| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
Department of Endocrinology and Medicine (Z.V.), Veterans Affairs Medical Center, Phoenix, Arizona 85012; and Medical Genetics Branch (Z.V., C.A.F.), National Human Genome Research Institute and Craniofacial and Skeletal Diseases Branch (D.J.W.), National Institute of Dental and Craniofacial Research, National Institutes of Health, Bethesda, Maryland, 20892
 |
Abstract |
|---|
 Top Abstract I. IntroductionII. Fibroblast Growth Factor...III. Clinical and Molecular...IV. Biochemical Analysis of...V. GH TreatmentVI. ImplicationsReferences |
|---|
Since the Gly380Arg achondroplasia mutation was recognized, similar observations regarding the conserved nature of FGFR mutations and resulting phenotype have been made regarding other skeletal phenotypes, including hypochondroplasia, thanatophoric dysplasia, and Muenke coronal craniosynostosis. These specific genotype-phenotype correlations in the FGFR disorders seem to be unprecedented in the study of human disease. The explanation for this high degree of mutability at specific bases remains an intriguing question.
|
I. Introduction |
|---|
|
TopAbstractI. IntroductionII. Fibroblast Growth Factor...III. Clinical and Molecular...IV. Biochemical Analysis of...V. GH TreatmentVI. ImplicationsReferences |
|---|
), the most common form of human dwarfism (1, 2). The achondroplasia
family of skeletal dysplasias, as described by Spranger (3), also
includes the mildly severe hypochondroplasia (Fig. 2) and the lethal thanatophoric dysplasia (TD) (Fig. 3). Recently, SADDAN (severe achondroplasia with developmental delay and
acanthosis nigricans) dysplasia (Fig. 4), a skeletal dysplasia with features of both achondroplasia and TD, has
been added to this family of disorders (4). These other disorders in
the achondroplasia family also result from mutations in the FGFR3 gene
(4, 5, 6, 7, 8, 9, 10, 11, 12). In individuals with achondroplasia the skeleton is the primary
system involved in the phenotype, and all of the disorders in the
achondroplasia family of skeletal dysplasias involve some degree of
short stature and/or abnormal ossification of bony structures.
|
|
|
|
) (15, 16, 17) and Crouzon syndrome with acanthosis nigricans (Fig. 6) (18). In general, the relationship between mutations in the FGFR3 gene
and other FGFR genes, and the phenotypes that result from these
mutations, have broken new ground in the understanding of human
mutations and genetic disorders. In the FGFR genes, more than any
other, there is a highly conserved relationship between mutations at
particular amino acids and resulting phenotypes (1, 2, 5, 6, 15, 17, 18, 19, 20). Moreover, the FGFR3 nucleotides mutated in the majority of
cases of achondroplasia and Muenke craniosynostosis are among the most
highly mutable nucleotides in the human genome.
|
|
|
II. Fibroblast Growth Factor Receptor 3 (FGFR3) |
|---|
|
TopAbstractI. IntroductionII. Fibroblast Growth Factor...III. Clinical and Molecular...IV. Biochemical Analysis of...V. GH TreatmentVI. ImplicationsReferences |
|---|
The FGFR3 gene maps to human chromosome 4p16.3 (53). The cDNA was originally isolated in the search for the Huntington disease gene on chromosome 4 (54, 55). The 4.4-kb cDNA contains an open reading frame of 2,520 nucleotides, encoding an 840-residue protein. The human and mouse FGFR3 genes have recently been characterized (56, 57, 58) and span approximately 16.5 kb and 15 kb, respectively. Both genes consist of 19 exons and 18 introns. In both genes, the translation initiation and termination sites are located in exons 2 and 19, respectively. The 5'-flanking regions lack typical TATA and CAAT boxes. However several putative cis-acting elements are present in the promoter region, which is contained within a CpG island (57, 58). The promoter regions of both the human and mouse FGFR3 genes are very similar, with several conserved putative transcription factor-binding sites, suggesting an important role for these elements and their corresponding transcription factors in the transcriptional regulation of FGFR3 (58). It has been demonstrated that the 100 bp of FGFR3 sequence 5' to the initiation site are sufficient to confer a 20- to 40-fold increase in transcriptional activity (59). FGFR3 sequences between -220 and +609 are sufficient to promote tissue-specific expression (59).
Proteins in the family of fibroblast growth factor receptors (FGFRs)
have a highly conserved structure (Fig. 7). The mature FGFR3 protein, like all of the FGFRs, is a
membrane-spanning tyrosine kinase receptor with an extracellular
ligand-binding domain consisting of three immunoglobulin subdomains, a
transmembrane domain, and a split intracellular tyrosine kinase domain
(Fig. 7) (60). Ligand binding requires dimerization of two monomeric
FGFRs and includes a heparin-binding step. Promiscuous dimerization is
observed; for example, in addition to dimerizing with itself, FGFR1 may
dimerize with FGFR2, FGFR3, or FGFR4. Similar dimerization combinations
of other FGFR monomers are also possible. Differing combinations of
dimers are observed in different tissues and different stages of
development, and this diversity of dimers probably plays an important
role in skeletal differentiation (60).
|
|
III. Clinical and Molecular Studies |
|---|
|
TopAbstractI. IntroductionII. Fibroblast Growth Factor...III. Clinical and Molecular...IV. Biochemical Analysis of...V. GH TreatmentVI. ImplicationsReferences |
|---|
The achondroplasia family, as described by Spranger (3), is characterized by a continuum of severity ranging from mild (hypochondroplasia) and more severe forms (achondroplasia) to lethal neonatal dwarfism (TD). The identification of FGFR3 mutations in each of the disorders in the "achondroplasia family" of skeletal dysplasias, as well as COL2A1 mutations in the "type II collagenopathies" (67), fortified Dr. Spranger’s remarkable power of clinical observation.
Achondroplasia and TD type II (see below) both appear to be genetically homogeneous (and, most of the time, homoallelic) conditions in that they are caused by a single nucleotide substitution in more than 95% of cases (1, 2, 5, 7, 10). Interestingly, the opposite situation was observed in association with mutations with other FGFR-related defects. In the craniosynostosis syndromes caused by mutations in FGFR1, FGFR2, or FGFR3, similar mutations, but in different receptors, have been found to cause distinct phenotypes: FGFR1 Pro252Arg results in Pfeiffer syndrome; FGFR2 Pro253Arg results in Apert syndrome; and FGFR3 Pro250Arg causes Muenke craniosynostosis (15). FGFR2 mutations are also associated with Crouzon, Pfeiffer, and Jackson-Weiss syndromes (19, 20, 68); interestingly, all three phenotypes can be caused by a FGFR2 Cys342Arg mutation.
1. Achondroplasia. Achondroplasia, the most common cause of
dwarfism in man, occurs in approximately between 1 in 15,000 and 1 in
40,000 live births. It is an autosomal dominant disorder with complete
penetrance, characterized by short-limbed dwarfism, macrocephaly,
depressed nasal bridge, frontal bossing, and trident hands (Fig. 1
)
(69, 70). X-rays show a shortening of long bones with squared-off iliac
wings, a narrow sacrosciatic notch, and distal reduction of the
vertebral interpedicular distance (Fig. 8) (69, 70). Physical and radiographic findings of the disorder are
remarkably consistent. Histopathology demonstrates a defect in the
maturation of the cartilage growth plate of long bones. More than 90%
of the cases are sporadic, and there is an increased paternal age at
the time of conception of the affected individuals, suggesting that the
de novo mutations are of paternal origin. Affected
individuals are fertile and achondroplasia is transmitted as a fully
penetrant autosomal dominant trait (21, 71). In contrast, homozygous
achondroplasia is usually lethal in the neonatal period and affects
25% of the offspring of matings between two parents with heterozygous
achondroplasia (72).
|
The first reports of mutations in FGFR3 causing achondroplasia (1, 2)
indicated that 37 of 39 mutations studied were exactly the same, a
G-to-A transition at nucleotide 1138 (G1138A). The remaining two
mutations were a G-to-C transversion at the same nucleotide (G1138C).
Both mutations result in the substitution of arginine for the glycine
residue at position 380 (Gly380Arg) in the transmembrane domain of the
protein (Figs. 7 and 9). Most analyses were performed on heterozygous achondroplasia patients,
but the Gly380Arg mutation was also detected in several cases of
homozygous achondroplasia, in which both parents of the proband had
achondroplasia. In 1995, Bellus et al. (74) confirmed the
remarkable degree of genetic homogeneity of the disorder by finding the
Gly380Arg mutation in 153 of 154 achondroplastic alleles. In this
series, the G-to-A transition accounted for 150 alleles, while the
G-to-C transversion was found in 3. [The last patient was later
rediagnosed as having SADDAN dysplasia, based on phenotypic findings
much more severe than those found in typical achondroplasia (see
below). Therefore Bellus et al. (74) found FGFR3 mutations
in 100% of their cohort, with the two achondroplasia mutations
observed in all 153 of their patients with true achondroplasia.] Thus,
the vast majority of cases of achondroplasia are caused by the same
Gly380Arg mutation. Exceptions include two cases, reported by
Superti-Furga et al. (75) and Nishimura et al.
(76), in which a Gly375Cys mutation was detected five amino acids away
from the common codon 380 mutation, and an achondroplasia patient with
a novel Gly346Glu mutation identified by Prinos et al. (77).
|
2. Hypochondroplasia. The findings in patients with
achondroplasia prompted the search for FGFR3 mutations in other
disorders considered related to achondroplasia. Hypochondroplasia (Fig. 2) is an autosomal dominant condition characterized by short stature,
micromelia, and lumbar lordosis. Clinical symptoms, radiological
features, and histopathological aspects are similar to, but milder than
those seen in achondroplasia (85, 86). Many cases are first referred
for endocrinological evaluation of short stature.
McKusick et al. (87) first proposed that achondroplasia and hypochondroplasia are allelic, based on the similarities in phenotype between the two disorders and the identification of a severely dwarfed patient whose father had achondroplasia and whose mother had hypochondroplasia. More than two decades later, molecular linkage studies supported allelism of achondroplasia and hypochondroplasia (14, 88). Subsequently, heterozygous FGFR3 mutations were detected in DNA from persons with hypochondroplasia: C-to-A or C-to-G transitions at nucleotide 1620 (C1620A, C1620G), resulting in an Asn540Lys substitution in the proximal tyrosine kinase domain (6). These observations have since been confirmed in several laboratories (8, 89, 90, 91). In 1996, Prinster et al. (92) also found the C-to-A and C-to-G changes at nucleotide 1620 in Italian hypochondroplasia patients, and a novel FGFR3 Ile538Val that results in hypochondroplasia was also identified (93). However, studies of other families with hypochondroplasia have shown the phenotype to be unlinked to chromosome 4p16.3 (94, 95). In three familial cases not linked to chromosome 4, Rousseau et al. (89) reported that the phenotype was milder, macrocephaly and shortening of the bones were less obvious, the hands were normal, and no metaphyseal flaring was noted, as compared with hypochondroplasia probands, due to the FGFR3 Asn540Lys mutation. Prinster et al. (96) also described nine cases of hypochondroplasia not due to the FGFR3 Asn540Lys mutation. The authors stated that although they could not identify firm genotype-phenotype correlations, in their study the Asn540Lys mutation was most often associated with disproportionate short stature, macrocephaly, and with radiological findings of unchanged or narrow interpedicular distance and fibula longer than the tibia (96). This observation supports the view that unlike achondroplasia, hypochondroplasia is a clinically and genetically heterogeneous condition (85, 86, 95, 96, 97).
3. TD. TD (Fig. 3) is one of the more common sporadic lethal
skeletal dysplasia, affecting approximately 1 of 60,000 births. The
features include micromelic shortening of the limbs, macrocephaly,
platyspondyly, and reduced thoracic cavity (98, 99). In the most common
subtype (type I, TD I), femurs are curved, while in type II (TD II),
straight femurs are present and cloverleaf skull may also be a feature
of the phenotype. Interestingly, mutational studies have confirmed the
classification of TD into these two subtypes (5). Affected individuals
usually die in the neonatal period. However, a limited number of cases
with prolonged survival have been reported (100, 101).
Mutations in the FGFR3 gene have been identified in both types of TD (5, 7, 9). Indeed, heterozygous mutations were found to cluster mainly to two different locations in the FGFR3 gene, depending on the phenotype. While TD II was accounted for by a single recurrent mutation in the tyrosine kinase 2 domain (Lys650Glu), TD I results from either a stop codon mutation or missense mutations in the extracellular domain of the gene (11). Interestingly, all missense mutations found so far created cysteine residues (9, 12).
In the first report of FGFR3 mutations in TD, Tavormina et al. (5) demonstrated a sporadic mutation causing a Lys650Glu change in the tyrosine kinase domain in 16 of 16 TD II patients. In the same study, the authors also report a mutation causing an Arg248Cys change in 22 of 39 TD I patients and a Ser371Cys mutation was found in one additional infant with TD I. Interestingly, the first 15 TD patients tested for the Lys650Glu mutation were not separated based on TD subtype. Of those 15, nine had the mutation. It was not until after the molecular analysis that the radiographs of the TD probands were reexamined and separated into subgroups based on straight or curved femurs. Nine patients had straight femurs, consistent with TD II. Those nine patients all had the Lys650Glu mutation. The remaining six had curved femurs, consistent with a TD I phenotype (5).
Subsequently, Rousseau et al. (7) reported mutations in the stop codon (stop807Gly, stop807Arg, and stop807Cys) in five additional patients with TD I. The latter mutations removed the normal translation stop signal and are predicted to result in a protein 141 amino acids longer than normal if translation continues to the next in-frame stop codon (7, 10). In 1996, Rousseau et al. (11) reported two novel missense mutations (Tyr373Cys and Gly370Cys), creating cysteine residues in the extracellular domain of the receptor in 9 of 26 TD I patients, giving further support to the view that newly created cysteine residues in the extracellular domain of the protein appear to play a key role in the severity of the disease (5, 7, 10, 11). Pokharel et al. (102), in late 1996, found the mutation most commonly reported in previous European and North American studies, the Arg248Cys substitution, in five of five Japanese TD I patients. The reported patients included cases with the usual presentation, and also, a case with a 9-yr follow up, representing an unusually mild clinical course for TD. This may suggest that, similarly to achondroplasia, TD I is a genetically homogenous condition (10, 102).
Histopathologically, cases with the Lys650Glu substitution demonstrated relatively more preservation of the physeal chondrocyte columns with identifiable proliferative and hypertrophic zones. The fibrous band was present only adjacent to the periosteum. In contrast, the fibrous band was more extensive and the column preservation poorer in cases with the Arg248Cys substitution (103). In this study, 91 cases of TD were examined for clinical, radiographic, and histological findings. Every case of TD examined had an identifiable FGFR3 mutation (103). Interestingly, radiographically, all of the cases with the Lys650Glu substitution demonstrated straight femurs with craniosynostosis and, frequently, a cloverleaf skull. In all other cases, the femurs were curved (103).
The platyspondylic lethal skeletal dysplasias (PLSDs) are a heterogeneous group of short-limb dwarfing conditions, with TD the most common form. Three other types of PLSD, or TD variants (San Diego, Torrance, and Luton), have been distinguished from TD. The most notable difference between TD and the variants is the presence of large rough endoplasmic reticulum inclusion bodies within chondrocytes of the variants. Brodie et al. (104) examined 22 cases of TD variants for the presence of missense mutations in the FGFR3 gene. All 17 cases examined of the San Diego type (PLSD-SD) were heterozygous for some of the same FGFR3 mutations that cause TD I. Of the 17 FGFR3 mutations identified, 7 were Arg248Cys mutations, 2 were Ser249Cys mutations, 6 were Tyr373Cys mutations, and 2 were stop codon mutations. No mutations were identified in the Torrance and Luton types. Large inclusion bodies were found in 14 cases of PLSD-SD, with the material retained within the rough endoplasmic reticulum staining with antibody to the FGFR3 protein. The authors speculate that the radiographic and morphological differences between TD and PLSD-SD may be due to other genetic factors (104).
4. SADDAN dysplasia. SADDAN dysplasia (Fig. 4) is a recently
described phenotype also belonging to the achondroplasia family of
skeletal dysplasias. SADDAN dysplasia was originally named SSB
dysplasia, for skeletal, skin, and brain dysplasia, as these are the
three systems predominantly affected in this condition (4, 105, 106).
SADDAN dysplasia is characterized by extreme short stature, severe
tibial bowing, profound developmental delay, and acanthosis nigricans
(4, 104). A novel mutation in the FGFR3 gene, A1949T (Lys650Met), has
been reported in three unrelated patients with SADDAN dysplasia (4, 107). These three patients have all survived past infancy, with two
patients now young adults, without the need for prolonged ventilatory
assistance. Individuals with the Lys650Met mutation have skeletal
findings distinct from both TD I and TD II. These findings included
absence of craniosynostosis or cloverleaf skull anomaly and moderate
bowing of the femurs with reverse bowing of the tibia and fibula.
Survival past infancy has led to the observation of phenotypic
manifestations that may not occur in surviving children with TD,
including development of acanthosis nigricans in the cervical and
flexural areas. Individuals with SADDAN dysplasia also had seizures and
hydrocephalus during infancy with severe limitation of motor and
intellectual development. The Lys650Met mutation has also been
identified in two patients with TD type I, (107, 108). Interestingly,
substitution of the identical amino acid residue by glutamic acid
(Lys650Glu) results in TD II.
B. Craniosynostosis disorders
FGFR3 mutations have also been identified in individuals with
disorders not in the achondroplasia family of skeletal dysplasias.
These include nonsyndromic craniosynostosis, recently referred to as
Muenke coronal craniosynostosis, and Crouzon syndrome with acanthosis
nigricans.
1. Muenke coronal craniosynostosis. Recently Bellus et
al. (15) identified a FGFR3 Pro250Arg amino acid substitution
caused by a C749G transversion in 10 unrelated patients with autosomal
dominant or sporadic cases of craniosynostosis (Fig. 5). This mutation
is in the region of the gene that encodes the extracellular domain of
the FGFR3 protein. The FGFR3 residue mutated in these individuals,
FGFR3 Pro250, corresponds to the exact residue in two other FGFR genes
in which mutations cause craniosynostosis syndromes (23, 24, 25, 26, 27, 28, 29, 30, 31, 32, 33, 34, 35, 36, 37, 38, 39, 40). FGFR1
Pro252Arg and FGFR2 Pro253Arg amino acid substitutions result in
Pfeiffer and Apert syndromes, respectively (23, 24, 25, 26, 27, 28, 29, 30, 31, 32, 33, 34, 35, 36, 37, 38, 39, 40).
Muenke et al. (17) provided extensive information on a series of 61 individuals from 20 unrelated families in which coronal craniosynostosis is due to the FGFR3 Pro250Arg mutation, defining a new clinical syndrome that might be referred to as Muenke coronal craniosynostosis (16). Considerable phenotypic variability is observed in individuals with this mutation. In addition to the craniosynostosis, some patients had radiographic abnormalities of their hands and feet, including thimble-like middle phalanges, coned epiphyses, and carpal and tarsal fusions. Brachydactyly was observed in some patients, as was sensorineural hearing loss. Developmental delay was observed in a minority of the patients. Reardon et al. (109) discussed the clinical manifestations in nine individuals with this mutation. Four of these individuals had mental retardation. Reardon et al. (109) suggested that there was a significant overlap between Saethre-Chotzen syndrome and the phenotype produced by this mutation. Saethre-Chotzen is caused by mutations in the TWIST gene (110), and patients originally diagnosed with Saethre-Chotzen in which an FGFR2 or FGFR3 mutation has been identified should be reclassified. Golla et al. (111) described a large German family with the Pro250Arg mutation in which there was also considerable phenotypic variability among individuals.
2. Crouzon syndrome with acanthosis nigricans. Crouzon
syndrome is characterized by cranial synostosis, hypertelorism,
exophthalmos and external strabismus, parrot-beaked nose, short upper
lip, hypoplastic maxilla, and a relative mandibular prognathism, and is
caused predominantly by mutations in the gene for FGFR2 (Fig. 6) (19, 23, 24, 26, 27, 28, 112). Recently, a FGFR3 Ala391Glu (G-to-A transition at
nucleotide 1172) substitution was identified in individuals with a
phenotype of Crouzon craniosynostosis in association with acanthosis
nigricans (18, 113). Meyers et al. (18) identified this
mutation in a mother and daughter and two sporadic cases with this
condition. This mutation is in the FGFR3 transmembrane domain, situated
close to the recurrent achondroplasia mutation. The patients had a
typical Crouzon syndrome phenotype. Skeletal survey showed no evidence
for the skeletal manifestations of achondroplasia, TD, or
hypochondroplasia, although they did have hydrocephalus, possibly
caused by stenosis of the jugular foramen (114), and some of the cases
had interpediculate narrowing (18).
The acanthosis nigricans in the patients with the FGFR3 Ala391Glu mutation was characterized by verrucous hyperplasia and hypertrophy of the skin with hyperpigmentation and accentuation of skin markings, distributed in a distinctive fashion including not only the axillae and neck, but also the chest, abdomen, breasts, perioral, and periorbital areas, and nasolabial folds (18). Meyers et al. (18) noted multiple melanocytic nevi over the face, trunk, and extremities of all four of their patients.
One of the patients with Crouzon syndrome with acanthosis nigricans due to the FGFR3 Ala391Glu mutation reported by Meyers et al. (18) has a second cousin with Crouzon syndrome. This individual does not have acanthosis nigricans. The phenotype in this patient is due to the FGFR2 Ser347Cys mutation (19).
|
IV. Biochemical Analysis of FGFR3 Mutations |
|---|
|
TopAbstractI. IntroductionII. Fibroblast Growth Factor...III. Clinical and Molecular...IV. Biochemical Analysis of...V. GH TreatmentVI. ImplicationsReferences |
|---|
). Cell surface-bound heparan sulfate proteoglycans are required to help
the ligand-receptor complex to form (115). Phosphorylation of the FGFR
tyrosine residues stimulates tyrosine kinase activity, possibly by
stabilizing the activation loop of the kinase in a conformation that
allows substrates and ATP to access the catalytic site (116, 117).
Furthermore, the phosphorylated tyrosine residues act as binding sites
for substrates containing Src homology or phosphotyrosine binding
domains, providing a means to recruit and phosphorylate other
molecules, furthering the FGFR signal transduction pathway.
|
Targeted disruption of the FGFR3 gene causes enhanced bone growth of long bones and vertebrae in mice, suggesting that FGFR3 negatively regulates bone growth (118, 119). Thus, FGFR3 mutations in the achondroplasia family of skeletal dysplasias can probably be interpreted as gain-of-function mutations that activate the fundamentally negative growth control exerted by the FGFR3 pathway (118, 120). The fact that the recessive loss-of-function mutation produces a phenotype in mice, which appears to be the opposite of those seen in achondroplasia, hypochondroplasia, or TD in humans, suggests that the human phenotype may result from a constitutive, or ligand-independent activation of the receptor (118).
Based on the current knowledge about signal transduction by the FGF pathway, activation of FGFRs normally occurs only after ligand binding (121). After studying Xenopus FGFRs, Neilson and Friesel (122) also found that different point mutations may activate FGFRs by distinct mechanisms, and that ligand-independent FGFR activation may be a feature skeletal dysplasias have in common.
Additional evidence for the gain-of-function hypothesis was provided by
Webster et al. (123), who recently demonstrated profound
constitutive activation of the FGFR3 tyrosine kinase (
100-fold above
the wild type) associated with the Lys650Glu mutation, which is known
to cause TD type II. The authors demonstrated a specificity for
position in FGFR3, as well as charge, in terms of amino acid changes
that result in altered kinase activation. The authors speculated that
the TD type II mutation in the FGFR3 activation loop mimicked the
conformational changes that activate the tyrosine kinase domain (123).
This activation is normally initiated by ligand binding and
autophosphorylation of the receptor. Using immunoprecipitation followed
by an in vitro kinase assay, Webster and Donoghue (124) also
found that the mutation in TD increased autophosphorylation activity of
the FGFR3 relative to the wild-type or achondroplasia mutant receptor.
Subsequently, Webster and Donoghue (124, 125) found similar constitutive FGFR3 activation associated with the Gly380Arg mutation, known to result in achondroplasia. Moreover, Naski et al. (126) demonstrated that the Gly380Arg, the Lys650Glu (TD II), and the Arg248Cys (TD I) mutations constitutively activate the receptor, as evidenced by ligand-independent receptor tyrosine phosphorylation and cell proliferation. Interestingly, but perhaps not surprisingly, the mutations that are responsible for TD activated the FGFR3 receptor more strongly than the mutations causing achondroplasia. It has further been demonstrated that the constitutive tyrosine kinase activity of FGFR3 containing the TD II mutation specifically activates the transcription factor Stat1 (signal transducer and activator of transcription) (127, 128). This mutant receptor also induced nuclear translocation of Stat1, induced expression of the cell-cycle inhibitor p21(WAF/CIP1), and resulted in growth arrest of the cell. Stat1 activation and increased p21(WAF/CIP1) expression was found in chondrocytes from a TD II fetus, but not in cells from a non-TD fetus. The authors suggest that in TD, Stat1 may be used as a mediator of growth retardation in bone development, and that abnormal STAT activation and p21(WAF/CIP1) expression due to the mutant FGFR3 receptor may be responsible for the resulting phenotype (127).
Naski et al. (129) examined the effects of an activated FGFR3 specifically targeted to growth plate cartilage in mice. The resulting mice were dwarfed, with axial, appendicular, and craniofacial skeletal hypoplasia (129). FGFR3 inhibited endochondral bone growth by disrupting chondrocyte proliferation and differentiation. The Indian hedgehog signaling pathway and bone morphogenic protein (Bmp) 4 expression were also down-regulated in growth plate chondrocytes from these mice, suggesting that FGFR3 is an upstream negative regulator of the hedgehog signaling pathway and that FGFR3 may coordinate the growth and differentiation of chondrocytes with the growth and differentiation of osteoprogenitor cells (129).
Wang et al. (130) and Li et al. (131) developed mouse models for achondroplasia. The mice are significant for their small size, including shortening of the long bones, especially the femur (130, 131). Also evident was a short craniofacial area, midface hypoplasia with protruding incisors, distorted skull with anteriorly shifted foramen magnum, and kyphosis (130, 131). Histological examination revealed narrowed and distorted growth plates in the long bones, vertebrae, and ribs of these mice, demonstrating that achondroplasia results from a gain of FGFR3 function, leading to inhibition of chondrocyte proliferation (130). Stat1, Stat5a, and Stat5b were activated by expression of the mutant receptor, and p16, p18, and p19 cell cycle inhibitors were up-regulated, also leading to inhibition of chondrocyte proliferation (131). Fewer maturing and hypertrophic chondrocytes were generated in the growth plates of these mutant mice, resulting in a "less-active" growth plate (131).
Thompson et al. (132) demonstrated that a chimera containing the transmembrane and intracellular domain of FGFR3 with the achondroplasia mutation fused to the extracellular domain of platelet-derived growth factor (PDGF), induces ligand-dependent differentiation of PC-12 cells. When stably transfected into PC12 cells, which contain no endogenous PDGF receptor, this chimera can be specifically activated by PDGF to signal through the altered FGFR3 intracellular domain. These chimeras induce ligand-dependent autophosphorylation of the chimera receptor and stimulated strong phosphorylation of mitogen-activated protein (MAP) kinase and phospholipase C. Compared with cells transfected with a chimera with normal FGFR3 sequences, cells transfected with the chimera with the FGFR3 achondroplasia mutation were more responsive to ligand, with less sustained MAP kinase activation, indicative of a primed or constitutively-on condition. This observation is consistent with the hypothesis that these mutations weaken ligand control of the FGFR3 receptor, and may provide a biochemical explanation for the observation that the TD phenotype is more severe than that of achondroplasia (132). Subsequently, using similar chimeras, this same group analyzed the effects of six FGFR3 mutations that result in skeletal dysplasias (133). The three tyrosine kinase domain mutations (Lys650Glu, Lys650Met, and Asn540Lys) all resulted in strong ligand-independent tyrosine phosphorylation, especially the Lys650Glu TD type II (133). Lys650Met (TD type I) and Lys650Glu mutations resulted in autoactivation of the receptor sufficient to produce partial differentiation of the PC-12 cells (133). Chimeras containing mutations in the transmembrane domain of FGFR3 (achondroplasia mutations Gly375Cys and Gly380Arg, and Crouzon syndrome mutation Ala391Glu) displayed normal expression and activation, but did exhibit a greater response to lower concentrations of ligand.
Similar autonomous receptor activation has been observed before with mutations in other tyrosine kinase receptors, such as FGFR2, epidermal growth factor, colony stimulating factor 1, and the RET oncogene (134, 135, 136, 137, 138, 139). Additional studies will need to be done before the cellular and biochemical consequences of these mutations are fully understood. It will be important to understand the transcriptional differences caused by FGFR3-mediated signal transduction in both normal and disease states.
|
V. GH Treatment |
|---|
|
TopAbstractI. IntroductionII. Fibroblast Growth Factor...III. Clinical and Molecular...IV. Biochemical Analysis of...V. GH TreatmentVI. ImplicationsReferences |
|---|
Growth has increased during the early phases of GH therapy in both patients with achondroplasia and hypochondroplasia: 34 patients with achondroplasia or hypochondroplasia in the National Cooperative Growth Study have been treated with an average dose of GH of 0.317 mg/kg per week for an average of 2.6 yr and have gained an average of 0.7 SD in height. These data suggest that the abnormal growth cartilage in patients with chondrodysplasia responds to GH therapy (144). Weber et al. (143) studied the effects of recombinant human GH treatment in six prepubertal children with achondroplasia, ranging in age from 2 to 8 yr. During the year of treatment the growth velocity increased from 1.1 to 2.6 cm/year in three patients, while in the others no variation was detected, confirming the individual variability in the response to GH treatment.
To clarify the effectiveness of GH treatment of short stature in achondroplasia, a long-term treatment study with a large number of patients was performed (140): 42 children (16 males and 26 females, age 3–14 yr) with achondroplasia were examined. After the evaluation, the children were treated with GH for more than 2 yr, and then posttreatment growth velocity and body proportion parameters were determined. The annual height gain during GH therapy was significantly greater than before therapy (3.9 ± 1.0 cm/yr before treatment vs. 6.5 ± 1.8 cm/yr for the first year, and 4.6 ± 1.6 cm/yr for the second year of treatment), and body disproportion was not aggravated during the treatment period. The authors concluded that GH might be beneficial in the treatment of short stature in children with achondroplasia in the first 2 yr of treatment (140).
In another study, 15 children with achondroplasia, 7 boys (4.8–12.2 yr of age) and 12 girls (5.7–2.2 yr of age), were treated daily with human GH at a dosage of 1 IU/kg/week (141). Auxological assessments were performed 6 months before, at initiation of, and at 6, 12, and 24 months after initiation of GH therapy. During the first semester of GH treatment, a significant increase in height velocity, from 3.2 to 8.3 cm/yr, was observed in all children. However, during the second semester, a relative decrease in growth rate was observed. By the end of the first year, height velocity had increased from 3.2 to 6.9 cm/yr (mean, 3.7 cm/yr; range, 1.1–8 cm/yr) in 13 children and remained unchanged in 2 children. Height velocity declined during the next 12 months and, by the end of the second year of treatment, had increased in only 7 of the 9 children who had completed 2 yr of therapy (mean increase, 3.1 cm/yr); 2 children did not respond to GH therapy. These studies demonstrate that GH treatment resulted in an increased growth rate in some children with achondroplasia; however, the amount of increase declined during the second year of treatment, and the final heights of these individuals is not yet known.
|
VI. Implications |
|---|
|
TopAbstractI. IntroductionII. Fibroblast Growth Factor...III. Clinical and Molecular...IV. Biochemical Analysis of...V. GH TreatmentVI. ImplicationsReferences |
|---|
Moreover, the high degree of phenotypic specificity associated with FGFR3 mutations is highly unusual in the study of human genetics and disease. That more than 97% of persons with achondroplasia have exactly the same amino acid substitution at nucleotide 1138 was a first in the study of human mutations and genetic disorders. Furthermore, the common Pro250Arg amino acid substitution, which causes Muenke coronal craniosynostosis, adds to the uniqueness of genotype-phenotype correlations in the FGFR disorders. The explanation for this high degree of mutability remains an intriguing question. Since the G1138A achondroplasia mutation was recognized, similar observations have been made in FGFR3 and other human FGFR genes regarding other skeletal phenotypes, including hypochondroplasia and TD, and Pfeiffer and Apert syndromes. Furthermore, it seems that particular nucleotides in FGFR genes are more highly susceptible to mutation than other nucleotides. There is a high degree of correlation in the locations of observed mutations from one FGFR to another. Again, this conservation of mutations at particular sites in the FGFR genes is a very intriguing biological phenomenon. It is possible that FGFR mutations in the same locations have been identified because mutations at these sites are capable of conferring constitutive activation of the receptor, while mutations at other sites do occur, but do not lead to severe phenotypic changes and, thus, have not yet been identified. However the different degrees of constitutive activation cannot explain all the differences in the resulting phenotypes. Furthermore, why do some FGFR3 mutations result in a relatively small amount of skeletal changes, such as in Muenke craniosynostosis and Crouzon syndrome with acanthosis nigricans? These questions remain to be answered.
The prenatal diagnosis of many skeletal dysplasias is difficult to make. A certain sonographic diagnosis of a de novo case is rarely possible. In fact, achondroplasia is almost never detected on prenatal ultrasound before the third trimester. In face of uncertainty, physicians sometimes elect to emphasize the most severe alternative diagnoses. In a recent retrospective study, 25% of achondroplasia patients were given an incorrect prenatal diagnosis of a lethal or very severe disorder (146). By identifying mutations responsible for skeletal dysplasias, mutational analysis can be offered when a short-limb disorder is detected by ultrasound; however, indiscriminate use of FGFR3 molecular testing cannot be recommended. Thus, the prenatal diagnosis becomes more effective, making it possible to reduce the amount of incorrect and potentially harmful information provided to the parents (146, 147), thereby helping to avoid unnecessary terminations. Therefore, the high degree of specificity of the FGFR3 G1138A mutation for the achondroplasia phenotype has profound implications for persons with achondroplasia, their families, and their physicians. Because the achondroplasia mutations are easily detectable by molecular means, the molecular diagnosis is one that can now be performed in many molecular diagnostic laboratories. One very positive outcome of the ability for molecular diagnosis is to provide couples at risk for children with homozygous achondroplasia with reliable prenatal diagnosis for the inevitably lethal condition. Individuals providing genetic counseling should keep in mind that there are other disorders with mild degrees of limb shortening that will not be diagnosed by FGFR3 molecular analysis, and that most cases diagnosed in the second trimester with short limbs and a small chest will have a lethal form of dwarfism, but, most likely, not TD or homozygous achondroplasia. These cases clearly do not have achondroplasia and there are many forms of lethal skeletal dysplasias other than TD; therefore, molecular testing for the common FGFR3 mutations cannot be recommended. The precise diagnosis in these cases is best made after birth or by radiographs and histology.
Additionally, as has been found with many genetic disorders in the past, understanding the physiology behind the achondroplasia family of disease, and other skeletal dysplasias, has the potential to help us understand the normal mechanisms of skeletal growth and development. As we gain a greater understanding of why a particular phenotype results from a particular, but specific, mutation in the FGFR3 gene, we should gain insight into the molecular mechanisms that distinquish one bone from another.
|
Acknowledgments |
|---|
|
Footnotes |
|---|
1 Supported by the Division of Intramural Research, National Human
Genome Research Institute, National Institutes of Health, and Division
of Intramural Research, National Institute of Dental and Craniofacial
Research, National Institutes of Health. 
|
References |
|---|
|
TopAbstractI. IntroductionII. Fibroblast Growth Factor...III. Clinical and Molecular...IV. Biochemical Analysis of...V. GH TreatmentVI. ImplicationsReferences |
|---|
