in apical and skein cells of both trunk and tail regions. In the trunk region, expression started to decrease at stage 58, until it completely disappeared at stage 62 (metamorphic climax). In the tail skin, however, expression persisted throughout the metamorphic stages. Treatment of larvae with thyroid hormone (TH) resulted in repression of expression of the 59 kDa molecule in a dose-dependent manner. Downregulation occurred earlier in the trunk than in the tail skin. These results suggest involvement in metamorphic events of an immunological mechanism: differential expression of the larval antigen in the trunk and tail skin cells due to their differing concentration of TH results in the tail, but not the trunk skin, being selectively attacked by the newly differentiating adult-type immune system.
During metamorphosis of anuran amphibians, the skin tissues are laid under a unique situation in that the epidermal cells show region-specific changes in histology. In the early stages of larval development, the epidermis of both the trunk and tail regions is comprised of larva-specific apical and skein cells. During metamorphosis, the trunk epidermal layer thickens and develops basal cells which survive and give rise to adult-type germinative basal cells after proliferation and differentiation (Robinson & Heintzelman 1987; Izutsu et al. 1993). In contrast, the tail skin remains as a larval tissue and is completely eliminated during metamorphosis (Izutsu et al. 1993).
The immune system also shows dramatic metamorphic changes. Functional differences between larval and adult immunocompetent cells have been shown by the mixed lymphocyte reaction (MLR) method (Kobel & Du Pasquier 1977; Du Pasquier et al. 1979), and larval-type immunocompetent cells are replaced by adult-type cells after hormone-induced cell death during metamorphosis (Katherine et al. 1997; Rollins-Smith 1998). An allotolerant state to grafted skin is easily induced in larvae, whereas it is difficult to induce in adult frogs (Chardonnens & Du Pasquier 1973; Barlow & Cohen 1983; Obara et al. 1983). The co-existence of larval- and adult-type tissue cells and immune cells in an individual metamorphosing amphibian means that complicated interactions of constituent cells are required to achieve coordinated tissue remodeling. In spite of these challenges in metamorphic transformation, involvement of the immune system in destruction of larval-specific tissues has scarcely been reported.
In our previous experiments using an inbred strain of Xenopus laevis, we found that trunk skin grafted from tadpoles older than the late metamorphic climax stage were accepted by young adults, but tail skin grafts were rejected irrespective of the metamorphic stage of the donor (Izutsu & Yoshizato 1993). This rejection was considered to be of an immunological nature because of an accelerated secondary graft rejection response, a significant proliferative response by splenocytes from adults and metamorphosing larvae against larval tail tissues, as well as an adult splenocyte-induced apoptotic destruction of larval tail pieces in vitro in the presence of a homologous serum (Izutsu et al. 1996). These results, together with in vitro analyses of T-cell function in these responses, have led us to propose that newly differentiating adult-type immune cells recognize and eliminate larval cells as ‘non-self’ targets, and predict that larva-specific antigens recognized by adult T cells may be expressed on all larval epidermal cells (Izutsu et al. 1996, 2000a).
More recently, by immunizing adult frogs with syngeneic larval skin grafts, we succeeded in producing anti-larval skin (LS) antibodies which react specifically with larval but not with adult skin lysates (Izutsu et al. 2002). A major 59 kDa protein detected by the anti-LS antibodies was selected for determination of partial amino acid sequence. Based on the sequence, a 19 amino acid peptide (larval peptide, LP) was chemically synthesized for raising a rat antiserum, which proved to be specifically reactive with a single band at 59 kDa in a western blot of larval skin, which corresponded to one of the larval antigens estimated by the anti-LS antibodies (Izutsu et al. 2002). This LP induced the augmented proliferative response of cultured splenocytes from LS-immunized adult frogs. Evidence so far obtained from western blotting, immunofluorescence studies, and adult splenic T-cell proliferative responses in vitro has indicated that the 59 kDa protein is one of the candidate antigens targeted to be eliminated by the immune system during metamorphosis.
In this study, we used the anti-LS and anti-LP antibodies to determine the temporal and spatial distribution of the 59 kDa larval antigen in the skin during premetamorphic and metamorphic stages, as well as the suppressive effect of thyroid hormone (TH) on expression of the larval antigen. We propose that the difference in the fates of trunk and tail skin during metamorphosis is based on differential expression of larval antigen in cells due to differential concentration of TH.
Materials and Methods
Individuals of X. laevis from a major histocompatibility complex (MHC) homozygous J strain were used throughout this study (Nakamura et al. 1985). In these animals there was no rejection of skin grafts exchanged between adult individuals (Izutsu & Yoshizato 1993). Larvae and adults were reared at 23 ± 1°C. Tadpoles were staged according to the normal table of Nieuwkoop & Faber (1956).
Generation of anti-LS frog antiserum and anti-LP antiserum
The frog antiserum against LS and the rat polyclonal antibodies against LP were produced as described previously (Izutsu et al. 2002).
Briefly, for immunizing adult frogs with LS, tail skin (including fins) excised from tadpoles at stage 55–57 was grafted under the dorsal trunk skin of a 1–2-year-old female frog, three times at 1 month intervals. Clean serum was obtained and was stored at −80°C until use.
A synthesized peptide (SFDNYAQAADITSALSDMK) that significantly induced adult T-cell proliferation (Izutsu et al. 2002) was coupled with keyhole limpet hemocyanin (KLH), mixed with an equal amount of complete Freund's adjuvant, and injected into a rat. Two weeks later and every 2 weeks after that, booster shots were given four times with the antigen and incomplete Freund's adjuvant. The highest titer was obtained 1 week after the final boost. The anti-LP rat serum thus obtained was stored at −80°C until use.
Skin tissues from both tail and dorsal trunk regions were dissected from tadpoles at stage 47–63. Frozen tissue sections (4 µm in thickness) of each tissue were dried without fixation and stored at 4°C until use (within 1–2 weeks). For immunohistochemistry, the sections were blocked with 10% fetal calf serum (FCS) in phosphate-buffered saline (PBS) for 30 min, and treated with 10-fold diluted anti-LS frog serum for 12 h at 4°C, followed by treatment with the monoclonal antibody 11D5 (a mouse anti-Xenopus IgY (IgG analog)) at a 1:200 dilution in PBS (Hsu & Du Pasquier 1984). Staining was detected by using Cy3-conjugated secondary antibodies against mouse Ig (Chemicon International, Temecula, CA, USA) at 1:300 in PBS. As a negative control for staining, sections were treated with normal Xenopus serum in place of primary anti-LS antibodies. After washing with PBS, counterstaining was done with 0.025% quinacrine dihydrochloride (Sigma, St Louis, MO, USA).
Sodium dodecylsulfate–polyacrylamide gel electrophoresis and western blotting
For gel electrophoresis and western blot analysis, skin tissues excised from the tail (including fins) and the trunk of individuals at stage 48–66 were quickly frozen in liquid nitrogen. The frozen tissues were homogenized using Polytron (Brinkmann Instrument, Westbury, NY, USA) in a lysis buffer (5% sodium dodecylsulfate (SDS), 100 mm Tris-HCl (pH 7.5) and 1 mm phenylmethanesulfonyl fluoride (PMSF)). The supernatant after centrifugation at 10 000 g for 10 min (4°C) was subjected to SDS–polyacrylamide gel electrophoresis (PAGE) on a 10% polyacrylamide gel under 2-mercaptoethanol (2-ME) conditions, with a total protein concentration of 50 µg per lane. The gel was electroblotted onto an immobilon membrane (Nihon Millipore, Tokyo, Japan). After blocking with 5% non-fat dry milk, the membrane was reacted with the anti-LP rat antiserum at 1:100. The membrane was washed, and visualized with alkaline phosphatase-goat antirat antibody (Zymed Laboratory, San Francisco, CA, USA) as previously described by Yamashita et al. (1992). Negative controls were treatments without primary antibodies. As a loading control for staining, tail tissue lysate was treated with antihuman muscle actin mouse monoclonal antibody (Dako, Carpinteria, CA, USA) at 1:100.
Immunohistochemical analysis of larval antigen molecules during development, studied by anti-LS frog antiserum
Expression patterns of larval antigens during metamorphosis of X. laevis were studied by immunofluorescent staining using the anti-LS antiserum. In the early stages of larval development, signal was seen in neither the tail nor trunk epidermis (Fig. 1A,C). Positive staining was first seen at the premetamorphic stage 54 when small hind limbs appeared, and many skein cells localized on the basement membrane of the epidermis were stained (Fig. 1E, arrowheads). However, approximately half of the skein cells were still unstained, and most of the apical cells localized at the surface of the epidermis were also negative. Similar staining patterns were observed in the dorsal trunk epidermis (data not shown). At metamorphic stage 59, when the fore limbs develop and the epidermal layer becomes thicker, most of the epidermal cells, including both apical and skein cells in the tail, were stained, showing the expression of antigen molecules both on the cell surface and in the cytoplasm (Fig. 1G,H). In the dorsal trunk skin, however, the cells localized in the bottom half of the epidermal layer became negative, suggesting their transformation into newly differentiated adult-type cells (Fig. 1H). At late metamorphic stage 63, when the tail starts to regress, the larval antigens continued to be expressed in all of the epidermal cells in the tail region, but disappeared completely in the trunk skin (Fig. 1I).
Western blot analysis of larval antigen molecule expression during development, studied by anti-LP rat antiserum
Skin lysates from various developmental stages were analyzed for change of the larval antigens during metamorphosis by western blotting, using the anti-LP rat antiserum. At larval stage 48, the antigen molecule was not detected at all in the tail (Fig. 2a). At stage 50, when the rudiments of hind limbs appear, a faint signal was first seen in both the tail and trunk lysate (Fig. 2a,b). The signal was detected as a clear band at stage 54, when positive staining was first observed in the section (Fig. 1), and increased thereafter until stage 57 (Fig. 2a). A similar pattern of expression was observed in both the trunk and tail skin lysates, indicating no apparent difference in the timing of appearance of the antigen molecules in these regions (Fig. 2b). The signal started to reduce in the trunk skin at stage 58, and disappeared by late metamorphic stage 62 (Fig. 2b). In the tail tissues, however, the antigen persisted with the same intensity during stage 58 and 62 (Fig. 2b). The band of actin (as a control) did not change its expression level during the course of larval development (Fig. 2a). The actin control was not shown for the trunk lysates because muscle tissues were included in the trunk skin used for preparation of the lysates (see Materials and Methods).
Effect of thyroid hormone on expression of larval antigen molecule
After 4 days of TH treatment of tadpoles at stage 49 and 54, there was no morphological change in the tail tissues (length and thickness; data not shown). However, TH repressed the expression level of the 59 kDa larval antigen in a concentration-dependent manner (Fig. 3a). During 4 days of treatment from stage 49, larvae developed to stage 51, at which time the larval antigen (LP) was weakly expressed in the lysates from whole tail tissues of control animals (Figs 2,3a). However, expression of the antigen was completely inhibited by TH at 1 × 10−8m (Fig. 3a), and inhibition of TH was weaker at lower TH concentrations, with no influence at 1 × 10−11m. The bands of actin were not notably changed during the course of TH treatment. Similar results were seen in treatment from stage 54, at which time expression of the antigen molecule had already started (Fig. 2). During 4 days of treatment, control animals remained at the same stage (54), and the tails of hormone-treated animals did not change in morphology, but expression of the larval antigen molecule was reduced significantly (Fig. 3b), although actin bands were not affected.
When the stage 50/51 larvae were treated with TH for 4 days, the tail did not change, whereas the trunk region showed morphological changes typical to spontaneous metamorphosis (Fig. 4a). Western blotting analysis of the 59 kDa antigen molecule indicated that on day three of treatment with 1 × 10−8m TH, the larval antigen molecule completely disappeared in the trunk skin but was still expressed in the tail (Fig. 4b, upper). After 4 days of treatment with the same concentration of TH, the molecule was not observed in the tail (Fig. 4b, lower). Thus, expression of the 59 kDa antigen was repressed by TH treatment 1 day earlier in the trunk than in the tail skin. At 1 × 10−8m TH for 4 days, there was no appreciable difference between tail and trunk region with respect to expression of the 59 kDa antigen molecule (Fig. 4b).
To be a target for immunological elimination as larval tissues, the expression pattern of pertinent molecules must correlate both temporally and spatially with the tissue remodeling processes during metamorphosis. We show in this study that the 59 kDa larval antigen molecule detected by our antisera fits well with this prediction in that it appears in association with metamorphosis and has a differential expression pattern in the trunk and tail skin, in accordance with the difference in their fates after metamorphosis. A number of previous observations have predicted the occurrence of larval antigen molecules whose expression patterns were quite similar to what our 59 kDa antigen showed in this study. First, a fragment peptide of the 59 kDa larval antigen molecule activates proliferation of adult T cells in vitro (Izutsu et al. 2002). Second, the tail tissue pieces from the metamorphic (stage 56/57; Izutsu et al. 1996) but not from the premetamorphic (stage 50) larvae elicit a pronounced proliferative response of the syngeneic adult T cells in vitro (unpubl. data, 2002). Third, skin from a donor at the metamorphic stage is rejected more acutely than skin from premetamorphic donors (Izutsu & Yoshizato 1993). Finally, the persistent expression of our 59 kDa molecule in the tail, in contrast with the stage-dependent repression in the trunk demonstrated in this study, provides strong support for the notion that this molecule contributes to successful metamorphosis by exposing itself as a target for elimination by adult immunocompetent cells. There may be other larval antigen molecules that were not detected by our antisera, as suggested by our previous observation that skin grafts from premetamorphic stage (46/47), when our 59 kDa antigen is not detectable, were also rejected (Izutsu & Yoshizato 1993).
It should be noted that the immune response to larval antigens depends also on the ability of recognition by immunocompetent cells in metamorphosing host individuals. Allogeneic rejection in the murine system is known to be prolonged in the absence of donor MHC molecules (Mannon et al. 2002). Thus, the first expression of our 59 kDa antigen in skein cells in the absence of MHC molecules (Izutsu et al. 2000a) will actually not stimulate immune responses at early metamorphic stages. The recognition of 59 kDa antigen molecules by the adult immune cells will take place for the first time at the metamorphic climax stage, when the antigen molecules are expressed in the skein cells localized in the tail, but not in the trunk, skin. In contrast, the apical cells expressing MHC molecules from early metamorphic stage 55/56 onwards (Izutsu et al. 2000a) will stimulate immune responses for the first time at stage 59, simply because the 59 kDa antigen expression commences at this stage.
Our results indicate that the 59 kDa antigen may be included in the list of epithelial tissue molecules, such as keratin (Wang & Brown 1993; Watanabe et al. 2001) and cadherin (Izutsu et al. 2000b), whose expression is regulated by TH during metamorphic periods. Perhaps a more important aspect in relation to the tissue remodeling process during metamorphosis is the demonstrated differential responses of the tail and trunk tissues to TH, that is conceivably responsible for the difference in the cell fates. It is pertinent to mention that the concentration of TH in tadpoles in vivo is lower in the tail than in the trunk tissues, due to the differential activities of type II iodothyroninedeiodinase (D2) in different tissues (Huang et al. 2001). Because the turning off of the 59 kDa antigen requires relatively higher concentrations of TH, the tail but not the trunk tissue may constitutively express the antigen due to lower concentrations of the hormone, resulting in exposure of a target molecule in the tail tissue for immunological elimination during metamorphosis. In contrast, the trunk tissue can escape attack by the adult T cells recognizing 59 kDa molecules.
The splenocytes from late metamorphic stage 63–65 but not from early metamorphic stage 56/57 tadpoles display a prominent proliferative response against stage 56/57 larval tail tissues, as the adult splenocytes do (Izutsu et al. 1996), indicating that the adult-type T cells recognizing the larval tissues as foreign newly differentiate at the metamorphic climax. The metamorphosing animal will fall into a problematic situation in that the larval and newly emerging antigens and larval and emerging adult-type T cells are intermixed together within the body. In this respect, a large scale death of larval lymphocytes concomitant with these metamorphic events (Rollins-Smith 1998) may well relieve newly differentiating adult tissue antigens from immunological attack by host immune cells. In contrast, the newly differentiating adult immune cells are exposed to the conditions in which larval antigens are confined to the tail region in an individual during metamorphosis. Under such conditions, the immune cells will not fall into tolerant states to decreasing larval antigens, as shown in the murine system where selective elimination of responder T cells is completely abrogated in cases where the number of tolerogen-producing cells is decreasing (Fukushi et al. 1990).
Our finding of the LS antigen expressed in the metamorphic period combined with its spatially regulated expression pattern provides a novel view of hormone action in metamorphosis. Previous interpretation of TH influence has been simply based on concentration-dependent, direct hormonal action on the target tissues. We propose that the higher concentration of hormone as in the trunk skin tissue turns off the larval antigenic molecules expressed in the metamorphic period, thereby rescues tissues from elimination by the immune system, in contrast with the tail tissue, which is targeted for immunological attack by newly emerging adult-type immune cells, due to persistent expression of larval antigens in lower hormonal environments. There is also a possibility that TH promotes differentiation of adult immune cells during metamorphic periods and the tail tissues are destroyed by the adult immune system rather than by direct action of hormone on the tail cells. The immune system conceivably plays more important roles than thought previously in the tissue remodeling process in amphibian metamorphosis.
We thank Dr Chiaki Katagiri for critical discussion. We are also grateful to Dr Masayuki Hatta at Ochanomizu University for helpful advice and suggestions during preparation of this manuscript.
This study was partially supported by a Grant-in-Aid to Y. I. from The Sumitomo Foundation, Japan, The Uchida Energy Science Promotion Foundation, Japan, and Narishige Zoological Science Award, Japan.