DEVELOPMENTAL DYNAMICS 207:157-170 (1996) Sonic Hedgehog in Feather Morphogenesis: Induction of Mesenchymal Condensation and Association With Cell Death SHEREE A. TING-BERRETH AND CHENG-MING CHUONG Department of Pathology, University of Southern California, Los Angeles, California 90033 Sonic hedgehog is involved in ABSTRACT vertebrate tissue interactions during development. During early feather development, Sonic hedgehog appears very early in epithelial placodes. During late feather development, Sonic hedgehog expression precedes the development of the marginal plates and is specifically localized in the marginal plate epithelium, which will later undergo cell death. By using retroviral vectors, exogenous Sonic hedgehog overexpression in devel. oping feathers induced enlarged feather buds that have either lost their anterior-posterior polarity or exhibited reverse orientation. The enlarged dermal condensations may be mediated through broader TGF-P2 expression and reduced protein kinase C (PKC) expression. Reciprocal mesenchymal interaction is required for the induction and maintenance of Sonic hedgehog in the epithelial placodes. In scaleless mutant, Sonic hedgehog is absent in the apteric region and aberrantly expressed in the mesenchyme of the abnormal feather ridge. These findings suggest that Sonic hedgehog mediates key interactions between the epithelium and mesenchyme during feather morphogenesis. o 1996 Wiley-Liss, Inc. Key words: Morphogenesis, Hair, Feather, Skin appendage, Scaleless, Epithelialmesenchymal interaction, Retroviral gene delivery, Pattern formation, Cell death. INTRODUCTION Pattern formation is a critical process during early vertebrate development. Coordinated regulation of specific molecular events and cellular signaling is required for the arrangement of functional tissues and organs. Many of these processes involve secondary inductions that require specific and complex tissue interactions during their development. For example, limb, tooth, kidney, and skin appendage inductions require epithelial-mesenchymal interactions (reviewed in Sawyer and Fallon, 19831, and the induction of the floor plate in the central nervous system requires interactions with the notochord (van Straaten et al., 1988; Placzek et al., 1990). The molecular basis underlying the inductive events during these tissue interactions remains one of the most challenging questions in developmental biology. 0 1996 WILEY-LISS, INC. Recent studies have shown that Sonic hedgehog (Shh),a member of the hedgehog (hh)multigene family that encodes putative signaling molecules, is expressed in tissues with inductive and polarizing activities in various vertebrate embryos (Riddle et al., 1993; Krauss et al., 1993; Echelard et al., 1993). In the developing limb, Shh is expressed in the zone of polarizing activity (ZPA). The ZPA is a region of posterior limb bud mesenchyme that is essential for anterior-posterior limb patterning. Ectopic expression of Shh in the anterior limb bud induced digit duplication with mirror image symmetry (Riddle et al., 1993; Chang et al., 1994). Expression of Shh in ventral neural structures such as the notochord and floor plate suggests a role in the dorsalventral patterning of the neural tube (Echelard et al., 1993). Indeed, the ectopic expression of Shh can induce ectopic floor plate formation (Roelink et al., 1994, 1995). Shh has also been implicated in the dorsal-ventral patterning of somites (Fan and Tessier-Lavigne, 1994; Johnson et al., 1994) and in several other sites of organogenesis (Bitgood and McMahon, 1995). These studies suggest Shh may play an inductive and fundamental patterning role in various vertebrate tissue interactions. We are interested in examining the molecular basis underlying secondary induction using chicken feather morphogenesis as the model (Chuong, 1993). Skin appendages develop from the inductive interactions between the epithelium and the underlying mesenchyme. Upon receiving inductive signals from the underlying mesenchyme, the epithelium segregates into two domains, the placode and the interplacode domains (Sengel, 1976). Mesenchymal cells then condense beneath the placodes to form dermal condensations. In mature feathers, the epithelium segregates further into alternating domains destined to become the barbs (barb plate epithelium) and the space between barbs (marginal plate epithelium). How are these morphogenetic processes regulated? With recent studies demonstrating the versatile role of Shh in different tissue interactions, it is reasonable that Shh should be involved in feather formation. We therefore started to explore the roles of Shh in feather morphogenesis. Recently, Nohno et al. (1995)described Received May 6, 1996; accepted June 7, 1996. Address reprint requestdcorrespondenceto Cheng-Ming Chuong, Department of Pathology, HMR 204,2011 Zonal Avenue, University of Southern California, Los Angeles, CA 90033, 158 TING-BERRETH AND CHUONG the presence of Shh in developing feathers. However, they described only a few stages of feather development. In this report, we examined the expression of Shh in all the major stages of early feather and filament development. Our detailed studies of the temporal and spatial distribution of Shh provided insight into the functions of Shh in skin appendage development and the formation of the stripe pattern of Shh in the feather filament. Furthermore, the functional role of Shh was explored using retrovirus-rnediated ectopic expression in ovo. The data suggest that Shh induces the formation of extra-large feather buds with aberrant or lost anterior-posterior polarity. This is preceded by the accumulation of many mesenchymal cells at the site of Shh overexpression. The regulation of Shh gene expression was also explored using skin explant cultures with epithelial-mesenchymal recombination. The results indicate that organized mesenchyme is required to maintain and induce Shh expression in the epithelium. Shh expression was also examined in avian scaleless mutants and showed abnormal expression that is consistent with the results from the retroviral experiments. These results show that Shh plays key roles in different stages of skin appendage morphogenesis. MATERIALS AND METHODS Chick Embryos Chick embryos used were from SPAFAS (Norwich, CT). These embryos are pathogen free and susceptible to infection by the A subgroup of RCAS virus. The embryos were staged according to Hamburger and Hamilton (1951). Scaleless chicken embryos were from Department of Poultry Genetics, University of Connecticut a t Storrs. In Situ Hybridization Nonradioactive in situ hybridization was performed according to Sasaki and Hogan (1993). Plasmid PHH-2 containing the full-length coding sequence of chick Shh (kindly provided by Drs. Randy Johnson and Cliff Tabin) was linearized with HindIII and transcribed with T3 RNA polymerase to synthesize the antisense riboprobe. After hybridization, the tissues were reacted overnight with preabsorbed antidigoxigenin Fab-conjugated alkaline phosphatase (Boehringer Mannheim). Color reactions were performed by incubating with NBT/BCIP substrates (Promega) in alkaline buffer. In situ hybridization on frozen sections was performed similarly. The bleaching step was omitted, and hybridization was carried out a t 48°C. Virus Production and Retroviral GeneTransduction RCASBP(A) plasmid containing the full-length coding region of chick Shh in the sense orientation is a generous gift from Drs. Robert Riddle and Cliff Tabin. The RCASBP(A)-Shh plasmid was transfected into stage 34 SPAFAS (line 15b) chicken embryo fibroblasts using lipofectamine (Gibco/BRL).The transfected cells were cultured in F-10 media containing 10% fetal bovine serum, 5% chicken serum, l x vitamin for MEM (Irvine Scientific), 20 pm folic acid, 0.5% dimethylsulfoxide (DMSO), and 1mM L-glutamine. Media containing the viruses were harvested when the cells were approximately 95% confluent and were stored at -80°C when not used immediately. Dorsal body ectoderm of stage 20 embryos was injected, in ovo, with microcapillary needles filled with 3-6 ~1 of virus-containing media. Embryos were harvested at stages 35-37 and analyzed for transgene expression and changes in skin morphology. Epithelial-MesenchymalRecombination Stage 34 dorsal skins were dissected in Hank's balanced salt solution (HBSS) and incubated in 2x calcium- and magnesium-free saline on ice for 5 min. The epithelium was then dissociated using watchmaker's tweezers and recombined with the epithelium. For examples shown in this report, the epithelium was rotated 90" to the mesenchyme. Recombined skins were transferred to culture inserts in six-well culture dishes (Falcon) containing DMEM with 10% fetal calf serum and incubated at 37°C in an atmosphere of 5% CO, and 95% air. Immunostaining Immunoalkaline phosphatase staining was performed on paraffin sections as described by Chuong et al. (1990). Rabbit polyclonal antibodies to Pan-PKC were from Upstate Biotechnology, Inc. Rabbit polyclonal antibodies to TGF-P2 were from Santa Cruz Biotechnology. Antibodies to NCAM and tenascin were described by Jiang and Chuong (1992). RESULTS Early Feather Morphogenesis: Shh Is Specifically Expressed in Epithelial Placodes The earliest morphological event detected during feather development is the formation of epithelial placodes from a flat and homogeneous ectoderm. With the formation of epithelial placodes, the mesenchymal cells underneath condense to form visible dermal condensations that will become the feather germs. It is from the interactions of the epithelium and mesenchyme in these feather germs that young feather buds develop (Fig. 7A; Lucas and Stettenheim, 1972). In the dorsal skin, there is a bilateral growth gradient where feathers at the midline develop first and feathers lateral to the midline develop subsequently. Shh is first detected as a tiny dot in the center of feathers at the early placode stage (Fig. l A , arrowhead). When the feather placode expands, Shh positive region also expands. In the short feather bud (the stage at which the long axis of the feather bud is shorter than the base; see Widelitz et al., 1996), Shh expression is stronger and is restricted to the tip (Fig. lA, arrow). As the feather buds develop, Shh expression pattern undergoes Shh IN FEATHER MORPHOGENESIS 159 Fig. 1. Distribution of Sonic hedgehog (Shh) transcripts during feather development. A-C and G H are wholemount in situ hybridization of embryonic chicken dorsal skin. D-F are nonradioactive in situ hybridization on sagittal frozen sections of dorsal skin. A: Top view of stage 35 dorsal skin. From left to right: early placode stage, placode stage, and short feather bud stage. The Shh transcript is first detected in a very small cluster of cells in the center of the early placode (arrowhead). This cluster enlarges in the short feather bud (arrow). B: Overview of stage 37 dorsal skin shows feather buds at various stages of development. Arrow indicates the midline of the skin. Inset: Outlined feather buds from the short bud stage to feather filament stage. C: Side view shows that Shh is expressed in the tips of the short feather bud epithelium and the mesenchyme is negative. D: Sagittal section of stage 33 dorsal skin. Shh is first detected in a few cells in the epithelial placode. Dashed lines outline the epithelium of the feather bud. E: Sagittal section of stage 36 dorsal skin shows that Shh is localized to the distal and posterior epithelium of the long feather bud. The expression of Shh has increased both in Intensity and in the number of positive cells. F: Cross section from the midshaft region of a feather filament shows that Shh is highly expressed in the marginal plate (mp) but not in the barb plate (bp). The basal lamina (bl) is located centripetal to the marginal plate epithelium (Chuong and Edelman, 1985a,b). p, Feather pulp. G: Higher magnification of a feather filament. Shh expression is in transition from the bud tip to the feather filament pattern. The arrow points to the transitional zone. Note that one feather filament stripe is made of two rows of epithelial cells. H: Shh expression in mature feather filaments from stage 40 chicken skin. Note the long stripes of Shh expression in the feather filament and the absence of Shh in the feather collar region (cl, equivalent to hair matrix). Scale bars = 100 prn in A-E, 50 pm in F-H. dramatic changes a t different stages of development (Fig. 1B).The side view showed that Shh is expressed only in the distal placode epithelium (Fig, 1 0 . We also performed in situ hybridization on 10 pm sections of developing feathers. This allowed us to view Shh expression at a higher resolution and to rule out artifacts due to inadequate probe penetration. The pat- tern is consistent with the wholemount in situ hybridization data. Shh is first detected at the center of the epithelial placode (Fig. 1D).The Shh signal increased as the feather developed. In the long feather bud (the stage a t which the long axis of the feather bud is longer than the base; see Widelitz et al., 19961,Shh expression is shifted to the distal end and posterior side of the 160 TING-BERRETH AND CHUONG feather bud (Fig. 1E). Shh was not detected in feather mesenchyme a t any of the stages studied, Late Feather Morphogenesis: Shh Is Expressed in the Marginal Plates Following the long feather bud stage, the developing feathers invaginate into the skin to form feather follicles. At this stage, feather growth is at the collar region of the follicle base and the morphogenesis of feather filaments begins. The feather epithelium starts to form alternating barb ridges. For each barb ridge, the epithelium produces a barb plate in the center and marginal plates bilaterally. The barb plate epithelium will later keratinize to become feather barbs, and the marginal plate epithelium will undergo programmed cell death to become the spaces between the barbs (summarized in Fig. 7A; Lucas and Stettenheim, 1972). Wholemount in situ hybridization of feather filaments showed that the pattern of Shh expression changed dramatically during the transition from the long bud to a filamentous feather. Prior to filament formation, Shh is localized in the distal tip epithelium. With the development of feather filaments, Shh is expressed in alternating “stripes” as epithelial compartmentalization begins (Fig. 1E). The “striped pattern suggests that Shh is either in the barb plate or in the marginal plate. A cross section through the feather filaments revealed that Shh is in the marginal plate (Fig. lF,G). This pattern of expression is maintained through stage 40 feathers, in which the filaments are longer and more mature. However, Shh has by now disappeared in the distal feather filament. Shh is negative in the collar epithelium (in which the new feather epithelial cells are added and compartmentalization has not started; Fig. 1H). We attempted to analyze the relationship between marginal plate formation and Shh expression. Because the feather is more mature toward the distal end, semiserial cross sections of feather filaments were cut from the base to the tip so that the different stages of barb ridge formation could be viewed. Barb ridges are initiated from the anterior feather by a series of epithelial invaginations a t the site of future marginal plates. Minute amounts of Shh are first detected in the smooth epithelium just prior to invagination (Fig. 2A, arrowhead). This is followed by the gradual increase of Shh in the invaginating cells that initiate the formation of a marginal plate (Fig. 2A, arrow). At this stage, the posterior side of the epithelium remains cylindrical and completely negative of Shh (Fig. 2A, asterisk). The epithelial invaginations propagate bilaterally around the circumference of the feather and Shh is turned on (Fig. 2B,C). As the bilateral waves of marginal plate/ barb ridge formation spread towards the posterior feather (Fig. 2D, solid curved arrows), a ring of Shh expression is seen in each marginal plate (Fig. 2E). The earliest Shh-expressing cells are in the valley of each marginal plate invagination, specifically in the basal layer of the stratified epithelium. As the invagination of the marginal plates deepens, Shh expression propagates bilaterally in the marginal plates toward the tip of the ridge (Fig. 2F, curved open arrows). Eventually the entire marginal plate epithelium becomes strongly positive for Shh expression (Fig. 2F). As the adjacent barb plate epithelium begins to keratinize, Shh-positive marginal plate epithelial cells will undergo cell death and disintegrate. This process leaves the space between the keratinized cells that creates the branched barbs (Lucas and Stettenheim, 1972). Our data suggest that Shh is associated with the epithelial compartmentalization process in feather filament morphogenesis. Ectopic Expression of Shh-Induced Mesenchymal Condensation To test the functional role of Shh in feather development, we adopted the retroviral gene delivery technique developed in chicken embryos (Morgan et al., 1992; Riddle et al., 1993; Fekete and Cepko, 1993). We first injected RCAS alkaline-phosphatase (Alk-P) virus-containing media into the dorsal ectoderm of stage 20 (embryonic day 3) chicken embryos. Embryos were sacrificed after 5 days and assayed for Alk-P activity (Fig. 3A). The patch of blue-purple staining represents the extent of transgene expression. Higher magnification revealed no abnormal feather development. This served as a control to show that the RCAS virus with an innocuous gene does not cause abnormal feather development. Ectopic overexpression of Shh was induced by injecting 3-6 ~1 of RCASBP(A1-Shh viral media into the dorsal ectoderm of stage 20 embryos. The embryos were incubated for an additional 7 days, harvested, and the skins examined. The development of enlarged feather buds was observed in the affected patch in 11 of 36 embryos that survived the injection. Many of the enlarged feather buds appeared to be symmetrical and Fig. 2 Expression sequence of Shh during the formation of marginal plates in the feather filament. In situ hybridization of serial cross sections from the base (A) towards the tip (F) of one feather filament. A: Shh expression is first detected (arrowhead) just prior to the invaginationof the epithelial sheet to form barb ridges and increases slightly as the epithelial invagination is more evident (arrow). The posterior side of the feather that has yet to undergo barb ridge formation is marked by the asterisk. B: Shh expression is initially seen when the filament epithelium is still smooth (arrowhead), and it is slightly stronger in the shallow indentations (arrows). C: The indentationsdeepen, and the developing marginal plates are more evident (arrows). D: Shh is enhanced in the developing marginal plates (straight arrows) between the forming barb ridges (bp). At this time, the posteriorside is still without barb ridges and is negative of Shh. In a similar fashion, a wave of barb ridge formation will spread bilaterally (curved arrows) toward the posterior side of the feather. E: Marginal plates (mp) have formed along the entire circumference of the feather, and Shh is expressed in all areas destined to become marginal plates. The marginal plates at the anterior side are more mature and are stronger in Shh expression. F: Within a more mature marginal plate, Shh expression spreads from the cells in the “valley” of the marginal plates towards the top of barb ridge (arrows). Later, these Shh- positive cells become the space between feather barbs. a-tp, Anterior-posterior axis of feather filament. Scale bars = 30 pm. Shh IN FEATHER MORPHOGENESIS Fig. 2. 161 162 TING-BERRETH AND CHUONG stunted in growth (Fig. 31, and some exhibited reversed anterior-posterior axis (Fig. 4). We verified the ectopic expression of Shh in the affected skin by wholemount and section in situ hybridization. Increased expression of Shh in and around the enlarged buds confirms the ectopic overexpression of Shh (Figs. 3B, 4A). It should be noted that, even though Shh is normally expressed only in the feather epithelium, our injection procedure ectopically expressed Shh in both the epithelium and the mesenchyme. The area occupied by this enlarged skin appendage in Figure 3B represents the space that would normally accommodate approximately eight feather buds. The mesenchymal cells in that area may all be incorporated into the enlarged bud. The size of the enlarged buds varied in different experiments, reflecting the degree of transgene expression. Some of the skins harvested (approximately 10%)showed no visible phenotypic changes even though they had some degree of Shh overexpression as judged from small patches of intense signals when hybridized with Shh antisense probe (not shown). Controls included embryos that were injected with RCASBP(A) virus only or with RCANBP(A)-Shhvirus (vector unable to produce transcripts due to the absence of a splice acceptor). These did not exhibit abnormal feather phenotype (not shown). Histological analysis of enlarged feather buds was performed. Sagittal sections stained with hematoxylin and eosin showed that there is a large accumulation of mesenchymal cells in the enlarged feather buds. This accumulation extended deeply into the dermis and the underlying muscle layer (compare Fig. 3C, a normal feather bud, and Fig. 3D).In situ hybridization with Shh antisense probe showed that there is overexpression of Shh in the accumulated mesenchymal cells and in the deep dermis (Fig. 4A). Downstream to Shh: Loss of Polarized NCAM and Tenascin Localizations, Increased Distribution of TGF-P2, and Suppressed PKC Expression To search for potential downstream molecular changes caused by Shh overexpression, sections were stained with various antibodies. Previously, we reported that tenascin and NCAM are expressed in developing feather buds (Jiang and Chuong, 1992). Tenascin is initially expressed in the anterior epithelium of early feather germs. In the short bud stage, both tenascin and NCAM are localized in the anterior feather mesenchyme. In the symmetrically enlarged bud, the asymmetric localization of these molecules is lost. Both tenascin and NCAM became symmetrically distributed compared to a normal bud (Fig. 3E,F, open arrow). The expression patterns of NCAM and tenascin remained intact in the reverse-oriented feather buds, but they were polarized to the new feather axis (not shown). Recently, we showed that TGF-p2 message is strongly expressed in the placode epithelium, and TGF-P2 protein can exogenous substitute for the epithelium by inducing dermal condensations from a mesenchyme stripped of its epithelium (Ting-Berreth and Chuong, submitted for publication). In the Shh- induced enlarged bud, there is a wider distribution of TGF-P2 (Fig. 4B, arrowheads) than in normal feather bud (Fig. 4C). However, the intensity of expression (per cell) between the normal and abnormal bud did not increase. We have also shown that enhanced PKC activity favored the expansion of the interbud domain (Noveen et al., 1995b). When TGF-P2 was delivered locally by an Mi-Gel Blue bead, PKC expression was suppressed in the induced mesenchymal condensation (Ting-Berreth and Chuong, submitted for publication). These results are consistent with our observation that PKC expression is suppressed in the region of Shh overexpression (compare Fig. 4D and El. Shh Is Induced and Maintained by the Feather Mesenchyme We explored the possible regulatory mechanisms of Shh expression in the skin. Feather formation requires continuous and reciprocal interactions between the epithelium and the underlying mesenchyme. When the epithelium and mesenchyme are separated, ongoing cellular and molecular organization is lost. However, if the epithelium and mesenchyme are recombined after separation, feather growth will resume (reviewed in Sengel, 1978). The position of these reformed feather germs depends on the mesenchyme (Sengel, 19761, whereas the orientation is determined by the epithelium (Novel, 1973). The regeneration capacity of the skin provides an excellent opportunity to determine whether the epithelial expression of Shh is autonomous or mesenchyme dependent. The epithelium and mesenchyme of stage 34 dorsal skins were dissociated and immediately recombined with a 90" rotation of the epithelium with respect to the mesenchyme. As a result, these two components of the skin were out of phase from their original positions and any existing interactions that may have been initiated prior to dissociation were reorganized. In addition, a new feather A-P orientation was established, allowing us to examine whether Shh induction is associated with feather orientation. Figure 5A shows that stage 34 skin expresses Shh in the posterior bud. Immediately after the recombination, Shh in the epithelium was still visible. However, the epithelium was out of phase and did not overlap with the condensed mesenchyme (Fig. 5B). Within 3 hr, Shh transcripts in the epithelium had completely disappeared (Fig. 5C), suggesting that close contact with organized mesenchyme is essential for the maintenance of Shh expression. Six hours after the recombination, new Shh expression was induced in the center of newly formed epithelial placodes (Fig. 5D). Subsequent expression patterns of specimens harvested a t 8, 10, 12, and 18 hr recapitu- Shh IN FEATHER MORPHOGENESIS 163 Fig. 3 Ectopic Shh expression, mediated by retroviral vectors, induces enlarged buds, accumulation of mesenchymal cells, and abnormal expression of adhesion molecules. Stage 20 SPAFAS chicken embryos injected with 3-6 pl viral media at the dorsal body wall. A: Chicken embryo injected with RCAS(A)-alkalinephosphatase viral media (RCASAlkP) was harvested at stage 34. The transgene expression is shown by a patch of skin expressing ectopic alkaline phosphatase (purple color, indicated by the arrow). The head is toward the right margin. 8: Wholemount in situ hybridization of a stage 37 embryo skin that was injected with RCASBP(A)-Shh. An enlarged bud is the result of Shh overexpression. Shh overexpression was verified by wholemount in situ hybridiration. The bud is abnormally large, round, and stunted. It is surrounded by normal buds that have already entered the feather filament stage. C: H and E staining of a normal feather bud (normal; H and E). D: H and E staining of the Shh-induced bud (RV; H and E). Note the accumulation of large numbers of mesenchymal cells in the Shh-induced bud. The accumulation extended deeply into the dermis (small open arrow). E: Sagittal section through an enlarged and symmetrical Shh-induced bud showed that NCAM (RV-NCAM), normally in the anterior of feather buds (solid arrows) has lost its polarized distribution. Adjacent normal feather is indicated by large open arr0w.F: Sagittal section of the enlarged and symmetrical Shh-induced bud showed that tenascin, also normally expressed in the anterior portion of normal buds, is induced in the mesenchyme of both anterior and posterior portions of the abnormal bud (RV-Tn). In the abnormal bud, tenascin is also lost in the epithelium. Anterior-posterior axis is designated a-p. Scale bars = 100 prn. lated the expression sequence in vivo (not shown). Thirty hours after recombination, the reformed feather buds were elongated and asymmetrical. A new anterior-posterior axis was established by the epithelium and Shh was located in the posterior-distal ends of the new bud orientation (Fig. 5E). The orientation of Shh and newly formed feather buds was still determined by the epithelium. These results showed that epithelial expression of. Shh is mesenchyme dependent. The entire piece of the epithelium is competent to respond to the mesenchymal signals to form placodes and express Shh. Shh Is Abnormally Expressed in Scaleless Mutants Scaleless is an avian autosomal recessive mutant that fails to develop scales on its feet and is mostly bald except for a few patches of feathers along the back, wing, gluteal, and head areas (Abbott and Asmundson, 1957). The genetic defects of scaleless have been localized in the epithelium using epithelial-mesenchymal recombination experiments (Sengel and Abbott, 1963; McAleese and Sawyer, 1981). At stage 34, Shh is mostly absent from the dorsal skin of the scaleless embryo (Fig. 6A). Some embryos at this stage developed 164 TING-BERRETH AND CHUONG Fig. 4 Ectopic Shh expression induces enlarged buds with abnormal orientation, wider TGF-gi2 distribution, and suppressed PKC expression. Sagittal sections through an embryo with enlarged feather growth due to Shh overexpression. A, overview. B,D, region of abnormal buds. C,E, region of normal buds. A: Shh induces a feather bud with reversed anterior-posterior polarity. The original anteroposterior direction is indicated by the straight open arrow. In situ hybridization with Sbh antisense probes was performed to verify the degree of overexpression. In the induced bud (open curved arrow), Shh is expressed in both the epithelium and mesenchyme, and extends into the muscle layer. In the normal buds (solid curved arrow), Shh is present only in the distal feather bud epithelium. B.C: In the enlarged bud, TGF-p2 protein expression is similar to the normal buds, but th6 area of expression is wider, The border of high TGF-p2 expression is flanked by large arrowheads in the abnormal buds and small arrowheads in the normal feather buds. In normal feather buds, TGF-P2 is limited to the distal-posterior feather epithelium. D,E: PKC is suppressed in the mesenchymal region of the enlarged bud (large star) as well as in the normal buds (small star). PKC is normally expressed in the interbud region of chick skin. Note that the region of PKC suppression correlates well with the region of Sbb overexpression (cf. A and C). Scale bars = 100 km. ridges of fused, feather-like structures along the lateral skin. It is possible that the scaleless gene is upstream of part of the dorsal skin. Interestingly, the mesenchyme Shh and plays a role in regulating Shh. of these fused structures strongly expressed Shh (Fig. DISCUSSION 6A, arrow), but the epithelium did not. At stage 38, bald regions of the scaleless skin remain negative of The elaborate morphogenetic processes in vertebrate Shh expression. A few normal feathers that had devel- epithelial-mesenchymal interactions create various oroped in the gluteal regions appeared normal and ex- gans with different functions in different species. Howpressed Shh in the normal striped pattern (Fig. 6B). ever, the fundamental molecular mechanism underlyThese results show that Shh is affected in the scaleless ing these processes is probably the same. To understand Shh IN FEATHER MORPHOGENESIS 165 Fig. 5 Feather mesenchyme induces Shh expression in the epithelium. Epithelial-mesenchymal recombinants with 90" rotation were prepared from stage 34 skin. Orientations of epithelium (E)and mesenchyme (M) are indicated by arrows in the lower left corner. The recombinants were cultured for the time indicated in the upper right corner, then fixed and processed for wholemount in situ hybridization with Shh. A: Stage 34 skin without recombination at time 0 showed that the hexagonal expression pattern of Shh is in the posterior-distal end of feather germs (arrow). The margin of some feather buds is outlined. 6: Immediately after E-M recombination, the epithelium and the mesenchyme are out of phase. The old Shh transcripts are still present in the epithelium. The mesenchyme not covered by the recombined epithelium (asterisk) does not express Shh. C: Three hours after the recombination, Shh expression has completely disappeared. D: Six hours after the recombination, Shh is reinduced in the epithelium at the center of each newly organized feather germ. Polarity is not clear at this stage. Note that Shh expression is more advanced in the midline region (arrow) of the epithelium. E: Thirty hours after recombination, normal Shh expression pattern is reestablished in the posterior-distalend of regenerated feather buds (arrow, pointing to posterior end of bud). The position of the buds is determined by the mesenchyme, but the orientation of the buds is determined by the epithelium. Scale bars = 250 pm. 166 TING-BEFtFtETHAND CHUONG Fig. 6 Shh in the skin of the scaleless mutant. Wholemount in situ hybridization of scaleless embryonic skins with antisense Shh probe. A: Stage 34, lateral and rniddorsal skin region. The middorsal region is towards the bottom of the figure. Shh is negative in the portion of the skin that does not form skin appendages (star). Large, fused skin appendages are formed in the lateral dorsal regions of some mutants (arrow). These ridges contain abnormally fused dermal condensations that express Shh in their mesenchyme. 8: Stage 37. gluteal skin region. The bare region remains Shh negative. A few morphologically normal feathers have formed. These exhibit the normal Shh alternating stripe pattern in the feather filament. This suggests that, in scaleless, Shh can be expressed normally in areas with normal feather buds, but it is deregulated in most regions. Scale bar = 150 pm. this mechanism, we have been studying various morphogenetically related molecules using the feather model (Chuong, 1993; Widelitz et al., 1996). For this report, we studied the roles of Shh during feather development. The process of feather morphogenesis and various Shh expression patterns are summarized in Figure 7A. Contrary to the mesenchymal expression pattern reported for the limb (Riddle et al., 19931, Shh is expressed exclusively in the epithelial structures of feather buds in all stages that we studied. This may be reasonable given the fact that the limb consists mainly of mesenchymal cells, whereas the feather consists mainly of epithelial cells. If Shh plays a major role in skin morphogenesis, it should be present in the epithelium, where major morphogenetic events occur. This is clearly evident during late feather development, when Shh is found in the marginal plates of feather filament epithelium with alternating stripes. Upon closer exam- ination, we found that Shh expression precedes the initiation of marginal plates, suggesting that Shh may be involved in specifying epithelial compartments that are destined to die. Shh Is an Early Event in the Molecular Cascade of Skin Appendage Induction We observed that, in the process of early feather development, Shh message is first detected during epithelial placode formation (summarized in Figs. 1, 7A). In reorganized epithelial-mesenchymal recombinations, old Shh transcripts in the epithelium disappeared while new Shh was induced by the feather mesenchyme. These findings suggest that 1)mesenchymal factor(s) is required to maintain the expression of Shh, 2 ) mesenchymal factors in the organized feather mesenchyme can initiate the expression of Shh, and 3 ) the entire epithelium a t this stage, whether it is placodal Shh IN FEATHER MORPHOGENESIS A B - Ectoderm Mcsenchvnie 1 Induction 1 1 1 Dermal Condensation Expression of Sonic Hedgehog in placode epithelia I___I__+ Placode 1 (E) 1 Axis and 4 Shon Fealher Bud 1 Growth 167 1 TGF Beta (E) and or other grow:h factors 1 AA&. +I 1 1 Enhanced NCAM Pr TGF Beta receptors (M) Long Feather Bud IInvagination I I feather hlament dermal papilla expression (M) and/or other adhesion molecules collar Feather Follicle 1 -Formation of dermal condensations (M) . _ I I basement membrane r-zLJ Feather Filament Differentiation and 1 I ,--- rachis Matun: Feather or interplacodal, is competent to respond to the mesenchymal factor(s) to express Shh and form new placodes. These data are consistent with earlier findings that the pattern is determined by the mesenchyme (reviewed in Sengel, 1978). Our expression data of Shh during early feather development showed that Shh appears after the formation of epithelial placodes. This implies that Shh is probably not the initiator of the molecular cascade of skin appendage induction but that it is one of the early events in the cascade. The early requirement of Shh during skin induction that came from the avian autosoma1 recessive mutant scaleless (Abbott and Asmund- Fig 7. Summary of Sbh expression pattern and a working model. A: Shh expression is schematically shown for different stages of feather development. Following induction, the formation of epithelial placodes creates the first recognizable morphological structure. Cells in the placode domain will proliferate, migrate, invaginate, and differentiate in ways totally different from those used by cells remaining in the interappendage domain. See text and Chuong (1993) for further explanation of the process of feather morphogenesis. The filament depicted is a cross section at the level shown for the feather follicle stage. The location of Shh transcripts is shown in black. Note that the marginal plate is an "altered" basal layer epithelium. Shh is expressed periodically in these basal epithelial cells. B: Hypothetical model of the epithelial-mesenchymal interaction during the formation of skin appendages that includes Shh and molecules recently studied in our laboratory (for NCAM and tenascin, see Jiang and Chuong, 1992; for TGF-pP, see Ting-Berreth and Chuong, submitted for publication). The epithelial or mesenchymal location of the molecular and cellular events is indicated by (E) or (M), respectively. Each arrow may include more than one step. See text for discussion. son, 1957) supports this hypothesis. The origin of the scaleless defect is in the epithelium; the mesenchyme is initially normal. However, if the scaleless mesenchyme does not receive the proper signals from the epithelium, it will become abnormal as well (Sengel and Abbott, 1963; McAleese and Sawyer, 1981). Because Shh is present in the placode epithelium and is involved in feather induction, we expect that Shh will be absent in the bare region. This is indeed the case (Fig. 6). In these mutants, there are fused skin appendage ridges in the dorsal regions that morphologically resemble the enlarged bud induced by ectopic Shh expression. We found abnormal Shh expression in the mesenchymal 168 TINGBERRETH AND CHUONG region of these ridges. Shh may act to induce mesenchymal cells to form the abnormal ridges. These data suggest that the scaleless gene probably acts upstream of Shh, and the defect may cause deregulation of Shh distribution and subsequent skin appendage defects. Shh May Induce Mesenchymal Condensations Through Wider Distribution Expression of TGF-62 and Suppression of PKC A major event during the placode and short bud stages of feather development is the condensation of mesenchymal cells toward the placode regions. 3H-thymidine and bromodeoxyuridine (BrdU) labeling experiments showed that cell proliferation has ceased during these stages (Wessels, 1965; Noveen et al., 1995a). Therefore, dermal condensations are mostly due to cell migration. When Shh was ectopically expressed in both the epithelium and the mesenchyme by retrovirus-mediated gene transduction, there was an enormous accumulation of mesenchymal cells toward this region that extended deeply into the muscle layer. This observation suggests that Shh transcript in the placode epithelium may be used to make Shh peptides that act on the surrounding epithelium and mesenchyme to induce the formation of dermal condensations (Fig. 3B). Morphologically, the Shh-induced feather buds appeared to be the result of the fusion of several feather fields. Therefore, Shh may have a strong ability to induce mesenchymal condensation. Recent studies have shown that Shh protein is proteolytically cleaved to produce amino- and carboxyl-terminal proteins that may play distinct biological roles (Lee et al., 1994; Porter et al., 1995; Johnson and Tabin, 1995). The aminoterminal end of Shh protein appears to be responsible for floor plate, motor neuron, and sclerotome induction (Roelink et al., 1995; Fan et al., 1995). In future studies, we will try to detect Shh peptides to see whether they are secreted into the mesenchyme from the placodes. What are the possible molecular events downstream of Shh? Recently we found that the activation of PKC expands the interbud domains (Noveen et al., 199513). We then searched for the extracellular signals that can induce dermal condensations and act on PKC. TGF-PZ protein-coated beads can substitute for epithelial placodes to induce dermal condensations. Around the bead, there is an inhibition of PKC expression (TingBerreth and Chuong, submitted for publication), suggesting that TGF-P2 may be involved in the induction of dermal condensations, at least partially, through the inhibition of PKC. In regions of Shh overexpression, we observed a wide distribution of cells that exhibit TGF-P2 and inhibit PKC expression (Fig. 4I3,D). Therefore, Shh may act early in the signaling cascade of mesenchymal condensations, upstream of TGF-P2 (Fig. 7B). With the current experimental procedure, both the epithelium and the mesenchyme are made to produce ectopic Shh. However, the ectopic expression data do demonstrate that Shh has a functional role in feather morphogenesis. In the future, higher resolution studies will be performed by directing overexpression only to the mesenchyme or epithelium. Shk and the Anterior-PosteriorPolarity of Feather Buds As the developing feathers grew from a symmetric short bud into an asymmetric long bud, Shh expression shifted from a symmetric, centrally localized pattern to a polarized, distal-posterior pattern. Therefore, Shh may also be involved in determining the anterior-postenor axis of feather buds. Among the Shh-induced abnormal buds, some are round and have lost asymmetry, whereas others exhibit reversed anterior-posterior orientation. In the round buds, the anterior localization of NCAM and tenascin in feather mesenchyme became symmetrical. The anterior expression of tenascin in the anterior bud epithelium (Jiang and Chuong, 1992; Fig. 3)was also lost in the symmetric buds. In the reverseoriented buds, NCAM and tenascin are distributed in a polarized manner, in accordance with the new anteriorposterior orientation. In the feather, there is a wider variation in the phenotypes observed than in the reported limb bud studies (Riddle et al., 1993). This is partially due to the fact that the feather field is much smaller than the limb bud. As a result, the region of ectopic expression influenced several feather fields rather than a single feather field. Future refinements in retroviral gene delivery may allow more specific and detailed studies of Shh function. The results presented here show that ectopic expression of Shh can cause abnormal anteroposterior orientation during feather morphogenesis. Shh Expression and Epithelial Compartmentalization A major cellular event in feather filament morphogenesis is the invagination of the cylindrical filament epithelium. This in essence is the compartmentalization of the feather filament epithelium into the barb plate and marginal plate. Among the numerous molecules we examined during feather morphogenesis (Widelitz et al., 1996), NCAM is the only molecule that is specifically expressed in the marginal plate (Chuong and Edelman, 1985b1, in a pattern similar to that of Shh. However, the onset of NCAM expression in the marginal plate lagged behind that of Shh by three marginal plates. Therefore, the order of appearance is Shh + marginal plate initiation + NCAM. What are the roles of Shh and NCAM in the marginal plate? Previous studies have shown that NCAM in the marginal plate epithelium will eventually “zip in” and form a collective cell group (Chuong and Edelman, 1985b). We hypothesized that the adhesive property of NCAM is used to define a group of cells to undergo a different developmental fate. The new fate can be cartilage, kidney epithelium, keratinization, or, as in this case, programmed cell death (Chuong, 1990). Shh may act upstream to regulate adhesion molecules to 169 Shh IN FEATHER MORPHOGENESIS define epithelial compartments. In this regard, it will be most exciting to study whether Shh and related molecules can be used to regulate the spacing and number of feather barbs and, therefore, the shape of the mature feathers. Nohno et al. (19951, using an independently derived Shh antisense probe, describe their observation of Shh expression in chicken limb feathers. The results from the representative stages of feather development that they studied are basically consistent with our data. In describing their data, the authors referred to the “zone of Shh expression” in the wing feathers. This, in fact, is the valley region of marginal plates, and its significance is discussed above. A WORKING MODEL FOR EPITHELIAL-MESENCHYMAL INTERACTION DURING SKIN APPENDAGE MORPHOGENESIS In this paper, we present a hypothetical model of how Shh may be involved in dermal condensations based on the data reported here and on our previous work (Fig. 7B). There are other molecules and signaling pathways involved, which remain to be elucidated. Our results indicate that initially unidentified molecule(s) from specific locations of the feather mesenchyme induce the epithelium to form epithelial placodes and that Shh message is subsequently turned on in these placodes. Secreted Shh proteins in the placode can induce dermal condensations through TGF-P2 (Ting-Berreth and Chuong, submitted for publication). The condensed mesenchyme is enriched in NCAM, tenascin, and other molecules and is distinctly different from the intercondensation mesenchyme. The condensed mesenchyme can signal back to the epithelium to maintain and induce Shh expression. In conclusion, during early feather development, Shh is a major component of feather placode epithelium and a major mediator of dermal condensations. Here, we have analyzed part of this molecular mechanism. In late feather development, Shh is a major component of the marginal plate epithelium and is associated with cell death. The distinct pattern exhibited by the feather provides a powerful model for dissecting the multifaceted roles of Shh in tissue interactions. ACKNOWLEDGMENTS We are grateful to Drs. Cliff Tabin, Robert Riddle, and Randy Johnson for providing us with RCASBP(A)Shh, RCANBP-Shh, and PHH-2 plasmids, and Dr. Bruce Morgan for RCASBP-Alk-P virus, We thank Dr. Ting-Xin Jiang for contributing Figure 3A and Dr. Randall B. Widelitz, Dr. Kevin Moses, and Ms. Florence Miyagawa for critical reading of the manuscript. This work was supported by NIH grant AR 42177, NSF grant IBN 9317397, a USC Zumberg Faculty Research Award, and a grant from the Council for Tobacco Research. REFERENCES Abbott, U.K., and Asmundson, V.S.(1957) Scaleless, an inherited ectodermal defect in the domestic fowl. J . Hered. 48:63-70. Bitgood, M.J., and McMahon, A.P. (1995) Hedgehog and BMP genes are coexpressed at many diverse sites of cell-cell interaction in the mouse embryo. Dev. Biol. 172:126-138. Chang, D.T., Lopez, A., von Kessler, D.P., Chiang, C., Simandl, B.K., Zhao, R., Seldin, M.F., Fallon, J.F., and Beachy, A. (1994) Products, genetic linkage and limb patterning activity of a murine hedgehog gene. Development 120:3339-3363. Chuong, C.-M., Oliver,G.,Ting, S.A., Jegalian,B.G., Chen,H.-M., and De Robertis, E.M. (1990) Gradients ofhomeoproteins in developing feather buds. Development 110:1021-1030. Chuong, C.-M. (1993) The making of a feather: Homeoproteins and retinoids and adhesion molecules. Bioessays 15:513-521. Chuong, C.-M., and Edelman, G.M. (1985a) Expression of cell-adhesion molecules in embryonic induction. I. Morphogenesisof nestling feathers. J. Cell Biol. 101:1009-1026. Chuong, C.-M., and Edelman, G.M.(1985b) Expression of cell adhesion molecules in embryonic induction. 11. Morphogenesis of adult feathers. J. Cell Biol. 101:1027-1043. Echelard, Y., Epstein, D.J., StJacques, B., Shen, L., Mohler, J., McMahon, J.A., and McMahon, A.P. (1993) Sonic hedgehog, a member of a family of putative signaling molecules, is implicated in the regulation of CNS polarity. Cell 751417-1430. Fan, C.-M., and Tessier-Lavigne,M. (1994) Patterning of mammalian somites by surface ectoderm and notochord: Evidence for sclerotome induction by a hedgehog homolog. Cell 79:1175-1186. Fan, C.-M., Porter, J.A., Chiang, C., Chang, D.T., Beachy, P.A., and Tessier-Lavigne, M. (1995) Long-range sclerotome induction by Sonic hedgehog: Direct role of the amino-terminal cleavage produd and modulation by the cyclic AMP signaling pathway. Cell 81:457465. Fekete, D., and Cepko, C. (1993) Replication-competent retroviral vectors encoding alkaline phosphatase reveal spatial restriction of viral gene expressiodtransduction in the chick embryo. Mol. Cell Biol. 13:2604-2613. Hamburger, V., and Hamilton, H. (1951)A series of normal stages in the development of the chick embryo. J. Morphol. 88:49-92. Jiang, T.-X., and Chuong, C.-M. (1992) Mechanism of feather morphogenesis: I. Analyses with antibodies to adhesion molecules tenascin, N-CAM and integrin. Dev. Biol. 15082-98. Johnson, R.L., and Tabin, C. (1995) The long and short of hedgehog signaling. Cell 81:313-316. Johnson, R.L., Laufer, E., Riddle, R.D., and Tabin, C. (1994) Ectopic expression of Sonic hedgehog alters dorsal-ventral patterning of somites. Cell 79:1165-1173. Krauss, S., Concordet, J.P., and Ingham, P.W. (1993) A functionally conserved homolog of the Drosophila segment polarity gene hh is expressed in tissues with polarizing activity in zebrafish embryos. Cell 75:1431-1444. Lee, J.J., Ekker, S.C., von Kessler, D.P., Porter, J.A., Sun, B.I.,and Beachy, P.A. (1994)Autoproteolysisin hedgehog protein biogenesis. Science 266:1528-1537. Lucas, A.M., and Stettenheim, P.R. (1972) Avian anatomy. Integument. In: “Agriculture Handbook 362.” Washington, DC: Agricultural Research Services, U.S.Department of Agriculture. McAleese, S.R., and Sawyer, R.H. (1981) Correcting the phenotype of the epidermis from chick embryos homozygous for the gene scaleless (sc/sc).Science 214:1033-1034. Morgan, B.A., Izpisua-Belmonte, J.-C., Duboule, D., and Tabin, C.J. (1992) Targeted misexpression of Hox-4.6 in the avian limb bud causes apparent homeotic transformation. Nature 358236-239. Nohno, T., Kawakami, Y., Ohuchi, H., Fujiwara, A,, Yoshioka, H., and Noji, S. (1995)Involvement of the Sonic hedgehog gene in chick feather development. Biochem. Biophys. Res. Commun. 206:33-39. Noveen, A., Jiang, T.-X., Ting-Berreth, S.A., and Chuong, C.-M. (1995a) Homeobox genes Msx-1 and Msx-2 are associated with induction and growth of skin appendages. J. Invest. Dermatol. 104: 711-719. Noveen, A., Jiang, T.-X., and Chuong, C.-M. (1995b)Protein kinase A 170 TING-BERFETH AND CHUONG and protein kinase C modulators have reciprocal effects on mesenchymal condensation during skin appendage morphogenesis. Dev. Biol. 1713377-693. Novel, G. (1973)Feather pattern stability and reorganization in cultured skin. J. Embryol. Exp. Morphol. 30605-633. Placzek, M., Tessier-Lavigne, M., Yamamda, T., Jessell, T.M., and Dodd, J. (1990)Mesodermalcontrol of the neural cell identity: Floor plate induction by the notochord. Science 250985-988. Porter, J.A., Ekker, S.C., Young, K.E., von Kessler, D.P., Lee, J.J., Moses, K., and Beachy, P.A. (1995)The product of hedgehog autoproteolytic cleavage active in local and long-range signalling. Nature 374:363-366. Riddle, R.D., Johnson, R.L., Laufer, E., and Tabin, C. (1993)Sonic hedgehog mediates the polarizing activity of the ZPA. Cell 7514011416. Roelink, H.,Augsburger, A., Heemskerk, J., Korzh, V., Norlin, S., Ruiz i Altaba, A., Tanabe, Y., Placzek, M., Edlund, T., Jessell, T.M., and Dodd, J. (1994)Floor plate and motor neuron induction by vhh-1, a vertebrate homolog of hedgehog expressed by the notochord. Cell 76761-775. Roelink, H., Porter, J.A., Chiang, C., Tanabe, Y., Chang, D.T., Beachy, P.A., and Jessell, T.M. (1995)Floor plate and motor neuron induction by different concentrations of the amino-terminal cleavage product of Sonic hedgehog autoproteolysis. Cell 81:445-455. Sasaki, H., and Hogan, B.L. (1993)Differential expression of multiple fork head related genes during gastrulation and axial pattern formation in the mouse embryo, Development 118:47-59. Sawyer, R.H., and Fallon, J.F. (eds.; 1983)“Epithelial-Mesenchymal Interactions in Development.” New York: Praeger Publishing. Sengel, P. (1976)“Morphogenesisof Skin.” Abercrombie, M., Newth, D.R., and Torrey, J.G. (eds). Cambridge: Cambridge University Press. Sengel, P. (1978)Feather pattern development. Ciba Found. Symp. 2961-70. Sengel, P., and Abbott, U.K. (1963)In vitro studies with the scaleless mutant: Interaction during feather and scale differentiation. J. Hered. 54:254-262. van Straaten, H.M.W., Hekking, J.M.W., Wiertz-Hoessells, E.L., Thors, F., and Drukker, J. (1988)Effect of the notochord on the differentiation of a floor plate area in the neural tube of the chick embryo. Anat. Embryol. 177:317-324. Wessels, N.K. (1965) Morphology and proliferation during early feather development. Dev. Biol. 12131-153. Widelitz, R.B., Jiang, T.-X., Noveen, A., Ting-Berreth, S.A., Yin, E., and Chuong, C.-M. (1996)Molecular histology in skin appendage morphogeneeis. Microsc. Res. Tech. (in press).