Sprouty 2 disturbs FGFR3 degradation in thanatophoric dysplasia type II: a severe form of human achondroplasia

Abstract

Thanatophoric dysplasia is a member of the achondroplasia family of human skeletal dysplasias, which result from FGFR3 mutations that exaggerate this receptor's inhibitory influence on chondrocyte proliferation and differentiation in the skeletal growth plate. We have previously reported that defective lysosomal degradation of activated receptor contributes to the gain-of-function of the mutant FGFR3. We now provide evidence that this disturbance is mediated by the receptor's kinase activity and involves constitutive induction and activation of Spry2. Our findings suggest that activated Spry2 may interfere with c-Cbl-mediated ubiquitination of FGFR3 by sequestering c-Cbl. They provide novel insight into the pathogenesis of this group of human skeletal dysplasias and identify a mechanism that potentially could be targeted therapeutically.

Fig. 1

Accumulation of FGFR3 is influenced by receptor tyrosine kinase activity and enhanced by lysosomal inhibition in transiently transfected cells. (A) Western blot of lysates from COS-7 cell cultures expressing WT or TDII FGFR3-V5-His with and without kinase activity using anti-V5 antibody indicates a reduced abundance of FGFR3 when kinase activity is abolished. Actin was probed as a loading control. (B) Autoradiograph resulting from a pulse-chase experiment documents a shorter half-life for kinase dead variants of WT and TDII FGFR3-V5-His. (C) Graphic representation of data represented in (B). The respective half-lives are determined from plotting normalized percent of receptor variants versus time. (D) Immunostaining of parallel cultures with anti-GFP antibody shows increased fluorescence intensity of TDII compared to WT FGFR3-GFP and reduced intensity for both receptors, especially the TDII FGFR3-GFP, when kinase activity is abolished. (E) Western blot of lysates from COS-7 cells expressing FGFR3-GFP variants indicates that treatment with the lysosomal inhibitor Bafilomycin (Baf) but not with the proteasomal inhibitor MG132 increases relative amount of the kinase active WT and TDII FGFR3-GFP. Kinase dead mutants may increase slightly with lysosomal inhibition. Tubulin was probed as a loading control. Ct = control, V = vector only, kd = kinase dead, tub = tubulin.

Fig. 2

Activation of FGFR3 by ligand or mutation induces expression of Spry2. (A) Real time RT-PCR shows a two-fold increase in Spry2 expression after treatment of COS-7 cells stably expressing WT FGFR3-GFP with FGF2. Results of three independent experiments are plotted showing standard deviation. (B) Representative western blots of endogenous Spry2 in lysates of TRex-293 cells expressing FGFR3-GFP variants using anti-Spry2 antibody and actin as a loading control. Spry2 is induced by FGF1 in presence of WT FGFR3-GFP, but is abundant and constitutively expressed in the presence of TDII FGFR3-GFP. Spry2 levels are low in cells expressing receptor lacking kinase activity. (C) Summarizes the results of four independent experiments. GFP-tagged WT and mutant FGFR3 expression was induced overnight using 10 μg/ml tetracycline. Cells were then made quiescent by serum starvation before stimulating with 10 ng/ml FGF1 and 10 U/ml heparin.

Fig. 3

Spry2 is more abundant in growth plate chondrocytes from mice and humans with TDII as compared to controls. (A) Immunostaining for endogenous Spry2 in proximal tibial growth plates from 1 week-old WT (left) and TDII (K644E) (right) FGFR3 knock-in mice is more intense for the TDII growth plate. The cell density is also higher in the TDII growth plate. (B) A higher magnification of a comparable region of both growth plates outlined in (A) more clearly demonstrates differences in both staining intensity and cell density. (C) Western blot using anti-Spry2 antibody and tubulin (Tub) as a loading control demonstrates greater amount of Spry2 in growth plate tissue extracts from a human fetus with TDII (K650E) compared to a normal control.

Fig. 4

Spry2 phosphorylation in response to FGFR3 signals leads to interaction with c-Cbl. (A) COS-7 cells stably expressing WT or TDII FGFR3-GFP were transfected with HA-hSpry2, immunoprecipitated for HA and blotted with antibody against phosphotyrosine (pY). A low level of tyrosine phosphorylation increased at 30 min following FGF2 treatment in the presence of WT FGFR3. A high level of phosphorylation of Spry2 independent of ligand stimulation was detected in cells expressing TDII FGFR3-GFP. (B) COS-7 cells stably expressing WT or TDII FGFR3-GFP were transfected with WT or Y55F HA-hSpry2, immunoprecipitated for c-Cbl and blotted for HA (HA-hSpry2). The c-Cbl immunoprecipitates contained WT, but not mutant HA-hSpry2 in the presence of both WT and TDII FGFR3-GFP. The pull-down was positive only after ligand treatment for WT FGFR3-GFP, while it was ligand-independent for TDII FGFR3-GFP. V = vector only.

Fig. 5

Spry2 acts through c-Cbl to influence FGFR3 ubiquitination, degradation and signaling. (A) COS-7 cells stably expressing WT FGFR3-GFP were co-transfected with WT or Y55F HA-hSpry2 and HA-Ub, and equivalent lysates were immunoprecipitated for GFP (FGFR3-GFP) and blotted with antibodies to HA for HA-ubiquitin. Receptor ubiquitination was substantially reduced by WT HA-hSpry2 and much less reduced by mutant HA-hSpry2 (Y55F). (B) COS-7 cells were co-transfected with WT or TDII FGFR3-GFP and increasing amounts of c-Cbl (c-Cbl-dsRed) or dominant-negative 70z-Cbl (for TDII FGFR3-GFP) and then blotted for GFP (FGFR3-GFP) and c-Cbl. Empty-vector DNA was added to keep total amount of transfected DNA constant. Both WT and TDII FGFR3-GFP show a progressive decrease in receptor levels with increasing amounts of c-Cbl, but not with 70z-Cbl. (C) Lysates from COS-7 cells stably expressing WT-FGFR3-GFP transfected with vector, WT or Y55F HA-hSpry2 and treated or not treated with FGF2 were blotted for pErk1,2 (upper panel) and total Erk1,2 (lower panel). Addition of FGF2 for 30 min slightly increases phosphorylation of Erk1,2. Erk1,2 phosphorylation is increased substantially in the presence of WT HA-hSpry2 compared to Y55F HA-hSpry2. V = vector only, HA-Ub = HA-ubiquitin, pErk1, 2 = phosphoErk1,2.