Fibroblast growth factor receptor 3 (FGFR3) plays important roles in cell proliferation, differentiation, and angiogenesis. FGFR3 is abnormally upregulated in hepatocellular carcinoma (HCC), where it correlates positively with clinicopathologic index, HCC differentiation, and advanced nuclear grade. In this study, we describe an aberrantly spliced transcript of FGFR3, termed FGFR3Δ7-9, was identified as a high frequency even in HCC. FGFR3Δ7-9 lacks exons encoding the immunoglobulin-like III domain and promoted the proliferation, migration, and metastasis of HCC cells both in vitro and in vivo. Coimmunoprecipation and surface plasmon resonance assays demonstrated that the binding affinity of the aberrant FGFR3Δ7-9 receptor to FGFs was significantly higher than wild-type FGFR3IIIc. Furthermore, FGFR3Δ7-9 could be self-activated by homodimerization and autophosphorylation even in the absence of ligand. Finally, FGFR3Δ7-9 more potently induced phosphorylation of the ERK and AKT kinases, leading to abnormal downstream signaling through the ERK and PI3K/AKT/mTOR pathways. FGFR3Δ7-9 also upregulated the metastasis-associated molecules Snail, MMP-9, and downregulated E-cadherin, which associated directly with FGFR3Δ7-9. Thus, as a ligand-dependent or -independent receptor, FGFR3Δ7-9 exerted multiple potent oncogenic functions in HCC cells, including proliferation, migration, and lung metastatic capacity. Overall, FGFR3 mRNA missplicing in HCC contributes significantly to its malignant character, with implications for therapeutic targeting. Cancer Res; 76(14); 4205–15. ©2016 AACR.

HCC is one of the most common cancers in the world, accounting for nearly half a million deaths worldwide (1). Our previous data indicated that the expression of FGFR3 was significantly upregulated in HCC, which was strongly correlated with clinicopathologic index, HCC differentiation, and advanced nuclear grade (2). Fibroblast growth factor receptor (FGFR) encode a set of transmembrane tyrosine kinase receptors and can be activated on binding to the extracellular domain by a number of related mitogenic fibroblast growth factors (3, 4). Previous investigations demonstrated that FGFRs might play important roles in a variety of processes, including cellular proliferation, differentiation, and angiogenesis (5–9).

The prototypical FGFR extracellular domain consists of three immunoglobulin (Ig) domains (D1–D3). Two membrane-proximal D2 and D3 domains, and interconnecting D2–D3 linker bear the determinants of ligand binding and specificity (10). The Ig-domain III is encoded by two separate exons, exon 8 and exon 9. Alternative usage of these exons determines the ligand-binding specificity of receptor, such as FGFRIIIb (exon 8) and FGFRIIIc (exon 9). This switch between alternatively spliced isoforms results in changes in FGFR signaling and promote carcinogenesis by the more oncogenic isoforms consequently (11–14). The disruption of regulated mRNA processing has emerged as a cellular function that could be disturbed during neoplastic transformation of cells (15–19). In this study, a novel aberrantly spliced variant of FGFR3, FGFR3Δ7-9, can be identified in HCC. This novel mutant transcript links exon 6 to exon 10 directly, which causes the deletion of the Ig-like-III domain. Although this aberrantly spliced transcript contains only parts of the ligand-binding domain, FGFR3Δ7-9 could be self-activated and autophosphorylated independent of FGFs binding. The aberrant ligand-binding affinity and signaling transduction pathways consequently determine significant changes in HCC biocharacters.

Patients, tissue specimens, and cell culture

Thirty-five sets of HCC tissues and 10 sets of normal liver tissues were collected from patients who underwent curative surgery in Ruijin Hospital. Human HCC cell lines HepG2 and human embryonic kidney cell 293T were purchased from ATCC. Human HCC cell lines Huh-7, SMMC-7721, QGY-7701, QGY7703, BEL-7402, BEL-7404, and BEL-7405, and normal human fetal liver–derived cell lines Chang liver and HL-7702 were also included in this study as described previously (20). After resuscitation, all cells were immediately expanded and frozen so that they could be restarted every 3 to 4 months from a frozen vial of the same batch of cells. The authors have not independently authenticated these cells. For details, see Supplementary Materials and Methods.

Nested RT-PCR and quantitative real-time PCR

The nested RT-PCR was employed to analyze the sequences of FGFR3. Steady-state FGFR3 mRNA expression was determined by real-time PCR using SYBR Premix Ex Taq (TaKaRa). More details about the primers and PCR reactions were described in Supplementary Materials and Methods.

Construction of recombinant plasmids of FGFR3IIIc, FGFR3Δ7-9, and FGFR3AT-I and transient transfection of cell lines

The sequences of FGFR3IIIc, FGFR3Δ7-9, and FGFR3AT-I were constructed in vitro by PCR-driven overlap extension. The establishment of lentivirus plasmids, pWPI.1-FGFR3IIIc, pWPI.1-FGFR3Δ7-9, and pWPI.1-FGFR3AT-I was described in Supplementary Materials and Methods.

For ligand-binding affinity assay, the sequences of FGF1 and FGF2 were amplified and then subcloned into the corresponding sites of pcDNA3-HA and pcNDA3-FLAG. To examine the affinity of alternative splicings of FGFR3 to FGFs, HepG2 cells were cotransfected with different FGFR3 expression plasmids (pcDNA3-HA-FGFR3IIIc, pcDNA3-HA-FGFR3Δ7-9, or pcDNA3-HA-FGFR3AT-I) together with pcDNA3-FLAG-FGF1 or pcDNA3-FLAG-FGF2 plasmid. To detect the difference between FGFR3 and its mutants in dimer formation and phosphorylation of tyrosine area, HepG2 was transfected by pcDNA3-FLAG-FGFR3IIIc, pcDNA3-FLAG-FGFR3Δ7-9, or pcDNA3-FLAG-FGFR3AT-I. For details, see Supplementary Materials and Methods.

Lentivirus packaging, virus transduction, fluorescent cell selection, and proliferation evaluation

More details were described in Supplementary Materials and Methods.

Cell migration assay

Cell migration ability of HCC cells expressing FGFR3 mutation was estimated by Transwell assay using Falcon Cell Culture Insert (BD353097). For details, see Supplementary Materials and Methods.

RNA interference

Two shRNAs against FGFR3Δ7-9 were synthesized to further understand the role of FGFR3Δ7-9 on HCC cells. For details, see Supplementary Materials and Methods.

In vivo tumor growth and metastases assay

HCC xenografts were established in nude mice. The stable clones of HepG2 cells with FGFR3IIIc and FGFR3Δ7-9 were subcutaneously injected into the right flank of nude mice for in vivo tumor growth assay. The tail vein injection assay was employed to assess the effect of FGFR3 and its mutants on tumor metastases. Tumor dimensions and the long-distance lung metastasis were measured using a digital caliper and micro-PET/CT. For details, see Supplementary Materials and Methods.

Immunoprecipitation and immunoblot analyses

Immunoprecipitation analyses were performed to detect the difference in the binding affinity between FGFR3s and their ligands in HCC cells. For details about immunoprecipitation and Western blot analysis, see Supplementary Materials and Methods.

Surface plasmon resonance binding analysis

Surface plasmon resonance (SPR) binding analysis was further employed to analyze the ligand affinity between FGFR3s and FGF1 quantitatively. For details, see Supplementary Materials and Methods.

Interacting proteins detection by mass spectrometry analysis

Mass spectrometry analysis was used to compare the difference of interacting proteins between FGFR3IIIc/FGFR3Δ7-9 and ligand. For details, see Supplementary Materials and Methods.

E-cadherin luciferase reporter assay

E-cadherin regulation by FGFR3 mutation was examined using E-cadherin promoter luciferase reporter assay. For details, see Supplementary Materials and Methods.

Statistical analyses

An ANOVA and Student's t test were used for comparison among groups. The Mann–Whitney U test was used for comparison of tumor volume. Categorical data was evaluated with χ2 test or Fisher exact test. A P value less than 0.05 was considered to be significant.

Novel RNA splicing variants of FGFR3 in HCC

The screening result of FGFR3 expression by nested RT-PCR assay was shown in Fig. 1A. The amplified products were separated by electrophoresis, and the abnormally migrating bands were purified and sequenced. Sequences of abnormally migrating bands revealed the presence of two distinct products in primary tumors, whereas 10 normal hepatic tissues showed normal-sized transcripts. In addition to FGFR3IIIb isoform (normal-sized transcript), sequence analyses of this distinct transcript revealed two abnormally spliced transcripts of aberrant cDNAs. The lower FGFR3 transcript was defined as FGFR3AT-I and was observed in 13 of 35 primary carcinomas (37.14%; Fig. 1A and B). The upper novel splicing variant of FGFR3 transcript was defined as FGFR3Δ7-9 and was observed in 4 of 35 cases (11.43%). FGFR3AT-I encoded a form of FGFR3 missing the second half of Ig-like-III domain, and exon 7 was spliced to exon 10, introducing a 1-base (−1) frameshift and resulting in a premature termination sequence at base 1217. In FGFR3Δ7-9, exon 6 was directly merged to exon 10 without frameshift, resulting in a missing of exon 7 to 9. The predicted open reading frame of FGFR3Δ7-9 coded an isoform of FGFR3 missing Ig-like-III domain.

Figure 1.

Analyses of aberrant FGFR3 transcripts in HCC and normal liver cell lines. A, expression of FGFR3 gene by nested RT-PCR amplification of mRNA from HCC and normal liver tissues. The primers span exon 6–10. The fragments in normal samples (lane N) and peri-neoplastic tissues (lane P) revealed an amplification of FGFR3IIIb (473 bp). The tumor samples (lane T) showed two other fragments (FGFR3Δ7-9, 131 bp; FGFR3AT-I, 322 bp) in addition to another wild-type fragment (FGFR3IIIc, 467 bp). B, sequence analyses of abnormal transcripts. Arrows indicate junctions between exon 6 and 10 in FGFR3Δ7-9 and junctions between exon 7 and 10 in FGFR3AT-I. C, schematic diagram of aberrant splicing. The location of exon 7–10 is shown with the schematic structure of FGFR3 (I–III, Ig-like domains; TM, transmembrane domain; Kinase, two tyrosine kinase domains). The lines connecting exons represent the splicing patterns used to produce the transcripts noted on the left.

Figure 1.

Analyses of aberrant FGFR3 transcripts in HCC and normal liver cell lines. A, expression of FGFR3 gene by nested RT-PCR amplification of mRNA from HCC and normal liver tissues. The primers span exon 6–10. The fragments in normal samples (lane N) and peri-neoplastic tissues (lane P) revealed an amplification of FGFR3IIIb (473 bp). The tumor samples (lane T) showed two other fragments (FGFR3Δ7-9, 131 bp; FGFR3AT-I, 322 bp) in addition to another wild-type fragment (FGFR3IIIc, 467 bp). B, sequence analyses of abnormal transcripts. Arrows indicate junctions between exon 6 and 10 in FGFR3Δ7-9 and junctions between exon 7 and 10 in FGFR3AT-I. C, schematic diagram of aberrant splicing. The location of exon 7–10 is shown with the schematic structure of FGFR3 (I–III, Ig-like domains; TM, transmembrane domain; Kinase, two tyrosine kinase domains). The lines connecting exons represent the splicing patterns used to produce the transcripts noted on the left.

Close modal

Effect on cellular proliferation of FGFR3 mutation

As shown in Fig. 2A and B, to minimize the interference from endogenous FGFR3, HepG2, and SMMC-7721, which expressed FGFR3 at a relatively low level, were selected to establish the cell models. The recombinant plasmid pWPI.1-FGFR3IIIc-GFP, pWPI.1-FGFR3Δ7-9-GFP, or pWPI.1-FGFR3AT-I-GFP was stably transfected into HepG2 and SMMC-7721, respectively. The results of CCK8 in Fig. 3A indicated that FGFR3Δ7-9 overexpression could promote cell proliferation ability in both HCC cell lines compared with FGFR3IIIc overexpression and empty vector control (HepG2 group at day 7: FGFR3Δ7-9: 1.17 ± 0.12, FGFR3IIIc: 0.94 ± 0.11, empty vector: 0.88 ± 0.12, FGFR3AT-I: 0.88 ± 0.12, FGFR3Δ7-9 vs. FGFR3IIIc: P < 0.05; SMMC-7721 group at day7: FGFR3Δ7-9: 1.12 ± 0.09, FGFR3IIIc: 0.95 ± 0.10, empty vector: 0.89 ± 0.09, FGFR3AT-I: 0.81 ± 0.08, FGFR3Δ7-9 vs. FGFR3IIIc: P < 0.05).

Figure 2.

Expression pattern of FGFR3 in HCC cell lines and identification of stable cell lines overexpressing FGFR3IIIc, FGFR3Δ7-9, and FGFR3AT-I. A, quantitative real-time RT-PCR was performed to detect the expression level of FGFR3 in HCC cell lines with GAPDH as an internal control; B, Western blot assay was performed to evaluate FGFR3 expression level in HCC cell lines, and GAPDH was used as loading control; C, Western blot analysis of FGFR3IIIc and its mutants in HepG2 and SMMC-7721 cells that stably transfected the recombinant plasmids pWPI.1-FGFR3IIIc-GFP, pWPI.1-FGFR3AT-I- GFP, and pWPI.1-FGFR3 Δ7-9-GFP by lentivirus.

Figure 2.

Expression pattern of FGFR3 in HCC cell lines and identification of stable cell lines overexpressing FGFR3IIIc, FGFR3Δ7-9, and FGFR3AT-I. A, quantitative real-time RT-PCR was performed to detect the expression level of FGFR3 in HCC cell lines with GAPDH as an internal control; B, Western blot assay was performed to evaluate FGFR3 expression level in HCC cell lines, and GAPDH was used as loading control; C, Western blot analysis of FGFR3IIIc and its mutants in HepG2 and SMMC-7721 cells that stably transfected the recombinant plasmids pWPI.1-FGFR3IIIc-GFP, pWPI.1-FGFR3AT-I- GFP, and pWPI.1-FGFR3 Δ7-9-GFP by lentivirus.

Close modal
Figure 3.

Effects of FGFR3IIIc and FGFR3Δ7-9 on cell proliferation in HCC cell lines. A, CCK-8 cell proliferation assay for HepG2/SMMC-7721 cells that stably express FGFR3IIIc, FGFR3Δ7-9, and FGFR3AT–I. Cells transfected with empty vector were used as a control. B, stable clones of SMMC-7721 cells expressing FGFR3IIIc and FGFR3Δ7-9 and control cells (transfected with empty vector) were injected subcutaneously into the right flank of nude mice (n = 5). Three weeks after implantation, FGFR3Δ7-9 stable clone produced larger tumors than FGFR3 IIIc and control clones. *, P < 0.05.

Figure 3.

Effects of FGFR3IIIc and FGFR3Δ7-9 on cell proliferation in HCC cell lines. A, CCK-8 cell proliferation assay for HepG2/SMMC-7721 cells that stably express FGFR3IIIc, FGFR3Δ7-9, and FGFR3AT–I. Cells transfected with empty vector were used as a control. B, stable clones of SMMC-7721 cells expressing FGFR3IIIc and FGFR3Δ7-9 and control cells (transfected with empty vector) were injected subcutaneously into the right flank of nude mice (n = 5). Three weeks after implantation, FGFR3Δ7-9 stable clone produced larger tumors than FGFR3 IIIc and control clones. *, P < 0.05.

Close modal

For in vivo confirmation, recombinant lentivirus harboring FGFR3Δ7-9 and FGFR3IIIc was transfected into SMMC-7721 cells. The stable clone expressing ectopic FGFR3IIIc or FGFR3Δ7-9 was subcutaneously injected into the flank of athymic nude mouse, and an equal volume of cells transfected with empty vector was injected into the opposite flank of the same mouse as the negative control. As shown in Fig. 3B, SMMC-7721 cells stably expressing FGFR3Δ7-9 formed larger tumors than those stably expressing FGFR3IIIc after 3 weeks' observation (FGFR3Δ7-9: 352.6 ± 131.9 mm3, FGFR3IIIc: 151.6 ± 97.6 mm3, empty vector: 148.2 ± 92.0 mm3, P < 0.05). Thus, data from in vitro and in vivo all revealed that FGFR3Δ7-9 overexpression did play a positive role in promoting the growth of HCC cells.

Effect of FGFR3Δ7-9 mutation on migration and metastasis

To investigate the role of FGFR3Δ7-9 in HCC progression, we employed wound-healing assays to evaluate the effect of FGFR3Δ7-9 on cell migration. The data showed that SMMC-7721 and HepG2 cells, transiently transfected with pcDNA3.1-FGFR3Δ7-9, could enhance the mobility of HCC cells, as compared with the empty vector (Fig. 4A). Furthermore, the Transwell assay was carried out to evaluate the effect of FGFR3Δ7-9 on cell migration. The results showed FGFR3 splicing mutation did indeed significantly promote migration of SMMC-7721 and HepG2 cells across the membrane when FBS was used as an attractant (SMMC-7721: FGFR3Δ7-9: 124.0 ± 19.3, FGFR3IIIc: 78.3 ± 17.6, empty vector: 66.0 ± 15.3, P < 0.05; HepG2: FGFR3Δ7-9: 75.5 ± 18.8, FGFR3IIIc: 37.1 ± 18.7, empty vector: 35.0 ± 12.8, P < 0.05).

Figure 4.

Effect of FGFR3IIIc and FGFR3Δ7-9 overexpression on migration and metastases of HCC cells. A, the migration of HepG2 and SMMC-7721 cells transfected with FGFR3IIIc and FGFR3Δ7-9 stable clones were assessed by Transwell assay, with empty vector as control. The experiments were repeated at least three times, and the histograms (right) represent mean numbers of transfected cells from triplicate tests (mean ± SD). B, effect on lung metastases of SMMC-7721 cells stably expressing FGFR3IIIc and FGFR3Δ7-9 via tail vein injection. 18F-FLT micro-PET/CT images of CEK-Sq-1MB mice are shown at the top. Arrow, 18F-FLT uptake positivity in thoracic metastatic lesions. Representative images of lung metastases are shown on the middle panel and pathologic study on the bottom panel. The scattergram (right) shows the numbers of tumor nodules in each of the five nude mice during 8 weeks of observation. Arrows, metastatic tumors. The ANOVA test was used to evaluate the statistical significance of these experiments. *, P < 0.05; **, P < 0.01.

Figure 4.

Effect of FGFR3IIIc and FGFR3Δ7-9 overexpression on migration and metastases of HCC cells. A, the migration of HepG2 and SMMC-7721 cells transfected with FGFR3IIIc and FGFR3Δ7-9 stable clones were assessed by Transwell assay, with empty vector as control. The experiments were repeated at least three times, and the histograms (right) represent mean numbers of transfected cells from triplicate tests (mean ± SD). B, effect on lung metastases of SMMC-7721 cells stably expressing FGFR3IIIc and FGFR3Δ7-9 via tail vein injection. 18F-FLT micro-PET/CT images of CEK-Sq-1MB mice are shown at the top. Arrow, 18F-FLT uptake positivity in thoracic metastatic lesions. Representative images of lung metastases are shown on the middle panel and pathologic study on the bottom panel. The scattergram (right) shows the numbers of tumor nodules in each of the five nude mice during 8 weeks of observation. Arrows, metastatic tumors. The ANOVA test was used to evaluate the statistical significance of these experiments. *, P < 0.05; **, P < 0.01.

Close modal

On the basis of the results above, in vivo model was employed to further confirm the enhancing effect of FGFR3Δ7-9 on tumor metastasis. SMMC-7721 cells stably expressing FGFR3IIIc or FGFR3Δ7-9 (1 × 106 cells) were injected via tail vein to observe long-distance tumor metastasis in vivo. After eight weeks of observation, lung metastasis was examined by Micro/PET-CT and pathologic study. Mice injected with those carrying empty vector were used as control. Intriguingly, 4 of 5 mice injected with SMMC-7721 cells expressing FGFR3Δ7-9 developed metastatic lung tumors with larger and greater numbers of nodules, whereas 1 of 5 visible metastatic tumors were found in FGFR3IIIc and 0 of 5 in control mice (Fig. 4B, P < 0.05). Histologic analyses also confirmed the presence of lung metastases in these mice. All these results supported that FGFR3Δ7-9 mutation can significantly enhance the migration and metastasis of HCC cells.

Effect of FGFR3Δ7-9 knockdown on tumor proliferation and metastasis

To further confirm the role of FGFR3Δ7-9 in the proliferation and migration of HCC cells, shRNA targeting FGFR3Δ7-9 was employed to inhibit FGFR3 mutation in SMMC-7721 cells transiently transfected with pcDNA3.1-FGFR3Δ7-9 (SMMC-7721FGFR3Δ7-9). The knockdown of FGFR3Δ7-9 was established by stable transfection of a lentiviral vector encoding shRNA1 and shRNA2 into SMMC-7721FGFR3Δ7-9 cells. The CCK8 results indicated that FGFR3Δ7-9 knockdown could significantly inhibit the viability of SMMC-7721FGFR3Δ7-9 cells (FGFR3-shRNA1: 0.56 ± 0.06, FGFR3-shRNA2: 0.12 ± 0.01) compared with that of the negative control group (NC: 0.78 ± 0.08, P < 0.01; Fig. 5A and B). Furthermore, the Transwell assay showed that FGFR3Δ7-9 knockdown induced by both shRNA1 and shRNA2 can significantly inhibit SMMC-7721FGFR3Δ7-9 migration as compared with that of empty vector for control (NC: 36.7 ± 4.4, shRNA1: 6.3 ± 1.2, shRNA2: 3.0 ± 0.5, P < 0.01, Fig. 5C).

Figure 5.

Effect of FGFR3Δ7-9 knockdown on HCC cells migration and metastases in vitro. A, the effect of FGFR3Δ7-9 knockdown by RNAi on SMMC-7721FGFR3Δ7-9 cells was confirmed by immunoblotting. B, CCK-8 cell proliferation assay for SMMC-7721FGFR3Δ7-9-NC, SMMC-7721FGFR3Δ7-9-shRNA1, and SMMC-7721FGFR3Δ7-9-shRNA2 cells. C, Transwell assay for SMMC-7721FGFR3Δ7-9-NC, SMMC-7721FGFR3Δ7-9-shRNA1, and SMMC-7721FGFR3Δ7-9-shRNA2 cells. The lower histograms represent mean numbers of transfected cells from triplicate tests (mean ± SD). Representative views (top) show the transfected cells. D, SMMC-7721FGFR3Δ7-9-NC cells and SMMC-7721FGFR3Δ7-9-shRNA2 cells were injected subcutaneously into nude mice. Mice were sacrificed 6 weeks after implantation. Tumor volume and tumor weight were measured after dissection. Tumor volumes were recorded 6 weeks after tumor cell inoculation. *, P < 0.05; **, P < 0.01.

Figure 5.

Effect of FGFR3Δ7-9 knockdown on HCC cells migration and metastases in vitro. A, the effect of FGFR3Δ7-9 knockdown by RNAi on SMMC-7721FGFR3Δ7-9 cells was confirmed by immunoblotting. B, CCK-8 cell proliferation assay for SMMC-7721FGFR3Δ7-9-NC, SMMC-7721FGFR3Δ7-9-shRNA1, and SMMC-7721FGFR3Δ7-9-shRNA2 cells. C, Transwell assay for SMMC-7721FGFR3Δ7-9-NC, SMMC-7721FGFR3Δ7-9-shRNA1, and SMMC-7721FGFR3Δ7-9-shRNA2 cells. The lower histograms represent mean numbers of transfected cells from triplicate tests (mean ± SD). Representative views (top) show the transfected cells. D, SMMC-7721FGFR3Δ7-9-NC cells and SMMC-7721FGFR3Δ7-9-shRNA2 cells were injected subcutaneously into nude mice. Mice were sacrificed 6 weeks after implantation. Tumor volume and tumor weight were measured after dissection. Tumor volumes were recorded 6 weeks after tumor cell inoculation. *, P < 0.05; **, P < 0.01.

Close modal

In in vivo study, SMMC-7721FGFR3Δ7-9 offspring subclone (1 × 106 cells) was subcutaneously injected into the right flank of athymic nude mice with empty vector for control. As expected, HCC tumors derived from the offspring subclones with inhibited FGFR3Δ7-9 were significantly fewer and smaller than those of SMMC-7721FGFR3Δ7-9 control after 5 weeks of observation (shRNA2:210.2 ± 94.63 mm3 vs. C:1546.0 ± 331.7 mm3, P < 0.05; Fig. 5D). Then, we injected SMMC-7721FGFR3Δ7-9-shRNA1 and SMMC-7721FGFR3Δ7-9 cells (1 × 106 cells) into nude mice via tail veins and then checked lung metastasis after 11 weeks. As expected, the number of metastatic tumors was higher in SMMC-7721FGFR3Δ7-9 group (5/5), whereas only 1 of 5 visible metastatic tumor could be found in shRNA2 group (P < 0.01). Histologic study confirmed the presence of metastatic tumors in the lungs of these mice.

Altogether the data from in vitro and in vivo assay confirmed that FGFR3Δ7-9 silencing can significantly inhibit the proliferation and metastasis of HCC cells.

Differential ligand dependence of alternative splicing of FGFR3

We next examined whether cells expressing wild-type and two FGFR3 alternative splicings, FGFR3Δ7-9 and FGFR3AT-I, were responsive to FGFs (FGF1 or FGF2) stimulation in terms of receptor activation. The binding of FGFs and FGFR3 can activate FGFR3 tyrosine kinase activity and result in cell proliferation or differentiation via signal transduction pathways (21). To detect the difference in binding affinity between FGFR3s and their ligands in HCC cells, HA-FGFR3IIIc, HA-FGFR3Δ7-9, or FGFR3AT-I was cotransfected with FLAG-FGF1 into HepG2 cells, respectively. The results in Fig. 6A showed that three HA-FGFR3s (HA-FGFR3IIIc, HA-FGFR3Δ7-9, and FGFR3AT-I) could all be immunoprecipitated by HA-antibody. The special binding by M2-FLAG mAb in Western blot analysis confirmed that FLAG-FGF1 or FLAG-FGF2 could interact with HA-FGFR3s. This result revealed that FGFR3IIIc and its mutants, HA-FGFR3Δ7-9 and FGFR3AT-I, could stably satisfy binding requests of both FGF1 and FGF2. More importantly, the binding affinity of FGF1/FGF2 to FGFR3Δ7-9 was apparently stronger than that to FGFR3IIIc.

Figure 6.

Differential ligand dependence of alternative splicings of FGFR3. A, FGF1 and FGF2 stably interact with FGFR3 and its mutants. HA-tagged FGFR3IIIc, FGFR3Δ7-9, and FGFR3AT-I proteins were expressed in HepG2 cell lines and examined with indicated antibodies. B, SPR binding analyses of ligand binding between FGFR3IIIc and FGFR3Δ7-9 with FGF1. The binding affinity (Kd) of FGF1 with FGFR3IIIc and FGFR3Δ7-9 was 32.4 ± 0.6 nmol/L and 2.9 ± 0.3nmol/L, respectively. C, self-dimerization of FGFR3Δ7-9. FLAG-tagged FGFR3IIIc, and FGFR3Δ7-9 proteins was expressed in 293T cells, treated with or without FGF1 for 2 hours. Whole-cell extracts were subjected to a non-reducing SDS-PAGE, followed by Western blotting analyses with an anti-FLAG antibody. D, activation of FGFR3IIIc, FGFR3Δ7-9, and FGFR3AT-I. HepG2 cells expressing FGFR3IIIc and its mutants was stimulated with (+) or without (−) FGF1. After SDS-PAGE, Western blot analysis was performed with an anti-phosphotyrosine antibody. Empty vector (EV) was used as control.

Figure 6.

Differential ligand dependence of alternative splicings of FGFR3. A, FGF1 and FGF2 stably interact with FGFR3 and its mutants. HA-tagged FGFR3IIIc, FGFR3Δ7-9, and FGFR3AT-I proteins were expressed in HepG2 cell lines and examined with indicated antibodies. B, SPR binding analyses of ligand binding between FGFR3IIIc and FGFR3Δ7-9 with FGF1. The binding affinity (Kd) of FGF1 with FGFR3IIIc and FGFR3Δ7-9 was 32.4 ± 0.6 nmol/L and 2.9 ± 0.3nmol/L, respectively. C, self-dimerization of FGFR3Δ7-9. FLAG-tagged FGFR3IIIc, and FGFR3Δ7-9 proteins was expressed in 293T cells, treated with or without FGF1 for 2 hours. Whole-cell extracts were subjected to a non-reducing SDS-PAGE, followed by Western blotting analyses with an anti-FLAG antibody. D, activation of FGFR3IIIc, FGFR3Δ7-9, and FGFR3AT-I. HepG2 cells expressing FGFR3IIIc and its mutants was stimulated with (+) or without (−) FGF1. After SDS-PAGE, Western blot analysis was performed with an anti-phosphotyrosine antibody. Empty vector (EV) was used as control.

Close modal

On the basis of the results above, SPR binding analysis was further employed to detect the binding affinity of FGF1 to FGFR3IIIc and FGFR3Δ7-9 quantitatively. The results indicated that the dissociation equilibrium constant (Kd) values of FGFR3Δ7-9 and FGFR3IIIc were 2.9 ± 0.3 nmol/L and 32.4 ± 0.6 nmol/L, respectively (P < 0.01, Fig. 6B). Altogether, the splicing mutation in FGFR3Δ7-9 could promote the ligand-binding ability between FGF1 and its receptors.

Self-dimerization and self-activation of FGFR3Δ7-9

FGFR3 transduces biochemical signals by lateral dimerization in plasma membrane (22). Therefore, we further compared the difference in dimer formation between FGFR3 and its mutants. As shown in Fig. 6C, FLAG-FGFR3IIIc could form dimer only in the presence of FGF1 stimulation. Interestingly, FLAG-FGFR3Δ7-9 showed constitutive dimerization in the absence of ligand FGF1, while FLAG-FGFR3AT-I was unable to form dimer no matter with the presence of FGFs or not. These results suggested mutation in exon 7–9 could lead to the overstabilization of unliganded FGFR3 dimers in plasma membrane, thus increasing ligand-independent activation.

After dimerization, FGFR3s must be phosphorylated on the tyrosine area upon stimulation with FGFs and followed by stimulation of downstream signaling cascades. We then detected the effect of exon7-9 mutation on FGFR3 ligand-independent phosphorylation within the context of full-length receptor. The result in Fig. 6D showed the phosphorylation of FGFR3IIIc required FGF1 cotransfection. However, FGFR3Δ7-9 could be self-phosphorylated on the tyrosine site even in the absence of ligand. As expected, FGFR3AT-I showed no phosphorylation in response to ligand stimulation (Fig. 6D). Thus, it appeared that mutation could trigger FGFR3Δ7-9 self-phosphorylation in the absence of ligand.

Taken together, these results revealed that RNA abbreviation splicing mutation from exon 6 to exon 10 in wild-type FGFR3 allowed FGFR3Δ7-9 self-activation even in the absence of ligand binding of FGFs, which may lead to cascade dysregulation of growth modulation.

Possible mechanisms of FGFR3Δ7-9 mutation in promoting HCC malignancy

To explore the molecular mechanisms, as the downstream signal pathways of FGFRs, the activation of PI3K/AKT/p-AKT and ERK/p-ERK pathway was detected (23). As shown in Fig. 7A, a great increase in ERK and AKT activation could be identified in HepG2 cells expressing FGFR3IIIc in response to FGF1, which revealed that ligand binding was the key to signaling activation for FGFR3IIIc. However, consistent with the ligand-independent characteristics of FGFR3Δ7-9, p-ERK and p-AKT could be readily detected no matter with or without FGF1 stimulation. These results indicated that FGFR3Δ7-9 mutation might influence the structure of binding domains, which led to a ligand-independent activation and aberrant downstream signaling pathway consequently. Meanwhile, as expected, FGFR3AT-I failed to stimulate ERK and AKT signaling transduction pathway in response to FGF1 treatment. In addition, we also identified that p-AKT and p-ERK were downregulated by FGFR3Δ7-9-knockdown in SMMC-7721FGFR3Δ7-9 cells (Fig. 7B). Thus, FGFR3Δ7-9 might be a ligand-independent or low ligand-dependent receptor, and FGFR3Δ7-9 mutation could consequently affect downstream signaling.

Figure 7.

Expression profile of downstream signal pathways of FGFR3 and epithelial–mesenchymal transition markers. A, HepG2 cells expressing FGFR3IIIc, FGFR3Δ7-9, and FGFR3AT-I were cultured with heparin and FGF1. After stripping, blots were reprobed with anti-AKT and anti-ERK antibody to examine equal protein levels. AKT and ERK activation were measured by immunoblotting with p-AKT and p-ERK antibodies. B, SMMC-7721FGFR3Δ7-9-NC, SMMC-7721FGFR3Δ7-9-shRNA1, and SMMC-7721FGFR3Δ7-9-shRNA2 cells were cultured with heparin and FGF1. AKT and ERK activation were measured by immunoblotting with p-AKT and p-ERK antibodies. C, HepG2 cells expressing FGFR3IIIc, Flag-FGFR3Δ7-9, and FGFR3AT-I were cultured without heparin and FGF1. Western blot analysis showed downregulation of E-cadherin and upregulation of Snail and MMP-9. D, E-cadherin promoter reporter assays in SMMC-7721 and HepG2 cells. The luciferase reporter constructed E-cad-Luc was cotransfected into SMMC-7721 and HepG2 cells with FGFR3 or its mutants, while normalized luciferase activity was determined. E, interaction of FGFR3Δ7-9 with E-cadherin endogenously in HepG2 cells. IgG was used as immunoprecipitation control. Phospho-tyrosine antibody was employed to detect the tyrosine phosphorylation modification of E-cadherin.

Figure 7.

Expression profile of downstream signal pathways of FGFR3 and epithelial–mesenchymal transition markers. A, HepG2 cells expressing FGFR3IIIc, FGFR3Δ7-9, and FGFR3AT-I were cultured with heparin and FGF1. After stripping, blots were reprobed with anti-AKT and anti-ERK antibody to examine equal protein levels. AKT and ERK activation were measured by immunoblotting with p-AKT and p-ERK antibodies. B, SMMC-7721FGFR3Δ7-9-NC, SMMC-7721FGFR3Δ7-9-shRNA1, and SMMC-7721FGFR3Δ7-9-shRNA2 cells were cultured with heparin and FGF1. AKT and ERK activation were measured by immunoblotting with p-AKT and p-ERK antibodies. C, HepG2 cells expressing FGFR3IIIc, Flag-FGFR3Δ7-9, and FGFR3AT-I were cultured without heparin and FGF1. Western blot analysis showed downregulation of E-cadherin and upregulation of Snail and MMP-9. D, E-cadherin promoter reporter assays in SMMC-7721 and HepG2 cells. The luciferase reporter constructed E-cad-Luc was cotransfected into SMMC-7721 and HepG2 cells with FGFR3 or its mutants, while normalized luciferase activity was determined. E, interaction of FGFR3Δ7-9 with E-cadherin endogenously in HepG2 cells. IgG was used as immunoprecipitation control. Phospho-tyrosine antibody was employed to detect the tyrosine phosphorylation modification of E-cadherin.

Close modal

As FGFR3Δ7-9 mutation might be increased in aggressive HCC cells, we examined some known metastasis-related molecules, including E-cadherin, Snail, vimentin, and MMP-9 in HCC cell lines expressing FGFR3IIIc and FGFR3Δ7-9. As shown in Fig. 7C, in accordance with the in vitro and in vivo results, FGFR3Δ7-9 showed a more strong ability in downregulating E-cadherin and upregulating Snail and MMP-9 than FGFR3IIIc in HepG2 cells.

Furthermore, evidenced by E-cadherin luciferase reporter assay, we confirmed that FGFR3Δ7-9 showed a significantly increased repression of luciferase compared with FGFR3IIIc in HepG2. FGFR3IIIc only lightly repressed E-cadherin luciferase activity (FGFR3Δ7-9: 5.0 ± 0.2, FGFR3IIIc: 10.0 ± 0.4, FGFR3AT-I: 10.1 ± 0.7, empty vector: 14.9 ± 0.8, P < 0.05). Similar results could be obtained from SMMC-7721 subclones (FGFR3Δ7-9: 3.2 ± 0.2, FGFR3IIIc: 9.7 ± 0.6, FGFR3AT-I: 10.4 ± 0.4, empty vector: 11.2 ± 0.5, P < 0.05). Data showed above suggested FGFR3Δ7-9 mutants contributed to the decreased expression of E-cadherin in transcription level (Fig. 7D), which might underline the aggressive metastases of HCC.

For further confirmation, tandem mass spectrometry was employed to compare the difference of interaction protein profiles between FGFR3Δ7-9 and FGFR3IIIc. Consistent with the above data, the results revealed that E-cadherin might be a downstream substrate of FGFR3Δ7-9, but not the interactive protein of FGFR3IIIc. Then, E-cadherin antibody was used to perform immunoprecipation in HepG2 subclones stably overexpressing FGFR3IIIc/FGFR3Δ7-9/FGFR3AT-I. As shown in Fig. 7E, compared with FGFR3IIIc, more FGFR3Δ7-9 could be pulled down by less quantity of E-cadherin protein. These results suggested that, compared with FGFR3IIIc, there was more affinity between FGFR3Δ7-9 and E-cadherin. It has been reported that E-cadherin phosphorylation was closely related to cell surface stability and adhesion, to separation of different tissue layers, and to cellular migration. Therefore, FGFR3Δ7-9 mutation might influence E-cadherin phosphorylation and promote HCC cells malignancy consequently.

Altogether, FGFR3Δ7-9 mutation could cause abnormal signal transduction and might interact with E-cadherin in HCC cell lines, which may lead to increased proliferation and migration ability.

FGFR family is composed of an extracellular ligand-binding domain, a transmembrane domain, and a split intracellular kinase domain (24). FGFR family can bind to FGFs, resulting in receptor dimerization, autophosphorylation, and downstream signal transduction (25). Four members of the family have been identified with domains of similar constitution (12). Since the alternative mRNA splicing of FGFRs, seven prototype receptors have been reported and each has a different ligand-binding capacity and tissue distribution (26, 27). Given the multiple forms of receptors, the enormous potential for diversity of FGFR family is involved in cell growth control, cell differentiation, and migration. Recently, many studies have proved FGFR family may be involved in carcinogenesis, cancer growth, and migration (5, 18, 28, 29). In our previous study, we detected mRNA levels of all members of FGFR family and found that only FGFR3 was frequently elevated in HCC specimens, which was correlated positively with clinicopathologic index, HCC differentiation, and advanced nuclear grade (2).

Activation and dimerization of FGFR3 result in cell proliferation or differentiation via signal transduction pathways (10). Point mutations in transmembrane domain of FGFR3 have been identified as a causative factor for skeletal development disorders such as achondroplasia and hypochondroplasia (7, 30). Activating mutations of FGFR3 are related with superficial bladder tumors and primary urothelial cell carcinomas (31–34). Several novel mutant transcripts caused by aberrant splicing and activation of cryptic splice sequences have been reported in digestive tract tumors (35). More recently, two novel mutations of FGFR3 in colorectal carcinomas have also been identified (18, 36). All identified mutations occur at highly conserved sequences, not only in the molecules of FGFR family but also throughout evolution and are clustered in the Ig-like loop-III domain, highlighting the functional importance of this domain. All the observations support that FGFR3 represents an important example of single-gene, causing different human developmental and tumor diseases (37).

In addition to wild-type, two novel transmembrane forms of FGFR3, FGFR3Δ7-9 and FGFR3AT-I, were identified in HCC in this study. Wild-type splicing of FGFR3IIIb creates a codon from the final base of exon 8 and the first 2 bases of exon 10. Similarly, FGFR3IIIc uses a codon created by the final base of exon 9 and the first 2 bases of exon 10 (38). The direct joining of exon 6 to exon 10 in FGFR3Δ7-9 is expected to be in-frame, leading to the expression of an intact intracellular TK domain. As to FGFR3AT-I, exon 7 is spliced to exon 10, introducing a 1-base (−1) frameshift, resulting in a premature termination sequence at base 1217, and causing a nonsense translation product, which provide a control for further understanding the significance of splicing mutation. Evidenced by both in vitro and in vivo experiments, compared with FGFR3IIIc, FGFR3Δ7-9 overexpression significantly promoted the proliferation, invasion, and distant lung metastasis of HCC tumor cells. Meanwhile, FGFR3Δ7-9 silencing induced by RNAi led to a decreased viability and migration of HCC cell lines. On the basis of our previous data, the findings in this study confirmed the possible mechanism underlying FGFR3 overexpression and HCC advanced malignancy. We proved amplification or activation of FGFR3 due to genomic imbalance may lead to its upregulation in HCC and can contribute to malignant behaviors of HCC.

The ligand-binding site of FGFRs was localized to Ig-like-II and -III domains (22, 27, 37). In this study, FGFR3Δ7-9 mutant can stably satisfy FGF1 or FGF2 binding requests, and showed a much more strong affinity to FGFs than wild-type FGFR3. More importantly, FGFR3Δ7-9 could constitutively form dimers by itself independent of FGF1 stimulation. Self-dimerization of FGFR3Δ7-9 resulted in receptor activation and overstabilization of unliganded FGFR3 dimers in plasma membrane thus increasing ligand-independent activation. Although FLAG-tag in ligand-binding affinity assay and SPR might have altered the 3D configurations of ligand, the unphysiologic binding properties were less apparent by systematic control. Therefore, the splicing mutation in Ig-like III domain of FGFR3 may influence both dimerization and receptor activation of FGFR3 or/and ligand-binding capacity. These results were correlated positively with an increased FGFR3 locus in HCC, suggesting that FGFR3 gain or amplification due to genomic imbalance could lead to advanced malignancy of HCC.

It has been reported that Ras/Raf/MAPK and PI3K could be downstream signaling pathways of FGFRs (39–41). Great increase in ERK and AKT signaling transduction pathway could be observed in HepG2 expressing FGFR3Δ7-9 even without FGF1 stimulation, while signaling activations in wild-type FGFR3IIIc were detected only with ligand binding. Thus, we consistently proved that FGFR3Δ7-9 mutation might constitute a ligand-independent or low ligand–dependent receptor, which consequently resulted in aberrant downstream signaling pathway.

The cadherin family of transmembrane glycoproteins and Snail play critical roles in cell-to-cell adhesion. In this study, FGFR3Δ7-9 could more apparently induce downregulation of E-cadherin and upregulation of Snail and MMP-9 in HCC cells than FGFR3IIIc. As is known, the ability of E-cadherin–catenin complex to mediate cell adhesion can be eliminated by posttranslational events, including phosphorylation, degradation, and steric hindrance (1). Our results showed that FGFR3Δ7-9 seemed to have a more strong ability to phosphorylate E-cadherin than FGFR3IIIc, which may lead to remarkable destabilization of cadherin/catenin complex. As tyrosine phosphorylation can initiate disassembly of the complex during normal cell migration (41) we may establish the hypothesis that E-cadherin functions as a substrate of FGFR3Δ7-9 and FGFR3Δ7-9 could regulate E-cadherin both at translational and posttranslational level.

Altogether, the fidelity of mRNA splicing could be perturbed in HCC and may generate abnormal aberrant FGFR3 transcripts. As a ligand-independent or low ligand–dependent receptor, FGFR3Δ7-9 functions importantly in HCC proliferation, migration, and distant metastasis.

No potential conflicts of interest were disclosed.

Conception and design: K. Li, W. Yang, W. Qiu

Development of methodology: K. Li

Acquisition of data (provided animals, acquired and managed patients, provided facilities, etc.): K. Li, B. Shen

Analysis and interpretation of data (e.g., statistical analysis, biostatistics, computational analysis): K. Li, B. Shen, X. Cheng, D. Ma, X. Liu, W. Qiu

Writing, review, and/or revision of the manuscript: K. Li, B. Shen, D. Ma, W. Qiu

Administrative, technical, or material support (i.e., reporting or organizing data, constructing databases): B. Shen, D. Ma, X. Jing, W. Qiu

Study supervision: B. Shen, X. Jing, C. Peng, W. Qiu

This study was supported by Nature Science Foundation of China (81172326) and Shanghai Charity Foundation for Cancer Research.

The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

1.
Waghray
A
,
Murali
AR
,
Menon
KN
. 
Hepatocellular carcinoma: From diagnosis to treatment
.
World J Hepatol
2015
;
18
:
1020
9
.
2.
Qiu
WH
,
Zhou
BS
,
Chu
PG
,
Chen
WG
,
Chung
C
,
Shih
J
, et al
Over-expression of fibroblast growth factor receptor 3 in human hepatocellular carcinoma
.
World J Gastroenterol
2005
;
11
:
5266
72
.
3.
Flippot
R
,
Kone
M
,
Magné
N
,
Vignot
S
. 
FGF/FGFR signalling: Implication in oncogenesis and perspectives
.
Bull Cancer
2015
;
102
:
516
26
.
4.
Du
X
,
Lin
BC
,
Wang
QR
,
Li
H
,
Ingalla
E
,
Tien
J
, et al
MMP-1 and Pro-MMP-10 as potential urinary pharmacodynamic biomarkers of FGFR3-targeted therapy in patients with bladder cancer
.
Clin Cancer Res
2014
;
20
:
6324
35
.
5.
Yin
Y
,
Betsuyaku
T
,
Garbow
JR
,
Miao
J
,
Govindan
R
,
Ornitz
DM
. 
Rapid induction of lung adenocarcinoma by fibroblast growth factor 9 signaling through FGF receptor 3
.
Cancer Res
2013
;
73
:
5730
41
.
6.
Zhu
L
,
Somlo
G
,
Zhou
B
,
Shao
J
,
Bedell
V
,
Slovak
ML
, et al
Fibroblast growth factor receptor 3 inhibition by short hairpin RNAs leads to apoptosis in multiple myeloma
.
Mol Cancer Ther
2005
;
4
:
787
98
.
7.
Herrera-Abreu
MT
,
Pearson
A
,
Campbell
J
,
Shnyder
SD
,
Knowles
MA
,
Ashworth
A
, et al
Parallel RNA interference screens identify EGFR activation as an escape mechanism in FGFR3-mutant cancer
.
Cancer Discov
2013
;
3
:
1058
71
.
8.
André
F
,
Bachelot
T
,
Campone
M
,
Dalenc
F
,
Perez-Garcia
JM
,
Hurvitz
SA
, et al
Targeting FGFR with dovitinib (TKI258): preclinical and clinical data in breast cancer
.
Clin Cancer Res
2013
;
19
:
3693
702
.
9.
Yuan
L
,
Liu
ZH
,
Lin
ZR
,
Xu
LH
,
Zhong
Q
,
Zeng
MS
. 
Recurrent FGFR3-TACC3 fusion gene in nasopharyngeal carcinoma
.
Cancer Biol Ther
2014
;
15
:
1613
21
.
10.
Acquaviva
J
,
He
S
,
Zhang
C
,
Jimenez
JP
,
Nagai
M
,
Sang
J
, et al
FGFR3 translocations in bladder cancer: differential sensitivity to HSP90 inhibition based on drug metabolism
.
Mol Cancer Res
2014
;
12
:
1042
54
.
11.
Aoyama
E
,
Kubota
S
,
Takigawa
M
. 
CCN2/CTGF binds to fibroblast growth factor receptor 2 and modulates its signaling
.
FEBS Lett
2012
;
586
:
4270
5
.
12.
Cha
JY
,
Maddileti
S
,
Mitin
N
,
Harden
TK
,
Der
CJ
. 
Aberrant receptor internalization and enhanced FRS2-dependent signaling contribute to the transforming activity of the fibroblast growth factor receptor 2 IIIb C3 isoform
.
J Biol Chem
2009
;
284
:
6227
40
.
13.
Du
X
,
Wang
QR
,
Chan
E
,
Merchant
M
,
Liu
J
,
French
D
, et al
FGFR3 stimulates stearoyl CoA desaturase 1 activity to promote bladder tumor growth
.
Cancer Res
2012
;
72
:
5843
55
.
14.
Schepeler
T
,
Lamy
P
,
Hvidberg
V
,
Laurberg
JR
,
Fristrup
N
,
Reinert
T
, et al
A high resolution genomic portrait of bladder cancer: correlation between genomic aberrations and the DNA damage response
.
Oncogene
2013
;
32
:
3577
86
.
15.
Li
Z
,
Li
X
,
Yu
C
,
Wang
M
,
Peng
F
,
Xiao
J
, et al
MicroRNA-100 regulates pancreatic cancer cells growth and sensitivity to chemotherapy through targeting FGFR3
.
Tumour Biol
2014
;
35
:
11751
9
.
16.
Jiang
H
,
Qu
L
,
Wang
Y
,
Cong
J
,
Wang
W
,
Yang
X
. 
miR-99a promotes proliferation targeting FGFR3 in human epithelial ovarian cancer cells
.
Biomed Pharmacother
2014
;
68
:
163
9
.
17.
Cao
K
,
Liu
W
,
Nakamura
H
,
Enomoto
H
,
Yamamoto
T
,
Saito
M
, et al
Vitamin K2 downregulates the expression of fibroblast growth factor receptor 3 in human hepatocellular carcinoma cells
.
Hepatol Res
2009
;
39
:
1108
17
.
18.
Jang
JH
,
Shin
KH
,
Park
YJ
,
Lee
RJ
,
McKeehan
WL
,
Park
JG
. 
Novel transcripts of fibroblast growth factor receptor 3 reveal aberrant splicing and activation of cryptic splice sequences in colorectal cancer
.
Cancer Res
2000
;
60
:
4049
52
.
19.
Sonvilla
G
,
Allerstorfer
S
,
Heinzle
C
,
Stättner
S
,
Karner
J
,
Klimpfinger
M
, et al
Fibroblast growth factor receptor 3-IIIc mediates colorectal cancer growth and migration
.
Br J Cancer
2010
;
102
:
1145
56
.
20.
Qiu
W
,
Zhou
B
,
Zou
H
,
Chu
PG
,
Lopez
R
,
Shih
J
, et al
Hypermethylation of growth arrest DNA damage-inducible gene 45 β promoter in human hepatocellular carcinoma
.
Am J Pathol
2004
;
165
:
1689
99
.
21.
Blanca
A
,
Requena
MJ
,
Alvarez
J
,
Cheng
L
,
Montironi
R
,
Raspollini
MR
, et al
FGFR3 and Cyclin D3 as urine biomarkers of bladder cancer recurrence
.
Biomark Med
2016
;
10
:
243
53
.
22.
Bocharov
EV
,
Lesovoy
DM
,
Goncharuk
SA
,
Goncharuk
MV
,
Hristova
K
,
Arseniev
AS
. 
Structure of FGFR3 transmembrane domain dimer: implications for signaling and human pathologies
.
Structure
2013
;
21
:
2087
93
.
23.
Xiao
Z
,
Huang
J
,
Cao
L
,
Liang
Y
,
Han
X
,
Quarles
LD
. 
Osteocyte-specific deletion of Fgfr1 suppresses FGF23
.
PLoS One
2014
;
9
:
e104154
.
24.
di Martino
E
,
L'Hote
CG
,
Kennedy
W
,
Tomlinson
DC
,
Knowles
MA
. 
Mutant fibroblast growth factor receptor 3 induces intracellular signaling and cellular transformation in a cell type- and mutation-specific manner
.
Oncogene
2009
;
28
:
4306
16
.
25.
Parish
A
,
Schwaederle
M
,
Daniels
G
,
Piccioni
D
,
Fanta
P
,
Schwab
R
, et al
Fibroblast growth factor family aberrations in cancers: clinical and molecular characteristics
.
Cell Cycle
2015
;
7
:
1
8
.
26.
Cha
JY
,
Lambert
QT
,
Reuther
GW
,
Der
CJ
. 
Involvement of fibroblast growth factor receptor 2 isoform switching in mammary oncogenesis
.
Mol Cancer Res
2008
;
6
:
435
45
.
27.
Itoh
H
,
Hattori
Y
,
Sakamoto
H
,
Ishii
H
,
Kishi
T
,
Sasaki
H
, et al
Preferential alternative splicing in cancer generates a K-sam messenger RNA with higher transforming activity
.
Cancer Res
1994
;
54
:
3237
41
.
28.
Jiang
X
,
Feng
K
,
Zhang
Y
,
Li
Z
,
Zhou
F
,
Dou
H
, et al
Sorafenib and DE605, a novel c-Met inhibitor, synergistically suppress hepatocellular carcinoma
.
Oncotarget
2015
;
6
:
12340
56
.
29.
Sung
JY
,
Sun
JM
,
Chang Jeong
B
,
Il Seo
S
,
Soo Jeon
S
,
Moo Lee
H
, et al
FGFR3 overexpression is prognostic of adverse outcome for muscle-invasive bladder carcinoma treated with adjuvant chemotherapy
.
Urol Oncol
2014
;
32
:
23
31
.
30.
Liao
RG
,
Jung
J
,
Tchaicha
J
,
Wilkerson
MD
,
Sivachenko
A
,
Beauchamp
EM
, et al
Inhibitor-sensitive FGFR2 and FGFR3 mutations in lung squamous cell carcinoma
.
Cancer Res
2013
;
73
:
5195
205
.
31.
Henriksson
ML
,
Edin
S
,
Dahlin
AM
,
Oldenborg
PA
,
Öberg
Å
,
Van Guelpen
B
, et al
Colorectal cancer cells activate adjacent fibroblasts resulting in FGF1/FGFR3 signaling and increased invasion
.
Am J Pathol
2011
;
178
:
1387
94
.
32.
Jang
JH
,
Chung
CP
. 
Loss of ligand-binding specificity of fibroblast growth factor receptor 2 by RNA splicing in human chondrosarcoma cells
.
Cancer Lett
2003
;
191
:
215
22
.
33.
Jang
JH
. 
Identification and characterization of soluble isoform of fibroblast growth factor receptor 3 in human SaOS-2 osteosarcoma cells
.
Biochem Biophys Res Commun
2002
;
292
:
378
82
.
34.
Pellegrini
L
,
Burke
DF
,
von Delft
F
,
Mulloy
B
,
Blundell
TL
. 
Crystal structure of fibroblast growth factor receptor ectodomain bound to ligand and heparin
.
Nature
2000
;
407
:
1029
34
.
35.
Dey
JH
,
Bianchi
F
,
Voshol
J
,
Bonenfant
D
,
Oakeley
EJ
,
Hynes
NE
. 
Targeting fibroblast growth factor receptors blocks PI3K/AKT signaling, induces apoptosis, and impairs mammary tumor outgrowth and metastasis
.
Cancer Res
2010
;
70
:
4151
62
.
36.
Bergholz
J
,
Zhang
Y
,
Wu
J
,
Meng
L
,
Walsh
EM
,
Rai
A
, et al
DeltaNp63alpha regulates Erk signaling via MKP3 to inhibit cancer metastasis
.
Oncogene
2014
;
33
:
212
24
.
37.
Li
R
,
Wei
J
,
Jiang
C
,
Liu
D
,
Deng
L
,
Zhang
K
, et al
Akt SUMOylation regulates cell proliferation and tumorigenesis
.
Cancer Res
2013
;
73
:
5742
53
.
38.
Kelleher
FC
,
O'Sullivan
H
,
Smyth
E
,
McDermott
R
,
Viterbo
A
. 
Fibroblast growth factor receptors, developmental corruption and malignant disease
.
Carcinogenesis
2013
;
34
:
2198
205
.
39.
Ahmad
I
,
Iwata
T
,
Leung
HY
. 
Mechanisms of FGFR-mediated carcinogenesis
.
Biochim Biophys Acta
2012
;
1823
:
850
60
.
40.
Vleminckx
KL
,
Deman
JJ
,
Bruyneel
EA
,
Vandenbossche
GM
,
Keirsebilck
AA
,
Mareel
MM
, et al
Enlarged cell-associated proteoglycans abolish E-cadherin functionality in invasive tumor cells
.
Cancer Res
1994
;
54
:
873
7
.
41.
Gumbiner
BM
. 
Cell adhesion: the molecular basis of tissue architecture and morphogenesis
.
Cell
1996
;
84
:
345
57
.