The differentiation patterns of animal cap explants from the Japanese salamanders Hynobius lichenatus and Hynobius nigrescens were examined after exposure to various concentrations of activin A. A wide range of concentrations of activin A (0.5-100 ng/ml) induced various mesodermal tissues such as ventral mesoderm, somitic muscle, and notochord. At concentrations higher than 50 ng/ml, yolk-rich endodermal tissue was induced in many of the explants. Activin A is also known to have mesoderm-and/or endoderm-inducing activity on the animal caps of Xenopus laevis and Cynops pyrrhogaster, but their response patterns are slightly different. The mode of differentiation of activin-treated Hynobius animal caps was compared with that of Xenopus and Cynops in relation to the structure of the animal caps.
INTRODUCTION
The animal cap assay (Yamada and Takata, 1961), a simple system using isolated ectoderm (animal cap) as the responding tissue, has enabled remarkable advances in the identification of inductive factors in recent years (reviewed in Asashima, 1994). Peptide growth factors belonging to the transforming growth factor-β (TGF-β) family can induce animal caps to differentiate into mesodermal and/or endodermal tissues that normally arise in the vegetal half of the embryo (reviewed in Ariizumi and Asashima, 1995a). Most studies on such “vegetalization” of animal caps by growth factors have been carried out with Xenopus laevis (reviewed in Klein and Melton, 1994). For example, we previously reported that activin A, a member of the TGF-β family, is capable of inducing various mesodermal tissues in Xenopus blastula animal caps in a concentration-dependent manner (Ariizumi et al., 1991a, b). Low doses of activin A induced ventral mesoderm (mesenchyme and coelomic epithelium), while higher doses induced dorsal axial mesoderm (muscle and notochord). However, the response patterns of animal caps from different species to activin A may differ slightly. No clear concentration-dependent effect of activin A is observed in the Japanese newt, Cynops pyrrhogaster, and the frequency of mesoderm differentiation (e.g., muscle) is generally lower than in Xenopus animal caps (Moriya and Asashima, 1992). On the other hand, marked induction of yolk-rich endodermal tissues is observed in Cynops animal caps (Ariizumi and Asashima, 1995b; Ninomiya et al., 1998), especially at a high dose of activin A (100 ng/ml).
As shown in Fig. 1, the late blastula animal cap of Xenopus is composed of more than one layer (i.e., an outer epithelial cell layer and an inner blastocoelic layer of 2-3 cells thick) whereas that of Cynops consists of just one layer. These structural differences in the animal caps may be responsible for their different responses to activin A. Although the Japanese salamanders Hynobius lichenatus and Hynobius nigrescens are classified as urodeles, the same as Cynops, their late blastula or early gastrula animal caps consist of more than one layer, the same as the animal cap of Xenopus. In this study, we investigated the response patterns of Hynobius animal caps to exposure to activin A and compared them to those of Xenopus (Ariizumi et al., 1991a) and Cynops (Moriya and Asashima, 1992) in relation to the structure of the animal caps.
MATERIALS AND METHODS
Embryo and preoperative treatment
Egg capsules of H. lichenatus and H. nigrescens were collected in Yamagata Prefecture in the Tohoku district of Japan on April 26, 1998. They were stored at 4°C until the embryos reached the desired stage. H. lichenatus embryos were staged according to Sawano (1947), and H. nigrescens embryos according to Usui and Hamasaki (1939). Egg capsules were opened with iridectomy scissors. Embryos were sterilized in 70% ethanol for 1 min and then washed with modified Holtfreter's solution (MHS; 60 mM NaCl, 0.7 mM KCl, 0.9 mM CaCl2, 4.6 mM HEPES, 0.1 g/l kanamycin sulfate, pH 7.6). Jellycoats were chemically removed with MHS containing 1% sodium thioglycolate (pH 9.0).
Operation and culture
Operations were performed under sterile conditions at 10°C. After removing their jellycoats, the embryos were placed in 3% agarcoated Petri dishes filled with MHS. Vitelline membranes were peeled off the embryos with watchmaker's tweezers. The H. lichenatus animal cap was dissected from early gastrulae (stage 11; about 3.0 mm in diameter) with tungsten needles. It was 1.8 mm × 1.8 mm in size and contained 5046 ± 120 cells. The H. nigrescens animal cap was dissected from late blastulae (stage 10; about 2.6 mm in diameter). It was 1.4 mm × 1.4 mm in size and contained 1668 ± 65 cells. The animal caps were then transferred to the activin A solutions with their surface cell side face down.
Human recombinant activin A was kindly provided by Dr. Y. Eto (Central Research Laboratories, Ajinomoto Co. Inc., Japan) and dissolved in MHS containing 0.1% BSA (A-7888, Sigma, USA) at concentrations of 0, 0.1, 0.5, 1, 5, 10, 50, 100 and 500 ng/ml. The solutions were placed in 96-well plates (SUMILON®; MS-309UR, Sumitomo Bakelite, Japan). The animal cap explants were cultured in the activin A solutions at 10°C for 3 weeks, during which time both the H. lichenatus and H. nigrescens control embryos reached stage 41.
RESULTS
At the end of the culture period, when the control embryos had reached stage 41 (Fig. 2A), the animal cap explants of H. lichenatus exhibited four different patterns of morphological changes depending on the concentration of activin A (Fig. 2B-E). The H. nigrescens animal caps generally showed similar patterns. The results of the histological analysis of H. lichenatus and H. nigrescens explants are summarized in Table 1 and Table 2, respectively.
Table 1
Differentiation of H. lichenatus animal caps treated with activin A
Table 2
Differentiation of H. nigrescens animal caps treated with activin A
Control explants cultured without activin A became smaller and very wrinkled (Fig. 2B). They formed atypical epidermis without any other differentiation (Fig. 3A). Similar explants were obtained when treated with 0.1 ng/ml of activin A. At 0.5 ng/ml of activin A, part or all of the explant swelled as a result of absorbing fluid (Fig. 2C). Well-developed epidermis lined with ventral mesoderm (mesenchyme and coelomic epithelium) was found in addition to atypical epidermis (Fig. 3B). Dorsal mesoderm, i.e., notochord and muscle, was also seen in 44% and 38%, respectively, of the H. lichenatus explants and in 59% and 24% of the H. nigrescens explants. Explants often formed a trunk and tail with fins when cultured in 1 ng/ml of activin A (Fig. 2D). Axial organs such as the notochord, somitic muscle, and the neural tube were organized along the axis (Fig. 3C). Similar explants with axial structures were obtained by treatment with 5 ng/ml of activin A. At 10 ng/ml of activin A, notochord and muscle were still induced at high frequencies (100% and 100% in H. lichenatus explants; 77% and 49% in H. nigrescens explants), however, most of them were not surrounded by epidermis (Fig. 2E, Fig. 3D). At concentrations of activin A higher than 50 ng/ml, yolk-rich endodermal tissue was frequently induced in addition to dorsal mesoderm (Fig. 3E). It occurred in most of the explants treated with 100 ng/ml of activin A (Fig. 3F). No other differentiation besides endodermal tissue could be seen in the H. nigrescens explants when incubated in 500 ng/ml of activin A (Table 2). The differentiation pattern of H. nigrescens explants was basically the same as that of H. lichenatus explants. However, the frequency of neural differentiation tended to be higher, and that of dorsal mesoderm (notochord and muscle) to be slightly lower, than in the H. lichenatus explants (Table 1). Although the frequency was relatively low, ventral mesoderm (mesenchyme and coelomic epithelium) was induced by a wider range of concentrations of activin A (0.5-50 ng/ml) in the H. nigrescens explants.
DISCUSSION
The array of tissues induced in Hynobius animal caps depended largely on the concentration of activin A added to the culture medium. As the concentrations increased, activin A induced ventral mesoderm, then dorsal mesoderm accompanied by neural tissue, and finally yolk-rich endoderm (Table 1 and Table 2). Activin A thus appears to have dose-dependent mesodermand endoderm-inducing activity on Hynobius animal caps, the same as in other species previously examined (Ariizumi et al., 1991a, b; Moriya and Asashima, 1992). Activin A had an effect on Hynobius animal caps even at 0.5 ng/ml, a minimum dose similar to that required to induce Xenopus and Cynops animal caps.
The differentiation pattern of activin-treated Hynobius explants, however is slightly different from that of Xenopus. In Xenopus, various mesodermal tissues, from ventral to dorsal, are induced at high frequencies at clear dose thresholds (Ariizumi et al., 1991a). In Hynobius, dorsal mesoderm was consistently induced at high frequencies by activin A at a broader range of concentrations (0.5-50 ng/ml). On the other hand, the frequency of ventral mesoderm differentiation was relatively low (e.g., coelomic epithelium was induced in less than 38% of H. lichenatus explants and less than 15% of H. nigrescens explants). Furthermore, the concentration range required for ventral mesoderm differentiation largely overlapped the range for dorsal mesoderm differentiation. These results imply that the Hynobius animal caps are in a more “dorsalized” state than Xenopus animal caps. In Xenopus, it is known that animal cap cells on the prospective dorsal side tend to form dorsal mesoderm when exposed to activin A, whereas cells on the ventral side form ventral mesoderm when they are exposed to the same concentration of activin A (Sokol and Melton, 1991, 1992; Ariizumi and Asashima, unpublished). The high sensitivity of Hynobius animal caps to induction of dorsal mesoderm differentiation can be ascribed to their dorsalized state, the same as the prospective dorsal region of Xenopus animal caps. This tendency was clear in the H. lichenatus animal caps, in which dorsal mesoderm was always induced at high frequencies (Table 1).
The differentiation pattern of Hynobius explants is very different from that of Cynops explants, in which the frequency of mesoderm differentiation is very low (Moriya and Asashima, 1992). The structural differences in animal caps seem to be related to the mode of their response to activin A. Although the Cynops animal cap consists of a single layer of homogeneous cells, Hynobius and Xenopus animal caps consist of more than one layer containing different types of cells (Fig. 1). Activin A cannot act on the deeper or superficial cells, only on the blastocoelic cells lining the animal cap. The high frequencies of mesoderm differentiation, especially of dorsal mesoderm, in Hynobius animal caps thus appear to be due to their heterogeneous structure. We recently confirmed that mesodermal tissues are frequently induced even in the activintreated Cynops animal caps (Ariizumi et al., 1999). The degree of mesoderm differentiation depended on the size of the animal caps. As the size of animal caps increased, more types of mesodermal tissue were induced in addition to endodermal tissue. Furthermore, a well-organized axial structure composed of dorsal mesoderm and a central nervous system was frequently induced in the “heterogeneous” explants in which activin-treated animal caps (fated to form endoderm) were combined with untreated animal caps. These findings suggest that further inductive interactions between induced and non-induced cells should be considered when the animal cap assay is performed with heterogeneous animal caps such as those of Xenopus, Ambystoma, and Hynobius.
Acknowledgments
This work was supported by Grants-in-Aid for Scientific Research from the Ministry of Education, Science, Sports and Culture of Japan, and by CREST (Core Research for Evolutional Science and Technology) of the Japan Science and Technology Corporation.