[Frontiers in Bioscience 3, d781 -794, August 1, 1998]
Reprints
PubMed
CAVEAT LECTOR




Table of Conents
 Previous Section   Next Section

FGF SIGNALING IN SKELETAL DEVELOPMENT

Michael C. Naski and David M. Ornitz

Department of Molecular Biology and Pharmacology, Washington University School of Medicine, Campus Box 8103, 660 S. Euclid Ave., St. Louis, MO 63110

Received12/22/98 Accepted 2/17/98

3. LIMB DEVELOPMENT

3.1. FGF signaling and limb initiation

The first step in the development of the vertebrate limb is the initiation of a site in the flank of the embryo where the presumptive limb will grow. Foil barrier and tissue transplant studies of early chick limb development have suggested the existence of a diffusible substance that regulates the initiation of limb outgrowth (76). Subsequent studies, placing an FGF soaked bead on the flank of the embryo, demonstrated that FGFs could induce the formation of a limb, suggesting that FGFs may be the diffusible element responsible for initiating the limb field (77-79). Several different FGFs, including FGF 2, 4, 8 and 10, applied to the flank of the embryo can initiate the pathway(s) leading to limb development (77-80). This raises the question as to which FGF(s) functions in vivo to initiate the site of limb formation. Recent studies of chick limb development showed that FGF 8 and FGF 10 are expressed temporally and spatially in a manner consistent with their function as the endogenous initiators of limb development (79). FGF 10 is localized to the lateral plate mesoderm at the site of the presumptive limb bud in chick (79) and mouse (26). The expression of FGF 10 precedes that of FGF 8 suggesting that FGF 10 may be the primary initiator of the limb bud.

3.2. FGF signaling and limb growth

FGFs also perform essential functions during the progressive outgrowth and patterning of the limb bud. The limb bud is a specialized structure consisting of three functionally and spatially defined domains (figure 4). The apical ectodermal ridge (AER) is a specialized thickening of epithelium at the tip of the growing limb. It functions to stimulate elongation of the limb and maintain signals required for patterning the limb. FGFs 2, 4, and 8 are expressed in the AER (81-84). The progress zone (PZ) is a domain of undifferentiated mesenchymal cells, expressing FGF 10, that lie beneath the AER. The proliferative cues elaborated by the AER stimulate the PZ and maintain the cells in an undifferentiated state. The zone of polarizing activity (ZPA) is a region of posterior mesenchyme, adjacent to the AER that functions as a molecular compass; providing spatial cues to orient the growing limb with respect to the anterior-posterior axis, resulting in the stereotyped anatomy of the limb. These three functional domains operate coordinately and interdependently. Epithelial-mesenchymal signaling, mediated by FGFs and their receptors, is essential for establishing and maintaining these specialized domains.

Figure 4. Limb bud Development A, Representation of early limb development following the formation of the apical ectodermal ridge (AER), zone of polarizing activity (ZPA), and progress zone (PZ). M, limb bud mesenchyme and Ec, surface ectoderm. B, Model for the initiation of the limb bud. Ec, surface ectoderm where FGF receptor 2 IIIb (FGFR 2b) and FGF 8 are expressed. M, mesenchyme wherein FGF 10 and FGF receptor 2 IIIc (FGFR 2c) are expressed.

FGF 10 is expressed in the lateral plate mesoderm and later in the limb mesenchyme and acts to initiate and maintain FGF 8 expression in the AER. Reciprocally, FGF 8 maintains expression of FGF 10 in the underlying mesenchyme (79). The proliferation of cells within the PZ and the subsequent outgrowth of the limb results from the actions of FGFs produced in the AER. This has been demonstrated by experiments showing that when the AER is excised from the growing limb, the limb becomes truncated in proportion to the developmental stage that the AER is removed (85-87) and that growth can be restored by replacing the AER with a source of FGF placed at the tip of the growth arrested limb (78, 88, 89). Establishment of the ZPA and the interdependence of the ZPA and AER also requires an FGF signal. Sonic hedgehog (SHH) is the molecular determinant of the ZPA that induces signaling cascades that control the anterior-posterior spatial orientation of the limb (90 , 91). Initiation of SHH expression is thought to require FGF. Removal of the posterior AER, and thus FGF 4 and FGF 8 in sites overlying the ZPA, results in a loss of SHH expression (92). Conversely, ectopic expression of SHH expands the expression of FGF 4 (92, 93). Thus, feedback signaling pathways exist between FGF4 in the AER and SHH in the ZPA. These signals maintain the AER and ZPA and are thus essential for both limb elongation and patterning.

3.3. FGF receptors and limb development

FGF signals in the growing limb are mediated by high affinity FGF receptors. The specific receptors and their splice variants expressed in the developing limb match the ligand binding specificity of the receptor with the ligand expected to act at that site. For example, FGF 10 is expressed in the mesenchyme underlying the AER and is proposed to maintain and initiate formation of the AER (79) (figure 4B). In support of this, the AER expresses the IIIb splice form of FGF receptor 2 (37), a receptor that can be efficiently activated by FGF 10 (94). FGF receptor 2 signaling is absolutely required for the initiation of the limb as evidenced by studies in which a presumptive null allele was introduced into the mouse Fgfr 2 gene by homologous recombination (26). In mice lacking normal FGF receptor 2 activity, development of the limb bud is completely abolished. This leads to a model for initiation of the limb, whereby FGF 10 induces expression of FGF 8 in the AER through its interaction with FGF receptor 2 IIIb. FGF 8, in turn, maintains FGF 10 expression in the mesenchyme through its interaction with FGF receptor 2 IIIc (figure 4B). Limb bud mesenchyme also expresses FGF receptor 1 (95). However, in vitro FGF 10 activates FGF receptor 1 IIIb poorly and FGF receptor 1 IIIc not at all (22). Therefore, FGF receptor 1 does not appear to be the target of FGF 10. Targeted inactivation the IIIc isoform of FGF receptor 1 is lethal during early embryogenesis (96). In contrast, using the same targeting strategy to delete the IIIb splice form yields a mouse that is viable and fertile (96), indicating that the IIIc isoform is required for most of the functions of FGF receptor 1 and is presumably the splice form present in the limb mesenchyme. Using this guideline, FGF 4 and 8 expressed in the AER act on the underlying mesenchyme to activate FGF receptor 1 IIIc and FGF receptor 2 IIIc, thereby promoting the elongation of the limb.

The significance of FGF receptor signaling in the developing limb is accentuated by the finding that certain human skeletal disorders with abnormalities of the digits result from mutations in FGF receptors. Apert syndrome and Jackson-Weiss syndrome result from mutations in FGF receptor 2 (58, 71). These syndromes are characterized by premature fusion of the cranial sutures and syndactyly of the hand and feet. Pfeiffer syndrome, which is characterized by broad great toes and thumbs in addition to premature fusion of cranial sutures, results from a single mutation in FGF receptor 1 or one of several mutations in FGF receptor 2 (60, 68, 69). Jackson-Weiss and Pfeiffer syndrome can result from several different mutations in FGF receptor 2 (74), including mutations of Cys342, a conserved residue in the third Ig-like fold of the receptor (figure 3). This residue is believed to normally be bound in a disulfide linkage that stabilizes the Ig-like domain. Substitution of other amino acids at this site (as in Jackson-Weiss and Pfeiffer syndromes) is then presumed to leave the opposing cysteine residue available for the formation of inter- rather than intramolecular disulfide bonds. Indeed, studies in Xenopus showed that mutation of this cysteine residue results in disulfide linked receptor homodimers and ligand-independent receptor signaling (97). Thus, these results indicate that some, and perhaps all, of the mutations causing Jackson-Weiss and Pfeiffer syndrome are the result of activating mutations in an FGF receptor.

Interestingly, the limb phenotypes of JWS and PS are restricted to the distal most portion of the limb. This implies that the outgrowth and patterning of the more proximal portions of the limb are insensitive to constitutive FGF receptor signaling and suggests that these portions of the limb are constructed under conditions of excess ligand. If so, a constitutively active receptor would be functionally equivalent to the normal conditions of proximal limb development where the receptor is fully activated by a saturating concentration of ligand. In contrast, the distal most limb appears to be more dependent on the degree of FGF receptor signaling. Here, the receptor may be functioning under conditions where the available ligand is limiting. Additional FGF receptor signaling supplied by an activated receptor would be expected to disrupt the normal growth and patterning of the digits. Recent fate mapping studies in chick limb development show that the digits are formed by cell populations originating in the mesenchyme arising just below the AER (98). A portion of these cells migrate to positions anterior to their site of origin, suggesting that altered FGF receptor signaling may perturb the migration of cells destined to form the digits. Interestingly, we and others have found that the S252W mutation of the Apert syndrome alters the ligand binding affinity of the receptor (75, 99). The increased affinity for ligand may result in enhanced activation of the receptor in mesenchymal cells destined to form the digits. This may lead to abnormal migration of cells toward a source of ligand which may result in aberrant mesenchymal condensations and, consequently, abnormal digits.