Skin is comprised of an epidermis (epithelium) derived from ectodermal tissue and an underlying dermis (mesenchyme) that is of mesodermal origin. During the formation of skin and its appendages, the epithelium and mesenchyme are inducers and targets of each other (Sengel, 1976). Two cutaneous structures, scales and feathers, have been studied extensively to define the molecular mechanism of their morphogenesis (Prin and Dhouailly, 2004; Lin et al. 2006).
Retinoic acid is known to regulate cell proliferation, differentiation, and morphogenesis during normal development of many tissues (Gudas, 1994; Duester, 2008). It induces a catagen-like stage in human hair follicles, presumably via up- regulation of TGF-β2 in the dermal papilla, resulting in hair loss (Foitzik et al., 2005). Retinol deficiency in rabbits can cause squamous metaplasia and keratinization in a wide variety of nonkeratinized and secretory epithelia (Mori, 1922; Wolbach and Howe, 1925). The effects of retinoic acid are mediated through 2 families of nuclear receptors, i.e., the retinoic acid receptors (RARs) and their isoforms and the retinoid X receptors (RXRα, β, and γ; Zelent et al., 1989). These receptors mediate ligand-dependent target gene transcription, typically by binding as heterodimers to cis-acting retinoic acid response elements (Chambon, 1996). Many of the abnormalities in pattern formation and organ formation that result from the exogenous addition of retinoic acid during embryogenesis are related in part to the ability of retinoids to change the pattern of expression of the clusters of homeobox (Hox) genes in the embryo (Chuong et al., 1992; Cardoso et al., 1996).
Fell and Mellanby (1953) first found that excess vitamin A can induce transdifferentiation of chick embryonic epidermis to a mucous epithelium. Earlier we examined the mechanism of the mucous transdifferentiation induced by retinoic acid and showed that in chick embryonic tarsometatarsal skin, this transdifferentiation to a mucous epithelium is induced by interaction of the epidermis with the dermis, when dermal fibroblasts are pretreated with retinol (Obinata et al., 1987, 1991a,b, 2001). Recently, we showed that retinoic acid induces transdifferentiation of epidermis of rat embryonic skin to an esophagus-like mucosal epithelium (Obinata et al. 2011).
Hox genes are master control genes that specify the body plan and regulate the development and morphogenesis of higher organisms (Gehring et al., 1994). Apart from these genes, there are other groups of homeobox genes (divergent homeobox genes), which are located outside the Hox loci and play an important role in regulating growth and differentiation during the development of many organs. Using a degenerate RT-PCR (reverse transcriptase-polymerase chain reaction)-based screening method, we previously isolated divergent homeobox genes Gbx1 (Obinata et al., 2001), Hex (Obinata et al., 2002, 2005a,b), and HB9 (Kosaka et al., 2000a,b) from chick embryonic tarsometatarsal scale skin in addition to Hox genes. Furthermore, we showed that among the many homeobox genes isolated, Gbx1 shows a strong increase in expression in the epidermis of tarsometatarsal scutate scale skin during the course of retinol-induced epidermal transdifferentiation to mucosal epithelium and that the expression is induced by interaction of the epidermis with retinol-pretreated dermal fibroblasts (Obinata et al., 2001).
The aim of this study was to elucidate (1) whether Gbx1 is expressed in the feather bud of dorsal skin, (2) whether retinoic acid increases the expression of Gbx1 mRNA in the dorsal skin, and (3) whether Gbx1 transient overexpression induces the transdifferentiation to a mucous metaplasia. This study showed that Gbx1 was expressed in the feather bud of dorsal skin and that treatment with 1 μM retinoic acid for 24 hr induced transdifferentiation of the epidermis to mucosal epithelium with a concomitant increase in Gbx1 mRNA expression in the epidermis in an organ culture of chick embryonic dorsal skin. Furthermore, we showed that transient transfection of the epidermis with Gbx1 cDNA by electroporation induced elongation of the feather buds, but did not lead to epidermal mucous transdifferentiation.
aa amino acid DAPI 4′, 6-diamidino-2- phenylindole DIG digoxigenin MV microvilli PAS periodic acid-Schiff PBS phosphate-buffered saline.
Expression of Gbx1 Gene and Gbx1 Protein in the Epithelium of the Feather Bud of Chick Embryos
By in situ hybridization analysis, we showed for the first time that in addition to its expression in the intestines and tarsometatarsal scale skin (Obinata et al., 2001), Gbx1 was expressed in the posterior region of the feather placode (primordium of the feather bud) as early as the placode stage (Fig. 1A, B). Subsequently, the Gbx1 expression became more intense in the epidermal basal cells of the posterior region of the bud at the short-bud stage (Fig. 1C, D) and then throughout the epidermis at the long-bud stage (Fig. 1E, F). However, the expression was very low in the interbud regions at all stages (Fig. 1B, D, F). Weak Gbx1 expression was also seen in the dermis of the long bud (Fig. 1F).
By Western blotting using anti-Gbx1 antiserum, we detected a specific band of Gbx1 protein (243aa, 26 kDa) in the chick embryonic dorsal skin (Fig. 2A). Gbx1 expression significantly increased during feather bud development, whereas actin expression did not change (Fig. 2A).
By using this antiserum, we examined immunohistochemically the distribution of the protein product of the Gbx1 gene. The immunostaining pattern of Gbx1 protein was not always consistent with the in situ hybridization pattern of Gbx1 mRNA. A weak immunoreaction indicated Gbx1 protein was observed in the whole epidermis of chick embryos at the placode stage (Fig. 2B, E) and, subsequently, an intense one was seen in the whole epidermis of both short (Fig. 2C, F) and long (Fig. 2D, G) buds, and its expression was increased during development. Gbx1 protein was detected in both nucleus and cytoplasm (Fig. 2C, D, F, G). Gbx1 protein expression was also seen in the dermis of both short and long buds (Fig. 2C, D).
Retinoic Acid-Induced Transdifferentiation of Epidermis to Mucosal Epithelium With Concomitant Inhibition of Feather-Bud Formation
When dorsal skin was cultured with or without retinoic acid at the optimal concentration of 1 μM in the presence of hydrocortisone for 1 day followed by incubation for 4 days without these chemicals, the elongation of the feather bud was inhibited in the retinoic acid-pretreated skin (Fig. 3A), whereas elongated buds were observed in the control skin (Fig. 3B). Cells in the upper layer of the epidermis in both bud and interbud regions of dorsal skin pretreated with retinoic acid were round and contained PAS-positive substances, but those in the corresponding regions of the control skin were flat and scarcely contained PAS-positive materials (Figs. 3C, D, 4A–D). Transmission electron microscopy indicated the presence of well-developed microvilli, Golgi apparatus, and a discontinuous basement membrane in the retinoic acid-pretreated skin (Fig. 3E, G). In the control skin, the microvilli and Golgi apparatus were not well developed, and the basement membrane was continuous (Fig. 3F, H). Interestingly, mucous-containing granules surrounded by a limited membrane were not observed in the retinoic acid-treated dorsal skin (Fig. 3E) but were previously observed in the retinoic acid–induced mucous epithelium in tarsometatarsal scale skin (Obinata et al., 1991a).
Retinoic Acid–Induced Gbx1 mRNA Expression in Dorsal Epidermis
As we had previously reported that retinoic acid increases the expression of Gbx1 mRNA in the epidermis of tarsometatarsal scale skin (Obinata et al., 2001), we examined whether retinoic acid could also induce such an increase in the dorsal skin. In situ hybridization analysis indicated that, while Gbx1 mRNA was expressed in the epidermis at placode, short-bud, and long-bud stages (Fig. 1B, D, F), Gbx1 expression in the control skin decreased during incubation for 5 days (Fig. 4F). On the other hand, retinoic acid increased Gbx1 expression throughout the whole epidermis and in the dermal cells compared with the expression in the control skin (Fig. 4E).
Elongation of Feather Bud by Transfection of the Epidermis With Gbx1 cDNA
To study whether transient overexpression of Gbx1 would induce the transdifferentiation of the epidermis to a mucous one (metaplasia), we transfected the epidermis of the dorsal skin with cGbx1-pEGFP or cGbx1-pcDNA3 for transient expression and then cultured it for 2 or 5 days in DMEM. Transfection with pEGFP was carried out by electroporation, as reported previously (Obinata et al., 2005a,b). To verify that the skin had been indeed transfected, we observed the fluorescent image of Gbx1-pEGFP protein (Fig. 5A, B) and Gbx1 mRNA expression by in situ hybridization (Fig. 5 C, D). Two days after the epidermis had been transfected with cGbx1-pEGFP or pEGFP, the cGbx1-pEGFP fusion protein (Fig. 5A) and pEGFP protein (Fig. 5B) were expressed throughout the epidermis. Although cGbx1 mRNA expression increased in the cGbx1-pEGFP-transfected epidermis (Fig. 5C), no increase in it was seen in the pEGFP-transfected epidermis (Fig. 5D). After 5 days in culture following transfection, the feather buds showed a significant increase in length compared with the control (Fig. 5E–I). However, neither rounded nor PAS-positive cells were observed among the superficial cells (Fig. 5G, H) of the Gbx1-transfected skin, the cells of which had been seen in the retinoic acid-pretreated epidermis (Fig. 4A, C), indicating that transdifferentiation to a mucous epithelium had not been induced by Gbx1. The same result was obtained with cGbx1-pcDNA3 (data not shown).
In the present study, we showed for the first time that the Gbx1 gene was expressed in the feather bud. Studies on Gbx1 have concentrated on the brain or neurons. For example, in zebra fish, Gbx1 and Otx2 are among the earliest genes expressed in the neuroectoderm, dividing it into anterior and posterior domains with a common border that marks the midbrain–hindbrain boundary primordium (Rhinn et al., 2004). In mice, during development Gbx1 is expressed in the central nervous system (Waters et al., 2003, Rhinn et al., 2004) and combined expression of Lhx7 and Gbx1 plays a role in the development of the cholinergic system of the basal forebrain (Asbreuk et al., 2002). However, few studies on Gbx1 in other tissues have been performed. Previously we reported that Gbx1 is expressed in the epidermis of tarsometatarsal scale skin of chick embryos (Obinata et al., 2001). Here, we showed that Gbx1 was also expressed in the feather buds of the dorsal epidermis of chick embryos as early as at the placode stage and that this expression increased during feather-bud development. Hence, Gbx1 might be required for morphogenesis of the epithelial structures of the integument, such as scales and feathers.
In the feather bud, homeobox genes other than Gbx1 are reported to be expressed, such as Gbx2 (Niss and Leutz, 1998); Msx-1 and Msx-2 (Noveen et al., 1995), HB9 (Kosaka et al., 2000b); Hox b-4, Hox a-7, and Hox c-8 (Reid and Gaunt, 2002); Dlx 2, 3, 5 (Rouzankina et al., 2004); and Hex (Obinata et al., 2005a,b). Gbx2 is a homeobox gene closely related to Gbx1 and is required for proper segregation of early regional identities anterior and posterior to the mid-hindbrain boundary in the case of vertebrates (Asbreuk et al., 2002). In the present study, we demonstrated that Gbx1 was expressed mainly in the epithelium of the feather bud. In contrast, Gbx2 is expressed in the mesenchyme of the feather bud (Niss and Leutz, 1998). The gene expression pattern of Gbx1 in the epidermis of feather bud was almost the same as that of the HB9 homeobox gene, as we reported previously (Kosaka et al., 2000b). Other than homeobox genes, the genes encoding morphogenetic proteins such as fibroblast growth factor (FGF; Widelitz et al., 1996), bone morphogenetic proteins (BMPs; Noramly and Morgan, 1998), sonic hedgehog (Shh; Bitgood and McMahon, 1995; Noveen et al., 1996; Ting-Berreth and Chuong, 1996), Wnt 7a (Widelitz et al., 1999), Notch, Delta, Serrate (Chen et al., 1997), and β-catenin (Noramly et al., 1999) are involved in the early stage of feather development (Lin et al., 2006). The present study indicated that the expression pattern of Gbx1 mRNA was not consistent with the patterns of these signals. Further study on the relationships among Gbx1, other homeobox genes, and morphogenetic signals has to be carried out to elucidate the processes underlying the development of feather buds.
In this study, we showed that treatment with 1 μM retinoic acid at stage 31 or 33 inhibited feather-bud formation and induced transdifferentiation of the epidermis to a non-keratinized stratified mucous epithelium concomitant with an increase in Gbx1 expression as microvilli developed on the upper surface of the epithelium. Also, a well-developed Golgi apparatus and PAS-positive materials were observed in the treated epidermis. These results are consistent with our previous reports stating that epidermal mucous transdifferentiation is induced by retinoic acid in chick embryonic tarsometatarsal skin at stage 39 along with an increase in Gbx1 expression and that typical goblet cells are induced in it by retinol (Obinata et al., 1987, 1991b, 2001). However, mucous granules surrounded by a limited membrane were not seen presently in the dorsal skin, indicating that retinoic acid acted differently depending on the area of the skin and on the developmental stages of the appendages, scale, or feather.
Modulation of the axis orientation by treatment with 1 μM retinoic acid throughout the culture period results in a higher frequency of small feather buds and many buds showing random orientations (Chuong et al., 1992). However, in the present study 1 μM retinoic acid was used for treatment for only 1 day, along with 10 nM hydrocortisone, and the skin was cultured for an additional 4 days without these chemicals, and a different result was obtained. Retinoic acid did not modulate the axis orientation, but rather caused the transdifferentiation of the epidermis to a mucous epithelium. Thus retinoic acid has important roles in transdifferentiation as well as in modulation of pattern formation.
As retinoic acid induced Gbx1 mRNA expression in the epidermis, we also examined whether Gbx1 alone (without retinoic acid) could induce epidermal transdifferentiation. Feather buds were elongated by transient transfection of the dorsal epidermis with Gbx1 cDNA, indicating stimulation of epidermal cell proliferation by Gbx1; however, the transdifferentiation was not induced. These results suggest that Gbx1 alone could not induce mucous transdifferentiation and that other signals playing a key role cooperatively with Gbx1 are involved in the epidermal mucous transdifferentiation.
We conclude that Gbx1 was involved in feather-bud formation, being one of the target genes of retinoic acid in the dorsal skin of chick embryos, and that transfection of the epidermis with Gbx1 cDNA induced elongation of the feather buds, but did not induce transdifferentiation. Further study is required to demonstrate the role of Gbx1 and other morphogenetic signals in the formation of feather buds and in retinoic acid–induced transdifferentiation. We are now studying the relationships among other signals and Gbx1 by using rat embryonic dorsal skin to elucidate the mechanism of retinoic acid–induced epidermal transdifferentiation as more information becomes available.
Fertilized White Leghorn eggs were incubated at 38°C. The embryos were staged according to Hamburger and Hamilton (1992). All animal experiments were carried out in accordance with the guidelines of the Committee of Animal Care and Experiments of Teikyo University. For the experiments, 7- to 9-day-old embryos (HH stage 31–35) were used.
Dorsal skin between the lower neck and tail was removed from chick embryos at stage 31 or 33, attached to a Millipore (Billerica, MA) filter (0.45-μM pore) with the dermis side toward the filter, and the filter was then laid on a culture insert (Corning, Lowell, MA). DMEM containing 5% delipidized FCS and 10 nM hydrocortisone with or without 1 μM all-trans-retinoic acid was placed in the outside well and in the inner chamber. A thin layer of the medium was left in the inner chamber to keep the explants moist and to provide an air-liquid interface. The explants were cultured for 1 day and, after having been washed with phosphate-buffered saline (PBS), were incubated in DMEM alone for 2 or 4 days at 37°C in a humidified incubator having an atmosphere of 5% carbon dioxide and 95% air.
Preparation of Digoxigenin (DIG)-Labeled RNA Probes
The RNA probe used for the Gbx1 homeobox gene was prepared as described previously (Obinata et al., 2001). Additionally, we prepared another type of Gbx1 RNA probe from a genomic cGbx1 clone, one corresponding to the upstream of cGbx1, which corresponds to the coding region of Gbx1 in the case of zebra fish and mouse. A DNA fragment consisting of 129 bp was amplified from the genomic cGbx1 clone by using the following primers from 5′ to 3′: cccca tgttcatgccgta and cgcagaacgtgttggtga, corresponding to 129 bp of the sequence. The fragment was then cloned into the pT7Blue T-vector, and the resulting clone was sequenced in its entirety.
In Situ Hybridization
In situ hybridization with the DIG-labeled probe was performed as described previously (Kosaka et al., 2000a). As the 2 types of DIG-labeled RNA probes had the same strength and pattern when used for in situ hybridization, we chose the Gbx1 RNA probe prepared from the genomic cGbx1 clone.
Rabbit antiserum was generated against a KLH-conjugated chick Gbx1 NH2 peptide (MVALTTALPSFSEPC). Rabbit monoclonal anti-actin antibody (clone EP184E) was obtained from Epitomics (Burlingame, CA). For immunostaining and Western blotting, anti-Gbx1 antiserum was diluted 1:100 and 1:300, respectively. Anti-actin antibody was diluted 1:1,000 for Western blotting.
Back skins of chick embryos at the placode stage, short-bud stage, and long-bud stage were taken, homogenized in PBS, pH 7.2, containing protease inhibitors. Following centrifugation, the supernatant was mixed with Laemlli sample buffer and boiled for 5 min. The same amount of proteins (10 μg) was loaded into each well, and the proteins were electrophoresed in 12% SDS-polyacrylamide slab gels and subsequently transferred to a PVDF membrane. The membrane was reacted with rabbit anti-Gbx1 antiserum, followed by horseradish peroxidase (HRP)-conjugated goat anti-rabbit IgG. The protein band was detected by use of enhanced chemiluminecence (ECL). The intensity of Gbx1 protein in the bands was quantified by scanning densitometry. Blots were stripped and then reprobed with anti-actin antibody to compare the actin expression with Gbx1 expression.
Skin explants were processed for light and electron microscopic observations as described previously (Obinata et al., 1991a). Sections were stained with periodic acid-Schiff (PAS), which stains glycogen and neutral and acidic glycoproteins.
Immunostaining was performed as described previously (Kosaka et al., 2000a,b; Obinata, 2002).
A Gbx1 cDNA fragment containing the entire cGbx1 coding region from a chick liver genomic library, which was kindly provided by Dr. A. Kuroiwa, was amplified with a sense primer with a HindIII site (5′-gaaaagcttaagatggtggccctcaccacg-3′) and an antisense primer with an EcoR1 site (5′-aaatagaattcaagggtcgggccc cctgctc-3′). The PCR product was ligated into the HindIII-EcoR1 site of the pEGFP-N3, pcDNA3, or pcDNA3.1/Hygro vectors. The sequence of the inserts was confirmed after cloning.
Electroporation of plasmids into the dorsal skin of chick embryos taken at stage 31 or 33 and attached to a Millipore filter (0.45-μm pore) was performed as described previously (Obinata et al., 2005b). The data of this electroporation experiment were compiled from 3 independent experiments. The length of feather buds was measured 5 days after electroporation. Student's t-test was used for statistical analysis.
We thank Dr. A. Kuroiwa (Division of Biological Science, Nagoya University, Japan) for providing a chick liver genomic library and Dr. S. Waters (Laboratory of Cancer and Developmental Biology, NCI-Frederick, National Institute of Health, Frederick, MD) for providing a mouse Gbx1 cDNA. We are grateful to Ms. S. Matubara, Ms. M. Kanai, Ms. T. Miura, and Ms. S. Koroishi (Laboratory for Electron Microscopy and Department of Anatomy, Kyorin University School of Medicine) for their technical support. This work was supported, in part, by Grants-in-Aid from the Ministry of Education, Science, Sports, Culture and Technology, Japan (O.A.), and from the Kazato Research Foundation (Y.A.).