Abstract
Background/Aim: Cholangiocarcinoma (CCA) is an aggressive tumor with limited treatment options especially in 2nd line or later treatments. Targeting fibroblast growth factor receptor (FGFR) 2 has recently emerged as a promising treatment option for patients with CCA harboring FGFR2-fusion. This study investigated the antitumor activities of tasurgratinib as an orally available FGFR1-3 inhibitor, in preclinical FGFR2-driven CCA models. Materials and Methods: Antitumor activities of tasurgratinib were examined in vitro and in vivo using NIH/3T3 cells expressing FGFR2-fusion as FGFR2-driven CCA models, and in vivo using a CCA patient-derived xenograft model. The molecular mechanism of action of tasurgratinib was elucidated through co-crystal structure analysis with FGFR1, manual complex model analysis with FGFR2, and binding kinetics analysis with FGFR2. Furthermore, the cell-based inhibitory activities against acquired resistant FGFR2 mutations in patients with CCA treated with FGFR inhibitors were evaluated. Results: Tasurgratinib showed antitumor activity in preclinical FGFR2-driven CCA models by inhibiting the FGFR signaling pathway in vitro and in vivo. Furthermore, cell-based target engagement assays indicated that tasurgratinib had potent inhibitory activities against FGFR2 mutations, such as N549H/K, which are the major acquired mutations in CCA. We also confirmed that tasurgratinib exhibited fast association and slow dissociation kinetics with FGFR2, binding to the ATP-binding site and the neighboring region, and adopting an Asp-Phe-Gly (DFG)-“in” conformation. Conclusion: These data demonstrate the therapeutic potential of tasurgratinib in FGFR2-driven CCA and provide molecular mechanistic insights into its unique inhibitory profile against secondary FGFR2 resistance mutations in patients with CCA treated with FGFR inhibitors.
Cholangiocarcinoma (CCA) is the second-most common primary hepatic malignancy worldwide. It accounts for approximately 30% of all primary liver cancers and is associated with a high mortality rate owing to few symptoms in its early stages and frequent diagnosis at an advanced stage (1, 2). Incidence rates have increased over the past decade, with an annual incidence of 2000-3000 cases in the United States (3). However, the incidence of CCA is much higher in Asian countries, including Japan and China (4). CCA is often diagnosed after the disease significantly progresses when it is unresectable. In addition to gemcitabine plus cisplatin as the standard of care, durvalumab plus gemcitabine and cisplatin was recently approved and recommended as first-line treatment for patients with advanced stage; however, the prognosis remains poor, and more effective treatment strategies are warranted (5, 6).
The Fibroblast growth factor receptor (FGFR) family of tyrosine kinases receptors consists of four genes that encode transmembrane receptors that bind to FGF on the extracellular domain. Upon binding to a ligand, a signaling cascade induces several cellular functions, including the proliferation, survival, and migration of cancer cells, tumor angiogenesis, and drug resistance (7). FGFR fusions or rearrangements occur at the 3’ region in several types of cancer, in which the kinase domain is intact and functionally conserved in the fusion protein (8). The fusion gene of FGFR3 and transforming acidic coiled-coil 3 (TACC3) can effectively induce spontaneous brain or lung tumors in mice with loss of the p53 gene, suggesting that FGFR fusions or rearrangements are oncogenic drivers (9). Several genetic abnormalities in isocitrate dehydrogenase (IDH) and FGFR have been reported, and 13% of intrahepatic CCA have FGFR2 fusions or rearrangements (10-13). Several nonclinical experiments and clinical studies show that FGFR inhibitors exhibit anticancer activity against cancers with FGFR2 fusions or rearrangements (10, 14-16). Furthermore, FGFR2 fusions or rearrangements are mutually exclusive of other genetic abnormalities occurring in CCA, such as KRAS, BRAF, IDH1/2 (10, 17). Considering the above, FGFR2 fusions or rearrangements are suggested oncogenic drivers of CCA and are appropriate targets of FGFR inhibitors. A pivotal phase II trial of tasurgratinib in patients with FGFR2 fusion gene-positive CCA is ongoing (NCT04238715) (18).
Here, we report the preclinical activities against FGFR2-fusion proteins of an orally available FGFR1-3 selective inhibitor (tasurgratinib, formerly known as E7090) (19, 20). These analyses included in vitro antiproliferative activities against NIH/3T3 expressing FGFR2-fusion genes, in vivo antitumor activities in mouse models harboring FGFR2-fusion, binding mode analysis, and the inhibitory profile of tasurgratinib against secondary FGFR2 resistance mutations.
Materials and Methods
Test agents. Tasurgratinib succinate (tasurgratinib), pemigatinib, futibatinib, infigratinib, and ponatinib were synthesized by Eisai Co., Ltd. (Tsukuba, Japan), in accordance with patent publications WO 2014129477 (US 20140235614), WO 2016027781, US 20130338134, WO 2013108809, WO 2007071752, and WO 2007075869. Gemcitabine hydrochloride was purchased from FUJIFILM Wako Pure Chemical Industries (Osaka, Japan). Erdafitinib was synthesized by ChemPartner (Shanghai, PR China) in accordance with the patent publication WO 2011135376 or purchased from MedChemExpress (Monmouth Junction, NJ, USA). For in vitro studies, all compounds were prepared as stock solutions in dimethyl sulfoxide (DMSO) and diluted in the relevant assay media. For in vivo studies, tasurgratinib was dissolved in distilled water or 10 mmol/l HCl.
cDNA cloning and establishment of NIH/3T3 cells expressing FGFR2-BICC1 type2. cDNAs of FGFR2-BICC1 type2 (a variant of FGFR2-BICC1) (13) were prepared from cancer tissues of a patient with biliary tract cancer. A nucleotide sequence encoding a FLAG epitope tag was ligated to the C-terminus of the cDNA and cloned into a pMX-neo retroviral vector to construct a retrovirus. The retrovirus was used to infect NIH/3T3 cells and express the FGFR2-BICC1 type2 gene into the cell.
Cell culture. NIH/3T3 cells expressing human FGFR2-fusion genes (FGFR2-AHCYL1, FGFR2-BICC1 type1, FGFR2-TXLNA, FGFR2-KCTD1) and NIH/3T3 cells expressing the human KRAS G12V gene were established using pMXs retroviral vectors, as previously reported (10, 13). Frozen NIH/3T3 cells expressing human FGFR2-fusion genes or human KRAS G12V gene were thawed, cultured, and maintained with Dulbecco’s modified Eagle’s medium (DMEM) high glucose (FUJIFILM Wako Pure Chemical) containing 10% fetal bovine serum (FBS) (Sigma-Aldrich, Saint Louis, MO, USA), penicillin/streptomycin (FUJIFILM Wako Pure Chemical), and 700 μg/ml of geneticin (FUJIFILM Wako Pure Chemical).
Anchorage-independent cell proliferation. Fifty microliters of DMEM (Sigma-Aldrich) containing 0.66% agar (BD Biosciences, Franklin Lakes, NJ, USA), 10% FBS, 1 nmol/l sodium pyruvate, and penicillin/streptomycin (0.66% agar-medium solution) were added to each well of a 96-well plate (Sumitomo Bakelite, Tokyo, Japan). The cells were washed with phosphate buffered saline, harvested with 0.05% trypsin-EDTA, and counted. NIH/3T3 cells expressing FGFR2-AHCYL1, FGFR2-BICC1 type1, FGFR2-BICC1 type2, FGFR2-KCTD1, and KRAS G12V were diluted to a density of 4×104 cells/ml in the assay medium, and FGFR2-TXLNA cells were diluted to a density of 8×104 cells/ml in the assay medium. Each cell suspension was mixed with an equal amount of a 0.66% agar-medium solution, and 50 μl of the resulting mixture were added to each well. The plates were cooled at 4°C for approximately 30 min. After returning the plates to room temperature, an additional 50 μl of the 0.66% agar-medium solution were added to each well. Tasurgratinib solution (0.549-400 nmol/l for FGFR2-fusion gene-expressing cells, 5.49-4,000 nmol/l for KRAS G12V-expressing cells), gemcitabine hydrochloride solution (0.549-400 nmol/l), or vehicle was added at 50 μl to each well and incubated at 37°C in a 5% CO2 atmosphere for 14 days. The tasurgratinib solution or gemcitabine hydrochloride solution on the top layer of each well was exchanged approximately one week after the initiation of incubation. Two experiments were performed in sextuplicate (blank, tasurgratinib, or gemcitabine hydrochloride) or in duodecuplicate (vehicle).
A cell proliferation assay was performed using Cell Counting Kit-8 (Dojindo Laboratories, Kumamoto, Japan). After 14 days of incubation, 10 μl of WST-8 (2-[2-methoxy-4-nitrophenyl]-3-[4-nitrophenyl]-5-[2,4-disulfophenyl]-2H-tetrazolium, monosodium salt) reagent of the kit was added to each well, followed by incubation at 37°C for approximately 70-95 min in each experiment. The optical density (OD) of each well was measured at 450 nm (reference wavelength: 650 nm) using a microplate reader (ARVO, PerkinElmer, Waltham, MA, USA).
The percentage inhibition of cell growth was determined using the following formula:
Inhibition % =(“average value of OD of control wells”-“average value of OD of test wells” )/(“average value of OD of control wells”-“average value of OD of blank wells” ) ×100
Blank well: with vehicle without cells; control well: with vehicle and cells; test well: with compounds and cells.
The 50% inhibitory concentration (IC50) values were generated from separate sigmoidal curves representing the degree of growth inhibition versus compound concentrations. The mean IC50 values were calculated based on two independent experiments. Statistical analyses were performed using GraphPad Prism software (GraphPad Software, San Diego, CA, USA).
In vitro signal inhibition analysis. Antibodies against phospho-FGFR (Tyr653/654), phospho-ERK1/2 (Thr202/Tyr204), ERK1/2, phospho-STAT3 (Tyr705), phospho-STAT3 (Ser727), phospho-S6 (Ser235/236), S6, phospho-AKT (Ser473), and AKT were obtained from Cell Signaling Technology (Danvers, MA, USA). FLAG antibody was obtained from Clontech Laboratories, Inc. (Mountain View, CA, USA). Antibodies against β-actin were purchased from Sigma-Aldrich. STAT3 antibody was obtained from BD Biosciences. FGFR2 antibody was purchased from Abcam (Cambridge, UK). Peroxidase-conjugated secondary antibodies were obtained from Cell Signaling Technology.
For in vitro signal inhibition analysis, NIH/3T3 cells were treated with the indicated concentrations of tasurgratinib in RPMI-1640 containing 10% FBS for 4 h, and cell lysates were prepared using RIPA buffer (Sigma-Aldrich) supplemented with a protease inhibitor cocktail (Sigma-Aldrich) and phosphatase inhibitor cocktail (Sigma-Aldrich). The expression of the indicated proteins was detected using western blotting. The blots were developed using Immobilon Western Chemiluminescent HRP Substrate (Millipore, Burlington, MA, USA) and chemiluminescence was detected using an Image Analyzer LAS-4000 (Fujifilm, Tokyo, Japan).
Nano-bioluminescence resonance energy transfer (BRET) assay. The NanoBRET assay was conducted by Carna Biosciences Inc. (Kobe, Japan), according to Promega’s instruction.
For the IC50 determination, BRET measurements were performed using the following settings: Integration time: 0.3 s/well, Donor filter: 450 BP, Acceptor filter: 600 LP.
The data were analyzed using nonlinear fitting to yield IC50 values. All experiments were repeated in triplicate.
For residence time determination, BRET is measured repeatedly with the GloMax® Discover Multimode Microplate Reader (Promega, Madison, WI, USA) under the following settings: Donor filter: 450 BP; acceptor filter: 600 LP; integration time: 0.3 s/well; measurement interval: 1 min; iterations: 61. The test concentration included the following concentrations: ponatinib, 0.1 μmol/l; erdafitinib, 0.3 μmol/l; and tasurgratinib, 0.03 μmol/l.
BRET was measured in triplicate and the averages were used for data analysis. The data were fitted to a kinetic rate equation to estimate the dissociation rate constants of the test compounds (21, 22).
Kinetic interaction analysis against FGFR2. Experiments to determine the binding affinities and kinetic rate constants of the interactions between the compounds and FGFR2 were performed as previously described (23, 24). Briefly, FGFR2 (final concentration, 13 nmol/l) was preincubated with the reporter probe at a concentration equal to its binding affinity (Kd) in a reaction buffer consisting of 20 mmol/l 3-morpholinopropanesulfonic acid (MOPS) (pH 7.0), 1 mmol/l dithiothreitol (DTT), and 0.01% Tween 20. The final reaction volume was 10 μl in black NBS 384-well polypropylene plates. After the transfer of serially diluted compounds, probe displacement was monitored for 60 min. The Kd values were calculated using the Cheng-Prusoff equation from the IC50 values obtained from the percentage displacement values measured at the last time point. The association rate constants of the inhibitors were calculated from the decay rate of the probe displacement. The dissociation rate constants were determined as the product of the Kd × association rate constants.
Co-crystal structure between FGFR1 and tasurgratinib. The crystallographic studies (FGFR1 protein preparation, crystallization, X-ray diffraction data collection, and structure modeling/refinement of FGFR1–tasurgratinib complex) were performed in Proteros biostructures GmbH (Martinsried, Germany).
FGFR1 protein production and purification. Synthetic DNA encoding the kinase domain of human FGFR1 (amino acids 461-774, numbered according to UniProt-entry P11362-1) harboring the C488A-mutation was cloned into pFastBac1 vector in frame with a hexa-histidine tag and TEV-protease cleavage site. Recombinant baculovirus was generated using the procedures outlined in the Bac-to-Bac manual (Thermo Fisher Scientific, Waltham, MA, USA). Protein was expressed in Sf9 cells cultured in Grace’s medium supplemented with 10% FBS at 26°C by infection with a high titer virus stock at a multiplicity of infection of 0.5. The cells were harvested by centrifugation 72 h after infection, and stored at −80°C until purification. The protein for crystallization was purified using affinity chromatography on Ni-NTA (Qiagen, Hilden, Germany), protease cleavage, negative affinity chromatography, and a final size exclusion chromatography step on a Superdex 200 26/60 column (GE Healthcare, Chicago, IL, USA) equilibrated in 10 mmol/l HEPES/NaOH (pH=8.0), 100 mmol/l NaCl, 3% glycerol, and 1 mmol/l (tris(2-carboxyethyl)phosphine) (TCEP). For crystallization experiments, the protein was re-buffered with 10 mmol/l Tris/HCl (pH 8.0), 10 mmol/l NaCl, and 2 mmol/l dithiothreitol (DTT) by repeated dilution steps during concentration using a 30 kDa cutoff ultrafiltration device (Millipore).
Crystallization of tasurgratinib–FGFR1 complex and X-ray diffraction data collection. The kinase domain of human FGFR1 was crystallized in hanging drops at 20°C by mixing 1 μl of protein (concentration 5.5 mg/ml as determined by UV-absorption) with 0.5 μl of reservoir solution (18.00% w/v PEG 3350, 0.2 mol/l di-Ammonium tartrate, 1 mmol/l GSH/GSSG; original condition: #43 from the JCSG Core I suite) and equilibrating the mixture against 500 μl of reservoir solution. Microseeding from imperfect crystals was used to improve the diffraction quality of the resulting crystals, which grew within 2-3 days. The complex with tasurgratinib was formed by soaking apocrystals in a reservoir solution containing 2 mM tasurgratinib (diluted from a 100 mM stock solution in DMSO) overnight. The crystals were cryoprotected in a reservoir solution containing 25% ethylene glycol and vitrified by plunging into liquid nitrogen. X-ray diffraction data were collected from the complex crystals of FGFR1 at the Swiss Light Source (SLS; Villigen, Switzerland) under cryogenic conditions (100 K). The crystal belonged to space group C2. All datasets were processed using XDS and XSCALE (Table I).
Data collection, processing, and refinement statistics for tasurgratinib-FGFR1.
FGFR1–tasurgratinib structure modeling and refinement. The phase information necessary to determine and analyze the structure was obtained via molecular replacement. A previously solved FGFR1 structure was used as the search model. Subsequent model building and refinement were performed according to standard protocols using the software packages CCP4 (version 6.2.0) and COOT (version 0.6). Approximately 4.1% of the measured reflections were excluded from the refinement procedure to calculate the free R-factor, a measure for cross-validating the correctness of the final model (Table I). Translation–libration–screw (TLS) refinement (using REFMAC5 and CCP4) resulted in lower R-factors and higher-quality electron density maps. Automatically generated local NCS restraints were applied (keyword “ncsr local” of newer REFMAC5 versions).
Ligand parameterization and the generation of the corresponding library files were performed using the CORINA program (Molecular Networks GmbH Computer Chemie, Erlangen, Germany).
The water model was built with the “Find waters” algorithm of COOT by putting water molecules in peaks of the Fo-Fc map contoured at 3.0; this was followed by refinement with REFMAC5 and checking all water molecules with the validation tool of COOT.
Statistics of the final structure and the refinement process are listed in Table I.
The coordinates and structural factors were deposited in the Protein Data Bank with the following accession numbers: FGFR1-tasurgratinib complex, 8YKI.
Protein structure preparation by homology modeling and creation of complex structure model. The crystal structure of FGFR1 and tasurgratinib was used as a template to model the FGFR2 structure. because the type V binding style is unique binding modes, tasurgratinib could not bind at the ligand binding site using the available crystal structures of FGFR2 (such as type I or type II). Maestro ver.11.5 software (Schrödinger, LLC, New York, NY, USA) was used for homology modeling and docking complex model creation. A homology model of the kinase domain of human FGFR2 was built using the Prime modeling program, with FGFR1 in complex coordinates and tasurgratinib as the template structure, according to the sequence alignments (25). Tasurgratinib from the complex crystal structure of FGFR1 was merged into the obtained FGFR2 model structure, resulting in the FGFR2- tasurgratinib complex model. Hydrogen atoms were added, and a brief relaxation was performed on each starting structure using the Protein Preparation Wizard. The FGFR2 and tasurgratinib complex model was refined using One-Step Glide Docking (Docking method-refine).
In vivo experiments. Animal experiments were performed in accordance with the guidelines approved by the Institutional Animal Care and Use Committee of Eisai Co., Ltd., or Crown Bioscience Inc. (Taicang, PR China). Tasurgratinib was dissolved in distilled water or 10 mmol/l HCl (depending on the experiment). Cultured NIH/3T3 cells expressing FGFR2-fusion gene were prepared using Hanks’ balanced salt solution at 1×107 cells/ml (NIH/3T3 cells expressing FGFR2-BICC1 type2 and NIH/3T3 cells expressing FGFR2-TXLNA) or 4×107 cells/mL (NIH/3T3 cells expressing FGFR2-KCTD1). A 0.1 ml aliquot of the cell suspension was subcutaneously inoculated into the right flank region of female BALB/c nude mice (6-7 week old, CLEA Japan, Tokyo, Japan or Charles River Laboratories, Yokohama, Japan). When the tumor volume (TV) reached approximately 100-200 mm3, mice were selected based on their TV and general condition, and randomly divided into groups according to their TV (n=5-6). Oral administration of tasurgratinib (6.25, 12.5, 25, or 50 mg/kg) or vehicle (10 mmol/l HCl for FGFR2-KCTD1; distilled water for the others) was started on Day 1, and the administration continued once daily for 7 to 11 days.
The tumor was measured in two dimensions, and the volume was calculated using the following formula:
TV (mm3)=1⁄2 length (mm) × [width (mm)]2
The relative body weight on each day was calculated according to the following formula:
Relative body weight=body weight (g)/body weight on Day 1 (g)
The patient-derived xenograft (PDX; CC6204) study was performed at Crown Bioscience, Inc. HuPrime® CCA model CC6204 derived from a 60 years old female Asian patient was selected for this efficacy study. The patient was originally diagnosed as intrahepatic CCA, grade II-III. CrownBio Pathology QC confirmed that CC6204 is a moderately to poorly differentiated adenocarcinoma. FGFR2-BICC1 gene was detected and confirmed using this model. Tumors from stock mice inoculated with CC6204 tumor tissues were harvested and subcutaneously inoculated into the right flank of female BALB/c nude mice. When the TV reached approximately 193 mm3, mice were randomly allocated to each group (n=7). Oral administration of tasurgratinib (5, 15, or 50 mg/kg) or vehicle (distilled water) at 20 ml/kg body weight was started on Day 1 and continued once daily for 15 days. Four hours after the final dose, tumor samples were collected to prepare formalin-fixed paraffin-embedded (FFPE) blocks and tumor lysates.
Differences in tumor volume between the vehicle-treated and tasurgratinib-treated groups were analyzed using one-way analysis of variance (ANOVA), followed by Dunnett’s multiple comparison test. The dose-responsiveness of tasurgratinib was determined using regression analysis. A p-value <0.05 (two-sided) was considered statistically significant. Statistical analyses were performed using GraphPad Prism software.
The expression of the indicated proteins in whole tumor lysates was detected using western blotting. The blots were developed using Immobilon Western Chemiluminescent HRP Substrate (Millipore) and chemiluminescence was detected using an Image Analyzer LAS-4000 (Fujifilm).
For immunohistochemistry, FFPE blocks from PDX tumors were stained for Ki-67 using an anti-Ki-67 (D2H10) rabbit mAb (IHC Specific) (Cell Signaling Technology) to assess tumor proliferation rates. Staining intensity was scored using Halo imaging analysis software (Indica Labs, Albuquerque, NM, USA).
Results
Effects of tasurgratinib on anchorage-independent cell proliferation of NIH/3T3 cells expressing FGFR2-fusion genes. The expression of FGFR2-fusion genes found in CCA in NIH/3T3 cells conferred anchorage-independent colony formation ability or tumorigenicity in nude mice (10, 13), indicating that FGFR2-fusion genes function as oncogenes. Initially, the effect of tasurgratinib on anchorage-independent cell proliferation was evaluated to confirm FGFR2 as a therapeutic target for CCA. Expression of the KRAS G12V gene in NIH/3T3 cells induces activation of the Raf-MEK-ERK pathway and anchorage-independent cell proliferation (26). Based on these results, NIH/3T3 cells expressing KRAS G12V were used as reference controls for anchorage-independent cell proliferation. We confirmed the anchorage-independent colony formation ability of NIH/3T3 cells expressing FGFR2-AHCYL1, FGFR2-BICC1 type1, FGFR2-TXLNA, and FGFR2-KCTD1, which were found in CCA. Furthermore, anchorage-independent colony formation ability was observed in NIH/3T3 cells expressing FGFR2-BICC1 type2 (data not shown), which is also found in CCA (13). Tasurgratinib inhibited anchorage-independent proliferation of NIH/3T3 cells expressing FGFR2-AHCYL1, FGFR2-BICC1 type1, FGFR2-BICC1 type2, FGFR2-TXLNA, and FGFR2-KCTD1 with IC50 values of 0.78, 2.4, 16, 2.3, and 2.1 nmol/l, respectively. In contrast, the IC50 value of tasurgratinib against NIH/3T3 cells expressing KRAS G12V was 170 nmol/l, which was weaker than that against NIH/3T3 cells expressing FGFR2-fusion genes. Gemcitabine hydrochloride (an anticancer chemotherapeutic agent used for CCA) also showed antiproliferative activity against NIH/3T3 cells expressing FGFR2-AHCYL1, FGFR2-BICC1 type1, FGFR2-BICC1 type2, FGFR2-TXLNA, FGFR2-KCTD1, and KRAS G12V genes, with IC50 values of 6.5, 7.0, 5.8, 6.7, 7.1, and 6.6 nmol/l, respectively (Figure 1 and Table II). These data suggested that tasurgratinib selectively inhibited the anchorage-independent cell proliferation of NIH/3T3 cells expressing FGFR2-fusion genes which are found in CCA.
Inhibitory activities of tasurgratinib and gemcitabine hydrochloride on anchorage-independent cell proliferation of NIH/3T3 expressing FGFR2-fusion gene or KRAS G12V gene. Data represents the mean from two independent experiments performed in more than quadruplicate. IC50: 50% inhibitory concentration.
Summary of in vitro activity of tasurgratinib and gemcitabine hydrochloride in anchorage-independent cell proliferation of NIH/3T3 expressing FGFR2-fusion gene or KRAS G12V gene.
Inhibition of the FGFR signaling pathway in NIH/3T3 cells expressing FGFR2-fusion genes. Subsequently, the effects of tasurgratinib on the FGFR signaling pathway were analyzed using western blotting in NIH/3T3 cells expressing FGFR2-AHCYL1, FGFR2-BICC1 type1, FGFR2-BICC1 type2, FGFR2-TXLNA, and FGFR2-KCTD1. Treatment with tasurgratinib inhibited the phosphorylation of FGFR2 and its downstream molecules ERK1/2, STAT3, and S6 in a dose-dependent manner (Figure 2). These results suggest that the inhibitory activities of tasurgratinib on anchorage-independent proliferation of NIH/3T3 cells expressing FGFR2-fusion genes were mediated by inhibition of the FGFR signaling pathway.
Inhibitory activities of tasurgratinib on the FGFR signaling pathway in NIH/3T3 cells expressing FGFR2-fusion genes. Western blot images of phosphorylation of FGFR2 and its downstream molecules in NIH/3T3 cells expressing FGFR2-fusion genes treated with tasurgratinib (1,000, 100, 10, 1, or 0.1 nmol/l) as indicated.
Antitumor activity of tasurgratinib on allografts of the FGFR2-fusion gene expressing NIH/3T3. In addition to anchorage-independent colony formation, subcutaneous inoculation of NIH/3T3 cells expressing FGFR2-BICC1 type2 induced tumor formation in nude mice (data not shown), indicating that FGFR2-BICC1 type2 works as an oncogenic driver, like FGFR2-TXLNA and FGFR2-KCTD1 (13). Subsequently, the antitumor activity of tasurgratinib on allografts of FGFR2-fusion gene-expressing NIH/3T3 cells in nude mice was evaluated. Tasurgratinib clearly inhibited in vivo tumor growth in all allografts of FGFR2-fusion gene-expressing NIH/3T3 cells in a dose-dependent manner, without severe body weight loss (Figure 3).
Antitumor activity of tasurgratinib in allografts of NIH/3T3 expressing FGFR2-fusion genes. Tumor volume (upper) and relative body weight (lower) are shown for allograft models of NIH/3T3 expressing FGFR2-BICC1 type2 (A), FGFR2-TXLNA (B), and FGFR2-KCTD1 (C). Results are presented as the means±SEM (n=5 per group for A and B, n=6 per group for C). *p<0.05, **p<0.01, ***p<0.001, ****p<0.0001 versus vehicle control on the last measurement day in each study (one-way analysis of variance followed by Dunnett’s multiple comparison test).
Antitumor activity of tasurgratinib in a human CCA PDX model. Since the above experiments used ectopically transfected murine cell lines, further experiments were performed using a patient xenograft model to allow for more clinically relevant studies.
Tasurgratinib showed clear dose-dependent antitumor activity in the CC6204 human CCA patient xenograft model harboring the FGFR2-BICC1 gene (Figure 4). Furthermore, tasurgratinib decreased the phosphorylation of ERK1/2 and S6 downstream of FGFR (Figure 5A) and significantly decreased the percentage of Ki-67-positive cancer cells (Figure 5B and C). These results suggest that the antitumor activity of tasurgratinib in the human CCA PDX model was accomplished by inhibiting the FGFR signaling pathway.
Antitumor activity of tasurgratinib in CC6204 human cholangiocarcinoma PDX model harboring the FGFR2-BICC1 gene. Tumor volume (left) and relative body weight (right) are shown. Results are presented as the means±SEM (n=7 per group). *p<0.05, ***p<0.001 versus the vehicle control on Day 15 (one-way analysis of variance followed by the Dunnett’s multiple comparison test). Two mice were sacrificed owing to ethical reasons in the 15 mg/kg and 50 mg/kg dosing groups.
Inhibitory activities of tasurgratinib on FGFR signaling pathway and proliferation of cancer cells in the CC6204 human CCA PDX model harboring the FGFR2-BICC1 gene. (A) Western blot images of phosphorylation of FGFR2 downstream molecules. (B, C) The number of Ki67-positive cells in the tumor area were analyzed using immunohistochemistry. Representative immunohistochemical images (scale bar=100 μm) are shown in (B). The data shown in (C) are the mean Ki-67-positive cancer cells (%) ±SEM (n=7 per group for vehicle and 5 mg/kg, n=6 per group for 15 mg/kg, n=4 per group for 50 mg/kg). **p<0.01 versus the vehicle control (one-way analysis of variance followed by the Dunnett’s multiple comparison test).
Tasurgratinib-FGFR1 co-crystal structure and manual complex model analysis of tasurgratinib-FGFR2. We previously reported that tasurgratinib exhibited type V-like binding kinetics to FGFR1, fast association, and slow dissociation (19). Co-crystal structural analysis was conducted to clarify the binding mode of tasurgratinib. The structure was solved at a resolution of 2.79 Å and revealed the detailed binding mode of the inhibitor to FGFR1. The amino acid residues located in the vicinity of tasurgratinib with a maximum distance of 3.9 Å were identified (Table III). All belonged to the ATP-binding site, a gatekeeper residue (a common site for protein kinases), or the neighboring region (a non-conservative allosteric region) (27), indicating that tasurgratinib binds to the ATP-binding site and the neighboring region. These bindings are characteristic of selective kinase inhibitors, which can exert their kinase inhibitory activity by binding to the ATP-binding site, while enhancing kinase selectivity by binding to the neighboring region (28-30). Tasurgratinib was situated in the cleft formed between the N-terminal- and C-terminal lobes and bound to the ATP-binding site and the neighboring region of the active site (Figure 6A). Four specific hydrogen bonds of tasurgratinib to the main chain atoms of Ala564 and Asp641 as well as the side chain atoms of Glu531 were identified based on a distance of less than 3.5 Å of the donor and acceptor atoms. All atoms involved in these hydrogen bonds were well defined by the electron density (Figure 6B and C). In protein kinases, the called DFG-motif (an N-terminal Asp-Phe-Gly) sequence of the kinase activation loop adopts the “DFG-in” conformation (open to substrates), and the “DFG-out” conformation (closed to substrates) (29). The DFG-motif (whose carbon atoms are shown with salmon color in the ball-and-stick model) adopted the “DFG-in” conformation in the FGFR1-tasurgratinib complex (Figure 6B). As far as known kinase inhibitors are concerned, most inhibitors fitting DFG-in kinases can only bind to the ATP-binding site, resulting in low kinase selectivity (type I), and those fitting DFG-out kinases bind to the ATP-binding site and the neighboring region, resulting in high kinase selectivity (type II) (31). In contrast, tasurgratinib could bind to both the ATP-binding site and the neighboring region of FGFR1 in the DFG-in conformation. The crystal structure analysis of the vascular endothelial growth factor receptor 2 (VEGFR2) complex with lenvatinib and a kinetic interaction analysis indicated that lenvatinib binds in a different and distinct mode to VEGFR2 compared to other approved VEGFR2 kinase inhibitors (32). This new binding mode was named type V based on the following three criteria: 1) it binds to the ATP-binding site, 2) it binds to the neighboring region, and 3) it fits the kinase adopting the DFG-in conformation (32-35). Crystal structure analysis of FGFR1-tasurgratinib revealed that the binding mode of tasurgratinib to FGFR1 also met the three criteria described above, indicating that tasurgratinib belongs to a new, unique class of tyrosine kinase inhibitors, type V. A complex model analysis of tasurgratinib with FGFR2 was conducted to confirm the binding mode of tasurgratinib to FGFR2, the main FGFR family in FGFR2-fusion gene-positive CCA. Homology modeling and complex model based on the tasurgratinib–FGFR1 complex suggested that the mode of binding to FGFR2 was identical to that of FGFR1 (Figure 6D).
FGFR1 amino acid residues located less than 3.9 Å from tasurgratinib.
Tasurgratinib-FGFR1 co-crystal structure and complex model analysis of tasurgratinib-FGFR2. (A) Overall structure of FGFR1 with tasurgratinib. Tasurgratinib is shown as a space-filling model (CPK® model) colored according to the chemical atom type; carbon, oxygen, and nitrogen atoms are shown in green, red, and blue, respectively. One monomer of FGFR1 is represented as a ribbon. The beta-sheet is shown in blue, the helixes in red, and other components in gray. (B) Binding pocket of FGFR1 complexes. Tasurgratinib and neighboring protein side chains are shown as ball- and-stick model colored according to the chemical atom type, as described above. Carbon atoms of the DFG domain are shown in salmon color. Residue names at the ATP-binding site, neighboring region, and gatekeeper residue are indicated by red, black, and blue letters, respectively. These residues (except for Gly643) are located with a maximum distance of 3.9 Å from tasurgratinib. (C) Interaction scheme of FGFR1 amino acid residues with tasurgratinib. Residues with some interactions with tasurgratinib are circled by different colors according to the type of interaction: those involved in hydrogen bonding, charge, or polar interactions are in pink circles. Residues involved in van der Waals interactions are in green circles. Hydrogen bond interactions with amino acid main- or side-chains are represented by dashed arrows directed toward the electron donor. Aromatic C-H bonds are indicated by blue lines. A Pi interaction is represented by orange lines indicating the interaction. (D) FGFR2-tasurgratinib complex model.
Binding kinetics of tasurgratinib to FGFR2. The binding kinetics of tasurgratinib to FGFR2 were assayed in cell-free and cell-based assays to confirm that the binding mode of tasurgratinib to FGFR2 is type V. The kinetic parameters of the interactions of FGFR2 with erdafitinib and ponatinib were also determined as representative type I and type II inhibitors, respectively (25, 36). The kon and koff values of tasurgratinib were 5.14×104 s−1 M−1 and 4.99×10−4 s−1, resulting in a Kd value of 9.71 nmol/l (Table IV). Erdafitinib had fast association and dissociation kinetics, with a residence time of 23 min, and ponatinib had relatively slow association and dissociation kinetics, with a residence time of 102 min. These data confirmed that erdafitinib and ponatinib are type I and type II inhibitors of FGFR2, respectively. In contrast, tasurgratinib showed intermediate features between the two representative inhibitors, with a residence time of 33 min. These results suggest that tasurgratinib binds to the receptor tyrosine kinase FGFR2 in type V binding mode, similar to the binding kinetics with FGFR1, as previously reported (19). The longer residence time of tasurgratinib compared to that of the type I inhibitor was also confirmed using the cell-based NanoBRET assay system (Table V).
Parameters of kinetic interactions with FGFR2.
Residence time to FGFR2 in cell-based assay.
Inhibitory activity of tasurgratinib against mutant FGFR2 which are found in FGFR inhibitor-treated CCA patients. Approved FGFR inhibitors for patients with CAA harboring the FGFR2-fusion gene include pemigatinib (US, EU, and Japan), infigratinib (US), and futibatinib (US, EU, and Japan) (37-40). Several acquired resistance mutations were detected in circulating tumor DNA (ctDNA) or tumor tissues upon disease progression following FGFR inhibitor therapy in FGFR2-altered cholangiocarcinoma patients (41). In a phase I study of tasurgratinib, ctDNA analysis revealed that secondary FGFR2 resistance mutations (L617F and M537I) also occurred upon treatment with tasurgratinib (42). Hence, the cell-based inhibitory activities of FGFR inhibitors against FGFR2 with several secondary resistance mutations were analyzed using the NanoBRET assay system (Figure 7). Tasurgratinib showed lower inhibitory activity against FGFR2 (L617F) and FGFR2 (M537I), whereas pemigatinib, infigratinib, and futibatinib showed inhibitory activities like those against FGFR2 wild-type. However, the inhibitory activity of tasurgratinib was not attenuated against FGFR2 (N549H/K), which is resistant to pemigatinib, infigratinib, and futibatinib, as previously reported in other preclinical studies (43-45).
Cell-based inhibitory activities of FGFR inhibitors against several secondary FGFR2 resistance mutations. (A) The IC50 values of FGFR inhibitors against FGFR variants were evaluated using the NanoBRET assay system. (B) Dose-dependent inhibitory activities of FGFR inhibitors on FGFR2 wild-type, FGFR2 N549H, and FGFR2 N549K. Data are presented as means of triplicate measurements±SEM. CI: Confidence interval; IC50: 50% inhibitory concentration; NE: not estimable.
Discussion
We confirmed that tasurgratinib has antitumor activity in preclinical tumor models with the FGFR2-fusion gene. In in vitro studies, tasurgratinib inhibited the anchorage-independent colony formation ability of NIH/3T3 cells expressing FGFR2-AHCYL1, FGFR2-BICC1 type1, FGFR2-BICC1 type2, FGFR2-TXLNA, and FGFR2-KCTD1, which are found in CCA (13). Tasurgratinib showed antitumor activity in allografts of NIH/3T3 expressing the FGFR2-fusion gene. We also used CCA PDX models to reinforce the preclinical evidence of tasurgratinib because these models are supposed to reflect the genetic, histopathological, and phenotypic characteristics of patients with CCA (46). Western blot analyses were performed to elucidate the molecular mechanism of action of tasurgratinib against preclinical CCA models with the FGFR2-fusion gene. These data suggested that the inhibitory activities of tasurgratinib on the FGFR signaling pathway led to antitumor activities in vitro and in vivo. Interestingly, nebivolol, a third generation β1-blocker was reported to synergistically inhibit the FGFR downstream pathway when combined with selective FGFR inhibitors in breast cancer cells (47). Progress in the study including the combination activity of tasurgratinib and nebivolol in preclinical FGFR2-driven CCA models might thus help to fulfil the unmet medical needs of patients with this type of cancer.
Pemigatinib, infigratinib, and futibatinib are approved for previously treated, unresectable, locally advanced, or metastatic CCA with FGFR2 fusions or rearrangements. Several acquired resistance FGFR2 mutations have been reported to occur after treatment with FGFR inhibitors in patients with CCA (41). Among them, FGFR2 (N549H/K) is one of the major mutations reported to occur after treatment with pemigatinib, infigratinib, and futibatinib (41). In concordance with clinical observations, the inhibitory activities of pemigatinib, infigratinib, and futibatinib against FGFR2 (N549H/K) were attenuated compared to that against FGFR2 wild-type in preclinical experiments (43-45). Here, we reproduced these observations, while the inhibitory activities of tasurgratinib against FGFR2 (N549H/K) were retained. We also confirmed that tasurgratinib has fast association and slow dissociation against FGFRs, binds to the ATP-binding site and the neighboring region and adopts an DFG-“in” conformation, which is determined as the type V binding mode. In contrast, most kinase inhibitors (including pemigatinib and infigratinib) only bind to the ATP-binding site, which is known as the type I binding mode (48). The binding mode of tasurgratinib is also different from that of futibatinib, which forms a covalent adduct with the cysteine side chain of FGFRs (49). Type V kinase inhibitors are suggested to have high affinity for target proteins and selectivity for a small number of receptor tyrosine kinases based on the kinetic interaction analysis of VEGFR2 and its inhibitors (32). Furthermore, tasurgratinib has a basic structure lacking the dimethoxyphenyl moiety, which is distinct from other known FGFR inhibitors (19). These features of tasurgratinib may lead to a unique inhibitory profile against mutated FGFRs. Further analyses using ctDNAs collected from patients after treatment with tasurgratinib and other FGFR inhibitors are warranted to determine the precise mechanism of resistance. Additionally, co-crystal structure and modeling analyses of each FGFR inhibitor used to acquire resistance FGFR2 mutations should elucidate the molecular mechanism more precisely.
In conclusion, we demonstrated the antitumor activity of tasurgratinib in a preclinical model of murine cells expressing the FGFR2-fusion gene. Furthermore, tasurgratinib showed antitumor activity in a CCA PDX model harboring the FGFR2-BICC1 gene. Our findings provide preclinical evidence for tasurgratinib treatment in patients with CCA harboring the FGFR2-fusion gene. A pivotal phase II trial of tasurgratinib in patients with FGFR2-fusion gene-positive CCA is ongoing (NCT04238715) (18).
Acknowledgements
The Authors gratefully acknowledge Proteros Biostructures GmbH, Carna Biosciences Inc., Crown Bioscience Inc., and Sunplanet Co., Ltd. for their excellent technical support during the experiments. The Authors also thank their colleagues at Eisai Co., Ltd. for their fruitful discussions and constructive comments and Editage (www.editage.com) for English language editing.
Footnotes
Authors’ Contributions
S.K., Y.A., T.S., and S.W.M. conceptualized the study. S.K., M.I.K., S.F., Y.A., and S.W.M. developed the methodology and conducted experiments. S.K., M.I.K., and S.W.M. contributed to data visualization. S.K. and S.W.M. prepared the manuscript. All Authors have verified and approved the final version of the manuscript.
Conflicts of Interest
S.K., M.I.K., S.F., and S.W.M. were employees of Eisai Co., Ltd. At the time the study was performed. No disclosures are reported by the other Authors.
Funding
This study was funded by Eisai Co. Ltd.
- Received March 29, 2024.
- Revision received April 12, 2024.
- Accepted April 15, 2024.
- Copyright © 2024 International Institute of Anticancer Research (Dr. George J. Delinasios), All rights reserved.
This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY-NC-ND) 4.0 international license (https://creativecommons.org/licenses/by-nc-nd/4.0).
References
In this issue
Jump to section
Related Articles
Cited By...
- No citing articles found.