Shh/BMP-4 signaling pathway is essential for intestinal epithelial development during Xenopus larval-to-adult remodeling

Authors


Abstract

During amphibian larval-to-adult intestinal remodeling, progenitor cells of the adult epithelium actively proliferate and differentiate under the control of thyroid hormone (TH) to form the intestinal absorptive epithelium, which is analogous to the mammalian counterpart. We previously found that TH–up-regulated expression of bone morphogenetic protein-4 (BMP-4) spatiotemporally correlates with adult epithelial development in the Xenopus laevis intestine. Here, we aimed to clarify the role of BMP-4 in intestinal remodeling. Our reverse transcriptase-polymerase chain reaction and in situ hybridization analyses indicated that mRNA of BMPR-IA, a type I receptor of BMP-4, is expressed in both the developing connective tissue and progenitor cells of the adult epithelium. More importantly, using organ culture and immunohistochemical procedures, we have shown that BMP-4 not only represses cell proliferation of the connective tissue but promotes differentiation of the intestinal absorptive epithelium. In addition, we found that the connective tissue-specific expression of BMP-4 mRNA is up-regulated by sonic hedgehog (Shh), whose epithelium-specific expression is directly induced by TH. These results strongly suggest that the Shh/BMP-4 signaling pathway plays key roles in the amphibian intestinal remodeling through epithelial–connective tissue interactions. Developmental Dynamics 235:3240–3249, 2006. © 2006 Wiley-Liss, Inc.

INTRODUCTION

During amphibian metamorphosis, the intestinal epithelium undergoes dramatic remodeling from larval to adult form in the adaptation to terrestrial carnivorous life. The larval epithelium degenerates by apoptosis (Ishizuya-Oka and Ueda,1996). On the other hand, a small number of undifferentiated cells appear as islets (adult primordia) between the larval epithelium and the connective tissue (McAvoy and Dixon,1977). The adult primordia are of unknown origin, although there is a growing body of circumstantial evidence that they are transformed from differentiated larval epithelial cells (Marshall and Dixon,1978; Amano et al.,1998; Schreiber et al.,2005). The adult primordia include stem cells that are analogous to those in the mammalian intestine (Ishizuya-Oka et al.,2003), rapidly grow in size by active cell proliferation, and then differentiate into the intestinal absorptive epithelium similar to the mammalian counterpart (Hourdry and Dauca,1977; Shi and Ishizuya-Oka,1996). We previously indicated that the connective tissue, which develops during metamorphosis, is necessary for adult epithelial development in the Xenopus laevis intestine (Ishizuya-Oka and Shimozawa,1992), similar to the requirement of mesenchyme for epithelial development in the embryonic gut of higher vertebrates (Kedinger et al.,1988; Yasugi,1993). Thus, the amphibian intestinal remodeling serves as a good model system for studying mesenchymal factors that regulate stem cells and/or their descendants to form the intestinal epithelium of terrestrial vertebrates.

It is well established that amphibian metamorphosis is triggered by a single hormone, thyroid hormone (TH; Dodd and Dodd,1976; Kikuyama et al.,1993; Shi,1999). In addition, in the X. laevis intestine, larval-to-adult epithelial remodeling can be organ-autonomously induced by TH in vitro (Ishizuya-Oka and Shimozawa,1991). This finding suggests that molecular mechanisms of epithelial–connective tissue interactions necessary for the adult epithelial development can be studied from the perspective of examining the expression cascade of TH response genes. Of interest, we previously found that, whenever primordia of the adult epithelium grow in size, fibroblasts possessing well-developed rough endoplasmic reticulum aggregate just beneath the adult primordia both in vivo and in vitro (Ishizuya-Oka and Shimozawa,1992; Shi and Ishizuya-Oka,1996). These results led us to consider that TH-inducible endogenous factor(s) produced by such fibroblasts may be important for the adult epithelial development. Among numerous TH response genes isolated from the X. laevis intestine (Shi and Brown,1993; Amano and Yoshizato,1998; Shimizu et al.,2002), bone morphogenetic protein-4 (BMP-4) is noteworthy, because its expression profile examined by in situ hybridization (ISH) agrees spatiotemporally with the presence of fibroblasts described above (Ishizuya-Oka et al.,2001a).

BMP-4 is known to be one of the targets of sonic hedgehog (Shh) signaling (Perrimon,1995; Ingham,1998). Shh induces BMP-4 expression in the gut mesenchyme during early development of higher vertebrates (Roberts et al.,1995,1998; Sukegawa et al.,2000). In the X. laevis intestine, our previous study indicated that TH up-regulates the expression of BMP-4 mRNA indirectly through unknown epithelial factor(s) (Ishizuya-Oka et al.,2001a). On the other hand, Shh has been identified as a direct TH response gene (Shi and Brown,1993; Stolow and Shi,1995), and its epithelium-specific expression is the highest in the adult epithelial primordia (Ishizuya-Oka et al.,2001b). Considered together, it is highly possible that Shh may up-regulate BMP-4 expression in the amphibian postembryonic intestine. To demonstrate it experimentally, we examined whether exogenous Shh protein could induce the expression of BMP-4 mRNA in epithelium-free intestines of X. laevis tadpoles.

In the avian and mammalian embryonic guts, the ectopic expression of BMP-4 generally causes thinning of the mesenchymal layer (Roberts et al.,1998; Smith et al.,2000). However, far less is known about the function of BMP-4 in the epithelium, especially in the postembryonic intestine. In the present study, to clarify its function in the adult epithelial development, we first examined by ISH whether receptors for BMP-4 were actually expressed in the epithelium. We used a probe for one of type I receptors of BMP-4, BMPR-IA, which binds BMP-4 with a high affinity (Suzuki et al.,1994; ten Dijke et al.,1994; Dewulf et al.,1995; Ikeda et al.,1996). Next, we investigated effects of BMP-4 protein and its antagonist, Chordin, on the X. laevis intestinal remodeling by using the organ culture system established previously (Ishizuya-Oka and Shimozawa,1991). Although there are more than seven known antagonists of BMP-4 in vertebrates (Balemans and Hul,2002), we used Chordin to inhibit BMP-4 activity, because it binds BMP-4 with a high affinity (Piccolo et al.,1996). In addition, we previously identified a Xenopus homolog of Drosophila Tolloid closely related to BMP-1 (Tolloid/BMP-1), which is known to degrade Chordin (Piccolo et al.1997; Wardle et al.,1999; Blitz et al.,2000; Balemans and Hul,2002), as a TH response gene in the X. laevis intestine (Shimizu et al.,2002). Here, we show that Chordin is transiently expressed during intestinal remodeling and, by using its protein in vitro, that BMP-4 plays important roles in the adult epithelial differentiation as a connective tissue signaling factor.

RESULTS

BMP-4 Expression Is Up-Regulated Indirectly by TH Through Shh

To clarify whether BMP-4 expression is induced by Shh, we cultured tadpole intestines at stage 57 (before metamorphic climax) and examined the in vitro expression of BMP-4 mRNA by ISH. In intact intestines, BMP-4 mRNA became detectable after day 3 and attained its maximum level around day 5 in the presence of TH (Fig. 1A; Ishizuya-Oka et al.,2001a) but did not in its absence (Fig. 1B). However, in epithelium-free intestines, BMP-4 mRNA induced by TH was very little, if any, throughout the cultivation (Fig. 1C). In such intestines, when Shh protein was added to the medium, BMP-4 mRNA became detectable. Cells expressing BMP-4 mRNA were localized in the connective tissue but not in muscles (Fig. 1D), similar to those in the intact intestines treated with TH. In addition, to clarify whether Shh alone is sufficient to induce BMP-4 expression, Shh protein was added to the medium deprived of TH. BMP-4 mRNA induced by Shh was very little, if any, both in intact (Fig. 1E) and epithelium-free intestines (Fig. 1F).

Figure 1.

In vitro expression of bone morphogenetic protein-4 (BMP-4) mRNA in Xenopus laevis intestines cultured for 5 days and analyzed by in situ hybridization. A,B: Intact intestines. A high level of BMP-4 mRNA is localized in the connective tissue (CT) just beneath the epithelium (E) in the presence of thyroid hormone (TH, A) but not in its absence (B). Small arrowheads show the boundary between the epithelium and the connective tissue. C,D: Epithelium-free intestines cultured in the presence of TH. TH up-regulation of BMP-4 expression does not occur in the absence of Shh (C). When Shh protein (500 ng/ml) is added to the medium (D), BMP-4 mRNA becomes detectable in the connective tissue but does not in muscles (M). E,F: Intact (E) and epithelium-free (F) intestines cultured in the absence of TH. Even though Shh is added to the medium, BMP-4 mRNA remains undetectable in any tissue. Scale bars = 20 μm.

Expression of Chordin and BMPR-IA mRNAs During Intestinal Remodeling

To examine whether BMPR-IA and Chordin are actually expressed in the X. laevis intestine, the developmental expression of their mRNAs was analyzed by reverse transcriptase-polymerase chain reaction (RT-PCR). The level of Chordin mRNA abruptly increased toward stage 59 and peaked around stage 61 (Fig. 2), when the intestine undergoes extensive remodeling (Ishizuya-Oka and Ueda,1996). Then, its level decreased toward the end of metamorphosis (stage 66). On the other hand, the level of BMPR-IA mRNA increased after stage 57 and then peaked around stage 61 (Fig. 2). Thereafter, although its level gradually decreased, BMPR-IA mRNA remained detectable until the end of metamorphosis. Furthermore, to identify what cells express BMPR-IA mRNA, we performed ISH analyses in the X. laevis intestine.

Figure 2.

Developmental changes in BMPR-IA, a type I receptor of bone morphogenetic protein-4, and Chordin mRNAs in Xenopus laevis intestine analyzed by reverse transcriptase-polymerase chain reaction. The levels of both mRNAs reach their peak around stage 61, when the intestine undergoes extensive remodeling. The products for EF-1α mRNA serve as a control.

BMPR-IA mRNA Is Expressed in Both the Epithelium and Connective Tissue

The tadpole intestine before metamorphic climax consists of the differentiated larval epithelium, a thin layer of connective tissue, and inner and outer thin layers of muscles. BMPR-IA mRNA was not definitely detected in any tissue by ISH throughout the larval period (Fig. 3A).Then, toward stage 61, when primordia of the adult epithelium were identified as islets stained red with methyl green–pyronin Y (MG-PY) between the larval epithelium and the connective tissue (Fig. 3B), BMPR-IA mRNA became detectable. Cells expressing BMPR-IA mRNA were localized in both the adult epithelial primordia and developing connective tissue (Fig. 3C) but neither in the muscles nor the larval epithelium, which became more weakly stained with MG-PY because of apoptosis (Fig. 3B; Ishizuya-Oka et al.,2000). Thereafter, as morphogenesis of multiple intestinal folds proceeded, BMPR-IA mRNA gradually decreased in the connective tissue but remained detectable in the single layer of the adult epithelium until the end of metamorphosis (Fig. 3E). In contrast, all control sections hybridized with the sense RNA probe showed only background levels of signals throughout metamorphosis (Fig. 3D,F).

Figure 3.

Developmental expression of BMPR-IA, a type I receptor of bone morphogenetic protein-4, mRNA in Xenopus laevis intestine analyzed by in situ hybridization. A–F: Cross-sections were hybridized with the antisense (A,C,E) and sense probes (D, F) and stained with methyl green–pyronin Y (MG-PY; B). A: Stage 56. No specific signals are detected in any tissue. B–D: Stage 61. Primordia of the adult epithelium (arrowheads) appear as islets stained red with MG-PY between the larval epithelium (LE) and the connective tissue (CT) (B). Specific signals are detectable in both the connective tissue and the adult primordia (C). E,F: Stage 66 after morphogenesis of intestinal folds (Fo) and the larval-to-adult epithelial replacement. Specific signals are mostly localized in the adult epithelium (AE) (E). In control sections hybridized with the sense probe, only background signals are weakly detectable throughout metamorphosis (D,F). M, muscle. Scale bars = 20 μm.

Effects of BMP-4 and Chordin Proteins on Intestinal Remodeling In Vitro

To clarify the roles of BMP-4 in intestinal remodeling, we used the X. laevis organ culture system, where larval-to-adult epithelial remodeling is reproduced by the inductive action of TH (Ishizuya-Oka and Shimozawa,1991). In control tadpole intestines cultured for 5 days of TH treatment, when both epithelial and connective tissue cells most actively proliferate (Ishizuya-Oka and Shimozawa,1992), the primordia of the adult epithelium became detectable as islets between the degenerating larval epithelium and the connective tissue (Fig. 4A)At this time, proliferating cells were numerous in the adult epithelial primordia but very few in the degenerating larval epithelium (Fig. 4D). On the other hand, the layer of connective tissue became thick (Fig. 4G) by active cell proliferation (Fig. 4D). These changes mimic those occurring in normal intestines around stage 61. Thereafter, the adult primordia gradually lost the proliferating activity and, by day 7, differentiated into a simple columnar epithelium expressing intestinal fatty acid-binding protein (IFABP), a marker for intestinal absorptive cells (Fig. 4J).

Figure 4.

A–L: Effects of exogenous bone morphogenetic protein-4 (BMP-4) (100 ng/ml) and Chordin proteins (1 μg/ml) on tadpole intestines cultured for 5 days (A–I) and 7 days (J–L). Cross-sections were stained with methyl green–pyronin Y (MG-PY, A–C) and immunostained with anti–proliferating cell nuclear antigen (PCNA, D–F), anti-smooth muscle actin (G–I), and anti–intestinal fatty acid-binding protein (IFABP) antibodies (J–L). A,D,G,J: Control intestines cultured without exogenous bone morphogenetic protein-4 (BMP-4) and Chordin. Primordia of the adult epithelium (arrowheads) appear between the degenerating larval epithelium (LE) and the connective tissue (CT) on day 5 (A). Proliferating cells positive for PCNA are numerous in any tissue. In the epithelium, they are mostly localized in the adult primordia (D). B,E,H,K: In the presence of BMP-4. Proliferating cells on day 5 are numerous in the single layer of the epithelium (arrows) and the muscles (arrowheads) but few in the connective tissue (E), which is thinner (H) than in the control intestines (G). K,J: On day 7, IFABP immunoreactivity of the epithelium is much higher (K) than in the control intestines (J). C,F,I,L: In the presence of Chordin. Epithelial cells stained red with MG-PY (arrowheads) are fewer (C) than those in the control intestine (A). Proliferating cells on day 5 are numerous in the connective tissue (arrows) but few in the epithelium (F). The epithelium on day 7 is negative for IFABP (L). Scale bars = 20 μm.

When exogenous BMP-4 protein was added to the medium, the activity of cell proliferation on day 5 significantly decreased only in the connective tissue but did not in the other tissues (Figs. 4E, 5). The connective tissue was thinner (Fig. 4H) and lower in cell density (Fig. 4B) than that in the control intestines (Fig. 4A,G). In addition, BMP-4 caused precocious differentiation of the intestinal absorptive epithelium. Epithelial cells formed a simple columnar epithelium on day 5 (Fig. 4B) earlier than in the control intestines (Fig. 4A). Then, on day 7, IFABP immunoreactivity of the epithelium was much higher (Fig. 4K) than that in the control intestine (Fig. 4J). In contrast, the addition of Chordin to the medium significantly increased the proliferating activity in the connective tissue, but decreased that in the epithelium on day 5 (Figs. 4F, 5). The primordia of the adult epithelium underneath the degenerating larval epithelium were smaller (Fig. 4C) than those in the control intestines (Fig. 4A). On day 7, the simple epithelium was smaller in cell number than that in the control intestine, and its IFABP immunoreactivity of was very little, if any (Fig. 4L). On the other hand, neither BMP-4 nor Chordin proteins affected the proliferating activity of muscular cells (Fig. 5), which continued to express smooth muscle actin before and throughout intestinal remodeling (Fig. 4G–I).

Figure 5.

Addition of bone morphogenetic protein-4 (BMP-4) and Chordin proteins to the medium changes the proliferating activity of intestinal tissues cultured for 5 days. Each value represents the mean + SD of more than five explants. BMP-4 significantly inhibits cell proliferation in the connective tissue, whereas Chordin promotes its proliferation. **P < 0.01, *P < 0.05 compared with control explants.

DISCUSSION

BMP-4 Activity Is Regulated by TH-Induced Shh and Chordin

We previously reported that the expression of BMP4 mRNA is not up-regulated directly by TH, but indirectly through unknown epithelial factor(s) in the X. laevis intestine (Ishizuya-Oka et al.,2001a). On the other hand, during spontaneous metamorphosis, the level of BMP-4 mRNA becomes higher toward the adult epithelial primordia expressing the highest level of Shh (Ishizuya-Oka et al.,2001a,b), which is up-regulated directly by TH in the epithelium (Stolow and Shi,1995). In the present study, we have experimentally shown that Shh alone can induce the expression of BMP-4 mRNA in the connective tissue of epithelium-free tadpole intestines in the presence of TH, but not in its absence. This finding strongly suggests that TH–up-regulated expression of Shh induces BMP-4 expression during amphibian intestinal remodeling, although the competence of the connective tissue to Shh induction is TH-dependent. Similarly, in the embryonic gut of higher vertebrates, the endoderm-specific expression of Shh induces BMP-4 expression in the mesenchyme (Roberts et al.,1995,1998; Ramalho-Santos et al.,2000; Sukegawa et al.,2000).

Another finding of the present study is that the expression of Chordin is transiently up-regulated during intestinal remodeling. Its expression temporally coincides with that of Tolloid/BMP-1 (Shimizu et al.,2002) and a little earlier than that of BMP-4 (Ishizuya-Oka et al.,2001a), both of which have been identified as TH response genes that are expressed only in the connective tissue. In X. laevis early embryos, Chordin is degraded by Tolloid/BMP-1 and antagonizes BMP-4 activity in dorsoventral patterning of the body (Piccolo et al.,1997; Wardle et al.,1999; Blitz et al.,2000). Taken together, it is highly possible that BMP-4 activity during intestinal remodeling may be spatiotemporally regulated by Tolloid/BMP-1 through Chordin. The expression profile of Chordin in the remodeling intestine is worth investigating by ISH in future.

We previously indicated that the adult epithelial primordia express not only Shh but Musashi-1 (Ishizuya-Oka et al.,2003), a stem-cell marker common to mammalian intestinal stem cells (Potten et al.,2003; Okano et al.,2005). Also in the human adult intestine, Shh expression is restricted to the region corresponding to stem cells (van den Brink et al.,2002; de Santa Barbara et al.,2003) and up-regulated in the neoplastic or inflammatory intestine when the stem cells compensate for the epithelial damage (Berman et al.,2003; Nielsen et al.,2004). Therefore, it seems likely that the stem cells may make their niche adequate for the new epithelial formation by expressing Shh, which in turn up-regulates its target genes such as BMP-4 in the surrounding tissues.

BMP-4 Suppresses Cell Proliferation in the Connective Tissue

There is a growing body of evidence that BMP-4 is involved in both organogenesis of the embryonic gut and homeostasis of the adult intestine (de Santa Barbara et al.,2003; He et al.,2004). However, only fragmentary information is available regarding precise functions of BMP-4 at the cellular and molecular levels. In the present study, we performed the functional analysis of BMP-4 in the X. laevis postembryonic intestine by adding BMP-4 and Chordin proteins to the culture medium in vitro. It is generally known that Chordin antagonizes BMP-2 and -7 in addition to BMP-4 (Balemans and Hul,2002). However, BMP-2 and -7 have not yet been identified as TH response genes, and their mRNAs were not detected by RT-PCR throughout X. laevis intestinal remodeling (our unpublished data). On the other hand, the expression of BMP-4 mRNA is up-regulated during intestinal remodeling both in vivo and in vitro, where BMP-4 mRNA becomes detectable after 3 days of TH-treatment and reaches its peak around day 5 (Ishizuya-Oka et al.,2001a). Therefore, Chordin added to the medium on and after day 3 in the present study is supposed to have mainly inhibited the activity of BMP-4, although there still remains a possibility that Chordin also inhibited the activity of the other BMPs.

The present culture study has shown that functions of BMP-4 vary in different tissues and/or states of cytodifferentiation during intestinal remodeling. In the connective tissue, exogenous BMP-4 protein represses the proliferating activity of cells, which are immature and actively proliferate in the presence of TH (Shi and Ishizuya-Oka,1996). This BMP-4 function agrees with our previous report that the proliferating activity of connective tissue cells in the remodeling intestine becomes lower toward the subepithelial region (Ishizuya-Oka and Ueda,1996), where BMP-4 mRNA is most highly expressed (Ishizuya-Oka et al.,2001a). In this regard, the function of BMP-4 antagonizes Shh, which promotes cell proliferation of the connective tissue (Ishizuya-Oka et al.,2001b). Similarly, in the embryonic gut of higher vertebrates, BMP-4 represses the proliferating activity of mesenchymal cells (Smith et al.,2000; Batts et al.,2006), suggesting that the BMP-4 function is conserved among the species and that its effects last in the postembryonic intestine. In contrast, in the present study, BMP-4 had no effect on cell proliferation of muscles, another mesenchymal tissue, in the X. laevis intestine, in agreement with no expression of BMPR-IA mRNA in the muscles. Because, unlike the mesenchyme of the embryonic gut, the muscles of stage 57 tadpole intestines used for the present culture experiments are morphologically (Kordylewski,1983) and functionally differentiated as shown by the immunoreactivity for smooth muscle actin (the present study), the competence of the mesenchyme to respond to BMP-4 may reduce with the progress of its differentiation into the muscles.

BMP-4 Promotes Adult Epithelial Differentiation

In comparison with the function of BMP-4 in the gut mesenchyme, its function in the endoderm and/or epithelium is still less defined. In the amphibian intestine (present paper), we have shown here that BMPR-IA mRNA is expressed in the adult epithelial primordia but not in the larval epithelium throughout metamorphosis. This finding is in agreement with our previous observations that the expression profile of BMP-4 mRNA spatiotemporally coincides well with the adult epithelial development both in vivo and in vitro (Ishizuya-Oka et al.,2001a) and suggests that BMP-4 mainly affects the adult epithelial cells through its receptors, but not the larval epithelial cells. More importantly, we used the organ culture system, where adult epithelial primordia grow by active proliferation on day 5 (Ishizuya-Oka and Shimozawa,1991,1992) and then differentiate into the intestinal absorptive epithelium expressing IFABP on day 7 (Ishizuya-Oka et al.,1997), and demonstrated that BMP-4 promotes differentiation of the adult epithelial primordia, whereas Chordin inhibits their differentiation. This BMP-4 function agrees with its expression profile that the level of BMP-4 mRNA peaks close to the adult epithelial primordia just before the beginning of differentiation (Ishizuya-Oka et al.,2001a). The level of BMP-4 mRNA then decreases but remains detectable in the subepithelial fibroblasts after metamorphosis, when the cell renewal system of the adult epithelium is established along the trough–crest axis of intestinal folds similar to that in the mammalian intestine (McAvoy and Dixon,1977). In the present study, BMPR-IA mRNA also remained detectable in the adult epithelium after metamorphosis. Therefore, it seems likely that BMP-4 may continue to function as a connective tissue signaling factor to maintain the cell renewal system of the intestinal absorptive epithelium. It is worth clarifying in the future the direct effects of BMP-4 on the epithelium alone by establishing a culture technique for the amphibian intestinal epithelium in the absence of the connective tissue.

In the mammalian intestine, recent studies reported that BMP-4 signaling represses crypt formation and polyp growth (Haramis et al.,2004; Batts et al.,2006) through suppression of Wnt signaling in epithelial proliferating cells (He et al.,2004). However, in the present study, exogenous BMP-4 had no effect on cell proliferation of the epithelium on day 5, when a small number of adult progenitor cells appear and most actively proliferate, whereas the larval epithelium undergoing apoptosis hardly proliferates (Ishizuya-Oka et al.,2001b,2003). Because several TH response genes other than BMP-4 are concomitantly expressed by the induction of TH during intestinal remodeling (Shi,1999) and some of them such as TH/bZip are known to activate cell proliferation of the epithelium (Ikuzawa et al.,2006), interactions with such genes may make it difficult to detect effects of BMP-4 alone on the epithelial cell proliferation. The only result we obtained concerning epithelial proliferation is that excessive Chordin, that is, BMP-4 deficiency, results in the reduction of proliferation and the total number of adult progenitor cells. This finding suggests a possibility that a certain amount of BMP-4 may be essential for the maintenance and/or self-renewal of the intestinal stem cells, as proposed in other organs such as the ovary (Xie and Spradling,1998; Fujiwara et al.,2001) and embryonic stem cells (Ying et al.,2003; Qi et al.,2004). To address its functions in the epithelial cell proliferation more precisely, further studies are necessary to distinguish effects of BMP-4 on stem cells from those on their descendant proliferating cells.

In conclusion, we have shown here that the Shh/BMP-4 signaling pathway is one of the key signaling pathways involved in epithelial–connective tissue interactions that are essential for amphibian postembryonic intestinal remodeling (Fig. 6). TH directly up-regulated the expression of Shh (Stolow and Shi,1995), which in turn promotes the proliferating activity of connective tissue cells as an epithelial signaling factor (Ishizuya-Oka et al.,2001b). In addition, Shh induces subepithelial fibroblasts near the adult primordia to express BMP-4 mRNA. Then, BMP-4 secreted by the fibroblasts represses cell proliferation of the connective tissue, whereas it promotes adult epithelial differentiation (the present study). Recently, in the human digestive tract, the growing body of evidence indicated that mutations in members of the Shh/BMP-4 signaling pathway are associated with different malformations and diseases such as juvenile polyposis syndrome (Howe et al.,2001; de Santa Barbara et al.,2003; Haramis et al.,2004). The amphibian intestinal remodeling, where this pathway can be experimentally activated by TH in vitro, provides a good model for understanding conserved roles of the pathway in the digestive tract of terrestrial vertebrates. Future studies should be directed toward more precise functional analysis of genes involved in this pathway by combining the culture system with recent transgenic technology (Fu et al.,2002; Buchholz et al.,2004; Ikuzawa et al.,2006).

Figure 6.

Schematic drawing showing potential roles of sonic hedgehog/ bone morphogenetic protein-4 (Shh/BMP-4) signaling pathway in amphibian intestinal remodeling. Thyroid hormone (TH) directly up-regulates the expression of Shh in the epithelium (Stolow and Shi,1995) and that of Tolloid/BMP-1 in the connective tissue (Shimizu et al.,2002). Shh protein promotes cell proliferation in the connective tissue and induces subepithelial fibroblasts to produce BMP-4, which in turn represses cell proliferation in the connective tissue and promotes differentiation of the adult epithelium. BMP-4 activity may be regulated by Tolloid/BMP-1 through degrading Chordin. Arrows indicate activation, and T-shaped bars indicate repression.

EXPERIMENTAL PROCEDURES

Animals

Tadpoles of the South African clawed frog (Xenopus laevis) at stages from 55 to 66 (completion of metamorphosis; Nieuwkoop and Faber,1967) were purchased from local suppliers and used throughout the experiments.

RT-PCR

Total RNA was extracted from the small intestine of the tadpoles using Trizol reagent (Gibco-BRL). RNAs were prepared from a mixture of more than three tadpoles at each stage. For each reaction, 1 μg of RNA was reverse transcribed to oligo (dT)-primed first-strand cDNA by using a cDNA synthesis kit (Amersham Pharmacia Biotech) and resulting cDNA was subjected to 30 cycles of PCR with Chordin-specific primers (5′-AACTGCCAGGACTGGATGGT-3′ and 5′-GGCAGGATTTAGAGTTGCTTC-3′; Blitz et al.,2000), to 30 cycles with BMPR-IA–specific primers (5′-GTTGAAGAGTAC- CAACTGC-3′ and 5′-AGTTCTGATACCAGTGTCC-3′; Suzuki et al.,1995), and to 25 cycles with elongation factor 1α (EF-1α)-specific primers (5′-CCTGAATCACCCAGGCCAGATTGGTG-3′ and 5′-GAGGGTAGTCTGAGAAGCTCTCCACG-3′; Suzuki et al.,1993). These PCR cycles were limited to the linear phase of amplification. The PCR products were electrophoresed through a 2% agarose gel and visualized by ethidium bromide staining.

Organ Culture

Intestinal fragments were isolated from the anterior part of the small intestine of X. laevis tadpoles at stage 57 (before metamorphic climax). Some of them were treated with 1,000 U/ml dispase (Godo) to remove their epithelial components. The intact and epithelium-free intestines were then cultured as described previously (Ishizuya-Oka and Shimozawa,1991). Briefly, they were placed on membrane filters (type HAWP; Millipore) on stainless steel grids and cultured in 60% Leibovitz-15 medium (Gibco-BRL) supplemented with 100 IU/ml of penicillin, 100 μg/ml of streptomycin, and 10% charcoal-treated fetal bovine serum (CTS medium, Gibco-BRL). To induce metamorphosis, T3, insulin, and hydrocortisone (Sigma) were added to CTS medium at 10 nM, 5 μg/ml, and 0.5 μg/ml, respectively (TH-containing medium). To examine effects of BMP-4 on intestinal remodeling, recombinant human BMP-4 (R&D Systems, 0–100 ng/ml), which is biologically active and 98% identical to Xenopus BMP-4 (Nishimatsu et al.,1992), and mouse Chordin (R&D Systems, 0–1,000 ng/ml), which is known to possess the functional domains for BMP binding, were used. They were added to TH-containing medium on and after day 3, just before TH–up-regulated expression of BMP-4 in vitro (Ishizuya-Oka et al.,2001a). In addition, to examine whether Shh can induce BMP-4 expression, recombinant mouse amino-terminal Shh (R&D Systems) was added to CTS medium or TH-containing medium at 500 ng/ml as used previously (Ishizuya-Oka et al.,2001b). The culture medium was changed every other day for 7 days at 26°C.

ISH

Procedures for ISH were the same as those described previously (Ishizuya- Oka et al.,2001a). In brief, digoxigenin (DIG) -11-UTP-labeled antisense and sense probes were prepared according to the manufacturer's instruction of a DIG RNA labeling kit (Roche Diagnostics) by using Xenopus BMPR-IA (Suzuki et al.,1995) and BMP-4 cDNAs (Nishimatsu et al.,1992). Intestinal fragments were fixed with 4% paraformaldehyde in phosphate buffered saline (pH 7.4) at 4°C for 4 hr, frozen on dry ice, and cut at 7 μm. Sections were treated with 0.2 N HCl, digested with 1 μg/ml proteinase K (Wako) at 37°C for 5 min, and fixed again with 4% paraformaldehyde. Hybridization buffer containing DIG-labeled antisense or sense RNA probe (200 ng/ml for BMPR-IA and 50 ng/ml for BMP-4) was applied to the sections. After hybridization at 40°C 18 hr, the sections were treated with 20 μg/ml RNase A (Wako) at 37°C for 30 min to remove excess unhybridized probes. They were then washed, processed for immunological detection of the hybridized DIG probes with a DIG-probe Detection Kit (Roche Diagnostics), and examined under the microscope.

Immunohistochemistry

Intestinal fragments were fixed with 95% ethanol at 4°C for 4 hr, embedded in paraffin, and cut at 5 μm. To identify proliferating cells, some sections were incubated with the mouse anti–PCNA antibody (Novacastra, 1:100) at room temperature for 1 hr. They were then incubated with peroxidase-labeled streptavidin (Nichirei) and reacted with 0.02% 3,3′-diamino-benzidine-4HCl (DAB) and 0.006% H2O2. The labeling indices were calculated as the ratio of labeled nuclei to total nuclei in more than five sections randomly selected from each explant cultured for 5 days. More than five explants were examined for each experimental point. Results were statistically analyzed by the analysis of variance (ANOVA) and Student's t-test. If the difference was significant based on the ANOVA analysis, the difference between the control and each group was further analyzed by Student's t-test.

To identify differentiated intestinal absorptive cells and muscular cells, other sections were incubated with the rabbit anti-Xenopus IFABP antibody (a generous gift from Dr. Y.-B. Shi, 1:500) and with the mouse anti–α smooth muscle actin antibody (Sigma, 1:400), respectively, at room temperature for 1 hr. They were then visualized by sequential incubation with streptavidin–biotin–peroxidase complex and DAB/H2O2 as described above. There was no positive staining when the same concentration of preimmune serum was applied as the specificity control (data not shown). In addition, to distinguish adult epithelial primordia from the larval epithelium conveniently, some of the sections prepared for ISH or immunohistochemistry were stained with MG-PY (Muto) for 5 min. Our previous studies amply showed that adult epithelial primordia are stained red intensely because of their RNA-rich cytoplasm, whereas the staining intensity of larval epithelial cells become weaker as they undergo apoptosis during larval-to-adult epithelial remodeling both in vivo and in vitro (Ishizuya-Oka and Ueda,1996; Ishizuya-Oka et al.,1997,2000).

Acknowledgements

We thank Profs. K. Yoshizato (Hiroshima University) and Y.-B. Shi (National Institutes Health) for their valuable advice and help.

Ancillary