|Home | About | Journals | Submit | Contact Us | Français|
Lacrimo-auriculo-dento-digital (LADD) syndrome is characterized by abnormalities in lacrimal and salivary glands, in teeth, and in the distal limbs. Genetic studies have implicated heterozygous mutations in fibroblast growth factor 10 (FGF10) and in FGF receptor 2 (FGFR2) in LADD syndrome. However, it is not clear whether LADD syndrome mutations (LADD mutations) are gain- or loss-of-function mutations. In order to reveal the molecular mechanism underlying LADD syndrome, we have compared the biological properties of FGF10 LADD and FGFR2 LADD mutants to the activities of their normal counterparts. These experiments show that the biological activities of three different FGF10 LADD mutants are severely impaired by different mechanisms. Moreover, haploinsufficiency caused by defective FGF10 mutants leads to LADD syndrome. We also demonstrate that the tyrosine kinase activities of FGFR2 LADD mutants expressed in transfected cells are strongly compromised. Since tyrosine kinase activity is stimulated by ligand-induced receptor dimerization, FGFR2 LADD mutants may also exert a dominant inhibitory effect on signaling via wild-type FGFR2 expressed in the same cell. These experiments underscore the importance of signal strength in mediating biological responses and that relatively small changes in receptor signaling may influence the outcome of developmental processes in cells or organs that do not possess redundant signaling pathway.
Fibroblast growth factors (FGFs) mediate their biological responses by binding to four receptor tyrosine kinases (RTKs) designated FGF receptor 1 (FGFR1) to FGF4 (25). The binding of FGF to FGFR in the presence of heparin sulfate glycosaminoglycan induces receptor dimerization and the activation of the protein tyrosine kinase domain (30). Tyrosine autophosphorylation and the recruitment of a complement of downstream signaling molecules result in the stimulation of various signaling cascades that play critical roles in mediating the pleiotropic responses of FGFs during development and in the adult organism (10).
Like all RTKs, FGFRs are composed of an extracellular ligand binding domain, a transmembrane region, and a cytoplasmic region containing a catalytic protein tyrosine kinase core and additional regulatory sequences. The extracellular domain is composed of three immunoglobulin-like domains (designated D1, D2, and D3), a stretch of negatively charged amino acids in the linker connecting D1 and D2 termed the acidic box, and a conserved positively charged region in D2 that serves as the binding site for heparin sulfate or heparin (13, 17, 30, 31). FGFR1, -2, and -3 transcripts are subject to alternative RNA splicing in which exon 7 of the FGFR gene codes for a common N-terminal half of D3 (referred to as ΙΙΙa) and exons 8 and 9 code for the C-terminal half of D3 to generate the ΙΙΙb and ΙΙΙc isoforms, respectively (21, 39). The ΙΙΙb isoforms are expressed exclusively in epithelial cells, while the ΙΙΙc isoforms are expressed only in mesenchymal cells (1, 9, 26, 38). Moreover, the IIIb and IIIc isoforms of FGFR1, -2, and -3 bind to different complements of FGFs that are expressed exclusively in mesenchymal or epithelial cells, respectively. For example, the FGFR2-ΙΙΙb isoform (also designated FGFR2b) binds FGF7, FGF10, and FGF22, while FGFR2-ΙΙΙc (also designated FGFR2c) binds FGF2, FGF8, FGF17, and FGF18 (14). FGF1, on the other hand, functions as a universal FGFR ligand, as it binds to all “b” and “c” FGFR isoforms. Strict lineage-specific expression of the two alternatively spliced isoforms of FGFR2 is essential for normal embryonic development.
Targeted disruption of the FGFR1 gene has shown that the FGFR1c isoform plays an essential role during early embryogenesis. The biological roles of FGFR2b, FGFR2c, and their specific ligands have also been explored by targeted disruption of isoform-specific genes fragments in the mouse by use of homologous recombination. Targeted disruption of the FGFR2b results in lethality at birth due to lung agenesis (7). Interestingly, the phenotype of the FGF10 null mice is similar to the phenotype of FGFR2b null mice (32). Characterizations of the phenotypes of mice deficient in FGF10 or FGFR2b have shown that FGF10 and FGFR2b play an essential role in the control of branching morphogenesis during the development of lung, pancreas, mammary gland, thyroid, lacrimal gland, and salivary gland. Moreover, aplasia of the lacrimal gland and hypoplasia of the salivary gland were observed for adult heterozygous FGF10 mice, indicating that the normal development of both glands depends on a precisely balanced dose of signaling stimulated by FGF10 (5, 17). Finally, human genetic studies and selective targeting of the FGFR3b and FGFR3c isoforms in mice have implicated the FGFR3c isoform in a variety of skeletal disorders (6, 8).
Recent studies have shown that patients with aplasia (or hypoplasia) of the lacrimal and salivary glands (ALSG) bear heterozygous mutations in the FGF10 gene (4, 5). Mutations in FGF10 were also detected for patients with lacrimo-auriculo-dento-digital (LADD) syndrome, which shows overlapping features with ALSG but in addition is characterized by facial dysmorphisms, outer and inner ear anomalies and hearing loss, teeth anomalies, distal limb malformations, and, more infrequently, impairment of kidney and lung development (2, 4, 12, 20, 22, 28). Genetic analysis has also revealed heterozygous mutations in FGFR2 and FGFR3 in LADD syndrome patients (28), implicating aberrant signaling by FGF10, FGFR2, or FGFR3 in this heterogeneous disorder.
In this report, we describe the biological properties of FGF10 and FGFR2b mutants implicated in LADD syndrome. We show that LADD syndrome mutations (LADD mutations) cause inactivation of FGF10 and that the tyrosine kinase activity of FGFR2b LADD mutants expressed in cultured cells is severely compromised. While the FGF10 mutation causes haploinsufficiency, the FGFR2b mutants may exert a dominant interfering effect on signaling via normal FGFR2b, causing LADD syndrome.
FGF1 and FGF2 were prepared and used as a stock solution at a concentration of 100 μg/ml with 5 mg/ml heparin for stimulation of cultured cells (33). Heparin agarose beads and Lipofectamine 2000 were purchased from Sigma and Invitrogen, respectively. Anti-FGFR2, anti-Grb2, anti-FRS2, and anti-Shc antibodies were previously described (6, 19). Antiphosphotyrosine (anti-p-Tyr) antibodies were purchased from Upstate Biotechnology. Anti-phospho-mitogen-activated protein kinase (anti-pMAPK) and anti-MAPK were purchased from Cell Signaling Technology. Horseradish peroxidase-conjugated protein A and horseradish peroxidase-conjugated goat anti-mouse antibodies were purchased from Kirkegaard & Perry Laboratories and Santa Cruz Biotechnology, respectively. Geneticin was purchased from GIBCO. Human FGF10 and FGF10 LADD mutants were expressed in Escherichia coli by use of the bacterial expression vector pET11c (Novagen) (3). Both Mirb and retroviral pBABE/neo expression vectors were used for human FGFR2 expression as previously described (11, 24). Point mutations in FGF10 and FGFR2 were generated using a QuikChange site-directed mutagenesis kit from Stratagene. pcDNA-3 expression vector (Invitrogen) was used for FGFR2b transient expression in 293 cells.
BL21(DE3) pLysS E. coli cells were transformed with expression vector for FGF10 or FGF10 LADD mutants and grown overnight in LB medium at 25°C. The bacterial cell pellet was resuspended in lysis buffer (20 mM HEPES, pH 7.4, 150 mM NaCl, 5% glycerol, 1 mM phenylmethylsulfonyl fluoride, and 2 mM EDTA), lysed by use of a French pressure cell (Thermo Electron Corp.), and centrifuged for 1 h at 32,000 × g. The cell supernatant was incubated with heparin-agarose beads for 1.5 h at 4°C, and FGF-bound beads were washed three times with lysis buffer. Washed beads were applied to a column and FGF10 or FGF10 LADD mutants were eluted using 20 mM HEPES buffer, pH 7.4, containing 1 M NaCl. Eluted proteins were further purified by fast-performance liquid chromatography (Amersham Biosciences) using a Mono S column (GE Healthcare). Protein purity was determined using sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) analysis.
L6 cells devoid of endogenous FGFRs were cultured in Dulbecco's modified Eagle medium (DMEM) containing 10% fetal bovine serum (FBS), 2 mM l-glutamine with 100 μg/ml penicillin and 100 μg/ml streptomycin. L6 cells were transfected with expression vectors for wild-type (WT) FGFR2b and FGFR2b mutants that were cloned into the Mirb or pBABE/neo expression vectors. Cells were transfected with Lipofectamine 2000 and selected in growth medium containing 1 mg/ml geneticin. Individual clones as well as cell pools were screened for FGFR2 expression using anti-FGFR2 antibodies. Prior to growth factor stimulation, cells were starved overnight in medium containing 0.1% FBS. 293 cells were transfected with Lipofectamine 2000 and incubated in transfection medium for 6 h; this was followed by changing the medium to DMEM containing 10% FBS. Cells were harvested and lysed 18 h later.
Human FGF10 (10 μg) was labeled with 0.5 mCi of 125I by use of Iodo-Gen iodination tubes (Pierce) following the manufacturer's instructions. For the displacement binding assay, L6 cells expressing FGFR2b were grown in 24-well plates in DMEM containing 10% FBS. Confluent cells were washed with DMEM containing 0.5% bovine serum albumin (BSA) and then incubated for 1 h at room temperature with 2 ng of 125I-labeled FGF10 in the presence of increasing concentrations of FGF1, FGF10, or the FGF10 LADD mutants. Cells were then washed three times with cold DMEM-BSA and lysed in 0.5 ml of 0.5 M NaOH for 30 min at room temperature, and 100 μl of the cell lysate was applied to 10 ml of Opti-Fluor scintillation cocktail (Perkin Elmer) in order to measure cell-associated radioactivity (using an LS6500 scintillation counter from Beckman Coulter).
Limited proteolysis analysis of FGF10 or FGF10 LADD mutants was carried out using factor Xa (12 × 10−5 U/μl), endoproteinase Glu-C (V8 protease) (0.04 μg/μl), and endoproteinase Lys-C (12 × 10−4 U/μl). All enzymes were purchased from Roche and used in a series of 10-fold dilutions. FGF10 or FGF10 LADD mutants were incubated with the enzymes for 2 h at 25°C, and the proteolytic products were visualized by SDS-PAGE followed by Coomassie brilliant blue staining.
FGF10 or the FGF10 LADD mutants (30 μg/ml in 20 mM HEPES, pH 7.4, 400 mM NaCl) were incubated at 37°C for various periods and then cooled to room temperature. To measure fluorescence emission, samples were excited at a wavelength of 285 nm, and emission was scanned at λ of between 300 and 380 nm by use of a fluorometer (Photon Technology International).
Several mutations in FGF10 were identified in LADD patients, including a missense mutation in which cysteine 106 is substituted by a phenylalanine (p.C106F), a missense mutation in which isoleucine 156 is substituted by an arginine (p.I156R), and a nonsense mutation (p.K137X) causing a deletion of 71 carboxy-terminal residues of FGF10 (22, 28). To study the biological activities of the FGF10 LADD mutants, WT or mutant FGF10 proteins were expressed in E. coli. While the expression of WT FGF10 and the I156R mutant was detected in cells that were induced by isopropyl-β-d-thiogalactopyranoside (IPTG) at 37°C, expression of the C106F mutant was detected only in cells induced at 25°C, suggesting temperature sensitivity of the C106F LADD mutant (data not shown). Expression of the K137X mutant could not be detected for any experimental condition that was tried. We surmised that because of the large truncation, the K137X FGF10 mutant was probably misfolded, resulting in rapid degradation. For large-scale production of WT FGF10 and the two FGF10 LADD mutants, cells were induced at 25°C overnight. WT or mutant FGF10 was purified on a heparin affinity matrix followed by cation-exchange chromatography using a Mono S column. WT FGF10 and the LADD mutants were eluted from either heparin or the Mono S columns (Fig. (Fig.1)1) at similar salt concentrations, indicating that the surface charge of the FGF10 LADD mutants was not substantially altered.
We first compared the capacities of WT FGF10 and the I156R LADD mutant to stimulate L6 cells expressing FGFR2b. Lysates of unstimulated or ligand-stimulated cells were subjected to immunoprecipitation with anti-FGFR2 antibodies followed by SDS-PAGE and immunoblotting with anti-p-Tyr antibodies. The experiment presented in Fig. Fig.2A2A shows that unlike WT FGF10, which stimulated L6 cells, the I156R mutant was unable to stimulate the tyrosine autophosphorylation of FGFR2b or the tyrosine phosphorylation of the downstream signaling molecules FRS2 and Shc. Furthermore, MAPK stimulation was not detected in L6 cells stimulated with the I156R LADD mutant. To understand why the I156R FGF10 mutant was unable to induce receptor autophosphorylation, we next examined its binding affinity towards FGFR2b by using a displacement assay in which cell-bound 125I-labeled WT FGF10 was displaced by increasing concentrations of native FGF10 (as a control) or the I156R FGF10 LADD mutant. The experiment presented in Fig. Fig.2C2C shows that the 50% inhibitory concentration of 125I-labeled FGF10 bound to FGFR2b of the I156R LADD mutant is approximately ninefold higher than the 50% inhibitory concentration of FGF10 towards FGFR2b expressed on the cell surfaces of L6 cells. This experiment demonstrates that the impaired biological activity of the I156R mutant is caused by compromised binding affinity towards FGFR2b.
During the course of the expression and purification of FGF10 and the FGF10 LADD mutants, the production of the C106F LADD mutant after the E. coli cells were induced at 37°C was very low compared to the production of WT FGF10 and of the I156R LADD mutant under the same conditions. The displacement assay presented in Fig. Fig.2C2C showed that the receptor binding profile of the purified C106F LADD mutant is similar to the binding characteristics of FGF10 towards FGFR2b expressed on the cell surfaces of L6 cells. Moreover, both WT FGF10 and the C106F LADD mutant showed similar levels of stimulation of tyrosine autophosphorylation of FGFR2b and similar MAPK responses upon cell stimulation when a broad range of FGF10 or C106F LADD mutant concentrations were applied (Fig. 2A and B). We next examined whether the C106F mutation affected FGF10 stability. Protein stability was first tested by comparing the susceptibilities of FGF10 and the C106F LADD mutant towards limited proteolysis by the enzymes V8 protease, factor Xa, and Lys-C. The experiment presented in Fig. Fig.3A3A shows that unlike WT FGF10, the C106F LADD mutant undergoes rapid degradation, resulting in the formation of low-molecular-weight degradation products when treated with these enzymes.
The stabilities of WT FGF10 and the C106F and I156R LADD mutants at 37°C were further tested by comparing the intrinsic fluorescence spectra of FGF10 and the FGF10 LADD mutant as a function of time at 37°C. The experiments presented in Fig. Fig.3B3B show that the fluorescence spectra of FGF10 and the I156R LADD mutant were stable over a period of 3 h of incubation at 37°C. By contrast, the fluorescence spectrum of the C106F LADD mutant changed within 2 min of incubation at 37°C. Moreover, a strong increase in the fluorescence intensity emitted from the C106F LADD mutant was observed after 20 min of incubation at 37°C. It is well established that fluorescence spectra of tryptophan and tyrosine residues of proteins can be used as a diagnostic tool to reveal local structural alterations that take place in host proteins (34). Since FGF10 and the FGF10 LADD mutants were excited at a wavelength of 285 nm, changes in fluorescence spectra will reflect local structural changes that take place in the vicinity of tryptophan residues of the proteins. The changes in the fluorescence spectrum of the C106F LADD mutant at 37°C may reflect the reduced stability of FGF10 mutant at a physiological temperature.
Since our results show that incubation of the C106F LADD mutant at 37°C affected the structural integrity of the mutant protein, we next examined the impact of these changes on cellular responses induced by the C106F LADD mutant at 37°C. Both FGF10 and the C106F LADD mutant were preincubated for various periods at 37°C in the absence or presence of heparin, and each sample was then used to stimulate, for an additional 5 minutes, L6 cells stably expressing FGFR2b. The experiment presented in Fig. Fig.3C3C shows that the capacity of C106F LADD mutant to induce the tyrosine autophosphorylation of the FGFR2b and MAPK response was strongly compromised and that within 30 min of incubation at 37°C the activity of the C106F LADD mutant nearly vanished (Fig. (Fig.3C).3C). Interestingly, in the presence of exogenous heparin the stability of C106F was protected. It is possible, however, that the majority of C106F LADD mutant molecules were not secreted from the cells after biosynthesis, as most of the unstable mutant molecules were degraded shortly after production.
In order to gain insights into the mechanism underlying the diminished receptor binding activity of the I156R mutant, we examined the potential impact of the substitution of isoleucine 156 by an arginine residue on the receptor binding region of FGF10 in the X-ray crystal structure of FGF10 (40). Figure Figure44 shows that isoleucine 156 is located in the β8 strand of FGF10, in a region that forms contacts with the βF-βG loop of FGFR2b (23, 40). It is expected that the substitution of isoleucine 156 with the larger arginine residue will cause a steric clash with critical amino acids in the ligand binding pocket of FGFR2b (Fig. (Fig.4B).4B). Moreover, the I156R LADD mutation may also destroy the highly conserved hydrogen bonds between Gly160 and Asn162 of FGF10 with Arg251 in the D2-D3 linker region of FGFR2, due to repulsion between Arg156 of the FGF10 LADD mutant and Arg251 of FGFR2b. The I156R mutation may also destroy the electrostatic interaction between Arg78 of FGFR10 with Arg251 and Asp283 of D3 of FGFR2b. Moreover, the Arg156 mutation in FGF10 may cause a steric clash and/or an electrostatic repulsion with Arg251 of FGFR2b that will disrupt three critical electrostatic interactions essential for FGF10 binding to FGFR2b.
We have also examined the potential impact of the substitution of cysteine 106 with a phenylalanine residue on the structure of FGF10 and its interactions with FGFR2b. Cysteine 106 is located in the β3-β4 loop of FGF10, a region that does not play a role in FGFR2b recognition (40). Indeed, the FGFR2b binding activity of the C106F LADD mutant remained unchanged. However, the replacement of cysteine 106 with a large hydrophobic residue such as phenylalanine may create a bulky region, which will become exposed, resulting in reduced stability and susceptibility to proteolytic digestion. Human mature FGF10 (amino acids 38 to 208) contains an additional cysteine residue at position 150. Inspection of the FGF10 structure (40) shows that Cys106 is located on the surface of FGF10, while Cys150 is buried in FGF10, and that the distance between Cys106 and Cys150 is 22 Å. The two cysteines are unlikely to form an intramolecular disulfide bond.
The mutations in FGFR2 that were identified for a variety of skeletal dysplasias have been mapped to the extracellular ligand binding domain in the vast majority of cases and less frequently in the tyrosine kinase domain, including the catalytic core and its regulatory activation loop. Biochemical characterization of mutant receptors has shown that FGFR2 mutations that are responsible for craniosynostosis and other severe bone disorders are gain-of-function mutations that enhance the tyrosine kinase activity of the receptor molecules. The mutations in FGFR2 that have been implicated in LADD syndrome were mapped to the activation loop or the catalytic loop of FGFR2. However, it is not clear whether the LADD mutations in FGFR2 are gain- or loss-of-function mutations.
In order to reveal the molecular mechanism of the FGFR2 LADD mutations, expression vectors that direct the synthesis of FGFR2b carrying the LADD mutations were prepared and tested for their biological activity following transient expression in 293 cells or by stable expression in L6 cells. The tyrosine kinase activities of FGFR2b carrying LADD mutations in the activation loop (A648T and R649S) or in the catalytic loop (A628T) were compared to the tyrosine kinase activities of WT FGFR2b, of a kinase-defective (KD) FGFR2b mutant (K508A), and of a Pfeiffer syndrome gain-of-function FGFR2b (K641R) mutant (18, 29) as controls. Cells expressing WT FGFR2b or the various mutants were stimulated with FGF10, and lysates from unstimulated or FGF10-stimulated cells were subjected to immunoprecipitation with anti-FGFR2 antibodies followed by SDS-PAGE and immunoblotting with anti-p-Tyr antibodies. The results presented in Fig. Fig.55 show FGF10 stimulation of the tyrosine autophosphorylation of WT FGFR2b and the K641R Pfeiffer syndrome FGFR2b mutant. By contrast, FGF10 stimulation of the three FGFR2b LADD mutants led to very weak tyrosine autophosphorylation of mutant FGFR2b. Different degrees of tyrosine autophosphorylation were detected for the three LADD mutants, with the R649S mutant having the highest tyrosine kinase activity and the A628T mutant having the weakest tyrosine kinase activity. In addition, the three LADD mutants failed to stimulate tyrosine phosphorylation of two well-characterized FGFR substrates, FRS2 and Shc, as revealed by immunoprecipitation with anti-Grb2, anti-FRS2, or anti-Shc antibodies followed by SDS-PAGE and immunoblotting with anti-p-Tyr antibodies (Fig. (Fig.5B).5B). We have also shown that MAPK stimulation in response to FGF10 stimulation could barely be detected in L6 cells expressing the FGFR2b LADD mutants. By contrast, robust FGF10-dependent or FGF10-independent MAPK responses were detected in L6 cells expressing WT FGFR2b or the K641R Pfeiffer syndrome FGFR2b mutant, respectively (Fig. (Fig.5C).5C). Furthermore, coexpression of WT FGFR2b with the R649S LADD mutant in transfected cells reveals a dominant interfering effect on the autophosphorylation of WT FGFR2 expressed in the same cells (Fig. (Fig.5D5D).
On the basis of these experiments, we conclude that the intrinsic tyrosine kinase activity of FGFR2 LADD mutants is strongly attenuated, resulting in impaired tyrosine phosphorylation of critical substrates and cell signaling. It is noteworthy, however, that weak tyrosine kinase activities could be detected for the LADD mutants that are above the background tyrosine kinase activity detected for the KD K508A mutant. Since all LADD mutations are clustered in the catalytic domain of FGFR2, a region common to both the b and c isoforms, these mutations will affect the tyrosine kinase activities of both the FGFR2b and FGFR2c isoforms.
Signaling pathways activated by FGFs and FGFRs have been identified in multicellular organisms from Caenorhabditis elegans to vertebrates. It is now well established that the FGFR family of RTKs and their numerous ligands play crucial roles in many developmental and physiological processes and that a variety of diseases are caused by aberrant signaling induced by FGFs or FGFRs (25, 30). The biological roles of individual FGFs and FGFRs have been analyzed by targeted disruption in mice of individual or combinations of FGF or FGFR genes or via analysis of disease-causing mutations in humans. In humans, both loss- and gain-of-function heterozygous mutations have been described. Several human skeletal dysplasias are caused by gain-of-function mutations in FGFR1, FGFR2, and FGFR3. Activating mutations mapped in the extracellular ligand binding domain were found in FGFR1 and FGFR2 associated with Pfeiffer, Crouzon, Jackson-Weiss, and Apert syndromes and in the FGFR2 kinase domain associated with Pfeiffer and Crouzon syndromes. Likewise, activating mutations in FGFR3 were mapped to the transmembrane and the tyrosine kinase domains found for achondroplasia, thanatophoric dysplasia type I (TDI), and TDII (reviewed in references 6, 27, 35, and 36). Interestingly, the nature and severity of the disease might depend on the mutated amino acid; replacement of the same residue by different amino acids may strongly influence the severity of disease. For example, replacement of lysine 650 in the tyrosine kinase domain of FGFR3 by a methionine results in short limbs and developmental delay (dwarfism, severe achondroplasia with developmental delay and acanthosis nigricans), while replacement of the same lysine by a glutamic acid results in lethality (TDII) (15, 16, 35). These observations emphasize the complexity of molecular change that takes place as the consequence of mutations in FGFRs that may influence receptor activity, receptor stability, and receptor localization, among other changes. It is striking that all syndromes described above are caused by gain-of-function mutations in the c isoform of FGFRs, although in some cases mutations were also found in a region common to both b and c FGFR isoforms.
Genetic studies of families and patients with sporadic LADD syndrome revealed mutations in the tyrosine kinase domains of FGFR2 and FGFR3 (28). The three missense mutations identified in FGFR2 are located in catalytic (A628T) and activation (A648T, R649S) loops, and a single mutation was found in the tyrosine kinase domain of FGFR3 (D513N). Several mutations in LADD syndrome patients were identified in FGF10 (C106F, I156R), including a nonsense mutation leading to a premature stop of translation (K137X) (22, 28). A nonredundant role of the FGF10-FGFR2b signaling pathway in lacrimal and salivary gland development was proposed based on the phenotypes of mice deficient in these genes. Aplasia in the lacrimal gland and hypoplasia in the salivary gland were observed for FGF10+/− mice as well as for mice heterozygous for FGF10 and FGFR2b (FGF10+/− FGFR2b+/− mice) (5). Despite the observation that mutations in either FGF10 or FGFR2 cause LADD syndrome, the underlying mechanism is not clear. Moreover, in the absence of biochemical data, modeling of the mutations in the structures of FGF10 and FGFR2 kinase domain did not provide conclusive insights concerning molecular mechanisms.
To reveal the mechanism underlying the molecular basis of LADD syndrome, we have compared the biochemical and biological properties of FGF10 or FGFR2b LADD mutants to the properties of their normal counterparts. Our results show that each of the three LADD mutations affects FGF10 activity by a different mechanism. While the I156R mutant is deficient in binding to FGFR2b, the C106F mutant is unstable at physiological temperatures and is most likely degraded shortly after synthesis before being delivered to its target cell. The K137X mutant, lacking a large C-terminal part of the molecule, was not produced; if it is produced, this mutant will not have any biological activity because its FGF core will have been severely disrupted.
The biological characterization of the FGF10 LADD mutants shows that the activity of the three LADD mutants is strongly compromised. Haploinsufficiency caused by the severely impaired FGF10 mutant leads to LADD syndrome, as the signal induced by FGF10 coded by the normal allele of LADD syndrome patients is not sufficient for mediating the normal development of the salivary and lacrimal glands. This conclusion is supported by genetic studies with mice demonstrating that two copies of FGF10 are required for the normal development of the salivary and lacrimal glands (5). Moreover, the description of two additional FGF10 mutants (R80S and G138E) for patients with ALSG further emphasizes the critical and nonredundant role of FGF10 in salivary and lacrimal gland development (4). The reason why FGF10 haploinsufficiency causes ALSG and the more severe LADD syndrome remains to be elucidated.
Analysis of the biological properties of ectopically expressed FGFR2 LADD mutants in the activation loops (A648T and R649S) or in the catalytic loop (A628T) shows that the FGFR2b LADD mutants are deficient in tyrosine kinase activity. The A628T mutant has the weakest activity, the A648T mutant exhibits an intermediate activity, and the R649S mutant has the highest activity, albeit lower than the tyrosine kinase activity of WT FGFR2 following ligand stimulation. Unlike the LADD mutation in the ligand molecule that is caused by the haploinsufficiency of FGF10, the FGFR2 LADD mutation will have a dominant negative effect on signaling mediated via WT FGFR2 expressed in the same cell (Fig. (Fig.5D).5D). Three types of FGFR2 dimers will be formed in cells expressing equal amounts of normal FGFR2 and the FGFR2 LADD mutant in response to FGF10 stimulation (Fig. (Fig.6):6): one-fourth of the molecules are homodimers of WT FGFR2 with normal tyrosine kinase activity, one-fourth are homodimers of the FGFR2 LADD mutant with a very weak tyrosine kinase activity, and one-half of the molecules are heterodimers composed of WT and LADD FGFR2 with attenuated tyrosine kinase activities. Since the activation of FGFRs is mediated by ligand-induced receptor dimerization and transphosphorylation, mutant receptors are unable to efficiently phosphorylate WT receptors on autophosphorylation sites in the activation loop of the tyrosine kinase core, a step essential for enhanced and sustained tyrosine kinase activity. Consequently, the defective LADD mutant will exert a dominant inhibitory affect on normal FGFR2, resulting in a strongly attenuated signal.
It has been reported that FGFR2+/− mice are normal, indicating that a single FGFR2 allele (providing 50% of the signal that take place in normal mice) is sufficient to support normal mouse development, including development at the lacrimal and salivary glands. The attenuated signal transmitted in cells coexpressing an FGFR2 LADD mutant together with WT FGFR2 in response to ligand stimulation (assuming that the WT and the FGFR2 LADD mutant are equally expressed) is expected to be larger than 25% and lower than 50% (25% < signal < 50%) of the signal transmitted by FGFR2 in normal mice. This conclusion underscores the importance of exact doses of receptor signaling in mediating biological responses; a small change in signal strength may have a strong impact on development and homeostasis in cells and tissues that do not possess a redundant signaling pathway.
On the basis of the previous genetic studies with and biochemical characterization of LADD mutations, it is possible to conclude that signaling pathways that are stimulated by FGF10 and mediated by FGFR2b play a critical role in the development and morphogenesis of branching organs such as salivary and lacrimal glands, kidneys, and lungs, which among other organs are affected by LADD syndrome. It has been shown that FGF10 expressed by mesenchymal cells will stimulate FGFR2b expressed in epithelial cells. Activation of FGFR2b in epithelial cells leads to the production of FGF8, which in turn stimulates the activity of FGFR2c and FGFR1c expressed in mesenchymal cells (37). Attenuation in signaling via FGF10 or FGFR2b will lead to the disruption of an important cell signaling circuit that takes place between epithelial and mesenchymal cells during development. Disruption of the epithelial-mesenchymal cell signaling circuit may lead to the dental and skeletal abnormalities seen for LADD syndrome patients.
We also conclude that normal development of lacrimal glands, salivary glands, ears, skeleton, and other organs relies on a correct dose of FGF10 signaling through FGFR2b and that both copies of the FGF10 gene are required for the normal development of these organs; these requirements are not met in the case of the ear, skeletal, and dental abnormalities associated with LADD syndrome. Unlike FGF10 mutations causing ligand haploinsufficiency without affecting the action of the product of the WT FGF10 allele, mutations in FGFR2 lead to more-severe diseases by exerting a dominant negative effect on WT FGFR2 and potentially also on other FGFRs that are expressed in the same cell.
No specific phenotypic differences were observed for patients with mutations in FGF10 and FGFR2 or by comparing phenotypes caused by different FGFR2 mutations. In general, a wide range of phenotypic variability of symptoms exists in LADD syndrome patients, even those within the same family and carrying the identical mutation. This fact makes genotype-phenotype correlations difficult. We propose that the phenotypic outcome of impaired FGF signaling caused by mutations in LADD genes is further modified by genetic, environmental, and stochastic factors which remain to be elucidated.
Finally, although both FGFR2b and FGFR2c carry the LADD mutations, LADD mutation is primarily mediated by the FGFR2b isoform, implying that the compromised signaling via the FGFR2c mutant seen for LADD syndrome is compensated for by other members of the FGFR family expressed in mesenchymal cells.
This work was supported by NIH grants AR 05141448 (J.S.), AR 051886 (J.S.), and P50 AR 054086 (J.S.) and by the European Commission FP6 Integrated Project EUROHEAR, LSHG-CT-20054-512063 (B.W.). Satoru Yuzawa was supported by a fellowship from the Uehara Memorial Foundation.
We thank N. Itoh for the FGF10 plasmid.
Published ahead of print on 6 August 2007.