Abstract

Activating mutations in the fibroblast growth factor receptor 3 (FGFR3) cause the most common genetic form of human dwarfism, achondroplasia (ACH). Small chemical inhibitors of FGFR tyrosine kinase activity are considered to be viable option for treating ACH, but little experimental evidence supports this claim. We evaluated five FGFR tyrosine kinase inhibitors (TKIs) (SU5402, PD173074, AZD1480, AZD4547 and BGJ398) for their activity against FGFR signaling in chondrocytes. All five TKIs strongly inhibited FGFR activation in cultured chondrocytes and limb rudiment cultures, completely relieving FGFR-mediated inhibition of chondrocyte proliferation and maturation. In contrast, TKI treatment of newborn mice did not improve skeletal growth and had lethal toxic effects on the liver, lungs and kidneys. In cell-free kinase assays as well as in vitro and in vivo cell assays, none of the tested TKIs demonstrated selectivity for FGFR3 over three other FGFR tyrosine kinases. In addition, the TKIs exhibited significant off-target activity when screened against a panel of 14 unrelated tyrosine kinases. This was most extensive in SU5402 and AZD1480, which inhibited DDR2, IGF1R, FLT3, TRKA, FLT4, ABL and JAK3 with efficiencies similar to or greater than those for FGFR. Low target specificity and toxicity of FGFR TKIs thus compromise their use for treatment of ACH. Conceptually, different avenues of therapeutic FGFR3 targeting should be investigated.

Introduction

Mammalian long bones grow at the epiphyseal growth plate cartilage, where chondrocytes proliferate and mature via a well-characterized sequence of events involving hypertrophic differentiation, extracellular matrix mineralization and, ultimately, apoptosis (1). Proper cartilage growth depends on complex spatiotemporal interactions among many different signaling systems involving known chondrocyte mitogens such as parathyroid hormone-related protein, C-natriuretic protein, bone morphogenetic proteins and leukemia inhibitory factor (2–5). Systems that inhibit cartilage growth are equally important for the establishment of final skeletal size, which is restricted in terrestrial vertebrates.

The receptor tyrosine kinase fibroblast growth factor receptor 3 (FGFR3) is one of these physiological negative growth regulators, as evidenced by skeletal overgrowth in Fgfr3 null mice, or the finding that loss-of-function mutations in the human FGFR3 gene cause camptodactyly, tall stature and hearing loss syndrome (6,7). Correspondingly, the introduction of constitutively active FGFR3 leads to dwarfism in mice, while gene duplication involving the FGFR3 ligand FGF4 is linked to short-legged phenotypes in multiple dog breeds (8–10). Activating FGFR3 mutations in humans are associated with four skeletal conditions that are collectively termed FGFR3-related skeletal dysplasias: achondroplasia (ACH), hypochondroplasia, severe ACH with developmental delay and acanthosis nigricans, and lethal thanatophoric dysplasia (TD) (11). ACH is the most common form of human dwarfism and is characterized by disproportioned short stature, a narrow trunk, short proximal limbs, a large head with frontal bossing and midface hypoplasia and trident hands. ACH complications originate from abnormal bone growth and include developmental delay, hydrocephalus, spinal stenosis, articulation problems, hearing loss, cardiorespiratory and sleep dysfunctions, obesity and reproductive complications (12). Successful targeting of FGFR3 during the postnatal growth period may restore normal skeletal size in those affected with ACH and may also be helpful in treating other skeletal conditions. Because FGFR3 is a major physiological negative regulator of bone growth, a safe and effective FGFR3 inhibitor will undoubtedly revolutionize the treatment of short-stature syndromes in general, possibly including many that are unrelated to FGFR3.

Tyrosine kinases represent relatively easily druggable targets, as proved by the existence of many therapies based on small molecule tyrosine kinase inhibitors (TKIs), which have enabled groundbreaking advances in the treatment of tumors such as chronic myeloid leukemia (13). The four FGF-receptor tyrosine kinases (FGFR1–4) represent attractive therapeutic targets because many cancers are associated with genetic lesions affecting FGFR genes. These include activating mutations, gene amplifications and chromosomal translocations that create fusion FGFR oncogenes (14). The first-generation FGFR TKIs targeting the ATP-binding site were reported in 1997 and are based on oxindole (SU5402) or pyridopyrimidine cores (PD173074) (15–17). Many other FGFR TKIs have since been developed, including the recently introduced compounds AZD4547, AZD1480 and BGJ398 (18–20). At least six FGFR TKIs are currently in Phase I/II of clinical trials for solid tumors (21).

It is accepted that FGFR TKIs may be useful for ACH therapy, although there is little experimental evidence supporting this claim (11,22,23). The present study was designed to critically evaluate the capacity of FGFR TKIs to target FGFR signaling in cartilage.

Results and Discussion

TKIs inhibit FGFR signaling in cultured chondrocytes

We evaluated two early-generation FGFR TKIs (SU5402, PD173074) and three late-generation inhibitors (AZD1480, AZD4547, BGJ398) to determine whether any significant improvement in anti-FGFR activity has been achieved during more than 10 years of development separating the two categories. It should be noted that AZD1480, AZD4547 and BGJ398 are currently undergoing cancer clinical trials and therefore represent the most likely candidates for an ACH therapy (21). Rat chondrosarcoma (RCS) chondrocytes, a well-characterized chondrocytic cell line (24), were used to initially evaluate the inhibitors' activity in a chondrocyte environment. RCS cells respond to experimental FGFR activation (via treatment with an exogenous FGF ligand) with potent cellular growth arrest. This process depends on activation of the extracellular signal-regulated/mitogen-activated protein kinase (ERK MAP) pathway and is accompanied by the induction of premature senescence, loss of extracellular matrix components and changes in cellular shape (25,26). We first evaluated the TKIs' activity in FGF2-mediated RCS growth arrest, which is a sensitive reporter of FGFR activity (27). Cells were treated with the inhibitors at concentrations spanning a range of 3–5 orders of magnitude in order to establish a basal activity profile for each TKI. All TKIs completely rescued RCS cells from FGF2-mediated inhibition of proliferation at concentrations ranging from as little as 5 nM for BGJ398 to 5 µM for SU5402 (Fig. 1A). All of the tested TKIs exhibited toxic effects at certain concentrations, manifested as inhibition of proliferation of FGF2-naïve RCS cells. These effects were particularly severe for SU5402 and AZD1480 and took place at concentrations below those required to rescue FGF2-mediated growth arrest. The optimal active concentrations for each TKI (i.e. concentrations at which the inhibitor completely restored RCS growth without causing toxic effects) were 15–20 nM for PD173074, 10–50 nM for AZD4547 and 5–10 nM for BGJ398 (Fig. 1). All TKIs restored growth even in the presence of 100 ng/ml FGF2, which is well above the FGF2 concentration required for maximum strength of the growth arrest phenotype (∼40 ng/ml; Fig. 1B) (27). Other features of FGF2-mediated growth arrest, e.g. changes in cell morphology and induction of premature senescence (26), were also completely reverted by TKIs (Fig. 2A–C). We also evaluated the TKIs' capacity to reverse an already established growth arrest phenotype; this would be important for a prospective ACH treatment, which would have to be effective in cartilage where growth had been chronically suppressed by aberrant FGFR3 signaling for months prior to the first administration of the treatment. Figure 2D shows that TKI treatment successfully reversed fully established FGF2-mediated growth arrest in RCS chondrocytes.

Figure 1.

TKIs rescue FGF2-mediated inhibition of chondrocyte proliferation. (A) RCS chondrocytes were treated with FGF2 and FGFR TKIs SU5402, PD173074, AZD1480, AZD4547 and BGJ398 as indicated, and cell numbers were determined by crystal violet staining 96 h later. Data represent averages for six wells. Results are representative of three independent experiments. Note the potent FGF2-mediated growth arrest, which is completely relieved by each of the FGFR TKIs at different concentrations. Also note that all TKIs show toxicity (growth inhibition in FGF2-naïve cells) at higher concentrations. (B) The optimal TKI concentrations determined in the experiments presented in (A) were applied to RCS cells treated with >40 ng/ml of FGF2, which induces the strongest possible level of FGF2-mediated growth arrest. All TKIs rescue this maximal growth arrest phenotype.

Figure 1.

TKIs rescue FGF2-mediated inhibition of chondrocyte proliferation. (A) RCS chondrocytes were treated with FGF2 and FGFR TKIs SU5402, PD173074, AZD1480, AZD4547 and BGJ398 as indicated, and cell numbers were determined by crystal violet staining 96 h later. Data represent averages for six wells. Results are representative of three independent experiments. Note the potent FGF2-mediated growth arrest, which is completely relieved by each of the FGFR TKIs at different concentrations. Also note that all TKIs show toxicity (growth inhibition in FGF2-naïve cells) at higher concentrations. (B) The optimal TKI concentrations determined in the experiments presented in (A) were applied to RCS cells treated with >40 ng/ml of FGF2, which induces the strongest possible level of FGF2-mediated growth arrest. All TKIs rescue this maximal growth arrest phenotype.

Figure 2.

TKIs rescue FGF2-mediated senescence in chondrocytes. (A) RCS chondrocytes were treated as indicated for 96 h and photographed (scale bar 200 µm). Note the FGF2-mediated effect on RCS colony sizes, which is rescued by TKIs. (B) Higher magnification of regions highlighted with black boxes in (A) demonstrating enlargement and flattening of cells treated with FGF2, which is rescued by SU5402. (C) FGF2-mediated premature senescence in RCS cells (26) is accompanied by induction of lamin A/C and caveolin, and suppression of ID2, as detected by WB. Actin was used as a loading control. (D) RCS cells were treated with FGF2 for 4 days, and cell numbers were determined by crystal violet staining. In some samples (3 days), TKIs were added after FGF2-mediated growth arrest was fully established in 24 h of incubation with FGF2 (28). Note that all of the tested TKIs reversed the established growth arrest. Data represent averages for six wells with indicated standard deviations. Results are representative of three independent experiments. Statistically significant differences are indicated (ANOVA, **P < 0.001).

Figure 2.

TKIs rescue FGF2-mediated senescence in chondrocytes. (A) RCS chondrocytes were treated as indicated for 96 h and photographed (scale bar 200 µm). Note the FGF2-mediated effect on RCS colony sizes, which is rescued by TKIs. (B) Higher magnification of regions highlighted with black boxes in (A) demonstrating enlargement and flattening of cells treated with FGF2, which is rescued by SU5402. (C) FGF2-mediated premature senescence in RCS cells (26) is accompanied by induction of lamin A/C and caveolin, and suppression of ID2, as detected by WB. Actin was used as a loading control. (D) RCS cells were treated with FGF2 for 4 days, and cell numbers were determined by crystal violet staining. In some samples (3 days), TKIs were added after FGF2-mediated growth arrest was fully established in 24 h of incubation with FGF2 (28). Note that all of the tested TKIs reversed the established growth arrest. Data represent averages for six wells with indicated standard deviations. Results are representative of three independent experiments. Statistically significant differences are indicated (ANOVA, **P < 0.001).

TKIs also abolished FGF-mediated activation of FGFR downstream signal transduction in RCS chondrocytes. Specifically, TKI treatment inhibited the FGF2-induced phosphorylation of the FRS2 adapter protein, MEK and ERK MAP kinases (all of which are components of the FGFR-RAS-ERK signaling pathway), as well as the phosphorylation of LRP6, which mediates crosstalk between FGFR and WNT/β-catenin signaling in chondrocytes (29). Similarly, potent suppression of FGFR signal transduction was achieved in cells where the FGF2-mediated FGFR activation was significantly enhanced by addition of heparin (Fig. 3). Altogether, our data show that all tested TKIs effectively suppressed FGFR signaling in cultured chondrocytes.

Figure 3.

TKIs suppress FGF2-mediated activation of FGFR signal transduction. FGFR signaling was activated in RCS chondrocytes by addition of exogenous FGF2 (20 ng/ml), and the cells were analyzed for activating phosphorylation (p) of the LRP6 WNT co-receptor, FRS2 adapter, and MEK and ERK kinases. As loading controls, the levels of actin and the total levels of the targeted kinases (irrespective of their phosphorylation status) were also determined. The electrophoretic mobility shift on FRS2 (total FRS2 blot) corresponds to the ERK-mediated feedback phosphorylation described previously (30). Exogenous heparin (1 µg/ml) was added to FGF2 since it potently enhances the RCS response to FGF2 stimulus (28). All TKIs completely abolish the FGF2-mediated activation of FGFR signal transduction.

Figure 3.

TKIs suppress FGF2-mediated activation of FGFR signal transduction. FGFR signaling was activated in RCS chondrocytes by addition of exogenous FGF2 (20 ng/ml), and the cells were analyzed for activating phosphorylation (p) of the LRP6 WNT co-receptor, FRS2 adapter, and MEK and ERK kinases. As loading controls, the levels of actin and the total levels of the targeted kinases (irrespective of their phosphorylation status) were also determined. The electrophoretic mobility shift on FRS2 (total FRS2 blot) corresponds to the ERK-mediated feedback phosphorylation described previously (30). Exogenous heparin (1 µg/ml) was added to FGF2 since it potently enhances the RCS response to FGF2 stimulus (28). All TKIs completely abolish the FGF2-mediated activation of FGFR signal transduction.

TKIs inhibit FGFR signaling in mouse tibia cultures

We next investigated tibia cultures, which are another well-characterized model system for studying FGFR signaling in cartilage. Tibias isolated from mouse embryonal limbs continue their growth program in vitro, and this growth can be suppressed by treatment with exogenous FGF2 (31). Tibias isolated from E18 embryos were cultured for 8 days, and their growth plates were analyzed by histology. Treatment with FGF2 caused significant attenuation of growth, accompanied by the absence of cells with typical hypertrophic chondrocyte morphology (Fig. 4A–C). This finding was confirmed by the loss of cells expressing hypertrophic chondrocyte marker collagen type 10 (Col10a1) in FGF2-treated tibias, whereas FGF2-naïve tibias exhibited normal Col10a1 expression.

Figure 4.

AZD4547 inhibits FGFR signaling in mouse tibia culture. (A) Tibias isolated from E18 embryos were cultured in media supplemented with FGF2 (100 ng/ml) and/or AZD4547 for 8 days, and their length was measured after extraction from the embryo and after 8 days of incubation. (B) Graphs representing differences between both time points (Δ length) in three independent experiments (four tibias for each treatment) are shown (1–3). Statistically significant differences are highlighted (ANOVA, **p < 0.001). Note the FGF2-mediated growth inhibition that is rescued by AZD4547. Also note the significant increase in lengths of tibias treated with AZD4547 alone in contrast to untreated controls, which reflects the inhibition of endogenous FGFR3 signaling. (C) Histological sections of representative tibias were stained with HE. Note the marked reduction of hypertrophic cartilage (H) in tibias treated with FGF2, with corresponding loss of Col10a1 expression (bottom panel, Col10a1 in situ hybridization counterstained with eosin). This phenotype is rescued by AZD4547. Cj, chondro-osseous junction. Scale bars 100 µM. Note the Col10a1 expression domain matching the zone of enlarged hypertrophic chondrocytes. Also note a band of hypertrophic cartilage at the end of the control growth plates negative for Col10a1 mRNA, which represents cartilage matrix lacunae containing apoptotic chondrocytes, which would normally be eliminated by invading bone; these accumulate here due to the disruption of bone growth in the tibia cultures. (D) Confirmation of the function of the Col10a1 in situ hybridization protocol on freshly isolated tibias; the signal corresponds to the zone of hypertrophic chondrocytes adjacent to the bone tissue. (E) Relative levels of FGFR1–4 expression in growth plates isolated from E18 tibias, determined by quantitative reverse transcriptase-polymerase chain reaction (qRT-PCR).

Figure 4.

AZD4547 inhibits FGFR signaling in mouse tibia culture. (A) Tibias isolated from E18 embryos were cultured in media supplemented with FGF2 (100 ng/ml) and/or AZD4547 for 8 days, and their length was measured after extraction from the embryo and after 8 days of incubation. (B) Graphs representing differences between both time points (Δ length) in three independent experiments (four tibias for each treatment) are shown (1–3). Statistically significant differences are highlighted (ANOVA, **p < 0.001). Note the FGF2-mediated growth inhibition that is rescued by AZD4547. Also note the significant increase in lengths of tibias treated with AZD4547 alone in contrast to untreated controls, which reflects the inhibition of endogenous FGFR3 signaling. (C) Histological sections of representative tibias were stained with HE. Note the marked reduction of hypertrophic cartilage (H) in tibias treated with FGF2, with corresponding loss of Col10a1 expression (bottom panel, Col10a1 in situ hybridization counterstained with eosin). This phenotype is rescued by AZD4547. Cj, chondro-osseous junction. Scale bars 100 µM. Note the Col10a1 expression domain matching the zone of enlarged hypertrophic chondrocytes. Also note a band of hypertrophic cartilage at the end of the control growth plates negative for Col10a1 mRNA, which represents cartilage matrix lacunae containing apoptotic chondrocytes, which would normally be eliminated by invading bone; these accumulate here due to the disruption of bone growth in the tibia cultures. (D) Confirmation of the function of the Col10a1 in situ hybridization protocol on freshly isolated tibias; the signal corresponds to the zone of hypertrophic chondrocytes adjacent to the bone tissue. (E) Relative levels of FGFR1–4 expression in growth plates isolated from E18 tibias, determined by quantitative reverse transcriptase-polymerase chain reaction (qRT-PCR).

The tibia culture faithfully reproduces cartilage growth but not the growth of the bone component of the developing limb. This is because cartilage is avascular tissue capable of autonomous growth, whereas bone depends on a blood supply that is disrupted in culture. The growth plate morphology of the tibias shown in Figure 4C differs from the normal morphology in that there is a band of hypertrophic cartilage at the end of growth plates that does not express Col10a1 mRNA. This appears to be the cartilage matrix lacunae containing apoptotic chondrocytes (therefore no longer expressing Col10a1 mRNA) that would normally be eliminated by the invading bone, but accumulate here due to the disruption of bone growth in culture. The authenticity of the Col10a1 in situ signal was confirmed in growth plates of freshly isolated tibias (Fig. 4D).

Treatment of tibias with 0.5 µM AZD4547 rescued the FGF2 growth-inhibited phenotype and restored normal growth plate architecture including Col10a1 expression. Similar results were obtained with 1 µM AZD4547 or with 0.5 µM BGJ398 (Supplementary Material, Fig. S1). Interestingly, tibias treated with AZD4547 and BGJ398 alone or in combination with FGF2 grew longer than untreated tibias (Fig. 4A–C, Supplementary Material, Fig. S1). The simple explanation for this phenotype is that the TKIs inhibited endogenous FGFR3 signaling, which restricts growth (6). The inhibition of physiological FGFR3 signaling using the C-natriuretic peptide analog BMN111 reportedly elevated growth in mice and cynomolgus monkeys (32), further supporting the hypothesis that inhibition of endogenous wild-type (wt) FGFR3 signaling accelerates growth in healthy mammals. This suggests that anti-FGFR3 agents could potentially be useful for treating other short-stature conditions unrelated to FGFR3-related skeletal dysplasia. Our data demonstrate that TKIs efficiently target FGFR3 signaling within the growth plate cartilage of tibia cultures despite the presence of a dense extracellular matrix and the lack of vasculature. Similar results were obtained by others (23), who reported the rescue of growth inhibition in tibias isolated from mice harboring FGFR3-Y367C (a homolog of the human TD mutation) following treatment with the PD173074 derivative A31 (33).

AZD4547 suppresses skeletal growth and causes lethality in vivo

We next sought to determine whether the positive effects of TKI treatment on bone growth in vitro could be replicated in vivo. AZD4547 was chosen for these experiments because of its most favorable effects in the in vitro experiments. Different concentrations of AZD4547 (0.5, 1, 5, 10, 50, 250, 1000 and 2000 µM) were intraperitoneally injected into newborn wt (CD1) mice, and their rate of postnatal growth was compared with control animals from the same litter injected with the AZD4547 vehicle dimethyl sulfoxide (DMSO) (Fig. 5A). Animals were injected every week for 28 days, and their weight, body and tail lengths were recorded. No significant changes in any of the tested parameters occurred at 0.5–250 µM AZD4547, although a slight increase in body weight was noted at 50 µM AZD4547 (Fig. 5D; Supplementary Material, Fig. S2). The animals' weight and length decreased, and lethality occurred in animals treated with 1000 or 2000 µM AZD4547 (Fig. 5B; Supplementary Material, Fig. S2). Lethality always occurred 2–3 days after the first injection; animals that survived this period always remained alive until the end of the experiment 28 days later. The signs of lethal toxicity were progressive muscle paralysis, altered breathing frequency, cyanosis and respiratory arrest. Histological analysis revealed infiltration of inflammatory cells and obscured hepatocyte borders in the liver, cellular hypotrophy and reduction of the convoluted tubules in the kidney, and hemorrhage accompanied by disturbances of the interalveolar septum in the lung (Fig. 5C). Surviving animals exhibited microscopic alterations in liver morphology including steatosis and accumulation of lipid droplets in cells. No changes in lung or kidney morphology were observed in surviving animals (data not shown).

Figure 5.

AZD4547 causes toxicity in mouse. (A–F) Newborn mice were injected intraperitoneally with 50 µl of AZD4547 or its vehicle DMSO (red lines in D and F) at indicated concentrations. (A) The overall appearance of animals after 28 days of AZD4547 application. Note the significantly shorter size in mice treated with 1000 and 2000 µM AZD4547. (B and C) Lethality in animals treated with higher AZD4547 concentrations, affecting the liver (necrotic areas, lymphocyte accumulation), lungs (alveolar hemorrhage) and kidneys (reduction of renal tubules). Scale bar 100 µm. (D) Body weight determined at the time of injection (every week) up to day 28. The length of body (from snout to tip of the tail) and tail (from radix to tip of the tail) was measured at the end of experiment at day 28 and plotted. Note the inhibitory effect of higher AZD4547 concentrations on body weight and length of body and tail (t-test; *P < 0.05, **P < 0.01, ***P < 0.001). (E and F) The skeleton was stained with alizarin red and alcian blue, and the lengths of each animal's scapula, pelvis, femur and humerus were measured. (E) A representative skull, scapula, pelvis and femur of mice treated either with DMSO or 1000 µM AZD4547. (F) Relative quantification of the scapula, pelvis, femur and humerus length reveals significant shortening following treatment with 1000 and 2000 µM AZD4547 (t-test).

Figure 5.

AZD4547 causes toxicity in mouse. (A–F) Newborn mice were injected intraperitoneally with 50 µl of AZD4547 or its vehicle DMSO (red lines in D and F) at indicated concentrations. (A) The overall appearance of animals after 28 days of AZD4547 application. Note the significantly shorter size in mice treated with 1000 and 2000 µM AZD4547. (B and C) Lethality in animals treated with higher AZD4547 concentrations, affecting the liver (necrotic areas, lymphocyte accumulation), lungs (alveolar hemorrhage) and kidneys (reduction of renal tubules). Scale bar 100 µm. (D) Body weight determined at the time of injection (every week) up to day 28. The length of body (from snout to tip of the tail) and tail (from radix to tip of the tail) was measured at the end of experiment at day 28 and plotted. Note the inhibitory effect of higher AZD4547 concentrations on body weight and length of body and tail (t-test; *P < 0.05, **P < 0.01, ***P < 0.001). (E and F) The skeleton was stained with alizarin red and alcian blue, and the lengths of each animal's scapula, pelvis, femur and humerus were measured. (E) A representative skull, scapula, pelvis and femur of mice treated either with DMSO or 1000 µM AZD4547. (F) Relative quantification of the scapula, pelvis, femur and humerus length reveals significant shortening following treatment with 1000 and 2000 µM AZD4547 (t-test).

Next, we analyzed the effect of AZD4547 on skeletal morphology. No significant changes were associated with 0.5–250 µM AZD4547 (Fig. S3). In contrast, injections of 1000 and 2000 µM AZD4547 caused significant shortening of long bones and the bones of the skull (Fig. 5E and F; Supplementary Material, Fig. S3). Interestingly, the general shape of the bones and the morphologies of individual processes and prominences in the analyzed bones were similar to those of controls, suggesting that the inhibition of skeletal growth may have been due to the general toxicity of AZD4547 rather than direct effects on bone. AZD4547 induced no changes in blood glucose, bile acids, vitamin D, inorganic phosphate, calcium, HDL, cholesterol and triacylglycerol levels (Supplementary Material, Fig. S4).

FGFR TKIs inhibit other tyrosine kinases

Taken together, our results demonstrate that while FGFR TKIs are potent inhibitors of FGFR signaling in cultured chondrocytes, they fail to accelerate skeletal growth in vivo and instead cause significant toxicity and lethality. This toxicity may stem from two principal sources: the TKI-mediated systemic inhibition of FGFR signaling and/or the TKI-mediated inhibition of other kinases unrelated to FGFR.

We used cell-free kinase assays to evaluate TKI activity toward all four FGFR tyrosine kinases (FGFR1–4) as well as 14 unrelated tyrosine kinases (AXL, TYRO3, DDR2, EGFR, IGF1R, INSR, MET, cKIT, FLT3, FMS, TRKA, FLT4, ABL, JAK3). The activity of recombinant kinases was assayed using either the phosphorylation of recombinant STAT1 or the autophosphorylation of the tested kinase as a reporter, with the TKIs being added directly to the kinase reaction mixtures. TKIs inhibited FGFR1–4 activity to the variable levels, ranging from almost complete suppression of FGFR1 activity by SU5402 to the as little as ∼20% suppression of FGFR2 activity by BGJ398 (Fig. 6; Table 1). The relative levels of FGFR1–4 inhibition by individual TKIs were as follows: SU5402 (FGFR1 = FGFR2>>FGFR3 = FGFR4), PD173074 (FGFR1 = FGFR2 = FGFR3>FGFR4), AZD1480 (FGFR1>FGFR2>FGFR3>FGFR4), AZD4547 (FGFR1>FGFR2 = FGFR3>FGFR4) and BGJ398 (FGFR3>FGFR4>FGFR1 = FGFR2). PD173074 and AZD4547 were moderately selective for the FGFRs over the other tested kinases, although the extent of their inhibition of other kinase activity was comparable with their anti-FGFR activity in many cases (INSR and FLT4 for PD173074; FMS, TRKA, FLT4 and ABL for AZD4547) (Table 1). Conversely, SU5402 and AZD1480 were potent inhibitors of other kinases and in some cases were more active against off-target kinases than against FGFR (FLT3, TRKA, FLT4 and JAK3 for SU5402; DDR2, IGF1R, FLT3, TRKA, FLT4, ABL and JAK3 for AZD1480). Interestingly, SU5402 and AZD1480 were also the most toxic TKIs in the RCS growth experiments (Fig. 1A).

Table 1.

TKI activity against FGFR1–4 and 14 other FGFR-unrelated tyrosine kinases

Tyrosine kinase Control (%) SU5402 (10 µM) PD173074 (50 nM) AZD1480 (5 µM) AZD4547 (50 nM) BGJ398 (20 nM) 
FGFR1* 100 5.5 ± 0.9 51.4 ± 5.3 7.3 ± 3.9 16.0 ± 3.6 74.1 ± 2.9 
FGFR2* 100 6.2 ± 1.1 48.0 ± 13.6 14.1 ± 5.8 35.0 ± 2.4 82.4 ± 13.4 
FGFR3* 100 18.2 ± 5.0 46.2 ± 5.5 28.6 ± 5.6 36.1 ± 4.9 33.5 ± 7.0 
FGFR4* 100 20.1 ± 2.8 69.4 ± 16.6 52.3 ± 5.9 75.2 ± 15.5 56.7 ± 12.6 
AXL* 100 52.0 ± 5.1 83.4 ± 10.0 74.5 ± 11.2 82.8 ± 2.3 88.6 ± 8.6 
TYRO3* 100 58.8 ± 14.2 92.6 ± 4.2 88.2 ± 9.2 89.4 ± 7.2 94.7 ± 11.7 
DDR2* 100 42.1 ± 1.7 89.5 ± 11.1 39.6 ± 6.1 82.7 ± 6.1 75.1 ± 15.8 
EGFR 100 89.6 ± 9.0 95.3 ± 1.9 76.7 ± 11.8 94.7 ± 14.2 93.2 ± 8.5 
IGF1R* 100 42.8 ± 3.2 81.5 ± 15.3 21.0 ± 7.8 81.9 ± 24.4 94.5 ± 11.6 
INSR* 100 48.8 ± 5.0 78.0 ± 4.0 56.9 ± 1.4 82.0 ± 15.0 78.4 ± 15.0 
MET* 100 105.1 ± 4.7 105.8 ± 1.8 102.4 ± 6.5 108.0 ± 10.4 96.8 ± 11.5 
C-KIT 100 36.3 ± 7.6 86.9 ± 11.5 84.3 ± 9.1 81.6 ± 9.4 97.5 ± 13.7 
FLT3* 100 0.0 ± 0.0 85.0 ± 6.6 36.9 ± 6.6 88.4 ± 9.1 87.5 ± 12.5 
FMS 100 58.8 ± 3.9 102.4 ± 4.7 99.7 ± 6.9 72.9 ± 3.3 78.9 ± 13.1 
TRKA 100 6.0 ± 0.2 100.7 ± 9.3 5.7 ± 1.1 75.2 ± 1.8 93.8 ± 5.2 
FLT4* 100 1.1 ± 0.1 73.6 ± 0.8 2.0 ± 0.5 79.5 ± 10.7 89.7 ± 3.1 
ABL* 100 13.4 ± 1.4 85.0 ± 6.8 13.5 ± 4.3 76.6 ± 3.7 93.8 ± 15.9 
JAK3* 100 0.0 ± 0.0 94.9 ± 12.8 0.0 ± 0.0 93.2 ± 5.1 84.6 ± 10.4 
Tyrosine kinase Control (%) SU5402 (10 µM) PD173074 (50 nM) AZD1480 (5 µM) AZD4547 (50 nM) BGJ398 (20 nM) 
FGFR1* 100 5.5 ± 0.9 51.4 ± 5.3 7.3 ± 3.9 16.0 ± 3.6 74.1 ± 2.9 
FGFR2* 100 6.2 ± 1.1 48.0 ± 13.6 14.1 ± 5.8 35.0 ± 2.4 82.4 ± 13.4 
FGFR3* 100 18.2 ± 5.0 46.2 ± 5.5 28.6 ± 5.6 36.1 ± 4.9 33.5 ± 7.0 
FGFR4* 100 20.1 ± 2.8 69.4 ± 16.6 52.3 ± 5.9 75.2 ± 15.5 56.7 ± 12.6 
AXL* 100 52.0 ± 5.1 83.4 ± 10.0 74.5 ± 11.2 82.8 ± 2.3 88.6 ± 8.6 
TYRO3* 100 58.8 ± 14.2 92.6 ± 4.2 88.2 ± 9.2 89.4 ± 7.2 94.7 ± 11.7 
DDR2* 100 42.1 ± 1.7 89.5 ± 11.1 39.6 ± 6.1 82.7 ± 6.1 75.1 ± 15.8 
EGFR 100 89.6 ± 9.0 95.3 ± 1.9 76.7 ± 11.8 94.7 ± 14.2 93.2 ± 8.5 
IGF1R* 100 42.8 ± 3.2 81.5 ± 15.3 21.0 ± 7.8 81.9 ± 24.4 94.5 ± 11.6 
INSR* 100 48.8 ± 5.0 78.0 ± 4.0 56.9 ± 1.4 82.0 ± 15.0 78.4 ± 15.0 
MET* 100 105.1 ± 4.7 105.8 ± 1.8 102.4 ± 6.5 108.0 ± 10.4 96.8 ± 11.5 
C-KIT 100 36.3 ± 7.6 86.9 ± 11.5 84.3 ± 9.1 81.6 ± 9.4 97.5 ± 13.7 
FLT3* 100 0.0 ± 0.0 85.0 ± 6.6 36.9 ± 6.6 88.4 ± 9.1 87.5 ± 12.5 
FMS 100 58.8 ± 3.9 102.4 ± 4.7 99.7 ± 6.9 72.9 ± 3.3 78.9 ± 13.1 
TRKA 100 6.0 ± 0.2 100.7 ± 9.3 5.7 ± 1.1 75.2 ± 1.8 93.8 ± 5.2 
FLT4* 100 1.1 ± 0.1 73.6 ± 0.8 2.0 ± 0.5 79.5 ± 10.7 89.7 ± 3.1 
ABL* 100 13.4 ± 1.4 85.0 ± 6.8 13.5 ± 4.3 76.6 ± 3.7 93.8 ± 15.9 
JAK3* 100 0.0 ± 0.0 94.9 ± 12.8 0.0 ± 0.0 93.2 ± 5.1 84.6 ± 10.4 

The percentages of kinase activity relative to activity in the absence of inhibitor (control, 100% active) are shown. Data represent averages from three independent experiments with indicated standard deviations. * indicates when phosphorylation of substrate (STAT1Y701) was used to detect the activity of given kinase (Fig. 6). Activation of EGFR, cKIT, FMS and TRKA was determined detecting the levels of their autophosphorylation with specific antibody.

Figure 6.

TKI activity against FGFR1–4 and 14 other unrelated tyrosine kinases. Cell-free kinase assays were conducted using selected recombinant tyrosine kinases and recombinant STAT1 as the substrate, with TKIs added directly to the reaction mixtures. Kinase activity was determined by monitoring STAT1 phosphorylation (p) at Y701. EGFR, cKIT, FMS and TRKA activity was determined by detecting the levels of their autophosphorylation with an antibody specific to the phosphorylated form of the given kinase. Each experiment contains two controls (no inhibitor) and a sample with omitted ATP as a negative control. Quantities of total kinase and STAT1 serve as loading controls; in some samples, the amounts of recombinant proteins were determined by Sypro Ruby staining of acrylamide gels. The results are representative of three independent experiments, as listed in Table 1. TKIs were used at 20 nM (BGJ398), 50 nM (AZD4547, PD173074), 5 µM (AZD1480) and 10 µM (SU5402).

Figure 6.

TKI activity against FGFR1–4 and 14 other unrelated tyrosine kinases. Cell-free kinase assays were conducted using selected recombinant tyrosine kinases and recombinant STAT1 as the substrate, with TKIs added directly to the reaction mixtures. Kinase activity was determined by monitoring STAT1 phosphorylation (p) at Y701. EGFR, cKIT, FMS and TRKA activity was determined by detecting the levels of their autophosphorylation with an antibody specific to the phosphorylated form of the given kinase. Each experiment contains two controls (no inhibitor) and a sample with omitted ATP as a negative control. Quantities of total kinase and STAT1 serve as loading controls; in some samples, the amounts of recombinant proteins were determined by Sypro Ruby staining of acrylamide gels. The results are representative of three independent experiments, as listed in Table 1. TKIs were used at 20 nM (BGJ398), 50 nM (AZD4547, PD173074), 5 µM (AZD1480) and 10 µM (SU5402).

Cell-free kinase assays indicated that BGJ398 was a much weaker inhibitor of FGFR1, FGFR2 and FGFR4 activity than the other TKIs. However, the effects of this compound in the RCS growth arrest experiments were comparable with or greater than those of the other TKIs (Fig. 1). We therefore evaluated its activity against individual FGFRs in vitro, in 293T cells transfected with FGFR1–4. BGJ398 inhibited activity of both wt FGFR1 and FGFR2 and their mutants associated with osteoglophonic dysplasia (FGFR1-Y374C), or Pfeiffer, Crouzon and Jackson-Weiss craniosynostoses (FGFR2-C342R) (Fig. 7B–E) (34,35). BGJ398 activity against FGFR3 and FGFR4 could not be examined because wt FGFR3 and FGFR4 did not activate after transfection into the 293 T cells (Fig. 7A). The FGFR3-K650M mutant (which is associated with TD, bladder cancer and plasma cell myeloma) was only partially sensitive to BGJ398, whereas the rhabdomyosarcoma-associated FGFR4-V550E mutant was completely insensitive (Fig. 7F and G) (36,37). As FGFR3-K650M is a strongly activating mutation, and FGFR4-V550E targets a gatekeeper residue that controls TKI access to the ATP-binding pocket (38,39), the resistance of these two mutants to BGJ398 cannot be used to estimate this inhibitor's activity against FGFR3 and FGFR4. The evidence obtained here is thus only sufficient to support the conclusion that BGJ398 is an effective FGFR1 and FGFR2 inhibitor in vitro and that its low activity in kinase assays (Fig. 6; Table 1) is an artifact of the method.

Figure 7.

BGJ398 activity against FGFR1–4 and their activating mutants. (A) 293T cells were transfected with vectors carrying C-terminally V5-tagged wt FGFR1–4 or their activated mutants, and the levels of FGFR activation were determined by WB with antibody that binds to FGFRs autophosphorylated (p) at Y653/Y654. The levels of total FGFRs and actin serve as loading controls. Note the spontaneous, ligand-independent activation of wt FGFR1 and FGFR2 in contrast to FGFR3 and FGFR4. (B–G) FGFRs were expressed in 293T cells that had been treated with BGJ398 for 18 h, and the levels of FGFR and ERK phosphorylation (p) were determined. Total FGFR and ERK levels are used as loading controls. Note the suppression of wt and mutant FGFR1 and FGFR2 activation by 20 nM BGJ398 (B–E) in contrast to full FGFR3-K650E inhibition, which required 200 nM BGJ398 (F); the activity of FGFR4-V550E was insensitive to BGJ398 (G).

Figure 7.

BGJ398 activity against FGFR1–4 and their activating mutants. (A) 293T cells were transfected with vectors carrying C-terminally V5-tagged wt FGFR1–4 or their activated mutants, and the levels of FGFR activation were determined by WB with antibody that binds to FGFRs autophosphorylated (p) at Y653/Y654. The levels of total FGFRs and actin serve as loading controls. Note the spontaneous, ligand-independent activation of wt FGFR1 and FGFR2 in contrast to FGFR3 and FGFR4. (B–G) FGFRs were expressed in 293T cells that had been treated with BGJ398 for 18 h, and the levels of FGFR and ERK phosphorylation (p) were determined. Total FGFR and ERK levels are used as loading controls. Note the suppression of wt and mutant FGFR1 and FGFR2 activation by 20 nM BGJ398 (B–E) in contrast to full FGFR3-K650E inhibition, which required 200 nM BGJ398 (F); the activity of FGFR4-V550E was insensitive to BGJ398 (G).

TKIs are not selective for FGFR1–4 in vitro or in vivo

We next evaluated the TKI effects on FGFR signaling in two other models relevant to bone development: primary cultures of mouse limb bud mesenchymal cells with chondrocyte differentiation induced via micromass culture, and chicken wing buds treated with TKIs and left to develop into wings in vivo (40). Micromass culture recapitulates the early stages of limb formation, including mesenchymal condensation, chondrocyte differentiation and cartilage nodule formation, which is observed after 7 days of culture (41). FGF signaling plays a central role in cartilaginous nodule formation due to its proliferative effects on limb bud mesenchymal cells (42). Treatment with AZD4547 significantly inhibited cartilaginous nodule formation in E12 mouse limb buds after 7 days of mesenchymal differentiation (Fig. 8A and B).

Figure 8.

TKIs inhibit FGFR signaling in different in vitro and in vivo models irrespectively of their FGFR outfit. (A) Micromass cultures established from E12 mouse limb buds were treated with AZD4547 for 7 days and stained with alcian blue to visualize cartilaginous nodules. Signal intensity is quantified and graphed (B). Scale bar 1 mm, statistically significant differences are highlighted (t-test, **P < 0.01). (C) Expression of Fgfr1–4 in mouse micromasses was quantified by qRT-PCR. (D–I) Limb buds of chicken embryos were injected with AZD4547 and analyzed for external (J) and skeletal (K) morphology 10–12 days later. Right wings (D, E) and left wings (D′,E′) of embryos treated with DMSO alone exhibited normal morphology. 1 mm AZD4547 caused the abolition of radii (F′) or a reduction in the sizes of all distal structures (G′). 10 mm AZD4547 abolished either distal structures (H′) or the entire wing skeleton (I′). (J) Treatment with 1 or 10 mm AZD4547 caused most of the embryo's limbs to develop abnormally, with the external phenotype ranging from shortening to a complete absence of the limb. (K) Detailed skeletal analysis revealed major morphological changes to the zeugopod and autopod bones of limb buds injected with 1 mm AZD4547 and total reduction of distal skeleton with significant changes on proximal wing bones at 10 mm AZD4547 (R, reduced; A, absent). (L) 293T and NIH3T3 cells were analyzed for FGFR1–4 expression by WB. Note the expression of FGFR1 and FGFR2 in NIH3T3 cells and FGFR1–4 in 293T cells. Actin is used as a loading control. (M) Treatment with AZD4547 or BGJ398 suppresses FGF2- (NIH3T3, 293T) or FGF22-mediated (MCF7) activation of ERK MAP kinase, detected by WB using an antibody against ERK phosphorylated at the activating T202/Y204 residues (pERK). Total ERK levels are used as a loading control.

Figure 8.

TKIs inhibit FGFR signaling in different in vitro and in vivo models irrespectively of their FGFR outfit. (A) Micromass cultures established from E12 mouse limb buds were treated with AZD4547 for 7 days and stained with alcian blue to visualize cartilaginous nodules. Signal intensity is quantified and graphed (B). Scale bar 1 mm, statistically significant differences are highlighted (t-test, **P < 0.01). (C) Expression of Fgfr1–4 in mouse micromasses was quantified by qRT-PCR. (D–I) Limb buds of chicken embryos were injected with AZD4547 and analyzed for external (J) and skeletal (K) morphology 10–12 days later. Right wings (D, E) and left wings (D′,E′) of embryos treated with DMSO alone exhibited normal morphology. 1 mm AZD4547 caused the abolition of radii (F′) or a reduction in the sizes of all distal structures (G′). 10 mm AZD4547 abolished either distal structures (H′) or the entire wing skeleton (I′). (J) Treatment with 1 or 10 mm AZD4547 caused most of the embryo's limbs to develop abnormally, with the external phenotype ranging from shortening to a complete absence of the limb. (K) Detailed skeletal analysis revealed major morphological changes to the zeugopod and autopod bones of limb buds injected with 1 mm AZD4547 and total reduction of distal skeleton with significant changes on proximal wing bones at 10 mm AZD4547 (R, reduced; A, absent). (L) 293T and NIH3T3 cells were analyzed for FGFR1–4 expression by WB. Note the expression of FGFR1 and FGFR2 in NIH3T3 cells and FGFR1–4 in 293T cells. Actin is used as a loading control. (M) Treatment with AZD4547 or BGJ398 suppresses FGF2- (NIH3T3, 293T) or FGF22-mediated (MCF7) activation of ERK MAP kinase, detected by WB using an antibody against ERK phosphorylated at the activating T202/Y204 residues (pERK). Total ERK levels are used as a loading control.

We also injected AZD4547 into the right wing bud of chicken embryos at stages HH20 and 21 and analyzed its effect on the external and skeletal phenotypes 10–12 days later (Fig. 8D–K). Wings treated with the AZD4547 vehicle (DMSO) and untreated wings exhibited normal morphology (Fig. 8D, D′, E and E′). Treatment with 50 or 250 µM AZD4547 produced no external phenotype, but treatment with 1 mm AZD4547 induced wing shortening, stump formation or total suppression of wing formation (Fig. 8J). Detailed skeletal analysis revealed morphological changes in the zeugopod and autopod bones, with abnormal phenotypes varying from the absence of radius and digitus allulae to reduction of all distal bones (Fig. 8F′, K and G′). These changes are further aggravated by 10 mm AZD4547, causing total reduction of distal skeletal components and major changes in proximal bones including the scapula and coracoid (Fig. 8H′, I′ and K).

The experimental models used in this work expressed various FGFRs. RCS chondrocytes express FGFR3 and FGFR2 (43); limb growth plate cartilages mainly expressed FGFR3 with some FGFR1 expression (Fig. 4E), in agreement with previous reports (44–47); micromass cultures derived from mouse limb bud mesenchyme mainly expressed FGFR1 and FGFR2 (Fig. 8C); chick limb buds express FGFR1 in the mesenchyme at stages HH20 and 21, and FGFR2 in the ectodermal cell layer covering limb mesenchymal cells. FGFR3 and FGFR4 are weakly expressed in chicken limb buds during their early development (48), MCF7 cells express FGFR1 and FGFR2 (43), 293T cells express FGFR1-4, and NIH3T3 cells express FGFR1 and FGFR2 (Fig. 8L). Despite significant differences in their FGFR outfit, all seven experimental models used here responded rather uniformly to treatment with AZD4547 and/or BGJ398, showing complete inhibition of FGFR signaling. The only exception was FGF2-mediated ERK MAP kinase activation in NIH3T3 cells, which was less inhibited by AZD4547 than by BGJ398 (Fig. 8M). Thus, while some TKIs may exhibit modest selectivity for individual FGFRs in cell-free kinase assays, no meaningful selectivity for specific FGFRs is shown in cells; all of the tested TKIs completely suppressed signaling mediated by different FGFRs.

FGFR TKIs represent poor therapeutic options for ACH and other short-stature syndromes

ACH therapy must satisfy several challenging criteria. First, it must be specific to the FGFR3 signaling in the cartilage. Since promoting chondrocyte proliferation and restoring hypertrophic differentiation are the two major objectives of ACH therapy, future ACH drugs should be highly specific for FGFR3 or components of its downstream signaling pathway in cartilage, even in the presence of closely related pro-growth pathways driven by receptor tyrosine kinases such as IGF1R (49); any interference with other processes vital for chondrocyte proliferation or differentiation will preclude successful therapy. Second, ACH therapy must be selective for FGFR3 over the other FGFRs. While FGFR3 signaling appears to be mainly limited to the growth plate cartilage, other FGFRs signal ubiquitously and regulate a plethora of metabolic, paracrine and autocrine processes in mammals. ACH treatment must therefore target FGFR3 signaling but not that of other FGFRs in order to avoid adverse effects in other tissues. The issue of specificity toward FGFR3 is even more important given that it may be necessary to administer high concentrations of an ACH drug to inhibit FGFR3 signaling in the growth plate cartilage, which is avascular and composed of a dense extracellular matrix. Third, the ACH therapy will be administered for extended period of postnatal growth and must therefore remain effective for years in order to achieve significant improvement in overall skeletal size. Resistance similar to that observed in cancer patients chronically treated with TKIs is unlikely to develop in ACH because FGFR3 inhibition provides growth advantages to chondrocytes. However, metabolic, hormonal and other changes associated with postnatal growth may complicate the drug's mechanism of action or affect responsiveness to the therapy during the prolonged treatment. Fourth, the ACH therapy must be without side effects. Because the ACH therapy will be delivered to growing children, with the key clinical outcome being an increase in stature, adverse effects during treatment or in the years after the therapy has been completed cannot be tolerated.

We conclude that FGFR TKI evaluated here are poor candidates for ACH therapy, mainly because of their lack of specificity for FGFR3 over other FGFRs, and their cell toxicity. These properties are similarly evident in both early- and late-generation FGFR TKIs, suggesting that further exploration of this TKI class (i.e. small molecules targeting the ATP-binding site of FGFR) is unlikely to yield drugs suitable for ACH therapy. New efforts in this area should thus be directed toward developing conceptually different methods for targeting FGFR3 signaling in short-stature syndromes.

Materials and Methods

Cell, micromass and limb organ culture

Cells were propagated in Dulbecco's modified Eagle's medium (DMEM), supplemented with 10% fetal bovine serum (FBS) and antibiotics (Invitrogen, Carlsbad, CA). For RCS growth assays, 2.5 × 102 cells per well were grown in 96-well tissue culture plates for 5 days, and cell numbers were determined by crystal violet staining according to a protocol described in detail elsewhere (27). The chemicals were obtained from the following manufacturers: FGF2 and FGF22 (R&D Systems, Minneapolis, MN); heparin (Sigma-Aldrich, St. Louis, MO); SU5402 and PD173074 (Tocris Bioscience, Bristol, UK); and AZD1480, AZD4547 and BGJ398 (Selleckchem, Houston, TX). Primary mesenchymal cultures were established from the forelimb buds of E12 mouse embryos after proteolytic digestion with dispase II (Sigma). Cells were spotted as 10 µl aliquots at 2 × 107 cells/ml and left to adhere for 1 h before differentiating media (60% F12/40% DMEM, 10% FBS, 50 µg/ml ascorbic acid, 10 mm β-glycerol phosphate) was added. Micromasses were cultivated for 7 days at 37°C in supplemented media that was replaced with fresh media every other day. Tibias were dissected out from mouse embryos at E18, placed on Millipore filters above a metal mesh and cultured in micromass differentiating media for 8 days at 37°C with daily media change. Tibias were photographed and measured at the beginning and end of each experiment. Their lengths were measured using Axio Vision (Zeiss, Germany). Statistical analyses were performed in Statistica 8.0 (StatSoft, USA), and ANOVA and Turkey's post hoc test were used to evaluate the significance of observed differences between individual treatments.

Western blotting, kinase assays and transfection

Cells were lysed in buffer containing 50 mm Tris–HCl pH 7.4, 150 mm NaCl, 0.5% NP-40, 1 mm EDTA and 25 mm NaF, supplemented with proteinase inhibitors. Samples were resolved by sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE), transferred onto a polyvinylidene fluoride membrane and visualized by chemiluminescence (Thermo Scientific, Rockford, IL). The following antibodies were used: lamin A/C, caveolin, ID2, LRP6, pFRS2Y436, pMEKS217/221, MEK, pERKT202/Y204, ERK, FGFRY653/Y654, AXL, pcKITY703, DDR2, EGFR, FLT3, FMS, IGF1R, JAK3, STAT1, pSTAT1Y701, TRKA and FGFR4 (Cell Signaling, Beverly, MA); pLRP6T1572 and pTyr (4G10) (Millipore, Billerica, MA); FRS2, actin, FGFR2 and FGFR3 (Santa Cruz Biotechnology, Santa Cruz, CA); V5 (Invitrogen); and FGFR1 (Sigma-Aldrich). Kinase assays were performed using 200 ng of recombinant kinases (FGFR1-4, AXL, TYRO3, DDR2, EGFR, IGF1R, INSR, MET, cKIT, FLT3, FMS, TRKA, FLT4, ABL, JAK3) together with recombinant STAT1 (SignalChem, Richmond, CA) as a substrate, in 50 µl of kinase buffer (60 mm 4-(2-hydroxyethyl)-1-piperazineethanesulfonic acid pH 7.5, 3 mm MgCl2, 3 mm MnCl2, 3 µM Na3VO4, 1.2 mm DTT) in the presence of 10 µM ATP for 60 min at 30°C. Kinase activity was determined by monitoring STAT1 phosphorylation or kinase autophosphorylation. The western blotting (WB) signal was quantified by densitometry using the ImageJ software package (National Institutes of Health, Bethesda, USA). Cells were transfected using the FuGENE6 reagent according to the manufacturer's protocol (Roche). Vectors (pcDNA3.1) expressing V5-tagged human FGFR1-4 enzymes and their activating mutants were generated as described previously (50).

Growth plate histology, collagen type 10 in situ hybridization and reverse transcriptase-polymerase chain reaction

Paraformaldehyde-fixed and paraffin-embedded tibias were sectioned at a thickness of 5 μm and stained with hematoxylin–eosin (HE) for morphological analysis. For in situ hybridization, the Col10a1 plasmid (IMAGp998B1114092Q, Biovalley, France) was linearized by polymerase chain reaction (PCR) with M13 primers. Sense and antisense digoxigenin (DIG)-labeled riboprobes were produced using the SP6 and T7 RNA polymerases, respectively. Hybridization was performed at 60°C overnight. The sense DIG-labeled probe was used as a negative control. In situ hybridization on tibia sections was performed as described before (51). Total RNA was isolated from micromass cultures and tibias (cleared of bone and soft tissues) using the Mini RNeasy Kit according to the manufacturer's protocol (Qiagen). Reverse transcription was performed with 200 ng of isolated total RNA using a SuperScript Vilo cDNA synthesis Kit (Life Technologies). Quantitative PCR (qPCR) was performed with Taqman Universal PCR Master Mix using probes obtained from Life Technologies: FGFR1 (Mm00438930_m1), FGFR2 (Mm01269930_m1), FGFR3 (Mm00433294_m1), FGFR4 (Mm01341852_m1) as well as for β-actin (ACTB, Mm00607939_s1). ΔCT was calculated, and the resulting values were normalized as described before (52,53).

AZD4547 injection into chicken limb buds and postnatal mice

Fertilized chicken eggs (ISA brown) were obtained from Integra (Zabcice, Czech Republic). Eggs were incubated in a humidified forced air incubator at 37.8°C, and embryos were staged according to Hamburger and Hamilton (54). AZD4547 was injected into the right wing bud using a micromanipulator (Leica, Germany) and microinjector (Eppendorf, Germany) to better target the selected area of application. Three injection sites (anterior, posterior, distal) were used. Left wings were treated with the TKI vehicle (DMSO). Embryos were collected after 10–12 days of incubation, fixed in ethanol, stained with alizarin red/alcian blue solution and cleared in KOH/glycerol. Mice were euthanized by cervical dislocation in accordance with the AVMA Guidelines on Euthanasia. Skeletal preparations were made using a procedure similar to that used for the chicken embryos. Blood tests were performed by an accredited clinical laboratory (CSN EN ISO 15189:2013) at the University Hospital, Palacky University, Olomouc, according to established procedures. All animal experiments were reviewed and approved by the Institutional Animal Care and Use Committee at the Institute of Animal Physiology and Genetics ASCR, Brno, Czech Republic (No. 144/2013).

Supplementary Material

Supplementary Material is available at HMG online.

Funding

This work was supported by the Ministry of Education, Youth and Sports of the Czech Republic (KONTAKT LH12004, CZ.1.074/2.3.00/30.0053, CZ.1.05/3.1.00/14.0324), the Czech Science Foundation (14-31540S), Grant Agency of Masaryk University (0071-2013), the European Regional Development Fund (FNUSA-ICRC No. CZ.1.05/1.1.00/02.0123) and the European Union (ICRC ERA-HumanBridge No. 316345). M.K. was supported by the European Social Fund and the state budget of the Czech Republic (CZ.1.07/2.3.00/30.0009). L.R. was supported by National Sustainability Program I (LO1304).

Acknowledgements

We thank Miriam Minarikova for excellent technical support and Jirina Medalova for help with manuscript preparation.

Conflict of Interest statement. None declared.

References

1
Provot
S.
,
Schipani
E.
(
2005
)
Molecular mechanisms of endochondral bone development
.
Biochem. Biophys. Res. Commun.
 ,
328
,
658
665
.
2
Kronenberg
H.M.
(
2006
)
PTHrP and skeletal development
.
Ann. N. Y. Acad. Sci.
 ,
1068
,
1
13
.
3
Pogue
R.
,
Lyons
K.
(
2006
)
BMP signaling in the cartilage growth plate
.
Curr. Top. Dev. Biol.
 ,
76
,
1
48
.
4
Pejchalova
K.
,
Krejci
P.
,
Wilcox
W.R.
(
2007
)
C-natriuretic peptide: an important regulator of cartilage
.
Mol. Genet. Metab.
 ,
92
,
210
215
.
5
Dagoneau
N.
,
Scheffer
D.
,
Huber
C.
,
Al-Gazali
L.I.
,
Di Rocco
M.
,
Godard
A.
,
Martinovic
J.
,
Raas-Rothschild
A.
,
Sigaudy
S.
,
Unger
S.
et al
. (
2004
)
Null leukemia inhibitory factor receptor (LIFR) mutations in Stuve-Wiedemann/Schwartz-Jampel type 2 syndrome
.
Am. J. Hum. Genet.
 ,
74
,
298
305
.
6
Colvin
J.S.
,
Bohne
B.A.
,
Harding
G.W.
,
McEwen
D.G.
,
Ornitz
D.M.
(
1996
)
Skeletal overgrowth and deafness in mice lacking fibroblast growth factor receptor 3
.
Nat. Genet.
 ,
12
,
390
397
.
7
Toydemir
R.M.
,
Brassington
A.E.
,
Bayrak-Toydemir
P.
,
Krakowiak
P.A.
,
Jorde
L.B.
,
Whitby
F.G.
,
Longo
N.
,
Viskochil
D.H.
,
Carey
J.C.
,
Bamshad
M.J.
(
2006
)
A novel mutation in FGFR3 causes camptodactyly, tall stature, and hearing loss (CATSHL) syndrome
.
Am. J. Hum. Genet.
 ,
79
,
935
941
.
8
Chen
L.
,
Adar
R.
,
Yang
X.
,
Monsonego
E.O.
,
Li
C.
,
Hauschka
P.V.
,
Yayon
A.
,
Deng
C.X.
(
1999
)
Gly369Cys mutation in mouse FGFR3 causes achondroplasia by affecting both chondrogenesis and osteogenesis
.
J. Clin. Invest.
 ,
104
,
1517
1525
.
9
Murakami
S.
,
Balmes
G.
,
McKinney
S.
,
Zhang
Z.
,
Givol
D.
,
de Crombrugghe
B.
(
2004
)
Constitutive activation of MEK1 in chondrocytes causes Stat1-independent achondroplasia-like dwarfism and rescues the Fgfr3-deficient mouse phenotype
.
Genes Dev.
 ,
18
,
290
305
.
10
Parker
H.G.
,
VonHoldt
B.M.
,
Quignon
P.
,
Margulies
E.H.
,
Shao
S.
,
Mosher
D.S.
,
Spady
T.C.
,
Elkahloun
A.
,
Cargill
M.
,
Jones
P.G.
et al
. (
2009
)
An expressed fgf4 retrogene is associated with breed-defining chondrodysplasia in domestic dogs
.
Science
 ,
325
,
995
998
.
11
Foldynova-Trantirkova
S.
,
Wilcox
W.R.
,
Krejci
P.
(
2012
)
Sixteen years and counting: the current understanding of fibroblast growth factor receptor 3 (FGFR3) signaling in skeletal dysplasias
.
Hum. Mutat.
 ,
33
,
29
41
.
12
Horton
W.A.
,
Hall
J.G.
,
Hecht
J.T.
(
2007
)
Achondroplasia
.
Lancet
 ,
370
,
162
172
.
13
Druker
B.J.
,
Tamura
S.
,
Buchdunger
E.
,
Ohno
S.
,
Segal
G.M.
,
Fanning
S.
,
Zimmermann
J.
,
Lydon
N.B.
(
1996
)
Effects of a selective inhibitor of the Abl tyrosine kinase on the growth of Bcr-Abl positive cells
.
Nat. Med.
 ,
2
,
561
566
.
14
Wesche
J.
,
Haglund
K.
,
Haugsten
E.M.
(
2011
)
Fibroblast growth factors and their receptors in cancer
.
Biochem. J.
 ,
437
,
199
213
.
15
Mohammadi
M.
,
McMahon
G.
,
Sun
L.
,
Tang
C.
,
Hirth
P.
,
Yeh
B.K.
,
Hubbard
S.R.
,
Schlessinger
J.
(
1997
)
Structures of the tyrosine kinase domain of fibroblast growth factor receptor in complex with inhibitors
.
Science
 ,
276
,
955
960
.
16
Mohammadi
M.
,
Froum
S.
,
Hamby
J.M.
,
Schroeder
M.C.
,
Panek
R.L.
,
Lu
G.H.
,
Eliseenkova
A.V.
,
Green
D.
,
Schlessinger
J.
,
Hubbard
S.R.
(
1998
)
Crystal structure of an angiogenesis inhibitor bound to the FGF receptor tyrosine kinase domain
.
EMBO J.
 ,
17
,
5896
5904
.
17
Hamby
J.M.
,
Connolly
C.J.
,
Schroeder
M.C.
,
Winters
R.T.
,
Showalter
H.D.
,
Panek
R.L.
,
Major
T.C.
,
Olsewski
B.
,
Ryan
M.J.
,
Dahring
T.
et al
. (
1997
)
Structure-activity relationships for a novel series of pyrido[2,3-d]pyrimidine tyrosine kinase inhibitors
.
J. Med. Chem.
 ,
40
,
2296
2303
.
18
Gavine
P.R.
,
Mooney
L.
,
Kilgour
E.
,
Thomas
A.P.
,
Al-Kadhimi
K.
,
Beck
S.
,
Rooney
C.
,
Coleman
T.
,
Baker
D.
,
Mellor
M.J.
et al
. (
2012
)
AZD4547: an orally bioavailable, potent, and selective inhibitor of the fibroblast growth factor receptor tyrosine kinase family
.
Cancer Res.
 ,
72
,
2045
2056
.
19
Guagnano
V.
,
Furet
P.
,
Spanka
C.
,
Bordas
V.
,
Le Douget
M.
,
Stamm
C.
,
Brueggen
J.
,
Jensen
M.R.
,
Schnell
C.
,
Schmid
H.
et al
. (
2011
)
Discovery of 3-(2,6-dichloro-3,5-dimethoxy-phenyl)-1-{6-[4-(4-ethyl-piperazin-1-yl)-phenylamino]-pyrimidin-4-yl}-1-methyl-urea (NVP-BGJ398), a potent and selective inhibitor of the fibroblast growth factor receptor family of receptor tyrosine kinase
.
J. Med. Chem.
 ,
54
,
7066
7083
.
20
Liang
G.
,
Chen
G.
,
Wei
X.
,
Zhao
Y.
,
Li
X.
(
2013
)
Small molecule inhibition of fibroblast growth factor receptors in cancer
.
Cytokine Growth Factor Rev.
 ,
24
,
467
475
.
21
Carter
E.P.
,
Fearon
A.E.
,
Grose
R.P.
(
2014
)
Careless talk costs lives: fibroblast growth factor receptor signalling and the consequences of pathway malfunction
.
Trends Cell Biol.
 ,
25
,
221
33
.
22
Laederich
M.B.
,
Horton
W.A.
(
2010
)
Achondroplasia: pathogenesis and implications for future treatment
.
Curr. Opin. Pediatr.
 ,
22
,
516
523
.
23
Jonquoy
A.
,
Mugniery
E.
,
Benoist-Lasselin
C.
,
Kaci
N.
,
Le Corre
L.
,
Barbault
F.
,
Girard
A.L.
,
Le Merrer
Y.
,
Busca
P.
,
Schibler
L.
et al
. (
2012
)
A novel tyrosine kinase inhibitor restores chondrocyte differentiation and promotes bone growth in a gain-of-function Fgfr3 mouse model
.
Hum. Mol. Genet.
 ,
21
,
841
851
.
24
Mukhopadhyay
K.
,
Lefebvre
V.
,
Zhou
G.
,
Garofalo
S.
,
Kimura
J.H.
,
de Crombrugghe
B.
(
1995
)
Use of a new rat chondrosarcoma cell line to delineate a 119-base pair chondrocyte-specific enhancer element and to define active promoter segments in the mouse pro-alpha 1(II) collagen gene
.
J. Biol. Chem.
 ,
270
,
27711
27719
.
25
Krejci
P.
,
Salazar
L.
,
Goodridge
H.S.
,
Kashiwada
T.A.
,
Schibler
M.J.
,
Jelinkova
P.
,
Thompson
L.M.
,
Wilcox
W.R.
(
2008
)
STAT1 and STAT3 do not participate in FGF-mediated growth arrest in chondrocytes
.
J. Cell. Sci.
 ,
121
,
272
281
.
26
Krejci
P.
,
Prochazkova
J.
,
Smutny
J.
,
Chlebova
K.
,
Lin
P.
,
Aklian
A.
,
Bryja
V.
,
Kozubik
A.
,
Wilcox
W.R.
(
2010
)
FGFR3 signaling induces a reversible senescence phenotype in chondrocytes similar to oncogene-induced premature senescence
.
Bone
 ,
47
,
102
110
.
27
Krejci
P.
,
Pejchalova
K.
,
Wilcox
W.R.
(
2007
)
Simple, mammalian cell-based assay for identification of inhibitors of the Erk MAP kinase pathway
.
Invest. New. Drugs
 ,
25
,
391
395
.
28
Krejci
P.
,
Bryja
V.
,
Pachernik
J.
,
Hampl
A.
,
Pogue
R.
,
Mekikian
P.
,
Wilcox
W.R.
(
2004
)
FGF2 inhibits proliferation and alters the cartilage-like phenotype of RCS cells
.
Exp. Cell. Res.
 ,
297
,
152
164
.
29
Buchtova
M.
,
Oralova
V.
,
Aklian
A.
,
Masek
J.
,
Vesela
I.
,
Ouyang
Z.
,
Obadalova
T.
,
Konecna
Z.
,
Spoustova
T.
,
Pospisilova
T.
et al
. (
2015
)
Fibroblast growth factor and canonical WNT/β-catenin signaling cooperate in suppression of chondrocyte differentiation in experimental models of FGFR signaling in cartilage
.
Biochim. Biophys. Acta
 ,
1852
,
839
850
.
30
Krejci
P.
,
Masri
B.
,
Salazar
L.
,
Farrington-Rock
C.
,
Prats
H.
,
Thompson
L.M.
,
Wilcox
W.R.
(
2007
)
Bisindolylmaleimide I suppresses fibroblast growth factor-mediated activation of Erk MAP kinase in chondrocytes by preventing Shp2 association with the Frs2 and Gab1 adaptor proteins
.
J. Biol. Chem.
 ,
282
,
2929
2936
.
31
Sahni
M.
,
Ambrosetti
D.C.
,
Mansukhani
A.
,
Gertner
R.
,
Levy
D.
,
Basilico
C.
(
1999
)
FGF signaling inhibits chondrocyte proliferation and regulates bone development through the STAT-1 pathway
.
Genes. Dev.
 ,
13
,
1361
1366
.
32
Wendt
D.J.
,
Dvorak-Ewell
M.
,
Bullens
S.
,
Lorget
F.
,
Bell
S.M.
,
Peng
J.
,
Castillo
S.
,
Aoyagi-Scharber
M.
,
O'Neill
C.A.
,
Krejci
P.
et al
. (
2015
)
Neutral endopeptidase resistant C-type natriuretic peptide (CNP) variant represents a new therapeutic approach for treatment of fibroblast growth factor receptor 3-related dwarfism
.
J. Pharmacol. Exp. Ther.
 ,
353
,
132
149
.
33
Le Corre
L.
,
Girard
A.L.
,
Aubertin
J.
,
Radvanyi
F.
,
Benoist-Lasselin
C.
,
Jonquoy
A.
,
Mugniery
E.
,
Legeai-Mallet
L.
,
Busca
P.
,
Le Merrer
Y.
(
2010
)
Synthesis and biological evaluation of a triazole-based library of pyrido[2,3-d]pyrimidines as FGFR3 tyrosine kinase inhibitors
.
Org. Biomol. Chem.
 ,
8
,
2164
2173
.
34
White
K.E.
,
Cabral
J.M.
,
Davis
S.I.
,
Fishburn
T.
,
Evans
W.E.
,
Ichikawa
S.
,
Fields
J.
,
Yu
X.
,
Shaw
N.J.
,
McLellan
N.J.
et al
. (
2005
)
Mutations that cause osteoglophonic dysplasia define novel roles for FGFR1 in bone elongation
.
Am. J. Hum. Genet.
 ,
76
,
361
367
.
35
Passos-Bueno
M.R.
,
Wilcox
W.R.
,
Jabs
E.W.
,
Sertié
A.L.
,
Alonso
L.G.
,
Kitoh
H.
(
1999
)
Clinical spectrum of fibroblast growth factor receptor mutations
.
Hum. Mutat.
 ,
14
,
115
125
.
36
Bamford
S.
,
Dawson
E.
,
Forbes
S.
,
Clements
J.
,
Pettett
R.
,
Dogan
A.
,
Flanagan
A.
,
Teague
J.
,
Futreal
P.A.
,
Stratton
M.R.
et al
. (
2004
)
The COSMIC (Catalogue of Somatic Mutations in Cancer) database and website
.
Br. J. Cancer
 ,
91
,
355
358
.
37
Taylor
J.G.
VI
,
Cheuk
A.T.
,
Tsang
P.S.
,
Chung
J.Y.
,
Song
Y.K.
,
Desai
K.
,
Yu
Y.
,
Chen
Q.R.
,
Shah
K.
,
Youngblood
V.
et al
. (
2009
)
Identification of FGFR4-activating mutations in human rhabdomyosarcomas that promote metastasis in xenotransplanted models
.
J. Clin. Invest.
 ,
119
,
3395
3407
.
38
Krejci
P.
,
Salazar
L.
,
Kashiwada
T.A.
,
Chlebova
K.
,
Salasova
A.
,
Thompson
L.M.
,
Bryja
V.
,
Kozubik
A.
,
Wilcox
W.R.
(
2008
)
Analysis of STAT1 activation by six FGFR3 mutants associated with skeletal dysplasia undermines dominant role of STAT1 in FGFR3 signaling in cartilage
.
PLoS One
 ,
3
,
e3961
.
39
Huang
Z.
,
Tan
L.
,
Wang
H.
,
Liu
Y.
,
Blais
S.
,
Deng
J.
,
Neubert
T.A.
,
Gray
N.S.
,
Li
X.
,
Mohammadi
M.
(
2015
)
DFG-out mode of inhibition by an irreversible type-1 inhibitor capable of overcoming gate-keeper mutations in FGF receptors
.
ACS Chem. Biol.
 ,
10
,
299
309
.
40
Horakova
D.
,
Cela
P.
,
Krejci
P.
,
Balek
L.
,
Moravcova Balkova
S.
,
Matalova
E.
,
Buchtova
M.
(
2014
)
Effect of FGFR inhibitors on chicken limb development
.
Dev. Growth Differ.
 ,
56
,
555
572
.
41
Daniels
K.
,
Reiter
R.
,
Solursh
M.
(
1996
)
Micromass cultures of limb and other mesenchyme
.
Methods Cell. Biol.
 ,
51
,
237
247
.
42
Kim
J.H.
,
Lee
M.C.
,
Seong
S.C.
,
Park
K.H.
,
Lee
S.
(
2011
)
Enhanced proliferation and chondrogenic differentiation of human synovium-derived stem cells expanded with basic fibroblast growth factor
.
Tissue Eng. Part A
 ,
17
,
991
1002
.
43
Buchtova
M.
,
Chaloupkova
R.
,
Zakrzewska
M.
,
Vesela
I.
,
Cela
P.
,
Zajickova
R.
,
Trantirek
L.
,
Martin
J.
,
Kostas
M.
,
Otlewski
J.
et al
. (
2015
)
Instability Restricts Signaling of Multiple Fibroblast Growth Factors
.
Cell. Mol. Life Sci.
 ,
72
,
2445
59
.
44
Deng
C.
,
Wynshaw-Boris
A.
,
Zhou
F.
,
Kuo
A.
,
Leder
P.
(
1996
)
Fibroblast growth factor receptor 3 is a negative regulator of bone growth
.
Cell
 ,
84
,
911
921
.
45
Peters
K.G.
,
Werner
S.
,
Chen
G.
,
Williams
L.T.
(
1992
)
Two FGF receptor genes are differentially expressed in epithelial and mesenchymal tissues during limb formation and organogenesis in the mouse
.
Development
 ,
114
,
233
243
.
46
Peters
K.
,
Ornitz
D.
,
Werner
S.
,
Williams
L.
(
1993
)
Unique expression pattern of the FGF receptor 3 gene during mouse organogenesis
.
Dev. Biol.
 ,
155
,
423
430
.
47
Delezoide
A.L.
,
Benoist-Lasselin
C.
,
Legeai-Mallet
L.
,
Le Merrer
M.
,
Munnich
A.
,
Vekemans
M.
,
Bonaventure
J.
(
1998
)
Spatio-temporal expression of FGFR 1, 2 and 3 genes during human embryo-fetal ossification
.
Mech. Dev.
 ,
77
,
19
30
.
48
Sheeba
C.J.
,
Andrade
R.P.
,
Duprez
D.
,
Palmeirim
I.
(
2010
)
Comprehensive analysis of fibroblast growth factor receptor expression patterns during chick forelimb development
.
Int. J. Dev. Biol.
 ,
54
,
1517
1526
.
49
Laron
Z.
(
2001
)
Insulin growth factor 1 (IGF-1): a growth hormone
.
Mol. Pathol.
 ,
54
,
311
316
.
50
Krejci
P.
,
Aklian
A.
,
Kaucka
M.
,
Sevcikova
E.
,
Prochazkova
J.
,
Masek
J.K.
,
Mikolka
P.
,
Pospisilova
T.
,
Spoustova
T.
,
Weis
M.
et al
. (
2012
)
Receptor tyrosine kinases activate canonical WNT/β-catenin signaling via MAP kinase/LRP6 pathway and direct β-catenin phosphorylation
.
PLoS One
 ,
7
,
e35826
.
51
Holland
L.Z.
,
Holland
P.W.H.
,
Holland
N.D.
(
1996
)
Revealing homologies between body parts of distantly related animals by in situ hybridization to developmental genes: amphioxus versus vertebrates
. In
Ferraris
J. D.
,
Palumbi
S. R.
(eds),
Molecular Zoology: Advances, Strategies, and Protocols
 .
New York
,
Wiley
, pp.
267
282
,
473–483
.
52
Bookout
A.L.
,
Mangelsdorf
D.J.
(
2003
)
Quantitative real-time PCR protocol for analysis of nuclear receptor signaling pathways
.
Nucl. Recept. Signal.
 ,
1
,
e012
.
53
Abramyan
J.
,
Wilhelm
D.
,
Koopman
P.
(
2010
)
Molecular characterization of the Bidder's organ in the cane toad (Bufo marinus)
.
J. Exp. Zool. B Mol. Dev. Evol.
 ,
314
,
503
513
.
54
Hamburger
V.
,
Hamilton
H.L.
(
1951
)
A series of normal stages in the development of the chick embryo
.
J. Morphol.
 ,
88
,
49
92
.

Author notes

These authors contributed to this work equally.