The fibroblast growth factor (FGF) family comprises a large group of about 20 polypeptides. These proteins are secreted and they signal through one or two of four types of cognate receptors, using heparin as a necessary cofactor for binding to the receptor. They have diverse properties, and act in modifying cellular behaviors, providing positional information, and generating localized instructive activity (Ornitz and Itoh, 2001). These properties are transduced in a number of ways, the best characterized of which is the MAP kinase pathway. Unsurprisingly, the family plays an important role during embryogenesis, with a number of members shown to play key roles in the development of several organ systems. For example, FGF-4, FGF-8, FGF-10, and FGF-18 have all been proposed to play roles in limb initiation and early patterning (Capdevila and Izpisua Belmonte, 2001; Martin, 2001). Moreover, FGF-8 is thought to participate in a signaling cascade that organizes the midbrain-hindbrain (Crossley et al., 1996), and FGF-10 has been shown to be essential for the development of the lungs (Bellusci et al., 1997; Sekine et al., 1999).
The existence of such a large number of FGF family members complicates the elucidation of the precise roles played by individual FGFs in organogenesis, especially since functional redundancy appears to occur among family members. As a first step in establishing the range of organ systems regulated by FGFs, we wished to catalog the expression of all of the available FGF family members during chick organogenesis, focusing on the head, trunk, limbs, heart, and tail at stages 18–25 (according to Hamburger and Hamilton (1951)). In the interest of brevity, this report does not seek to be an exhaustive analysis of the expression of FGFs; rather, we provide a survey highlighting sites of expression of fgf-2, -3, -4, -8, -10, -12, -13, -14, and -18 during chick organogenesis.
MATERIALS AND METHODS
Fertilized hen's eggs were obtained from Dunlap Hatchery (Caldwell, ID) and incubated in humidified incubators at 38°C until they reached stages 18–24 (Hamburger and Hamilton, 1951). Embryos were then dissected from the eggs, washed with 123 mM saline, and fixed overnight with 4% paraformaldehyde in PBS.
Whole Mount In Situ Hybridization
Antisense riboprobes, labeled with digoxygenin, were transcribed from plasmids containing the relevant cDNA corresponding to a member of the FGF family. Probes were used for the following genes (numbers in parentheses indicate the Genbank accession numbers): fgf-2 (M95706), -3 (Z47555), -4 (U14654), -8 (U41467), 10 (D86333), -12 (AF199602), -13 (AF108757), -14 (AF199606), and -18 (AB030229). These were hybridized to fixed embryos overnight at 65°C, as previously described (Ladher et al., 2000). The localization of transcripts was revealed using an antibody directed against the digoxygenin moiety, with alkaline phosphatase covalently attached. After extensive washing, a staining mixture of NBT/BCIP was used to produce a purple precipitate at sites of alkaline phophatase activity.
Stained embryos that had discrete patterns of expression were sectioned. Embryos were embedded in a buffered solution of 30% gelatin at 60°C, which was then allowed to harden. A block of gelatin containing the embryo was excised and placed into 4% paraformaldehyde overnight at 4°C. After the block was rinsed with PBS, it was mounted onto the base plate of a Leica VT1000 vibratome using Krazy Glue (Elmer's Products, Columbus, OH), and sections were cut using a Gillette razor blade. Sections were then mounted onto glass slides, drained of excess liquid, and coverslipped using Aquapolymount (Polysciences, Warrington, PA).
Stained whole embryos were partially cleared with a 50% solution of glycerol in PBS. These were then photographed with a dissecting microscope using a Nikon (www.nikonusa.com) digital camera (Coolpix 990). Sections were photographed, also with a Nikon digital camera (Coolpix 990), which was mounted on a compound microscope. Sections were viewed and photographed without counterstaining using Nikon Hoffman optics. Photographic plates were assembled using Adobe Photoshop, with adjustments made for color balance and to remove background artifacts when necessary.
All FGF family members were expressed in discrete patterns at the stages we examined. Figure 1 shows overviews of the patterns of expression of each of the nine FGFs examined, and results are summarized in Table 1. More detailed descriptions of the patterns of FGF expression in the head, trunk, limbs, heart, and tail bud are given in Figures 2–7. In all figures, we used a consistent labeling system whereby the FGFs are indicated as follows: (A) fgf-2, (B) fgf-3, (C) fgf-4, (D) fgf-8, (E) fgf-10, (F) fgf-12, (G) fgf-13, (H) fgf-14, and (I) fgf-18.
Table 1. The expression patterns of nine members of the fibroblast growth factor family were examined by in situ hybridization, focusing on particular regions of the embryo
+ Indicates the presence of expression in a particular structure;– indicates that expression could not be detected. Figure numbers and panel letters illustrate these results.
We first describe the expression of FGF family members in the head. The rather broad category of “head” includes the brain, sensory organs, and branchial arches.
Fgf-8 (Figs. 1D and 2D), -13 (Fig. 2G), -14 (Figs. 1H and 2H), and -18 (Fig. 2I) are all discretely localization within the rostral brain. Fgf-8, as was previously described (Crossley et al., 1996; Stolte et al., 2002), is expressed strongly in the isthmus, the border between the midbrain and the hindbrain (Figs. 1D and 2D). In addition, it is expressed in the most rostral prosencephalon (Figs. 1D, and 2D and d: arrowheads) and optic vesicles (Figs. 1D, and 2D and d: arrows). Fgf-13 is broadly expressed in a salt-and-pepper pattern throughout the dorsal midbrain (Fig. 2G: arrowhead), and in small domains of expression in the dorsal prosencephalon (Fig. 2G and g: arrowheads). Fgf-14 is expressed in the diencephalon (Fig. 2H: arrow); this expression can be traced back to stage 8, when a broader, but still localized, pattern is observed encompassing the nascent neural tube of the future brain (Karabagli et al., in press). Fgf-18 is expressed in the caudal midbrain (Fig. 2I and i).
Fgf-12 and -14 have domains of expression in the caudal brain, with both fgf-12 and -14 being expressed broadly in the hindbrain (Fig. 2F: arrow, 2f: arrows; Fig. 2H: at section line, 2h: arrowheads).
Fgf-12, -13, and -14 are expressed within the cranial sensory ganglia. Fgf-13 expression is strong and is restricted to the trigeminal (semilunar) and acousticofacialis ganglia (Fig. 2G: arrows). Its expression in the trigeminal ganglia persists in both the ophthalmic and maxillo-mandibular lobes as the trigeminal nerves form at later stages (data not shown). Fgf-12 and -14 are expressed at lower levels in the cranial sensory ganglia than is fgf-13, with fgf-12 being expressed in the trigeminal and acousticofacialis ganglia (Fig. 2F: arrowheads) and fgf-14 being expressed in the trigeminal, acousticofacialis, and superior ganglia (Fig. 2H: arrowheads).
Fgf-3 and fgf-10 are expressed in the otic placode (Fig. 3B and E: arrowheads). Fgf-3 is expressed weakly throughout the walls of the otic placode (Fig. 3B). Fgf-10 expression is restricted to two spots within the otocyst: one in the rostral wall and the other in the caudal (Fig. 3E).
Fgf-8 (Fig. 2D: just above the “fgf-8” label), fgf-14 (Fig. 1H: arrows), and fgf-18 (Fig. 2I: arrowhead) are all expressed in the developing olfactory placodes. Fgf-14 expression in this area was described previously (Munoz-Sanjuan et al., 2001). We note that in the present comparative study the onset of fgf-14 expression lagged behind that of fgf-8 and -18, beginning at stage 22.
In the branchial arches, fgf-4, -8, -12, -14, and -18 are widely expressed. Fgf-4, -8, -12, and -14 are all expressed within the pharyngeal pouch endoderm of pouches 1–3 (e.g., Figs. 2c: arrowheads). Fgf-4, -8, -14, and -18 also show restricted expression within the facial primordia. The localization of both fgf-4 and fgf-8 in the face has already been characterized (McGonnell et al., 1998; Barlow et al., 1999); briefly, both are expressed in the epithelium at the region where the maxillary primordia meet the mandibular primordia (Fig. 2D and d: asterisks). Fgf-8 is also expressed within the caudal edges of branchial arches 2–4 (Fig. 2D: arrrow). Similarly, fgf-14 has been well characterized within the developing chick face (Munoz-Sanjuan et al., 2001). It is expressed beginning about stage 22 within the epithelium of the frontonasal primordia, and a little later at the junction between the maxillary and mandibular primordia, again in the epithelium. Fgf-18 is expressed in the epithelium of the caudal edge of branchial arches 3 and 4 (Fig. 2I: arrow). Finally, fgf-2 is expressed at low levels throughout the cephalic mesenchyme (Fig. 2a: arrows).
In this section, we include both somites and the nervous system of the trunk. Of the FGFs that we tested, Fgf-4, -8, -13, and -14 are discretely expressed in the trunk. Fgf-4 and fgf-8 are expressed strongly in the myotomal compartment of the somites by stage 18 (Fig. 4C, c, D, and d). This pattern of expression is consistent with those found in previous studies (Shamim and Mason, 1999; Stolte et al., 2002). The expression of fgf-4 marks the entire dorsoventral extent of the myotome, whereas the expression of fgf-8 marks a more central expanse of myotome. Fgf-13 is expressed somewhat weakly in the myotome, but it is strongly expressed in the dorsal root ganglia and adjacent neural tube (Figs. 4G and g). Fgf-14 is expressed strongly in the dorsal lip of the somites and less strongly more ventrally in the myotome, as well as being strongly expressed throughout the dorsoventral extent of the intermediate zone of the spinal cord (Fig. 4H and h).
Fgf-4, -8, -10, -12, -13, -14, and -18 are expressed in the limb buds during the stages that were examined (stages 18–24). The expression patterns for all of these FGFs have been characterized in previous studies, but in the present work we compare these patterns. Both fgf-4 and fgf-8 are expressed in the apical ectodermal ridge (AER) of the limb buds (Fig. 5C and D), a region controlling proximodistal outgrowth of the buds (Niswander et al., 1993; Vogel and Tickle, 1993; Mahmood et al., 1995). FGF-10 has been shown to initiate outgrowth of the limb buds (Ohuchi et al., 1997; Sekine et al., 1999), but at these stages it is expressed in a proliferating region of the limb bud known as the progress zone (Fig. 5E). Expression persists in some of these distal cells during progressively later development, remaining in the most immature skeletal elements (Fig. 5E, and data not shown). Fgf-13 is also expressed in the progress zone (Fig. 5G), more proximally then fgf-10 (Munoz-Sanjuan et al., 1999). No other expression domain could be detected outside this region. Fgf-12 is expressed in the caudal mesenchyme of the limb bud (Fig. 5F), a region known as the zone of polarizing activity (ZPA), which imparts digit identity to the autopod. This expression domain commences at stage 19 and remains present until the latest stages we examined, consistent with previous results (Munoz-Sanjuan et al., 1999). Fgf-14 expression in the limb bud is first detected at the base of the limb bud (Fig. 5H), where it has been postulated to play a role in limb innervation. At the stage shown (stage 25), a smaller expression domain can also be detected in the caudal distal sector of the hindlimb bud, again consistent with previous results (Munoz-Sanjuan et al., 2000). Fgf-18 is expressed in a domain just proximal to the progress zone, displaced rostrally within this region (Fig. 5I) (Ohuchi et al., 2000).
Only three of the FGFs we tested at stages 18–25 were expressed in the heart: fgf-2, -3, and -14 (Fig. 6A, a, B, b, H, and h). Both fgf-2 and fgf-3 are broadly expressed in the heart, and expression is detectable in both the atrium and ventricle (Fig. 6a and b). The expression of fgf-2 has already been well described (Zhu and Lough, 1996). Fgf-14 expression is restricted to the ventricle (Fig. 6h). This restriction occurs relatively early during development and is visible even at stage 10 (Karabagli et al., in press).
The tail bud of embryos at stages 18–25 is still actively segregating into mesodermal, ectodermal, neuroectodermal, and endodermal derivatives, and thus provides insight into the mechanisms that dictate these processes. Fgf-4, -8, -13, and -18 are expressed in the tail bud at these stages. All are expressed in the undifferentiated tail bud blastema, at the caudal-most end of the tail bud. Fgf-4 expression is largely restricted to this region, with some expression in more lateral mesoderm (Fig. 7C). Fgf-8 expression extends rostrally from the undifferentiated blastema into the caudal neural tube and notochord (Fig. 7D). Similarly, fgf-13 expression extends from the undifferentiated blastema into the lateral mesoderm and dorsal neural tube (Fig. 7G). No expression can be detected in the axial or paraxial mesoderm of the tail bud, closely mirroring the situation at earlier stages (Karabagli et al., in press) when fgf-13 is expressed in the more lateral mesoderm and not in the unsegmented paraxial or axial mesoderm. Fgf-18 extends from the undifferentiated blastema as a paired paraxial mesodermal domain flanking the neural tube. This domain expands rostrally as a region of higher expression, corresponding to the position at which the next pair of somites will form (Fig. 7I). This area of higher expression is not obvious at younger stages (Karabagli et al., in press).
In this report, we compare the expression patterns of nine FGF genes during organogenesis in the chick embryo. We show that even this subset of the vast FGF family is expressed in a diverse number of tissues. At the stages we studied, patterning, morphogenesis, proliferation, and differentiation all occur within various organ rudiments. Below, we discuss three major sites of expression of the FGFs described in this study, and the possible functions of FGFs in these regions.
Expression of FGFs in the Brain
One feature of our analysis of FGF expression in the brain is the patterned expression of several FGF family members in the rostral brain. Fgf-8, -18, and -14 are all expressed in progressively more rostral territories within the brain (i.e., the isthmus, midbrain, and forebrain, respectively). These expression domains are established during earlier development (Karabagli et al., in press), and they point to a pivotal role in rostral brain patterning for these factors. Indeed, data for the key roles that fgf-8 and fgf-18 play in midbrain patterning have already been well described (Ohuchi et al., 2000). Fgf-13 can be detected in a salt-and-pepper pattern in the dorsal part of the midbrain. It is unclear which cells are marked by fgf-13 expression, although lineage labeling and expression analysis at later stages should answer these questions.
Fgf-8 and -18 are also expressed in the rostral prosencephalon in similar expression domains. These factors likely play a patterning role in this region, as suggested by a comparison of the expression and function of fgf-8 and fgf-18 (Ohuchi et al., 2000). Recent knockout data suggest that fgf-8 and -18 may be functionally redundant during development of the midbrain, as fgf-18 nulls do not have a midbrain phenotype (Liu et al., 2002; Ohbayashi et al., 2002), presumably because fgf-8 compensates for the loss of fgf-18 function.
Fgf-12 and -14 are expressed through the hindbrain. As yet, the function of these molecules is unclear; however, the close structural similarity between the two molecules, and the fact that they are expressed, respectively, in the dorsal and ventromedial parts of the hindbrain, may indicate that they have distinct functions during hindbrain development.
Strikingly, a number of FGFs are expressed within the cranial sensory ganglia. At these stages of development, the ganglia undergo proliferation and morphogenesis—processes that can be mediated through FGF activity. Particularly noteworthy is the strong expression of fgf-13 within two sensory ganglia: the trigeminal and the acousticofacialis.
Expression of FGFs in the Trunk
Four of the FGF family members we surveyed were expressed in the trunk region of the embryo. Fgf-4, -8, -13, and -14 are expressed in the somites. The expression of fgf-4 in the myotome was characterized previously (Shamim and Mason, 1999); interestingly, however, fgf-8 and fgf-13 are also expressed in subsets of cells within the myotome. Why these FGFs are restricted to a subset of the myotome is unclear, but it could be due to a modification of fgf-4 activity by these signals, thus promoting the differentiation of one subtype over the other. Fgf-14 is restricted to the dorsal-medial lip of the somites.
In other regions of the trunk, fgf-13 and -14 show remarkable expression patterns. Fgf-13 is expressed in the dorsal root ganglia, and weakly in the adjacent neural tube. Again, postulating on the function of FGF-13 is difficult, but migration and/or proliferation may be an important process mediated by FGF-13 in this structure. Fgf-14 is expressed in the nonventricular zone of the neural tube. Cells in this region of the neural tube have left the proliferative ventricular zone and these cells are now beginning terminal differentiation. It is possible that FGF-14 promotes this important change in these cells.
Expression of FGF in the Limbs
The limb buds are a well characterized developmental system and much data exist on the function of FGFs during limb development (Capdevila and Izpisua Belmonte, 2001; Tickle and Munsterberg, 2001). Of the FGFs we analyzed, only two (fgf-2 and -3) were not detected in the limb buds. Fgf-4 and fgf-8 are both expressed in similar regions within the apical ectodermal ridge, a region of the limb bud responsible for its proximodistal outgrowth. Indeed, both have been shown to direct outgrowth of the limb, either in experiments in which they were substituted for the AER (Niswander et al., 1993; Mahmood et al., 1995) or through genetic ablation (Lewandoski et al., 2000; Moon and Capecchi, 2000; Moon et al., 2000; Sun et al., 2000). Fgf-10 has been shown to play a role in the initiation of the limb; however, this process occurs prior to the stages examined in this report. During the stages studied here, fgf-10 is expressed in the progress zone, a region necessary for the correct proximal-distal progression of the limb. Although a role for FGF-10 in the progression of the limb has not yet been ascribed, because the knockout never has limbs, conditional mutants of fgf-10 could address this issue. Similarly, fgf-13 is expressed within the progress zone, and overexpression studies have that it has a role in the development of the limb, with ectopic expression leading to a shortening of some of the skeletal elements (Munoz-Sanjuan et al., 1999). It is therefore probable that FGF-13 maintains cells in the blastema-like progress zone, and that down-regulation of expression allows differentiation to proceed.
Fgf-12 is expressed in the posterior mesoderm of the limb, a region that has been shown to confer positional identity to the limb bud. Again, the functional characteristics of fgf-12 have not yet been elucidated. Fgf-14 is expressed at the base of the limb bud, where it has been proposed to play a role in the innervation of the limbs. Although a previous overexpression study (Munoz-Sanjuan et al., 2000) did not show any changes in the nerve pattern of the limbs, the authors proposed that fgf-14 is under additional regulation which prevents changes in limb innervation from becoming apparent. Fgf-18 is expressed in a region of the limb mesenchyme just proximal to the progress zone. This domain overlaps with the site of the first skeletal elements to form within the limbs, and recent genetic ablation studies in mouse have found that fgf-18 is an important molecule in the regulation of chondrogenesis and osteogenesis (Ohbayashi et al., 2002; Liu et al., 2002).
The following investigators generously provided us with probes: John Lough (fgf-2), Ivor Mason (fgf-3), Gail Martin (fgf-4 and -8), Sumihare Noji (fgf-10), Jeremy Nathan and Ignacio Munoz-Sanjuan (fgf-12, -13, and -14), and Nobuyuki Itoh (fgf-18). We thank the members of the Schoenwolf laboratory, and Suzi Mansour and Tracy Wright for their advice.