Answers are grounded in this paper and cite their sources.
Bone is a highly complex tissue. Recent studies of humans and mice with various skeletal abnormalities involving many important genes are providing an understanding of some of the molecular mechanisms that control the activity, differentiation, and maintenance of bone.1 Much effort has been put into analyzing the transforming growth factor-β (TGF-β) superfamily of secreted signaling molecules.2-5 The papers of Opperman et al. (this issue)6 and Roth et al. (this issue)7 together with some of their earlier work8-10 have extended this research to cranial sutures in an attempt to understand sutural development, maintenance of suture patency, normal suture closure, and craniosynostosis. A suture is a craniofacial articulation in which contiguous margins of bone approximate each other, being united by a thin layer of fibrous tissue. During development, a presumptive suture becomes a definitive suture with interdigitations. Patent sutures—both cranial and facial—eventually close with bony bridging11 (Table 1). In this editorial review, both TGF-βs and fibroblast growth factors (FGFs), together with their receptors (Figs. FIG. 1, FIG. 2 and 2), are contrasted because both are expressed at the osteogenic fronts of developing calvarial sutures (as well as at other anatomical locations) and, under certain conditions, can play a role in early pathological closure (craniosynostosis). Signaling serine/threonine kinase receptors for TGF-β superfamily members. It is likely that both types of receptors are required for normal signaling, probably forming a heteromeric complex. Vertical bars = variable number of extracellular cysteines; C = cysteine box; TM = transmembrane region; GS = series of tandem glycine/serine residues; STK = serine/threonine kinase domain. Most mutations for major craniosynostosis syndromes are on FGFR2 with only one known mutation on FGFR1. Note mutations for short limb skeletal dysplasias (achondroplasia, hypochondroplasia, thanatophoric dysplasia, type I, and thanatophoric dysplasia, type II) on FGFR3. Fibroblast growth factor receptors 1, 2, 3. Hatched square = signal peptide; solid oblong = acid box; solid oval = CAM homology domain; open square = transmembrane domain; dotted oblongs = kinase domains 1 and 2. Three loops from left to right are immunoglobulin-like domains (IgI, IgII, IgIII). IIIc is alternatively spliced form of second half of IgIII. From Cohen.13 TGF-βs are potent growth regulatory molecules involved in the proliferation and differentiation of mesenchymal precursor cells, chondrocytes, osteoblasts, and osteoclasts; regulation of chondrogenesis and osteogenesis; and induction of extracellular matrix proteins. TGF-βs together with insulin-like growth factors (IGFs) are the most abundant growth factors stored in bone. With respect to TGF-βs per se, cross-talk among bone cells is partially understood.1-5 Although a role for TGF-βs in sutural biology is evident, no etiologic factor resulting in craniosynostosis has been identified. Fibroblast growth factor receptors (FGFRs) have an alternatively spliced loop in the second half of the third immunoglobulin-like domain (IgIII). The spliced form IgIIIc (Fig. 2) is the most abundant and is expressed primarily in mesenchymal tissue. Transcripts of FGFR2 with spliced loop IgIIIc may be expressed in the interdigital mesenchyme of the limbs and initially in the region of the future calvarial bones and then down-regulated except at the periphery of the calvarial bones (presumptive sutures).12 Many specific etiologic mutations on FGFRs are known for craniosynostosis syndromes (often with various limb defects) (Table 2), but little is known about pathogenesis.13, 14 The TGF-β superfamily consists of more than two dozen polypeptides that mediate many key events in normal growth and development, including cell differentiation, motility, organization, and death. Some molecules of this family can be grouped into distinct subfamilies of highly related sequences. The term TGF-β is somewhat misleading because many effects can be produced in different systems. TGF-βs can inhibit proliferation of different cell lines, and “transforming” activity may result from the secondary effects on matrix production and synthesis of other growth factors. The wide range of activities includes the basic body plan during early embryogenesis in Drosophilia, Xenopus, and mammals; formation of cartilage, bone, and sex organs; suppression of epithelial cell growth; a role in wound repair; and regulation of important immune and endocrine functions.2, 3 TGF-β is synthesized as an inactive precursor consisting of a latency-associated protein (LAP) and a cysteine-rich latent TGF-β binding protein (LTBP). After cleavage, LAP remains noncovalently bound to that part of TGF-β encoding the mature polypeptide. When present, LTBP is bound to the LAP/TGF-β complex by disulfide bonds, but it cannot bind TGF-β once TGF-β becomes activated. Furthermore, LTBP is not found in inactive TGF-β complexes in murine bone cultures, suggesting that LTBP is not essential for latency. The absence of LTBP may help segregate secreted TGF-β into free pools available for activation, LTBP-bound TGF-β being reserved for bone matrix storage.4 All TGF-β superfamily members show sequence similarity to the prototype TGF-β1, especially the active domain in which the most conserved feature is the spacing of seven cysteines. The three-dimensional structure of TGF-β2 includes six cysteines and four additional amino acids that form a cystine knot; four segments of antiparallel β-sheet; one long α-helix; and two short α-helices. The monomers dimerize, and a 7th cysteine forms an interchain disulfide bond.3 Recently, two transmembrane receptors for TGF-β have been isolated (Fig. 1). TβR-I (sometimes called type I) has a short extracellular region containing seven cysteines, a nine amino acid cysteine box, a transmembrane region, and an intracellular region composed of a short series of tandem serine/glycine residues and other conservative amino acids, and a serine/threonine kinase domain. TβR-II (sometimes called type II) also has a short extracellular region containing six to nine cysteines, a nine amino acid cysteine box, a transmembrane region, and an intracellular serine/threonine kinase domain. A variety of experiments suggest that TβR-I requires TβR-II for ligand binding, and both receptors are required for signaling, probably forming a heteromeric complex.2, 3 TGF-βs also bind with lower affinity to betaglycans (initially described as Type III TGF-β receptors). Binding to cell surface betaglycans may involve TGF-β storage, activation, or binding to TβR-I and TβR-II. Betaglycan core proteins have at least two subgroups; some permit cell membrane anchoring; others are found in the extracellular compartments.4 From a skeletal perspective, it is important to recognize that TGF-β superfamily members include TGF-β isoforms (TGF-β1, TGF-β2, TGF-β3), bone morphogenetic proteins (BMPs), and activin-A. These are expressed in bone cells, stored in bone matrix, and increase bone cell activity both in vivo and in vitro.4 Centrella et al.5 have addressed the problems of “dialogue” among osteoblast-like cells, osteoblast and osteoblast precursors, and osteoblast and marrow-derived hemopoietic cells. Opperman et al.,6, 10 Roth et al.,7, 9 and Levine et al.8 have extended bone-related research activities to sutural biology. To date, only one mutation has been described in a TGF-β superfamily member. The Hunter-Thompson type of chondrodysplasia15 is attributable to a mutation in cartilage-derived morphogenetic protein (CDMP-1), but craniosynostosis is not a feature. Furthermore, no receptor (TβR-I or TβR-II) mutations are known at the present time. Opperman et al.6 and Roth et al.7 in these papers (this issue) and in their earlier work(8-10 implicate a TGF-β dural signaling role for maintaining suture patency. This raises an interesting question: If TGF-β signaling from the dura is required for cranial suture patency, how is the patency of facial sutures maintained without dura? Facial sutures differ from cranial sutures in several other respects besides the presence or absence of dura. First, early synostosis of facial sutures is uncommon compared with premature synostosis of calvarial sutures. Second, developing facial bones have fibrous periosteal capsules surrounding them that are fully established by week 17 in utero. In contrast, cranial bones develop in a preformed continuous fibrous membrane, the ectomeninx, and do not develop capsules until after birth. Third, facial sutures11 Except the midpalatal suture. do not fuse before the seventh to eighth decades. In contrast, cranial sutures fuse relatively early in adult life15-20 (Table 1). It is postulated that continued patency of the circum-maxillary sutures may have a shock-absorbing function related to the muscles of mastication, but weasels produce powerful masticatory shocks and have fused facial sutures. Similarly, peccaries with their stressful rooting and feeding activities have early fusion of facial sutures.15 The answer may hinge on the ubiquitousness of TGF-βs. Opperman et al. indicate that TGF-β3 transcripts are present within cranial sutures. Are transcripts of all three TGF-β isoforms expressed within facial sutures? If so, this needs to be demonstrated. Further research should also begin to address the differential timing of normal suture closure. FGFRs are a subset of receptor tyrosine kinases, each composed of an extracellular ligand-binding domain, connecting a transmembrane region, and an intracellular kinase domain, responsible for receptor dimerization and cell signaling. Kinase activation takes place when an FGF ligand forms a trimolecular complex with an FGFR and heparan sulfate proteoglycans.21-26 To date, there are four known FGFRs and nine FGFs. Among the FGFRs, mutations for various craniosynostosis syndromes are associated with FGFR1, FGFR2, and FGFR3, but mostly with FGFR227-53, 56, 57 (Table 2 and Fig. 2).22 Mutations for achondroplasia,66 hypochondroplasia,67 and thanatophoric dysplasia49—three short-limb skeletal dysplasias—are on FGFR3. A few craniosynostosis syndromes map to non-FGFR chromosomes54,55,58-65 (Table 2 Figure 2 shows that mutations have been identified in the linker region between IgII and IgIII on FGFR1, FGFR2, and FGFR3. The two amino acid substitutions on FGFR2 that result in Apert syndrome are both bulky mutations (Ser252Trp and Pro253Arg). The analogous Pro252Arg change, resulting in some cases of Pfeiffer syndrome, is located in the same linker region on FGFR1.43, 44 FGFR3 associated coronal synostosis syndrome(56, 57) and thanatophoric dysplasia, type I,49 are associated with Pro250Arg and Arg248Cys substitutions, respectively, in the linker region between IgII and IgIII on FGFR3. These amino acid substitutions in the linker region between IgII and IgIII occur within the putative ligand-binding domain that is highly conserved among species and between different human FGFRs. Wilkie et al.27 proposed that such substitutions may alter the relative orientation of IgII and IgIII, affecting the binding of FGF ligands or dimerization with other FGFRs. Other possible effects of these amino acid substitutions include higher or lower affinity for certain ligands, altered kinase activity, altered receptor turnover, or autonomous receptor activation.68 Mutations for Crouzon syndrome frequently involve loss or gain of a cysteine residue (Table 2). Such changes may alter the structure of the IgIII loop, which is maintained by a disulfide bond between two cysteine residues.34 Wilkie et al.69 suggested that the creation of free, exposed cysteine residues by IgIII mutations might lead to receptor dimerization by intermolecular disulfide bonding, resulting in ligand-independent activation. One common mutation, 1211G → A, does not result in an amino acid substitution (Ala344Ala). However, the mutation produces a cryptic donor splice site, causing a deletion of 17 amino acids and shortening the distance from the disulfide bond of IgIIIc to the transmembrane domain.42 Mutations on FGFR2, resulting in Pfeiffer syndrome, cluster in one of two locations—the acceptor splice site for exon 9 or a cysteine residue essential for forming the disulfide-bonded loop of IgIII (Table 2). Splice site mutations may disrupt the regular splicing mechanism of exon 9, and an absent cysteine from position 342 may alter the IgIII loop. Both types of mutations may affect ligand binding or dimerization.45 Stop codon FGFR3 mutations in thanatophoric dysplasia, type I (Table 2), result in a receptor elongated by 141 amino acids, producing a hydrophobic domain in the C-tail. Such a hydrophobic region could act as a second transmembrane domain and affect either dimerization or kinase activity.52 Heterogeneity has been demonstrated for craniosynostosis by genes on different chromosomes (8, 10, and 4 for FGFR1, FGFR2, and FGFR3, respectively; 5 and 7 for other craniosynostosis syndromes)29, 33a, 33b, 43-45, 54, 55, 58-65 (Table 2). There is also heterogeneity shown by mutations in different genes for the same syndrome (FGFR1 and FGFR2 for Pfeiffer syndrome)30, 44, 45 and by different amino acid substitutions in the same gene for the same syndrome (e.g., Cys342Ser and Tyr328Cys from exon 9 and Cys278Phe and Gln289Pro from exon 7 in FGFR2, all resulting in Crouzon syndrome)29, 33c, 34, 36, 37, 39 (Table 2). Furthermore, it is clear that causal homogeneity may lead to phenotypic heterogeneity; a mutation such as Cys342Arg in exon 9 of FGFR2 can result in either Crouzon syndrome or Pfeiffer syndrome,29, 30, 34, 45 although both breed true within families. With respect to Crouzon and Pfeiffer syndromes (Table 2 and Fig. 2), there are many more mutations than there are developmental pathways for craniosynostotic syndrome phenotypes. Thus, heterogeneous mutations for the same phenotype must be canalized during development. An earlier concept of a single mutation resulting in a single disease is outmoded. However, the same mutation producing different phenotypes that breed true, such as Cys342Arg, which may result in either Crouzon syndrome or Pfeiffer syndrome, suggests the possibility of interaction with other gene products or of other alterations within the same gene. Mutations identified on FGFRs may be located extracellularly, affecting ligand binding, or intracellularly, affecting activation of the signaling pathway. Almost all nucleotide changes are of the missense or splice-site types. Occasionally, insertions39 and deletions37 are reported (Table 2). Jabs et al.55 were the first to delineate a craniosynostosis syndrome at the molecular level. The locus for craniosynostosis, Boston type, was mapped to 5qter.54 Then Jabs et al.55 detected a C → A transversion at nucleotide 64 of exon 2, resulting in the substitution of a histidine for a highly conserved proline at amino acid position 7 of the homeodomain of the MSX2 gene. Jabs et al.55 also studied the distribution of Msx2 transcripts in the calvaria of neonatal mice and localized its expression to the osteogenic front of calvarial sutures. Lui et al.70 engineered a Pro7His mutation in the mouse Msx2 gene and studied its expression in the developing calvaria of transgenic mice. They found premature suture closure and ectopic cranial bone formation, concluding that: (1) Msx2Pro7His mutation causes craniosynostosis of the Boston type, (2) ectopic bone formation indicates that the mutant gene can initiate a bone formation program, and (3) Msx2Pro7His may be a gain-of-function mutation that enhances normal activity of the Msx2 gene product.33 There are several interesting animal mutations. Carlton and Evans71 reported a Crouzon-like mutant (Bey) in the mouse; the cause is an insertional mutation in the Fgf4 region. Two mouse mutations are known in the TGF-β superfamily but craniosynostosis is not a feature; short ear (se) results from a mutation in bone morphogenetic protein-5 (Bmp-5)72 and brachypodism (bp)73 is caused by a mutation in growth differentiation factor-5 (Gdf-5). Simple craniosynostosis—by far more common than syndromic craniosynostosis—remains a big mystery. Both its etiology and pathogenesis are unknown. With respect to TGF-βs per se, cross-talk among bone cells is partially understood.(5) Roth et al.,7, 9 Levine et al.,8 and Opperman et al.6, 10 have correlated TGF-βs isoforms with their differential immunoreactivity patterns in developing and fusing cranial sutures, with special attention paid to the underlying dura. Although TGF-βs play a role in normal suture formation, maintenance, and closure, the mechanisms by which they exert their effects at suture sites is not known. They may possibly play a role in premature craniosynostosis. It is not clear whether these autocrine or paracrine growth factors play a primary or secondary role in synostosis. Can such growth factors initiate synostosis sui generis under special circumstances, such as a growth factor mutation per se or in one of their receptors, or are they simply part of the complex chain of events resulting in bone formation that is accelerated because of some unknown initiating factor earlier in the chain? I favor the latter interpretation, but further research is necessary. Molecular etiologies of several common craniosynostosis syndromes have been delineated (Table 2), but pathogenesis is poorly understood in each case. FGFR mutations (Fig. 2) cannot be loss-of-function mutations because chromosomal deletions [del(8p), del(10q), del(4p)] that result in multiple anomalies do not include craniosynostosis.44 or a short limb skeletal dysplasia, which is found with a number of mutations on FGFR3.14, 49, 66, 67 Thus, such mutations (Table 2) are of the altered or gain-of-function type. All human mutations on FGFRs known to date allow human development to proceed normally but interfere with later development, particularly bone. Much work remains to be done in unraveling the pathogenesis of these phenotypes. The creation of transgenic mouse models for FGFR mutations and their analysis during development should aid in this understanding.
No takes yet. Share an insight, caveat, or question.
Maimon M. Cohen (1997) studied this question.
Synapse has enriched 10 related papers. Consider them for comparative context: