Expression of Sonic hedgehog, Patched and Gli1 in developing taste papillae of the mouse

Authors: Joshua M. Hall¹*, Joan E. Hooper¹, and Thomas E. Finger¹

¹Department of Cellular and Structural Biology and
Medical Scientist Training Program,
University of Colorado Denver, Denver, CO 80262.

Indexing terms: developmental signaling, epithelial patterning, in situ hybridization, PGP 9.5

Correspondence to:

Joshua M. Hall
Department of Cellular and Structural Biology
4200 E. 9th Ave., Box B111
Denver, CO 80262
Email: halljosh@york.uchsc.edu

Lingual taste buds form within taste papillae, which are specialized structures that develop in a characteristic spatial and temporal pattern. In order to investigate the signaling events responsible for patterning and morphogenesis of taste papillae, we have examined the time course and distribution of expression of several related developmental signaling genes as well as the time course of innervation of taste papillae in E12-E18 mouse embryos. Lingual expression of the signaling molecule Sonic hedgehog (Shh), its receptor Patched (Ptc), and the Shh-activated transcription factor Gli1 were assayed by in situ hybridization. Shh is broadly expressed in the lingual epithelium at E12 but becomes progressively restricted to developing circumvallate and fungiform papillary epithelia. Shh is expressed specifically within the central cells of the papillary epithelium starting at E13.5 and persisting through E18. Ptc and Gli1 expression follow a pattern similar to that of Shh. Compared to Shh, Ptc is expressed in larger regions surrounding the central papillary cells and also in the mesenchyme underlying Shh-expressing epithelium.

Innervation of taste papillae was examined by the pan-neuronal antibody to ubiquitin carboxyl terminal hydrolase (PGP 9.5). Nerves reach the basal lamina of developing taste papillae at E14, to densely innervate the papillary epithelium by E16. Thus, the pattern of Shh expression within developing taste papillae is established prior to innervation, ruling out neuronal induction of papillae. Our results suggest that the Shh signaling pathway may be involved in: (1) establishing papillary boundaries in taste papilla morphogenesis, (2) papillary epithelial-mesenchymal interactions, and/or (3) specifying the location or development of taste buds within taste papillae.

INTRODUCTION

The gustatory system is unique in that its receptor cells develop from local epithelium rather than neurogenic ectoderm as in other sensory systems (Barlow and Northcutt 1995; Stone, Finger et al. 1995). In mammals, taste receptor cells are organized into taste buds which, on the tongue, are located within specialized papillae (Mistretta 1991). In rodents, taste papillae develop prenatally while taste buds subsequently appear within these papillae at about the time of parturition (Paulson, Hayes et al. 1985; Mistretta 1991). Thus, the formation of papillae is an important first step in development of the gustatory system in mammals.

Four types of lingual papillae are present on the tongues of mice (and other rodents): fungiform, circumvallate, foliate, and filiform -- the first three of these contain taste buds and are referred to as taste papillae. Morphogenesis of fungiform and circumvallate papillae occurs in a stereotyped pattern on the tongue during prenatal development, from E12-E16 in mice (Paulson, Hayes et al. 1985; Mistretta 1991)). Fungiform papillae form in longitudinal rows on the anterior portion of the tongue with the medial rows forming prior to more lateral ones and the more anterior papillae developing first within each row (Paulson, Hayes et al. 1985; Farbman and Mbiene 1991). A single circumvallate papilla forms on the midline at the oral-pharyngeal border of the tongue. The foliate papillae form on the lateral edge of the tongue at the level of the intermolar eminence.

Although each type of papilla is morphologically distinct, the initial events in their development are histologically similar. Taste papillae begin as placodal thickenings in the lingual epithelium (Paulson, Hayes et al. 1985; Farbman and Mbiene 1991; Mistretta 1991; Fujimoto, Yamamoto et al. 1993). The placodal epithelium then begins to grow into the underlying mesenchyme and evaginates into a raised structure. At the same time, nerve bundles begin to grow into the tongue and ultimately reach the lingual epithelium (Farbman and Mbiene 1991; Whitehead and Kachele 1994). Importantly, studies of cultured explants of embryonic rat tongue have shown that the fungiform papillae initially develop normally despite the absence of innervation (Farbman and Mbiene 1991; Mbiene, Maccallum et al. 1997), suggesting that papillary morphogenesis is independent of innervation. However, innervation does seem to be necessary in mammals for completion of normal taste bud development(Oakley, Brandemihl et al. 1998).

The early stages in the formation of taste papillae resemble those of other specialized epithelial structures, particularly teeth and feather buds. In both of those developing systems, a cascade ofintercellular signaling interactions specifies organ position and induces morphogenesis (Chuong 1993; Kratochwil, Dull et al. 1996). One signaling molecule known to operate in vertebrate patterning and tissue induction is Sonic hedgehog (Shh), a homolog of the Drosophila Hedgehog (Hh) signaling molecule. Shh signaling utilizes the transmembrane Patched (Ptc) protein as a receptor and the Gli1 transcription factor as part of its signal transduction mechanism, (Chen and Struhl 1996; Hahn, Christiansen et al. 1996; Marigo, Davey et al. 1996). Shh is well known for inducing ventral neural and somitic tissues and for specifying limb anterior-posterior polarity (Hammerschmidt, Brook et al. 1997). Shh also appears to be involved in many other developmental processes, including whisker, tooth, and even lung formation (Bitgood and McMahon 1995; Bellusci, Furuta et al. 1997; Hammerschmidt, Brook et al. 1997).

Recent studies in mice have reported expression of Shh in the tongue during the period of taste papillary morphogenesis prior to the formation of the taste buds (Bitgood and McMahon 1995). This raises the possibility that Shh is one of the signals active in patterning the lingual epithelium. Because Shh is involved in several developmental processes (Hammerschmidt, Brook et al. 1997), it could play multiple roles with regard to papillary development. For example, Shh could act as a polarizing signal as it does in the limb bud (Johnson, Riddle et al. 1994), establishing tongue polarity for subsequent papillary development. It also could act as an inductive signal as it does in the neural tube (Echelard, Epstein et al. 1993; Roelink, Porter et al. 1995), inducing development of lingual epithelium along papillary cell fates. Shh could be active in epithelial-mesenchymal interactions or establishment of papillary spacing or borders as well. Finally, Shh could be involved in later processes of lingual gustatory development, potentially specifying the location of and/or inducing taste receptor cell differentiation.

In order to clarify the role that Shh plays in the development of taste papillae, we have conducted a detailed study of the timing and distribution of lingual expression of members of the Shh signaling pathway from E11 to E16.5 in mice. This corresponds to the period of lingual organogenesis and morphogenesis of fungiform and circumvallate papillae. We also have utilized a pan-neuronal antibody to ubiquitin carboxyl terminal hydrolase, PGP 9.5, to study papillary innervation in relation to timing of expression of Shh pathway genes. We report here that Shh, Ptc, and Gli1 expression occurs throughout the early lingual epithelium and becomes progressively restricted to the regions in and around the developing fungiform and circumvallate papillae prior to innervation of the papillary epithelium.

MATERIALS AND METHODS

Embryo collection and staging

Timed pregnant CD-1 mice were obtained from Charles River (Wilmington, MA). Use of these animals was approved by the UCD Animal Care and Use Committee and conformed to NIH guidelines. The mice were overdosed with halothane, decapitated on appropriate days from 10-18 days post coitus and embryos were collected into 4% phosphate buffered (0.1 M, pH 7.2) paraformaldehyde (PFA). Embryos were fixed overnight at 4° C and staged by crown-rump (CR) length according to (Rugh 1968) and by other morphological features (Kaufman 1992). The tissue was dehydrated by sequential washes in 50%, 100%, 100% methanol and stored at -20° C. Upon use, embryos were washed sequentially with 50% methanol, and phosphate buffered saline (PBS: 0.01 M Sodium phosphate, 0.13 M NaCl; pH 7.4) plus 0.1% Tween-20 (PBST; three times). Heads, jaws or tongues were then dissected out for hybridization. For sections, some dissected tissue was equilibrated with 20% sucrose in 4% PFA, then cut into 15-20 µm sagittal sections and placed on Superfrost slides (Fisher, Pittsburgh, PA).

Following whole mount expression analysis, the lingual developmental stage was assessed by morphology and tongue length (circumvallate to tip). Comparison of these data to CR length showed CR length to be an accurate predictor of relative lingual development.

Whole-mount and section in situ hybridization

Whole or sectioned tissue was washed in PBST and treated 10-20 minutes with proteinase K (10 µg/ml in PBST). It was washed twice in PBST and refixed for 20 minutes in 4% PFA. Embryos were then prehybridized 2 hours at 42° C in hybridization buffer containing: 50% formamide; 5x hybridization buffer (20x: 3M NaCl, 100 mM EDTA, 100 mM PIPES, pH 6.8); 1x Denhardts, 250 µg/ml sheared salmon sperm DNA; 250 µg/ml poly-A and 0.1% Tween-20. Digoxigenin-labeled sense and antisense riboprobes were produced from pBluescript II vectors as described eleswhere (Goodrich, Johnson et al. 1996, Hui, 1994 #27). Details about the probes are listed in table 1. Hybridization was performed overnight at 60° C in hybridization buffer containing 5.5% dextran sulphate and 0.2-0.5 µg/ml riboprobe. Excess probe was removed by sequential washes in 2x SSC (3 times, 60° C), 0.2x SSC (three times, 60° C), 1:1 0.2x SSC: 0.1 M phosphate buffer (PB), and PB (2 times). Nonspecific binding in the tissue was blocked 1-2 hours in 10% sheep serum, 4% dry milk, 2 mg/ml BSA, 0.3% Triton X-100 in 0.1 M PB. After this treatment, the tissue was incubated overnight with anti-digoxigenin antibody conjugated to alkaline phosphatase diluted 1:1000 in blocking solution. Excess antibody was removed by washes in 0.1 M PB and the tissue was equilibrated with color buffer containing: 100 mM Tris, pH 9.5; 50 mM MgCl₂; 100 mM NaCl; and 0.1% Tween 20. Antibody was visualized by the NBT/BCIP blue color reaction. Prior to photography, the tissue was refixed in 4% PFA. Following hybridization, some whole-mount tongues were cryoprotected by incubation in 25% sucrose in 4% PFA overnight and then cut on a cryostat to 20 µM sections.

TABLE 1:ISH probes used

PGP 9.5 Immunohistochemistry

After rehydration and sectioning, some tongues were analyzed for presence of the neuronal marker Protein Gene Product 9.5 (PGP 9.5, ubiquitin carboxyl terminal hydrolase). Sections were rinsed in 0.1 M PB and blocked for 1 hour in 10% donkey serum in 0.1 M PB + 0.3% Triton X-100. Incubation with primary rabbit anti-PGP 9.5 (1:1000; Biogenesis, Poole, UK) occurred overnight at 4° C in a humidified chamber. After rinsing with 0.1 M PB, sections were incubated with Lissamine Rhodamine Sulfonyl Chloride (LRSC)-conjugated donkey anti-rabbit IgG (1:100; Jackson, West Grove, PA) 1-2 hours at room temperature. Slides were rinsed with 0.1 M PB and coverslipped with Fluoromount G (Southern Biotechnology Associates, Birmingham, AL). Immunofluorescence and Nomarski images were collected with an Olympus Fluoview confocal microscope.

All figures were composed using Adobe Photoshop 4.0.1 with only overall adjustment of levels and color balance for each image. Immunofluorescence/Nomarski overlay images were created within Photoshop using layering effects.

RESULTS

Lingual morphogenesis in the mouse

The basic progression of mouse tongue development can be seen in figures 1 and 2. As described by (Paulson, Hayes et al. 1985), our studies in mice confirm that murine tongue development is morphologically similar to that described in rat (Slavkin, Bringas et al. 1989; Farbman and Mbiene 1991; Mbiene, Maccallum et al. 1997). A tongue bud is first visible early in the 12th embryonic day and quickly becomes a distinct tongue by E12.5. It lengthens and takes on both lateral and dorsal-ventral curvature during E13-14, achieving the adult spatulate shape by E16.

Lingual expression patterns of Shh signaling pathway members

Sonic hedgehog

Figure 1a and 1b show the results of whole mount ISH for Shh in developing tongues of embryos age E12 and E12.5. At E11, there is no distinguishable tongue and the hybridization signal from the oral floor is not significantly above background (not shown). Early in the 12th embryonic day, a tongue primordium is visible and Shh is expressed diffusely throughout this early tongue. A dark staining region of higher Shh expression is present just anterior to the foramen caecum, corresponding to the future location of the circumvallate papilla. Broad lingual expression persists as the tongue grows until around E12.5. At this time, Shh expression is comparatively diminished along the midline of the tongue and areas of higher expression can be seen in rows parallel to the long axis of the tongue. This is the same pattern and location in which the first fungiform papillae develop (Paulson, Hayes et al. 1985; Farbman and Mbiene 1991). Sections of hybridized tongues at this age show that Shh expression is localized to lingual epithelial cells and is absent from the underlying mesenchyme (Fig. 3a). The region of dark staining at the location of the developing circumvallate papilla persists. As the late E12 tongue grows, Shh expression continues to be focused into the developing papillae.

During the 13th and 14th embryonic days, the tongue takes on the curvature and shape of a fully-developed mouse tongue. At E13, the rows of higher Shh expression at the presumed sites of fungiform papillae, first seen at E12.5, have condensed into more discrete regions of expression (not shown). Beginning at E13.5, and continuing throughout gestation, Shh expression is tightly localized to the regions of developing fungiform and circumvallate papillae (Fig. 2a). The broad Shh expression seen previously has diminished and is not significantly above background outside of taste papilla regions. Sections of hybridized tongues at this stage show that Shh expression is localized to small groups of columnar epithelial cells, 5-7 cells across (Fig. 3c), which apparently correspond to the placodal thickenings that eventually give rise to fungiform papillae (Mistretta 1991).

The restricted Shh expression in taste papillae persists through the end of gestation. From E16 onward, Shh is expressed in the central core of the developing papillary epithelium (Figs. 2e, 4a). Older tongues are not amenable to whole mount ISH, but hybridization to sectioned tissue shows that Shh expression remains localized to the central set of cells on the surface of the fungiform papillae through E18 (Fig. 4e).

Patched

The expression of Ptc in the developing tongue correlates with that of Shh, as observed in other systems (Goodrich, Johnson et al. 1996). The results of whole mount ISH with Ptc antisense RNA probes are shown infigures 1 and 2. Ptc expression is not detectable above background during fusion of the mandibular arches at E11.5 (not shown). Broad Ptc expression occurs throughout the tongue primordium early in the 12th embryonic day. By E12.5, expression is obviously higher in the regions that will form taste papillae (Fig. 1d). Thus, in the earliest stages of tongue development (E12-E13), the pattern of Ptc expression mirrors that of Shh. In sections of tongues at these stages, Ptc expression is present in both the lingual epithelium and its underlying mesenchyme (Fig. 3b). This is the first time at which Shh and Ptc are expressed in different cell populations within the tongue.

Ptc expression, similar to Shh expression, continues over time to coalesce into regions around taste papillae (Fig. 2b). By E13.5, significant Ptc expression is present only in regions of developing taste papillae. However, the expression pattern around these papillae is distinct from that of Shh. The regions of Ptc expression near the fungiform papillae are more diffuse and larger than the regions expressing Shh centered in the developing papillae. Also, Ptc expression near the circumvallate papilla forms a ring surrounding the center of the papilla, rather than being centered in the papilla like Shh. In sections, expression of Ptc is less discrete than that of Shh. Ptc expression extends beyond the columnar epithelial cells of the developing fungiform papillae laterally 10-15 µm in the lingual epithelium and vertically 5-10 µm into the mesenchymal cells of the tongue (Fig. 3d).

At later stages of lingual development, Ptc expression changes slightly while remaining localized to taste papillae. By E16.5, expression near the circumvallate papilla is within the entire region of the papilla (Fig. 2f). In the fungiform papillae, Ptc expression at E16.5 is clearly in an annular pattern around each papilla: lower in the center of the papillae and darker around the edge (Fig. 2g). This is the reverse of Shh expression at this age, which is focused discretely in the center of the developing papilla (Fig. 2e). Sectioned tongues show that Ptc expression continues to be localized to a larger region within fungiform papillae, extending both laterally and vertically beyond where Shh is expressed. Lower Ptc expression is seen in the epithelium in the center of the developing papillae, consistent with the annular pattern seen in whole tongues (Fig. 4b). Ptc expression in the circumvallate papillae is in both the epithelium and underlying mesenchymal cells at this stage (Fig. 4d). Again like Shh, Ptc is expressed in fungiform (Fig. 4f) and circumvallate (not shown) papillae through E18.

Gli1 and Gli3

Gli1 is the transcription factor responsible, at least in part, for cellular responses to Shh (Martí, Takada et al. 1995; Hammerschmidt, Brook et al. 1997). In the tongue, Gli1 expression mirrors that of Ptc, shown in figures 1 and 2. Gli1 is broadly expressed in the early tongue primordium at E12 and is progressively localized to taste papillae in a pattern resembling that of Ptc. Gli1 expression does not decrease in non-papillary regions as quickly as Shh or Ptc; some broad Gli expression persists through the thirteenth embryonic day (Fig. 2c). At later stages of lingual development, Gli1 expression exists in an annular pattern surrounding each fungiform papilla, again similar to Ptc (not shown).

Gli3 is another member of the Gli transcription factor family. It is not thought to be involved in Shh response and can act antagonistically to Shh and Gli (Marigo, Johnson et al. 1996). We do not see Gli3 expression above background levels in the tongue.

Innervation of developing taste papillae

The time course of innervation of papillary epithelium from E13-E16 was examined by immunohistochemistry for the neuronal marker PGP 9.5 (ubiquitin carboxyl terminal hydrolase). At E13, nerve fibers can be seen growing towards the surface of the anterior tongue and just below the circumvallate placode (Fig. 5a-d). Nerves have not yet reached the epithelium at this time. At E14, immunofluorescence extends to the basement membrane just below developing fungiform and circumvallate papillary epithelium (Fig. 5e-h). The sites of nerve growth to the epithelium are restricted to the developing circumvallate and fungiform papillae; no fibers appear to be innervating non-papillary epithelium. Nerves begin to penetrate into the papillary epithelium by E15 (Fig. 6a-d). By E16, the entire circumvallate and central core of the fungiform papillary epithelia are densely innervated by nerve fibers (Fig. 6e-h).

Taken together with the ISH data presented above, it appears that Shh expression is localized to taste papillae prior to nerve fiber contact with the lingual epithelium. Shh expression is specifically localized to the developing circumvallate papillae as early as E12 and to the developing fungiform papillae by E13.5. Nerve fibers are not present at the papillary epithelium until E14 and do not innervate the entire papillary epithelium until E16.

DISCUSSION

Development of lingual taste receptors can be divided into two stages. First, taste papillae form on the tongue in a characteristic spatial and temporal pattern. This process occurs from E12.5 to E16 in mice and requires several steps: an inductive event specifying the locations of taste papillae; proliferation and evagination of the developing papillary epithelium; and growth of a supportive mesenchymal core (Farbman and Mbiene 1991; Mistretta 1991). Second, taste buds develop within the papillary epithelium following papillary morphogenesis. This process may require an additional inductive event to specify taste receptor cell differentiation and must be coordinated with innervation of the lingual epithelium (Farbman and Mbiene 1991; Oakley 1991). The lingual expression patterns of Shh signaling pathway members relative to morphological stages are summarized in figure 7. These results indicate that Shh may function in both stages of gustatory development.

Sonic Hedgehog signaling in papillary morphogenesis

Expression of Shh, Ptc, and Gli1 begins broadly in the tongue primordium early in the 12th embryonic day and becomes progressively localized to regions in and around taste papillae. Shh is detectable only in epithelial cells throughout lingual development and its expression is exclusively located in the center of fungiform and circumvallate papillae by E13.5. Ptc expression has a slightly different distribution than Shh as it is expressed in larger regions surrounding developing fungiform papillae than Shh. Ptc expression is lower in, or absent from, the Shh-expressing centers of fungiform papillae at later stages of papillary development. Ptc is also expressed in mesenchyme underlying Shh-expressing areas throughout papillary morphogenesis. Both Ptc and Gli1 are transcriptionally activated in response to Shh in tissues adjacent to Shh signaling centers (Goodrich, Johnson et al. 1996; Marigo, Davey et al. 1996; Marigo, Scott et al. 1996; Sasaki, Hui et al. 1997). Thus, expression of Ptc or Gli1 indicates those cells or tissues that are actively responding to the Shh signal. Because Ptc is expressed both within the lingual epithelium and in the underlying mesenchyme, Shh may play a role in papillary morphogenesis by signaling to one or both of these tissues.

In the epithelium, Shh may be involved in establishing papillary boundaries. This situation is exemplified in Drosophila where Hedgehog (Hh) is responsible for specifying embryonic parasegment borders (Kalderon 1995). Reciprocal signals between Hh and Wingless establish two adjacent rows of cells that follow separate developmental fates. In the same way, reciprocal signals between Shh and another signaling molecule, produced by lingual epithelium around the perimeter of the developing papillae, may establish a boundary for papillary development. The tight localization of Shh to an easily identified set of papillary epithelial cells suggests that a boundary formation or restriction process is active in papillary development.

Alternatively, Shh could specify papillary boundaries within the lingual epithelium in a concentration-dependent manner as in the developing neural tube. In neural tube development, high levels of Shh from the notochord promote floor plate differentiation while lower concentrations induce motor neuron cell fates (Roelink, Porter et al. 1995). Shh may act in a similar fashion by specifying papillary epithelial growth and differentiation at a certain threshold concentration. In this model, cells at the boundaries of the developing papillae would follow nonpapillary fates as the Shh concentration decreases with increasing distance from Shh-expressing centers.

Shh might also play a role in the epithelial-mesenchymal interactions of papillary morphogenesis. Papillary development requires coordination between epithelium and mesenchyme, which grows and differentiates to form the supportive core of the papilla (Farbman and Mbiene 1991; Mistretta 1991; Fujimoto, Yamamoto et al. 1993). Expression of Ptc in the mesenchyme underlying developing taste papillae indicates that this tissue is responding to the epithelial Shh signal. The epithelial Shh signal could be responsible for induction or support of growth of the mesenchymal core of taste papillae.

Shh-mediated epithelial-mesenchymal interactions are well-known in other systems. During tooth formation,Shh is expressed in the epithelial enamel knot, which acts as a signaling center for organization of dental epithelial-mesenchymal interactions (Vaahtokari, Aberg et al. 1996). Epithelial Shh expression with corresponding mesenchymal Ptc expression also has been observed in developing whisker barrels (Bitgood and McMahon 1995; Iseki, Araga et al. 1996; Platt, Michaud et al. 1997) and feather buds (Nohno, Kawakami et al. 1995; Jung, Francis-West et al. 1998). These structures develop similarly to taste papillae in that they begin as placodal epithelial thickenings and gain a mesenchymal core as they grow (Fristrom 1988). The involvement of Shh in epithelial-mesenchymal interactions in such similar structures supports the idea that it is also involved in epithelial-mesenchymal signaling in taste papillae.

Sonic Hedgehog in taste bud determination

From E15 through E16.5, Shh expression remains in the lingual epithelium, restricted to a small set of cells in the center of each fungiform papilla and throughout the epithelium of the circumvallate papilla. This distribution coincides with the eventual locations of taste buds. Taste buds form postnatally, initially within the dorsal epithelium of the circumvallate papilla and in the centers of fungiform papillae (Mistretta 1991). These are the same sites where Shh expression concentrates at E16.5 in fully formed lingual papillae. Thus, Shh may be involved in morphogenesis of taste buds.

Support for this proposed role of Shh comes from studies of neurotrophic factors within the developing tongue. The distribution of lingual Shh expression after E15.5 matches that of the neurotrophic factor BDNF at the same stage (Nosrat and Olson 1995). BDNF is believed to support gustatory nerves that grow into developing taste buds (Nosrat, Ebendal et al. 1996; Zhang, Brandemihl et al. 1997). Taking BDNF to be an early marker of differentiation of the gustatory epithelium, Shh is in the right place at the right time to be responsible for specification or support of gustatory development. Shh is, however, probably not solely responsible for taste bud histogenesis because that process requires innervation to proceed normally in mammals (Oakley 1991; Fritzsch, Sarai et al. 1997; Mbiene, Maccallum et al. 1997; Oakley, Brandemihl et al. 1998). Shh could be inducing competence to form taste receptor cells within the central papillary epithelial cells which then differentiate further due to interaction with gustatory nerves.

Taste Papilla Patterning

Gustation is the only sensory system whose receptors are not derived from neurogenic ectoderm. Instead, the receptors of the gustatory system arise from local epithelium (Barlow and Northcutt 1995; Stone, Finger et al. 1995). While taste buds develop in the absence of papillae in the epiglottis and palate, lingual taste buds develop solely within taste papillae (Mistretta 1991). Thus, the positioning of, and possibly the development of, taste receptor cells in the tongue may be dependent upon the formation of taste papillae.

Taste papillary morphogenesis is a process intrinsic to the tongue. In rats, taste papillae develop in the normal spatial and temporal sequence in the absence of innervtion (Farbman and Mbiene 1991; Mbiene, Maccallum et al. 1997), so neural induction is not responsible for taste papilla patterning. This conclusion is supported by comparing the time course of papillary innervation to Shh expression (Figs. 2, 5). Discrete localization of Shh to developing papillae occurs by E12.5 in the presumptive circumvallate papilla and by E13.5 in fungiform papillary precursors. However, nerve fibers have not yet reached the lingual epithelium by E13 and extend only to the level of the basement membrane at E14. So, Shh expression in the papillary precursor regions is established prior to their innervation. Papillae are thus specified independent of neuronal influence.

If papillary development is initiated independently within the tongue, how is the pattern established? If Shh were to be patterning papillae by self-restriction, then one would expect to see an evenly-spaced, random distribution of papillae. The actual distribution of taste papillae is too regular and stereotyped for this. As monitored by Shh expression, the process of papilla patterning occurs in at least two steps. First is restriction of precursor regions to parasagittal rows along either side of the tongue. Following this, individual papillae are generated within these rows. This is reminiscent of the patterning of feathers in birds, another type of epithelial specialization. Feather buds, too, arise as Shh-expressing placodal thickenings in a patterning process intrinsic to the epidermal epithelium (Chuong 1993; Nohno, Kawakami et al. 1995). Prior to expression of Shh within feather placodes, Shh and FGF-4 are expressed in a stripe running along the dorsal midline. This stripe then changes to become discrete Shh-expressing regions that form feather buds (Jung, Francis-West et al. 1998). This process is thought to occur through epithelial-mesenchymal interactions mediated by Shh and other signaling molecules (Jung, Francis-West et al. 1998).

A similar process involving Shh may be responsible for taste papilla patterning. Taste papilla, feather, and tooth development are all alike in that these structures begin as placodal thickenings and then evaginate or invaginate to form a raised structure with a mesenchymal core (Fristrom 1988; Lumsden 1988; Mistretta 1991; Chuong 1993; Fujimoto, Yamamoto et al. 1993). Additionally, Shh is expressed in tooth and feather primordia as well as taste papilla precursors (Bitgood and McMahon 1995; Nohno, Kawakami et al. 1995); Fig 1a). Feather position is thought to be specified by interaction between FGF-4, BMP-2, BMP-4 and Shh (Jung, Francis-West et al. 1998). Recent studies also indicate that tooth position is specified by overlap of mesenchymal FGF-8 and BMP signals (Neubüser, Peters et al. 1997). A patterning process involving overlap of several intercellular signals may be active in the developing tongue. Investigation of the role of these other signaling families in the tongue may help to further explain papillary morphogenesis and lingual development.

ACKNOWLEDGMENTS

The authors thank K. Anderson for expert technical assistance and sectioned in situ data. We are grateful to L. Goodrich at Stanford University for the Shh and Ptc cDNA plasmids, and A. Joyner at New York University for the Gli1 and Gli3 clones. Thanks to L. Barlow for critical reading and comments on this manuscript. We also thank A. Ribera for use of her laboratory facilities and photographic equipment.

This work was supported by NIDCD grant DC00244 to T.E.F. and training grant GM08497 to J.M.H.

Barlow, L. A. and R. G. Northcutt (1995). Embryonic origin of amphibian taste buds. Developmental Biology 169(1): 273-285.

Bellusci, S., Y. Furuta, et al. (1997). Involvement of Sonic hedgehog (Shh) in mouse embryonic lung growth and morphogenesis. Development 124(1): 53-63.

Bitgood, M. J. and A. P. McMahon (1995). Hedgehog and Bmp genes are coexpressed at many diverse sites of cell-cell interaction in the mouse embryo. Developmental Biology 172(1): 126-138.

Chen, Y. and G. Struhl (1996). Dual roles for patched in sequestering and transducing Hedgehog. Cell 87(3): 553-563.

Chuong, C. (1993). The making of a feather: homeoproteins, retinoids and adhesion molecules. Bioessays 15: 513-521.

Echelard, Y., D. J. Epstein, et al. (1993). Sonic hedgehog, a member of a family of putative signaling molecules, is implicated in the regulation of CNS polarity. Cell 75(7): 1417-1430.

Farbman, A. I. and J. P. Mbiene (1991). Early development and innervation of taste bud-bearing papillae on the rat tongue. Journal of Comparative Neurology 304(2): 172-186.

Fristrom, D. (1988). The cellular basis of epithelial morphogenesis. A review. Tissue & Cell 20(5): 645-690.

Fritzsch, B., P. A. Sarai, et al. (1997). Mice with a targeted disruption of the neurotrophin receptor trkB lose their gustatory ganglion cells early but do develop taste buds. International Journal of Developmental Neuroscience 15(4-5): 563-576.

Fujimoto, S., K. Yamamoto, et al. (1993). Pre- and postnatal development of rabbit foliate papillae with special reference to foliate gutter formation and taste bud and serous gland differentiation. Microscopy Research & Technique 26(2): 120-132.

Goodrich, L. V., R. L. Johnson, et al. (1996). Conservation of the hedgehog/patched signaling pathway from flies to mice: induction of a mouse patched gene by Hedgehog. Genes & Development 10(3): 301-312.

Hahn, H., J. Christiansen, et al. (1996). A mammalian patched homolog is expressed in target tissues of sonic hedgehog and maps to a region associated with developmental abnormalities. Journal of Biological Chemistry 271(21): 12125-12128.

Hammerschmidt, M., A. Brook, et al. (1997). The world according to hedgehog. Trends in Genetics 13(1): 14-21.

Iseki, S., A. Araga, et al. (1996). Sonic hedgehog is expressed in epithelial cells during development of whisker, hair, and tooth. Biochemical & Biophysical Research Communications 218(3): 688-693.

Johnson, R. L., R. D. Riddle, et al. (1994). Sonic hedgehog: a key mediator of anterior-posterior patterning of the limb and dorso-ventral patterning of axial embryonic structures.Biochemical Society Transactions 22(3): 569-574.

Jung, H.-S., P. H. Francis-West, et al. (1998). Local inhibitory action of BMPs and their relationships with activators in feather formation: Implications for periodic patterning. Developmental Biology 196: 11-23.

Kalderon, D. (1995). Morphogenetic signalling. Responses to hedgehog. Current Biology 5(6): 580-582.

Kaufman, M. H. (1992). The Atlas of Mouse Development. London, Academic Press.

Kratochwil, K., M. Dull, et al. (1996). Lef1 expression is activated by BMP-4 and regulates inductive tissue interactions in tooth and hair development. Genes & Development 10(11): 1382-1394.

Lumsden, A. G. (1988). Spatial organization of the epithelium and the role of neural crest cells in the initiation of the mammalian tooth germ. Development. Supplement 103: 155-169.

Marigo, V., R. A. Davey, et al. (1996). Biochemical evidence that patched is the Hedgehog receptor. Nature 384(6605): 176-179.

Marigo, V., R. L. Johnson, et al. (1996). Sonic hedgehog differentially regulates expression of GLI and GLI3 during limb development. Developmental Biology 180(1): 273-283.

Marigo, V., M. P. Scott, et al. (1996). Conservation in hedgehog signaling: induction of a chicken patched homolog by Sonic hedgehog in the developing limb. Development 122(4): 1225-1233.

Martí, E., R. Takada, et al. (1995). Distribution of Sonic hedgehog peptides in the developing chick and mouse embryo. Development 121(8): 2537-2547.

Mbiene, J. P., D. K. Maccallum, et al. (1997). Organ cultures of embryonic rat tongue support tongue and gustatory papilla morphogenesis in vitro without intact sensory ganglia. Journal of Comparative Neurology 377(3): 324-340.

Mistretta, C. M. (1991). Developmental neurobiology of the taste system. Taste and Smell in Health and Disease. T. V. Getchell. New York, Raven Press: 35-64.

Neubüser, A., H. Peters, et al. (1997). Antagonistic interactions between FGF and BMP signaling pathways: a mechanism for positioning the sites of tooth formation. Cell 90(2): 247-255.

Nohno, T., Y. Kawakami, et al. (1995). Involvement of the Sonic hedgehog gene in chick feather formation. Biochemical & Biophysical Research Communications 206(1): 33-39.

Nosrat, C. A., T. Ebendal, et al. (1996). Differential expression of brain-derived neurotrophic factor and neurotrophin 3 mRNA in lingual papillae and taste buds indicates roles in gustatory and somatosensory innervation. Journal of Comparative Neurology 376(4): 587-602.

Nosrat, C. A. and L. Olson (1995). Brain-derived neurotrophic factor mRNA is expressed in the developing taste bud-bearing tongue papillae of rat. Journal of Comparative Neurology 360(4): 698-704.

Oakley, B. (1991). Neuronal-epithelial interactions in mammalian gustatory epithelium. Ciba Foundation Symposium 160: 277-287.

Oakley, B., A. Brandemihl, et al. (1998). The morphogenesis of mouse vallate gustatory epithelium and taste buds requires BDNF-dependent taste neurons. Brain Res. Dev. Brain Res. 105(1): 85-96.

Paulson, R. B., T. G. Hayes, et al. (1985). Scanning electron microscope study of tongue development in the CD-1 mouse fetus. Journal of Craniofacial Genetics & Developmental Biology 5(1): 59-73.

Platt, K. A., J. Michaud, et al. (1997). Expression of the mouse Gli and Ptc genes is adjacent to embryonic sources of hedgehog signals suggesting a conservation of pathways between flies and mice. Mechanisms of Development 62(2): 121-135.

Roelink, H., J. A. Porter, et al. (1995). Floor plate and motor neuron induction by different concentrations of the amino-terminal cleavage product of sonic hedgehog autoproteolysis. Cell 81(3): 445-455.

Rugh, R. (1968). The mouse: its reproduction and development. Minneapolis, Burgess.

Sasaki, H., C. Hui, et al. (1997). A binding site for Gli proteins is essential for HNF-3beta floor plate enhancer activity in transgenics and can respond to Shh in vitro. Development 124(7): 1313-1322.

Slavkin, H. C., P. Bringas, Jr., et al. (1989). Early embryonic mouse mandibular morphogenesis and cytodifferentiation in serumless, chemically defined medium: a model for studies of autocrine and/or paracrine regulatory factors. Journal of Craniofacial Genetics & Developmental Biology 9(2): 185-205.

Stone, L. M., T. E. Finger, et al. (1995). Taste receptor cells arise from local epithelium, not neurogenic ectoderm. Proceedings of the National Academy of Sciences of the United States of America 92(6): 1916-1920.

Vaahtokari, A., T. Aberg, et al. (1996). The enamel knot as a signaling center in the developing mouse tooth. Mechanisms of Development 54(1): 39-43.

Whitehead, M. C. and D. L. Kachele (1994). Development of fungiform papillae, taste buds, and their innervation in the hamster. Journal of Comparative Neurology 340(4): 515-530.

Zhang, C., A. Brandemihl, et al. (1997). BDNF is required for the normal development of taste neurons in vivo. Neuroreport 8(4): 1013-1017.